Earthquake Geotechnical Engineering for Protection and Development of

Environment and Constructions – Silvestri & Moraci (Eds)

© 2019 Associazione Geotecnica Italiana, Rome, Italy, ISBN 978-0-367-14328-2

Development of the second generation of Eurocode 8 – Part 5:

A move towards performance-based design

A. Pecker
Ecole des Ponts ParisTech, France

ABSTRACT: The Member States enquiry about Eurocode 8 Part 5, the geotechnical part of
Eurocode 8, clearly pointed out some deficiencies and lacks in the present version of the code.
One of the main criticisms was the overconservatism introduced in the code by the pseudo–
static approach. The ongoing revision aims at providing more practical and easier to use recom-
mendations and design rules but also at reducing this conservatism. The intent of the revision is
to give consideration to this aspect by attributing more importance to a displacement–based
approach rather than to a force–based approach. Displacement checks for defining the dynamic
failure of geotechnical structures like foundations, retaining structures, slopes, underground
structures will be introduced in the revised document. These criteria will be tentatively depend-
ent on the limit states to check: significant damage (SD), near collapse (NC) and damage limita-
tion (DL) to move towards performance–based–design. All these topics are presently under
discussion within the project team in charge and within the SC8 committee and it is expected
that at the time of the conference a consensus will be reached that can be presented.

1 INTRODUCTION

The Member States enquiry about Eurocode 8 Part 5, the geotechnical part of Eurocode 8,
(CEN. European Standard EN 1998-5, 2004) clearly pointed out some deficiencies and lacks
in the present version of the code. The ongoing revision aims at providing practical and easy
to use recommendations and design rules as well as to complement some technical aspects
improperly, or not, covered in the present document. For Part 5 (CEN/TC250 Eurocode 8:
Design of structures for earthquake resistance —Part 5) the revision has been initiated in Sep-
tember 2017 and is intended to be completed in June 2020 with the final draft of the code sub-
mitted to TC250 after having been reviewed by the Sub–Committee 8 (SC8) and the mirror
groups in each national country. Due to this planning, it is obvious that the present paper
cannot present the future document; furthermore, only general trends are presented with the
intent to give more details at the conference as almost two year would have passed since the
inception of the revision. The first orientations address the comments provided by the
National Standardization Bodies (NSBs) during the enquiry and result from the first discus-
sions initiated within the Project Team.
The choice of the project team’s members was dictated by the need of having a good balance
between practice and academy and between regions of low and high seismicity. It is composed
of Pr. Alain Pecker, AP Consultant and Ecole des Ponts ParisTech (France), who acts as PT
leader, Pr. Luigi Callisto, University La Sapienza, Rome (Italy), Pr. George Gazetas, National
Technical University of Athens (Greece), Dr. Amir Kaynia, Norwegian Geotechnical Institute
(Norway) and Pr. Kyriazis Pitilakis, Aristotle University of Thessaloniki (Greece).

2 REQUIREMENTS FOR THE REVISION

The general objectives of the revision of Eurocode 8 were clearly stated in the mandate from
CEN to TC250:

273
• enhance the ease of use of the document
• and reduce the number of NDPs (Nationally Determined Parameters) which should only
be related to safety; all other former NDPs that can be described by mathematical/mechan-
ical models should no longer be considered as NDPs and should be attributed a fixed value.
Only the first bullet is really relevant for Part 5 as the number of NDPs is limited and they
represent true NDPs. At the end, the only NDPs that will remain are the material factors on
the strength parameters: cohesion, friction angle, unconfined compressive strength and
undrained cyclic shear strength for cohesionless soils (presently called factor of safety against
liquefaction).
In addition to these general requirements applicable to all parts of Eurocodes, specific
requirements have been set for Part 5. These requirements have two origins and generally
concur: the comments raised by the NSBs during the enquiry and the items specifically identi-
fied in the mandate from TC250 to SC8.
During the enquiry the member states were invited to formulate comments, criticisms, sug-
gestions for the future revision of Eurocode 8. As expected, these comments were based on the
feedback from the use of Eurocode 8 in the members countries and represent valuable infor-
mation towards an improvement of the standard. 94 comments were collected by TC250 deal-
ing with almost all chapters of EN 1998–5. The main issues, relevant to the topic of this
lecture, can be summarized as follows:
• Several approaches in EN 1998–5 were judged overconservative because they are derived
from a pseudo–static force–based approach.
• The material safety factors γm need to be revisited. Presently the material factors for the
seismic situation are equal to those of the persistent and transient situations. Several coun-
tries complained that these values are overconservative; the new proposed values are still
open to discussion, but it must be noted that in any case the recommended values will
remain NDPs subject to the choice of each country.

3 BACKGROUND AND GENERALITIES ON EUROCODE 8

Although the existing Eurocode 8 was already a performance–based standard, the concept is
clarified in the new revision: the level of expected reliability of the structure is estimated and
performance objectives are defined for specified seismic actions.

3.1 Classification of structures

The choice of an appropriate level of reliability for the structure should take account of the
following:
• possible consequences of failure in terms of risk to life, injury, and potential economic losses;
• possible cause and mode of attaining a limit state;
• public aversion to failure;
• the expense and procedures necessary to reduce the risk of failure.
Examples of classification of buildings and geotechnical structures are provided in Table 1,
which describes the level of consequences and provides a classification in terms of consequence
classes (CEN/TC250 Eurocode 8. Design of structures for earthquake resistance —Part 1-1).

3.2 Performance objectives

Performance objectives are defined so that the designed structures meet specified limit
states under prescribed seismic actions. The state of damage is referred to the following four
limit states (LS) (CEN/TC250 Eurocode 8. Design of structures for earthquake resistance —
Part 1-1):

274
Table 1. Qualification of Consequences Classes
Description
Consequence of Examples for Examples for
Class consequences buildings Geotechnical structures

CC3–b Highest Buildings of installations of vital

importance for civil protection, e.g.
hospitals, fire stations, etc. and their
equipment
CC3–a Higher Buildings whose seismic resistance is of Large earth dams, deep
importance in view of the consequences excavations, tunnels for
associated with a collapse, e.g. schools, major infrastructures
assembly halls, cultural institutions
CC2 Normal Residential and office buildings Natural slopes, small earth
small buildings dams, tunnels, retaining
structures
CC1 Lower Agricultural buildings Embankments and retaining
Storage buildings walls, with height less than
3m

• LS of Near Collapse (NC) is defined as one in which the structure is heavily damaged, with
large permanent drifts, but retains its vertical-load bearing capacity; most non-structural
components, where present, have collapsed.
• LS of Significant Damage (SD) is defined as one in which the structure is significantly dam-
aged, possibly with moderate permanent drifts, but retains its vertical-load bearing cap-
acity; non-structural components, where present, are damaged (e.g., partitions and infills
have not yet failed out-of-plane). The structure is expected to be repairable, but, in some
cases, it may be uneconomic to repair.
• LS of Damage Limitation (DL) is defined as one in which the structure is only slightly dam-
aged and economic to repair, with negligible permanent drifts, undiminished ability to with-
stand future earthquakes and structural members retaining their full strength with a limited
decrease in stiffness; non-structural components, where present, exhibit only minor damage
that can be economically repaired (e.g. partitions and infills may show distributed cracking).
• Fully Operational LS (OP) is defined as one in which the structure is only slightly damaged
and economic to repair, allowing continuous operation of systems hosted by the structure
remain in continuous operation.
SD and NC limits states are considered as Ultimate Limit States. DL and OP limit states
are considered as Serviceability Limit States.

3.3 Prescribed seismic actions

Seismic actions should be specified in terms of their return periods. The attainment of the per-
formance requirements should be achieved by selecting appropriate return periods, TLS,CC,
depending on the specified limit states, LS, and consequence class, CC, of structures.
The proposed values of TLS,CC are given in Table 2 for buildings (CEN/TC250 Eurocode 8.
Design of structures for earthquake resistance —Part 1-1) and geotechnical structures (CEN/
TC250 Eurocode 8. Design of structures for earthquake resistance —Part 5); it should how-
ever be noted that these values are only proposals and that the return periods are NDPs left to
the choice of the member states.

3.4 Application to geotechnical structures

Aside from the requirements for the revision of the standard, Part 5 of Eurocode 8 (CEN/
TC250 Eurocode 8. Design of structures for earthquake resistance —Part5) presents some spe-
cific aspects and challenges that need to be addressed.

275
 Table 2. Return periods of seismic action in years
Consequence Class

Limit state CC1 CC2 CC3–a CC3–b

NC 800 1600 2500 5000

SD 250 475 800 1600
DL 50 60 60 100

First, as Part 5 is a transverse document, applicable to all other Eurocode 8 Parts, it must
cover at least both the NC limit state (for existing structures) and the SD limit state for new
structures. It has therefore been decided to provide criteria for those two limit states plus the
DL limit state. The latter is not mandatory but may be useful for the owner in some specific
situations when he wants to preserve, e.g. for economic reasons, the integrity of his structure.
Second, on the one hand Part 5 deals with purely geotechnical structures; natural slopes,
levees, retaining walls are examples of such structures. On the other hand, it also covers buildings
or equipment interacting with geotechnical structures; structure at the crest of a slope, retaining
walls supporting a crane, foundations of a building are examples falling in that category. The
latter are designated as geotechnical systems and the former as geotechnical structures.
For geotechnical systems the performance objectives are those defined for the structure itself;
for instance, foundations of a building will be classified in the same consequence class as the build-
ing and will be designed for the same seismic actions. Obviously enough, criteria to satisfy for
geotechnical systems cannot be set in the code: acceptable settlement, sliding or tilt of a founda-
tion is clearly dependent on the typology of the supported structures and it is not possible to cover
all situations. Therefore, the numerical choices of the criteria are left to the parties involved in the
project: "Rocking can be allowed provided the induced residual rotation and settlement of the founda-
tion are compatible with the performance criteria of the superstructure at the relevant limit state".
For geotechnical structures, a classification in consequence classes is proposed (Tab. 1,
CEN/TC250 Eurocode 8. Design of structures for earthquake resistance —Part 5) and seismic
actions are defined by their return periods provided in Table 2 (in fact, the return periods are
identical to those of buildings). In addition, criteria are proposed which depend on the conse-
quence class and/or on the considered limit state, e.g. strain limits in piles.

3.5 Choice of material factors

In Eurocode 8, the safety in design is introduced with the material safety factors which divide
the strength parameters to yield the design strength values. These material factors are called
γM and affect the cohesion, the tangent of the friction angle and the cyclic undrained shear
strength, also named resistance to liquefaction. A major difficulty faced in Part 5 is the con-
tinuity with Eurocode 7, which must be considered first, Eurocode 8 being the complement
for the seismic situation. Different approaches for the safety are used in Eurocode 7 which
makes a smooth transition from Eurocode 7 to Eurocode 8 difficult. The choice that is made
to make the transition smother is to fix all material safety factors, except the one on the cyclic
undrained shear strength, to 1,0. Furthermore, it is more consistent with the displacement–
based approach followed in Eurocode 8–Part 5. The user of the code must be warned that,
although material safety factors are NDPs and can be changed by the Member States, the cri-
teria that are listed in Part 5 (e.g. the parameter χ introduced in paragraph 4.1.1) are calibrated
on the assumption of material safety factors equal to 1,0.

4 METHODS OF ANALYSES AND CONSIDERATION OF PERFORMANCE–

BASED DESIGN IN PART 5

One of the main criticisms addressed to Part 5 was the over conservatism introduced by the
force–based pseudo static approach. Although it was duly recognized in the document that

276
the limit states should be expressed in terms of permanent displacements, the way it was
implemented was not fully transparent to the user. In an attempt to address this issue, the new
version of the standard introduces two approaches:
• a force–based approach in which the limit states are checked with safety factors; the safety
factors being chosen to limit the permanent displacements to acceptable values they may
take values smaller than 1,0, thereby implicitly implying permanent displacements. The
smaller the factors, the higher the permanent displacements. The force–based approach is
the reference analysis method in the standard.
• a displacement–based approach which is allowed whenever the force–based approach does
not allow the designer to satisfy the criteria or when an explicit evaluation of the permanent
displacements is required.

4.1 Force–based approach

4.1.1 General concept

In the force–based approach the seismic actions are converted in equivalent static forces and
the design actions in the seismic situation are compared to the design resistance of the system.
The fundamental inequality to satisfy is:

1
Sd
Rd ð1Þ
γRd

where Sd is the seismic design action, Rd the seismic design resistance and γRd a model
factor; γRd can also be viewed as a safety factor on resistance and, as such, is greater than or
equal to 1,0. Specifying a value of γRd smaller than 1,0 is equivalent to accepting that the
resistance be smaller than the action. However, it must be realized, as it has long ago been
pointed out by Newmark (1965), that this situation will occur for small time intervals as the
actions are not acting permanently as assumed in the force–based approach. The conse-
quences will not be a generalized instability of the system but the occurrence of residual, per-
manent deformations or displacements. By appropriate calibration of the model factor γRd
(<1,0) permanent displacements can be accepted and controlled; obviously, γRd depends on
the limit state to be checked and possibly on the consequence class of the structure. This is the
approach followed in the revision of Part 5 where instead of bearing on the resistance a coeffi-
cient χ = 1/γRd > 1,0 divides the action. Eq. (1) is rewritten as

SEd
SG 
Rd ð2Þ
χ

where it should be understood that χ only affects the seismic component of the design
action and not the other actions SG which are combined (⊕) with it. Examples on the use of
Eq. (2) are provided below.
At the present stage of development of the document, calibration of χ is not yet achieved
and will be the purpose of the forthcoming year. Values indicated in the draft document, and
copied in this paper, should be viewed as tentative values. Calibration will be achieved will
numerical analyses, involving simple rigid block models and more advanced finite element
analyses, as well as from existing statistical relationships (e.g. Bray & Travasarou, 2007; Whit-
man & Liao, 1985).

4.1.2 Seismic action

One peculiarity of the seismic action in soils is its extension. Stability of geotechnical struc-
tures involves a large volume of soils across which the seismic action is not necessarily

277
uniform: spatial incoherence and variation with depth of the seismic motion contribute to a
non-uniform motion in the volume of interest. To account for such effects, Part 5 defines the
seismic action for the study of geotechnical structures or systems by a conventional horizontal
ground acceleration defined by:

βH Sα
aH ¼ ð3Þ
χ FA

where:
Sα defined in EN 1998–1 (CEN/TC250 Eurocode 8. Design of structures for earthquake
resistance —Part 1-1) is the maximum response spectral acceleration (5% damping) corres-
ponding to the constant acceleration range of the elastic response spectrum;
FA defined in EN 1998–1 (CEN/TC250 Eurocode 8. Design of structures for earthquake
resistance —Part 1-1) is the ratio of Sα to the zero-period spectral acceleration;
βH is a coefficient reflecting the spatial variation with depth and the spatial incoherence of
the horizontal ground motion within the soil mass under consideration;
χ is the coefficient introduced in 4.1.1 reflecting the amplitude of accepted residual displace-
ments of the soil–structure system induced by the horizontal ground motion for the considered
consequence class and limit state. In displacement–based approaches χ = 1,0.
Figure 1 illustrates the rational behind the introduction of coefficient βH for a slope stability
analysis.
Annexes, currently under development, will provide simple means of computing approxi-
mate values of βH.

4.2 Displacement–based approach

Since the rules of TCEN/TC250 do not allow for alternative methods of analyses, the displace-
ment–based approach constitutes an alternative to the force–based approach when the latter is
no longer valid and does not permit the designer to fulfill the criteria. Not so many indications
are provided in the standard because use of a displacement–based approach is rather straight-
forward: displacements or deformations need to be computed and compared to appropriate cri-
teria. As explained in paragraph 3.4 the choice of the numerical values of criteria, which cannot
be specified without consideration of the structure and its destination, are opened and left to the
parties involved in the project: “the values of the acceptable residual displacements can be agreed
for a specific project by the relevant parties”. The only exception is the design of piles for which
limit strains are provided for the concrete and steel reinforcement.
However, some guidance is provided in the document for dynamic response analyses:
choice of input time histories which should be preferably chosen in a database of natural
records, modelling of the soil behavior which should at minimum be a viscoelastic linear

Figure 1. Illustration of the meaning of coefficient βH in eq. (3)

278
equivalent constitutive model, spatial discretization of the soil volume with at least 6 (fre-
quency domain) to 10 (time domain) elements per wave length, use of transparent boundaries
at the model boundaries.

5 ILLUSTRATIVE EXAMPLES

The force–based approach, whose principles have been described in paragraph 4.1 is illus-
trated below for two examples: slope stability analysis and foundations. Attention is drawn on
the fact that although numerical values are given in the following paragraphs they shall not be
regarded as definitive values. Additional calibration is required and they will be discussed fur-
ther within the PT and SC8.

5.1 Slope stability analyses

In the force–based approach the equivalent horizontal static seismic forces are evaluated by
multiplying the gravity forces by a horizontal seismic coefficient defined by eq.(3) (αH = aH/g).
The coefficient βH in eq.(3) should be related to the slope height and predominant wave length
of the seismic motion and on the location where Sα is defined. One annex (under development)
gives a relationship to calculate βH. The coefficient χ≥ 1,0 in eq.(3) should be related to the
residual ground displacements and should take different values depending on the considered
limit state (DL, SD or NC). The seismic resistance of the slope is expressed by its critical seis-
mic coefficient αC, defined as the minimum value of the horizontal seismic coefficient that
leads to pseudo–static failure. Tentative values for χare given in Table 3. As indicated before,
the values are not yet related to any permanent displacement value and are given only to illus-
trate how they should vary with the considered LS (and possibly Consequence Class); future
calibration will precise the actual values.
The same methodology is used for retaining structures, although with possible different
values for χ: the earth pressure acting on the retaining structure are calculated with the seismic
action defined in eq.(3) and retaining structures that cannot displace with respect to the
ground are calculated with χ = 1,0 while retaining structures that can displace are attributed χ
values that depend on the limit state and Consequence Class.
Displacement–based approaches should be used when evaluation of displacements is
needed. In that case, the performance of a slope should be evaluated based on acceptable
residual displacements for the considered limit states. These residual displacements due to the
seismic action may be calculated either by means of established rigid block models or by non-
linear dynamic analysis with modelling of the soil domain.
The seismic demand should be expressed as the residual displacement produced by the seis-
mic action and the seismic resistance should be expressed as the maximum acceptable residual
displacement.

5.2 Foundations
In a force–based approach, the design action effects should be calculated with eq.(4).

Table 3. Values of χ for slope stability analyses

Limit State
Consequence class DL SD NC

CC1 2,0 2,5 3,0

CC2 1,5 2,0 2,5
CC3 1,0 1,5 2,0

279
 Ω γRd
EFd ¼ EF;G þ EF;E ð4Þ
χ

where:γRd is an overstrength factor, χ reflects the amplitude of accepted residual displace-

ments as introduced in eq.(3), EF,G is the action effect of the non-seismic actions included in
the combination of actions for the design seismic situation and EF,E is the action effect of the
design seismic action calculated from the seismic analysis of the structure.
For use of eq.(4), distinction is clearly made between the verification of the foundation
elements (pile cap, footing, raft) and the verification of the soil capacity (sliding, bearing
capacity).
For foundation elements and soil capacity, when the structure is designed with low or mod-
erate ductility (DC1 and DC2) the product ΩγRd is taken equal to 1,0 (no capacity design for
the foundation elements or the foundation).
For structures designed with high ductility (DC3) the product ΩγRd is greater than 1,0 with
χ = 1,0 for checking the foundation elements. For checking the soil capacity of these struc-
tures, ΩγRd = 1,0 for sliding and 1,2 for bearing capacity; the intent here is to favor sliding
before bearing capacity failure.
The product Ω γRd is equal to 1,0 unless a capacity design approach is used for structural
design (ductility class DC3 as defined in CEN/TC250 Eurocode 8. Design of structures for
earthquake resistance —Part 1–1). For ductility class DC3, Ω γRd is larger then 1,0. In eq.(4) Ω
is the value of Rdi/Edi (design resistance of element i/design value of the action effect on elem-
ent i) of the dissipative zone of element i of the structure which has the highest influence on
the effect EF under consideration. Ω depends on the typology of the structure–foundation
system. For instance, for common foundations of more than one vertical element (foundation
beams, strip footings, rafts, etc.), the value of Ω used in eq.(4) is derived from the vertical elem-
ent with the largest horizontal shear force in the design seismic situation, or, alternatively, if a
value Ω = 1,0 is used in eq.(4) with the value of the overstrength factor γRd taken equal to 1,4.
The way to calculate Ω for other structural systems (shear walls, columns with concentric or
eccentric braced frames..) is indicated in the document.
Regarding the coefficient χ in eq.(4), it is equal to 1,0 for checking the foundation bearing
capacity of shallow foundations, to prevent excessive tilt of the foundation which may affect
the building stability, but takes values larger than 1,0 for checking the sliding stability.
Table 4 provides tentative values at this stage.
In a displacement–based approach χ should be taken equal to 1,0 and sliding may be
accepted at the SD or at the NC limit states, provided that:
a. the calculated displacements due to sliding are acceptable for the superstructure;
b. and do not adversely affect the performance of any lifelines (e.g. water, gas, access or tele-
communication lines) connected to the structure.
As explained in several instances, numerical values for acceptable displacements are not
provided because they depend on the typology and destination of the structure.
The same philosophy applies to pile foundations: the designer is allowed to design piles to
remain elastic (non–yielding piles) or to develop a plastic hinge under certain conditions
(yielding piles). For checking the soil lateral resistance of yielding pile χ values larger than 1,0

Table 4. Values of χ for sliding analyses of foundations

Limit State
Consequence class DL SD NC

CC1 1,5 2,0 2,5

CC2 1,2 1,5 2,0
CC3 1,0 1,2 1,5

280
 Table 5. Values of χ for soil lateral resistance of piles
Yielding piles

Limit State
Consequence class Non–yielding piles DL SD NC

CC1 1,0 1,2 1,5 2,0

CC2 1,0 1,0 1,2 1,5
CC3 1,0 1,0 1,0 1,2

are allowed; for non–yielding piles or for checking the axial soil resistance of any type of piles
χ is taken equal to 1,0 (see Table 5).
For displacement–based analyses of yielding piles, criteria are provided in terms of limit
compressive strength of concrete and limit tensile strain of reinforcement. These criteria
depend on the limit state to check.

6 CONCLUSIONS

The purpose of this paper is to provide the geotechnical earthquake engineering community
with the trends guiding the development of the revised version of Eurocode 8 Part 5. The
option that is clearly chosen is to reduce the conservatism of the present version by giving
more importance to performance–based design. Even in the force–based approach, which
remains the reference method of analysis, due consideration is given to this aspect by specify-
ing “safety factors” less than 1,0 to reflect the possibility, under certain circumstances and
depending on the limit state to check and consequence class of the structure, of developing
permanent (residual) displacements. A true displacement–based approach is allowed as an
alternative to the force–based approach when explicit values of permanent displacements are
needed; however, no numerical criteria are provided (with the exception of strain limits in
piles) because they are strongly dependent on the typology and destination of the superstruc-
ture; the definition of the criteria are left to the relevant parties involved in the project.
As a final conclusion, it must be stressed that the concepts presented in this paper are still
under discussion within the project team and the Sub–Committee 8 and must be regarded as an
instantaneous picture of the document at the present date of writing. In particular, all numerical
values given in the paper are tentative values which may evolve in the coming months before
publication of the final draft, due next year; one possible (probable) evolution may consist in
having the χ values introduced in the document function only of the limit states and not of the
consequence classes of the structure. It is therefore recommended that the readers do not take
them into consideration at this stage and wait for the new official draft of EN 1998–5.

REFERENCES

CEN. European Standard EN 1998-5 2004. Eurocode 8: Design of structures for earthquake resistance.
Part 5: Foundations, retaining structures, geotechnical aspects. Comité Européen de Normalisation.
Brusells.
CEN/TC250 Eurocode 8. Design of structures for earthquake resistance —Part 1-1: General rules and
seismic action, draft document.
CEN/TC250 Eurocode 8: Design of structures for earthquake resistance —Part 5: Geotechnical aspects,
Foundations, Retaining and Underground structures, draft document.
Newmark, N.N. 1965. Effects of earthquakes on dams and embankments. Geotechnique 15(2): 139–159.
Bray, J.D. & Travasarou, T. 2007. Simplified Procedure for Estimating Earthquake-Induced Deviatoric
Slope Displacements. ASCE Journal of Geotechnical and Geoenvironmental Engineering 133(4): 381–392.
Whitman, R.V., Liao, S. 1985. Seismic design of gravity retaining walls. Report GL–85–1, Dept. of Civil
Eng. Massachusetts Institute of Technology, Cambridge.

281