Govpub C13

A 11 ID 2 t> 23B7
NBS
PUBLICATIONS
NATL
All 102623870
NBS IR 86-3364
*°*eav o*
U.S. DEPARTMENT OF COMMERCE/National Bureau of Standards
Wind and Seismic

Effects
Proceedings of the
Seventeenth Joint UJNR
Panel Conference
QC
100
. U56
86-3364
1986
TMm
he National Bureau of Standards' was established by an act of Congress on March 3, 1901 The
Bureau’s overall goal is to strengthen and advance the nation’s science and technology and facilitate
their effective application for public benefit. To this end, the Bureau conducts research and provides: (l)a
basis for the nation’s physical measurement system, (2) scientific and technological services for industry and
government, (3) a technical basis for equity in trade, and (4) technical services to promote public safetv
The Bureau’s technical work is performed by the National Measurement Laboratory, the National
Engineering Laboratory, the Institute for Computer Sciences and Technology, and the Institute tor Materials
Science and Engineering
The National Measurement Laboratory

Provides the national system of physical and chemical measurement; • Basic Standards'
coordinates the system with measurement systems of other nations and • Radiation Research
furnishes essentiaJ services leading to accurate and uniform physical and • Chemical Physics
chemical measurement throughout the Nation’s scientific community, in- • Analytical Chemistry
dustry, and commerce; provides advisory and research services to other
Government agencies; conducts physical and chemical research; develops,
produces, and distributes Standard Reference Materials; and provides
calibration services. The Laboratory consists of the following centers:
The National Engineering Laboratory

Provides technology and technical services to the public and private sectors to • Applied Mathematics
address national needs and to solve national problems; conducts research in • Electronics and Electrical
engineering and applied science in support of these efforts; builds and main- Engineering*’
tains competence in the necessary disciplines required to carry out this • Manufacturing Engineering
research and technical service; develops engineering data and measurement • Building Technology
capabilities; provides engineering measurement traceability services; develops • Fire Research
test methods and proposes engineering standards and code changes; develops • Chemical Engineering*'
and proposes new engineering practices; and develops and improves
mechanisms to transfer results of its research to the ultimate user. The
Laboratory consists of the following centers:
The Institute for Computer Sciences and Technology

Conducts research and provides scientific and technical services to aid • Programming Science and
Federal agencies in the selection, acquisition, application, and use of com- Technology
puter technology to improve effectiveness and economy in Government • Computer Systems
operations in accordance with Public Law 89-306 (40 U.S.C. 759), relevant Engineering
Executive Orders, and other directives; carries out this mission by managing
the Federal Information Processing Standards Program, developing Federal
ADP standards guidelines, and managing Federal participation in ADP
voluntary standardization activities; provides scientific and technological ad
visory services and assistance to Federal agencies; and provides the technical
foundation for computer-related policies of the Federal Government. The In
stitute consists of the following centers:
The Institute for Materials Science and Engineering

Conducts research and provides measurements, data, standards, reference • Ceramics
materials, quantitative understanding and other technical information funda- • Fracture and Deformation
mental to the processing, structure, properties and performance of materials; • Polymers
addresses the scientific basis for new advanced materials technologies; plans • Metallurgy
research around cross-country scientific themes such as nondestructive • Reactor Radiation
evaluation and phase diagram development; oversees Bureau-wide technical
programs in nuclear reactor radiation research and nondestructive evalua-
tion; and broadly disseminates generic technical information resulting from
its programs. The Institute consists of the following Divisions:
'Headquarters and Laboratories at Gaithersburg, MD, unless otherwise noied; mailing .iddu
Gaithersburg, MD
20899.
’Some divisions within the center are located at Boulder, CO 80303.
’Located at Boulder, CO, with some elements at Gaithersburg, MD.
Wind and Seismic Effects
PROCEEDINGS OF THE 17th JOINT

PANEL MEETING OF THE U.S.-JAPAN
COOPERATIVE PROGRAM IN WIND AND SEISMIC EFFECTS
Noel J. Raufaste, EDITOR

Center for Building Technology
National Engineering Laboratory
National Bureau of Standards
Gaithersburg, MD 20899
U.S. DEPARTMENT OF COMMERCE, Malcolm Baldrige, Secretary

NATIONAL BUREAU OF STANDARDS, Ernest Ambler, Director
Issued May 1986

PREFACE
The U.S. -Japan Cooperative Program in Natural Resources (UJNR) was
established in 1964 for the exchange of technical information and
experience mutually beneficial to the economics and welfare of both
countries. The Panel on Wind and Seismic Effects held its first
joint meeting in 1969 in Tokyo, Japan. The joint meeting has been
held annually ever since, alternating in Japan the odd numbered
years and the U.S. in the even numbered years.
The Seventeenth Joint Meeting was held in Tsukuba, Japan May 21-24,
1985. Under five themes, forty-seven technical papers were presented
for discussion. Ten Panel task committees held their meetings during
this period.
These proceedings include the program of the Seventeenth Joint Meeting,
the Panel resolutions, and the technical papers. Acknowledgement is
extended to those U.S. Panel members who helped edit for clarity the
technical papers.
Preparation of the proceedings was partially supported by funds from
the U.S. Federal Agencies providing members to the U.S. Panel.
Noel J. Raufaste, Secretary

U.S. Panel on Wind and
Seismic Effects
SI CONVERSION UNITS
In view of the present accepted practice for wind and seismic technolog, common uni is
of measurements were used through this publication. In recognition of the position
of the United States as a signatory to the General Conference on Weights and Measures,
which gave official status to the International System of Units (SI) in I960, the
table below is presented to facilitate conversion to SI Units. Readers int<*res:ei i
making further use of the coherent system of SI units are referred to NBS SR l' 1 1
,
l
1
'
Edition, The International System of Units; and ASTM Standard for Metric l> r.n tu*-.
TABLE OF CONVERSION FACTORS TO SI UNITS
International Conversion
Customary Units (SI) Unit Approximate
Length inch ( in) meter (m) a 1 in - 0.0254 m*
foot (ft) meter (m) 1 ft 0.3048 m
Force pound (lbf) newton (N) 1 lbf - 4.48 N
kilogram (kgf) newton (N) 1 kgf 9.807 N
Pressure pound per square newton/per square

or Stress
inch (psi) meter 1 psi - 6895 H/m 2
kip per square newton per square
inch (ksi) meter 1 ksi 689SxlO^N/«*
Energy inch’pound (in-lbf) joule (J) 1 In* lbf - 0.1130 J
foot ’pound (ft-lbf) joule (J) i ft’ lbf - 1 .3558 1
Torque pound inch

• newton-meter (N'm) 1 lbf *ln • 0.1130 N*a
or
Bending pound'foot (lbf-ft) newton-meter (N*m) 1 1 bf • ft » 1.3558 N *•>
Moment
Weight pound (lb) kilogram (kg) 1 lb - 0.4536 kg
Unit Weight pound per cubic foot kilogram per cubic pcf • ^
1 |fi ,Olft k £ / n
(pcf meter (kg/m^)
Velocity foot per second meter per second If t/fl 0 . 3048 fli/ii
( ft/sec) (m/s)
Acceleration foot per second meter per second 1 ft/#’ “ 0.3048 m/n
square (ft/s 2 ) square (m/s 2 )
a Meter may be subdivided. A centimeter (cm) is 1/100 n and a ml 111 -, , •

,
1/1000 m.
*
Exactly
IV
ABSTRACT
The Seventeenth Joint Meeting of the U.S. -Japan Panel on Wind and Seismic
Effects was held in Tsukuba, Japan from May 21 through 24, 1985. This
publication, the proceedings of the Joint Meeting, includes the program,
list of members, formal resolutions, and technical papers.
Papers were presented under five themes:
Theme I WIND ENGINEERING

Theme II - EARTHQUAKE ENGINEERING
Theme III - STORM SURGE AND TSUNAMIS
Theme IV - U.S. -JAP AN COOPERATIVE RESEARCH PROGRAM
Theme V - REPORTS ON THE NIHONKAI-CHUBU EARTHQUAKE
AND THE NAGANOKEN-SEIBU EARTHQUAKE
Subjects covered in the papers presented include:
(1) characteristics of strong winds;

(2) wind loads on structures and design criteria;
(3) earthquake ground motions;
(4) soil liquefaction studies and methods to improve
liquefaction resistance;
(5) seismic loads on structures and design criteria;
(6) stress analyses of pipelines during earthquakes;
(7) full-scale seismic experiments;
(8) earthquake hazard reduction program;
(9) use of the microcomputer for earthquake studies;
(10) quantitative evaluation of damages caused by winds
and earthquakes; and,
(11) tsunami research projects.
KEYWORDS: Accelerograph; codes; design criteria; disaster; earthquakes;

earthquake hazards; geotechnical engineering; ground failures;
liquefaction; pipeline; seismicity; solids; standards; struc-
tural engineering; structural response; tsunami; wind loads;
and winds.
V
CONTENTS
PAGE
PREFACE iii
SI CONVERSION iv
ABSTRACT v
CONTENTS vii
SCHEDULE OF THE 17 th JOINT MEETING xi
PROGRAM OF THE 17th JOINT MEETING xii
OPENING SESSION xvi
LIST OF MEMBERS xvii
LIST OF TASK COMMITTEE MEMBERS xxii
LIST OF U.S. PARTICIPANTS xxiii
FORMAL RESOLUTIONS xxiv
THEMES AND TECHNICAL PAPERS 1
Theme I: WIND ENGINEERING

Interpretation of Section Model Aeroelastic Results in the Case of
Cable-Stayed Bridges 3
R. H. Scanlan, A. M. Abdel-Ghaf f ar
Characteristics of Aerodynamic Forces on Bridge Decks in Smooth Flow

and in Turbulent Flow 15
N. Narita, H. Tada, K. Yokoyama and H. Sato
Field Measurements on the Deer Isle-Sedgwick Suspension Bridge 25

H. R. Bosch, H. S. Lew
Wind Tunnel Modeling for Flow Around Buildings 48

T. Murota, T. Okuma and H. Okada
Automatic Detection of Gust Fronts with a Single Doppler Radar 58

H. Uyeda, Dusan S. Zrnic' and H. Takahashi
Theme II: EARTHQUAKE ENGINEERING

Earthquake Hazard Reduction Program of the Department of State 70
P E Gurvin
. .
Strong Motion Instrumentation in Bureau of Reclamation Program 83

R. B. MacDonald and A. Viksne
Array Observac ion of Strong-Motion Earthquakes in the Fuchu Area 98

K. Ohtani S. Kinoshita and T. Mikoshiba
,
Implications of Recent Advances in Instrumentation for Strong-Motion

Studies 106
R. D. Borcherdt
VII
CONTENTS
PAGE
Dynamic Behaviours of a Rockfill Dam During Earthquakes 118

N. Matsumoto, N. Yasuka and M. Shiga
Vibration Test of Richard B. Russell Concrete Dam After Reservoir

Impoundment 132
R. J. Smith
Earthquake Resistant Bridge Design in California 146

J. H. Gates
Seismic Analysis, Instrumentation and Dynamic Testing of Cable-Supported

Bridges 154
A. M. Abdel-Ghaf far
—
Earthquake Resistant Design of Building Foundation Introduction and
Commentary on "Design Guide for the Building Foundation against
Seismic Force" 168
Y. Sugimura
Shear Resistance of Unreinforced Hollow Concrete Block Masonry Walls 202

K. Woodward and F. Rankin
An Experimental Study on Sliding Rigid Body in Water During Earthquake 215

T. Uwabe N. Higaki and S. Noda
,
A Rapid Seismic Analysis Procedure for Anchored Sheet Pile Bulkheads 233
T. K. Lew
Design of Piles for Slopes 259

H. Tada, K. Ohshima, Y. Sato and R. Abo
Dynamic Stresses and Displacements in a Buried Tunnel 27 3

S. K. Datta, K. C. Wong and A. H. Shah
Theme III: STORM SURGE AND TSUNAMIS

Damages of Coastal and River Structures Caused by the Nihonkai-chubu
Earthquake Tsunami 299
T. Uda and A. Omata
Progress in Developing Tsunami Data Base q ] 4

J. F. Lander and P. A. Lockridge
Tsunamis in the Japan Sea with Numerical Calculation 328

Y. Tsuji, T. Konishi and H. Takahashi
Theme IV: U.S.-JAPAN COOPERATIVE RESEARCH PROGRAM

Recommended U.S. -Japan Joint Research on Seismic Behavior of Buried
Pipelines and Telecommunications Systems 344
M. Shinozuka
Experimental Investigation on Hysteretic Behavior of Reinforced Concrete

Bridge Pier Columns v< 7
T. Iwasaki, K. Kawashima, R. Hagiwara, K. Hasegawa, T. Koyama
and T. Yoshida
VII
CONTENTS
PAGE
Re-evaluation of Soil Liquefaction Resistance Curves 382

R. M. Chung, H. B. Seed, A. G. Franklin, and J. A. Farrar
In-situ Test for Assessing Liquefaction Potential Using Vibratory Cone

Penetrometer 396
Y. Sasaki, Y. Koga Y. Itoh, T. Shimazu and M. Kondoh
,
U.S. -Japan Joint Earthquake Research Program Involving Large-Scale Experi-

—
ments on Masonry Building Structures Japanese Side Research Plan 410
S. Okamoto, Y. Yamazaki, A. Baba, M. Teshigawara and H. Hiraishi
Inelastic Seismic Tests on a Full-Scale Six-story Eccentric-K Braced

—
Steel Building U.S. /Japan Cooperative Earthquake Research Program 422
H. Yamanouchi, M. Midorikawa, I. Nishiyama and M. Hirosawa
Seismic Performance of Existing Buildings —A Progress Report 435

R. N. White and P. Gergely
Theme V: REPORTS ON THE NIHONKAI-CHUBU EARTHQUAKE AND

THE NAGANOKEN-SEIBU EARTHQUAKE
Study on Economical Losses Caused by Change of Cargo Flow at Port of
Akita Due to the Nihonkai-chubu Earthquake, 1983 459
A. Yonezawa and H. Tsuchida
Landslides and Damages Due to the Naganoken-seibu Earthquake of

September 14, 1984 471
M. Tominaga, Y. Sasaki, K. Senoo,
H. Yoshimatsu, E. Taniguchi
K. Nakano, N. Obinata, Y Hara, and Y. Kuwabara
Geomorphological Characteristics and Factors of the Large-Scale Landslide

and Detritus Flow on the South Slope of Ontake Volcano Caused by the
Naganoken-seibu Earthquake, 1984 490
M. Nagaoka, Y. Hoshino, Y. Takei, S. Ichikawa and F. Saitoh
Seismological Studies on the Naganoken-seibu Earthquake, 1984, based on

JMA Seismic Network 503
M. Takeo, N. Mikami and M. Ichikawa
PAPERS NOT PRESENTED ORALLY BUT INCLUDED IN THE PROCEEDINGS

Features of Ocean Wind Fluctuations During Typhoon Passages 518
G. Naito and H. Takahashi
The Variation of Gust Structure with Height up to 200 m 532

K. Naito, T. Hanabusa and T. Fujitani
Seismic Vulnerability of Lifeline System Using Computervision CADDS Color

Graphics Software 547
C. V. Chelapati and S. K. Takahashi
Vertical Distribution of Subsurface Horizontal Earthquake Accelerations 559

P. C Chen
.
Shaking Table Test on the Effects of Thickness of Liquefied Layer on Deformation

of Embankments 569
T. Iwasaki, Y. Sasaki, H. Matsumoto and M. Kondoh
IX
CONTENTS
PAGE
Study on Aseismic Measures to Embankments on Soft Clav Ground 580

Y. Koga and 0. Matsuo
Draft of Guidelines for Post-Earthquake Inspection and Evaluation of Earthquake

Damage in Reinforced Concrete Buildings and Its Application to the Namioka
Town Hospital Ruildlng 593
T. Okada, M. Hirosawa, H. Hiraishi, and M. Yoshimura
Dense Strong Motion Earthquake Seismometer Array at Site with Different

Topographic and Geologic Condition in Sendai 618
Y. Kitagawa, T. Kashima, and M. Hirosawa
Influence of Soil Conditions Upon Response Spectrum of Strong Ground Motion 631
I. Ohkawa and Y. Ishiyama
Regional Tsunami Warning System (THRUST) 644

E. N. Bernard and R. R. Behn
Studies on Storm Surges in Tidal Estuaries 650

T. Konishi, T. Kinoshita, and H. Takahashi
A SDOF Model to Analyze Earthquake Response of a Full-scale Seven-story

Reinforced Concrete Building Structure 662
M. Yoshimura; S. Okamoto, and V. Sigmund
Settlement of Approaching Embankments to Bridges Caused by the Nihonkai-chubu

Earthquake 679
Y. Koga and E. Taniguchi
Rock Avalanches and Slope Failures Induced by the Naganoken-seibu Earthquake,

1984 688
K. Tanaka, N. Oyagi , T. Inokuchi, S. Uehara and H. Takahashi
APPENDIX: TEN TASK COMMITTEE REPORTS 701
X
SCHEDULE OF THE 17th JOINT MEETING
May 19 (Sun) U.S. members arrive at Narita; other U.S. members who attended
Task Committee D join in Tsukuba
20 (Mon) Visit Tsukuba EXPO '85
21 (Tue) Opening Session

Technical Sessions
Reception (Hosted by the Engineer General of the
Ministry Of Construction)
22 (Wed) Technical Sessions

Task Committees
23 (Thu) Technical Sessions

Task Committees
24 (Fri) Technical Sessions

Reports of Task Committees
Adoption of Final Resolution
Closing Session
25 (Sat) Tsukuba-Tokyo
26 (Sun) Tokyo-Matsumoto-Kamikochi ,
Nagano Pref.
27 (Mon) Kamikochi — (Kisofukushima

29 (Wed) Kisofukushima — Nakatsugawa I.C. — Nagoya — Kobe
30 (Thu) Kobe —Naruto (Honshu-Shikoku Bridges) — Kobe (Port Island,

Harbor)
31 (Fri) Kobe — Kyoto
June 1 (Sat) Kyoto — Tokyo
2 (Sun) U.S. members leave Japan
XI
PROGRAM OF THE SEVENTEENTH JOINT MEETING
Tuesday - May 21
Opening Session 10:30 - 12:00
Call to order by Dr. Nobuyuki Narita, Secretary-General, Japan Panel
Remarks by: Tetsuo Kutsukake, G. H. Hlfrich, Hirofumi Satake, Masateru

Tominaga, and R. N. Wright
Technical Session 13:00 - 17:20

Session 1: Theme 1 - WIND ENGINEERING Chairman: M. Tominaga
13:00--13:20 Interpretation of Section Model Aeroelastic Results in

the case of Cable-Stayed Bridges
R. H. Scanlan (A. M. Abdel-Ghaf f ar)
13:20--13:40 Characteristics of Aerodynamic Forces on Bridge Decks in

Smooth Flow and in Turbulent Flow
N. Narita, H. Tada, K. Yokoyama and H. Sato
13:40--14:00 Field Measurements on the Deer Isle-Sedgwick Suspension

Bridge
H. R. Bosch (H. S. Lew)
14:00--14:20 Discussion
Session 2: Theme I - WIND ENGINEERING Chairman: M. Tominaga
14:20--14:40 Wind Tunnel Modeling for Flow Around Buildings

T. Murota, T. Okuma and H. Okada
14:40--15:00 Automatic Detection of Gust Fronts with a Single Doppler

Radar
H. Ueda, Dusan S. Zrnic and H. Takahashi
15:00--15:20 Discussion
15:20--15 :40 Break
Session 3: Theme IV - U.S.-JAPAN COOPERATIVE RESEARCH PROGRAM

Chairman: M. Tominaga
15:40--16:00 Recommended U.S. -Japan Joint Research on Seismii Behavior

of Buried Pipelines and Telecommunications Systems
M. Shinozuka
16 GO--16:20
: Experimental Investigation on Hysteretic Behavior of
Reinforced Concrete Bridge Pier Columns
T. Iwasaki, K. Kawashima, R. Hagiwara, K. Hasegawa,
T. Koyama and T. Yoshida
16: 20--16:40 Re-Evaluation of Liquefaction Resistance Curves

R. M. Chung, H. B. Seed, A. G. Franklin and J. A. Farrar
lb :40--17:00 In-situ Test for Assessing the Liquefaction Potential

Using Vibratory Cone Penetrometer
Y. Sasaki, Y. Koga, Y. Itoh, T. Shimazu and M. Kondoh
17 :00--17:20 Discussion
18:30- Reception
XII
Wednesday - May 22
Session 4: Theme IV - U.S.-JAPAN COOPERATIVE RESEARCH PROGRAM
Chairman: R. N. Wright
9:00-9:20 U.S. -Japan Joint Earthquake Research Program Involving Large-

Scale Experiments on Masonry Structures —
Japanese Side
Research Plan —
S. Okamoso, Y. Yamazaki, A. Baba, M. Teshigawara and H. Hiraishi
9:20-9:40 Inelastic Seismic Tests on a Full-Scale Six-story Eccentric-K

Braced Steel Building
— U.S. /Japan Cooperative Earthquake Research Program —
H. Yamanouchi, M. Midorikawa, I. Nishiyama and M. Hirosawa
9:40-10:00 A Summary Report of the U.S. -Japan Workshop on Evaluation of

Structural Performance
R. N. White and G. R. Fuller
10:00-10:20 Discussion
10:20-10:40 Break
Session 5: Theme II: EARTHQUAKE ENGINEERING Chairman: R. N. Wright
10:40-11:00 Earthquake Hazard Reduction Program of the Department of State

P. E. Gurvin
11:00-11:20 Strong Motion Instrumentation in Bureau of Reclamation Program

R. B. MacDonald and A. Viksne
11:20-11:40 Array Observationof Strong-Motion Earthquake in Fuchu Area

S. Kinoshita, K. Ohtani and T. Mikoshiba
11:40-12:00 Implications of Recent Advances in Instrumentation for Strong-

Motion Studies
R. D. Borcherdt
12:20-13:20 Lunch

Session 6: Theme II - EARTHQUAKE ENGINEERING Chairman: R. N. Wright
13:20-13:40 Dynamic Behaviours of a Rockfill Dam During Earthquakes

N. Matsumoto, N. Yasuka and M. Shiga
13:40-14:00 Vibration Test of Richard B. Russell Concrete Dam After

Reservoir Impoundment
R.J. Smith
14:20-14:30 Break
14:30-17 :00 Task Committee Meetings

B: Large-Scale Testing Programs
C: Repair and Retrofit of Existing Structures
E: Land Use Program for Controlling Natural Hazard Effects
H: Soil Behavior and Stability During Earthquakes
I: Storm Surge and Tsunamis
18:00- Reception
XIII
Thursday - May 23
Session 7: Theme II - EARTHQUAKE ENGINEERING Chairman: M. Tominaga
9:00-9:20 Earthquake Resistant Bridge Design in California

J. H. Gates
9:20-9:40 Seismic Analysis, Instrumentation and Dynamic Testing of

Cable-Supported Bridges
A. M. Abdel-Ghaf f ar
9:40-10:00 Earthquake Resistant Design of Building Foundation

— Introduction and Commentary on "Design Guide for the
Building Foundation Against Seismic Force"
Y. Sugimura
10:00-10:20 Shear'Resistance of Unreinforced Hollow Concrete Block

Masonry Walls
K. Woodward and F. Rankin
i0:40-ll:00 Break
Session 8: Theme II - EARTHQUAKE ENGINEERING Chairman: M. Tominaga
11:00-11:20 An Experimental Study on Sliding Rigid Body in Water During

Earthquake
T. Uwabe ,
N. Higaki and S. Noda
11:20-11:40 A Rapid Seismic Analysis Procedure for Anchored Sheet Pile

Bulkheads
T. K. Lew
11:40-12:00 Design of Piles for Slopes

H. Tada, K. Ohshima, Y. Sato and R. Abo
12:00-12:20 Dynamic Stresses and Displacements in a Buried Tunnel

S. K. Datta, K. C. Wong and A. H. Shah
12:40-13:40 Lunch

Session 9 Chairman: M. Tominaga
13:40-14:00 A Short Slide Presentation on Survey of the Recent Chilean

Earthquake
J. H. Gates
14:00-14:20 A Short Slide Presentation on Survey of the Recent Chilean

Earthquake
H. Tsuchida
14:20-14:30 Break
Task Committee Meetings 14:30 - 17:20

A: Strong-Motion Instrumentation Arrays and Data
D: Evaluation of Performance of Structures
F: Disaster Prevention Methods for Lifeline Systems
G: Wind Characteristics and Structural Response
J: Wind and Earthquake Engineering for Transportation
Systems
XIV
Friday - May 24
Session 10: Theme III - STORM SURGE AND TSUNAMIS Chairman: R. N. Wright
9:00-9:20 Damages of Coastal and River structures Caused by the Nihonkai-

chubu Earthquake Tsunami
T. Uda and A. Omata
9:20-9:40 Progress in Developing Tsunami Data Base

J. F. Lander and P. A. Lockridge
9:40-10:00 Tsunamis in the Japan Sea with Numerical Calculation

Y. Tsuji and H. Takahashi
10:20-10:40 Break
Session II: Theme V - REPORTS ON THE NIHONKAI-CHUBU EARTHQUAKE AND THE

NAGANOKEN-SEIBU EARTHQUAKE Chairman: R. N. Wright
10:40-11:00 Study on Economical Losses Caused by Change of Cargo Flow at

Port of Akita Due to the Nihonkai-chubu Earthquake, 1983
A. Yonezawa and H. Tsuchida
11:00-11:20 Landslides and Damages Due to the Naganoken-seibu Earthquake

of September 14,1984
M. Tominaga, Y. Sasaki, K. Senoo, H. Yoshimatsu, E. Taniguchi,
M. Nakano, N. Obinata, Y Hars and T. Kuwabara
11:20-11:40 Geomorphogical Characteristics and Factors of the Large-Scale

Landslide and Detritus Flow on the South Slope of Ontake Volcano
Caused by the Naganoken-seibu Earthquake, 1984
M. Nagaoka, Y. Hoshino, Y. Takei, S. Ichikawa and H. Saito
11:40-12:00 Seismological Studies on the Naganoken-seibu Earthquake, 1984,

based on JMA Seismic Network
M. Takeo, N. Mikami and M. Ichikawa
Reports of Task Committees 13:20 - 15:00 Chairman: M. Tominaga

A: Strong-Motion Instrumentation Arrays and Date
B: Large-Scale Testing Programs
C: Repair and Retrofit of Existing Structures
D: Evaluation of Performance of Structures
E: Land Use Program for Controlling Natural Hazard Effects
F: Disaster Prevention Methods for Lifeline Systems
G: Wind Characteristics and Structural Response
H: Soil Behavior and Stability During Earthquakes
I: Storm Surge and Tsunamis
J: Wind and Earthquake Engineering for Transportation Systems
15:00-15:10 Break
15:10-15:40 Adoption of Final Resolution Chairman: M. Tominaga
15:40- Break
16:00-16:20 Closing Session

Closure by Dr. Nobuyuki Narita
XV
OPENING SESSION
Tuesday - May 21, 1985 (10:30-12:00)
Conference Hall (8th Floor), Public Works Research Institute
Call to order by Dr. Nobuyuki Narita, Secretary-General, Japan Panel

Assistant Director-General, Public Works Research Institute
Remarks by Mr. G. H. Helfrich, Counselor for Scientific and Technological

Affairs, Embassy of the United States of America
Remarks by Mr. Hirofumi Satake, Director, International Affairs Division

Promotion Bureau, Science and Technology Agency
Remarks by Dr. Masateru Tominaga Chairman, Japan Panel, Director-General

,
Public Works Research Institute, Ministry of Construction
Remarks by Dr. Richard N. Wright, Chairman, TJ. S. Panel, Director,

Center for Building Technology, NEL National Bureau of Standards,
,
Department of Commerce
Introduction of Japan Panel Members by Japanese Chairman and U.S. Panel

Members by U. S. Chairman
Election of Conference Chairman
Adoption of Agenda
Adj ourn
Group Photograph
XVI
U.S. PANEL MEMBERS-WIND AND SEISMIC EFFECTS
1985
Dr. Richard N. Wright, Chairman Dr. John J. Burns, Jr.

Director, Center for Building Technology Head, Structures Section
National Bureau of Standards Office of Research
Gaithersburg, MD 20899 U.S. Nuclear Regulatory
(301) 921-3377 Commission
Washington, DC 20555
Dr. H. S. Lew, Secretariat (301) 443-7860
Leader, Construction Safety Group
Structures Division Dr. James Choromokos ,
Jr.
Center for Building Technology Director
National Bureau of Standards Research and Development
Gaithersburg, MD 20899 Department of the Army
(301) 921-2647 DAEN-RDZ-A
Dr. S. T. Algermissen (202) 272-0254
Office of Earthquake Studies
Branch of Earthquake Tactonics, TJSGS Dr. Charles G. Culver
Denver Federal Center Chief, Structures Division
Stop 978, Box 25046 National Bureau of Standards
Denver, CO 80225 Gaithersburg, MD 20899
(303) 236-1611 (301) 921-2196
Dr. Celso S. Barrientos Dr. A. G. Franklin

Supervisory Physical Scientist Chief, Earthquake Engineering
noaa/nesdis and Geophysics Division
3300 Whitehaven Street, NW Geotechnical Laboratory
Washington, DC 20235 U.S. Army Engineer Waterways
(202) 634-7379 Experiment Station
Vicksburg, MS 39180
Dr. Eddie N. Bernard (601) 634-2658
Director, Pacific Marine
Environmental Laboratory Mr. G. Robert Fuller
7600 San Point Way, NE Chief, Standards Branch
Seattle, WA 98115 Manufactured Housing Construction
(206) 526-6239 FTS 292-6800 Standards Division, Room 9156
Housing and Urban Development
Dr. Roger D. Borcherdt Washington, DC 20411
Chief, Branch of Engineering (202) 755-6920
Seismology and Geology
Office of Earthquake Studies, USGS Mr. James H. Gates
345 Middlefield Road Office of Structures Design
Menlo Park, CA 94025 Department of Transportation
(415) 323-8111 X2755 1120 N Street
Sacramento, CA 95814
Dr. A. Gerald Brady (916) 445-3810
Physical Scientist
Office of Earthquake Studies Dr. Riley M. Chung
U.S. Geological Survey Leader, Geotechnical Engineering
345 Middlefield Road Group, Structures Division
Menlo Park, CA 94025 National Bureau of Standards
(415) 323-8111 Gaithersburg, MD 20899
(301) 921-2647
XVII
Dr. Michael P. Gaus Mr. James F. Lander
Section Head of Critical Engineering Deputy Director
Division of Emerging and Critical National Geophysical Data Center
Engineering Systems NOAA/NESDIS
National Science Foundation 325 Broadway
1800 G Street, NW Boulder, CO 80303
Washington, DC 20550 FTS 320-6215 or (303) 497-6215
(202) 357-9500
Dr. E. V. Leyendecker
Mr. Peter E. Gurvin Leader, Earthquake Hazards Reduction
A/FBO, SA-6 Group, Structures Division
Department of State Center for BuiLdlng Technology
Washington, DC 20620 National Bureau of Standards
(202) 235-3689 Gaithersburg, MD 20899
(301)921-3471
Dr. Walter W. Hays
Deputy for Research Application Mr. Robert MacDonald
U.S. Geological Survey Geologic Service Branch
Office of Earthquake Survey Code D
905 National Center 1630 Bureau of Reclamation
Reston, VA 22043 Building 67, Denver Federal Center
FTS 928-6472 or (703) 860-6471 Denver, CO 80225
FTS 786-4195 or (303) 236-8428
Dr. William B. Joyner
Geophysicist Dr. Richard D. McConnell
Office of Earthquake Studies Office of Construction
Branch of Ground Motion and Veterans Administration
Faulting, USGS 811 Vermont Avenue, NW
342 Middlefield Road Washington, DC 20420
Menlo Park, CA 94025 (202) 389-3103
(415) 323-8111
Dr. Francis G. McLean
Mr. Richard W. Kramer Chief, Division of Research and
Civil Engineer Laboratory Services, D-1500
Technical Review Staff, D-3210 Rureau of Reclamation
Bureau of Reclamation P. 0. Box 25007
P. 0. Box 25007 Denver, CO 80225
Denver, CO 80225 FTS 776-5981 or (303) 234-3303
FTS 776-8539 or (303) 236-8539
Dr. Richard D. Marshall
Mr. Richard Krimm Structural Engineering Group
Office of Natural and Structures Division
Technological Hazards Center for Ruilding Technology
Federal Emergency Management Agency National Bureau of Standards
Washington, DC 20472 Gaithersburg, MD 20899
(202) 646-2871 (301) 921-2170
Mr. Paul K. Krumpe Mr. John F. Meehan

Office of Foreign Disaster Assistance Principal Structural Engineer
Agency for International Development Research Director
Department of State Structural Safety Section
Washington, DC 20523 Office of the State Architect
(202) 632-5916 Department of General Services
Sacramento, CA 95805
(916) 445-8730
XVIII
Dr. John B. Scalzi Alternates
Program Director, Systems
Engineering for Large Structures Mr. Michael Changery
National Science Foundation
'
National Climatic Center
1800 G Street, NW, Room 1130 Federal Building
Washington, DC 20550 Asheville, NC 28787
(202) 357-7710 (704) 258-2850 ext. 217
Mr. Joseph T. Schaefer Mr. Roger Kenneally

Chief, Techniques Development Unit Structures Section
National Severe Storm Forecast Center Office of Research
National Weather service, NOAA U.S. Nuclear Regulatory Commission
601 E. 12th Street Washington, DC 20555
Room 1728 Federal Building (301) 443-7860
Kansas City, MO 64106
(816) 374-3367 Mr. Robert R.Ledzian
Senior Staff Assistant for
Dr. Charles Scheffey Research, WO-220
Science Advisor, Offices of Research, Bureau of Reclamation
Development and Technology Department of the Interior
Federal Highway Administration 18th at C Streets, NW
6300 Georgetown Pike Washington, DC 20240
McLean, VA 22101 (202) 343-6703
(703) 285-2054
Mr. Ronald J. Morony
Mr. Robert J. Smith Program Manager
Department of the Army Building Technology Research Staff
Office of the Chief of Engineers Department of Housing and Urban
DAEN-CWE-DS Development
Washington, DC 20314 Washington, DC 20410
(202) 272-0219 (202) 755-0640
Dr. Lawrence A. Soltis

Supervisory Research Engineer
Engineering design Criteria
Forest Products Laboratory
Forest Service, Dept, of Agriculture
Madison, WI 53705
(608) 264-5910
Mr. K. Thirumalai
Earthquake Hazards Mitigation Section
National Science Foundation
1800 G Street, NW, Room 1130
(202) 357-9500
Mr. Stanley K. Takahashi

Structures Division
Naval Civil Engineering Laboratory
Port Hueraene, CA 93043
(805) 982-4684
XIX
JAPANESE PANEL MEMBERS - WIND AND SEISMIC EFFECTS
1985
Dr. Masateru Tominaga, Chairman Dr. Yuji Ishiyama

Director-General Head, Building Engineering Division
Public Works Research Institute International Institute of Seismology
Ministry of Construction and Earthquake Engineering (l.E.S.E.E.)
Asahi 1-banchi, Toyosato-machi Tsukuba-gun
,
Building Research Institute
Ibaraki-ken 305 Ministry of Construction
0298-64-2211 Tatehara 1-banchi Oh-ho-machi Tsukuba-gun
,
Ibaraki-ken 305
Dr. Nobuyuki Narita, Secretary-General 0298-64-2151
Assistant Director-General
Public Works Research Institute Dr. Toshio Iwasaki
Ministry of Construction Director, Earthquake Disaster Prevention
Asahi 1-banchi, Toyosato-machi, Tsukuba-gun Department
Ibaraki-ken 305 Public Works Research Institute
0298-64-2211 Ministry of Construction
Asahi 1-banchi Toyosato-machi Tsukuba-gun
,
Mr. Tsunekazu Fukui Deputy Secretary-General

,
Ibaraka-ken 305
Director, Planning and Research 0298-64-2211
Administration Department
Public Works Research Institute Dr. Kazuhiko Kawashima
Ministry of Construction Head, Earthquake Engineering Division
Asahi 1-banchi, Toyosato-machi, Tsukuba-gun Earthquake Disaster Prevention Department
Asahi 1-banchi ,ToyosatO-ma< hi ,TsuV uba-gui)
Dr. Yoshimi Goda Ibaraki-ken 305
Deputy Director-General 0298-64-2211
Port and Harbour Research Institute
Ministry of Transport Mr. Yasuyuki Koga
3-1-1, Nagase, Yokosuka-shi Head, Soil Dynamics Division
Kanagawa-ken 239 Construction Method and Equipment .
0468-41-5410 •
PubLic Works Research Institute

Ministry of Construction
Dr. Sadaiku Hattori Asahi 1-banchi Toyosato-machi .Tsukuba-gun
,
Director, International Institute of Ibaraki-ken 305

Seismology and Earthquake Engineering
Building Research Institute Dr. Keizo Masamura
Ministry of Construction Head, Typhoon Research Division
Tatehara 1-banchi, Oh-ho-machi Tsukuba-gun
, Meteorological Research Institute (mpi)
Ibaraki-ken 305 Japan Meteorological Agency (JMA)
0298-64-2151 1—1, Nagamine, Yatabe-machi, Tsukuba-gun
Ibaraki-ken 305
Dr. Masaya Hirosawa 0298-51-7 111
Director, Research Planning
and Information Department Dr. Norihisa Matsumoto
Building Research Institute Head, Fill Type Dam Division
Tatehara I-banchi, Oh-ho-machi, Tsukuba-gun Dam Department
Dr. Masaji Ichikawa Asahi i-banchi ,Toyosato-mai hi .TsuWub.i gun
Head, Seismology and Volcanology Division Ibaraki-ken 305
Meteorological Agency (JMA) 0298-64-2211
1-1, Nagamine, Yatabe-machi , Tsukuba-gun
Ibaraki-ken 305
0298-51-7111
XX
Mr. Tatsuo Murota Dr. Yasushi Sasaki
Head, Building Aerodynamics Division Head, Ground Vibration Division
Structural Engineering Department Earthquake Disaster Prevention Department
Building Research Institute Public Works Research Institute
Ministry of Construction Ministry of Construction
Tatehara 1-banchi Oh-ho-machi Tsukuba-gun
, ,
, ,
Ibarai-ken 305 Ibaraki-ken 305

0298-64-2151 0298-64-2211
Mr. Masatoshi Nagaoka Mr. Hiroshi Shinohara

Head, Geographical Research Office Head, Bridge Division
Geographical Research Institute Structure and Bridge Department
Ministry of Construction Public Works Research Institute
Kitazato 1-banchi, Yatabe-machi Tsukuba-gun
,
Ibaraki-ken 305 Asahi 1-banchi Toyosato-machi Tsukuba-gun
, ,
Ibaraki-ken 305
Dr. Keikichi Naito 0298-64-2211
Head, Meteorological Satellite Division
Meteorological Research Institute (MRI) Dr. Yoshihiro Sugimura
Japan Meteorological Agency (JMA) Head, Foundation Engineering Division
1-1, Nagamine Yatabe-machi, Tsukuba-gun
,
Structure Engineering Department
Ibaraki-ken 305 Building Research Institute
0298-51-7111 Ministry f Construction
Tatehara 1-banchi ,0h-ho machi Tsukuba-gun
,
Dr. Setsuo Noda Ibaraki-ken 305

Chief, Earthquake Resistant Structures 0298-64-2151
Laboratory, Structure Division
Port and Harbour Research Institute Dr. Hirohiko Tada
Ministry of Transport Director, Structure and Bridge Department
3-1-1, Nagase, Yokosuka-shi Kanagawa-ken 239
,
Public Works Research Institute
, ,
Mr. Kazuya Ohshima Ibaraki-ken 305

Head, Foundation Engineering Division 0298-64-2211
Structure and Bridge Department
Public Works Research Institute Dr. Hiroshi Takahashi
Ministry of Construction Director-General
Asahi 1-banchi, Toyosato-machi Tsukuba-gun
,
National Research Center for Disaster
Ibaraki-ken 305 Prevention Science and Technology Agency
,
0298-64-2211 3-1, Tennodai, Sakura-mura, Niihara-gun

Ibaraki-ken 305
Mr. Keiichi Ohtani 0298-51-1611
Chief, Earthquake Engineering Laboratory
Second Research Division Dr. Haime Tsuchida
National Research Center for Director, Structure Division
Disaster Prevention Port and Harbour Research Institute
Science and Technology Agency Ministry of Transport
3-1, Tennodai, Sakura-mura, Niihari-gun 3-1-1 Nagase , Yokosuka-shi Kanagawa-ken 239
, ,
Ibaraki-ken 305
0298-51-1611 Dr. Takaaki Uda
Head, Coastal Engineering Division
Mr. Shin Okamoto River Department
Diroctor, Production Department Public Works Research Institute
Building Reseach Institute Ministry of Construction
Ministry of Construction Asahi 1-banchi Toyosato-machi Tsukuba-gun
, ,
Tatehara 1-banchi Oh-ho-machi Tsukuba-gun

, ,
Ibaraki-ken 305
Ibaraki-ken 305 0298-64-2211
0298-64-2151
Mr. Koichi Yokoyama, Secretary
Head, Structure Division, Structure and
Bridge Dept., Public Works Research Institute,
Ministry of Construction, Asahi 1-banchi,
Toyosato-machi, Tsukuba-gun, Ibaraki-ken 305
XXI
MEMBERS OF TASK COMMITTEES
Task Committee 1J. S. Side Japanese Side
A. Strong-Motion Instrumentation * A.O. Brady * Haj ime Tsuchida

Arrays and Data W.B. Joyner Sadaiku Hattori
J.F. Lander Masaj i Ichikawa
F.G. McLean Yuji Ishiyama
Keitchi Ohtani
Yasushi Sasaki
B. Large-Scale Testing Programs * E.V. Leyendecker * Keiichi Ohtani
H.S. Lew Yuji Ishiyama
J.B. Scalzi Yasuyuki Koga
C.F. Scheffey Setsuo Noda
Shin Okamoto
Koichi Yokoyama
C. Repair and Retrofit of * J.B. Scalzi * Masaya Hirosawa
Existing Structures P.E. Gurvin Yuji Ishiyama
H.S. Lew Toshio Iwasaki
R.D. McConnell Shin Okamoto
C.F. Scheffey Hirohiko Tada
D. Evaluation of Performance * G.R. Fuller * Shin Okamoto

of Structures R. D. McConnell Yuji Ishiyama
J.F. Meehan Kazuya Ohshima
J.B. Scalzi Hirohiko Tada
S. K. Takahashi
E. Land Use Program for * S.T. Algermissen * Kazuhiko Kawashima

Controlling Natural A.G. Brady Sadaiku Hattori
Hazard Effects G.R. Fuller Toshio Iwasaki
R. D. Marshall Masatoshi Nagaoka
Yoshihiro Sugimura
F. Disaster Prevention Methods * K. Thirumalai * Toshio Twasaki
for Lifeline Systems E.V. Leyendecker Kazuhiko Kawashima
J.B. Scalzi Tatsuo Murota
S. K. Takahashi Setsuo Noda
Keiichi Ohtani
Yasushi Sasaki
G. Wind Characteristics and * R.D. Marshall * Keikichi Naito

Structural Response C.S. Barrientos Keizo Masamura
M.P. Gaus Tatsuo Murota
S.K. Takashashi Nobuyuki Narita
C.F. Scheffey Koichi Yokoyama
H. Soil Behavior and Stability * A.G. Franklin * Yasushi Sasaki

During Earthquakes R. Chung Yasuyuki
F.G. McLean Norihisa Matsumoto
R.J. Smith Kazuya Ohshima
Yoshihiro Sugimura
Haj ime Tsuchida
I. Storm Surge and Tsunamis * C.S. Barrientos * Yoshimi Coda
E.N. Bernard Keizo Masamura
M.p. Gaus Masaji Tchlkawa
J.F. Lander Hiroshi Takahashi
Takaaki Mda
J. Wind and Earthquake Engineering * C.F. Scheffey * Hirohiko Tada
for Transportation Systems A.G. Franklin Toshio Twasaki
H.S. Lew Kazuhiko Kawashima
E.V. Leyendecker Nobuyuki Narita
Kazuya Ohshima
Hiroshi Shinohara
Jf
Chairman Yasushi Sasaki
XXII
U.S. PARTICIPANTS
AT THE
SEVENTEENTH JOINT MEETING
1985
Dr. Richard N. Wright, Chairman Dr. Michael P. Gaus

Director Section Head of Critical Engineering
Center for Building Technology Division of Emerging and Critical
National Bureau of Standards Engineering Systems
Gaithersburg, Maryland 20899 National Science Foundation
(301)921-3377 1800 G Street, NW
Dr. H. S. Lew, Secretariat (202) 357-9500
Leader, Construction Safety Group
Structures Division Mr. Peter E. Gurvin
Center for Building Technology A/FBO, SA-6
National Bureau of Standards Department of State
Gaithersburg, Maryland 20899 Washington, DC 20520
(301) 921-2647 (202) 235-3689
Dr. Roger D. Borcherdt Mr. James F. Lander

Chief, Branch of Engineering Deputy Director, National Geophysical
Seismology and Geology Data Center, NOAA/NESDIS
Office of Earthquake Studies, USGS 325 Broadway
345 Middlefield Road Boulder, CO 80303
Menlo Park, CA 94025 FTS 776-5981 (303) 234-3303
(415) 323-8111 Ext. 2755
Mr. Robert MacDonald
Dr. A. G. Franklin Geologic Service Branch
Chief, Earthquake Engineering and Code D-1630
Geophysics Division Bureau of Reclamation
Geotechnical Laboratory Building 67, Denver Federal Center
U.S. Army Engineer Waterways Denver, CO 80225
Experiment Station FTS 776-4195 (303) 236-4195
Vicksbrug, MS 39180
(601) 634-2658 Mr. Robert J. Smith
Department of the Army
Mr. G. Robert Fuller Office of the Chief of Engineers
Chief, Standards Branch DAEN-CWE-DS
Manufactured Housing Construction Washington, DC 20314
Standards Division (202) 272-0219
Department of Housing and Urban
Development, Room 9156 Mr. Stanley K. Takahashi
Washington, DC 20410 Structures Division
(202) 755-6920 Civil Engineering Laboratory
Naval Construction Battalion Center
Mr. James H. Gates Port Huemene, CA 93043
(805) 982-4684
Office of Structures Design
Department of Transportation TEMPORARY MEMBERS
State of California
1120 N Street Dr. A. M. Abdel-Ghaf f arAssociate
,
Sacramento, CA 95814 Professor, Princeton University

(916) 445-3810
Dr. M. Shinozuka, Prof., Columbia Univ.
Dr. R. N. White, Prof., Cornell Univ.
XXIII
RESOLUTIONS OF THE SEVENTEENTH JOINT MEETING
U . S .-JAPAN PANEL ON WIND AND SEISMIC EFFECTS (UJNR)

Tsukuba Science City, Japan
May 21-24, 1985
The following resolutions are hereby adopted:
1. The Seventeenth Joint Panel Meeting provided an extremely valuable

exchange of technical information which was beneficial to both
countries. In view of the importance of cooperative programs on
the subject of wind and seismic effects, the continuation of Joint
Panel Meetings is considered essential.
2. The following activities have been carried out since the Sixteenth
Joint Meeting:
a. Several technical documents, including strong motion

records, research reports, post-earthquake damage
records, and workshop proceedings were exchanged.
b. Several workshops were held to develop joint research

activities and methodologies for wind and seismic
disaster mitigation.
c. A joint post-earthquake investigation of damage due to

the 1984 Naganoken-Seibu Earthquake was conducted.
These activities are considered by the Panel to be essential in

carrying out its objectives. Similar activities should be en-
couraged and supported by each side.
3. The Panel recognizes the accomplishments of the U.S. -Japan Coop-

erative Program on Large Scale Testing. The Panel endorses the
Implementing Arrangement for the Coordinated Research Program on
LargeScale Testing of Masonry Structures and establishment of a
Joint Technical Coordinating Committee on Masonry Research
( JTCCMAR)
4. The Panel endorses the following workshops proposed by the

Task Committees:
a. In-situ Testing Methods for Evaluation of Soil

Liquefaction Potential Task Committee (H), (August,
17-20, 1985).
b. Second U.S. -Japan Bridge Workshop Task Committee (j),

(August 19-22, 1985)
c. Repair and Retrofit of Existing Structures Task

Committee (C), (August, 1986 or May, 1987)
d. Evaluation of Structural Performance of Buildings Task

Committee (D), (May, i986)
XXIV
e. Lifeline Earthquake Engineering Task Committee (F),
(May, 1986)
f. Computational and Experimental Methods in Tsunami and Storm

Surge Task Committee (I).
Dates and locations are to be determined by the chairmen of each Task

Committee with concurrence by chairmen of the Joint Panel. Results of
each workshop shall be presented at subsequent Joint Panel Meeting.
5. The Panel recognizes the importance of continued exchange of personnel

technical information, research data and recorded data on significant
earthquakes and strong winds, and use of available facilities in both
countries. Thus, these activities should be strengthened and expanded
To facilitate the exchange of personnel, the Panel will continue to
assist such exchanges by providing official Panel sponsorship and
channels of communication for proper arrangements to be made by each
side
6. The Panel recommends that for future Joint Panel Meetings, priority
attention be given to following items:
a. Coordination and liaison between Task Committees. Task Committee

Chairmen should furnish copies of appropriate Task Committee
correspondence to the Secretariat of both sides for proper
dissemination.
b. Increased emphasis on wind engineering research and damage

investigations
c. Task Committee meetings should give attention to development

of research needs, priorities, and recommendations for future
joint research programs.
7. Recognizing that wind and seismic disaster mitigation plans benefit

disaster-prone countries worldwide, that many developing countries
suffer from natural disasters such as extreme winds, storm surge,
earthquakes and tsunamis, and that both the U.S. and Japan have been
involved in international technical cooperation programs, the Panel
will continue efforts to develop and coordinate projects which provide
aid and training to developing countries and maintain the exchange of
technical information.
8. The Eighteenth Joint Meeting of the UJNR Panel on Wind and Seismic
Effects will be in May 1986 at Washington, D.C. (USA). Specific
dates, program and itinerary will be proposed by the U.S. Panel
with concurrence by the Japanese Panel.
XXV
'
Theme I
WIND ENGINEERING
,
INTERPRETATION OF SECTION MODEL AEROELASTIC RESULTS

IN THE CASE OF CABLE-STAYED BRIDGES
By
Robert H. Scanlan*
ABSTRACT device used in this endeavor, has been the elas-

The paper focuses on the aeroelastic wind tunnel tically sprung deck section model, but the problem
sectional model of a bridge deck and close inter- recurs as to what extent such a model represents
pretation of its relation to the prototype and to the dynamics of the prototype bridge. The correct
other models. The inadequacy of such model for answer to this is that such a model cannot ever
cable-stayed bridges, where vibration modes are truly represent the action of the full bridge.
more complex than for the older suspension bridges, This is more particularly apparent for cable-
is apparent. The deserved popularity of the sec- stayed bridges, where vibration modes are more
tion model rests, however, on the important in- complex than for the older suspension types. The
sights that can be gained from it notwithstanding. deserved popularity of the section model rests,
The most important of these is the information however, on the important insights that can be
that leads to aerodynamical ly favorable modifica- gained from it notwithstanding. The most import-
tions of sectional form. Furthermore, an effort ant of these is the information that leads to
is made, in this paper, to distinguish categori- aerodynamical ly favorable modifications of sec-
cally between mechanical-dynamic and purely aero- tional form. Yet, in the process of investigation,
dynamic effects. Finally, the theory which is certain dynamical misunderstandings have contin-
outlined and applied to a cable-stayed bridge de- ued to arise, notable in failing to distinguish
sign offers means for a systematic, mode-by-mode categorically between mechanical-dynamic and pure-
investigation of bridge vulnerability to wind- ly aerodynamic effects. It proves to be especial-
induced instability. In addition, the central ly effective to identify the latter alone through
role of torsional motion in driving an instability use of the section model.
is underlined, and its aerodynamic interaction
It has been assumed, for example, that endowing
with vertical and sway effects are delineated. In
the dynamic model with the lowest bending and
conclusion, the paper emphasizes the expanded use
torsion frequencies (to scale) of the full bridge
of the flutter derivatives, particularly those
provides the model with enough freedoms and con-
related to damping, as the most fundamental analog
servatism so that its eventual response amplitudes
results obtainable from section model test.
and possible instabilities mimic the critical ones
KEY WORDS: Aerodynamic Damping, Aeroelasticity
Cable-Stayed Bridge, Flutter, Mechanical Damping, of the prototype. It has further been argued that
Mode Shapes, Section Model, Stability (Aerodynam- the model must possess not simply the mass and
ic), Wind-Tunnel (Tests), Wind Velocity (Critical)
mass moment of inertia of the chosen representa-
1. INTRODUCTION
tive deck section, but those corresponding to what
The wind tunnel has long been employed to investi- that section effectively manifests when moving in
gate the potential effect of wind on long, flex- certain particular "lowest" modes of the prototype.
ible bridges. The most common, and economical, For example, a strong argument has been made [1]
*Professor, Department of Civil Engineering, for locating and using, for the model, the correct
Princeton University, Princeton, NJ 08544
effective rotation point of the representative
. :
deck section when it is participating in the ex- based at its local c.g., namely h (vertical),
pected full-bridge torsional instability mode. a (twist about c.g.), and p (lateral sway in
the alongwind direction); see Fig. 1. If the sec-

It is the purpose of the present paper to enter
tion undergoes sinusoidal motion in the freedoms
into a more fundamental discussion of this and
noted at circular frequency to, let K = Bcu/lJ
related issues. This will be concerned with the
be a dimensionless frequency, where B is deck
two aspects: dynamics and aerodynamics .
width, U is lateral approach velocity of the

2. BEAM EXAMPLE
wind. Under these conditions, the respective
It can easily be demonstrated that a dynamic load (per unit
motion-dependent lift, drag, and moment
that acts over only a limited portion of a given air of
span) on the section are given [5,6] is
beam nonetheless excites all complete vibration
density p, by
modes of that beam. To extrapolate from this
example to bridges, an aerodynamic action confin-
L = y p U B [K H
Z
1
(K) jy + K
U -
(K) — U
+ n H, (K) a ]
(1)
ed to the bridge deck alone (or portion thereof)
i pU 2 B[K
will excite all full-bridge modes. This is a
D = P*(K)
{j
+ K P* ( K)
^ + lCP*(K)a]
( 2 )
consequence of the dynamics alone and is indepen- * *

* R *>
M ± pU 2 B 2 [K
1 h *v
dent of any special character that the aerodynam-

= Aj(K)
fj
+ K A, (10
^ K‘A.(k) «]
ics may exhibit. Nevertheless, it is a matter of (3)

* * *
experience with full-bridge models, and with a where FL(K), P^(K), A^(K) are experimentally
documented case like the Tacoma Narrows incident, determined dimensionless "flutter" coefficients
that flutter -- when it does occur -- takes place that characterize the self-excited aerodynamics
preponderantly in a single mode that closely re- of the bridge deck geometric form.
sembles an eigenmode of the structure. What cir- 2 . 2 Full Bridge Modal Dynami c s
cumstances are effective in "choosing" this mode?
The full bridge responds in (dimensionless vihra
It is probable that it is the mode susceptible of
tion modes $.(x,y,z) that have x,y,z displace
absorbing the most destabilizing energy from the
ment components at each point of the entire struc-
wind. It is also a mode in which most of the
ture. If the bridge is modeled as a set of dis-
energy is concentrated in the bridge deck. Some
crete point masses m^, then the total kinetic
light is shed on this question by bridge flutter
energy of the system is, summing over all masses:
aerodynamics
2
r* • *2 *2
t2 L)m(u
1
T =
r r
v + v
r
+
r'
w)
2 . 1 Bridge Deck Unsteady Aerodynamics r
A number of wind velocity observations [2,3,4],
where (u ,v ,w ) are the displacement 1
ment
together with recent experience of the author and r r r
of mass m .
A. Abdel-Ghaffar on the Golden Gate Bridge (to be r
reported elsewhere), have indicated that condi- Any motion u(x,y,z,t) of the strue tun
tions of very low turbulence (2 to 3%) often occur expressed as a sum of modal responses
in the wind approaching at bridge height over
u(x,y,z,t) = l (t )
^(x.y.zlB
water fetches. This argues for testing bridge i
models under laminar approach flow as a conserva-

Use of this form permits writing the alternative
tive design measure. Thus, while the buffeting
form for T
of models is an important study, it may be consid-
ered separately. The present remarks are there-

fore first confined to bridge stability under lam-
inar approach flow. where the generalized inertias I ire let im 1
f f
Let a bridge deck section at spanwise deck station , 2
1 . = B
1
x be considered that has three degrees of freedom. .
entire bridge
4
) , . .
m(x,y,z) being the continuously distributed mass The increments in h, p, and a for a virtual
per unit volume. In practice, this integral is displacement 6£. are, respectively.
usually evaluated in discrete terms. The equa- 6 h = h (x)B6q (15)

i
tions of motion of the system become
6p = pi (x)B6^ (16)
= i
1. + 2C-u.£
1 1 'i
. + w. £ ] Q ± (t) ( 8 )
6 a = a (x) 6 E,
i (17)
i
where are uncoupled from each other, and
and or, respectively, represent the damping ratio so that the net generalized aerodynamic force in
and the natural circular frequency of the i^ mode. mode i is (neglecting intermodal cross-couplings):
f
For convenience in developing the form of the
Q ± (t) [L (x,t) h (x)B +
i i
generalized force Q^, let the motion of the span
deck section at x be described by the following,
D Cx,t) P (x)B + M.(x,t)a.(x)]dx
where ln(x), p^(x), ol (x) are the (dimension- ± i ( 18 )
less) components (those portions of the total

bridge mode
only) of <fu(x,y,z)
tively in the vertical, lateral, and twist

<jr (x,y,z) that represent deck motions
at deck section x respec-

y
Qi(t)
r-2-2
P u B
= W P— *
K
C • b
G
h.h.
ii
degrees of freedom:
h(x,t) = q(t) h.(x)B (9)

iH 2 0y tS A *
K H tK )t i G h a
i 3 i
1 i
*
l
i
.
qB , bc
P(x,t) = ill 4-G

K.P (K.)
U
a.P.(L)
i 2
V
if
-iG
U p,a.
+
I (t) p i (x)B (10) p p i i
i
a(x,t) = 1
l
^(t) ai (x) (ID K
i
P *3 (K
i’
5
i
G
p.a.
1 1
4
W K
P if
U
G
h „
h.a.
l l
*
Then, for any single mode i, the value of Q^(t) C4 B

K A 2< K G
will be developed from the following expressions i P -D- a.a.
1 1
G
a a (19)
i i
for motion-related lift, drag, and moment, respec-
where the following definitions have been employed:
tively, appropriate to that mode:
f
1 2 * C-h (x)B (a) G,

L = PU BtK-H^K.) h h h^ (x)dx
i 2 + .
1 1 span
qa.(x)
WV B
+ K H
i 3
(K
i
}
^ a iW] C12)
(b)
h.a.
1 1
.
span
h
i
(x) a (x)dx
i
n * ?eP4(x)B
pU 2 B
1
D r
-
i =
2 t K .-
i
P
r K iJ
i ( ;
U
(c) G
p.p.
r
i 1 span
P i (x)dx
B |.a. (x) r
i i
K P_ (K + K.P (K.) qa.(x)] (13)
l 2
.
l
r
.
U 3 (d) P A (x) a i (x)dx
r ia1
p .
span
M
i
c
7
1
p u B
2 2
twv *
-i-i,
(
—
x) B
(e) G
r
ou (x)dx
——
a. a.
* B La.(x) *
1 1 span
K A (K u + ^(K.) ZfiW] d 4 )
(20)
i 2 i^
.th
The equation of motion for the i mode can be
where = Bar/U and total lift, drag, and
moment are rewritten
L., J Lu , J M^, respectively.
5
. . .
flutter. The latter is invariably a bluff-bodv

liUi * 2^.5. * - 0 ( 21 )
instability associated with separated flow, and

where the net effective damping ratio and it is usually driven -- even when torsion and
frequency uh are given by the expressions vertical motion are in fact coupled -- by a
(
separately unstable torsional self-excitation
co.
l
r
1
£ji
4 I;
r
V K )G
i h.h.
l l
+
WV i
a.
i
mechanism [7]. Such a flutter mechanism is radi-
cally different from classical, stiffness-coupled
Bridge structures tend to be both
+ P, (K- )G
1 1
Pi P i WVa. rl 1
flutter.
strong and massive relative to wind forces; they
~
1
are thus little deformed away from their natural
+ A, (K.)G, + A.(K.)G
1 i h.a.
ii
^
2 V a.ii
a vibration modes by wind effects. On the other
~2 hand, wind damping forces -- i.e., forces linked
0 ). 22 )
tWV
(
l
to the velocities of structural deflection
= 1 -
iV l i
a.
i
become comparable in magnitude to mechanical
P„(K.)G A-(K.)G (23) damping forces. Therein lies the apparent "para-
]
J
o l p.a.
r 3 i a. a.
i i i i dox" that was once such a mystery in cases like
Note that co. U). . However, this is not of great the Tacoma one.
l
importance to the present discussion, since the
In any single mode, = 0 signifies instabil
shift in K-value from mechanical eigen frequency
ity. Hence the flutter velocitv II for that
'
cr
to aerodynamical ly influenced characteristic fre-
mode is such that K. = K = B( . /U and, from
i cr i cr
quency is small. On the other hand, the shift in
Eq. (24)
value of to as given by Eq. (22), is of
considerable importance to the condition of flut- H* (K )G, + P* (K )G

1 cr' h.h. ,
cr p.p.
'
l
"
l 'll
ter stability.
4 C- 1.
l l
If aerodynamic cross-coupling terms associated A,(K )G ~ ( 25 )
2 cr a. a.
i
„4
* * * ii p B
with H9 ,
A ,
P.,, are provisionally neglected
defines the flutter condition.
as small,
becomes
and co.
ii - co. , the result for c.
l
In the study of the stability of a »iven i
• •
cross-section, the flutter coefficients ll. I’
,
* *
q q -
IV i
ttwq.h.
ii
A
?
are first experimentally obtained a funct ions
of k. For the particular mode i, the parameters

P, (K. )G + A 0 (K.)G
1 i p. p. 2
v
a. a. ]
J o I., G, , ,
G , and G are calculated,
l h.h. p.p. a. a.
i i i i (24) ii 'ii ii
The left hand side of expression i then
Since most flutter depends upon effective torsion-
i
(
evaluated for a series of k values. If one

al damping, the role of will be especially
such value exceeds the right-hand ide >f
important
the critical flutter velocit
2 . 3 Single-Mode Bridge Flutter
exceeded. The process permits systematii e\ a 1 .
Equation (24) emphasizes the principal aerodynamic '
ation of the stability of each of the 1 ri U;e

damping mechanisms that function in any given mode
modes successively, the one providin tin uest
of the bridge. The present paper focuses in fact,
value of U being considered the ; nt tl.
cr
on aerodynamic damping.
This procedure offers a physical rationale l
Bridge flutter is almost invariably a damping- the selection of the correct flutter • I t
related phenomenon. The classical frequency- and bridge

mode-coalescence associated with high-speed, flat-
The case of modern cable- staved hri :, >• I.
plate flutter occurs seldom, if ever, in bridge

presents new circumstances. In t \« . •, ,
•
configurations, the 3D bridge eigenmodes are typ- reference are a set of reduced velocities U/NB,
ically seen to contain, along the deck span, based on a high-wind velocity U = 160 mph.
simultaneous contributions to bending, torsion,

A first remark about the results of Table 1 is
and lateral sway. It is then no longer possible
that, while a number (8) of modes exhibit about
to categorize a given eigenmode simply as pure
90% or more of their generalized inertia as being
bending or pure torsion; it will simultaneously
due to deck action, certain other modes are not-
contain elements of all possible motions. A
ably low in deck contribution, for example, modes
major conclusion from this fact is that, in the
6 and 7. In fact, these two modes are strong in
general case, a simple section model is incapable
torsion (see Figs. 6 and 7) and might even have
of accomplishing a fully correct one-to-one
been selected on a purely intuitive basis as can-
analog simulation of the prototype. The model
didates for section model torsional simulation.
may continue to serve, however, as the analog
It is clear for such modes that application of
test bed for extracting the aeroelastic coef-
the method of Ref. [1] would lead to serious
ficients.
underestimation of the modal inertia. Instead,
The choice of prototype mode most "eligible" to
mode 13, with the highest
& value of the G
a. a.
be the eventual flutter mode should not depend
parameter, was chosen for this modeling. I^locle 3
merely on the visual or judgmental selection of
was chosen for model simulation in the bending
one mode, from among many, that appears to be
degree of freedom, although two other modes (9
"mostly bending" or "mostly torsion". Such a
and 12) have slightly higher values of the G^
^
choice requires the guidance offered by theory, 1 1
parameter.
using measured dimensionless aerodynamic coef- 1
It is particularly worth noting that, in mode 13:
ficients. These latter permit the complete
separation of the aerodynamic side of the problem I. (deck only)
l 0.955 5
x 10
from the dynamic side, whether model or prototype. ~G 0.3640
2 4 .Example: Modal Properties of Cable- a -a.
Stayed Bridge 262,363 mugs in“/in

The first 20 modes of a cable-stayed bridge de- 262,363 slugs ft~/ft
sign [8] were made available to the author. The
first 13 modes of this set are sketched in Figs.
and, in mode 3:
1 through 13. The deck, approximately uniform
1^ (deck only)
throughout the span, had a mass of 434.4 slugs/ft 0.935
0.3031
and a mass moment of inertia per unit span of
259,100 slugs ft~/ft about its c.g. Modes 3 3.0848 mugs in“/.
° m3
(largely deck bending) and 13 (largely deck tor-
3.0848 mugs/in
sion) were selected on a judgmental basis to be
modeled via a rigid, spring- supported deck sec- 444.2 slugs/ft

tion model for wind tunnel study. Full-bridge
generalized inertias I ^ , calculated for each
These two calculated values are good approxima-
mode i, were normalized at unit value. Similar
inertias, calculated over the deck only, were *Tn tne computer program used, modes were normal
ized on the basis that I. = 1 for all modes.
l
also calculated. Further, the values of certain
Input mass units used were "mugs" (1 mug weighs
modal integrals G , G, , , and G ,
’ in 386.4 lb); input length units used were inches.
a. a. h. h p p
consistent units,*
11
were
11 .
1 r .
1
.
1 Thus, nominal units for I. can be assessed as

additionally calculated.
2
mug in nominal units for G G, G
The results for the first 13 modes are presented * ;
a.a.h.h.p.p.
ii ii
, , ,
1
i i
in Table
bridge deck.)
1. (B = 100 ft.
Included in Table
= 1200 in. for the are inches; the modes a.,
edj dimensionless.
j • ii‘i h. , p. are consider-
1 for later
7
3
tions, respectively, to the deck inertia values
(259,100 slugs ft“/ft and 434.4 slugs/ft) of

A
2
* °- 004
O iO
2
4] (28)
the type given at the introduction of this exam- frequency in

where U/NB = 2 tt/K, N is the modal
ple. It is of further interest to note that sim-
Hz. These formulas are chosen to correspond to
ilar calculations hold for all modes that exhibit
fairly typical experimental values of these coef-
strong character in any one of the freedoms: however, that the formula
ficients. It is noted,
k
bending, sway, or twist. In Table 1, the largest
given for A1 describes a particularly severe
modal integral value in each mode is underlined.
case, or one with greater than average instabil-
In one case (mode 11), two modal integrals are of
ity tendencies. This is introduced intention-
comparable magnitude. In particular, modes 1, 3,
ally to illustrate the influence of such tenden-
4, 5, 9, 10, 12 are distinct enough in either
cies in torsion and to perceive, in particular,
bending or sway to yield calculations of deck
how they may participate in the full criterion,
mass that are of the same order as that given by
Eq. (25), for instability.
mode 3.
By Eq. (28), A* > 0 for (U/NB) > 2. The
The above facts emphasize that, particularly for
method of procedure used here is to fix the max-
cable-stayed bridges, problems can exist in the
imum expected cross-wind velocity at 160 mph.
selection of appropriate prototype modes that are
Thus U/NB = 2.35/N. This value is listed for
to be simulated in the section model. One possi-
each mode in the last column of Table 1. With
k k *
bility for such selection might be to choose,
this value and the formulas given for Hj , P ,
A,,
among all modes considered, the one for which the
the value of the left-hand side of Eq. (25) is
value of G , G , G, , was the largest
b
a. a. p.p. h.h. calculated.
i l
r
ii l i
(i.e., the largest among the underlined values of
Note first that, taking t,. = 0.01, 1. = 1,
Table 1). However, while these integrals offer a 1
p = 0.002378 slugs/ft = (12) x 0.002378 mugs/

quantitative indication that the mode is "mostly
in^, and B = 100 ft = 1200 in, yields the
bending" or "mostly torsion", etc., they also
criterion
directly influence the degree of participation of
4 C T 4
aerodynamic damping mechanisms. These effects i i _ 4(0,01)1 x 12
4 4 8
are not as easy to judge intuitively and require P B 0.002378 12 x 10
a sound theory. 7
= 1.6821 x 10 (inches).
This situation is characterized by the form of
All modes of Table 1, except one, result in cal-
Eq. (25), wherein aerodynamic damping derivative
culations that do not meet the criterion for
terms are multiplied respectively by the modal
flutter. However, the flutter criterion is rv
integrals given in Table 1. The aerodynamic
* k
ceeded at 160 mph for mode 13, which yields
damping terms and are usually negative,
effectively enhancing stability. The coefficient
h.h. (-1.25) (3.25) (0.13751 * 10
usually most likely to become positive, or de- i i
k i= 13
stabilizing, is A.,. We examine therefore in the
present bridge case, the effect of typical values -0.5586 x 10

k k k
of H^, and A9 upon the stability of vari-
ous modes. The following convenient formulas for
k k k (-0. 064) (3. 25) (- 8
. 2006) * lo
,
P^ and A^ will be employed:
i=1
k
1 o 25 JL)
(
(26) = 8
-0.04172 * io'
*
- 0.064 ( (27)
NB
8
3 i .
•k -k k
cients , P , A^.
A. G
2 a. a.
i i
i=1 3. SUMMARY AND CONCLUSIONS
The paper focuses on the aeroelastic wind tunnel
2 -5
0.004(3.25) [(3.25) -4][0. 3640] * IO sectional model of a bridge deck and close inter-
pretation of its relation to the prototype and to
= + 3. 10540 x 10" 7 other models. The viewpoint is particularly
7 espoused that bridge dynamics and unsteady aero-
sum = 3.1012 x io"
dynamics may usefully be divorced, conceptually,
from one another and reunited via theory. It is
7 '
Since 3.1012x10 > 1.6821x10 flutter is particularly useful to have in hand, as separate
possible in this mode at 160 mph. data, the flutter aerodynamic damping coefficients.
Another interesting mode is mode 1, which is When the theory outlined is applied to a cable-
unstable in torsion but simultaneously damped stayed bridge design, it points out needed refine-
via the sway action. For this mode. ments of interpretation regarding wind tunnel test
results for this newer type of configuration,
*
~ particularly in view of the greater modal complex-
H G = C-1.25)(13.72)(0.1414) x io" 11 ities observed. The method offers means for a
1 h.h
i i
i=l systematic, mode-by-mode investigation of bridge
= -2.4250 x 10"
U vulnerability to wind-induced instability.
The central role of torsional motion in driving
P, an instability is underlined, and its aerodynamic

G = “6
1 pi p i (-0.064) (13.72) (0.2242) x
0 interactions with vertical and sway effects are
i=l
delineated. The paper emphasizes the expanded
7
= -1.9687 xicf
use of the flutter derivatives -- particularly
those related to damping -- as the most fundamen-
tal analog results obtainable from section model
*
A- G test
2 a. a.
l i
4. REFERENCES
2
1. Irwin, H. P. A. H. "Centre of Rotation for:
0. 004(13. 72)[(13.72) -4] (0.1816) x io" 7

Torsional Vibration of Bridges," Jnl. Indus .
Aerodynamics 4_ (1979), pp. 123-132.

,
-7
= 1.8362 x io
2. Scanlan, R. H. "On the State of Stability Con-
:
-8 -7
sum = -1.3278 x io < i o 6821 x io siderations for Suspended-Span Bridges under
Wind," Practical Experiences with Flow- Induced
In this example,
Vibration Univ. Karlsruhe, Germany, 1979,
the proclivity for flutter in
Paper FI.
torsion (which alone exceeds the criterion) is
3. Teunissen, H. W. "Validation of Boundary- Layer
:
annulled by the damping effect of sway. This Simulation: Some Comparisons between Model and
illustrates how an effect (sway) usually neg- Full-Scale Flows," Wind Tunnel Modeling for
Engineering Applications, Cambridge Univ. Press,
lected may aid stability. It also illustrates U.K. pp. 217-235.
,
how near to flutter vulnerability a low mode may 4. Shiotani, H. and Ari, H. "Lateral Structures :
come if it exhibits appreciable torsion. of Gusts in High Winds," Proceedings, Third

Inti. Conf. on Wind Effects on Bldgs, and
In general, given the data and procedures of the Structures, Saikon Co., Ltd., Tokyo, 1971,
p. 535-543.
present example, all modes of a bridge may be
5. Scanlan, R. H. and Tomko, J. J.: "Airfoil and
systematically examined in the manner illustrated Bridge Deck Flutter Derivatives," Jnl. Eng.
based upon experimental knowledge of the coeffi- Mech. Div., ASCE, Vol. 97, No. EM 6, Dec. 1971,
pp. 1717-1737.
9
. 0
2
3
6. Scanlan, R. H. "The Action of Flexible Bridges

:
8. Christopher, B. and Stroch, S., Private Commun-
Under Wind," Parts I and II, Jnl. Sound and ication, Greiner Engineering Sciences, Inc.,
Vibration (1978), Vol. 60(2), pp. 187-199 and Tampa, Florida, February 19S5.
201 - 211 .
7. Nakamura, Y., and Ybshimura, T. "Binary :
Flutter of Suspension Bridge Deck Sections,"

Jnl Eng. Mech. Div ASCE Vol. 102, EM4,
. , ,
Aug. 1976, pp. 685-700.
TABLE 1
Mode Freq ! G G Description V/NB P If

Number (Hz)
(deck
i a Q
i i Vi p
i
p
i of form = 2.3!
inches inches
only)
1 0.171 0.992 0.1816 E- 07 0.1414 E-l 1 0.2242 E-06 lateral 13.72
2 0.188 0.781 0.3354 E-l 1 0.0896 E-07 0.4358 E-l vertical 12.48
*3 0.270 0.935 0.3193 E-l 3 0.2105 E-06 0.2064 E-l vertical 8.69
4 0.349 0.890 0.3614 E-l 2 0.2010 E- 06 0.1662 E-ll vertical 6.72
5 0.389 0.894 0.4876 E-07 0.1685 E-ll 0.1975 E-06 lateral 6.03
6 0.516 0.050 0.7846 E- 07 0.3021 E-l 3 0.0739 E-07 torsion 4.57
7 0.522 0.062 0.9214 E-07 0.3644 E-l 3 0.0795 E-07 torsion 4.50
8 0.552 0.694 0.5313 E-07 0.0729 E - 09 0.1317 E-06 lateral 4.25
9 0.555 0.947 0.1778 E-l 0 0.2133 E-06 0.4478 E-l vertical 4.23
10 0.645 0.975 0.5484 E-l 4 0.2201 E-06 0.3812 E-l vertical 3.64
11 0.696 0.653 0.1239 E- 06 0.4637 E-l 2 0.1 130 E-06 tors.-lat 3.37
12 0.714 0.988 0.2264 E-l 1 0.2238 E-06 0.1817 E-l vertical 3.29
*13 0.721 0.955 0.3640 E- 05 0.1375 E-l 2 0.2006 E- 08 torsion 3.25
Selected to be simulated via wind tunnel sectional model
PLAN
10
PLAN
Figure 2
PLAN
ELEVATION
MODE 3
f = 0.270
Figure 4
II
PLAN
Figure 5
PLAN
ELEVATION MODE 6
Figure 6
f = 0.516 Hz
f = 0 522 Hz
Figure 7
12
PLAN
ELEVATION f = 645 Hz
Figure 10
13
Figure 13
ELEVATION
14
. .
CHARACTERISTICS OF AERODYNAMIC FORCES ON BRIDGE DECKS
IN SMOOTH FLOW AND IN TURBULENT FLOW
by
1 2
Nobuyuki Narita Hirohiko Tada
, ,
Koichi Yokoyama^ and Hiroshi Sato 4
ABSTRACT 1. INTRODUCTION
The wind-resistance of long-span bridges in Japan The wind-resistance of long-span bridges in Japan
has been mostly investigated and proved through has been mostly investigated and proved through
the wind tunnel tests in smooth flow. Natural the wind tunnel tests in smooth flow. The wind-
wind, however, is turbulent flow, whose wind induced vibrations of long-span bridges are
speeds vary with time and space. Recent studies classified into self-excited vibrations (vibra-
have shown that the characteristics of aerodynamic tions caused by the fluctuating aerodynamic
forces on bridge decks in turbulent flow are dif- forces which are generated by the vibrations of
ferent from those in smooth flow. Therefore, the the structure itself and which promote the vibra-
effects of turbulence on wind-induced vibrations tions) and forced vibrations (vibrations caused
of bridges should be investigated in order to by the fluctuating aerodynamic forces which are
establish the rational wind-resistant design generated by vortices behind the structure or by
method. approaching turbulence)
In this paper are described the characteristics Buffeting, the forced vibrations caused by ap-
of fluctuating lift forces, which cause vertical proaching turbulence, cannot be observed in the
bending vibrations in bridge decks. The fluctu- smooth flow, and it has been shown in recent
ating lift forces acting on some typical bridge studies
12
’
’
3
that the characteristics of aerody-
decks were measured both in smooth and turbulent namic forces on bridge decks in turbulent flow
flow, and the effects of turbulence and cross- are different from those in smooth flow. There-
sectional shape of decks on the fluctuating lift fore, the effects of turbulence on wind-induced
forces were clarified. vibrations of bridges should be investigated in
order to establish the rational wind-resistant
KEY WORDS: Aerodynamic (forces). Bending (verti-
cal vibration), Bridge (decks). Damping (aerodyn- design method.
amic), Lift (forces). Smooth (flow). Suspension
Bridges, Turbulent (flow). Vibration (wind- induced) From these points of view, the characteristics of
Wind Tunnel (testing).
fluctuating lift forces, which cause vertical
bending vibrations to bridge decks, were investi-
gated. The fluctuating lift forces acting on
Dr. of Eng., Asst. Director General, Public some typical bridge decks were measured both in
Works Research Institute, Ministry of Construc-
tion, Japan.
smooth and turbulent flow.
2. CHARACTERISTICS OF TURBULENCE USED IN
Dr. of Eng. Director, Structure and Bridge
,
EXPERIMENT
Department, Public Works Research Institute,
Japan The experiments were made in the Low Speed Wind
'Head, Structure Division, Structure and Bridge Tunnel-B of Public Works Research Institute, whose
Department, Public Works Research Institute, specification is shown in Table 1. The turbulence
Japan.
was generated by a coarse grid whose mesh size is
Research Engineer, Structure Division, Public
Works Research Institute, Japan. 0.25m and whose bar size is 0.05m. The position
15
. . . . , . ,
and the shape of the grid are shown in Figs. 1 2 . 1 Relations Between The Wind-Induced Vertical
and 2, respectively. The mean wind speeds of the Bending Vibrations and the Fluctuating Lift Forces
turbulent flow were uniformly distributed along Wind-induced vibrations of vertical bending mode
the y-axis (x-axis is along the main flow direc- of long-span bridges can be estimated using the
tion, and y-axis is horizontal and perpendicular following equation of motion.

to x-axis, and z-axis is vertical). The charac-
M.fq. + 2£ w q + oo.q.l = P. ( 1 )
4 . .
l
L
i iî
.
iî l
teristics of the turbulent flow are as follows.
1 2
where M. / m(x)$7 (x)dx, Nh; generalized mass
1
Fluctuating Kind Speed th
of i mode, m; mass per unit length, 1; bridge
u x-comp
: w z-comp
:
length, x; axis along the bridge length, 4 ;
Intensity of Turbulence I =6.2% I =5.0% t '1

u w i mode shape, h(x,t) = Iq.(t)4-(x), h; dis-
Scale of Turbulence:
placement of vertical bending vibration, t; time,
U W
y-direction L =0.09m L =0.04m th
X X q^; i generalized coordinate, ^ ; structural
U W
z-direction L =0.04m L = 0.04m damping ratio of i^ mode, 0).
l
;
natural circular
y y
frequency of i mode, (•) = d( )/dt.
.th
The Power Spectral Density Functions of u and P^ on the right hand side of Eq. (1) means i
w and S
(S respectively)
f
at the position
, r
generalized force due to the fluctuating lift
uu ww
where the lift forces were measured, were com- force. In general, the fluctuating lift force
pared with those of natural wind. In the compar- is the summation of self-excited component P^..
)
ison, frequency f was reduced by deck width B ‘
Ei
and mean wind speed U. S and S were P. = P_. + P_. ( 2 )
r Si Ei
uu ww l
reduced by J and f. They are shown in Figs.

where, P„. = /*p c . (x)dx
/ n p r.4. (x)dx
Si CP S l 0 E l Ei
3 and 4. In the figures, the value of one box
Ps ;
self-excited component of fluctuating lift
girder bridge was applied to B (18.25m for real
force per unit length, p^; forced component of
bridge and 0.30m for model). S and S of
uu ww fluctuating lift force per unit length.
natural wind were estimated from the data of
The self-excited component can be expressed as
ESDU.4 According to ESDU,
and S of S
uu ww
fol lows
natural wind depend on the height and the rough-
2
ps = (l/2)pU B[C^(h/B) cJnOi/u)]
ness length of the terrain surface (z^). In the
figures, the height was assumed to be 50m and z where p; air density, U; mean wind speed, B;
U
-4
was changed from 10 m to lm. The relations be- bridge deck width, C j ; coefficient of the self
^ |
tween Zq and terrain description are as follows. excited lift force in phase with the vibration*]
Zp=lm: center of large towns, displacement. coefficient of the self-
z =10 ^m: outskirts of town or farmland excited lift force in phase with the vibrati mal
U
-2
z =10 m: fairly level grass plains velocity.
U
-4
Zq= 10 m: calm open sea Pg^ can be written as follows, using the orthog
The turbulence intensity of u generated by the onality condition of mode shapes.
grid corresponds to that of the calm open sea.
The scale ratio (the ratio of the scale of turbu-
P
Si
qp/Jo/npu^dx] .
lence to the bridge deck width) is about 1/30 of q.[/J(l/2) P UBc[ || «Tdx] (4)
natural wind. The intensity of w corresponds

Substituting Eq. (2) into Eq. ill and moving I’
to that of fairly level grass plains, and the

to the left hand side, then the foil '»in.: oqua
scale ratio is about 1/10 of natural wind.
tion is obtained
j
M [
l q . + 2 ( C C .)(i). q. P .
.
i î l ai iî Ei
16
) ^
ment meter at the same time as the lift force.

where,
^
aerodynamic damping ratio of
= - [f
1
(1/2) pUBC^dx]/ (2M.U).)
i
j.L
tn mode,
, C
&i
;
The coefficients of self-excited lift force C

R
LH
I
= ' [/J(l/2)pU“C^ H $^dx]/M i , a ; apparent and C or logarithmic aerodynamic damping 6
“ai “i &i LH £L
can be calculated from the frequency response

circular frequency of i^ mode in wind.
function between vibrational displacement and lift
As is shown in Eq. (5), both self-excited compo-
forces. For linear system with input x and
nent v[which results in £ . and w ) and forced
the frequency response function
.
ai at output y, H (f)
component P act on bridge decks. When
^ can be obtained from Power Spectral Density
C,. + £ .
ai
=0, that is when the summation of
Function (PSDF) and Cross Spectral Density
l
structural damping and aerodynamic one is 0 or
S
^
Function (CSDF) S as follows.
yx
negative, then the aerodynamic instability will
take place even if
vibrations caused by this mechanism is called
. is 0. The wind- induced V
The fluctuating lift forces
f) = S
yx
(f)/S
XX
(f)
L acting on the
( 6)
self-excited vibrations. Galloping is the diver- vibrating model whose length is 1 can be
gent self-excited vibration of vertical bending written as
mode. The characteristics of galloping can be 2 R
L = (l/2)pU Bl[C (h/B) + Cj^Ch/U)] + p dx
LH £
predicted if the characteristics of self-excited
(7)
component of fluctuating lift forces or the aero-
If the independence between the vibrational dis-
dynamic damping are clarified.
placement h and the forced component of the lift
Vibrations will also take place due to the forced forces is assumed, the C.S.D.F. of L and
p^
component of lift forces even if the summation of h ’ S ls
Lh
structural damping and aerodynamic one is positive. 2
S = (l/2)pU Blfc|J /B i27Tf<:' /II)S (8)
Lh H rt hh
Buffeting is the forced vibration due to approach-
R I
Therefore, C and C TII can be calculated from
ing turbulence. The magnitude of the forced vi- LH LH
the frequency response function of L and h.
bration depends on the forced component and aero-
H. as follows.
dynamic damping. It increases with the magnitude Lh
, ,
R 2
of the forced component, and it decreases as the C [f = Real (H )/((l/2)pU l) (9)
H r) Lh
summation of structural damping and aerodynamic
2
C (f = Im[H )/[((l/2)pU l)-(2uf (10)
one increases. } )]
LH r Lh r
2 . 2 Measurement Method of Fluctuating Lift
where f = (fB)/U. Real; real part, Im; imagi-
Forces
The fluctuating lift forces were measured using nary part. Aerodynamic damping depends on ,
the dynamic balance^ developed at the Structure C* and mode shape in general. When the effect
LH
^
Division of PWRI. of is negligible and the mode shape is
Sectional models of bridge
decks were kept at rest or shaken in heaving mode uniform, it can be calculated as
2
with an amplitude of B/100. The measurement was 0-6 = -Im(H )/[(8Trfp((l/2)pU l)] (11)
a Lh
made both in smooth laminar flow and in turbulent 2
where, a; reduced mass (=m/ (pB )).
flow whose characteristics are described in Chap-
ter 2. Before the measurement of the lift forces The self-excited component of fluctuating lift
on oscillating models, inertia forces are balanced forces was transformed into 0*6^ in this paper.
out by adjusting the mass and the position of On the other hand, the forced component L act-
dummy weights to measure the lift forces only. ing on the model at rest was reduced to C. by
L
2
Shaking frequencies were swept from 0.1Hz to 5.0 dynamic pressure (l/2)pU , model width B and
Hz, and the models were shaken more than 100 cycles model length 1. Then its standard deviation
at every frequency. The vibrational displacement and P.S.D.F. were calculated.
of the models were measured by optical displace-
17
. 6 . 6
3. EXPERIMENTAL RESULTS 1. Negative aerodynamic damping acts on the rela-

From Fig. 5 to Fig. 12, shown are cross-sectional tively thick bridge decks in high wind speeds
shapes of the bridge decks tested, the product of of smooth flow, that is, self-excited vibra-
reduced mass and aerodynamic damping a* , and tions of vertical bending mode, galloping, may
9.
P.S.D.F. of the coefficient of the forced compo- occur. In turbulent flow, however, aerodynamic
nent The standard deviation of C^, damping remains positive up to higher wind
are also shown in the figures. speed, and the possibility of galloping
becomes smaller.
a* 6 were measured in a range of relatively
a
2. Turbulence hardly changes the characteristics
small f , that is relatively high wind speeds.
of aerodynamic damping in strong winds in case
0*6 of the relatively thick bridge decks (model
3
of very flat bridge decks and stiffening
.
A to D) becomes negative at small f in smooth

trusses. The aerodynamic damping is positive
flow. It corresponds to the onset of self-excit-
and may increase the apparent damping.
ed vibrations of vertical bending mode, what is
3. The forced component of the fluctuating lift
called galloping. In turbulent flow, however,
forces acting on the relatively thick bridge
the f at which a* turns negative becomes
r a
decks becomes much larger in turbulent flow
much smaller (model B) ,
or a *6 remains posi-
cl
than in smooth flow.
tive even at the smallest f in the measurement
r
4. The forced component of the fluctuating lift
range (model A, C, D) . It suggests that the
forces in turbulent flow decreases as bridge
possibility of galloping is much smaller in tur-
deck becomes thinner.
bulent flow than in smooth flow.
5. REFERENCES
On the other hand, a *6 of the relatively thin
1. A. G. Davenport, N. symov and T. Miyata:
I The
bridge decks (model E to G) and a *6 of the
3 experimental determination of the response of
stiffening truss (model H) are positive both in suspension bridge to turbulent wind. I'rod. 3rd
Int. Conf. on Wind Effects on Bldg', and Sts.,
smooth and turbulent flow. It suggests that no
Tokyo, 1971.
galloping will take place for these bridge decks.
2. W. H. Melbourne: West Gate Bridge wind tunnel
The effect of turbulence on 0*6 is negligible tests. Aug. 1973.
3.
in case of very thin bridge decks (model G) and 3. H. P. A. Irwin and G. D. Schuyler: Experi-
H.
ments on a full aeroelastic model of ions I
stiffening truss (model H)

Gate Bridge in smooth and turbulent flow.
Laboratory Technical Report, National Re- ran h
S and C were measured both in smooth and
Vj LjLj Li Li Council Canada, Oct. 1977.
turbulent flow for model B. It is found that
4. ESDU: Engineering Science Data. Wind Engineei
in turbulent flow becomes about 3 times as large ing. Volume 1, Wind and turbulence data.
Engineering Science Data Unit Ltd., 19S2.
as Cr in smooth flow, and that the f where
L r
5. T. Okubo, N. Narita and k. Yokoyama Some :
large power is concentrated becomes smaller in

approaches for improving wind stability of
turbulent flow. For other models, measurements cable-stayed girder bridges. I’roc 1th Int.
.
Conf. on Wind Effects on Bldgs, and Sts.,

were made only in turbulent flow. It is found
Heathrow, 1975.
that decreases as bridge deck becomes
6. J. S. Bendat and A. G. Piersol: Random pat.a,
thinner Wi ley- Intersc ience , 1971.
4. CONCLUSIONS
The fluctuating lift forces acting on some typical

bridge decks were measured in both smooth flow
and in turbulent flow generated by the coarse
grid. The following characteristics were clari-
fied.
18
Table 1. The Specification of the Low Speed Wind Tunnel B
Type Gottingen Type
Dimension Test Section
Width 1.000 mi
Height 2.000 mm
Length 3.000 mm
Contraction Cone
Inlet 3.000 mm X 5.000 mm
Outlet 1.000 mm X 2.000 mm
Contraction Ratio 7.50 : 1.00
Total Path 55,000 mm
Blower Type Axial with Fixed Blades
Diameter 1.500 an
Total Pressure 40.0 mm Aq
Total Wind Volume 70.0 mVsec
Maximum Rotation 1,250.0 rpm
Maximum Power 50.0 KW
Flow Maximum Wind Speed 30.0 m/sec
Turbulence Intensity less than 0.5%
Contraction Cone Test Section

3.5 m 3.0 m
1.1m T79H
Coarse Grid Bridge Deck Model
Fig.l. Position of the Coarse Grid
19
/ H
r—
o
200 0
uy
n
s
in
(/>
2D
o
o
O'
CM
in smooth flow
in smooth flow
in turbulent flow
o in turbulent flow
LT>1
o E iil
o'
,
; t
O! I ii a
O LT)
CO
6 6
o
‘
10 10 °
Mrl
10 °
o.
tr> fB/U
fB/U
o
o tn
o
i
o
CM
q
CM
Fig. 6. Model B Fig. 7. Model C
21
16.7
o
LO
o
in smooth flow I
in turbulent flow 'J
o
in
co
b
°10' 3 10
-2
107 V 10 °
A
o
O fB/U ° 10' 3
in'
10' 2 10' 1 10 °
fB U
Fig. 8. Model D Fig. 9. Model E
22
fB/U
CM
o
Fig. 10. Model F Fig. 11. Model G
25
107.1
o
CO
° 10' 3 10' 2 °
10' 1 10
fB/U
Fig. 12. Model H
Zi
"Field Measurements on the
Deer Isle-Sedgwick Suspension Bridge
by
Harold R. Bosch*
ABSTRACT With the objective of obtaining a better

This paper first provides a general characterization of the wind environment in
introduction to Federal Highway Administra- the vicinity of highway structures, bridges in

tion's Wind Engineering Research Program which particular, and a clearer understanding of the
deals, in part, with full-scale measurement of structure's response to the wind loadings,
wind characteristics and bridge response to major effort within Task 1 of FCP Project 5A
wind loadings. It then focuses on a current has been and is being directed toward
program of field testing that is being development of automated bridge instrumentation
conducted in cooperation with the Maine systems and evaluation of data collected by
Department of Transportation on the Deer such systems. Information collected
Isle-Sedgwick Suspension Bridge. Details of the traditionally consists of wind parameters such
bridge structure, bridge site, and instru- as velocity components, direction, turbulence
mentation system are presented. Data analysis intensity, turbulence scale, turbulence
is discussed and samples of the results are spectrum, and associated bridge response
provided. Comments are made regarding future parameters such as displacements, accelera-
work and plans. tions, strains, damping, and frequencies.
Keywords The FHWA installed its first field wind

Suspension bridge, full-scale measurements, instrumentation system on the Newport, Rhode
wind characteristics, bridge vibration, wind Island Suspension Bridge in cooperation with
turbulence, aerodynamic stability the Rhode Island Turnpike and Toll Bridge
Authority. Sensors were located on outrigger
1. INTRODUCTION booms and positioned symmetrically along the
During the past 15 years, the Federal Highway main span. While in operation, the system
Administration (FHWA) has sponsored as well as successfully recorded the passage of Hurricane
conducted many research studies addressing the Doria and analysis of this data is documented
guest-ion of wind and earthquake loadings on in several research reports [2]. The
highway structures. As part of Federally instrumentation was then relocated to Sitka,
Coordinated Project 5A, "Bridge Loading and Alaska where it was modified to incorporate
Design Criteria," the research has covered such improvements and installed on the Sitka Harbor
areas as design procedures, computer modeling,
field instrumentation systems, wind tunnel
testing, shake table testing, retrofitting, and *Research Structural Engineer,
fuil-scale field testing [1], Federal Highway Administration,
6300 Georgetown Pike, McLean,

Virginia 22101, U.S.A.
25
Table-Stayed Bridge. In a 5 year cooperative coast of Maine. Built in 1938, the bridge
effort with the Alaska Highway Department, data spans Eggemoggen Reach and consists of a
was collected on the Nation's first vehicular girder-stiffened deck similar in cross section
cahle-stayed bridge. Vortex excitation of the to the original Tacoma Narrows Bridge. The
free-standing pylon towers, response of the main span is 1080 ft (329m) in length with side
deck, and wind characteristics at the site have spans of 484 ft (148m) and approach spans of
been evaluated and are reported in various 130 ft (40m) for a total bridge length of 2308
papers and reports [2, 3]. Following the Sitka ft (704m). As shown in figure 2, the structure
investigation, the wind instrumentation system is symmetrical with two towers and the roadway
was installed on the Perrine Memorial Bridge has a 6.5% grade to provide vertical naviga-
in Twin Falls, Idaho to study wind induced tional clearance of 85 ft (26m) at midspan.
vibration of the bridge columns. This bridge The 2-lane deck consists of a 4.5 in (llcml
structure, spanning the Snake River Canyon, concrete slab having a width of 20 ft (6m).
consists of a four-lane deck on plate girders Stiffening girders are 6.5 ft (2m) deep and
supported by slender, unbraced columns atop a spaced 23.5 ft (7m) apart, figure 3 provides
steel arch. Column motion was measured before additional deck and tower details.
and after installation of special damping
devices [A]. The instrumentation system was 2 . 1 Bridge Site
next moved to the Pasco-Kennewick Intercity The bridge site is situated in Hancock County
Bridge in the State of Washington. This approximately 55 miles (88 km) east of Augusta
structure is the first concrete cable-stayed and 35 miles (56 km) south of Bangor. This
bridge in the United States. Wind location is noted on the map of Maine which has
characteristics at this inland site and bridge been included as Figure 4. The bridge carries
response measurements have been documented in a State Route 15 over Eggemoggen Reach and
recent report and several papers [5], In 1981, connects Sedgwick on the mainland with Deer
the FHWA instrumentation system was completely Isle and Stonington on the Islands. The
refurbished and several new types of sensors longitudinal axis of the structure lies on a
were added to upgrade and expand its capability NE-SW orientation with the mainland being at
[6]. The system was then installed on the Deer the north end of the bridge and Little Deer
Isl e-Sedgwick Suspension Bridge in State of Isle at the south end. The countryside to the
Maine. Through a cooperative arrangement with north consists of heavily forested, rolling
the Maine Department of Transportation (MEDOT), hills varying in elevation to a few hundred
FHWA is investigating the aerodynamic feet. To the south and east are many forested,
stability of the Deer Isle Bridge in order to low-lying islands, Deer Isle being the largest,
provide recommendations regarding bridge with the Atlantic Ocean beyond. To the west
rehabilitation. This study is currently there is Penobscot Bay with a small mountain
underway and will be described in more detail range beyond.
in what follows.
3 . INSTRUMENTATION - S ENSOR S
A large array of instruments has hern insta
2 . DEER ISLE-SEDGWICK SUSPENSION BRIDGE on the Deer Isle Bridge to monitor th<- wind
The Deer Isle-Sedgwi ck Bridge, illustrated in environment and structural response tn wind
Figure 1 ,
is a conventional suspension loadings. The positions of sensors on the
structure located near Penobscot Bay on the bridge are indicated in Figure 5. Mean wtnd
26
speed and direction are measured by two skyvane To examine tower motion, three accelerometers
sensors mounted on the top and bottom of the are used at the top cross girder of the north
north bridge tower. The upper skyvane is 18 ft tower. Two of these are mounted at the top of
(5.5m) above the tower top while the lower one each tower leg with their axes oriented to
is located approximately 30 ft (10m) above the point northeast. This defines bending and
water. Air temperature is measured by three twisting of the tower top. The third is
thermistor probes installed at the top and pointed southeast to indicate tower sway.
bottom of the same tower and at the instrument
house north of the bridge. Six tri-axis 3 .1 Instrumentation - Data Recorder
anemometers are used to monitor wind turbulence Data recording eguipment as well as lab test
guantities (speed, direction, intensity, scale) eguipment, parts, and supplies are housed in a
at the elevation of the bridge span. These small instrument house located just off the
anemometers are mounted on 12 ft (3m) north end of the bridge. This building is
outriggers cantilevered from the east side of outfitted with heat, air conditioning,
the superstructure and spaced in a logarithmic telephone, security system, and motor-generator
array. Each anemometer set is aligned with the (for conditioning incoming power). Analog
bridge so that the "U" component is normal to signals from all sensors located on the bridge
the longitudinal bridge axis (facing south- are routed to this instrument house via
east), "V" component is parallel to the bridge multi-conductor, shielded cable installed on
axis (facing northeast), and "W" component the west side of the bridge. Approximately
points upward. The tri-axis anemometers 20,000 ft (6096m) of cable was reguired to
employed are dynamic instruments which are instrument the Deer Isle Bridge. To insulate
capable of tracking velocity fluctuations in the instrumentation from lightning strikes,
the freguency range of interest. protection devices were installed in each
accelerometer enclosure and at the input panel
Bridge deck response is monitored by six pairs in the instrument house.
of single-axis, servo accelerometers installed
along the north side span and main span. The analog signals from the bridge sensors are
Accelerometer stations on the side span are passed through signal conditioning and
located at midspan and the guarter point amplified to provide maximum resolution in the
nearest the north tower. For the main span, recording system. Anemometer gains were
stations are located at midspan, both guarter adjusted to accommodate wind velocities up to
points, and the eighth point nearest the north 100 mph (161 Km/H) and accelerometer gains
tower. At each station, an accelerometer adjusted to handle .25g. Amplified signals
mounted inside a weathertight enclosure is from the deck anemometers and all acceler-
clamped to the bottom flange of each stiffening ometers are low-pass filtered at 10 Hz to
girder. The axis of each accelerometer is avoid aliasing of the data. All signals are
oriented to point upward to measure vertical then routed to the data recorder.
motion. The pair of accelerometers at each
station define the vertical displacement and The recorder is an automated data acguisition
rotation of the span at that point. The whole system (DAS) eguipped with dual 9-track digital
array of stations are used to determine mode tape drives providing enough capacity for 8
shape and freguency for the structure.
27
hours of recording [6]. Figure 6 illustrates direction obtained from the skyvane sensors on
the data path and control logic employed by the the tower are displayed in polar form and
data acquisition system. The DAS continuously evaluated to detect any trends in the recorded
scans up to 60 channels of analog input at a data. Figure 7 is a polar plot of wind
rate of 20 Hz. Data recording does not begin activity at the top of the north bridge tower.
until either wind velocities or bridge This graph represents an accumulation of
accelerations exceed preset levels for a recorded data and is up dated each time new
selected period of time. Once initiated, information is obtained. Hourly means are
recording continues until signals drop below obtained from the continuous chart recordings
these threshold values. This mode enables and plotted in time-history fashion for each
unattended operation and optimizes use of month. As demonstrated in Figure 8, this
available data storage capacity. While technique is a useful means of compressing
recording, the system continuously scans all large volumes of data and enables identifica-
channels, converts the analog signals to tion of significant storm events.
digital data, and stores this data on magnetic
tape for later processing. When one tape Detailed data analysis is being performed for
fills, the recorder automatically switches to FHWA by a research contractor. As events of
the other drive and continues without loss of interest are identified during the pre-
data. Wind speed and direction measured by the processing, copies are made and forwarded to
two tower anemometers are also recorded on the contractor for detailed analysis. The data
strip-chart recorders. This recording is processing sequence is illustrated in Figure 9.

continuous and documents the day-to-day wind This analysis is accomplished in three stages.
conditions at the site regardless of whether or In the initial stage, raw data is converted to
not the DAS is active. engineering units and summarized using program
TAPSUM much the same as was done during
3 .2 Data Reduction and Analysis preprocessing by FHWA. The intermediate stage
After wind and bridge response data has been prepares anemometer data from selected events
recorded on digital tape, the tapes are for detailed analysis by applying response
forwarded to Turner-Fairbank Highway Research corrections and rotating the data into the
Center (TFHRC) for preprocessing and evalua- desired coordinate system. This is
tion. This data reduction is accomplished by accomplished using program MASTER. The final
FHWA staff using various microcomputers. Data stage of data processing involves spectral
tapes are first scanned for errors. Raw data analysis of the time series data. In ANAL2,
is converted to engineering units using sensor time series data is first prepared for analysis
calibrations obtained in the field and channel by point averaging, detrendinq, and tapering.
statistics such as minimum, maximum, mean, and Next, fast Fourier transform techniques are
standard deviation are compiled for 10/20- employed to process data from 1, 2, or 3
minute blocks of data. A typical 10-minute channels simultaneously. This analysis provides
summary is provided in Table 1. This statistics on all the data as well as power
information is used to evaluate system spectral densities and auto correlations for
performance and to identify events which the individual series. In addition,
warrant detailed study. Mean wind speed and co-spectral and quadrature spectral densities,
28
,
coherence, squared coherence, phase difference, damping values, identifying vortex shedding
and lagged cross-correlations are available parameters (including Strouhal number) and
for each pair of series. evaluating the along-span coherence of

turbulence.
Bridge displacements are extracted from

acceleration time histories using a The existing data set provides extensive wind
sophisticated double integration program called coverage from the west, some from the east, and
DAISMA [7], Figure 10 is an example of a very little from the north or south. Wind
measured acceleration record and the computed velocities are generally in the 20-30 mph
displacement time history. Bridge mode shapes (32-80 Km/H) range. So far, no major bridge
are determined using a semi-automated process motions or vortex induced response have been
which examines spectral amplitude data identified. More information is needed
searching for pronounced peaks. Figure 11 is a regarding severe storms containing high wind
spectral plot of deck acceleration at velocities or strong turbulence as well as weak
accelerometer 9 with 7 peaks identified. fronts containing mild, steady winds. Better
Similar results for all accelerometers are directional coverage is also desirable. To
listed in Table 2. Bending modes are derived accomplish this, a newer and more flexible data
using the average signal from the accelerometer acquisition system with remote communications
pair at each bridge station. The difference capability was installed at the instrument
between signals from each pair is used to house in the fall of 1984 and data acquisition
develop torsional modes. Bending modes which is continuing.
have been identified so far are illustrated in

5. REFERENCES
Figure 12.
(1) O'Fallon, 0., "FCP Annual Progress Report
Wind spectra obtained for each selected event - Year Ending September 30, 1984," Project
are compared with three theoretical models. 5A, FHWA, September 1984
Typical spectral comparisons for the "U" and
"W" wind components are shown in Figures 13 and (2) Gerhardt, C. L., et al ,
"Selected Wind and
14, respectively. In addition, turbulence Bridge Motion Data Recorded at Long-Span

intensity and scale are computed for comparison Bridge Sites, " Report No. FHWA-RD-76-1 80,
with existing meteorological data. Typical FHWA, August 1976
values obtained from two wind events are

included in Table 3. (3) Nelson, D. ,
et al ,
"Wind and Bridge Motion
Data Recorded at the Cable-Stayed Bridge,
4. FUTURE WORK Sitka, Alaska, " Report No. FHWA-RD-78-57
All data collected to date has been pre- FHWA, May 1977
processed and summarized; however, only a small

portion has received detailed analysis and (4) Cheng, D., "Aerodynamic Stability of Long
evaluation. Therefore, analysis will continue. Columns With Elastic End Constraints," Report
Attention will be focused on determining the No. FHWA7RD-81/133, FHWA, February 1982
full range of bridge displacements, developing
torsional mode shapes, extracting structural
29
References (continued)
(5) Bampton, M., et al ,
"Pasco-Kennewick
Cable-Stayed Bridge Wind and Motion Data,"
Report No. FHWA/RD-82/067 ,
FHWA, February
1983
(6) Strope, L., "FHWA Data Acquisition

System, " Unpublished Report, FHWA,
January 1982
(7) Miklofsky, H., "Digitization and

Integration of Strong Motion
Accelerographs, Report No. FHWA-
RD-77-92, FHWA, August 1977
30
: I
TABLE 1. Tape Summary - 10 Minute Data Block
DEER ISLE - SEDGWICK BRIDGE
Tape Number: 119 Block Size: 1925 Cho nne

Start Record: 4301 End Record: 5200
Day: 290 Tine: 18:33:30 Pwr: 0

Day: 290 Ti ne 18:43:43 Pw r : 0
HEL MINIMUM MAXIMUM MEAN STD DEV
0 -0.0326 0.0131 -0. 0021 0. 0045

1 -0.0268 0.0096 -0. 0003 0.0037
2 -0.0233 0.0117 -0. 0054 0.0033
3 -0.0114 0.0095 -0 0007 . 0.0034
4 -0.0114 0.0114 -0. 0003 0.0032
5 -0.0110 0.0073 -0.0003 0.0030
6 -0.0106 0.0106 -0.0001 0.0034
7 -0.0112 0.0130 -0.0006 0.0036
3 -0.0129 0.0129 -0.0008 0 0034 .
Cl
-0.0141 0.0121 -0.0008 0.0033
10 -0.0113 0.0094 -0.0009 0.0037
ii -0.0 1 65 0.0110 -0.0002 0.0035
12 -0.0113 0. 0094 0.0018 0.0029
13 -0.0033 0. 0152 0.0081 0.0029
14 0.0011 0.0118 0. 0073 0.0016
15 0 0000 . 0 0000 . 0. 0000 0.0000
IS 0 0000 . 0 0000 . 0 0000 . 0 0000 .
17 0 0000 . 0 0000
. 0 0000 . 0.0000
13 0 .0000 0 0000 . 0.0000 0. 0000
19 0 0000 . 0 0000
. 0.0000 0. 0000
20 -21 5469 . 0.7695 -11.5911 5.3622
21 -5.7268 10.7377 1.4961 2.1137
-3.3629 10.5970 -2.0139 3. 2206
23 -25.6511 3.0173 -12.8086 6. 2303
24 -3 . 4361 10.3005 1 3423
. 2. 3343
“ 3 12.7035 -2.5232 3.5603
'? . 1 63
2K -24. 6368 2.3144 -13.6838 5.6273
27 -4. 6756 3.5720 1.3423 1.9319
25 -9.0240 12.2330 -1.9314 3.2273
V -26. 7996 0.7657 -13. 4075 5.7323
30 -5.6221 7.7304 1 2650
. 2. 1762
31 -10. 7949 11.5659 -2.7344 3.6344
J w -24.3110 7. "534 -10.5602 6. 3790
C* -• -7.2033 9.3714 1.5625 2.2305
34 -10.2116 10.9823 -2.4041 3.6457
C|
-23.0859 5.3867 -11. 2640 5.9269
J -7 2427 . 12. 3125 1.0122 2.6451
37 -11. 3280 13.3240 -2 4072
. 3.5433
3S 0. 0000 0.0000 0.0000 0 . 0000
35* 0. 0000 0. 0000 0 0000
. 0 . 0000
40 0000
0. 0 0000. 0 . 000O 0 . 0000
41 0000
0. 0 0000. 0 0000
. 0 . 0000
42 0.0000 0. 0000 0.0000 0 0000
.
43 -5.0781 -5.0781 -5.0781 0 0001

.
44 5.0731 5. 1563 5. 1355 0. 0345

45 0 .0000 0. 0000 0 0000
. 0. 0000
46 0.0000 0 0000, 0 0000
. 0 0000
.
47 0.0000 0 0000. 0.0000 0 0000

.
43 0. 0000 0.0000 0. 0000 0 0000

.
49 0 0000 . 0 0000. 0 0000

. 0 0000
.
50 17. 1375 25.0000 21.0977 1 6296

.
51 253. 7500 236.3750 271. 1313 4. 4200

c '2.0375
Jw 67. 5000 77. 5000 1.3911
53 47 5000 . 50.0000 47.9500 0 9617
.
CI0
54 0000 . 50.0000 50.0000 0.0000
55 -1.7969 -1.7969 -1.7969 0 0000
.
56 13.7500 28. 1250 23.5859 1 7707

.
Cj T 196.8750 219. 3750 203.3000 3.3015

co
0.0000 0 0000. 0. 0000 0. 0000
59 0. 0000 0 . 0'000 0.0000 0.0000
31
1
TABLE 2. Spectral Amplitudes Deck Acceleration
lAf I 40 - ACCf EROHElFfS RET ERE MCI' !• 10 ACrfLEROMC lEf* B

l
-- TAi-ir nr awliiuI'E5
bond frcq'iency occel 1 occel 2 o< cel 3 •i* cel 4 occel 5 occel 6 occel 7 •ICC el* 0 occel 9 occel 10 occel 11 occel 12 occel 13 occel 14 occel 13
VARIANCE : 0.000277 0.000234 0.000417 0.000370 0.000260 0.000242 0.000108 0.000220 0.000325 0.000325 0.000204 0.000201 0.000002 0.000002 0.000002
1 9.94784 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
2 9.07254 0.049556 0.04537 0 0 0 0 0 0 0 0 0 0 0 0 0
3 8.27425 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
4 7.54621 0 0 0 0.01617 0 0 0 0 0 0 0.042929 0.042102 0 0 0
5 6.88222 0 0 0 0 0 0 0 0 0 0 0 0 0 0 -1.20177
6 6.27666 -0.00307 -0.06432 0 0 0 0 0 0 0 0 0 0 -0.90054 0 0
7 5.72430 0 0 0 0 0 0 0 0 0.031116 0.031116 0 0 0 0 0
0 5.2207 0 0 -0.27710 0.07744 0.111763 -0.09936 0 0.046334 0 0 0 0 0 0 0
9 4.6134 0 0 0 0 0 0 -0.07 763 0 -0.02097 0.02897 0.093993 -0.07547 0 0 0
to 4.3423? 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
11 6.96031 0 0.349635 0. 1851 0.130356 0 0 0 0 0 0 0 0 0 0 0
i: 3.61104 -0.31024 0 0 0 -0.1372? 0 0 0 -0.05102 -0.05102 0 0 0 0 0
13 3.27404 0 0 0 0 0 -0.17991 0.1 19957 0.120212 0 0 -0.14323 -0. 14924 0.370077 0.303215 -0.32802
1 4 3.0042 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
15 2.73906 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
16 2.49078 0 0 -0.19929 0.248074 -0.09751 -0.1005? 0 0 0 0 0 0 -0.42110 0 -0.37186
17 2.7892 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
18 2.0784 -0.24435 -0.27924 0.059032 0 0 0 0 0 0 0 0 0 0 0 0
19 1.87552 0 0 0 0 0 0 0 0 0. 108172 0.100172 0 0 0 -0.19712 0
20 1.72874 0 0 0 0 -0.12020 -0.07330 0.203703 0.275347 0 0 -0.21213 -0.30208 0 0 0
21 1.57663 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
22 1.437? 0.005030 0 5.558478 5.407030 0 0 0 0 0 0 0 0 0 0.497961 0
23 1.31130 0 0 0 0 0 0 0.337305 0.367365 -0.73223 -0.73223 0.335793 0 0.291323 0 0.246490
24 1.1959? 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
25 1.09076 0 0 0 0 2.3942 2.166553 0 0 0 0 0 0 0 0 0
26 0.99478 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
27 0.90725 0 1.064544 0 0.317226 0 0 0 1.499881 0 0 1.300167 1.410314 0 -3.44955 0.172498
28 0.82743 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
2? 0.75462 0 0 0 0 0 0 0.920562 0 0 0 0 0 0 0 0
30 0.68022 -2.17816 -2.25422 -0.55150 -0.62990 0 0 0 0.035072 -1.97208 -1.97200 0 0.014751 0.693053 0.606495 0.930250
31 0.62767 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
32 0.57244 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
33 0.52207 -0.94077 -0.91274 0 0 0 0 0 0 0 0 0 0 0.941201 0 0.952794
34 0.47613 0 0 -0.23416 -0.24404 0.398616 0.386744 1.530000 1 .38761 0 0 -1.50154 -1.4962 0 0 0
35 0.4J424 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
36 0.39603 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
37 0.36118 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
38 0.3294 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
39 0.30042 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
40 0.27399 -0.13495 -0.11991 -0.03231 -0.03596 0.057960 0.059173 0.380941 0.352899 1 .130804 1.130804 0.398917 0.404177 0.255802 -0.00530 0.287724
41 0.24980 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
42 0.22789 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
43 0.20704 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
44 0.18955 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
45 0.17287 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
46 0.15766 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
47 0.14379 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
48 0.13114 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
4? 0.1196 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
50 0.10908 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
51 0.0??48 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
52 0.09073 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
53 0.00274 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
54 0.07546 0 0 0 0 0 0 0 0 0 0 0 0 0 0 8
55 0.06882 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
t1itlonce :
242 242 363 363 619 619 754 754 1024 1024 1294 1294 404 484 484
32
TABLE 3. Turbulence Properties for Selected Events
TURBULENCE INTENSITIES AND LENGTH SCALES
TURBULENCE LENGTH
TAPE # ANEMOMETER COMPONENT RECORDS INTENSITY SCALE (m)
15 4 U 1-1024 0.4260 16.0
15 5 U 1-1024 0.4631 18.0
15 6 U 1-1024 0.4649 22.0
15 1 U 1-1024 0.3998 23.0

15 5 U 1-1024 0.4631 18.0
15 2 U 1-1024 0.1925 14.0
40 4 U 5-1029
40 5 U 5-1029 0.1132 674.0

40 6 U 5-1029 0.1040 534.0
40 1 U 5-1029 0.1147 451.0

40 5 U 5-1029 0.1132 674.0
40 2 U 5-1029 0.1131 452.0
40 4 W 5-1029 0.0503 74.0

40 5 W 5-1029 0.0582 92.0
40 6 W 5-1029 0.0542 66.0
40 1 W 5-1029 0.0498 85.0

40 5 w 5-1029 0.0582 92.0
40 2 w 5-1029 0.0622 134.0
33
3
V
BRIDGE
>
a
SEDGWICK
"2
0 )
tc
ISLE
DEER
FIGURE
35
tofaZ*
6'-6*b
23.5 ft |
(7.2m)
FIGURE 3. Bridge and Tower Cross Section
56
FIGURE 4. Site Location Map
37
Deer Isle - Sedgwick Bridge
Skyvane 1
A12 A 5
A10 A8 A6 I A4 A2
n r
— —m ,
.
9 — w a m m fnl
u
) .
m
.
m ... m -- —
m |
( )
m
m-
.
m-
~m- . .
u
A11 A9 A 7 A5 A A3 A1
A13
Accelerometer Layout
(Plan)
FIGURE 5. FHWA Instrumentation Layout
38
FIGURE 6. DATA ACQUISITION SYSTEM SCHEMATIC
39
0
081
WIND ANGLE VS. RESULTANT WIND VELOCITY

TOP OF NORTH TOWER
FIGURE 7. Polar Plot of Recorded Wind Events
40
DEER ISLE - SEDGWICK BRIDGE
JANUARY 1382
(DEG)
Direction
ncJ
1
W
(MPH)
Speed
Wind
i irre (Days)
Figure 8. Monthly Summary of Wind Activity

FIGURE 9. Data Processing Sequence
42
0.00 5.00 10.00 15.00 20.00 25.00 30.00 35.00 ( 10.00 45.00 50.00
0.06
G
0.02
IN
RCCEL
0.02
0.06
00
-
111
. TRPE 07 — RCCELEROMTER 9 -- STRRT TIME = 50.112
6.00
CM/SEC.
IN
2.00
VEL.
-
10.00
3.00
1.00
CM.
IN
1.00
DISP.
3.00
TIME IN SECONDS
FIGURE 10. Double Integration Input and Output

TAPE 40 REC. 992-1024 - ACCEL. 9 - 25bpd
NORMALIZED
FREQ MULT
T
O
4
rCflN - 3.3633 xlO*"
VRRIflNCE: - 3.2506x10""
FIGURE 11. Accelerometer Spectral Amplitude

NOIS 310NV 3SVHd danind^v ivdiodds
45
(N+Z/U)
FREQUENCY
NSIONAL
vit
*6
SPECTRA
AVE
ORIGINAL
PTS
4
(N*Z/U)
-
5W
FREQUENCY
ANEMOM.
1-1029
NONDIMENSIONAL
REC.
AO
TAPE
A1ISN30 IVdlOddS lYNOiSNd AIONON
K7
WIND TUNNEL MODELING FOR FLOW AROUND BUILDINGS
BY
Tatsuo MUROTO 1
,
Hisashi OKADA 2 ,
Takeshi OHKUMA 3
ABSTRACT
In Japan since 1965, the wind flow around tall (2) a
buildings has caused many disputes between resi- (3) ap /U
dents who live near tall buildings and building

owners. In these disputes, detailed and accu- 1.2 Field Observation Data
rate wind speed distribution data are requested Three cases of field observation data shown In
from the concerned parties. Table 2 were selected as similarity criteria.

Since the wind speed distribution around tall
Those data are usually obtained by wind tunnel buildings change remarkably according to the
tests, using at hundredth scale, scale-models atmospheric stability condition, field observa-
set in a turbulent boundary layer. Similarity tion data used for comparison were limited to
parameters chosen are: those observed in neutral atmospheric conditions

or in high wind conditions.
a. average wind speed profile;

b. turbulence intensity profile of wind. 1.3 Wind Tunnel Tests
Taking into account that the geometrical simil- Five wind tunnel models were made for tests. They
itude of scale models can not be satisfied per- are scale models of Shinjuku, Tsuklshima and
fectly and that wind profiles in a specific area Ichikawa area corresponding to field observations
can not always be known with sufficient accuracy, stated above. Scale models were as follows:
we must conclude that the accuracy of conven- Sinjuku 1/500, 1/1000

tional wind tunnel test data is not clear quan- Tsuklshima 1/200, 1/500
titatively. Ichikawa 1/300

To solve this problem, the Building Research These wind tunnel models were tested in thirteen
Institute and the Building Center of Japan con- different wind tunnels.
ducted cooperative research from 1981 to 1984.
Twenty-three research groups in Table 1 partici- 2. TEST RESULTS
pated in this project. 2 . 1 Modeling of Exterior Surface of Buildings
A 14 story apartment house located in the en:er
(1)
1. INTRODUCTION of Tsuklshima area has floors pro)e, : <*d l.s-n from
1 . Similarity Parameters Selected an exterior wall horizontally it im. h 1 ** v «• 1 . These

In this research, the following parameters were projected floors were modeled in four types. c >r
selected and their effects on similarity were responding to the difference in modeling, wl id
in vestigated by wind tunnel tests. speed distribution near ground changed sligh;lv
A. Geometric similitude of wind tunnel model
(1) modeling of exterior surface of buildings
(2) modeling of trees on ground surface * Head, Aerodynamic Sec., Structure]
(3) modeling of built-up areas surface around Division, Building Research Institute
9
tall buildings Senior Research Officer, ditto
B. Flow similarity parameters (Fig. 1) 2 Professor, Faculty of Engineering, Ka.i.igawa
R/H University
48
.4
at points near windward corners of the center 2 Wind Velocity Profile
building. When we observed wind speeds U and simultan-

eously at a point i near the ground and reference
2 .2 Modeling Of Trees point respectively, we call the ratio U/Uh as

Field observation wind data in Tsukishima area "wind speed ratio" at point i. When the wind ve-
were subjected to the influence of roadside locity profile of approaching flow is expressed
trees. Six types of model trees were included by an exponential function with exponent 1 /n, it
in a Tsukishima model. Wind speeds along the is noted as wind speed ratio, Ri/ n .
street were measured for each type. Test re- Fig. 3 shows the correlation between R 1 3 ,
r i /6
/
sults lead to the conclusion; similarity of wind and R 1/4 obtained in Tsukishima model for R/H of
speed shapes and solidity ratios of model trees 3 and reference height of 600mm (300mm in full
have poor agreement with prototypes. scale). Presenting the relation between Ri/ n and
Rl /4 as
2 .3 Modeling of Built-up Area Around Tall ^1/n ^1/n R l/4
~
Buildings Coefficient A becomes larger than 1 for n larger

than 4 and becomes smaller than 1 for n smaller
Effects of R/H (Fig. 1) on the wind speed dis- than 4. This means that the wind speed ratio de-
tribution in wind tunnel model center were increases with the increasing exponent of wind
vestigated for Shinjuku, Tsukishima and Ichikawa speed profile.
model. The following conclusions were obtained. On the other hand, the coefficient A changes its
(1) After the wind passed the leading edge of a value when the reference height changes as shown
wind tunnel model, wind velocity profile of in Fig. 4. This figure also shows that if the
approaching flow begins to break at its reference point is set at height between 30 to
lower parts and new boundary layer (inner 70m, the value of A will not change so largely.
boundary) begins to develop in the place. This suggests that setting the reference point at
(2) Thickness of inner boundary layer increases this height, we will obtain the same wind speed
with the distance x from the leading edge distribution independent of the wind velocity
at the rate of 0.05x. profile of approaching flow. This is seen clear-
(3) At the area located in the leeward side of ly in Fig. 5, which shows correlations of wind
the leading edge of wind tunnel models by speed ratios for reference height of 58.1m.
lOh (for Shinjuku model, h is the average Fig. 6 shows wind velocity profiles at some
height of building models) or 25h (for points in Tsukishima and Ichikawa model. In this
Ichikawa model) ,
wind speed distribution figure, each of wind velocity profiles looks sim-
near ground surface is not subject to the ilar at height from 30 to 70m. This agrees with
effect of change in approaching flow pro- the height range stated above.
file. In such areas, wind speed distribu- In the case of Shinjuku model, different features
tions agree very well with those in field were observed. Figs. 7 and 8 show ratios r s /r at
observations points LOC. 10, 17, 7 and 8 (r s : wind velocity
(4) In wide open areas or around buildings pro- ratio for exponent of approaching flow a = 1/4,
jecting higher than inner boundary layer, reference point: top of a high rise building). ©
wind speed distribution does not agree with in the figure corresponds to field observation
those in field observation. In order to get data. LOC. 10 and 17 in Fig. 7 located on the wind
good similarity in those areas, buildings in ward side of high rise buildings and LOC. 7 and 8
the windward city area must be modeled more located on the leeward. According to the figures
widely than lOh or 25h. wind speed ratios at the windward points do not
49
change so largely for a of 1/4 to 1/3 but for a high rise buildings are influenced mainly by the
less than 1/4 it increases with decreasing a . wake of high rise buildings rather than by inner
At the leeward points, on the other hand, wind boundary layer and approaching flow.
speed ratios are nearly equal to field observa-
tion value and do not change for all a . This 2.5 Boundary Layer Thickness
suggests that the wind speed ratio at points The effect of boundary layer thickness on simi-
leeward of high rise buildings is independent of larity is small when the modeling area is suffi-
approaching flow profiles. This is also known ciently large to develop the thick inner boundary
from Fig. 9 showing wind velocity profiles at layer. Therefore, for smaller modeling area ( R/
some point leeward of high rise buildings. H=3^8) ,

the influence of boundary layer thick-
We can derive the following inference from these ness is observed.

facts Fig. 10 shows the correlation of wind speed ra-
The windward area of high rise buildings is cov- tios for boundary layer thickness 6 of 70, 100,
ered by the inner boundary layer developed by and 180m in Tsukishima model. These values cor-
windward buildings and the corresponding wind respond to 6 /H of 1.75, 2.5, and 4.5. From this
speed distribution is realized there. If the figure no special effect of boundary layer thick-
reference point is set outside of the inner ness is observed.
boundary layer, wind speed ratio observed at a Fig. 11 shows that wind speed profiles at a point
point depends on the wind speed profile of ap- leeward of the high rise buildings change largely
proaching wind. This is the same phenomenon ob- when 6 /H changes 0.8 to 2.5.
served in cases of Tsukishima and Ichikawa mod-

els. For the leeward side of high rise build- 3. CONCLUSIONS
ings, wind speed distributions are strongly in- The following are conclusions from the work pre-
fluenced by wakes induced by high rise buildings sented in this report:
rather than by both inner boundary layer and the (1) To expect good similiarity, wind Lunnel mod-
approaching flow. els must involve sufficiently wide area
With regard to the effect of wind velocity pro- around the tall buildings undergoing testing.
files of approaching flow on the similarity of (2) The area of modeling must he enough for the
wind tunnel tests, we obtain the following con- inner boundary layer developed from the lead-
clusions : ing edge to involve the tall building con-
(1) When all the points including reference cerned and the surrounding buildings.
point are involved in the inner boundary (3) The minimum distance of modeling area ( dis-
layer, the wind speed ratios observed are tance from an observation point to the lead-
not influenced by the wind speed profile ing edge) was 10 and 25 times the averaging
of approaching flow. building height in Shlnjuku and Ichikawa
(2) When the reference point is located outside area, respectively.
of inner boundary layer, wind speed ratios (4) When the modeling area is small, the wind
for points near the ground depend on the profile must be made as similar to the rei:
wind speed profile of approaching flow. condition as possible.
Therefore, when wind tunnel models do not Wind speed ratios obtained in wind tunnel tr'.;«
have sufficient area around tall buildings, conducted under experimental conditions al -
the wind speed profile of approaching flow Table 3 and field observation data ire . uspared
must be made as similar as possible to the in Figs. 12 to 14 for refero
real condition. £ and O correspond to the average wind speed
(3) Wind speed ratios in the leeward area of ratio for wind tunnel tests and field nf<«rrva: 1
>
50
respectively, and the length of bars extending
up and downward from the circle correspond to
the standard deviation of wind speed ratios.
These figures show the following:

(1) Shinjuku (Fig. 12)
Wind speed ratios for wind tunnel tests are
larger than field data at several points but
wind speed distribution patterns are in good
agreement
(2) Tsukishima (Fig. 13)
The difference in wind tunnel tests and field ob-
servation is large at point A for wind direction
N and point E for wind direction SW and SSW, but
at the other points both are in good agreement.
A is a special point influenced by trees and E

is a point very near to the corner of the apart-
ment house.
(3) Ichikawa (Fig. 14)
Wind tunnel test results are in good agreement
with field observation data. Larger differences
are seen in case of S and SSW winds. This is
due to the unsimilarity of separation flow at
the building corner.
51
Table 1 List of Research Groups participated
Institution Member Symbol
Osaka City University Eiji Kimoto N
Ohbayashi-Gumi Ltd., Akihisa Kawaguchi A

Technical Research Institute
Kajima Corporation, Hayatoshi Sanada B

Kajima Institute of Construction Technology
Wind Engineering Institute Masaaki Yoshida 0
Kanagawa University Takeshi Ohkuma c

Hisao Marukawa
Kanazawa Institute of Technology Takeo Matsumoto p
Kyoto University, Junji Katsura Q

Diseaster Prevention Research Institute Yoshihito Taniike
Kumagaya-Gumi Co., Ltd., Institution of Technology Hiroki Kashiwagi S

Kazutami Kohashi
Building Research Institute, Hisashi Okada D

Ministry of Construction Tatsuo Murota
Shimizu Construction Co. ,Ltd. Kunio Fujii E

The Research Laboratory
Central Glass Co., Ltd. Tomohisa Ishikawa T
Taisei Corporation, Technical Research Institute Yutaka Asami U
Takenaka Koumuten Co., Ltd., Yoshiteru Iwasa P

Technical Research Laboratory Yoshihiro Mataki
Tokyu Construction Co., Ltd., Nobuo Kato G

Technological Research Institute
Tokyo Institute of Polytechnic Nobuyuki Kobayaahi H

Yukio Tamura
Institute of Industrial Science, University of Tokyo Shuuzo Murakami I
Tokyo Denki University Hiromitsu Kavai J
Toda Construction Co., Ltd. Kyouichi Higaki V
Nippon Sheet Glass Co. Ltd. Saburou Kavabata K
College of Industrial Technology, Nihon University Eizou Maruta L
Hazama-Gumi Co., Ltd. Shuuichi Takakura W
Fujita Corporation, Technical Research Laboratory Yasuyuki Utsumi M

Youichi Shimada
Mitsui Construction Co., Ltd. Kazuo Tsukamoto X
52
Table 2 Description of Field Observations at Three Areas
Name of areas
Item Shinjuku Tsukishima Ichikawa
Address Nishishinjuku 1-Tsukishima Gy out oku

Shinjuku-ku, Tokyo Chuoh-ku , Tokyo Ichikawa-city , Chiba
Observation in and around a city along streets with in a small open

points are area involved trees around lUstory park next to 7story
located 100-200m high apartment house apartment house
buildings
Surrounding city area involves city area involves city area involves
area middle and low rise middle and low rise middle and low rise
buildings buildings buildings
Reference point tops of 2 high rise top of- the apartment top of a two story
buildings, of 237m house of 58.1m high school building,
and 187m high 2km apart from the
observation field.
l6. 7m above ground
Observation
points
number 10 17 12
height 2-9m 1. 5-8m 5m
Anemometer
reference vane type urtra sonic vane type
point type
other points 3-cup type urtra sonic and 3-cup type
3-cup type
Period of 1977.7-1978.1 1978.9-1979.8 1976.10-1977.9

observation i 979 .i 2 -i 98 O.ll
Table 3 Wind Tunnel Test Condition
Parameter Shinjuku Tsukishima Ichikawa
Modeling of trees no Type A no
Extent of modeling
of region roundabout
R 1200m 200m 320m
R/H 6 5 16
Approaching flow
6 220m l40m 270m
6 /H 1.1 3.5 14.3
a 1/3 1/4 1/4
Reference Point
location top of a building top of a building center of turn
table
height 187m 58.1m 16.7m
Model scale 1/500 1/200 1/300
53
Fig.l Schematic Description of Boudary Layer Wind Tunnel Tests
Fig. 2 Layout of Observation Points
n< 4
n>4
A i/n
Fig.U Relation between A.
refered to Wind Speed
at the Gradient Height
»
Rl/3
of High-rize Buildings
55
a a
Fig. 8 Effect of a on Wind Speed Ratio on the Leeward of High-rize Buildings
(Shinjuku)
1 .0 r
Ui/Uo FOR 6 = 100m LOCATION

Fig 10 Effect of Boudary Layer Thickness
o Fig. 12 Wind Tunnel Test Result:' coraj ir" :
on Wind Speed Ratio (Tsukishiraa) with Field Observation Data (Shinjuku
56 .
I.Or 1.0
SW wsw
o o
3
^0.5h 5§0.5r
3 •3 *
.5 5-?
5 .
x*
0 0
v
CAB CsAs E Bs CAAsGBBsE
LOCATION .. LOCATION
I.Or I.Or
SSW N
o o
3
^0.5 . - 5
. 5:05
3 *
5 3 5*
5
*5.5
a
§ .
n n
r
0 -J LI It II
0 -I I I 1 I L_
C A B E GA^B? B C GA E B 2 A 2
LOCATION LOCATION
Fig. 13 Wind Tunnel Test Results compared with Field Observation Data
(Tsukishima)
Ui/Uo
Ui/Uo
Fig.lU Wind Tunnel Test Results compared with Field Observation Data
( Ichikawa)
57
AUTOMATIC DETECTION OF GUST FRONTS WITH A SINGLE DOPPLER RADAR
by
2 1
Hiroshi Uyeda^ , Dusan S. Zrnic' and Hiroshi Takahashi
Abstract associated with gust fronts (Zrnic' and Lee,
A procedure is developed that detects and tracks 1983)
gust fronts automatically. It does not rely on

Although recognition of these signatures by
a single method but requires simultaneous opera-
trained human observers is not complicated, there
tion of two related algorithms. The convergence
are strong reasons to develop automated tech-
algorithm measures radial convergence and hence niques that would help or in some cases even
only gusts propagating along radials can be When several patterns
replace the human operator.
readily detected. The mesocyclone-shear algorithm
need to be examined, quantified, and remembered
measures azimuthal shear and is suitable for
for correlations in space or time, the machines
detecting gusts parallel with radials as well as Furthermore, machine's per-
become superior. a
low-level vortices. Long shear lines that these
formance is consistent regardless of the environ-
algorithms detect are classified as gusts whereas
mental conditions and is not subject to the bore-
symmetric shear features are rejected if their dom and fatigue that affect humans.
shear and flux or "momentum" are insignificant;
With procurement plans for Next Generation
otherwise they are classified as low-level vor-
Weather Radar (NEXRAD) well under way in the
tices. To locate gusts we use second order poly-
United States of America, a need exists to pro-
nomials in the range-azimuth plane. It is shorn
vide as much assistance as possible to interpre-
that predicted gust locations from simple linear
tation of radar displays. So far, several algor-
projections of the least square fitted curves
ithms have been developed and tested to serve
agree very well with actual gust locations.
that purpose (NEXRAD Algorithms Report, 1983).
KEY WORDS: Algorithm; Convergence; Doppler
This paper concerns algorithms that detect and
(spectral moments); Gust (wind engineering); Least
Square (fitting); Momentum; Shear; Vortex. track gust fronts; major part of it is condensed
from a report (Uyeda and Zrnic', 198)1 prepared

1. INTRODUCTION
for the Federal Aviation Administration and the
Early displays of Doppler spectral moments demon-
National Weather Service. Further detail', on
strated that many hazardous weather features pro-
case studies are described in Uyeda and Zrnic'
duce characteristic signatures that human obser-
(1985)
vers can readily interpret (Wood and Brown, 1983).
Among the most dramatic signatures are those of 2. GUST FRONT CHARACTERISTICS FROM DOPPLER
RADAR MEASUREM ENTS
mesocyclonic circulations (Sirmans and Doviak,
1973); tornado vortex signatures (Brown et al, A gust front is associated with a thunderstorm
1978); divergence signatures (Lemon and Burgess, downdraft. Gust fronts and thunderstorm down-
1980; Zrnic' and Gal-Chen, 1983; Doviak and Zrnic' drafts are hazardous to aircraft because of i
1984; Wilson et al . ,
1984 ); and shear lines sudden change in the horizontal wind component
National Research Center for Disaster Prevention, (Fig. 1). The convergence occurs about the
Science and Technology Agency, Japan.
frontal discontinuity because fluids with two
2
National Severe Storms Laboratory, Environmental
different velocities are moving toward each
Research Laboratories, National Oceanic and
Atmospheric Administration, Japan. other. The cool outflow air 1 flowing underneath
58
the warm moist inflow. Ahead of the outflow the terminates, a pattern vector is formed; its six
environmental air has a strong velocity component components are beginning velocity v^, ending
toward the front. velocity v , beginning range r^, ending range

r ,
azimuth, and maximum gradient g defined
An intense front propagating forward or away from e m
by
the radar produces at low levels a very well de-
fined line of convergent radial velocities (Zrnic' v - v
n-4, n+4
gm max ( 1 )
and Lee, 1983). The frontal position is also de- r - r
n+4 n-4^
tectable from the Doppler spectrum widths, which
where v^ is a smoothed (i.e., nine-point aver-
are large at the front, behind it, and sometimes
age) velocity corresponding to a range r
even ahead of it. Alas, the reflectivity field n
between beginning and ending ranges of a pattern
offers almost no clue to the precise position of
vector. We estimate this gradient from data
the front and none whatever about the turbulence
spaced nine points apart because the spacing fur-
ahead.
ther smooths the gradient over the same (matched)
Frontal characteristics have been documented for
length as the one used for averaging the veloci-
several fronts (Zrnic' and Lee, 1983), and a de- ties .
tailed discussion of their dynamics in the lower

Before the pattern vector is saved for further
400 in can be found in the report by Goff (1975).
processing, its gradient defined as
On the basis of their observations and our previ-
ous experience with the divergence algorithm

( 2 )
(Zrnic' and Gal-Chen, 1983), we decided to attempt
detection of the convergence line in the Doppler
and "flux"
velocity field. We have applied the principles
from the divergence algorithm to detect conver- f = Wb - v
e
Hre - r
b
) O)
gence lines instead.
are checked against a set of thresholds. If the
For most orientations of the gust front except
vector fails low gradient L or low "flux"
when it is parallel to a radial, a Doppler radar g
L^ thresholds, it is discarded; otherwise, a
would detect convergence as a decrease in the
check against high thresholds is made. If the
velocities along a radial. Our algorithm capital-
vector passes either high "flux" H^. or high
izes on this simple fact. However, fronts para-
gradient H thresholds, it is saved for later
llel to radials do not produce radial convergence g
sorting. We caution the reader that the term
signatures. To detect such fronts the mesocyclone-
"flux" used here is a carry-over from the diver-
shear detection algorithm (Hennington and Burgess,
gence algorithm. All the thresholds and their
1981) must be used.
values used in the present study are summarized
3. PATTERN RECOGNITION ALGORITHM DETECTION in Table 1.
OF RADIAL CONVERGENCE
After the completion of a scan, all saved pattern
Our gust front pattern recognition algorithm is
vectors are sorted into features according to
modeled after the mesocyclone and divergence
their proximity. Two vectors belong to the same
algorithms (Zrnic' et al., 1984; Zrnic' and Gal-
feature if the range and azimuth centers r
Chen, .1983). Therefore only a brief description
az. satisfy
J
l
is given here. r ’
1 r .
- r . 1 < 1 and 1 az . - az . 1 < 1
The algorithm starts by searching along a radial

1
ci cj 1
r 1
i j
1
a
4)
for runs of decreasing velocities, which character- (
ize radial convergence.

where 1 is a threshold for radial distance and
This search is done by r
comparing consecutive velocities. 1 for azimuthal distance. Features with very

When a run a
59
few pattern vectors are immediately discarded; percent deviation between two adjacent azimuths.
the remaining features are further processed. For these reasons we have added a seven-point look
Features with azimuthal length L, larger by a ahead capability to our vector construction rou-
factor of 2 from average width W, are classified tine. The procedure looks at seven adjacent (in
as fronts (F). Features with smaller azimuthal range) velocities and selects the maximum of the
lengths are classified as shear lines (S) . We minima, which is less than or equal to the current
have often observed that pattern vectors repre- velocity. Thus, the algorithm leaps from a maxi-
senting secondary surges are classified in the mum location to next maximum point and connects
same feature as the primary gust front. This is the segments until there is an increase (over
because pattern vectors are statistical and out more than seven consecutive points) of velociti
distance criterion is liberal. Therefore pattern with range.
vectors in a feature that lie on the same radial
Next we show a typical example of Doppler spectral
(i.e., have the same azimuths) are separated. So
moments along a radial that is cutting through a
if the gust front is approaching the radar, all
gust front (Fig. 2). An increase Ln tivit)
pattern vectors along common radials with centers where the radial is perpen-
is evident in Fig. 2
farthest from the radar are taken out of the When the radial tangent
dicular to the front. is
feature and classified as belonging to a secondary no clear signatures are obtainable
to the front,
surge. Our dealiasing consisted of checking the
from the spectral moments along a radial (a fig-
continuity of radial velocities. We compared a
ure of this is not shown). Otherwise the smoothed
nine-point running average with the first non- region of
mean velocities clearly exhibit a
averaged value outside the nine points and made pattern
large gradient, which is included in
corrections whenever the difference between the indicated
vectors whose beginning and ending is
two exceeded the unambiguous value v (Nyquist
3 . in Fig. 2. It is apparent in this figure that
velocity)
the location of maximum gradient is a better
3 . 1 Construction of Pattern Vectors indicator of the front's position than the mul
The method by which pattern vectors are obtained point of the pattern vector.
in the gust front algorithm is somewhat different
3 . 2 Gradient and "Flux Thresholds
from that used in the mesocyclone algorithm.
Because our detection algorithm relies hesvily Oil
Namely, a nine-point running average of the veloc- gradi-

kinematic properties of gust fronts l i.e.,
ities along a radial is first performed so that
ent and "flux"), we state here some statistical
the effective range resolution is 1.35 km. This
data concerning these quantities that influenced
equals the minimum width of a 2-dBZ gust contour
the choices in Table 1. We had examined the grad
(Zrnic' and Lee, 1983). Averaging reduces the fronts
ient versus "flux" data for several gust
fluctuations due to estimation errors and turbu- the
in order to obtain the detection region in
lence. But even with this smoothing, we observed
gradient - "flux" plane. This region i> also
that deviations of velocities about a linear fit
described by the thresholds in Tabic 1
in range were enough to break up what seems to be

From several examples it became evident that de-
a single, continuous pattern vector.
tection is not very sensitive to small variation
We contrast these results with those of mesocy-
of thresholds about the chosen value >incc the
clone algorithm where comparisons of adjacent
detection algorithm sorts pattern vectors SCCOI I
velocities are sufficient to yield valid pattern

ing to proximity with
vectors. For equal shear the percent deviation
(Table 1), we note that a loss of several vectors
of incremental velocity (due to noise) with res-
does not affect detection. Similarly, false
pect to the mean between two adjacent range loca-
vectors are extremely unlikely to for- elongated
tions is considerably larger than the corresponding
60
features and thus are rejected by the sorting propagation velocity equals the velocity at which
routine. the center is moving, the front's position can be

projected for later times. This is done by ad-
3 . 3 Detection of Azimuthal Shear
vecting the curve (5) at a time t 0
with the
It is possible to detect fronts that are tangent
estimated propagation velocity.
to radials if the mesocyclone algorithm (Zrnic'
et al., 1984) is used. This is because the con- For azimuthal shear lines detected by the meso-
verging winds on either side of the line are not cyclone-shear algorithm, it is natural to regress
perpendicular to the line; the environmental 0 on r because 0 has a quadratic dependence
flow usually has a component along the gust; be- (arc) on r. So we use
hind the gust the flow has a cyclonic curvature 2

0 = Ar + Br + C (6)
so that a cyclonic shear is seen in the mean
velocity display. Anticyclonic shear may also where the data pairs (0^, r^) are the centers
characterize the transition between the environ- of pattern vectors.
ment and the outflow, especially if the wind 4. CASE STUDIES

shear vector progresses counterclockwise with We examined the performance of our algorithm on
height five different cases: April 13, 1981, May 29,
We decided to adapt the mesocyclone-shear algor- 1980, April 26, 1984, May 9, 1981 and April 10,
ithm (Hennington and Burgess, 1981) so that it 1981. The data were obtained by the Doppler radar
can detect azimuthal shear near the ground. Be- (A = 10cm) of the National Severe Storms Laboratory
cause cyclonic shear is prevalent and because at Norman, Oklahoma. In this section, three of
of time constraints, we made modifications to them are explained precisely. As ground truth we
detect only cyclonic shear. The main changes use the positions of the gust as indicated by
needed for low-level shear are the thresholds the velocity discontinuities on color displays and
listed in Table 2. These were obtained after on contour plots. We also show examples of track-
several trials with a limited number of data. ing and outputs such as feature area, "flux", and
gradient as a function of range.

3 . 4 Tracking of Gust Fronts
Once the gust front feature has been identified, 4.1 April 13, 1981
there are several options for displaying and A very intense front developed in a squall line
tracking it. In order to display the front at that passed by some surface stations and over
sites remote from the main computer, it is de- Cimarron on this day (see Fig. 10 in Zrnic' and
sirable to characterize the front's shape with a Lee (1983)). We applied the algorithm to seven
small number of parameters. We have found that scans and chose three consecutive scans at 1.2°
with a convergence algorithm a least-squares in elevation to illustrate detection and tracking.
fitted curve, We superposed the gust front position on the
2 velocity contours to obtain Fig. 3. There the

r = A0 + B0 + C (5)
deduced front's shape follows quite well the
can be used very effectively both to display the -10 m-s 1
velocity contour and is at the locations
front and to extrapolate its position into the
that one would subjectively determine from the
future. In (5) 0's are azimuths of pattern color display. Note that the arc stops at 40 km
vectors and r's can be either the average value
to the west of radar and about 30 km to the north.
(r^ + r )/2, of the pattern vector range or the
From the color display we inferred that the front
range to the maximum gradient of a pattern vector.
extends beyond these ranges, but because it be-
In order to track the front, its range center
comes tangential, ô the radial, it cannot be de-
and azimuth center at two consecutive times t ,
tected with radial convergence measurements.
t are needed. Then, assuming that the front's
2
61
These gusts exhibit height continuity as seen in situation. (2) It had a "line echo wave pattern"
Fig. 4, where the secondary surges (2 and 3) are with a cyclonic couplet. This front is also dis-
found at four elevation angles and the primary cussed by Zrnic' and Lee (1983), who show the
surge is present at all five elevations examined. three moments on their Fig. 7. The gust front
The continuity should extend through two to three algorithm has no provision for circulation detec-
lowest elevations for practical use of the detection, but in an operational setup several algor-
tion method. In order to establish track the ithms that detect and track hazardous weather
front must be detected at least in two consecu- features must be used concurrently. Thus, a com-
tive scans at the same elevation. posite of hazards needs to be presented to opera-
tors for further interpretation and warnings.
Three consecutive positions of the front are
Therefore, we also subjected these data to the
shown in Fig. 5. We used the locations of maxi-
mesocyclone algorithm (Zrnic 1
et al, 1984).
mum gradients to trace the front and perform the
fitting curve. Finding maximum gradients requires As expected, the gust front algorithm detected
an additional pass through the velocity data con- the southern and northern portions of the wave
tained between r. and r of a pattern

r vector, but failed to connect the two. The azimuthal
b c
limits of the radar antenna were such that I
The forecast position of the front (Fig. 5) is in

southern portion of the line was not covered
reasonable agreement with the actual position;
during the first two scans and that is why the
maximum displacements between the two are less
front from 140° to 180° was detected only in the
than 4 km. To improve further the forecast posi-
third scan (Fig. 7). The northern part of the
tions would require intervals shorter than 5-7 min
line tracks as well as in the previous examples,
between scans. The values of average gradient,
with the standard error between the fitted curves
average "flux", average of maximum gradients, and
and the data below 1 km. From the position of
area depend on the development stage of the front.
2 the center at three consecutive elevations we
The area (here 280-430 km ) tends to increase
deduce that the southern portion of the cut had
closer to the radar if elevation is constant be-
a slope of about 7.5°, which is quite close to
cause more pattern vectors that would otherwise
the slope of 8.5° that Zrnic' and Lee (1982) show
fall below the beam can be detected. Average
^
in a cross section of another gust on their I lg.
gradients between 1.5 and 2 m-s • km ^
,
maximum
6. We have found (Uyeda and Zrnic', 1984) that
gradient larger than 5 m-s *-km \ and fluxes
1 the area, average "flux", average gradient, aver-
between 100 and 400 m-s • km are typically en-
4.2 age of maximum gradient, and rms error-, arc also
countered in strong gusts; this is illustrated in
consistent with the previous two examples (Table
Table 3.
3)-
Because the frontal discontinuity on this day ex-
The mesocyclone algorithm 1 meso.
tended also along a radial, we ran the modified
shear (M) between the two lines dig. 7). It
mesocyclone shear algorithm to see how it would
estimated the azimuthal extent (diameter t-> be
detect and connect the azimuthal shear to the
2.3 km and the radial 2.4 km, which to
radial convergence. Results for the three scans
draw the core with the small circle.
are extremely promising (Fig. 6). Gust positions
estimated from convergence and from azimuthal 4 . 3 April 26, 1984
shear join smoothly and even overlap a bit. Because the winds of this front w>r*- . erv t irbu-
lent (Uyeda and Zrnic', 1984) and disrupted

May 29, 1980
flights from and to the Will Rogers \lrp<<rt in
We chose this gust (Fig. 7) for two reasons: (1)
Oklahoma City, we included this case in ur tests
it was moving away from the radar, and we wanted
A TWA flight 163 encountered severe turbulence on
to see how the algorithm would perform in such a
62
approach and had to be diverted to Tulsa. have separately tested the various subsets of a
general low level (less than 3 km above ground)
A meteorological feature prominent on the veloci-
hazard detection algorithm. The flow chart in
ty display was a long (80 km) convergence bound-
Fig. 9 depicts the low level hazard algorithm as
ary (gust front). Strong southerly winds ahead
we envision it. In order to discriminate between
of the gust front apparently identified a low-
gust fronts and short convergence lines, a conver-
level jet associated with the storm system. Winds
gence algorithm developed in this report is used.
were gusting up to 25 m-s ^
at the surface and
*
Calculations commensurate with radial update
gusting higher than 35 m-s at the topmost
rates are required on one radial at a time for
level (444 m) of the KTVY instrumented tower.
constructing pattern vectors that are later (at
Doppler velocities confirm that strong flow ex-
the end of a scan) sorted into features. Our
tended to the aircraft altitude (2 km and beyond).
convergence algorithm measures the radial conver-
Doppler spectrum width display showed large
gence, and hence only gusts propagating along
widths at the frontal discontinuity as well as
radials can be readily detected. For detecting
ahead of the tower. This was confirmed by high
gusts that are aligned along radials, we have
variability in the recorded tower winds, which
employed the mesocyc lone-shear algorithm, which
^
change by as much as 14 m-s in 10 s.
requires storage and operation on two consecutive
The frontal outflow was rather shallow (<1 km) and radials of velocity data. The long shear lines
the discontinuity could be seen clearly only at that this algorithm detects are classified as
0.5° in elevation (about 0.3 km height at a range gusts whereas symmetric shear features are re-
of 40 km). At the next elevation, 1.5°, it was jected if their shears and momentum are insignif-
not visible. The shape of the primary front de- icant or are classified as low-level vortices
picted by a five-point running average of maximum that are known to accompany wavelike gusts.
gradient location in Fig. 8 reproduces very well Shears of intermediate length (10 to 20 km) may
the pattern seen on color displays (Uyeda and be classified as spurious.
Zrnic', 1984). The forecast position agrees with
Downdraft and tornado vertex signatures (TVS)
the actual location of the front even though the
represent low level hazards that can be detected
forecast center and the actual center are dis-
with the same basic technique of building pattern
placed by about 10 km. Their displacement is
vectors, and that is why they are included in the
along the frontal discontinuity and thus does not
flow chart (Fig. 9) even though they have not
affect the front's position in range but only
been developed yet. Other low-level hazards such
influences its azimuthal extent. Tabulated param-
as turbulence and hail require spectrum width
eters are generally consistent with previous
and reflectivity data as well as different algor-
cases. The rms error is on the high side because
ithms for identification; that is why they are
the front is long and has kinks; therefore, a
not included in the chart.
second-order polynomial is not a best model for
After sorting of pattern vectors, the resulting
its shape.
features are classified and tracked. It appears
5. SUMMARY AND CONCLUSIONS
that convergence lines (i.e., lines obtained from
The procedure we have developed to detect and
the convergence algorithm) shorter than 10-20 km
track gust fronts automatically does not rely on
can be discarded. Distinguishing by length and
a single method but requires simultaneous, opera-
symmetry between hazardous and spurious features
tion of several related algorithms. These algor-
is not unique with the mesocyclone-shear algorithm
ithms process one or two radials of Doppler veloc-
because mesocyclones are symmetric and of small
ity data in real time. We have not incorporated
extent as are spurious shear lines. Therefore,
in our programs the simultaneous operations but
checking the length and continuity in time may
be a most effective means to eliminate spurious (National Weather Service) and by the Federal
shears Aviation Administration Grant No. DTFA01-8-Y-

10524. One of the Authors (Hiroshi Uyeda) was
Overall detection of gust fronts from measurements
supported by the Science and Technology Agencv,
of radial velocity gradients and azimuthal shear
Japan and Cooperative Institute for Mesoscale
is feasible. The results exceeded our expecta-
Meteorological Studies for his stay at the
tions; there was no undetected front among the
National Severe Storms Laboratory.
test cases. However, spurious shear and false
7. REFERENCES
alarms due to overlaid echoes were inevitably
present. These can be significantly reduced Brown, R. A., L. R. Lemon, and D. W. Burgess,
1978: "Tornado detection by Doppler radar,"
by checking their length as well as time-height Mon. Wea. Rev., 106, 29-38.
continuity. Doviak, R. J., and D. S. Zrnic', 1984: Doppler
Radar and Weather Observations. Academic
Here, we discuss the applicability of developed Press, New York, NY.
algorithms and algorithms shown in Fig. 9 to Goff, R. D. 1975: "Thunderstorm outflow kine-
observations of Japanese weather phenomena. The matics and dynamics," NOAA Tech. Memo IRl.-
NSSL-75. Available from National Technical
heavy rainfall associated with Baiu and Typhoon Information Service, Operations Div., Spring-
along the seacoast of the Pacific Ocean and the field, VA 22151.
Hennington, L. and D. W. Burgess, 1981: "Auto-
heavy snowfall and snowstorm along the seacoast ,
matic recognition of mesocyclones from

of the Japan Sea are severe weather phenomena in single Doppler radar data," Preprints, 20th
Japan. For detection and forecast of them, ob- Radar Meteorology Conference. American
7
Meteorological Society, Boston, pp. 7 01- 06.
servation of wind field is essential.
Lemon, L. R., R. J. Donaldson, Jr., D. K. Burgess,
When we apply a Doppler radar for low-altitude and R. A. Brown, 1977: "Doppler radar appli-
cation to severe thunderstorm study and poten-
weather observation in mountainous Japan, we have tial real-time warning," Bulletin, Am. Meteor.
to think about proper radar wavelength. A 3 cm Soc., 58, 1187-1193.
radar would not be sufficient for the observation and D. W. Burgess, 1980: "Magnitude and
,
implications of high speed outflow at severe

of heavy rainfalls and snowfalls because of at- storm summits." 19th Conference on Radar
tenuation. A 10 cm radar would not be practical Meteorology, American Meteorological Socictv,
Miami, Florida, pp. 364-368.
for low altitude observation in Japan because
NEXRAD (Next Generation Weather Radar) Algorithm
of ground clutter, however it is very powerful Report, 1985: Prepared by the NEXRAD Joint
for the observation of large area. A 5 cm System Program Office, Gramax Building,
Room 900, 8060 13th, Silver Spring, MI) 20910.
Doppler radar would be appropriate for the obser-
Sirmans, D. and R. J. Doviak, 1973: "Pulsed-
,
vation of Japanese severe weather. Donpler velocity isotach display of storm

winds in real time," J. Appl. Meteor., 12.
For understanding of weather effects to building, 694-697.
it would be very useful to observe low altitude Uyeda, H. and D. S. Zrnic', 1984: "Automatic
,
wind field by a Doppler radar. It would also be detection of gust fronts." Report prepared
for the Federal Aviation Administration and
necessary to examine wind shear distribution and the Joint System Program Office. National
spectrum width distribution by a Doppler radar Severe Storms Laboratory, 1313 Halley Circle,
Norman, OK 73069.
for the understanding of wind effects.
Uyeda, H. and D. S. Zrnic', 198S: "Automata
,
6. ACKNOWLEDGEMENTS tection of gust fronts." J. Atmospheric and

Oceanic Technology. (To be published.
We appreciate the help of Yosefa Gal-Chen during
Wilson, J. W. R. D. Roberts, D. Kcssinger, and
,
the initial stages of algorithm development. Don J McCarthy, 1984: "Microburst wind Structure
.
Burgess provided us with the information concern- and evaluation of Doppler radar for airport
wind shear detection," J. of Clim. and Appl.
ing the weather situation on April 26, 1984, when Meteor., 23, 898-915.
TWA flight 163 encountered turbulence. This work Wood, V. T. ,
and R. A. Brown, 1985 "'-ingle
was supported by the Joint System Program Office Doppler velocity signatures: \n atlas of
to
patterns in clear air/widespread precipitation Zrnic', D. S. and J. T. Lee, 1983: "Investigation
and convective storms," NOAA Tech. Memo. ERL of the detectability and lifetime of gust
NSSL-95. Available from National Technical fronts and other weather hazards to aviation,"
Information Service, Operations Div. Spring- ,
FAA Report No. DOT/FAA/PM-83/33
field, VA 22151 Zrnic', S., D. W. Burgess, and L. D. Hennington,
D.
Zrnic '
D. S.
,
and J. T. Lee, 1982: "Pulsed Dop-
,
1984: "Automatic detection of mesocyclonic
pler radar detects weather hazards to avia-
1
- shear," accepted by the J. Tech.
tion," J. Aircraft, 19, 183-190.
Zrnic', D. S., and Y. Gal-Chen, 1983: "Divergence
measurement in storm tops," Report for the
Joint Systems Program Office. National
TABLE 1
Severe Storms Laboratory, 1313 Halley Circle,
Norman, OK 73069. Thresholds for Detecting Radial Convergence Lines
THRESHOLD NUMERICAL VALUE
1 1
High Gradient Hg = 0.65 m-s *km
1 1
Low Gradient Lg = 0.50 m*s *km
High "Flux" -1,
Hf = 100 m-s km
Low "Flux" _1 -km
Lf = 30 m-s
Max. azimuthal separation T - 4°
J-a **
between vectors
Max. range separation lr = 4.0 km
between vectors
Reflectivity Threshold -15 dBZ
Height Threshold 2 km (and below)
TABLE 2
Thresholds for Detecting Azimuthal Shear Lines
THRESHOLD NUMERICAL VALUES
1
Ls Low shear 1.5 m*s -km
Hs High shear 2.0 m s” -

*km
Lm Low momentum 25 nrs" -km
Hm High momentum 50 ra*s" *km
M Minimum number of vectors 10

allowed in a feature
la Maximum distance between the 3.3°
azimuthal centers of two vectors
for classification in the same feature
lr Maximum radial distance between 4 km
two vectors
Reflectivity threshold -15 dBZ
Zt
TABLE 3
Minimum, Average, and Maximum Value of Gust Front Parameters
Range of Avr of
Center Length Area Avr. Grad. Max. Grad. Avr. “Flux" RMS Error
2 1 1 -1 1
(km) (km) (km ) (ms" , km" ) (ms * km" ) (m3"'*km) (km)
DATE min. avr. max
,
min avr max
, , min, avr, max min, avr, max min, avr, max min avr max
,
, min, avr, max
April 13, 1981 25, 28, 39 12, 30, 44 280, 340, 4 30 1.3, 1.6, 1.9 5.3, 5.9, 6.5 83, 255, 365 0.8, 1.3, 1.5
May 9, 1981 47, 68, 83 13 48, 63 100, 430, 760 1.4, 1.7, 2.1 4.6, 5.2, 6.4 79, 231, 321 0.8, 2.1, 3.3
May 29, 1981
April 10, 1981*

April 26, 1984
24,
52,
41
34,
55,
45,
48
57
49
12,
70,
77, 85
25,
73,
33
79
91
200,
—317,
840, 965,
480
1075
1.5,
4.6,
2.5, 2.6,
2.0, 3.0
5.0, 5.5
2.6
—
8.4, 9.1, 9.9
9.0, 9.4, 9.7

108,
66,
351,
,
,
335 , 438
63,
383
68
, 437
0.2, 0.6,
1.2, 1.5,
1.6, 2.0,
1.0
2.2
2.6
This front was detected with the mesocyclone-shear algorithm; the rest were detected with the
convergence algorithm
65
Figure 1. Schematic of a thunderstorm downdraft and associated gust front
on the approach path to an airport. Note the sudden change in the
horizontal wind component at the distance of about 11 km. In particular
cases and at particular stages in the life of a storm, the horizontal scale
of the disturbance may be substantially smaller or larger. (From Zrnic and
Lee (1983))
Figure 2. Profiles of equivalent reflectivity factor, mean velocity and

Doppler spectrum width along radials for a gust of May 9, 1981. Both raw
and nine-point averaged mean velocities are shown. The spectrum -
a nine-point running average. The beginning, B, and endinq, E, of tha
pattern vectors are indicated.
66
2123 CST, 13 APR 1981
EL = 1.2°
o
Z 20
8
4
\ iiuu l
IIHUU L I I 1 1 J-LJ l i lll U HI
l klllLL ' i ' 1 1 1 1
60 AO 20 0
RANGE (km) Figure 5. Three consecutive
Figure 3. Contours of mean Doppler positions of the front (April 13,
velocities for the gust front of 1981). The smooth curves are least
April 13, 1981. The contour interval squares fits to the data, and the
_1 forecast position is indicated. The
is 10 m*s Objectively determined
.
location of the front (thick line) is range mark is at 40 km from the radar.
superposed. The square next to C is Centers of the pattern vectors were
at the center (r c az c ) of the front. ,
used to trace the gust, squares indi-
cate the centers of these three front
lines.
Figure 6. Detected gust fronts.

Dashed line is a five-point running
Figure 4. Detected gust lines in one average of maximum gradient locations
volume scan of April 13, 1981. (from convergence algorithm) and the
solid line is a five-point running
average of pattern vector centers
(from mesocyclone-shear algorithm)
67
Figure 7. Three consecutive
positions of the front on May 29, 60 60 40 ?o
1980. The smooth curves are least RANGE (km)
squares fits to the data, and the

forecast position is indicated. The
range mark is at 40 km from the Figure 8. Three consecutive
radar. The small circle indicates positions of the front for April 26,
the core of a mesocyclonic vortex, 1984. Five-point running average,
and the dashed circle has a diameter least squares fitted curves, and a
three times larger than the core. forecast position are shown. Square
indicates the center of each curve.
Figure 9. Algorithm for detecting

weather hazards near the ground.
68
Theme II
EARTHQUAKE ENGINEERING
EARTHQUAKE HAZARD REDUCTION
PROGRAM OF THE DEPARTMENT OF STATE
BY
PETER E. GURVIN
AB STRACT Perhaps the most significant aspect of

the Department’s facility posture is the
The Department of State owns or leases a sheer magnitude of these aspects. They
large inventory of buildings throughout total nearly 10,000 buildings covering 46
the world some of which are located in million square feet in 256 cities in the
areas of high seismic risk. The world. These facilities are worth 3
Department has initiated a seismic hazard billion dollars. The percentage mix of
reduction program to upgrade the seismic our buildings are 23 percent U.S.
resistance of our buildings through Government owned and 77 percent short
either replacement or retrofit. The term lease.
seismic design requirements for our major
buildings may have been overtaken by a A typical post consists of three
greater design condition that relates to office buildings, five "other" community
the terrorist threat! and support facilities, and 30
residential buildings ranging from the
1. INTRODUCTION Ambassador’s residence to 25 individual
homes
The Office of Foreign Buildings
Operations of the Department of State was 2 - HISTORICAL EARTHQUAKE DAMAG E
created by the Foreign Service Buildings
Act of 1926 and is responsible for the It may be of interest to review the
design, construction, acquisition, and past historical performance of tho
maintenance of our diplomatic and Department of State’s buildings during
consular facilities overseas. past earthquakes Many thousands of
people lost their lives in those
devastating earthquakes.
The following is a summarization of performance
of the
Embassies and Consulates as far as known:
DATE LOCATION MAGNITUDE COMMENT
9-1-1923 Tokyo M-7 . Embassy buildings destroyed by

f ire
See Appendix A
3-1-1931 Manaqua M= American Legation destroyed
7-29-1967 Caracas M-6 . Non structural damage to

Embassy
12-23-1972 Managua M-6 . Embassy damaged beyond repair
2-4-1976 Guatemala CityM-7 . Non structural damage
3-4-1977 Bucharest M-7 . Non- structural damage
6-20-1978 Thessaloniki M-6.5 Non structural damage
3-3-1984 Santiago M-7 .4 Non structural damage
1) Chief, Civil Structural Engineering,

Office of Foreign Buildings,
Department of State. Washington, DC
70
3. DEPARTMENT OF STATE SEISMIC DESIGN CRITERIA and Dames /Moore (2) to zone all Foreign
Service Posts according to the 1976
Uniform Building Code format. This
The National Earthquake Hazards global zoning has proved to be very
Reduction Program, developed in useful to obtain a first-cut on the
accordance with the National Earthquake seismic design requirement for new design
Hazards Reduction Act of 1977 (PL95-124), of buildings for various cities
stated that this program "should proceed worldwide. More in-depth seismic zoning
on the basis of a balanced strategy would be accomplished on a case by case
reflecting relative seismic risk and basis depending on the size of the
mitigation of highest hazards on a project. Secondly, this global zoning
priority basis". "A Federal example in provided a basis for preliminary
construction: the Federal Government recommendations for action toward
should set a strong example in the reduction of the earthquake risk of
construction of its own facilities and existing buildings in FBO's inventory.
develop guidelines and standards for
Federally assisted or licensed The preliminary seismic-exposure
facilities .
estimates as given by Dames/Moore and

Wooward-Clyde are summarized in the
This Federal Law was primarily Appendix B. These estimates place
directed towards internal U.S. Government approximately 15 percent of the cities
Agencies. However, the Department of with Foreign Service Posts in zones
State believed that to the maximum extent higher than zone 3 and over 40 percent in
possible the intent of the program should zones higher than 2. In general, the
be incorporated into the management of zoning appears to be conservative. As an
the State Department's design and example, let's look at the zone
construction program outside of the U.S. recommendation for Lisbon which was given
as zone 4. Karnik (3) indicates that
First of all, I would like to give damage evaluated at MM1 IX (Modified
you some insight into how our design Mercalli Intensity) has been observed in
program for major office and staff the region which provides some support
housing is accomplished. All major toward a rating of 4. Sozen (4) noted
designs, with a few exceptions, are that in a study of comparative seismicity
accomplished by U.S. design firms. Our as given in Figure 1 (5), the seismicity
basic structural design guidance, of Portugal is only a fraction of that of
including earthquake design, refers to California. Seismic zone 4 is assigned
the current Uniform Building Code. to areas within seismic zone 3, depending
Almost 99 percent of our buildings are of on the proximity to certain major fault
reinforced concrete so the American systems. Seismic zone 3 is assigned to
Concrete Institute Standards also form a areas where major damage corresponding to
fundamental reference for design. MM Intensity VII or higher can occur.
An exception to the above is the recently
completed U.S. Embassy housing project at FBO recently completed a new Chancery
Mitsui Compound in Tokyo which was in Lisbon. At the beginning of the
structurally designed by the Japanese design we took a more in depth look at
firm of Kirmura Structural Engineers. In the seismic estimate for Lisbon. This
addition to U.S. design standards, the study (6) based on the seismicity of the
construction plans and specifications region and evaluation of local faults
have to take into consideration the resulted in the recommendation to use
country where the facilities are to be zone 3. FBO has made some other site
built and therefore relate to available specific studies which also resulted in a
materials, level of construction lowering of the zone estimates as given
expertize, and also the construction in the Appendix.
norms of the country.
It should be noted that further specific
4. EARTHQUAKE HAZARD REDUCTION PROGRAM
site studies are useful only if their is
a real payback for the reduction in
The initial stumbling block for
zoning. The reduction in a zone from 4
overseas building designs for FBO was to
to 3 may have a nominal effect on the
obtain a proper zoning according to the overall cost of the building, and the
zones used in the Uniform Building Code
design could be carried out for the
as used in the U.S. based on zones 0
initial assessment. In any event the
through 4. In order to obtain an idea or
preliminary assessment reports have
gross zoning for the world, FBO
proved useful in the structural design of
contracted with the soil consulting
new buildings.
engineering firms of Woodward-Clyde (1)
71
The second purpose of the seismic Our building program has dramatically
zoning report was to obtain a basis for been enlarged, Congress authorized 13 new
development of a list indicating which office buildings this year at a cost of
buildings at our 256 Posts should be $170 million, mainly in the Middle East.
investigated on a priority basis. The In addition, a special Panel on Security
steps in the preliminary process were for the Secretary of State has been
defined as follows: recently quoted in the Washington Post
newspaper as recommending that the
Step 1 Relative Seismic Department replace over 140 Embassies at
Exposure a cost of 3.3 billion dollars.
Step 2 Occupancy
Step 3 Vintage and Location In conclusion, the earthquake threat
Step 4 Classification as a governing design parameter may have
According to Building Material been overtaken by the terrorist threat
Step 5 Preliminary Structural with regard to State Department
Evaluation facilities .
In our hazard reduction

program it was decided to
concentrate on the high occupancy
buildings, which would be the
Chanceries and office annexes.
Since 1982, our office has
1. "Seismic Exposure Assessment", report
received about $85,000 per year
for preliminary structural
prepared for Department of State, Office
of Foreign Building Operations by
Woodward- Clyde Consultants, San
Francisco, California. June 1979.
As a result of these evaluations, our 2. "Seismic Exposure Assessment", report

office has used the outcome to prioritize prepared for Department of State, Office
of Foreign Building Operations by
Dames
seismic strengthening or make
recommendations for total replacement of and Moore, Cranford, New Jersey.
buildings. Currently, new Embassy Office June 1979.
building designs are about to begin or
are under design for these locations of 3. "Seismicity of the European Area",
high seismicity; Caracas, Damascus, V. D. Reidel Publishing Co.,
Kamik,
Dhaka, La Paz, Nicosia, Osaka-Kobe, Dordrecht, Holland. 1969 (two volumes).
Ottawa, Rangoon, and Santiago. While
$85,000 is not a large sum the seismic 4. "Review of Seismic - Exposure
reduction program has provided valuable Estimates for Foreign Service Costs and a
input for justification by management for Preliminary Plan for Identification of
new projects where existing buildings Hazardous Buildings", report preparod for
have proved to be totally deficient. the Department of State, Office of
Foreign Building Operations, by Mete
5. CURRENT CAPITAL PROGRAM IMPACT A. Sozen, University of Illinois.
August 1979.
As a prologue to our prior discussion
on seismic hazard reduction program, the 5. "Structural Safety", V. Ferry Borges
Beruit bombing of July 1983, the Kuwait and M. Castanheta, Laboratoria Nacional
Chancery bombing of September 1983, and de Engenharia Civil, Second Edition,
the Beruit Chancery Annex bombing of Lisbon, Portugal. March 1971 pp 235 239
.
September 1984 have changed our

structural design guidance for new office 6. "Seismic Considerations for New U.S
buildings and also the size of our Embassy, Lisbon", report preparod by
building program. 87 persons lost their Woodward- Clyde Consultants, New York,
lives in these terrorist attacks and of NY. November 1979.
course, many more were severly injured.
Among other design initiatives being
implemented as a result of the terrorist
threat are that all new office buildings
are designed for Uniform Building Code
Zone 3 earthquake design regardless of the
location. Blast loading is a different
loading condition than earthquake loading
but ductility in structural design is
important to both situations.
72
APPENDIX A
Teli ye
Historical Sketch of the Embassy Property
The present Diplomatic and Consular buildings of the United States replace wooden
buildings that were used by the United States Legation and Embassy froa 1889 until they
were destroyed by earthquake and fire on Septeaber 1, 1923. There were three houses on
what is now the lower compound, which was all that the United States held at the tiae ot
the disaster. One was a two storied fraae building in which the Ambassador had his
residence. To it was attached the chancery which occupied a wing extending towards the
Reinanzaka hill road. The archives were housed in a brick vault in the chancery. This
vault withstood the fire. The aain entrance at that tiae was by a gate on the north side
near the Reinanzaka road. There was a gravel roadway to the house which stood on an oval
and a flower bed. At the eastern end of the oval was a bungalow usually occupied by the
Counselor. In the southeastern part of the property was a house built by Mr. R. S. Mille
for his own use when he was Japanese Secretary - a grade now specially provided for.
This was a two-story structure and was approached froa the street by a separate driveway
which ran ran behind the Counselor's house. The property sloped down this driveway to
the street and near the gate of the back driveway was a sort ot stable and storehouse.
The property was formerly a graveyard. It was purchased about 1888 by Baron
Okura, the merchant, founder of Okura Guai, who erected the aain building on the under-
standing that it would be used by the Aaerican Minister (as he then was) as an office
and residence. The bungalow was erected subsequently. In 1889 the Japanese Government
purchased the property froa Baron Okura and leased it formally to the Aaerican Legation.
It was the property of the Japanese 6overnaent until 1696, when the United States
purchased the buildings and took the land on perpetual lease. The consideration was
$16,000. It was later discovered that a saall piece of land at the northern corner of
the Reinanzaka hill road had not been included in the perpetual leasehold. As perpetual
leaseholds were no longer permissible after 1899, this property has been held as an
ordinary lease. The Consulate 6eneral was aoved to Tokyo froa Yokohama after the earth-
quake of 1923.
The old buildings were painted white Ind the property was separated froa the
street on the north by a low wall suraounted by a hedge. The garden was well kept. A
large wisteria vine was trained over the arbor at the back of the Ambassador's house
where there was a covered porch. There were trees and bushes, the latter aostly flower-
ing, and the grounds had a distinctly pleasing .appearance.
The "upper compound" at the top of the hill where the Ambassador's house now
stands belong to the Iaperial Household and was coaaonly known as the "Ito Property"
froa the fact that the first Prince Ito lived there when hl a 8 onnected with the Iaperial
_
Household. There was a brick house on the lot which was destroyed in thV earthquake.
The Iaperial Household sold the property to the United States in 1925.
73
APPENDIX B
SEISMIC EXPOSURE EVALUATION TABLES
A. Is SCOPE: This listing Is a preliminary evaluation of seismic exposure

of foreign service posts throughout the world where the U. S. Department of
State has the responsibility for owned and/or leased buildings. In this
listing, seismic exposure is defined as the estimated level and likelihood
of seismic effects of earthquakes in the area expressed in terms of seismic
zones as defined in the Uniform Building Code of 1979.
A. 2 BASIS OF EVALUATION: This evaluation is based on available informa-

tion on the regional tectonic condition and an assessment of the historical
seismic activity in the area of each location. Where the available informa-
tion was incomplete, it was supplemented by the judgement of experts in
geology, seismology, and earthquake engineering based on their work on
research and projects in the respective regions. An indication of the
confidence in the assessment exposure is included in the classification of
sites into zones. The evaluation also included identification of the latest
local earthquake code and a listing of relevant seismic activity.
A. 3 SUMMARY: A summary of the results of the seismic exposure evalua-

tion is given in the table following this section.
DEGREE OF
COUNTRY CITY ZONE CONFIDENCE
AFGHANISTAN Kabul 4 M
CHILE Sant 1 ago 4 L-M
CHINA Peking 4 M
CHINA Tihwa 4 M
COLOMBIA Bogota 4 M
ECUADOR Quito 4 M-H
EL SALVADOR San Salvador 4 M
GREECE Thessaloniki 4 L
GUATEMALA Guatemala 4 H
ICELAND Reckjavik 4 M-H
INDONESIA Bandung 4 L
INDONESIA Jakarta 4 L
INDONESIA Surabaya 4 L
IRAN Tabriz 4 H
JAPAN Okinawa, Naha 4 M-H
JAPAN Osake-Kobe 4 M-H
JAPAN Tokyo 4 H
JAPAN Yokohama 4 H
NEPAL Kathmandu 4 M-H
NICARAGUA Managua 4 H
PERU Lima 4 M
PHILIPPINES Cebu 4 M
PHILIPPINES Manil 4 M
PORTUGAL Ponta Delgado 4 M
SICILY Palermo 4 M-H
TIAWAN Taipei 4 M-H
TURKEY Istanbul 4 M
TURKEY Izmir 4 M
CANADA Quebec 3+ M-H

COLUMBIA Buenaventure 3+ L-M
COLUMBIA Cal i 3+ L-M
JAMAICA Kingston 3+ M-H
NEW ZEALAND Well ington 3+ M-H
PAKISTAN Isl amabad 3+ M
PERU Piura 3+ L-M
VENEZUELA Caracas 3+ H
ALBANIA Tirana 3 H
ALGERIA Algiers 3 L-M
ALGERIA Oran 3 L-M
ANTILLES Curacao 3 M
BANGLADESH Dacca 3 L
BOLIVIA Lapaz 3 L-M
BULGARIA Sofia 3 M
BURMA Mandal ay 3 L
BURMA Rangoon 3 L
75
DEGREE OF
BURUNDI Bui umbura 3 L

CANADA Vancouver 3 M-H
CHINA Mukden 3 L-M
COLUMBIA Medellin 3 L-M
COSTA RICA San Jose 3 L
CYPRUS Nicosia 3 M
DOMINICAN REPUBLIC Santo Domingo 3 M
ECUADOR Guayaquil 3 M
ETHIOPIA Asmara 3 L
FIJI Suva 3 L-M
FRANCE Marseille 3 L-M
FRANCE Nice 3 L-M
FRENCH WEST INDIES Mart inique 3 M
GREECE Athens 3 L-M
HAITI Port au Prince 3 M
HONDURAS Tegucigal pa 3 M
INDIA Bombay 3 L-M
INDIA New Delhi 3 M
INDONESIA Medan 3 L
IRAN Isfahan 3 H
IRAN Shiraz 3 H-H
IRAN Tehran 3 M-H
ISRAEL Haifa 3 M
ISRAEL Jerusalem 3 M
ISRAEL Tel Aviv 3 M
ITALY Florence 3 M
JAPAN Fukuoka 3 M-H
JAPAN Sapporo 3 M
JORDAN Amman 3 M
LEBANON Beirut 3 M
MALAWI B1 ant yre 3 L
MALAWI Lilongwe 3 L
MALAWI Zomba 3 L
MALTA Valletta 3 L
MEXICO Guadal ajara 3 M
MEXICO Mexico City 3 M
MEXICO Tijuana 3 M-H
MOROCCO Tangier 3 M
NEW ZEALAND Auckl and 3 M-H
PANAMA Colon 3 M
PANAMA Panama 3 M
PAPAU NEW GUINEA Port Moresby 3 L
PHILIPPINES Baguio 3 M
PORTUGAL Lisbon 3 M
PORTUGAL Opporto 3 L
RUMANIA Bucharest 3 M
RWANDA Kigal 3 L
SAIPAN Saipan 3 l
SOUTH AFRICA Capetown 3 M-H
SOUTHERN RHODESIA Sal i sbury 3 L
76
DEGREE OF
SYRIA Aleppo 3 L-M

SYRIA Damascus 3 L-M
TERR. OF AFARS AND ISSAS 3 L
TRINIDAD Port-of-Spain 3 M
TUNISIA Tunis 3 L-M
TURKEY Adana 3 L-M
YUGOSLAVIA Zagreb 3 M
ZAIRE Bukavu 3 L
AUSTRIA Vienna 2+ M
CANADA Montreal 2+ H
CHINA Tsingtao 2+ M
CZECHOSLOVAKIA Bratislava 2+ L-M
FRANCE Strasbourg 2+ M
GERMANY Stuttgart 2+ M
GHANA Accra 2+ L
HUNGARY Budapest 2+ M
IRAQ Baghdad 2+ L-M
ITALY Genoa 2+ L-M
ITALY Naples 2+ L-M
ITALY Trieste 2+ L-M
PAKISTAN Peshawar 2+ M
YUGOSLAVIA Belgrade 2+ M
ADEN Aden 2 L
BARBADOS Bridgetown 2 M
BELGIUM Brussels 2 M
CANADA Hal if ax 2 M-H
CANADA Ottawa 2 H
CANADA St. Johns 2 M-H
CHINA Canton 2 L-M
CHINA Nanking 2 L-M
CHINA Shanghai 2 L-M
CUBA Santiago de Cuba 2 M
ENGLAND London 2 L
EGYPT Alexandr ia 2 L-M
EGYPT Cairo 2 M
EGYPT Port Said 2 L
ETHIOPIA Addis Ababa 2 L
FINLAND Helsinki 2 M-H
FRANCE Bordeaux 2 M
GERMANY Bonn 2 M
GERMANY Frankfurt 2 L-M
HONG KONG Hong Kong 2 L
INDIA Calcutta 2 M
ITALY Mil an 2 L-M
ITALY Rome 2 L-M
ITALY Turin 2 L-M
KENYA Nairobi 2 L
LIBYA Tr ipol i 2 L
MEXICO Ciudad Juarez 2 M
77
DEGREE OF
MEXICO Hermosillo 2 M
MOROCCO Casablanca 2 M
MOROCCO Rabat 2 M
NORWAY Oslo 2 L
PAKISTAN Karachi 2 M
PAKISTAN Lahore 2 M
POLAND Krakow 2 L-M
SAUDI ARABIA Jeddah 2 M
SOUTH AFRICA Johannesburg 2 L
SOUTH AFRICA Pretoria 2 L
SPAIN Barcelona 2 L-M
SPAIN Bilbao 2 L-M
SPAIN Seville 2 L-M
SWEDEN Goteborg 2 M-H
SWITZERLAND Bern 2 M
SWITZERLAND Zurich 2 M
TURKEY Ankara 2 M-H
UGANDA Kampal a 2 L
VENEZUELA Maracaibo 2 M-H
YEMEN Sanaa 2 M
ZAIRE Lubumbashi 2 L
ZAMBIA Lusaka 2 L
AUSTRIA Sal zburg 1+ M

DENMARK Copenhagen 1+ L-M
MEXICO Mazatl an 1+ M
AUSTRALIA Canberra 1 M
AUSTRALIA Melbourne 1 M
AUSTRALIA Perth 1 M
AUSTRALIA Sydney 1 M
BELGIUM Antwerp 1 M
BOLIVIA Santa Cruz 1 L-M
CANADA Calgary 1 M
CANADA Toronto 1 M-H
CANADA Winni peg 1 M-H
COLUMBIA Barranquil 1 a 1 L-M
CUBA Havana 1 M
CZECHOSLOVAKIA Prague 1 L-M
ENGLAND Edinburgh(Scotl and) 1 L
ENGLAND Hamilton(Scotl and) 1 L
ENGLAND Liverpool 1 M
FRANCE Lyon 1 M
GERMANY Dusseldorf 1 M
GERMANY Mun ich 1 M
IRAQ Basra 1 L-M
KUWAIT Kuwait 1 L-M
LAOS Vient iane 1 L
LESOTHO Maseru 1 L
LIBERIA Monrov i 1 L
LUXEMBOURG Luxembourg 1 M
MALAYSIA Kuala Lumpur 1 L
78
DEGREE OF
MOZAMBIQUE Maputo 1 l
OMAN Muscat 1 L-M
POLAND Poznan 1 M
POLAND Warsaw 1 M
SINGAPORE Singapore 1 L
SOUTH AFRICA Durban 1 L
SOUTH AFRICA Natal 1 L
SWAZILAND Mbabane 1 L
SWEDEN Stockholm 1 H
SWITZERLAND Geneva 1 M
TANZANIA Dar es Salaam 1 L
TANZANIA Zanzibar 1 L
THAILAND Chiang Mai 1 L
TOGO Lome 1 L
U.A.E Abu Dhabi 1 L
ANGOLA Luanda 0 M
ARGENTINA Buenos Aires 0 M
AUSTRALIA Brisbane 0 M
BAHAMAS Nassau 0 L
BAHRAIN Manama 0 L
BELIZE Belize City 0 L
BERMUDA Hamilton 0 L
BOTSWANA Gaberone 0 L
BRAZIL Belem 0 L
BRAZIL Belo Horizonte 0 L
BRAZIL Brazilla 0 L
BRAZIL Manaus 0 L
BRAZIL Porto Allegre 0 L
BRAZIL Recife 0 L
BRAZIL Rio de Janeiro 0 L
BRAZIL Salvador 0 L
BRAZIL Sao Paulo 0 L
CAMAROON Yaounde 0 L
CENTRAL A.R Bangui 0 L
CHAD Ndjamena 6 L
CONGO Brazzaville 0 L
COTONOU Benin 0 L
EQUATORIAL GUINEA Mai abo 0 L
FRANCE Paris 0 M
GABON Librevil le 0 L
GAMBIA Banjul 0 L
GERMANY Berl in 0 M
GERMANY Bremen 0 M
GERMANY Hamburg 0 M
GUINEA Conakry 0 L
GUINEA Bissau 0 L
GUYANA Georgetown 0 L
INDIA Madras 0 L
IRELAND Bel fast 0 L
79
DEGREE OF
IRELAND Dubl In. 0 L

IVORY COAST Abidjan 0 L
KOREA Seoul 0 M
MALAGASY REPUBLIC Tananarive 0 L
MALI Bamako 0 L
MAURITANIA Nouakchott 0 L
MAURITIUS Port Louis 0 L
MEXICO Matamoros 0 M
MEXICO Merida 0 L
MEXICO Monterrey 0 M
MEXICO Nuevo Laredo 0 M
NETHERLANDS Amsterdam 0 L-M
NETHERLANDS The Hague 0 L-M
NETHERLANDS Rotterdam 0 L-M
NIGER Niamey 0 L
NIGERIA Ibadan 0 L
NIGERIA Kaduna 0 L
NIGERIA Lagos 0 L
PARAGUAY Asuncion 0 H
QUATAR Doha 0 L
SAUDI ARABIA Dhahran 0 L
SAUDI ARABIA Riyadh 0 L
SENEGAL Dakar 0 L
SEYCHELLES Victoria 0 L
SIERRA LEONE Freetown 0 L
SOMALI REPUBLIC Magadiscio 0 L
SPAIN Madrid 0 M
SRI LANKA Colombo 0 L
SURINAM Paramaribo 0 L
THAILAND Bangkok 0 L
THAILAND Songkhl 0 L
THAILAND Udorn 0 L
U.A.E. Dubai 0 L
UPPER VOLTA Ouagadougou 0 L
URUGUAY Montevideo 0 L
USSR Kiev 0 L
USSR Leningrad 0 L
USSR Moscow 0 L
VIETNAM Saigon 0 L
ZAIRE Kinshasa 0 L
80
TABLE 1
STEPS IN THE PRELIMINARY RANKING PROCESS
STEP 1: RELATIVE SEISMIC EXPOSURE

Eliminates buildings In Zones 2 and below
STEP 2 : OCCUPANCY
Eliminate buildings which are seldom occupied and
which do not
house essential equlprrent
STEP 3: VINTAGE AND LOCATION

For each location, rank buildings according to date
of construction
STEP 4: CLASSIFICATION ACCORDING TO BUILDING MATERIAL
STEP 5: PRELIMINARY STRUCTURAL EVALUATION
icaa Ttaair m«» OP laarwouuci op mowtuM MCaTd m
Fig. i Mean Yearly Number of Earthquakes of Magnitudes

Greater Than M in Portugal, California, and
New Zealand
81
LIST OF SLIDES SHOWN WITH PRESENTATION OF PAPER
SI ide 1 FBO INTRODUCTION

SI ide 2 FBO FACILITIES, SCALE OF MAGNITUDE
SI ide 3 FBO SQUARE FOOTAGE BREAKDOWN OF FACILITIES
SI ide 4 TYPICAL FOREIGN SERVICE POST
SI ide 5 1931 MANAGUA CHANCERY
1976
SI ide 6 1967 CARACUS CHANCERY
SI ide 9 GUATEMALA CITY CHANCERY
SI ide 10 1976 GUATEMALA CITY CHANCERY
SI ide 11 1976 GUATEMALA CITY RESIDENCE
SI ide 12 1976 GUATEMALA CITY RESIDENCE
SI ide 13 1977 BUCHAREST CHANCERY
SI ide 14 1977 BUCHAREST CHANCERY
SI ide 15 1978 THESSALONIKI CONSULATE
SI ide 16 1978 THESSALONIKI CONSULATE
SI ide 17 1984 SANTIAGO CHANCERY
SI ide 18 1984 SANTIAGO CHANCERY
SI ide 19 TOKYO MITSUI HOUSING COMPOUND
SI ide 20 DAR ES SALEM CHANCERY
SI ide 21 DAR ES SALEM CHANCERY
SI ide 22 1983 BERUIT CHANCERY
SI ide 23 1983 BERUIT CHANCERY FLOOR PLAN
SI ide 24 1983 BERUIT CHANCERY COLUMN DAMAGE
SI ide 25 1983 KUWAIT CHANCERY ANNEX
SI ide 26 1983 KUWAIT CHANCERY ANNEX
SI ide 27 1984 BERUIT CHANCERY ANNEX
SI ide 28 1984 BERUIT CHANCERY ANNEX
SI ide 29 SCHEMATIC, PRESSURE ON BUILDING
SI ide 30 SCHEMATIC, BLAST IMPACT ON FRAME
82
STRONG MOTION INSTRUMENTATION IN
BUREAU OF RECLAMATION PROGRAMS
Robert Mac Donald and Andy Viksne

Bureau of Reclamation
ABSTRACT reevaluation for optimum deployments regarding

the many site and instrumentation character-
The Bureau (Bureau of Reclamation) deployed its istics are constantly underway. The following
first strong motion earthquake instrument array paragraphs discuss Bureau guidelines for
in 1937; however, in the intervening years until instrumentation and deployment.
the 1971 San Fernando earthquake, only a few
additional sites were instrumented. The near-
2. SITE SELECTION CONSIDERATIONS
failure of the Lower San Fernando Dam (Lower Van
Norman Reservoir) gave new impetus to the Bureau dams are located throughout the 17
installation of strong motion instruments in westernmost states of the contiguous United
Bureau dams and related structures and 31 sites States. Strong motion earthquake instrument
are now instrumented. Bureau considerations in deployment at selected dams and other water-
the selection of sites to be instrumented and the resource related structures is taking place and
deployment of strong motion instruments at the will continue for many years to come. Discuss-
various sites are discussed. These considerations ions of site selection factors for strong motion
have evolved to meet the data requirements for instruments and the distribution of such
response analyses, and generally conform to instruments in the United States have been made
typical state-of-the-art practices dealing with by Viksne (1979) and Hudson (1984). The present
strong motion instrumentation of water resource goal of the Bureau strong motion program is to
structures continue deploying instruments considering the
Keywords; Earthquake; Instrument Array: following factors:
Strong-motion
1. INTRODUCTION (1) Seismic zoning and proximity to faults

capable of causing earthquakes
In 1937, Hoover Dam, located in an area that is
considered relatively aseismic, except for
(2) Dimensions of the dam and reservoir
occasionally intense but low levels of seismicity,
became the first Bureau dam to have strong motion
(3) Foundation materials
earthquake instruments installed. Until 1970,
strong motion instruments were installed at only
(4) Method of construction/type of dam
four other sites in Montana, California, Utah,
and Arizona. From 1970, and especially since
(5) New versus existing dams
the near- failure of Lower San Fernando Dam
during the 1971 San Fernando earthquake, the
(6) Water- resource lifeline features of
Bureau has accelerated its strong motion program
special interest
and added 26 installations with an additional
12 sites planned for the near future (figure 1).
The most important factor controlling the Bureau
These installations represent only a fraction
instrumentaton program is the location within
of the nearly 300 storage dams constructed and
prioritized seismic zones. The Bureau uses the
operated by the Bureau. As the Bureau network
Uniform Building Code (1982) Seismic Zone Map
of strong motion instrument sites expands,
as the basic guideline for this purpose
83
influence strong motion instrument siting be placed at a distance beyond any significant
priorities. Foundations that contain unconsoli- influence of the dam to the recorded ground
dated silts or fine-grained sands and are, there- motion. For structural analyses purposes, the
fore, potentially liquefiable are important Bureau does not require free-field installations
considerations. Hydraulic fill structures are if an array that includes input and response
vulnerable to failures under strong ground motion (see below) instrumentation is in place.
motion (Seed, et al, 1977) and may be prioritized Some of the Bureau installations are free-field
for instrumention over earth embankments emplaced simply because of access or siting limitations
with modern compaction methods. Jackson Lake Dam around the dams, especially downstream. There
in northwest Wyoming represents a hydraulically are a large number of dams in the United States
emplaced embankment on a potentially liquefiable with nearby "free-field" instruments, for example
foundation in seismic zone 3 and very close to in California (figure 4). More study is required
Teton fault, considered capable of causing a to determine the extent of the influence of large
magnitude 7.5 earthquake. Jackson Lake Dam is structures, such as dams and their concomitant
not only instrumented, but plans are underway reservoir loads and ground water effects, on
for a more earthquake-resistant structural earthquake strong motion to be certain of
modification of the dam. structure-independent free-field sites.
3 . DEPLOYMENT AT STRUCTURES 3.2 Input Motion Instruments - Input motion

instruments may involve surface or subsurface
The deployment of strong motion instrumentation emplacements, or both. For surface arrays at
(generally referred to as " ins trument ( s )" in the dams, the Bureau locates the input motion
following paragraphs) in dams requires special components of the strong motion array at the
site-specific evaluations and consideration of downstream toe and on the abutments as close to
dynamic analysis results (Bolt and Hudson, 1975; the dam as possible. Most of these instruments
Negmatullaev, et al, 1978; Viksne, 1979; Fedock, are placed in prefabricated housings on concrete
1982, 1984; Maley and Etheredge, 1984; Chugh, pads firmly secured to the underlying foundation
1985) when installations are planned. Fedock rock or surficial material. Instruments at the
(1982, 1984) subdivides a hypothetical instrument toe are located as close to the maximum section
array into three categories, free-field, input, as possible. A field review of a few typical
and response motion instrumentation. dams selected at random will demonstrate that
exterior siting conditions are difficult at the
3.1 Free Field Instruments -An ideal free-field toes of most dams due to backwater from down-
strong motion installation will obtain ground stream reservoirs, tailwater from power plants,
motion records that are unaffected by nearby spillways, stilling basins, plunge pools, outlet
man-made structures, by any enclosure for the works, or other features not amenable for
instruments, or by unusual topographic sitings. instrumentation sites. Similar conditions can
Under these criteria, the installation of ideal exist at the dam/abutment contact area where
free-field instruments may be an elusive goal. topographic and access considerations frequent lv
The Bureau has located free-field instruments allow little room for satisfactory instrument
near a number of structure sites, and at some sites. In the abutment areas, an lde.il Installa-
of these sites, free-field is the only type .of tion in a spatially restricted area would be ..
installation. Fedock (1984) recommends free- small chamber in the natural material where
field instruments near both abutments and at the maintenance problems could be minimized.
toe of an "ideal" dam; such instrumentation to Interior or subsurface input motion emplacement#
(figures 1 through 4). There is no doubt that free field strong motion installation, will be
such zoning is very broad and that there are instrumented upon completion of construction
"zones within zones" that must be considered in in the near future.
the Bureau's sequential deployment of strong
motion instruments. Within zones 3 and 4, the Newer dams, outside of seismic zones 3 and 4,
zones of highest instrumentation priority, dam have been or are planned for strong motion
locations may vary greatly in respect to instrumentation because of engineering concerns
proximity of faults capable of causing earth- for a complete suite of dam instrumentation,
quakes and earthquake magnitude and recurrence including strong motion. This concern may
exposure levels. Within seismic zone 4 in occasionally transcend seismic zone boundary and
California, for example, the USGS (Geological seismic exposure considerations. Bureau strong
Survey), CDMG (California Division of Mines and motion instrumentation planning does, however,
Geology) and the University of Southern take into account site-specific detailed seismo-
California have located arrays of strong motion tectonic studies throughout the Western United
instruments where historical seismicity States that are done by the Bureau, other
indicates strong motion events will occur in the agencies, and private industry. These studies
relatively near future (Rojahn and Borcherdt, are extending our knowledge of causative earth-
1983; Maley and Etheredge, 1984). An example quake structures such that existing seismic
of a subzone of seismic zone 3 where the Bureau zoning, in terms of magnitude and seismic
considers significant seismicity possible during recurrence to be expected, will require modifica-
the lifetime of its dams is in east-central tions and refinements. The Bureau, therefore,
California, where Boca, Stampede and Prosser views seismic zone maps as state-of-the-art
Creek Dams are located. Both Boca and Stampede guidelines which can be modified by engineering
Dams have been instrumented for a number of years judgement. Strong motion instrumentation, as
and the instrumentation of nearby Prosser Creek related to water resource structures throughout
Dam is planned (figure 2). the Western United States, is concentrated in
seismic zones 3 and 4, but still widely
The basic program with regard to seismic zoning distributed throughout seismic zone 2 (figures 3
calls for the installation of strong motion and 4 )

instruments on all existing storage dams in
seismic zone 4 and on most of the significant Early in the Bureau's strong motion program,
storage dams in seismic zone 3. Two of the large dams and reservoirs had high priority for
Bureau's large pumping plants, Dos Amigos and strong motion instrumentation because of size
Pleasant Valley, in seismic zone 3 have strong and not because of seismic zoning. Therefore,
motion instruments and instrumentation is planned Hoover, Hungry Horse, Bradbury, Flaming Gorge,
for two additional pumping plants, Pacheco and and Glen Canyon Dams had strong motion
Coyote, now under construction and located in instruments installed as a construction-related
seismic zone 4. These two additional pumping activity from 1937 to 1960. Only one of these
plants are of special interest because of their dams, Bradbury, located in southern California,
lifeline functions and proximity to the major, is in seismic zone 3 or 4.
active Calaveras and San Andreas faults. The
Santa Clara pipeline, which crosses the Calaveras The present Bureau practice is to consider
fault between the Pacheco and Coyote pumping seismic zoning as the predominant siting factor,
plants, will also be instrumented. The Yuma but size, foundation materials and method of
desalizination plant, which now has a nearby construction are other parameters which may
ss
consist of borehole instrumentation in the locations would be where lower safety factors
foundation of earth dams and installations in and higher loads are expected. These locations
selected galleries of concrete dams. Drainage would be site specific for each structure and
and grouting galleries, when excavated in earth for earth dams depend upon: (1) zoning geometry
dam foundations, can be utilized as input motion of the dam, (2) types of materials used in the
sites for strong motion instruments. zones; and (3) the nature of the foundation.
Many Bureau dams are built on deep deposits of Figure 5, showing a plan and section of Sugar
unconsolidated surficial materials in the Pine Dam in California, is an example of a
bottoms of valleys and canyons. If such relatively large instrument array. There are
deposits represent the typical foundation four downhole accelerometer units in Sugar Pine
conditions at the base of a dam, then input Dam. One is located on bedrock for input motion
motion is best represented from instrumentation and the others in the dam in zones 1 and 3A for
located on the surficial materials as close to response motion.
the base of the dam as feasible, and not, for
example, on any bedrock that might be out- For embankment dams, instruments are installed
cropping in the area. However, at some sites, on or in selected embankment material zones
both bedrock and surficial deposits may be through the use of surface sites or boreholes.
instrumented Access is important in such selections. The
galleries in many concrete dams usually allow

3.3 Response Motion Instrumentation -The Bureau a choice of internal locations. Ideally, the
ideally installs one or two response instruments response and input strong motion instrumentation
on the crest of both earth and concrete dams. will be located in the same crosssection normal
The primary location is where maximum deformation to the longitudinal axis of the dam and thus
during strong motions is expected. This primary provide the basis for analyses that compares
location is, based on past analyses, usually at calculated with actual response of the dam to
the maximum section. Many Bureau dams are earthquake loadings.
asymmetrical because of differences in abutment
slopes, stream channels not in the center of the 3.4 Payback Records- Payback records, meanlnjt
valley, or other topographic features often the Bureau received good strong motion records
related to geologic structures. The maximum from its investment in instrumentation, were
section, therefore, may be far from the actual obtained at Pleasant Valley Pumping Plant, one
center of the dam crest. Other locations may of two Bureau pumping plants currently having
be selected such as about one-third of the crest strong motion instruments. The Pleasant Valley
length from an abutment and/or on the downstream Pumping Plant had a maximum input (at the basr '
of the face of the dam. These locations are acceleration of 0.33g recorded during the Mav ,
basically for backup and special applications. 1983 Coalinga earthquake (figure 6). Minimum
Additional strong motion instrument locations instrument placement at these pumping plant 1
*
throughout concrete or embankment dams may be require instruments to be located at the ground
desired for dynamic analyses after earthquakes. surface or near the foundation/structure contact
and at the top of the steel super s t rue t un t*>
If a dynamic analysis of a structure has been obtain data for an adequate response analysis
done prior to strong motion instrument deploy- (figure 7).
ment, the response instrument locations would be
specified based on the analysis. Specific
86
In comparing the pumping plant foundation strong eliminated by making the basement instrument the
motion records with those of the switchyard primary starter and interconnecting it to the
(figure 6 and 8), it is seen that the switch- other instruments in the plant. An option to
yard or "free-field" instrument recorded peak eliminate false triggering would have been to
horizontal accelerations slightly over 50 percent reset the threshold acceleration of the first
larger than those of the foundation in both the floor and "roof" instruments at a higher level
main magnitude Ml = 6.5 earthquake of May 2, than 0.01 g, at which (presently) all Bureau
1983 and the Ml = 5.1 aftershock of May 9, 1983. accelerographs are set.
The epicentral and hypocentral distances were
9 km and 14 km, respectively, for the May 2 In California, the Pacheco and Coyote pumping
earthquake and 10 km and 16.4 km, respectively, plants, as well as the Santa Clara pipeline,
for the May 9 aftershock. The topographic will be situated much closer to the active San
effects on ground motion are emphasized in this Andreas and Calavaras faults than the Dos Amigos
case. The input motion peak acceleration is and Pleasant Valley pumping plants, which
lower than the "free-field" motion because the already have strong motion instruments. The
input motion instrument is at the base of the Santa Clara pipeline, located between the
pumping plant in a deep excavation in piedmont Pacheco and Coyote pumping plants, will cross
alluvial fan deposits, whereas the switchyard the Calaveras fault. Instruments that will
"free-field" instrument is located near the top alert the pipeline operations center of
of the cut slope of the pumping plant excavation predetermined levels of strong ground motion
(figure 7). As previously discussed under (probably 0.25g) at the pipeline-fault crossing
"free-field instruments", the pumping plant mass are planned, such instrumentation to utilize a
could have had some influence on the ground typical electromagnetic trigger to actuate a
motion and could have dampened it during both warning device and to give an alert. This
the May 2 earthquake and the May 9 aftershock. instrumentation will also automatically cause
a shut off of water flows in the Santa Clara
The two response motion instruments located on pipeline until onsite inspections can be
the first floor and on a steel beam just beneath performed after triggering- level earthquakes.
the roof of the pumping plant experienced Similar electromagnetic triggers are used as
instrument malfunction and false triggering due switches in nuclear power plants (Rihn and
to crane operation vibrations before the main Beckmann, 1979). Seismic alarm devices
earthquake shock and no records were obtained. (prototypes) using vertical accelerometers as
The film was replaced and accelerations from some accelerograph triggers and used to actuate
of the aftershocks were recorded. In the after- warning horns and lights, have been designed
shock of May 9, 1983 the basement instrument and are in use by the Corps of Engineers for
recorded a maximum horizontal acceleration of water resource related life-line structures
0.14g while the roof instrument recorded 0.24g. (R. Ballard, personal communication). These
Since, during the main shock, a horizontal Corps of Engineers seismic alarm devices are
acceleration of 0.33g was recorded in the base- installed at seven sites in California.
ment, a proportional extrapolation would indicate
that a horizontal acceleration of over 0.5g could 4 . TYPES OF STRONG MOTION INSTRUMENTS DEPLOYED BY
THE BUREAU AND SOME INSTALLATION CONDITIONS
*
have occurred during the main earthquake in the

superstructure of the plant. False triggering
Except for a few older model RFT-250 and RFT-350
at the Pleasant Valley Pumping Plant has been
accelerographs, the Bureau utilizes SMA-1
87
accelerographs on all surface installations. Downhole accelerometers originally were
Models FBA-3 and FBA-13 accelerometers are installed in boreholes and then grouted into
utilized in boreholes. All borehole accelero- place. In order to deal with instrument
meters are installed in conjunction with CRA-1 malfunction, current practice is to attach a
recorders. Manufacturer's specifications and stress cable to the instrument and to backfill
other papers (Pauly, 1984) discuss details of around the triaxial downhole units with pea
these instruments. gravel. In this way an option of washing the
instrument free of the gravel pack and then
All of the self-contained accelerographs are removing the instrument from the hole is
three-component, optical mechanical units which retained. Borehole strong motion instruments
produce analog records. A component system is have been successfully installed and removed by
used for several borehole systems, wherein the Bureau using a gravel pack and cable during
accelerometers and a starting device are placed installation.
in the borehole, and are hard-wired to a
centralized recording unit at some remote OPERATION AND MAINTENANCE CONST DERATIONS
location. The self-contained and borehole

systems photo-optically record the data. The Optical-mechanical, self-contained accelero-
systems remain in a standby condition until graphs which record on 70 mm photographic film
ground motions (O.Olg or more) are detected by are the most widely used strong motion type
the seismic trigger which then actuates a worldwide due to their low cost, high
recording lamp and a film (70 mm) drive motor reliability, simple operation, and straight
contained in every system. forward maintenance/repair procedures (Pauly,

1984). They are preferred over both analog and
As mentioned earlier, most of the surface digital tape and solid state recording methods;
instruments are located in prefabricated housings however, improvements in these alternate types
and bolted onto concrete pads. The pads are continue to be made. After triggering, up to
usually designed so that they have a natural 25 cumulative minutes of motions equal or
frequency (30-32 H^) which is higher than the stronger than O.Olg can be stored (recording
s
continues for 10 seconds after motions drop

frequency of the accelerometers (20 range)
below O.Olg level).
and considerably higher than seismic frequencies
of interest (Bureau of Reclamation, 1985).
Several calibration tests, both laboratory and
Instruments placed in the interior of dams, such
field, must be conducted before the instruments
as those in the galleries of a concrete dam or,
are installed. Instrument sensitivity, or the
less frequently, in drainage or grouting
galleries under embankment dams, may be bolted deflection of the seismograph trace under l.Og
loading, is generally performed by the
to the floor of the gallery or gallery alcoves.
The three component accelerographs contain three manufacturer under laboratory conditions before
the equipment is brought to the field. Calibra-
accelerometers in orthogonal arrangement with
tion tests for damping and natural frequency are
the horizontal units both parallel and
easily performed in the field. These calibra-
perpendicular to the longitudinal axis of the
tion tests are recorded at the beginning and end
dam (tangentially and perpendicularly if an arch
of each roll of film.
dam) .
88
Two timing traces are continuously recorded 6 . DISCUSSION - STRONG MOTION INSTRUMENTATION IN
with the earthquake ground motions. One of BUREAU PROGRAMS
these traces is always a square wave pulse

The safety of its dams and continued research
produced internally. The second trace may be
into design, construction, and maintenance of
another square wave pulse, a binary time code
water resource features are primary concerns
controlled by an accurate oscillator and set
of the Bureau's mission. The relationship of
according to an external clock, or, more
the installation of strong motion instruments
commonly, a WWVB Bureau of Standards radio
with structural safety is not quite direct.
signal time code. The timing is necessary so
For Bureau purposes, strong motion instruments
that each motion event can be clearly identified
are primarily research tools for understanding
for correlation with events on other records.
structural response to earthquakes.
Most of the maintenance of Bureau strong motion

In order for an instrument array to have any
instruments is carried out through a cooperative
payback, it is necessary for the structure in
effort between the Bureau and the USGS.
question to be subjected to strong earthquake
Typically, a technician checks the instrument
motions. The closer those motions are to
installation every six months and removes the
damaging motions, the more knowledge is to be
exposed portion of the film if the instrument
gained
has been triggered. If a significant earthquake
occurs in the vicinity of an instrumented dam,

Strong motion records of seismic events wherein
an additional check is made for strong motion
significant shaking occurred in dams are very
records. If strong motion has been recorded,
few. Those records that are applicable to dams
the analog record is routinely digitized,
of modern construction are even fewer. Thus,
corrected for film distortion, variable film
available strong motion data, in relation to
speed, and instrument response, and the record
dam design, are small in quantity and poorly
filtered to remove long period noise.
applicable. Nevertheless, existing strong
motion data are valued for the confidence they
The maintenance of the Bureau's strong motion
produce when applying, in design considerations,
arrays is complicated by the diversity of the
parameters judged to be adequate dynamic
sites, such as extremes in elevation and climate.
loadings
Both heat and cold have to be dealt with; the
addition of sun roofs and improved air
The preceding discussion has presented a perspec-
circulation by venting has lessened heat
tive of strong motion instrumentation in Bureau
problems, while electric heating and insulated
programs. This is useful in further discussion
shelters have dealt with extremely cold condi-
that follows involving strong motion issues such
tions. Drainage is an important aspect of
as utilization of data, cost-effective equipment,
surface instrument sites, and poor drainage has
optimum deployment, function completion and
temporarily caused operational problems and shut-
additional research needed.
downs. In attempts to locate instrument housings
out of the way of access roads and work areas,

6.1 Utilization of Data - The primary value of
the housings have been located close to cut
strong motion data to water resource structure
slopes. Sloughing material and minor slides have
design engineers is that the accrued data can
both threatened and caused damage to housings at
be used to check that actual response of
these sites. Small concrete revetments now
structures to earthquakes are within design
protect several housings.
89
limits and that the actual response conforms to Optimum deployment will be achieved when
assumed response. This was accomplished analysis requirements have been satisfied by
successfully at Pleasant Valley Pumping Plant. the most efficient and dependable instrumenta-
For example, structural periods determined by tion siting pattern. The initial cost of the
a dynamic analysis of a finite element model equipment, the cost of the site preparation,
conformed to structure periods (response motion) which is relatively high if borehole instruments
recorded during the 1983 Coalinga earthquake at are utilized, and maintenance costs must be
the Pumping Plant. reasonable
6.2 Cost - Effective Equipment - The use of Optimum deployment regarding instrument locations
relatively low cost and reliable optical- relative to dams and other related structures
mechanical self-contained accelerograph is requires considerable study of each site. The
planned for all surface sites. Digital effect of topography on ground motion is not
recording acce lerographs will continue to be well understood; this aspect of strong motion
considered as alternatives, especially as is considered in more detail under the
technical modifications are constantly being discussion of additional research needs which
made in improving efficiency, reliability, and follows
sensitivity of the digital system.
6.4 Function Completion- In most Bureau studies
There has been a less then desirable operating and investigation programs the following
reliability of the borehole units. Even though question has to be answered: "When has this
the Bureau's installation plan for "sanding in" investigation program collected enough data to
these borehole instruments with pea gravel and be satisfactory for the purpose intended?" All
an attached stress cable is satisfactory for too often time and funding factors make the
retrieval and replacement, the deployment of decision. Refined investigations after bas
borehole instruments is presently being data are gathered frequently belong in the
reevaluated "level of diminishing returns" and in Bureau
design the lack of such refinements can often
6.3 Optimum Deployment- The general requirements be relegated to an area synonymous with
for a strong motion instrument array at a damsite "engineering risk."

have been previously discussed. When dealing
with the site-specific aspects of instrumenting Strong motion instrumentation in the Bureau is
a dam, however, detailed consideration of subject to the same scrutiny as any other
constraining factors in surface and subsurface program in the name of economy and "payback",
deployment are necessary. These considerations and with the advent of significant ground motion
may be as simple as maintenance access or may and subsequent records at a site, the function
require a detailed study of an embankment dam's completion of that strong motion instrumenta-
material and zoning, available gallery locations tion should be considered. In this context,
in concrete dams, or a knowledge of traffic the 1983 actuation of the Pleasant Valley
patterns and operational activities with their Pumping Plant array by strong motion raises an
subsequent limitation of sites. issue regarding the value to the Bureau of
continuing deployment at this location. The
following facts are pertinent: (1) The ground
motion and subsequent loadings on the pumping
90
plant are probably the maximum that can be [2] Borcherdt, R.D., "The Coalinga Earthquake
expected. Even though the earthquake main Sequence Commencing May 2, 1983,"
shock was of moderate magnitude, the epicentral Geological Survey Open-File Report 83-
distance was only 9 km from the pumping plant, 511, 1983.
(2) Another similar earthquake is not expected

within the lifetime of the pumping plant (the [3] Bolt, B.A. and Hudson, D.E., Seismic
return period for similar or greater earthquakes Instrumentation of Dams," Journal of the
on the causative fault structure is estimated Geotechnical Engineering Division, ASCE,
at 350 years [Geological Survey, 1984]), November 1975.
(3) Analyses have indicated an acceptable
structural response and no structural damage [4] Bureau of Reclamation, "Bureau of
resulted Reclamation Strong Motion Instrumenta-
tion Site Characteristics," Department
6.5 Research Needs- Much research is currently of the Interior, Denver, Colorado, 1985.
underway in the study of earthquake strong motion;

however, three research areas stand out as being [5] Chang, F.K. ,
"The Effects of Elevation and
particularly applicable to optimal array deploy- Site Conditions on Ground Motion of the
ment at dams. The first area involves the site San Fernando, California, Earthquake,
topography, especially the effect of the 9 February 1971," Proceedings of the
proximity of valley slopes or canyon walls on Seventh World Conference on Earthquake
ground motion. The shape and height of valley Engineering, Volume II, Istanbul, Turkey,
slopes are probably factors. Davis and West, September 1980.
1973, Chang, 1980, and Chang, 1985, discuss
topographic effects on strong ground motion from [6] Chang, F.K., "Analysis of Strong-motion
empirical observations. The second area involves Data from the Mount Borah, Idaho, Earth-
the effects on strong ground motion of surficial quake of 28 October 1983," Miscellaneous
[1]
deposits (soils), especially, under large water Paper GL-85-12, U.S. Army Engineer Water-
resource structures. The third area involves ways Experiment Station, Corps of
the effects of dams and reservoirs on strong Engineers, Vicksburg, Mississippi, May
ground motion. Long-term, detailed, empirical 1985.
studies are visualized as necessary to gather

adequate data for analyses. [7] Chugh, A.K. ,
"Dynamic Response Analysis of
Embankment Dams," International Journal
7 • REFERENCES for Numerical and Analytical Methods in
Geomechanics, 1985.
Ballard, R.F. ,
Jr., Earthquake Engineering
and Geophysics Division, Waterways [8] Davis, L.L. and L.R. West, "Observed
Experiment Station, Corps of Engineers, Effects of Topography on Ground Motion,"
Vicksburg, Mississippi, personal Bulletin of the Seismological Society
communication, April 1985. of America, Volume 63, No. 1, pp. 283-
298, February 1973.
[9] Fedock, J.J., "Strong-Motion Instrumenta-

tion of Earth Dams," Geological Survey
Open-File Report 82-469, 1982.
sa
[10] Fedock, J.J., "Strong-Motion Instrumenta- [17] Rihn, W.J. and W.J. Beckmann, "Electro-
tion of Earth Dams," Proceedings of the magnetic Starters for Strong Motion,"
Eight World Conference on Earthquake Kinemetrics Application Note No. 15,
Engineering, Volume VII, San Francisco, Pasadena, California, August 1979.

California, July 1984.
[18] Rojahn, C. and R.D. Borcherdt, "On the
[11] Hudson, D.E., "Strong Motion Accelero- status of In Situ Strong Ground Motion
graph Systems - Problems and Prospects," and Structural Response Investigations,"
Proceedings of the Eighth World Proceedings of the Fourteenth Joint UJNR
Conference on Earthquake Engineering," Conference, pp. 617-654, 1983.
Volume II, San Francisco, California,
July 1984. [19] Seed, H.B., F.I. Makdisi and P. DeAlba,
"The Performance of Earth Dams During
[12] Kuribayashi, E., H. Tsuchida and M. Watabe, Earthquakes," Report No. UCB/EERC-77/20,
"Maintenance of the Strong-Motion Earthquake Engineering Research Center,
Accelerograph and the Data Processing of College of Engineering, University of
the Records," Proceedings of the Seventh California, Berkeley, California.
Joint UJNR Panel Conference, Tokyo,
Japan, May 1975. [20] Uniform Building Code, Seismic Zone Map of
the United States, 1982.
[13] Maley, R.P., "The United States Strong-

Motion Network: Field Operations," [21] Viksne, A., "Bureau of Reclamation Strong
Proceedings of the Seventh Joint UJNR Motion Instrumentation Program,"
Panel Conference, Tokyo, Japan, May 1975. Proceedings of the Third National
Congress on Pressure Vessels and Piping,
[14] Maley, R.P., and E.C. Etheredge, "The San Francisco, California, June 1979.
Development of Ground and Structural
Response Strong-Motion Instrumentation
Arrays in the United States," Proceed-
ings of the Eighth World Conference on
Earthquake Engineering, Volume II,

San Francisco, California, July 1984.
[15] Negmattulaev , S. Kh., G.S. Seleznyov, D.W.

Simpson and C. Rojahn, "Engineering and
Seismological Observations at Dams,"
Proceedings of the Second International
Conference on Microzonat ion ,
Volume II,
San Francisco, California, 1978.
[16] Pauly, S.E., "Strong Motion Accelerograph

Selection," Proceedings of the Eighth
World Conference on Earthquake
Engineering, Volume II, San Francisco,
California, 1984.
92
Fig. 1 Distribution of Bureau of Reclamation strong motion installations.
Fig. 2 Distribution of Bureau of Reclamation strong motion

installations in California.
S5
Fig. 3 Distribution of strong motion installations at water resource
features throughout the western United States — excluding California.
Fig. 4 Distribution of strong motion installations at water
resource features throughout California.

eUBMNKUCNT EXPLANATION
StlO€fO0 C'tff, im. »#*# 0»1 gruwl tO»«f '##
O dr 'tmgiog 'Ot/or '0 Otim 'orort
St'octd atr. tuf. u*0 aod g't*t> f*oro*f*tf mn*o >0
© Ongm.H mitt ’ \
tr -o'tmo of Off 0OmO»rtd boot « '#
»*tf compoetoo #r tompitg ro"or to Otim o/tr-
© Sond filtor -compocfod bf fbeotorp rot'or * 0 Sm tro't

-mimmt Timm- compact to br
© G'bwot tilfor
,0 a 5* 'Off*.
*+cttm -m,*mt Timm -eompoe'td br •
• ’bnto'r
'brofrf '« •'

10 0 6m ’Of or ».
doe* fill-plot Timm -co»po*tod bf • •brofmrr 'O' •r
,0 0 9 m 'oror t
Fig. 5 Plan and section of Sugar Pine Dam showing strong

motion instrument locations.
95
530 gal
ig r Basement 1 35‘
TJF
o -
— rr
045‘
-ig 320 gal
I 5 seconds f
Fig. 6 Strong motion records from Pleasant Valley Pumping Plant during main shock of
Coalinga Earthquake, May 2, 1983 (after Borcherdt, 1983).
Fig. 7 Section through Pleasant Valley Pumping Plant showing location of strong motion
instruments.
96
Switchyard
19 r 215 gal
135
ywyvw
UP
0
045°
-19
seconds
Basement
q 140 gal 135°
V
UP
- 045°
Igr First Floor

o *130
VV -'
' —
gal
- '
135'
UP
045'
-ig
Plant array
Fig. 8 Strong motion records from Pleasant Valley Pumping
during Coalinga Earthquake aftershock of May 9, 1983
(after Borcherdt, 1983)
97
ARRAY OBSERVATION OF STRONG-MOTION
EARTHOUAKES IN THE FUCHU AREA
BY
Keiichi OHTANI ,
Shigeo KINOSHITA,
Tadashi MIKOSHIBA
ABSTRACT
In order to investigate the earthquake response study at the Fuchu area located west of Tokyo.
of thick sedimentary layers in the Tokyo metro- The purpose of observation at the above-mentioned
politan area, the National Research Center for area is to investigate the earthquake response
Disaster Prevention has been performing strong- of the thick sedimentary layers. The seismic
motion observations. For that purpose, two waves within the period of 7 to 10 seconds are
types of the observational systems arrays at predominant because of the modulation effects
the Fuchu area, i.e., the vertical array by four by the thick sedimentary layers. In such period
triaxial accelerometers and the horizontal array ranges, the velocity waves are emphasized is
by five velocity-meters, have been established. compared with the acceleration waves. Also, the
velocity-type strong-motion seismographs are now
In this report, we will introduce the outline of accurate enough for practical usage. Therefore,
the observational study at the Fuchu area locat- the velocity-type strong-motion seismograph arrav
ed in the west of Tokyo, the velocity-type has been constructed at the Fuchu area.
strong-motion seismograph, and the observed
examples of seismic waves propagating in the
thick sedimentary layers. 2. ARRAY OBSERVATION OF STRONG-MOTION F.ARTHQ
Keywords: Earthquake; Strong-motion 2.1 OUTLINE OF OBSERVATION
1. INTRODUCTION
In the Tokyo metropolitan area, the sedimentary The observational sites are located at the Fuchu
layers of several kilometers in thickness on the area in the west of Tokyo (see Figure 1). This
Pre-Tertiary basement, are among the most impor- array consists of the center site FCH (Fuchu
tant factors of earthquake disasters. Therefore, crustal activity observatory) and five satellite
several studies have been performed to estimate sites: ING, KFC ,
TMA ,
MFC, AND CHF. A triaxial
the velocity structures of the thick sedimentary velocity-type seismograph is Installed at all
layers and base layer, to observe the strong- sites .
motion waves on and in such layers, and to simu-

late the observed waves theoretically. The At site FCH, three boreholes were drilled with
National Research Center for Disaster Prevention depths of 200 m, 500 m, and 2,750 m, respec-
(NRCDP) has proceeded with the observational tively. The bottom of the deepest borehole of
study of strong-motion earthquakes in the metro- 2,750 m is 750 m below the upper boundary of
politan area. the Pre-Tertiary basement. A triaxial for..-
In this report we will introduce an observational balanced type accelerometer Is Installed at

the bottom of each well. The same accelero- 2.2 Velocity-Type Strong-Motion Seismograph
meter is also installed on the surface ground We have used a new velocity-type strong-motion
at the site. Each seismometer is connected to seismograph in this array observation. This
a digital event recorder. The seismic signals seismograph is installed on the ground surface
are digitized as 12-bit words at 50 samples at each site.
per second. Each recorder has digital delay In this section, we mention simply the content
memory which stores the seismic signals for 10 of the new velocity-type seismograph. The
seconds, and a common time base. The crystal fundamentals of this seismometer are shown in
clock of each instrument is corrected every Fig. 2. Voltages e

s
(t) is generated across
day by radio. By January 1983, all instru- the terminals of a signal coil that is fixed
ments were in place. to the mass m and is moved through a magnetic
The actual extent of velocity structure vari- field with velocity x(t). The motion x(t) of
ations is already known in some measure, from a pendulum is relative to the seismometer
borehole information (velocity loggings) from frame. The voltage e (t) is proportional to
s
site FCH and the results of seismic prospect- this relative velocity x(t) between mass and
ing by using the explosions. Table 1 is the magnet. The constant of proportionality is
result of estimated velocity structures of the generator constant G $ .
shear-wave from site FCH. The sedimentary The induced voltage e (t)
s
is fed into a servo-
layers over the Pre-Tertiary basement have a amplifier with a gain of so that the output
thickness of about 2 kilometers. In this voltage of amplifier is e (t) = Aje s (t). By

v
array area, the Pre-Tertiary basement is dip- the voltage e (t), feedback current if(t)
v
ping from the west to the east. Based upon flows through the feedback circuit as follows:
the results shown Table

in 1, shear waves with i
f
(t) = A
1
G
$
x(t)/z t
^
the period of from 7 to 10 seconds are predom-
where z
t
is impedance of the feedback circuit.
inant at site FCH. But, the basement is dip-
The current if(t) produces the restoring force
ping from the west to the east in this area so
Gfif(t) acting on the mass. Gf is the motor
that the above-mentioned predominant shear
constant of the feedback coil.
waves have different characteristics , as com-
It follows that the equation of the pendulum's
pared with the response characteri sties of SH-
motion is
waves propagating in horizontally layered 2

x + wo x = -y (t) - G (t)/m
(
t ) (
t )
f
i
f (2)
model s
where y(t) is ground displacement and wq is
natural angular frequency.
When the spring constant is k, coq is natural
angular frequency.
99
When the spring constant is k,wQ equals direct shear wave is modulated by the thick
to ifak/m. From equation (1), equation (2) has sedimentary layers.
a following solution for h>>l: The record was obtained during the earthquake
*(t) = 3 which had origin at the Eastern Yamanashi Pre-

- ( )
2h
h
where fecture. It is a well known fact that earth-
A.G G-
= quakes whose origins are located in the East-
The
h
wo t
feedback
\ current is fed into a
<4)
current ern Yamanashi Prefecture are convenient for
amplifier with the gain of F° r ^2 >> investigation of the effects of wave propaga-
the output current of the amplifier is tion in the thick sedimentary layers of the
represented as follows: metropolitan area. This fact is based on the
V 4
* c R
f
XV 1 *
dt ’ ^ total reflection of seismic waves at the upper
where Rf is output impedance of the amplifier. boundary of basement.
With a load resistance R , the final output The earthquake of Feburary 14, 1984 (M=5.2)
0
voltage eg(t) is following results from which produced the NS-component record shown
equations (1), (2), (3), and (5): in Fig. 4, is regarded as an aftershock of an
= earthquake of August 8, 1983 (M=6.0). The N-S

e
o
(t)
VU*)
y component is almost identical to SH-component .
G C R
f f
The voltage e (t) is in proportion to ground The wave train shown in Fiq. 4 has five parts:
0
velocity. The absolute value of seismometer P-wave and its coda, direct S-wave (
Sq , two
sensitivity is determined by mass m, load re- total reflected pulses of Sq (Sj and Sp), lat-
sistance Rg/C Rf and motor constant Gf. er large amplitude wave packets, and final
Amplifiers and signal detector do not influ- coda wave. The waves Sq, Sj, S^ and later
ence the accuracy of measurement. packets appear to be necessary to explain the
The actual seismometer used in practical seis- shear wave modulated by the sedimentary
mometery has specifications shown in Table 2. 1 ayers
Fig. 3 shows the overall frequency character- The pulses corresponding to Sq, Sj, and Sp In
istics. In our seismometry, the sensitivity the wave trains recorded by velocity-afters
of the seismograph is 50 mV/kine. Photo 1 i nstal 1 ed on the ground
shows the velocity-meter and the recorder. seen in the records by bore-hole observat'on.
Two pulses Sj and are phase nodulated wive
3. EXAMPLES OF OBSERVED WAVES of Sq, propaqatinq between the surface of the
Fig. 4 is an illustration of recorded wave- basement and the qround surface, a kind of
forms across the whole array. This record multiple reflection. Phase velocities of Sq,
demonstrates clearly the process in which a Sj, and S in Fig. 4 are 3.9 km/s, 2.1 k»/s
2
100
5.
and 1.7 km/s, respectively. These values show ACKNOWLEDGMENTS
that these pulses are propagating in down-dip The present work would not have been possible
di rection without the aid of the members of the second
The SH-waves with predominant periods of about research division of NRCDP. In particular,
8 seconds will be modulated to equal the pul- the authors would like to thank Dr. H. Taka-
6.
hashi Mr. M. Takahashi Mr. H. Suzuki, and
ses Sq, S}, and S
2
. To observe this, the , ,
bandpass filtering with the central frequency Mr. K. Itoh.
0.125 Hz and the damping factor 0.2 is applied

REFERENCE
to the transverse component records obtained
at the site TMA. Fig. 5 shows these results Yamamizu, F., Takahashi, H., Goto, N., and
and also those of two records obtained during Ohta, Y., 1981, Shear wave velocites in
the earthquakes of the Western Nagano Prefec- deep soil deposits, Zisin , v. 34, p. 465-
ture. Fig. 5, then, shows that when the mag- 479 (in Japanese).
nitude of an earthquake is large, the seismic
wave with a period of about 8 seconds devel-
ops, and that the wave shapes with a period of Table 1 .Velocity structure of the S-waves
4. at the site FCH.
about 8 seconds are almost equal, independent (after Yamami/'u et al . 1981) ,
of the magnitude of earthquake. Thickness Velocity Density

No. (m) (km/s) (t/m 3 )
This fact is explained by making use of Hil-
1 4 140 1.7
bert transform. An example of model wave, 2 12 350 1.8
3 6 420 1.8
which is obtained by considering the dipping 4 22 360 1.8
5 6 530 1.8
structure of basement, is also shown in the 6 6 600 1.9
7 20 440 1.8
same figure. 8 14 460 1.9
9 115 540 2.0
10 856 780 2.0
11 963 1190 2.2
CONCLUDING REMARKS 12 726 2530 2.5
13 2530 2.5
This report is an introduction of the Fuchu
array observation progression by NRCDP. We

Table 2. Specifications of seismometer
explained the outline of observation and the
examples of seismic waves propagating the (1) Mass (m) 43 gram

(2) Natural frequency (fn) 1.43 Hz
thick sedimentary layers. (3) Generator constant (G s )
2.5 volt/kine
(4) Motor constant (Gf) 2.86 x 10 6
Quantitative study is underway now, and will dyne/A
(5) Ampl fiersi
be reported at a later date. Voltage gain of A^ 54 dB

Current gain of > 90 dB
(6) Load constants
C 1 uF
R
f
1 m
R* 0.1 - lkn
IDOL
Fig. 1 Layout of seismographs at the Fuchu area.
Fig. 2 Fundamentals of velocity type sc i smome t < r.
HE
(m\l/ k ne )
i
1
100 -n
50 -
u— 1 1
— i— -
i
1
i t r|
1
1
— —rn
i
tf] i
1
——
i
0.01 0.1 1 5
Frequency in Hz
( rad)
Photo 1 The velocity-meter and the recorder.
103
CHOFU (CHF)
2.0
2.0 -|
4 -
-2.0 -i
HIGHS 1 —
2.0 -
- 2.0
)
2.0
I
fSS kkh AAT\/V~^v-^vvsxwv
cm/s
- 2.0 i Down-_hpl_e_
( f|
FUCHUj--200n
Q B
* 1
-0.6 - f
FUCHU -FOQM
0.6 -
-
0.6 -
t
r lJ|HU -275m
>3 1
4 ——
-0 .3
,r " 1
ki r h-f 1
i
2.0 KFC I
( ) I
- ** - -
- 2.0
T
HMH
2.0 -
- 2.0 -
10s
Fig. 4 Velocity waves obtained by array observation

shown in Fig . 1
Filtered Waves (f 0 = 0.125 Hz)
T MA
1984.2.14
M5.2
Original Waves
1983.8.8
Nj /""\y gh/lAv/V/.
'
M6.0
1984.9.15
M6.2
1984.9.14
M6.8
10s
Model Wave
( Ricker Wave.
H = 1.75km)
Fig. 5 Velocity waves and its filtered waves observed

at the site TMA.
IMPLICATIONS OF RECENT ADVANCES IN
INSTRUMENTATION FOR STRONG-MOTION STUDIES
BY
R. D. Borcherdt
ABSTRACT
Studies in seismology and engineering within the accessibility of strong-motion data. Improved
last decade have emphasized the importance of hardware components permit greater frequency
available strong-motion data sets recorded near bandwidth, wider dynamic range, and less power
moderate eartyhquakes and the need to record im- consumption. Microcomputers allow software con-
proved data sets near large earthquakes in the trol of various hardware components, efficient
future. Recent advances in instrumentation per- and reliable execution of system tasks, increas-
mit strong-motion signals to be recorded over ed flexibility in system design, and increased
broader bandwidths, with wider dynamic range and digital processing capabilities. Modern storage
signal resolution, and greater data accessibil- media extend data capacity and provide compatible
ity via computer. These improvements in data format for mini-computer systems, which when de-
quality suggest that seismic displacements cal- ployed in the field permit extensive pre-proces-
culated from accelerometer recordings can be in- sing of large volumes of data during field exper-
ferred to longer periods, more stable estimates iments.
of seismic source parameters (e.g., moment, cor-
ner frequency, seismic source dimension) can be This paper describes a General Earthquake Obser-
developed, effects of crust and local site vation System (GEOS) developed to incorporate re-
structure on seismic wave propagation (e.g., cent advances in technology. Implications of
-1
wave conversion, f
x Q ,
and scattering) can
,
these advances for strong-motion studies, espe-
be determined more accurately, and non-linear cially bandwidth and signal resolution, are dis-
structural response, including high-frequency cussed in the contest of date recently recorded
(50-100Hz) signals associated with failure of on the GEOS.
structural components, can be better resolved.

A general microcomputer controlled data acqui- 2 . A BROAD BAND MICRO- COMPUTE R-BASED DATA
sition system (GEOS) has been developed by the ACQUISITION SYSTEM

U.S. Geological Survey and deployed to acquire The portable digital data acquisition system was
improved-quality strong-motion data sets. Data developed by the U.S. Geological Survey for use
sets acquired with the GEOS illustrate the sig- in a wide variety of both passive and active sel
nificance of broad frequency bandwidth and wide smic experiments (3). Versatility in system ap-
dynamic range for future strong-motion studies. plication was achieved by isolatlonof the appro-
priate system functions on hardware modules con-
Keywords: General Earthquake Observation Sys- trolled by a central microcomputer via a general
tem (GEOS), Strong-Motion, Instru- computer bus. CMOS hardware components wer.- uti-
mentation, Microcomputer, Data lized to reduce quiescent power consumption to

Acquisition, Earthquake Engineering, less than 2 watts for use of the system as either
Engineering Seismology a portable recorder in remote locations « i r in an
observatory setting with inexpensive backup i"wr

sources. The GEOS, together with two sets <>f
1. INTRODUCTION three-component sensors (force balance accelero

Recent advances in technology offer new oppor- meter and velocity transducer) and ferrite WVv
tunities for improving the quality and computer antenna, is shown in Figure 1.
106
The hardware modules comprising the system are achieved by using a ringed software architec-
shown in Figure 2. ture and modular hardware components. Incor-
poration of new hardware modules is accom-
The signal conditioning module for the GEOS is
plished in a straightforward manner by replac-
configured with six input channels, selectable
ing appropriate module and corresponding seg-
under software control, to permit acquisition
ments of controlling software. The flexibili-
of seismic signals ranging in amplitude from a
ty afforded by micro-computer technology to
few nanometers of seismic background noise to
modify the system for specialized applications
2g in acceleration for ground motions near
and to incorporate changes in technology al-
large events. The analog-to-digital conversion
lows seismic signals, detectable by a wide
module is equipped with a 16 bit CMOS analog-
variety of sensors to be recorded over a broad
to-digital converter which affords 96 dB of
band of frequencies with high resolution.
linear dynamic range or signal resolution;
this, together with two sets of sensors, im-
2.1 Implications of Broad Bandwidth and High
plies an effective system dynamic range of
Signal Resolution for Strong-Motion Studies
about 180 dB. A data buffer with direct mem-
ory access capabilities allows for maximum The system response designed for strong-motion
throughput rates of 1200 sps. With sampling applications of the GEOS (Figure 3) was in-
rates selectable under software control as any tended to allow large amplitude near source
integral quotientbroad and variable
of 1200, signals of 1-10 Hz as detected by a forced-
_5 2
system bandwidth ranging from (10 - 5x10 Hz) balance accelerometer (FBA) to be recorded on
is achieved for use of recorders with a wide scale, while at the same time permitting much
variety of sensor types. smaller amplitude high-frequency signals (30-
100 Hz) as might be detected on FBA's or velo-
Modern high density (1600 - 6400 bpi) compact
city transducers to be recorded with high sig-
tape cartridges offer large data storage capa-
nal resolution. The design system response
cities (1.25 - 33 Mbyte) in ANSI standard for-
together with that for two types of sensors
mat, to facilitate data accessibility via
frequently used for aftershock studies in the
mini-computer systems. Read capabilities of
near source region of large earthquakes is
cartridge tape recorders is utilized to allow
shown in Figure 3. With digitization rates
recording parameters and system operational
and anti-aliasing filters, selectable under
software to be changed automatically. Read
software control, the design system response
capability also allows systems equipped with
allows signals ranging from essentially DC to
modems to transmit data via telecommunications
500 Hz to be recorded at high resolution with-
to a central data processing laboratory.
out aliasing. Gain settings, selectable under
Microcomputer control of time standard pro-
software control, allow two sets of three-con-
vides capability to synchronize internal clock
ponent data, ranging in amplitude over 180 dB,
to internal receivers (such as WWVB and satel-
to be recorded simultaneously.
lite), external master clock, or conventional
digital clocks. Convenient system set-up and To date, the most extensive application of the
flexibility to modify the system in the field GEOS has been the acquisition of near-source
for a wide variety of applications is achieved aftershock data following earthquakes near
using a 32-character alphanumeric display un- Mammoth Lakes, California (M 5.2), Coalingn,
der control of the micro-computer. California (M 6.5), Newcombe ,
New York (M
5.2), Borah Peak, Idaho (M 7.3), and

Flexibility to modify the system to incorpor- g
Santiago, Chile (M 7.8). Several thousand
ate future improvements in technology is
events, recorded and analyzed during these
.107
studies, have provided an opportunity to test resolution and frequency bandwidth allow seis-
and develop many of the design features under mic signals as high as the corner on the anti-
field conditions. The several hundred after- aliasing filter (50 Hz) to be resolved on the
shocks recorded at ten GEOS stations in the GEOS recording.
epicentral region of the mainshock (M 6.5)
The velocity time history and displacement
near Coalinga, California provided on-scale
spectra inferred by McGarr et at. (1984) for a
three-component recordings of ground motion
recording obtained at another GEOS station of
for essentially all events ranging in magni-
the M 5.1 Coalinga aftershock is shown in Fig-
tude from less than 1 to 5.1 with the events
ure 5. The inferred displacement spectra with
greater than M 3.5 providing recordings with
a Nyquist frequency of 100 Hz, shows discern-
signal resolution greater than 40 dB. De-
ible seismic displacements ranging over more
tailed descriptions of the Coalinga data sets
than 7 orders of magnitude. The Increase of
are provided by Borcherdt [2,3]. Examples of
signal resolution and bandwidth illustrated In
data from these studies illustrate the signif-
Figures 4 and 5 provides a basis for inferring
icance of broad bandwidth and high signal
seismic displacement fields to longer periods,
resolution for studies of seismic source and
a basis for more stable estimates of seismic
wave propagation characteristics.
source parameters such as moment and corner
The largest aftershock of the Coalinga se- frequency, and an improved basis to estimate
quence (M 5.1, 5/9/83), was recorded at a effects of seismic wave propagation such as
distance of about 14 km on a standard strong- intrinsic absorption, f

m , and effects of
motion recorder (SMA-1) and a GEOS with sen- local site conditions. The improved capabili-
sors for the two recorders, separated by a ty to resolve high frequency strong motions
distance of about 25 m (Figure 4). The SMA implies improved estimates of non-linear
time histories were digitzed at 600 sps ,
fil- structural behavior, as well as high frequency
tered to prevent aliasing, and decimated to signals associated with structural failure.
200 sps. The GEOS time histories were re- as
Capability to resolve seismic frequencies
corded at 200 sps with high-cut anti-aliasing
high as 300 Hz can be useful for near-source
filters corners set at 50 Hz. As sensors with
studies in high Q environments. An example of
similar responses were used to record at each
such data collected in New Brunswick, Canada
location, comparison of time histories and
by E. Cranswick (pers. commun.) is shown In
acceleration spectra illustrate improvements
Figure 6. The velocity time history, recorded
in data quality permitted by improvements in
with high resolution at 1200 sps, was utilized
recording instrumentation. Utilization of a
to infer the displacement time history by re-
data buffer or pre-event memory, allows the
moving the base line, using the average of the
initial P wave onset to be recorded as well as
first 512 points, and straighforward Integra-
background /instrument noise levels. This in-
tion. The resulting displacement spectra sug-
formation, if routinely recorded on strong
gests seismic radiation as high as 2nn-25'i H/
motion recorders, could augment studies of
and emphasizes the significance of both haul-
seismicity and crustal structure as well as
width and dynamic range in defining the nature
studies of differential ground motion as re-
of near-source seismic radiation fields.
corded on closely spaced instruments. Com-
The ability to resolve periods of strong mo-
parison of the acceleration spectra shows that
tion longer than about 20 seconds have been
the spectra inferred from the two recorders
limited by capabilities of the sensor, sensor
are approximately the same for the frequency
coupling to the Earth, and
band 0.8-25 Hz, but that increased amplitude
available recorders. Recent advances in tech- Mueller, and G. Sembera played key roles in
nology to measure Earth strain over 7 decades acquiring the data sets for the aftershock
in frequency [6] up to frequencies of 5-10 Hz, sequences. M. Johnston has supervised the
may also prove useful in documenting the na- successful deployment of numerous Sacks-
ture of long period seismic radiation near the Evertson dilatometers throughout the State of
source of a great earthquake. For example, California, making the dilatometer data set
the Sacks-Evertson strainmeter [9] is sensi- discussed herein possible. Data for Figures
tive to strain signals over about 9 orders of 3, 5, 7, and 8 were prepared by R. Warrick, C.
magnitude in amplitude and periods ranging Mueller, E. Cranswick, and M. Johnston, res-
from months to 0.1 sec. Utilization of this pectively. Manuscript reviews by W. Joyner
sensor together with appropriate band-pass and G. Brady are appreciated.
filtering and a recorder with wide dynamic

range would allow examination of possible long 5. REFERENCES
period, near-source signals not discernable
[1] Aki ,
K. and P.G. Richards (1980). Quan-
using conventional recorders and accelerome-
titative Seismol ogy Theovy and Meth-
ters. As an example, Figure 7 shows the Earth
ods, vol. 1, W. H. Freeman and Co.,
strain as detected at a depth of 200 m near
San Francisco, Calif., 557 pp
San Juan Bautista, California on a Sacks-
[2] Borcherdt, R.D., C. Mueller, and L. Wen-
Evertson dilatometer and recorded on the GEOS.
nerberg (1983). Effects of local geo-
The upper time history shows the Earth tidal
logical conditions in the vicinity of
strains at 0.1 sps for a 28-hour interval and
Coalinga, California in Workshop on
the seismic energy radiated by a nuclear de-
Site-Specific Effects of Soil and Rock
tonation at a distance of about 400 km. The
on Ground Motion and Implications for
lower trace shows strain variations radiated
Earthquake-Resistant Design, July 26-
by the nuclear explosion plotted at an ex-
28, 1983, ed . by Walter Hays, U.S.
panded scale and at the sampling rate of 50
Geol . Sunv . Open-File Rept. 83-845.
sps used to record the data.
[3] Borcherdt, R.D., E. Cranswick, G. Max-
The broad bandwidth and wide dynamic range of well, C. Mueller, R. McClearn, G. Sem-
both the sensor and the recorder allow the bera, and L. Wennerberg (1984). Digi-
short period strain variations of 0.5-20 sec- tal strong-motion data recorded by
onds radiated by the nuclear explosion to be U.S. Geological Survey near Coalinga,
resolved as well as the large amplitude tidal California, in Coalinga, California,
variations of 12 and 24 hour periods. Such earthquake of May 2, 1983, reconnais-
signal detection capabilities offer new op- sance report, Fanthquake Fng . Res.
portunities for detection of near-source long- Inst. Rpt. 84-03, 89-98.
period variations as might occur in the epi- [4] Borcherdt, R.D., J.B. Fletcher, E.G.
central region of great earthquakes. Jensen, G. L. Maxwell, J. R. Van-
Schaack, R.E. Warrick, E. Cranswick,
3. ACKNOWLEDGEMENTS M.J.S. Johnston, and R. McClearn, A
This effort general earthquake observation system,

is the result of dedicated con-
tributions by number
submitted. Bull. Seismol . Soc. Am.
a of poeple to develop
and [5] Cranswick, E. R. Wetmiller, and J.
produce the GEOS (listed in alphabetical ,
order): Boatwright (1985). High-frequency

J. Fletcher, G. Jensen, G. Maxwell
(software design observations and source parameters of
and development), J. Van-
Schaack, small aftershocks recorded at hard
and R. Warrick. E. Cranswick, C.
109
rock sites, Bull. Seismol . Soc. Am.
(in press )
Johnston, M.J.S. and R.D. Borcherdt
(1984). Earth strain in the period
range 0.1 to 10,000 seconds at six
borehole sites within the San Andreas

fault system in California, FOS,
"Brans., Am. Geophys . Un., 65, p. 1015,

Fall Mtg., San Francisco, CA.
McGarr, A., C. Mueller, J. B. Fletcher,
and M. Andrews (1985). Ground motion
parameters of the 1983 Coalinga, Cali-
fornia earthquakes: implications for
crustal strength, U.S. Geol . Sum).
Open-File Ppt. 84-6ZQ, (in press).
Mueller, C.S., and Cranswick, E. (1985).
Source parameters from locally-
recorded aftershocks of the January 9,
1982, Miramichi earthquake. Bull.

Seismol. Soc. Am. ,
(in press).
Sacks, I.S., S. Sugehiro, D.W. Evertson,
and Y. Yamaguchi (1971). Sacks-Evert-
son strainmeter, its installation in
Japan and some preliminary results
concerning strain steps, Pap. Met.
Geophys., Tokyo, 22, 195-208.
Figure 1. Side and front-panel view of the General Earthquake Observation
System (GEOS), together with a WWVB radio antenna and two sets of three-
component sensors commonly used to provide more than 180 dB of linear dynamic
range. System operation for routine applications requires only initiation of
power. Full capability to reconfigure system in the field is facilitated by
simple operator response to English-language prompts via keyboard.
in
Figure 2. General system configuration for the GEOS, showing required system
functions isolated on separate hardware modules. Central CPU control of each
module via general computer bus permits full capability to configure system
for most field applications and facilitates incorporation of improvements in
technology.
112
Figure 3. Relative amplitude response of the GEOS recorder, GEOS with L4-C
velocity transducer and force balanced accelerometer (FBA) and square root of
power spectral density for Earth noise [1], Amplitude responses were deter-
mined for recorder with constant input voltage of 10 V, for recorder with L4-C
10 cm/sec or 10 V with sensitivity of 1 V/cm/sec at 10 Hz, and for recorder
with FBA with constant input acceleration of 2£. Two sets of sensors operat-
ing simultaneously and linear dynamic range of 96 dB for Earth gain setting
allow system to record 10 Hz signals with amplitudes ranging from 2 nm in
displacement to 2 in acceleration on scale without operator intervention.
113
COALINGA
3 f- 5/9/83
2
-i M - 5.2
100 § AZM- 135°
•
-
• i
! SMA SUB
A'y/V'AA/Vw-v^
CM/S**2
. o
01
* n 1 1 1 1 1 1 1 1 1 1 1 1 n
23456789
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
TIME-SECS
1 1 1 1 1 1 1 ii I 1 n 1 1 1 1 1 1 1 1 n 1 1 1 1 n I
10
CM/SEC
LOG
Figure 4. Strong ground motions and spectra recorded on GEOS and SMA-1 units
separated by a distance of about 25 m for 135° horizontal component of motion
generated by a magnitude 5.1 aftershock near Coalinga. The time histories and
spectra obtained on the two instruments are similar between 0.8 and 25 Hz.
Seismic energy is discernible in the SMA-1 spectrum as high as 25 Hz, and in
the GEOS spectrum as high as 50 Hz.
m
Figure 5. Time history of ground velocity
and displacement spectra
calculated
y c arc et al.
,
[7] from a recording of a H 5.1
aftershock near Coalinga,
° r " la ' The t,me h,stor
y inferred from a GEOS recording
J
mo ion as detected by a
of ground
forced balance accelerometer
at a distance of about 12
km from the hypocenter.
The seismic displacement field
is discernible over 7
orders of magnitude in
amplitude and up to freguencies as
high as 60-80 Hz.
115
0.1763E+05
VELOCITY
M<1
o A= 4 km
LU
New Brunswick
1200 SPS
« j
|;
(
>(i
L
I
g
Q
0.1680E +03
DISPLACEMENT
O
v
OK-
\>
g
Q
TIME (secs)
,s>
"O
X*
>
t
cn
2
LU
Q
<
cr
O
LU
a
w
O
CL
0.29 0.59 1.17 2.34 4.69 9.37 18.7 37.5 75.0 150 300 600
FREQUENCY (Hz)
Figure 6. Horizontal ground velocity, inferred displacement (integration,

only dc removed), and displacement spectra determined from GEOS recording of a
small earthquake (M < 1, A = 4 km) near Miramichi, New Brunswick. The com-
puter displacement spectra, with a Nyquist frequency of 600 Hz, shows a dynam-
ic range of 102 dB and detectable seismic frequencies as high as 300 Hz.
116
Figure 7. Variations in Earth strain detected by Sacks-Evertson dilatometer
and recorded on a GEOS unit at a site near San Juan Bautista, California.
Upper plot shows tidal strains at 0.1 sps for a 28-hour interval, and strains
associated with detonation of a nuclear device in Nevada at a distance of 400
km. Lower figure shows strain variations recorded from nuclear explosion,
plotted at an expanded scale and at 50 sps used to record the data.
117
DYNAMIC BEHAVIOR OF A ROCKFILL DAM DURING EARTHQUAKES
Norihisa Matsumoto
Nario Yasuda
Minori Shiga

ABSTRACT dam is thin and most of the dam is composed of

rockfill materials. The rockfill consists of
The Dam Department of PWRI, Ministry of Con- three zones, i.e. smaller, intermediate and ,
struction, has been continuing to measure the larger rocks. Figure 2 shows the grain size
earthquake motions at Miho Dam, which is a distribution of each zone. Since the fifty
rockfill dam with central impervious core, and percent grain size of the smaller rock zone is
has conducted geophysical investigations to about 5.00 mm, all zones are pervious, having
obtain dynamic deformation properties of the permeability coefficients larger than 1.0 x
_3
dam body. One of the authors presented the 10 cm/sec, which indicates that the rock
dynamic analysis of the dam by using the zones have free drainage ability. Core mate-
recorded accelerograms, at the 13th^ and rials of the impervious zone are well graded
2)
15th Joint Meeting of UJNR (both at Tsukuba) soils with gravels which were produced at the
At that time, however, we had not yet obtained stockpile by blending weathered rock with
laboratory test data on dynamic shear moduli scoriaceous tephra. The core zone was placed
and damping ratios for the materials of Miho in layers of 20 cm thickness and compacted by
Dam, and hence we used the material properties 12 passes of a sheepsfoot roller. The filter
of a similar rockfill dam. This time, we con- zone was compacted in layers of 40 cm thickness
ducted large-scale cyclic triaxial testing for by 12 passes of a 32 ton bulldozer. The rock
rockfill materials borrowed from the quarry zone was compacted in layers of 100 cm thick-
site of Miho Dam, and succeeded in obtaining ness by 4 passes of vibratory roller. The void
the dynamic shear moduli (or moduli of rigi- ratio, specific gravity and density of each
dity) and damping ratios for the strain range zone are listed in Table 1 . Each zone is com-
-5 -3
of 10 to 10 In this paper, we describe
. pacted to high density as shown in Table 1
the results of response analysis computed by
using the obtained laboratory material proper- The rock forming the foundation of Miho Dam is
ties and recorded base accelerograms as an andesitic pyroclastics resulting from submarine
input volcanic activity of the early Neogene, partly
intruded by pyroxene andesite dyke. The bed
Key Words: Earthquake Engineering, Rockfill rock is well indurated hard rock with a P-wave
Dams, Dynamic Analysis, Dynamic Behavior. velocity of 4.0 to 4.5 km/sec.
1. INTRODUCTION 2. LARGE SCALE CYCLIC TKTAXTA1 TESTS
Miho Dam is a 95.0-m-high rockfill dam with a Cyclic triaxial tests were conducted op th<-
central earth core; its maximum section is rockfill materials of Miho Dam. The main fr.i
shown in Figure 1 . The impervious zone of the tures of performance are shown in Table 2,
118
and Figure 3 illustrates the equipment. The h 1/2-ir • AW/W ( 2)
displacement transducer and load cell are

2)
placed inside the triaxial cell to exclude Here the value 0.35 of Poisson's ratio was
mechanical errors caused by friction from mea- used for Equation (1), because this value has
sured values. The two separated sensors, each been obtained independent of the depth from the
of which is a disk-shaped coil, are used as a -esults of geophysical investigation in Miho
displacement transducer. Since the capacity of Dam.
this displacement device is limited to 5 mm.
the position of the target is adjusted by Figure 6 shows the relation of shear modulus to
shifting it to conform to the axial displace- shear strain. The shear modulus becomes larger
ment during consolidation. The gradation of with increasing confining pressure. The data
the materials tested is shown in Figure 4, and points are rather scattered in the strain range
—6
is almost identical with the gradation of the smaller than 5 x 10 because of the inadequate
smaller rock, as shown in Figure 2. The diame- performance of the pickups. The shear moduli
ter of the specimen is 30 cm, and the height is obtained from the geophysical investigation are
60 cm. The materials are in an air-dried con- also plotted in Figure 6. Comparing the shear
dition, with a water content of 0.3 percent. moduli obtained from laboratory tests with the
The specimens were at first uniformly consoli- field geophysical investigation, the laboratory
2
dated with a confining pressure of 1.0 kgf/cm ,
testing gives us a little bit higher shear mod-
and then cyclic deviatoric axial stress a ,

of ulus than field testing at a given depth. How-
ad
0.2 Hz was applied to the axial stress ° ,
ever, the difference is so small that it can be
after the volumetric change rate was confirmed considered that the two testing methods give us
as negligibly small, i.e. less than 0.05 per- similar results. Here, we assumed that the
cent a minute. The twelfth stress-strain loop confining pressure in the laboratory tests rep-
of the cyclic loading was used for obtaining resents the stress in the dam at the depth of
dynamic deformation properties at each stress Z, through the following relations.
level. The deviatoric stress was increased

gradually at the same confining pressure, and Mean principal stress o' can be expressed by
the specimen was subjected to next confining the following equation by assuming the plain
pressure when the strain reached around strain deformation in the dam axis direction.
3
1 x 10~ .
o' = 1/3 (1+v) • ( 1+K) o'

m l
The four levels of confining pressures were

2
taken for a specimen as 1, 2, 3 and 4 kgf/cm \
where K = o'^/a'^ = principal stress ratio
The shear modulus G was computed as follows: The observed earth pressure in Miho Dam gives
us
E = o /e
d ad ad
a\ = 0.9 • o
v (3)
G = E /2 ( l+ v ) (1)
a
where a = overburden load
V
Y = (l+v) •
E
d
Therefore, o' will be related to depth Z as
m
The damping ratio h was obtained as follows (see
Figure 5)
119
'
= 0.9/3 (l+ v ) . (1+K) y . Z 4. DYNAMIC RESPONSE ANALYSIS
m
3
Y
= weight of unit volume (t/m ) The two-dimensional dynamic response analysis
Z = depth(m) from the dam surface was performed by using the results of labora-
tory cyclic triaxial tests and the field geo-
In Equation (3), the values of v > K and Y for physical investigation, and recorded earthquake
Miho Dam are respectively, 0.5, 0.35 and 2.31. motions. The computed response was to be com-
pared to the recorded response. QUAD-4 was
Figure 7 shows the normalized shear moduli used in the computation.
obtained by divding by G , a shear modulus cor-
q
responding to very small strain. The relation The initial shear modulus in the dam was given
between normalized shear modulus and shear according to the overburden depth based on the
strain in Figure 7 is almost the same for four results of geophysical investigation shown in
different confining pressures. Figure 11. Figure 12, in which two curves were
obtained from cyclic triaxial tests, was used
Figure 8 illustrates the relationship between for the G to y relation.
damping ratio and shear strain. The effects of
confining pressures on damping ratio are not The damping of the material itself was evalu-
obvious in this coarse granular materials, dif- ated by the hysteretic damping of the cyclic
fering from the case of fine clean sand such as triaxial testing, and another dissipation damp-
Toyoura standard sand. ing at the boundary between dam and foundation
was added to this hysteretic damping. After
Kanai^ , the dissipation damping between two
3. RECORDED EARTHQUAKE MOTIONS
layers is given by the following equation, when
Miho Dam was instrumented with two strong the excitation frequency is the resonant one.
motion accelerometers at the crest and down-
stream toe of the dam as in Figure 1 . Since 1+q
h (4)
1 -a
the accelerometer at the toe was placed on the
where
bedrock, the accelerograms recorded by this
p'V’s
accelerometer were used as an input motion in pVs
the dynamic response analysis.
q = impedance ratio
Miho Dam was subjected to considerable earth- 3
p' = density of upper layer (kgf/cm )
quake shaking at three times, although it suf-

G' = shear modulus of upper layer
3 4)
fered no damage from the shaking ’
. The 2
(kgf/cm )
main features of each event are listed in

p = density of lower layer (kgf/cm )
Table 3, and the epicenters are shown in Fig-

G = shear modulus of lower layer
ure 9. The recorded accelerograms were digi- 2
(kgf/cm )
tized at 0.01 second intervals and were

replotted in Figure 10. The peak acceleration
In Miho Dam, the S-wave velocity of the <Lim and
and predominant frequencies are shown in
foundation are 0.5 km/sec and 2.0 km/sec,
Table 4.
respectively; and the density of the d.iir .nn]
3 3
foundation are 2.3 kgf/cm and 2.5 kgf/cm ,
respectively. Therefore, the Impedance rath '
is 0.23, and h becomes about 15 percent. As
120
the dissipation damping at the boundary is not 5. CONCLUSIONS
taken into account in the computer program,
this 15 percent damping was added to the 1 . Due to the improvements of the large-scale
internal hysteretic damping. cyclic triaxial test apparatus, the shear

modulus and damping ratio of rockfill mate-
The finite element discretization of the dam is rials were obtained for the strain range of
illustrated in Figure 13. 5 x 10 ^ to 1 x 10
2. The shear modulus resulted from laboratory
Figure 14 shows the frequency response function testing was a little bit larger than that
of the dam crest to the foundation after the from the field geophysical investigation.
final iteration. The first and second natural 3. There was a good coincidence between mea-
frequency of the analytical model are 1.8 to sured and computed response in the case of
1.9 Hz, and about 3.3 Hz, respectively. The the East Izu Peninsula Off event by using
first natural frequency of the model is coinci- the QUAD-4 computation program. However,
dent with the measured data. much consideration should be given in the
case of high frequency input motion in the
Figure 15 shows the response acceleration of use of this program.
the dam crest. In the case of the East Izu
Peninsula Off Earthquake, the peak value of 6. R EFERENCES

acceleration is 70 gal in computation and
66 gal in measurement, so there is good coinci- 1) N. Matsumoto, S. Kondo, "Observed behavior
dence in peak acceleration, and an average and analysis of rockfill dam during earth-
level of amplitude during the main motion from quake," 13th UJNR, 1981.
7 sec to 12 sec. On the other hand, in the 2) N. Matsumoto, M. Toyoda, M. Shiga, "Distri-
case of the East Yamanashi Prefecture Earth- bution of the dynamic shear modulus in a
quake, the peak value from the computation is prototype rockfill dam," 15th UJNR, 1983.
34 gal, while the measured value is 87 gal, and 3) N. Matsumoto, S. Kondo, H. Katahira and
in the case of the West Kanagawa Prefecture M. Shiga, "Earthquake motions at rock
Earthquake, the computed value of 187 gal is foundations of dam sites," Technical memo-
smaller than the measured value of 253 gal. randum No. 1789, P.W.R.I., 1982 (in
The foundation motions of the East Yamanashi Japanese)
Prefecture and West Kanagawa Prefecture Earth- 4) N. Matsumoto, N. Yasuda, M. Shiga, "Char-
quakes contain much high frequency components, acteristics of earthquake accelerograms
because these two events occurred within ten recorded at damsites," Large Dam, No. 108,
and several km from the dam site. The computa- June, 1984 (in Japanese).
tion procedure might not be appropriate for 5) Katsutada Sezawa, Kiyoshi Kanai, "Damping
evaluating high frequency input motion. In the in Seismic Vibrations of a Surface Layer
case of the East Izu Peninsula Off Earthquake, due to an Obliquely Incident Disturbance,
1
the input motion did not contain much high fre- Earthquake Research Institute, Sep., 1936
quency components, and hence the computation No. 14.
gave us a good coincidence. However, the high

frequency components during the initial part of
the motion in the measured accelerogram do not
exist in the computation results.
121
6) Y. Sawada, T. Takahashi, A. Sakurai, and
H. Yajima, "Material Property Distribution
and Dynamic Characteristics of Rockfill
Dams — A consideration based on elastic
wave — Central Research Institute of
Electric Power, Report No. 377008, 1977
( in Japanese)
Zone
Core
— Void
Ratio
0.42
Table 1 Void Ratio, Specific Gravity
and Density for Each Zone
Specific
Gravity
2.86
Dry Density
3
(t/m )
2.01
Saturated
Density (t/m
2.31
1
)
Filter 0.25 2.88 2.30 2.50

Rock 0.26 2.92 2.31 2.52
Table 2 Main Features of Performance
Specimen size ^30cm x h60cm
Control system Electro-Hydraulic-Servo
Wave form Sinusoidal, triangular, and earthquake waves
Maximum loading capacity compression - 20 tf
Axial loading Maximum loading capacity extension - 10 tf
Frequency range 0.2 - 10 Hz

2
Max. 15 kgf/cm
Confining pressure
Frequency range 0.2 - 5 Hz
Table 3 Generals of Earthquakes
Epicentral
Date Magnitude Depth(km)
Dis tance(km)
East Izu Peninsula Off '80,6,29 6.7 10 56
East Yamanashi Prefecture ’81,4,14 4.5 20 13
West Kanagawa Prefecture '83,8,8 5.8 30 14
122
Table 4 Maximum Acceleration and Predominant Frequencies
Component Max. Acceleration Predominant Frequencies
F. Up/Downstream 30.9 (gal) 3.0, 4.0, 6.6, 7.9

<4-1
4-1
o F . Axis 33.0 2.2, 3.0, 4.1, 8.3
3
N F. Vertical 12.0 0.8, 2.2, ^3_, 6.8
CvJ
1—1 i—
v 3
u w D. Up/Downs tr earn 66.4 2.0, 3.2, 3.8
c/d c
-H
W 3 D . Axis 81.1 1.6, 2.7, 3.1, 5.8
0)
D. Vertical 37.6 2.4, 3.2, 3.9, 7.0
•H F. Up/Downstream 32.0 6.9, 8.5, 9.6

D5
3
<d
S-i
F . Axis 38.0 1.2, 3.1, 3.5, 6.8, 8.8
3 4-1 F. Vertical 11.0
E 0 1.2, 2.0, 4^, 6.8
3 0
>H
0 D. Up/Downstream 87.5 2.7, 5.3, 6.8
4-) J-i
Pu
C/D
3 D . Axis 104.1 2.9, 4.0, 5.2, 7_^

D. Vertical 73.0 3.3, 7.4, 8.5, 9.9
F. Up/Downstream 149.0 1.0, 2.6, 6.5, 7.5, 9.0

§ 0 F Axis 91.4
3 J- .
1.3, 2.3, 6.0, 8.0
W) 3
3
C 0
4—1
F. Vertical 53.6 3.3, 4.7, 5.9, 6.3, 11.3—
3 0
X
4J
C/3
0
0^
0
U
Ph
D. Up /Downstream
D . Axis
253.7
206.4
1.8,
1.7,
2.6,
2.3, 5.0, 6.2

—
5.3, 6.2, 6.9, 8.3
2
D. Vertical 165.8 2.8, 6.2, 6^, 8.3
Photo 1 Target for Displacement Photo 2 Target and Displacement

Sensor Sensor
123
ACCILEROGRAPH
SCALE Of METER
Fig. 1 Standard Cross Section
Groin Size ,
mm
Fig. 2 Grain Size Distribution
Fig 3 Cyclic Triaxial Testing Apparatus
m
<0 o CO OJ uO O
1ô CVJ CO o lO ro O 00
o - o o cr>
’£)
~ 2
o o
Gram size (mm)
Fig. 4 Grain Size Distribution of Specimen
Fig. 5 Stress Strain
025
Fig. 6 Shear Modulus (G) ^ Shear Strain
Fig. 7 G/Gq ^ Shear Strain
026
Ul
lO
6
2 5 1(T 5 2 5 1(T 4 2 5 lO' 3 2 5 10“
Shear strain y
Fig. 8 Damping Ratio Shear Strain
Fig. 9 Geographical Locations of Epicenters
127
Prefecture)
Earthquake
wnstrean,
Yamanashi
Do
/
Up
Observed
(East
Crest
10(b)
Fig.
Off)
Peninsula
Izu
(East
Earthquake
*
Observed
10(a)
Fig.
128
300.000
200.000
z
o
100.000 ll
£
cr
J
0.000
-100.000
aa.Ai
fiwyp
nil
f W 11 ,7 V
1 ll lln |)u
II
v
rr
^ - 200.000
<
V 11
*
-300.000
0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00 9.00 10.00 11.00 12.00 13.00 14.00 15.00
TIME IN SECONDS
Crest Up /Downs tr earn

75.000
50.000
z
o
25.000
£ 0.000
cc
UJ
J
cr
-25.000
-50.000
i 'vVV'
-75.000
0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00 S.00 10.00 11.00 12.00 13.00 14.00 15.00
time in seconds
Foundation Vertical
150.000
100.000
1 r
z 50.000
o
Si o.ooo
<d
cc
J -50.000
o
a -lOO.OOOl
-iso.oool
0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00 9.00 10.00 11.00 12.00 13.00 14.00 15.00
time in seconos
Foundation Up /Downstream
Fig. 10(c) Observed Earthquake (West Kanagawa Prefecture)
S-Wave Velocity m s
200 400 600
100 300 500 700
h
ratio
Damping
Shear Strain V
Fig. 11 Determination of Initial Fig. 12 Strain Dependency of Shear Modulus

Shear Modulus and Damping Ratio for Computation
(S-Wave Velocity)
129
5.33
4.67
4.00
3.33
2.67
2.00
1 . 33
0.67
0.00
0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00 9.00 10.00 1.00
FREQUENCY IHZ)
East Izu Peninsula Off
6.00
5.33
4.67
4.00
3.33
2.67
2.00
1 . 33
0.67
0.00
0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00 9.00 10.00 I 1.00
FREQUENCY (HZ)
East Yamanashi Prefecture
6.00
5.33
4.67
4.00
3.33
2. 67
2.00
1 . 33
0.67
0. 00
0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00 9.00 10.
FREQUENCY IHZ
West Kanagawa Prefecture
Fig. 14 Calculated Frequency Response Fun< t i >n .it Crest
130
I Kail
20
20
0.00 2.00 .1.00 0.00 8.00 10.00 12.00 11.00 10.00 18.00 20.00
Vertical
1
loo
80 i
00
j
.
10
20
o aA a. a :
L J\
rj|ww Srv
JJl Mk M Ai wv A A/l. JL
M fWr
ft/p
20
wjn
T
IF /*\fv V
10 i
00
80 i
| O0 !
0.00 2.00 1.00 0.00 8.00 10.00 12.00 11.00 10.00 18.00 20.00
Up/Downstream isi-.ci
Fig. 15(a) Calculated Response Acceleration at Crest

(East Izu Peninsula Off)
20. 00 [ Ra 1 I
o.oo . A._ Av ^^
\J
'
A/\
|*|J
- 20.00
0.00 2.00 4.00 6.00 8.00 10.00
TIME IN SECONDS
Vertical
a 1
40.00 I
20. 00 L
0.00 [\ IS
^ A A aA AAA Ai
Jin /I L
vv \|yirv
V \Jv
- 20. 00
-40.00
0.00 2.00 4.00 6.00 8.00 10.00
„ _ time in seconds
Up/Downstream
Fig. 15(b) Calculated Response Acceleration at Crest
(East Yamanashi Prefecture)
lga11
100.00
Aa . AAA/iA*. A . JUl rtAA. V!rt

-V’A/V pyvWSrY |
0.00 2.00 4.00 6.00 8.00 10.00

TIME IN SECONDS
Vertical
o.oo 2.00 4.00 6.00 8.00 10.00 12.00

TIME IN SECONDS
Up/Downs t ream
Fig. 15(c) Calculated Response Acceleration at Creast

(West Kanagawa Prefecture)
131
VIBRATION TEST OF RICHARD B. RUSSELL CONCRETE DAM
AFTER RESERVOIR IMPOUNDMENT
ov
UX
R. J. SMITH
ABSTRACT with the foundation and reservoir can

significantly affect earthquake response
A second forced vibration test of Richard B. calculations. The vibration tests provide the
Russell Dam was conducted after a reservoir was means for determining the dynamic properties of
impounded. The dynamic properties of a concrete the Richard B. Russell concrete dam and
gravity dam with hydrodynamic interaction were evaluating the various parameters that influence
experimentally measured. Richard B. Russell Dam these properties.
has recently been built by the Corps of Engineers
approximately 170 miles (272 kilometers) from the
mouth of the Savannah River between the States of The concrete dam includes a spillway section, a
Georgia and South Carolina. The crest of the non-overflow section, and an intake section. The
concrete gravity part of the dam is 1884 feet crest of the concrete gravity part of the dam is
long (574 meters) and is composed of 13 1884 feet long (574 meters) and is composed of 13
nonoverflow, 8 intake, and 11 spillway monoliths, non-overflow, 8 intake and 11 spillway monoliths,
the tallest of which is approximately 200 feet the tallest of which is approximately 200 feet
high (61 meters). The average reservoir height (61 meters) high.
was 174 (52 meters)feet during the tests.
Forced vibration tests were made on the Richard
B. Russell Dam before and after impoundment of
The structure was excited at three locations by a the reservoir to determine the natural
crest mounted 17,000 pound (7711 kilograms) frequencies, mode shapes, modal damping ratios,
inertial mass which was driven by an and relative joint movements. The first
electrohydraul ic servo-controlled actuator. The low-level forced vibration test was conducte
force input to the dam was computed as the January and February 1982 which seized a rare
product of the measured acceleration and the opportunity to experimentally measure the dynamic
17,000-pound (7711 kilograms) mass. Servo properties of a concrete gravity dam without
accelerometers with sensitivities ranging from hydrodynamic interaction. Another forced
0.25 to 5.0 volts per g measured the horizontal vibration test was performed in June and July
accelerations of all 32 monoliths and the 1984 after impoundment of the reservoir. This
distributions of horizontal acceleration with provides an experimental measure of the prototype
elevation in the three drive point monoliths. hydrodynamic interaction and a comparison of the
Dynamic pressures were measured during two tests changes of the dynamic properties. The validity
at the upstream face of the tallest nonoverflow of 2-D and 3-D finite element analyses were
monolith. Piezoelectric pressure transducers checked against the experimentally measured
with a charge sensitivity of approximately 150 dynamic properties.
pico-coulombs per psi measured the dynamic
pressures The dynamic properties of the dam will also ho-
used to provide a record of the dam's structural
condition as built. In the event an earthquake
The results reveal the dynamic characteristics of should damage the dam in the future, the tesr
the complex structure with hydrodynamic data will be useful in determining the
interaction. The natural frequencies, mode significance of the damage.
shapes, and damping ratios were determined.
Results were compared with the previous test of
the structure without the reservoir impounded. 2 . EXPERIMENTAL EQUI PMENT
Relative joint motion was observed above a low
response threshold. Richard B. Russell Dim was
(7,711— kg . ) inertial mass which was m
steel plate which was epoxied to the n.e,'.
Key Words: Forced Vibration Test; Frequencies; concrete of the dam. The force input to t d.m
Mode Shapes; Damping Ratios; Transfer was computed as the product of the measured
Functions; Accelerations acceleration and the 17,000-lb. mass.
Accelerometers were used to measure the

1 . INTRODUCTION horizontal accelerations of the concrete he*.
Most of these accelerometers have sennit vlt>
. vnl
understanding of the significant parameters that per g accelerometers. The accelerometers have a
influence the dynamic properties of such
structures is necessary. Assumptions regarding Office Chief of Engineers
geometry, boundary conditions, and interaction Washington, D. C.
132
frequency response range from zero to 500 Hz. condition to provide enough data in a given time
Analog signals were recorded on a tape and later interval for adequate digital time resolution.
digitized on a structural dynamics analyzer. Each test lasted about 5 minutes; but there were
many hours of preparation required to conduct a
test for a given array. The electrical signals
Peizoelectric pressure transducers were used to output from the accelerometers and the vibrator
measure the hydrodynamic pressures. The were
4. recorded as analog signals on a 32-track
3.
transducers were mounted in special waterproof tape recorder.
canisters bolted to the dam face by divers.
DATA REDUCTION AND EXPERIMENTAL RESULTS

VIBRATION TESTS
To conduct a modal analysis (obtaining the
The dam was excited at three different monoliths natural frequencies, damping ratios, and mode
by the crest-mounted inertial mass. The force shapes) the recorded data were played back into a
input was in an upstream-downstream direction. structural dynamic analyzer to digitize the
The three exciter locations were at monolith 7, a analog signals. The force input and one of the
nonoverflow monolith with the mass mounted on the response-accelerometers output formed an
crest at elevation 494 feet (150 meters); input/output (I/O) pair to compute a frequency
monolith 16, a monolith between intake and response (transfer) function and coherence
spillway with the mass mounted in a recess at function. Typical frequency response (FRF) and
elevation 483 feet (147 meters); and monolith 22, coherence functions are shown in Figures I and II
a spillway monolith with the mass mounted on the respectively. The coherence function is used to
crest at elevation 494. reveal the existence of any signals which may be
contaminating the transfer function measurements.
Modal parameters are identified in the analyzer
Ideally it would be desirable to excite the dam by mathematically matching (curve fitting) an
at one location and mount enough transducers at analytical form of a transfer function to
all points of interest to measure the response of measured data.
the structure. This was not possible because of
the limited number of accelerometers available. Hydrodynamic pressures were measured at monolith
The first set of crest and cantilever tests were 7 with the exciter positioned on monolith 7. The
performed starting with the 17,000-lb. inertial piezoelectric transducers were placed at six
mass located on the crest of monolith 7. The elevations on the upstream face. The
first array of accelerometers was placed along hydrodynamic pressures were successfully recorded
the crest at the center of monoliths 1 to 16 to during the low level vibration test in the
monitor the horizontal upstream-downstream frequency range of 2 to 25 hertz. To date the
motion. The 17,000-lb. mass was then moved to data has not been reduced. This will be included
monolith 16 and the accelerometers were placed in the final report.
along the crest at the center of monoliths 7 to
22. So the accelerometer locations overlap on
the crest of monoliths 7 to 16. The inertial Ambient vibrations were present during the test.
mass was next moved to monolith 22 and the 5.
The main cause of these vibrations was water
accelerometers were placed along the crest at the flowing over the spillway and through the low
center of monoliths 16 to 32, so the level sluices. The ambient vibrations were
accelerometer locations overlap on the crest of recorded and will be accounted for as the
monoliths 16 to 22. spectral density functions are prepared.
The accelerometers were placed at different Work on the modal analysis is still underway.
elevations on both the upstream and downstream Preliminary results indicate a shift in three
faces of monolith 7 to measure the horizontal natural frequencies as shown in Table 1. Also
cantilever motions. The accelerometers were shown are the frequencies arrived at by three
placed at different elevations in the access dimensional finite element analysis of the dam
shaft of monoliths 16 and 22. without the reservoir. This analysis is
discussed in detail in the attached Appendix A.
The forced vibration test results were obtained
Relative joint motion was measured at three by using a peak-picking method on tjie imaginary
locations and consisted of two accelerometers part of the frequency response function. The
closely spaced on each side of the monolith joint preliminary results also indicate no significant
being monitored. The measurements were made at change in the damping ratios from values obtained
non-overflow monolith joint 15-16 and intake in the first forced vibration test.
crest joints 13-14 and 14 and 15.
DISCUSSION OF RESULTS
All the tests were 1/2 decade/minute sine sweeps
for 20 Hz to 2 Hz. This rate was found to be Figure III shows the mode shapes in plan view of
slow enough for a required quasi-steady state the crest of the dam for monoliths 8 to 21.
133
Figures IV, V, and VI show the comparative
results of the dam crest mode shapes for both
forced vibration tests and the 3-D finite element
analysis. Changes in the maximum response
location of the dam and a shift in frequencies
occurs when there is a reservoir impoundment.
6. CONCLUSIONS
Since gathered data are still being reduced, only

preliminary conclusions can be reached at this
time. The final report for this study will
include a complete analysis of the prototype
tests results and a three-dimensional finite
element analysis comparison including
hydrodynamic interaction effects. To date we can
conclude that:
a) Added water mass caused a decrease in the

frequencies evaluated thus far.
b) The abutment on the Georgia side of the

dam was not complete at the time the first
test was made. The added mass of the
additional soil caused a slight increase in
the stiffness of the structure at the
Georgia-side abutment.
c) The powerhouse is located at the downstream

toe of the structure. It was not completed
at the time of the first test but was at
the time of the second test. This affected
the rigid body modes the greatest.
d) The 3-D finite element analysis estimated

frequencies higher than those measured
experimentally. This is reasonable since
the FE models are an approximation of the
dam using a discrete number of degrees of
freedom
e) The best mode shape correlation was between

the results from the experiment and the 3-D
FE flexible foundation model for the dam
with no reservoir.
SUMMARY OF FREQUENCIES
(ALL IN HERTZ)
MODE PROTOTYPE TESTS 3-D FE ANALYSES

NO. 1ST 2ND FIXED BASE FLEXIBLE BASE
1 5.9 4.0 7.3 6.9
2 6.8 5.3 8.2 7.7
3 7.6 6.6 9.7 9.1
c
x
r\i
1353 12 031
TRANSFER FUNCTION 'HZ)
FIG. l EL 495, INPUT MONOLITH 16, OUTPUT MONOLITH 13-
135
FIG. 2 EL 495, INPUT MONOLITH 16, OUTPUT MONOLITH 13-
136
UPSTREAM
o
O
o
o
° o
o o o o
o o oo
4.0 Hz
8 16 21
o o o
O
o
° o
8 O 16 21
° o 5.3 Hz
o
°
o o °o
GA. S.C.
o o
o o
o
o o
o o
o 6.6 Hz
8 o 16 21
o
O
DOWNSTREAM
o
CREST PLAN VIEW OF MODE

SHAPES FROM MONOLITHS 8-21
FIG. 3
157
CREST PLAN VIEW OF LEGEND
MODE SHAPE 1
1ST TEST
FIG, H
CREST PLAN VIEW OF

MODE SHAPE 2
13S
CREST PLAN VIEW OF
MODE SHAPE 3 LEGEND
138
APPENDIX A
LINEAR FINITE ELEMENT COMPARISON WITH EXPERIMENTAL MODAL ANALYSIS

FOR A CONCRETE GRAVITY DAM
Vincent P. Chiarito
Research Structural Engineer
U. S. Army Engineer Waterways Experiment Station
P. 0. Box 631
Vicksburg, Mississippi 39180
ABSTRACT The crest length of the concrete dam is

1.884 feet (574 meters) and is composed
Three-dimensional finite element <FE> of 13 nonoverflow. 8 intake. and 11
analyses of a concrete gravity dam were spillway monoliths. The tallest monolith
conducted to determine the dam's natural is approximately 200 feet (61 meters)
frequencies and mode shapes. The results high.(l) Two-dimensional (2-D) and 3-D
of the linear elastic FE analyses were parameter studies were conducted to
compared to the dynamic properties obtain appropriate FE grid sizes for
computed from experimental results. Two each of the three differently shaped
models for the foundation were used. The monoliths. A compatible grid was found
first model used a fixed base for the for all of the monoliths . thus making
foundation (or a rigid foundation) and possible the generation of the nodal and
the second used vertical and horizontal element data with the computer. Figures
spring elements. The springs were used 1 through 3 show typical sections of the
only to qualitatively model the effect monoliths and the resulting 3-D grids of
of dam-foundation interaction. A modal each monolith. In Figure 4 the resulting
assurance criterion was used to evaluate 3-D FE grid of the Richard B. Russell
the correlation between mode shapes Dam is shown. The 3-D FE model is
computed from the FE dynamic analyses composed of 3-D solid brick. beam. and
and experimental results. plate elements. The comparison between
the analytical (FE) and the experimental
INTRODUCTION mode shapes was evaluated using a modal
assurance cr l ter i on. 2
The results of linear elastic
three-dimensional (3-D) dynamic finite FINITE ELEMENT MODEL
element (FE) analyses of the Richard B.
Russell Dam with an empty reservoir were The grid for the FE model was developed
compared to the experimental results of by conducting 2-D and 3-D modal analyses
the first prototype vibration test.(l) of each isolated monolith. For each
This comparison provides one basis for monolith uniform 2-D grids were
assessing the accuracy of linear elastic constructed and the FE modal analysis
FE models for predicting dynamic results were compared to those resulting
properties of concrete gravity dams. A from the grids shown in Figures 1. 2.
subspace iteration method was used to and 3. The parameter studies involved
solve for the undamped natural using dynamic material properties
frequencies and mode shapes in the FE determined by previous tests on standard
ana vses
1 6- bv 12-in. (152- by 305-mm. concrete
)
cylinders. (3) Favorable results were

The dairi was modelled with its in situ obtained when the 2-D and 3-D modal
boundary conditions present at the time analyses using the uniform grids were
of the vibration test.(l) Two FE models compared to analyses using the grid for
bounded assumptions for the stiffness of the entire 3-D FE model .
the foundation! (l)fixed
f oundat l on-c omp e te v
1 1 rigid and (2) Three element types were used to model
flexible foundation-modelled with the dam. (4) Var 1 ab e-numiber-node s (8
1 to
vertical and horizontal springs. No 21 nodes) thick shell and 3-D elements
attempt was made in this paper to were used to model the mass concrete of
completely deter mine the effects of the nonoverflow .intake and seillwav
.
dam-foundation interaction during monoliths. Beam elements were used t

earthquakes. model the roadways connecting the
concrete piers of the spillway monoliths CONCLUSIONS
at the crest elevation. Plate elements
were used to model the concrete snllwav A 3-D linear FE analysis has been
piers. Figure 2b shows the penstock tube compared with previous experimental
openin 9 modelled as elements with a estimates of the modal parameters of a
smaller stiffness than the mass concrete gravity dam. Usin9 available
concrete's stiffness throu9h the width dynamic material properties of the dam
of the intake monoliths. Fi9ure 3b shows concrete and the foundation. a 3-D FE
the locations of the beam and plate model was successfully developed to
elements for a 3-D FE grid of two estimate the linear elastic modal
spillway monoliths. Once the compatible properties of the dam. The modal
rid was found for the interfacing assurance criterion was computed for
Sections imons the three monoliths. the mode shape comparisons between the
nodal point and the element data were experimental and the FE results. The
generated with a computer. values computed using the modal
assurance criterion ran9ed from 0.29 to
The boundary conditions for the fixed 0.72. Quantitative comparisons indicate
base model were such that all nodes that a reasonable 3-D FE model was
al ons the base had all degrees of developed from available material
freedom deleted. The nodes at the two properties for computing the dam's first
dam-abutment sections . except at the natural mode of vibration. A mode shape
base. had degrees of freedom defined in that was not measured experimentally was
the plane of each section. (3) Vertical computed by the 3-D FE analyses. The
and horizontal springs were attached to modal assurance criterion was useful for
the previously fixed nodes of the base. evaluating the correlation among the
The spring constant was computed from mode shapes. The best mode shape
available results for the dynamic correlation was between the results from
elastic modulus of the foundation. (3) the experiment and the 3-D FE flexible
The springs allowed motion only in the foundation model.
vertical plane perpendicular to the
longitudinal axis of the dam. The ACKNOWLEDGEMENTS
degrees of freedom at the dam— abutment
interfaces were still allowed in the The results presented herein, unless
plane of those two sections of the dam. otherwise noted, were obtained from
A total of 5.190 and 5.606 degrees of research conducted under the Structural
freedom were used for the fixed base and Engineering Research Program of the U.
the flexible base FE models, S. Army Engineer Waterways Experiment
respectively. Station. Permission was granted by the
Office. Chief of Engineers to publish
RESULTS this information.
The 3-D shapes shown in Figure 5 are the

first five normal modes of the dam's
vibration assuming a rigid foundation REFERENCES
model. At 8.1 Hz the major response is
associated with the sideswav of the (1) Chiarito. Vincent P. , and Mlakar,
roadway and concrete spillway pier Paul F."Modal Test of a Concrete
system. The other mode shapes show that Gravity Dam, “ Proceedings of the 2nd
the major responses of the dam are International Modal Analysis Conference.
associated with bending of the dam Volume 1. pp. 142-148. 1984.
perpendicular to the longitudinal axis
of the dam. Mode no. 5 at 10.3 Hz is the (2) Allemans. R. J. .and Brown, D. L.
shape computed by the FE analysis but "A Correlation Coefficient for Modal
not measured exper imental y. This was Vector Analysis." Proceedings of the 1st
suspected due to the possibility of the International Modal Analysis Conference,
exciter locations existing at nodes of pp. 110-116. 1982.
that mode shape during the vibration
test . 1< The fixity of the dam abutment
)
(3) Chiarito, Vincent P. and Mlakar.
.
interfaces was not modelled so the crest Paul F. "Vibration Test of Richard B.
of the dam at the ends is free. Russell Concrete Dam Before Reservoir
Impoundment," TR-SL-83-2. May. 1983.
Comparisons of the mode shapes are shown
in Figure 6. The solid symbols represent (4) Bathe. K. J. , Wilson, E. L. and
,
the normalized mode shape vectors of the Peterson. F. E. "SAPIVi A Structural

experimental results. The solid and Analysis Program for Static and Dynamic
dotted lines respectively represent the Response of Linear Systems." University
normalized mode shapes for the flexible of California, Earthquake Engineering
and fixed foundation models. Table 1
Research Center. Richmond, California.
summarizes the results of the 1973.
comparisons in Figure 6. In Figure 6a
the comparison of the results from the (5) Chiarito. Vincent P. Wang, Xianru.
.
three estimates of mode no. 1 is and Chen. Yangyou . "Linear Elastic.

favorable. Modelling the foundation Three-Dimensional Finite Element Modal
flexibility improved the correlation Analysis of Richard B. Russell Concrete
between mode shapes as computed bv the Dam Before Reservoir Impoundment."
modal assurance criterion value. (2) The Technical Report SL-85-nn. U. S. Army
frequencies estimated by the FE models Engineer Waterways Experiment Station.
are higher than frequencies measured Vicksburg. Ms., 1985.
experimentally. This is reasonable since
the FE models are an approximation of
the dam using a discrete number of
degrees of freedom. The FE model in this
case Is stiffen than the actual dam. For
a qualitative comparison the mode shapes
all seem to have the largest amplitude
in the same general location on the
crest of the dam. The modal assurance
criterion does indicate quantitatively
how good the comparison is among modes.
141
TYPICAL NONOVERFLOW
FIGURE 1 a . View of typical nonoverf lou b. View of typical 3-D FE and of one
section# nonoverflow aonolith.
b
b
b
b
,*5
f-IbUh'E 2j. View of typical intake b. View of typical 3 d FE arid of o»t

y ec 1 1 on intake aonolith.
VQ
i Sonnet rood ana
^15
FIGURE 3 a. View of typical spillway b. Uieu of typical 3-1' FE and of one

section. spillway monolith.
monoliths
nonoverflow
~ ~~
•
n uj
-s.c. Ga
UNDEFORMED SHAPE FROM
u
U.S
rAnni INA SIDE
FIGURE 4. Uiew of 3-D FE arid of entire dam from downstream D . S . ) Georgia Ga.
and upstream U.S. South Carolina S.C.' sides.
143
NORMALIZED onn lenoth
FIGURE 6. Plan view of normalized mode shapes at the crest.
1 ABLE 1. Summary of mode shame comparisons.

REFER
TO PROTOTYPE U I BRAT 1 ON 3-D FE DYNAMIC ANALYSES
FIGURE TEST 3) l FIXED BASE flexible BASE
6 MODE FREQUENCY MODE F REuUENCY M ACN * MODE F RE QUENL M ACN »
PAR T NUMBER H: NUMBERV . ) i Hr . ) NUMBER (Hr. )
a 1 b 9 7 0.67 1 6 0
b — 6.8
. 1
2 8.0
.
0 . 29 7
7.7
.
v
.
.
72
J3
c •
i 7.6 7 9 7 0 S4 A
. .
9. 1 i) 6 C*
LUKKELll 1UH tV 1 L Ut L U ME' U 1 ED El MUDAL ASSURANLt UKlItK ION 2
116
EARTHQUAKE RESISTANT BRIDGE DESIGN IN CALIFORNIA
by
James H. Gates (1)
ABSTRACT
2. DEVELOPMENT OF SITE-SPECIFIC RESPONSE SPECTRA
The current seismic design criteria for bridges
in California at CALTRANS is discussed. A The CALTRANS seismic criteria (3), permits the
second criteria which now being used in the development of special design response spectra
United States outside of California, which was for specific sites. The spectrum is a 5% damped
developed by the Applied Technology Council, is elastic acceleration spectrum which describes
also discussed This second criteria was recently the maximum expected response at the surface of
adopted by the American Association of State the site. This spectrum is referred to as an
Highway and Transportation Officials (AASHTO), 'ARS' spectrum; where AR is the acceleration
as a national guideline specification. The spectrum in rock-like material at the site and S
procedure used in California for the preparation is the soil amplification factor (9,10,17).
of site-specific response spectra is also
discussed. Other recent developments discussed The use of site-specific spectra is usually
include the application of interactive computer limited to large projects in high seismic
programs to column and foundation design, regions. For example: The Route 105 Freeway,
current column and footing details, the current (or Century Freeway), to be located in Los
status of the California seismic retrofit Angeles will extend eastward from the Los
program for bridges and information on results Angeles International Airport approximately 18
from the the instrumentation of California miles (29 km). The estimated cost of this
bridges for strong-motions. project is about $1.5 billion (US), including
approximately $450 million (US) for over 200
1. INTRODUCTION structures. The project is expected to be
completed in late 1992.
This paper is an extension and update of the
previous paper presented at the UJNR in 1983 A series of site-specific response spectra were
( 11 ). developed by CALTRANS TRANSLAB (6), for use in
the design of the structures along the 18 mile
The current seismic design criteria for bridges (29 km) length of the project.
in California is based on work at CALTRANS after
the 1971 San Fernando earthquake (9,10,17). An extensive geotechnical investigation was
Later efforts by the Applied Technology Council performed, evaluating the materials along tin-
(16), resulted in a second criteria which now length of the project (12). The surflcial
being used in the United States outside of Cali- deposits along the length of the project wen-
fornia. This second criteria was recently then characterized into three regions (6):
adopted by the American Association of State
Highway and Transportation Officials (AASHTO), 1. Recent dune dep osi t s at the westerl y e nd of
as a national guideline specification (13). the project.
Although there are differences between these two
criteria, the underlying philosophy of each are 2. Upper-Pleistoce ne terrace deposit s In the
identical. Both utilize elastic force level middle portion of t he project.
spectra with a post-analysis reduction for
ductility. The design force in a particular 3. Recent alluvial de posits at the e/i st<e r l y
member is thus dependent on the seismicity, site portion of the proj ect
conditions and the post-yield characteristics of
the member itself. Both criteria permit the Eleven active fau Its were considere d In the
development of special site-specific elastic evaluation of the sei smlcity of the p ro Ject
spectra. area. One of the se f aults, the wp o rt- Ingle-
fJ c*
wood, crosses the proposed freeway AliK nme nt In

This paper will describe the procedure used in a NW-SE direction.
California for the preparation of site-specific
response spectra Other recent developments A tabulation of the active faulIts c oris ld<p red
discussed include the application of interactive showing the distances t o the proj«»ct are a and
computer programs to column and foundation as sumed magnitudes and durations are sh OW1a in
i
design, current column and footing details, the Table 1.
current status of the California seismic retro-

fit program for bridges and further information Sensitivity studies wer e conducted to dot r»• lne i*
on results from the the instrumentation of Cali- the effect of the d 1st ant faults r>n the m ot ions
fornia bridges for strong-motions. at the project site and It was ccmr 1 o <i.-d that 1
(1) Structural Mechanics Engineer California Department of Transportation O M I- \\

Division of Structures, Office of Structures Design
the motions from the nearby Newpor t-Inglewood eration at the site, A, and the computed
fault represented a conservative envelope of the soil amplification, S.
response spectra for the range of rock motions
at the site. All soil columns were analyzed using a depth to
rock-like material of 300 feet (92 m) . This
The field investigation for the ground response depth was selected after performing a series of
analysis consisted of thirteen, 6 inch (152 mm) sensitivity studies which evaluated the effects
holes drilled to depths of 200 to 350 feet (61 of depth variations to 600 feet (183 m) (6).
to 107 m) (6).
,
In addition to densities and
standard penetration tests, shear wave velocity Smooth envelope spectra were prepared for each
profiles as deep as 350 feet (107 m) were deter- of the six zones. A comparison plot of the ARS
mined using the procedure developed and docu- curves developed for the project is shown in
mented by Beeston and McEvilly in 1977 (1). This Fig. 1.
procedure determines the velocity of a surface
generated shear wave which is propagated down- 3. COLUMN AND FOOTING DESIGN
ward to a geophone located in the bore hole.
Several soil and shear velocity profiles were Recent criteria for the seismic resistant design
averaged to obtain a series of composite of bridges (3,13) requires that any connections
profiles for a series of six zones along the which are expected to undergo plastic deforma-
length of project. These zones are shown in tion be able to develop the plastic moment and
Table 2. Boring logs for structures along the resultant shear as a limiting value. In Califor-
project length were reviewed to assure that nia, this means that the plastic forces usually
computed spectra would be representative of become the limiting design values to be used in
ground motions for the sites within the zone. the design of the connections. This load case
when evaluated alone, tends to simplify the
Bedrock motions within the six zones were estab- design of the connection.
lished by considering the closest distance from
the causative fault to the zone and the maximum In reality however, this requirement adds anoth-
credible magnitude (See Table 1). The charac- er level of complexity to the design process and
teristics of the bedrock motions used for each the design engineer must now consider plastic
zone were computed based on work by Seed et al. loads in addition to evaluating loads at the
(22,23). factored and service level.
Two earthquake time histories were utilized to At CALTRANS the design of the column member is
describe the bedrock motions; a deconvoluted highly automated. An interactive computer
1940 El Centro rock motion by Seed and an arti- program, YIELD, (7) is used to evaluate the
ficial rock motion generated by Romstad (20). column member for factored, service and plastic
loads and determine vertical, shear and confine-
Surface motions were computed by propagating ment reinforcement.
both of these rock motions to the surface of a
one-dimensional soil column. The current CALTRANS design specification (3),
requires the extension of the confining spiral
The computer program, SHAKE (21) was used for reinforcement into the adjacent cap or footing.
this analysis. The details for a typical single column/footing
and column/cap connection are shown in Fig. 2.
The computation of the surface (or ARS) spectra Shear and confinement reinforcement consists of
was performed using a procedure similar to that continuous closely spaced spirals (usually #6 or
used to develop the CALTRANS standard ARS curves 19 mm bars). These spirals are spliced by weld-
(3,17). Recent investigators (2,14) have ing or 80 diameter lap splices with 135 degree,
described this procedure as capable of obtaining 10 inch (254 mm) hooks at ends. This will
a reasonable approximation of the motions at a assure that the spiral will remain effective
site. The procedure is as follows: when the column is subjected to large excursions
into the ductile range. Rectangular columns are
1. Analyze a representative soil column for reinforced for shear and confinement by using a
the bedrock motion using the SHAKE series of interlocking spirals spaced not great-
program ( 21 ) er than 0.75 times the spiral diameter.
2. Compute the 5% damped elastic response The design of the footing is also automated at
spectra for both the applied rock motion CALTRANS using the program, FOOT, (8). Footings
and the computed surface motion. are evaluated for factored, service and plastic
loads from the column. Plastic moments are
3. Compute the soil amplification curve applied at 15 degree intervals to consider the
(S curve) by dividing the surface spec- effect of plastic action in any direction.
trum by the rock spectrum.
Top footing reinforcement is designed for all
4. Compute the surface ARS spectrum by imposed loads including uplift on piles and the
multiplying the CALTRANS normalized rock weight of footing and overburden.
spectra, R (17), by the peak rock accel-
147
Allowable soil stresses or pile loads vary
depending on the class of loading:
Allowable soil stresses or Pile Loads In Terms of Ultimate (3)
Foundation
Material Loading Allowable (Times) Ultimate
Soil Service 0.33 Soil Stress

4. Soil Factored 0.50 Soil Stress
5.
Soil Seismic & Plastic 1.00 Soil Stress
Pile Service 0.50 Pile Load
Pile Factored 0.75 Pile Load
Pile Seismic & Plastic 1.00 Pile Load
PILE SHAFT FOUNDATIONS

COLUMN/PILE SHAFT DESIGN PROCEDURE
Drilled pile shaft foundations are being used
at an increasing number of locations on Califor- 1. Determine an equivalent column length. The
nia bridge projects. The drilled shaft founda- equivalent length of a column/shaft member
tion is generally cheaper than other foundation is defined as that length of column plus
types and permits the location of columns in pile shaft which when fixed at the bottom
tight locations with a minimum of disturbance to will produce the same deflections at the top
existing facilities. The use of this foundation of the column for a given load as the actual
is generally limited to areas where soil condi- column plus shaft surrounded by soil.
tions permit economical excavation for the shaft
and where ground water is not encountered. The 2. Analyze the bridge with the equivalent
presence of ground water does not prohibit the column lengths and determine column loads to
use of the drilled shaft, however the cost be used in programs YIELD(7) and PILE(8).
becomes considerably higher in this case. The superstructure may be designed at this
time
It is more expensive to maintain the shaft walls
underwater (usually thru the use of bentonite 3. Design the TOP of column reinforcement for
slurry) and underwater concrete placement is all load groups I thru VII using the YIELD
required. Vertical load tests by CALTRANS on program(7) and the equivalent column
carefully installed slurry seal piles have shown lengths
that the resulting pile is equal in strength to
a normal dry constructed shaft (13). 4. Using the PILE program( 9 24 ) determine the
1
maximum moments in the pile shaft for the

Drilled shaft dimensions vary from 6 to 8 feet components of group loads I thru VII.
(1.8 to 2.4 m) in diameter and 30 to 60 feet (9 Detailed soil data from Engineering Geology
to 18 m) into the ground. is required for use of the PILE
program( 19,24)
Construction of the foundation consists of
drilling of the shaft excavation, installation 5. Using the YIELD program(7), determine the
of the reinforcement cage and placement of the amount of vertical reinforcement required to
concrete resist the maximum moment in the pile shaft.
The plastic moment capacity of the column
Column reinforcement is continuous but may be and shaft is also determined at this time.
spliced by welding or approved connectors,
however staggered lap splices are permitted for 6. Using the PILE program( 19 24 analyze the
, ,
#11 (36 mm) bars or smaller in columns over 34 pile shaft for the plastic condition. The
feet (10 m) high. Lap splices are not permitted plastic moment, the associated axial load,
near areas where plastic hinges are expected to and the assumed plastic shear are applied it
form. Fig. 3 shows the reinforcement details for the top of the column. The program Is then
a typical 7 foot (2.1 m) diameter drilled shaft run interactively by incrementing the she.ir
and column. until a plastic hinge forms in the pile
shaf t
The design problems involved with the use of the
pile shaft are slightly different from those of The shear reinforcement is then designed for
ordinary pile and spread footing foundations. the lesser of the shears resulting from
The pile shaft has a softer lateral stiffness seismic plastic hinging or group loads 1 t"
and therefore requires more refined foundation VI and unreduced elastic group load VII.
data at an earlier stage in the design process.
This softer lateral stiffness must also be 7. Perform a final check of the overall stabil-
considered in the design and analysis of the ity of the pile shaft using the PILE program
superstructure as well as the substructure (19,24). The pile shaft is considered
components stable when a substantial decrease in pi !«•
shaft length does not result in excessive state. The Statute has since been expanded
deflection. The amount of reserve shaft (January 1, 1977), to mandate the instrument
length is an indication of the factor of maintenance and record processing. The Division
safety against overturning. A stability has been collecting data and archiving records
ratio greater than 1.0 is mandatory. A good since September 1976.
rule of thumb is; the greater the uncertain-
ty of the soil, then the greater the stabil- Funding for the California Strong Motion Instru-
ity ratio. A stability ratio below 1.5 is mentation Program comes from a tax on building
not recommended without extremely reliable permits and currently amounts to about
soil data. $1,000,000 annually.
6. THE CALIFORNIA SEISMIC RETROFIT PROGRAM Current instrumentation consists of approximate-

ly 316 ground response stations, 71 building
The California Department of Transportation has, stations, 22 dam stations, 4 bridge stations and
since 1971 been involved in the retrofit of 1 tunnel station.
existing highway bridges.

A catalog of strong motion records recovered by
The California retrofit program began almost the Strong Motion Instrumentation Program thru
immediately after the 1971 San Fernando Earth- 1983 (4,5), is available. Four California
quake . bridges are currently instrumented under the
program:
The California Department of Transportation
identified about 1247 bridges (out of about Meloland Road Overcrossing (El Centro
13,000) which are primarily deficient in seat area)
width. These unrestrained joints represented
the prime focus of the Departments retrofit Route 154/101 Separation (Hollister
program. The average retrofit project consists area)
of the addition of steel restrainer cables at
expansion joints. To date 1070 of the 1247 Painter Street Overcrossing (Eureka
bridges have been retrofit at a cost of $40.5 area)
million. The $54 million project, which started
in 1971, is expected to be completed in 1985. Yincent Thomas Suspension Bridge (Los
Angeles area)
7. STRONG MOTION INSTRUMENTATION
The Painter Street Overcrossing, (Bridge Number
The California Department of Conservation, Divi- 4-236) is a two span cast-in-place prestressed
sion of Mines and Geology has been involved in box girder bridge on a two column monolithic
the development of data on the characteristics bent over Route 101 in the town of Rio Del (near
of earthquake generated strong motions for about Eureka) in northern California. The bridge is
twelve years. A legislative statute brought the located on a 39 degree skew. The bridge is
Division's Strong Motion Instrumentation Program instrumented to record 20 channels of data
into being on January 1, 1972. This statute collected on two CRA recording systems (18).
gave the Division the role of procuring and
installing strong motion instruments on repre- Strong shaking has been recorded at this bridge
sentative structures (including bridges) and in from three separate earthquakes:
various geologic environments throughout the
EARTHQUAKES RECORDED AT PAINTER STREET OVERCROSSING (18)
Mag Epicentral Maximum

Earthquake Date, Time (GMT) (ML) Distance Accel
Trinidad- 8 Nov 1980 ,

10:27 6.9 72 mi 0.34g
Offshore
Rio Dell 16 Dec 1982 ,

06:53 4.4 15 mi 0.60g
Cape Mendocino- 24 Aug 1983 ,

13:36 5.5 61 mi 0.27g
Offshore
Magnetic tape No. RIODEL8083 with 18 files of

data including uncorrected and corrected accel- Send to
eration data and response and Fourier Amplitude
spectra are available for reproduction and California Division of Mines and Geology
handling costs. A Preliminary report describing Office of Strong Motion Studies
the strong motion data recorded at the Rio Dell 2811 '0' Street
Overpass (18) is also available, (a final report Sacramento, CA 95816
of this data is in preparation). USA
8.
FUTURE NEEDS AND DEVELOPMENTS
2. Bell, J.M. and Hoffman, R.A. ,
"Design Earth-
Criteria quake Motions Based on Geologic Evidence.",
Proceedings of the ASCE Geotechnical Engi-
Some effort should be directed toward a neering Division Specialty Conference,
unified worldwide seismic criteria for bridg- Earthquake Engineering and Soil Dynamics,
es. Efforts in the United States, Japan and Pasadena, CA, June 19-21, 1978, Vol 1, pp
New Zealand are going in somewhat different 231-271.
directions and if the criteria were closer,
then resultant research would be more capable 3. "Bridge Design Specifications Manual", Stan-
of cooperative efforts. dard Specifications for Highway Bridges
Adopted by the American Association of State
Columns and Footings Highway and Transportation Officials,
Twelfth Edition, 1977, With Interim
There is an increasing need improved analysis Revisions Thru 1982, Washington, D.C., With
methods and criteria which considers the Revisions by the California Department of
effects of both vertical accelerations and the Transportation, Office of Structures Design,
variations of vertical loads on columns during Sacramento, CA, July, 1984.
large seismic events.
4. "Catalog of Strong Motion Accelerograph
Standard details with proven seismic resist- Records Recovered by Office of Strong Motion
ance, especially for columns about 3 to 6 Studies before January 1, 1982", California
foot, (0.9 to 1.8 m) diameter should be devel- Division of Mines and Geology, Special
oped with connection and footing details of Report 154, 1982.
proven toughness.
These members could be used in moderate seis- 5. "Catalog of Strong Motion Accelerograph
mic regions with little design effort and they Records Recovered by Office of Strong Motion
would provide excellent seismic resistance. Studies during 1982", California Division of
Mines and Geology, Special Report 154
The largest unknown in current design criteria Supplement A, 1983.
is the non-linear effects of both the soils
and the plastic hinging in columns. These 6. "Century Freeway Ground Response Study", by
effects should be more explicitly addressed in the Office of the Transportation Laboratory,
the criteria rather than built in implicitly. California Department of Transporta-
tion Sacramento CA, August, 1983.
, ,
Research and Testing

7. "Column Design by Yield Surface 'YIELD'",
Full-scale testing continues to be the only Instructions to Users by the California
way to definitely answer questions about the Department of Transportation Office of
ability of bridge components to bridge to Structures Design, Bridge Computer Manual,
withstand large excursions into the plastic Chapter 5-3, July, 1983.
9.
region.
1. "Footing Analysis/Design Program 'FOOT'",
8.
Cooperative research projects should be devel- Instructions to Users by the California
oped to test full sized, prototype structures Department of Transportation Office of
to destruction. The magnitude of such an Structures Design, Bridge Computer Manual,
effort is so large that cooperative projects Chapter 8-2, July, 1983.
can spread the costs among a number of spon-
sors . 9. Gates, James H., "California's Seismic
Design Criteria for Bridges", Journal of tin-
Better definition of the ground motions and structural Division, ASCE, Vol 102, No.
force levels close to faults continues to be ST12, December, 1976, pp. 2301-2313.
needed. Current motions are justifiably
conservative and until proven otherwise will 10. Gates, James H., "Factors Considered In the
be exaggerated in order to be safe. Development of the California Seismic Design
Criteria for Bridges", Proceedings of a
Procedures for defining and designing for the Workshop on Earthquake Resistance of Highway
out-of-phase displacements over the length of Bridges held on January 29~31 (
a bridge needs to be developed. The overall Applied Technology Council, Palo Alto, CA,
significance of these displacements on the 1979, pp 141-162.
total seismic forces in the system is unknown.
11. Gates, James H., "Seismic Resistant Bridge
REFERENCES Design in California", UJNR, Proceedings >!
the 15th Joint Meeting of the U.S. -Japan

Beeston, H.E., and McEvilly, T.E., "Shear Panel on Wind and Seismic Effects, Tsukuba,
-
Wave Velocities From Down-Hole Measure- Japan, May, 1983, Paper
ments", Earthquake Engineering and Struc-
tural Mechanics, Vol 5, 1977, pp. 181-190.
150
12. "Geotechnical Report for the Route 1/105 El Plots to Accompany Tape RIODEL8083" Cali-
,
Segundo-Norwalk Transportation Corridor", A fornia Division of Mines and Geology,

Report by the District 7 Materials Section, Sacramento, CA.
California Department of Transportation, Los
Angeles, CA April, 1974. 19. Reese, Lymon C. and Sullivan, W. Randall,
"Documentation of Computer Program COM624",
13. "Guide Specifications for Seismic Design of Geotechnical Engineering Center, Bureau of
Highway Bridges 1983", Highway Subcommittee Engineering Research, The University of
on Bridges and Structures 1982. American Texas, Austin, Texas, Aug., 1980.
Association of State Highway and Transporta-
tion Officials, Washington, D.C. 20. Romstad, K.M. ,
Bruce, J., and Hutchinson,
J.R., "Site Dependent Earthquake Motions",
14. Hays, Walter W., "Procedures for Estimating Journal of the Geotechnical Engineering
Ground Motions.", United States Department Division, ASCE, Vol 104, No. GT11, November,
of the Interior, Geological Survey, Profes- 1978, pp 1389-1400.
sional Paper 1114, Washington, DC, 1980.
21. Schnabel, P.B., Lysmer, J., and Seed, H.B.,
15. "Load Test - 48 inch Diameter Pile, "SHAKE: A Computer Program for Earthquake
Constructed by Slurry Displacement Method, Response Analysis of Horizontally Layered
Guadalupe River Viaduct, Bridge No. 37-308 Sites." Report EERC 72-12, Earthquake Engi-
R/L, 04-Scl-87-5 .8"
,
California Department neering Research Center, University of Cali-
of Transportation Division of Structures, by fornia, Berkeley, CA, 1972.
R.C. Wilhelms, Engineering Geology Branch,
May 5, 1975. 22. Seed, H.B., and Schnabel, P.B., "Acceler-
ations in Rock from Earthquakes in the West-
16. Mayes, Ronald L. and Sharpe, Roland L. ern United States." Report EERC 72-2,
"Seismic Design Guidelines for Highway Earthquake Engineering Research Center,
Bridges.", Applied Technology Council, Palo University of California, Berkeley, CA,
Alto, CA and The Federal Highway Adminis- 1972.
tration, Office of Research and Development,
Washington, DC, Report No. FHWA/RD-81/081 23. Seed, H.B., Idriss, I.M., and Kiefer, F.W.
October, 1981. "Characteristics of Rock Motions During
Earthquakes." Report EERC 68-5, Earthquake
17. "Memos to Designers", 15-10, Earthquake Engineering Research Center, University of
Design (Commentary), California Department California, Berkeley, CA, 1968.
of Transportation, Office of Structures
Design, Sacramento, CA, July, 1984. 24. "Shaft Pile Analysis", Instructions to Users
by the California Department of Transporta-
18. "Processed Strong-Motion Data from the Pain- tion Office of Structures Design, Bridge
ter Street Overpass —
Preliminary Report — Computer Manual, Chapter 8-1, July, 1983.
151
Table 1. — Active Faults and Associated Magnitudes (6,12)
Distance Maximum Maximum Credible

to Historic
Proj. Limits Magnitude Mag. Duration
Fault Mi (km) (Richter ( Richter) (sec)
Cucamonga
Malibu-Santa Monica
24-40
8-18
(39-64) — 6.6 18
(13-29) 5.2 7.5 30
Newport -Ingle wood
Palos Verdes
Raymond Hill
3-12
8-14
12-14
(5-19)
(13-23)
(19-23)
—
6.3
—
6.5
7.2
24+
24+
7.5 30
San Andreas 37-44 (60-71) 8.3 8.3 40+
San Fernando
Santa Susanna
23
26
(37)
(42) —
6.4
—
6.6
6.7
18
18+
Sierra Madre
Simi-Northridge
Whittier-Elsinore
18-24 (29-39)
22 (35)
8-14 (13-23)
—
—
6.6
7.5
18
30
7.6 30
Table 2 — Zones covered by ARS Spectra (6)
Curve Coverage
No* General Description Mi. (km)
1 Airport Viaduct 0.9 1.4

2 San Diego Frwy. (Rte 405) Interchange 3.9 6.3
3 Harbor Frwy. (Rte 11) Interchange 4.3 6.9
4 Alameda Viaduct 3.0 4.8
5 Long Beach Frwy. (Rte 7) Interchange 3.6 5.8
6 San Gabriel R. Frwy. (Rte 605) Interchange 2.5 4.0
Figure 1 Final ARS Curves
'
\
V
Curvt 1 T*0. aac KmO .69
1
Curva 2 T*0.3 aac A-O .69
—
1
i i
i
Curva 3 T-0.) aac A-O .69
i
Curva 4 T-0.) aac A-O .59
r\
71
/ Curva 5 T-0. aac A-O .49
1
v -
1
\ Curva 6 T-0. 35 aac A-O .49
j Li \\ l
1 / ‘
\
•*.... >
1
/
I / / |
/
t ,
7 / VO 5 % Da ® p i ny
li i
V
V"
^
J
ACCELERATION
it /
^
y t
ij
/
1
//.I
tu r j
MAXIMUM
0 2 4 * 810 IS 20 SO 40 SO
PE RIO 0 OF ST R U CT URE ( SE C )
152
Figure 2 Typical Single Column Bent on Footing
Figure 3 Typical Single Column Bent With Shaft Foundation
153
SEISMIC ANALYSIS, INSTRUMENTATION AND DYNAMIC
TESTING OF CABLE-SUPPORTED BRIDGES
BY
Ahmed M. Abdel -Ghaf far*
ABSTRACT
This overview paper summarizes two phases of ex- In general, there are significant deficiencies in
tensive research that was concerned with: (l) our present understanding of how reliably to de-
full-scale ambient vibration measurements of sus- sign, construct, analyze, and retrofit earthquake-
pension bridges, and (2) the earthquake response resistant structures. It is believed that these
analysis of long-span cable-supported bridges when deficiencies stem from a lack of reliable data on
subjected to mu 1 1 i
p 1 e-suppor t seismic excitations. the behavior of full-scale structures subjected
In the first phase extensive experimental investi- not only to forces from earthquakes but to ambient
gations were conducted on the Golden Gate Bridge forces from environmental factors such as wind,
to determine its dynamic characteristics. Compar- traffic, microtremors and functional activities.
ison with previously computed mode shapes and fre- There is no doubt that proper strong-motion in-
quencies showed good agreement with the experimen- strumentation of major structures in anticipation
tal results. Based on a comprehensive earthquake- of the occurrence of earthquakes, and the dynamic
response analysis of cable-supported bridges rele- full-scale testing of structures will provide the
vant observations and conclusions were made. In most valuable information for a rapid and cost-
addition, the paper sheds some light on the recom- effective improvement in the se smi c- res
i i s tance
mendations regarding permanent strong-motion in- of structures.
strumentation of these long-span bridges.

Earthquakes in the United States and other parts
KEY WORDS: Ambient Vibration, Cable-Stayed
Bridges, Damping, Dynamic Testing (Full-Scale), of the world have consistently demonstrated that
Earthquake (Response), Measurements, Mode Shapes, bridges are vulnerable to damage by earthquakes.
Multiple-Support Excitations, Natural Frequencies,
Suspension Bridges. And there is a strong need for continued evalua-
tion of the seismic design, performance, and
1. INTRODUCTION
Devastating earthquakes will unquestionably occur retrofitting of bridges as well as for the dissem-
in the future in the seismically active zones in

ination of research developments and pertinent
our planet. Uncertainty exists only as to the practical experience among researchers, design
scale of the damage and when and where such earth- engineers and government representatives con-
quakes will occur. The economic and human costs fronted with the earthquake hazard.
of major earthquakes are predominantly associated
with the behavior of structures and the lifelines Long-span bridges, such as suspension and cable-
that supply and connect them. In both urban and stayed bridges represent a very important compo-
suburban areas, the consequences of severed life- nent in 1 i e 1 ine-transportat ion structures. The
lines (e.g., bridges, water lines, etc.), in

relatively flexible and extended- in-plane c-
...,
addition to collapsing structures can be devasta- uration of such structures makes them susceptit lr
to a unique class of vibration problems. Earth-

ting.
quake damage to such long-span bridges interrupt-
‘-Associate Professor, Dept, of Civil Engineering, their primary function -- the safe and economical
Princeton University, Princeton, NJ 085^ movement of people and materials. The d-> tge is
154
even more serious when an earthquake causes a modal frequencies and modes of vibration of the
large number of casualties and hampers the gener- tower are determined. Finally, comparison with
al ability of people to respond to the emergency. previously computed mode shapes and frequencies
Thus, significant efforts should be made to ensure shows good agreement with the experimental results.
that these essential long-span structures will 1.1.2 Description of the Bridge
continue to function after an earthquake and be
The Golden Gate Bridge (Figs. 1, 2, 3) which lies
structurally serviceable after a strong earth-
across the entrance to San Francisco Bay and joins
quake .
the northern and southern peninsulas was completed

in 1937. The main span is 4200 ft, and side spans
This overview paper summarizes two phases of ex-
are 1125 ft long each and are suspended from the
tensive research that was concerned with: (l)
main cables. The width of the roadway is 90 ft,
full-scale ambient vibration measurements of the
and provides six traffic lanes and two sidewalks.
Golden Gate Suspension Bridge (San Francisco, CA),
The towers (Fig. 3) are made up of two shafts
and (2) the earthquake- response character st cs i i
that are connected by horizontal struts in the

of long-span cab e- supported bridges
1 (such as
panels that comprise the upper 500 ft of each
suspension and cable-stayed bridges) when subject-
tower. The towers are anchored to massive con-
ed to multiple-support seismic excitations. In
crete piers which are founded on rock. The two
addition, the paper sheds some light on the rec-
cables are 36.5 inches in diameter.
ommendations regarding permanent strong motion
instrumentation of these long-span bridges. 1.1.3 Description of the Instrumentation
1.1 Full-Scale Ambient Vibration Measurements Figure 4 summarizes the instrumentation used in
Performing tests on full-scale structures is the the tests, the measuring procedures, the data
only sure way of assessing the reliability of the processing, and the experimental set-up of the
various assumptions employed in formulating mathe- ambient vibration tests. More details of the
matical or finite-element models of structures. description of the instrumentation can be found
It is also a very reliable way of determining the in Ref. 1. The following is a brief summary of
parameters of major interest in structural dynamic, the instrumentation:
wind, and earthquake engineering problems. Fur- Motion-Sensing Transducers: bridge response
1.
thermore, amb ent-v brat on data can provide valu-

i i i
was measured at various locations using
able information pertaining to the optimal loca- Kinemetrics' Model FBA- 1 and FBA- 1 I Force-
tions for strong-motion instrumentation. Balance Accelerometers.
1.1.1 Ambient Vibration Measurements of the 2. Signal Conditioning: signals from the acceler-
Gate Bridge ometers were amplified and filtered using
Extensive experimental investigations [1] were three Kinemetrics' Model SC-1 four-channel
conducted on the Golden Gate Bridge to determine, Signal Conditioners.
using ambient vibration data, the mode shapes,
3. Recording: the amplified and filtered data
the associated frequencies, and the damping of the Honeywell
were recorded simultaneously on a
bridge vibration. The ambient vibration tests in- Model 5600E 14-channel FM tape recorder.
volved the simultaneous measurement of vertical, A Hewlett-Packard 3582A
4. Spectrum Analyzer:
lateral, and longitudinal vibrations of the sus- dual channel Spectrum Analyzer was used in the
pended span and the measurement of longitudinal
field to observe the frequency content of the
and lateral vibration of the main tower. Measure- data
ments were made at selected points on different
5. Electronic Analog-Digital Converter: the
cross sections of the stiffening structure and the
recorded analog signals were digitized using
tower. A total of 91 modal frequencies and mode the Kinemetrics DDS-1103 electronic analog-
shapes of the suspended span and a total of 46 digital converter
for the suspended spans and tower; certain simi-
1.1.4 Center and Side Span Measurements
larities and differences between the measurements
The vertical, longitudinal, and lateral motion of
and the computations are apparent, but in general
the center and side spans on half of the bridge
there is an excellent agreement between the meas-
were measured using 12 channels of accelerometers.
ured and computed mode shapes and their associated
During each testing session, six accelerometers
natural periods.
were mounted at one of the stations ( 1
- 1 8) indica-
1.2 Earthquake Response of Lo ng-Span Cable-
ted in Fig. 1. The positions and orientations of
Supported Bridges
these six accelerometers were located at the cross
section indicated by R (Station 8 in Fig. Significant improvement in the earthquake resis-
2),
where they remained throughout the tests on both tance of long-span bridges requires additional
spans. Summing the outputs of accelerometers information and understanding in three primary
C
and E (Fig. gives the purely vertical motion areas

2)
while subtracting their outputs gives the torsion- 1. improved knowledge of the forces and deforma-
al motion. Similarly, summing the outputs of actions generated by strong ground shaking,
celerometers A and B gives the purely lateral particularly those induced by multiple-support
motion while subtracting their outputs provides seismic excitations; the relatively flexible
information on the torsional motion. Figure and ext ended- n-p ane configuration of such
i 1
shows a type of Fourier amplitude spectra of the structures makes them susceptible to a unique
recorded motion. class of vibration problems.

2. empirical data on the behavior of structural
1.1.5 Tower Pier Measurements
components and on the interaction of structures
The longitudinal and lateral motions of the south with foundation-soil systems, when subjected
tower were measured using 9 channels of acceler- to propagating strong-ground motions, and
ometers. The reference station was chosen to be 3. development and verification of mathematical
at the roadway level (Station 5 of Fig. 3). Pure- models that can be used to calculate the res-
ly longitudinal vibration was obtained by summing ponse of these structures to seismic forces
the outputs of accelerometers A and C (Fig. 3) ranging from low-amplitude vibrations to dam-
while torsional motion was obtained by subtracting age initiation and ultimately to collapse.
their outputs Obviously, the best sources of data and informa-
tion on the above-mentioned areas, for improved
1.1.6 Natural Frequencies, Modes of Vibrations,
and Damping earthquake engineering of existing long-span
The procedure for determining mode shapes was to bridges, are heavily instrumented structures that
divide the spectral amplitude of the response at experience natural earthquakes. The need for
a given station by the spectral amplitude of the such information has been extensively noted in
simultaneously recorded response at the reference the technical recommendations and in the procfd-
station. In this way, an amplitude proportional ings of numerous meetings dealing with the prob-
to the mode shape amplitude at that station lems of earthquake hazards including those spon-
is
obtained for a given frequency of vibration. Re- sored by the National Academy of Sciences and t
peating this procedure for every station, the National Science Foundation [6,7,8]; the economic
mode shapes are determined. The phase of the return on capital invested in full-scale dyna ic
response was compared to that of the reference testing and strong-motion instrumentation i
.
instrument to determine the signs of the modal thought, by the research community, to be high.
displacements. The measured mode shapes and

The accuracy of calculated earthquake-r<- M>n.f
natural periods are compared to those obtained by
theoretical analyses [1,2]; Figs. 6 and 7, res- characteristics for long-span bridge structures
pectively, show some of these comparative results depends upon a sound knowledge of the expected
156
ground motions at different supporting sites. mode shapes are typically seen to contain,
Earthquake ground motions in the three orthogonal along the deck span, simultaneous contributions
directions of a long-span bridge may be transmit- to bending, torsion, and lateral sway. Thus
ted to the superstructure through the tower bases in that case, it is then no longer possible to
(piers) and the abutments or anchorages, as il- categorize a given eigenmode simply to pure
lustrated by Fig. 8. The bridge may be long with bending or pure torsion; it will simultaneous-
respect to the wave-lengths of motion in the fre- ly contain elements of all possible motions.
quency range of importance to its earthquake res- This distinct difference should have an impact
ponse. Because of this fact, different portions on the approach to the earthquake- response
of the bridge can be subjected to significantly ana lysis.
different excitations, a problem not normally 2. Multiple-support seismic response of the cable-
important for buildings. The essential feature supported bridges requires that the total dis-
of this problem is that the motion at the support- placements (or response quantities) may be
ing points a given distance apart on the surface decomposed into quasi- (or pseudo-) static
of the ground may receive long period earthquake displacements and relative (or vibrational)
motions that are nearly equal, but experience displacements. For such highly statically in-
short period motions that are dissimilar and un- determinate structural systems quasi-static
correlated. In general the correlation of the displacements are those resulting from the
motion at these points is extremely complicated, static application of support displacements.
particularly in the case of a long-span bridge, Figure 9 shows the quasi-static function asso-
1.
with different foundation conditions, subjected ciated with the vertical displacement of the
to seismic waves with different angles of inci- four supports of a suspension bridge, while
dence and different travel paths. In that case, Fig. 11 shows the vibrational and total dis-
marked differences in amplitude as well as phase placement response of a cable-stayed bridge
could occur over distance of the order of a few tower (the bridge was treated as a 3~D struc-
wave-lengths of the motion. ture and ground motion inputs were taken from
Based on a comprehensive earthquake- response anal- existing ground motion records recorded at
sis [2] of both suspension and cable-stayed time-synchronized close-spaced stations from
bridges, the following relevant observations and the 1979 El Centro earthquake).
conclusions merit mentioning: 3. It was found that a relatively large number of
Both theoretical dynamic analyses [6] and full- modes, closely-spaced in the frequency domain,
scale ambient vibration tests of suspension participate in the earthquake response of a
bridges have indicated that modes of vibration long-span bridge (see Figs. 10 and 12). Uni-
of the structure can be separated into two form ground motion for such a long-span struc-
groups. In one group, the displacements of ture is not a good assumption since it results
the stiffening structures and cables are pre- in nonconservative responses. The vibrational
dominant, and in the other group, the displace- stresses induced in the cable-suspended struc-
ments of the towers are predominant. Further- ture under multiple-support seismic excitation
more, the vibrations of the suspended structure are significant live loads and may come close
can be distinctly categorized into vertical, to or exceed design yield stresses. Further-
torsional, and lateral vibrations; that is, more, it was found that the P-A effects of the
one occurs mainly independently of the other. tower response have significant role in the
It was then possible in such cases, to view design values of the base bending-moment.
the truly 3~D modes of the bridge as being In the analysis of longitudinal tower-pier
"essentially bending" or "essentially torsion" vibration, the effect of the soil flexibility
without appreciable error. On the other hand, underlying and surrounding the pier upon the
for modern cable-stayed bridges the _ bridge
3 D
157
mode shapes and natural frequencies as well as the Midwest and Eastern U.S. (where there are
the response displacements, stresses, and seismically active zones that are subject to
shear forces is very important (Fig. 13). Thus frequent, sma 1
1
-to-moderate earthquakes). In
the estimation of soil properties underlying addition, more research is needed to acquire more
the foundation is essential in design. The understanding of the earthquake problem of long-
response stresses in the tower-pier system span cable-supported bridges.
under earthquake excitation are significant
but are still below their yield values. In conclusion, full-scale, sma 1
1
-amp 1 i tude dynam-
5. Finally, it should be emphasized that assur- ic testing is essential in working toward the
ance of the aerodynamic stability of a suspen- above-mentioned goals because the data generated
sion bridge does not in any way imply the -- consisting essentially of a definition of the
safety of these structures during earthquake natural vibration characteristics of the struc-
loading. Both the inputs and the responses, ture under study, and the interaction of these
as well as the possible modes of failure, are with the immediate environment -- would be useful
different for the two kinds of excitation. A (l) in confirming the reliability of analytical
mu 1 1 i
p 1 e-support analysis methodology is essen- and numerical models and, (2) in identifying the
tial in the earthquake resistant design of optimal locations for strong-motion instrumenta-
such a long-span structure. More details on tion to yield the utmost s t ructu ra 1
- response
this phase of research can be found in Ref. [2]. information. The data may also provide the basis
for research on fundamental problems related to
2. INSTRUMENTATION AND FUTURE RESEARCH NEEDS
earthquake engineering.
The ambient vibration studies of both the Vincent-
Thomas Bridge [3l and the Golden Gate Bridge [l] 3. ACKNOWLEDGMENTS
will be of great value in the development of a The authors are grateful to the people of the
set of recommendations regarding a comprehensive Golden Gate Bridge, Highway and Transportation
strong-motion instrumentation system for the District. This research is supported by Grants
national long-span cable-supported bridges, in (CEE8 1
- 1 3567 , CEE-8017736, and CEE-83^2559) from
general, and the Californian three suspension The National Science Foundation and the Depart-
bridges, in particular. These latter bridges are: ment of Transportation.
(l) San Franc sco-Oakl and Bay Bridge: which has
i
A. REFERENCES
the most urgent instrumentation priority; as a
1. Abdel -Ghaf far A. M. Scanlan, R. H. and Diehl,
, ,
prerequisite to proper permanent instrumentation
J., "Analysis of the Dynamic Characteristics
both comprehensive ambient vibration survey and of the Golden Gate Bridge by Ambient Vibration
Measurements," Civil Engineering Dept. Report
an extensive earthquake- response analysis of the
SM- 85 -I, Princeton University, Princeton, NJ
bridge should be conducted, (2) Golden Gate February 1 98 5 -
Bridge: which needs a back-up earthquake-engi-

2. Rubin, L. I., Abdel -Ghaf far ,
A. M. and Scanlan,
neering system to complement and support the cur- R. "Earthquake Response of Long-Span Sus-
H.,
pension Bridges," Civil Engineering Dept.
rently proposed space-technol ogy system, and (3)
Report SM- 83 -A, Princeton University, Prince-
the V ncent-Thomas Bridge:
i which needs upgrading ton, NJ June 983 ,
1 .
of its current strong-motion instrumentation sys-

3. Abde -Ghaf far A. M. and Housner. G. W.
,
"Am- ,
tem (of Fig. 1 k) bient Vibration Tests of Suspension Bridge,"

Journal of the Engineering Mechanics Division ,
Proceedings of the American Society ol Civil

The practical experience gained from these ambient Engineers, Vol. I0A, No. EM5, October 1978,
pp. 983-999.
vibration studies should be utilized to test more
important U.S. long-span bridges, such as suspen- 1*. Rojahn, C., Ragsdale, J., and Cooper, J.,
"Strong-Motion Instrumentation Schrmr for :it
ion and cable-stayed bridges, particularly in
158
V ncent-Thomas Suspension Bridge," Nov.
i 1 98 0
Open-file USGS, Engineering Branch.
CABLE
5. Abdel -Ghaf far Ahmed M. "Dynamic Analysis of

, ,
Suspension Bridge Structures," Earthquake

Engineering Research Laboratory, EERL 76-01,
California Institute of Technology, May 1976,
Pasadena, CA.
6. National Research Council, Earthquake Engineer-

ing Facilities and Instrumentation National ,
Academy Press, Washington, DC, 198*1, 33 PP-

7020
7. Committee on Earthquake Engineering Research

(1982), Earthquake Engineering Research- -
1982 National Academy Press, Washington, DC.
,
8. Earthquake Engineering Research Institute

(1984), Experimental Research Needs for Improv -
ing Earthquake-Resistant Design of Buildings :
Overview and Recommendations Report No. 8 *t- 02 ,
Earthquake Engineering Research Institute,

Berkeley, CA.
MEASUREMENT STATIONS OF THE GOLDEN GATE BRIDGE TOWER
Fig. 3 Measurement stations of the

Golden Gate tower
Fig. 1 Measurement stations

along the suspended
structure
the recorded motion at different point? of a typical cross section
Fig. 2 Positions and orientation of the six

accelerometers at a typical cross-
section of the suspended structure.
159
GOLDEN GATE SUSPENSION BRIDGE
AMBIENT VIBRATION STUDIES EXPERIMENTAL SET UP
and the experimental set-up of

the ambient vibration tests.
GOLDEN GRIE HMRIEN1 VIBRfitlON SURVE T
Fig.5 Fourier amplitude spectra of the recorded

acceleration oh Station A of the span and the
corresponding reference
160
GOLDEN GATE BRIDGE GOLDEN GATE BRIDGE
VERTICAL 5 TMM'E T R C MODE I 51 VERTICAL ANT -5TMMETR] C MODE A51
I
COMPUTED TEST COMPUTED TEST

T=8. 1450 SEC. T=8. 1900 SEC. T= 0. 55B0 SEC.
I T = 1 0 . 917 0 SEC
S Y”VE
T
VERT 3 CA_ RIC MODE S2 VER~ I C 0_ A\ T i-S YM M£*c;£
COMPU T ED TEST COMPUTED tjjt

T = 6 . 46 1 0 SEC. T=5 64 97 SEC T = 5 51 06 SEC.
.
. T= 5.]?03 SEC
GOLDEN gate bridge

GOLDEN gate bridge
VERTICAL STMMETRIC MODE S3
VERTICAL ant 5TMMETRJ c MODE A53
]
COMPUTED TEST COMPUTED TEST

T = 3 93 36 SEC. T = 3.8095 SEC. T=4 92 36 SEC. T= m.u.257
. .
SEC.
vertical ant i
- stmme tr j c mode a^4
CO“-U T ED TEST COMPUTED TEST

T = 3 4 380 SEC.
. T = 3 . 4 1 30 SEC. Tr2.9759 SEC. T=2. 7307 SEC.
Fig. 6 Comparison between computed and measured f rs t four natural i
periods and mode shapes of symmetric and antisymmetric vertical

vibration of the stiffening structure of the Golden Gate Bridge.
m
GOLDEN GATE BRIDGE TOWER
Lateral Vib ration
THIRD MODE
COMPUTED MEASURED COMPUTED MEASURED COMPUTED MEASURE
PERIOD IN SECONDS 23070 2.1845 0.7143 0.6206 0.3983 0 3744
BRIDGE TOWER Torsional V bration
FIRST MODE SECOND MODE THIRD POPE
COMPUTED MEASURED OCMPUTED PEASURED OCMPUTED HASURED
PERIOD IN SECONDS 1.1000 1.2136 0 4399 0 4950 0.2613 0 3290
BRIDGE TOWER L ngitudinal Vibration
SECOND PODE THIRD PPOE FOURTH MCI*.
MEASURED OCMPUTED MEASURED OCMPUTED hCASURED COPUTED PEASURED
1.3320 0.4974 0.5461 0.2782 0 2655 0 1971 0 1625
Fig. 7 Comparison between computed and measured natural periods and

mode shapes of vibration of the Golden Bridge tower.
162
GROUND MOTION IN THREE
ORTHOGONAL DIRECTIONS
(a) SUSPENSION BRIDGE
(b) CABLE-STAYED BRIDGE
Fig. 8 Typical three-dimensional cable-supported bridges subjected to

multiple seismic input motions at their supporting points.
163
?• *200
GOLDEN GRTE BRIDGE /
OURSI-STRTIC FUNCTIONS
VERTICAL GROUND MOTION
Fig. 9 Quasi-static functions of the vertical

displacements at the four supporting
points
Fig. 11 Vibrational (relative) and total displacement response of t wt f ,
cable-stayed bridge due to multiple-support seismi <v it .it i ri

GOLDEN GRTE BRIDGE VINCENT THOMAS BRIDGE
LEFT SPAN TORSIONAL STRESSES SYMMETRIC VERTICAL MODES
VIBRATIONAL CABLE TENSION
KSI’xSEC
CENTER SPAN TORSIONAL STRESSES

CORRELRTEO CR5E
ru
EL-CENTRO 1979 EARTHQUAKE
5 ARRAYS 6.7.
KSI'xSEC
Fig. 10 Autospectra of midspan (side and center

spans) torsional stress response due
to multiple- support seismic excitations.
SYMMETRIC VERTICAL MODES

FREQUENCY - CYCLES/SEC
Fig. 12 Autospectra of
vibrational cable-
tension response
due to different
earthquake inputs.
165
0 40 80 40 80 40 80
DISPLACEMENT RESPONSE (CM)
GOLDEN GATE TOwER

DISPLACEMENT RESPONSE AT X = 0.2H
EARTHQUAKE 61 VieRATJONAi
TOTAL
Fig. 13 Displacement response

of a suspension bridge
tower
166
SAN PEDRO-TERMINAL ISLAND SUSPENSION BRIDGE
r h—
u
—
PLAN AT TOP OF TOWERS
^
u
+ 25
1 o —
1
7 1 1 20
J
;i2 h*-9 J21 1
^ PLAN (TOP OF TRUSS) BELOW DECK

| 10 M6
26
hs
n D
1
2 LU. 3
r i plan ground level i* i i
28 9—*~ 29
F i
g . 1 4 Locat on of strong-motion instrumentation, on the
i F ree- F idi
^27
V ncent-Thomas Bridge, proposed by Rojahn, Ragsdale
i
site
and Cooper in 1980 (Ref. 4).
167
EARTHQUAKE RESISTANT DESIGN OF BUILDING FOUNDATION
- Introduction and Commentary on "DESIGN GUIDE FOR

THE BUILDING FOUNDATION AGAINST SEISMIC FORCE" -
by
Yoshihiro SUGIMURA*
ABSTRACT to 8, precautions for design of pile head
joints, design of underground walls, and cutting

Some important points and bases of the "Design methods for pile heads are described.
Guide for the Building Foundation against Seis-
mic Force" enacted in September 1984 are intro- KEY WORDS: Earthquake Resistant Design; Founda-
duced in this report. As historical tions; Design Guide; Earthquake Engineering;
background, two events gave impetus to estab- Pile Foundation; Spread Foundations.
lish the Guide. The first was the enactment of
the New Earthquake Resistance Standard for 1. INTRODUCTION
buildings in 1981, and the second was the
experience of damage to concrete piles due to Since the enactment of the New Earthquake Resis-
the Off-Miyagi prefecture earthquake in 1978. tance Standard for the Enforcement Ordinance of
the Uniform Building Code in Japan, revised in
The Guide is composed of eight chapters. In June 1981, the preparation of a building founda-
Chapter 1 , basic policies are described to pro- tion guide conforming to the Standard had been
vide structural safety of building foundations strongly requested. Under such circumstances,
that is equal to or greater than that of the the Ministry of Construction held several meet-
superstructure in allowable capacity design. ings with the Building Foundation Review Sub-
Chapter 2 prescribes the setting of design Committee (chairman: Professor Ohsaki,
external forces to be applied to building foun- University of Tokyo, at that time) of the Build-
dations. For pile foundations, the reduction ing Engineering Examination Committee, and the
of horizontal forces on piles due to effects of "Proposed Design Guide for the Building Founda-
the embedded portion is also described. tion against Seismic Force" was formulated in
March 1982. The contents of this Guide had hem
The design of spread foundations is introduced investigated by many organizations involved in
in Chapter 3, and design of pile foundations in the design of foundations and was revised and
Chapter 4, respectively. Since more problems enacted from September 1984 as the official
involve pile foundations than spread founda- design guide, i.e., "Design Guide for the Build-
tions, more pages are given to pile founda- ing Foundation against Seismic Force". Some
tions, i.e., bearing capacity, pull-out important points and bases of the concept • of
resistance, horizontal resistance and combined the Guide are introduced with some commentaries
stresses of piles. in this report.
Allowable stresses for various pile materials

•
are summarized in Chapter 5. In Chapters 6
* Head of Geotechnical Engineering Division, Structural Engineering Department, Hu i 1 < .

. .
•
Institute, Ministry of Construction
168
2. HISTORICAL BACKGROUND OF THE GUIDE buildings by considering dynamic effects were
started, based on the experience of the past
The historical background of the Guide is summa- earthquake damage, earthquake observation
rized in Table 1 . From the latter half of the results and so on. These research works were
1950's into the 1960's, the Japan Industrial mainly concerned with superstructures, but
Standard (JIS) provided production methods for there was awareness of the significance of
precast concrete (RC) piles, steel pipe (SP) establishing new systematic design methods for
piles and prestressed concrete (PC) piles, building foundations as well as superstructures.
successively . This fact shows the structural
properties of piles to be changing gradually After the earthquakes in the Izu peninsula area
from the preceding primitive condition to the accompanied by damage due to landslides and
higher strength condition in these two decades. faults in 1974 and in 1978, the Off-Miyagi pre-
fecture earthquake occurred in 1978. This
On the other hand, two important earthquakes earthquake made, as it were, an epoch to the
occurred in the 1960's, i.e., the Niigata aseismicity of building foundations, because
earthquake in 1964 and the Off-Tokachi earth- some examples were found in Sendai city of piles
quake in 1968. Although some examples of dam- supporting buildings being severely damaged
age to piles were found due to these two without any effects of change of ground condi-
earthquakes, the more conspicuous damage was tion, such as liquefaction, landslide, or
overturning or collapse of buildings due to fault. Furthermore, the important point is that
liquefaction in the Niigata earthquake or due the piles that were heavily damaged were pre-
to landslide in the Off-Tokachi earthquake. In stressed high strength concrete (PHC) piles
particular, liquefaction phenomena in sand which had been developed recently to be of
deposits in Niigata city gave a shock to higher strength than other types of concrete
researchers and engineers in soil mechanics and piles. This experience of damage gave impetus
foundation engineering, so that, since then to establishing an aseismic design method for
research works have been continued mainly for building foundations, and the activity of the
liquefaction phenomena. Building Foundation Review Sub-Committee was
started in 1979.
In the first half of the 1970's, two kinds of
regulation were revised and executed. The In 1981, the New Earthquake Resistance Standard
first is the Notice No. Ill of the Ministry of was enacted, and the Proposed Design Guide was
Construction in 1971, and the second is the exhibited in 1982 as the outcome of the activity
Building Foundation Design Standard of the of the Building Foundation Review Sub-Committee.
Architectural Institute of Japan (AIJ) in 1974. The Japan Industrial Standard also prescribed
These are mainly prescribing bearing capacity t'.e production method for PHC piles in 1982,
of spread foundations and pile foundations based on the experience of damage due to the
against permanent load. Of course, effects of Off-Miyagi prefecture earthquake. After that,
liquefaction and precautions are included in damage to piles of similar type was found in
these regulations, but aseismic design of subsequent earthquakes, i.e., the Off-Urakawa
foundations is not always described earthquake in 1982 and the Nihonkai Chubu earth-
systematically quake in 1983, which showed the necessity of
enforcement of the Guide widely as soon as pos-
From the middle of the 1970's, research works sible. Finally, the contents of the Guide were
preparing for a new aseismic design of revised on the basis of various opinions from
069
many organizations, and enacted from 1984 as the stress in the pile section to be calcu-
official design guide. lated, and also provisions are made with
emphasis upon the relation to ultimate
3. CONFIGURATION AND ENGINEERING POSITION capacity design and the relation to empir-
OF THE GUIDE ~ ~~ 5)
ical concerns such as workability.
The configuration of the Guide is as illustrated Precautions for design of pile head
in Fig. 1. The features of the contents of the joints, for execution of work for handling
Guide are as follows: pile heads, and for design of underground
walls are indicated.
1) As basic policies, the Guide is to be
applied to building foundations in allow- Though it was originally intended to arrange

able capacity design against seismic systematically the present findings on earth-
force, and its purpose is to provide quake resistance of foundations in the Guide, it
structural safety equal to or higher than has been not possible to organize a complete
that of the superstructure. The Guide system. The subject matter of the Guide
also emphasizes that possibility of col- includes items which are still in research work
lapse of the ground such as liquefaction, and thus cannot be applied, items which are
landslide or settlement, as those examples still not clarified and cannot be applied, and
shown in Photo 1 to Photo 6, should be items which are still unsolved but adopted in
separately investigated and countermea- the Guide because of the necessity in the admin-
sures taken in advance of the application istration and enforcement. Thus, the Guide is
of the Guide in order to maintain the in a position to function as a tentative index
safety of building sites. until a better design guide is prepared in the

future
2) Design external forces acting on build-

ing foundations are established, and the 4 . DESIGN EXTERNAL FORCES TO FOUN DATION
influence of combined external forces of DURING EARTHQUAKE
both horizontal and vertical forces is

emphasized. Also, the effects of decrease The design external forces on the foundation
in horizontal force due to embedment in during an earthquake are outlined in Fig. 2.
the case of pile foundations are clearly The horizontal force is equal to the value of
described the shearing force at the lowest floor (includ-
ing basement) prescribed in the New Earthquake
3) Since pile foundations involve more Resistance Standard plus the horizontal force .it
problems than spread foundations, most of the foundation (usually determined by the seis-
the description of the design calculation mic coefficient method). The vertical force is
method is for pile foundations. In addi- equal to the value of permanent axial force plus
tion to investigating bearing capacity and or minus the axial force due to overturning
pull-out force, the calculation formulae moment. The design load is, therefore, con- i <1
—
for elastic combined stresses in pile sec- ered as an inertia force of the superstructure.
tions are 'introduced In the case of a pile foundation, the horizontal
force may be reduced by the value of a (but
4) The allowable stress for the pile mate- shall not be more than 0.7) given by Eq . (11
rial has been determined for each kind of
170
because of the effects of embedment such as a in Fig. 5, and the results are shown in Fig. 6,
basement in form of the participation ratio for hori-

zontal force of each resisting element. It is
a 0.2
h CD
recognized that the participation ratios of
piles are similar for any types of ground. The
participation ratios of the embedded portion,
i.e., sum of passive resistance and frictional
where
resistance, are also similar for any types of
a : participation ratio of horizontal
ground, even though each resistance element
force of embedment
itself varies depending on the ground
H : height of the portion above the
conditions
ground surface (m)
: depth of embedment (m)
Similar analytical studies were made by Kishida
et al. (1979) and Kotohda et al. (1980). The
The relations of Eq. (1) are shown in Fig. 3
participation ratios of the embedded portion
for the case of larger than 2 m. The basis
studied by Kishida et al. (1979) are slightly
of these relations is as follows. If the equi-
larger than the results shown in Fig. 6, but the
librium of the horizontal forces on the embed-
results obtained by Kotohda et al. (1980) are
ded portion is assumed to be as shown in
very close to those shown in Fig. 6.
Fig. 4, then the following equations are
obtained
The participation ratio for piles has character-
istics as follows. If there is any embedment of
lb - % + Q + Q
f p
(2)
even one story basement, the participation ratio
suddenly decreases to the range of 0.8 to 0.6.
a (3) But this ratio does not decrease in proportion
to the number of stories of basement. If the
number of stories above the ground increases,

Each resisting element is indicated in Table 2.
the participation ratio also increases but not
The passive resistance at the front surface
in the proportion to the number of stories. And
(one side only) is assumed to be an elastic
then, if the increasing and decreasing tendency
spring. The frictional resistance at side sur-
of the participation ratio of pile is expressed
faces (both sides) is assumed to be of bilinear
in the form of functions, the participation
type that reaches the maximum value when the
ratio will approximately become proportional to
displacement is 1 cm. The unit frictional
the square root of the above-ground height H and
force is used as 80% of the normal value
inversely proportional to the one-fourth power
because both sides do not act effectively
of Df , the depth of embedment. Equation (1) was
depending on the direction of the seismic
derived based on the facts mentioned above.
force. Frictional resistance at the base sur-
Figure 6 can be redrawn as shown in Fig. 7 and
face is neglected here because of the possibil-
thus Eq. (1) indicates the average relation from
ity that an opening will occur between the base
the results of the trial calculations.
of foundation and the soil surface due to set-
tlement of the ground.

If the depth of embedment is small, the passive
reaction and friction resistance of the ground
Under the conditions described above, numerical
cannot be predicted sufficiently well, so that
analyses were performed for the examples shown
the range of less than 2 m is omitted in
171
Fig. 3. The Guide recommends the application distribution and shape of the contact pressure
of Eq. (1) only for the case of larger than and the rigidity of the foundation beam. Also,
or equal to 2 m. no special provisions are set forth for the
ratio of the area of the uplifted portion to
With respect to the seismic force on building the total bottom surface area, because uplift
foundations and the participation ratios of the rarely occurs or is normally considered to be
embedded portion and piles, as described above, very small under the conditions of design
there still remain many factors to be further external forces. Thus, uplift is not pro-
studied and investigated. For example, the hibited but is desired to be extremely small.
behavior of the foundation during an earthquake
is definitely affected by the effects of the 4.2 Design of Pile Foundation s
ground such as dynamic earth pressure during an
earthquake. Thus, it seems to be not adequate Items to be investigated for pile foundations
to evaluate the design force merely with are summarized in Fig. 9. For the vertical
respect to the inertia force of the superstruc- direction, it is required to examine the short-
ture, or the relation of Fig. 6 seems to be time allowable bearing capacity and the short-
insufficient since the behavior of the founda- time allowable pull-out resistance of the pile.
tion is also affected by the plan shape as well The former should conform to Notice No. Ill of
as the depth of embedment. It should be noted, the Ministry of Construction and the latter to
therefore, that the relation of Eq. (1) may be Article 27 of the Building Foundation Design
possibly improved or revised depending on the Standard of the Architectural Institute of
progress of future studies. Japan. That is to say, if design capacity is
not determined based on loading test results,
4 . 1 Design of Spread Foundations the following equations must be satisfied for
each case.
For spread foundations, bearing capacity of the
ground and slip-out of the foundation should be For bearing capacity per pile,
investigated (see Fig. 8). The bearing capa-
city should be examined as the short-time
o N q
allowable unit bearing capacity, since the R
a 43 (aNA p + (
0
L
S
+ ^
2
L
c
) tM (4)
duration of action of external forces during an

earthquake is very short. If there is possi-
where
bility of liquefaction, it should be sep-
R : short-time allowable bearing
a
arately investigated and the necessary
capacity (t/pile) (weight of pile
countermeasures should be taken as stated pre-
should be reduced in case of cast-
viously. For spread foundations, the influence
in-place concrete pile)
of inclined loads as combined external forces
a : coefficient of point resistance
must be taken into account and, in certain
(30 for driven pile, 20 for
cases, an increase of about 10 to 20% in con-
embedded pile, and 15 for cast-
tact pressure may occur.
in-place concrete pile)
N : average value of Standard Penetra-
With respect to uplift of foundation, the Guide
tion Test (SPT) blow count around
does not specifically prohibit occurrence of lower
the pile tip (range of IB in
any uplift in the calculations, but foundations
part and 4B in upper part from
must be designed by taking full precautions for
172
pile tip should be considered in for embedded piles and cast-in-
the case of driven piles; place concrete piles)

B = diameter of pile) N : SPT blow count (values of not
2
A : section area of pile tip (m ) more than 50 are applied for
_P
N : average value of SPT blow count driven piles and values not more
s
in sandy soil layer than 25 for embedded piles and
L : total thickness of sandy soil cast-in-place concrete piles)

s
layer (m) W : weight of pile (t)
P
q : average value of unconfined com-
pression strength in clayey soil For pull-out resistance per pile as a pile
layer (t/m ) group
L : total thickness of sandv soil

c
layer (m) = L
R - (AW + \ / ip S dz) ( 6)
t a
ip : perimeter of pile cross section
(m)
where
For pull-out resistance per pile as a single

R Short-time allowable pull-out
pile. t a
resistance (t/pile)
n number of piles
=
2 L
R -y •/ ijjrdz + W (5)
t a 3 o p A section area of polygon made by
connecting the outside of the
2
where pile group (m )
weight of pile and soil per unit

R : short-time allowable pull-out
K area at the bottom of pile group
t a
2
resistance (t/pile) (t/m )
L : length of pile (m) pile length (m)

ip : perimeter of pile cross sec- distance around the polygon made
tion (m) by connecting the outside of the
t : pull-out resistance per unit area pile group (m)
2 2
(t/m ) shear strength of soil (t/m )
It should be noted that the details of the

for clay : t = ^ method of determination of soil constants set
N forth in the Guide may be slightly different
for sand : t = (driven pile)
j
from those of the Standard of the Architectural
x =
N
— (embedded pile Institute of Japan.
and cast-in-
place con- For the horizontal direction, necessary parame-
crete pile) ters such as displacement and bending moment at
q : unconfined compression the pile head should be calculated by the theory

U
2
strength (t/m ) (values of of bending of a beam on an elastic foundation
not more than 20 are under the conditions of fixing ratio of the pile
applied for driven piles head as shown by the following equations:
and values not more than 10
173
: coefficient of subgrade reaction
P
y0 = R (7) 3
4EI6 3 Yo (kg/cm )
: fixing ratio at pile head (1 for

fixed condition, 0 for pinned
M condition)
o -Ih. (8)
With respect to the pile length L (cm) , the

M
max
.s.
26 Hi
(9) condition of BLS3 must be satisfied. If it is
not satisfied, the pile must be separately

investigated as a short pile.
Z R ( 10 )
m 6 z.
It is recommended that intermediate values of
where the fixing ratio be determined by experi-
mental results between 1 and 0, which correspond

to a completely fixed condition and completely
(ID pinned condition, respectively. However, if

experiments are not performed, a fixed pile head
condition is adopted as a rule to make calcula-
( 12 )
tions on the safe side, based on the examples of
damage due to the Off-Miyagi prefecture earth-
(13) quake in 1978 reported by Sugimura (1981), some
of which are shown in Photo 7 to Photo 12. Var-
ious types of damage are recognized in these
-1
= eXp tan photos, i.e., complete crushing, direct shear or
max
inclined shear cracks, tensile breaking of PC
(14)
bars, and buckling at the near part of the pile
+ 1
head. Photo 13 to Photo 16 show other types of

earthquake damage of the pile head joint, foun-
-1
R tan (15) dation slab, pile cap, and foundation beam.
z
m
Hair cracks are often found at the pile head as
shown in Photo 17 and Photo 18, which were not
where
caused by an earthquake but by the construction
displacement at pile head (cm)
work of cutting the pile head. All of these
M bending moment at pile head
o examples show the importance of details of the
(kg. cm)
pile head joint and its degree of fixing In
M maximum bending moment in the
max earthquake resistant design of pile foundations.
ground (kg. cm)
z depth of the position where
m However, as design philosophy of the Guide, It
M occurs (cm)
max is not intended to encourage the complete fixing
horizontal force at pile head
of the pile head. Instead, it is rather
(kg)
intended to recommend calculations on the basis
B diameter of pile (cm)
of fixed condition unless the design Is based n
El flexural rigidity of pile
special studies or experiments on pile head
2
(kg. cm )
joints. For the detail of Joints, it Is natural
that there will be a design method which tales

2
account of some degree of freedom of pile head, E :
elastic modulus of soil (kg/cm )
q
and there should be room left for judgement by B : diameter of pile (cm)
the designers. Therefore, further progress in
research and study of pile head joints includ- For determination of the elastic modulus of the
ing the concept of not connecting the pile head soil, there are three methods, i.e., direct
to the foundation slab is especially desirable. measurement by pressuremeter in the borehole,

by unconfined or triaxial compression test, or
No special allowable values for horizontal dis- for sandy soil deposits only, the empirical
placement are set forth in the Guide. By rea- relation E = 7N where N is the SPT blow count,
o
son that the allowable horizontal displacement
is primarily determined by the intended use of The above review of pile foundations deals
a building, it is desirable to determine the mainly with properties of the ground, and now it
displacement individually as required. Never- becomes necessary to investigate the stress con-
theless, if any rough idea is needed, about dition of the pile material itself. The outline
3 cm may be recommended as the value corre- of the investigation is indicated in Fig. 10 and
sponding to Eq. (7). Of course, a smaller Table 3. Basically, it is required to calculate

allowable value should be established for a the combined normal stress and shearing stress
building containing precision machines and thus based on axial force and bending moment by
requiring smaller displacement. applying the theory of elasticity. Some note-
worthy features are as follows:
The Guide indicates some other comments, i.e.,
that there is a possibility of projection of 1) While the concrete at the tension side
the pile head due to settlement of the ground, of bending is ignored for cast-in-place
that larger stress in the pile may occur at concrete piles and precast concrete piles,
much deeper locations, compared with the stress as in the case of columns of the super-
distribution derived from the calculation, and structure, the total section is considered
that a different calculating method may be to be effective for precast-prestressed
required for a section which varies consider- concrete piles, i.e., prestressed concrete
ably with depth. The Guide also shows the piles and prestressed high strength con-
results of comparison and review of many crete piles.
studies made on coefficients of horizontal
subgrade reaction. The most desirable method 2) Effective prestress is used as existing
is, of course, to determine the coefficient of stress in the equations for calculating
horizontal subgrade reaction by horizontal stress in the pile section.
loading test of piles. If such experiments
cannot be performed, the following equation is 3) For shearing force, the value of the
recommended to be used as well as the Road mean stress multiplied by shape factor is
Bridge Standard Regulation in Japan. basically used, by calculations based on
oblique tensile stress are applied for
-3/A
'
k = 0.8E B (16) precast-prestressed concrete piles. Fur-
h o
thermore it is recommended that the value
where of shearing force be increased by 50%,
i.e., = 1.500^ is used as the design

k, : coefficient of horizontal sub- shear force for all types of concrete
3
grade reaction (kg/cm ) piles, in order to avoid brittle failure.
175
4) About 1 mm of thickness as allowance F -0.6a
c e
°
against corrosion is provided outside for a = (A.I.J.) (17)
4>5
steel pipe piles, based on the results of
corrosion tests for steel material in F
O = (this Guide) (18)
various types of ground during the last a
ten years by Ohsaki (1981).

where
4.3 Allowable Stress of Pile Material
F^ : compressive strength of concrete
2
Allowable stresses for various kinds of pile (kg/cm )
materials are listed in Table 4. The range of o : long-time allowable compressive

ultimate compression strength of concrete, and stress in concrete (kg/cm")
2
the representative design values usually used a :
effective prestress (kg/cm )
g
in each case of concrete piles in Japan are as
follows The main reason for changing the function from

Eq. (17) to Eq. (18) is as follows: The effec-
1) In the case of cast-in-place concrete tive prestress is to be adopted in the equations
piles, concrete is usually designed with for calculating stress of the pile section as
ultimate compression strength in the range existing stress, as mentioned before, and is
2
of 180-270 kg/cm , and the design value considered to be independent of the setting of
2
most frequently used is 210 kg/cm or the allowable stress in the concrete. The
2
240 kg/cm . allowable bending-tensile stress is also sped-
fied because the whole section is considered to
2) As for the kinds of precast concrete, be effective against bending. Therefore, under
the Japan Industrial Standard (JIS) pre- conditions where the bending stress is mixed in,
scribes that the ultimate compression the stress of the pile section should be inves-
strength of concrete must not be less than tigated with the allowable compressive stress
2 2 2
400 kg/cm , 500 kg/cm and 800 kg/cm for for the compression side and with the allowable
precast concrete (RC) piles, prestressed bending-tensile stress for the tension side.
concrete (PC) piles and prestressed high
strength concrete (PHC) piles, respec- The process of determining the short-time allow-
tively. Thus, these values are usually able bending-tensile stress as a function of
selected as design strengths. But the effective prestress is as follows: The ultimate
real strength in most cases lies in the axial force-bending moment interaction curve of
2
ranges of 400-500 kg/cm (RC piles). the pile material can be determined by bending
2
500-600 kg/ cm (PC piles) and 800- failure theory of reinforced concrete, as shown
2
900 kg/cm (PHC piles) in Fig. 11. The relation of axial force-bending
moment interaction to the allowable stresses is
It should be noted that the long-time allowable determined by elastic stress superposition
compressive stress in concrete for precast- theory; as shown, for instance, by the triangle
prestressed concrete piles is different from shape in Fig. 11. Then, the value of the
that conventionally set forth in the customary short-time allowable bending-tensile stress is
formula in the Foundation Structure Design set so as to satisfy the following relation.
Standard of Architectural Institute of Japan,

as shown by the following equations
176
Allowable shearing stresses in concrete are
rra s1 - 75 (19 > shown in Fig. 13 for the case of the long-time
state. Some piles conform to the standards for
columns and beams of the superstructure, while
LA^. is the extension line connecting the state
other conform to the method for calculating
of long-time axial force (point L) to the state
oblique tensile stress. The upper limits have
of load condition under the design external
been specified in order to avoid excessive dif-
force (point A ); i.e. , a change of bending
e
ferences. But, as for shearing characteristics,
moment and axial force from point L to point A^
there still remain problems such as insufficient
occurs due to the design force in the case of
experimental data and extrapolation of results
the pull-out state. The extension is made
obtained for columns and beams with ordinary
under the assumption that the fluctuation in
strength concrete to piles with high strength
axial force and in bending moment will not
concrete
depend on the magnitude of the seismic force.
Therefore, Eq. (19) indicates an allowance
Actually, the amount of reinforcing against
ratio up to failure, and corresponds to the
shear force is extremely small and the differ-
securing of a ratio greater than 1.75. Since
ence between piles and columns as structural
the ratio LB./LB of the compression side,
f a
members is great. Photo 19 and photo 20 show
determined by the short-time allowable compres-
the test results of PHC piles recently performed
sion stress, is always larger than 1.75 for
in the Building Research Institute, the purpose
this kind of pile, the relation shown in
of which was to reproduce earthquake damage and
Eq. (19) need be specified only for the pull-
to investigate bending-shear properties of PHC
out side.
piles. Cyclic loads in both directions were
applied to the approximately central part of the
The allowable bending-tensile stress of con-
pile supported by a fixed end (footing) and
crete established by the method stated above
another pinned end. As the final state, shear
can be approximately expressed by the functions
failure of the pile occurred within such a range
of the amount of effective prestress as shown
that the stress condition of the pile did not
in Table 4.
reach the maximum bending strength. According
to this test result, the values of allowable
Figure 12 shows the axial force-bending moment
shearing stress presented here may not be always
interaction relation of the failure state and
required or sufficient. For shearing character-
the short-time allowable stress for each class
2 istics of piles, therefore, it is highly desir-
of prestress, i.e., 40 kg/cm for class A,
2 2 able to accumulate sufficient data similar to
80 kg/cm for class B and 100 kg/cm for
the test results mentioned above.
class C. If the axial force fluctuation ratio
to the long-time axial force exceeds 1 (tensile

The allowable stress for steel pipe piles may be
force occurs in pile), it becomes very diffi-
as same as that conventionally used, but the
cult to satisfy the conditions of Eq . (19),
thickness allowance against corrosion should be
particularly for the case of a small amount of
handled from a different viewpoint. According
prestress. To avoid this kind of situation, it
to the results of observations by Ohsaki (1981)
is recommended that the pile be used within the
for the last ten years, the average annual
shaded range illustrated in Fig. 12.
amount of corrosion is found to be about
0.02 mm. Thus, about 1 mm is appropriate as the
total thickness of corrosion for 50 years, from
277
been revised to 1 mm. the findings available to date, there still
remain many unknown factors and items to be
4 . 4 Design of Underground Walls examined in the future with respect to earth-
quake resistance of building foundations. The
If horizontal force 'is reduced on the basis of author hopes that the Guide will be a starting
effects of embedment in the case of a pile point for breaking the present situation where
foundation, the reduced portion of the force there has been no appropriate design guide, and
must be supported by underground walls. If will function as a seed which may develop fur-
this force is defined by aQ, then aQ is sup- ther studies in this field.
ported by the passive resistance component
of the front surface and frictional resistance 6. ACKNOWLEDGEMENT
component of side surfaces. As seen in
Fig. 6, participation ratios corresponding to The author wishes to express his gratitude to
these components will vary considerably depend- the members of the Building Foundation Review
ing on the properties of the ground, depth of Sub-Committee listed below, for discussions
embedment, plan of the building, horizontal about the Guide. The author also wishes to
displacement of the underground walls, and so thank the members listed below of the Working
on. In particular, it is necessary to assign a Group established under the Sub-Committee and
small participation ratio to the frictional managed by himself, for preparing and amending
resistance because it has more indeterminate the draft of the Guide.
factors than the passive resistance. There-

fore, the participation ratios of about 3/4 and Members of the Building Foundation Review Sub-
about 1/4 are recommended for Q and Q,., Committee (as of March 1982)
w f
respectively.
Chairman Dr. Yorihiko Ohsaki (University of
The force corresponding to the passive resis- Tokyo)
tance is to be added to the ordinary design Members Dr. Akira Enami (Nihon University)
external force for the underground wall (gen- Dr. Hideaki Kishida (Tokyo Institute
erally, earth pressure at rest) in the form of of Technology)
a uniform distribution with depth. According Dr. Osamu Sakaguchi (Building

to the results of measurements of earth pres- Research Institute)
sure during an earthquake at the basement of a Mr. Yoshihiro Maekawa (Save Energy
building in Yokohama City, reported by Ikuta Organization for Residence and

et al. (1979), the seismic increment of earth Building)
pressure has a constant distribution with Dr. Makoto Watabe (Building Pesenrch
depth, or a distribution pattern almost equal Institute)
to an inverse triangle shape, as shown in Expert Mr. Masaaki Endo (Takenaka Komuten
Fig. 14. Members Company)

the late Dr. Yasunori Koizumi (Tokyo
5. ADDITIONAL COMMENT Metropolitan University)
Prof. Kikuo Kotohda (Wanedn
Points and bases of "Design Guide for the University)

Building Foundation against Seismic Force" are Mr. Katsumi Yano (Nlkker Sekkei
summarized above. Although the Guide was Limited)
178
Members of the Working Group of the Sub- References
Committee (as of March 1982)
Ikuta, Y., Maruoka, M. , Minota, T. and Naga-
Members Dr. Hiroshi Oh-oka (Building Research nou, M. (1979). "Record of Lateral Pressure
Institute) Taken During Earthquake," Soils and Foundations,
Mr. Norimichi Ohki (Takenaka Techni- Vol 19, No. 4, pp . 85-92.
cal Research Laboratory)

Dr. Shin Okamoto (Building Research Kishida, H. ,
Yamato, S. and Nakai, S. (1979).
Institute) "Horizontal Resistance of Pile by Considering
Dr. Tetsuo Kubo (Building Research Embedment of Foundation" (in Japanese) , Kenchiku
Institute) Gijutsu, No. 329, pp. 147-163.
Mr. Fumio Chatani (Ohbayashigumi
Technical Research Laboratory) Kotohda, K. , Kazama, S. and Muroki, M. (1980).

Dr. Shoich Nakai (Shimizu Construc- "Research on Horizontal Resistance of Basement-
tion Company) pile System, Part 1" (in Japanese), Proceedings
Mr. Tsuneyoshi Hagiwara (Kajima of 15th Annual Meeting of Japanese Society of
Institute of Construction Technology) Soil Mechanics and Foundation Engineering,

Mr. Jyoji Yamashita (Taisei pp. 1009-1012.
Corporation)
Dr. Manabu Yoshimura (Building Ohsaki, Y. (1981). "Corrosion of Steel Piles
Research Institute) Driven in Soil Deposits," Research Report
81-04, Department of Architecture, University
of Tokyo.
Sugimura, Y. (1981). "Earthquake Damage and

Design Method of Piles," Proceedings of 10th
International Conference on Soil Mechanics and
Foundation Engineering, Stockholm, Vol 2,
pp. 865-868.
179
Table 1 Historical Background of the Guide
Earthquake and main damage Japan Industrial Standard Administrative activity
Precast concrete pile

(1955)
Niigata Liquefaction Steel pipe pile

(1964) Pile damage (1963)
Off-Tokachi Landslide Prestressed concrete pile

Notice No. Ill of t lie Ministry

of Construction (1971)
Off-Izu peninsula Landslide Building Foundation Design

(1974) Fault Standard of AIJ (1974)
Izu-Ohshima kinkai Landslide

(1978) Fault
Off-Miyagi pref. Pile damage Building Foundation Review

(1978) Landslide Sub-Committee (1979)
New Earthquake Resistance

Standard (1981)
Off-Urakawa Pile damage Prestressed high strength "Proposed Design Guide

(1982) concrete pile (1982) (1982)
Nihonkai chubu Liquefaction Design Guide

180
Embedment
at
Element
Resisting
of
Condition
Table
181
Table 3 Method of determination of Stresses in pile material
Axial force-bending Shear

Kind of pile
moment
Cast-in-place concrete In accordance with the 4Q < 1£

P =
pile review for columns of 3A s
superstructure (con-
Crete at tensile side 9p
Precast concrete pile 2 - £
is ignored.) A s
QP S° 2
- 2
<iV(o e + 2 Ofi ) Op
2tl
Precast-prestressed -f
concrete pile b gr- + Oe + y-Y^fc
L N
e
°8 " °e + a7
N M
-f 3£<f
Steel pipe pile b^ T + T y^ fc 2
z
A = s
N: Axial force (positive for compression and negative for tension);

M: Bending moment; Qp Shearing force;
: t: Thickness of wall of
pile; A: Cross sectional area; Ag Equivalent cross sectional
:
area of concrete; I: moment of inertia of pile;

I
e
:Equivalent geometrical moment of inertia of concrete;
S0 :moment of area of pile; Y: Radius of pile (positive at bend.'ng
compression side, negative at bending tension side);
og :Effective prestress; o^: Allowable oblique tensile stress;
f^: Allowable tensile stress or allowable bending tensile stress:
f
c
:Allowable compressive stress or allowable bending compressive
stress; f
s Allowable shearing stress
:
182
allownce
thickness
lie
t
excluding
corrosion
Materials
Thickness
against
Pile
of
Kinds
Various
for
Stresses
Allowable
Table
IS
Fig 1 Contents of the Guide
Q - iFj + Ff Qp= ( 1
- <*) Q/
M=I FjHj N = P/n
P =Pu ±Pe
o< Participation ratio
ofembedment
M
n : Numbers of pile
Ff
Fig. 2 Design External Forces during Earthquake
Df ( m)
Fig. 3 Values of Horizontal Force Reduction Ratio a
185
Acting external force
Passive resistance at
Qw
the front surface
: Frictional resistance
at the side surfaces
Q : Horizontal resistance
^ of pile
Fig. 4 Equilibrium of Force at the Embedded Portion
Numbers of story : 3, 6, 10
Basement stories : 1 to 4
Ground condition :
Sand : N = 10, 20
Clay : q = 0 .
6 , 1 kg/ cm
Pile : Steel pipe pile
<t>
= 609 .
t = 12.7
P = 116 t/pile
a
Fig. 5 Combination of the Exercises for Trial Calculation

(Building with a square plan)
186
0.8 Ground A:
N = 10
0.6
ratio
0.4
Participation
02 n— — _ 6F cTjo
- -_^^0C7.
-a -alOF
ojo
c
]ct
0
B1 B2 B3 B4
ratio
Participation
n. 3F
e
6F
-a®F fe;^L^=a 6F
. ~o 10F
B1 85 B3 B4~ B1 B2 B3 B4
Fig. 6 Horizontal Force Participation Ratios of Resisting

Elements Obtained from the Numerical Analysis
187
embedment
of
ratio
Participation
Fig. 7 Relation between Eq . (1) and the Results of the Trial Calculations
Conditions of SI ip-out
Design forces
ftttt tttttttt tmtmttttt
Allowable Bear- Influence of

ing Capacity Oblique Force
Fig. 8 Design of Spread Foundation
188
Fixing ratio of
t
Allowable Bear- Pull-out Horizontal
ing Capacity Resistance Resistance
Fig. 9 Design of Pile Foundation
Axial force- Shearing

bending moment
Fig. 10 Investigation on the Stress of Pile Material
JS5
M (tm)
1 00 r
00
00
c
•rH
0)
x:
co
Fig. 13 Comparison of Allowable Shearing Stresses

in various Kinds of Pile (Long-time)
190
Seismic earth pressure (t/m 2 )
0 02 Q4 06 0.8
/
/
/
/
/
/
/ Q.
/
/
/
/
/ /
o'
\
' /
/
/
X '
'
/ \
/ '
/ \
/
/ \
\
{
\
\
\
\
\
't 1
s
}
/
20 cr
Fig. 14 Results of Measurement of Seismic Earth

Pressure [after Ikuta et al. (1979)]
191
Photo 1 Overturning of Building due to
Liquefaction (Niigata earthquake 1964)
Photo 2 Overturning of Building due to Liquefaction

(Niigata earthquake 1964)
192
Photo 3 Landslide of the Hill at the Back of Houses
(Izu-Ohshima kinkai earthquake 1978)
Photo 4 Detail of Damage of Photo 3 (Izu-Ohshima

kinkai earthquake 1978)
133
kinkai earthquake 1978)
Photo 5 Collapse of Fill Ground (Izu-Ohshima
194
Photo Crush of PHC Pile (Off-Miyagi prefecture earthquake 1978)
Photo 8
li
Photo 9 Shear Failure of PHC Pile (Off-Miyagi prefecture earthquake 1978)
Photo 10 Shear Failure of PHC Pile (Off-Miyagi prefecture earthquake
196
Photo 12 Damage of PC Piles (Off-Miyagi prefecture earthquake 1978)
237
Photo 13 Damage of Joint between Steel Pipe Piles and
Foundation Slab (Off-Urakawa earthquake 1982)
Photo 14 Failure of Pile Cap (Off-Miyagi prefecture earthquake L978)
198
Photo 15 Cracks of Foundation Slab (Off-Tokachi earthquake 1968)
Photo 16 Cracks of Foundation Beam (Off-Tokachi earthquake 1968)
1SS
Photo 17 Pair Cracks due to Cutting Work of Pile Head (PHC Pile)
Photo 18 Hair Cracks due to Cutting Work of Pile He.'d ( PHC Pi 1
200
Photo 19 Bending-Shear test of Prestressed High Strength Concrete Pile
Photo 20 An Example of Shear Failure as the Final State
201
SHEAR RESISTANCE OF UNREINFORCED HOLLOW
CONCRETE BLOCK MASONRY WALLS
BY
Kyle Woodward and Frank Rankin

National Bureau of Standards
Gaithersburg, Maryland 20899
ABSTRACT
An experimental investigation is described which program of experimental research on the shear re-
has as its primary focus the determination of sistance and behavior of shear-dominated masonrv
the shear resistance exhibited by unreinforced, walls. The parameters under study are applied
ungrouted, hollow concrete block masonry walls. vertical (axial) compressive stress, wall aspect
Thirty-two wall panel tests are reported. The ratio ( length-to-height) ,
block and mortar type,
parameters in the investigation include the horizontal and vertical reinforcement, out-of-
amount of applied vertical compressive stress, plane loadings, and loading history. This paper
wall aspect-ratio, block strength, and mortar reviews the test program on ungrouted, unrein-
type. All of the wall panels have heights and forced hollow concrete block masonry walls. Sup-
thicknesses of 64 in. and 8 in. ,

respectively. plemental information on the research may be
Four different wall lengths are used to vary the found in references 1, 2, and 3.
aspect ratio: 48 in., 64 in., 80 in., and 96 in.
The vertical compressive stress varies from ap- The investigation reported herein involved tests
proximately 100 psi to 500 psi based on the wall on thirty-two wall panels whose nominal dimen-
net cross-sectional area. The block have gross sions are 8 in. thick, 64 in. high, and either 48
area unit strengths of either 1800 psi or 1300 in., 64 in., 80 in., or 96 in. long. The walls
psi, while the mortar is proportioned as either are subjected to a vertical compressive stress in
a Type S or a Type N. The wall panels are test- combination with in-plane lateral displacement.
ed in the NBS Tri-directional Test Facility which In addition to the variation of aspect ratio, the
imposes well defined boundary conditions at the vertical compressive stress, block strength, and
top and bottom of the walls. The resulting data mortar type are also varied in this series of
presented in this paper include the lateral load tests. The resulting data include in-plane lat-
(shear) versus wall displacement relationships, eral load resistance, wall displacement, and
crack pattern maps, and wall strains along their strains measured at discrete locations on the
diagonals. One of the principal findings of the wall faces.
investigation is that the relationship between
increasing amounts of applied vertical compres- This paper presents a brief description of the
sive stress and the resulting increased shear materials properties, wall panel details and fa-
resistance is significant and nearly linear. The brication, test setup, and instrumentation. The
other major observation is that there exists a data include load-dlspLacement relationships,
critical diagonal tension strain, independent of crack pattern maps, and wall diagonal strain
the varied parameters, which defines the onset of measurements
wall diagonal tension cracking. 2 . WALL PANEL DETAILS AND TE ST _8ETUP
1. INTRODUCTION 2 .1 Materials
The Center for Building Technology of the Nation- All materials used in constructing the wall pa
al Bureau of Standards is currently undertaking a nels and companion prisms are commercially avail
2Q2
able aad are representative of those commonly psi. The A term indicates the block strength with
used in building construction. The concrete ma- A being replaced by either H for the high strength
sonry units are two-core hollow block and have block or L for the low strength block. Similarly,
nominal dimensions of 8 in. x 8 in. x 16 in. the B term denotes the mortar strength (type) us-
The block labeled as high strength have a gross ing H for high strength (Type S) and L for low
area of 119.2 sq. in., a net solid area of 61.5 strength (Type N) . The terms high and low strength
sq. in., and a gross area compressive strength are used only in a relative sense and do not imply
of 1813 psi based on the average of six unit an absolute measure. That part of the identifier
tests. The block labeled as low strength have following the hyphen is a construction code and
a gross area of 119.8 sq. in., a net solid area provides for unique identification of each wall.
of 60.4 sq . in., and a gross area compressive
strength of 1304 psi based on the average of 2 .3 Test Setup
nine unit tests. The mortar is either Type S The test setup (Fig. 2) is the NBS Tri-directional
proportioned with 1 part by volume of portland Test Facility (NBS/TTF), a permanent loading appa-
cement, 3/8 part by volume of lime, and 4 parts ratus designed to test building components using
by volume of sand or Type N proportioned with 1 three-dimensional loading histories. The NBS/TTF
part by volume of portland cement, 1 part by is a computer-controlled loading apparatus which
volume of lime, and 5 parts by volume of sand. applies forces/displacements in all six degrees of
freedom at one end of a test specimen. The other
2 .2 Details and Fabrication end of the specimen is fixed. The six degrees of
A typical wall panel is shown in Fig. 1. The freedom are the translations and rotations in and
wall panels are constructed in running bond about three orthogonal axes. The application of
with a 50 percent overlap of block in alternate such actions is accomplished by seven closed-loop,
wall courses. The panels are fabricated and servo-controlled hydraulic actuators which receive
cured in a controlled environment laboratory their instructions by means of computer-generated
from materials stored in the same environment commands. A simplified sketch of the test method
for at least 30 days. The wall panels are for applying in-plane displacement to a wall panel
built using face-shell bedding except for the is shown in Fig. 3.
two end cross-webs. The joints are struck flush,

but not tooled. Mortar cubes and prisms are 2 .4 Instrumentation
also built and later tested as companion speci- In the interest of brevity, only that instrumenta-
mens to the wall panels. The mortar cubes are tion which provides data used in this paper is de-
2 in. x 2 in. x 2 in. and the prisms are made scribed. The loads imposed on the wall panels are
by stack bonding three stretcher units. measured by load transducers incorporated in the
hydraulic actuators. The readings from these de-
The details of each wall panel are listed in vices are adjusted to account for geometry changes
Table 1. The wall panel identifier is a two- and the combined effects of all of the hydraulic
part mnemonic with the two parts separated by a actuators. The wall panel displacement (in-plane)
hyphen. That part of the identifier preceding is measured by displacement transducers mounted
the hyphen is descriptive and has the form mABn. horizontally between the top course of the wall
The m term denotes the walL length in inches, and an external fixed reference (Fig. 4). The di-
while the n term specifies the approximate ver- agonal strain referenced in future discussion is
tical compressive stress applied to the wall. computed from the displacement measured bv a dis-
The stress is based on the net cross-sectional placement transducer mounted diagonally on the
area of the walL and is expressed in units of wall panel (Fig. 4). All of the instrumentation
205
is connected to a computer-based, analog-to- The 48 in., 64 in., and 80 in. walls tended to
digital converter having a sample rate of 50,000 form a consistent corner-to-corner diagonal crack
samples per second. pattern which essentially separated the wall into
two triangular segments. The shear transfer be-
3. TEST RESULTS tween the two segments took place along the dia-
3 . 1 General Behavior gonal crack by shear friction. The 46 in. Long
In general, the primary form of distress exhibit- wall, by contrast, did not form the corner-to-
ed by the wall panels was a diagonal-tension type corner crack. The crack pattern which formed
of failure as illustrated by the typical crack (Fig. 6) did not separate the wall into triangu-
patterns in Fig. 5. However, there were excep- lar segments but, instead, included a horizontal
tions. Flexural distress in the form of horizon- crack in the high flexural compression region of
tal flexural tension cracks in the mortar bed the wall. The combination of favorable crack
joints occurred if the applied vertical compres- orientation (horizontal) and high normal compres-
sive stress was insufficient to suppress this sive stress permitted a much more efficient shear
failure mode. It should be noted that even the friction mechanism to form than was possible in
walls exhibiting flexural distress, did, in fact, the shorter walls. In fact, 96 in. walls were
suffer a local diagonal-tension distress in a able to resist more lateral load through the
corner block and when the vertical compressive shear friction mechanism than from the diagonal-
stress was increased, the primary mode of dis- tension mechanism, thus first cracking was not
tress changed to a general diagonal-tension fail- coincident with the maximum lateral Load resis-
ure. In all of the walls tested, the final fail- tance .
ure of the wall was the result of crushing in the

high flexural compression zone of the wall. The 3.2 Shear Stress-Displacement Relationships
term final failure denotes an inability to sus- The shear stress versus wall displacement curves
tain the applied vertical compression load in for all of the tests are shown in various combi-
combination with the imposed lateral displace- nations in Figures 7 through 9. Shear stress Is
ment. computed by dividing the measured ln-plane lat-
eral load by the net cross-sectional area of the
While the cracking pattern was relatively insen- wall. There appears to be a common form to the
sitive to the parameters studied in the test pro- curves regardless of the variation in parameters.
gram, it appeared that the orientation of the The effect of the applied vertical compressive
cracking was influenced by both the level of ver- stress is much more pronounced on the maximum
tical compressive stress and the wall aspect ra- stress achieved than on the initial stiffness of
tio. The tendency of the diagonal cracking to the wall. This observation excludes the gross
follow the mortar joints, or to pass through the differences in the stress-displacement relation-
units, was affected by vertical compressive ship for the walls having flexural distress prior
stress, block strength, and mortar strength. to diagonal-tension distress. As the applied
Walls built with the low strength mortar exhibit- vertical compressive is increased, the shear
ed a pronounced tendency to exhibit mortar joint stress, or conversely the wall displacement, at
cracking while the walls built with the high which the curve exhibits nonlinearltv also In-
strength mortar exhibited much more cracking creases. Thus, higher applied vertical compres-
through the block. Increased vertical compressive stress has a negligible Impact on the abso-
sive stress increased the likelihood of block lute wall displacement at which the wall reaches
cracking, especially for the high strength block its maximum resistance. Therefore, the dlspl n <•-
and mortar combinations. ment which causes diagonal crack lm 7

is relatively
2M
independent of vertical stress, but the maximum stress and the vertical stress is affected by the
resistance and initial stiffness are increased wall aspect ratio and the combination of material
by vertical compressive stresses. strengths. Figure 10 illustrates the significant
The effect of the block and mortar strengths on effect of aspect ratio, but the effect becomes
the initial stiffness of the walls is interac- pronounced only for the higher levels of vertical
tive. The walls made with high strength block stress. Larger aspect ratios produce higher val-
tend to exhibit significant differences in ues of maximum shear resistance. The impace of
stiffness when the low strength mortar data material strengths on the maximum shear stress-
are compared with the high strength mortar data applied vertical stress relationship is also
(Fig. 7). Similarly, though not shown in the shown in Fig. 10. The most advantageous combina-
figures, differences are pronounced when the tion in terms of maximum resistance is, not sur-
stiffness of walls made with high strength mor- prisingly, high strength block and mortar. The
tar, but low and high strength block are compar- lease advantageous is the low strength block and
ed. However, comparison of the data from walls mortar, but as with aspect ratio, the effect of
with low strength block and low strength mortar material strength becomes significant only for
versus walls with low strength block and high the higher levels of applied vertical compressive
strength mortar (Fig. 8) show no discernible stress

differences in stiffness. This lack of apparent
effect is also true for the comparison of high The lines plotted in Fig. 11 are best-fit approx-
strength block-low strength mortar and low imations developed using regression analyses of
strength block-low strength mortar (not shown in the data. The particular data used in developing
figure). The overall behavior tends to suggest each line are identified by the line labels. Gen-
that the stiffness of the wall is strongly rela- erally, the approximations are linear with the
ted to its probable final crack path (mortar exceptions of the line 64LH which is quadratic.
joint or through block) and ultimately its weak- The label 64XX is the line including all of the
est material component. data for the 64 in. long walls. The shaded re-
gion is bounded by ordinates 2e above and below
The aspect ratio of the walls has a strong influ- the 64XX line. The symbol e denotes the standard
ence on the maximum shear resistance, but the error of estimate of the regression analysis. The
data do not indicate a similar influence on the curves exhibit a definite tendency to converge at
stiffness of the walls (Fig. 9). The actual ef- lower levels of vertical compressive stress and
fect may exist, but may not be discernible with- diverge at higher vertical compressive stresses.
in the scatter of the test data. The equation of the line 64XX is
v = 70.8 + 0.313a
3 .3 Maximum Shear Stress Resistance where v is the maximum shear stress and a is the
The maximum shear stresses computed from the applied vertical compressive stress both in units
maximum measured lateral loads along with the of psi. The value of e is 17.4 psi and the cor-
other pertinent data are listed in Table 2 for relation coefficient is 0.91. The vertical lines
all the wall panels. The maximum shear stresses on the plot in Fig. 11 denote the range of verti-
for the 48 in., 64 in., and 96 in. long wall cal stresses for which data is actually available.
panels are plotted versus the applied vertical The regression lines extending beyond these lim-
compressive stress in Fig. 10. Clearly, there its are extrapolations to show possible trends.
is an increase in maximum shear resistance with
increased vertical compressive stress. The 3 .4 Diagonal Strain

nature of the relationship between the shear The general mode of distress exhibited by the
ZB
test specimens is diagonal cracking, indicative o The maximum shear resistance was affected by
of a diagonal tension failure. It is reasonable aspect ratio and material strengths for higher
to expect that the diagonal tension strain levels of applied vertical compressive stress,
should provide a reasonable predictor of the on- generally being higher for larger aspect ratio
set of wall distress. The relationships between and increased material strengths.
shear stress and the diagonal wall strain comput-
ed from measured displacements are shown in Fig. REFERENCES
12 for all of the 64 in. long wall tests. With
certain exceptions, the curves in Fig. 12 sug- 1. Woodward, Kyle and Rankin, Frank, "Influence
gest the existance of a common diagonal tension of Vertical Compressive Stress on Shear Resis-
strain, in the range of 100-150 microstrain, at tance of Concrete Block Masonry Walls," Report
the onset of diagonal cracking. Diagonal crack- No. NBSIR 84-2929, National Bureau of Standards
ing is indicated in the relationships by the sud- Gaithersburg, MD, October 1984, 61 pp.
den large increase in diagonal tension strain
(displacement) with little or no increase in 2. Woodward, Kyle and Rankin, Frank, "Influence
load resistance. The exceptions are wall tests of Aspect Ratio on Shear Resistance of Concrete
having high vertical compressive stresses. The Block Masonry Walls," Report No. NBSIR 84-2993,
commonality of the threshold diagonal strain National Bureau of Standards, Gaithersburg, MD,
suggest a material criterion which defines the January 1985, 64 pp.
onset of diagonal cracking independent of
strength and vertical compressive stress. 3. Woodward, Kyle and Rankin, Frank, "Influence
of Block and Mortar Strength on Shear Resis-
4. CONCLUSIONS tance of Concrete Block Masonry Walls," Report

The following conclusions are based on the exper- in preparation. National Bureau of Standards,
imental test data obtained in the investigation Gaithersburg, MD.

described in this paper.
° A nearly linear relationship existed between

increased amounts of vertical compression
and the resulting maximum in-plane shear
resistance of the wall panels when other
parameters are held constant.
0
The lateral displacement coincident with the
attainment of diagonal cracking was not signi-
ficantly influenced by the amount of vertical
compression applied to the wall panel, or the
wall aspect ratio, or the material strenghts.
0 Tensile strain across the wall panel diagonal

was the critical determinant of diagonal crack-
ing and there appeared to exist a threshold
strain of about 150 microstrain above which
diagonal cracking occurred.
206
MOTE AB dmmione are nonenai dwnenaons Shaded area identities typrcal
horizontal mortar fO«n (t>eo <xrrt)
mortarbeddmg - face shell
beddmg except to» end crosa-weto*
Runreng Pond constructor
Mortar head
font (typ)
Monar P«d (omf

Wa« co»#ee one Wock and oea jomi
l
(typ)
Half Half
Streicher Stt etcher Stretcher Stretcher
hen kerf
40*. 64' and 90' tong wall panels
Whote kerf Stretcher Stretcher Stretdher Whole kerf are similar except for length
Figure 1. Typical wall panel.
Figure 2. Test setup (NBS/TIF)
207
&Q * ttw to*x>Md m-p*ana lateral d*ap*acamant
W» *»ticaJ (Jmanatana ra<*irad t°

Ai. A2 ^
achiava daatrad vertical load Ai naad not Uppaf
aquai A2 croaahaad
Uppar croaahaad
Lowar croaahaad
SIMPLIFIED DESCRIPTION OF IMPOSED DISPLACEMENT
Figure 3. Test method.
Direction of Displacement
r’H
Figure 4. Instrumentation.
206
Crack partem Mtl*40-«U» Otorth faoe) CftACft TOW
PAT •4MM1M-SL01 to)
Figure 5. Typical crack patterns.
CRACK PATTERNS: 96HH400-4L04 (North Face)
Figure 6. Crack pattern for 96 in. wall.
209
Figure 7. Effect of a block strength-mortar type conbination on
stress-displacement relationship.
Figure 8. Effect of a block strength-mortar type conbination on

stress-displacement relationship.
m
Figure 9. Effect of aspect-ratio on stress-displacement relationship.
Figure 10. Maximum shear stress versus applied vertical stress.
211
Figure 11. Regression analysis of data.
Figure 12. Stress versus diagonal wall stress.
212
Table 1. Wall Panel Details
Wall Panel Mortar Cube 28 Day Prism Prism Conpressive

Identifier Conpressive Strength Bedding Strength
(psi) (psi)
64HH120-2L04 2437 Face Shell 1839

64HH160-3L01 1825 Face Shell 1820
64HH240-3L04 2237 Face Shell 2132
64HH300-2L05 2160 Face Shell 1870
64HH320-3L03 2095 Face Shell 2091
64HH400-3L02 2139 Full Area 2813
64HH400-2L03 2232 Face Shell 2074
64HH500-2L06 2191 Face Shell 2005
48HH150-3L06 1847 Full Area 2661
48HH450-3L05 2055 Full Area 2645
80HH250-3L07 1994 Face Shell 1867
80HH400-4L01 3254 Face Shell 2050
96HH200-4L03 3076 Face Shell 1917
96HH300-4L02 2746 Full Area 2615
96HH400-4L04 2425 Full Area 2783
64HL160-5L01 1826 Face Shell 2049
64HL240-5L02 1809 —
64HL320-5L03 1761 — —
64HL400-5L04 1490 Face Shell 1923
64LL170-6L07 1987 Face Shell 1522
64LL250-6L06 1841 Full Area 1955
64LL340-6L08 1591 Full Area 1983
64LL420-6L09 1505 Face Shell 1443
64LH105-6L01 2646 Face Shell 1630
64LH170-6L02 2657 Full Area 2033
64LH250-6L03 2772 Face Shell 1522
64LH340-6L04 3127 Full Area 2133
64LH420-6L05 3110 — —
48LH170-6L10 2985 Face Shell 1447
48LH450-6L11 2892 Full Area 2094
96LH220-6L12 2700 Face Shell 1537
96LH320-6L13 2810 Full Area 2025
Note: Hie mortar cube stress is based on an area of 4 sq. in. Hie
values listed are the average of at least three cube tests.
Mortar cubes are removed from the molds after 24 hours and
air cured in the laboratory environment until tested.
Hie prism stress is based on the unit net solid area. Hie
values listed are the average of at least three prism tests.
213
Table 2. Wall Panel Maxinum Shear Stresses
Wall Panel Wall Net Cross- Applied Vertical Maxinum Shear

Identifier Sectional Area Conpressive Stress Stress Resistance
(sq. in.) (psi) (psi)
64HH120-2L04 246.0 122 113

64HH160-3L01 246.0 162 123
64HH240-3L04 246.0 243 167
64HH300-2L05 246.0 305 186
64HH320-3L03 246.0 325 206
64HH400-3L02 246.0 406 205
64HH400-2L03 246.0 406 227
64HH500-2L06 246.0 507 260
48HH150-3L06 184.5 163 117
48HH450-3L05 184.5 434 175
80HH250-3L07 307.5 228 178
80HH400-4L01 307.5 390 202
96HH200-4L03 369.0 217 157
96HH300-4L02 369.0 312 208
96HH400-4L04 369.0 407 251
64HL160-5L01 246.0 163 122
64HL240-5L02 246.0 243 151
64HL320-5L03 246.0 316 171
64HL400-5L04 246.0 407 190
64LL170-6L07 241.6 162 115
64LL250-6L06 241.6 246 144
64LL340-6L08 241.6 332 158
64LL420-6L09 241.6 413 174
64LH105-6L01 241.6 103 101
64LH170-6L02 241.6 161 128
64LH250-6L03 241.6 248 152
64LH340-6L04 241.6 327 166
64LH420-6L05 241.6 418 177
48LH170-6L10 181.2 165 116
48LH450-6L11 181.2 430 162
96LH220-6L12 362.4 220 163
96LH320-6L13 362.4 315 203
214
AN EXPERIMENTAL STUDY ON SLIDING RIGID BODY
IN WATER DURING EARTHQUAKE

BY
Tatsuo UWABE ,* Norihiro HIGAKI** and Setsuo NODA***
ABSTRACT
Shaking table tests of model blocks were conduct- necessary to develop the method to evaluate a
ed to study sliding behavior of a rigid body in permanent sliding displacement of gravity type
water during earthquakes. Accelerations and hy- structures
drodynamic pressures on the model block decreas-
ed at the onset of sliding and remained the same As use of the ocean space expands ,
a scale of
during sliding. Static and dynamic coefficients offshore structures becomes large with the in-
of friction obtained from vibration tests, with crease of the waiter depth at the construction
intertia/f orces from added mass corresponding to site. A large composite type breakwater which
hydrodynamic pressures on the model block, were consists of a rubble mound and concrete caissons
almost the same values as those measured by is planed for construction at mouth of Kamaishi
Euler's method. A method to estimate displace- bay in 60 m water depth. During earthquakes the
ments of the sliding block during earthquakes hydrodynamic pressures increase with the water
was presented. Calculated displacements based depth. It is necessary that the width of the
on horizontal and vertical base accelerations caisson be large to prevent it from sliding. In
showed relatively good agreement to the measured the design of the usual breakwater, the sliding
values stability of the caisson et al is not discussed

because the ground motions do not occurr in one
Keywords: Damage Deformation; Earthquakes direction but turn and because the duration of
the earthquake is not long. And in a foreign
1. INTRODUCTION design standard of the offshore structures the
1 . 1 General slide and overturn stability are not analyzed.
Many studies for earthquake disaster prevention Judging from the above description it is consid-
have been done in the Tokai area where a large ered that the gravity type structures may be de-
earthquake is predicted to occur in the near signed without the sliding stability analysis.
future. This investigation of seismic stability But in case of the important large structures it
of port and coastal facilities evaluates whether is necessary to know an allowable sliding dis-
each facility will be damaged by the Tokai earth- placement during earthquakes in the design be-
quake or not by means of tbe evaluation method cause there are many earthquakes in Japan. A
reported by Tsuchida et al.^D In order to take calculation method of sliding displacement dur-
effective countermeasures for earthquake disast- ing earthquakes is desired to be discussed in
er prevention, it is necessary to know how each full detail. It is also necessary to investigate
facility is damaged; that is, to know the damage the effect of the waters on sliding gravity type
deformation characteristics. One of the struc-
tures whose damage deformation values are desir- *Chief and **former member of Earthquake Disaster
ed to be evaluated is a gravity type structure. Prevention Laboratory, and ***Chief, Earthquake
A principal earthquake damage to a gravity type Resistant Structures Laboratory, Structures Divi-
structure is a sliding failure. Thus, it is sion, Port & Harbour Research Institute
215
structures in the waters. study, models identified by A, Bl ,
B2 ,
C, and D.
Models Bl and B2 are similar, but Table 1 shows
In his 1965 Rankine Lecture, Newmark 2) described the unit weight (rounded to one decimal), the di-
simple concepts for computing the displacement mension, the weight (the weight that includes
of a sliding mass in an embankment subjected to pickups is also shown) and the kind of test. Mod-
earthquake accelerations. Franklin and Chang 3) el A is a low model to clearly illustrate the
reported the calculation results of Newmark sliding behavior. Model D is also low and was
method for 169 strong motion earthquake records. used for preparatory tests to select a proper ac-
Richards and Elms 4) presented a new method for celerometer. Model C is heavier than the others.
designing gravity retaining walls considering al-
lowable sliding displacements. Sim and Berrill 3. STATIC COEFFICIENT OF FRICTION TESTS
5) studied shaking table tests of a model gravity 3 . 1 Slide Test Setup
retaining wall to verify the Richards and Elms The test setup, as shown in Fig. 1, is used to
design method. In addition to these there are a measure static coefficients of friction. 4 pull-
study of Makdisi and Seed 6) and a study of Nadim ing load is applied by using a container of lead
and Whitman 7). In Japan Fujino, Sasaki and shot as shown on the left side of Fig. 1. A
Hakuno 8) reported the slip of a friction con- strain-gauge type load ceil, whose rated capacity
trolled mass on a horizontal foundation excited is 50kgf (490N) ,
was used to measure the pulling
by earthquakes. Toki ,
Sato and Miura 9) report- load. An inductive-type displacement transducer,
ed a separation and a sliding between soils and whose amplitude limit is 25mm, was used to meas-
structures during strong ground motions using a ure displacement. The contact surface between
joint element. the model and the acrylic sheet is either dry or
wet. The water was put to 1 cm from the acrylic

In the studies mentioned above values of static sheet surface for the wet contact surface. A full
and dynamic coefficients of friction, the rela- infiltration of the water between the model and
tion between them were not discussed in detail, the acrylic sheet was confirmed before the test.
and the validity of the calculation methods for The tests of the wet surface were matched with
sliding displacement during earthquakes was not the shaking table tests of rigid body models In
checked,
2. except the study of Sim and Berrill. the water described after.
The sliding of gravity type structures in water

during earthquakes was not analyzed in past 3 .2 Test Results
studies. Therefore, this paper is concerned with Table 2 shows the test results. According to
these points. Accordingly, presented are the Table 2, the average static coefficients of fric-
sliding behavior of model blocks in the water, tion for the model Bl is 0.52 (a standard devia-
static and dynamic coefficients of friction, and tion is 0.02). As the static coefficient of
validation of the calculation method used for friction for the model Bl with the wet surface,
sliding displacement. Simple rigid body models at an average of 4 times of tests were 0.52, It
were examined as the first step of sliding analy- is considered that there Is no change of the sta-
sis for gravity type structures. Shaking table tic coefficients of friction In surface condi-
tests of rigid body models, made of mortar, on tions .

an acrylic sheet, or on an inclined mortar board,'
with or without water, were examined. 4 . EULER'S TEST FOR STATIC AND DYNA MIC CO-
EFFICIENT OF FRICTION
RIGID BODY MODEL DESCRIPTION 4 .1 Euler's Test
Four kinds of rigid body models were used in this The well-known Euler's test gives static and 1v
216
namic coefficients of friction from a sliding 4 ,
the static coefficients of friction on the
mass on a slope. The coefficients of friction mortar board are larger than those on the acrylic
for a mass on a slope, p s static and p d dy- sheet
namic are as follows 10).

p s = tan 9 s As shown in Tables 3 and 4, the dynamic coeffi-
2 • S cient of friction is smaller than the static one.
p d = tan 0s- ... A dynamic to static ratio is about 0.7 on the a-

2 ° „
& crylic sheet and about 0.8 on the mortar board.
g•t *cos
s The difference between dynamic and static coef-
where ficients of friction for the acrylic sheet are

0 s: Angle of inclination to start sliding larger than that for the mortar board.
S: Distance of sliding down (m)

ts : Time to slide distance of S (s) 5. SHAKING TABLE TESTS OF SLIDING RIGID BODY
g: Acceleration of gravity MODEL IN WATER
5 . 1 Model and Instrument Description
Fig. 2 shows the Euler's test setup. The inclin- Table 5 shows accelerometers used in this test.
ed board was pushed up by a oil jack to make a A waterproof inductive-type displacement trans-
rigid body model start sliding. Two kinds of ducer whose amplitude limit is +10mm was used to
tests, the rigid body model B2 on the acrylic measure the sliding displacement of the rigid
sheet (No. 1) and the rigid body model Bl on the body model. The accelerometers and the displace-
mortar board (No. 2) were examined. ment transducers were fixed by an adhesive. A
2
strain-gauge (rated capacity of 0.5 kgf/cm (49
The angle of inclination was obtained from mea- KPa) and pressure gauges were used to measure the
suring the sliding direction of acceleration of hydrodynamic pressure on the model. The pressure
gravity by an accelerometer on the inclined gauges were embedded in the rigid body model.
board. The time for the rigid body model to Figs. 3, 4 and 5 show models and pickups. Models
slide certain distance was measured as follows. used for these tests were A, B2 and D. Model D
An instance to start sliding was detected by the was used for preparatory tests to select a proper
displacement transducer on top of the inclined accelerometer. The models were put on the acry-
board shown in Fig. 2. The instant the model lic sheet or on the mortar board fixed on the
comes to the end of sliding its distance was re- shaking table. The water depth in tests of
corded by a displacement transducer at the lower models in the water was 25 cm.
part of the inclined board shown in Fig. 2. The
time to slide a distance was calculated from re- 5 .2 Shaking Table and Type of Excitation
cords of the two displacement transducers and a The shaking table of the Port and Harbour Re-
recorder. search Institute, Ministry of Transport was used.
A glass-sided rectangular box on the shaking
4 .2 Test Results table is 5ra long, 1.5m wide, and 1.5m high. As
Test results are shown in Tables 3 and 4. Table a duration time of an excitation was short, and
3 shows test results of Model B2 on the acrylic as it was considered that the distance between
sheet, and Table 4 shows those of the Model Bl the model and the wall in exciting directions is
on the mortar board. According to Table 3, the long enough, a wave absorber was not placed in
average static coefficients of friction obtained front of the wall in exciting directions.
from Euler's tests is smaller than the average
value measured in Section 2. As shown in Table Types of excitation are a periodic 10-sine-wave
217
and an earthquake wave. The 10-sine-wave means a erometers shown in Figs. 4 and 5 are all Al . As
periodic excitation of 10-wave with a certain the rate of error is probably high in small ac-
frequency and a uniform amplitude. The frequency elerations measured by this accelerometer whose
of 10-sine-wave excitation was selected from 5, rated capacity is + 1(1 G (9800 Gal.), attention
7, 10, 20 and 30 Hz. The amplitude of the 10- was given measuring small accelerations.
sine-wave was given to be strong enough to cause
model sliding. The wave used for the earthquake 6. RESULTS OF SHAKING TABLE TESTS
excitation is a strong-motion earthquake record 6 .1 Slide of Rigid Body Model During Vibration
(S-1210, E-W component) obtained on the rock at Fig. 7 shows test results of model A on the acry-
Ofunato port in 1978-Miyakiken-oki earthquake. lic sheet by 10-sine-wave excitation with a fre-
The duration time of the record was not modified, quency of 10Hz. Horizontal accelerations at each
and the amplitude was scaled to cause model slid- measuring point, displacement and vertical accel-
ing. erations are shown in order in Fig. 7. According
to the acceleration waveforms of model A an amp-
5 . 3 Selection of Accelerometer Used in Tests litude of model accelerations decreased instan-

An accelerometer used in tests was selected from taneously at the moment to cause a sliding dis-
small and water-proof instruments with a proper placement and remained the same level during
rated capacity and frequency characteristics. sliding. The reason why accelerations of the mo-
The important point in the selection was the del decrease is a change of coefficient of fric-
frequency characteristics of the accelerometer. tion from static to dynamic. A complete discus-
A flat response range of the frequency was un- sion is described below. A high frequency compo-
certain to measure the accurate acceleration nent of about 100Hz was observed in vertical ac-
waveform of sliding model. Then a preparatory celeration shown in Fig. 7. It was observed that
test to select a proper accelerometer was con- vertical accelerations show a tendency of an in-
ducted for the 5 kinds of accelerometers whose crease in the amplitude with an increase of the
frequency range is different as shown in Table 5. frequency of input sine waves. Though Lhe accel-
eration waveforms of models on set of the first
The accelerations of 5 kinds of accelerometers sliding are mentioned above, the acceleration
on the model D shown in Fig. 3 were compared by waveforms on set of the second sliding was not
a periodic excitation test. Fig. 6 shows a test the same as the first one.
result. Solid lines shown in Fig. 6 are the
waveforms of a accelerometer A3 whose upper Fig. 8 shows the test results of model B2 in the
limit of a frequency range is highest; dotted water on the acrylic sheet by lG-sine-wave exci-
lines are those of other accelerometers. Ac- tation with a frequency of 3Hz. In Fig. 8 hydro-
cording to Fig. 6 waveforms of accelerometer A2 dynamic pressures on the rigid body model are
and A5 are different from that of the accelero- also shown. These waveforms show the same tend-
meter A3, and waveforms of accelerometer A1 and ency in waveforms of horizontal accelerations.
A5 are similar to that of A3.
6.2 Hydrodynamic Pressure on Rig id B ody Mode l
As mentioned above there was no difference in Waveforms of hydrodynamic pressures on the r 1 I d
the acceleration waveform of the accelerometer body model are shown in Fig. R. Fig. 9 shows the
whose upper limit of frequency range is more maximum amplitude of hydrodynamic pressures I ;i
than 150Hz. Therefore, it was decided that the right half and maximum accelerations in left halt
accelerometer Al whose frequency range is from versus the water depth. The dotted line shows
0 to 150Hz was used for this study. The accel- the calculated values of Westergaard ' s formula
213
with a seismic coefficient that is equal to a ra- displacement and remained at the same levels dur-
tio of the maximum acceleration at the middle ing sliding. As described before the model accel-
height (AH4) to acceleration of gravity. This eration decreased instantaneously at the moment to
calculated value is similar to the observed cause a sliding displacement and remained at the
values. Though it is considered that the wall same levels during sliding. Therefore the above
in exciting directions have an effect on the ob- assumption is likely valid to calculate a sliding
served hydrodynamic pressures, this effect seems displacement. The calculation method of a sliding
to be small because of the long distance between displacement is discussed in the next section.
wall and the model that is nine times longer than
the water depth. Coefficients of friction based on the above as-
sumption were obtained from shaking table tests.
The amplitude of hydrodynamic pressures decreased The test results of model A which shows a sliding
instantaneously at the moment of sliding and re- behavior clearly were first discussed. As shown
mained at the same level during sliding as did in Fig. 7 the amplitude of model accelerations
the accelerations. Then hydrodynamic pressures reached a maximum at the moment to cause a sliding
on set of sliding were compared with those during displacement, and it was considered that the model
sliding. Fig. 10 shows ratios of hydrodynamic started sliding at this instant. Then the static
pressures during sliding to those at the instant coefficients of friction were calculated from the
of sliding. According to Fig. 10 the decreasing maximum horizontal accelerations and the vertical
rates are from 20 to 30% and are almost the same accelerations. The horizontal accelerations of
at each height. the model remained at almost the same levels after
sliding started. It is assumed that dynamic co-
6 .3 Coefficient of Friction Obtained from efficients of friction hold constant during slid-
Shaking Table Tests ing; the dynamic coefficient of friction were cal-
A concept of static and dynamic coefficients of culated from the average of horizontal accelera-
friction for a rigid body model during vibration tions during sliding. This dynamic coefficient of
is described at first before discussions of co- friction is called the "average dynamic coeffi-
efficients of friction obtained from shaking cients of friction" in this report. Fig. 12 shows
table tests. Fig. 11 shows a slide of a rigid static and average dynamic coefficient of friction
body subjected by a periodic loading without a of a model on the acrylic sheet or on the mortar
vertical motion. As shown in Fig. 11 the rigid board. According to Fig. 12 the ratio of dynamic
body starts sliding by an inertia force that is to static coefficient of friction is about 70% on
equal to a static friction force, and continues the acrylic sheet and about 90% on the mortar
sliding under a dynamic friction force that is board
smaller than static one until a relative velo-
city becomes zero. Though it was considered Secondly, test results of the model B2 with the
that coefficients of friction decrease continu- water depth of 25 cm and without the water de-
ously from static to dynamic values with in- scribed. In calculations of coefficients of fric-
crease of a relative velocity and approach a tion of a model the added mass of which inertia
certain value (10), it is accumed in this re- force corresponds to twice the hydrodynamic pres-
port that the coefficients of friction decrease sure on a side wall of the model was considered.
discontinuously from static to dynamic value and As the acceleration of the model top was about 20%
remain dynamic values during sLiding. As de- larger than that of the model bottom in some cases
scribed before the model accelerations decreased the average of accelerations at each level were
instantaneously at the moment to cause a sliding used to calculate coefficients of friction. The
219
method to calculate static and dynamic coeffi- sliding when the input acceleration A(t) is equal
cient of friction for model B2 is similar to that to p s*g. As the acceleration of the rigid body
of model A. In calculations of dynamic coeffi- becomes p d’g during sliding, the difference be-
cient of friction with the water the reduced hy- tween the input acceleration and the rigid body
drodynamic pressures during sliding were used. acceleration results. It was assumed that the
coefficients of friction decrease discontinuously

Fig. 13 shows static coefficients of friction for from static to dynamic value and remains a dynarni
model B2 with and without the water versus input value during sliding. As the input acceleration
maximum accelerations. Static coefficients of decreases and turns to another direction during
friction for model B2 with the water are a little sliding, the rigid body does not continue sliding
smaller than those without the water. It was to one direction but moves in one united founda-
guessed that one reason to cause this differ- tion when the relative velocity is equal to zero.
ence was the characteristics of the pressure This difference of accelerations between the foun
gauge and the accelerometer (an accelerometer dation and the rigid body during sliding gives
whose rated capacity is jtlOG (9800 Gal) was used a relative permanent displacement. The permanent
in case of small input accelerations from re- displacement is obtained from the integral of
striction of the frequency characteristics). the relative velocity that is the integral of the
Both coefficients of friction obtained from slide difference of accelerations between the founda-
tests in Section 3 and Euler's tests shown in tion and the rigid body during sliding. In this
Fig. 13 are likely the same as those without the report the sliding displacement of the rigid hodv
water. In view of the results described, it was were calculated by this method.
concluded that the sliding behavior of a rigid
body in the water was able to be discussed in In this case without the vertical motion shown in
consideration of the added mass that is twice the Fig. 11 the rigid body returns to the initial
hydrodynamic pressures on a side wall of a rigid place after one period of sine excitation. But
body divided by acceleration of gravity. the rigid body does not always return to the ini-
tial place with a certain vertical motion. 4s
Fig. 14 shows static and dynamic coefficient of the permanent residual displacement to one side
friction obtained from shaking table tests. In was observed in some cases of shaking table tests
Fig. 14 the results of Euler's tests are also the effect of a vertical acceleration should be
shown. The ratios of dynamic to static coeffi- considered. Therefore the vertical accelerations
cient of friction for Euler's tests were slightly were also used for the displacement computation
small compared with those of shaking table tests. of the sliding rigid body in this report.
7. DISPLACEMENT COMPUTING METHOD OF SLIDING 7 . 2 Comparison of Observed and Calculated Re s ulf

RIGID BODY DURING EARTHQUAKES The relative displacements of a sliding rigid
7 . 1 Concepts for Displacement Computing Method body were calculated under the following.
of Sliding Rigid Body
i) The input motions are horizontal and
In this report the displacement of a sliding rig- vertical accelerations (AH1, AVI),
id body during earthquakes was computed from the. li) Coefficients of friction obtained in
concepts shown in Fig. 11. In Fig. 11 A(t) is Chapter 6 were used,

the input acceleration, p s and p d are the iii) The first sliding of 10-slne-wnve
static and dynamic coefficient of friction and g excitation was discussed,

is acceleration of gravity. A rigid body starts iv) The effect of the water on a sliding
220
rigid body was evaluated by the added based on the above assumption were obtained from
mass that is twice the hydrodynamic the acceleration waveforms of a rigid body model.
pressures on a side wall of a rigid Both values are nearly the same as those of
body divided by acceleration of Euler's tests (see Fig. 14).
gravi ty (3) Coefficients of friction obtained in consid-
eration of the added mass that is twice the hydro-
Table 6 shows the computed results. The time dynamic pressure on a side wall of a rigid body
histories of computed and observed sliding dis- divided by accleration of gravity were nearly the
placement are shown in Fig. 15 for the model B2 same as those of a rigid body model without the
on the acrylic sheet and in Fig. 16 for the water (see Fig. 13). Therefore, it may be con-
model A on the acrylic sheet. cluded that the added mass that is twice the hy-
drodynamic pressure on a side wall of a rigid
Though there were two computed displacements body divided by acceleration of gravity can be
which were 3 and 0.5 times respectively than ob- used to calculate the sliding displacement of a
served ones, other computed displacements were rigid body during earthquakes.
almost the same as observed displacements. (4) The hydrodynamic pressure on the rigid body
Therefore it was concluded that the displacement model decreased instantaneously at the moment of
computing method presented here was valid for displacement and remained at the same level dur-
the the sliding rigid body that does not show a ing sliding (see Fig. 8). The waveform of the
large response. The displacement computing hydrodynamic pressure was similar to that of ac-
8.
method of a sliding rigid body that shows large celeration.
response will be discussed in the future. (5) The displacement computing method considerat-
ing paragraphs (1) to (4) revealed the same ob-
Coefficient of friction used in the calculations served results. Therefore it is concluded that
were estimated from shaking table tests. It is this displacement computing method of a sliding
troublesome that shaking table tests are always rigid body that does not show a large response is
conducted to obtain coefficient of friction. valid (see Table 6, Fig. 15 and Fig. 16).
Then Euler's test is useful to obtain coeffi-
cients of friction. 9. ACKNOWLEDGEMENT
The authors wish to thank Dr. Hajime Tsuchida,
CONCLUSION Director of Structures Division, Port and Harbour
Research Institute, Ministry of Transport for
Shaking table tests of rigid body models and com- many helpful discussions.
puted displacement of a sliding rigid body report
were discussed. Results, conclusions follow: 10. REFERENCES
(1) According to the shaking table tests, the 1) Tsuchida, H. ,
Inatomi, T., Noda, S., Uwabe
amplitude of acceleration decreases instantan- T., Yagyu, T. and Murata, T.: Method of
eously at the start of displacement and remains evaluation for Seismic Stability of Port and
at the same level during sliding (see Fig. 7). Coastal Facilities, Technical Note of the
Therefore it may be assumed that the coefficient Port and Harbour Research Institute, No.
of friction charges from the static to the dyna- 336, June 1980
mic value and remains a dynamic value during
sliding 2) Newmark, N.M.: Effects of Earthquakes on
Dams and Embankments, Geotechnique, Vol. 15,
(2) Static and dynamic coefficients of friction No. 2, Jan. 1965
221
3) Franklin, A.G. and Chang, F.K.: Permanent
Displacements of Earth Embankments by
Newmark Sliding Block Analysis, Miscella-
neous Paper S-71-17, U.S. Army Engineer Wat-
erways Experiment Station, Nov. 1977
4) Richard, R. ,
Jr. and Elms, D.G.: Seismic
Behavior of Gravity Retaining Walls,
Proc. of ASCE, Vol. 105, No. GT4 , pp. 449-
464, Apr. 1979
5) Sim, L.C. and Berrill, J.B.: Shaking Table
Tests on A Model Retaining Wall, Proc. S.
Pacific Reg. Conf. Earthquake Engineering,

Wellington, May 1979
6) Madkisi, F.I. and Seed, H.B.

Simplified Produre for Estimating Dam and
Embankment Earthquake-induced Deformations,
Proc. of ASCE, Vol. 104, No. GT7 , pp. 849-
867, July 1978
7) Nadim, F. and Whitman, R.V.: Seismically

Induced Movement of Retaining Walls, Proc.
of ASCE, Vol. 109, No. GT7 ,
July 1983
8) Fujino, Y., Sasaki, Y. and Hakuno, M.

10)
Slip of a Friction-controlled Mass Excited
by Earthquake Motions, Bull. Earthq. Res.
Inst., Vol. 53, 1978
9) Toki ,
K., Sato, T. and Miura, F.:
Separation and Sliding Between Soil and
Structure During Strong Ground Motion,
Proc. of JSCE ,
No. 302, 1980
Soda, N.: Friction Story, Iwanami Shinsho

1980
222
Table 1 Rigid body model
Unit HeightxWidth Weiahtfkaf Kind of test

Model we i g ht xLength No Inc! ude Tensil Euler’s SI iding
3
(gf/cm ) (cm) pickups pickups test test test
A 1.6 5 x 23 x 30 5.5 5.9 0
B1 1.6 28 x 23 x 30 30.7 31.9 0 0
B2 1.6 28 x 23 x 30 31 .3 32.5 0 0
C 2.1 28 x 23 x 30 40.4 41 .6 0
D 2.1 10 x 23 x 30 14.3 14.9 0
Table 2 Results of slide test
Static coef. of friction

Model Test Observed Average Standard
No Val ue Deviation
1 0.47
2 0.50
3 0.49
4 0.50
B1 5 0.53 0.52 0.03
6 0.53
7 0.54
8 0.55
g 0.54
1 0.62
2 0.63
3 0.63
4 0.59
C 5 0.64 0.63 0.02
6 0.65
7 0.64
8 0.62
9 0.61
10 0.67
Table 3 Results of Euler’s test ( No.l )
Model B2 on acrylic sheet. Sliding distance 150 cm

No Angl e of Time Velocity Coef. of friction
si idinc board elapsed (m/s) Static Dynamic
(Degree) (s)
1 29. 0 1 .57 0 .96 0 55 0 41

2 29. 9 1 .43 1 .05 0 58 0 40
3 29. 6 1 .62 0 93 0 57 0 44
4 30. 1 1 .31 1 .15 0 58 0 37
5 29. 6 1 .48 1 .01 0 57 0 41
6 27. 8 2 .20 0 .68 0 53 0 46
7 33. 1 1 .17 1 28 a 65 0 39
8 29. 8 1 .48 1 01 0 57 0 41
9 29. 9 1 .51 0 99 0 58 0 42
10 29. 9 1 .47 1 02 0 58 0 41
11 29. 8 1 .49 1 01 0 57 0 41
12 30. 9 1 .35 1 .11 0 60 0 40
13 31. 0 1 .44 1 04 0 60. 0 43
Averaae 0 58. 0 41.
Standard
0 03. 0 02.
deviation
2?3
Table 4 Results of Euler's test ( No. 2 )
Model B1 on mortar board. Sliding distance 146 cm

No. Angle of Time Velocity Coef. of friction
sliding board elapsed Static Dynamic
(Degree) (s) (m/s)
1 36.9 1.50 0.97 0 75 0.58

2 32.2 1 .71 0.85 0 63 0.51
3 32.8 2.01 0.73 0 65 0.56
4 35.6 1 .47 0.99 0 71 0.54
5 29.8 1 .84 0.79 0 57 0.47
6 32.9 1 .42 1.03 0 66 0.47
Average 0 66. 0.52
Standard
0 06. 0.04
deviation
Table 5 Accelerometer
Symbol Type Ampl itude Size (mm) Weignt Frequency

limit (gf) range(Hz
A1 Strain-gauge +10G 18x18x24 40 0-150 Water-proof
A2 Semi -conductor ± 5G 3. 7x3. 7x9 2 0- 80
element
A3 Servo + 2G 46x46x109 530 0-370
A4 Strain-gauge +10G 16x16x16 17 0-350
A5 Strain-gauge +10G 18x18x24 40 0- 60 Water-proof
Table 6 Calculated and observed relative displacement
Case No. Model Water Input Input Coef. of frictior Relative

depth wave max. acc for calculation disolacement(mm)
(cm) (Gal) 1
Observes
Static Dynamic Calculated
01 D2
Sine wave
1 9 A* 0 10 Hz 842 0.73 0.55 0.22 0.73
Sine wave
1 12 •» * 740 0.76 0.55 0.05 O'. 07 0.08
20 Hz
Earthq.
1 22 ¥ *
wave 682 0.59 0.52 0.26 0.26 0.29
Sine wave
5 8 * 690 0.68 0.53 0.40 0.41 0.26
7 Hz
Sine wave
5 4 "
10 Hz
726 0.69 0.57 0.18 0.27 0.30
Sine wave
3 34 B2 *
573 0.57 0.49 0.46 0.45 0.36
5 Hz
Sine wave
4 6 25 130 0.45 0.34 0.15 0.15 0.08
5 Hz
6 3 A** 0 Sine wave 631 0.65 0.61 0.04 0.03 0.02
10 Hz
6 12 * Sine wave 650 0.61 0.56 0.02 0.01 0.02
20 Hz
A*,B2 : on acrylic sheet. A**: on mortar board
224
MODEL
'
ACRYLIC SHEET
DISPLACEMENT
— | !
— TRANSDUCER
I OAD-CF I I. 7
Fig.l Slide test setup
DISPLACEME’-r
’ransducer
ACCE-EROMETER
PLAN VIEW ,,
MODEL
— D'S place men t
DISPLACEMENT ’ransducer
TRANSDUCER
\ v-.
v MODEL
—ACCE_EROME~E =
-acrylic
SHEE' OIl. JACK
SLIDING
DISPLACEMENT 80ARD
TRANSDUCEP-
Fig. 2 Euler’s test setup
-MODEL D
A i~A5 ACCELEROMETER
Fig. 3 Model D and pickups
225
-ACRYLIC SHEET
Fig. 4 Model A and pickups
AH ACCELEROMETER (HORIZONTAL)
AV » (VERTICAL)
W PRESSURE GAUGE
Fig. 5 Model B2 and pickups
A I /A 2 i CHZ
GC
a.
A 2 /A 2 i 0HZ
SCO -
n
c l
i
a\jr \nA \
1 cn o o l
A 4 /A 2 i 0HZ
SCO -
o
'
i A f i r
i <n o o
'
1 \J
-
A 5 /A 2 i 0HZ
6 0 0.
0
1 la r A
|
\
-500. - 1
1
— \J XJ 1
0 ! 0 2 0.3 0. i
7 I ME -SECONDS
Fig. Observed waveforms of accelerometers
226
:ase:-s *h: :oh/ accelera * ; cn -oa_
~
-
a A A A A, A />
/\
CASE
•v
-
- 9
.'V v
AH? 0H/
'v V \J '\J
AC CE L
\j
E RA *
I K -Z A
! i
'
I
“E -SECCNCS
; A »£ - r a. . CHZ ACCELERA" I Z - CA _
; *.: :jhi *cct.£B»-::K--A.
'y ‘wH 11
Vh’ i '
j * *>^‘*
v^“
'
•
’-
/V-
:ase.-? a.:- .:hz acce.epa : r.- a.
Ij
:a.EE. 5 A.A :3H7 ACCELERAT ION-CA.
;r.E -SECONDS
Fig. 7 Results of sliding test ( Model A )
MODEL B2 ON ACRYLIC SHEET

WATER DEPTH. 25 cm * OBSERVED VALUE
SINE EXCITATION 5Hz WESTERGAARD
(K. "0 12)
200 100 0 10 20 30
MAXIMUM ACCELERATION HYDRODYNAMIC PRESSURE
(Gol) 2
( gf /cm x 98 Po)
on model B
Fig. 9 Hydrodynamic pressures
227
ACCELERAT ION-CA
^ A A A A A A
^ v v2 V7 \y'
vy v v
CASE * -5 AH? SHZ ACCELERATION-GAL
nu^ Cj cj
n
—
r~, ^
C- ^
CASE 4 -G AH 3 SHZ ACCELERAT ION-CAL

/\
CASE
^
4 -S
/-^
AH 4
wuv SHZ
Xt XX
/-* r*
CCELERAT ION-CAL
1
\ f
j
-£2.
\^-j
CASE 4 -S AH S SHZ ACCELERAT ION-CAL

\ /*—
:ASE4-S AHS SHZ ACCELERAT ION-CAL

‘VI 1^—
*
/»»** r-*' w* /w* r*-*" ,
T I n£ -SECONDS
ACCELERA* ICN-CA
ASE4-S 01 SHZ OISPLACEHENT -nn
CASE 4 -5 D2 SHZ 0 I SPL ACE HE NT -H
CASE 4 -S Rl SHZ RATER pres C/Cnu?
"\7— — *=:
CASE 4 -6 V? SHZ rater pres, c/cnu;

H r\ /x r\ /s a
U U U ^
TinE -SECONOS
CASE 4 -S AVI SHZ ACCELERAT ION-CAL
CASE 4 -S A V? SHZ ACCELERAT ION-CAL
CASE 4 -S AV3 SHZ ACCELERAT ION-CAL
1
-tJ
case 4 -6 a V4 shz ACCELERAT ION-CAL
,o -]
»•
.,..1
+- 4»h«
4 M a
CASE 4 -S AVS SHZ ACCELERAT ION-CAL
—
< • * i i»
p «i
CASE 4 -S AVS SHE ACCEL I SAT I OH - CAL
'
P* 4 ^ * f*
r n,f
MnE-SECOMOS
Fig. 8 Results of sliding test ( Model B2 >
228
MODEL B2
JSINE EXCITATION 5 Hz
ON ACRYLIC SHEET £
” 10Hz
WATER DEPTH = 25cm'
W I
-
10
W2 —
W3 20 bo
0 0.5 10
HYDRODYNAMIC PRESSURE RATIO
Fig. 10 Ratio of hydrodynamic pressure
after sliding to that before sliding
A (t
D (t )
Fig. 11 Sliding block under horizontal

sine excitation
229
MODEL A
Fig. 12 Static and dynamic coef. of friction
( Model A )
TENSILE TESKMODEL 81)

MODEL B2 * Euler’s TEST
10 v * SINE EXCITATION 5 Hz
ON ACRYLIC SHEET 0 , . |QHz
L!_wiTH WATER
— WITHOUT WATER
ij- 0
Id
O
O
0 500 1000
INPUT MAXIMUM ACCELERATION (Gal)
Fig. 13 Static coef. of friction

( Model B2 with and without water )
230
I 0
MODEL B 2 ON ACRYLIC SHEET
SINE EXCITATION 5 Hz y
u
cc
U_ r WATER DEPTH 0cm
" 25cm,/
/
o
u_ Euler’s TEST
/
/
o /
o I
/
/
2 0.5 i
/ 7
/
>
Q
cr /
/
/:
LxJ /
/
! /
1 /
0 0.5 1.0
STATIC COEF. OF FRICTION
Fig. 14 Static and dynamic coef. of friction
( Model B2 )
<
a CAS -3AAVI SHZ
00 .
-•
-7CC. J
cj
<
CALCULATED
CASE3-3AD1 SHZ
Fig. 15 Calculated and observed relative

displacement ( Model B2 )
251
c
o CASE -22AV
i
CO. -I
o
c 0.
cr
-700.
ca
CJ
<
CALCULATED
CASE ;
-2 2D 1
0. S
u S o.
<
0-
CO
-c. s
C. i. 2. 0. A. S. G 7.3 9. .0
TIME-SECONDS
Fig. 16 Calculated and observed relative
displacement ( Model A )
232
A RAPID SEISMIC ANALYSIS PROCEDURE FOR
ANCHORED SHEET PILE BULKHEADS
BY
K . Lew*
involve the risk of many human lives, such
Typical failures modes of steel sheet pile damage can definitely interrupt the ability or
bulkheads in past major earthquakes are present- reduce the effectiveness of a Naval activity on
ed to provide insight into the failure mecha- performing its intended function. In past major
nisms involved. Because they have a significant earthquakes, damage to port and harbor facili-
influence on the response of the bulkhead, the ties includes [1, 2]:
lateral active and passive earth pressures and • Settlement of filled land
residual water pressure acting on the bulkhead • Settlement or collapse of breakwaters
under earthquake conditions are given. Soil • Sliding, tilting, or collapse of
considered include cohesionless soils, soils quaywalls
with cohesion and angle of internal friction • Tilting, bulging at the face of the
such as silty sand, and cohesive soil such as revetment, sliding, and collapse of
clay and silt. A rapid seismic analysis proce- sheet pile bulkheads
dure is presented for anchored steel sheet pile • Buckling of pier piles
bulkheads. In the procedure, the adequacy of • Collapse of pier sheds
the depth of penetration of the sheet piles, the Many of the older Navy port and harbor struc-
sheet pile section in bending, tie rod section tures located in relatively high seismicity
and connections, tie rod anchorage are evalu- regions were designed and constructed with
ated. Concepts for strengthening inadequate little or no seismic provisions as was common
bulkheads are given. In addition, guidelines with the prevalent engineering practice during
are given for the seismic analysis of steel their design. Consequently, many of these older
sheet pile bulkheads with relieving platform. waterfront facilities may be susceptible to
Keywords: Seismic analysis, Bulkheads, Earth damage from major earthquake motions at the
pressure, Failure modes. Tie rod site. There is a need for relatively rapid
anchorage, Cohesive soil, Cohesion- analysis procedures to evaluate the Navy's
less soil existing waterfront structures. Those structures
1. INTRODUCTION found to be inadequate are then analyzed in
The Navy's port and harbor facilities play an detail to determine the degree of strengthening
important role in the U.S. military system. required
Cargoes of military material and personnel are
As a part of an ongoing investigation of the
loaded, unloaded, shipped, and stored at such
earthquake safety of Naval facilities, the Naval
facilities. In addition, these facilities
Civil Engineering Laboratory (NCEL) has develop-
provide berthing for the repair and maintenance
ed a rapid seismic analysis procedure for gravi-
of the Navy's ships that support this function.
ty quaywalls [ 3]
The interruption of this vital function could
reduce the effectiveness of the U.S. military in This investigation was authorized by the Naval
responding to the nation's interest around the Facilities Engineering Command (NAVFAC) under
world Project Order N0002584WR1052W of 19 Oct 1983.
1 . 1 Objective
Many of the Navy's port and harbor facilities The primary objective of this investigation is
are located in regions with relatively high to develop a relatively rapid seismic analysis
seismicity. Experience from past major earth- procedure for anchored steel sheet pile bulk-
quakes indicates that these facilities are heads. The secondary objective is to provide
susceptible to severe damage from earthquake concepts for strengthening those bulkheads found
ground motions at the site. Although damage to to be seismically inadequate.
port and harbor facilities generally does not "Naval Civil Engineering Laboratory, Port llueneme, CA 93043
253
1 . 2 Approach 2. TYPICAL FAILURE MODES IN PAST EARTHQAKES
Typical failure modes of steel sheet pile bulk- In this section, the typical failure modes of
heads in past major earthquakes are presented to sheet pile bulkheads and their causes under
provide insight into the failure mechanisms ordinary loading conditions are presented.
involved. Because they have a significant Then, typical failure modes of sheet pile bulk-
influence on the response of the bulkhead, the heads in past major Japanese earthquakes and
lateral active and passive earth pressures and their probable causes are given to provide
residual water pressure acting on the bulkhead insight into the mechanisms involved.
under earthquake conditions are given. Soils
considered include cohesionless soils, soils Before discussing the typical failure modes of
with cohesion and angle of internal friction <|> sheet pile bulkheads under ordinary loading
such as silty sand, and cohesive soils such as conditions, it is instructive to examine the
clay and silt. Earth pressure coefficients for behavior of anchored bulkheads under load. The
cohesionless soils are obtained from the bulkhead sheet piling is loaded horizontally on
Mononobe-Okabe [4,5] equations. A procedure is the backface by the active earth pressure and
presented for the rapid seismic analysis of the residual water pressure. The sheet piling
anchored steel sheet pile bulkheads. In the acts as a beam supported by the tie rod near the
procedure, the adequacy of the depth of penetra- top and by the soil below the dredge line in
tion of the sheet piles, the sheet pile section front. Both of these supp rts are not rigid and
in bending, tie rod section and connections, and move and deform under load. The bottom support
tie rod anchorage are evaluated. The earthquake of the sheet piling is derived from the passive
stability of the bulkhead as a gravity wall is pressure developed in the soil*.
evaluated. Where appropriate, stability against
circular slip is examined. Finally, concepts Rotation about the base of a rigid retaining
for strengthening bulkheads that are found to be wall from the at-rest condition is required to
seismically inadequate are presented. develop active and passive pressures in a cohe-
1 . 3 Scope sionless soil. The active (minimum) earth
The analysis procedure presented is for steel pressure is developed by rotating the wall away
sheet pile bulkheads embedded in medium to dense from the soil. By contrast, passive (maximum)
cohesionless, c — 4> ,
and stiff cohesive soils. earth pressure is developed by rotating the wall
For relatively stiff sheet piles, such as those into the soil. For a given cohesionless soil,
fabricated from precast concrete, the analysis the rotation required to develop the active
procedure presented is not valid and the free- pressure is smaller than that required to devel-
earth support method [6] with moment reduction op the passive earth pressure. This relationship
using Rowe's curves [7] is recommended. However, holds even if a safety factor (SF) of 2.0 is
the procedure presented for computing the active used for the passive earth pressure (resis-
and passive earth pressures and residual water tance). For example, using the experimental
pressure is valid. Furthermore, the procedure results given in Figure 1 of Reference 8, the
implicitly assumes that the cohesionless soil following wall rotations are tabulated:
encountered does not lose more than 15% of its Wall Rotation (rad)
static strength under earthquake ground shaking Condition
Dense Sand Loose Sand
and large pore pressure buildup is not a prob-
lem, i.e., soil liquefaction is not a problem. Active 0.0005 0.002
Passive (SF = 2.0) 0.012
The surface of the ground adjacent to the bulk-
head is assumed to be horizontal, which is
*For clarity, the contribution of the horizontal
commonly the case. bearing capacity of the soil is neglected
The wall rotations for the passive case above ciples of static equilibrium, the percentage of
correspond to the maximum pressure divided by a the horizontal load supported by the tie rod is
safety factor of 2. Thus, the rotation of the smaller than that supported by the soil in front
bulkhead head about the toe in dense sand re- of the bulkhead below the dredge line. As
quired to develop the passive resistance is mentioned earlier, both the top and bottom
24 times that required to develop the active supports of the sheet piling are not rigid. The
pressure. The corresponding rotation required tie rod support moves outward due to the removal
to develop the passive resistance in loose sand of tie rod sag, the stretching of the tie rod,
is 3.5 times that required to develop the active and movement of the tie rod anchor required to
pressure develop resistance under load. The anchorage

moves due to the deformation of the supporting
Generally, a rotation about the toe of the sheet soil. The tie rod/anchorage system is generally
piling of the order of 1 degree (0.01 rad) can not stiff enough to prevent the sheet piling
be expected for bulkheads in loose or dense sand from moving outward at that level. The soil
even under ordinary loading conditions. below the dredgeline in front of the bulkhead
also deforms to develop the needed resistance
By contrast, the lateral displacement at the top under load.
of the bulkhead is controlled by the lateral
stiffness of the tie rod anchorage system, i.e., In the great majority of cases the failure modes
the axial stiffness of the tie rod, the axial of sheet pile bulkheads under ordinary loading
and/or bending stiffness of the anchor, and the conditions [6, 9] are:
stiffnesses of the soil supporting the tie rod • Excessive tilting of the bulkhead
anchorage. The stiffnesses of the tie rod and • Lateral translation of the bulkhead and
anchor are determined by the design stresses of fill as a whole
the materials used. Although they do contribute Essentially all the bulkhead failures under
to the overall stiffness of the tie rod anchor- ordinary loading conditions are due to one or
age system, their effect on the overall stiff- two causes [6]
ness of the system is relatively small. The • The designer has estimated the earth
lateral stiffness of the tie rod anchor system pressure and earth resistance on the
is controlled primarily by the stiffness of the basis of the "angle of repose" concept,
soil supporting the anchor. In the case of and
concrete wall or vertical pile or sheet pile • he has failed to notice the source of
anchors, the anchorage system stiffness is weakness in the ground below the toe of
controlled by the compressibility of the soil in the sheet piles.
front of the anchor. For the A-frame pile Except for loose dust-dry sand, there is no
anchor, the lateral stiffness of the anchor relationship between the angle of repose and the
system is controlled primarily by the shear angle of internal friction. Nonetheless, several
stiffness of the soil along its contact surfaces bulkheads have been constructed in which soft
with the piles and the compressibility of the clay was assigned an angle of internal friction
soil at the tip of the compression pile. of 11 degrees based on the approximate slope of
one in five commonly observed on front of a
The combined active earth pressure and residual sheet of hydraulic fill. All these bulkheads
water pressure distribution on the inside face have failed [ 6]
of the bulkhead is generally trapezoidal in

shape with the pressure increasing toward the In some cases, excessive tilting of the bulkhead
bottom of the sheet piling. Based on the prin- was due to the shifting of the tie rod anchorage
235
resulting from the placement of a heavy sur- • Settlement of the ground behind the
charge immediately behind the anchorage. This sheet piling
problem can be avoided by placing the surcharge • Cracking of the concrete apron
far away from the anchorage or supporting the • Fracturing of the tie rod
surcharge by vertical piles. In other cases, • Tie rod anchoring plates or walls shift-
excessive tilting of the bulkhead was caused by ed forward with the ground surface
failure of the tie rod anchorage resulting from behind the plates or walls caved in
the inadequate compaction of the backfill mate- • Parapet wall tilting and collapsing
rial in front of the anchorage. For bulkheads with a relieving platform, the
platform is shifted toward the sea. If the
Some sheet pile bulkheads driven into sand platform has a tie rod anchorage, the anchorage
underlain by a soft clay stratum have failed is also damaged with the tie rod fracturing or
because of an outward movement of the bulkhead the anchor plate being pulled forward.
and fill as a whole [6] . Although the bulkhead
and anchorage designs were satisfactory, the The tie rod anchor is the most critical element
designer had failed to consider the subsoil of an anchored bulkhead. The anchor pull from
stratum that laterally supports the portion of the bulkhead sheet piles are transferred through
the sheet piles below the dredge line. It would the tie rods to the tie rod anchorage. In turn,
be difficult to prevent such a translation. In the tie rod anchorage transfers the anchor pull
such situations, driving the sheet piles through to the soil laterally supporting the anchorage.
the soft stratum and installation of the A-frame A large number of tie rods have been fractured
type of tie rod anchorage is indicated. Alter- and anchor plates have slid out in major Japa-
natively, a relieving platform supported on nese earthquakes (1). Once the tie rod has
piles with point bearing below the soft stratum fractured or the anchor has slid out sufficient-
may be used. ly to reduce the tie rod tension to a small
fraction of its former value, the anchored

If the submerged natural earth in front of the bulkhead behaves like a cantilever retaining
bulkhead slopes downward, the slope can fail wall and tilts excessively.
along a sliding surface below the tie rod an-
chorage and toe of the sheet piles. Several However, it is not practical nor economical to
bulkhead failures of this type have occurred on design and construct a cantilever sheet pile
the Wangpoo River at Shanghai, China. Most of bulkhead for heights greater than 15 feet above
these failures occurred within 2 hours after low the dredge line. Maintaining the alignment for
tide following spring tide [6]. This type of such a bulkhead is a problem under ordinary
failure also suggests the importance of consid- conditions even for heights less than 15 feet.
ering draw down of the water level from tsunami The alignment problem will be acerbated by the
in the seismic analysis of bulkheads. earthquake loading. Anchored bulkheads are
designed for heights as high as 50 feet. Thus,
The majority of the earthquake damage data for special attention should be given to the safety
steel sheet pile bulkheads comes from Japan of the tie rods and their connections and pre-
[2,10-12]. Typical failure modes of steel sheet caution should be taken to prevent the tie rod
pile bulkheads observed in past major Japanese anchorage from sliding excessively due to pas-
earthquakes are as follows: sive type soil failure.

• Sheet piles tilting toward the water
with the revetment face bulging out It is hypothesized that the outward tilting and
bulging of the revetment face observed in past
256
earthquakes was caused by the failure of the tie within the active failure zone. Furthermore,
rod/anchorage system, i.e., the tie rods, their the active soil failure zone spreads away from
connections, and the tie rod anchorage. The the bulkhead with increasing earthquake ground
fracturing of the tie rods and their connections acceleration. That is, the angle that the
can occur suddenly, leading to excessive tilting active failure plane makes with the horizontal
of the bulkhead and bulging of the revetment plane decreases with increasing earthquake
face ground acceleration. Similarly, the angle that
the passive failure plane makes with the hori-
The tilting and bulging of the bulkhead is zontal plane decreases with increasing ground
expected to be a gradual process. This is acceleration. For the concrete wall anchorage
because the active and passive failure of the to be fully effective, the active failure plane
soil surrounding the sheet piling and tie rod drawn from the intersection of the sheet piling
anchorage is progressively gradual. The active with the dredge line and the passive failure
and passive failure surfaces in the soil do not plane drawn tangent to the bottom of the tie rod
form instantaneously nor does sliding along anchor wall or plate should intersect at or
these planes take place instantaneously once above the ground surface (see Figure 2)
they have formed. Sliding along the failure
planes occurs as a series of outward increments Even when the tie rod anchorage remains intact
[3] The anchor pull is equal to the sum of the during a major earthquake, some tilting and
active earth pressure and residual water pres- bulging of the revetment face from the earth-
sure that act over the height of the sheet quake ground motions can still be expected due
piling minus the passive pressure exerted by the to the rotation about the toe of the sheet
soil in front of the sheet piling below the piling and displacement of the tie rod anchorage
dredge line. Each outward displacement increment required to develop the passive resistance of
of the sheet piling occurs when the anchor pull the supporting soil. Sheet pile bulkheads in
plus the active pressure acting on the back of Japan have experienced outward tilting of be-
the tie rod anchorage, induced by the inward tween 5 to 10 degrees without collapse. While
earthquake ground acceleration pulse, exceeds asthetically undesirable, rotations of such
the passive resistance of the soil in front of magnitude are not expected to have a serious
the anchorage. By contrast, an outward directed effect on the function of the bulkhead. However,
earthquake ground acceleration pulse will not utility lines connected to the bulkhead must be
cause the bulkhead to tilt and bulge inward designed to accommodate the relative displace-
because such a movement requires forces that are ments associated with such rotations and remain
about ten times larger than that required to functional
produce the outward displacement. However,
inward tilting of the bulkhead can occur in a Settlement of the ground immediately behind the
soft soil site as a result of sliding instabili- sheet piling and the tie rod anchorage, and the
ty due to circular slip of the bulkhead and fill cracking of the concrete apron is generally
as a whole. caused by the outward and downward movement of
the active soil wedge (Figure 2) by sliding
In some cases, the tie rod anchorage failed along the failure surface. Granted, a portion
because it is located too close to the bulkhead. of this localized settlement may be caused by
The resistance of the tie rod anchorage is the uneven consolidation of the soil during the
reduced if it is located near the active soil earthquake ground shaking. Uneven settlement of
failure zone. The resistance of the tie rod the soil from earthquake ground shaking can only
anchorage is nil when it is located entirely be avoided by selecting sites that are not
257
susceptable to such effects and carefully con- 3- LATERAL EARTH PRESSURES AND RESIDUAL WATER
trolling the compaction of the backfill soil. PRESSURES
In this section, the active and passive lateral
Parapet attached to the top of the bulkhead earth pressures recommended for the rapid seis-
sheet piles will tilt with the piles. If the mic analysis of sheet pile bulkheads are pre-
connections of the parapet to the sheet piles sented. To facilitate calculations, equations
are inadequate, the parapet can collapse into for computing the pressures at the top and
the water during a severe earthquake. bottom of each soil layer, the resultant force
acting on each layer, and its location with
For sheet pile bulkheads with relieving plat- respect to the top of each layer are given. The
form, the shifting of the platform toward the angles that the active and passive failure
sea is caused by the earthquake induced inertia planes make with the horizontal plane are pre-
forces on the platform and the fill and sur- sented to aid in the determination of the re-
charge that it supports and the inadequate quired location of the tie rod anchorage rela-
lateral stiffness and resistance provided by tive to the bulkhead. Procedure for computing
vertical piles supporting the platform. The the resulting horizontal force on the bulkhead
inertia force on the platform can be decreased from line and point loads on the ground surface
by reducing the total weight (mass) of the in the vicinty of the bulkhead is given. Final-
platform. Lateral stiffness and resistance ly, a procedure for computing the residual water
required by the platform are best provided by pressure acting on the bulkhead is presented.
the installation of batter pile under the plat- 3.1 Lateral Earth Pressures
form. A tie rod/deadman anchorage is not effec- For cohesionless soils, the active (K^) and
tive for a relieving platform supported on passive (Kp) earth pressure coefficients are
vertical piles because such an anchorage gener- obtained from computer generated curves based on
ally cannot accommodate the large outward move- the Mononobe-Okabe [4,5] equations for a verti-
ment of the platform and supporting piles with- cal wall with a wall friction angle (6) to angle
out fracturing the tie rod and/or shifting the of internal friction ($) ratio, delta/phi of 0.5
anchorage (Figures 3 and 4). The angles that the active
(Z^) and passive (Zp) failure planes of the soil

The yielding or fracturing of the sheet pile make with the horizontal plane (see Figure 2)
sections of the bulkhead due to bending in past are also obtained from the computer generated
major earthquakes is rare. The observed yield- curves, based on the Mononobe-Okabe equations,
ing of the sheet pile section in bending is given in Figures 5 and 6. The validity of the
associated with excessive tilting. The local- Mononobe-Okabe equations for computing the
ized yielding occurred at only one section along lateral earth pressures on bulkheads under
the length of the sheet piling. Furthermore, simulated seismic excitation has been verified
the observed yielding did not lead to the col- by a number of laboratory experiments (13-18).
lapse of the bulkhead. Fracturing of the sheet However, it should be noted that all the experi-
pile section has occurred at welded splices due ments were performed on dry sand instead of
to excessive residual stresses from welding. saturated sand commonly encountered in sheet
Thus, care must be taken during field welding to pile bulkhead construction.
reduce the resulting residual stresses. The
rare occurrence of yielding or fracturing of the In Reference 3, a parametric study was performed
sheet pile section by bending suggests that the on the effects that the earthquake .nn'ln.ii
Japanese design procedure used to determine the coefficient, k, the soil angle of internal
bending moments in the sheet piling is adequate. friction, <J> ,

and the wall friction angle to
238
angle of internal friction ratio, 6/<|), have on Of the observed trends, the increase in active
the active and passive pressure coefficients (K^ earth pressure coefficient and the decrease in
and Kp) and the angles that the active and passive earth pressure coefficient with earth-
passive failure planes of the soil make with the quake acceleration coefficient have the greatest
horizontal plane (Z^ and Zp) . The following are effect on the seismic response of sheet pile
the trends observed from the results of the bulkheads

study pertinent to the behavior of anchored
sheet pile bulkheads According to Figure 3, the K^ value for a soil
The active lateral earth pressure coefficient with 4> (phi) = 35 degrees increases from 0.25 to
0.38 (about 52%) when the earthquake accelera-

• Increases with the earthquake acceler- tion coefficient increases from 0 to 0.20.
ation coefficient, k Because the lateral earth pressure on the bulk-

• Decreases with increasing soil angle of head increases with K. ,
the bulkhead must be
A
internal friction, <|) able to accommodate this increase and remain
• Decreases somewhat as the wall friction functional. According to Figure 4, the Kp value
angle to soil angle of internal fric- for a soil with 4> = 35 degrees decreases from
tion ratio 6/<{> increases 7.3 to 6.1 (about 16%) for the same increase in
The passive lateral earth pressure coefficient, earthquake acceleration coefficient. Because
the sheet pile bulkhead generally depends on the
• Decreases with increasing earthquake passive resistance of the soil below the dredge
acceleration coefficient, k line and of the soil in front of the tie rod
• Increases with increasing soil angle of anchor to prevent outward translation and rota-
internal friction, 4> tion about the toe of the sheet piling, its
• Increases with the absolute value of the resistance to outward displacement would be
wall friction angle to soil angle of reduced under earthquake ground shaking.
internal ratio, 6/4>
The angle that the failure plane of the active The observed increase in K. and decrease in
A P
soil wedge makes with the horizontal plane, Z^: values with increasing earthquake acceleration
• Decreases with increasing earthquake coefficient suggest that the tie rod pull will
acceleration coefficient, k also increase with the earthquake acceleration
• Increases with the soil angle of coefficient. This is because the tie rod pull
internal friction, 0 is generally determined from the difference
• Decreases slightly as the ratio of wall between active and passive pressures acting over
friction angle to soil angle at internal the height of the sheet piling. On the other
friction (6/<}>) increases hand, the capacity of a concrete wall or short
The angle that the failure plane of the passive sheet pile tie rod anchor is equal to the dif-
soil wedge makes with the horizontal plane Zp: ference between the passive pressure acting in
• Decreases with increasing earthquake front and the active pressure acting on the back
acceleration coefficient, k of the anchor. Thus, the capacity of a tie rod
• Increases or decreases with increasing anchor decreases with increasing earthquake
soil angle of internal friction, <{>, acceleration coefficient while the anchor pull
depending on the range of earthquake it resists increases.
acceleration coefficient.
• Decreases as the ratio of the wall Futhermore, decreases with an increase in 4>
friction angle to soil angle of internal and 6/<}> ratio. The opposite is true for Kp.
friction (6/<{)) increases Whenever there is an uncertainty in the backfill
239
angle of internal friction 0, the smaller <}) of representative samples. When such data are
value should be used to determine the seismic not available, the data presented in Table 1 can
adequacy of the bulkhead. only decreases be used.
slightly with increasing 6/(J) ratio. By contrast,
Kp increases significantly with the 6/4> ratio. Reliance should not be placed on the passive
Although experimental data indicate that a resistance provided by soil backfill unless its
6/4> = 2/3 can be obtained under favorable condi- placement is well controlled and thoroughly
tions, 6/ <J)
= 0.5 was selected for the analysis compacted. Backfill soils for bulkheads are
procedure to provide static Kp values that usually deposited under water and their struc-
closely match values recommended by Terzaghi tures are likely to be loose.
[ 6 ].
3.1.1 Active Pressure- Soil in the field generally
The observed decreases of the angles and Zp consists of different layers and the earth
with increasing earthquake acceleration coeffi- pressures are computed at the top and bottom of
cient can shift the deadman anchor from the each layer (Figure 7). The computation for each
fully passive zone (to the right of the passive pressure component is carried out from the
failure plane in Figure 2) to the intermediate uppermost layer downward. In cases where the
zone between the active and passive failure ground water table is located in a soil layer,
planes where the lateral resistance of the the layer is separated into two artificial
anchor may be reduced. layers, one above the water table and one below
the water table. Computations are carried out
In this study, a sand is considered as silty if for a unit width of bulkhead.

it has more than 5% by weight that passes the
number 200 mesh.
3.1.2 Cohesionless Soils
For soils with cohesion (c) and internal fric- = K

A.
0 )
ti l ti
tion angle ($) ,
c-<{> soils, such as silty sand,
the active and passive pressures are computed =
approximately by substituting the K^ and Kp \i
K. q + p7
Z
ti
.
+ Y.
li
h. ( 2 )
values from Figures 3 and 4 into the Rankine

equations for active and passive pressures. The where: p. active earth pressure at top of
A
ti layer i; lb/ft 2
Z. and Z„ angles are assumed to be the same as
A P
those for cohesionless soils with the same angle p = active earth pressure at bottom
^bi of layer i; lb/ft 2
of internal friction.
K^ = active earth pressure coefficient
i for layer i (See Figure 3)
For cohesive soils, the active earth pressure is
2
assumed to be a function of (1 + k) ,
where k is q = surcharge; lb/ft
the earthquake acceleration coefficient. By

contrast, the passive earth pressure is assumed
to be a function of (1 - k). The cohesion, c, effective vertical soil pressure
at the top of soil layer i;
is assumed to be unaffected by the earthquake lb/ft 2 This is equal to the aum
.
acceleration coefficient. Approximate equations of the effective vertical soil

pressures from the soil layers
for Z^ and Zp were developed for k < 0.2.
above. Use the dry unit weight,
y, for soil layers above the
water table and the submerged
Whenever possible, it is best to determine the
unit weight, y' ; for soil layers
needed soil properties from laboratory testing below the water table; lb/ft 2
240
y. = dry unit weight of the soil in ic. Consequently, the location of the seismic
1
layer i, if it is located above
resultant force is located at a height above the
the water table. If it is below
the water table, use the sub- base higher than one-third times the height of
merged unit weight y' lb/ft 3
;
the wall considered. However, the location of
h^ = thickness of soil layer i; ft the resultant force above the base increases
with the rigidity of the structure. It is
located at about H/3 above the base for a very

= ka q + p (lb/ ft) (3)
flexible structure and at about 2H/3 above the
2
where: c ;
= cohesion of soil layer i; lb/ft base for a very rigid structure, where H is the
2
= Ka (q -r
p + Y* h.'l - 2 c.jK (ib/ft ) (4)
z height of the wall or bulkhead. For simplicity
Cohesive Soils. and convenience, both the static and seismic
q + pz (1 + k) - 2 c. (lb/ft ) (5)
\
lateral earth pressure distributions are assumed
to be triangular and increase linearly with
PA = 0.5 / q + p \
(1 + k) (lb/ft )
( 6)
z
ti ' ti / depth in this study. Accordingly, the resulting
where: k = earthquake acceleration coefficient
pressure distribution for each soil layer will
Use the larger value from Equations 5 and 6.
be triangular or trapezoidal.
(1 + k) - 2c. (lb/ f t
P
V q + p
z
ti
+ V
i
h.
'
)
) (7)
+ k
For ease of computation, artificial layers are
PA = 0.5 (q + p z + Yi h. j(l ) )
( 8)
bi
sometimes created within a given soil layer and
Use the larger value from Equations 7 and 8.
the force resultants and their locations are
3.1.3 Passive Pressure then computed for the newly created layers.
Neglect the effect of the
surcharge on the passive resistance. For a triangular pressure distribution, the
Cohesionless Soils .
resultant lateral force of soil layer i is
pp
t
.
= V Pz
t
.
(lb/ft2)
(9) P.
1 2
P bi h.i (15)
= 2
Pp K + V H (lb/ft
Pi ( i i ) ) (10) and located at
bi
where: Pp = passive earth pressure at top of
ti soil layer i; lb/ft 2
2
h. h. (16)
pp = passive earth pressure at bottom i 3 i
bi of soil layer i; lb/ft 2
Kp = passive earth pressure for soil below the top of the layer.
i layer i (See Figure 4)
C- Soils
<j> .
For a trapezoidal pressure distribution, the

2
P
P
ti
.
= K
P
P
Z
+ 2 (lb/ft ) (ID resultant lateral force of soil layer i is
l ti \ i
P. + (17)
Pbi’ h i
2
*
' ") * 2t .\[v (lb/ft ) (12) x 2
(p
ti
Cohesive Soils.
and located at
2
P = p (1 - k) + 2 c (lb/ft ) (13)
P.
ti
. Z
ti
i h (p ti * 2 p
. _ i bi>
" (18)
Pp = + Y h. - k) +
2
1 3
(
P,
Z
\ (1 2 c. (lb/ft )
(14)
*bi V
ti
1
*/ 1
3.1.4 Resultant Force & Location below the top of the layer.
Results from lab- where: p 1 = lateral earth pressure at top of
soil layer i; lb/ft 2
oratory experiments on model retaining walls or
anchored sheet pile bulkheads subjected tb = lateral earth pressure at bottom
p^
of soil layer i; lb/ft 2
simulated earthquake excitation [13-18] indicate
that the distribution of the seismic component h^ = thickness of soil layer i; ft
of the earth pressure is approximately parabol-
3 • 2 Angles that the Active and Passive Soil two-thirds of the tidal range above the low
Failure Planes Make with the Horizontal Plane water level (LWL) (see Figure 7). Under earth-
quake excitation, the pore water in the soil is
3.2.1 Cohesionless and c-<b Soils Determine Z
A assumed to move with the surrounding soil.
and Zp from Figures 5 and 6, respectively. To
Hence
check the required location of the deadman
2
p RW = 0 (lb/ft ) (21)
anchor relative to the bulkhead for full effec-
tiveness, plot the active soil failure plane 2
--P RW = (1 + k) yw h u (lb/ft ) (22)
angles, Z^, starting with the layers that con-
b
tain the intersection of the bulkhead sheet where: P = residual water pressure at the
R^
t top of the ground water table;
piles with the dredge line (see Figure 2) and
lb/ft 2
work upward. Then, plot the passive soil failure
P = residual water pressure at the
plane with KW
Zp, starting the soil layer that b LWL; lb/ ft 2
contains the bottom of the anchor. Draw the Zp unit weight of seawater =
yW =
line for that layer so that it is tangent to the 64 lb/ft 3
bottom of the anchor. h = height of the residual water

u
level above the LWL; ft
3.2.2 Cohesive Soils ( Determine Z. and Z from
A p k = earthquake acceleration
Equations 19 and 20 given below. These approxi-
coefficient
mate equations were developed by the writer for
4. PROCEDURE FOR ANCHORED SHEET PILE BULKHEADS
earthquake acceleration coefficients, k, of less
In this section, a rapid seismic analysis proce-
than or equal to 0.2 using the computer generat-
dure developed for the analysis of anchored
ed curves from the Mononobe-Okabe equations.
sheet pile bulkheads is presented. The analysis
They are not valid for earthquake accelerations
procedure includes the following major steps:
coefficient k greater than 0.2.
• Determination of the loads acting on the
45° - 1.4 tan 1
k (19)
sheet piling
45° - 0.66 tan *
k (20) • Determine the adequacy of the depth of
penetration of the sheet piling

3 . 3 Line and Point Loads
• Determine the maximum bending stress in
Heavy line and point loads near the bulkhead are
the sheet piling
best supported by vertical piles to minimize
• Determine the adequacy of the tie rod
their effect on the bulkhead. For line loads
and tie rod connections.
and point loads not supported by vertical piles,
• Determine the adequacy of the tie rod
use the procedure given in Figure 8 to compute
anchorage
the equivalent horizontal force and its location
• Check for other potential modes of fail-
on the bulkhead. The curves given in the figure
ure that have not been considered pre-
were developed by Terzaghi [6] on the assumption
viously that can have a significant
of an unyielding wall and are approximately
effect on the behavior of the bulkhead
double those values from elastic equations. A
considered, such as circular slip for
sheet pile bulkhead is generally flexible and
sheet piling embedded in soft soil
yields (displaces) under load, reducing the
sites
loading exerted by the line or point loads.
Because of the conservatism built into the After an examination of the various design and
curves, no attempt was made to modify the curves analysis procedures for anchored sheet pile
for seismic effects. bulkheads in the literature [1,2,6,8,9,19-22),
3 . 4 Residual Water Pressure the procedure given in Chapter 5 of Reference 19
Unless more accurate information is available, for sheet pile type quaywalls was selected for
assume the residual water level, h ,

to be adaptation in this studv The procedure was
u
242
selected because of its simplicity. Furthermore, can be expressed by the following generalized
the adequacy of the procedure has been verified equation

sheet
by
pile
surveys
bulkheads
of postearthquake
designed or
damage
analyzed
to
by the
SF =
^resistin g,
driving
> ^ (23)
procedure [1,11,12].
Loading where M resisting moment about

4.1 resisting
the tie rod setting point
The loading on the anchored sheet pile bulkhead from the passive earth
pressure and the active
included the active earth pressure, the residual
earth pressure above the
water pressure, and the passive earth pressure tie rod level
(Figure 7). In addition, the loading should M driving moment about the
driving
tie rod setting point
include the earthquake-induced inertia force on
from the active earth
the concrete parapet if appropriate. This pressure and residual
water pressure
outward directed force is equal to the mass of
the parapet per unit width of bulkhead times the NOTE: Include the moment contribution from line
earthquake acceleration coefficient, k. Berthing or point load to H. . . or M . . .
driving resisting
forces on the bulkhead is neglected in the A line or point load contributes to
analysis because the probability of the maximum M, . if the resulting lateral force
.
driving
berthing force acting simultaneously with the acts on the sheet piling below the tie
earthquake-induced forces is very remote. rod setting point. Otherwise, it contri-
butes to M ...
resisting
The procedure for computing the lateral earth
pressures and residual water pressure acting on In normal design, the required depth of penetra-
the sheet pile bulkhead for cohesionless, c-<j>, tion for the sheet piling is computed from an
and cohesive soils were given previously in the equation similar to Equation 23. Then, the
section on "Lateral Earth Pressures and Residual actual depth of penetration is determined by
Water Pressure." increasing the computed depth by a percentage of
the computed value to account for uncertainties
Currently, there are no established seismic in the soil properties from boring samples
coefficients for the pseudo-static analysis of (typically spaced at 100-foot intervals) and
anchored sheet pile bulkheads in the United potential scouring at the dredge line. In the
States. It is suggested that the seismic accel- U.S., the computed required depth of penetration
eration coefficients given in Table 2 be used is commonly increased by 20%. By contrast, the
tentatively for the rapid analysis procedure Japanese authorities [9] specify a 20% increase
until more precise data becom available. above the computed depth for cohesionless soils
4 • 2 Depth of Penetration and a 50% increase above the computed depth for
The steel sheet piles are driven to a sufficient cohesive soils. In determining the seismic
depth below the dredge line to limit the outward adequacy of an existing bulkhead, use the com-
movement of the sheet piles under load to a puted depth of penetration whenever possible.
sufficiently small value. Otherwise, use 83.3% (1/1.2) or some appropriate
fraction of the actual depth of penetration.
The required depth of penetration is determined
by taking the summation of moments about the tie For a site consisting of different soil layers,
rod setting point of the forces due to the it is convenient to compute the resultant active
active earth pressure and residual water pres- and passive forces acting on each soil layer and
sure and the passive pressure (Figure 7). This their location relative to the tie rod setting
20
point before computing M. . . and M moment occurs is located by summing algebra-
driving resisting
in Equation 23. ically the horizontal forces acting on the
4 . 3 Bending Stress virtual beam sequentially from the top. The
In the determination of the maximum bending maximum bending moment is located in the soil
moment in the sheet piling, the sheet piling is layer below the tie rod where the net shear
assumed to be a virtual simply supported (SS) force on the virtual beam becomes zero. The
beam with supports at the tie rod level and at location of the maximum bending moment within
the dredge line (Figure 9). The beam is loaded that layer is obtained by summing all the hori-
by the active earth pressure and the residual zontal forces from the top of the bulkhead to
water pressure upward from the dredge line. If the top of the soil layer and adding the force
3.
appropriate, include the outward earthquake-in- within the layer expressed as an unknown, x,
duced inertia force acting on the concrete distance from the top of the layer and setting
4.
parapet and lateral forces acting on the sheet the resulting equation to zero and solving for
piling from a point or line loads. x .
The above SS beam assumption and loading condi- Determine the maximum bending moment corre-
tion is valid for relatively flexible sheet sponding to location x on the virtual beam.
piles embedded in medium to dense cohesionless

soil and stiff cohesive soil sites. For sites Determime the maximum bending stress in the
where the compact material below the dredge line sheet piling (f^ ) and compare it against the
is located below a layer of soft material, permissible value.

assume the lower support of the virtual beam to
be located at the bottom of the soft layer. For
f
bs
= -5n|2L-§.
S
<
-
12 F (24)
y
s
special situations such as relatively stiff where: M = maximum bending moment in the
max s . . .
sheet piling; in-lb

steel sheet piling embedded in a soft soil site
or the toe of the sheet piling embedded in rock, S sections modulus of the sheet
s
piling per foot of bulkhead;
the free-earth or other appropriate methods
should be used in determining the maximum bend-
F = minimum specified static yield
ing moment [8] ^ strength of the sheet piling
material; lb/in 2
Evaluation of the adequacy of the sheet piles in The maximum stress of 1.2 F under seismic
y
bending consists of the following steps: loading was considered acceptable because the
1. Compute the reactions on the virtual beam maximum moment is located in a section below the
shown in Figure 9 by summing moments about the low water level (LWL) where corrosion is not a
top and bottom beam supports according to the problem. The yield strength of steel increases
principles of statics. Check the answers by by about 10% above the static value under rapid
taking a summation of the horizontal forces. load with a rise time between 0.01 second and
Remember to include other appropriate forces 0.1 second such as those from a major earthquake
mentioned earlier but not shown in Figure 9. [24,25). The actual yield strength of steel is
Neglect any loading on the beam below the dredge generally about 20% higher than the minimum
line specified static value. Furthermore, a maximum
bending stress of 1.2 F corresponds to a duc-
2. Determine the location of the maximum bend- tility factor* of about 1.2 which is considered
ing moment by locating the point of zero shear acceptable functionally. As mentioned earlier,
on the sheet piling below the tie rod setting

di spl acement /d -
*Ductility factor = maximum i »
point. The soil layer in which the maximum placement at yield.

2
yielding of bulkheads in bending in past major where: A = net area of tie rod; in.
net
earthquakes has been rare. = static yield strength of tie rod
F
4 . Tie Rod and Tie Rod Connection Stresses F material; psi
Once the tie rod and/or its connections have
4 . Wale Stress
failed, an anchored sheet pile bulkhead behaves
1. Determine the maximum bending moment in the
as a cantilevered retaining wall and tilts
wale, M ,
from
max w
outward excessively. The tie rod may be sub-
T d
M (ft-lb) (27)
jected to secondary bending stress from the max w 10
uneven settlement of the surrounding soil as a

2. Determine the maximum bending stress in the
result of earthquake ground shaking. Moreover,
wale, f, :
bw
the tie rod and its connections are generally
M
located near the tidal zone where corrosion is
f = (12 in -/ ft ) 1 °- 6 F (P si )
bw y
y
most severe. Thus, it would be desirable to w
= section modulus of wale; 3
where: S in.
assign a higher safety factor to the tie rod and w
its connections to ensure that they will not F = static yield strength of wale
^ material; psi
fail under earthquake loading.
To remain functional, the tie rod and its con- 4 . Tie Rod Anchorage
nections must be able to transmit the static and The tie rod cannot carry out its intended func-
earthquake induced forces without fracturing. tion if its anchorage is located too close to
Accordingly, the maximum stresses in the tie rod the bulkhead where the capacity of the anchorage
and its connections are limited to less than or is reduced. As mentioned earlier in the section
equal
4.5 to 0.6 F^, where is the static yield on "Lateral Earth Pressures and Residual Water
strength of the material considered. No attempt Pressure," the active soil failure zone (wedge)
will be made to cover the various configurations spreads away from the bulkhead and the passive
of the tie rod connections . Equations for soil failure zone spreads toward the bulkhead
computing the maximum axial stress in the tie with increasing earthquake acceleration coeffi-
rod and the maximum bending stress in the wale 4.8
cient. This, in effect, reduces the distance
are as follows. between the anchorage and the bulkhead. In
addition, the tie rod anchorage must provide

Tie Rod Stress adequate resistance against lateral displacement
1. Determine the maximum tension acting on the if the bulkhead is to remain functional. This
tie rod, T, from requirement is satisfied indirectly by providing
T = Ap d/cos 0 (lb) (25) the tie rod anchorage with a specified safety
where: A reaction of the tie rod setting factor against ultimate soil failure.
P point from the determination of
the maximum bending moment in the
sheet piling (Figure 9); lb/ft Location
d = tie rod spacing; ft

4.8.1 Concrete Wall Anchorage . To be fully effec-
0 = angle that the tie rod makes with
tive, the position of the concrete wall anchor-
the horizontal plane; degree
age should be located such that the active
2. Determine the maximum axial stress in the failure plane drawn from the intersection of the
tie rod, f ,
from sheet piling with the dredge line and the pas-
at
sive failure plane drawn tangent to the lower
edge of the concrete wall anchorage does not
f = < 0.6 F (psi) (26)
at J
net intersect below the ground surface as shown in
245
Figure 2. If the active and passive failure ratio L/T > 5, where T = (EI/f)^^, E = modulus
planes intersect below the ground surface, the of pile material, I = moment of inertia of pile,
effective height of the wall anchorage is equal and f = coefficient of variation of subgrade
to the actual height, h^, minus the height of reaction of the supporting soil [8). When the
the top of the wall anchorage above the inter- sheet pile anchorage cannot be considered as
section of the two planes (Figure 10). The long piles, the portion of the sheet pile an-
effective height is used for computing the chorage below a distance of 1 ,/2 should be
ml
lateral resistance of the anchor. neglected in the analysis. Here, 1^ has the
same meaning as for vertical pile anchorage.
4 . 8 . 2Vertical Pile Anchorage- To be fully effec-
tive, the position of the vertical pile anchor- 4.8.4A-Frame Pile Anchorage- To be fully effec-
age should be located such that the passive tive, the position of the A-frame piles should
failure plane drawn from a point 1/3 below the be located behind the active failure plane drawn
tie rod setting point at the anchorage and the from the intersection of the bulkhead sheet
active failure plane drawn from the intersection piling with the dredge line, Figure lib. Other-
of the bulkhead sheet piling with the dredge wise, neglect the soil support for the portion
line intersects at or above the plane containing of the A-frame piles above the active failure
the tie rod as shown in Figure 11a. Here, 1 plane in the analysis.
ml,
is the depth from the tie rod setting point at
the anchorage to the point where the bending 4.9 Resistance
moment in the vertical pile first becomes zero.
The pile head is assumed to be free and the 4 . 9 . IConcrete Wall Anchorage- Determine the
virtual ground surface is assumed to be located safety factor of the concrete wall anchorage
at the tie rod setting point of the anchorage. against sliding under the action of the active
A detailed discussion of the analysis of pile and passive pressures and inertia force (Figure
behavior and capacities under lateral loads is 12) using the equation:
beyond the scope of this study. Readers unfa-

SF = > 2.0 (29)
miliar with the subject can refer to References A + E + !
P A w
26 to 30. For readers familiar with the subject, where: A^ = anchor pull; lb
materials needed for pile analysis are given in
E^ = active force; lb
Reference 8.
Ep = passive force; lb
When the active and passive failure planes of I = wall inertia force; lb
w
the soil intersect below the horizontal surface
= k x W
w
containing the tie rod, assume the ground sur-
k = earthquake acceleration coefficient
face to be located at the same level as the
intersection of the two planes in the deter- W weight of wall anchor; lb
w
mination of the response and capacity of the
vertical pile anchorage. The procedure for determining the earth pres-
sures acting on the concrete wall anchorage it
4.8.3Sheet Pile Anchorage- When the sheet pile the same as that for the bulkhead sheet piles,
anchorage can be considered as a long pile, it except that the residual water pressure on each
should be located relative to the bulkhead as side of the wall anchorage cancel each other
for vertical pile anchorage for full effective-

ness. A sheet pile is considered long if its Evaluate the adequacy of the concrete wall in
length, L, to relative stiffness factor, T, bending in the horizontal and vertical direr-
tions by assuming that the earth pressure is When the sheet pile cannot be considered as a
distributed approximately uniformly and that the long pile, it should be analyzed as a concrete
anchorage is a continuous beam in the horizontal wall anchorage with the assumption that the
direction and a '.cantilever beam in the vertical sheet pile terminates (not effective) at a point
direction. l,/2
ml
below the tie rod setting point at the
T d anchorage (Figure 13) . Here 1 is the distance

(30)
“H 12 between the tie rod setting and the first point
T h where the bending moment is equal to zero in the

=
A
(31)
“V 8~T" sheet pile when the sheet pile anchorage is
where: ft, = maximum horizontal bending moment;
considered as a free-headed long pile with the
ft-lb
tie rod setting point as the virtual ground
= maximum vertical bending moment;
"v surface
ft-lb/ft
T = tie rod tension; lb

4 . lOA-Frame PileAnchorage A-frame pile anchorage
d = tie rod spacing; ft should be analyzed as batter piles subjected to
the external loading from the tie rod tension
h height of concrete wall anchorage;
A Whenever possible, the utlimate capacity
ft [8] .
of the piles should be determined from field
The concrete wall anchorage should have a safety tests. The piles should be provided with the
factor of equal to or greater than 1.5 against following safety factor against ultimate failure
ultimate failure in flexure with a strength of the soil:
reduction factor of 4> = 1.0 for flexure. Safety
Loading Type Factor
(Readers unfamiliar with the design and analysis
of reinforced concrete structures can refer to Compression
Bearing 1.5
Reference 31.)
Friction 2.0
Tension 2.5
4 . 9 . 2Vertical Pile Anchorage. Vertical pile
anchorage should be analyzed as vertical piles For A-frame piles located behind the active soil
subjected to an external load from the tie rod failure plane (Figure lib), the piles are pri-
tension, T [8]. Whenever possible, the ultimate marily subjected to axial stresses. The soil
lateral resistance of the vertical piles should strengths generally control the pile design.
be determined from field tests. A safety factor However, the axial strengths of the pile and its
(SF) of >2.0 should be provided against the connections should have a safety factor of equal
ultimate soil failure. If concrete piles are to or greater than 2.0 against ultimate failure
used, they should have a SF of >1.5 against to ensure that the anchorage remains functional
ultimate failure in flexure with a strength under the earthquake loading. Where appropriate,
reduction factor (|) = 1.0. If steel sheet piles the bending strengths of the A-frame piles
are used, the maximum bending stress should be should be considered. The recommended safety
limited to 1.2 F where < F = minimum speci- factors for concrete and steel piles in flexure
- y
,
y
fied static yield strength. were given earlier.
4.11 Other Considerations

4.9. 3Sheet Pile Anchorage. When the length of Besides failing by excessive outward tilting due
sheet pile below the tie rod setting point is to failure of the tie rod anchorage system, an
long enough to be considered as a long pile anchored sheet pile bulkhead may fail by sliding
(L/T > 5), the sheet pile should be analyzed as instability of the soil and bulkhead as a whole
a vertical pile anchorage. due to lateral movement of the soil that pro-
2k
vides support to the buried portion of the
bulkhead sheet piling. In this section, concepts for strengthening
sheet pile bulkheads that are found to be seis-
Circular Slip in Soft Soil Sites . Sheet pile mically inadequate by the analysis procedure are
bulkheads embedded in soft clay soil sites presented. Theoretically, concepts for strength-
should be evaluated for sliding instability ening inadequate sheet pile bulkheads are only
against circular slip. A safety factor of equal limited by the imagination of the engineer and
to or greater than 1.2 against circular slip is his experience. However, the strengthening
adequate. An example of the analysis procedure concepts that are practical for a particular
presented in the previous sections including a bulkhead at a particular site are controlled by
procedure for performing circular slip analysis the soils at the site, configuration of the
is given in Reference 32. utility lines in the vicinity of the bulkhead,
and economics.
Slide Along Soft Stratum Beneath the Toe of Depth of Penetration
Sheet Piling . Sheet piles embedded in sites It is not practical to increase the depth of
where the relatively compact soil beneath the penetration of the sheet piles for those exist-
dredge line is underlain by a shallow layer of ing bulkheads found to have inadequate depth of
soft material beneath the toe of the sheet piles penetration. The only thing that can be done is
should be checked for sliding instability along to increase the passive resistance of the soil
the bottom of the soft layer. The bulkhead is below the dredge line in front of the sheet
analyzed as a virtual gravity type wall whose piling and/or to reduce the active earth pres-
lower edge is defined by the bottom of the soft sure acting on the back of the sheet piling by
soil layer and right boundary defined by the chemical or cement grout injection of the soil.
right-most extent of tie rod anchorage and whose Chemical or cement grouting, whose effectiveness
left boundary is defined by the bulkhead sheet is generally limited to granular soil sites, are
piles (Figure 14). Potential outward sliding of very expensive and may only be justifiable in
the wall is driven by earthquake inertia force cases where the lead time for the construction
induced on the mass bounded by the virtual wall of a new bulkhead is a problem.
boundaries and the active earth pressure and Bending Capacity of Sheet Piles
residual water pressure acting on the right Theoretically, steel cover plates can be welded
boundary of the virtual wall. This potential by divers at the critical section of the sheet
movement is resisted by the frictional and piles found to be inadequate. However, this
cohesive forces along the bottom of the virtual generally cannot be justified economically.
wall and the passive earth pressure exerted by Fortunately, yielding of the steel sheet pile
the soil on front of the virtual wall below the due to bending has occurred only rarely in past
dredge line. A safety factor, resisting force major earthquakes. Yielding of the sheet piles
divided by driving force, of equal to or greater from excessive bending during earthquakes is
than 1.2 is adequate. expected to be gradual and not catastrophic.

Tie Rod/Anchorage
If tsunami draw down is a potential hazard at New tie rods can be installed at above or be-
the bulkhead site, the increase in hydrostatic tween the existing tie rods if the existing tie
pressure on the bulkhead from the tsunami draw rods are found to be inadequate. New wales will
down should be considered in the seismic evalu- be needed if the new tie rods are to be instal-
ation of the bulkhead. led at a level above the existing ones. The
capacity of the tie rod anchorage will generally

5- CONCEPTS FOR STRENGTHENING INADEQUATE BULKHEADS those
be found to be inadequate for bulk!
with inadequate tie rods. Existing concrete 8. Naval Facilities Engineering Command.
wall anchors can be strengthened by driving NAVFAC Design Manual 7.2: Foundations and earth
short sheet piles in front of the concrete structures, Alexandria, Va . ,

May 1982.
anchorage and pouring new concrete between the 9. James R. Ayers and R.C. Stokes. "The design
two to make the old concrete wall and new steel of flexible bulkheads,” American Society of
sheet piles act compositely together as a unit. Civil Engineers, paper no. 2676, vol 119, 1954,
This technique was proposed by Casagrande [33] pp 373-383.

for improving the anchorage of an existing 10. R. Amano, et al. "A seismic design of quay
bulkhead at Cleveland, Ohio. Alternatively, the walls in Japan," in Proceedings of the First
new tie rods can be anchored to new tie rod World Conference on Earthquake Engineering,
anchorage constructed behind the existing one. Berkely, Calif., Jun 1956, paper 32-1.
The location of the new anchorage will most 11. Japanese Port and Harbour Research Insti-
probably be determined by the location jf the tute. "The damage to port structures by the
existing utility lines in the vicinity of the 1978 Miyagi-Ken-Oki earthquake," by H. Tsuchida,
old anchorage and the location of the active and et al., Technical Note no. 325, 1979.
passive failure planes in the soil. 12. Japanese Port and Harbor Research Insti-
tute. "Damage to harbour structures by the 1968
6. ACKNOWLEDGMENT
Tokachi-Oki earthquake, the investigation of the
The continual support by Mr. J. Tyrrell (NAVFAC Tsunami caused by the 1968 Tokachi-Oki earth-
Code 04BA) is appreciated. quake,” 1968. (in Japanese)
13. N. Mononobe and H. Matsuo. "On the deter-
7. REFERENCES
mination of earth pressures during earthquakes,"
1. S. Okamoto. Introduction to earthquake in Proceedings of the World Engineering Confer-
engineering. New York, N.Y. ,
John Wiley and ence, vol 9, 1929, p 176.
Sons, 1973. 14. L.S. Jacobsen (1939). Described in Appen-
2. S.D. Werner and S.J. Hung. Seismic response dix D of "The Kentucky Project" Technical Report
of port and harbor facility, Agbabian Associ- no. 13, Tennessee Valley Authority, 1951.
ates, Report no. R-8122-5395. El Segundo, 15. H. Matsuo. "Experimental study on the
Calif., Oct 1982. (National Science Foundation distribution of earth pressure acting on a
Grant no. CEE-8012337) vertical wall during earthquakes," Journal,

3. Naval Civil Engineering Laboratory. Tech- Japanese Society of Civil Engineers, vol 27,
nical Memorandum M-51-83-11: A rapid seismic no. 2, 1941.
analysis procedure for gravity quaywalls, by 16. Y. Ishii, H. Arai, and H. Tsuchida.
T.K. Lew. Port Hueneme, Calif., Aug 1983. "Lateral earth pressure in an earthquake," in
4. N. Mononobe. "Earthquake-proof construction Proceedings of the Second World Conference on
of masonry dams," in Proceedings of the World Earthquake Engineering, vol I, Jul 11-18, 1960.
Engineering Conference, vol 9, 1929, p 275. Tokyo, Japan.
5. S. Okabe. "General theory of earth pres- 17. H. Matsuo and S. Ohara. "Lateral earth-
sure," Journal, Japanese Society of Civil quake pressure and stability of quay walls
Engineers, vol 12, 1926, p 1. during earthquakes," in Proceedings of the
6. K. Terzaghi. "Anchored bulkheads," Transac- Second World Conference on Earthquake Engineer-
tions of the American Society of Civil Engi- ing, vol II, Jul 11-18, 1960. Tokyo, Japan.
neers, vol 119, 1954, pp 1243-1280. 18. S. Kurata, H. Arai, and T. Yokoi. "On
7. P. Rowe. "Anchored sheet pile walls," in earthquake resistance of anchored sheet pile
Proceedings of the Institute of Civil Engineers, bulkheads," in Proceedings of the Third World
London, vol 1, Pt 1, Jan 1952, pp 27-70.
Conference on Earthquake Engineering, 1960. New 26. W.C. Teng. Foundation design. Prentice-
Zealand Hall International, 1962.
19. The Overseas Coastal Area Development 27. A.W. Skempton. "The bearing capacities of
Institute of Japan. Technical standards for clays," in Proceedings of the Building Research
port and harbour facilities in Japan, Tokyo, Congress, London, 1951.
Japan, 1980. 28. M.F. Tomlinson. "The adhesion of piles
20. U.S. Army Waterways Experiment Station. driven in clay soils," in Proceedings of the
Instruction Report K-81-2: User's Guide: Fourth International Conference on Soil Mechan-
Computer program for design and analysis of ics and Foundation Engineering, London, 1957.
sheet pile walls by classical methods (CSHTWAL) 29. Canadian Geotechnical Society. Canadian
Report I, Computation process, by William P. Foundation engineering manual, Canadian Geotech-
Dawkins, Vicksburg, Miss., Feb 1981. nical Society, 1978.
21. Department of the Army, Office of the Chief 30. L.C. Reese and H. Matlock. "Non-dimensional
of Engineers. Draft EM 1110-2-2906: Engineering solutions for laterally loaded piles with soil
and design of pile structures and foundations, modulus assumed proportional to depth," in
Washington, D.C., Nov 1970. Proceedings of the Eighth Texas Conference on

22. Naval Facilities Engineering Command. Soil Mechanics and Foundation Engineering,
NAVFAC Design Manual (DM) 25.4: Seawalls, Austin, Tex., American Society of Civil Engi-
bulkheads, and quaywalls, Alexandria, Va . neers, 1956.
Jul 1981. 31. C.K. Wang and C.G. Salmon. Reinforced
23. Naval Facilities Engineering Command. concrete design. Harper and Row, New York,
NAVFAC P-355: Seismic design for buildings, 1979.
Alexandria, Va ,
Feb 1982. 32. Naval Civil Engineering Laboratory. Tech-
24. U.S. Air Force Weapons Laboratory, Air nical Memorandum M51-85-01: A Rapid Seismic
Force Systems Command. AFWL-TR-74- 102 : The Air Analysis Procedure for Anchored Sheetpile Bulk-
Force manual for design and analysis of hardened heads, by T.K. Lew, Port Hueneme, CA. ,
March,
structures, by R.E. Crawford, C.J. Higgins and 1985.
E.M. Bultmann. Kirtland Air Force Base, N.M., 33. Leo Casagrande. "Comments on conventional
Oct 1974. design of retaining structures," American Soci-

25. Naval Civil Engineering Laboratory. Tech- ety of Civil Engineers Journal of the Soil
nical Report R642: Dynamic tests on selected Mechanics and Foundations Division. Feb 1973,
structural steel by W.L. Cowell, Port Hueneme SM2 , pp 181-198.

Calif., Sep 1969. (AD 695375)
250
|nterscc-t © ni
Figure 2. Location of the tie rod anchorage relative to active and

passive failure planes in order to he fully effective.
ACTIVE EARTH PRESSURE COEFFICIENTS
0ELTA/PHI * 0.500
PHI r 20.
PHI = 25.
PHI s 30.
PHI :
— — — PHI =
35.
10.
<
1C
0.00 0.10 0.20 0.30 0.10 0.50 0.60

E0 ACCEL C0EFF
Figure 3. Active lateral earth pressure coefficient (K ) versus earthquake
acceleration coefficient (k) for DELTA/P1II = 0.500 (from
Reference 3)
251
PASSIVE EARTH PRESSURE COEFFICIENTS
DELTA/PHI « - 0.500
PHI a 20.
PHI i 25.
PHI = 30.
PHI = 35 .
PHI s 50.
o.
EQ ACCEL COEFF
Figure A. Passive lateral earth pressure coefficient (K^) versus
earthquake acceleration coefficient (k) for DIEI.TA/ I’ll I
= 0.500
)
(from Reference 3).
ACTIVE FAILURE PLANE U/R TO HORIZONTAL
90 .
80.
DEITATPHI i 0.500
• PHI = 20.
TO. PHI = 25 .
PHI s 30.
PHI : 35 .
-PHI z 50.
60.
DEGREE)
( 50.
<
50.
30.
20 .
10 .
0 .
0.00 0.10 0.20 0.30 0.50 0.50 0 (0
EQ ACCEL COEFF
Figure 5. Active soil failure plane angle (Z^) versus earthquake
acceleration coefficient (k) for DEI.TA/PHI 0.50(1 I 1 1 uni
Re ference 3 )
252
PASSIVE FAILURE PLANE U/R TO HORIZONTAL
OELTA/PHI i - 0.500
PHI = 20.
PHI ; 25 .
PHI = 30.
PHI = 35
) - PHI j 5 0
.
DEGREE
0.00 0.10 0.20 0.30 0.10 0.50 0.(0

EQ ACCEL COEFF
Figure 6. Passive soil failure plane angle (Zp) versus earthquake
acceleration coefficient (k) for DELTA/P1II = 0.500 (from
Reference 3)
Figure 7. Lateral earth pressures and residual water pressure acting

on anchored sheet pile bulkhead.
253
Figure 8. Horizontal pressures on rigid wall from surface load (from ,
Reference 8)
bending moment in the
Figure 9. Virtual beam used for determining the
sheet piling.
= height of utoll anchoraê qLouo "the

Intffrrcction of ocfiutf Q nci passive
-fa 1 1 u re planes
hA - actual height of u>all anchorage
Figure 10. Effective height of wall anchorage when the intersection of

the active and passive soil failure planes is below the
ground surface.
255
a. Vertical p'lle ancliorage b. A - frame p > Ic anchorage
Figure 11. Location of pile anchorage relative to bulkhead for full

effectiveness (after Reference 19).
I ulr\ e-VG. Ap - jnclior- pull ( lb*)
Ea = acibw* fore Ob)

ET
p — pos5 1 -f ore. c ( /b>
nr rtia (orcc
T
w == t uall i ( lb)
fa x W *
K - €“a r ttitfu akc accelecatien coîieie.if

W n uJei^/rf' <=>L C/tl
-777S771=77.7-
T i t L J. J . i ii L r
Figure 12. External forces acting on wall anchorage for tie rod.
256
“1
_JL 1
Figure 13. Virtual earth pressure acting on short sheet pile anchorage
for tie rod.
Figure 14. Anchored sheet pile bulkhead analyzed as virtual gravity

type wall.
257
1
DESIGN OF PILES FOR SLOPES
Eirohiko Tada^, Kazuva Oshima^ , Yukio Sato"^^

III
and Ryoichi Abo
ABSTRACT In the model test, meanwhile, we have confirmed
that it is logical for the horizontal resis-

Generally there is no problem with the vertical tance of piles on sloping ground to be deter-
bearing capacity of piles provided on mountains mined by a calculating method taking the
and other slopes if they are driven into a sat- plasticization of the ground into consideration
isfactory bearing layer, but their horizontal and that the method to find the coefficient of
bearing capacity in the direction of the slope ground reaction (k-value) and plastic resis-
for the force of inertia in an earthquake, for tance used for this calculation is basically
example, is much smaller than in level ground. the same with soil cement ground as it is wT ith
Moreover, there is no established way of deter- sandy ground.
mining this horizontal bearing capacity and in
fact, no definite method of calculation is Key Words : Earthquake Engineering, Founda-
shown even in the Specifications for Highway tions, Pile Foundations, Lateral Loads.
Bridges (1)
1. INTRODUCTION
The Foundation Engineering Division, Public
Works Research Institute, has investigated the Recently, with the progress of road improve-
realities of sinso (deep bed method) pile ment, road structure foundations have come to
design and execution and conducted model tests be more frequently constructed on mountain
to see the effects of the ground slope on the slopes. Sinso piles are commonly used for
behavior of piles upon which horizontal force these foundations provided in sloping ground
works and establish a method to design them. for reasons concerning the execution of work,
Of these, the results of the model test with such as the relative freedom from the fear of
sandy ground has already been reported (2) . ground loosening by pit excavation, and the
This time, we will report on the results of our saving of the trouble of bringing and install-
fact-finding survey for design and execution ing heavy equipment. Sinso piling involves
and of our model test with soil cement ground. digging shafts into the ground while sheathing
with corrugated sheets or liner plates and
According to the fact-finding survey, the com- forming cast-in-place reinforced concrete piles
monly used design values for sinso piles are in the shafts.
2-4 m for pile diameter, 10-20 m for pile
length and about 5 for the pile length/pile When a horizontal force works on an ordinary
diameter. From the design classification, this long pile, the ground is gradually destroyed
places them in the category of limited-length from the surface and its displacement pro-
piles intermediate between caissons and long gresses with the increase of load and, gen-
piles erally, no definite yield point exists.
Therefore, the horizontal bearing capacity of
I) D. Eng.; Director of Structure and Bridge Department, Public Works Research Institute,
Ministry of Construction.
II) Head of Foundation Engineering Division, Structure and Bridge Department.
Ill) Member of Foundation Engineering Division.
259
the pile is set so that the pile-head displace- the Hokkaido Development Bureau, and the
ment cannot exceed the prescribed level. Okinawa General Bureau. The cases covered
include various types of superstructure, in
In the case of a highway bridge foundation, this ratios as indicated in Fig. 2.1 in terms of the
prescribed displacement is normally so small as number of places of executed work (total: 38)
to prevent large statically indeterminate and in Fig. 2.2 in terms of the number of sinso
stress on the superstructure, and to give nec- piles used (total: 230). These superstruc-
essary rigidity to the substructure itself. tures were mostly for bridges.
Here, since the ground can practically be
regarded as an elastic body, the stress, dis- We shall see hereunder about the bridge founda-
placement and ground reaction in different tions. The inclination angle of the ground was
parts of the pile can be determined by the the- as indicated in Fig. 2.3 and about 3/4 repre-
ory of beams on an elastic bed. sented sloping ground. A considerable number
of the foundations were constructed on steep
However, in the case of sinso piles (limited- slopes of more than 40°.
length piles) on sloping ground, it is consid-
ered necessary to check not only the pile-head The per-substructure number of piles used was
displacement but also the pile stability with as indicated in Fig. 2.4. About half of them
respect to horizontal forces because the hori- represented single-pile cases and about 1/3
zontal bearing power of the ground in the double-pile cases. The work was designed so
direction of the slope is small. It is also that the number of piles could be minimized by
deemed necessary to use a somewhat large design using large piles.
pile-head displacement by determining ground
reaction in the consideration of the ground Pile diameter D, pile length J!
, and pile length
plasticization because, in this case, the use ratio £/D are shown in Figs. 2.5 - 2.7.
of small design pile-head displacement is some-
times very uneconomical, as it is with ordinary The pile diameter was 2-4 m in more than 907 of
piles the cases while the pile length was 10-20 m in
more than 70%. The £/D was commonly about 5
However, all has not yet been thoroughly eluci- and, in about 90% of the cases, ranged from 3
dated on matters such as the effect of the to 9. From the design classification, most
ground slope on the behavior of piles on which piles belonged in the category of limited-
a horizontal force works and so the Foundation length piles — intermediate between caissons
Engineering Division, Public Works Research and long piles.
Institute, is conducting experiments and
research to clarify these issues and to estab- If horizontal resistance is expected In the
lish a method for the design of sinso piles for design, then the gap developing between the
sloping ground (3) strutting material and the natural ground must
be carefully backfilled. The backfill material
2 . REAL IT IES OF DESIGN AND CONSTRUCTION used was cement mortar with the amount of cement
OF SINSO PILES
and the flow value as shown In Figs. 2.8 and 2.9.
For backfilling mortar, importance is attached
A survey was made on sinso piles constructed not so much to strength as to the capacity t‘
during fiscal 1982 by the Ministry of Construc- fill the gap.
tion's construction offices throughout Japan,
260
T ype of
rn Type of
super-
- super
structure
structure
(Ratio
(Ratio
by
by place number Angle of incli-
of work) nation of
of piles)
ground
40 -
20 -
-
pi
O o
nn ha 0
Ini
' * ' '
10 20 30 40
'
TJ JZ
s £
CD CTi ^ I I I I
'
I
c 1- " ' '

10 20 30 40
cc £
Fig. 2.1 Fig. 2.2 Fig. 2.3
60 -
20 -
I I I I I
10 15 20 25
Fig. 2.4 Fig. 2.5 Fig. 2.6
Quantity of
cement in
grout mor-
3
tar (kgf/cm
i \ ) i i
3 5 7 9
Fig. 2.7 Fig. 2.8 Fig. 2.9
Fig. 2. Characteristics of cases surveyed
251
3. MODEL TEST Since the dimensional matrix rank of these nine
factors is 2, the physical quantity ratio
3.1 Purpose of Test between the model and the prototype (scale
factor) can be freely selected for only two
Fig. 3.1 shows one of the ordinary models used factors. Scale factors for the other factors
to calculate the behavior shown by sinso piles may be set so that seven independent non-
on sloping ground when a horizontal force works dimensional quantities tt - can be equal
1 7
on them. The ground constants used for this between the model and the prototype.
model are the lateral k-value kH , vertical
k-value £v ,
bottom rotation spring constant K^, From the results of the fact-finding survey. It
bottom shearing spring constant and the is considered appropriate to use 1/20 as the
plastic resistance of ground q. However, the scale factor for the length and the diameter of
primary purpose of the test was to clarify k sinso piles in general, providing that their
H
and q which are presumed to profoundly affect length and diameter are 10 m and 2 m, respec-
the horizontal bearing capacity of piles. tively. Also, if the scale factor of y is 1,
the scale factor of other factors can be set as
indicated in Table 3.1.
Ty77y7T77.
-AAAA-
-A/W- Table 3.1 Scale Factor of Factors
^WV-
-A/W'-
Ground Non-dimensional
reaction P = ( ku \yKq Factor Scale factor
kr =
Here, lg
¥b is effective geometrical
quantity
Z 1/10
moment of inertia
y 1
VW —
-/ V
*i = t/D D 1/10
tt
2 = P/yZ 3 p 1/10*
Tt,
*4
=
=
E
p
E 0 /y£
I/Yt
S
V
Eo
1
1/10
/10
s
TT
s
= k /Z k 1
Fig. 3.1 Model of Pile ”6 = 1
tt 7 = c/yZ c 1/10
= &Z Bi
3.2 Test Conditions TT
8 1
When making a model test, it is necessary to

satisfy the condition of similarity between the Selecting soft-rock ground as the object ground
prototype and the model. for the prototype, the physical quantities of
prototypes and models of pile and ground can be
The factors to be considered when studying the set as in Table 3.2, which also lists the
law of similarity concerning the horizontal values of non-dimensional quantities 8*t

bearing capacity of piles would probably obtained by dimensionally analyzing the gov-
include horizontal load P (kgf) , flexural erning equation based on the assumption that
2
rigidity of pile E I (kgf "cm ) , modulus of piles are beams on an elastic bed. Here, t Is
P 2
deformation of ground E (kgf/cm ) , k-value k called the characteristic value of the pile"
°
3 4
(kgf/cm ), pile length L (cm), pile diameter and 6 = (k-D)/(4-E -I).
P
3
D (cm), unit weight of ground Y (kgf/cm ),
shearing resistance angle of ground <f>, and The model ground, from the target value of F ,
2
cohesion of ground c (kgf/cm ). corresponds to loose sandy ground but was
252
Table 3.2 Physical Quantities of Ground
and Piles
Physical quantity Prototype Model

2
Eo (kgf /cm ) 1500 150
3
k (kgf /cm ) 10 10
3
Y (kgf /cm ) 1 . 7 1.7
<t>
40° 40°
2
C (kgf /cm ) 3 0.3
i (cm) 1000 100
D (cm) 200 20
2 13 s
Ep • I (kgf -cm ) 2. 12 x io 2. 12 x io
B-l 2.2 2.2 Fig. 3.2 Equipment of Experiment
Cracks developed on the ground surface in the

loading process were noted and underground
prepared, using lean-mix soil cement, to pro- sliding surfaces were examined by digging the
vide ground approximating soft rock and liable ground after loading.
to fragile destruction.
The test cases are shown in Table 3.3. In all
From the target value of E I, a vinyl chloride test cases, data for model piles were: outside
P
pipe is suitable for use as a model pile. How- diameter 21.6 cm, pipe thickness 0.4 cm, and
ever, a steel pipe pile was used instead, for setting length 100 cm.
the purposes of approximating the prototype (?•£,
the main factor governing the behavior of the Table 3.3 Test Cases
limited-length pile for horizontal force, so as
to not cause yielding of the pile body prior to Angle of inclination
Type of ground of ground (deg.)
the destruction of the soil cement ground, and
also to improve precision in the model manufac-

Sand (S) 0, 10, 20, 30, 40
ture and stress measurement.
Soil cement (R 1 - R6)
C/S = 1, 2, 3, 4, 6% 0, 20, 30, 40
3.3 Test Method
Note) Figure for R represents C/S (%)

In the test, a model pile was installed in the
ground prepared in a test soil tank, as in the 3.4 Test Results

case of sandy ground, and a horizontal force was
exerted against the pile head, using a hydraulic 3.4.1 Ground Constants
jack and a wire rope (see Fig. 3.2)
Unconfined compression tests were conducted
Loading was made by the 4-cycle formula using using compacted samples of the soil cement used
load control with 1/16 of the design maximum to produce the model ground, or samples cut out
load as the load stage. The load, the dis- of the model ground and shaped after the load-
placement on the pile head and the ground ing process.
surface, and the bending strain in the pile
body were measured by, respectively, load cell, There were considerable variations in the test
displacement gauge and strain gauge. results but it is deemed that the unconfined
compression strength qu and E^ of the model
ground prepared were generally within the range The ultimate horizontal bearing capacity tended
in Table 3.4, according to the ratio c/S of the to decrease with the increase of the inclina-
soil cement. tion angle of the ground. This is a sign that
it is necessary for sinso piles in sloped
Table 3.4 Physical Constants of Test Ground ground to be checked for stability of horizon-
tal force.
Type of ground qu E
°
c
2 2 2
C/S (7.) (kgf/cm ) (kgf/cm ) (kgf/cm ) (deg.
1 0.8 400 0.19 40 3.4.3 Inversely Calculated k-Value

2 5.7 1850 1.33 40
3 1.6 950 0.37 40

4 4.3 1300 1 .00 40
Assuming the piles to be beams on an elastic
6 10.9 3150 2.54 40 bed, the fundamental equation for their stress
and displacement can be shown by Equation (3.1)
Note) c was calculated from qu by the following expression
on the assumption that = 40°. <j>
1 - sin <p
c •qu d^y 1
2 • cos if
+ = 0
kI) y 3
dx
4 rr
p
c - 1)
3.4.2 Horizontal Bearing Capacity
where
In the test, loading was continued until either
y = Horizontal displacement of pile (cm)
the ground in front of the pile was destroyed
x = depth in direction of pile axis from
or the stress of the pile body approached the
ground surface (cm)
yield point. It was deemed that the horizontal
force reached its ultimate when, in the load vs.

The stress and displacement of the limited-
pile-head displacement curves thus obtained,
length pile can be obtained by solving Equa-
load decreased or almost ceased to increase as
tion (3.1) under the boundary conditions at the
displacement increased.
pile head and the pile tip.
Test cases where ultimate horizontal bearing cap-

From the measured values of horizontal force
acity was obtained are as indicated in
and pile-head displacement, the k-value,
Fig. 3.3.
assumed to be constant with depth, can be
obtained by the solution of Equation (3.1). An
O R3
A example of the relation between the inversely
* R 2
O
A R 1
calculated k-value and the pile displacement at
the ground-surface (D, R6) is shown in Fig. 3.4
A
O
A
A A
ill 1 1 1
211 .in 4n
Angle of inclination of ground (deg.)
Fig. 3.3 Horizontal Bearing Capacity vs. Pile displacement on ground surface (cm)
Angle of Inclination
Fig. 3.4 Relation Between Inversely Calculated
k-Value and Pile Displacement
264
In this example, the k-value decreases with the 3.4.4 Distribution of Ground Reaction
increase of pile displacement. Other cases are and Distribution of k-Value
omitted for reasons of space but are much the
same as this example. In Equation (3.2), show- The bending moment in the pile section can be
ing the decrease of this k-value, n is gener- obtained from the measured value of bending
ally within the range of 0.1 to 0.6 and strain in the pile body by Equation (3.3).
indicates the nonlinearity of the k-value based
on the inelastic behavior of the ground. E r
E -E-
M = (3.3)
k = k )' n (3.2)
(y/y o
o
where
where M Bending moment (kgf*cm)
k = k-value at time when pile displacement e Bending strain
3 2
is y cm (kgf/cm ) E Young's modulus of pile (kgf/cm )
P
k = k-value at time when pile displacement r Distance from neutral axis to point of
o
3
is y cm (kgf/cm ) measurement (cm)
I Geometrical moment of inertia of pile
, 4
The relation between the inversely calculated (cm )
k-value and the inclination angle of the ground

is shown in Fig. 3.5. This k-value is the value Assuming piles to beams on an elastic bed, the
relation of pile displacement and ground reac-
tion with the bending moment can be shown by
Equation (3.4) .
E I = -M
P 2
dx
(3.4)
O H ’ •
• H 3
A H - • o
* HI
E I -P
a S
P
kQ : k-Value of level ground
Angle of inclination of ground (deg.)

where p = Ground reaction (kgf/cm^)
If M is approximated by a polynomial in x by
Fig. 3.5 Inversely calculated k-value vs.
Angle of Inclination the method of least squares, from the distribu-
tion of M at different points of measurement,

at the time when the pile displacement is the distribution of p can be obtained by Equa-
0.2 cm. In the case of soil cement ground, tion (3.4). Also, the distribution of k-value
ground strength varies widely, but there is the can be obtained by Equation (3.5) from the dis-
general tendency that the k-value deceases with tributions of y and p.
the increase of the inclination angle of the
ground k = p/y (3.5)
Examples of the distribution of ground reaction

and the distribution of k-value ( S— 0 , 30, R6-0,
265
Ground reaction (kgt/cm J )
Ground reaction (kgf/cm J
)
P=450kgf
Ground reaction (kgf/cm 2 )

Ground reaction (kgf/cm J )
10 •
0 10 5 0 10
l
KX)-*- 100 J-
Fig. 3.6 Distribution of Ground Reaction (No. 1)
Fig. 3.7 Distribution of k-Values
266
Ground reaction (kgf/cm 2 ) Ground reaction (kgf/cm 3 )
o CM
CM
100 kgf
200 kgf
300 kgf
400 kgf
500 kgf
600 kgf
700 kgf
800 kgf
850 kgf
900 kgf
Dep. (cm)
Ground reaction (kgf/cm 3 ) Ground reaction (kgf/cm 3 )
-10
o o o
Fig. 3.8 Distribution of Ground Reaction (No. 2)
26
30) are shown in Fig. 3.6 and Fig. 3.7. These 3.4.5 Inversely Calculated k-Value
are the results of approximation of M by the (Trapezoidal Distribution)
6th-degree polynomial in x. Mention of the
other cases is omitted for reason of space. The distribution of the k-value becomes greater
with the increase of depth, as seen in 3.4.4.

The center of rotation of the pile (the point But here, let us simplify it and assume it to
at which the ground reaction is 0) is deeper in be a trapezoidal distribution, as shown in
sloping ground than in level ground and its Fig. 3.9. Since, in this case, i in the
K
depth increases with increase of the load. fundamental equation (3.1) for the stress and
This indicates that the relative rigidity of displacement of the pile is a function of x,
the pile increases due to the decrease of the the solution cannot be easily obtained. How-
k-value by the slope of the ground in front of ever, if each k-value at n-points of equal
the pile, the decrease of the k-value by non- division of the pile are given, it is possible
linearity, and the apparent decrease of the to solve this numerically by the method of
k-value by the plasticization of the ground in redundant reaction.

front of the pile near the ground surface; and
the pile has assumed a behavior close to that
of a caisson.
The position at which the maximum ground reac-

tion of the maximum k-value occurs is deeper in
sloping ground than in level ground. This
indicates that the k-value is generally small
near the ground surface and increases with
depth because the ground is unrestrained on its
surface. However, in the case of sloping
ground, the k-value decreases further because
the ground in front of the pile is limited in
its horizontal direction near the ground
surface
The ground reaction near the ground surface Fig. 3.9 Distribution Shape of k-Values
tends to decrease with the increase of load in

the case of sloping ground. This seems to First, k is obtained inversely from the mc.i-
max
indicate that the ground is plasticized from sured values of the horizontal force and pile-
the ground surface to a certain depth and head displacement for the cases with 0, 20, 40
ground reaction for this extent does not and 60 cm as values of £ in Fig. 3.9. Out of
K
increase with the increase of load. This is the different cases of £ k obtained,
and
icmax
more manifest in Fig. 3.8 which shows the dis- the case where the calculated value of the
tribution of the ground reaction obtained by bending moment distribution of the pile agree'
geometrically differentiating measured values well with its measured value Is used for the
of bending. inversely calculated k-value (distribution).
An example of the relation bstwssn t lie calcu-
lated value and measured value of bending
268
moment in the pile thus obtained is shown in 2. The k-value is small near the ground sur-
Fig. 3.10. It shows £ of the inversely calcu- face and increases with depth.
lated k-value to be larger in sloping ground 3. With the increase of load, the ground is
than in level ground. The values in agreement plasticized from the surface to a certain
with the experimental value for R6, for depth.
instance, are in the vicinity of £^ = 60 cm in

sloping ground (30°), compared with the vicin- We shall discuss more about the quantitative
ity of only £^ = 20 cm for level ground. This evaluation of design constants hereunder.
is similar to the tendency in the position of
occurrence of maximum k-value in the k-value Let us first see about the k-value distribu-
distribution in 3.4.4. tion. Assuming that the k-value distribution

is trapezoidal, as indicated in Fig. 3.9, it is
possible to have a numerical solution from
load vs. pile-head displacement, as described
in 3.4.5. The ratio between the inversely cal-
Moment (kgf*cm) Moment (kgf-cm)
0 20000 0 100000 200000
culated k-value in 3.4.3 and k obtained from
max
S-0 °r R6-0 the load vs. pile-head displacement falls with
xC*
_ /* = 0 cm o a x
>04* aUala> ,
*
=20cm 0 the increase of £ /D, as indicated in Fig. 4.1.
O' C* lated 8
<X
• , =40cm 0
value
o
a * =60cm
O
o
X
x a
A «
From both Fig. 4.1 and Fig. 3.5, it
x
g 50 -
0>& %
cm e*
1 50 [*
Experimental
value
°
o
“ xa**
o a x» *
O A 111
Obu X k/kmax
P„-400kgf P«=6890kgf
jqq
1
^ S— 30 0
R6-30
OCMX
P«=300kgf P H =4973kgf
100 100
Fig. 3.10 Comparison of Pile Bearing

Moment Distribution
4 . DESIGN P OLICY FOR SINSO PILES

ON SLOPED GROUND
Fig. 4.1 Ratio of Inversely Calculated

4 . 1 Design Constants k-Value and k
max
What was known about constants from the model is probably reasonable for £^ to be obtained by
test is consolidated below. Equation (4.1) according to the inclination

angle of the ground, 0.
1. The inversely calculated k-value decreases
with increase of the inclination angle of £ /D = a*0 (4.1)
K
the ground.
269
where where
a = Coefficient of k-value decrease by Q = Plastic resistance of ground (kgf)
P
3
slope of ground. Believed to be V = Volume of soil wedge (cm )
2
approximately a = 0.05 - 0.1. A = Bottom area of soil wedge (cm )
0 = Angle of inclination of ground (deg.) V = Angle formed by sliding surface with

horizontal plane
Let us next see about the plastic resistance of 4 = Spread angle of sliding surface
the ground. For cases where the ground surface z = Depth from ground surface to sliding
is sloping and the ground is highly cohesive, surface (cm)
as it is with soft-rock ground, it seems to be
advisable to obtain the plastic resistance by Fig. 4.3 shows the plastic resistance of ground
Coulomb's theory of earth pressure, where the obtained by Equation (4.2) from the physical
equilibrium of forces of the soil body held constant of the model ground, using 4 1 = 40° and
2
between the pile and the sliding surface (soil c = 1.27 kgf/cm . Since the plastic resistance
wedge) is considered. Here, let us consider of ground in Fig. 4.3 is quite close to the
the equilibrium of force indicated in Fig. 4.2, ground reaction distribution in Fig. 3.8, it
assuming the sliding surface to be plane and seems to be reasonable for plastic resistance
of the ground to be obtained by expression
(4.2).
Plastic resistance (kgf/cm 1 )
Q,
r -V
<P
Depth (cm)
Fig. 4.2 Equilibrium of Forces of Soil Wedge
ignoring resistance on the sides of the soil

Fig. 4.3 Calculated Value of Plastic
wedge and friction between the pile and the Resistance of Ground
ground. In this case, plastic resistance

can be obtained by Equation (4.2) providing Here, the cohesion of the model ground is esti-
that the angle of the sliding surface ^ is the mated at 1/2 in consideration of such factors
angle at which Q is minimal. as the variation of ground. Considering also
the great impact of the value of c on calcu-
Y’V + c’A’sin V lated by expression (4.2), the ground cohesion

Q sin (V + <f>) + c*A*cos. V
P cos (V + <f>)
must be estimated carefully.
,D cos 0*tan 4 z N cos 0*cos V 2

+ Z
(
2 sin (6 + V)
*
3 sin (6 + V)' 4.2 Design Method
(4.2)
cos 0»tan 4 cos 6
A = (D + z) 2 Basically, the elasto-plast ic ground reaction
sin (9 + *0 sin (0 + V)
method must be used for soft-rock ground, too,
270
Horizontal force (tf)
to design sinso piles on sloping ground. For

8
a
analysis, "the Pile and Caisson Calculating a
Program Taking the Ground Plasticization into Q
6
Consideration," (4) prepared by the Foundation

a
Engineering Division, Public Works Research 4 a
Institute, is used. E x Experimental value
Q
2 Calculated value
Q
The characteristics of this program are as '
Gk
follows
0 2 4
Pile-head displacement (mm)
1. It assumes the use of a limited-length
elastic body as a pile.

Horizontal force (tf)
2. It assumes the use of a completely elastic
body as ground. In other words, the rela- x

6
tion between ground reaction force and dis- x
X
placement is bilinear.
O
3. The pile and the ground are subdivided in
x Experimental value
the depth direction. Therefore, the pile
2
data and ground constants can be changed Calculated value
x
with depth. Also, a ground constant can be
changed rectilinearly even within each of 0 2 4 6 8
Pile-head displacement (mm)

the layers obtained by subdividing.
4. The calculating method used combines the

characteristic matrix method and the method
Horizontal force (if)
of redundant reaction.
x
4
x
6. x
x
Fig. 4.4 compares the experimental values of Lk
2 •
0
horizontal force vs. pile-head displacement in 0
x Experimental value
a
Calculated value
soft-rock ground and the values calculated by a
0 2 4 6
this program. The values by elasto-plastic
Pile head displacement (mm)
calculation agree well with the experimental

values. This indicates that if design con- Fig. 4.4 Comparison of Horizontal Force vs.
Pile-head Displacement
stants are properly obtained, the behavior of
piles on sloping ground can be satisfactorily
design constants and the applicability of
analyzed by this program.
elasto-plastic calculation in accordance with
the results of real-pile tests, and establish a
5 . CONCLUSIONS
method to design piles for sloping ground.
It has been confirmed that the behavior of

BIBLIOGRAPHY
piles on sloping ground can be elucidated by a
model test and analyzed well by elasto-plastic
(1) Japan Road Association: Specifications
calculation
for Highway Bridges, IV: Substructure,
1980.
In the future, we are planning to continue ana-
lytic studies on the quantitative evaluation of
271
(2)
Misato Murakami, Kazuya Oshima, Yukio Sato (3) Hideya Asanuma et al.: A Study on Pile
and Ryoichi Abo: Experiments and Research Foundations in Slopes, Public Works
on Lateral Resistance of Piles in Slopes, Research Institute Technical Memorandum
Sixteenth Joint Meeting, U.S. -Japan Panel No. 1949, February 1983.
on Wind and Seismic Effects, 1934. (4) Yukitake Shioi et al.: Estimation Scheme
for Inversely Calculated k-Values in Large
Diameter Piles, Public Works Research
Institute Technical Memorandum No. 1817,
September 1981.
272
DYNAMIC STRESSES AND DISPLACEMENTS IN A BURIED TUNNEL
S. K. Datta
K. C. Wong
Engineering
Department of Mechanical
University of Colorado, Boulder, CO 80309
A. H. Shah
Department of Civil Engineering
University of Manitoba, Winnipeg, Canada R3T 2N2
ABSTRACT 2. GOVERNING EQUATIONS AND SOLUTION
Dynamic three-dimensional response of a tunnel It is assumed that the tunnel is infinite in

of noncircular shape in an infinite elastic length and is perfectly bonded with the
medium has been studied in this paper. Also surrounding ground, which is assumed to be
studied is the two-dimensional (plane strain) linearly elastic, isotropic and homogeneous.
response of a tunnel buried in a semi-infinite Assuming harmonic time dependence, the displace-
elastic medium. In both cases it is found that ment u(x,y ,z)e""' satisfies the equation
significant dynamic amplification of stresses
y
a v2 u + (x a + y a V V • u = - a 2 u (1)
) P a)
and displacements occurs in a tunnel embedded in
a soft ground when the wavelength is long. In where A, y are Lame constants and p the density
addition, in the latter case the peak values are of the material. Superscript a takes the values
significantly influenced by the depth of embed- I or II depending on whether the point is in the
ment . surrounding ground or in the tunnel wall. Note
that the region inside the tunnel is empty.
1. INTRODUCTION Thus the boundary conditions are
Dynamic three-dimensional response of a tunnel

x
H. n. - 0 on S, 2
embedded in an elastic medium is a subject of ij J
( )
interest for proper earthquake resistant design.

I II
u = u
Although this subject has received some atten-
• on S, (3)
tion, no careful analysis is available that 11
t!. n. = x n.
TJ J 1 J J
takes into account the interaction of the tunnel
with its surrounding ground. Our recent studies S\ and S 2 are the inner and outer surfaces of
[1-14] of seismic response of pipelines have the tunnel, x-jj is the stress tensor and n is
shown the importance of the interaction effects. the unit normal vector to or S 2 . The
Here we solve by a combined finite element geometry of a tunnel of general cross section is
method and eigenfunction expansions the full shown in Fig. la and the end view of the tunnel
three dimensional equations of motion of the considered here is shown in Fig. lb.
tunnel and the surrounding infinite ground. The
problem gets considerably more complex when the We will consider seismic waves propagating along
tunnel the length of the tunnel with speed c = u>/e,

is embedded in a semi-infinite medium.
For the sake of simplifying the computations we where A(=2ir/e) is its wavelength. Then it can
be shown that the general solution to (1) is
then consider the two-dimensional (plane strain)
case and show that the depth of embedment plays given by
an important role in determining the dynamic u = v <j>

+ v ~ ( \|i e )
+ ?~(xe 2 (4)
)
amp! i f ication
273
with where K n is the modified Bessel function of the
= f(x,y) ei (ez-mt) second kind and

<t>
'P
= 93(x»y) e] ( eZ-a)t
1 (ez-ust) (5)
x = 9 i(x,y) e
K2 = u)/C2
Here C2 is the shear wave speed in the material.
The equations satisfied by f, and g 3 are The displacements and stresses in Rj can be
gj_,
obtained by substituting Eqs. (8) to (10) in (5)
2 2 2 and (4) and using the stress - strain relations.

V f + (k - e )f = 0 (6)
Details of these expressions can be found in
[16].
The interior region R2 (including the tunnel) is
divided into finite elements having Nj interior

nodes and Ng boundary nodes on B. The displace-
ment field within each element is expressed in
terms of the shape functions Nj and the nodal
0
displacements q.
and ki = w/ci, ci being the longitudinal wave
speed in the material.
e
u = [N] (11)
| ) (qj )
If the cross section of the tunnel is circular,
then Eqs. (6) and (7) can be solved by separa-
where the superscript stands for an element and
tion of variables in terms of Bessel functions
the subscript j for the jth node of the element.
for each circumferential mode number n.
The strain within an element is given by a 6x1
However, for noncircular cross sections this is
matrix as
not useful. But the eigenfunction solution in
polar coordinates can be used outside a circle

of radius Rg
nel.
(Fig.
The solution in the interior of this

lb) that surrounds the tun-
[f
- [L] [N]
j
qj ) [B1
1 ?! 1
(12)
circle (B) is obtained by a finite element

where [L] is an operator matrix given by
method and matched with the exterior solution on
B by using a variational principle. This is an "
“a/3x 0 0
extension to three dimensions of the method used 0 3/3y 0
for two dimensions in [5, 10, 15]. 0 0 i e
[L] = 0 i e 3/ 3y (13)
In the exterior region Rj the solutions f, gj,
i e 0 3/ 3x
and g3 are
_3/3x 3/3y 0
in6
f = l C K Y r) e (8)
(
n n
n=-oo
100
9 = l c K
(
6r ) e ( 9 )
X i n n
n=-oo
«3
’ I
n = -oo
C
3n
K
n
^ (10 >
274
Note that in writing (12) and (13) it has been ( s
v
where q is found from (4) by using (5) and
~X
assumed that the displacement of the jth node is B
e i ez ( 8 )
and q is the incident seismic disturbance
~X
6 B
Sj
assumed to have originated outside B.
and the shape functions are independent of z.
Substituting (17) in (15) and taking variation,
The energy functional F can be written as we obtain
+
F
’II
\ + q
Xj
S
IB
% (14)
II
S T
d
IB
G
xy
*T *T
+ S„ T G S__
RI q
Xj
% S
BB
% xy BI xy BB
- S
in which q = q and qX = q . The super- IB ' a
~X ~ B
'll I B ~ B
scripts * and T designate complex conjugate and ,*T ,( i ,(D (18)

"xy BB TB
transpose, respectively. The elemental impe-
dance matrices are defined as
Flere pg^ is the interaction force acting at the

T T
[S
e
] = // {[B*] [D][B] - P J [N] [N]} dxdy (15) nodes on B on the material in R2 due to that in
R
e R l* Matrix [G X y] is defined by the equation
for an isotropic material [D] is given by
S>
[G {a} (19)
( q(
~X
l
xy ]
B
[D] = (16)
T
{a} jc
|
c c c c c
- 1 n 31
...
N 1n 3N
Note that in writing (18) the series appearing

in Eqs. ( 8) to (10) have been summed from
where =
n " + 1 to ancl we ^ ave wr tten "’
\ ^B \ ^B
D. = X + Z\s ,
D0 = X ,
= p
1 e e ’
2 e 3 e
c
i
1
' c
. I n + i
,c li _c l - — n + i
’
The continuity of the displacement on B implies 2 B ’

2 B
that
= =
C
3i
C
3» " N + 1
^ l * C ° n_
7 B
(17) sistent with (19) it can be shown that, by using

the principle of virtual work.
275
to setting e = 0 in Eqs. (5) to (7) and x = 0 in
m fa) + { B
(1)
}
( 20 ) Eq. (4). The corresponding solution has
recently been discussed in [10] and some of the
numerical results presented there for this case
where [R] and are approximated by
are discussed in the following section.
T
DO R
b
48 [G*
0
] [F
r9 ],
(21)
3. NUMERICAL RESULTS AND DISCUSSION
1
1
- R 48 [G* 0 ] fa* '). (22) Numerical results are first presented for the
Ip*' )
b
B
case of a tunnel embedded in an infinite medium.
In writing (21) and (22) it has been assumed It is assumed that the tunnel is excited by
that the circle B is divided into N equidistant plane P, SV and SH waves propagating in the yz-
points. The expressions for [G r q] and [F r e ] can plane. The potentials <j>, x> and (Eq. 4) for
be found in [16], these three incident waves are given by:
Substituting (20) in (18) and transferring all (1) P wave

the terms contained in {a} to the left-hand
(i)
= e
1 <“ + ’V H“ t
, *
(i)
= 0.
side, we get the following equation to solve for
(2) SV wave
{q x j} and {a} simultaneously.
(i ) _ ^2_ i (ez+S.yH u>t

= 0.
x
i6
x
S r T
S T qG
II I B xy ( qx )
(3) SH wave
*T *T 1
G S DT G S G - R
|_xy BI xy ddBB xy _ { a ) k
!23' i ) - 2 i ( EZ+6,y)-i uit
v ~ = 0.
"S i5j
IB
Note that if 6o is the angle made by the inci-
*T
%
^
-G *S„ n ^
+ n '
dent wave propagation direction with the y-axis
xy BB ?B
D
then e = k j s i n q for the P wave and e = kpsinOg
for the S wave. Thus y\ = kcos0g and
Once the interior nodal displacements {q X }
are
~
I Si = k2cos0g.
known they can be used in (12) to calculate the
The material making up the tunnel is assumed to
strains, which in turn can be used to calculate
be concrete with the properties
the stresses. The numerical results thus
obtained are discussed in the next section.
p = mass density = 2.24 kg/m^
E = Young's modulus = 1.6xl0^kN/m^

In the above derivation it was assumed that the
a = Poisson's ratio = 0.2
surrounding ground was infinite in extent. If
the ground is semi -i nf inite and the tunnel is

Two types of ground materials were considered.
buried in it at a depth H (Fig. 2), then the
The properties are given in table 1.
representations (8) to (10) have to be modified
to include the free surface effects. This Different incidence angles were considered. For
causes considerable algebraic complications. each angle of incidence axial and hoop stresses
The analysis is somewhat simplified if it is were computed at the outer surface of the tunnel
assumed that the waves are moving perpendicular wall and were normalized with respect to the
to the length of the tunnel and there is no maximum stress associated with the incident
displacement along the length. This corresponds wave. In Figs. 3 and 4 maximum normalized axial
276
and hoop stresses for incident P wave are shown plane P wave moving at an angle of 40° with the
against the incident angles. Here case I refers negative y-axis. It is seen that the hoop
to tunnel in soft soil and case II that in rock. stresses increase with frequency at small
It is seen that in case I the hoop stress is depths, but increase initially with frequency
nearly constant and is always greater than the and then decrease for large depths. It is also
axial stress. In case II, however, the hoop seen that larger stresses occur at smaller
stress decreases rapidly with increasing 9g and depths. For comparison purposes the results for
becomes less than the axial stress when the wave SV wave for tunnels in rock and soft soil are
is moving nearly along the axis of the tunnel. shown in Fig. 13. Much larger stresses are
Also, larger stresses are induced in case I than found to occur in case I than in case II. Also,
in case II. Fig. 5 shows the comparison of the the variation of stresses along the ci rcum-
hoop stresses in cases I and II at two different ference is significantly different in the two
frequencies. Here the incidence angle is 5°. cases. This was found to be true for the P wave
Fig. 6 shows the corresponding results for 0q = also. Variation of hoop stresses in a tunnel
85°. with the changes in the angle of incidence of SV
wave is shown in Fig. 14. Large variations
We then show in Figs. 7 and 8 the maximum axial
are found to occur with different angles of
and hoop stresses for incident SV wave. It is
incidence and these are more pronounced in case
found that much larger axial stresses are
II than in case I. For P wave it was found that
induced in case I than case II. Also axial
in case II results were not too sensitive to
stresses are much larger than the hoop stresses
changes in the angle of incidence. Finally,
in case I, whereas they are nearly the same in
Fig. 15 shows the variation of hoop stress with
case II. Figs. 9 and 10 show the results for SH
frequency for an incident Rayleigh wave moving
wave. Here it is found that maximum hoop stress
along the positive x-axis. This figure is to be
decreases with increasing 6q both in case I and
contrasted with Fig. 12, where incident P wave
case II. On the other hand, axial stress
is considered. Very large stress concentrations
increases with 0g in case I but decreases in
are caused by Rayleigh waves near the corners of
case II.
the base of the tunnel , more so on the shadow
side than on the illuminated side. This was
Comparison of the results for the three types of
found also for SV waves (Fig. 14).
incident waves shows that when the waves are
moving nearly along the axis, maximum stress is
4. CONCLUSION
induced by a P wave. At an intermediate angle,
however, SV wave induced the largest axial In this paper we have presented a combined
stresses. At nearly vertical incidence, it is finite element and eigenfunction expansion tech-
the SH wave that causes the largest hoop stress. nique to solve the problem of seismic response
of a noncircular shaped tunnel. Three-
In the next set of figures we show the hoop
dimensional dynamic response analysis of the
stresses induced in a tunnel embedded in a semi-
tunnel in an infinite medium shows that large
infinite elastic medium. As mentioned before,
hoop stresses develop in the tunnel wall when
we considered now the plane strain case (e=0).
the surrounding ground is soft. In fact, these
Plane P, SV, and Rayleigh wave incidences were
are always larger than the axial stresses for
considered. The material properties of the tun-
incident P waves. It is also found that the SV
nel and the ground were the same as described
waves induce the largest axial stresses in the
before. Figs. 11 and 12 show the results for a
wall when the surrounding ground is soft. The
277
effect of depth of embedment of the tunnel in a 4. S. K. Datta, T. Chakraborty, and A. H. Shah,
half-space is found to be quite significant. "Dynamic Response of Pipelines to Moving Loads,"
For the sake of simplifying our calculations we Proceedings of the Eighth World Conference on
have considered only the plane strain problem in Earthquake Engineering , July 21-28,
this case. San Francisco, CA, Vol . VII, pp. 295-302
This study suggests that future research should 5. S. K. Datta, A. H. Shah, and K. C. Wong,
take into account the effect of layering in a "Dynamic Stresses and Displacements in Buried
half-space on the dynamic three-dimensional Pipe," ASCE Journal of Engineering Mechanics ,
response of buried tunnels. It appears that the 110 (1984), pp. 1451-1466.
presence of soft layers overlying hard substrate
6. K. C. Wong, A. H. Shah, S. K. Datta, and P.
may significantly influence the response.
M. O'Leary, "Dynamic Amplification of
5. ACKNOWLEDGMENTS Displacements and Stresses around Buried
Pipelines and Tunnels," in Earthquake Source
The work reported here was supported in part by
Modeling, Ground Motion, and Structural
a grant (CEE 81-20536) from the National Science
Response , ed. S. K. Datta, AMD-Vol . 10,
Foundation under the Earthquake Hazards
PVP-Vol. 80, The American Society of Mechanical
Mitigation Program. Partial support was
Engineers, New York, 1984, pp. 147-161.
received also from the Natural Science and the
Engineering Research Council of Canada by grant 7. K. C. Wong, A. H. Shah, and S. K. Datta,
A-7988. "Dynamic Amplification of Stresses and
Displacements Induced in a Buried Tunnel,"
6. REFERENCES
CUMER 84-1, January 1984, Department of
Mechanical Engineering, University of Colorado,
1. S. K. Datta, A. H. Shah, and T. Chakraborty,
Boulder.
"Dynamic Response of Pipelines to Moving Loads,"
in Earthquake Behavior and Safety of Oil and Gas
8. P. M. O'Leary and S. K. Datta, "Dynamic
Storage Facilities, Buried Pipelines, and
Response of Buried Pipeline at Low Frequencies,"
Equipment , ed. T. Ariman, PVP-Vol . 77, ASME
Journal of Pressure Vessel Technology,
New York, 1983, pp. 246-253.
Transactions of ASME, 107 (1985), pp. 44-50.
2. S. K. Datta, A. H. Shah, and K. C. Wong,

9. S. K. Datta, P. M. O'Leary, and A. H. Shah,
"Dynamic Amplification of Stresses and
"Dynamic Response of Buried Pipelines to
Displacements Induced in a Buried Pipe in a
Incident Longitudinal and Shear Waves," CUMER
Semi -i nf i nite Medium," CUMER 83-3, Department of
84-3, January 1984, Department of Mechanical
Mechanical Engineering, University of Colorado,
Engineering, University of Colorado, Boulder.
Boulder.
(Also to appear in Journal of Applied
Mechanics .
3. S. K. Datta, T. Chakraborty, and A. H. Shah,
"Dynamic Response of Pipelines to Moving Loads,"
10. K. C. Wong, A. H. Shah, and S. K. Datta,
Earthquake Engineering and Structural Dynamics ,
"Dynamics Stresses and Displacements in a Buried
12 (1984), pp. 59-72.
Tunnel " ASCE Journal of Engineering Mechanics ,
111 , (1985), pp. 218-234.
278
11. P. M. O'Leary and S. K. Datta, "Dynamics of
Buried Pipelines," to appear in the Advances in
Underground Pipeline Engineering , ASCE, 1985.
12. P. M. O'Leary and S. K. Datta, "Dynamics of
Buried Pipelines," to appear in the

International Journal of Soil Dynamics and
Earthquake Engineering , 1985.
13. K. C. Wong, A. H. Shah, and S. K. Datta,

"Dynamics of Underground Pipelines and Tunnels,"
to be presented at the 1985 PVP Summer
Conference, New Orleans, LA, June 23-27, 1985.
Also to appear in the Proceedings.
14. S. K. Datta, A. H. Shah, and P. M. O'Leary,

"Seismic Response Analysis of Embedded
Pipelines," to be presented at the 2nd
International Conference on Soil Dynamics and
Earthquake Engineering, June 28-July 3, 1985.
Also to appear in the Proceedings.
15. S. K. Datta, K. C. Wong, and A. H. Shah,

"Dynamic Stresses and Displacements around
Cylindrical Cavities of Arbitrary Shape,"
Journal of Applied Mechanics , Transactions of
ASME, 51 (1984), pp. 798-803.
16. K. F. Chow, Three Dimensional Motion of a
Tunnel Buried in an Infinite Medium , M.S.

Thesis, Department of Civil Engineering,
University of Manitoba, Winnipeg, Canada.
279
TABLE 1
Material Properties of Soft Soil and Rock
Material Case I II
Properties Soft Soil Rock
3
p kg/ra
3
2.665 * 10 2.665 * 10 3
2 8 9
E N/n 6.9 * 10 7.567 * 10
a 0.45 0.25
Fig. la Geometry of the Pipe or Tunnel
280
Fig. lb Geometry of Tunnel (End View)
281
0
Fig. 2 Tunnel in a Half-space
282
03
TUNNEL IN SOFT SOIL

P-WRVE
K
2
fl = 0.118
I
SZZ| |
STT|
O A
o
o
0.00 15.00 30.00 45.00 60.00 75.00 90.00 ?05. 00
INCIDENT ANGLE
283
Fig. 4 Hanimum axial stress and hoop stress in
tunnel vail vs incident angle (or P Wave
k, a «0.
1 32 (Case II)
o
o
"
4 .00 15.00 30.00 115. 00 60.00 75.00 90.00 105.00

INCIDENT RNGLE
284
P-WflVE (5 DEG.)
STT (CONCRETE)
TUNNEL IN ROCK TUNNEL IN SOFT SOIL
—
Ulfl
TT
(M/SEC) NF —
UJfl
tt
(M/SEC) NF
2 2
89.50 + 0.277 89.50 0 3.530

134.33 x 0.623 134.33 A 7.930
Fig. 5 Comparison between Case I and Case II for

normalized hoop stress in tunnel wall for
P-Wave, incident angle 5*
285
P-WflVE 185 DEG.)
5TT (CONCRETE)
TUNNEL IN ROCK TUNNEL IN SOFT SOIL
—
UJfl
TT
(M/SEC) NF —UJfl
(M/SEC) NF
2 2 TT
B9.50 + 0.094 89.50 0 2.910

134.33 X 0.212 134.33 A 6.540
Fig. 6 Comparison between Case I and Case II (or

normalized hoop stress in tunnel wall (or
P-Wave, incident angle 85*
286
Pd
ZT
CD
TUNNEL IN SOFT SOIL
SV-WRVE
LD
CD
t—
—
1
00
in c\j
o
az
o
rvi Pd
rvi
in o
X zt
CL
CD
r~
o
o
o
o
4 Too 15.00 30.00 4 5.00 60.00 75.00 90.00 105.00
INCIDENT ANGLE
287
rvj
INCIDENT RNGLE
288
cvi
o
TUNNEL IN SOFT SOIL
SH-HflVE
K A = 0.118
2
o
C\J
SZZI ISTTI
o I
O a
Fig. 9 Maximum axial stress and hoop stress in

tunnel vail vs incident angle (or SH-Wave,
k 2 a 3 0. 1 18 (Case 1)
o
o
“
4 . 00 15.00 30.00 U5. 00 60.00 75.00 90.00 105.00

INCIDENT RNGLE
289
PJ
290
TUNNEL IN SOFT SOIL
P WAVE
STT; CONCRETE
F. E.>l.
o
CD
Fig. 11 Hoop Stress in a Tunnel in a Half-space Due to

Incident P Wave. Depth Ratio = 3.33
291
TUNNEL IN SOFT SOIL
P WAVE
STT; CONCRETE
F.E.H.
o
00
Fig. 12 Hoop Stress in a Tunnel in a Half-space due to

Incident P Wave. Depth Ratio = 6.67
292
TUNNEL IN ROCK
S V WAVE
STT; CONCRETE
F.E.M.
and
Rock
in

Embedded
C
QJ
oj -o
C *H
c a
3 C
H hh
CD i-H
U *H
^ O
00 OO
O-
O u-i
-U TUNNEL IN SOFT c L
o o S v WOVE
CC CO STT; C C N C R E T
F.E.M.
CO
00
•H
fin
295
T U JV N E L IN ROCK
2 v wave
S T T ; C C N c A £
F . E . M
n
Z)
DO
C
<
and
Rock
in
Embedded
Tunnel
eg o
u
C
3
</> -o
in
TUNNEL IN SOFT SOIL 0) *“H
5 V WAVE *-» o
STT; CONCRETE co co
F.E.M. c- *->
O U*
O O
3: co
00
u-
294
TUNNEL IN SOFT SOIL
R WAVE
STT; CONCRETE
F . E . M*.
o
rr
Fig. 15 Hoop Stress in a Tunnel due to Rayleigh Wave

Propagating From Left to Right.
2S5
Theme III
STORM SURGE AND TSUNAMIS

DAMAGES TO COASTAL AND RIVER STROCTORES
CAUSED BY THE NIHONKAI-CHUBU EARTHQUAKE TSUNAMI
Takaaki Uda, Public Works Research Institute

Atsushi Omata, Public Works Research Institute
Abstract part, and the Shibikawa River in southern part of

region are investigated. Most of these rivers are
The 1983 Ni honkai -Chubu Earthquake Tsunami damaged located in the northern part as one group and the
many coastal structures and river administration rest in the south. No rivers enter the sea between
facilities. These were investigated by using pho- the Yoneshiro River and Kotohama Coast as shown in
tographs taken right after the tsunami. The dominant Figure 1 and sand dunes extend along the coastline.
types of damage were scouring around the structure, There is a sea cliff between Kotohama Coast and Oga
collapsing of the structure, and scattering of the Peninsula. This cliff is interrupted at some loca-
blocks of detached breakwater or wave dissipation tions along the coast by small streams emptying into
breakwater. Damage patterns are classified with the sea producing a lowland around the mouth. The
respect to the kind of structure and estimated exter- Shibikawa River is one of the rivers having such a
nal force. Finally, all instances of damages are characteristic. In the following sections, damage to
counted. In total, 63 coastal structures and 12 structures due to the tsunami are shown by some pho-
river administration facilities were damaged. tographs and a pattern classification of damages is
made.
1. INTRODUCTION
2. DAMAGES TO COASTAL STRUCTURES
The Ni honkai -Chubu Earthquake Tsunami which occurred
at noon on May 26, 1983, caused devastation along the In this section damage to coastal structures due to
Japan Sea Coasts. The author reported on the tsunami the tsunami are discussed. The kinds of structures
inundation and the general view of damages due to the damaged are coastal revetments, detached breakwaters
tsunami in the previous paper*). Many coastal struc- and wave dissipation breakwaters only. Some charac-
tures and river administration facilities were teristic cases will be discussed, since it is im-
damaged by the tsunami. In order to consider coun- possible to investigate all of the cases of damages
termeasures to future tsunamis, it is important to to the structures.
classify the damages to the structures. In this

study, photographs illustrating the characteristics An example of a damaged coastal revetment is the
of the damage are presented and a pattern classifica- stair-type revetment in Noshiro Port. The coast is
tion is done. located north from the Yoneshiro River. Figure 2
shows the damage to the stair-type revetment. At the
The northern coast of Akita Prefecture, as shown in end joint of the revetment a depression occurred and
Figure 1, is selected for the study area as extensive a large crack was formed between the stair part of
tsunami damage was reported there. The damage to the revetment and the top. The foundation of the top
coastal structures and river administration facili- of the revetment was scoured by the tsunami. The
ties are investigated in detail. Typical examples of cause of the damage is considered to be due to both
the damaged structures are the cases of the stair- the earthquake and the tsunami, judging from the form
type revetments in Noshiro Port and Kotohama Coast, of the damage to the revetment.
the detached breakwaters of the Hachiinori Coast and

wave dissipation breakwaters of the Minehama Coast. A second example is the damage to the stair-type
revetment on the in Kotohama Coast. Figure 3 shows
The circumstances of the damage occurring to the the situation behind the revetment. The top of the
river administration facilities on the Mizusawa, the revetment, which is made of concrete 20 cm thick,
Hanawa, the Tako and the Yoneshiro Rivers in northern collapsed. A small drainpipe existed at this site.
299
The return flow of the tsunami that overflowed the 2 . 1 Damage to River Structures
top of the revetment concentrated on the pipe.
Therefore, it is believed that the earth around the Damages to the river administration facilities in
drainpipe was washed away, forming a cavity, and five rivers in Akita Prefecture are shown. The
allowing the top of the revetment to collapse. locations of these five rivers are depicted in
Figure 1. In what follows, damage in each river is
One characteristic damage due to the 1983 investigated in detail by using photographs.
Ni honkai -Chubu Earthquake Tsunami was the scattering
of concrete blocks. The damage is revealed in pho- In the Mizusawa River the river revetment which was
tographs. A general view of these has already been located 300 m from the river mouth was broken for a
given in the previous paperl). Figure 4 was taken distance of 148 m. The revetment collapsed and
from the shore near the detached breakwater along the extensive scouring occurred behind the revetment.
Hachimori Coast right after the first onset of the The foot protections in front of the revetment were
tsunami. The seabed beyond the detached breakwater carried away. Figures 8 and 9 show the charac-
is visible since the tsunami has retreated offshore. teristics of the damages. These photographs were
The blocks are scattered around the detached break- taken from the downstream and upstream ends of the
water. Figure 5 shows the detail of the scattering revetment, respectively. The earth behind the
of the blocks, including many 8-ton blocks scattered revetment was washed away. The downstream side of
toward the shore. the revetment has fallen down and the upstream side
was scoured severely due to the tsunami overflowing
Next, an example of damage to wave dissipation break- the revetment. The cause of the breakdown of the
waters is shown. Wave dissipation breakwaters differ revetment is as follows. First, the tsunami caused
from detached breakwaters on the position of con- scouring behind the revetment and the stability of
struction relative to the shoreline. A detached the structure was decreased. Next, it was collapsed
breakwater is built off the initial shoreline. On by the return flow of the tsunami.
the other hand a wave dissipation breakwater is built

on the beach inland from the shoreline. Figure 6 In the Flanawa River the revetment which was located
shows the scattering of blocks of wave dissipation 250 m from the river mouth was broken for a distance
breakwater at the Minehama Coast. Concrete blocks of 202 m. Damages due to the collapsing of the slope
weighing 4 tons were scattered deep into the protec- protection and the scouring of the anbankment are
tion forests. The forests were damaged, but they noted. Figures 10 and 11 show the damage to the
were effective in preventing concrete blocks or embankment. Figure 10 was taken from the downstream
floating debris from being transported deep into the side of the damaged embankment. The anbankment was
land. The return flow frequently concentrated on a broken severely at the location adjacent to the river
narrow zone along the sand dunes. In such a region mouth. Figure 11 shows the details of the damaged
the sand dunes were eroded due to the return flow of revetment of combined blocks. The combined blocks
the tsunami as shown in the figure. Figure 7 shows were gathered together. Because the body of the dike
the same kind of situation of scattering of blocks on was made of earth, it was broken easily and the
the Minehama Coast. embankment was scoured.
Regarding the scattering of blocks, it is thought The Tako River meanders in the counterclockwise
that the buoyancy effect caused by the rise of sea direction near the river mouth. The dike of the
water level and the high velocity of the tsunami meandering part of the river was reinforced with a
front have important roles, though the exact nature revetment. The revetment facing to the sea was
is not known. The tsunami propagated on the gently severely broken. Figure 12 was taken from the river
sloping bottom as shown in Figure 1, and then in the mouth looking in the upstream direction. The river
form of a bore with short period osci 1 lations^). dike was eroded at two locations: right and left
This must effectively increase the impulsive force from the center of Figure 12. The embankment shown
exerted on structures. near the center of Figure 12 was severely damag<-
300
and a bridge crossing the river was carried away. after the tsunami. Here the pattern of the damage
The slope protection of the embankment was also and the estimated external force will be discussed.
broken. Figure 13 shows the details of damage In order to generalize the result of the classifica-
to the embankment at the downstream side. The blocks tion, some patterns of damage which were not observed
were scattered, and the top of the embankment was in the present tsunami will be included.
eroded severely. Figure 14 shows the damage at the

upstream site. A large scouring hole was formed First, the patterns of damage to coastal dikes and
behind the dike. The hole formed by scouring was revetments are expressed in Table 1. Damage to
55 m long, 18 m wide and 2.5 m deep at maximum. structures that was caused mainly by the return
Figure 15 shows the scouring of the abutment shown in flow of the tsunami consist of scouring of the toe,
the lefthand side of Figure 14. The ground under the forward inclining or falling down of the revetment
abutment was scoured to a depth of about 0.7 m. and damage to the parapet. Damage to structures due
to tsunami run-up include the falling of the upper
In the Yoneshiro River the river revetment was broken structure of the coastal dike and damage to the
due to the vibration of the earthquake and to the parapet. These were frequently found when rein-
tsunami at a location about 600 m upstream from the forced concrete was not used between the body and
river mouth. Figure 16 shows the damage to the the upper structure, and when there existed a weak
revetment. The low-water concrete revetment was point in the structure.
moved forward and scouring occurred at the top of the

revetment although the exact cause is not known. The There are other kinds of damage caused by both the
surmised events are as follows: first, the revetment vibration of the earthquake and by the tsunami. The
was partially destroyed by the earthquake, and then damage to the stair-type revetment occurring at
the scouring and the subsidence occurred due to the Noshiro Port is an example of these kinds of damage.
tsunami attack after the earthquake. Near the same
location as shown in Figure 17, a high-water revet- Regarding detached breakwaters and wave dissipation
ment of combined concrete was carried away. This may breakwaters, a dominant feature of the damage was the
have been due to the tsunami current. scattering of the constituent blocks. This must be
one of the characteri sties of the present tsunami.
Finally, the damage to the revetment in the Shibikawa In addition, the subsidence of the blocks was fre-
River is discussed. The Shibikawa River is located quently found. Table 2 sums up these types of
south of the other four rivers as is shown in Figure damage. Though it was found that a large number of
1. Sea cliffs stand around the river mouth and a low the blocks had subsided, there was a problem in
land is confined near the river mouth. The tsunami counting the number. This is because it was not
strongly invaded into the river mouth from the north always known whether the cause was due to the tsunami
and a paddy field was destroyed by the flow over the when the subsidence of the blocks was noted or due to
river revetment. The revetment which extended along wind waves sometime before the tsunami attack.
the river was extensively broken. The damage to the
revetment was severe on the lefthand side of the Patterns of damage to river administration facilities
river. Figure 18 was taken from the right embankment are summed up in Table 3. The kinds of damaged
and shows the damage to the revetment on the lefthand structures include the body of the enbankment, the
side of the river. The main cause of the damage may revetment and the foot protection. Slope failure
be due to the return flow of the tsunami, since the caused by the vibration of the earthquake which
revetment collapsed and the fill under the revetment occurred near the river mouth was included in the
was carried away. damages to the body of the enbankment. In addition,
scouring of the slope and the breakdown of the dike
2.2 Pattern Classification of Damage were found. Regarding the types of damage to the
revetment, breakdown of the slope of the low-water
In the previous sections, the types of damage to revetment, scouring behind the revetment, forward
coastal structures and river adini nistration facili- inclining or falling of the revetment and collapsing
ties were disclosed by using photographs taken right of the slope protection were found. The causes are
301
thought to be as follows: regarding the inclining or
falling down of the revetment, scouring behind the
revetment is caused by the overflow or the return
flow of the tsunami decreasing the stability of the
structure. Then the structure collapsed due to the
current of the tsunami. The slope protection made of
combined blocks was lost when the blocks were carried
away by the current of the tsunami.
Finally, Table 4 counts up the number of instances of

damage with the classification of the damage pattern.
In the table each case was counted as one when two
kinds of damages occurred at the same location. The
total number of instances of damage to coastal struc-
tures and river administration facilities are 63 and
12, respecti vely.
3. SUMMARY
In this paper the damage to coastal structures and

river admi ni stration facilities due to the 1983
Ni honkai -Chubu Earthquake Tsunami were investigated
by using photographs taken right after the tsunami.
Next, the damage patterns were classified with

respect to the kind of structures and estimated
external force. As a result, tables showing the pat-
tern classification of damages were obtained

regarding coastal dike, revetment, detached break-
water, wave dissipation breakwater and river admin-
istration facilities. Finally, all of the cases were
counted. It is found that the number of instances of
damage to coastal structures and river admin-
istration facilities are 63 and 12, respectively.
4. REFERENCES
1) Uda, T. , Inundation damages due to the 1983

Ni honkai -Chubu Earthquake Tsunami, The 16th
Joint Meeting U.S. -Japan Panel on Wind and
Seismic Effects, pp. 739-762, 1984.
2) Shuto, N. , Behavior of the 1983 Ni honkai-Chubu

Earthquake Tsunami in the northern part of Akita
Prefecture, 31 Japanese Conf. on Coastal Eng. ,
pp. 247-251, 1984 (in Japanese).
302
Mizusawa River
Fig. 1 Location map of coasts and rivers in Akita Prefecture
303
Fig. 2 Damaged stair-type revetment of Ochiai Coast in Noshiro
Port
Fig. 3 Damage behind the stair-type revetment in Kotohama ( oast
304
305
Fig. 5 Scattering of concrete blocks of detached breakwater in
Hachimori Coast
Fig. 6 Scattering of concrete blocks of wave dissipation break-

water in Minehama Coast (1
Fig. 7 Scattering of concrete blocks of wave dissipation break

water in Minehama Coast (2 )
306
Fig. 8 Damaged revetment of the Mizusawa River (looking up-
stream)
Fig. 9 Damaged revetment of the Mizusawa River (looking down-

stream)
Fig. 10 Damaged embankment of the Hanawa River
307
Fig. 1 1 Damaged blocks and eroded embankment of the river
Fig. 12 Damaged embankment of the Tako River
Fig. 13 Situation of the damaged embankment in the lako River
308
Fig. 14 Scouring behind the dike of the Tako River
Fig. 1 5 Scouring under the abutment
Fig. 16 Damaged revetment of the Yoneshiro River
309
Rjver
Shibikawa
the
of
revetment
Damaged
310
Table 1 Pattern classification of damages of coastal dike and re-
vetment
Table 2 Pattern classification of damages of detached breakwater

and wave dissipation breakwater
301
Table 3 Pattern classification of damages of river administration
facilities
ft.
312
Table 4 Summary of pattern classification
1
Ex. force Faci 1 i
ty Pattern of damage Number
Scouring of the toe 3
Tsunami Forward inclining

of the revetment 2
(reteat) Revetment
Falling down (forward) 3
C/5
05
and Damage of the parapet 1
O Falling of upper structures

CS Tsunami Embankment of the coastal dike (backward) 2
c (run-up)
c Damage of the parapet 1
o
q; Earthquake Damage of the
l! & Tsunami stair-type revetment 2
05 Breakwater Scattering of the blocks 5
(Feet
protection)
Subsidence of the blocks 1 4
Wave Scattering of the blocks 6
Tsunami dissipation
breakwater Subsidence of the blocks 1
Betat.ched Scattering of the blocks 1 7
breakwater Subsidence of the blocks 6
Slope col lapsing 1
Body of the
c/5 Scouring of the crown
05
embankment & backs lope 1
O'
Breaking 3
- Breakdown of the slope
c
c Tsunami of the low- water revetment 1
L.
-«-d
Revetment Scouring behind the revetment 1
C/5
c Forward inclining 1
g
s Falling 1
> Collapsing of slope protection 2

Foot Scattering of the
protection concrete blocks 1
* Estimated external force causing the damage
313
PROGRESS IN DEVELOPING TSUNAMI DATA BASE
by James F. Lander and Patricia A. Lockridge
National Geophysical Data Center, NOAA

Boulder, CO
ABSTRACT Data Center agreed to produce an educational map

depicting the history of occurrence of tsunamis in
In support of the Tsunami Hazard Reduction Using the Pacific and to develop a data base of available
System Technology (THRUST) demonstration project for data to support the program in its demonstration pro-
Chile, the National Geophysical Data Center (NGDC) ject for Chile. For convenience, a digital data
has published a multi-colored map of Tsunamis in the base on the occurrence and effects of tsunamis in the
Pacific Basin (1900-1983) (approximately 110 x 150 Pacific was developed from available sources (I, 2,
cm). NGDC is also preparing a compilation of infor- 4, 5, 6, 7). The map, "Tsunamis of the Pacific
mation about Chilean tsunamis including occurrences, Basin, 1900-1983" has been published and a com-
local and distant effects, marigraphic data, tide pilation of data and information on tsunamis which
station locations and characteristics, photographs have affected Chile is being readied for publication
and a selected bibliography for joint publication with the Chilean Hydrographic Institute later this
with the Chilean Hydrographic Institute later this year. The digital data base has proved to be a
year. valuable resource in itself and considerable pro-

gress has been made to extend its coverage in time
Substantial progress has been made in producing a and areas included and of details of effects. There
digital data base of tsunami occurrences, including are, of course, limitations on the completeness and
reported wave heights, effects and references. For types of data available from around the Pacific and
example, this file can be used to produce lists or from earlier times. Earthquake recording devices
maps of destructive tsunamis, tsunamis which produced and recorders were few and primitive until well into
waves and/or damage beyond the source region, tsu- the 20th Century.
namis caused by or associated with volcanic erup-

tions, tsunamis effecting certain regions from both Also, much of the area has only a short historical
local or remote sources, and tsunamis producing runup record. The rapid development of much of the
heights of a specified level or higher. It can be Pacific coastal regions makes it essential to deter-
used for a variety of statistical purposes. This mine the risks from tsunamis and develop counter
paper illustrates some of these. measures against them.
The file will be completed for the Pacific Basin and 2 . DIGITAL TSUNAM T DATA BASE
extended to include tsunamis in the Medi terranean and
Caribbean Seas, and Atlantic Ocean. Information on The current file consists of data drawn essentially
source dimensions as inferred from earthquake after- from published references for all events except the
shocks and information on focal mechanism will also smaller tsunamis (less than 1.5 meters wave heights
be added to the file over the next several years. or undamaging) prior to 1900 (1, 2, 4, 5, 6, 7, 8,
9, 10). The file contains information on date,

Key Words: Tsunamis, Seismic Sea Waves, Natural location, effects, reference to original source, and
Hazards, Pacific Basin, Chile, Japan, Hawaii a validity code. Where available, it will have
information on wave height, earthquake magnitude,
and tsunami magnitude and intensity. The validity
1. INTRODUCTION
code ranges from 4 (certain) to 0 (no tsunami). A
As part of an AID-sponsored project to develop a validity code is necessary as It Is not always clear
comprehensive program to mitigate tsunami devastation from accounts whether the observed wave activity was
in developing countries, the National Geophysical due to a meteorological source (storm surge) or a
3M
tectonic source such as an earthquake. Also, some area. The columns represent the damage caused in
tsunamis are listed under an erroneous date in some each region by tsunamis originating in a region.
catalogs due to mi sconversion from the several calen-
dar systems in use over the historical period. These Regions 0, 1, 2, 3, 4 and 8 have not generated
are left in the file with intensity 0 to prevent tsunamis destructive beyond their region in the last
their being re-entered in the file. 100 years while region 9, South America, has gen-
erated tsunamis destructive to five other regions.
Entries for some areas such as Hawaii and Chile have The rows show the history of destructive tsunamis in
been completed for all known events. The file is a region. Hawaii, for example, has generated only
being actively expanded to complete the entries from three locally damaging tsunamis but has experienced
available sources, add new sources including, in par- damage from nine more tsunamis generated remotely.
ticular, the Japanese history and other areas such as Every source region for remotely damaging tsunamis
the Atlantic Ocean and the Medi terranean and has generated tsunamis damaging to Hawaii.
Caribbean Seas.
Figure 5 shows a similar diagram for fatalities.
Table 1 illustrates the present contents. Japan, while not a notable producer of destructive
tsunamis (13 in the last 100 years), has by far the
2.1 Examples From Digital Data most fatalities. Even excluding the 1896 Sanriku
tsunami, it has had over 9,000 fatalities. It is
The following examples have been drawn from the file noteworthy that 99 percent of the approximately
to illustrate some types of questions which can be 50,000 fatalities are due to locally generated tsuna-
addressed. The file is still incomplete so changes mis. Meaningful damage values are rarely available
in detail may be expected. The principal conclu- but the ratio of damage from regionally generated
sions, however, should be valid. Also, other tsunamis versus damage from tsunamis from remote
researchers using other methods may have previously sources is probably between 90 percent (frequency
published results similar to these illustrations of of occurrence of damaging tsunamis ratio) and 99 per-
the uses of the file. cent (number of fatalities ratio). Except for Hawai
which has most of the fatalities from remote source
Much of the interest in tsunamis focuses on their tsunamis and a far greater risk from these tsunamis
capability to cause damage. Table 2 lists 94 than from local tsunamis, all other regions have a
destructive tsunamis in the last 100 years, for an far greater risk from local and regional tsunamis.
average of approximately one per year. Figure 1 It seems clear that improved preparedness and
shows the location of these events. Most lie along warnings for local tsunamis have the greatest poten-
the western margin of the Pacific. Figure 2 shows a tial for saving lives but it remains necessary to
map of boundaries defined for tsunami source regions have accurate warning for remote sourced tsunamis
in this report. for which hours to nearly a day's warning times are
possible.
Most damaging tsunami effects are limited to the
generating region. Of the 94 damaging tsunamis in The frequency of occurrence of damaging tsunamis
the last 100 years, only 9 (10 percent) caused decreases rapidly with the lower earthquake magni-
damage beyond their source region. These are shown tude. Figure 6 shows that two-thirds of the damaging
on Figure 3. At least in the last 100 years all such tsunamis are associated with earthquakes of magnitude
tsunamis have originated in Japan, Kuril-Kamchatka, 7.5 or greater. Still, for local tsunami warning
Aleutian Islands and South America. Table 3 lists purposes, earthquakes as small as magnitude 6 or less
these events, exclusive of the Sea of Japan events. may need to be considered. About 1600 people have
Note that in almost every instance Hawaii receives been killed this century by tsunamis from these
the largest waves outside of the source region. smaller earthquakes, three times more than those
killed from remote sourced tsunamis. Table 4 lists
Figure 4 summarizes the frequency of occurrence of these lower magnitude destructive tsunamis. This
tsunamis by region. The diagonal represents the understates the risk since magnitudes are not
total number of damaging tsunamis originating in an always reported nor are damages always cataloged for
315
smaller events, particularly in the first half of 3. Solov'ev, S. L., and Ch. N. Go, "Map of the
this century. Source Regions and Heights of Tsunamis in the
Pacific Oceans," Moscow, 1977 (Russian legends).
The file can be used to locate the tsunamis which
have effected an area. Figure 7 shows the location 4. Solov'ev, S. L., and Ch. N. Go, "A Catalogue of
of tsunamis effecting Hawaii. Hawaii is at risk Tsunamis of the Eastern Shore of the Pacific
from all sources of tsunamis capable of producing Ocean," Nauka Publishing House, Moscow, 1977.
destructive tsunamis beyond the source region and, in
fact, probably the wave heights observed there could 5. Solov'ev, S. L. ,
and Ch. N. Go, "A Catalogue of
be used to predict wave heights elsewhere around the Tsunamis of the Western Shore of the Pacific
Paci fic. Ocean," Nauka Publishing House, Moscow, 1974.
The file is being used to determine the historical 6. I TIC Newsletter, International Tsunami Informa-
effects of tsunamis originating in one region have on tion Center, Honolulu, Hawaii, published quar-
another region. Japanese tsunamis, again exclusive terly 1968-1976.
of the Sea of Japan earthquakes, are shown to have
caused damage only once (1933, Hawaii). The file can 7. Tsunami Newsletter, International Tsunami Infor-
also be used to identify the threat that remote mation Center, Honolulu, Hawaii, 1977-1983.
sourced tsunamis pose for a region. Chile, for
example, has experienced eight tsunamis from remote 8. Cox, Doak C. and George Pararas-Carayanni s with
sources in the last 100 years but only the 1964 revisions by Jeffrey P. Calebaugh, "Catalog of
Alaskan tsunami caused damage and that was minor. Tsunamis in Alaska," SE Report-1 - National
Oceanic and Atmospheric Administration, World
Tsunamis caused by volcanic sources may not be fully Data Center-A, Boulder, Colorado, 1976.
included in hazard warnings. The file contains
information on over 50 volcanic tsunamis in the 9. Pararas-Carayannis , George with revisions by

Pacific in the last 100 years. Over 100 fatalities Jeffrey P. Calebaugh, "Catalog of Tsunamis in
from this source are listed. Hawaii," SE Report-4, National Oceanic and
Atmospheric Administration, World Data Center-A,
The above examples illustrate some of the questions Boulder, Colorado, 1977.
which the digital file can treat. While caution must
be used in applying the historical record to future 10. United States Earthquakes, U.S. Department of
events and the deficiencies of the data must be the Interior, Geological Survey and U.S. Depart-
remembered, the information has potential for ment of Commerce, National Oceanic and Atmos-
improving the local and Pacific tsunami warning pheric Administration, annual volumes, 1944-1980.
system.
11. Lockridge, P. A., R. H. Smitn, "Tsunamis in tne
3. REFERENCES Pacific Basin," National Geophysical Data Cen-

ter, National Oceanic and Atmospheric Adminis-
1. Iida, K. , D. Cox and G. Pararas-Carayanni s tration, Boulder, Colorado, 1984.

"Preliminary Catalog of Tsunamis Occurring in
the Pacific Ocean," Hawaiian Institute of Geo- 12. Bernard, Eddie N., James F. Lander, Gerald T.
physics, University of Hawaii, 1967. Hebenstreit, "Feasibility Study on Mitigating

Tsunami Hazards in the Pacific," NUAA Technical
2. Cox, Doak C. , Kumizi Iida, S. L. Solov'ev, and Memorandum ERL PMtL-37, December 1982.
George Pararas-Carayanni s ,
"Pacific Tsunami
Historic File," National Oceanic and Atmospheric 13. Lander, James F., “Data Services Supporting the
Administration, Boulder, Colorado, unpublished THRUST Project," United States-Japan Panel on
manuscri pt. Wind and Seismic Effects, 16th Joint Meeting,

May 15-18, 1984.
316
TABLE 1
TSUNAMI DATA FILE CONTENTS
NR. OF EVENTS NR. OF EVENTS NR. OF LOCATIONS NR. OF EVENTS NR. OF LOCATIONS
(validities 2-4) REPORTING WAVE REPORTING WAVE REPORTING DEATHS REPORTING DEATHS
HEIGHTS HEIGHTS OR DESTRUCTION OR DESTRUCTION
CURRENT FILE* 789 268 1825 259 644
CHILE ONLY 250 82 295 34 126
2 OTH CENTURY ONLY 407 170 1453 87 255
* DOES NOT YET INCLUDE MINOR TSUNAMIS BEFORE 1900
EXCEPT FOR CHILE AND HAWAII.
317
TABLE 2
OBSTRUCTIVE TSUNAMIS OF LAST 100 YEARS SORTED BY REGION WITH DAMAGE CODE AND NUMBER OF DEATHS, 1885-1984
DATE LAT. LONG. E.MAG RUNUP SOURCE VALID. CAUSE DAMAGE DEATHS
REGION 0 3 TOTAL; 1 CONSIDERABLE DAMAGE; 35 DEATHS
1903 11 29 10.0 HAWAII 4 C

1951 08 21 19.5 -156.0 12 6.90 3.6 HAWAII 4 T C 33
1975 11 29 19.4 -155. 05 7.20 8.0 HAWAII 4 T B 2
1917 06 25 -15.6 -173.0 SH 8. 30 12.0 SAMOA ISLANDS 4 T B

194 03 26
7 -38.8 178.6 SH 7.00 4.6 N. NEW ZEALAND 3 T C
1953 09 14 -18.6 178.6 60 6. 80 2.0 FIJI ISLANDS 4 TSiL C C
REGION 2 15 TOTAL; 6 CONSIDERABLE DAMAGE; 3 EXTENSIVE DAMAGE; 407 DEATHS
1895 03 06 -8.5 150.0 7.5 W. SOLOMON SEA 4 T A 26

1906 09 15 -7.0 149.0 SH 8.00 1.5 W. SOLOMON SEA 4 T B
1919 05 07 -6.0 153.0 7.75 2.5 BISMARCK SEA, NEW GUINEA 4 T B C
1926 09 17 -11.6 160.0 50 7.00 2.0 SOLOMON ISLANDS 4 T B
1930 12 24 - 1.3 144.3 5.75 2.5 MELANESIA 4 T A 5
1931 10 04 -10.6 161.8 SH 7.90 9.0 SOLOMON ISLANDS 4 T A 50
1938 05 13 -6.0 147. 7 SH 7.50 BISMARCK SEA, NEW GUINEA 3 T C
1939 04 30 -10.6 158.6 50 8.00 10.5 SOLOMON ISLANDS 3 T C 12
1951 02 22 -3.6 142.6 3.5 BISMARCK SEA, NEW GUINEA 2 T c
1970 11 01 -4.9 145.5 20 7.00 3.0 BISMARCK SEA, NEW GUINEA 4 T c 3
1971 07 14 - 5.5 153.9 47 8.00 3.0 BISMARCK SEA, NEW GUINEA 4 T c 1
1971 07 26 -4.9 153.2 48 8.00 3.5 BISMARCK SEA, NEW GUINEA 4 T B

1974 02 01+ -7.4 155.6 40 7.00 4.0 SOLOMON ISLANDS 4 T C
1975 07 21 -6.6 154.9 50 6.70 2.0 SOLOMON ISLANDS 4 T B 200
1979 09 12 - 1.7 135.9 05 8. 10 2.0 W. IRIAN, INDONESIA 2 T B 100
REGION 3 18 TOTAL; 6 CONSIDERABLE DAMAGE; 6 EXTENSIVE DAMAGE; 5,378 DEATHS
1885 04 30 -2.5 127.5 7.2 CERAM 3 T&V C C

1889 09 06 3. 125.6 8.0 N. MOLUCCA IS., INDONESIA 4 T&V B C
1899 09 30 -3.5 128.5 7.8 BANDA SEA, INDONESIA 4 T6V A 3,620
1907 01 04 1.5 97.0 7.50 2.8 SW. SUMATRA 4 T A 400
1907 03 30 3.0 122.0 7.25 4.0 CELEBES SEA, INDONESIA 2 T A
1918 08 15 5.6 123.0 SH 8. 30 7.0 CELEBES SEA, INDONESIA 4 T B B
1921 05 14 0.0 118. 6.25 MAKASSAR STRAIT, INDONESIA 2 T B
1927
1928
1938
12 01
08 04
05 20
-0.5
-8.3
-0.7
119.5
121.5
120. 3
—
6.00
7.50
15.0
10.0
2.8
CELEBES SEA, INDONESIA
FLORES SEA, INDONESIA
N. MOLUCCA ISLANDS INDONESIA
,
2
4
4
T
V
T
C
B
C
14
128
17
1939 12 22 0.0 123.0150 8.60 N. MOLUCCA ISLANDS INDONESIA
, 3 T C
1965 01 24 - 2.4 126.0 6 7.60 4.0 CERAM ISLAND, INDONESIA 4 T B 71
1967 04 11 -3.3 119.4 33 5.50 3.0 MAKASSAR STRAIT, INDONESIA 3 T C 13
1967 04 12 5.5 97.3 58 7.50 2.0 MALAY PENINSULA 3 T C
1968 08 14 0.2 119.8 23 6. 10 10.0 BANDA SEA 4 T A 200
1969 02 23 -3. 1 118.8 33 6.88 MAKASSAR STRAIT, INDONESIA 4 T A 600
*1977 08 19 -11.0 118.4 SH 8.00 30.0 SUNDA ISLANDS 4 T B 100
1979 07 18 -8.5 123. 5 10.0 LOMBLEN ISLAND, INDONESIA 2 T4L A 187
REGION 4 7 TOTAL; 5 CONSIDERABLE DAMAGE; 5.059 DEATHS
1897
1933
1934
09 21
12
02
25
14
6.8
12.8
17.6
122.5
124.0
119.0
—
8.5
SH 7.60
1.4
1.0
SULU SEA, PHILIPPINES
E. SAMAR ISLAND, PHILIPPINES
W. LUZON ISLAND, PHILIPPINES
4
2
4
T
V
T
B
B
B
13
9
1948 01 25 10.6 122.0 SH 8.20 2.0 SULU SEA, PHILIPPINES 4 T 20

1949 12 29 18.0 121.0 SH 7.20 2.0 E. LUZON ISLAND, PHILIPPINES 3 T C 16
1975 10 31 13.0 126.2 33 7.40 PHILIPPINE TRENCH 4 T B 1
1976 08 16 6.3 124.0 SH 8.00 4.9 MORO GULF, PHILIPPINES 2 T B 5,000
REGION 5 13 TOTAL; 3 CONSIDERABLE DAMAGE; 7 EXTENSIVE DAMAGE; 36.500 DEATHS
*1896 06 15 39.6 144.2 7.6 SANRIKU, JAPAN 4 T A 27, 122

1911 06 15 29.0 129.0 8.90 2.0 RYUKYU TRENCH 4 T B 6
1914 01 12 31.0 130.4 6.20 3.0 SEIKAIDO, JAPAN 4 T C 35
318
TABLE 2 PAGE 2
DATE LAT. LONG. E. MAG RUNUP SOURCE VALID. CAUSE DAMAGE DEATHS
1923 09 01 35.3 139.5 SH 8.20 12. TOKAIDO, JAPAN 4 T A 2, 144

1927 03 07 35.6 135.0 10 7.60 11.3 SW. HONSHU ISLAND, JAPAN 4 T A 1, 100
*1933 03 03 39.2 143.0 10 8.30 28.0 SANRIKU, JAPAN 4 T A 3,000
*1944 12 07 33.8 136.0 8. 30 8.4 TOKAIDO, JAPAN 4 T A 1,040
1946 12 21 33.0 135.6 30 8.00 6.0 NANKAIDO , JAPAN 4 T A 1,997
1952 03 04 42.2 143.8 45 8.00 6.5 SE. HOKKAIDO ISLAND, JAPAN 4 T A 33
1964 06 16 38.3 139.2 40 7.60 6.4 NW. HONSHU ISLAND, JAPAN 4 T C 26
1968 04 01 32.3 132.5 32 7.75 2.3 SEIKAIDO , JAPAN 4 T B
1968 05 16 41.4 142.9 33 7.05 5.0 JAPAN TRENCH 4 T B
1973 06 17 43. 2 145. 8 33 6.50 3.0 KURIL, US SR—HOKKAIDO , JAPAN 4 T C
REGION 6 7 TOTAL; 4 CONSIDERABLE DAMAGE; 2 EXTENSIVE DAMAGE; 258 DEATHS
1918 09 08 45.6 151.6 SH 8.30 12. 1 S. KURIL ISLANDS, USSR 4 T C 23

192 3 02 04 54.0 161.0 SH 8. 30 8.0 KAMCHATKA PENINSULA, USSR 4 T B 3
1923 04 14 56.6 162.6 SH 7.20 20.0 KAMCHATKA PENINSULA, USSR 4 T B 18
1940 08 02 44.0 139.6 10 7.00 3.5 W. HOKKAIDO ISLAND, JAPAN 4 T A 7
*1952 11 05 52.8 159.6 45 8. 30 18.0 KAMCHATKA PENINSULA, USSR 4 T A A
1958 11 07 44.6 148.9 1008. 30 4.0 S. KURIL ISLANDS, USSR 4 T B
*1983 05 26 40.2 140.0 24 7.80 14. NOSHIRO, JAPAN 4 T B 107
REGION 7
F
7 TOTAL; 3 EXTENSIVE DAMAGE 371 DEATHS
1899 09 10
1936 10 27
*1946 04 01
60.0
58.6
52.8
140.0
-137.
-163.5
—
8. 2
50 7.40
150.0
32.0
GULF OF ALASKA
S. ALASKA
E. ALEUTIAN ISLANDS
4
4
4
T&L
L
T
C
C
A 247
*1957 03 09 51.3 -175.8 8.30 16.0 CENTRAL ALEUTIAN ISLANDS 4 T A
1958 07 09+ 58.0 -138.8 SH 7.50 525.0 S. ALASKA 4 T&L C 2
*1964 03 27+ 61.0 -147.7 33 8.40 32.0 GULF OF ALASKA-ALASKA PENINSULA 4 T A 122
1965 02 03 51. 178.6 40 8.70 10.0 W. ALEUTIAN ISLANDS 4 T C
1887 05 03 -31.0 109.0 8.0 N. MEXICO 2 T C 44

1902 02 26 13.5 - 89.5 5.0 GUATEMALA-N IC ARAGUA 4 T B 185
1907 04 14 17.0 -100.0 8.00 1.8 S. MEXICO 4 T C 8
1928 06 16 16.3 - 98.0 SH 7.90 2.0 S. MEXICO 4 T C 4
1932 06 03 19.8 -104.2 SH 8.00 2.8 CENTRAL MEXICO 4 T B 425+
1932 06 22 18. 7 -104.7 SH 6.90 6.0 CENTRAL MEXICO 4 T B C
1934 07 17 8.0 -82.5 7.75 2.0 COSTA RICA-PANAMA 4 T C
REGION 9 1
14 TOTAL; 5 CONSIDERABLE DAMAGE; 3 EXTENSIVE DAMAGE; 2.878 DEATHS
*1906 01 31 1.0 - 81.5 SH 8.60 5.0 COLOMBIA -ECUADOR 4 C 500+

*1906 08 16 -33.0 - 72.0 SH 8.40 3.6 S. CENTRAL CHILE 4 T C
1914 01 12 -12.0 - 76.6 3.0 PERU 4 B
1918 12 04 -2 6.0 - 71.0 SH 7.80 5.0 N. CENTRAL CHILE 4 T C
*1922 11 10 -29.0 - 71.0 8.25 9.0 N. CHILE 4 T A A
1927 11 21 -44.6 - 73.0 SH 7.00 2.8 S. CHILE 4 T B
1943 04 06 -30. 8 - 72.0 55 7.90 N. CENTRAL CHILE 4 T B C
1955 04 19 -30.0 - 72.0 30 7.00 1.0 N. CENTRAL CHILE 4 T A 1
1958 01 19 1.5 - 79.5 40 7.50 4.0 COLOMBIA -ECUADOR 4 T C 5
*1960 05 22+ -39.5 - 74.5 8. 30 22.6 S. CHILE 4 T A 1,590
1960 11 20 - 6.9 - 80.8 55 6.75 9.0 PERU 4 T B 13
1966 10 17 -10.7 - 78.8 24 8.00 3.0 PERU 4 T B
1971 07 09 -32.5 - 71.2 58 6.80 1.2 CHILE 2 T C
1979 12 12 1.5 - 79.3 24 7.70 4.9 COLOMBIA -ECUADOR 4 T A 259
DAMAGE CODE:
C = MINOR DAMAGE I.E. FEW HOUSES, BOATS, ONE LOCATION
B = CONSIDERABLE DAMAGE, I.E. 100 HOUSES, ONE LOCATION
A = EXTENSIVE DAMAGE, I.E. 1000 HOUSES, ONE LOCATION OR MANY HOUSES AT SEVERAL LOCATIONS
DEATH CODE: CAUSE: VALIDITY:

C = FEW T = TECTONIC (EARTHQUAKE) 4 = DEFINITE
B = MANY L = LANDSLIDE 3 = PROBABLE
V = VOLCANIC 2 = QUESTIONABLE
*INDICATES TSUNAMIS THAT CAUSED DESTRUCTION OR FATALITIES BEYOND SOURCE REGION

DEATHS AND DAMAGE FIGURES INCLUDE EFFECTS CAUSED OUTSIDE REGION OF ORIGIN
219
"J O' Co 4^ UJ NJ 0 H > z r— p-* — (— r— r- p- p- G O H r -3
G
CO CO VO vO VO vo VO VO so vO vo X EE
x 0 z ON ON VO Cn 4> CO ro ro O G G G PH
G
> X G M z z
x g > X zc CO CO 3
>
z CO 4> O -O ro ON UJ co ro O' Z G
p> gi CO r-3 G
WEST WEST
w
M g
PH 0 Si 2
JDJ
G 0
- 2: s
M
po O OO — O0 O 1 <“ 0 O S
O G
3 t- & > H O
U) Cn Co UJ ro p- CO
z z 0 IN P-t M > 0 > O 3> ?>
G CO x CO 00 O 0 O •— H
5 H
A 1 1 rn CD •O ro ro p-
£ G M Z >
ra CO r-H r* ro vo cn CO P' 0 O' G
COAST Z r 1
CO H G C
COAST
G z > CO >> p z PH
-4
M g > z > s a G
CO £ CO 0 CO CO G H
Z >»
CO O G O G G G G Z G £ rn
O i§
CO > G to• b* >•
£ • • G H a Z
SOUTH NORTH z CO X 0 r z s z 3 O EE O
o **
C CO£ Tl O H O > G O O O G G EC
g CO H 0 0 G G 3 F M G sc G G G O
0 EC EC O H > > G G > G Z O G PH G
r
0 'X "0 C- G H X G H G H G G O
AND 0 PO G •H " O EE H
£ 0 z £ CO CO > £ £ £ G M H G
> 3 0 O 0 F* £ c_ G G G G
z G G Z »> G O Z O Z
AMERICA
CO > 'n pH > G

H G O 0 G G G O rH G J>
£ M
CO O
r-i pu z Z M
G Z —
Z»** G O O
£ £ rn 1 rn Z G H PH
£ z z EE G
CENTRAL
s> > G
Z G G G
|
£ CO
M X £> £ H G
z Z X r G Z G jg EE O
0 Jt>CO £> H G a G Z G G
£ C/3 c_ > G G i> G > G £
Z 0 po H
o X CO £ £ £ £ Z
CO Z X a G H G
a G G
AMERICA
M G G G G
H G
z CO EE G
z G G £ r-3
rn G G
II Z > G H
= =
II II 11 II II II II
G G
CO NJ P-* 0 O O 0 O G
tr
£ Z
c O
P> G G O
3 0 G
G O 3 K G
z z G CO 00 00 00 00 00 00 .O G O >
> § X g M EE 0 X X z • O X
Z z > G G
5 M 0 G
r > ro CO CO ro UJ UJ u> 42 CO co ro
SOUTH NORTH > Z G M 0 2: 2. 2: z G O
CO PH u> IT (J > CO O 0 CD O G G PH
IT z — z
t O CM G PH X ^3 CO • • C Z
£ — G CO
CO c CO rn 0 G a PO 3 C
& G M > CO CO CD G
rH G z r G
1
CO
Z > CO >
G 0 O r< TSUNAMIS
• H pc H *o O
* CO > Z O O G G G z 3 G G G G G PH
£ 3 > G H H ph G z
AMERICA
CO O H G 1 1
> Z G £
O
H g G G G C O 0
G G
/—
z O 2. SS G
CENTRAL
> S O *O Z c pH
O X
,<C Z CO P» O > G PH
O > ro CO CO < z H z
x z G G £ G G
3 3 CO
C J
£ CO rn H C-. H 5 H Grn rn ph 0 G O
H H C O 3 £ rn rn > G Z pc
Z Z G 0 O Z Z GV G
DESTRUCTIVE
Z M
•
£ > 3 G H H > £ G z
H G G 9 H
AMERICA
H po po <* rn
M G
C/3 K > rH G PH
£ G G G G M **
O £ CD G
V*/ Z Z S rn rn G z
p> G M O <*
G G
r—i r<
O > Z G c
M T Z
3 rn
G r -3 z £
c O
II II
M
II II II II II II
=
oO0 0 O 0O 00
3
G
BEYOND
H H S
G
=1 G £ H
G
0 G S
G SOURCE
O' O ON O O' O' ro u>
N. G G M Z
ft* z EC r-3 zpo
CO cn O cn O cn »— p- O' G >
G
i> x z CO CO G
3 y M 0 g g s C3 0 § C REGION
SOUTH
& CO PH > t - 0 *
> z CO CD 8 Z
f z rn z O CM G rH CO pp
C. £ G CO c 0 O O S G
M
CO
03 CO G
G M Z
CO
>
r*
rn
M 3
O O
*3
£
rn
££ w 2:
£££££
C G 2:
G
>
S
2
Z £
O
<*•
Z ;> CO > O > H

G > > > > > >
• >0 rn ph
> z rn rn rn Z 0 G
3 G
» CO O "O >
AMERICA
w CO CJ ?> H > • • rn rn rn »
*-< rn G G £ PH
CO Z O O CM
AMERICA
2 G
[fc»
O
CO
JJ>
rn
O
S
C) CO
2: G G G
|> G £
ro £2
W P-t
rn
G
p-t p H
G
u» CO CO CO O — n G Z O
G G G
1
£ M 0 O O G M > •
H 2: > H ^0 > > > rn w £

p> £ G - G 0 H G r- H
Q Z
Z £ Q O H P < G G > n G
v^/ H 3 G c 0
rn rn
G G
G >
H
0
Z > ?0 G M G 3
>
O £
G £
/->
II II II II 11 II II II
S 8 ^M >•*
O G z H
Z
0O CO t— 0 0 co VO M
O G
rH
O
>
•
G
O' > G
**
£ r> (/jZ
Z O Z ^-N P- G G M 3
po |— w
-
^ G
Z C
G
•
O
> % >
c
G »
> G O £
G £ G
G G > X £
3
£ > Z rn
s: z > G > > H
>
rH
M G r-n
rn O
G £
rn G *H
VO O £
sO G
H G O > O fi
Z
CO Z £ CO O
CE 0 O > C G G
NJ CO z 3 G H G >
O 0 p> G rn £
> H
G Z
CO CO
f G
CO § 8 M O
0 rH z O H
G c G p-« G
CO z ho ro > G Z
O' H G
on O X 2
G
520
*— G _ G —
t H- t— t—t *— h- —
i— H- G
VD VO VO VO VO vO vO vO VO VO VO vO VO vO vO vO VO vo VO vo
N VO vO VO vo VO <
1975
4--O
O' -O
4>
UJ VO
CO •-4
U> O
c_n
v-n
4> tN ro on OO CSV
O av ro VO
on O' 'O
OJ o
v_n VO
ro
4>G
av ro VO av
'O o
G
Z
O o O
4>
O f— O
)— 4^
o
CO
o O O Oo O O O
VO 00 OV ro 'O vO
o
Ov on
O o
CO
t— e-
ro ro 4^
o
10 -o U> N3 fs3
$2
o U3 N3 N3 — o o j— o N> H- N3 ro N3 o ro ro (— H- o ro j— H
ro OO VO 4> o t— 'O 4^
i
o vO vO 4> vO K3 to 4^ G
31
Ov av av av av
G
'O 'O -O "O ''J 'M 'M "N -M av av CSV CSV av av O' On on /D
•
4> K3 JO K> o oO
IM» o O o vO VO CO CO G -M -o on ro o
ro t— -o on 2
o on O O oo oO o O oo o CO o o on o O on o o o On O a
7.40
po
vo ro
o
1— H— cc3
z >
TABLE
fO 4>
• •
00 N3
• •
o u>
• • •
I—*
•
03 ro N3 03 O' t
1
ro vo ro VO vo
• •
On ro oo
• • • c X
oO o O oo oo On CO o CSV o ro oO o o o o o On O 2
4
-
g O X PS CO CO z £ G G X o O G G G 75 G o 3
os G
c G o > z G G G G G [> O
> G
G Ch pa o G
•
PS • • • • DAMAGING
G £ 2 G G r £ z G po G z r r 75 O
> Pi > f— 1
o O o X o o > G > G G G a G > 75o G > > >
F G G G X ^4 > G o GX 3 G pj G G 3 p" G > u> X z G H
PHILIPPINE
g PS G CM > o X z 75 o G > G G C3 G ># G G G G G

a CO o H z C 3 G 75 r z G > G Z > O G G G > o
H z 75 > G po G po
G o G G po z
G CO G G G G > G p>
G >
> 3 > *
P> m G G a G G G z G G G G o
G *r>
C/J
Z > CO G G o r X
G H G a r ?o G
1
G TSUNAMIS
G G > *> O > > 30 > 30

G /> Z Z X G z o ?> z X >> 1
m
(73G Z G G G G a o G a o G G G >
TRENCH
r Z a Z G z G G G G G 75 G Z Z H po
> a <• CO G G > *
X >•
a >•
H
z o G s Z > a PC
a z *T3 G o G G G z G <© WITH
C/3 PS X /> o z a z G z G
CO G - G a o o G o 3>
G G G C-t o G G o 75
i> G G z ?> z Z ?> z G
-X3 CO G G <— G G
2 CO 3> G p> G G
G po z G G G G
z > p> > >
MAGNITUDE
G Z
CO
G G
• £ LESS
X g G o o G G z r O a X > <n o £ < OS H X G z 3 G

I a G
pc
NE.
G z 75
?> > > § C/3 > o o G G G cC S > G 3 O o G ,> o O
r z G po H o H% X O < G G > z Ki G O pc pa 75 G G o z G CD
C3 G G > o G G o o Q a n s
SAMAR
SLA
o H s > .> > > o o > c o G > G o X > THAN
> J> 75 G G o O G G G G G Z G o > 75 > X O G o

G G > G PC g CD o PC H
IQUIQUE,
75 PS G G /> 5> 3*; > G G X r* o G > G

pc O Z z CO 2 G G O G o G G G Z a X G > > G o O
G G z G >
C 3
IS., JUAN
COAST
> Po CO O P> G r H S G 75 3 c > o G 3 H G z

a z po O G & ?> G o Z G >*
G < > Ga >* •* 7.5
> N a G G Z* G < > G O o G G

X > G G Z H o
•
G «*CO *•
o G O £ g G G • * 2 G G G
»#
z > G G
CHILE G M» • § ?0 H > G Z>• m G G o ?> oa Z G H a Z IN
O > 1/3 G CO H o G G C 3 X H G n G G G > PO r > P
PHIL HAWAII
> G • > 75 r
•*
O /O G 3 G G G > O oo PC O u> z G
• G > CD 3 2
g o > G G Z o ?> O G G 3 Z G p> O £ a w
G G z G G > o G r o G Z H G G a
G G C3 G
FERNANDEZ,
p> pa
H G G Z s O G pi z G G Z z CD 20TH
P CM G os G Z
G s G o PC 3 G G :> G PS o G
> O G .> o 'm G G G >
Z z G G K r G Z G
G z
• PO
o pc X O G •G
M G G o < •
G z G CENTURY
CO C/3
CHILE • G • 73
G
G
-
BY
<r> G o
G G
-</> -c/>
— osG o
o G J> 3 u> a O G H G 3 3 G 4> G OO 5 G G DO G os H
2 H G G G G O O
jg
>
G o Z G Q o
> O q
O
30 ro N3 1
Mj
BOATS
$10
O' Q G
X H
**
o
• > CO z G <
z G G o r
G o CS
G o z
pa z < <
G G G > G G H G >
G r H G
G
DAMAGE
Ol H X
CO G G O G
G o H G X po H
S
G
H M oX C3 pc > Z os p X X G G > G G PC G
>•
G »>
O
< o G
z ?C G
G
G > G
z G
C/3
G o G
U> G
z G G r p
?> >
O Z G
G CD o o
G CP
O
G
MAGNITUDE
G
G G G g aG G
HOUSES
G <n 3 g
CO C 3
t
-1
G H a G G G < G G £i> > rs G G G
M z CO CO H 3 G H O > G > 2
MILLION
C/3g i> G 3 z a oo> 3 G G G P

O PS b 2 G X H > G *•
G z g
C/3 2 G C3 G o 2 > G G H
S O o S 2
DAMAGED
z o i> H G > r > >

g G O G > o g PC
o ?> C3 z H as G G O H G o n G H H o G n n 2 O
D PS ?o PS G O o O CO a o F G G G a G G G a G G K o
>> CO G
g O > z > G G j> G o O G o G G CD G G
2 H > >
2
z H Z G a 03 G z G G H H G CP
G
DESTROYED
DAMAGE
,> CO 3 o C/3 »*
G C3 PO G po PO a
o o G G O O C Z O O
g a
•-< 03 G o > oo n K K
PS PS PS CO G o G o o G ri M
G < H O G >
C/3 > a CP a
> i» G G G G G
CO o G C/3 IN G G
H -5 G G
G C/3 C SI
H CP >
G a G
CP G G
CO G
H a
G
o >
K
PS
a
>
CP
G G
a>
G
ro ro
>
H
1
173 ON o S t— o VO O G X
oo On ro CSV 03 o t— o On on o On OO G
321
years
100
region
-
tsunamis
source
beyond
local
aestructive
or
onal
of destructive
regl
Sources
= =
O <9
-
Figure
322
Basin
Pacific
the
in
regions
source
Tsunami
2
Figure
323
years
100
region
source
their
beyond
fatalities
damage
and
causing
damage damage
Tsunamis
=
®=
O
-
Figure
32
Figure 4
NUMBER OF TIME S TSUNAMIS HAVE CAUSED DAMAGE
BY REGION - IOO YEARS
SOURCE REGION
0 1 2 3 4 5 6 CO 9
0 3 1 2 3 3
1 3 1 2
2 15
3 18
4 7 1
5 13 2* 1 2
6 7*
7 7
00 3 1 1
9 1 1 14
* Includes Hokkaido and Noshiro tsunamis
Figure 5
DEATHS FROM TSUNAMIS
BY REGION - 100 YEARS
SOURCE REGION
0 1 2 3 4 5 6 7 8 9
0 35 1 241 61
1 10
2 407
3 5,400
4 5,000
5 36,500 199
6 257
7 115
8 15 676
9 2,618
Key:
0 - Hawaii
1 - South Pacific
2 - New Guinea/Solomon Islands
3 - Indonesia
4 - Philippines
5 - Japan
6 - Kuri 1 -Kamchatka
7 - Alaska & Aleutian Islands
8 - West Coast, North & Central America
9 - South America
525
EARTHQUAKE MAGNITUDE FREQUENCY
FOR DAMAGING TSUNAMIS
1885-1984
MAGNITUDE
(Cumulative Percent)
Figure 6
526
1900-1903
Hawaii,
in
recorded
Tsunamis
7.
Figure
327
TSUNAMIS IN THE JAPAN SEA WITH NUMERICAL CALCULATION
Yoshinobu Tsuji, National Research Center for Disaster Prevention

Tatsuo Konishi, Japan Meteorological Agency
Hiroshi Takahashi, National Research Center for Disaster Prevention
ABSTRAC T 2. NIHONKAI -CHUBU EARTHQUAKE-TSUNAMI (1983-

V-26 M = 7.7, m = 2.5 )
,
Earthquake-tsunamis frequently occur in the northeast

part of the Japan Sea. Among these tsunamis, the Losses of life and property caused by the Ni honkai
Ni honkai -Chubu Earthquake-Tsunami of 1983 (M = 7.7, Chubu Earthquake-Tsunami are listed in Table 2. A
tsunami magnitude m = 2.5), the Kamuimi saki -Oki total of 100 people, out of 104 killed in Japan, were
Earthquake-Tsunami of 1940 (M = 7.5, m = 2), and the the victims of the tsunami. Major damage was done on
Kampo Earthquake-Tsunami of 1741 (M = 7.5, m = 3) are the coasts of Hokkaido Island, Aomori and Akita
discussed. Traces of these big tsunamis were Prefectures. The coasts of Shiinane Prefecture, in
recorded not only at the coast near the epicenters, the western part of the Honshu Island, and of the
but also on remote coasts, such as Noto Peninsula and east coast of the Korean Peninsula were also badly
the Korean Peninsula. A detailed survey was made of damaged. Those coasts are 700 km to 1,000 km,
inundation height of the Ni honkai -Chubu Tsunami along respecti vely , from the tsunami source area.
the coasts of the Japanese Islands and the south part
of the Korean Peninsula. Just after the occurrence of the tsunami , surveys of
effects and interviews of witnesses were made by
Also, a magnetic tape file of the sea bottom topog- scientists and engineers, and several kinds of
raphy has been compiled and the numerical calculation reports were published. Inundation heights of the
of the propagation of the tsunamis for the whole tsunami were measured at about 600 points in total.
Japan Sea has become possible. The surveyed area covers almost all of the coasts
facing the Japan Sea on the Japanese side.

In the present study, we discuss the distribution of
the height of these tsunamis and the result of the Baek (1983) reported the distribution of tsunami
numerical calculations for them. height and the statistics of damages on the Korean
side. One person was killed and two were lost in
1. INTRODUCTION South Korea.
The dates of 21 tsunamis occurring in the Japan Sea, No information has come from the northern part of
in the area off the west coasts of the northern part the Korean Peninsula, and only a few details have
of Honshu Island and Hokkaido Island are given in been gathered for the coast of the USSR.
Table 1. Earthquake-tsunamis have occurred only in a
narrow zone in the sea which runs along the north Figure 2 shows the outline of the distribution of
part of the Japanese Islands. They occurred fre- the tsunami height. The maximum height of 14 meters
quently during the periods of 1741 through 1834 and was reported at Minehama Village, Akita Prefecture.
after 1939. No earthquake-tsunami occurred during The largest peak appears in the north part of Akita
1614 to 1740, and 1834 to 1938. The reason for such Prefecture and smaller peaks appear at Shakotan
a pattern of occurrence is not clear. Peninsula (Hokkaido Island), the northern tip of Sado
Island, outside coast of Noto Peninsula, Uku-Tango
In the present paper, we discuss the distribution of Peninsula, Oki Islands (Shimane Prefecture), and at
tsunami inundation height of the 1983 Ni honkai -Chubu Shimane Peninsula. Prominent peaks also appear at
EarthquakeTsunami , which is compared with the result Imweon Port, Gangweon-Do of South Korea, where
of a numerical calculation. We also discuss the 1741 heights of 3. 6-4.0 meters were reported.
Kampo and the 1940 Kamuimi saki -Oki Earthquake-
Tsunami .
328
Figures 3 through 6 show the detailed maps of the the shelf is a and inundation height at the coast is
distribution of tsunami heights. R, then.
2 ( 4ttL ) (
4ttL ) *
2.1 Numerical Calculation R/a = 2 {J } (i)
o '/go t'
V^ T
Numerical calculations of the tsunami propagation where, L, T, and D are the length of slope, wave
with grid size of 252 X 311 in the mesh intervals of period, and depth at outside boundary of the slope
5 km X 5 km, which covers the whole area of the Japan (200 m). If we put = then a /R converges to
L 0, 2.
Sea were made. The initial sea bed deformation was That is, in the case that the length of slope is
assumed as shown in Figure 7, which had been given by zero, incident wave would be perfectly reflected, and
Aida (1984), who decided on the parameters for two therefore sea water climbs up to the height of twice
faults on the basis of the data of release of stress, the amplitude of incident wave, which is the case of
direct measurements of land, the distribution of the the condition of the computer system. The actual
epicenters of aftershocks, distribution of initial topography of the coasts facing to the Japan Sea has
motion of sea level observed at tide gauge stations, a length of the shelf L of 10 to 30 km in general.
and so on. Influences of the rotation of the Earth, We write the inundation height for the slope length
the surface curvature of the Earth, viscosity of sea being zero as R 0 •, and that for finite slope length
water, non-linearity in the equations of motion, and as R, and the ratio R/R 0 as r, which we call the
of wave dispersion were neglected. As we cannot "amplification ratio." Figure 10 shows the ratio of
simulate the behavior of waves in shallow water, sea amplification for T = 5, 7, 10, 15, and 20 minutes.
areas with depths shallower than 200 meters were It had been indicated that the predominant period of
replaced with seas of 200 m. depth. We assumed that the tsunami is 7 to 10 minutes and therefore, it is
there were vertical walls situated at the coastline, reasonable that the actually observed tsunami height
and that the wave was perfectly reflected there. is three to five times the calculated height.
Calculations were made using the finite difference
method with time steps of 12 seconds for a total time 2.3 Effect of "Yamato Rise"
of 2 hours 40 mi nutes.
In the center of the Japan Sea, there is a
Figure 8 shows the result. Short lines on the graphs remarkable table-shaped shallow sea region called the
show the maximum sea level at the coast. One mesh "Yamato Rise," the top of which is about 300 m depth
interval denotes 50 cm wave height. Tsunami heights while the depth of the surrounding sea is about
on the coast of smaller islands are shown with lines 3,000 m. It has the dimensions of 300 km in the
beginning on the middle column. east-west direction and 150 km in the north-south
direction. The prominent peak of wave heights on the
We can recognize that the distribution pattern is coast of Korea as seen in Figure 2 or Figure 6, and
generally well simulated, except that the absolute the calculation result of Figure 8 is located just at
values do not agree. the counter point of the tsunami source with the
Yamato Rise at the center. We are apt to jump to the
conclusion that the rise acted as a lens for the pro-

2. 2 Amplification of Tsunami Height On
Continental Shelf Region pagation of tsunami (Miyoshi , 1984, Tsuji et al .
1984). In order to test that idea, we made another

As the effect of amplification of a tsunami wave numerical calculation removing the rise and replaced
by the continental shelf region was not taken into by the sea of 3,000 meters depth. We obtained the
account, the absolute value of the calculated tsunami result as seen in Figure 11. The tsunami height on
height is as small as one-third or fifth of the the coast of Korea becomes higher than that in Figure
actual height (Figure 9). Shuto (1972) gave the 8 showing that "Yamato Rise" does not act as a lens
amplification formula under the condition that the at all. We should not use the phrase "lens effect"
depth changes linearly, deeper sea is flat, and that without quantitative verification.
the incident wave is a stationary sinusoidal wave.
We assume that tsunami height at the outside edge of
529
2 . 4 Calculation With Simplified Initial Sea 3. CONCLUSION
Bed Deformation
Comparing the observed tsunami height with the

We made a numerical calculation also under the result of numerical calculations for three
condition that the initial sea bed deformation is earthquake-tsunamis in the Japan Sea, we might
simplified as shown in Figure 12, instead of Aida's conclude that:
model. The result is shown as Figure 13. Comparing
it with Figure 8, we can recognize that on the i) There are several specific coasts where the
coasts, except in the area close to the source, both tsunami height becomes higher than that of the
tsunami height distribution patterns resemble each neighboring coasts for any earthquake tsunami whose
other, which shows that we can calculate the general source is located in any sea region of the northeast
pattern of the tsunami distribution for almost all part of the Japan Sea.
coasts around the Japan Sea except that close to the
source even if we assume the initial sea bed defor- ii) Most of those "specific coasts" are coasts
mation to be simple as that in Figure 12. having the shape of a peninsula protruding in the
direction of the tsunami source region.
2 . 5 Kamuimisaki-Oki Earthquake-Tsunami (1940-

VIII-2 M = 7.5, m = 2)
,
iii) Pattern of the distribution of tsunami
height can be simulated by a calculation using a
Distribution of the 1940 Kamuimisaki-Oki Earthquake- grid in which the shelf region was replaced by a sea
Tsunami height is shown in Figure 14. We assumed the of 200 m depth, and the grid interval is set as 5 km
initial sea bed deformation simplified as Figure 12, x 5 km or so. If we consider the effect of amplifi-
and made a numerical calculation. The result is cation on the shelf region, we might explain the
shown as Figure 15. Peaks of distribution of height difference of absolute values between calculated and
at Rishiri Island, Teshio Port (northern Hokkaido), observed tsunami heights.
appear in both figures, and moreover the distant
coasts of Oki Islands, Uljin Port on the Korean iv) Except for the coast near to the source, the
Peninsula, and Oliga Bay in USSR were hit by a rather distribution pattern of tsunami height can be simu-
high tsunami, which is also well simulated. lated by simplified source deformation model like
that shown as Figure 12.
2 . Kampo Earthquake-Tsunami (1741-VIII-29

M = 7.5, m = 3) v) Without quantitative examination, we should
not try to explain the convergence of tsunami energy
Tsunami source and distribution of tsunami height is with the phrase "lens effect."
shown in Figure 16. The source area was decided by
Hatori (1984). On the coast of the southern district
of Hokkaido, 1,467 people were killed. Recently an

old record from the Korean side was discovered, and
it reported that houses were washed away and vessels
were destroyed in Gangweon-Do Province. The result
of the numerical calculation with the simplified ini-

tial deformation of sea bed is shown as Figure 17.
Higher peaks at Sado Island, Noto Peninsula, and east
of the Korean Peninsula correspond well with the
information of historical documents.
530
Table 1. List of earthquake tsunamis in Japan Sea
4 0 Da t e Earthquake Ep i cen tre Tsunami Renarks
Magn i lude lati N Ion)- E Magn i lude
1 850 7 7 7 0 39, 1 140 O' 2 Yamftgata Prefecture
2 887 VI 1 1 2 6 5 37 5. 138 1 2 several thousands killed
3 1092 IX 13 unrer tain
4 1341X31 Tsugaru Pen Aomori Pref
5 1614 XI- 26 7 7 37 5. 138 0 2 uncertain
6 1 74 1 -VI 11-29 7 5 41 5. 139 4 3 about 1.500 killed, see text
7 1 762-X-- 3 6 6 38 1. 138 7 1 26 houses washed
8 1766 111 8 6 9 40 8 140 6 in Mutsu Hay. Aomori Pref
9 1792 VI 13 6 9 43 6 140 3 2 near Shakotan Pen Hokkaido
10 1793 11-8 6 9 40 7. 140 0 1 Tsugaru Pen . Aonor i Pref
1 1 1804 -VII 10 7 1 39 0. 140 0 1 Kisakata Yamagata Pref
12 1810 IX-25 6 6 39 9 139 9 1? Oga Pen . Ak i t a Pref
13 1833 XII 27 7 4 38 7, 139 2 2 off Yanagata Pref
H 1834- 11-9 6 4 43 3 141 4 - Isikari. Hokkaido
15 1939 V 1 7 0 40 0. 139 8 -1 Oga Pen Akita Pref
IB 1940 VIII 2 7 5 44 1. 139 5 2 10 killed see text
17 1947 XI 4 7 0 4 3 8. 141 0 1 o f f Rumo i . Hokka do i
18 1964 V 7 6 9 4 0 3. 139 0 1 o f f Aomor i Pref
19 1964 VI-16 7 5 38 4 139 2 2 N i i gata Ear thquakc
20 1 983-V 26 7 7 40 4 139 1 2 5 Nihonkai Chtihu Earthquake
21 1983 VI 21 7 0 41. 3 139 2 0 largest aftershoek of No 20
Fig 1 Locations of the Earthquake-
Tsunamis listed in Tablr 1
O. . M S> 7 0. o M ^ 6 9
331
Table 2 Statistics of damages of the N i honka -Chubu Earthquake Tsunami
i
After National Land Agency. Japan and Central Meteorological
Observatory. Republic of Korea
country Japan Korea
prefecture/ dokka do Aomor

i i Akita Sh mane i other total Gangweon Do
prov nee i pre f ec tures Kyeongsangbug Do
persons
killed 4 17 83 0 0 104 1
lost 0 0 0 0 0 0 2
injured 24 25 265 5 5 324 2
damaged houses
entirely &
swept 5 447 1. 132 0 2 1. 584 1
partially 16 865 2. 622 0 2 3. 505 0
si i
ght 1
69 3. 108 2 867 0 0 5. 954 22
submerged
above floor 27 62 65 141 3 298
below f 1 oor 28 152 277 277 8 742 19
damaged
vessels 637 853 681 319 161 2 651 81
Fig. 2 Schematic distribution of inundation height of th** Nlhonkal ' hut u
Earthquake-Tsunami of May 26th, 1983.
532
141 °15'
I
— 45°30'
45°20'
Fig. 3-a Detailed map of Rishiri

and Rebun Islands.
2.9m
Fig* 3-b Detailed map of Shakotan

Peninsula.
•
ilf
100km
Fig. 3 Distribution of inundation height

of the Nihonkai-Chubu Earthquake-Tsunami
along the vest coasts of Hokkaido Island.
Black circles show the locations of

the measured points, and triangles
5.0m show tide gauge station with tsunami
height (cm)
Fig. 3-c Detailed map of Okushiri Island.
333
2 3 m
334
— >*•
555
IS** I40'E
Fig. 7 Initial sea bed displacement
caused by the Nihonkai-Chubu
Earthquake. After Aida(l98U).
Full line shows the area of
upheaval and broken line shows
that of subsidence.
Fig. 8 Result of numerical calculation

for the Nihonkai-Chubu Earthquake-
Tsunami. One mesh corresponds to
50cm of tsunami height. Notice that
the depth in the shelf region is
assumed uniform as 200m.
336
Fig. 9 Comparison of observed tsunami height
with numerically calculated height(h /.
Full line denotes h = h

. , and broken line
obs cal
denotes h obs = 3.0 x h ca i.
To 20 30 So
- 50
SLOPE LENGTH! KM)
337
Sakhalin
Is .
Fig. 11 Result of numerical calculation of the Nihonkai-

Chubu Earthquake-Tsunami under the condition that Yamato
Rise is removed.
338
Fig. 13 Tsunami height distribution for the symplified model.
239
Sakhalin Is.
Fig. lb Distribution of height of the Kamuimihaki-Oki Earthquake-

Tsunami of 19^0-VIII-2, M = 7.5, m = 2. unit:m.
the Kamuimisaki ki
Result of the numerical calculation of
1
Fig. 15
Tsunami.
340
Earthquake-
of the Kampo
16 Dl3tr i'b n , heisht - ‘ 3-
Fig. ;° 1-29! M = 7.5,
f TO
Tsunami of lTtl-vii
T
341
"•
. ,
V.
f ,
Theme IV
U.S.-JAPAN COOPERATIVE
RESEARCH PROGRAM
RECOMMENDED U.S.-JAPAN JOINT RESEARCH
ON
SEISMIC BEHAVIOR OF BURIED PIPELINES AND TELECOMMUNICATIONS SYSTEMS
BY
M. SHINOZUKA
ABSTRACT
The U.S.- Japan Workshop on Seismic Behavior of systems place the highest priority on the follow-
Buried Pipelines and Telecommunications Systems ing areas of research.
was held under the sponsorship of the U.S. Na-
tional Science Foundation with Dr. K. Thirumalai 1 . 1 Definition of Ground Motions Including Ground
as Program Director in cooperation with Task Strain Measurements for Lifeline Engineering
Committee F Disaster Prevention Methods for Ground motion information for lifeline engineer-
Lifeline Systems (Co-Chairman Dr. W. W. Hakala, ing purposes is relatively scarce. Acquisition
and Dr. T. Iwasaki, PWRI) ,
UJNR Panel on Wind and development of information to define a proper
and Seismic Effects (Co-Chairmen Drs. R. N. intensity of ground motions explicitly for Life-
Wright, NBS, and Dr. R. Iida, PWRI) at PWRI, lines should advance the state-of-the-art of
Tsukuba Science City, Japan on December 5-7, lifeline engineering and reliability analysis.
1984. Nippon Telegraph and Telephone's cooper- The deployment of dense strong motion instrument
ation is also noted. The workshop reviewed the arrays, e.g., at PWRI and elsewhere In Japan,
current knowledge, analytical, experimental and provide data to improve analytical models for
field observation, of seismic effects on buried predicting the relative motions and ground
pipelines and telecommunications systems, and strains
presented technical papers by the participants
on the subject matters. This paper summarizes The proposed joint activity is to expand existing
the recommended U.S. -Japan joint research as a Japanese dense arrays for direct measurement of
result of the Workshop. ground strains and response of representative
lifeline configurations (e.g., welded steel
Keywords: Telecommunications Systems; Buried pipelines, jointed pipelines, conduits, etc.).
Pipelines; Ground Motions; Seismic The cooperative program should also include the
Analysis; Reliability Analysis following related tasks: (a) data reduction and
analysis (including spectral and correlation

1. RECOMMENDATIONS studies) of available and future array data to
On the basis of the paper presentations, compre- develop the power spectral density of differen-
hensive discussions were conducted during the tial motions as functions of separation distance;
resolution session to reach a consensus on the (b) the development of related attenuation rela-
research topics that need immediate attention tionships as a function of earthquake magnitude
and to identify the specific research areas that and epicentral distance; and (c) seismic hazard
are most suitable for future coordinated reanalysis techniques should be developed to eval-
search between the U.S. and Japan. The discus- uate hazard curves in terms of the pertinent
sions led to the recommendation that, in the im- intensity scale. The proposed improved defini-
mediate future, coordinated research and infor- tion of ground motions is of greatest signifi-
mation exchange between the U.S. and Japan dealing cance for disaster preparedness that require re-
with buried pipelines and telecommunications liability of lifeline systems; for this purpose,
m
the demand for such information is critical and mile with peak acceleration of MMI suggests that
urgent. only some damage appears immediately after the
earthquake. More damage, often extensive, ap-
1 .2 Improvement of System Serviceability of pears later. We propose to extend examination of

Telecommunications Field Facilities And leak rates in water and sewer pipes by including
Equipment Under Seismic Conditions By More Japanese records and by collecting and interpret-
Advanced Earthquake-Resistant Design ing data on post-earthquake performance. Data
It appears that the current contingency plans collection and interpretation would be performed
for telephone services in the case of an earth- jointly by the US and the Japanese groups. The
.5
quake emergency are largely dependent upon the specific objectives of the joint project are pri-
routing redundancy of networks. While this is marily information exchange on (a) collected data
probably adequate enough in most cases, the on leak rates in water and sewer lines before,
following questions remain: (1) are such plans immediately after and for a longer period after
possible for all physical equipment and outside an earthquake; (b) observed correlation between
facilities? (2) is the common cause nature of post- and pre-earthquake performance to help un-
earthquake events reflected in such plans? and derstand how pre-earthquake condition of the
(3) has a cost-benefit analysis been carefully pipes influences seismic performance.
performed? A joint research project to address
these questions is proposed. 1 Model and Prototype Testing of Soil Restraint
on Buried Pipelines in Large Displacement
1 .3 Analytical and Experimental Investigation of Algorithms for modeling soil restraint on buried
Continuous and Segmented Buried Pipelines in pipelines subjected to large ground distortions
Liquefield Areas (e.g., faulting, landslides) are based on a lim-
One of the most important geotechnical hazards ited number of model tests of soil-pipeline in-
for buried pipelines involves those associated teraction and selected anchor pull-out tests re-
with liquefaction. Liquefaction induced move- ported in the literature. Although these algor-
ments include (a) lateral spreading, (b) flow ithms have been used extensively in nonlinear
failure, (c) loss of bearing, and (d) subsidence finite element models, the need for verification/
and buoyancy effects. At present in the US, calibration is well-recognized. A comprehensive
little work is devoted to the behavior and fail- program of model testing (possibly some centri-
ure mechanism of pipelines in liquefield regions, fuge and a soil box) is recommended. In addition
even though it has been recognized that there several well-planned field tests on prototype-
exists a great need for analytical and experi- size pipeline segments are needed to validate
mental research on this subject matter. The model test results. Even though no clearly de-
behavior and failure of both continuous and seg- fined faults exist near-field in Japan, it should
mented pipelines should be determined and de- be mutually beneficial for the US and Japan to
sign procedures must be developed. In this re- pursue a joint program in this area which could
spect, a joint project is proposed in which involve a number of university and research org-
technical data be exchanged and coordination of anizations in both countries. Since the proposed
analytical and experimental efforts by the US testing requires a substantial effort and budget,
and Japanese groups established. our immediate goal is to exchange experimental
data so that we can maximize the utility of our
1 .4 Seismic Performance of Water and Sewer current resources and at the same time, begin to
Pipelines coordinate our experimental efforts in the US and
Present work to correlate number of leaks per Japan.
345
1.6 Study of Current Seismic Protection Methods
For Telecommunications, Related Electronic,

and Peripheral Equipment Supported on Raised
Floors
The objective of the proposed research project is
to provide better seismic protection for sensi-
tive telecommunications, electronic and computer

and other equipment supported on raised floors,
possibly in a clean room environment. These
types of equipment are expected to remain opera-
tional during and immediately after a strong
earthquake. As a matter of fact, during the
1970s, lateral bracing of equipment frames was
widely adopted as a cost-effective method of
stiffening against toppling. However, the in-
creasing reliance on computer-based telecommuni-
cations justifies considering alternative pro-
tection systems, including base isolation for
raised floor systems. We propose a joint pro-
ject to identify more cost-effective support

systems among various alternatives. The steps
we might follow in this project are: (a) iden-

tify equipment which can be advantageously sup-
ported by alternative systems; (b) develop re-
quirements for alternative systems; (c) develop
and screen analytically candidate systems for
their cost-effectiveness; (d) construct and
test the most promising system by a shaker

table; and (e) investigate if active control
concepts might be applicable.
346
EXPERIMENTAL INVESTIGATION ON HYSTERETIC
BEHAVIOR OF REINFORCED CONCRETE BRIDGE PIER COLUMNS
1
BY
Toshio Iwasaki- -, Kazuhiko Kawashima , Ryoji Hagiwara-*, Kin;ji Easegawa ,
A C.
Tatsuhiko Koyama , and Takeshi Yoshida^
ABSTRACT
Substantial damage to reinforced Earthquake Disaster Prevention

concrete piers were observed during Countermeasures of Highway. Although
recent earthquakes in Japan and these seismic resistant design by means of
failures were most likely caused by seismic coefficient method has been used
large lateral seismic forces experienced for highway bridges, substantial damage
by the bridge. It is therefore to reinforced concrete piers were
necessary to evaluate nonlinear observed in recent earthquakes such as
hysteretic behavior of reinforced the Miyagi-ken-oki Earthquake of 1978
concrete piers. This report summarizes and the Urakawa-oki Earthquake of
the experimental studies on dynamic 1982 (refer to Photos 1, 2 and 3).
strength and deformation characteris- Such failures were most likely caused
tics of reinforced concrete piers, by lateral seismic forces, experienced
which were conducted between 1981 and by the bridge during large earthquakes
1984 fiscal years at the Earthquake of a short epicentral distance to the
Engineering Division, Public Works site, which are far larger than the
Research Institute. Tentative seismic force currently considered in
conclusions obtained through the the design in terms of a seismic
past studies as well as future experi- coefficient of 0.2 to 0.3.
mental research programs are presented.
Because it is anticipated that
1. INTRODUCTION substructures including piers would
experience large shaking when
Japan is one of the most earthquake-prone subjected to severe ground motions,
countries in the world, and it is of the bridge should be designed
particular importance to assure safety to avoid the collapse of the primary
of highway bridges against seismic bridge system such as falling-off of
disturbances in accordance with the
Director, Earthquake Disaster Prevention Department, Public Works Research

Institute, Ministry of Construction, Tsukuba Science City, Japan.
2
Head, Earthquake Engineering Division, Earthquake Disaster Prevention
Department, Japan.
3
Research Engineer, Earthquake Engineering Division, Japan.
* Research Engineer, Earthquake Engineering Division, Japan.
5
Assistant Research Engineer, Earthquake Engineering Division, Japan.
the superstructure from the substruc- to static and/or dynamic loadings at
ture. To maintain the integrity of the the top by an electro-hydraulic servo
bridge during severe earthquakes, it actuator (the maximum loading capacity
is particularly important to evaluate of +125 tf and the maximum stroke of
nonlinear hysteretic behavior of rein- +125mm) The applied load to the pier
.
forced concrete piers. simulated the force which would occur

at the pier top during an earthquake
Because, up to now, much of seismic including the effects of the inertia
resistant design has been conducted on force due to the mass of super - and
the basis of elastic behavior, substructures. Such a loading
comparatively few investigations have condition is appropriate for testing
been made on nonlinear hysteretic of bridge piers in which the inertia
behavior of reinforced concrete piers. force associated with the superstructure
Consequently, comprehensive experimental is pronounced.
investigation on dynamic strength and
deformation characteristics of reinforced Reversed-cycl ical lateral force was
concrete piers in inelastic range was applied to the pier with a loading
initiated at the Earthquake Engineering time history of displacement as shown
Division of the Earthquake Disaster in Fig. 2. Up to the yielding of the
Prevention Department, the Public Works longitudinal reinforcement, the load
Research Institute, in 1981 using the was increased step by step to about
Dynamic Structural Testing Facility. 10 increments. The displacement
developed at the top of the pier when
This report summarizes the experimental the longitudinal reinforcement
studies performed between 1981 and 1984 exhibited yielding is defined
fiscal years, and tentative conclusions hereinafter as the yield displacement,
5
obtained from the studies as well as y, and is used as a reference value
future experimental research program to specify the displacement loading
needs. subjected to the pier. After initial
yielding of the longitudinal reinforce-
2. EXPERIMENTAL INVESTIGATIONS CONDUCTED ment, displacement loading with n-cycles
BETWEEN 1981 AND 1984 FISCAL YEARS were applied to the pier by successively
increasing the displacement amplitude
2.1 Ex perimental Procedure and Specimen as ^y, 2 y, 3 y, . According
of Model Pier to such loading scheme, envelopes of
the load-displacement hysteresis loops
Model specimens of reinforced concrete can be obtained as shown in Fig. 3.
pier were fixed at the footing which in Although there are many definitions
turn was tied down to the reaction floor on failure of a structural element,
as shown in Fig. 1. The footing was in this study the pier is assumed to
designed as large as possible and have failed when the strength becomes
stirrups were densely spaced so that smaller than the yield strength after
failure in the footing would be the pier had attained its maximtr
prevented. The pier model was subjected strength.
348
The scale of the specimens used in flexural failure with loading velocity
this study was approximately 1/5 of as high as 70cm/sec. However, hysteretic
typical highway bridge piers, and the damping ratio increases with the
standard cross section was selected increase in loading velocity after
as 50cm x 50cm. Table-1 summarizes yielding of longitudinal reinforcement
the dimensions and loading procedure (after approximately 2 -y)
of the 34 model piers tested in this

study. It should be noted here that 2.3 Effect of Dynamic Loading
although bridge piers are subjected to an d L o n g itu d i n a l
.
nominal axial stress of 5 - 20Kgf/cirr Re in fQEcero ent Rat i o . .1 1 9 82

associated with dead weight of super- Fiscal Year)
structure, it was disregarded in this
experiment, because their effects are To investigate the effects of
relatively insignificant. dynamic loading and longitudinal
reinforcement ratio on flexural
2.1 Effect of Loading Velocity failure of reinforced concrete piers,
(1981 Fiscal year) both static loading test (loading
velocity of 2.4cm/sec.) and dynamic
To investigate the effects of loading loading test (loading velocity of
velocity on the dynamic strength and 25cm/sec.) were conducted on model
deformation of reinforced concrete piers with longitudinal reinforcement
piers, dynamic loading tests were ratio of 1.79%, 0.87% and 0.48%.
conducted by varying the loading Fig. 7 represents six specimens used
velocity as 17.5cm/sec., 35cm/sec. and for the experiments (Model P-4 -
70cm/sec. For this purpose, three P-9, refer to Table-1).
model piers with the same dimensions as
shown in Fig. 4 were used (Model P-1 Figs. 8 and 9 show the effect of
P-2, P-3, refer to Table-1). dynamic loading and longitudinal
reinforcement ratio as shown by the
Figs. 5 and 6 show the effects of envelope of load-displacement
loading velocity in terms of the hysteresis loops and the equivalent
envelope of load-displacement hysteresis hysteretic damping ratio. Table 3
loops and the equivalent hysteretic summarizes the dynamic performance of

damping ratio. Table-2 summarizes the six specimens.
dynamic performance of the three
specimens. The conclusions derived from this
series of tests are as follows:
The conclusion derived from this
series of tests is as follows: 1) The yield strength, the maximum
strength and yield displacement
1) The variation in loading velocity on associated with flexural failure
the maximum strength and ductility developed at the bottom of the pier
factor (displacement ductility factor) increased with increase in longitudinal
is not significant in the case of reinforcement ratio.
349
2) Although significant differences factor in case of flexural failures
between dynamic loading and static developed at the bottom of the pier.
loading are not observed, the maximum
strength, ultimate displacement and 2.5 Effect of Cut-off of
ductility factor in the dynamic Longi tu din a l R einforc e ment
loading were slightly larger than 11983 Fiscal Year)
those in the static loading.
To investigate the effects of cut-off
3) Plastic hinge region, in which of longitudinal reinforcement, dynamic
extensive spalling and cracks of loading tests were conducted on three
concrete were developed, increased model piers (Model P-14, 15 and 16,
with the increase in the longitudinal refer to Table-1) . The bars were
reinforcement ratio. cut-off at the height of 110cm, 135cm
and 160cm from the bottom as shown in
2.4 Effe.c.t-Q£..Tifi. P g in fprce- Fig. 13. Because the height from the
ingn b ..Rflt.3LQ...(.J-.983...Fisgal bottom of pier from which longitudinal
Year) reinforcement can be reduced was set
at 110 cm, the height of cut-off at
To investigate the effects of tie 110cm, 135cm and 160cm corresponds,
reinforcement ratio on flexural failure respectively, to no anchoring length
of reinforced concrete piers, dynamic (£=0), anchoring length of half width
loading tests were conducted on four of pier (£-d/2, d: width of pier),
model piers (Model P-10 - P-13, refer and anchoring length of pier width (
to Table-1) . The tie reinforcement =d) . Dimensions of the model P-14,
ratios ranged from 0.1% to 0.3%, and 15 and 16 were the same with those of
the height of pier from the bottom model P-10 (no cut-off) so that the
where tie bars were placed densely was effect of cut-off could be investigated
also taken as a variable as shown in by comparing the two results.
Fig. 10.
Figs. 14 and 15 represent the effect
Figs. 11 and 12 represent the effect of of cut-off as shown by the envelope
tie reinforcement ratio as shown by the of load-displacement hysteresis loops
envelope of load-displacement hysteresis and the equivalent hysteretic damping
loops and the equivalent hysteretic ratio. Table 5 summarize the dynamic
damping ratio. Table-4 summarizes the performance of the four specimens.
dynamic performance of the four
specimens. The conclusions derived from this
The conclusion derived from this test
series is as follows: 1) Flexural failure occurred at the
bottom of pier in model P-10 (no cut-off)
1) Tie reinforcement ratio between and P-16 =d) (In the model P-15
. (
0.1% and 0.3% does not have significant =d/2), although cracks were initiated
effect on dynamic strength and ductility both at the bottom of pier and at the
390
cut-off point, flexural failure was pier with a shear-span ratio of 2.2.
predominantly developed at the bottom
at failure. Model P-14 (£=0) failed 2) Hysteretic damping ratio of the
due to progress of diagonal cracks at model pier with ductility factor of
the cut-off point. 2.2, which failed due to shear failure,
is substantially smaller (50 ^ 60%)
2) Equivalent damping ratio and than that of the model piers with
ductility factor were very small in shear-span ratio of 3.8 and 5.4.
model P-14 (£=0), which failed at the
cut-off point, as compared with the 3) Ductility factor decreases in
other model piers. accordance with decrease in the
shear-span ratio.
3) Failure at the cut-off point
could be prevented if the anchorage 2.7 Effect of Number of Loading
length of longitudinal reinforcing (1983 Fiscal Year)
bars are greater than or equal to the
width of the pier. To investigate the effects of number
of loading cycle, n, dynamic loading
2.6 Effect of Shear-Span tests were conducted on seven model
Ratio (1983 Fiscal piers (Model P-10, 17 % 22, refer to
Year Table-1) as:
• model P-10, 19 and 22 (shear-span
To investigate the effects of shear-span ratio of 5.4) n=3, 5, and 10.
ratio, dynamic loading tests were • model P-17 and 20 (shear-span
conducted on three model piers (model ratio of 3.8); n=3 and 10.
P— 10, 17 and 18, refer to Table-1). •
model P-18 and 21 (shear-span
The
1) shear-span ratio was 5.4, 3.8 and ratio of 2.2); n=3 and 10.
2.2 as shown in Fig. 16.
Fig. 19 represents dimensions of the
Figs. 17 and 18 represent the effect of seven specimens.
shear-span ratio in terms of the
envelope of load-displacement hysteresis Figs. 20 and 21 show the effect of
loops and the equivalent hysteretic number of loading in terms of the
damping ratio. Table-6 summarizes the envelope of load-displacement
dynamic performance of the four hysteresis loops and the equivalent
specimens hysteretic damping ratio. Table-7
summarizes the dynamic performance of
The conclusions derived from this the seven specimens.
The conclusions derived from this
Flexural failure was developed at series of tests are as follows:
the bottom for model piers with a
shear-span ratio of 3.8 and 5.4 whereas 1) In the case where flexural
shear failure occurred in the model failure developed at the bottom of
3a
pier , the effect of loading cycle, n, directional loading could be investigated
on hysteresis loops is not significant in comparison with the test results of
until spalling of concrete cover Model P-10 and 22.
occured. However,'- after the spalling
has developed, breaking of longitudinal Fig. 23 and 24 show the effect of two
reinforcing bars occurred at early directional loading in terms of the
stages as the number of loading cycle envelope of load-displacement hysteresis
increased. This, in turn, caused loops and the equivalent hysteretic
substantial reduction in the resistance damping ratio. Table-8 summarizes the
to lateral load. The hysteretic dynamic performance of the four specimens.
damping ratio is also significantly
affected by the number of loading cycle The conclusions derived from this
after spalling of concrete cover, and study are as follows:
it became larger for a small number of
loading cycles. 1) Spalling of concrete cover, and
breaking of longitudinal reinforcing
2) The effect of the number of bars as well as the reduction in
loading cycle on load-displacement strength occurred earlier in the
hysteresis loops and equivalent two-directional loading than in the
hysteretic damping ratio is much more one-directional loading. Ductility
pronounced in shear failure than in factor under the two-directional
flexural failure described above. loading was about 60 80% of that
under the one-directional loading.
2.8 E££figfc..a£..Tttfi..R 3 LcfigtiQna 3 ...LfiadjLn.g However, the ultimate strength of
j_l£8_3__F i£c.a.l_y.£ajJ pier was not significantly effected
by the two directional loading.
To investigate the effects of two
directional loading on flexural failure 2) Effect of the number of loading
of reinforced concrete piers, dynamic cycles under the two-directional
loading tests were conducted for two loading is the same with that under
model piers (Model P-23 and 24). The the one directional loading as
model piers were subjected to the described in the preceding section.
loading condition as shown in Fig. 22,
in which the number of loading cycle, 2.9 Effect of Shear Reinforcement
n, was also taken as parameter (n=5 and Squaie Cioss
fp.i--Eifei.s..Yath
10). It should be noted here that two Section (1984 Fiscal Year)
directional loading was not applied
to the pier simultaneously, but applied The effectiveness of shear reinforce-
alternatively due to the limitation of ment against brittle shear failure of
the experimental facility. The reinforced concrete piers with short
dimensions of Model P-23 and 24 were shear-span ratio was investigated
the same as Model P-10 and 22, which using three model piers with a
were tested under one directional shear-span ratio of 2.5 (Model P-25,
loading, so that the effect of two 26 and 27, refer to Table-1). The
352
shear reinforcement was provided in 2.10 Effect of Cross Sectional Shape
three ways, i.e., placing tie bars (1984 Fiscal Year)
densely at 5cm intervals (Model
P-25), spiral hoops with a pitch of Dynamic tests were conducted on three
2.5cm (Model P-26), and diagonal model piers having a circular cross
reinforcing bars with a reinforcement section to investigate the effect of
ratio of 25 percent of the longitudinal cross sectional shape on the lateral
reinforcement ratio (Model P-27) load resistant capacity (Model P-28,
Fig. 25 presents the dimensions of 29 and 30, refer to Table-1). The
model piers studied. shear-span of the specimens were 4.7,
3.3 and 1.9 as shown in Fig. 28. The
The effectiveness of shear reinforcement cross sectional area, height, longitu-
is compared with the test result of dinal and tie reinforcement ratios, and
Model P-18, which had a shear-span loading procedure were the same as
ratio of 2.2 and ties spaced at 25cm on the model P-10, 17 and 18 (square cross
centers section) described in the preceding
section. It should be noted here that
Figs. 26 and 27 represent the effective- because the diameter of round piers
ness of shear reinforcement in terras of and the width of square pier are
the envelope of load-displacement different when the cross sectional
hysteresis loops and the equivalent area is taken as the same, apparent
hysteretic damping ratio. Table-9 shear-span ratio is not identical
summarizes the dynamic performance of between the round pier and the square
the four specimens. pier
The conclusions derived from this Figs. 29 and 30 represent the effect
series of tests are as follows: of cross sectional shape on the seismic
behavior of piers in terms of the
1) Placing tie bars densely, spiral envelope of load-displacement
hoops and diagonal reinforcing bars are hysteresis loops and the equivalent
all effective for increasing ductility hysteretic damping ratio. Table-10
factor of piers with small shear-span summarizes the dynamic performance of
ratio. However, shear reinforcement the six specimens.
has no effect in increasing the maximum
strength of the piers. The conclusions derived from this
2) Among the three procedures for shear
reinforcement described above, spiral 1) In the case where flexural
hoops is most effective in increasing failure is developed at the bottom, the
ductility. effect of cross sectional shape
between circular and square is not
The effectiveness of spiral hoops for significant if cross sectional area,
round pier will be described in section height, and longitudinal and tie
2 . 12 . reinforcement ratios are the same.
353
2) In the case where shear failure significantly in the piers subjected
occurs, the maximum strength is to the loading in the direction
substantially larger in the round pier rotated to the axis of cross section.
than in the square piper.
2.12 Effect of Spiral Hoops of
2.11 Effect of Loading Direction Eie t w 3.th C jlLGulai.-Ciosg
.
U9.8 4 Fiscal Tear.)

.
Section (1984 Fiscal Year)
Three model piers (Model P-10, 33 and Two model piers with shear-span ratio
34, refer to Table-1) were tested of 4.7 and 3.3 (Model P-31 and 32, refer
under dynamic loading to evaluate to Table-1) were tested to investigate
effect of loading direction on the the effect of spiral hoops on the
seismic behavior of piers. Fig. 31 seismic behavior of piers. They were
shows the three specimens in which reinforced by spiral hoop with a pitch
shear-span ratio, longitudinal and of 2.5cm within the range of 50cm in
tie reinforcement ratios, and cross height from the foot as shown in
section are the same. However, the Fig. 34. Effect of such spiral hoop
principal axes of the cross section was investigated in comparisons with
was rotated with respect to the the test of Model P-28 and 29, which
footing at an angle of tt/6 and tt/ 4 in have the same dimensions and shear-span
the Model P-33 and 34, respectively. ratios as Model P-31 and 32, respectively,
Only the principal axes of the pier and are reinforced with ties at 25cm
were related, while the position of on centers.
the actuator remained parallel to the
major axes of the footing. Figs. 35 and 36 represent the effect
of spiral hoops in terms of the envelope
Fig. 32 and 33 represent the effect of load-displacement hysteresis loops
of loading direction in terms of the and the equivalent hysteretic damping
envelope of load-displacement hysteresis ratio. Table-12 summarizes the dynamic
loops and the equivalent hysteretic performance of the four specimens.
damping ratio. Table-11 summarizes the
dynamic performance of the three The conclusion derived from this
specimens series of tests is as follows:
The conclusions derived from this 1) Significant increase in ductility

series of tests is as follows: factor and hysteretic damping ratio
can be achieved by placing spiral
1) In the case where flexural failure hoops having a pitch of 2.5cm for both
is developed at the bottom of pier, the shear-span ratio of 4.7 and 3.3 as
effect of loading direction is less compared with the reinforcement with
significant until spalling of concrete ties. The effect of spiral hoops on
cover. However, after the spalling of the maximum strength is small.
concrete cover, reduction of the
resistance to lateral load occurred
354
3. FUTURE EXPERIMENTAL PROGRAM longitudinal and transverse directions
to the bridge axis. It is, therefore,
3.1 Investigation on Structural necessary to investigate the effect
Characteristics of twisting loading on the dynamic
performance of reinforced concrete piers.
(1) Effect of Cross Sectional Shape
(Hollow and Rectangular) (2) Effect of Loading Hysteresis
Following the study described in The loading time history as shown in

section 2-10, the effect of cross Fig. 2 has been adopted in the experi-
sectional shape, hollow and rectan- ments, i.e., reversed cyclic loading
gular which are frequently adopted at with n repetitions was applied to the
sites, on the seismic behavior of pier, and amplitude of displacement is
piers needs to be investigated. increased monotonically . Because it
became apparent that the dynamic
(2) Effect of Diameter of Longitudinal performance of reinforced concrete
Reinforcement piers is significantly affected by the
history of loading, it is necessary to
It is necessary to investigate the investigate such effects on the dynamic
effect of diameter of longitudinal performance of reinforced concrete
reinforcement, which affects buckling piers
fracture of longitudinal reinforcing
bars, and height of plastic hinge (3) Effect of Force-controlled Loading
zone.
In the experiments presented in Chapter
(3) Scale Effect 2, the bridge piers were subjected to
displacement-controlled loading. In
The model piers, which have been used this loading procedure, no matter how
for the experimental study, had a much the pier sustained damage, it was
(1)
scale factor of about 1/5 in length forced to return to the original rest
to typical prototype highway bridge position. Thus, progressive failure
piers. Therefore, it is necessary to of the pier in one direction can be
assure the scale effect on dynamic avoided. However, the actual inertia
performance obtained from the model force developed in the pier during severe
pier tests. earthquakes is more likely to be
simulated by force-controlled loading.
3.2 Investigations on Loading It is, therefore, necessary to assure
Characteristics the difference of ultimate failure
modes developed under displacement-
Effect of Torsional Load controlled loading with those developed
under force-controlled loading.
Bridge piers such as skewed and curved
bridges are subjected to twisting load
as well as lateral load in the
355
(4) Comparison with Shaking Table Public Works Research Institute,
Tests Vol. 159, 1983.
During an actual earthquake, the 2. Narita, N., Murakami, M. and

structural response of a bridge depends Asanuma, H.: Report of the
on deterioration of stiffness and damping Investigation on Earthquake Damage
characteristics of piers, which, in Shizunai Bridge, 15th Joint Meeting,
turn, affects the lateral loading U.S. - Japan Panel on Wind and
subjected to piers. It is, therefore, Seismic Effect, U.J.N.R., Tsukuba,
necessary to take such hysteretic Japan, 1983.
behavior into account in evaluating
failure modes of the pier. From this 3. Kuribayashi, E., Iwasaki, T.,
point, it is necessary to conduct Hadate, T. and Hagiwara, R.:
shaking table test and assure the test Experimental Studies on Seismic
results presented in the preceding behavior of Structural Members
chapter in comparison with shaking Using a Dynamic Structural Testing
table tests. Facility At PWRI, 14th Joint
Meeting, U.S. - Japan Panel on Wind
4. CO NCLU D I NG JEffiM and Seismic Effects, U.J.N.P.,
Washington, D.C., U.S. A., 1982.
This paper summarized the experimental
study, conducted between 1981 and 4. Iwasaki, T., Hagiwara, R., Hasegawa,
1984 fiscal years, on dynamic strength K., Koyama, T. and Yoshida, T.:
and deformation characteristics of Model Experiments on Dynamic
reinforced concrete bridge piers. It Behavior of Reinforced Concrete
1.
is necessary to promote the experimental Bridge Pier Columns, 16th Joint
study for the remaining problems from Meeting, U.S. - Japan Panel on Wind
both structural and loading character- and Seismic Effects, U.J.N.R.,
istics point of view. It is further Washington, D. C., U.S. A., 1984.
needed to conduct precise examination
on the important factors, which have
pronounced effects on the dynamic
behavior of piers to get rational
seismic design procedure for reinforced
concrete piers of highway bridges.
5. REFERENCES
Saeki, S., Ota, M., Kobayashi, S.,

Shioi, Y., Kawashima, and
Nishikawa, K.: Damage of Bridges,
Chapter 9 of Report on the Disaster
caused by the Miyagi-ken-oki
Earthquake of 1978, Report of the
556
Photo 1 Damage of Pier 6 in the Sendai Bridge by
Miyagi-ken-oki Earthquake of 1978
Photo 3 Damage of Pier 6 in

the Shizunai Bridge by
Urakawa-oki Earthquake
of 1982
357
CO X ,
1— ON £ cj
O * 00 00 03 4-1 CD m
Ob 00 o on o o — £
1
O CO Mi CN o
X LO CO CN 4—1 CD CN Ln •H £ •
o o o
f 1
PM CN NO Q Pd Mi b CN Q O CN
<t CO CO o CD •rH
CM U rH 4—1
CO X 1
1
— ON 0 CJ
o r 00 r. 00 03 4-1 CD <1-
m o o o £ O o o
00 oo ON r— Mi
X CO 1 CN 4-1 qj CM ON •H £ CN rH
o o
1
P-1 o CNJ NO Q cd Mi
o
e CN o O
<t CO CO <U •H
CM u i— 4-1
NO X .
1
— ON £ U
o o
r» n- •N 00
o
03 4-1 CD o
m O o o
00 On 00 — £ 00 Mi
m
i
X CO CN 4-1 QJ CN in *H £ rH
o o o
i 1
PM CX) NO Q Pd Mi b CN o O CN
<t CO CO O CD •rH
CM CJ 4-1
X "
T
NO £ CJ
o o
T— ON oo
o
03 4-1
o O <D
vD CO
X m <r
ON
o
r* 00 r r—
4-1
£
OJ CN ON *H
Mi
£ CN
o
r-4
o o
r
o
1
CM CNJ NO CO CN Mi b CN Q O
<r a cd O CD •rH
CO CO CM U r— 4-1
Procedures
ON
— X 1
ON c CJ
o O
1 1
* ON r 00 03 4-1 QJ
in 00 o ON o <r o i
— £ 1 o 00 Mi
m o
X LT) -<3- CO CN 4-1 CD CN in •H £ f-H
o o
1 i
CM CnJ NO a r— Cd Mi B CN n O CN
CO CO O CD •rH
&. U i — 4—1
Loading
X 1
ON a CJ
o
00 o ON
1
—
r
1 ON ON
r>
oo
o r—
03 4-1
£ o O QJ
M
<r
o
X in o
m o
Ml CD CN ON •H £ CN rH
o
i i
and CM CNJ NO i — CN Ml B CN a O
Q pd o CD •rH
CO CO CM CJ f-H 4—1
ON X
o O
1
1 — CN ON 00 C u
CO oo o ON
o
r* oo r o 03 4-1
£
O QJ
m o
H
oX
Dimensions
LO <r CN r— Mi
f
CM OJ NO co
i
O CN o 4-1
1
qj m •H c
Q cd Mi b Q o
CO CO O (D
CM u rH 4-1
of
X \
ON c U
O 1 — CN ON 00 03 4-1
O o 0) O o
<N 00 O ON * oo r>
o — £
i Mi
m
1
o
X LO -cf
NO
o
CO
i
o CN o
4-1
m b 91
CN f-H
CO
•r-4
c
c
0 m rH
CM CNJ
Q Pd o QJ •H
Summary
CO CO CM u F—— 4-1
X 1
m
o ON CN oo
o c
O
u
o
00
X
O
m CN
ON 1
—
*
00 ON
* r—
03 4-1
£ CN
r—
i—4
0)
Mi rH
1
Ph o NO o i
o NT o G CD m *H c rH
CO CN G b o o
Q Cd o QJ •rH
CO CO * U r-M 4-1
Table-1.1
—
,
£
u CD
1
qj ur
'
1
a G 4-H u X 1
1 00 U 1
O OO 03 X
c o «T3
QJ ' 1
QJ rJ O X
cd
X £ X *H
—g 4—* 4-1 o QJ o.
C B
— QJ £ QJ od «
C/3 o 4-4 CJ j CM cr,
cd •
CJ Pd 03 i »
Pd o QJ C/3
— — c f-H
— c V) 4-< o
,
— I .
i— 1 Mi 4-1 i
i—l Mi 4-1 • f-H •H

QJ U3 Mi
UP 00 4-1 QJ £ 1
4-1 QJ e •H 4-J QJ Mi 1 1 QJ 0 UP QJ *h
£ 4-1 •H 03 4-1 03 4-1 4-1 £ 4-1 03 4-J c 4J X 00 •H or M 00 f-H 00 B Mi 4-1 C
O Pd £ •H cd oo £ •H c c £ 0 cd g
•H
PM
QJ
QJ
PC E Q
aj -h
CJ
03
Pd
QJ
£
03
E Q
cd
od
QJ
E 2
03
E < PP> ’ — a •rH
QJ
> •fH 7. 4^ J E
4-1 « £
03 X
*1
CJ QJ
QJ 00 > PM 1 f Mi QJ a QJ
CO £ •rH CO £ M 03 E •r- 4-1 r-H
•H 4—1 1 g o aa M H QJ
M
<d
M
03 Mi CJ Ml •Hi — 1 4h C QJ
03 QJ QJ o3 ocn) c 00 4-4 N a U QJ CJ
O > 4-1 QJ £ £ *r- £ QJ H c £ 4H Mi
M4 O 4-1 x O-H QJ •H c 03
X
c O cd 0
u
1
CJ w C/3 njisoa CJ CQ CJ J-.
358
X a
1
Cl C
r— ON 03 4-J CD
O * CO o rH c
o
Vj
O o
oo mX m o CN o O -<r —
1 4-J a)
H
•H
a
c
i-H o Cl \
• CN
o
• Vj 6 CO 0
m
• rH
Du
o
m
ro CN a
CO
CM Pi
CO Du
O
U
CD CO
rH
•H
4-1 CM
XI 1
CO c CJ
r— O' 03 4-J CD
o *> co o rH c Vj
o
mX m m 00 o
co
o CM
H 4—1
Vj
CD
£
O
H
00
CO
•H
Q
d
o •
o
rH
o CO Cl Q
f
CN Cd O O CD CO
H *H m
CM
CO CO CL CJ 4-J
X )
co 4-) C a
1— 4-J O' 03 4—1 CD
o CO CM Vj o * O — a
i
O
Vj
o o
mX in o
in
<T o
CO
o O O
vO • •
(D
4-J
1
4-J CN
rH 4-J
Vj
CD
£ —
i
u-1
CM
*H
o
c
o rH
o co CM m Q i-H CN -H 4-1 2 Pi
CO
o O
CJ
CD CO
rH
*H m
CM
in CO o3 CJ 0-4 4-1
X 1
a
co 4-J c
i— CO CM Vj 4-J O' o 03 4-1
o
CD
O o
LO m O
m
** in o o a) o rH rH c CN
H
Vj
rH
CM m o
CO
co
rH
•
CN iH
• 4-J
4-4
1
4-J CM o
4-J
Vj
CD
£
rH ro
CO Q
c
o m
rH
Cl, a n3 o Pi O CD •H CN
CO CJ CO Ou CJ 4-J
Procedures
X 1
CO 4-1 c - u
— CO CN H 4-J O' O 03 4-J
o
CD
o o
m o o O O
I
r- CD O — rH c Vj
m
i
o i-H • • 4-J <r

o
4-J CD rH CM *H c
m
rH
m
|
Cl C\] CO r— 1 CN i —14-1 4-J CN Vj £ CO Q o

a 03 3 Pi O (V
rH
*H CM
CO CJ CO Du CJ 4-J
Loading
X |
CO ON E C o
o 1
— CO
o
r> >
— o e 03 4-1
o
CD
o o
in ln o <r rv Mf CO 4-J in rH c Vj
and
X
o co
• «n
CM m o
CO
1
CN
CN
Pi o
• 4-J
o o
• 4-J
Vj
CD
B
—
1 CO
CO
•H
Q
c
o m
—
t
m Q CO CD O
U
CD
rH
•H CM
CO P Du 4-J
X |
CO £ c a
r— o B 03 4-J CD
m o
r. CO r>
O 4J o rH c
o
Vj
o O
Dimensions
in .
o
CO
o •
<3-
CN
CM
•
4-J
o
•
rH
4-J
Vj
CD
h rH
rH
CM
*H
Q
c
o rH
Cl CM m P
1
CM O PQ O (D CO *H m
CO CO '•—x
P Du U t— 4-J CM
of
CO X 1
rH ON B c a
*4 CO r. o o B 03 4-J CD
Ln o o
CO
O <r CM 4—1 m rH c O C- Vj O o
rH
in • CM • 4-1 4-J CD rH CM *H c
Cl CM m Q
1
CM X O O o Vj £ CO Q o m
CO CO CQ O QJ •H CM
Summary
p Du u rH 4—1
X
CO C 1
1
— O' 03 4-J U
CO r> O rH C 0)
o—
m o
in •
o
CO
o • CM
rH
•
4-J
Vj
CD
£
O
iH
O'
rH •H
Vj
a
o 1
Cl CM in Q
1
<N o O CD CO Q o m
CO CO Du a rH
•H
4-J
CM
Table-1.2
"e | 1
CJ
CD CD rC
O CJ 4-J 4-J X 1
PD a
CD Vj Vj o PD 03 X l
C *H
i
CD
4-J O O C O 03 •H 4-4 a
CD
1—1
4-1
— 4-4 (D XT CD h4 o X •H 03
c rH
HC g
Vj 1
CM Vj 03 CJ
U X 6 i i
HDg •H •H
—£ 4-J 4-4 • O CD CL
C r— C E £ CD C E
— CD CO '
CO O 4-j a p CL C/3
O 6 1
Pi O CO *H
—
'
03 CJ 03 ‘ Pi CD
CJ CJ '
— C OJ rH
— a C/3 4-1 Q
o 1 1
P S-l 4-1 »
«H Vj 4-J i
H •H 4-J 03 1
o X'- o X
4-) n) -i oo
<
4-J •h
—a Vj
W
<0 oi
a)
x pu
x:
00
•H
4-J
03
u
QJ
4_i
94
£
o3
O
\
4-J
•H
H a
o
4-J
Vj
cj
•u
CD
B
to
•H
o
4_J
4-J
C
Vj
CD
u X
1 •H c Q) C 0 o « CD
CD x S P CJ oi E s: a oi £ F*. S <d X ' 1
p •H > -H z >4-1 P B
Q
O CD CT3 X
< 00 > CL 1 1 Vj <D CD CD
CO c •H CO a u 03 £ *r- 4J rH
•H 4-J 1
4-J o ca u E- CD 03
CO u CJ VJ •H rH 4-1 o Vj Vj CD
CO CD CD 03 00 cO c 00 4-4 Du U <D U
o > 4-4 (D c C -H C CD Vj C C 4-J Vj
vj o 4-1 X. o •H <D •H P
w co C O 03 O
CJ CJ W CO p XI oi O CQ X CJ p Liu
359
Procedures
Loading
and
Dimensions
of
Summary
Table-1.3
360
1
-d-
, , o
—
1
in O
2.03 10
3.5
Portland
406 C <T
50x50 250
m
1
o Cement
O II m
P-34
SD30 SR24
H CD
X)
CO C
o I— CO cr> o 03 4-J
co
CO
mX in o
in • o
* o —
1
•
rH
4-J
c
03
o
rH
vD
O C
o
rn
o CO C4 m CO
1
CNl CNJ o 5-i

E NT O m
m
1 II
P-i o
CO
ol
CO
O 03 H CD CM
CL CJ
Procedures
CO T3
rH CO CT5 CM iH C
OJ
<t in
•
m
r>.
co
o
r o • -d"
r O .
03
5-i rH
03 4-J o
—
1
vD
O
1
C2
CO vO CO i— CO CO CNl CNl rH *H 4-J 03 Nf
CL
f m a
CO
cl;
CO
CL
CO
J-l
o
E
03
CL U
Loading
CO T3
NT rH CO c^ CNJ — CJ
m o
i
r—
CO M3 m • o
*>
o • Nt
«v
O .
03
5-i
03
— C
4-J o
—
1
vD
O o
m
i
m
1
and | co CNl CO CM CNJ i— 1 *H 4-J 0) Nf

Ph -©. Q cd CL 5-4 E
CO CO CO O 03
CL U
CO 03
o
Nt
m o
1
— CO Ob o C
Dimensions
CO vD
•
o •
o>
o
r.
o • <3"
r —
1
03
— c
4-J o
rH
vD
O
m
i
o
1
1 in CO i— i— CO CNJ CM 4-J 03 NT
CL, 'O a pC 5-i E CNJ
CO CO O 03
CL CJ
CO 03
of
NT
m
rH CO On o C
o
a> in CO * o e> rH 03 4-J o vD
O
CNl
1
CD
m CO
• r-*
i— CO
•
o
CO
1
CNJ
• Nf
CM o
• i — c
4-J
—
1
m
CL Q pci 5-i
<1]
E
Summary
CO co O 03
p4 CJ
03
CO C
NT
m o
1
— CO O 03 4-J
o
co
CM VD m •
o
* o Nt
—
i rH
4-J
c
03
O
>H
M3
O
} in co CNl NT CO
a
i
CM CM o 5-i E m
P-i hsi Pd O 0)
CO CL CJ
Table-1.4
, ,
e 1 1
03
CJ
0)
CJ 4-1
X 4-J GO O |
03 V-I 5-i o GO 03 C *H <D

o O u
4—1
03
u — 4-1 4-1 03
N
• — C3
03
03
O
•—
T3
4-1
H 03
1 1
c 1
1
c £ 5-i -I 03 CJ rH
CJ T3 E ' 1
•H T3 E •H •H
—g 4-J O q} a
C
o
r—
£
c
—E E 03 C E
— 03 C/3 •
C/3 4—1 CJ J a- co
a3 1 1
—CJ Oi 03 1
PC O 03 c/3 -H
O
cj
—o rH
1 1
rH — 'H
c a i — 0) 4j Q
' • 5-1 4—1 • '
5-1 4-1 « 1
4-J 03
» 1
o x:
^
i
z
1
E *4— 4-4 o o •H Is U
— u o 4-1 O
03
•H
03
4-J 4-1 — •
03 03
•H
03 (fl co
GO X E
CM ij
•H O u 03
iH
4-1 5-i
—
1 1
r-i i 1
cd
rH > •
JP •H 5-i (1) o o 5-4 a) o 5-i O G M v-i
03 Si 00 4—1 <l) E 1 •H 4-1 a> E •H 4-J 0) 5-i O n (U U

T3 c 4-1 •H oJ .u a3 4-J 4-1 C 4-J 03 H a 4-J XM •H GO rH GO E H 4-J C
o o CL Cl) cd 03 *H CJ co a) 03 -H 03 03 03 GO
03 a & 0J C 3 O (U 3J
2 •H 04
Q
X 2 Q CJ x e s a pj e 2 2 < C3 *— > -H 2 ^ hJ e
4-1 c
U 03 n3 03
03 GO > CL 1 1 u 03 03 03
CO C •H CO D Sj a3 e •H 4-J
•H
U
4—1 1 4-J
•Hi—
O CQ U H 03 03
C/3 CJ 5-4 1
4-1 O 5-i u a;
CO 03 03 03 GO 03 H CxO 4-1 /—s CJ 03 a
o > U-l 03 c c H C •
03 5-i o C .u u
H o U-l cc O -H 03 -H Q 03 o O oj o
cj u CJ CO i-i x x o CQ CJ j p-
361
Table 2 Seismic Performance of Pier Model P-1
P-2 and P-3 (Effect of Loading Velocity)
Model No. p-i P-2 P-3
Loading Velocity [cm/sec.] 17.5 35 70
Yield Strength
Py [t] 12.2 10.0 12.0
Maximum Strength
Pu [t] 13.6 13.0 13. 3
Pu
1.11 1.3 1.11
Py
....
Yield Dispalcement
6 y mm [ 22.5 22.5 22.5
r
Ultimate Displacement
6 u [mm] 84.0 90.0 84.0
6 u
3. 73 4.0 3. 73
57
Table 3 Seismic Performance of Pier Model P-4 ^ P-9

(Effect of Dynamic Loading and Longitudinal
Reinforcement Ratio)
1
Model No. P-4 P-5 P-6 P-7 P-8 P-9
Longitudinal Reinforment
1 79 0.87 0 48
.
Ratio P£ [%]
Loading Procedure Static Dynamic Static Dynamic Static Dynamic

Loading Loading Loading Loading Loading Loading
1
Yield Strength 3.9

14.3 13. 8 7.1 7.1 4. 3
Py [t]
Maximum Strength 4.8 5.0

15. 16.2 8. 8 9.2
Pu [t]
Pu 1.28
1.08 1.17 1.24 1. 3 1.12
Py
Yield Displacement
19.0 19.0 12.2 12.2 5.4 5.4
5 y [mm]
Ultimate Displacement 55.0

69.0 79.0 80.0 86.0 55.0
5 u [mm]
5 u
3.6 4.2 6.6 7.0 10.2 10. 2
5 y
362
Table 4 Seismic Performance of Pier Model P-10 ^ P-13
(Effect of Tie Bar)
Model No. P-10 P-11 P-12 P-13
25cm (Bottom 12.5cm (Bottom 12.5cm (Bottom 8.3cm (Bottom

^ crest) ^ 0.5m) and ^ 1.0m) and 't' and
0.5m_)
Interval of Hoop
25cm (0 5m ^. 25cm (0 5m ^
. 25cm (0.5m ^
crest) crest) crest)
Yield Strength
13.9 13.2 13. 2 12.2
Py [t]
Maximum Strength
17.0 16.8 16.9 16.0
Pu [t]
Pu 1.22 1.27 1.28 1.31

Py
Yield Displacement
15.4 14.8 14. 7 15.0
6y [mm]
Ultimate Displace-
74.1 78.4 77.6 88.3
ment 5 y [mm]
5 u
6“y
4. 81 5. 30 5.28 5.89
Tablr 5 Seismic Performance of Pier Model P-10, P-14, P-15

and P-16
(Effect of Cut-off)
Model No. P-10 P-14 P-15 P-16
Height of Cut-off
No Cut-off 110 135 160
from Bottom [cm]
Yield Strength 12.7

13.9 12.0 12.9
Py [t]
Maximum Strength 16.9

17.0 15.1 16.4
Pu [t]
Pu
1.22 1.26 1.27 1.33
Py
Yield Displacement 17.9 15.6

15.4 16.2
6y [mm]
74.1 70.6 83.5 80.0
6 u [mm]
6 u
4.81 4.36 4.66 5.13
67
365
Table 6 Seismic Performance of Pier Model P-10, P-17 and P-18
(Effect of Shear-span Ratio)
Model No. P-10 P-17 P-18
Shear-span Ratio h/d 5.4 3.8 2.2
Yield Strength
13.9 21.7 37.8
Py [t]
Maximum Strength
17.0 23.8 41.1
Pu [t]
Pu
1.22 1.1 1.09
Py
Yield Displacement
15.4 10.1 6.0
6y [mm]
74.1 33.1 11.2
6u mm [
6u
4.81 3.28 1.87
6y
Table 7 Seismic Performance of Pier Model P-10 and P-17 ^ 22

(Effect of Number of Loading)
Model No. P-10 P-17 P-18 P-19 P-20 P-21 P-22
Number of Loading 10 3 5
Shear-span Ratio 5.4 3.8 2.2 5.4 3.8 2.2 5.4
Yield Strength
13.9 21.7 37.8 13.6 22.1 40.0 13.0
Py [t]
r
Maximum Strength
17.0 23.8 41.1 16.7 24.2 42.6 16.7
Pu [t]
Pu
1.22 1.1 1.09 1.23 1.1 1.07 1.28
Py
Yield Displacement
15.4 10.1 6.0 16.4 11.1 6.7 16.4
6y [mm]
74.1 33.1 11.2 72.3 40.4 11.1 73.0
6u [mm]
fill
4.81 3.28 1.87 4.41 3.64 1.66 4.45
6y
L
364
Table 8 Seismic Performance of Pier Model P-10 and P-22 'U P-24
(Effect of Two Directional Loading)
Model No. P-10 P-22 P-23 P-24
Loading Procedure One Directional Two Directional

Excitation Excitation
Number of Loading 10 5 10 5
Yield Strength
13.9 13.0 14.0 13.1
Py [t]
Maximum Strength
17.0 16.7 16.6 17.2
Pu [t]
Pu
1.22 1.28 1.19 1.31
Py
Yield Displacement
15.4 16.4 16.3 17.4
Sy [mm]
74.1 73.0 51.3 64.3
6u [mm]
6u
4.81 4.45 3.15 3.7
6y
Table 9 Seismic Performance of Pier Model P-18 and P-25 'u P-27
(Effect of Reinforcement for Shear Failure)
Model No. P-18 P-25 P-26 P-27
Reinforcement for Hoop with Hoop with Diagonal

_ Shear Failure Interval Interval Spiral Reinforce-
of 25cm of 5cm ment
Yield Strength
37.8 28.8 30.7 32.0
Py [t]
Maximum Strength
41.1 37.5 38.6 42.8
Pu [t]
Pu
1.09 1.3 1.26 1.34
Py
Yield Displacement
6.0 6.2 7.4 6.0
6y [mm]
11.2 36.0 46.1 27.9
6 u [mm]
6u
1.87 5.81 6.23 4.65
6y
565
Table 10 Seismic Performance of Pier P-10, P-17, P-18 and P-28
^ P-30
(Effect of Shape of Cross Section)
Model No. P-10 P-17 P-18 P-28 P-29 '
P-30
Shape of Cross Section Square Circle
Pier Height [cm] 250 175 100 250 175 100
Yield Strength
13.9 21.7 37.8 13.4 20.4 37.0
Py [t]
Maximum Strength
17.0 23.8 41.1 17.3 25.2 42.8
Pu [t]
Pu
1.22 1.1 1.09 1.29 1.24 1.16
Py
Yield Displacement
15.4 10.1 6.0 15.2 10.0 4.9
6y [mm]
74.1 33.1 11.2 63.8 43.5 17.0
6u [mm]
6u
4.81 3.28 1.87 4.2 4.4 1 3.5
6y
Table 11 Seismic Performance of Pier P-10, P-33 and P-34

(Effect of Direction of Lateral Loading)
Model No. P-10 P-34
Direction of Lateral
0 tt/6 tt/4
Loading [radian]
Yield Strength
13.9 16.0 16.7
Py [t]
Maximum Strength
17.0 18.6 18.7
Pu [t]
Pu
—
1.22 1.16 1.12
Py
Yield Displacement
15.4 17.9 18.5
6y [mm]
74.1 54.4 <-n
v-n
6u mm [
6u
4.81 3.04 2.99
6y
366
Table 12 Seismic Performance of Pier Model P-28, P-29, P-31
and P-32
(Effect of Spiral Reinforcement)
Model No. P-28 P-29 P-31 P-32
Reinforcement at Spiral
Hoop
Bottom
Shear-Span Ratio 4. 7 3. 3 4. 7 3. 3
Yield Strength 24.0

13.4 20.4 15.6
Py [t]
Maximum Strength 25.4

17. 3 25.2 17.0
Pu [t]
Pu
1.29 1.24 1.09 1.06
Py
Yield Displacement
15.2 10.0 19.5 12.5
5 y [mm]
Ultimate Displace- 63.6

63. 8 43.5 93.0
ment 6u [mm]
<5 u
4.2 4.4 4.77 5.09
6y
367
Strength
CO
aj
u
u
u
a)
•H
Ultimate
P-.
O
4-1
X)
CD and
•H
t—
CL
CL
<
Maximum,
Tests
Yield,
a
CO
•H
Q of Loading
V4

V4
o
CO
>> Its
3
of •4-4 CO
.
c
60 60
XJ •H
Direction
Excitation
C c
•H U-
X>
d
a.
O o
t-1
o
(J
*4-4
£
O ctf
C •-
Cl > >
u
3 Q V
4-1 XJ
u
(/)
oj d
co d X
(N01 ) OVOI
568
800
Fig. 4 Pier Models for Studying Effect of

Loading Velocity (P-1, P-2 and P-3)
50
40
30
20
10
0
LEGEND
-10
LOADING VELOCITY
-20 U= l7 5 OT/sec I PI)
V= 35 cm/$ec IP2)
-30
— 17 = 70 cm/sec ( P3 )
-40
-50
-150 -120 - 90 - 60 - 30 0 30 60 90 120 150
DISPLACEMENT 8 (MM)
Fig. 5 Effect of Loading Velocity on Envelope of

Load-Displacement Hysteresis Loops
LEGEND
LOAOING VELOCITY
V - 1 7.5 cm/sec (PI I
'
V 35 cm/sec ( P2 )
'
V 70 cm/sec I P3
Fig. 6 Effect of Loading Velocity on Equivalent

Hysteretic Damping Ratio
569
(TON)
LOAD
DISPLACEMENT 8 (MM)
(a) Longitudinal Reinforcement Fig. 7 Pier Models for Studying Effect

Ratio of 1.79% of Dynamic Loading and Longitudinal
50
Reinforcement Ratios
40
UCCND
30 STATIC LOADING IP4I
‘
DYNAMIC LOACH NC ‘
20
[TON)
10
P 0
- 10 LEGEND
LOAD
STATIC LOADING (P6I

-20
DYNAMIC LOADING IP7)
-30
-40
(a) Longitudinal Reinforcement
-50
Ratio of 1. 79%
-150 -120 -90 -60 -30 0 30 60 90 (20 150
DISPLACEMENT 8 (MM)
(b) Longitudinal Reinforcement

Ratio of 0.87%
50
40
30
20
[TON) (b) Longitudinal Reinforcement
10
P Ratio of 0.87%
0
LOAD
-10 LEGEND
STATIC LOADING IP8I
-20
- DYNAMIC LOADING (P9I
-30
-40
-50
-150 -120 -90 -60 -30 0 30 60 90 120 150
DISPLACEMENT 8 (MM)
(c) Longitudinal Reinforcement

(c) Longitudinal Reinforcement
Ratio of 0.48% Ratio of 0.48%
Fig. 8 Effect of Dynamic Loading and Fig. 9 Effect of Dynamic Loading and
Longitudinal Reinforcement Longitudinal Reinforcement
Ratios on Envelope of Load- Ratios on Equivalent Hysteretic
Displacement Hysteresis Loops Damping Ratio
570
on
Ratio
Ratio
Damping
o
•H
U Reinforcement
oz
Hysteretic
c
<u
E
aj
Tie
a
u
o
4-1 of
C
•H
Equivalent
CD
Pi
Effect
CD
•H
H
4-1
O 12
a
CD
4-1 Fig.
.
4-)
%OI'0 = M
d
-I'
% 02
.
0 = *d
w o o o
00 Oliva 9NldWV0 3I13H31SAH
c
•H
(%£0Z- »d) O fr X El a "a

00S 2
d
4-»
cn
o
4-1
on
co
CD
T3 Ratio
O
S
CD
•H
Loading-Displacement
,
(%£0'Z =
d) Ofr *£l 0
Reinforcement
oos e s'zsz
OJ O
00
•H
o
o
Loops
m Tie
o of
of
Hysteresis
Envelope
Effect
( OOS x 00Q 3JDnbs )
uo|p9S uwnjoo
11
(N01) d OVOl
Fig.
371
u
0 )
u
0)
•U
01
>>
X
c
01
cd
>
•H
3
O'
W
c
o
U-l
u-l
0
1
u O
CJ U
cd
U-t C£
o
&0
4-J u C
CJ -H
o 0) c.
U-J
u-i e
u-i cd
U-l
0 w Q
1
U
u
>4-1
ci
o 3M Oliva ONIdWVO 0 13831 SAH
I
U-.
u
a»
U-<
U-4
w
&0
c
H3
cd
o
o
0) 0)
a. a
o
ô o
O kJ
>
C 0)
UJ
C 0)
o *-
O U
CJ c
O 1)
4_» cd
0) Q-
(N01) d ovo
572
o
o r-— co Equivalent
Q_ Q_ Q_
T CO CM
IT) to c\i
on
Ratio
Ratio
Shear-span
Damping
of
Hysteretic
Effect
18
Fig.
of
oos 2 Ts'zq i Envelope

Loops
= *d
( %£0‘Z 0t> * £
•
) 1
Ratio
juawaDjojuiay |Duipni;6uo~|
i
, , . . .
-Span
on
Hysteresis
Shear
oi|°M ]ueiueoio)U|ey eij.

Ratio
i
(009*005 eJDnbs
uoiioes uiunioo
span
Shear
of
Effect
17
Fig.
373
o
O
X>
JO
CO
C*2
C
c0
Cl
CO
II 00 I
Lj
e c cO
' N •H
oo m X) d)
X
«H II cO
C o C/5
CL
1
^ hJ
4-1
o Loops
a [ N 01) d OV01
Envelope
Number
oo
Hysteresis
of
m on
CM
o 4-1
o
-L>
1
a X3
CL 0) Loading
2SZ I 4-1
**
4-1
JC
(%£0 2 =
*d ) Ot- * El a
O w o
~L i — 00 of
II
c
'
C '
•H CO
r^. co XJ
•H ||
c Number
ê
c0
i 4-J
CL O.
CO
CO
Li
O of
CO
rQ 44
 0)
(N01) d 0V01
r—\
I CSI
005 (X CN
2 5 25 L
I
Ratio
(%£0 2 - *d)0fr* £10 - CL ^
)ueuj8DJ0ju|ea iDuipnjibuon
O X
rH
m
Span
C
II CO
00 4h
C
V-/ •H O
v
Shear-
%0I 0 = *d
o
•H
n in X
II ||
ojioa )U8iueoiO)U|ea 9 !i
(005 * 005 eJdnbS)

CL
I
wC wC J=
uoiioes uuinioo
o
c
«0
Cl
CO
I
Li
CO
0)
JO
CO
(noi) d oven
3W
LEGEND
NUMBER OF LOADING n
n =10 (PIOI
- n = 3 ( P 19)
n = 5 ( P 22)
(a) Shear-span Ratio h/d of 5.4
LEGEND
NUMBER OF LOADING n
n = 10 ( P 17)
n = 3 ( P20)
(b) Shear-span Ratio h/d of 3.8
LEGEND
NUMBER OF LOADING n
n = 10 ( PI8)
-
n = 3 ( P2I
(c) Shear-span Ratio h/d of 2.2
Fig. 21 Effect of Number of Loading on Equivalent

375
c
O o
•H
60 U
C Cd
•H a£
T3
Cd 60
o G
hJ -H
CL.
* B
cd G
C Q
O
•h a
4J*H
U AJ
a) 
OJ 0) -H
a 4-1 3
aj 4-j cr
a
cd
W W
r—
a
CO
•H 60
Q C
•H 60
4-4 T3 •H
O cd
^ OllVd DNIdWVO DI13H31SAH U-
o
>. hJ
u
o i-H
4J cd
CO a
•H o
•hi
4-J

CM
u
cd
CM Ha
co
60 *H
luawaooidsiQ •H
Cn
Q
I
60
c
•H
T3 CO
td a.
o o
I-J o
4-i
J
O CO
<V (0
o. a;
o u
0)
CD 4-1
> CO
C >>
W X
(N01) d 0V01
60
Lu
376
Ratio
Damping
Hysteretic
co
-U
c
a)
6
0)
o
(%£OZ = 'd)Otr*£ia }-l
S Z9I 1 S£8 o
4-4
C Equivalent
*rH
cu
*-4
MO d
CM 0 )
I
x
Ch CO
4-1
o
CO %Z0’l=*d »!tDd
„ dooH lDjids
o
<D
=5 (009*009) 4-1
S uoijoas umn|oo 4-1
W
00
c
•H
(%£0 Z ? -a
=
d) 0 fr*£l a
S Z9I I 9£8
3
4-J
Loops
uo
J-l
o
4-1
0]
-
I Hysteresis
cu
T3
O
u
cu
•H
D-.
LD
CM
(% £0'Z= y d)0fr *£l0

00
0001 ! T zw •H
fa
Ratio
Span
-
of
Shear
377
B
U
o
LO
CM
U-l
o
U
u
ac
V-i
a)
•rH
(fr9S*l JDinojia
uojioeg uionioo on
6
u
(%£0'Z = »d) O-b * El 0 Envelope
m
U-i of
Ratio
0
3.3
Span
= jr
UJ M
h/d
u •H Section
<U
Shear-
J %OI'0=*d 0} jDy a 2C
co
luouueojojuiey e||
o Vj
'J
E ( t79S 96 )
JDinojjO •H
o uotpa$ uujnjoo CL Cross
of
Loops
Shape
of
Hysteresis
Ratio
Effect
4.7
= o
o
h/d 3
Shear-Span
3 29
CO
JC Fig.
60
0)
-X.
(noi) d oven
378
LEGEND
SECTION
CIRCULAR ( P30)
SQUARE ( PI 8
(a) Pier Height of 100 cm
LEGEND
SECTION
CIRCULAR P29I
SQUARE P17)
(b) Pier Height of 175 cm
LEGEND
SECTION
CIRCULAR IPI9I
SQUARE (PIO)
(c) Pier Height of 250 cm
Fig. 30 Effect of Shape of Cross Section on Equivalent

579
ro 'd-
o ro ro
CL CL on
Ratio
Q_
O^ ^
II HII
CD CD ©
Loading
Damping
I
of
J % £0 2 = Id) Ot^xeiO
”^525^
0052
00
C Hysteretic
•H
X Direction
U-i
O of
e
o Equivalent
•H
4-1 Effect
U
OJ
>-<
•H
Q 33
(%C02=ld)0t>*£l0_ 91 0I1VH DNIdHVQ 0I13H31SAH

005 2 5 25Z" O Fig-
O)
M-i
00
c
•H
X)
D
Envelope
Loops
(%co 2 = id) ot?xeia

_ 525Z" on
005 2
Hysteresis
Loading
of
,
% 010 --*d
gi oiioy luswaajojima ail
§ (00S»005I
uoiiaas uuin |03
Direction
of
Effect
of
[NOll A 0 V 0
32
Fig.
380
Equiva
on
(%£0'Z =
»d)0fr*£ia
05 / 1 S'ZOSI
Ratio
Ratio cm
ro
Reinforcement
=3.3
P-.
4-J
h/d Damping
Shear-Span
a
_ J a>
n
aj
%OIO=*d °iZ0l=*d u
o|ioy juawaojojujay an u Spiral
Loading
o
(t?9Stf) JDinojjo 4-4
uojpas uainioQ c Hysteretic
•H
<U of
C*!
(%£0 3 = *d ) otr * £1
05/ 1 5 305 1 cd
Effect
•H lent
Ratio cn Q*
C/5
3.3
1
OJ
<4-1
= PU
O 36
h/d
Shear-Span
o
o o o
juaiuaojojujea ajj.
a a) 3M OllVd ONIdlNVO 0ll3d31SAH
E 4-4
4-1
Fig.
(b9S0) JDinotn
uojpas utun|oo w
oo
c
•H
(%£0'Z =
»d) ot^ x £i a "d
2
005 Z 5 35/
cn
p
Ratio
o Envelope
Loops
4-4
1
4.7
Shear-Span h/d <u on

nj
1
O
'r % oi o --
*d °lzo'i=*d
S Hysteresis
oipy juaujaojopiey 01 x
oj
( t7 9Q ^ JDinoilD ) •H
uojpas uiunioo Ph
Reinforcement
<r
I % £0 3 = *d ) 0 tr<£l 0 OP
005 Z
oo
•H
P-l
Ratio
oo
C\J
47 Spiral Loading-Displacement
= PH
Shear-Span
h/d
n 1
oipy j
% 0l'0= M d
ueiueo jo^ u 1 ey 9(1
of
(fr 9 Q**) J d 1
no j o |
Loading
uo|joas uuj n 1 03 Effect
of
35
Fig.
381
REEVALUATION OF LIQUEFACTION RESISTANCE CURVES
by
Riley M. Chung^
H. Bolton Seed
A. G. Frankliijiy
J. A. Farrar
of the most challenging tasks that geotechnical

ABSTRACT
engineers face today. Due to the unavoidable
This paper updates the status of the U.S. -Japan consequences of sample disturbance, design
Cooperative Program in SPT testing and the approaches using laboratory test results are now
correlation of SPT with liquefaction resistance beginning to receive less emphasis as compared
of saturated soil deposits. with methodologies developed from in situ
testing techniques. The Standard Penetration
SPT field practice in Japan was found to result Test (SPT), due to its popular use in site
in higher energy ratios (ER^) than those from exploration and its very large accumulated
the U.S. practice for the same rope and pulley world-wide data base, has proved to be very
method used (67% in Japan vs 60% for the safety useful as an index of soil liquefaction resis-
hammer widely used in the U.S.). The use of an tance during earthquake shaking.
ASTM sampler without liner resulted in N-values
which are about 80% of those measured using a However, it has long been recognized that there
sampler of straight barrel configuration. are problems associated with the use of SPT and
its results for engineering design in general,
From an assessment of the overall worldwide SPT and for liquefaction resistance evaluation in
practices based on the most recent available particular, due to the large number of variables
information, a standard energy ratio of 60% is that affect the N-values obtained from SPT
recommended. The field data relating SPT to tests. Schmertmann in 1976 [16] and Kovacs,
liquefaction susceptibility, accumulated from et al. in 1983 [8] showed that the SPT test is
previous earthquakes throughout the world, have in fact conducted, in different parts of the
been reinterpreted and replotted with respect world, using different kinds of hammers, dif-
to this recommended standard energy level. ferent energy delivery systems with differing
degrees of efficiency, different borehole
Key Words : Earthquake Engineering, Soils, fluids, and different kinds of sampling tubes,
Liquefaction, In Situ Tests, Standard Penetra- in addition to differences caused by operator
tion Test idiosyncrasies. This situation certainly raises

a legitimate question as to the consistency of
the SPT N-values which have been used In
1. INTRODUCTION
developing the correlations between liquefaction
Evaluation of liquefaction resistance of satu- resistance and modified penetration resistance,
rated soil deposits during earthquakes is one N [21, 22, 27].
l
I U.S. National Bureau of Standards

II University of California, Berkeley, CA
III U.S. Army Engineer Waterways Experiment Station
IV U.S. Bureau of Reclamation
382
Research to measure the amount of energy Additionally, this study provided the
delivered by a hammer to the sampling spoon opportunity to examine in detail the suit-
through a SPT test set up has been an important ability of the SPT vs. liquefaction
program in the geotechnical engineering resistance curves now used extensively by the
operation at the National Bureau of profession for the liquefaction potential
Standards (NBS) . In 1981, a field measurement determination [23]. Findings from this
system and procedure which measures the energy examination will be briefly presented in this
delivered by a drill rig operation were paper
developed and successfully used to study the
factors that affect delivered energy [7]. The 2. FIELD SPT TEST PROGRAM IN JAPAN, OCT. 1983
developed measurement system has been used by
the researchers at NBS, with cooperation and Standard penetration tests were conducted at
support from the Nuclear Regulatory Commission, Akita, Niigata, Tokyo, and Yokohama sites in
Bureau of Reclamation (BuRec) , and Waterways Japan during the Japan trip using the SPT
Experiment Station (WES) , to calibrate drill technique typical of the Japanese practice. In
rigs throughout the U.S. In this energy cal- essence, three different procedures in raising
ibration study, summarized by Kovacs, et al. in and releasing the 140 lb (63.5 kg) hammer are in
1983 [8], the energy ratios measured for use [28]

56 drill rigs and operators in the U.S. were
reported. Based on the data from this study, a. a mechanical trip device (the tonbi
the mean energy ratio delivered by a safety method) which is essentially a free-fall
hammer is about 61 percent, whereas the mean technique
energy ratio for a donut hammer is about
45 percent. It should also be noted that the b(l). the rope and pulley technique, and
safety hammer is the most popularly used hammer

type in the U.S. b(2). the rope and pulley technique with a
special throw of the rope in which an
In October, 1983, under the auspices of the effort is made to throw the rope at least
UJNR Panel on Wind and Seismic Effects and partially off the pulley during the
through the financial support of many organi- release
zations both in Japan and in the U.S.,
engineers from the NBS and Bureau of Reclama- All three methods were used; the majority of the
tion carried out a successful SPT program to tests were conducted using either method (bl)
a) document how the SPT is performed and used or (b2) . The amount of energy delivered to the
in present Japanese engineering practice; sampling spoon, in terms of energy ratio, ER^,
b) compare the Japanese SPT results with those with respect to the theoretical energy delivered
found in the U.S.; and c) comment on the SPT by a 140 lb (63.5 kg) hammer falling 30 inches
vs. liquefaction resistance curves with respect (760 mm) was measured using the measurement
to the Japanese data base. Preliminary system described in Ref. 7. Figure 1 shows a
findings from this study were presented to the frequency diagram of the data which indicates an
16th UJNR panel meetings held at NBS in May overall average of 68 percent with a standard
1984. An NBS report [9] has been prepared to deviation of 9.6 percent and a coefficient of
present in detail the conclusions with respect variation of about 14 percent. When the data
to the objectives of the study. The results from tests using the tonbi method are separated
will be summarized herein. from that using the rope and pulley methods, the
383
averages would be 80 and 67 percent, respec- which illustrates the N-value profile with depth
tively. Data from the U.S. practice averages for borings Cl and C2 at the Niigata site. Open
about 55 percent with a standard deviation of circles are used for data points obtained using
12 percent and a coefficient of variation of the JIS sampler, and triangles for the data from
22 percent [8]. the ASTM sampler without liner. The numbers in

parentheses adjacent to the data points are the
The standard penetration test in Japan is per- average values of the measured energy ratios in
formed according to the Japanese Industrial the drill rod, ER^ . A solid line connects the
Standard (JIS) A1219, which is based on the data points for a given boring. The dashed
present version of the ASTM D1586. It lines in the figure represent the profiles of
specifies the size of the SPT sampler to a the normalized N-values with respect to the
2 in. (51 mm) O.D. and 1-3/8 in. (35 mm) I.D. averaged energy ratio of 68 percent versus
throughout its length. However, in the current depth
ASTM standard, the I.D. of the SPT sampler is
specified to be 1-1/2 in. (38 mm) to allow Using the data from Figure 2 and other similar
insertion of a brass liner. With the liner in data obtained from the tests at other sites, it
place, it would have a uniform inside diameter is possible to compare the N-value results at a
of 1-3/8 in. (35 mm) . Even though the ASTM given depth using a JIS sampler and an ASTM sam-
standard calls for a uniform 1-3/8 in. (35 mm) pler without liner to study the effect of sam-
inside diameter split barrel, many U.S. opera- pler type. Figure 3 gives the results of this
tors do not use the liner, probably because of comparison. The N-values in this figure have
the greater ease with which the soil sample can been normalized for a 68 percent energy ratio.
be examined and removed from the sampler with- It is apparent from this figure that at the same
out the liner. depth the JIS sampler produces greater N-values
than the ASTM sampler without liner.
An ASTM sponsored survey conducted in July 1980
among members of the Association of Soil and In his review of the data from this study,
Foundation Engineers (ASFE) showed that only Schmertmann suggested that additional data
52 percent of the member firms who responded to points could be obtained by comparing the
the questionnaire said that they used the liner N-values from the JIS sampler with those from
(Kovacs, 1981). Without the liner, the fric- the ASTM sampler without liner by interpolation
tion is reduced, which could account for a 10 on the data sets where direct depth comparisons
to 30 percent reduction in N-values in sands, are lacking. Figure 4 shows

according to Schmertmann [19, 20]. the results when all the data points are plotted
together and the N-value correlation between JIS

To investigate the sampler effect on N-value, and ASTM samplers without liner may be
the SPT tests were conducted under similar con- established as follows:
ditions with the sampler as an only variable.
In many cases, tests were conducted in a pair N - 0.8 N , (1)
us j
of borings which were about 3 to 5 m apart.
The JIS sampler and the ASTM sampler without if a straight line relationship is assumed.
liner were used alternately in both borings so

that the SPT tests could be carried out at the In general, it can be stated that the use of in
same depth with two different types of sam- ASTM sampler without liner leads to N-values
plers. A typical plot is given in Figure 2 lower than those measured with a ‘ampler of a
584
constant inside diameter, by about 10 to hammers such as the Borros AB drop-hammer and
30 percent. the Pilcon hammer. A comparative field study

was conducted by Douglas and Strutznsky [3] to
2 . 1 Average Energy Ratios for SPT Procedures evaluate the difference between the Chinese free
Countries Where SPT is Most Popularly Used fall hammer and the Pilcon hammer and no sig-
nificant difference was found between the
The rope and pulley technique is still the most N-values by the two types of hammers.
widely used method for hammer raise and release
in the U.S. and other Pan-American countries. Thus, while we do not have direct energy ratio
Two types of hammers are currently used: the measurements on the free fall hammers used in
donut hammer and the safety hammer. As stated the Chinese practice, it is logical to use the
earlier, a study by Kovacs, et al. [8] found an energy ratios measured for the Pilcon type ham-
average energy ratio of about 45 percent mer as the reference for the Chinese data.
delivered by the donut hammer whereas the Several studies of the energy delivered by the
average energy ratio for the safety hammer is Pilcon hammer were conducted and their results
about 60 percent. This difference in energy are summarized [23] in Table 2. An average of
ratios means that a donut hammer could be 60 percent is derived.
expected to result in about 30 to 35 percent
higher N-values than a safety hammer. This Shi-Ming [24] also reported that N-values
range of difference is also found in field com- measured by a mechanical trip system are about
parisons between the two different hammer types 83 percent of those measured using a manual
[1, 4, 13, 14, 25, 30]. system. Thus, the energy ratio for hammers
operated with the rope and pulley method in
Major hammer release mechanisms used in the China is likely to be about 50 percent, which is
Japan SPT practice are discussed earlier in in good agreement with that determined for
this paper. Several investigators have U.S. hammers using the same hammer releasing
attempted recently to quantify the energy mechanism (55 percent)
ratios delivered to the drill rods with the
different types of releases. Their findings, 3. RECOMMENDED STANDARD ENERGY RATIO
including that from the 1983 Japan study
described earlier, are summarized [23] in Several authors [5, 15, 18] have in the past
Table 1 . The results show generally good recommended that an energy ratio of 55 percent
agreement. Also shown in the table are the be adopted as the standard for SPT data cor-
overall energy ratios considered most relation since this seems to be the average
representative of the Japanese practice and energy delivered by hammers used in the U.S.
thus adopted for the present study. This means that SPT data obtained from both
safety and donut hammers can be correlated using
Until about 20 years ago, the Chinese were the following equation:
using a manual method analogous to the rope and
pulley method for SPT hammer release.
the normal method today in China in performing
However, N
55
= N
mm
(ER /55) (2)
the SPT tests uses an automatic mechanical trip where N = N-value for method used in
m
to release the hammer [24]. The automatic investigation
"free fall" hammer used in the Chinese practice ER = Energy ratio for method used in
m
is similar in configuration to "free fall" investigation
385
It should be noted that neither the safety Another correction which is sometimes necessary
hammer not the donut hammer produces an average with respect to comparison of Japanese and U.S.
energy ratio of 55 percent. The task of cor- data is the effect of different borehole
recting field SPT data would be minimized if an diameters. In Japan, the most commonly used
energy ratio of 60 percent is adopted as a borehole diameter is 2.6 in. (66 mm) whereas in
standard [23], If this were done, data the U.S. it is normally not less than 4 in.
obtained by safety hammers, the most commonly (100 mm). The use of a smaller borehole would
used hammer type in the U.S., would not need tend to increase the penetration resistance of a
correction, nor the SPT data from the Chinese deposit, thus leading to the conclusion that
practice using an automatic trip hammer. The
SPT N-value determined in this way is N N
( *
) ah >
60
(
us >4A
60
j
designated as N^q, and is determined by:
There is little data available that could be
N,_ = N (ER / 60) (3) used to quantify the two effects described
dU m m
above. Furthermore, it should also be noted
The energy ratios for SPT tests from the that the two effects are compensating for dense
countries included in this discussion have been sands while for loose sands they are cumulative.
summarized by Seed et al. [23] as shown in Thus, it is suggested [23] that for soils with
Table 3. Table 3 also gives the correction (N^p) < 20, the following relationship may be
factors for each condition, derived using used provisionally for the combined influence of
eq. 3. these two effects.
Other corrections may sometimes be required to (N ) = 0.9 (N, n ), (4)

60 us 60 j
the SPT data, particularly with respect to com-
parison of N-values obtained from Japanese and Corrections should also be given for the case
U.S. practices. For example, the frequency of where very short drill rod length (less than
hammer drops is significantly lower in Japan 10 ft) is used. In this case, studies [13, 17]
than in the U.S. They range from 17 to have shown that there is a reflection of energy
20 blows per minute in the Japanese rope and in the rods which reduces the energy available
pulley method using a special throw, but for driving the sampler, resulting in
typically about 30 to 40 blows per minute in fictitiously high N-values. Seed, et al. [22]
the U.S. practice. Consideration of the effect recommended that the measured N-values for tests
of the hammer drop rate on the rate of dissipa- made at depths between 0 and 10 ft be corrected
tion or accumulation of excess pore water by a multiplier of 0.75, which is also used in
pressure during sampler penetration suggests the present study.
the following relationships:
4. REINTERPRETATION OF SPT LIQUEFACTION

(N.) > (N ), n for loose sands, (N. ) < 20 RESISTANCE DATA BASE
.
j bU us 60 1
and
(N.) < (N ), for dense sands, (N, ) > 20 Liquefaction resistance curves such as the one
j bU us bU 1
shown in Figure 5 have been developed [22], com-

where N is the blow count corrected to 1 tsf paring the earthquake-induced cyclic stress
^2
(1 kg/cm ) overburden pressure [21, 22] ratio in the soil at any depth to the normal l/r.l
soil penetration resistance, N ^ , for cases of
both occurrence and nonoccurrence of
386
liquefaction-related ground failure phenomena The available body of data used in reinterpre-
during the earthquake. tation of the correlation between cyclic stress

ratio and the normalized standard SPT value,
The cyclic stress ratio for soil at any depth (N, is obtained from sites in North and
during an earthquake can be estimated using the South America, Japan and China. Appropriate
following equation [21, 22]: corrections, as discussed earlier, were applied
to all the field values to determine the cyclic
a a stress ratio corresponding to M = 7-1/2 earth-

max o
= 0.65 (5)
quake and the corresponding values of (N ) .
1 oU
The results for clean sands with fine contents
less than 5 percent are plotted as shown in
where a = maximum ground surface
max
Figure 6. Soil deposits known to have liquefied
acceleration
are plotted as solid points, those which did not
a = total overburden pressure at
o
show any evidence of liquefaction are shown as
depth under consideration
open points and sites of marginal liquefaction
o ' = effective overburden pressure at
o
by half-solid points. The boundary line is
depth under consideration
drawn to separate conditions associated with
r = a stress reduction factor which
d
liquefaction from conditions for which liquefac-
decreases from a value of 1 at
tion occurrences have not been observed, which
the ground surface to a value of
is the same principle used in developing
about 0.9 at a depth of 35 ft
previous resistance curves. In Figure 6, the
boundary line from Figure 5 is also superimposed
The value of N^ can be obtained from the value
for comparison. It may be noted that the cor-
of N measured in a standard SPT investigation
relation line drawn for the case of clean sand
by using the relation [22]:
is very close to that proposed by Seed et al. in
= C 1983 [22] indicating that the values used in

N * N, (6)
! N
the earlier work correspond closely to (N,),„
values. In fact, the two boundary lines merge
where is a correction factor determined as a
N
at (Nj-n = 15. Beyond this level, the original
function of overburden pressure and relative
boundary line is slightly more conservative.
density of the soil under investigation.
For example, at a cyclic stress ratio of 0.3,
the standardized values would have to be 25
All available field performance data for soil
from the revised boundary line, rather than 27
deposits which have or have not liquefied under
known conditions of past earthquakes have been from the original boundary line.
reviewed, reinterpreted and related to the

recommended standardized values of (N ) as 5. CONCLUSIONS
1 60
outlined earlier by Seed et al. [23]. Most of
the available field data were collected for
The following conclusions can be drawn from the
earthquake magnitudes close to M = 7-1/2. results of the studies presented herein:
Resistance curves for other earthquake mag-

1. The differences in SPT practice in Japan and
nitudes can be obtained from the curve for
7-1/2 magnitude by using the correction factors the U.S., caused by the variations in equipment
given in Ref. 21. and procedure, result in differences in the

actual energy passing through the drill rods,
ER^ . The cooperative study in Japan indicated
387
that the average ER^ values are 67 and 80 per- two boundary lines that there is no difference
cent for the rope and pulley method and the in the correlation for (N^)^q values up to 15.
tonbi method, respectively, used as the hammer Above (Nj)gQ = 15, the original boundary line
release mechanism. Both values are sub- is slightly more conservative.
stantially higher than the U.S. average of

55 percent, using the rope and pulley method. 6- ACKNOWLEDGEMENT
2. The Japanese energy data are found to have The studies presented in this paper were
less scatter than those from the U.S. practice. sponsored by the National Science Foundation,
the Naval Facilities Engineering Command, and
3. The effect of sampler type on the SPT the Japan Society for the Promotion of Science,
N-Value is significant. On the average, SPT in addition to the parent organizations of the
7.
N-values obtained using an ASTM sampler without authors. The support of these organizations is
liner were found to be approximately 20 percent gratefully acknowledged. The authors also wish
lower than those obtained using a sampler of a to thank Messrs. W. D. Kovacs, K. Tokimatsu, and
constant inside diameter. The sampler effect L. F. Harder for their valuable advice provided
is smaller for loose sands, on the order of in the course of the study.
10 percent. However, its effect may be as

large as 25 to 30 percent for dense sands. REFERENCES
4. Examination of energies delivered to the 1. Clark, K. R. (1969), "Research on
sampler by different methods of performing SPT Undisturbed Sampling of Soils, Shales, Air
tests in different countries and with different Drilling Techniques, and Data on Penetration
equipment led to the choice of a standard Resistance Testing - Third Progress Report on
energy ratio, ( N ^)gQ» determined in SPT tests Soil Sampling Research," Report No. EM-770,
where the driving energy in the drill rods is U.S. Department of Interior, Bureau of
60 percent of the theoretical free fall energy. Reclamation.

The selection of (N.),. will allow us to
1 60
minimize the task in converting any measured 2. Decker, M. D., Holtz, R. D. , and Kovacs,
N-value to this recommended standard. W. D., "Energy Transfer of SPT Hammers," In

Preparation
5. Field performance SPT data with respect to
liquefaction resistance have been reexamined, 3. Douglas, B. J., and Strutznskv, A. (1984),
reinterpreted, and replotted with respect to "SPT Hammer Energy Measurements and Piezocone,"
(N.),., Other corrections considered in the Ertec Western, Inc., Report to U.S. Geological
1 60
reinterpretation included the effect of hammer Survey.
drop rate, the borehole diameter, the condition

where very short drill rod length was used, and 4. Harder, L. F., Hammond, W. D., Ross, P. S.,
the sampler effect. Kassel, J. W. and Motumah, L. K. (1983),

"Thermalito Afterbay Dam Seismic Evaluation,"
6. A liquefaction resistance curve, shown in California Department of Water Resources, Final
Figure 6, is developed for an earthquake Draft
magnitude of M = 7-1/2. The figure also

includes the boundary line from the original
design curve. It can be concluded from these
388
5. Kavazanjian, E. , Jr., Roth, R. A., and 12. Oh-oka, H. (1984), "Comparison of SPT
Echezuria, H. (1983), "Probabilistic Evaluation N-values by Cathead-and-Rope Method and Trip

of Liquefaction Potential for Downtown Monkey Method (Concerning Diluvial Sandy
San Francisco," Report No. 60, Department of Deposits)," Proc. of 19th Annual Meeting,
Civil Engineering, Stanford University, 147 pp. JSSMFE, (in Japanese).
6. Kovacs, W. D. (1980, "Results and Inter- 13. Palacios, A. (1977), "The Theory and
pretation of SPT Practice Study," Geotechnical Measurement of Energy Transfer During Standard
Testing Journal, American Society of Testing Penetration Test Sampling," Dissertation sub-
and Materials, September, pp. 126-129. mitted in partial fulfillment of Doctoral of
Philosophy Degree, University of Florida,
7. Kovacs, W. D., Salomone, L. A., and Yokel, Gainesville, FL.
F. Y. (1981), "Energy Measurement in the
Standard Penetration Test," Building Science 14. Rebertson, P. K. (1983), "In situ Testing
Series 135, National Bureau of Standards, of Soil with Emphasis on its Application to
Gaithersburg, MD, August, 99 pp. Liquefaction Assessment," Dissertation submitted
in partcal fulfillment of Doctoral of Philosophy
8. Kovacs, W. D., Salomone, L. A., and Yokel, Degree, University of British Columbia,
F. Y. (1983), "Comparison of Energy Measure- December.
ments in the Standard Penetration Test Using
the Cathead and Rope Method," National Bureau 15. Rebertson, P. K. , Campanella, R. G. and
of Standards Report to the U.S. Nuclear Regula- Wightman, A. (1982), "SPT-CPT Correlations,"
tory Commission, Report No. NUREG/CR-3545 Soil Mechanics Series No. 62, University of
November, 80 pp. British Columbia, Department of Civil Engi-
neering, October, 19 pp.
9. Kovacs, W. D. and Salomone, L. A. (1984),
"Field Evaluation of SPT Energy, Equipment, and 16. Schmertmann, J. H. (1976), "Predicting the
Methods in Japan Compared with the SPT in the q /N Ratio-Interpreting the Dynamics of the
c
U.S.," NBSIR 84-2910, National Bureau of Standard Penetration Test," University of
Standards, Gaithersburg, MD, August, 72 pp. Florida Report to the Department of Transporta-
tion, FL, October.
10. Liang, Nancy (1983), "An Examination of
the Standard Penetration Test with Comparison 17. Schmertmann, J. H. (1977), "Use of the SPT
to Other In Situ Tests," Senior Thesis, to Measure Dynamic Properties? — Yes, But...!"
University of British Columbia, Canada, April. Proceedings, American Society for Testing and
Materials Symposium on Dynamic Field and Labora-
11. Nishizawa, Y. , Fuyuki, S., and Uto, K. tory Testing of Soil and Rock, Denver, CO, June.
(1980), "In Situ Tests and Efficiency Con-
cerning SPT," Proceedings of Sounding 18. Schmertmann, J. H. and Smith, T. V. (1977),
Symposium, JSSMFE, (in Japanese) "A Summary of SPT Energy Calibration Services
Performed for the Florida Department of Trans-

portation Under Service Contract 99700-
7150-010," Final Research Report 245*D73 College
of Engineering, University of Florida,
Gainesville, FL, 21 pp .
plus appendices.
389
19. Schmertmann, J. H. (1979), "Statics of 25. Steinberg, S. (1982), "Energy Calibration
SPT," Journal of the Geotechnical Engineering and Hammer Influence on SPT," Paper presented
Division, American Society of Civil Engineers, at the Engineering Foundation Conference on
Vol . 105, No. GT5 , May. Updating Subsurface Sampling of Soils and Rocks
and Their In situ Testing, Santa Barbara, CA,
20. Schmertmann, J. H. (1980), "Guidelines for January.
Cone Penetration Test - Performance and
Design," Report FHWA-TS-78-209 , Federal Highway 26. Tokimatsu, K. and Yoshimi, Y. (1981),
Administration Implementation Division, 145 pp. "Field Correlation of Soil Liquefaction with SPT
and Grain Size," Proceedings of International
21. Seed, H. B. (1979), "Soil Liquefaction and Conference on Recent Advances in Geotechnical
Cyclic Mobility Evaluation for Level Ground Earthquake Engineering and Soil Dynamics,
During Earthquakes," Journal of the Geo- St. Louis, MO, Vol. 1, pp. 203-208.
technical Engineering Division, ASCE, Vol. 105,
No. GT2, Proc. Paper 14380, February, 27. Tokimatsu, K. and Yoshimi, Y. (1983),
pp. 201-255. "Empirical Correlation of Soil Liquefaction

Based on SPT N-value and Fines Content," Soils
22. Seed, H. B., Idriss, I. M. , and Arango, I. and Foundations, Vol. 23, No. 4, December.
(1983), "Evaluation of Liquefaction Potential
Using Field Performance Data," Journal of the 28. Yoshimi, Y. and Tokimatsu, K. (1983), "SPT
Geotechnical Engineering Division, Vol. 109, Practice Survey and Comparative Tests," Soils
No. GT3, March. and Foundations, Vol. 23, No. 3, September,
pp. 105-111.
23. Seed, H. B., Tokimatsu, K., Harder, L. F.,

and Chung, R. M. (1984), "The Influence of SPT 29. Yoshimi, Y., Tokimatsu, K. and Oh-oka, H.
Procedures in Soil Liquefaction Resistance (1983), "Comparison of SPT N-values by
Evaluations," Report No. UCB/EERC-84/15 Cathead-and-Rope Method and Trip Monkey Method,"
Earthquake Engineering Research Center, Proceedings of the 18th Annual Meeting, JSSMFF.,
University of California, Berkeley, CA, (in Japanese)
October
30. Youd, T. L. and Bennett, J. J. (1983),
24. Shi-Ming, Huang (1982), "Experience on a "Liquefaction Sites, Imperial Valley,

Standard Penetration Test," Proceedings of the California," Journal of the Geotechnical Engi-
Second European Symposium on Penetration neering Division, ASCE, Vol. 109, No. GT3,
Testing, Amsterdam, May. March
390
Table 1. Summary of Rod Energy Ratios for Japanese SPT Procedures [23]
Study Mechanical Trip System (tonbi) Rope & Pulley

Nishizawa et al. (1980) 80 - 90 63 - 72
Decker, Holtz, and Kovacs 76

(In press)
Kovacs and Salomone (1984) 80 67
Tokimatsu and Yoshimi (1981) 76*

Yoshimi and Tokimatsu (1983),
Yoshimi et al. (1983),
Oh-oka (1984)
Adopted for this study 78 67
* Equivalent rod energy ratio if rope and pulley method is assumed to have an
[
energy ratio of 67 percent and values for mechanical trip method are different
from this by a factor of 1.13.
Table 2. Summary of Rod Energy Ratios for Pilcon-Type Free-Fall Hammer [23]
Study Hammer Energy Ratio (%)

Decker, Holtz, and Kovacs (1984) Pilcon 55
Douglas et al. (1984) Pilcon 62
Liang (1983) Pilcon-type 58
Overall average = 60
391
Table 3. Summary of Energy Ratios for SPT Procedures [23]
Hammer Estimated Rod Correction Factor
JAPAN*
Country
*
I.
A.
Type
**
Donut
Hammer Release
Free-Fall 78
Energy (%) for 60% Rod Energy
78/60 = 1.30
B. Donut Rope & Pulley with
special throw release 67 67/60 = 1.12
II. USA
* A. Safety Rope & Pulley 60 60/60 = 1.00
B. Donut Rope & Pulley 45 45/60 = 0.75
III. ARGENTINA
* A. Donut Rope & Pulley 45 45/60 = 0.75
IV. CHINA
* A. Donut Free-Fall*** 60 60/60 = 1.00
B. Donut Rope & Pulley 50 50/60 = 0.83
* Prevalent method in this country today

** Japanese SPT results have additional corrections for borehole diameter and
frequency effects.
*** Pilcon Type hammers develop an energy ratio of about 60 percent.
58.8 68.3 77.8
Figure 1. Frequency Diagram of ER^ Results
392
N-BLOWS PER FOOT
0 10 20 30 40 50
0 “ —i
1 1 1 1
2 -
Akita Site C
Data from borings C-1
and C-2
4 -
O JIS sampler
A ASTM sampler (with no liner)
( ) ERj-%
6 -
8 -
(meters)
10 -
DEPTH
12 -
14 -
16 -
18 -
20 -
Figure 2 . N-Value vs Depth Profile for Akita Area

Borings C-1 and C-2
395
Figure 3. N-Value Correlation Between JIS Sampler and ASTM
Sampler Without Liner, Normalized to Average Energy
Ratio of 68 Percent, Direct Measurement Data
Figure 4. N-Value Correlation Between JIS Sampler and ASTM

Sampler Without Liner, Direct Measurement Data and
Data Obtained from Interpolation
594
Figure 5. Correlation Between Field Liquefaction Behavior
of Sands (D > 0.25 mm) under Level Ground Condi-
tions and Standard Penetration Resistance (Ref. 22)
Frorn Fig. 5
New Boundary . Old Boundary

1 \
A
ol
/
/
/
r o
• . / /
/
•
/
• / /
0
•
—zLl
•
• //.
• •
•
v/
/ <?
o 0
" 0 // o
*\ •a • a J»' *°o
M = 7.5
— a /
0
a
'
tv o
o
$
Chines s Building
FINE S
Co de(clay cont ent

L que-
a ction
CONTENT
Margi nal
Liq.
5%
= 0)
No
Liq
(•) —
A f
Pan- Ame ic a n data

r a a
Japanese data • Q o
Chinese a a t
A
1
0 10 20 30 40 50
( N i
)
6 o
Figure 6. Relationship Between Stress Ratios Causing Liquefaction

and N^-values for Clean Sands for M = 7-1/2 Earthquakes
396
IN SITU TEST FOR ASSESSING LIQUEFACTION POTENTIAL
USING VIBRATORY CONE PENETROMETER
Yasushi Sasaki
Yasuyuki Koga
Yoshihiro Itoh
Takao Shimazu
and
Masuo Kondo

ABSTRACT The development of the VCPT and some results of
tests were introduced in the past UJNR Joint

In situ tests using the newly developed vibra-
Meeting [1, 2]. The in situ test results by
tory cone penetrometer were carried out at sites
this VCPT are partially described in the
where liquefaction of ground took place during
U.S. -Japan joint research report submitted to
past earthquakes.
the 16th Joint Meeting [2], in which a rela-
tionship between results obtained by the VCPT
The results obtained by these tests were com-
and liquefaction strength by triaxial tests is
pared with the cyclic triaxial test results on
proposed.
undisturbed samples taken from the same sites.
The test results by the vibratory cone penetrom-
In this present paper, the same relationship is
eter at Akita sites were compared with the
reexamined; in addition, the relation between
earthquake-induced settlement of Hachirogata-
VCPT results and earthquake-induced settlements
Dikes which were severely damaged by the
of embankments is discussed.
Nihonkai -Chubu Earthquake of May 26, 1983.
2 . OUTLINE OF VIBRA TORY CONE PENETROMET ER

1. INTRODUCTION
The vibratory cone penetrometer has been devel-
In Japan, the SPT (standard penetration test) is
oped in order to satisfy the following objec-
the most popular test used as the in situ test
tives:
method for assessing liquefaction potential of
soils during earthquakes. Blow counts (or N- (1) To obtain liquefaction strength of soils
values) from the SPT, however, are obtained continuously along depth,
di scontinuously with respect to depth, generally (2) To get liquefaction strength of soils
at every one meter of depth. Therefore, the SPT accurately, even for loose silty fine sands
is not always the best way to get continuous having much fine content,
information of soils, especially for assessing (3) To know precise liquefaction strength of
liquefaction potential of complicated deposits soils, in their existing state, at sites,

which are composed of sandy and clayey layers, (4) To get liquefaction strength of soils at
alternatively. sites rapidly, and to get data at many sites
easily,
In order to solve this disadvantage, the VCPT
(5) To carry out in situ tests in an inex-
(vibratory cone penetrometer test) has been
pensi ve way.
newly developed at the PWRI (Public Works
Research Institute) and field tests using this The VCPT apparatus is composed of parts similar
apparatus have been carried out in the labora- to the Dutch Cone Penetrometer. A basic dif-
tory and at several sites. ference between them, however, is that a snail
396
during earthquake. Among the vibratory cones
vibrator is installed inside the vibratory cone,
shown in Fig. 1, Cone-II has only the vibrator
as illustrated in Fig. 1.
and is without any sensors to measure pore water
When the installed vibrator is turned on, the pressure or cone resistance. Table 1 summarizes
whole body of the vibratory cone vibrates around the main features of the vibratory cone.
its axis. By this excitation, the pore water
pressure in soils surrounding the cone may build 3. SOIL STRENGTH AGAINST LIQUEFACTION AND
DEGREE OF PENETRATION RESISTANCE DECREASE
up as in the case of a saturated sandy ground.
Therefore, penetration resistances under the
In Fig. 3 is shown the relation between the soil
condition in which the vibrator works, will
strength against liquefaction and the degree of
become smaller than those without excitation.
penetration resistance decrease which was orgi-
nally proposed in the previous paper [2].
Although the effects of the vibrator on soils
surrounding the cone might be different from
The data points shown in this figure were
those of actual earthquakes, the pore water
obtained at Hachirogata site where the river
pressure change is deemed to represent an indi-
dikes were severely damaged during the
cator of the vulnerability of the in situ soils
Nihonkai -Chubu Earthquake of May 26, 1983, and
to liquefaction. It is considered that the dif-
Kawagishi -cho site and Showa-Ohashi bridge site
ference between penetration resistances with and
damaged during the Niigata Earthquake of
without cone excitation will reflect in situ
June 16, 1964.
liquefaction resistance of soils.
The following equation was proposed in the pre-
Figure 2 shows an example of in situ measurement
vious paper, based on the data obtained from
of penetration resistances. As shown in this
these sites. In this figure, the stress ratios
figure, the penetration resistances decrease
to cause liquefaction, which were measured by
when the vibrator works. This difference can be
cyclic triaxial tests for undisturbed samples
expressed by the degree of penetration
taken from the sites, are plotted against the
resistance decrease which is defined by the
degrees of penetration resistance decrease at
fol lowing equation.
the same depth.
Rjj.20 = 0.26 + 0.16 1-D) 1 - 5

( (2)
Where, RÔ = liquefaction strength (stress

ratio) obtained by triaxial tests
where D = degree of penetration resistance
decrease D = degree of penetration resistance
penetration resistance without decrease [Eq. (1)]

Q =
c
operation of vibrator (kgf/cm 2 )
y In situ tests were carried out at several addi-
Q = penetration resistance with operation
c
of vibrator (kgf/cm 2 ) tional sites other than the above mentioned
The value of D, the degree of penetration sites. The data from these tests are plotted in
resistance decrease, Fig. 4, together with the data shown in Fig. 3.

is nondimensional and can
have any value between zero and one. In Fig. 4 the data points are plotted with dif-
When a
soil has a large value of D, close to one, the ferent marks according to effective overburden
soil can be regarded vulnerable to liquefaction pressure. Table 2 summarizes the test results.
337
From Fig. 4, it is known that the Eq. (2) ru = N /N*
0 (7)
expresses fairly well the relation between
Considering Eq. (8), Eq. (9) is reduced
stress ratio to cause liquefaction and the
degree of penetration resistance decrease, as
N* = n o/ r u (8)
most points are distributed close to the curve
obtained by Eq. (2). However, the points are R 20 = R0 * (1/N*) b • (20)b
scattered around the curve. Reasons for this

= R
0
' (20/N o )b . (r u )b (9)
scatter are discussed below.
If the relations of R 20 = t20/ a v'> Ro = T 20/°v'

The following are some assumptions on liquefac-
tion characteristics: are taken into account and the pore water
pressure ratio caused by the penetration of the
Assumption 1 . - Logarithm of stress versus vibratory cone is supposed to be equal to the
logarithm of stress cycles has a linear rela- value of the degree of penetration resistance
tion, namely decrease, then the following equations are redu-
ced :
log R = a + b log N (3)
ru = D (10)
where R = stress ratio
N = stress cycles T 20 = t o /(20/N o )b • (D)b (11)
a, b = constants
When the dynamic stress is caused by the
Assumption 2 . - Pore water pressure ratio penetrating vibratory cone acts constantly on
versus stress cycles ratio has a linear rela- soils surrounding the cone ( = const.), the
0
tion, namely value of R0 (= x 0 /a v ') in Eq. differs by
(9)
changing the depth of penetration. On the

ru = rN (4)
contrary, since the terms in Eq. (11) have
where ru = pore water pressure ratio (=u/a v

'
constant values, except for (D) h , X 20 is given
u = pore water pressure generated by by Eq. (11) by substituting the measured value
the cyclic stress (kgf/cm2) of D.

'
ac = initial confining pressure
Considering this fact, the results shown in
r[\j = stress cycles ratio = N/N^
Fig. 4 might be better illustrated by plotting
N = stress cycles
the stress itself to cause liquefaction rather
Nj£ = stress cycles to cause lique-
than the stress ratio. Figure 5 shows the
facti on
result plotted for the stress.
Futhermore, it is supposed that the stress

Comparing with Fig. 4, the plotted points in
ratio R0 = x 0 /a v ' acts for N 0 cycles on soils
Fig. 5 are spaced more closely if the confining
surrounding the vibratory cone that has
pressure has the same condition. It seems,
penetrated undergound and generates the pore
however, that the relation between x* and D is
water pressure of u0 = ru • aw '
affected by the confining pressure. This can he
Then the following relations are reduced explained by the fact that the supposed relation
expressed by Eq. (10) is different from the real
R 2o = 10 a • (20) b (5) relation and the assumption established might
differ slightly from the actual condition. This
R0 = 10 a • (N*)b ( 6 )
596
may imply that for estimating more precise liquefaction of the ground. The authors
liquefaction strength by utilizing the VCPT intended to find out a relation between the
results, the effect of the overburden pressure damage to the dikes and the liquefaction degree
should be considered. Regarding this point, obtained by the VCPT.
further studies should be carried out in the
The settlements of the crest of the dikes are
near future.
used as an index value of the damage to the
4. DAMAGE TO HACHIROGATA DIKE AND DEGREE OF dikes. There are recorded elevations of the
PENETRATION RESISTANCE DECREASE
crest measured in July 1976, so the differences
between the elevations of the crest after the
In situ tests by the vibratory cone penetrometer
earthquake (measured in June 1983) are defined
were carried out near the toe of the Hachirogata
as the settlements caused by the earthquake.
river dikes, which were severely damaged during
the Nihonkai -Chubu Earthquake of May 26, 1983. Several factors, including ground conditions,
The number of test points were 32 and their height of dikes, seismic effects and so on, are
locations are shown in Fig. 6. generally considered to affect complex
earthquake-induced settlements of dikes. In the
Although the geological conditions at these 32
case of the damage at Hachirogata, however, the
points differ from each other as illustrated in
heights of the dikes are comparatively identical
Fig. 7, they can be classified into four types
and the formation of soil layers of the ground
considering ground response characteristics
is considered to be almost the same in a fairly
during the earthquake. Four ground types and
wide area. So, the estimated thicknesses of
their peak surface accelerations calculated by
liquefied ground are taken as a principal factor
dynamic analysis of the material based on the
to explain the degree of the settlements.
multi reflection theory are shown at the bottom
of Fig. 7. In estimating liquefied thickness, integrated
values of the degree of penetration resistance
Calculated values of peak surface accelerations
decrease are taken as index values. Since it is
have a different range for each ground type.
considered that liquefaction of the shallow part
Peak surface accelerations at these points are
of the ground affects the settlement more inten-
estimated between 160 and 220 gal during the
sively than liquefaction of deeper layers, the
earthquake.
calculation of integrated values is made con-
Due to this earthquake, the Hachirogata river sidering a weighting function as shown in Fig.
8. In this calculation, however, the effects of
dikes settled and suffered from the cracks on
the overburden pressure on the VCPT results men-
their slopes. Although the height of the dikes
tioned in a previous section are neglected.
vary from place to place, front dikes are some 4
Integrated values of the degree of penetration
to 6 m high. The earthquake-induced settlements
vary from place to place; the maximum dike
resistance decrease with the weight function in
settlement induced by the earthquake was about this manner are tentatively denoted as the
1.5 m in this area [3], thickness of liquefiable layers.
Fig. 9 shows a relation between the settlements

As sand boils were observed along the toe of the
dike at most of these settlement places, the of dikes and the thickness of liquefiable
settlement layers. The ground types mentioned in the above

is considered to be caused by
399
(5)
are designated by different marks in this Settlements tend to get larger for sites
fi gure. exposed to larger accelerations, even if the
thicknesses of liquefiable layers are the
In Fig. 9 is seen a fairly good correlation bet-
same.
ween the settlements of the dikes and the
thickness of the liquefiable layer for each 6. ACKNOWLEDGEMENTS
ground type. Settlements of dikes are affected
The in situ test data and the triaxial test data
by thickness of liquefiable layers, as well as
introduced in this paper are obtained by
ground types. This implies that for estimating
cooperation of several organizations. The
possible damages to embankments during earth-
authors express sincere acknowledgment to indi-
quakes, it is important to know the liquefiable
5.
viduals concerned at Tohoku Regional
thickness of ground in addition to the degree of
Construction Bureau, Kanto Regional Construction
liquefaction potential. Since the VCPT can give
Bureau, Hokuriku Regional Construction Bureau,
continuous information on liquefaction poten-
Chubu Regional Construction Bureau, Metropolitan
tial, it is considered that the VCPT is one of
Highway Public Corporation, Akita Prefectural
the more useful in situ tests.
Government, and Miyagi Prefectural Government.
CONCLUSIONS
7. REFERENCES
The following are concluded from the results of
1. Yasushi Sasaki and Yasuyuki Koga, "Vibratory
in situ tests using the vibratory cone penetrom-
Cone Penetrometer to Assess the Liquefaction
eter which was developed at PWRI:
Potential of the Ground," 14th Joint Meeting,
(1) A good correlation is noticed between JUNR , 1982.
the degree of penetration resistance decrease
2. Yasushi Sasaki, et al . U.S. -Japan
and the shear stress to cause liquefaction. ,
Cooperative Research on In-situ Testing Procedure

(2) As the degree of penetration resistance for Assessing Soil Liquefaction (No. 1), 16th
decrease becomes larger, the soil strength Joint Meeting, UJNR, 1984.
against liquefaction gets smaller.
3. Public Works Research Institute, Report on
(3) The effects of the degree of penetration Damages due to the Nihonkai -Chubu Earthquake,
resistance decrease on the liquefaction Report of Public Works Research Institute,
strength of soils are affected by the effec- No. 165, March 1985.
tive overburden pressure at tested depth.
(4) A good correlation is noticed between
the thickness of the liquefiable layer and

the settlement of the dikes damaged by the
Nihonkai -Chubu Earthquake. The larger the
thickness of the liquefiable layer is, the
larger the settlement is.
400
Table 1 Vibratory Cone
• (a) Comparison of Vibratory Cone
Standard Vibratory Cone

Cone
I n nr IV V VI
Diameter (mm) 35.7 41 41 42 41 42 44
Length (mm) 205 789.6 789.6- 760.2 333.2 1177.8 743.8
o 60° 60°
Apex Angle <T>
O 60° 60° 60° kQ
OO
i i
Weight (kg) 1— 1
—
6.3 6.3 5.2 2.7 8.6 6.7
Vibrator No Yes Yes Yes No Yes Yes
Sensor No Yes No Yes Yes Yes Yes
(b) Specification of Vibrator
Size, etc.
Item
i, n nr, v, vi
Diameter 41 mm 31 mm
Length 5 76 mm 431.5 mm
Frequency 200 Hz
Power Source AC 3(p 48V 200 Hz
Centrifugal Force 160, 80, 32 kgf 80 kgf
(c) Specif ication of Transducers
Capacity
Symbol Item Type
H H a < VI
Load Transducer
qc Strain Gauge 2tf 5tf
of Point Resistance
Load Transducer
fc Strain Gauge 2tf
of Side Friction
Pore Pressure
p Strain Gauge 2kgf/cm 2 5kgf / cm 2
Transducer
Displacement
s Transducer of Potentiometer 106 cm
Penetration
Table - 2 Summary of VCPT & Cyclic Triaxial Test Result — 1
Site Depth (m) Qc Qc D or; Rfi20 t£ FC

n i 2 3 - 64 . 12 0 . 0 813 . 0 27. 0. 280 0 U76 . 4."
n-2 33 0 « 30 0 .. 0 43 . J 36
(_) . 0. 4oO 0.144 8. 0
A~3 4 . 84 . 67 . 0. 202 0 49. 0 4 . .1
( o„ 201 12.0
B- 2 42 0 .. 12 0 . 0 .714 0 38. 0. 330 0 . ) 25 2 .
D-2 4 . 31 0 .. 5 . 0.839 0. 52 0 325 . o. 169 0 5 .
B-3 1 0 80 23 0 .. 0 713 . 1 05
. 0 28 .. 0. 294 0 5 .
C-l 3 8 . 60. 0 94 . 0 000 - 0 53- 0. 365 o 93 . J 4 .
D— 1 1 . 59. 2 . 0. 966 0 25. 0 304 . 0. 07 6 1 .
D-2 4 . 44.. 0 1 4 . 0. 682 0 53. 0. 368 0. 195 0.

D-3 6. 46.0 1 4 . 0. 696 0.68 0 31 « 0 .21 l 1 .
D— 9.8 26 0 « 26 . 0. 000 0. 98 0 31 . o 309 . 1."

E-l 2 3 a 4 0 . 0 . 0. 988 0 35. 0.348 0. 22 1 S3."
E— 4 . 55 0 . 0 . 0. 991 0 49.
(*)
. 27 *\
0.11 4 20 .
E-3 3. 4 65 . 4 . O 9 ill 0 58. 0 5. J 0.1 25 13. 5

E— / . 3 75. 0 6. 0 0 .920 0. 77 0 21 7. 0 . 67 J 1 9 . ' >
E— g, | 9 0 1. . 1 3 . 0.857 0 84. 0 233 . 0 200 . 23.

E-6 9 4 . 72 . 1 6 0 . 0. 778 0 .96 0 .21" 0. 202 1 9 .
E— 1 i . 4 92 0 .. 36 0 - 0 609 . 1.15 0. 90 1 0 2 9 . 1 15.0

F- 1 2. 4 65 0 .. 42 . 0 354 . 0 32. o. 5.;8 0. 169 60 o .
F--2 3 . 17.0 1 0 . 0 . 4 1 2 0 38. o. 268 0. 102 5 1 .
F-3 5. 0 68. 0 50. 0 0 2 5 . O 52m 0. 275 0. 43 J 1 8 .
G- 1 1 4 . 24 . 1 7 . 0. 293 0 83.. 0. 307 ( ) .. 255 6" 0 .
6-2 1. 7 . 4
”•
~J
”f
1 2 0 . 0. 560 1 07 0. 33< ' i
. 353 49. o
G— 1 9 - 4 40. 33 3 „ 0 1 35 . 1 22 0. 295 Q '» g 36. 0
H-l 3. 4
<—, cr j_
11.9 0.535 0. 36 0. 4 22 0 152 .
H-2 3- 3 70 8 .. 52. 8 0 254 - 0 57. 0. 368 0 2 o . 1
H— h> . L"; 93 3 „ 75. 9 0 1 86 . 0.63 0- 278 0. 175

1-1 6 .
9'
52. 4 35. 1 C t "T ‘T ||")
0 65. 0. 373 0. 24 2
J-l 6. 9 .1 33 . 5 122. 2 0.085 0.65 0. 373 0. 24 2
K-J 5. 4 4 8.5 43. 3 0 07 . J 0 52. 0. 4 i 0 . :< iB
| —2:
7 3 - / 3! - 6) 60. 9 0 161 . o. 69 0 . 5R5 o. 26 4
2’
L-l 8.3 58.3 34 . 0 41 . 0. 78 ". 34 5 !',
- j/ C?
L— 2 1 O 3 . 14 7. 128.9 0 1 23 . 0. 96 0. 575 0. 552

|_— 3 J 2 3 . 1 22 3 . 96. 1 0. 2 1 4 1 1 4
. 4"7
<4 . 0. 16 1
11-1 3 3 # 33 7 _ 2 0 . 0. 377 0 33. 0 .377 0. .‘4 1
li -2 12-5 104 . 0. 16). 1.16 0. 483 o. 560
(JQ2
Table - 2 Summary of VCPT & Cyclic Triaxial Test Result — — 2
Site Depth(m) Qc Qc
v
D
000 0 35
cr;
50
Rj220 U75 FC
3(9 7
~7
i i-i 3. 4 .
!'• T
0. . 0. 1. 0. 1
M-2 7 3 . 1 1 4 5 . 1 02. 1 0. 08I 0. 69 0 350 - 0. 242

1 J— cp
a
~r
150. 1 1 43. B 0. 042 U 0J. < 0 51 . 0. 449
0- 1
cr -?•
42,. 1 38 4 . 0 . 088 0 50. 0 450 0 225
-2 6 1 . 72. 72. 1 0 . 00 0 50n 0. 442 0 256 -
P— 1 1 6 3 . 40.9 42.7 0 . 000 1 57

. 0 272 .. 0. 427 1 0.
Q— 2. 4 1 5.0' 3 0 . 0. BOO 0. 36 0 41 . 0. 148 83. 0

Q-2 B. Li S1 4 . < 47.0 0 . 4 20 0 „ 9 0-330 0 300 , 38 .
R-1 5 3 . 121. 98 4 . 0 187

. 0 - 60 0.435 0.261 5. 0
R-2 9 4 . 1 0 6 J. . 90.7 0. 107 0 - Vb 0.385 0 366 . 21.0
B— 1
pi;
45. 6 2B 8 . 0. 368 0 - 35 0.340 0 .119 49 . 0
S-2 cp
93. Q '7 o 0. 000 0 . 90 0.305 0 2 75 . 32 . 0
T- l 6 7 « 33 3 . 25. 1 0. 546 0 . 96 0.255 0.245 9 . 8
T— 7 2 a
O Z. ,L
t
34. 0 64 4
. 1 . 1 0 0.352 0 387 - 5 . 0
I— 12.3 205. 4 123. 2 0. 400 1 47
. 0.345 0 50 7 . 4. 1
T —4 1 5 4 . VO o ™ 6 *1 - x.*.' 0 290
. 1 74
. 0.290 0 505 .. 52 6 -
1-1 4 4 . 43. 5. 1 0 88 1
. 0. 59 0 270 m 0. 159 37 0 .
LI-2 7 9 . 1 20 9 . 1 01.6 0.160 0. 90 0 310 „ 0 n 279 2. 8

V- 1 3 0 . 71 5 45. 5 0.364 0. 46 0 260 0 J 20 . 2 3.
i » _ '3 3. 9 4 1 . 3 52. 2 0 000

r 0. 55 0 290 . 0 . 1 60 1 9.
v— B. 3 165,3 81.4 0 500

. 0. 94 0.235 0 22 . 9 3.
l
J. 0 6 .. 1 67. 1 160. 0.04 1 1 . 1 0.280 0 322 . 23. 4
H-i V.) H / 4 1 . 2 32. 5 o 2 i
. :l 0 . 63: 0.2B0 t.
76
'
. 1. 74.
x-t B 67 0 . 74 8 . 0 j00
.. ( 0 82. 0.290 0 238 , 5 4.
X 2 9 4 . 1 1 3 7 . 1 02 . 0. 103 0. 92 0u300 0 276 . 5 . si.
V- 2 5 . 1 7 1 J,4 . 0 333
- 0 4 .. J. 0.315 0. 129 4 2.
V— "T ^ 13.5 7 0 . 0. 481 0. 48 i

j j
i ^—
0. 146 25 0 .
Y-3 5 6 „ 72. 2 1 . 0.654 0 63. 0. 270 0 1 70 . 3 .
V —4 7 1 02 0 . 80. 0 0 21 Oh 73 0. 290 O l.zo „ V.i 5 0.
Z-i 4 . 7 5?. 0 5 0 . 0 136

„ 0 80- 0. 200 0 1 60 .
403
m iv
(T) Porous metal
(D Pore pressure transducer

(3) Load transducer of
point resistance
(4) Load transducer of

side friction
(5) Take-out cable for

transducer
(6) Vibrator
( 7) Power source cable for

vibrator
(o) Penetrating rod
Fig. 1 Vibratory Cone
404
Fig. 2 Example of in-situ Measure of VCPT
— Showa-Ohashi Bridge Site —
Fig. 3 Relation between Liquefaction Strength and

Degree of Penetration Resistance Decrease
— U.S. -JAPAN Cooperative Research (No.l) —
405
0.8
O 0.0£CTV'<0.3
0.7 h a o.3<;av'<o.5
o.5^cr;<o.7
• ov^CTvô.g
Hi
* a o.9â;<i.i
.
0.6
£ « m 1.1 Sff,'<1.3
_ 2o
cn
V 1.3SCX<1.5
X >> o i.5s cr;< i.7
a O 0.5
1.7S<T,’<1.9
>
ô CM
Unit ; kqf/crrp
O — 0.4
=5
^ .2£
O
*7 O
O a3 0.3
O 'cd
'— 3
03 cr
CC
0.2
c/3 CD
CO c/3 . = 0.26 + 0. 1 6( 1 -D)
03 3
_t CO
CO O 0.1
0.0
0.0 0.2 0.4 0.6 0.8 1.0
Degree of Penetration Resistance Decreace ( 1-D )
Fig. Relation between Liquefaction Strength and

- R (1-D) -
120
0.8 -I 1
|
£
—o
'
oo.osa;<o.3 i.isa;<i,3
0.7 - A0.3^CTv'<0.5 V1.3iCrv‘<1.5 -
w a o.5£CTv'<o.7 oi.s^cr;<i.7
O • o.7scrv'<o.9 Ti.7scr;<i.9
zr - A 0.9S (7v< kgf/cm J -
CO 2 o CO 1 .1 Unit;
0 aJ
H CO
0
ycl
o bi - T V “
O •
o O
2 o -
it
cj
>» c
O o
0.3
£ ^3
ff
0.2
a A*n
CD - 1
A
i_ CD
ct3 c/3 aA
CD 3 0.1
J3 CO
CO o o
0.0 02 0.4 0.6 0.8 1.0
Degree of Penetration Resistance Decreace D

Fig. Relation between Liquefaction Strength and
— T D —
120
406
6 Location of VCPT Points at HACHIROHGATA
(0 ~ .ts ~
-
w
8 .
- o « Profile
0 * Q. M
<0 _
"O
>
10
" o «
-
>» w Soil
o> C -a
>.
S! » and
O ru
.c -
•— + .•e +
J - ? * Points
i_ a *.
o
0 l«.
0 ta.
*
>. *
>.
iS - i9 - VCPT
g *
> «“
the
3
—CO; to of
_
.a o
Loaction
CM
7
I
(0
a
>.
Fig.
•a
c
o
k—
O
408
HI-
Fig. 8 Weight Function and Integration of D
Fig. 9 Relation between Settlement of Dikes and the Thickness

of Liquefiable Layer
408
U . S .-JAPAN JOINT EARTHQUAKE RESEARCH PROGRAM
INVOLVING LARGE-SCALE EXPERIMENTS ON MASONRY

BUILDING STRUCTURES-JAPANESE SIDE RESEARCH PLAN
BY
Shin Okamoto^, Yutaka Yamazaki^ ,
Akio Baba^,
Masa-omi Teshigawara^ ,
and Hisahiro Hira-ishi^
ABSTRACT
A joint earthquake research program of building and was effectively employed in this research
structures involving large-scale experiments was program. In addition to the large-scale test,
carried out since 1979 between Japan and the several other types of experimental work is in-
U.S. under the auspices of the Panel on Wind and cluded in the program. They are 1) component
Seismic Effects of the U.S. -Japan Natural Resour- test of structural beam and column elements, 2)
ces Development Program. During the first two planar test of frame subassemblages, 3) shaking
years (1979-1980), earthquake response behavior table test of replicas of building structures,
of a reinforced concrete building was investi- (4) correlation study between shaking table and
gated based on experiments of a full-scale seven- pseudo dynamic tests.

story specimen. This work was carried out at The research program including these experimental
the Building Research Institute, Japan. A simi- programs is aimed at investigating the effective-
lar experimental research program on steel build- ness of a variety of procedures developed to
ing structures was held during the next three evaluate the seismic behavior of structures such
years (1981-1983). At this time, as a third as analytical simulation procedures, design pro-
phase of the joint program, masonry building cedures, and safety evaluation procedures. Such
structures, is being developed for the five year a comprehensive effort in this ioint program is
period 1984-1988. This paper introduces the believed to be the first to achieve the above ob-
Japanese-side research objectives and research jective. The expected outputs of this study are:
plan for masonry building structures for the (1) to establish design methods to ensure higher
U.S. -Japan joint research program. safety;
(2) to establish more economical design methods
1. INTRODUCTION without losing structural safety;

A joint research program on earthquake engineer- (3) to develop more accurate procedures to esti-
ing of building structures was performed since mate the seismic safety of structures;
1979. This research was conducted between the (4) to develop experimental and analytical pro-
U.S. and Japan under the auspices of the Panel cedures to simulate the earthquake response
on Wind and Seismic Effects of the U.S. -Japan behavior of structures with a greater accu-
Natural Resources Development Program. racy;
Testing earthquake response simulation of large- (5) to make the overseas construction by U.S. and
scale building structures is the objective of Japanese contractors more competitive; and
this program. It was carried out at the Build-
ing Research Institute of the Ministry of Con- -

^Director, ^Division Heads, ^Research Engineer,
struction, Japan. To perform the large scale ^Senior Research Engineer, Structural Engineer-
test, a new earthquake simulation test technique, ing Department, Building Research Institute,
designated as the pseudo dynamic test method, Ministry of Construction, Tatehara, Oho-machl
was developed by the Building Research Institute Tsukuba ,
Ibaraki ,
30S ,
JAPAN
410
(6) to enhance information and personnel ex- new masonry construction systems have been devel-
change. The joint program was initiated in 1979 oped to make the construction more reliable and
as shown in Fig. 1. In the first phase of this competitive. The reinforced concrete block sys-
program, the reinforced concrete structural system, named RCB, is one example. The Building Re-
tem was examined; a full-scale seven-story rein- search Institute has contributed much to the
forced concrete structure was tested using the development of the masonry structural system. One
pseudo dynamic test technique. In the second current project, "Application of Small PC Unit to
phase, the steel structural system was studied Urban Housing," involves developing quality ma-
with a test of a full-scale six-story steel sonry units to better resist earthquakes.
braced frame. For the third phase of this pro-
gram, the masonry structural system was selected. This is a brief review of the background of the
This third phase masonry program was initiated masonry research in Japan. Considering this situ-
in 1984 and is to continue for five years. The ation, the immediate target of this program as
first joint workshop on the masonry program was the third phase of the U.S. -Japan Joint Research
held in March, 1984 at the Building Research Program will be the development of a masonry con-
Institute. struction system fit to low- to medium-rise
building structures.
2. BACKGROUND OF MASONRY STRUCTURES AND NEEDED
AREAS OF RESEARCH 3. JAPANESE RESEARCH PLAN
The masonry structural system is known worldwide 3 . 1 Organizations
to be a traditional but efficient structural sys- Fig. 2 shows the organization for promoting U.S.-
tem having high capacity in durability, fire re- Japan cooperative research on masonry structures.
sistance, sound and heat insulation, and high The Japanese have three committees: Technical
flexibility in construction. The structural sys- Coordinating Committee on Masonry Research
tem was first imported to Japan from Europe, and, (TECCMAR) ,
Promoting Committee on Masonry Re-
therefore, little attention was given to the search (PROCMAR), and Building Construction Com-
earthquake resistance of the constructed mason- mittee on Masonry Research (BLDCMAR). The first
ry structures. Accordingly, those structures committee, TECCMAR, coordinates technical re-
sustained great damage in the Great Kanto Earth- search planning on materials, components, assem-
quake. Since then, masonry structures practi- blies and full-scale experiments. They are draft-
cally disappeared in Japan, and very little ing a proposal of aseismic design guidelines of
effort was made for the study of seismic resis- low/medium-rise reinforced masonry building
tance of the masonry structural system. structures in Japan. The second one, PROCMAR,
was organized under domestic cooperation research
Because of large demand of building construction agreement among the Building Research Institute,
after World War II, Japan developed various mass the Japan Association for Building Research Pro-
produced construction systems. During those motion and the Building Contractors Society for
periods, however, the masonry structural system promoting U.S. -Japan cooperative research on
was not used mainly because masonry systems masonry structures. The committee members are
require tedious labor work. Today, more value composed of government officials, local self-
is put on quality than quantity, and versatil- governing body's officials, and representatives
ity is more appreciated than uniformity. To from many industrial organizations. The last
comply with this new demand, it is natural to committee, BLDCMAR, was organized under PROCMAR.
reconsider benefits of the masonry structural Problems concerning masonry building construction
system. Since the late 1960s various types of techniques, except structural matters, are dealt
\
with in this committee. 2) to propose decreased wall length rate:
21-cm/ra^ at present regulation to;
3 .2 Research Plan 15-18cm/m^ for five-story buildings and

The Japanese research items of interest for the 12-15cm/m2 for three-story building in
five year program are: proposal
1) material test; 3) not to use reinforced concrete collar beams:
2) static test of walls; Reinforced concrete collar beams are re-

3) static test of beams; quired in reinforced masonry buildings
4) static test of wall and beam assemblies; under present regulations
5) full-scale planar static test; 4) to simplify joint works of reinforcing bars:
6) shaking table test of scale model specimen; Effectiveness of lap joint or mechanical
7) full-scale five-story structure test; and joint works of vertical reinforcing bars
8) establishment of aseismic design guidelines. in walls is experimentally investigated.
Research time schedule for the above items are Through Japanese research, high strength block
shown in Fig. 3. During the first two years, units in which design prism strength is more than
fundamental tests on materials, wall and beam 180kg/cm^ are now used. In the present regula-
components, assemblies and full-scale planar tions, 80kg/cm2 design strength for full section
static tests will be carried out. The program area of a block unit and 135kg/cm2 design
for shaking table tests of a scale model strength for grout concrete are required. Stan-
specimen and a full-scale five-story structure dard concrete and clay block units used in
|
test will be performed in the third year of the Japanese research, which were newly designed and
program. produced for the program, are shown in Figs. 4
and 5
4. JAPANESE RESEARCH OBJECTIVES

The joint research program is to propose aseis- To allow setting relatively large amount of rein-
1)
mic design guidelines for low/medium-rise reinforcing bars into unit holes, large size holes
forced masonry building structures in Japan. are prepared for the units, and are fully grouted
At present, only the reinforced masonry build- R/C slab system will be applied to future full-
ings with height up to 12 meters (one to three scale five-story specimen which is to be tested
stories buildings) are permitted to be con- under 1987 Japanese side research plan.
structed in Japan. Under this program in Japan,
the design guidelines for low- to medium-rise 5 . 1984 RESEARCH WORKS IN JAPAN
reinforced masonry buildings (up to about five 5 . 1 Test Plan on Materials and Construe t Ion
stories maximum) are being proposed. Method
The following items are Important to realize low'
In addition to this final target, the following medium-rise masonry buildings in this research
objectives in structural design of masonry program.
buildings are considered: a) The prediction of the compressive strength of
to ensure sufficient aseismic performance: prism specimens from properties of block units
Aseismic performance which is substan- grout concrete, and Joint mortar:
tially equal to the one to be required for 1) basic poperties of units;
present wall type R/C buildings in Japan 2) the effect of particle size distribution
is expected for newly developed reinforced of aggregate on the strength of lolnt
masonry buildings under the program. mortar;
412
3) the effect of joint mortar on the prism marized in Table 3. It is expected to be clari-
strength; fied from these test series how much effect each
4) the effect of grout concrete or grout mortar factor has on deformation performance and shear
strength on the prism strength. capacity.
b) Possibility of high lift grouting to execute 5 .3 Masonry Walls

construction works fast: Test specimens of walls are listed in Table 4 and
It is necessary to confirm fullness of grouting an example, Specimen WSRl (shear failure type) is
and bond between reinforcing bars and grout con- shown in Fig. 7. Photo 1 shows the specimen WSRl
crete or grout mortar. Eight specimens, six after testing. The following test series were
concrete block and two clay block specimens were considered
tested. Test specimen is shown in Fig. 6. a) Tests on shear capacity and deformation per-
c) Applicability of lap joints on vertical rein- formance (WS specimens)
.2 o
forcing bars: ‘Average compressive stress a = 20kg/cm ,
Now, in Japan, lap joint in masonry structures and for specimens with symbol "N" a 0 = 40
is inhibited. It is necessary for easy execution and 60kg/ cm^
of works to confirm effectiveness of lap joints. ‘Main reinforcing bars a t = Dl6 , 2-D19 and
Six types of twelve specimens as listed in Table 2-D25
1 were tested. •Horizontal shear reinforcement ratio Pwh =
d) Anchoring strength of reinforcing bars: 0 to heavy (0.668%)
Large amount of reinforcing bars would be set in •Shear span ratio M/OD = 0.45, 0.75 and 1.13
a small unit hole to resist seismic horizontal b) Test on flexural deformation performance (WF
loading to vertical components of buildings. specimens)
Tests on the effect of face shell and the thick- ‘M/QD = 0.75 and 1.13
ness of covering grout concrete were carried out •Ratio of shear strength to bending strength
using the specimens as listed in Table 2. Qsu/Qmu = variable
2
‘Average compressive stress a = 5kg/cra
5 Test Plan on Masonry Walls and Beams c) Test on effect of transverse wall on strength
In 1984 static tests of isolated walls and beams and deformation performance (WT specimens)
constructed by Japanese practice were conducted *T and + types
to obtain basic knowledge of five-story masonry d) Test on effect of ladder steel on shear capa-
structures. The major purpose is to investigate city and deformation performance (J specimens)
their deformation performance and shear capacity. e) Test on effect of unit shape on shear capacity
Therefore, the shear-span ratio, amount of cross and deformation performance (Bl specimens)
sectional area of main reinforcing bars and f) Test on effect of open joint work on shear
amount of shear reinforcement of test specimens capacity and deformation performance (0 speci-
were mainly varied in this test program. Prism mens)
compressive strength, f'm, and ratio of vertical g) Reinforced concrete wall test to compare shear
shear reinforcement sectional area to the gross capacity and deformation performance with
concrete area of a horizontal section of the those of masonry walls (RC specimens).
wall, PWV ,
are constant values and those are
f'm = 200-250kg/cm^ (would be 240kg/cm^) and 5 .4 Masonry Beams
Pwv - 0.25%, respectively. Test specimen of beams are listed in Table 5, and
the test specimen, GSRl (shear failure type) is
Test series of masonry walls and beams are sum- shown in Fig. 8. The following test series were
413
considered 7 . ACKNOWLEDGMENT
a) Test on shear capacity and deformation perfor- The authors wish to express their sincere thanks
mance (GS specimens) to TECCMAR members for their coordination of
• at = 2-D16, 2-D19 and 2-D25 Japanese technical research planning under the
*Pw = light (0.16%) to heavy (0.67%) U.S. -Japan Joint Program on masonry building
•M/QD = 0.71 and 1.18 structures. The authors also express their
b) Test on flexural deformation performance (GF gratitude to PROCMAR members for their contribu-
specimens) tion to promote U.S. -Japan cooperative research
•Qsu/Qmu = variable on masonry structures in Japan.
c) Test on beams with slab (GFl, GSl and GF2

specimens)
•Slab thickness = 15cm
•Slab width = lm on both sides of a beam
•Slab reinforcement = 2-Dl0@300
d) Test on effect of unit shape (Bl specimens)
e) Test on effect of Hoop shape (H specimens)
5 .5 Loading System and Sequence

The loading system used in wall and beam tests
is shown in Fig. 9 and in Photo 2. This loading
system consists of three actuators. A and B
actuators are controlled by a computer in order
not to produce any angle of rotation at top and
bottom of a specimen under constant axial load.

C-actuator is controlled by another computer in
order to apply horizontal force to a specimen
under predetermined loading sequence as shown in
Fig. 10.
6. CONCLUSION
The Japanese research plan for masonry building
structures under the U.S. -Japan Joint Earthquake
Research Program involving LargeScale Experi-
ments was presented. There are few masonry
buildings of concrete and clay block in Japan
compared with the other types of structural
systems such as timber houses, reinforced con-
crete and steel buildings. Through the five
years epoch-making research program concerning
low/medium-rise reinforced masonry buildings,
masonry structural systems in Japan are being
considered as a viable and safe construction
system and are increasing in number.
414
Table 1 Table 2
Test Specimen of Lap Joint Applicability Test Specimen of Anchoring

of Reinforcing Bars
LENGTH OF WIRE JOINT SPIRAL SECTION
REINFORCEMENT REINFORCEMENT TEST

SPECIMEfK LAP JOINT
METHOD DRCING BAR
1
- - -
- \S / S 7 > face SHELL
SECTION
n
^FACE
!n n n n in n
2 - - -
~T r SHELL
^REINFORC ING
LSI i
:
;
x:
3 40d - - U L
BAR u u
I
LOCATION c= CENTER
OF BAR a=1cm a=3cm b=1cm b=3cm FACE SHELL OF
4 40d 400400 - +1cm BLOCK
D- 16 - 3 - - - 1*
- 4050 PITCH 0-19 3 3 3 3 3 1*

5 40d
tyr 0-22 - 3 - - - 1*
6 40d 400400 4050 PITCH
Table 3 Test Series of Wall & Beam Specimens
SPECIMENS
SERIES TOTAL
STANDARD C/B B1 C/B CLAY BLOCK R/C
A ) WALL A-1)M/QD WS1 ,WS4 ,WS7 - WSR ,WSR4 ,WSR7

1
WSRC ,WTRC
WTT1.WTT2 (5) (3) (2) 10
A-2 ) FLEXURE WF1 ,WF2 ,WTC1 WFB1 ,WFB10 WFR1 ,WFR2 WFRC
WFL1 ,WFLM ,WFJ
WFJ2 (7) (2) (2) (1) 12
A- 3 ) AX I AL FORCE WSN1 ,WSN2 WSB1 WSB

, 1
- -
(2) (2) 4
A-4)Pw WS2 ,WS4 ,WS5 WSR2 ,WSR4 ,WSR5

WS6 ,WS8 ,WS9 - WSR6 -
WS10 (7) (4) 11
A- 5) At WS3,WS4,WS1 WSR3
WS2.WSJ1 ,WSJ2 - -
(6) (1) 7
SUB-TOTAL 27 4 10 3 44
B ) BEAM B-1)H00P GS2 ,

GSH1.GSH2 GSB1 - -
(3) (1) 4
B-2 ) Pw GS2 ,GS3 ,GS4 • - GSR1 ,GSR2 -

(3) (2) 5
B-3 ) FLEXURE GF3 ,GF4 ,GF5 - GFR1 ,GFR2 ,GFR3 -

GF6 (4) GFR4 (4) 8
B-4 ) SLAB GF1 ,GS1 ,GF2 - - - 4

GF3 (4)
SUB-TOTAL 14 1 6 - 21
TOTAL 41 5 16 3 65
415
Table 4 Test Specimen List of Walls
Pw D1 30400 D1 30200 2D1 30200

0
(0.167%) (0.334%) (0.668%)
TRANSVERSE^\
WALL
M/QD At(cm 2 )
X O X 0 X O X O
D16
0.45
(1.99)
h/1 - 2D1 WS1 ,WSR1
180/200 (5.74)
2D25
(10.1)
D16 WS3.WSR3
0.75
WSR2 WS4*3,WSB1 WTT1 WS5 WF2.WFB1 WTT2
2D1 WSJ1 WSR4*2 ,WSN1 WTC1 WSR5 WFR2
h/1 =
WS2x2 WSN2.WF1 ,WFR1 WTRC WFJ2
180/120 WFL1 ,WSRC,WFJ1 WFB10
WELM.WFRC WS6
WSB10.WSJ2 WSR6
2D25 WS8 WS11 WS9 WS10
D16
1.13
h/1 = 2D1 WS7.WSR7
180/80
2D25
PRISM STRENGTH fm’ =250Kg/cm 2 VERTICAL REINFORCEMENT D160400
WS ;
SHEAR FAILURE TYPE B1 ; 300mm LENGTH C/B SPECIMEN
WF ; FLEXURAL FAILURE TYPE RC ; R/C SPECIMEN
WT ; TT:T-SHAPE TC: CROSS-SHAPE SPECIMEN
, 0 ; OPEN JOINT SPECIMEN
R ;
CLAY BLOCK N ; AXIAL FORCE SERIES SPECIMEN
J ;
EFFECT OF LADDER STEEL
Table 5 Test Specimen List of Beams
D1 30400
Pw D1 30200 2D1 30200
(2D1O04OO)* 1
D130150* 2 D160150* 2
D 1 30400 *
2
R/C SLAB \
M/QD At(cm 2 )
o X O X O X
2D16 GF1
0.71
(3.98)
l/h= 2D19 GF6.GFR4

120/95 (5.74)
2D25 GS1 GS2.GSR1 GS3.GSR2

(10.1) GSH1 , GSB
GSH2
1.18 2D16
l/h= 2D1 GF2 GF3.GFR1 GF4.GFR2 GF5.GFR3

200/95
2D25 GS4
PRISM STRENGTH fm' =250Kg/cm 2 HORIZONTAL REINFORCEMENT 0160400

*1 FOR HOOP SERIES SPECIMEN
* 2
FOR 300mm LENGTH UNIT SPECIMEN
GS ; SHEAR FAILURE TYPE
GF ; FLEXURAL FAILURE TYPE B1 ; 300mm LENGTH C/B SPECIMEN
R ; CLAY BLOCK H ; HOOP SERIES SPECIMEN
IQ6
First Workshop for masonry project was held on March 16, and 17, 984 1
Two planning meetings, one in the middle of 1985, and the other in early 1986, are proposed.
The first one will be held In Japan, and the second In USA.
NSF : National Science Foundation, STA : Science Technolongy Agency, HOC : Ministry of Construction
Fig.1 Schedule of U.S. -Japan Joint Research Program
Fi g . 2 Organization for Promoting U.S. -Japan Joint Research

on Masonry Building Structures
<07
Fig. 3 Japanese Research Plan on Masonry Building Structures
GROUTING
1 » 1.000 !
Fig. 6 High-Lift Grouting

Test Specimen
i
1
s.
4-
i
s'
265
Fio.5 Standard Clay Block Unit

in Japanese Research Program
<03
Hq. 7 Test Specimen of Wall : WSR1 Fig. 8 Test Specimen of Beam : GSR1
Fig. 9 Loading System
419
a) Shear Failure Type
b) Bending Failure Type
Fig. 10 Loading Sequence
420
Photo 1 Specimen WSR1 after Tested
Photo 2 Loading System
421
INELASTIC SEISMIC TESTS ON A FULL-SCALE
SIX-STORY ECCENTRIC-K BRACED STEEL BUILDING
- U . S .-JAPAN COOPERATIVE EARTHQUAKE RESEARCH PROGRAM -
BY
Hi royuki YAMANOUCHI 1 ,
Mitsumasa MIDORIKAWA 2
Isao NISHIYAMA 3 and Masaya HIROSAWA4
ABSTRACT
As part of the U.S. -Japan Cooperative Earthquake 500 gal in the 2nd stage test. Sinusoidal excita-
Research Program Utilizing Large-Scale Testing tion input test followed to determine the
Facilities full-scale seismic tests on a steel strength, def ormabi li ty and failure mechanism of
building were performed as a three-year research the eccentric-K braced structural svstem in the
project from 1981. This research involves com- ultimate stage.

paring full-scale building behavior with scale
model behavior and assesses the damage and 1. INTRODUCTION
safety levels of buildings subjected to earth- As part of the U.S. -Japan Cooperative Earthquake
quakes designed by current design practices. Research Program Utilizing Large-Scale Testing
Facilities (Ref. 1), full-scale seismic tests on
Reported are results of tests on a full-scale a six-story steel office building were carried
six-story steel building with eccentric-K out at the Building Research Institute of the
braces which followed the tests on a full-scale Japanese Ministry of Construction to compare ac-
six-story steel building with concentric-K tual full-scale building behavior with scale
braces. The latter results were reported pre- model or member behavior and to assess the dam-
viously. age and safety levels of buildings subjected to
earthquakes, which were designed to satisfy the
A six-story, 2x2 bay, steel building constructed requirements of current seismic design codes.
inside the Large-Size Structure Laboratory of
the Building Research Institute of the Japanese Reported are the results of the tests on a full-
Ministry of Construction, had its buckled con- scale six-story steel building with eccentric-K
centric-K braces cut off. The field welding of braces which followed the tests on the building
the new eccentric-K braces to the structure with concentric-K braces (Ref. 2). Overall tests
followed the repair of the cracked R/C slabs by were carried out according to the test sequence
epoxy injection. listed on Table l.
All of the seismic tests were run as a six de- ^Head of Structure Division, Department of Stru.
gree of freedom pseudo-dynamic system (computer tural Engineering, Building Research Institute
on-line system). The input excitation was the ^Chief Research Structural Engineer, Department
NS component of the 1952 Taft Earthquake and of Structural Engineering, RRI
the maximum intensity was set at two levels. ^Research Structural Engineer, Department "t
To simulate working load conditions, the earth- Structural Engineering, RRI

quake motion was scaled to 65 gal in the 1st 4 Director of Research Planning and Information,
stage test. The maximum earthquake was run at Department, RRI, Tsukuha, Tharakl IDS ,
lAÂN
422
All of the concentric-K braces, some of which idual vertical displacement was about 2.5 cm at
had buckled in PHASE I TEST (Ref. 2), had been the midspan of the girder where braces are con-
cut off and the cracked R/C slabs were repaired nected. The residual deformations were in the
by epoxy injection. The eccentric-K braces were allowable range, only the epoxy injuection of the
newly installed and welded to the existing principal cracks in the composite R/C slabs was
frames by field welding. conducted before PHASE II TEST. The epoxy injec-
tion was performed within 1 m width in both sides
All of the seismic tests were run as a six de- of the girder parallel to the loading direction
gree of freedom pseuo-dynamic system (computer and within 0.5 m width in both sides of the gird-
on-line system) that was applied to the er parallel to the loading direction and within
concentric-K brace tests. The input excitation 0.5 m width in both sides of the girder trans-
was the NS component of the 1952 Taft Earthquake verse to the loading direction. Other cracks
and the maximum intensity of it was set at two wider than 0.2 mm were also repaired. Epoxy in-
levels. To determine the strength, deformabil- jection repairing was carried out after removing
ity and failure mechanism of the eccentric-K the concentric-K braces by gas.
braced structural system in the ultimate stage,

sinusoidal excitation input test followed these 3. DESIGN OF ECCENTRIC-K BRACES AND THEIR
seismic response tests with the test buiLding INSTALLATION
deformed into the ultimate stage. The plan and elevation of the eccentric-K braced
building are shown in Fig. 1. The detail of the
2. REPAIR OF FULL-SCALE TEST BUILDING eccentric-K braced frame is shown in Fig. 2.

Prior to the tests on the concentric-K braced Principal member size is tabulated in Table 2.
building (PHASE I TEST), discussions were held The eccentric-K braces were installed in the
between Japanese and U.S. researchers on the braced frame between Lines 2 and 3 after the con-
allowable degree of damage of the test building centric-K braces between Lines 1 and 2 were re-
in PHASE I TEST. The following agreements were moved. The sectional shapes of the braces were
made on the test conditions in PHASE I TEST and rectangular tubes in inch size and they were ar-
on the repair conditions before PHASE II TEST. ranged so their weak axes coincided with the
plane of the braced frame. The braces were se-
1) Maximum story drift angle should be less than lected so they would not buckle until the shear
about 1/60 radian. panels in the griders yield due to the shear
2) Member change or deformation reform is not force produced by the brace axial force. Each
necessary if residual member rotation angle of the shear panels was the web plate of the
is less than about 1/300 rad. in the prin- girder of its length 711 mm, which was surround-
ciple member. ed by both of the flange plates of the girder
3) If the residual vertical displacement at the and a pair of vertical stiffeners (2-PL19) to
midspan joint of each girder (girder in the transmit the brace axial force into the girder.
braced frame between lines 1 and 2 in Fig. 1)
where the concentric-K braces are connected This structural system is aimed to make the shear
is larger than about 3 cm, the girder should panels yield prior to the buckling or yielding of
be changed by new one. the eccentric-K braces, to absorb the energy in
4) Large cracks in the composite R/ C slabs these panels, and to show stable restoring force
should be repaired by epoxy injection. characteristics. As shown in Fig. 2, three ver-
tical stiffeners (PL12) were welded on one side
Immediately after PHASE I TEST, the maximum res- of each of the shear panels to prevent the panels
i
23
\
from shear buckling in the early stage of load- number of measurements was 980.
ing. Sub beams in the transverse direction were
installed to restrain the distortion of the gird- 6. NATURAL PERIODS AND DAMPING RATIOS
er at both ends of the shear panel on one side Table 5 shows the natural periods and damping ra-
of the girder (Fig. la). All the gusset plates. tios observed in the Free and Forced Vibration
stiffeners and braces were welded to the central Tests, for Each Floor Level Loading Test, for the
frame in the field. PSD Free Vibration Tests and for the PSD Pulse
Response Tests which were compared with those
After the repair of the R/C slabs and the instal- calculated by frame analysis (Ref. 3).
lation of the eccentric-K braces, straightness
along the height of the test building was mea- Natural periods of each of the modes obtained by
sured before the Each Floor Level Loading Test the tests conducted prior to the seismic response
(Table 4). The measured data are listed in tests were a little smaller but coincided well
Table 3. The residual roof horizontal displace- with those calculated using frame analysis.
meat was 1.1 cm and the maximum residual story Periods of the 1st mode were about 4 to 9% short-
drift was 0.53 cm (story drift angle: 1/645 er than those of the concentric-K braced building
rad.) in the 6th story. These initial un- (Ref. 2). The fundamental natural period after
straightnesses of the test building were about PHASE II TEST was almost 1.2 times longer than
the same order of magnitude as the construction that before PHASE II TEST. The Increase in na-
accuracy. Therefore, it may be assumed that the tural period was a little smaller than those (1.3
structure before the PHASE II TEST was in almost - 1.4 times) observed in the test of the concen-
the same condition as those of the test build- tric-K braced building (Ref. 2).
ings just after construction.
The damping ratios of 1st and 2nd modes obtained
5. TEST SEQUENCE by the Forced Vibration Test conducted just be-
The tests on the eccentric-K braced building fore the seismic response tests were 0.35% and
(PHASE II TEST) were carried out in the sequence 0.31% respectively. They were almost the same as
shown in Table 4. The main tests consisted of those (about 0.5%) obtained in the PHASE l TEST
two seismic response tests the Elastic Pseudo- (Ref. 2). On the basis of these results, Rayleigh
Dynamic (PSD) Test and the Inelastic PSD Test. type damping was used in the PSD tests; both 1st
The maximum intensity of the input excitation and 2nd damping ratios were set at 0.35%. How-
was set at two different levels for earthquake ever, the apparent 1st damping ratios observed 1 .i
excitation and at three levels for the Inelastic the PSD Free Vibration Test and the PSD Pulse Re-
PSD Sine Wave Input Test. Other tests were per- sponse Tests were 1.17-3.4% and they were larger
formed before and after the main tests so as to by 0.82-3.05% than those (0.35%) assumed. This
get the dynamic characteristics of the test difference seems to come from the additional
building and also the characteristics of the PSD damping produced by the testing system as ex-
test system as was done in the PHASE I TEST. In plained in the previous paper (Ref. 2). The
the PHASE II TEST, the PSD Pulse Response Test amount of this additional damping depanda botl
was also performed to examine the dynamic char- the control conditions of the testing system and
acteristics in the higher modes. Displacement the displacement amplitude of response of the
and strain measurements in the PHASE II TEST test building.
were almost the same as those in the PHASE I
TEST, and several measurements around the shear

panel in the girder were newly added. The total
424
7 . PSEUDO-DYNAMIC SEISMIC RESPONSE TESTS test was continued up to 17.92 second. The time
AND SINE WAVE INPUT TESTS histories of the horizontal displacements at the
The pseudo-dynamic seismic response tests and 3rd, 5th and roof floor levels are shown in Fig.
the sine wave input tests were carried out by 4. The maximum roof horizontal displacement was
using a six degree of freedom pseudo-dynamic 1.41 cm at 12.39 sec., the maximum story drift
testing techniques the same as that used in angle was 1/1339 rad. in the 3rd story (story
the concentric-K brace tests. In the pseudo- drift: 0.25 cm) and the maximum shear force in
dynamic seismic response tests, the input ex- the 1st story was 59.6 tonf, which corresponds
citation was the NS component of the 1952 Taft to 0.114 of the base shear coefficient (ratio
Earthquake as shown in Fig. 3 and the maximum between the shear force in the 1st story and the
intensity of the excitation was set at two total weight of the building - only the actual
levels, that is, 65 gal to simulate the working weight of the test building is used in this esti-
Load conditions (Elastic PSD Test) and 500 gal mation) . These maximum responses were about one-
as the maximum earthquake (Inelastic PSD Final half of those obtained in the concentric-K brace
Test) (Table 4). In the sine wave input tests, test (PHASE I TEST - Elastic Test (3) (Ref. 2).
the maximum acceleration was set at three The dotted lines in Fig. 4 show the results of
levels; 92.6-100 gal, 270 gal and 320 gal. the frame
9. analysis (Ref. 3). As the damping ra-
These sine wave input tests were carried out tios in the analysis (0.5% in the 1st mode) dif-
in place of the static cyclic loading tests. fered from those in the test (1.2-1. 3% in the
In these tests, the periods of the input sine 1st mode) ,
the maximum responses obtained in the
waves were so selected that they coincided test were about less than half of those calcu-
with those of the 1st mode of the test build- lated in the analysis. The relationship between
ing in the stationary response conditions story shear force versus story drift in each of
under the different excitation levels. The the stories was almost linear.
tests were continued for about 0. 5-1.0 cycle

on each of the excitation levels. INELASTIC PSD FINAL TEST
After the Elastic PSD Test, the NS component of
The maximum responses obtained from these tests the 1952 Taft Earthquake scaled to 500 gal was
are tabulated in Table 6. The time intervals input and the test was continued for 17.10 secs.
used for the calculation in PSD tests are 1/100 The time histories of the horizontal displace-
sec. in the Elastic PSD Test, 1/200 sec. in the ments at the 3rd, 5th and roof levels are shown
Inelastic PSD Final Test and 1/100 sec. in the in Fig. 5 and the relationships of story shear
Inelastic PSD Sine Wave Input Test. In the force vs. story drift in the 1st to 4th stories
following sections, the test results and the are shown in Fig. 6. In Fig. 7 are shown the
frame analysis results (Ref. 3) are described. shear force vs. shear deformation relationships
In the frame analysis, the damping ratio of the in the shear panels of the web plates of the
1st mode was assumed 0.5% in the eleastic re- girders where the eccentric-K braces are connect-
sponse and 2% in the inelastic one. The effect ed. The shear force is estimated from the verti-
of the yielding of the members which occurred cal components of the axial forces of a pair of
in the PHASE I TEST was not considered in the braces. The roof horizontal displacement reached
frame analysis. almost 8.6 cm at around 7.5 sec. and reached its
maximum (8.8 cm) at 14.505 sec. As shown in Fig.
8. ELASTIC PSEUDO-DYNAMIC TEST 6, the relationship of story shear force vs. story
The input wave was the NS component of the 1952 drift obtained in the first story showed stable
Taft Earthquake motion scaLed to 65 gal and the hysteresis loops without deterioration. As shown
425
in Table 6, the maximum responses of the story levels are shown in Fig. 8, and the relationships
drift angle are larger in lower stories, that is, of story shear force vs. story drift in the lowest
1/150 rad. in the 1st story, 1/156 rad. in the four stories are shown in Fig. 9. In these fig-
2nd story, and 1/225 rad. in the 3rd story. The ures, the test results obtained in the three tnde-
maximum shear force in the 1st story was 312.8 dependent tests are simply combined and shown.
tonf, which corresponds to 0.597 of the base Therefore, the figures don't represent continuous
shear coefficient. In this estimation, only the responses
actual weight of the test building was consider- In Figs. 8 and 9, the origins of the figures coin-
ed as the building weight. The maximum shear cide with the residual displacements at the begin-
force corresponds to 0.383 of the base shear co- ning of the sine wave input tests; 0.423 cm,
efficient if live loads for seismic design and 0.426 cm, 0.383 cm, 0.337 cm, 0.303 cm and 0.185
the exterior wall weight plus actual dead weight cm from the roof level to the 2nd floor level.
of the test building (816 tonf) are considered. Therefore, the maximum displacements in these
The maximum displacement responses obtained were figures are smaller than those shown in Table 6
about one-half or less than those obtained in by the residual displacements after the Inelastic
the concentric-K brace test (PHASE I TEST - PSD Final Test.
FINAL TEST, Ref. 2). The shear deformations in As shown in Fig. 9 the story shear force vs. storv
the shear panels, where several yieldings were drift relationships showed slight successive de-
observed, were relatively large in the lower gradation in the last one-and-a-hal f cycle. This
three stories as shown in Fig. 7, and the shear degration was due to the local deformation In the
deformation angle reached 0.047 rad., 0.029 girder web and the out-of-plane deformation of the
rad. and 0.015 rad. in each of the stories from gusset plate near the joint (Fig. 2) of the braces
the lowest at around 14.5 sec. Although these in the 1st story and the girder of the 2nd floor
shear panels yielded by a large amount, they level. Because of the out-of-plane deformation
retained a relatively large stiffness in the of the gusset plate, lateral-torsional deformation
plastic range. In addition to the yielding in occurred in the girder and the braces In the 1st
the shear panels, other yielding was observed story started to buckle. Therefore, the test was
in the column and in the gusset plates connect- stopped when the roof horizontal displacement
ing braces and girder. However, the apparent reached its maximum 23.2 cm. The story drift
damage is not so severe. In Fig. 5, the re- angles in the lowest three stories were larger
sponses calculated by frame analysis (Ref. 3) than those in the other stories as in PHASE I
shown by dotted lines are about 30% larger TEST and they finally reached 1/46. 4rad. in the
than those obtained in the test. This is be- 1st story, 1/52.7 rad. in the 2nd story and 1 / 96.7
cause the damping ratio used in the frame an- rad. in the 3rd story as shown in Table 6.
alysis (2%) is small compared with that ob- The maximum story shear force in the 1st story was
observed in the test (3.4%). 359.9 tonf which corresponds to 0.687 of the ham-
10. INELASTIC PSD SINE WAVE INPUT TEST shear coefficient, and at this stage the st<>rv
In the Inelastic PSD Final Tests, the damage of drift in the 1st story was 4.4 cm (storv drift
the test building was rather smaller than ex- angle: 1/102 rad.). In this estimation, only tie-
pected. In order to find out the strength, de- actual weight of the test structure was considered
formability and failure mechanism of the eccen- as the building weight. If live loads for s.isii
tric-K braced structural system in the ultimate design and exterior wall weight besides ictunl
stage, three levels of sine wave input tests weight of the test structure are considered, tin-
were conducted. The time histories of the hor- base shear coefficient changes to 0.44 1. Th<-
zontal displacements at the 3rd, 6th and roof imum horizontal displacements were aim. -. th«-
426
as chose obtained in the PHASE I TEST (Final gusset plate at the brace-girder junction were
Test). With respect to the deformation of the observed. Lateral- tors ional deformation of the
shear panel in the girder and its surroundings, girders occurred associated with the out-of-plane
out-of-plane buckling of the gusset plates con- buckling of the braces.
12.
necting 1st story braces and 2nd floor girder
and the yielding of the bottom of the 1st story ACKNOWLEDGMENTS
columns (C5-column in Fig. 1(a)), some photo- The authors wish to express their gratitude to
graphs
11. showing these effects are presented in the members of the J.T.C.C. (the Joint Technical
Ref. 3. Coordinating Committee of the U.S. -Japan Coopera-

tive Research Program: Co-chairman; Profs. H.
CONCLUSIONS Umemura and J. Penzien) who encouraged the authors

From the results of the seismic response tests and cordially gave advice. The authors are also
on a full-scale six-story eccentric-K braced thankful to the visiting researchers at B.R.I.

steel building using the multi-degree of freedom and from the general contractor companies, for
pseudo-dynamic testing system at the Building their excellent help in the analysis, tests and
Research Institute ( Tsukuba-Japan) ,
the shear data processing. The research staff of the Struc-
panels in the web plates yielded by a large tural Engineering Department and Production Dept.,
amount as was intended in the design philosophy of the International Institute of Seismology and
for an eccentric-K braced structural system. It Earthquake Engineering in V.R.I., and the U.S.
13.
was verified that this type of structural system visiting researchers are also acknowledged. This
has a remarkable capacity for energy absorption. research was supported in part by the Japanese
According to the test results obtained in the Ministry of Construction, the Kozai Club, the
Inelastic PSD Final Test (the maximum intensity Building Contractors Society, and the National
of the NS component of the 1952 Taft Earthquake Science Foundation of the U.S. A.
was scaled to 500 gal) ,
the maximum base shear
coefficient was 0.597, the roof horizontal dis- REFERENCES

placement reached its maximum 8.8 cm and the
1. U.S. -Japan Planning Group,
maximum story drift angle reached 1/150 rad. in "Recommendations for a U.S. -Japan
the 1st story. At the same time, a large amount Cooperative Research Program Utilizing
Large-Scale Testing Facilities," Report
of shear yielding was observed in the lowest No. UCB/EERC-79/26, University of
three stories and the shear deformation angle California, Berkely, Californis, Sept., 1979
2. Yamanouchi, H. Midorikawa M.
, ,
attained 0.047 rad. in the shear panel of the Nishiyama, 1. and

Hirosawa, M. " Inelastic Seismic Tests
second floor level. In spite of a large amount ,
on a Full-Scale Six-Story Concentric-K

of yielding in the shear panels their hysteresis
Braced Steel Building - U.S. -Japan
loops were quite stable and the observed damage Cooperative Earthquake Research Program,
Proceedings of the 16th Joint Meeting of
of the test building was modest.
U.S. -Japan Panel on Qind and Seismic
In the PSD Sine Wave Input Test, the maximum Effects, U.J.N.R., National Bureau of
Standards,
base shear coefficient was 0.687, the maximum
Gaithersburg, Maryland, May 15-18, 1984
roof horizontal displacement was 23.2 cm and the 3. Goel, S. C., and Foutch, D.A.,
"Preliminary Studies and Test Results of
maximum story drift angle was 1/46.4 ra. in the
Eccentrically Braced Full-Size Steel
1st story. The strength of the test building Structure, U.S. -Japan Cooperative
Earthquake Research Program Utilizing
gradually decreased and the test building almost
Large Scale Test Facilities," Proceedings of
reached its ultimate strength stage. Finally, the 16th Joint Meeting of U.S. -Japan
Panel on Wind and Seismic Effects,
local deformation in the girder web next to the
U.J.N.R., National Bureau of Standards,
shear panel and out-of-pLane buckling of the Gaithersburg, Maryland, May 15-18, 1984
427
1.
TABLE 1 WHOLE TEST PROGRAM

PHASE I TEST Apr. 11 - Oct. 21 1933
2. Concentrically K-3raced Frame Test
Input Motion: 1978 Miyagi -Ken-Ok i Earthquake
Tohoku University NS
Maximum Input Acceleration:
Elastic Test - 65gal
Inelastic Test - 250aal and 500gal
3.
PHASE II TEST Nov. 24 1983 - Mar 9 1984 .
Eccentrically K-3raced Frame Test

Input Motion: 1952 kern County Earthauake
4. Taft NS(N21°E)
and Sine Wave
Maximum Input Acceleration:
Elastic Test - 65oal Inelastic Test - SOOgal
Sine Wave Test - 92.6- 320gal
PHASE III TEST Mar. 21 - 28 1984

Unbraced Moment-Resisting Frame Test
Input Motion: 1940 Imperial Valley Earthquake
El Centro NS
Maximum Input Acceleration: Inelastic Test - 350gai
PHASE IV TEST Apr. 23 - July 19 1984

Unbraced Frame With Nonstructural Elements Test
Loading Condition- Displacement Control
Maximum Interstory Drift Angle: 1/40
TABLE 2 MEMBER SIZE AND PROPORTION
||
(A) GIRDER SCHEDULE (C) 3RACE SCHEDULE
\MARK mark
G1 Gc Gj G4 ' BRt
FLOOR 1 1
STORY"'
A 16 x 31 ;
W 16 x 31 !
A 18 x 35 W 21 x 50 6 - 5 Tube 3x6x0.313
RFL-oFL H-403 .4x140.3 H-A03 4x 40. 3 . 1 H-449 5x1 52 4 . .
'
H-529. xl 65. 9
1 \
Box-203. 2x1 52 4x7. 95 .
______
x6. 98x1 1. 13, x6. 93x11. 13 x7. 62x10. SO
_ _
x9. 55x13. 59
_
4-1
. ,
.
uoe 8xox0.375
3ox-203 2x SC 1x9 53
5FL |
W 16 :< 31 j
W 18 x 35 W 18 :< 35 ;
M 21 x 50 . 1 .
4FL W 18 x 35 |
A 18 x 35 A 18 x 35 U 21 x 50
‘
lote: AS ™
A50 ° GSAC;; 3
1
_ Structural ste»l
W 18 x 35 W 18 x 40 j
M 18 x 35 W 21 x 50
3FL H-454. 7x152.3
x3. 00x13. 34
2FL A 13 x 40 |
A 18 x 40 A 18 x 35 W 21 x 50
Note: ASTM A36 structural steel
(S) COLUMN SCHEDULE
\MARK
Cl C2 C3 C4 C5
STORY
A 10 x 33 A 10 x 33 A 10 x 49 A 10 x 33 W 12 x 40
6-5
H-247 . 1 x2C2.2 H-253. 5x254.0 H-303. 3x203.
x7. 37x1 1 .05 x8. 64x14. 22 x7. 49x13. 08
W 10 x 39 A 12 x 53 W 12 x 65 A 10 x 60 W 12 x 72
4-3
H-252.0x202.3 H- 306.3x253.9 H-307.3x304.3 H- 259 5x256.0 . H-311 .2x305.3
x8. 00x13. 46 x8 76x 4 60
. 1 . x9. 91x15. 37 x 1 0 57x
. 1 7.27 xlO. 92x17. 02
A 12 x 50 A 12 x 55 W 12 x 79 A 12 X 79 A 12 x 106
2
H-309.5x205.2 1 H-314.5x306.3 H-327.4x3l0.4
x9. 40x16. 26 I xll. 94x13. 67 x 1 5 .49x25 . 1
W 12 x 65 W 12 x 37 A 12 x 37 W 12 x 106 U 12 X 136
1
H-318. 3x308.0
xl 3. 08x20. 57 x20 0/x.ll 7*.
Note: ASTM A36 structural steel

TABLE 3 RESIDUAL HORIZONTAL TABLE 4 TEST SEQUENCE
DISPLACEMENTS BEFORE
PHASE II TEST Gas Cutting Works of Concentric K-8races
Repair Works of Concrete Floor Slabs by Epoxy Injection
INTERSTORY DRIFT Welding Works of Eccentric K-Braces
DISPLACEMENT AT
1983
FLOOR FLOOR LEVEL STORY DISPLACEMENT ANGLE
1. Free and Forced Vibration (VT) Test #3 Nov. 24-25
(cm) (cm) (radian)
2. Each Floor Level Loading (FLL) Test # Jan. 17
3. Elastic Pseudo-Dynamic (PSD) Test Jan. 18-26
RFL 1.124 6 0.527 1/645
(Maximum Input Acceleration 65qal ) 1984
6FL 0.597 5 0.199 1/1709
4. PSD Free Vibration (F-Vt) Test y4 Jan. 30
5FL 0.398 4 -0.390 1/872
1/3778 5. PSD Pulse Response Test rl Jan 31
4FL 0.788 3 0.090
6. PSD Pulse Response Test 42 Feb. 1-3
3FL 0.698 2 0.260 1/1308
1/1027 7. PSD Pulse Response Test -3 Feb. 2
2FL 0.438 1 0.438
8. PSD Pulse Response Test 44 Feb. 4
9. Inelastic PSD Final Test Feb. 7-27
Note: Displacements in this table are a little Acceleration 500qal
(Maximum Input 1984
different from those measured just after
10. Inelastic PSO Sine Weve Input Test Mar. 1-5
Concentrically K-Braced frame (PHASE I) Test.
(Maximum Input Acceleration 92.6- 320qa 3 1984
11. Free and Forced Vibration (VT) Test 44 Mar. 8-9
TABLE 5 NATURAL PERIODS AND DAMPING RATIOS
NATURAL PERIOD(sec) DAMPING RATIO (%)

TEST
1st 2nd 3rd 1st 2nd 3rd
Analysis 0.595 0.214 0.119
Free VT Test #3 0.565x0 567 0.35x0. 37(0. 6x2. 2ton)*‘

Forced VT Test # 0.568 0.201 0.35 0.31
FLL Test #4 0.545 0.193 0.106
Elastic PSD Test - 6 5ga 1 peak input
PSD F-VT Test #4 0.553 1 ,25(6xl0mm)* 2

PSD Pulse Test #1 0.550 3.10(1 ,4x4mm)* 2
PSD Pulse Test #2 0.554 2. 68(2. 5x4. 5mm)* 2
PSD Pulse Test #3

PSD Pulse Test #4
0.550
0.553
—
1.17(4. 5x1 2mm)* 2
5.90(2.3x5mm)* 2
3.40(5xl2mm)* 2
Inelastic PSD Final Test - 500gal peak i nput

Inelastic PSD Sine Wave Input Test - 92.6x320gal peak input
Free VT Test A4 0.679x0.686 0.50x0.55(0.5x1 ,7ton)*‘

Forced VT Test #4 0.680 0.229 0.48 0.34
Note: *1) Pulling force *2) Roof displacement amplitude
429
TABLE 6 MAXIMUM RESPONSES
NOMINAL DISPLACEMENT INTERSTORV DRIFT BASE

TEST DAMPING AT FLOOR LEVELS DISPLACEMENT ANGLE SHEAR
COMBINATION (cm) (sec) (cm) (sec) (radian (ton)
RF -1.405 12.39 6 -0.226 12.39 1/1504

ELASTIC 0.350.35 6F -1.204 12.38 5 -0.240 12.40 1/1417 +59.6
PSD TEST 2 90 5F -0.988 12.37 4 -0.239 12.36 1/1423 12.66sec
90 90% 4F -0.749 12.37 3 -0.254 12.36 1/1339
65gal (1 st-6th) 3F +0.512 12.65 2 -0.228 12.37 1/1491 -59.2
peak input 2F +0.305 12.66 1 +0.305 12.66 1/1475 !2.93sec
INELASTIC RF -8.841 14.505 6 -0.935 7.275 1/364

PSD FINAL 0.35 0.35 6F -8.337 14.510 5 -1.236 7.285 1/275 +293.6
TEST 5 90 5F -7.560 14.515 4 -1.279 7.295 1/266 14. 175sec
90 90% 4F -6.586 14.515 3 +1.512 7.040 1/225
500gal ( 1 s t-6 th 3F -5.170 14.520 2 -2.173 14.525 1/156 -312.3
peak input 2F -3.004 14.520 1 -3.004 14.520 1/150 14 . 51 5sec
INELASTIC RF -23.239 1.75 6 + 1 .090 1.32 1/312

PSD SINE WAVE 0.35 0.35 6F -22.454 1.75 5 +1.428 1.36 1/238 +359.2
INPUT TEST 5 90 5F -21.308 1.75 4 +1.832 1.25 1/186 0.85sec
90 90% 4F -19.671 1.75 3 -3.517 1.75 1/96.7
92.6- 320ga ( 1 s t-6 th 3F -16.154 1.75 2 -6.457 1.75 1/52.7 -359.9
peak input 2F -9.697 1.75 1 -9.697 1.75 1/46.4 1 .03sec
Note: 1) Displacement and shear force to the left in Fig. are positive. 1
2) The origins of displacements at floor levels in Inelastic PSD Sine Wave

Input Test are the same as those in Inelastic PSD Final Test.
430
7,500
7,500
(b) ELEVATION FRAME J)
FIG. 1 PLAN AND ELEVATION OF TEST STRUCTURE
FIG. 2 DETAIL OF ECCENTRIC-K BRACED FRAME
431
TAFT GROUND ACCELERATION
TAFT
NORMALIZED ACC- RESPONSE max. imput aCC. 174.62 CAL
FIG. 3 INPUT MOTION AND RESPONSE SPECTRA
ELASTIC TEST 6S GAL FINAL TEST 500 Gal

3FL
: TEST ANALTS1S I
-•i
\ t '- y ,j v - AftrHV
- A A ft ft A A
* V u u J'V
U v «'*v
\i \j
v '
' ./
.. t/ v u
I; ^ v i/ij w; rr^)T
- M \j\lyVV7
i/
,/T; 1717 i/ .
"
. , A a A A a a < . 4 A ft A A A A a . A fl l\ ft A A
•* j
*r v 'J
y v, V w v < v
y
v
-
'vvy.r
' '
vw
FIG. 4 TIME HISTORIES OF HORIZONTAL FIG. 5 TIME HISTORIES OF HORIZONTAL

FLOOR DISPLACEMENTS FLOOR DISPLACEMENTS
(ELASTIC PSD TEST) (INELASTIC PSD TEST)
452
FIG. 6 STORY SHEAR FORCE VS. STORY DRIFT RELATIONSHIPS
(INELASTIC PSD TEST)
FIG. 7 SHEAR FORCE VS. SHEAR DEFORMATION RELATIONSHIPS

IN SHEAR PANELS (INELASTIC PSD TEST)
433
SINE WAVE INPUT TEST
TIME HISTORIES OF HORIZONTAL

FLOOR DISPLACEMENTS
(INELASTIC PSD SINE WAVE
INPUT TEST)
FIG. 9 SHEAR FORCE VS. STORY DRIFT RELATIONSHIPS

(INELASTIC PSD SINE WAVE INPUT TEST)
434
SEISMIC PERFORMANCE OF EXISTING BUILDINGS
A PROGRESS REPORT
by
Richard N. White and Peter Gergely

Department of Structural Engineering
School of Civil and Environmental Engineering
Cornell University
Ithaca, N.Y. 14853
ABSTRACT 1. INTRODUCTION
This progress report is based on two workshops
held in 1983 and 1984 on "The Evaluation of the The problem of the assessment of the safety of
Performance of Existing Buildings for Resistance various types of existing buildings in seismic
to Earthquakes." The workshops were organized zones has received increasing attention in
through Task Committee D, Evaluation of the recent years. Engineering design offices in
Performance of Structures, and were supported areas with high seismic risk as well as appro-
by the National Science Foundation (Grant CEE- priate government agencies devote a growing
8217190 to Cornell University). fraction of their efforts to the evaluation and
strengthening of existing buildings and other
The first workshop, held in Tsukuba in May 1983, structures
had thirteen representatives from Japan and
twelve from the U.S.; seventeen papers were The existence of a large number of unsafe
presented in five theme areas: (I) Overview of buildings, or buildings with unknown safety,
evaluation of buildings for seismic resistance, makes it necessary to conduct a thorough,
(II) Practical methodology, (III) Computer critical evaluation of current techniques, and
programming, (IV) Evaluation of structural to plan research that will further advance the
performance through full scale tests, and state-of-the-art of evaluation methodology.

(V) Post-earthquake evaluations. The second Extensive discussions of this subject within
workshop took place in July 1984 in Berkeley, UJNR Task Committee D, "Evaluation of the
California with twelve representatives from Performance of Structures" led to a recommen-
Japan and eighteen from the U.S.; eighteen dation that a workshop be held on evaluation
papers were presented in three theme areas: methods. The entire UJNR Panel, at its I4th
(I) Safety evaluation methods, (II) Evaluation joint in 1982, adopted the following resolution:
methods applied to benchmark structures, and "This panel agrees with the recommendation
( III) Damage and analysis of other structures. of Task Committee (d) on 'Evaluation of
Names of participants and paper titles are given Performance of Structures' that workshops
in Appendices A and B in this paper. be held on the 'Evaluation of the Perfor-
mance of Existing Buildings for Resistance
Highlights of the workshops are presented here, to Earthquakes'. The first workshop is
along with discussion and conclusions. tentatively planned to be held in Tsukuba
Science City, Japan, in May 1983, prior
Keywords : Buildings, Building screening, to the 15th Joint US-Japan Panel Meeting
Damage, Damaged buildings, Evaluation, Existing on Wind and Seismic Effects. Specific
buildings, Full scale tests, Post-earthquake
details and agenda of the workshops will be
evaluation, Rapid evaluation, Safety evaluation,
established by each side and coordinated by
Seismic performance
435
correspondence. A final report is expected (for evaluation of timber structures).
5.
to be published in 1985 and reported to the Special approaches still under active
17th Joint Panel Meeting in Japan." development, including rule-based
inference procedures for formal damage
With the support of the National Science assessment, and the use of expert
Foundation through Grant CEE-8217190 to Cornell systems.
University), two workshops have been held
(May 1983 in Japan and July 1984 in Berkeley, One of the difficulties in placing a given paper
California). The proceedings of these workshops from the workshops into a category is that the
have been published in two volumes [1] and [2]. multi-level procedures in use overlap into at
2.
A third workshop devoted primarily to critical least two different categories. Hence we will
comparisons of evaluation results and to plan- begin this section with a general discussion of
ning of the final year's activities was held on methodologies, and then move into review of the
May 17, 1985, in Tokyo. individual contributions for the workshops.
EVALUATION METHODOLOGIES It is convenient to divide the several levels of

approximate analysis into two groups: those
There are many evaluation methods currently in that require little analysis and are used mainly
use for assessing the seismic performance of to classify buildings, and those that require
buildings. They range from very approximate load-deflection and limited modal analysis.
screening methods for identifying potential These approaches are reviewed briefly here.
problem structures in large groups of buildings,
to detailed analytical approaches for studying .2.1 Rapid Evaluation Methods
performance of structural members, components,
and systems on an individual structure basis. The purpose of rapid classification is to be
Some of these detailed methods involve dynamic able to judge the safety of typical buildings.
analysis and others are done on equivalent These methods are useful In surveying large
static idealizations. groups of buildings at the municipal or regional
level. The calculations typically require about
Evaluation methodologies discussed at the work- two to four hours per building, and often much
shops may be classified into several categories: less
1. Screening methods for quick evaluation

of large groups of buildings. The key aspect of these approaches is that
2. Rapid evaluation procedures for various allowable stress levels are assigned to
individual buildings. supporting elements (columns or walls) depending
3. Multi-level procedures for progres- on their composition, length, and ductility.
sively more accurate evaluation of These stress levels were established from tests
individual structures, including the and from studies of actual damage. The great
1977 Japanese Standard for Seismic advantage of these rapid classlf lcatloo
Capacity Evaluation of Existing Rein- is that they can easily be modified to account
forced Concrete Buildings, and the for new knowledge gained about the behavior >f
U.S. Navy Procedure. systems. The interaction of various typer of
4. Detailed analytical approaches, usually structural elements can he accounted for. Pour
involving dynamic analysis, and behavior of certain materials or elements, such

computer-based methods such as WOODAM as the problems with short (captive) columns
*06
are readily included in the evaluation. increased damping to account for the energy
absorption due to hysteretic response. The
These simple classification approaches are not changes in period and damping are functions of
usable or at least are not reliable for unusual estimated damage indices which represent a
structures. For example, highly unsymmetrical measure of the inelastic deformation level of
structures or poor soil conditions are so the structure.
variable that only much more rigorous analysis

can assess safety in these cases. The rapid This method is relatively simple because only an
evaluation methods are normally quite conserva- elastic model analysis is required. However, it
tive, and higher level approaches are necessary has been calibrated only for concrete frames and
for screening structures that fail the rapid more work is needed to extend it to other types
test. of structures. In essence, the use of inelastic
response spectra is replaced by elastic spectra
2 . 2 Rapid Analysis Methods but with changed periods and damping. In the
use of inelastic spectra the period is not
The common characteristic of these approaches is changed. Of course, this approach can be used
that the nonlinear load-deflection relationship only on the high-period (descending) side of the
for the structure, or for each story, is re- spectrum.
quired as shown in Fig. 1. Structures behave
inelastically when subjected to a major earth- b. The capacity spectrum method
quake and it is necessary to know the deflec- In this method the inelastic load-deflection
tions beyond the elastic limit. This is curve is first estimated; often rough calcula-
especially important for buildings composed of tions will suffice. This curve is best drawn as
several types of lateral resisting elements the roof displacement versus the base shear for
(frames, stairways, walls) because there is a lateral load representing the inertia force
then a second line of defense. vector (Fig. 1). For low buildings a straight
line load vector is acceptable. A static linear
Some engineers question whether we can establish computer program can be used as follows: the
the load-displacement curve for complex struc- load is increased until a few hinges form and
tures. In most cases only a rough estimate of then the rotational stiffnesses are greatly
the elastic limit and the stiffness is neces- reduced at these locations. The load is
sary in these evaluation methods. The second increased until another group of hinges form
point is that since the structure will deform and in this manner a piece-wise linear load-
inelastically, with various post-elastic stiff- displacement relation is produced. In some
nesses, a realistic estimate of the safety must cases a rough estimate of the initial stiffness,
rely on an approximate knowledge of the load- the elastic limit, and the subsequent stiffness
displacement relationships. can be done without a computer.
Three different rapid analysis methods are In the next step the roof displacement is
summarized here: related to the corresponding spectral displace-
ment, S, , and the base shear is related to the
cl
a. The substitute structure method spectral acceleration S through the well-known

In this approach (developed by Sozen and his modal equations. These transformations are
associates) the inelastic response is estimated ordinarily based on the first mode only. The
by calculating an increased period and an period corresponding to and values are
calculated 3 . SUMMARIES OF PAPERS IN W ORKS HOP VOLUM ES
The plot of S versus the period T is called 3.1 Screening Methods

a
the capacity spectrum (after Freeman). It is
next superimposed on the design spectrum curve Methodology for screening a group of buildings
(Fig. 2). The relative position of the capacity is given by Agbabian in Paper II-4, Vol. 1. His
and response spectra determines the safety of approach includes the initial preparation of a
the structure. An interpolation is necessary detailed seismic safety plan to specify the
because the damping changes as inelasticity evaluation criteria, methods of analysts to be
develops used, basis for damage estimates, practical

methods for upgrading the earthquake resistance
The capacity spectrum method is a relatively of the structures being studied, and a manage-
simple and attractive method when the first mode ment plan for implementation of the seismic
dominates, although the contribution of higher safety requirements. Important steps in the
modes to the roof deflection can be incorporated evaluation process include: (1) study of
in an approximate manner. It is usually more as-built drawings plus visual inspection to
satisfactory to estimate the bilinear stiffness categorize the buildings, (2) determination of
of the structure. the level of refinement to be used in analyzing
the buildings, (3) establishment of consistent
c. The reserve energy technique criteria for degrees of damage, (4) estimates of
This approach was proposed by John Blume in the damage to each building for different levels of
early 1960's but it has not been used widely earthquake intensity, and (5) definition of
(perhaps because its title is not descriptive). acceptable level of risk based on the evaluation
It is also based on the nonlinear load-displace- of the expected damage. The end product Is
ment curve but instead of developing a capacity information on percentage of damage and asso-
spectrum, the input kinetic energy is calculated ciated costs for each building for differing
using the velocity spectrum and modal masses levels of maximum ground acceleration. An
2
0.5raS^. The area under the load-displacement example illustrating the cost of repair and
curve represents the absorbed strain energy and replacement due to earthquakes of different
by equating the two energies, one can estimate severities, for six buildings, Is given In
the deflection and the corresponding damage. Table 1.
In both the capacity spectrum method and the Three screening procedures are given by Lew In
reserve energy technique, the static load- Paper II-3, Vol. 1. These methods, developed by
displacement curve is estimated. Cyclic the National Bureau of Starfdards, the General
behavior is not accounted for directly, although Services Administration, and the State of
it may be possible to modify the curve to do California Seismic Safety Commission, can he
that. As with all approximate methods, it is used for a large number of buildings to Identify
important for the engineer to have a good feel hazardous conditions and to supply decision
for the real behavior of structures. makers with Information on the extent of hazard
and the feasibility of retrofit measures.
The three methods are based on structural

performance indices which are derived fr.»m past
performance records of particular types of
438
buildings, from engineering judgement, and on tive for ductile frames. The second level
seismicity considerations for particular sites. procedure concentrates on potential problems in

The structural performance indices are related vertical members by identifying five types of
to structural characteristics, structural vertical elements: extremely brittle short
configurations, and the degree of deterioration columns, shear-critical columns, shear walls,
of buildings, whereas the seismicity index is flexure-critical columns, and flexural walls.
obtained by considering the distance from active Simple equations are given for treating each
faults and expected magnitude and frequency of type of element.
earthquakes
The third level procedure recognizes flexural
3 . 2 Multi-Level Procedures hysteresis by considering the response of three
additional types of vertical members: columns
a. Japanese Standard, 1977 governed by shear beams, columns governed by
The Japanese "Standard for Seismic Capacity flexural beams, and rotating walls with uplift
Evaluation of Existing Reinforced Concrete effects from overturning. Equations are again
Buildings" was compiled in 1977 under the provided for these calculations.
Ministry of Construction. It is described in
some detail in Paper 1-2, Vol. 1, by Aoyama. Results of the application of these procedures
It is intended to be applied primarily to low to damaged structures are given in Fig. 3, where
rise buildings up to 6 or 7 stories. the correlation between first and second level
procedures applied to buildings from 1978

The three-level procedure involves calculation earthquakes in Japan is given in Fig. 3a. A
of a Seismic Index comparison of damage levels and seismic index I
s
I = E GS„T values calculated for both the longitudinal (L)
s o D
and transverse (T) motions of a group of school
where E = 4>(CxF) is the basic seismic index,
o
buildings is shown in Fig. 3b.
Within the term E^ , C is a strength matrix, F
is a ductility matrix, and <j> is a story index
Additional examples of damage level and index
that relates single DOF system response to that
values are given in Paper II-2, Vol. 1, and in
of the multi-story system. Thus it is apparent
some of the benchmark structure analyses to be
that this method evaluates both strength and
described later in this paper.
ductility capacities and expresses the result
E^ as a product of these capacities.
While seismic capacity is best judged by compar-
ing I to an E^_ index that is tied to specific
The other necessary definitions are G = geo-
ground acceleration levels and other critical
logical index, S = structural design index
Q factors, general guidelines on performance may
ranging from 0.4 to 1.2, and T = time index
be stated. Superior performance is indicated by
ranging from 0.4 to 1.0 to cover such items
I values of 0.9, 0.7, and 0.6 for levels 1, 2,
as previous damage, shrinkage and settlement
and 3 evaluation procedures, respectively. I
effects, and deterioration of the structure
values less than 0.4 at the 2nd and 3rd levels
from loadings and environmental factors.
indicate questionable performance, and the
borderline of damage for level 1 is often taken
The first level procedure is very easy to apply
as 0.8 (Paper II-2, Vol. 1).
— it is based on the average stress in the
column and wall areas. It is well-suited for
b. U.S. Navy procedure
wall-dominated structures but is quite conserva-
A methodology is presented by Freeman (Paper
439
II— 1 ,
Vol. 1) for the preliminary evaluation of has been used a number of times for specific
the safety of existing buildings for earthquake buildings in the U.S. The paper gives a case
effects. The procedure was developed for the study on a building constructed in 1900 and
U.S. Navy to evaluate its existing structures upgraded recently to meet 1979 UBC Seismic
throughout the naval shore establishment that Zone 3 forces.
are located in areas of high seismicity. The
scope of the methodology includes a screening Tabulated cost data related to degree of compli-
process for a large number of buildings, a rapid ance with the Uniform Building Code have proven
procedure for approximating building capacity, to be invaluable to program administrators and
and a graphical procedure for estimating damage building officials who must make decisions on
due to postulated earthquake motion and the strengthening or demolishing buildings. This
costs of their corrections. The fundamental data not only helps in economic and engineering
ideas in this approach have been described above planning, but it is also a major factor in
in the discussion of the capacity spectrum determining levels of acceptable risk.
method.
b. Computer-based approaches
Freeman states that it cannot be expected that The computer program W00DAM, developed for the
the rapid evaluation method will give an accu- U.S. Navy to estimate damage to wood frame
rate prediction of building response, but it buildings, is described by Hart in Paper II 1—2 ,
is highly useful in giving a general picture Vol. 1. The method provides determination of
of behavior and it is also useful as an aid in expected damage for a given structural reli-
determining which buildings will require a full- ability over an entire range of earthquake
scale dynamic analysis. levels, and it is implemented on a microcomputer
(such as the HP 83). Thus the evaluation pro-
3. 3 Other Methodologies cedure can be done on-site.
a. General The program modules are used to calculate the

The United States Department of Housing and earthquake loading on critical components of the
Urban Development (HUD) published a three-volume lateral force resisting system, taking into
manual in November 1978 titled "A Methodology account uncertainties in the seismic loading and
for Seismic Evaluation of Existing Multistory in the analytical model for the building. The
Residential Buildings" to be used to satisfy the next step is to calculate the load carrying
HUD requirement that all multistory residential capacity of the lateral load system of the
buildings in Seismic Zones 3 and 4 (Uniform structure, including uncertainties in material
Building Code) shall be evaluated for earthquake properties, geometry, and the capacity model.
hazard and seismic resistance. Since that time, From these two steps, the reliability index for
HUD has published a series of additional guide- each possible failure mode for each critical
lines on building rehabilitation. This HUD element in the lateral force resistant system is
methodology for evaluation and structural determined
assessment of building rehabilitation is sum-
marized by Fuller in Paper II-5, Vol. 1. A damage matrix is next constructed the!
the damage to the magnitude of seismic load for
The methodology has proven to be a valuable tool a determined value of reliability index. The
in analyzing existing buildings that have been final step is to calculate the expected damage
proposed for renovation in HUD programs, and it on a probabilistic basis for given levels >1
w
seismic load. The latter are expressed in terras status of this promising area for the future
of maximum response spectra for different time is captured in the following quote:
periods up to 50 years. "The fundamental question regarding
the use of inference machine evalua-
Meyer and Roufaiel (Paper 1-5, Vol. 2) describe tion methods is whether such methods
an analytical method for determining the reli- can be relied upon without a careful
ability of reinforced concrete frames that may review by a human expert. It must be
or may not have been damaged by earlier earth- recognized that subjective knowledge
quakes. A model is described for simulation of and experience of an expert imply a
reinforced concrete frame elements to large higher level of intellectural organi-

cyclic loads, and ground motion input is zation that is possible in data bases
simulated as a non-stationary random process. containing large sets of information
Damage probability matrices are constructed . . . Nevertheless, the use of
using Monte Carlo simulation, accounting for the appropriate evaluation techniques
variability of the ground motion and the build- utilizing inference machine
ing strength parameters. The methodology is approaches can greatly enhance the
applied to a three-story, two-bay reinforced effectiveness and reliability of
concrete frame building. The results indicate human expert evaluations. Further-
that it is unusual for even strong ground more, it may also contribute to
motions to appreciably extend damage sustained improvements in human expertise
under earlier seismic loadings. through increased communication among
experienced engineers."
The wealth of computer programs available for
analyzing structures subjected to seismic loads 3 . 4 Miscellaneous
is illustrated by the compilation of Japanese
programs by Aoyama in Paper III-l, Vol. 1, where a. Structures to be tested by actual
earthquakes
nearly 70 programs are described by name,
Okada and Tamura report in Paper 1-6, Vol. 2 on
company or organization responsible for the
the behavior of model structures that have been
program, and critical features of the program.
constructed deliberately weak (either weak beams
or weak columns) at the Chiba Experimental Field
c. Rule-based inference procedures
Station. The experiment has yielded more than
The rule-based inference procedure for struc-
20 sets of acceleration and displacement records
tural damage assessment called SPERIL-I is
reviewed by Yao in Paper 1-3, Vol.
up to June 1984. The information from these 5
1. This
story model structures should be of substantial
approach is a highly structured formal damage
benefit in calibrating analysis methods for
assessment method utilizing structural reli-
imperfect structures and to study the influence
ability, the theory of fuzzy sets, and expert
systems to arrive at consistent results to of repeated seismic loadings on the response of
the structures.
imperfectly formulated problems. Fig. 4
indicates the numerous factors to be considered

and the complex linking between these factors.
b. Post-earthquake evaluation
Several papers were presented in Session V of
the First Workshop on this important topic.
Yao continues his treatment of expert systems in
safety evaluation in Paper I— 1 Johnston (Paper V-l) reported on the post-
,
Vol. 2, where he
summarizes recent results since May 1983. earthquake evaluation of a large hospital in
The
California. He summarized the major factors
involved in this type of engineering work: 3 . 5 Applications to Benchmark _St rue tur e
"(1) A careful assessment of the physical
damage caused by the earthquake. This Six benchmark structures were selected for
assessment should be made by a structural evaluation by the various methods; they are
engineer, (2) A knowledge of the historical described in Appendix A of Ref. (2). The
background of earthquake engineering code buildings are:
requirements and construction practice, 1. 7 story reinforced concrete frame
(3) An evaluation of the ultimate behavior tested under the U.S. -Japan Cooperative
of the structural system and materials, Research Program
particularly its toughness, (4) A recogni- 2. Imperial County Services Building,
tion of what constitutes acceptable risk, California
(5) Confirmation of the assessment by a 3. 9 story ATC benchmark structure,
peer review, (6) Consultation with medical reinforced concrete, 1964
experts to establish the operational 4. 7 story reinforced concrete building

feasibility and the ability to remain (Holiday Inn), instrumented
functional under repair, and (7) An honest 5. Reinforced concrete building, Japan
and objective decision based on the 6. 6 story steel frame tested under
engineering data." U.S. -Japan Cooperative Research Program
Hirosawa and Murota (Paper V-2) provided a Inasmuch as the evaluation procedures are still
summary of many projects being conducted as part being applied to these structures, and critical
of the Japan National Project on Development of comparisons of all results are not yet complete,
Post-Earthquake Measures for Buildings and we will provide only selected results in this
Structures Damaged by Earthquake. Each summary paper. Final conclusions will be given in a
includes the objectives of the study, progress later report on this project.
made to date, and the product of the research
task. a. 7 story U.S. -Japan reinforced concrete
building
This combination frame-shear wall structure had
Paper V-3 by Gergely was the closing paper in
three bays of 19.7, 16.4, and 19.7 ft (6, 5, and
the First Workshop; it dealt with the many
6 m) in the longitudinal direction and two spans
planning and coordinating efforts needed in
of 19.7 ft each (6 m) in the transverse direc-
emergency safety evaluation and rating of
tion. A shear wall was placed parallel to the
structures after a major earthquake in a large
direction of loading in the middle bay of the
urban environment. Issues needing the combined
center frame, continuous from the first to the
attention of city, state, and federal government
seventh story.
agencies, as well as police and other enforce-
ment agencies, are discussed.
Several papers (Yoshimura, IV-2 and Kabeyasawa,
IV-3, both Vol. 1), provide detailed experi-

c. Analysis methods and damage levels
mental and analytical results for this
There were a number of papers that reported on
structure. This wealth of data is extremely
analysis methods for structures (in particular
valuable for comparison with both approximate
dynamic methods), and on damage analysis of
and detailed analytical techniques.
other structures. Included are Papers 1-2,
1-4, III-l and III-2 in Vol. 2 of the workshop
Okamoto (Paper IV- l, Vol. 1) evaluated the seven
proceedings
story reinforced concrete building using the
^2
Japanese Standard described earlier. The using a structural identification technique.
evaluation was made only for the direction in The study showed that these relations are highly
which the seismic forces were applied during the useful in studying the structural response both
loading in the laboratory, and the weight of qualitatively and quantitatively; it is possible
each story was taken as the dead weight of the to extract such information as the range and
floor level plus the weight of the loading number of cycles of inelastic response for use
frames. The Ground Motion Index G, the Shape in damage expressions. The planned extension of
Index S, and the Time Deterioration Index T this research includes identification of damage
were all taken as unity. expressions to more effectively use the quanti-
tative information from the force-deformation
Results are quoted directly from Okaraoto: relations, calculation of energy dissipation and
"The Basic Structure Performance Index, E ,

its correlation to structural damage, comparison
o
was computed as 0.36 (1st Screening), 0.68 of results with other investigators, and possi-
(2nd Screening), and 0.59 (3rd Screening). ble extension of the methodology to include
The index given by the first screening is found torsional effects and overturning.
to be significantly lower than those given by
the second and third screenings. This low Aoyama (Paper II-8, Vol. 2) applied the Japanese
value of index was obtained primarily because Standard procedures (levels 1 and 2) to the
of small areas of the columns and shearwalls. structure, using = 0.95, G = 1.0, and T =
In the second and third screenings, on the 1.0, and obtained the following results:
other hand, E was much larger because the
o E-W I N-S I
Story s s
structure was estimated to exhibit flexural
6 0.65 1.06
failure. Nevertheless, if the weight is
5 0.31 0.52
2
increased to 1 ton/m by taking live load (First Level 4 0.23 0.37
Procedure) 3 0.18 0.31
into account, E^ will decrease significantly
2 0.17 0.29
to a level at which the structure is considered 1 0.11 0.48
to be unsafe.” (Second Level 1 0.22 0.48
Procedure)
b. Imperial County Services Building In comparison with accepted Japanese standards,
The Imperial County Services Building is well- it was found that the Imperial County Services
known for its unsatisfactory performance in Building fell in a region regarded to be
the 1979 Imperial Valley earthquake, and has dangerous under earthquake loads.
received considerable attention in various
post-mortem analyses. It was designed in 1968 By the Japanese standard, the upper level walls
in accordance with the 1967 UBC and construction would be regarded as "miscellaneous walls" with
was completed in 1971. only 1/3 the effectiveness of regular shear
walls. The average axial stress level in the
The poor performance of this building under vertical elements was about twice that of
seismic loads has been attributed partially to typical Japanese building designs. While there
the poor shear transfer properties at one end is a question on the validity of using the
of the building. Japanese approach on the N-S framing, the E-W

direction framing does fit within assumptions.
Force-deformation relations for this building
were obtained by Stephens and Yao (Paper 1-3, The analysis by Hart (Paper 11-10, Vol. 2)
Vol. 2) from structural acceleration records showed that the normal assumption of fixed base
shear walls was not valid and that the soil- framing is concrete pan-joist on 20 by 24 ft
building system m^st be modeled. He also showed (6.1 by 7.3 m) bays framing into girders and
that the floor diaphragms were flexible, not columns, with shear walls along both axes of the
rigid, and that using a rigid diaphragm analysis building to resist lateral loads. The plan
is misleading in terms of the distribution of dimensions of the main portion of the building
loads at the ground floor column level. are 141.5 by 80.67 ft (43.1 in by 24.6 m)
Sozen (Paper II-9, Vol. 2) summarizes the key Lew (Paper II-7, Vol. 2) applied the rapid,
findings from the PhD dissertation of M. Kreger Level 1 Field Evaluation Method of the MBS
(U. Illinois, August 1983) who analyzed the approach, and concluded that for an earthquake
Imperial County Services Building with par- intensity of 9 MMI ,
the building rated fair .
ticular emphasis on the causes of the column The unsymmetrical distribution of shear walls in
failures. It is shown that it is difficult to the transverse direction controlled the rating
explain the failures by considering forces in of the structure.
only one direction at a time. Sozen provides

some very interesting philosophical discussion Okamoto and Watanabe (Paper II-2, Vol. 2)
at the end of his paper: applied the Japanese Standard 1st level pro-
"Evaluation of existing buildings is better cedures. In the longitudinal direction, I
g
kept simple. A simple analysis tempered by values ranged from 1.37 at the top floor to
judgement will suffice in most cases. In 0.34 at the base; the fifth floor had I - 0.37.
the interest of the community, it is a In the transverse direction, 1^

ranged from 2.30
better buy to invest less in reanalysis to 0.67, with 0.68 at the 5th floor. These I
s
at the risk of missing a few potential values are low by Japanese standards.
disasters than to invest more with the
possibility of identifying some of those Freeman (Paper II-3, Vol. 2) used the capacity
few. Vulnerable buildings are in general spectrum method for the ATC building, in a very
like bad wine. Even a naive analysis can quick one page calculation. He estimated the
identify the vinegary taste. It is very static capacity for the uncoupled walls and the
unlikely that spending more professional spandrel frames and plotted the capacity
resources on analysis routines will spectrum curve as two straight lines. The
automatically identify the buildings with intersection with a design spectrum is at about
subtle vulnerabilities. The Imperial 0. 19g.
County Services Building is a good example.

It is very unlikely that calculation alone d. 7 Story Holiday Inn
would have identified the problem and it is The Orion Holiday Inn building in Los Angeles Is
very likely that a structural engineer who a reinforced concrete flat plate structure with
was informed of similar events would have 8 bays (18.75 ft or 5.71 m) in the F.-W direction
avoided the conditions at the east end of and 3 bays (about 20 ft or 6.1 m) In the N-S
the structure." direction, with spandrel beams in the periphery.
The building is instrumented and measured re-
c. 9 story reinforced concrete ATC benchmark sponse during the 1972 San Fernando earthquake
building
is available.
This structure was designed with the 1964
UBC provisions, and has been the subject of
This structure was studied by Freeman, Shlmlru
re-design in the ATC project. The building
(Paper 1 1— 1 ,
Vol. 2) and Moehle (Paper I
m
Vol. 2). Shimizu applied the Japanese Standard Hart (Paper II-6 , Vol. 2) did a detailed
and obtained E values between 0.20 and 0.36 by structural analysis of the system, based on a
o
rapid evaluation, and 0.24 to 0.45 by detailed perspective which recognizes the use of site
evaluation. By Japanese standards, E should be dependent earthquake response spectra. The
q
at least 0.6. analysis required about 100 engineering labor
hours. It provides a rather detailed prediction
Moehle made a relatively simple analysis plus of the development of plastic hinges in the
an assessment of connections and details. He structure, the lateral displacement (nonlinear),

concluded that (a) the structure is fairly and the variation of period with base shear.
flexible and excessive nonstructural damage Typical U.S. steel frames have a much flatter
could be expected in a moderate earthquake, curve for change in period with change in base
(b) inadequate ductility may be a problem shear.

during a more intense earthquake, (c) analysis
underpredicts initial stiffness by nearly two Paper II-5, Vol. 2 (by Yamanouchi et al) pro-
to one, and (d) the first floor column details vides detailed test results for this structure.
are considered inadequate.
4. SUMMARY AND CONCLUSIONS

The Capacity Spectrum method results in spectral
accelerations in the 0.15g - 0.21g range, Methods of seismic safety evaluation of existing
depending on the contribution of nonstructural buildings were studied in three U.S. -Japan
elements, in the transverse direction. The workshops. The methods range from rapid classi-
range is 0.20g to 0.26g in the longitudinal fication of groups of buildings to detailed
direction. dynamic analysis of individual structures.
e. 6 story steel frame (U.S. - Japan Program) Methods developed in Japan and in the U.S. are
The six story steel frame building tested in "tuned" to the characteristics of typical build-
Tsukuba, Japan consisted of three two-bay frames ings in each country. Comparative analyses of
of 24.6 ft (7.5 m) spans with a first story benchmark structures reflect the relatively more
height of 14.8 ft (4.5 m) and upper story conservative Japanese seismic design approaches.
heights of 11.1 ft (3.4 m) . The two exterior
frames were rigid and the interior frame had The comparisons reveal that the various methods
K-bracing in one bay. The structure was de- are suitable for the expected level of evalua-
signed to satisfy the requirements of both the tion accuracy for buildings with good arrange-
1976 UBC and the 1980 Japanese Aseismic Design ment of lateral strength elements, and that a
Code, with a design base shear coefficient of general idea of expected damage level can be
0.197. estimated. However, additional developments are
required to validate and improve the methods to
Lew (Paper II-7, Vol. 2) rated this building make them more useful, especially for buildings
using the NBS rapid FEM, and concluded that the of unusual geometry or with special secondary
structure rated fair for a MMI 9 earthquake. elements, and to include effects of soil-
The semi-rigid nature of the steel deck with structure interaction.
concrete topping was a critical factor in
reaching this rating.
445
5. REFERENCES
( 1 ) Proceedings of the First Workshop on

Seismic Performance of Existing Buildings,
May 1983 , edited by Richard N. White and
Peter Gergely, Department of Structural
Engineering, Cornell University, Ithaca,
N.Y. 14853, April 1985, 370 pp.
(2) Proceedings of the Second Workshop on

Seismic Performance of Existing Buildings ,
July 1984, edited by Richard N. White and

Peter Gergely, Department of Structural
Engineering, Cornell University, Ithaca,
N.Y. 14853, April 1985, 516 pp.
6. ACKNOWLEDGEMENTS
The financial support of the National Science

Foundation and the personal support and
encouragement provided to the series of
workshops by Dr. John B. Scalzi, NSF Program
Manager, are greatly appreciated. The efforts

of Dr. M. Watabe, Dr. S. Okamoto, and G. R.
Fuller in helping to organize the workshops are
also much appreciated, as is the enthusiastic
involvement of the many individuals from Japan
and the United States who participated in the
workshops
The opinions, findings, conclusions, and

recommendations expressed in the paper are those
of the individual contributors and do not
necessarily reflect the views of the NSF and
other private or governmental organizations.
446
At Cost (in $ Millions) of
Replacement Percent Damage •
Repair or Replacement At
Building Cost
($ Millions)
0.20 g 0.30 g 0.40 g 0.20 g 0.30 g 0.40 g
A 1.000 35 100 100 0.350 1.000 1.000
3 5.000 23 100 100 1.150 5.000 5.000
C 0.500 5 23 85 0.025 0.140 0.425
D 1.000 10 26 58 0.100 0.260 0.580
E 2.000 2 18 48 0.040 0.360 0.960
F 3.000 0 0 15 0 0 0.450
Total 12.500 1.665 6.760 3.415
TABLE 1. EXAMPLE ILLUSTRATING COST OF REPAIR AND

REPLACEMENT DUE TO EARTHQUAKES OF
DIFFERENT SEVERITIES
w
Force (base shear)
Fig. 1 Roof displacement vs. base shear for building
Fig. 2 Capacity spectrum method plot
m
ProceAjre
Level
2nd
I
Index
a. Seismic index and damage in b. Seismic index and damage of

1978 earthquakes in Japan school buildings in 1968
Tokachi-Oki EQ (2nd level
procedure)
Fig. 3 Seismic indices and damage, Japanese Standard (1977)

APPENDIX A - FIRST WORKSHOP, MAY 1983
List of Participants:
From Japan:
H. Aoyama University of Tokyo
M. Hirosawa Building Research Institute
M. Izumi Tohoku University
T. Kabeyazawa University of Yokohama
K. Karaimura Building Research Institute
T. Murota Building Research Institute
T. Okada University of Tokyo
S. Okamoto Building Research Institute
M. Watabe Building Research Institute
M. Yoshiraura Building Research Institute
Observers
M. Ishizuka University of Tokyo
T. Kaminosono Building Research Institute
M. Teshigawara Building Research Institute
From the United States:
M. Agbabian Agbabian Associates
D. Foutch University of Illinois
S. Freeman Wiss Janney Elstner Associates
G.R. Fuller Department of Housing and Urban Developmei
P. Gergely Cornell University
P. Gurvin Department of State
G. Hart University of California, Los Angeles
R. Johnston Brandow and Johnston Associates
H.S. Lew National Bureau of Standards
J. Scalzi National Science Foundation
R.N. White Cornell University
J. Yao Purdue University
<60
Papers Presented:
I OVERVIEW OF EVALUATION OF BUILDINGS FOR SEISMIC RESISTANCE
1. EVALUATION OF BUILDINGS FOR SEISMIC RESISTANCE - AN OVERVIEW OF 1
PROBLEMS AND ISSUES

R. N. White
2. A METHOD FOR THE EVALUATION OF THE SEISMIC CAPACITY OF EXISTING 17

REINFORCED CONCRETE BUILDINGS IN JAPAN
H. Aoyama
3. SAFETY EVALUATION OF EXISTING BUILDINGS 43

J.T.P. Yao
II PRACTICAL METHODOLOGY
1. A METHODOLOGY TO EVALUATE THE VULNERABILITY OF BUILDINGS TO 61

EARTHQUAKE DAMAGE
S. A. Freeman
2. PRACTICAL APPLICATION OF THE EVALUATION STANDARD IN SPECIFIC 85

EARTHQUAKE HAZARDOUS LOCALITY
T. Okada
3. SEISMIC HAZARD EVALUATION METHODS OF EXISTING BUILDINGS 105

H. S. Lew
4. EVALUATION AND SCREENING OF LARGE GROUPS OF BUILDINGS 135

M. S. Agbabian
5. STRUCTURAL EVALUATION OF MULTISTORY RESIDENTIAL BUILDINGS 145

FOR SEISMIC RESISTANCE
G. R. Fuller
III COMPUTER PROGRAMMING
1. SUMMARY OF VARIOUS COMPUTER PROGRAMS TO EVALUATE STRUCTURAL 159

PERFORMANCE AGAINST EARTHQUAKE EXCITATIONS
H. Aoyama
2. SEISMIC EVALUATION OF EXISTING BUILDINGS INCORPORATING 165

UNCERTAINTY AND UTILIZING MICROCOMPUTERS
G. C. Hart
3. COMPONENT MODE METHOD IN EIGENVALUE ANALYSES AND ITS 181

APPLICATIONS TO BUILDING STRUCTURES OF NUMEROUS
DEGREE-OF-FREEDOM
M. Izumi, H. Katukura, and M. Takita
451
IV EVALUATION OF STRUCTURAL PERFORMANCE THROUGH FULL SCALE TESTS
1. COMPARATIVE STUDY ON THE APPLICATION OF THE EVALUATION 199

STANDARD TO 7 STORY FULL-SCALE REINFORCED CONCRETE
TEST STRUCTURE
S. Okamoto
2. ANALYSIS ON THE BEHAVIOR OF 7 STORY FULL-SCALE REINFORCED 215

CONCRETE TEST STRUCTURE
M. Yoshimura
3. COMPUTER ANALYSIS OF THE 7 STORY FULL-SCALE REINFORCED 255

CONCRETE TEST STRUCTURE
T. Kabeyasawa, H. Shiohara, S. Otani, and H. Aoyama
V POST-EARTHQUAKE EVALUATIONS
1. POST-EARTHQUAKE EVALUATION OF AN ESSENTIAL FACILITY 333

R. G. Johnston
2. METHODOLOGIES FOR IMMEDIATE JUDGEMENT ON THE SAFETY OF 345

DAMAGED STRUCTURES
M. Hirosawa and T. Murota
3. POST-EARTHQUAKE EVALUATION METHODOLOGIES OF SAFETY OF 361

STRUCTURES
P. Gergely
162
APPENDIX B - SECOND WORKSHOP, JULY 1984
List of Participants:
From Japan :
H. Aoyama University of Tokyo

T. Fukuta Building Research Institute
T. Kabeyazawa University of Yokohama
T. Okada University of Tokyo
S. Okamoto Building Research Institute
S. Otani University of Tokyo
H. Umemura Shibaura Technical University
M. Watabe Tokyo Metropolitan University
H. Yamanouchi Building Research Institute
Observers
T. Karainosono Building Research Institute
S. Nakata Building Research Institute
M. Yoshimura Building Research Institute, on leave
From the United States :
M. Agbabian Agbabian Associates

G. Brandow Brandow and Johnston Associates
S. Freeman Wiss Janney Elstner Associates
G.R. Fuller Department of Housing and Urban Development
P. Gergely Cornell University
G. Hart University of California, Los Angeles
H. S. Lew National Bureau of Standards
J . Moehle University of California, Berkeley
C. Meyer Columbia University
J. Scalzi National Science Foundation
M. Sozen University of Illinois
R.N. White Cornell University
J. Yao Purdue University
Observers
A.J. Eggenberger National Science Foundation
J. Noland Atkinson Noland and Associates
R. Nutt Applied Technology Council
M. Wagner H. J. Degenkolb & Associates
S. Wood University of Illinois
<65
Papers Presented
I SAFETY EVALUATION METHODS
1. EXPERT SYSTEMS FOR SAFETY EVALUATION OF EXISTING STRUCTURES 1
J.T.P. Yao
2. ULTIMATE-STATE DESIGN ANALYSIS OF R/C WALL-FRAME STRUCTURES 7

T. Kabeyasawa
3. SAFETY EVALUATION OF SEVERAL DAMAGED BUILDINGS — A PROGRESS 31

REPORT
J. Stephens and J.T.P. Yao
4. EVALUATION OF DYNAMIC CHARACTERISTICS OF A LONG SPAN STEEL 73

TRUSS STRUCTURE EXCITED BY VERTICAL GROUND MOTIONS
M. Watabe, H. Yamanouchi, T. Minami, Y. Maekawa,
H. Miyanaga, T. Morimoto, and K. Mizutori
To be introduced by S. Okamoto
5. RELIABILITY OF DAMAGED REINFORCED CONCRETE FRAMES 103

C. Meyer and M.S.L. Roufaiel
6. OBSERVATION OF EARTHQUAKE RESPONSE OF REINFORCED CONCRETE 133

WEAK-MODEL STRUCTURES
T. Okada and R. Taraura
II EVALUATION METHODS APPLIED TO BENCHMARK STRUCTURES
1. RAPID AND DETAILED EVALUATION ON A MEDIUM-TALL R/C BUILDING 131

K. Shimizu
2. A QUICK ANALYSIS OF THE ATC BENCHMARK R/C NINE STORY BUILDING 191
S. Okamoto and K. Watanabe
3. RAPID EVALUATION METHODOLOGIES

S. Freeman 207
4. SEISMIC HAZARD EVALUATION OF HOLIDAY INN 235

J. P. Moehle
5. INELASTIC SEISMIC TESTS ON A FULL-SCALE SIX-STORY CONCENTRIC-K 247

BRACED STEEL BUILDING — U.S.-JAPAN COOPERATIVE RESEARCH
PROGRAM
H. Yamanouchi, M. Midorikawa, I. Nishiyama, and T. Fukuta
6. DETAILED SEISMIC EVALUATION OF A FULL-SCALE SIX STORY STEEL 267

BUILDING
(U.S. -Japan Cooperative Program)
G. Hart
7. SEISMIC PERFORMANCE EVALUATION OF CONCRETE AND STEEL STRUCTURES 295

H. S. Lew
8. A QUICK ANALYSIS OF THE IMPERIAL COUNTY SERVICES BUILDING FOR 335
SEISMIC RESISTANCE
H. Aoyama
10.
9. AN EXAMPLE AND AN OPINION ON SCOPE OF STRENGTH EVALUATION FOR 357
BUILDING STRUCTURES
M. Sozen
EVALUATION OF THE IMPERIAL COUNTY SERVICES BUILDING 373

G. C. Hart
III DAMAGE AND ANALYSIS OF OTHER STRUCTURES
1. DAMAGE AND ANALYSIS OF NAMIOKA TOWN HOSPITAL DAMAGED BY THE 381

NIHONKAI-CHUBA EARTHQUAKE, 1983
S. Okaraoto
2. KURAYOSHI CITY EAST BUILDING DAMAGED BY THE 1983 TOTTORI EQ 407

S. Otani and K-N. Li
APPENDIX A - DESCRIPTION OF BENCHMARK STRUCTURES A1
455
Theme V
REPORTS ON THE
NIHONKAI-CHUBU EARTHQUAKE
AND THE
NAGANOKEN-SEIBU EARTHQUAKE
A STUDY OF THE ECONOMICAL LOSSES CAUSED BY A CHANGE OF CARGO FLOW AT
THE PORT OF AKITA DUE TO THE NIHONKAI-CHUBU EARTHQUAKE, 1983
BY
Akira Yonezawa*
Hajime Tsuchida**
ABSTRACT
Port facilities that are damaged by an earth- 2) The methodology for a survey had not been
quake result in a decreased cargo handling established.
capacity of the port since the cargo flew is
interrupted. Therefore, users of the port will The 1983 Nihonkai-Chubu Earthquake (magnitude
experience economical losses caused by the time 7.7) occurred offshore Akita Prefecture on
delays in handling cargo, i.e. reduced cargo May 26, 1983 (Fig. 1) . As a result, the coast
flew. along the Sea of Japan (Nihonkai) experienced a
strong ground shaking and Tsunami. The loss of
This paper addresses the Port of Akita and the lives exceeded 100, and the total damage to
decreased cargo handling capacity of that port facilities amounted to 180 billion yen.
as a result of the major Nihonkai-Chubu Earth-
quake of May 26, 1983. For the first tirre at Damage which occurred in Akita Prefecture
this port, economic losses due to the change of (which was the nearest to the epicenter)
cargo flew were surveyed. These losses were amounted to 110 billion yen. The damage to
evaluated to have amounted to 570 million yen ports in Akita Prefecture amounted to
during the period of one year after the earth- 10 billion yen in total, and Akita Port
quake. suffered the severest damage among them.
The results of this study will be used for For the first time, the economic losses were
planning of restoration work at port facilities surveyed at Akita Pori: after the Nihonkai-Chubu
after future earthquakes, for port planning and Earthquake. The survey was possible because of
earthquake port: preparedness planning. the availability of data on cargo flow from the
preceding years and the cooperation of the
Keywords: Earthquakes; Port facilities; enterprises which have been using Akita Port.
Economic losses
It was one year after the earthquake occurred
1. INTRODUCTION that the cargo handling capacity was restored
to the original level.
Damage to the port facilities by an earthquake
has a great impact on the community. Cargo In this paper, the results of the survey will
flew through the port has to be changed because be presented on the change of cargo flew and
of the decrease of cargo handling capacity. the accompanied economic losses during the one
This change results in economic losses to the year after the earthquake occurrence.
enterprises which have been using the port.
Proper port planning and earthquake prepared- 2. DAMAGE, RESTORATION AND USE
ness planning have to be established to OF PORT FACILITIES
maximize the investment efficiency, taking
social and economical losses into consider- Akita Port is located in Akita City which is
ation. The restoration work also has to be the center of industry and economics in Akita
planned to minimize those losses. Prefecture.
It was pointed out that those plannings re- The volume of cargoes handled in Akita Port is
quired sufficient information on the change of 5 million tons per year, which represents
cargo flew and the occurrence of the social and 80 percent of the total cargo that is trans-
economical losses caused by the earthquake. ported through all of the ports in Akita
Prefecture. The goods of large amounts are
Until new, earthquake damage of a port was oil, wood and ore. They are foreign or
surveyed only from the viewpoint of structural domestic imports (Fig. 2) . ^ They support
engineering, and was not surveyed from the
viewpoint of reduced cargo flow and accompanied
economic losses. * Researcher, Port Planning Laboratory, Port
and Harbour Research Institute, Ministry of
The reasons were as follows: Transport
** Director, Structures Division, Port and
1) The cargo flew through the port is very Harbour Research Institute, Ministry of
complex. Transport
459
manufacturing facilities of many enterprises 3. ECONOMIC LOSSES CAUSED BY CHANGE
nearby the port. OF CARGO FLCW 4
Port facilities were seriously damaged by the The 43 enterprises which had been using Akita
earthquake. This damage amounted to 8 billion Port were interviewed concerning their economic
yen. On the other hand, the nanufacturing losses caused by the change of cargo flew
facilities of the enterprises were only (Table 2). These enterprises can be divided
slightly damaged. The nanufacturing activities into two groups: manufacturers/wholesalers and
were interrupted after the earthquake to port service enterprises.
inspect the facilites, but were restored in a
short period. In this paper, the economic losses are defined
as follows:
2.1 Damage of Port Facilities ^
(1) The additional charge and cost of trans-
None of the 11 private berths (steel pile-type portation caused by the damage to port
dolphin) for oil handling were damaged, but 20 facilities.
out of the 25 public berths (caisson or sheet
pile-type structure) were very seriously (2) Reduction of income, as income is closely
damaged and became unserviceable (Table 1, related to the number of vessels arriving
Fig. 3) and the volume of cargoes handled at Akita
Port.
Three ship unloaders and four warehouses became
unserviceable because of the damage. The results of the survey are classified into
the three categories according to the cargo
For a week after the earthquake, the arrivals flow:
of vessels were restricted due to the outflow
of logs caused by the Tsunami. Case 1 Change of transportation route
2.2 Restoration Case 2 Change of facilities (berth, warehouse,

others)
The restoration works started in August 1983.
From June 1983 to March 1984, the 11 public Case 3 Others
berths were able to be used (Table 1)
Each case will be described in more detail as
In fiscal year 1983, one ship unloader and one follows:
tcwer-crane were set up and two warehouses were
repaired. In addition, the principal facili- Case 1 Change of transportation route
ties were restored.
This case occurred due to the congestion of
After April 1984, cargo handling returned to vessels and the restriction on the vessels
almost the same level as that before the arriving and on the utilization of the berths.
earthquake. This case was observed intensively from May to
July 1983, but continued to some extent for a
2.3 Use 3 year. The details were as follows (Fig. 6,
Table 3)
The nuirber of vessels arriving at Akita Port
and the volume of cargoes handled is shown in 1. Foreign Irrport
Fig. 4. From May to June 1983, they decreased
due to the restrictions on the vessels arriving 1) Zinc ore: Zinc ore which was scheduled
and on the utilization of the berths. to be discharged at Gaiko -13-meter-
depth wharf was transported to Akita
After July 1983, they were in the sane or Port after the transshipping to domes-
higher levels in comparison to those before the tic small-size vessels at Hachinohe
earthquake. This fact shows that the service- Port. The volume of the ore trans-
able berths were used very excessively, and ported in this way was 19 thousand tons
'this excessive use required extra expenditures and the additional cost was 42 million
to many enterprises, and those extra expendi- yen.
tures are considered as an economic loss in
this paper. 2. Domestic Export
The volume of cargoes handled at Gaiko 1) Zinc ingot: 1,000 tons of zinc ingot
-13-meter -depth wharf (damaged) and O-hama produced in Akita City were transport.^!
-10- rreter-depth No. 1 wharf (not damaged) are to Funakawa Port by motor trucks, and
shown in Fig. 5. It clearly shows that the loaded on a ship there. The additional
volume of cargoes handled at each berth changed cost was 1 million yen. Some other
before and after the earthquake. ingots were transported to Kurobe City
460
by rail. (The volume and the cost 4) Zinc ore: The additional cost caused
were unknown.) by the changes of handling berth, ware-
house used and the decrease of handling
efficiency was 55 million yen.
3. Domestic Irrport
Gement: Before the earthquake, the 5) Phosphorus ore: The additional cost
1)
cement had been transported to the caused by the change of warehouse used
users from Akita Port by motor trucks. was 3 million yen.
After the earthquake, it was trans-
ported by motor trucks or rail from 6) Flourite: The handling efficiency was
nearby ports and factories without lowered by the change of handling
passing through Akita Port. The volume berth, so that it had to be unloaded at
of cement transported by the alterna- night. The additional cost was
tive means was 130 thousand tons, and 3 million yen.
the additional cost was 260 million

yen. 2. Domestic Import:
2) Aluminum hydroxide: 1,400 tons of 1) Cement: Handling berth was changed. A

aluminum hydroxide were transported netfcement duct was constructed in a
from Funakawa Port to the factory in newly specified berth. Furthermore,
Akita City. The additional cost was the cement was transported by small-
2 million yen. size ships and the transportation fre-
quency had to be increased. The addi-
3) Bean cake: 600 tons of bean cake were tional cost was 150 million yen.
transported by motor trucks from
Funakawa Port to the factory in Akita 2) Potassium chloride: The additional
City. The additional cost was cost caused by the change of warehouse
1 million yen. used was 1 million yen.
Note: Log and flourite (foreign import) and Note: The berth for handling cryptomeria and
gasoline (domestic irrport) were unloaded at pine log (foreign import) and rice (domestic
other ports, but their volumes and additional export) vere changed, but the economic losses
costs were unknown. caused by this change were unknown.
The total additional cost amounted to The total economic loss amounted to 240 million
300 million yen. yen.
Case 2 Change of facilities in Akita Port Case 3 Others
The handling of half of the total volume of Other economic losses ware as follows:
cargoes was affected by the damage to the
facilities. The economical loss extended over 1) Demurrage was 8 million yen, as the
a period of one year, until nain facilities vessel arrivals were restricted after
were restored. the Tsuanmi.
Details to interruption of specific goods were 2) Pilot income decreased by 6 million

as follows (Table 3) yen, as numbers of entry of big-size
vessels (over 10 thousand dead-weight
1. Foreign import tonnage class) arriving at Akita Port
decreased.
1) Lauan: The additional cost caused by
the change of handling berth, the 3) Some enterprises bought their materials
decrease of handling efficiency and from adjoining enterprises because of
the congestion of vessels was 18 mil- lack of stock. The additional cost due
lion yen. to such unusual supply of the materials
was 9 million yen.
2) Wood chip: The additional cost caused
by the change of handling berth, the The total economic loss covering the three
congestion of vessels and the increase cases amounted to 570 million yen. It was
of transportation distance from the equivalent to 7 percent of the total cost to
berth to the factory was 8 million yen. reconstruct the damaged facilities.
3) Illuminate: The handling efficiency 4. CONCLUSIONS

was lowered by the change of handling
berth, so that the illuminate had to be For the first time, the economic losses caused
unloading even at night. The additional by the change of cargo flow were surveyed at
cost was 3 million yen. Akita Port, which was seriously damaged by the
1983 Nihonkai-Chubu Earthquake. It was
concluded that the total economic loss was
570 million yen.
It is considered that the economic losses at

Akita Port were minimized since the damaged
port facilities were restored within the short
time period of one year.
5. AFTERWORD
It vas not possible to evaluate the economical

losses of some enterprises. In the near future
this survey will be evaluated for its
effectiveness.
The results of the survey presented in this

paper will be helpful for evaluating several
alternatives on priorities of damaged
facilities for reconstruction work and on the
arrangement of high earthquake-proof berths in
7.
port planning and earthquake preparedness
planning.
6. ACKNOWLEDGMENT
Mr. Tomoo Ishiwata, Director of Design Standard

Division, Port and Harbour Research Institute
reviewed the draft of this paper and made many
valuable comments. The authors wish to express
their sincerest thanks to him.
REFERENCES
1) Port and Harbour Research Institute,

et al.; The Report on the Damage to Port
Structures and the Investigation of the
Tsunami Caused by the Nemuro Han to Oki
Earthquake on June 17, 1973; October
1973.
2) Civil Engineering Division, Akita Pre-

fecture; Damage Report of Civil Engi-
neering Facilities by the 1983
Nihonkai-Chubu Earthquake; May 1984.
3) Civil Engineering Division, Akita Prefec-

ture; Year Book of Post Statistics, 1982
through 1984.
4) Akita Yonezawa; Study on the Change of

Cargo Flow and Economical Losses at Port
Caused by Earthquake; Technical Note of the
Port and Harbour Research Institute;
No. 497; September 1984.
462
E 140
Fig. 1 Epicenter of the 1983 Nihonkai-Chubu Earthquake
463
others
others
3 )
Fig. 2 Trade cargoes at Port of Akita (1982)
Sea of Japan
Legend O-hama
'
Area
Marks Facil ities
Pubulic berth ( damaged )
f i ditto (not damaged)

—1 Private berth (not damaged)
mam Unloader ( damaged )
t Public warehouse ( damaged )
2
Fig. 3 Damage of main port facilities
4G5
o o Number of vessels arriving
(unit . vessel
0 20 40 60 80 100 120 140 160 180 200220
(unit 1,000 tons)
Volume of cargoes handled
Fig. k Number of vessels arriving and volume

3 )
of cargoes handled by month
466
3 o Gaiko -13 meter depth wharf ( damaged )
> ~
D O-hama- 10 meter depth No. wharf
I (not damaged)
Fig. 5 Comparison of volume of cargoes handled at wharves^

Legend
Fig. 6 Transportation map of goods after the 1983 Nihonkai-Chubu Earth

\
^ )
(They were handled at Port of Akita before the earthquake)
468
Table 1 Number of serviceable public berths
Existing mooring facilities Number of berths Number of serviceable berths
Objective From
scale of From June Aug 1983 From Apri 1 From July
Depth vessel Number of not damaged to July, to to June, to Dec
(meters) (dead weight berths damaged 1983 Mar. 1984 1984 1984
tonnage )
- 4 5 700 2 1 1 1 1 2 2
-5 0 1 ,000 4 0 4 4 2 4 4
- 5 5 2,000 3 1 2 1 1 3 3
- 7 5 5,000 7 2 5 7 4 7 7
-9 0 10,000 1 0 1 1 0 0 1
- 10 0 1 5,000 7 1 6 5 3 3 5
- 13.0 50,000 1 0 1 1 0 i 1
Tota 1
— 25 5 20 20 1 1 20 23
After Jonuary 1985, all the berths were oble to handle cargoes
Table 2 Number of enterprises surveyed
Item Act ivit ies Number
Manufacturing Chemical industry 3

industry Refining industry 3
and Paper industry 2
wholesaler Lumbering industry 3
Cement wholesaler 7
Petrolium wholesaler 10
Others 2
>
Port servise Shipping broker

industry Stevedoor
Warehousing 6
r
Pilot, etc
Nonlife insurance
company 7
Total 43
4G9
Table 3 Economical losses at Port of Akita
Economical loss Period of loss

Item Goods ( million Yen ) ( from) ( unt1 1 )
(1) Change of transportat ion route 3 0 6
Foreign import Zinc ore 4 2 May 1983 July 1983

Domestic export Zinc ingot 1 May 1983
Domestic import Cement 2 6 0 May 1983 May 1984
ditto Aluminum hidroxide 2 May 1983
ditto Bean cake 1 May 1983
(2) Change of faci 1 i t ies

in Akita Port 24 1
Foreign import Lauan 1 8 June 983 Jan 1984

1
di tto Wood chip 8 June 1983 Nov. J983

ditto I llminate 3 Aug 1983 Feb 1984
ditto Zinc ore 5 5 May 1983 Mar 1984
ditto Phosphrus ore 3 Aug 1983 Mar 1984
ditto Flourite 3 May 983 Nov 1983
1
Domestic import Cement 1 5 0 June 983 May 1984

1
di tto potassium chloride 1 Aug 1983 May 1984
( 3) Others 2 3
Dumurrage 8 May 1983

Decrement of income 6 May 1983 Dec 1983
Others 9 May 1983 May 1984
Total 5 7 0
m
LANDSLIDES AND DAMAGES DUE TO THE NAGANOKEN-SEIBU
EARTHQUAKE OF SEPTEMBER 14, 1984
Masateru Tominaga
Yasushi Sasaki
Katsumi Senoo
Hiroyuki Yoshimatsu
Eiichi Taniguchi
Kimiaki Nakano
Naomi Obinata
Yoshifumi Hara
Tetsurou Kuwabara

ABSTRACT Hypocenter : Latitude 35 49.3'N

Longitude 137°33.6'E
Focal Depth 2 km
On September 14, 1984, a severe earthquake with
a magnitude of 6.8 on the J.M.A. scale occurred
Magnitude : (Mj) 6.8 (Mj: J.M.A.
in central Japan. The magnitude had been Magnitude)
tentatively reported as 6.9 just after the
earthquake but was modified to be 6.8 later. Fig 1.1. shows seismic intensities of various
stations and Fig. 1.2 shows the number of
As a result of this earthquake, a large scale aftershocks observed at Iida Station of J.M.A.
landslide named Ontake-kuzure and many other The largest aftershock of magnitude 6.2 is
landslides occurred near the epicentral area reported to have occurred at 7:14 a.m.
causing damage to human lives and civil engineer- September 15. Fig. 1.3 shows distribution of
ing facilities. epicenter of aftershocks observed by J.M.A.
In this paper the authors report the damages to According to the investigation carried out by
civil engineering facilities and three main land- the Science and Technology Agency, the seismic
slides induced by the earthquake. intensity around the Matsukoshi Area in the
epicentral region had to be almost 6 in J.M.A.
1. INTRODUCTIO N scale. In addition, it is reported that seismic

motion around the epicenter of the main shock was
A severe earthquake occurred in the central part motion in which the short period component was
of Japan on September 14, 1984. Details of this predominant, rather large vertical motion was
earthquake were reported by Japan Meteorological included, and the duration time of the main shock
Agency (J.M.A.) as follows: was relatively short.
Earthquake Name : The Naganoken-seibu Fig. 1.4 shows maximum acceleration recorded by
Earthquake of
accelerographs installed at various structures
September 14, 1984
and on the ground near these structures. At the
Makio Dam near the epicenter no records of strong
Date and Time : 8:48 a.m. September 14,
1984 (Japan Standard Time) motion by the main shock were obtained because the
earthquake shocks were beyond the scales of the
Epicentral Region: Western Nagano Prefecture electromagnetic seismograph recorders. However,
the maximum acceleration is known to have exceeded
200 gal from the gain which was set at 200 gal 290,000m 3 . The third one, whose volume was about
for maximum acceleration. 150,000m 3 , was "Takigoshi Landslide". It
occurred on the left bank of the Ohtaki River at

Fig. 1.5 is an example of an acceleration record the upper reach of Ohtaki Dam reservoir. The
of an aftershock obtained by the seismograph fourth largest landslide occurred in the south-
which was installed at Makio Dam by P.W.R.I. east slope of Ontake Plateau and the volume was
(Public Works Research Institute, Ministry of about 100,000m 3 (Fig. 2.1). Small-scale land-
Construction) after the main shock. slides were concentrated in the area along the
Ohmata River and in the upper reaches of the
Although there is no record of ground motion Nigorisawa River and the Senzawa River. The
near the epicenter, the maximum horizontal landslides along the Ohmata River and in the
acceleration is estimated by Eq. (1.1). 2) upper area of the Nigorisawa River occurred in
the slopes of valleys incising the stratovolcano
Amax = 2.67R + 200 (1.1) The geology of the landslides in the upper area
A;,,^: Maximum Horizontal Acceleration (gal) of the Senzawa River is a shattered Paleozoic
R : Overturning Ratio of Gravestones (7») layer
As overturning ratio of gravestones at Matsukoshi Some rainfall had been recorded five days before
Area was reported over 907, by some investiga- and on the day of the earthquake (Table 2.1).
tions, maximum horizontal acceleration of main The mountain slopes were undoubtedly wet on the
shock is estimated to exceed 400 gal. days of the rainfall.
2. OUTLINE OF THE LANDSLIDES 3. DAMAGE TO CIVIL ENGINEERING FACILITIES
Due to this earthquake, numerous landslides Most damage due to this earthquake was concen-
occurred in many slopes near the epicentral trated in an area of radius not farther than
region. Landslides which are larger than 100m 3 10-15 km from the epicenter, and most of the
in area were caused at about 500 locations by the damage was related to landslides.
earthquake
The damage statistics are summarized in Table 3.1
Four of them were huge landslides whose collapse The total amount of damages was 46.3 billion yen.
volumes were larger than 100,000m 3 ;

others were The cost of damages to civil engineering
rather small and were slides whose collapse facilities and to forestry amount to 12.4 billion
volumes were about 800m 3 on the average. The yen and 29.4 billion yen, respectively, and the
largest landslide, which occurred at the south- sum of these two damage costs is about 907 of the
east slope of Mt. Ontake, was named "Ontake- total damage cost.
kuzure" and the collapse volume is estimated to
be 36,000,000m 3 . The debris from this landslide Roads were closed at many points because road
flowed down about 14 km along the valley and embankments had collapsed, because the roads
caused serious damage to people, houses, and were covered by debris from collapsed slopes
forestry. The second largest landslide was adjacent to the roads, because several bridges
"Matsukoshi Landslide", which occurred on the were swept away by debris flow and because a
left bank of the Ohtaki River at the upper reach road tunnel was filled by a debris flow.
of Makio Dam reservoir. The volume was about

As shown in Photo 5.1, the main entry route to landslide was the second largest in Japan in
Ohtaki Village collapsed at two places due to this century after the 150,000,000 m^
landsliding. Photo 3.1 shows the Kohrigase Hiedayama landslide of 1911.

Tunnel which was buried by the flow of debris
from "Ontake-kuzure" The ridge in which the landslide occurred was
composed of alternate strata of andesite lava
The large amount of debris from the "Ontake- and pyroclastic material underlain by a pumice
kuzure" dammed up the Ohtaki River at Yanagase, stratum. This pumice stratum is believed to
causing a natural reservoir, and this made local have been the slip surface.
inhabitants worry about a secondary disaster such
as a flood which might occur by failure of this It is presumed from the testimony of those who
natural dam. heard the sound of the landslide that the
collapse occurred almost simultaneously with the
Other disasters, besides landslides, which were earthquake
directly caused by seismic motion, were the
cracking of road surfaces, the collapse of road 4 . 2 Outline of Sediment Runoff
shoulders, settlements at the access portions

to bridges, damage to reinforced concrete bridge The debris avalanche caused by the Ontake-kuzure
piers, cracking of concrete girders, buckling of reached the Ohtaki River by the route shown in
the side corners of steel girders, damage to Fig. 4.3. It first flowed into the Denjo River
footings, and others. Photo 3.2 shows the Valley. Most of the debris avalanche flowed
damages to Matsubara bridge piers. The Makio down this river valley, passed through the
Dam (rockfill dam) which is located near the Nigorisawa River Valley and reached the Ohtaki
epicenter, had minor damage consisting of a small River Valley. Part of the debris avalanche
crack along the crest. Water supply was inter- passed over a ridge about 100m in height on the
rupted by damage to pipes and muddiness of the right bank of the Denjo River before entering
sources the Nigorisawa River Valley, and this part flowed
down to the confluence of the Denjo River where
4. SEDIMENT RUNOFF DUE TO SLOPE COLLAPSE it combined with that part of the debris
avalanche that had continued in the Denjo River
4 . 1 Ontake-kuzure Valley. The combined debris avalanche reached
the Ohtaki River, passing over the ridge near
The Ontake-kuzure is the largest landslide that the confluence of the Danjo and Ohtaki Rivers,
occurred on a ridge of the southeastern slope of passed through two narrow parts of the Ohtaki
Mt. Ontake (elevation: 3,063m). The elevation of River and flowed to a stop in the Kohrigase area.
the top of the landslide collapse was 2,550m, the Meanwhile, some of the debris avalanche from the
elevation of its lower end was 1,900m, the length source area passed over a ridge on the left bank
of collapse slope was about 1,500m, the maximum of the Denjo River and flowed into the
width was about 700m, the maximum depth was about Higashimatazawa and Nakamatazawa Rivers above
150m, its collapse area was 7,500,000 m^ and the Suzugasawa River. The debris that flowed into
volume of the collapse was 36,000,000 m^ . The the Nakamatazawa River was deposited above its
slope gradient of the collapsed part was originally confluence with Higashimatazawa River, while the
about 25° (Fig. 4.1, Fig. 4.2, Photo 4.1). This debris that flowed into the Higashimatazawa River
passed through the Suzugasawa River and reached
the Ohtaki River.
The volumes of erosion and deposition were as was uneven and looked like that of roughly
indicated in Fig. 4.4. In the Denjo River, loosened soil. The hummocks structurally
erosion was predominant and caused a riverbed retained the texture of the original natural
scour of 10-20m downstream to the point where ground and appeared to be dry. When artificially
the debris flowed Over the ridge into the crumbled, they were very brittle. Since this
Nigorisawa River. Below that point, riverbed structure was retained during the flow, it can be
scour was l-10m. In the Nigorisawa River Valley imagined that the flow wasn't mixed violently
there was a tendency to deposition with a river- (Photo 4.3).
bed rise of 0-14m. The bed of the Nigorisawa-
gana River rose 20-30m. In the Ohtaki River, In the Suzugasawa River, the amount of riverbed
the large debris deposits caused a riverbed rise deposits was small above elevation 1,450m. Below
of 30-40m. These deposits caused the Ohtaki this elevation, there were debris avalanche
River to be blocked above its confluence with deposits and mud-flow deposits. The mud-flow
the Nigorisawa River and to form a dammed lake deposits were predominant in the downstream
with a depth of about 30m (Photo 4.2). direction.
The volumes of erosion and deposition in the 4 . 4 Flow Marks

Suzugasawa River were smaller than those in the
Nigorisawa River; erosion was the tendency in The debris flow stripped vegetation and soil off
the upper reaches of the stream and deposition the valley walls and left flow marks as it
was the tendency in its lower reaches. proceeded. Fig 4.5 is a profile of the levels
of flow marks on both the right and left banks
The volumes of deposits shown in Fig. 4.4 have of the Nigor isawagawa River where vegetation and
not been multiplied by such factors as volumetric soil had obviously been stripped by the debris
increase rate (expansion factors). If the avalanche. It shows that the highest level of
deposits flowing into the Makio Dam reservoir flow marks was about 100m above the old riverbed
was zero, the volume of erosion was larger than and that the marks were highest on the sides of
the volume of deposition by 1320m^. It is the gorges of the river.
considered that, due to volcanic geology, the

porosity of the material involved was smaller The height of flow marks in the Suzugasawa River
after the landslide than before the collapse. (Fig. 4.6) was 20-40m in the upper reaches and
less than 10m in the middle reaches and suddenly
4 . 3 Riverbed Deposits decreased in the vicinity of the elevation of

1,350m. This may indicate that a large amount
After the debris avalanche, hardly any deposits of sediments was deposited in the vicinity of
were left on the bed of the Denjo River, debris this elevation.
avalanche deposits and small hummocks resulted

in the Nigorisawa River bed, and debris deposits 4 5 Movement of Deb ris Aval anch e
and large hummocks occurred in the Nigorisawa

River bed. In the Ohtaki River, the riverbed The debris avalanche velocity was presumed from
deposits consisted of debris avalanche deposits the observations of people watching some of the
covered by hummocks and mud- flow deposits. series of sediment movements that attended the
Ontake-kuzure . First, two people witnessed the
The surface of the mud-flow deposits was very initial debris avalanche in the vicinity oi the
smooth but the surface of the debris deposits upper gorge (named Gakiganodo) of the OhtaWi
River; the time of their observation was 5 . MATSUKOSH I LANDSLIDE
estimated to be about 8:56 a.m. Assuming that

the Ontake-kuzure occurred simultaneously with 5.1 Outline of Lan dslide
the earthquake, the average flow velocity for
the distance from the base of the landslide to The Matsukoshi landslide occurred along the right
Gakiganodo was about 23 m/sec. Then, the bank of the Ohmata River which flowed with the
average flow velocity for the distance from Matsukusa River. The landslide scale was 30m in
Gakiganodo to Kohrigase can be estimated to be maximum depth, 170m in width, 225m in length,
4-7 m/sec from testimony that the sediment had and 290,000m^ in volume. The collapsed area is
reached Kohrigase area on the Ohtaki River some located at the southerly- facing hillside slope
time between 9:00 and 9:05 a.m. It can be of Mt . Kurakoshi (EL 1100m) and is composed of
imagined from these facts that, in the upper andesite layers originating from Ontake Volcano.
reaches forming the Denjo and Nigorisawa Rivers, Two clear scarps, like the shape sharply cut by
the flow velocity exceeded the velocity of a a spoon, are successively observed through the
debris flow with a normal wave height and that terrace plain as shown in Photo 5.1 and Fig. 5.1.
from Gakiganodo down the Ohtaki River the flow After the landslide pushed a concrete plant
was rather slow. factory, which was located near the Ohmata river-
bed, up the opposite riverside slope about 35m
Another important testimony is that an aqueduct vertically, the greater part of these soils then
crossing the gorge of Gakiganodo had not fallen flowed down along the Ohmata River and were
in the debris avalanche at about 8:56 a.m., but deposited widely in the upper area of Ontake
fell in the first aftershock of the earthquake, reservoir, about 890m in horizontal distance
causing water to pour from the pipe. The pipe from the landslide collapse area. On that day
fell about 15m, onto newly deposited sediments much of the lake bed was seen because the water
on the right bank. The debris avalanche seems level was low (the reservoir water volume only
to have occurred more than once since the height attained 23.5% of filling).
confirmed in a subsequent field survey was 2-3m
(Fig. 4.7). The deposited soils consisted mainly of volcanic
ash, but partially contained some hummocks about
Calculated from the time 8:55 a.m., when a person 2m in diameter. From the geotechnical character-
saw mud-flow runoff in the lower reaches of the istics of this debris, the deposited form with
Suzugasawa River, the average flow velocity along ripple marks along the margin of the deposit
this river from the toe of the Ontake-kuzure was area, and also from the observations of a person
about 20 m/sec. Later, sediment runoff was who happened to watch the landslide, the soils
observed several times in the lower reaches. riding over the opposite slope behaved like
cement milk, and the flowing soil behavior is
K.J. Hsu charted the relation between collapse estimated to be the same as a mud-flow type.
volumes in great landslides of the past and their
equivalent friction coefficients (Fig. 4.8).^) 5.2 Outline of Ge ology
If the Ontake-kuzure landslide is included in
this chart, it is at a position with a smaller The landslide scarp forms a vertical wall shaped,
equivalent friction coefficient than the rest. from above, like a horseshoe, and layered
The Ontake-kuzure landslide apparently took a structure is clearly observed on the scarp wall.
runoff form having fairly small friction A geological profile of this structure is shown
resistance in Fig 5.2. The base rock consists of Paleozoic
475
layers containing slate, sandstone and chert It is estimated that a slip surface existed in
strata striking N70°E. This outcrop was the weathered pumice layer from observations
partially observed in the collapsed surface of resulting from the collapsed slope and soil
the head zone after the landsliding. Volcanic investigations. Triaxial tests were performed
products originating from Mt . Ontake lay on this by using undisturbed soil samples taken from the
base rock in layers. These products are classi- weathered pumice layer. Cohesion of the soil,
fied into two parts. One part which is called Cc 9
was 0.29 - 0.76 kgf/cm^ and angle of shear
new product consists of volcanic cobble, tuff resistance, 0 CU ,
was 22 - 24 degrees. Those
breccia and andesite lava layer. The other is values are relatively small. The weathered
old product containing layers of pumice, gravel, pumice contains a wide range of soil particle
volcanic ash and clastic deposits of mud-flows. sizes from clay to gravel; the mean particle size
By the boring investigation which was done around D 50 is 0.02 - 0.3 mm and the uniformity
the periphery of the landslide area, the new coefficient U c is over 180.
product layers cannot be correlated with layers
observed only a short distance from the landslide. Observations of the collapsed slope and measure-
Therefore, great erosional processes which ments of strike and dip of strata indicated that
created a very irregular erosional surface are the Paleozoic rock contours under the weathered
estimated to have taken place before the pumice layer formed a valley and water may
sedimentation of new product. The landslide slip easily have permeated into the buried valley
surface was observed successively in the strongly slope. The Matsukusa River flows into contact
weathered pumice layer. with the slope on its north side and it is
estimated that the weathered pumice contained a
5 . 3 Causes of Landslide lot of water when the earthquake occurred.
Soil investigations including borings were No strong motion record was obtained near
performed at the collapsed slope of Matsukoshi, Matsukoshi during the main shock, but, as
and triaxial tests were conducted by using mentioned before, it is estimated from the
undisturbed soil samples taken from the slope. collapse of gravestones that the horizontal
The following three causes may be pointed out maximum acceleration was over 400 gal. There-
for the landslide at Matsukoshi based on the soil fore, considerably large accelerations must have
investigation, geographical condition and the been applied to the slope at Matsukoshi during
characteristics of the earthquake motion: the earthquake.
(1) The weathered pumice layer which formed 6. TAKIGOSH I LAN DSLIDE
a slip surface had very low shear resistance.
6.1 Out line of t he Landslide
(2) The contours of the top of the Paleozoic

rock which underlies the weathered pumice The escarpment of this landslide Ib found at the
layer formed a valley and water was likely to level ridge (about 1 , 200 m elevation) on the vc k i
infiltrate into the slopes of this buried facing slope behind Takigoshi hamlet on the left
valley. bank of the main stream of the Ohtaki River. Thin

slope dipped about 50° before the landslide.
(3) A considerably strong earthquake motion The landslide collapse was 60m in maximum height,
occurred 150m in average width, 50m in average thickness,
and the volume of soil involved was about
476
7. REHABILITA TION WORK TO REPAIR DAMAGED
150,000 m^ It occurred on the left bank of
FACILITIES 4 )
.
a gully running at a right angle to the main
stream of the Ohtaki River. The plane configura-

The rehabilitation of facilities under prefectu-
tion of the escarpment is rather straight and the
ral control has cost a total of about 6,000
profile is almost perpendicular (Fig. 6 1).
million yen for 90 sites, namely: about 1,940
million yen for 75 road sections, 120 million
Some of the collapsed soil remained immediately
yen for 4 bridge sections, 40 million yen for
below the escarpment and in the gully, but most
an erosion control site, and 3,950 million yen
flowed away. Most of the soil flow went down
for 10 river sections. The rehabilitation of
along the gully for about 850m to the main
facilities under municipal control has cost a
stream of the Ohtaki River. It then flowed into
total of about 2,000 million yen for 171 sites,
the reservoir of Ohtaki River Dam, spread out for
namely: 1,880 million yen for 158 road sections,
a width of about 300m and was deposited there.
130 million yen for 10 bridge sections, and 50
Many hummocks measuring about 2.5m in height and
million yen for 3 river sections. Thus, the
about 10m in diameter can be seen in this
grand total is about 8,000 million yen for 261
area of deposition, particularly at the upstream
sites
apex of the area. These hummocks consist mainly
of boulders about 30cm in diameter.
Of these public works facilities, the first to
require urgent rehabilitation was the Matsukoshi
6.2 Details of the Escar pment
and Azuma area of Ohtaki Village, and sections
of the Ontake-Ohtaki-Kurosawa prefectural road

The upper part of the escarpment contains brown
line as a road necessary for local life. This
to dark brown volcanic ash soil (marked A in
rehabilitation work was delayed for several
Fig. 6.2) with a thin layer of light grey clay
reasons, such as the search for missing people
near the ground surface. A brown sand layer (B)
buried in the landslide in the Matsukoshi area.
is below the ash soil. Under the sand layer is
However, the Matsukoshi section of the road was
light gray to bluish gray andesite tuff (C and D)
opened on October 14, after the Matsukoshi area
and in its center there is apparently hard rock
temporary road construction (length: 310m,
with distinct columnar jointing (E). The lowest
width: 6m, approximate cost: 65 million yen)
part of the escarpment seems to be composed of
including the erection of a temporary New Ohmata
andesite tuff and volcanic effusive rock. It is
Bridge using a transportable built-up bridge
presumed from outcrops around the collapsed area
rented from the Ministry of Construction
that a layer of lacustrine sand and a clayey layer
(length: 40m, width: 6m). The Azuma section of
exist underneath.
the road was opened on October 29 after the Azuma
area temporary road construction (length: 600m,

Under these geological conditions and considering
width: 3 - 6m, approximate cost: 38 million yen).
that the configuration of deposition of the
collapsed soil at the toe of the escarpment is

From its confluence with the Nigorisawagawa River,
flat, the slip plane of collapse is concluded to
and downstream for a section of about 3 km in
have been formed in an aquiferous sand layer of
length, the channel of the Ohtaki River was filled
the Takigoshi lacustrine stratum distributed
to a depth of 30 - 40m with 20,000,000 m^ of
almost horizontally near the elevation of 1,140m.
sediments discharged from the Nigorisawagawa
River. These deposits dammed the river and
caused a lake of about 3,700,000 m^ in the upper
reaches of the Ohtaki River. Since this abnormal check dam, were carried out by the Nagano
deposition of sediments might be scoured in a Prefectural government (Fig. 7.2, Table 7.1).
future freshet and cause a washout of this new
dam with subsequent flooding, channel excavation 8. CONCLUSIONS
(length: 1,580m, volume: 89,500 m^) was carried
out to release water from the dammed-up lake. Characteristics of the damages due to the
This work assured adequate channel capacity. A Naganoken- seibu Earthquake of September 14, 1984
groundfill, using 349 concrete riprap blocks, was can be summarized as follows:
constructed in the Kohrigase area to stop
movable sediments. The approximate cost was 240 1) Since the epicenter was in the middle of
million yen (Fig. 7.1). Honsyu Island, the shaking was felt widely
in all directions. This felt area was
About 10,400,000 m^ of debris deposited in the similar to those in the past due to earth-
general area of the confluence of the Nigorisawa quakes with the same magnitude.
and the Denjo Rivers and in the Nigor isawagawa
River from the Ontake-kuzure may well cause 2) Severe damages were caused by landslides
damage in the lower reaches by abnormal sediment on slopes near the epicenter. The number of
flows during future freshets. Therefore, 14 slopes where the sliding area exceeds 100 m-’
small check dams were scheduled to be constructed was about 500.

near the area of the confluence of the Nigorisawa
River and the Denjo River. This work was planned 3) The severest of all damages to civil
by the Nagano Regional Forestry Office to be engineering facilities was the deposition of
accomplished by the end of fiscal 1984 at a cost debris in the Ohtaki River channel, where
of 1,150 million yen as an emergency forestry the volume was estimated to be 20,000,000
conservancy and damage rehabilitation project. Adding to this, damage to sabo dams, slope
Further, two sabo dams with a height of 14 m failures adjacent to roads, and damage to
(Nigorisawa Dam No. 1 and No. 2) are scheduled bridges took place.
to be constructed at the Nigorisawa River gorge
mouth by Nagano Prefecture under an emergency 4) The major damages to civil engineering
erosion control project and a countermeasure for facilities were caused mostly by landslides
landslides project. and debris from landslides.
Also scheduled is the construction of a few sabo 5) The big landslides were "Ontake-kuzure",
dams each on the main Ohmata River, on the landslides at Matsukoshi, at Takigoshi and
Matsukusa River, which is a branch of the Ohmata at Ontake Plateau. The shape of these land-
River, and in the Suzugasawa River, where hill- slides were different from each other.
side collapses along its upper reaches during Although failure mechanisms of each landslide
the earthquake were remarkable (Fig. 7.2, are not yet clarified, the existence of a
Table 7.1). weathered pumice layer among the strata which

constitute a slope is considered to be one
In, the Matsukoshi and Takigoshi areas, where of the main reasons.
large landslides occurred, and in the Uwajima

area where large hillside collapses must be 6) The movement of the debris from the
prevented, emergency landslide countermeasures Ontake-kuzure was presumably affected bv

including drainage, pile driving and a small water, especially in the lower reaches of
478
the Nigorisawa River, as determined from 4) Kiso Construction Office: Disaster of the
the study of the flow marks. Naganoken-seibu Earthquake, Nagano Prefecture,

November 1984 (In Japanese).
7) It was learned, from the damage caused by

this earthquake, that the method of estimating
the stability of mountain slopes during earth-
quakes, the rehabilitation of river channels
filled by huge amounts of debris, and the
countermeasures to increase the stability of
roadside slopes during earthquakes, are
important engineering subjects to be investi-
gated in the near future.
9 . ACKNOWLEDGEMENT
The authors are very grateful to all concerned in

the Chuubu Regional Construction Bureau, Nagano
Prefectural Government and Ohtaki Village Office

for their cooperation given to this earthquake
survey. The authors are also very grateful to

Science and Technology Agency, National Land
Agency and Fill Type Dam Division of Public Works
Research Institute for their supplying precious
data
10. REFERENCES
1) Investigation of Seismic Intensities

Distribution: Study Progessing Data of
Urgent Study on Sediment Disaster caused by
Naganoken-seibu Earthquake of 1984, March,
1985 (In Japanese).
2) Ohhashi, Iwasaki, Kawashima: Strength of

Earthquake Motion of the Izu-Ohshima-Kinkai
Earthquake of 1978 estimated from Overturning
Ratio of Gravestones, Civil Engineering
Journal 20-11 (1978).
3) Hsu, K.J.: Catastrophic Debris Streams

(Sturzstrons Generated by Rockfalls),
Geological Society of America Bulletin,
vo 1 . 86, 1985.
iffS
Unfelt
Felt
Note The numerol inside 'he Dor wesenfs

the number of fell ollercnocks
(Including momsnockl
15 16 l? 18 19 20 21
fMTrfl
22 23 24 25 26 27 28 2 9 30 I 2 3
1984 SEP. OCT
Fig. 1.2 Daily Frequency of

Aftershocks observed
Fig. 1.1 Seismic Intensity Distribution of at Iida Station
Various stations (After J.M.A.) (After J.M.A.)
IWAYA 0AM HIR0SE 0AM It)

Basei Foceofi InsxJeilnsideiGroutiRx* of i- . Dowry**' Upside
Ran Slope Core Core Tunrai Slope Bend* Grout
t k
1
1 1
S. Si S4 , S, 1 Sr Ss S. Ofhce funnel
8, , O*'
78 29 ,
IC
1
1
Sn,99 35 1 163 2001200 NS «3 14
® rW
U0 1
16 16 j
101 U0 /o'wn 64 j
lg 1 1 i“n£i; 4 ° 35 1 3 1
77
_c jj 0»eo» - 9reo*
XatL
-
3 1 12
Y0K0YAMA oam
Base Bose Com
Rocx Rock Bod* 0-sms®
R R
TR 30 10 80 8l
TU 9
UD 10 8 30 ,
9
LG Little 10 40 9 9
MNAT08RDGE USOGAWA DAM MANAGAWA OAM

(Pier 1 (Crownl iCrowni
TR 6 TR '
60 LG 60
UO 4 UO 50 UO 1 5
LG 10 LG 40 TR 35 *
«OSm!0u OAM
ta.1
•T i 1 1 41 1
& 7T "Is in :
_L«_-Li 18J
MARUYAMA 0AM _
K
JO I 34 1
3Q ,
LG 1 35 1
[UDI 0 _
'
- - -
TP 1 29 1
Note
I) 0 represents electromognetic seismogropn
0 represents small site analog record type susmogropn
Monmum acceleration volue m the parenthesis has io« r*i-oti |y because of •
of record.
Maximum acceleration volue of KUROBE 0 AM «as beared 07 pnone
Maximum occelerotion value of IWAYA 0 AM «ere meosuved Or do* off-cer
Colum marked by an oblique line s not observed component
Fig. 1.3 Distribution of

Aftershocks Fig. 1.4 Maximum Acceleration of Various Site
(After J.M.A.) based on Seismogram
480
100
0AM AXIS MAX. ACC =85 gat
50
-50
- 100 *-
0 5
(GAL)
TRANSVERSE DAM AXIS MAX. ACC.= 58gal
;,4u, •"
,inu i
. r
-
< a sm
0 5
(GAL)
UP -DOWN MAX. ACC. = -27 gal

-
Note : Item of Earthquake

Date and Time 9-18 am Oct. 9. 1984 (JST)
Hypocenter 35*51' N 137° 37' E Depth *2 km
-
i i(SEC) Magnitude 4.1 (J.M.A. Magnitude)
Fig. 1.5 Acceleration Record of Aftershock observed

on Rock beside Makio Dam
_L Mt. Ontoke ( 3,063 m )
V
\ o Large Landslide
Fig. 2.1 Plane of Landslides Locations
\
481
Fig. 4.3 Sediment Runoff from "Ontake-kuzure
482
ONTAKE LANDSLIDE
? „
UNIT : 10 m3
OHTAKI river
Fig. 4.4 Diposition and Erossion

Volume
ELEVATION
OHTAKI RIVER ond SUZUGASAWA RIVER
Fig. 4.5 Flow Marks Vertical Fig. 4.6 Vertical Section of

Section of Nigorisawagawa Suzugasawa River
River
EX PLANATION
• sedimentary rock o volcanic rock
a limestone ^ volcomclostic motenol
0.4 •
H /L •
£ • # O
*d* o &
0.2
**
a
a
Ontoke °
St Helens
1000
I10'm 3 l
Fig. 4.7 Cross Section near Fig. 4.8 Relationship between

Aqueduct Equivalent Coefficient
(H/L) and Rockslide Volume
483
<r
^ Cleaf
knick line to steeper slope
Maximum
of traces
dip direction
due to landslide r-Y:'y Flat plane

-<'r
< Unclear Cut
Ridge
* knick line lo steeper slope
Gully
Knick line to gentler slope
sgT. Layered structure
Fig. 5.1 Collapsed Area at Matsukoshi observed

from Oblique Photograph
ELM ELM
(Right side (Lett side
01 river I
of river)
ttoo HOP
1 050 1050
1000 1000
7"
'
- 1,
/
8 o Ovc ' 7'
950 m.
—y\
900 900
Olm Loom
Oam Andesite
Otb Tuff breccio

dt Detritus
850 850
Ovc Volcanic cobble
Ml Clostic deposit of
Opt Pumice tuff
Psl Slote
dr Collapsed detritus
Mgr Grovel 8-3. 0-5. 8-6 Borings
Fig. 5.2 Geological Profile at Matsukoshi
484
Fig. 6.1 Profile of Collapsed Area
and Geological Conditions
Legend
Escorpment
Port with developed columnor joint
Gully
Very smoll crock
Dork brown to blockish brown soil
Port of which qrovet ond sond ore moin components
Port of which soil is moin component
Fig. 6.2 Oblique Photograph of Collapsed

Area and Its Sketch
485
Fig. 7.1 Repair Works of Damage in Ohtaki River
A ^
Ml Onloke
3.063 \
\ N
Legend
IU»0)i»fll 2**o
o«i»»
Fig. 7.2 Sabo and Landslide Repair Works Plan

Table 2.1 Rainfall during 7-14 Sep. 1984
Unit : mm
Day Sep 14 A.M. Total
Rain-gauge station 7 8 9 10 11 12 13 (0:00 8:00)
Mt. Ontake (Mt. Mikasa) 0 0 119 10 1 0 1 27 158
3' 24 110
Miura Dam 0 1 80 3 0 0
Ohtaki 0 0 27 2 0 0 0 20 49
Table 3.1 Damage Statistics (After National Land

Agency, As of December 20, 1984)
Damage Cost
Category Unit Number of Damage
(Million Yen)
Fatality Person 14
Human Obscurity Person 15
Injury Person 10
Complete Failure House 14

Damage to
Residential Half Failure House 73
Houses
Part Failure House 565
Highway Site 264 6,940

Bridge Bridge 14 443
Public
River Site 15 4,930
Works
Sabo Site 1 40
Total Site 294 12,353

Forestry 29,419
Others 4,556
Total 46,328
Table 7.1 Statement of Item on Sabo and Landslide Works
Erosion control works Landslide countermeasures

Urgent rehabilitation Sabo works Most urgent and urgent
Most urgent rehabilitation Sabo works rehabilitation Sabo "works
Name of Name of Name of Name of Name of
Scale Scale Scale
torrent dam torrent dam area
H=6 5m . H=6 0m 10 0)
. ( Drainage, pile driving,
.
Ohmata River Ohmata No.l H=41 . Ohmata River Ohmata No. L=124 .Matsukoshi smail check dam, etc.
(steel) (steel)
H=4. 0(10.0) H=13 .
Matsukusa Drainage, pile driving,

Ohmata River Ohmata No. L=96.0 (124.0) Matsukoshi L=76 . Uwa j ima
River piling works, etc.
(steel) (concrete)
H=9 H=14 small check dam
. .
Matsukusa Suzugasawa Suzugasawa

Kurakoshi L=32 . L=99 . Takigoshi
River River No. 2 Drainage, etc.
(steel) (concrete)
H=7 . H=8 .
Suzugasawa Suzugasawa Suzugasawa Suzugasawa
River
L=62 . L= 109.0
No. 1 River No. 3
(steel) (concrete)
Nigorisawa H=4 . H=10 .

Nigorisawa Mizoguchi
River L=101. Mizoguchi L=112 .
No. 1 River
(steel) (apron) (concrete)
Nigorisawa Nigorisawa H=14 .
gawa River No.l L=101 .
Nigorisawa - Nigorisawa H=14 .
gawa River No. 2 L= 111 .
5 torrents 7 torrents 3 areas

Estimated total project cost: Estimated total project cost: Estimated total project cost:
740 million yen 2,800 million yen 1,020 million yen
(including 200 million yen
for emergency works)
Photo. 3.1 Damage to Kohrigase Photo. 3.2 Damage to
Tunnel Matsubara
Bridge
Photo. 4.1 Ontake-kuzure
488
Photo. 4.2 Natural Reservoir Photo. 4.3 Hummock in Ohtaki
in Ohtaki River River
Photo . 5.1 Landslide at Matsukoshi
489
GEOMORPHOLOGICAL CHARACTERISTICS AND FACTORS OF THE LARGE-SCALE LANDSLIDE AND
DETRITUS FLOW ON THE SOUTH SLOPE OF ONTAKE VOLCANO CAUSED BY THE
NAGANOKEN-SEIBU EARTHQUAKE, 1984
Masatoshi Nagaoka
Yoshihisa Hoshino
Yoshimichi Takei
Seiji Ichikawa
Fumio Saitoh
Geographical Survey Institute

ABSTRACT 2 . GENERAL BA CKGROUND
The largest landslide caused by the 1984 .2.1 The Naganoken-seibu Earthquake
Naganoken-seibu earthquake was 3.4xl0^m^ in
volume. It is the third largest in Japan in the The earthquake of M 6.8 occurred in the west-
past 100 years. This landslide consisted mostly ern part of Nagano Pref. at 8:48 a.m. on
of scoria and lava beds deposited in an old September 14, 1984. The epicenter was at the
buried valley on the slope of Ontake Volcano. southeastern foot of Mt . Ontake, very close to
During the earthquake shaking, these beds Otaki Mura (village). Although the magnitude
collapsed and slid along the old valley surface. of the earthquake was not very large, Otaki Mura
The characteristic microtopography of the area suffered heavy damage, losing 29 lives and many
and its changes before the landsliding are houses and facilities, due to the very shallow
recognized as precursors from a viewpoint of hypocenter. According to the analysis of the
historical development of landform. earthquake mechanism by Japan Meteorological
Agency, the fault or underground ;lip plane
The second largest landslide of 2.3xl0^m^ caused causing the earthquake is a normal plane of
by this earthquake is also ascribed to similar ENE-WSW strike. Widespread damage was caused
topographical and geological factors. by the landslides and accompanying detritus and
mud flows derived from the volcanic products

1. INTRODUCTION of Ontake Volcano and fluvial gravel beds.
The landslides caused by the 1984 earthquake are 2.2 Topography a nd Ge ology
different in mechanism from those caused by heavy

rainfalls. These ear thquake- triggered landslides The topography of the studied area is shown
have the potential to take place repeatedly in in Fig. 1. The landslides and the subsequent
areas where earthquakes are common. detritus flow and mud flows, excepting small
scale landslides, are displayed in Fig. 1.
In this paper, the processes and causes of the

landslides triggered by the earthquake are The studied area is in the southern foot of
discussed. Large scale maps of pre- and post- Ontake Volcano, which has formed in the late
landslides have been produced by photogrammetr i- Quaternary. Volcanic products are widely
cal methods for this purpose. The progressive distributed in this area. The basal rocV* of
topographical changes preceeding the earthquake Ontaki Volcano consist of two major type:
are analysed by photo interpretation and separated by an inferred fault extending from
characteristic microtopography related to the the central part of the volcano to the south-
landslides is discussed. southeast. One rock type is the Chichibu

Palaeozoic formation (to the east) and the
490
other is Nohi Cretaceous rhyolites, including 3. TOPOGRAPHIC CHARACTERISTICS AND FACTORS
O F THE LARGEST LANDSLIDE
intrusive granites, to the west.
3 . 1 Topographical Features
2.3 Precise L arge Scale Maps
Pre- and post-landslide precise large scale maps

The Geographical Survey Institute made pre- and
and aerial photos are shown in Fig. 2 and in
post-landslide large scale topographic maps
Photo 1-10. The results of the field surveys
(1/5,000 for the largest-scale landslide,
and interpretations of aerial photos (1948 to
1/2,500 for Matsukoshi and Takigoshi landslide)
the time of the landslide) in the area of the
for the areas involving the largest landslide
largest landslide are as follows:
and two other smaller landslides. The mapping
is specially noted as follows:

(1) A ridge collapsed and became a valley
with a downslope dip of 25°. A pumice layer
(1) It was taken into account to obtain
is the slip plane, and lava-scoria layers,
sufficiently high accuracy for determining
with a maximum thickness of 160-170 m, were
topographic changes from the pre- and post-
deposited above the pumice layer.
landslide maps.
(2) Comparison of several aerial photos

(2) For financial reasons, 1/1,000 topo-
taken since 1948 shows that a small scale
graphic maps made by Nagano Pref. were
landslide at the foot of the collapsed
reduced and compiled for making the post-
ridge had been gradually growing (Photo
landslide topographic maps (field surveys
10-1 to -4), i.e. the right-side wall of
were not carried out since mapping had to
the valley in the upper part of Denjo Gawa
be done urgently).
(River). This is particularly apparent by
comparing those photos taken in 1959 and
(3) Locations and altitudes on post-land-
1963.
slide maps were adopted as the control
points in conducting aerial trianglation
(3) Some lineaments (observed as small
for pre-landslide maps.
stepped cracks) striking ENE-WSW were present
at elevations from 2,025 to 2,150 m in the
(4) The aerial photos taken in the season
upper part of a small scale landslide. They
of defoliation (Forestry Agency, Yama-56,
are interpreted stereoscopically in Photos 2
taken in April 1969) contributed to the
and 3
accurate indication of the topographic
configuration.
3 . 2 The Process of the La r gest 1984 Landslide
For the geomorphological studies, many aerial

Based on the above-mentioned facts, the causes
photos taken by various organizations (1948, '59,
and processes of the largest 1984 landslide are
'63, '69, '74, '75, '79, and '80) were used for
inferred as follows:
the pre- lands lide survey. For the post-
landslide survey, many photos (scale is around
(1) Deposits of pumice and lava scoria
1/10,000, but some have a scale of 1/4,000)
layers had filled up the buried valley,
taken by survey companies soon after the land-
which was exposed after the landslide.
slides were used. Also, the maps mentioned
above were used.
ifil
(2) The foot of the collapsed ridge was 4. OTHER LANDSLIDES
intersected by the upper part of the Denjo

Gawa (River) valley, where a knickpoint, seen Many other landslides occurred during the 1984
as a waterfall, was present. Assuming that Naganoken-seibu Earthquake. Matsukoshi land-
the fall kept receding, the foot of the ridge slide (2.3xl0^m^) and Takigoshi Landslide
was subjected to the downward erosion of the ( 1 . 7xl0^m^ ) are the next largest. Ontake Kogen
valley, giving rise to instability. landslides are slightly smaller in size (the
largest of them is 3xl0^m3)* There were many
(3) Items 3-1(2) and (3) indicate that bed- other smaller landslides.
rock creep was progressing in the lower part

of the ridge. The position of enlargement of The process of Matsukoshi landslide is similar
the small scale landslide coincides with that to the largest landslide previously discussed.
of the downstream portion of the buried valley As to Takigoshi landslide, it is inferred that
exposed after the landsliding. This seems to a lava layer slid due to liquefaction of under-
suggest that groundwater flows in the buried lying lacustrine sediments. The straight scarp
valley enlarged the small scale landslide. of Takigoshi landslide might be of structural
origin. At Onkake Kogen, landslides were mostly
(4) The processes mentioned gave rise to of tephra layers.
instability in the ridge, and the landslide

was induced by the earthquake. The land- Precise large scale topographic maps and aerial
slide's plane overlies the pumice layer on photos of pre- and pos t- lands 1 ides are shown in
the buried valley slopes. In the process of Fig 2-4 and Photo 11-20.
landsliding, sliding of the lower part of the

valley fill induced the major landslide above.
These landslides seem to have occurred
simultaneously.
(5) A similar landform of an old landslide

can be observed to the west of the 1984 land-
slide. This landform has probably been formed
following the same processes as already
mentioned, i.e. the receding fall in the
lower part, the unstable foot, and, finally,
landsliding.
The culmination of various destabilizing

conditions such as the receding waterfall,
downward erosion, and increased instability of
the foot of ridge gave rise to both the large
1984 landslide and the prehistoric landslide

mentioned in (5).
492
Fig. 1 Distribution of landslide and accompanying deposits
( except small landslides )
f notice San,Yama mountain or peak

\
Gawa.Sawa river or valley
495
Fig. 2 Large — scale Landslide on the south slope of Ontake Volcano
500 109 0m 0
500 IQQO m
A) Post— landslide, 1/ 5,000 topographic map B) Pre— landslide , 1/ 5,000 topographic map
based on 1984. 9. 22 aerial photo, x0.5 based on 1969. 5 aerial photo, x 0.
C) Profile of the landslide
494
3 Matsukoshi Landslide S' S// //~>v ' BililOSim 1
mm
Post — landslide,
^v , -¥ j x/j
>1/ i-L
1/2 500 topographic map
~ - - •
'
based on 1954. 9 22 aerial photo, x 0.5 based on 1969
0 200 (V:H =1:1 )

Fig. 4 Takigoshi Landslide
A) Post— landslide, 1/ 2,500 topographic map B) Pre— landslide, \/ 2 , 500 topographic map
based on 1984. 9. 22 aerial photo, x 0. based on 1969. 5 aerial photo, x 0.
496
20 Cm
Legend
Earth floe deposit aocoopanymg Matsukoshi Landslide unconsolidate
( sand and gravel )
Holooene Ob Debris
[
brooola )
Olm :
Quaternary
M f Mud floe deposits and debris
( sand and
gravel, interbeddod pualoe layers)
Pleistocene
Lqr Lacas trine sediments
^ (
(utrvl and gravel, in ter bedded fxnloe layers)
Voloanloa or Ckuaite voioj
unoonsol Idate
end derived and.
J
Opm ; Pualoe layer ( thin )
•eal consolidate
i
Qvg Uud floe deposit and laoastrlne nnd.( underlying ) •ealoonsol Idate
Fig. 3 • D) Surface geologic map (
sand and gravel
)
consolidate
Palaezolo
of Matsukoshi Landslide Psl : 31»t.
oraok oeuned by ths oar tfxûake
497
Photo. 1 Large-scale landslide on the south slope of Ontake Volcano taken on 1984. 9. 22 8 days after the sliding
$#8 EH£<7)1984. 9. 22 T H C2 4715-4716. fjl/13.000. > 0 40
Photo. 3 The photograph before the sliding (36 years ago . taken on 1948. 9
1948. 9 X Ml 171 215-216. n 1/36.000 *1.0
498
Photo. 4 Source area and accompanying detritus flow Photo. 5 Main scarp of the landslide
&±n\ hxn\<^ 1984. 9. 30

sag®, zm\*ffc<nmPMv'if&r>myk&. 1984.9.17
+ a
Photo. 6 Lower part of the landslide and overflow area Photo. 7 Scarred transportation area of detritus flow
STSiSliiSajL ÛSTtKIc «fc & SS»«0
1984. 9.22 *•0-, 1984. 1 1. 11 mi
Photo. 8 Middle reaches of the detritus flow Photo. 9 Lower reaches of the detritus flow
«•)«« (jop m lit; f i'i #j a w<r>mm (7,: k®l i»iK ( fcj- ,® )

• i m ftiAftgi: Hit b
1984. 9. 17 I'll ipftttiSU ifiidl.'l. Lfc l.fi«WlttHt«iK, 1984. 9. 17 MlHiiM?
499
Photo. 10-1 Widening of shallow slide scar on the valley side since Photo. 10-2 Same place as 101 taken on 19f»9. 11.
36 years ago taken on 1948. 9

**>®.
Is),
1/15.000
1959.
txi.o)
11 m . MJfrr lU-148 In 08.
1948. 9 mi. ifcSjt M 1 1 71—215, 1/36.000 1X2.3)
Photo. 10-3 Same place as 10-1 taken on 1969. 9. Photo. 10-4 Same place as 10-2 taken on
r»i, *<7>®. mi M. -tco®. 1979. 10 «»i. Htfi'

Photo. 2 14
O’
3 #M. 1969. 5 Hi! . 1/17,000 (Xl.l)
Ifcttlt 01-560 C15-23, 1/17.000 ( x 1. 2 )
Photo. 11 Landslides of tephra layer on Ontake kdgen

1984 . 9. 30
500
501
landslide
*B*«t£*^
the
of 14
9.
view
1985.
Panoramic
ft.
***>$•
20
Photo.
502
SE I SMOLOG I CAL STUDIES ON THE NAGAN0KEN-SEI8U EARTHQUAKE, 1984
BASED ON JMA SEISMIC NETWORK
Minoru Takeo, Naoya Mikami and Masaji Ichikawa

Meteorological Research Institute, Japan Meteorological Agency (JMA)
ABSTRACT Focal parameters determined by the Japan Meteoro-

logical Agency (JMA) are as follows:
The Naganoken-seibu Earthquake (M; 6.8) accompanied

by a large-scale slope failure occurred on 14 Epicenter: 35°49.3'N, 137°33.6'E
September 1984. The source process of this earth- Focal depth: 2 km
quake was investigated in detail by using various Magnitude: 6.8
seismic data obtained by the JMA seismic network.

This earthquake was named “the Naganoken-seibu
Distribution of seismic intensities for the main Earthquake, 1984" by JMA.

shock suggests that the main shock was a peculiar
earthquake which radiated short-period waves In the past, two earthquakes (Zenkoji Earthquake
strongly. By using strong motion seismograms which 1847 and Hida Earthquake 1858) with large landslides
were recorded at stations within one hundred kilome- have occurred near the present epicenter. Detailed
ters of the epicenter, the distribution of disloca- focal processes for these two earthquakes, however,
tions and the character of the rupture propagation on were not obtained because of lack of instrumental
the fault were investigated. Results from this observations. On the other hand, for the present
analysis indicate that large dislocations ranging earthquake there are many data observed at stations
from 1.5 m to 2m occurred at the shallower part on around the source region which belong to JMA, uni-
the fault plane. Such a localized large dislocation versities and other research institutes. In this
might have caused the strong radiation of short- study, we investigate the distribution of seismic
period waves. intensities, focal mechanisms and the aftershock
distribution of the main shock and of the largest
The size of the fault plane is 12 km X 6 km, and the aftershock, based on the data of JMA seismic net-
seismic moment of the main shock is 3 X lO? 5 dyne*cm. work. Furthermore, by the use of strong motion
seismograms obtained by JMA stations near the epi-
Ground motions at the summit of Mt. Ontake and at center ( A< 100 km), we estimate the source process
Ki sofukushima were estimated by using the source proof main shock in detail. From this result, ground
cess models obtained in this study. The results motions, which had periods longer than 1 second for
obtained are as follows: the nearby source region are calculated. The
results of this study provide useful fundamental
maximum maximum data about the mechanism of the landslide.
di splacement vel ocity
2 . DI STRIBUTION OF SEISMIC INTENSITIES
Summit of Mt. Ontake 15 (cm) 12 (cm/s)

Ki sofukushima 18 (cm) 37 (cm/s) The distributions of seismic intensities for the
main shock and the largest aftershock are shown in

1. INTRODUCTION Figure 1. Radii of felt area (R) of these two
earthquakes are both about 400 km. According to the
On 14 September 1984 at 0848 JST (Japan Standard relation between R and JMA magnitude (M) for shallow
Time) an earthquake accompanied by a large-scale earthquakes in Japan [Ichikawa (I960)]:
slope failure, called "Ontake-kuzure," occurred at
the western part of Nagano-Prefecture, Honshu, Japan. M = -1.0 + 2. 7*1 og (R)
503
magnitudes of the main shock and the largest (pre-event epicenter) by Ishikawa et al . (1985). We
aftershock are both about 6.0. Though this value is determined the mechanism solutions of the main shock
consistent with the magnitude of the largest (main event) and the largest aftershock by using the
aftershock (M = 6.2), it is too small for the main P-wave first motion data of JMA and of WWSSN long-
shock magnitude (M = 6.8). period records. In Figure 4 are shown the lower
hemisphere equal-area projection of the fault plane
The maximum intensity reported by JMA was 4 (JMA solutions. Both earthquakes have strike-slip
scale) for the main shock and 3 for largest faulting with NW-SE maximum pressure axis. The
aftershock, respectively. It is reported that in the nodal plane of the main event striking in N70°E
source region the acceleration of the ground motion direction agrees well with the strike of the
5.
of the main shock exceeded gravitational acceleration aftershock distribution. For this reason, we con-
[Kuroiso et al . (1985), Ito et al . (1985)]. These sider this nodal plane to be the fault plane. The
facts suggest that the main shock is characterized by fault plane of the largest aftershock is considered
high-frequency seismic waves radiated rather strongly to be the nodal plane striking in N-S direction.
compared with the ordinary inland shallow earthquakes
in Japan. ANALY S IS OF THE SOURCE PROCESS
3 . A FTERSHOCK DISTRIBUTION The source process of the main shock was analyzed by
using the records obtained by the JMA strong motion
Figure 2 shows the spatial distribution of seismographs at Iida ( 1 1 D) , Takayama (TKY),
aftershocks that occurred from September 14 to Matsumoto (MTM) and Gifu (GIF), and by the long-
October 20. Their hypocenters were routinely deter- period seismographs at Matsushiro Seismological
mined by JMA. Most of the aftershocks are distri- Observatory (MAT) of JMA. The strong motion
buted on an almost vertical plane striking in the seismograms are shown in Figure 5. As exhibited in
N70°E direction. In order to see the aftershock Figure 6, the above four stations are distributed
distribution in detail, the aftershocks are plotted around the epicenter of the main shock, and the
in Figure 3 by dividing the period of time. The main distances from the epicenter to the stations are
shock is located at the center of the aftershock area 43 km, 46 km, 60 km and 86 km, respectively.
(region A) that occurred in 3 hours immediately after
the main shock. Assuming that region A is the source In order to carry out the analysis of the source
area of the main shock, the fault length is about 12 process, the seismograms were digitized at a rate of
km. The largest aftershock occurred in the western 0.1 mm on each record. The distortion of the origi-
edge of region A. Judging from the aftershock area nal records due to a mechanical recording system
(region B) following the largest aftershock in 3 were corrected after the digitization, except for
hours, its fault is conjugate to the main shock large traces just after the direct S wave arrival on
fault. After September 27, the number of aftershocks the records at I ID and MTM. After these correc-
increased in the eastern part of region A, and many tions, the digitized records were also interpolated
events occurred in regions C and D. The aftershock at a rate of 4 samples per second.
distribution is deeper eastward and the activity is

lower near the main shock hypocenter. The digitized records for the vertical component at
each station were not used in the analysis of the
4 . FOCAL ME CHANISM source process, because the records were too small
to use.
Records obtained by JMA 59-type seismographs
(
0 = 5 sec, h = 0.7, V - 100) indicated that the In addition to these strong motion seismograms,
main shock consisted of two events and that a small records obtained by long-period velocity type
pre-event preceded the large fracture (main event) seismographs (natural period; about 30 seconds)
about 1.5 seconds [Ishikawa et al . (1985)]. The installed at Matsushiro Seismological Observatory
itarting point of the main fracture was determined to were used in the analysis of the source process, in
ie about 3.7 km west of the main shock epicenter order to cover the eastern side of the epicenter.
The records were transformed to those equivalent to functions were calculated for the solid line struc-
the JMA strong motion seismograms. Arrows in Figure tures (structure-B) [Takeo (1984)]. The assumed Q
7 indicate times when the records were off scale. structure of the crust is shown in the right side of
Figure 8. A fairly low velocity in the surface layer
In order to investigate the distribution of the is assumed, and the effect of unelastic attenuation
dislocation and the character of rupture propagation, is not taken into account in the structure-A.
a nonlinear inversion method [Takeo (1985)] was used. Therefore, the seismic moments investigated by using
In the present study, the fault of the main shock was the structure-A may be nearly minimum values. On
divided into twelve sections at 3 km intervals, and the other hand, the effect of unelastic attenuation
double couple point sources were arranged in the in the surface layer of the structure-B is fairly
center of the sections. The source process of the low-Q and the seismic moments investigated by using
earthquake was assumed to be expressed by an integra- the structure-B may be nearly maximum values.
tion of the double couple point sources having the

same source time functions. The source mechanism, The resolution of frequency for the analysis data
obtained by the distribution of compressions and was about 1 Hz, because the records were digitized
dilatations of P waves of the main event, was used at a rate of 4 samples per second. In this study,
for that of each source on the fault. Horizontally we divided the fault plane into twelve sections at
stratified velocity structures were assumed for the intervals of 3 km. Judging from the average value
calculation of Green's functions, too. of rupture velocity and rise time of shallow earth-
quakes, and from the distances and the solid angles
t? and m-j denote the starting time of fracture and from the fault to the stations, this division may be
J J
the seismic moment at j-th point source, respec- enough to analyze the waves over the frequency range
tively. f-jj denotes Green's function which formu- from zero to one Hertz. The distribution of after-
lates the i-th seismogram due to the j-th point shocks occurring before the largest aftershock and
source. Then, the synthetic seismogram (F -j
(
t )
the point sources are shown in Figure 9. The
$
corresponding to the i-th observed seismogram parameters mj and tj represent the source process in
( 0-j (
t ) )
can be expressed as: the three square kilometers region round about the
j-th point source. The rise time of each point
F-f(t) = ?m *f ij (t-t ?). source represents the duration of the source process
j J
in each region.
We determine t? and m-j by minimizing the least-square
error (S(P)) defined by Twenty four parameters of mj , tj (j=l-12) were
investigated by the use of the two different velo-
S(P) = Ew-j *[0-j - F (p)]2 city structures, independently. The rise time of
i
i
double couple sources was assumed to be 1.5 seconds.
where w-j is the weight for the i-th seismogram and p These final solutions are shown in Figure 10.
is the column vector of parameters. S(p) is a non- Instead of the distribution of seismic moments, the
linear function for tj, so that this inverse problem average dislocations at each region are shown in
is nonlinear. We employ the modified Marqurdt method Figure 10.
in order to solve this nonlinear inverse problem.
The solutions are influenced by the initial values
In this study, we also investigated the difference of of parameters because of the nonlinearity in this
final solution due to the difference in velocity problem. There were several solutions which showed
structures used in the calculation of Green's func- much small residuals in comparison. Large minus
tions. Green's functions were calculated by the use seismic moments were given in some solutions. The
of two velocity structures as shown in Figure 8. In minus moment may be caused by the difference of
order to calculate Green's functions, DWFE method velocity structure for the real and assumed crust,
[Olson (1982)] was used for the broken-line structure irregularity in the rise time function of disloca-
(structure-A). Furthermore, taking into account the tion, variation in the source mechanisms on the
effect of unelastic attenuation of the crust, Green's fault plane, and the inverse dislocation. In view
505
of the evidence that the shortest epicentral dis- The seismograms at TKV , I ID , and GIF, calculated from
tance used in the present study was about 40 km, the final solutions, are shown in Figure 12. High
further discussion on this matter is meaningless. frequency waves in the later part of the seismograms
Particularly, source processes with large minus of the mode-A are caused by the calculation without
seismic moments in the middle part of the fault plane the unelastic attenuation in the crust. However,
may require one to suppose there were very compli- these synthetic and observed seismograms agree very
cated distributions of the initial stress and frac- well. In particular, the seismograms of the vertical
ture strength in and around the fault. Such a components are fairly consistent with the observed
complicated dynamics situation seems to be unnatural seismograms which were excluded from the analysis
so that the solution having smaller residual and no data.
large minus moments was adopted as a final one.
6. ESTIMATION OF GROUND MOTION IN SOURCE
REGION
The synthetic seismograms calculated by using the
s
final solution are compared with the observed Using the source process model ' obtai ned in the
seismograms. As shown in Figure 11, the synthetic former section, the ground displacement and velocity
and observed seismograms agree very well. The whose period is longer than 1 second at the summit
structure-B was slightly revised, as shown in Figure of Mt. Ontake (A = 16.3 km) and at Ki sof ukushima
8 with the thin solid lines, in order to fit the S-P (A= 13.0 km) were estimated. Since the divided
time in the Green's function to the observed S-P time sections at intervals of 3 km used in the former
at MTM and MAT. section were too large to calculate the ground
motion of these short epicentral distances, we
In spite of the difference of the velocity structures divided the fault plane into forty sections at
used in the calculation, the overall characteristics intervals of 1.5 km. Moments and fracture starting
of the final solutions are similar to each other. times of each source were interpolated from the
More concrete, the fracture starts at the shallower model-A and -B. Figure 13 shows the calculated
point in the central part of the fault and spreads ground displacement and velocity. At the summit of
over the whole source region within about 3 seconds. Mt. Ontake the estimated maximum displacement and
The large dislocation occurs in the near surface velocity are 8.3 cm (model-A), 15 cm (model-B), and
region. The field investigation suggests that the 11 cm/sec (model-A), 12 cm/sec (model-B), respec-
acceleration may be larger than the gravitational tively. The location of Ki sofukushima is in the
acceleration in the region ranging from 1 km to 4 km direction of strong SH radiation, so that the N-S
westward of the epicenter [Kuroiso et al . (1985), Ito component of ground motion is predominant. The
et al . (1985)]. This evidence is consistent with the maximum displacement and velocity are estimated to
dislocation estimated in this study. be 14 cm (model-A), 18 cin (model -B) and 33 cm/sec
(model-A), 37 cin/sec (model-B), respectively.
The dislocation at the deeper part of the fault seems

to be small. In particular, the final solutions 7. CONCLUSION
obtained by the two different velocity structures
suggest that no dislocation may occur at the deeper The Naganoken-seibu Earthquake, 1984 was analysed on
part of the western edge on the fault where the the basis of the data obtained by the JMA setsnic
largest aftershock occurred. network.
The total seismic moments estimated from the final Distributions of seismic intensities for the win
solutions for the structure-A and the structure-B, shock and the largest aftershock suggested that the
which are named model-A and model-B, are 2 X 10^5 main shock is a peculiar earthquake which radiated
dyne*cm. The maximum dislocations on the fault esti- short-period seismic waves strongly. Results fro«
mated from the model-A and the model-B are about the analysis of strong motion seismograms Indicate
1.5 m and about 2 m, respectively. The consistence that dislocation is i nhomogeneously distrlbut'-i »n
between the two models suggests that the final solu- the fault plane and has a maximum of about 1.5-2 m.
tions are highly reliable.
506
Such a localized large dislocation might cause the Ito, K., A. Kuroiso and Y. Umeda (1985), Region of
strong radiation of short-period waves. Large Acceleration Accompanied with Naganoken-seibu

Earthquake, 1984, Abstr. Seismo. Soc. Japan, No. 1,
From the distribution of dislocation on the fault, 18.
the size of the source area is approximately 12 km X
6 km. The total seismic moment of the main shock is Kuroiso, A., K. Ito, Y. Iio, Y. Umeda and I.
estimated to be 3(^1) X 10 ^ dyne*cm. Muramatsu (1985), Stones and Bogwoods Jumped due to
Seismic Waves, Abstr. Seismo. Soc. Japan, No. 1, 17.
Using Aki ' s formula

Olson, A. H. , (1982), Forward Simulation and Linear
1 og(M 0 ) = 1. 5*M S + 16.0 [Aki (1972)], Inversion of Earthquake Ground Motion, Ph.D thesis,
University of California, San Diego.
surface wave magnitude (M s )
is calculated to be
6. 3(+0. 1) Takeo, M., (1984), Synthetic Seismograms near Source
Region Taking into Account Unelastic Attenuation of

Ground motion, with periods longer than 1 second, Crust, Abstr. Seismo. Soc. Japan, No. 2, 255.
accompanying the main shock were estimated by using

the final solutions. The results obtained as Takeo, M. , ( 1985), Inversion of Source Process by
fol 1 ows Using Near-Field Seismograms, Abstr. Seismo. Soc.
Japan, No. 1, 168.
maximum maximum
di spl acernent vel oci ty
Summit of Mt. Ontake 15 (cm) 12 (cm/s)
Ki sofukushima 18 (cm) 37 (cm/s)
8. REFERENCES
Aki, K. , (1972), Scaling Law of Earthquake Source

Time-Function, Geophys. J.R.,_31, 3-25.
Public Works Research Institute (1982), Analysis of

Long-period Seismic Motion on the Basis of JMA Strong
Motion Seismograms.
Ichikawa, M. , (1960), On the Relation between Earth-

quake Magnitude and Maximum Value of Epicentral
Distance at Limit of Perceptibility. Quarterly
Journal of Seismology, _25, No. 3, 13-17.
Ishikawa, Y., N. Mikami, N. Hamada and S. Kashiwabara

(1985), Naganoken-seibu Earthquake and Its Preceding
Seismic Activity, Abstr. Seismo. Soc. Japan, No. 1,
1 .
907
Fig. 1. Distributions of seismic intensities (JMA scale) for the main
shock (M = 6.8, Sep. 14) and the largest aftershock (M = 6. ,
Sep . 15)
508
137.20 137.30 137. 40
137.20 137.30 137.40
Fig. 2. Spatial distribution of aftershocks that occurred from Sep. 14

to Oct. 20.
509
Fig. 3. Epicenter distribution for four periods: (a) 3 hours period
immediately after the main shock. (b) 3 hours period immediately
after the largest aftershock. (c) Sep. 27 to Sep. 29.
(d) Oct. 3 to Oct. 4.
Sep. 14 1984 Sep. 15 1984
Fig. 4. The focal mechanisms of the main shock (main event) and the
largest aftershock plotted onto the lower hemisphere by equal-
area projection. Solid and open circles indicate compression and
dilatation.
510
Takayama
D
Obs.
Gifu
Fig. 5. Strong-motion records used in this analysis.
I*,
j i
Matsushiro
'
°
O Matsumoto
U
r
:
Takayama
Takaî
w
4
s~
Gifu S O .Jk
Q..' t lida
j^ i
) \k
"L_12
u O""
h-
100 km
Fig. 6. Location of the main shock epicenter (Star) and the JMA stations
used in the analysis of the source process.
511
Matsushiro
Strong Motion
AM UD
AM NS
Fig. 7. The long-period velocity type seismograms at Matsushiro (left)

and the corresponding strong motion seismograms (right)
Vp.V s Qp,Q s
20 40 60 80 km/s 500 lOOO
Depth
Fig. 8. Models of velocity and Q structure used for calculation of Creen

functions. Dashed and solid lines indicate structure A and 11,
respectively. Thin solid lines indicate structure B revised nr I
MTM and MAT.
502
-e-
e
o o
0 c
» 0 0a( c o°
.0
V/
°
o
0
© \ c
©CfcO
cS>
o ?Pp
O o
to 0
e
„
•V*
°°oOOO
c^"
Q
b
3
° (« c
(
L
o0
Fig. 9. Arrangement of the point sources plotted on cross-section with

aftershocks that occurred before the largest aftershock.
Fig. 10. The starting time of fracture of each point (top) and the average
dislocation at each region (bottom)
513
Model A I I D TKY MTM
S N N N
0BS A
I J 4
1
STN. J
1 4
\
M
GIF MAT
S E N E
0BS
aM A/y
20
SEC
STN .
'/v p/W, -A-/
1
lf
Model B 1 1 D TKT MTM
s w N W
0BS •
t 1
STN . J A
GIF MAT
0BS .
-aM -vj
20 8
SEC
STN. -aA^ -aa/
Fig. 11. Comparison between observed seismograms and the synthetics from
the final solutions.
sia
Syn.
Model A — —
Model B A/Afjf-
—
3 cm
1 min
Fig. 12. The whole strong motion records and synthetic seismograms
calculated from the final solutions
Displacement Velocity
Kisofukushima
Z E
Model A
Model B —
Mt.Ontake
Model A —J\l\f^~
Model B
Fig. 13. The estimated ground displacement and velocity at the summit of
Mt. Ontake and at Kisofukushima for model A and B.
505
PAPERS
NOT PRESENTED ORALLY
BUT INCLUDED
IN THE PROCEEDINGS
FEATURES OF OCEAN WIND FLUCTUATIONS DURING TYPHOON PASSAGES
BY
Gen'ichi Naito* and Hiroshi Takahashi**
ABSTRACT 1 . INTRODUCTION
Characteristics of atmospheric fluctuations have Observations of sea winds under typhoons
been investigated concerning two typhoons which passages have been made at an observation tower
made passage around the Japan Islands. The wind in Sagami Bay. The tower is situated about 1 km
observation were made on the off-shore tower in off the Hiratsuka Coast, and is founded on the
Sagami Bay, and the waves were also seabed in approximately 20 m of water. Sagami
simultaneously observe. Bay is about 50 km wide in the east-west direc-
tion, and is open toward the south. Therefore,

Features of the ocean surface wind are
an infinite fetch can be expected for southerly
considered to be that behaviours of the large
winds
scale disturbance are added to ordinary turbu-
lent characteristics of the atmospheric boundary The observation tower is about 20 m high above
layer under high winds. Furthermore, when the the sea surface, and various sensors for the
high waves interfere with the air flow of the measurements of the atmospheric phenomena are
surface layer, this interference should be taken attached around it. Winds have been measured by
into account. a sonic anemometer and by a vane and anemometer.
The former was fixed at 685 cm high above the

Under typhoon conditions, the abnormally large
mean sea level, and the latter at 23 m high (see
gust factor is often shown, and the turbulent
Fig. 1). The tower proper is fairly large, and
intensities are very large. The surface drag
so a sonic anemometer at the lower level is
co-coefficient is also large, and depends
-
subject to blockage effects for certain dire
strongly on the mean wind speed.
tions. Therefore, winds can be observed only
Spectra of the wind vectors and of the vertical for the upward direction in order to exclude the
momentum flux are analyzed under strong winds, disturbances induced by the tower. In the pre-
and those behaviours are discussed in the fre- sent observation, a sensor of the soni •
anemo-
quency field. Additionally, the wave spectrum meter was positioned for the measurement f
is shown for the storm. southerly winds.
* Head of Second Coastal Disaster Laboratory, Features of the air flow over the sea ire
Hiratuska Branch, National Research Center for
Disaster Prevention. strongly influenced by waves. Theref r< ,
•
**Director-General National Research Center for

,
Disaster Prevention.
533
clarify the structure of the atmospheric boun- 2. TYPHOON TRACK AND SEA WIND
dary layer, it is very important not only to We have presently analyzed data from Typhoons
observe wind fluctuations but also to observe 8305 and 8410. Both typhoons moved northward
wave heights. Wave heights have been measured off Hiratsuka, in which the observatory was
by a capacitance-type wave gauge whose sensor located. Therefore, the strong sea winds could
responds rapidly to wave variations. be observed at the observational tower. Each
typhoon had the following development and

Three-dimensional wind components by the sonic
history
anemometer and a wave component are digitized in
the tower and transmitted to the data processing 2.1 Typhoons 8305 (Abby)
system on land by a cable buried in the seabed. The typhoon began to form in the sea near the
Finally, the gathered data are processed by var- Mariana Islands, on August 5, 1983, and devel-
ious statistical techniques. Other data are oped to extremely strong and large disturbances.
recorded on an analogue recorder. The minimum central pressure was temporarily
below 900 mb. The track of the typhoon near the

Wind fluctuations and wave heights are usually
Japan Islands is shown in Fig. 2(a). The cen-
sampled every 0.06 sec. A duration of one run
tral intensity decreased with movement toward
is 30 min, and the FFT method is used for
the southern coast of the islands and remained
spectral analysis of the turbulent eddies.
large in size. The typhoon made landfall near
Under the condition of high winds, a wind sensor Hamamatsu city on the afternoon of August 17.
is covered with a lot of water spray and, occa- Her forward speed was slow and she passed
sionally, is destroyed by high waves. Thus, the through the northern region of the Kanto dis-
very long observation is difficult, and the fea- trict. She decayed off the Sanriku coast. Fig.
ture of the air flow over a long time can be 2(a) shows the weather map for 0300 JST of
rarely investigated. August 18. The central pressure was 970 mb
before 24 hours.
The main purpose of the present report is to
estimate characteristics of the atmospheric Fig. 3 shows the time history of the mean wind
fluctuations above the sea during typhoon pas- direction W*D, the mean wind speed U, and the
sages near the observation site, and to compare maximum wave height C max during 10 min periods.
these characteristics with characteristics The wind data were obtained by a sonic
observed under ordinarily wind conditions. anemometer at a height of about 7 m above the
mean sea level after 0000 JST of August 17. The
wind speed gradually increased after 0200 JST.
519
The wind direction changed to the southeast at Korea Peninsula before moving into the Japan
about 0300 JST, and simultaneously the speed Sea.
came up to the high value beyond 10 m/s. Since

As the typhoon passed to the north of the
the typhoon developed vigorously in the southern
islands, the southern wind direction was
sea and approached the entrance to Sagami Bay,
observed much earlier in Hiratsuka, and then the
severe high waves were generated. As shown in
large change of the mean wind direction did not
the figure, before high winds arrived, there
occur. Fig. 4 shows the history of the 10 min
existed wave heights of about 4 m. Waves devel-
wind direction and wind speed as the typhoons
oped with time and destroyed the sensor of the
moved near the islands. As shown in the figure,
anemometer of 0530 JST. After that time, wind
the wind speed at Hiratsuka increased after the
fluctuations near the sea surface could not be
typhoon passage through the Tsushima Strait.
observed. The high waves resulted from wind-
The mean wind direction was southward over 24
generated waves being superimposed on large
hours, and showed considerably large variations
swells. We find also a beat of the ocean waves
with a short period. Moreover, the wind speed
with a relatively long period. The mean wind
also varied abruptly, corresponding to the vari-
speed was 1,4 49 cm/s at 0500 JST, and then it
ation of the direction. When a well-developed
became much higher. But we could not quantita-
extratropical cyclone passes in the Japan Sea,
tively observe the wind because of the above
its track resembles that of the present typhoon.
mentioned occurrence.
In this case, high winds also blow for a long
2.2 Typhoon 8401 (Holly) time. But we can scarcely find the abrupt vari-
No typhoons made landfall in the Japan Islands ation of the wind direction and speed (Naito,
on 1984. Typhoon 8410 is the typhoon which 1984). Characteristics of the atmospheric dis-
passed nearest to the Japan Islands. She was of turbance induded by a typhoon are considered to
average size, and brought southern high winds differ from those associated with an extratropl-
for a long time. The typhoon began to form into cal cyclone. When the typhoon moved nearer to
the sea around the Minamidaito Island, and the observatory from 0200 to 0900 JST, the wind
passed through the Tsushima Strait after direction ^gd speed varied must abruptly. Th<’
approaching Japan. Then she moved north- maximum mean wind speed of 10 min was recorded
eastward in the central area of the Japan Sea at 0600 JST, and had the value of 0 - 1,426
and turned into a extratropical cyclone on 2100 cm/s. The correspond ing gust factor was G - 2.4
JST of August 22 (see Fig. 2(b)). The minimum and the maximum instantaneous wind speed wi
central pressure was 965 mb in the south of the 3,434 cm/s. The center of the typhoon moved
520
over the sea and was about 460 km from respond to wind variations shorter than about
Hiratsuka 0.5 sec, even in high winds. A sonic anemometer
can be sensitive to much faster variations, and

As the typhoon did not approach Sagami Bay
so presently the sampling has been done every
directly, there existed only wind waves gener-
0.06 sec. Therefore, the present gust factor is
ated by southern winds and there was no swell in
somewhat larger than would be obtained using an
the bay. Accordingly, the maximum wave height
aerovant-type anemometer.
during 10 min was not large, that is, from 130
to 220 cm. Fig. 5 shows the gust factor G during 10 min
against the mean wind speed. Each symbol corre-

3. STATISTICAL FEATURES OF ATMOSPHERIC
TURBULENCE sponds to a different typhoon. Usually, G is
Turbulent characteristics of the air flow near small over the sea surface covered with wind
the ground surface are dominated by the surface waves only, and slightly increases with U. We
roughness. But we cannot ignore the long vari- have reported that the averaged value is G =
ation of winds as regards the structure of the 1.33 for moderate winds in the previous observa-
surface boundary layer. In other words, it tions (Naito, 1984). A group of the data
should be considered that atmospheric distur- plotted in the part of the figure corresponds to
bances having a scale much larger than the sur- the above described value, and agrees with it
face layer affect behaviour of the turbulent very well. But the atmospheric disturbance by
eddies above the ground surface. Therefore, typhoon gives gusts which do not relate to the
atmospheric fluctuations associated with a surface roughness. In the present observation,
typhoon may add special properties to the its gust factor is in the neighborhood of G =
turbulence in the boundary layer. 2.8, and has the maximum value of G = 3-0. This
air flow essentially differs from the forced

3.1 Gust Factor
convection near the sea surface.
The gust factor G of the ocean wind is small in
comparison with that of the land wind. More- In the case of Typhoons 8305, the gust factor
over, the factor varies as waves develope, has the special feature in high winds, and is
because the surface roughness increases with larger than the gust factor associated with
wave height. The observed gust factor also wind-generated waves. The fluctuations induced
depends upon the response of the anemometer. An by high waves form the greatest part of wind
aerovane-type anemometer is not sensitive to fluctuations. The maximum instantaneous wind
short gusts because the distance constant is speed during the observational time shown in the
usually about 8 m and the anemometer cannot figure is almost constant and is from 24.5 to
521
25.0 m/s. Therefore, G decreases with
o
u
/U = 0.122, o
v /U
= 0.096, o
w /U
= 0.053
increasing mean wind speed U.
for U = 11 m/s. We can find for the u component
3.2 Turbulent Intensity that the large fluctuations beyond the scale of
The standard deviation of wind vector (u,v,w) is the surface boundary layer greatly contribute to
denoted by (o u ,o v ,o w ), and the turbulent inten- the turbulent intensity.
sity can be expressed by (o /U, o

v /U, o
w /U).
3.3 Friction Velocity and Surface Drag
Fig. 6 shows the turbulent intensity of 30 min Coefficient
against the mean wind speed. The values When the momentum of the atmospheric
observed in Typhoon 8410 are divided into two fluctuations transfers toward the sea surface,
groups in the figure. One group was measured the scale of the flux is represented by the
before 1000 JST of August 22, at which the cen- friction velocity:
ter of the typhoon approached nearest, and the u# = A/-UW.
other group was measured after that time. The

In the boundary layer near the surface, u* is
lines in the figure represent the approximate
constant with respect to the height. As the
expression of the turbulent intensity in the
atmospheric stability is neutral under the con-
previous observations (Naito, 1983). As the
dition of high winds, the vertical profile of
values observed in Typhoon 8305 are few, we can-
the mean wind speed U can be described by using
not compare them with values obtained under
u# in the following expression:
ordinary wind conditions. However, we find that
u*
the intensity of the w component does not behave
-
U = — —
<
In
z
Zq
, <-0.39
as well as the characteristics of the turbulent
eddies in the surface boundary layer. The where < is Vou Karman's constant, z the height
fluctuations induced by. waves determine the fea- above the mean sea level, and Zq the aerodynamic
ture of the w component in this case. The pre- roughness. Zq gradually increases with the wave
sent plotted values are obviously larger than development, and is very small in comparison
the previous ones. As shown in Fig. 5, it is with that on land. The coefficient of the sea
the cas.e that the air flow contains very strong surface drag, which produces the shear stress >f
gusts under the typhoon. We scarcely indicate the air flow, is defined by
the difference before and after the nearest

2
c
D
- (u,/U) .
approach of Typhoon 8410, although the intensity
under the later condition is a little small. Therefore, we deduce the expression fir th<
Approximately, the averaged intensity is surface roughness by wind waves to have the r rn
522
From the figure, the average C D can be expressed
z = z exp(- ~j== ) ,
0 vu D as
and then we can obtain Zq if the drag C

D =
0.0001 9U for 7<U<14 m/s,
coefficient C D at the height of z is known.

where U is expressed in m/s. We can obtain from
Usually, the reported C D is the value at 1 0 m
the above formula C D = 2.28 x 10 ^ for U = 12
high, and Zq is of the order of 0.1 mm from the
m/s, and then the friction velocity is u* = 57
above equation. We are not aware that the
cm/s. Under the same condition, the sea surface
roughness under a severe storm such as a typhoon
roughness is Zq = 2.0 mm from those values.
has been directly measured because of very
difficult observation. Various statistical quantities of wind vectors
and waves in Typhoon 8410 are shown in Table 1.

Fig. 7 shows C D of 30 min as a function of mean
Those were obtained by processing the data for
wind speed U. The wind fluctuations in Typhoon
10 min every one hour. We can point out the
8305 were strongly influenced by ocean waves, as
characteristics from 1440 to 1700 JST of August
described in the section on wind gusts. There-
21. In the table, S../S,. and K../K.. are the third
fore, the covariance of the u and w components
and fourth moments of the u/w component, and are
has a very large positive value. In other
called the skewness and kurtosis factors respec-
words, the height of the anemometer was too low
tively. and T mov are the maximum wave
to obtain the atmospheric momentum flux which is
height and its corresponding period, o denotes
^
transferred to the sea surface. Consequently,
the standard deviation of the wave height and
C
Q
could not be obtained in this case.
indicates the intensity of the wave energy. We
The curve in the figure expresses the find from the table that the w fluctuation devi-
dependency of the average C

D
in the previous ates from the Gaussian distribution. When K u
observations (Naito, 1978). Also shown is the becomes very large, G is large. But when G is
drag coefficient reported by Davies and Flather large, K u is not always abnormal. This fact
(1978) who expect to derive the surface wind indicates that the strong gust has various
from the geostrophic wind over the ocean. C

D
types. The waves are not large, and those per-
observed in Typhoon 841 0 is larger than those of iods are from 4 to 6 sec. Consequently, the
the two examples. Moreover, the different fea- wind field is typically in the circumstance of
ture is that C increases with the wind speed wind waves.

D
and resembles the wind dependency of C by

D
Davies, et al.
523
4. SPECTRAL BEHAVIOUR OF WIND FLUCTUATIONS AND ponent along the wind direction is not strong-
WAVES
est, and the horizontal wind field becomes
4 . 1 Spectra of Wind Vectors
almost uniform. The decay corresponding to the
As is well known, the power spectrum of wind
-5/3 power law cannot be found in the inertial
fluctuations in the surface layer is represented
subrange of high frequencies, and the spectrum
as a function of the frequency normalized by the
decays slowly with increasing frequency. This
height and the mean wind speed, and of the
dependency is considered to be caused by the
atmospheric stability. In other words, the geo-
interference of high waves. The power spectra
metric similarity of the air flow is suggested
in Typhoon 8305, including the example of Fig.
to exist (Lumley and Panofsky, 1964). But the
7, are the special cases in which the interfer-
similarity cannot be obviously applied to the
ence of waves superimposes large scale
air flow such that the large scale fluctuations
disturbance.
are superimposed on the turbulence of the
surface boundary layer. Fig. 9 shows the power spectrum of the u
component in Typhoon 8410. In the figure, the

In high winds, the atmospheric stability is near
case of A-28 is under high winds before the
neutral. Additionally, except for summer, it is
nearest typhoon approach, and the cases of B-3
unstable in the ocean region close to the
and B-20 are after that time. Those spectra are
Kuroshio such as Sagami Bay. Fig. 8 shows the
of 30 min. In the figure, the spectrum extend-
frequency power spectrum P(n) of the wind fluc-
ing over the lower frequencies is analyzed for
tuations (u,v,w) in Typhoon 8305. The analyzed
the wind fluctuations during 720 min (equal to
example is obtained under high wind where the
12 hours), including B~3 and B-20. Its beha-
30-min averaged speed is 1,380 cm/s. In this
viour corresponds to the wind variation of the
case, as the atmospheric disturbance induced by
latter duration of Fig. 4. While the observa-
high waves is superimposed on the essential wind
tion was made, the wave condition scarcely
fluctuations, the spectrum is extraordinary in
varied, but the deviation of the tide was about
form. For the w component, with which the waves
60 cm. As the level of the anemometer Is .ibout
interfere strongly, the spectrum resembles the
7 m, the variation of 60 cm is large. But its
form produced by white noise, although it decay
influence could not be made clear as a result of
at high frequencies show a little turbulent
analysis. We find from the figure that the
behaviour. For the u and v components, the
spectra at high frequencies decay according to
spectra behave almost the same except in the
the -5/3 powers law. In the Inertial subrange,
high frequency region. Consequently, it is
the spectrum is described as follows,
noted in severe storms that the fluctuating com-
S2H
where f is the non-dimensional frequency, and z
nP(n) = £ e 2/3 (
JL)2/3
2 2im the level above the mean sea surface. Many
empirical formulas have been proposed for the u

where a = 0.M8 and e- is the viscous dissipation
component. However, most of them are based on
rate. The above formula assumes Taylor's frozen
the wind observation over land. The present
turbulence. Hence, we indicate that the three
example shows that the sea surface wind has a
spectra in the figure do not satisfy the above
large spectral density deviated toward the
formula according to the mean wind speed U and
longer period side in comparison with the land
have deviations larger than the values given by
surface wind.
the formula. This tendency can be seen in the
other examples. Therefore, considering the tui Fig. 10 shows the power spectrum nP
w (n) of the w
bulent structure of the surface layer, we note component and the co-spectrum nC uw (n) of the
that the relation of the reduction of the wave momentum flux by the turbulence. Each smoothed
number (or the wave length) to the frequency has curve with the observed values is given by the
not been sufficiently established yet. following formulas. The power spectrum of the w
component can be approximated by the formula of

One of the spectral characteristics during the
Naito (1978):
long period is that there exists a minimum in
the neighborhood of n = 10~3 Hz. Now, the wind nP w (n) 1 . 2f
disturbances weaken about every 17 min. A a

W
2
1+11f 5/3
smoothed curve with the observed values is given

The above formula does not agree well with the
by the empirical formula of the spectrum under
observed data around the frequency n = 0.5 Hz,
neutral atmospheric conditions from Naito
but is in good agreement elsewhere.
(1978). The long-period case and B-3 are
connected to apply the formula in the figure. The spectral scales and w are given by the
X X
u
peak frequency of the spectrum, and can be

Since the spectrum is non-dimensionalized using
obtained from Figs. 9 and 10. As the measured
the standard deviation o = 109 cm/s the case
u
height is f
= 685 cm, these scales are
B-3, the appropriate curve of the figure
X
u =
380 m for the u component;
represents
X
w = 23 m for the w component.
nP u (n) 74f nz
In a severe storm, such as seen in Fig. 8, the
2
(1+8i)f) 5/3 U
’
c
spectral peak cannot be determined and so the
spectral scale cannot be reduced.
525
In general, there exists a large scatter of the Fig. 11 shows a power spectrum of the waves in
covariance uw which determines the vertical Typhoon 8305. The maximum wave height during 30
transport of the atmospheric momentum. Then, min is 4 max = 352 cm, and its corresponding per-
the co-spectrum C uw (n) abruptly varies about the iod is T max = 11 sec. As the wave has been
frequency n. Fig. 10 shows a example of the recorded 4 max = 455 cm before this run, the sea
smaller variation. The uw co-spectrum decays circumstance is not severest, but is under the
more rapidly at high frequencies. Moreover, as ordinary condition of high waves. We indicate
the spectrum is wide and flat around its peak from the figure that the spectral peak corres-
frequency, its form resembles a trapezoid. ponds to the period of the maximum wave height.
Therefore, it is very difficult to obtain a suit- The spectrum in the high frequency region decays
able approximate expression. The curve in the with the -5 power of the frequency. Accord-
figure has the form ingly, although the sea surface is not covered
with only wind waves, the -5 powers law can be

nC uw (n) lUf
applied in the equilibrium range.
u#
2
+1
7/3
( 1 1 f)
In order to clarify the degree of interference

The observed form decays more rapidly than the
of the waves with the wind fluctuations, the
above formula, a high frequencies, and the u and
coherences between the wave height c and the u/w
w components correlate effectively at frequen-
components were computed (see Fig. 12). The
cies of less than about n = 6 Hz. This cut-off
coherence represents the correlation with fre-
frequency corresponds to the eddy scale of about
quency. The sensor of the wave gauge was not
2 m. Eddies smaller than this case usually do
just below the anemometer, but about 6 m off.
not contribute significantly to the shear
Therefore, we cannot obtain a more accurate cor-
stress.
relation. However, we notice the relatively
4.2 Interference of Waves with Air Flow high correlation in the range from 0.07 to 0.15
The layer in which waves interfere with the air Hz. From Fig. 11, these frequencies correspond
flow is close to the sea surface and is thin to the strongest energy range of the wave
(Kondo, et al., 1972). Therefore, the atmo- height
spheric fluctuations induced by waves are

5. CONCLUSIONS
usually ignored as regards to discussions of the
Features of the sea surface wind in Typhoons
characteristics of the wind turbulence. But we
8305 and 8410 have been investigated from th<-
can never find examples observed under the con-
measured values at the off-shore tower. Th<-
dition of severe high waves which are considered
former storm is high winds with high wav> : i, m'.
to influence the behaviour of the air flow.
526
the latter is long lasting high winds moderate November-December 1973. Dtsch. Hydrogr.
wind waves. Remarkable results are shown as: Z.Eng.-H. A. 1_4, 1-72.
(1) The gust factor above the sea surface with 2) Kondo, J., Y. Fujinawa, and G. Naito, 1972:
high waves is much larger than the ord Wave-induced wind fluctuation over the sea.
inary one due to turbulence alone. The J. Fluid Mech., 51_, 751-711.
factor under typhoon conditions is 3) Lumley, J. L. and H. A. Panosky, 1964: The
irregular and abnormally large. structure of atmospheric turbulence. Mono
(2) The turbulent intensity under the graphs and Texts in Physics and Astronomy.
condition of typhoon is large, and has the Interscience, New York.
average value that o

u
/U = 0.122 and o
w/U
= 4) Naito, G. , 1978: Direct measurement of
0.053. momentum and sensible heat fluxes at the
(3) The sea surface drag coefficient is large tower in the open sea. J. Meteorol. Soc
and increases with increasing wind speed. JPN , 56, 25-34.

-
We obtain that C 2.28 x 10 ^ for U = 5) Naito, G. 1983: Spatial structure of
D = ,
12 m/s. surface wind over the ocean. J. Wind Eng.
(4) The previously reported formulas can be Industr. Aerodyn., 1 >

67-76.
applied to the power spectra of the three 6) Naito, G. , 1984: Characteristics of long
wind components and the co-spectrum of the lasting winds over the ocean. Proc. 8th
momentum flux, except for the condition of Natl. Symp. Wind Eng., Tokyo, 1984, 1-6.
1)
strong interference of the waves. The
power spectrum of high waves is shown.
The study of sea winds under typhoon conditions
continues to clarify the more detailed feature
of gusty winds. Many observations will contri-
bute to obtain the basic data of the atmospheric
disturbance for the disaster prevention of winds
and waves in coastal regions.
6. REFERENCES
Davies, A. M. and R. A. Flather, 1978:
Application of numerical models of the North-
west European Continental Shelf and the North
Sea to the computation of the storm surges of
527
„ —
o
UJ' E
u
1
C\J
OsJ
OsJ OsJ
o
CO
i
CO
cn
Osl
1
CO OsJ
CO
CO
O o o
CO CO CO
00
l\J OJ
CT*
OJ
cn
OJ OJ
1
’
X
a3
E
.
V)
' LO o- Cn LO o Osl Osl cn o r-s r— 00 00 OJ
1 LO LO LO LO LO LO LO LO ^3* LO LO LO LO LO LO
X
03
E
'
E
p—
LO
LO
r—
c CT)
00 CT 1^
o 00
co
OsJ co
LO
LO
CO
o-
o LO
LO
^3-
LO
o
cn
OJ
CO
W' O (
r— i r— r— 1
I^S
1 i i
— r— f— OJ r— r— , r—
CD
00
C\J
CT
CO
Osl
co
o
Cn
CO
CO
co
co
o CO
o
*3- CO
cn
co
CT* LO
CO
CO
CO *:
cn
co
CO
co
«— <— Osl r- <— r- CO — OJ r- •— r- r- r-
o
o 1
o OsJ r— cn o LO LO oo LO r— LO r— f"-
O o
Q o o o
i
cn r— oo LO LO LO p— LO
— 1
1
•
OsJ CT*
•
X OJ OsJ 1 C\J 1 Osl Osl Osl OJ 1 * 1 OJ OJ 1 1 1
8410.
3
* CO 1
o- CO
CO
I^S o LO
LO
o LO
LO
OsJ
LO
*3"
LO
LO CO o
C3-
LO
IT)
LO
LO
o
^3- ^3"
U
Typhoon
—
3
co
*3"
CO
LO
oo
o- o-
, CO
LO
LO 00
LO
00
^3-
CO
CO
co
00
LO
LO
00
OJ
CO O
LO
Lt
LO
CO CO CO CO co co co LO CO co co co CO co co CO CO
in
cn LO LO cn 00 LO LO LO ^3- 00
waves
23 cn 00 Cn LO LO LO cn LO O'- 00 LO o ^3" CT* cn
OJ OsJ cn OsJ OsJ OsJ OJ 0J OJ OJ CO OJ OJ OJ OJ
_ o co — — 00
— o LO CO
OJ OJ 00
o LO
and r>s c3* ^3"
3 OsJ OsJ CO r— OsJ i i 1 r— OJ p—OJ O OJ r—
CD
o o o o o o o o o i
CD o o o o o o O
O — o
vectors
lo LO 00 OJ 00 OJ
LO
3 OsJ o o O J-
Osl
OsJ — J-
OsJ
,
o 00
o o f—
co o ,
CT*
1 i p—
o o o o o o o O o o o o o o o o
i 1 1 1 1 1 1
CD
wind
CO
3 —v LO CO O'* CO o- OsJ r- - CT* LO OJ ^3" cn r— LO
of D E LO LO LO 00 LO LO LO LO LO LO LO lT)
u
'
CO
> — cn CO 00 — — CT* LO CT* r- CT* o LO OJ co OJ LO
quantities
D E o- o-
1
O'* o
i
r—
— OO CO CO 00 oo cn
—o •
i i r
CO
3 " — LO co OJ co
Oh 00 00 LO
o OJ
CO CVJ
D E
u
00 1
1 1
OsJ
—
i
CO
f i
CT*
LO
i
OJ
—
Osl i
co
i
LO
r
o
*
CO
»
co
—
t
o
t—
r—
r
o
r—
Statistical
\E CO
CT
o 00 00
LO o LO LO
Osl cn
co
LO
LO
LO
CT*
LO
LO
00 LO
O- CT*
CT*
o-.
00
o
OJ
OJ
LO
^3-
o
CO
— oo o
o
1 r— — 1
OsJ
r
CT*
co
i
o
—
1
r—
i
O o
i •
CT* o
r—
r—
r—
cn r—
•
o
—
t
1.
Z
LO
OsJ CO
cn
LO
co
o-
LO
00 CT*
cn
O OJ
CO
CO
^3- LO
—
i
00
OJ
o OJ
LO
^3
LO
vO
ZD p— 1—
Table
02
c < i 1 i
c c c 1 1
1 i i 1 1 1 1 i 1 1 1 1
<c <c CO CO CO CO CO CO CO
528
180
aerovane
Fig. 1 Schematic diagram showin the spacing of

anemometers
0 i i i i i i i ————
i i i i
0 1 2 3 4 5 6
8/17 JST
Fig. 3 History of mean wind direction and

speed, and maximum wave height
during 10 min in Typhoon 8305
(deg
D
W
Fig. 4 History of mean wind direction and speed

in Typhoon 8410
Fig. 2 Track of the typhoon: (a) Typhoon 8305,

(b) Typhoon 8410
529
1 1 1 1 1 1 1 i i i i i i i T
o
0
• TVPOON 8305 ° 0
o
o 0
8410
- o -
- o -
• •
• •
•
•
o V
o
- -
- -
0
_
0 0° 0
0. ° 0
o _ <P
00
: _
o 8b
0 o
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
0 5 10 15 20
0 ( m / s )
5 Gust factor against mean wind speed Fig. 7 Surface drag coefficient as a function
of mean wind speed at 685 cm high
0-20
’
u-COMPONENT ’
v- COMPONENT
'
U) *
;
(b) •
4
*
o® 0 15
*eo *
0 e
#
f
. *' /
• •
f °^Vfo O "
a, ° *
^ 'rJ ,
NA1T0O983)
° A, 0 05
l TYPHOON 8410
*
8305
5 10 15 5 10 It
_
U(m/s
Fig. 8 Frequency power spectra of wind vectors
in Typhoon 8305.
’
w- COMPONENT a
;
(c)
i
%X)
005
5 10 It
U( m/s )
Fig. 6 Turbulent intensity of wind fluctua-

tions: (a) u component, (b) v compo-
nent, (c) v component. Solid lines are
from Naito (1983).
530
Fig. 9 Power spectra of longitudinal wind
component in Typhoon 8410. Smoothed
solid curve is from Naito (1978).
)
1
/s
,
Fig. 11 Same as Fig. 8 except for wave height
(n)(cm
uw
nCo
&
(n)
w
nP
n ( Hz
Fig. 10 Power spectrum of vertical wind com-

ponent, and co-spectrum of momentum Fig. 12 Frequency coherence of wave height with
flux in Typhoon 8410. Smoothed curves longitudinal wind component (solid
correspond to empirical formulas. line), and with vertical wind component
(dotted line)
5a
THE VARIATION OF GUST STRUCTURE WITH HEIGHT UP TO 200 M
BY
K. Naito* , T. Hanafusa**, and T. Fujitani***
ABSTRACT bridges have made the consideration of the wind
The variation of 'gust structure with height up to forces and earthquake equally important.
200 m was investigated by using the ciata from a

Accordingly, the study of the characteristics of
meteorological observation tower equipped with
high wind at only one point is not sufficient and
sonic anemometer-thermometers at six levels.
spatial and time variations of high wind's
The vertical and horizontal scales of turbulence characteristics are required.
for three wind components were discussed with the

This paper will concentrate on the variation of
integral scales obtained by the integration of
gust sturcture with height by using wind data
the auto-correlation coefficient of wind fluctua-
from a meteorological observation tower with 213
tions and the cross-correlation coefficients at
m height.
two different levels.
2. INSTRUMENTATIONS AND DATA REDUCTION

The variation of the spectral scales with height METHOD
was obtained by the spectral analysis of wind The principal source of the data for this study
fluctuations. is wind fluctuations measured by three dimen-
sional sonic anemometei — thermometers located at

All terms in the energy balance equation except
10, 25, 50, 100, 150, and 200 m on a meteorologi-
the pressure transport term were estimated
cal tower. The signals from the sonic
directly and the energy balance was discussed.
anemometei — thermometers were digitalized at every
1 . INTRODUCTION 0.1 sec and recorded on magnetic tapes. The data
Earthquake forces have been more important for analysis was carried out on the MRI large elec-
planning the most structures in Japan than have tric computer (HITAC-M-200H) . The data discussed
wind forces. But recently, the reduction of tall in this paper were the same ones presented in t’.*
buildings' weight, the utilization of outer walls previous UJNR session (Naito, et al. (1983)) ).
to achieve strength and stiffness in buildings, The tower and the observation system were pr»-
and the construction of long space suspension sented in detail in another paper (Hanafusa, et
2
al. (1981)) ).
* Head of the Meteorological Satellite Research

Division
** Physical Meteorological Research Division,
Chief of the Second Research Laboratory.
552
3. RESULTS AND DISCUSSIONS mean wind speed by a gust factor. On the
3.1 Frequency Distribution of Wind other hand, the maximum wind speed can be
Fluctuation Maximum Instantaneous Wind
Speed infered from the fact that the frequency
Frequency distribution of wind fluctuation distribution of the wind fluctuation is nor-
is considered to be almost normally distri- mal. As mentioned previously, the frequency
buted, regardless of the atmospheric stabi- distribution of the u component is found to
lity and of the roughness length. Examples be normal. Accordingly, the probability of
of the frequency distributions of wind com- a maximum 5 sec wind speed for 10 min samp-
ponent are shown in Fig. 1 and the values of ling duration is indicated by the shaded
skewness and kurtosis are plotted in Fig. 2. region as shown in Fig. 3. From the normal
Though the observed points vary a little distribution table, it is about 2.4 and the
widely, the values of skewness are central value of the 5 sec maximum wind speed for 10
to zero and the values of the kurtosis are min sampling duration, U max can be expressed
close to 3- Accordingly, the frequency dis- as follows,
tribution of each component is considered to Umax = M + to = M(1+2.4 a/M)
be adequately described by the normal where M is mean wind speed and a/M,
distribution. intensity of turbulence. The calculated
values obtained by using observed mean wind

In the case of studying wind load, it is
speed and intensity of turbulence are com-
preferable to adopt the maximum instantane-
pared with observed maximum wind speeds.
ous wind speed rather than the mean wind
The results are shown in Table 1 and Fig. 4.
speed. As the response of the anemometer has
From these results there is no remarkable
a limit, even the maximum instantaneous wind
difference between them. It is very easy to
speed indicates the mean wind speed which is
estimate the maximum wind speed if the
averaged for a certain degree of time.
intensity of turbulence can be estimated.
Accordingly, as the value of maximum instan-
taneous wind speed will depend on the aver- The power index of the vertical distribution of
aging time, it is very important to make a the maximum wind speed is usually smaller than
description of the averaging time when the that of the mean wind speed. It will depend on
maximum instantaneous wind speed is mea- the averaging time that the index in this experi
sured. The gust factor is defined as the ment is relatively smaller than the root of that
ratio of the maximum wind speed to the mean for mean wind speed a shown in Fig. 5.
wind speed for some long period. Then the
maximum wind speed can be estimated from the
533
3.2 Correlation and Scale of Turbulence LV - z °-3
0 7
The spatial scale of turbulence can be determined
v - »
'
by the integration of the auto-correlation func- In order to examine the spatial characteristics
tion for wind fluctuations. An example of the of wind fluctuations, we must measure wind
auto-correlation coefficient for u, v and w com- fluctuations by the sme type of instruments at
ponents vs time lag is shown in Fig. 6. The the same time and at several points.
values for v and w components become smaller com-

The cross correlation coefficient between two
pared with that of u component at same time lag
different heights can be estimated by the sonic
and the higher the measurement height is, the
anemometer's data at six levels. Its value is
larger the lag time at which R ( t ) becomes zero.
considered to be a function of the averaging
The values of R(t) plotted against the height are
time. The dependence of this coefficient on the
shown in Fig. 7 for each wind component. For the
averaging time is shown in Fig. 8 and Table 2.
u component, the values of R(t) increase linearly
The longer the averaging time becomes, the larger
with height up to 1 00 m and are nearly constant
the coefficient is and the tendency of the
at higher levels. On the other hand, the value
increase becomes large abruptly at more than 3
for the v component are almost constant for the
sec. This indicates that there are phase differ-
entire 200 m layer while the values of w compo-
ences between wind fluctuations at two levels.
nent increase linearly with height. This ten-
dency can be found in the previous paper (Naito, Table 3 indicates the time lag when the cross
1
et al. (1983)) ) for the case of t
q
where R(t q ) correlation coefficient has a maximum value.
is 0.6. For each component, /R(t)dt can be From this table, the maximum value can be found
expressed as follows, at plus time lag (the gust wind at higher levels
u component /R u (t)dt - z 0 ’^
(except the is faster than that at lower levels) for u and v
higher two levels)
components, though there is no remarkable differ-
v component /R y (t)dt - const
ence for w component. The time lag between 1 >< m
w component /R w (t)dt - z^'^
and 50 m is largest when the wind shear is
If the Taylor's hypothesis can be assumed to be largest
applied for every component, the integral scale

As shown in Fig. 9, the cross correl itlon
of the turbulence of each component can be
coefficients among three wind fluctuations a*
expressed as follows, as the vertical mean wind
three different levels are approximately
0 3
D
profile can be approximated as z ,
(Naito, et
expressed as follows,
1
al. ( 1 983) )
r " r1 * r
1 ,3 ,2 2,3
L
U - z°-8
534
= cross correlation coefficient between The dependency of the vertical scale of
j
height z^ and Zj
turbulence on height is similar quantitatively
This suggests that it should be possible to
with that of the horizontal scale of turbulence.
express the cross correlation coefficient by the
following formula, 3.3 Spectral Character istics of Wind

Fluctuations
r exp { — k —
2 = ( 4> (
2 ) (
z-| )}
1
In turbulent flow, various scale eddies mix in

As there has been no definite function for <f>
( z )
confusion and the distribution of their energy is

the relation between the correlation and <f>(z) can
not homogeneous. It is a very important problem
be examined by changing the value of n from 0.2
to examine how the energy is distributed over the
n
to 0.7, assuming <}> ( z ) is z . Results are shown
frequency range in order to consider the dynamic
in Fig. 10. As observed values fairly scattered,
character istics of the structures. One of the
the function of <f>(z) can not be determined uni-
methods used to examine the distribution of wind
17 ^
formly. For w component, 4>(z) is almost z as
energy is spectral analysis of wind fluctuations.
mentioned by Panofsky and Singer (1965). For u
Spectral analysis was carried out by the Blackman
and v components, the power index larger than 1/3
and Turkey method and the sampling duration was
is seen to be fitted for the entire range. From
10 min for all components.
the figure, the cross correlation coefficient can
be expressed as follows, The averaging time is 2 sec for u and 0.2 sec
4 ~ 4
R
u =
exp{-0.5(z2* z°* )} for v and w components. Example of spectra for
R
y =
exp{-0.3(z2’ 6 “ z°‘6)} each component are shown in Fig. 11. The dotted
z® ^ -
R
w = exp{-1 .2(
z®’3)}
’
line in the figure for u and v components indi-
As mentioned before, the correlation will vary cates the experimental formula proposed by
4 5
according to the averaging time, and the values Davenport and for w, by Busch and Panofsky .
of n and k will change. Accordingly, the aver- The spectra were shifted in order to match the
aging time is a very important factor in deter- peak frequency. For the u component, the experi-
mining the correlation and scale of turbulence. mental formula was in agreement with the spectra
obtained in the lower levels (50, 25, and 10 m)

Finally, the scale of turbulence by the above
i this experiment. But for higher levels there
formula can be expressed as follows,
are remarkable spectral peaks with frequency
L“U) - 20 6 -
about 0.006 Hz at which the experimental formula

- 0 ' 4
Lj(z) z
can not be fitted.
L”(z) - z 0 - 7
On the other hand, Davenport's spectra coincide
with observed v spectra and the -2/3 power law
535
can be observed in the higher frequency range. shown in Table 4. The height variation of e are
_1
For w spectra, the experimental spectra proposed shown in Fig. 14 indicates that c ~z for heights
by Busch and Panofsky well fitted the observed less than 50 m and e-z for height .greater than
one. It is found that the peak frequency, n in 50 m. It may indicate that the height of the
m ,
which nF(n) has a peak shifts to a higher fre- constant flux layer is at most 50 m.
quency range as the height decreases. The spec-

The distribution of w'E has a maximum at 1 00 m
tral scales, A
m (=u/n m ), are shown in Fig. 12. For
for all runs and then the direction of the turbu-
0 q
J
up to 50 m layer, it is shown that A -z and
lent energy flow will change at 1 00 m height
o J
3
A
y , A
w ~z and at higher levels than 50 m, the
(Fig. 15). The balance proposed by Busch and
spectral scale of each component is proportional
Panosfky (1968)'’, namely that - = M+B, seems
to height, -z.
appropriate in this case as is shown in Fig. 16.
The relative magnitudes of the spectral scale Accordingly, insofar as the residual represents
hardly change with respect to height and it is the pressure term, P, the results then indicate
found that An
U : A,.
V
: A,
w ,
= 6 : 2 : 1 . that -P=T , i.e., that the divergence of pressure
transfer and divergence of the turbulent energy

3.4 Energy Balance Equation
flux sum to about zero. This conclusion is in
The energy balance equation can be written as
agreement with that by McBean and Elliott (1975)
follows, assuming homogeneity of turbulence and
in which these two terms were estimated at a
zero mean vertical velocity,
height of 5.77 m above dry prairie grassland.
—
—
3
3t
(E) = -u'w' — 3u
3z
g
+ -Sw'T'
T
3
3z
1-
Kw'p')
p
+ —3
3z
(w'E)-e
4. CONCLUSION
M B P T
The variation of gust structure with height up to

2 2 2
where E = 1 /2( u +v +w )
200 m was investigated by using the data from

M: mechanical product term
the meteorological observation tower with sonic
B: Buoyant production term
anemometei — thermometers at six levels and the
P: Pressure transport term
major levels results obtained in this experiment
T: Divergence of the flux of turbulent kinetic
energy term were as follows,
All terms except the pressure term can be estimated (1) It is found that the wind fluctuations for
and the values of e can be obtained indirectly each component, u, v, and w are normally
from the spectra of wind fluctuations. distributed and there is good agre^n.-nt
The pressure transport term is estimated as the between observed and calculated maximum
residual. An example of the height variation for instantaneous wind speed.
each term is shown in Fig. 13 and r'esults are
536
(2) The height dependency of the integral 3) Panofsky, H.A, and I. A. Singer, 1965:
scales obtained in this experiment is shown Vertical structure of turbulence. Quart. J.
to be approximately as follows, Roy. Met. Soc . , Vol. 91, pp. 339~344.
8
L^~z° , l£-z°- 3 L«-z°- 7 4) Davenport, A.G., 1961: The spectrum of hori
(3) The height dependency of the vertical scale zontal gustiness near the ground in high
obtained from the cross correlation winds, quart. J. Roy. Meteor. Soc., 87, 194
coefficient at two different points is 21 .
expressed approximately as follows, 5) Busch, N.E. and H.A. Panofsky, 1968: Recent
0'6
L^-z ,
L^z 0 '^ L^-z 0 ’ 7
Spectra of Atmospheric Turbulence. Quart. J
(4) The spectral scale obtained by using the Roy. Meteor. Soc., Vol. 94.
spectral peak and the mean wind speed can 6) McBean, G. A . and J.A. Elliott, 1975: The
(5)
be expressed by the following equations, Vertical Transports of Kinetic Energy by Tur
bulence and Pressure in the Boundary Layer.

(z < 50 m), A A A - z (z > 50 m)
u w
V V
,
v ,
z 3 J. Atmos. Sci., 32, 753~ 766
In the energy balance equation, it is found
that the sum of the mechanical production
and buoyant production is nearly equal to
the rate of dissipation and that the diver-
gence of pressure transport and the divei
gence of turbulent energy flux sum to about
zero.
6. REFERENCES
1) Naito, K., T. Hanafusa, and T. Fujitani, 1983:
On spatial structures of high winds. 15th
Joint Meeting, U.S. Japan Panel on Wind and
Seismic Effects, UJNR, Tsukuba.
2) Hanafusa, T., T. Fujitani, 1981: Characteris-
tics of high winds observed from a 200 m
meteorological tower at Tsukuba Science City.
Papers in Meteorology and Geophysics, Vol. 32,
No. 1, pp. 19-35.
557
Table 1. Comparison the observed maximum wind speeds with calculated ones
^\^Height
200m 150m 100m 50m 25m 10m
Term'\^
U 16.39 14.39 12. 82 10.39 9.44 7.26
°u 1.95 2.14 2.08 2.18 1.83 1.78

O /U
u
0.119 0.149 0.162 0.210 0.194 0.245
MW , 21.70 19.26 17.59 15.15 15.43 11.75
obs
MW 21.07 19.54 17.80 15.63 13.84 11.53
cal,
U 16.92 15.67 14.62 11.61 10.44 7.99

a 2.10 2.00 2.33 2.53 2.41 2.08
u
a /u
u 0.124 0.128 0.159 0.218 0.231 0.260
MW t 22.30 21.07 21.22 18.21 16.53 13.43
obs
MW _ 21.96 20.48 20.20 17.68 16.23 12.98
cal
u 16.85 15.26 14.32 10.73 8.85 6.59
°u 1.99 2.23 2.59 2.60 2.28 2.07

a /u 0.118 0.146 0.181 0.242 0.258 0.305
u
MW t 20.58 19.25 18.54 17.28 16.39 13.19
obs
MW , 21.62 20.61 20.54 16.96 14.33 11.41
cal
538
Table 2. Dependency of the cross correlation coefficients
on the averaging time
U U U u U U U u U U
20(n- 150 15Cn> 100 100^ 50 5CKy. 25 25^ 10
0.2 sec 0.62 0.54 0.38 0.42 0.50
0.4 sec 0.63 0.55 0.37 0.43 0.52
0.8 sec 0.65 0.56 0.40 0.45 0.55
1.6 sec 0.67 0.58 0.42 0.47 0.59
3.0 sec 0.69 0.61 0.45 0.50 0.62
5.0 sec 0.73 0.63 0.47 0.54 0.66
10.0 sec 0.78 0.68 0.54 0.62 0.73
20.0 sec 0.84 0.72 0.59 0.68 0.82
60.0 sec 0.90 0.77 0.64 0.79 0.88
^\C.C. v <
Vv V 'Vv v 'Vv v 'W V 'Vv
A.T\^'\ 200 150 150 100 100 50 50 25 25 10
0.2 sec 0.40 0.37 0.20 0.24 0.32
0.4 sec 0.41 0.38 0.21 0.25 0.33
0.8 sec 0.42 0.40 0.22 0.26 0.36
1.6 sec 0.45 0.42 0.24 0.29 0.40
3.0 sec 0.49 0.46 0.27 0.33 0.47
5.0 sec 0.51 0.50 0.31 0.38 0.55
10.0 sec 0.60 0.57 0.38 0.43 0.67
20.0 sec 0.66 0.60 0.49 0.44 0.73
60.0 sec 0.74 0.68 0.68 0.48 0.84
W '"V'W W r
V/W W 'X'W W '"Vaj W r
VA*7
'â7tT^£^ 200 150 150 100 100 50 50 25 25 10
0.2 sec 0.44 0.37 0.23 0.31 0.31
0.4 sec 0.46 0.39 0.25 0.33 0.34
0.8 sec 0.49 0.42 0.27 0.36 0.39
1.6 sec 0.54 0.46 0.31 0.41 0.46
3.0 sec 0.60 0.54 0.36 0.46 0.52
5.0 sec 0.64 0.59 0.43 0.51 0.58
10.0 sec 0.70 0.64 0.48 0.61 0.65
20.0 sec 0.76 0.67 0.65 0.62 0.70
60.0 sec 0.83 0.78 0.82 0.69 0.78
C.C.: Cross correlation coefficient

A.T.: Averaging Time
539
Table 3. Lag time when the cross correlation coefficient is maximum
Mean wind
200-150 150-100 100-50 50-25 25-10
speed at 10m
Run 1 u 0.4 8.0 8.0 0.4 2.6

V 4.0 4.8 6.4 2.8 1.8 7.26
w -0.4 0.2 -1.2 0.2 0.4
Au(m/ s) 2.01 1.57 2.43 0.95 2.18 7.26
Run 2 u 2.2 2.2 4.0 0.0 0.8

V 4.0 2.0 4.0 2.2 2.6 8.00
w 0.6 0.8 0.0 0.2 0.0
Au (m/s) 1.23 1.04 3.02 1.17 2.44 8.00
Run 3 u 2.0 4.8 8.0 4.0 0.6
V 2.4 3.2 4.0 2.4 2.4 6.59
w 0.2 0.0 -0.6 0.2 0.0
Au(m/s) 1.59 0.93 3.60 1.88 2.26 6.59
Unit; wind speed: m/sec

time: sec
Table 4. Observed values of each term iii the energy balance equation
''-'-Height (m)
Run No. 173 122 71 35 16
Term
M 0.032 0.025 0.089 0.089 0.232
B 0.003 0.004 0.006 0.006 0.005
1 T -0.023 -0.020 0.026 0.054 0.125

e 0.011 0.016 0.046 0.109 0.206
(P) 0.047 0.033 0.023 -0.068 -0.094
M 0.050 0.038 0.156 0.204 0.368

B 0.007 0.008 0.008 0.006 0.007
2 T -0.015 -0.004 0.022 0.168 0.154
e 0.010 0.017 0.054 0.154 0.276
(P) 0.072 0.033 0.088 -0.112 -0.055
M 0.027 0.018 0.069 0.070 0.150

B 0.004 0.004 0.006 0.006 0.006
3 T 0.001 -0.020 0.042 0.032 0.035

e 0.014 0.021 0.063 0.103 0.158
(P) 0.016 0.021 -0.009 -0.059 -0.037
M 0.023 0.012 0.065 0.072 0.175

B 0.004 0.005 0.006 0.007 0.006
4 T -0.007 -0.024 0.028 0.030 0.136

£ 0.012 0.019 0.050 0.103 0.241
(P) 0.022 0.044 -0.007 -0.054 -0.194
M 0.025 0.019 0.096 0.146 0.274

B 0.002 0.003 0.004 0.005 0.003
5 T 0.003 -0.016 0.024 0.129 0.059
£ 0.010 0.018 0.056 0.135 0.220
(P) 0.014 0.020 0.020 -0.113 -0.002
M 0.023 0.016 0.048 0.062 0.116

B 0.000 0.000 0.001 0.001 -0.001
6 T 0.013 -0.008 -0.002 0.034 0.029
£ 0.009 0.016 0.050 0.115 0.193
(P) 0.001 0.008 0.001 -0.086 -0.107
541
Fig. 3 Normal frequency distribution (the
shadow part indicates 1/120 (5=/600))
#
a
£
Fig. 1 An example of the frequency of wind

fluctuations
Ku tosis
+4 0
U comp
10
Skewness
Fig. 4 Comparison of observed maximum wind

Kurtosis speeds with calculated ones
j40
V comp
10
Skewness
Kurlosts
40f
•V comp
-10 10
Skewness
Fig. 5 Height distribution of mean wind t
(M) and maximum wind speed (G)
-*-20
Fig. An example of the frequency of skewness

and kurtosis
542
Fig. 8 Dependency of the cross correlation
coefficient on the averaging time
Fig. 6 Auto correlation of wind fluctuation
Fig. 9 The relation among the cross correlation

coefficients at different levels
543
Fig. 10 Relation between the cross correlation Fig. 11 Examples of spectra >f wl
coefficient and cj>(z) f luctuat ions
5W
2000
200
100
£
500 - ^ 50
'e
25
0)
a
o
l/l
10
100 -
0.01 0.1
E ( m2 /s 3 )
50 - Fig. 14 Height variation of the dissipation

rate
10 50 100 200
Height (m)
Fig. 12 Height variation of the spectral scale,
3
10 In^/s )
200f~
’
100 r
50-
(m
Heighl
10 -
i
-03
Fig. 13 Height variation of turbulent energy
96
*
546
SEISMIC VULNERABILITY OF LIFELINE SYSTEM USING
COMPUTERVISION CADDS COLOR GRAPHICS SOFTWARE
by
C. V. Chelapati*
Stanley K. Takahashi**
ABSTRACT Color graphics are an excellent way of visually

inspecting the status of each utility line and
Seismic vulnerability of lifeline utility sys- the consequences of damage to utility lines in
tems are highly 'graphics' intensive. Several the event of an earthquake. This pilot study
maps involving the layouts of utility systems, shows that it is efficient in studying the
site plans, soil parameters, fault zones and problem associated with seismic vulnerability of
buildings are involved in the study. The efforts lifeline systems.
involved in the analysis part is relatively sim- Keywords: Seismic, Lifelines, Computer graphics,
ple when compared to the management of extensive Computervision, CADDS, Damage crite-
'graphics' data and reporting. ria, Strains, Pipes, Faults
In this pilot study, the color graphics soft-

1. INTRODUCTION
ware, "CADDS", developed by "Computervision"
Company is used to develop the Seismic Vulnera- 1.1 Background
bility Model (SVM) of a lifeline utility system The United States Navy has several installations
at a typical Navy facility. This model displays throughout the world on the CircumPaci f ic belt
in color the land base map, including site out- where the facilities could be subjected to po-
line, roads, runways, buildings, utility system tentially strong earthquake motions. The need
layout, fault break, liquefaction and accelera- for basic operations at any installation to pro-
tion zones. Pertinent information regarding ceed with minimal disruptions after an earth-
each utility segment is encoded onto the SVM as quake is obvious. The operations of a base de-
"Intelligent" data on each pipe. pend not only on the safety of the buildings but
An earthquake is simulated on an assumed fault also on the lifeline utilities serving the
break, and the strain levels in each pipe is building and its functions in the building. Most
computed using "NEWVAR"; a programming language lifeline utilities should remain operational or
for Computervision applications. A damage be repaired with minimum delays.
criteria is established for failure/ breakage of Most utilities have a source, distribution net-
the pipes as a function of strain in the pipes, work, collection and disposal points. They serve
acceleration due to shaking and the susceptibil- various buildings on a Naval base, and the phys-
ity of the liquefaction zone. Using the ical characteristics of components in a utility
'Boolean' operations, a damage state for each line frequently changes. They may pass through
segment of the wastewater system is estimated. different soil zones. Thus, it is very difficult
This information is displayed in various ways either to describe or process the utility system
using color and the Area Information Management information without appropriate maps. The recent
(AIM) software. advances in computer graphics and Computer Aided
Each building is encoded with the designation of Design and Drafting System (CADDS) software
pipeline segments that serve its functions. A make it more convenient now to investigate and
comparison of these with the damaged pipeline study the seismic vulnerability of lifelinp
segments, will identify the buildings that will uti 1 i ties.
not be served due to the occurrence of an earth- This report utilizes the North Island Naval Air
quake. This information is presented using a Station (NINAS) as a pilot project to study the
color display monitor and by generating a report seismic vulnerability of lifeline systems. The
using the Data Extract/Data Merge software of seismic vulnerability of buildings at NINAS was
Computervision. presented in Reference 1. The soil studies and
the liquefaction potential at the NINAS site was
*California State University Long Beach
**Now with L. I. Dimmick Corp., Oxnard, CA
presented in Reference 2. The Seismic Vulnera- be damaged during an earthquake. The resulting
bility Model (SVM) will eventually include all data will be merged into the graphics database.
buildings, roads, airfields, wharfs, piers, soil The buildings that will be denied utility ser-
conditions, liquefaction zones, fault traces, vices due to potential failure will be identi-
groundwater profile, zones that might be sub- fied; information will then be merged into the
jected to flooding in case of tsunami, all graphic database and displayed.
lifeline systems that serve the facility, loca-
tion of emergency supply measures such as water 1 • 2 Ob j ective
storage tanks, power generators, portable gas The objective of this study is to develop
and all other items necessary for functioning of seismic vulnerability lifeline system model
the base. The SVM can also capture the data on using computer color graphics and the Computer-
the damage potential to buildings in case of vision CADDS software.
potential earthquakes. At NINAS, the site
selected for this study, the following lifeline 1 . 3 Scope of Work
utility systems are identified: In Phase I of this study, the methodology to
a) Fresh Water analyze the lifeline systems will be developed

b) Sanitary Sewer and applied to the sanitary system. The metho-
c) Industrial Waste dology developed will be applied to other
d) Potable Water between NINAS and utilities during the next phase.
City of Coronado
e) Salt Water 2. SITE DESCRI PTION
f) Steam/Condensate Distribution
Compressed Air Distribution 2 . 1 North Island Naval Air Station (NIN AS)
g)
h) Oily Waste - Recovered Oil NINAS occupies about 2,580 acres of land and
i) Natural Gas Distribution System constitutes a major portion of Coronado Island,
j) Telephone System which is situated between San Diego Bay and the
k) Electrical Power System Pacific Ocean. Portions of the Naval Air
Utility systems a) through i) have pipes with Station are hydraulically filled, reclaimed
either gravity flow or are under pressure in areas. Two areas called Spanish Bight (Region
their distribution system. In the telephone or I) and Whalers Bight (Region II) are prone to
electrical system, the distribution system is liquefaction (Figure 1).
cable and/or circuits with equipment at various This region is affected by four major fault
nodes. Thus, the main features and the vulnera- systems (Figure 2). These are offshore the San
bility analysis of these two utilities are quite Clemente, Elsinore, San Jacinto and San Andreas
different from other utility systems. faults. There is a 25 year, 90% probability of
A graphic database for the SVM containing most not exceeding peak ground acceleration levels of
of the above information is to be established 0.05g, 0.13g, 0.06g and 0.07g due to earthquake
using the Computervision CADDS System. A magnitudes of 5.9, 6.9, 6.6 and 7.3 occurring
seismic vulnerability criteria for lifelines respectively from the four fault systems (Ref.
should be based on the site characteristics 1). From past experience with earthquakes,
which might have an adverse effect on the these acceleration levels are too low to cause
utility pipes, the physical characteristics of any disruptions in the utility systems; however,
the pipes and the potential for earthquake the Spanish Bight Fault is assumed to run
hazard at the site. Vulnerability analysis on through the NINAS (Reference 3), and the length
this SVM will be performed using the established of the fault is estimated to be 10 miles. There
criteria to locate the pipes that might break or
548
is no recorded history of earthquakes occurring 4 . USE OF COMPUTERGRAPHICS FOR ANALYSIS
on this fault. Since this assumed fault passes For their analysis and design, engineers exten-
through the site, a sharp high acceleration jolt sively use computations, figures, drawings and
confined to a small area could occur and rupture reports. Therefore, the ability to access
utility lines. Further, the assumed Spanish information from a graphic database and to
Bight Fault runs through the Spanish Bight area display the results in graphic form provides
which has a potential for liquefaction. Utility tremendous advantage. In addition, with the aid
lines which serve many buildings traverse this of color, visualization is very effective. The
area. current trend of practice is to capture graphic

In this pilot study to develop a methodology for information in the form of drawings. Storing
3.
the analysis of Seismic Vulnerability Model, the the drawings on paper for retrieval and updating
Spanish Rose Fault break, high acceleration purposes is very cumbersome and time consuming.
levels and liquefaction are assumed and this is However, when this information is captured on a
a very unlikely event. graphic database, it is extremely versatile and

efficient for many engineering uses. Computer-
UTILITY INFORMATION graphics is extensively used in aerospace appli-
Information on all utilities was gathered by cations and this segment of computer industry is
visiting the Public Works Office and other destined to be a multi-billion dollar industry.
appropriate administrative offices at the NINAS The US Navy has acquired several Computervision
site. The information is in the form of several "CV" workstations and has begun to capture much
sheets of detailed drawings for each utility. information on Computergraphics. One of the
Site visits were also made to observe the areas which has extensive application is Auto-
utility locations. mated Mapping and Facility Management (AM/FM).
From the site plan and other available informa- Aerial photography and stereo mapping is used to
tion, the outline of the island, roads, aircraft capture the "footprint" of an installation. This
runways, liquefaction zones, fault rupture due information is then converted using Automated
to a hypothetical event on the Spanish Bight Mapping procedures into a "Computervision"
Fault, estimated acceleration zones and "foot- graphics database and is used for facility
prints" of most important buildings are extract- management and updates (Ref. 4). Computer-
ed and a graphic database for the SVM is graphics are widely used in many other areas of
established using the Computervision CADDS application including mechanical design, finite
software. From the detailed utility maps, the element analysis, numerical control, architec-
needed information on the sanitary pipe elements ture/engineering/construction, piping design,
is extracted and transferred to a key map on a printed circuit and wiring diagrams to name a
Size E (34"x44") drawing. For this study, the few. Since this field is new, several applica-
key map for sanitary system provides the dia- tions are constantly emerging and better soft-
meter of the pipe, pump stations, lift stations, ware packages are being written. For this
type of material and direction of flow from study, extensive use of Computervision CADDS
collection of sewage from various buildings to a Color Graphics System is made to study the
main collection point to be treated by the City Seismic Vulnerability of a lifeline utility
of Coronado Sewage System. Then, from the key system at a Naval installation.
map of the sanitary sewer utility system, the
layout of the pipes and other elements are 5. COMPUTERVISION WORKSTATION DESCRIPTION
entered manually into the graphics database. In The Computervision "CV" graphics computer system
addition, the "intelligent" information on each consists of hardware and software. The hardware
pipe is coded into graphic database separately. consists of Computervision Graphics Processor
549
Unit (GPU) with several microprocessors (to do data inserted into various graphic entities
graphic functions), a tape drive, disk drive and making up the map image (lines, strings, nodal
a workstation (Figure 3). A workstation consists lines, symbols and polygon nodes). Each of
of display monitor, digitizing tablet, tablet these coded entities serves as a "hook" for
menu, digitizing pen, image control unit and posting or reporting geographically distributed
hard copy graphics printer. Computervision has A/N data "through" the map. Pertinent feature
developed its own graphics language. There are codes can be added manually at a graphic work-
several thousand commands; each command is station. More important, however, A/N data
structured with a verb, noun and modifiers as records resident on any computer can be merged
necessary (Fig. 4). The input needed is sup- into the map. This can greatly expedite the
plied via the keyboard or digitizer. The Com- coding of utility systems if there are already
putervision workstation has two levels; one is pertinent computer data files. If an Automated
an Operating System (OS) level and the other is Mapping/Faci lty Management (AM/FM) system is
Computer Aided Design Drafting System (CADDS) already in place or is being established for a
level. Different functions are performed at Naval facility, the SVM can be included by
different levels. adding additional feature codes and map layers

The Computervision supports several programming to those essential to a full AM/FM system.
languages which are used for different environ-

6 . 3 Database Management (DBM) Components
ments. Some of the languages are FORTRAN,
NEWVAR and VARPR02. The following is a description of the system's
Although Computervision is extensively used for individual components of SVM.
graphics intensive applications, the use of • Land Base
CADDS for Computer Aided Engineering (CAE) is • Area Information Management

relatively new. The present application of Appl ications
CADDS software for the Seismic Vulnerability • Utility System Network

Analysis of Lifeline Systems is a new develop- • Report Extraction and Merge
ment (Figure 5). Each of these components are discussed In
general detail.
6. SEISMIC VULNERABILITY MODEL (SVM)
6 . 4 Land Base Map
6 . 1 Database Management (DBM) Schema
The graphic map for NINAS was prepared by cali-
Figure 6 shows the overall Database Management brating the E size "paper" base map to a
configuration. Central to this schema is a Computervision Instaview Digitizing Tablet.
"virtual map". This virtual map is a computer Control points were defined at the four corners
generated image of the NINAS "physical map". The of the map based on a local grid. The base map
physical map is a conventional two-dimensional elements include:
"Paper Map" limited to displaying lines, surface • Runway Edges

textures, symbols and graphics text. The virtu- • Road Center Lines
al map is enhanced with geo-coded nongraphic • Building Footprints

attribute data (feature codes) associated with • Property Boundary
individual graphic elements. • Approximate Location of Earthquake

Fault
6 2 . Graphic/Nongraphic Interface • Liquefaction Zones
Nongraphic attribute data resides in the virtual • Approximate Earthquake Acceleration
map as "property values" — Alpha/Numeric (A/N) Zones
550
Each of these elements reside on different quake impact conditions. Each subpolygon carries
layers; there are 256 separate layers available automatically generated sets of feature codes
for use. describing the combination of conditions

The land base map is a fair representation of present.
the physical features of the NINAS. However, 2) Polygon CrossHatchi ng. Polygons were
the precision of this automated map is limited crosshatched in various colors according to the
by the precision of the originally drafted lines findings of Boolean operations searching out
that make up the "paper" map. A far more precise specific conditions or combination of conditions
and practical approach to "capture" this data is present. These include:
the direct conversion of digital photogrammetri- • Presence of liquefaction conditions
cal data to a Computervision Graphic Database. • Intensity of earthquake acceleration
This technique has recently been proven to be potential
practical in a pilot project being conducted by • Varied combinations of acceleration and
the Mare Island Naval Shipyard (Ref. 4). It is 1 iquefaction
not only possible to directly derive highly pre- • Building subject to interruption of
cise representations of building "footprints," service due to failure of specific utility line
pavement edges and other visual physical fea- segment(s)
tures, but the graphic data can be converted to 3) Polygon Projections. Edges of polygons
"polygon" entities without any human interven- were projected in 3D spaces to display relative
tion. "Polygon" entities are used to create, conditions. Any "real" property value placed on
categorize, manipulate and report map proper- the polygon can be used as a basis. Areas so
ties. However, all utility lines and other projected included:

associated components have to be manually • Actual estimated height of buildings
inserted. The photogrammetrical data has far Relative intensity of earthquake impact
more information than is needed for the SVM. potential. In this instance, the projected
Thus, special editing has to be done to create distance was accentuated by a scaling factor;
the base map. i.e., the horizontal acceleration values were
multiplied by a factor of 1,000 so that they
6 5
1)
Area Information Management (AIM)
.
appeared projected 300 to 2,000 feet vertically
The AIM software makes it possible to convert to visually dramatize the relative potential
boundaries of area information critical to impact.
earthquake impact assessment into "intelligent" 4) Data Extraction/Data Merge. Feature
polygon entities. Each polygon has a node that codes automatically placed on polygon nodes
contains feature codes describing conditions (area, perimeter, unique database number desig-
pertinent to the SVM. Area information so nator) or manually added (building heights,
converted to polygons included: liquefaction conditions, etc.) can be extracted
• Building "Footprints" by using the "Extract Definition File" (EOF) and
• Known Liquefaction Zones by generating User Defined reports. Most impor-
• Estimated Limits of Earthquake tant, information that may already reside in a
Acceleration in Various Zones main frame data set (i.e., real property record
Polygon manipulations performed with these data files) may be automatically merged into polygons
enhanced areas include: adding (potentially vast) arrays of data to the
Polygon Overlays. Acceleration zones Virtual Map without human intervention. This
were spatially correlated (intersected) with capability will be described later.
liquefaction zones to derive a mosaic of sub-
areas representing varied combinations of earth-
551
This work to date has shown that the Computer- • Acceleration level at midpoint of pipe
vision AIM software is an appropriate tool for • Liquefaction condition at midpoint of
performing spatial analysis, displaying and pipe
reporting extraction essential to the SMV's DBM • Length of pipe before earthquake
needs. A capability proven possible, but not • Length of pipe after earthquake
employed at this time, is the overlay of the •
"Dummy Value" regarding damage potential
building "footprint" polygons with the earth- of pipe due to earthquake
quake impact polygons to determine varied impact The first two features, segment designator and
potential on a single building. This is impor- sequence number provide a unique address of each
tant when large buildings straddle liquefaction pipe segment critical to report generation and
zones. Reference 1 gives the amount of damage data merge/analysis operations. The sequence
to each building as a function of ground accel- number (an interger property value) makes it
eration at the site. It is also possible, using possible to determine which buildings along the
AIM software, to display the amount of potential same network segment experience service inter-
damage to each building for each level of ruption due to a pipe breaking at the beginning,
earthquake. middle or end of the segment (Figure 7).
6 . 6 Utility System Networks Accelerations of pipe elements were added using
Nodal lines were used to represent piping net- Entity Classification software which automati-
works to utilize their "rubberbanding" capabili- cally adds feature codes associated with polygon
ties. Line ends are connected to "Connect to any components origin present inside the
Nodes" (CNODES). It is possible to manipulate polygon's edges.
the locations of the CNODES, and regardless of The before and after lengths were automatically
CNODE displacements, the nodal line(s) will added as feature codes using a macro (NEWVAR
"snap" to their new end position. Since the language) written specially for the SVM study
length of nodal lines can be extracted to a effort.
thousandth of a foot, strain in pipes induced by Pipe condition data (type, year installed,
earthquake generated ground movement can be diameter) is manually encoded. However, this
computed. data, if already available in A/N form on a
To best represent the before and after condi- mainframe, can automatically be merged into the
tions of these pipes, their length was broken by pipes.
a CNODE at each vertex represented in the Data capture cost in digitization of data, is
utility map from which they were digitized. not prohibitive but will require substantial
Significant distribution collection system amount of work hours to digitize and code each
facilities were also defined with nodal symbols utility line. This study proved that nodal
(pump stations, lift stations, etc.). Feature lines can be used to successfully measure pre-
codes pertinent to the SVM were added to the cise before and after conditions of utility
pipes including: networks.

• Unique network segment designation A key management consideration is that the SVM
• Sequence of pipe in network segment could very well be another dimension of a full
starting at furthest distance from central plant AM/FM system. The graphic and nongraphic DBM
(sewage treatment, central supply point, water, requirements of both are similar. Based on the
gas, electrical) or point of generation (com- lessons learned from the Mare Island Naval Ship-
pressed air, steam, other) yard AM/FM system prototypical concept model
• Type of pipe material (Ref. 4), it is reasonable to say that a com-
• Year pipe was installed plete SVM represents costs which could have been
552
reduced if a full AM/FM system were already 7 . RESULTS
available. As such, the SVM cost would be less Using the Computervision "CADDS" workstation,
expensive if performed in conjunction with a all the needed information for the seismic
full AM/FM system. This would save time in analysis from the key is entered into the
capturing the raw data of Land Base Map from graphics database. Figure 8 shows a reproduc-
drawings. tion of some of the features of the SVM. Differ-
ent types of information is on different layers
6 . 7 Report Extraction and Merge so that the needed information can be displayed
As previously discussed, the virtual map is in a variety of ways.
designed with an array of graphic hooks for The seismic vulnerability criteria for the
extracting and merging A/N data from and to the lifeline utility system is developed as follows.
map. This capability is supported by existing In many cases of actual earthquakes, the damage
CADDS 4/4x Data Extract/Data Merge software. to underground utilities is mostly limited to
These "instruments" set the property values areas near the fault break zone.
(feature codes) to be included in and the for- There are three basic parameters associated with
matting parameters of the desired reports. the rupture of pipes: excessive strains caused
Report extract/merge operations can be performed by the rupture of potential fault, liquefaction
for local use (within the C V system exclusively) and the level of acceleration due to shaking. A
or be used as a mainframe (via modem) and other potential fault rupture scenario was created on
computers in the communication network. Like- the assumed Spanish Bight Fault. Using the
wise, the "merge" capabilities of the data ex- coordinates of each node point which resides in
tract software makes it possible to automatical- the graphics database, the length of each pipe
ly (and regularly) update the "virtual map" with before the earthquake was computed using a
information encoded exclusively at the "Macro" written in "NEWVAR" language. The

mai nframe. deformed length after the fault rupture is
The data extract form prepared for the SVM is computed and the resulting strain in each pipe
designed to: is determined. From this explicit strain, a
• Extract all information in a utility rating of 1 to 5 is given (5 being the most

network pertinent to the impact -assessment. severe to 1 being a very low strain). The
• This data is used as a source file by explicit acceleration levels are converted to
the NEWVAR program that processes this informa- Zone 1 being the lowest level of acceleration
tion to determine the estimated pipe strain and Zone 5 being the highest. The graphics
(deformation), strain level and the damage database also contains liquefaction zones.
rati ng. Table 1 shows a utility damage level interaction
• Once these findings are generated, these matrix as a function of liquefaction and accel-
pertinent fields in the EDF are overwritten. eration zones. If a pipe is located in the
• The data extract file is "merged" into liquefaction zone and subjected to high and long
the pipe nodes in the virtual map. duration acceleration, then it is assumed that
• The pipes meeting a user defined damaged the soil support for pipes is weakened, and high
rating are then identified with a "Mark Ent damage to pipes occur. Table 2 shows the utility
Cond" Command. damage level interaction matrix as a function of
• The extract is then run again. The list the acceleration zone and strain level. A
of pipe addresses are the unique addresses of subjective judgement criteria is shown where the
the damaged pipes. damage level will be highest with increasing
• The buildings that have interrupted strains and increasing levels of acceleration.
services due to a damaged pipe(s) are then
marked (or cross hatched).
553
From the three variables, a damage level to each 9. RECOMMENDATIONS
pipe is determined. This information can be It is highly recommend that the remaining utility
displayed in various forms and in color to aid lines be included in the Seismic Vulnerability
visualization. Thus, one could easily identify Model study so that a complete system can be
the "damaged" or "ruptured" pipes. studied for the vulnerability of lifeline facil-
Each building is coded with all the pipes in the ities at NINAS.
utility that are necessary for performing the It is also recommended that the complete AM/FM
designated functions. For this report, this is system be created for NINAS so that the manage-
a network/branch of pipes that carry the ment can study other functions (such as mainte-
sanitary sewage from a building to the main nance and facilties management) in addition to
trunks that lead to the City of Coronado Sewage the lifeline vulnerability of utility system.
System. When the complete AM/FM/SVM is created, it would

By comparing this information with the list of be very easy to update this system on a regulr
damaged pipes, the buildings that will not be basis. The computer management of facilties
served by the damaged lifelines can be displayed would not only be highly accurate but it will
visually. Figure 9 shows an assumed earthquake also provide the capabilty of fast information
fault, partial area where the sanitary sewer retrieval and manipulation.
pipes are damaged and the buildings that will
not be serviced by the damaged utility. 10. ACKNOWLEDGMENTS
Since all the information is in computer The authors wish to thank Mr. J. V. Tyrrell
graphics format, it is very convenient to study (NAVFAC, Code 04BA) for his continual financial
8.
the situation in case of emergency and to devise and technical support and Mr. Dennis H. Klein
alternate plans expeditiously. Since time is of and Mr. Hugh Williams (AM/FM consultants) for
essence, in case of a potential earthquake, the their participation in the development of the
graphical information will be extremely useful Seismic Vulnerability Model. Appreciation is
as an aid in bringing up the facilities to a also extended to Mr. Jerome H. Hopkins of NCEL
normal functioning level. for providing Computervision system assistance
Table 3 summarizes a typical output of the during very crucial times.
information stored by Computervision. Addition-
11. REFERENCES
al information on the SVM may be obtained from
Reference 5. 1. Naval Civil Engineering Laboratory, Techni-
cal Report M51-78-08: "Earthquake Hazard Reduc-
SUMMARY AND CONCLUSIONS tion Program, North Island Naval Air Station,
This pilot study shows that the emerging San Diego, Calif." by C. V. Chelapati, S. K.
computer graphics and Computervi sions CADDS Takahashi and T. K. Lew, Port Hueneme, Calif ,
software can be used with great advantage in April, 1978.

conducting the Seismic Vulnerability of Lifeline 2. Naval Civil Engineering Laboratory, Techni-
Utility Systems. The development of the Seismic cal Report R847: "An Earthquake Analysis of
Vulnerability Model (SVM) will provide tools to the Liquefaction Potential at the Naval Air
predict damage in case of an earthquake. By Station, North Island," by J. Forrest and J.
studying several scenarios, intelligent deci- Ferritto, Port Hueneme, Calif., Sept., 1976
sions can be made to reduce or to mitigate 3. "Report of Geotechnical Studies Investiga-
potential damage. In case of an earthquake, the tion/Evaluation of Seismic Vulnerability of
SVM will provide instant visualization of Structures, Naval Ocean Systems Center and Naval
damaged conditions to utility networks and can Communications Station, San Diego, CA for the
be used as a tool to make decisions to alleviate United States," By Leroy Crandall & Associates
the damaged state. Contract No. 247483C2466, Dec., 1984
554
4. "Mare Island Naval Shipyard Automated EDF Extract Definition File
Mapping/Facility Management System Prototypical FORTRAN Formula Translation Programming
Concept Model," by Dennis Klein & Associates, Language used mostly for scien-
Mill Valley, Calif. ,
March, 1985. tific and engineering applica-
5. "Seismic Vulnerabilty of Lifeline Systems tions
Using Computervision CADDS Color Graphics Soft- GPU Graphical Processer Unit
ware - Phase I" (In preparation), by C.V. Macro A series of CADDS commands group-
Chelapati Huntington Beach, Calif., 1985. ed together to perform a certain
function
12. DEFINITIONS NEWVAR An interactive programming langu-
age to generate CADDS commands
AIM Area Information Management NINAS North Island Naval Air Station
AM/FM Automated Mapping and Facility OS Operating System
Management Polygon Entity used to create categories,
A/N Alpha/Numeric manipulate and report map
CAD/CAM Computer Aided (Design)/Computer properti es
Aided Manufacturing SVM Seismic Vulnerability Model
CADDS Computer Aided Design and Drafting VARPR02 A Computervision CAD/CAM Language
System Virtual map Computer generated image of a
CAE Computer Aided Engineering physical map

CV Computervisi on
Table 1. Utility Damag e Level Interaction Matrix,

Acceleration Zone/ Liquefaction
Acceleration Zone
Liquefaction
1 2 3 4 5
No
0 0 0 0 0 0
Yes
1 0 0 5 5 5
Table 2. Utility Damage Level Interaction Matrix,

Acceleration/Strain
Acceleration Zone
Strain Level
1 2 3 4 5
1 0 0 0 0 0
2 0 0 0 1 2
3 0 0 1 2 3
4 0 1 2 3 4
5 1 2 3 4 5
555
Table 3. Typical Report Output from Computervision
North Island Naval Air Station Automated Earthquake Assessment Model

Using Coraputergraphics-Computervision CADDS System
Siesmic Hazard Evaluation of Utility Systems
MM-DD-YY Page ////
5- 1-85 1
A B C D E F G H I J K
System Matl Strain Quake Accel Liq Length Length Damage
Segm Type Strain Level Accel Ratg Ratg Before After Rating
A1 1 TC 0.773 5 0.20 3 Y 129.2250 129.2250

A2 1 TC -0.658 5 0.20 3 Y 108.7710 108.7700
A3 1 VC -0.274 5 0.20 3 Y 66.5170 66.5150
A3 2 VC 0.842 5 0.20 3 Y 122.5180 122.5180
A3 3 VC -0.428 5 0.20 3 Y 70.0410 70.0410
A3 4 VC 0.149 5 0.20 3 Y 50.8090 50.8100
A352 Cl -0.244 5 0.20 3
'
153.9990 154.0000
A353 TC 9.258 5 0.20 3
'
203.7180 201.4730
A4 1 VC -0.884 5 0.20 3 Y 45.9670 27.6560
A4 2 VC 1.934 5 0.20 3 Y 144.8800 134.8700
A4 3 VC -0.731 5 0.20 3 Y 79.3960 79.3940
A4 4 VC 2.001 5 0.20 3 Y 238.2710 238.2710
A8 2 TC -0.330 5 0.20 3 Y 72.8900 72.8890
A8 3 TC -0.776 5 0.20 3 Y 91.6050 91.6040
A8 4 TC -0.809 5 0.20 3 Y 114.8830 114.8830
A9 1 TC 1.769 5 0.20 3 Y 318.1410 318.1390
B1 1 TC -0.388 5 0.20 3 Y 108.0490 108.0480
B1 1 TC 3.910 5 0.20 3 Y 249.4510 249.4510
B1 2 TC 0.366 5 0.20 3 Y 176.5270 176.5270
B1 3 VC 0.185 5 0.20 3 Y 295.5350 295.5340
B1 4 TC 2.711 5 0.20 3 Y 408.4520 401.0180
E2 1 TC 1.053 5 0.20 3 19.6410 297.5090
1>
Figure 1. Condition map of NAS North Island.
556
Figure 2. Faults used for estimating ground accelerations at
site of NAS, North Island.
PERMANENT STORAGE
GRAPHICS NUMERICAL
CAD/ —> CONTROL
CADDS CAM
Computervision :
Basic
Graphics
Commands
Several Thousands of Commands

Manual 3 in. thick
Commands more like language -- sentences
Verb Noun Modi f iers Data/Digitizer
i * /Explicit
Insert Line Horizontal

—
Figure 3. Computervision graphics CADDS Figure 4. Computer graphics command

work station. language structure.
557
ARRAY OF PIPES SERVING
BUILDING "l"
PROPERTY ROOTS EQ - 2
Figure 7. Coding of utility line and building

schema
Figure 5. Computer aided design and drafting
systems (CADDS).
Figure 8. Snapshot of Computervislon monitor

screen showing NAS, North Island and
portion of sanitary sewer.
Figure 9. Snapshot of Computervi sion monitor screen '
of earthquake fault and damaged and undama 1

r ! I in
558
VERTICAL DISTRIBUTION OF SUBSURFACE
HORIZONTAL EARTHQUAKE ACCELERATIONS
P. C. Chen
Interpacific Technology, Inc.
180 Grand Avenue, Suite 900
Oakland, California 94612
(415) 444-7081
Abstract A total of 192 sets of peak horizontal various sites to investigate their engineering
accelerations of subsurface earthquake motions characteristics. A comprehensive and systematic
recorded at 20 sites in Japan were studied to study was therefore initiated using as many re-
determine the vertical distribution with depth corded subsurface earthquake motions as possible
of peak subsurface horizontal accelerations. at various sites to investigate the vertical
The 20 sites are classified as either soft or distributions of accelerations and response
medium sites depending on the blow count number, spectra of subsurface earthquake motions and the
N, near the ground surface and the average blow effects of site geological conditions and earth-
count number, "N, of the soil layers between the quake parameters on these distributions. This
ground surface and the depth at which N first paper presents some of the results with regard
reaches a value of 50. The 192 sets of peak to the vertical distribution of subsurface hori-
accelerations were separated into five groups zontal earthquake accelerations.
according to different peak ground-surface ac-
celerations. The mean values of peak accelera- 2. Data Base
tions at various depths for each acceleration
group were computed for each site classifica- The subsurface horizontal acceleration data of
tion. The amplification factors of the mean earthquakes with peak horizontal accelerations
peak accelerations between the ground surface at or near the ground surface equal to or larger
and the depth at which N first reaches a value than 10 gals (as reported in References 1 to 16)
of 50 were studied. These amplification factors were studied. All the data were recorded at 20
were compared with the calculated site amplifi- Japanese sites. No U.S. data were used in the
cations using the average normalized accelera- study, as their recorded peak accelerations were
tion response spectra for soft and medium sites. less than 10 gals.
1. Introduction References 1 to 16 contain information on the
site locations; site geological information;

With the recent trend of constructing more deep- instruments used to record the subsurface earth-
ly embedded and buried structures, there is in- quake motion data; earthquake information such
creasing need for better knowledge of the as date of occurrence, magnitude, etc.; and the
engineering characteristics of subsurface earth- peak horizontal accelerations at different
quake motions such as accelerations and response depths of each site caused by different earth-
spectra. Many sites have been instrumented to quakes. The site classification, instrument
record subsurface earthquake motions, either to locations (depths), and the number of earthquake
develop seismic design criteria for structural data sets used in the study are summarized in
facilities at these sites or to observe the pat- Table 1. The earthquake data sets are divided
terns of these motions. Some observed earth- into five groups according to the peak accelera-
quake motions at individual sites have been tion values at or near the ground surface. One
reported; however, no study has been performed set of data contains all horizontal peak ac-
using recorded subsurface earthquake motions at celerations recorded for an earthquake event at
559
different instrument depths of the site. For sites, the instruments are not located at a
example, for Site 1, there are 11 data sets in depth when N first reaches 50; the mean peak
Group (a) and 5 data sets in Group (b) recorded acceleration at this depth is computed from the
at depths of 3m, 21m, 50m, and 90m below the mean peak accelerations at the neighboring
ground surface. instrument locations, again by assuming a linear
relation.
3. Analysis Approaches
The ratios of the mean peak accelerations at

3.1. Site Classification different depths of the same site are the peak
acceleration amplification factors (amplifica-
All the 20 sites are classified as either soft tions factors) calculated for this site. Be-
or medium sites depending on the blow count num- cause these amplification factors were obtained
ber, N, near the ground surface and the average from recorded data, they are called observed
blow count number, N, of the soil layers between amplification factors to differentiate them from
the ground surface and the depth at which N calculated site amplifications. The amplifica-
first reaches a value of 50. If the N value tion factors were studied for three cases: (1)
near the ground surface is less than or equal to amplification factors between the ground surface
5, and N is equal to or less than 10, the site and the depth at which N first reaches a value
is considered soft. A site is classified as of 50; (2) amplification factors between the
medium when its N value near the ground surface ground surface and depths at which N is less
is less than 50 but larger than 5, and N is than 50; and (3) amplification factors below the
larger than 10. Because the depth at which N depth at which N first reaches a value of 50.
first reaches 50 varies from site to site, each Once the amplification factors for each site
site is further classified as either shallow or were examined, they were grouped according to
deep. If this depth is less than 15m, the site the three cases listed above. Of the three
is considered shallow; otherwise, the site is types of amplification factors discussed above,
considered deep. In Japan, the depth at which N only the amplification factors between the
first reaches a value of 50 is generally consid- ground surface and the depth at which N first
ered the base rock for design purposes. The reaches 50 are presented herein.
other factor that suggests the selection of N is
that Japanese bole-hole data normally stop at a 4. Amplification Factors Between Ground Surfacp
depth when N reaches a value of 50. and the Depth at Which N First Reaches $0
The amplification factors of the mean peak ac-
3.2. Data Analysis celerations between the ground surface and the
depth at which N first reaches 50 for both the
The mean values of the peak horizontal accelera- soft and medium sites are shown in Figure 1 and
tions (peak accelerations) at different depths 2, respectively. They are the ratios of the
for each acceleration group of each site were mean peak accelerations at the ground surface to
computed. In order to have a common base, the that at which N first reaches a value of 50.
mean peak accelerations of some sites were For each site, various acceleration groups are
extrapolated to the ground surface (0m); because- identified, together with the site's average
not all sites have instruments located at the blow count number, N. For example, the amplifi-
ground surface. For example, one of the cation factors of various acceleration groups of
instruments for Site 1 is at -3m. The Site 6 are identified with different symbols and
extrapolation is achieved by assuming a linear site numbers. The average blow count number n
relation between 0m and -3m. Also, for many for the site is given in parentheses.
560
The amplification factors for soft and medium amplifications that are largely affected earth-
sites are discussed below. quake parameters and by the depths and the
material properties of soil layers between the
4.1. Soft Sites ground surface and base rock of these sites.
The observed amplification factors are the am-
Observation of Amplification Factors plifications of the mean peak accelerations

4.1.1.
between the ground surface and a depth at which
N first reaches 50. Because these two amplifi-
o Amplification factors are not necessarily
larger for acceleration groups with larger cations have similarities, a comparison was made
of their amplification values. Ideally, site
mean peak accelerations at the ground surface
(Sites 1, 6, 14, 16 and 17). This is because acceleration response spectra and site geologi-
when the peak accelerations at the ground cal data would be used to determine site ampli-
surface are larger, the peak accelerations at fications. However, except for Sites 1 and 9,
depths also are larger. The resulting ampli- site acceleration response spectra were not
fication factors between the ground surface available. The average normalized site ac-
and the depth at which N first reaches 50 can celeration response spectra for soft sites
therefore be smaller compared with cases (Reference 17) were therefore used for calculat-
where the peak accelerations at the ground ing site amplifications.
surface are smaller.

o In Figure 1, a weighted solid line that takes The period of each site was calculated using
into account the number of data sets in each site geological properties between the ground
acceleration group is drawn through the am- surface and the depth at which N first reaches
plification factors of various sites. This 50, and assuming that earthquake waves propagate
line is extended by a dash line to the ground vertically. The calculated periods of all the
surface, because at ground surface, the am- soft sites are shown in Figure 1. They range
plification factor is 1.0. It can be seen from 0.268 second for Site 6 to 1.17 second for
from the solid line that the amplification Site 11. In general, the period increases as
factors for all soft sites fluctuate about a

the depth increases; but there are exceptions
value of 2.5. where the site period decreases when the depth
o The average blow count number, N, of all soft increases and the "N is larger (Sites 1 and 11).
sites except for Site 11 are less than or

equal to 10. The top 25m of Site 11 has a N Once the site periods were calculated, the site
value less than 5, a N value of 10 between amplifications were obtained from the average
the depths of 25m and 37m, and an average N normalized acceleration response spectrum with a
value of 30 between 37m and 56m, at which N 5% damping for soft sites (Figure 3). They are
first reaches a value of 50. Although the marked by small circles in Figure 1, and are
overall average N is larger than 10 for Site connected by a dash-dot line. The portion of
11, it is classified as a deep soft site be- the dash-dot line from the site amplification of
cause the N value of the top 25m is less Site 6 to that of 1.0 (ground surface) was ex-
than 5. tended.
4.1.2. Comparison of Observed Amplification Fac- The calculated site amplifications (dash-dot
tors and Calculated Site Amplifications line) fluctuate about a value of 2.5. The ob-
served amplification factors (solid line) and
The average normalized acceleration response the calculated site amplifications are seen to
spectra at given sites can be considered as site be very close. The calculated site amplifica-
561
tions fluctuate about 2.5 for two reasons: The for soft sites. The calculated periods of
average normalized acceleration response spectrum medium sites are shown in Figure 2. The site
with a 5% damping for soft sites (Figure 3) has amplifications obtained from the average
an amplification factor of 2.5 in the period normalized acceleration response spectrum with a
range of 0.25 to 1.0 second, and the periods of 5% damping for medium sites (Figure 4), using
the sites studied also are mostly in this range. the calculated site periods are marked by small
circles in Figure 2. These small circles are
It should be noted that the data base for devel- connected by a dash-dot line. The portion of
oping the average normalized acceleration the dash-dot line from the site amplification of
response spectra for soft sites is different Site 15 to that of 1.0 (ground surface) was
from that used to obtain the observed amplifica- extended.
tion factors of mean peak accelerations. The
values of IT are unknown for the sites used to An examination of the observed amplification
develop the average normalized acceleration re- factors (solid line) and the calculated site
sponse spectra. In spite of the different data amplifications (dash-dot line) indicate that for
base, the comparison between the two amplifica- deep medium sites, the observed and calculated
tions is good. This suggests that the average amplifications are close. However, there exist
normalized site acceleration response spectra larger differences between observed ampli-
with a 5% damping for soft sites may be used to fication factors and calculated site amplifica-
estimate the amplification factors of the mean tions for shallow medium sites.
peak accelerations between the ground surface
and the depth at which N first reaches 50. The calculated site amplifications were obtained
using the thickness and soil properties of the
4.2. Medium Sites soil layer, measured from the ground surface to
the depth at which N first reaches 50, and the

4.2.1. Observation of Amplification Factors average normalized horizontal response spectra,
assuming that earthquake waves propagate verti-
o Amplification factors are not necessarily cally. The above soil layer thickness for shal-
larger for acceleration groups with larger low medium Sites 15, 10, 13, and 4 is less than
mean peak accelerations at the ground sur- 15m. The periods of these sites are 0.175
face. second, 0.179 second, 0.269 second, and 0.295
o A weighted solid line that takes into account second, respectively. Because these periods
the number of data sets in each acceleration fall in the range at which large site amplifica-
group is drawn through the amplification fac- tions occur (Figure 4), the calculated site
tors of various sites. This line is extended amplifications are therefore large.
by a dash line to the ground surface. The
amplification factors start with 1.0, However, the average blow count number, N, for
gradually increase to about 2.5, and then Sites 15, 10, 13, and 4 varies from 30 to 35
fluctuate about 2.5. (Figure 2). These large N, together with the
small thickness of the soil layer (less than 15m
4.2.2. Comparison of Observed Amplification Fac- for these four sites), would suggest that the
tors and Calculated Site Amplifications
site amplification factors for these sit*»s would
be small when the earthquake wave’ are assumpd
An effort was made to compare the observed am-
to propagate vertically. The dif.erences
plification factors with the calculated site
between the observed amplification factors and
amplifications, using the same approaches taken
the calculated site amplifications for shallow
562
medium sites are probably due to the different 6. Acknowledgment
data base used to obtain the observed amplifi-
cation factors and that used to develop the This study was supported by the U.S. National
average normalized horizontal acceleration Science Foundation under Grant No. CCE-8304066
response spectra for medium sites. Further in- and an in-house research and development program
vestigation of the data base, including the N of Interpacific Technology, Inc. The author is
values of sites used to develop the average nor- indebted to Dr. Y. Ohsaki for his inspiration;
5.
malized horizontal response spectra, will clari- Drs. Y. Osawa, Y. Sakai, T. Okubo, T. Tanaka, T.
fy the above differences. Iwasaki, K. Ishida, Y. Kitagawa, S. Kawamura, K.
Kawashima, T. Ohta, J. Jido, T. Tsunoda, T.
Conclusions Suzuki, M. Nakamura, and Messrs. Y. Saito and
I. Katayama for their assistance and discus-
Observed amplification factors were obtained for sions; and to Drs. H. Tsuchida and T. Morioka,
both soft and medium sites of the mean peak ac- J. Litehiser, Y. K. Lin, and B. Schmidt for
celerations between the ground surface and the their assistance, discussions, and review of the
depth at which N first reaches a value of 50. study results.
For soft sites, the amplification factors fluc-
tuate about a value of 2.5. For deep medium 7. References
sites, the amplification factors are also about
2.5. For shallow medium sites, the amplifica- 1. Aria, H., and Saito, S., "Observation of
tion factors are less than 1.5. Earthquake Motions in Foundation Soil Layers
of Shore Protection Facilities," Technical
The observed amplification factors for both the Note No. 251, December 1977, Port and Harbour
soft and medium sites were compared with the Research Institute, Ministry of Transport,
site amplifications calculated using previously Japan.
published average normalized acceleration re-
sponse spectra with a 5% damping for soft and 2. Yokota, H., "Observation of Underground
medium sites. Although different data bases Earthquake Motions at Shibaura, Minato-ku,
were used to obtain amplification factors and Tokyo," Proc. 7th Symposium on Ground
site amplifications, the comparisons are good Vibrations, A.I.J., March 1979.
for soft sites and deep medium sites.
3. Yokota, H. and Watanabe, H., "Results of

Because of the above observations, it appears Earthquake Observation at Hibiya in Tokyo,"
that the average normalized acceleration re- Proc. 3rd Symposium on Ground Vibrations,
sponse spectrum with a 5% damping may be used A.I.J., Nov. 1974.
for soft sites to estimate the amplification
factors of the mean peak accelerations between 4. Nasu, N. et al., "An Interim Report on the
the ground surface and the depth at which N Results from Underground Earthquake Observa-
first reaches a value of 50. The average nor- tion at the Site of Grand Heights, Narimasu,
malized acceleration response spectrum with a 5% Tokyo." June 1975. Published by Tokyo Soil
damping for medium sites can also be used for Research Co.
this estimate, provided the differences between
the observed amplification factors and the cal- 5. Saito, T. et al. "Underground Observation of
culated site amplifications for shallow medium Earthquake Motion in the Uptown Sections of
sites are accounted for. Tokyo," Taisei Technical Report, No. 6.
565
6. Iwasaki, T. et al., "Observation of Under- in a Soi 1-Structure System," Kenchiku Kenkyu
ground Earthquake Motions Around Tokyo Bay," Shiryo, No. 12, March, 1976, Building Re-
Report No. 1, Public Works Research Insti- search Institute, Ministry of Construction.
tute, Ministry of Construction, March, 1981.
13. Science & Engineering Research Lab., Waseda

7. Private Communication, Earthquake Research University, "Results from Earthquake Observa-
Institute, Tokyo University, 1981. tion at Tatemachi, Hachioji City," Proc. 3rd
Symposium on Ground Vibrations, A.I.J., Nov.,
8. Sugimoto, M. and Abe, Y., "Underground Obser- 1974.
vation of Earthquake Motion in Soft Ground,"

Proc. Annual Meeting, A.I.J., Sept., 1979. 14. Yokota, H., "Observation of Underground
Earthquake Motions at Toyosu," Proc. 3rd
9. Private Communication, Hokkaido University, Symposium on Ground Vibration, A.I.J., Nov.,
1981. 1974.
10. Tsunoda, T. and Seo, K., "An Approach to In- 15. Private Communication, Science & Engineering
put Seismic Waves for Structural Design (Part Research Laboratory, Waseda University, 1982.
2. Observation of Ground Motion)," Proc.
Annual Meeting, A.I.J., Oct. 1973. 16. Private Communication, Kajima Corporation,
1981.
11. Matsushima, Y. and Iwashita, M., "Results
from Earthquake Observation in Kawaguchi 17. Hayashi, S., Tsuchida, H., and Kurata, E.,
City," Proc. 3rd Symposium on Ground Vibra- "Average Response Spectra for Various Subsoil
tions, A.I.J., Nov., 1974. Conditions," Third Joint Meeting, U.S. -Japan
Panel on Wind and Seismic Effects, UJNR,
12. Report on Digitized Earthquake Accelerograms Tokyo, May 10-12, 1971.

s_
CD
cn
CL
rO >5 CD
4-> i
rO ro
"O
Q CC O S-
O' LO O
LU C
O TD
CD
v
e
LO LO
— ro 1952o
r-H
1 CXI LO CX
LO
cxi ro 00 »-H <0- f-H CXJ CXI < — 1
o
f-H
i-
o o
4-
rO
cn
Information
CL
O
ro
O O
oo oo oo
f-H
o X
1
3 f—
CD E
—
1 1 1
LO O LO o
o
X
-M
o
co
1
u c *>
o
cn
oo o
cn
LO
cn
* O CO CXJ
cn
•
CN
LO
1
lo coo
CN CN
CT> f— CL
LO LO LO O LO O o
f 1
ro a;
4— oo CO 1^ cn ro
f-H • LO CXI 1 ro cxi lo 1 1 • CD --
c LO CO LO f-H 1 — LO CXJ CN ro CN CN LO TD
—
CD
o
1 1 1 1 1 1 1 i i i
Data
3 o ^ LO -
r< " f— CN ro Cn Cn cn LO v
O O o LO
<r>
| 1 1 1 1 1 |
S-
V
O LO o
f— f-H ^ LO ro * •* •> CN CN co • ro s»
00 4->
-Q 00 rO
4-> CXI CXI CXI OJ LO ro 1 f-H f-H
LO
1
CN
•
o
ro
CO
f-H
f-H i C0 f-H f-H f—
ro
LO i
4—
o CN
3 C CJ
1 1 1
r
CXI 1 1 *• 1 CXI i
* LO •
1
o
CN CXI 1
CN
1
CN ,
1
CN
1
CN ro
CN
00
O 1 2 7 1 r—
CN •'CXI •* CXI 1 CN 1 CXI
OO O ro ro ro
1 1
*
i i i |
r—
O LO o ro ro *
I
*
O^f-H f-H CXI CXJ f-H f-H cn CN CN 1
CXI f—
o *vt*
oo
1
1 I l l 1 1 1 i 1 I 1 i 1 i i 1 1 1 I 1 1 1 o ro
Earthquake
CD
CO
+->
CD O
LO
CD 00
4-> TD
E
Subsurface
00 3
4-> TD *
4- CD to
O E r—
00 ro
S- cn
o
and O
4-
4- o
LO
O O
LO o 1
Site o LO *3"
to
CL 00 i
CD
1.
x
a;
lo
i—
rO
3
ro
3
cr
—CJ
OO 4— CT CD CN
O CD *
OO
Table ro
s- Or acceleration.
4 O
«o-
SZ
DO
O 4->
o o
CXJ
I
4~>
"e
00 H-
oo
on
’vf
fOMH 00
oo
00
00
cd —
_Q
-
a;
CD C surface
c C -X
-X CJ *
CJ T— .00
-C E »—
r
4->
-p lo ro
«— CD
LO
LO
c\j
C\J
ro
LO LO
LO ro
—
i i
o
ro
O
LO
CO
ro S-
S-
CD C
I
o ground
CD >, ro CXI
>) ro .c
rcj
f—”
r— 4-> o
r-H
I
CD 1
4-> *Peak
ih oj ro ^ lo cd co cn
— o CXJ + —ns
OO « I f-H C\J CXJ + +
565
Amplification Between Om/Depth with N=50
1.0 2.0 3.0 4.o
Figure 1. Amplification Factors for Soft Sites

Legend :
• - x - 20-40, + - 40-60, xx - 60-80, x+ - 80-

10-20, 1 - 0
peak ground surface acceleration of various groups
566
Amplification Between Om/Depth with N=50
Figure 2. Amplification Factors for Medium Sites

Legend
• - 10-20, x - 20-40, + -
40-60, xx - 60-80, x+ - 80-130 gals
peak ground surface acceleration of various groups
567
ratio
Acceleration
Figure 3- Average Response Spectra for Soft Sites
ratio
Acceleration
Figure 4. Average Response Spectra for Medium Sites
568
SHAKING TABLE TEST ON THE EFFECTS OF THICKNESS OF
LIQUEFIED LAYER ON DEFORMATION OF EMBANKMENTS
BY
Toshio IWASAKI ,* Yasushi SASAKI,**

Hideo MATSUMOTO*** and Masuo KONDO****
ABSTRACT
Earthquakes have caused tremendous damages to Niigata Earthquake. Even for low embankments
earth structures such as road embankments and openings and large longitudinal cracks develop-
river dikes. Typical damages of earth struc- ed as the embankments settled. Embankments thus
tures are settlements, openings and crackings damaged were mostly resting on old river beds or
of embankments. Many of these damages to em- lowlands between sand dunes where the ground
bankments occurred in areas where ground layers layers are liable to liquefaction. This was
are liable to liquefaction, namely in old river also the case for the Nihonkaichubu Earthquake
beds and lowlands between sand dunes. of 1983. Thus, it is considered essentially
The amount of seismic deformation of embankments important to assure earthquake resistance of

on liquefied ground layer depends, largely, on earth structures standing on easily liquefied
several factors such as the intensity of earth- grounds
quake motion, the degree of liquefaction and the The amount of seismic deformation of embankments
thickness of the liquefied layer. The authors on liquefied grounds is known to depend largely
of this paper have conducted a series of shaking on the intensity of earthquake motion, the de-
table tests to observe how the thickness of the gree of liquefaction and the thickness of the
liquefied layer affects the amount of seismic liquefied layer. Yet, there are many other fac-
deformation of the embankment. The following tors that remain to be clarified in this connec-
describes the results of these tests. tion.

Under these circumstances, the authors conducted
Keywords: Earthquakes, Embankments, Ground a series of shaking table tests to see how the
Motion, Liquefied Layers, Seismic amount of seismic deformation of an embankment
Deformation, Shaking Table is affected by the thickness of the liquefied
1. INTRODUCTION layer. This paper briefly describes the method

Earth structures, such as road embankments and and the results of the experiments.
river dikes, were frequently damaged in the
Niigata Earthquake of 1964 (magnitude=7 .5) ,
the 2. METHOD OF EXPERIMENTS
Tokachi-oki Earthquake of 1968 (M=7.9), the Vibration tests were carried out, using models
Miyagiken-oki Earthquake of 1978 (M=7.4) and the consisting of a lower subground layer of loam
Nihonkai-chubu Earthquake of 1983 (M=7.7). In which is not liable to liquefaction, an upper
the plains, these earthquakes have caused heavy subground layer of mountain sand which is likely
damages to embankments constructed on loose al- to liquefy during vibration, and an embankment.
luvial sandy grounds or soft clayey grounds; Those models were set up in a container (1.8 m
while in the mountainous areas, high embankments long, 0.6 m wide and 1.1 m high) placed on a
on steep slopes were greatly damaged.
One remarkable damage in the plains is the col- *Director, Earthquake Disaster Prevention Dept.,
lapse of embankments which occurred extensively Public Works Research Institute, Ibaraki-Ken, Japan
in the suburbs of Niigata City during the ** ,
*** ,
and **** Ground Vibration Division, do.
969
shaking table. were applied. In each of them constant sinu-
soidal acceleration with a frequency of 5 Hz was
Six models with different thicknesses of liquefi- employed for 8 seconds (40 cycles). After vanish-
able ground layers were used for the experiments. ment of the pore water pressure was confirmed, a
test with higher acceleration was done. Succes-
2 .1 Ground Models and Embankment Models sive tests were conducted until the embankment
As shown in Fig. 1, each of models had a ground completely failed. The relation between the
thickness of 80 cm, an embankment top width of stage of excitation and the measured maximum ac
20 cm and an embankment height of 15 cm and the celeration of the shaking table was indicated in
side section was symmetrical along the central Table 3 for 6 models. Values at the bottom exci- -
vertical line. tation step for the models, indicate accelera-
tions when failure occurred.
As for the model ground, a loam layer which is
not liable to liquefy was first prepared in a Accelerometers and pore water pressure meters
container, a certain amount of water was poured were installed in the ground and the embankment
and then a liquefiable sand layer was prepared (as illustrated in Fig. 1) to measure vibration
by means of the underwater drop method, using amplitudes and pore water pressure changes. Mesh-
mountain sand taken from Mt . Sengen in Chiba es of white-color sand at intervals of 10 cm were
Prefecture. Upon completion of the model prepared on the side of the model to watch the
ground, the model embankment was prepared, using deformation of the model during and after the
air-dried mountain sand and making the water shaking test. Also displacement gauges and bench
level equal to the surface of the ground. The mark points were seen on the surfaces of the
characteristics of materials used for the exper- ground and the embankment to measure the movement
iments are indicated in Table 1. of the surface.
Regarding model conditions, the same compaction 3. RESULTS OF EXPERIMENTS

condition was maintained for the embankment and 3 . 1 Resonance Test
the ground. Thickness of the liquefiable moun- Fig. 2 shows the resonance curve for the crest of
tain sand layers ranged from 15 to 60 cm. Also the embankment. The ordinate indicates response
the relative density of the actual embankment amplification factor representing the ratio of
models ranged 25-40% and that of the ground 61- acceleration amplitude at the crest of the em-
80% (see Table 2). bankment to that of the shaking table.
2 .2 Shaking Table Tests It can be seen from this that the resonance fre-
In each test, a low-amplitude resonance test was quency of the model embankment is about 17-18 Hz
first made to obtain frequency characteristics and has almost identical vibration characteris-
of the model, and then large-amplitude excita- tics regardless of thickness of liquefiable lay-
tion tests intending to identify failure charac- ers (varying from 15 to 60 cm).
teristics of the models. In the resonance test 3 .2 Failure Test

sinusoidal exceitation with a constant shaking (1) Relation Between Acceleration and Deformation
table acceleration of 10 gal was used and the Fig. 3 shows the relation between the accelera-
excitation frequencies ranged 5 to 40 Hz. tion of the shaking table and the accumulative
settlement of the embankment crest. From this
In the excitation tests for getting failure it can be seen that the accumulative settlement
characteristics, several levels of accelerations increases as table acceleration Increases. It is
570
observed that the accumulative settlement of the zontal displacement was observed up to a depth of
embankment models with layer thickness of 30 cm 30-40 cm from the ground surface. The maximum
or less tends to increase in proportion to the horizontal displacement in the ground was observ-
thickness of the liquefiable layer, when sub- ed not on the ground surface but at a depth of
jected to the same table acceleration. For 20-30 cm from the surface. Also, in the case of
models with layer thickness larger than 30 cm, A-l, embankment settlement was caused by the
however, accumulative settlement nearly equals movement of embankment slope toward the toe of
regardless of layer thickness. Fig. 4 shows the slope. In the case of A-5, however, the entire
relationship between the response acceleration embankment settled uniformly into the ground.
of the ground just below the embankment and the Photographs of failure patterns of each model
accumulative settlement of the embankment crest. after the final stage when the excitation were
Similarly to the fact seen in Fig. 3, accumula- finished, are shown in Photos 1-6.
tive settlement tends to increase in proportion (2) Relation Between Excessive Pore Water Pres-
to the thickness of the liquefiable layer. For sure and Embankment Deformation
these experiments where accumulative settlements Fig. 6 shows settlements of the embankment crest
are only about 7 cm, characteristics of accumu- for each model at the excitation stage 5 (shaking
lative settlement may be divided into two cate- table acceleration of 159-171 gal). From this,
gories, namely the case of shallow liquefiable it can be seen that embankment settlements in-
layer with thickness of 15-20 cm and the case of crease with increases in the thickness of the
deep liquefiable layer with thickness of 30cm or liquefiable layer.
more. In the case of shallow liquefiable layer, Figure 7 shows the settlement speed at the em-
the rate of increase of settlement tends to be- bankment crest for each model at the excitation
come small when accumulative settlement is large, stage 5. In the case of A-l with a shallow li-
however, in the case of deep layer, accumulative quefiable layer with thickness of only 15 cm,
settlement tends to increase in proportion to settlement speed takes its maximum at about one
the response acceleration in the ground just be- second after the start of excitation. Character-
low the embankment. This is presumably because istics of settlement speed can be generally di-
of the difference of liquefiable layer thickness vided into these two types.
which may affect the deformation characteristics Similarly, Fig. 8 shows time histories of exces-
of the embankments and grounds. sive pore water pressure in the ground just below
Fig. 5 shows successive changes in the mesh the embankment, at excitation stage 5. Excessive
lines at each end of excitation stages 3 (shak- pore water pressure becomes the maximum at about
ing table acceleration of about 100 gal) ,
4 2.5 seconds after the start of excitation in the
(about 130 gal) and 5 (about 160 gal). Those case of A-l. For the liquefiable layer thickness
figures were drawn from photographs taken of 20 cm or more, however, excessive pore water
through the observation glass window on the side pressure takes its maximum at about 0.5 to 1 se-
of the container. The glass window is equipped cond after the start of excitation. It is seen
for watching the deformation of the ground and that excessive pore water pressure well corres-
the embankment. In the case of A-l with a shal- ponds to the settlement speed of the embankment.
low liquefiable layer with thickness of only 15 Also, it can be seen from Fig. 8 that the settle-
cm, horizontal displacement in the ground was ment speed of the embankment crest monotonously
observed for the entire depth of the liquefied increases until the excessive pore water pressure
layer, with the maximum displacement at the ratio in the ground just below the embankment
ground surface. In the case of A-5 with a deep reaches 0.5-0. 6, but after that it becomes near-
liquefiable layer with thickness of 50 cm, hori- ly constant or decreases. This decrease in set-
571
tlement speed occurs for the case of A-2 and which are likely to cause embankment deformation.
subsequent models where the settlement of the
embankment crest is large. This probably indi- 5. REFERENCES
cates that the decrease in settlement speed can 1) P.W.R.I., "Report on the Damage during the
attribute, largely, to the decrease in embank- Niigata Earthquake of 1964," Public Works Re-
ment load resulting from the decrease in embank- search Institute Report, Vol. 125, 1965.
ment height. 2) P.W.R.I . , "Report on the Damage During The
Tokachi-oki Earthquake of 1968," Public Works
4. CONCLUSIONS Research Institute Report, Vol. 141, 1970
The following can be concluded from shaking 3) Sasaki, Y., et al ,
"Seismic Damage Examples
table tests on embankment models with various of Earth Structures," Technical Memorandum of
liquefiable layer thicknesses: Public Works Research Institute, Vol. 1756,

(1) The amount of embankment settlement for the 1970
same input acceleration gets large as the 4) Iwasaki, T. ,
Tokida, K., Kimata, T. , "Studies
thickness of liquefiable ground layer in- on Liquefaction of Sands during Earthquakes,"
creases . 13th Joint Meeting, UJNR, 1981
(2) In case where the thickness of the liquefi- 5) Iwasaki, T., et al., "Estimation Procedure
able layer exceeds a certain level, the ef- of Liquefaction Potential and Its Application
fects of the thickness on embankment settle- to Earthquake Resistant Design," 14th Joint
ment become minor. Meeting, UJNR, 1982

(3) This threshold thickness generally agrees 6) Osterberg, J.O., "Influence Values for Verti-
with the depth at which the horizontal dis- cal Stresses in a Semi-infinite Mass due to
placement of the ground can be observed for an Embankment Loading," Proc. 4th ICSMFE,
the case of deep liquefiable layer. Vol. 1, 1957
(4) Generation of excessive pore water pressure 7) Arakawa, T., Kimata, T., Kondo, M. ,
"Experi-
in the ground just below the embankment well mental Study on Failure Pattern of Embankment
corresponds with the settlement speed of the Supported by Sandy Layers During Earthquakes"
embankment crest. The settlement speed mon- 15th Joint Meeting, UJNR, 1983
otonously increases as long as the excessive 8) Arakawa, T., Kawashima, K., Matsumoto, H.,
pore water pressure ratio remains up to about Kondo, M. ,
"Experimental Study of Embankment
0.5-0. 6. Rut for the pore water pressure ex- on Sandy Layers During Earthquakes - Charac-
ceeding this level settlement speed decreases teristics of Reconstructed Embankment on Liq-
with the decrease in embankment height. uefied Layers," 16th Joint Meeting, UJNR,
The test results point out that for an embankment 1984
with a height of 15 cm, the range of layer thick-

nesses that affect settlement is 30-40 cm (ap-
proximately double the embankment height). Since
the above is the results from the shaking table
tests on model embankment with a certain section

shape made in a container, the results cannot be
directly applicable to actual embankments. Yet,
it is considered that the above tendencies may
apply to actual embankments. The above conclu-

sions may be helpful when determining the depth
of soil improvement of liquefiable ground layers
572
Table 1 Characteristics of materials used for the Experiments
Sign Un Sand Loam

I terns
— t
i
Specific Gravity Soil Particle G 2.641 2.556
Optimum Moisture
of
Content
s
18.0 —
—
W0 p t 96
Maximum Dry Density 1.574
—
7 dial tf/rrf
Maximum Void Ratio e »ax 1.007 2.543
Minimum Void Ratio em i n

— 0.647 1.665
60% Grain Size D 60 mm 0.301 0.225
10% Grain Size D j o mm 0.217 0.042
Mean Grain Size D so mm 0.290 0.165
Coefficient of Uniformity Uc — 1.387 5 .3 a
Table 2 Characteristics of Models of Embankments and Grounds
Model No. A- 1 A- 2 A- 3 A—4 A- 5 A- 6

Thickness of the
15 20 30 40 50 60
Liquefiabl Layer [cm]
Dry Unit Weight Embankment 1.39 1.38 1.41 1.41 1.42 1.38
7 d [tf/m 3 ] Ground 1.49 1.50 1.54 1.48 1.49 1.49
Void Ratio Embankment 0.905 0.915 0.877 0.878 0.861 0.917
e Ground 0.769 0.762 0.718 0.786 0.778 0.768
Relative Density Embankment 28.2 25 .4 36.0 35 .7 40.4 24.9
Dr [96] Ground 66.0 67.8 80.2 61.3 63 .5 66 .3
Table 3 Measured Maximum Acceleration of the Shaking Table

Mode 1 No A- 1 A- 2 A- 3 A— 4 A- 5 A- 6
Thickness of the
15 20 30 40 50 60
Liquefiabl Layer [cm]
Excitation Stage 1 36 37 36 32 38 36
Mesured Excitation Stage 2 71 74 68 60 68 68
Maximum Excitation Stage 3 103 103 103 100 100 100
Acceleration Excitation Stage 4 139 133 134 132 132 128
[gal] Excitation .Stage 5 171 164 163 159 161 167
Excitation Stage 6 204 197 — — — 195
573
A Pore Water Pressure Mater
V Displacement Gauge
Fig. 1 Ground and Embankment Model
574
(cm)
Settlement crest
Embankment
Accumulative
the
of
Acceleration of the Shaking Table (gal)
^™ 3 Relation between the Acceleration of the Shaking Table and
the accumulative Settlement of the Embankment crest
(cm)
Settlement crest
Embankment
Accumulative
the
of
r~i m r— i * 1 1 1 1 1—
0 100 200 300 400 500 600 700
Response Acceleration of the Ground
just below the Embankment (gaO
Fig. 4 Relation between the Response Acceleration of the Ground just below
the Embankment and accumulative Settlement of the Embankment crest
575
Fig. 5 Successive changes in the mesh lines at each end
of excitation stages 3,4 and 5
576
0
•+- <n
M-0
o »-
c +-
0 C
E Q
0 E
0 0
CO ID
E
LU
0 2 4 6 8 10 12
Time (sec)
Fig. 6 Relation between Time and Settlement of the Embankment crest
577
<<r 0.04
E
i_ ô 0.00
-*-•
co ê.
O) avo ,=: 0.025kgf/cm 2
-0.04
o ID
i_
0.04
n
c
03 0.00
0
0vo '=O.O 31 kgf/cm ;
-0.04
0.02
*fe
O
0 0.00
co
O)
-0.02 a v0 -0.025kgf/cm 2
Z 2 °-04
O m0
£
CL yj 0.00
0
^ - 0.04
Oyo— 0.03 1 kgf/cm 2
Time (sec)
Fig. 8 Example of Records of Pore Water Pressure
578
Photo. 1 Failure pattern of Case A-l Photo. 2 Failure pattern of Case A-2
(excitation stage 5) (excitation stage 5)
Photo. 3 Failure pattern of Case A-3 Photo. 4 Failure pattern of Case A-4
SiSfiBU
1
Photo. 5 Failure pattern of Case -A- Photo. 6 Failure pattern of Case A-6
579
STUDY ON ASEISMIC MEASURES TO EMBANKMENTS
ON SOFT CLAY GROUND
by
Yasuyuki Koga (1) and Osamu Matsuo (2)
ABSTRACT soft clay ground, such as the failure at the

Yoshida-River embankment caused by the
The effects of four types of aseisinic measures Miyagi-ken-Oki Earthquake (1978). Therefore,
to embankments constructed on soft clay ground aseismic stability and earthquake-resistant
were studied by model shaking table tests and measures to embankments on soft clay ground
2.
finite element analysis. Aseismic measures should also be studied. This paper describes
investigated in this study were "Counterweight the results of both experimental and analytical
fill method", "Sheet pile method", "Soil studies on the effects of several types of
improvement method", and "Reinforced earth meth- aseismic measures to embankments on soft clay
od". Shaking table tests were performed on a ground .
model embankment on soft clay ground prepared in

a 2.5m x0.6m x 0.7m steel box. Analyses to OUTLINE OF STUDY PROGRAM
obtain earthquake-induced deformation were also
conducted. Settlements of the various measures The objective the study described herein is
of
were compared with unmodified embankments. The to estimate the effects of several types of
experimental and analytical results are summa- aseismic measures to embankments on soft clay
rized as follows: ground in a qualitative and quantitative manner.
1. Shake table tests and analyses showed that For this purpose, a series of shaking table
all the measures reduced the settlement of tests and finite element analyses were undertak-
embankment on soft clay ground. en. Aseismic measures considered in this study,
2. Calculated settlements were in a range from numbers of shaking table tests and parameters
a half to twice the observed settlements. considered in the analysis are shown in Table 1.
The "earthquake induced permanent deforma- Fig. shows the flow of this study.
1 Firstly, a
tion analysis method" was considered appli- series of shaking table tests on clay ground-em-
cable . bankment models with and without aseismic meas-
3. Application of the "earthquake induced ure were performed to estimate the qualitative
permanent deformation analysis method" effects of the measures. Secondly, finite
1. revealed the following: element analyses were applied to the test
i) Counterweight fill method: As the width models, and by comparing calculated and observed
and height of the counterweight fill are results the applicability of the analysis method
increased the earthquake-induced settle- to predict the earthquake-induced deformation
ment is reduced. was investigated. Lastly, applying this analy-
ii) Sheet pile method: As the length of piles sis method to the hypothetical proto-scale
is increased, the effect becomes larger. ground-embankment model with and without several
The effect of this method was not types of aseismic measures, these effects were
confirmed in shake table tests. estimated. The procedure shown in Fig. 1 seems
iii) Soil improvement method: Improving the to be generally used to study such geotechnical
ground outside of the toe of the embank- problems, and in this case the followings .ire
ment showed comparatively larger effect pointed out as the advantage.
than improving the portion inside the toe. 1. The scale model need not necessarily satisfy
iv.) Reinforced earth method: Even a very high the similarity conditions of the prototype.
modulus geotextile showed little effect on 2. Even if the physical and mechanical condi-
the settlement, though lateral spreading tions of the models are not the same, the
of the ground was reduced. effect of the difference can be corrected by
analysis, so we can indirectly know the
INTRODUCTION effect of the remedial measures.
3. Because FEM simulation is far more conven-
The Nihonkai-Chubu Earthquake (1983) caused ient than model tests, we can easily examine
damage to many earth structures such as road and many different cases by using simulation
river embankments. Most of this damage was analysis
caused by liquefaction of loose saturated sandy
deposits However, as the analysis is applied to the cases
other than those verified by experimental
For the last decade, liquefaction-induced results, the simulation is essentially an
damages have attracted the attention of "extrapolative method", therefore the applica-
researchers because of their numerous and bility of the analysis method should always h<-
catastrophic examples. But there sometimes considered .
occur similar heavy damages to embankments on
(1) Head, Soil Dynamics Div. P.W.R.I., Ministry of Construction, Tsukuba, Japan
(2) Researcher, Soil Dynamics Div. P.W.R.I., Ministry of Construction, Tsukuba, lapan
580
3. EXPERIMENTAL STUDY part. This indicates that a primary consol-
idation was not completed in this part. The
3.1 Test Procedure degree of consolidation in terms of strain, U ,
was 80-90% in all cases. It is also shown that

Shaking table tests were carried out o 4 models
. the shear strength was relatively low in the
as shown in Fig. 2. Each model consisted of a upper part in Case 4 compared with that in other
clay ground, embankment with or without a reme- cases. This fact seemed to influence the test
dial measure. The ground was prepared by results that the Case 4 model was not so stable
consolidating the slurry clay in a steel box seismically.
3.2.2
with the length of 2.5 m, a height of 0.6 m and
width of 0.68 m. The material was an alluvial Results of Strong Vibration Tests
clay, of which physical properties are as
follows It was found that in all cases the amplitude and
Soil classification: CH the distribution of response acceleration was
Natural water content: Wn approx. 60% relatively stable and the amplification factor
Liquid limit: LL = 51% of the embankment was 1.5 - 2.0, up to the input
Plasticity index: PI = 24 acceleration level of about 100 gal where a
Specific gravity: Gs = 2.686 significant deformation did not occur in the
The consolidation was performed for three months models. However, at the acceleration level
under both sides drainage condition, however, where a plastic deformation began to occur in
the primary consolidation was not completed. the model, the amplification of response accel-
The embankment was constructed on the clay eration increased rapidly during vibration. On
ground by compacting the material with a wooden the contrary, at a greater input acceleration
plate, just after the consolidation weight was level, it decreased. It would be because the
removed. The embankment material was prepared weaker portion of the ground yielded by the
by mixing lead shots (Gs - 11.6, 2.0 mm dia) and shear stress induced by the vibration. Fig. 4
clay (w approx. 50%) with weight proportions of illustrates the deformed shape of all models
25 :3. The sheet piles and tie rods in case-4, after final vibration (350 - 400 gal). The
waveform steel plates (El = 500000. kgf.sq cm/m) vertical lines in the ground shown in Fig. 4 are
and 2 mm diameter steel wires were used respec- the deformation of the line markers which were
tively, considering the similarity of stiffness. inserted after the consolidation of the ground
Main sheet piles and anchor sheet piles are 35 ended. This figure indicates that the embank-
cm and 10 cm in length respectively. They were ment settles as the ground is compressed verti-
connected with 4 tie rods on each side. The cally and spreads laterally. In Case 1 of no
sheet piles and tie rods were installed into the measure, the horizontal deformation and internal
ground prior to embankment construction, and tie tension crack at the center portion can be seen
rods were pretensioned to 1.0 kgf respectively. in the embankment . In Case 2 and 3 of counte-
For the model ground, prior to embankment weight-fill method, it can be seen that lateral
construction, a set of vane shear tests (with 4 spreading of the ground at the filled side were
cm height by 2 cm diameter vane), cone pene- comparatively restrained and this tendency is
tration tests (with shape angle of 30 degrees more evident in Case 3 with wider fill. More-
and section area of 1.0 sq cm cone), shear wave over, because the tension crack of the main
loggings and water content tests were performed. embankment was restrained, large cracks appeared
The shaking table test consisted of a resonance at the approaching portion of the fill. This
test and strong vibration test. Sinusoidal phenomenon seems to be supported by the deforma-
horizontal vibration was applied. Firstly, a tion pattern of the ground surface. In Case 4
resonance test was performed at a constant of the sheet-pile method, the outside ground was
acceleration amplitude of 10 gal. Fundamental rather uniformly deformed, whereas in Case 1 the
frequency obtained was in the range of 15 - 18 lateral spreading was concentrated in the upper
hz in all cases. The strong-vibration test was part of the ground, due to the existence of
subsequently performed. In this test ,
input sheet piles. Further, the tensile stress ampli-
acceleration was a sinusoidal wave with a tude of the tie rods during vibration were not
frequency of 5 hz and duration time of 5 sec. large and after the vibration ended. Therefore,
Acceleration amplitude was changed intermittent- it is concluded that the supporting sheet piles
ly from 50 gal to 350 or 400 gal with a pitch of were not effective for Case 4. In addition, the
50 gal and pauses of 20 - 30 minutes. During test results described herein are considered to
the test, displacement, acceleration, pore water appear qualitatively as the deformation pattern
pressure and tensile stress of tie rods (only in of an embankment which is subjected to earth-
Case 4) of the model were recorded. quake just after construction (i.e., the consol-
idation has hardly progressed.) Fig. 5 shows
3.2 Test Results the relationship between the input acceleration
and the settlement at the top of the embankment.
3.2.1 Properties of Models It should be noted that the settlement shown in
Figure 5 is accumulated and includes settlements
Fig. 3 shows some properties of the model induced by previous vibration stages, therefore
grounds. It is shown that in the middle part of the settlement shown should be reduced to some
the ground the water content ratio was high and extent to obtain the settlement from the
shear strength was low compared with the other vibration stage. Although there were some
581
differences in shear strength of each model
ground as shown in Fig. 3, it can be seen in 4.2 Results of Analysis
Fig. 5 that in the case of the counter-weight
fill method, the settlement of the main embank- Fig. 6 shows the amplification factor of accel-
ment was decreased as the width of the fill was eration in the middle section of the model
increased. But in the case of the sheet pile obtained by the dynamic response analysis of
method the settlement was larger than that in Case 1, being compared with observed one. This
the case with no measure (Case 1). The follow- figure shows that the analysis provides fairly
ing two reasons can explain this result. One is good agreement between calculated and observed
the fact that the shear strength of the upper response accelerations at 10 gal shaking but the
part of the ground was lower in Case 4 than that calculated value is greater than observed at 20
in other cases. Another is that the response gal shaking. This would be because of yielding
acceleration at large input shaking was also of the ground at from middle to upper portion,
larger in Case 4, which might indicate the subjected to large shear stress caused by large
possibility that the shear force during shaking shaking. On the other side, the calculated
was transmitted to the upper portion on the response acceleration is slightly increasing at
model through the main sheet piles and this 200 gal shaking compared with 100 gal shaking.
accelerated the deformation. Finally, it should This is because the non-linear characteristics
be noted that the test results described herein influenced the response of the model and a larg-
were derived under the test condition and all er strength parameter for the material was used
the models were prepared in a constant size test in the non-linear dynamic analysis to obtain a
box whereas the size and location of embankment stable solution. Overall, it was concluded that
and fill are different among models, which would the dynamic response analysis gave satisfactory
have influenced each test result differently as and practical results. Fig. 7 shows the
a boundary condition. This kind of problem deformed finite element mesh after 200 gal shak-
peculiar to model testing will be solved by the ing obtained from the permanent deformation
analysis described in the following chapters. analysis. Such deformation can be compared with
the test results in Fig. 4 qualitatively,
4. ANALYSIS FOR THE MODEL TESTS although the shaking acceleration level is
different from each other. In all cases, the
Finite element analysis was applied to the test settlement of the embankment occurs accompanied
models described previously. This analysis was by lateral spreading of the underlying ground,
intended to confirm the applicability of the at the same time a heaving occurs beyond the toe
analysis method. "Ear thquaked-induced permanent of the embankment. Comparing the deformation of
deformation analysis method" was used in the Case 2, 3 with Case 1, lateral spreading is
analysis, which can calculate the permanent relatively reduced in the ground side where
deformation of the earth structure induced by counterweight fill is constructed, and it is
earthquakes and was developed and improved by more effective when a larger fill is
the authors for these years. constructed. In Case 4, the sheet piles are
restraining the local lateral spreading of the
4.1 Method of Analysis ground. These analytical results correspond
well to those of the model tests.
This method is described in detail in Ref. 1.
Parameters of the analysis were provided as In addition, in Case 4, as the stresses of tie
follows. Most physical and mechanical proper- rods were calculated as compression, re-analysis
ties were determined from the results of phys- was done using the model without the tie rods
ical property tests, cone penetration tests, because the tie rods do not resist compression
vane shear tests and triaxial compression tests. stress. This also corresponds very well to the
Moreover, the cyclic stress-permanent strain test result that the stresses of tie rods were
relationship of the clay material, which is one minimally increased by shaking. A comparison
of the most important parts of the analysis between observed and calculated settlement at
method, was given by performing some 20 cyclic the top of the embankment in all cases is shown
triaxial compression tests and compiling these in Fig. 8. But it should be noted that the
test results. Poisson's ratio of 0.499 was observed settlement shown in this figure is
used in the deformation analysis considering accumulated and includes that having already
that the deformation of the models during shak- occurred at the end of previous shaking stage,
ing would be occurring in substantially in an so the settlement will be underestimated to some
undrained condition. Using the parameters extent. It is found in this figure that the
described above, permanent deformation was both settlements appear to be in good agreement
calculated under the condition of input acceler- with each other except that in Case 4.
ation of 100 and 200 gal, for all the cases. A likely that effect of sheet piles would be over
computer program SADAP was used in the analysis, estimated because a slippage between the sheet
which was developed by the first author and can pile and the ground was not considered in the
conduct static and dynamic analysis of earth analysis. Further study will be needed in till-,
structures considering non-linear stress-strain area
characteristics of the soil(2).
582
5. ANALYSIS FOR COMPARISON OF EFFECTS OF COUN- not significant for a constant input acceler-
TERMEASURES ation level.
5.1 Models of Analysis Fig. 10 illustrates the deformed finite element

mesh due to 200 gal shaking. The followings are
Analytical investigation was performed on the found from Fig. 10. Deformation patterns of
effects of several types of aseismic measures (a), (b) and (c) in Figs. 4 and 7. It can be
assuming that they were applied to a hypothet- seen from figure (d) that the improved walls
ical prototype embankment. Fig. 9 shows analyt- tilt outwards, at the same time resisting the
ical models used in the analysis. The local spreading of the ground. The effect of
fundamental model is the embankment with dimen- the improved walls seems to look like that of
sions of 5 m in both height and top width and sheet piles. In figure (e), a slippage occurs
with side slopes of 1:2 constructed on the clay between the ground surface and the geotextile,
ground (thickness = 10 m) underlain by a sand whereas it does not occur between the geotextile
layer (thickness = 5 m) . In other models coun- and the embankment, and the amount of lateral
ter-measured were added to be the fundamental spreading of the ground is slightly reduced by
model. The sand layer was assumed to be good the geotextile compared with that in figure (a).
enough to resist liquefaction during shaking and But the effect of reducing the settlement is
normally consolidated so that the strength insignificant. Fig. 22 summarizes the results
distribution increased with depth. The Cu/p of the analyses which show the relationship
ratio was assumed to be 0.34. All the other between the settlements and maximum input accel-
properties were given realistic values that can eration in all cases. The result indicates that
be observed in the field. A safety factor of most types of countermeasures are more or less
the fundamental model in a static condition is effective for reducing the earthquake induced
calculated as 1.2 by the circular sliding permanent deformation of the embankment by
surface method. Sheet piles were used at both comparing the deformation with that in the
sides. As a soil-improvement method, a fundamental model without measure. Fig. 12
wall-shaped-type "Deep Mixing Method" was taken shows how the calculated settlement varies by
which consists of a row of three walls with a changing the dimensions of the counterweight
thickness of 1.0 m and a pitch of 1.6 m. Two fill. This indicates that the counter-weight
cases were assumed in the analysis where the fill method is more effective when its dimension
walls are driven into the ground inside and is larger. In Fig. 11 again, it is found that
outside of the toe of the embankment. The walls the soil improvement method is slightly more
are not perfectly supported in the sand layer, effective in Case A than in Case B of Fig. 9(d).
and the clay between the walls were not This might be due to the ground condition, that
improved is, shear strength of the ground is gradually
weaker depending on the distance from the center
5.2 Conditions of Analysis because of the absence of consolidation stresses
by the embankment load. Moreover, the weaker
The clay layer was considered to be fully portion of the ground is improved in Case A than
consolidated by the embankment load and therein Case-B. This might explain why the soil
fore the shear strength increased to the depth improvement is more effective in Case A. Next,
according to the ratio of strength increase, the sheet-pile method appears most effective
Cu/p of 0.34. Counterweight fill was considered among all the cases shown in Fig. 11, however,
to be made of the same material as that of main it should be noted that the analysis in this
embankment. The improved walls were considered case has the problem mentioned previously.
elastic (G = 400kgf/cm2) without yield strength. Therefore further study will be needed in the
In case of the reinforced-earth model, joint analysis of the sheet-pile method. Fig. 13
elements were arranged at both sides of the shows the settlement and Young's modulus of
geotextile to permit slip deformation between geotextile obtained from the analysis of the
geotextile and soil, where the interface shear reinforced-earth method. When a value of EA is
strength parameter was given as the normal smaller than 100 tf/m, there is no effect, but
stress times the tangent of 20 degrees, whereas when it is 1,000 tf/m, the settlement is reduced
joint elements were never used in other models. by about 20%.
An acceleration record obtained at the' ground
surface near Tsugaru-Ohashi Bridge at the Nihon- 6. CONCLUSIONS
kai-Chubu Earthquake, 1983, was used as an input
acceleration, with the maximum acceleration of Effects of four types of aseismic measures to
200 gal in most cases. In each case, parameters embankments constructed on soft clay ground have
presented in Table 1 were considered in the been studied experimentally and analytically.
analysis Countermeasures investigated herein were as
follows: counterweight fill method, sheet-pile
5.3 Results of Analysis method, soil improvement method, and reinforced
earth method. The latter two methods were
It was found as a result of the dynamic response studied only analytically. The main conclusions
analyses that the differences of maximum obtained from this study are as follows:
response accelerations among all the cases were
583
1. Each countermeasure showed the effect more 7. ACKNOWLEDGEMENTS
or less to reduce the earthquake induced
settlement of embankment. The study described in this paper was sponsored
2. The method of analysis adopted in this paper by the Arakawa-Karyu Construction Office, Kanto
provided satisfactory agreement between District Bureau of Construction, Ministry of
calculated and observed settlement of the Construction. The authors wish to acknowledge
embankment due to earthquake motion. their financial support. Acknowledgement is
3. The counterweight fill method was more also given to Mr. Yasuaki Karasawa for his
effective, as the dimension of the fill participation in this study.
becomes large.
4. The sheet-pile method was effective in 8. REFERENCES
restraining the ground to make local lateral
spreading, but the effect to reduce the 1. Sasaki, Y., Matsuo, 0. and Karasawa, Y.:
embankment settlement was not confirmed in Analysis method of earthquake-induced defor-
the experiment. mations of earth structures and its applica-
5. The soil improvement method, of which the tion, Civil Engineering Journal, Vol. 27, No.
Deep Mixing Method was adopted in this 3, pp. 46- 51, 1985. (in Japanese)
study, showed the effects that are similar
to that of the sheet-pile method. 2. Koga, Y., Tateyama, S. and Karasawa, Y. :
6. The reinforced earth method which uses the SADAP-1, A computer program for the static
geotextile under the embankment showed and dynamic analysis of earth structures
little effect to reduce the embankment considering nonlinear stress-strain relation-
settlement ship of soils. Memorandum of the Public Works
Research Institute, No. 1688, 1981. (in Japa-
nese )
584
Table 1. Aseismic measures considered in the study
Case number Parameters considered

Aseismic measure
of model test in analysis
No measure Case-1
width, height of
Counterweight-fill method Case-2, 3
counterweight-fill
Sheet-pile method Case-4 length of sheet piles
Soil improvement method location of improved zone
Reinforced earth method modulus of geotextile
Fig. 1. Study flow
585
Case- 1 (No measure)
^XÊrânkment^C.
Clay ground
250
Case- 2 (Counterweight fill method)
Case- 3 (Counterweight fill method)
Case - 4 ( Sheet pile method)

Anchor sheet pile ( ( — 10cm)
Fig. 2. Schematic diagram of test models
586
Water content ratio ;
W {%) Shear strength: Cu(tf/ m ) Shear wave velocity: Vs (m x 's)
Fig. 3. Properties of model ground
CASE-2 (Counterweight fill method)

Counterweight fill
(h=15cm, w=30cm)
ImWnTTTTTlT
CASE-3 (Counterweight fill method)
CASE-4 (Sheet pile method)

Tie rods
Sheet piles
Fig. 4. Observed deformations of the test models

(after 350-400gal shaking)
587
Input acceleration ( gal
Fig. 5. Relationship between input acceleration and settlement

of embankment in the experiment
Fig. 6. Comparison of calculated and observed amplification factor

of acceleration
588
Case— 1
Case— 3
Case — 4
Fig. 7. Calculated deformations of the test models

(after 200gal shaking)
589
Fig. 8 Comparison of calculated and observed settlements of
embankments
(a) No measure (Fundamental model)
(b) Counterweight fill method ( c) Sheet pile method
-77777777 777/777
(d) Soil improvement method (e) Reinforced earth method
Improved
«> zone® /
(Deep Mixing Geotextile
Method) 77
77777777 7777777
Fig. 9 Models in simulation analysis
590
10m
Scale of model
1 m
Scale of deformation
Sand
Counterweight fill
Sheet pile
(c) Sheet pile method II
o 3 —
.
^ r- Fj
Improved walls
(d) Soil improvement method
/A T
Li mi L L
Ml: v
nr
H. 1 |
Geotextile
Fig.10 Deformed finite element mesh
5B1
Maximum input acceleration (gal)
0
6
"
Sheet pile method (^ = 15m)
10
" (^ = 10m)
20 Counterweight fill method
I
( h = 3 m, d =30m
I 30 ( h = 5 m, d - 10m)
s
a; (hi 3 m, d —10m)
O 40 Soil improvement method
(outside)
50 (inside)
| Reinforced earth method
(EA = lOOOtf)
w 60
(EA= 1 ~100tf)
No measure
70
Fig. 11 Calculated settlement v. s. maximum input acceleration
(cm)
embankment
of
Settlement
Fig. 12 Relationship between the dimensions of the fill and the

settlement of embankment
Modulus of geotextile, EA (tf/ m)
Fig. 13 Relationship between modulus of geotextile and settlement

of embankment (calculated)
592
DRAFT GUIDELINES FOR POST-EARTHQUAKE INSPECTION
AND EVALUATION OF EARTHQUAKE DAMAGE IN
REINFORCED CONCRETE BUILDINGS AND ITS
APPLICATION TO THE NAM I OKA TOWN HOSPITAL BUILDING
BY
I O O
Tsuneo Okada, Masaya Hirosawa, Hisahiro Hiraishi,
and Manabu Yoshimura^
ABSTRACT
The Ministry of Construction, Japanese evaluation for risks of damaged

Government has been promoting a five- buildings to aftershocks, 2) the
year project which is called "Develop- guideline for the first level
ment of Repair and Strengthening Techni- inspection and evaluation identi-
ques for Buildings and Structures fying damaged buildings and treatment
Damaged by Earthquakes" since 1981. for continued occupancy, and 3) the
guideline for the second level
The guidelines for post-earthquake inspection and evaluation for
inspection and evaluation of earthquake necessity of structural strengthening.
damage in buildings, which is one of The judgement for these is based on
the objects of the project, was foundation settlement and listing of
established in 1985. buildings and degree of damage to
columns and walls.
They are applied for wooden structures,
steel structures, reinforced concrete As a result of application of the
structures and artificially field above three guidelines for inspection
ground. This paper describes the and evaluation of damage to the
summary of the guideline which is Namioka Town Hospital, the following
applied to reinforced concrete build- decisions were drawn.
ings, and also describes the results
When the guidelines were applied to the a. Entry prohibition (based on the
inspection and evaluation of the emergency inspection and evaluation)
earthquake damage of the Namioka Town
Hospital Building in Aomori Prefecture b. Severe damage (based on the first
which was severely damaged during the level inspection and evaluation)
1983 Nihonkai-Chubu Earthquake.
c. Structural strengthening needed
The guidelines consist of 1) the (based on the second level
guideline for emergency inspection and inspection and evaluation)
1
Professor, Institute of Industrial Science, University of Toyko.
O
Director, Research Planning and Information Department, Building Research
Institute, Ministry of Construction (B.R.I.).
O
Senior Research Engineer, Structural Engineering Department, B.R.I.
^ Special Assistant of Director for International Corporation, B.R.I.
593
These results are the same as the ones This is to judge the degree and
which had been reached by the authority location of damage of a building and
based on detailed inspection and the required strengthening for continued
evaluation. This paper finally occupancy.
describes newly recognized several
matters of concern. c. Guideline for the Second Level
Inspection and Evaluation
1. THE GUIDELINE FOR INSPECTION AND
EVALUATION OF EARTHQUAKE DAMAGE This is to judge the necessity for
OF REINFORCED CONCRETE BUILDINGS strengthening the damaged buildings
which is unclear to be judged only by
1.1 The Objective of This Guideline the guideline for the first level
inspection and evaluation.
The objective is to make possible the
prevention of incidents of deaths and Figure 1.1 shows the general flow to
injuries to occupants from aftershocks be applied for each guideline. Each
immediately after a major earthquake guideline was prepared by the subcom-
and the selection of methods for mittee on reinforced concrete
restoration. structures and the working group
affiliated with the subcommittee
1.2 The Outline of This Guideline under the Committee on the Development
of Post-Earthquake Measures of
There are three types of guidelines to National Project based on the draft
be established for identification of written by Prof. Shunsuke Otani of the
earthquake damages. They are called 1) University of Tokyo, Prof. Katsumi
Guideline for Emergency Inspection and Takiguchi of The Tokyo Institute of
Evaluation, 2) Guideline for the First Technology, and Prof. Masamichi
Level Inspection and Evaluation, and 3) Ohkubu of the Kyushu Institute of
Guideline for the Second Level Inspection Design. The chairman of the subcommittee
and Evaluation. Each guideline is made on reinforced concrete structures is
1.3
based on the consideration of emergency Prof. Tsueno Okada of the University
degree of Tokyo, the chairman of the working
group is Prof. Masaya Murakami of the
a. Guideline for Emergency Inspection Chiba University, and the chairman of
and Evaluation committee of the National Project is
Prof. Hajime Umemura of the Shibaura
This is to judge safe entry to a Institute of Technology. In the
building by investigating potential following chapters, the detail of
risk of occupancy of buildings from each guideline is mentioned.
aftershocks immediately after a major
earthquake. Guideline for Emergency
Inspection and Evaluation
b. Guideline for the First Level
Inspection and Evaluation This is to make the urgent judgement
594
of the entry prohibition and the (severe) and V (collapse) are
limited entry to reinforced concrete classified based on (1) crack of
frame structures and wall structures concrete, (2) spalling of concrete of
in damaged areas of a major earthquake, columns, and (3) bending or fracture
with considering hazards produced by of reinforcing bars.
dropping of fixtures and the falling
of debris, risk to lives of people due Table 1.1 shows the investigation items
to collapse of buildings caused by of foundation settlement, inclination
aftershocks. This guideline is also of buildings and of superstructures,
applied to public buildings to judge and corresponding hazard rank.
the safety of the building. The
investigation should be executed The classification of judgement for
immediately after a major earthquake by danger, caution and safety is decided
building and civil engineers from as follows, and each decision is
governmental organizations in concern, accompanied by the posting of "entry
mainly through visual observation of prohibited", "entry limited" or
exterior appearance of damaged buildings, "safety"
in accordance with Emergency Earthquake
Damage Inspection Form: the outline of DANGER
a building, damage condition of a
superstructure, foundation settlement, 1. If there is one "rank C" in
inclination, appendage dropping hazard, the investigation items of
debris falling hazard, etc. The superstructure of a building,
investigation takes about half an hour and/or there are two or more
per building. At public buildings, the than two "rank B" in the same
following items are also to be investi- investigation items, the whole
gated: the damage condition of building is judged to be in
interior columns and walls, the falling danger
hazard, the dropping hazard in the
building and the adequacy of accommo- 2. If there is found damage degree V
dating the sufferers from a major in the interior of exterior columns
earthquake. The degree of danger is and walls, the area of "danger" is
determined according to each hazard from the damaged columns and walls
identification rank of each investiga- to the undamaged columns and walls
tion item. The rank is stated like less than damage degree III
rank A, rank B, and rank C. An nearest to the damaged ones, and
identification of damage of superstruc- the upper floor of the overhead of
ture is defined based on damage ranks the damaged columns and walls.
of exterior columns of the frame
structures and exterior walls of the 3. Every floor which support damaged
wall structures, of most severely stairways
damaged floor, whose damage rank are
equal to or higher than moderate. The 4. If there is any rank C in
damage ranks III (moderate) , IV the investigation items of
596
the dropping hazard, the area this guideline is applied to revise the
which is estimated to have the results of the emergency inspection and
hazard is decided as "danger". evaluation. If the judgement of repair
5. If there is any rank C in for restoration or strengthening is
the investigation items of difficult, the guideline for the second
the falling hazard, the area level inspection and evaluation is
which is estimated to have applied. The treatment for continuous
the hazard is decided as occupancy has the following four types.
"danger "
1. REUSE WITHOUT ANY STRUCTURAL REPAIR
CAUTION
Only cosmetic repair is done. Repair
The building with one or more than one such as injecting epoxy resin, for
rank B in each investigation item of example, is not necessary.
supersturcture of a building, the
dropping hazard of appendage and the 2. REPAIR FOR RESTORATION
falling hazard of debris.
The goal of this repair is to make the
SAFETY seismic performance of damaged
structures the same level as the
This judgement is applied only for condition before damage.
public buildings. Its determination
is made only for the buildings without 3. STRENGTHENING
any rank B and C in the investigation
items of the dropping hazard and the The goal of this strengthening is to
falling hazard, and with a very slight improve the seismic performance of
damage which is less than damage degree damaged structures better than the
III in columns of frame structures and condition before damage.
walls of wall structures.
4. DEMOLITION OR WITHOUT REUSE
1.4 Guideline for the First Level
The investigation is made by building
and civil engineers from governmental
This guideline is used to investigate organizations in concern, by persons
the outline of damaged buildings admitted as an investigator by
suffered from a major earthquake, to governmental organizations and by
identify the degree of damage of the persons whom the organizations
building, to determine the safety required the investigation. They make
against further earthquakes through the measurements of the inclination of
investigation of the intensity of the buildings and of foundation settlements,
earthquake ground motion which has and investigate damage conditions of
caused the damage, and to decide the superstructures (as to frame structures,
treatment of the damaged buildings for every columns and beams, and at to
continuous occupancy. If necessary, wall structures exterior walls should
596
be inspected as much as possible) The inspection is conducted on dairiage
If earthquake damage of a structure is of columns and walls, tilitng and

obviously identified, inspection is foundation settlement of the building,
stopped at that stage. The inspection and damage to piles, etc. The
may take, at most one day for two in- necessity for strengthening is
spectors per building. evaluated based on the tilting of the
building, soil condition and intensity
Identification of damage of the of earthquake ground motion etc., if
building is defined as the worst of the foundation or pile is damaged.
inclination, foundation settlement This is done based on earthquake
and damage of superstructure. The damage index being function of the
damage of beams of frame structures potential of the seismic capacity of
is assessed by replacing that by the building to resist further
damage of columns. Identification of earthquakes, if the superstructure is
the damage of the superstructure is damaged. Table 1.7 (omitted) shows
judged based on damage rate (shown in relationship between the tilting of
Table 1.2) defined by damage rank of building caused by damage of foundation
columns, beams and walls of the most and pile and the need for strengthening
severely damaged floor. of foundation and pile.
Table 13 through 1.5 show the each Table 1.8 shows upper limit values of
relationship between inclination, earthquake damage index requiring
maximum foundation settlement and strengthening. The earthquake damage
maximum damage rate, and identifica- index is calculated according to
tion of damage. Table 1.6 shows the following procedures.
relationships between the identifica-
tion of damage of the building and the 1. Seismic index I
£
before damage
treatment of continuous occupancy. is calculated by "Guideline for
evaluation of Seismic Capacity of
1.5 Guideline for the Second Level Existing R/C Building," issued by
Inspection and Evaluation Japan Association for Building
Disaster Prevention, April 1977.
This guideline is used to judge the
necessity for strengthening the 2. In the same manner, seismic index
damaged buildings, which are unclear to Is' after damage is calculated by
be judged according to the guideline the following equation.
for the first level inspection and
evaluation, or requires detailed
analysis for seismic capacity of a
V V - * G x SD '
x T
building to resist anticipated earth- where,

quakes before restoration for continuous
occupancy, and to obtain basic informa- EQ '
= basic seismic index after
tion about strengthening of the damage given by the evaluation
building for restoration. of seismic capacity of existing
997
R/C building in which reduction 1.6 Other .Contents and Application
coefficient of seismic ftf.t.he... G ui de lin e
capacity shown in Tables 1.9

and 1.10, depending on damage This draft of guideline was prepared
level of columns and walls, is as one of the intermediary research
multiplied to the basic seismic products of the national project.
index before damage, Although, here in this report, inspec-
tion and evaluation methods of
G, T = geological index and structural damages are mainly
time index, respectively, described, the following data should
which are the same as those be obtained with the report.
before damage,
a. hazard inspection for falling
SD '
= Structural design index of exterior cladding of build-
after damage which is the ings .
same value for the first b. hazard inspection for overturning
level evaluation of seismic of stairways for escape and equip-
capacity and for S D for the '
ment .
second level evaluation of c. inspection for buildings for places

seismic capacity the effect of refuge.
of damage is considered in d. many photographs of damaged
items of the eccentricity be- buildings and portions as
tween center of gravity and references to evaluating damage,
center of stiffness in and
plan and irregularity of the e. application examples of inspection
stiffness along the height of and evaluation of actually damaged
the building. Here the stiffness buildings
of columns and walls after
damage is given by multiplying However, as this guideline was compiled
that before damage by as one of the first drafts for each
structural kind, the following problems
3. Earthquake damage index $ is should be sufficiently discussed until
given by the following equation. the guideline is actually applied
widely to damaged buildings. Up to that
$ = {1-(I S '/I S ) } x 100 (%) time, this guideline will be used as
an available reference in case by case.
Earthquake damage index $ should be
calculated in both X and Y directions a. To consolidate data related to
of each story. However, damage to a several numerical constants with-
certain story should be focused if out sufficient back data.
damage to that story is obviously b. To adjust the evaluation standard
severe compared to other stories. for safety among the guidelines for
each kind of structure.
598
c. How to qualify persons to longitudinal direction are
inspect and evaluate earthquake pronounced and those in the
damages. transverse direction are fairly
d. How to treat the guidelines slight. The 3rd and 4th stories
administratively. were damaged severely; the 5th
story was damaged moderately; the
2. OUTLINE OF THE,. DAMAGES 2nd story was damaged slightly as
OBSERVED IN THE NAN I OKA well as the 1st and basement.
TOWN HOSPITAL BUILDING Damage of each story did not cause
a total collapse to the story as
2.1 Outline. of the Hospital ^Building a whole.
b. In the longitudinal direction,
The observed damage, analyses and pronounced shear failures and
restoration of the hospital were bond splitting failures were
already reported at the 16th Joint observed in many columns with
Meeting^) 3) 4) Therefore, the data spandrel walls at the 3rd, 4th
.
necessary to apply the guideline to the and 5th floors. In the same
hospital building are presented here. ways, shear failures or shear
cracks appeared in the reinforced
The hospital, as shown in Fig. 2.1, concrete walls at the 2nd, 3rd,
is a 5-story reinforced concrete 4th and 5th floors.
structure with one story basement. c. Additionally, a reinforced
The first floor is wider than the concrete chimney and parapets on
other floors. The structural system the roof tilted and many window
of the 5-story portion of this building glasses were broken and fallen
comprised of frames with few shear down. Out of the above-mentioned
walls in the longitudinal direction. damage, crack patterns of the
The external two frames, out of the exterior frame in the longitudinal
four frames, are distinguished by direction are shown in Fig. 2.2
uniformly existing spandrel walls. and examples of the detailed
Those in the transverse direction are crack patterns of columns are shown
mainly consisted of multi-storied frame in Fig. 2.3.
shear walls. This building was
constructed from 1968 to 1970 and had (2) Outline of Restoration
not suffered from other earthquakes until
the May 1983 Nihonkai-Chubu Earthquake. This building had been functioning
normally until the earthquake
2.2 Outline P.f_JB.amagea..and. Restoration occurred. After the earthquake, use
of the building was stopped partially
(1) Outline of Damage and the overall reuse was started
from August, 1984. Outlines of
The damage of this building is limited occupancy of the building
summarized as follows. during this period are as follows,
a. On the whole, damage in the a. The patients in the sick rooms on
999
the 3rd, 4th and 5th floors were frames are generally high, those
evacuated to outdoor by nurses columns of the interior frames are
immediately after the earthquake ranked low except columns in the
and these floors were sealed off 3rd and 4th stories,
except authorized persons until c. Although damage ranks of the
reuse of the entire building. interior columns in the exterior
b. The function of operation rooms was frames at the 3rd and 4th floors
stopped during the period of are classified to the rank IV, their
strengthening construction work damage of the several columns are
from Feb. 1984 to June 1984. very close to the rank V as the
c. The use of the other portions at bucking of longitudinal bars and
the first and basement floors were cracking of the floor slab were
stopped for about one through three partially observed.
months because of the repair and
strengthening work. However, even 3. EYftLU AT ION--QF...m.. E ARTHQUAKE
in this period, function as a hospital DAMAGE OF THE HOSPITAL
except the above-mentioned sick rooms BUILD I MG
and operation rooms was continued by
using temporally other portions of 3.1 Qut.iirLb
the building.
The guideline described in the first
In Fig. 2.4, main strengthening chapter is applied to evaluate the
construction methods adopted and their earthquake damage observed in the
location are shown. hospital building. That is, the
guideline for emergency inspection
2.3 Observed Damage of Columns and walls and evaluation is applied to determine
safe for entry to the building etc.;
As shown in Fig. 2.2, several kinds of the first level inspection and
damage, classified into several damage evaluation is applied to classify the
ranks, were observed in columns and damage of the building; and the
walls. Concerning these damages, second level inspection and evaluation
classification to several damage ranks is applied to judge whether the
were carried out based on Table 1.2 and building should be strengthened or
Table 1.11 and the results are shown in not
Fig. 2.5.
3.2 EmfcDafciiQy and First Level
As shown in the figure, damages ranks In£E£g.t.iP.n..and Eyflluat.ioh
are as follows:
a. In the portions higher than the 3rd (1) Emergency inspection and evaluation
floor, many members are ranked
higher degree than those in the In absence of any tilting and
lower portion. settlement of a building, safety of
b. Although damage ranks of the the building is judged as a whole based
interior columns in the exterior on the degree of damage exterior
GOO
columns in roost severely damaged 3.3 The Second Level Inspection
stories using Table 1.1. In case of and F^.&iuatiop
.
this hospital building, the 3rd story

was roost severely damaged and 8 columns (1) Calculation of residual seismic
out of 12 exterior columns in the 3rd capacity
story is rated as V. Accordingly, as
the hazard rank of the superstructure The objective of this level of
is rated as rank C by Table 1.1. Thus, evaluation is to further refine the
this building is judged to be dangerous residual seismic capacity of a
and the entry was prohibited. damaged building for strengthening.
In the case where foundation piles
(2) The first level inspection and suffered no damage, the residual
evaluation seismic capacity is decided by the
damage ranks of all vertical load
After identifying the most severely carrying members at all floors. That
damaged story, and rating each column is, after deciding damage rank of
in the story, repair of the structure each member and coefficient to reduce
for occupancy was decided by the seismic capacity of the member by the
resultant identification and the damage rank, residual seismic
seismic intensity of the earthquake. capacity using the coefficients
of each member by the damage rank,
The result obtained by applying the residual seismic capacity of the
first level evaluation is shown in building is calculated by the first
Table 3.1. In this table, the results or second level evaluation of seismic
are shown of all stories from the first capacity using the coefficients n of
to the fifth where damage of all each members.
columns were inspected. Furthermore,
the results judged only from rating of In Table 3.2, shear coefficients, C,
all exterior columns, assuming all of all vertical members in the 2~5th
interior columns are not accessible for stories are listed with the shear
t
inspection, are also shown in the coefficients, C , of those after-

table. As shown in the table, identi- earthquake. And as the basic seismic
fication of the damage at each story is index, Eo, of each story is shown in
generally reasonable but when there are the following expression in the case
many damaged columns unable to be without any extremely short column,
investigated, a different identification the value of Cc(Cw + 0.7C c E C ) ,
may be obtained because of significant and the ratio of EC '/£ C are also
difference in the sum of damage rates. shown in the table.
That is, although the result of the 4th
story was judged to be "severe" based Eo - (c + 0 - 7 c (i)
-"-fr w o>
only from the inspection of the
exterior columns, the result from the where,
all columns was rated as "moderate". Cw , C c = sum of shear coefficient of
each wall and of each column,
6Q1
respectively, and Sp index of this building after the
n, i = number of stories above ground earthquake is almost same as the
and location of the story under before plan the earthquake so
investigation counting from the far as the eccentricity of the
first floor, respectively. building is concerned. But stiffness
distribution along the height became
As understood from the values of IC/IC lower after the earthquake became
in Table 3.2, deterioration ratios of lower due to severe damage in the
seismic capacity of this building are third story. However, neglecting
I
0.97, 0.54, 0.67 and 0.89 at the 2nd, this change of S D ,

$ values are
3rd, 4th and 5th story, respectively, calculated assuming I
s
/I s = C / C.
and this means that damages of 3rd
floor was most severe. The results are listed in Table 3.3
with strengthening requirements
(2) Determination for strengthening according to Table 1.8. Also listed
are the results obtained from the
Necessity to strengthen a damaged first and second level evaluations of
building is judged from Table 1.8 based seismic capacity. As shown in the
on the earthquake damage index $ shown table, the two results by the first
in the following expression and the and second level evaluations agree
intensity of earthquake. with each other and it is concluded
from the investigation on the 3rd and
$ = {1- (I '
s /I s ) } x 100 (%) (2) 4th floors that this building should
be strengthened.
where,
t
l
I
£ = EQ x G x SD x T = Seismic Index 3.4 Comparison Between Results of
after earthquake Judgement. a nd. .Actual Treatment-.
.
G = geological index and Discussions

i
SD = structural design index after

earthquake (1) Entry prohibition
T = time index
Whole building is prohibited for
As shear reinforcement was not adequate entry as a result of seismic evaluation
both in columns and walls in the case of according to this guideline.
this building, the value of ductility However, entry prohibition for the
index, F, of each member is less than floors above 2nd floor was actually
1.27. Therefore, the basic seismic recommended for individuals who
index, E 0 , in the second level evaluation conducted investigations. Entry to
of seismic capacity is also shown in the floors below the 2nd floor was
Eq. (1). However, in this case, the permitted because this was the only
strength of each member should be hospital in the town. This was also
based on calculated ultimate flexural the reason for strengthening.
strength and shear strength.
GQ2
(2) Identification of damage of the developed by the National Project
building promoted by the Ministry of Construction,
and described the results when this
The building was classified as "severe guideline was applied to the Namioka
damage" to the whole structure (but, Town Hospital which was damaged
"moderate damage" is concluded if all during the Nihonkai-Chubu Earthquake
columns is inspected and evaluated) in May 1983.
On the contrary, experts actually class-
ified 5th, 4th, 3rd and 2nd floors as The result of the judgement agreed
moderate, severe, severe and small reasonably well with the actual
damages, respectively. Although both strengthening work which was actually
results agree with each other if they carried out. However, the following
are investigated in detail, in the subjects are decided to be investigated
former classification - severe damage as to the guideline.
it is necessary to distinguish whether
damage is partial or total. (1) General Comments
(3) Necessity or unnecessity of a. The treatment for the

strengthening of the building judgement on structural
for restoration members whose damage is
likely to be unclear
Strengthening of the building for the because of existence
restoration is necessary only for the 3rd such as wooden finishing
and 4th stories as a result of the should be explained.
application of this guideline to the
building. However, other floors were b. Judgement whether the
also repaired and strengthened while treatment of entry prohibition,
the repair was mainly concentrated on identification of the
the 3rd, 4th, and 5th floors. The damage and necessity of
degree of repair depended largely on strengthening of the
the difference in the expected intensity building for restoration,
of earthquake motion considered for the etc. is applied or net for
seismic structural analysis. A careful whole building should be
consideration should be paid in explained in detail.
carrying out strengthening work in
areas other than damaged floors, c. Selection of a most severely
particularly strengthening continuously damaged floor from the
from lower floors to upper floors. damage of exterior columns
should be explained (the
4. CQNCLUSIP N S emergency and first level
inspection and evaluation)
This paper described the outline of the
guidelines for inspection and evaluation
of damaged reinforced concrete building
( 2 ) The guideline for emergency a. Quantitatively values for
inspection and evaluation the earthquake damage index
representing a potential of
a. Exceptional cases, where seismic capacity to resist to
identification of the further earthquakes $ should be
damage is not classified by prepared
the damage of only exterior
columns, should be explained b. The relationship between
as example. the limit value of for
strengthening for restoration
b. There should be more examples and intensity of suffered
about dropping hazards and earthquake ground motion
falling hazards, and definite should be discussed further
examples of the area of these from the viewpoint of the
hazards should be added. target of strengthening.
c. It should be explained that c. It should be explained that

there are exceptional cases it is not rare that
of partially entry prohibition, strengthening might be
etc conducted for even slightly
damaged floor.
(3) The guideline for first level
inspection and evaluation 5. ACKNOWLEDGEMENTS
a. Judgement for continuous The authors would like to extend

occupancy in such a case wherein their heartily gratitude to many
building suffers moderate experts of universities and of
damage, moderate earthquake Building Research Institutes, for
ground motion should be giving much cooperation in the
discussed further (see Table investigation of earthquake damage of
1 . 6 ) . the Namioka Town Hospital and on the
preparation of the draft of the
b. It should be explained that guidelines for inspection and
there is a difference in the evaluation which were quoted in
treatment among floors (see this paper.
Table 1.6).
6. REFERENCES
c. The method of identification
of the flexural failure 1. "Guideline for Post-Earthquake
column and shear failure Inspection and Evaluation
column should be explained. of Earthquake Damage in
Buildings," Report of
(4) The guideline for second Development of Repair and
level inspect ion and evaluation Strengthening Techniques
604
for Buildings and Structures
Damaged by Earthquakes,
Building Research Institute,
1984.
2. Fasaya Hirosawa, et.al.,
"Damage and Repair of Namioka
Town Hospital Damaged by the
Nihonkai-Chubu Earthquake,
1983 - Fart 1: Outline of
Damage and Screening of
Structural Safety," Sixteenth
Joint Meeting, UJNR, May 1984.
3. Yoshikazu kitagawa, et.al.,

The same as the above -Part
2: Analysis of Damage by
Elasto-Plast ic, Static and
Dynamic Analyses," ditto.
4. Masaya Hirosawa, et.al., "The

same as the above -Part
3: Retrofit Design and
Estimation of the Effect,"
ditto
5. "A Seismic Diagnosis Standards for

Existing RC Building Structures,"
Japan Building Disaster Prevention
Association.
605
Settlement
Foundation
of
Rank
Hazard
Superstructure
and
Inspection
and
of
Inclination
Items
1.1
Table
606
Table 1.2 Damage Rank of Column, Beam and Exterior Wall
by Observed Damage Pattern and Damage Rate
by Observed Damage Ratio
Damage Rank Damage Pattern Observed Damage Damage Rate

of Column, Beam Ratio*! (di : % )
and Exterior Wall
I Visible Narrow Crack of Cone. Less than 10% 0.5

(Crack Width; less than 0.2mm)
10 'V 20 1
20 'U 30 2
30 % 40 3
A0 'U 50 A
More than 50% 5
II Visible Clear Crack of Cone. Less than 5% 0.5

(about 0.2 'v, 1.0mm)
5 'V 10 1
10 15 2
15 % 20 A
20 'V 25 5
25 % 30 6
30 % 35 8
35 % 40 9
40 'U 45 10
45 % 50 11
More than 50% 13
III Clear Crumbling of Cone, or Less than 3% 2

Remarkable Wide Crack of Cone. %
(about 1.0 ^ 5.0mm)
5 % 10 5
10 % 15 8
15 'V 20 10
20 'V, 25 13
25 30 15
30 'U 35 18
35 % 40 20
40 % 45 23
45 % 50 25
More than 50% 30
I Remarkable Crumbling of Cone. Less than 3% 3

with Exposure of Reinforcing 5
3 5
Bar
5 'U 10 9
10 % 15 14
15 20 18
20 25 23
25 % 30 27
30 % 35 32
35 'U 40 36
40 % 45 41
45 % 50 45
More than 50% 50
V Bending of Reinforcing Bars, Less than 3% 4

Crumbling of Core Cone, and
Visible Vertical Level Down
(Distinguished by settlement 5 10 14
and / or Inclination)
10 15 21
15 'U 20 27
20 % 25 34
25 % 30 40
30 % 35 48
More than 35% 50
Note *1 Observed Damage Ratio D.R.

(Frame Structure)
_
Number of Columns of the Same Damage Ra nk
D ^
Number of Columns inspected
(Wall Structure)
^
Total Length of Exte rior Wall of th e Same Damage Rank in Same Direction
D ^
Total Length of Exterior Wall inspected in Same Direction
607
Table 1.3 Identification of Damage due to Total Inclination
1 3 3 6
Inclination Angle ( 0 6 < e fla
e a fia < 6
100 ioo - S ioo 100 « = 100 100
Identification Small or Moderate Severe Collapse

of Damage Slight (Overturned)
Table 1.4 Identification of Damage due to Total Settlement
Maximum Settlement (S:m) S < 0.2 0.2 £ S £ 1.0 1.0 < S
Identification Small or Moderate Severe

of Damage Slight
Table 1.5 Identification of Damage due to Maximum

Damage Rate out of Ones of each Story
Maximum Damage Rate

Edi i, 5 5< Edi 6 10 10 < Edi £ 50 50< Edi
Edi(%)
Identification Severe
Slight Small Moderate
of Damage
Note) Building with a story where members with damage rank V occupy more
than 50% of all members shall be judged as collapsed.
Table 1.6 Permanent Treatments of Damaged Buildings

Based on the Identified Result and Intensity of Suffered
Earthquake Ground Motion
Intensity of Earthquake Indentif ication of Damage of Building
(JMA Degree of Seismic Slight Small Moderate Severe Collapse

Intensity)
Repair Strengthening Strengthening
Less than Strengthening Demolition Demolition
or Equal to V Second Level
Check
Repair Strengthening
Greater than Strengthening Demolition
Reuse Reuse Repair
or Equal to VI Second Level
Check
Table 1.8 Upper Limit Value of Earthquake Damage Index (Z)
Requiring Strengthening for Restoration
2)
Construction Intensity of Earthquake Ground Motion (JMA Degree)
1)
year Less than IV Lower V Upper V Greater than VI
2.
Before 40
1971 20 30 50
After 1971 30 40 50
Note) 1. t-values corresponding to ones after 1971 may be applied to

building of which hoop spacing is less than 10cm, even if it
was constructed before 1971.
In order to divide the intensity V into Lower V and Upper V,
ISOgal of ground acceleration may be used as boundary.
Table 1.9 Coefficient to Reduce Seismic Performance

of Members, n (For First Level Evaluation)
Damage Rank Kind of Memhpr
j. 4^
of Column or Wall Column-M Column-S Wall
I 1.0 1.0 1.0
II 1.0 0.8 0.9
III 0.6 0.4 0.6
IV 0.3 0 0.3
V 0 0 0
Note) 1. These damage ranks are classified according to Table 1.11

2. Column-M : flexure failure column
3. Column-S : column failing in other modes than flexure mode
4. All kinds of reinforced concrete wall, i.e., wing wall with
column and wall out of frame line etc.
Table 1.10 Coefficient to Reduce Seismic Performance

of Members, n (For Second Level Evaluation)
2. Classification of members with damage is fundamentally same as that in the

second level diagnosis before damage. However, in case that observed failure
mode is obviously different from those before damage, obserbed one has priority.
609
C r-i H X »H U M AJ cO
rH
co
>
*H
•H O U rH CO CO O 3 3 U-l E
H W O O co co uj
-a co
*H c C
x
U d C AJ
GJ u-i CO *H 1 co( ) C u-l
d AJ -H o *H -H CL c u u
5 co co co rH QJ
00 Q>
U
CO •
U H
<D cn GJ M UJ X -H
CO
E 3
CL
c 3 0 3
J*2 AJ to
c
o
H
X
to
CO
CO
TO *H
rH CO
O H
CO XI E
3
AJ
co C
H
CO a. o co •H C
U u-i X c
E
3 C 3
E
> m oo > -* AJ 00 C0 <H CO rH
0) 0 r3 C O CO 3 d _C O 15
*H
00
XCO
CO CM *H E X CO
:*
3 C
<H *H
CO x-v
-Dud QJ
CL tH
U g 1 E 5
CO — QJ
-• -H o O
*H <— l O •3
<h
3 00
•H C CO C
CO CJ
AJ
H U-i CL TO
O O CO *H
E .
5
00
J2 u-l C
O *H
rH •h in
.. o C xE AJ 3
C 00 co co a
-X X) 3 U CH C a>
u CO O 3 TO O -H - -x
co TO rH
H 60 (0 C
C
CO
X
CO
5 O' O' w
d cl o X rH —t TO
•H O O
H X 2 J2 rH O
CO AJ ooj:
HU3 x 4)
O) TO
CO U-i
U
O
X X
CO CO
O'
•
—3 X O U H H rH
E I
*
O TO 3 3
O X 00 rH
O
O
c
O u_<
d rH
o
N C
-H CO GJ
U O
H XAJ «H
CO
O O CS O C C C
AJ d -H -H -H
d
•H X O *H
d
O
•H
-h -h
(0 j o
H XXX
HUH
U <U TO CO AJ C X CO
CO 3 C E
OO
- CO
qj d
TO
•H TO
•H
3 HU
H X •H O -H TO H O X U-l QJ
CO AJ AJ X rH CO AJ *H H
QJ TO to TO
C
AJ
CO
O
AJ
3 &
QJ 3 CO 0) V) 0}
CO E W V)
X
rH
Q
CO
-H XXX—
*H
ru cn «j irt
OO
CO
E C
610
Fig. 1.1 General Flowchart of Post
Earthquake Inspection and Evaluation
01
Fig. 2.1 Perspective View of the Naroioka Town Hospital Building
3
C G3 4C G3 5 C G3
Fig. 2. 3 Examples of Detailed Crack Patterns of Columns
612
floor
below
floors basement
the
at
at
line
line
©and®
©and®
between
0
E between
03
U
lx.
wall
1
walls
the
of of
K
CU that
inspected)
«4H Pattern
o
c and
l_
0)
be
4->
4-J
03 not
second
Q- (NoteCCrack
o
03 could
L. the
o
cnj
cî
lx.
ZsPHR
03
Streugthening Method
Position and Number of Members
Streugthened
•Interior Columns of Exterior Frames (3~5th floor) Replacing Cane 24
Column •Exterior Columns of Exterior Frames (3~5th floor) Wire-Mesh

•All Columns of Interior Frames (3— 5th floor) Reinforceng 48
•Isolated Columns at 1st and 2nd floor Wrapping by steel

angles and plate 10
•Hand I frames (B~5th floor) New Walls or Increase
of Thickness 29
Wall •J frame (B— 5t h floor) Replacing Walls

Cone 12
•2 through 7 frames(B~5th floor) New Walls or Increase
of Thickness 32
*12 Boundary beams between new walls were reinforced by many stirrups
0~
0—
Fig. 2.4 List of Adopted Strengthening ConstructionMethod and
Key Plan Showing Location of Shear Walls newly Constructed
614
0—0 O « (0
O O -C
J-. J_ o
—
>
C/3
13
> > 2
-©
-©
of
O O
-a 55
E
>» >»
00 00 Q walls
e I and
columns
of
rank
0 D—i0 0 -O— o-
-H -©
Damage
> JL —
-0
i —0
-Q 9~
IT
- 2.5
>
-9—9- 4-i -© Fig.
>
-0
>
- h -0
LL P ; r~n^
3 e© © 0 © 0 ^<£> © ©©
605
Tab le3 . 1 Identification of Damage of Each Story
by First Level Evaluation
Damage Rank of Column 1 dentif icati
Story 0 I D m IV V Total of Damage

*'
n 1 (13) 2 ( 2) 5 ( 5) 4 ( 4) 0 ( 0) 0 ( 0) 12 (24)
tZ
5 R 8 (53) i 7 ( 8) 42 (21) 33 (17) 100 Noderate
!
di* 0 ( 0) i (0.5) 10 ( 5) 18 (10) 0 ( 0) 0 ( 0) 29 (15.5) (Do)
n 1 ( 7) 2 ( 2) 0 ( 6) 1 ( 1) 8 ( 8) 0 ( 0) 12 (24)
4 R 8 (29) 17 ( 9) (25) 8 ( 4) 66 (33) 100 Severe
di 0 ( 9) 1 (05) 0 ( 5) 5 ( 3) 50 (32) 0 ( 0) 56 (405) (Moderate)
n 1 ( 3) 0 ( 0) 1 ( 8) 1 ( 4) 9 ( 9) 0 ( 0) 12 (24)
3 R 8 (13) 8 (33) 8 (17) 75 (38) Severe

di 0 ( 0) 0 ( 0) 1 ( 8) 5 (10) 50 (36) 0 ( 0) 56 (54.0) (Do)
n 0 ( 0) 1 ( 1) 3 ( 4) 0 ( 1) 0 ( 0) 0 ( 0) 12 (24)
2 R (75) 8 ( 4) 25 (17) ( 4) Small
di 0 ( 0) 0.5 (05) 5 ( 4) 0 ( 3) 0 ( 0) 0 ( 0) 55 (75) (Do)
n 12 (25) i ( 1) 1 ( 2) 0 ( 0) 0 ( 0) 0 ( 0) 14 (28)
1 R 86 (81) 7 ( 4) 7 ( 7) 100 Slight
di 0 ( 0) 0.5 (0.5) 1 ( 1) 0 ( 0) 0 ( 0) 0 ( 0) 15 (15) (Do)
Note)»l n:Number of columns classified into same rank

•2 R = n/2nX 100 (%)
•3 di:Damage Rate(?)of column classified into same damage rank

•4 Values out of parenthesis show the result by exterior columns only
and those with parenthesis show the result by all columns
Table3. 3 Judgement for Necessity of Streugthening using the
First and Secoud Level Seismic DiagnosisNethod
(By Second Level Evaluation)
Eva luat ior

1
Is/Is (S)
'
trenRthcni nr.
Story Level Is Is <>
mJ
i 0.183 0.163 0.89 11 Jnneccssar y
0.265 0.233 0.88 12 do

5 2
1 0.140 0.094 0.67 33 Necessary
0.231 0.145 0.63 37 do

4 2
1 0.145 0.078 0.54 46 Necessary
3 2 0.243 0.112 0.46 54 do
1 0.243 0.234 o.a7 3 L'nnecessa ry
0.273 0.262 C.96 4 do

2 2
616
Table3.2 Sahear coefficient (C) of Each Member and
Those after Damage Calculated by the First
Level Seismic Diagnosis Method
Type of Member
Story Damage
C8 Cs Wi „
w& WS .yO
Rank 172 (2 0) *2. (1 0) EC. Ec'
(W) *1 (T-l 0)* - (Z= 10) (3 0)
2x502 - 3 9 0 0
— E C=
1 2x5(f 0.0 6 2 + 0.0 3 9 +
(1.0. 0.0 4 0) ( 1.0. 0.0 6 2)
(1.0. 0.2 4 1)
0. !
0.7 (0 2 4 1 +
- x 5 0
2 - — 4 9 0 0 0.2 2 1) = 0.4 2 4
5 5
(0.8. 0.0 8 0) (0.9. 0.0 3 5)
n
(12 4 6)
4 X5 (f
— _
in (0.4. 0.0 3 2) E C= 0.3 7 6
Tota 1 12x50* 1 1 x S 0
2 - 3 9 0 0 4 9 0 0
(C. d) (0.2 4 1. 0.2 4 1) (0.2 2 1. 0.1 5 2) (C.0 6 2. 0.0 6 2) (0.0 3 9. 0.0 3 9) Ec'/EC = 0.8 9
6x5 5’ 2x5 5
2 - — -
0. I (1.0. 0.0 8 6) (1.0. C.0 2 9) EC = 0.2 9 1
4 6x55 - - 3 9 0 0
D (1.0. 0.0 8 6) (0.9. 0.0 3 4) EC= 0. 1 9 5
(2 10 0)
- 1 x5 5
2 - - 4 9 0 0
m (0.4. 0.0 0 6) (0.6. 0.0 I 6)
- 8 x 5 5
2 - — E C/E C = 0.5 7
IV (0. 0)
Total 1 2 x 5 5
J
1 1 x 5 5
2 - 3 9 0 0 4 9 0 0
(C. CIt (0.1 7 2 x 0.1 7 2) (0.1 5 8. 0.0 3 5) (0.0 3 9. 0.0 3 4) (0.0 2 3. 0.0 1 6)
2x6 0
2
1x6 0’ - - -
0. ! (1.0. 0.0 2 4) (1.0. 0.0 1 2) EC= 0.2 £ 5
3 ( x 5 u 1 x 6 0
2 - - -
(i C. 0 0 8 5) (0.8. 0.0 1 0) EC= 0.1 4 4
(2 9 7 4)
3 x 6 i? - - 7 14 8
-
“ (0.6. C.C2 2) (0.3. C.0 2 9)
- 9x6 0
2 - — 7 4 6 3 EC/EC=0.5 3 S
IV' (0. 0) (0 3. 0 0 0 8)
Total 1 2x6 C
2
1 1 x S C
2 - 7 14 8 7 4 5 3
(C. Ci (0.1 4 5. 0.1 3 1) (0.1 3 3. 0.0 2 2) (0.0 4 3. 0.0 2 9) (0.0 2 5. 0.0 0 8) 1
9 x 6 f- *x (5 5 /2/ 7 x 6 5
2 _ 2 4 9 6 0
2
C. I + 5 x 5 O
(1.0. 0.1 3 2) (1.0. 0.0 7 0) (1.0. C.l 1 7) (1.9. 0.0 3 0) EC= 0.3 9 3
2 1 xS 5
2
3 x 6 5
2
9 9 0 0
-
3 (1.0. 0.0 1 0) (0.3. 0.0 2 4) (0.3. 0.0 5 3) SC-0.27S
(4 2 5 8)
1 x 6 5
2 - - -
2 (0.5. 0.0 0 6)
1
1 1x6 5!- = x (5 5/2/ 10x65" 9 9 0 0 2 4 9 6 0 1 2 7 2 G

2
Total - 5 x 5 O
tc. c; (0.1 5 2. 0.1 4 6) (0.0 9 9. 0.0 9 4) (0.0 7 0. 0.0 6 3) (0.1 1 7. 0.1 1 7) (0.0 3 0. 0.0 3 0)
1
Note) * 1 W:Total veight inton from top floor to thefloor concerned.

•2 t :L'ltimate average shear stress in kg/csf
assumed for simplified calculationhere,Wl ,V2 and V3 mean

Vail with sur rounding columns at both eud va 11 with column ,
at one eud and wall without sur rounding column , respectively
*3 Total cross sectional area of columns (Acin cnf);or of walls(Aw in an)
*4 Two values within parenthesis mean coefficient to reduce seismic
ca pa city, ^ ,and sheav coef f icent , c of each member ,
617
DENSE STRONG-MOTION EARTHOUAKE SEISMOMETER
ARRAY AT SITES WITH DIFFERENT TOPOGRAPHIC
AND GEOLOGIC CONDITIONS IN SERDAI
BY
Yoshikazu KITAGAWA, Toshihide KASHIMA

and Masaya HIROSAWA
ABSTRACT
Strong ground motions induced by an earthquake occurrence are known to be affected not only by
occurrence are known to be affected not only by the earthquake process itself, but by the propa-
the earthquake process itself, but by the prop- gating pass of seismic topographies and proper-
agating path of seismic topographies and properties of ground surface layers as well. These
ties of ground surface layers as well. These factors apparently influece the type of struc-
factors apparently influence the type of structural damages. Therefore, many efforts, both
tural damages. In order to have reasonable theoretical and experimental, have been directed
aseismic standards, therefore, it is necessary in clarifying the complex effects of surface
to examine how each of those factors would con- soil-layers during earthquake ground motion. The
tribute to the actual ground motions. Based on advent of large-scale structures and important
these circumstances, the dense seismometer array facilities since the 1960s has led to a remarkable
at various sites with difference topographic and increase in research with special reference to
geologic conditions was started in the 1983 fis- the underground earthquake observation. At the
cal year and has been constructed on a five-year present time, more than 200 vertical arrays for
plan in the Sendai area, the Northeastern part earthquake engineering purposes exist in various
of Japan. places in Japan.
The array network is composed of eleven sta- With rapid progress in the field of seismology and
tions with spacing of approximately 3 to 4 km earthquake engineering, demand for observations
on the E-W line passing through the center of of more highly qualified strong motion earthquake
Sendai City. At each station, three observation records is keenly requested on an international
points are arranged vertically, respectively, on scale. Under these circumstances, the Interna-
the surface, at about 20 to 30 m underground, tional Workshop on Strong Motion Earthquake In-
and in the structural base rock lying at a depth strument Array was heldin Honolulu, Hawaii in
of about 50 m. A control and monitoring center 1978, sponsored by the International Association
is set up in the Building Research Institute for Earthquake Engineering. As a goal of the
(BRI) at Tsukuba via public telephone lines. workshop, the delegates approved a resolution on
the immediate establishment of arrays which would

1. INTRODUCTION be capable of resolving the nature of earthquake
source mechanisms, wave propagations, and local

Strong ground motion induced by an earthquake site effects.
613
Around the same time, a national workshop for trict is generally classified into three areas
the deployment and observation of strong- as shown in Fig. 1: (1) the hilly tertiary
motion seismometer arrays was organized, and terrane; (2) the terrace area; and (3) the
the first ten-year plan was made. In May alluvial plain. The oblique NE-SW line pass-
1980, the Science Council of Japan adopted ing near the center of the map is called the
this plan and recommended the Prime Minister Ri fu-Nagamachi tectonic line.
implement it immediately. National and pri- The area west of the tectonic line is charac-
vate research projects on earthquake observa- terized by hilly tertiary terrain and several
tion using seismometer arrays were put into levels of terraces. The surface deposit of
practice in the 1981 fiscal year. Then the this terrace is loam, which is underlain by
purposes of conducting observation were deter- hard clay, gravels, pelite, and shale. The
mined according to the interest of organiza- hilly terrain is either of very hard andesite
tions carrying out those individual projects. or shale, but the surface is covered with loam
Considering this circumstance, BRI began a at several places.
project on earthquake observation by seis- The alluvial plain develops east of the tec-
mometer arrays, as a five-year plan, in the tonic line and is composed mostly of sand,
1983 fiscal year. silt, and gravels. The depth of the tertiary
The Sendai area has a wide variety of topo- base rock varies abruptly near this line.
graphy and geology, and many types and scales There are several areas in the plain which are
of structures exist. The Sendai area suffered covered by very soft peat or mud.
severe damage from the 1978 Miyagi-Ken-Oki The values of shear-wave velocity at intervals
Earthquake (M=7.4), and various studies were of 0.5' in both latitude and longitude in the
performed to examine them. Sendai area are estimated from the soil pro-
In this report, the dense seismometer array. file, results of the penetration tests, and
placed at various sites with different topo- the geological conditions. The profile of
graphic and geologic conditions in the Sendai shear-wave velocity on each section from line
area, is briefly introduced. Particular at- of N38°13'30" to N38°17' are shown in Fig. 3.
tention is paid to the geological conditions 2.2 Microtremor Measurements - The microtremors
in the Sendai area, the plan configuration, are measured at several sites shown in Fig. 1,
and system of seismometer array. with the electromagnetic seismometer having
the natural period of 1.0 sec in obtaining
2. Outline of Geological Condition dynamic properties of soil layers. The area
2.1 Topographical Aspect s- As for the geological where the measurements were performed belongs
conditions of surface soil, the Sendai dis- to the two types of geological conditions, (2)
619
terrace area, and (3) alluvial plain, men- ployment of strong motion seismometer arrays.
tioned above. The array network is composed of eleven sta-
The representative Fourier spectra of micro- tions with spacing of approximately 3 to 4 km
tremors in the E-W and N-S directions are on the E-W line passing through the center of
shown in Fig. 3, where sites G-06 through G- Sendai City, and the N-S line passing through
08, and sites G-17 and G-18 are on the outcrop Nigatake and Oroshimachi, which suffered se-
of the terrace and on the alluvial plain. vere damages of structures in the 1978 Miyagi-
respectively. The peaks at long periods more Ken-Oki Earthquake. The layout of array con-
than 1.0 sec reflect the deep ground charac- figuration is shown in Fig. 6.
teristics, whereas the peaks at short periods At each station, three observation points are
less than 1.0 sec reflect the shallow ground arranged vertically; one on the surface, one
ones. at about 20-30 m underground with the shear-
2.3 Distribution of Dynamic Characteristics for wave velocity of 300-400 m/s, and one in the
Underground Structure - The structural base structural base rock having the shear-wave
rock with shear-wave velocity of 700 m/s was velocity of 700 m/s and lying at a depth of
set up to obtain the ground characteristics in 50-60 m. A control and monitoring center is
the range of short periods less than 1.0 sec. set up in the BRI. The center is connected to
The predominant period and magnification fac- a sub-center at Sendai, and the sub-center is
tor of the elastic media were calculated at an further connected to each observation station
interval of 0.5 in both latitude and longitude via public and exclusive telephone lines.
by Haskell's Method. Figure 7 shows a block diagram of the dense
The distribution maps for the predominant per- strong motion seismometer array system. Fig-
iod and magnification factor of the under- ure 8 shows a plan and a section of the Stan-
ground structure above the structural base dard observatory.
rock are shown in Figs. 4 and 5, respectively. The array observation system consists essen-
As shown in these figures, the predominant tially of an accelerometer sensor, an ampli-
period is around 0.2-0. 3 sec in the terrace fier, an A-D converter, a pre-event memory, a
area and 0. 3-0.8 sec in the alluvial plain. digital magnetic tape recorder, and a time-
Similarly, the magnification factor is about code generator. In order to obtain large dy-
2-3 in the former and 3-5 in the latter. namic range and high resolution in recording.
a digital system is used. Specifications of
3. ARRAY CONFIGURATION AND OBSERVATION SYSTEM the array observation system are as follows:
The Sendai area is designated as one of the 3.1 Accelerometer-Triaxial force— bnl.ino’ tvpr 1
•
ten high-priority sites in Japan for the de- used. Recording and overall frequency ranqes
are ± 1G and 0.05 to 30 Hz.
620
3.2 Amplifier and A-D Converter (ADC) -16 bit ADC criteria. Two observation stations have been
is used. The final dynamic range is 96 dB. installed respectively at Miyagino (MIYA
The signals from the accelerometer are digi- station) and Nakano (NAKA station) primary
tized AD and a sampling rate of 1/100 or 1/200 schools. In addition to these stations,
sec. another station is scheduled to be installed
3.3 Pre-Event Memory-Delay device utilizing IC in fiscal year 1985 at the Tamagawa secondary
memory with 5 sec delay time is adopted. school (TAMA station).
3.4 Time-Code Generator-Quartz with an accuracy 4.1 MIYA Station-The station is located near the
of 10"'7 to 10“8 is used for tracking time. eastern part of the Ri fu-Nagamachi tectonic
The absolute time is corrected automatically line shown in Fig. 6. Locations of the seis-
to the Japanese standard time within an error mometers and observatory are shown in Fig. 9.
of ±5 m sec by receiving the time signal Photo 1 shows overviews of the observatory.
broadcasted by NHK (the Japanese Broadcasting The soil profile and the result of seismic
Corporati on) prospecting tests are shown in Fig. 10. As
3.5 Digital Date Recorder-Digital magnetic tape seen in this figure, the underground structure
with 9 tracks, half-inch in width and 1600 BPI is classified into eight soil layers in P- and
in recording density is used for the accelero- S-wave velocities. Three seismometers are
graph of a signal set. placed along the vertical line at depths of 1
A telemetric monitoring device is used. The m, 22 m, and 55 m below the ground level.
device transmits data, such as peak accelera- The transfer function (Uq/Uj N ), defined as the
tion, triggered time, duration of recording, ratio of the wave on the ground surface, Uq,
and the operating condition of the observation to the incident wave, Uj^, for the underground
system. The control system enables one to structure at the observation site is shown in
calibrate the recording sensitivity and to Fig. 16.
adjust remotely the triggering level of the 4.2 NAKA Station-As shown in Fig. 6, the station
recording device at each station. is located at the far eastern part of the
Ri fu-Nagamachi tectonic line, which is the al-
4. OBSERVATION STATION luvial plain near the lower reaches of the
The dense strong-motion earthquake seismometer Nanakitada River and the Sendai Bay. The
array project was started in fiscal year 1983, locations of the seismometers and observatory
for the purposes of studying the effect of are shown in Fig. 12. Photo 3 shows overviews
geological and topological conditions of of the observatory. The soil profile and the
earthquake ground motions and establishing the resulting seismic prospecting test are shown
standard earthquake motions for dynamic design in Fig. 13. As seen in this figure, the un-
derground structure is classified into 6 and 9 6. CONCLUSIONS
soil layers in P- and S-wave velocity, respec- The dense seismometer array at various sites
tively. Three seismometers were placed along with different topographic and geological
the vertical line at depths of 1 m, 29.5 m. conditions is briefly introduced in this pa-
and 61 m below the ground level. The transfer per. The numer of observation sites has in-
function ( Uq/U in) is shown in Fig. 14. creased year by year, and high-quality array
data will be accummulated in the near future.
5. EXAMPLE OF EARTHQUAKE RECORD In relation to the increase of data, dissemin-
Operation of the MIYA station was initiated in ation of the data is necessary for cooperation
March 1984. A total of four earthquakes have and coordination between the earthquake obser-
been observed. As an example, the E-W and N-S vations under individual projects. Because of
components of the acceleration records, during this, establishment of a national data bank
the Southeast--Of f Boso Peninsula Earthquake system to aid in exchange is hoped for in the
of September 1984 (Date and Time: 02.03 a.m.. near future.
September 19, 1984, Magnitude=6.8, Epicenter
Lat. 33°56‘N, long. 141 °51 E 1
, Focal depth: 46 7 . ACKNOWLEDGMENTS
km) are shown in Fig. 15. As seen in this The dense strong-motion earthquake seismometer
figure, the maximum values of accel erations at array observation is implemented by a work
the ground level are 7.7 gal and 7.2 gal in group in the Subcommittee of Earthquake Obser-
the E-W and N-S directions, respectively. The vation in and around buildings organized in
magnification factors of the surface maximum BRI. Members of the working qroup for the
acceleration relative to that observed 54 m array observation are Y. Kitaqawa, Head of
below the ground level are 1.7 times in the E- Structural Dynamics Div.; Y. Yamazaki , Head of
W and N-S directions. Large-Scale Structural Testinq Div.; H.
The Fourier spectra of the accel erograms are Mizuno, Senior Research Staff of USEE; T.
shown in Fig. 16. Figs. 16 and 17 show the Okawa, Research Staff of IISEE; and T. (Cash-
spectral ratio and the phase lag between the ima, Assistant Research Staff of Structural
surface and underground layer 54 m below the Dyanmics Div. In order to discuss the array
ground level. Judging from these figures, the observation plan, the committee of dense
predominant frequency of the soil layers is strong-motion earthquake seismometer array
around 2 Hz and 7 Hz. These values correspond (Chairman: Y. Osawa, Professor of Toyko Uni-
to the values in the transfer function shown versity) is organized in KKSK. The authors
in Fig. 11. wish to express their sincere thanks to mem-
bers of the committee, working qroup, and
persons concerned.
8. REFERENCES
(1) Kitagawa, Y., and Tanaka, T., Observation

of Underground Earthquake Motions in Japan,
Some Recent Earthquake Engineering and Prac-
tice in Japan: Japanese Nat. Comm, of IAEE
pp. 14-26, 1980
(2) Tanaka, T., and Kitagawa, Y., Strong Motion

Earthquake Seismometer Arrays in Japan,
Some Recent Earthquake Engineering and
Practice in Japan: Japanese Nat. Comm, of
IAEE, pp. 17-31, 1984
(3) Report on the Damage by 1978 Off-Miyagi

Prefecture Earthquake (in Japanese):
Report of BRI ,
No. 86, pp. 75-81, 1979
(4) Kitagawa, Y., and Matsushima, Y., Evalua-

tion of Dynamic Ground Characteristics
and Seismic Microzoning: Proc. of Third
South Pacific Regional Conference on

Earthquake Engineering, New Zealand,
pp. 73-84, 1983
6Z5
tfSBa
i
Fig. 6 Layout of Array Configuration in Sendai Area
625
! 2700 J
PLAN
Fig. 8 Plan and Section of Standard Observatory
Soil Depth P-wave 5-wave Dens ty 1
Velocity Velocity
J
(ml (m/s 1 («/s) t t/m )
0.0
ir 380 iKL. 1.6
—
580 J50
1000
1.9
7.0 4 30
1200
Gravel 11.0
wi tn
Clay
480
2.0
ir 22 0 1600
2b .5
Tufa
640
30.
1 o 1
Mudstone
39 9 Ibuo
band- 46 0 1 9
stone
.mi 4mO
Muds tony
u
Shale 1 9LHJ 630 1 0 ,
MIYAGINO site
^ U«l(On oi
9 Location of Seismometers and Fig. 10 Soil Profile

Observatory at MIYA Station at MIYA Station
626
Fig. 11 Transfer Function of Soil-layers at MIYA Station
NISHIHARA Park Soil Depth P-wave S-wave Oensi ty

Veloci ty Veloci ty
J
(m) (m/s) (m/s) lt/m )
320 94 1 .7
2.2
1 .6
3.8 170
1 .8
5.6 970
Fine
Sand
215
1.9
15.5
1550
18.7 170
Clay 1.7
F i ne
23.8
Sand 1 .9
26.0
225
Clay
1.7
28.8
29.6
340 1 .9
1650
Sandy 320 1 .8
Clay
Clayey 2.0
Grave 1
Fig. 12 Location of Seismometers and Observatory 2150 490
at NAKA Station 57.8

1 .9
61 .
2600 720 2.0
NAKANO site
^ Lorn Mon of
Fig. 13 Soil Profile

at NAKA Station
627
Fig. 14 Transfer Function of Soil-layers at NAKA Station
GL-lm COI* 9/19/04 Off BOUbOU Pill
EW Component
GOII 9/I9/U4 Off UOUSOU I'M.
NS Cpmponent
001/ 9/19/64 OH BOllbOU l*» h
UD Conoonent
GL-54m CS4* 9/19/64 Off BOubOU 1*1
EW Component
Gb4f 9/19/64 Off UOu'.n r

\SU iU!
1W
'frVfiv- 'f j)>\ A
NS ComDoncnc
b4/ 9/19/64 Off on.r.nu r\ I,
m* UD Component
Fig. 15 Accelerograms at GL and GL-54m
628
FOURIER SPECTRUM FOURIER SPECTRUM
G54 X MTGN 9/19/84 icm/sec) G54T MTGN 9/19/84 icti/SEC)
bPLGTIUJII
HIUHLtil
FOURIER SPECTRUM FOURIER SPECTRUM

GO X MTGN 9/19/84
1 icm/secj GOIT MTGN 9/19/84 (cn/SEO
SPtCIMUM
f-OUMILM
(a) EW Component (b) NS Component

Fig. 16 Fourier Spectra of Accelerograms
SPECTRUM RATIO ( T / X ) SPECTRUM RATIO ( T / X )
X-COMP. = G54X MTGN 9/19/84 X-COMP. = G54T MTGN 9/19/84
RATIO

Fig. 17 Fourier Spectral Ratio between GL and GL-5Am
629
'’I! AST-. .. \ G :
X - T i PHASE LAG ( X - Y
< -l liMP. -54X MTGN 9/19/84 X -COMP. = GS4 T MTGN 9/19/B4

Fig. lb Phase Lag :ween GL and GL-54n
Photo 1 Overviews of Observatory Photo C Overviews of Observatory

at IIIYA Station at NAKA Station
630
INFLUENCE OF SOIL CONDITIONS UPON RESPONSE SPECTRUM
OF
STRONG GROUND MOTION

BY
Izuru OHKAWA*
Yuji ISHIYAMA*
ABSTRACT
The intensity of the earthquake ground motion is hard ground and presented the following equa-
often expressed in terras of some parameters such tion.
as maximum acceleration, maximum velocity and

_ in 0. 47 2M-( 1.97+1. 80/X) log (X)+
response spectra. These characteristic values .
mnax iU
are generally a function of eathquake magnitude, (2.20-11.1/X)
epicentral distance, hypocentral distance, and Many other researchers proposed so-called atten-
the soil conditions of the sites where the ation formulae of maximum amplitudes of strong
earthquake ground motions are observed. earthquake motions, Ohsaki et al (1980), at
In this paper, the response spectra are related first, classified the records according to the
with especially the soil conditions, i.e., shear wave velocities at observation site into
stiffness of soil deposits etc., and the magni- the regression analysis for records of each
tude and the distance from the fault. group. They presented the formulae to obtain
Keywords: Earthquakes; Response spectrum; Soil the maximum velocity.
<>.54M-1.311ogCX)-0.95
conditions; Strong ground motion 'max ' 10 for soft r „ ck
= 0* 65M - 1 - 361 °glll)' 2 -00
1. INTRODUCTION V max 10 Cor hard rock
The maximum amplitudes and the response spectra Watabe and Tohdo (1981) obtained the formulae
have been related with eathquake magnitude and for maximum acceleration and maximum velocity,
the distance from the fault. These relation- using the records observed on hard rock with
ships ever presented are originated from the shear wave velocity greater than 0.6 km/sec.
definition of the earthquake magnitude. In _ xu 0 .44M-1 381og(X)+l .04
.
nmax
Japan, Kanai (1968) presented the following V
v
max
=10 0.607M-1.191og(X)-1.40
semi-empirical relationship to provide the P.W.R.I. (1983) presented the equations to give
maximum acceleration and maximum velocity amp- the maximum amplitudes by analyzing with many
litudes at so-called seismic base layer as, cases of mathematical models involving the soil
conditions. For example,
1
‘GiM-CO .66=3.60/X) log(X)+(0 167-1. 83 /X)
.
10°’ 221M
x =T i()0
Amax= 107 3
v =ln 0. 61M-(0. 66=3. 60?X)log(X)-(0. 631+1. 83/X) )~1* 2 51
A +30 for stiffer soil
max (
where Amax =227.3 10°* 308M ( A +30)' 1 * 201 for med. soil
-1 * 208
Ajnax : maximum acceleration in cm/sec 2 Amax =403.1 io°* 262M ( A +30) for soft soil
Am ax : maximum velocity in cm/sec where, A : epicentral distance in kilometer
T : fundamental period of the ground in secoqd As for response spectra, the equations with
M : earthquake magnitude
X : hypocentral distance in kilometer *International Institute of Seismology and
Ohsaki and Watabe(1977) performed the regression Earthquake Engineering, Building Research
analysis by using the accelerograms recorded on Institute, Ibaraki Prefecture, 305, JAPAN
651
the coefficients of same types except being 2.1 Analysis I
functions of period T are assumed are obtained The earthquake records are classified into
by the regression analyses. three groups according to the shear wave
Kobayashi and Nagahashi (1976) presented the av- velocity Vs of the soil layer of the obser-
erage response spectra at the seismic base layer vation sites. The classifications are made
the shear wave velocity of which is larger than as (1) Soft: Vs less than 0.75 km/sec, (2)
3.0 km/sec by using the accelrograms observed at Hard: Vs greater than 1.5 km/sec. The num-
the ground surface. They adopted the equation bers of available accelerograms classified
of the following type. into these groups are 30, 18 and 60 for Soft,
S v ( T) = 10
a(T)M - b(T)log(X) - c(T) Hard and HH
0 ,
respectively.
Hisada et al.(1978) presented the design spectra Then, the following equation is used to ex-
for stiff structures such as nuclear power reac- press the average response spectra for each
tor buildings. The shear wave velocities of the group at each natural period.
“ b(T) log(X) - c(T)
ground to which this spectra can be applied are pSv(T) = 10 a(T)M
greater than 0.7 km/sec. where
Thereafter, many researchers are concerned about pSv : pseudo velocity response spectrum in
the influence of soil (rock) conditions upon the cm/sec
response spectra of earthquake ground motion. M : earthquake magnitude (JMA)
X : hypocentral distance in kilometer
Ohsaki et al.(1980) reported the results that a,b,c: coefficients determined by regression
the velocity response spectra of the records on analyses
hard rock sites (Vs greater than 1.0 km/sec) are T : natural period of SDOF system in
about half of those on soft rock sites (Vs less second
than 1.0 km/sec). The accelerograms are classified as summar-
ized in Table 1. In that table, magnitude.
Ohta et al.(1982) presented the similar results epicentral distance, focal depth and shear
that the response spectra for hard rock are sim- wave velocity are also described. In Fig. 1
ilar than those for soft rock, from the regres- the distribution of the data with respect to
sion analyses considering the site effects. earthquake magnitude and epicentral distance
is shown. It should he noted here again that
P. W.R.I . ( 1983) presented the equations to give number of data of each classification and the
the acceleration response spectra with many distribution of magnitude and distance are
cases of mathematical models involving the soil different.
conditions of observation site. Fig. 2 shows the coefficients a(T), and b(T)
and c(T) in the above equation derived from
2. COMPARISON OF RESPONSE SPECTRA OF DIFFERENT the Soft data using the least square method.
SOIL CONDITIONS Fig. 3 and Fig. 4 also show those coefflc-
Two types of analyses are performed in this ients derived from the Hard and HH rock data,
part. One is based on the analyses of the respectively. From those figures, the coef-
accelerograms which are classified into three ficient a(T) the multiplier of magnitude takes
groups according to the soil conditions in a value from 0.4 to 0.8 and is larger at the
advance. The other is based on the analyses longer period. The coefficients b(T) and r
with dealing those data in the gross applying (T), the multiplier of the hypocentral dis-
the model involving the effect of soil condi- tance and the constant, respectively, have the
tions in the equation. maximum and minimum. Using these coefficients
632
we plotted the response spectra corresponding ficient c(T), when the magnitude and distance
to the magnitudes 6 and 7 ,

the hypocentral dis- are same. Fig. 9 shows the ratios of the spec-
tance 100 and 200 km in Fig. 5(a) - 5(d). In tra of those for Vs-1.0 km/sec and 1.5 km/sec to
each figure, the calculated spectra for Soft, that for Vs-0.7 km/sec. From this figure, when
Hard and HH soils are. plotted for the quantita- the shear wave velocity of the ground is 1.5 km/
tive comparison. sec Vs, the spectral ratio becomes 0.3 to 0.5
From these figures, it is seen that the spectral

values are smaller, when the soil becomes stif- 3. DISCUSSIONS AND CONCLUSIONS
f er. From the Analysis I and Analvsis II, it can be
When we compute the ratio of the spectra to generally said that the spectral values are re-
those for Soft rock and plot the ratio in Fig. duced when the shear wave velocity of the ground
6, the ratio of the Hard rock to the Soft rock becomes higher. Though the spectral value for
varies from 0.5 to 1.0, whereas the ratio of the Hard rock partly exceeds that for Soft rock,
HH rock to the Soft rock varies from 0.2 to 0.6. this may be due to the shortcomings of the re-
cords to be used, as is mentioned before. The
2 .2 Analysis II feature of the Analysis II is the introduction

The shear wave velocity of each observation site of Vs term into the equation. The ambiguous ex-
is known, then we can make the shear wave velo- pression such as "soft” or "hard" which is fre-
city represent the soil condition. In this case quently used in the past research works is in-
of analysis, the accelerograms are not classifi- convenient for the practical applications. But
ed. The mathematical model of the empirical for- the authors are not fully convinced of the re-
mula we adopted herein is as follows: sults. There are many problems to be solved,
pSv(T) = 10 a ( T)M-b( T) log(X)-c( T) log(Vs)-d( T) i.e., the analyses by using more number
where of data, and the theoretical explanation of the
A,b,c,d: coefficients determined by the regres- property, etc.

sion analyses. It should be noted that
c(T) has different meaning from the ob- 4. REFERENCES

previous analysis. 1. Kanai ,K. Suzuki
, ,T. ( 1968 ), "Expectancy of
Vs: shear wave velocity at the observation the maximum velocity amplitude of earthquake mo-
site in km/sec, other parameters are tions at bed rock,” B.E.R.T., Vol. 46
,
same as used in the Analysis I. 2. Ohsaki ,Y .
,Watabe ,M. ( 1977 ) "On the peak
The regression analyses are applied to the model values of earthquake ground motions," Proc. of
by using whole data in Table 1. The coefficients Annual Meeting of A.I.J.
a(T) ,
b(T) , c(T) and d(T) thus obtained are shown 3. Ohsaki ,Y. ,Sawada,Y. et al.(1980), "Spec-
in Fig. 7. The coefficient a(T) shows a similar tral Characteristics of Hard Rock Motions,"
feature with that obtained in Analysis I. How- Proc. of 7th W.C.E.E., Istanbul, Turkey
ever, coefficients b(T) and constant term d(T) 4. Watabe ,M. ,Tohdo , M.(1981), "Research on
are changing, because of the introduction of the the design earthquake ground motions," Transac-
Vs term. tions of A.I.J. ,
Vol. 303
Using these coefficients, we plotted the average 5. P.W.R.I. (1983), "Estimation of peak
response spectra of magnitudes 6 and 7, hypocen- ground motions and response spectra," Research
tral distances 100 and 200 km, in Fig. 8(a) - 8 Material, P.W.R.I., v 0 l. 1993
(d). It is apparent from the above equation that 6. Kobayashi ,H. ,Nagahashi,S.(1976), "Amplifi-
the ratio of the spectral values between the dif- cation characteristics of the ground and spec-
ferent shear wave velocities depends on the coef- tral characteristics of earthquake motions on
653
the seismic bedrock inferred from spectral char-
acteristics of earthquake motions observed on
the ground surface," Trans, of A.I.J., Vol. 240
,
7. Hisada,T. et al .( 1978) "Design spectra
for stiff structures on rock," Proc. of 2nd Mi-
cro zonation Conference, Vol. 3
,
8. Ohta,T. et al .( 1982) "Characteristics of
peak acceleration and response spectra for ob-
served records at hard and soft rock sites,"
Proc. of 6th J.E.E.S.
634
Table 1 Earthquake Records Used in This Study
HH (Hard--Hard) Rock
Soft Rock
DEP. DIS. Vs DEP. DIS. Vs
NO NAME M (km) (km) (km/sec) NO. NAME M (km) (km) (km/sec
1 Cl-1-2 4.9 0 145 0.60 1 G2-2-2 5.3 10 51 1.50
2 Cl-2-2 4.7 20 170 0.60 2 G2-3-1 4.6 20 220 1.50
3 Cl-3-2 4.4 60 100 0.60 3 G2-3-2 4.6 20 220 1.50
4 Cl-4-2 4.4 40 120 0.60 4 G2-4-1 4.3 10 80 1.50
5 Cl-5-2 4.2 40 85 0.60 5 G2-4-2 4.3 10 80 1.50
6 Cl-6-1 5.8 80 270 0.60 6 G2-5-2 6.1 10 52 1.50
7 C 1-7-1 4.7 50 50 0.60 7 G2-6-2 5.5 10 55 1.50
8 Cl-8-1 6.2 40 180 0.60 8 G2-9-1 4.0 0 56 1.50
9 Cl-9-1 6.0 50 145 0.60 9 G2-9-2 4.0 0 56 1.50
10 Cl-10-1 7.0 60 475 0.60 10 Tl-2-1 3.9 10 28 1.50
11 Cl-11-1 7.0 70 460 0.60 11 Tl-2-2 3.9 10 28 1.50
12 Cl-12-1 5.3 50 50 0.60 12 Tl-4-1 5.1 10 51 1.50
13 C 1 — 1 3— 7.2 50 460 0.60 13 Tl-5-1 5.1 10 94 1.50
14 Cl-14-1 5.3 50 60 0.60 14 Tl-6-1 4.7 0 112 1.50
15 Cl -1 5-1 6.3 40 110 0.60 15 Tl-7-1 6.1 10 137 1.50
16 Cl-16-1 5.9 30 80 0.60 16 Tl-7-2 6.1 10 137 1.50
17 C2-1-1 5.4 10 90 0.60 17 N1— 1 — 1 6.1 10 67 2.00
18 C2-2-1 5.5 40 155 0.60 18 Nl-1-2 6.1 10 67 2.00
19 C2-3-1 6.0 50 150 0.60 19 C6-1-1 4. 7 10 8 2.00
20 C2-4-1 4.1 50 35 0.60 20 C6-2-1 4.4 0 6 2.00
21 C2-5-1 5.5 40 55 0.60 21 C6-2-2 4.4 0 6 2.00
22 C2-6-1 6.3 50 320 0.60 22 C6-3-1 5.0 0 8 2.00
23 C2-7-1 5.3 50 50 0.60 23 C6-3-2 5.0 0 8 2.00
24 C2-8-1 6.3 40 100 0.60 24 C6-4-1 6.6 0 70 2.00
25 C2-9-1 5.9 30 85 0.60 25 C6-4-2 6.6 0 70 2.00
26 C4-1-1 7.2 50 245 0.75 26 C6-5-1 4.0 0 8 2.00
27 C4-1-2 7.2 50 245 0.75 27 C6-6-1 6.0 10 90 2.00
28 C4-2-1 6.9 10 110 0.75 28 C6-6-2 6.0 10 90 2.00
29 C4-2-2 6.9 10 110 0.75 29 El-1-1 5.1 0 85 1.50
30 C5-1-1 6.9 10 120 0.55 30 El-1-2 5.1 0 85 1.50
31 El-2-1 4.5 20 90 1.50
32 El-2-2 4.5 20 90 1.50
33 El-3-1 5.5 0 145 1.50
34 El-3-2 5.5 0 145 1 . 50
35 El-4-1 4.0 0 56 1.50
36 El-4-2 4.0 0 56 1.50
37 FI — 1 — 1 6.6 0 145 2.20
Hard Rock 38 G1—1—1 4.2 0 50 1.80
39 Cl-2-2 5.1 10 45 1.80
40 Kl-1-2 6.6 0 95 1.60
DEP. DIS. Vs 41 Pl-2-1 5.1 5.8 24.5 1.50
NO. NAME M (km) (km) (km/ sec 42 Pl-2-2 5.1 5.8 24.5 1.50
1 Dl-1-2 6.6 0 160 1.00 43 Pl-3-1 4.3 18.3 21.5 1.50
2 Dl-2-1 5.5 30 40 1.00 44 Pl-3-2 4.3 18.3 21 . 1.50
3 Dl-2-2 5.5 30 40 1.00 45 Pl-4-1 5.8 8.0 26 1.50
4 Dl-3-1 6.9 10 60 1.00 46 Pl-4-2 5.8 8.0 26 1.50
5 Dl-3-2 6.9 10 60 1.00 47 PI —5—1 5.4 8.3 27.2 1.50
6 Al-2-2 7-.0 60 330 1.30 48 Pl-5-2 5.4 8.3 27.2 1.50
7 Al-3-1 6.4 80 270 1.30 49 Pl-6-1 4.7 9.4 24.7 1.50
8 A 1 —3 — 2 6.4 80 270 1.30 50 Pl-6-2 4.7 9.4 24.7 1.50
9 Al-4-1 5.4 60 240 1.30 51 Q1—1—1 4.9 13.9 33.2 1.50
10 Al-4-2 5.4 60 240 1.30 52 Q1—1—2 4.9 13.9 33.2 1.50
11 Al-5-1 7.4 40 440 1.30 53 Ql-5-1 4.0 10.6 15.6 1 . 50
12 Al-5-2 7.4 40 440 1.30 54 Ql-5-2 4.0 10.6 15.6 1.50
13 Al-6-1 7.1 40 505 1.30 55 Q 1—8— 4.3 9.0 10.2 1.50
14 Al-6-2 7.1 40 505 1.30 56 Ql-8-2 4.3 9.0 10.2 1.50
15 Al-7-1 6.6 30 500 1.30 57 Ql-9-1 4. 8.7 3.5 1.50
16 Al-7-2 6.6 30 500 1.30 58 Ql-9-2 4. 8.7 3.5 1.50
17 Al-8-1 6.5 100 120 1.30 59 Ql-10-1 4.2 5.0 33.3 1.50
18 Al-8-2 6.5 100 120 1.30 60 Q 1 — 1 0—2 4.2 5.0 33.3 1.50
M magnitude, DEP. : focal depth, OIS. : epicentral distance, Vs : shear wave velocity
635
SOFT - * , HARD - . HARD HARD - O
8
a
* *
LU
eg —* i
O
ZD
^
< 6 -V
LlJ
*
c 00
o 5
HI
I—
cc
c
LlJ
10 20 50 100 200 500 1000
EPICENTRAL DISTANCE (KM)
Fig. 1 Distribution of Earthquake Records
PERIOD (SEC)
Fig. 2 Coefficients a(T), b(T), c(T) in Analysis I (Soft Rock)
636
COEFFICIENT
OF
VALUE
COEFFICIENT
OF
VALUE
PERIOD (SEC)
Fig. 4 Coefficients a(T), b(T), c(T) in Analysis I (HH Rock)
657
Analysis
in
Spectra
(M=6.0,X=200km)
Computed
5(b)
050
0-
=
Fig.
Analysis
in
Spectra 100km)
0,X-
imputed
1=6.
050
638
\
Analysis
in
Spectra X=200km)
0,
.
Computed
(M=7
5(d)
050
0.
= Fig.
Analysis
in
Spectra 100km)
X=
0,
.
Computed
(M=7
5(c)
.050
Fig.
639
RATIO
SPECTRAL
PERIOD ISEC)
Fig. 6(a) Spectral Ratio to the Soft-Rock Spectra

(M=6 0 X=100km)
. ,
RATIO
SPECTRAL
PERIOD (SEC)
Fig. 6(b) Spectral Ratio to the Soft-Rock Spectra

(M-7.0,X=200km)
WO
COEFFICIENT
OF
VALUE
PERIOD (SEC)
Fig. 7 Coefficients a(T), b(T), c(T) and d(T)

in Analysis II
PERIOD (SEC)
Fig. 9 Ratios to the Vs=0. 7km/ sec-Spec t ra

(numeral indicates shear wave velocity in km/sec)
641
II
Analysis
in
Spectra X=200km)
0,
.
Computed
(M=6
COEFFICIENTS
8(b)
4
Fig.
II
Analysis
in
)
Spectra
OOkm
1
X=
0,
.
Computed 6
M=
(
COEFFICIENTS
8(a)
Fig.
642
o
o
o
COEFFICIENTS
C/3
C/1
X
03
C
<
c
03
J-i ^
AJ E
u ^
Q) O
Q- O
CO t-H
II
T3 X
01
•U O
D •
Q. r-
E ll
O 2
w
O
COEFFICIENTS
00
Ci4
(3NIVI) UI3013A
643
REGIONAL TSUNAMI WARNING SYSTEM (THRUST)
BY
Eddie N. Bernard and Richard R. Behn*
ABSTRACT
The National Oceanic and Atmospheric Administra- generation of a tsunami. A gap, the 4 to one
tion (NOAA) has embarked on a 3-year project, hour after tsunami generation, exists in the
Tsunami Hazard Reduction Using System Technology present warning structure.
(THRUST), to create a pilot regional tsunami
warning system. THRUST, which is being funded The purpose of the THRUST (Tsunami Hazard
by the Agency for International Development Reduction Using Satellite Technology) project
(AID) and developed by the Pacific Marine Envi- has been to examine existing technology and to
ronmental Laboratory (PMEL) for the country of ascertain if an early warning system can be
Chile, is utilizing existing instrumentation designed that would fill the "gap" in nations
connected to satellite communication to es- with minimal or no regional warning system.
tablish an early warning system. All pre-event Specifically, the objective is to design, assem-
work has been completed including hazard map, ble, test, install and evaluate a system that
numerical modeling simulations, and the creation can deliver early warnings to a developing
of an emergency operating plan. Instrumentation country (Bernard et al., 1982).
design has been completed and bench testing is
expected to begin in summer of 1985. The Valparaiso, Chile area was selected as the
site in which the THRUST project will take place
because of its great tsunami threat and its
1. INTRODUCTION
existing warning infrastructure.
One of the most destructive natural hazards
within the Pacific Basin is the seismic sea
2. THRUST SYSTEM
wave, or tsunami. More than two million people
support themselves or reside in the tsunami Warning systems can be divided into two time
prone areas of the Pacific Basin. Since the frames, the pre-event stage--the period (days,
early 1850's, more than 70,000 of these people weeks, months, years, etc.) prior to the event
have lost their lives due to the devastation of (tsunami); and the real-time stage--the first
tsunamis (Iida et al . , 1967). hours after the event (tsunami generation). The
pre-event time frame of this warning system

Developing countries within the Pacific Basin, determines the potential danger the event pre-
with minimal or no regional warning system, sents to that area and the solutions to these
cannot be alerted of tsunamis originating close dangers. The real-time time frame's efforts are
to their shores. The present operating limita- focused on the collection and analysis of seis-
tions of the existing Pacific-wide tsunami warn- mic and water level data and the dissemination
ing network (the Pacific Tsunami Warning Center of the warning information.
(PTWC) near Honolulu, Hawaii, USA) allows a *Pacific Marine Environmental Laboratory
warning to be issued 4 to one hour after the National Oceanic and Atmospheric Administration
Seattle, WA 98115 USA
m
The technologies that are being examined in Utilizing the above data files, a "Tsunami in
THRUST can be categorized into three areas: the Pacific Basin" map has been assembled and
data collection, data analysis, and information published. This map details earthquake origin
dissemination. - and magnitude, tsunami run-up height, related

deaths, dollar damage, etc. of all tsunamigenic
The THRUST system, being an early warning earthquakes in the Pacific Basin from 1900 to
system, can be conceptually described as com- 1983. The data files have also been utilized
bining the time frames of a warning system with for validating the numerical models and
the areas of technology to form a working assisting in the development of the Standard
matrix. Utilizing Table 1, we can describe the Operating Plan.
system in the fallowing fashion (Bernard et al .
1983).
3 . 2 Data Analysis
This task has produced computer simulations
3. PRE-EVENT which have provided estimates of potential in-

undation levels, flood hazard areas, and worst
3 . 1 Data Collection case effects of observed and potential
This task consisted of compiling, cataloging and tsunamis
synthesizing all available data relating to

tsunami effects in the Pacific Basin, concen- The SURGE II model was applied to all modeling
trating on the country of Chile. efforts (Reid et al . ,

1977). A h km grid was
utilized with a seafloor uplift (within the

To date, three files utilizing this data have grid) being the source of the tsunami. These
been assembled. The first file is a pre- simulations provided information that the threat
twentieth century file that include 382 events to Valparaiso from local tsunamis approaching
(categorized into run-up heights, magnitudes, from the west and/or northwest is quite great.
origin, etc.) of which 178 have caused death or Data obtained from the Chilean Navy Hydrographic
destruction. The second file is a twentieth Institute (IHA) in Valparaiso, Chile on the
century file which includes 405 events May 22, 1960 tsunami in Corral, Chile was
(categorized as mentioned above). The third utilized to verify the model (Hebenstreit
file is a file of all Chilean tsunamis since the 1984a).
16th century. This file contains 249 events of

which 34 have caused death or destruction.
THRUST SCHEMATIC PILOT STUDY

FUNCTIONAL AREA
TIME FRAME DATA COLLECTION DATA ANALYSIS DISSEMINATION
PRE-EVENT DEVELOP TSUNAMI EVALUATION OF DEVELOPMENT OF

DATA BASE HAZARD USING EMERGENCY OPERATIONS,
SIMULATIONS PROCEDURES
REAL-TIME SENSOR DEVELOPMENT OPERATIONAL INTEGRATION OF

INSTRUMENT "PREDICTIVE" EARLY WARNING
+ MODEL DEVICE INTO
PROCESSING EMERGENCY
+ SYSTEM
TRANSMISSION
Table 1
645
Results of these numerical simulations will be THRUST will utilize a satellite-based communi-
used to formulate the evacuation plan in the cations system, in the form of the Geostationary
Standard Operating Plan (SOP) and for the Operational Environmental Satellite (GOES)
development of the Real Time Processor (RTP). system. Utilizing the GOES system will allow
the lag time between the event and the receipt
of initial data to be reduced to the order of

3 3 Information Dissemination
.
minutes — enough time to provide regional early

This task includes the development of the warnings (see Figure 1).
Standard Operating Plan (SOP) for the city of
Valparaiso, Chile. The purpose of the SOP is to After reviewing the requirements of the seismic
achieve effective preparedness and to organize a system needed for THRUST, it was decided that
coordinated program of tsunami warning dis- strong motion triggers would be well suited for
semination which will ensure a prompt and our needs. Two seismic triggers, made by Kine-
flexible response by the local population, thus metrics, will initiate the system when a seismic
minimizing loss of life and property. Mr. event occurs. One of the triggers will be in
Emilio Lorca, IHA, has assisted THRUST in formu- Valparaiso and the other will be in Santiago
lating the SOP. Information obtained from the (see Figure 2).
data files and numerical simulation were
utilized to formulate an effective SOP. Pressure digital water level sensors are best
suited for this project because of their Large
dynamic range. Paroscient i f ic Digiquartz
4. REAL TIME
pressure transducers will be used as the water
level gauges. Both water level sensors will be
4 1 . Data Collection
placed on the concrete pier located in the
This task will coordinate the collection and Valparaiso harbor (see Figure 2).
dissemination of all data. The instruments for
GOES
Figure 1.
THRUST instrumentation design.
CDA refers to Command Data and Acquisition.
646
printer that will print a stored message (see
COES Figure 3) when it receives an event alert signal
through GOES from the seismic trigger. The RTP
will:
1) Alert Warning Center personnel that

the seismic trigger was turned on by
an event;
2) provide some level of assessment of

the tsunami threat;
3) remind Warning Center personnel of the

procedures to follow (Hebenstreit
1984b).
1 . Accelograph: Kinemetrics Vertical Seismic Trigger (VS- 1
2. Water Level Gauge: Paroscientific Digitiquartz Pressure Transducer

(2100AS-002) 4.3 Information Dissemination
3. Microprocessor /Terminal: Commodor 64 with Dot Matrii Printer
4. GOES Radio Receiver: Kinemetrics True Time Receiver (468-DC) This final task deals with the transmission of
5. GOES Radio Transmitters: Synergetics (3401 A) warning information to threatened population
Figure 2. areas. In this final step of the warning pro-
THRUST instrumentation.
cess, the incoming data and the accompanying
CDA refers to Command Data and Acquisition.
analysis are used to determine which areas to
warn, in what order to warn them, and what
The GOES radio sets that will be utilized by the
instructions to issue (SOP).
THRUST system will be manufactured by Syner-
getics. These radio sets will be transmitters
The THRUST project has received eight satellite
(TX) only. The radio sets will be connected to
addresses for the GOES System. Four of these
the seismic triggers and water level sensors.
Kinemetrics True Time Receivers will be utilized VALPARAISO EARTHQUAKE ALERT

as the Commodore 64 with a dot matrix printer A STRONG EARTHQUAKE OCCURRED AT z ON IN THE VICINITY
OF VALPARAISO. CONTACT THE FOLLOWING AUTHORITIES:
(see Figure 2). The RTP will be located at IHA
- OFICINA NACIONAL DE EMERGENCE TELEPHONE No 718333
in Valparaiso, Chile. - DEPARTMENT OF GEOPHYSICS TELEPHONE No 6968586
AND AVISE THEM OF THIS EARTHQUAKE ALERT AND DISSEMINATE THE

Procurement of the above mentioned items has ATTACHED TSUNAMI WATCH MESSAGE.
been initiated. After PMEL receives all the TSUNAMI WATCH

equipment (Summer 1985) it will be assembled and ESTABLISHA TSUNAMI WATCH. A TSUNAMI MAY ACCOMPANY THIS
EARTHQUAKE. IF A TSUNAMI OCCURRED, THE WAVE WILL RETARD
tested for approximately 6 months. After a IN REACHING THE COAST AT THE COAST AT THE SHOWN LOCALITIES THE
NEXT SPECIFIED TIMES:
successful test, all equipment will be installed
in Chile (May 1986) for one year of testing and
VALPARAISO minutes
evaluation COQUIMBO minutes

CALDERA minutes
TALCAHUANO minutes
CHANARAL minutes
CORRAL minutes
4 . 2 Data Analysis ANTOFAGASTA minutes
ISLA DE CHILOE (ANCUD) minutes
TOCOPILIA minutes
This task deals with the development of the Real IQUIQUE minutes
ARJCA minutes
Time Processor (RTP). The RTP is the first link
REMEMBER. THIS IS ONLY A TSUNAMI WATCH. NO TSUNAMI HAS
between the THRUST instrumentation and the BEEN OBSERVED BUT ONE MAY OCCUR
Chilean Tsunami Warning Center personnel. The

Figure 3.
RTP will consist of a microprocessor and a RTP message format.
647
addresses will be for the daily test made of the This information can then be accessed via tele-
system. The remaining four will be for the phone line.
standard operation of the system. NOAA, manager
of the GOES system, has assigned THRUST a random Note that this process is entireLy automatic and
report channel with very little usage for the should take no more than 10 minutes to com-
demonstration of the THRUST system to ensure plete. But no real decisions have been made,
successful transmissions of messages. except possibly a predetermined one to sound a
general alert. Final authority to make deci-

Every seismic trigger (uplink) message, whether sions, to issue further alerts or sound the
test mode or standard operation, will be sent evacuation alarm, must rest with human offi-
through the GOES system four times to ensure the cials. This task should be made simpler because
receipt of the message. Each message will be of several factors. The decision makers will
sent within a 15 second window with a 60 second now be familiar with the historical and
delay after each window. Thus, a message could potential hazard analyses performed during the
be received by GOES in as little as 15 seconds pre-event phase and will now be aware of the
or as much as 7 minutes after the occurrence of general type of threat the coast of Chile
the event. Upon receipt of this event triggered faces. Also, because of the Standard Operating
message, the GOES-CDA (Command Data and Plan implemented during the project, IHA will
Acquisition) would be alerted that this is an have a set of procedures to follow and should
emergency message and immediate action is to be not find it necessary to improvise. In addi-
initiated. This action (to be initiated within tion, because of the public awareness program
1 minute) is a message sent back through GOES to established during the project, the Chileans
turn on the water level sensors and the RTP (see should be confident that the threatened popula-
Figure 1) (THRUST Pilot Study, 1984). tion know how to respond to ensure their own
safety. And lastly, the Chileans can be sure
that the sensors and the real-time analysis
5. THRUST SCENARIO
package are providing them with the most up-to-
Once the THRUST study has been installed and date information available.
operating, a typical event scenario should occur
in the following manner (see Figure 1).
6. CONCLUSIONS
An earthquake will activate the seismic trig- The goals of THRUST are to show that such a
ger. This instrument then transmits four mes- system can be built, to work with the Chilean
sages through the GOES system which responds by government to integrate the technical system
initiating an alert code back through the GOES into its disaster control structure, and to
system to the RTP located at the Hydrographic train the Chilean personnel in the operation and
Institute in Valparaiso. The RTP instantly maintenance of the system. Each phase of
responds by initiating a prerecorded message THRUST, then, will be conducted in conjunction

based on the Standard Operating Plan and proce- with personnel from Chile. In this way the
dures established prior to the tsunami. The technology behind THRUST can be demonstrated to
message format THRUST will use is similar to other tsunami-prone (and geophysical hazard-
that used in Figure 3. prone, in general) nations, while concurrently
enhancing the technological capabilities of
In addition. The GOES alert code will initiate Chile. Successful completion of the THRUST
the water level sensors in the Valparaiso harbor project will not only enhance the tsunami
to begin sending data via satellite to CDA. protection of Chile but will, by adding addi-
643
tional input to PTWC, improve the protection of Keywords :
the entire Pacific community.

1 Tsunami
2 GOES
7. R EFERENCES .
3 Hazard Reduction
4 Numerical Model
Bernard, E.N., G.T. Hebenstreit, J.F. Lander and 3 Seismology
P.F. Krumpe, Regional. Tsunami Warnings Using 6 Oceanographic Instruments
Satellites. Proceedings of 1983 Tsunami
Symposium, U.S. Dept, of Commerce, PMEL,
Seattle, WA, 1983.
Bernard, E.N., J.F. Lander, G.T. Hebenstreit,

Feasibility Study on Mitigating Tsunami
Hazards in the Pacific. NOAA Technical
Memorandum ERL PMEL-37, 1982.
Hebenstreit, G.T., THRUST Model Results for

Valparaiso, Chile. Science Application
International Corporation, McLean, Virginia,
1984a.
Hebenstreit, G.T., Basic Criteria for the THRUST

Real-Time Processor. Science Application
International Corporation, McLean, Virginia,
1984b.
Iida, K., D. Cox and G. Pararas-Carayanni s

Preliminary Catalog of Tsunamis Occurring in
the Pacific Ocean. H.I.G. University of
Hawaii, 1967.
Reid, R.O., A.C. Vastano and T.J. Reid,
Development of SURGE II program with

application to the Sabine-Calcas iev area for
hurricane Carla and design hurricanes.
Coastal Studies Inc., College Station, TX,
Report TP77-13, 1977.
THRUST Pilot Study for Pacific Marine

Environmental Laboratory, Task I - Final
Report, CyberLink Corp, Boulder, CO, 1983.
THRUST Pilot Study Prototype System Work Plan,

Task II - First Quarter Report, CyberLink
Corp., Boulder, CO, 1984.
649
STUDIES ON STORM SURGES IN TIDAL ESTUARIES
BY
Tatsuo Konishi
Takeo Kinosita
Hiroshi Takahashi
National Research Center for Disaster Prevention
Science and Technology Agency, Japan
ABSTRACT
Characteristics of storm surges in rivers and 1970 has not yet been determined. Storm surge
estuaries in connection with the discharge of modification in a river reach as well as an
rivers are studied in this report. estuary and a port and harbor has not been
solved hydraulically. On the other hand, it has
The time difference between storm surges and a been noticed that the flood characteristics
flood is dependent upon the drainage area of the change remarkably due to urbanization. Comple-
river if it has not rained before the onset of tion of an urban drainage system, such as a
the typhoon. Three types are categorized by sewerage, makes the time of concentration short
means of drainage areas, as follows. and induces a big runoff coefficient. Consider-
ing these changes in the relation to storm
In a river whose drainage area is over 3000 km 2 surges, the time difference between storm surges
(for example, the Kiso river), the flood is de- and a flood, occurring late, becomes shorter.
layed from the storm surges. The character- The probability of their superposition becomes
istics of storm surges during the non-flood greater. From the view point of disaster pre-
period are clarified in this type. The storm vention, it is necessary to determine the risks
surges occur simultaneously with the flood in a due to superposition. This paper describes the
river whose drainage area is smaller than effects of superposition. In addition, though
1000 km 2 (for example, the Tsurumi river). The the storm surges in embayments were well
analysis of this type explains combined effects studied, the behavior of the storm surges in
of storm surges with floods. A river whose rivers and estuaries, even in the non-flood
drainage area is from 1000 to 3000 km 2 (for period, are not well known.
example, the Arakawa river) shows the inter-
mediate effect of storm surges taking place in The first purpose of this paper is to clarify
the rivers mentioned above where the storm surge the nature of storm surges in rivers during the
meets the flood in the middle reach. non-flood period. The second is to make clear
the combined effects of floods and storm surges
Finally, other effects not obtained from the in rivers. These problems are discussed using
observed data are discussed in light of numer- the observed data, water levels and discharges.
ical experiments. From the analyses, the nature of storm surges in
rivers may be understood as a whole.

Keyvords: Seismology, storm surge
1 . INTRODUCTION
Research on storm surges is not active because a

2. STORM SURGES IN III NON -FLOOD PERIOD
I
'.
serious storm surge has not occurred recently. ( FROM CASES OF THE KISO RIVER)
This does not imply that there are no problems The Kiso river system is composed of the K>to
in storm surges. For example, the cause of the (drainage area, 5275 km 2 ), the Nagara ( 1985 *m- )
big storm surge that occurred in Tosa Bay in and the Ibi (1840 km 2 ) rivers. As the Kiso
690
river has a Large river basin, the flood peak the rising stage of the floods in the Ibi and
will be deLayed from the peak of storm surges, the Nagara rivers, especially in the Ibi, whose
if it has not rained heavily before the typhoon drainage area is smaller than other rivers'.
onset. Therefore the characteristics of storm Therefore the maximum height is partly composed
surges can be discussed in the non-flood period of the flood. But the peaks of storm surges in
from the results obtained in the Kiso river. the Kiso river are separated from the peaks of
The location of the Kiso river is shown at point floods, as seen in Figs. 2 and 3. The maximum
"a" in Fig. 1 and the water level stations and water level at the downstream of the Kiso river
weather stations are shown in Fig. la. Typhoons is mainly caused by the storm surges. The
7220 and 7916 are chosen for analysis. The amplitude of storm surges is defined as the
tidal anomalies at the tide gauge station of difference between the maximum water level of
Nagoya are observed at 2 m and 1.4 m respective- the storm surge and the mean water level at a
ly. Fig. 2 shows the hydrographs of the water certain station as shown in Figs. 4 and 5. The
level stations along the Kiso river during the amplitude at a distance of about 13 km from the
period of typhoon 7220. The peak, which is mouth is 20 or 40 percent higher than the ampli-
observed at 24:00 on 16 September at Yokomakura, tude at the mouth. It decreases to 0.9 percent
is produced by the storm surges, and the peak at of the amplitude at the mouth at 20 km. The
15:00 on the 17th at Naruto is produced by the results show that storm surges are amplified in
flood. The case of typhoon 7916 is shown in rivers and estuaries.
Fig. 3. The peak of the storm surge is observed
at 24:00 on 30 September at Yokomakura and the Why are storm surges amplified in the lower
flood peak is observed at 17:00 on 1 October at reach? Two kinds of phenomena are considered as
Naruto. It is understood that the water levels causes, such as run-up of the storm surge wave
are varying as the sum of the astronomical tide, and wind stress. In order to evaluate the
the storm surges and the flood. The details effectiveness of these causes, numerical simula-
about meteorological conditions and hydrographs tions are applied. The following equations of
of the Ibi and the Nagara rivers are explained motion and continuity in one dimension,
by Konishi and Kinosita (1983).
3U
+ j.
3U _ _ 3_D
_ xb +
ta
3t 3X ® 3X pwR pwR
The most important data for disaster prevention
concerning storm surges are maximum heights in 12 = _ I 12 ( 2 )
3 1 B 3X
rivers and estuaries. Figs. 4 and 5 show these
heights along the Kiso river system. These are used for the basic equations, where U is the
graphs include the high water levels of astro- mean flow velocity over the vertical cross
nomical tide and the 25-hour average of the section of the river, D is the water level
water level before the onset of the storm measured from the reference level, R is the
surges. Mean levels and high water levels are hydraulic radius, Q is the flow flux, xb is the
almost horizontal on the lower reaches and bottom friction, xa is the wind stress, B is the
increase upstream according to the rise of the width of the river, g is the gravitational
river bed. These trends are already mentioned acceleration and pw is the density of the water.
by Unoki (1968). On the other hand, the maximum They are transformed into the difference equa-
heights of the storm surges increase with the tions. A is the area of the cross section and
distance from the river mouths. A difference of the following relation,
1 m of water level was observed between the
river mouth and the station 18 km from the Q = U • A (3)
mouth. Surge peaks on upper reaches occur at
651
is introduced. By the survey of the river Two roughness coefficients, n = 0.02 and
channel, the width of the river, A and R are n = 0.025, are used in the computation. The
determined as functions of the water leveL for drag coefficient of wind is varying within the
each cross section. The simple shape of the range of (0,2, A, 6) * 10“ 3 . 2.6 * 10“ 3
has often
river channel is actually used in this compu- been used in numerical simulations for similar
tation. The part of the river channel higher phenomena, but it does not agree with the obser-
than "H" and the part lower than -2 m T.P. are vation in the cases described below. The wind
supposed to be constant in width. "H" means the observed at Kuwana is applied over the entire
water level above which the channel width is model, but only the component parallel to the
practically constant. The bottom friction, xb river channel is used. More detailed infor-
is expressed by the Manning formula, mation is found in Konishi and Kinosita (1983).
xb _
,
= pw
gn 2 U U |
—L
|
The longitudinal profile of the maximum storm

R^ surge height from the river mouth is shown in
Figs. 6b, 7a and 7b for different drag coeffi-
where n is the roughness coefficient. Normally,
cients. Figs. 6 and 7 are the profiles produced
xa is proportional to the square of wind speed,
by typhoons 7220 and 7916, respectively. The
W, as follows,
zero drag coefficient means that wind stress
plays no role in the river and the result shows
xa = pa Cd W 2
horizontal water level. This is quite different
from the real profile of maximum height of the
where pa is the air density and Cd is the drag
storm surges. The wind stress in the river is
coefficient. The boundary conditions are as-
important in such a river as the Kiso. The drag
signed for the inflow flux at the upper boundary
coefficient adequate to explain the observed
and the tide level at the river mouth. The
profile may be A-6 x 10~ 3 . These values are
arrangement of temporal and spatial grid points
somewhat larger than usually reported. The
is similar to Unoki ' s (Unoki, 1968). Main terms
reason for the large difference is not well
of differential equations are approximated by
understood. The simulated hydrographs are shown
centered differences.
in Figs. 2 and 3 by circles and dots. Cd and n
are equal to A x 10 -3 and 0.025, respectively.

The water level in the Kiso river is simulated.
Good agreement may be seen in the figures.
The storm surges appear separately from a flood
in the river. Therefore the inflow hydrographs
from the upper reach need not be considered.
3 . CASES OF COMBINED STORM SURGES W1 1
The inflow at the upper boundary for the case of FLOOD: 1 ( FROM CASES OF THE TAUK M 1
typhoon 7220 is 2A2 m 3 /s, which is the daily

Small rivers which pour into the vulnerable bay
mean discharge on 15 September at Okoshi, 3A km
of storm surges always have the combined effects
from the river mouth. The inflow for typhoon
of storm surges and floods, because their basins
7916 is 6A6 m 3 /s, which is observed at the
are small and therefore the time of concen-
Umakai weir on 30 September. The water level at
tration is short. Such a case is the flood in
the lower boundary is chosen to be the water
the Tsurumi river, whose drainage area is
level hydrograph at Yokomakura, 2 km upstream
235 km 2 . In this section, the water levels
from the mouth. The river channel form used in
during the storm surge period in a small-scale
the computation is obtained by the survey. The
river, the Tsurumi river, are considered. The
spatial grid is 1 km and the temporal grid is
location of the Tsurumi river i9 shown at point
100 seconds. The upper boundary is located at
50 km from the river mouth.
652
"b" in Fig. 1 and the water level stations are a small variation. The variation comes from
shown in Fig. lb. water of the tributaries.
Fig. 8 shows hydrographs observed along the The water levels of Ashiho-Bashi and Sueyoshi-
Xsurumi river during the period of typhoon 8218 Bashi, in the case of a small variation, are
in September 1982. In Fig. 9 tidal anomalies mostly determined by the assumption of station-
observed at Kawasaki, Yokohama and the river ary state. The boundary conditions are given as
mouth of the Tsurumi river are shown. A flood the discharge from the upper reach at Tsunashima
peak is found at 15:00 at Tsunashima station and and as the tidal level at the mouth. The dif-
at 18:00 at Ashiho in Fig. 8. On the other ference between the observed water level and the
hand, the tidal anomaly at 18:00 is about 50 cm, computed one is about 20 or 30 cm. In the case
which is 80 percent of the maximum tidal of a large variation, the water levels at the
anomaly, as shown in Fig. 9. Thus, these stations can be reproduced by the assumption of
figures indicate the superposition of storm stationary state. But the upper boundary con-
surges and a flood in the Tsurumi River. dition, the water level at Tsunashima, is not
available, because the discharge from the tribu-

At Tsunashima, which is 11 km upstream from the tary, the Yagami river, is important. Two
river mouth, the relation of observed water discharges, Qa and Qs, are introduced. Qa is
level and discharge is shown in Fig. 10. Two given by the surface gradient between the Ashiho
cases are mixed in the figure. One is the case and the river mouth with the assumption of
of superposition of storm surges and a flood, stationary state. Qs is given by the surface
and another is the case of a flood only. A gradient between the Sueyoshi and the river
single relationship is found on the figure. mouth with the same assumption. If Qa equals
Consequently, it may be expected that the tidal Qs, the assumption can be acceptable in the
oscillation has no influence on the water level reach. The comparison is shown in Fig. 11.
at Tsunashima. The relation can be reproduced Fig. 12 compares the calculated result with the
by the numerical simulation. The curve in the observed value at Sueyoshi-Bashi. Data include
figure is drawn by the numerical solution of both cases of superposition and a flood only.
Eqs. (1) and (2) with the assumption of station- Fairly good agreement is found in these figures.
ary state. The upper boundary is Kamenoko-Bashi The water levels at Ashiho-Bashi and Sueyoshi-
which is 16 km upstream from the mouth. The Bashi can be calculated by the assumption of
lower boundary is the river mouth, where the stationary state. In this scale of the river,
water level is supposed to be constantly T.P. the past conditions have no influence on the
0 m. The spatial grid is 400 m. The river bed water level, where the water level can be com-
form is based on the survey in 1982. The curve puted by tide level at the mouth and the dis-
obtained by the solution of Eqs. (1) and (2) charge from the upper reach at the same time.
with n = 0.025 agrees well with the observed
values. It is concluded that the water level at
Tsunashima is simply determined by the discharge 4 . CASES OF COMBINED STORM SURG E S WITH A
FLOOD: 2 (FROM CASES OF THE ARAKAWA R IVER)
from the upper reach.
The combined characteristics of storm surges in
The water levels at Ashiho-Bashi , which i s 4 km both the Kiso and the Tsurumi are studied in
upstream from the mouth, and Sueyoshi-Bashi medium scale rivers. Wind stress amplifies
8 km from the mouth, are analyzed. There are storm surges in the river and the amplified
two kinds of flood discharge variations at these storm surges meet a flood in the middle reach.
stations, a large variation along the river and Such cases are found in the Ibi river and the
653
Arakawa river whose drainage areas are 1840 km 2 station, Shinsuna, is applied over the Arakawa
and 2940 km 2 respectively. Attention ls direc- river model. The spatial grid is 1 km and the
ted to the Arakawa river during the period o£ drag coefficient of 4 x 10 -3 is used, based on
typhoon 7920, because sufficient data were the calculation of the Kiso river. The appro-
acquired. The location of the Arakawa river is priate roughness coefficient used to reconcile
shown as "c" in Fig. 1 and the water level the numerical results with observed data is
stations are shown in Fig. lc. The maximum 0.015 at the lower reach of 22 km and 0.02 at
tidal anomaly is 116 cm at the Tokyo tide gauge the upper reach of 22 km. The results are shown
station. Detailed information on the typhoon, in Fig. 13. The agreements are good. At
tides at the various stations and disasters are Akigase, the weir is not considered in this run
collected in Tsuji (1981) and Konishi and and the difference may be fairly large. The
Kinosita ( 1985 ) surge component is calculated by the method
mentioned above. The ratio of the maximum surge
Fig. 13 shows the hydrographs observed at the component at each station to that at the mouth
stations along the Arakawa river during the is shown in Fig. 14. In the middle of the
typhoon period. At Minamisuna, the peak at figure, the white circles indicate the ratio
15:00 on the 19th is produced by the storm mentioned above, while the black circles are the
surges and the peak of Chisuibashi at 24:00 on ratio of the simulated water level minus E.L. at
the 19th is by the flood. There is a weir at each station to that at the river mouth.
Akigase. The water level at the upstream side Crosses are the storm surge amplitude, defined
("Akigase-kami" ) is different from that at the in section 2, by the difference between the peak
downstream side ("Akigase-shimo" ) before a of the storm surges and the mean water level
flood. But the difference disappears during the before the storm surges and therefore the ampli-
flood period because the gate is opened during tude includes the flood effect. In Fig. 14, the
the period. surge component increases 40 percent at about
22 km upstream from the river mouth. The longi-
In order to know whether the storm surges are tudinal profile of the river bed is also shown
amplified or not, the separation of the storm in the figure. Comparing the ratio with the
surge component from the flood is required. The profile, maximum storm surges may occur near the
following steps lead to it. First, observed place where the profile bends. Similar rela-
water level and discharge variations are simu- tionships between the ratios and the profile are
lated numerically by Eqs. (1), (2) and (3). In seen in other rivers. The surge components are
the process the roughness coefficient, which is calculated in the Edo river during the period of
the most appropriate to the observed data, is typhoon 7920 and in the Kiso river during the
determined. Using the roughness coefficient, periods of typhoons 7220 and 7916. As both
the expected water level (hereinafter called rivers have large basins, it takes a long time
E.L.) is computed under the condition of no for a flood to arrive at the estuary. The water
storm surges, including a flood and an astro- level during the period of storm surges is
nomical tide. The difference between observed mainly influenced by both surge component and
water level and E.L. may be caused by storm astronomical tide. Harmonic analysis is per-
surges only and is defined as surge component. formed for the water levels before the storm
The river bed form is determined by the survey surge to determine the contribution of astro-
of 1979. The discharge hydrography is given at nomical tide and to separate the surge component
the upper boundary, Chisuibashi, and the vari- from the observed data. The period of analysis
ation of tidal level is given at the mouth, is 25 hours, about a lunar day. The differences
Minamisuna. The wind measured by the AMEDAS between the observed storm surge height-, ind •hi-
654
tidal Levels obtained from the harmonic analysis ted around the place where the Longitudinal
give the surge components. They are shown in profile bends.

Fig. 15 for the Kiso river and Fig. 16 for the (3)
Edo river, as the ratio of the maximum surge River with a small drainage area (below
component to the value at the river mouth. In 1000 km 2

)
the Kiso river, the ratio is 1.4 at 12-18 km

upstream from the river mouth, while the ratio A flood occurs simultaneously with storm surges
is 1.2 at 14 km upstream from the river mouth in in the lower reach. The water level in the
the Edo river. The position where the maximum river is determined by the calculation with the
ratio appears is near the place where the longi- assumption of stationary state. The boundary
tudinal profile bends. This trend is also found conditions are the flood discharge at the up-
in the Arakawa river. It is therefore concluded stream and the tidal level at the river mouth at
that
5. the storm surge is 20 or 40 percent larger the same time.
at the place where the longitudinal profile
bends than at the mouth. The flood superposes Several other effects are discussed by Konishi
over the storm surges, and consequently the and Kinosita (1985) using the numerical experi-
water level must be expected high at that point. ments in an ideal river. Two important results
are pointed out. (1) The water level calculated
by the assumption of the stationary discharge is

DISCUSSIONS sometimes underestimated at the back water
From the above analysis, if it has not rained region. (2) Supposing that storm surges and a
before the onset of the typhoon, the character- flood occur separately, T t
is defined as the
istics of the storm surges in rivers may be time when storm surges only attain to the maxi-
classified by drainage areas, as follows. mum at the place where the longitudinal profile
of the river bed bends, while T 2 is the time
(1) River with a large drainage area (over when a flood only attains to the maximum at the
3000 km 2
) place where the river bed height is the mean sea
level. If T t
is T 2 ,
the combined water level
A peak of storm surges is normally separated attains to the maximum.
from a flood. The surge component increases by
20 or 40 percent over that at the river mouth.
The maximum amplification occurs near the place 6. ACKNOWLEGEMENTS
where the longitudinal profile of the river bed The authors would like to express hearty thanks
bends. It is mainly caused by wind stress in to the Kanto Regional Construction Bureau, the
the river. The drag coefficient obtained from Kisogawa Work Office, the Mie Work Office, the
the observed data is 4-6 x 10~ 3 . Keihin Work Office, the Arakawa Work Office, the
Edogawa Work Office of the Ministry of Construc-
(2) River with a medium drainage area (from tion and the Oceanographical Division of the
1000 to 3000 km 2 ) Japan Meteorological Agency for providing us
with valuable data and support.
A flood meets storm surges in the middle reach.
In the moderate flood, the surge component has 7. REFERENCES
the same characteristics as in the case of the Konishi, T. and T. Kinosita (1983): Studies on
large drainage area, (1). The astronomical tide the river invasion of the storm surge
and flood components are superimposed on the (I). Report of the National Research
storm surges and the high water level is expec- Center for Disaster Prevention, 31, 67-87.
C5
Konishi, T. and T. Kinosita (1985): Studies on
the river invasion of the storm surge
(II). Report of the National Research

Center for Disaster Prevention, 34,
13-42.
Tsuji, Y. (1981): On the high tide and wind

waves caused by typhoon 7920. Report of
the National Research Center for Disaster
Prevention, 25, 155-168.
Unoki , S. (1968): Studies on the tide in

rivers. Report of the 15th Coastal
Engineering in Japan, 228-235.
Fig. 1 Storm surges occurred at (a) the Kiso

river, (b) the Tsurumi river, and (c)
the Arakawa river. The tracks of the
typhoons: 1 indicates typhoon 20 in
1979 (7920), 2 = 7916, 3 = 7220, 4 =
8218.
Fig. la Water level stations (A) and weather

stations (•) around the Kiso river
system.
656
Fig. lb Water level stations (A) and tidal
stations (•) around the Tsurumi
river
Fig. 2 Solid lines show observed water levels

along the Kiso river, typhoon 7220.
0's show the computed water levels for
Cd = 4 x 10 -3 and n = 0.025.
Fig. lc Water level stations (O) around the

Arakawa (or Ara) river.
7916.
657
Fig. 4 Storm surge peaks, astronomical tide
peaks and mean water leveLs with the
distance from the river mouths, Fig. 6b Same as Fig. 6a, n is equal to 0.02.
typhoon 7220.
a
£<
Fig. 5 Same as Fig. 4, except for typhoon

7916.
Fig. 7a Same as Fig. 6a, except for typhoon
7916.
Fig. 6a Computed results of the surge peaks,

typhoon 7220. 0's are observed
values, n is equal to 0.025. The
numerals in the figure correspond to
the drag coefficients with the unit
- 3 .
10
658
Fig. 10 Relation between water level and
discharge at Tsunashima. The curve is
Fig. 8 Observed water levels along the obtained by the calculation with
Tsurumi river, typhoon 8218. An arrow Eqs. (1) and (2) under the assumption
indicates the time of the nearest of stationary state.
passing of the typhoon.
SUEYOSmI Cal
tidal anomaly
ASH1HO COJ
Fig. 9 Tidal anomalies observed at the river Fig. 11 Calculated discharge by the wate.-
mouth, Kawasaki and Yokohama, typhoon level gradient, Ashiho-the mouth vs.
8218. on that of Sueyoshi-the mouth.
Fig.
d i scharge
659
Fig. 13 Water Level hydrographs along the Arakawa river. 0's are calculated and *'s are observed
at the time of the discharge observation.
7920 Arakawa River

surge component ratio
Fig. 14 Amplitude ratio(X), the surge component( O is refered to the ob-

calculation) and the river bed and its
servation and • to the
from the river mouth.
running mean with the distance
660
Kiso River
b
3
Fig. 15 Same as Fig.lU for the Fig.l6 Same as Fig.lU for the Edo river,
Kiso river, the ty- the typhoon 7920.
phoon 7220(0, A) and
the typhoon 79l6(#,A)
A SDOF MODEL TO ANALYZE EARTHQUAKE RESPONSE OF A FULL-SCALE
SEVEN-STORY REINFORCED CONCRETE STRUCTURE
BY
Manabu Yoshimura*
Shin Ckamoto**
Vladimir Sigmund"
ABSTRACT
If initial assunptions about the single-degree- common-type hysteretic models. It is also

of -freedom (SDOF) modeling of a multi-story attempted in this paper to predict the response
reinforced concrete (RC) structure are of the structure for different base motions with
reasonable, and the hysteretic model used is different intensities.
capable to reasonably simulate inelastic
behavior of the structure under earthquake 2. DESCRIPTION OF THE STRUCTURE
excitation, it is believed that earthquake
responses of the structure, such as tine history The full-scale seven-story RC structure was
of displacements and maximum displacements, base built and tested as part of the US-Japan
shear and base moment can be reasonably well Cooperative Research Program utilizing Large-
simulated by the use of a SDOF system of the Scale Testing Facilities at the Building
structure. Research Institute, Ministry of Construction,
Tsukuba, Japan. 1,2 Ihe building was designed
In order to verify this, a SDOF model of a and constructed in accordance with normal
full-scale seven story RC structure built and construction specifications and practices
tested at BRI, Tsukuba, Japan, was studied. In (Fig. 1). It had three longitudinal bays and
1.
the stud/, input acceleration was used, which two transverse spans. Cross sections of columns
was the same as the one used during the test, are 500 by 500 mm, and cross sections of beams
and initial skeleton curve characteristics of are 300 by 500 rnn and 300 by 450 nm,
the structure were also derived from the test. respectively, in longitudinal and transverse
Feasibility of common hysteresis models for directions. Ihe structure has a shear wall
force-displacement relationship of reinforced thickness of 200 rnn in the middle frame. It was
concrete was studied by comparing the calculated placed parallel to the loading direction. 'Ihe
response to the measured. building was tested by performing the "SDOF
Pseudo Dynamic" (SPD) earthquake response test
In the second phase of the study, results from procedure. Load was applied by use of
the chosen SDOF model of the structure subjected eight actuators which were controlled to keep
to different base motions with different inverted triangular lateral load distribution
intensities, were used to predict the general during the test. The details of the structure
response behavior of the structure. and test can be found in Reference 1.
Keywords: Earthquakes; Hysteretic model; SDOF 3. INPUT GROUND MOTION
INTRODUCTION The intensity of earthquake motions was varied

in four test runs of the SPD test to yield
The response of a structure to the excitation expected roof level displacements (Fig. 2) . The
caused by an earthquake are fundamental to earthquake record used in the test was modified
earthquake resistant design. If a variety of from the original record, and free vibration
parametric studies can be easily and economi- tests were performed between pseudo dynamic tost
cally performed using a simple SDOF model and runs to study the change in period and damping.
the results obtained adequately reproduce the In this way, by combining the consecutive pseudo
behavior of the real structure, then the prac- dynamic tests, in which the next test stage
tical importance of such a model is verified. started immediately after the previous free
However, there are many practical problems in vibration test is finished, the behavior of the
reducing a MDOF structure to a SDOF model and in structure from its initial intact stage to the
evaluating its hysteretic properties. In most
pratical cases, we can only make reasonable
estimates of the deflection mode shape, proper- * Building Research Institute, Planning
ties of the initial skeleton curve and the Department.
hysteretic behavior of a structure. ** Building Research Institute, Production
Department.
This paper studies the possibility to simulate + Building Research Institute, Visiting
earthquake responses of the SDOF system of a Scholar, on leave from University <>f X.i ih i
,
full-scale seven-story RC structure by using Yugoslavia.
662
last damaged stage, was simulated. The second, 0 = 1.422
third and fourth pseudo dynamic tests, together 2
Me = 0.643 (t * sec /cm)
with free vibration tests, were considered in
4.
this paper.
Q1 = 1.457 Pe
MEASURED STRUCTURAL BEHAVIOR
X7 = 1.422 Xe
Observed behavior of the test structure is
presented in the form of the roof level dis- Initial equivalent stiffness Ke of the SDOF
placement vaveform and hysteretic behavior of system was derived from the first mode period
the roof level displacement and base shear
T of the structure:
5.
during the test sequences of SPD-2, 3 and 4. 6. ‘
Measured tirre history of the displacement Ke = Me * (2ir/T)
response is given in Fig. 3 and hysteretic

behavior, together with interpolated initial By substituting T = 0.448 sec, it becomes:
curve, is shown in Fig 4.
Ke = 126 t/cm
DERIVATION OF EQUIVALENT SDOF MODEL
HYSTERETIC MODELS
The differential equation of motion for an
undamped equivalent SDOF model representing MDOF In this paper, three different common hysteretic
system, as derived in Reference 1, is: models were used to simulate hysteretic behavior
of the structure. They dominantly represent
Me * Xe + Pe = -Me * Xo flexural behavior of the RC. The primary curve
is symmetric about its origin and is represented
where: either by bilinear or trilinear lines. The
models are extensively described in the
Me = EMrUr (equivalent mass) literature and just their main characteristics
are outlined herein.
2
B = Ffe/EMrUr (mode participation factor)
a) Degrading Trilinear (D-tri) Hysteretic
Xe = (1/3 )X (equivalent displacement) Model (Fukada, 1969)
Pe = ZPrUr (equivalent force) It is presented in Fig. 5 and its min

characteristics are: 1) the envelope curve
(Ur) (mode displacement vector) is trilinear, 2) stiffness continuously

degrades with increasing maximum
(Mr) (story mass vector) displacement amplitudes after yielding and
the cracking point controls fatness of a
(Pr) (external force vector) hysteretic loop (hysteretic energy
.
dissipation)
Xo (ground acceleration)
b) Modified Clough Degrading Stiffness Model
X (time-varying function) (Clough, Johnston, 1966)
Numerical integration of the differential It is described in Fig. 6. Its min char-

equation of motion is performed by Newmark s 0
1
acteristics are: 1) the envelope curve is
method with the value of 0 equal to 0.167. bilinear, 2) stiffness degrades with the
Mode displacement vector, story mass vector and increase of a maximum displacement and
external force distribution vector, taken from unloading stiffness degrades in a form of
Reference 1, are as follows: Kr = Ky * (2Dy/ (Dmax - Dmin)) A (meaning
of the symbols used in obvious from the
(Ur) :
= (1.00; 0.85; 0..696; 0.540; 0. 384; 0.234; figure, and when the value of A = 0 , then
0.102) Kr is equal to Ky)
(Mr) :
= (0.156; 0.173; 0.173; 0.173; ^ 0.173 /
0.173; 0.187) (t * sec^/cm) * Using the symbols shown in Fig. 5, unloading

stiffness Kr is expressed as:
(Pr) = (21.75; 18.75; 15.75; 12.75; 9.75; 6.75;
3.75) (ton) Kr = Kc * ALF
3 Me and coefficients which correlate Pe

,
where:
and Xe to the base shear Q1 and roof level
displacement X7 of the structure are: ALF = (Fmax - Fmin) * Cy/Fy/ (Dmax - Dmin)
663
c) Takeda Hysteretic Model (Takeda, Sozen, c) A comparison is presented in Table 3.
Nielsen, 1970) Maximum calculated displacements were much
underestimated during the SPD-3 phase of
Its rrain characteristics are shewn in Fig. 7 the test, while somewhat more reasonable
and are summarized as follows: 1) the results vere obtained for the SPD-4 phase.
envelope curve is trilinear, 2) stiffness
degrades with increasing maximum displace- d) As shown in Table 4, the free vibration
ment, and unloading stiffness degrades in a period after the SPD-3 phase of the test was
form of: calculated much shorter than the measured
value. Such shorter periods calculated
*
A after SPD-3 are closely related to the fact
Kr = (Fc + Fy) / (Dc + Dy) ABS(Dy/Dm)
that the maximum calculated displacement
where Dm defines Dmax or Dmin with respect during SPD-3 are much smaller than the
to the direction. measured value. Stiffness degradation of
the SDOF model was considerably different
7. RESULTS OF ANALYSIS (smaller) even in SPD-4 than the observed
behavior of the structure.
A primary curve for different hysteretic models
was chosen so that stiffness properties and e) Calculated roof level displacement and base
hysteretic energy dissipation capacity are as shear relations are shown in Fig. 12. It
similar as possible to the measured values bears not much resemblance to the measured
(Figs. 8, 9 and 10) The effect of different
.
behavior; the calculated hysteresis shew
hysteretic models on the response is investi- considerably smaller stiffness degradation
gated through comparisons of the measured and properties and larger hysteretic energy
calculated responses, such as: dissipation properties than the measured
one.
a) roof level displacement time histories,
Through the above discussion, it is made clear
b) maximum displacements and base shear, that behavior of the analytical SDOF model using
the D-tri hysteresis rule poorly represents
c) maximum displacements at two phases of the behavior of the RC structure even when the
test (SPD-3 and 4) skeleton curve properties are precisely knewn.
d) free vibration periods, and 7.2 Behavior of the SDOF Model with Modified
Clough Hysteresis Rule
e) roof level displacement and base shear
relationship. Assumed initial curve is shown in Fig. 9, and
properties of the models are presented in
7 . 1 Behavior of SDOF Model with D-tri Table 5. Comparisons on the previously
Hysteresis Rule mentioned terms (a) through (e) are described
below.
Fitting of the initial curve to the measured is
shown in Fig. 8, and properties of the models a) The calculated roof level displacement time
selected are presented in Table 1. Comparisons histories of the selected three models are
on the above terms are described below. presented in Fig. 13. High-amplitude
oscillations were favorably represented in
a) The calculated roof level displacement time all models. When the initial stiffness was
histories are shown in Fig. 11. For all low (model Cl), the SPD-2 phase was poorly
three models shown in the figure, the represented; in other words, the calculated
response waveforms were similar, but they displacements were much larger than the
poorly matched the measured waveforms, in measured ones. This is because the struc-
other words, response amplitudes were much ture remained dominantly in the elastic
underestimated especially in medium- and range in this test run. Mien the initial
low-level amplitude oscillations. The cal- stiffness was equal to stiffness of the
culated response was usually shifted due to intact structure derived from free vibr.r
residual deformations remaining at the end tests (model C2), SPD-2 phase oscillation:;
of SPD-3. were well represented but the model failed
to represent medium-level oscillation:.
b) A comparison is presented in Table 2. Cal- (SPD-3) after pronounced excursion:, in the
culated displacement maxima were very inelastic domain. For the initial stiffnes:
sensitive upon the assumption of cracking value set equal to 0.6 of the intact stage
level and yielding level (D1 and D3) Base . (model C3), the model favorably represented
shear was less sensitive to the assumed measured waveforms throughout the ent re
:
force level. period of the test.
664
b) Maximum displacements and base shears ware measured one. Cbserved and calculated
favorably represented for all five models response waveforms showed no appreciable
with different initial stiffness values, see sensitivity of the shape when the parameters
Table 6. The bigger unloading stiffness of the initial curve were changed within
value (A) considerably increased displace- 20 percent in area. The study included the
ments and base shear in both directions effect of changes in the following
(comparison of models Cl and C4). Note that parameters: yield level, yield displace-
a smaller unloading stiffness (bigger ment, post-yield stiffness, unloading
coefficient) means smaller hysteretic energy stiffness degradation parameter and mass of
dissipation properties, hence, leading to the model, which could be different if a
larger response values. Lower yield level different assumption about the displaced
decreased base shear but did not have much mode shape was used. Within this area
effect on displacement (comparison of models (+— 10 percent of the ideally fitted initial
Cl and C5) curve) , the calculated response waveform is
a favorable representation of the observed
c) As shown in Table 7, all models except C4 one.
failed to represent maximum displacements
during the SPD-3 phase of the test, and b) Comparison of the calculated and measured
especially underestimated negative maxima is presented in Table 10. For all
displacements considerably. models whose initial skeleton curve was
different with +10 percent range of the
d) Calculated amd measured free vibration ideally fitted one, calculated maximum
periods are presented in Table 8. Cbser- values of the displacement and base shear
vation of the stiffness degradation through were within +10 percent of the measured
comparison of those periods at the end of values with few exceptions. Variation of
each test phase shows favorable represen- raxima, ewing to the specific changes in the
tation of the stiffness degradation of all initial curve, is presented in Fig. 16. The
models. ratio of the calculated and measured rraximum
values is plotted against the assumptions of
e) Calculated roof level displacement and base the initial curve. Hiose figures show us
shear relationship are shown in Fig. 14. only a trend of the influence of certain
Cbserved and calculated behavior compare parameters, and general characteristics
favorably. cannot be derived as this trend is changing
with the changes in the input motion
Based on this analysis using the Clough hyster- characteristics. Note that this ratio tends
esis rule, it was found that: 1) the overall to increase as the unloading stiffness
response of the structure can be favorably coefficient increases. This trend can be
represented except for lew-level anrplitude explained by the fact that assumption of a
oscillations, and 2) response is not very larger coefficient results in smaller
sensitive to the estimation of the yield level hysteretic energy dissipation properties.
or unloading stiffness.
c) Comparison of the calculated and measured
7.3 Behavior of SDOF Model with Takeda values of displacements is presented in
Hysteresis Rule Table 11. Calculated negative maximum
displacement values for the SPD-3 phase of
Fitting the initial skeleton curve to the the test showed greater sensitivity of input
measured one is shown in Fig. 10. As this parameters. This was especially pronounced
hysteresis rule was deemed as representing the for changes in yield displacement (T2 and
observed behavior most favorably among the three T8) and unloading stiffness degradation
hysteresis rules, more extensive studies were parameters (T2, T9 and T10) . Variations of
made for this rule than the others. the ratio of calculated and measured maximum
Characteristics of selected models are presented displacements for two phases of test (SPD-3
in Table 9 in which main parameters are yield and 4) are plotted in Fig. 17 against the
level, post-yield stiffness and unloading change in the initial skeleton curve
stiffness, etc. Comparisons on the previously parameters. This also verifies the
mentioned terms (a) through (e) are described previously stated conclusion under b) . This
belcw. behavior can be explained on the basis of
the sensitivity of the response upon the
a) Calculated waveforms of different models frequency content of the input motion.
are presented in Fig. 15. They represented Stiffness behavior of the model sometimes
the measured one favorably during all failed to represent the change of the
stages, except for lew- level oscillations, structural stiffness, in other words,
say, 11 to 13 sec. in some models (T1 and stiffness of the analytical model degrades
T9). The best results were obtained for only when the response point exceeds the
model T2 when the stiffness characteristics previously obtained maximum displacement,
of the model fitted in the best way with the while the structure degrades continuously in
865
stiffness. If in this case the response Obtained rraximum values of displacements and
displacement spectra for a certain ground base shear are shown in Table 13. Calculated
motion has drastic changes within the area values of the response of Miyagi Cki and Tokachi
of interest, slight changes in the effective Oki Earthquakes were larger than that of the
stiffness of a model can have perceptible El Gentro record, although the acceleration was
influence on calculated waveforms. The of the same intensity. This result is believed
sensitivity of the response to the unloading to be correlated to different energy contents of
stiffness coefficient, stated under b), is those earthquakes. Behavior of the structure
also observed in this table. would be satisfactory if it was exposed to those
base motions with the intensities of 300 and
d) Comparison of the calculated and measured 400 gals, considering maximum obtained dis-
free vibration periods is presented in placement and base shear. Furthermore, it is
Table 12. From that, it can be concluded believed that this structure can withstand with
that the models represented change of sufficient safety even the base motions with
stiffness properties favorably. It is also 500 gal intensity.
obvious, hcwever, that those models which
failed to represent maximum values of the 9. CONCLUSIONS
SPD-3 test phase (T4, T9 and T13) , showed
rather poorer representation of the change By using a reasonable hysteretic model and
of stiffness. assuming an appropriate initial skeleton curve
of the structure, we favorably simulated the
e) Calculated roof level displacement and base behavior of a full-scale seven-story structure
shear relation are presented in Fig. 18. As using a SDOF response analysis. Gommon
can be seen by comparing it to the measured flexural-type hysteretic models of RC were used,
one in Fig. 4, the overall hysteretic shape namely: D-tri, Clough and Takeda models. Hie
8. Takeda model showed the most favorable results,
calculated for the SDOF model has favorable
similarities with the measured one. and calculated response with this model was less
sensitive to changes in the input parameters.
Through the above discussion, it can be
concluded that the SDOF model of the full-scale In the second part of the study, the response of
structure using the Takeda hysteresis rule with the tested structure was predicted using the
initial stiffness properties and hysteretic SDOF model under different base motions with
energy dissipation capacity as similar as different acceleration anplitudes. From the
possible to the measured one with allowable study, the tested structure was found to
tolerance of +10 percent, gave favorable withstand safely based motions with a 400 gal
representation of the earthquake response of the intensity and even with a 500 gal intensity.
structure.
10. REFERENCES
EARTHQUAKE RESPONSE OF THE FULL-SCALE
STRUCTURE AGAINST DIFFERENT BASE 1) S. Ckamoto, et al. ; A Progress Report on The
MOTIONS Full-Scale Seismic Experiment of a Seven-
Story Reinforced Concrete Building;" BRI
In the preceding part, it was proven that by Research Paper No. 94.
reasonably estimating the initial curve, and
using a reasonable hysteretic model, v«e could 2) T. Kaminosono, et al.; SDOF Pseudo Dynamic
favorably simulate the earthquake response of a Test;" US-Japan Cooperative Research; VIII
full-scale seven-story structure. In this part, WCEE; San Francisco, CA; 1984.
model T2 was used which fitted initial stiffness
properties and the hysteretic behavior of the 3) S. Okamoto, et al.; Techniques
structure in the best manner to stud/ the Scale Testing at BRI Large Structure Test
response of the full-scale test structure under Laboratory;" BRI Research Paper No. 101.
different base motions with different
intensities. 4) Y. Kitagawa, et al.; Dynamic Characteristics
of a Full-Scale Seven-Story Reinforced
Input base motions were modified in terms of the Concrete Building;" BRI Research Paper No.
peak acceleration from the following earthquake 108.
records
5) M. Yoshimura; Ultimate Moment Resisting
El Centro 1940 N-S with intensities of 300, 400 Capacity; US-Japan Cooperative Research;
and 500 gals. VIII WCEE; San Francisco, CA; 1984.
Tokachi Cki 1968 E-W with intensities of 300,

400 and 500 gals.
Miyagi Cki 1978 N-S with intesnities of 300, 400

and 500 gals.
666
Table 1. Model Characteristics (D-tri Model)
Model Kc Me Dc Dy Fy Y
t/cm ton cm cm ton
D1 126 630.8 0.85 3.83 201 0.045

D2 126 630.8 0.59 3.83 201 0.045
D3 126 630.8 0.59 3.83 188 0.045
Note y : stiffness reduction coef. after
yielding (relative to Kc)
Table 2. Comparison of Response Maxima (D-tri Model)
Model Calc. Meas . Error Calc Meas Error

shear shear % disp disp %
ton ton cm cm
• D1 371.0 432.5 -14.3 18.9 34.2 -44.7
-411.3 -439.0 -6.3 -25.8 -32.1 -10.0
D2 440.2 1.8 30.8 -10.0
-433.3 -1.3 -29.6 -7.8
D3 389.6 -9.9 25.4 -25.7
-402.7 -8.3 -27.6 -14.0
Table 3. Comparison of Displacement Maxima at Two

Phases of the Test (D-tri Model)
Model Calc, Meas Error Calc, Meas Error

disp disp % disp disp %
cm cm cm cm
SPD-3 SPD-4
D1 12.0 23.8 -49.6 18.9 34.2 -44.7
-4.1 -22.3 -81.6 -25.8 -32.1 -19.5
D2 18.7 -21.4 30.8 -10.0
-7.3 -67.3 -29.6 -7.8
D3 17.6 -26.1 25.4 -25. 7
-5.3 -76.2 -27.6 -14.0
667
. 8. . . . . . . . .
Table 4. Comparison of Free Vibration Periods

(D-tri Model)
Model Calc, Meas Error Calc Meas Error

sec sec % sec sec %
SPD-3 SPD-4
D1 0.53 1.16 -54.3 1.13 1.36 -16.9
D2 0.66 -43.2 1.23 -9.6
D3 0.61 -47.5 1.15 -15.4
Table 5. Model Characteristics

(Modified Clough Model)
Model Ky Me Fy Y A
t/cm ton ton
Cl 52.5 630.8 201 0.11 0.40
C2 126.0 630.8 201 0.05 0.60
C3 74.5 630.8 201 0.05 0.60
C4 52.5 630.8 201 0.11 0.60
C5 52.5 630.8 188 0.11 0.40
Note y : stiffness reduction coef. after yielding
(relative to Ky)
A : coef. defining unloading stiffness
Table 6. Comparison of Response Maxima

Model Calc Meas Error Calc, Meas. Error

shear shear % disp disp. X
ton ton cm cm
Cl 437.8 432.5 1.2 29.9 34.2 -12.6
-434.0 -439.0 -1.1 -29.3 -32.1 -8.7
C2 459.8 6.3 28.1 -17.8
-467.5 6.5 -29.3 -8.7
C3 406.1 -6.2 33.5 -2.1
-388.6 -12.5 -28.9 -10.0
C4 505.2 16.8 41.3 20.8
-449 . 2.5 -32.0 -0.3
C3 419.8 -2.9 29.8 -12.9
-414.5 -5.6 -27.3 -15.1
668
.. .. . .. .. .
Table 7. Comparison of Displacement Maxima at Two Phases

of the Test (Modified Clough Model)
Calc, Meas Error Calc, Meas Error

Model
cm cm cm cm
SPI >3 SPD-4
23.8 -10.5 29.9 34.2 -12.6

Cl 21.3
-16.7 -22.3 -25.1 -29.3 -32.1 -8.7
-37.4 28.1 -17.8

C2 14.9
-53.4 -29.3 -8.7
-10.4
-4.6 33.5 -2.1
C3 22.7
-14.7 -34.1 -28.9 -10.0
24.0 0.8 41.3 20.8

C4
-24.0 7.6 -32.0 -0.3
-5.9 29.8 -12.9

C5 22.4
-17.8 -20.2 -27.3 -15.0

Calc, Meas Error Calc, Meas Error

Model
sec % sec sec 1
sec
SPD-3 SPD-4
1.16 -13.8 1.20 1.36 -11.8

Cl 1.00
-13.8 1.13 -16.9
C2 1.00
-7.8 1.33 -2.0
C3 1.07
-10.3 1.33 -2.0
C4 1.04
-6.0 1.23 -9.6
C5 1.09
(Takeda Model)
Table 9. Model Characteristics
Dc Dy Fy Y A
Model Kc Me
ton cm cm ton
t/cm
3.83 201 0.045 0.60
T1 126 630.8 0.85
3.83 188 0.045 0.60
T2 126 630.8 0.85
3.83 188 0.040 0.60
T3 126 630.8 0.85
3.83 188 0.040 0.50
T4 126 630.8 0.85
3.83 188 0.050 0.55
T5 126 630.8 0.85
3.83 188 0.050 0.60
T6 126 630.8 0.85
3.83 176 0.045 0.60
T7 126 630.8 0.85
4.36 188 0.045 0.60
T8 126 630.8 0.85
3.83 188 0.045 0.50
T9 126 630.8 0.85
3.83 188 0.045 0.70
T10 126 630.8 0.85
3.83 188 0.045 0.60
Til 126 655.0 0.85
3.83 201 0.045 0.60
T12 126 655.0 0.85
3.83 188 0.045 0.60
T13 126 610.2 0.85
yielding
Note y : stiffness reduction coef. after
(relative to Kc)
A : coef. defining unloading stiffness
669
. . . . . . .

(Takeda Model)
Model Calc. Meas Error Calc, Meas Error

sec. sec % sec sec
SPD-3 SPD-4
T1 1.07 1.16 -7.8 1.27 1.36 -6.6
T2 1.09 -6.0 1.29 -5.1
T3 1.07 -7.8 1.29 -5.1
T4 1.00 -13.8 1.29 -5.1
T5 1.04 -10.3 1.20 -11.8
T6 1.08 -6.9 1.28 -5.9
T7 1.08 -6.9 1.31 -3.7
T8 1.11 -4.3 1.32 -2.9
T9 1.04 -10.3 1.21 -10.1
T10 1.11 -4.3 1.37 0.7
Til 1.09 -6.0 1.31 -3.7
T12 1.04 -10.3 1.31 -3.7
T13 1.04 -10.3 1.28 -5.9
Table 13. Maximum Response Values against Different

Base Motions
Earthquake rec. 300 gal o o gal 500 gal

disp shear disp shear disp shear
cm ton cm ton cm ton
El Centro
1940 NS 10.9 305.4 16.5 337.9 16.4 349.0
-9.7 -298.9 -17.8 -345.5 -25.7 -391.5
Tokachi Oki
1968 EW 29.5 413.8 43.2 493.2 48.3 522.7
-29.6 -414.5 -37.9 -462.7 -45.5 -506.5
Miyagi Oki
1978 NS 20.2 359.5 35.3 447.6 47.0 515.4
-25.0 -387.6 -35.4 -447.7 -48.0 -521.6
671
. . . . 6 . .. 1
. . 55
Table 10. Comparison of Response Maxima (Takeda Model)
Model Calc Meas Errpr Calc Meas Error

shear shear % disp disp %
ton ton cm cm
T1 461.6 432.5 6.7 34.5 34.2 0.9

-447.8 -439.0 2.0 -32.1 -32.1 0.0
T2 445.5 3.0 35.0 2.3
-416.3 -5.2 -30.0 -6 .
T3 423.0 -2.2 34.3 0.3

-394.6 -10.1 -28.8 -10.3
T4 399.4 -7.7 29.7 -13.2
-389.4 -11.3 -27.8 -13.4
T5 455.7 5.4 33.6 -1.8
-433.1 -1.3 -30.1 -6.2
T6 472.8 9.3 36.3 6.0
-437.2 0.5 -30.7 -4.3
T7 427.5 -1.2 34.8 1.8
-390.6 -11.1 -28.5 -11.1
T8 443.8 2.6 35.4 3.5
-408.1 -7.0 -29.3 -8.7
T9 415.4 -4.0 29.8 -12.9
-410.3 -6.5 -28.9 -10.0
T10 483.4 11.8 41.5 21.3
-423.4 -3.5 -31.2 -2.8
Til 445.8 3.1 35.0 2.3
-407.3 -7.2 -28.4 -11.5
T12 467.7 8.1 35.5 3.8
-438.2 -0.2 -30.5 -5.0
T13 445.8 3.1 35.2 2.9
'1
-407.3 ~7 2 - -31.2 j
-z.o
i
Table 11. Comparison of Displacement Maxima at Two

Phases of the Test (Takeda Model)
Model Calc Meas Error Calc Meas. Error

cm cm cm cm
SPD -3 SPD-4
T1 20.4 23.8 -14.0 34.5 34.2 0.9
-17.6 -22.3 -21.1 -32.1 -32. 0.0
T2 22.0 -7.6 35.0 2.3
-19.9 -10.8 -28.8 -6.
T3 21.7 -8.9 34.3 0.3
-19.5 -12 . -28.8 -10.3
T4 21.2 -10.9 29.7 -13.2
-16.0 -28.3 -27.8 -13.4
T5 21.3 -10.5 33.6 -1.8
-17.5 -21.5 -30.1 -6.2
T6 21.7 -8.8 36.3 6.0
-19.6 -12.1 -30.7 -4.3
T7 23.7 0.4 34.8 1.8
-21.3 -4.5 -28.5 -11.1
T8 22.9 -3.8 35.4 3.5
-16.3 -27.0 -29.3 -8.7
T9 21.3 -10.5 29.8 -12.9
-16.0 -28.3 -28.9 -10.0
T10 23.1 -3.0 41.5 21.3
-24.0 7.6 -31.2 -2.8
Til 22.8 -4.2 35.0 2.3
-21.1 -5.4 -28.4 -11.5
T12 21.3 -10.5 35.0 3.8
-18.1 -18.8 -30.5 -5.0
T13 20.1 -15.5 35.2 2.9
-16.4 -26.5 -31.2 -2.8
670
© © _l! P' II II 1
© 9 1 1 1! II 1 FUL-I-.
17.000 4500
QQ0 5 QQQ 6QOQ
6-
i 1 II II 1
3000 3000 1250012500 [3000 [3QQQ

1
i i i i i oj §j
1 II 1
©- ^S_G]_ _ _X_ _ _G2
Bl_
Gi
GS
X Il_ 1 II II
1LJ1.T]
:s si;
1
X
1 1 II
,
'(
!>
Q|
9 si
i Gl!
1
<S, 1
_ JiQ?
~r
-fi-
p
1 II 1
si
'
®!'
l
-J i i 1 1 3
!|
i
''
;sool ?soo
iP 3' '' !
-Jt -- i°-l 4= = J'9 , 6000 I sooo

:s O., 0.1 s! s 600
© © © ©
a) Plan b) Elevetion
Fig.l Outline of the Test Building
Fig. 2 Input Acceleration Data
Fig . 3 Measured Time History of Roof Level Displacement
672
Displacement Level Fixst Yield Of the Reinforcement
or 500
Column
® R £ 1/^80
© 1/1)80< R £ 1/320
© 1/320 < R S 1/170
© 1/170 < R i 1/130
® 1/130 < R 5 1/100
© 1/100 < R £ 1/ 6 8
-400
Roof-Level Displ.(mm)
TEST-2
TEST-3
TEST-4
-500
Fig. 4 Measured Roof Level Displacement and Base Shear Relationship
Fig. 5 Hysteresis Rule of Degrading Fig. 6 Hysteresis Rule of

Trilinear Model Modified Clough Model
a) unloading after
cracking
b) unloading after
yielding
c) small amplitude
reversals
d) small amplitude
reversals
Fig. 7 Hysteresis Rule

Takeda Model
673
674
Fig. 12 Calculated Roof Level Displacement - Base Shear Relationship
(D-tri Model)
Fig. 13 Calculated Time History of Roof Level Displacement
675
Fig. 14 Calculated Roof Level Displacement - Base Shear Relationship
calc calc
meas
©
1.1
1.0
0.9
Me
610,2 630,8 655,0
calc calc
meas
©
1,1
z
1.0 .
0.9
r
0.04 o o 45 0,0 5
Fig 16 Sensitivity of Maxima to Different Parameter

.
(Takeda Model)
676
400
200
-200
400-
200 -
- 200 -
400 ;
200 -
- 200 -
400-
200 -
- 200 -
400-
200
-?nrv
Fig. 15 Calculated Time History of Roof Level Displacement
(Takeda Model)
677
calc
meas
©
SPD-4 SPD-3
calc
calc
meas
meas
©
1
i n
I.U
09
1 .
-
V
— p*"*
1
0.9
1
0
-
i
/./ /
/
Fy r
25 9 274 29 3 0.04 Q045 0.05
Fig. 17 Sensitivity of Maximum Displacements to Different Parameters
during Two Phases of the Test (Takeda Model)
Fig. 18 Calculated Roof Level Displacement Base Shear Relation
(Takeda Model)
675
: . . .
SETTLEMENT OF APPROACHING EMBANKMENTS TO BRIDGES

CAUSED BY THE NIHONKAI CHUBU EARTHQUAKE
by
Yasuyuki Koga (1) and Eiichi Taniguchi (2)
ABSTRACT approaching embankments at both sides, it is

counted as two sites If an approaching part to
The Nihonkai Chubu Earthquake, May 26, 1983 bridge was in a cut slope, it is was not inves-
caused settlements of approaching embankments to tigated .
bridges. This paper evaluates the relationship
between the damage rate of approaching embank- A questionnaire form was sent to Aomori, Noshiro
ments and various factors based on the results and Akita Local Construction Office of Ministry
of an investigation on actual damage to of Construction and Aomori and Akita Prefecture.
approaching embankments. Factors influencing These questionnaires were filled out by staff
the damage to approaching embankments are members of such organizations and were collected
epicentral distance, soil condition of embank- and analyzed.
ment, soil condition of ground and micro-relief.
It is concluded that an approaching concrete A questionnaire form also included figures of
slab is very effective to maintain a smooth road the investigation location, plan, longitudinal
4.
surface after earthauakes. and transverse sections and photographs of the
Keywords: Earthquakes, embankments, settlements damaged embankments
1 . INTRODUCTION
Items of investigation included the character-
The Nihonkai Chubu Earthquake, May 26, 1985, istics of the embankment (height, soil name of
caused severe damage to roads, especially to embankment, etc.), subsoil conditions (micro-re-
approaching embankments to bridges as shown in lief, depth of water table, etc.), and the char-
Photo 1. Investigations were performed on the acteristics of the abutments and failure
actual damage to all the approaching embankments patterns
to bridges of national roads, main local roads
and prefectural roads in Aomori and Akita RESULTS AND DISCUSSIONS
Prefecture by means of questionnaires. The
results of investigations are described. Table 1 shows the number of investigation
locations. A location is judged damaged if one
2. PURPOSE of the following three conditions are satisfied:
1) The maximum settlement of the crest of
The purpose of the investigation was to find embankment is over 5 cm, or cracks with the
factors which affect earthquake-induced damage width of over 5 cm developed at the crest of
to approaching embankments and to obtain basic embankment
data for developing countermeasures to prevent 2) The slope of the embankment collapsed, or
settlement of approaching embankments by analyz- cracks with the width of over 5 cm developed
ing those factors. at the slope.
3) The retaining wall tilted or collapsed or
3. METHOD cracks with the width of over 1 cm developed
at the retaining wall.
The investigation was performed on all the
approaching embankments to bridges, with a Table 1 shows that the number of all the inves-
height of over m and a length of over 15 m, on
1 tigation locations is 1673 and 97 locations were
national roads, main local roads and prefectural damaged, therefore the damage rate is 5.8%. The
roads in Aomori and Akita Prefecture. All the damage rate of approaching embankments (which
approaching embankments under above mentioned will be called merely "the damage rate" in this
conditions were investigated regardless of the paper) is defined as follows:
occurrence of damage. If a bridge had two
Number of Damaged Locations

Damage Rate of Approaching Embankments = (1)
Total Number of Approach Embankments
Fig. 1 shows the relationship between the epicentral distance is over 170 km, and that the
epicentral distance and the damage rate. The damage to approaching embankments mainly
epicentral distance of the investigation occurred in the region with the epicentral
locations is between 80km and 240 km and the distance less than 170 km. In Fig. 1 the damage
maximum epicentral distance of damaged locations rate in the region with an epicentral distance
is 219 km. Fig. indicates
1 that the damage less than 100 km is relatively low. This is
rate is less than 5% at locations where the because the region is located in Oga Penninsula
(1) Head, Soil Dynamics Div. P.W.R.I., Ministry of Construction, Tsukuba, Japan
(2) Senior Research Engineer,
Soil Dynamics Div. P.W.R.I., Ministry of Construction, Tsukuba, Japan
€79
:
which consists of mountainous and hill sites damaged approaching embankments had earthquake-
with good subsoil conditions. induced settlements which is less than 10% of
the height of embankment, and that the settle-
Fig. 2 shows the relationship between the soil ment at only one location exceeds 20% of the
condition of embankments and the damage rate. height of embankment
This figure indicates that the damage rate is
low in the case for embankments made up of grav- Table 4 shows the installation rate and damage
elly material. The gravelly material is defined rate of the approaching concrete slab. The
as material whose content of gravel (grain size approaching concrete slab is a reinforced
=2-75 mm) is over 50% in weight. The Earth concrete slab which is installed to prevent the
Work - Construction Manual (Japan Road Associ- gap between abutment and embankment as shown in
ation, 1977) describes good material for back- Fig. 6. Table 4 shows that there were two
fill of retaining structures or buried locations where approaching concrete slab were
structures as shown in Table 2. Although this installed and earthquake-induced settlement was
regulation was determined with consic oration of observed. Actually at these two locations nonu-
settlement of embankments during earthquakes, it niform settlement occurred for a long distance
is effective in preventing seismically induced along the road, but a car could negotiate the
settlements of embankments by requiring soil roadway after earthquake. It can be concluded
with a lot of gravel for backfill material. from Table 4 that the approaching concrete slob
is very effective to keep the road open to traf-
Fig. 3 shows the relationship between the thick- fic after an earthquake, because it is able to
ness of soft soil layer and the damage rate for prevent an earthquake-induced gap between abut-
gravelly and sandy ground and clayey ground. ment and embankment. However attention should
The thickness of soft soil layer is defined as be paid to the fact that caving was observed
follows under the approaching concrete slab as shown in
Photo 2.
Thickness of Soft Soil Layer (m) =
Therefore it is necessary to inspect if there is
Sum of Thickness of Layers with N value < 10 at caving under the approaching concrete slab after
the depth of less than 10 m. a big earthquake and to fill the hollow with
soil or mortar if it is found.
Fig. 3 indicates that the damage rate is higher
at locations with soft layers over 4 m thick in 5. CONCLUSIONS
gravelly and sandy ground and at locations over
8 m thick in clayey ground. It may be said that (1) The damage rate of approaching embankm.
the damage rate increases for the softer ground. to bridges was 5.8% in the Nihonkai Chubu
Earthquake .
The relationship between the micro-relief and (2) The damage to approaching embankments to
the damage rate is shown in Fig. 4. The micro- bridges mainly occurred in the region with
relief classification was carried out by using an epicentral distance of less than 170 km.
"micro-relief classification map for flood (3) The damage rate of approaching embankm
control". When this map was not available at to bridges is low in the case that embank-
certain locations, the "land classification map" ment consists of gravelly soil and high in
by the Economic Planning Agency was used. the case when the embankment is located in
Figure 4 shows that reclaimed land, swamp or reclaimed land, swamp or former swamp and
former swamp has higher damage rate and that the sand dune in micro-relief.
damage rate is as high as 9.4% in flood plains (4) The settlement of 90% of all the damage
where most damaged locations were. The damage approaching embankments to bridges is less
rate in sand dune is high because of liquefac- than 10% of the height of embankment.
tion of sandy ground during earthquake. (5) The approaching concrete slab is very
effective to keep the smoothness of road
Table 3 shows the relationship between abutment surface after earthquakes.
foundation type and the damage rate. In this
figure the damage rate is high for spread foun- 6 . ACKNOWLEDGEMENTS
dation and it is low for pie and caisson founda-
tion. This is because a rock layer is located The authors express their appreciation to
near the ground surface which has a large staff members of the Disaster Prevention
resistance against earthquake loadings at Section of Road Bureau, Ministry of
locations where the spread foundation is used, Construction, Planning Division and F"ud
and the ground has small resistance at locations Division of Tohoku Local Construction
where pile and caisson foundation is used. Bureau, Aomori, Noshiro and Akit < L" il
Construction Office, Aomori and Akita
Fig. 5 shows the relationship between the maxi- Prefecture for their cooperation on th.-
mum settlement of embankment and the height of investigation. The authors also than* Mr.
embankment. The maximum settlement of approach- Y. Sasaki, Head of Ground Vibration Division
ing embankment obtained in this investigation of PWRI for his advice.
was 80 cm. Fig 5 indicates that about 90% of
680
Fig.l Epicentral distance and damage
rate of approaching embankments
to bridges
oVO
o
P
U
0)
tn
d
£
<d
'd
soil condition
of embankment
Fig. 2 Soil condition of embankment

and damage rate of approaching
embankments to bridges
681
rate
damage
cT~ 2 4 6 8 10 0 2 4 6 8 10
thickness of soft thickness of soft
soil layer (m) soil layer (m)
Fig. 3 Thickness of soft soil layer and damage

rate of approaching embankments to bridges
3 C E c
Hc
0)
flj 3 « E <o
 -P Jo T3 U U)
c f0 T3 O
3 C U l-l T3 u
O (0 (D o O a oi
E in E O E E
P TD q u
O 0) 3 O
o-i XI in o-i
Fig. 4 micro relief classification

and damage rate of approaching
embankments to bridges
682
settlement
earthquake-induced
Fig. 5 Earthquake-induced settlement of approaching

embankments to bridges and the height of
embankments
approaching
concrete slab
embankment
Fig. 6 Schematic view of

approaching concrete
slab
683
)
Table 1 Number of investigation points
A B A + B B/(A + B)
organization not damaged daraged damage ^te
Aomori Local Con-
K-i o 120 0 120 0.0
struction Office
Akita Local Con-
sf struction Office
100 3 103 2.9
•H CO
Noshiro Local Con-
56 4 60 6.7
Is struction Office
total of Ministry
276 7 283 2.5
of Construction
Akita Prefecture 991 35 1026 3.4
Aomori Prefecture 309 55 364 15.1
total 1576 97 1673 5.8
Table 2 Characteristics of good

material for back fill
(Earth Work-Construction
Manual
maximum grain size 100mm

percent finer 25 - 100%
(4.76mm sieve)
percent finer 0 - 25%
(0.074mm sieve)
plasticity index less than 10
Table 3 Kind of foundation of abutment and

damage rate of approaching embank-
ment to bridges
kind of number of damage

total damaged
foundation points rate (%)
direct 907 19 2.1
pile 522 51 9.8
caisson 10 1 10.0
unknown 234 26 11.1
684
I 1
<D
O
(0
ip
P
3
CO
T3
ITJ
O
i t3
£ o
p
P a OHdm eg
ij
,
ip
>p a ti c
0
O ZBA 1
-P
c | K| 3 l.
CD 8-5 9!
e
0)
I—
-p
-p
CD um
CO
-p
c
CD TJ
> CTi cn d
CD
3 g 8 8
P •5 U X P j,
o
a 0
'O
8 8p §
, w cj
o pO §
-p Ch d) a)
eu P P E CD
H p d p co
a) p a) u t:
‘
.
nm
xi id
fO u a P p 4-1
rp g a) c vi p p
3 x o c a) 3
CO c 3 o -p w w
CD c
P o
CD ll dP
P CO
O
c CQ <
o 3
u in
.5
IT>
c
•p
x:
u
c0
o
p
a,
a,
co
ip
o
VO
-p (N VO
o vo
CD
cp
ip
W
CD
rH 3
XI s
c0 D>
Eh M
O
685
Photo. 1 Example of settlement of approaching
embankment to bridge (settlement
= 50 cm, height of embankment = 10 m)
Photo. 2 Cave under the approaching concrete

slab (this photograph was taken
after the earthquake)
686
.
ROCK AVALANCHES AND SLOPE FAILURES INDUCED BY THE NAGANOKEN-SEI BU EARTHQUAKE, 1984
BY
Kohei Tanaka*, Norio Oyagi*, Takashi Inokuchi*, Shigetsugu Uehara*, and Hiroshi Takahashi*
ABSTRACT 2. THE NAGANOKEN-SEIBU EARTHQUAKE, 1984
A strong earthquake occurred at the southern The earthquake occurred at 8:48 in the morning
foothills area of Mt . Ontake on September 14, on September 14, 1984, at a latitude of 35.807
1984. A ridge, near the top of Mt. Ontake, N, longitude of 137.554 E, 1.1 km depth (data
collapsed and an immense amount of debris fell from the National Research Center for Disaster
13 km. This was one of the largest rock ava- Prevention), and magnitude of 6.8. The epi-
lanches in history. Several large slope fail- center was situated in anactive area which has
ures also occurred at Matsukoshi, Ontakekogen, suffered many small earthquakes since 1976.
and Takikoshi. In the majority of the failures, Judging from the analysis of the earthquake
rupture surfaces were situated in layers of vibration, this earthquake was caused by the
pumice from the Ontake volcano. These pumice lateral fault whose main stress direction Is
layers contained halloysite which has less cohe- northwest to southeast, but the fault could not
sion and a smaller friction angle than pumice be located on the ground surface. The after-
layers containing allophane. shock area had a width of 10 km from ENE to WSW
and a depth of about 5km. The seismic intensity

Key words: Avalanches, earthquake, landslides,
of the main shock was 4° at the Japan Meteoro-
slope failure.
logical Agency at Iida, Suwa, and Kofu. How-
1 . INTRODUCTION ever, at the mountainfoot area of Mt. Ontake it
The Naganoken-Seibu earthquake in Japan on was though to be 6 ° or more, because the
September 14, 1984, triggered several hundred maximum acceleration was estimated as about
landslides. One landslide was particularly 370-390 gal, based on the condition of grave
large, with debris forming one of the largest stones. Most slope failures were caused by this
rock avalanches in history. Other relatively main shock (Fig. 1, 2, and 3. Uehara, et *1.
large slope failures occurred in Matsukoshi, 1985).
Ontakekogen, and Takikoshi. A total of 29

3. THE GEOLOGY OF MT . ONTAKE
people were killed. This paper is a report of a
The earthquake occurred at the foothills area
survey conducted by the National Research Center
Mt. Ontake which is 3063 m high. The mounts n Is
for Disaster Prevention.
supported by Paleozoic formations and rhyolite
*National Research Center for Disaster in Cretaceous formations. The activity of Mt

Prevention.
688
Ontake is divided into four stages (Kobayashi, 4.1.2 Topography
1975). In the first period, andesitic material The source area is situated on a ridge which
formed a large stratovolcano . Then it was stretches in a southeasternly direction from the
dormant for a long period, and the mountain body central cone of the volcano. Its elevation Is
became heavily eroded. About eighty thousand from 2550 m to 1850 m. The topography of this
years ago the second period began. Activity of failure follows the valley, so that it looks as
rhyolitic magma characterized this new stage. At if a steep valley which borders the main scrap
first, an enormous volume of pumice (Pm-I layer) begins from the ridge. Its form is a long strip
erupted and simultaneously the caldera was and at the lower part it curves to the west.
formed at the center of the mountain body. An

4.1.3 Geology
eruption of rhyolitic lava and pyroclastic rocks
This failure occurred in the area which is
buried the caldera and formed a new mountain
covered by the material from the third active
body during the third period. This was a period
period of Mt. Ontake. The order of rock layers
of active andesitic magma. Almost all the
is: volcanic breccia, tuff including pumice,
mountain body which we see now was then formed
scoria and then lava from the lower part. The
in the fourth period. As mentioned later, the
source area of the failure is broadly covered by
pumice layers from the second period are related
the tuff layer including the pumice as mentioned
to large slope failures.
above. This tuff layer is very soft and wet,
4. SLOPE FAILURES AND ROCK AVALANCHES and striations are carved on the surface of it
4 . 1 Ontake Kuzure - Failure at the Head of the (Photo 1). This layer was formed by deposits
Denjo River and - Rock Avalanche
falling on the eroded valley and then scoria and
4.1.1 The Situation of the Disaster
lava layers covering it. Erosion then produced
The severest damage caused by this earthquake is
a ridge there. The cause of the failure is
that which occurred near the top of Mt. Ontake
thought to be that the earthquake caused a
and brought a rock avalanche from the Denjo
fracture in the tuff layer, which worked as the
River to the Otaki River through the Nigorigawa
rupture surface so that the rock body above it
River. This rock avalanche killed at least
fell down as one large block. Judging from a
fifteen persons. This failure was the biggest
longitudinal section of the source area, if the
one induced by this earthquake. The scale of
initial velocity had been great, almost all the
this failure was approximately 1300 m long, 450
debris would have jumped over the front valley
m wide, and 1 80 m in maximum depth; an estimated
and then flowed toward the Suzugasawa River.
3.6x10^ m^ of material was removed (The Japan
Yet, actually almost all the debris flowed down
Geographical Survey Institute, 1984).
689
the Denjo River. Therefore, the initial velocity. Because the top level of denudation
velocity could not have been great (Fig. 5). is undulating, it is thought that the rock
avalanche flowed down irregularly. As there is

4.1.4 Flow and Sedimentation of Debris
little sedimentation of debris in the flow area,
4. 1.4.1. Rock Avalanche. The debris flow which
most of the debris flowed down to the
was caused by this great failure flowed down
sedimentation area with the aggregation (Photo
along the Denjo River - Nigorigawa River -
2).
Ontake River and finally arrived at Korigase 13
km down the river. This debris flow was thought 4. 1.4. 3. Banded Sedimentation Structure. One of
to have been a "mud flow", "pulverulent body the characteristic features of the flow area is
flow" or "rock avalanche" by many researchers found on flat places where the rock avalanche
right after the disaster. There is a strong overtopped the ravine wall. That is the banded
possibility judging by the observed data, that sedimental structure in which different colored
the main body of debris contained very little rocks are disposed in a striped structure. It
water. And it was observed from the inside is observed on the opposite flat ridge of the
structure of the debris that big blocks with source area putting into the Denjo River, on the
cracks had flowed so that it cannot be thought flat ridge 2 km downstream from the source area
that all the debris was in a powdery condition. and on a flat part of the valley wall 3 km
Thus the debris flow is described as a rock downstream. This structure has patterns like
avalanche. In this paper, the area of the flow geological banded folding. The cause is thought
means from the lower part of source area to the to be a part of the rock aggregates kept
Denjo River and the Nigorigawa River. And the their geological structure and overtopped the
area of sedimentation means along the Otaki bend of the valley without becoming mixed. The
River. But actually debris sedimentation was difference of the color in the banded structure
observed along the Nigorigawa River, too (Fig. must have come from difference of original rock
4). faces (Photo 3)-
4. 1.4. 2 Denudation of the Valley-Wall. The 4. 1.4. 4. Double Layer Structure Sediment of Rc-k
denudation of the valley-wall was observed in Avalanche. The sedimentation of rock avalanche
the whole of the flow area. Especially, it appears as a double layer structure. The upper
reached a point 100 m from the bottom of the thin layer contains much water and appears as a
valley in the Denjo River. This shows that a muddy layer, but the lower layer consists of
huge debris aggregation, baring the valley as debris aggregation with little water. The upper
high as that place, flowed down at high muddy layer has fragments of wood and little
690
2
roots, but these are not observed in the lower plant was damaged, and several houses were
layer. The upper layer is generally thin destroyed (Fig. 6).
(under 1 m), but the thickness of the lower

4.2.2 Topography and Geology
layer is thought to be '30-40 m, by comparison
The Matsukoshi area is situated on the river
with the original topography. The material in
terrace of the Otaki River near the junction of
the lower layer must have come from the source
the Omata River. The basement of this area is
area of the failure, but the origin of the
paleozoic formation and debris deposits with
upper layer is not clear. However, judging from
volcanic ash from Mt. Ontake forms the river
the many sticks in the upper layer and its
terrace. There is terrace gravel between the
tendency to become thicker and clearer near the
basement and debris deposits too. Almost all
valley-wall, the origin of the upper layer may
areas near the source area covered with a layer
be related to the denudation of the valley wall.
of volcanic ash, included pumice. This layer of
The time gap in the sedimentation of these two
volcanic ash does not appear as a horizontal
layers may not be great, because the form of the
layer but is distributed along undulations of
upper layer is thought to be closely related to
the source area, the volcanic ash layer was
the movement of the lower layer (Photo 4).
deposited to cover an original valley and then
4. 1.4. 5. Hummock . There are hummocks which form debris sediments were deposited thickly
small prominences on the sediment in the flow horizontally. It was a buried valley before the
and sedimental areas. The hummocks mainly con- failure occurred. But in this earthquake the
sist of rock aggregate and area covered with volcanic ash layer was sheared and then
thin mud. They appear at the places where the sediments above it flowed down along it. This
velocity of the rock avalanche was decreased by valley structure must also have gathered ground
the topography or other factors. The velocity water (Fig. 7, Photos 5 and 6).
of the downward flow was approximately 80 km per

hour according to eyewitness reports (Geological
The debris, which failed on the right river terrace
Survey of Japan, 1984).
flowed across the conjunction of the Omata River
4 . Landslide Failure of Matsukoshi and the Matsukusa River, ran against a wall of
4.2.1 The Situation of the Disaster the left river terrace, then changed direction
A large failure, approximately 150 m wide and to the right and climbed up the right terrace
200 m long, happened on the right bank of a con- wall about 200 m downstream from the source
jection of the Omata River and Matsukusa River. area. Finally, the debris flowed into the Otaki
This failure killed 13 persons. A cement-mixing River along the Omata River.
691
4.3 Landslide Failure at Ontake Kogen 4.4 Landslide
4.3.1 The Situation of the Disaster 4.4.1 The Situation of the Disaster
Five rather large scale failures occurred at This failure occurred on the north-western wall
Ontake Kogen; identified as A, B, C, D, and E in of one ridge extending north-east from Otaki
Fig. 8. No one was killed but the road was River Dam. The failure is 150 m wide and 50 m
badly damaged by these failures. long. The debris attacked the east edge of
Takikoshi village, destroying four houses and

killing one person (Photo 8).
Judging by the topography at Ontake-Kogen the
failure did not occur easily, because it is 4.4.2 Topography and Geology
gentle slope (about 15°) covered by new tephra. The main scrap is straight, which suggests that
But many traces of slope failure were found in there is a fault along the main scarp. The rup-
this area, which suggests that this area had ture surface is situated between a layer of
suffered slope failures of same type in this impermeable volcanic ash and a layer of volcanic
time. The geology of Ontake Kogen is a pumice sand composed of lacustrine sediments. Antece-
and ash sequence. The geologic column made in dent precipitation was thought to have had an
failure A is shown in Fig. 9. effect in this case too. Ground water had col-
lected on the impermeable volcanic ash layer at

These five failures appear characteristic of
the time when earthquake happened. In this sit-
translational slides, of which rupture surfaces
uation liquefaction occurred in the sand layer
are flat, and the depth of failures is shallow
above the layer of volcanic ash. Therefore, the
and constant. It means that the failures were
rupture surface was in an almost horizontal
caused by a rupture in a specific layer which
plane (Fig. 10).
was formed parallel to the slope. This specific
layer was a white pumice layer which is Pm-I 4.4.3 Flow and Sedimentation of Debris
layer of Mt. Ontake volcano (Photo 7). The main body of debris was deposited fanwlse on
the reservior, which contained no water. Debris

cones are concentrated at the front of the fan
The debris from Failure A flowed about 1 km down
shape and the thickness of debris decreases at
two streams, respectively. The debris from
the center of sedimentation. On the whole, the
Failure B did not move much, remaining in the
shape is very similar to that of the Oik«' land-
source area; a white pumice layer was observed
slide induced by the Izu-Oshima Kinkal <-.irth-
in it. Failures of C, D, and F were small and
quake, 1978. This sedimental situations
there was little debris flow.
suggests .that the majority of debris moved in a
692
p
mass and has deposited at the front leaving a 5 . 1 X-ray Diffraction Analysis
very thin layer on the flow area. A sample was picked up at each rupture surface,
but samples at Takikoshi were also taken from

4 .5 Other Smaller Failures
the volcanic ash below the rupture surface.
4.5.1 The Situation of the Disaster
Each sample was about 150 g, material under 2
Many small failures were recognized along the
was separated by hydraulic elutration and ana-
swamp near the tollgate of the road, along the
lyzed. The diffraction pattern of hallyosite
upper stream of Nigorizawa River or along the
was recognized in each sample. In particular,
trail to Nigorigawa hot spring. These failures
strong peaks of halloysite were detected in
were small scale but numerous. Nobody was
samples from Matsukoshi and Ontake Kogen (Fig.
killed (Fig. 3).
11 ).

5 . 2 Observation Under an Electron Microscope
These failures occurred on the steep slopes cut
Samples analyzed by the x-ray diffraction method
by the river erosion or the road. They were
were observed under an electron microscope with
concentrated in the thick lava layer with joints
an acceleration voltage of 200 kv and magnifica-
or at the fracture zone in the Palaeozoic
tion of 50,000. That data coincides with the
formation.
data of x-ray diffraction analysis. Ball type
4.5.3 Flow and Sedimentation of Debris halloysites were common in samples of Matsukoski
The debris was deposited directly under the and Ontake Kogen and there were few in the
source area and moved a very short distance. sample from Ontake Kuzure. But it was not
Most of the sedimentation is talus. detected in the sample from Takikoshi (Photo 9
and 10).
5. MINERAL ANALYSES AND SHEARING TESTS ON THE
PUMICE LAYER
5 . 3 Shearing Tests
At the large scale failures of Ontake Kuzure,
The shearing test was formed against Pm-I pumice
Matsukoshi and Ontake Kogen, the rupture sur-
layer's sample from Ontake Kogen and another
faces were situated in layers of pumice or vol-
pumice layer (M-Pm) from Ontake Kogen which was
canic ash from the Ontake volcano. Therefore,
not connected with the failure. The Pm-I layer
these pumice layers were examined by x-ray dif-
contained halloysite and the M-Pm layer
fraction analysis, observation under an electron
contained allophane. The cohesion and the
microscope, and the shearing tests.
friction angle of the Pm-I layer was
tf 2 7°
C = 1 .9 /m , <\> =
693
and the M-Pm layer was 6 . 3 Failures of Steep Slopes
tf /m 2 29°
C = 8.7 , 4> = Contrasted with big failures, small failures or
rock falls are caused by weathered surface soil

6. DISCUSSION
or weathered rocks with developed joints. These
6 . 1 Classification of Slope Failures
failures are surface phenomena; topographical
All the failures which occurred in the
factors are the main cause.
Naganoken-seibu earthquake, 1984 are classified
into two groups according to the topographical 6 . 4 Precipitation
and the geological features of the source area. The nature of failures in the Tokachi-oki
Failures in the first group were on a large earthquake, 1968, is thought to be the same as
scale but not numerous, whereas failures in the that of huge failures in the Naganoken-seibu
second group are the reverse. The first ones earthquake, 1984. That case had precipitation
were related to the geological structure, of about 200 mm and then the earthquake occur-
especially pumice layers, and the second ones red. Many failures occurred on gentle slopes
were affected by high slope angles. and 49 persons were killed. There was preci-
pitation of 1 65 mm at the observation point of

6 . 2 Pumice Layers
Mt. Ontake, and it is thought that ground water
Pumice layers tend to be very sensitive to
was concentrated by the geological structure at
vibrations from earthquakes. We have several
the head of the Denjo River, Matsukoshi and
examples of this problem: the white clay layer
Takikoshi. Judging from these observations, it
in the Imaichi area in the Imaichi earthquake,
is necessary to check the existence of pumice
1949 (Morimoto, et al. 1957); Shirasu, which is
layer, effect of precipitation, and concentra-
a kind of pumice layer, in the Ebino-Yoshimatsu
tion of ground water, in order to predict large
earthquake, 1968 (Ando, 1971); and Hachinohe
slope failures occurring as a result of
pumice layer in the Tokachi-oki earthquake, 1968
earthquakes.
(Hotta, et al. 1968). As in the case under dis-
7. REFERENCES
cussion, halloysite was detected in them except
Hachinohe pumice layer which has not yet been 1) Ando, T. (1971), Landslides in "Shirasu"
examined. It is possible that the presence of area caused by Ebino-Yoshimatsu Earthquake,
halloysite affects the physical property of Reports Cooperative Research Disaster
pumice layer. Prevention, No. 26, pp 147-156, National
Reseach Center for Disaster Prevention (In
Japanese) .
694
.
2) Geographcial Survey Institute (1984),
Disaster Caused by the Naganoken-Seibu
Earthquake, 1984, Bull. Geographical
Survey institute, No. 60, pp 48-52 (In
Japanese)
3) Geological Survey of Japan (1985), Two
debris-avalanches associated with the
Naganoken-seibu Earthquake, in 1984, Report
of Coordinating Committee for Prediction of
Volcanic Eruption, No. 32, pp 48-51 (In
Japanese)
4) Hotta, H., Miura, 0., and Tamura, T. (1968),
Slope Failures Caused by the Tokachi-oki
Earthquake in the Southeastern part of
Aomori Prefecture, Annals of the Tohoku
Geographical Association, Vol. 20, No. 4, pp
195-201 (In Japanese).
5) Kobayashi, T. Ohmori, E., and Ohmori, T.
(1975), Petrochemistry of Ontake Volcano,
Bull. Geological Survey Japan, Vol. 26, No.
10, pp 7-22 (In Japanese).
6) Morimoto, R., Ossaka, J., and Fukuda, T.
(1957), Geology of Imaichi District with
Special Reference to the Earthquake of Dec.
26, 1949 (III), Bull. Earthquake Res.
Institute, Vol. 35, Part 2, pp 359-375.
7) Uehara, S., Tanaka, K., Inokuchi, T.,
Kinoshita, S., Imoto, S., Kishi, S., and
Iwasaki, M. (1985), Report on the Survey of
the Naganoken-seibu Earthquake, 1984, Report
of the Investgation of Principal Natural
Disaster, Vol. 25 (In Japanese).
695
• compression
o dilatation (Lower hemisphere projection)
Fig. 2 Farthquakc mechanism of main shock

(Ueharaetal. 1985)
Photo. 1 Striation on rupture surface

(Ontake Ku/urc)
Photo. 2 Denudation of the valle>
(Ontake Ku/urel
696
* . -S'*
•«% .
r
.*
yt
V
Takikoshi ***
• +
"
»/. ^kMatsukoshi
2km
I ••
«>
Fig. 3 Distribution map of slope failures.
Old slope failures
Slope failures induced by the Naganoken-seibu Earthqueake,

1984
637
A
CROWN
Fig. 5 Longitudinal section (Ontake Kuzurc)
Photo. 3 Banded Structure (Photo. Asahikoyo) Photo. 4 Double layer structure of sediment of rock
avalanche.
Debris
*
.
‘
(included thin volcanic ash laycrl
S in
Volcanic ash
Debris (Terrace Deposit!

Volcanic ash (included scolla)
Volcanic ash (included putnicei
Terrace gravel ($ J- 30cm l

Volcanic ash (included pumnx
Paleozoic formation
Fig. 6 Slope failure at Matsukoshi. Fig. 7 Geologic column of Matsukoshi
698
Photo. 5 Source area (Matsukoshi) Photo. 6 Rupture Surface (Matsukoshi)
(photo. Shinano-mainichi).
Yellow volcanic ash.
Yellow pumice
Bluish gray volcanic ash
Purple yellow volcanic ash
White pumice
Crashed pumice
Pumice thin layer
White pumice included

pebbles
Purple volcanic ash

Yellow volcanic ash
White pumice (Pm-1)
Yellow volcanic ash
Fig. 9 Geologic column of Ontake Kogen

Fig. 8 Slope failure at Ontake Kogen
Photo. 7 Slope failures at Ontake Kogen. Photo. 8 Slope failure at Takikoshi
Failure A (right) and Failure B (left) (Photo. Nakanihon-koku)
699
Volcanic ash
Sand
„ t>
20 in —
Volcanic breccia
(or autobrecciated lava)
15 in -
Sand
R. S.-»
Volcanic ash
10 ml
Blue volcanic ash
included pebbles
5 m—
Brown conglomerate
Fig. 1 1 X-ray diffraction patterns of clay minerals

on rupture surfaces
0 m-
Fig. 10 Geologic columan of Takikoshi
Photo. 9 Electron microscope image of clay Photo. 10 Electron microscope image <>l lay
minerals (Ontake Kogen) minerals (Ontakc Kuzurc)
700
TASK COMMITTEE REPORTS
Report of Task Committee on
(A) STRONG-MOTION INSTRUMENT ARRAYS AND DATA
Date : May 23, 1985
Place : Public Works Research Institute, Ministry of Construction

Tsukuba Science City, Ibaraki Japan
,
Attendees : Japan Side- H. Tsuchida (Chairman) (PHRI)

S. Hattori (BRI)
M. Ichikawa (MRI)
S. Noda (PHRI)
K. Ohtani (NRCDP)
Y. Sasaki (PWRI)
K Tamura
. (PWRI)
T. Kuwabara (PWRI)
U.S. Side- R.B. Borcherdt (Temporary Chairman) (USGS)

A. G. Franklin (WES)
J,F. Lander (NOAA)
R.B. Mac Donald (USBR)
A.M. Abdel-Ghaf far (Princeton Univ.)
I . Activities and Principal Accomplishments to Date
Catalogs of strong-motion earthquake records observed both in the

United States and Japan are being exchanged. U.S. data are published
in the "Seismic Engineering Program Report," and the Japanese data are
published in the "Strong-Motion Earthquake Records in Japan."
Dr. Krinitzky of WES visited PWRI NRCDP and PHRI, exchanged the
,
reports, and discussed strong motion records and their analysis.
II . Future Programs
Both sides of the Task Committee plan to continue their respective

efforts in the following nine points 1) to 9) and to initiate two
points 10) and 11).
1) After an earthquake which has caused damage to structures

or an earthquake during which maximum acceleration
exceeding about 0.1G has been recorded, the task committee
of the country where the earthquake has taken place will
provide a list of the strong-motion earthquake records for
^the counterpart of the task committee. The list contains
maximum component acceleration of each record. If there is
such a list compiled by any organization, the list
mentioned above may be replaced by it.
2) Every year the task committee will exchange catalogs of the

strong-motion earthquake records during the previous year.
The catalog contains maximum component accelerations and
wave forms of major records. If there is such a catalog
compi'led by any organization, the catalog mentioned above
may be replaced by it.
3) The Task Committee makes appropriate arrangements to

provide digitizable copies of records when they are
requested. In addition, arrangements will be made to
provide information on the characteristics of the site and
structures at the locations where such records are obtained.
7®
. .
4) When the organizations taking part in the panel publish

reports on the strong-motion earthquake records, the
organizations will distribute copies of the reports to the
organizations of the panel interested in them. The Task
Committee exchanges lists of the organizatipns which wish
to receive the reports.
5) The Task Committee will continue to exchange digitized data

on all the major strong-motion earthquake records recovered
in both countries, under the condition mutually acceptable
to the agencies involved in aquisition of the data.
6) The Task Committee will exchange technical information

concerning seismometers to be developed, and deploying and
installing instruments with emphasis on the instal-lation
of bore-hole instruments and the design of seismic arrays.
7) The Task Committee plans to assist and cooperate, where

possible, in the following areas:
a. Assistance and cooperation with governmental organiza

tions in other seismic areas, in order to promote high
quality strong earthquake motion observations in all
seismically active areas of the world.
b. Assistance and cooperation with any international

effort to record strong ground motion close to the
source of a large magnitude shock.
8) The Task Committee feels the UJNR Panel on Wind and Seismic
Effects should play a major cooperative role in the imple-
mentation or relevant parts of the Resolution of the Inter-
national Workshop on Strong-Motion Earthquake Instrument
Arrays held May 1978, in Honolulu, Hawaii. The exchange of
complete information on all aspects of the program, as it
develops particularly in Japan and the United States, will
be carried out in the manner of our standard exchange when
appropriate. Particularly with unique data from arrays, or
deep bore-hole instruments in the U.S. and Japan, Task
Committee Chairmen will retain the responsibility of
relaying information on their existence and subsequently of
transferring the digital data from such institutions as in
the U.S., the USGS, CDMG USC; and in Japan, PWRI, BRI,
,
PHRI MRI and NRCDP

, ,
9) The Task Committee will make a preliminary study on

comparison of digitizing and computer processing techniques
of member organizations on strong-motion earthquake
records. The Task Committee chairmen will exchange their
ideas on a proposed plan for this study to possibly include
digitizing specific records, and computer processing of all
such digitizings; followed by a workshop to discuss the
results
10) The Task Committee recommends collaborative studies

involving strong-motion instrumentation and interpretation
of resulting strong-motion data sets for purposes of earth-
quake hazard reduction.
11) The Task Committee will study the possibility of exchange

of scientists, engineers, and appropriate equipment in a
timely fashion for the observations following large earth-
quakes in either country. As the largest and most
interesting aftershocks occur immediately following the
main earthquake, recovery of the most useful aftershock
data requires that the response to major earthquakes be
initiated as rapidly as logistics permit.
704
. )
(B) LARGE-SCALE TESTING PROGRAM
Date : May 22, 1985
Place : Public Works Research Institute, Ministry of Construction,

Tsukuba Science City, Ibaraki, Japan
Attendees : Japan Side- K. Ohtani (Chairman) (NRCDP)

S. Noda ( PHRI
S. Okamoto (BRI)
K. Yokoyama (PWRI)
U.S. Side - H.S. Lew (Temporary Chairman) (NBS)

J.H. Gates (Cal. Tran.)
R.N. White (Cornell Univ.)
The main activities related to this task committee during the past year
are as follows:
1) All of the experimental and analytical studies related to

the reinforced concrete structure have been completed. The
results have been published in Special Publication 84 by
the American Concrete Institute.
2) Testing of the full-scale steel structure and other suport

tests have been completed in Japan. Support tests in the
U.S. is in progress. Analytical studies are currently in
progress
3) Active exchanges of U.S. and Japanese researchers have

occurred under the reinforced concrete, steel and masonry
programs and have been very useful.
4) The Joint Planning Committee on the masonry program met in

the U.S. in July, 1984. Researches on the coordinated
masonry structures were discussed.
5) Ministory of Construction and National Science Foundation

signed on December, 1984, the Implementing Arrangement for
Cooperation in the U.S. -Japan Earthquake Research Program
Involving Large-Scale Experiments of Masonry Structures.
II. Future Programs
The status of the U.S. -Japan Joint Program of reinforced concrete and
steel structures and the results of the masonry planning committee were
reviewed. The Task Committee adopted the following resolutions as a
result of discussion.
Each country should collect copies of

reports for all
coordinated research conducted in that country in
a single
set. Each country should exchange a set of these
reports.
705
2) The major aspects and conclusions of the reinforced
concrete and steel test programs in each country should be
synthesized into two separate reports. Since it is
critical to disseminate the program results to engineering
profession in both countries, the Joint Technical
Coordinating Committee is strongly encouraged to expedite
this effort.
3) To assure effective exchange of information on masonry

structures, immediate and expanded personnel exchange is
encouraged under the Large-Scale Testing Program.
4) The Task Committee will review progress of the Joint

Research Program and continue to reassess the need for
testing of other structures and report its findings and
recommendations to the Panel.
5) The Task Committee recommends that the Panel endores the

Implementing Arrangement on the coordinated masonry program
between the Ministry of Construction and the National
Science Foundation.
6) Task Committees on (B), Large-Scale Testing Program and

Task Committee on (D), Evaluation of Performance of
Structures should maintain closer liaison to facilitate the
planning of instrumentation and analysis for full-scale
test structures.
7) Technical exchange of information on masonry structures

should be continued through this Task Committee. The
following information will be exchanged;
a. bibliography of research reports, and

b. copies of existing reports.
8) Exchange of information on large-scale testing facilities

and large-scale testing programs should be encouraged.
706
. . )
Report of Task Committee
(C) TASK COMMITTEE ON REPAIR AND RETROFIT OF EXISTING STRUCTURES
Date : May 22, 1985
Place : Public Works Research Institue, Ministry of Construction,

Tsukuba Science City, Ibaraki Japan,
Attendees : Japan Side- M. Hixrosawa (Chairman) (BRI)

T. Iwasaki (PWRI)
H. Tada (PWRI)
U.S. Side - M. P. Gaus (Temporary Chairman) (NSF

S. K. Takahashi (NCEL)
R. N. Wright (NBS)
I . Activities and Accomplishments to Date
In Japan the following activities have been carried out:
1) The Japan National Project - Post - Earthquake Measures of

Buildings and Structures Damaged by Earthquake is
now carried out as a five-year plan from 1981 to 1985.
Research products in 1984 are
a. Inspection Method of Damaged Structure

b. Investigation Results of Repairing Methods
2) Following handbook was compiled and published by the Japan

Concrete Institute in October, 1984.
3) "Hand book of Seismic Retrofitting of Existing Reinforced
Concrete Structures"
3) "Total System for Aseismic Design, Inspection and Repair

4) and Retrofit of New or Existing Building, Equipment and
Non-Structural Element for Official Use" was compiled by
Government Buildings Department, Ministry of Construction
in March, 1985 and was under trial use.
In the U.S. the following activities have been carried out:
1) Initial studies on behavior of unreinforced masonry

structures were concluded. A conference on this subject
was held in Charleston, South Carolina.
2) Studies on base isolation devices have been in progress

which may be of benefit in repair and retrofit of existing
structures
Research has been initiated on techniques to assess current

material and structural properties in structures which are
aging or showing the effects of weathering or heavy use.
Such information is essential in developing retrofit
measures
Research on methods to assess the structural capabilities

of existing and damaged structures is continuing along
with research on methods to repair and retrofit such
structures
707
. . .
II. Future Programs
The task committee recommends the following items for future activities:
1) Continue and expand the exchange of reports, books and

special publications related to repair and retrofit of exis-
ting structures.
2) A cooperative workshop on repair and retrofit should be

held in the near future. Possible dates might be in August
1986 in conjunction with the National earthquake meeting
to be held in Charleston, South Carolina or in conjunction
with the May 1987 UJNR meeting to be held in Japan. The
Charleston, SC meeting was considered to be particularly
appropriate because of the very strong interest in assess-
ment and retrofit in the eastern U.S. at the present time.
3) Exchange information and give consideration to possible

cooperative efforts to study methods to quickly and econo-
mically assess the characteristics and residual structural
capabilities of material which may have suffered degradation
due to weathering, aging, misuse or change of use and on
methods to upgrade the capabilities of such materials where
required. Develop methods of analysis for predicting the
structural capacities of structures containing such
materials
4) Share information such as handbooks and on specific

completed projects such as evaluations of particular designs
for which retrofit studies have been carried out and which
may provide specific validation of retrofit design methods.
5) Arrange for the evaluation of the reports and manuals on

seismic retrofit of concrete structures and of museum build-
ing and contents and of historical structures which were
prepared under the UNIDO Balkans Project.
6) Carry out research to evaluate the potential structural

benefits and relative economics of using base isolation
devices for retrofit of existing structures. Such devices
may be of particular interest for historical or monumental
structures
•
7) Give increased attention to developing methods to assess

nonstructural items in existing structures which could
pose severe hazards and/or cause loss of service. Develop
methods for retrofit or repair of important non-structural
elements
8) Coordinate with Committee D "Evaluation of Performance of

Structures" to identify and possibly encourage a joint
U.S. -Japan paper on computer programs which are usefull for
evaluation of existing structures and which could be used
to guide repair and retrofit decisions.
708
, , , ) , .
(D) EVALUATION OF STRUCTURAL PERFORMANCE
Date : May 23, 1985

Tsukuba Sience City, Ibaraki, Japan
Attendees : Japan Side- S. Okamoto ( Chairman) (BRI
U.S. Side - G. R. Fuller (Chairman) (HUD)
1) Second Workshop July 1984, Berkeley Marina Marriott Hotel,

:
California. Analysis and evaluation case studies of six

benchmark structures were presented by U.S. and Japanese
representatives. Three types of analysis were conducted
by each side 1) Rapid Evaluation Method, 2) Detailed
:
Structural Analysis, and 3) Research Computer Analysis

Method. Participants included 12 Japanese and 18 Americans.
2) Third Workshop May 17, 1985, Housing and Urban Development

:
Corp., Tokyo, Japan.

Papers were presented on the following subjects:
a. Progress Report on Seismic Performance of Existing Build-

ings (R.N. White)
b. Rapid Evaluation Methods (S.A. Freeman),
c. Comparison of U.S. Spectrum Capacity Method and Japan
Standard Evaluation Method (S. Okamoto),
d. Summary and Recommendations for Improvement of Building
Seismic Design and Construction (M. Ozaki),
e. Seismic Capacity Evaluation of Existing SRC Buildings
(H. Aoyama)
f. Status Report on U.S. Activities of Interagency Committee
on Seismic Safety in Construction (ICSSC), Applied
Technology Council (ATC-14), and Federal Emergency
Management Agency (FEMA) program for Abating Seismically
Hazardous Non-Federal Buildings (ATC, BSSC, EERI)(G.R.
Fuller)
g. Guideline for Post-Earthquake Evaluation of Earthquake
Damaged R/C Buildings (H. Hiraishi),
h. Emergency Standard for Classification of Risk Levels
of Earthquake Damaged R/C Buildings (S. Otani),
i. Standard for Evaluation of Seismic Performance of R/C
Block Masonry School Building (T. Okada)
j. Instrumentation of Building Nc.5, Ministry of Public
Welfare (Y. Yamazaki),
k. Insrumentation of Buildings (M. Agbabian)
709
3) Workshop Proceedings A two-volume publication containing
:
the proceedings of the First and Second Workshops was

completed by Cornell University under a research contract
with National Science Foundation. Copies were furnished
to Task Committee (D) members at the Third Workshop on
May 17, 1985.
4) UJNR Panel Meeting A "Summary Report on the U.S. -Japan

:
Workshops on Evaluation of Structural Performance" was

presented by Dr. R.N. White in Session 4 of the 17th UJNR
Panel Meeting on Wind and Seismic Effects.
5) Coordination of U.S. and Japan Activities Discussions

:
were held by Task Committee Members on the coordination

of several activities in each country concerning evaluation
of existing buildings for earthquake resistance. A closer
liaison with Task Committee (B), Large Scale Testing
Program will be maintained to coordinate the testing,
instrumentation and analysis of the full-scale masonry
structure.
II. Future Programs
1) The Task Committee reconfirmed the objectives and recommend-

ations of May 17, 1984 concerning analysis, instrumentation,
and post-earthquake evaluation of benchmark structures.
2) Encourage the continued exchange of literature, analyses,

and research reports relative to assessment of seismic
resistance of buildings.
3) Expand the inventory of existing buildings in each country

that have been evaluated and instrumented, and then
observed after any future seismic activity. Continue to
evaluate other buildings in each country using Japanese
and American methods of analysis.
4) Continue to evaluate differences and similarities in various

analysis methods used in each country.
5) Develop recommended evaluation and screening methods for

determination of seismic resistance of existing buildings
and identify research needs to improve evaluation techno-
logy.
6) Consider planning for a final workshop to be held in the

U.S. in conjunction with the 18th Joint Meeting, and arrange
for* publication of a Final Repoht for distribution at the
18th Joint Meeting.
710
" )
(E) LAND USE PROGRAMS FOR CONTROLLING NATURAL HAZARD EFFECTS
Date : May 22, 1985

Tsukuba Science City, Ibaraki Japan ,
Attendees : Japan Side- K. Kawashima (Chairman) (PWRI

S. Hattori (BRI)
T. Iwasaki (PWRI)
M. Nagaoka (GSI)
Y. Sugimura (BRI)
U.S. Side- G. R. Fuller (Temporary chairman) (HUD)

R. D. Borcherdt (USGS)
1) The US-Japan Workshop on "Urban Earthquake Hazards Reduc-

tion," sponsored by the U.S. Earthquake Engineering
Research Institute (EERI) was held at Stanford Univ. in
July 1984. Attendees included UJNR members Dr. Kazuhiko
Kawashima and Dr. John B. Scalzi, and representatives of
the Government of Shizuoka Prefecture: Messrs. Toshiroh
Sugiyama and Katsunori Imado. The final report from EERI
is to be published in July 1985 and will be furnished to
Task Committee (E) members.
2) EERI Report on "The Anticipated Tokai Earthquake" was

published in June 1984 and was distributed to Task Commit-
tee (E) members. Copies will be obtained from EERI for new
Task Committee (E) members.
3) Several papers pertaining to Task Committee (E) were

presented at the 17th Joint Panel Meeting:
a. Session 3, Theme IV, "Liquefaction."

b. Session 9, "Survey of Recent Chilean Earthquake."
c. Session 10, Theme III, "Storm surge and Tsunamis."
d. Session 11, Theme V, "Nihonkai-Chubu and Naganoken-Seibu
Earthquakes .
Following the discussion, the Task Committee recommends the following

coordinated works between the U.S. and Japan.
1) Both sides will continue to exchange technical information

on the following subjects:
a. Seismic risk maps, socio-economic programs, and studies

of recent earthquakes.
b. Planning methodologies for earthquake disaster mitiga-
tion in large populated cities.
c. Geological conditions, soil liquefaction potential,
seismic risk, and tsunami risk maps.
711
2) Both sides will continue to promote coordinated research
and exchange of specialists in land use programs for earth-
quake hazard mitigation.
3) The U.S. Chairman will furnish to Task Committee members

future reports on land use programs being coordinated or
sponsored by the Federal Emergency Management Agency
(FEMA), the Interagency Coordinating Committee of the
National Earthquake Hazards Reduction Program (ICC/NEHRP),
the U.S. Geological Survey (USGS), the National Science
Foundation (NSF), or the National Academies of Science and
Engineering.
712
(F) DISASTER PREVENTION METHODS FOR LIFELINE SYSTEMS
Date : May 23, 1985

Attendees : Japan Side- T. Iwasaki (Chairman) (PWRI)

K. Kawashima (PWRI)
T. Murota (BRI)
S. Noda (PWRI)
K. Ohtani (NRCDP)
Y. Sasaki (PWRI)
U.S. Side - M.P. Gaus (Temporary Chairman) (NSF)

S.K. Takahashi (NCEL)
A US-Japan Workshop on Seismic Behavior of Buried Pipelines and

Telecommunications Systems was held on December 5-7, 1984, at the
Public Works Research Institute, Tsukuba Science City, Japan, with the
participation of more than 40 experts on lifeline earthquake
engineering from both countries The workshop reviewed the current
.
knowledge, analytical, experimental and field observations of seismic

effects on buried pipelines and telecommunications systems, and had
presentations of 17 technical papers on current reserach in the area of
buried pipelines and telecommunications systems. On the basis of these
reviews and paper presentations, comprenensive discussions were
conducted, and specific research areas that are most suitable for
future coordinated research between the US and Japan were identified.
The proceedings of this workshop including the 17 technical papers and
the resolutions was published by Columbia University, and will be
distributed to appropriate members of the UJNR Panel on Wind and
Seismic Effects, participants of the Workshop, and other interested
parties. Also an outline of the workshop was presented at the 17th
Joint Meeting of the UJNR Panel on Wind and Seismic Effects.
In Japan, the Public Works Research Institute is conducting a 5-year

(1981-1985) research project on the development of repairing technology
of seismically damaged structures which includes assessment of damage
degree and proposal of appropriate repair methods for buried sewer
pipelines. PWRI is also conducting a dense instrument array
measurement program to identify ground motions applicable to seismic
design of buried pipeline structures. An evaluation of the economic
losses due to minimal cargo movements because of structural damages to
port and harbcr facilities following the Ninonkai-Chubu Earthquake of
1983 was conducted by the Port and Harbour Research Institute.
713
:
In the U.S., the National Science Foundation has continued to sponsor

research on risk analysis of lifeline systems and effects of ground
deformation waves on underground pipelines. Papers submitted at the
17th UJNR meetings by the U.S. were on dynamic stresses and displace-
ment in a buried tunnel and seismic vulnerability of lifeline facili-
ties using CADDS-CAE system. A study on gas and liquid lifelines has
been completed by a special committee of the American Society of Civil
Engineers. The committee report on a manual of recommended practice
for seismic design of gas and liquid lifelines has been published by
the American Society of Civil Engineers.
1) The members of the Task Committee will facilitate the

continued exchange for technical information and
cooperation in research on specialized problems listed as
follows
a. Seismic damages to lifeline systems

b. Seismic design procedures for lifeline systems
c. Procedures to detect and inspect damages to buried
pipelines
d. Repair and retrofit methodology for lifeline systems
e. Estimation of reliability of lifelines
f. Management and public education on importance of lifeline
systems
g. Investigation of needs for large-scale testing of life-
line systems
2) It was agreed that the knowledge derived from the US-Japan

Workshop of December 1984 was significant. Therefore, it
is recommended to hold another u'5-Japan Worksnop on
lifeline earthquake engineering in conjunction with the
18th UJNR Joint Meeting in Washington, D.C. in May 1986.
It is also proposed that Task Committee (F) members be
encouraged to participate in the lifeline earthquake
engineering session at the Third U.S. National Conference
on Earthquake Engineering in Charleston, South Carolina in
August 1986.
m
. . ) )
(G) WIND CHARACTERISTICS AND STRUCTURAL RESPONSE
Date : May 23, 1985

Tsukuba Science City, Ibaraki Japan ,
Attendees : Japan Side- K. Naito (Chairman) (MRI

K. Masamura (MRI)
T. Murota (BRI)
N. Narita ( PWRI
K. Yokoyama ( PWRI
U.S. Side- M.P. Gaus (Temporary Chairman) (NSF)
Noting that
High winds frequently cause loss of life and extensive

a.
property damage,
b. The knowledge derived from the exchange of high wind
data and information on wind effects can be useful
inreducing loss of life and property damage,
c. Insufficient knowledge exists on the effects of high
windson structurres and methods for modeling these
effects, and
d. The needs for wind data for various applications and new
projects are increasing.
The Task Committee carried out the following activities;
1) The following documents were exchanged:
a. Report entitled "Building Damage in South Carolina

Caused by the Tornadoes of March 28, 1984", published by
Committee oh Natural Disasters Commission on Engineering
and Technical Systems, National Research Council, U.S. A.
b. Paper entitled "Features of Ocean Wind Fluctuations
during Typhoon Passages" by G. Naito and H. Takahashi,
c. Paper entitled "The Variation of Gust Structure with
Height up to 200m" by K. Naito, T. Hanabusa and T.
Fuj itani
2) Some qualified instruments to measure high winds at coastal or

sea regions were installed or prepared in Japan: Kashima coast
observation with several masts 20 to 30 m high, an off-shore tower
20 m high at Hiratsuka, wind measurements at long span bridges in
the Setonaikai Sea, et al
715
3)
Coastal wind measurements were considered in the U.S.A. with the use
of portable equipment. Concerning high wind effects, particular
investigations were started in the U.S.A. on various projects, such
as
wind loading and strength prediction of glass cladding, wind loading
and response of transmission towers and wind pressure difference
between the interior and exterior of buildings.
I I. Future Programs
The Task Committee hereby proposes to carry out the following programs:
1) Exchange observation reports of high winds.
2) Encourage interaction between meteorologists and engineers

to identify the types of wind data required for future use
in establishing extreme wind distributions; in determining
wind loadings on buildings and structures; in understanding
the urban wind climate; and in considering design issues
involved with the wind generation of energy.
3) Encourage the establishment of minimum requirements for the

simulation of atmospheric boundary layers in wind tunnels
and the exchange of boundary layer wind tunnel test
results, including comparisons with corresponding
full-scale situations.
4) Exchange information on critria, techniques and instrument-

ation for structural modeling in boundary layer wind
tunnels. Also encourage the exchange of information of
measurements made on actual structures arc on coimpu*' isons
with predictions from wind tunnel studies of *hese
structures.
5) Continue the exchange of engineers and meteorologists, and

make available research facilities for mutual use.
716
. . )
(H) SOIL BEHAVIOR AND STABILITY DURING EARTHQUAKES
Date May 22, 1985

Tsukuba Science City Ibaraki Japan
, ,
Attendees : Japan Side - Y. Sasaki (Chairman) (PWRI)

T. Iwasaki (PWRI)
Y. Koga (PWRI)
N. Matsumoto (PWRI)
K Oshima
. (PWRI)
Y. Sugimura (BRI)
S. Iai ( PHRI
T. Kuwabara (PWRI)
M. Matsumoto (PWRI)
U.S. Side - A.G. Franklin (Chairman) (WES)

I . Activities and Accomplishments to Date
1) Exchange of relevant documents on soil behavior during

earthquakes was made. Eight papers were submitted to the
Joint Meeting from the Japanese side and three papers were
submitted from the U.S. side.
2) Under the auspices of the UJNR, Dr. E. L. Harp and

Dr. D. K. Keefer of the USGS inspected damaged areas in
,
Nagano Prefecture with Japanese Panel members from the PWRI

immediately following the Naganoken-Seibu Earthquake of
September 14, 1984.
3) Members from both sides of the Task Committee exchanged

information for preparing the workshop on in-situ testing
for soil liquefaction susceptibility.
4) The PHRI provided to representatives of the U.S. side

information on Japanese practice in seismic design of
breakwaters and jetties.
1) The Task Committee will hold the workshop on in-situ

testing methods for evalution of soil liquefaction
potential from 17 through 20 August, 1985, in Berkeley,
California
2) Information on U.S. practice in seismic design of

breakwaters and jetties will be provided to the Japanese
side
717
3 ) When requested, the Task Committee will assist in
arrangements for visits to earth embankments including
fill dams and foundations which have been subjected to
significant ground motions.
4) The Task Committee encourages the maximum exchange of

information on the following areas of earthquake
engineering which are currently of intense interest
to both U.S. and Japanese engineers.
a. Liquefaction potential of soils

b. In-situ testing methods for evaluation of soil
liquefaction potential
c. Seismic stability of embankment structures and natural
slopes
d. Remedial treatment of embankments and foundations to
improve seismic stability
e. Seismically induced permanent displacements in earth
structures
f. Seismically induced settlements of soils and
foundations
g. Earthquake effects on semi-buried concrete structures
h. Earthquake effects on retaining walls
i. Centrifuge testing facilities and methods
j . Field performance data showing full-scale seismic
response of earth dam
5) The Task Committee requests that the Panel encourage visits

and exchanges of engineers and scientists between U.S. and
Japanese goverment research organizations, and that the
Panel assist such exchanges by providing the channels f
communication for arrangements to be made.
718
)
(I) STORM SURGE AND TSUNAMI
Date : May 22, 1985

Attendees : Japan Side- Y. Goda (Chairman) (PHWR)

M. Ichikawa (MRI)
K. Masamura (MRI)
T. Uda ( PWRI
U.S. Side - J.F. Lander (Temporary Chairman) (NOAA)
1) In July, 1984, Dr. Goto from Tohoku University completed a

six-weeks visit to Texas A & M University under an
exchange program between the universities funded by the
National Science Foundation (U.S.).
2) A copy of the Proceedings of the Tsunami Symposium of the

Hamburg IUGG General Assembly published by NOAA (U.S.) was
sent to the Japan side.
3) The U.S. side received a copy of "Survey of the

Nihonkai-Chubu Earthquake", Ministry of Education, Pub.
No. 58022002 from Prof. Kajiura.
4) N0AA(U.S.) published multi-color wall map "Tsunamis of the

pacific 1900-1983" and sent copies to the Japan side.
5) Technical information of data basis and studies on tsunamis

produced by the Nihonkai-Chubu Earthquake were presented
during the 17th Joint Meeting.
1) The Task Committee recommends that the program of exchange

tsunami-related information will be continued and to do the
same for storm surge. The U.S. side will specifically
supply the Japan side with a copy of its digital history of
tsunami occurrence which the Japan side will review
particularly for accuracy and completeness of data on
Japanese tsunamis. This may include a working visit to the
National Geophysical Data Center (NGDC) by a Japanese
expert upon the request of the U.S. side.
2) The Japan side (PWRI) is planning to send copies of

photographs or slides taken right after the Nihonkai-Chubu
Earthquake Tsunami to NGDC. This will improve the under-
standing on tsunami effects in the U.S. side.
3) The Task Committee proposes that investigators in the U.S.

and Japan explore the possibility of undertaking specific
joint projects by the two countries, results of which may
benefit disaster-prone countries worldwide. The U.S. side,
in particular, will try to organize a workshop composed of
experts of the both sides on computational and experi-
mental methods in tsunami and storm surge.
719
)
(J) WIND AND EARTHQUAKE ENGINEERING FOR TRANSPORTATION SYSTEMS
Date : May 23, 1985

Tsukuba Science City, Ibaraki Japan
,
Attendees : Japan Side- H. Tada (Chairman) (PWRI

T. Iwasaki (PWRI)
K. Kawashima (PWRI)
N. Narita (PWRI)
K. Oshima (PWRI)
Y. Sasaki (PWRI)
H. Shinohara (PWRI)
K. Yokoyama (PWRI)
U.S. Side - H.S. Lew (Temporary Chairman) (NBS)

J.H. Gates (Cal. Tran.)
I . Activities and Principal Accomplishments to Data
The major accomplishments of Task Committee (J) during the past year
were as follows;
1) Summaries of wind and earthquake research on bridges con-

ducted by both countries were presented to the 17th Joint
Meeting, and exchange of information was made.
2) The coordinated research study on the seismic performance

of bridge piers and columns was commenced in 1983, and the
following experiments were made.
a. the performance of reinforced concrete piers and columns

subjected to dynamic cyclic loading
b. model tests on the failure of reinforced concrete piers
c. testing full-scale concrete columns and associated small-
scale model tests
d. the behavior of concrete filled steel tubes
3) The report of the U.S. group on the U.S. National Plan to

improve the seismic performance of highway bridges has been
published and sent to the Japanese Task Committee.
II. Future Programs
As a result of comprehensive discussions, the Task Committee (J) hereby

resolves to carry forth with the following programs:
1) Continue with the coordinated experimental research study

on the seismic performance of bridge piers and columns.
2) Initiate a coordinated research study both theoretical

and experimental on the seismic, aeroelastic and aerodynamic
response of cable-supported bridges.
720
. .
3) The Second U.S. - Japan Bridge Workshop is going to be held

in Berkeley, California, in August 19-22, 1985 with the
cooperataion from organizations concerned in both countries
in order to exchange the technical information
on bridge engineering and discuss the future cooperative
programs. Some of the subjects to be discussed include;
a. Research on aerodynamic stability of suspension bridges

and cable-stayed bridges
b. Laboratory experiments on the seismic behavior of
reinforced concrete piers and columns including
concrete-filled steel tubes
c. Repair procedure for seismically damaged bridge
structures
d. Experimental field testing of bridges including forced
and ambient vibration measurements
5) e. Experimental and analytical studies on the effects of
soils on the behavior of bridge structures
f. Strengthening procedures for existing bridges
g. Strong motion instrumentation of bridges, correspond-
ing data processing and parameter identification
h. Seismic design and construction procedures of highway
bridges
4) Encourage the continued exchange of information and research

personnel through the Panel on Wind and Seismic Effects of
UJNR
Dr. H.Tada, Chairman of Task Committee (J), will stay in

the U.S. for one month beginning in August 9, 1985 and
visit laboratories and organizations related to the Panel
on the basis of personnel exchange program of the UJNR
Panel
721
NBS
Periodical
Technical Publications
Journal of Research— The Journal of Research of the National Bureau of Standards reports NBS research
ana development in those disciplines of the physical and engineering sciences in which the Bureau is active.
These include physics, chemistry, engineering, mathematics, and computer sciences. Papers cover a broad
range of subjects, with major emphasis on measurement methodology and the basic technology underlying
standardization. Also included from time to time are survey articles on topics closely related to the Bureau’s
technical and scientific programs. Issued six times a year.
Nonperiodicals
Monographs — Major contributions the to on various subjects

technical literature the Bureau’s related to scien-
tific and technical activities.
Handbooks— Recommended codes of engineering and practice (including
industrial codes) developed safety in
cooperation with professional organizations, and regulatory bodies.
interested industries,
Special Publications— Include proceedings of conferences sponsored by NBS, NBS annual and other reports,
special publications appropriate grouping such
to this wall pocket cards, and bibliographies.
as charts,
Applied Mathematics — Mathematical

Series manuals, and
tables, of studies special interest to physicists,
engineers, chemists, biologists, mathematicians, computer programmers, and others engaged in scientific and
technical work.
—
National Standard Reference Data Series Provides quantitative data on the physical and chemical properties
of materials, compiled from the world’s literature and critically evaluated. Developed under a worldwide pro-
gram coordinated by NBS under the authority of the National Standard Data Act (Public Law 90-3%).
NOTE: The Journal of Physical and Chemical Reference Data (JPCRD) is published quarterly for NBS by
the American Chemical Society (ACS) and the American Institute of Physics (AIP). Subscriptions, reprints,
and supplements are available from ACS, 1155 Sixteenth St., NW, Washington, DC 20056.
Building Science Series — Disseminates Bureau on building materials,
technical information developed at the
components, systems, and whole structures. The methods, and perfor-
series presents research results, test
mance criteria related to the structural and environmental functions and the durability and safety
characteristics of building elements and systems.
Technical Notes — Studies or reports which are complete in themselves but restrictive in their treatment of a
subject. Analogous to monographs but not so comprehensive in scope or definitive in treatment of the subject
area. Often serve as a vehicle for final reports of work performed at NBS under the sponsorship of other
government agencies.
—
Voluntary Product Standards Developed under procedures published by the Department of Commerce in
Part 10, Title 15, of the Code of Federal Regulations. The standards establish nationally recognized re-
quirements for products, and provide all concerned interests with a basis for common understanding of the
characteristics of the products. NBS administers this program as a supplement to the activities of the private
sector standardizing organizations.
Consumer Information Series —

Practical information, based on NBS research and experience, covering areas
of consumer. Easily understandable language and illustrations provide useful background
interest to the
knowledge for shopping in today’s technological marketplace.
Order the above NBS publications from: Superintendent of Documents, Government Printing Office,
Washington, DC 20402.
—
Order the following NBS publications FIPS and NBSIR ’s—from the National Technical Information Ser-
vice, Springfield, VA 22161.
Federal Information Processing Standards Publications (FIPS PUB) — Publications in this series collectively
constitute the Federal Information Processing Standards Register. The Register serves as the official source of
information in the Federal Government regarding standards issued by NBS pursuant to the Federal Property
and Administrative Services Act of 1949 as amended, Public Law 89-306 (79 Stat. 1127), and as implemented
by Executive Order 1 1717 (38 FR 12315, dated May 11, 1973) and Part 6 of Title 15 CFR (Code of Federal
Regulations).
NBS Interagency Reports (NBSIR) —

A special series of interim or final reports on work performed by NBS
for outside sponsors (both government and non-government). In general, initial distribution is handled by the
sponsor; public distribution is by the National Technical Information Service, Springfield, VA 22161, in paper
copy or microfiche form.

Govpub C13

Uploaded by

Copyright:

Available Formats

Govpub C13

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Govpub C13

Uploaded by

Copyright:

Available Formats

A 11 ID 2 t> 23B7

U.S. DEPARTMENT OF COMMERCE/National Bureau of Standards

Wind and Seismic

The National Measurement Laboratory

The National Engineering Laboratory

The Institute for Computer Sciences and Technology

The Institute for Materials Science and Engineering

PROCEEDINGS OF THE 17th JOINT

Noel J. Raufaste, EDITOR

U.S. DEPARTMENT OF COMMERCE, Malcolm Baldrige, Secretary

Issued May 1986

The U.S. -Japan Cooperative Program in Natural Resources (UJNR) was

established in 1964 for the exchange of technical information and

experience mutually beneficial to the economics and welfare of both

held annually ever since, alternating in Japan the odd numbered

years and the U.S. in the even numbered years.

1985. Under five themes, forty-seven technical papers were presented

These proceedings include the program of the Seventeenth Joint Meeting,

the Panel resolutions, and the technical papers. Acknowledgement is

Preparation of the proceedings was partially supported by funds from

the U.S. Federal Agencies providing members to the U.S. Panel.

Noel J. Raufaste, Secretary

TABLE OF CONVERSION FACTORS TO SI UNITS

Length inch ( in) meter (m) a 1 in - 0.0254 m*

foot (ft) meter (m) 1 ft 0.3048 m

Force pound (lbf) newton (N) 1 lbf - 4.48 N

kilogram (kgf) newton (N) 1 kgf 9.807 N

Pressure pound per square newton/per square

kip per square newton per square

inch (ksi) meter 1 ksi 689SxlO^N/«*

Energy inch’pound (in-lbf) joule (J) 1 In* lbf - 0.1130 J

foot ’pound (ft-lbf) joule (J) i ft’ lbf - 1 .3558 1

Torque pound inch

Weight pound (lb) kilogram (kg) 1 lb - 0.4536 kg

a Meter may be subdivided. A centimeter (cm) is 1/100 n and a ml 111 -, , •

Papers were presented under five themes:

Theme I WIND ENGINEERING

Subjects covered in the papers presented include:

(1) characteristics of strong winds;

KEYWORDS: Accelerograph; codes; design criteria; disaster; earthquakes;

SCHEDULE OF THE 17 th JOINT MEETING xi

PROGRAM OF THE 17th JOINT MEETING xii

OPENING SESSION xvi

LIST OF MEMBERS xvii

LIST OF TASK COMMITTEE MEMBERS xxii

LIST OF U.S. PARTICIPANTS xxiii

FORMAL RESOLUTIONS xxiv

THEMES AND TECHNICAL PAPERS 1

Theme I: WIND ENGINEERING

Characteristics of Aerodynamic Forces on Bridge Decks in Smooth Flow

Field Measurements on the Deer Isle-Sedgwick Suspension Bridge 25

Wind Tunnel Modeling for Flow Around Buildings 48

Automatic Detection of Gust Fronts with a Single Doppler Radar 58

Theme II: EARTHQUAKE ENGINEERING

Strong Motion Instrumentation in Bureau of Reclamation Program 83

Array Observac ion of Strong-Motion Earthquakes in the Fuchu Area 98

Implications of Recent Advances in Instrumentation for Strong-Motion

Dynamic Behaviours of a Rockfill Dam During Earthquakes 118

Vibration Test of Richard B. Russell Concrete Dam After Reservoir