Fundamentals of The Design of Bamboo Structures

Fundamentals of the design of bamboo structures
Citation for published version (APA):

Arce, O. A. (1993). Fundamentals of the design of bamboo structures. Eindhoven: Technische Universiteit
Eindhoven. https://doi.org/10.6100/IR402687
DOI:
10.6100/IR402687
Document status and date:

Published: 01/01/1993
Document Version:
Publisher’s PDF, also known as Version of Record (includes final page, issue and volume numbers)
Please check the document version of this publication:

• A submitted manuscript is the version of the article upon submission and before peer-review. There can be
important differences between the submitted version and the official published version of record. People
interested in the research are advised to contact the author for the final version of the publication, or visit the
DOI to the publisher's website.
• The final author version and the galley proof are versions of the publication after peer review.
• The final published version features the final layout of the paper including the volume, issue and page
numbers.
Link to publication
General rights
Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners
and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights.
• Users may download and print one copy of any publication from the public portal for the purpose of private study or research.
• You may not further distribute the material or use it for any profit-making activity or commercial gain
• You may freely distribute the URL identifying the publication in the public portal.
If the publication is distributed under the terms of Article 25fa of the Dutch Copyright Act, indicated by the “Taverne” license above, please
follow below link for the End User Agreement:
www.tue.nl/taverne
Take down policy

If you believe that this document breaches copyright please contact us at:
openaccess@tue.nl
providing details and we will investigate your claim.
Download date: 03. Jul. 2020

bouwstenen
24
FUNDAMENTALS OF THE DESIGN
OF BAMBOO STRUCTURES
IR. O.A. ARCE-VILLALOBOS
faculteit t(!3 bouwkunde

technische universiteit eindhoven
BOUWSTENEN is een publikatiereeks van de Faculteit Reeds verschenen in de serie BOUWSTENEN
Bouwkunde, Technische Universiteit Eindhoven.
Zij presenteert resultaten van onderzoek en andere akii- nr.1
viteilen op het vakgebied der Bouwkunde, uitgevoerd in Elan, a computermodel tor building energy design,
het kader van deze Faculteit. theory and val i dation
M.H. de Wit
BOUWSTENEN zijn verkrijgbaar bij: H.H. Driessen
R.M.M . van der Velden
Publikatiewinkel 'Legenda' Eerste druk: februari 1987
Hoofdgebouw 4.92 Tweede druk: augustus 1987
Faculteit Bouwkunde
Technische Universiteit Eindhoven nr.2
Postbus 513 Kwaliteit, keuzevrijheid en kosten
5600 MB Eindhoven Evaluatie van experiment Klarendal, Arnhem
drs J. Smeets
of telefonisch te bestellen: 040- 472529 C. Ie Nobel, arch. HBO
040-472193 M. Broos, J. Frenken, A. v.d. Sanden
Eerste druk: februari 1986
Kernredaktie Tweede druk: augustus 1988
Prof. drs G.A. Bekaert nr.3

Prof. dr dipl. ing. H. Fassbinder Crooswijk
van 'bijzonder' naar 'gewoon'
Prof. ir J .W.B. Stark
drs V. Smit
Prof. dr H.J.P. Timmermans
ir K. Noort
Prof. ir J.A. Wisse
Tweede druk: augustus 1988
International Advisory Board
nr.4
Dr G. Haaijer PhD
Staal in de woningbouw
American lnstitute of Steel Constructions, lnc.
ir E.J.F. Delsing
Chicago, U.S.A.
Tweede druk: augustus 1988
Prof. ir N.J. Habraken
• Uitverkocht
Massachusetts lnstitute of Technology
Cambridge U.S.A.
nr.5
Mathematica! theory of stressed skin action in
Prof. H. Harms proflied sheetlng with various edge conditlans
Techische Universität Hamburg ir A.W.A.M.J. v.d. Bogaard
Hamburg, Duitsland Eerste druk: februari 1987
Prof. dr G. Helmberg nr.6
Universität lnnsbruck Hoe berekenbaar en betrouwbaar is de coëfficiënt
lnnsbruck, Oostenrijk k in i- ko en i- ks?
ir K.B. Lub
Prof. dr H. Hens drs A.J. Bosch
Katholieke Universiteit Leuven Eerste druk: februari 1987
Leuven, Belgie • Uitverkocht
Prof. dr S. von Moos nr.7

Universität Zürich Het typologisch gereedschap
Zürich, Zwitserland Een verkennende studie omtrent typologie en
omtrent de aanpak typologisch onderzoek
DrM. Smets J.H. Luiten arch. HBO
Katholieke Universiteit Leuven Eerste druk: februari 1987
Leuven, Belgie Tweede gewijzigde druk: februari 1990
Prof. ir D. Vandepitte nr.B

Rijksuniversiteit Gent Informatievoorziening en beheerprocessen
Gent, Belgie ir A. Nauta I drs J. Smeets (red)
prof. H. Fassbinder (projectleider)
Prof. dr F.H. Wittmann ir A. Proveniers, drs J.v.d. Moosdijk
ETH- Zürich Eerste druk: februari 1987
Zürich, Zwitserland
o mslag illustratie u it :
Repo rt on Globa l Observance
World Habitat Day 1986 . UNCHS (Habitat)
English Translation of the official text on the title page
FUNDAMENTALS
OF THE DESIGN
OFBAMBOO
STRUCTURES
THESIS
to obtain the degree of doctor at the Eindhoven University of Technology by the

authority of the Rector Magnificus, prof.dr.J.H. van Lint, to be defended in public
in the presence of a committee nominated by the Council of Deans on Tuesday
21"* September 1993 at 16.00 hrs.
by
Oscar Antonio Arce-Viiialobos
bom in San José, Costa Rica.

CJP-GEGEVENS KON1NKLUKE BIBUOTHEEK, DEN HAAG
Arce-Villalobos, Oscar Antonio
Fundamentals of the design of bamboo structures I Oscar

Antonio Arce-Villalobos. -Eindhoven : Faculteit
Bouwkunde, Technische Universiteit Eindhoven. -
(Bouwstenen; 24)
Thesis Eindhoven. -Met index,ref.
ISBN 90-6814-524-X
Subject headings: bamboo; structural design.
Photographs by B.G.J. Elfrink

English editing by Mr.S. Ralston
Cover photo: Bamboo culm connected by gluing to a wood fitting, aftera bending test.
FUNDAMENTALS
OF THE DESIGN
OFBAMBOO
STRUCTURES
PROEFSCHRIFf
ter verkrijging van de graad van doctor aan de Technische Universiteit Eindhoven,
op gezag van de Rector Magnificus, prof.dr.J.H. van Lint, voor een commissie
aangewezen door het College van Dekanen in het openbaar te verdedigen op
dinsdag 21 september 1993 om 16.00 uur
door
Oscar Antonio A ree-Viiialobos
geboren te San José, Costa Rica.

Dit proefschrift is goedgekeurd door de promotoren:
ir. W.R. de Sitter

prof dr. J. Gutiérrez
Copromotor:
Dr.ir.J.J.A.Janssen
English Translation of the official text on the title page
FUNDAMENTALS
OF THE DESIGN
OFBAMBOO
STRUCTURES
THESIS
to obtain the degree of doctor at the Eindhoven University of Technology by the

authority of the Rector Magnificus, prof.dr.J.H. van Lint. to be defended in public
in the presence of a committee nominated by the Council of Deans on Tuesday
21"1 September 1993 at 16:00 hrs.
by
Oscar Antonio Aree-Viiialobos
bom in San José, Costa Rica.

Thls thesis has been approved by the promoters:
ir. W.R. de Sitter

prof. Dr. J.Gutiérrez
Copromoter:
Dr.ir. J.J.A.Janssen
To Mylenne, Carlos-Andrés and Mauricio,
who deserve all credit for all these years of sacrifice,
continuous support and con.fidence.
They are my family, they are also my ciosest friends.
Acknowledgements:
This research project would not have been possible without the financial support of the
Eindhoven University of Technology in The Netherlands, and The Costa Rican lnstitute of
Technology (ITCR).
The author wishes to express bis gratitude to bis fellow Ph.D. students in Eindhoven, who
always had a word of support, especially to Harry Vekemans, whobas been a real friend and
a very supportive person.
The collaboration of the staff members of the Pieter van Musschenbroek Laboratory was very
important during the many days of arduous testing.
The personnet of the secretary of the Department of Structural Design of the Eindhoven
University of Technology, bas been specially helpfull. Mrs. A. H. M. Verhoef-Kanters helped
with the fmal editing of the thesis. Mrs. M. Alblas bas been a wonderfuil co-worker and
friend, helping the author and bis family in many ways, from the initial steps in Eindhoven
onwards.
The author was honoured by the acceptance of memhership by the distinguished members of
the Thesis Comission. Gratitude is expressed to them, fortheir great contribution to the quality
of this thesis, as well as their generous comments and teachings.
Thanks is given to Dr.J.J.A.Janssen, the co-promoter. He patiently listened to the author's

proposals during countless sessions, in a respectfut and positive manner.
Special acknowledgement is made to Prof.ir. W.R. de Sitter, the thesis promoter, for his
permanent encouragement, support and teaching. His broad view on engineering matters, and
bis personal qualities, have been a souree of inspiration for the author.
Table of Contents
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . v
Samenvatting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . v
Chapter 1: Introduetion . . . . . . .. . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1-Introductory remarks . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2-Thesis organization . .. . . . . . . . •. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.3-General remark . . . .. . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
REFERENCES . . . . . . .. . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
Chapter 2: Tensile strength of bamboo . . . . . . . . . . . . . . . . . . . . . .. . . . . . .. . . . . 9

2.1- Introduetion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . •. . . . . 9
2.1.1-Bibliographic background . . . . . . . . . . . . . . . . .. . . . . . .. . . . . 9
2.2.-Tangential tension capacity of bamboo . . . . . . . . . . . . .. . . . . . .. . . . . 17
2.2.1-Test specimen . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . .. . . . . 17
2.2.2-Experimental observations: . . . . . . . . . . . . . . . .. . . . . . .. . . . . 22
2.3- Parallel tension . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . .. . . . . 27
2.4-Some notes related to the modelling of bamboo elements . . . . . . .. . . . . 32
REPERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . .. . . . . 35
Chapter 3: Compression Behaviour of Bamboo Culms . . . . . .. . . . . . . . . . . . . . . . . 38

3.1- lntroduction: . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . 38
3.2 Structural behaviour: . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . 43
3.2.1.-Failure Modes hypothesis: . . . . . . . . . . .. . . . . . . . . . . . . . . . . 43
3.2.2- Lateral buckling analysis: . . . . . . . . . . .. . . . . . . . . . . . . . . . . 46
Chaptcr 4: Glueability of bamboo wood phases . . . . . . . . . . . . . . . . . . ... ..... . . 54

4.1- Introduetion : . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... ..... . . 54
4.2.-Structural performance of bamboo wood glued phases: . . . . ... ..... . . 54
4.2.1-Description of some factors and their influence on gluing
capacity: . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... ..... . . 54
4.2.2- Cyclic performance of glued bamboo wood phases ... ..... . . 59
REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... ..... . . 65
Chapter 5: Design of bamboo connections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66

5.1.-Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
5.2-Design Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
5.3-0bjective function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
5.4- Constraints . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
5.4.1-Intemal Constraints. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
5.4.1.1-Material properties. . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
5.4.1.2-Shape . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
5.4.2-Extemal constraints. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
5.4.2.1-Maximization of the use of bamboo . . . . . . . . . . . . . . . 74
Fundamentals of the Design of Bamboo Structures
5.4.2.2-Simplicity: ....................... ·. . . . . . . . . 74
5.4.2.3- Stability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
5.4.2.4- Adaptability of dimensions toa modular system: . . . . . 75
5.4.2.5- Strengtil predictability: . . . . . . . . . . . . . . . . . . . . . . . . 75
5.4.2.6-Cost effectiveness: . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
5.5-Design Profile . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
5.6.-Joint evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
5.7-Proposal fora Conneetlog System for Bamboo Elements . . . . . . . . . . . . . 76
5.8-Evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
5.8.1-0bjectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
5.8.2-Design profile . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
5.8.3-Extemal constraints . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
REPERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
Chapter 6: Design of bamboo structures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83

6.1-Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
6.2-Load systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
6.3-Bamboo's major assets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
REPERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
Chapter 7: Conclusions and further research recommendations . . . . . . . . . . . . . . . . . 97

7.1-The structure of bamboo culms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
7.2-Tangential tension capacity of bamboo . . . . . . . . . . . . . . . . . . . . . . . . . . 97
7 .3-P.arallel tension . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
7.4-Compressive capacity of bamboo . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
7.5-Criticalload of bamboo columns .....................•........ 100
7.6-Glueability of bamboo ..................................... 101
7.7-Bamboo glued conneetloos .................................. 101
7.8-Connections for bamboo structures ............................. 102
7.9-The Design of bamboo structures .............................. 102
7.10-General remark regarding bamboo as a structural material ............ 104
7 .11-General remark on future research . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
Appendix A: Experimental analysis of parallel tension . . . . . . . . . . . . . . . . . . . . . . . 106

A.1-lnfluence of the speed of loading on the tensite capacity of bamboo . . . . . 106
A2-Influence of density ...................................... 108
A3- Influence of nodes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
Appendix B: Experimental analysis of short elements ....................... 114

B.l- Introduetion ........................................... 114
B.2-Experimental results ....................................... 121
Appendix c. : Experimental analysis of bamboo wood glued phases .............. 129

C.l-Bonding specimen ........................................ 129
C.2- Load deformation relations and influencing factors ................. 131
C.3- Glued phases under the effect of dynamic loads ................... 139
ii
C.3.1-Load type one. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140
C.3.2-Load type two..................................... 140
C.3.3-Load type three ................................... 141
C.3.4-Load type four..................................... 142
C.4-General Remarks ......................................... 143
REPERENCES ...... ; ..................................... 150
Appendix O:Glued connectors for bamboo................................ 150

0.1-Calculation of bonding stresses ............................... 150
0.1.1-Basic equations of elasticity ........................... 150
0.1.2-Stress due to shear. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155
0.1.3-Effect of bending on the value of bonding stresses ........... 159
0.2-Parametric study ......................................... 163
0.3-Stresses due topure tension or pure compression .................. 165
REPERENCES ............................. ·................ 166
Appendix E: Bamboo columns ....................................... 167

E.l-Calculation of the critica! load . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167
E.l.l-Introduction ...................................... 167
E.l.2-Influence of initial twisting deformations ................. 167
E.2-Buckling Ioad of an initiaUy crooked tapered bar with variabie modulus of
elasticity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171
E.3-Theoretica! bockling load of bamboo culms . . . . . . . . . . . . . . . . . . . . . . 176
E.4- Buckling Experimental Analysis of Bamboo Columns .............. 181
E.4.1-Theoretical basis .................................. 181
E.4.2-Experiment set-up ................................. 182
E.4.2.l-Altemative procedure ........................ 185
E.4.2.1.1-Extension of the Southwell Plot to the Plot of
Strains ............................... 185
E.4.2.l.2 Experirnentallay-out ................... 187
E.5-Experimental resu1ts ...................................... 189
E.6-Calculation of the criticalload as a function of average properties ....... 191
REPERENCES ............................................. 194
Appendix F: Structural modelling and analysis of bamboo structures ............. 194

F.1-Introduction ............................................. 194
F.2-Theoretical model ........................................ 196
F.2.1-The effect of conicity ............................... 196
F.2.1.1-Case one .................................. 200
F.2.l.2-Case two .................................. 201
F.2.1.3-Case three ................................. 202
F.2.1.4-Case four ................................. 203
F.2.1.5-Solution of the equations of displacement and rotation .. 204
F.3-Parametric study ......................................... 209
F.4-The effect of the nodes ..................................... 226
F.4.1-Parametric study on node influence on the stiffness. . . . . . . . . . . 229
F.5-Simplified ca1culation of bamboo structures. . ............' ......... 233
F.6-Experimental analysis of the behaviour of bambOQ structures. . ......... 235
iii
REPERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . -......... 247
Appendix G; Geometrie parameters of bamboo culms . . . . . . . . . . . . . . . . . . . . . . . 248

G.l-Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 248
0.2-Extemal diameter and thickness . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 248
0.3-Area and second moment of area . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 250
SYMBOLS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 254
Curriculum Vitae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 261
iv
Summary
Summary
This thesis presents the results of the theoretica! and experimental analysis of bamboo culms,
subjected to tensile and compressive loads, or acting as beam columns in frameworks.
Glueability of bamboo wood phases is included as wen. The conneetion of bamboo culms is
also analyzed and a proposal based on the use of glue is made.
Evaluation and synthesis is made by presenting some general considerations for the structural
design and analysis of bamboo frameworks.
Keywords: Bamboo I mechanica! properties I structural design I compression capacity I

tensile capacity I bamboo connections I gluing I structural analysis I bamboo research I
bamboo construction I bamboa testing I buckling I
Cover photo: Test specimen of a bamboa wood glued conneetion
Samenvatting
Dit proefschrift geeft de resultaten weer van de theoretische en experimentele analyse van
bamboe stammen, onderworpen aan trek en drukbelasting, of gebruikt in raamwerken.
De mogelijkheid van het lijmen van bamboe aan hout is onderzocht. De verbinding van
bamboe stammen is geanalyseerd en een voorstel gebaseerd op het gebruik van lijm is
gemaakt.
Evaluatie en synthese is verkregen door het weergeven van enkele algemene overwegingen
met betrekking tot het constructief ontwerpen en analyseren van bamboe raamwerken.
Trefwoorden: Bamboe/ mechanische eigenschappen I konstruktief ontwerp I drukcapaciteit

I trekcapaciteit I bamboe verbindingen /lijmen I constructive analyse I bamboe onderzoek I
bamboe bouw I bamboe proeven I knik
Foto voorpagina: Proefstuk van een lijmverbinding tussen bamboe en hout
V
Chapter 1: Introduetion
1.1.-Introductory remarks
This thesis reports the results of a foor year research project on bamboo as a structural
material carried out at the Eindhoven University of Technologyin The Netherlands. This is
of course not the ftrst time such an enterprise has been undertak:en. On the contrary, as shall
be seen from some of the bibliographic references accompanying the thesis, the subject of
structural properties of bamboo has been a research matter since the beginning of this century.
Bamboo has attracted people's attention for centuries, and many publications have been
devoted to the description of the hundreds of applications that mankind has given to this
Gramminae1•
Applications in the fteld of civil engineering are neither rare nor new. Bridges, panels for
roofing and walling, fun or split culms as earth supporting structures, fun culms as beams or
columns, as wen as elements for the construction of frames and trusses are among the many
applications described in the literature.
Bamboo grows very easily and yields high productivity levels. Moreover, its cultivation does
not harm the environment, because among other teasons culms can be taken from the plant
from time to time onder a rotational scheme. In this way, the plant itself does not die. lnstead,
careful harvesting revitalises its growing. lt can be cultivated in relatively poor soils, helping
to control erosion by water and wind. 1t is therefore a renewable resource.
Nevertheless, people have systematicany decrèased the use of bamboo for construction
purposes. When the author of this thesis asked a Taiwanese experf, ten years ago, why less
and less bamboo was used in construction in south-east Asia, he answered that it had already
been considered to be a poor rnan's material forsome time. Indeed, one sees how the peasants
of Thailand have practically abandoned the practice of building with bamboo though the
material is available and handy as wen, in spite of the fact that this material used to be a
primary construction material all over south-east Asia and Japan. Now, one can observe that
1
The grass family.
2
Dr. Wei-chin Lin, head of the Taiwanese mission for the capacitation of bamboo
craftmen and craftwomen in Limón, Costa Rica.
as soon as rural people have access to wood or other materials, they change to them. Some
social factors are often put forward as the cause for this shift in attitude, among which is the
fact that other matenals give a sort of status to their users, while bamboo emphasizes their
position as the poorest merobers of society.
Social factors are important and they have to be accounted for in the promotion of a product
or material, but it is equally important to look at the causes of human reactions in the
matenals or technologies themselves. What can be observed in the rural areas of south-east
Asia is that people still only use bamboo as a last resource, on a "better than nothing" basis.
Is there any objective reason for this change of attitude ? When people are asked about the
properties of the material for construction purposes, the first problem that is identified by them
is the durability of the culms. Indeed, some field observations and the testimony of people
lead to the condusion that bamboo rots under some conditions in less than one year. When
complexity of the construction and availability allow, this is not a serious problem, because
replacement is cheap and straightforward. In terros of large-scale housing projects, however,
this creates a long term problem of maintenance, unless effective preservation is put in place.
So far no preservation process is readily available to the rural people, at a reasonable cost and
with a sufficient level of efficiency and security. A very important underlying assumption
in this project is that this problem is being attended to by other researchers in different
parts of the world, and that a solution for this problem will be soon available3•
One other major reason for the social neglect of bamboo is the lack of acceptance by
professionals in the construction field. Engineers and architects prefer to work with the
determinacy of a well-known system or material, supported by solid knowledge of its
properties, backed by the existence of a minimum of code specifications on which they can
base their judgement and design decisions. Besides these elements, most engineers and
architects of the third world have a bias in favour of matenals that are traditional ones in the
North, the ones they are taught about. Actually, professionals in the Third world very
frequently suffer from the same problem of social status observed among the rural people of
3
Research has been ondertaken in several places. Construction oriented research is under
way in the Costa Rican Bamboo National Project, see for example, González ( 1992 ).
2
Fundamentals ofthe Design of Bamboo Structures
Thailand. In other words, they fmd it a bit degrading to work with such a 'rough' materiallik:e
bamboo.
There is another group too, as observed in Latin America, for whom tradition is proof enough,
that tries to promotetheuse of bamboo based more on 'beliefs' than on technica! and rigorous
knowledge, some times producing a vicious circle of repeated mistakes.
A similar problem of 'acceptance' is encountered among researchers in so-called developing
countries, who endlessly repeat the subjects colleagues from the North tackle in their research
projects, refusing to put their talent and intelligence to the salution of their own nation's
problems. In the ftrst place this occurs because in doing so, they get far more recognition
among the public and the scientiftc community.
Another social issue around bamboo is the sort of target population associated with it.
In that respect, Prof. Krijgsman4 , a member of the Committee of this thesis, once said to the
author that if one wants to promotetheuse of a certain system or material among the needy,
one should flrst demonstrate that it is also acceptable toother [ rich] memhers of a society.
The author very much agrees with that. Poor people are poor because of economie problems,
not because they lack the incentive to be better. So long as a material is regardedas a 'poor
rnan's material' it is condemned to be refused, or to be used only on a very temporary basis.
It is the author' s belief that to some extent there is a conneetion between social and technica!
problems relating to the use of bamboo in the field of civil engineering, building and
architecture. Or even more generally, in relation to overall construction purposes. The link is
the lack of a sufftcient onderstanding of the material among professionals and the general
public. As a matter of fact, it is indeed a rèal difftculty to fmd consistent, ready-to-use and
complete information in this respect
One objective fact is a clear reality, resources are becoming more and more scarce, needs are
growing faster than ever, and bamboo is available, breaking the soil and reaching towards the
sky, as vital as it bas always been, producing a substantial amount of biomass that calls for
our attention. But perhaps more important, bamboo calls for our creativity and scientiftc
4
Prof.K.J. Krijgsman, Faculty of Building and Architecture, Delft University of
Technology, The Netherlands.
3
curiosity and ingenuity.

There is information, perhaps not as abundant and wen codified as for other materials, that
tells us of the wondertul properties of this material, and in one aspect it is consistent enough
to spark imagination and admiration. The tensile capacity of bamboo is claimed to be as high
as that of reinforcing steel. Besides that, certain species grow up to 30 m high and gain an
extemal diameter of 200 mm in one year, with no need for special cultivation care or the like.
The promise of good properties and availability is a powerful reason to undertake continued
research into this material.
In this research project, the emphasis is not just on the exploration of the basic mechanical
properties, but an attempt is made to correlate these to the design feasibilities of the material.
The final goal of structural engineering is the design of effective, safe and socially
accepted structures. The flnal goal of the research reported bere is to contribute to the
design ot bamboo structures in that same spirit
Research results are presented in this thesis in a manner hopefully consistent with the above
stated aim. To explain the way in which the research is presented, the reader is referred to
figure 1.1. The processof design itself, at the bottorn of the figure, is founded on a number
of previous steps collectively named the justification process (Addis, 1990). This can be
divided into three different levels in which information about structural systems is gathered
and processed ftrSt, as indicated. It is interesting to note that at these three levels the 'design
codes' (which most of the time could better have been called 'calculation codes') are started
and developed, and specifications are given.
The intensity of the design effort during each of the steps, strongly depends on the level of
development and availability of information and the extent to which it can be regarded as
being consolidated. The researcher, on the other hand, is compelled to start at the level where
information needs are assessed as not being fulfilled. The more developed a structural system
or material is, the more attention can immediately be paid to the fmal phase of the design
process.
As statedat the International Bamboo Workshop in Cochin, Kerala, India, 1988 (Anonymous,
1988), in reference to the lack of information in certain areas, attention should be concentraled
4
Chapter 1: introduetion
[;;] êJ ~
l'l.l
INTI!RMI!DIA1B
LBVBL
MATERIAL BI..BMENTS ~
STR.UCl'URES
PROPBRTIBS CONNBCTIONS
~
~
AHALYSIS
THESIS Q 11!NSION
mMPRBSSION
BUCIWNG
CONNI!Cl10NS llUIJ!S
I
OLliBABil1TY
Figure 1.1: Stepsipbases in the structural design process as employed in this research.
on " joints with bamboo to facilitate construction; strength properties as affected by specific
applications; development of a design code for bamboo; establishment of an engineering data
base to facilitate the use of bamboo in the construction and building industry .. ". All of the
points addressed in this statement of purposes fall within the justification process of design,
in clear recognition of the lack of information at this stage.
In view of this, the major effort, both in terms of time and attention during the development
of the research reported in this thesis, has been put into the investigation of important points
related to the structural nature of bamboo. Design proposals are based on the analysis of the
results of this research.
5
Chopter 1: Introduetion
1.2-Thesis organization.
The thesis follows two different semantic approaches. In the body of the thesis itself, the
major emphasis has been placed upon the description of phenomena and results in qualitative
terms, dealing mainly with those concepts that have direct consequences for the design
process. Quantitative analysis, both theoretica! or experimental, is dealt with in the appendixes.
This set up is only diverged from when necessary for ciarity's sake.
Chapters 2, 3 and 4 describe actlvities at the micro level of material properties, dealing with
tensile and compressive strains and stresses, and the glueability of bamboo, as seen in
figure1.2. Related appendixes are A, B, and C 5•
At the micro level problems are investigated to help find an explanation for the way bamboo
is able to sustain strains and stresses. Tension both perpendicular and parallel to the fibres are
included. The roles of the matrix and of the fibres are discussed and analyzed. The analysis
of bamboo culms under compressive loading is presented in chapter three and appendix B.
Chapter 4 describes the work carried out on the glueability of bamboo wood phases, and the
possibilities of using this approach as a solution to the problem of bamboo connections.
Experimental results on gluing are summarized in appendix C.
A synthesis about design is done at two different instances. Firstly, some considerations are
made at the material element level. Chapter 5 presents a design methodology I proposal for
the conneetion of bamboo structural components based on gluing. A full description of the
designprocessis included. The analysis of a glued conneetion is presentedinappendix D.
Secondly, the designprocessis forther considered at the level of the structure as a whole. In
order to onderstand the important features of bamboo structures, the results of the investigation
of a few features are presented as follows.
Boelding and the calculation of criticalloads are included in appendix E, the analysis of
bamboo trusses and frames is described in appendix F. The content of these appendixes
5
Shear has not been included as a research subject in this thesis. The reader is referred
to Janssen, (1981) fora good discussion on this subject.
6
TENSION ~ CBAPI'ER2 !:::) APPENDIX A
[::] COMPU'.SSION
GL1JE.UILITY
~ CBAPI'ER3
~ CBAPI'ER4
!:::) APPENDIX B
!:::) APPENDIX C
[;;J~
LEVEL
DESIGN
~ CBAP1'ERS !:::) APPENDIX D
BUCILING
~ EVEL CALCULATION
RULES
;::.:::.::::... APPENDIX Jl'
APPENDIXG
I
DESIGN > CBAPI'ER'
CONCWSIONS AND
I'
CBAP1'ER 7
RESEAB.CB RECOMMKNDATIONS
Figure 1.2; Thesis content description
supports the design considerations in chapter 6. Appendix G contains a description of

statistica! analysis of relevant geometrie parameters for bamboo culms. Finally, chapter 7
summarizes the main conclusions, and discusses ideas for further research.
1.3-General remark
Evidently bamboo is a very complex material, and the culms have a complex geometry. In this
workan effort bas been made to accountforthese complexities in an inclusive manner, but
at the same time to keep modelling and descriptions as simple as possible. Parameters are
included in the resulting proposed formulae so that their effect can be studied, and the
proposed models can be adapted to specific conditions after observation of experiments. The
latter were insufficient in our case due to,the lack of sufficient experimental material (only
two species could be explored).
7
In this sense the proposed models are meant to indicate possible lines of thinking, rather than
to be final and conclusive explanations about complex relations between parameters in a
complex material. Models are proposed with the clear intention of motivating extensive
discussions among bamboo scientists and structural designers. lt is the author's hope that such
a broad discussion would allow clarification of the way these results can be used in every
specific circumstance, and that it would encourage the gathering of fundamental data needed
to validate some of the results described here. In turn it is hoped that in this way, some code
specifications will be agreed upon for bamboo, based on a better onderstanding of the
structural properties of the material.
REFERENCES:
1 Addis,W.;1990. Structural Engineering: The nature of theory and design. Ellis

Horwood, London.
2 Anonymous; 1988. Proceedings ofthe International Bamboo Workshop. Kerala, India,

in the conclusions.
3 González, G.; 1992.Técnicas de preservación de culmos y esterillas de bambû.

Seminario Centroamericano usos del Bamb11. Pocora, Costa Rica.
4 Janssen, J.J.A.; 1981. Bamboo in building structures. Ph.D. Thesis, University of

Technology of Eindhoven, The Netherlands.
8
Chapter 2: Tensite strength of Bamboo
Cbapter 2: Tensile strengtil of bamboo

2.1.· Introduetion
2.1.1-Bibliographic background
One of the more abundant pieces of information about the mechanica! properties of bamboo
is related to its tensile strength, although often very little accountis given on the procedure
followed to measure this quality, nor enough data included in reports or textbooks on the
matter.
In the following section literature is referred to, to show the way the approach to the study
of the tensile capacity of bamboo has evolved over the years. Results are quoted with the
same purpose. Much of the carried out research deals with biological aspects, which are
covered in many good papers. A complete and thorough review of this research is beyond the
scope of this thesis, though reference is made to some of the most comprehensive and
authoritative studies to support the bere reported research , but readers with a specific interest
in this field are strongly encouraged to review the existing litetature.
Meyer (1923) reports a first set of results, coming from the testing of several concrete beams
reinforeed with bamboo. lt is not clear in the paper what the appearance of the.reinforcement
was like, or whether full bamboo sterns or strips were employed. The tensile strength of
bamboo was calculated by means of formulae similar to those used for steel reinforeed
concrete, and a maximum value of 100 N/mm2 is reported for the tensile strength of bamboo.
The botanical species is not mentioned in the paper.
The same author reports that in a visit totheUpper Yangtze river, the working stress of the
twisted and plaited bamboo ropes used for towing the junks, was calculated to be about 70
N/mm2 but it could be doubled in some cases. These ropes were made of the outer layer of
bamboo culms.
Duff (1941) carried outsome tests on the different basic mechanica! parameters of bamboo
on Phyllostachys pubescens. It seems that he tested strips taken from the cuhns, and that were
reinforeed at the extremes to avoid damage by the grips. Reinforcing was enlarging the areas
in the support region gluing some extra materiaL The exact shape of the specimens is not
reported. He studied the varlation of the tensile strength according to both the position along
9
Chapter 2: Tensite strengthof Bamboo
the culm and the cross-section in the radial direction. The author reports a maximum tensile
strength in the outer layer of 342 N/mm2 and a minimum in the inner layer of 54 N/mm2 • He
also found a consistent varlation for the elastic modulus. In relation with the position along
the culm, this author found that 'the strength does not vary much from the base to about half
the full height, but that it încreases toward the top where the strength was found to be 30%
to 40% higher. The tensile strength at the nodes was found to be only 80% of the
strength at the internodes.
Ota (1954) reports what seems to be a first attempt to relate mechanica! properties to physical
properties or biologica! factors. He studied the relation between tension at failure and the
sample moisture content on Phyllostachys edulis and Phyllostachys pubescens. He studied
bamboos from different localities, finding a linear experimental formula to relate tensile
strength to moisture content. This same author summarized as well the values of average
density and tensile strength for all the specimens tested, which may allow for comparlsons
with other results. Nevertheless, careful judgement is necessary since no detail is given about
the test procedure.
Karamchandani (1959) examined the suitability of bamboo as a construction material from
a more systematic approach, accounting for a series of engineering factors. In his paper he
reports that the tensile strength for the outer layers varies from 100 to 335 N/mm2 and from
lSO to 160 N/mm2 for the inner layers. This is in very poor agreement with the result
reported by Duff (opus cit). The species tested was Dendrocalamus strictus.
A more comprehensive study on this matter was undertaken by Cox (1969) on Arundinaria
tecta. From his literature review he reports that tensile strength of bamboo depends on age,
physiological varlation of individual culms, habitat, liquid content of soil in the habitat, and
"external physical forces". In the sam'e report it is established that the tensile strength of
individual culms increases from the flrst node to the middle node and then decreases towards
the top, and that specimens from the longest internodes usually had the largest tensile strength
as well. In most cases the node was the weakest section. It is also reported that the Bamboo
Research Committee for the Bureau of Public Highways from Manila stated that tensile
strength increases with "age of the specimen and from the basal to the distal section of
10
Chapter 2: Tensile strengthof Bamboo
individual culms. Besides that, it is reported that in studies both in the United States and in
India it was found that peak values for tensite strength occurred for specimens three to four
year old. The author performed a set of tests on full culms, with special grips to avoid failure
by emshing in the supports. Rate of loading was controlled to a constant of 7 N/mm2/minute,
and the results were 110 N/mm2 on average for the maximum stress and 18670 N/mm2 for
the modulus of elasticity.
~I ~A I:r-1~.......--, -
I I L ~us-100
~140~ 50
~A PLAN VIBW
wood support
bamboa strip
ctimensions in mm
wood support
A-A
Ffgure 2.1: Specimen by Atrops.
A detailed description of a specimen for the tensile test is given by Atrops (1969), as shown
in figure 2.1.
The loading rate was set at 20.27 N/mm2/mln, and the test was carried out for both the outer
skin and the inner part of the culm, with results of 290 N/mm2 and 153 N/mm2 respectively.
No details are given on the studied species .
More recently, McLaughlin (1978) accepted the at the time well-established principle
according to which the mechanical properties of cellulose based matenals correlate very well
11
with density, and he proved that this was indeed a good approach for bamboo as well by
studying the relation between density and tensile strength for very thin samples of bamboo
tissues. This experimental work was performed on Bambusa vulgaris.
r-- 120 -----j

[1~......---__ >
C_ ____.l~
114-----.------- -----...~.1 250
dimensions in m.m
Figure 2.2: Specimen by Xiu.
Figure 2.2. shows the specimen used by Xiu-Xin (1985) to study the mechanical properties
I
of Phylostachys glauca from four different locations in China. His report includes data on
tensile strength, moisture and density, and regressionsof tensile stress and age are calculated.
Widjaja and Risyad (1985) studied the relation between anatomical and mechanica! properties
for some Indonesian bamboos.
Testing procedures were as recommended in ASTM 0143-52, 72, though the type of
12
modification and specific conditions of the tests are not described.

Soeprayitno et al (1988) carried out an analysis of certain physical and mechanica! featnres
for Gigantochloa pseudoarundinacea from two different localities. Results are summarized
in the following table.
origin av.strength av.E modulus av.density [kg/m3]

[N/mm2] [N/mm2] node i.nod
slope hili 177.9 27631 770 609
valley bottorn 149.4 19643 693 565
Table 2.1 Results from Soeprayitno (opus cit).
Prawirohatrnodjo (1988) also studied some Indonesian bamboos, ftnding average tensile
strengtbs of 297 N/mm2 and 31S N/mm2 for green and dry bamboo respectively. There is no
complete description of test procedures in the paper. Species were Bambusa arundinacea,
Bambusa vulgaris, Dendrocalamus asper, Gigantochtoa apus, Gigantochtoa ater, and
Gigantochtoa verticilliata.
Sharma (1990) reports on an experiment with the objective of investigating the possibilities
of bamboo as reinforcement for concrete T-beams. He studied Bambusa vulgoris and found
parallel tension strengtbs of 145 N/mm2 for specimens with nodes, and 200 N/mm2 for
internode samples. No details are given on the test procedure or other mechanica! aspects.
No reference was found on the tangendal strain or stress1 capacity of bamboa canes,
though many researchers and users mention the fact that bamboo is cleavage prone, being one
of its main advantages when used for handieraft purposes.
This cuerent chapter is dedicated to the description of the tensile characteristics of bamboo.
Experimental observations are described as an appendix (A), and some resulting data are
1
in the direction perpendicular to the fibres.
13
Fundamentals
introduced, to establish some reference points of further use in the following chapters.
2.1.2-Anatomical and physico-chemical background :

The properties of the culm are mainly determined by its anatomical structure. Basically the
composition at any transverse section is determined by the shape, arrangement, size and
number of vascular bundles.
V ascular bundies consist of the xyiem2 with one or two smaller protoxylem elements3 and
two large metaxylem vessels4 and the phloem 5 • 1t consists of sieve elements and companion
cells andlor parenchyma cells, often with fibres or sclereids (short sclerenchyma cell) with
thin walled, unlignified sieve tubes connected to companion cells. Four major types of
vascular bundiescan be differentiated according to Liese (Liese, 1985).
The density of vascular bundies increases with height (Grosser; Liese; 1971).
Fibres, that constitute the sclerenchymatous tissue, occur as caps of vascular bundles, but in
some species also as isolated strands. They contribute 40-50% to the total culm tissue (Liese,
opus cit) and 60-70% to the weight. They have a polyiamellate wall structure, especially the
outer skin, leading to very high tensite strength. This does not exist in other cell walls of
fibres, nor the tracheids ()f wood from dicotyledons, which gives bamboo a very high tensile
strength in the direction of the fibres.
From the point of view of the chemistry of bamboo, it can be said that the nodes contain less
water soluble extractives, pentosans ash, and lignin but more cellulose than the internodes
(Liese, opus cit).
2
Vascular tissue with the primary function of water transport and that consists of vessels
and tracheids and associated parenchyma and fibres.
~e frrst-formed primary xylem.
~he last-formed primary xylem, that matures after the organ has ceased to elongate and
bas reticulated thickened or pitted walls.
srissue with the major function of transporting metabolites, especially sugars, from
sourees to sinks.
14
According to the same author, the ash content is 1-5%, the silica content varles .5-4% on
average increasing from bottorn to top. Most silica is deposited in the epidermis.
The cellulose amounts to more than 50% of the chemical constituents. The lignin is the
second most abundant constituent
6
Wood normally contains 40-50% cellulose and 20-30% lignin, 10-30% hemicelluloses •
Cellulose is thoroughly mixed and somelimes covalently linked with lignin.

From the point of view of the physics of the material, it can be said that young one-year old
shoots have a high relative moisture content of about 120-130%7• The nodes show a lower
value than the intemodes.
The above-mentioned author explains that 'shrinkage begins right at the beginning of
seasoning as observed in different studies. It affects both the size and shape of the cross
section. From green condition to about 20% of moisture content shrinkage goes from 4 to
14% in thickness and 3 to 12% in the diameter, but no shrinkage is noticed between 70 to
40%'. One wonders about the amount of prestressing that this process creates, which would
be important in correlating small specimen testing to full culm testing, particulary in bending.
The fibres in bamboo culms are confmed to the fibre strands and sclerenchyma sheaths of the
vascular bundles.
Thick-walled mature bamboo is specially liable to crack. Collapse is a most serious seasoning
defect. Liese (opus cit) argues that this is due to the fact that the outer fibre bundies are
pressed together while the inner ones are stretched. A possible explanation may be that
because tangentlal and longitudinal stresses are developed during drying, the different
anatomical composition widthwise the thickness locates the resultant force towards the outer
skin, creating inward bending moments along the circumference, thus producing tension
perpendicular to the fibre in the inner skin, which then cracks, producing collapse.
This fact may play a significant role in dealing with the concentration of stresses at the
6
Plant cell component that is soluble in dilute alkali or hot dilute mineral acids with the
formation of simple sugars.
There are reports of 195% for Guadua s.p. in Costa Rica.
15
bearings, or at load application points, where it is therefore possible to find less than the
expected capacity.
Liese (opus cit) says that a close correlation exists between specific gravity and maximum
croshing strength, a claim that is widely supported in the literature. Ghavami (1988) also
maintains that the fibres are the main souree of strength. But fibre content and density are
rather variable. Liese (opus cit) says that besides marked variations from bottorn to top within
one culm, there are differences between individual culms from the same stand. This bas to
be taken into account in attempts to classify this material for engineering purposes.
It has been argued that bamboo is much harder than wood. In actual fact, it is a real problem
for saws and sawing. Probably the major difference between bamboo and wood in this respect
is not hardness, but the characteristics of its fibres8 • First of all, fibres are all aligned to the
axis, which means a perpendicular encounter with the facesof the saw. Second, because of
the concentration of fibres at the cutting planes, the abrasiveness of bamboo is much higher
than that of wood. Tests on this were perforrned in Malaysia ( Latiff et al; 1988). Differences
in wearing capacity between bamboo and rubber wood were as much as five-fold were found.
Friction is a relevant factor in saw-ability, thus bamboo saws should be fabricated bearing this
in mind.
Some ideas can be actvaneed in relation to the mechanism of failure when strains grow in the
direction perpendicular to the fibres.
It has been found (Jeronimides, 1976) that, for wood, the work of fracture for cracks
propagating across the grain is about 100 times greater than those propagating in the grain
direction. On the other hand, in fibre-reinforced materials, fracture working perpendicular to
the fibre is mainly due to the frictional work dissipated in pulling out fibres from the matrix.
It is argued that in the case of wood this mechanism is not available or it is very small. In
any case, since bamboo fibres are longer than wood fibres, and exhibit large diameters as
well, and on top of it all they fully contribute to the tensile capacity because of their
8
Some people believe that the difficulties for sawing are caused by the high silica content
of the skin of bamboo, but laboratory observations of the author during this research indicate
that difficulties remain, even when the skin has been removed.
16
Fundamentals Structures
orientation, it is logica! to assume that this friction mechanism is far more important in the
case of bamboo. In the radial and tangentlal directions this mechanism is not at all available,
so therefore the amount of crack energy is much less. In other words, in the transversal
direction only the matrix plays a direct role in the strength.
2.2.-Tangentlal tension capadty of bamboo

2.2.1.-Test specimen
The experimental study of this parameter is complicated because of the round shape of
bamboo, thus making it difficult to perform a test keeping control of eccentricities within the
thickness of the specimen.
A test specimen as shown in figure 2.3 was used for such purposes.
Porces are applied to the specimen by means of steel bars as indicated. In our case a constant
rate of deformation was preferred as a controlling method, and 0.6 mmlmin was maintained
allalong the.different experiments. Dimensions are indicated in the figure.
The test is actually a three-point bending test, carried out on an element that probably behaves
more like a deep beam, and that has two comparatively large holes where forces are applied.
The analysis of stresses in this type of specimen is not straightforward, because the proposed
shape creates complicated distributions.
So that this type of problem could be studied, a finite element model of the specimen was set
up, and the calculations were done using the Diana Finite Element Code (Anonymous,l991).
Axisymmetric quadratic elements were employed, and orthotropy was simulated by assuming
different valnes for the elastic moduli in the directions parallel and perpendicular to the fibres.
Completely proportional behaviour was assumed, as most of the experimental evidence
suggests is the case. Though fibres can be regarded as reinforcement, they were not modelled,
under the assumption that a brittie failure would occur in the matrix before any important
bending deformation could occur in the specimen, allowing no change for the 'reinforcement'
to act (it has to be noted that bamboo is a unidirectionally reinforeed composite, so that
reinforcement would be effective only when stresses make an angle different from 90 degrees
with the fibres).
17
force is transmitted by steel Pars

10 mm èl at tb.e arrow locations
dimensions in mm
45----1
Figure 2.3: Specimen for tangential tension [mm].
The first element of concern was to know whether it was possible to find an appropriate place
in the specimen, where strains perpendicular to the fibres could be measured. In other words,
a place where these strains might occur without major disturbances from other effects.
Figure 2.4 shows a typical distribution of tangential strains, and figure 2.6 shows the same
for shear stresses in the specimen. lt should be noted that there is an area at the mid-point
between the holes where conditions are good, since stresses there can be considered as being
purely normal to the fibres.
At the beginning of the analysis, a value for the tangentlal E modulus (perpendicular to the
fibres) had to be guessed. It could be assumed that the tigure is close to that of lignin and,
according to Janssen (1981), it bas been taken between 1000 and 2000 N/mm2 in this
research. The specimen was analyzed under these two extreme conditions, and it was found
that the relation between peak vertical stresses and average ones was equal to 2.97 on
avemge. Now, the peak value occurs, as indicated in figure 2.5, as a compressive or contact
18
0.2298-2
N.J94E-2
M.JSI!B-2
L .J23B-2
K.872B-3
J .SJ7E-3
I .162B-3
H-.193B-3
G-.S48B-3
F -.903B-3
E -.126B-2
D-.161B-2
C-.J97B-2
B -.232B-2
A-.268B-2
Figure 2.4: Tangentlal strain distribution in the specimen ( mm /mm ).
stress (see the bottorn of the left circle, where concentration of lines can be observed). On the
other hand, peak shear values occur in the region indicated in figure 2.6 by a concentration
of lines at the bottorn right of the left circle. lt was later observed that, most of the time, the
crack tends to start right at that point
Por either of the two extreme values of the tangentlal E modulus, stresses along the horizontal
19
02.44
N2.06
MI .lil!
Ll.l
K.924
I .545
I .166
H·.212
G·.S91
F-.<r!
E ·1.35
D-1.73
c ·2.11
B ·2.48
A-2.86
...
Figure 2.5: Tangentlal stress distribution in the specimen ( N/mm...2).
mid plane were found to be very small for deformations like the ones observed later in the
laboratory, close to the moment of failure. During the experimental phase, it was decided to
calculate average tangentlal stresses in the plane of failure, so that the tendency of this figure
could be compared to that of density and moisture, or any other physical property , under tbe
assumption that if such a relation existed, it would be observed even by cbecking this gross
20
ii ::~:::
~ .784E-2
~ .209E-2
I -. 367E-2
: ::~~~
. ·.267E-1
I ·.324E-1
. -. 382E-1
-.044
: -.497E-1
-.SSSE-1
: -.612E-1
Figure 2.6:Shear stress distribution.
result (because the true values were just scaled). So in what fellows it must be borne in mind
that the values of strengths are only nominal ones.
2.2.2.-Experimental observations:
In the results reported here density has been measured by immersion of the whole specimen
21
Chapter 2: Tensite strength of Bamboa
so that a better average could be achieved, but evidence suggests that the average of two
smal! samples from both sides of the specimen would equally do.
Table 2.2 presents a summary of statistics of two samples of bamboa with different origins,
Guadua s.p. from Costa Rica, and Gigantochtoa scortechinii Gamble9 from Malaysia.
The sample of Gigantochtoa scortechinii Gamble was brought to Eindhoven only for the sake
of having some indication of the possible stability of the results over the maximum strain
found for Guadua s.p., as will be explained later, but it is acknowledged here that the sample
is smal! (12 specimens picked at random from a stock in the Forest Research Institute,
Malaysia) and that more systematic tests must be undertaken befare being conclusive on
characteristic parameters for this or any other species.
In respect to the results for Guadua s.p., experiments were exhaustive, many samples were
carefully taken on different occasions, and resulting data proved to be consistent enough to
be considered reliable (in total 50 tests were done for specimens with different positions in
the culm and of different diameter, these properties being random ones).
The frrst striking figures in the table are the non-existent correlation between density and
strength and maximum strain. It is important to mention that maximum strain is an
unmistakable feature, since failure occurs in an absolutely brittie way (in the case of non
brittle failure there is always room for discussion about this figure), and density was carefully
deterrnined in this investigation.
9 Gratitude is expressed to Mr. Latiff Mohamad from the Forest Research Institute
Malaysia, who supplied us with a smal! sample.
22
Guadua s.p. Gigantochtoa sconechinii

av.strength 2.608 2.445
E modulus 2122.8 2444.7
maximum ·strain 0.001386 0.00111
av. density 795.7 ·. 788.68
r density-strength 0.3568 -0.5134
r density-strain c0.1362 0.0035
Table 2.2: Summary of results. [N/mm1,[Kg/m3],

Average ·stresses are taken.
It is convenient at this time to take a look at the possible mechanism of failure, befere
discussing other results.
Figure 2.7 shows a close-up of the radial plane of failure for a specimen of Gigantochtoa
scortechinii Gamble and figure 2.8 does the same for a specimen of Guadua s.p.. Cracks
develop joining weak points within the matrix, nonnally being situated around the
intervascular bundles. When the direction of growth of the crack crosses through a bundle,
it follows it surface or that of a fibre strand along the interface with the parenchyma cells and
other weak structures.
It seems clear that the more fibres there are present, the more likely it is that the crack is
Jonger. If, on the other hand, an extreme situation of only fibres and no matrix is taken, then
probably the strength would be very Iow as well. In other words, like in any other composite
material, there has to be an optimum combination of the number of fibres and the amount of
matrix.
But yet another factor that plays a significant role in the strength is the density of the matrix
itself, since the presence of voids would imply less energy needed to split the element.
23
Chapter 2: Tensile strength of Bamboo
Figure 2.7: Radial crack in Gigantochloa scortechinii.
The above argumentation suggests that a sort of ceiling exists in relation to density and
strength, that is to say, it is logica! tothink that density would increase the capacity, but only
to a certain maximum, beyond which it may even have a negative effect. In the other
direction, about the same can be said, because there can be a combination of few fibres and
a dense matrix (and therefore very straight cracks), which would then produce another peak
capacity, but the latter would decrease downwards from there if the density of the matrix
decreases (giving rise to an imperfect matrix).
The above argument is put forward more with the intention of motivating specialists in the
24
Figure 2.8: Radial crack in Guadua s.p ..
field because the current research project can not be exhaustive on this matter. Laboratory
observations shed some light on the fracture fashion of bamboo, enabling to indicate that the
above remarks have some objective basis.
The correlation values in table 2.2 thus indicate that we are in the presence of samples which
are very different in nature from the point of view of the way they resist cracking, though
they do have something in common. In both cases, the maximum strain is about the same.
In fact, a hypothesis test was carried out comparing the two samples, only to prove that there
is no difference between their maximum strains. A similar conclusion was reached when these
two samples were compared to another one of Bambusa blumeana from the Philippines.
25
Though more studies are needed on more species, some evidence has been found to support
the idea that bamboo is able to carry an amount of strain that is rather invariable, or at least
would be so within a smal! range.
As a matter of fact, it was this finding what finally made that focus of attention was
concentrated into the examination of critica! strains, as an interesting design criterion.
Janssen (opus cit) found that the value of Poisson modulus is very stabie for bamboo, and
with this it is possible to relate longitudinal strains and tangential ones, closing a circle that
allows one to establish a design criteria for the material.
Therefore, it is proposed here that a criteria of maximum admissible tangential strain be
eventually adopted as a limiting criteria for design purposes. This would greaüy simplify
design procedures, because, as has been previously observed by many authors, and as is
plainly validated in this research report, bamboo behaves very proportionally. Thus the
calculation of strains would be an easy matter and could function as an easy check of ultimate
capacity. This will be the subject of more discussion in the following chapters, since
validation of this must be found in different stress conditions.
The tangential maximum strains for Guadua s.p., from several samples from different
experiments, were pooled together so that a better appreciation could be achieved, especially
over the variability of the figures. Results are summarized in table 2.3.
item figure
average 1.1 057x 10" 3
standard deviation 1.6170x 10-4
standard error 5.7175xJ0· 5
Table 2.3: Descriptive stalistics for Guadua s.p. from Costa Rica
Ultimate tangential strains
It is recommended here, that similar studies be carried out for culms from other species and
origins, since a lot more data are needed to be conclusive on the generality of these findings.
26
1t is possible that different critica! values may be found for different bamboo types, but in any
case, specific design values can be calculated for each characteristic condition.
2.3.· Parallel tension:

As reported by Janssen (opus cit), a large amount of problems arise when dealing with testing
the tensite strength of bamboo, mainly due to the very weak: transversal direction of the
material (as shown in the previous section). When pressure from a grip is applied, the most
common result is a emshing effect in the area of the supports, generating very high shear
stresses in the interlibre planes.
Most specifications for the tensite test of other materials, including wood (see for example,
ISO 3345), recommend that the cross section of the specimen be reduced (gauge length), so
that the area or section of faiture can be kept under control and therefore final stresses can
be accurately calculated.
For the purpose of structural capacity, the most relevant datum is the average tensite strength
for the whole cross-section, and from this point of view it was decided that the full thickness
had to be taken into account in the specimen (as opposed to examining the variation
depending on the position across the thickness).
In figure 2.9 the transition zone from the gauge and the supporting area on a nonna! specimen
is shown. Two small sections of dy high are taken, one on each side of the limiting section
a-a in the figure and there is an increment of
tA eq.2.(1)
x=-+my
2
on both sides of strip B, for every increment y. Then, once the tensile load F is applied to the
specimen, defonnations in the y axis will occur. If perfectly proportional behaviour is
assumed then the defonnations in each one of the elementary strips are
27
Chopter 2: Tensile strength of Bamboo
ldàx
-,--+------lf----r-- t
-&' .
.I
__l_
I
I
I
I
l~.j
Figure 2.10: Transition zone between the gauge and the supporting section.
eq.2.(2)
eq.2.(3)
Due to the Poisson effect, each section of the specimen will also deform in the lateral or, in
lhis case, x direction. For strips A and B respectively this deformation is
28
Fundnmentals of the Design of Bamboa Structures
Chapter 2: Tensile strength of Bamboa
eq.2.(4)
and
eq.2.(5)
From equations 2.(4) and 2.(5) it can beseen that there is a difference between the unit lateral
deformations in each one of the successive elementary strips. This duferenee leads to the
appearance of shear defonnations which have to be mainly taken by the matrix of the
material, because of the type of structural array in bamboo. From figure 2.10 it can be seen
that the shear angle due to the difference in defonnation is
dlie"
eq.2.(6)
dy =-2-
dy
and the corresponding shear stress is
Ex eq.2.(7)
dt:=--dy
2(1 +v)
From equations 2.(4) and 2.(5)
29
eq.2.(8)
Back-substitution in 2.(7) gives
eq.2.(9)
which after integration turns out to be
eq.2.(10)
Integration gives
ENv
~ ~ y +~ (11)
4(1 +v )(tA +2my)E,t 1
Since it can be assumed that no change occurs in the section below line a-a, then
eq.2.(12)
In order to know the value of shear stresses at the root of the gauge sec ti on, then a limit value
of eq. 2.(12) has to be calculated, so
eq.2.(13)
which for more generality can also be written as
30
E"v o, eq.2.(14)
4(1 +v)E1
For average values of material parameters10 the above equation gives
eq.2.(15)
From data in this chapter it can be seen that this is too close to the maximum stress that
bamboa can carry in the interfibre planes.
That is to say, the shape proposed in ISO 3345 for wood might not be suitable fora material
lik:e bamboo, since there wil! be a tendency to produce a failure of the specimen at the root
of the gauge zone.
It seems, therefore, that a straight specimen may be more promising in the case of bamboo,
or at least one with only a very slight change in section.
An additional conclusion that can be reached from the above result, is that matenals with a
structure lik:e that of bamboo behave very poorly when initia! cracks or sudden changes in
section occur, especially in areas with a high concentration of tensile stresses. An initia! crack
may be regarded as a sudden change in section and therefore a simtlar phenomenon to the one
just described wil! occur.
In choosing a way to measure parallel tension, another problem has to be taken into account,
since large compressive stresses in the transversal direction are developed by the grips. What
is more, the difference in toughness between the steel grips in the testing machine and
bamboo is so important, that it is actvisabie to place a transition material between the grips
and the specimen to avoid damage by contact.
Based on the above and after some trials, the specimen of figure 2.11 was used.
During the development of the specimen, it was found that hard wood is necessary for the
10say tension capacity is 270 N/mm2 , Young's modulus 2000 and 14000 in x and y
respectively. ·
31
Figure 2.11: Parallel tensionspecimen for bamboo [mm].
making of the supports, since soft woods tend to split and slip under the application of loads.
Of special importance is the choice of gluing length. This very much depends on the type of
glue and of course on the amount of force to be expected 11 • Therefore, some trials are
always actvisabie to get an insight into this problems, and calculations are needed to finally
determine the total length of support.
2.3.-Some notes related to the modelling of bamboo elements

Appendix A includes the results of some experiments carried out at the Eindhoven University
of Technology, to observe some of the structural features of bamboo in parallel tension.
Figure 2.12 shows a typical stress strain curve for tension that supports the fact that bamboo
behaves proportionaJiy, and fails in a brittie manner under tension.
11 See appendix C fora way of determining glue capacity in the lab.
32
PARALELL TENSION
300 ------ -----········· ·---·-- ........................ .
250 ···················-··-··········--···································-·················· ·······-··· ···········-·······
200
~
~ 150
Ul
t3
a:
lii
100 -······-·················-····-- .. ···-························-·-·--·······-·· --·-·-····-··········· ········--····-··--
50 ····--- ....... ·······--·--··-··--····················----·----- ··-·-·· ----·········-·······---·---·-
0 3 6 9 12 15
STRAINS (X 1E-3)
Figure 2.12: Stress strain curve from a parallel tension test.
A problem that wil! later prove to be of great importance is the determination of mechanica!
and geometrie parameters for the rnadelling of the nodes. lt has on occasion been argued that
the thickness of the node is an x number of times the thickness of the internode, a fact that
should be precisely defined to be taken into the analysis.
Figure 2.13 shows a longitudinal section across the node region of a sample of Guadua s.p..
It can be observed that the real shape of the node is complex in nature, ranging from a well
defined disk to a sort of hilly plate. Thickness is smal! compared to that of the internode, and
the transition region between internode and node occurs in a smal! region. In this transition
region fibres bend towards the centre of the diaphragm, or they just stop, or, in the case of
the outermost group, they point outwardly.
All this discontinuity and complexity in the shape of the node region calls for a more indirect
way of determining its properties.
To calculate the elastic modulus of the nodes as reported in this chapter, strain gauges of 30
33
mm were located as shown in figure 2.14, so that the readings averaged the effects there. In
this way the node Young's modulus refers to that average, and whenever used for calculation
purposes, a node length of 30 mm should be used. The problem of the detennination of the
thickness can be overcome since for the length of 30 mm thickness can be considered
constant, because whatever value it takes in the diaphragm, its weight within the 30 mm
length can not be relevant.
Figure 2.13: Longitudinal section along a node.
Another important reason for following the above recommendation is that the fibres in the
transition zone can not play an important role in the longitudinal stiffness of the element,
34
because before they can take any direct tension energy they have to be straightened, and for
that purpose it is expected that very little energy is needed.
As presented in appendix A, nodes are the weakest and the softest components of the culm.
Average capacity is only about 30% of that of the intemode, and the average elastic modulus
is only 40%.
Since they are the weakest link, their properties are of paramount importance in the deflnition
of design capacities. Their influence in the elasticity of the members becomes very
bamboo strip
~
strain gauge
-- --1mm
node (re:~~=--
the test) -···---·..l
Figure 2.14: Location of strain gauges at the node region.
important as well. The relevanee of this effect is a matter of discussion in appendix F of this
thesis.
REFERENCES:
1 Anonymous; 1991. DIANA Finite Element Analysis-User's Manual. TNO Building
35
Chapter 2: Tensite strength of Bamhoo
and Construction Research, Rijswijk, The Netherlands, July.
2 Atrops, J.L.; 1969. Elastizitaet und Festigkeit von Bambusrohren. Der Bauingeniur,
Vol 44, pp. 220-225.
3 Cox & Geymayer; 1969. Expedient reinforcement for concrete for use in S.E. Asia.
Monograph.
4 Duff, C.H.;1941. Bamboo and its structural use. The Engineering Society of China,
session 1940-41, paper# ICE.. 1. Institution of Civil Engineers of Shanghai, pp.l-27.
5 Ghavamy, K.; 1988. Application of Bamboo as a Low-cost Construction Material.

Proceedings of the International Bamboo Workshop. Nov. 14-18, Cochin, India, pp.
270-279.
6 Grosser, D.; Liese, W.;1971. On the Anatomy of Asian Bamboos, with special
reference to their vascular bundles. Wood Science and Technology. Vol. 5, pp.290-
312.
7 Janssen, l.J.A.; 1981. Bamboo in Building Structures. Ph.D.Thesis, University of

8 Jeronimides, G.; 1976. The fmcture of wood in relation to its structure. Leiden
Botanical Series. N.3, pp.253-265.
9 Karamanchandani, K.P.; 1959. Role of Bamboo as a Constructional Material. Natural

Building Organization of India. Proceedings of a Symposium on Timber and Allied
Products. New Delhi, 18-22 May; pp-430-435.
10 Latif, M.A.; Tamizi, M.M.; Rashid, M.S.; Shukari, M.M.; 1988. Wear resistance of
two commercial bamboo species in peninsular Malaysia and their suitability as a
flooring material.Proceedings of the International Bamboo Workshop, Nov. 14-18,
Cochin, India, pp.223-230.
11 Liese, W.; 1985. Anatomy and Properties of bamboo. Proceedings of the International
Bamboo Workshop, october 6-14, Hangzhou, People' s Republic of China, pp. 196-208.
12 McLaughlin, E.C.; 1979. A noteon the strengthof Jamica grown barnboo. Wood and
Fiber, Vol 11, #2, pp 86-91.
13 Meyer, H.F.; Ekelund, B.; 1923. Tests on the mechanica! properties of bamboo.
Engineering Society of China. Session 1922-1923, Vol22, paper 7, pp 141-169.
36
Fundamentals of the Design of Bamhoo Structures
14 Prawirohatmodjo, S.; 1988. Comparative Strengths of Green and Air Dry Bamboo.
Proceedings of the International Bamboo Workshop, Nov. 14-18, Cochin, India,
pp.218-222.
15 Ota, M.; 1950. Studies on the properties of bamboostem (part 10). On the relation
between the tensile strength parallel to the grain and the moisture content of splint
bamboo. Bulletin Kyushu University of Forestrv. Vol 23, pp. 155-164.
16 Sharma, A.; 1990. Bamboo Reinforeed Concrete Beams. Proceedings of the IABSE
Symposium, Brussels, pp. 677-682.
17 Soeprayitno, T.; Tobing, L.; Widjaja, E.A.; 1988. Why the Sundanese of West Java
Prefer Slope-inhabiting Gigantochtoa pseudoarandinacea to those Growing in the
Valley. Proceedings of the International Bamboo Workshop, Nov. 14- 18, Cochin,
India, pp. 215-217.
18 Widjaja, E.A.; Risyad, Z.; 1985. Anatomical Properties of some Bamboos Utilized in
lndonesia. Proceedings of the International Bamboo Workshop, October 6-14,
Hangzhou, People's Repoblie of China.
19 Xiu-Xin, L.; Liu-Keqing, W.; 1985. A study on the physico mechanical properties of
culm wood of Phylostachys glauca of Shandong. Joumal of Bamboo Research, Vol
4, #2, July.
37
Fundamentals
Chapter 3: Campression behaviour of bamboo culms
Chapter 3: Compression Behaviour of Bamboo CuJms

3.1· lntroduction:
Meyer and Ekelund (1923) tested a total of 11 specimens under compressive stress, reporting
an average strengthof 39.27 N/mm2, for specimens of 120 mm height and about the same
in diameter. These authors report that two types of support conditloos were employed: a lead
cushion in between the specimen and the testing apparatus, or a support in which the
specimen was placed in direct contact with the steel plates of the machine. The type of
failure was described as a set of vertical cracks which opened wider and wider with
increments of the load. When lead was used, the outer layer of the culm penetrared the lead
cushion, showing that this layer is much stronger than the inner part of the culm. The authors
report that when no cushion was used, the outer layer came apart from the inner section by
shear, which was immediately followed by break-down.
Espinosa (1930) carried out compression tests using specimens 1200 mm and 350 mm in
height, with a speed of deformation of 0.02 mm/s, on Bambusa spinosa trom the Philippines.
The author says that to resembie "actual" construction conditions, he U-shaped one of the
ends, while the other was kept flat at an angle of 90 degrees. The results ranged from 27 to
32 N/mm2 for the 1200 mm specimens and from 51.8 to 82.8 N/mm2 for the shorter ones.
He also tested small pieces of a cross-section (9 by 15 mm and 25 mm in height), getting a
mean ultimate compressive stress of 53.5 N/mm2•
In a summary report Roudakoff (1947) presents the average compressive valnes for bamboo
from four different localities. No further details are given about the species, physical
properties or metbod of testing.
Ota (1950) studied the relationship between the percentage of structural elements and the
density and compressive strength of bamboo specimens. For that purpose specimens were
taken from the outer, the inner and the full section of the wall thickness, with the approximate
shape of a very small cube, and compression tests were carried out at a speed of 0.04 mm/s.
For moisture contents between lO and 14.7% Ota found correlations between the mass per
volume and the ultimate compression stress of the outer skin, the inner skin and the middle
part of the cross-section, for Phyllostachis edulis and Phyllostachis reticulata. The mean
38
Fundamentals of the Structures
Chopter 3: Gompression behaviour of bamboo culms
values obtained for pieces containing the whole section were 81.6 and 83.1 N/mm1
respectively for both species, for related masses of 965 and 866 kg/m3• He found that
compressive strength decreases towards the node.
Limaye (1952) carried out a matricial analysis of different factors presumably related to the
mechanica! properties of bamboo, namely moisture content, presence of node in the specimen,
position along the culm, and age.
Some comments will be made to summarize the author' s conclusioos.
In re gard to the position of the specimen along the culm and its influence, the author states
that " the main effect between the three positions is significant in all cases. But the
interactloos between ages and positions are not significant for compression and the
interactloos between the position of nodes and positions along the culm are not significant for
modulus of elasticity and compression". A condusion in this research is that the modulus of
Young is much higher for the bottorn position than for both the middle and the top position
in bamboo culms, a fact that the researcher associates with " the greater wall thickness of the
culm near the ground level" (judgement on this affirmation is left to the reader !). Seasoning
was found to be an important factor, increasing the strengthof green bamboo by more than
40%, and it was also found that there is an increase in the strength with age up to three years,
though no further increase in age was studied. The species was Dendrocalamus strictus.
Sekhar and Rawat (1956) made a first attempt to establish a standard test for compression of
bamboo specimens, calling fora ratio of 10 or less for length to wall thickness. This comes
from the review of the work by Limaye (opus cit) and others and the comparison with
specifications for wood. They also suggested that two specimens should be taken, one in the
green condition and the other being kiln dried to 12% moisture content, both originating from
the internode section of the culm.
Atrops (1969) studied the compression and elastic characteristics of bamboo by performing
a total of 108 tests using three types of specimens: a) internode b) nodes at the extremesof
the specimen and c) node at the middle of the specimen . Tests were done at a toading rate
of .33 N/mm2/s and a moisture content of 18.1% on average. The size of the specimens was
chosen according to a relation of diameter to height of 1 to 4. Maximum and minimum
39
Chopter 3: Campression behaviour of bamboo culms
strengths found for specimens type a, band c were, respectively, 522- 402, 537- 407, and
528 - 433 N/mm2•
lanssen (1981) performed a very extensive literature review and a comprehensive set of tests
on the compression capacity of bamboo culms on Bambuso blumeana. Some aspects deserve
special attention and will be summarized in the following.
The height of the specimens was 50, 100 and 200 mm, the diameter varied from 70 to 90 mm
and the wall thickness from 5 to 9 mm. The following variables were taken into account,
besides the height of the specimen:
-Moisture content
-position along the culm
-node and internode
The loading speed was set in the range of 2.7 to 3.2 kN/s or about 1.5 N/mm 2, for an
approximate speed of deformation of 0.02 mm/s.
This author reports that failure was either splitting or croshing and that no relation was
found between this and the moisture content. Results between 60 and 176 N/mm2 are
reported. 1t was found that the parameters height of the specimen and node were not
significant and a recommendation was given to search for an explanation for this effect.
The tests showed that there is no difference between the compressive strength of split
bamboos and full culms, at least for the 5 % lower boundary, and a recommendation was
made to carry out compression tests for split bamboos only, using a smaller machine.
A relation between compression strength and density was found as:
fc = 0.094 p eq.3.(1)
for 12% moisture content, and
fc=.CYJ75p eq.3.(2)
for green condition
40
Chapter 3: Gompression behaviour of bamboo culms
A partlal summary of the conclusions is:

-The moisture content is significant
-Node or internode in the specimen is notsignificant
-Height of the specimen within the area 1 to 3 times the diameter is not significant
-Instead of a full bamboo, a small piece of split bamboo may be used for the
prediction of the 5 percent lower boundary.
Xiu-Xin et al (1985) carried out a set of experiments to study the relation between
compression strength and the geographical conditions of four localities, for Phyllostachys
glauca. To account for the fact that the culms are not perfectly circular in cross-section, a
correction factor was introduced to the formula for the calculation of stresses, so that nominal
stresses were divided by a factor of
eq.3.(3)
A polynomial was found to calculate the compressive strength as a function of the age of the
culm at the time of cutting as
fc =588.4 +S3.9t -4.75t 2 eq.3.(4)
N eq.3.(S)
s=±3.72-
cm2
tnu =4.97 years eq.3.(6)
Espiloy (1985) carried out a very complete statistica! analysis of the physical and basic
mechanica! properties of Bambusa blumeana and Gigantochloa levis from the Phillipines,
41
Chapter 3: Compression behaviour of bamboo culms
following the recommendations of ASTM1 for small specimens of wood. Variables included
fibre length, fibre diameter, lumen diameter, cell wall thickness, vessel dimensions related to
moisture content, relative density, shrink:age, compression strength, modulus of ropture and
modulus of elasticity, and in turn these mechanic variables were statistically related to outside
diameter, lengtb-to-span ratio and culm wall thickness.
Prawirohatmodjo (1988) performed tests on both full culm 100 mm height and split bamboo
30 x 10 mm x thickness of the culm. The author reports that in his experiment the presence
of a node .in the specimen does not seem to affect the strength, and that moisture content
is a significant factor. No data are included in the paper about the relation of compressive
strengthand density.
By request of the Costa Rican Bamboo National Project (Chaves and Gutiérrez, 1988), Sotela
(1990) produced a handhook for 'physical and mechanica! testing of bamboo'. It includes
recommendations for the determination of density, contractions, bending, compression and
shear capacity. In general, the specifications for bending shear and compression follow
lanssen (opus cit), except that details are given on the number of specimens, the preparation
of specimens prior to testing, and the protocol for reporting results. One important feature is
the use of strain gauges on compression specimens though in a paper by the same author
(Sotela, 1992) he maintains that 'it is difficult to calculate strains from compression tests,
even with electronk devices'. In this paper the physical and mechanica! properties of Guadua
s.p. from two different localities in Costa Rica are compared. This author reeommencts that
the elastic modulus of bamboo is calculated from bending tests only according to the
recommendations of the above mentioned handbook. A comparison is made between bamboo
and some hardwoods from Costa Rica, as summarized in the following table.
1
American Society of Testing Materials.
42
-------
density [kg/m3] E modulus [kg/mm2]

Cupressus
lusitonicus 460.0 9800 41.1
Hieronyma
achoneoides 790.0 14300
Playmiscium
pinnaturn 860.0 16700 81.3
Terminalia
lucida 770.0 13000
Guadua s.p. 790.0 27700 42.0
Table 3.1: Bamboo and som~ Costa Rican wood types, after Sotela {opus cit).
The moisture content of each of these wood species was 12%, and it was 15% for the bamboo
sample.
3.2 Structural behaviour:

3.2.1.-Failure Modes hypothesis:
As described in chapter 2, the anatomical structure of bamboo culms consists of fibres and
vessels running parallel to each other and to the axis of the culm, with no radial structural
elements as in the case of wood. The culm is made of a composite material which is the
result of the combination of very strong fibres and a relatively weak matrix of lignin.
The examination of short elements from compressive tests points in the direction of at least
four possible ways failure can occur; these are the so-called failure modes.
In the first place, if a specimen is set in a testing machine such that friction is negligible,
and therefore, no horizontal reaelions are possible, the state of stresses can be pictured as
shown in figure 3.1. The top left of the figure shows the full specimen. It will be assumed
43
Fundamentals
in thls case that the length of the specimen is comparable to the diameter, so that
eccentricities due to tapering are negligible.
Detail A shows the fact that the axial stiffness varles along the thickness of the culm,
increasing towards the outside. This means that axial stresses will be mainly taken by the
outermost rings or layers of the culm, creating an uneven dis tribution of stresses, and inducing
relative tensions in the innermost rings. The overall effect of this will be the generation of
radial bending moments as shown in detail B.
The resulting compression in the outermost layers will generate, expansion of the rings in the
tangentlal direction, due to the Poisson phenomena, as pictured in detail C, and it will create
contraction in the innermost rings. The expansion of the outermost rings will eventually lead
to longitudinal crack:ing as shown in figure B.2. The relative difference between the
displacements in the outermost and the innermost rings also creates some sort of annular
crack:ing, indicated by the radial arrows in detail C, as was observed in the laboratory.
In this way bamboo in compression will fait when a critica! value of tangentlal stresses is
reached in the outer skin, initiating a longitudinal crack. This mode of faiture bas been
reported by, for example, Meyer and Ekelund (opus cit), and by Janssen (opus cit).
This faiture mode leads to the conclusion that fibre density could contribute negatively to the
strength of the specimen, since tangentlal tensile strength seems to be the more logical
mechanism, and strength can be better related to the matrix in this case (see chapter 2, section
2.2).
Secondly, as mentioned in appendix B, it is possible that due to partienlar initial conditions
of the geometry of the specimen, with an especially large influence from tapering in
combination with the presence of friction, radial inward-bending moments occur, as illustrated
in figure 3.2.
Thirdly, as explained in appendix B, when the.specimen is short enough, it is possible to
have failure by contact stresses, something close to a compression failure in a cylindrical or
cubic specimen as observed in other types of materials. This mode of failure, which might
be of interest from the point of view of the material, does not describe the way bamboo
44
ential ._.......
~ale~
ftn
axial
stiffnessc:;
Figure 3.1:State of stresses of a specimen under frictionless compression
culms are likely to react.

Fourthly, figure 3.3 shows what may happen when the slenderness is high and friction is
generared between the plates and the specimen. As the quantification of this phenomenon
might be of some interest in determining the role of the different parameters involved, some
detailed attention is given to its analysis jn the following lines.
45
Figure 3.2:Longitudinal section of a tapered specimen under compression.
3.2.2- Latera1 boelding analysis:

Por the analysis of the element shown in figure 3.4 the following premises are assumed to
hold:
1- The cross-section of the culm is a circular ring with a constant diameter and wall
thickness.
2- Ex is constant along the element.
3- E8 is constant
4- Deformations are small

5- Material is perfectly elastic.
46
Fundamentals of the Design of Bamboa Structures
Chapter 3: Campression behaviour of bamboa culms
friction
CROSS SBCilON BLOW UP

Figure 3.3:Lateral buckling
As shown in figure 3.4, two lateral displacements can be considered, e due to the lateral
expansion of the specimen, and w, due to the bending deformation of the wall of the culm.
It can be shown that the relation between tangentlal strains and radial expansion is given by
w, eq.3.(7)
E =-
e r
or
47
t
t
DETAIL A
Figure 3.4:Bearn on elastic foundation model
eq.3.(8)
Tangentlal stresses are given by
or
eq.3.(10)
Equilibrium of radial forces ón an infinitesimal sector gives
48
Chapter 3: Campression behaviour of bamboo culms
eq.3.(11)
from which we gain, after simplification and substitution
eq.3.(12)
which is the value of the reaction coefficient in the radial direction.

The sum ofmoments gives
am
..__2:!-v +n
8(w,. +e} = 0 eq.3.(13)
öx ,. x öx
which after denvation becomes
&mxx _àv,. + ni1(w, +e) =

0 eq.3.(14)
&2 öx &2
Equilibrium of forces in the radial direction implies that
àv,. eq.3.(15)
--kw
ÖX r r
0
Substitution of this value in eq.3.(14) and the fact that
&w, eq.3.(16)
mJtX = -Dx -.-
&2
gives
é1w &(w +e) eq.3.(17)

D --'+kw -n ' = 0
~ &4 7 7 x &2
49
A solution of the above equation is
w +e = li+w \.<rin 1tx eq.3.(18)

r ~ rr-·z
thus eq.3.17 becomes
eq.3.(19)
from where it can be seen that
eq.3.(20)
Taking into account that
wr = 00
then, after the proper substitutions
eq.3.(22)
The first component of the right hand side part of this equation can be recognized as the
bockling load of the culm when initial cracking exists, or when the tangentlal capacity of the
culm is negligible. The second component is, therefore, the contribution of the lateral stiffness
of the specimen. To get some idea of the participation of these teems in the whole
50
Fundamentals
Chapter 3: Campression behaviour of bamboa culms
phenomenon, an example can be examined.

Suppose that a specimen with the following characteristics is taken
N N
r ""40mm t·= 6mm E = 16000- E6 = 2000-- l = 60mm
" mm 2 mm2
For this case
Ocr = 144+456
This means that the normal recommendation, to take the length as ten times the thickness of
the specimen to prevent buckling, is far too conservative for bamboo, since the first
component of the criticalload (cracked specimen) already gives a value that is almost three
times the average compressive capacity of bamboo.
It can be concluded that lateral buckling of the walls of the specimen is not a critica! mode
in compression tests. It has tobetaken into account that the model of eq.3.(17) is even more
conservative, because it does not consicter the effect of radial reacting bending moments in
the surface of contact between the specimen and the plates of the machine. Experience
indicates that the height of the specimen should be equal to the diameter at most, in order to
avoid the influence of tapering.
Finally, some comments about the influence of the nodes on the compression capacity.
Janssen (opus cit), Limaye (opus cit), and Sekhar and Rawat (opus cit), among others,
reported that no influence of the nodes on the compression capacity could be found. In
compression tests the node impedes the lateral expansion of the specimen working in tension.
lf the low tensile capacity of the nodes is recalled, on the one hand, and on the other, the fact
that nodes are not flat disks but curved surfaces then it is expected that indeed their
contribution to the lateral stiffness of the walls is small. Even though this is the case, a
simplification can be made by assuming that they are flat disks. The complete analysis is not
51
Fundamentals
presented here, but it can be shown that under these conditions, at the maximum compressive
capacity, average tensile stress in the nodes are equal to about 30 N/mm2, about ten times
their expected capacity, which would explain the experimental findings.
REFERENCES:
1 Atrops, J.L.; 1969. Elastizitaet und Festigkeil von Bambusrohren. Der Bauingeniur,
Vol 44, pp. 220-225. ·
2 Chaves, A.C.; Gutiérrez, J.; 1988. The Costa Rican Bamboo National Project.
Procedings of the International Bamboo Workshop, Cochin, India, November 14-18,
pp.344-349.
3 Espiloy, Z.B.; 1985. Physico-Mechanical Properties and Anatomical Relationships of

Some Philippine Bamboos. Proceedings of the International Bamboo Workshop,
October 6-14, Hangzhou, People's Republic of China, pp 257-264.
4 Espinosa, J.C.;1930. Bending and Compressive strengtbs of the common Philippine

bamboo. Philippine Joumal of science.Vol 41, N. 2, ppl21-135.
5 Janssen, J.J.A.; 1981. Bamboo in Building Structures. Ph.D.Thesis, University of

6 Limaye, L.D.;l952. Strengthof Bamboo (Dendrocalamus strictus). Indian Forester,

Vol 78, pp.558-575.
7 Meyer, H.F.; Ekelund, B.; 1923. Tests on the mechanica! properties of bamboo. The
engineering Society of China. Session 1922-1923, Vol 22, paper 7, pp 141-169.
8 Ota, M.; 1950. Studies on the properties of bamboostem (part 4). Study of the form
of the specimen in the test of compressive strength. Joumal of the Jaganese Forestry
Society. Vol 32, pp. 65-69.
9 Prawirohatmodjo, S.; 1988. Comparative Strengtbs of Green and Air Dry Bamboo.
Proceedings of the International Bamboo Workshop, Cochin, India, November 14-18,
pp.218-222.
10 Roudakoff,G.; 1947. Note Sur les constructionsen bambou. Bulletin de l'Université

De l' Aurore.
52
-------~
11 Sekhar, A.C.;Rawat, B.S.; Bhartari, R.K.;l956. Strength ofbamboo: Bambusa nutans.

Indian Forester, leaflet # 147, pp.67-73.
12 Sotela,J.; 1990. Manual de ensayos ffsico-mecánicos para especfmenes de bambû.

Laboratorio de Productos Forestales, Instituto de Investigaciones en Ingenieria,
Universidad de Costa Rica, febrero.
13 Sotela,J.; 1992.Propiedades ffsicas y mecánicas de la Guadua de Costa Rica. Seminario

Centroamericano Usos del Bambó. Pocora, Limón, Costa Rica.
14 Xiu-Xin, L.; Liu-Keqing, W.; 1985. A study on the physico-mechanical properties of

culm wood of Phylostachys glauca of Shandong. Joumal of Bamboo Research,
Nanjing, Vol 4, #2, July. ·
53
Chapter 4: Glueability of bamboo wood phases

4.1· Introduetion :
No indication of previous studies on the gluing characteristics of bamboo phases could be
found in the literature. This is not a surprising fact, since the use of bamboo in furniture
making, for example, follows the tradition of joints being finely adapted pieces made to fit
each other. Elements are connected by insection through bored holes, or binding, or shaping
by heating and so forth. The grade of structural redundancy in these pieces of furniture, on
the one hand, and the refmed details on the other, make it possible to avoid fittings other than
those made of bamboo itself, and so far the possibility of gluing has not been seen as a
necessary impravement in the process. This perhaps explains why little or no attention bas
been paid to the gluing properties of bamboo.
Resins have been used in experiments for the fabrication of bamboo composites (Jindal,
1988). Urea formaldehyde and Cardanol-Phenol-Formaldehyde have been in use in the
production of ply-bamboo in India as well (Zoolagud, 1988), but no report was found on the
properties of glued phases.
As has been mentioned in previous chapters, the mechanism of failure in bamboo is splitting,
when strains reach a maximum sustainable value. From this point of view gluing seems an
attractive solution to a number of problems, like the reparation of cracks, or, more
importantly, the conneetion of elements. Glue may not only be useful for force transmission,
but to keep the fibres together at points of Ioad concentration as well.
In this chapter the structural possibilities of gluing are discussed. A more detailed description
of testing procedures and quantitative results can be found in appendix C.
4.2.-Structural performance of bamboo wood gloed phases:

4.2.1-Descrlption of some factors and their inBuence on gluing capadty:
As is explained in appendix C, forces and deformations are perfectly proportional in bamboo
wood phases (referred to as 'glued phases' in the remaining part of the thesis for the sake of
simplicity). A description of the test specimen employed during this research can also be
found in the same appendix.
54
DEFORMATION BONDING STRESS

0.59 ·-.. ... ·-......................... - .._, ......... --- -
0.49 ....................._ ..____ _
0.39
5
i 0.29
12
l':l 0.19
2 B a
0
"
STRESS N/1111'1"'2
Figure 4.1: Bonding test showing all registered defonnations. Glue type: PVAc ( white
wood glue ).
Figure 4.1 shows results of a bonding test. Each one of the curves is codified to identify the
corresponding point of measurement. The codification corresponds to pins, in such a way that
those beginning with 'wood' are related to pins onthe surface of wood, and those beginning
with 'bam' topins located on bamboo. The depth difference between the layer where bamboo
pins and wood pins were located was about 1 mm (see figure C.2). It can be seen that at the
depth where the bam pins are located very little defonnation occurs, compared to that of the
wood pins. This is an indication of the fact that only the innennost layer of bamboo is able
to take stresses from the glue , due to the low shear capacity of the matrix. Glued phases
actually fail because of shear in the interfibre surfaces.
lt can be expected that the shear behaviour of bamboo differs from point to point because of
the different concentration of fibres. Different crack lengtbs may occur as a consequence.
55
Despite of this fact attention is paid only to the gluing capacity of the inner layer of the culm
in this research report, because of a predetermined interest in transmitting forces to or from
the inside of the culms
BONDING STRESS DEFORMATION CURVE

8 --- --- --- --- -- -- --- --- ---- --- - -- ---- - -- - -- -- - - -
2 -------- --------------------------------------------------------------------------------
0 0.1 0.2 0.3 0.4 0.5

OEFORMATION m-n
Figure 4.2: Stress deformation curve.
In figure 4.2 it can be seen that the relation between stresses and deformations is very
smooth, almost directly proportional for most of the range. Failure always occurs in a rather
sudden way, and it looks as if the pretoading loop does not seriously affect the characteristics
of the curve.
An experiment was set to examine the influence of certain parameters on the bonding capacity
of the phases. Considered factors were :
- Bamboo density (represented as the average along the thickness).
- Bamboo thickness.
- Bamboo initial diameter.
56
- Type of wood either hard Of sofe .

-Type of glue, either PVAc or Araldite (in what follows called 'resin'). 2
An analysis of varianee showed that none of the above factors on its own, nor in combination
to the other ones, significantly affects the results of bonding capacity as shown in table 4.2
below.
In relation to the type of glue, improvements on its quality do not alter bonding capacity
unless peneteation pattems change, because of the low shear capacity of the inner layers of
bamboo. The experiments showed that glued phases always fail within the frrst mm of the
internal part of the culm, bamboo is thus the weakest link in the glued phase.
Table 4.1 shows the results of an analysis of varianee for all the studied factors.
Although changes in the considered factors do oot by themselves influence the bonding
capacity, it is of interest to obtain the 95% confidence limits for the means after an analysis
of variance. Such aresult is sbown in table 4.2 below.
Such separate analysis helps to visualise the effects of the diverse variables, to confrrm the
previously stated result.
The figures show that in spite of the fact that no significant difference exists between the two
types of glue, the resin seems to be less stable, yielding more scattered results than PV Ac
glue.
The experiments showed that the strength distribution was different for the two types of glue
examined, so though the average value fortheresin was slightly higher than wood glue (6.33
against 5.78 N/mm2), a material factor calculated according to equation C.l, appendix C,
produces hardly any difference in the design capacity.
Experimental results suggest that glued phases made with resin would tend to be stiffer than
those made with wood glue, but the analysis of the correspondent values at the design level
1
soft wood represented by an European conifer of 420 kgr/m 3 , hard wood represented by
a tropical hardwood of 780 kgr/m 3•
2
PV Ac following the Oerman lnstitute of Normalization ( DIN ) norm 68602B2, Araldite
AW 106 N( 3-dimethylaminopropyl)l-3 propylenediamine.
51
of stresses shows an average deformation of about 0.1 mm, without a significant difference
between the two types of glue.
For all practical purposes this deformation is little enough to produce a very stiff connection,
though this fact has to be examined in more detail for every specific application.
Allalysis of Varianee for STRENGTH
Souree Sum of d.f. Mean F-ratio Sig. level

of varlation squares square
COVARIATES .9626674 3 .3208891 .246 .8617

density .0010824 1 .0010824 .001 .9780
diameter .0043521 1 .0043521 .003 .9559
thickness .7708084 1 .7708084 .592 .4717
MAIN EFFECTS .9739144 2 .4869572 .374 .6994

woodtype .0757670 1 .0757670 .058 .8179
gluetype .7096751 1 .7096751 .545 .4891
2-INTERACTIONS .2622303 1 .2622303 .201 .6702

GLUE-WOOD .2622303 1 .2622303 .201 .6702
RESIDUAL 10.416414 8 1.3020517
TOTAL (CORR.) 12.615226 14
d.f.:degrees of freedom
F-ratio: Fisher' s coefficient.
Tab1e 4.1: Allalysis of varianee for the whole factors.
58
Std. Error Std. Error 95% Confidence

internal pooled Level for mean
Average
WOODTYPE
h 6.20106 .407181 .4312857 5.20623-7.19588
s 5.80141 .300461 .4034309 4.87084-6.73199
GLUETYPE
r 6.21919 .401665 .4312857 5.22436-7.21401
w 5.78555 .302878 .4034309 4.85498-6.71612
INTERACTION WOOD-GLUE
h r 6.34300 .732939 .5705374 5.02697-7.65903
h w 6.01180 .234167 .6587998 4.49218-7.53142
s r 6.05410 .176344 .6587998 4.53448-7.57372
s w 5.64980 .477621 .5103042 4.47271-6.82689
TOTAL 5.98791 .294624 .2946242 5.30832-6.66751
Table 4.2 : 95% confidence limits for the mean S1RENGTHS after an analysis of
variance.
4.2.2- Cyclic performance of gloed bamboo wood phases :

A study was ondertaken to observe the performance of gloed phases under prescribed cyclic
loadings.
A ftrst experiment was set to determine whether the application of a certain number of cycles
onder the rate of toading specifted in ISO 6891-1983 (E) would weak:en the static bonding
capacity.
Initially, it was supposed that a top ftgore of 20 cycles would be enough and satisfactory for
the comparison, but actually as many as 90 cycles were applied to the specimens, with an
amplitude of between 10 and 100% of the design load. Following this specimens were
unloaded and final toading ondertaken till failure.
59
Final stress valnes were then compared to those of a sample subjected only to steadily
increasing load, where no significant difference between the two samples was found. Final
deformation proved not to be affected either. A typical stress deformation plot from this
experiment is shown in figure 4.3.
CYCliNG LOAO 10-10~ DESIGN LOAO
-0.1 0.9 1.9 2.9 3.9

DEFORMATION rrrn
Figure 4.3:Cyclic load result
The experiment was then repeated on another sample for a larger amplitude of the load,
maiutaining the same loading speed. The chosen range was between 10 and 80% of the
expected ultimate stress level. Figure 4.4 shows that deterioration occurs at this level of stress
and rigidity quickly deteriorates leading to the failure of the specimen.
An obvious question arises here. What is the level of stresses for which the behaviour changes
from that reported in figure 4.3 and that of figure 4.4 ?
Because of time limitations, it was not possible to answer this question, but this matter is
worthy of further research.
60
Chopter 4: Glueability of bamboo wood plitJses
Another important aspect considered was load reversion. Structures under cyclic loadings
frequently sustaio some degree of reversion of the toading as well. This is especially true of
light structures, because in those cases dead load is small compared to live and other types
of load.
CYCLIC LOAOING 10-90~ ULTIMATE CAPACITY
DEFORMATI ON mm
Figure 4.4: Cyclic load 10-80% of ultimate capacity.
This argument also applies to bamboo structures, and thus it is of interest to know what is
the behaviour of glued phases under the action of reversing loadings.
An experiment was undertak:en for such a purpose. Specimens were loaded with a si~nal of
amplitude -1 00,+ l 00% of the design load. An example of the results is shown in tigure 4.5.
Mter 20 cycles specimens were completely unloaded, and then reloaded till failure, and final
stresses were again compared to those of the steadily increasing load specimens, resulting in
no significant difference between the two sets of results. Final deformation was also the same,
indicating no deterioration of the stiffness. It can be seen that the. amplitude of the loops
61
remained very stabie too, at this level of stresses. Even though no attempt was made to study
the real hysteresis of the glued phases evidence as shown in ligure 4.4 suggests not much
energy dissipation is available and therefore stress design levels should be kept in the
elastic range. Structures in areas prone to seismie activity should be designed for the
maximum earthquake response without any allowance for plastic deformation, that is,
without considering any energy dissipation mechanism, unless other dissipating points are
included, of course. Because of the lightness of bamboo, peak loads are expected to be low
in any case.
REVERSING LOADS -100, +100% DESIGN LOAD

6 ------ -------------- ------------------------------------------ --------------------- -------------- ------
111
w
!lla:
In - 2 -----------------------------------------------------------------------------------------------------------------------
- ..q ---------------------------------------------------------------------------------------------------------------------
-6 ------ ---------------------------------------------------------- ---------------------- ------------- -------

-0.5 -0.2 0.1 0.4 0.7
OEFORMAT I ONS mm
Figure 4.5: Completely reversing load.
1t was considered important to observe the sensitivity of the above results to a change in the
speed of the application of the load, because the speed recommended by ISO is very slow as
compared to a real earthquake signal. In the case of the signal of figure 4.3, frequency is
0.008 hertz, while it is considered that the possible range of responses for bamboo structures
62
are between 0.5 to 20 hertz, depending on a number of facts (mass distribution, stiffness,
support and soil conditions, etc.). In order to observe the effect of this type of load on the
glued phase, the signal of figure 4.6 was chosen. The period of the different trains was
decided after examination of some earthquake registers (Arce,O.A., 1992). Different locations
under different soil conditions, including extremesof stiff and very soft soil were looked into,
so that the experimental input covered those extreme conditions. The basic idea was to imitate
the initial, the intermediale and the final frequency trains that may typically occur in a real
earthquake, for different combinations of amplitudes and frequencies. Because of limitations
on the equipment, digitalization of the signals was not possible, so no real earthquakeinput
was employed. The random character of peak loads that are typical of earthquake signals
could not be replicated either because of the same limitations
It was also found that the equipment was only sensitive to frequencies of or below 0.5 hertz,
after which there was no forther controL Therefore it was not possible to go beyond this level.
Further, it was not possible even at this frequency, to record deformations, because the data
acquisition speed was found to be very slow for this signal, so the experiment was only
checked with a continuous plot of force against time.
Again it was found that the final load of the specimens was not affected by the imposition
of the dynamic signal. What is more, three specimens were subjected to more than fifty cycles
of complete reversal between -100 and +100% of the design load at 0.5 hertz, with no
apparent damage. Specimens were reloaded again up to failure, and the average result showed
no difference with that of the reference sample.
Thoughthese results are only indicative, and more research is needed to be conclusive about
the dynamic properties of glued bamboo wood phases, results so far are encouraging and can
be taken as a confrnnation of the degree of safety of the recommended design level.
The final choice of glue on a given design of course depends on a number of aspects. So far,
evidence suggests that strength is not the most important issue, because, ~ explained before,
faiture occurs inside bamboo.
Three factors must be carefully examined though. First, durability, especially in relation to
weathering, is a problem that requires close examination. In relation to this it has to be
63
force as a ~ of design load

1
40
cycles
0..5 hertz 0..5 hertz 0..5 hertz ftequency
Figure 4.6:Design signa! for high frequency tests.
remembered that bamboo requires preservaûon, and, as a complete definiûon of this problem
is yet to be gained, preservants may react with glue, or may impede correct bonding. Second,
glues are normally expensive, and someûmes unavailable in eertaio regions of the third world.
So a cost effecûve choice is always a priority. Third, workability may be important,
depending on the specific design application. It can be, for example, that a filling glue type
is desirabie for some reason. Among other soluûons, a mixture of resin and portland cement'
may turn out te be attractive, because in this case the resin volume is decreased. As a result
there are lower costs with the advantages on the durability side. In addiûón, the need for
3
Work done by Dr. Jurek Jassienko, a staff member of the University of Technology of
Wroclaw, Poland, duringa sabbatical in the University of Technology of Eindhoven, showed
the potentials of this applicaûon in the case of wood restoration work.
64
pressure during the setting period is eliminated, and surface irregularities or are compensated
for.
REFERENCES:
1 Arce, O.A.; 1992. Conneetion of bamboo elements. Paper presented to the IV

International Bamboo Workshoo, Chiang Mai, Thailand, 27-30 November.
2 Jindal, U.C.; 1988. Tensile strengthof bamboo fibre-reinforced plastic composites with
different stacking sequences. Proceedings of the International Bamboo Workshop.
Nov. 14-18, Cochin, India, pp.231-234.
3 Zoolagud,S.S.; 1988.Recent developments in bamboo board manufacture and future

research needs. Proceedings of the International Bamboo Workshop, Nov. 14-18,
Cochin, India, pp.291-293.
65
Chapter 5: Design of bamboo connections

5.1.-Introduction
In the course of time a number of joints for bamboo construction have been developed, and
some of the solutions in use can be traeed back thousands of years. As in the case of wood
initial trials consisted of two or more elements tied together with rope or string. Initially,
organic fibres like bamboo itself were used, and more recently steel wires. W arner (?)
describes some of the lashing type joints normally found in the field, and some of them are
depicted in figure 5.1.
The lashing technique is popular among users of bamboo. The Chinese people are able to join
together elements of bamboo to build structures as high as forty storeys, as in scaffolding
(Anonymous, 1979). In Thailand this system is still common practice in the construction
industry. Plastic bands are employed there to connect·the culms, and a small piece of wood,
bamboo or a metallic pipe is utilized to twist and fasten the lashing.
In more formal construction though, lashing does not produce enough stiffness in the
conneetion and better options are needed.
An alternative has been to clamp the culms between pieces of wood or plywood connected
with bolts, as shown in figure 5.2. Rigidity can largely be improved making use of the in
plane moment of inertia of the pieces of plywood or the like. There is no need for much
precision because the presence of plates relaxes this requisite, allowing for a certain amount
of flexibility in the dimensions of the culms.
A different proposal is described by Sonti (1959) in which bamboo elements are lasbedtoa
steel plate. The author reports that a full-scale dome was constructed in India using this joint
with good results. The advantage of the design is that advantage is taken of the rigidity of the
plate, but that damage to the culms is avoided. No report was made on the rigidity of the joint
itself, but it is very likely that this characteristic largely depends on the quality of the lashing.
A detailed description of common lashing and lashing-like joints is given by Janssen in his
Ph.D. dissertation (Janssen,l981), foliowed by an analysis of the way the examined joints
react to the imposition of loads. The author carried out a set of tests on full-scale joints to
66
Chopter 5: Design ofbamboo conneedons
IIOilllOIITIIl; tilrED
tiE tiOlE!I 10 VEil TICliL kiEI.totn
IlORED 1\T 1\ll<llE
lOTIUCIIOS!l
--.....l~'!T" vtnnc"L
TOP VIE\'1 MEMiti!A
.!.~.-B
SI!Cllotl
1ENOII KEY JOINT Fon IIEI\VV
llUTY on FINE WonK
.-r
t.•
DOUBLE bUlT liEb JOINT lDIIGIIE !!!!In TIClilT 1\CIIOSS
llt!JOtll AlfD 11EO
·-~
~""'Jfh~-\
SIDE VIEW
Tllllt: on VIII!! IJIIlDIIlO
.. .__ .....
JOJII; fOII COIIIIECIIQII
or SI.IAll I.II::Unrms IIMDWOOD
tUl TO tiT I"IIHI lO:IO TOIIGIIt
,~, ""'
SEC1ION COIIIItCttQII Or !IAI.IUOO
SIIJE VIEW
to IIOUIIO 1'1119, Uc.
mne: on
VIIIE TIE
SIIIGLE
.bUTt JOINt
DOUBLE DUTT VIIIEon ~
fJEIITJOIUT ~
IIOniZOini\L SECTION
IIIITII!Illl' ton tUilfiiTUIII!'
IIOrtiZONTI\L SECTIOtl
AltO !III.IILAII USU
Figure S.l:Some joints as reported by Wamer, reference 13.
67
Chapter 5: Design of bamboa conneetlans
Figure 5.2:Plated joint.
simulate the under-loading behaviour of a king-post type truss.

The joint in figure 5.3 was reported by DANIDN (Anonymous, 1985).
It is claimed to be especially serviceable in the case of axial forces, for which a set of design
rules are described in the said document. The basic concept is to use extemal ties to take
tension forces transmitted by pins through bamboo. Splitting of the culms is avoided by
binding the zones of stress concentration with wires, though the authors report that during the
test of full-scale trusses this type of reinforcement d.id not prove to be etlective. Similar
conclusions in this respect are reported by Janssen (opus cit), who compared the bending
behaviour of bamboo culms with that of a sample that was reinforeed by a wire wound
around the elements.
The use of precast steel or aluminium fittings was suggested by Duff (1941). A detail is
1
The Danish agency for development.
68
Fundamentals the Design of Bamboo Structures
8
. L.
tlrUT
REtHrORC!M[HT
- - ---r
S~CTION 11·11
Figure 5.3:Connecting system proposed by DANIDA,

from reference 2.
shown in figure 5.4.

More recently, Spoer (1982) tested somespace structures in bamboo where the joint of figure
5.5 was employed. The author reports that the conneetion performed very well during the
tests.
A system has been developed by the Costa Rican Bamboo National Project (see reference 2,
chapter 3 for more information on this project) in which a steel bar runs inside the culm. The
bar is welded at the extremes to steel plates. The extreme internodes are filled with mortar
to prevent insects from entering and to keep the bar in position. The steel bar has the purpose
of carrying tensile forces, whereas the culm takes on compressive forces. The element-to-
etement conneetion is achieved by welding or by machining the ends of the steel bar, so as
to be able to screw it to some type of receptacle.
In the present chapter the description of a systematic approach to the design of a feasible
69
Chapter 5: Design of bamboa connections
-d--~
Figure 5.4:Steel fitting proposed by Duff,

from reference 3.
bamboo mortar material ingluing tube mm screw nodal sphere ---

~
=------=---
stopper rope reinforcement

/ conus
Figure 5.5: Proposal from Spoer, reference 12.
70
Chapter 5: Design of bamboo co~:mections
conneetion system for bamboo elements is undertaken. The experience summarized in thls
introduetion is employed as background information in the process.
5.2-Deslgn Methodology
A systematic design approach is now given according to the description in figure 5.6.
Basically, the design problem is frrst defined by means of eertaio objectives to be
accomplished by the resulting object, system or technique. In this case structural performance
is the chosen semantic level, since the object to be produced is fundamentally a structural
object Next, constraints of the design are clearly stated. Basically, these frrst two elements
should completely define the design problem, to which the resulting design should be a
plausible solution. An instrument .for the evaluation of the resulting design bas to be
developed, so that decisions can be made over a number of options. The actual application
of the approach is described below.
5.3-0bjective function
" To achieve structural continuity between elements "
In modern construction technology, the concept of conneetion involves the search for
mechanisms such that deformations can be kept largely under control, and that predictions are
, also possible. Structural continuity also conveys the idea of force transmission according to
a certain prescribed and desirabie manner.
Perhaps the main role of the conneetion is indeed to safely transmit loads in a prescribed
manner from element to element, and finally to the ground.
5.4- Constraints.
5.4.1-Intemal Constraints.
5.4.1.1-Material properties.
In order to define material constraints all relevant features of the material are described and
a synthesis is made, based on the descriptions made in previous sections of the thesis, and on
other sourees where necessary.
71
IOBJECTIVE FUNCTION I
~§ ~
~
~
!:l
(I)
u
-
-(I)
(I) (I) r:alil .... 1-t-
t-w
z-
i~
11~
xt ~
ä -- (l)z
~~
..... ~<[
<[~
~w ~ §! w~
1-t-
.........
~~t
1-a_ u (I)
(I)Ci
~~ []~ <z
z~
Da_
u §i~ IL ~~ ~D
<u
:c
...J....J .... ~[] EJ
u <[
....J
<:!i
z~
~ ..... il!: wz
...J ~
~w
Wt- ~w
1-<[
~~ ~0 <t-
$x
I
Z:::t: uo (I) w
1-t
... .... w
t: ...1 R
... ~
..(J.-
..(J.-
DESIGN PRDFILE
UDGEMENT
CRITERIA
til HAD..S 111.TS
IR SCREYS
AVDIB IPEID ENDS
SIZE ADAPTAIDUTY Q
AXIAL TRANSFERENCE [F I..DADS
Figure 5.6:Design approach.
72
Bamboo is an anisotropic material which shows its best mechanica! properties in the direction
parallel to the grain, coïncident with the longitudinal axis within very small margins of
variation. Loaded in tension fibres face stresses with comparatively little shear disroption of
the weak lignin in between them.
Under compression, lignin does play a significant role and failure occurs because of the
presence of tangentlal expansive forces in the specimen leading to critica! tangentlal strains.
This mixture of high-strength fibres running almost perfectly parallel (but certainly without
crossing), and a weak lignin matrix in between, produces a material with little tensile and
shear capacity in the direction parallel to the grain. Bamboo is a unidirectionally reinforeed
composite with comparatively litOe tangential capadty.
These facts ·are well-known to people in the villages and they actually use weaknesses of
bamboo to their own advantage. They split bamboo longitudinally by wedging a knife in
between the fibres, with very little effort, to produce the raw material they need for baskets
and furniture.
5.4.1.2-Shape
Some of the most promising species are hollow and nearly round. This means there are a
number of problems for the design of connections. Besides that, bamboo canes are tapered
and thickness varlation is important.
Forces generate high tangentlal stresses due to the round and hollow shape. As mentioned
above, the absence of tangentlal or radial fibres creates a set of natura! failure planes along
the longitudinal direction and this stops bamboo from taking advantage of the dome effect,
that could have been expected as a response to forces perpendicular to the culm. Open ends
are then vecy easily crushed and this is no doubt the reason why common practice
reeommencts the location of transversal loads on the nodes, to take àdvantage of their
diaphragm effect. As node distribution along the cane is rather varia~le (Espiloy,l985), this
is not always possible and in any case it works out to be restrictive and rigid from a
constructional point of view.
Another inconvenience associated with the round shape of bamboo is the difficulty of aligning
73
Clwpter 5: Design of bamboo connections
elements to produce a neat joint. This has led to element overlapping, but this produces a very
massive joint, creates serious problems to the modulation and prefabrication of elements and
what is equally important, prevents an easy solution for the problem of interaction between
bamboo frames and other elements of the building. Besides that, overlapping leads to the
creation of ect:entricities, increasing bending moments and secondary effects in the
construction.
But the main drawback: of bamboo's shape need not be the stresses involved, but the
difticulties that arise on the geometrie side of the connection. Since sizes occur in a rather
random fashion, each conneetion presents a specific case and produces specific cutting
problems. So the conneetion of elements is a very Iabour intensive task:, that requires highly
qualified work:ers. The amount of sk:ill needed is such that the furniture industry workers are
real artists, and their wages follow suit In Costa Rica, for example, this phenomenon is such
that bamboo furniture is a luxurious product, in spite of the low cost of the material, far
beyond the buying capacity of the low iocome levels of general society (Manger,l991).
Thus bamboo bas a high variability in size, thickness and shape. HoJJowness creates littJe
croshing strength.
5.4.2-Extemal constraints.
Extemal constraints are set up as a list of desirabie characteristics to be fulfilled by the design
proposal.
5.4.2.1-Maximization of the use of bamboo.

Lik:e any other material bamboo has a combination of good and bad material properties. Joints
should take full advantage of the material and should support the use of bamboo to its full
structural potential.
5.4.2.2-Simplicity:
Bamboo construction is fundamentally aimed at the solution of infrastructure problems in
areas where sophisticated equipment and technica! capacity are not readily available or are
74
Fundamentals
expensive. Therefore, good design must be simple in terms of the amount of skill and
equipment involved in its production. These conditions are very important because solutions
to building problems in developing countries normally eaU for the participation of untrained
volunteers or users in self-construction projects.
Field tests would be necessary to fully understand and qualify the grade of simplicity of a
design. Factors like those explained by the Theory of Learning and Learning Curves can then
be quantified and accounted for, and, in the same way, the interaction between component and
total costs can be better understood.
5.4.2.3- Stability:
Joints should bestabie in relation to time. Design must take durability into account. in terms
compatible with the expected life of the entire structure.
5.4.2.4- Adaptability of dimensions to a modolar system:

So far the need for modulation on the basis of individual construction appears to be an
unnecessary sophistification. But it has to be realised that without this important tooi it is very
difficult to achieve the goal of coping with the increasing problem of housing in the third
world, which urgently needs the introduetion of modem management to speed up construction,
to improve quality and to lower costs. Besides, modolation has potentials to initiate systematic
approaches to production problems, which may help users to improve their own solutions,
may enable the participation of unskilled workmanship, may concentrate teehoical problems
in areas of easy control Iike specialized shops, and may also allow the formation of
production units. On top of this, modolation is a principle in its own right, with a wide range
of applications in production systems, worthy of being tried out and leamed with benefits
beyond the partienlar solution of a building problem.
5.4.2.5- Strength predictability:

From an engineering point of view what is perhaps the greatest problem found in the
lirerature on joints, is the absence of design rules to allow extrapolation, strength prediction,
75
and safety estimation. Not even rules of thumb seem to be available. Plausible mechanical
roodels andlor sound experimental data should be the natural companions of proposals for
joints.
5.4.2.6-Cost effectiveness:
Taking the interests of bamboo users into account, any submission or proposition must fulfil
requirements of cost-effectiveness.
In themselves joints are one of the highest components of structural costs in a building. But
it is common knowledge that joints affect the overall structural costs as wen, because they
change the amount of structural material needed, construction time, Iabour needs, architectonic
design possibilities, and so forth. So, the real impact of joints on the total cost is not a simple
matter to determine, and it certainly would be a mistake to simply compare joint to joint by
the cost of, for example, only the matenals involved.
5.5-Design Profile
The analysis of internal constraints indicates certain elements that should be present in a
'Design Profile'. This can be summarized as a set of specifications for the designer and in
this specific case works out to be :
-Avoid peneteation by nails, screws or bolts.
-Avoid open ends.
-Solve the probieros of size adaptability.
-Transfer forces by axially distributing them to the fibre system of the culm.
5.6.-Joint evaluation
In order to compare different design options or joints already available, it is necessary to
produce an instrument, so that the decision on the quality of the different possibilities can be
madeintbemost objective way.
At the end of this chapter an assessment is made of the design proposal, by summarizing its
most important features, and comparing them with the objectives and con,straints of the
76
Chapter 5: Design of bamboo conneenons
objective function.
5.7-Proposal fora Connecting System for Bamboo Elements

As indicated before, the use of pins, nails or screws entails some drawbacks and
disadvantages for the purpose of making joints for bamboo structures.
The above argument explains why it bas been of common practice to use natural or artificial
matenals to rope the elements together. No damage is done to the culm in this way, butsome
restrictions to the transfer of loadings become apparent.
Some of the natural fibres used for this purpose are not strong enougb to generate the
necessary tensile force, and besides that, there is very little friction with bamboo surfaces,
whicb are rarely sand papered or cleaned. Another souree of softening in tbis type of joint is
the use of green bamboo, which shrinks and slips out of the bond made by the rope-like
materials.
lmprovements can be made by using devices like the Delft wire-lacing tooi (Huybers, 1988-
1,2, 1989, 1990). Basically, this tooi winds steel wires around round pieces of wood, and bas
been suggested for the making of bamboa connections as wel1.
Some trials at the University of Technology of Eindhoven showed though that the amount of
tying is not sufficient to avoid slipping of the culms in the case of bamboo, and sometimes
the wires cause early cracking by emshing of the ends. The major problem would be the
calculation of the safety level of such connections, especially because uniformity in quality
is difficult to achieve.
The use of steel or aluminium fittings has given good results according to several authors
(Duff, Spoer, opus cit). Indeed, it is possible to provide the right shape to avoid all, or most,
of the problems mentioned problems above. However, those fittings are usually costly and
unfeasible in the types of applications bamboo is aimed at.
A feature or characteristic is common among proposals like the ones made by Duff and
Spoer. The problem of a low crushing capacity of open ends is eliminated, because the inner
chamber is ftlled, in one case by wood and in the other by mortar. Both solutions required
some extemal reinforcement, Duffs proposal in the way of a complex steel fitting, and
77
Spoer' s by using a tightened wire that is wound around the culm until it is completely
covered in spite of the author's claim of its dubious effectiveness. So, in both solutions an
attempt is made to take care of the problem of the low tangentlal capacity of bamboo.
Experience shows that taking forces out from inside the eulm solves a number of problems.
For example, the alignment of elements, distribution of emshing stresses, etc. So far, the
problem remains how to transmit forces to other elements without using expensive fittings,
on the one hand, and how to proteet bamboo in the tangentlal direction on the other.
Filling the eulm seems to be a flrst logica! step. To transmit forces from the eulm to the
filling, and at the same time to reinforce bamboo to sustain tangentlal strains, some sort of
adhesive material is needed. Better still, a piece of wood ean be used and glue can be
employed to stick it to the inner surface of bamboo. As shown in chapter four and appendix
C, any normal glue provides a capacity far larger than that of bamboo in the tangentlal
direction. The piece of wood can be regarded as a wood fltting. This component of the
conneetion can be extended outside the culm to meet the outcoming piece of wood from other
elements, and the design of an appropriate way of joining them together can be achieved, as
it is normally done in wood construction. In more demanding structural conditions, some
other type of solution ean be developed.
Figure 5.7 shows a proposal. The main principle bebind it is, as explained before, that the
system takes forces out from inside the culm, distributing contact stresses over as much area
as needed. In order to achieve this a cylindrical piece of wood is glued to the culm internally.
The frrst gain comes from the fact that the presence of the piece of wood changes the intemal
distribution of shear stresses, because of the enlargement of the net second moment of area
in the region of the connection. Obviously, the cross-section is also enlarged. Th is means that
bending stresses are diminished too, compared to those in the vicinity of the connection,
where bamboo acts alone. Normally the thickness of bamboo is small compared to the
diameter of the full combined section. This then produces a low demand on bonding stresses,
as shown in appendix D, where an analysis of stresses is presented for the derivation of rules
for calculation.
Two slots are needed to control cracking during the insection of the wood cylinder into the
78
Chapter 5: Design of bamboo conneenons
culm, as shown in the figure, because, as said before, bamboo is rarely perfectly round.
Besides this fact, the inner surface must be carefully cleaned prior to gluing. The preparation
of the inner surface can be done using a hand-held drilling machine with sand paper attached
to the drill. Up to 5 mm can be easily removed this way, n'l.aking it possible to use a flxed
diameter siie of the wood cylinder in a variety of bamboo sizes.
It bas been observed in the laboratory that up to 10 mm of difference between diameters can
be dealt with normally by this means.
The way in which forces are transmitted from one structural memher to another very much
depends on the type of structure and the predominanee of one force type upon another. Thus,
a discussion about the actual conneetion of elements is presented in chapter 6, once some
criteria for the structural possibilities of bamboo have been gathered.
Figure 5.7: Proposed system, inner part.
79
J
5.8-Evaluation
5.8.1-0bjectives
As shown in appendix D, it is possible to predict the level of defonnations in the connecting
system, and it is also possible to calculate material and geometrie requirements. On the other
hand, the study of glued phases in chapter 4 and appendix C shows that structural continuity
is also achieved with the system.
5.8.2-Design profile
All the specifications of the design profile are met by the proposed system. Though size
adaptability is not completely solved, some degree of achievement can be claimed in this
respect as well. However, no complete evaluation of this matter can be made until some
classification system is put in place, because otherwise the designer would always face an
unlimited number of sizes and shapes.
5.8.3-External constraints
The making of the wood cylinders may present some problems in the field, but they can be
made in small workshops to the specifications of the designer. This is economically important
also. The construction industry generates a significant contribution to the economies of
developed countries, mainly because of the amount of economie actlvities that it is able to
generate. In the same way, if bamboo construction proves to be technically feasible, it should
significantly contribute to the development of the economies of developing countries, by
creating business and job opportunities, as well as forward and backward economie linkages.
Once the wood fittings are shop tailored and made, the construction process in the field is
greatly simplified. There is no need for skilied Iabour. Bamboo elements need but Httle
preparation, sand-papering and cleaning of the interlor surface in the region of the connection.
Cutting requires no precision at all, since differences can be taken on by the conneetion itself.
Angle cutting on bamboo is completely eliminated, since this can be done at the extreme end
of the piece of wood.
The cost of wood can also be significantly reduced, since pieces of the required diameters are
80
Fundamentals
often the result of plantation thinning in tropical forests. The amount of glue involved is very
small and good quality can be achieved with normally available types. So, though this can
only be evaluated under real site conditions, it is believed that the cost of the proposal is low
and makes it feasible from an economie point of view.
The system makes the best use of bamboo by tak:ing advantage of its good properties, but
also helps to proteet and improve conditloos in the r!!gion of the connection, not only from
the point of view of stresses, but also from that of protection. No open ends means no place
for insects to hide, neither places of difficult maintenance.
The system allows for the design of modolar systems, concentrating the problem of
dimensions in the conneetion itself, so bamboo becomes sort of neutral in the system, which
is a real relief, since, because of the wide range of shapes and sizes, putting bamboo elements
together would otherwise require very skilied workmanship.
REFERENCES:
1 Anonymous, 1979. Chinese bamboo artisans preserve ancient craft Engineering News
Record, 202(79) Nr.7 Febr. pp.44-45.
2. Anonymous, 1985. Tiles Bamboo and Cyclones. Project Report. NRD-II Project
DANIDA/BSCIC, 1985.
3 Duff, C.H.,l941. Bamboo and its structural use. Institution of Civil Engineers
(Shanghai Assodation). Engineering Society of China.
4 Espiloy, Z.B.,l985. Physico-Mechanical Properties and Anatomical Relationships of

some Philippine Bamboos. Proc. of the International Bamboo Workshop Oct6-14
Hangzhou, Republic of China.
5 Huybers,P.; 1988. The use of roundwood poles for agricultural building structures.
Conference on Wooden Buildingsin Agriculture, Seminar section III, Växjo, Sweden,
Nov.14-17. Publisbed by the Swedish University of Agricultural Science, Lund,
Sweden.
6 .Nuevos métodos para la construcción de estructuras con Guaduas

y madera de sección cilfndrica. Primer Congreso de desastres y Arguitectura. Ibague,
81
Colombia, Nov .4-7, Sociedad Colombiana de Arquitectos.
7 ; 1989. Round Tîmber Poles as struts forspace structures. Proceedings

of the IASS Congress, Madrid, Sept.
8 ; 1990. Thin poles of roundwood for Structural Engineering

Applications in Building. Structural Engineering Review, 2, pp.169-l82.
9 Janssen, J.A., 1981. Bamboo in building stiuctures. Ph.D. Thesis. Technische

Universiteit Eindhoven. The Netherlands.
10 Manger, Y.,199l. Muebles Económicos de BambiJ en Costa Rica. M.Sc. Thesis.

Eindhoven University of Technology, Faculty of Philosophy and Social Sciences.
11 Sonti V.R., 1959. A Noteon Dome Structures Constructed from Short Bamboo and
Timber Elements. National Buildings Organization of India. Proceedings Symposium
on Timber and allied Product§. New Delhi 18-22 May 1959, pp.319-321.
12 Spoer, P., 1982. The Use of Bamboo in Space Trusses. M.Sc. Thesis, Technica!
University of Denmark.
13 Warner, J.R., ?. Handhook of Construction. The Di vision oflnternational Affairs, U.S.

Department of Housing and Urban Development. Washington O.C. 20410.
82
Chapter 6: Design of bamboa structures
Chapter 6: Design of bantboo structures

6.1-Introdnction
In the previous chapters bamboo was dealt with from the point of view of the material
(chapters 2, 3 and 4 and appendixes A, B and C), and from the point of view of the properties
of the culms (chapter 5 and appendixes D, E, F, anf G).
In this chapter some ideas related to the design of bamboo structures are based upon the basic
concepts described there. The approach in this chapter is basically descriptive, remarks are
made based on the physical relation of variables, that has been introduced in previous sections
of the thesis, on the observations of different phenomena in the laboratory, and on some
limited field observations. It is acknowledged, though, that probably the most important
souree of input to the development of structural design, namely, engineering experience, is
very limited as a souree in this case, due to circumstances which will be explained in the
following. As an eventual 'fully engineering' material, bamboo is still at the very beginning
of its development. Moreover, the crude application of 'common' engineering knowledge may
be misleading, because of substantial differences between bamboo and other more normally
used (and more normalized) materials.
Though basic mechanica! properties have been dealt with by many authors and were also dealt
with during this research project, field work should still be ondertaken from the point of view
of the practical applications of the findings. In wood, for example, and to talk about a
material that may resembie bamboo, properties are related to species, broadly speaking, but
in the case of bamboo it looks as if things are not so straightforward, because of the very
random nature ,of the material. Basic valuable design data for tropical wood is still in the
process of being developed (see Nakai, 1992, only to mention one example of this effort), so
it would be too much of a simplification to believe that the bamboo data base is complete.
The way bamboo should be classified still has to be developed, and only then will it be
possible to determine, in a cyclic process, design levels for the different parameters, and the
relation of these design levels to species, location, and so forth.
So far, though, one interesting possibility pointed out in this thesis (chapter 2) is that it may
be possible to control the maximum tangential strain in the culms as a way of fJXing desirabie
83
Chapter 6: Design of bamboo structures
levels of safety. For that purpose all one needs to do is to check the maximum tangentlal
capacity for a certain batch or species, and to detennine the elasticity modulus in an
appropriate and reliable way. This can not be done without serious statistkal studies and
carefut laboratory investigations, but in this way the work of the designer can be greatly
simplificated and grow in meaning.
In the following lines, therefore, consequences of the properties of the material and of the
culms are connected to design possibilities from several points of view, covering matters like
loads, shapes, connections and others that may have some relevanee in the detennination of
the design procedure and the final product: the bamboo structure.
There might be some discussion upon the fact that perhaps practice should lead the way in
the development of bamboo use in construction, as some rnaintaio is the case in engineering.
That criticism may be plainly justified, especially because in general tenns a theory in
engineering is not ~really) engineering till it is put to work, that is, until there is some
concrete engineering evidence of it, otherwise it just remains a theory.
In the case of the general methodology of this thesis, theory precedes practice to some extent
(some people will not be surpriscd about this in any case, they probably even applaud the
approach, especially manyin favour of theory first and practice later). But practice in ban1boo
construction 1 is practically nonexistent, so there was no choice but to try to simulate a broad
reality by modelling it, accepting that models may be incomplete, but that at least they may
give some idea of the extreme situations that may arise from extreme valnes of the parameters
involved. The design consequences that are referred to in the next sections contain this
methodological bias, so the reader should keep this in mind.
6.2-Load systems
Loads can be classified in many ways, and it is not the intention of the author to propose
criteria in that sense, or to extensively discuss ideas currently available on this matter. It is
important to acknowledge, though, that there must be a conneetion between the properties of
'Meaning 'engineered' construction.
84
the structural elements that are to be nsed in a construction, and the type of loads that those
partienlar elements may carry in an efficient way.
Perhaps this point is not so obvions in the case of shapeable materials (those that are
susceptible to forming or shaping, the final shape of which is not an intrinsic property),
becanse those materials are precisely shaped to meet certain structural and architectonic
requirements, to mention just two demanding conditions.
From the perspective of the initia! and basic scope of this research project, bamboo is
considered nnshapeable2, that is to say, the accepted objective is to nse the culms as they are
produced by nature, as tubular tapered elements. This shape and the mechanica! properties of
the material shonld in the first place determine the type of loads that a partienlar structure
may carry.
For simplicity's sakeforcescan be grouped according to the effect that they produce in the
element, such as axial forces, transversal forces, bending moments, and twisting moments.
Axial forces can produce compression or tension states in the element.
The final effect on the element depends on a number of conditions, lik:e the shape of the
structure, the position and direction of the forces, the conditions at the supports, initia!
imperfections, creep, time dependency of the loads, etc.
According to the arguments in chapters 2 and 3, bamboo is an elastic brittie material, with
a relatively low modulus of elasticity3• Though the culms have nodes that sometimes are
thought of as strong links, it has been shown that actually nodes are the weakest component
of the culms, and the points where attention should be mostly concentrated. As a general rule,
the more nodes a cnlm has, the less effective it is carrying loads, both from the point of view
of strength and of deformation.
Mostly, it is therefore desirabie that culms carry axialloads only, or at least that transversal
loads and bending moments can be kept to a minimum. Bending moments will be present
anyway, because of initia! imperfections, even in visually straight culms, and becanse of some
~ough, in a broader sense, this is not true of course.
~at is, the overall phenomena in the fnll culm, as explained in appendix F.
85
time dependency of the shape of the culms, namely creep, that has been observed as an
important factor. Though lanssen (1981) has shown that creep in bamboo is less than in
wood, evidence suggests that the initia! conditions of storage in the process of drying, play
an important role in this matter, because culms supported at two extremes during this period,
usually show important sagging effects. Moreover, bending and shear is introduced by the
connections, which can never behave as perfect hinges.
As said before, nodes also generate bending moments, normally accompanied by twisting. In
the experiment of appendix E, though, it was found that the angle of twisting was not
considerable (but culms were selectedon the criteria of straightness). In terms of twisting, in
any case, the major obstacle remains the fact that research in this area is practically non-
existing. This is of course limiting for the design and calculation of space trusses, especially
those where, because of asymmetrie loading or shapes, torsion may be important.
Though this research project did not cover the study of shear in an explicit manner, we
explained in chapters 2 and 3 how tangentlal strains could be the single most critical
parameter in the definition of the type of faiture in bamboo culms. Transversal shears can be
associated to tangentlal strains (perpendicular to the fibres) , of course, and therefore they
should be avoided. It was pointed out by lanssen (opus cit), that shear calculated according
to bending theory related poorly to the measured capacity of a sample of culms tested in
bending, the roeasored shear capacity being about five times as large. Oddly enough, the
figure found in bending calculations is much closer to the maximum capacity that has been
found in this research project for tension perpendicular to the grain (chapter 2), confirming
how critica! this can be. From the point of view of its composition, bamboo is not natorally
reinforeed for shear, because, compared toa reinforeed concrete beam, 'stirrups' are located
on the longitudinal instead of the transversar direction in a bamboó culm.
So it is important to consider shapes that produce axial loads in the elements. In addition,
account has to be given to bending moments due to initial eccentricities and those generated
at the connections.
The discussion on whether to prefer compressive or tensile forces is a bit more difficult. The
design value for tension and compression works out to be quite similar, but compression
86
capacity is rather irrelevant from a structural point of view, because the culms are usually
slender. Therefore, elastic instability may play a very important role, bringing down the
maximum admissible stresses, depending on a number of parameters as indicated in appendix
E.
A comparison of the values given by eq.E.(74) and the maximum compressive capacity,
typically found for bamboo, indicates that the critica! value of average slenderness at which
the compression changes to buclding is 1 .... = SO • It means that for the more common
species the maximum length for a double hinged strot should be kept under 250 mm, because
under this condition compression would govern the maximum capacity of the element
Soit is clear that elements under compression loads are bound to be much shorter (or less
slender, in any case) than those under tension loads, and that the overall shape of the structure
should reflect this fact.
One possibility that can be examined in relation to the need for augmentation of the second
moment of area concerning elements under compression, would be the use of arrays of two
or more culms.
The left-hand side of figure 6.1 shows an array of four culms of 80 mm in external diameter.
This is perhapsnot the most feasible shape for constructional purposes, but it is proposed bere
only for the sake of the cuerent argumentation, though it will be shown in the next section
that some constructional possibilities do exist.
When the second moment of area of the left side array is compared to that of a single culm
of approximately the same size, as shown on the right hand side of the figure, we find that
they are very simHar to each other. On the other hand, it is always much simpter to build
using single elements, especially if we consider the complexities arising in the connections.
So, bamboo structures should tend to be formed by using single elements instead of groups
of smaller ones.
It can be concluded, therefore, that bamboo structures deliver their best structural
performance onder axialloads.
87
Figure 6.1: Four culm and one single culm arrays
6.2-Structural sh~pes and eonnections

In terms of the very limited recent experience, different approaches to the use of bamboo have
been on trial, mostly at the laboratory level, but also in actual construction, particularly in
housing projects. There is on the one hand a sort of a traditional approach, where bamboo is
used for walling and partitions, and where some level of transformation is given to the culms.
Woven strips, culm frames ftlled with these woven strips, and sometimes plastered faces, are
examples of these types of applications. Part of this approach is the use of the culms to
produce trusses and frame trusses as structural shapes, where the elements are conneeled to
each other using natural fibres, crossing the elements over each other. This is the solution one
typically comes across in traditional construction in South-East Asia.
More recently the literature reports the studies of Hidalgo (1981), Spoer (1982), Ghavamy
(1988), tomention just the best doeurnenled cases, where connections follow the sbapes used
in the 'West' in steeland wood construction. Structural shapes also resembie the ones used
88
in wood and steel.

Perhaps the most complete coneetion of design ideas can be found in " Bamboos"
(Dunkelberg, 1985), a book publisbed by the Institute for Lightweight Structures, University
of Stuttgart. under the direction of Prei Otto.
The observations of these authors, and the examples of appendix F, seem to suggest that one
of the most promising applications may be the building of trusses, frame trusses and spatlal
structures. Bamboo frames are definitely not a good option, teading to very flexible and weak
structures, with very low degrees of redundancy, as is discussed in the last section of
appendix F.
In accordance with the results of chapter three and the discussions of appendix E, structural
shapes should be such that the length of compression elements is kept short to avoid buckling,
and the connections should be as stiff as possible for the same reason.
In this sense the choice of shapes has been limited in the past by the type of connections
available. Spöer (opus cit) tried the use of steel fittings bolted to a spherical receptacle, that
allows the building of lamellae and grid-like structures. Ghavamy (opus cit) reports works in
the same direction, with positive results.
The use of steel or roetal to transfer loads from element to element seems to be a promising
approach. As discussed in appendix F, wood-to-wood connections add a great deal of
flexibility to bamboo structures, because the second moment of area that is available, due to
the size of the incoming elements, is very low, though the effective elasticity modulus of
wood is normally much higher than that of bamboo culms.
So, before going ahead with the discussion on feasible structural shapes, it is necessary to
develop some ideas about the possible joining systems, departing from the fact that the basic
problem of taking forces from the culm can be solved according to the proposal made in
chapter 5.
A simple way of dealing with the problem is depicted in figure 6.2.
The steel plate C is introduced in the slot of the wood cylinder and glued to it with a mixture
of epoxy resin and Portland cement. The plate is projected, so that its outer extreme can be
adapted for different applications, as indicated in the details D and E in the figure.
89
Figure 6.2: Connecting system.
In plane trusses, the plates from two or more incoming elements can be pre-welded to each
other and then the rest of the conneetion can be assembled.
The specific way by which plates or steel tips are joined together is left to the creativety of
the designer, but an idea is depicted in figure 6.3 just to show the kinds of possibilities that
can be made available.
Figure 6.3 shows a conneetion in which a small box is made of steel plates, so that the faces
are perpendicular to the axis of the incoming elements. The steel tips are then welded directly
to those surfaces. Welding is thought of bere because it is cheaper than machining of the tips,
but in some cases this can be achieved as well.
Figure 6.4 shows a different solution, where a piece of steel pipe is used as a central element
so that incoming plates, in this case, can be welded to the required angles. lt can be seen that
this type of generic solution gives great flex.ibility to the designer because there are few
restrictions from the perspeelive of the structural shapes that it may support. The amount of
90
Figure 6.3: Conneetion for space truss, with centre steel box element.
steel that is needed is very small and the fabrication of the components is simple compared
to precast steel fittings. A rough calculation of the thickness of the plate shows that a
thickness of only 2.0 mm is needed to match strength and stiffness needs in the connection,
for the expected capacity of bamboo culms between 70 and 120 mm diameters. Experimental
evidence on the use of the mixture of resin and Portland cement at the laboratory of the
Department of Structural Design of the University of Technology of Eindhoven4 very much
supports this idea both from the point of view of strength and stiffness. Though the cost bas
to be evaluated in a field situation, the amount of resin involved suggests that this would not
be a major problem.
4
From the workof Dr. Jerzy Jasienko, invited researcher, staff member of the University
of Technology of Wroclaw, Poland, who has studied this possibility for the reinforcement of
wood beams with embedded steel bars.
91
Figure 6.4:Connection made with a piece of steel pipe and plates.
Several problems are dealt with by means of this proposal.

In the ftrSt place, all the advantages described in chapter five are kept intact, both those
related to the properties of the material and those related to dimensioning, detailing, and
building.
In the second place, the proposal is more of a concept than a definitive rigid fitting or system,
because it allows the designer to have choices in relation to the final specific shape of the
connection. In that way it allows the interconnection of bamboo culms, but also their
conneetion to elements of other materials, like a concrete beam or a foundation system, as
shown in figure 6.5.
In this way .the structural shape is not restrained or defined by the connecting system. So,
what should the shape of bamboo structures be then?
lt has been shown how the use of the glued wood cylinder plug proposed in chapter five
opens possibilities for the designer, as have been shown in the previous paragraphs. This fact
92
o. ~ o o. ~ o o. ~ o
-e?O -e?e? -e?O
~ ~
())- 0 ~ ())_ ~ ())_ ~ 0
-e:::? 0 -te::>
Figure 6.5: Conneetion to a foundation or concrete element.
combined with the recommendations of section 6.2 leaves the designer free to cespond to the
specific ciccumstances of the problem to be solved.
In spite of that, some other recommendations can be derived from the statements in previous
chapters.
Though it was argued that the use of a single element, instead of a combination of several
ones, is more efficient from the point of view of the stiffness, there may be other reasons why
it is desirabie to proceed in that manner.
In figures 6.6 and 6.7 we can see that the proposed system can be adapted to cope with this
type of demand. There, a combination is shown of three and four culms, which together form
a chord, or a diagonal element. The culms probably need to be kept together at midspan, and
for this purpose the figures show the use of thin steel bands.
The conneetion is achieved by projecting steel tips out of the wood cylinder, so that these tips
can be welded to a plated or any other central component. As shown in the figures, stiffening
93
Figure 6.6:Structural element composed of three culms.
plates can be welded to the tips, to ftx the relative position of the culms.
6.3-Bamboo's major assets

As explained in chapter one, bamboo is a renewable resource. So far, its 'silviculture' has
been rarely deftned in terrus of construction purposes, except in the case of the Costa Rican
Bamboo National Project (a bamboo housing project supported by the Govemment of Costa
Rica and the Government of The Netherlands, see for example, Chaves and Gutiérrez, 1988).
Bamboo supplies can easily be sustained in a process that is envîronmentally friendly, and
this one of bamboo's major assets.
The second major advantage of bamboo is its relative lightness, a property that comes from
the hollowness of the culms. Designers can take great advantage of this characteristic. It
seems that the major field of applications is in structures for roofs and sheds, where the
demands of stiffness of conneetloos are not that high. This is an important factor to take into
94
Figure 6.7: Structural element composed of four culms
account. because if the weight of the connections becomes an important factor in the total
weight of the structure (which may happen, due to the lightness of bamboo), then little
capacity may be left to carry the rest of the loads in the structure. An appropriate design of
the shape may allow it to carry quite important loads with only a small amount of bamboo
mass in the structure. Therefore, designer should pay most attention to the relation between
bamboo and total weight so that an optimum can be reached. This fact may in some cases
define the viability of the use of bamboo for a predetermined purpose.
KEFERENCES :
1 Dunkelberg, K; 1985. Bamboo as a Building Material. IL 31 Bamboo. lnstitute of
Lightweight Structures, University of Stuttgart, Germany.
2 Ghavamy, K.; 1988. Application of Bamboo as a low cost construction material.
95
Proceedings of the International Bamboo Workshop, Cochin, India, 14-18 November.

pp.270-279.
3 Hidalgo, 0.; Manual de Cob.strucción con BambU. Ediciones Técnicas Colombianas,

Bogotá.
4 Nakai,T.; 1992. Mechanica! Properties of Tropical Woods. Proceedings 2nd

Cffi/W18B International Conference on Tropical and Hardwood Tirober Stmctures,
Kuala Lumpur, Malaysia, September 7th-10th, pp. D-23,0-29.
5 Janssen, J.J.A.; 1981. Bamboo in building structures. Ph.D.Thesis, University of

6 Spöer, P.; 1982. The use of bamboo in space trusses. M.Sc. Thesis, Technica!
University of Denmark, Stmctural Research Laboratory.
96
Chopter 7: Conclusions and further research recommendations
Chapter 7: Conclusions and forther research reconnnendations

7.1-The structure of bamboo culms
7.1.1-Most bamboo species have hollow culms, though some solid stemscan be found. The
hollow internodes are conneered toeach other by nodes, or transversal walls (see section 2.1.2
and related references).
7.1.2-Bamboo culms are composed by a matrix that holds fibre strands together. These fibres
run parallel to the axis of the culm, being more concentrared at the top of the culm and
towards the outer layer of the cross-section. The fibres are confined to the fibre strands and
to sclerenchyma sheaths (see section 2.1.2).
7 .1.3-Fibre strands are not continuons along the length of the culm, instead they bend towards
the inside of the nodes (see section 2.3).
7.1.4-No radial fibre strandsexist in the case of bamboo (see section 2.1.2).
7.1.5-Culms are tapered, the wall thickness decreases from bottom to top, and the internode
length varles along the culm.
7.2-Tangentlal tension capacity of bamboo
7 .2.1-The elasticity modulus in the tangentlal direction is about one-eighth the elasticity
modulus in the longitudinal direction (see chapter 2).
7.2.2-No clear correlation could be found between tangentlal tension capacity and density of
the specimen (see section 2.2.2).
97
Chapter 7: Conclusions and further research recommendations
7.2.3-A maximum tangentlal strain was observed for different samples of Guadua s.p.,
Bambusa blumeana, and Gigantochloa scortechinii, indicatlng the possible existence of a
characteristlc value for bamboo. More research is needed on more species and samples to
verify this aspect This is a matter of importance not only for constructional and structural
purposes but a1so for handieraft (see section 2.2.2).
7.2.4-Fibres contribute indirectly to the tangentlal strength, because they change the 'texture'
of the crack path. The more fibres the crack trajectory encounters, the more energy is needed
to fracture the material. Research on this matter should be ondertaken to determine the
relation between fibre content, density of the non-sclerenchymatose components, and
tangentlal tension capacity (see sectlons 2.2.2 and 5.4.1.1).
7.2.5-A specimen for the determination.of maximum strain at faiture in tangentlal tensionwas
developed. Research is needed to develop a specimen which is also suitable for a simple
determination of strength (see section 2.2.1 ).
7.2.6-Bamboo exhibits an elastlc behaviour in tangential tension, with a brittie mode of failure
(see section 2.2.2).
7.3-Parallel tension
7.3.1-A strong correlation between parallel tension strength and density was found (see
appendix A, eq.A.(l)).
7.3.2-A specimen for the determination of tensile capacity of bamboostrips was developed
(see section 2.3).
7.3.3-Nodes are the weakest components. The capacity of the nodes fora sample of Bambusa
blumeana was found to be only 40% of that of the internode section (see section A.3).
98
7.3.4-Nodes are the more flexible part ofthe culm. The average elastic modulus in the region
of the node (measured over a length of 30 mm, centred at the node position) was found to
be 40% of that of the internode section for a sample of Bambusa blumeana.
7.3.5-No study was undertaken on the structural behaviour of the full culm under tension.
Research should be done on this matter, taking the role of the nodes into account.
7 .3.6-The capacity of strips from the internode section is very high, found to be 270 Nlrrtfl'l
for a sample of Bambusa blumeana. This fact should stimulate research into the use of the
internode sections of bamboo culms in such a way that full advantage is taken of this
capacity. In the use of full culms this capacity can not be used because the nodes are the
limiting factors (see section A.3).
7.3.7-Further research into the utilization of the fibres of bamboo should include the way
different factors affect fibre content and density.
7.3.8-Bamboo presents a perfectly elastic behaviour in parallel tension, with a clearly brittie
mode of failure.
7.4-Compressive capacity of bamboo
7.4.1-The conditions of the test set-up in compression are a critica! point tobetaken care of.
Friction and stiffness of the loading plates contribute to the modification of the mode of
failure of the specimens (see section 3.1, 3.2 and B.l).
7.4.2-A distinction should bemadebetween the compressive capacity of bamboo as a material

and the compressive capacity of the culms. In the latter case, capacity is affected by the
geometry of the specimen. Measured parameters may show a strong influence of
imperfections in the specimen and the conditions of the test (see sections 3.2 and B.l).
99
7.4.3-When the full culms are intended to be used for structural purposes compressive
capacity should be measured on rings with a ratio of 1:1 for the height and the diameter, and
the test should be run in such a way that friction is limited to the minimum possible (see
sections 3.2 and B.l).
7.4.4-The relation between longitudinal and tangentlal strains is about 0.3. The maximum
tangentlal strain on the external surface of bamboo prior to faiture is about 0.0012 (see
sections 2.2.2 and B.l ).
7.4.5-It is possible to predict the stress level at faiture when the elastic modulus is known,
by using 0.0012 as a limitlog figure for the maximum tangentlal strain (see section B.l).
7.4.6-The secant value between 10 and 80 % of the maximum deformation is a good

estimator of the elastic modulus, being more stabie and less affected by interactions between
the specimen and the plates of the testing machine (see section B.2).
7.5-Criticalload of bamboo columns
7.5.1-The major souree of lateral deformation in a bamboo column is the initial lateral
deformation (see sections E.2 and E.5). This bas to be taken into consideration in laboratory
experiments on buckling.
7.5.2-It is possible to measure the criticalload of bamboo culms by using a Southwell Plot
procedure in the laboratory (see sections E.2 and E.5).
7.5.3- The criticalload of bamboo culms is affected by a combination of the change in the
cross-section an4 the change of the elasticity modulus along the length (see eq.E.36 and
E.37).
100
Chapter 7: Conclusions and further research recommendt;ltions
7.5.4- lgnoring the influence of the nodes, the calculation of the criticalload as a function of
average properties of the culm gives a conservative estimate (see eq. E.36 and E.37).
7.5.5-Since the nodes are more flexible than the internoctal sections, they decrease the value
of the critica! load of bamboo columns (see chapter 2).
7.5.6-Nodes are a very random feature, so their average influence should be studied for
specific conditions. The results of the Southwell plot procedure can be statistically adjusted
to take the influence of the nodes into consideration, and to simplify the calculation. The
study of the influence of the nodes on the stiffness of culms indicates that the critical Ioad
is overestimated when the nodes are not taken into account (see section E.7 and appendix F).
7.6-Giueabllity of bamboo
7.6.1-The bonding capacity of bamboo glued phases is not affected by bamboo density, nor
by the type of wood, the thickness or the diameter of the culm. A normal wood type of glue
is enough to produce a high-quality conneetion (see section 4.2.1).
7.6.2-Bamboo glued phases can resist cyclic loadings below the design levelwithno damage
or deterioration of the final capacity (see section 4.2.2).
7.6.3-Failure in bamboo glued phases always occurs inside bamboo, therefore bamboo is the
most critical component of the phase (see section 4.2.2).
7.6.4-Failure of bamboo glued phases is brittle (see section 4.2.2).
7.7-Bamboo gloed conneedons
101
Chapter 7: Conclusions and further research recommendatións
7.7.1-Glued wood fittings are a feasible and effective way of taking or transmitting forces to
bamboo beam columns, taking into account the weaknesses of the culms and making use of
their full capacity {see section 5.1, 5.4, 5.5 and 5.6).
7.8-Connections for bamboo structures
7.8.1-The use of nails, screws or bolts to conneet bamboo culms entails a number of
disadvantages, because this type of solution concentrates stresses in regions and directions
where bamboo is very weak (see section 5.4).
7 .8.2-The insertion and gluing of a wooden cy1indrical connector in the extremes of bamboo
culms not only allows us to take and transmit forces, but also to reinforce the culm in those
regions (see section C.4).
7.8.3-It is necessary to use steel fittings as central elements in the connections of bamboo
structures. Simpte nailing of the wood fittings leads to low capacity and high rotation of the
connections (see section F.6).
7.8.4-Simple cheap steel fittingscan be designed and developed to conneet the wood fittings
glued to the inner side of the culms. Some examples can be seen in section 6.2. Research
should be concentraled in developing a complete construction system, that is based on the
utilization of the possibilities of the glued wood fitting, allowing for a speedy and efficient
utilization of bamboo in structures.
7.9-The Design of bamboo structures
7.9.1-When the problem of the connections has well been taken care of, it will be possible
to produce bamboo structures that are light, stiff and strong, as shown in section F.6.
102
7.9.2-In general, the varlation of the diameter and that of the modulus of elasticity do not
compensate each other in termsof the final stiffness of the culm (see section 0.2).
7.9.3-The intluence of the variation of the cross-section and of the elasticity modulus
produces a reduction of no more than 15% in the bending and the axial stiffness of bamboo,
as compared to the tigure resulting from using average properties of the culms.
7.9.4-Nodes largely affect the bending and the axial stiffness of bamboo culms. Therefore,
it is better from this point of view to use culms with long internodes (see section 0.4).
7.9.5-In general, it can be said that the combined effect of the varlation of cross section and
the presence of nodes results in a rednetion up to 50% of the bending stiffness calculated
using average properties1, and up to 80% of the axial stiffness (see section 0.4).
7.9.6-Both from the point of view of capacity and deformations, trusses and frame trusses are
a much better application of bamboo than frames (see section 6.1 and 6.2).
7 .9. 7-Structural bamboo componentsin compression should be kept under a slendemess of 50.
In this way buckling is controlled, but also bending moments arising from initia!
imperfectloos are kept toa minimum (see section E.7).
7.9.8-The cmde application of 'commom' engineering knowledge to the structural design of

bamboo may be misleading, because of substantial differences between bamboo and other
more commonly used materials.
7.9.10-As a general mie, the morenodes a culm bas, the less effective it is for carrying loads,
both from the point of view of strength and of deformation.
1
the average of the properties at the extremes of the culms
103
Chopter 7: Conclusions and further research recommendations
7.9.11-The structural behaviour of bamboo culms under twisting loads is an area of research
that is yet to be unexplored.
7.10-General remark regarding bamboo as a structural material
Structural materials have strengtbs and wealrnesses, and it is up to the structural engineer and
architect to use them in the best possible way for the benefit not just of a specific elient or
idea, but for that of mankind as a whole.
The facts found during this research show that bamboo can be given some restricted structural
use. It is the hope of the author that some elements of relevanee are put forward in this thesis,
so that new structural shapes and applications can be devised. The gatbering of controlled
field experience should be the major focus of attention in the future, and would be the best
way to promote the use of bamboo.
The results of the testing of trusses in the laboratory are encouraging, especially because the
comparison of the weight of the structure with its capacity and stiffness gives a clear vote of
conftdence as regards the potential of the material.
During this research effort, bamboo proved to be a difficult material, sometimes it was even
evasive and impenetrable. Laboratory work on the preparation of specimens of all types has
shown how unworkable it can be, perhaps due to the fact that no tools, at the construction
level, are yet available for bamboo.
During the very last phase of the research project, though, the making of a few trusses
demonstrared that it is possible to build attractive, light, strong and well-engineered structures
from this materiaL As a ftnal statement the author wishes to share bis enthusiam in this sense,
and to encourage engineers and architects, especially those colleagues who have bamboo
available, to give it a chance by using all their inventiveness. Bamboo will pay back.
7.11-General remark on future research
The attention of bamboo research should move from the speciftcs of the properties of bamboo
104
as a material, to the applications in the construction industry.

It must be bome in mind that the major reason some institutions and individuals are paying
attention to this matenals is its abundance and easy reproduction, and the possibilities that it
may hold for people in the third world. But the results of research actlvities will never reach
these populations as long as they are kept in the sometimes ambiguous terrain of testing
compression and bending capacity in the laboratory only. Communication between field
engineers and researchers is needed in order to focus our attention on the real issues
concerning the utilization of bamboo in construction.
The current efforts in looking for a solution for the problem of preservalion and treatment
should be strongly encouraged and supported.
105
Appendix A: Experimental analysis of parallel tension

A.l-lnfluence of the speed of toading on the tensile capadty of bamboo
Table A.1 shows the descriptive stalistics for the results of three treatments (speeds of
deformation).
Variable: 0.2 mm/min 0.6 mm/min 1.0 mm/min

Sample size 11 11 11
Average 272.135 294.144 309.983
Median 300.82 291.96 328.0
Mode 282.48 261.0 284.0
Geometrie mean 264.718 288.377 307.832
Varianee 3733.93 3910.7 1468.22
Standard deviation 61.1059 62.5356 38.3173
Standard error 18.4241 18.8552 11.5531
Minimum 143.68 195.81 263.44
Maximum 345.76 419.88 374.58
Range 202.08 224.07 111.14
Lower quartile 224.0 259.68 272.25
Upper quartile 314.64 317.0 339.6
Interquartile range 90.64 57.32 67.35
Table A.1: Results from three loading speeds.
These results show that strengtbs tend to be higher for higher rates of loading, although it is
clear that more studies are needed to completely onderstand this phenomenon.
The goal of the study regarding the influence of the Ioading rate was to be able to gain some
flexibility in the eventual specifications of the test. This is because most of the laboratories,
where tensile tests may eventually be carried out, may not have conditions good enough to
106
assure a permanent and non variabie rate of loading. So, the question is to what extent some
varlation in the rate ofloading may lead to varlation in the tensile strength measured this way.
Two-Sample Analysis Results

Sample Statistics: 0.2 mm/min 1.0 mmlmin Pooled
Number of Obs. ll ll 22
Average 272.135 309.983 291.059
Varianee 3733.93 1468.22 2601.08
Std. Deviation 61.1059 38.3173 51.0007
Median 300.82 328. 301.24
Difference between Means =-37.8482
Conf. Interval For Diff. in Means: 95 Percent
(Equal Vars.) Sample 1 - Sample 2 -83.2222 7.5258 20 D.F.
Ratio of Variances = 2.54317
Conf. Interval for Ratio of Variances: 0 Percent
Sample 1 + Sample 2
Hypothesis Test for HO: Diff = 0 Computed t statistic = -1.7404
vs Alt: NE Sig. Level =0.0971534
at Alpha =0.05 so do uot reject HO.
Table A.2: Comparison of two loading speeds.
Table A.2 reports the test results of the hypothesis that the difference in means between the
two extreme speeds tested is zero. It can be seen that for a confidence interval of 95% this
hypothesis is correct, in spite of the difference in loading rates. Therefore, at least in a
preliminary step, it may be concluded that the proposed test could be done in a eertaio range
of toading speeds without affectlog the results thus obtained in a significant way.
107
Fundamentals
A.2-Inftuence of density
It was found that when moisture and density were taken tagether in the analysis, moisture
accounted for only 20,29% of the variance. In our case, moisture ranged between 12.80 and
17.71 % as it can be seen in table A.3. It can be observed that specimens were not
specifically conditioned from the perspeelive of constant moisture, nor was a wide enough
range selected for this variable, since as we established before, our main interest was to see
the influence of density under some flexibility regarding the values of other parameters 1• No
attempt has been made in this experiment to find a correcting factor for moisture, but its
influence is analyzed anyway, of course, as explained in the following lines. Density and
moisture were measured according to ISO 3131-1975(E) and ISO 3130-1975(E) respectively,
except that the shape of the specimens was changed to smalt strips of 3 mm x 50 mm x
2
thickness • Table 2.4 gives us some appreciation of the high tensile strength of bamboo
specimens from the internode section of the culm.
Two probieros related to moisture and density become apparent after analyzing table A.3.
First, both density and moisture tend to distribute in the way of a long right tail, but this
factor is even more pronounced in the case of moisture, as revealed by the values of the
skewness coefficients. Second, the distribution of density tends to be flat with short tails, but
this feature is more pronounced in the case of moisture, as shown by the values of the
Kurtosis coefficients. What is more important, the standard values of these coefficients also
show that the normality assumption is very good for density but poor for moisture, facts that
have to be taken into consideration in the analysis.
In spite of the above statement the influence of moisture was also checked by calculating a
regression of tensile strength vs. density and moisture, as shown in table A.4.
Analysis of the significanee levels and the t-factors leads to the conclusion that the regression
1
The relation between density and tensile capacity was also studied for conditioned
specimens at 12% moisture content, but, again, it was of interest to see if some deviation
from this value wou1d significantly affect the resu1t, which was not found to be the case.
~us density refers to the average along the thickness.
108
is not a good one, since the possibility of failing to predict is very high. Perhaps this was to
be expected due to the combination of two variables with two very different distributions as
was shown before.
This result supports the possibility of looking for a simple relation between tensile strength
and density, insteadof complicating the equation by including moisture.
Table A.5 below shows the results of calculating a linear regression of tensile strength and
density. The improverneut in the significanee level can be seen. It can therefore be concluded
that density can indeed be a good estimator of the tensile strength of this batch of bamboo.
An analysis of variances confirnled the accuracy of the model and, therefore, it could be
concluded that
!, = 0.4 p eq.A.(l)
l09
Variabie: density moisture strength

Average 686.56 13.63 275.16
Median 666. 13.27 300.82
Mode 634. 13.24 300.82
Geometrie Mean 682.17 13.57 269.79
Varianee 6869.42 1.89 2978.84
Standard Deviation 82.88 1.37 54.58
Standard Error 24.98 .41 16.45
Minimum 585.4 12.8 177.0
Maximum 827.3 17.71 345.76
Range 241.9 4.91 168.76
Lower Quartile 622. 13. 224.
Upper Quartile 783.8 13.57 314.64
Interquartile Range 161.8 .57 90.64
Skewness .66 3.11 -.57
Standarized skewness .90 4.21 -.77
Kurtosis -1.01 10.02 -1.01
Standardized Kurtosis -.68 6.78 -.69
Units : density = kg/m3 moisture =% tensile strength =N/mm 2
Table A.3: Deseriptive statistlesfora sample of B.blumeana from

the Philippines.
110
Model fitting results for:TENSILE STRENGTH

Independent coefficient std. error t-value sig.level
variabie
DENSITY 0.200101 0.210589 0.9502 0.3668
Moisture 10.03717 10.62858 0.9444 0.3696
r =o.9623
Table A.4: Density and moisture related to strength.
Model fitting results for: TENSILE STRENGTH

variabie
DENSITY 0.397578 0.024755 16.0606 0.0000
r =o.9627
Table A.5: Regression model of density strength.
A.3- Influence of nodes.

To examine the effect of the nodes, samples were checked in such a way that internode and
node samples were taken from neighbouring positions along the circumference , and tests ran
under the same circumstances.
The results in table A.6 show that the tensile strength of the node region is only about 30%
that of the intemode.
Table A.7 shows that there is also a significant difference in the value of theelastic modulus.
Node elastic modulus is about 40% the internode's.
lll
Variable: TENSION INTERNODE TENSION NODE

Sample size ,u 11.
Average 270.586 79.8936
Median 287.09 80.71
Mode 273.13 80.7
Geometrie mean 263.66 78.6725
Varianee 3597.13 201.64
Standard deviation 59.9761 14.2019
Standard error 18.0835. 4.28204
Minimum 154.76 52.48
Maximum 334.27 101.89
Range 179.51 49.41
Lower quartile 224.37 69.4
Upper quartile 326.4 88.67
Interquartile range 102.03 19.27
Skewness -0.76207 -0.29442
Standardized ..
skewness -1.03185 -0.398646
Kurtosis -0.446007 0.172994
Table A.6: Strengtbs from node-intemode tests.
112
Variable: INTERNODE MODULE NODEMODULE

Sample size 11 11
Average 18857.8 7452.62
Median 19444.7 7730.
Mode 17600. 6752.86
Varianee 2.99612xl07 2.2288xl06
Standard error 1650.38 450.131
Minimum 8928.0 5511.0
Maximum 26794.9 9794.0
Range 17866.9 4283.0
Skewness -0.403314 0.216891
Standardized skewness -0.546089 0.293671
Kurtosis -0.380636 -1.28877
Standardized kurtosis -0.257692 -0.872502
Table A.7: Biastic Moduli from node-intemode tests.
113
Appendix B: Experimental analysis of short elements
Appendix B: ExperimentaJ analysis of short elements

B.l· Introduetion
Several problems need to be taken into consideration in dealing with compression tests of
bamboo elements.
Perhaps the ftrSt aspect that is usually noted by the experimenter is the fact that the cross-
section of bamboo elements is not a perfect circle, but rather an irregular shape. Since it is
customary to calculate stresses as simply being the ratio of the normal force to the area of an
assumed cylinder of constant thickness, it is relevant to fmd out what influence this
simplification bas on the values measured in this way. As proposed by Xiu (ref. 11, chapter
3) a factor can be derived, like that of equation 3.3, to correct the cross sectional area of the
specimen.
From a different stand-point when the cross-section resembles an elliptical ring, its area can
be related to that of a circular ring by the equation
A._ 1 +r4 -2r1

eq.B.(l)
A 2-2r,
A sample taken from the stock of Bambusa blumeana and Guadua s.p. in the laboratory at
Eindhoven University of Technology, showed that the range of values for the parameters r4
and r, ranged, respectively, from 1.0 to 1.3 and from 0.05 to 0.15. Evaluation of equation
B.(l) for the cesuiting extreme values shows that the amount of error in assuming a constant
circular section ranges between 0 and 16 %. Therefore, a directly correspondent range also
occurs in the estimation of stresses. The important point bere is that the deviation of the
cross-section from a circular one should be observed and its importance never underestimated,
and that the actual conditions of the sample in this respect should be added to the report, to
make comparisons possible.
A second problem, very much related to the above-mentioned irregularity of the cross section
114
is tllat the thickness of the specimen is not constant, and neither is the density on the surface
of the cross-section. This means that some concentration of stresses may occur during testing
and this produces very different faiture patterns, some of them merely due to higher contact
stresses at the points of larger stiffness.
A third problem is that the steel plates of the apparatus act as a lateral reinforcement, teading
to possible overestimations of the mechanic properties, as found by Meyer (ref. 7 chapter 3).
Tests done in machines, which are very stiff in the movement of the heads, are likely to be
influenced by the rigidity of the plates. In some cases the plates are supported by a ball-like
connection, but even then some other measures should be taken to eosure that the test is a fair
approximation to the envisaged conditions.
Figure B.l shows the results of a test using a machine in which the upper plate was hinged,
but no special measures were taken in relation to friction. Results are stresses vs. vertical
strains calculated using the vertical displacement of the plates, as has traditionally been in the
case of bamboo.
The result can be misteading in many ways. Firstly, the initial section of the curve registers
the 'consolidation' of the plates. In other words, all the movement that takes place till all the
pieces fit each other, including the specimen in relation to the plates. So, this is not a plot tllat
gives information about the material only, but about the whole mechanism present in the test
set-up.
Secondly, some recovery is observed after failure. Insome cases even higher loads than the
ftrst peak can be recorded, but this is due to the fact that, because of the high normal stresses,
friction k:eeps the specimen trapped between the plates, so that it can sustaio some extra force.
Besides the above-mentioned factors, and purely from a teehoical point of view, the making
of the specimen is a very laborious and difficult task, due to the tapering and to the fact that
the cross-section is oot circular. Most of the time this leads to specimens in which the
surfaces in contact with the plates are oot parallel to each other, thus inducing accidental
eccentricities during the test What is more, the problem is worsened by the fact that an
increase in the height of the specimen implies increments on the eccentricity of the load as
well, inducing radial bending moments towards the inside of the specimen resulting as in
115
figure B.2.
Control of the conditions during a compression test is therefore a factor of the major
relevance.
MACHINE INTERACTION
50 ··--···-----·· . ··-·--····················-······-·····-·················································
"'40
~
~30
tf)
tf)
w
0:
lï;20
10
a~~~~~~~~~~~~~~~~~~
0 0.01 0.02 0.03 0.04 0.05 0.06

STRAINS
Figure 8.1: Campression test stress strain curve.
Figure B.3 shows a test specimen tested using the special support shown in figures B.4, B.5
and B.6. 1t can be seen that in these cases the culm could expand freely in the lateral
direction, displaying longitudinal cracking as a main feature.
116
Figure B.2: Bamboo specimen after a compression test with no special consideration for
the machine interaction problem.
117
Figure 8.3: Bamboo specimen after a compression test with special supports.
118
Figure B.4: Special hinged support.
119
Appendix R Experimental analysis of short elements
Figure B.S: Hinged support showing thin teflon layers and greased surfaces.
120
Figure 8.6: Hinged support showing radial sliding elements made of steel.
The radial elements one can observe in tigure B.6, displace radially with the lateral expansion
of the specimen, as a result of remaining friction. They are made of pieces of galvanized steel
l mm thick, that are independent from each other, so that they completely fill the surface of
the supporting fitting. A layer of grease and teflon separates the galvanized steel radial sheets
from the support.
121
8.2-ExperimentaJ resuJts
Table B.l shows the descriptive stalistics for the maximum strains, both Iongitudinal and
transversal, for a sample of Bambusa blumeana.
Variable: long.strains tang.strains E modulus

Average 0.00317 0.001149 18784.7
Varianee 7.32x10. 7 2.68x10. 7 2.196x107
Standard dev. 8.56xl04 5.18x104 4686.7
Standard error 2.58x104 1.56xl04 1413.09
Table B.1: Descriptive stalistics for Bambusa blumeana in

compression. [ N/mm2 ]
One-Sample Analysis Results

Sample Statistics: Number of Obs. 25
Average 1.22804x 1o-J
Varianee 1.07117xl0-7
Std. Deviation 3.27288x10-4
Median 1.28xl0-3
Confidence Interval for Meao: 95 Percent
1.36317x10-3 24 D.F.
Table B.2: Analysis of pea.k tangential strains for Guadua s.p.
It was found that at a 95% confidence level, the peak value of tangential strains found for
Bambusa blumeana in compression is equal to the average value found for the pea.k tangential
strain from tangential tension tests. This suggests that it is possible to estimate the maximum
compression capacity by using the values of strains from perpendicular tension tests, if
122
--------------------
average values of the E modulus are known for a given batch.

The results on Guadua s.p. from Costa Rica show a highly similar maximum tangential strain
as can be seen in table B.2. The 95% confidence limit for the mean is also included in the
table. The similarity of the results is interesting, and it indicates the importance of exploring
these parameters for other species. It is fairly likely that a critical value for tangential strain
can be found for each species, or perhaps groups of species, allowing for a simplification in
the calculation of the capacity of the culms.
Further, no correlation could be found between peak strains and the elastic modulus. Figures
B.7 and B.8. illustrate these last two results. In the ftrst ftgure one can see that, culms tend
to fail at a commom maximum strain, as mentioned before, in spite of the wide range of
observed strengtbs in the sample.
No reliable correlation could be found between the peak stress of compression tests and
density, though, as indicated in section 3.1, several authors report that such a correlation
exists. A possible explanation for this is that density does correlate with the amount of friction
that can be generated between the specimen and the plates of the testing machine', on the
one hand, and on the other, that the reinforcing effect that the interaction machine-specimen
produces does increase this phenomenon. Figure B.9 presents data from Guadua s.p.
concerning the above. Figure B.lO further illustrates the relation between densities and
strengtbs found.
Descriptive statistles are included in table B.3 below. Moisture content was set at 12%.
1
which would obviously depend on the number of fibres
123
Variable: dry density streng tb E

Average 691.121 50.4958 18464.6
Median 718 50.85 18098
Mode 713.91 49.97 18097
Geometrie mean 684.869 49.8506 18248.5
Varianee 7918.63 66.2256 8.50777x106
Standard deviation 88.9867 8.13791 2916.81
Standard error 15.9825 1.46161 523.874
Minimum 448 36.5 14062
Maximum 843.5 67.15 25233
Range 395.5 30.65 11171
Lower quarille 644.61 42.16 15808
Upper quartile 742.63 57.69 20619
Interquartile range 98.02 15.53 4811
Skewness -l.l3173 0.038994 0.520005
Standardized skewness -2.57245 0.0886346 1.18199
Kurtosis 1.45706 -0.810767 -0.485961
Standardized kurtosis 1.65597 -0.921449 -0.552302
Coeff. of varlation 12.8757 16.116 15.7967
Table B.3: Descriptive stadstics for the sample of Guadua s.p.[kglm3 , N/mm 2]
The skewness and the Kurtosis coefficients of the distributions give further evidence on the
differences between the sample in relation to density, on the one hand, and strength and
elastic modulus on the other.
124
CCMI<'RHS tON TlSST

loW< IWW IIT1'IA I NIS VRII ln'RENGTH
•t- TAN3.1!11"'MtN
33
····································ii·UliiG~STIIII<IN-
22 ---.-------------------.-- --------------------------
... .. ............. .
11 -------.~·-:;fk ........."•.:t--------·'-.:··------!".. _
z
~ 0 ------------------------------ ----------~-----------
------------------------------------------- .. -------
... .. .
---------- _"""_--- ~-- ---------- ïi --..-------------
.
----------- ..... ----------iii -:--··· .• --...... ------ --
... ~
~~~~~~~·~~~~~~~·~==~~~
~
..
~ ~ ~
&TREIGTH w....-e
GUAOIIA 8. P.
Figure B.7: Strains and stresses compared
c:áMI'R&!I!II ON TIHIT
CX 11!•4:1 NI\XIM.M TANGöi!NTIAL STAAlN VR11 e K'IOUI..U8
2lJ - - - - - - - - - -- - - - • • - - • - - -
20 •••••••••••••••••••••••••••••••••••••••••••••••••••
+ t
~110 ............... ,F··t··;·············.;..t............... .
•+
·~
10 •••• !:••••••••••••••• ·+·· ----- ·····+··········--···
·•
11 - •
14 111 1a aa :za a.. 2&
e MODULUS NIIMI"'I! ex 10DO;l
Figure B.S:Tangential strains vs. E modulus
125
·•
800 .. -..............................- ................ ~ ......................... ............................ -
~
... .. .. .. ... +·!- ...
i 700 1-·························'···F···:~.. ~--~---··········

.x .. ... ::,.
i: ·• ....
ï 600 t-··········---------------·-·····--·····--------------
soo ••••••••••••••••••••••••••••••••••••••••.••••••••.•••
·•
400 • •• •• • •• •• •• • •• •• • • -- ••••••••••••
3D ?D
STR&NClml N/11111"2 ""
GVACUA IS.P,
Figure B.9:Density vs. strength for a sample of Guadua s.p.
COMI"RESSION TEST QUIIDUA S.P.
a
5
"' ..
~
J 3
è
7'1
DENSITY C lc0frn"3J
Figure 8.10:3-D histogram for density and strength, Guadua s.p.
126
1t is important to add some comments on the way the elasticity modulus was calculated,
because, though it is customary to report its value from compression tests, the way it is
calculated is rarely stated with clarity. Most of the time the elasticity modulus is calculated
as the secant value between zero and the peak load, though somelimes the 'limit of
proportionality' is taken for the calculation.
Figure B.ll shows the resulting plots of strains and stresses for four specimens of Guadua
s.p., from tests done using the frictionless set-up. As explained before, friction is eliminated
somelimes so that the failure is absolutly brittle, but, as is shown in the examples, it may be
that the cracked specimen is trapped between the supports, giving the impression of some
plasticity, which is not the case. The point is that under these conditions, which can be
ambiguous, it is difficult to establish what the 'limit of proportionality' is, which probably
leads to results that are difficult to compare. As can be appreciated from the comparison of
figures B.ll and B.l. it is even more difficult to avoid ambiguity where the test is run in a
traditional way.
The results presented in figure B.ll show that there is also some ambiguity at the beginning
of the loading path, which makes it seem as if the specimen was stiffer.
Tak:ing into account the purpose of using the elasticity modulus in the prediction of
deformations, and the need to make results comparable and self-consistent, it would perhaps
be actvisabie to calculate theelastic modulus asthesecant value, between ranges of 10 and
80% of the strength, or in any case, in a region where the curve strain stress is less influenced
by extreme conditions. During this research project the values calculated in this way were
compared tothesecant values between 0 and 100%, and between 0 and 50%. It was found
that there can be very big differences betweenone criterion and another, and therefore it is
of the major importance to clearly state the way the modulus is calculated in each report.
Finally, gratitude is expressed to the staff memhers of the Costa Rican Bamboo National
Project, who kindly supplied the studied sample of Guadua s.p..
127
Fundamentals
-L.DHIJ.S'1"MIN ...,. LQG.aTRAIH

+TNIO.IIIROIN + TAIG.STRAIN
15 --------- ............... ------ .................... ---------~-
-ao
Q 110 ...
S"'"JlUlW_...2
Bil Bil
·311
Q
"" ...
sn:ESIS H/ ......2
•• ••
SPECUI!Il F101CS1 SPECIIIIEN F1D1Ct2
ctM'IIlSS ION '!!!81" ClliOPIIEI!SIONTEI>T

ex 1B-<)
ao
- LDfrG, """IN
+ TANGO,STMIN
,. ...
!1
i ~
_" ....
·80
•• .., •• 80
•:!Hl
" 15TRE151!5 Wllll""à .......... w .......

ll'li!CU&N F1a1CI4S IFEC IIIIE!N t=:&1e:d
Figure B.ll:Strain stress plots for four specimens of Guadua s.p.
128
Appendix C: Experimental analysis of bamboo glued phases
Appendix C : Experimental analysis of bamboo wood gJued pbases

C.l-Bonding specimen.
Small samples of bamboo were fust taken out of the culm, their inner surface flattened, and
specimens as the one shown in tigure C.l were prepared.
Several probieros were found in practice. For example, the removal of part of the inner
surface (in order to have a flat face) introduces variations in the surface characteristics of the
specimens. This is provoked by the fact that the number of tibres present varies with the
radius, so there are more tibres at the borders of the samples because there the remaining
thickness is smaller. It is very difticult to keep this fact under control in a manner that allows
for the comparison and repetition of tests.
On the other band, the preparation of the specimens proved to be rather cumbersome, because
it is very difficult to keep dimensions and relative positions of the components under control
Because of partienlar interests in the gilling possibilities of the inner surface of the culms, it
was considered best to try to use sections of the culms kept as intact possible. So, the
specimen shown in tigure C.2 was considered to be a better option.
As shown in the tigure, two glue strips lO mm wide conneet the two pieces of bamboo to the
inner core of wood. While setting, pieces are kept together by clamps, or most preferably by
steel screw rings. The surfaces of both wood and bamboo are carefully cleaned and freed
from wax, dust or whatever other substance or condition that may be deleterious to gluing.
The bottorn surface of bamboo, which is in contact with the supporting plate in the testing
machine, is carefully levelled and made parallel to the upper surface of the wood cylinder to
avoid eccentricities during loading. Grease and teflon are placed between bamboo and the
steel support to avoid friction resistance, though this fact is not so important where the upper
surface of the wood cylinder is concemed. The specimen is located between two freely
rotating plates.
Deformation is measured by controlling the movement of wood. This is done by means of
pins located at opposite sites half way down the length of glue. Initially two pins were located
on bamboo as well, at about 1 mm from the inner surface, and two were located through
holes on the wood cylinder. This was done in order to observe slipping between the two
129
glued surfaces
j_
wood
10J
bamboo
û
cross section lateral elevation
Figure C.l:Specimen for determination of gluing capacity. Dimensions in [mm].
matenals and to be able to rate the quality of the specimen. For normal measurements, only
the pins on wood are considered necessary. It has to be stated though that the readings from
such pins correspond to the movement of a partienlar point on the skin of the wood cylinder,
but not to the actual strains at the gluing surface, though the metbod does give an indication
of the shape of the load deformation path for these glued phases. The area of glue is kept as
small as possible, so that the wood cylinder does not significantly deform during loading.
The speed of toading was set according to ISO 6891-1983 (E)(Anonymous, 1983).
130
Appendix C: Experimental analysis of bamhoo glued phases
glue strip 10 nnn

widtho
I 50
_l
PLAN VIEW 0
LATBRAL ELEVATION
Figure C.2: Specimen used for bonding test. Dimensions in [mm].
C.2- Load deformation relations and influencing factors

An example of a stress deformation curve is given in figure C.3.Most of the observed results
showed a very smooth and continuous relation between stresses and deformations, as seen in
the figure. Failure occurs in a brittie manner, and was found within the frrst mm of the layer
of bamboo for all cases studied, as will be explained in the following lines. Before going
ahead with the description of results on capacity, deformations and other features, it is
necessary to explain the results of some experiments, to determine the degree of influence of
certain factors on the qualities of the phases.
Studied factors were :
- Bamboo density.
- Bamboo thickness.
- Bamboo initial diameter.
131
BONDING STRESS DEFORMATION CURVE

e . .. .. ... ... ... ·-· ·-· ... ... .... ... .... .... -· ... ........ -·· -- .. .. ..........
SPECIMEN GOS
~
~ 4 -···-··-··-··-···--·····-···-·····-· ··········-····· -·····-·····-···-···-···-···········-······--·
"'"'
~
0 0.1 0.2 0.3 0.4 0.5

DEFORMATION mm
Figure C.3: Typical stress deformation curve.
- Type of wood either hard or soft.

- Type of glue, either PVAc or Araldite.
A total of 32 specimens were prepared, 16 for each type of glue. Half of each block was
made using hard wood, and the other half using soft wood. The initial diameter, the thickness,
and density were vatlables that were entered mndomly into the experiment. The analysis of
vatiance for the experiment is shown in table 4.1. The results have been already discussed in
chapter 4, but a few other elements are added bere for the sake of completeness.
Table C.l shows the result of pooling each one of the glues together (that is, involving the
two types of wood for each type of glue). This was done in order to have some indication of
the variability of the data in this way, and to be able to calculate material factors for the two
glues, so that design values could be compared.
The material factor in the table is worked out by using the expression
132
W = (1+(t-o,9375 (1-6,25 v:)~T1 eq.C.(l)

1-6.2SV:
(for a reference on this equation see for example, Huybers, 1990). Here v. is the
coefficient of varlation and T is a time factor, normally taken as 0.8.

1
To add an extra element to the analysis, a statistica! comparison of the two sets of results was
undertaken, and the result of the test of hypothesis is given in table C.2.
Variable: wood glue re sin

Sample size 16 16
Average 5.78555 6.32779
Median 6.0937 6.1795
Mode 5.548 5.905
Varianee 0.733879 1.0612
Standard error 0.302878 0.364212
Minimum 3.929 4.85
Maximum 6.57 8.348
Range 2.641 3.498
Coefficient of varlat 0.1469 0.1627
Material factor 2.985 3.1629
Design valne 1.94 2.00
Table C.l: Summary of stalistics for strength [N/mm 2].
133
Appendix C: Experimental analysis of bamboo glued phase$
Two-Sample Analysis strengtbs
wood glue resin Pooled

Sample Statistics: Number of Obs. 16 16 32
Average 6.32779 5.78555 6.05667
Varianee 1.0612 0.733879 0.897541
Std. Dev. 1.03015 0.856667 0.947387
Median 6.1795 6.0937 6.1387
Hypothesis Test for HO: Diff = 0 Computed t statistic = 1.1447

vs Alt: NE Sig. Level = 0.271526
at Alpba = 0.05 so do not reject HO.
Table C.2: Comparison of the two glue types.
As can be deduced from the table it can be stated that there is no difference between the
strengtbs of the two types of glues. The difference is even less significant for the design
values, as shown in the previous table.
One fmal element concerning the influence of the two glues is examined.
When dealing with structures, the stress level is not the only concern matter, deformations
should be looked at too. In order to see the effect of the different factors on this parameter,
an analysis of varianee for the final deformations was undertaken. Table C.3 shows the
resnlts, and table C.4 shows the results of calculating the 95% confidence limit for the means.
Certainly, it is clear that though the results are below significance, different glues give
different end deformations. The same can be said of bamboo density, and the table also shows
134
as that the type of wood is even less of an infiuence in this matter.
Analysis of Varianee for DEFORMATION
Souree Sum of d.f. Mean F-ratio Sig. level

of varlation squares square
COVARIATES .0066216 3 .0022072 .773 .5409

density .0056442 1 .0056442 1.977 ,1974
diameter .0000204 .0000204 .007 .9356
thickness .0000734 1 .0000734 .026 .8783
MAIN EFFECTS .0056313 2 .0028157 .986 .4142

woodtype .0001708 1 .0001708 .060 .8155
gluetype .0055857 1 .0055857 1.956 .1995
2-INTERACTIONS .0001179 1 .000118 .041 .8461

GLUE-WOOD .0001179 1 .000118 .041 .8461
RESIDUAL .0228419 8 .0028552
TOTAL (CORR) .0352127 14
Table C.3: Infiuencing factors on fmal deformation.
The results seem to suggest that phases tend to fail at about the same level of deformation.
At the level of the design stress, deformations are also similar.
The analysis of the corresponding values shows that the average deformation at the design
level of stresses is about 0.09 mm, without a noticeably significant difference between the two
135
Appendix C: Experimental analysis of bamboo glued phasis
types of glue, as found in a statistica! comparison.

It is also of interest to see if tilere is any correlation between tlle maximum stress and tlle
maximum deformation.
Such an analysis was ondertaken for wood glue only. A correlation analysis for strengtil and
deformations gives a coefficient of (-)0.3731, but tlle significanee level of a Stuctent's test is
0.1546.
Std. Std. 95% Confidence

error error for tlle mean
Level Average (internal) (pooled s)
woodtype
h 0.27349 .012573 .0201963 0.22690 0.32007
s 0.26610 .022508 .018891 0.22252 0.30968
glue type
r 0.25301 .010497 .0201963 0.20643 0.29960
w 0.28401 .021957 .0188919 0.24044 0.32759
INTERACTION WOOD-GLUE
h r 0.25277 .014465 .026717 0.19115 0.31440
h w 0.30110 .004563 .030850 0.22994 0.37226
s r 0.25330 .018782 .030850 0.18217 0.32449
s w 0.27376 .035687 .023896 0.21864 0.32888
Total 0.26955 .013797 .0137967 0~23772 0.30137
Table C.4: Confidence limits for tlle mean maximum deformation.
In otller words, evidence shows that no correlation exists between strengtil and fmal
deformation.
136
Appendix C: Experimentol analysis of bamboo glued phases
In order to observe the way strains vary along the Iength of glue, a different test specimen
was designed as shown in figure C.4. Small strain gauges were glued along the length of the
specimens, and readings of the strains were taken constantly during loading. Data were later
summarized so that strains at each one of the strain gauges could be plotted against the
position of the strain gauges along the length, as seen in tigure C.S. In this way it is possible
to obtain a plot for specific levels of force. Two of these plots are given in the figure with
an indication of the force level as a percentage of the maximum reached value. The graph
gives an indication of the approximate linear varlation of the strains from one end to the
other. The angle of the linear varlation increases with increments in load.
D : strain gauge positi.on
D
D
D bamboo
D culm
I DI
I D I
L..!!.J
ELBVATION
_____"sl:::._ued_surfaoes
Figure C.4:Test specimen for the distribution of strains. Dimensions in [mm].
The plot helps to visualize to what extent one can be precise in assuming a certain
distribution of strains along the length of glue. lt can be seen that at the design level a
137
Appendix C: Expertmental analysis of bamboo glued phase$
LONGITUDINAL STAAlN DISTRIBUTION

ex 1E-"1)
8 ----- --------------
- 2~ END LOAD
-+- END LOAD
0 --------------------
0 20 "'O ao ao 100
POSITION ALONG THE LENGTH ~
Figure C.S:Strain distribution along the lengtil of glue.
constant distribution of strains (and therefore of stresses) can be justified, but for higher levels
of load the distribution becomes more complex. According to the plot, strains tend to be
larger towards the side where the wood cylinder projects beyond the connecûon. A possible
explanation for this is that the culm is completely glued to the piece of wood on this side ,
so that it is constrained to move radially. At the other extreme, the culm expands more freely,
due to compressive forces, so therefore less stresses can be transferred along the glued phase.
The higher the load, the more important this mechanism becomes, and thus the more
pronounced the effect shown in figure C.5 is.
C.3- Gloed phases onder the effect of dynamic loads

Four different types of experiment were set to study the effect of dynamic loads on glued
phases, as is explained in the following.
138
Five groups of six specimens each were prepared using soft wood and PVAc glue, as
explained insection C.l. The frrst group was set as a reference, each specimen was steadily
loaded till failure, and relevant infonnation was recorded.
Each one of the remaining groups was subjected to different load conditions, and the result
of fmal capacity and defonnation was compare(l to that of the reference group. In other
words, each load condition was regarded as a different treatment from a statistical point of
view. As will be seen, statistica! comparisons were not always necessary, because of the
clarity of the results.
C.3.1-Load type one.

At the speed of toading prescribed by ISO specimens were loaded up to 100% of the
calculated design load\ then unloaded to 10% of that same value, and finally the cycle was
repeated 20 times. A result was presented in figure 4.3. Mter the 20th cycle, loading
continued till failure. 1t can be observed that though, some degree of consolidation occurred,
as was mentioned before, it looks lik:e the curves retumed to that of the reference group, after
the last cycle. The statistica! comparison of both ultimate load and ultimate deformation
showed no significant difference between this group and the reference group, though final
defonnation was found to be slightly higher for specimens under the repeated load condition.
C.3.2-Load type two.

Another group of specimens was loaded following the same load pattem as that of load type
one, except that the maximum load in the cycling phase was increased to about 250% of the
design load, or about 80% of the ultimate capacity. Speed of toading was kept the same.
Figure C.6 below shows a typical result As seen in the figure, the effect is clear enough in
showing an accumulative deterioration. Plastic permanent deformations are reached from the
flfSt cycle, and the specimen fails after just a few loops. As in all other cases, failure occurs
1
the ultimate load reduced by tbe material factor calculated following eq.C.(l).
139
Appendix C: Experimental analysis of bamboo glued phase$
within the frrst mm inside bamboo, however, in this case far more deterioration is observable
in the matrix of lignin. 1t was thought that perhaps the total amount of deformation, and not
necessarily the amount of stress was the souree of this early failure, so the next experiment
was set to observe if that was the case when the stresses were kept under the design value.
CYCLIC LOADING 10-80% ULTIMATE CAPACITY

8 -- -- --- ------ - --- --- -- --------- --- -- - -- -- -- --- ---- --- ---· --· ·-
6 ----------·-· ·-·---··--·-----------------·---------------------··----·-·---·
~
~ "' ···-·· -···- . . ..........................................·-···--···--·--·----·-----·-··-··-··-···
w
In
0~====~==========···=··=·=···-~·-=···=···=···=···=···=--··~··=···=····=···=··=···=····~
0 1 2 3 5
OEFORMATION mm
Figure C.6:Cyclic load for second load case.
C.3.3-Load type three

The third group of specimens was subjected to a completely reversing signa!, keeping the
same speed of loading, with amplitudes between -100% and +100% of the design value.
Again, after 20 cycles results were compared to those of the reference group. No significant
duferenee could be observed between the capacities or the ultimate deformations. A typical
result can beseen in figure C.7.
140
REVERSING LOADING -100,+10~ DESIGN LOAD

8.'1 - · ......- ....... ·- - - ,. ___ .... ·-
111
OEFORMATIONS rrm
Figure C.7:Cyclic load for third load case.
C.3.4-Load type four.

As reported in reference 1 chapter four, eertaio load signals were chosen after the examination
of several earthquake registers. In principle the idea was to try to replicate as closely as
possible the type of frequency trains that may be typical of the initia! phase of seismie
signals, the intensive phase, and the final one.
It was determined that registers could be represented by the extreme conditloos of frequency
and amplitude indicated in figure C.8, but because of equipment limitations only the second
signal could be implemented, with restrictions in the possibilities of data acquisition. Only
time load records were possible, and only realtime plots of that were possible, so no data ftles
were kept for post-processing of the results.
Actually, in the end, each specimen was subjected to a different load pattem in a random
way, with some level of control over the peak amplitude of the signal only. In some cases the
141
Appendix C: Experimental analysis of bamboa glued pháses
experiment was allowed to run up to 100 cycles in a kind of random combination of

amplitudes between 40 and 100% of the design load, at approximately 0.5 hertz. In any case,
since no damage was observable in the specimens, they were allloaded to faiture afterwards,
and the capacities and deformations were compared to those of the reference group. Again,
no significant difference was found between the two groups.
FORCE AS A 5 OP DESIGN CAPACITY
100
TIME
CYa.BS 11)1-- 20 - - - 1 - - 10 _.....,__ _ 20

O.S HERTZ O.S HERTZ 0.5 HERTZ LOAD 1
FRBOUBN-mt
CY 20 HERTZ 20 HERTZ 20 HERTZ LOAD 2
Figure C.8:High frequency signals.
C.4-General Remarks
A large number of observations were additionally made in the laboratory, related to the
response of gluing phases to dynamic loads, beyond the initial goal of the experiments.
Results are, clearly, indicative, including the calculation of the design level of stresses.
The major goal of this phase of the project was to study the feasibility of gluing in
bamboo construction, and not the prescription of design parameters, because to reach
142
tbat point more stndies are necessary.

One more example of laboratory observations is included in figure C.9 below.
REVERS I NG LOAD I NG
5.2 ..... ··-······- ................................ ·····-· .. ·- ·-- ·- -··-··-
3.2
Ul
Ul
~-0.8
-0.7 0.1 0.5 0.9 1.3 1.7

DEFO~TIONS
Figure C.9: Response of a glued phase to different amplitudes.
This specimen was subjeeled frrst to two cycles of -100,+100% of the design load, then two
cycles of -200,+200% of the design load, and then loaded to failure. The plot shows how
deterioration happens soon after stresses get closer to expected non-dynamic strengths.
But after the observation of the results of the different types of loads, evidence suggests that
up to the design value glued phases are able to sustaio cyclic and repetitive loading without
significant loss of properties, though, again, somewhere above this limit deterioration happens
after a very low number of cycles, even at very low speeds of loading. The same evidence
suggests that the speed of loading is not a critical condition in the testing of this type of
specimen. 1t can be concluded that gluing can be an interesting option for the solution of
certain problems pertaining the use of bamboo for applications like fumiture mak:ing, but in
143
construction actlvities as well, though more facts are shown regarding this respect in appendix
D.
Finally, some tests were run to observe the behaviour of full-scale connections made by
gluing, following the design proposal described in chapter 5. Some traditional connections
were also tested to have a reference.
TRADITIONAL CONNEeTION
200 - .. ·-· -· ..•.... --·- ··-
120 ····-·····-···-·········· ···-······-···-····--···-···-···-·--·-----···-··-··-··········--··

z
~
~
0 ··-····
0 20 '10 60 60 100
0 I SPLACEMENTS nrn
Figure C.lO:Force deformation curve for a traditional connection.
All tests were done on cantilever specimens. Traditional connections were made using
galvanized wire 3 mm in diameter, wound around the culm and a wooden support, as shown
in tigure C.12. The surfaces of contact between the culm and the support were made to
perfectly fit each other, to avoid undesirable concentration of stresses. Figure C.IO shows an
example of one of the resulting force deformation diagrams. It is noticeable that a large
amount of rotation happened during the test. Actually, the culm became loose, being unable
to snstain any reversion of loads, deformation at the ends of the culm were very large under
144
a comparatively low load, as seen in the figure.
GLUED CONNECTION
ex 1ooo' SPECIMENS BEAM01 ANC TRUOI01
2~~========~====--=====~====~==~
O.B
0.4
20 so BO 100 120
OEFORMATIONS AT LOAD POINT mm
Figure C.ll:Force deformation for glued ( upper) and traditional ( lower) supports.
Figure C.ll shows one example of a force deformation plot for a glued conneetion (as
indicated in tigure C.l3), together with the previously shown result, that is included here to
facilitate the comparison. The straightness of the glued conneetion plot indicates the small
degree of or nonexistent rotation at the support up until the failure load was reached.
To confirm this, a plot of maximum strains at 20 mm from the support is presented in figure
C.14. There, one can see that continuity is achieved to a good level, with no observable
participation of rotations in the end deformation of the culm.
Figure C.15 shows maximum strains at sections located 20 mm from the end of the wooden
support, one of them in the support itself, and the other on the free standing part of the culm.
The distance between the two sections is comparatively small, so the difference between the
two levels of strain is due to the change in the second moment of area, but it can be seen that
145
plots remain straight, indicating üttle rotation in this other example. Tested beams had a
gluing length of about two times the diameter, which, as will be shown in appendix D, should
leave no room for failure in the connection, as happened in reaüty. Failure occurred by
compression. Maximum strains were well within the same range of those observed during
compression tests for bamboo, as indicated in chapter 3 and appendix B.
Table C.4 contains a description of the two beams presenled in the previous plots .
FORCB
steel frame
TBSTSBTUP
wood support ..
CONNBC110N DBTAD..
Figure C.12: Traditional conneetion test set-up.
146
specimen Extemal diameter thickness length E

[mm] [mm] [mm] [N/mm2]
Traditional 86.68 6.98 1500 18627
Glued 84.44 7.60 1485 18303
Table C.4: Properties of the compared beams.
steel
frame
glued wood fitting
CONNECTION DETAIL
Figure C.13: Glued conneetion test
147
Gl.UEC COIIINECT ION
(X 1000) IIEAMD1'
2 ~====~====~-===~-===~~~==~====~
1.8
a a.s 1.5 2.5 3

ex 1e-a'
Figure C.14: Glued connection, beam 01, maximum strains at 20 mm from the support on
the free standing culm.
GLUEO CONNECTIONS
ex 1E-3:l eeAM os
..
4~·=··=-=--=-~==-==--~-~--~-~-=-·~-~---F·--~-~.m.==~·-=-~---m=--~--=--·=··=--·=·=··~-
- SAMeco
-t- CXJNNECTION
o.s 1.5 2 2.5 3

FORCE N c:x 1000)
Figure C.l5:Comparison of strains at both sides of the limit between the support region
and the culm.
148
KEFERENCES
l Huybers, P.; 1990. Thin Poles of roundwood for Structural Engineering Applications
in Building. Structural Engineering Review.2, pp.l69-182.
149
Appendix D: Glued connectors for bamboo
Appendix D:Glned connectors for bamboo.

D.I-Calcnlation of bonding stresses
D.I.I-Basic equations of elasticity
The distribution of bonding stresses, on the surface of contact between the two materials, is
examined in the followîng sections, according to the reference system of tigure D.l.
Figure D.I:Reference system.
The analysis is based on the following basic assumptions :
150
FUndamentals of the Design of Bamboo Structures
1- Bernooili's assumptions bold true

2- Each material is homogenons and isotropie
3- There is no slipping in the interface between the two matenals
4- Sections are perfectly circular
5- There are no discontinuities in the glued surfaces
6- The maximum strain capacity of bamboo is predictabie
7- The tbree matenals behave elastically in tbe range of analysis
Stresses on tbe surfaces of the infinitesimalelementof figure D.l (bottom) are
iJoz o-cJCY o<t:IQ:

1; 0% + - 2: 'tJCY + - - 3: "CJQ: +--
ox ax ax
ao, 0-cy.t 0-c~
4: o, + -.- S: 'ty.t + - 6: 't~ +-
8y 8y 8y
ao:t 0-cu 0-cey
7: or + - 8: 'tu + - 9: "Cey + -
Oz Oz Oz
For an infinitesimal element in the body of the connection, as the one shown in figure D.l
(bottom), the following well known equations of elasticity are examined (see for example,
Timosbenko,l984).
If no body forces are present tben the state of equilibrium of the element is given by the
relationships
eq.D.(l)
ao1 0-cJt1 a ..ey eq.D.(2)

- + - + - =0
ay ax az
eq.D.(3)
151
Fundamentals
On the other hand, for a particular element on the extemal boundary of the conneetion it
holds that
eq.D.(4)
eq.D.(S)
eq.D.(6)
Compatibility of deformations is given by
&ex &e., _ fPy yx eq.D.(7)

-
aya+axa
--- -
ayax
&e., &ez. _ éPy

14 eq.D.(8)
-+----
& 2 ayz c3yäz
&ez. &ex fPy .tt eq.D.(9)

-+-=--
&2 azz axaz
eq.D.(lO)
eq.D.(ll)
eq.D.(12)
Under a state of only axial stresses
152
-------
o=o=~ =0 eq.D.(13)
' y ;'/
Bending stresses follow the rule
M,z
0 =- eq.D.(14)
% I
."
I." is the equivalent second moment of area of the two material section, and will be dealt
with later.
Substitution of the above values in equations D.l to 0.3 implies that
0 eq.D.(lS)
ék eq.D.(16)
__l:! =0
öx
z ék a~
-M '+ ___!:! + --l:! = 0 eq.D.(17)
1." ' az ay
Since forces are zero on the surface of the connection, then
eq.D.(18)
Now, because I= dy m = -.-d%

- then
ds ds
1
Note that this is true when the length of the conneetion is small compared to the length
of the culm.
153
Fundamentals
eq.D.(19)
Using Hooke's law, the equations of compatibility can now be written as
eq.D.(20)
eq.D.(21)
eq.D.(22)
eq.D.(23)
eq.D.(24)
eq.D.(25)
ae M'
-àz =-'
I.,
eq.D.(26)
eq.D.(27)
This set of equations reduces to
154
, Appendix D: Glued connectors for bamboa
eq.D.(28)
and
M' eq.D.(29)
~1: = .,
rx (v + l)I..,
lt can be seen that the equations of equilibrium are satisfied if
eq.D.(30)
and
1: = -at, eq.D.(31)
"' az
<p is a stress function on y and z, and f(y) depends on the boundary conditions.
D.1.2-Stress due to sbear.

Transversal· forces generate shear stresses that affect bonding stresses. This effect can be
accounted for in the following way.
Back-substitution of equations 0.(30) and 0.(31) on equations 0.28 and 0.29 gives
eq.D.(32)
155
àcp
&
= __z!_
[ 21.,
l
M' 2 -Jf:y) dy
da
eq.D.(33)
Since <p must be constant along the boundary, then j(y) must be such that the right side
of equation 0.33 becomes zero. If the attention is now concentrated on the glued region of
the connection, then, for a circular section of extemal radius re , the extemal boundary is
defmed by
eq.D.(34)
Therefore
eq.D.(35)
and thus
M'y
f'(y) = _=z:.. eq.D.(36)
I.,
Appropriate substitutions show that
êPcp êPcp _ v M,'y M,'y eq.D.(37)

- + - - ----+--
&2 ay2 1+v I., I.,
or
156
iP~p + iP~p = ( 1+2v) M,'y eq.D.(38)

at2 ayz 1+v 1.,
A solution of this equation is
cp _
- -1 (1+2v)M,'(
-- - z2+y 2 -r 2\.,
V eq.0.(39)
8 1 +v 1., •
In order to obtain stress values and taking the convenience of using know polar coordinates
into account, this boundary function can be now back substituted in 0.30 and 0.31 to obtain
-r M,'[
_- (3+2v) - r 2 -r 2~-~2. el-2v 2~•-'Ze]
CO"> - - r sur eq.D.(40)
~ 8 (1+v) I*" " 3+2v
'r
1+2v
=----r
M,' 2 case sine eq.D.(41)
yz 4(1 +v) I .,
Forces can now be calculated for an elemental sector sucb that
lf a unit are is defined on the gluing surface, tben dA = dr and the force acting on sucb
elementary sector is
157
d , "., _
v~
.3+2v
- - --- - r 2-r 2---2"'
M,'[ - - r 2...:-ze]dr
con::J -1-2v llW.l
eq.D.(42)
8(1 +v) I • 3 +2v
"''
1 +2 v M'' r 2 cos8 sin8dr eq.D.(43)

dv)'X dx =
4(1+v) I"'~
After double integration the above equations become
3+2v V" [r 2 r -r -- ~ cosz

1 (r 3 -r •)
8(1 +v) 1"'1 ' (' .) , 3 t •
e
eq.D.(44)
V
u =
_.! 1-2v (r3-r3)sin2e]
1
3 3+2v '
and
v = - (1+2v) yx(r,3-r,3)cose sin8 eq.D.(4S)

yx 4(l+v) 1"'1 3
lt is conventent to convert shear stresses into tangentlal and nonnal stresses on the sarface
of contact between the two materials, which can be done as
eq.D.(46)
eq.D.(47)
After proper substitutions, the above two equations become
158
eq.D.(48)
3+2v y%[r 2tr -r) sin8-.!(r3-r 3)

v _ 8(l+v)I.., •n • 3 ' ' eq.D.(49)
T - (cos28 sin8 (1 +2 3+2v
1 +2v) -( 1-2v
3+2v
smsa)]
The maximum valnes of these shear forces occur at the value of e , for which the frrst
derivatives of the above equations are zero. It can be shown that the maximum normal force
occurs at 8 =0 and it takes the value of v~ , and that the maximum value of the
tangentlal fórce occurs at 8 = 2!. , and that it takes the value of v~ . Now, in order to
2
see the significanee of these stresses, a parametrie analysis is ondertaken in the next section.
D.1.3-Effect of bending on tbe value of bonding stresses

Under a condition of pure bending, stresses acting on an elementary sector will be Iike those
indicated in figure D.2, where
159
Fundamentals
bamboa BENDINGAND
T.RANSVBRSAL
SHBAR
1M Cl
.!JL~-----==-------oliiii~~..._~>-"---==..!...
BOUNDA:R.Y oX -1 1- dx
LONmroDINAL
SHBAR
[3 lt v'b.b 181
N
RADIAL RBACI'ING
TBNSION
.b.. •
Figure D.2:Stresses on an elementary bamboo sector.
eq.D.(50)
eq.D.(Sl)
B=l!..e". E", eq.D.(S2)
The total resistant moment of the section can be calculated by taking moments around the
neutral axis, so that
160
Eu.x T sin9 .
dM.1 = drdS .! E111 r sin9 eq.D.(S3)
2
Since
dS = rd9
then the total moment is given by
eq.D.(S4)
which afler integration and grouping results in
eq.D.(SS)
The term in brackets is the equivalent stiffness of the composed section of two materials.
Equation D.55 can be simplilled as
eq.D.(S6)
where
eq.D.(S7)
and
161
eq.D.(58)
The maximurn strain in the section is given by
M,r" eq.D.(59)
E =--
m E6 11Ml
and therefore
(J
M,r"
=-- eq.D.(60)
.. I
eq
Axial stresses of a partienlar point on the section can thus be calculated as
eq.D.(Cil)
The corresponding force, after integration works out to be
M
N
x
= __l_.oosO(r
2I i
2
-r• 2) eq.D.(Ci2)
eq
(force per unit sector)

Along the length of glue, the bonding stress will be
eq.D.(Ci3)
Therefore
162
eq.D.(64)
0.2-Parametric study
Since normal and tangentlal forces reach their maximum at 0 =0 and 0 = ~

2
respectively, then
ll = 3+2v v"'[r2fr-r)-..!fr3-r3)] eq.D.(65)

N 8(1+v)I." ·~· • 3n •
and
eq.D.(66)
1t can be assumed that v = 0,3 . On the other hand, the author found that for Bambusa
Blumeana from the Philippines, and for Guadua s.p. from Costa Rica, t =0,09 «!> f),or
r1 = 0,82 r. . Substitution of these values gives
(8=0) eq.D.(67)
2
See appendix G.
163
In the same way it can be shown that
eq.D.(68)
and
(8=0) eq.D.(69)
Equilibrium indicates that
eq.D.(70)
and
l 1 0,23V"
1
ob-- = --=--_;;;,_..---= eq.D.(71)
22 'ltr., (o.4s-P;l]
This equation can be reordered as
0,2944 vCIIIrn
eq.D.(72)
[o.45 _p;I] 06t~t~m
The extreme right tenn of the equation is the ratio of maximum design values, for the
164
Furulamentals of the Design of Bamboo Structures
transversal shear in the culm, and bonding capacity in the glue, respectively.
It is convenient to see the effect of changes in the parameters involved, what can be done via
a plot li.k:e shown in figure 0.3.
It can be seen that for average material properties a glue length of about 1.5 times the
extemal diameter of the culm will guarantee that failure will not occur within the region of
the conneetion due to bending alone.
GLUE INOEX PLOT

15 - - • - • - • • • • • • • • • •• •• •• • • • • • • • -
+ p.1.5
12
* p.1
•••• ,(ji......,.O..$........................................... .
0 ••••••
0 1 2 ,3
STREIGTH RATIO
Figure D.3:Parametric analysis of equation D.72.

p=E/E.,.
D.3-Stresses doe to pure tension or pure compression

If stresses are to be kept equal or below the design valnes found in appendix C, then it is
reasonable to assume a linear distribution along the length of glue. In that case the calculation
of stresses becomes very simple, because their average value can be obtained by dividing the
165
total tensile force by the area of the connection.
REFERENCES :
Timoshenko,S.P.; Goodier,J.N., 1984. Theorv of Elasticity. Third Edition. Me. Graw

Hill International Book Company.
166
Appendix E: Bamboo columns

E.l-Calculation of the criticalload
E.l.l-Introduction
In what follows the criticalload of bamboo columns is investigated. The role of the geometry
and of the changing mechanical properties of the culm is analysed. The influence of the nodes
is not studied in this chapter.
A direct link is established between the theoretica) analysis and the experimental phase, since
both are based upon an adaptation of the Southwell Plot procedure, which enables the
experimenter to take care of the problem of initial deformations during testing. A detailed
description is given in the following sections.
Since the measuring procedure is crucial for good results, a complete description is given of
two options devised during the development of this study.
E.1.2-InDuence of initial twisting deformations

As experience shows, bamboo culms are naturally twisted to some extent. Fibres run almost
parallel to each other within the internode region, but they are not necessarily parallel from
internode to internode, so that this initial condition leads to twisting under axial loading.
It is well-known that twisting and lateral deformation participate in an uncoupled manner in
a column, as will be shown in the following lines.
The initial deformation can be described by a sine series of the form
.
cf>(x)=:E C,.sin(n7t~) eq.E.(l)
n=l l
The figure shows how fibres elongate under the twisting of the culm. It is assumed though
that straight fibres before loading remaio so afterwards.
It can be seen that
167
Fundamentals
dll.x =~(x) eq.E.(2)

rdcl>(x) dx-dll.
which after elimination of second order differentials becomes
eq.E.(3)
or by introducing the value of the derivative,
Figure E.l:Geometric considerations on twisting.
--l.nn ~)r
Ë c,.",../
dll.x =[r 1t eq.E.(4)
dx l,.=l l
1Q8
The work done by twisting is therefore
eq.E.(S)
or
eq.E.(6)
The bending moment at a distance x along the culm is
eq.E.(7)
or simply
eq.E.(8)
The work done by bending can be calculated as
eq.E.(9)
and since
df.J=K dx eq.E.(lO)
this results in
eq.E.(ll)
On the other hand the work done by the axialload is, using eq.E.4
169
eq.E.(12)
since a stationary value is looked for then
eq.E.(l3)
Since the second integral vanishes only the last one remains associated with the initial
twisting deformation of the culm. If a further assumption is taken, in the sense that it is
always possible to express the angle of twisting using the ftrst term of the series only, the
contribution of this term only remains important depending on the values of deformations
actually occurring during loading. It is assumed bere for the time being that this value is
random with respect to the force and that it is very small, but this point will be discussed
later on in the section on experimental results.
With this in mind and due to the fact that
170
Fundamentals of the Structures
eq.E.(14)
in order to have a maximum value of the force, then
eq.E.(15)
or
eq.E.(16)
which is the first Euler Ioad, unaffected by the presence of initiaJ twisting of the culm.
E.2-Buclding load of an iuitiaJiy crooked tapered bar with variabie modulus of elastidty
Let us soppose that the bar shown in figure E.2 is initially deformed in a random way, that
can be described by the series
eq.E.(17)
Let us also suppose that the solution of the differentlal equation for the pin-ended strot can
be approximated by
eq.E.(18)
Two further assumptions will be made in relation to the elasticity of the element and its cross
section. First, soppose that the Young's modulus varles from the smallest section following
the rule
171
Figure E.2: Initially crooked bar.
eq.E.(19)
and second, that the moment of inertia of the section follows the rule'
eq.E.(20)
it can be noticed that
1
Actually a linear relationship would do as well, as explained in appendix G, though at
the time this section of the research project was undertaken evidence for this was not
available. An option in that sense is added at the end of this appendix.
172
Fundamentals
E(l)- E(O) = -& E eq.E.(21)
and that
eq.E.(22)
and in the same way
eq.E.(23)
The critlcal load of the strut can be calculated by finding the maximum potentlal energy of
the system.
The potentlal energy due to bending in the bar is equal to
Ub = - 1 I'E(x)l(xlJtP<w-wJf
dx dx
eq.E.(24)
2 0
2
It also can be shown that the potentlal energy due to the shortening of the bar is given by
eq.E.(25)
Substltutlon of the respective values of the derivatlves and products and the fact that the
derivatlve of this functlon with respect to b,. has to be zero gives
173
Fundamentals
eq.E.(26)
After integration and some further simplification eq.E.26 can be reduced to
0 = A (b,. -a,.) - b,.NJl eq.E.(27)
where
eq.E.(28)
and
eq.E.(29)
therefore
Aa',._
b = __ eq.E.(30)
" A-Njl
The lateral deformation due to the application of the load is
174
Appendix E: Bamboa columns
w = b,.-a,. eq.E.(31)
w an w
--- =- eq.E.(32)
Nz A A
B B
A,. B
which is the equation of a straight line with intersection at and slope
A A
B
Eq.E.32 is a varlation of the Southweli Plot (see for example, Gregory, 1967) procedure that
accounts for variations in the second moment of area and the modulus of Young.
Forther examination of the valnes of eq.E.27 can be done taking into account that in order
to have a maximum value for the equation of energy
eq.E.(33)
so that
eq.E.(34)
or
eq.E.(3S)
Substitution of the corresponding values gives
175
Fundamentals
eq.E.(36)
which is therefore the magnitude of the criticalload of a perfectly straight bar with a tapered
cross-section and variabie Young's modulus, as a function of properties at the extreme of the
elements and their variations.
'E.3-Theoretical boelding Joad of bamboo culms

It can be noted that the flrst component of eq.E.36 corresponds to the buckling load of a bar
with constant elastic properties along its length. The other three components can be regarded
as corrections due to the presence of changes in these properties.
In the case of bamboo it is possible to plot the result of eq.E.36 as a sequence of graphs each
repcesenting the influence of each component for a family of valnes of the variations in these
parameters for the ranges found in the field.
Such a set of graphs is presented in flgures E.3 to E.6 in the following pages. In the figures
units are Nlrmrr, and
1
À=--
~
eq.E.(37)
176
BUGKLING OF BAMBOO CULMS

FIRST COMPONENT OF GRITICAL STRESS
+ E1=10000
"'*" E1=12DDD
80 ------------------------------------=&- ~-1=-1-4.000.------
* E1=16000
~ E2=18000
C\1 -- E3=20000
(
E
60 -----------------------------+--Ë4;;2-2öoö____ _
E
"z
(/)
(/)
w
([
1- 40
(/)
20 ---------------------
30 50 70 90 110 130 150

LAMBDA
Figure E.3:First component of the critica! stress for bamboo [N/mm 2].
177
Fundamentals
BUGKLING LOADS OF BAMBOO CULMS

SECOND COMPONENT OF CRITICAL LOAD
+ E1-E2=1000
* E1=E2=2000
12 -------------.. -... -..... ------%-- -E-1- E2-4DOO.-·-------.. -
* E1-E2=6DOO
+ E1-E2=8000
(\J
< 9
E
E
'z
(fJ
(fJ
w
er 6
1-
(f)
3 ·········-
30 50 70 90 110 130 150
LAMBDA
Figure E.4:Second component of critical stress[N/mm 2].
178
Fundamentals of Bamboo Structures

THIRD COMPONENT OF CRITICAL STRESS
+ E=1DOOO
* E=12000
-& E=14000
* E=16000
~ E:::18000
C\1 -- E=20000
(
E
E
+ E=22000
'z 9 ------
(/)
(/)
w
(I
r-
(f)
3 --------·--····--····-
60 100 140 180 220 260

LAMBDA 1 - LAMBDA 2
Figure E.S:Third component of critical stress [N/mm1].
179

FOURTH COMPONENT OF CRITICAL LOAD
40
+ E1-E2=1000
"* E1- E2=2000
-& E1-E2=4000
30 ·····································*·-E-1-E2=fi00&·······--··
~ E1-E2=8000
(\J
{
E
E
'z 20
(/)
(/)
w
a:
l-
Ul
10
60 100 140 180 220 260

LAMBDA 1 - LAMBDA 2
Figure E.6:Fourth component of critical stress [N/mmz:!.
180
Fundamentals
E.4- Bockling Experimental Analysis of Bamboo Columns

E.4.1-Tbeoretfcal basis
The value in eq.E.36 can be regarded as
eq.E.(38)
Each term represents the different contributions to the boelding load of the ideal strut
In this way eq.E.32 can be written in a simpter way as
w a w
--=- eq.E.(39)
This is the equation of aso-called" Southweli Plot ofDefonnations", as shown in figure E.7.
Figure E.7:Southwell Plot of defonnations.
So it is possible to run an experiment on an initially crooked strut and get rid of the influence
181
of initial deformations, by ta.king the inverse of the slope of the line shown in figure E. 7. In
the case of bamboo culms this may be a powerfut instrument, because crookedness is more
the normal case than the exception. In the following lines an experimental set-up for the
determination of force deformation relations in the case of bamboo culms is presented.
E.4.2-Experiment set-up
Due to the fact that both rotations and displacements are expected to occur during a bockling
test, it is not possible to read lateral displacements by simply attaching a device to a point on
the culm skin. This is because the movement of the point of contact between the device and
the culm will be affected by both the lateral movement of the piece and whatever rotation of
its section has.
In the following lines a test set-up is proposed. 1t takes into account the above mentioned
facts and therefore allows us to make a Southweli Plot for bamboo columns.
The general lay-out is shown in figure E.8. lt is basically an arrangement of four rotating
potentiometers, each attached to the extreme of a 6 mm diameter needle, which passes
through the centre of the culm. The attachment points between the ropes and the needie are
set equally apart from the centre of the culm. When the needie moves due to any son of
movement of the specimen, it makes the potentiometers rotate thus allowing for registrations
of the angle of rotation of each potentiometer, rogether with the force producing the change,
the longitudinal deformation of the culm and whatever other variabie of interest.
From geometrie relations as shown in figure E.8, it follows that
eq.E.(40)
Therefore, the value of a is
182
2 2 d2
r2 -rl + 13 eq.E.(41)
a=---,--
2di2
With this value it is now possible to calculate the coordinates of point number 1 on the needie
as
eq.E.(42)
and
eq.E.(43)
From the triangle on the left it follows that
eq.E.(44)
which after simplification becomes
2 2 2
r3-r4 +~ eq.E.(4S)
b = -2d:w--
and this leads to the calculation of the coordinates of point 2 of the needie as
eq.E.(46)
and
eq.E.(47)
If the coordinates of the end-points on the needie are known, then it is easy to calculate the
coordinates of the centre. They are
183
eq.E.(48)
and
eq.E.(49)
· The initia! angle between the needie and the horizontal is
eq.E.(50)
The lateral absolute displacement of the middle of the culm at any time is
eq.E.(51)
po-t: 2
I
.
0.
C:j •
.._",_..._ - - ---' $lj
o.-~"a
"i
pot 4
Figure E.S:Cross-section of the experimentallay-out.
184
and the total rotation of the cross-section is
eq.H.(52)
E.4.2.1-Aiternative procedure
The experimental lay-out of the previous section may present some limitations or may be
found too cumbersome, and therefore it is convenient to have a second option. In order to
find out whether this alternative does exist, it is frrst necessary to return to the basic
tormulation of the Southweli Plot
E.4.2.1.1-Extension of the Southweli Plot to the Plot of Strains

If the corresponding value of the coordinates is substituted in the equation of deformations
then the result is
eq.E.(53)
If the initia! deformation of the bar can also be described by the series
w =a SUl1t-
. x eq.E.(54)
t 1 l
then the lateral deformation can be expressed as
al . x
w= --SUl1t-
N% l eq.E.(SS)
1--
Ncr
%
and thus the maximum lateral deformation at half the length of the culm is
185
Fundamentals
Appendix E: Bamboo colunms
eq.E.(56)
The bending moment associated with this deformation is therefore
eq.E.(S7)
The maximum stress somewhere in the periphery of the culm at this level can easily be
calculated by superposition of the bending and the compressive effects. This stress is
N
0 =2
"' A eq.E.(SS)
The maximum strain corresponding to eq.E.58 is
E
N"'
= --+--
Np1 r"
"' AE N IE eq.E.(S9)
1--"'
Ncr
;:c
Eq.E.59 can be rearranged to get
N"
E"-E
-A- = _1_ [E N:s ]
_ + a,r, eq.E.(60)
N:s N:!
:s " A E EI
a1r,
It should be noted that eq.E.60 is the equation of a straight line with intersection at
El
186
---~
Nz
e--
1 , in a plot of " EI against e - -
N"- as indicated in figure E.9.
and slope
N: N" "A E
e - Nx
x AE
Figure E.9: Southweli plot of strains.
E.4.2.1.2 Experimental lay-out

The point of maximum strains can hardly be anticipated dwing a buckling test on a strut of
bisymmetric cross-section. In the past some readings were approximated during bending tests,
by using a large number of strain gauges located longitudinally around the periphery of the
culm, si that the maximum strain could be guessed by observation after the test, and then
some sort of interpolation could take place.
In order to produce a continuous plot of eq.E.60, it is necessary to find a simpler and more
187
Fundamentals
precise way of calculating the position of the point of maximum strains and of calculating this
value too.
·.f.,>
c-+
e (+) ci:-
-
.....=-::---s::_tro.ln
go.uge <:'e
vx"'{
stro.ln go.uge
e2
go. u ge
Figure E.lO: Three-point strain gauge Iay-out.
Figure E.lO shows an option to the solution of this problem. Three strain gauges are attached
to the prepared surface of the culm. If great precision is to be achieved, then different sets
can be glued at different levels of the culm near the point of maximum initia! lateral
deformation.
In the figure, the position ofthe y axis on which the points of maximum and minimum strain
lay has been guessed.
It is a matter of simple geometrie relations that, under the conditions shown, the value of the
maximum strain can be calculated as
188
eq.E.(61)
in this expression
ç = tan-1[ asin30 ] eq.E.(62)

r.-acos30
and
a = 2 r. cos30
eq.E.(63)
E
1+-3-
e(+)
The value of e( +) in the equations is chosen from any positive value of the readings of
strains, and e3 is one of the remaining different sign strains (or the different sign strain if
there is just one).
E.S-Experimental results
A sample of bamboo canes was selected from a larger random sample, so as to have a good
description of the possible varlation in the value of l . Following the procedure of the
above paragraphs, a Southweli Plot was generated as can be seen in tigure E.ll.
The value of the combined criticalload N: was determined in each case by calculating
a linear regression model for each one of the graphs, after which results were compared with
those predicted by the theory.
189
Fundamentals
The two sets of results were statistically compared, resulting in a correlation of r 2 = .81
between predicted and measured critical load. As can be seen the amount of agreement
between the theory and the experimental results is remarkably high, and therefore gives a
good indication of the potentlal of the proposed model to explain the relation between
slendemess and buckling load of bamboo culms.
SOLrrHWELL PLOT SOUTHWELL PLOT

c>< 1e-e) SPECIMEN 8UC.006 CX 1E-5) SPECIMEN BUC.OOS
es
7!5
a
~:5 ~
~ ...
I
u
::Jss
~
u
<
--'
c..
VJ D:i
Cl Cl
45
35~==~==~==~~~==~ Ob:==============~===d
0 20 40 50 80 100 0 30 150 so 120 1:50
LATERAL DISPLACEMENT mm LATERAL DISPLACEMENT mm
Figure E.ll:Resulting Southweli plots of deformations.
A final experimental result of interest is the influence of twisting in the phenomena of

buckling . It was already said that this effect would be of no interest if the amount of rotation
of the transversal section was limited enough, so that the assumptions made in the frrst
sections of the appendix hold true.
Figure E.12 illustrates the relation between twisting of the cross section and axialload. There,
it can be seen that the amount of rotation registered during this test was very little, in spite
of the fact that the culm was completely free to rotate at the ends. This fact lends support to
190
---------------------
the decision not to take the twisting energy into account in the general equation.
AXIAL LCY.O VS TWISTING AXIAL LOAO VS TWISTING

ex '100) SPECIMEN 8VC.004 CX '1000) SPECIMEN BUC.OOB
'11löl 40~==~~==~~~~~~
z z
D~==~====~==~====d
o 0.2 o.4 o.e o.e -0.4 -0.2 0 0.2 0.4
ANGLE OF TWISTING degreea ANGLE OF TWIST I NG degreas
Figure E.l2:Twisting of the cross-section.
E.6-Calculation of the critical load as a function of average properties

An alternative way of calculating the critical load of bamboo culms can be proposed if the
simplifications suggested in appendix G are followed. If, besides that,. the elastic modulus and
the moment of area are written as a function of the average values such that
where
191
and
then the result of integrating eq.E.26, and further simplifying it as in eq.E.28, is that the
critical load can be calculated as
eq.E.(64)
In this equation the term in brackets represents a 'correction factor'. In other words, a factor
by which the average critical load bas to be corrected, to account for the varlation in
geometrie and mechanical properties.
A sample of Guadua s.p. was taken to run a parametrie study using this last equation, and see
the effect of combinations of some real values on the criticalload.
Figure E.13 shows the results of such study. In the ftgure the valnes of the right term of
eq.E.(64) are plotted against the product a1 6 8 for each specimen in the sample.
As can be observed, for this partienlar batch, the calculation of the criticalload using average
properties gives conservative figures for most of the specimens, but this is not a general rule,
since, as can be seen in the plot, there are culms for which the average valnes give
underestimations of the critical load.
Eq.E.(64) can be used for the study of partienlar batches of bamboo, so that, with the right
input, the range of valnes of the factor of figure E.13 can be estimated, and proper design
192
factors can be agreed upon and recommended.
BUCKL liG OF BAM300 CULMS

CORRECTION FACTOR FOR GUADUA S.P.
1. 2 -- --- --- -- •••••• -- -- --- •• -- •••••• --
1.1 ·························:~································
1: ·~
+
.a.+ +
·11'~
1
....
rr·+·······························------------·· ........ .
·~
o. 9 r-·······················································---
0. 8 •• --- ••• •• ••• ••• •• • •••••••••••••••••••••••

0 0.02 0.04 0.0!1 o.oe
Figure E.13:Result of the estimation of the correction factor for Guadua s.p.
REFERENCES:
1 Gregory, M.;l967. Elastic Instability: Analysis of Buckling Modes and Loads of
Framed Structures. E.& F.N. Spon Itd., London.
193
Appendix F: Structural mndelling and analysis of bamboo strui:tures
Appendix F: Strnctural modelling and analysis of bamboo strnctnres

F.l-Introduction
In thls appendix the influence of bamboo's geometrie and mechanical properties on the
deformations of structures is evaluated.
Bamboo culms are tapered, the · cross-section deercases from bottorn to top, the elastic
modulus increases from bottorn to top, and they have nodes with complîcated geometrical and
mechanical properties, different from those of the neighbouring intemodes. The spacing of
these nodes is not a constant feature, it varles along the axis of the element and the pattem
of variations is different in samples from the same origin.
The differentiation between node and internode properties bas been explained by several
researchers, for example, Grosserand Liese (Grosser, D.; Liese, W.; 1971; Liese; W.; 1985).
These and other scientists have found significant anatomical differences between the nodes
and the internodes. However, it was not possible to find any quantification of the effect of
these differences in the literature, on the overall structural properties of the culms. As
indicated in chapter 2, the nodes constitute the weakest link in bamboo culms.
In the next sections the intention is to explain and to quantify, on a theoretica! basis, the
essential role that conicity and nodes play in the structural performance of bamboo culms. The
foliowed methodology is summarized in figure F.l. As seen there, the stiffness matrixfora
tapered element with varying modulus of elasticity is calculated first of all, adding to it the
effect of nodes so thàt bamboo can be studied. Parameters that show any relevancy in the
definition of the matrix, are evaluated by studying real-ease culms in a sensitivity analysis.
This approach is chosen because there can be combinations of the parameters that just do not
exist in real cases, therefore, a simpte mathematical analysis may lead to unreal situations.
Secondly, the stiffness of some sample elements is calculated following two different
approaches, ·using the generated matrix, and using average properties only (i.e. average section
194
Appendix F: Structural modeUing and analysis of bamboo structures
..
THBORBTICAL MODEL
Bquations to calcu1ate forees and displaceorms
..
Measurem.tm of parameters for the selected sample - ·
..
CtlcuJation of sliffness matrix
COMPARISON OF STIPF.NESS VALUES AND

DERIVATION OF ASIMPLIFIBD MBTHOD
t
<ilcuJation of av~~f'I.S for the sample -
Figure F.l: Method
195
Appendix F: Structural modelling and analysis of bamboo struètures
and average elastic modulus1).

Thirdly, the two sets of stiffnesses are compared to derive a simplified approach for the
calculation of bamboo trusses and frames (and therefore to avoid the use of complicated
equations).
From a methodological point of view, conicity and variation of the elastic modulus, on the
one hand, and node influence, on the other, are treated in a separate way, as will beseen later
in the chapter. The validity of this procedure is of course subject to doubt in the first instance,
but on the pragmatic side, it simplifies the problem. In this way, a theoretica! model is
developed Orst to explain the combined effect of conicity and varlation of the elasticity
modulus. The model is then used to undertake a parametrie and a sensitivity study. A second
model is developed based on the results, to take account of the influence of nodes on the
structural properties of bamboo culms.
From the perspective of the behaviour of bamboo structures, comparlsons between three
different structural types are presented in the last sections of the chapter, in accordance with
laboratory observations of their performances, with the purpose of giving input for the
decision about the best structural options for bamboo constructions.
F.2·Theoretical model
F.2.1-The effect of conicity
Several assumptions are made in order to set up a theoretica! model.
-.The material is elastic.
-.Elements are perfectly straight.
-.The conicity can bedescribed as in eq.F.(l-3). 2
-.The varlation of the elastic modulus can be described as in eq.F.(4)
1
average properties are calculated by taking the average of the valnes at the extremesof
the culm.
2
see appendix G for more details on this matter.It is also possible to express the varlation
in terms of the average values, but it can be proved that that way does not contribute to the
simplification of the final equations.
196
Appendix F: Structural modelling and analysis of bamboo structures
• (x) = •
• ~A
(1-.,•l x) eq.F.(l)
eq.F.(2)
eq.F.(3)
eq.F.(4)
Figure F.2 Degrees of freedom of the element
The element stiffness depends upon a combination of its elasticity, Poisson ratio and
geometry. The stiffness matrix of prismatic elements is readily available in many books (for
197
Appendix F: Structurall'tWdelling and analysis of bamhoo strui:tures
. ---'
Figure F.3: Relations between displacements and rotations.
instance, Norris,Ch.; Wilbur, J.B.; 1960; Willems, N.; Lucas, W.M.; 1978) and it permits the
full analysis of such elements and of structures made of those elements.
In what follows details of the calculation of the stiffness matrix for non prismatic elements
are shown, according to the assumptions depicted by equations F.l to F.4.
198
Soppose that the beam-column of figure F.2 is displaced and rotated at the extremes as
shown.
Following the principles of the Reduced Area Moment metbod the equations of displacements
and rotations for the beam-column can be set as
UA =U8 - [ti-j 0
M
E(x)l(X)
(l-x)d:x] eq.F.(5)
or
I
Mx
UA = Us-+J-f0 E(x)J(x) dx eq.F.(6)
eq.F.(7)
or
eq.F.(8)
and the rotations are related such that
+A -- c!»s+J E(x)l(x)
M dx eq.F.(9)
0
•· = ... ~ IE(Z(x)dx eq.F.(lO)
199
If the element is allowed to displace or to rotate at the extremes, in such a way that only one
possible displacement or rotation is permitted at a time, then four cases can be distinguished
as explained in the following lines.
F.l.l.l-Case one
Figure F.4: Equilibrium conditions for case one.
The state of displacements and rotations is described by
UA = "z = 1
"• =u, = 0
+.t = u3 = 0 eq.F.(ll)
+s = u6 = 0
u1 = 0; u4 = 0
From the equilibrium of the beam-column
200
eq.F.(12)
eq.F.(13)
eq.F.(14)
F.2.1.2-Case two
Figure F.S: Equilibrium of the beam column for case two.
Displacements and rotations are described by
UA "' ":z = 0
"• =u, = 0
4».. = "3 = 1 eq.F.(IS)
4». = "6 = 0
u1 = 0; u4 = 0
201
Fundamentals of the Design. of Bamboo Structures
eq.F.(l6)
eq.F.(l7)
eq.F.(l8)
F.2.1.3-Case three
Figure F.6: Equilibrium of the beam. column for case three.
The state of displacements and rotatîons is described by
202
· Fundamentals of the Design of Bamboo Structures
Appendix F: Structural nwdelling and analysis of bamboo structures
UA = Uz = 0
u8 = u5 = 1
.Á = 143 = 0 eq.F.(19)
•• = "6 = 0
"t = 0; "4 = 0
eq.F.(20)
eq.F.(21)
eq.F.(22)
F.2.1.4-Case four
Figure F.7: Equilibrium of the beam column for case four.
Displacements and rotations are described by
203
UA"'~= 0
"• =u, = 0
t ... = u3 = 0 eq.F.(23)
t 8 = u6 = 1
u1 = 0; u4 = 0
From the equilibrium of the beam-columnn
eq.F.(24)
eq.F.(25)
eq.F.(26)
F.2.1.5-Solution of the eqnations of displacement and rotation

Substitution of the values of eq.F.ll into eq.F.5 to F.lO gives
eq.F.(27)
and
eq.F.(28)
For the sake of simplification the following notation will be kept throughout :
204
eq.F.(29)
eq.F.(30)
eq.F.(31)
eq.F.(32)
If the values of equations F.l5 to F.26 are substituted one case at a time into equations F.5
to F.lO then it is possible to write the following expressions:
From case one,
eq.F.(33)
eq.F.(34)
From case two,
eq.F.(35)
eq.F.(36)
From case three,
eq.F.(37)
205
Appendix F: Structural modelling and analysis of bamboo struetures
eq.F.(38)
From case four
eq.F.(39)
eq.F.(40)
The solution of the above equations can be expressed in terms of K32 and of the following
relations
z1 =F.Ft- eq.F.(41)
2
F.
z.2 = ___!
F.
eq.F.(42)
2
to fmd that
eq.F.(43)
eq.F.(44)
eq.F.(45)
eq.F.(46)
206
Kn=Klz eq.F.(47)
K,, = Z2Klz eq.F.(48)
Kg= -Kn eq.F.(49)
R$ =(l-Z1) K,a eq.F.(SO)
x"= -K,a eq.F.(Sl)
&2, =-zt KD eq.F.(52)
K" = zt KD eq.F.(53)
~ = (1-Z1 ~K32 eq.F.(54)
KM= (1-~) KD eq.F.(SS)
Ki. = (lZt-1) KD eq.F.(56)
KM= (1-Zt l) KD eq.F.(57)
K66 = (Zz +l2zt -21) x,2 eq.F.(58)
207
Integration of equations F.30, F.31, and F.32 allows to state that:
eq.F.(S9)
F. = In
2
(1 +aE' Xl-&/) ,z eq.F.(60)
(li/+l'JE1 ) E,IA
eq.F.(61)
F. = _1_
' aE'a'I
The members K11 and K44 of the stiffness matrix can be found taking
eq.R(62)
and all the other displacements and rotations equal to zero. Then
eq.F.(63)
and therefore
~ I +~ I
B A
.J 1-~ IJ eq.F.(64)
lt+&J
Equations F.43 to F.58 and F.64 describe the members of the stiffness matrix of a tapered
beam-columnn element with linearly varying elastic modulus. It is possible to show that, for
example,
208
eq.F.(65)
and
[&'']
HmA-OK=~
EA
&
B
I 1l l
eq.F.(66)
which are tbe corresponding merobers of tbe stiffness matrix of a prismatic beam-column witb
constant modulus of elasticity.
F.3-Parametric study
The major difficulty in tbe use of tbe stiffness matrix, to see tbe effect of changes in tbe
various parameters, rises from the fact tbat valuable data to facilitate a practical exploration
is not available. Data do exist, of course, but tbey have been gatbered and produced under a
different focus of attention.
Basically, tbe metbod followed consisted of changing tbe values of the different parameters
involved, so tbat a more complete picture of tbeir effect could be produced.
To set an adequate reference to tbe valnes tbat tbe different parameters may take, a sample
out of a larger sample of 120 culms of Guadua s.p. from Costa Rica was studied.
A total of 24 pieces of different lengtbs were taken. The sample was not completely
randomized, because part of tbe current stock at Eindhoven University has been attacked by
borers, so all tbe damaged culms were regarded as unsuitable. Every one of tbe culms was
closely examined and its relevant properties determined, namely tbe geometrie and mechanica!
characteristics as described in previous chapters.
In tbe search fora way to easily carry out a parametrie study, it was of initia! interest to
know whetber tbere was a way of relating tbe stiffness calculated using tbe proposed matrix,
and tbat calculated taking only average properties for every culm into account, because tbe
formulae suggest tbere should be some straightforward relation between tbem. Should tbe
209
formulae be close to reality, that type of relation would then come forward in an obvious way
after the examination of a sample. Results are encouraging, as can be seen in the following
paragraphs.
Some basic information about element properties is given to put results in the right
perspective.
statistic extemal diam. E (10-80%)

average 83.01 18710.0
varianee 25.32 4724710.0
standard dev. 5.03 2173.52
standard error .726 443.67
minimum 71.82 14930.0
.maximum 97.68 22235.0
Table F.l : summary of element properties. Diameter [mm].

Average elasticity modulus [Nintnf].
Axial deformation stiffness was exarnined first.

The values of K11 for each one of the elements were calculated according to equation F.64,
as well as the corresponding average value. The results can be better visualized in a plot like
the one shown in figure F.8.For this sample it was found that the ratio between axial
stiffnesses according to the proposed formulae and that calculated based on average properties
is .97. Thus, for the examined sample it looks as if the error in the estimation of the axial
stiffness is an average of approximately -3% in relation to the axial stiffness basedon average
properties.
(For the sake of clarity, in what follows, stiffness calculated according to the proposed matrix,
will be referred to as 'theoretica! stiffness', though it is ack:nowledged that the average one
is also theoretica!. Average properties are calculated by taking the valnes at the extremes
A and B of the culm and averaging them).
The same type of comparison can be made for other factors in the matrix. If for example K22
is considered then the result is like that shown in figure F.9.
210
K11 av VRS K11 thaory

(X 10000) BAMBUSA GUAOUA
10 .. ... ·-·· ··-· -··· •.. ··- -· ... ··- ··-· •.... ... ..
8 ··---·-··-··-···--···········-····-····-·--···-··-··-·············-··-;··-·----·-··-
B ···-·-··-··-·---·-·········-···············-···-·····;···:---···-········-···-······--·-·---
~
w
;:
2 ·····-···-···-···········... ···-···-····-··-··-····-···-·············-··-··---·-···-··-···-·
0 .....
0 2
"' 6
a 10
THEORY ex 1oooo::~
Figure F.8 Average and theoretical axial stiffness compared
Next, the most relevant elements of the stiffness matrix were compared to those from average
values for all the elements in the sample, so that a more general picture could be gained.
Figure F.lO shows the ratio between theoretical values and average ones, on the y axis and
the identification label number of the element is located on the x axis.
There it cao be seen that, at least for this sample, the range of varlation of the coefficients
for different stiffness elements is within -.92 and + 1.05 times the average value. The plots
show that in general, values tend to fall below 1.0, meaning that the average approach tends
to overestimate stiffness.In order to broaden the picture on these effects, and to see how
general this .conclusion cao be, a limited sensitivity analysis was carried out
6'1
Culms were studied in such a way that the ratio changed while the remaining factors
a•.'
211
Appendix F: Structural nwdelling and analysis of bamboa struàures
were kept constant. The results of the calculations of the ratlos for different stiffness merobers
are given in figures F.ll to F.l4.
The studies of the author on samples of Bambusa Blumeana from the Philippines and Guadua
s.p. from Costa Rica allow us to conetude that in these two cases the conicity rates spread
over a wide range of values (though the conicity itself is small). More refmed anatomical
studies show that conicity rates change along the length. Therefore, equations F.l to F.4 are
only a simplification, of course, but as indicated in appendix G they are precise enough for
structural calculations, as long as the length of the element is bellow say one third of the
total lengtb of the original culm , which seems to be the case in structural applications.
The figures show that in most of the cases the error, in the estimation of the stiffness by
taking average properties, is about 15% and that average values overeslimate tbe sliffness
of bamboo elements.
Further calculations were undertaken in order to show the individual effects of different
combinations of parameters, as shown in figures F.l5·to F.32.
Hopefully this analysis covers all the extreme possibilities that can be found in structural
applications. of bamboo, though it is acknowledged that further data should be gathered to be
conclusive about on these results. Nevertheless, some observations can be made.
6'1
-The sample of Guadua s.p. showed that the ratio has a large variation, in the
&BI
range from 60 to about 400, with some concentration of valnes in !}le region of 350.
According to the plots, it is possible to have overestimations as large as 25% on the bending
stiffnesss, though a typical figure would tend to be closer to the 5% mentioned above, for all
ranges of initial diameters and initia! elasticity moduli examined in this study.
-So, for most of the cases of structural interest, it probably does not pay to go into any
refinement in taking the influence of the variation of the diameter and the elasticity modulus
in the bending stiffnesss of the culms into account. In relation to the axial stiffness, this
212
characteristic seems less sensitive to the value of
K22 av VRS K22 theory

ex 1ooo:> BAMBUSA GUAOUA
6~-~--~---~·-=-~~==->--=r-·=··=·~·-~·-~·-=·~-~-~-==~-=-=-~-~·=···~···c···=··~
5 ··-··-··-·-····-·-·-..- · - - · - -.. -···-·············-···-··--·· ..--·-·········-
" ··-··-··--------··-···-···---·····-···-·-----··-···-···-········-·-···-···-······
~ g !-······-···-··-··-----··-··--··-·····:-··-··-·-·---··-··-···-······-···-
~
2 ...................................:~·-·--·-··-·--···-······-···-···--·-·-·-··--···-·-··-·
1 !-··-··-·---;·-····-···-········-··········-··-···-·---··-···-·-···-·--·-·--·--····
0
/
................ .
0 2 3 " 5 6
THEORY ex 1ooo)
Figure F.9 Average and theoretica! bending stiffness compared.
213
Appendix F: Structural modelüng and analysis of bamboa struètures
THEOAY AND AVERAGE COMPARED

~LE BtZE ~B GUAOUA
"f,"' ••• ~.-.. ••• •• h• •••-•• • - • on on ••• •~-·~-·-·---
"1 , OB •-•-••~•••-••---•·•-•••-•••--•••-•••••••••••••••-•••••n••-••-•-••••-••••••••••••••••••••••-••n••••-•--
0.9 ......................
0 10 20 30 40 50
t!I..E:MENT LABEL
Figure F.lO: Ratio of theoretica! stiffness to average ones for different stiffness elements.
-CM!
t.1. --- -·· ...... -···· ·-··· -·· - · -· ............ ···-· ·-··· ·-·
0.7
--
---·--···-··-·····-······-···-···-·-·-··-·-·-·-·-····""'··.IIIIIL-·-··--··-···-·
Figure F.ll Case 1: l= 4500; èf>;=llO mm; E;=14000 N/mm2•
214
t.18 .. ···- ·-··· ···-· ·-·· -··· ···- ·-·· ··- ····- ··-- - - --·
'·' --·-··-·-·-··-···-····-··-··-·-···--······-···-··-···-·-··-··-···-··-··-·-··-
0.711
113
~~0=======.
~======~
. .==~==~·--=·-~--~·=·-=··=···-==····~
. ..
----
-·--·-·-·····-····-·····················-······························································-···
--······-··-···-···---·-·····--···----···-·····-···-···-·-··-·-···········-·······
3'
-=.1.
3'
B
Figure F.12 Case 2: 1=3000 mm; Cll1= 82 mm; E1=22034 Nlmm!.
~.~. !~'==·=····=·
1.1
--
··=···=··=···""···=····""·=····""··=··-""·=-·""···""···=···=·=·-=··-=·===-'=·=·-·=·=··-=·=···-"'-"
..•... ············································································-·························-···
~
0.?
0
·-··--·· ·········-·
.. .. ---
··-···-···---··-···-·····-···-·····-······--···-··-···-···-·---·-·-··-···-······
...
- 3'
-=.1.
3'
B
Figure F.13 Case 3: 1=2000 mm; Cll1 110 mm; E1=10000 Nlmm!.
215
1.18 -··· ···-· -··· . --
----
0."111 ····-··-··-····-···-·····-····-···-······-·-······-····-..·--..·-··-......._ ........_ .._,
..·~~=··=·==-=·=··-=·=-~=-·~····=·=···=-=·-=.. =-=-·~=-~·-·=·=·-·=·-=·..=·=·..~~
~~0==·=·"=""="·=·=
Figure F.l4 Case 4: 1=3000,· <1>1=82 mm; E1=15000 N!mni.
-MT!Da-INI'IW._"._,_
1.18 ................. ·-.. ·-·· ........... -·..............
1.1 ··-····-··--·--.. - .... -_.................................. - ........................................ .
0
·- -............. ,_.. _,..
-
- ---- - 3'
_::x..
6'
B
Figure F.lS: Correction factors for average stiffness, conditions as indicated.
216
1.1a
1.1
--
-ftAnoltPOI'IINIIIIIL--
-==========<=-=·""··=···=··-=··=-"'F"··-=·=··=···=·=""'
·····--·--·-··--··--·-··-····-----······-··-··---···-······-·
--
~~o=·=·~··-~·~··-=··=--~-~~-=···=·=···-~·~·-=-=··~~~-=··====~~~~======~~·
&'
-=t.
&'B
Figure F.16: Correction factors for average stiffness, conditions as indicated.
1.1
--
-MnoltPOI'IIIII"'W.--
----- -------
-·--··-·····--·---·····-··-·--·····-·······-··--···-··-···-·-··---
I u
o:rs ·······-···--·--·········--·-····-··············--····-····-=·..Ji:!l$1111111..........- ..
03
--
- 11!-1-
·--···-········--····-·······-···-----·····-···-·-··--·-··-······
=·-=-=·-=··=-·~ïiao!::"ö"····~-=·-=-·=-=····~~==·=-=··=·=··-=·=·--=·=·-=-=·=-··=·=···=·~
~ '!'o=····""· ....
217
-l'limOII--~-
u• ··- --·· ...... -·· ...... ...... ..... ..... ·-·- -- ...... ...... ...... ..... .....
1.1 ···············-····-·-···--····-··-···-·················--··--···--····-···-···--·············
0.7
---
--··--·----··-······-···-·····-···--··--···--···-··-··--·-···-·-·
~~o~====~.u======···=···=·-~~~-·-=··=-=-··==-·=··=··Ma~·-=···~···=···=···=···~·=·-~-
6'
_::J_
6'
B
Figure F.l8: Correction factors for average stiffness, conditions as indicated.
u• -- --· -··· . . . . . . ...... ·--- . . . ...... . . . ..... ............ ·---- ···-· -·-·
1.1 ··-····-···-···········-·····-····--··············-------------·-··········-····-········-···-··
.. ·······-··-----------------------·····-·········-···---=--Jii!!UIIIIIIL._______ _
0.7 -------·-··········--··----·--·-----------···------·-···-··
--
~~o~=====·-~.o~-~--~-=-=··-=··=··=---~~~-=-=-=-=··=-·==---~~~--=--·=---=·-=·-=·-~-=--*»~·
6'
--=.L
6'
B
Figure F.l9: Correction factors for average stiffness, conditions as indicated.
218
..,. ro===="""'==·-=·""···""-"'···=···F-··=-""·-=··==--=·"'·""-""··'-"-·"'·=···=·=----
1.1 ·----·-·--·--·--·-·---------·--···········-··-··-·--
---
o.e -·····-·-----·-··--··--·-··--··-·················-··------···-··-·········
G.7& ···-··-···--·-·-··--·-········--··-··--·--·--=-II'!!!JIIIAQ.............
----·-··-·-········-------·-·---·--·-···---3
G.7
.............. ~
~~o==~==~.a~==~==~._~==~~~aD========~.a
t.1& -··· ·-··· •• - · - - ·-·· •••••• ····- ·-·· • ···- --·
1.1 -------------··-···········------·-·--------···-···········-·-···-
1.GII
_
..,.._ .._______
-- ----·-·-·----···-···-··-···-······-····-·-·-=-Ji,!IJIIIAA.
G.7
11.8
0
··- .....................
.a .. --
-
--···-···-·-·-·-···-··-···-···--·------··-··---···-·-··-·
-
Aj
a•B
219
t.ta ..... --·· ..... ---· ..... --
- -
. ........................-··-·---..-.........................=...11!!11111111-..- .....
- -- -
o.7
.....
· - · - - · - · -..- · - - - - - · - · - - - - - · - - -..- - · - · - - - - ·
a•
....:.L
0 .. -
a•B
-MTIOIII'ORINI!W.-
ua ~!='==="""""-=··=-=--~·-·=-·=-=·""-=·-=·=-==-=
..".....=..·=··=···=··=..·=-~..."",.
..
,..
1.1 ..................................... - ........................... ____...........-·-·--·····-··
................:.:ii"""";;3"""~'~---"""'·
~
...~.. ........._ ..........
; - f/
-
o.•
j
OA
0.711
I _,__
_ .................- ..................._ .. _,............- ...........=-~...............
- - - --
G.7
- -
0
·-···-···-·······-~···--·-·--·-·-·-~··-·-·······......_....__

...- ..-...
220
Appendix F: Structural rnadelling and analysis of bamboa structures
M. - - - - - - - - - - - - - - - - - - - - - - - - -
1.1
- .....
·····-···-·-·············--·-·-···-·-·-··························-···································
1.1111 ··········································-··················-------·····-·····--···---······-····
u: ~:~::. .: : :_· · -· · · · · :· · · · · · · · · · · · · · · ·-· · · · · · · · ·: : : : : : :

u ······················-·-·······················-----·-···-················-···························
G.M ··············································································-·-·······-···-····-···········
--
u ···············································································································-
·-··-·············--····--············································=·-~--··············
---
Q."IW
0.7 ......
···············-········-----·-············----------·--·-·····--·----...........,.·-------------
o.a~o·=··=···=·=··=···=···~~~~··=··=··=···=·=···~-~~--~·=···=··=··=···=··=öä~···=·=·=···=·=-~-ÄA~-·~·=···=··=·=···=·~~
Figure F.24: Correction factors for average stiffness, conditions as indicated .
.._ .....
M. - - - - - - - - - - - - - - - - - - - - - - - - -
1.1 ····-·····-············-··-·-······-·······································································
1.1111 ·································································-··-···········································
G.M ···································································--······-·············-·-··············-·
--
u······································································································-··········
··-················-·-···········-········································=··-···············
---
Q."IW
0.7 ......
··················-----------------······------···-··············......·· ..........................
o.a~·=··=···=·=··=···=··7~7~··=··=··=···=·=··~··a.~···~·=·-=-=··=···=·7a.~···=·-==···=··=··~··DA~-~-=····=··=··=···=··~~
0
Figure F.2S: Correction factors for average stiffness, conditions as indicated.
221
Appendix F: Structural modelling and analysis of bamboo struttures
ua FF=="""'"""'===-~.......rn --
1Ct1M'IIOIII'OIIINIIW.--
•• =···""··=··=···""··"'·=~=-====-!!
,..
1.1 ·······················-·····-···-·-···--···-··--····-·-······-··-··-·--·--·--···-··-
---··-···-···-·-··-·-·--·-··-··-····-··--··················-·--·--·······
-
o.a --·--·-··-····-·····-···-····-····--..·-···-····-········-···-····-··-·-···-·-·
0.711
·-····-·----····-·················-···--··-·-····-····-··-·-···--············-···--
·············-··-·-·········-···························-····-·-~·=··············"j
j
:: c:::::::::~:::::::::~~:~~~=::::::::::::::::~:~;:;:.::~=~J
0 ~ u u u u
ua
,..
1.1
-· . .... ... ... .... ... --
... - .. ... -· . ............
··-·-···-··-····-····-···-····-····-··--··--···-···-·-·····-·-··-···-·
···-··-···--·-·--·-··-·--··--···--·-·---··---···-···-···-···-········
I a..:- =~--~···--·=-·-·--·--~--=····-~------·-·-···-·~===:
··---·--···-·····-···-·····--·-···-··-···-··-·-···-········-···-··-·-··-
o.a ·····-··-·-----···-·--··········-···-····-··--·---·--··-·····················
---
u -··--····---·--···--·-·····-····-······················-···-····--·-···-··..···········
0.711 --···-···-···············-····-···-·-·--··--··················=·~---·--
......
.....
o.a~o====~a.~,~==·=·-~··Ö.~-·~··=···~··=-~··~·--~-~-=·~-=·-~··u~·~···~···~·~···=··=-~-.
222
ua 11""'==='<'=-=--=--=-=·=-·=·=··=··=·=··=·"<·-=···=··=··=··=·=·.".
•• -=··=·-=·=·-=·-="A··
1.1 ·-··-·-·-·-··-··-·-----------------·--·-··-··-···-----------····
1.011 --------------------------------------------·-···-···-····
c.-..
o:l
-------------------------------=---------
__________________________........,...._ _ _ _ _ _
----
1(11~----
--
1.1all""'==~===-·=---=r--=-=--=-=-=·=---=---~--=·=··=··"'·
1.1
."..=--=--=="""
--------·-··-···-···-······-···-··-····-··-········-··-··-··-·······-·······-···-··
1.GII ··-··-····-···-··-···-···------------------------··-···-·---------------····
~:ft-~===
- ·-·-·-··········-·--···-···--···-·-···-···----·-··-··-···-···········-··-······-···-···
u ·-···-···--···-···-···-·····-···-·-···-···-········-···--·-···-···-···-·-··-···-·-···-···
·-···-···--···-···-·--···--······-···--·-···-···-···-···-···=-.",--··············
--
Cl.'N
0.7 ··············-·············-··············--··--···-···-·····------·············
7
-~o===··=··=---~-~~~=-·=····=··=·=···=··n.~--~·-=··=··=··=···=·a.~·-~···=··=·=··=··=···~èA
-·=·=···~··==~u
223
Appendix F: Structural nwdelling and analysis of bamboa struétures
ua
1.1
1.cll
-·········-·······-········ -- --····-··-··
-··-··-······-···-···-···--···-····-···-···--·-····-··-··-··········-·-·-··-··
·-···-···-···--------····-··-~---········-·-·-··-·······-···-···-············-·
03 --
--
·······-····-···-·····-··--·····--····-··················-····--=-.II!!UIIIIIII...............
-···-····-········-·····--···-···-······-····-···-···-··········---·-··-··-··
~~o==··=·~··=·-~-~~-~~--~···~··=·=···=~~--~··=·==·=·~~--~···=··=·=··=····=··OA~-~·-=·-=·=··=-~o.
3'
--=.A
3'
E
1.1
--
ua ..... -· ··- -··· -- .................... ··- . ---· ............ -
-···-··················---·---··-·····-···················-···-··---·--··-···-···
1M ··--····-··············-·····-··-·-·--····-··-···················-······--·--·---···
u: -~~:-·····-~----·········:. .:. . .:. .-.~----··-·:. . . . . . . ~::~~:=~

o.e ···········-·····--···-··--·-···-····-························-····-·-·---···--·················
OM ··-···-·······-··-·····-···-----····-···········-··········-···-·---·--···-······
o.e ··-··-··-···--··········-···--··--··----·-····-···············-·-······-···-···-···-··-
0.7 ·-··-··----··--···-···-·······-----·-·------····----------·-- ---

II.'N . ··-·--··--···-················-····-··-·-·-····-··-····-=··.".,-·-·-·
--..
.. ~--·=·==--=--~4A7=--·=
o.a~o======~~,~,~---~·--==---~~~-~-=·--=-=··=--~·~~-~ ..
..=·--=--·=-~à. ~
3'
3'
E
224
Appendix F: Structural modelling and analysis of bamboa structures
1.1 ---------------------------------------------------······
O.lll ---------------------------------=--II!!1GIIII)________ _
0.7 -+- - -
----------------------------------------------
~~o====~~~;·=··=--=---~·-=-~-Öa7···=··=·==~ ---
. .~====u~====~.. a•
~
225
Appendix F: Structural11Wdelling and analysis of bamboo struètures
F.4-The effect of the nodes

As indicated in cbapter 2, tbe nodes of bamboo culms are mucb softer tban the intemodes,
so it is reasonable to expect that tbey decrease tbe general stiffness of the canes.
The amount of stiffness decrement depends upon tbe number of nodes and upon tbe stiffness
relation between internodes and nodes. As said before, the statements in chapter 2 in this
respect are based on the examination of a sample available in the University of Eindhoven,
but the description of anatomical features of the nodes by Grosser (opus cit) leads one to
believe tbat tbose remarks are widely valid. The specific relations for each species or group
of species will have to be worked out to be used for structural purposes.
On tbis basis nodes can be incorporated in tbe model of figure F.2.
Equation F.27 can be rewritten so tbat
l
eq.F.(67)
O=f
0
and equation F.28 now becomes
eq.F.(68)
Integration and some reorganization of terms permits us to write these equations as
eq.F.(69)
and
226
eq.F.(70)
where
eq.F.(71)
eq.F.(72)
eq.F.(73)
The solution of these equations is
K = B eq.F.(74)
32 Bz+CA
and
eq.F.(75)
where
1
A= EI (-l+a,.N (1-REI)) A REI = internode stilfness
node stiffness eq.F.(76)
1 1
227
Fundamentals of the Design ofBamboo Structures
Appendix F: Structural modelling and analysis of bamboo struétures
eq.F.(77)
eq.F.(78)
This solution has been written in terms of the average properties, so that comparisons can be
done in a more direct way.
One aspect that can be notleed after examination of the above equations is that, as common
sense dictates, the length of the nodes a,. plays an important role in the calculation, but the
determination of that tigure is a matter subject to great discussion, as has been already pointed
out in chapter 2. In what follows, the assumptions made in this respect in that chapter also
hold for the. sake of the present arguments.
The recalculation of the longitudinal stiffness can be done in an easy way as well. It is
possible to model the element as indicated in tigure F.33.
Using basic principles the sum of the stiffness of the different components along the cane can
be written as:
eq.F.(79)
or
A R = axial stiffness internode eq.F.(80)

E axial stijfnes node
228
Appendix F: Structural mndelling and analysis of bamboo structures
Figure F.33 Model for the calculation of axial stiffness
F.4.1-Parametric study on node influence on the stiffness.

Figures F.34 to F.39 show the results of a parametrie study on the influence of nodes on the
bending stiffness of bamboo elements.
As can be observed, nodes have a very important impact on the stiffness of culms, theireffect
averaging out to a reduction of 40% on the bending stiffness, and of 12% on the axial
stiffness, being the major souree of flexibility.
A small word has to be said about the node width indicated in the plots. The value of an
corresponds to the indications that are made in chapter 2, and therefore it is only a nomina!
figure. The fact that some of the curves are saw-lilce may be confusing, but it is only due to
the discrete character of the number of nodes.
229
NOCE INPLUENCE ON K3a

NODE WICTH- 30 Re-2.B
., •• ··~• ••- ..... • - ·-· •• .... •• -•• •••• -··· •• ~• ••o *"' ••• ··- ••- •••• """" -••
- I NTe!RNOCE!-30
0.9
...,.... t NTERNOOE-20
o.e
o.
. 0 ,
~ r-·······--........... ~-== .
-4l_. ---r-·v··.,---&--·r·...--- .........., ...,,,.. __,.....
-··=·---...-ï:;--·-...·:;------:.-.......-:...--....
·"t"'X"'""""""""'*"'1""X"""""<'"j
J
0 3 -4 5
L..ENGTH mm ex -.ooo::>
Figure F.34 Influence of node on bending stiffness.

Internode lengths in mm. Conditions as indicated.
NOOE INF~UENCe ON K32

NOOE WIOTH- 3C ~~-2.S
~ ~==m=~==m-~-=-~--=--·=·=--=--~--=--~--~--=-··=·==~=*-~--~--=·=--~· =--·=--=--·=·=··=--=-=-=R
-INTERNOCE•::iiJC
-+-
f NTERNOOE•2!5
o.a ~-----·-··-····-·············-··.,-------·-············-···········-···-···-···-····--~-------··-········
-11t- INTERNODE==20
o .a -----···-·-·-------------------·-··-----------··--·-----------·-·-·-·--·-----··----·----------·-
0
~ 0 .7 ..............................-··-···-······---·---·-·--·---·--·--·---····--·--·--------------·----·"""'""'"""''''
0.-4 ~-·=========-==·--=·=--·=--=--·=--=--c-=·=-=·=--=-=--=·=-~--=·=··=--~--·=·=·--=·=·-=+·=·-=--=··=--=-=--=--~-~
0 -4 5
L..ENGTH mm ex -.ooo:>

Internode lengths in mm. Conditions as indicated.
230
NOOE INF~UENCE ON K~1
., -.
NOOe WiCTH - 300 Re-2.0
~ ... ..... ..
---------·---~·-~~~~~
o.e
o .e
-
--·------·-····-··-···---·-···-··-··-···-·····-········--·-·-···-·-··-·-··-··-··-··-··-·--·-
0
~ 0 .7 ....................-·-···..........- ..--..··---··-··..--------·--·--·-···--···-·..-·.................._,_
0 • :5 - · · - · - · - · - · - - - - · - - -..---·-·-··-------+N'I'ERNOOE•=EII!tt11-..····-·-
-+- INTERNODE•250
...... INTERNCDE•2tl0
0
LENGTH mm
3 . 5
ex -1ooo:.

Internode lengtbs in mm. Conditions as indicated.
NCOS •NFLUENCE ON K1~
., .......
NOCE WtOTH- 300 Re•2.C
o.s
0
~ 0.7
o.e
0.5 -....... -... -... -... - .. -·-------·······-------------+N'I'e~eee-- ..·-

rNTERNOOS•2SO
INTERNOce-aoo
Q...
0
.. ...........
2
LENGTH mm
3 . 5
ex 1ooo;,
Figure F.37 Influence of node on axial stiffness.

231
NOCE INF~UENCE ON ~~~

NOOe WIOTH- 300 Re-2.~
o . e -··-·················-··········-·····-··-·-·········-·················-···-···-····--·--·············-········-·····
0
~ 0 .7 ··--···-·······-····--·-----················-···-···-·········-······-···-·······················-···-···------
0 . e ------w·············-·················--·-···----···········-·····················~ ···--········•n••············-········
0.5 r-········••••••••····--··-·--·-·..·-·--·····--··· ............... --..·-·--·---..·i·N~RNeDE=·SS<~ ......... ~

-+- I NTERNOCE""'250
."... 1r:-rreRNOoe-aoc
=.. ..
..·=·=···~·-=·-~·=
0.4 ~~~~~~~-~-.. =-·~--~~-~-·~···=·=-~--=·-~--~·-=·=··=··= ..·=··=-~~~
0 ~ 2 5 ...
L..5NGTH mm ex ~ooo::>

NOOE INFLUENCE ON K~~
., NODE WIDTH • 300 Re•2.S

... .... .... ··- ··- . ·- ... ... .... -·· .... .... ··-· ··- .... ·-- -- . .... ... ... --- --- -·
0 .8 ........-·-------···-········-·····-···---·----·--·-·--·------·-······--·-.. ··-·-----
0 .B ......................... .. ...............-................... ................... ···-···--··..-··-··-..·· -·····
0
~ 0 .7 ···············-·--···---·-·..--·-····-···--·····-·"····--................ - ......................................
0. B --------·-··-···-·· ....................- .........- .....- ....................................--·--·-··
o. '!5 ··-···-······-···-······· ..~-··-~-----··-·····-····----·-······-····t-~~-se.e······

~ INTERNOOE=250
~ JNTE~Nooe-aco
0.4 ........•
0 2 ... 5
LENGTH mm ex ~ooo::>

232
Appendix F: Structural rnadelling and analysis of bamboa structures
F .5-Simplified caleulation of bamboo structures.

So far the question bas in the first place been, to what ex tent conicity, variation in the elastic
modulus and nodes should or should not be taken into account in the calculation of bamboa
structures, and, in the second place, as a corollary, ü using average properties would be
precise enough for structural purposes.
Some theories have been adapted in this appendix to be able to simulate a wide range of
possible situations, in relation to the combination of the different parameters involved.
In the fitst place it has been shown that in principleconicity accounts for an average reduction
of 5% in the bending and axial stiffness.
On the other hand, a much more importantroleis played by the nodes, amounting to as much
as a reduetion of 40% in the bending stiffness and 12% in the axial one.
This suggests that within certain margins of error it is perhaps possible to reduce the average
stiffness of bamboo elements by a eertam factor, to account for the overall influences. It is
not the intention of the author to state precisely what that factor should be. In the frrst place
this is because such a proposal should come from a broad experimental validation, for which
the data at the Eindhoven University are probably not representative enough, and second,
because it is believed that such an approach is absolutely unfeasible without a correct
conneetion to classifieation criteria, whieh does not exist yet for bamboo.
However, it is considered that light has been shed on the importanee and relevanee of the
different parameters, and a general approach has been put forward, so that it ean be forther
supported for any particular situation.
A general condusion is that just taking the average values for the parameters of the culms,
is indeed an over-simplification that may lead to underestimations of the deformations(though,
not necessarily, to unsafe designs).
Just for the sake of a practical appreciation of what is said above, the result of figure F.40
is presented.
In this case the connections were made following the proposal of chapter 5, in such a way
that wood was projected outside the culms, and then the joints were made by screwing and
gluing them to steel plates, to sandwich the pieces of wood, so that the joint was as stiff as
233
Appendix F: Structural rnadelling and analysis of bamboo struétures
possible, as indicated in figure F.41.

It can beseen in figure F.40 that certainly the estimation of deformations basedon average
properties is worse than those made when the effects of the variations are taken into account.
In this case, of course, culm properties were determined specifically for each of the elements
of the truss, so this is strictly an academie exercise, because it would be unfeasible to do that
in a real construction case. This does, however, give us an idea of the validity of the fmdings
presented in this chapter.
FRAME 1-02
40 ············· ···~· ................ ..
- EXPERIMENT
~ AVERAGE PROPERTIES
~ MDDlFIEO PROPERTIES
0 2 4 s s
FORCE N ex 1000)
Figure F.40: Deformation load curves for a bamboo truss.
234
Figure F.41: Double-plated connection.
F.6-Experimental anaJysis of the behaviour of bamboo structures.

Figures F.42 to F.45show relations between displacements and forces for three different
structural configurations.
The frrst of these figures shows the complete load deforroation curve for a truss type of
structure. The same type of conneetion as the one indicated in tigure F.41 was used for all
these cases, except that the steel plates were not glued to the wood fittings.
1t can be seen that, but for some settiement at the beginning, the behaviour is very elastic,
ending with a brittie type of failure. The structure is pretty stiff, and the reached capacity is
very satisfactory in terros of the employ of the capacity of bamboo, but also in terros of the
relation of the load carried and the weight of the structure. Faiture occurred by longitudinal
splitting due to tangentlal strain near the conneetion of element number 5 and joint 3. One
interesting, though accidental observation, that could be made during this test, was that faiture
occurred in the first trial through buckling of the steel plates in the conneetion with the stiffly
reacting steel frame, near the maximum capacity of the truss (because of an underestimation
of the thickness of the plates, of course). This is interesting. Though the main advantage of
bamboo in areas prone to seismie activity is its lightness, this observation suggests that in the
235
FRAME 1-01
ex 1ooo)
a~-~-·=·-=--=----=---=---r.-~--~---~-=---~·-=-·=·--=-·=--·-=-·-~---~---=···=-~---=---~---~---=--~~==~
0 10 20 30 40
DISPLACEMENT nrn
Figure F.42: Test result for a truss.
case energy dissipation is needed, this can be concentraled in the region of the steel
connectors. (lt is good to remember that such a possibility does not exist anywhere else in
the structure of bamboo culms).
Figure F.43 .presents the result on a frame-truss. In this case, and in relation to the elements,
transversal forces are applied, increasing the amount of bending moments. Still, it can be seen
that performance is very good. Capacity and stiffness are high. Failure happens in a brittie
way here too, in the highly compressed diagonal element
Finally, figures F.44 and F.45 show the result of testing a portal frame. In figure F.44 the
scale is kept the same as in figure F.43 to facilitate the comparison, but the plot is repeated
in figure F.45 to permit further observation of the behaviour. lt is clear that this type of
236
Áppendix F: Structural modelling and analysis of bamboo structures
FRAME 2-01
ex 1ooo)
24~-=
...r..=··=···~·-=·-~-===-=·-==-~·-=··-=-==·=··=·==·-=·=--==--=--=-=~====~
z
~ 12 -·· --···--·-·--···········--········-············-·········-···-····-··-···-···-···········-···-····
~
L .... -J
0 40 ao 120 160 200

0 I SPLACEMENT mm
Figure F.43: Test result for a frame-frame.
application is not the most promising for bamboo culms. Deformations are very high for very
low rates of loading. The induced rotations of the conneetloos are also high, leading to early
local failure.
One aspect that deserves attention is the fact that in all these tests, the amount of rotation of
the conneetloos was very high, though it always happened outside the region of the glued
bamboo wood connection. This is a fact that bas to be given special attention in the
calculation of bamboo structures, though its treatment lies beyond the scope of this research
project.
This observation led us to make the proposal in chapter 6, that a connecting system in which
elements are put together using a steel fitting, helps to keep rotations under more control
237
FRAME 3-01
ex 1ooo)
a4~~-r..=--~---~···=--~--=--·~·=···-~--=-·+=··r.·=-~-~--=--·~·=···-~·-=···~--~~-=--~--~-·-~·=···~
20 r-··-··-·-···········--·-···----·---·-······-··········-·········--·--·-····-··-··········-···-·
1s r································----·-····-················-······-··-·-----·········-·-·······-·······-···-
z
~ 12 -·-··-··-·-···---··-··-····-···-···--····-········-·--··--··---·········-·····-··-··-···········-
u.
8 -··-·-·---·-················-··-··--·--·-···-············-·-·····-····--·--·-·-···-··-
4 f----···-···-···-················-····-···-·---·····-···············--··-······-·-----··················
0~~~~~~~---·~····~-·~···~····~····~··-=···~·-···~···=····=····=·=·~
..
0 40 80 120 160 200
t:liSF'LACEMENT
Figure F.44: Test result for a frame test.
Some limited laboratory experience suggests that even in the case of frames, this proposal
might significantly improve their structural performance. In spite of that, cost may be the
major limiting aspect in this type of application, thus the recommendation of using bamboo
for trusses and frame-trusses is still valid.
From figure F.46 onwards, details of the tested frames are given. The respective indications
are found at the bottorn of each of the figures.
238
FRAME 3-01
990 ............................ ······ ····························· ········-···· ...... ............................
z
~
er
a
lL
390 ................................ ··········-·····················································································
190 ........... ······················································································································
- 1a .............. ...... ....... ··-··· -····· ........ ···-·· ...... ···-·· ··-··· .................... ......
a "10 80 120 160
DISPLACEMENT rrm
Figure F.45: Frame test, enlarged vertîcal scale.
239
Appendix F: Structuralrrwdelling and analysis of bamboo structures
Figure F.46: Typical L-type conneetion employed.
The above figure shows the 1 mm thick steel plate used for joining the wood components of
the connection. The conneetion was made in two steps. First, the waoden components were
glued to the culms, and temporaly aligned to each other. Once the glue settled, the steel plates
were screwed in position.
240
Appendix F: Structural modelZing and analysis of bamboo structures
Figure F.47:Culm-to-support connection, showing diameter adjustment. The steel ring-

eramp holds bamboo against the wood fitting until the glue is settled.
This figure shows how, within certain margins, size differences between the culm and the
wood fitting, mean no major problem for the making of the connection.
241
Appendix F: Structural rnadelling and analysis of bamboo structures
Figure F.48:Truss placed in the steel reacting frame, prior to testing.
The load was applied as indicated in figure F.42. In this particular case a maximum load of
9.5 kN was reached, with a deformation of about 35 mm. The total weight of the truss was
120 N.
242
Appendix F: Structural modelZing and analysis of bamboo structures
Figure F.49: Frame-truss in the process of testing.
As seen in the figure, vertical loads were hung using collom towing bands, Lo keep their
position on the beam. AL the momentthe photo was made, the test was about half way before
failure.
243
Appendix F: Structural TrUJdelling and analysis of bamboo structures
Figure F.SO:Bamboo frame being tested.
The tigure shows the beginning of the test of a bamboo frame. The structure proved to be
very flexible, as will be shown in tigure F.52.
244
Appendix F: Structural rnadelling and analysis ofbamboo structures
Figure F.Sl :Detail of the load application point in frame tests.
245
Figure F.52: Bamboo frame after failure.
The above figure shows a bamboo frame right after failure. Large plastic rotations of the top
connections initiated early during the test. Final mechanism of failure included explosive
splitting of the culms, in the bottorn connections to the steel reacting frame.
246
KEFERENCES
1 Grosser, D.; Liese, W.; (1971). On the Anatomy of Asian Bamboos with special
reference to their vascular bundles.
Wood Science and Technology. Vol 5, pp 290-312.
2 Liese, W.; (1981). Anatomy and Properties of Bamboo. Proceedings of the
International Bamboo Workshop.
October 6-14, Hangshou, People's Republic of China,
pp 196-312.
3 Norris, C.; Wilbur, J.B.; (1960). Elementarv Structural Analysis. Me Graw Hili, New
York.
4 Willems, N.; Lucas, W.M.; (1978). Structural Analysis for Engineers. Me Graw Hill,
New York.
247
Appendix G: Geometrie parameters of bamhoo culms
Appendix G: Geometrie parameters of bamboo cuJms

G.l-Introduction
Nature supplies bamboo in all sorts of diameters, thicknesses and shapes. The exact way the
extemal diameter changes along the length of the culm is very complex. In a strict sense, it
is specific for every culm.
Though for structural purposes it is important to determine the way geometrie parameters
change along the length, an exact description of this may seem rather irrelevant.
A mixed sample of Bambusa Arundinaria and Guadua s.p. was studied at the Eindhoven
University of Technology to determine a simple way to relate extemal diameter to thickness,
and to see if the area of the cross-section and the second moment of area could be
appropriately approximated by a simple relation to be used for structural purposes.
G.2-External diameter and thickness

Finding an easy relation between the thickness and the diameter may contribute significantly
to the simplification of calculating fonnulae, but the level of inaccuracy must be kept under
control so that this simplification does not lead lack of safety.
Table G.l on the next page shows the result of calculating descriptive statistles for the
relation between the thickness and the extemal diameter for a large sample, as indicated in
the introduction.
As can be seen, the average ratio is 0.09. In other words, we can say that
t = 0.09 c~J.. eq.G.(l)
To have a better idea of the level of accuracy in this assurnption, an estimation of the 97%
limits for the mean was undertaken, as shown in table G.2. Here, one can see that such a
simplification can not lead to any significant change in calculation output, because the level
of varlation of the probable error is far below that related to other variables.
248
Appendix G: Geometrie parameters of bamboo culms
Variable: · ratio thiekness/diameter
Sample size 87
Average 0.0919237
Median 0.0849339
Mode 0.06875
Geometrie mean 0.0897555
Varianee 4.44906xl0-4
Standard deviation 0.0210928
Standard error 2.26138xto·3
Minimum 0.0632911
Maximum 0.154762
Range 0.0914708
Lower quartile 0.0763889
Upper quartile 0.103659
Interquartile range 0.0272696
Skewness 0.981581
Standardized skewness 3.73775
Kurtosis 0.328252
Standardizeà kurtosis 0.624972
Table G.l:Deseriptive statistles for ratio between thickness

and diameter.
Sample Statistics: Number of Obs.: 87

Confidence Interval for Mean: 97%
Sample 1 0.0869324-0.0969149
Table G.2: 97% confidenee limit for the mean ratio

thicknessldiameter.
249

variabie
diameter 0.0927350 .00226840 .8834 0.0000
r= o.9stt
Table 0.3: Regression thickness diameter.
A further look can be taken by calculating a regression of thickness and diameter. The result
of such a calculation is shown in table 0.3, where the resemblance between the regression
coefficient and the average of the ratio in the previous tables turns out to be self-evident
G.3-Area and second moment of area

It is possible to describe the variation of the external diameter with a linear rule like
eq.G.(2)
In the same way the area and the second moment of area can be calculated respectively as
eq.G.(3)
eq.G.(4)
It can be proved that. from the above equations
0.16381t4>A6t 2 eq.G.(S)
6A = • - 0.0819'11:6
l ••
and
250
eq.G.(6)
Now, though these relationships may represent the varlation of geometrie parameters in
bamboo in an appropriate way, it is worth seeing whether it is possible to say that, for
example, both the area and the moment of area vary according to a linear relationship with
the length of the culm.
A sample was studied with the purpose of estimating the possible error of such a
simplification. The errors in the estimation of the middle area and the middle moment of area
for the culms are statistically analyzed in table G.4.
The 97% confidence limits for the means are also calculated. It can be seen that a linear
relation between moment of area and length is good enough for structural purposes, and the
same can be said for the area, though the 97% confidence limit for the mean error is much
wider, due to two outlying values in the sample. Without them, the results are equally good.
It can in any case, be seen facts suggest that it is possible to assume linear relations between
area and moments of area and the length, without significantly affecting the accuracy of
structural calculations. In this simplified approach, of course no rational relation would exist
between the expression for area and that for moment of area, and thus the respective
parameters a1 and aA must be statistically determined for every batch of bamboo
according to pre-established classification criteria.
251
Varlable: moment area error % area error%

Sample size 51 51
Average error 1.52957 0.254269
Median 0.426717 0.0711111
Mode 0.0 0.0
Varianee 12.256 0.337543
Standard error 0.490217 0.0813541
Minimum 0.0 0.0
Maximum 19.1415 3.17346
Range 19.1415 3.17346
Lower quartile 0.135776 0.0226286
Interquartile range Ll8136 0.196813
Skewness 3.70161 3.69561
Standardized skewness 10.7919 10.7745
Kurtosis 14.4306 14.3744
Standardized kurtosis 21.036 20.9541
Table 0.4: Descriptive statistles for the errors in the

estimations of areas and moments of areas.
moment of area

Confidence Interval for Meao: 97%
1 0.434294-2.62484 %
Table 0.5: 97% confidence interval for the meao

error in the estimation of the ·moment of area.
252
area
Confidence Interval for Mean: 97%
Sample 10.0725023-0.436036 %
Table G.6: 97% confidence interval for the mean

error in the estimation of the area.
253
Fundamentals
Symbols
SYMBOLS
:correction factor for the calculation of cross sectionat area
y :shear deformation angle [radians]
6 :coordinate angle [radians]
4 :change
-6.1 , but m
. section
. E6. , where tt. ts
. 2 2 Iz
14 3 3 !<IV
&'I
254
Fundamentals
Symbols
&'E
. :strain in direction indicated by i
:strain in direction indicated by , at the location I
em :maximum strain
v :Poisson's ratio
p :density [kg/m3]
~. :external diameter [mm]
~~ . :internal diameter [mm]
(~.}max :maximum external diameter [mm]
(~.)mln :minimum extemal diameter [mm]
~(x) :x function of the initial twisting deformation
:slenderness
oc :compressive stress [N/mm2_1
255
Fundamentals
Symbols
o8 :tangential stress [N/mm~
o" :stress in the x direction [N/mm 2]
oer :Euler stress, or critical stress [N/mm2]
o", :maximum stress [N/mm2]
o, :tensile stress [N/mm2]
'C :shear stress [N/mm2]
:stress function
A :when used as an index, it refers to the culm extreme with the largest cross
sec tion
A., :area of an elliptical ring [mm2)
a11 :node length [mm]
B :when used as an index, it refers to the culm extreme with the smallest cross
section
256
Fundamentals of the design of Bamboo Structures
-------
Symbols
E"" :average elasticity modulus, calculated as the average of the correspon-
dent values at the extremes of the culm
:elasticity modulus at the extreme with the smallest cross section
:elasticity modulus in the direction indicated by the subindex
:bamboo elasticity modulus [N/mm2]
:wood elasticity modulus [N/mm2]
e :Iateral expasion [mm]
:lateral expansion coefficient [mm]
F, :as defined in eq.F.(59-61) and F.(71-73)
fc :compressive strength [N/mm 2]
G :shearing modulus [N/mm2]
h :hard-wood code in table 4.2

HO :null hypothesis
HE :alternative hypothesis
I"" :average second moment of area [mm4]
Itlf :equivalent second moment of areafora composed section [mm4]
257
Symbols
:polar moment of area [mm4]
k, :reaction coefficient in the radial direction [N/mm3]
:stiffness coefficient
K :curvature [1/mm]
l :length [mm]
M, :bending moment around the y axis [N-mm]
:bending moment per unit length [N-mrn/mm]
m :slope
n, :stress in the direction indicated by the subindex
N :number of nodes
:force in the direction of the subindex [N]
. :criticalload [N]
r :resin code in table 4.2
258
Symbols
r :middle radius [mm]
:external radius [mm]
t
r,
4»._
:square of the correlation coefficient

s :standard deviation; soft wood code in table 4.2
:internode length [mm]
t :time [years]; thickness [mm]
:time coefficient
T :torsional moment [N-mm]
u, :displacement in the direction indicated by the subindex [mm]
:bending potentlal energy [N mm2_1
:axial load potentlal energy {N mm2]
:shear force in the direction indicated by the subindex [N]
259
Fundamentals of the design of Bamboo Structures
Symbols
:coefficient of varlation
:bonding stress due to bending [N/mm]
:shear stress in the directionj and perpendicular to i [N/mm~
vrs :versus, in figures and plots

VS :versus, in statistical tables
w :material factor in eq.C.(l)
w :PVAc glue code in table 4.2
w, :radial bending displacement [mm]
:radial bending displacement constant [mm]
z, :as defined in eq.F.(41-42)
V
a a a
-+-+-
axöyaz
260
Curriculum Vitae
The author graduated as a civil engineer from the University of Costa Rica. He obtained a
Post-graduate Diploma in Advanced Structural Engineering from the University of
Southampton in England, and he later undertook a Course on Seismie Engineering in the
National Autonomous University of Mexico.
He was the coordinator of the Structural Design Section of the Bridge Department of the
Ministry o(Public Works of Costa Rica, Head of Construction of the Program of Rural
Health eentres of the same Ministry, and Chief Engineer of the Civil Aviation Authority
of Costa Rica. For many years he has been a private consultant on structural and seismie
engineering. His professional work includes structural design and field supervision of the
construction of buildings and bridges, and the structural design of especialized port
facilities in Costa Rica. The evaluation of buildings after earthquake damage and the
reinforcing of existing ones has been an area of professional concentration in the last
-
years. This includes the reinforcing of hospitals and telecommunication centres.
He is Senior Lecturer on Concrete and Masonry Structures at the Department of Construc-
tion Engineering of the Costa Rican Institute of Technology. He has also served as Head
of this Department, as Dean of the Faculty of Construction Engineering and Forestry, and
as Vice-Rector for Research and Extension, in that University. He has contributed papers
to international technica! congresses on bamboo and masonry structures, having been the
chairman of the First Central American Seminar on Masonry Design and Construction. He
has been merober of organizing committees of regionat congresses on Science and
Technology Policy Making.
He has worked as advisor for the Ministries of Housing and of Science and Technology of
Costa Rica. With the support of the Organization of American States he carried out an
evaluation mission of Housing Research eentres in Latin America. In 1988 he was invited
by Appropriate Technology Intemational, to be guest speaker of the Committee of Foreign
Affairs of the Congress of the United States of America, when he spoke about Develop-
ment and the Role of Universities.
Since 1989 he joined the Department of Structural Design of the Faculty of Architecture
and Planning of the Eindhoven University of Technology, where he carried out the
research project reported in this thesis.
261
Statements
pertaining to the thesis of O.A.Arce-Villalobos
Fundamentals of the Design of Bamboo Strnctures
1-Tests on the tangentlal tensile capacity and on compression for three different species show
the possible existence of a common maximum for tangentlal strain in bamboo (see sections
2.2.1 and appendix B).
2-Bamboo culms do not fail in compression, in bending or shear, but do fail when a
maximum tangentlal tensile stress is reached (see chapters 2, 3, and appendices B and D}.
3-The majority of fittings based on some sort of peneteation normally used in construction,
are not suitable for bamboo because they create high tangentlal stresses (see section 2.2.2).
4-The use of glue and a wooden fitting has not only proven to be a way of taking and
transmitting farces to the culms, but it actually reinforees them (see chapters 4 and 5, and
appendices C and D).
5-Contrary to general belief not the fibres, but the matrix, is the major souree of strength in
bamboo culms (see sections 2.2.2, 5.1, 5.4 and 5.6).
6-The nodes contribute negatively to the axial and bending stiffness of the culm, and to the
parallel tension capacity of bamboo (see sections 2.2.2 and 0.4}.
7-The calculation of bamboa structures can easily be made by using a factor to take the
variation of the modulus of elasticity, and the cross-section along the culm into account (see
appendices G and H, and chapter 7).
8-The results of tests on bamboo trusses are highly encouraging. The balance between
lightoess, on one hand, and capacity and stiffness, on the other, gives a vote of confidence
to the pnssibilities of this material (see appendix 0 and section 7.10).
9-Bamboo culms can be appropriately employed in trusses, space trusses and frame trusses,
provided connections are well taken care of. Bamboo culms are not adequate for frames nor
for beams.(see appendix G and chapter 7).
lO-lt is necessary to pay more attention to practical design problems in future bamboo
research actlvities (see chapter 7).
11-Some people would like to 'preserve third world cultures' like keeping monkeys in a cage.
There has to be a balance between cultural preservation and the right of people to decide
about their own future. This is only possible through education and knowledge. No lasting
freedom is possible in the darkness ofignorance.
12-Through the power of science we have managed to develop our knowledge to unthinkable
horiwns. Our brain has reached formerly unimaginable frontiers. Now it is time we start
developing our hearts, and to embrace ethics as the centre piece of our behaviour.
13-0ur ego is so huge that we believe it is we who will destray this planet with our careless
treatment of the environment. The truth is that we will destroy ourselves ftrst, not the earth,
by making it uninhabitable, unless we seriously incorporate environmental considerations as
a routine part of everything we undertake.
14-We no langer elect our leaders, we only 'buy' images. Like with other things in the
market, we do not buy what is the best for our interests, but just what is sold in the best
manner. No wonder actors become politicians. No wonder politicians are becoming actors.
15-0n top of the unfair division of the world according to the level of one' s possessions, or
according to the colour of our skin, there is the unfair distinction between men's and women's
chances of success and happiness. Even in the most developed countries wamen are still
destined to 8econdary roles, violence of all kinds, even the right access to knowledge and
education, being men's main instrument of domination. Men's appreciation of wamen is
limited to Mother' s day and to the sexist concession that bebind every great man there is a
great woman.
ii
16-Some people blame science and technology for everything that is going wrong on our
planet. But even if they are right, science and technology, in a proper moral context, are the
most powerfut instruments to make a better life possible.
17- My family and I have been favourably impressed by the Dutch people. We have found
people who are warm, hospitabie and generous. Everywhere we went, people made us feel
wen treated, respected. Part of our hearts will remain bere, with our friends, for ever.
iü
nr.9 nr.19
Strukturering en verwerking van tijdgegevens Supporting deelsion making processes
voor de uitvoering van bouwwerken A graphical and interactive analysis of multivariate
ir W.F. Schaeter data
ir P.A. Erkelens drsW. Adams
Eerste druk: februari 1988 Eerste druk: maart 1991
• Uitverkocht
nr.20
nr.10 SeH-help bullding productivity
Stedebouw en de vorming van een speciale A method for improving house bullding by low-
wetenschap income groups applied to Kenya 1990-2000
K. Doevendans ir P.A. Erkelens
Eerste druk: februari 1988 Eerste druk: mei 1991
nr.11 nr.21
Informatica en ondersteuning van ruimtelijke Oe verdeling van woningen:
beslultvorming een kwestie van onderhandelen
dr G.G. van der Meulen drs V. Smit
Eerste druk: februari 1988 Eerste druk: mei 1991
nr.12 nr.22
Staal in de woningbouw, korrosiebescherming van Flexibiliteit en kosten in het ontwerpproces
de begane grondvloer Malthijs Prins
ir E.J.F. Delsing Eerste druk: augustus 1992
nr.23
nr.13 Spontane nederzettingen begeleid
Een thermisch model voor de berekening van Po Hin Thung
staalplaat-betonvloeren onder Eerste druk: juni 1993
brandomstandigheden
ir A.F. Hamerlinck
nr.14
Oe wijkgedachte In Nederland
Gemeenschapsstreven in een stedebouwkundige
context
dr ir K. Doevendans
dr R. Stolzenburg
nr.15
Oiaphragm effect of trapezoidally proflied steel
sheets. Experimental research into the influence of
force application
ir A.W.A.M.W. v.d. Bogaard
Eerste druk: juli 1989
nr.16
Versterken met spuit-ferrocement.
Het mechanische gedrag van met spuit-ferrocement
versterkte gewapende betonbalken
ir K.B. Lub
ir M.C.G. van Wanroy
nr.17
De traetaten van Jean Nlcolas Louis Ourand
Ir G. van Zeyl
Eerste druk: oktober 1990
nr.18
Wonen onder een plat dak.
Drie opstellen over enkele vooronderstellingen
van de stedebouw
dr ir K. Doevendans
Eerste druk: januari 1991
ISBN 90-6814-524-X

Fundamentals of The Design of Bamboo Structures

Uploaded by

Document Informationclick to expand document information

Document Informationclick to expand document information

Copyright:

Available Formats

Fundamentals of The Design of Bamboo Structures

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

Fundamentals of The Design of Bamboo Structures

Uploaded by

Copyright:

Available Formats

Fundamentals of the design of bamboo structures

Citation for published version (APA):

Document status and date:

Please check the document version of this publication:

Take down policy

Download date: 03. Jul. 2020

IR. O.A. ARCE-VILLALOBOS

faculteit t(!3 bouwkunde

Prof. drs G.A. Bekaert nr.3

Prof. dr S. von Moos nr.7

Prof. ir D. Vandepitte nr.B

to obtain the degree of doctor at the Eindhoven University of Technology by the

Oscar Antonio Arce-Viiialobos

bom in San José, Costa Rica.

Fundamentals of the design of bamboo structures I Oscar

Photographs by B.G.J. Elfrink

Oscar Antonio A ree-Viiialobos

geboren te San José, Costa Rica.

ir. W.R. de Sitter

to obtain the degree of doctor at the Eindhoven University of Technology by the

Oscar Antonio Aree-Viiialobos

bom in San José, Costa Rica.

ir. W.R. de Sitter

Thanks is given to Dr.J.J.A.Janssen, the co-promoter. He patiently listened to the author's

Chapter 2: Tensile strength of bamboo . . . . . . . . . . . . . . . . . . . . . .. . . . . . .. . . . . 9

Chapter 3: Compression Behaviour of Bamboo Culms . . . . . .. . . . . . . . . . . . . . . . . 38

Chaptcr 4: Glueability of bamboo wood phases . . . . . . . . . . . . . . . . . . ... ..... . . 54

Chapter 5: Design of bamboo connections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66

Chapter 6: Design of bamboo structures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83

Chapter 7: Conclusions and further research recommendations . . . . . . . . . . . . . . . . . 97

Appendix A: Experimental analysis of parallel tension . . . . . . . . . . . . . . . . . . . . . . . 106

Appendix B: Experimental analysis of short elements ....................... 114

Appendix c. : Experimental analysis of bamboo wood glued phases .............. 129

Appendix O:Glued connectors for bamboo................................ 150

Appendix E: Bamboo columns ....................................... 167

Appendix F: Structural modelling and analysis of bamboo structures ............. 194

REPERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . -......... 247

Appendix G; Geometrie parameters of bamboo culms . . . . . . . . . . . . . . . . . . . . . . . 248

Curriculum Vitae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 261

Keywords: Bamboo I mechanica! properties I structural design I compression capacity I

Cover photo: Test specimen of a bamboa wood glued conneetion

Trefwoorden: Bamboe/ mechanische eigenschappen I konstruktief ontwerp I drukcapaciteit

Foto voorpagina: Proefstuk van een lijmverbinding tussen bamboe en hout

curiosity and ingenuity.

TENSION ~ CBAPI'ER2 !:::) APPENDIX A

supports the design considerations in chapter 6. Appendix G contains a description of

1 Addis,W.;1990. Structural Engineering: The nature of theory and design. Ellis

2 Anonymous; 1988. Proceedings ofthe International Bamboo Workshop. Kerala, India,

3 González, G.; 1992.Técnicas de preservación de culmos y esterillas de bambû.

4 Janssen, J.J.A.; 1981. Bamboo in building structures. Ph.D. Thesis, University of

Chapter 2: Tensite strength of Bamboo

Cbapter 2: Tensile strengtil of bamboo

Chapter 2: Tensite strengthof Bamboo

Chapter 2: Tensile strengthof Bamboo

Ffgure 2.1: Specimen by Atrops.

Chapter 2: Tensile strengthof Bamboo

r-- 120 -----j