Time Dependent Behaviour of FRP Rebar in Reinforced Concrete

m m
U N I V E R S I T E DE
m S H E R B R O O K E
Faculte de genie
Departement de genie civil
TIME-DEPENDENT BEHAVIOUR OF FIBRE REINFORCED

POLYMER (FRP) BARS AND FRP REINFORCED CONCRETE
BEAMS UNDER SUSTAINED LOAD
LE COMPORTEMENT A LONG TERME DES BARRES EN
POLYMERES RENFORCES DE FIBRES (PRF) ET DES
POUTRES EN BETON ARMEES AVEC DES BARRES EN PRF
SOUS CHARGE PERMANENTE
A Thesis Submitted in Partial Fulfilment of the Requirements for the Degree of
Doctor of Philosophy
Speciality: Civil Engineering
Tarik A. YOUSSEF
Sherbrooke (Quebec) Canada
August 2010
1 * 1
Library and Archives

Canada
Biblioth^que et
Archives Canada
Published Heritage
Branch
Direction du
Patrimoine de l'6dition
395 Wellington Street

Ottawa ON K1A 0N4
Canada
395, rue Wellington

Ottawa ON K1A 0N4
Canada
Your file Votre reference
ISBN: 978-0-494-70621-3
Our file Notm reference
ISBN: 978-0-494-70621-3
NOTICE:
AVIS:
The author has granted a nonexclusive license allowing Library and

Archives Canada to reproduce,
publish, archive, preserve, conserve,
communicate to the public by
telecommunication or on the Internet,
loan, distribute and sell theses
worldwide, for commercial or noncommercial purposes, in microform,
paper, electronic and/or any other
formats.
L'auteur a accorde une licence non exclusive

permettant a la Bibliotheque et Archives
Canada de reproduire, publier, archiver,
sauvegarder, conserver, transmettre au public
par telecommunication ou par I'lnternet, preter,
distribuer et vendre des theses partout dans le
monde, a des fins commerciales ou autres, sur
support microforme, papier, electronique et/ou
autres formats.
The author retains copyright

ownership and moral rights in this
thesis. Neither the thesis nor
substantial extracts from it may be
printed or otherwise reproduced
without the author's permission.
L'auteur conserve la propriete du droit d'auteur

et des droits moraux qui protege cette these. Ni
la these ni des extraits substantiels de celle-ci
ne doivent etre imprimis ou autrement
reproduits sans son autorisation.
In compliance with the Canadian

Privacy Act some supporting forms
may have been removed from this
thesis.
Conformement a la loi canadienne sur la

protection de la vie privee, quelques
formulaires secondaires ont ete enleves de
cette these.
While these forms may be included

in the document page count, their
removal does not represent any loss
of content from the thesis.
Bien que ces formulaires aient inclus dans

la pagination, il n'y aura aucun contenu
manquant.
Canada
Abstract
ABSTRACT
Greater attention has been given recently to the long-term performance of FRP-reinforced
concrete elements (beams in particular). Despite the conducted effort in the past two decades,
the area of long-term behaviour of FRP reinforced concrete is virtually considered virgin
grounds. Decision makers and designers, alike, are in dire need for such hard-to-obtain
information content. Such necessity renders this type of research a core requirement to
promote the widespread of FRP-internal-reinforcement usage.
In this respect, an extensive experimental/research program has taken place at the University
of Sherbrooke FRP Durability Facility. The program, consisting of four phases, studies the
creep performance of FRP bars as well as the overall long-term behaviour of FRP reinforced
concrete beams. Phase 1 deals with the creep performance of two types of GFRP bars
subjected to different levels of sustained axial load; causing creep rupture at higher levels. In
Phase 2, six different types of GFRP bars are tested under two levels of allowable service
load, according to the currently available North American standards. The test duration, for the
two phases, exceeded 10000 hours (417 days) wherein regular monitoring of creep strain
evolution took place and the creep coefficient of GFRP bars was calculated. Residual tensile
tests and microstructural analysis followed the long-term testing period. It was found that 45
% of the GFRP bars' tensile strength, fUiave, is a safe limit for GFRP exhibiting sustained load,
in standard laboratory conditions. Microstructural analysis shows that the increase in creep
strain, after the 10000 hour period, is negligible for GFRP bars under allowable service load.
Phase 3 consists of twenty reinforced concrete beams (ten pairs) comprising GFRP, CFRP,
and steel reinforcing bars. The dimensions of which are 100 mm x 150 mm x 1800 mm,
installed under third-point sustained load, for a period exceeding one year. Exhibiting a
maximum applied moment of 25 % of their nominal moment capacity, Mn, all beams were
regularly monitored in terms of (i) time-dependent deflection, (ii) strain increase in concrete
and reinforcement and (iii) crack widths. Theoretical predictions for immediate deflection
were calculated, using three methods (ACI 440.1R-06, CAN/CSA S806-02 and the ISIS
Canda Design Manual (2007)), and compared to the obtained experimental results. Results
Abstract
showed that the calculations, regarding immediate deflection, under estimate by 67 %;

underestimate by 10 %; overestimate by 11 %, for the aforementioned methods, respectively.
The long-term to immediate deflection ratio, X, was calculated for all beams and compared to
ACI 440.1R-06 and CAN/CSA S806-02 predictions. Results showed that the North American
standards are conservative as regards long-term deflection prediction. Immediate crack width
results were compared to the prediction equations adopted by ACI 440.1R-06 and CAN/CSA
S6-06, on the one hand, and by the ISIS Canada Design Manual (2007) on the other hand.
Satisfactory results were found when the kb bond-coefficient factor is taken as 1.2 and 1.0,
respectively. From the obtained data, the time-dependent kt multiplier, accounting for crack
width increase after one year, was deduced as 1.7 and 1.5 for both models, respectively.
Phase 4 deals with four full-scale GFRP reinforced concrete beams, of dimensions (215 mm x
400 mm x 4282 mm), subjected to uniform distributed load for a period of six months.
Sizeable concrete blocks (of dimensions 610 mm x 762 mm x 1219 mm and weight = 13334
kN) were arranged on top of the beams to simulate sustained uniform distributed load. The
main study parameters, of this phase, are (i) bottom reinforcement ratio and (ii) type of
upper/compression reinforcement (GFRP and/or steel). The applied moment ranges from 15 to
21 % of the nominal moment capacity for the beams. Numerical modelling took place using a
computer program (Fortran-2003) based on the age-adjusted effective modulus method, to
predict the long-term deflection of the beams. The creep and shrinkage coefficients were
calculated based on the ACI Committee 209 recommendations (1992) and CEB-FIP Model
Code (1990). The theoretical curves were in very good agreement with the measured values.
Furthermore, the empirical models available in ACI 440.1R-06 and CAN/CSA S806-02 were
used for long-term deflection prediction. These predictions showed that both models can serve
as upper bound and lower bound limits for the measured long-term deflection curves,
respectively. As regards crack width prediction, the equations adopted by ACI 440.1R-06
(same as that of CAN/CSA S6-06) and by the ISIS Canada Design Manual (2007) yield
satisfactory results when the kb bond-coefficient factor is 1.2 and 1.0 respectively (similar to
phase 3). For both equations the time-dependent kt multiplier is deduced as 1.4, after six
months.
ii
Abstract
All four phases function in synergy to better understand the long-term behaviour of FRP
reinforced concrete elements under different types of sustained load. This study presents the
long-term performance data of 99 bar samples, 20 FRP and steel reinforced concrete beams
under third-point load and four full-scale GFRP reinforced beams under uniform distributed
load. Creep coefficients for different types of GFRP bars were calculated; microstructural
analysis was conducted. The obtained data was used to verify existing empirical models in
North American standards; numerical (finite difference) modelling was conducted and showed
good agreement with the experimental results. The information content made available in this
study, regarding the long-term behaviour of FRP reinforced beams is unprecedented.
iii
Resume
RESUME
Une attention accrue a ete accordee recemment a la performance a long terme des elements en
beton renforces de PRF. Malgre l'effort realise au cours des deux dernieres decennies, le
domaine du comportement a long terme de beton arme de PRF est pratiquement considere
comme un motif vierge. Les decideurs et les concepteurs, aussi bien, ont besoin de cette tres
importante information. Cela rend la necessite de ce type de recherche une exigence
fondamentale pour promouvoir la generalisation de l'utilisation de PRF comme renforcement
interne.
A cet egard, un vaste programme experimental a eu lieu au laboratoire de durability de PRF de

l'Universite de Sherbrooke. Le programme, compose de quatre phases, a ete mis pour etudier
le comportement au fluage de barres de PRF, ainsi que le comportement global a long terme
des poutres en beton arme de barres de PRF. La phase 1 porte sur la performance au fluage de
deux types de barres en PRFV soumis a differents niveaux de charge axiale constante; qui,
pour des niveaux eleves, a cause la rupture de la barre en fluage. Dans la phase 2, six types
differents de barres en PRFV sont testes dans deux niveaux de charge (15 % et 30 % de
fu,ave)
de service (conformement aux codes disponibles en Amerique du Nord). Un suivi regulier de

revolution de la deformation de fluage a eu lieu et le coefficient de fluage de barres en PRFV
a ete calcule pour toute la duree des deux phases de l'essai (plus de 10000 heures ~ 417 jours).
Des essais residuels de traction et une analyse microstructurale a suivi la periode d'essai a long
terme. II a ete constate que a des fms de conception, 45 % de la resistance des barres en
PRFV a la traction,
fu,ave,
est une valeur securitaire, dans des conditions standard de
laboratoire. Les analyses microstructurales montrent qu'apres la periode de 10000 heures sous
charge de fluage egale ou inferieure a 45 % de / Mave , il n'existe aucune degradation dans les
barres testees.
La phase 3 se comporte vingt poutres en beton (dix paires) renforcees de barres de PRFV (14
poutres), PRFC (4 poutres), et d'acier (2 poutres). Les dimensions des poutres sont de 100 mm
x 150 mm x 1800 mm, installes sous une charge soutenue en flexion quatre points (trois tiers),
pour une periode d'un an. La charge constante appliquee correspond a 25 % du moment
iv
Resume
nominal, Mn, tous les poutres ont ete regulierement suiviez en termes (i) devolution des
deformations dans le beton et le PRF, (ii) fleche en fonction du temps, et (iii) la largeur des
fissures. Les predictions theoriques pour la fleche immediate ont ete calculees, en utilisant
trois methodes (ACI 440.1R-06, CAN/CSA S806-02 and the ISIS Canda Design Manual
(2007)), et comparees aux resultats experimentaux. Les resultats ont montre que les calculs,
de fleche immediate, sous-estimation de 67 %; sous-estiment de 10 %; surestimation de 11 %,
pour les methodes ci-dessus, respectivement. La fleche a long-terme par rapport la fleche
immediate, X, a ete calculee pour toutes les poutres et comparee aux predictions du ACI
440.1R-06 et CAN/CSA S806-02. Les resultats ont montre que les codes Nord Americains
surestiment la prediction de fleche a long terme pour les poutres testees. Les resultats de
largeur de fissures ont ete compares aux equations de prediction adopte par ACI 440.1R-06 et
CAN/CSA S6-06, d'une part, et par le Manuel de conception d'ISIS Canada (2007) d'autre
part. Des resultats satisfaisants ont ete obtenus lorsque le facteur kb d'adhesion est pris comme
1,2 et 1,0, respectivement. D'apres les donnees obtenues, le multiplicateur en fonction du
temps kt, de la comptabilite pour l'augmentation de la largeur des fissures apres un an, a ete
deduit comme 1,7 et 1,5 pour les deux modeles, respectivement.
La phase 4 traite de quatre poutres en beton armees de PRFV de dimensions (215 mm x 400
mm x 4282 mm), soumises a une charge uniformement distribute pour une periode de six
mois. La charge repartie a ete applique par le biais de blocs de beton de dimensions 610 mm x
762 mm x 1219 mm (le poid de chacun = 13334 kN) disposes sur le dessus des poutres. Les
parametres principaux de cette phase, sont (i) le taux d'armature en traction, et (ii) le type de
renforcement de compression (PRFV ou acier). Le moment applique varie de 15 a 21 % du
moment nominal des poutres, M. La modelisation numerique a eu lieu en utilisant un
programme informatique (Fortran-2003) base sur la methode du module effectif, ajuste en
fonction de l'age (age adjusted effective modulus method), pour predire la fleche a long-terme
des poutres. Les coefficients de fluage ont ete calcules sur la base des recommandations ACI
Committee 209 (1992) et CEB-FIP Code model (1990). La courbe theorique a donne des
resultats tres proches de ceux des valeurs mesurees. En outre, les modeles empiriques
disponibles dans ACI 440.1R-06 et CAN/CSA S806-02 ont ete utilises pour la prediction de la
fleche a long terme. Ces predictions ont montre que les deux modeles peuvent servir de limite
Resume
superieure et limite inferieure des limites pour les courbes de fleche a long terme mesuree,
respectivement. En ce qui concerne la prediction de la largeur de fissures, les equations
adoptees par l'ACI 440.1R-06 (identique a celles de la CAN/CSA S6-06) et par le Manuel de
conception d'ISIS Canada (2007) donnent des resultats satisfaisants lorsque le facteur kb
d'adhesion est pris comme 1,2 et 1,0 respectivement (similaire a la phase 3). Pour les deux
equations le multiplicateur en fonction du temps kt est deduite que 1,4, apres six mois.
Ensemble, les quatre phases, aident a mieux comprendre le comportement a long terme des
elements en beton renforces de barres de PRF sous differents types de charge soutenue. Cette
etude presente les donnees de performance a long-terme de 99 echantillons de barres, 20
poutres en beton, renforcees par des barres de PRF et d'acier, sous un chargement en flexion
quatre point (trois tiers) et quatre poutres a grande echelle, renforcees par PRFV sous charge
uniformement repartie. Les donnees obtenues ont ete utilisees pour verifier les modeles
empiriques existants dans les codes Nord Americains ; la modelisation numerique (differences
fmies) a ete realisee et a montre un bon accord avec les resultats experimentaux. Le contenu
des informations mises a disposition dans cette etude, concernant le comportement a long
terme des poutres renforcees de PRF est sans precedent.
Mots cles : PRF, fleche a long terme, fluage, largeur de fissures, deformation.
vi
Bibliography
BIBLIOGRAPHY
AUTHOR'S RESEARCH CONTRIBUTION
The candidate has conducted experimental and analytical investigations concerning the longterm (creep) behaviour of concrete beams reinforced with carbon and glass FRP bars.
Furthermore, the candidate has participated in some research activities and publications
concerning the combined effect of sustained load and adverse (alkaline) environment on
GFRP bars. During this research work at the University of Sherbrooke the following papers
were published/accepted or submitted for publication:
Journal Papers:
Direct results from PhD work
1. Youssef, T. and Benmokrane, B., (2010), "Creep Behaviour and Residual Properties of
GFRP Bars Exhibiting Different Levels of Sustained Load," Journal of Composites for
Construction, ASCE Journal of Composites for Construction, (Submitted).
2. Youssef, T. and Benmokrane, B., (2010), "Long-term Performance of Different Types
of GFRP Bars under Sustained Service Load," ASCE Journal of Materials in Civil
Engineering, (Submitted).
3. Youssef, T. and Benmokrane, B., (2010), "Long-term Performance of Third-Point
Loaded FRP-Reinforced Concrete Beams after One Year of Continuous Loading,"
ACI Structural Journal, (in progress).
4. Youssef, T. and Benmokrane, B., (2010), "Modelling the Long-Term Deflection of
Full-Scale GFRP-Reinforced Concrete Beams under Uniform Distributed Load,"
ASCE Journal of Composites for Construction, (in progress).
Conference Papers:
Direct results from PhD work
5. Youssef, T., El-Gamal, S., El-Salakawy, E. and Benmokrane, B., (2009), "Deflection
and Strain Variation of GFRP-Reinforced Concrete Beams after One Year of
Continuous Loading," Proceedings of The Ninth International Symposium on Fibre
vii
Bibliography
Reinforced Polymer Reinforcement for Concrete Structures (FRPRCS-9), Sydney,

Australia, July 13-15, 287p.
6. Youssef,
T., El-Gamal,
S.,
El-Salakawy,
E.,
and
Benmokrane,
B.,
(2008),
"Experimental Results of Sustained Load (Creep) Tests on FRP Reinforcing Bars for
Concrete Structures," Proceedings of the 37th CSCE Annual Conference, Quebec City,
Quebec, Canada, June 10-13, (CD-ROM), lOp.
7. Youssef, T., El-Gamal, S., El-Salakawy, E. and Benmokrane, B., (2008), "Assessment
of the Long-term Performance of FRP Bars under Different Sustained Load Levels,"
Proceedings of the 5 th Middle East Symposium on Structural Composites for
Infrastructure Applications MESC-5, Hurghada, Egypt, May 23-25, 179p.
8. Youssef, T., El-Gamal, S. and Benmokrane, B., (2009), "Time-Dependent Deflection
and Strain Variation of GFRP Reinforced Concrete Beams under Sustained Load,"
Proceedings of the 4 th International Conference on Construction Materials CONMAT
09, Nagoya, August 24-26, (CD-ROM).
9. Youssef, T., El-Gamal, S. and Benmokrane, B., (2009), "Long-Term Deflection of
GFRP Reinforced Concrete Beams under Uniform Sustained Load," Proceedings of
the 38th CSCE Annual Conference, St. John's, Newfoundland-Labrador, Canada, May
27-30, (CD-ROM), lOp.
Results from other research work

10. ISIS Canada (2006), "Durability of Fibre Reinforced Polymers (FRP) in Civil
Infrastructure," Durability Monograph, The Canadian Network of Centres of
Excellence on Intelligent Sensing for Innovative Structures, ISIS Canada, University
of Winnipeg, Manitoba, pp. 3.1-3.27.
11. Bouguerra, K., Youssef, T., and Benmokrane, B., (2007), "Durability of GFRP
Internal
Reinforcement:
Proceedings
Mechanical
of the Third
Properties
International
after
Moisture
Conference on Durability
Absorption,"
and
Field
Applications of FRP Composites for Construction (CDCC'2007)," Quebec City,

Quebec, Canada, May 22-24, pp. 557-566.
viii
Bibliography
Technical Reports:
12. Youssef, T., (2006), "Creep Testing of FRP Bars and FRP Reinforced Beams under
the Combined Effect of Sustained Load and Environmental Conditioning," PhD
Definition of Research Project - Internal Document, University of Sherbrooke,
Sherbrooke, Quebec , Canada, October, 59p.
13. Youssef, T. and Benmokrane, B., (2006), "Mechanical and Physical Property Testing
of Schock's FRP Rebar," Technical Report, University of Sherbrooke, Sherbrooke,
Quebec, March 17, 20p.
14. Youssef, T. and Cousin, P., (2008-2009), "A Series of Technical Reports on the Creep
Behaviour of Asiatic GFRP Bars," Technical Reports 1 to 5, University of Sherbrooke,
Sherbrooke, Quebec.
ix
Acknowledgement
ACKNOWLEDGEMENTS
The author yields thanks to Dr. Brahim Benmokrane, NSERC Research Chair Professor in
Innovative FRP Composites for Infrastructures (PhD supervisor); Dr. Ehab El-Salakawy,
Canada Research Chair Professor in Advanced FRP Composite Materials and Monitoring of
Civil Infrastructures, Department of Civil Engineering, University of Manitoba, for his effort
as co-supervisor at the inception of this program and Dr. Sherif El-Gamal, former Postdoctoral fellow at the Department of Civil Engineering, University of Sherbrooke.
The author would also like to thank the structural laboratory technical staff at the Department
of Civil Engineering at the University of Sherbrooke, especially, Mr. F r a n c i s Ntacorigira and
Nicolas Simard for their help in constructing and testing the specimens.
The financial support received from the Natural Sciences and Engineering Research Council
of Canada (NSERC), Fonds quebecois de la recherche sur la nature et les technologies
(FQRNT), Pultrall Inc. (Thetford Mines, Quebec, Canada), the Ministry of Transportation of
Quebec (MTQ), the Network of Centres of Excellence on the Intelligent Sensing of Innovative
Structures (ISIS Canada), and the University of Sherbrooke is greatly acknowledged.
Sincere gratitude and respect is given to Dr. Hussein Abdel-Baky, Dr. Ehab Ahmed, Dr.
Walid Abdel-Raouf, Dr. Kheireddine Bouguerra, Dr. Ahmed Godat, and Mr. Wael Asem
(PhD Candidate) for their genuine help while conducting this PhD effort.
The patience, love, support, and encouragement of my parents, siblings and extended family
cannot be praised enough; to them this thesis is dedicated.
Table of Contents
TABLE OF CONTENTS
ABSTRACT
RESUME
iv
BIBLIOGRAPHY
vii
ACKNOWLEDGEMENT
TABLE OF CONTENTS
xi
LIST OF FIGURES
xv
LIST OF TABLES
xx
NOMENCLATURE
CHAPTER 1
xxii
INTRODUCTION
1.1
Background and Problem Definition
1.2
Obj ectives and Originality
1.3
Methodology
1.4
Structure of the Thesis
CHAPTER 2
LITERATURE REVIEW
2.1
Background
2.2
Creep of FRP Composites and FRP Components
2.2.1
Composition of FRP Materials
2.2.2
Creep of Resin (polymer matrix)
12
2.2.3
Factors Affecting the Creep of FRP
13
2.2.4
Creep rupture stress limits for FRP
18
2.3
Creep and Shrinkage of Concrete
20
2.3.1
Classification of Different Types of Time-dependent Strains
20
2.3.2
Theories of Creep of Hardened Cement Paste
24
2.3.3
Factors Affecting the Creep of Concrete
26
2.3.4
Methods of Modelling the Creep of Concrete
32
2.4
Modelling the Deflection Behaviour and Crack Width Evolution of FRP Reinforced
Concrete Beams
2.4.1
36
Theoretical Prediction of Immediate Deflection
xi
36
Table of Contents
2.4.2
Theoretical Prediction of Long-term Deflection
2.4.3
Deflection Prediction (Immediate and Long-term) in Codes and Guidelines.. 44
2.4.4
Permissible Computed Deflections and Deflection Control (FRP-RC Beams)47
2.4.5
Effect of Sustained Loads on Flexural Crack Width in FRP-RC Beams
2.5
38
50
Long-term Behaviour of FRP Reinforced Concrete Beams under Sustained Load.. 52
2.5.1
Long-term Behaviour of FRP-RC Beams subjected to 2-point Loading
2.5.2
Long-term Behaviour of FRP-RC Beams under Uniform Distributed Load... 57
2.6
Summary
CHAPTER 3
52
59
CREEP BEHAVIOUR AND RESIDUAL PROPERTIES OF GFRP

BARS EXHIBITING DIFFERENT LEVELS OF SUSTAINED
LOAD
61
3.1
Introduction
61
3.2
Research Purpose
66
3.3
Experimental program
66
3.3.1
Materials
66
3.3.2
Study Parameters
67
3.3.3
Fabrication, Instrumentation and Installation of Samples
70
3.3.4
Long-term Monitoring
72
3.4
Results and Discussion
73
3.4.1
Creep Tensile Strain
73
3.4.2
Residual Tensile Properties (Strength and Young's Modulus)
77
3.4.3
Microstructural Analysis
81
3.4.4
Pitfalls in Creep Testing
82
3.5
Summary and Conclusion
CHAPTER 4
84
LONG TERM PERFORMANCE OF DIFFERENT TYPES OF GFRP

BARS UNDER SUSTAINED SERVICE LOAD
86
4.1
Introduction
86
4.2
Research Purpose
86
4.3
86
4.3.1
Materials
86
4.3.2
Study Parameters
91
xii
Table of Contents
4.3.3
91
4.3.4
93
4.4
94
4.4.1
94
4.4.2
99
4.4.3
4.5
104
CHAPTER 5
105
LONG TERM PERFORMANCE OF THIRD-POINT LOADED FRPREINFORCED CONCRETE BEAMS AFTER ONE YEAR OF
CONTINUOUS LOADING
.107
5.1
Introduction
107
5.2
Research Purpose
113
5.3
Experimental Program
114
5.3.1
Materials
114
5.3.2
Sample Description
118
5.3.3
Instrumentation...
118
5.3.4
Sustained Load Frame Calibration and Beam Installation
118
5.3.5
Design of FRP and Steel Reinforced Concrete Members
124
5.3.6
Long-term Test Measurements and Monitoring
125
5.4
126
5.4.1
Creep Induced Strain in FRP Reinforcement and Compression Concrete
126
5.4.2
Deflection of Beam Elements
136
5.4.3
Crack Width Evolution for Beam Elements
144
5.5
Conclusions
CHAPTER 6
155
MODELLING THE LONG TERM DEFLECTION OF FULL-SCALE

GFRP-REINFORCED CONCRETE BEAMS UNDER UNIFORM
DISTRIBUTED LOAD
157
6.1
Introduction
157
6.2
Research Purpose
159
6.3
159
6.3.1
Materials
159
xiii
Table of Contents
6.3.2
Description of Test Specimens
161
6.3.3
Instrumentation and Test Setup Description
162
6.3.4
Design of FRP Reinforced Concrete Members
165
6.3.5
166
6.4
167
6.4.1
Creep Induced Strain in Compression Concrete and FRP Reinforcement
167
6.4.2
DeflectionofBeam Elements
171
6.4.3
182
6.5
Conclusions
CHAPTER 7
188
CONCLUSIONS AND RECOMMENDATIONS
190
7.1
Summary
190
7.2
Conclusions
191
7.2.1
Long-term Performance of GFRP Bars under Axial Sustained Load
7.2.2
Long-term Performance of Third-Point Loaded FRP-RC Beams After One

Year of Continuous loading
7.2.3
193
Long-term Performance of Full-Scale FRP Reinforced Beams under Sustained

Uniform Distributed Load for a Six Month Duration
7.3
Recommendations for Future Work
CHAPTER 8
191
194
196
CONCLUSIONS ET RECOMMANDATIONS
197
8.1
Resume
197
8.2
Conclusions
198
8.2.1
Performance a long terme de barres en PRFV sous charge axiale soutenue.. 198
8.2.2
Performance a long terme des poutres renforcees en PRF avec des charges aux
tiers de la portee pour une duree d'un an sous charge soutenue
8.2.3
Performance a long terme des poutres a pleine echelle renforcees en PRFV

sous charge soutenue pour une duree de six mois
8.3
200
Recommandations pour des travaux futurs
xiv
202
203
List of Figures
LIST OF FIGURES
Figure 2.1: Basic material components of an FRP composite (ISIS Canada Educational
Module 2, 2003)
10
Figure 2.2: Stress-strain relationships for fibrous reinforcement and matrix (ISIS Canada
Design Manual No.3 2007)
11
Figure 2.3: Properties of different fibres and typical reinforcing steel
11
Figure 2.4: Schematic of creep strain and recovery strain in polymer
12
Figure 2.5: Tensile creep curves for SMC-R25 polyester laminates (Cartner et al., 1978)
14
Figure 2.6: Comparison of creep curves for (45) and (90/45) laminates (Sturgeon 1978).. 15
Figure 2.7: Coupled effect of applied stress and environment on failure mechanism
(schematic)
16
Figure 2.8: Potentially harmful effects of FRP materials in civil engineering application (ISIS
Canada Educational Module 8, 2006)
17
Figure 2.9: Degradation of FRP material due to harsh (alkaline) environmental conditions
(schematic)
18
Figure 2.10: Typical strain history curve during creep deformation (fib 2007)
21
Figure 2.11: Creep and shrinkage strains in concrete subjected to sustained load (Neville,
1996)
23
Figure 2.12: Schematic illustration of the evolution of the deformation of a concrete loaded by
constant sustained load, then unloaded (Byfors, 1980)
24
Figure 2.13: Creep of mortar specimens cured and stored continuously at different humidities.
(Neville, 1959)
27
Figure 2.14: Stress-strain-time relationship. (Riisch, 1960)
28
Figure 2.15: Relation between the creep coefficient and volume/surface (Hansen 1966)
29
Figure 2.16: Creep of concrete cured in fog for 28 days, then loaded and stored at different. 30
Figure 2.17: Flowchart for calculating deformations of GFRP reinforced concrete beams .... 41
Figure 2.18: Moment-curvature (M-K) relation for FRP reinforced concrete (CSA S806-02) 44
Figure 2.19: Average multiplier curves for normal strength concrete (top) and high strength
concrete (bottom) (Gross et al. 2003)
54
Figure 2.20: Comparison of experimental and predicted long-term deflections for two GFRP
reinforced beams (Hall and Ghali 2000)
55
xv
List of Figures
Figure 2.21: Longitudinal and side views of concrete beams reinforced with GFRP bars under
sustained load (Vijay and GangaRao 1998)
56
Figure 2.22: Test setup for beams (Arockiasamy et al. 2000)
57
Figure 2.23: Comparison of experimental and predicted deflections for beams B3 (left) and B4
(right) (Arockiasamy et al. 2000)
58
Figure 3.1: University of Sherbrooke FRP durability facility (1)
62
63
Figure 3.3: Typical strain history curve during creep deformation
64
(schematic)
65
Figure 3.5: GFRP bars
67
Figure 3.6: Specimens ready for installation
70
Figure 3.7: Schematic of bar sustained load frame (magnifying frame)
71
Figure 3.8: Regular levelling of lever arms
12
Figure 3.9: Upper bound creep strain evolution in GFRP-1 bars (above) and GFRP-2 bars
(below) under different sustained load levels
74
Figure 3.10: Residual tensile strength for GFRP-1 bars (above) and GFRP-2 bars (below) after
the 10000 hour duration
78
Figure 3.11: Magnified samples' cross section after exhibiting 10000 hours of loading
81
87
91
Figure 4.3: Sustained load frames comprising GFRP bars
92
Figure 4.4: Magnifying sustained load frame (schematic)
92
Figure 4.5: P-3500 portable strain indicator
93
Figure 4.6: Creep strain evolution for samples at 15 % fu,ave-
9.5 mm GFRP bars (above); 12
mm and 12.7 mm GFRP bars (middle); 15.9 mm GFRP bars (below)

Figure 4.7: Creep strain evolution for 9.5 mm GFRP bars (above) at 30
95
%fu,ave',
12 mm and
12.7 mm GFRP bars at 25 %fu,ave (middle); 15.9 mm GFRP bars at 30 %fu>ave (below)
95
Figure 4.8: Residual tensile strength for samples that exhibited 15 % fu ave (above); Residual
tensile strength for samples that exhibited 25 and/or 30
%fu,ave
(below)
Figure 4.9: Enlarged samples' cross section before applying sustained load
xvi
99
104
List of Figures
Figure 4.10: Enlarged samples' cross section after exhibiting 10000 hours of loading under
different sustained load levels
105
Figure 5.1: Moment-curvature (M-K) relation for FRP reinforced concrete (CSA S806-02) 109
Figure 5.2. GFRP bars (left) and CFRP bars (right) examined in the current study
115
Figure 5.3. Typical beam sample instrumentation (left) and cross section (right)
118
Figure 5.4. Schematic of sustained load frame
119
Figure 5.5. Calibration of sustained load frame
120
Figure 5.6. Beams in comprising sustained load frames
121
Figure 5.7,. Long term deflection measurement
126
Figure 5.8. Deformometer (P-3500 Portable strain indicator) for strain measurement
126
Figure 5.9. Creep-strain evolution (Frame 1) of GFRP-1 reinforcement (left) and creep-strain
evolution of concrete (right)
127
127
128
128
129
129
130
Figure 5.16. Creep-strain evolution (Frame 8) of CFRP-1 reinforcement (left) and creep-strain
130
131
Figure 5.18. Creep-strain evolution (Frame 10) of steel reinforcement (left) and creep-strain
131
Figure 5.19: Stress/Strain distribution at a cross-section
134
xvii
List of Figures
Figure 5.20. Time dependent deflection (Frame 1 and Frame 2) of GFRP-1 reinforcement
(The result of two sets/pairs)
137
Figure 5.21. Time dependent deflection (Frame 3) of GFRP-2 reinforcement
137
(left) and GFRP-4 reinforcement (right)
138
138
Figure 5.24. Time dependent deflection (Frame 8 and Frame 9) of CFRP-1 reinforcement
(left) and CFRP-2 reinforcement (right)
139
Figure 5.25. Time dependent deflection (Frame 10) of steel reinforced concrete beams
139
Figure 5.26. Time crack width test results for GFRP-1 reinforced beams (The result of two
sets/pairs)
144
Figure 5.27. Time crack width test results for GFRP-2 reinforced beams
145
Figure 5.28. Time crack width test results for GFRP-3 reinforced beams (left) and GFRP-4
reinforced beams (right)
145
146
Figure 5.30. Time crack width test results for CFRP-1 reinforced beams (left) and CFRP-2
146
Figure 5.31. Time crack width test results for Steel reinforced beams
147
Figure 5.32. Typical crack pattern for FRP reinforced concrete beams at three time intervals
(GFRP-5)
150
Figure 5.33: Crack width growth over 1 year of sustained loading
151
Figure 5.34: Crack width growth over one year of sustained loading as a multiple of initial
crack width
151
Figure 5.35: Measured versus calculated initial crack widths using various fa values
152
Figure 5.36: Measured versus calculated initial crack widths using fa values (1.4 and 1.6). 153
Figure 5.37: Measured versus calculated crack widths after one year using optimal fa and kt
values
154
Figure 5.38: Calculated bond coefficient kb for the different types of FRP reinforcement... 154
Figure 6.1: GFRP-1 and steel reinforcing bars
xviii
160
List of Figures
Figure 6.2: Reinforcement details of the beams
162
Figure 6.3: Reinforcement cages prior to casting
163
Figure 6.4: Transportation of concrete blocks to laboratory
163
Figure 6.5: Forklift truck loading blocks on beams
163
Figure 6.6: Instrumented beams under uniform sustained block load
164
Figure 6.7: Deformometer (P-3500 Portable strain indicator) for strain measurement
166
Figure 6.8: Crack detection and crack width measurement (high precision microscope)
166
Figure 6.9: Data acquisition system; LVDT connected to data acquisition system
167
Figure 6.10: Strain evolution for FRP reinforcement at midspan for beams A, B, C and D.. 168
Figure 6.11: Creep-strain evolution of compression concrete for beams A, B, C and D
169
Figure 6.12: The stress strain distribution at a cross-section
170
Figure 6.13: Deflection results for all beams
172
Figure 6.14: Moment-curvature (M-K) relation for FRP reinforced concrete
174
Figure 6.15: Flowchart for calculating the deformations of GFRP-RC beams
176
Figure 6.16: Experimental versus empirical and numerical deflection results (Beam A)
179
Figure 6.17: Experimental versus empirical and numerical deflection results (Beam B)
179
Figure 6.18: Experimental versus empirical and numerical deflection results (Beam C)
180
Figure 6.19: Experimental versus empirical and numerical deflection results (Beam D)
180
Figure 6.20: Maximum crack width evolution for beams A, B, C and D
182
Figure 6.21: Crack propagation pattern after 1 month, 3 months and 6 months (Beam A)... 183
Figure 6.22: Crack propagation pattern after 1 month, 3 months and 6 months (Beam B)... 183
Figure 6.23: Crack propagation pattern after 1 month, 3 months and 6 months (Beam C)... 184
Figure 6.24: Crack propagation pattern after 1 month, 3 months and 6 months (Beam D)... 184
Figure 6.25: Crack width growth over 6 months of sustained loading
185
Figure 6.26: Measured versus calculated initial crack widths
187
Figure 6.27: Measured versus calculated crack widths after 6 months of sustained load
187
xix
List of Tables
LIST OF TABLES
Table 2.1: Maximum allowable stress in FRP bars (CAN/CSA S6-06)
19
Table 2.2: FRP-reinforcement creep rupture stress limits (ACI 440.1R-06)
19
Table 2.3: FRP-reinforcement creep rupture stress limits (CNR-DT 203/2006)
19
Table 2.4: Sustained stress limits as a percentage of design strength (fib Bullettin 40)
20
Table 2.5: Maximum deflection formulas using the deflection approach (ISIS Canada Design
Manual 2007)
45
Table 2.6: Values for coefficients ft and m (Eurocode 2 and Model Code 1990)
47
Table 2.7: Recommended minimum thickness of non-prestressed beams or one way slabs... 47
Table 2.8: Maximum permissible computed deflections
48
Table 3.1: Mechanical properties of the two types of GFRP as provided by manufacturer and
as obtained through tensile tests
69
Table 3.2: GFRP-1 creep-test details (15 %, 30 %, 45 % and 60 %/ H , aw )
75
Table 3.3: GFRP-2 creep-test details (15 %, 30 % and 45 %fu,me)
76
Table 3.4: GFRP-2 creep-test details (60 %/, ave )
76
Table 3.5: Residual tensile property results for GFRP-1 samples
79
Table 3.6: Residual tensile property results for GFRP-2 samples
80
Table 4.1: Physical properties of the six types of GFRP bars
88
Table 4.2: Mechanical properties of the tested GFRP bar-types indicated by manufacturers. 89
Table 4.3: Mechanical properties of the tested GFRP bar-types obtained from tensile tests... 90
Table 4.4: Creep-test details of GFRP bars subjected to 15 %fu,ave- (Group 1)
97
Table 4.5: Creep-test details of GFRP bars subjected to 25 % and 30 % fUiave. (Group 2)
98
Table 4.6: Residual tensile property results for 15 %fUiave. (Group 1) samples
100
Table 4.7: Residual tensile property results for 25 % & 30 % fu,ave (Group 2) samples
102
Table 5.1. Physical properties of FRP bars (GFRP and CFRP)*
116
Table 5.2. Mechanical properties of FRP bar-types as obtained from tensile tests
117
Table 5.3. Tested concrete beams and relevant details
122
Table 5.4. Details on strain evolution within reinforcement
132
Table 5.5. Details on strain evolution of top concrete fibre
133
Table 5.6: Experimentally obtained initial deflection and long-term deflection values
140
xx
List of Tables
Table 5.7: Theoretical calculation of immediate deflections using different approaches
141
Table 5.8. Long-term deflection results compared North American standard predictions.... 143
Table 5.9: Average crack widths for tested beams
149
Table 5.10: Average calculated kb bond coefficient for the different types of FRP bars
155
Table 6.1: Mechanical properties of the reinforcing bars
160
Table 6.2: Physical properties of GFRP bars
161
Table 6.3: Test specimens and relevant details
164
Table 6.4: Details on creep strain evolution within reinforcement (at mid-span)
168
Table 6.5: Details on creep strain evolution of compression concrete (at mid-span)
169
Table 6.6: Comparison between strain increase due to flexural loading and axial sustained
load
171
Table 6.7: Immediate deflections (experimental results versus theoretical predictions)
175
Table 6.8: Comparison of long-term deflection results with ACI 440.1R-06 and CAN/CSA
S806-02 predictions
175
185
xxi
List of Tables
NOMENCLATURE
a
Depth of equivalent rectangular stress block, (mm)
Ac
Area of concrete in compression (mm 2 )
A'
Area of the age-adjusted transformed section (mm )
Widths of rectangular cross section, (mm)
The ratio of the distance from the neutral axis to extreme tensile fibre
and distance from neutral axis to centre of the tensile reinforcement =
(h - c)/(d - c)
Pd
Reduction factor used in calculating deflection
Distance from extreme compression fibre to neutral axis, (mm)
Ce
Environmental reduction factor for various fibre type and exposure

conditions.
Distance from extreme compression fibre to centroid of tension

reinforcement, (mm)
dc
Thickness of concrete cover measured from extreme tension fibre to

centre of bar or wire location closest thereto, (mm)
Ec
Modulus of elasticity of concrete, (MPa)
Ee
Age-adjusted effective modulus
Ef
Design or guaranteed modulus of elasticity of FRP defined as mean

modulus of sample of test specimens (E/= E/,ave), (MPa)
Efiave
Average modulus of elasticity of FRP, (MPa)
Es
Modulus of elasticity of steel, (MPa)
Ec
Modulus of elasticity of concrete, (MPa)
E'c(t, t 0 )
Age adjusted modulus of elasticity of concrete, (MPa)
Strain in FRP reinforcement, (microstrains)
frp,0
Initial strain in FRP due to sustained load, (microstrains)
Zfu
Design rupture strain in FRP reinforcement, (microstrains)
Guaranteed rupture strain in FRP reinforcement (microstrains)
fu
Creep strain, (microstrains)

cu
Ultimate strain in concrete, (microstrains)
cs
Shrinkage strain, (microstrains)
Ei
Initial strain, (microstrains)
ER
Strain in reinforcement, (microstrains)
u,ave
Average ultimate tensile strain for FRP, (microstrains)
xxii
List of Tables
Concrete compressive strength, in (MPa)
ff
Stress in FRP reinforcement in tension, (MPa)
ffs
Creep rupture stress limit, (MPa)
fu.ave
Average ultimate tensile stress, (MPa)
fju
Design tensile stress, (MPa)
f fu
Guaranteed tensile stress, (MPa)
Height of beam cross section, (mm)
/'
Moment of inertia of a the age adjusted transformed section using the

modular ratio 77'(t, t 0 ), (mm 4 )
Ic
Moment of inertia of a the concrete about the centroid after age

adjustment, (mm 4 )
ICr
Moment of inertia of a cracked transformed concrete section, (mm 4 )
Ie
Effective moment of inertia, (mm 4 )
Ig
Moment of inertia of a gross uncracked concrete section, (mm 4 )
kb
Bond-dependent coefficient
kt
Multiplier to account for crack width increase with time
Length of span (between two supports) (mm)
Modular ratio for reinforcement, (Efrp/ Es)
Ma
Maximum actual applied moment, (KN.m)
Mcr
Cracking moment, (KN.m)
Mcr
Nominal moment capacity, (KN.m)
M^
Moment due to sustained load, (KN.m)
rif
Modular ratio between FRP and concrete
RH
Relative humidity
Spacing between reinforcement (mm)
Crack width (mm)
yc
Distance between centroids of the effective concrete area and the ageadjusted transformed area, (mm)
dcp+sh
Additional deflection due to creep and shrinkage, (mm)
Aj
Initial deflection, (mm)
(A ,')SUs
Initial deflection due to sustained load, (mm)
(A/l)max
Limiting deflection-span ratio
<pu
Ultimate creep coefficient of concrete
xxiii
List of Tables
4>(t,to)
Creep coefficient
<pcorr
Correction factor for creep coefficient
(p.
Creep coefficient at time t
ri'(t, t 0 )
Modular ratio of modulus of elasticity of GFRP to the age adjusted

modulus of elasticity of concrete
Reinforcement ratio
p'
Upper reinforcement ratio
Pf ip/rp)
Reinforcement ratio of FRP
Long-term deflection multiplier
if/
Curvature (rad)
X (1,1.)
Aging coefficient function; % = 0.8 is normally used for concrete
Time dependent factor for computing deflection
xxiv
Chapter 1: Introduction
CHAPTER 1
INTRODUCTION
1.1
Background and Problem Definition
Fibre reinforced polymer (FRP) materials are gaining wider acceptance for use as primary
reinforcement in concrete structures. Due to its high strength and non corrosive nature, FRP
provides an alternative to steel reinforcement. The use of FRP as structural reinforcement, in
turn, provides the potential advantage of lowered maintenance costs and extended service life
for several types of structures, including bridge deck slabs, abutments, walls and other
structures exposed to corrosive environments (Gross et al. 2003). In the past two decades, an
abundance of research has taken place on fibre reinforced polymer reinforced concrete (FRPRC). A better understanding is how available on paramount characteristics such as strength,
stiffness, bending, shear and FRP-concrete bond.
Nevertheless, the area of durability of FRP remains needy for information; mainly due to the
time consuming nature and/or costly equipment necessary for durability experiments. The
ISIS Canada Durability Monograph (2006) defines durability for FRP as "the ability of an
FRP element to resist cracking, oxidation, chemical degradation, delamination, wear and/or
the effects of foreign object damage for a specified period of time, under specified loading
conditions". As regards the durability of FRP materials, the commonly known adversities can
be categorized under two branches, physical effects and environmental effects. The former
includes sustained load (creep) and fatigue, whereas the latter includes moisture, alkalinity,
fire, freeze-thawcycles and UV rays. The negative influence on an FRP bar can be due to a
combination of any of the former adversities. Sustained load (creep), however, remains a
common aspect in most if not all FRP reinforced structures; the typical purpose of designing
such structures is to withstand sustained load and resist deterioration.
Under sustained load, FRP bars suffer plastic (permanent) deformation, typically occurring
under unfavourable environments over a long time. This phenomenon is what is commonly
referred to as "Creep". Creep typically increases the long term deflection of FRP reinforced
concrete elements and may, under certain circumstances, cause catastrophic failure. Despite
its higher tensile strength over conventional steel, FRP exhibits less tensile and shear stiffness.
As a result of the relatively lower axial stiffness of the FRP bars, FRP reinforced concrete
members deform more than their steel reinforced counterparts. Consequently the design of
FRP reinforced concrete members is predominantly governed by serviceability requirements
(Gaona 2004). Thus, allowable deflections and cracking govern the design of FRP reinforced
concrete elements, particularly GFRP reinforced concrete elements.
In 2003, a protocol was set to remedy this lack-of-information drawback, where seven
subcommittees were assembled; each of which is to focus on a particular FRP adversity
(Karbhari et al. 2003). Research priority was given to the following seven conditions:
moisture/solution, alkali, thermal (including temperature cycling and freeze-thaw), creep,
fatigue, ultraviolet, and fire. For this study the creep phenomenon within FRP bars was
chosen; the aforementioned synopsis ranks this phenomenon as highly critical yet the relevant
data, to date, is sparse and questionable. Despite continuous research, the information content
regarding this issue did not increase significantly.
Furthermore, it has become a necessity to ratify standard test methods by which the
international research community can recommend to civil engineers to use as a basis for
selecting the FRP rebar appropriate for a certain application. With more manufacturers getting
into the FRP-production industry, certifying such products according to the sought standard
test methods is also of great importance.
1.2
Objectives and Originality
Having mentioned that the information content associated with the long-term behaviour of
FRP reinforced concrete under sustained load is scarce, the main objective of this study is to
further the knowledge in this particular research domain. The long-term behaviour of the
material, itself, under sustained load and the overall long-term behaviour of the concrete
element comprising FRP bars are the main focus of this study.
Keeping this focus in mind, the following objectives are addressed in this study:
To perform a critical review on the available contributions related to the study of FRP
creep due to imposed stress and adverse environmental conditions. A similar review is
necessary regarding the long-term deflection, strain variation and crack width
evolution of FRP-reinforced concrete members.
To conduct long-term (creep) experiments on the currently available commercial FRP

bars under different sustained load levels. Through the obtained information, the
available codes and guidelines can be modified or confirmed appropriate as regards the
indicated creep rupture stress limits.
Long-term (creep) data for FRP bars is inherently noisy; means of ameliorating FRP
creep tests are necessary.
To study the performance of FRP-reinforced concrete beams under sustained load for
long-term duration, in terms of serviceability; there is yet a lack of information for
different types of sustained load and different types of commercial bars.
To calibrate and compare between the models available on immediate and long-term
deflection of FRP-reinforced beams in light of the data obtained. Numerical modelling
is to be conducted to gain insight into the mechanism beyond the long-term behaviour
of an FRP reinforced concrete section.
1.3
Methodology
To achieve the aforementioned objectives, an extensive experimental program, followed by

microstructural analysis as well as empirical and numerical modelling, was conducted. The
experimental program included the preparation, installation and long term monitoring of 99
GFRP bar samples and 24 FRP reinforced concrete beams. The materials^ testing frames and
instrumentation, used for this study are mentioned below:
Materials:
(i)
GFRP-1: Sand-coated glass FRP bars of diameter = 9.5 mm and fibre content
ratio = 56.8 % by volume.
(ii)
GFRP-2: Helically wrapped sand-coated glass FRP of diameter = 9.5 mm and

fibre content ratio = 50.6 % by volume.
(iii) GFRP-3: Sand-coated glass FRP bars of diameter = 12.7 mm and fibre
content ratio = 64.5 % by volume.
(iv) GFRP-4: Ribbed glass FRP bars of diameter = 12 mm and fibre content ratio
= 75.2 % by volume.
(v)
GFRP-5: Sand-coated glass FRP bars of diameter = 15.9 mm and fibre

(vi) GFRP-6: Helically wrapped sand-coated glass FRP of diameter = 15.9 mm

and fibre content ratio = 57.5 % by volume.
(vii) CFRP-1: Sand-coated carbon FRP bars of diameter = 9.5 mm and fibre
(viii) CFRP-2: Helically wrapped (textured) carbon FRP of diameter = 9.5 mm
(ix) Conventional 15 M steel of diameter = 1 6 mm.
(x)
Conventional 10 M steel of diameter = 11.3 mm.
(xi) Ordinary concrete with a target concrete compressive strength = 35 MPa.
Means of maintaining sustained load:

(i)
Three series of bar sustained load frames: Series 1 contains 20 frames of

inner clear height = 1206 mm; Series 2 contains 24 frames of clear height =
1486 mm; Series 3: contains 20 frames (in heating tunnel) with a clear height
of 1486 mm.
(ii)
Beam sustained load rigs: Ten sustained load rigs, each with the capacity of
two mirror beams of dimensions (100 m x 150 mm x 1800 mm)
(iii) Four full-scale beams: Steel supports and concrete blocks (of dimensions 610
mm x 762 mm x 1219 mm and weight = 13334 kN) to simulate sustained
uniform distributed load.
Instrumentaion:
(i)
A data acquisition system comprising 60 ports connected to strain gauge

wires for regular strain measurement.
(ii)
A P-3500 portable strain indicator for discrete strain gauge measurements.
(iii) Electrical strain gauges of 10 mm length and
120 ohm resistance
manufactured by Kyowa limited.

(iv) Electrical strain gauges of 67 mm length and 120 ohm resistance
manufactured by Kyowa limited.
(v)
A high precision vernier caliper for deflection reading.
(vi) Demec points to be fixed on concrete beams to serve as references for

deflection reading.
(vii) A high precision microscope for crack width reading.
Analysis:
(i)
Numerical (finite difference) modelling is performed using a Fortran-2003

environment, using a system of equations (based on the age-adjusted effective
modulus method).
(ii)
Empirical (simplified) equations are developed through regular linear or nonlinear regression using Microsoft Excel.
1.4
Structure of the Thesis
The thesis comprises eight chapters, four of which (Chapters 3 to 6) represent stand-alone
studies that include their own literature review, set of objectives, results, analysis, discussion
and conclusion. A portion of information may be repeated in more than one chapter to give
each chapter/study an independent sense and convenience for the reader. The chapters are
interdependent, serving together towards the goal of better understanding the long term
behaviour of FRP reinforced concrete. The following is a brief description of each chapter's
content:
Chapter 1: This chapter includes (i) the background and problem definition of the thesis, (ii)
research objectives and originality, (iii) methodology and (iv) thesis structure.
Chapter 2\ In this chapter, a detailed literature review is conducted on (i) the creep behaviour
of materials comprising FRP-RC (i.e., FRP reinforcement and concrete) and the factors
affecting them; (ii) the theoretical models available regarding immediate and long-term
deflection of FRP-RC beams; (iii) previous efforts related to the topic of long-term deflection
of FRP-RC beams. Selected parts of this literature are made available, within chapters 3 to 6,
to provide convenience for the reader while going through the material of this thesis.
Chapter 3 : The creep behaviour and susceptibility to creep rupture of two types of GFRP bars
is studied. Different levels of sustained axial load are applied to 37 GFRP samples. Regular
monitoring of the creep strain evolution of the bars took place over a 10000 hour (417 days)
duration. This was followed by residual tensile tests and microstructural analysis. The creep
coefficient of GFRP bars was calculated.
Chapter 4\ The creep behaviour of six different types of GFRP bars was studied under two
levels of allowable service load, according to the currently available North American
standards. The test duration, for the two phases, exceeded 10000 hours (417 days) wherein
regular monitoring of creep strain evolution took place and the creep coefficient of GFRP bars
was calculated. This was followed by residual tensile tests and microstructural analysis. The
creep coefficient of GFRP bars was calculated.
Chapter 5: The long term performance of twenty concrete beams (ten pairs) comprising
GFRP, CFRP, and steel reinforcing bars was studied in terms of deflection, strain increase and
crack width evolution. The beams of dimensions 100 mm x 150 mm x 1800 mm were
installed under third-point concentrated load, for a period exceeding one year. Exhibiting a
maximum applied moment of 25 % of their nominal moment capacity, Mn, all beams were
regularly monitored for the one-year duration. Theoretical predictions for immediate
deflection were calculated and compared to the obtained experimental results. The long-term
to immediate deflection ratio, 2, was calculated for all beams and compared to ACI 440.1R-06
and CAN/CSA S806-02 predictions.
Chapter 6. Four full-scale GFRP reinforced concrete beams, of dimensions (215 mm x 400
mm x 4282 mm) were subjected to uniform distributed load for a period of six months. The
main study parameters are (i) bottom reinforcement ratio and (ii) type of upper/compression
reinforcement (GFRP and/or steel). The maximum applied moment ranged from 15 to 21 % of
the nominal moment capacity for the beams. Numerical modelling took place using a
predict the long-term deflection of the beams. The creep and shrinkage coefficients were
calculated based on ACI Committee 209 recommendations (1992) and CEB-FIP Model Code
(1990). The theoretical curves using different empirical models (available in CSA A.23.3-04
and ACI 318-08) as well as the numerical models mentioned above were compared to the
measured values.
Chapter
7: This chapter contains the overall conclusion for all previous chapters.
Recommendations are proposed for subsequent research studies in the same domain.
Chapter 8: This chapter contains the overall conclusion for all previous chapters (in the
French language). Recommendations are proposed for subsequent research studies in the same
domain.
Chapter 2: Literature Review
CHAPTER 2
LITERATURE REVIEW
2.1
Background
In the past two decades, an abundance of research has taken place on fibre reinforced polymer
reinforced concrete (FRP-RC). Due to that, a better understanding is now available on
paramount characteristics such as strength, stiffness, bending, shear and FRP-concrete bond.
While the behaviour of steel reinforced concrete beams under sustained loads has been studied
for nearly half a century, the area of long-term/creep behaviour of FRP reinforced concrete
beams remains barely touched. This is mainly due to the time consuming nature of creep
experiments. The sought knowledge - of long-term serviceability aspects - is extremely
important for the infrastructure facilities constructed using FRP.
It is commonly known that concrete creep, shrinkage and other factors contribute to the
increase of strain, curvature, and deflection over time. For conventional reinforced concrete,
steel reinforcement acts to restrain these effects. On the other hand, FRP reinforcement with
its lower modulus of elasticity and higher susceptibility to creep may be a concern for its
comprising concrete members, as regards serviceability. Whether it be steel or FRP
reinforcement, the concrete cross section exhibits strain in concrete within the compression
zone due to creep. This, in turn, leads to an increase in neutral axis depth and redistribution of
stresses between the concrete and reinforcement (ACI 435, 2003). Associated with the latter
time-dependent behaviour, is an increase in curvature. This curvature increase, along with the
formation of additional flexural cracks over time, leads to increase in beam deflection.
In this chapter, a thorough literature review is provided on all aspects that lead to a better
understanding of the aforementioned time dependent behaviour of FRP-RC beams under
sustained flexural load. The main topics addressed in this literature-review chapter are (i) the
creep behaviour of FRP bars (comprising fibres and polymer matrix); (ii) the creep behaviour
of concrete; (iii) the creep-rupture stress limits made available by commonly used codes and
guidelines; (iv) the long-term behaviour of FRP-RC beams under sustained load and finally
(v) the theoretical (empirical and numerical) methods of prediction available in codes and
guidelines. Nevertheless, every chapter in this study serves as a stand-alone project containing
its relevant literature. This provides convenience for the reader who seeks the information
content of a particular portion of this study.
2.2
Creep of FRP Composites and FRP Components
Creep of a bar is defined as the increase in length of a bar loaded with a constant force over
time, beyond the initial (elastic) deformation (Gere and Timoshenko 1990). In this section a
brief review is presented on the properties of FRP materials, in general, as well as the
properties of the individual components of FRP bars. The creep of the polymer (matrix) will
be emphasized upon as well as the creep of the FRP bars and the relevant influencing factors.
The tensile strength of FRP is known to exceed that of steel. Nevertheless, the stiffness of
FRP is lower than the stiffness steel. FRP materials behave essentially linearly until fracture
when loaded along the fibre direction, and thus FRP is brittle by nature. For many fibre
reinforced polymers, the thermal expansion coefficients are much lower than those of metals.
Thus, composite structures may exhibit better dimensional stability over a wide range of
temperatures. Non-corroding behaviour is another major advantage attributed to FRP
composites. Nevertheless, a disadvantage for many polymeric matrix composites is their
moisture-absorbing nature, from their surrounding environment. This induces adverse internal
stresses within the material and creates dimensional changes.
2.2.1
Composition of FRP Materials
In simple terms, composites are a combination of two or more non homogeneous materials
that do not chemically interact. This combination creates a new material that exhibits better
properties than the individual components. The combination of fibre and resin in FRP bars
yields a reinforcement of higher tensile strength than that of steel. Nevertheless, unlike steel,
FRP composites are non-isotropic, since they depend on the orientation of the fibres and the
applied load relative to the fibre orientation.
FRP materials can be classified into three main types (unidirectional, bidirectional, and bidirectional). Unidirectional FRP is the most popular and defines most of the commercial FRP
bars available. FRP composites, in general, consist of two or more distinct physical phases;
the fibrous phase impregnated in a continuous matrix phase (See Figure 2.1). The fibres carry
the load while the surrounding matrix (the resin) acts as a load transfer medium between the
fibres and keeps them in the desired location/orientation; the resin also protects the fibres from
abrasion and environmental damage (chemicals and humidity). Figure 2.2 shows the stressstrain relationship of the individual components of fibres, resin and the overall stress-strain
behaviour of FRP material.
FIBRES
POLYMER
MATRIX
FRP
Figure 2.1: Basic material components of an FRP composite (ISIS Canada Educational
Module 2, 2003)
2.2.1.1
Fibres
The most commonly used fibres in the civil engineering domain are carbon, aramid, and glass.
There are two types of carbon fibres. The first type is of high resistance (3000 to 5000 MPa),
whereas the second has a high modulus of elasticity (400 GPa) (See Figure 2.3). Aramid fibres
have very high stress at rupture, yet the use of this type is limited due to its low resistance in
compression and shear. This is attributed to the poor adherence between the fibre and the
matrix. All three types of fibres have very good creep properties and their mechanical
properties can be maintained for a temperature range that varies from 50C to 300C. Glass
fibres are the most commonly used and emphasized upon in research due to their lower cost.
Combinations/hybrids of glass fibres and carbon or aramid fibres are used, in certain
applications, to increase the stiffness of a structural member.
10
Figure 2.2: Stress-strain relationships for fibrous reinforcement and matrix (ISIS Canada
Design Manual No.3 2007)
Strain (%]
Figure 2.3: Properties of different fibres and typical reinforcing steel.
2.2.1.2
Resins
Thermoplastics and thermosets are the two main types of polymers used for resins. For the
former, the resin molecules contain rigid aromatic rings yielding them a relatively high glass
transition temperature (Tg), and dimensional stability at high temperatures. Thermosetting
resins, such as vinylesters and epoxides, are the types found more suitable for the construction
11
industry. Thermosetting resins (vinylester, epoxy and phenolic resins) are less viscous than
thermo-plastic resins (polypropylenes, nylon, polycarbohydrates, etc.).
2.2.1.3
Fillers and Additives
A polymeric matrix (the resin) may have fillers added to enhance its properties. Some of the
main advantages of filler addition are (i) increasing the modulus of elasticity, (ii) reducing
cost, (iii) reducing mould shrinkage, (iv) controlling viscosity and (v) producing smoother
surface. One of the most commonly used fillers is Calcium carbonate (CaCOs). It is used for
polyester and vinylester for reduction of cost as well as mould shrinkage.
2.2.2
Creep of Resin (polymer matrix)
The typical long-term (creep) behaviour of polymers under sustained load is shown in Figure
2.4. This behaviour is influenced by a variety of factors elaborated upon in the following subsection.
A
Constant Stress
60
.S
5
ca
J
C
' 1O3_
g_
5'
era
Time
Figure 2.4: Schematic of creep strain and recovery strain in polymer
Since a considerable magnitude of creep may occur at low temperatures, the creep of
polymers is considered a vital area. The creep of polymers is a function of temperature and
12
stress level. Creep is a result of viscoelastic deformation (the combination of elastic

deformation and viscous flow). As shown in Figure 2.4, creep strain exhibits nonlinear
increase with time. Nevertheless, the elastic deformation is immediately recovered when the
stress is waived. The accumulated deformation due to viscous flow recovers asymptotically to
a value called recovery strain. For non-reinforced polymers, large creep strain may take place
at room temperature and at low stress levels. The creep phenomenon becomes more critical at
elevated temperature or high stress levels. The commercial reinforcing bars, however, do not
exhibit significant creep strain. Nevertheless, aramid fibres and glass fibres have a greater
susceptibility to creep rupture at lower stress levels than carbon fibres (Chiao and Moore
1971).
2.2.3
Factors Affecting the Creep of FRP
The creep behaviour of FRP bars is a function of several factors; the most influential are
presented below.
2.2.3.1
Sustained-load
Magnitude
The polymeric behaviour of FRP bars is highly dependent on the loading magnitude. When
the magnitude is high, the behaviour of the reinforced polymer is rigid and adopts a brittle
manner; the sample may rupture in a short period of time (eg., the rupture of a GFRP bar
under high sustained tensile stress). Nevertheless, the same specimen may exhibit a ductile
manner and show high toughness values when the applied load is low in magnitude. The creep
behaviour of an FRP bar under uniaxial stress can be expressed by the following creep
compliance equation:
(2.1)
where s (t) is the measured strain at time t and o is constant stress level in a creep experiment.
In Figure 2.5, it is obvious how the rate and magnitude of creep increases when the applied
load increases, for identical SMC-R25 laminates tested at room temperature.
13
TIME
(WIN.)
Figure 2.5: Tensile creep curves for SMC-R25 polyester laminates (Cartner et al., 1978)
Gaona (2003) and Youssef et al. (2008) have conducted creep behaviour experiments on a
variety of GFRP bars for lengthy periods. The former study concluded that the creep strain
increase for three different types of 15.9 mm GFRP bars, subjected to sustained load levels
ranging from 23 to 27 % fUiave, is in the order of 2 to 6.6 % of initial elastic strain SfrP,o',
surrounding environment was controlled at (31C and 67 % relative humidity) for a test
duration of six months. Showing fair agreement, the latter study (Youssef et al. 2008) yields
creep strain values that range from 2 to 8.7 % efrp.o for two types of 9.5 mm commercial GFRP
bars; the test duration ranged from 4000 to 8600 hours (5.5 to 12 months approx.); sustained
load levels ranged from 15 to 45
%f
ve
at standard laboratory atmosphere (23 3 C and 50
10 % relative humidity).
Studies such as that of Budelman & Rotasy 1993 indicate that if sustained stress is less than
60 % of the average ultimate tensile strength
(fu,ave),
creep rupture is less likely to occur.
Greenwood (2002), however, deduced through series of creep-rupture testing that the creep
stress limit of two types of 6.4 mm GFRP rods, in air at 23 C, is approximately 45 %/, ave for
a 50 year structure survivability. The latter claim is supported by the findings of Youssef et al.
14
(2008) as well as the findings of the study at hand, where two commercial GFRP bars were
tested in similar surrounding environmental conditions. It was realized that there is no sign of
creep rupture at 45 % fu,ave\
a realization confirmed by microstructural analysis. On the other
hand, rupture susceptibility is evident on exhibiting 60
2.2.3.2
%fu,ave,
at different endurance times.
Fibre Content and Fibre Orientation
Since fibres barely exhibit creep under relatively high loads, the abundance of fibres in an
FRP bar significantly decreases deformation under sustained load. Nevertheless, when an FRP
sample is low in fibre content or the load is applied transversally to the fibres, the creep
phenomenon is more pronounced. With the increase of stress level, temperature and time,
creep compliance increases. Creep is also a function of the fibre orientation. For unidirectional
fibres (of orientation angle 0 = 0) creep compliance is nearly constant; creep in the
longitudinal direction of a unidirectional laminate, roving or bar is insignificant. Nevertheless,
for other fibre orientation angles, creep strain can be significant. In multidirectional FRP
rovings, creep depends on the construction of the element.
2000
10
130
1000
DURATION OF LOADING (HR.)
Figure 2.6: Comparison of creep curves for (45) and (90/45) laminates (Sturgeon 1978)
The creep-strain magnitudes vary noticeably when using (45) versus (9045) oriented fibres
(Figure 2.6); this is despite the fact that the mechanical properties of the two rovings are
almost equal. The addition of 90 layers to a 45 construction causes restraint to the
rotational tendency of 45 fibres towards the direction of loading direction, reducing creep
15
strain significantly. This concludes that fibre orientation has an impact on the creep properties
of FRP laminates. Nevertheless, creep behaviour is governed by the fibres when aligned with
the loading direction, (Dootson 1972).
2.2.3.3
Environmental
Effects
Environmental effects may have a serious impact on the durability of FRP composites,
especially when coupled with sustained or cyclic loading. The durability of FRP is defined by
Karbhari et al. (2003) as the ability of an FRP element to resist cracking, oxidation, chemical
degradation, delamination, wear and/or the effects of foreign object damage for the specific
period of time under the appropriate load conditions, under specified environmental
conditions. Figure 2.8 illustrates all adverse environmental and physical effects that may harm
an FRP element. The potential harm may be a combination of two or more adversities (See
Figure 2.7).
(schematic)
Environmental factors such as high temperature, temperature fluctuation, high humidity,

corrosive fluids and water absorption can adversely affect the behaviour of some polymer
composites. Moisture (water absorption) can reduce the strength and stiffness of some
polymeric composites by as much as 30 % compared to the original dry material. Water
absorption breaks down the interface between the reinforcing fibre and resin matrix leading to
16
a loss of strength and rigidity. Harsh environmental conditions such as high alkalinity, freezethaw, chloride exposure and degradation from ultraviolet light also have important effects on
FRP.
Environmental Effects
Figure 2.8: Potentially harmful effects of FRP materials in civil engineering application (ISIS
Canada Educational Module 8, 2006)
Masmoudi et al. (2003) studied the combined effect of alkaline environment and sustained
load on the durability of GFRP composite bars (See Figure 2.9). Elevated temperature
(between 45 to 63C) was used to accelerate the aging of bars under sustained tensile loads
(varying from 20 to 29 % of the ultimate tensile s t r e n g t h , f u , a v e ) for 104 days. The residual
tensile strength was evaluated after the test; a strength-reduction of 13 to 15 % was found.
17
Figure 2.9: Degradation of FRP material due to harsh (alkaline) environmental conditions
(schematic)
Twenty 9.5 mm GFRP bars were investigated by Nkurunziza et al. (2005) under the coupled
effect of sustained load and alkaline solution (pH = 12.8) or deionized water. Two levels of
sustained load were applied (25 to 38 % of the guaranteed tensile stress) for a period of 417
days. An increase in creep strain of 3 and 5 % of the instantaneous strain was observed.
Greenwood (2002) studied the environmental creep rupture performance for two different
types of pultruded rod samples (diameter = 6.4 mm) of similar mechanical properties. Tests
were conducted in air (to provide a baseline for comparison); in road-salt water solution and
cement-extract solution (pH = 12.8). Using the obtained data, a linear regression was
conducted and the 50 year creep rupture stress limit was extrapolated. Results showed that the
creep rupture stress limit, in air, was 0.45; the partial reduction factor for the adverse effect of
cement extract was 0.54 for one type of bar and 0.34 for the other. It is well known that if high
temperature is coupled with the adverse effect of alkalinity or moisture, the degradation rate
would be much faster. This is why temperature is typically used as an accelerated aging factor
for FRP-durability experiments.
2.2.4
Creep rupture stress limits for FRP
CAN/CSA S806-02 indicates that the tensile stress in GFRP reinforcement, under sustained
load should not exceed 0.3 (30 %) of its tensile failure stress. The Canadian highway bridge
design code (CAN/CSA S6-06), however, states that the maximum stress in FRP bars or grids
under SLS (serviceability limit state) loads shall not be more than FsLsfFRPu, where the
18
dimensionless factor
FSLS
is indicated in Table 2.1 below and
/FKPU,
is the specified tensile
strength of an FRP bar. This allowable stress value would later be multiplied by a resistance
factor <Pfrp that accounts for the manufacturing method and application.
Table 2.1: Maximum allowable stress in FRP bars (CAN/CSA S6-06)

Fibre type
Maximum allowable stress at SLS
GFRP
AFRP
CFRP
0.25 /FRPU
0.35 /FRPU
0.65 /frpu
ACI 440.1R-06 indicates the creep rupture stress limit ff iS as a fraction of the design tensile
strength of FRP, fju, that considers reduction for service environment. In other words, the
guaranteed tensile stress of the bar,/*^
=ffu,ave
-3a = mean tensile strength minus three times
the standard deviation, is multiplied by an environmental reduction factor, Ce, prior to

multiplication by the creep rupture stress limits in Table 2.2.
Table 2.2: FRP-reinforcement creep rupture stress limits (ACI 440.1R-06)

Fibre type
GFRP
AFRP
CFRP
Creep rupture stress l i m i t s f f s
0.20 jfr
0.30 ffu
0.55 ffu
As regards European recommendations for sustained stress limits, the corresponding values
for the two tables above, in the Italian guide for construction with FRP bars (CNR-DT
203/2006), are shown in Table 2.3.
Table 2.3: FRP-reinforcement creep rupture stress limits (CNR-DT 203/2006)

Loading mode
Type of fibre/matrix
m (SLS)
Glass/Vinylesters or epoxy
0.30
Aramid/Vinylesters or epoxy
0.50
Carbon/Vinylesters or epoxy
0.90
Quasi-permanent and/or cyclic

(creep, relaxation and fatigue)
The fib task group bulletin 40 states that they have not yet fully-defined the stress limits of
FRP until bar classification methods develop (FRP Classification is now available in the
19
S807-10 specifications for FRP polymers). Nevertheless, Table 2.4 shows the available stresslimits available in fib Bullettin 40 to-date:
Table 2.4: Sustained stress limits as a percentage of design strength (fib Bullettin 40)
FRP type
GFRP
Class 3 (worst case)
Class 2
Class 1
AFRP
Class 2 (worst case)
Class 1
CFRP
Class 2
Class 1
2.3
50 years
100 years
30
to come
to come
25
to come
to come
45
to come
40
to come
80
to come
75
to come
Creep and Shrinkage of Concrete
One of the main factors affecting long-term deflection is the creep of concrete. As a matter of
fact, concrete creep, is more influential than the creep of FRP rebar, as regards the overall
long-term behaviour of an FRP-RC beam. All concrete elements subjected to sustained load
are, in turn, subjected to creep. Creep is a physico-chemical phenomenon that may
significantly affect the durability, stability and serviceability of concrete structures. To explain
the mechanism of the creep process, several theories have been proposed; some of which are
presented in the current chapter. Shrinkage also plays a role in the short and long-term
behaviour of reinforced concrete elements.
2.3.1
Classification of Different Types of Time-dependent Strains
Typically, the strain-versus-time curve for any material, under sustained load, mimics the
form shown in Figure 2.10. Once the sample is loaded (i.e. at zero time), the induced strain is
elastic yet may comprise a plastic (non-elastic) component. As shown in Figure 2.10, the
creep curve has three stages. In the first stage - known as the primary creep stage - the rate of
creep evolution decreases with time. Should the rate of creep be very low, the secondary creep
phase would designate the range of steady state creep (known as stationary creep). Depending
on the level of applied stress, the tertiary creep phase may or may not exist. In concrete, the
20
Chapter 2: Literature
Review
increase of micro-cracking at high levels of stress may lead to the tertiary phase. The
summation of shrinkage strain, elastic strain, basic creep strain, and drying creep strain yield
the total time-dependent strain in concrete. Creep strain can be obtained by subtracting the
initial strain and measured shrinkage strain from the total deformation.
I ^ Tertiary _ |
\
Primary ^ |
Secondary
I
|
I-
/1 \
\
Fracture
Initial Elastic Strain
Time
Figure 2.10: Typical strain history curve during creep deformation (fib 2007)
The terms presented below are frequently found when introducing subjects concerning the
creep phenomenon and creep behaviour of concrete:
2.3.1.1
Creep
Creep is defined by Neville as the strain increase of a material with time, when subjected to
constant stress (Neville and Meyers 1964). Most if not all structural materials are subjected to
a certain level of sustained loading in their life time, mainly due to the dead load of the
structure itself.
2.3.1.2
Specific Creep
Specific creep is defined as the amount of creep per unit applied stress. It is used to compare
the creep potential of different types of concrete loaded at different stress levels. A value of
1.5><10~4 per MPa (1 xl0~ 6 per psi) is often used in the absence of a specific value.
21
2.3.1.3
Shrinkage
Shrinkage is defined as the time-dependent decrease of concrete volume, mainly due to the
change in moisture content. Shrinkage is also a function of other physico-chemical changes
that take place without external stress. Shrinkage includes: (i) Drying shrinkage due to the loss
of moisture in concrete; (ii) Autogenous shrinkage due to the hydration of cement and (iii)
carbonation shrinkage which is a result of the carbonation of various cement hydration
products in the presence of CO2. Ultimate shrinkage is the shrinkage that occurs in 20 years
(Neville and Brooks 1987); 14-34 % of shrinkage occurs in the first 2 weeks after casting
concrete; 40-80 % occurs in the first 3 months, and 66-85 % in the first year.
This paragraph presents ranges of shrinkage strain values, presented by different researchers:
Mehta (1986) indicated that the ultimate shrinkage of concrete is in the range of 500-1200
microstrains. For Nilson and Winter (1986), the ultimate shrinkage of ordinary concrete
ranged from 400-800 microstrains. The data of Troxell et al. (1958) yielded an ultimate
shrinkage value of 1100 microstrains for concrete stored at 50 % relative humidity, and 800
microstrains for concrete stored at 70 % relative humidity. Russell and Larson (1989) found
that the ultimate shrinkage of concrete of 62 MPa and 52 MPa is 800 and 700 microstrains,
respectively. The shrinkage of concrete was found almost independent of concrete strength, by
Ngab et al. (1980). The results obtained by Smadi et al. (1982) show that shrinkage is greater
for low strength concrete (23 MPa) than for medium (38 MPa) and high strength concretes (62
MPa). Yet, the shrinkage of high strength concrete may be marginally more than that of
medium strength concrete.
2.3.1.4
Elastic/Initial Strain so
The initial strain (elastic strain) is typically measured once sustained load is applied. It
depends on the loading rate. The more rapid the rate of applied load, the smaller the initial
strain and the larger the apparent modulus of elasticity (Neville 1996). It is also evident that
creep starts once loading takes place, yet it takes a period of time to take the elastic strain
measurement. For a significant number of concrete samples of this thesis, the initial/elastic
strain measurements were taken manually by a portable strain indicator (See Chapters 5 and 6)
22
2.3.1.5
Basic Creep
In the absence of moisture movements, the time-dependent strain caused by sustained stress
represents basic creep (Bazant et al. 1997). Basic creep is a function of the age of concrete
which also implies that it is a consequence of long-time chemical reactions associated with the
hydration of cement.
2.3.1.6
Drying Creep
When the same concrete subjected to sustained load is allowed to dry, an additional creep
strain results, in excess of basic creep. This creep component is called drying creep (See
Figure 2.11). In turn, the total resulting creep is the sum of basic creep and drying creep.
However, it is not easy, in practice, to separate exactly creep strain from elastic/initial strain.
Figure 2.11 illustrates the different types of creep strain in concrete.
Srmnkoge
to
Trmâ> Sfrnwojr os or unloaded specimen
nn fckssis
of A d d r t v e
Definition
Shrinkage o>
on Unloaded
Specimen
Nominal E;a c
Stroki
Tirne
b> C n o n g * in Strain of u x o e a ono
Drying Speimsr
Creep
Nomina i E *q1 c
Strain
10
Time
c> Creep Of G i <v?r?rct ^perime^ tf Hygroi
E{joMitar-njni w i t l the- A n i e n t
Medium
Drying Creep
Sosi Creep
Nominal Elostta,
Strain
Ttmo
a) Choice in Strain of o Loaded cod Drying 5p mri

Figure 2.11: Creep and shrinkage strains in concrete subjected to sustained load (Neville,
1996)
23
Creep strain can be differentiated into recoverable and irrecoverable creep strain. When
unloading a concrete specimen, the accumulated creep strain is only partially recovered,
whereas the elastic recovery is nearly complete (Yunping and Jennings 1992). Byfors (1980)
classifies creep strain, in concrete, into instantaneous elastic, delayed elastic, viscous, and
permanent strain (see Figure 2.12).
Stress ^
tl
t*
Time
Strain
CccCto)
Time
Figure 2.12: Schematic illustration of the evolution of the deformation of a concrete loaded by
constant sustained load, then unloaded (Byfors, 1980)
2.3.2
Theories of Creep of Hardened Cement Paste
The material available on the mechanism of creep and shrinkage of concrete is extensive.
Nevertheless, the chemical and physical processes involved, and their time-dependent nature,
are not yet fully comprehended (Tamtsia and Beaudoin 2000). This sub-section presents a
briefing on some of available explanations for this phenomenon.
24
Feldman (1972) justified the occurrence of creep through the gradual crystallization (aging) of
layered silicate material. This leads to the increase of the extent of layering. In Feldman's
hypothesis, the movement of water was not considered to be the major mechanism. Powers
(1968), however, states that the cause of creep is the diffusion of load-bearing water and that
the free energy of adsorbed water changes due to external load. Wittmann (1970) mentioned
that in the former two hypotheses (of Feldman and Powers), creep will cease to exist if the
hydrated cement paste fully dries out. The measurements obtained from Wittmann's effort,
though, disagreed with the latter hypothesis since the creep values were significant, for the
tested fully-dried samples. On the contrary, the samples tested by Mullen and Dolch (1964)
were oven-dried of cement paste, yet they exhibited no creep at all. In other words, the notion
that water movement is the dominant mechanism responsible for the concrete creep is yet
questionable.
On the basis of the available theory, Bazant and Prasannan (1989) indicated that the total
strain in concrete et is the sum of several components:
e = ^r + e c r +e h
(2.2)
where
cr = v +f
(2 3)
where Eo is the modulus of elasticity, a is stress (i.e. a/ Eo = elastic strain), e cr is the creep
strain and /, is the sum of the hydrothermal strain which includes drying shrinkage, thermal
dilatation chemical strain such as autogeneous shrinkage and cracking strain at high stresses.
Creep strain, however, is the sum of viscoelastic strain, ev, and viscous/flow strain, /. The
elastic strain, a/ Eo, occurs due to the deformation of mineral aggregate pieces in concrete and
microscopic elastic particles in the hardened cement paste. Due to the physco-chemical nature
of such microscopic components, their elastic properties stay constant (i.e. do not age).
Nevertheless, the elastic strain of concrete is typically taken as a/ E(t), where the elastic
modulus Eft) is a function of the age of concrete. The basis of all previous explanations, of
creep, is a variety of known theories. These theories are briefly presented below:
25
The seepage theory. This theory predicts the loss of water from the sample during basic creep.
On applying external load, the change in volume takes place due to the change in internal
vapour pressure. Powers (1968) describes the creep process where external load squeezes
some of the water out into areas of unhindered adsorption through a time-dependent diffusion
process. When this takes place, the volume of the paste decreases and the spacing between the
particles is reduced.
The interlayer theory. Proposed by Feldman and Sereda (1968), this theory states that creep
within a cement paste is the result of gradual crystallization (aging) of a poorly crystallized
layered silicate material, accelerated by drying or stress. In this theory, the adsorbed water
plays no sizeable role in the mechanism.
The adsorption theory. Ruetz (1968) proposed the adsorption theory where creep occurs
through slippage between Calcium Silicate Hydrate (C-S-H) particles in a shear process. In
this process water acts as a lubricant.
The thermal activation theory. This thermal activation theory was proposed by Wittmann
(1970). This theory indicates that water diffusion does not play a major role in the creep
mechanism; water plays an indirect role due to its effect on the disjoining pressure which
weakens the inter-particle bond. Creep takes place (increases) when the strength of these
forces is reduced and the particles slide apart with respect to each other.
2.3.3
Factors Affecting the Creep of Concrete
Concrete is a complex material where the change of one factor would almost certainly affect
another. This fact makes studying the creep of concrete, and interpreting its data, a difficult
task. Some of the main factors influencing the creep of reinforced concrete are the level of
sustained load, concrete compressive strength, member size as well as type and magnitude of
reinforcement; all of which are addressed in this study. Nevertheless, there exist other creepinfluencing factors such as humidity, type of cement, mix proportions, additives, aggregate
type ancLdegree of saturation and hydration at loading. Whether due to external conditions or
inherent properties of the concrete mix, these factors are addressed below.
26
2.3.3.1
Influence of stress & strength (sustained load and concrete compressive
strength)
Except for concrete specimens loaded at a very early age, there is a direct proportionality
between creep and applied stress. Whether subjected to low or high stress levels, creep of
concrete takes place. Thus, there is no lower limit for the occurrence of the creep of concrete.
However, the upper limit (typically between 40 to 60 % of the concrete ultimate strength) can
be determined by the appearance of severe micro-cracking (Neville 1995). Within the typical
range of stresses, for structures in service, creep and stress are linearly proportional (Bruegger
1974) (See Figure 2.13); this fact is adopted by the majority of creep expressions. At higher
stress (sustained load) to strength ratios, creep produces time failure (Figure 2.14).
1000
Ambient relative
humidity
- 95%
800
0 - 75%
A - 32%
0
r- 600
A
m/
jfo
A
1
a
u
4>
7 O
& Of
Ljr Qj
m
t
6 400
0
200
20
40
60
80
Stress/Strength Ratio-per cent
100
0
Figure 2.13: Creep of mortar specimens cured and stored continuously at different humidities.
(Neville, 1959)
27
Xi
10
^
08
2 0-6
/
/ y
"^BggU^jfeir
/ / /
to
0'4
e
^
o
a 0-2
ou
o
a
%
cc
I = Time und^r load
0 002
0-004
o ooe
Concrete strain.
JL
O'OdB
010
Figure 2.14: Stress-strain-time relationship. (Rtisch, 1960)
2.3.3.2
Concrete
Strength
The strength of concrete, f'c , has a considerable impact on creep behaviour. Creep is inversely
proportional to the strength of concrete at the time of application of load. This applies for a
wide range of results. Paulson et al. (1991) studied the long-term deflection of high-strength
concrete beams reinforced with steel. The study, based on a large body of experimental data,
confirmed that the creep coefficient of high-strength concrete under sustained load was
significantly less than that of ordinary concrete. A similar conclusion was rendered by Gross
et al. (2003), where GFRP reinforced beams (6 of normal concrete and 6 of high strength
concrete) were tested under sustained - two point - loading for 180 days; the creep coefficient
of high-strength concrete under sustained load was again significantly less than that of
ordinary concrete. In turn, the ratio of time-dependent deflection to immediate elastic
deflection of high-strength concrete beams under sustained loads was lower.
2.3.3.3
Type and Quantity of
Reinforcement
Creep time-dependent deformation is not the free creep of concrete but a value modified by
the quantity and the position of reinforcement (Neville 1995). Regarding steel reinforced
beams, it is proven that the type and amount of steel bars have an impact on the bonding force
28
of the interface between concrete and steel. According to ACI Committee 209, the creep
coefficient of steel reinforced concrete is calculated as:
(t-t
V6
10+(t-t0f6
(2.4)
where (p(t, t 0 ) is the creep coefficient at time t for age at loading to; 4>u is the ultimate creep
coefficient ranging from 1.3 to 4.15 and (typically taken as 2.35). As regards FRP reinforced
concrete, a new correction factor, (f)corr - 0.55, was proposed by Brown (1997). The creep
coefficient (Eqn. 2.4) is necessary to calculate the long-term deflection of steel or FRP
reinforced concrete through numerical modelling.
2.3.3.4
Member Size
The magnitude and rate of concrete creep is also a function of concrete member size. Moisture
and temperature gradients between the surfaces and the interior of the member, which is a
form of micro-curing, plays a role in shrinkage and creep. Creep, for specimens of big size, is
known to be less than for small scale specimens. The shrinkage effect and the occurrence of
creep at the surface of a concrete specimen, under drying conditions, may explain this
phenomenon; creep on the surface is, thus, greater than within the core of the specimen
(Bruegger 1974). Figure 2.15 illustrates the relationship between creep and the member size
(Hansen and Mattock 1966). It is generally expressed by the volume/surface ratio of the
concrete member.
Time urWerbat - days:
Uf
t*
<
* Cy irsJw
I- ihose
Specimens
SO
CO
BO
WAime -Surface Ratio - mm
200
Figure 2.15: Relation between the creep coefficient and volume/surface. (Hansen et al.1966)
29
2.3.3.5
Humidity
Neville (1996) states that at lower relative humidity levels creep increases, for a given type of
concrete (See Figure 2.16). An earlier hypothesis assumes an analogy between the creep of
concrete and the consolidation of clay. In other words, the theory assumes that creep is a result
of the squeeze-out of retained water due to compression loads. However, this hypothesis was
contradicted by L'Hermite and Mamillan (1968b) who conducted tests on saturated concrete
specimens. Results showed that, whether the samples were loaded or load-free, the difference
between the obtained curves was statistically negligible; given that all samples were identical
and exposed to the same environmental history. Thus, the conclusion was: concrete does not
behave like a sponge which loses its water when compressed (L'Hermite and Mamillan
1968b).
1200
8Q0
IQ
T"
i
a
m
13
u 400
10
28
Days-
90
Tim since
Loading (log
5
Years
10
20 30*
scale)
Figure 2.16: Creep of concrete cured in fog for 28 days, then loaded and stored at different
relative-humidity levels (Troxell et al.1958)
2.3.3.6
Hydration at Loading
Commonly called aging, the process of cement hydration has a pronounced influence on the
creep of concrete. If loaded at an early age, concrete would exhibit much greater creep than
concrete loaded well after casting. At an early age, the capillary porosity is higher and the
degree of hydration is low. This, in turn, causes the higher capacity of creep. Furthermore, the
micro-porosity of C-S-H (Calcium Silicate Hydrate; the main product of the hydration of
Portland cement which is primarily responsible for the strength in cement based materials)
30
might be greater at early ages than in mature pastes. The reaction between the silicate phases
of Portland cement and water is typically expressed as:
2 Ca 3 Si0 5 + 7 H 2 0 > 3 CaO 2 Si0 2 4 H 2 0 + 3 Ca(OH) 2 +173.6 kJ

2.3.3.7
Other Intrinsic properties of the Concrete Mix
Other intrinsic properties are known to affect the magnitude and rate of creep of concrete.
Some of them are the type of cement, mix proportions, additives and aggregate type:
Type of cement: The rate of strength gain is a function of the type of cement. It was reported
by Alexander et al. (1986) that creep reduces with the increase in the SO3 content. According
to the ACI committee 209, if the cement content is increased and the water-cement ratio is
kept constant, creep should increase.
Mix proportions:
Concrete composition is mainly comprised of the water-cement ratio (W/C),
aggregate, cement type and quantities of all the above. The effect of mix proportions on creep
characteristics was investigated by Collins (1989) in an effort where five mixes of 28-day
design strengths, ranging from 60 MPa to 64 MPa, were applied. The obtained results showed
that concrete mixtures with lower cement paste and larger aggregate size exhibited less creep.
Furthermore, the use of high-range water-reducing admixtures had little impact on creep
behaviour.
Additives: In a state-of-the-art report by Zia et al. (1991), several studies were summarized,
regarding the creep of concrete comprising fly ash, silica fume, and water reducer. Results
showed that there were very marginal differences between the specific creep of Portland
cement concrete and that of silica fume concrete or of fly ash concrete. Nevertheless, for a
particular load level and at any loading-time, the creep coefficient, creep strains, and specific
creep, were all lower in value for high strength concrete than for conventional concrete.
Aggregate type: The creep behaviour of a variety of very high strength concrete (YHS) with
different types of aggregate (crushed granite, marine marl, and rounded gravel) was studied by
31
Zia et al. (1993e). For each type of aggregate, creep strain measurements were taken for 90
days. The resulting creep strains, of the different VHS-concrete groups, ranged from 20 % to
50 % of that of ordinary concrete. For concrete comprising marine marl aggregate, the specific
creep was much greater than that of concrete with crushed granite or washed rounded gravel.
Thus, the aggregate content is an influential factor in the creep of concrete. Neville (1996)
claims that increasing the aggregate content from 65 to 75 %, by volume, typically decreases
creep by 10 %.
2.3.4
Methods of Modelling the Creep of Concrete
Most available creep laws are derived from one another with some differences related to the
range in which the parameters vary, or to the theory in which the law relies on. Some of these
laws take many parameters into consideration, whereas others have a limited number of
parameters with a goal of simplifying calculations. Furthermore, the definitions of these
parameters are sometimes different. Some of the most popular creep laws are presented
below:
2.3.4.1
Rate of Creep Method (RCM)
This method was originally developed by Glanville (1930b). The method proposes a simple
solution by assuming the creep law as the first-order differential equation, and that the rate of
creep is independent of the age at loading, as
de _
dt
1 da ^ a(t) dtp + dssh

E(t) dt
E(t0) dt
dt
^ 5)
This method overestimates the creep recovery of concrete, yet this deficiency is overcome in
the rate of flow method to follow.
2.3.4.2
Rate of Flow Method (RFM)
England and Illston (1965) proposed to present the specific creep function J(t f ) as the sum of
two components; the irrecoverable creep or flow (/) and recoverable creep or the delayed
elastic component (ed). They also concluded from their experiments that the delayed elastic
component is independent of the age at loading and reaches a finite value after a relatively
small period of time that is much faster than the flow component. Both components of creep
32
are, however, directly proportional to stress. It should be noted that the RFM method is also
based on a known control creep curve whether predicted or measured in the lab. In this
method, to represents the time when the control creep specimen was initially loaded, whereas t
is the time at which any other specimen under study has been loaded.
Applying the same concept discussed in the effective modulus method, Nielsen (1970)
proposed to add the delayed elastic component to the instantaneous elastic strain component
resulting in a fictitious effective modulus called the dynamic modulus (EJ) as
1 _ 1
d
E(t0)
E(t0)
.6)
where, fy is the creep coefficient for the delayed elastic strain. The creep function is treated
the same way as in the RCM method, but based on the new E j value, as given by
j
M L M
E(t)
(2.7)
where, (j)f{t)= creep coefficient for the creep flow strains at any time t, given by
Ât) = (t>{t,t0)-(t)d
2.3.4.3
(2.8)
Effective Modulus Method (EMM)
This method is considered the simplest, oldest, and the most widespread creep model. It
consists of a single elastic solution using the effective or sustained modulus (Ee) as follows:
1
=
J(t,t0)
ô)
l+
(2.9)
where E (t) is the initial elastic modulus, and (p{t,tt) is the creep coefficient. This method is
based on an assumed specific creep function, and provides good results when the aging effect
is negligible as in old concrete. In the presence of aging, creep due to stress changes after (t0)
is overestimated, since the method cannot account for stress variation with time.
33
2.3.4.4
Age Adjusted Effective Modulus Method (AEMM)
This method was originated by Trost in (1967). It takes into account the aging effects and the
variation of the applied stress at any time ( f ) after a loading age (to) without overestimating the
creep strain. This method also reduces the problem to a single elastic analysis, as follows:
go=g(0-gQ
g(0 =
<>
(2.10)
where Go is the applied sustained stress, a(t) is the stress at time t, Ee is the effective modulus
as in the effective modulus method (EMM), and Ee is the age-adjusted effective modulus
including the aging affect, as follows:
E(t0)
E =
1 +Z(t,t0)4>(t,t0)
(2.11)
where % (t,t0) is the aging coefficient function. The inconvenience within this method is due to
the lack of age coefficient data, which is needed in the calculation. In lieu of test data, an
approximate value of % = 0.8 is normally used for concrete, but may not be appropriate for all
cases.
2.3.4.5
Creep Equations in Codes and Guidelines
For compressive stresses less than 0.4 f'c the total time-varying (creep) strain, sca(t,t0)
due to
constant stress, <rc(t0), applied at time to are calculated as follows in CAN/CSA-S6-06 and
CEB-FIP Model Code 1990:
cr7(t,t0)
ac(t0)
1 +<f>(t,to)
f. (fo)
ii. i
r, 28
E
c
(2.12)
where Ec(to) is the modulus of elasticity of concrete at loading time; ECi28 is the modulus of
elasticity of concrete at 28 days. The creep coefficient of concrete, <p(t,to), is calculated as
follows:
4(t>to) = 4mPfPtPc(t>to)
(2.13)
where
,
fRH ~
1-(//)/(100%)
'
, ,
0.46[(2r v )/100]
34
(2
M)
In Eqn 2.14, RH is the mean annual relative humidity and rv is the volume per unit length of a
concrete section divided by the corresponding surface area in contact with freely moving air
(in millimetres). As regards the factors /?/, /?, and /?c (tJo)'
P
13
=
/
[ ( / ; + )/10] 5
(2.15)
where "a " is the difference between mean concrete strength and specified strength, f ' , at 28
days;
1
P,=
(2.16)
o.i+(0 02 '
0.3
t-tn
fin+t-to
(2.17)
where
18"
PH
=150 1 + 1.2
100%
2r
100
+ 250
(2.18)
NB. pH, shall not be taken larger than 1500.
As regards ACI 209, a power creep function was proposed to predict creep strain as follows:
it-t
' 1 0 + (t-t0)
\0.6 TU
( 2 1 9 )
where t is the time in days after loading; <pu is the ultimate creep coefficient for concrete
which varies between 1.3 and 4.15 (8U is typically taken as 2.35); ecr is the creep strain and 8,
is the initial immediate strain. Creep can also be expressed in terms of the creep coefficient (pt,
defined as the ratio of creep strain to initial immediate strain. According to ACI Committee
209, the creep coefficient is determined using the expression given in Eqn. (2.20). This
expression is applicable to normal, semi-low and low-density concrete.
V6
(t-t
Kl
l }
6=
MM
10 +
'
0 ~ 0
(2.20)
The above equation was developed for sustained compressive stresses not exceeding 50 % of
the concrete strength. It consists of an expression for creep under standard conditions
35
multiplied by the correction factor (f)corr to modify for non-standard conditions. Up to date, no
creep law is appropriate enough to be used for predicting the creep strain of confined concrete.
Several numerical models have been developed for this purpose but all need to be validated
with experimental data.
2.4
Modelling the Deflection Behaviour and Crack Width Evolution of FRP-RC Beams
The information laid out, in the previous sections, provides some of the ingredients by which
the long-term deflection of FRP reinforced concrete beams can be calculated. The total
deflection at any time t is the sum of the instantaneous deflection at time to and the timedependent deflection. This section provides the theoretical means by which the deflection of
FRP-RC beams is calculated (immediate and long-term) using different methods, most of
which are available in codes and guidelines. Furthermore, in this section, models dealing with
deflection-control and crack-control of FRP-RC beams are displayed and explained.
2.4.1
Theoretical Prediction of Immediate Deflection
Two well-known approaches are available for the prediction of initial FRP-RC deflection: the,
well known, effective moment of inertia approach (Branson's equation) and the mean
curvature approach. Long-term deflection and its prediction was initially the main objective
for this part of the current study. However, the obtained data regarding instantaneous
deflection (for 22 FRP-RC beams) was found to be of great benefit; especially that the wellknown Branson's equation, used to predict the instantaneous deflection for reinforced
concrete beams, significantly under-estimates experimental values of FRP-RC beams loaded
at McrIMa (cracking moment to actual applied maximum moment) low enough to approach
unity. Thus, the models available were compared to the obtained experimental data and a
modification was proposed for Branson's equation to yield better predictions.
2.4.1.1
The Empirical Moment of Inertia Approach (Branson's Equation)
The first approach, developed by Branson, is known for its use in calculating the postcracking deflection of steel reinforced concrete members (CSA A.23.3-04 and ACI 318-08);
later modified to accommodate FRP reinforced concrete (ACI 440.1R-06). This equation
36
interpolates between the moment of inertia of the gross uncracked concrete section, I g , and the
moment of inertia of a transformed cracked section, I cr , to account for the tension stiffening
effect of the concrete in tension.
/ =
1+
\MaJ
L, < L
(2.21)
where Mcr and Ma are the moment just sufficient to produce cracking and the maximum
applied moment, respectively. Considering the value of I e as constant over the length of the
cracked member, a simple equation can be used to calculate the midspan deflection of the
beam as a function of Ie, load magnitude, length of member and modulus of elasticity of the
concrete. A modification factor, ftd, was introduced by ACI 440.1R-06 to account for the
reduced tension stiffening exhibited by FRP-RC members, to yield the following equation:
I
i+
'A/
/ <1
(2.22)
In the current ACI 440.1R-06 guide, (id is calculated as a function of the relative reinforcement
ratio:
Pf_
<1.0
(2.23)
Pfi
In the earlier ACI 440.1R-03 version, the latter factor was calculated as follows:
Pd
-+1
(2.24)
where Ef is the modulus of elasticity of FRP; Es is the modulus of elasticity of steel bars and a
is taken as 0.5 for all types of FRP bars.
2.4.1.2
The Moment-Curvature
Approach
This approach based on the direct relationship between moment and curvature. It depicts the
real life behaviour of the structure in service conditions. The moment-curvature model allows
the effect of different parameters such as the reinforcement, cracking, creep and shrinkage,
which allows for long term deflection to be calculated (Favre and Charif, 1994). A mean
curvature, y/m, at a cracked section, is calculated as a function of two curvatures for uncracked
37
transformed sections and fully cracked transformed sections. Assuming parabolic variation of
curvature, midspan deflection can be calculated as a function of i//, and the member length.
The mean curvature at time of loading, y/(to), is given by:

V M = (l - Z M + Z i V i
(2-25)
where y/j is the instantaneous curvature for an uncracked section (1/77 = M /E c I g ); 1//2 is the
instantaneous curvature for a cracked section (if/2 = M /E c I cr ); C/ is the coefficient for
instantaneous curvature.
1-
K-
for a cracked section

(2.26)
for an uncracked section
v.
In turn, the initial deflection is calculated by assuming a parabolic variation of curvature, with
zero deflection at the ends and maximum at midspan, for a simply supported beam. It can be
roughly expressed as a function of the mean curvature at time of loading, y,{to), and the
length of span, L:
A =VML
9.6
(2.27)
For simplicity, this method is applied using only one curvature at the midspan of the beam.
For refined results, it can be applied on several sections along the beam and numerical
integration would take place.
2.4.2
Theoretical Prediction of Long-term Deflection
The time dependent behaviour of reinforced concrete is intrinsically complex due to the many
influencing
factors.
Some
of
these
factors
are
concrete
compressive
strength,
upper/compression reinforcement, ultimate creep coefficient, humidity, time of loading and

duration of loading. There are two common approaches for predicting the long-term deflection
of FRP-RC beams, due to creep and shrinkage. The first is an empirical approach which
involves multiplying the initial deflection J , by a multiplier X, where X varies according to the
age of loading. The second approach involves numerical (finite-difference) modelling. In the
38
latter approach, the time-dependent curvature, strains and deflections induced by creep and
shrinkage are calculated based on the age adjusted modulus method.
2.4.2.1
The Empirical Multiplier Approach (X)
The time-dependent beam deflection multiplier, X, is a function of two effects: increasing

curvature at cross-sections along the span due to creep, and the effects of additional cracking
along the span (Gross et al 2003). CAN/CSA S806-02 and ACI 440.1R-06 adopt the empirical
multiplier approach indicated below:
creeps shrinkage)
^ ( ^ i ) sustained
1 + 50 mp'
(2.28)
(2.29)
where the factor accounts for the time-dependent concrete behaviour and the denominator
accounts for the restraint against creep provided by compression reinforcement; p is the
compression reinforcement ratio and m is the modular ratio of the reinforcement (m Esteel-for
EFRP
FRP reinforced concrete). CAN/CSA S806-02 attribute to the time-dependent factor,
the values of 1, 1.2, 1.4 and 2.0 for three, six, twelve and sixty months, respectively. ACI
440.1R-06 multiplies the latter values by a 0.6 modification factor (in light of earlier studies
conducted by Kage et al. (1995), Brown (1997), Vijay and GangaRao (1998), Arockiasamy et
al. (1998) and Gross et al. (2003)). Where no compression reinforcement exists or with FRP
compression reinforcement, the long-term deflection equation reduces to:
(creeps shrinkage)
= 0.6(A,)
(2.30)
It is worth noting that CAN/CSA S806-02 adopts the same values for as ACI 318-08 and
CSA A.23.3-04 (for steel reinforced concrete).
2.4.2.2
The Time-dependent Mean Curvature Approach (Numerical Modelling)
The second approach, for computing long-term deflection, is based on the age-adjusted
effective modulus method in which the properties of the section change with time and
calculation at each time step relies on its predecessor. The foundation of this approach was
first introduced by Bazant (1972) and later presented using a system of equations found in the
textbook of Ghali, Favre and Elbadry (2002). The age-adjusted effective modulus method is
39
sophisticated and requires numerical (finite-difference) modelling; it is encouraged as an

application for researchers to gain insight into the mechanism underlying the long-term
deflection phenomenon.
Using the age-adjusted effective modulus method, a system of equations provided by Ghali,
Favre and Elbadry (2002) was used to code a user-friendly Fortran-2003 program (verified by
Fortran-2003). In this program, beam curvature, deflections and strains are calculated based
on a theoretical approach. The flowchart of this program is shown in Figure 2.17.
40
Figure 2.17: Flowchart for calculating the deformations of GFRP reinforced concrete beams.
The system of equations describing the long-term deflection calculation is explained below:
41
Mean curvature at time t is given by:

y/{t) = {\-^2\y/x+Ay/x)
+ ^2{y/2+Axi/2)
(2.31)
where A ^ is the change in curvature of the uncracked section; AI//2 is the change in curvature of the
cracked section;
is the coefficient for long-term curvature.
M.
1-0.5

(2.32)
6 =
The total deflection at the midspan at time t is given by

A
{t)l}
9.6
(2.33)
The time dependent change in strain at the centroid of the section and the time dependent
change in curvature due to creep and shrinkage are given by
Ae0=7i[<f>(t,t0)e0+ecs(t,t0)]
(2.34)
Ay/ = K 0(t,to)y/ + cs(t,to) yc
(2.35)
where
T - ' s - ' r
(2.36)
where 0 ( t , t 0 )is the creep coefficient from time to to time t\ cs(t, t 0 ) is the shrinkage strain
from to to time /; yc is the distance between centroids of the effective concrete area and the
age-adjusted transformed area; Ac is the area of concrete in compression; A' is the area of the
age-adjusted transformed section; I c is the moment of inertia of concrete about the centroid
after age adjustment; / ' is the moment of inertia of the age adjusted transformed section using
the modular ratio r)'(t, t0) in which r\'{t, t 0 ) = EGFRP/E'c(t,
elasticity, E'c(t, t 0 ), is calculated as follows:
42
t 0 ) . The age adjusted modulus of
EM
(2.37)
where Ec is the modulus of elasticity of concrete at time to and % (/,/) is the aging coefficient
function. In this study, the creep and shrinkage coefficients are first calculated based on ACI
Committee 209 and CEB-FIP Model Code 1990 approaches. The same coefficients were later
multiplied to a 0.55 correction factor as proposed by Brown (1997). This multiplier accounts
for the difference between the creep of FRP-reinforced concrete and steel reinforced concrete.
The theoretical predictions dramatically improved when Brown's, 0.55, multiplier was
incorporated.
The ACI Committee 209 equations describing the creep and shrinkage coefficients of concrete
are:
(2.38)
(2.39)
where 0 ( t , t 0 ) is the creep coefficient at time t for age at loading to; c s (t, t ' ) is the free
shrinkage which occurs between t'at the end of the curing period and time t; (pu is the ultimate
creep coefficient (~ 2.35);
(ECS)U
is the ultimate shrinkage strain at infinite time and (pcorr is
Brown's correction factor for creep of FRP reinforced concrete. Coefficients cf}u and (e c s ) u
depend on atmospheric humidity, member size, slump of concrete, cement content, fine
aggregate percent, air content in percent and type of curing. On the other hand, the creep
coefficient according to CEB-FIP Model Code 1990 is given as:
icorr
(2.40)
where the factors 0 R H , / ? ( / c m ) , / ? ( t o ) and /? c (t t 0 ) take into account the corrections for the
atmospheric humidity, mean compressive strength of the concrete, time of loading and
duration of loading respectively. The strain due to shrinkage and swelling according to CEBFIP Model Code 1990 is calculated as follows:
cs csoPs(.t
43
t )
(2.41)
where cs0 is a function of the mean compressive strength of concrete, type of cement and
relative humidity; /? s (t t ' ) is a function of the loading duration.
2.4.3
Deflection Prediction (Immediate and Long-term) in Codes and Guidelines
In this subsection, the deflection-prediction methods available in popular codes and guidelines
(CAN/CSA S806-02; ACI 440.1R-06; ISIS Canada Design Manual (2007); Eurocode 2 and
CEB-FIP Model Code 1990; CNR-DT 203 (2006)) are highlighted upon. The approaches
explained above, for short term and long term deflection prediction, are adopted in most of
these provisions.
2.4.3.1
Deflections in Accordance with CAN/CSA S806-02
CAN/CSA S806-02 recommends the integration of curvature along the span to determine
curvature. On knowing the value for curvature if/, the virtual work method can be used to
calculate the deflection of FRP-RC beams under any load level.
(2.42)
To calculate the curvature at any section, a tri-linear moment-curvature relation is assumed
with the flexural stiffness being EcIg for the first segment, zero for the second, and EcIcr for the
third (See Figure 2.18).
K
Figure 2.18: Moment-curvature (M-k) relation for FRP reinforced concrete (CSA S806-02)
44
This method typically requires more computational effort than the empirical equations made
available in ACI 440.1R-06 and the ISIS Canada design manual (2007). This approach
assumes no tension stiffening. However, the actual inertia for each section along the beam,
whether cracked or uncracked, is considered. To enhance the accuracy of deflection
calculation, numerical integration should take place over several sections. Table 2.5 provides
formulas derived from this method for simply supported beam cases.
Table 2.5: Maximum deflection formulas using the deflection approach (ISIS Canada Design
Manual 2007)
Load/Beam Type
Formula
P - Point Load
I
-L/2-
-L/2-
PU
4SEf/.r
P - Point Load
1
P - Point Load
1
J max =
A1
I
q - Distributed Load
I < 1 I i I i "J I' S I J 4 ,(
T
HI
<5L
PU
24 E J W
5qlt
3S1
IL s
I-8/7M
L j
f.1-4
X )
8/7
v-^ /
19:
3
Where L length of beam still uncracked, ran
n = (i-f1)
As regards the calculation of long-term deflection in CAN/CSA S806-02, the long-term
multiplier approach, explained in section (2.4.2.1).
45
2.4.3.2
Deflections in accordance with ACI 440.1R-06
The short-term deflection of a steel RC cracked beam can be simply obtained by applying the
standard linear-elastic approach and using a constant effective moment of inertia [Branson
(1966), (1977)] (See Eqn. 2.21). For FRP-RC modifications were undergone to account for the
reduced tension stiffening exhibited by FRP-RC members (See Eqn. 2.22 and Eqn. 2.23).
Several research efforts, including the current, realized that further modification for this
approach is necessary to yield better predictions.
As regards long-term deflection, the long-term deflection prediction can be obtained by

multiplying the short-term (immediate) deflection due to sustained load by a factor X (See
section 2.4.2.1).
2.4.3.3
Deflections in Accordance with ISIS Canada Design Manual (2007)
To calculate deflections, ISIS Canada Design Manual (2007) recommends the following
equation to be used in design:
(2.43)
Mota et al. (2006) examined a number of the suggested formulations for I e and found that the
former yielded the most consistent and conservative results over the entire range of madeavailable specimens. This equation was reported to work well with different types of FRP
reinforcement.
2.4.3.4
Deflections in Accordance with Eurocode 2 and CEB-FIP Model Code 1990
Eurocode 2 and Model Code 1990 adopt the following approach for the calculation of short
and long-term deflection "<T for steel RC. It is worth noting that CNR-DT 203/2006 (the
Italian guide for FRP-RC) endorses Eurocode 2 for FRP reinforced concrete.
s =
sir+s2(i-r);
(2.44)
(2.45)
46
In the above expressions, Si and 82 are calculated assuming constant un-cracked and cracked
sectional moments of inertia along the element. The coefficient [1 takes into account the
duration of the load and bond, while the coefficient m allows for the tension stiffening effect.
Values recommended for these coefficients by the two European codes of practice are shown
in Table 2.6.
Table 2.6: Values for coefficients /? and m (Eurocode 2 and Model Code 1990)
Eurocode 2
Model Code 1990
0.8
For FRP rebars the coefficients ft and m should be evaluated experimentally. Zhao (1999)
concluded that both the Eurocode 2 and Model Code 1990 prediction equations for the
instantaneous deflection of steel RC elements could be adopted directly for GFRP RC
members in bending. Pecce et al. (2000) pointed out that the model proposed for steel by
Eurocode 2 is reliable and could be used for GFRP RC beams if the bond performance is
comparable.
2.4.4
Permissible Computed Deflections and Deflection Control for FRP-RC Beams
Typically, there are two methods available in provisions for the control of deflection due to
immediate and sustained static loads. The first is an indirect method that mandates the
minimum thickness of the member such as that available in ACI 440.1R-06 (See Table 2.7):
Table 2.7: Recommended minimum thickness of non-prestressed beams or one way slabs
Solid one-way slabs

Beams
Simply
Supported
LI 13
Z/10
Minimum thickness h
Both end
One end
continuous
continuous
LI 17
LI 22
LI 12
L116
Cantilever
LI 5.5
LI4
The second - direct method - limits the computed deflections. CSA S806-02 indicates that the
computed deflections shall not exceed the limits stipulated in Table 2.8. Due to the brittleelastic nature, and particular bond features of FRP reinforcement, the deflection of FRPreinforced concrete members is more sensitive to the influencing variables than their steel
47
counterparts. Deflections of FRP-RC beams also tend to be greater in magnitude because of

the lower stiffness associated with commercially available FRP bars. Thus, the use of the
direct method of deflection control is required (the second method); the minimum thicknesses
for FRP-reinforced members can be obtained from the indirect method (the first method) for
convenience in establishing member proportions for design only.
Table 2.8: Maximum permissible computed deflections

Deflection to be considered
Type of member
Flat roofs not supporting or attached to Immediate deflection due to
non-structural elements likely to be specified load, L.
live damaged by large deflections.
Floors not supporting or attached to Immediate deflection due to
non-structural elements likely to be specified load, Z.
damaged by large deflections.
Roof or floor construction supporting
or attached to non-structural elements
likely to be damaged by large
deflections.
Roof or floor construction supporting
or attached to non-structural elements
not likely to be damaged by large
deflections.
The part of the total deflection

occurring after attachment of
the non-structural elements
(sum of the long-time deflection
due to all sustained loads and
the immediate deflection due to
any additional live load) J
Deflection limitation
L /180*
LI 360
L /480f
L /240
*Limit not intended to safeguard against ponding. Ponding should be checked by suitable calculations of
deflection, including added deflections due to ponded water, and consideration of long-time effects of all
sustained loads, camber, construction tolerances, and reliability of provisions for drainage.
f Limit may be exceeded if adequate measures are taken to prevent damage to supported or attached
elements.
JLong-time deflections shall be determined in accordance with Clause 8.3.2.4 but may be reduced by the
amount of deflection calculated to occur before the attachment of nonstructural elements.
Not to be greater than the tolerance provided for nonstructural elements. Limiting deflection may be
exceeded if camber is provided so that the total deflection minus camber does not exceed the limit shown
in this Table.
It is worth noting that the use of the recommended minimum thicknesses in Table 2.7 does not
guarantee that all deflection considerations will be satisfied for a particular project. Values in
Table 2.7 are based on a generic maximum span-depth ratio limitation (Ospina et al. 2001)
corresponding to the limiting curvature associated with a target deflection-span ratio:
48/7 1 -k
h
5K,
48
(2.46)
J\ l j
where r] = d/h\ {All)max is the limiting service load deflection-span ratio; Kt is a parameter that
accounts for boundary conditions and is taken as 1.0, 0.8, 0.6, and 2.4 for uniformly loaded
simply supported, one end continuous, both ends continuous, and cantilevered spans,
respectively. The term e/ is the strain in the FRP reinforcement under service loads, evaluated
at midspan except for cantilevered spans. For cantilevers, Sf shall be evaluated at the support.
The factor k can be calculated from the following equation:
k = J2pfnf+(pfnf)2
-pfnf
(2.47)
where pj and nj are the reinforcement ratio and the modular ratio between FRP and concrete,
respectively. Tabulated values are based on an assumed service deflection limit of //240 under
total service load, and assumed reinforcement ratios of 2.0/3/j and 3.0 pjb for slabs and beams,
respectively.
An alternative approach to the above (developing minimum thickness criteria) was originally
proposed by Rangan (1982) for one-way members and extended to two-way systems by
Gilbert (1985). This approach has the advantage of being rationally based and can be linked to
a pre-specified permissible deflection criteria. The critical deflection for concrete members
satisfying ACI Code permissible deflections is usually the time-dependent deflection
occurring after installation of nonstructural elements. The deflection equation for a uniformly
distributed load can be written as:
A
mc
(2.48)
384E c I e
384 c / e
This equation can be re-arranged in the new form of an expression for span-to-depth ratio in
terms of deflection-to-span ratio,
I
h
32 aEb
I Jaua.
1/3
(2.49)
K(AWs+WL(add))
where a = Ie/Ig, A, is a long-time multiplier, Ws is the sustained load including a portion of live
load and Wi{add) is the remaining portion of live load. An equation of this general form was
incorporated in the Australian code based on Rangan and Gilbert's work with an expression
for a as a function of reinforcement ratio. Unfortunately, the reinforcement ratio is unknown
49
at the time the thickness is selected. The other term in the equation that is not known at the
time the thickness is computed is the dead load which obviously depends on the member
thickness. Scanlon and Choi proposed a constant factor for a for one way members and
demonstrated that a reasonable estimate of the member depth quickly converges to the desired
member depth.
A modified form of the Scanlon-Choi span-to-depth function was subsequently incorporated
in the Australian Design Standard (2001). Scanlon and Lee (2006) extended this approach to
cover both one-way and two-way systems. The advantage of this approach is that it takes into
account the significant factors affecting deflection. This approach can accommodate other
assumptions such as long time multiplier, loads, and specified deflection to span ratio, and is
relatively easy to apply.
2.4.5
Effect of Sustained Loads on Flexural Crack Width in FRP-RC Beams
The control of flexural cracking takes added importance in the design process of FRP-RC
beams than steel reinforced beams. The lower elastic modulus for FRP bars leads to larger
crack widths and deflections than identical members with an equal amount of steel
reinforcement. Thus, serviceability considerations often govern the design of FRP-RC
whereas steel-reinforced concrete is governed by the capacity at strength limit state. Two of
the most popular methods for predicting the instantaneous flexural - most probable maximum
- crack width of FRP-RC beams are Frosch's (1999) equation and the expression of Gergely
and Lutz (1968). The former equation, adopted by ACI 440.1R-06 and CAN/CSA S6-06, is
expressed as:
(2.50)
In this equation, /? is the ratio of distance from neutral axis to extreme tensile fibre and
distance from neutral axis to centre of the tensile reinforcement = (h - c)/(d - c); where h is
beam depth; c is the neutral axis depth under sustained load and d is the effective to centre of
reinforcement; er is the reinforcement strain at service; dc is the concrete cover from extreme
tension fibre to the centre of the closest reinforcing bars; s is the centre-to-centre spacing of
reinforcing bars within outermost layer. The coefficient kb accounts for the degree of bond
50
between the FRP bar and the surrounding concrete. According to ACI 440.1R-06, kb - 1.0 for
deformed steel bar reinforcement; < 1.0 for superior bond relative to deformed steel bars; >
1.0 for inferior bond relative to deformed steel bar. It is typically assumed as 1.2 to 1.4 when
the actual value is unknown. For an analysis of crack width data performed by the ACI
440.1R-06 committee on a variety of FRP bars, the average kb values ranged from 0.60 to 1.72
with a mean of 1.10 (ACI 440.1R-06).
On the other hand, the Gergely-Lutz (1968) equation is adopted by the former ACI 440.1R-03
guidelines and the ISIS Canada Design Manual (2007). It also serves as the basis for a
modified expression used to compute the "z factor" used for crack control calculation in CSA
S806-02 (clause 8.3.1.1). The Gergely-Lutz equation is expressed as follows:
w=
2.2/3kbsr{dcA),1/3
(2.51)
Nevertheless, these two equations do not account for the increase in crack width, with time,
due to sustained loading. A long-term study conducted by Gross et al. (2009) proposed an
additional coefficient/multiplier kt to account for the time-dependent increase in crack width:
(2.52)
Based on the long-term test results of 8 GFRP reinforced beams and 2 CFRP reinforced
beams, kt is taken as 2.0 for FRP reinforced beams exhibiting sustained load for a one-year
duration. The latter factor can be applied to the Gergely-Lutz equation, as will be shown in
Chapter 5 and Chapter 6.
As shown in the two crack-width models, above, w is a function of the strain induced in the
reinforcement bars. Typically, the measured strain in the bars is higher than the values
obtained from calculations. This increase also varies according to the relative distance from a
concrete crack. According to MacGregor and Wight (2005), reinforcement strain/stress
exhibits minimal increase (~3 % for GFRP and 6 % for steel) at the cracked section, whereas
the reinforcement strain/stress would increase significantly between cracks (expected to
triple). On the other hand, concrete stress on the tension side changes negligibly because creep
is effectively redistributing stress from the concrete to the reinforcement. This phenomenon is
51
explained elaborately in the discussion section of the study of Gross et al. (2009), indicating
its impact on the growth of crack width with time.
As regards the acceptable crack width limits in provisions, CAN/CSA S806-02 implicitly
identifies the maximum acceptable crack widths of 0.02 inches (0.5 mm) for exterior exposure
and 0.028 inches (0.7 mm) for interior exposure conditions. Other European guides such as
CNR-DT 203/2006 indicate that under no circumstance should the crack width of FRP
reinforced structures be higher than 0.5 mm. ACI 440.1R-06 endorses the use of the
CAN/CSA S806-02 limits. These limits are more relaxed than those associated with
conventional reinforced concrete design due to the corrosion-free nature of FRP
2.5
Long-term Behaviour of FRP Reinforced Concrete Beams under Sustained Load
Since the advent of FRP reinforced concrete beams, the number of studies regarding the long
term performance/deflection of FRP reinforced concrete beams is yet very limited (Brown and
Bartholomew (1996), Brown (1997), Vijay and GangaRao (1998), Arockiasamy et al. (2000),
Hall and Ghali (2000), Gross et al. (2003), Gross et al. (2006) and Gross et al. (2009)). A
thorough literature review was conducted to find and summarize all projects dealing with this
research area. It was found that the experimental work, to date, amounts to a dozen studies at
most; some of which, are summarized and presented in the paragraphs below. It was found
that most of these studies adopt a two-point loading setup where the longterm behaviour of
FRP reinforced beams (designed for compression failure) are monitored in terms of deflection,
strain variation and crack width, for time durations varying from 3 months to two years. In the
majority of the previous studies, FRP reinforced beams had no top reinforcement. Aside from
the current study, the author of this thesis does not know of any experimental program that has
FRP-RC beams under uniform sustained load but that of Arockiasamy et al. (2000).
Nevertheless, the latter study does not represent the scenario of full-scale beams such as those
cast and tested in Chapter 6.
2.5.1
Long-term Behaviour of FRP-RC Beams subjected to 2-point Loading
Midspan deflection and crack width results of six high-strength-concrete beams (two mm
GFRP reinforced, two CFRP reinforced and two steel reinforced) were presented by Gross et
52
al. (2006). The beams of dimensions (114 mm x 184 mm x 1880 mm) were maintained under
four-point bending sustained load for a period of three months (ninety days). The testing/clear
span was 1828 mm and the constant moment region was 102 mm. The reinforcement used
was 2-15.9 mm GFRP bars, 2-9.5 mm CFRP bars and 2-15.9 mm steel bars, respectively,
yielding similar moment capacity for the three sets of beams. All beams had no shear
reinforcement. The applied sustained loads amounted to 30 % of the beams' nominal capacity
which approximately corresponds to 2 times the cracking moment (MJMcr
= 2). When
compared to earlier efforts (Brown (1996); Vijay and Gangarao (1998); Arockiasamy et al.
(1998)), the obtained long-term deflection results were relatively lower due to the use of high
strength concrete. The average deflection-increase was 33 %, 16 % and 24 % from the initial
induced deflection, for GFRP and CFRP and steel reinforced beams, respectively. The former
two values - for GFRP and CFRP - are far less than the predicted 60 % increase and 100 %
increase indicated in ACI 440.1R-06 and CAN/CSA S806-02. Measured maximum crack
widths - at test inception - for GFRP and CFRP reinforced beams were significantly greater
than steel reinforced beams. The ratio of time-dependent to initial crack width, however,
showed relative proximity for all three sets; the average crack width increase was found to be
13 %, 15 % and 20 % for GFRP, CFRP and steel reinforced beams, respectively.
Time dependent strains, deflections and cracking were monitored for three normal and three
high strength concrete beams; reinforced with 2-12.7 mm and 3-12.7 GFRP bars, respectively
(Gross et al. 2003). The beams were subjected to four-point sustained bending for a period of
6 months (180 days). The beams contained no shear reinforcement and had a (121 mm x 235
mm) cross section; the test-span was 3050 mm and the two loading points were located 252
mm apart at midspan. Three sustained load levels were applied; specimens were loaded to
approximate top fibre concrete compressive stresses of 0.40 / c ' , 0.45 / c ' and 0.50 f'c at an
age of 28 days. The overall behaviour of the GFRP reinforced beams was similar to steel
reinforced beams; where a downward shift of the neutral axis would occur over time
corresponding to redistribution of internal stresses.
The obtained results rendered six values of time-dependent deflection multipliers; such
multipliers are defined as the time-dependent deflection (total minus initial) divided by the
53
initial deflection. The multipliers associated with normal concrete beams were far greater than
their high strength concrete counterparts (See Figure 2.19). This was mainly due to the higher
creep coefficient of normal concrete. Furthermore, formation of additional flexural cracks,
with time, had a significant impact on the time-dependent deflection multiplier.
Bemns, Avg Df 2 sfieeimara
Cyw
i >dMS
14
28
dSKwnwu
42
66
?Q
84
t8
112 126 140 154 188 182
T ! m (Oaya A f t e r L o a d i n g )
B M m s N1 ( 4 0 f c >
- - B w m N2 ( 45rc)
- - B m N 3 (.50fc>
Cylnaof*
- * ~ A C t 31S-89
- - - A C ! 440 {2001)
a*m k f j . 01 J ipc>ir*n
OfimOm*
14
28
v> 3 pocn
42
36
70
84
T i m * (Ocya
112 128
140 154 186 1S2
AHt Loading)
FLILITIWWI* HI ( i o r e )
a B m * w a T t i f c ) " A 8 m H3 f.SOfe)
I -<>-CYndW
- - A C I 31M9
- A C I 440 (2001)
Figure 2.19: Average multiplier curves for normal strength concrete (top) and high strength
concrete (bottom) (Gross et al. 2003)
The experimental immediate and long-term deflections of four shallow concrete beams, two
reinforced with 3-15 mm GFRP bars and two reinforced with 3-15.9 mm steel bars, of crosssection (280 mm x 180 mm) and 3500 mm in length were presented by Hall and Ghali (2000).
The stirrup-less beams were simply supported and exhibited third-point loading for a period of
approximately 8 months. The two equal-in-magnitude applied concentrated loads were 1067
mm apart centreed in a 3200 mm loading span. Two levels of sustained loading were
considered (Maximum applied moment to cracking moment MJMcr = 1.5 and 3.0).
In the analysis phase, the authors used equations found in Ghali, Favre and Elbadry (2002) to
calculate the long term deflections. These equations are based on the first principles of
equilibrium and compatibility. The creep coefficients and free shrinkage values used in these
54
equations were calculated using equations from CEB-FIB Model Code 1990 and/or ACI
Committee 209 recommendations (1992). Immediate and long-term deflection values obtained
from experimental work were compared with theoretical values. It was found that CEB-FIB
Model code 1990 accurately predicts both the immediate and long-term deflections for both
steel and GFRP reinforced shallow beams . The ACI Committee 209, however, overestimated
the deflections of the tested steel and GFRP beams (Figure 2.20).
50
CEB-FIP Model Code 1990
CEB-FIP Model Code 1990
45
ACI Committee 209
40
Lg-S-3-1,5
ACI Committee 209

ACI Committee 209 (modified)
35 +
'"""O"** LgG-3*1,5
30
25
20 +
&
15
10
5
0
100
150
200
50
Time since loading (days)
100
150
200
250
Time since loading (days)
Figure 2.20: Comparison of experimental and predicted long-term deflections for two GFRP
reinforced beams (Hall and Ghali 2000)
In 1998, Vijay and GangaRao presented the long-term behaviour results for four GFRPreinforced concrete beams that were monitored for an average of 360 days (12 months
approximately). The beams were of dimensions (150 x 300 x 3050 mm) exhibiting third-point
loading where the testing span (distance between supports) was 2670 mm approximately (i.e.
distance between two concentrated loads = 890 mm) (See Figure 2.21). For the first pair of
beams, the concrete compressive strength, / c ', was 24 MPa; 2-12.7 mm sand coated bars were
used as bottom reinforcement; the maximum applied moment Ma, at midspan, was 35 % and
50 % of the nominal moment capacity of the beams Mn; the beams were designed for tension
failure. For the second pair of beams, the concrete compressive strength, / c ', was 28 MPa; 215.9 mm ribbed bars were used as bottom reinforcement; the maximum applied moment Ma,
at midspan, was 20 % and 35 % of the nominal moment capacity of the beams Mn\ the beams
were designed for compression failure. For all beams, 9.5 mm sand coated bars were used for
55
compression and shear reinforcement. The main findings of the study were that the creep
coefficient is less in GFRP concrete beams as compared to the available data on steel
reinforced beams; the long-term to initial deflection of GFRP reinforced concrete beams is
less than that of steel reinforced concrete beams; crack width increase is proportional to the
creep coefficient and can be estimated with suitable reduction factors on the latter.
1. Generate team
2 Distribtiicn l-Beam
3. Steel beam
4. Load cell
5. Loacing piston
6. Swivel arm for
adjusting ruts
7. Dywidagbar
8. Threaded nut for
9. RoBer assembly
10. Supporting block
Figure 2.21: Longitudinal and side views of concrete beams reinforced with GFRP bars under
sustained load (Vijay and GangaRao 1998)
Gross et al. (2009) conducted an experimental study to evaluate the time-dependent crackwidth increase in FRP-RC beams as a result of sustained loading. Twelve beams (eight GFRP,
two CFRP, and two steel-reinforced) were kept under a constant sustained service load for a
period approximating three years. For each beam, three flexural cracks were monitored over
the test duration. It was found that the increase in flexural crack widths for FRP-reinforced
specimens was greater than the steel-reinforced specimens. Flexural crack widths in FRPreinforced concrete specimens were observed to approximately double over one year of
sustained loading. A simple design approach, based on modification to the existing ACI
440.1R-06 crack control procedure (also available in CSA S6-06), was proposed to account
for this observed increase in crack widths with time.
56
2.5.2
Long-term Behaviour of FRP-RC Beams under Uniform Distributed Load
The typical scenario, adopted by researchers, for experiments dealing with the long-term
behaviour of FRP reinforced concrete beams, is as follows: The beams are of relatively small
scale, cast using normal
concrete without top/compression reinforcement or shear
reinforcement. The rendered information would typically deal with (i) the strain variation
within concrete and reinforcement with time, (ii) the time dependent deflection at midspan
and (iii) the maximum resulting crack widths, at midspan. Calibration/modification of
equations and models designed for steel reinforced concrete would take place based on the
experimental findings. As per the recommendations of Brown 1997, there is a need to study
the long-term behaviour of full-scale FRP reinforced beams as well as the moderating effect
of FRP and/or steel compression reinforcement. Regarding the loading method, a more
realistic approach would be to apply uniform distributed load, in agreement with the typical
design philosophy for reinforced concrete beams.
Figure 2.22: Test setup for beams (Arockiasamy et al. 2000)

In 2000, (Arockiasamy et al. 2000) studied the long-term behaviour of concrete beams
reinforced with carbon FRP bars (for a period exceeding 470 days). Four rectangular concrete
beams (two sets) were cast in two sizes; 152 mm x 203 mm x 2438 mm (for Set 1 of / c ' = 32
MPa) and 152 mm x 152 mm x 2438 mm (for Set 2 of f c = 43 MPa). Each beam contained 27.5 mm CFRP bars for bottom reinforcement, likewise for top reinforcement (hanger); 9.5
57
steel stirrups spaced at 75 mm to 150 mm for ends to midspan, respectively. The beams were
simply supported and loaded with concrete blocks to simulate uniform sustained load (See
Figure 2.22). Strain gauges were installed on the top concrete compression surface as well as
attached, at midspan, onto the embedded FRP reinforcement. The applied moment Ma was 77
% and 123 % of the respective cracking moment Mcr for Set 1. For Set 2 beams, Ma was 110
% and 123 % of the respective Mcr. After a 470 day period, it was observed that the relative
deflection increase of long-term (accumulated) to initial deflection was 15 %, 115 %, 65 %
and 71 % for beams B l , B2, B3 and B4, respectively. The increase in top surface (concrete)
strains, relative to the instantaneous values, was 101 %, 151 %, 209 % and 245 % for the same
beams, respectively. The authors used a computer program based on the age-adjusted effective
modulus method (made available in Ghali, Favre and Elbadry 2002) to predict the long-term
deflection of the beams. The creep and shrinkage coefficients were calculated based on ACI
Committee 209 recommendations (1992) and CEB-FIP Model Code (1990). The theoretical
curve rendered good agreement with the measured values (Figure 2.23).
Figure 2.23: Comparison of experimental and predicted deflections for beams B3 (left) and B4
(right) (Arockiasamy et al. 2000)
58
Furthermore, a more-convenient (simplified) equation - a modification of the empirical

equation available in ACI 318-08 and CSA A.23.3-04 - was proposed. The latter model
addressed two main parameters (i) the age of loading and (ii) the upper/compression
reinforcement. The study of Arockiasamy et al. (2000) is the closest available to full-scale
FRP reinforced concrete beams. Yet, more effort is necessary increase the information content
regarding
the
long-term
performance
FRP
reinforced
concrete.
Different
bottom
reinforcement ratios; different levels of sustained load; different types of upper reinforcement
are parameters that need to be addressed further.
2.6
Summary
This chapter contains a thorough literature review on the long-term behaviour of the
components comprising an FRP reinforced concrete element (i.e., FRP bars and concrete).
The information in this chapter can be summarized in the following points:
1-
The long-term (creep) behaviour of FRP bars is influenced by a variety of factors:

sustained load magnitude; fibre content and orientation; adverse environmental
effects.
2-
Studies were conducted, earlier, on the creep behaviour of commercial FRP bars
under sustained axial load. For the bars subjected to sustained load levels, within the
service load range (25 to 30 % fu,ave),
creep strain values ranged from 2 to 8.7% of the
immediate initial strain sjrP,o- At 60 %
fu,ave,
and below, it is stated that creep rupture
is less likely to occur to GFRP bars at standard laboratory atmosphere (Budelman and
Rotasy 1993).
3-
The time dependent deflection behaviour of FRP reinforced concrete beams is highly
dependent on the creep of concrete. Some factors that affect the creep of concrete are:
(i) the sustained load, (ii) concrete strength, (iii) type and quantity of reinforcement
and (iv) the size of the reinforced concrete member.
4-
The long-term deflection prediction of FRP reinforced concrete beams is done by

either empirical methods available in North American provisions (such as CAN/CSA
S806-02, ACI 440.1R-06 and ISIS Canada Design Manual (2007)) or numerical
methods such as the age-adjusted effective modulus method (Ghali, Favre and
Elbadry (2002)).
59
5-
Two popular methods for the instantaneous crack width prediction of FRP reinforced
concrete beams, are Frosch (1999) and Gergely-Lutz (1968), adopted by North
American provisions. The bond coefficient fa - a parameter in both equations - has
ranged from is suggested to be taken as 1.4 by ACI 440.1R-06 when the actual fa
value is unknown. Both equations do not account for the increase in crack width with
time.
6-
Regarding the long-term behavior studies of FRP reinforced concrete beams. There is
yet a lack of data. Most of the previous efforts deal with relatively small scale beams
under constant two point load. A more realistic approach is to apply sustained
uniform distributed load on full-scale FRP reinforced concrete beams.
60
Chapter 3: Creep Behaviour & Residual Properties of GFRP Bars Exhibiting Different Levels of Sustained Load
CHAPTER 3
CREEP BEHAVIOUR AND RESIDUAL PROPERTIES OF GFRP BARS
EXHIBITING DIFFERENT LEVELS OF SUSTAINED LOAD
3.1
Introduction
In 2003, a synopsis of gap analysis report related to the durability of fibre reinforced polymers
(FRP) composites in civil infrastructures, was presented (Karbhari et al. 2003). Backed up by
the aegis of the Civil Engineering Research Foundation (CERF), the Market Development
Alliance (MDA) and attendees from the user, owner, construction, design, materials, research
and manufacturing sectors, this study heavily emphasized on the lack of comprehensive and
validated durability data dealing with FRP composite materials, at whole. A protocol was set
to remedy this lack-of-information drawback, where seven subcommittees were assembled;
each of which is to focus on a particular adversity (environmental condition). Research
priority was given to the following seven environmental conditions: moisture/solution, alkali,
thermal (including temperature cycling and freeze-thaw), creep, fatigue, ultraviolet, and fire.
For the current study, the creep phenomenon within GFRP bars was chosen; the
aforementioned synopsis ranks this phenomenon as highly critical yet the relevant data, to
date, is sparse and questionable. Despite continuous research, the information content
regarding this issue did not increase significantly.
Not only the creep and shrinkage of concrete, but the creep of GFRP bars, account for the total
long-term deflection of GFRP reinforced concrete beams (Gaona 2003). Creep behaviour of
GFRP bars in general, whether creep deformation or creep rupture, has an influence on the
design of GFRP reinforced concrete beams. The accumulation of creep deformation/strain
within GFRP reinforcing bars affects the longterm performance of the comprising concrete
element.
This
manifests
in
crack
width/propagation
and
longterm
deflection
(i.e.
serviceability). As for creep rupture, it is the creep-rupture stress limit that defines the
magnitude of allowable strain for a particular GFRP bar-type and in turn the amount/cost of
necessary reinforcement. The lack of comprehensive data regarding creep of GFRP, in
general, keeps creep rupture stress limits unchanged and generalized for all commercial GFRP
61
bars, regardless their very variant performance. Remedying this issue would help designers,
encourage consumers and promote the usage of GFRP reinforcement worldwide.
In this respect, an FRP durability facility was constructed at the University of Sherbrooke over
a 300 m 2 area with a capacity of seventy bar sustained-load frames and ten beam sustainedload rigs (each rig accommodates two beams). The purpose of this facility is to study the longterm/creep behaviour of commercial FRP bars, of different types and diameters, under
different levels and forms of sustained load (axial sustained load and flexural sustained load).
Moreover, a 20 m 2 environmental chamber was constructed within to simulate the synergistic
environmental effects (e.g., temperature; humidity; alkalinity) that a loaded FRP reinforced
concrete may encounter. Over a three year period, a comprehensive series of creep rupture and
creep (axial and/or flexural) tests were conducted (See Figure 3.1 and Figure 3.2).
62

Figure 3.3 illustrates the three phases of a creep strain curve. After exhibiting initial elastic
strain, a GFRP bar goes through the first phase where creep strain grows rapidly in a short
time period. The second phase is typically lengthy and characterized by a constant flat slope.
The third phase may never occur unless high stress level is applied (fib 2007). Creep
behaviour, of GFRP, also synergizes with other adverse environmental conditions yielding a
more pronounced effect on GFRP reinforced concrete. Gaona (2003) and Youssef et al. (2008)
have conducted creep behaviour experiments on a variety of GFRP bars for lengthy periods.
The former study concluded that creep strain increase for three different types of 15.9 mm
GFRP bars, subjected to sustained load levels ranging from 23 to 27 %fUitzve, is in the order of
2 to 6.6 % of initial elastic strain frP,o; surrounding environment was controlled at (31C and
67 % relative humidity) for a test duration of six months. Showing fair agreement, the latter
study (Youssef et al. 2008) yields creep strain values that range from 2 to 8.7 % /rPio for two
types of 9.5 mm commercial GFRP bars; the test duration ranged from 4000 to 8600 hours
(5.5 to 12 months approx.); sustained load levels ranged from 15 to 45 %f U i a v e at standard
laboratory atmosphere (23 3 C and 50 10 % relative humidity).
63
1CO
o
H
Tertiary |
Secondary
Primary
Fracture
Initial Elastic Strain
Time
Figure 3.3: Typical strain history curve during creep deformation
Based on the findings of researchers such as Yamaguchi et al. (1997) and Seki et al. (1997)
and using the most conservative results available in literature, ACI 440.1R-06 design
guideline has assigned GFRP reinforcement the creep rupture stress limit of 20 % of the bar's
tensile strength. The corresponding values for CAN/CSA S806-02 and CAN/CSA S6-06 are
30 % and 25 %, respectively. Nevertheless, studies such as that of Budelman & Rotasy (1993)
indicate that if the sustained stress is less than 60 % of the average ultimate tensile strength
(fu.ave), creep rupture is less likely to occur. Greenwood (2002), however, deduced through
series of creep-rupture testing that the creep stress limit of two types of 6.4 mm GFRP rods, in
air at 23 C, is approximately 45 % f,iaVe for a 50 year structure survivability. The latter claim
is supported by the findings of Youssef et al. (2008) as well as the findings of the current
study, where two commercial GFRP bars were tested in similar surrounding environmental
conditions. It was realized that there is no sign of creep rupture at 45 % fu,ave, a realization
confirmed by microstructural analysis. On the other hand, rupture susceptibility is evident on
exhibiting 60 % fu,ave, at different endurance times.
The creep behaviour mechanism, however, is not exclusively stress-related; the surrounding
environmental conditions may affect the time to failure (See Figure 3.4). In earlier efforts, the
coupled effect of creep-inducing sustained load and creep-exasperating environmental
aggressivities (moisture and alkalinity) were studied, as regards to their effect on the creep
behaviour of GFRP bars as well as the overall allowable stress limits (Nkurunziza et al. 2005;
Greenwood 2002; ACI 440.1R-06 (report)). Combined-effect studies on GFRP render
64
important partial environmental
reduction factors related to different environmental
adversities (e.g. alkalinity and moisture). Nevertheless, it is sustained load tests (creep
behaviour and creep rupture), in air, that serve as a baseline for calculating such
environmental reduction factors.
(schematic)
After lengthy exposure times (10000 hours approx.) in air or alkaline solutions (pH 12.8) at
room temperature, most results obtained from earlier studies indicate that the modulus of
elasticity of GFRP bars and hybrid glass/carbon FRP bars exhibit no significant change (e.g.
Nkurunziza et al. 2005 and Rahman et al. 1995) under a variety of sustained load levels. As
for the loss in tensile strength, it is typically negligible for tests conducted in air yet variant for
bars saturated in de-ionized water (pH 7.0) or alkaline solutions (pH 12.8). No consistent
relationship was found between residual tensile strength and the intensity of sustained load.
CAN/CSA-S807-10 has set a standard for conducting creep behaviour tests where GFRP bars
are subjected to two levels of sustained load (20 % and 40 % of the ultimate tensile strength
fu.ave);
strain readings are to be reported at 1000, 3000 and 10000 hours from test inception. On
the other hand, the ACI 440.1R-06 guide and the CAN/CSA S806-02 code adopt the same
method for creep rupture testing of FRP bars where five sustained load levels are applied to an
FRP bar-type (of the same batch). ACI 440R-07 refers to the traditional methods for creep
65
testing. The adopted method for the current study is that of gravity loading using a lever arm so called creep frame - to permit greater loads. Nevertheless, the extraction of generalized
design criteria is hindered by a lack of standard creep test methods and reporting as well as the
diversity of constituents and processes used to make proprietary FRP products (ACI 440.1R06). In addition, little data is currently available for endurance times beyond 100 hours. These
factors have resulted in design criteria judged to be conservative until more research has been
done on this subject.
3.2
Research Purpose
This study provides essential data on the creep deformation of GFRP bars, under different
levels of axial sustained load, in ambient temperature; creep rupture takes place at higher
sustained load levels. The residual tensile properties (modulus of elasticity and tensile
strength), after a 10000 hour test-period (417 days), are observed as well. The main objectives
herein are to: (i) assess the creep behaviour of two different types of GFRP bars (of two
different brands/manufacturers) under a variety of sustained load levels, (ii) gain insight into
the GFRP bars' residual tensile properties (modulus of elasticity and tensile strength) after a
lengthy duration (417 days) and (iii) ameliorate the means by which creep tests on FRP bars
are being conducted along with highlighting the pitfalls that experimenters may fall into whilst
creep testing. All objectives above work towards mending the information gap regarding the
long-term behaviour of different types (brands and sizes) of GFRP bars exhibiting different
levels of sustained load. This study serves as a baseline to subsequent tests where GFRP bars
are tested under the coupled effect of sustained load and environmental aggressivities. It
adopts the reporting methodology determined by CAN/CSA-S807-10 and endorses it as a
universal standard for creep testing.
3.3
3.3.1
Materials
Thirty seven samples of two commercial GFRP bars are examined in this study. The first
commercial bar (GFRP-1) is made of high-strength E-glass fibres (77.3 % fibres by volume;
59.9 % fibres by weight) impregnated in vinylester resin. The bar's circular cross section has a
66
9.5 mm diameter; manufactured in a process that couples pultrusion with sand-coating along
the external surface of the bar. Likewise, the second type (GFRP-2) is 9.5 mm in diameter;
made of E-glass fibres that constitute 74.2 % of the bar's volume (55.7 % by weight). The
glass rovings (fibres) are drawn/pultruded through a vinylester resin bath; surface undulation
(helical wrapping) and sand are applied prior to thermosetting of the polymeric resin. See
Figure 3.5 below.

The main mechanical properties, as provided by the manufacturer and as verified by tensile
tests conducted prior to creep testing (average ultimate tensile strength fu,ave, modulus of
elasticity Ej and average ultimate tensile strain e uave ), are displayed in Table 3.1. Good
agreement was found between manufacturer and verified results.
3.3.2
Study Parameters
The two parameters of interest are the bar-type and the axial sustained load level. The two
bars at-hand exhibit similar characteristics in terms of fibre content and type of resin, yet each
is of a different manufacturing process. For each bar-type, four levels of sustained tensile load
(15 %, 30 %, 45 %, 60 % fUlave) were applied, giving a range of initial strain frp,o ranging from
2000 to 14000 pe. Three to five samples are assigned to each sustained load level per bartype. The maximum allowable strain for the sustained load given by ACI 440.1R-06 is 2549
and 2541 ps for GFRP-1 and GFRP-2, respectively (See Table 3.1). Corresponding values
given by CAN/CSA S806-02 and CAN/CSA S6-06 can be obtained by multiplying the latter
values by 1.5 and 1.25, respectively. The surrounding environment whilst conducting the
67
long-term creep tests is standard laboratory atmosphere (23 3 C and 50 10 % relative

humidity). The purpose of testing the material at load levels far beyond the allowable is to
explore the true capacity of such bars; the available codes and guidelines may be conservative
when it comes to GFRP reinforced concrete elements not exposed to earth and weather.
68
CO
4>
00
ci
^
1
Jh
-H
'3 oo
c4>
C
N
O o oo
43
SO
o
Ja
0)
-o <N S

-H
oo
Tl-
-H
00
Tt-
o
00
oo
in
(U
t+H
o
O
X
1)
in
CM
CN
o
v
o
oo
o
-H
oo
<
CN
00
00
ON
CN
ON
CD
-O
^ *T3
i g o
-H
Hi 43 in
O O oo
s
.>
T3
4>
-o
o
03
o
43
o
(N
IN
r--
CN
o
\o
4>
O
o
-H
82.8 VO
CO
N
03 Ô 3 Ioo
[Xh
O
0
OT
0)
ft 4>
1

4>
Oh
O
<N
I
o
c
0)
a
o
Z,
+1
ON
vd
ro
-H
(N
O
o
ON
in
in
r-
in
CN
CN
CN
00
NO
O
o
CN
-H
tT
in
-H
o
o
ON
ro
ci
in
C)
ON
CD
>
^^
-A
o
I
o
"3
rtL-L T 3 rt"
c
d
co O H-H
4)
3
O
MD
in
<
m
0
T3
O
4)
1
O
4>
ON
43
4>
(N
O
I
S3
O
G
43
CR
o
00
in
4)
4)
<
o
C
O
CO
CS
CO
T3
II
00
co"
CN
O
I
VO
L-H
O
O
T+H
00
I
CO
o
in
< a
C
4)O
<D
t-i
O
^
S3
C VO
'3
oI
i-l
03
4
C
CO
sifl ^ CO
<D <ZS
4> "to U W o
o O * 5 (L) i ' c/>
d Tt
CO "'a
i) HH
J3
C
3
u
3
CL <
00
-o
I
o
O
C
oO
Q
4>
"E,
iCO
4)
in
O
D,
X
4>
T3
<D
a,
CO
in
CN
a,
03
CL,
s
V
3O
efl
v ai'
i-i
w
^
8 S 05 ^
<D
4> O
a <
S3
CO
a
(U
Q -
4)
&
00
D
ci
C3
4+-
3
Y)
KH
CQ
u
OO
4)
ON
X
II
<
CN
CN
4)
4)
M
(D
<
m
43
<+H
o
-o
$o
0
o
(k
o
o
Tf
HH
u<
CN
00
<
<
4)
-T3
<
1
U
S3
O
T3
<L>

4)
o
CO
O
T3
44)>
I
S3
OH
co
CTF
c/T
4)
"c3
>
a
' 3
BO
KH
o
fl S
.SP
' CO O
4)
13
-A
4)
4)
43
H
4>
O
4>
a,
M
4)
(-H
4T
T3
'3
bt)
vo
oI
O
3.3.3
All samples were prepared according to CAN/CSA S806-02 and ACI 440.3R-04. The GFRP
bars, of both types, were cut into a variety of lengths (1170 mm, 1270 mm and 1470 mm) to fit
into three different frame sizes of heights 1550 mm, 1750 mm and 1880 mm, respectively. Each
one of the two ends, of a bar sample, was fitted into a 410 mm-long steel tube (grip) using
Bristar 100 expansive grout. The steel pipe-grips on both ends have a 50 mm hollow portion,
threaded on the inside, to screw/fit onto spherical nuts that keep the sample intact with the frame.
Two Kyowa 10-mm strain gauges (120 ohm resistance and a gauge factor of 2.10) were attached
- o n opposite sides- at the middle of the free portion of each sample. For gauge installation, Mbond AE adhesive was used and gauges were properly aligned in the longitudinal direction of the
bar. Each gauge has an attached 3 meter long wire, the tip of which gets connected to a portable
strain indicator when strain measurements are to be taken (Figure 3.6).
The schematic of a sustained-load frame, shown in Figure 3.7 illustrates the comprising frame
elements and location of the sample within. The aim of such frames is to have a constant tensile
load sustained along the bar's length for extended time durations. The load should be maintained
perfectly axial; assurance that no eccentricity or bending occurs to the bar is imperative. The
associated load magnifying system (the two lever arms and sustained weight-pan) multiplies the
kept-on-pan weight to reach a significant percentage of the sample's ultimate tensile capacity
fu,ave-
Prior to installation, all frames were calibrated using a standardized steel rod and five sets
70
of different weight magnitudes. Further details are available in section 3.4.4 that deals with
pitfalls in FRP creep testing.
Figure 3.7: Schematic of bar sustained load frame (magnifying frame)
The typical scenario for bar installation, in chronological order, is as follows: (i) The two lever
arms are upheld in a pseudo-horizontal position that allows placement and fitting of an FRP
sample between the base and the swivel, using threaded bolts and spherical nuts; (ii) The
necessary weight (calculated earlier from the calibration process) is placed on the pan; this may
disturb the horizontal state of the lever arms and thus cause the applied load to be eccentric onto
the bar's cross section; (iii) A trade-off takes place between manually tightening (or loosening)
the spherical nuts at the pipe-grip ends and adjusting the load calibration system at the back of
the frame; this restores the horizontality of the lever-arms (aided by high-precision spirit levels)
and consequently the concentricity of the load onto the sample.
The three sizes of frames available, at the University of Sherbrooke durability facility, adopt the
same design as in Figure 3.7. It is the length of the lever arms, though, and not the height that
influences the frame's magnifying capacity. Out of 70 frames, thirty seven were dedicated to the
71
experiments of this study. Whilst long-term testing, a verification test was conducted using two
sets of frames, each of a different size (height and arm-length). On both sets, the same bar-type
(GFRP-1) at the same sustained load level (45 % fUtCrve) was installed. The results from both
frame-sets showed were close, indicating that neither the size of the frame nor the length of the
sample has a significant impact on the outcome (See Set a and Set b in Table 3.2).
3.3.4
Measurements are taken as soon as the GFRP bars are installed under the prescribed load, then at
following times for the extended test duration (10000 hours at least). The device used is a P-3500
portable strain indicator (with an accuracy of 3 ps), manufactured by Measurements Group Inc.
Strain measurements were taken on regular basis. The horizontality of the two lever arms were
checked every time prior to strain reading (using a spirit-level on each arm) to assure that the
load is maintained concentric and that the load-magnitude is kept constant (Figure 3.8).
Figure 3.8: Regular levelling of lever arms
72
The frequency at which strain-measurement is necessary decreases with time. For the tests at
hand, readings were taken every six hours for the first two days; every day for the first week;
biweekly when the rate of change of creep strain significantly decreases. Each measurement,
indicated on the graphs (See Figure 3.9), is actually an average of the two back-to-back gauges
installed onto the bar-sample. It was not necessary to strictly abide by the reading times
suggested by the ACI 440.3R-04 test method's guide; the rate of strain increase with time was
typically low for all bar-samples. Three replicates were taken for each reading, and then
averaged, to yield a point on the sample's creep-evolution curve. As per the CAN/CSA-S807-10,
creep strain values are recorded and (Table 3.2. to Table 3.4).
After sufficing the 10000 hour duration, all samples were uninstalled from the test frames. Static
tensile tests were then conducted on the samples using an MTS tensile test machine. Residual
tensile properties, indicated as percentage of the average ultimate tensile strength ( % fUtave) and
as percentage of the average Young's modulus ( % /), were determined (Table 3.5 and Table
3.6).
3.4
In this section, the creep-strain measurements obtained during the long-term testing phase
arepresented and discussed. In compliance with ACI 440.3R-04, measurements were taken
regularly; readings at test inception as well as at specific milestones (1000 hours, 3000 hours and
10000 hours) were also recorded as per CAN/CSA-S807-10 (Table 3.2. to Table 3.4). Residual
tensile testing was carried out to determine residual tensile properties of the creep-tested bars;
this includes residual tensile strength and residual Young's modulus (Table 3.5 and Table 3.6).
Furthermore, the outcome of microstructural analysis on both commercial GFRP bars is
presented.
3.4.1
For any GFRP bar, the creep characteristics are known by monitoring the change in axial strain
with time under constant applied stress. The creep behaviour, of the two commercial GFRP bars,
is displayed in Figure 3.9 as well as Table 3.2. to Table 3.4. It was observed that for both bartypes, exhibiting standard laboratory conditions, there was no sign of creep failure for sustained
73
load levels amounting to 45 % fu,ave and less. Creep rupture, however, took place at 60 % fuave
after variant time-periods. For GFRP-l, the upper-bound creep-strain percentages (maximum
values in each dataset) after 10000 hours are 8.8, 3.8, 4.6, 7.6 and 3.6 % for sets 15 %, 30 %, 45
%-a, 45 %-b and 60 % fu,ave,
respectively. The corresponding values for GFRP-2 are 2.5, 11.8
and 12 % for sets 15 %, 30 % and 45 %/, ave , respectively (See Figure 3.9).
120
180
240
300
360
14000
420
480
540
davs
(Creep rupture failure point)

12000
"
4.3
10000
7.3%
AAA A
- - 1 5 % fu.ave
a . 8000
- * - 3 0 % fu.ave
- * - 4 5 % fu.ave
6000
55
- - 6 0 % fu.ave
3.8%
<M. 'A
4000
8.8%
2000
hours
00
1440
2880
4320
5760
7200
8640
10080
11520
12960
60
120
180
240
300
360
420
480
540
davs
(Creep ruptur efailurc point)

16000
m 1d 10/.
F
14000
12000
- - 1 5 % fu.ave
*
10000
12.1 %
a
co
- * - 3 0 % fu.ave
- * - 4 5 % fu.ave
8000
- - 6 0 % fu.ave
6000
r 5.5/
TT"
2.5/
2000
hours
1440
2880
4320
5760
7200
8640
10080
11520
12960
Time
Figure 3.9: Upper bound creep strain evolution in GFRP-1 bars (above) and GFRP-2 bars
(below) under different sustained load levels
No consistent relationship is evident between the magnitude of accumulated creep strain and
sustained load level. It is important to consider that the rate of creep-strain increase tapers down
greatly with time when extrapolating the obtained measurements over the service life of a
concrete structure (75 years). The 10000-hour period has gained consensus as the period that
74
captures most of the resulting creep strain. The variance, in replicates, of the same data-set is
significant. However, this is typical for creep tests regardless the devoted degree of precision and
caution.
Table 3.2: GFRP-1 creep-test details (15 %, 30 %, 45 % and 60
^T ,
Applied
Load
Sample
F
frpo
'
W / W
ratio
(%f u , a v e )
Cre Strain CSt in

Increase)^) aftT
ratio
(%//)
1000
hrs
1
2
3
15
/ofu ave
4
5
1
2
j u /O
3
r
a
Ju,ave
4
5
1
2
oz /"
"
Ju,ave
3
ôei a j
4
5
1
4 5 /o
fu,ave
2
(Set b)
3
4
1
60 % f ,
2
3
a
For this sample
l
/0
a v e
10000
hrs
ve)
Creep Strain-Initial
Strain ratio (^/o of actual
initial strain) after
1000
hrs
3000
hrs
2264
12.4
14.2
46
68
93
2.0
3.0
2268
12.4
14.2
101
28
176
1.2
4.5
14.4
2631
1
181
132
16.5
0.0
6.9
2306
12.6
14.5
8
75
166
0.3
3.3
2703
14.8
17.0
155
145
239
5.4
5.7
3809
20.9
23.9
-6
48
46
-0.2
1.3
5674
31.1
35.6
-166
-132
-163
-2.9
-2.3
5214
28.6
32.7
74
39
29
1.4
0.7
6021
33.0
37.8
-269
-286
-288
-4.8
-4.5
5255
28.8
33.0
114
176
199
2.2
3.3
8595
47.1
54.0
115
97
229
1.3
1.1
8195
44.9
51.4
183
232
295
2.2
2.8
8086
44.4
50.8
128
-21
83
1.6
-0.3
7438
40.8
192
179
339
46.7
2.6
2.4
8897
48.8
55.8
196
118
373
2.2
1.3
8443
46.3
381
53.0
230
406
2.7
4.5
7496
41.1
47.1
254
432
486
3.4
6.5
8149
44.7
51.2
747
757
623
9.2
9.3
7812
42.8
49.0
38
65
237
0.8
0.5
10416
57.1
65.4
179
292
-101
-1.0
1.7
11056
60.6
69.4
283
397
135
1.2
2.6
11561
63.4
72.6
391
495 b
3.4
4.3b
set, resin was used instead of cement grout to adhere the GFRP bars within the
-
steel tubes (grips).

b
3000
hrs
%f
The sample failed after 2964 hours (ie. 4.1 months)
75
10000
hrs
4.1
7.8
5.0
7.2
8.8
1.2
-2.9
0.6
-4.8
3.8
2.7
3.6
1.0
4.6
4.2
4.8
5.8
7.6
3.0
2.8
3.6
-
Table 3.3: GFRP-2 creep-test details (15 %, 30 % and 45 %/, ave ).
Nominal
Applied
Load
Sample
Number
( %
15 /o fUiave
m
o/ f
30
Voju.ave
/ir 0/
f
/oJu,ave
t j
u,ave
1
2
3
1
2
3
4
1
2
3
4
Sfrpfilz JU
ratio
ratio
frp,0
2032
1941
2105
6946
5377
9006
5600
8188
8977
9594
9237
Creep Strain (Strain

Increase) (pe) after
fu.ave)
11.0
10.5
11.4
37.6
29.1
48.7
30.3
44.3
48.6
51.9
50.0
3000
hrs
-8
29
34
758
63
37
230
313
757
686
1066
1000
hrs
30
11
35
702
78
-22
206
248
633
613
939
12.8
12.2
13.3
43.7
33.9
56.7
35.3
51.6
56.5
60.4
58.2
10000
hrs
-12
33
53
819
108
220
308
388
843
718
1111
Table 3.4: GFRP-2 creep-test details (60 % f
Nominal
Applied
Load
sfrp.ofs
Sample
Number
frp,o
u.a
ratio
(%
frp,ofefu
ratio
fu,ave)
Creep Strain-Initial
Strain ratio ( % of actual
Creeprupture
time
(hours)
1000
hrs
1.5
0.6
1.7
10.1
1.5
-0.2
3.7
3.0
7.1
6.4
10.2
3000
hrs
-0.4
1.5
1.6
10.9
1.2
0.4
4.1
3.8
8.4
7.2
11.5
10000
hrs
-0.6
1.7
2.5
11.8
2.0
2.4
5.5
4.7
9.4
7.5
12.0
,ave).
Creep
Strain
Increase at
rupture
time (pe) a
1
14954
94.2
13.8
80.9
2
11512
72.5
56.8
62.3
60%
3
13995
75.7
88.1
231
fu.ave
54
4
10839
68.3
58.6
Creep strain readings were taken manually; the measurements taken
Creep StrainInitial Strain

percentage at
rupture time
2.2
332
1.3
155
1488
10.6
6.4
696
are expected to be less than
the actual value at rupture time.
Creep coefficients can be determined by linearizing the creep-strain curve into strain versus log
time. When plotted in such a manner, the curve becomes a linear relationship. The equation for
the total strain of the commercial bars can be written as:
Zfrp(t)=P log t + efrpto
76
(2.1)
where
Sfrp{f)
is the total strain in the material after a time period t (10000 hours for this study),
frPio is
the initial (elastic) strain value and /? is the creep rate parameter that is equal to
ds(t)/
dt.
Using linear regression for the obtained data, the creep coefficients for the upper boundary
values of GFRP-1 samples were determined as 44.8, 28.7, 116.0 and 149.7 for 15 %, 30 %, 45
%, and 60
%fu,ave,
respectively. Similarly, the creep coefficients for GFRP-2 samples were 9.7,
68.2, 273.2 and 504.9 for 15 %, 30 %, 45 %, and 60 % f
,ave,
respectively. It is evident that the
coefficient [1 increases significantly with the increase of applied stress. At 60
%fu,ave,
however,
one sample of the GFRP-1 bars as well as all GFRP-2 bars exhibited creep rupture at variant
time durations (See Table 3.2 and Table 3.4).
3.4.2
After the elapse of the test duration (10000 hours), all bars were dismantled from their
comprising frames and tensile tests were conducted to obtain residual mechanical properties. For
GFRP-1 samples the rupture mode was burring of the fibres, whereas GFRP-2 exhibited helical
wrap unwinding. The average residual strength was barely affected creep tests (Figure 3.10). The
max percentage loss, in strength, for GRFP-1 bars was 4 % for 60 % f
ll>ave.
The corresponding
value for GFRP-2 was 5 % for 45 %/L,ave (Figure 3.10). It is worth noting that GFRP-1 bars have
a higher fibre content that allowed more of its samples to sustain 60 % f ,
u aVe-
As for the modulus
of elasticity, the residual values showed barely any change from the original values. The average
residual modulus for all bars ranged between 44.3 and 48.5 GPa; whereas the original GFRP-1
fu.ave
is 46.9 1 . 2 GPa. Even for the data-set exhibiting the maximum applied stress (60 %
fUlave),
from which one sample ruptured after a 3000 hour duration, the Young's modulus remained
unaffected for the surviving bar-samples. Likewise for GFRP-2; there was no change in the
modulus of elasticity after 10000 hours of loading.
77
110
100
"J 90
an
-Q
70
6n
I>
a 50
TS. 40
t> 30
20
M
10
0
A
w
i
i
30%
15%
45%
60%
110
"i
-
.:*
100 -
90 -
1--
80
70
S
'3
s
M
d
m
60
55
20
50
40
30
10
0
15%
30%
45%
Nominal stress applied lor a 10000 hour period (%/L
Figure 3.10: Residual tensile strength for GFRP-1 bars (above) and GFRP-2 bars (below) after
the 10000 hour duration.
78
o
DO
cd
LTi 00
On
O O O O
O O O O
0)
o
CO
CO
IT)
CO
CN
O O O
ON
00
-H
o
(N
o
(N
ON
ON
00
CO
o
CN
o
ON
VO
^T
OO
ON
ON
T>
R ^ CN
CN| T--.
T}-'
1)
Ph
CO
2 u 3
t m H
c o <-H CN On c o
K h r^
<8
'+3
so
0s
o
00
CN ^
NO ON
in 00
(>
t ^ ^f ^f
00 in oe
co
^
co
CN
co
'st Tl-
00
t-^
CO
^
0 0 NO
CO ^
10
^
CO CO
CN
R-;
^t t
Tf
-rj00 10
iri r^ K
vo (N h 00
in
00 co
CO
>0
CO
Tl-
00 00 00 Tt
ON
NO
CO CO T f CO
^
CO Tl-
"St
no
00
TL1
r--
00
in
CO
CO
<D
r-; o o p
tn
o
,o
c

iri
iri
rf
^H rt
^H
CN
00
ON
Oi
o 00
o
T^
co
C
0)
0s
flu
3
M
<D
t-i
Tf
O
^
co
rt
co
O
O O O O
o
a>
Dh
00
ON
CO CO
uo ^ r- n
in
in in
rON
ON
ON
ZZ,
ON
10 p
n
iON
NO
IN
ON
ON
<>5
in
ON
ON
r-
l-l
<D
'w
C
<D
00
"2 fl
OH

CN
CO
_
ON
(N i/i
i/i vi
ifl
^
00 00 00 00 00
R-
ON
IN
LO
UO
(N 1O
V)
11 1
1
00 00 00 00
00
CN
00
O_
ON
T}- in
CO MD
00 00
ON
RF
ON
CN
ON
m
vo
00 )0
00
00
CO
NO - H
^
^F
Tf
00
ON
ON
CN
CO
CN
NO
00
o
tjK
00 00 00
CO
<D
Dh
in
co
<D
I
H
1
<
3 &
00
r-H vq 00
CN
CN
Tt
CN
- H (JO
CO CN
M
CO
OO
CN
CN
CO
TT
*-<
r^ rf ^t 00
^ ^ ^f
r- 00
(N
Tj"
rco
M
NO
CO ^J"
CN
CO
NO
a
u
o '
1> I
a
<N
CO
CN
T3

bO
a
a
>
<
s-
<D
>
O
CO
<
Ph
vf
m
0)
Of)
cd
u
>
<
Hh
to
S^
ed
sU
>
0s
ir>
u00
K
!
1-1
(>
U
00
<
<
<L>
BO
at
E(D
O
S-i
in in
O
t> co in
ON
ON
Tj"
M
CO
<U
OH
CO
S u 3_ ^ co in
OO T f
in
$ o O no
'sT f
JQ vo m vo
M
2 Oin N^ T} -
on f -
f )
<4
CO
On CN
in oo
ON
CN
CN
<N
m
co co in 5
r-- r-- no
CN
ON
ON
^R
ON
Tt no
in -^j- co co
<D
R 00 C N
^ t^ VD NO
00
CN
CO
*>
<L)
CD
S
4>
A

A ,
ON
ON
"3-
ON
On
Tt rf
K CN Tt-'
On On
no
^
.
O
R--
ON
NO
ON ON 2
on
On
oo
KJ
in m tN rH
bo
^ &cd a <>
co H S CN So
4>
^
P< CO 00
"O
a 8
"S"
NO
IN
vo NO
oo oo
vo
RN
< 3 &
00
W
ON
CN'| vo P
oo r--
ON
o co
ON
CN
00
NO
00
CO Ô ON CO
oo ^ 00
f -n- in
CO
a4>
I
CN
CN
CN
CO
CO

C2
4-
<
cn 4)
3
CO >0)
a-s
a
o
(N
4)
00
ea
L-H
4)
>
<rU O O m <
CN
Oh
K
O
4)
00
cj
4>
CN
OX)
al-H
4)
>
4>
>
CO
<
O m
ti-
<
3.4.3
The formation of microcracks in the resin and the debonding at the interface of fibres/matrix
are the most common phenomena occurring in a GFRP material under sustained load and/or
adverse environment. In this respect, investigation (Scan Electron Microscopy - SEM analysis) took place using selected samples of the tested bars after the 10000 hour test
duration to have a better understanding of the causes behind strength loss, if any.
(c)
Figure 3.11: Magnified samples' cross section after exhibiting 10000 hours of loading: (a)
GFRP-1 at 45 %fu
ave\
(b) GFRP-2 at 45 %; (c) GFRP-1 at 60 %fu,ave (top, middle and bottom,

respectively)
Micrographs, in Figure 3.11, show magnified cross-section images of the rupture zone of
GFRP-1 bars that exhibited 45 %/UjOTe and 60 %f UMve as well as GFRP-2 after exhibiting 45 %
fu,ave,
for a 10000 hour test duration. Results show that both bar types show no signs of
debonding between the fibres and vinylester resin, and no induced microcracks, for the lower
81
sustained load level (45 % fu,ave)-
As for GFRP-1, subjected to 60 % fu,ave,
thin voids appear
around the fibres; one sample of the same set ruptured after 4.1 months. All GFRP-2 bars
subjected to 60 % fu,ave, ruptured after durations as early as 13.8 to 231 hours. Subsequent
SEM tests have also indicated that the external surface of all bars remained unchanged.
3.4.4
Pitfalls in Creep Testing
The experimenter was confronted by many challenges whilst conducting the aforementioned
tests. Creep-testing is intrinsically sensitive and prone to error. This error, though, can be
minimized by caution and avoiding pitfalls that may render creep testing of FRP bars
worthless. A prepared GFRP bar sample, suspended under its own-weight, typically exhibits a
strain increase in the order of 50 to 100 microstrains; so may a GFRP bar-sample under axial
service load, after a 10000 hour duration. This meager value appears more insignificant when
compared to the initial induced strain. As found in earlier efforts/literature, creep-strain is
typically in the range of 2-10 % eUtave. Due to this inherent nature, it is of dire importance to
highlight all aspects/pitfalls that long-term (creep) researchers may fall into and provide
remedy. This section discusses all potential sources of error and means of early prevention or,
at least, minimization. The information provided below is not typically found in written
documents; it is the merit of numerous trials and errors, experience and acquired dexterity at
the University of Sherbrooke FRP durability facility.
Sound sample-preparation is the cornerstone of a successful and informative long-term

experiment. Otherwise, it may be the source of loss after lengthy test duration. It may happen,
when samples are tested for residual properties after 10000 hours of sustained loading, that
slippage would occur in the grip. This essentially means that the strain-readings obtained are
indicative of the ongoing creep within the bar as well as the creep of resin in the steel sleeve.
This adversity was remedied by using Bristar 100 expansive grout to well adhere the bar to the
steel sleeve. The typical resin mix used earlier consists of epoxy resin (west-system 105), fast
hardener (west system 205) and sand (accuasand) holding a ratio 100:20:60 respectively. Its
convenience lies in the capability of recuperating the steel grips after testing; however, the
trustworthiness of the results is compromised.
82
It is important to keep the FRP bar concentric within the comprising steel grip/sleeve at both
ends. On the inner end, the steel tubes are soldered and a concentric hole, having the same
diameter of the GFRP bar, is drilled. On the outer end, the bar's tip is also maintained
concentric within the steel tube aided by pre-customized plastic rings. The concentricity of
both ends keeps the bar concentric along the steel-tube's shaft. This in turn guarantees that, on
bar-installation in magnifying frames, any applied load will be maintained axial along the
bar's length. As to gauge installation: An extreme level of care should be given to proper
removal of sand-coating without excessive abrasion of the bar to avoid localized stress at the
gauge. Two gauges are to be installed back-to-back, along the bar's length, at the centre of the
bar. Averaging both strain-gauge measurements minimizes the error due to a probable
curvature in the bar.
The design philosophy of a magnifying frame implies that a small magnitude of weight is kept
on the weight pan (Figure 3.7); the two lever arms then magnify the weight significantly and
maintain it axial onto the erected bar-sample. However, without calibration the magnifying
factor cannot be precisely known. The calibration process takes place in two phases. In the
first phase, a standard steel bar is subjected to different tensile magnitudes, on a universal
testing machine, and the resulting deformation in the bar is recorded. A linear regression line
(Steel rod calibration) is drawn using the obtained stress-strain data. For phase 2: The same
steel bar is used for each magnifying frame. Five different weights are placed, one at a time,
and the resulting deformation on a standard steel rod is again recorded. The generated linear
regression, this time, is a relationship between applied pan-weight and resulting strain in the
steel bar. Both regressions can be clustered into one single linear relationship between the
weight applied on the pan and the resulting force on the bar. For proper creep-test results later
on, the steps above should be executed with great care. Soundness of calibration can be
verified during installation of a GFRP bar. Once a GFRP bar is installed in a magnifying
frame under a particular load, the ratio between the initial strain and average ultimate tensile
strain of the bar
(sfrp,o
l<-u,ave)
should equate to the sustained load percentage anticipated from
prior calculation. In this manner, the initial strain was verified for all tested samples within
this study.
83
On installation, it should be ensured that a GFRP bar is erected in a perfectly vertical manner
and that the magnifying arms are perfectly horizontal. The spirit levels used to maintain the
arms' horizontality should be of high precision. A potential source of error is the base of the
self-reacting frame. If not fully intact with the floor, the tilting of the base implies the frame's
entire inclination and consequently further error in strain measurements. It is highly important
to take frequent measurements and observe the lever-arms' horizontality for the first fourty
five minutes after the prescribed load is attained. Along the 10000 hour duration the lever
arms would relax, from time to time, losing their horizontality. Re-leveling the arms maintains
the sought load constant and axial onto the bar. Thus, delayed checking on arm horizontality
affects the test's integrity.
The following recommendations are to ameliorate creep-testing and minimize error: (i) Turn
the top threaded bolts, holding the bar-sample to the swivel (Figure 3.7), into load cells by
installing strain gauges on them; this enables monitoring of the actual applied tensile load onto
the bar and allows for restoring of load during the test; (ii) The presence of a data acquisition
system (connected to all strain gauges and load cells) would yield more accurate
measurements and minimize the errors caused by human measurements; (iii) Continuous
monitoring and leveling by the experimenter is inevitable.
3.5
Creep behaviour tests were conducted on thirty seven GFRP bar-samples, of two different
commercial brands, over a period of 10000 hours (417 days). The E-glass/vinylester GFRP
bars, 9.5 mm in diameter, were tested in air (23 3 C) at different levels of axial sustained
load, nominally (15 %, 30 % 45 % and 60 % of the average ultimate tensile strength f ,ave),
u
in
air. All obstacles/pitfalls dealing with FRP creep tests were discussed and means of
ameliorating creep test results were proposed. The following conclusions can be drawn from
this phase of the research study:
1- Creep strain, for both commercial types, shows no significant strain increase for
sustained load levels 15 % and 30
%fu,ave
after 10000 hours. At 45
%f,ave,
however,
the upper-bound (maximum) creep strain percentage for GFRP-1 and GFRP-2
amounts to 7.6 % and 12 % of the initial strain frPin , respectively; accumulation of
84
creep strain decreases asymptotically with time. Aside from the exceptional upperbound values, creep strain is considered low and inconsequential; the creep coefficient
(percentage) of concrete itself varies between 150 % and 400 %.
2- At 60 %
fu,ave,
creep rupture was witnessed in one sample of the GFRP-1 bars; all
GFRP-2 bars ruptured at periods ranging from 13.8 to 231 hours.

3- Microstructural analysis was conducted on samples from both commercial bars that
exhibited 45 % fu,ave for 10000 hours. No microcracks were found indicating that 45 %
fu,ave is an accepted creep rupture limit for the bars, in air.
4- The residual tensile strength and modulus of elasticity, for all samples that survived
the 10000 hour duration, were not changed (almost retaining their full strength).
5- The presence of outliers and uncertainty is inevitable; creep tests rely heavily on
human monitoring. However, such issues can be minimized when pitfalls regarding
creep tests are treated.
85
Chapter 4: Long-Term Performance of Different Types of GFRP Bars under Sustained Service Load
CHAPTER 4
LONG-TERM PERFORMANCE OF DIFFERENT TYPES OF GFRP BARS
UNDER SUSTAINED SERVICE LOAD
4.1
Introduction
In the previous chapter, two commercial GFRP bars were examined. Different sustained load
levels were applied (15 %, 30 %, 45 % and 60 % f
u>ave)
for lengthy time duration (10000 hours).
Results showed that 45 % of the average ultimate tensile strength, / ave, is a safe creep rupture
stress limit for the aforementioned materials. ACI 440.1R-06, CAN/CSA S806-02 and
CAN/CSA-S6-06, however, have assigned GFRP creep-rupture stress-limit values ranging from
20 % to 30
% fu,ave-
In this chapter, a comparative study takes place for six different types and
sizes of GFRP bars. The applied axial stress is within the allowable stress-limit levels given by
North American standards.
4.2
Research Purpose
This study mainly aims to address the lack of data regarding the creep deformation of six
different types and sizes of GFRP bars under two different levels of service load. The residual
tensile properties (modulus of elasticity and tensile strength), after a 10000 hour test-period (417
days), are determined. The main objectives herein are to: (i) assess and compare the creep
behaviour of six different types of GFRP bars under service load and (ii) gain insight into the
GFRP bars' residual tensile properties (modulus of elasticity and tensile strength) after a long
time period (417 days). This study serves as a baseline to subsequent tests where GFRP bars are
tested under the coupled effect of sustained load and environmental adversities/aggressivities.
4.3
4.3.1
Materials
Fifty two samples of six commercial GFRP bars, pertaining to four different sizes and three
different manufacturers (Figure 4.1), are examined in this study. For manufacturer A, GFRP, is
made of high-strength E-glass fibres impregnated in vinylester resin. The bar's circular cross
section is manufactured by a process that couples pultrusion with sand-coating along the external
86
surface of the bar. The GFRP bars of manufacturer B are also made of E-glass fibres that are
drawn/pultruded through a vinylester resin bath; surface undulation (helical wrapping) and sand
are applied prior to thermosetting of the polymeric resin. As for manufacturer C, the core of the
rebar is a patented pultrusion process. In this continual process high-strength glass fibres are
drawn through a tool where they are impregnated with liquid synthetic resin. The impregnated
fibres are sent through a profiling mould and are hardened. The ribs are ground and hardened
into the bars. Further detail as to the physical properties of the six bar-types and their appearance
are available in Table 4.1 and Figure 4.1, respectively. The main mechanical properties, as
obtained through tensile tests conducted prior to creep testing and as indicated by the
manufacturers (average ultimate tensile strength
ultimate tensile strain
u,ave),
K a v e
modulus of elasticity Ef and average
are given in Table 4.2 and Table 4.3, respectively. The measured
properties show good agreement with those of the manufacturer.
87
Table 4.1: Physical properties of the six types of GFRP bars

GFRP-1
9.5 mm
Sand
coated
GFRP-2
9.5 mm
Helical wrap/
sand coated
GFRP-3
12.7 mm
Sand
coated
GFRP-4
12mm
56.8
50.6
64.5
7.0
6.58
28.6
GFRP-5
15.9 mm
Sand
coated
GFRP-6
15.9 mm
Helical wrap/
sand coated
75.2
66.6
57.5
6.7
5.7
8.5
7.8
33.7
27.2
18.2
29.9
43.5
1.94
1.89
1.99
2.25
2.01
1.95
0.31
N.A
0.15
0.12
0.18
0.52
Glass transition
temperature (C)
105
106
105
170
112
80
Cure ratio (%)
98
99
98
>95
95
95
Surface
Fibre content
( % volume)
Longitudinal coefficient
of thermal expansion
(x 10"6 C)
Transverse coefficient
(x 10"6 C)
Density
(g/cm 3 )
Water absorption
(%)
88
Ribbed
<
& C3N
-
-H
o
IT)
co
<
in
o
o
CO
c-NO
NO
NO
<
<
CO
CO
CO
CO
CO
CO
<N
NO
m B
p-t
K 0\
in
-H
o
<
in
in
o
CN
in
-H
00
<
2
in
o
NO
CO
oo
(N
ON
NO
in
<N
d
<
I
S
a
<
o
v
<u
r--
-H
oo
oo
o
NO
CN
NO
00
(N
o
ON
o
00
ON
CN
x
T3
ID
13
O
i
<
co
CN
-H
gn
O
N
T3
.g
IT)
CO o ^
cd
CO
00
n
ô
Tf
-H
(N
00
^t-
o
On
ON
OO
-H
rsi
oo
in
in
OO
in
(N
CD
NO
<N
CN
cd
^
O
Ph
O
0)
co
(U
<d
X
O o^
cID
/2
a
a
o
o
0)
oo
r-
CO
-H
N
-H
co
NO
<3-
<
in
CN
TI
NO
in
oo
II
a z
* <
s,
NO
rON
NO
o
ON
CN
in
00
X
(N
CO
<N
CN
NO
Tl<N
NO
NO
-H
in
in
oo
oo
u
13
o
<D
'O
"3
00
N
oO
I
ON
NO
CO
CO
o
CN
3,
co
X
3
CN
d
II
ID
co
s3
<d
T3
0)
ID
cd
SJ
o
s
<+H
o
CO
X)
co
M
o<
CN
a>
3
o
Sh
^
CJ
CO
<
CU
3/
<D
Dh
cc3
Xi
oo
00
G
FH
a
o
f1-1
T3
<D
S<3
cSH
o
<L>
13
>
C
CO
O
oSH
M
u
oCD
C
D3/
CO IC
U
cd
w
w
I
D
3 l-i
PH
i
IH
a,
13
S
>
IhD
W
O
a^H NO <2
CO N
cd
oI o
o
c
S
o
o
T3
tIw
cd
O f ID
D
CO
S-I
S-i c ts
ck 2
S-H
<D u o
C
O
<D cd <d /w "O
on o
U
H
CO o 5+
u
w
H PH
o
in i o <D T3 o
a)
-L
ID
Tl"
"sf
o
O
^
5
c
s
C
Xcfl (SD3 Tf M ID1 .3
'xjn
<D
3 HH w ' a
M <43H CO ID
i
H
<D u
i u
o
aID S3
< 3-a
C C O CO <
=3
>
00
-t-j
O
o
r(
.P
co
C
I
ID
co
<D
<D _00 XI
<D
H
Q
H <D
T3
~^
T3
IcD
o
!<D
i
ID
o
&
o
CD
s > l-i
S3
2 2
cCoO
Oh
U
D
SH
- <
ItD
O
S
o
cd X
S3
a
ID O
S3
CN
NO
>
ID
T3
T3
<
uu
I
a
00
<D
S3
'co 0)
C
O X
I)H
D
U 00
XI (D CO
-a
13
S3
I
D
<D CI O
Oh
O
<a C
ID
oo ao
X
o
(D
I<D
o 13
& >
ID
13
ID
X
<D
0
a X O
&
<o -o
1 IgD U
x
fl
ID
oCI
(D
X
X
o H
co
T3
(D t- I>D
<D ID
'o
co
oco
O
CO J C
ID
1
CO
O
1I
ID c3
X
>
ts
X
T3
<D
"d
<D
o 13 O
<+H o
c -3
a i CI
o
+
'.O
-
ooCO
o I3D
S
cs
(D
-T3
> J3
ID 13
13
>
U
cd
X
X
H H 13
T3
NO
o
tk
o
TjHH
<
<D
kH
ID
Oh
CO
CD
S
"3Eh
1CO
<D
+->
l+H
CO
3
O
Sh
S3
cID
d
g
ID
ON
00
CO
cd
T3
ID
S3
t+H
CO
+->
CO
JD
13
o
CO
S
3
-T3
og
t3
I(D
D
"cS
d
H
cd
00
S-i
O
S3
CD
ID ID
X
H
0
CO
s<D 5
OH
-H
00
Tt-
CO
tt
NO
CN
rr
in
-H
-H
m
<N
r-
CN
NO
Timl-
Tf
rin
m
CO
CN
r-
CO
<D
cind
& i
E <N
CO o
U
'co
C
0>
CO
<D
ft
Tiro
CO
NO
NO
CO
CN
tJr-
00
NO
NO
CN
in
in
CN
NO
in
NO
CN
CO
NO
o
in
NO
CN
NO
On
-H
o
-H
On
CN
NO
CN
in
CN
CN
o
(N
O
O
Tt"
o
CN
m
00
NO
00
CN
CO
-H
Tl"
00
(Jh
O
00
-H
rm
00
00
Tt
Tt-
t1-
CN
(N
Ph
O O
<D
co
<L>
<L>
-H
t|-
in
+1
T
OO
Tt"
00
00
in
CN
in
t^
CN
CN
r--
-H
-H
NO
ON
CN
NO
CN
TT
rn
in
CO
ON
in
Titl-
es!
00
ro
*CO
^
X o
ON
Tl-
in
CN
X
CN
o
II
tf 'L,
II o
<<
to
OH
o
o
 r j O
3 3 
OH
O
S-l
-H
00
Tf
>
-H
10
CO
NO
NO
00
-H
o
T3
O
gJ
co
oj
CN
O
C(H
'3
-O
0
co
CN
-H
13
O
H
G
op
I I3 SU
tS
<U
CO
O =5. < CO
'co ^v
4.3.2
Study Parameters
The parameters in this phase are the bar-type, bar size/diameter, and the axial sustained load
level. For each bar-type, two levels of sustained tensile load (nominally 15 % and 25-30 % fUiave)
were applied, inducing a range of initial strain frPio ranging from about 2000 to 7000 [ie. Four to
five samples are assigned to each sustained load level per bar-type; the maximum allowable
sustained load level indicated within the ACI 440.1R-06, the CAN/CSA-S6-06 and the
CAN/CSA S806-02 guides, for GFRP reinforcement, are 20 % fu>ave, 25 % fUMVe and 30 %fu,ave,
respectively. The surrounding environment whilst conducting the long-term creep tests is
standard laboratory atmosphere (23 3 C and 50 10 % relative humidity).
4.3.3
All samples were prepared in accordance to CAN/CSA S806-02 and the ACI 440.3R-04 guides.
Each one of the two ends, of a bar sample, was fitted into a 410 mm-long steel tube (grip) using
Bristar 100 expansive grout. The steel pipe-grips were screwed onto spherical nuts that keep the
sample intact with the frame (See Figure 4.2 and Figure 4.3). Two Kyowa 10-mm strain gauges
(120 ohm resistance and a gauge factor of 2.10) were installed - o n opposite sides- at the middle
of the free portion of each sample. Strain measurements were taken using a portable strain
indicator.

The schematic of a sustained-load frame, shown in Figure 4.4, illustrates the comprising frame
elements and location of the sample within the test frame. The aim of such frames is to have a
constant tensile load sustained along the bar's length for extended time durations. The load
should be maintained perfectly axial assuring that no eccentricity or bending occurs. The load
magnifying system (the two lever arms and sustained weight-pan) multiplies the kept-on-pan
weight to reach a certain percentage of the sample's ultimate tensile capacity fu,ave-
91
Figure 4.3: Sustained load frames comprising GFRP bars
Prior to sample installation, all frames were calibrated using a standardized steel rod, five sets of
different weight magnitudes and a strain indicator. The typical scenario for bar installation, in
chronological order, is as follows: (i) The two lever arms are upheld in a pseudo-horizontal
position that allows placement and fitting of an FRP sample between the base and the swivel,
using threaded bolts and spherical nuts; (ii) The necessary weight (pre-calculated on the basis of
92
frame calibration) is placed on the pan; this may disturb the horizontality of the lever arms and in
turn apply eccentric load onto the bar's cross section; (iii) A trade-off takes place between
manually tightening (or loosening) the spherical nuts at the pipe-grip ends and adjusting the load
calibration system at the back of the frame; this step restores the horizontality of the lever-arms
(aided by high-precision spirit levels) and consequently the concentricity of the load onto the
sample.
4.3.4
Strain measurements are taken once the GFRP bars are installed under the prescribed load, then
at following regular intervals for the extended test duration (10000 hours ~ 417 days). The
device used is a P-3500 portable strain indicator (with an accuracy of 3 (as), manufactured by
Measurements Group Inc (See Figure 4.5). The horizontality of the two lever arms are checked
every time in advance (using a spirit-level on each arm) to assure that the load is maintained
concentric. The frequency of the strain readings was typically as follows: every six hours for the
first two days; every day for the first week; biweekly when the rate of change of creep strain
significantly decreases. Each measurement, indicated on the graphs (Figure 4.6 and Figure 4.7),
is actually an average of the two back-to-back gauges installed onto the bar-sample. As per
CAN/CSA-S807-10 requirements, creep strain values are recorded and presented at 1000, 3000
and 10000 hours (Table 4.4 and Table 4.5).
Figure 4.5: P-3500 portable strain indicator

After sufficing the 10000 hour duration, all samples were uninstalled from their comprising
frames. Static tensile tests were then conducted on the samples using an MTS tensile test
93
machine. Residual tensile properties, indicated as a percentage of the average ultimate tensile
strength ( % fu,ave)
and as percentage of the average Young's modulus ( % Ej), are presented in
Table 4.6 and Table 4.7 in the discussion section below. Some samples were reserved for
micro structural analysis to know if the matrix or fibre-matrix interface within the bars were
affected after the lengthy duration.
4.4
In this section, the creep-strain measurements obtained during the long-term testing phase are
plotted (Figure 4.6 and Figure 4.7) and discussed. In compliance with the ACI 440.3R-04 guide,
measurements were taken regularly; readings at test inception as well as at specific milestones
(1000 hours, 3000 hours and 10000 hours) were also recorded as per CAN/CSA-S807-10 (Table
4.4 and Table 4.5). A following phase is that of residual tensile testing from which residual
tensile properties of the creep-tested bars are obtained; this includes residual tensile strength and
residual Young's modulus (Table 4.6 and Table 4.7). Furthermore, the outcome of
micro structural analysis of the six commercial GFRP bars is presented.
4.4.1
For any GFRP bar, the creep characteristics are known by monitoring the change in axial strain
with time under constant applied stress. The creep behaviour of all GFRP bars is displayed in
Figure 4.6 and Figure 4.7 as well as in Table 4.4 and Table 4.5. It was observed that for all bars,
exhibiting standard laboratory conditions, there was no sign of creep failure for the assigned
sustained load levels. For sample sets at nominal tensile stress of 15 % fu,ave,
the upper-bound
creep-strain percentages (maximum values in each dataset) after 10000 hours are 8.7, 2.5, 3.2,
4.1, 15.7 and 8.3 % for bar-types GFRP-1, GFRP-2, GFRP-3, GFRP-4, GFRP-5 and GFRP-6,
respectively (Figure 6). These values represent the accumulated to initial elastic strain. The
corresponding values for (25-30 %) fU:ave are 11.8, 3.8, 12, 5.6, 15.8 and 8.6, respectively (Figure
7). The values displayed above are significantly greater than the average value for each data set,
in certain cases. In general, no consistent relationship is found between the magnitude of
accumulated creep strain and sustained load level.
94
120
180
240
300
360
420
480
540
5000
GFRP-1 15%
' 4000
- X GFRP-2 15%'
3000
2000
2.5%
1000
0
1440
2880
4320
5760
7200
8640
10080
11520
12960
60
120
180
240
300
360
420
480
540
5500
A GFRP-3 15%
' 4500
GFRP-4 15% -
3500
2500 ^ U i . * ! i
A 4
4.1%
1500
1440
2880
4320
5760
7200
8640
10080
11520
12960
60
120
180
240
300
360
420
480
540
6000
GFRP-5 15%
'w 5000
a 4000
53 3000
\f
oCOOO
-JJ 0 o
>
'J1
o o
15.7%
8.3%
2000
1440
2880
4320
5760
7200
8640
10080
11520
12960
Time (hours/days)
Figure 4.6: Creep strain evolution for samples at 15 %fUiClve: 9.5 mm GFRP bars (above); 12 mm
and 12.7 mm GFRP bars (middle); 15.9 mm GFRP bars (below)
120
180
240
300
360
420
9000
480
540
GFRP-1 30%
- GFRP-2 30%
8000
11.8%
7000
6000
3.8%
5000
1440
60
2880
120
4320
180
5760
240
7200
300
8640
360
10080
420
11520
480
12960
540
5500
4 GFRP-3 25<X
"a 5000
GFRP-4 25/!
I
"a 4500
v v t
12%
5.6%
4000
3500
1440
60
2880
120
4320
180
5760
240
7200
300
8640
360
7000
10080
420
11520
480
12960
540
GFRP-5 30%
6500
c
6000 0
55
o o 15.8%
0
"J
8.6 %
5500
5000
1440
2880
4320
5760
7200
8640
10080
11520
12960
Time (hours/davs)
Figure 4.7: Creep strain evolution for 9.5 mm GFRP bars (above) at 30 % fu,ave', 12 mm and 12.7
mm GFRP bars at 25 % fu,ave (middle); 15.9 mm GFRP bars at 30 %f a v e (below)
95
It is important to consider that the rate of creep-strain increase tapers down greatly with time
when extrapolating the obtained measurements over the service life of a concrete structure (75
years). The 10000-hour term has gained consensus as the period that captures most of the
resulting creep strain. The variance, in replicates, of the same data-set is significant. However,
this is typical for creep tests regardless of the devoted degree of precision and caution.
Creep coefficients can be determined by linearizing the creep-strain curve into strain versus log
time. When fashioned in such a manner, GFRP materials approximate to a linear relationship.
The equation for the total strain of the commercial bars can be written as:
Efip (t) = fi log
t + Efip,o
(3.1)
where jrp (t) is the total strain in the material after a time period t (10000 hours for this study),
Efrp.o is the initial (elastic) strain value and [3 is the creep rate parameter that is equal to de(t)/dt.
Using linear regression for the obtained data, the creep coefficients for the upper boundary
values of GFRP samples, installed under nominal 15
%fu,ave
were determined as 46.7, 15.7, 23.5,
38.5, 138.1 and 77.0 for GFRP-1, GFRP-2, GFRP-3, GFRP-4, GFRP-5 and GFRP-6,
respectively. Similarly, the creep coefficients for nominal 30 % fUiave samples were 32.5, 216.6,
105.9, 98.9, 106.9 and 344.7 for the same GFRP bar-types respectively. It is evident that the
coefficient /3 increases significantly with the increase of applied stress.
96
Table 4.4: Creep-test details of GFRP bars subjected to 15
Bar Type/
Diameter
GFRP-1
9.5 mm
( 1 5 % fu.ave)
GFRP-2
9.5 mm
(15 %/,)
GFRP-3
12.7 mm
( 1 5 % fu.ave)
GFRP-4
12 mm
( 1 5 % fu.ave)
GFRP-5
15.9 mm
(15
%f,ave)
GFRP-6
15.9 mm
(15
%fUMVe)
Sample
No.
1
2
3
4
5
1
2
3
1
2
3
4
1
2
3
4
1
2
3
4
1
2
3
4
Elastic
Strain
S
frp,0
2264
2268
2631
2306
2703
2032
1941
2105
2536
2525
2304
2497
2865
2766
3412
2775
2495
2789
2723
2497
2959
2848
3657
3298
Jrp.O/
Su.ave
ratio
{%
fu.ave).
12.4
12.4
14.4
12.6
14.8
11.0
10.5
11.4
16.2
16.2
14.8
16.0
13.4
12.9
16.0
13.0
14.1
15.7
15.4
14.1
15.4
14.8
19.0
17.1
,*
Efrp.OI fu
ratio
( %/fu)
Creep Strain (Strain

Increase) (ps) after
1000
hrs
46
28
1
8
155
30
11
10
8
26
-34
50
5
17
74
42
84
47
86
132
349
193
424
282
14.2
14.2
16.5
14.5
17.0
12.8
12.2
13.3
19.9
19.8
18.1
19.6
17.3
16.7
20.6
16.8
17.2
19.3
18.8
17.3
16.8
16.1
20.7
18.7
97
%fUMve-
3000
hrs
68
101
181
75
145
-8
29
-4
22
47
-7
83
-8
44
86
66
114
78
120
165
394
237
480
313
10000
hrs
93
176
132
166
239
-12
33
53
51
92
15
102
19
61
110
72
145
118
134
207
464
276
573
375
(Group 1)
Creep Strain-Elastic
Strain ratio ( % of actual
1000
hrs
2.0
1.2
0.0
0.3
5.7
1.5
0.6
0.5
0.3
1.0
-1.5
2.0
0.2
0.6
2.2
1.5
3.4
1.7
3.2
5.3
11.8
6.8
11.6
8.6
3000
hrs
3.0
4.5
6.9
3.3
5.4
-0.4
1.5
-0.2
0.9
1.9
-0.3
3.3
-0.3
1.6
2.5
2.4
4.6
2.8
4.4
6.6
13.3
8.3
13.1
9.5
10000
hrs
4.1
7.8
5.0
7.2
8.8
-0.6
1.7
2.5
2.0
3.6
0.7
4.1
0.7
2.2
3.2
2.6
5.8
4.2
4.9
8.3
15.7
9.7
15.7
11.4
Table 4.5: Creep-test details of GFRP bars subjected to 25 % and 30 % fVMVe. (Group 2)
Creep Strain-Elastic
gj
fl
GFRP-2
9.5 mm
7
1
a)
GFRP-3
12.7 mm
(Set b ) b
(jh Kr-4
b
VL
1
2
3
4
5
1
2
3
4
1
2
3
4
1
2
3
1
2
3809
5674
5214
6 o21
5255
6946
5377
9006
5600
3790
3922
4040
4542
4144
3835
4005
4328
4266
209
31.1
28.6
33.0
28.8
37.6
29.1
48.7
30.3
24.3
25.1
25.9
29.1
26.5
24.6
25.6
20.3
20.0
3680
172
219
35.6
32.7
37.8
33.0
43.7
33.9
56.7
35.3
29.7
30.8
31.7
35.6
32.5
30.1
31.4
26.1
25.8
1000
3000
10000
1000
3000
10000
hrs
hrs
48
-132
39
-286
176
758
63
37
230
46
343
127
263
-120
58
61
181
156
hrs
46
-163
29 .
-288
199
819
108
220
308
201
472
255
227
-83
61
N/A
166
201
hrs
^02
-2.9
1.4
-4.5
2.2
10.1
1.5
-0.2
3.7
1.5
1.5
2.3
6.0
0.1
2.7
1.1
-0.2
2.6
hrs
L3
-2.3
0.7
-4.8
3.3
10.9
1.2
0.4
4.1
1.2
8.7
3.1
5.8
-2.9
1.5
1.5
4.2
3.7
hrs
1.2
-2.9
0.6
-4.8
3.8
11.8
2.0
2.4
5.5
5.3
12.0
6.3
5.0
-2.0
1.6
N/A
3.8
4.7
57
61
-166
74
-269
114
702
78
-22
206
56
57
93
271
3
102
45
-9
112
222
~19
m m
"-5
L5
4
4026
18.8
24.3
59
112
77
1.5
2.8
5
4091
19.1
24.7
189
199
231
4.6
4.9
1
6613
37.3
45.7
49
101
210
0.7
1.5
GFRP-5
2
5475
30.9
37.8
257
246
470
4.7
4.5
15.9 mm
3
5162
29.1
35.7
294
321
339
5.7
6.2
4
5214
29.4
36.0
157
258
321
3.0
4.9
1
6946
37.6
43.7
702
758
819
10.1
10.9
GFRP-6
2
5377
29.1
33.9
78
63
108
1.5
1.2
15.9 mm
3
9006
48.7
56.7
-22
37
220
-0.2
0.4
4
5600
30.3
35.3
206
230
308
3.7
4.1
a
For this sample set, resin was used instead of cement grout to adhere the GFRP bars within the
steel tubes (grips).
Sample-sets enduring 25 % fu,ave', other sets endure 30 % fu
98
axe.
L7
1.9
5.6
3.2
8.6
6.6
6.2
11.8
2.0
2.4
5.5
4.4.2
After the creeep test duration (10000 hours), all bars were tested in tension to obtain residual
properties. For all samples, the rupture mode was burring of the fibres; preceded by and helical
wrap unwinding in the case of GFRP-2 and GFRP-6 bars. The average residual strength for all
bars was negligibly affected by the lengthy period under constant load; the loss percentage
ranged from 0 to 5.4 % / u ave. Typically the loss for all data sets, if any, was less than the standard
deviation yielded by mechanical property testing (Figure 4.8; Table 4.6 and Table 4.7).
Similarly, the average loss of the modulus of elasticity, for each dataset, was as low as 0 to 8 %
E/.ave The maximum loss was that of GFRP-4 (12 mm) at 15 %f u>ave . However, that percentage
loss of 8 % is still less than the standard deviation obtained from static tensile tests. The level of
sustained load had no impact on the residual tensile properties for all bars types.
110
100
1
90
80
70
-f
rt-i
Sa 60
H
50
u
j=
40
| 30
OT 20
10
GFRP-1
GFRP-2
GFRP-3
GFRP-4
GFRP-5
GFRP-6
9.5nun
9 5nun
12."mm
12nun
15.9mm
15 9 m m
110
100
90
2
r#-|
rl
r*i
80
70
60
50
40
30
20
10
0
GFRP-1
GFRP-2
GFRP-3
GFRP-4
GFRP-5
GFRP-6
9 5mm
9 5mm
12.7mm
12mm
15 9 m m
15.9mm
Bar type (Group B: 25-30%f
Figure 4.8: Residual tensile strength for samples that exhibited 15 % fU:ave (above); Residual
tensile strength for samples that exhibited 25 and/or 30 % fu,ave (below)
99
SP
<
D
O
(-1

00
Tt
Tt
oo
in
oo
ON
R -
CO c o
Tt Tt
CN
Tt
co
Tt
O
Tt
Tt
Tt
K
Tt
i> . o o Tt
ON
Tt
in
CN
oo
5>3
00
Tt
NO
IN
o
Tt
NO
CN
i> t^
m
NO
NO
NO
Tt
O
NO
00
NO
CN
CO
CN
00
CO
00
NO
Tt
Tt
Tt
CN
CO
CN
oo
r- Tt
NO
NO
in
ON
CN
no
CN
i-J
co
oo
Tt
in
NO
Tt
CN
CO
in
CN CN
5
O
o
t/3
in
Tt
in
CN
(U
+-*
o o o
r-H CN
JD
'co
C

ON
ON
ON
o o o
ON
Dh
Tt
r--
o o
NO
ON
ON
o
o
<D
l-H
ta<u
-e
m
DH
l-l
Dh
<D
<
'co
(2
r-;
CN
CN
in
NO
NO
oo
oo
r^
m
oo
Tt
Tt
CO
in
oo
Tt
in
Tt
oo
Tt
NO
00
Tt
CN
Tt
CN c o
Tt
in
NO
a ^
oo
CN
o r-
ON
OO
r--
0 0
NO
-H
O
oo
CO
Tt
CO
in ON
m ON oo
ON
CN
NO
CO
CN
CO
Tt
CO r- NO NO
T t co N
O oo
CO CO

at
3
'w
<L>
&
T3
g M ,
CN CN
NO
Tt
1)
in
CN
CN
NO
NO
oo
Tt
NO
CN
*-I
CN
CO
Tt
3 35
CO
c
<D
3

.w
0>
S1
t?
^ a
00
I-I
<L>
>
ao
<D
<
Hh
O
0)
00
2
<D
>
<
PH CN
o
-
U
00
BJ
Uh
<
>U
<
<D
00
cd
<
u
o
sh
(D
Ph
in
ON
OO
in
o
IIII
oo
on
in oo o CO
< C N
CO
o o o C
N
oo
on
C3/
1 .a
CN
oo
o
o
cn on
i n
a
-rt
^
in co -< r-<
c/3 -in
ri o5 O t f t
^
C
3/
<D
a
Im
NO
f
no
cn
o
o
CN
^ Tl"
cn
cn
cn
p CN in CN
ON OO NO o
ro ro co ^
^ oo o
00 00 CN NO
ro co co
(N (S rH
r- o
r-J
o
5
C
3/
<D
j
-h
o
h
ON
o
cn
r-h
CN
ON
o
j d
in
3C3/
<
SD
-I
t;
u
Oh
O
s-i
'c/3
C
I
<L>
ca>
a
3
CD
Oh
in
*a
NO
ON
ro
<N oi
o, o
in in ro N O
ON ON ON ON
ta

O'1

>
g-3
o
C
D
00
cs-d<
<
>D
<
NO g
ph S
S
GinON
O
D
<
<u
00
cd
<D
ii
on

î- ^r n- ^
co
3
t3
o
'5cd3
i n
no
CO
in
in
oi
ON in oo o n
in "o ni c o t >^
TJ^t lo Tf
no
cn
00
no
ir! oo oo
m
rf
Tt
in
co
oN
CM
&
00
oo NO
t> co
in
CO
CO
CO
CO
CN
Tf
t-i
t t
no
rj-
CO
CO
on
co
cn
i n
CN
cn
in -^r in
cn
co
oo
co
r-h
NO ^ ON p
K t^ n o o
^ m
<"3
cn
cn
cn
co
i-
cN
"S,
co
CN t-; 00 Tf
o ^t > oo
T"< o o
CN t-H
CN t^ On
On
CN
m
t-3
cn
t t
NO
cn
no
*->
cn
CN
On
in
CN
on
no
On
on
on
on
on
00
co
00
co
rt
rt
r- oo o
no
co
no
co
no
co
on
co
co in -sf
in in n o
m
cn
no
on
cn
cn
^t
00
CN
ON
r-;
cn
cn
co
lo
'

i-
(N CN in
NO 11 <
I S 0 00
00 oo
co
Pi
cu
ft
<D
in CO CN y-<
xt n o in O n
no
t
00 00 t^ t
r^ co
NO ON
oo
cn
On
OO NO
co n
cn
cn
cn
OO
is
oo

cl-hd
s - s
s - 2
o
&
00
(u
>
On
<
CN a
a
Hh
^
O C3N
0)
00
cd
s-i

<
r?
4)
00
cd
co
S-H
<d
>
fe
c-
ci -di
"5
<d
<
' : 00
< O CJ
cu
00
>
& I
FE
CN
^j-
4>
00
a
34>
okl
4>
On
oo O
ON
oo oo in
co
co
c^
cn
Tt
CN
PH
O
m Oah
<3 O
W co
in m (N
NO CO CN
rI
Tt
T3 3
I -a
IjO o
4>
Pi
co
4>
3h
ico
CO
Tt"
ON
CO
CO
t>
t--
cn
no
CO
Tj-
CO
NO
cn
<-h
Tt
00
Tt- Tt Ti-
r-N
(N
&
O
S-i
co
in co
ro OO NO00
in
cn
i n
r^
Tt
CO CO
CO
a
o
co
ON
O
C3/
^4>
Jh
' CO
aN c
4>
3
4l-H>
oo h- Isco
i oo
t*-
NO OO CO
CO
I
ON
o
NO
o4>
CO O N
Tt
Tt
ON ON
ON
ON
CN
Tt
<
T
t
CO
Oh
00
>>
tS
4>
o.
o
00
C
i>
Oh
'm

^
CN
00
ON
Tt
oo
r
Tt
ON
<l>
-a
4>
fl fl s
3 C3 0
a
o
U
oTt
NO
CN
in CN CO
CO
2 r- r- r3 w
r2
Sw
<
co
3
co
CO ON
l>
co
co
on
cn)
cn
co
CN
ON
OO
--H
r-H
CN
CO
NO
CN
CO
CN
CO
r-0\
CN
wa 3
e4>
9 I1
O1
1)
Oh
00
4>
a
co
0)
m g
V5 <D
dj a
a-a
E
o
3 >
a-2
o
4>
oo
a
ii
4)
^ a
^s C*
a
[in
S.!2
4>
00
ci d4>
>
<
4.4.3
The formation of microcracks in the resin and the debonding at the interface of fibres/matrix
are the most common phenomena occurring in a GFRP material under sustained load and/or
adverse environment. In this respect, investigation (Scan Electron Microscopy - SEM analysis) took place using selected samples of the tested bars after the 10000 hour test
duration to have a better understanding of the causes behind strength loss, if any.
Micrographs, in Figure 4.9 show a magnified cross section of the samples prior to testing.
Figure 4.10 shows magnified cross-section images of the rupture zone for all tested bar types.
Samples of GFRP-1, GFRP-2, GFRP-3, GFRP-4, GFRP-5 and GFRP-6 were tested after the
aforementioned duration after exhibiting 30 %, 30 %, 25 %, 25 %, 30 % and 30 % fu,ave,
respectively. Results show that all bar types show no signs of debonding between the fibres
and vinylester resin, and no induced microcracks. Subsequent SEM tests have also indicated
that the external surface of all bars remained unchanged.
Figure 4.9: Enlarged samples' cross section before applying sustained load
104
Figure 4.10: Enlarged samples' cross section after exhibiting 10000 hours of loading under
different sustained load levels
4.5
A comparative study was conducted regarding the creep behaviour of fifty two GFRP barsamples, of six different commercial brands, over a period of 10000 hours (417 days). The
glass/vinylester GFRP bars (of diameters 9.5 mm, 12 mm, 12.7 mm and 15.9 mm), were
tested at room temperature (23 3 C) and subjected to constant sustained load service levels,
nominally 15 % and 30 % of the average ultimate tensile strength / uave . The following
conclusions can be drawn from the longterm tests, residual tensile tests and microstructural
analysis at the end of the test duration:
1 - The creep strain evolution, for the tested six commercial types, typically starts with a
high rate at the first 80 to 120 hours then tapers down asymptotically with time. For
bars that exhibited 15 % flL ave the upper-bound (maximum) values for accumulated
creep strain after 10000 hours were 8.8 %, 2.5 %, 4.1 %, 3.2 %, 8.3 % and 15.7 % of
the initial strain frp,o, for bar-types GFRP-1, GFRP-2, GFRP-3, GFRP-4, GFRP-5 and
GFRP-6, respectively. Similarly, the upper-bound values for 25-30 %/,',,<,? were 3.8 %,
11.8 %, 12.0 %, 5.6 %, 8.6 % and 11.8 % for the same bar-types, respectively.
105
2- The upper-bound values in certain cases are much higher than the average accumulated
strain calculated for the replicates of each test-set. Nevertheless, these high upperbound values are low and inconsequential compared to the creep coefficient of
concrete itself that varies between 150 % and 400 %.
3- There is no evident/clear relationship between the applied load level and resulting
creep strain after 10000 hours. Moreover, due to this irregular nature of the data, an
isochronous stress-strain relationship between the magnitude of sustained stress and
accumulated strain cannot be derived.
4- In general, bars of bigger diameters exhibit greater creep strain values; possibly due to
the curing factor which is better for bars of small diameter.
5- The residual tensile strength and modulus of elasticity, for all samples that survived the
10000 hour duration, were found barely changed (almost retaining their full strength).
The loss percentage ranged from 0-5.4 % fUMVe and 0-8 % E/ave for tensile strength and
Young's modulus respectively. In both cases, the loss was less than the standard
deviation yielded by mechanical property testing.
6- Microstructural analysis was conducted on samples from both commercial bars that
exhibited 25 % and 30 % fu>ave for 10000 hours. No microcracks were found indicating
that there is no sign of degradation for GFRP bars exhibiting service load.
7- This study serves as a baseline for similar tests to be conducted on the same
commercial GFRP bars under the combined effect of stress and sustained load. The
impact of environmental conditioning should be considered for GFRP reinforcement
exhibiting outdoor exposure.
106
Chapter 5: Longterm Performance of Third-Point Loaded FRP. RC Beams after One Year of Continuous Loading
CHAPTER 5
LONG TERM PERFORMANCE OF THIRD-POINT LOADED FRPREINFORCED CONCRETE BEAMS AFTER ONE YEAR OF
CONTINUOUS LOADING
5.1
Introduction
In the past two decades, an abundance of research has taken place on fibre reinforced polymer
reinforced concrete (FRP-RC). Due to that, a better understanding is now available on
paramount characteristics such as strength, stiffness, bending, shear and FRP-concrete bond.
Where the behaviour of steel reinforced concrete beams under sustained loads has been
studied for nearly half a century, the area of long-term/creep behaviour of FRP reinforced
concrete beams remains barely touched; mainly due to the time consuming nature of creep
experiments. This sought knowledge - of long-term serviceability aspects - is extremely
important for the infrastructure facilities constructed using FRP.
It is commonly known that concrete creep, shrinkage and other factors contribute to the
increase of strain, curvature, and deflection over time. For conventional reinforced concrete,
steel reinforcement acts to restrain these effects. On the other hand, FRP reinforcement with
its lower modulus of elasticity and higher susceptibility to creep may be a concern for its
comprising concrete members, as regards serviceability. Whether steel reinforcement or FRP
reinforcement, the concrete cross section exhibits strain in concrete within the compression
zone due to creep. This, in turn, leads to an increase in neutral axis depth and redistribution of
stresses between the concrete and reinforcement (ACI 435, 2003). Associated with the latter
time-dependent behaviour, is an increase in curvature. This curvature-increase, along with the
formation of additional flexural cracks over time, leads to an increase in beam deflection.
The total deformation at any time t is the sum of the immediate and time-dependent
deformations. As regards the calculation of initial deflection, two main approaches are
available: the effective moment of inertia approach and the mean curvature approach. The first
approach - the effective moment of inertia I e , approach is adopted by ACI 440.1R-06 wherein
107.
the proposed equation interpolates between the moment of inertia of the gross uncracked
concrete section, I g , and the moment of inertia of a transformed cracked section, I cr , to account
for the tension stiffening effect of the concrete in tension.
/
1+
cr
(5.1)
\MaJ
where Mcr and Ma are the moment just sufficient to produce cracking and the maximum
applied moment, respectively, [id is calculated as a function of the relative reinforcement ratio:
/
Pf_ <1.0
(5.2)
Pfl>
ISIS Canada Design Manual (2007) recommends the following equation to be used in design:
7 =L.+
(5.3)
1-0.5
M,
Mota et al. (2006) examined a number of the suggested formulations for I e and found that the
equation above yielded the most consistent and conservative results over the entire range of
made-available specimens. This equation was reported to work well with different types of
FRP reinforcement.
The second approach - the curvature approach - is adopted by CAN/CSA S806-02. wherein
the integration of curvature along the span to determine curvature is recommended. On
knowing the value of the curvature y/, the virtual work method can be used to calculate the
deflection of FRP-RC beams under any load level.
A, = jmy/dx
(5.4)

third (See Figure 5.1). It is worth noting that this method assumes no tension stiffening.
108.
Figure 5.1: Moment-curvature (M-k) relation for FRP reinforced concrete (CSA S806-02)
On the other hand, the time dependent behaviour of reinforced concrete is intrinsically
complex due to the many influencing factors. Some of these factors are concrete compressive
strength, upper/compression reinforcement, ultimate creep coefficient, humidity, time of
loading and duration of loading. There are two main approaches for computing long-term
deflections (due to creep and shrinkage). The first, as per ACI 318-08 and CSA A.23.3-04 for
steel reinforced concrete, involves multiplying the initial deflection AT by a multiplier X. The
time-dependent beam deflection multiplier, X, is a function of two effects: increasing curvature
at cross-sections along the span due to creep, and the effects of additional cracking along the
span (Gross et al 2003).
a ( A , ) sustained
(creep+sh rinkage)
X
1 + 50 mp'
(5.5)
(5.6)
where the factor ^ accounts for the time-dependent concrete behaviour and the denominator
accounts for the restraint against creep provided by compression reinforcement; p is the
compression reinforcement ratio and m is the modular ratio of the reinforcement (m =
EFRP
Esteei-for FRP reinforced concrete). The time-dependent factor, is given the values of 1, 1.2,
1.4 and 2.0 for three, six, twelve and sixty months, respectively. ACI 440.1R-06 modified Eqn
109.
5.5, for both GFRP and CFRP reinforced concrete beams, in light of earlier studies conducted
by Kage et al. (1995), Brown (1997), Vijay and GangaRao (1998), Arockiasamy et al. (1998)
and Gross et al. (2003); a modification factor of 0.6 was attributed to concrete beams with no
compression reinforcement or with FRP compression reinforcement, keeping the same c
values for the aforementioned time durations. CAN/CSA S806-02, however, adopts the same
model as in Eqn 5.5 without modification.
{creep+shrinkage)
= 0.6^(A,)sustamed
(5.7)
The second approach, for computing long-term deflection, is based on the age-adjusted
effective modulus method in which the properties of the section change with time and
calculation at each time step relies on its predecessor. The foundation of this approach was
first introduced by Bazant (1972) and later presented using a system of equations found in the
textbook of Ghali, Favre and Elbadry (2002).
The former approach is more common and user-friendly. Gross et al. (2006) indicated that it is
rational to use simple empirical methods, as of the multiplier approach, for the prediction of
time-dependent deflections and to expect a degree of variability in the results; the long-term
behaviour is both complex and highly unpredictable. The latter method (the age-adjusted
effective modulus method), however, is more sophisticated and requires numerical (finitedifference) modelling; it is encouraged as an application for researchers to gain insight into
the mechanism underlying the long-term deflection phenomenon. Further details of the
aforementioned theoretical calculation of initial deflection and long-term deflection models
are discussed in detail in the results and discussion section below, along with their application
using the obtained experimental data.
Since the advent of FRP reinforced concrete beams, the number of studies regarding the long
term performance/deflection of FRP reinforced concrete beams is yet not abundant (Brown
and Bartholomew (1996), Brown (1997), Vijay and GangaRao (1998), Arockiasamy et al.
(2000), Hall and Ghali (2000), Gross et al. (2003) and Gross et al. (2006)). A thorough
literature review was conducted to find and summarize all projects dealing with this research
area. It was found that the experimental work, to date, amounts to a dozen studies at most;
some of which, are summarized and presented in the paragraphs below. It was found that most
110.
of these studies adopt a four-point loading setup where the long-term behaviour of FRP
reinforced beams (designed for compression failure) are monitored in terms of deflection,
strain variation and crack width, for time durations varying from 3 months to two years. In the
majority of the previous studies, FRP reinforced beams had no top reinforcement.
Midspan deflection and crack width results of six high-strength-concrete beams (two GFRP
reinforced, two CFRP reinforced and two steel reinforced) were presented by Gross et al.
(2006). The beams of dimensions (114 mm x 184 mm x 1880 mm) were maintained under
four-point bending sustained load for a period of three months (ninety days). The testing/clear
span was 1828 mm and the constant moment region was 102 mm. The reinforcement used
was 2-15.9 mm GFRP bars, 2-9.5 mm CFRP bars and 2-15.9 mm steel bars, respectively,
yielding similar moment capacity for the three sets of beams. All beams had no shear
reinforcement. The applied sustained loads amounted to 30 % of the beams' nominal capacity
which approximately corresponds to 2 times the cracking moment ( M J M c r = 2). When
compared to earlier efforts (Brown (1996); Vijay and Gangarao (1998); Arockiasamy et al.
(1998)), the obtained long-term deflection results were relatively lower due to the use of high
strength concrete. The average deflection-increase was 33 %, 16 % and 24 % from the initial
induced deflection, for GFRP and CFRP and steel reinforced beams, respectively. The former
two values - for GFRP and CFRP - are far less than the predicted 60 % increase and 100 %
increase indicated in ACI 440.1R-06 and CAN/CSA S806-02. Measured maximum crack
widths - at test inception - for GFRP and CFRP reinforced beams were significantly greater
than steel reinforced beams. The ratio of time-dependent to initial crack Width, however,
showed relative proximity for all three sets; the average crack width increase was found to be
13 %, 15 % and 20 % for GFRP, CFRP and steel reinforced beams, respectively.
Time dependent strains, deflections and cracking were monitored for three normal and three
high strength concrete beams; reinforced with 2-12.7 mm and 3-12.7 GFRP beams,
respectively (Gross et al. 2003). The beams were subjected to four-point sustained bending for
a period of 6 months (180 days). The beams contained no shear reinforcement and had a (121
mm x 235 mm) cross section; the test-span was 3050 mm and the two loading points were
located 252 mm apart at midspan. Three sustained load levels were applied; specimens were
111.
loaded to approximate top fibre concrete compressive stresses of 0.40 f'c, 0.45 f'c and 0.50
f'c at an age of 28 days. The overall behaviour of the GFRP reinforced beams was similar to
steel reinforced beams; where a downward shift of the neutral axis would occur over time
corresponding to redistribution of internal stresses. The obtained results rendered six values of
time-dependent deflection multipliers; such multipliers are defined as the time-dependent
deflection (total minus initial) divided by the initial deflection. The multipliers associated with
normal concrete beams were far greater than their high strength concrete counterparts. This
was mainly due to the higher creep coefficient of normal concrete. Furthermore, formation of
additional flexural cracks, with time, had a significant impact on the time-dependent
deflection multiplier.
The experimental immediate and long-term deflections of four shallow concrete beams, two
reinforced'with 3-15 mm GFRP bars and two reinforced with 3-15.9 mm steel bars, of crosssection (280 mm x 180 mm) and 3500 mm in length were presented by Hall and Ghali (2000).
The stirrup-less beams were simply supported and exhibited third-point loading for a period of
approximately 8 months. The two equal-in-magnitude applied concentrated loads were 1067
mm apart centreed in a 3200 mm loading span. Two levels of sustained loading were
considered (maximum applied moment to cracking moment
MAIMCR
= 1.5 and 3.0). In the
analysis, the authors used equations found in Ghali, Favre and Elbadry (2002) to calculate the
long term deflections. These equations are based on the first principles of equilibrium and
compatibility. The creep coefficients and free shrinkage values used in these equations were
calculated using equations from CEB-FIB Model Code 1990 and/or ACI Committee 209
recommendations
(1992). Immediate and long-term deflection values obtained from
experimental work were compared with theoretical values. It was found that CEB-FIB Model
code 1990 accurately predicts both the immediate and long-term deflections for both steel and
GFRP reinforced shallow beams. The ACI Committee 209, however, overestimated the
deflections of the tested steel and GFRP beams.
In 1998, Vijay and GangaRao presented the long-term behaviour results for four GFRPreinforced concrete beams that were monitored for an average of 360 days (12 months
approximately). The beams were of dimensions (150 x 300 x 3050 mm) exhibiting third-point
loading where the testing span (distance between supports) was 2670 mm approximately (i.e.
112.
distance between two concentrated loads = 890 mm). For the first pair of beams, the concrete
compressive strength, / c ', was 24 MPa; 2-12.7 mm sand coated bars were used as bottom
reinforcement; the maximum applied moment Ma, at midspan, was 35 % and 50 % of the
nominal moment capacity of the beams Mn; the beams were designed for tension failure. For
the second pair of beams, the concrete compressive strength, / c ', was 28 MPa; 2-15.9 mm
ribbed bars were used as bottom reinforcement; the maximum applied moment Ma, at
midspan, was 20 % and 35 % of the nominal moment capacity of the beams Mn; the beams
were designed for compression failure. For all beams, 9.5 mm sand coated bars were used for
compression and shear reinforcement. The main findings of the study were that the creep
coefficient is less in GFRP concrete beams as compared to the available data on steel
reinforced beams; the relative long-term deflection of GFRP reinforced concrete beams is less
than of steel bars; crack width increase is proportional to creep coefficients and can be
estimated with suitable reduction factors on the latter.
5.2
Research Purpose
This study provides essential data on the behaviour of FRP reinforced concrete beams, under
constant sustained load (strain increase, deflection and crack width). The main objective is to
set a longterm-performance study for a variety of concrete beams reinforced with different
types and sizes/ratios of FRP (GFRP and CFRP) bars as well as identical beams reinforced
with steel bars. This implies: (i) observing the creep strain evolution, with time, for the FRP
bottom reinforcement as well as compression concrete fibre; (ii) observing the initial
deflections and comparing them to the theoretical predictions of the effective moment of
inertia method (Branson's equation) and the curvatures' method (available in CI 440.1R-06
and CAN/CSA S806-02); (iii) observing the long term deflection values and comparing them
with the predictions of North American standards (CAN/CSA S806-02 and ACI 440.1R-06)
and (iv) calculating the anticipated maximum crack width and modelling its evolution with
time.
For this phase, the main parameters are the type of reinforcement and reinforcement ratio
within the beams. The sustained load level, which is calculated as a percentage of the nominal
113.
moment capacity of the beams, is kept constant for all beam samples. Furthermore, the
increase in crack width over time as a result of sustained load was studied.
5.3
5.3.1
Materials
The materials used in this study are ordinary concrete, a variety of FRP (six types of GFRP
bars and two types of CFRP bars) and 15M steel. Two concrete batches were used for the 20
concrete elements. For both batches, the concrete used was MTQ Type-V with a target
compressive strength of 35 MPa after 28 days. The mixture proportion per a cubic meter of
concrete was as follows: coarse aggregate content of 646 kg with a size ranged between 10
and 20 mm, 341 kg with a size ranged between 2.5 and 10 mm and fine aggregate content of
717 kg, cement content of 455 kg, water-cement ratio (w/c) of 0.35, air entrained of 5.0-8.0 %,
and water-reducing agent. The slump of the fresh concrete was measured before casting and
was about 100 mm (4.0 in.). Six concrete cylinders (100 x 200 mm) were cast for each batch
and tested the day the corresponding beams were installed under sustained load. Beam testing
took place after 3 to 7 months from the date of casting, to minimize the effect of shrinkage on
the test results. The average concrete compressive strength f'c, for two batches, turned out as
35 and 40 MPa and the corresponding measured modulus of elasticity, Ec, was 28.5 GPa and
31 GPa, respectively.
Six commercial GFRP bars and two commercial CFRP bars, pertaining to four different sizes
and three different manufacturers, are examined in this study (Figure 5.2). For manufacturer
A, GFRP and CFRP bars are made of high-strength E-glass fibres or carbon fibres
impregnated in vinylester resin. The bar's circular cross section is manufactured by a process
that couples pultrusion with sand-coating along the external surface of the bar. Similarly, the
GFRP and CFRP bars of manufacturer B comprise of E-glass fibres and carbon fibres,
respectively. The fibres are drawn/pultruded through a vinylester resin bath; surface
undulation (helical wrapping) is applied prior to thermosetting of the polymeric resin. The
GFRP bars of manufacturer B have a sand coated surface along with the helical wrapping. As
for manufacturer C, the core of the rebar is a patented pultrusion process. In this continual
process high-strength glass fibres are drawn through a tool where they are impregnated with
114.
liquid synthetic resin. The impregnated fibres are sent through a profiling mould and are
hardened. The ribs are ground and hardened into the bars. Further detail as to the physical
properties of the bars and their appearance are available in Table 5.1 and Figure 5.2,
respectively. The main mechanical properties, as obtained through tensile tests conducted
prior to casting the beams (average ultimate tensile strength fu,ave, modulus of elasticity Ef and
average ultimate tensile strain su,ave), are displayed in Table 5.2. The mechanical properties of
steel is 15 M steel (diameter = 15.9 mm) of yield stress fy = 421.3 MPa and modulus of
elasticity Ef = 200 GPa.
Figure 5.2. GFRP bars (left) and CFRP bars (right) examined in the current study
115.
i n
(n
<
<
oo
CN
<
co
<
<
T3
<d
>
i n
NO
PQ
oo
in
ro
CN
oo
o
ON
S-H
r -
o
o
in
<D
OH
in
PQ
m
CO
1
00
k
m
in
CN
in
ON
00
in
On
Id
&
T3
fa
u
T3
NO
00
NO
ooh
on
in
NO
ON
(n
CN
ffi1
CN
in
ON
fa
o
t )
CN
S3
CN
in
CN
oo
i n
r -
in
CN
CN
>r,
o
r -
ON
CN
(jh
o
n
CD
NO
NO
CN
CN
r -
OH
in
ON
ON
in
00
ON
OH
&
oo
<
00
U
cd
43
CL,
a
rUn
oo
T5
4)
"-b
<d
o
>>
C/3
C/3
T3
&
"O
S
Cd
CN
CQ
oo
in
NO
in
in
On
00
CO
CO
NO
4)
NO
i d
X
&
ON
cod
On
in
CD
S
T3
00
NO
in
NO
oo
ro
ON
CN
in
oo
ON
o
o
o
o
3
T3
<D
S3
o
S
aCD
fl
c/3
3
Si
M 5Oh nJ
n
"S
C3 o
O
X O
s3
o
o o h
2 ^
o
s3
rt
A
fe v p T3 g
x
:S ^
fa w
S
00 +3
S3 <+H
o o
hJ
>
.2
i-H
o * v3
cl) o h u
o
o
<d
2
c/3 3
x
0> vo
l-i
<d
>
c/3
U
J3 00
-a
fl s3
o 0
o
0)
1
o
<D
%
o
.8
c/3
S
l-i
CD
<4-1
o
o
o
S-I
(D
C
C33//
cd
o
cd cS-ld
S-I
<d
oh

g
oo
in
NO
s-c

CN
-H
co
00
r-
CN
CN
C--
oo
in
co
4>
ON
in
CO
c
'5g3
H
>
o
on
03
CO
4
> CO
&
r--
o
oo
4>
4
>
&
o
CN
o
4>
co
CN
-H
o
Tt
m
-H
C
O
-H
in
ON
m
ON
CO
41
oo
CN
00
Tt
CO
-H
Tt
m
oo
ON
NO
ON
Tt
NO
o
00
CN
ON
ON
C
OM
O
r-CO
Tt
NO
Tt
m
Tt
CO
CO
ro
CN
-H
in
CN
Tt
-H
CN
rNO
-H
oo
NO
CO
CN
NO
-H
in
NO
NO
oo
CN
m
CN
o
NO
CN
Tt
rin
oo
CN
oo
>n
CN
CO
NO
-H
oo
Tt
Tt
oo
NO
OO
41
Tt
O
O
Tt
00
o
00
00
in
rNO
f41
CN
CO
CN
00
CO
ON
m
ON
T1
in
r-
in
Tt
-H
CN
CN
in
NO
in
ON
41
o
C
N
N
inO
CN
ON
Tt
Tt
Tt
NO
Tt
O
NO
CN
NO
NO
00
CO
-H
Tt
r-
CN
co
co
00
CN
a0
1.3
m
CO
-H
oo
Tt
r--
CN
vi
O
in
-H
r-
CN
in
NO
in
Tt
-H
Tt
NO
CN
-H
ON
NO
Tt
CN
CN
00
o
CN
o
NO
Tt
o
o
CN
CN
00
CN
Tt
r--
00
CN
CN
in

I %
4> ^ W
vi Itt:
i
a
w
'ss o CO
5 I
D
8 I7 4>
H<
h
I
M
CO '
1 3 II
a , ' -O Si
D a S X. , HH^ o
+4->
4) C^s
4)
co
5 N
-s
w 3
Js ^ OX) II
3
a
o
4> ^
O
to
fl
'I
co
4)
s
to
=1
4)
c B
CO
1P
VI i O
T3
toa -a
4)
4
-t-t II 4)
4)
to
a
Co
S
ai
si
to +3 to d>S
S X
*
53 O
t2
03 13
3 03
t3 O
to 1ii1 ^
O 4> M 8
TO *
l-H OX)
3 o
aC3 to >
03-1
4)
O
O
Q
5.3.2
Sample Description
Twenty beams, of dimensions 100 x 150 x 1800 mm, were cast using the aforementioned FRP
bars as bottom reinforcement. The initial intent was that all beams have the same concrete
compressive strength (f'c = 3 5 MPa). However due to damage whilst transportation, two
samples were recast using a different concrete batch; the obtained concrete compressive
strength, fc, was 40 MPa. All the beams have: (i) two bars of bottom reinforcement, (ii) a
clear concrete cover of 25 mm, (iii) no upper reinforcement and (iv) no shear reinforcement.
Each pair (2 beams) are identical in terms of reinforcement and concrete properties (See
Figure 5.3).
5.3.3
Instrumentation
Prior to concrete pouring, the bottom reinforcing bars were instrumented at the middle/centre
of each bar with 10 mm Kyowa strain gauges of 120 ohm resistance. Beam specimens were
left out to cure for a period of two weeks after which 2 kyowa concrete strain gauges (67 in
length) were installed at the midspan of the compression surface of the beams. Furthermore,
six demec points were installed per side of each beam (12 demecs in total) in the manner
shown in Figure 5.3 below. The demec points serve as fixed points for deflection
measurement using a high precision vernier caliper.
25 mm
P/2
500 mm
P/2
500 mm
_L
. 1 -
Demec points.
500 mm
"""Concrete gauges
FRP gauges
^]50 rnrn^
25 mm
S
om
Demec
point -
I T
^150r
1500 mm
Reinforcing
bars
100 mm
Figure 5.3. Typical beam sample instrumentation (left) and cross section (right)
5.3.4
Sustained Load Frame Calibration and Beam Installation
The schematic in Figure 5.4 shows how a sustained load frame comprises two mirror beams of
identical characteristics. For all frames, the provided testing span is 1500 mm whereas the
constant moment region is 500 mm. Installation of the beams in each frame takes place in the
following chronological order: (i) The two beams are assembled on top of the frame base in
118.
mirror-fashion (i.e. the lower beam is centred on metal cylinders - 500 mm apart - with its
bottom reinforcement upwards and vice versa for the upper beam); the beams are kept apart
by metal cylinders spaced - 1500 mm apart - serving as beam supports; (ii) the beams are
closed upon by a steel piece applying concentrated load at two points 500 mm apart; (iii) an
upper lever arm is then installed and attached to a threaded rod; the lower portion of the
threaded rod is mechanically attached to a lower lever arm on which dead weight may rest
(See Figure 5.4). To maintain constant sustained load magnitude on the tested beams, it is
mandatory to keep both the upper lever arm and the lower lever arm - with applied weight perfectly horizontal. This is made possible by a trade-off of tightening or loosening both the
upper-arm screw and threaded-rod nut shown in Figure 5.4.
Threaded rod
nut
Upper arm s c r e w .
Vertical sleeve
Steel piece maintaining

, two point load
Upper lever arm

Threaded rod
Steel cylinders
Upper beam (B2)
(100 x 150 x 1800 mm)
Lower beam ( B l )
(100 x 150 x 1800 mm)

'fr
nr
Lower lever arm M

Lower arm weight
Frame base
Laboratory floor
P/2
P/2
Upper beam (B2)
T
P/2
P/2
P/2
P/2
-1.
Lower beam (Bl)
I
P/2
P/2
Figure 5.4. Schematic of sustained load frame
119.
The two mirror beams of each frame are anticipated to yield close results (self weight is of
negligible effect). For all beams, once loading took place cracks were observed. The obtained
measurements in the subsequent section show very good agreement for the two beams of each
frame. Prior to beam installation, thorough calibration was done for all frames available using
a load cell, a portable strain indicator, and dead weights of different magnitudes (See Figure
5.5). The purpose of this calibration process is to calculate the necessary weight and lever arm
length for a particular sought load to be applied on the beams. Furthermore, a frame
verification test was conducted over the longterm course of the study where two calibrated
frames were loaded with two pairs of beams of identical properties. Likewise, very good
agreement was observed, indicating reproducibility of the frames and the soundness of the
calibration.
Figure 5.5. Calibration of sustained load frame
Having calculated the nominal moment capacity for each beam Mn, the sustained load (Msus)
to be applied was taken as 25 % of Mn (approximately twice the cracking moment Mcr) (See
Table 5.3). In this respect, ten pairs of beams - each pair of identical properties - were
installed in ten sustained load frames (See Figure 5.6) and monitored regularly for a twelve
month (one year) duration. The twenty beams can be categorized into two groups; Group 1
has relatively lower reinforcement than Group 2 leading to two ranges of nominal moment
capacity Mn of (7.55 to 7.98 kN.m) and (10.01 to 11.13 kN.m), respectively (See Table 5.3).
The continuous loading of the beams took place in the laboratory at ambient temperature of 23
120.
1C and relative humidity of 50 2 %. The casting and loading dates of the beams are
available in Table 5.3.
Figure 5.6. Beams in comprising sustained load frames
121.
O
fa fai
in
^h CN
PQ PQ
ON
ON
ON
faI faI
ON
in
T-<
CN
ON
CQ CQ
00 00
fa
oo fa
I
I
T-H (N)
ro
ro
O
in
ro
ON
ro
O
N
in
co
(N
ON
fa fa
ON
i
^H (N
Tf.
in
ro
CQ CQ
ON
fa
N fa
ri CN
rl-
in
in in
fa fa
i
n-
CN
CN
in
ro
N
00
in
ro
00
ro
DQ PQ
co co
fa fa
i
'
CN
fa
o
r-CN
in
CN
ON
o
oo
o
CN
CN
i
r-H
(N
CQ CQ
CN
in
m
ro
CN
ON
in
CN
ON
<D
<
&u<
C/3
0)
o op
i-i cd
S 8
at, 
T3

N
in
CN
CN
ON
CN
in
ro
CN
a
<D
ON
CN
CN
ON
fa fa
00
ON
ON
CN
FFL CQ
=tt
NO
CN
CQ CQ
CN
in
ro
ON
ro
CN
PQ CQ
fa fa
i
(N
Tj"
ON
ON
ON
NO
CQ CQ
ON
CN
r- o-
in
CN
CQ PQ
\ O
ON
in
ro
CN
ro
oo
ro
in
CO
<3.
CO"
ON
co
cdI
&<<
O
<=>
CO"
O-
%
<D
<L>
CO*
<u
CO
Co
-a, u
Co 00
cd
a
1;
I c1-ou 5i u
- <U AP. uS-l
. a
00
CO
cd
CO
CO
ft! P H
I
I
^ C
N
TT
in
ON
PH
ON
P I
OO
NO
ON
I-
00
CN
CQ CQ
ON
CO
OO
r--
CO
CO
-h (Ni
ON
ON
C
N
CQ CQ
CN
CO
CN
NO
CN
CN
oo oo
00 ft, ft,i
CO
CO
-h C
N
TT
NO
CQ CQ
ft
CO
CO
PH
o
o
CN
CN
i
CN
CQ CQ
O
o
r-
00
00
in
CN
in
NO
CN
CN
&
o1a

>
CO
NO
"-C3
C
N
O
D
<
NO
PH
NO
T3
s
in m
m
oo
ON
ON
ft,
r^l (N
PH
o
o
oo
oo
in
<
U
s
ON
TJ
CN
s
'B
m
3
C
3
/
M
CN
NO
CN
Tt Tt
ft,
i
-h
CN
Tt
PH
cm
in
Tt
in
CN
CQ CQ
ON
r-
r
-
t--
in
a
CO
<L)
oo
CN
co
C
O CO
ft, P H
i
CN
Tt
oo
On
CN
Tt
o
CN
5 ffl
O
o
03
CO
CO
CQ CQ
3
3
i i
i CN
o
C
N
<
D
lo
co
o
T3
1>
CO
1)
H
T3
<
L
>
O
CN CN
PH
PH
CN i i
-h (N
CO
CO
in
in
ON
CN
00
Tt
o
SO
CQ CQ
O
N
t>
O
CN
'c|H
s
CO
o
00 ' c<
aL
>
fl
ft, ft.
. i
V
M CN
CO
CO
ON
CN
TT
00
in
CQ CQ
o
00
op
'm
u
T3
W
ON
O
1O
=tt

CO
CN
5.3.5
Design of FRP and Steel Reinforced Concrete Members
In Table 5.3, each beam is noted according to its location and comprising frame (for example
B2-F5 represents the upper beam in frame number 5). The beams were designed according to
the ISIS Canada Design Manual (2007) as follows:
The balanced reinforcement ratio, pfrPb, is calculated as:

Pfrpb =
//
+
h ffrpu V cu frpu)
(5.8)
where f'c is the concrete compressive strength (MPa); the ratio of average concrete strength in
rectangular compression block to the specified concrete strength a\ = 0.85 - 0.0015 f'c > 0.67;
the ratio of depth of rectangular compression block to the depth of the neutral axis fi\ = 0.97 0.0025 f'c
> 0.67; (pc and ty are the material resistance factors for concrete and FRP,
respectively; ffrpu is the ultimate tensile strength of FRP (MPa); scu and frpu are the ultimate
strain in concrete in compression and the ultimate strain in FRP in tension, respectively.
Reinforced concrete beams comprising a reinforcement ratio, pjrP, greater than the calculated
value, pjrpb, are designed to fail due to concrete compression (known as over reinforced
beams). Whereas, reinforced concrete beams are considered under-reinforced (designed to
fail due to rupture of reinforcement) when the actual reinforcement ratio, pfrp, is less than the
calculated value, pfrpb. All beams, but the pair reinforced with steel, are over-reinforced; the
actual to balanced reinforcement ratio is indicated in Table 5.3.
The beams' cracking moment and nominal moment capacity (for over-reinforced concrete
beams), Mcr and Mn, respectively are calculated using the two equations below:
(5.9)
Mcr=0.6jfy,
M = P/rpf/rp
v
2
1 - 0 . 5 9 pfrpf f r /P , f' bd
(5.10)
J c
where It is the moment of inertia of the transformed uncracked section; y, is the distance from
the centroid of the uncracked section to the extreme surface in tension; b and d are the cross
124.
section's width and depth, respectively. The axial stress in reinforcement at failure, ffrp, is
calculated using Eqn 5.11:
(5.11)
where Efrp is the modulus of elasticity of FRP, scu is the ultimate strain in concrete and /?/ is
the stress-block factor for concrete.
5.3.6
For any sustained load frame, once a pair of beams was installed and both lever arms were
made horizontal, initial readings were taken. Strain measurements were taken by attaching the
strain gauge wires to a portable strain indicator (used for both surface and embedded strain
gauges). The average of three individual measurements was recorded. The digital vernier
caliper - of 0.01 mm precision - was used to read the resulting deflection. The external jaws of
the caliper would close upon two demec points to take a single reading; one demec point
would be that installed on the beam whereas the other is fixed onto the base of the frame.
Three demec points are fixed on each side of every frame-base; lying vertically underneath the
demec points installed onto the beams (See Figure 5.7 and Figure 5.8). As for crack width, a
microscope of 0.01 mm precision was used and the width of the biggest crack (at midspan)
was measured. This manual measurement process (for strains, deflection and crack width) was
conducted three times for the first week; once every two weeks for the consecutive month and
once a month until the end of the one year duration. The applied load was maintained constant
by adjusting the upper-arm screw and the threaded-rod nut periodically.
125.
Figure 5.7. Long term deflection measurement
Figure 5.8. Deformometer (P-3500 Portable strain indicator) for strain measurement
5.4
5.4.1

Creep Induced Strain in FRP Reinforcement and Compression Concrete
In this sub-section, Figure 5.9 to Figure 5.18 display the measured strain evolution with time
for bottom reinforcement as well as the creep evolution of compression concrete. As
mentioned earlier, two gauges were installed (one on each reinforcing bar), at midspan, and
likewise on the concrete compression surface. For Figure 5.9 to Figure 5.18, below, the gauge
and the beam are indicated as "G" and "B"; i.e., the nomenclature G2-B1 means "gauge 2 beam 1".
126.
60
120
180
240
300
360
180
240
300
360
4320
5760
7200
8640
420
days
hours
1440
hours
1440
2880
4320
5760
7200
8640
2880
10080
Time (hours/days)
10080
Time (hours/days)
0
60
1 20
180
240
300
360
420
60
Time (hours/days)
120
180
240
300
360
420
Time (hours/days)
evolution of concrete (right).
127.
120
180
240
300
360
420
60
120
180
240
300
360
420
1800
days
16.2%
g1-b1
-k32-b1
-*-g1-b2
g2-b2
600
hours
1440
2880
4320
5760
7200
8640
hours
400
0
10080
1 440
2880
4320
5760
7200
8640
10080
Time (hours/days)
Time (hours/clays)
60
120
180
240
420
6000
60
120
180
240
300
360
420
days
days
5500
5000
4500
w 4000
a,
17.8%
GT-31
"-^24.5%G2-B1
3500
G1-B2
m 3000
20.5%
G2-B2
2500
2000
1500
hours
1000
1440
2880
4320
5760
7200
8640
hours
1440
10080
t i m e (hours/days)
2880
4320
5760
7200
8640
10080
Time (hours/days)
Figure 5,12. Creep-strain evolution (Frame 4) of GFRP-3 reinforcement (left) and creep-strain
128.
60
120
180
240
300
360
60
420
120
180
240
300
360
i '
days
420
days
-G1-B1
--G1-B1
-G2-B1
->-G2-B1
-G1-B2
-*~G1-B2
. G2-B2
- G2-B2
12.5%
hours
hours
1440
2880
4320
5760
7200
8640
1440
10080
Time (hours/days)
2880
4320
5760
7200
8640
1 0080
Time (hours/days)
120
180
240
300
360
420
5000
days
4500
4000
3500
" t j 3000
=L
2 9 . 3 % Q2-B1
. S 2500
av*
1 9
'
0 %
. O 2 - B 2
2000
1500
1000
500
hours
1440
2880
4320
5760
7200
8640
1440
10080
2880
4320
5760
hours
7200
8640
10080
Time (hours/days)
Time (hours/days)
129.
60
120
180
240
300
360
5000
60
420
120
2400
days
4500
2200
4000
2000
3500
1800
180
240
300
360
420
days
"w 3000
8 . 5 % G2-B1
>-G1-B2
G2-B2
1500
1000
500
hours
1440
2880
4320
5760
7200
8640
hours
11X180
1440
Time (hours/days)
2880
4320
5760
7200
8640
10080
Time (hours/days)
hours
hours
1440
2880
4320
5760
7200
8640
1440
10080
Time (hours/days)
2880
4320
5760
7200
8640
10080
Time (hours/days)
130.
60
120
160
240
300
360
420
5000
120
180
240
300
360
420
2880
4320
5760
7200
8640
10080
davs
4500
4000
3500
^ 6.8%
i" 3000
",7.7%'
S 2500
' 9.8%
'
55 2000
1500 1000 500
bout's
0 1440
2880
4320
5760
7200
8640
hours
10080
1440
T i m e (hours/days)
F r a m e S : Steel 15-M beam samples
Frame 5: Steel 1S-M beam samples

FRP Strain evolution (Initial Strain = 980 m;.)
60
120
180
240
Concrete S t r a i n evolution
360
300
days
3500
3000
-.2500
w
3
"-G1-B1
2000
5 1500
-15%
1000
500
hours
hours
0
0
1440
2880
4320
5760
7200
8640
10080
1440
2880
4320
5760
7200
8640
10080
Figure 5.18. Creep-strain evolution (Frame 10) of steel reinforcement (left) and creep-strain
131.
1024
One year
average
NO
in
in
One year
Strain increase in reinforement (pe) after
OS OS 7*
OS NO NO
IT) r -
3000
Hrs
average
IT) T ?
0S
N
(N
rNO
-15.4
r - co CO
N CO CN O S 00
CO Os CN
OS t - oo in
CN m rf- f - CN co co
CN
NO
oo
CN
<N
VI
in o *i O S O S
N CN OO O S r co
^CN co
i O 00 ro oo oo n
T T 'ST- V V CO CN
^
CO r - in m
CO
CN ro ro
ro
m O S NO in
N
00 ro 00
CN m
OS
O S ; t> OS OS in
CO m ( N CN
SSH CN NO
OS
T--
OS 00 Os
Os
Os
CO CN
CN
00 oo
N
1
ON
CO N CN m m
CN
ro ro
rr 1
i t>
r-
CO
CN k> CN CN in O S SO
in
oo in m T 1 oo
in in 1
OO ( N in NO in CN
OS t 00 in Os
in
in
m
CO 'O ( N in
CN CN ii
CN CN
OO
NO
277
CN
<N
00
:
N
r - OO
00
CN OS
CN v .O
CN
co
in rtCO in (N
3000
hrs
CN CO 00
m Tf
> CO in V
m r - (N CN CO CN
N
O CN
fWM| N
t r - 00
CN N CN VO CO r -
v
in f>
N
r-C
N
V
CN
CO
CO
rf 71
NO
10.6
17.6
14.6
15.4
os
23.6
24.6
22.6
19.9
14.4
17.9
11.5
16.7
10.1
CN
I
1
|
Strain
increase
(%)
after
10000 hr
duration
Os
CN
CN
:
T>
CS
CO CO ^
CO
CN CO
00 00
CO
T-H
V SO
"3-
R-OS
CN CN
M RCO CN
00
in
cN
ro ro
OS
I
m
~i
VO
rlY
<
r^)
O 00 CN CN
CN OS T1 in 00 N ro
TF
^3- TT -<3CN CO
CN
<D
OO
4169
4291
4048
in
CO
00 s 1'
OS
ro ro
r- 00
CN O N
'
Bl-Fl
GFRP-1
B2-F1 11 GFRP-1
GFRP-1
B1-F2
GFRP-1
B1-F3
GFRP-2
B2-F3
GFRP-2
B1-F4 | GFRP-3
B2-F4
GFRP-3
B1-F5
GFRP-4
B2-F5
GFRP-4
B1-F6 | GFRP-5
B2-F6
GFRP-5
B1-F7
GFRP-6
B2-F7
GFRP-6
B1-F8
CFRP-1
CFRP-1
B2-F8
B1-F9
CFRP-2
CFRP-2
B2-F9
1
B1-F10
Steel
B2-F10 i;
Steel
Reinf.
type
CO
Beam
00 i 00
CO CN
o1I
ro
'
4299
4038
4009
4573
4249
3846 1
I 3862
4623
4137
3275
4564
3987
!1 2729 ; 2964
3196 i1 3004
3416
3078
3551
2782
| 2064
2368
| 1751 1| 2087
2107
2409
2251
2366
2098
2411 1, 2679
2459 |
2321
Avg
Actual Initial strain

(pe)
TD
OO
CO
CN
<3-
00
CO
in
CO r - V r - r - SO m
CO CO
CO ro
298 1
<N
OS
3706
4276
2847
3100
3247
3167
2216
1919 ]
2107
2330
2232
2545
2459
2526
>
00
in
oo
CN CN m rj-
NO
CN r - *co
CN in
CN in co
1000
hrs
1000
hrs
average
OS
CN
fa
CN
CQ
CN
co
Strain
increase
(%)
after a
one year
duration
^g?
ao
4>
00
G
3a
a
5
s*
l4-H
>
>
4)
C
S
4)
N O
rCO Tt l-H
OS m 00 m
m
Tt r- in CO CN Tt OinN NO
Tt in it CN 00 00 Tt
in co
r- t^ CO
r-
NO
CN o Tt co in N O m CO CN CN HCN O N rTt ON o in oo CN 00 Tt c- Tt
CO CN
l-H r-- N O
m CO CN co in N O Tt CO
CO in 00
00 N O Tt 00
CN CN CN CO
Tt N O
r^ O Tt N O Tt m
CN
r CO Tt rin 00 C
NO o
o CN ô
CN o m
m Tt CN co m N O in CO CO
r- c--
00
in
l>
o ON
m in
ON
CO N O
os r-j
vO N O
NO
>>
4)
<n
S
uK
ta.
I?
o
or
s
4>
>H
O
C
o
o
Dh
O
<4-1
O
c
0
1
"o
>
0)
g
'S
^l-H
OO
C
O
.O
Os 00
oo -V
o in r- NinO oin
C-- o CO l-H
>n Tt 00 NO M O
OS
in CN C
N as
g co
l-H
co N O N O o
NO
CO ^ 00 o
in in CN Tt in O Tt Tt CO CN V O Tt
M
Tt
CN NO l-H in CN M in O S 00 N O r- in
in I CN Tt
ON r- o o in
NO in CN Tt in R - in co CO iC
N NO NO
co
r--
ON
Tt o
in
CN n
Tt in 00
NO
'
OO
NO
NO
in O M
N
O S r- co 00
l-H
in Tt NO NO
OS
o CO
o O
Tt NO
ON
CO
in
CN -Tt
00
VO N
O
2
M
co
N oo co i-H o O N Tt in o
CO
CN in oo O S
in CN NO . . 1C
CN NO CN CO
in o Tt 00 o CO CO oo
r- <N
in NO Tt Tt CN CN N
Tt in CN
O NO m Tt
co in Tt m r-
4)
&D
S i ^
~ s
i-H CO m O N O S
N O oo in oo o
CN NO CO O S O N N O r- o
Tt CN in CO O N CN 00 o N O m O N oo m l-H
CO Tt r- Os
CN CO CN CN CO Tt CO CN CN l-H CO co CO CN Tt CN CO CN CO co
o NO CO OS 00 O O 00 Tt r- Tt CO NO
OS C
O CN in CO NO V NO o 00 m ON
CN CO CM CN CO M CO CN CN
CO co
CN NO
in CN
CN CO
r-r 00
CN i
CO CO CO
2
o
C3

00
>
<
4)
00
O
4)
00
Beam
J}
Actual Initial strain

(H)
SH
X)
0)
4>
oo I-H
in Tt CO CO
r- in o t O N O N oo CN oTt
in CO CN
in m CO CO
4>
Reinf.
type
03
U
f<
S3
o
Strain increase in reinforement ((is) after
NO CO O O o in N O NO L-H O NO NO Tt NO oo r- m CO
r- Tt
CO CN CN CO o Tt NO oo NO ON o oo 00 CN 00 C
N ro
o
O
o
o
o
r- l-H Tt
r-Tt
O
S
O
S
O
S
Os
OS 1
Tt
l-H t-
T-H:
HH l-H
iH
NO 00 in o CO N O m OO CO OS m CN 00 00 O O
in
ON
CO CN o CN in r- NO Tt OS OS in in NO 1
Tt l-H
l-H
T
T
C
O
O
o
o
O
S
c^
O
S
oo
oo
O S l-H
1H
1H r-- >n NO Tt N O CO r-
oo R CO r^
r- Tt
m
l-H I Tt Tt
NO
o oo s ON o os
T-H'
NO
i fa g
fa
fa
O o o
FA
I
tI
CQ
CN CN co
g
s
fa fa fa
o o o
co
fa
o
vO O
Tt oo
CO
OS
NO Tt 00 in m o
m cO 00 00
CO CO in o oo C
00
ol-H 00 O S o O N o
CO Tt
&
Tt Tt mI m NO N O
CN CN
&
g
G
13
4)
I
fa fa fa fa fa fa fa yfa fa fa 0 0
U
U
O o O O o O U
&
in m VO NO r 00 oo O N O S o
<N CN CO CO Tt
fa fa fai fa1 fa fat 21 fa fa i 'fa1 fa fa fa1 fa fa fa fa
CN T-H CN
CN TH CN
CN i-H1
CN
CN
CN i-H CN
CQ CQ OQ s
CQ
CQ s
m 3 CQ s CQ CQ CQ s CQ
m
13
4)
00
o
fa
R'I
CQ
co
co
Regarding time dependent behaviour, MacGregor and Wight (2005) indicated that
reinforcement strain/stress exhibits minimal increase (~3 % for GFRP and 6 % for steel) at the
cracked section, whereas the reinforcement strain/stress would increase significantly between
cracks (the value may triple). On the other hand, concrete stress at the tension surface changes
negligibly because creep is effectively redistributing stress from the concrete to the
reinforcement. This phenomenon is explained elaborately in the discussion section of the
study of Gross et al. (2009), indicating its impact on the growth of crack width with time. On
the other hand, the calculated compressive concrete strains showed reasonable agreement with
the measured values.
With time, the initial cracks propagate and more cracks gradually formed. Strain gradually
increases over time in both the concrete and FRP (Table 5.4 and Table 5.5) causing the neutral
axis to shift downwards (away from the extreme compression fibre) with time. This behaviour
is indicative of a redistribution of internal stresses over time, in FRP reinforced beams, similar
to that is found in steel reinforced beams. This redistribution is accompanied with an increase
in reinforcement stress and reduction in top fibre concrete stress required to maintain
equilibrium (Gross et al 2003). Strain increase takes place as well for both top fibre concrete
and FRP reinforcement (See Figure 5.19).
So I
Vo
//
Vt
Strain
Stress
Figure 5.19: Stress/Strain distribution at a cross-section

Significant creep activity was observed in the initial days of loading, for all beams comprising
different types of reinforcement, which is usually referred to as primary creep. The rate of
134.
creep strain accumulation tapers .down after a period of about 180 days. The graphs on the left
for Figure 5.9 to Figure 5.17 show the strain evolution within FRP reinforcement at midspan;
specific milestone values at 1000 hours, 3000 hours and 10000 hours are indicated in Table
5.4. For the GFRP and CFRP bottom reinforcement, the initial strain (ejrp_c) is in the vicinity
of 1393 to 3227 ps once the load was applied. This roughly corresponds to 16.9 % to 25.1 %
of the design tensile strain /. The strain increase, after one year, was in the range of 4 to 25
% (Table 5.4). On average and given comparable initial strain ejrp.o, the gradual strain increase
Aefrp for FRP under sustained flexural load is higher than that due to sustained axial load
(applied in Chapters 3 and 4). This difference is mainly due to the gradual curvature-increase
of the beams with time which, in turn, causes strain increase within the reinforcing FRP bars.
Thus, strain increase in FRP reinforcement comprises of two components; a creep component
and a due-to-curvature component.
As regards the creep of concrete, there is a significant difference between the time-dependent
increase in creep strain for steel reinforced beams, (Figure 5.18) and FRP reinforced concrete
beams of comparable nominal moment capacity Mn (Figure 5.14; Figure 5.15; Figure 5.16;
Figure 5.17). From the obtained results, the ratio of accumulated creep strain after 1 year to
initial strain,
ECP/EC0,
is on average 158 % for steel reinforced concrete and 60 % for concrete
beams reinforced with GFRP-5, GFRP-6, CFRP-1 and CFRP-2 (i.e. of comparable Mn). Thus,
the aforementioned ratio for FRP reinforced concrete is 0.4 times that of steel reinforced
concrete.
(5.12)
This implies the necessity of using a different value for the creep coefficient of concrete, or
introducing a correction factor, when referring to concrete reinforced with FRP. However, this
only relates to the behaviour of reinforced concrete elements under flexure. According to ACI
Committee 209, the creep coefficient of steel reinforced concrete is calculated as:
(5.13)
135.
where <p(t, t 0 ) is the creep coefficient at time t for age at loading to; 4>u is the ultimate creep
coefficient ranging from 1.3 to 4.15 and typically taken as 2.35); 4>corr is the proposed
correction factor for FRP reinforced concrete. If the time factor in the equation above applies
to both FRP and steel reinforcement, the ratio of
(pUFRP
to <pu steei will be equal to (p(t,to)
FRP/
<p(t,to) steel Specific creep (p(t,to) defines the ratio of to Acp to J which leads to:
(5.14)
In this respect, Brown (1997) proposed a value of 0 c o / r = 0.55. The latter fraction is calculated
based on the assumption that creep deflections are proportional to concrete compressive strain
(Acp
FFIP!
Acp
STEEI
SCP
FRP7
steel ) The reduction factor, (pcorr, is the outcome of a parametric
study of rectangular beams with p = 0.0033 - 0.03, and f'c = 20.5 - 67 MPa; results showed
that compressive strain will be 1.55 to 2.18 times greater in FRP-reinforced sections than in
similar steel-reinforced members. In addition, the cracked transformed moment of inertia will
be 3 to 4.25 times larger for steel-reinforced concrete sections, indicating that FRP-reinforced
concrete members should have deflection levels 3 to 4.25 times larger than in similarly
proportioned steel-reinforced beams. Thus, the analytical value for
4>U,FRP
4>U,steei should range
from 1.55/4.25 = 0.36 to 2.18/3 = 0.73, with the 0.55 average value providing best
approximation for the full range of parametric values. It is worth noting that the 0.4 ratio (ecp
FRP / ec steel) obtained from the current study falls within the aforementioned, 0.36 to 0.73,
range for 4>corr. In the coming chapter, it is shown how using (pcorr dramatically improves the
long-term deflection predictions for GFRP reinforced beams (Figure 6.16 to Figure 6.19).
5.4.2
In this sub-section, Figure 5.20 to Figure 5.25 display the initial deflection values, at midspan,
as well as the evolution of time dependent deflection of all tested concrete beams (14 GFRP
reinforced beams, 4 CFRP reinforced beams and 2 steel reinforced beams). For each beam, the
values displayed represent the average of measurements taken on both sides using a high
precision digital vernier calliper. The values written on the curves indicate the long-term to
initial deflection percentages.
136.
60
120
180
240
300
360
420
60
120
180
240
300
360
44.3%
420
'>s
42.6%
-*-Beam1
-Beam2
hours
hours
1440
2880
4320
5760
7200
8640
10080
1440
2880
4320
5760
7200
8640
10080
(The result of two sets/pairs)
0
60
1 20
1 80
240
300
360
420
42.8% davs
33.3%
-Beaml
Beam2
hours
0
1440
2880
4320
5760
7200
8640
10080
Figure 5.21. Time dependent deflection (Frame 3) of GFRP-2 reinforcement
137.
60
120
180
240
300
360
60
420
120
180
240
300
360
9.0
days
420
days
8.5
8.0
7.5
30%
7.0
28.9%
S, 6.5
i-Beaml
6,0 -
>-Beam2
V 5.5 i
E
I
O
5.0
4.5
4.0
3.5
hours
1440
2880
4320
5760
7200
8640
hours
3.0
10080
1440
Time (hours/days)
2880
T
4320
m
5760
7200
8640
10080
(hours/days)
60
120
180
240
300
360
420
days
44.4%
55.6%
*-Beam1
Beam2
: hours
hours
0
1440
2880
4320
5760
7200
8640
1440
10080
Time (hours/days)
2880
4320
5760
7200
8640
10080
Time (hours/days)
138.
hours
1440
2880
4320
5760
7200
8640
hours
10080
1440
Time (hours/days)
2880
4320
5760
7200
8640
10080
Time (hours/days)
Figure 5.24. Time dependent deflection (Frame 8 and Frame 9) of CFRP-1 reinforcement
(left) and CFRP-2 reinforcement (right).
60
120
180
240
300
360
420
6.0
d
j
5.5
5.0
4.5
0.5
00
J
0
->
hours
1440
2880
4320
5760
7200
8640
10080
Figure 5.25. Time dependent deflection (Frame 10) of steel reinforced concrete beams.
The deflection values in the graphs above (available in Figure 5.20 to Figure 5.25) are
summarized in Table 5.6. It is evident that the initial deflection of FRP reinforced beams
ranges from 0.36 % to 0.53 % of the 1500 mm loaded span (i.e. ~ L/276 to L/187). The total
deflection of the same beams, after one year of sustained loading, ranged from 0.47 % to 0.77
% of the loaded span (i.e. ~ L/212 to L/130). The latter values are considered high. In practice,
139.
when the magnitude of initial deflection is unacceptable, design would be controlled by

serviceability and the service load capacity of the section would be reduced until deflections
were reasonable in magnitude. As regards the steel reinforced beams, the average initial
deflection and total deflection after one year was as low as 0.13 % and 0.25 % respectively
(equivalent to L/790 and L/408, respectively).
Table 5.6: Experimentally obtained initial deflection and long-term deflection values
Type of
reinforcement
Bl-Fl
GFRP-1
B2-F1
GFRP-1
B1-F2
GFRP-1
B2-F2
GFRP-1
B1-F3
GFRP-2
B2-F3
GFRP-2
B1-F4
GFRP-3
B2-F4
GFRP-3
B1-F5
GFRP-4
B2-F5
GFRP-4
B1-F6
GFRP-5
B2-F6
GFRP-5
B1-F7
GFRP-6
B2-F7
GFRP-6
B1-F8
CFRP-1
B2-F8
CFRP-1
B1-F9
CFRP-2
B2-F9
CFRP-2
B1-F10
Steel
B2-F10
Steel
Average (FRP-RC)
Average (Steel-RC)
Beam
notation
5.4.2.1
Immediate
A initial
A longterm
A total
înitial
(mm)
(mm)
(mm)
7.24
7.84
7.5
8.02
6.8
7.4
7.87
7.21
5.5
5.43
5.82
6.23
6.85
6.53
6.09
5.98
6.12
6.25
1.87
1.93
6.71
1.90
2.69
3.64
3.19
3.56
2.27
3.18
2.85
2.94
1.72
1.59
2.58
3.23
2.49
2.72
2.51
2.65
3.00
2.68
1.61
1.96
9.45
3.69
9.93
11.48
10.69
11.58
9.07
10.59
10.72
10.15
7.15
7.09
8.4
9.46
9.34
9.25
8.6
8.63
9.12
8.93
3.48
3.89
2.75
1.79
/ 0/o)
0.48
0.52
0.50
0.53
0.45
0.49
0.52
0.48
0.36
0.37
0.39
0.42
0.46
0.44
0.41
0.40
0.41
0.42
0.12
0.13
0.45
0.13
,o,al
O/x
0.66
0.77
0.71
0.77
0.60
0.71
0.71
0.68
0.48
0.47
0.56
0.63
0.62
0.62
0.57
0.58
0.61
0.60
0.23
0.26
0.63
0.25
deflections
Immediate deflections occur once sustained load is applied. The experimental immediate
deflections are compared to the theoretical values calculated using the empirical and curvature
based methods, of ACI 440.1R-06, CAN/CSA S806-02 and the ISIS Canada Design Manual
(2007) (See equations 5.1 to 5.4 and Table 5.7).
140.
Table 5.7: Theoretical calculation of immediate deflections using different approaches
Beam
notation
Bl-Fl
B2-F1
B1-F2
B2-F2
B1-F3
B2-F3
B1-F4
B2-F4
B1-F5
B2-F5
B1-F6
B2-F6
B1-F7
B2-F7
B1-F8
B2-F8
B1-F9
B2-F9
BMTO
B2-F10
Type of
reinforcement
GFRP-1
GFRP-1
GFRP-1
GFRP-1
GFRP-2
GFRP-2
GFRP-3
GFRP-3
GFRP-4
GFRP-4
GFRP-5
GFRP-5
GFRP-6
GFRP-6
CFRP-1
CFRP-1
CFRP-2
CFRP-2
Steel
Steel
Aitexp
(mm)
ACI 440.1R-06
(fid = 0.2 pj pjb)
Aijheo
(mm)
7.24
1.55
7.84
1.55
7.5
1.55
8.02
1.55
6.8
1.59
7.4
1.59
7.87
2.33
7.21
2.33
5.5
2.29
5.43
2.29
5.82
2.58
6.23
2.58
6.85
2.36
6.53
2.36
6.09
2.59
5.98
2.59
6.12
2.49
6.25
2.49
1.87
1.85
1 1.93
1.85
Average ratio (FRP)
Average ratio (Steel)
Average ratio (FRP + Steel)
Ajjheo/
A exp
ratio
0.21
0.20
0.21
0.19
0.23
0.21
0.30
0.32
0.42
0.42
0.44
0.41
0.34
0.36
0.43
0.43
0.41
0.40
0.99
0.96
0.33
0.98
0.39
CAN/CSA S806-02
(20 increments)
A=fydx
Aijheo
(mm)
7.03
7.03
7.03
7.03
7.67
7.67
5.79
5.79
5.03
5.01
6.42
6.42
6.26
6.26
4.86
4.86
4.39
4.38
2.15
2.15
Aijheo/
A exp
ratio
0.97
0.90
0.94
0.88
1.13
1.04
0.74
0.80
0.93
0.91
1.10
1.03
0.91
0.96
0.80
0.81
0.72
0.70
1.15
1.12
0.90
1.14
0.93
ISIS Canda Design

Manual (2007)
A
Aijheo
(mm)
8.50
8.50
8.50
8.50
8.99
8.99
7.04
7.04
6.31
6.27
8.17
8.17
7.88
7.88
6.20
6.20
5.51
5.51
2.74
2.74
i.theo/
A exp
ratio
1.17
1.08
1.13
1.06
1.32
1.21
0.89
0.98
1.16
1.14
1.40
1.31
1.15
1.21
1.02
1.04
0.90
0.88
1.47
1.42
1.11
1.45
1.15
Comparing the theoretical immediate deflection to the measured values, it is realized that the
empirical model adopted by ISIS Canada yields the most conservative results for the FRP
reinforced beams; CAN/CSA S806-02 slightly underestimates the immediate deflection
values; ACI 440.1R-06 significantly underestimates the latter values. It is a known fact that
the modified Branson's equation adopted by ACI 440.1R-06 yields unsound deflection results
when the Mcr to Ma ratio approaches unity (Bogdanovic 2002). As regards the steel reinforced
beams, the original form of Branson's equation (without the ft factor) yields the best results
whereas the ISIS Canada model significantly overestimates the experimental deflection
values.
141.
5.4.2.2
Long-term
deflections
As mentioned earlier, ACI 440.1R-06 and CAN/CSA S806-02 have attributed to the factor X
(the long-term to initial deflection multiplier) certain values based on steel reinforced concrete
codes (ACI 318-08 and CSA A.23.3-04) as well as a limited number of earlier studies dealing
with the long-term deflection of FRP reinforced concrete (Brown 1997, Vijay and GangaRao
1998, Arockiasamy et al. 2000 and Gross et al. 2003). In this section, X is calculated based on
the obtained long-term and initial deflection results of the 14 GFRP and 4 CFRP reinforced
concrete beams from the current study. The calculated X's of earlier studies are also made
available in Table 5.8. The previous and current multiplier values are compared to ACI
440.1R-06 and CAN/CSA S806-02 predictions. As mentioned earlier, X is reduced to the time
dependent factor c for beams containing no compression reinforcement (See Eqn 5.7); is
equal to 1, 1.2, 1.4 and 2.0 for three, six, twelve and sixty months, respectively, by CAN/CSA
S806-02. ACI 440.1R-06 multiplies the latter values by 0.6 for the same time durations
respectively.
Except for the concrete beams reinforced with GFRP-1 and GFRP-2 of 9.5 mm diameter, all
other GFRP, CFRP and steel reinforced concrete beams in this study have an almost equal
nominal moment capacity M (See Table 5.3). As shown in Figure 5.22 to Figure 5.25, the
initial deflection of the FRP reinforced beams was, on average, 3.3 times that of the steel
reinforced beams. Whereas the FRP- to steel reinforced beams' long term deflection ratio was
1.1. Brown and Bartholomew (1996), reported that the corresponding values were 3.78 and
1.8, respectively, using 9.5 mm GFRP bars. Therefore, the long-term to initial deflection
percentages for FRP reinforced concrete specimens are significantly less than that of steel
reinforced beams.
The values obtained show that the value of X is significantly over-predicted by both North
American standards; yet ACI 440.1R-06 yields closer predictions (See Table 5.8). There exists
fair agreement between the X values of this study and earlier studies. The fact that ACI
440.1R-06 and CAN/CSA S806-02 overestimate the long-term deflection predictions for the
samples above cannot be generalized due to the absence of important parameters such as the
full-scale effect as well as top reinforcement.
142.
O A
S .A
-f! ^
<0tU "
II
oTt
Tf
Os
5
u
f
"-S
u
_
~
SiE
nS
I B,
ON
P<N.
B
_o
a
u01 3 ?
S,
s C
V
o
NO
cn
CN
NO
2 S'oS
S -a a
E 3- .2 so
H
t,
-8
y
>o
<N
CN
<N CN
CN
(N CN
CN
CN
CN
CN
CN
NO
CN
CO
bO
e
O S
SI
o
o
in
c 'v
ffi a
S O '-S S
09
" B
u a w
O
O
CN
CN
i
a
CN
CN
NO
CN
4( B.
PS H
ON
B
u
s
u
<>
l
B.
u
so
Vi
o
NO
ON
ON
00
ON
>> 5
.s. a
> g?
o
L.
CS
CQ
U
03
CQ
w
os
<n
2 r
s
N
O
O
O
su n
O
o
Se
or
<U CN
g. U w
V
CQ
5.4.3
The crack widths for all beams were measured over the one year test duration. The maximum
crack-widths, at midspan, were clearly marked and regularly monitored using a high precision
- jeweler's - microscope. Figure 5.26 to Figure 5.31 illustrate the maximum crack width
evolution for all beams. Similar to the strain evolution and deflection increase, the rate of
crack width increase is significant in the first 180-240 days then tapers down asymptotically
with time, to become almost negligible when approaching the end of the one year duration.
All FRP and steel reinforced beams cracked instantaneously once the sustained load was
applied.
Table 5.9 gives the average initial crack width at midspan for each beam; the crack-width
magnitude after 12 months; the percentage increase and ratio of crack-width to-date to the
limits specified by ACI 440.1R-06 and CSA S806-02 (0.028 in ~ 0.7 mm for FRP-RC beams
exhibiting internal exposure). For steel, this value corresponds to 0.016 in. (i.e. 0.4 mm). For
the FRP reinforced beams, the number of flexural cracks increased over time from just a few,
to as many as seven to twelve, after a one year duration. For the steel reinforced beams, the
number of cracks was 5 after 6 months. Indicated on each curve, in the graphs below (Figure
5.26 to Figure 5.31), is the immediate crack width and crack width after one year.
60
120
180
240
300
360
420
120
180
240
300
360
420
2880
4320
5760
7200
8640
10080
days
-B1-Cr1
5
J*
u
te
:U>
hours
hours
1440
2880
4320
5760
7200
8640
10080
1440
Time (hours/days)
Time (hours/days)
Figure 5.26. Time crack width test results for GFRP-1 reinforced beams (The result of two
sets/pairs)
144.
120
180
240
300
360
420
I days
0.3 - 0.68
-*-B1-Cr1
B1-02
-*-B2-Cr1
0.2 - 0.45
-B2-Cr2
0.2-0.45
'hours
1440
2880
4320
5760
7200
8640
10080
Figure 5.27. Time crack width test results for GFRP-2 reinforced beams
180
240
300
360
420
60
120
180
240
300
360
420
days
days
-B1-CM
-B1-Cr2
*Beam1
0.2-0.49 U-B2-Cr1
Beam2
0.15-0.40 ^*"B2-Cr2
ff.tjM3.34
0.15-0.33
hours
1440
2880
4320
5760
7200
8640
10080
hours
0
1440
2880
4320
5760
7200
8640
10080
Time (hours/days)
145.
60
120
180
240
300
360
1 :
60
420
days
0.9
0.9
0.8 -
0.8
0.7
0.7
= 0.6
180
240
300
360
B1-Cr1
B1-Cr2
0.5
0.5 -
$
a, 0.4
o
0.2-0.45
0.35-0.45
0.3-0.38
0.4
s
-
420
days
0.6
120
-*-B2-Cr1
B2-Cr2
0.3
0.3
0.2-0.31
0.2
0.1
hours
1440
2880
4320
5760
7200
8640
hours
1440
10080
Time (hours/days)
2880
4320
5760
7200
8640
10080
Time (hours/days)
60
120
180
240
300
360
60
420
1
days
0.9
0.9
0.8
0.8
0.7
-s- 0.7
e
5
0.6
0.6
--Bl.Crl
-*-B2-Cr1
0.4
180
240
300
360
420
; days
a
0.5
XL
0.4
V
B1-Cr2
0.5
1 20
-B2-Cr2
0.3
0.3
0.2
0.2 -
0.1
0.1
.0
hours
1440
2880
4320
5760
7200
8640
horn's
1440
10080
Time (hours/days)
2880
4320
5760
7200
8640
10080
Time (hours/days)
Figure 5.30. Time crack width test results for CFRP-1 reinforced beams (left) and CFRP-2
146.
60
120
180
240
300
360
0.32
420
days
0,28
0,24
"a
S
0.2
;-*-B1-Cr1
B1-Cr2
0.16
-*-B2-Cr1
|
*B2-Cr2
0.12
U
0.08
0.04
1440
2880
4320
5760
7200
8640
10080
Figure 5.31. Time crack width test results for Steel reinforced beams
The crack patterns for FRP-reinforced beams and steel reinforced beams were found to be
similar (See Figure 5.32). Figures 5.33 and 5.34 illustrate the crack width growth after 1 year
as a difference as well as a ratio of the crack width to initial crack width.
As mentioned earlier in section 2.4.5, there are two popular methods for predicting the
instantaneous flexural - most probable maximum - crack width of FRP-RC beams are Frosch's
(1999) equation and the expression of Gergely and Lutz (1968). The former equation, adopted
by ACI 440.1R-06 and CAN/CSA S6-06, is expressed as:
(5.15)
In this equation, ft is the ratio of distance from neutral axis to extreme tensile fibre and
distance from neutral axis to centre of the tensile reinforcement = (h - c)!(d - c); where h is
reinforcement; er is the reinforcement strain at service; dc is the concrete cover from extreme
reinforcing bars within outermost layer. The coefficient fa accounts for the degree of bond
between the FRP bar and the surrounding concrete. According to ACI 440.1R-06, fa = 1.0 for
deformed steel bar reinforcement; < 1.0 for superior bond relative to deformed steel bars; >
147.
with a mean of 1.10.
w = 22/3kbr{dcA)m
(5.16)
2
w = 2f3ktkbr.\d
+
(5.17)
beams, kt is taken as 2.0 for FRP reinforced beams exhibiting sustained load for a one-year
duration. The kt multiplier can also be applied to the Gergely-Lutz equation, yielding the
following equation:
w = 2.2/3ktkhr(dcA)m
(5.18)
Equation 5.15 and 5.16 were applied to all tested beams using values for the kb bond
coefficient ranging from 0.8 to 1.6. Given the obtained crack width measurements, it was
observed that the optimal kb values are 1.2 and 1.0 for equations 5.15 and 5.16, respectively
(See Figure 5.35 and Figure 5.36). Consequently, the optimal kt value, for each equation, was
calculated and found to be 1.7 and 1.5, respectively; the corresponding values for steel
reinforcement are 2.5 and 1.9, respectively (Figure 5.37). Alternatively, the kb bond factor was
calculated for. each type of reinforcement yielding average kb values for equations 5.15 and
5.16 (See Figure 5.38 and Table 5.10).
148.
5
Tt
ro
0.30
0.42
0.23
0.20
0.18
0.15 j1
0.20
0.21
0.21
0.19
0.33
0.20
0.41
0.28 |
0.28
0.28
0.04
0.02
in in -co
o
0.52
0.45
0.41
0.37 1
0.41
0.40
0.35
0.35 !
0.42
0.38
0.50
0.44
0.42
0.40
0.13 !
0.08
0.82
0.70
0.70
0.88
0.72
0.63
0.58
0.52
0.58
0.56
0.49
0.49
0.58
0.53
0.70
0.61
0.58
0.56
0.33
N
r- C
88.5
87.5
97.5
117.5
137.5
o O O l-H os
-
00 NO r-- 00 r- vO m,
!
1
i
1
1
Tt ro in Tt Tt in in NO Tt in Tt CO Tt m CO Tt Tf Tt CN l-H
m m in o o in in o in in
o Os ol-H <N CN oi-H ON Os CN i-H
GFRP-1
GFRP-1
GFRP-1
GFRP-1
GFRP-2
GFRP-2
GFRP-3
GFRP-3
GFRP-4
GFRP-4
GFRP-5
GFRP-5
GFRP-6
GFRP-6
CFRP-1
CFRP-1
CFRP-2
CFRP-2
Steel
Steel
Ratio (@ 12
months/limit)
Crack
spacing
after one
year(5)
r- SO 00 00 Tt NO oc O*
in in
00 ON
Bl-Fl
B2-F1 1I
B1-F2
B2-F2 !I
B1-F3
B2-F3
B1-F4
B2-F4 iI
B1-F5
B2-F5
B1-F6
B2-F6
B1-F7
B2-F7
B1-F8
B2-F8
B1-F9
B2-F9 1
B1-F10 1
B2-F10 1
Reinforcement
type
oK
a
g
t
o,
g
Specimen
O
Number
of cracks
after 1
year
oo OS OS;: 00
Number
of cracks
after 6
months
5
m
(N
Number
of cracks
after 3
months
Percentage
(width
increase)
Crack
width
after 12
months
0.58
g?
ao
-J
o
a
5
s
Initial
crack
width
(mm)
"3
o oON
ON
00 OS o os
in
NO Tt CO
CN
ON
Tt
OS in Tt r-. in o r-- NO 00 o CN m oo
CO in
IT) r-- .OS
O Tt o OS NO 00 CN OS CN m Tt Tt l-H
Tt NO in (N CN C
l-H
CN CN
Strain gauge on FRP bar
- - t - U / j
-V--J--
K '
Strain gauge on FRP bar
3 molnths
!
a
- 4 - - -/
- i,
6 mdnths
!
-1-1-
JJ
4-14--)-
<7
' i n - 12 mlonths
jJ 50 mm^.
500 mm
Jf
500 mm
500 mm
J 50 mm
Figure 5.32. Typical crack pattern for FRP reinforced concrete beams at three time intervals
(GFRP-5)
150.
0.35
CL
Li-
Q.
u.
CS V
u.
u.
CN
CO
CQ
CL
CL
LI.
CL
CL
<r
LL
0
ce
LL
O
CM
IX.
LL
1
CN
m m
a.
a:
LL
0
LL
CN
CD
t
CL
11
LL
0
LO
LL
CD
d.
01
LL
<3
lf>
LL
CM
DO
lO
l>
<>
CL
CL
CL
w
CL
LL
LL
O
LL
LL
CD
CD
to
LL
ii
CD
CD
LL
CN
DQ
r^
LL
r^
LL
CN
CD
03
LL
CO
LL
CN
m
(I)
LL
U)
LiCN
CO
OC ec
OC ec
CJ
CL
CL
CL
CN
CL
1L
LL
LL
LL
dC
CD
ec
OC ec
Figure 5.33: Crack width growth over 1 year of sustained loading.
CL
st
LL
- v CM* CN
P
a.
CL CL CL a
oc Q:
OC a: ec
LL L
L LL LL
V
O
(5 O
CN
CN
IL LL ci
LL er>
LL
t
CD
CN
CD
CD
CO
CL CL OL
OC ec CC
LL IL LL
V
O
o
-t
t
LL LL lO
u.
P
CE
L d
C
Li. ec
LL
o
0
in
LL ll>
IL
in" to w
CL CL CL
OC cc CC
LL IL IL
O
o 2<>
LL LL LL
CN
CD
CD
Csl
CN
CN
(D
CD
CO
^L
CD
CD
CD
CN
CQ
CL
ec
IL
o
00
LL
^
p CN
CL CL
OC OC
IL LL
O
tiro cr>
LL LL
CN
CD
<D
OL CD
LL O
o
LL
i
<T>
LL m
CN
CN*
LL
I
CM
CD
CQ
Figure 5.34: Crack width growth over one year of sustained loading as a multiple of initial
crack width.
151.
Frosch {1999) (kb = 0.8 for FRP and 1 for steel)
Gergely-Lutz (1968) (kb = 0.8 for FRP & 1 for steel)
0.8
0.8
R-V
c
s
0.7
o
Jf
-i
f
o
3
'
06
\/
0.6
GFRP-1
0.5
0.4
s
M
ss 0.3
GFRP-1
GFRP-2
GFRP-3
GFRP-3
GFRP-4
GFRP-4
GFRP-5
GFRP-5
GFRP-6
xx / >
: /
<u0.2
a
s
xx
;
/
GFRP-6
CFRP-1
0.3
CFRP-2
!/ * d
0.1
GFRP-2
0.5
CFRP-1
E
a
'a
I
0.7
CFRP-2
Steel
0.2
0.1
+ !
"
A
D,
Steel
y/
0.0
0.0
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
Measured maximum crack width
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
Frosch (1999) (kb = 1.0 for FRP and 1 for steel)
0.8
0.8
0.7
-S
!H
0.6
<j
f 0.5
<m>
s3
S
a
B
a*
GFRP-1
!_/ :
.a
X GFRP-2
0.4
GFRP-3
GFRP-4
A GFRP-3
XX
O GFRP-5
0.3
0.2
GFRP-1
X GFRP-2
GFRP-4
O GFRP-5
GFRP-6
GFRP-6
&
CFRP-1
CFRP-1
CFRP-2
Steel
"
CFRP-2
Steel
3
!S
V
0.1
I"""
U
0.0
0.0
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.8
;
:
J
5
0.7
/
: /'
X--
-a
0.6
is
(J
0.5
E
3
E
04
K-L
0.3
E
3
GFRP-1
X GFRP-2
A GFRP-3
au
a
A -
7T
A--
GFRP-6
CFRP-1
CFRP-2
Steel
0.2
0.3
0.4
0.5
0.6
0.7
0.8
J...A>..Uli
" ~ i * /A".
Ia
;
; /'
i
./*
;
0
GFRP-4
O GFRP-5
z.
0.1
A GFRP-3
"3
GFRP-1
X GFRP-2
KJ 0.1
U 0.0
/ i
xx
0.4
o
a
GFRP-4
O GFRP-5
0.2
1
j
0.6
0.7
GFRP-6
A CFRP-1
O CFRP-2 .
Steel
i
0.1
0.2
0.3
0.4
0.5
0.8
Figure 5.35: Measured versus calculated initial crack widths using kb values of 0.8, 1.0 and 1.2
152.

0.8
\
\
\ /
'/
\/
!
0.7
0.6
/\
1
...... L
0.5
<x
_
/
- i
0.3
0.4
;
1
Ata
v i
GFRP-1
GFRP-2
j . .
GFRP-4
O GFRP-5
!
.... J
A
1
GFRP-6
CFRP-1
CFRP-2
Steel
/
/
0.5
r
i
0.4
: +
A GFRP-3
4
:
GFRP-4
O GFRP-5
0.2
GFRP-1
X GFRP-2
0.3
GFRP-6
CFRP-1
CFRP-2
Steel
0.1
0.1
*'
0.0
0.0
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.1
0.2
0.3
i
:
0.8
:
0.7
0.7
0.6
V
0.4
I
:
GFRP-2
GFRP-4
O GFRP-5
\y \
/
GFRP-1
0.3
! ' A
- i._. / D.
0.5
A GFRP-3
i
1
GFRP-6
CFRP-1
CFRP-2
Steel
0.4
0.3
0.2
0.1
0.2
0.3
0.4
0.5
0.6
0.0
0.7
0.8
0.8
0.1
';
0.2
0.3
0.4
i.
!
0.5
7~
.......
/*
/
0
....
:
J/
i
1"
L
i
0.1
0.7
L .
P :+
~ f A; / A
,
: n|z*
0.1
0.0
0.6
;
i
0.6
Xx
0.5
>< x
0.5

0.8
0.4
0.2
0.6
A GFRP-3
0.2
0.7
T3

0.8
GFRP-1
GFRP-2
A GFRP-3
GFRP-4
O GFRP-5
- -
'
GFRP-6
CFRP-1
CFRP-2
Steel
0.6
0.7
0.8
Figure 5.36: Measured versus calculated initial crack widths using kb values of 1.4 and 1.6
153.
Frosch (1999)
(1^ = 1.2 for FRP and 1 for steel)
(k,= 1.7 for FRP and 2.5 for steel)
Gergely-Lutz (1968)
(kb = 1.0 for FRP and 1 for steel)
( k t = 1.5 for FRP and 1.9 for steel)
0.8
X X j y/i
a
w
2
e
GFRP-1
GFRP-2
0.4
0.3
GFRP-4
O GFRP-5
A A
l/i
GFRP-6
CFRP-1
CFRP-2
Steel
0.6
0.3
0.4
0.5
0.6
GFRP-1
X GFRP-2
A GFRP-3
GFRP-4
O GFRP-5
\ / "A l
-en-
0.8
0.7
.J
U
0.2
0.4.
aa o.i
' 0.1
s
o
0.1
it
10 '5
A GFRP-3
':
\ XX
GFRP-6
CFRP-1
CFRP-2
Steel
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
Figure 5.37: Measured versus calculated crack widths after one year using optimal kb and kt
values
Frosch (1999)
Gergely-Lutz (1968)
2.5
a.
cc
l l
15
a.
a:
l l
13
T
| T
a. a.
cl
cc cc cc
Li.
l l
u.
O O O
tn|
1
ix
CC
u_
o
c>
cl
cl
cc
l l
CC
l l
O
CC
cl
cm
1
cl
cm
1
cl
l l
l l
l l
t
cl
m
cl
it)
a.
cl
cl
cl
l
cl
l l
l l
l l
l l
l l
l l
l l
"t
cc
CC
CC
CC
0
CC
CC
CC
CC
CC
Figure 5.38: Calculated bond coefficient kb for the different types of FRP reinforcement
154.
Table 5.10: Average calculated kb bond coefficient for the different types of FRP
reinforcement
Type of FRP reinforcement
GFRP-1
GFRP-2
GFRP-3
GFRP-4
GFRP-5
GFRP-6
CFRP-1
CFRP-2
Frosch (1999)
1.21
0.66
0.74
1.10
0.92
1.24
1.99
1.90
Gergely-Lutz (1968)
0.92
0.50
0.57
0.85
0.72
0.97
1.51
1.45
NB. The latter two kb values for CFRP-1 and CFRP-2, in Table 5.10, are exceptionally high
since the selected bars were of non-typical batches that had inferior bond properties.
5.5
Conclusions
In this chapter, 20 (14 GFRP-; 4 CFRP- and 2 steel-) reinforced concrete beams were tested in
third-point sustained load set-ups for one-year within standard laboratory conditions. The
dimensions of the beams were 100 x 150 x 1800 mm. Along this duration, the long-term
performance of all beams was monitored as of (i) strain variation for bottom reinforcement as
well as top concrete fibre; (ii) long-term deflection and (iii) crack width. Based on the
obtained data and conducted analysis, the following conclusions can be drawn:
1
The rate of increase in strains and deflections is high in the initial periods of loading
and reduces asymptotically with time under sustained loading. The strain increase in
GFRP and CFRP reinforcement after one year ranges between 4.2 % and 24.6 % of
the initial strain, mainly due to the gradual increase in beam curvature rather than
creep strain.
For Group-2 beams of approximately equal nominal moment capacity, the initial
deflection of FRP reinforced beams is on average 3.3 times that of the steel
reinforced beams. As regards long term deflection, the long-term to initial deflection
percentage is 41 % and 94 % on average for FRP-RC beams and steel reinforced
beams after the one year duration, respectively.
The creep coefficient of FRP reinforced concrete (p(t,to) under flexural loading was
found to be 0.4 times that of steel reinforced concrete. Thus, a correction factor <pcorr
155.
of 0.4 should be introduced when calculating the the former value. The 0.4 value falls
in the (0.36 to 0.73) range propsed by Brown (1997).
4
Three methods were used to calculate the immediate deflection of the FRP and steel
reinforced concrete beams. On average, the ACI 440.1R-06 empirical method
significantly underestimated the measured FRP-RC deflection results (by 67 %). On
the other hand the CAN/CSA S806-02 curvature method was inferior by 10 % and
the ISIS Canda Design Manual (2007) overestimated FRP-RC immediate deflection,
on average, by 11 %. As regards the steel reinforced concrete the calculations
underestimated by 2 %, overestimated by 14 % and overestimated by 45 % for the
three methods, respectively.
Though favourable, the long term deflection multiplier method (>1), indicated in ACI
440.1R-06 and CAN/CSA S806-02, significantly over-predicts the long-term
deflection values of the FRP reinforced concrete beams. The long-term to initial
deflection ratio, on average, was underestimated by 51 % and 70 %, respectively.
This fact, however, cannot be generalized due to the missing parameters that are
typically available in full-scale FRP reinforced concrete beams.
After the one year duration, the crack width for all beams was yet below the 0.7 mm
limit indicated by ACI 440.1R-06. The crack-width-increase percentage varied
according to the reinforcement ratio and type of reinforcement.
The crack width prediction equations adopted by ACI 440.1R-06 and CAN/CSA S606, on one hand, and by the ISIS Canada Design Manual (2007) on the other hand,
yield satisfactory results when the kb bond-coefficient factor is 1.2 and 1.0
respectively. From the obtained data, the time-dependent kt multiplier is deduced as
1.7 and 1.5 for both models, respectively.
For the crack width prediction equation (Frosch 1999) adopted by ACI 440.1R-06
and CAN/CSA S6-06, the kb bond-coefficient was calculated as 1.21, 0.66, 0.74, 1.1,
0.92, 1.24, 1.99 and 1.90 for GFRP-1, GFRP-2, GFRP-3, GFRP-4, GFRP-5, GFRP6, CFRP-1 and CFRP-2, respectively. The latter two values for CFRP-1 and CFRP-2
are exceptionally high since the selected bars were of non-typical batches that had
inferior bond properties.
156.
Chapter 6: Modelling the Long-term Deflection o f Full-scale GFRP-RC Beams under Uniform Distributed Load
CHAPTER 6
MODELLING THE LONG TERM DEFLECTION OF FULL-SCALE
GFRP-REINFORCED CONCRETE BEAMS UNDER UNIFORM
DISTRIBUTED LOAD
6.1
Introduction
In the previous chapter, the long term deflection of 20 GFRP, CFRP and steel reinforced
concrete beams were studied under sustained third point load, for a one-year duration. The
initial deflection values, obtained through experiments, were compared with the theoretical
predictions of initial deflections using two empirical methods (ACI 440.1R-06 and ISIS
Canada design manual) as well as the curvature method available in CAN/CSA S806-02. As
regards the long term deflection of FRP reinforced concrete beams, the X multiplier (the ratio
between long-term to initial deflection), for all FRP reinforced beams, was calculated using
the obtained initial and long-term deflection values.
&(creep+shrinkage)~
^(î)sustained
-(6.1)
The multiplier/ratio, X, values were compared to the predictions of ACI 440.1R-06 and
CAN/CSA S806-02. In light of the obtained data and data made available by earlier efforts,
the North American standards are conservative when it comes to predicting long-term
deflection of FRP reinforced concrete, as a function of initial deflection.
In Chapter 5, however, as well as the majority of long-term deflection studies to date, a
symmetrical four-point loading setup is adopted. A typical scenario is that FRP reinforced
beams of relatively small scale are cast of normal concrete without top/compression
reinforcement or shear reinforcement. The rendered information would typically deal with (i)
the strain variation within concrete and reinforcement with time, (ii) the time dependent
deflection at midspan and (iii) the maximum resulting crack widths, at midspan.
Calibration/modification of equations and models designed for steel reinforced concrete
would take place based on the experimental findings. As per the recommendations of Brown
1997, there is a need to study the long-term behaviour of full-scale FRP reinforced beams as
well as the moderating effect of FRP and/or steel compression reinforcement. Regarding the
157
loading method, a more realistic approach would be to apply uniform distributed load, in
agreement with the typical design philosophy for reinforced concrete beams.
Arockiasamy et al. (2000) studied the long-term behaviour of concrete beams reinforced with
carbon FRP bars (for a period exceeding 470 days). Four rectangular concrete beams (two
sets) were cast in two sizes; 152 mm x 203 mm x 2438 mm (for Set 1 of / c ' = 32 MPa) and
152 mm x 152 mm x 2438 mm (for Set 2 of f'c = 43 MPa). Each beam contained 2-7.5 mm
CFRP bars for bottom reinforcement, likewise for top reinforcement (hanger); 9.5 steel
stirrups spaced at 75 mm to 150 mm for ends to midspan, respectively. The beams were
simply supported and loaded with concrete blocks to simulate uniform sustained load. Strain
gauges were installed on the top concrete compression surface as well as attached, at midspan,
onto the embedded FRP reinforcement. The applied moment Ma was 77 % and 123 % of the
respective cracking moment Mcr for Set 1. For Set 2 beams, Ma was 110 % and 123 % of the
respective Mcr.
After a 470 day period, it was observed that the relative deflection increase of long-term
(accumulated) to initial deflection was 15 %, 115 %, 65 % and 71 % for beams B l , B2, B3
and B4, respectively. The increase in top surface (concrete) strains, relative to the
instantaneous values, was 101 %, 151 %, 209 % and 245 % for the same beams, respectively.
The authors used a computer program based on the age-adjusted effective modulus method
(made available in Ghali, Favre and Elbadry 2002) to predict the long-term deflection of the
beams. The creep and shrinkage coefficients were calculated based on ACI Committee 209
recommendations (1992) and CEB-FIP Model Code (1990). The theoretical curve rendered
good agreement with the measured values. Furthermore, a more-convenient (simplified)
equation - a modification of the empirical equation available in ACI 318-08 and CSA A.23.304 - was proposed. The latter model addressed two main parameters (i) the age of loading and
(ii) the upper/compression reinforcement. The study of Arockiasamy et al. (2000) is the
closest available to full-scale FRP reinforced concrete beams. Yet, more effort is necessary to
increase the available database regarding the long-term performance of FRP reinforced
concrete. Different bottom reinforcement ratios, different levels of sustained load, different
types of upper reinforcement are parameters also need to be addressed further.
158
6.2
Research Purpose
This phase of the study provides essential data and analysis regarding the behaviour (strain
increase, deflection and crack width) of GFRP reinforced concrete beams, under constant
sustained load. The main objective is to observe and model the long-term deflection of fullscale concrete beams reinforced with GFRP, under sustained load. In this study, three
parameters are addressed: (i) the type of compression reinforcement (GFRP or steel); (ii) the
GFRP bottom reinforcement ratio and (iii) the concrete compressive strength. Two modelling
methods, for theoretical prediction of long-term deflection, are applied and compared: (i) the
an empirical approach adopted by ACI 440.1R-06 and CAN/CSA S806-02, to which different
factor values are multiplied to account for the long-term deflection and (ii) the age adjusted
effective modulus approach (a numerical approach) that involves programming to calculate
beam curvature, deflections and strains on the basis of the theory made available by ACI
Committee 209 and CEB-FIP Model Code 1990. Furthermore, the increase in crack width
over time as a result of sustained load was studied and modelled.
6.3
6.3.1
Materials
The materials used in this study are two batches of ordinary concrete; 9.5 mm sand-coated
GFRP bars (noted as GFRP-1); No. 10M steel bars. The two batches (one batch per two
beams) were MTQ Type-V with a target compressive strength of 35 MPa after 28 days. The
mixture proportion per a cubic meter of concrete was as follows: coarse aggregate content of
646 kg with size ranging between 10 and 20 mm, 341 kg with a size ranging between 2.5 and
10 mm and fine aggregate content of 717 kg, cement content of 455 kg, water-cement ratio
(w/c) of 0.35, air entrained of 5.0-8.0 %, and water-reducing agent. The slump of the fresh
concrete was measured before casting and was about 100 mm (4.0 in.). Six concrete cylinders
(100 x 200 mm) were cast for each batch and tested the day the beams were installed under
sustained load (ie. after 5 months from beam casting). The average concrete compressive
strength f'c was 36 and 32 MPa and the corresponding measured modulus of elasticity, Ec, was
30 GPa and 29 GPa for each batch, respectively.
159
The 9.5 mm sand-coated GFRP-1 bars were used as bottom reinforcement for all tested beams
(Figure 6.1). For two of the four beams, No. 10M steel bars were used as top reinforcement.
No. 10M steel bars were also used to fabricate stirrups for all beams. The mechanical
properties of the GFRP and steel reinforcing bars used in this study are summarized in Table
6.1. The GFRP bars are made of high-strength E-glass fibres impregnated in vinylester resin.
The bar's circular cross section is manufactured by a process that couples pultrusion with
sand-coating along the external surface of the bar. Further information, as to the physical
properties of the GFRP reinforcement, is presented in Table 6.2.
Figure 6.1: GFRP-1 and steel reinforcing bars.
Table 6.1: Mechanical properties of the reinforcing bars.

Reinforcement type
Diameter (mm)
Area (mm )
fu,ave (MPa)
/v(MPa)
GFRP-1
Steel (10M)
9.5
11.3
71
100
85434
N/A
N/A
412
46.91.2
201
Ef,ave & Es ( G P a )
18232767
N/A
ave *
N/A
15931
*
aw is the mean tensile strain at rupture; s ju is the guaranteed rupture strain (e/ = e,
3a)
160
Table 6.2: Physical properties of GFRP bars

Diameter (mm)
Surface
Fibre content
( % volume)
Longitudinal coefficient
(x 10'6 C)
Transverse coefficient
(x 10'6 C)
Density
(g/cm3)
Water absorption
9.5
Sand
coated
56.8
7.0
28.6
1.94
-1
(%)
6.3.2
Glass transition
temperature (C)
105
Cure ratio ( % )
98
Description of Test Specimens
Two sets of four rectangular concrete beams were reinforced with GFRP bars, as bottom
reinforcement. The beams measured 4282 mm in length with a rectangular cross section of
(215 mm x 400 mm). Set A & B, w i t h / ' c = 36 MPa, contained 3 GFRP (9.5 mm) bars as
upper/compression reinforcement, per beam; whereas the set C & D, withf' c = 32 MPa, had 2
No.lOM steel bar. Each set comprised of two beams, with 5 No. 3 GFRP (9.5 mm) in one
beam and 8 No. 3 GFRP (9.5 mm) in the other. Figure 6.2 shows the dimensions and
reinforcing details of the four beams. Figure 6.3 shows the reinforcing cage in the formwork.
Sizeable concrete blocks of dimensions 610 mm x 762 mm x 1219 mm (2' x 2.5' x 4'), and
3000 lb in weight (13334 kN), were used to simulate the sustained distributed loads (Figure
6.4 and Figure 6.5).
161
Applied load =12.5 kN/m'
35
35
No. 10M(ffi200mm
?no ?on 700 . 7.00

5 GFRP 0 9.5 mm or 8 GFRP 0 9.5 mm
4224
3810
.?36.i
4282 mm
3 GFRP 0 9.5 mm
3 GFRP 0 9.5 mm
2 No. 10 M
2 No. 10 M
5 GFRP 0 9.5 mm
8 GFRP 0 9.5 mm
8 GFRP 0 9.5 mm
5 GFRP 0 9.5 mm
|. 215 .|
121M
Beam B
Beam C
Beam A
[.
215
Beam D
Figure 6.2: Reinforcement details of the beams.
6.3.3
Instrumentation and Test Setup Description
After casting and curing, the beams were simply supported on a hinge and a roller in a
standard laboratory environment (ambient temperature of 23 1C and relative humidity of 50
2 %). Electrical strain gauges (10 mm Kyowa strain gauges of 120 ohm resistance) were
installed, at midspan, on the surface of two of the bottom reinforcing GFRP bars. Likewise,
three concrete gauges were installed, at midspan, on the upper/compression surface of each
beam (67 mm Kyowa strain gauges of 120 ohm resistance). All gauges were protected from
the ambient surroundings by two kinds/layers of epoxy coating. For an initial period of 35-38
days after placement of sustained load, the wires attached to the strain gauges were connected
to a data acquisition system (manufactured by Vishay instruments). One LVDT per beam, of
accuracy of 0.01 mm, was installed and connected to the same system, where timedependent deflection would be monitored for the same time period.
162
Figure 6.3: Reinforcement cages prior to casting
Figure 6.4: Transportation of concrete blocks to laboratory
Figure 6.5: Forklift truck loading blocks on beams
163
Figure 6.6: Instrumented beams under uniform sustained block load
The four beams were assembled in two pairs over simple supports in a parallel manner. The
concrete blocks (610 mm x 762 mm x 1219 mm), simulating the distributed load, were carried
by a forklift and arranged on the beams, in a manner where each two beams carry six blocks
of concrete on the 3810 mm loading span, as shown in Figure 6.5 and Figure 6.6. Caution was
exerted, on loading, to ensure that each beam would carry half the block weight; in turn all
beams would be subjected to the same applied moment Ma (See relevant details in Table 4.4).
The dates of beam-casting and sustained-load application are made available in Table 4.4. The
duration of block loading was 3 to 4 hours. Hair cracks were evident during the loading
process and crack monitoring took place along test duration.
Table 6.3: Test specimens and relevant details

Beam
Concrete f'cu (MPa)
Bottom Reinf.
Top Reinf.
A bottom-reinf(jnm
Reinf.
Mcr
Mn
Ma
Ratio ( % )
(kN.m)
(kN.m)
(kN.m)
Ma/Mcr
Ma/Mn(%)
Casting Date
Loading Date
Beam A
36
5 GFRP (9.5 mm)
3 GFRP (9.5 mm)
355 (5 bars)
0.44 (balanced)
17.2
106
23
1.3
21.6
10/12/2007
13/5/2008
Beam B
36
8 GFRP (9.5 mm)
3 GFRP (9.5 mm)
568 (8 bars)
0.7 (over-reinf)
17.2
163
23
1.3
14.1
10/12/2007
13/5/2008
164
Beam C
32
8 GFRP (9.5 mm)
2 steel No. 10 M
568 (8 bars)
0.7 (over-reinf)
15.4
160
23
1.5
14.4
7/12/2007
10/5/2008
Beam D
32
5 GFRP (9.5 mm)
2 steel No. 10 M
355 (5 bars)
0.44 (balanced)
15.4
105
23
1.5
21.9
7/12/2007
10/5/2008
6.3.4
Design of FRP Reinforced Concrete Members
Regarding the design of the FRP-reinforced concrete beams, beams A and D where designed
as balanced sections, whereas beams B and C were over-reinforced (pfrp~ 1.5 pjrpb)-
The balanced reinforcement ratio, Pfrpb, is calculated as:

\
= a\P\
Pfrpb
(6.2)
\cu + frpu J
<t>f ffipu
where f'c is the concrete compressive strength (MPa); the ratio of average concrete strength in
rectangular compression block to the specified concrete strength a\ = 0.85 - 0.0015 f'c > 0.67;
the ratio of depth of rectangular compression block to the depth of the neutral axis
0.0025 f'c
= 0.97 -
> 0.67; (f>c and <pf are the material resistance factors for concrete and FRP,
respectively; ffrpu is the ultimate tensile strength of FRP (MPa); ecu and frpu are the ultimate
strain in concrete in compression and the ultimate strain in FRP in tension, respectively.
For the values in Table 6.3, Mcr (cracking moment) and Mn (nominal moment capacity) are
calculated using the two equations below (ISIS Canada Design Manual 2007):
M c r = f /(/y,
(6.3)
Mn = pfrpffrp(\-0.59pfipffrp/fc)bc?
(6.4)
where fr, the modulus of rupture of concrete, ranges from 2.4 to 2.6 MPa (obtained from
laboratory splitting tests); I t is the moment of inertia of the transformed uncracked section; jîs
the distance from the centroid of the uncracked section to the extreme surface in tension; pfi.p
is the reinforcement ratio; b and d are the cross section's width and depth, respectively. The
axial stress in the reinforcement at f a i l u r e , ^ , is calculated using Eqn 6.5:
\{EfrpsJ
=
J frp
0.85A/;
p
,
frp
_ 0 5 ,
cu
frp
(6 5)
cu
^ . J )
where Efrp is the modulus of elasticity of GFRP, ecu is the ultimate strain in concrete and /?/ is
the stress-block factor for concrete.
165
6.3.5
Following this initial phase, measurements took place manually until the end of the six month
duration. A portable strain indicator P-3500 (manufactured by Intertechnology Inc.) was used
to take strain measurements for all gauges (Figure 6.7); deflection was measured by a high
precision vernier-caliper (of 0.01 mm accuracy) aided by demec points installed on all
beams. Crack widths were taken using a jeweller's microscope of 0.01 mm accuracy (Figure
6.8). For the first two weeks, readings were taken every half-hour using the data acquisition
system (Figure 6.9); manually every two weeks for the consecutive two months then every 30
days until the elapse of the six month period.
Figure 6.7: Deformometer (P-3500 Portable strain indicator) for strain measurement
Figure 6.8: Crack detection and crack width measurement using a high precision microscope
166
Figure 6.9: Data acquisition system; LVDT connected to data acquisition system
6.4
6.4.1

Creep Induced Strain in Compression Concrete and FRP Reinforcement
In this sub-section, Figure 6.10 and Figure 6.11 display the tensile strain evolution for the
bottom reinforcement as well as the creep evolution of compression concrete. The strain
gauges were installed for the reinforcement and likewise on the concrete compression surface
at midspan.
167
6000
months
5500
5000
4500
d-1
4000
i
Bea m D
3500
Bpsm P
3000
et
JBrr-
2500
2000
Beam b
Beairib
ff --
1500
1000
500
davs
0
0
15 30 45
60 75 90 105 120 135 150 165 180 195 210
Time (days/months)
Figure 6.10: Strain evolution for FRP reinforcement at midspan for beams A, B, C and D
Table 6.4: Details on creep strain evolution within reinforcement (at mid-span)
Beams
Beam
Beam
Beam
Beam
A
B
C
D
Top
reinforcement
Bottom
reinforcement
3 GFRP-1
3 GFRP-1
2 10M
2 10M
5
8
8
5
f'c
36
36
32
32
Calculated
Initial
strain
(pe)
Actual
Initial
strain
(HE)
Strain
after 6
months
(pe)
Strain
increase
to date
(%)
3056
1919
1925
3164
2465
1814
2354
3661
2859
2195
2717
4045
16.0
21.0
15.4
10.5
168
Strain Increase (ps) after

1
month
268
315
185
290
3
months
305
343
267
380
6
months
394
381
364
384
2000
months
3ea DC
1800
1600
1400
1200
S 1000
'3
cz5 800
ti B
600
400
3eain A
IM
200
davs
0
0
15
30
45
60
75
90 105 120 135 150 165 180 195 210
Time (days/months)
Figure 6.11: Creep-strain evolution of compression concrete for beams A, B, C and D
Table 6.5: Details on creep strain evolution of compression concrete (at mid-span)
Beams
Beam
Beam
Beam
Beam
A
B
C
D
Top
reinforcement
Bottom
reinforce
-ment
f'c
3 GFRP-1
3 GFRP-1
2 10M
2 10M
5
8
8
5
36
36
32
32
Initial
strain
348
279
1410
1512
Strain to
date
(l^s)
Strain
increase
to date
(%)
400
484
1740
1710
14.9
73.5
23.4
13.1
Strain increase (|j,s) after

1
month
-47
51
163
119
3
months
-6
33
238
158
6
months
52
205
330
198
After initial cracking and the gradual formation of more cracks, strain gradually increases over
time in both the concrete and FRP causing the neutral axis to shift downwards with time. This
behaviour is indicative of a redistribution of internal stresses over time, in the GFRP
reinforced beams, similar to that is found in steel reinforced beams. This redistribution is
accompanied with an increase in reinforcement stress and reduction in top fibre concrete stress
169
required to maintain equilibrium (Gross et al 2003). Strain increase takes place as well for
both top fibre concrete and FRP reinforcement (See Figure 6.12).
f ft
Strain
Stress
Figure 6.12: The stress strain distribution at a cross-section

Significant creep activity was observed in the initial days of loading for the four beams;
typically known as primary creep. The rate of creep strain accumulation tapered down after a
period of about 180 days (See Figure 6.10 and Figure 6.11). Table 6.4 and Table 6.5 give
milestone values for the strain increase of FRP bottom reinforcement and compression
concrete at 1 month, 3 months and 6 months. It is worth noting that the calculated concrete
compression strains and reinforcement strains show good proximity with their measured
counterparts, indicating strain compatibility for the four beams (Table 6.4).
The strain increase, in the bottom GFRP reinforcement after six months was lower for the two
beams (A and B) of higher concrete compressive strength (f'c= 36 MPa); despite the fact that
beams C and D contain upper compression steel; compression steel has a moderating effect on
beam deflection and in turn strain increase in the bottom reinforcement. As shown in Table
6.4, the induced initial strain values in the GFRP bottom reinforcement for beams A, B, C and
D ranged between 1814 and 3661. The strain increase after six months shows significant
proximity for GFRP reinforcement with similar initial strain values. The percentage of strain
increase after six months to initial strain is 16, 21, 15 and 10.5 % for beams A, B, C and D,
respectively. As indicated above, the initial induced strain for beams A, B and C are all in the
vicinity of 1800 to 2500 pe and their corresponding strain increase after six months were very
close. The initial strain values are averaged along with their corresponding strain increase;
170
then compared to the outcome of five samples of the same type of GFRP-1 bars (See Chapter
1) of comparable initial strain values yet erected under axial sustained load (see Table 6.6):
Table 6.6: Comparison between strain increase due to flexural loading and axial sustained
load
No. of samples
Mean (Initial strain) (as
Mean (Strain increase)
(j, after six months
Flexural loading
3
22111348
Axial sustained load

5
24271218
379 1 5
2 2 2 1 110
In Table 6.6, the resulting creep strain, due to flexural loading, exceeds that due to axial
sustained load experiments. This difference (approximately 70 %) is mainly due to the gradual
deflection-increase of the beams with time which, in turn, causes strain increase within the
reinforcing GFRP bars. Thus, strain increase in the FRP reinforcement comprises of two
components; a creep component and a due-to-curvature component.
As regards the creep of concrete, the accumulated creep strain for the four beams was
insignificant after 6 months. Nevertheless, the creep coefficient correction factor, (pcorr= 0.55,
proposed by Brown (1997) is applied in the long-term deflection analysis of the following
subsection (For further details regarding the FRP creep correction factor refer to sections
2.3.3.3 and 5.4.1).
6.4.2
6.4.2.1

Immediate
deflections
Immediate deflections occur once the sustained load is applied. Figure 6.13 shows the initial
deflection results as well as the long-term deflection results over the duration of six months. It
is observed that deflection at midspan is greater for the two beams of lesser concrete
compressive strength despite their comprising of upper compression steel reinforcement
~ 4.3
EGFRP)-
(ESTEEI
This is due to the lower modulus of elasticity and lower stiffness. It is also
noticed that the increase in bottom GFRP reinforcement decreases the initial deflection at
midspan by 16 % and 29 % for set A-B and set C-D, respectively (See Table 6.7).
171
In Figure 6.13, the initial deflection values, as a percentage of the loaded span of 3810 mm,
are 0.096 %, 0.074, 0.118 and 0.137 for beams A, B, C and D, respectively. These values
correspond to L/1038, L/1346, L/849 and L/731, respectively. As regards, the total deflection
of the same beams, after six months of sustained loading, the corresponding ratios were L/560,
L/256, L/491 and L/439 (Table 6.8).
months
14
13
12
11
10
Bea m D
8
&
'C
o
4>
5=
r-
jk
m.
pea fG
BeanrA
Bea m B
4
3
2
davs
0
15
30
45
60
75
90 105 120 135 150 165 180 195 210
Time (days/months)
Figure 6.13: Deflection results for all beams
The immediate deflection values for the four beams were calculated and compared to the
measured values. Two common approaches are available for immediate deflection calculation:
the effective moment of inertia approach and the curvature approach. The first approach - the
effective moment of inertia I e , approach is adopted by ACI 440.1R-06 wherein the proposed
equation interpolates between the moment of inertia of the gross uncracked concrete section,
I g , and the moment of inertia of a transformed cracked section, I cr , to account for the tension
stiffening effect of the concrete in tension. The I e approach is given by Eqn 6.6 as follows:
/ =
'M.
1+
kM.J
172
L < /.,
(6.6)
where Mcr and Ma are the cracking moment and the maximum applied moment, respectively.
Pd is calculated as a function of the relative reinforcement ratio:
f
(6.7)
<1.0
ISIS Canada Design Manual (2007) recommends the following equation to be used in design:
Vcr
/
I...+
1-0.5
M,
( W , )
(6.8)
Mota et al. (2006) examined a number of the suggested formulations for Ie and found that Eqn
6.8 yielded the most consistent and conservative results over the entire range of available
specimens reinforced with different types of FRP bars.
The second approach - the curvature approach - is adopted by CAN/CSA S806-02 wherein
the integration of curvature along the span to determine the curvature is recommended.
Knowing the value for curvature y/, the virtual work method can be used to calculate the
deflection of FRP-RC beams under any load level.
A, = | m y / d x
(6.9)

third segment (See
Figure 6.14).
173
Figure 6.14: Moment-curvature (M-K) relation for FRP reinforced concrete (CSA S806-02)
Table 6.7 displays the immediate deflection values rendered from the ACI 440.1R-06 and ISIS
Canada (2007) empirical equations as well as the curvature method adopted by CSA S806-02.
The ISIS Canada (2007) yields very good agreement followed by the CSA S806-02 curvature
method, whereas the modified Branson's equation adopted by ACI 440.1R-06 significantly
underestimates the immediate deflection values (a similar scenario to what took place in
Chapter 5).
6.4.2.2
Long-term
deflections
Figure 6.13 displays the maximum measured deflection at mid-span versus time for the four
tested beams. A common trend draws the trajectory of deflection-to-time for all four beams.
The obtained long-term to initial deflection percentage was 98, 95, 78 and 81 % for the beams
A B, C, and D, respectively (Table 6.8).
The two main approaches for computing the long-term deflection of FRP reinforced concrete
beams are: (i) the empirical A multiplier approach available in ACI 440.1R-06 and CAN/CSA
S806-02 and (ii) the age-adjusted effective modulus method in which the properties of the
section change with time and the deflection calculation is conducted using numerical methods.
174
NO NO CN CN
co ro ro ro
r-
in 00 oo
< m O Q
r-H
Qh
o
U
00
NO
O
1H
(i
-r&io)
S
Ho K
2
0.096
0.074
0.118
0.137
6.81
5.04
7.76
8.67
(mm)
(mm)
(mm)
<f ^
P. c?
3.14
2.21
3.27
3.46
A total
fi
o
o
CI
oryi
A long-term
<D
A initial
<D
T3
Total
deflection
EH
Long-term
deflection
o
o<u
Initial
deflection
3O
C
<kHD
0.97
0.91
0.67
0.82
0.84
3.58
2.58
3.03
4.27
0.179
0.132
0.204
0.228
"ao
<r
3.67
2.83
4.49
5.21
TO
T3
NO
O
ck
o
Tj"st
HH
U
<
4.06
2.93
3.37
4.74
long-term^ Îi
1.11
1.04
0.75
0.91
0.95
= /(1+50j
ACI
CSA
(440)
(S806)
(b)
(c)
1.16
0.86
0.70
0.78
0.70
1.16
0.73
0.65
1.10
0.66
1.10
0.65
Average
XC
S
pq
w
Oh
ro
0.44
1?
1U
"
N
2 - 10 M
o O1
S |
^ -S
^
O
NO
0
00 <]
<
t/3
u
CN
0.44
II
^
0.44
0.35
0.43
0.29
0.27
0.33
g & O
1.27
1.21
1.32
1.39
3.67
2.83
4.49
5.21
Average
Ci CN
^
-g c
xfS
en
C
O
3 GFRP-1
0.44
S>
& O
-S H 4-
Zr -r-T* TO
<U
H
s
Bottom
Rft
Bottom
reinforce
-ment
o^ft.
Si O
3 GFRP-1
3 GFRP-1
2 10M
2 10M
Top
reinforcement
Beam
Beam
ACI
(440)
(a)/(b)
1.22
1.12
1.12
1.02
1.12
AS75 Canda Design

Manual (2007)
(uirn)
oai/;'/p
CSA S806-02
(20 increments)
A=fydx
CSA
(S806)
(a)/(c)
0.84
0.82
0.71
0.74
0.78
o g
+3 P-i
*& "ri
&
W
o
?
NO NO CN CN
CO ro ro CO
00 oo
< CQ U Q
in
r-
The former approach indicates that ACI 440.1R-06 under-predicts the factor X by an average
of 12 %. The corresponding CAN/CSA S806-02 predictions, however, are conservative by an
average of 22 % (Table 6.8).
As regards the age-adjusted effective modulus method, the system of equations provided by
Ghali, Favre and Elbadry (2002) was used to code a user-friendly Fortran-2003 program. In
this program, beam curvature, deflections and strains are calculated based on a theoretical
approach. The flowchart of this program is shown in Figure 6.15; program source-code is
available in the appendix.
/ Print the
results
Figure 6.15: Flowchart for calculating the deformations of GFRP reinforced concrete beams.
176
The system of equations describing the long-term deflection calculation is explained below:
Mean curvature at time t is given by:

V(t) = (1 - & Wx +A Vx) + (2 + A y/2)
(6.10)
where A y/} is the change in curvature of the uncracked section; A y/2 is the change in curvature of the
cracked section;
is the coefficient for long-term curvature.
1-0.5!
6 =
M,
M,a
(6.11)
0
The total deflection at midspan at time t is given by
{t)L2
A, =
9.6
(6.12)
The time dependent change in strain at the centroid of the section and the time dependent
change in curvature due to creep and shrinkage are given by
Af 0 = 77 [(j)(t, t0 )e0 + sa (t, t0)]
y / - K
</>(t,t0)+cs(t,t0yf
(6.13)
(6.14)
where
h
1=
A'
1'
(6.15)
where (pit, t 0 ) i s the creep coefficient from time to to time t\ cs(t, t 0 ) is the shrinkage strain
from to to time t; yc is the distance between centroids of the effective concrete area and the
age-adjusted transformed area; Ac is the area of concrete in compression; A' is the area of the
age-adjusted transformed section; I c is the moment of inertia of concrete about the centroid
after age adjustment; / ' is the moment of inertia of the age adjusted transformed section using
the modular ratio 77'(t, t 0 ) in which r]'(t, t 0 ) = EGFRP/E'c(t,
elasticity, E'c(t, t0), is calculated as follows:
177
t 0 ) . The age adjusted modulus of
E'c(t,t0) =
(6.16)
where Ec is the modulus of elasticity of concrete at time to. In this study, the creep and
shrinkage coefficients are first calculated based on ACI Committee 209 and CEB-FIP Model
Code 1990 approaches. The same coefficients were later multiplied to a 0.55 correction factor
as proposed by Brown (1997). This multiplier accounts for the difference between the creep of
FRP-reinforced
concrete
and
steel
reinforced
concrete.
The theoretical
predictions
dramatically improved when Brown's, 0.55, multiplier was incorporated.
The ACI Committee 209 equations describing the creep and shrinkage coefficients of concrete
are:
0.6
<6 18)
'
where <p(t,tQ) is the creep coefficient at time t for age at loading to', scs(t,t')
is the free
shrinkage which occurs between t'at the end of the curing period and time t; (ptl is the ultimate
creep coefficient (~ 2.35); (_scs)u is the ultimate shrinkage strain at infinite time and <pcorr is
Brown's correction factor for creep of FRP reinforced concrete. Coefficients (pu and ( c s ) u
depend on atmospheric humidity, member size, slump of concrete, cement content, fine
aggregate percent, air content in percent and type of curing. On the other hand, the creep
coefficient according to CEB-FIP Model Code 1990 is given as:
(l)(t,tQ) = (t)RHp(fcrn)l3{t0)Pc(t-tQ)(Pcorr
(6.19)
where the factors <pRH>(3(.fcm)>fi(t0) and /? c (t t 0 ) take into account the corrections for the
atmospheric humidity, mean compressive strength of the concrete, time of loading and
duration of loading respectively. The strain due to shrinkage and swelling according to CEBFIP Model Code 1990 is calculated as follows:
ecs = ecs0Ps(t
~ t')
(6.20)
where c s 0 is a function of the mean compressive strength of concrete, type of cement and
relative humidity; f$s(t t') is a function of the loading duration.
178
14
months
13
12
(0 igina
11
fmf0'
0tea
10
CEB FIP
w
4>
5=
v
Q
CSA S80 5-02
8
&
3dett
ti
3l COd e l 390
(modified
- t
t-ACI Corn
Exp grlmi
1
6
5
AC 440. 1R-06
4
3
2
1
davs
0
15
30
45
60
75
90
105 120 135 150 165 180 195 210
Time (days/months)
Figure 6.16: Experimental versus empirical and numerical deflection results (Beam A)
0
months
d e H o d e 1990
ginaj)
a
o
"O
w
<u
C
<u
Q
P M o d e l C o d e 1990
(modified)
6-02
ACI C o r r m i t t e e 209
(modified)
days
0
15
30
45
60
75
90
105 120 135 150 165 180 195 210
Time (days/months)
Figure 6.17: Experimental versus empirical and numerical deflection results (Beam B)
179
months
Mo<jel Code 1990
F l P ^ o d e 1990
modified)
Experimental (C)
"C
<y
G
davs
0
15
30
45
60
75
90 105 120 135 150 165 180 195 210
Time (days/months)
Figure 6.18: Experimental versus empirical and numerical deflection results (Beam C)
months
^ C^B-FIP Mqdel Code :

t
(original)
CEB-FIP M o d e l t o d e 1990
(modified)
CSA S805-02
AQ Committee :!09 (modified)
s
'C
o>
C
<u
G
days
0
15
30
45
60
75
90 105 120 135 150 165 180 195 210
Time (days/months)
Figure 6.19: Experimental versus empirical and numerical deflection results (Beam D)
180
Figure 6.16 to Figure 6.19 display the measured deflection-versus-time values for the four
beams. In each graph, five means of calculating the initial and long term deflection are
displayed along with the obtained experimental results. The five calculating means are
described in the points below:
(ii)
ACI 440.1R-06: The long-term deflection is calculated according to the X

multiplier values made available by ACI 440.1R-06 based on the actual
measured initial deflection values.
(iii)
CAN/CSA S806-02: The long-term deflection is calculated according to the

X multiplier
values made available by CAN/CSA
S806-02
(more
conservative than ACI 440.1R-06) based on the actual measured initial

deflection values.
(iv)
CEB-FIP Model Code 1990 (Original): Initial and long term deflection
values are calculated using the bi-linear (mean curvatures method); the
creep coefficient, <p(t, t 0 ), is calculated according to CEB-FIP Model Code
1990 for conventional steel reinforced concrete.
(v)
CEB-FIP Model Code 1990 (Modified): Same as the above yet a creep
correction factor, (<pfrp 0.55), is introduced.
(vi)
ACI Committee 209 (Modified): Initial and long term deflection values are
calculated using the bi-linear (mean curvatures method); the creep
coefficient,0(t, t 0 ), is calculated according to ACI Committee 209 with the
4>frp correction factor.
By comparing the results of the empirical and numerical calculation methods above, it is
obvious that the modified versions of CEB-FIP Model Code 1990 and ACI Committee 209
yield the best initial and long-term deflection predictions. A very reasonable degree of
accuracy is exhibited despite the inherently complex nature of concrete creep. In general, -FIP
Model Code 1990 is more conservative than ACI Committee 209. As regards ACI 440.1R-06
and CAN/CSA S806-02, the empirical (multiplier) methods for predicting long-term
deflection gave reasonable results serving as lower bound and upper bound values for the
measured deflection results, respectively (See Table 6.8).
181
6.4.3
Once load was applied, all beams cracked instantaneously. Along the six-month duration, the
maximum crack widths, in the midspan region, were measured for all beams. Regular
monitoring, of the cracks, took place using a high precision (jeweller's) microscope. Figure
6.20 illustrates the maximum crack width evolution for all beams; it is evident that the rate of
crack-width-increase tapered down after 120 days. Beams comprising bottom reinforcement
equal to 1.5 pbalanced exhibited less cracks (that also propagated with a lesser rate) than the
beams with the balanced cross section. It is also evident that the pair of beams with higher
concrete compressive strength (36 MPa) exhibited tighter cracks than the pair of 32 MPa
concrete compressive strength.
15
30
45
60
75
90
105 120 135 150 165 180 195 210
davs
0.9
0.8
07
(U.46-1t.bb)
0.6
1.60] -*-Beam C
1.45] --Beam A
(0.35-(
A
-x-Beam B
--Beam D
4J (.33;
Cm
0.5
0.4
Xy
0.3
m-m --
SSSSfci
J2
0.2
0.1
hours
0
360
720 1080 1440 1800 2160 2520 2880 3240 3600 3960 4320 4680 5040
Time under sustained load (days/hours)

Figure 6.20: Maximum crack width evolution for beams A, B, C and D
The crack patterns for FRP-reinforced beams were found to be similar to steel reinforced
beams. Figure 6.21 to Figure 6.24 display the crack pattern for all four beams at milestones of
1 month, 3 months and 6 months.
182
\
;
1 month
/
^ ^
, (,
i
I
^ ^
3 months
^ ^
6 months
4282 mm
Figure 6.21: Crack propagation pattern after 1 month, 3 months and 6 months (Beam A)
Figure 6.22: Crack propagation pattern after 1 month, 3 months and 6 months (Beam B)
183
> I I 1 H
<
1 month
i l l
) )
3 months
i i f t
6 months
4282 m m
Figure 6.23: Crack propagation pattern after 1 month, 3 months and 6 months (Beam C)
( J J
1 month
1 r-
6 months
4282 m m
Figure 6.24: Crack propagation pattern after 1 month, 3 months and 6 months (Beam D)
Table 6.9 indicates the average initial crack width at midspan for each beam; the crack-width
magnitude after 6 months; the percentage increase and the ratio of crack-width after 6 months
to the limits specified by ACI 440.1R-06 and CSA S806-02 (0.028 in ~ 0.7 mm for FRP-RC
beams for internal exposure). Most cracks occurred during initial loading; the number of
flexural cracks increased over time from just a few (five to seven) to as many as nine to
184
twelve after 6 months (See Figure 6.21 to Figure 6.24). The crack width growth, after six
months of sustained loading is illustrated in Figure 6.25.
Beams
Beam
Beam
Beam
Beam
A
B
C
D
Top
reinforcement
Bottom
reinforce
-ment
3 GFRP-1
3 GFRP-1
2 10M
2 10M
5
8
8
5
f'c
36
36
32
32
Crack
width
after 6
months
(mm)
0.6
0.33
0.45
0.65
Initial
Crack
width
(mm)
0.4
0.24
0.35
0.45
0.25
Percentage
of crack
width
increase
Ratio (@ 6
months/limit)
50.0
37.5
28.6
44.4
0.86
0.47
0.64
0.93
1.55
Beam A
Beam B
Beam C
Beam D
Beam A
Beam B
Beam C
Beam D
Figure 6.25: Crack width growth over 6 months of sustained loading
In a similar manner to Chapter 5, the two commonly used methods for predicting the
instantaneous flexural - most probable maximum - crack width of FRP-RC beams are applied
to the beams at hand (Frosch's (1999) equation and the Gergely and Lutz (1968) expression).
The former equation, adopted by ACI 440.1R-06 and CAN/CSA S6-06, is expressed as:
w = 2J3kbsr,d2c +
'X2
(6.21)
In this equation, /? is the ratio of distance from neutral axis to extreme tensile fibre and
distance from neutral axis to centre of the tensile reinforcement = (h - c)/(d - c); where h is
185
reinforcement; sr is the reinforcement strain at service; dc is the concrete cover from extreme
reinforcing bars within outermost layer. The coefficient kb accounts for the degree of bond
between the FRP bar and the surrounding concrete. According to ACI 440.1R-06, kb - 1.0 for
deformed steel bar reinforcement; < 1 . 0 for superior bond relative to deformed steel bars; >
with a mean of 1.10.
w = 2.2fikbsr
(dcA)
1/3
(6.22)
(6.23)
beams, kt is taken as 2.0 and 1.7 for FRP reinforced beams exhibiting sustained load for a oneyear duration in the effort of Gross et al. (2009) and Chapter 5 of the current study,
respectively. The kt multiplier can also be applied to the Gergely-Lutz equation, yielding the
following equation:
w=
1/3
2.2/3ktkbsr{dcA)
(6.24)
Confirming the kb values obtained from the previous chapter, it was found that the values of
1.2 and 1.0 when attributed to the kb bond coefficient for equations 6.23 and 6.24,
respectively, good agreement is obtained with the measured crack width values (See Figure
6.26). Furthermore, the optimal kt value, for both models, was calculated and found to be 1.4
(Figure 6.27).
Frosch (1999) (kb = 1.2)
Gergely-Lutz (1968) (kb = 1.0)
0.8
0.8
0.7
0.7
0.6
0.6
-c
0.5
- - X
0.5
Beam A
s3
X Beam B
04
S
a
s
0.4
Beam D
cu
0.2
0.2
a
s
o
0 1
0.1
Beam A
X Beam B
0.3
TJ
W'
A Beam C
0.3
A Beam C
Beam D
Beam A
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.1
0.7
0.2
0.3
0.4
0.5
0.6
0.7
Figure 6.26: Measured versus calculated initial crack widths
Gergely-Lutz (1968) (kb = 1.0; k, = 1.4)
Frosch (1999) (kb = 1.2; k, = 1.4)

0.8
JS
J=
-M
5
u
cs
0.7
w .
0.6
Beam A
X Beam B
V,
g
u
g
3
A Beam C
K
|
Beam D
<u
w
73
1
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
i j
A Beam C
0u 30
'
"3
.X
fl4
X Beam B
0.1
0.2
0 2
0.1-
0.8
0.3
0.4
0.5
0.6
0.7
0.8
0.0
Figure 6.27: Measured versus calculated crack widths after 6 months of sustained load
187
Beam D
6.5
Conclusions
Four full-scale concrete beams of dimensions (215 mm x 400 mm x 4282 mm) were subjected
to uniform distributed block load on a span of 3810 mm. All beams contained bottom GFRP
reinforcing bars of 9.5 mm diameter; 2 reinforcement ratios were used (pbaianced and 1.5
pbalanced);
the top reinforcement used was either 2 No. 10M or 3 GFRP of 9.5 mm diameter; 2
concrete batches were used / c '= 36 MPa and 32 MPa (one batch per beam-pair); all beams
were tested in standard laboratory conditions. Based on the obtained measurements and
analysis, the following conclusions can be drawn:
1-
The rate of increase of strain for GFRP reinforcement and compression concrete beams
typically starts high at the initial loading period (60 to 90 days approx) and tapers down
asymptotically with time, under constant uniform sustained load. The gradual strain
increase in the bottom reinforcement is mainly due to the gradual change in curvature
rather than creep strain.
2-
The immediate deflections were predicted using three methods: the ACI 440.1R-06
empirical equation, the CSA S806-02 curvature method and the ISIS Canada (2007)
empirical equation. On average, all methods underestimated the measured deflection
values by 67 %, 16 % and 5 %, respectively. Thus, the ISIS Canada (2007) empirical
equation rendered the best (most conservative) results.
3-
Despite having compression steel, the pair of beams with higher concrete compressive
strength (by 12 % more than the other pair) exhibited less initial deflection than their
counterparts
with
compression
steel
reinforcement
(a
difference
of
22
approximately). Thus, high strength concrete is a solution for minimizing the deflection
of FRP reinforced concrete beams.
4-
For beams A,B and C, the overall initial and long-term deflection predictions rendered
by CEB-FIP Model Code 1990 and ACI Committee 209 are outstanding when
compared to experimental values, given that the <pcorr ( = 0.55) correction factor for the
creep coefficient (pit,to) is incorporated in the numerical model. The overall deflection
predictions for beam D over-estimated the experimental measurements by 40 % to 30 %,
roughly, for the two methods respectively.
188
5-
The empirical 2 -multiplier method adopted by ACI 440.1R-06 and CAN/CSA S806-02
gives good long-term deflection results; serving as lower bound and upper bound
curves, respectively. The X -multiplier method is convenient and less tedious in
application than the former finite-difference based models.
6-
The crack width prediction equations adopted by ACI 440.1R-06 and CAN/CSA S6-06,
on one hand, and by the ISIS Canada Design Manual (2007) on the other hand, yield
satisfactory results when the kb bond-coefficient factor is 1.2 and 1.0 respectively. For
both equations the time-dependent kt multiplier is deduced as 1.4, after six months. This
indicates good agreement with the bond coefficient calculations of the previous chapter
as well as earlier efforts in the same research venue.
189
Chapter 7: Conclusions and Recommendations
CHAPTER 7
CONCLUSIONS AND RECOMMENDATIONS
7.1
Summary
This PhD program investigates a variety of areas that deal with the long-term (creep)
behaviour of FRP bars as well as the long-term (creep) performance of concrete beams
reinforced with FRP bars. The study, as a whole, aims to further the very limited knowledge,
to-date regards the long-term behaviour of FRP-RC beams.
At the level of the reinforcing material (FRP), two experimental studies were conducted on six
different types of GFRP bars (a total of 99 bars). In the first study, the creep behaviour and
susceptibility to creep rupture of two types of GFRP bars (9.5 mm in diameter) were
addressed. Different levels of sustained axial load (15 %, 30 %, 45 % and 60 %fu,ave) were
applied to 37 GFRP samples. In the second study, the creep behaviour of six different types of
GFRP bars (52 bars of three different manufacturers having diameters ranging from 9.5 mm to
15.9 mm) was studied under two levels of allowable service load (15 % and 30 %
fUMve)
according to the currently available North American standards (CAN/CSA S806-02, CSA-S6Addendum (2009) and ACI 440.1R-06). Both studies adopted standard laboratory conditions
(ambient temperature of 23 1C and relative humidity of 50 2 %). The test duration
amounted to 10000 hours (417 days) wherein regular monitoring of creep strain evolution
took place and the creep coefficient of GFRP bars was calculated for all sustained load levels.
This was followed by residual tensile tests and microstructural analysis.
As regards the long-term performance of concrete beams comprising FRP bottom

reinforcement, a total of 24 beams were studied in standard laboratory conditions; under
sustained load; for lengthy time durations. The beams are classified into two categories: (i) 20
beams under third-point load for a one-year duration and (ii) 4 beams under uniform
distributed load for six months. For the former set, the long term performance of twenty
concrete beams (ten pairs) comprising GFRP, CFRP, and steel reinforcing bars was studied in
terms of deflection, strain increase and crack width evolution. The beams of dimensions 100
190
mm x 150 mm x 1800 mm were installed under third-point concentrated load, for a period
exceeding one year. Exhibiting a maximum applied moment of 25 % of their nominal moment
capacity, Mn, all beams were regularly monitored for the one-year duration. Theoretical
predictions for immediate deflection were calculated and compared to the obtained
experimental results. The long-term to immediate deflection ratio, k, was calculated for all
beams and compared to CAN/CSA S806-02 and ACI 440.1R-06 predictions.
Regarding the set of GFRP reinforced beams: the beams were of dimensions (215 mm x 400
mm x 4282 mm) subjected to uniform distributed load for a period of six months. The main
study parameters were (i) bottom reinforcement ratio and (ii) type of upper/compression
reinforcement (GFRP and/or steel). The maximum applied moment ranged from 15 to 21 % of
the nominal moment capacity for the beams. Numerical modelling took place using a
predict the long-term deflection of the beams. In this numerical model, the creep and
shrinkage coefficients were calculated using the ACI Committee 209 recommendations (1992)
as well as the CEB-FIP Model Code (1990). Furthermore, two versions of the empirical
multiplier method, adopted by ACI 440.1 R-06 and CAN/CSA S806-02, were applied. The
outcome of the numerical and the empirical models were compared to the experimental
values.
7.2
Conclusions
The main findings from the experimental work and analysis, within this study, can be
summarized as follows:
7.2.1
Long-term Performance of GFRP Bars under Axial Sustained Load
1- For GFRP-1 and GFRP-2 bars, no significant strain increase was found for sustained
load levels 15 % and 30
%fu,ave
after 10000 hours. The upper bound strain increase at
15 % fu.ave was 8.8 % and 2.5 % of the initial applied strain, frPt0, for the two bar-types,
respectively. The corresponding values at 3.0 %
fu,ave
were 3.8 % and 5.5 % ejrpo,
respectively. At 45 % fu, ave, the upper-bound creep strain percentage reached 7.6 % and
191
12 % jrP,o , respectively. However, at 60 % fu,ave, creep rupture occurred at different

time durations; 2964 hours for GFRP-1 and 13.8 to 231 hours for GFRP-2.
2 Microstructural analysis shows no microcracks indicating no signs of degradation for
the GFRP bars tested at levels inferior to 60 %
fUiave.
Degradation of bond was
exhibited for GFRP-1 at the latter sustained load level, after a 10000 hour test duration.
3 The residual tensile strength, for all samples of GFRP-1 and GFRP-2 that survived the
10000 hour duration, was found barely changed (the loss ranged from 0 to 4.3 % of the
ultimate tensile stress fu,ave)- As regards the modulus of elasticity, there was no loss for
all samples. It is also evident that there is no consistent relationship between the
sustained load level and the loss in tensile strength.
4
For the six commercial GFRP bar-types exhibiting a 10000 hour duration under
service load (15 % and 30
% fu,ave),
the strain increase was high at the first 80 to 120
hours then tapered down asymptotically with time. For bars that exhibited 15 %
fUiave
the upper-bound (maximum) values for accumulated creep strain after 10000 hours
were 8.8 %, 2.5 %, 4.1 %, 3.2 %, 8.3 % and 15.7 % of the initial strain e M 0 , for bartypes GFRP-1, GFRP-2, GFRP-3, GFRP-4, GFRP-5 and GFRP-6, respectively.
Similarly, the upper-bound values for 25-30 % fuave
were 3.8 %, 11.8 %, 12.0 %, 5.6
%, 8.6 % and 11.8 % for the same bar-types, respectively. In general, bars of bigger
diameters exhibit greater creep strain values; possibly due to the curing factor which is
better for bars of small diameter.
5 For the latter six commercial bars, the residual tensile strength and modulus of
elasticity, for all samples that survived the 10000 hour duration, were found barely
changed (almost retaining their full strength). The loss percentage ranged from 0-5.4 %
fu,ave and 0-8
EF>AVE
for tensile strength and Young's modulus respectively. In both
cases, the loss was less than the standard deviation yielded by mechanical property
testing. Furthermore, microstructural analysis showed no sign of degradation after the
samples were dismantled.
192
7.2.2
Long-term Performance of Third-Point Loaded FRP-RC Beams After One Year

of Continuous loading
Twenty (14 GFRP-; 4 CFRP- and 2 steel-) reinforced concrete beams were tested in thirdpoint sustained load set-ups for one-year within standard laboratory conditions. The
dimensions of the beams were 100 x 150 x 1800 mm. Along this duration, the long-term
performance of all beams was monitored as of (i) strain variation for bottom reinforcement as
well as top concrete fibre; (ii) long-term deflection and (iii) crack width. Based on the
obtained data and conducted analysis, the following conclusions were drawn:
The rate of increase in strains and deflections is high in the initial periods of loading
and reduces asymptotically with time under sustained loading. The strain increase in
GFRP and CFRP reinforcement after one year ranges between 4.2 % and 24.6 % of
the initial strain, mainly due to the gradual increase in beam curvature rather than
creep strain.
On comparing FRP and steel reinforced beams of approximately equal nominal

moment capacity, the initial deflection of FRP reinforced beams was found, on
average, 3.3 times that of the steel reinforced beams. As regards long term deflection,
the long-term to initial deflection percentage is 41 % and 94 % on average for FRPRC beams and steel reinforced beams, after a one year period, respectively.
The creep coefficient of FRP reinforced concrete 4>(t,to) under flexural loading was
found to be 0.4 times that of steel reinforced concrete. Thus, a correction factor (pcorr
of 0.4 should be introduced when calculating the former value. The 0.4 value falls in
the (0.36 to 0.73) range proposed by Brown (1997).
Three methods were used to calculate the immediate deflection of the FRP and steel
reinforced concrete beams. On average, the ACI 440.1R-06 empirical method
significantly underestimated the measured FRP-RC deflection results (by 67 %). On
the other hand the CAN/CSA S806-02 curvature method was inferior by 10 % and
the ISIS Canada Design Manual (2007) overestimated FRP-RC immediate deflection,
on average, by 11 %. As regards the steel reinforced concrete the calculations
underestimated by 2 %, overestimated by 14 % and overestimated by 45 % for the
three methods, respectively.
193
Though favourable, the long term deflection multiplier method (A), indicated in ACI
440.1 R-06 and CAN/CSA S806-02, significantly over-predicts the long-term
deflection values of the FRP reinforced concrete beams. The long-term to initial
deflection ratio, on average, was underestimated by 51 % and 70 %, respectively.
This fact, however, cannot be generalized due to the missing parameters that are
typically available in full-scale FRP reinforced concrete beams.
After the one year duration, the crack width for all beams was yet below the 0.7 mm
limit indicated by ACI 440.1 R-06. The crack-width-increase percentage varied
according to the reinforcement ratio and type of reinforcement.
The crack width prediction equations adopted by ACI 440.1 R-06 and CAN/CSA S606, on the one hand, and by the ISIS Canada Design Manual (2007) on the other
hand, yield satisfactory results when the fa bond-coefficient factor is 1.2 and 1.0
respectively. From the obtained data, the time-dependent kt multiplier is deduced as
1.7 and 1.5 for both models, respectively.
For the crack width prediction equation (Frosch 1999) adopted by ACI 440.1 R-06
and CAN/CSA S6-06, the fa bond-coefficient was calculated as 1.21, 0.66, 0.74, 1.1,
0.92, 1.24 and 1.99 for GFRP-1, GFRP-2, GFRP-3, GFRP-4, GFRP-5, GFRP-6,
CFRP-1 and CFRP-2,respectively.
7.2.3
Long-term Performance of Full-Scale FRP Reinforced Beams under Sustained

Uniform Distributed Load for a Six Month Duration
Four full-scale concrete beams of dimensions (215 mm x 400 mm x 4282 mm) were subjected
to uniform distributed block load on a span of 3810 mm. All beams contained bottom GFRP
reinforcing bars of 9.5 mm diameter; 2 reinforcement ratios were used (p/i, and 1.5 pji); the top
reinforcement used was either 2 No. 10M or 3 GFRP of 9.5 mm diameter; 2 concrete batches
were used / c '= 36 MPa and 32 MPa (one batch per beam-pair); all beams were tested in
standard laboratory conditions. Based on the obtained measurements and analysis, the
following conclusions were drawn:
194
The rate of increase of strain for GFRP reinforcement and compression concrete beams
typically starts high at the initial loading period (60 to 90 days approx) and tapers
down asymptotically with time, under constant uniform sustained load. The gradual
strain increase in the bottom reinforcement is mainly due to the gradual change in
curvature rather than creep strain.
The immediate deflections were predicted using three methods: the ACI 440.1R-06
empirical equation, the CSA S806-02 curvature method and the ISIS Canada (2007)
empirical equation. On average, all methods underestimated the measured deflection
values by 67 %, 16 % and 5 %, respectively. Thus, the ISIS Canada (2007) empirical
equation rendered the best (most conservative) results.
The pair of beams with higher concrete compressive strength (by 12 % more than the
other pair) exhibited less initial deflection than their counterparts with compression
steel reinforcement (a deflection difference of 22 % approximately). Thus, high
strength concrete is a solution for minimizing the deflection of FRP reinforced
concrete beams.
For three of the four beams (A,B and C) the overall initial and long-term deflection
predictions rendered by CEB-FIP Model Code 1990 and ACI Committee 209 were in
very good agreement with the experimental values, given that the (j)corr ( = 0.55)
correction factor for the creep coefficient <p(t, to) was incorporated in the numerical
model. The overall deflection predictions for the fourth beam (D) over-estimated the
experimental measurements by 40 % and 30 %, roughly, for the two methods
respectively.
The empirical X -multiplier method adopted by ACI 440.1R-06 and CAN/CSA S80602 gives good long-term deflection results; serving as lower bound and upper bound
curves, respectively. The X -multiplier method is convenient and less tedious in
application than the former finite-difference based models.
The crack width prediction equations adopted by ACI 440.1R-06 and CAN/CSA S606, on the one hand, and by the ISIS Canada Design Manual (2007) on the other hand,
yield satisfactory results when the kb bond-coefficient factor is 1.2 and 1.0
respectively. For both equations the time-dependent kt multiplier is deduced as 1.4,
after six months.
195
7.3
Recommendations for Future Work
Despite the extensiveness of this study, the field of long-term performance of FRP reinforced
concrete remains lacking for information content. In light of the conducted literature review,
experimental work and relevant analysis, the following recommendations for future work are
proposed:
1- The sustained load frames used for the creep testing of FRP bars at the University of
Sherbrooke need further enhancement, should FRP creep rupture tests be conducted as
per the B.8 test of ACI 440.3R-04 and Annex J of CAN/CSA S806-02. The available
frames are currently of a maximum capacity barely enough to apply a sustained load
level equivalent to 60
%fu,ave
of a typical 9.5 mm GFRP bar; the maximum capacity of
the frames in the meantime is 36 kN and the sought capacity is 60 kN (i.e. a forty
percent increase). Higher capacity is necessary since, at least, five different sustained
load levels and their corresponding creep rupture time form the regression line that
yields the sought 1 million hour creep-rupture load.
2- For further knowledge regarding the impact of adverse environments on the creep
performance and creep rupture stress limits of FRP bars, FRP sustained load tests
should be conducted whilst FRP bars are surrounded by field-like conditions (for
example, saturated concrete). This will also allow calculating the partial reduction
factors that relate to particular adverse environments, when calculating FRP creep
rupture stress limits.
3- More experimental work is needed regarding the long term behaviour of full-scale
FRP reinforced concrete beams exhibiting environmental adversities such as freezethaw cycles and elevated temperatures.
4- More experimental work is needed regarding the long term behaviour of full-scale
FRP-RC beams using high strength concrete to gain insight into its impact on
ameliorating long-term deflection results.
5- It is recognized that the approach of using a simple multiplier is simplistic in nature;
however it is convenient for designers. It is recommended that long-term sustained
load tests be conducted on all commercial FRP bars and a database of long-term
deflection multipliers is made available for each bar type.
196
CHAPTER 8
CONCLUSIONS ET RECOMMANDATIONS
8.1
Resume
Ce programme de doctorat a etudie une variete de domaines qui traitent le comportement a

long terme (fluage) des barres de PRF, ainsi que la performance a long terme (fluage) des
poutres en beton arme de barres en PRF. L'etude, a l'ensemble, vise a faire progresser les
connaissances tres limitees a-jour en ce qui concerne ce sujet.
Au niveau du materiau de renforcement (PRF), deux etudes experimentales ont ete realisees
sur six differents types de barres de PRFV (un total de 99 barres). Dans la premiere etude, le
comportement au fluage et la susceptibilite a la rupture de fluage de deux types de barres en
PRFV (9,5 mm de diametre) ont ete abordes. Des niveaux differents de charge axiale soutenue
(15 %, 30 %, 45 % et 60 % fUAVe) ont ete appliques a 37 echantillons de PRFV. Dans la
seconde etude, le comportement au fluage de six differents types de barres en PRFV (52
barres de trois fabricants differents ayant des diametres allant de 9,5 mm a 15,9 mm) a ete
etudie sous deux niveaux de charge de service admissible (15 % et 30 %
fu,ave)
selon les
directives actuellement disponibles en Amerique du Nord (CAN/CSA S806-02, la norme

CSA-S6-Addendum (2009) et l'ACI 440.1R-06). Les deux etudes ont adopte des conditions de
laboratoire standard (temperature ambiante de 23 1 C et une humidite relative de 50 2
%). La duree de tous les essais a ete elevee a 10000 heures (417 jours) dans laquelle un suivi
regulier de 1'evolution de la deformation de fluage a eu lieu et le coefficient de fluage de
barres en PRFV a ete calcule pour tous les niveaux de charge soutenue. Apres cette periode,
des essais de traction residuelle et de l'analyse des microstructures ont ete effectues sur ces
echantillons.
Autrement, la performance a long terme de 24 poutres renforcees (22 en PRF et 2 en acier) a

ete etudiee sous les conditions standards du laboratoire. Ces poutres ont ete soumises a une
charge soutenue pour une longue duree. Les poutres sont classees en deux categories: (i) 20
poutres avec des charges aux tiers de la portee pour une duree d'un an et (ii) 4 poutres sous
197
une charge repartie uniformement pour six mois. Pour le premier ensemble, 14 poutres
renforcees en PRFV, 4 poutres renforcees en PRFC, et 2 en acier d'armature ont ete etudiees
en termes de la fleche a long terme, de l'augmentation de la deformation et de la largeur des
fissures. Les poutres ont dimensions de 100 mm x 150 mm x 1800 mm. Elles ont exposees a
une charge constante (soutenue) qui cause 25 % du moment nominal, Mn; toutes les poutres
ont ete regulierement suivies pendant toute la duree d'un an. Les predictions theoriques pour la
fleche immediate ont ete calculees et compares aux resultats experimentaux. Le rapport, A,
entre la fleche a long terme et la fleche immediate a ete calcule et compare aux valeurs
correspondantes des normes CAN/CSA S806-02 et ACI 440.1R-06.
En ce qui concerne l'ensemble des poutres renforcees en PRFV: elles etaient de dimensions
(215 mm x 400 mm x 4282 mm) soumis a une charge distribute uniformement pour une
periode de six mois. Les parametres principaux de l'etude ont ete (i) le taux d'armature de
traction et (ii) le type de renforcement de compression (PRFV et/ou acier). Le moment
maximal applique variait de 15 a 21 % de la capacite nominale du moment des poutres. La
modelisation numerique a eu lieu en utilisant le logiciel Fortran-2003 base sur la methode du
module effectif, ajuste en fonction de l'age (age adjusted effective modulus method), pour
predire la fleche a long-terme des poutres. Les coefficients de fluage ont ete calcules sur la
base des recommandations ACI Committee 209 (1992) et CEB-FIP Code model (1990). En
outre, deux versions de la methode empirique multiplicateur, adoptee par l'ACI 440.1 R-06 et
CAN/CSA S806-02, ont ete appliquees. Les resultats de la numerotation et les modeles
empiriques ont ete compares aux valeurs experimentales.
8.2
Conclusions
Les conclusions principales du travail experimental et de l'analyse, dans cette etude, peuvent
etre resumees comme suit:
8.2.1
1-
Performance a long terme de barres en PRFV sous charge axiale soutenue
Pour les barres PRFV-1 et PRFV-2, aucune augmentation significative de la

deformation a ete trouvee pour les niveaux de charge soutenue 15 % and 30 %
198
fUiClve
apres 10000 heures. La limite superieure de l'augmentation de la deformation a 15 %

fu.ave a ete 8.8 % et 2.5 % de la deformation initiale appliquee, ep-p.o, pour les deux
types de barres, respectivement. Les valeurs correspondant a 30 % fUiave
% et 5,5 % Sfrp.o, respectivement. A 45 % fu.ave,
etaient de 3,8
la limite superieure de la pourcentage
de la deformation de fluage a ete 7,6 % et 12 % Sfrp.o, respectivement. Cependant, a

60 % fu.ave, la rupture en fluage est survenue a des durees differentes, 2964 heures
pour PRFV-1 et de 13,8 a 231 heures pour PRFV-2.
2-
L'analyse des microstructures ne montre aucune microfissure indiquant aucun signe

de degradation de barres de PRFV teste a des niveaux inferieurs a 60
% fUMve-
La
degradation de 1'adherence a ete exposee pour PRFV-1 au niveau de la charge

soutenue dernier, apres une duree 10.000 heures de test.
3-
La resistance residuelle a la traction de tous les echantillons de PRFV-1 et PRFV-2

qui ont survecu a la duree 10000 heures, a ete trouvee a peine change (la perte varie
de 0 a 4,3 % de la contrainte de traction ultime fu,ave)-
En ce qui concerne le module
d'elasticite, il n'y avait pas de perte pour tous les echantillons. II est egalement
evident qu'il n'existe pas de relation systematique entre le niveau de charge soutenue
et la perte de resistance a la traction.
4-
Pour les six barres types commerciales en PRFV presentant une duree de 10000
heures sous la charge de service (15 % et 30 %
fu,ave),
l'augmentation de la
deformation a ete elevee aux premieres 80 a 120 heures, puis a diminue

asymptotiquement par rapport au temps. Pour les barres qui ont montre 15 % fu,ave de
la borne superieure (maximale) pour les valeurs de la deformation de fluage
accumulees apres 10000 heures ont ete de 8,8 %, 2,5 %, 4,1 %, 3,2 %, 8,3 % et 15,7
% de la deformation initiale frP,o pour les barres types PRFV-1, PRFV-2, PRFV-3,
PRFV-4, PRFV-5 et PRFV-6, respectivement. De meme, les valeurs de la borne
superieure de 25-30 % f , a v e ont ete 3,8 %, 11,8 %, 12,0 %, 5,6 %, 8,6 % et 11,8 %
pour la meme barre type, respectivement. En general, les barres de gros diametres
manifestent des valeurs elevees de deformation de fluage; ?a peut etre a cause du
facteur de cure qui est mieux pour les barres de petit diametre.
5-
Pour les six dernieres barres commerciales, la resistance residuelle a la traction et le

module d'elasticite, pour tous les echantillons qui ont survecu a la duree 10000
199
heures, ont ete a peine changes (elles ont presque conserve leur pleine resistance). Le
pourcentage de perte varie de 0 a 5,4 % fu,ave
et 0-8 % Efave en ce qui concerne la
resistance a la traction et le module d'elasticity respectivement. Dans les deux cas, la

perte a ete inferieure a l'ecart-type donne par des essais de proprietes mecaniques. En
outre, l'analyse microstructurale n'a presente aucun signe de degradation apres que
les echantillons ont ete demontes.
8.2.2
Performance a long terme des poutres renforcees en PRF avec des charges aux
tiers de la portee pour une duree d'un an sous charge soutenue
Vingt poutres en beton arme (14 PRFV-4; PRFC-et 2-acier) ont ete testees sous une charge
soutenue aux tiers de la portee pour une duree d'un an, sous des conditions standards de
laboratoire. Les dimensions des poutres ont ete de 100 x 150 x 1800 mm. Le long de cette
duree, la performance a long terme de toutes les poutres a ete suivie a partir de (i) la variation
de deformation de 1'armature inferieure ainsi que la surface superieure du beton (ii) la fleche a
long terme et (iii) la largeur des fissures. Base sur les resultats obtenus et l'analyse effectuee,
les conclusions suivantes ont ete tirees:
1- Le taux d'augmentation de la deformation et de la fleche a ete eleve aux periodes
initiales de chargement et a reduit asymptotiquement avec le temps, sous une charge
soutenue. L'augmentation de la fleche de l'armature en PRFV et en PRFC apres une
annee qui varie entre 4,2 % et 24,6 % de la fleche initiale est principalement due a
l'augmentation progressive de la courbure de poutre plutot que la deformation de
fluage.
2- En comparant les poutres renforcees en PRF et en acier ayant approximativement la
meme capacite nominate de moment, la fleche initiale des poutres renforcees en PRF a
ete, en moyenne, 3,3 fois celle des poutres renforcees en acier. En ce qui concerne la
fleche a long terme, le pourcentage de celle-ci par rapport a la fleche initiale a ete en
moyenne 41 % et 94 % pour les poutres renforcees en PRF et en acier, respectivement.
3- Le coefficient de fluage du beton renforce en PRF (f)(t,to) a ete 0,4 fois celui du beton
renforce en acier. Ainsi, un facteur de correction de 0,4 <pcorr devrait etre introduit lors
200
du calcul de la premiere valeur. La valeur 0,4 se situe dans la gamme (0,36 a 0,73)
proposee par Brown (1997).
4- Trois methodes ont ete utilisees pour calculer la fleche immediate des poutres en PRF
et en acier. En moyenne, la methode empirique d"ACI-06 440.1R sous-estime
considerablement le fleche des poutres en PRF mesurees (par 67 %). D'autre part, la
methode de courbure presentee par CAN/CSA S806-02 a ete inferieure de 10 % et le
manuel canadien de conception d'ISIS (2007) surestime la fleche immediate des
poutres en PRF par 11 % en moyenne. En ce qui concerne le beton renforce en acier, le
calcul sous-estime de 2 %, surestime de 14 % et surestime de 45 % pour les trois
methodes, respectivement.
5- Quoique favorable, la methode de multiplicateur de la fleche a long terme (X),
presentee dans l'ACI 440.1 R-06 et CAN / CSA S806-02, surestime significativement
les valeurs de la fleche a long terme des poutres renforcees en PRF. Le rapport entre la
fleche a long terme et celle immediate a ete sous-estime, en moyenne, de 51 % et 70
%, respectivement. Ce fait, cependant, ne peut pas etre generalise en raison des
parametres manquants qui sont generalement disponibles dans les poutres en beton
renforcees en PRF en vrai grandeur.
6- Apres la duree d'un an, la largeur des fissures de toutes les poutres a encore ete
inferieure a la limite de 0,7 mm indiquee par ACI 440.1R-06. Le pourcentage
d'augmentation de la largeur des fissures varie selon le ratio et le type d'armature.
7- Les equations d'estimation de la largeur des fissures adoptees par l'ACI 440.1R-06 et
CAN / CSA S6-06, d'une part, et par le Manuel canadien de conception d'ISIS (2007)
d'autre part, donnent des resultats satisfaisants lorsque le coefficient d'adherence kb est
1,2 et 1,0, respectivement. D'apres les resultats obtenus, le multiplicateur en fonction
du temps kt est deduit 1,7 et 1,5 pour les deux modeles, respectivement.
8- Quant a l'equation d'estimation de la largeur des fissures (Frosch 1999) adoptee par
l'ACI 440.1R-06 et C A N / C S A S6-06, le coefficient d'adherence kb a ete calcule 1,21,
0,66, 0,74, 1,1, 0,92, 1,24 et 1,99 pour les PRFV- 1, PRFV-2, PRFV-3, PRFV-4,
PRFV-5, PRFV-6, PRFC-1 et PRFC-2, respectivement.
201
8.2.3
Performance a long terme des poutres a pleine echelle renforcees en PRFV sous
charge soutenue pour une duree de six mois
Quatre poutres en pleine echelle de beton de dimensions (215 mm x 400 mm x 4282 mm) ont
ete soumises a une charge repartie uniforme de bloc sur une portee de 3810 mm. Toutes les
poutres figurant en bas des barres d'armature en PRFV de 9,5 mm de diametre, 2 taux
d'armature ont ete utilisees (pp, and 1.5 pjb); l'armature superieure utilise etait soit 2 No. 10M
ou 3 PRFV de diametre de 9,5 mm, 2 lots de beton ont ete utilises (un lot pour chaque paire de
poutres) = 36 MPa et 32 MPa, toutes les poutres ont ete testees dans des conditions de
laboratoire standard. Sur la base des mesures obtenues et de l'analyse, les conclusions
suivantes ont ete tirees:
1
Le taux d'augmentation de la deformation d'armature de PRFV et le beton de

compression des poutres commence generalement eleve a la periode de chargement
initial (60 a 90 jours environ) et diminue avec le temps asymptotiquement avec le
temps, sous charge constante soutenue uniforme. L'augmentation progressive de la
deformation d'armature inferieure est principalement due a revolution progressive de
la courbure de la poutre, plutot que la deformation de fluage.
Les fleches immediate ont ete predites en utilisant trois methodes: l'equation empirique .
de l'ACI 440.1R-06, la methode de la courbure de CSA S806-02 et l'equation
empirique d'lSIS Canada (2007). En moyenne, ces methodes sous-estiment les valeurs
de deflexion mesures de 67 %, 16 % et 5 %, respectivement. Ainsi, l'equation
empirique d'lSIS Canada (2007) a rendu la meilleure (les plus conservateurs) des
resultats.
La.paire de poutres de resistance a la compression de 36 MPa (de 12 % de plus que

1'autre paire de 32MPa) presentaient de deformation initiale mo ins de la paire avec
l'armature d'acier de compression (une difference de fleche de 22 % environ). Ainsi, le
beton a haute resistance est une solution pour reduire la fleche de poutres en beton
armees de PRF.
Pour trois des quatre faisceaux (A, B et C) les predictions de la deflexion initiale et a
long terme donnee par le CEB-FIP Model Code 1990 et ACI Committee 209 ont ete
en tres bon accord avec les valeurs experimentales lors que le facteur de correction
202
(pcorr (= 0.55) pour le coefficient de fluage (p(t,to) a ete constitute dans le modele
numerique. Les predictions de deflexion initiale et long terme pour la quatrieme poutre
(D) a sur-estime les mesures experimentales de 40 % et 30 %, environ, pour les deux
methodes respectivement.
5
La methode empirique, de multiplicateur de X, adopte par ACI 440.1 R-06 et CAN

CSA S806-02 donne de bons resultats de deflexion a long terme; siegeant en tant que
limite inferieure et superieure des courbes limites, respectivement. La methode de
multiplicateur de X est pratique et moins fastidieux a appliquer que les modeles de la
courbure.
Les equations, utilises pour predire la largeur des fissures adoptee par ACI 440.1 R-06
et CAN / CSA S6-06, d'une part, et par le Manuel de conception d'ISIS Canada (2007)
d'autre part, donnent des resultats satisfaisants lorsque les facteurs de coefficient
d'adherence de 1,2 et 1,0 respectivement. Pour les deux equations du kt, dependant du
temps, est deduite comme multiplicateur de 1,4, apres six mois.
8.3
Recommandations pour des travaux futurs
Malgre l'ampleur de cette etude, le domaine de la performance a long terme du beton renforce
de PRF demeure defaut pour le contenu de l'information. A la lumiere de la revue de
litterature
effectuee,
les
travaux
experimentaux
et
des
analyses
pertinentes,
les
recommandations suivantes pour les travaux futurs sont proposees:
1-
Les batis de charge soutenue utilises pour l'essai de fluage des barres en PRF a
l'Universite de Sherbrooke ont besoin d'etre renforces encore, si les essais de la
rupture de fluage de PRF sont effectues selon le test de B.8 de ACI 440.3R-04 et
l'annexe J de la norme CAN/CSA S806- 02. Les batis ont une capacite maximale a
peine suffisante pour appliquer une charge soutenue equivalant a 60 % fUiave
d'une
barre typique de PRFV de 9,5 mm; la capacite maximale des batis est dans l'intervalle
de 36 kN et la capacite demandee est de 60 kN ( une augmentation de quarante pour
cent).
203
2-
Pour approfondir la connaissance sur l'impact des environnements defavorables sur la

performance de fluage, des tests de charge soutenue doivent etre effectues sur les
barres de PRF quand elles sont entoures par des conditions sur le terrain, (par
exemple: satures de beton).
3-
Un travail experimental complementaire est necessaire en ce
1 qui concerne le
comportement a long terme des poutres pleine echelle en beton renforcees en PRF
presentant l'effet des adversites environnementales tels que les cycles de gel-degel et
des temperatures elevees.
4-
Un travail experimental complementaire a 1'echelle reelle est necessaire en ce qui

concerne le comportement a long terme des poutres en beton a haute resistance
renforcees en PRF, pour mieux comprendre son impact sur l'amelioration des
resultats de fleche a long terme.
5-
Malgre le fait que l'approche d'utiliser un simple multiplicateur est reconnue

simpliste dans la nature, il est pratique pour les concepteurs. C'est preconise
d'effectuer les tests a charge soutenue a long terme sur les barres en PRF disponibles
et commerciales; une base de donnees des multiplicateurs de fleche a long terme est
disponible pour chaque barre type.
204
References
REFERENCES
ACI Committee 209, (1992), "Prediction of Creep, Shrinkage and Temperature Effects in
Concrete Structures," ACI 209R-92, American Concrete Institute, Farmington Hills,
Mich., 47p.
ACI Committee 318, (2005), "Building Code Requirements for Reinforced Concrete and
Commentary," ACI 318-05/ACI 318R-05, American Concrete Institute, Farmington
Hills, Mich., 430p.
ACI Committee 318, (2008), "Building Code Requirements for Reinforced Concrete and
Commentary," ACI 318-08/ACI 318R-08, American Concrete Institute, Farmington
Hills, Mich., 473p.
ACI Committee 435, (2003), "Control of deflection in Concrete Structures," ACI 435.R-03,
American Concrete Institute, Farmington Hills, Mich., 40p.
ACI Committee 435, (2003), "Control of deflection in Concrete Structures (Appendix B),"
ACI 435.R-95 (Appendix B [2003]), American Concrete Institute, Farmington Hills,
Mich., 13p.
ACI Committee 440, (2003), "Guide for the Design and Construction of Concrete Reinforced
with FRP Bars," ACI 440.1R-03, American Concrete Institute, Farmington Hills, Mich.,
44p.
ACI Committee 440, (2004), "Guide Test Methods for Fibre Reinforced Polymers (FRPs) for
Reinforcing or Strengthening Concrete Structures," ACI 440.3R-04, American Concrete
Institute, Farmington Hills, Mich., 40p.
with FRP Bars," ACI 440.1 R-06, American Concrete Institute, Farmington Hills, Mich.,
42p.
with FRP Bars," ACI 440.1R-06, American Concrete Institute, Farmington Hills, Mich.,
44p.
ACI Committee 440, (2008a), "Specification for Construction with Fibre-Reinforced Polymer
Reinforcing Bars," ACI 440.5-08, American Concrete Institute, Farmington Hills, Mich.,
5p.
205
References
ACI Committee 440, (2008b), "Specifications for Carbon and Glass Fibre-Reinforced
Polymer Bar Materials for Concrete Reinforcement," ACI 440.6M-08, American
Concrete Institute, Farmington Hills, Mich., 6p.
ACI Committee 440, (2007), "Report on Fibre Reinforced Polymer (FRP) Reinforcement for
Concrete Structures," ACI 440.R-07, American Concrete Institute, Farmington Hills,
Mich., 228p.
Al-Salloum, A., Almusallam, H., (2007), "Creep effect on the behaviour of concrete beams
reinforced with GFRP bars subjected to different environments," Journal of Construction
and Building Materials, Vol. 21, Issue 7, July, pp. 1510-1519.
Arockiasamy, M., Amer, A., Shahawy, M. and Chidambaram, S., (1996), "Long-term
Behaviour of Concrete Beams Reinforced with CFRP Bars under Sustained Loads,"
Proceedings of the 2 nd International Conference on Advanced Composite Materials in
Bridges and Structures (ACMBS-II), Montreal, Quebec, Canada, August 11-14, pp. 673680
Ando, N., Matsukawa, H., Hattori, A., and Mashima, A., (1997), "Experimental Studies on the
Long-term Tensile Properties of FRP Tendons," Proceeding of the Third International
Symposium on Non-Metallic (FRP) Reinforcement for Concrete Structures (FRPRCS3), Sapporo, Japan, October 14-16, Vol. 2, pp. 203-210.
Arockiasamy. M.., Amer. A., and Shahawy, M., (1998), "Environmental and Long-term
Studies on CFRP Cables and CFRP Reinforced Concrete beams," Proceedings of the
First
International
Conference
on Durability
of Composites
for
Construction,
Sherbrooke, Quebec, Canada, August, pp.599-610

Arockiasamy, M., Chidambaram, S., Amer, A., and Shahawy, M., (2000), "Time-dependent
Deformations of Concrete Beams Reinforced with CFRP Bars," Composites Part B:
Engineering, Oxford, UK, Vol. 31, October, pp. 577-592
Bazant. Z. P., (1972), "Prediction of Concrete Creep Effects Using Age-Adjusted Effective
Modulus Method," American Concrete Institute (ACI) Journal, USA, Title No. 69-20,
pp. 212-217
Bazant, P., and Prasannan, S., (1989), "Solidification Theory for Concrete Creep I:
Formulation," ASCE Journal of Engineering Mechanics, Vol. 115(8), September, pp.
1691-1703.
206
References
Benmokrane, B. and El-Salakawy, E., (2002), "Durability of Fibre Reinforced Polymer (FRP)
Composites for Construction," Proceedings of the Second International Conference on
Durability
of
Fibre-Reinforced
Polymer
(FRP)
Composites
for
Construction
(CDCC'2002)," Montreal, Quebec, Canada, May 28-31, 715p.

Benmokrane, B., Chaallal, O., and Masmoudi, R., (1996), "Flexural Response of Concrete
Beams Reinforced with FRP Reinforcing Bars," ACI Structural Journal, Vol. 93, No. 1,
January-February, pp. 46-55.
Bogdanovic, B., (2002), "Deflection Calculations of FRP-Reinforced Concrete Beams,"
Technical Report, University of Manitoba, Manitoba, Canada, 23p.
Bouguerra, K., Youssef, T., and Benmokrane, B., (2007), "Durability of GFRP Internal
Reinforcement: Mechanical Properties after Moisture Absorption," Proceedings of the
Third International Conference on Durability and Field Applications of FRP Composites
for Construction (CDCC'2007)," Quebec City, Quebec, Canada, May 22-24, pp. 557566.
Branson, D. E., (1977), "Deformation of Concrete Structures," McGraw-Hill, New York,
USA, 546 p.
Brown, V., (1997), "Sustained Load Deflections in GFRP-Reinforced Concrete Beams," NonMetallic (FRP) Reinforcement for Concrete Structures (FRPRCS-3), Proceeding of the
3rd International Symposium on Nonmetallic (FRP) Reinforcement for Concrete
Structures, Sapporo, Japan, October 14-16, Vol. 2, pp. 495-502.
Brown, V., and Bartholomew, C., (1996), "Long-term Deflections of GFRP-Reinforced
Concrete Beams," Proceedings of the First International Conference on Composites in
Infrastructure (ICCI-96), Tuscon, Arizona, USA, January 15-17, pp. 389-400.
Bruegger, J. P., (1974), "Methods of Analysis of the Effects of Creep in Concrete Structures,"
Ph.D thesis, Department of Civil Engineering, Universty of Toronto, Toronto, Ontario.
Canadian Standard Association (CSA), (2004), "Design of Concrete Structures," CAN/CSAA23.3-04, Rexdale, Ontario, Canada, 240p.
Canadian Standard Association (CSA), (2006), "Canadian Highway Bridge Design Code,"
CAN/CSA S6-06, Rexdale, Ontario, Canada.
Canadian
Standard
Association
(CSA),
(2010),
"Specification
for
Fibre-Reinforced
Polymers," CAN/CSA S807-10, Rexdale, Ontario, Canada, Final Draft, 35p.
207
References
Canadian Standard Association (CSA), (2009), "Canadian Highway Bridge Design Code
(CHBDC)," CAN/CSA S6-06, Addendum, Rexdale, Ontario, Canada.
Canadian Standards Association (CSA), (2002), "Design and Construction of Building
Components with Fibre Reinforced Polymers," CAN/CSA S806-02, Rexdale, Ontario,
Canada, 177p.
Chiao, T. T., and Moore, R. L., (1971), "Stress-Rupture of S-Glass/Epoxy Multifilament
Strands," Journal of Composite Materials, Vol. 5, No. 1, pp. 2-11.
Comite Euro-International du Beton (CEB) - Federation Internationale de la Precontrainte
(FIP), (1990), "Model Code for Concrete Structures," CEB-FIP 1990, Thomas Telford
House, London, 437p.
Debaiky, A., Nkurunziza, G., Benmokrane, B., and Cousin, P., (2006), "Residual Tensile
Properties of GFRP Reinforcing Bars after Loading in Severe Environments," ASCE
Journal of Composites for Construction, September-October, pp. 370-380.
European Committee for Standardisation, (2002), "Eurocode 2: Design of Concrete
Structures-Part 1: General Rules and Rules for Buildings," European Standard, Brussels,
Belgium, November, 226p.
Favre, R., Charif, H., (1994), "Basic Model and Simplification Calculations of Deformations
According to the CEB-FIP Model Code 1990," ACI Structural Journal, Vol.91, No.2,
pp. 169-177.
Feldman, R. F., (1972), "Mechanism of Creep of Hardened Portland Cement Paste," Journal
of Cement and Concrete Research, Vol. 2, No. 5, pp. 509-520.
Feldman, R. F., and Sereda, P. J., (1968), "A Model for Hydrated Portland Cement Paste as
Deduced from Sorption-Length Change and Mechanical Properties," Journal of
Materials and Structures, Vol. 1, No. 6, pp. 509-520.
Federation Internationale de Beton (FIB), (2007), "FRP Reinforcement for Reinforced
Concrete Structures," FIB Bulletin 40, Federal Institute of Technology Lausanne
(EPFL), Switzerland, June, 160p.
Karbhari, V., Chin. J., Hunston, D., Benmokrane, B., Juska, T., Morgan, R.., Lesko, J.,
Sorathia, U. and Reynaud, D., (2003), "Durability Gap Analysis for Fibre-Reinforced
Polymer Composites in Civil Infrastructure," Journal of Composites for Construction,
August, pp. 238-247
208
References
Gere, J. M., and Timoshenko, S. P., (1990), "Mechanics of Materials," PWS Publishing
Company, Boston, USA, 807 p.
Gao, D.; Benmokrane, B. and Masmoudi, R., (1998), "A Calculating Method of Flexural
Properties of FRP-Reinforced Concrete Beam: Part 1: Crack Width and Deflection,"
Technical Report, Department of Civil Engineering, University of Sherbrooke,
Sherbrooke, Quebec, Canada, 24 p.
Gaona., F., (2003), "Characterization of Design Parameters for Fibre Reinforced Polymer
(FRP) Composite Reinforced Concrete Systems," PhD Thesis, Texas A & M University,
Texas, USA, December, 270 p.
Gerritse, A., (1998), "Assessment of Long-term Performance of FRP Bars in Concrete
Structures," Proceedings of the First International Conference (CDCC-98), Sherbrooke,
Quebec, Canada, pp. 285-296.
Greenwood, M., (2002), "Creep-Rupture Testing to Predict Long-Term Performance,"
Proceedings of the 2nd International Conference on the Durability of Fibre Reinforced
Polymer (FRP) Composites for Construction," Sherbrooke, Quebec, Canada, pp. 203212.
Ghali. A., Favre. R., and El-Badry, M., (2002), "Concrete Structures: Stresses and
Deformations," Third Edition - Chapman and Hall, London, England, 584p.
Ghali, A., and Neville, A., (2003), "Structural Analysis: A Unified Classical and Matrix
Approach," Fifth Edition - Spon Press, London and New York, 845p.
Gross, S.P., Yost, J.R. arid Stefanski, D.J., (2009), "Effect of Sustained Loads on Flexural
Crack Widths in Concrete Beams Reinforced with Internal FRP Reinforcement," ACI
Special
Publications:
Serviceability
of
Concrete
Members
Reinforced
with
Internal/External FRP Reinforcement, SP-264, American Concrete Institute, Farmington

Hills, Mich., Edited by C. E. Ospina, P. H. Bischoff and Tarek Alkhrdaji, October 2009,
pp. 13-32.
Gross, S., Yost, J. and Crawford, J., (2006), "Serviceability of High Strength Concrete Beams
with Internal FRP Reinforcement under Sustained Load," Proceedings of the Third
International Conference on FRP Composites in Civil Engineering (CICE 2006), Miami,
Florida, USA, December 13-15, pp.:203-206.
209
References
Gross. S., Yost. J., and Kevgas. G., (2003), "Time-Dependent Behaviour of Normal and High
Strength Concrete Beams Reinforced with GFRP Bars Under Sustained Load," High
Performance Materials in Bridges, American Society of Civil Engineers, pp. 451-462.
Hall, T. and Ghali, A., (1997), "Prediction of the Flexural Behaviour of Concrete Members
Reinforced with GFRP Bars," Society for the Advancement of Material and Process
Engineering, Vol. 42, pp. 298-310.
Hall, T. and Ghali, A., (2000), "Long-Term Deflection Prediction of Concrete Members
Reinforced with Glass Fibre Reinforced Polymer Bars," Canadian Journal of Civil
Engineering, Ottawa, Ontario, Canada, 27: 890-898.
Hansen, T. C., and Mattock, A. H., (1966), "Influence of Size and Shape of Member on the
Shrinkage and Creep of Concrete," ACI Journal Proceedings, Vol. 63, pp. 267-290.
Hughes Brothers Inc., (2007), "Fibreglass Rebar," Technical Datasheet, Nebraska, U.S, 4 p.
ISIS Canada, (2006), "Specifications for Product Certification on Fibre Reinforced Polymers
(FRPs) as Internal Reinforcement in Concrete Structures," Product Certification of FRP
Materials, the Canadian Network of Centres of Excellence on Intelligent Sensing for
Innovative Structures, ISIS Canada, University of Manitoba, Winnipeg, Manitoba,
Canada, 21 p.
ISIS Canada, (2007), "Reinforcing Concrete Structures with Fibre Reinforced Polymers,"
Design Manual No. 3 (ISIS-M03), The Canadian Network of Centres of Excellence on
Intelligent Sensing for Innovative Structures, ISIS Canada, University of Winnipeg,
Manitoba, 81 p.
ISIS Canada, (2006), "Durability of FRP Composites for Construction," ISIS Educational
Module 8, The Canadian Network of Centres of Excellence on Intelligent Sensing for
Innovative Structures, ISIS Canada, University of Winnipeg, Manitoba, 20p.
ISIS Canada, (2006), "Durability of Fibre Reinforced Polymers (FRP) in Civil Infrastructure,"
Durability Monograph, The Canadian Network of Centres of Excellence on Intelligent
Sensing for Innovative Structures, ISIS Canada, University of Winnipeg, Manitoba, pp.
3.1-3.27.
Italian National Research Council (CNR), (2006), "Guide for the Design and Construction of
Concrete Structures Reinforced with Fibre-Reinforced Polymer Bars," CNR-DT 203,
Rome, Italy, 35p.
210
References
Japanese Society of Civil Engineers (JSCE), (1997), "Recommendation for Design and
Construction of Concrete Structures Using Continuous Fibre Reinforcing Materials,"
Concrete Engineering Series 23, Edited by A. Machida, Tokyo, Japan, 325 p.
Kage, T., Masuda, Y., Tanano, Y. And Sato, K., (1995), "Long-term Deflections of
Continuous Fibre Reinforced Concrete Beams," Proceedings of the Second International
Symposium on Non-Metallic (FRP) Reinforcement for Concrete Structures (FRPRCS2), Ghent, Belgium, October 23-25, pp. 251-258.
Karbhari, V., Chin, J., Hunston, D., Benmokrane, B., Juska, T., Morgan, R., Lesko, J.,
Sorathia, U., and Reynaud, D., (2003), "Durability Gap Analysis for Fibre-Reinforced
Polymer Composites in Civil Infrastructure," Journal of Composites for Construction,
August 2003, pp. 238-247.
Laoubi, K., El-Salakawy, E., and Benmokrane, B., (2006), "Creep and Durability of SandCoated GFRP Bars in Concrete Elements under Freeze/Thaw Cycling and Sustained
Loads," Journal of Cement and Concrete Composites, Volume 28, pp. 869-878.
Laoubi, K., El-Salakawy, E., and Benmokrane, B., (2006), "Behaviour of Concrete Beams
Reinforced with GFRP Bars under Sustained Load and Freeze/Thaw Cycles,"
Proceedings of the 2nd International Conference on the Durability of Fibre Reinforced
Polymer (FRP) Composites for Construction, Sherbrooke, Quebec, Canada, pp. 453464.
L'Hermite, R. G., and Mamillan, M., (1968b), "Retrait et Fluage des Betons, Annales, Inst.
Techn. Du batiment et des Travaux Publics," Paris, France, pp. 1317-1337.
MacGregor, J. G., and Wight, J. K., (2005), "Reinforced Concrete: Mechanics and Design,"
Pearson Prentice-Hall, 1132 p.
Masmoudi, R., Nkurunziza, G., Benmokrane, B., and Cousin, P., (2003), "Durability of Glass
FRP Composite Bars for Concrete Structure Reinforcement under Tensile Sustained
Load in Wet and Alkaline Environments," Proceedings of the 31st Annual Conference
(Canadian Society for Civil Engineering), Moncton, Nouveau-Brunswick, Canada,
GCA-400: lOp.
Mehta, P. K., (1986), "Concrete: Structure, Properties, and Materials," Prentice-Hall, New
Jersey, USA, 450p.
211
References
Mota, C., Alminar, S., and Svecova, D., (2006), "Critical Review of Deflection Formulas for
FRP Reinforced Concrete," ASCE Journal of Composites for Construction, pp. 183-194.
Neville, A., (1996), "Properties of Concrete," Fourth Edition - John Wiley and Sons Inc., NJ,
USA, 844p.
Neville, A. M., and Brooks, J. J. (1991), "Concrete Technology," Prentice-Hall, New Jersey,
USA, 438p.
Neville, A. M., (1959), "Some Aspects of the Strength of Concrete," Civil Engineering and
Public Works Review, London, 54 (Part I, Oct. 1959, pp. 1153-1156; Part II, Nov. 1959,
pp. 1308-1310; Part III, Dec. 1959, pp. 1435-1439.), pp. 1153-1156.
Neville, A. M., (1960), "The Relation between Creep of Concrete and the Stress-Strength
Ratio," Applied Scientific Research, Section A, Springer Publishers, Amsterdam,
Netherlands, pp.285-92.
Neville, A. M., (1970), "Creep of Concrete: Plain, Reinforced, and Prestressed," NorthHolland Publishing Company, Amsterdam, Netherlands, 622p.
Neville, A. M., and Meyers, A. L., (1964), "Creep of Concrete; Influencing Factors and
Prediction," Symposium on Creep of Concrete, American Concrete Institute (ACI), SP9, Detroit, Michigan, USA, pp. 1 -31.
Ngab, A. S., Nilson, A. H., and Slate, F. O., (1980), "Behaviour of High-Strength Concrete
under Sustained Compressive Stress," Report No. 80-2, Department of Structural
Engineering, Cornell University, Ithaca, NY, USA, 49p.
Nilson, A. H., Darwin, D. and Charles, W. D., (2010), "Design of Concrete Structures," 14th
Edition, McGraw-Hill, Boston, USA, 795 p.
Nkurunziza, G., Benmokrane, B., Debaiky, A., and Masmoudi, R., (2005), "Effect of
Sustained Load and Environment on Long-Term Tensile Properties of Glass FibreReinforced Polymer Reinforcing Bars," ACI Structural Journal, July-August 2005, pp.
615-621.
Ospina. C. E. and Bakis. C. E., (2006), "Indirect Crack Control Procedure for FRP-Reinforced
Concrete Beams and One-Way Slabs," Third International Conference on FRP
Composites in Civil Engineering (CICE 2006), December 13-15, Miami, Florida, USA,
pp. 535-538.
212
References
Powers, T. C., (1968), "Mechanism of Shrinkage and Reversible Creep of Hardened Portland
Cement Paste," Proceeding of the International Conference on the Structure of Concrete,
London, England, pp. 319-344.
Pultrall Inc.. (2007), "V-Rod Composite Reinforcing Rods," Technical Datasheet, Thetford
Mines, Quebec. Canada, 7 p.
Rahman, A., Kingsley, C., Crimi, J., (1995), "Behaviour of FRP Grid Reinforcement for
Concrete under Sustained Load," Proceedings of the Second International Symposium
on Non-Metallic (FRP) Reinforcement for Concrete Structures (FRPRCS-2), Ghent,
Belgium, October 23-25, pp. 90-99.
Russell, G. H., and Larson, S. C., (1989), "Thirteen Years of Deformations in Water Tower
Place," ACI Structural Journal, Farmington Hills, Michigan, USA, Vol. 86, No. 2, pp.
182-191.
Ruetz, W., (1968), "A Hypothesis for the Creep of Hardened Cement Paste and the Influence
of Simultaneous Shrinkage," Proceeding of the International Conference On the
structure of Concrete, London, England, pp. 365-387.
Scanlon,
A.,
(2009), "Deflection Control for FRP-Reinforced Concrete Members: An
Overview," ACI Special Publications: Serviceability of Concrete Members Reinforced

with Internal/External FRP Reinforcement, SP-264, American Concrete Institute,
Farmington Hills, Mich., Edited by C. E. Ospina, P. H. Bischoff and Tarek Alkhrdaji,
October 2009, pp. 1-12.
Seki. H., Sekijima. K., and Konno. T., (1997), "Test Method on Creep of Continuous Fibre
Reinforcing Materials," Proceedings of the Third International Symposium on NonMetallic (FRP) Reinforcement for Concrete Structures (FRPRCS-3), Sapporo, Japan,
October 14-16, Vol. 2, pp. 195-202.
Shoeck Bauteile GmbH, (2007), "Technical Information Schoeck Combar," Technical
Datasheet, Vimbucher Strabe 2, 76534 Baden-Baden, Germany, October, 32 p.
Smadi, M. M., Nilson, A. H., and Slate, F. O., (1982), "Time-Dependent Behaviour of HighStrength Concrete under High Sustained Compressive Stresses," Report 82-16,
Department of Structural Engineering, Cornell University, Ithaca, NY, USA, November
1982, 264 p.
Sturgeon, J. B., (1978), "Creep of fibre reinforced thermosetting resins," Creep of Engineering
213
References
Materials, C. D. Pomery, Editors, Mechanical Engineering Publishing Ltd., London,

UK, pp. 175-795.
Theriault, M., and Benmokrane, B., (1997), "Effects of FRP Reinforcement Ratio and
Concrete Compressive Strength on the Flexural Behaviour of Concrete Beams
Reinforced with FRP C-Bar rods," Technical Report No. 3, Universite de Sherbrooke,
Sherbrooke, Quebec, Canada.
Tamtsia, B. T., and Beaudoin, J. J., (2000), "Basic Creep of Hardened Cement Paste, A Reexamination of the Role of Water," Cement and Concrete Research, Elsevier, pp. 14651475.
Theriault, M., and Benmokrane, B., (1998), "Effects of FRP Reinforcement Ratio and
Concrete Compressive Strength on Flexural Behaviour of Concrete Beams," Journal of
Composites for Construction, ASCE Journal of Composites for Construction, Vol. 2,
No. 1, February, pp. 7-16.
Troxell, G. E., Raphael, J. M., and Davis, R. E., (1958), "Long-Time Creep and Shrinkage
Tests of Plain and Reinforced Concrete," ASTM Proceedings, Vol. 58, USA, pp. 11011120.
Vijay. P., and GangaRao, H., (1998), "Creep Behaviour of Concrete Beams Reinforced with
GFRP Bars," Proceedings of the First International Conference on Durability of
Composites for Construction (CDCC'98)," Sherbrooke, Quebec, Canada, August,
pp.661-667.
Wittmann, F. H., (1982), "Creep and Shrinkage in Concrete Structures," Wiley Series in
Numerical Methods in Engineering, New York, USA, 363 p.
Yamaguchi, T., Kato, Y., Nishimura, T., and Uomoto, T., (1997), "Creep Rupture of FRP rods
made of Aramid. Carbon, and Glass Fibres," Proceedings of the Third International
Symposium on Non-Metallic (FRP) Reinforcement for Concrete Structures (FRPRCS3), Sapporo, Japan, October 14-16, Vol. 2, pp. 179-186.
Youssef, T., El-Gamal, S., El-Salakawy, E., and Benmokrane, B., (2008), "Experimental
Results of Sustained Load (Creep) Tests on FRP Reinforcing Bars for Concrete
Structures," Proceedings of the 37th CSCE Annual Conference, Quebec City, Quebec,
Canada, June 10-13, (CD-ROM), lOp.
214
References
Youssef, T., El-Gamal, S. and Benmokrane, B., (2009), "Long-Term Deflection of GFRP
Reinforced Concrete Beams under Uniform Sustained Load," Proceedings of the 38th
CSCE Annual Conference, St. John's, Newfoundland-Labrador, Canada, May 27-30,
(CD-ROM), lOp.
Yost, J.; Gross, S.; and Dinehart, D., (2003), "Effective Moment of Inertia for GFRP
Reinforced Concrete Beams," ACI Structural Journal, Vol. 100, No. 6, Nov.-Dec., pp.
732-739.
Yunping, X., and Jennings, H. M., (1992), "Relationships between Microstructure and Creep
and Shrinkage of Cement Paste," Materials Science of Concrete, American Ceramic
Society, Westerville, Ohio, USA, pp. 37-63.
Zia, P., Leming, M. L., Ahmad, S. H and Washington, D. C., (1991), "High-Performance
Concrete: A State-of-the-Art Report. Strategic Highway Research Program," National
Research Council, Washington, USA, 251 p.
Zia, P., Ahmad, S. H., Leming, M. L., Schemmel, J. J., and Elliott, R. P., (1993e),
"Mechanical Behaviour of High Performance Concretes, Very High Strength Concrete,"
Strategic Highway Research Program, National Research Council, Washington, USA,
101 p.
215
APPENDIX
216
C
C
C
C
C
C
mai n.txt
1-THIS PROGRAM IS MEANT TO COMPUTE THE LONG TERM DEFORMATION OF
RC BEAMS REINFORCED WITH FRP BARS.
2-THE ANALYSIS IN THE CODE IS BASED ON NONLINEAR SECTION ANALYSIS
FOR BEAM CROSS SECTION AND THE CREEP EVOLUTION OF THE CONCRETE
3-THE ANALYSIS IN THE CODE IS BASED ON NONLINEAR SECTION ANALYSIS
FOR BEAM CROSS SECTION AND THE CREEP REVOLUTION OF THE CONCRETE
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
C
C
C
C
Q
TARIK YOUSSEF
CIVIL ENGINEERING DEPARfEMENT
FACULTY OF ENGINEERING
UNIVERSITY OF SHERBROOKE
PHD CANDIDATE
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
IMPLICIT DOUBLE PRECISION ( A-Z )

INTEGER I,J,t
COMMON /DATAl/
COMMON /DATA2/
COMMON /DATA3/
COMMON /DATA4/
b,h,1,dl,d2,w,Mext
AS,ES
Afrp, Efrp
fc, EC
PRINT *, "XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX"
PRINT *, "XXXXXX THIS . PROGRAM IS MEANT TO COMPUTE THE LONG TERM
CDEFORMATION XXXXXX"
PRINT *, "XXXXXX OF RC BEAMS REINFORCED WITH FRP BARS.
C THE ANALYSIS IN THE XXXXXX"
PRINT *, "XXXXXX CODE IS BASED ON NONLINEAR SECTION ANALYSIS
CFOR BEAM CROSS XXXXXX"
PRINT *, "XXXXXX
SECTION AND THE CREEP REVOLUTION OF
CTHE CONCRETE
XXXXXX"
PRINT *, "XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX"
PRINT *, ""
PRINT *, "XXXXXX
CALCULATE THE INTIAL STRAIN, CURVATURE AND
CDEFLECTION
XXXXXX"
1050
1055
open (9,FILE='DATA.txt 1 )
OPEN (8,FILE='OUT.txt')
WRITE (8,1050) DEF_0
FORMAT (4f25.10)
FORMAT (14)
PRINT *, "TIME
STRAIN"
DO 1=1,24
CALL INPUT
t=0
CALL STRAIN_0 (Cu TV_0,STRAIN_F0,DEF_0)
PRINT*, I, DEF_0*1000000
WRITE (*,1050) DEF_0

ENDDO
PRINT *', ""
PRINT *,
"XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX"
PRINT *, "XXXXXX
CALCULATE THE INTIAL STRAIN, CURVATURE AND

Page 1
mai n.txt
C
CDEFLECTION
XXXXXX"
C
C
C
PRINT *, "XXXXXX PRINT THE CREEP STRAIN, CURVATURE AND

CDEFLECTION VALUES XXXXXX"
PRINT *, "XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX"
C
C
C
C
PRINT *, ""
PRINT *, "XXXXXX
END OF THE PROGRAM
C
XXXXXX"
PRINT *,
C*I
C*I
C*I
END
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
XXXXX
XXXXXX
XXXXX
XXXXXX
C*I
C*I
XXXXX
XXXXX
C*I
"XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX"
END
MAIN PROGRAM
XXXXXX
XXXXXX
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
Page 2
STRAIN_0.txt
SUBROUTINE STRAIN_0 (Curv_0,STRAIN_F,DEF)
C
*****
it*****************************************************
C
{ THIS SUBROUTINE IS USED TO COMPUTE THE INSTANTANEOUS STRAIN, DEFLECTION,
AND CURVATURE }
C
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *

INTEGER 1,3
COMMON /DATAl/
COMMON /DATA2/
COMMON /DATA3/
COMMON /DATA4/
b,h,1,dl,d2,W,Mext
AS,ES
Afrp, Efrp
fc, EC, rohb
E=EC
C
e
c
c
c
c
C
CALL CG (E,YCG_0,IC_0,YCG,lC,STRAIN_C)
PRINT*, STRAIN_C
ft=0.6*DSQRT(fc)
Mc=ft*lc_0/YCG_0
xi1=1-(Mc/Mext)**2
Curvl=Mext/(Ec*lc_0)
Curv2=Mext/(Ec*lc)
Cu rv_0=(1-Xi1)*Cu rvl+xil*cu rv2
STRAIN_F=STRAIN_C/YCG*(h-YCG-dl)
DEF=Curv_0*l**2/9.6
rohf=Afrp/b/(h-dl)
beta= O.lOOdOO
beta=(rohf/rohb)/5
beta = 1 !
0.5*((Efrp/200000d00)+l) le=lc+C(beta*lc_0-Ic)*(Mc/Mext)**3)*(Efrp/200000d00)
le=ic+Cbeta*ic_0-ic)*(Mc/Mext)**(3*200000d00/Efrp)
le=(lc_0*lc)/(lc+(l-0.5*(Mc/Mext)**2)*(lc_0-lc))
IF (le .GT. IC_0) THEN
le=lc_0
ENDIF
Load=w*l/2
DEF=23.0D0*1oad*l**3/(648.OdO*le*E)
DEF=5*w*l**4/(384*Ie*E) !*1.06
DEF=STRAIN_C
defl=5*w*1**4/(384*lc_0*E)
def2=5fcw*1**4/(384*ic*E)
print *, def1,def2,YCG,E
END SUBROUTINE STRAIN_0
Page 1
INPUT.tXt
SUBROUTINE INPUT
*****************************************************************************
C
C
. { This subroutine is used to read the input data from an external file }
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
IMPLICIT DOUBLE PRECISION (A-Z)

INTEGER I
DIMENSION DATA_GEO(6), DATA_STEEL(2), DATA_FRP(2)
DIMENSION DATA_CON(l)
COMMON /DATAl/
COMMON /DATA2/
COMMON /DATA3/
COMMON /DATA4/
b,h,1,dl,d2,W,Mext
AS.ES
Afrp, Efrp
fc, EC, rohb
READ (9, *) DATA_GEO, DATA_STEEL, DATA_FRP,

DATA_CON, rohb
b=DATA_GEO(l)
h=DATA_GEO(2)
l=DATA_GEO(3)
dl=DATA_GEO(4)
OF TENSION REINFORCEMENT
d 2=DATA_GEO(5)
CG OF TENSION REINFORCEMENT
W=DATA_GEO(6)
AS=DATA_STEEL(1)
ES=DATA_STEEL(2)
REINFORCEEMNT
! BEAM WIDTH
! BEAM HEIGHT
! BEAM FLEXURE SPAN
! CONCRETE COVER IN TENSION ZONE MEASURED FROM CG
! CONCRETE COVER IN COMPRESSION ZONE MEASURED FROM
! APPLIED DISTRIBUTED LOAD
! AREA OF COMPRESSION REINFORCEEMNT (STEEL OR FRP)
! YOUNG'S MODULUS (MPa) OF COMPRESSION
AfrpDATA_FRP(1)
Efrp=DATA_FRP(2)
! AREA OF TENSION REINFORCEEMNT (STEEL OR FRP)

l.YOUNG'S MODULUS (MPa) OF TENSION REINFORCEEMNT
fc=DATA_CON(l)
! CONCRETE COMPRESSIVE STRENGTH
EC=4500*DSQRT(fO
Mext=w*l**2/8
END SUBROUTINE INPUT
Page 2
CREEP.txt
SUBROUTINE CREEP (t,E,Curv_0,Curvt,STRAIN_Ft,DEFt)
*********************************************************
{***********************************************************************
THIS SUBROUTINE IS USED TO COMPUTE THE CREEP COEFFICIENTS AT TIME t }
INTEGER I,J,t
COMMON /DATAl/
COMMON /DATA2/
COMMON /DATA3/
COMMON /DATA4/
b,h,1,dl,d2,W,Mext
AS,ES
Afrp, Efrp
fc, EC
PhiU=2.35
EPS_SHRU=0.0
Phi=((t)**0.6/(10+t**0.6))*PhiU
EPS_SHR=((t)/(35+t))*EPS_SHRU
CALL CG (EC,
YCG_0,IC_0,YCG,IC,STRAIN_C)
CALL CG (E,
YCG_0t,lC_0t,YCGt,let,STRAIN_Ct)
k=lc/lct
Eta=(YCG)/(YCGt)
rc2=ic/(YCG*b)
STRAIN_FRP=(STRAIN_Ct/YCGt)*(h-YCGt-dl)
STRAiN_c0=0.OdOO ! case of non prestressed STRAiN_C/YCG*(h/2-YCG)
yc=0.OdOO !(YCGt/2-YCGt)
print *, YCG, YCGt
Xi1=1-(Mc/Mext)**2
Curvl=Mext/(E*lc_Ot)
Curv2=Mext/(E*lct)
Cu rv=(1-xi1)*Cu rvl+xil*Cu rv2
DELCUR=k*(Phi*(curv_0+STRAlN_C0*(yc/rc2))+EPS_SHR*(yc/rc2))
DELEPS=Eta*(Phi*(STRAIN_CO+Curv*yc)+EPS_SHR)
DELCURl=k*(Phi*(Curvl+STRAiN_cO*(yc/rc2))+EPS_SHR*(yc/rc2))
DELCUR2=k*(Phi *(cu rv2+STRAiN_cO*(yc/rc2))+EPS_SHR*(yc/rc2))
ft=0.6*DSQRT(fc)
MC=ft*IC_0/YCG_0
Xi1=1-(Mc/Mext)**2
Xi2=1-0.5*(Mc/Mext)**2
Page
CREEP.txt
Cu rvt=(1-xi2)*(Curvl+DELCURl)+xi 2 *(cu rv2+DELCUR2)
STRAIN_Ft=STRAIN_FRP !Curvt*(h-YCGt-dl)
!Curv2*(h-YCGt-dl)+DELEPS
DEFt=Curvt*l**2/9.6
DEFt2=DELCUR*l**2/9.6
END SUBROUTINE CREEP
Page 2
CONCRETE_STRESS.txt
SUBROUTINE CONCRETE_STRESS (E,EPS,STR)
C
*************************************************************
C
C
{ This subroutine is used to calculate is used to calculate the concrete
stresses}
C
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * "
IMPLICIT DOUBLE PRECISION (A-Z)

CHARACTER MESSAGE*60
COMMON /DATAl/
COMMON /DATA2/
COMMON /DATA3/
COMMON /DATA4/
b,h,l > dl,d2,w,Mext

AS,ES
Afrp, Efrp
fc, EC
E_CONCRETE=E
C
xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
IF (E_CONCRETE .EQ. O.OdOO) THEN
MESSAGE= "CHECK INPUT CONCRETE DATA"
CALL ERROR (MESSAGE,60)
ENDIF
xxxxxxxxxxxxxxxxxxxxxxxxxxxx>9oooooodbooooooooooooooooooooooooooooooooooooooooo<
EPS0L0N_0=2*fc/E_C0NCRETE
xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
IF (EPS0L0N_0 .EQ. 0.0035d00) THEN
MESSAGED "CHECK CONCRETE STRESS-STRAIN RELATIONSHIP"
CALL ERROR (MESSAGE,60)
ENDIF
C
xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
IF (EPS.LE.EPS0L0N_0) THEN
STR=fc*((2*EPS/EPS0L0N_0)-(EPS/EPS0L0N_0)**2)
ELSE
STR=fC-0.15 *((EPS-EPS0L0N_0)/(0.003 5-EPS0L0N_0))
ENDIF
END SUBROUTINE CONCRETE_STRESS
Page 1
CG.tXt
SUBROUTINE CG (E,YCG_0,IC_0,YCG,IC,STRAIN_C)
C
***********************************************************
C
C
{ THIS SUBROUTINE IS USED TO COMPUTE THE UNCRACKED AND CRACKED CG OF THE
BEAM CROSS SECTION}
C
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *

INTEGER 1,1
CHARACTER MESSAGE*60
COMMON /DATAl/
COMMON /DATA2/
COMMON /DATA3/
COMMON /DATA4/
b,h,1,dl,d2,W,Mext
AS,ES
Afrp, Efrp
fc
TOL=l.0
MOMENT=0
AREA=0
Nfrp=(Efrp/E)-l
NS=(ES/E)-1
C
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
CALCULATE THE UNCRACKED CG (YCG_0) AND THE MOMENT OF INERTIA (lC_0)
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
*
*
*
YCG_0=((Afrp*Nfrp*(h/2-dl)-As*Ns*(h/2-d2))
/(b*h+As*Ns+Afrp*Nfrp))+h/2
lc.JD=(b*h**3/12)+(b*h)*(YCG_0-h/2)**2
+ As*Ns*(YCG_0-d2)**2
+ Afrp*Nf.rp*(H-dl-YCG_0)**2
C
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
CALCULATE THE CRACKED CG (YCG) AND THE MOMENT OF INERTIA (Ic)
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
Eps_elastic=(Mext*YCG_0/ic_0)/E
C=YCG_0*(2.0d00/3.OdOO)
EPSlLON_c=Eps_elastic
DO 1=1,100
c
DO 11=1,100
print *, EPSlLON_c,Mint
CALL FORCES (E,C,EPSILON_C,Cl,C2,T,Mint)
EQU=Cl+C2-T
IF (DABS(EQU) .GT. TOL) THEN
Page 1

Time Dependent Behaviour of FRP Rebar in Reinforced Concrete

Transféré par

Informations du document

Copyright

Formats disponibles

Partager ce document

Partager ou intégrer le document

Options de partage

Avez-vous trouvé ce document utile ?

Ce contenu est-il inapproprié ?

Droits d'auteur :

Formats disponibles

Time Dependent Behaviour of FRP Rebar in Reinforced Concrete

Transféré par

Droits d'auteur :

Formats disponibles

m m

TIME-DEPENDENT BEHAVIOUR OF FIBRE REINFORCED

Speciality: Civil Engineering

Sherbrooke (Quebec) Canada

Library and Archives

395 Wellington Street

395, rue Wellington

The author has granted a nonexclusive license allowing Library and

L'auteur a accorde une licence non exclusive

The author retains copyright

L'auteur conserve la propriete du droit d'auteur

In compliance with the Canadian

Conformement a la loi canadienne sur la

While these forms may be included

Bien que ces formulaires aient inclus dans

showed that the calculations, regarding immediate deflection, under estimate by 67 %;

A cet egard, un vaste programme experimental a eu lieu au laboratoire de durability de PRF de

de service (conformement aux codes disponibles en Amerique du Nord). Un suivi regulier de

est une valeur securitaire, dans des conditions standard de

Reinforced Polymer Reinforcement for Concrete Structures (FRPRCS-9), Sydney,

Results from other research work

Applications of FRP Composites for Construction (CDCC'2007)," Quebec City,

Background and Problem Definition

Obj ectives and Originality

Structure of the Thesis

Creep of FRP Composites and FRP Components

Composition of FRP Materials

Creep of Resin (polymer matrix)

Factors Affecting the Creep of FRP

Creep rupture stress limits for FRP

Creep and Shrinkage of Concrete

Classification of Different Types of Time-dependent Strains

Theories of Creep of Hardened Cement Paste

Factors Affecting the Creep of Concrete

Methods of Modelling the Creep of Concrete

Theoretical Prediction of Immediate Deflection

Theoretical Prediction of Long-term Deflection

Deflection Prediction (Immediate and Long-term) in Codes and Guidelines.. 44

Permissible Computed Deflections and Deflection Control (FRP-RC Beams)47

Effect of Sustained Loads on Flexural Crack Width in FRP-RC Beams

Long-term Behaviour of FRP Reinforced Concrete Beams under Sustained Load.. 52

Long-term Behaviour of FRP-RC Beams subjected to 2-point Loading

Long-term Behaviour of FRP-RC Beams under Uniform Distributed Load... 57

CREEP BEHAVIOUR AND RESIDUAL PROPERTIES OF GFRP

Fabrication, Instrumentation and Installation of Samples

Results and Discussion

Creep Tensile Strain

Residual Tensile Properties (Strength and Young's Modulus)

Pitfalls in Creep Testing

Summary and Conclusion

LONG TERM PERFORMANCE OF DIFFERENT TYPES OF GFRP

Fabrication, Instrumentation and Installation of Samples

Results and Discussion

Creep Tensile Strain

Residual Tensile Properties (Strength and Young's Modulus)

Summary and Conclusion

Sustained Load Frame Calibration and Beam Installation

Design of FRP and Steel Reinforced Concrete Members

Long-term Test Measurements and Monitoring

Results and Discussion

Creep Induced Strain in FRP Reinforcement and Compression Concrete

Deflection of Beam Elements

Crack Width Evolution for Beam Elements

MODELLING THE LONG TERM DEFLECTION OF FULL-SCALE