PHD Laursen-Masonry Prestressed Walls

http://researchspace.auckland.ac.
nz
ResearchSpace@Auckland
Copyright Statement
The digital copy of this thesis is protected by the Copyright Act 1994 (New
Zealand).
This thesis may be consulted by you, provided you comply with the
provisions of the Act and the following conditions of use:
Any use you make of these documents or images must be for

research or private study purposes only, and you may not make
them available to any other person.
Authors control the copyright of their thesis. You will recognise the
author's right to be identified as the author of this thesis, and due
acknowledgement will be made to the author where appropriate.
You will obtain the author's permission before publishing any
material from their thesis.
To request permissions please use the Feedback form on our webpage.

http://researchspace.auckland.ac.nz/feedback
General copyright and disclaimer
In addition to the above conditions, authors give their consent for the
digital copy of their work to be used subject to the conditions specified on
the Library Thesis Consent Form.
Seismic Analysis and Design
of
Post-tensioned Concrete Masonry Walls
By Peter T. Laursen
A thesis submitted in partial fulfilment of the requirements

for the Degree of Doctor of Philosophy
Supervised by Dr. Jason M. Inghant
University of Auckland
Department of Civil and Environmental Engineering
New Zealand
December 2002
ABSTRACT
This thesis explores the seismic analysis and design of post-tensioning concrete masonry
(PCM) walls. Using unbonded post-tensioning, walls are vertically prestressed by means of
strands or bars which are passed through venical ducts inside the walls. As the walls are sub-
jected to lateral displacements (in-plane loading), gaps form at the horizontal joints, reducing
the system stiffness. As long as the prestressing strands are kept within the elastic limit, or at
least maintain a considerable amount of the initial prestressing force, they can provide a restor-
ing force, which will return the walls to their original alignment upon unloading. The key feature in this behaviour is attributable to the tendons being unbonded over the entire wall heighg
allowing for distribution of tendon strain over the entire length of the tendon.
An extensive literature review found that post-tensioning of masonry has had limited application in seismic areas and that there currently are no specific code requirements for it's use for
ductile seismic design, largely as a consequence of little knowledge about the ductility capacity
and energy dissipation characteristics.
It was concluded that structural testing of PCM walls
and concrete masonry creep and shrinkage testing were essential to advance the understanding
of this construction type.

Creep and shrinkage experiments confirmed that long term preshess losses are considerable in
both grouted and ungrouted concrete masonry, and must be taken into account in design. It was
concluded that
it is essential to use high strength steel for
prestressing
of PCM in order to
reduce long term losses.
Structural testing confirmed that fully grouted unbonded post-tensioned concrete masonry is a
competent material combination for ductile struchral wall systems. In particular, PCM walls
strengthened in the flexural compression zones with confining plates are expected to success-
fully withstand
severe ground shaking from an earthquake.
It was
suggested that partially and
ungrouted PCM walls may suitably be used in strength design (non-ductile).
The proposed prediction method for wall in-plane behaviour was validated by experimental
results. Good correlation between predictions and results was found. Displacement spectra
were developed for ductile seismic design of PCM walls. These can be used to accurately estimate the displacement demand imposed on multi-storey PCM cantilwer walls.
-l-
DISCLAIMER
This thesis was prepared for the Department of Civil and Environmental Engineering at the
University of Auckland, New Zealand, and describes analysis and design of post-tensioned
concrete masonry walls. The opinions and conclusions presented herein are those of the author
and do not necessarily reflect those of the University of Auckland or any of the sponsoring par-
ties to this project.
ACKNO\ryLEDGEMENTS
This research was funded by direct industry sponsors W. Stevenson and Sons Ltd., Firth Industries Ltd., Ready Mix Concrete Ltd., and by the industry organisations New Zealand. Concrete
Masonry Association (NZCMA) and the Cement and Concrete Association of New Zealand
(CCANZ). VSL (Aust.) Pfy. Ltd., Construction Techniques Group Ltd., Fletcher Steel Ltd.,
Alan H. Reid Engineering Ltd. and SIKA (NZ) Ltd. generously provided their products free of
charge. Their contributions are gratefully acknowledged.
The author also wishes to acknowledge the contributions by Hank Mooy, Tony Daligan and
Mark Byrarni who were responsible for the practical aspects in relation to construction and
testing of the walls in the Civil Test Hall. Furthennore, the important contributions made by
final-year undergraduate students in designing and constructing the Series
testing setup, and
in testing of t'wo walls reported herein are acknowledged.

Special thanks are given to those who found time to comment on this thesis. This includes
Assistant Professor R.C. Fenwick and Dr. B.J. Davidson of the University of Auckland, E.
Lapish of Auckland, and particularly my primary supervisor and friend Dr. J.M. Ingham.
Finally, a very special thanks to my wife, Jill, for her lasting love and support, particularly
throughout the period ofresearch and preparation for this thesis.
-lt-
ADVISORY COMMITTEE
Dr. J.M. Ingham, SeniorLecturer (principal advisor)

Dr. B.J. Davidsoq Senior Lecturer
Assistant Professor R. C. Fenwick
Professor B. Melville
Mr. E. Lapish, ConsultantEngineer
EXAMINATION COMMITTEE
Assistant Professor C.E. Ventura, Department
of Civil Engineering, University of British
Columbia, Canada (chair)

Dr. J.M. Ingham, Senior Lecturer, University of Auckland
Dr. H.R. Ganz,VSL (Switzerland) Ltd.
Professor M.J.N. Priestley, University of California, San Diego, USA
-iii-
-iv-
LIST OF PUBLICATIONS
Journal articles:
Laursen, P.T. and Ingham, J.M. (2001), Structural Tbsting of Single-Storey Post-Tbnsioned
Concrete Masonry Walls, TMS Journal, Vol. 19, No.
l,
The Masonry Society, Boulder, Colo-
rado, Sept.2001.
Ingham, J.M., Laursen, P.T. and Voon, K.C. (2001), Appropriate Material Values
for
Use in
Concrete Masonry Design,Joumal of the Strucnral Engineering Society, Vol. 14, No. 1, Auckland, New Zealand,
April200l.
Laurseno P. and Ingham, J. M. (1999), Design of prestressed concrete masonry walls, Journal
of
the Stnrctural Engineering Society, Vol. 12, No. 2, Auckland, New Zealand, October 1999.
Conference papers:
Laursen, P.T., Davidson, B.J. and Ingham, J.M. (2002), Seismic analysis of prestressed concrete masonry shear walls,Proceedings of the l2th European Earthquake Conference, London,
England, September 2002, paper reference247.
Laursen, P.T. and Ingham, J.M. (2001), Research on multi-storey post-tensioned concrete
masonry walls, Proceedings of the NZCS Combined Concrete Industry Conference, Thupo,
New Zealand, October 2001.
Laursen, P.T. and Ingham, J.M. (2001), Material prcperties
for
New Zealand concrete
masonry, Proceedings of the 6th Australasian Masonry Conference, Adelaide, Australia, July
2001.
Laursen, P. and Ingham, J. M. (2001), Seismic Resistance of Prestressed Concrete Masonry

Shear Walls,Proceedings of the Structural Engineering Institute Congtess & Exposition 2001,
American Society of Civil Engineers, Washington DC, May 2001.
Laursen, P. and Ingham, J. M. (2000),In-plane Response of Post-tensioned Concrete Masonry,
Proceedings
of l2th lnternational Brick/Block Masonry Conference, p. 905, Madrid,
October 2000. (Alan H. Yorkdale, FASTM Memorial Award winning paper 2000)
Spain,
Laursen, P., Ingham, J. M. and Voon, K.C. (2000), Material Testing Supporting a Study of Posttensioned Concrete Masonry, Proceedings
ence, p.
937,Madid,
of 12ft International Brick/Block Masonry Confer-
Spain, October 2000.
Laursen, P. and Ingham, J. M. (2000), Cunent Research on Post-tensioned Concrete Masonry

Walls, Proceedings of the New Zealand Concrete Society Conference, p. 56, Thupo, New Zea-
land" October 2000.
Laursen, P. and Ingham, J. M. (1999), Ductile response of post-tensioned concrete masonry

walls, Froceedings of the IPENZ Technical Conference, Auckland, New Zealand, July 1999.
LaurEen, P.T., Seible, F. and Hegemeir, G.A. (1995), Seismic retrofit and repair of masonry
walls with carbon overlays, Non-Metallic (FRP) Reinforcement for Concrete Structures. Proceedings of the Second international RILEM Symposium (FRPRCS-2), pp. 616-623., Gent,
Belgium, July 1995,
Technical reports:
Laursen, P.T. and Ingham J.M. (2000), Cyclic In-plane Structural Tbsting of Prestressed Concrete Masonry Walls, Phase
I:
Simple Wall Confgurations, School of Engineering Research
Report #599, Dept. of Civil and Resource Engn., University of Auckland, New Zealand, September 2000.
Laursen, P.T. and Ingham J.M. (1999),
Mateial
Tbsting of Dricon Superset
Mortar, Auckland
Uniservices Research Report 7796.00, University of Auckland, New Zealand, July 2000.
Laursen, P.T. (1999) Progress Report and Feasibility Study into Prestressed Concrete Masonry
in New Zealand, Report prepared for the New Zealand Concrete Masonry Association, Auckland,
April 1999.
Laursen, P.T., Seible, F. and Hegemeir, G.A. (1995) Seismic repair and retrofit of masonry
walls with carbon overlays, Structural Systems Research Project, Report No. SSRP-95101,
University of California, San Diego, USA, June 1995.
-vl-
TABLE OF CONTENTS
ABSTRACT .............
DISCLAIMER ...........
ACKNOWLEDGEMENTS ........
ADVISORY COMMITTEE ...........
EXAMINATION COMMTTTEE ...........
LIST OF PUBLICATIONS
TABLE OF CONTENTS ........
LIST OF FIGURES
LIST OF TABLES
NOTATION
............i
..........ii
..........ii
..... iii
............... iii
...................v
............vii
......... xiii
..........xvii
.....................xix
INTRODUCTION
I.
1.I
t.2
PRINCIPLE OF UNBONDED POS}TENSIONED WALLS

........1
MOTIVATION FOR PCM RESEARCH ...........
............2
scopE AND ORGANISATION OF THESrS............... .................3
LITERATURE REVIEW
2. OVERVIEW
2.1 CODIFICAf,ION OF PRESTRESSED MASONRY
2.2 CURRENT STATE OF R8S8ARCH............
2.2.1 Materials
2.2.1.1 Concrete masonry....
2.2.1.2 Prestressing steel
2.2.2 Structural behaviour.
2.2.2.1 Out-of-plane response
2.2.2.2 In-plane response
2.2.2.3 Joints
2.2.2.4 Prestress loss...........
2.3 CONCLUSrONS........
2.4 REFERENCES..........
.............7
..........8
................9
.....10
...................10
......10
.....10
................10
......11
.......14
...................14
..............14
...............16
MATERIAL PROPERTIES
3.
3.1
3.2
3.3
3.4
OVERVIEW.............
STRESS-STRAIN CHARACTERISTICS
CRUSHTNG STRENGTH............
3.2.1 Grade dependent strength
3.2.2 Characteristic strength
3.2.3 Experimental data
3.2.4 Australian approach
3.2.5 USA approach............
ELASTIC MODULUS.,.............
3.3.1 Current New Zealand codification............
3.3.2 Literature and other codes........
MASONRY STRAIN CAPACTTY .....,........
-vii-
................ 19
...................20
..............21
...............22
....................23
,........24
......27
.....................2g
................28
........2g
.......2g
.........29
...............30
recommendations..
3.4.2 PCM testing results
FLEXURAL STRESS DISTRIBUTION...........
3.4.I
3.5
3.6
J.t
3.8
3.9
Code rules and
...................30
......
3l
.........32
THEORETICAL STRESS-STRAIN RELATIONSHIP......... ....... 33

3.6.1 Unconfined concrete masonry....
....34
3.6.1.1 Failure mode
............34
.....................35
3.6.1.2 Theoretical stress-sfain curve
........ 36
3.6.2 Confined concrete masonry....
............37
3.6.2.2 Theoretical stress-strain curve
.....................37
3.6.3 Evaluationofstress-strainrelationship.......... ....................38
.....................40
3.6.3.1 U200, Unconfined prisms.......
.....41
3.6.3.2 CP200, Confined prisms.......
prisms.......
3.6.3.3 CPl00, Confined
.....43
3.6.4 Effect of strain rate...........
...........-..44
3.6.5 Tensile strength.....
........45
....45
3.6.6 Unloading properries
....................45
PRESTRESSING STEEL,.....
prestressing
steel....... ..................... 46
3.7.1 Types of high strength
3.7 .2 Material properties
.......46
CONCLUSIONS
.....................47
.............. 49
REFERENCES..........
TIME DEPENDENT EFFECTS
4.
4.1
4.2
4.3
4.4
4.5
LOSSES...
CONSIDERATIONS.
concrete............
masonry
expressions........-.
stain.........
strain
steel
Loss
EXPERIMENT.........
experiments................
instrumentation
results
results..
experiments...........
VALUES
work....
ASSESSMENT OF PRESTRESSING
......5I
....52
THEORETICAL
4.1.1 Creep and shrinkage of
........-.52
4-I.2 Creep and shrinkage of concrete
.....-. 55
4.1.3 Mathematical
.... 56
4.1.3.1 Creep
.....57
.......58
4.1.3.2 Shrinkage
...................58
4.1.3.3 Relaxation of prestressing
4.1.3.4 Total Prestress
...................59
....60
CREEP AND SHRINKAGE
4.2.I Design of the
................ 60
.......62
4.2.2 Test setup and
4.2.3 Series I experimental
.......... 65
4.2-4 Series 2 experimental
........67
..................... 7l
4.2.5 Discussion of
....................74
EXPERIMENTAL VS. CODE
............74
4.3.1 Other experimental
4.3.2 Code values for creep and shrinkage.............. .................... 75
.....................76
4.3.3 Comparison
RECOMMENDATIONS FOR CREEP AND SHRINKAGE .............. ........... 76
4.4.1
..........76
4.4.2
.....................77
4.4.3
.....................77
TYPICAL PRESTRESS
.........,.......77
...............
Creep
Shrinkage.
Comments
LOSS
- vlll -
4.5.1 Typical losses - additive approach

4.5.2 Incremental approach
4.6 REFERENCES..........
....""""""""78
......'."""""""80
..............'85
STRUCTURAL TESTING. SERIES I

5.
5.1
5.2
s.3
5.4
5.5
WALLS............
TNTRODUCTION.....
CONSTRUCTION DETAILS
5.2.1 Wall specifications
5.2.2 Wall construction
5.2.3 Material properties.
TESTING DETAILS
5.3.1 Test setup
5.3.2 Testing procedure.
5.3.3 Predicted nominal flexural strength
5.3.4 Predicted masonry shear strength............
TEST RESULTS
5.4.1 Fully grouted walls
5.4.1.1 Force-displacement response
5.4.1.2 Damage pattern and failure mode
5.4.1.3 Sliding
5.4.1.4 Prestressing force.........
5.4.2 Partially and ungrouted walls....
5.4.2.3 Sliding
5.4.2.4 Prestressing force
5.4-3 Vertical masonry strain
STNGLE-STOREY PCM
...........
DrscussloN
5.5.1.1 Flexural strength.....
5.5.1.2 Ultimate drift capacity.............
5.5.1.3 Tendon yielding....
5.5.1.4 Vertical masonry strain........
5.5.1.5 Masonry shear strength............
5.5.1.6 Sliding propensity

5.5.1.7 Axial load ratio..
5.5.2 Partially and ungrouted walls...

5.5.2.1 Masonry shear strength............
5.5.2-2 Flexural strength.....
5.5.2.3 Vertical strain
5.5.2.4 Axial load ratio..
5.6
5.7
CONCLUSIONS AND FUTURE

5.6.1 Fully grouted
5.6.2 Partially and ungrouted
walls
...................87
..............' 87
....................88
...'....88
..""'."89
...""89
.............."89
'-.--...'............'89
'."""'90
.....-.'..'...'-..90
.'...'."91
.----.--.-..----....--92
-'-'."92
.--..92
.......-..-.....94
'.......'........-....95
........-.....95
--..--96
""'97
...-----.......'97
........-.............98
--..'98
'...'...........--.98
...............99
....'..99
........'..-."""'99
........--.........100
.....-'......-'....' 101
-....102
................-.103
..-.....--.......... 104
'....104
"...105
.................. 105
...-..-..-........106
.........106
RESEARCH..............
....- 106
............107
."..107
walls...
-lx-
-....108
.............109
STRUCTURAL TESTING - SERIES 2

6.
6.1
6.2
6.3
6.4
6.s
6.6
6.7
ENHANCED SINGLE-STOREY PCM
7.1
7.2
7.3
7.4
........... 1I1
............,.... I 11
................. I12
.................. 112
................. I 16
..... I 16
............. I 17
................... 117
....... 119
............:.................... 119
............... 120
................... 120
TNTRODUCTION
6.2.1 Wall specifications......
6.2.2 Wall construction.........
6.2.3 Material properties
TESTING DETAILS
6.3.1 Test setup.
6.3.2 Testingprocedure
6.3.4 Predicted masonry shear strength....
TEST RESULTS.
6.4.1 Force-displacement response
6.4.2 Damage pattern and failure mode
6.4.3 Sliding
6-4.4 Prestressing force
6.4.5 Vertical masonry strain........
DISCUSSION............
6.5.1 Flexural response...
6.5.2 Tendon yielding....
6.5.3 Vertical masonry strain........
6.5.4 Hysteretic energy dissipation
6.5.5 Sliding propensity
6.5.6 Initial stiffness
6.5.7 Comparison with Series I wall tests..........
CONCLUSTONS
REFERENCES ..........
STRUCTURAL TESTING. SERIES

7.
WALLS.............
.......I2I
..................123
......125
.......126
......... 128
............ 135
..... 135
......137
.........137
....... 138
....... 139
............ 139
..... 140
................... r42
............144
WALLS.............
INTRODUCTION
7.2.1 Wall specifications......
7.2.2 Wall construction.........
TESTING DETAILS
7.3.1 Testsetup
7.3.2 Instrumentation.........
7.3.3 Testing procedure
7.3.4 Flexural strength prediction.
7.3.5 Predicted masonry shear stength............
TEST RESUTJTS
7.4-l Force-displacement response
7.4.3 Sliding
7.4.4 Prestressing force ........
3-STOREY PCM
-x-
.......... l4s
................. 145
................. 146
.................. 146
................. 149
..... l5l
.............152
..152
.. 153
....... 154
......... 156
.......157
.. 158
....... 159
..................162
......162
.................162
7.4.5 External axial force

7.4.6 Vertical masonry strain
7.4.7 Lateral wall displacement profi1e.......
7.5
7.6
"""""'
response
behaviour.
strain
dissipation
propensity
7.5.6 Initial stiffness....
COMPARISON WITH RCM WALL TESTING
7.6.1 Priestley and Elder wa11s.........
7.6.2 Strategy for comparison of results...........-.
7.6.3 Comparison of results
DISCUSSION
7.5.1 Flexural
7.5.2 Tendon
7.5.3 Vertical masonry
7.5.4 Hysteretic energy
7.5.5 Sliding
7.6.4
7.7
7.8
.....167
--...-.......'...167
..--.....--..174
""""""'174
-.-..-..174
.......178
....'............178
......-.179
-......181
......... 181
.......182
.....' 182
.....183
........-..-.......186
Conclusions from comparison of PCM with RCM testing.................188
............ 188
.............190
CONCLUSIONS........
ANALYSIS AND DESIGN

8.
8.1
8.2
PREDICTION OF WALL IN-PLANE BEHAVIOUR............ .....191

FLEXURAL RESPONSE OF CANTILEVER WALLS................................192
8.1.1 Limit states""""
"""""192
.-.................194
8.1.2 First cracking.............
.'. 195
8.1.3 Maximum serviceability moment................
198
.......-..--8.1.4 Nominal flexural strength
.-----..-.202
8.1.5 Yield strength....
.---.-.-............2M
8.1.6 Flexural overstrength
----205
8.1.7 Ultimate displacement capacity
..--.-.---.-..-.-.209
8.1.8 Calculation example....
................2I1
DESIGN CONSIDERATIONS
----------2ll
8.2.1 Required prestressing force and area....
8.2.2 Un-cracked section analysis.....

8.2.3 Cracked section analysis
8.2.4 Ultimate limit state..
8.2.4.1 Strength design.......
8.2.4.2 Capacity design
8.2.4.3 Displacement capacity
8.2.5 Base sliding
8.3
VERIFICATION OF PREDICTION
8.4
METHOD
---..213
..-.-.'........213
.-.-214
..'-.--.........-.214
--...-214
.............215
..--.........-..-215
.......2I8
.............219
DYNAMIC ANALYSIS
9. TNTRODUCTION.....
9.I MULTI-DEGREE OF FREEDOM STRUCTURE....
9.1.1 Prototype structures
9.1.2 Finite element modelling
9.1.3 Push-over analysis....
-xi-
.............22r
..................221
.-..22I
..-...........223
.--224
9.2
9.3
EQUIVATENT STNGLE-DEGREE OF FREEDOM STRUCTURE ........... 226

9.2.1 Transformation ............
.................226
9.2.2 Definition of SDOF structure
.......227
DYNAMIC ANALYSIS ..............
...........229
9.3.1 Earthquake records
.....231
vs.
9.3.2 Time-History response: MDOF
SDOF
.....233
9.3.2.1 Displacement response
............235
9.3.2.2 Drift demand...............
.............237
response...............
response
SPECTRAL ANALYSIS...............
9.4.I Elastic response
9.4.2 Ductile response
9.4.3 Spectral displacement..............
9.4.4 Displacement ductility demand....
USE OF DISPLACEMENT SPECTRA
NOTES
CONCLUSIONS
9.3.2.3
9.3.2.4
9.4
9.5
9.6
9.7
9.8
Base moment
Base shear
............... 238
.................238
..........240
..........241
.........241
.....241
..................245
.,246
................247
...................247
............248
SUMMARY OF CONCLUSIONS
IO.
l0.l
IO.2
10.3
IO.4
10.5
10.6
CONCLUSIONS AND FUTURE RESEARCH.............. ...........249

MATERTAL PROPERTTES - CHAPTER
................ 2s0
TIME DEPENDENT EFFECTS - CHAPTER
.......250
STRUCTURAL TESTING - CHAPTERS 5, 6, AND 7 ...............................251
ANALYSIS AND DESIGN - CHAPTER
..253
........253
DYNAMIC ANALYSIS - CHAPTER
..............254
FUTURE
3................
4
8............
9:..........
RES8ARCH..............
APPENDIX A
A.
DEFORMATION AT NOMINAL FLEXURAL STRENGTH.....................257
APPENDIX B
B.
FORCE-DISPLACEMENTPREDICTION-CALCULATIONEXAMPLE265
APPENDIX C
C.
C.1
COMPARISON OF PREDICTED VS. EXPERIMENTAL BEHAVIOUR.. 269

FLEXURE/ROCKING PREDICTION
..........269
...................269
C. L
C.1.2 Comparison
...................269
.........270
C.I.2.l Series I
.........276
C.1.2.2 Series 2
C.1.2.3 Series 3
.........276
..................280
C.1.3
METHOD...........
I Assumptions..............
...............
walls
walls
walls
Conclusion
-xll-
LIST OF FIGURES
..................2
Fig. 1.1-Post-tensioned concrete masonry waII..........
.......13
Fig.2.t-schematic of PRESSS wall test 12-17)
Fig. 3.l-Typical stress-strain relationship for grouted concrete masonry.......................20
Fig. 3.2-Typical PCM cantilever wall force-displacement characteristics......................2l
Fig. 3.3-statistical compressive strength distribution, 20 series concrete masonry........24
Fig. 3.4-Prism and confining plate dimensions and specification..............'...................26
..........32
Fig. 3.5-Flexural stress b1ock........
..........34
Fig. 3.6-Priestley-Elder compression stress-strain curves t3-21.........
Fig. 3.7-Theoretical stress-strain relationships, definition of symbols...........................36
................39
Fig. 3.8-Typical failure modes for prisms.
..-.........40
Fig. 3.9-U200, Unconfined masonry stess-strain curve
............-...42
Fig. 3.1O-Variation of masonry strain e.o with masonry strength
Fig.
Fig.
Fig.
Fig.
Fig.
Fig.
Fig.
3.1l-CP200, Confined masonry stress-strain curye

3.12-CPl00, Confined masoffy stress-strain curve
............42
.........-..44
....-..-............47
3.13-Prestressing steel stress-strain relationship ...............
4.I-Change of strain in axially loaded and drying concrete masonry.....................53
.....61
4.2-Wallette dimensions and test setup.........
................64
4.3-Wallette test setup and instrumentation.........
..........65
4.4-series 1, creep + shrinkage strain......
Fig.4.5-series
1, creep
strain
..-...........-..66
variation.............
strain......
Fig. 4.G-Series 1, ambient air temperature

Fig.4.7-series 2, creep + shrinkage
Fig. 4.8-series 2, creep
Fig. 4.9-series 2, ambient air temperature and humidity
Fig. 4.1O-Experimental Creep coefficient
strain
......67
..........69
......-.-..-.....70
............71
variations.
.............-...73
Cr.............
.....................74
Fig.4.1l-Experimental Specific Creep lq. ............
..........79
Fig.4.l2-Expected prestress loss in fully grouted concrete masonry
....-.........-79
Fig. 4.13-Expected prestress loss in ungrouted concrete masoffy
Fig. 4.14-Time variation Fn(t) for creep, shrinkage and re1axation................................81

Fig. 4.1S--Comparison of prestress loss, incremental vs. additive approaches................83
................90
Fig. 5.I-Test setup and
......................93
Fig. 5.2-Fully grouted walls; Force-displacement
....................94
Fig. 5.3-Fully grouted walls; Extent of wall damage at
............96
Fig. 5.4-Fully grouted walls; Tendon force histories (all
Fig.5.5-Partiallyandungroutedwalls;Force-displacementhistories.... ......97
............98
Fig. 5.6-Partially and ungrouted walls; Extent of damage at
...................99
Fig. 5.7-Fully grouted walls; F-D envelopes
Fig. 5.8-Fully grouted walls; F-D envelopes, normalised with Vr................................100
instrumentation....
histories....
failure
tendons)...
...........
failure
...106
............
...113
4).............
.............114
5).........
..........115
Fig. 6.3-Wall CP-CA-ED energy dissipation bars..........
Fig. 5.9-PG and UG Walls, F-D envelopes, nonnalised with Vr.

Fig. 6.1-Wall geometry for CP walls (number I Fig. 6.2-Wall geometry for HB wall (number
xllr -
instrumentation
histories
failure...............
base..........
histories...........
Fig. 6.4-Typical
.......... 118
Fig. 6.5-Force-displacement
....................122
Fig. 6.6-Damage accumulation at
..................124
Fig. 6.7-Relative sliding between wall and
............125
Fig. 6.8-Prestressing force
............127
profiles,
pull direction on1y.............129
Fig. 6.9-FG:L3.0-W15-Pl-CR Horizontal strain
Fig. 6.IG-FG:L3.0-WI5-P2-CP, Horizontal
...... 130
Fig. 6.11-FG:L3.0-W15-P2-CP-CA, Horizontal stain
............ l3l
Fig. 6.12-FG:L3.0-W15-P2-CP-CA-ED, Horizontal strain
.....132
.................. 133
Fig. 6.13-FG:L3.0-Wl5-P2-HB, Horizontal strain
Fig. 6.14-F-D envelopes, P2-CP, P2-CP-CA and
......... 135
Fig. 6.15-F-D envelopes, P2-CP-CA and P2-CP-CA-ED.......... ................ 136
Fig. 6.lHumulated hysteretic energy
.................. 138
Fig. 6.17-Initial wall
........... 140
Fig. 6.18-F-D envelopes, P2-CB P2-HB, P2C and
.................. 141
Fig. 7.l-Wall geomebry for Series 3
........147
Fig.7.Z-Dimensions and location of confining
........... 148
Fig. 7.3-{a) Prestressing ducting and shear reinforcing and (b) confining plate.......... 150
Fig. 7.4-Reinforcing layout for RC
...... l5l
Fig. 7.5-Testing
......... 153
Fig. 7.6-External axial load
.................. 154
........... 155
Fig.7.7-Photograph of 53-l in the testing setup ready to be
Fig. 7.8-Instnrmentation, (a) in-plane lateral displacement and (b) load ce11s............. 156
Fig. 7.8-Instrumentation (Cont.), (c) vertical deformation and (d) sliding displ. ........157
.................... 160
........... 16l
Fig. 7.10-{yclic loading
............... 163
Fig. 7.1l-{a) and (b), 53-l Damage accumulation at
................... 164
Fig. 7.11{c) and (d), S3-2 Damage accumulation at
................. 165
Fig. 7.l3-Prestressing force
... 166
................... 168
Fig. 7.l4-Extemal axial load
Fig. 7.15-53-1, Horizontal strain profiles, levels 0-200 mm and 200400 mm............ 169
Fig. 7.15{Cont.) S3-1, Horiz. strain profiles, levels 400-600 mm and 600-800 mm.. 170
Fig. 7.1G-S3-2, Horizontal strain profiles, levels 0-200 mm and 200-400 mm............ l7l
Fig. 7.lflCont.) S3-2, Horiz. strain profiles, levels 400-600 mm and 500-800 mm..I72
............ 175
Fig.7.l7-S3-1, Lateral displacement
............ 176
Fig. 7.18-33-2, Lateral displacement
................177
.................. 180
Fig.7.2}-lumulated hysteretic energy
........... 181
Fig.7.2l-lnitial wall
................... 183
Fig.7.22-Priestley/Elder wall specification, from
....... 185
Fig.7.23-PEl experimental result and PCM-l prediction
....... 185
Fig.7.24-pE2 experimental result and PCM-1 prediction
....... 186
Fig.7 .25-PE3 experimental result and PCM-2 prediction
strainprofiles
profiles.....
profiles.....
profiles.....
P2-HB
stiffness....
dissipation
P2E...........
walls......
plates...
setup
elements...
system......
tested
histories
histories...............
failure.......
failure...
Fig.1.l2-Slidingdisplacement..............
histories
histories
profi1e................
profi1e................
envelopes
dissipation
stiffness....
[7-5]..........
envelope...
envelope...
envelope...
-xlv-
Fig. 8.I-Definition of cantilever wall dimensions, forces and deformations........ ........194

Fig. 8.2-First
............195
Fig. 8.3-Maximum serviceability moment
...............196
Fig. 8.4-Curvature distribution at maximum serviceability moment..........................,197
Fig. 8.5-Wall equilibrium at nominal flexural
............199
Fig. 8.6-Wall deformation at nominal flexural
.......200
Fig. 8.7-Wall rocking
.............203
Fig. 8.8-Vertical sffain evaluation at ultimate displacement
.........206
Fig. 8.9-Post-tensioned concrete masonry cantilever wall
...............210
Fig. 8.lG-Predicted in-plane wall
.................211
Fig. 8.ll-Cracked section
.......213
Fig. 9.I-S-Storey PCM wall prototype and
...................222
Fig. 9.2-Masonry constitutive
................224
Fig. 9.3-Cyclic push-over
...........225
...................225
Fig. 9.5-5-storey wall, Analytical prediction model based on Chapter 8.....................229
Fig. 9.6-Bi-linear elastic
............229
Fig. 9.7-Uniform Earthquake Hazard
........230
Fig. 9.8-Normalised earthquake records
.................232
Fig. 9.9-Spectral response of Nl'N4,3% and5Yo
....................233
Fig. 9.lG-5-storey, Displacement response MDOF vs. SDOF,
.......234
Fig. 9.1l-8-Storey, Displacement response MDOF vs. SDOF, Nl:N4.........................236
Fig.9.12-5-storey wall, storey
..............239
Fig. 9.13-8-storey wall, storey
..............239
Fig. 9.l4-Displacement spectra, cr:
...............242
Fig. 9.l5-Displacement spectra,
...........243
Fig. 9.lfDesign displacement Spectra, cr:
...2M
Fig. 9.l7-Displacement ductility demand
...............245
Fig. A.l-Definition of cantilever wall dimensions, forces and deformations...............257
Fig. A.2-Flexure / rocking cracking
..........258
Fig. A.3-Wall deformation at nominal flexwal
.......259
Fig. A.4-Moment-curvature
..........260
Fig. A.5-Wall deformation at Mn for unconfined concrete
..........262
Fig- A.G-Wall deformation at Mn for confined concrete
..............263
crackin9...............
................
response
strength
strength.....
capacity
..........
response..............
analysis
modelling............
relationships..............
response..............
relationships.............
model
Spectra
Nl:N4......
damping...
Nl:N4
shear.........
shear.........
0.01
a: 0.02....
0.01
Fa.............
patterns
strength....
relationships.....
masonry
masoffy
Fig.
Fig.
Fig.
Fig.
Fig.
Fig.
Fig.
Fig.
Fig.
Fig.
C.l-Series I FG-L3.0-W20-P3: Prediction vs. experimental behaviour...............271

C.2-Series I FG-L3.0-Wl5-P3: Prediction vs. experimental behaviour...............272
C.3-Series I FG-L3.0-W15-P2C: Prediction vs. experimental behaviour............273
C.4-Series I FG-L3.0-Wl5-P2E: Prediction vs. experimental behaviour.............274
C.5-Series 1 FG-L1.8-Wl5-P2: Prediction vs. experimental behaviour...............275
C.6-Series 2 FG-L3.0-Wl5-P2-CP: Prediction vs. experimental behaviour. ........277
C.7-Series 2 FG-L3.0-Wl5-P2-CP-CA: Prediction vs. experimental behav.........278
C.8-Series 2 FG-L3.0-Wl5-P2-FIB: Prediction vs. experimental behaviotn........279
C.9-Series 3 53-l: Prediction vs. experimental behaviour........... ...,...281
C.1G-Series 3 S3-2: Prediction vs. experimental behaviour........... .....282
xv-
-xvi-
LIST OF TABLES
TABLE 3.1-NZS 4230:1990 Masonry grades
..........22
TABLE 3.2-Measured crushing strength
..................25
prism
TABLE 3.3-{onfined
testing
TABLE 3.4-Ultimate compression strain
....................................31
TABLE 3.S-Vertical masonry strain, test results
.......31
TABLE 3.6-Material properties of prestressing stee1.........
........47
TABLE 4.l-Relaxation of prestressing steel (20"C)................ .....................58
TABLE 4.2-Specifications for series I ...............
......62
TABLE 4.3-Specifications for series 2 ...............
......63
TABLE 4.4-1reep and shrinkage testing, schedule of events...
....................63
TABLE 4.5-Series l, summary of results
.................68
TABLE 4.6-Series 2, summary of results
.................70
TABLE 4.7-Experimental maximum values of C", k, e.r,.
.........74
TABLE 4.8-Experimental values of k", normal weight concrete masomy.....................75
TABLE 4.9-Experimental values of lq, light weight concrete masonry.........................75
TABLE 4.l0-{ode defined creep and shrinkage coefficients............ ...........75
TABLE 4.1l-Recommended values of C., h, E r,.
......................77
TABLE S.I-Wall specifications.............
.......,...........88
TABLE S.2-Predictions and results
........91
TABLE 5.3-Measured masonry vertical strain
.........99
TABLE 5.4-Masonry shear strength
.....103
TABLE 6.1-Wall specifications
..,.........112
TABLE 6.2-Material properties ............
..................116
TABLE 6.3-Predictions and results
......120
TABLE 6.4-Measured masonry vertical strain
.......I34
TABLE 7.I-Wall specifications .............
.................146
TABLE 7.2-Material properties ............
.......-..........152
TABLE 7.3-Strength predictions ..........
..................158
......158
TABLE 7.S-Vertical stain
....................173
TABLE T.FTheoretical and experimental results for PE and 33 wa11s........................184
TABLE 7.7-Theoretical performance of PCM walls for comparison
.........184
TABLE 8.I-Basic plastic deformation zone parameters
...........207
TABLE 8.2-Predicted force and displacement
.......210
TABLE 9.1-Equivalent SDOF properties, m at each floor and roof.............................228
TABLE 9.2-Equivalent SDOF properties, m at each floor,m/2 at roof........................228
TABLE 9.3-Analysis results for S-storey wall models, maximum values....................235
TABLE 9.4-Analysis results for 8-storey wall models, maximum values....................235
- xvll -
xvlll
NOTATION
Roman characters:
a
as
As
ao
Ae
Ao.
Aprj
A,
,\
)
b*
c
C"
ce
d
d
D
4,
4,n
d"oh
d"
4n
4sr,
dn*
dj
d,
4n
4sr,
do
4
4,
d,
q
equivalent rectangular stress block compression zone length

seismic ground acceleration
gross area of
wall cross section
flexural compression zone length at Mo

confining plate area for Priestley-Elder stress-strain curve
total prestressing steel area in wall cross section
prestressing tendon area of tendon
J'
wall aspect ratio

shear (horizontal) reinforcing steel area
flexural compression zone length at M,
wall thickness
distance from exherne masonry fibre in compression to flexural neutral axis
concrete masonry creep coefficient
distance from extreme masoffy fibre in compression to flexural neufral axis
at M"
effective wall length for calculation of
V*
lateral displacement at h"

displacement level
total lateral displacement at h" due to V",

lateral flexural displacement at
due to V",
lateral shear displacement at h" due to V".
total lateral displacement at h" due to V"

lateral flexural displacement at h" due to V"
lateral shear displacement at tq due to Vu
lateral displacement demand at the equivalent SDOF structure height
lateral displacement of
floor'j'
total lateral displacement at tr. due to Vg

lateral flexural displacement at tq due to V1
lateral shear displacement at h" due to V1
lateral displacement increment at flexural overstrength
roof lateral displacement demand

target displacement demand
lateral displacement increment at first tendon yield

lateral displacement at fust tendon yield
-xrx-
d"
{,no
EI
E,o
En'
Eo,
et
et"
ot"
e1.;
f'"b
f'"8
f',
f'j
f'*
f'u"
f.
Lr
f'oi
Fn
Fn"
Fn'h
fe,
tr
fu
f'
{"
q"
fy
&r
h*
h"
h.,
h.
hn*
ho
ultimate wall displacement capacity

measured wall displacement at V*u*
elastic stiffness of wall cross section

masonry elastic modulus
average masonry elastic modulus related to Priestley-Elder stress-strain curve
prestressing steel elastic modulus
total tendon force eccentricity with respect to wall centre line

shortest horizontal distance between any tendon and compression end of wall
horizontal distance between extreme tendon in tension and compression end

of wall
horizontal distance between tendon
J'
and compression end of wall
concrete masoffy unit crushing strength

grout cylinder crushing strength, specific to AS 3700
grout cylinder crushing shength
mortar cylinder crushing strength

masonry crushing strength
unconfined masoffy unit crushing sfrength, specific to AS 3700

axial mason-ry stress
final prestress at termination of creep and shrinkage experiment

masonry stress immediately after prestressing
generic rymrbol for'function of'
time development function for creep
time development function for shrinkage

instantaneous prestressing steel stress
prestressing steel stress immediately after anchorage lock-off
ultimate (rupture) strength of prestressing steel
yield strength of prestressing steel

average tendon stress at Vs for wall in unloaded state
nominal yield strength of shear reinforcing steel

nominal yield strength of reinforcing steel
nominal yield strength of confining steel
equivalent height of SDOF structure
horizontal confined dimension

location of fust cracking height for applied moment of M"
wall equivalent height

equivalent height of modal mass of nth mode
vertical extent of plastic deformation zone at wall ultimate displacement
h,
h*
k
K
Kr
K,
lq
Kd
lq
I
l*
M
M*
M'f
M*
N4
M*o
Mo
N4*
Mo
M
Mty
M,
M,
Mr'
N
N*
N"
N,"
P
P*
Pi
Pj
Pl
Py
vertical distance between floors n and n-l
wall height
defining the maximum permissible extreme masonry strain kf,n at Ir{"
prism strength and strain enhancement factor for Priestley-Elder stress-strain
cuwe
initial stiffiress of bilinear elastic SDOF

post-yield stiftress of bilinear elastic SDOF
concrete masonry specific creep
similar to K but related to high strain rate

prestressing steel relaxation parameter
length of tendon
J'
wall length
base moment
applied factored moment

equivalent mass of SDOF structure
first cracking moment

maximum serviceability moment
maximum developed base moment from time-history analysis
nominal strength of wall (moment)
modal mass of nth mode
wall base moment overstrength

total seismic horizontal mass
wall base moment increase at first tendon yield

wall base moment at first tendon yield
required yield moment strength of nominally elastic SDOF structure
provided yield moment strength for SDOF structure

axial load due to wall self-weight" and live and dead load from suspended
floors
applied factored axial force
externally applied axial load in Series 3 experiments
wall self-weight in Series 3 experiments

prestress force in wall
applied factored prestress force

prestressing force immediately after anchorage lock-off
prestressing force in tendon
J'
(unloaded state)
prestressing force after all loss has occurred
total tendon force when stress in all tendons is fo,
-xxl-
R
ra
force reduction factor associated with ductile seismic design

concrete roughness amplitude of construction joint (shear friction calcula-
tion)
s
Su"
Sd
Sh
T
Tl
t
h
tl
t
ue
q
us
V
V*
Vu"."
V.,
%
Vr
vm
V,
v**
V.*
Vr"*
vertical spacing of shear reinforcing steel

elastic spectral acceleration
spectral displacement
vertical confined dimension (confining plate spacing)

fundamental structural period
first mode natural period

time
time of application of prestress to masonry
time at which all time dependent prestress loss has occurred
time at time step n
vertical extension of 'tension' end of wall at Vt
vertical displacement of top anchorage point of tendon
vertical shortening of ocompression' end of wall at Vt
applied lateral force at
'j' at V1
applied factored shear force

base shear due to lateral forces
wall lateral force at t5 corresponding to M.,

wall lateral force at tq corresponding to M"
lateral force applied at h" corresponding to Mn
masoffy shear strength (stress)

wall shear strength (force) due to masonry
maximum measured wall shear stress
maximum experimental lateral force
estimated maximum base shear at wall overstrength (shear friction calcula-
tion)
V,no
vneh,p
Vn,.
%
Vpp
V.
V,,
Vry
V"
maximum developed base shear from time-history analysis

shear strength (stress) predicted by NEHRP provisions
shear strength (stress) predicted by NZS 4230:1990 provisions
wall lateral force at h" corresponding to Mo

shear strength (stress) predicted by Paulay and Priestley provisions
wall shear strength (force) due to contribution from
V,
and Vr,
wall shear strength (force) due to horizontal reinforcing steel

wall lateral force at h" at first tendon yield
wall lateral force at h" corresponding to M"
- xxll -
Vy
!
y:
Z
Z^
Zâ
th)
{m}
required yield strength of nominally elastic SDOF structure
provided yield shength for SDOF structure

horizontal location of tendon
J' with respect
to wall centre line
seismic zone factor CNZS 4203:1992)

slope of descending branch of Priestley-Elder stress-strain curve
similar to Z^ but related to high strain rate

vector indicating location (height) of masses given by {m}
mass vector of MDOF structure
Greek characters:
cr
o
P
Aeo,
Ai,
Afpr
Afp,
Afr,
AP
APj
APry
APtvj
At
8",
%
en"
eni
tmp
gnu
%u
sh
0
0
defines equivalent rectangular stress block average stress
c[f'*
strain hardening ratio for bilinear SDOF structure

defines equivalent rectangular stress block length
a:
pc
strain increase in extreme tendon at first tendon yield
long term prestress loss due to creep

total long term prestress loss
long term prestress loss due to prestressing steel relaxation
long terrn prestress loss due to shrinkage
total tendon force increase at Mn
tendon force increase of tendon
J'
at Vg
total tendon force increase a fust tendon yield

force increase in tendon
J'
at first tendon yield
length of time step n

masonry axial strain
long term concrete masonry creep strain
masonry strain at maximum prism strength
f'-
extreme fibre strain in wall section due to M,

elastic masonry strain immediately after prestressing
masoffy axial strain at initiation of post-peak strength plateau for PriestleyElder stress-strain curye
maximum dependable masonry strain
ultimate elongation strain ofprestressing steel
concrete masonry final shrinkage shain
curvature at wall section due to applied moment M
wall average curvature in the plastic deformation zone atultimate displacement capacity
- xxlll -
0
0..
0"
0r
T"
Tr,*
Ti
Ti,,nu*
shength reduction factor

curvature at wall section due to M",
curvature at wall section due to M.
flexural strength reduction factor

non dimensional crack length at M"
drift demand at equivalent SDOF structure height

interstorey drift demand
interstorey drift limitation
yo
Tu.*
T,"
A
wall ultimate drift capacity
Fa
v
e
pu
p.
(&
x
q
q,,
E,
{0"}
displacement ductility demand
wall drift corresponding to dn,o

roof drift demand
Loss ratio between prestress loss calculated with additive and incremental
methods
poisson's ratio
wall rocking rotation

fransverse confining ratio for Priestley-Elder sbess-strain curve
volumetric confining ratio for Priestley-Elder stress-strain curve

seismic dynamic base shear amplification factor
ratio of net concrete masonry unit area to total area ofmasonry unit and void
wall axial load ratio

viscous damping ratio
wall axial load ratio at M"

wall axial load ratio at wall ultimate displacement capacity
mode shape vector for n6 mode
- xxlv -
Chapter 1
INTRODUCTION
I.
PRINCIPLE OF UNBONDED POST:TENSIONED WALLS
Using unbonded post-tensioning, walls are vertically prestressed by means of strands or bars
which are passed through vertical ducts inside the walls. As the walls are subjected to lateral
displacements (in-plane forces), gaps form at the horizontal joints, reducing the system stiffness.
As long as the prestressing strands are kept within the elastic limit, or at least maintain a
considerable amount of the initial prestressing force, they can provide a restoring force which
will return the walls to their original alignment upon unloading. Thus, the lateral force-displacement response may be idealised by a nonJinear elastic relationship. The integrity of the
walls is maintained as no tensile strains form in the wall units and there are no residual postearthquake di splacements.
The key feature in this behaviour is attributable to the tendons being unbonded over the entire
wall height, allowing for distribution of tendon strain over the entire length of the tendon. This
results in a large tendon elongation capacity before yielding which directly results in a large
elastic deformation capacity of the wall panel. It is important to use high strength prestressing
steel
(f,, > 900 MPa) because of its larger elastic strain capacity when compared to ordinary
steel grades (300-500 MPa). High strength prestressing steel improves the elastic wall displacement capacity and reduces the potential prestressing force loss due to creep and shrink-
A single large gap opening is expected to occur at the base of unbonded post tensioned
concrete walls, this location being a constnrction joint and subjected to the largest flexural
age.
demand.
Fig.
l.l
shows two Post{ensioned Concrete Masonry PCM) wall solutions schematically.
Both consist of a foundation beam, the masonry wall panel built of hollow core concrete
blocks, and a beam (bond beam) on top of the wall accommodating prestressing anchorages.
Typically grouting of the wall cavities is required. When grouting is required, prestressing
-l-
(o)
(b)
Loop sysiem
End onchoroge syslem
Fig. 1.1-Post-tensioned concrete masonry wall

ducts are placed in the wall cavities during construction. Solution (a) is prestressed with flexi-
ble high tensile strength strands looping through the foundation beam, while solution (b) is
prestressed with high tensile strength bars anchored in the foundation beam. Both solutions are
post-tensioned, i.e. the prestressing is applied after construction of the wall. The prestressing
force is normally applied using a hydraulic jack.
1.1 MOTIVATION FOR PCM RBSEARCH

Prestressed concrete masoffy can be regarded as a new construction form. The individual com-
ponents, concrete masonry blocks and prestressing steel, are well known and widely available,
however the combination of the two brings a new constnrction form with a set of unique properties amalgamating the compression strength of masonry and the tensile strength of the prestressing system. Through the pre-compression provided by the prestressing, cracking under
serviceability loading may readily be eliminated. Effectively one has a material with efficiency
similar to prestressed concrete.
In the last three decades prestressing has become widely used in the construction industry for
buildings, e.g. for floor beams and floor slabs, and for other types of structures, notably bridges
and offshore platforms. This extensive usage has led to the development of a variety of compet-
ing prestressing systems with reasonable pricing.

Post-tensioning of masonry has had limited application in New Zealand. As
will be expanded
on in the following chapter, there are currently no specific code requirements for the use of prestressed masonry in New Zealand. As a consequence of little knowledge about the ductility and
energy dissipation characteristics of preshessed concrete masonry, ,rmong other factors, this
-2-
technology is currently not applied for seismic design in New Zealand,. Designers are effec-
tively discouraged from speciffing prestressed masonry, partially due to uncertainty about the
material and panially because of the liability risk.
The lack of detail on the use of prestressed concrete masoffy in the current New Zealand
masoffy design standard is, in fact, due to a worldwide lack of understanding of the in-plane
seismic behaviour
of PCM, which differs
considerably from the behaviour
of reinforced
masoffy. There is a distinct lack of experimental data illustrated by the fact that the tests on
PCM shear wall subjected to cyclic in-plane loading, as reported in this document, constitute
the only testing of this kind yet to be carried out worldwide. Seismic design principles for
PCM, therefore, need to be developed through a research programme. Theoretical concepts
need to be refined and subsequently verified by laboratory experiments.
Recently, considerable attention has been given to the use of post-tensioned concrete walls.
This research effort underlines the large potential of using post-tensioning in conjunction with
concrete masonry walls because concrete masonry material properties are similar to those
of
concrete.
Motivation for the research reported herein derives from the fact that
(l)
the understanding
of
PCM walls subjected to seismic loading is limited, (2) the seismic structural performance is
likely to be satisfactory (or better) (3) little structural testing has been conducted, and (a) PCM
presents a potential industry opportunity.
In the wake of the above findings, an indusfiy supported research programme on Prestressed
Concrete Masonry (PCIO was initiated in 1997 at the University of Auckland. Initial research
was funded by the Univenity of Auckland and by the Cement and Concrete Association
of
New Zealand (CCANZ). The research actMties in the past three years were primarily funded
by the New Zealand Concrete Masonry Association (NZCMA) through a student grant.
I.2
SCOPE AND ORGANISATION OF THESIS
Scope:
The principal scope of the thesis is to provide detailed guidelines for complete analysis and
design
of ductile post-tensioned concrete masonry walls, including assessment of
demand and time dependent effects.
-3-
seismic
The focus in this thesis is on in-plane response of post{ensioned concrete masonry walls with
un-bonded tendons and fully grouted wall cavities. Three structural testing series were conducted to create an experimental data base. Analytical procedures were developed for predic-
tion and design of PCM walls subjected to lateral in-plane seismic forces. Material properties
of concrete masonry were investigated in relation to strength, strain capacity and time dependent losses.
Organisation
The thesis is organised in the following main subjects:
Literature Review: Chapter 2 presents a literature review of the curent state of research and
codification of prestressed masonry. It focuses on concrete masonry because it is the dominant
structural masonry material in New Zealand.Issues associated with materials and structural
response are discussed. The conclusions identifu the areas where research is needed, and serve
as background and
justification of the current prestressed concrete masoffy research activities
at the University of Auckland.
Material Properties: The objective of Chapter 3 is to discuss appropriate material properties

for grouted concrete masonry manufactured in New Zealand,, with particular emphasis given to
the use of these parameters when designing prestressed concrete masonry GCM) walls. The
discussion targets concrete masonry crushing strength, modulus of elasticity and strain capac-
ity, and assesses criteria from current international research and masonry design standards.
Time dependent e/fects.' Assessment of prestressing losses for prestessed concrete masonry
structures is presented in Chapter 4. The effective prestress level decreases over time due to
creep, shrinkage and steel relaxation, hence reducing structural efficiency. A detailed review
of
the current state of research on the topic is presented and recommendations are given for New
Zealand conditions with support from experimental results.
Stntctural ksting: Three series of structural testing were carried out. Chapters 5, 6 and 7 are
concemed with interpretation of the testing results. The principal intent with these in-plane
wall tests was to validate the use of PCM in a realistic sfuctural configwation. Furthennore,
the tests explored means of masonry confinement or strengthening that are expected to increase
the reliable wall drift capacities.
-4-
Analysis and design' Chapter 8 is concerned with analytical methods for predicting the inplane force-displacement relationship for unbonded post-tensioned concrete masonry. The
scope of such methods is to describe the wall behaviour throughout the entire loading range,
including non-linear large displacement response that can be anticipated in a seismic event.
Dynamic analysis: Chapter 9 is concemed with non-linear time-history analysis of unbonded
post-tensioned concrete masonry GClv! cantilever walls. Analyses were carried out in order to
clari$
the relationship between
wall stiftress and sfiength, and the wall displacement demand
due to dynamic seismic excitation. Pseudo displacement specha

design use, and a displacement focused design procedure is outlined.
-5-
(S)
were dweloped for
-6-
Chapter 2
LITERATURE REVIEW
2.
OVERVIEW
Post-tensioning of masonry has had limited application in New Zealand. There are currently no
specific code requirements for the use of presffessed masonry, largely as a consequence of little
knowledge about the ductility capacity and energy dissipation ofprestressed concrete masonry,
among other factors. Designers are effectively discouraged from specifuing prestressed
masoffy, partially due to uncertainty about the construction form and partially because of the
liability risk.
There are, however, a few examples of the use of prestressed concrete masonry in New Zealand. The classic example is a S-storey building erected in Christchurch in 1970. The apartment
complex was supported by 12
n. cavity walls with vertical
prestressing strands
time the solution was found to be both structurally and economically feasible.
l2-ll.At
It
that
should be
noted that the design requirements at that time were less stringent. The earthquake loading was
lower than seen today and the building was designed according to the allowable stress limit
state.
In the United Kingdom prestressing of masonry is mostly used for wall and pier designs. The
walls are primarily designed to resist out-of-plane loading, i.e. face loading such as strong
winds or earth prsssure. The in-plane forces in non-seismic areas are normally negligible in
comparison to the face loading. Post-tensioning of masonry has typically been used for stmctures of unusual dimensions, e.g. for large masonry beam spans and tall walls. There are, for
example, many reports on the use of prestressed masonry fin-walls [2-2] with eccentrically
placed post-tensioning because of unidirectional loading applied to such stnrctures as storage
tanks and retaining walls. The fins (or pilasters) greatly increase the lateral strength of the wall,
allowing for thin wall sections between the fins.
The use of prestressed masonry in the United States of America has been limited so far. There
are reports on some two-storey homes, factory buildings, retaining walls and sound walls, e.g.
l2-3,2-4). All of these were in non-seismic
areas. Currently in the U.S.A., prestressed concrete
masoffy is designed in accordance with allowable sfress design principles based on elastic
response to seismic excitation. Therefore, prestressed masonry is presently not applied in seis-
mic zones where design loads potentially can be reduced significantly by means of ductility
and energy dissipation.
Post-tensioning can be used advantageously for rehabilitation of old buildings. Such applica-
tion used for seismic strengthening is discussed by Ganz[2-5] who provides several case studies on masonry stmctures located on the American west coast- It is beyond the scope of this
thesis to discuss the use of post-tensioning for masonry building rehabilitation, so no further
details are given here on that subject.
Detailed discussions of the development of prestressed masonry research and application can
be found in Schultz and Scolforo p-61and Lissel et al.12-71.
2.I CODIFICATION OF PRESTRSSSED
MASONRY
The development of prestressed masonry was pioneered in the United Kingdom P-61, begin-
ning in the late 1950's and primarily promoted by daring structural engineers. The combined
efforts of development through the years led to a draft code in the late 1970's, which was even-
tually approved and incorporated into the general masonry code BS 5628,1985: 'Code of Practice for Use of Masonry'. This code is based on limit state design principles. The early uses
of
prestressed masoffy predominantly involved clay brick, which is widely available in the U.K.,
but more recent efforts have incorporated prestressed concrete masonry. BS 5628 part2, "Code
of Practice for Use of Masonry, Stnrctural Use of Reinforced and
Prestressed Mason.ry" was
revised in 1995 [2-8].
In the U.S.A., masonry codes have recently incorporated provisions for design of preshessed
masonry. The process of developing the appropriate chapters for amendment to the codes had
been proceeding since the early 1990's. The proposals were explicitly recognised with the pub-
lication of the MSJC 'Building Code Requirements for Masonry Structures' in 2002 12-91.
These were inspired by the relevant masonry codes and prestressing codes, and to some extent
by BS 5628. The MSJC 2002 is partially based on the allowable stress concept, which simplifies calculations but also produces unnecessarily conservative results. MSJC recognises the use
-8-
of prestressed masonry for ductile seismic design. A minimum amount flexural reinforcement
(mild steel) is required. Bonded prestressing steel may count towards the minimum amount of
reinforcement but unbonded prestressing steel may not. This means that mild steel is required
in the plastic hinge zone for unbonded post-tensioned masonry walls.

Prestressed masonry was first adopted in the Australian masonry design code with the issue
of
AS 3700-1998 [2-10]. This code mainly targets constnrction in non-seismic regions, but can be
applied for design of structures in earthquake zones given design to the elastic response of the
seismic forces. The prescriptions
for
preshessed masonry
iue largely inspired by
BS
5628:1995 [2-8].
There are currently no specific code requirements for the use of prestressed masonry in New
Zealand.In the New Zealand masonry design code NZS 4230:1990 [2-11] reference is made to
the design of prestressed masonry, being conceptually parallel to design of prestressed concrete, but no guidance is given. However, the fact that prestressing of masonry is incorporated
in the Australian masonry design standard is of significant importance in the New Zealand context due to the expected harmonisation of the New Zealand and Australian codes in the future.
It
seems natural that the design concepts for prestressed masonry in seismic zones for a future
common masoffy code should be developed in New Zealand, primarily because of New Zealand expertise in seismic design and because of less concern about earthquakes in Australia.
This is clearly a strong incentive to push forward the research into prestressed masonry in New
Zealand.
There has been significant interest in developing codes in many other countries. ln Canada, for
example, code development is well advanced and is coordinated with the U.S. effort.
2.2 CURRENT STATE OF RESEARCH

This chapter focuses on concrete masoffy because it is the dominant sfiructural masonry mate-
rial in New Zealand. Fired brick is widely used for wall cladding, but rarely for load bearing
purposes. [n order to describe the current state of research, issues associated with materials and
structural response are discussed. The subsequent conclusions identifo the areas where
research is needed, and serve as background andjustification
crete masonry research activities at the University of Auckland.
-9
ofthe current prestressed con-
2.2.1 Materials
2.2.1.1 Concrete masonry
The strength and constitutive properties of grouted concrete masonry are well established as
documented in Chapter 3. With the introduction of pre-compression, the creep and shrinkage
parameters of concrete masonry become more important, and the higher magnitude of unrecoverable deformation over time leads to significant prestressing losses. There have been sev-
eral studies on creep and shrinkage of concrete masoffy, most recently by Hamilton and
Badger 12-12) who studied U.S. concrete masonry.
It is unclearwhether
these studies apply
directly to the New Zealand type of concrete block, which may have significantly different
creep and shrinkage behaviour due to its unique aggregate composition. The studies by Hamil-
ton and Badger did not include experiments on fully grouted concrete masonry.
2.2.1.2 Prestressing steel

The constitutive properties of high strength prestressing steel are well established as documented in Chapter 3. In terms of time dependent effects, steel relaxation for high-strength steel
also affects the effective prestress level over time. This depends on the type of high-strength
steel, generally categorised as high relaxation or low relaxation. The steel relaxation effect is
well defined and may be extracted from most codes for prestressed concrete, e.g. NZS
3l0l
[2-
131.
2.2.2 Structural behaviour

2.2.2.1 Out-of-plane response
Current codes L2-8,2-9,2'10] that incorporate criteria on prestressed masonry, apply directly to
non-seismic application where out-of-plane lateral forces typically govem design. For nonseismic application, these codes, supported by extensive laboratory research, appear to perform
satisfactorily despite some differences. A selected list of out-of-plane wall tests can be found in
12-61and,12-71.
An extensive literature review suggests that no tests of prestressed concrete masoffy walls subjected to simulated oucof-plane seismic loading have yet been conducted. This lack of test data
limis the knowledge about the ductility capacity and energy dissipation properties of PCM
out-of-plane response, and necessarily leads to the use
-10-
of design forces
based on elastic
response to ground motion for design
of PCM walls carrying significant out-of-plane seismic
forces. Thus no force reduction due to ductility is permitted.
In most New Zealand applications, walls do not carry significant out-of-plane lateral seismic
forces, other than those arising from wall self weight and the imposed overall structural displacements (seismic forces are resisted by other structural elements). In that case the require-
ment for ductility and energy dissipation will be low and design to the elastic seismic response
will probably
be satisfactory.
2.2.2.2 In-plane response

The in-plane response of prestressed masonry walls for non-seismic design has been studied in
some detail. Page and Huizer [2-14,2-15] and Huizer and Shrive [2-16] have reported on
monotonic lateral in-plane load tests of post-tensioned hollow clay masonry shear walls.
Page and Huizer
l2-I4l tested three clay masonry walls. The walls were 3.0 m high and 2.5 m
long, constructed in running bond of hollow clay masonry units. The walls were reinforced
vertically with four VSL prestressing bars. Walls A and B had the four VSL bars prestressed
vertically to approximately 390IdlIupon construction. These walls were not grouted. The VSL
bars in wall C remained unstressed but were bonded to the wall by grouting of the wall cavities.
Furthermore wall A had four VSL bars embedded horizontally which were post-tensioned to a
total of about 200 kN. Testing of the walls in a monotonic push excursion revealed significantly different behaviour. The maximum applied horizontal shength was 146 kN, 175 kl.{ and
I 15
kN for walls A, B and C, respectively. Wall A failed prematurely due to a construction
error but the actual wall strength was assessed to be significantly higher than 175 kN. The typical failure mode for all walls was tensile splitting of the diagonal compression strut. Applica-
tion of prestress demonstrated that the vertical precompression in wall B and more so the
vertical and horizontal precompression in wall A deferred wall tensile splining to higher horizontal loads when compared to the non-prestressed wall C. Effectively larger strength and displacement were reached for wall B (and also for wall
A had it not failed prematurely) during
testing. These wall tests attest to the significant increase in masonry shear strength due to prestressing. However, they did not provide much information on prestessed wall behaviour in
the context of ductile seismic design where reversing displacement excursions far beyond the
wall yield displacement are expected. Similarly, the tests reported in [2-15] and [2-16] demon-
l1-
strated that prestressed masonry walls provide higher strength and comparable deformation
capacity, when compared to reinforced masonry walls.
The design procedures for shear walls subjected to in-plane forces, suggested by l2-l4l,12-151
and [2-16], largely follow those for the out-of-plane response as given by 12-8,2-9,2-L01. These
codes have drawn on research into both prestressed concrete and prestressed concrete masonry
t2-6).The procedures treat presfressed masonry walls as conceptually similar to unreinforced

masonry walls with the prestress acting as an additional extemally applied axial load.
Currently, limited experimental data exists for prestressed concrete masonry walls under inplane seismic loading.
It is in fact believed, that the only
tests to have considered unbonded
post-tensioned concrete masonry walls subjected in-plane seismic loading are those conducted
at the University of Auckland, as reported in Chapters 5, 6 and 7. For proper in-plane ductile
seismic design of unbonded post-tensioned concrete masonry walls, it is imperative to understand the ultimate behaviour of prestressed masonry. It is important to be able to categorise var-
ious failure modes, such as flexural failure characterised by concrete compression failure or
steel rupture, shear failure characterised by diagonal tension cracks or bed joint sliding, joint
failure and axial instability. As knowledge of these modes of failure, and their associated
strength and ductility capacity is heavily dependent on experimental work, it is a cumbersome
process to establish an appropriate design procedure. The experimental database must consider
cyclic load tests, simulating seismic motion. Clearly more in-plane testing is needed.
On the confary, there has been extensive research undertaken on other types of prestressed
systems under seismic loading, notably prestressed concrete. The culmination of this research
was testing of a precast prestressed concrete wall with unbonded tendons carried out by the
"PRESSS" coordinated research group in the USA [2- 17,2- I 8]. The elevation of the large scale
wall is shown in Fig.
l'
2.l.lt
consisted of four panels each 9' (2.7a m) long,
l8'-9" (5.72 m) tall,
(0.305 m) thick, and stacked in a 2 x 2 pattem. The wall was prestressed vertically with
unbonded prestressing tendons that were located centrally in the wall panels. The wall panels
were clamped together vertically by the prestressing such that panels I and 3 (P13) could move
independently in the vertical direction from panels 2 and 4 (P24\. UFP connectors (energy dissipators) were located in the interface between Pl3 and P24 such that relative vertical displacement between Pl3 andP24 due to wall lateral displacement would force the UFP connectors to
dissipate energy. The wall panels were lightly reinforced vertically. Panels
confinement reinforcement embedded in the compression toes.
-t2-
and 2 had special
Applying cyclic lateral load to the wall showed that high displacement capacity could be
achieved, far beyond zoh drift, and that this displacement capacity relied on integrity of the
wall toe regions which acted as pivot points for the observed rocking behaviour. The onset of
rocking could be compared to a reinforced concrete wall reaching its yield strength. After
onset of rocking the
wall strength only increased at a low rate
as
little stress increase occurred
in the unbonded prestressing tendons as a result of the lateral displacement. Failure never
occurred to the wall despite being taken to drift ratios far beyond 2Yo and only damage of cos-
metic character in the compression toe regions occurred.
A similar study was conducted at the University of Canterbury [2-19,2-20].

In conclusion, testing of unbonded post-tensioned concrete walls proved that extremely satisfactory in-plane behaviour could be achieved with the prestressing tendons remaining elastic
for interstorey drifts of the order of 2%. Only damage of cosmetic character occurred to the
confined corners of the walls. The PRESSS study also found that the displacement response
Flclrtcr.d
TCYFmrc
Urbodrd
UFPC,:@dd!
I
Fnrnc
I
{-
X
X
E
Wrll
Pnd
wdl
Prcl4
Wrll
hsol I
Wrtl
hd
X
X
OpGtr
I5'-0'
-13-
X
X
I5'-0"
Fig. 2.I-Schematic of PRESSS wall test
l2-l7l
amplitude could be effectively confolled (limited) by incorporation of enerry dissipators.

lnsignificant residual wall displacement was observed after sustaining the peak drift. The
structural behaviour of this type of wall resembles that of PCM, so results are highly relevant to
the current study.
2.2.2.3 Joints
Wall to floor joints could potentially be simplified with the use of prestessing. As suggested
by Ganz 12-211, a system with precast concrete floor slabs clamped between prestressed concrete masonry walls from the below and above storeys could potentially reduce the amount
mild steel in
the
of
joint. Reinforcement congestion in the joint can be avoided and speed of con-
struction improved. There has been some research into this type ofjoint for concrete solutions
for non-seismic application. There is, however, a total lack of experimental data for prestressed
concrete masonry joints of this type when subjected to seismic loading.
2.2.2.4 Prestress
loss
Guidelines for assessing the prestress loss are provided by existing masonry codes, e.g. BS
5628l2-8), MSJC [2-9] and AS 3700 [2-10]. Application of the provisions in these codes to
similar materials results in considerable difference in the calculated prestress losses. It has
been estimated [2-6] that the prestressing losses over time due to creep, shrinkage, steel relaxa-
tion etc. for concrete masonry may amount to up to 30% of the initial tendon stress.
2.3 CONCLUSIONS
The following conclusions on the current state of prestressed concrete masonry wall research
are drawn from the above discussion. These conclusions focus on the use of unbonded post-
tensioning, which is the topic of this thesis. Furthermore, the focus of this document has been
limited to studying prestress loss due to time dependent effects and in-plane wall response.
(l)
Material properties for concrete masonry and prestressing steel, in terms of constitutive
behaviour (stress-strain relationship), have been researched extensively and are rela-
tively well understood.
(2)
There is a need for better understanding of the creep and shrinkage properties for concrete masonry subjected to axial loading, with a view to New Zealand conditions. This
is particularly the case for fully grouted concrete masonry, which has not been thoroughly examined experimentally.
-t4-
Nunnerous laboratory tests have been conducted

stressed concrete lnasonqr walls, but
m the out-of-plane
response of pre-
only few in-plane tests. Little research has tieen
done on in-plane seismic response, neither experimentally nor theoretically, in particrr-
lar for the case of rmbonded post-tensioned concrete lunonry walls.
-15-
2.4 REFERENCES
L2-Il Hanlon,
J. R. G., Concrete Masonry in New Zealand: Prestressed Concrete Masowyo
Concrete, The Concrete Society (England), Vol. 4, No. 9, September 1970.
t2-21 Curtin, W.G.,
Shaw, G., Beck, J.K. and Parkinson, G.1., Masonry-frn walls, The Struc-
tural Engineer, Vol. 624, No. 7, July 1984.
12-31 Biggs, D.T., Planning Projects to Use Prestressed Masonry, Proceedings of the 2001
Structures Congress and Exposition (CD-ROM), American Society of Civil Engineerso
Washington D.C., May 2001.
l24l
Hamilton III, H.R. and Woodham, D.B., Prestress Losses in Post-tensioned Concrete
Masonry Sound Walls, Proceedings of the 2001 Structures Congress and Exposition
(CD-ROM), American Society of Civil Engineers, Washington D.C., May 2001.
12-51 Gatu, H.R., Strengthening of masonry stntctures with post-tensioning,Proceedings of

the 6th North American Masonry Conference, Philadelphia, Pa., 1993,YoL.2, pp. 645655.
12-61 Schultz, A. E. and Scolforo, M. J., An Overview of Prestressed Masonry, The Masonry
Society Journal, August 1991, Vol. 10, No. l, pp. 6-21.
12-71 Lissel, S.L., Sayet-Ahmed, E.Y. and Shrive, N.G., Prestressed Masonry - The Last
Ten
Years,Procedings of the 8fr North American Masonry Conference, Austin Texas, USA,
June 1999.
[2-8] 855628:
1995, Part
2:
Code of Practice
for
use
of Masonry. Structural Use of Rein-
forced and Prestressed Masonry, British Standards Institution, London.

12-91 MSJC: Building Code Requirements
for
Masonry Structures and SpeciJication
for
Masonry Stntctures, ACI 530-02/ASCE 5-02/TMS 402-02, The Masonry Standards

Joint Committee, USA, 2002.
[2-10] AS 3700-1998, Masonry Structures, Standards Australia, Homebush, NSW Australia.

[2-11] NZS 4230:1990, Code of Practice
for
the Design of Masonry Structures, Standards
Association of New Zealand, Wellinglon.
l2-I2l Hamilton, H.R. III
and Badger, C.C.R., Creep Losses in Prestressed Concrete Masonry,
The Masonry Society Journal, July 2000, Vol. 18, No.
- 16-
l.
l2-I3l NZS 3101:1995, Code of Practice for the Design of Concrete Stnrctures,
Standards
Association of New Zealand, Wellington.
l2-I4l
Page,
A. W. and Huizer, A., Racking
Tbst on Reinforced and Prcstressed
Hollow Clay
Masonry Walls, Proceedings of the Eighth International Brick/Block Masonry Conference, Dublin,Ireland,
Yol.2, September 1988, pp. 538-547.
[2-15] Page, A. W. and Huizer, A., Racking Behaviour of Prestressed and Reinforced Hollow
Masonry Walls, Masonry International, The British Masonry Society, Vol. 2, No. 3,
Winter I 988/1 989, pp.
97
-102.
12-16l Huizer, A. and Shrive, N. G., Performance of a Post-tensioned Single Wythe, Clay Brick
Masonry Wall Tested in Shear, Proceedings of the fourth Canadian Masonry Symposium, Fredericton, N.8., Vol. 2, June 1986, pp. 609-619.
l2-l7l Nakaki, S.D.,
Stanton, J.F. and Shritheran, S., An overview of the PRESS,S Five-Story
Precast Test Building,PCl Journal, March-April 1999,pp 26-39.
[2-18] Priestley, M.J. Nigel, Shritheran, S., Conley, J.R. and Pampanin, 5., Preliminary Results
and Conclusions From the
PRESSS
Five-Story Precast Concrete Test Building, PCI
Journal, Yol.44, No. 6, Nov.-Dec. 1999.
I2-l9l Rahman, A.M. and Restrepo,

Performance
J.1., Earthquake Resistant Concrcte Buildings: Seismic
of Cantilever Walls Prestressed Using
Unbonded Tbndons, Research
Report 2000-05, Department of Civil Engineering, University of Canterbury, Christchurch, New Zealand,.
12-201Holden, TJ., A Comparison of the Seismic Performance of Precast Wall Constntction:
Emulation and Hybrid Approaches, Research Report 2001-04, Deparfinent
of Civil
Engineering, University of Canterbury, Christchurch, New Zealand,.
[2-2I] Ganz, H. R., Post-tensioned Masonry Stntctures, Report 2, VSL lntemational Ltd.,
Berne, Switzerland, 1990.
-t7
- l8 -
Chapter
MATERIAL PROPERTIES
3.
OVBRVIEW
The objective of this chapter is to discuss appropriate material properties for grouted concrete
masonry manufactured in New Zealand, with particular emphasis given to the use of these
parameters when designing prestressed concrete masonry
eCM) walls. The discussion
targets
the concrete masonry crushing strength, modulus of elasticity and strain capacity, seen from a
New Zealand perspective. However, a review of criteria from current international research and
masonry design standards is included. Strength and strain enhancement of the wall compression zone due to confinement is also discussed, and a concrete masonry constifutive relation-
ship model for detailed structural analysis is described and erraluated. Prestressing steel
material properties are discussed at the end of this chapter. Time dependent effects in PCM,
notably masonry creep and shrinkage, and prestressing steel relaxation are discussed in Chapter 4-
It is noted that the term 'PCM' in this chapter refers to post-tensioned concrete masoffy walls
with unbonded tendons, suitable for ductile seismic design.
Design of PCM generally benefits from an accurate estimation of the material properties. Currently, the New Zealandstandard for design of masonry structures NZS 4230:1990
[3-l]
speci-
fies unrealistically low masonry compression strength. This may not greatly affect the design
of reinforced masonry, but certainly affects the design of PCM, where increased axial wall
forces place greater demand on the available compression strength of the masoffy. Similarly,
the current New Zealand practice is to adopt unrealistically high values of the masonry elastic
modulus, thereby underestimating the structural period and overestimating the design lateral
forces resulting from earthquake response. Structural testing of PCM walls has shown large
masonry strain capacity of the flexural compression zone. Strain values far beyond those spec-
ified in NZS 4230 have consistently been measured, suggesting that current code criteria are
unnecessarily disadvantageous when designing PCM.
-19-
It is argued that there is little value in
190 mm high blocks

confined 90 mm high blocks
+/
,m
UI
a
q)
l-
unconfined
190 mm
high blocks
a
0m t-,
Fig.
3.l-Typical
Stroin
stress-strain relationship for grouted concrete masonry
using either an unrealistically high or uffealistically low assessment of material properties

when conducting structural design, and that attention should instead be given to formulation
of
accurate predictive methods to establish characteristic material properties.
3.1
STRESS-STRATN CHARACTERISTICS
Grouted concrete masonry is a composite material consisting of hollow precast concrete blocks
laid up with horizontal and vertical mortar joints and subsequently filled with grout. The prop-
erties
of such material
consequently uary considerably depending on the proportions and
geometry of the ingredients, and the quality of workmanship.
Fig. 3.1 shows typical stress-strain relationship curves for grouted concrete masonry subjected
to uniaxial compression [3-2]. Three curves are shown: one for plain grouted masonry (uncon-
fined) and two for grouted masonry with confining plates in the bed joints. Four distinct
parameters describe the material behaviour:
(l)
the initial elastic modulus
E,, (2) the peak
strength f',n and (3) it's associated strain capacity em, and (4) the ultimate strain capacity e*u.
Plain grouted masonry (unconfined) typically reaches the peak strength at a strain of 0.00150.002 and provides reliable strength at strains up to about 0.003-0.004.
Confining the prisms with plates in the horizontal bed joint improves masonry performance
dramatically l3-2l.The peak strength is improved, but most importantly the ultimate strain
capaciry enu, is increased significantly to provide reliable strength at strains of up to 0.0080.010 for 190 mm high masonry blocks confined in each bed joint and higher for 90 mm high
masonry blocks confined at each bedjoint.
-20 -
3.2 CRUSHING STRENGTH

Meaningful estimation of the masonry compression strength, f',n, is important for accurate pre-
diction of PCM wall behaviour. Fig. 3.2 shows the predicted in-plane lateral force-displacement (F-D) relationship for a realistic PCM wall, 3.6 m long, 15 m high and 190 mm thick,
post-tensioned with four 15 mm high strength prestressing strands.
dead and live load of above floors
A self-weight of 200 kl.l,
of 665 kN and an initial prestress of
70Yo of the
yield stress
were assumed. The force-displacement relationships were predicted based on the analytical
method presented in Chapter 8 for an effective height, tL,
of 10 m using confined masoffy
(CPl00) compression strengths, fn' of 8 MPa, 12 MPa, 18 MPa and24 MPa, with all other
parameters the same.
f'r,
It is seen that the wall strength increases significantly with an increase of
as a direct result of compression zone length shortening and an associated increase of lever
arm between the flexural compression zone centroid and the cenfioid of the wall axial load.
Fig. 3.2 also shows that the wall displacement capacity at nominal flexural strength (symbol A)
and the ultimate displacement capacity (symbol O) increase with an increase of
f'-.
This effect
is caused by shortening of the compression zone, allowing higher curvature and rotation of the
hinging zone given the extreme fibre strain associated with the respective limit state. The
example shown in Fig. 3.2 suggests that the effective wall stiffness at maximum serviceability
moment (extreme fibre compression of 0.55fl,") (symbol
x)
reduces with an increase of
because of a shorter compression zone lenglh and a longer wall base crack.
300
o
250
.a'
h*=15m
he=10m
l*=3.6m
b* = 0.19 m
22oo
J
o
E. = 80OfAp = 600 mm2

Ep" = 190 GPa
fpy = l58lMPa
150
N=865kN
P=683kN
tn
o
in
Masonry compressive
E roo
strength
E = 0.129
f'-
fp" = 1138 MPa

f. = 2.3 MPa
hp = o,7o m
tmu = 0 015
K = r,05
50
a=0.9K=0.95
p = 0.e6
50
100
150
200
250
300
Displacement (mm)
Fig.
3.2-Qpical PCM cantilever wall force.displacement characteristics
-2t
fl*
The masonry compression strength,
f'*,
must be assumed for design. The designer has the fol-
lowing options: (i) adopt a code-defined characteristic strength based on material choice and/or
workmanship, (ii) specifu higher strength justified by a record of consistent high strength construction provided by the contractor/mason or
(iii) specify higher
strength to be validated by
testing of prism or grout cylinders made on site during construction.
Using options (i) or (ii) eliminates the need for material testing for the particular job. Option
(ii) most likely will allow assumption of a higher strength than using (i), given a consistent
level of workmanship. Option (iii) requires material testing specific to the job. Testing of
masonry prisms is a cumbersome aflair and concrete testing machines rarely are able to
accommodate the prism dimensions. Testing of grout cylinders is much easier to handle.
In the New Zealand context, f'n' is associated with the concrete masonry compressive strength
found by testing of prisms with a minimum height-to-thickness ratio of 3. Typical NZ prisms
are three blocks high and have height-to-thickness ratios of 4.3 and3.2, for 15 series and20
series masoffy blocks respectively.
3.2.1 Grade dependent strength

Currently, the New Zealand masoffy design code NZS 4230:1990 specifies the compressive
strength for fully grouted masoffy according to the level of supervision on the construction
site. Three grade levels are stipulated as shown in Table 3.1.
It is observed
that the design
strength is not related to the actual material strengths of blocks, mortar and grout, except for a
minimum concrete block strength requirement of l2 MPa. The material specifications for mortar and grout are found in NZS 4210:2001 [3-3], that stipulates a minimum mortar strength
12 MPa and a
of
minimum grout strength of 17.5 MPa.
The New Zealand system of grade dependent strength is essentially of historical character and
will most likely be replaced in the near future by a system based on characteristic strength of
the actual materials, similar to the approach found in the Australian standard for masonry
structures AS 3700 t3-41.
TABLE 3.1-NZS 423021990 Mason

Compressive strength
f'*
(MPa)
8 or higher when confirmed by testing

Supervised construction
Non-supewised construction
-22-
3,2,2 Characteristic strength

Modem manufacture of concrete products is associated with strict quality control. A statistical
record of strength is required so that the most likely strength (average sfength) and the charac-
teristic strength (typically 5 percentile value) of the product can be established.
The New Zealand standard for manufacturing of typical hollow normal-weight concrete
masonry blocks, NZS 3102:1983 [3-5], details that 49 out of wery 50 blocks attain a strength
of at least l2 MPa and that 45 out of every 50 blocks attain a strength of at least 14 MPa.
NZS 3109:1997 [3-6] details the minimum production standards for various concrete categories. Assuming high grade grout production with a performance record and the lowest available
specified strength
of
17.5 MPa, the statistical descriptors are target mean strength
and coefficient of variation
of
of 25 MPa
15.5.
Assuming that both block and grout strengths are normal distributed random variables, one
(l)
finds the following characteristics:

2.59 MPa, characteristic strength
MPa, standard deviation
:3.94
block mean strength
17.3 MPa, standard deviation
13.0 MPa (5 percentile) and (2) grout mean
strength:25
MPa, characteristic strength: 18.5 MPa (5 percentile).
Priestley and Chai [3-7] found that the grouted masonry compressive strength for 190 mm high
units with 10 mm mortar bed thickness could appropriately be calculated from knowledge
of
the block strength (uniaxial strength of one block), f'"6,arrd grout strength (typically 100 mm
diam. by 200 mm cylinder), f'n, according to the following formula:
f; :
0.59xf,'6 + 0.90(1
-*)fi
(3.1)
where x is the ratio of the net concrete block area to the total area of the block and void, taken
as .r
: 0.45 for 20 series blocks and 0.55 for l5 series blocks. Eqn. 3.1 was based on prism te st
data from New Zealand and North America. Assuming that the block and grout strengths are
independent random variables, it can be shown that
iable with mean strength
of
masonry, and mean strength
f'*
is also a normal distributed random var-
17.0 MPa and standard deviation
of
of 2.07 MPa for 20
15.7 MPa and standard deviation
of
1.80 MPa
series
for 15 series
masonry. Fig. 3.3 illustrates the strength distribution of grout and mortar, and the resulting
sfength distribution for grouted 20 series concrete masonry. Characteristic masonry strengths
f',o(char)
of
13.6 MPa and 12.8 MPa for 20 series and 15 series masonry respectively, are
determined based on the 5 percentile value of its normal distribution; effectively values that
-23 -
s
t,
510
3
CT
l!
15
20
25
Compre*sion Stength (MPa)
Fig.3.}-Statistical compressiye strength distribution,20 series concrete masonry

can safely be specified for supervised constnrction in New Zealand without any need for mate-
rial testing.
Eqn. 3.1 also provides a method for verification of f'n, using the characteristic block strength
provided by ttre manufacturer in conjunction with the characteristic grout strength found by
cylinder testing. Using Eqn. 3.1 in this way, it was recorlmended [3-7] to apply a strength
reduction factor of 0.75 to derive a lower-bound value.

The experimental data presented here was recorded in conjunction with PCM wall testing at
the University of Auckland. Results from testing of unconfined and confined grouted concrete
masonry were taken from wall test Series
I and2 described in Chapters 5 and 6.
Thble 3.2 shows the average strength of unconfined masonry prisms, grout and mortar measured for 10 wall tests. Generally, three prisms, three mortar cylinders and three grout cylinders
were tested for each wall. Prism construction was performed by registered masons using stand-
ard 140 mm precast grey hollow core concrete blocks that were provided by a range of manufacturers. The blocks were filled with standard 17.5 MPa target strength grout. Prior to
grouting, SIKA CAVEXTM shrinkage compensating agent was added to the grout. All test cylinders were 100 mm diam. and 200 mm high.
-24 -
It is observed from Table 3.2 that an average prism strength of at least 14.4 MPa was measured
in all cases. An average prism strength of
l8.l
MPa and a standard deviation of 2.3 MPa was
found based on the total number of prisms tested. Assuming normal distribution of the
strength, a 5 percentile strength for
each
f'-
of
14.3 MPa is found; thus the individual averages
for
wall and the 5 percentile strength based on all wall tests exceeded the minimum charac-
teristic design strength of 12.8 MPa derived from Eqn. 3.1 above. It is noticed that there is considerable scatter in the measured grout strength, ranging from 11.1 MPa to 35.0 MPa.
It is of
some concem that the measured grout strength in 4 out of 8 cases fell well below the code min-
imum specified strength of 17.5 MPa, even with consideration of the presence of the shrinkage
compensating agent. The mortar strength, f'.;, also exhibited considerable scatter; however the
measured minimum average
minimum strength of
of
10.5 MPa was in reasonably good agreement with the expected
12 MPa.
Interestingly, little scatter was measured in the prism tests, despite the large amount of scatter
measured for both grout and mortar strengths. This may suggests that unconfined masonry
strength is inlluenced more by the block strength than the grout and mortar strengths.
Table 3.3 shows results from testing of 9 prisms.
All
prisms were constnrcted
of
15 series
masonry blocks (140 mm thick) using 10 mm bed joints with blocks, mortar and grout from the
same batches. The series consisted
of 3 unconfined prisms with 190 mm block height (U200),
3 confined prisms with 190 mm block height (CP200) and 3 confined prisms with 90 mm
block height (CPl00). The confining plates were fabricated from 25 mm wide and 3.1 mm
ABLE 3.2-Measured cru

Wall Test
ft m
prism
frg
fi
cylinder
cylinder
l.l
16.6
20.5
23.7
26.1
20.6
t4.4
I8.8
13.9
0.9
18.2
8.6
3.5
17.8
t2.3
4.9
15.9
t6.2
0.5
35.0
l0
19.0
23.6
14.6
mean
18.l
t7.5
st. dev.
2.3
8.0
5.6
MPa
MPa
MPa
-25 -
, s8o -{rl
l_l
Kl
El
lr--lT-tl-l
,5
ll
-'n<-thick
3.1 mm
mild steel
conf ining plotes
nconf ined
u200
Conf ined
Conf ined
CP2OO
CP1 OO
Conf inin g
Plotes
Fig.3.4-Prism and confining plate dimensions and specification

thick mild steel. Fig. 3.4 provides details of nominal geometry and specifications of the prisms
and confming plates. Compressive strains were measured with four displacement transducers
attached to the prisms using a gauge length of 400 mm.
It is seen from Table 3.3 that the con-
fining plates did not increase the masonry crushing strength considerably.
TABLE 3.3-Confined prism testing

Prism
u200
CP2OO
cPl00
fm
e**
c**
Ynu
E-*
Em/fm
7.6
0.00209
6229
922
7.7
0.00218
1755
8.1
0.00218
3492
664
745
mean
7.8
0.00215
382s
777
2256
t32
st. dev.
0.3
0.0000s
t7.5
0.00262
2190
697
t6.4
0.00243
0.00653
2405
756
t7.6
0.00251
0.00701
2690
72r
mean
t7.2
0.00252
0.0067't
2428
725
st. dev.
o.7
0.00010
0.00034
251
30
t9.2
0.002M
0.02820
884
20.0
0.00242
0.0 195 I
t6975
ts442
18.7
0.00231
0.01681
14847
794
mean
19.3
0.00239
0.02151
15755
817
st. dev.
0.7
0.00007
0.00595
1098
59
772
strain
MPa
*strain
at peak force, **strain
Mortar strength: 14.9 MPa; Grout strength 12.3 MPa,
Fbased
on slope in 0.05-0.33f* interval
at approx. 50% loss ofstrength,
MPa
strain
-26 -
3.2.4 Australian approach

According to the Australian masonry standard AS 3700-1998 [3-4], the designer has two
options for specifying the masonry design strength:
(l)
use
f',
derived on a statistical basis and
related to the utilised block type or (2) derive f'n' from prism testing.
AS 3700 associates f''n with the 'Unconfined Compressive Sfrength', in the Australian context
meaning the compressive strength found by testing of slender prisms with a height+o-thickness
ratio of 5. Prisms of such proportions are deemed unaffected by the confining effect of the test-
ing machine platens. Compressive strength for prisms of lower height-to-thickness ratio are
converted to the oUnconfined Compressive Strength' by means of code defined aspect ratio
factors (Appendix C in AS 3700). Note that the expression 'unconfined' refers to plain grouted
concrete masonry (no confining plates embedded) in the general context of this thesis; only in
the context of AS 3700 is the expression related to testing of slender prisms.
Statistical base: Knowing the characteristic unconfined strength of the masonry units, fu", and
the bedding type, AS 3700 section 3.3.2 is applied to determine the characteristic strength (5
percentile) for ungrouted concrete masonry. The manufacturer generally provides
f* based on
continuous sampling. There is no direct dependency on the grout properties, however it is spec-
ified that the characteristic grout strength, f"r, must be higher than 12 MPa and lower than
[.3fu". Using fu" of l5 MPa, a level of strength often encountered in New Zealand, and typical
concrete masoffy dimensions,
f, :
8.0 MPa is found. Using AS 3700 section 7.3.2, defining
the axial strength of grouted masonry based on
f',
and
f'.*, a grouted
concrete masonry
sfength of 6.0 MPa was calculated. That strength falls short of both the minimum characteris-
tic strength for NZ masonry derived above of 12.8 MPa and the NZS 4230 grade B masonry
strength of 8 MPa. It is noted that the aspect ratio factor defined by AS 3700 is approximately
0.9 for typical NZ prism dimensions, thus not explaining the low AS 3700 compressive
stength estimate.
Prism testing can be used to establish the characteristic 5 percentile strength,
f-.
It appears that
prism sampling can be done by the contractor/mason and that the sampling does not have to be
directly associated with the project in mind. It is also stated that for wall strength verification,
i.e. testing of prisms made on site during construction, the testing average strength should be
used for verification. In that case, with only a few prisms tested, the most likely strength (the
average) is used instead
ofthe characteristic strength (5 percentile).
-27 -
3.2.5 USA approach

The Masonry Standards Joint Committee, constituted by the American Concrete Institute
(ACD, the American Society of Civi! Engineers (ASCE) and The Masonry Society (TMS),
proposes an approach [3-8] for specification of the masonry compression strength similar to
that of AS 3700. There are two options:
(l) to specify
f',
from tabulated values that depend
solely on the concrete block strength and the mortar type (Table 2,page S-10). It is required
that the grout strength at least equals f,n and exceeds 13.8 MPa; (2)
f'^may
be detennined
from prism testing. Specifying mortar type N and a characteristic block strength of 15 MPa
results in a design compression strength of 10.3 MPa. That strength exceeds the tabulated pre-
dictions of both NZS 4230 and AS 3700, but falls short of the minimum characteristic strength
of 12.8 MPa for NZ grouted concrete masonry.
3.3 ELASTIC MODTJLUS

In the context of seismic design, it is important to accurately estimate the stmctural stiffiress.
This requires precise knowledge of the elastic modulus, E,o, of the masoffy.
A high estimate of
En, results
in a low estimate of the fundamental period of vibration and
therefore generally higher seismic design loading, while the corresponding structural displace-
ments are underestimated. The result is a conservative estimate
of seismic loading but
an
unconservative estimate of structural displacements. A low estimate of E,n results in the oppo-
site'effect, i.e. overestimation of struchrral displacement and underestimation of the seismic

actions. Not only does
E- influence the flexural
response,
it
also governs the prediction
of
shear deformation.
Unbonded PCM walls subjected to in-plane seismic loads are expected to develop a single hor-
izontal crack along the wall/foundation interface and no significant cracking above, suggesting
elastic response of the panel above the hinging zone adjacent to the foundation. The stiffiress
of
the uncracked PCM panel therefore depends almost solely on E,n, while e.g. a reinforced
masonry panel with extensive cracking is less sensitive to
En.,
because of the considerable
influ-
ence of the reinforcing steel.
3.3.1 Current New Zealand codification

The current New Zealand masonry code NZS 4230:1990 simply specifies that E.o :25 GPa,
irrespective of masonry characteristics, unless established otherwise by testing. In reality this
-28 -
clause overestimates Em almost by a factor of two for grouted concrete masoffy. The value was
set this high deliberately to ensure conservative estimates of the seismic design forces.
In addi-
tion, little emphasis was placed on accurate calculation of displacements at the time of writing
of the code.
3.3.2 Literature and other
codes
Extensive research has been carried out on establishing appropriate expressions for E*. It was
found that the elastic modulus is not clearly related to any property of block, mortar, grout or
prism dimensions, but influenced by all of these. For simplicity
En.,
is generally related to
f',
because the crushing strength is also influenced by these parameters. Examples of recommen-
dations and code equations are:
E, :
900f;
AS 3700-1998 [3-4]
(3.2)
E^:
900f;
MSJC 2002 [3-81
(3.3)
1000f
;
E^ : 7s0f;
Paulay and Priestley [3-9]
(3.4)
UBC 1997 [3-10]
(3.5)
E, = 750f ;
NEHRP 1997 [3-l
E^
E^
: 850f;
l]
(3.6)
CSA S304.1-M94 [3-12]
(3.7)
The above formulae were generally developed to reflect the elastic modulus in relation to
measured prism strength.
It is thus incorrect to use the specified masonry strength (5 percen-
tile) for evaluation of E* because the strength likely to be encountered (mean strength) is
expected to exceed the specified strength.
A typical
5 percentile strength to mean shength ratio
is 80%. It is also mentioned that the formulation of En' is sensitive to the testing procedure, i.e.
the stress range on which the elastic modulus is based. The typical stress range is 0.05-0.33f'-.

The experimental data shown in Table 3.3 represent typical values of
E. for grouted
concrete
masonry in the Auckland region. The results, measured in the stress range of 0.05-0.33f,o,
suggest little effect of confinement on the
E./f',
ratio. Based on the above equations and the
experimental data in Table 3.3, the following relationship based on the actual crushing strength
is proposed for New Zealand conditions:
-29 -
E. : 800f; (mean)
(3.8)
When using the characteristic value of
f'., it follows from the discussion in section 3.3.2 that
the best estimate of the elastic modulus would be:
E,
800f
(char)/O.8
1000f; (char)
(3.e)
3.4 MASONRY STRAIN CAPACITY

Accurate estimation of the ultimate masonry strain capacity is important for unbonded PCM
walls that are expected to undergo large in-plane structural displacements due to seismic
action.
It is shown in Chapter 9 that structural
displacement limitations are likely to govern
ductile seismic design of PCM walls, e.g. limitations to the overall structural drift angle and
interstorey drift angle.
The masonry crushing strain is highly influential on the wall lateral force-displacement charac-
teristic. lncreased strain capacity normally results in slightly higher wall strength due to short-
ening of the compression zone, but more importantly
it
most likely
will
promote higher
displacement capacity because of larger curvature capacity of the hinging zone.

The definition of ultimate strain capacity related to flexural response can be differentiated from
the definition related to uniaxial loading.
It is clear that a uniaxial test unit reaches the maxi-
mum strength when all masonry fibres in the cross section attain
e,n
simultaneously (refer to
Fig. 3.1). Structural testing of PCM walls subjected to in-plane flexural loading generally indicated that the overall maximum
wall strength was achieved for masonry strains in the extreme
compression fibre beyond er. In fact the maximum wall strength was often associated with a
peak strain larger than the code defined ultimate strain (often termed omaximum usable strain')
listed in Table 3.4.It is therefore argued that the ultimate masonry strain, e-u, should be based
on results from wall flexural testing.
3.4.1 Code rules and recommendations

Table 3.4 summarises a series of design code requirements and recommendations for the ultimate masonry strain to be used for the design of reinforced concrete masonry. The values for
unconfined masonry vary between 0.0025 and 0.0035, depending on the source, with the New
Zealand codified value of 0.0025 as the lowest in the field. The two estimates of ultimate strain
for confined masonry of 0.008 originate from the silne research [3-2]. It is noted that the strain
values given in Table 3.4 are related to flexural action and nominal flexural strength, and are
-30-
therefore higher than the strain values associated with the masonry peak shength found by
uniaxial compression testing, as suggested in Fig. 3.1 and Table 3.3. It appears that Australia
and New Zealand should consider both moving to ,nu :0.003 for unconfined masonry,
facili-
tating harmonisation and providing consistency with other international codes, and with the
New Zealand concrete code NZS 3101:1995 [3-13].
TABLE 3.LUltimate compression strain

Source
NZS 4230:1990
AS 3700-1998
Unconfined
Confined
u200
cP200
0.0025
0.0080
[3-l]
t34l
0.0035
MSJC 2002 [3-8]
0.0025
uBc
1997 [3-10]
NEHRP 1997 [3-l l]
0.0030
Paulay and Priestley [3-9]
0.0030
0.0080
strain
strain
0.0025
3.4.2 PCM testing results

Table 3.5 shows results from testing of 6 unconfined PCM walls (refer to Chapter 5), reporting
the vertical strain in the exheme fibre at nominal flexural strength l\do, at first observation
distress and at maximum strength, Vr"*.
It is seen that the strains at nominal
of
strength core-
spond well with the values in Table 3.4 with an average strain of about 0.003. At first distress
(initiation of vertical splitting cracking of block face shell), the strain values had doubled and
at V,no the values had attained at least 0.01. It is noted that the values given in Table 3.5 were
measured over the lowest 100 mm above the foundation, thus some degree of confinement
of
the first course of masonry was caused by the reinforced concrete foundation.
Confined masonry is expected to sustain even larger strains at maximum strength. Table 3.3
shows that uniaxial testing of CPl00 resulted in a post-peak stength of 50% of f'n. for r,ou:
TABLE 3.S-Vertical masonry strain. test results

at nominal
shength, Mn
at first distress
at maximum
strength, Vro
I
a
0.0025
0.0027
0.0067
0.0137
0.0039
0.0104
0.0104
0.4027
0.0104
0.0104
0.0033
0.0074
0.0224
0.0031
strain
shain
strain
- 3t -
6mu
Stroiri prof ile
0.85f;
NZS4230:1 990
0.85c
1.5f
AS5700- 1 998
0.80c
0.80f;
MSJC droff
Rectongulor stress block

Fig. 3.S-Flexural stress block
0.021 (mean); indeed a very large strain capacity for a concrete material. Results from testing
of confined CP200 PCM walls showed strains measured over the lowest 100 mm of 0.01 to
0.02 at
Vrr"
and testing of confined CP100 PCM walls showed average strains measured over
the lowest 200 mm of 0.025 at Vn'* (refer to Chapter 7). A discussion of maximum useable
strain for ductile PCM wall design is given in Section 8.1.7.
3.5 FLEXURAL
STRESS DISTRIBUTION
Despite general consensus on the definition of the uniaxial masonry compression strength,
f',.,.,,
notably when derived from prism testing, there is considerable variation of the definition of the
stress distribution at nominal flexural strength. Fig. 3.5 shows three code definitions of the
equivalent rectangular stress block for unconfined concrete masonry related to f"n and the neu-
tral axis position, c. The MSJC stress block is specified for in-plane design, while no distinction between in-plane and out-of-plane design is made in NZS 4230 and AS 3700.
The use of larger rectangular stress block parameters leads to a reduced neutral axis depth,
resulting in a larger flexural strength and a larger curvature capacity. The AS 3700 stress block,
l.3f'., appears mainly to be associated with out-of-plane response. It is understood that the term 1.3f'. stems from the Australian concrete design code provisions and
with a height of
reflects the conversion from 'Unconfined Compressive Strength' to test cylinder strength using
32
a height-to-thickness ratio of 2. Given the origin of this tenn, the suitability of the AS 3700
flexural stess block for limit state design of grouted concrete masonry is questionable. The
American and New Zealand stress block parameters arc comparable, and in reality predict sim-
ilar behaviour. In-plane testing of PCM walls reported in Chapter 5 supports using the NZS
4230 or MSJC definition.
3.6 TI{EORETICAL STRESS-STRAIN RELATIONSHIP

Further to the above discussion, accurate modelling of the concrets masonry constitutive rela-
tionship is essential for detailed calculation of the global response
of PCM walls with
unbonded tendons. This material model should be of continuous nature to allow for capture
of
the wall response in the service load regime (elastic response), at nominal wall strength and at
large displacement response, including possible strength degradation.
The NZ and US design code approaches for calculation of concrete masonry response are concemed with the elastic modulus and with the equivalent rectangular masonry stress block and
associated ultimate strain capacity for calculation of nominal strength. These parameters do not
provide sufficient information for establishing the concrete masonry constitutive relationship
because only the initial elastic response and the nominal flexural response are captured. In
order to select a suitable constitutive relationship, a literature review on concrete masonry

stress-strain behaviour was conducted. Amongst the various research efforts in the past decades, the
following two contributions stand out:
A study by Drysdale and Hamid [3-14] considered the behaviour of concrete block masonry
under axial compression and provides a good description of the masonry compression failure
mechanism in the context of allowable stress design. However, it does not offer essential infor-
mation on critical sffain parameters necessary for ultimate strength calculations, which are
essential for capturing the wall ultimate response, such as ultimate stain capacity and strength
degradation characteristics.
A concise formulation of the stress-strain relationship for concrete masoffy was established by
Priestley and Elder [3-2] in the context of strength design. This relationship, covering both
unconfined and confined grouted concrete masoffy, was based on prism testing and the KentPark stress-strain curve for concrete [3-15]. The Priestley-Elder formulation is considered suitable for the present PCM wall study because it describes concrete masonry behaviour under all
loading conditions and is based on laboratory testing utilising New Zealand materials. Addi-
-33-
a-
Izo
qtfined prisms
vl
t4
trJ
R
tn
STRAIN I%)
Fig. 3,G-Priestley-Elder compression stress-strain curves [3-21
tionally, the Priestley-Elder formulation reflects the stress-strain relationship for both low and
high strain rates. Fig. 3.6 shows the experimental and theoretical stress-strain curves for unconfined and confined masonry, derived by Priestley and Elder.
3.6.1 Unconfined concrete masonry

Unconfined concrete masonry stress-strain behaviour is shown in Fig. 3.6. It is seen that the
experimental relationship (solid line) is nearly linear elastic to a stress of 0.5f'*, followed by
softening before the maximum strength,
f', :
26.5 MPa, is reached at a strain of approxi-
mately 0.0015. Subsequently, rapid strength degradation to approximately 0.2f'* at a strain
of
0.004 is observed.

Prisms without confining plates failed by separation of the masonry block face shell, resulting
in rapid strength degradation l3-zj.Vertical splitting initiated shortly before reaching maximum strength, prior to separation of the face shells. This failure mechanism was attributed to
incompatibility of the material properties of the concrete units, grout and mortar. The generally
lower strength of the mortar relative to the masonry units and grout, causes the mortar in the
-34-
bed
joints to expand laterally. Lateral expansion of the mortar towards the cente of the prism is
inhibited by the grout, resulting in a net displacement of the mortar towards the outside of the
prism, thus pulling the masonry face shells away from the grout core and instigating premature
failure.
Priestley and Elder concluded that vertical reinforcement included in the grouted cells did not
signifi cantly affect the experimental behaviour.

The theoretical stress-strain curye for unconfined masonry (dashed line in Fig. 3.6) is defined
by Eqns. 3.10 and 3.1l, using the notation adopted in this document. A parabola governs the
loading branch up to the maximum strength
f', at a strain of e. :0.0015. Beyond maximum
stength, the masonry strength decreases linearly to a plateau of 0.2f'. at a strain of
enp, as
dic-
tated by the slope 2n'.
f^(e)
: t o67f;
t# - (o#r)'] , ,. ,,,
f^(e) :
f; lI -Zm(E- 0.0015)l , t^St<tnp
f^(e):
o-zf;
(3.10)
, E)E^P
The slope of the descending branch is given by:

0.5
Z^:
. E^r: 0.8
an0
T+8,
'
zzm
Eqn. 3.10 directly implies an initial elastic modulus
(e:
(3.1r)
0) of E^ and an average elastic modu-
lus for the stress range 0-0.33f'- of En,:
E^ = 1067f;
Em
(3.r2)
976f;
(3.13)
Fig. 3.7(a) schematically defines the theoretical stress-strain curve parameters for unconfined
masonry.
-35-
1.O67Kf
o
q,
n
o
q)
v',
tn
'^
o.2x1.067Kfl
tm
tm
tm slroin -P
(a) Unconfined
masonry
slroin
Stroin
(b) Confined masonry
Fig. 3.7-Theoretical stress-strain relationships, defi nition of symbols
3.6.2 Confined concrete masonry

Confinement of concrete members is uzually achieved by means of closely spaced steel reinforcement hoops (spiral or square) [3-15]. In a retrofit situation, conlinement can be achieved
by external steel or fibre (glass or carbon) jacketing. Neither of these options appear to be
appropriate for confinement of concrete masonry.
It is difficult to incorporate reinforcement hoops in ordinary concrete block masomy construction because of the modular geometry. The hoops would necessarily have to be placed in the
bedjoints, where it is difficult to find the required space considering
10 mm- Placing steel reinforcement hoops
typical joint thickness
of
in the grout core would result in ineffrcient confine-
ment because confinement of the grout core would not restain the masonry block face shells
and prevent premature failure by spalling of the face shells.
It was decided to revisit the
use
of confining plates of the type shown in Fig. 3.4. This tech-
nique was developed for concrete masonry in the early 1980s l3-2). The technique has not been
widely used, mainly due to the increased complexity of construction. Ductile reinforced concrete masonry walls subjected to in-plane loading generally develop a large plastic hinge zone
with a vertical extent of the order of 50% of the wall length. Confining plates are therefore
potentially required over the entire first storey. Fortunately, a much smaller confining zone is
required for PCM walls because the unbonded tendons and associated rocking response tend to
concentrate plastic deformation in a short zone just above the wall base, with a length of the
order
of I
meter.
It
follows that the shorter plastic zone results in higher stain demands
imposed on the lower corners of an unbonded PCM wall.
-36-
As stated above, efficient confinement of concrete masonry can be achieved by embedding of

steel plates in the horizontal bed joints dwing construction as shown in Fig. 3.4. As a result
of
the confining effect of the embedded steel plates, the concrete masonry behaves in a ductile
manner as shown in Fig. 3.6
l3-2l.It is seen from
the figure that the measured stress-strain
curye (solid line) is nearly linear elastic in the 0-0.5f'. stess range, followed by softening
before reaching the maximum strength of 32 MPa at a strain of approximately 0.0020. Subse-
quently gradual strength degradation to approximately
0.2f', at a strain of approximately
0.012 is observed. It is noted, that the prisms used to established the experimental curves for
both the unconfined and confined masonry were constructed at the same time of blocks and
grout from the same batches. Comparing the strength of the confined prisms with that of the
unconfined prisms, a strength enhancement of approximately 1.2 is found.

Including confining plates in the bed joints changed the prism failtue mode. Vertical splitting
was
virtually eliminated due to constraint of the mortar by the confining plates. Ductile failure
resulted from gradual strength degradation associated with a shear-compression tlpe of failure,
generally occurring within one masonry course.

The theoretical stress-strain curve (dashed line in Fig. 3.6) is defined by Eqns. 3.14. A parabola
govems the loading curve up to the maximum strength
1.067Kf, at a shain of e-:0.002K.
Beyond maximum strength, the masonry strenglh decreases linearly
to a
plateau of
0.2x1.067Kf, at a strain of gnp, as dictated by the slope Zr.
r^(e):
where
r.067Kr;l#--(o#*)'] ,
f^(e)
1.067
f^(e)
0.2
Kf; [l - Z^(e- 0.002J()] ,
x 1.067 Kfl
E1E^
En,3s1Enp
(3.14)
, E)Erp
f', is the unconfined strength related to testing of unconfined prisms. This equation is
K that accounts for the prism strength
analogous to Eqn. 3.10, except for the factor

and strain increase due to the confining plates:
-37
increase
K: 1 *rr*
(3.15)
where p, is the volumetric confining ratio and
Normally the transverse confining ratios,
pu,
tr, is the yield strengttr of the confining
steel.
for confined rnasonry in the two orthogonal direc-
tions are different due to the masonry unit dimensions. Priestley and Elder therefore proposed
to use:
2A^
Pr: 2Po: 6
(3.16)
where p" is the minimum ratio of the two directions and Ao is the area of the confining plate cut
by a vertical section of h"xSn flength by spacing, respectively). The slope of the descending
branch is given by:
. E^p:
anc
0.5
zn
0.8
TnE,
'm
In the Priestley-Elder study it was calculated from Eqn. 3.16 that pr:0.00766 where
(3.r7)
h":
390
mm (long direction of prism), 56:200 mm and f1,r:316 MPa were used.ln Eqn. 3.17, Z',, was
based on the values defined above, except h" was taken as the short dimension of the prism,
140 mm.
Eqn. 3.14 directly implies an initial elastic modulus

lus for the stress range 0-0.33x1.067Kf', of
E^:
(e:
0) of E* and an average elastic modu-
E,:
1067f;
(3.18)
E. : l02lf;
(3.1e)
where
f'*
is the unconfined prism strength. Fig. 3.7(b) schematically defines the theoretical
stress-strain curve parameters for confined masonry.
3.6.3 Evaluation of stress-strain relationship

A series of nine masonry prisms were tested
as described in section 3.2.3. Prism and
confining
plate dimensions are shown in Fig. 3.4 and the testing results are shown in Table 3.3. Continuous stress-strain curves were measured in all of the tests. The strain measurements were relia-
ble for strains up to e-, corresponding to the maximum strength. Beyond this point, the failure
-38-
Unconfined U200
Confined CPl00
Confined CP200
F'ig.3.8-Ilpical failure
modes for prisms
pattern did in some cases influence the measurement readings. The explosive failure mode
of
the unconfined prisms resulted from rapid release of elastic strain energy stored in the testing
machine, that could not be resisted by the prism. Ductile behaviour of the confined prisms was
experienced, i.e. the prism were able to absorb the elastic strain energy stored in the testing
machine, that was released after reaching prism maximum strength.
Fig. 3.8 illustrates the typical failure modes for the unconfined and confined prisms tested. The
failures resemble those described above 13-21: U200 exhibited splitting failure and CP200
exhibited shear/compression failure that occurred within one masonry course. The highly confined prisms, CP100, failed as a result of weld fracture of the confining plates, allowing for lat-
eral expansion of the masonry and initiation
of shear/compression failure. Failure for
the
CPl00 prisms extended over one or two half height courses.
It is noted that the prism
tests conducted by the author had a height of 600 mm (140 mm thick)
and that those tested by Priestley and Elder had a height
of
1000 mm (190 mm thick). The
shorter prism height used by the author may have caused some degree of confinement due to
end plattern restraint with the expectation of some augmentation of maximum strength and
-39-
20
/x
16
-Elperimental
-Theoretical
- average
\
a
2p
o
gt
o
0+0.003
0.000
0.004
0.005
0.008
Strain
Fig. 3.9-U200, Unconfined masonry stress-strain curve

associated strain in comparison with those
of
1000 mm high prisms
of similar material. This
may also affect the unloading branch of the prism stress-strain curve with the expectation
increased e,no. Comparison
of
with the Priestley and Elder research should therefore be done with
the above observation in mind.
3.6.3.1 U200, Unconfined prisms

Fig. 3.9 shows a typical measured stess-strain
flrve for unconfined masonry
prisms. Only the
rising branch of the curve was measlued reliably because of explosive prism failure. A maximum strength of 17.8 MPa was measured at a strain of 0.0021. The theoretical stress-strain
curye is also plotted in Fig. 3.9 and appears to describe the rising branch of the stess-strain
curve reasonably accurately. The greatest deviation is seen at maximum strength, where the
strain theoretically attains 0.0015 while the experimental strain amounts to 0.0021. This discrepancy may be caused by plattern restaint for 600 mm high unconfined prisms resulting in
some confinement. The theoretical curve predicts that the falling curve reaches the residual
plateau of 0.2f'* at ,np :0.0066. The experimental value for e*o of about 0.0052 is associated
with a high degree of uncertainty and cannot be relied on conclusively. As noted above, confinement effects from the plattern restraint may have increased the measured maximum
strength.
-40-
In the Priestley-Elder study an average compressive strength of 26.5 MPa was found, and was
used for developing the above formulae. Fig. 3.10, showing
function of
f'.,
predicted by Eqn.
3.il
as a
reveals drastic increase of enp for lower masonry strength when compared to
the value associated with
f',
:26.5 MPa. It is the opinion of the author that Eqn. 3.1 I exagger-
ates the masonry strain capacity for unconfined concrete masonry
MPa. For example it is found that
e*o:
0.0083 for
f'-:
of strength of
16 MPa, a strength
less than 25
typically encoun-
tered for NZ North Island concrete masonry. The explanation for prediction of such high values of s,np can be found in the origin of the equation. The Kent-Park curve was determined
from concrete research which rarely deals with specified strengths of less than 30 MPa. The
equation, having been calibrated for compressive strengths higher than 30 MPa, therefore only
serves as an extrapolation for compressive strength below 30 MPa. Consequently,
posed to
it is pro-
limit tro to 0.0040 (approximately the value predicted for f',n :26.5 MPa) for uncon-
fined masonry in this document, a realistic value for masonry compressive strength below 26.5
MPa.
Both Eqn. 3.9 and the prism testing results reported in Table 3.3 support the elastic modulus
given in Eqn.3.l3, although a discrepancy of about 20o/ois found between Eqns. 3.8 and 3.13.
It is mentioned that the 0.2f'. plateau may not be appropriate for unconfined masonry. Both
these deficiencies are, howwer, expected to not affect prediction
of flexural behaviour signifi-
cantly.
With due consideration to the limited amount of stess-strain data available from prism testing
of North Island concrete masonry using pumice aggrcgate, the Priestley-Elder theoretical
stress-strain relation is found acceptable for detailed analysis of unconfined PCM.
3.6.3.2
CP200o Confined prisms
Fig. 3.11 shows a typical measured stress-strain curve for confined masonry, using the confining plates shown in Fig. 3.4 (p,
: 0.00577, fy1.,:240 MPa). Shess and shain measurements
for
the loading part of the curve are associated with little uncertainty. The accuracy of the unload-
ing curve was, howeveq uncertain because masonry crushing and associated strength degradation occurred over approximately one block height, corresponding to approximately 0.25 to 0.5
times the gauge length of the strain measurement. The post-m strain measurements shown in
Fig. 3.11 were corrected according to the observed location and extent of failure.
-4t-
S
EG
0.0,,
art
e
E
o.oos
taronry Comprcsslve Strangth (llPa)
Fig.3.l0-Variation of masonry strain qnp with masonry strength
PdesfeytEHor
Ptiostley-Eldor
,/
,
r{\t./
| -uoamea
|
I
//
16
!l
rz
o
o
-Experimental
- average
-Theoretical
I
\'
0+0.000
0.006
0.008
0,010
0.012
0.016
Strain
Fig. 3.11-CP200, Confined masonry stress-strain curre

Fig. 3.1I also shows the theoretical stress-strain relationship predicted by the Priestley-Elder
equations using
f'. :
17.8 MPa determined from the U200 prisms. As suggested by the figure,
the predicted shength enhancement due to the confining plates was not achieved experimen-
tally. This observation can to some degree be explained by using
-42-
f'*
from the U200 prisms. It
was stated above that the testing machine plattem may have provided a confining effect to the
U200 prisms. In that case the measured U200 strength was superficially increased relative to
the true unconfined strength. With that in mind
it
cannot be excluded that CP200 confining
does result in strength enhancement.
In terms of the unloading branch, the experimental curve follows the theoretical curve reasona-
bly closely within the margin of error of the strain measurements. For better approximation of
the experimental result, the theoretical Priestley-Elder stress-strain curye was modified by
using
Kf', : f'.
in Eqn. 3.14 (K retained in other equations). The resulting curve is shown in
Fig. 3.11 labelled 'Priestley-Elder Modified' and reveals a better fit with the average experimental cun/e.
The initial elastic modulus given by Eqn. 3.19 appears to be approximately l7o/o higher than
the value found by experiment and listed in Table 3.3. Note that
maximum strength measured while
f',
f', in Eqn. 3.19 is the unconfined
in Table 3.3 denotes the

prism strength.
3.6.3.3 CPl00, Confined prisms

Further confinement of the concrete masoffy was achieved by using half height masonry
blocks with confining plates spaced vertically at 100 mm, as outlined in Fig. 3.4
f9:240
(p,:0.0[
154,
MPa). The experimental and theoretical curves for CPl00, shown in Fig. 3.12, exhibit
trends similar to those of CP200: exaggerated theoretical shength enhancement but reasonably
good correlation between the falling branches

Assuming little strength increase, by taking
of the experimental and theoretical
Kf'-: f'.
curves.
in Eqn. 3.14, resulted in generation of
the curve labelled 'Priestley-Elder Modified'. It is seen that very good correlation between the
experimental and modified theoretical curves was obtained. It is also seen from Fig. 3.12 that
the predicted initial elastic modulus corresponds well with the theoretical cun/es, particularly
for the 'Priestley-Elder Modified' curve.
It is observed, given the confining plates used in the experiment, that the strength enhancement
predicted by Eqn. 3.14 was exaggerated by a factor of approximately K, but that the strain
capacity,
%., given by Eqn. 3.17 was in agreement with the experimental results. It is again
pointed out that
f'*
from U200 prisms may have been superficially enhanced due to restraint
by the testing machine plattern, making comparison wittr the Priestley and Elder theory less
conclusive.
-43-
20
tl
A\
16
Priestley-
TEldar
vl
3t,
o
Experimental - average
-Theoretical
tl
PriesUel rElder
("*":
t0
\
\
aD
\
0
0.000
0.016
0.032
Straln
Fig. 3.12-CP100, Confined masonry stress-strain curve
3.6.4 Effect of strain rate

The dynamic nature of seismic loading results in masonry strain rates in the order of 0.005
strain/sec [3-15]. The general trend is that a high strain rate results in higher masonry peak
strength and lower strain capacity. Priestley and Elder [3-2] determined an average strength
of l7% for a strain rate increase from 0.000005 strain/sec to 0.005 strain/sec for
unconfined as well as confined masonry. Experimental stress-strain curves for concrete
increase
masonry are normally obtained at low sfain rates.

For unconfined masonry, it follows that the theoretical peak strength at high strain rates, to be
used in Eqn.3.l0, should be taken as 1.17 times

increased l7%o. For confined masonry,
and
f-.
Zîn
Similarly,Z^ in Eqn 3.11 should be
Eqns. 3.14 and 3.17 should be replaced
with:
Kd: r.rz[r.r,t+]
7
Lmd
-
(3.20)
t.l7 x 0.5
(3.21)
lffiv;*oo].1'',F,-ooozKd
-44-
It is again emphasised that f'- is related to unconfined
masonry prism testing at low strain
rates.
3.6.5 Tensile strength

The tension strength of grouted concrete masonry is generally negligible in comparison to the
compressive strength, is associated with high uncertainty, and nomrally exerts little inJluence
on wall design in the context of limit state design. The tensile strength is therefore ignored by
most limit state design codes, e.g. NZS 4230:1990. Estimates for tensile capacity of grouted
concrete masonry subjected to in-plane flexural loading can be found
in
codes that allow
design to elastic theory (allowable stress design), e.g. MJSC [3-8], that defines the tensile
capacity (modulus of rupture) as f,
1.72 MPa
A tensile capacity of the order of l-2 MPa
appears to be appropriate for design.
3.6.6 Unloadingproperties
The unloading properties describe the unloading path following development of strains larger
than
, corresponding to the peak
stength,
f'r.
These properties may inlluence dynamic ana-
lysis of PCM walls, which may be subjected to load reversal many times during a seismic
event. There
is little information available on unloading
properties specific
to
concrete
masonry. The reader is therefore referred to section 9.I.2 for a brief discussion of unloading
properties adopted in relation with FEM analysis of PCM walls.
3.7 PRESTRESSING STEEL

High strength prestressing (HSP) steel is the preferred material for prestressing of concrete
masonry walls. HSP steel has considerably higher elastic strain capacity than low strength pre-
stessing steel, simply because HSP steel possesses a much higher yield stress than low
strength steel, while having similar elastic modulus. The higher strain capacity allows for large
elastic tendon elongation and makes the HSP steel prestress less sensitive to losses due to creep
and shrinkage of the concrete masonry.
Steel is not the only material option for prestressing of walls. Prestressing of masonry walls
using carbon fibre tendons was invesfigated by e.g. Sayed-Ahmed et al. [3-16] and Holden [317]. The investigation by Sayed-Ahmed et al. featwed testing of a prestressed clay masonry
wall and proved that carbon fibre tendons are a viable option for non-seismic design where
lir
tle displacement capacity is required. Holden investigated a prestressed concrete wall and sug-
-45-
gested that carbon fibre tendons are a viable option for seismic design when the carbon fibre
tendons are designed to remain elastic for the required displacement demand. Carbon fibre tendons behave in a linear elastic manner in the entire loading range until maximum shength is
attained, followed immediately by rupture. The material is inherently of brittle nature and gives
no warning before failure. HSP steel, on the other hand, is ductile and allows for plastic strain
far beyond nominal yield strain.
While emphasis in this document is put on tendons remaining elastic during maximum credible
ground shaking (design earthquake with approx. 500 year return period), tendon yielding can
potentially occur in the event of an extraordinarily strong earthquake (return period of more
than 500 years), thus the ability for tendons to deform plastically (yielding) will prevent sudden tendon failure, entail energy dissipation and therefore further protection of the structure.
In
addition, the focus of this thesis is on application of commonly used and commercially available materials. Therefore, only the use of HSP steel is considered for design of unbonded PCM
walls in this document.
3.7.1 Tlpes of high strength prestressing steel

Prestressing steel basically comes in two forms: as flexible strand and as rigid bars. In general
terms, flexible strands are appropriate for undulated prestressing profiles and long distanceso
and for high loads, as they typically have higher yield stress, whereas bars primarily are useful
for straight prestressing over shorter distances. Installation of strand solutions is more complex
than installation of bar solutions, because strands are anchored individually with wedges or the
equivalent and require specialised hydraulic jacks for stressing. Installation of bar solutions,
consisting of threaded bars, anchor plates and nuts, requires stressing the bars with a hydraulic
jack and tightening of the nul Alternatively, the bars can be stressed by tightening the nut with
a torque wrench or by using calibrated washers and a wrench.

The material properties of strands and bars vary considerably, as suggested in Table 3.6, which
presents typical material properties [3-
l8]. Fig 3.13
shows a typical stress-strain curve for high
strength prestressing steel. It is noted that there is no yield plateau as in the case for mild reinforcement. The curve is approximately linear elastic up to the yield strength f,n then the curve
bends over, gradually levelling out, until failure at the stress of f,u and corresponding shain
%u. The
of
yield strength fon provided by the manufacturers, is defined as the steel stress corre-
-46-
sponding to a non-proportional strain of 0. I %, i.e. the intersection of the loading curve with a
line offset of 0.1% strain (or 0.2Yo depending on applicable testing standard and material) and a
slope of Eo., refer to Fig. 3. I 3 . This stress is also called the 0. I % proof strength. Prior to calcu-
lating Qn the elastic modulus Eo, is found as the slope of the linear part of the loading curve.
TABLE 3.LMaterial properties of prestressing steel

rype
Strand
sl5mm
Bar s23mm
tu
%u
Eo,
r480
1750
3.5%
l 80-205
930
1080
60/o
t70
MPa
MPa
$t
GPa
3.8 CONCLUSIONS
It
was found that accurate prediction of f,.,, and e.u is necessary for design of PCM walls
because of the significant influence of these parameters on the
wall strength and displacement
capacity. NZS 4230:1990 needs to move towards a system for specification of
f-
based on the
characteristic strength of the blocks and grout, because the curent grade dependent approach
poorly reflects actual masonry sfrength. In fact, a characteristic strength
f', of 12.8 MPa can
safely be specified by designers in New Zealand without the need for verification through
material testing. Prism testing should, where possible, be employed to exploit the available
strength fully, which often exceeds 16 MPa for standard products.
Using confinement plates in the bed joints is an efficient way of improving the masonry strain
capacity. Uniaxial strain capacity of beyond lYocanreadily be achieved. The ultimate masonry
,Pu
?
'pv
1% Proof Strength
n
o
0)
t+
V)
Slope
0.001
Ep.
Stroin
tP,
Fig. 3.I3-Prestressing steel stress-strain relationship
-47 -
strain, e-u, should be related to the PCM wall flexural behaviour because the expected maxi-
mum wall strength is associated with an extreme masonry fibre strain far larger than the codi-
fied limits. Appropriate equivalent rectangular stress block parameters need to be established
for PCM wallso with due consideration to confinement.
It is recommended to use Eqn. 3.8 for prediction of E* for New Zealand conditions
based on
the likely masonry crushing strength, and Eqn. 3.9 when using characteristic strength.
Theoretical stress-strain curves may be established using the Priestley-Elder fomrulation. The
only parameters required to define the curves are the masonry geometry the expected unconfined strength,
f'r, and if present
the confining plate dimensions and yield strength- For uncon-
fined masonry, it appears that e-o is overpredicted for masonry strengths below 25 MPa, thus
q,o should be limited to 0.004. For low volumetric confining steel ratios, ps, the Priestley-Elder
formulation was found to over-predict the sfength enhancement, however, the predicted strain
capacity appears reasonable. The strength enhancement
servatively be taken as just
Kf'.
given in Eqn. 3.14 could con-
f*.
The tensile strength of grouted concrete masonry is unreliable and should be neglected for
limit state design.
It was concluded that it is essential to use high strength steel for prestressing of PCM walls
because
of the beneficial effect of the large elastic elongation capacity and reduction of the
time dependent losses.
-48-
3.9 REFERENCES
t3-11 NZS 4230:1990, Code of Practice for the Design of Masonry Structures,
Standards
Association of New Zealand,, 1 990, Wellington.
t3-21 Priestley, M.J.N. and Elder, D.M., Stress-strain curves
for
Unconfned and Confined
Concrete Masonry, ACI Journal, Vol. 80, No. 3, May-June 1983, pp.192-201.
t3-31 NZS 4210:2001,
Code of Practicefor Masonry Construction: Materials and Worhnan-
sftlp, Standards Association of New Zealand,l989, Wellington.
t34]
AS 3700-1998, Masonry Structures, Standards Association of Australia, Homebush,

NSW Australia.
t3-51 NZS 3102:1983, Concrete masonry units, Standards Association of New Zealand,
1983, Wellington.
[3-6] NZS 3109:1997, Concrete Constructio,n,
Standards Association of New Zealand, 1997,
Wellington.
13-71 Priestley, M.J.N., and Chai, Y.H., Prediction of Masonry Compression Snength, Part 2,
New Zealand Concrete Construction, April 1984,pp.21-24.
[3-8] MSJC: Building Code
Requirements
for
Masonry Structures and Specification
for
Masonry Stntctures, ACI 530-02/ASCE 5-02/TMS 402-02, The Masonry Standards

t3-91 Paulay, T., and Priestley, M.J.N., Seismic Design of Reinforced Concrete and Masonry
Buildings, John Wiley & Sons Inc.,1992,744p.
[3-10] UBC, Unifurm Building Code l99T,Intemational Conference of Building Officials,
Whittier, Califomia, USA.
[3-lll
NEHRP, Recommended Provisions

Other Structures, Part
for
Seismic Regulations
I - Provisions, Building
for
New Buildings and
Seismic Safety Council, 1997,Washing-
ton, D.C., U.S.A.
[3-12] CSA 5304.1-M94, Masonry Designfor Buildings (imit states design), Canadian Standards Association, Rexdale, Ontario, Canada.
[3-13] NZS 3101: 1995, Concrete Stntctures Standard Standards Association of New Zealand,
1995, Wellington.
-49-
[3-14] Drysdale, R.G. and Hamid, A.A., Behaviour of Concrete Block Masonry Under Axial
Compression, ACIJournal, Proceedings V. 76, No. 6, June 1979, pp.707-721.
[3-15] Scott, 8.D., Park, R and Priestley, M.J.N., Stress-Strain behaviour of Concrete Confined
by Overlapping Hoops at Low and High Strain Rates, ACI Journal, Proceedings, V. 79,
No. l, Jan.-Feb. 1982,pp. 13-27.

[3-16] Sayed-Ahmed, E.Y., Lissel, S.L., Tadros, G., and Shrive, N.G., Carbonfibre reinforced
polymer (CFRP) post-tensioned masonry diaphragm walls; prestressing, behaviour,

and design recommendations, Canadian Journal of Civil Engineering, Vol. 26, No. 3,
1999, pp.324-344.
[3-17] Holden, T.J., A Comparison of the Seismic Performance of Precast Wall Construction:
Emulation and Hybrid Approaches, Research Report 2001-04, Deparhnent of Civil
Engineering, University of Canterbury Christchurch, New Zealand.
[3-18] VSL Construction Systems, VSL Prestressing (Aust.) Pty. Ltd., Australia, 1995.
-50-
Chapter 4
TIME DEPENDENT EFFECTS
4.
ASSESSMENT OF PRESTRESSING LOSSES
Assessment
of prestressing losses is of great importance for prestressed concrete masonry
structures. The effective prestress level decreases over time due to creep, shrinkage and steel
relaxation, hence reducing structural efficiency. Research has indicated that preshess losses in
prestressed concrete masonry may nmge as high as25o/o of the
initial stress level, depending on
the masonry, the initial stress level and the prestressing hardware [4-1].
Losses are typically attributed to the following factors:
L Shrinkage
2. Creep
3. Relaxation of tendons
4. Elastic shortening
5.
Anchorage'draw-in' (seating losses)
6.
Friction due to undulation of tendons
7.
Thermal effects
Elastic shortening (4) occurs when a tendon is stressed due to the masonry elastic properties.
When more than one tendon is used in a wall and these are stressed one at the time, stressing
any tendon but the first one
will theoretically affect the tendon
of
stress in all previously stressed
tendons. Elastic shortening can readily be estimated using elastic analysis, but is expected to
have minimal influence for PCM because prestressing area ratios typically are relatively low
As a consequence, (4) is only discussed in this chapter as a benchmark for shortening in rela-
tion with creep strain. As this chapter primarily is concerned with post-tensioned concrete
masonry (PCM) walls with unbonded straight tendons, item (6) is a non-issue. Anchorage
draw-in (5) may be estimated according to the utilized prestressing hardware and will not be
- 5l
discussed further. Short term thermal effects (7) are disregarded in this study because of their
reversible nafure. Nevertheless, tendon stress change may arise from temperature difference
between the masonry and the prestressing steel, which may readily be evaluated. It is noted that
the issues 4 through 7 have little or no time dependency. Consequently, this chapter solely discusses the
time dependent issues affecting prestress losses in concrete masonry: shrinkage,
creep, and prestressing steel relaxation.
Equations for evaluating creep and shrinkage are presented, complying with current practice
for design of prestressed concrete and the masonry code provisions from the British standard
BS 5628 14-21, the Australian standard AS 3700 [4-3] and the North American code MJSC [441.
A creep and shrinkage experimental program conducted at the University of Auckland is
thereafter presented and compared with creep and shrinkage code provisions. Equations for
evaluating steel relaxation are presented, followed by current code recommendations. Lastly
recommendations are made for calculation of prestress losses under New Zealand conditions.
4.I
THEORETICAL CONSIDERATIONS
4.1.1 Creep and shrinkage of concrete

The following is a short description of the creep and shrinkage mechanisms in concrete (water
based) as reported by Neville [4-5]. It is emphasised that the description serves only to provide
sufficient understanding of the creep and shrinkage mechanism for development of relevant
equations presented below.
Despite the fact that creep and shrinkage are interactive processes and cannot be completely
dissociated, the effects of creep and shrinkage traditionally are dissociated for simplicity as
illustrated in Fig. 4.1. This is of course an extemely simplified description of a very complex
interaction between time dependent effects- The following shortening of a concrete specimen
is observed when subjected to sustained axial load:
Initial elastic shortening
The initial elastic shortening, denoted
'5' in Fig.
4.1, results from application of axial load at
time to and is often used as reference for describing the magnitude of creep and shrinkage. The
initial elastic shortening reflects the stress level applied to the concrete and the initial elastic
modulus of the concrete.
-52-
'I .
Drying creep
Bosic creep
3: Totol creep
4. Totol shrlnkoge
2-.
5:
lnitiol elosiic shortening
L
+
to
t1
Time i
Fig.4.l-Change of strain in axially loaded and drying concrete masonry
Shrinkage:
Shrinkage of drying concrete, denoted
'4' in Fig. 4.1, is chamcterised by volume change in the
cement paste and can be described by two phsnomena: (i) drying shrinkage and (iD shrinkage
due to carbonation.
(i) Drying shrinkage occurs when water is removed from the cement paste because of surrounding unsaturated air. Only water lost from the cement paste causes drying shrinkage. Free
water evaporating from cavities, etc., induces little or no shrinkage. Many factors affect drying
shrinkage. The physical dimensions, notably the surface to volume ratio, and the concrete
aggregate are highly influential.
Storage conditions play an important role. Neville [a-5] states that the relative humidity of the
medium surrounding the concrete greatly influences the magnitude of drying shrinkage, thus
the drying shrinkage strain of concrete stored in 50% relative humidity can amount to about
L50% of the drying shrinkage strain encountered at70%o relative humidity.
The issue of reversibility of shrinkage due to moisture movement is not considered here
because
it is assumed that PCM walls will not be subjected to drastic change of relative humid-
ity. This effect is only distinct if concrete, which has been allowed to dry in air with a given relative humidity, is subsequently placed in water (or subjected to nearly l00o/o humidity).
(ii) Shrinkage due to carbonation is caused by a chemical process where carbonic acid reacts
with CO2. The rate of carbonation depends on the moisture content of the concrete, the relative
-53-
humidity of the ambient medium and the size of the specimen. Carbonation increases the
shrinkage at intermediate relative humidities (RH), but not at high RH (100%) or RH below
35Vo, either because of pores
full of water prohibiting diffi.rsion of
CO2 or because
of insuffi-
cient water to sustain the chemical process.

Creep:
Creep is defined as gradual increase in strain with time under sustained load. According to
Neville [4-5], the nature of creep is still controversial, but is generally attributed to internal
pressure caused by external loads in absorbed and interlayer water within the microstructure
of
the cement paste. This pressure causes migration of the water and resulting volumetric change.
In reality, it is the cement paste that undergoes creep. The aggregate undergoes linle creep and
only affects the overall creep through it's function, primarily being that of restraint to the
cement paste. Fig. 4.1 schematically illustrates the change in strain of an axially loaded and
drying concrete specimen as a function of time. It is seen that the total creep is divided into two
components: (i) basic creep and (ii) drying creep.
(i) The basic creep, denoted '2' ln Fig. 4.1, is related to the moisture content in the concrete
paste that is available for evaporation (no moisture movement to or from the ambient medium).
A reduction of the available moisture content reduces the propensity for viscoelastic movement
and therefore results in lower basic creep.
(ii) The
second term, drying creep denoted
'l'
in Fig. 4.1, occurs when an axially loaded spec-
imen is not in hygral equilibrium with it's surroundings, meaning that moisture is transferred
between the specimen and the surrounding air; thus drying while under load increases the
creep of concrete. As concrete dries, the rate of creep reduces and approaches zero. One of the
most important external factors influencing creep is the relative humidity. The lower the
humidity" the larger the creep rate. At this stage both basic and drying creep occurs. At a later
age when the concrete has reached hygral equilibrium
with the surrounding medium, the creep
rate stabilises, and remains the same irrespective of relative humidity level. The latter creep
rate is attributed to basic creep because of hygral equilibrium.
It appears that the rate of creep is
little affected by ambient temperature when the fluctuations about room temperature (20 "C)
are small. Howeveq the temperature directly affects relative humidity.
Alternating wetting and drying may result in an increase of the magnitude of creep. This phenomena, termed 'wetting creep', suggests that results of laboratory tests may underestimate the
-54-
creep under normal weather conditions [4-5]. Bryant et al. [4-61came to the opposite conclusion from comparison of creep strains in concrete specimens placed
(l)
on top of a bridge deck
(alternating wetting and drying) and (2) inside the box girder of the same bridge (sheltered
environment). Bryant et al. found that creep strain after 600 days for
(l)
amounted to about
90% of the creep strain of (2).
Long term creep and shrinkage:
Fig.4.l
illustrates that at a certain age, t,, nearly all creep and shrinkage have occurred.
According to Neville [4-5] creep and shrinkage take place over long periods. Typically, the
shrinkage after
year amounts to about 7 5% of the shrinkage after 20 years.
If the shrinkage
stain at I year is taken as unity, the shrinkage strain after 20 years amounts to
413.
Similar
observation is made for concrete creep, where the creep shain after 20 years may amount to 4/
3 of the creep strain after
year. This has significant influence on the estimate of prestress loss
in a structure, when a service life of more than 20 years commonly is considered, given that
most creep and shrinkage experiments last
- 2 years.
According to Bryant and Vadhanavikkit [4-7], the final amount of creep and shrinkage strain
depend strongly on size of the concrete member. Thicker members, i.e. members with further
distance between the concrete core and the exterior surfaces, show slower creep and shrinkage
strain rates and lower long term creep and shrinkage strain, than thin members subjected to
similar stress level and environment. Modelling of the strain rate and long term sfiain dependencies of member dimensions can for example be found in the CEB-FIP model code [4-8] and
Eurocode 2 l4-9l.The ACI approach in [4-10] does not take concrete dimension into account
in estimation of creep and shrinkage properties.
4.1.2 Creep and shrinkage of concrete masonry

Creep and shrinkage of concrete masoffy, a material combination composed of water-based
cementicious materials, are expected to eventuate in a fashion similar to those of concrete as
described above. However, there are differences between concrete and concrete masonry, such
as the higher porosity
of concrete masonry, the fact that concrete masonry blocks are precast
and normally autoclaved (steam cured and moisture controlled) and the presence of mortar
joints. The nature of creep and shrinkage for firlly grouted masonry is expected to be similar to
that of concrete because of the presence of wet concrete (grout, block
cross section after grouting.
-55-
fill) in about 50% of the
It appears that drying shrinkage in concrete masoffy can be reversed to a considerable extent
because of the inherent porousness of the concrete masonry blocks, allowing water to easily
penetrate deep within the specimen. This can be prevented with a suitable water repellent coat-
ing.
According to Lenczner
[4-ll], who studied clay masonry, about 95% of all creep can be
expected to occur within
I year. This information
is not confirmed to apply for concrete
masoffy. As discussed above for concrete, perhaps 75% of the total creep and shrinkage (after
more than 20 years) for concrete occurs within
I year (strongly
dependent on member thick-
ness). For concrete masonry a larger proportion of the long term shortening is expected to
occur within one year because of thin dimensions and high porosity which lead to hygral equi-
librium with the surroundings at an early stage. This is in particular the case for the concrete
masoffy shrinkage. Consequently, nearly all shortening due to both creep and shrinkage takes
place within 5 years of construction and load application. This statement does not hold true
if
concrete masonry is subjected to extreme humidity variations which, as noted above, may
reverse the shrinkage process and may instigate wetting creep.
It is assumed that creep and shrinkage of concrete masonry can be identified as independent
and additive effects, as outlined in Fig. 4.1; a simplification of a very complex interaction
between the time dependent effects. It appears, however, that more complex theoretical consid-
eration is little warranted in the context of concrete masonry because of significant scatter
of
creep and shrinkage results from laboratory testing 14-121, and because of wide variation
of
material properties for concrete blocks, mortar and grout.
4.1.3 Mathematical expressions

This section considers the total prestressing loss due to creep, shrinkage and prestressing steel
relaxation, meaning the long term loss after say 20 years, when nearly all axial movement in
the concrete masonry and all relaxation of the prestressing steel have taken place. The initial
prestressing force after anchorage
lock-ofi Pi, and the long term prestressing force, P1, iu
important quantities to the structural designer, who must ensure adequate structural capacity at
any given time in a structures service life. The variation of the prestressing force with time is
not discussed in this section because such detail is not warranted for this study. However, the
implications of time variation of shrinkage, creep and relaxation are briefly discussed in section 4.5.2.
-56-
4.1.3.1 Creep strain

The total creep strain may be expressed by Eqn. 4.1.
creep strain
It is assumed that
q, is proportional to the initial elastic defonnation
en.,i
the concrete masonry
by the factor
C'
called the
'creep coefficient'.
8",
C"E^r
nfrr
-- wcF
um
(4.1)
In this equation, [6 is the initial axial stress on the masonry cross section after anchor lock-off
(prestressing + gravity load) and E,n is the initial elastic modulus of masonry defined in Chapter 3. Assuming that
E, remains nearly constant, Eqn. 4.1 can be reduced to:
Err: krfî
(4.2)
where k" is termed the 'specific creep'. Eqn. 4.2 is a greatly simplified expression. While perhaps troublesome to researchers,
it has considerable appeal to codification and design work
because of it's simplicity. The total prestressing loss due to creep, Af"., is derived from Eqn. 4.1
as:
4f"r:
e"rEp,
(4.3)
fC"f^,
where Eo, is the elastic modulus of the prestressing steel defined in Chapter 3. When assuming
that E,n remains nearly constant this equation reduces to:
L.f"r: erQpr:
Erskcfî
(4.4)
The material creep parameters C" and k" are specific to the implicated materials (concrete
block, mortar and grout) and are derived experimentally. In the event of an experimental inves-
tigation, the initial elastic shortening strain considered in Eqn. 4.1, e-i, is evaluated as strain
arising immediately from application of axial load.
It is clear that the Eqns. 4.3 and 4.4
are
reflecting prestress loss due to creep for walls with uni-
form stress applied to any cross section. If significant permanent moment is applied to a cross
section, the end of the wall with the largest compressive stress will experience more shortening
than the end with the lowest compressive stress. Such differential creep may lead to non-uni-
form prestress loss, i.e. tendons placed in opposite ends of a wall with a permanent moment
applied may experience different magnitude of prestress loss. For the case of ordinary rectangular PCM walls, this issue is normally of little concern because of symmetrical tendon distri-
-57 -
bution in the wall cross section and insignificant permanent wall moment. In fact this issue is
mostly of concern in prestressed beams and girders that carry permanent load in flexure. The
'effective modulus' method, e.g. Sritheran and Fenwick [4-13], may be employed to capture
creep effects due to combined axial and flexuml loading.
4.1.3.2 Shrinkage strain

Eqn. 4.5 defines the prestressing loss due to shrinkage of the masonry, Ath. It is assumed that
the loss is proportional to the total shrinkage strain .6, which is independent of axial load on
the masonry.
Lfrn
(4.s)
erhEp,
The material shrinkage parameter
ts
is obtained experimentally.
4.1.3.3 Relaxation of prestressing steel

Relaxation of prestressing tendons (general term for both strand and bar) is a well documented
property which may be exfiacted from the prestressing codes. The steel relaxation properties
should preferably be provided by the manufacturer. The following values presented in Table
4.I
may be considered typical of modern prestressing strand and bars
l4-l4l which normally
are of the 'low-relaxation'-type. These values represent minimum requirements to prestressing

steel. The relaxation is given as a percentage of the
70Yo
initial reference prestessing steel stress of
of the breaking load, 0.7fou, and is measured after 1000 hours. Long term relaxation loss
may conservatively be estimated to 3 times the loss after 1000 hours 14-91.
When
it is necessary to evaluate
relaxation as a function of time, the Prestressed Concrete
Institute [4-15] proposes the following equation for calculating the prestessing loss Afo, due to
relaxation of low-relaxation prestressing strands.
^fp,(t)
: fo,,lrye
o.5s)]
r",H>
(4.6)
0.55
TABLE 4.I-Relaxation of prestressing steel (20oC)
Type
Relaxation
lg
Time
Stress
Strand sl5mm
2.5%
1000
0.7fpu
Bar s23mm
4.0%
1000
0.7fpu
% of
0.7fou
-58-
Hours
In this equation t.i is the initial steel stress after anchor lock-off, t is the time given in hours
after stressing and
f,,
is the yield stress for the steel. Eqn. 4.6 only applies to prestressing
strand with presfress levels above
0.55f,r For application to prestressed structural components
with prestressing of less than 0.55L" insignificant relaxation is anticipated. Relaxation after 50
years calculated with Eqn. 4.6 amounts to 3.4%o. This is significantly less than the previously
discussed and probably conservative estimate of 3 times k (1000 hours) :7.5o/o.
The above information may be synthesised by the following equation that evaluates the long
term prestressing force loss in stand due to relaxation,
at 0.55$, and O.7k,fou at 0.7fp,, assuming that
A,fo,
: k,fp,ifilc-
o.5s)
tu
{yt$"
AL
bV interpolation between zero loss
:0.85:
(4.7)
E>0s5
4.1.3.4 Total Prestress Loss

The total long-term prestressing loss due to creep, shrinkage and steel relaxation may be evaluated by:
Lfil : Lf"r+ Lfrn+ Lfp,
(4.8)
Eqn. 4.8 is of conservative nature because the prestress losses are based on the initial masonry
stress
f.i. h reality,
the gradually occurring creep, shrinkage and relaxation gradually reduces
the masonry stress which then reduces the creep and relaxation rates. Section 4.5.2 discusses
these implications in some detail.
Mild steel reinforcement parallel to the presfessing steel effectively resists the vertical
creep
and shrinkage strain movement and consequently reduces the prestress loss. This effect is not
considered herein because it is assumed that prestressing steel constitutes all flexural reinforcement for PCM walls.
For later discussion of structural performance, it follows that the total long term tendon force,
P1,
following all losses is given by:
Pt : Pi-AorLfpt
(4.e)
where P; represents the total initial preshess force after anchor lock-offand
stressing steel area-
-59-
A*
is the total pre-
4.2
CREEP AND SHRINKAGE EXPERIMENT
A considerable number of
research projects have been conducted in various countries on creep
and shrinkage of concrete masonry, assessing the magnitude of creep and shrinkage as a func-
tion of materials, axial load, time and ambient temperature and humidity L4-12,4-16,4-171.
Results from these experimental investigations, that all lasted about one year, constitute the
basis for codification of creep and shrinkage in BS 562814-21, AS 3700 [4-3] and MSJC 14-41.
The precast hollow core concrete masonry blocks ordinarily used in the upper North Island
of
New Zealand are composed of a porous pumice type of aggregate. With a mass of approximately 1850 kN/m3, this type of block can be classified as light-weight in the New Zealand,
context. The MSJC working
draftl44l classifies concrete masonry in three weight groups: (l)
light-weight with mass less than 1680 kd.', (2) medium-weight with mass 1680-2000 kg/m,
and (3) normal-weight with mass higher than 2000 kg/m3. The North Island concrete masonry
block falls in the medium-weight category according to the North America weight specification. Normal-weight concrete masonry blocks are commonly used elsewhere in New Zealand,,
for example in the Canterbury region where granite aggregate (or other dense aggregate types)
is readily available.
None of the studies previously reported apply directly to the New Zealand light-weight type
of
concrete masonry and none of the creep and shrinkage experiments directly addressed fully
grouted concrete masonry, to the knowledge of the author. It was therefore considered necessary to conduct a creep and shrinkage study on locally manufactured concrete masonry. The
University of Auckland study presented herein was initiated in collaboration with researchers
at the University of Wyoming U.S.A., who initiated a parallel creep study
in
1996 and later
reported the findings in [a-12].
4.2.1 Design of the experiments

The experiments described in this chapter were conceived to reflect common concrete masoory
construction practice in New Zealand and to reflect realistic axial prestress levels expected in
PCM. In the year of 1998 it was decided to manufacture testing equipment for 10 wallettes, 1.0
m high, 0.59 m wide and 0.14 m thick. The wallette dimensions are shown in Fig. 4.2.
-60-
FRONT
VIEW
SIDE VIEW
Fig.4.2-Wallette dimensions and test setup

All walls in the first testing
series were constructed
of ordinary grey 15 series precast hollow
block concrete masonry manufactured to the specifications ofNZS 31021983 [4-18]. The wallettes were built in running bond. The mortar used was pre-blended Trade MortailM, a mortar
used for nearly all concrete masonry constnrction in New Zealand,, consisting of cement, sand
and a plasticiser with a mix ratio cement:sand
of l:3.
Trade Mortarru conforms to NZS
4210:200I [4-19] with a minimum compressive strength
of
12.5 MPa. Five wallettes were
grouted, one was partially grouted (2 out of 3 flues grouted) and four remained ungrouted. Presffess was applied at day 17 after construction (day I
after grouting) with levels ranglng
between 0.97 MPa and 2.8 MPa. The grout was specified according to NZS 4210:2001with a
minimum compressive strength
of
17.5 MPa.
shrinkage compensating agent, SIKA
CAVEXTM, was added to the grout shortlybefore pouring to compensate for immediate shrinkage caused by water removed from the grout by the dry concrete masonry blocks. TWo wallettes were not stressed, providing an estimate of pure shrinkage shortening. Shortening of the
wallettes and ambient temperature were monitored on a regular basis throughout the 373 day
test duration (counted from the day of stressing). Specifications for testing series
are given
in
Table 4.2.
The following net areas were used for computation of the initial stress, fi1: For ungrouted wallettes it was assumed that only the block flanges (30 mm wide) and the ends (30 mm wide)
were carrying load, as the intermediate webs alternated location and therefore did not carry
-61
TABLE 4.2-Specifications for series

Prestress
oll,o
| "';J'
Wall
Grouting
-UGI
Ungrouted
0.00
0.00
-UG2
Ungrouted
1.40
t.32
-UG3
Ungrouted
-UG4
.PGI
Ungrouted
Partially grouted
.FGI
.FG2
Fully grouted
Fully grouted
Fully grouted
Fully grouted
Fully grouted
-FG3
-FG4
-FG5
*ocfi
estimated,
measured on
I
Concretr masonry properhes
change
day 28
% of fn,i
fm,28*
60/o
dav 373
fm fl
E-n
Load ratio
day 0
fri/f*,2g
5.0
0.000
5.0
0.093
l0
|.92
8o/o
5.0
0.140
2.80
2.60
7o/o
5.0
0.187
2.70
2.22
t8%
6.7
0.00
0.00
0.97
0.93
r.22
1.
t,46
t.94
MPa
MPa
2.
0.t62
't\
20. I
17.4
0.000
4%
7.5
20.1
17.4
0.056
t0%
7.5
20.1
0.070
t.34
8%
7.5
20.1
t.57
l9o/o
7.5
20.1
t7.4
t7.4
t7.4
MPa
MPa
GPa
l0
0.083
0.lll
cut
load. A resulting net area of 40200 mm2 was found. The cross-sectional areas of the fully
grouted and partially grouted wallettes were taken as 80300 mm2 (gross area) and 55100 mm2
(2/3 of the gross area), respectively.
A second test series of l0 wallettes, using similar materials and construction practice, was initiated in the year of 2000 with the purpose of confirming and expanding the findings from
series
l. A wider range of prestress
was applied to the wallettes, 5 being grouted and
ungrouted. The testing setup used for testing of the series I wallettes was reused and featured
improved instrumentation. Prestress was applied at
l0
days after construction (9 days after
$outing). These wallettes were subjected to prestress levels ranging between 1.0 MPa and 4.0
MPa. Two wallettes remained unstresse4 providing assessments of the pure shdnkage strain.
Shortening of the wallettes was monitored on a regular basis throughout the 388 day test dura-
tion, while ambient temperature and humidity were measured continuously. Specifications for
testing series 2 are given in Table 4.3.
Table 4.4 provides schedules of events for both testing series, again with day zero taken as the
day of stressing. The age of the blocks could not be obtained accurately but presumably ranged
between 28-52 days at the day of construction.
4.2.2 Test setup and instrumentation

Fig. 4.3 shows schematically the test setup and instrumentation configuration for the two test-
ing series. Heavy coil springs were installed in series with the walls between the tendon
-62-
TABLE 4.3-,Specifications for series
2
Concretr masoffy properoes
Prestress m
day0 |
day3S8
Load ratro
change
day 28
day 388
day 0
of f^1
fm28
fm- | n-
&r/f'n:s
Wall
Grouting
{-
frnf
2-UGI
2-UG2
Ungrouted
0.00
0.00
Ungrouted
|.t4
l.l6
-]."/o
15.0"
0.076
2-UG3
Ungrouted
2.t4
2.to
lo/o
15.00
0.143
2-UG4
2-UG5r
Ungrouted
2.99
2.96
0%
15.0"
0.199
grouted
0.00
0.00
grouted
1.00
|.07
2-FGI
t-yul.
2-FG3
2.FG4
2-FG5
Yo
0.000
15.0"
Ungrouted
Fully
Fully
Fully
Fully
f ully
8.2-
19.5
0.000
18.2-
19.5
0.055
-6%
grouted
t.E2
t.67
8o/o
t8.2
19.5
0.100
grouted
2.78
2.31
t7%
t5%
18.2
19.5
0.1 53
grouted
4.00
3.41
MPa
MPa
*failed by vertical splifting during stressing sequence,
tr
t8.2
19.5
MPa
MPa
estimated,
0.220
GPa
measured by testing of 3 prisms
anchorages in order to keep the prestressing force approximately constant over time. Both wal-
lette series were built in the Civil Test Hall. Series I wallettes were stored in the Civil Test Hall
while the series 2 wallettes were relocated to the basement of the Test Hall after application of
axial load.
Axial shortening strains were measured via a digital gauge between measurement points. Each
measurement sequence (one measurement over each set of measurement points for all walls)
that was taken at a particular time was repeated three times to improve the reliability of the
data. Test series
had the Demec measurement points attached to the 32 mm end-plates in
each of the four corners (measurements
Sl-1 to S1-4), resulting in a gauge length of 1000 mnr
The Demec instrumentation layout for testing series 2 was improved so that, for each wallette,
8 measurements were taken directly on the
masoffy (measurements S2-l to S2-8) over
a gauge
TABLE 4.4-Creep and shrinkage testing, schedule of events

(a) Series I
(b) Series 2
Day:
Event:
Day:
Event:
consfructron
-10
constructlon
grouting
-9
grouting
-J
cappung
-3
cappmg
application of axial load
applrcatton ot axlal loac
373
data collection concluded
400
testing of prisms cut from wallettes
2-UG4 split vertically between two cells

2-UG3 split vertically between two cells
70
230
2-FC5 split vertically down the middle
388
data collection concluded
388
testing of prisms cut from wallettes
-63-
. Meosurement
point
s2s2-5
s2-5
s2-7
s2-2
s2-4
s2-8
sl -1
st -5
-2
sl -4
s1
s2-6
SERIES
SERIES 2
Fig. 4.3-Wallette test setup and instrumentation

length of 200 mm. Several improvements were anticipated: (i) strain was measured in a uniform stress region, (ii) measurements were not influenced by potential deformation of the endplates, and (iii) more measurements were obtained with improved reliability.
A reference gauge length of 200 mm was provided by a reference bar made of INVr\R
steel
with a very low coefficient of thermal expansion. This bar was measured 3 times for
each
measurement sequence for testing series
l0) and 6 times for
(before wallette
each measurement sequence
l, after wallette
5 and after wallette
for series 2 ftefore ahd after every two wal-
lettes). It was found using the Demec system that length measurements (and changes) could be
replicated with good accruacy for repeated measurement sequences taken at the same day.
The following procedure was used to determine the axial shortening strain of a wallette at a
particular time, t, having acquired three measurement sequences at time t:
l.
For each sequence: veriff the reliability of reference measurements (comparison between
the three sequences); subtract the average reference measurement from the wallette measurements taken between these two reference measurements. Then
veriff the reliability of
the corrected wallette measurements (comparison between the three sequences).
2. For each data set (three corrected measurement sequences for a particular day), find the
average measurement for each wallette (i.e. average of 3
of 3 x 8 measurements for series 2).
-64-
x 4 measurements for series I and
800
(a) Fully grouted wallettes
700
E
ooo
.9 500
9r+oo
.Y
1,
E 3oo
EE
+
o
zoo
o
100
tA
f^.^-
tt
^
ffi
F
t
F
'-
-d
H
+--
---84
-E{
,-u
A
-/H
F-
-6
+.0.00
\-._"
tl
MPa, Fully grouted
+F0.97 MPa,
Fully grouted
-l<-'t.22 MPa, Fully grouted
+ 1.46 MPa, Fully grouted
-O- l.94,MPa, Fully grouted
tl
50
100
150
200
250
300
350
lime (days)
800
s-
(b) Partially and un-grouted wallettes

-e- O.OO MPa, UngrouteC -e- t.+O MPa, Ungrouted
--F 2.10 MPa, Ungrouted -# 2.80 MPa, Ungrouted
400
700
.E
g600
-.-2]O
MPa, Part.
I|
o
.9 500
h/
Irqoo
xco
'F
z+ 300
o-
-q200
o
100
./
/
grouted
{r-n| \.'-l-+'--Y
pa
MA
--*. alr--<l >---+

EJ-fI
N&Wr
\
eF; ry
o
50
-H-
1oo
.E
-r
ts-x
-l
I
4,/
4r
----r-
tv'
'150
r,|n"rt8"}'5)
Fig.4.4-,Series 1, creep
2s0
"_J
300
3so
4oo
shrinkage strain
3. The shortening at time, t, relative to the day of reference (day of stressing), to, is found by
subtraction of the wall average shortening at time to from the average obtained at time
Finally, the shortening shain is found by division of the shortening at time t with the measurement gauge length (1000 mm for series I and 200 mm for series 2).
4.23
Series
I experimental
results
Fig. 4.4 shows the shortening strain (shrinkage + creep) curves for all wallettes, where the two
curves for the unstressed wallettes reflect the pure shrinkage strain. Fig. 4.5 shows creep strain
curves for all data, sampled over 373 days. The creep results presented for grouted, partially
grouted and un-grouted concrete masonry are not corrected for temperature and humidity variations throughout the sampling period.
-65-
400
tl

350
300
re--w.-O\ -.
/,r
c
$zso
I
.9
zoo
F'uo
o)
100
JU
^
o------.E
v/tr{'=g **.E-4|
rn&a^*.q
d
-*-*-4
-#
-e---
J>-'
\-,-
.Y
Xn
=r
Er--
-+-0.97 MPa,
r-o
tsd
\---A
Fully grouted
-}-1.22 MPa, Fully grouted
200
150
-X-
1.46 MPa, Fully
--
1.94 MPa, Fully
grouted
grouted
250
Time (days)
400
(b) Paftially and un-grouted wallettes

350
hrAr, -\
300
at'
v'
A/-'
c
$zso
0
.9zm
['uo
o
100
50
0
{v
ffi-lf
,lt\
z5
-K
--*-
*t
-4
EI
4-E
F-*
+l-.1.40 MPa, Ungrouted
--F2.10 MPa, Ungrouted
-e
\t
ad
so
\---4
2.80 MPa, Ungrouted
+2.7O
'roo iso
ri-iffi"y.l
MPa. Part. orouted
25o
Fig.4.5-Series l, creep strain

Fig. 4.6 shows the variation of the of the ambient air temperature measured concurrently with
the creep and shrinkage measurements.
It is noted
that the measurements and the smoothed
curve (thick line) reflect the daytime temperature in the Civil Test Hall. The mean 24 hour temperature was probably a few degrees Celsius lower on average.
From Fig. 4.4 it appears that the shortening strain curves stabilised over the last 50 days of
measurements, and for the unloaded wallettes (pure shrinkage) approximately levelled out.
This observation is supported by Fig. 4.6, showing increasing temperature over the last 50 days
of the experiment, a condition that generally would have indicated an increase of the shrinkage
rate. It would be reasonable to assume that the maximum shrinkage strains for un-grouted concrete masonry amount to approximately 200x10-6 and for grouted concrete masonry amount to
approximately 450x10-6. Th" shrinkage strain for the partially grouted wallette was estimated
-66-
22
flrtr
.:
g
-0! 20
dt
q,
9rg
o
I(!
bro
,r
=(
u
{6
/e
\
\
(,
12
R.
\
-
so
1oo lso
Fig.4.ASeries
n.iil"*r
1, ambient
as 213 of the grouted wallette value plus
l/3 of
\f
25o 3oo 3so
4oo
air temperature variation
the ungrouted wallette value, reflecting the ratio
of filled to unfilled flues in the partially grouted wallette (same assumption was used in Fig.
4.5(b)). Fig. 4.6 shows that the ambient temperature at the beginning and end of the test were
similar. This suggests that thermal expansion of the wallettes did not affect the final strain readings because of no temperature difference.
Considering Figs. 4.5(a) and (b), it is evident that the creep strain curves have approximately
levelled out by the end of the experiment. In fact, the creep curves for the fully and partially
grouted wallettes levelled out as early as 4 months (200 days) after application of axial load.
It
is therefore assumed that most of the creep and shrinkage took place within the373 day period.
Table 4.5 summarises the results from test series 1. It is seen from this table that the measured
initial elastic shortening of the wallettes, eni, was inconsistent with the applied force level for
both ungrouted and grouted wallettes. This inconsistency was probably caused by the location
of the measurement points on the base plates. It is likely that the base plates deformed during
the stressing sequence, thus causing error to the measurements. This effect is not believed to
have influenced measurements after application
of axial load because the axial force in the
wallettes remained nearly constant thereafter. Consequently, the creep coefficient and the spe-
cific creep given in Table 4.5 were based on the theoretical initial elastic shorteningo en'r
4.2.4 Series 2 experimental results

Fig.4.7 shows the shortening strain (sbrinkage + creep) curves for all wallettes with the two for
the unstressed wallettes representing the pure shrinkage strain. Fig. 4.8 shows creep strain
-67 -
TABLE 4.S---Series 1, summary of results

Applied
Elastic lhortening
Shortening under sustained load*
total I
I
theoretical
Specific
creep
Coeff.+
creep+
t",
co
kc
stress
measured
Wall
fmi
tmi
I.UGI
0.00
203
203
I.UG2
I.UG3
I.UG4
I-PGI
r.40
tt7
355
203
t58
.44
r20
2.t0
97
t75
4t6
203
234
34
ll
2.80
62
233
469
203
293
26
105
2;t0
55
203
694
364
346
7l
128
444
444
70
627
444
225
3.23
231
shortening
g,,,.,
shrinkage
Creep
r,
I-FGI
0.00
I-FG2
I-FG3
I-FG4
I-FG5
0.97
t.22
123
87
680
444
246
2.83
203
1.46
t23
104
68r
444
257
2.47
t'76
444
307
2.21
pe
|.94
37
139
"t20
MPa
pe
f rased
f,niEm where
r58
pelMPa
Et
curves for all data sampled over 388 days. The creep results presented for grouted, partially
grouted and un-grouted concrete masonry are not corrected for temperature and humidity variations throughout the sampling period.
Fig.4.9 shows variation of the ambient air temperature and relative humidity measured continuously throughout the experiment (24 hour mean). It is noted that temperature and humidity
measurements were taken every half how throughout the test period and that the wallettes were
located in the Civil Test Hall basement.
It is also noted that the ambient temperature
at the
beginning and end differed by about 2 degrees. Assuming a coefficient of thermal expansion
for masonry of l0pe/"C a thermal expansion of 20pe is calculated, suggesting that therrral
expansion of the wallettes did not affect the final strain readings significantly.
From Fig.4.7 it appears that the shrinkage curves (unloaded wallettes) started to level out after
150 days of measurements, and for the ungrouted wallettes approximately levelled out. In a
short time intenral after day 220 all curves indicated expansion. This coincided with a rapid
increase in relative humidity and temperature as shown in Fig. 4.9. As the relative humidity
resumed normal levels of about 600/o, the wallettes resumed the shortening process. It is noted
that the expansion of 2-FG5 (a.0 MPa) was significantly larger than that for the other fully
grouted wallettes. This abnormality was attributed to the splitting of 2-FG5 at day 220 which
affected the subsequent measurements. The drop in strain of 2-FG5, just after day 220, of 160
microstrain should have been of the order of 80 microstrain, as was measured for the fully
-68-
(a) Fully grouted
Fa
eoo
,/,
6t
o| 600
.t
.g
?o.
loo
&'
o
t,
zoo
#
#
'# w fv
ry
--H
-qt'
t'*
.$
h=*- --4 --ffi
14.
1000
?
ffi
Tt
>a-
-{
s----"-
-1
-_-t_rA
-*-0.00
MPa
+F1.00 MPa
+e1.82 MPa
-e-2.78 MPa
+-,+.0OMPa
150
200
250
400
350
Iime (days)
1200
tl
(b) Ungrouted wallettes
tl
^100o
.E
.$
E
o)
aoo
600
.tr
?rL qoo
o
g
zoo
f'
.j 4 u--'x
.g
-.6
#
a*-
4F'4
--g
-d
A---A
-e--'
- --'/ -\
/
x____x_
.* E-----E-
2r-------6\
)r-
--+--g
...*.-{ss6ypu
+F i .14 MPa
++ 2.14 MPa
trf
->e--x
-2.99
so
1oo
lso
Fig.4.7-,Series
nrj?i.yrl
2o
250
3oo
MPa
35o
4oo
creep + shrinkage strain
grouted unstressed wallefte 2-FGl. The true shortening and creep for 2-FG5 after day 220
should therefore be approximately 80 microstrain higher than shown in Figs.
.7@) and 4.8(a).
Considering Figs. a.8(a) and (b), it is seen that the creep strain curves appear to have approximately levelled out by the end of the experiment. It is therefore assumed that most of the creep
and shrinkage took place within the 388 day period.
From Series 2 test results,
it would be reasonable to assume that the morimum
shrinkage
stains for un-grouted concrete masonry amount to approximately 450x10-6 and for grouted
concrete masonry amount to approximately 750x10-6.
Table 4.6 summarises the results from test series 2. It is seen from this table that the measured
initial elastic shortening of the wallettes, rr;, was reasonably consistent with the applied force
level. For this test series, there was good agreement between measured and estimated initial
-69-
600

I
500
4.
F+oo
(!
J4
;o
$ zoo
'100
\d
*"*
.= 300
E
-#
-F
tr
-)<^
r{
)<-'4
-tr|-Ff
H
fr-
-*-^/
rA- 1.00
{F1.82
150
200
250
300
tl
(b) Un-grouted wallettes
Fcoo
o
,r
roo
o
o
$ zoo
'100
-r
500
,l
350
,---a
4=,}
-*'---
a#
^r
-{
-g'
lw {;r"*
50
MPa
-+-4.00 MPa
lime (days)
600
MPa
-X-2.78 MPa
100
50
"*t'
--e---O
)e-
-#
")<'^
\
-rEE
-EE--z
J\"'A -A'-
/^
-x
400
---*.-...Z\
.+
-B
--_t---+
---E--
-F 1.14 MPa
100
r50
200
lime (daF)
250
300
--F
2. 14
-*-
2.99 MPa
MPa
350
400
Fig.4.8-Series 2, creep strain

TABLE 4.HSeries 2, summary of results
Applied
Etastic
hortening
Shortening under sustainr ld load*
Ureep
Specific
theoretical
total
shrinkage
creep
Coeff.+
creep*
Qni,r
shortening
sh
ecr
cc
kc
435
435
95
741
435
321
3.39
852
435
430
2.43
282
202
l0t4
435
592
2.40
200
stress
measured
Wall
fmi
grni
2-UG1
0.00
2-UG2
l.14
t26
2-UG3
2.14
208
t78
2-UG4
2.99
261
249
738
738
2-UG5
2-FGI
2-FG2
0.00
1.00
9l
69
950
738
220
3.19
219
2-FG3
1.82
It8
125
t022
738
292
2.34
t6l
2-FG4
2.78
183
l9l
l09l
738
361
1.89
r30
2-FGs
4.00
283
274
I 159
738
46r
l.68
lt5
MPa
lre
pe
a !3se4
lues, +
tased on
eni.r:
fmiEm where Ern:
800f,
-70-
pe/MPa
s0
100
150
250
200
300
350
23
.2
o
lv
21
Mean
19.2 deg.
t^r\
ct
9rg
"L.
*'1
vv"
(1,
o
q,
\,
CL
E 1"
o
"
^4
\j
15
l^
80
n
a
:)70
E
tr
/l,r
i60
o)
-g
o
r\
/ul'li\
50
ill
il
\lh
WA
An
^
\l
l'
l/l.
t]
tl
r/\
I
I
1r
|llh/h'
/l
ll./
nr
l\i
|/il 'l
il
'll
1/
40
50
'ro0
Jt/ban
58.9%
^,
|/V
150
200
250
300
350
Time (days)
Fig. 4.9-Series 2, ambient air temperature and humidity variations

elastic shortening values. To be consistent with Series
creep given in Table 4.6 were based on the theoretical
l,
the creep coefficient and the specific
initial elastic shortening,qr,i.t.
As indicated in Table 4.4,2-UG3,2-UG4 and 2-FG5 split vertically down the middle after various durations under axial load. Only the splitting of 2-FG5 appears to have affected the creep
results significantly, partly because of the late occurrence after about 230 days under load. The
creep curves for 2-UG3 and 2-UG4 in Fig. 4.8(b) appear smooth and consistent.
4.2.5 Discussion of experiments

Shrinkage:
Comparing the measured shrinkage strain for the two test series,
it is clear that the highest
amount of shrinkage occurred in the 2nd test series. This can be explained by studying Figs.4.6
and 4.9, showing the temperature and humidity values throughout the tests.
-7r
It is clear that the
mean temperature was higher for the second test series which was stored in the Test Hall basement. The climate in the basement was clearly dryer than that of the Test Hall floor where test-
ing series
was stored (humidity was not measured for series
l). Both higher temperature
and
lower humidity are expected to increase the drying shrinkage rate and also the long term
shrinkage magnitude.
ln fact, the environment in the Test Hall basement is considered to result in shrinkage strains
exceeding those expected for walls in a realistic outdoor environment. This is because an out-
door environment (in New Zealand) is characterised by lower average temperature and higher
relative humidity than those of the Civil Test Hall indoor environment. The maximum shrinkage strain measured
and tr6
in series 2 of
:750 lre for grouted
er6
450 pe (microstrain) for ungrouted concrete masoffy
concrete masonry may therefore be considered as conservative.
When considering the shrinkage results it should be kept in mind that the measurements only
include shrinkage recorded after application of prestress. Series
day after construction, and the prestress was applied I
I wallettes were grouted
one
days after construction. The series 2
wallettes were grouted 3 days after construction and stressed l0 days after construction. During this period the walls dried considerably. This is in particular the case for the grouted walls,
which were heavily water saturated directly after the grouting process. In case of final applica-
tion of prestress to a relatively wet wall a considerably higher amount of shrinkage strain
should be expected.
The age of the mason-ry units at the time of wall construction also affects the anticipated
shrinkage strain. Even though the masonry units have attained nearly all of the 28-day strength
after perhaps 7 days, the drying shrinkage strain rate has not yet slowed down. In the exteme
case
it is possible that the masoffy units are 7 days old at the time of construction and perhaps
14 days
old at the time of final prestressing, hence in such situation a higher amount of shrink-
age strain could be anticipated.

Creep:
Fig.4.10 presents the experimentally derived creep coefficient C., calculated according to Eqn.
4.1 as a function of the axial force in the concrete masonry. The results are based on the modu-
lus of elasticity defined by Eqn. 3.8 in Chapter 3. Note that this equation must be applied for
estimation of
E*
when using the creep coefficients presented below.
strain is based on specified strength,
If the estimate of creep
E should be based on Eqn. 3.9. It is clear that compari-
(a) Ungrouted
lo
o3
(t
co
sr"*'Tl
Series 2
(b) Fully Grouted
FS-GI
Series 2
lo
o3
C'
c
-9
t
6A
o
a,
oa
Ea
o.
o0
CL
o
E
(,1
CL
E
o1
40123
(b)
123
Arial Stress (MPa)
(a)
Atlal Stress (ilPa)
Fig. 4.l0-Experimental Creep coefficient Cc.

son of experimental results and code-defined values may potentially be misleading because
significant differences in definition of
It is seen from Fig. 4.10 that there
En.,.
are significant differences in the results from the two test
series. For ungrouted masonry (Fig.
series
a.l0(a)) it is clear that smaller values were measured in
and larger values measured in series 2. Also, the values of C" reduced with increase
fi;, a maximum
of
of
value of 3.4 being measured.
Regarding Fig. 4.10(b) for grouted concrete masonry it is seen that these test results showed a
similar trend with reduction of C. for an increase of
f.', but there was little scatter
between the
values from the two test series. The maximum value for C" for grouted concrete masonry
recorded in these experiments was 3.2.
Fig. 4.1
shows the specific creep calculated with Eqn. 4.2 as a function of
Li.
These curves
show similartrends to those of C" in Fig. 4.10. The maximum specific creep forungrouted concrete masonry was
l<":282 pr/MPa and for grouted concrete masonry
I\:231
pe/MPa.
Thble 4.7 summarises the maximum values of the creep coefficient C", specific creep
and
shrinkage strain .1. These values should theoretically be amplified to account for the creep and
shrinkage expected to occur between years
I and 20. No magnification has been adopted
herein because no research so far has indicated it's applicability to concrete masonry.
-73
300
300
rS eries
(a) Un grouted
vo
IL
200
CL
200
o(,
rtr
(,
roo
(a)
roo
Arial Stress (MPa)
At<ial Stress (MPa)
Fig.
E
(,
.S rrie"
.s sries 2
(,E
&
l?^
(t
(b) Fully Groute d
.s eries 2
o)
4.ll-Experimental Specific Creep lq.
TABLE 4.7-Experimental maximum values of Cr,lt,

cc
kc
tsh
Ungrouted
3.4
282
43s
Grouted
3.2
231
738
pe/IvlPa
tl
rr,.
4.3 EXPERIMENTAL VS. CODE VALUES

4.3.1 Other experimental work
A compilation of recent experimental work on creep of
concrete. masonry was prepared by
Hamilton and Badger 14-121. The results were given in terms of k" and were specific to
ungrouted concrete masonry. Values for light-weight and normal-weight units were presented
and categorised
with respect to the utilised mortar, M,
S and
N (refer to [4-4] for definition of
mortar types), with typical mortar compressive strengths N: 5.2 MPa, S: I2.4MPaand
M:
17.2
MPa being adopted.
In New Zealand only one type of mortar is commonly used: Trade MortarrM. It's a pre-mixed
and pre-bagged product characterised by a specified strength
of 12.5 MPa. The Austalian
des-
ignation for mortar of such properties is M3 or M4. In comparison with the US mortar designation, the New Zealand Trade Mortarru can be classified as an S-mortar.
It is noted that the mortar type influences significantly the creep and shrinkage properties for
ungrouted concrete masonry because much of the shortening takes place in the mortar joints.
-74-
TABLE 4.8-Experimental values of k,, normal weight concrete masonry

Duration of
Mortar type
Mlslu
Researcher
Grouting
Malaoud [4-16]
no
55.1-74.0
400
Hamilton & Badger [4-12]

Laursen, Series I
no
92.8- 189
300
no
105-120
373
Laursen, Series 2
no
200-282
388
Laursen, Series I
yes
l 58-23
373
Laursen. Series 2
yes
115-219
388
pre/MPa
days
loading
TABLE 4.9-Experimental values of k., light weight concrete masonrT

Mortar rype
Researcher
Harvey &Lencmer14-17)
no
Maksoud [4-16]
Hamilton & Badger [4-12]
no
Duration
MlslN
Grouting
162-186
46.4-51.6
78.7-82.4
74.0-92.8
pe/MPa
Fe/MPa
300
400
34.8-74.0
no
of
loading
300
pelMPa
days
TABLE 4.10-Code defined creep and shrinkage coefficients

Source Document
cc
kc
sh
BS 5628
3.0
210,
500
AS 3700
)5
70"
700
MSJC,r
0.5
36
100
pe/MPa
Units
moisture controlled units, o using
:800f'm
and
f'-:
l8 MPa
The inlluence of mortar type on the creep and shrinkage of fully grouted concrete masonry is
small because of the presence of the grout core that occupies much of the cross section.
Tables 4.8 and 4.9 show a compilation of research results, including the results found by exper-
iment at the University of Auckland. Only results from experiments lasting more than 300 days
are included. Results from the 'Laursen' experiment (University of Auckland) are associated
with normal weight concrete masonry and S-mortar.
4.3.2 Code values for creep and shrinkage

Table 4.10 summarises values from creep and shrinkage provisions of BS 5628,4,5 3700 and
MSJC f4-2,4-3,441. These provisions do not clearly distinguish between gouted and ungrouted masonry, so it is presumed that the values are primarily applicable to ungrouted concrete masonry. It should be noted that the calculation of creep strain due to C. is dependent on
E*, which is obtained in different ways by the individual codes.
-75 -
4.3.3 Comparison
Creep:
It is seen from Thble 4.8 that the range of k" values for ungrouted
concrete masonry by
Laursen, ranging from 105 gs/MPa to 282 pe/MPa, correspond reasonably well with those
of
Hamilton & Badger. The extreme value of 282 pr/MPa was measured for 2-UG2 that was subjected to a relatively low axial load making the strain measurements less accurate. Furthermore, the Laursen values for grouted concrete masonry range somewhat higher than those
measured for ungrouted masonry, from
ll5
F/MPato 231 pdMPa. The relevance of the
measurements by Laursen is thus confirmed. Thble 4.9 shows that light-weight units generally
result in less creep than normal weight units.
It is seen from Table 4.10 that the BS 5628
and AS 3700 values of C" are of same magnitude,
but that the MSJC value is much lower. Comparison of the Laursen C" values to code values
reveals that C" of the order of 3.0 is relevant. The MSJC value is unrealistic and is disregarded.
Shrinkage:
In terms of estimated shrinkage e,1, it appears that the BS 5628 and AS 3700 values of 500 pe
and 700 p, respectively, are similar, but that the MSJC value is much lower. The values meas-
ured by Laursen for ungrouted concrete masonry of 203 pe to 435 pE are lower than the BS
5628 and AS 3700 values. For grouted concrete masonry higher shrinkage was measured,,4M
Us
for series I and 738 pe for series 2, values that fall near the BS 5628 and AS 3700 values.
4.4 RECOMMENDATIONS FOR CREEP AND SHRINKAGE

4.4.1 Creep
It is found that the creep coefficient C. based on experimental
results for grouted concrete
masonq/ is in agreement with the values stipulated by BS 5628 and AS 3700. For grouted concrete masonry it is acceptable to assume
C":3.0 for all prestress
levels.
For ungrouted concrete masonry, it is clear that the values predicted by BS 5628 and AS 3700
are on the conservative side, especially for axial load levels higher than2
ficient can safely be estimated at C"
MPa The creep coef-
3.0 for design purposes. The value of C" proposed by
MSJC appears unrealistically low.
-76-
4.4.2 Shrinkage
Based on the experimental results and criteria provided in BS 5628 and AS 3700,
it
appears
reasonable to use the following shrinkage strains for design under New Zealand conditions:
For grouted walls
t 6:
700 UE and for un-grouted walls q6
:400
pe. For ungrouted concrete
masonry, the value is significantly lower than the value from AS 3700. For grouted concrete
masoffy the value is similar to the value from AS 3700. On the contrary, the value of
er6
pro-
posed by MSJC appears unrealistically low.
4.4.3 Comments
It is noted that the results presented
above were obtained under laboratory conditions. The
masonry units were not sealed or surface treated, but all wallettes, including those un-stressed"
were capped to simulate a typical construction scenario. Other factors such as storage conditions and age of the masonry units at the time of construction, were not taken into account. Prudence should be exercised when estimating creep and shrinkage losses for concrete masonry
applied in extreme conditions.

The creep and shrinkage properties for partially grouted concrete masonry, which is becoming
increasingly commonly used in New Zealand, should be based on interpolation of the results
given in Table 4.11. Interpolation between the results for grouted and un-grouted concrete
masonry may be done according to the ratio of grouted cells to the total number of cells.
The material parameter, C., should preferably be used for design, rather than k" which does not
reflect actual material properties.
TABLE 4.ll-Recommended values of C", hr r,n.
4.5 TYPICAL
cc
kc
tsh
Ungrouted
3.0
250
400
Grouted
3.0
250
pe/MPa
700
pe
PRESTRESS LOSS
The following section investigates typical trends using Eqn. 4.8, which is termed the 'additive
approach' because of addition of shrinkage, creep and relaxation components without regard to
interaction. Subsequently section 4.5.2 briefly explores an incremental approach for estimation
of the total prestressing loss taking into account time variation and interaction between shrink-
-7't
age, creep and relaxation. That approach is termed 'incremental approach' and is compared
with the additive approach.
4.5.1 Tlpical
losses
- additive approach
The magnitude of prestressing loss can be estimated using Eqn. 4.8 by considering typical
material properties for concrete masoffy and prestressing steel, and typical prestress and axial
load levels. Thble 4.11 gives the recommended values for calculation of prestress losses due to
shortening of the concrete masoffy. Long term relaxation after 50 years was taken as 2 times lq
from Table 4.1.

Four cases are considered: (1) fully grouted wall prestressed with high strength prestressing
strand, (2)
fully grouted wall
prestressed
with high strength prestressing thread-bars, (3)
ungrouted wall prestressed with high strength prestressing strand, and (a) ungrouted wall prestressed
with high strength prestressing thread-bar. It is assumed that the crushing strength,
f'r,
for fully grouted masonry is l8 MPa and forungrouted masonry is l5 MPa. Realistic axial load
ratios, (, ranging from 0.05 to 0.30 are examined, with
E _ P+N
tJffi
(4.10)
Material properties for prestressing strand and thread-bar are taken from Table 3.6. It is furthermore acknowledged that the maximum achievable (permissible) presfress in the prestress-
ing steel after lock-off is in the vicinity of 0.7f," (NZS
101 [4-20D.
The total prestress loss (long term loss) based on the above properties is represented in Figs.
4.12 and 4.13, defined as a percentage of the initial prestress for varying levels of axial load
ratio. It is noted that the losses in terms of stress are independent of the actual prestressing
force (kN) and that the axial force (N+P) can be constituted by any ratio of P vs. N. Prestressed
retaining walls typically have low N, thus P amounts to nearly 100% of the axial load. Prestressed strucfural walls often are subjected to
more than 50Yo of the axial load.
N larger than R where N easily can amount to
It is also noted that if the axial force (N+P)
has reduced sig-
nificantly over time (P relatively large) an axial load based on the average of the initial and
long term preshessing force may be used for estimation of the prestressing loss (iterative process).
-78-
+.Bar
- 0.5 fpu
-+.Bar - 0.7 fpu

-+Sband - 0.5 fpu
-.-Slrand - 0.7 hu
3lo
o
o
6
g
=3s
o
tso
a
a
o
o
825
o
o
o
tzo
15
0.05
0.1
0.15
0-2
0.25
0.3
Atlal load rado E- H?^
Fig. 4.l2-Expected prestress loss in fully grouted concrete masonry
--
Bar - 0.5 fru

-+-Bar - 0.7 tsu
-.-Strand - 0.5 fpu
Brc
o
-+-Strand - 0.7 fpu
o
E
s
=35
o
tso
o
o
e25
e
o
o
tzo
15
0'05
0'1
tur
o''
ro"l
ljto E '
o'25
0'3
'J"
Fig.4.f}-Expected prestress loss in ungrouted concnete masonry

It is seen from Fig. 4.12 that a high amount of loss can be expected for firlly grouted concrete
masonry. For the case of high strength strand with an initial stress of 0.7fpu and an axial load
ratio of0.l5, about 25% loss ofprestress can be expected after all losses have occurred. Forthe
case
of a high strength bar solution with similar initial stress and axial load ratio a prestress
loss ofabout35%o can be expected.
-79 -
Fig. 4.13 shows that considerably less loss can be anticipated using ungrouted concrete
masonry in comparison with fully grouted masonry. The main reason for the lower relative prestress loss for ungrouted masonry is that the anticipated shrinkage strain is lower than that
of
grouted masonry. For the case of high strength strand with an initial stress of 0.7$" and an axial
load ratio of 0.15, about 20% loss of prestress can be expected after all losses have occurred.
For a high strength bar solution, with similar initial stress and axial load ratio, a prestress loss
of about 29o/o can be expected.
In both grouted and ungrouted concrete masoffy,
losses are considerable and must be taken
into account in design. It is seen that lower initial stress levels result in higher relative losses.
This is mainly the result of concrete masonry shrinkage strain becoming relatively more
impodant at lower stress levels.
High prestress loss can be countered to some extent by restressing of the prestessing steel at
some later time after the
initial stressing. Figs. 4.4 and 4-7 suggest that approximately 50% of
all creep and shrinkage takes place within 3 months. It is therefore in theory possible to reduce
the prestress loss by about 50%
if the restressing takes place about 3 months after the initial
stressing.
Finally, if a prestressing steel stress (long term) of less than
0.7t,
is required for ductility
(drift
capacity) considerations, it can in some cases be argued that application of an initial steel stress
of 0.7fp" is acceptable because creep, shrinkage and relaxation possibly will reduce the steel
stress to the desired design level after, say, 3 months.
Ductility requirements arise from seismic
loading, which normally corresponds to the worst credible ground shaking with a return period
of about 500 years. The risk of a seismic event during the period of 'over-stress' of three
months (or even one year) is therefore negligible and likely to be acceptable.
4.5.2 Incremental approach

This section explores the total prestress loss when time effects have been taken into account.
The term 'incremental' refers to evaluation of the total prestress loss by stepping through a
finite number of time steps.
Initially, the time dependency of prestress losses due to shrinkage strain e,6(t) and creep strain
q,(t) are defined as [4-10] (t in days):
-80-
Lf"r(t)
: er,(t)Eo, : Fn"'(tffr"ro :
Lfrn?):
esh(t)Ep,
Fnth(t)ErnEp,
# Hrr*
(4.11)
(4.r2)
7L5*rrroEo,
The prestress loss time dependency for high strength low relaxation strand is given by Eqn. 4.6
and repeated below with time dependency notation added (t in hours):
Afo,(t)
It
: fo,@lryw-
is noted from Eqns. 4.1
and
o.ss)]
4.l3 that prestress
",
ff>
0 55
(4.13)
loss due to creep and relaxation not only are
functions of direct time variation Fn(t) but also function of the instantaneous masonry stess
f",(t) and prestress fo(t). tnfrerent in Eqns. 4. I I and 4.13 is the assumption that any increment
in prestress loss due to creep and relaxation is a direct function of the instantaneous masonry
stess
t(t)
and instantaneous prestress for(t) and not the initial masonry stress
f6
and initial
prestress fp.i. The shrinkage rate is per definition independent of the masonry stress.
Fig. 4.14 shows Fn(t) for shrinkage, creep and relaxation for time between zero and 1000 days
(approximately 3 years). It is seen that the shrinkage rate is higher than the creep rate and that
relaxation initially occurs at a higher rate than shrinkage and creep.
100 200
300
400
500
600
700
800
t (days)
Fig.4.l4-Time variation Fn(t) for creep, shrinkage and relaxation

-81 -
The total preshess loss at any instant can be written as:
Lfplt)
Fn{Lf,,.lt, Lfp{t)J, L,frl?), Lfp,lt, Lfp{t)l}
(4.r4)
which indicates a complex relationship with Afor(t) occurring on both sides of the equation.
This equation can be approximately solved numerically, e.g. assuming that the masoffy stress
fr(t)
and the prestress for(t) remain constant in small time intervals:
interval (step)
At : !*r - tn, the creep, shrinkage and prestress
to h*r.Within each time
relaxation increments are
defined as follows:
Lf"r(Ltn+
l) :
1Fn"1tn+ r)
Lf,h(Ltn+
l) :
1Fn'h 1tn+ r ) - Fn'h 1t))erhEp,
(4.16)
Afo,{Lt,*,
W t'r"(+)W-
(4.r7)
*1r"r"@->
Jw
rr"{t)@cpo,
(4.15)
o.ss)
0.55
From Eqns. 4,15,4.16 and 4.17 it is found that:

P(t
) : P(t) - L\fcr(Ltn* ) * Lfrr(Ltn*

r
fp,(tn*l)
P(tna
* a'fo,(a'tn * 1)\Ap,
(4.18)
(4.1e)
N+
f^(tn* l) :
P(rn+
As
l)
where
Ar: lnb-
(4.20)
The solution to time step n+l is thus solely based on the previous time step n and the given
functions of Fn(t). The total presfiess loss is evaluated, using the initial conditions:
and
fo.(O):
fpsi,
fr(O):
f",i
by stepping through the Eqns.4.l5 to 4.20 a total of M times. As the time step
length diminishes (number of increments M increases), the solution converges to the exact
solution.
Comparison between the incremental and additive approaches may be found in Fig. 4.15. Prestress loss after 50 years was calculated according
4.1
to Eqn. 4.8 (additive approach) and Eqn.
8 (increments approach) using the following input:
-82-
f ,n :
18 MPa,
E^
14.4 GPa,
t6
0.91
0.90
tt
H 0.8s
9g
b 0.88
tr
"F
.. 0.87
o
Axial load ratio
(, =
(Pi+N/(f.&)
g 0.86
0.84
0-1
o-2
0.3
o.4
oiu=
r^rnfl,u
0.7
0.8
0.s
1.0
Fig. 4.I5-Comparison of prestress loss, incremental vs. additive approaches

0.000700, C"
3.0, f,ri
0.7q0* fou
1750 MPa,
iy
1488 MPa, Eor
190 GPa and k (50
: 0.034. The axial load ratio ( was varied in the interval 0.05 to 0.25 and the ratio of initial prestressing force to total initial axial force A: Py'(P;+N) was varied from 0.1 (very light
years)
prestressing) to 1.0 (axial force purely due to prestressing). A constant time step
of I day was
used for the calculation.
Fig. 4. 15 displays a series of plots of the ratio betrveen total prestess loss calculated with the
additive and incremental approaches. This ratio is termed ol-oss ratio'. For a given axial load
ratio, the loss ratio was calculated for a series of prestress force to total force ratios A. In
essence, the figrne can be interpreted as the relative benefit of conducting prestress loss assess-
ment using the more cumbersome and presumably more accurate incremental method.
The following observation are made from Fig. 4.15:
(l)
the loss ratio reduces with increasing
L, (2) the loss ratio reduces with increasing ( for A greater
than about 0.3, (3) the prestress loss
calculated using the incremental method is at least 10% less than that calculated with the addi-
tive method and (4) at the best predicts about 15% less presfress loss, and (5) a loss ratio of
0.88 is found for typical PCM wall values A
: 0.5 and 6: 0.15.
In conclusion, the incremental method gives favourable results when compared with the additive method in terms of structural efficiency and economy. A typical reduction in predicted pre-
-83-
stress loss
of
l2%o can be expected using the incremental method. Given the small magnitude
of achieved improvement using the incremental method in comparison with the significant
uncertainty of the shrinkage and creep properties for concrete masonry,
it is questionable
whether use of the incremental method can be justified. It is therefore recomrrended to use the
additive method for prediction of prestress losses in PCM walls as defined by Eqn. 4.8. This
section confimrs that typical prestress losses as calculated and discussed in section 4.5.1 are
relevant to typical PCM wall design.
-84-
4.6 REFBRENCES
t4-1]
Schultz, A. E. and Scolforo, M. J., An Overview of Prestressed Masonry, The Masonry

Society Joumal, August 1991, Vol. 10, No. 1.
[4-2)
BS 5628: L995,Part2: Code of Practicefor use of Masonry. Structural Use of Rein-
forced and Prestressed Masonry, British Standards Institution, London.
[4-3] AS 3700-1998,
Masonry Structures, Standards Association of Australia, Homebush,
NSW 2140.
t44l MSJC: Building Code Requirements for Masonry
Stntctures and, Specification
for
Masonry Structures, ACI 530-02/ASCE 5-02ITMS 402-02, The Masonry Standards

[4-5] Neville, A.M.,
[4-6] Bryant, A.H.,
Properties of Concrete, John Wiley & Sons, Edition 4F,,844 p.,1996.
Wood, J.A. and Fenwick, R.C., Crcep and Shrinlmge
tn
Concrete
Bridges, RRU Bulletin 70, National Roads Board, Wellington, New Zealand,1984.
l4-7) Bryant. A.H
and Vadhanavikkit, C., Creep, Shrinkage - size and age loading effects,
ACI Material Journal, Vol. 83, No. 2, March-April 1987,pp II7-123.
[4-8]
CEB-FIP model codefor concrete structures, Commite European du Beton - Federation

lnternationale de la precontrainte, 1978.
[4-9]
Eurocode 2: Design of Concrete Sfiactures, Part
General rules and rules for build-
ings, ENV 1992-l-1, European Committee for Standardization, Brussels, December

1991.
[4-10] ACI Special Publication - 76, Designingfor Creep & Shrinkage in Concrete Structures,
Published by American Concrete lnstitute, 1982,484p.
[4-11] Lenczner, D., Creep and Prestress Losses in Brick Masonry, The Structural Engineer,
Institution of Structural Engineers (England), Vol. 648, No. 3, September 1986, pp. 5762.
l4-l2l Hamilton, H.R.
and Badger, C.R., Creep Losses in Post-Tbnsioned Concrete Masonry,
TMS journal, July 2001, Vol. 18, No.
l.
[4-13] Sritharan, S. and Fenwick, R.C., Creep and shrinkage efects in prestressed beams,
Magazine of Concrete Research, Yol.47, No. 170, March 1995, pp. 45-55.
[4-14] YSL Construction Systems, VSL Prestressing (Aust.) Pty.Ltd., Ausfialia, 1995.
-85-
[4-15] Recommendations
for
Estimating Prestress Losses, Report
of Prestressed Concrete
Institute (PCD committee on Prestress Losses, Joumal of the Prestressed Concrete Institute, Vol. 20. No. 4, July-August 1975, pp.43-75,
[4-16l Marksoud, A., Short and Long Term Capacities of Slender Concrete Block Walls,PhD
thesis, McMaster University, Hamilton, Ontario, 1994.
l4-l7l
Harvey, RJ. and Lenczner, D., Creep Prcstrcss Losses in Concrcte Masonry Proceedings, 5h RILEM Intl. Symp. On Creep and Shrinkage of Concrete, 1993.
[4-lS] NZS 3102:1983, Concrete masonry units, Standards Association of New
Zealand,,
1983, Wellington.
[4-19] NZS 4210:2001, Masonry Constntction: Materials and Worlonanshlp, Standards New
Zealand, 200
l, Wellington
!4-201NZS 3101 1995, Concrete Structures Standard, Standards Association ofNew Zealand,
1995, Wellington.
-86-
Chapter 5
STRUCTURAL TESTING - SERIES
5.
SINGLB.STOREY PCM WALLS
In-plane response of post{ensioned concrete masonry walls with unbonded tendons is examined by means of structural testing. A description of the Series
I wall testing prognunme is fol-
lowed by a presentation of the results from structural testing of six fully grouted walls, one
partially grouted wall and one ungrouted wall. Discussion of the results is concerned with wall
structural response in terms of flexural strength, shear strength, displacement capacity, tendon
stress and masonry vertical strain.
5.I
INTRODUCTION
This chapter describes the results from structural testing of eight unbonded PCM cantilever
walls. The walls were tested to explore the behaviour of single-storey rectangular post-tensioned walls of various aspect ratio, thickness and tendon layout. The objective of this study
was to obtain data for use in design; however, design examples are not considered herein.
Using unbonded post-tensioning, walls are vertically prestressed by means of strands or bars
which are passed through vertical ducts inside the walls. As the walls are subjected to lateral
in-plane displacements, gaps form at the horizontal joints reducing the system stiffiress. As
long as the prestressing strands retain a significant force they will return the walls to their initial position. Thus, the lateral force-displacement response may be described by a nearly nonlinear elastic relationship. The integrity of the walls is maintained as no plastic hinges form and
there are no residual post-earthquake displacements. Consequently minimal structural and non-
structural wall damage can be expected.
-87 -
5.2 CONSTRUCTION DETAILS

Wall dimensions are specified in Table 5.1. All walls were 2.6 m high. The wall designations
may be deciphered in the following way: FG indicates fully grouted, PG indicates partially
grouted, UG indicates ungrouted, L3.0 indicates that the wall was 3.0 m long,
Wl5
indicates a
nominal wall thickness of 150 mm and P3 means that three prestessing bars were embedded in
the wall.
FG:L3.0-W20-P3 was the only wall constructed of 20 series masonry units and was the only
wall that had horizontal shear reinforcement embedded. All other walls were constnrcted of l5
series
or l0 series concrete masonry units, without
shear reinforcement. The actual thickness
of the walls was approximately 10 mm less than the nominal masonry unit thickness. The par-
tially grouted wall had the end flues and two intermediate flues grouted; all other flues
remained unfilled.
All walls were prestressed with two or three 23 mm VSL thread bars. For the long walls (L3.0)
with three bars, two were placed 300 mm from the wall ends and the other was placed on the
wall centreline. FG:LI.8-WI5-P3 had two bars placed concentrically at +400 mm from the
wall centreline and the last bar placed on the wall centreline. For the case of walls with trvo
bars, FG:L3.0-W15-P2C, FG:LI.S-Wl5-P2 and UG:LI.8-W10-P2 had bars placed concenhi-
cally at +400 mm from the wall centreline, while FG:L3.0-W$-P2E and PG:L3.0-W15-P2
had bars placed eccentrically at 300 mm from the wall ends. The prestressing tendons
remained unbonded over the entire wall height and had a typical unbonded length of 3.5 m,
TABLE S.I-Wall specifi cations

Wall
l.
FG:L3.0-W20-P3
2. FG:L3.0-W 5-P3
3. FG:L3.O-W 5-PzC
4. FG:L3.0-W 5-P2E
5. FG:L1.8-W 5-P2
initial
initial
axral
steel
prestr.
force
n ickness
self
masonry
shength
shess
force
ratio
lw
bw
weight
N
f't
fo.
fm/f'm
3.0
190
50
13.3
468
580
0.083
)J5
690
0.1 15
622
0.077
length
tendons
3.0
140
l5.l
a
J
3.0
140
20.6
757
3.0
t40
17.8
757
628
u.uu9
t.8
140
20.5
534
M5
0.092
614
760
0.t67
t76
146
0.070
0.t62
6. FG:L1.8-Wl5-P3
1.8
140
18.4
7. PG:L3.0-Wl5-P2
8. UG:LI.8-Wl0-P2
3.0
140
32
15.5
1.8
90
l0
15.0
308
256
mm
KN
MPa
MPa
KN
-88-
including unbonded lengths in the foundation (300 mm), loading beam (400 mm) and load
cells on top of the loading beam (200 mm). The wall prestressing force and axial force ratios
(due to prestressing and self-weight) are given in Table 5.1.
5.2.2 Wall consfruction

Prior to wall construction, the prestressing bars were inserted in the foundation and encased by
ducting to ensure proper isolation from the grout. The walls were consfiucted in running bond
by experienced blocklayers under supervision, using standard grey precast concrete masonry
units and DRICONTM trade mortar. Open-ended concrete masonry units were used for most
of
the walls in order to avoid having to thread the masonry units onto the 4 m long prestressing
bars. Grouting
with 17.5 MPa ready-mixed regular block fill, containing SIKA
CAVEXTM
expansion agent, was perfiormed the subsequent day.
5.2.3 Materialproperties
Masonry crushing strength was determined by material testing of masonry prisms (height by
length of 600 mm by 400 mm) for walls FG:L3.0-W1,5-P2C, FG:L1.8-WI5-P2, FG:L1.8-
Wl5-P3 and UG:L1.8-W10-P2. Strengths of walls FG:L3.0-W20-P3, FG:L3.O-Wl5-P3,

FG:L3.0-W15-P2E and PG:L3.0-W15-P2 were estimated on a statistical basis from knowledge
of masonry block, mortar and gtout strengths.
VSL 23 mm diameter high stength threaded bars were used in all wall tests and had the typical
properties: yield strength of 970 MPa, ultimate strength of 1160 MPa, ultimate strain of
and modulus of elasticity
of
8o/o
190 GPa.
5.3 TESTING DETAILS

5.3.1 Test setup
Wall testing reported herein was conducted using the setup shown in Fig 5.1(a), consisting of a
reusable foundation and loading beam. Notably, no effort was made to replicate gravity and
live loads from suspended floors, with the assumption that these loads were directly analogous
to a component of the applied post-tensioning force. Fig. 5.1(b) schematically shows typical
wall instrumentation. Lateral force was measured by two load cells positioned in series with
the hydraulic actuators, denoted as instruments 1 and 2 in Fig.
5.lO). Wall flexural deforma-
tion and vertical strain were measured by instruments 5-18 (portal type LVDTs) and panel
-89-
L4
l5
I
l0
ll
4l
l6
l7
t8
Fig. S.l-Test setup and instrumentation

deformation was measured by instruments l9-38. Relative sliding displacements between loading beam/wall, wall/foundation and foundation/strong floor were measured by instruments 39,
40 and 41, respectively.

Cyclic structural testing was conducted according to the procedure outlined by Park [5-1], with
the ductility one displacement here termed 'displacement level one' or
Dl. After definition of
Dl, the walls were cycled to D2,D4,D6, etc. until failure, with Dn being defined as n times
Dl. Prior to executing this loading procedure, serviceability limit state level force cycles were
applied to some walls in order to investigate the uncracked stiffness, the lateral displacement
profile and the initial crack formation.

For a symmetric distribution of post-tensioning and gravity actions, the nominal flexural
strength, Mn, is defined by the New Zealand Masonry code NZS 4230:1990 [5-2] as given in
Eqns. 8.17, 8.18 and 8.29. Details of these equations can be found in Chapter 8. The base
shear, Vs, corresponding to Mo may be calculated using Eqn. 8.17, where P represents the
ini-
tial prestress force (no lateral load applied to wall), AP is the total tendon force increase due to
wall deformation at Vr, l* is the wall length, h. is the height of the wall and N is the axial load
at the base of the wall due to live and dead load. The equivalent rectangular compression zone
length, a, is defined by Eqn. 8.18 as a function of the axial forces on the section, the masonry
compressive strength, f',n, the wall width
b* and a
0.85 (unconfined grouted concrete
masonry). Eqn. 8.29 defines the effective tendon force eccentricity, e,, with respect to the initial
-90-
prestressing force and the tendon force increase based on the assumption of uniformly distributed prestressing steel. ln the absence of suitable procedures for estimating AP at the time
of
testing, the tendon force increase at nominal strength was estimated to be 50 kN for the fully
grouted walls and zero for the partially and ungrouted walls. The latter assumption of zero
force increase was based on expectation of minor flexural deformation for these wall types.
5.3.4 Predicted masonry shear strength

The predicted masonry shear strength (stress) was calculated according to the following formulae: Eqn. 5.1 - NEHRP [5-3], Eqn. 5.2 - Paulay and Priestley
[54]
and Eqn. 5.3 - NZS
4230:1990, using the notation adopted herein.
u,
: 0.083(4- t.?s*)lf; +0.25|J,.;=r-:3fr
(5.1)
v^
(5.2)
v^:
0.86ffi
+ o.z4*
l'
Ao
=1.04
MPa
0.024ffi +0.24V=0.576
MPd
(s.3)
The masonry shear strength in terms of force, V,, is calculated by multiplication of

gross cross-sectional area,
A* for fully grouted walls. For the partially
v- by the
and ungrouted walls,
due consideration was made for ungrouted cavities (refer to section 5.2). The predicted
masonry shear strengths, V*, listed in Table 5.2 were based on Eqn. 5.1, this equation being
most recently developed. The initial prestress force was used in this calculation.
TABLE S.2-Predictions and results

Wall
l.
Preclrctrons
vslv.
Test results
V*u* lduro
d"
Yu
lbehaviow
FG:L3.0-W20-P3
336
567
561
29.2
>38
2. FG:L3.0-W 5-P3
3. FG:L3.0-W 5-P2C
4. FG:L3.0-W 5-P2E
5. FG:LI.8-W 5-P2
6. FG:LI,8-W 5-P3
7. PG:L3.0-W 5-P2
8, UG:L1.8-W 0-P2
367
504
465
5.3
20.4
>1.4 rocking, compr. toe crushing

0.73 rocking, compr. toe crushing
3s6
542
373
t2.2
26.3
0.93
356
515
373
8.4
28.3
l.0l rocking,
r.J) rocking,
rocking, compr. toe crushing

compr. toe crushing
154
332
178
25.3
37.9
218
355
266
I4.l
l4.l
9l
180
t20
3.9
10.0
0.50 rocking, incl. compres. strut splitting

0.36 inclined shear cracking, shear failure
73
140
102
7.7
7.7
0.28
mm
o/o
KN
KN
KN
mm
-91 -
compr. toe crushing
inclined shear cracking, shear failure
5.4 TEST RESIJUTS

This section describes the behaviour of the eight walls and presents the measured masonry
compression toe strains at various wall states. General wall behaviour is summarised in Table
5.2, where
V**
is the maximum lateral force recorded and
4*o
is the corresponding displace-
ment. The ultimate displacement capacity, du, is defined as the point at which the lateral wall
strength had degraded below 80% of
V.o.
The ultimate drift capacity is defined as y,
- q,/L.

Rocking response was recorded for all fully grouted wall tests, characterised by opening of a
single large crack along the wall-foundation construction joint. No distributed flexural cracks
were observed and distinct diagonal compression strut cracking was only observed for
FG:LI.8-Wl5-P3. High initial stiffiress was recorded. At nominal flexural strength, Mn, wall
softening occurred due to the initiation of rocking. Tendon yielding was experienced for all
walls that failed in flexure.
The failure mode was characterised by localised masonry crushing in the compression toe
regions, resulting in gradual strength degradation (with the exception of wall FG:LI.8-W15-P3
that failed in shear).
FG:L1.8-Wl5-P3 exhibited rocking response for displacements up to approximately
Dl
(4.3
mm). Displacing the wall towards D2 (8.6 mm) initiated diagonal cracking, eventually leading
to failure in both directions of loading.

Fig. 5.2, depicting the force-displacement response for all fully grouted walls, demonstates
the nearly non-linear elastic behaviour exhibited by unbonded post-tensioned walls prior to toe
crushing, It is seen that the walls retumed to their original alignment, even after large displacement excursions. Therefore the response may be categorised as "origin oriented". The individual curves for each excursion appear pinched, implying little hysteretic energy dissipation. It is
noted that the term 'elastic' in a rigorous sense indicates reversibility and no accumulation
of
damage. In reality the walls did accumulate damage from the onset of tendon yielding and
masonry crushing, causing a reduction in wall stiffness. Thus the behaviour was termed
'nearly' non-linear elastic.
-92 -
Ddft(%)
Ddn (%)
.1.5 -1.0 -0.5
o.5
*[
0.0
0.0
.t500
600
zv
,*oQ
2
4J
400
5m;;
o Pg
6a
(g
soo
o
6
*f
400
.50
-40
-30 -20 -10 0
10 N
V=367kN
v. =50{ lil
*of
-40
30
-400
V. = 567 kN
600
500
OF
TS
-1(x)0
Vr = 336 kN
t(xl0
J6
.200
ao
lo
400 F
.g)
-20 -10
-1.2 {.8
1.5
300
1200
4@
900
300
600
200
$0
,to0
-100
-300
C'
.200
(X)
dl
-3{X)
400
-50
-40
-30 .20 -10 0
-900
gs0 tt't
10 20 30 40
-0.4 0.0
Ezzm
y
EF'*
E.g
6s)
-1.5 -1.0 .0.5 0.0 0.5 1.0
3"t
'o
-300
00
-3{n
E
z
x
c
o
=o
@
dt
-900
V=356kN
-4{n
-1200
.1200
40-30.m-10010,o3040
50
Drift (%)
1.5
4fi
aoo
600
l-
4m:5c
2oo ;E
'50
f'
-2m
4oo
opo
0l-
-50
ao
.1(x)
..*& A
.150
-450
-r0 0 10 20 30 40
8(x,
r
r
50r
15o
100
150;;
-rso
0.6
250
l20o
zz
x!
o Fi
6.h
-50 40 -30 -20
.o,4 4.2 0.0 0.2 0.4

3n; -0.6
50
.N
1z)0
900
3@
o
o
o
1.2
(d) FG:L3.0-w15-P2E
100
0.8
Lateral Displacement (mm)
150
o
o
0.4
@0
Drift (70)
-1500
(c) FG:L3.0-W15-P2C
z
I
dl
Lale]al Displaoement (mm)
?@
=oo
-r@0
Drift(%)
Drift (%)
-1.5 -r.0 -o.5 0.o 0.5 1.0
o
o
-500
(b) FG:L3.0-wl5-P3
(a) FG:L3.0-W20-P3
zI
c
o
E
o
Lateral Oisplacornont (mm)
Lsteral Dsplacament (mm)
400
10
zz
'lm
r
-20o
-r5o
-250
-3{X'
F
lL
-n
50
Vt
-15 -t0
-5
"
00
218 kN
=o
{00
10
15
Latral Displacemnt (mm)
FG:L1.8-wl5-P2
(f) FG:L1.8-W15-P3
(e) FG:LI.8-Wl5-P2
Fig. 5.2-Fully grouted walls; Force-displacement histories
All walls, except for FG:L3.0-W15-P3
and FG:L3.0-WI5-P2E, exhibited nearly symmetrical
respoffie in the push and pull directions. FG:L3.0-WI5-P3 was accidentally displaced -34 mm,
as shown in Fig. 5.2(b), causing serious damage to the wall and preventing effective further
testing in the pull direction. The asymmetric behaviour of FG:L3.0-WI5-P2E, seen in Fig.
5.2(d), was caused by extensive sliding. Note that the data depicted in Fig. 5.2(d) was corrected
for sliding displacement, thus representing only the wall flexural and shear deformation behaviour.
-93 -
-+Push
(o)
Push
FG:13,0-w20-P3
Push
(c)
'+Push
(d) FG:LJ.o-wts-p2E
FG:15.0-Wl5-P2C
Push
Push
-+
(r)
(a) FG:LI.8-Wt5-P2
FG:L|
.s-wr s-PI
Fig. 5.3-Fully grouted walls; Extent of wall damage at failure

Gradual strength degradation was observed for all FG walls (except for FG:L1.8-WI5-P3),
attributed to spalling of face shells and crushing of the grout core. Large displacement capacity
was observed with reliable wall resistance for maximum drifts ranging from 0.73% to more
than I .4Yo,refer to Table 5.2. For all FG walls except FG:L1.8-W15P3, the crack patterns after
failure, depicted schematically in Fig. 5.3, indicated that the extent of damage was confined to
-94-
the lowest two masonry courses in the toe regions. The total lack of shear cracking for
FG:L3.0-W20-P3 indicated that the shear reinforcement was not engaged at all, explaining the
omission of this reinforcement from all subsequent tests in this series.
Fig. 5.3(f) shows the final crack pattern for the only FG wall that failed in compression strut
splitting, FG:LI.8-Wl5-P3. The failure mode was characterised by diagonal cracking due to
tensile splitting of the masoffy compression stnrts forming between the prestressing anchorage
locations in the loading beam and the wall flexural compression zone.
It is seen in Fig. 5.2(f)
that the wall strength exceeded Vl before failure, thus indicating a brittle failure mode influenced by flexure. This type
of failure was unexpected
as the predicted masonry shear strength
was higher than the predicted flexural strength, as indicated in Table 5.2.
Fig. 5.3(b) reveals an inclined crack that appeared while pulling FG:L3.0-W15-P3 to -D4 (-12
mm). This crack is believed to have been caused by large localised splitting forces associated
with the prestress anchorage at the centre of the loading beam. The crack did not develop further during testing, possibly due to transverse restraint provided by the loading beam, and did
not appear to affect the remainder of the test.
5.4.1.3 Sliding
There was no indication
of wall sliding relative to the foundation for
FG:L3.0-W20-P3,
FG:L3.0-W15-P3 and FG:LI.8-W15-P3. Walls FG:L3.0-WI5-P2C and FG:L1.8-WI5-P2

showed sliding displacements of less that 5oh of the total displacement, while sliding of more
than l7 mm was measured in the pull direction for FG:L3.0-WI5-P2E. Sliding became more
pronounced for the last walls tested. This was because of gradual deterioration of the founda-
tion beam roughness as the test series progressed, thus reducing the shear friction along the
wall/base construction j oint.

The prestressing tendon force histories (sum of force in three tendons) for all FG walls are
plotted against lateral displacement in Fig. 5.4. It is seen that the first significant total tendon
force loss (more than 57o of the initial force) for the walls that failed in flexure occurred after
lateral wall displacements ranging from +8 mm to *12 mm, in all cases roughly corresponding
to first yield of the extreme tendons when considering the wall as a rigid body rocking about
the wall corners. At the conclusion of testing, between 25Yo and75% of the initial total pre-
-95-
Drift(%)
.{r.5
-1.0
^
z
5
o
g
o
0.0
800
600
o*
o
Ert
.s0
40
-30 -20 -10
0 10 20
30
-30 -N
Lateral Displacemenl (mm)
-'10 0
Drifi(%)
Drift (%)
-1.5 -r.0 .0.5 0.0 0.5
-1.2 4-8
r.0
75{l
z '*
J
6m
450
uo
300
o
e
&
e
C
Eo
E
t'-
o
F
150
-40 -30 -20 -10 0
-50
10 m
0.0
",o
0.8
1-
*lI
15o
oL
o -30 -20 -10 0
.40
30
0.4
r,
10 n
30
(d) FG:L3.0-wl5-P2E
(c) FG:L3.O-W15-P2C
Ddfr (%)
Ddft P/")
.0.5 0.0 0.5
4.4
*[ tt
Lateral Dsplacement (mm)
-0.6 .0.4 4.2
't.0
p#
zT
o
I
I
o
I
5
c
o
F
L_\
F
-50 -40 -30 -20 -10
20
(b) FG:L3.0-w1s-P3
(a) FG:L3.0-W20-P3
z!
10
Lateral Displacernent (mm)
0 10 20 30 40
-[
*[
'*f
t*F
,*L
50
0.0
lnitial
Tcftbn
Fd6 -
760 ldJ
-15-10-5051015?{J
Latsral Displacemnt (mm)
FG:L1.8-WI5-P2
(f) FG:L1.8-W15-P3
(e) FG:LI.8-Wl5-P2
Fig.S.4-Fully grouted walls; Tendon force histories (all tendons)

stressing force remained. Fig. 5.4(f) shows the prestressing tendon force history for FG:LI.8-
Wl5-P3 and reveals no prestressing force loss before failure.
5.4.2 Partially and ungrouted walls

These walls clearly exhibited shear dominated response, signified by development of shear
cracks spanning diagonally across the wall face. Flexural action was observed at force levels
-96-
Drift (%)
150
"0.8 -0.6 -0.4
4.2 0.0 0.2
0.4
0.6
0.8
100
xz
(fl.
(!
-50
o
o
dl
------c=-
.100
-150
.25
-m
125
300
1m
200
50
400
100
/'
V=9rkN
-15-r0-5051015?o25
-'t00
-200
-300
ts
=5
Eb25
oo
FSU
da
fl
co
300
20
so
r00
o
.100
s'50
-ro
.200
-100
.300
-4m
-125
-r0-8-6-4-20246
(b) UG:L1.8-w10-P2
(a) PG:L3.0-Wl5-P2
Fig. S.S-Partially and ungrouted walls; Forcedisplacement histories

up to approximately nominal flexural strength, V1,
BS
indicated in Fig. 5.5. Distress of the flex-
ural compression zone was not observed and tendon yielding never occurred.

Fig. 5.5(a) shows the force-displacement history for PG:L3.O-WI5-P2. At displacements of
approximately +3.91-2.3 mm, diagonal shear cracking initiated with significant strength degradation in both directions of loading, however PG:L3.O-Wl5-P2 did not exhibit sudden failure
at this point and retained more than 600/o of its maximum strength up to displacements of
approximately t10 mm.
Wall UG:L1.8-W10-P2 initially exhibited flexural and rocking response, signified by opening
of one single crack at the base of the wall for load levels up to Vs. Inclined shear cracking then
initiated, resulting in wall failure. Fig. 5.5(b) shows the non-linear elastic wall response
of
UG:L1.8-W1,0-P2 with nearly symmetrical response in the push and pull directions for applied
lateral forces up to V1

The failure modes for both the PG and UG walls were characterised by diagonal shear cracking
initiated by tension splitting along the axis of the diagonal strut. For the PG wall, the cracks
followed the vertical and horizontal mortar joints as shown in Fig. 5.6(a). For the UG wall the
cracks took a more direct path through the concrete masonry units, as shown in Fig. 5.6(b).
This failure mode was anticipated for both walls, despite prediction of nominal shear strength
higher than nominal flexural strength (Table 5.2), because the relatively high steel content was
-97 -
Push
Push
(o)
(b)
PG:1J.0-wl5-P2
uc:Lr .E-wr o-P2
S.fPartially and ungrouted wallsl Extent of damage at failure

to develop significant flexural overstrength, thus exceeding the nominal shear
Fig.
predicted
strength.
5.4.2.3 Sliding
Sliding was recorded in both wall tests, although the sliding displacement at any given stage
amounted to less than
5%o
of the total displacement for the PG wall and to less than
7o/o
of the
total displacement for the UG wall.

Significant tendon force increase due
to flexural action was measrued during testing of
PG:L3.0-WL5-P2, prior to the onset of stength degradation, but tendon force eccentricity was
determined to be insignificant. Testing of UG:L1.8-W10-P2 showed little prestressing force
variation and the associated tendon force eccentricity was insignificant compared with the wall
length.

Vertical strains in the flexural compression zones (the toe) of the walls were evaluated from
data acquired during testing for the following wall states:
(l)
at nominal flexural strength, Vs,
(2) at the peak of the excursion causing the first signs of wall toe distress, (3) at maximum
strength,
V**,
and (4) at ultimate displacement, du. It is noted that (1) was based on Eqns.
8.17, 8.18 and 8.29 and (2) was based purely on visual inspection. Refer to Thble 5.2 for Vs,
Vn'u* and du. The measured strains
in the extreme compression fibre at the wall states described
above are listed in Table 5.3 as the average for the push and pull directionsn as available. The
strains were measured in the lowest 200 mm of the walls.
-98-
Drift(%)
0.0 0.5 1.0
-1.5 -1.0 .O.5
H==v
600
^z
5
400
269
6
go
tt,
E
6o
1.5
- - FG:L3.GWa}P3
-- FG:L3,GWl3,P3
FG:L3,0-W13.P2C
-.5 - FG|UISW1*2E
FG:L1_8;W15.P2
+ F6:L.t.&Wr3-P3
-..-+_.<-4
-zoo
400
00
-50
40 -30 -20 -10
Lateral Displacemont (mm)
Fig. S.7-Fully grouted walls; F-D envelopes
5.5 DISCUSSION
This section discusses how design parameters, such as height, length, preshessing and masoDry
properties affect the wall behaviour in terms of flexural strength, displacement capacity, tendon
stress, masonry vertical strain, shear strength and sliding propensity. Selection
of appropriate
axial load ratio is also discussed.

5.5.1.1 Flexural strength
The wall flexural strength (maximum moment) is dependent on the wall dimensions, in partic-
ular wall length
l* This is illustrated in Fig. 5.7, which shows the force-displacement enve-
lopes (FDE) for all walls. Walls FG:LI.8-Wl5-P2 and FG:L3.0-WI5-P2C, which had similar
parameters except for wall length, clearly showed that wall strength increased from about 185
TABLE S.3-Measured masonrv vertical strain

Sbains mr nsured at
Wall
l:
at first dis-
I
Vf I tress
V,ol
0.0253
FG:L3.0-W20-P3
0.0025
2: FG:L3.0-Wl5-P3
0.0027
0.0067
0.0137
0.0228
3: FG:L3.0-Wl5-P2C
0.0039
0.0104
0.0104
0.0168
4: FG;L3.GWl5-P2E
0.0027
0.0104
0.0104
0.01l2
5: FG:LI.8-Wl5-P2
0.0033
0.0074
0.0224
0.0192
6: FG:L1.8-Wl5-P3
0.0031
/: ru:LJ.0-wl5-PZ
0.0016
8: UG:L1.8-WlGP2
0.0018
-99 -
Drift (%)
-1.5 -1.0 -0.5
0.0
t_5
t
o
tu
a
s
o
Vnom
1.0
0.5
1.0
1.5
,''--fi*+-
0.0
FG:L3.GWm.P3
- -
{,J
TD
-1.0
\--\
rr
.1.5
-50 -40 .30 -z)
--.FG:L3-GW1#3
- -
FG:L3.Gt/yt59zc
FG:l-3.0'Wt5+:lE
FG.Lt.6,W15+2
FGiLl.&W15f3
-10 0 10 2A 30 40
50
Fig. 5.8-Fully grouted walls; F-D envelopes, normalised with Vp
kl.l to approximately 367 kN. The effect of wall thickness on wall strength is illustrated by
comparison of the FDEs for FG:L3.0-W20-P3 and FG:L3.0-W15-P3. It is seen that for these
walls, having the same quantity of prestressing steel, the thicker wall (W20) ultimately developed higher shength than the thin wall, ca. 549 kN vs. 472kJ{.
The effect of the reinforcing ratio (area of presfessing steel,
area,
Ar:
pp,
\,
relative to the gross sectional
Apr/Ar) on wall strength can be illustrated by comparison of walls FG:L3.0-
Wl5-P3 and FG:L3.0-W15-P2C, which share the same length and thickness. Similarly walls
FG:L1.8-WL5-P2 and FG:L1.8-WI5-P3 may be compared. From Fig. 5.7 and Table 5.2
it
is
clear that the walls with the highest reinforcing ratio (3 prestessing tendons) developed the
highest strength.
Fig. 5.8 shows the wall FDEs normalised with predicted nominal flexural stength in terms
base shear, Vr.
It is noted that the actual wall strengths for all walls
of
exceeded the nominal
strength with a margin of 5o/o to 66Yo. Walls with low initial prestressing tendon stresso e.g.
FG:L3.0-W20-P3, developed the largest overstrength, emphasising the significance of tendon
reserve elastic capacity.
5.5.1.2 Ultimate drift capacity

The wall ultimate drift capacity, %, is dependent on wall dimensions. An increase of wall
length is likely to reduce the displacement capacity because a longer wall tends to develop sig-
nificant tendon force increase at a smaller displacement. This tendon force increase is then
likely to promote distress of the masonry compression zone at a lower displacement. Comparison of wall tests FG:L3.0-Wl5-P2C and FG:L1.8-W15-P2 in Fig. 5.8 suggests that for walls
-100-
with identical po, and location of presfressing steel, the shorter wall had considerably larger
ultimate displacement capacity. Increase of wall height is expected to increase the wall ultimate displacement capacity, partially due to the longer unbonded length of the prestressing
tendons. The ultimate drift capacity thus increases with an increase in wall aspect ratio,
.\ :
hJl*
An increase of reinforcing ratio, pp' is expected to reduce the wall ultimate drift capacity as a
consequence
of an increase of the flexural compression zone length, which imposes higher
strain on the extreme masonry fibre at any given wall displacement. This behaviour is illustrated by comparison of walls FG:L3.0-Wl5-P3 and FG:L3.O-WI5-P2C, which share the same
masonry dimensions. The initial prestress in the tendons is not expected to influence the wall
ultimate displacement capacity greatly because yu is rather controlled by the magnitude of the
tendon yield force.
5.5.1.3 Tendon yielding

Tendon yielding was generally expected for all walls because of the relatively short unbonded
tendon lengths, and was recorded in all tests except FG:L1.8-W15-P3. The wall drift corresponding to tendon yield is largely affected by three parameters, wall aspect ratio, A", tendon
initial stress and tendon location.
A longer wall typically results in a lower tendon yield displacement
because of comparatively
larger tendon strains at a given lateral displacement. An increase in wall height results in
longer unbonded tendon length and therefore in higher tendon elongation and base rotation
before yield. Thus an increase of the wall aspect ratio, A,, results in higher wall drift before
tendon yield, assuming that the tendons remain unbonded over the entire wall height.
In order to effectively avoid yielding of the prestressing tendons, it is typically necessary to

keep the tendons unbonded over several storeys and to keep the initial tendon sfess relatively
low, say in the range of 40%-60% of the tendon yield stress. In typical masonry structures,
it
may not be feasible to require elastic tendon response at high displacements. Still, adequate
wall performance can be achieved despite yielding of tendons, as has been demonstrated by the
research described herein.
It is noted that perhaps 50% the restoring moment for typical struc-
tural walls originates from the selfireight of the wall and supported floors. A parallel can be
drawn to research by Kurama et al. [5-5], who investigated unbonded post-tensioned concrete
101
walls and concluded that adequate seismic performance could be achieved despite yielding of
tendons.
Also, the tendon distribution along the length of the wall significantly affects the tendon yield
displacement. Considering the two extreme tendon locations, at the wall ends and at the wall
centreline, it becomes evident that the strain in the tendon on the wall centreline is approximately half of the strain in the tendon at the wall end for a given lateral displacement, assuming
wall rocking about the comers. It may not be feasible for masonry walls to have all tendons
placed near the wall centreline. Use of distributed tendons may be necessary due to considera-
tion of e.g. wall penetrations and out-of-plane forces.
5.5.1.4 Vertical masonry strain

Table 5.3 shows the vertical strain recorded in the walls at different stages of the tests. It is seen
that the calculated masonry strains corresponding to the nominal flexural condition ranged
between 0.0025 and 0.0039. These results suggest that the above equations capture the codedefined nominal flexural strength condition in a realistic and reasonably consistent fashion. An
average strain of 0.0030 was found when discarding the two extreme values. This result can be
compared with the strain limit at nominal flexural strength (unconfined masonry) given by the
New Zealand masonry code, NZS 4230:1990, of 0.0025, refer to Chapter 3.

Masonry strains associated with the first sign of toe reglon distress may be regarded as an alternative definition of the ultimate flexural condition.
It is seen from
Table 5.3 that these strain
values vary between 0.0067 and 0.0104, about 2 to 3 times higher than those associated with
nominal flexural strength.

Masonry strains associated with the maximum wall strength were for walls FG:L3.0-WI5-P2C
and FG:L3.0-WI5-P2E similar to those associated with the first visual sign of toe distress.
Walls FG:L3.0-W15-P3 and FG:L1.8-W15-P2 showed that the masonry strains at maximum
strength were three to four times as high as those associated with the fust visual sign of toe distress.
Thble 5.3 indicates that masonry extreme fibre strains ranged between 1.1% and 2.5o/o atthe
ultimate displacement. These values are of course of theoretical character as the toe regions
were damaged at this stage, so that those strains should not be regarded as the reliable ultimate
strain capacity of concrete masonry. The results do nevertheless suggest that masonry strains
t02 -
far beyond those related to nominal flexural strength can be expected for rocking wall systems,
while still providing significant axial strength.
5.5.1.5 Masonry shear strength

As suggested by the lest results and Eqns. 5.1 to 5.3, the masonry shear strength is greatly
enhanced by the additional axial force due to prestressing. Previous research on clay masonry
describes similar increase in masonry shear strength due to axial prestressing, e.g. Page and
Huizer t5-61. The first wall test, FG:L3.0-W20-P3, showed that the embedded horizontal shear
reinforcement was not engaged at all. Therefore, suf,ficient shear strength from the masonry
alone was assumed and the rest of the walls were constructed with no shear reinforcement.
Five of the six frrlly grouted walls failed in flexure and only one wall failed in a brittle manner.
It was unexpected that wall FG:LI.8-W15-P3 would fail due to tensile splitting in the diagonal
compression struts, given the satisfactory performance of FG:LI.8-Wl5-P2. The lone diagonal
crack that occurred in FG:L3.0-WI5-P3 may indicate that this wall was close to failing in a
similar manner.
Table 5.4 shows masonry shear strength predictions for the six FG walls according to NEHRP,
Paulay and Priestley, and NZS 4230:1990 (shear reinforcement in W20 was not included in
strength prediction), and the maximum wall strength recorded during the tests. It is seen from
Table 5.4 that NZS 4230:1990 underestimates the actual masonry shear strength in all cases,
despite the failure mode. The Paulay and Priestley prediction appears to underestimate the
shear strength
of walls FG:L3.0-W20-P3 and FG:L3.0-WI5-P3, while providing an upper
bound for the rest of the FG walls. It is interesting that the Paulay and Priestley equation accu-
TABLE S.4-Masonry shear strength

Wall
l. FG:L3.0-W20-P3
Measured
demand
NEHRP
Vmax
Vnehm
vnn
vnzs
0.984{,
0.994
0.763
0.s76
Paulay and
Priestley
NZS
4230:1990
2. FG:L3.0-Wl5-P3
l.107*
1.199
o.947
0.576
3. FG:L3.0-Wl5-P2C
0.888'r
t.290
1.000
0.576
4. FG:L3.0-Wl5-P2E
0.ggg*
1.227
0.956
0.s76
5. FG:L1.8-Wl5-P2
0.706{'
t.317
1.040
0.576
6. FG:L1.8-Wl5-P3
1.056f
1.408
1.040
0.s76
7. PG:L3.0-Wl5-P2
0.635"
0.928"
0.952
0.720
0.545
l,278
1.040
0.576
MPa
MPa
MPa
MPa
8. UG:L1.8-WI0-P2
*flexural failure; "shear fai ure; -compression strut failure
-103-
rately predicts the strength at which FG:L1.8-Wl5-P3 failed. The NEHRP expression does in
all cases predict shear strengths higher than the maximum measured strengths.
It is noted that the geometry of the utilised concrete masonry units is likely to significantly
inlluence the masonry shear strength. For example, one can compare a wall constructed
entirely from masonry 'open end' units, resulting in horizontal discontinuity of the grout core,
and a wall constructed with H-units that allow both vertical and horizontal grout continuity.
It
is clear that the latter solution provides a more monolithic wall, being likely to possess higher
masoffy shear strength. Note however that no testing was conducted specifically to validate
this conclusion.
5.5.1.6 Sliding propensity

Wall sliding was generally not a problem. However, for wall FG:L3.0-W15-P2E significant
sliding occurred. This phenomenon may (and should) be eliminated by ensuring adequate
shear friction between wall and foundation by requiring a certain magnitude of roughness
of
the wall to foundation interface, refer to Chapter 8 for an in-depth discussion. Alternatively, the
wall could be recessed for example 100 mm into the foundation beam, or dowel bars could be
incorporated in the construction joint to arrest any sliding.
5.5.1.7 Axial load ratio

The initial axial load ratio in the wall highly affects the wall performance. Realistic ratios are
found in the interval between, say, 0.05 and 0.20. Self-weight constitutes a significant component of the axial load ratio. In a multi-storey masonry building, self-weight of the above floors
may cause masonry stress at the foundation level of 1.0 MPa or more, i.e. an axial load ratio
component of 0.06 for
l8 MPa concrete masonry.
The lower bound is due to practical considerations in terms of prestressing losses (i.e. shrinkage, creep, etc.) and due to serviceability
limit
state considerations such as first cracking and
lateral wall stiffness. The lowest ratio for this test series was 0.077 for wall FG:L3.0-W15P2C, which performed satisfactorily.
The upper bound serves to keep the axial reinforcing ratio reasonably low. A high ratio also
reduces the efficiency of the wall in terms of flexural strength, due to a relatively long compression zone, and reduces the wall displacement capacity. It is furthermore difficult to achieve
satisfactory distribution of high prestressing anchorage forces in a concrete masonry wall, e.g.
104 -
Page and Shrive [5-7]. Wall FG:LI.8-W15-P3, t]rat failed due to tensile splitting
of diagonal
compression struts, had a relatively high axial force ratio of 0.167. The concentration of pre-
stessing tendons in the middle of that wall may have caused localised stress concentration in
the top of the wall, potentially initiating the diagonal cracking.

5.5.2.1 Masonry shear strength
Prediction of shear strength of partially grouted and ungrouted walls follows the same principles as outlined for fully grouted walls, using the initial total tendon force, P. Evaluation of the
wall width for the purpose of calculating shear strength for partially and ungrouted masonry
should, however, be based on the sum of the masonry unit flange thicknesses. Singular grouted
cells or masonry unit webs are not included in the shear flow through the panel and therefore
cannot be expected to carry shear (Paulay and Priestley
[5-a[. For walls PG:L3.0-WI5-P2
and
UG:L1.8-W10-P2, the effective wall thickness for shear calculation amounted to 63 mm and
6l
mm respectively.
Partially grouted and ungrouted wall shear demand and masonry shear strength comparison is
provided in Thble 5.4. The shear demand is determined from the maximum shear carried by the
walls before shear failure. Predicted masonry shear strength was calculated according to Eqns.
5.1 through 5.3 using a masonry crushing strength
of 13.3 MPa for PG:L3.0-WS-P2 and 15.0
MPa for UG:LI.8-Wl0-P2. It is noted that the masonry crushing strength for PG:L3.0-WI5P2, given in Table 5.1, was specific to a grouted cell, while the above value constitutes an average for the partially grouted wall.
It is seen from
Table 5.4 that the masonry shear strength predicted by NZS 4230:1990, Eqn.
5.3, gives conservative estimates of actual masonry shear strength. The prediction was reason-
ably close for the squat wall (L3.0) while rather low for the slender wall (L1.8). Comparison
of
test results with the masonry shear strengths predicted by the NEHRP provisions and that
advocated by Paulay and Priestley, indicate that
Eqns.5.l and
5.20 overestimate the actual
shear strength considerably, and are therefore unconservative for partially and ungrouted walls.
-105-
Ddfr (yo)
4.25
-0.75
1.5
0.00
1"*l----
Vnorm
1.0
b
o
600
o
E o.s
d)
/r
t.0
-1.5
.15 -10 -5
10
15
Fig.5.9-PG and UG Wallsn F-I) envelopes, normalised with VF
5.5.2.2 Flexural strength

Fig. 5.9 shows the force-displacement envelopes for PG:L3.0-WI5-P2 and UG:L1.8-Wl0-P2,
normalised with respect to Vs.
exceeded V1. This is
It is
observed for both walls that the measured wall strength
of significant relevance when conducting non-ductile strength design,
where nominal flexural strength must exceed the flexural demand, but there is no requirement
for reliable inelastic response.
5.5.2.3 Verticalstrain
The masonry vertical strain was evaluated at nominal flexural strength, which was achieved for
both walls. Table 5.3 shows that strains of 0.0016 to 0.0018 were measured at the predicted
nominal shength, Mn.
5.5.2.4 Axial load ratio

The axial load ratio for partially and ungrouted walls directly affects both the nominal flexural
strength and the masonry shear strength, as suggested by Eqns. 8.17 and 5. I to 5.3. These walls
are not expected to perform
in a ductile manner, therefore there is no need to provide reservc
tendon strain capacity allowing non-linear elastic rocking behaviour- Consequently, the prestressing steel area should be determined from the desired serviceability limit state capacity or
nominal flexural capacity using the maximum allowed tendon stress, taking due consideration
for prestress losses.
106 -
The axial load ratio that partially and ungrouted walls can sustain is probably of the same mag-
nitude as that suggested for fully grouted walls. It is however emphasised that proper diffirsion
of the concentrated force under the prestressing anchorages is critical.
5.6 CONCLUSIONS
A]\ID FUTURE RESEARCII
It is concluded that fully
grouted unbonded post-tensioned concrete masonry is a competent
material combination for ductile structural wall systems. The PCM walls reported herein
exhibited a nearly non-linear elastic behaviour dominated by rocking response. Large drift
capacity of more than I.4% was measured and still larger drift capacity can be expected from
higher walls.
Only localised damage occurred, as shown in Fig. 5.3, making earttrquake damage easy to
repair. Possible tendon force loss could be compensated for by re-stressing of the tendons. This
appears feasible because tendon strains are expected to stay below lVo for typical
wall dimen-
sions and realistic drift ratio demands when tendon yielding occurs over the entire unbonded
lenglh.
Tendon force loss due to yielding should, for short (squat) walls, be expected at relatively low
wall drift ratios. Even after tendon yielding, reliable and self-centring wall behaviour is
expected.
Measurements at nominal flexural strength, Mn, suggest an extreme masonry fibre sfain of the
order of 0.003. Strains as high as 0.007 may be considered for design if the tme ultimate condi-
tion is assumed as being the stage at which toe distress initiates.

Clearly, the additional axial load from the prestressing enhanced the concrete masonry shear
stength.
It
was concluded that fully grouted PCM walls of dimensions similar to those
described above, in most cases did not need horizontal shear reinforcement in order to develop
flexural/rocking response. However, one fully grouted wall with high axial presffessing force
failed by tensile splitting of diagonal compression struts, suggesting that this issue should be
pursued further.
The Paulay and Priestley [5-4] recofilmendation for masonry shear strength accurately predicted the strength of FG:L1.8-WI5-P3, however this failure was atfibuted not atfibuted
-r07-
directly to shear but to tensile splitting of diagonal compression struts. The NZS4230:1990
shear strength provision is highly conservative and seems to unsatisfactorily reflect the large
increase in masonry shear stength due to the axial prestressing force.
Relatively little energy dissipation was observed during wall cycling. This may impact the use
of such walls for ductile seismic design in the sense that additional damping may be required to
control the structural drift demand in an earthquake event. Further research is needed to investigate the relationship betrveen PCM wall shength and drift demand arising from seismic exci-
tation.

Testing clearly showed that partially and ungrouted PCM walls were capable of developing
significant strength, exceeding the predicted nominal flexural strength. This suggests that par-
tially and ungrouted PCM walls may suitably be designed using strength design.
The wall tests initially displayed flexural behaviour but ultimately developed shear response.
Diagonal shear cracking was the failure mode. Despite this brittle failure mode, significant displacement capacity was recorded. These walls did not develop reliable nonJinear elastic
behaviour.
The additional axial force due to prestressing clearly enhanced the concrete masonry shear
strength. Test results suggest that none of the shear strength equations perfonned well.
It
is,
however, safe to use the NZS 4230:1990 code requirements when predicting concrete masonry
shear strenglh for ungrouted and partially grouted walls, basing masonry shear strength on the
concrete masonry unit flange thickness.
-108-
5.7 REFERENCES
[5-l]
Park, R., Evaluation of Ductility of Stnrctures and Structural Assemblages
for Labora-
tory Tbsting, Bulletin of NZNSEE, Vol. 22, No. 3, Sept. 1989, pp. 155-166.
[5-2] NZS 4230:1990, Code of Practice
for
Association of New Zealand, Wellington.
[5-3] NEHRP, Recommended Provisionsfor Seismic Regulationsfor New Buildings and Other
Structures, Part
I - Provisions,
Building Seismic Safety Council, Washington, D.C.,
U.S.A., 1997, pp. 159-162.
[54]
Paulay, T., and Priestley, M.J.N., Seismic Design of Reinforced Concrete and Masonry
Buildings, John Wiley & Sons Inc., L992.
[5-5] Kurama, Y., Sause, R., Pessiki, S., and Lu, L-W, Lateral Load Behavior and Seismic
Design of Unbonded Post-tensioned Precast Concrete Walls, ACI Structural Journal,
July-August 1999, pp. 622-632.
[5-6] Page, A.W. and Huizer, A., Racking Tbsts on Prestressed Hollow Clay Masonry Walls,
Proceedings of the 8ft International Brick/Block Masonry Conference, September 1988,
Dublin Ireland, pp. 538-547.
[5-7] Page, A.W. and Shrive, N.G., Concentrated Loads on Hollow Concrete Masonry, ACI
Stnrctural Journal, July-August I 990, pp. 43 6444.
- 109-
-ll0-
Chapter 6
STRUCTURAL TESTING - SERIES 2
6. ENHANCBD
SINGLE-STOREY PCM WALLS
In-plane response of post-tensioned concrete masonry walls with unbonded tendons, incorporating strengthened masonry and enhanced energy dissipation, is examined by means of stuc-
tural testing. An introduction to the Series 2 wall testing programme is followed by

presentation of the results from structural testing of the five fully grouted walls. Discussion
of
the results is concerned with wall structural response in terms of flexural strength, displacement capaciry tendon sfiess and masonry vertical strain, and makes comparison with Series I
testing.
6.1 INTRODUCTION
Previous testing presented
in
Chapter
5 explored post-tensioned concrete masonry
(PCNO
walls made with un-confined masonry. It was found that such walls typically had reliable displacement capacity (drift capacrf) of up to 17o, however,
in some cases the limited
strain
capacity of un-confined masonry resulted in masonry crushing and subsequent strength degra-
dation, and prohibited development of further reliable drift capacity.

One of the principal intents of this testing series was to irwestigate means of masonry strength-
ening that were expected to allow for reliable drift capacity, say of the order
demands
of
such magnitude are likely
to be encountered in structural
of 1.5%. Drift
systems utilising
unbonded post-tensioned walls when subjected to seismic loading because of the inherent lim-
ited structural damping, see Chapter 9.
As a consequence of limited hysteretic energy dissipation, additional structural damping may

be required to control the wall displacement demand during seismic activity. As a potentially
simple solution, 'dog-bone'-type dampers were incorporated in one wall test to explore the
efficiency of such a device.
lil
Two of the tested walls incorporated a hydraulic system for application of constant axial load
in the prestressing bars, thus entirely avoiding prestress loss due to tendon yielding.
The single-storey wall testing results reported herein enabled further development and verifica-
tion of structural behaviour prediction methods for prestressed concrete masonry walls which
are currently being developed at the University of Auckland. Recommendations from this test-
ing phase, having explored strengthening and energy dissipation devices, are incorporated in
the design and testing of the two large scale multi-storey walls, as reported in Chapter 7. These
213-scale walls represented a realistic 4-5 storey masonry prototype structure.
testing Series
ln that respect,
and 2 may be regarded as component testing enabling the author to select and
design an appropriate structural configuration to be used for the large scale testing.
Detailed descriptions of materials, wall construction, test setup, instrumentation and testing
procedure, pertaining to similar testing, may be found in Chapter 5.

Wall dimensions are specified in Thble 6.1. The wall designations follow the convention presented in the previous chapter. Further additions are: CP, HB, ED and CA, which represent
confining plates, high strength blocks, enerry dissipation and constant axial load, respectively.
In the following discussion of Series 2 walls, the designations FG:L3.0-W15 have been left out
for convenience, however, when comparing with Series
results the proper wall designations
have been resumed.
TABLE 6.1-Wall specifi cations

initial
initial
steel
presrr.
axtal
force
stress
force
ratio-
fo,
fm/f'm
0.044
thick-
self
lless
weight
nasonry
strength
lw
bw
f'-o
l. FG:L3.0-Wl5-Pl-CP
3.0
40
4T
18.2
720
299
2.
3.
4.
5.
FG:L3.0-W15-P2-CP
3.0
40
r5.I
703
584
0.099
FG:L3.0-Wl5-P2-CP-CA
3.0
40
20.6
708
588
0.073
FG:L3.0-W15-P2-CP-CA-ED
3.0
40
t7.8
727
604
0.086
FG:L3.0-Wl5-P2-HB
3.0
40
4l
4l
4l
4l
12.5
743
6t7
0.1 25
MPa
KN
Wall
length
'unconfined strength, full height units,
-f.:
mm
(P+Ny(lwbw)
KN
-ttz-
MPa
endons
Woll bose
joint confinement plofes

25mm x 3.2mm flof mlld sl.6l
Gr. 250, full pcnctrollon reldcd
Bed
Section A -
Confinemenl olole
2xHD10 Sheor
reinf orcemenf '
:f, eor.u -,fl

TF--J
rF;rEf
rs2.s/
l5 series unif ,
2 cells, f ull
2oo2oo
helght
* Wolls CA ond CA-ED only

*t Woll CA-ED only
*r* Full helght unils both
15 series unit,
1cell. full
height
ends for PI-CP
Preslressing d
15 serles unif.
helght*"
'oT
lr
Bed joint
conflnemenl ploies
Energy dissipolion
borsr*
Fig.6.l-Wall geometry for CP walls (number I - 4)

All test units were constructed to the same dimensions, a length, l* of 3.0 m, a height, h* of
2.6 m and a thickness,
ness
b*
of 0.14 m using l5 series concrete masonry units. The actual thick-
of the walls was approximately 10 mm less than the nominal masonry unit thickness of
150 mm. The main part
of each wall was constructed with full height concrete masonry units
(nominally 200 mm high). One lower corner of all P2-CP walls (corner in compression under
push loading) was constructed with half-height concrete masonry units (100 mm nominal
height), as indicated in Fig. 6.1, with an extend of 600 mm vertically and 1200 mm horizontally. The asymmetrical use of half height units in the bottom corners enabled a direct comparison between the effectiveness of different confining plate details. The half height units were
obtained from the masonry block supplier, though some webs needed to be modified to accommodate the shear reinforcement. Both lower corner units of the HB wall were fabricated spe-
cially for this test using high strength steel fibre reinforced concrete, as shown in Fig. 6.2.
Confining plates embedded in the toe regions (flexural compression zones) are indicated on
Fig. 6.1. These plates were placed in all bed joints in the lowest 600 mm above the foundation,
embedded with mortar on both sides for best integrity. In the following U200, CP200 and
CP100 rofer to 'unconfrned full height $outed concrete masonry' (main body of walls), 'con-
-113-
Woll bose
Seclion A
Speclol high
slrength unll
2xHDlO
in
top
courcc only
1
5 serles end
unit, 2 cells
1
5 series end
unit, I
cell
Prestresslng duct
Speciol high
strength unil
Fig.
^r
la
6.2-Wall geometry for HB wall (number
5)
fined full height grouted concrete masonry' (confined region bottom left corner in Fig. 6.1) and
'confined half height grouted concrete masoffy' (confined region bottom right corner in Fig.
6.1), respectively. Further information on enhancement of the masonry performance caused by
confining plates may be found in Chapter 3.
High stength comer units were manufactured to ttre exact dimension of ordinary 15 Series
concrete masonry units. Using these units, it was attempted to provide durable pivoting points
for wall rocking, thus isolating the above units from extreme strain. The high strength steel
fibre concrete measured an average cube srength of 131 MPa, further details below.
Walls P2-CP-CA and P2-CP-CA-ED had 2
x HD10 (10 mm
bar, 430 MPa nominal yield
strength) embedded horizontally at 400 mm vertical intervals. No shear reinforcing was provided in the other walls (except for 2x HDl0 placed in top course to provide a bond beam).
Energy dissipators were embedded centrally in P2-CP-CA-ED as shown in Fig. 6.1. These
were of the so called 'dog bone'-Upe, and were expected to provide hysteretic damping as a
result of yielding in tension and compression. The milled part of the bars were designed such
that the axial prestressing force (kept constant throughout testing) comfortably could yield the
bars in compression, thus forcing the wall back to its original alignment upon unloading. An
-114-
Bor Diomeler reduced to 16mm
20 --rl
Sfeel
rl/
Unbonded
Lengfh \
ol
'l
confining tube
+
T
-Tol
Rl-==---Embedmenf
I
Fig. 6.3-Wall CP-CA-ED energSr dissipation bars
energy dissipator yield force to prestressing force ratio of approximately
l:3 was chosen for
this test, reflected by nominal yield strength of the reduced diameter section of each energy
dissipation bar of approximately 100 kN and axial force in the wall due to the prestressing
of
617 kN. Fig. 6.3 illushates the dimensions of the energy dissipation bars. The bars were made
of 20 mm reinforcing steel with nominal yield strength of

rangingbetween
&:500
MPa and ultimate strength
l.l5&and 1.4t Thebardiameterwasreducedto 16mmoveralength of 375
mm, to ensure that the bar yielded in a well defined region adjacent to the wall base, and to
limit strain such that an extreme wall displacement excursion would not cause bar rupture. The
bar was confined by heavy steel tube with an inside diameter of 27 mmand a length of 555 mm
to ensure that the section of bar intended to yield would not buckle when put into compression.
The tube was filled with dental plaster and the dissipation bars remained unbonded over a
length of 465 mm, as shown in Fig. 6.3.
All walls were preshessed with two 23 mm VSL thread bars placed concentrically
at +400 mm
from the wall centreline. The prestressing tendons remained unbonded over the entire wall
height and had a typical unbonded length of 3.5 m, including unbonded lengths in the founda-
tion (300 mm), loading beam (400 mm) and load cells on top of the loading beam (200 mm).
An additional tendon length of 300 mm due to hydraulic jacks in the vertical loading system
applied to the CA walls. The wall initial prestressing force and axial force ratios (due to prestressing and self-weight) are given in Table 6.1. Unfortunately, the left prestressing bars (bar
closest to strong wall) in test unit Pl-CP tore out of the foundation anchorage during the
stressing, thus the designation
'Pl'.
initial
The damage could not be repaired with the wall in the test-
-115-
ing setup. As a result an alternative testing procedure was instated for this wall, consisting of
testing the wall in the pull direction only to maximise the prestressing bar eccentricity.
Constant axial load in the prestressing bars was applied for the walls CP-CA and CP-CA-ED.
For these particular walls, the prestressing force was kept constant, such that the bars did not
yield
as the
wall was displaced laterally. Two identical hydraulic jacks were connected in paral-
lel to the same pump and remained in place throughout the test. This ensured that the force
applied in each bar was kept the same and constant as the wall was rocked back and forth.

Prior to wall consffuction, the prestressing bars were inserted into the foundation and encased
by ducting to ensure proper isolation from the grout. The walls were constructed in running
bond by experienced blocklayers under supervision, using standard grey precast concrete
masonry units and DRICONTM trade mortar. Open-ended concrete masoffy units were used to
avoid having to thread the masonry units onto the 4 m long prestressing bars. Grouting with
17.5 MPa ready-mixed regular block
fill, containing SIKA CAVEXTM
expansion agenf was
generally performed the following day. Vibration of the grout was carried out for the CA walls.

Average masonry material properties were determined by material testing, typically in samples
of three. The masonry crushing strength
f'., elastic modulus E., grout crushing strength f',
(200 mm high, 100 mm diam. cylinder), and average age at day of testing are given in Table
6.2.
High strength masonry units were fabricated with steel fibre reinforced concrete of the following mix ratio:
l0l0 kg of dry silica sand, 705 kg of ordinary portland
cement, 230 kg Microsil-
TABLE 6.2-Material properties

Wall
f,n
E,,,
l. FG:L3.0-WI5-Pl-CP
u200 lcP200lcPl00 u200 lcP200lcP100

18.2
2r.7
2.
3.
4.
5.
20.6
t 7.8
12.5*
19.8
19.9
18.6
MPa
MPa
MPa
FG:L3.0-Wl5-P2-CP
FG:L3.0-W I 5-P2-CP-CA
FG:L3.0-W I 5-P2-CP-CA-ED
FG:L3.0-Wl5-P2-HB
I CAVEX
t5.l
fls
Age at
18.6
DOT
l3
r6.2
t5.0
t4.9
3.8
14.3
r3.6
GPa
GPa
GPa
3l
19.0
not applied
-116-
63
38
MPa
32
days
ica 600, 210 kg of crushed silica sand, 16 kg of super plasticiser (Viscocrete-5, SIKA NZ,
Ltd.),200 kg of water and 60 kg ofFortex steel fibres produced by Novocon International (dosage rate of approximately 2.5o/o by weight). The mix resulted in an average 50 mm cube crush-
ing strength of 131 MPa and a standard deviation of 2 MPa.
VSL 23 mm diameter high stength threaded bars were used in all wall tests and had the typical
properties: yield strength of 970 MPa, ultimate strength
and modulus of elasticity
of
190 GPa,
of
1160 MPa, ultimate strain of
8o/o
all of which were confi.rmed by material testing.
6.3 TESTING DETAILS

6.3.1 Test setup
Wall testing reported in this chapter was conducted using the setup discussed in Chapter 5 and
shown in Fig 5.1(a), consisting of a reusable foundation and loading beam. Notably, no effort
was made to replicate gravity and live loads from suspended floors, with the assumption that
these loads were directly analogous to a component of the applied post-tensioning force.
Fig. 6.4 schematically shows typical wall instrumentation. Lateral load was measured by two
load cells positioned in series with the lateral hydraulic actuators, LCHI and LCH2, and lateral
displacement at the top of the wall (height of 2.8 m above base) was measured by the displacement transducers DISP1 and DISP2. Wall flexural deformation and vertical strain were meas-
ured by instruments denoted 'F'. Relative sliding displacements between loading beam/wall,
walVfoundation and foundation/strong floor were measured by instruments denoted 'S'.
'DIGI' denotes digital display displacement tansducers which were employed to confirm lateral displacement measurements. Axial force in the prestressing bars was measwed by the load
cells LCPRI and LCPR2. Flexural instrumentation F13-F24, shown in Fig. 6.4(a),
was
mounted on one side only, whereas instruments F1-F4, F9-Fl2 and Fl7-F28 shown in Fig.
6.4(b) were mirrored on the wall back side.
Dowel bars were embedded in the foundation for wall tests P2-CP-CA and P2-CP-CA-ED, as
indicated in Fig. 6.1. The dowels were made from HD20 reinforcing steel, had an embedment
length into the foundation of 200 mm and protruded 80 mm into the wall. The deformations
were removed from the 80 mm Iength and the bar was greased and encased in a sleeve to
ensure that no bond was formed between the dowel and surrounding grout.
rt7 -
E
9o
tsF
FA
f;*
g
(a) Walls PI-CP andP2=IIB
Push
.d>
9o
55
=
g-}
fi.:
tr
o
rr
ng
Troor
(b) \Valls m-CP" PZ-CF-CA aad P2-CP-CA-ED

X'ig.
6.#Ilplcal instrumentation
Sliding blocking devices were muntcd on ths foundation beam by each end of the wall. These
devices were intendetl to blsck
from
7O
mm for walls Pl -CR
wall sliding, should this occur. The height of the deviees waried
P2{B
P2-CP-CA and HBj to 150 mm for P2-CP-CA-ED. G,aps
- 118 -
of 3 mm (nominal) were present between the wall ends and the sliding blocking devices in the
unloaded state.
6.3.2 lbsting procedure

Cyclic structural testing was conducted according to the procedure outlined by Park [6-l], with
the ductility one displacement here termed 'displacement level one' or
Dl. After definition of
Dl, the walls were cycled toD2,D4,D6, etc. until failure, with Dn being defined as n times
Dl. Prior to executing this loading procedure, serviceability limit state level force cycles were
applied to some walls in order to investigate the uncracked stiffness, the lateral displacement
profile and the initial crack formation.

For a symmetric distribution of post-tensioning and gravity actions, the nominal flexural
stength, Mn, is defined by the New Zealand Masonry code NZS 4230:1990l6-21as given in
Eqns.8.17,8.18 and 8.29. Details on these equations can be found in Chapter 8. The base
shear, Vs, corresponding to Mn may be calculated using Eqn. 8.17, where P represents the
ini-
tial prestress force (no lateral load applied to wall), AP is the total tendon force increase due to
wall deformation at Vs,
lu,
is the wall length, lL the height of the wall and N is the axial load at
the base of the wall due to live and dead load. The equivalent rectangular compression zone
length, a, is defined by Eqn. 8.18 as a function of the axial forces on the section, the masonry
compressive strength,
f'.,
the wall width
b*
and cr
0.96 (confined grouted concrete
masonry). Eqn. 8.29 defines the effective tendon force eccentricity, e., with respect to the initial
prestressing force and the tendon force increase based on the assumption of uniformly dishibuted prestressing steel. tn the absence of suitable procedures for estimating AP at the time
of
testing, the tendon force increase at nominal strength was estimated to be 50 kN for walls P2-
CP and P2-HB and zero for PI-CP. Walls P2-CP-CA and P2-CP-CA-ED had constant axial
load applied to the prestressing bars, i.e.
AP:
0.
Flexural strength increase of wall P2-CP-CA-ED, due to the yielding energy dissipators, was
taken into account when estimating Vr by assuming a steel stress of f,
- l19-
: 500 MPa.

The predicted masonry shear strength,
V.
(force), was calculated according to Eqn. 5.2 as
advocated by Paulay and Priestley [6-3]. Shear strength due to shear reinforcing was calculated
using the equation:
Vrt
where
nrfrr4,
(6.1)
,\ is the steel area with centre to centre spacing s, d: 0.81* and fn, is the nominal yield
strength of the shear reinforcing. The predicted wall shear strengths V, =
V. *
V., are listed in
Thble 6.3. The initial prestress force was used in this calculation.
6.4 TEST RESULTS

This section describes the behaviour of the five walls and presents the measured masonry compression toe strains at various wall states. General wall behaviour is summarised in Table 6.3,
where Vr", is the maximum lateral force recorded and du*r* is the corresponding displacement.
The ultimate displacement capacity, du, is defined as the point at which the lateral wall strength
had degraded below 80% of V.u*. The ultimate drift capacity is defined as yu:4,/h"
with
4:
2.8 m.
Rocking response was recorded for all wall tests, with a single large crack opening up along
the wall-foundation construction joint. No distributed flexural cracks were observed. A singu-
lar inclined crack was observed for wall P2-CP as a result of loading in the pull direction.
similar crack was observed for wall P2-HB. Both of these walls contained no shear reinforcing.
Cracking occurred in the centre of wall CP-CA-ED, due to bond stress between the energy dis-
sipation bars and the masonry. High initial stiffiress was recorded.
At
nominal flexural

Wall
l.
FG:L3.0-Wl5-Pl-CP
Predictior
vf v3
216 325
2. FG:L3.&Wl5-P2-CP
3. FG:L3.0-Wl5-P2-CP-CA
4. FG:L3.0-Wl
5-P2-CP-CA-ED
5. FGrL3.0-Wls-P2-llB
Note: sliding displacement subtacted from
372
3ll
815
400
842
339
360
KN
KN
Results
Dir./conf. I ot
PulVCP200
Pusb/CPl00
PulVCP200
Push/CPl00
PulYCP200
Push/CPl00
PulVCP200
Push/HlJ
PUIVHB
1.7
2.5
3.0
4.3
2.5
mr
d,.", and du
-120-
lv*"1 a"*
-249
-7.1
J95 22.2
-345 -29.3
338
9.1
-301
4.0
4t4
-396
14.4
-7.0
5t6
0. 5
-384 -10.1
KN
mm
d"
-32.1
Comments
y"
30.4
-55. I
0.012 Masonry crushing, flexural

failure^ null test onlv
0.0r I Masonry crushing, diagonal
0.020 crack. flexural failure
73.1
0.426 Masonry crushing, flexural
18.8
0.007 flailrrre
0.0 2 Masonry crushing, fl exural

-35.0 0.0 J failure
J 1.4 0.0 I Masonry crushing, diagonal
-38.1 0.0 4 nrecL flevrrml failrrre
mm
34.5
stength,
Vr: Mnft*
wall softening occurred due to the initiation of rocking. Tendon yielding
was experienced for walls PI-CR P}-CP and P2-HB, which all had tendons permanently
anchored in the foundation and loading beam. The tendons in walls P2-CP-CA and P2-CP-CA-
ED remained elastic during testing.
regions, resulting in gradual strength degradation. Using CP200 masoffy, inelastic masonry
sffain were found to be concentrated in the lowest course (200 mm), whereas when using
CP100 masonry inelastic vertical strain penetrated into the second and third cotuse above the
foundation (200-300 mm). Failure of the compression toes for wall P2-HB was initiated by
crushing of the ordinary masoffy course (U200) just above the high strength units.
Considerable sliding displacement between wall and foundation was recorded for PI-CP, P2CP and P2-CP-CA.
Hysteretic energy dissipation was increased considerably for P2-CP-CA-ED because of the
embedded 'dog-bone' energy dissipation bars, in comparison with the other walls.
It is pointed out for PI-CP that, while results have been included in this chapter, relatively little
attention has been given to this wall due to its unrealistic dimensions and prestressing configuration.

Fig. 6.5, depicting the force-displacement response for all fully grouted walls, demonstrates
the nearly non-linear elastic behaviour exhibited by the tested unbonded post-tensioned walls
prior to toe crushing, with the exception of P2-CP-CA-ED. It is seen that all walls returned to
their original alignment, even after large displacement excursions. This type of response may
be categorised as 'origin oriented'. For all walls, except P2-CP-CA-ED, the individual curves
for each excursion appear pinched, implying little hysteretic energy dissipation. It is noted that
the term 'elastic' in a rigorous sense indicates reversibility and no accumulation of damage. In
reality the walls did accumulate damage from the onset of tendon yielding and masonry crush-
ing, causing a reduction in wall stiffness. Thus the behaviour was termed 'nearly' non-linear
elastic.
-tzr-
Drift (7o)
0.50
0.75
Drift (06)
r.25
1.00
'lr*
-lr*
250
zJ
l* Eb
EE
'lo*
l* HE
-l*
150
o
6
100
60
-.q
dl
50
-l
t*
'100
of
-300
-200
600$dt
-300
- {0 .50 -40 -3{t -20 -r0 0 l0
-1.5
9m
kN
-'r.0
.0.5
0.0
0.5
1.0
1.5
1200
,$O
6me ^*
22200
200
900
600
y,*
o Fg
6a
-soo ?
E
s'*
H
-600 d!
3oo--
roo
(!
300
at,
'1oo
-roo
-20iJ
-300
-600
Vr= 311 kN
-300
-10 0 10 20 30 40 50 60
-900
c
o
E
-90o
400
40 -30 -20 -10 0 10 20 30
70
zJE
co
-300
-30 -20
-1200
2t
Drift (%)
2-s
C{r
d)
U'333 !l'l- -gfi)

V. = 17t kN
(b) FG:L3.0-Wl5-P2-CP
Ddft e6)
E
o
Latsral Displacement (mm)
-r.0 -o-5 0.0 0.5 1.0 r.5 L0
-100
(a) FG:L3.0-W1s-Pl-CP
600 2
J
3oo
200
Latenal Displacement (mm)
$o ' V'=311
1mo
900
-400
r0
05101520253035
0.0
V=333kN
300
e
z2
l6m 5g
200
-g
400
-2.0 -1.s -1.0 -0.5

!clz3!!!
Lateral Dsphcement (mm)
-1200
4,0
Lateral Displacsment (mm)
(d) FG:L3.0-W I 5-P2-CP-CA-ED
(c) FG:L3.0-W I 5-P2-CP-CA

Dritt(%)
-1.5 -r.0 {.5 0.0 0.5

z!
s()
v,
1.0
1.5
400
4=s6oku
1200
300
V' = 3:19 kN
900
o02
3m;:
200
100
-o
U
0
-100
-300
-6{F
dl -mo
V=339kN
v" = 36o.kN
-100
-50
-40
-30 -20
-r0 0 r0 m 30 40
o
ft
c0
-900
.1200
50
Laleral Displacement (mm)
(e) FG:L3.0-Wl5-P2-HB
Fig. 6.S-Forcedisplacement histories

None of the walls exhibited symmetrical response in the push and pull directions. PI-CP was
tested in the pull direction only. PZ-CP was tested cyclically (alternating push and pull excursions) for displacements up to D2 (5 mm) and thereafter cyclically, first in the push direction
and then in the pull direction, because of an inclined crack (Fig. 6.6(b)) that occurred while
pulling to -D2 (-5 mm) which caused concern about the wall integrity in the pull direction.
-t22-
Continued testing in the push direction was adopted to at least capture large displacement
response in one direction.
All
P2-CP walls had asymmetric strengthening in the wall corners as
indicated in Fig. 6.1, thus different performance in push and pull directions. P2-HB had a symmetric layout as indicated in Fig. 6.2, nevertheless, its performance in the directions of loading
differed. The asymmetric behaviour of P2-CP-CA, as seen in Fig. 6.5(c), was partially caused
by extensive sliding in the push direction. Note that
(l)
all lateral displacement data depicted in
Fig. 6.5 was corrected for sliding displacement, thus representing only the wall flexural and
shear deformation behaviour and (2) the corrected lateral displacement has been used as refer-
ence in all graphs and tables in this chapter. Sliding displacement data is presented in section
6.4.3.

Gradual strength degradation was observed for all walls (excepting P2-HB), attributed to
spalling of face shells and crushing of the grout core. Large displacement capacity was
observed with reliable wall resistance for drifts,1u, ranging from0.1o/o to more than2.6o/o,refer
to Thble 6.3. A relatively more sudden sfength degradation was observed for P2-HB in the
push direction when going beyond D12 (30 mm), caused by rapid strength degradation of the
ordinary masonry block (U200) just above the high strength block (lower right wall corner as
shown in Fig. 6.6(e)). Crack pattems after failure, depicted schematically in Fig. 6.6, indicated
that the extent of damage was confined to the lowest two masonry courses in the toe regions,
with the majority of damage concentrated in the lowest courso. Damage to the masonry units
immediately above the lowest course generally was limited to vertical splitting cracking near
the wall end. The inclined cracking observed in P2-CP (no shear reinforcing) and P2-HB (no
shear reinforcing) suggests that shear reinforcement is necessary
for walls of present dimen-
sions, in order to limit sporadic cracking in the diagonal compression strut. Cracking observed
in P2-CP-CA-ED (with shear reinforcing) was mainly due to localised shess concentration
arising from mobilisation of the energy dissipator strength.
Figs. 6.6(b) and (e) showing inclined cracks that appeared while pulling P2-CP and P2-HB to -
D2 (-5 mm) and -D4 (-10 mm), respectively. These crack are believed to have been caused by
large localised splitting forces associated with the prestress anchorage at the centre of the loading beam. The cracks did not develop further during testing, possibly due to transverse restraint
provided by the loading beam.
-t23-
Push
__>
+Push
(b)
(o) Fc:13.0-w15-P1-CP
+Push
(c)
FG:13.0-wt 5-P2-cP
+Push
(d)
FG:13.0-Wt 5-P2-CP-CA
FG:13.0-Wl 5-P2-cP-CA-ED
Push
->
Crushlng of
ordlnory
foce
lost
(e) FG:15.0-W15-P2-HB
Fig.
6.fDamage accumulation at failure
-r24-
Drift(%)
-r.0 {r.5 0.0
.1.5
4.0
0.5
uF
I
E
E
"r
lo 'o8lc
E
o . l-l
o
o
ao
t:
4r-
,l;
oF
051015?s253035
-60 -50 -ro
-$
-20
.10 0 l0
20
Lateral Displaoement (mm)
(b) FG:L3.0-wl5-P2-CP
(a) FG:L3.0-Wl5-Pl-CP
Orift(%)
-'t.0
.30 .20
-t0 0 10 20 30 40 50 60
-0.5
0.0
-10 -30 -20 -10
70
0 't0 8 3[, &
Leteral Displacemnt (mm)
(d) FG:L3.0-W I 5-P2-CP-CA-ED
(c) FG:L3.0-W I 5-P2-CP-CA

Drift(%)
.'t.5
E
E
sBo
o
o
6
o
dt
:F
-1.0
.0.5
0.0
-rz
t.0
ffiP
t.5
'f
"f
"I
4l.,L
0
.{0 -30 .20 -10 0 10 m
30
Lateral Dsplaoement (mm)
(e) FG:L3.0-W15-P2-HB
Fig.6.7-Relative sliding between wall and
base
6.4.3 Sliding
Fig.6.7 presents the recorded sliding displacement for the five wall tests; again it is noted that
lateral displacement on the horizontal axis has been corrected for sliding displacement. It
appears from Figs. 6.7(a) and (b) that significant sliding occurred
-r25-
for PI-CP and P2-CP
throughout the tests, and that sliding increased gradually with lateral displacement. Both tests
were conducted in one direction only at large displacement levels which naturally did tend to
increase the total sliding displacement due to the absence of excursions in the opposite loading
direction that could reverse the sliding that occurred. Sliding displacements were negligible for
P2-CP-CA for all cyclic excursions, see Fig. 6.7(c). Once the wall had failed in the pull direc-
tion (CP200) and the testing in the pull direction was terminated at -Dl0 (-30 mm), the wall
slid further and further in the push direction. This was partially due to reduced efficiency of the
dowel bars as the compression zone integrity deteriorated. There was no indication of significant wall sliding relative to the foundation for P2-CP-CA-ED as indicated in Fig. 6.7(d), attest-
ing to the efficiency of the dowel and energy dissipator bars to block sliding. Fig. 6.7(e) shows
that negligible sliding occurred for P2-HB until after
Dl6
(40 mm), where push testing was ter-
minated. Still sliding displacement was relatively small.

Gradual deterioration of the foundation beam roughness had occurred as the testing series progressed (foundation beam also used for Series
I testing), thus reducing the shear friction along
the wall/base construction joint.
6.4.4 Prestressing force

The prestressing tendon force histories (sum of force of all tendons) for all walls are plotted
against the lateral displacement in Fig. 6.8. For reference,
it is noted that the nominal yield
strength for one 23 mm VSL CT stress bar is approximately 385 kN.
Walls PI-CP, P}-CP and P2-HB had the tendons rigidly anchored in the foundation and on top
of the loading beam, and did exhibit tendon yielding during testing, as indicated in Figs. 6.8(a),
(b) and (e). It is seen that the first significant total tendon force loss (more than 5% of the initial
force) for these walls occurred after lateral wall displacements to about *10 mm, in all cases
roughly corresponding to first yield of the exheme tendons when considering the wall as
rigid body rocking about the wall corners. At the conclusion of testing of walls PI-CR P2-CP
and P2-HB, between 20%o and 50% of the
It is noted in Fig.
initial total prestressing force remained.
6.8(e) that the total preshess loss never approached the nominal yield
strength of the two tendons of approximately 770 kN despite recording prestress loss. This was
due to uneven force in the tendons due to flexural action and inaccuracy in the tendon force
measurement.
-126-
Drift (%)
zJ
0,
o
lto
o
'6
o
o
4E
([
*[
.'8r
35o
-1.5
500
r-
,*f
l-
L
0
0
tmL
100
05101520253035
-50
40
-30 -20 -10
Drift (%)
Drift (oi)
.1.0
1.s
675
b8(n
,Ee.u
gq
f;
E
o
6
F
s75
oso
ldialPBlrs
0.0
600
Iru
589 kN
Eo*
F
s25
500
-0.5
3ut
o
3*u
(b) FG:L3.0-w15-P2-CP
(a) FG:L3.0-W15-P1-CP
-1-0 4.5 0.0 0.5 1.0
0 l0
Latral Disphcment (mm)
Lateral Osplscement (mm)
L
o
1.0
*f
*f
-l-
650
0.5
'*f
*F
i.
;F
150
.1.0 -0.5 0.0
-30 -20
-10 0 10 20 30 40 50 60
525
70
40 -30 -20 .10 0 10 20 30 &

Lateral Displacernent (mm)
(d) FG:L3.0-w I 5-P2-CP-CA-ED
(c) FG:L3.0-W 1 5-P2-CP-CA

Drift(%)
-1.0
{.5
0.0
zJ '*|1
g
k
5o0
o
o
q
o
a
'*F
*[
o-
o
o
F
*f
'*r
0L
-50
-50
40
-30 -20 -10
0 10 m
30
(e) FG:L3.0-Wl5-P2-HB
Fig. 6.8-Prestressing force histories

Figs. 6.8(c) and (d) show the prestressing tendon force history for P2-CP-CA and P2-CP-CA-
ED. These figures reveal that the prestressing force at the excursion displacement
peaks
amounted to approximately 610 kN and 640 kN for P2-CP-CA and P2-CP-CA-ED, respec-
tively, or about 5% higher than the initial prestress force value.
It was concluded that the
hydraulic system controlled the prestressing force with satisfactory accuracy.
-127 -

Vertical strain recorded along the bottom of the walls during testing is shown in Figs. 6.9
through 6.13. Each figure shows strain plots for instrument levels
and 2 above the base, i.e.
average strain for 0-100 mm and 100-300 mm above the foundation for Pl-CP and P2-HB, and
average strain for 0-200 mm and 200-400 mm above the foundation for the other walls. For
walls with flexural instrumentation on both sides, the average strain of both sides was used.
The sign convention in these figures defines compression strain as negative. The vertical
dashed lines in the figures indicate the extremities of walls, i.e. the extreme masonry fibres at
*1.5 m from the wall centre line. Figures (a) and (c) relate to low displacement response, all
excursions up to and including displacement level D2, and (b) and (d) relate to response
beyond D2.
Regarding vertical strain measurements for PI-CP and P2-HB,
it is suggested in Fig. 6A@)
that the instruments F6, F16, F20,F24 and Fl2 were spanning between the foundation beam
and the middle
of the first masonry course. Likewise, instruments F5, Fl5, Fl9, F23 and
Fll
and
spanned between the middle of first and second courses. This implies that both level
level2 vertical strain measurements for these walls included a component of vertical deformation of the first course, the course where most inelastic action occurred. Strain measurements
for all other walls were recorded between bed joints of one course for HB and CP200 configurations and between bed joints of two courses for CP100.
No plots of vertical stain above level 2 have been included in this report because the recorded
vertical strain indicated elastic masoffy response at these locations (vertical strain significantly
lower than 0.001). This attests that the vertical extent of the compression zone undergoing
plastic deformation remained rather short, of the order of 200 mm to 400 mm. Furthermore, the
resolution of the utilised instruments of perhaps 0.0005 strain excluded meaningful interpreta-
tion of data recorded above level 3. It is acknowledged that vertical strain readings at level I
were influenced by base sliding to some degree. It is remarked that the data plotted in Figs- 6.9
through 6.13 are associated with some degree of uncertainty, especially for displacement levels
above D2 where significant deterioration of the masoffy units may have affected instrument
reading.
r28 -
0.020
Lovel
(0.100 mm
abo{t
barl
0.018
|
|
|
|
|
0.0 16
0.0r4
+-0.5v6
(0.15 mm)
+vc(0.ram)
-r- 1,33/6 {0.77 mm)
-r-0.75\/|(1.2
+O2
m)
(2.9 mm)
0.012
0.0r0
p
t
o
E
o.ooo
0.006
'I
to
o.004
0.002
------------]
0.000
-0.002
{.001
.0.00
-1.6 -1.2 4.8 -0.4 0.0

Lodtld along ull
0.4
0.8
1.2
t.6
0.0 0..1 0.8

rlll l.ogtn (mm)
(b) Above D2,lettel0-100 mm
4.4 0.0 0.4 0.8
Lootldr elong wrll l.ngth
{.4
Locrtlon dong
(a) V",-D2,level 0-100 mm
-1.6 -1.2 {.0
-1.0 -1.2 {.8
length (mmt
-1.8
1.2
-1.r
{.6
{.1
0.4
0.9
Loc.tlon elone wrll l.nglh (mm)
(mm)
(d) Above D2, level 100-300 mm
(c) Vo-D2,level 100-300 mm
Fig. 6.9-FG:L3.0-WI S-PI-CB Ilorizontal strain profiles, pull direction only
-r29-
0.020
0.018
0.016
0,ot4
o.o12
0.000
0.010
E 0.008
E
o
o.oro
6
d
0.006
.0.020
0.004
-0.030
0.002
0.000
{.040
-0.002
-o.050
.0.004
-0.006
-t
.0.060
.6
-1.2 -0.8 -0.4 0.0
0.4
0.8
1,2
1,6
-1.6 -1.2 {.8
Locatlon alon9 wall llngth (mm)
-0.4 0.0
0.4
0.8
1.2
Locltlor rlong wall lcngth (mml
(b) Above D2, level0-200 mm
(a) V";D2, level 0-200 mm

0.0003
0.0002
0.0001
0.0000
6
o
5o
-0.0001
-0.0002
-1.6 -1.2 -0-8 -0.4 0.0 0.4
0.8
-1
Locetlon .long wllt langlh (mml
.6
-1.2 .0.8 {.4
0.0
0.4
0.E
Locallon along wall l.ngth (mml
(d) Above D2, level 200-400 mm
(c) V",-D2,level 200400 mm
Fig. 6.10-FG:L3.0-Wf 5-P2-CB Horizontal strain profiles
-130-
1.2
1.6
0.020
0.016
0.016
0.014
0.012
0.010
"E
o.ooa
TA
?o
E
o
0.006
0.004
0.002
0.0@
-0.002
-0.@4
-0.006
-1.6 -1-2 -0.8 {.4
0.0
0.4
-r.6 -r.2 -0.8 .o.4 0.0 0.4 0.8 1.2
0,8
Locllon.lone w.ll l.ngilh
Lodllm along vrall l.nglh (mml
1.6
(mm)
(b) Above D2, lwel0-200 mm
(a) V",-D2, level 0-200 mm
0.0004
0.0002
0.0000
4.0002
#..*
o
o
E
o
> {.0006
{.0008
{.0010
-0.0012
.0.0014
-t.6 -1.2 -0.8 {.4
0.0 0.4 0.8
-1.6 -1.2 {.8
1.2
6.1
0.0
0.4
0.8
(d) Above D2, level200-400 mm
(c) V",-D2, level 200400 mm
Fig.
4.4
Loc.tlon alofiO wlll lcnglh (mml
Locallon elong w.ll length (mml
1-FG:L3.0-W15-P2-CP-CA, Horizontal strain profiles
-131-
1.2
1.6
0.020
0.050
0.018
0.040
0.0r6
0.030
0.014
0.020
0.012
0.010
cG
o
0.010
0.008
ie
0.006
o.ool
.E
t
o
0.000
to
E o.oro
{.020
0.002
0.000
4.030
'0.002
{.0,10
{.tx}4
{.050
{.006
-0.008
-1,6 -1.2 4.8 -0.,0 0.0 0.4
4-00
0.8
-1.6 -1.2 {.8
Locrtlon along wrll longth (mm)
4.4 0.0 0.4
0.8
Locatlon along wall longth (mml
(b) Above D2, level0-200 mm
(a) V..-D2, level0-200 mm

0.0002
L.vel
(awrage200{10
0.0001
0.0000
LoYel2 (we.rg 200.||!0 m]|t)
ml
r"ll:::;:==='-E
4.0001
\s
F.
E
0
ie-
-'r'--------'--:'.
----
't
'B-E
to
o
-0.0004
---r
$ ".*0,
o
'a
,--\:-..\
A-
'i.
;o 4.0O03
tr/
-*r
':.
tr:
Fi
---
L\
4.0002
5i
{.000,1
\p
''b
dE
\-o
rlI
--.r--O2(-8.5m)
---il--F
G3 mm)
---r(--Vq (4.84 mm
- {-.Vs (1,6 mm)
- -.Vf (6-7 mhl
--a--\4(-7 m)
--{--Vo(-1.8m)
-*-.Vq (0,57 m)
-+-.Fl (3.1 mm)
-F.O2(8.3 m)
.0.0006
il
l.l
ll
!l
--.--O8 t35 mm) --.--DO C23 mm)

-a--D4 (-17 mm) -{--D4 (14 nm)
{.0008
i\rO
-{-.D6
(23 mm)
-F.Dlo
-e-.DO
(30 mm)
(40 mm)
-0.m10
-1.6 -1.2 -0.8 -0.4 0.0
0.4
0.8
1.2
1.6
L'oatlqr slong wrll lcngth (mm)
6.
-0.4 0.0 0.4 0.8
Locatlm along wsll lcngth (mm)
(d) Above D2, level200-400 mm
(c) V",-D2, level 200400 mm
Fig.
-r.6 -1.2 {.8
l2-FG:L3.0-Wl 5-P2-CP-CA-ED' Ilorizontal strain profiles
-r32-
1.2
0-020
0.018
0.0r6
0.014
0.012
cG
ta
-E
0.010
0.008
o.oou
0.004
0.002
0.000
-0.002
{.fix
J.6
-1.2 {.8
{.4
0.0
0.4
-1.6 -1.2 -O.E {.4
0.8
0.0
0.4
0.8
1.2
Loc.tlon tlong w.ll hngth (mtnl
Locrtlm .long wlll lsngth (mmt
(b) Above D2,level0-100 mm
(a) V",-D2, level 0-100 mm
0.003
g
I
0!
E
0.0O2
o.oor
0.000
-0.001
{.002
-1.6 -1.2 {.8
-0.4 0.0 0.4 0.8
Loc.llon dong
s.ll
1.2
-1.6 -1.2 {.8
-0.4 0.0 0.4 0.8

r.ll longfh (mm)
Locltlm slong
l.ngih (mml
(d) Above D2,level 100-300 mm
(c) V.,-D2,level 100-300 mm
Fig. 6.13-FG:L3.0-W15-P2-HB, Ilorizontal strain profiles
-133-
1.?
1.6
Figs. 6.9(a)+(b) through 6.12(a)+(b) reveal that vertical sfain varied nearly linearly along the
length of the walls (except P2-HB) for all displacement levels. Figs. 6.13(a)+(b) show that this
was not the case for P2-HB for loading above *V6 presumably because of the stiffness incom-
patibility between high strengh and ordinary masoffy units.

The general trend for CP200 and CP100 masonry for low level response (Figs. (a)) were
recorded extreme fibre strain of approximately 0.003 for readings below D2 and approximately
0.004-0.006 for D2. The extreme fibre sfrain for large displacement response ranged from 0.01
to 0.08. Strains at level 2 remained below 0.0012 for P2-CP, P2-CP-CA, and P2-CP-CA-ED,
thus indicating elastically responding masonry. Measurements at level 2 for Pl-CP and P2-HB
suggest strains higher than 0.0015, however, as noted above, these measurements were taken
by instrumentation that also picked up deformation in the first masonry course.

Table 6.4 presents a sunmary of vertical strain measurements related to specific events occur-
ring during testing. Vertical strains in the flexural compression zones (the toe) of the walls
were evaluated from data acquired during testing for the following wall states:
(l) at nominal
flexural strength, Vn (2) at maximum strength, Vn
and (3) at ultimate displacement, du.

"",
Refer to Table 6.3 for Vp V*"* and dr. The measured stains in the extreme compression fibre
at the wall states described above are listed in Table 6.4 for the push and pull directions, as
available. The strains were measured in the lowest 100 to 200 mm of the walls. Results for P2-
HB are not included.
TABLE 6.LMeasured masonrv vertical strain

Wall
vf
Dir./Conf.
disol.
FG:L3.0-Wl5-Pl-CP
Pull/CP200
-1.7
2. FG:L3.0-Wl5-P2-CP
Push/CPl00
PulliCP200
Push/CPl00
Pull/CP200
Push/CPl00
Pull/CP200
3. FG:L3.0-Wl 5-P2-CP-CA
4. FG:L3.0-Wl5-P2-CP-CA-ED
Note: +extrapolation based on instruments Fl6 and F20
-r34-
1.7
-2.0
4.0
-4.0
6.7
7.0
nrm
strain
0.003s
0.0016
0.0019
0.00r6
0.0029
0.0027
0.0055
Vr-
d"
strain
strain
0.070*
0.0107
0.0230
0.0190
0.0037
0.0052
0.0053
0.01l6
0.036
0.055
0.068
0.024
0.033
0.050
Drift (%)
-2.0
-1.5
500
-0.5 0.0
0.5
1.0
1.5
2.0
2.5
400
300
z.Y
G
200
100
-c,
o
(t,
o
-100
U)
-200
-.]--.+
-300
FG:L3.0-W$P2-CP
FG:L3.o.nrt $ P2-CP-CA
FG:L3.Gu|I+P2-HB
400
-500
-50
-40
-30 -20 -10
o 10 20 30 40 50 60
70
Fig.6.14-F-D envelopes, P2-CP, P2-CP-CA and P2-HB
6.5 DISCUSSION
6.5.1 Flexural response
Fig. 6.5 indicates that the developed flexural strength Vn'", (base shear corresponding to the
maximum moment) exceeded the predicted flexural strength Vs in all wall test. Comparing the
force-displacement envelopes (pea,k response of first excursion cycle for each displacement
level) for walls p2-cp,p2-cp-cA and P2-HB, which had similar dimensions and prestressing
properries, shown in Fig. 6.14, it is seen that P2-CP and P2-}IB developed higher strength than
p2-Cp-CA. This directly resulted from the tendon force in P2-CP-CA being kept constant during testing, prohibiting tendon force increase due to flexural action. The performance of P2-CP
in
and p2-CP-CA in the push direction (positive displacement), both featuring CPt00 masonry
the compression zone, differed. It appears from Fig. 6.14 that P2-CP-CA sustained far larger
displacement in this direction of loading. This was expected as a result of the lower axial load
in p2-Cp-CA. It is difficult to compare the pull response of P2-CP and P2-CP-CA because P2Cp test results in that direction were significantly influenced by the testing regime which consisted in pull direction testing upon completion of push direction testing.
-135-
It is of considerable
Drift (%)
-2.0
500
-1.0
-0.5
300
100
o
U'
-100
1.0
2.0
1.5
2.5
{R
200
o
o
0.5
[l-l--rr
400
z-Y
0.0
a-c,
ct
(D
-200
--+-
--.-
-300
FG:L3.0-Wl +P2-CP-CA
FG:L3.GWI $P2-GP-CA.ED
400
-500
-50 -40 -30 -20
-10 0 10 20 30 40
50
70
Fig. 6.11-F-D envelopes, P2-CP-CA and P2-CP-CA-ED

interest that P2-FIB perfonned similar to P2-CP in the push direction and possibly better than
P2-CP in the pull direction. The suggestion is that using a single high strength block in each
wall comer performs as well as the confining plate solution. Embedding high strength blocks
in the lower corners appears to be a less laborious solution compared with embedment of confining plates.
Comparing the responses for P2-CP-CA and P2-CP-CA-ED in Fig. 6.15, it is clear that the latter developed the highest strength. This was due to the energy dissipators built into P2-CP-CA-
ED that had a nominal yield strength of 2 x 100 kN

prestressing force
:200 kN. Adding this force to the applied
of approximately 600 kN suggests that the P2-CP-CA-ED wall strength
would be approximately (600 kN + 200 kNy600 kN = 1.33 times higher than that of P2-CPCA. This is confrmed by Fig. 6.15 and Table 6.3. The displacement capacity of P2-CP-CA-ED
was lower than that of P2-CP-CA in the push direction (CPl00 masonry), a direct effect of the
higher stress imposed on the compression zone by the addition of energy dissipation bars. It
appears that the displacement capacity of P2-CP-CA-ED in the pull direction inexplicably was
superior to that of P2-CP-CA. Comparison of the perfiormance of CP200 vs. CPl00 for these
two tests favours CPl00 in terms of displacement capacity.
136 -
6.5.2 Tendon yielding

Tendon yielding was expected for all walls without axial load control. Significant prestress loss
occurred for these walls, as is illustrated in Fig. 6.8. At the peaks of excursions Pl-CP and P2-
Cp generally developed a total tendon force higher that the initial tendon force (Figs. 6.8(a)
and (b)). P2-HB exhibited some tendon force increase at peak response. Howevero as wall toe
degradation initiated, the total tendon force at displacement peaks started to decline to levels
below the initial total tendon force. Despite tendon force loss due to yielding, the walls continued to retum to their initial alignment (disregarding sliding), as suggested in Fig. 6.5'
No tendon yielding occurred for P2-CP-CA and P2-CP-CA-ED as a result of active control of
the tendon force. Modest tendon force variation was recorded as shown in Figs. 6.8(c) and (d)
as a result
of friction in the hydraulic system.

Table 6.4 shows the vertical strain recorded in the plate confined walls at different stages of the
tests. It is seen that the calculated mason-ry strains corresponding to the nominal flexural condi-
tion, Vs, ranged from 0.0019 to 0.0055 for CP200 and 0.0016 to 0.0027 for CPl00. These
results do not conform with the code defined nominal flexural strength condition set out by
NZS 4230 for confined masonry, that specifies a useable extreme fibre masonry strain at nominal flexural strength of 0.008.
The masonry strain recorded at V*u* ranged from 0.0052 to 0.0190 for CP200 and 0.0037 to
0.023 for CP100. Thus a large variation in vertical strain at this limit state. It was felt at this
limit
state that the masonry fibres
in the exheme ends of the walls still provided some axial
stength, resisting the overtuming moment'

Table 6.4 indicates that masonry extreme fibre stains ranged from 0.024 to 0.070 for CP200
and from 0.033 to 0.068 for CP100 at the ultimate displacement, 4.,.These values are of course
of theoretical character as the toe regions were damaged at this stage, so that those strains
should not be regarded as the reliable ultimate strain capacity of concrete masonry. The results
do nevertheless suggest that masonry strains far beyond those related to nominal flexural
strength can be expected for rocking wall systems, while still providing significant axial
stength. At this limit state the effective flexural neufial axis at the wall base had migrated
towards the middle of the wall. Clear evidence of this is found in Figs. 6.9(b) through 6-13(b)'
t37 -
Drift (%)
? ooI
P
d)
u,l
o
(D
*f
*r
'r
15
0L
llJ
*r
'u l-
0)
o
T
t--
E
f
f uof
I0)
*f
0.0
or
"fur
-o
{.5
-1.0
,or
tuf
E'
o ro
c,
lE ur
E
J
-40 -30 -m -10 0
10 m
gL
-40 .30 -20 -10 0

-40
30
r0
30
Lateral Displacemenl (mm)
(b) P2-CP-CA-ED
(a) P2-CP-CA
Fig. 6.lG-Cumulated hysteretic enerry dissipation
showing that the crossing of the strain profiles with the zero strain axis moved closer to the
wall centre as the walls were displaced to larger and larger displacements.
6.5.4 Hysteretic enerry dissipation

Comparison of Figs. 6.5(c) and (d) illustrates the effect of the energy dissipation bars in P2CP-CA-ED. The area of all displacement cycles for P2-CP-CA-ED were increased considera-
bly by yielding of the energy dissipation bars, as significant yielding commenced going to displacement level
4 (17 mm). The energy
dissipated
in
each excursion (half cycle) due to
hysteresis can be quantified by integration of the area enclosed by the loading and unloading
curves. Results of such integration applied on the P2-CP-CA and P2-CP-CA-ED are illustrated
in Fig. 6.16. The figure shows the cumulated energy dissipation as a function of the wall latoral
displacement history. Fig. 6.16(a) illustrates that the energy applied to displace P2-CP-CA
from zero displacement (original alignment) to the peak of an excursion, nearly was recovered
upon subsequent unloading. On the contrary, the P2-CA-ED response shown in Fig. 6.16(b)
suggests that
only about 30-50% of the energy exerted by the loading system displacing the
wall from zero displacement to a displacement peak was recovered upon subsequent unloading. As a numerical example, the total energy dissipated in the last two cycles shown in Figs.
6.16(a) and (b) has been calculated. The circular and square markers illustrate the beginning
and end of the two displacement cycles to
t30-35 mm, respectively. Calculation shows that 4.8
kJ was dissipated by P2-CP-CA while 15.5 kJ was dissipated by P2-CP-CA-ED. Clearly, the
hysteretic energy dissipation was increased substantially by the energy dissipation bars. Little
sliding was measured for the walls depicted in Fig. 6.16 in the displacement range shown, thus
this was not a source of 'parasitic' energy dissipation.
- 138-
6.5.5 SlidingpropensitY
Wall sliding was measured for all walls, except P2-CP-CA-ED. This phenomenon was sought
eliminated by the addition of sliding blocking plates placed adjacent to each end of the walls
and by embedment of dowel bars. Testing results of PI-CP and P2-CP with sliding blocking
plates mounted on the foundation showed that solution inadequate, because, once distress
of
the compression zone had initiated, little lateral force could be resisted by the compression
zone as the wall slid towards the sliding blocking plates. Sliding for P2-HB did not initiate
before integrity of the high strength blocks was lost. This occurred after approximately 40 mm
of displacement. Again, the sliding blocking devices did not offer much restraint.
Walls p2-Cp-CA and P2-CP-CA-ED had dowel bars embedded in both compression zones
of
the wall as shown in Fig. 6.1. These bars offered sufficient sliding restraint for P2-CP-CA in
the pull direction (CP200), however when the CP200 end had failed and monotonic testing in
the push direction only was instated for the remainder of the test, sliding initiated gradually as
the CPl00 compression zone integrity gradually degraded.
It
appears that the dowel bars
worked satisfactorily in the push direction until displacements of about 45 mm (1.6% drift)
was reached.
It is concluded that
adequate shear friction between wall and foundation requires intentional
roughening of the wall to foundation interface (refer to Chapter 8 for an in-depth discussion)Unbonded dowel bars, possibly in conjunction with longer continuous bars (in this case energy
dissipation bars), appeared a viable option for adding further sliding resistance to the shear
friction provided by the masonry/concrete foundation interface.

Estimation of initial stiffness is of considerable interest for serviceability loading. Fig. 6.17
shows curves for all P2 walls from Series 2 testing and for two walls from Series
I testing, P2C
and p2E (refer section below and to Chapter 5), relating to the wall secant stiffiress at the theo-
retical first cracking limit state (refer to Chapter 8). It is seen in the figure that the initial
stiF
ness varies littte bet'ween the walls. This is not surprising because the wall initial stiffiress
theoretically only depends on the wall dimensions and the masonry elastic properties, and
these were nearly identical for all tests. The average initial stiftress for the walls shown in Fig.
6.17 was 238 kN/mm compared to a theoretical elastic stiffiress of 354 kN/mm (based on Eqns'
139 -
150
125
100
zlz
(E
-c
o
a
o
(D
75
50
25
0
-25
-50
-75
-100
-',25
-150
-0.75
-++
+
+
+
-
FG:L3.0-Wl5-P2.CP
FG:L3.0-W1 5-Pz-CP-CA
FG:L3.O-Wi 5-P2{P.CA-ED
FG:L3.0-Wi5-P2+|B
FG:L3.0-Wi5-P2G (Serles l)
FG:L3.0-Wi5.P2E (Serle8 1)
-0.50 -0.25 0.00 0.25
0.50
0.75

Fig. 6.17-Initial wall stiffness
8.2 and 8.4 using
Er:
15 GPa,
v :0.2 and P*N:641 kN). The measured average wall initial
stiffness was thus 33% lower than the theoretical one.
6.5.7 Comparison with Series I wall tests

The primary goal with this testing series was to investigate several means of strengthening the
compression toe regions in order to improve wall displacement capacity. In this section com-
parison is made with two
1ss1
rrnifs from Series
I testing: FG:L3.0-WI5-P2C
and FG:L3.0-
WI5-P2C. Both walls had two prestressing bars embedded, the P2C bars concentrically at
locations identical to all Series 2 wall tests (*400 mm) and the P2E bars eccenhically at +1200
mmo and both walls featured unconfined ordinary masonry units throughout (U200). P2C
results should be used for direct comparison with P2-CP and P2-HB because of similar dimension and prestressing layout. Fig. 6.18 shows the force-displacement envelopes for P2-CR P2-
HB, P2C and P2E.

Examining the wall responses in the push direction (positive displacement), the following
observations are made: all walls developed similar strength; strength degradation for P2C and
P2E initiated at approximately 0.4% drifr while strength degradation for the other walls initi-
- 140-
Drift (%)
-2.0
-1.5
-1.0
500
-0.5 0.0
-0.8vmax
300
z
G
o
E
@
o
o
o
m
2.0
1.5
1.0
r ,n-\
400
0.5
200
100
0
-100
-200
-300
_
-:-</"Yt--"'/
- ----}t-tt+'-h-;z
400
-s00
r
cPm0
<l-r--?-*-
FG:13.0-lt'l5.HI-CP
FG:Lil.GWl$Pil-HB
FG:L3.&wlsP2c (Sstl.! 1l
F9:L3.0-W1$P1 E (S.dot tl
-0.Evmax
rrtll
-50 40 -30 -20 -10 0 10 20 30 40
50
Fig. 6.f 8-F-D envelopes, P2-CB P2-HB' P2C and P2E

ated at about 0.9%
drift; the ultimate displacement capacity (at strength degradation to 0.8Vn'",
shown as dashed lines) for P2-CP, P2-HB and P2E were similar at approximately
had a displacement capacity of about
l.2yo,Pzc
0.5%.It is therefore concluded that CP100 and HB solu-
tions performed significantly better than U200 for both P2C and P2E in terms of initiation
of
strength degradation, and CP100 and HB performed better in terms of ultimate displacement
capacity in comparison with U200 for P2C.
Examining the wall responses in the pull direction (negative displacement), similar observations are made: all walls developed similar strength (except for P2-CP that was influenced sig-
nificantly by base sliding rendering the results difficult to compare with those from the other
wall tests); rapid strength degradation for P2E initiated at approximately 0.3% drift while
strength degradation of P2-HB and P2C was rather gradual; the ultimate displacement capacities (at strength degradation to 0.8V,.* shown as dashed lines) for P2-llB, P2C and P2E were
approximately l.4o/o,l.lo/o, and 0.5%, respectively. It is therefore concluded that the HB solu-
tion performed significantly better than U200 for both P2C and P2E in terms of initiation
of
strength degradation and ultimate displacement capacity. It cannot be directly concluded from
Fig. 6.18 that CP200 performed better than U200, however, the large displacement capacity in
-l4r-
the pull direction for P2-CP suggest that also CP200 masoffy outperforms U200 masonry with
a significant margin.
6.6 CONCLUSIONS
It is concluded that strengthening of the flexural compression zones of fully grouted unbonded
post-tensioned concrete masonry walls successfully improved the wall displacement capacity
and delayed the onset of stength degradation in comparison with a PCM wall of similar
dimensions made with unconfined masonry. The maximum wall stength remained insensitive
to strengthening of the compression zone.

The PCM walls reported on in this chapter behaved similarly to the behaviour observed for
testing Series
l,
notably they exhibited a nearly non-linear elastic behaviour dominated by
rocking response. The exception was P2-CP-CA-ED that exhibited rocking response with nonlinear elasto-plastic behaviour due to the presence of energy dissipation devices.
Large drift capacities were measured, 0.7o/o to
2o/o
for directions with CP200 comers engaged,
l -Io/o to 2.60/o with CP100 corners engaged and I.IVo
to
l.4o/o
with high strength block corners
engaged. This clearly attests to consistent improvement of wall displacement capacity using
confining plates or high strength corner units. The highest performing solution was CP100.
Only localised damage occurred, as shown in Fig. 6.6, making earthquake damage easy to
repair.
It is seen that nearly all damage occurred to the lowest masonry course for CP200
and
CPl00. This is supported by the vertical strain plots shown in Figs.6.9 to 6.12,that suggest
that little inelastic masonry response was measwed above the lowest 200 mm of the walls. The
high strength masonry units appear to have shifted masonry crushing to the second masonry
course (U200) above the foundation because of the very high strength of the high strength
units.
Use of confining plates is recommended for PCM wall construction based on the general
observations of increase of wall displacement capacity and reduction of wall damage when
compared with the performance of unconfined PCM walls. The use of high strength blocks in
the lower wall corners appears to improve the wall performance. Still brittle masonry failure
occurs in the masonry blocks just above the HB blocks which does not suggest increased wall
toughness. The use of HB corners is not recommended for ductile design if HB blocks only are
used to replace single masonry unit in the wall corners.
-r42-
Relatively little energy dissipation was observed during cycling of all walls, except for P2-CP-
CA-ED. This may impact the use of such walls for ductile seismic design in the sense that
additional damping may be required to control the structural drift demand in an earthquake
event. Therefore one wall had 'dog-bone' energy dissipation dwices embedded. This simple
device proved successful and worked as expected.
A threefold
increase of hysteretic damping
was achieved at large displacement levels. The inherent simplicity and insensitivity to construction tolerance of the 'dog-bone' type device makes it a desirable option for 'real life' construction. Buckling restraint achieved by casing of the reduced diameter section of the 'dogbone' energy dissipation bars with heavy pipe allows for placing of the energy dissipation bars
inside the wall without risk of out-of-plane bursting of the masonry because of bar buckling.
Tendon force loss due to yielding should be expected at relatively low wall drift ratios, for
squat walls. Even after tendon yielding, reliable and self-centring wall behaviour is expected.
The hydraulic system used for control of the prestessing force for P2-CP-CA and P2-CP-CA-
ED performed successfu lly.

Measurements at nominal flexural strength suggest that an exfeme masonry fibre strain in the
order of 0.0035 was found for CP200 and 0.0020 was found for CPl00. Both values are substantially lower than 0.008 as stipulated by NZS 4230:1990 for confined masonry. At ultimate
displacement, sffains as high as 0.046-0.050 were measured for both CP200 and CPl00. At
this stage the extreme masonry fibres no longer carried axial load as face shell spalling had
occurred.
It was concluded that adequate shear friction between wall and foundation requires intentional
roughening of the wall to foundation interface. Dowel bars are effective for adding firther slid-
ing resistance.
Shear reinforcing should be embedded
in PCM walls to reassure integnty of the masonry
panel, though not necessarily needed for strength.
-r43-
6.7
REX'ERENCES
[6-U
P,ark,
k,
Evaluation of Ductility af Stru,eturcs and Stnrefrrd As;emblages for Labora-
tory Testing,BulletinofNZX.ISEE" Yol.22,No.3n

[6-2] NZS 4230:1990" Cade o{ Pracfice
for
Sqt
1989, pp 155-165.
the Detign qf Masonry Stmrctitws, Standards
Assoeiation of Neii' Zealaad, Wellinglon.
[6.3] Paulay; T., and hiestley, I\{.JN.,
Selrrlnlc Design
Buildings,John Wiley & Sons Inc.,1992,744p.
-t44-
af Reinforced Cona'zte and Mosonry
Chapter 7
STRUCTURAL TESTING - SERIES
7. 3-STOREY PCM WALLS

Two 3-storey high unbonded post-tensioned concrete masonry (PCM) cantilever walls were
tested in the
Civil Test Hall at the University of Auckland. The 67%o scale wall units were
designed to model a typical cantilever wall from a 4-5 storey high office or apartment building.
A detailed account of the wall construction, test setup, testing procedure and test results is provided in this chapter.
The principal intent with these wall tests was to validate the use of PCM in a realistic structural
configuration. The test units, incorporating RC slabs at the intermediate floor levels, were subjected to a realistic moment gradient. Furthermore, the tests explored means of masonry confinement or strengthening that are expected to allow for reliable drift capacities beyond' l%.
7.I
INTRODUCTION
Previous testing by Laursen and Ingham, reported in Chapters 5 and 6, explored PCM walls
made with unconfined and confined masonry.
It was found that unconfined (ordinary) walls
typically provided displacement (drift) capacities of up to lVo, however the limited strain
capacity of unconfined masonry resulted in masonry crushing and subsequent sfrength degradation, and prohibited development of further reliable drift capacity. Using confined masonry,
it was found that further displacement capacity could readily be developed, typically of the
order
of lYo to l.5o/o, depending on the confinement
demands
of
method. As mentioned before, drift
such magnitude are likely to be encountered
in
unbonded post-tensioned walls when subjected to seismic loading.

The objectives of this testing series were to:
Evaluate
1. Strength and displacement capacity of the walls
- 145-
structural systems utilising
2. Effectiveness of using confinement plates in the bed joints to enhance the concrete masoffy
strain capacity
3. Energy dissipation during
cyclic testing
Determine
l.
Ultimate useable masoffy strain
2. Vertical extent of
the plastic deformation zone in the wall corners
Compare
I
Chapter 7 test results with testing of RCM walls of similar proportions

The walls were designed to sustain large in-plane displacement arising from severe ground
shaking. This was ensured by keeping the three prestressing strands (Fig. 7.1) unbonded over
their entire length for maximum elastic elongation capacity and by enhancing the wall strength
and strain capacity in the critical lower corners.
Wall dimensions are specified in Table 7.1. The wall designations do not follow the convention
presented in the previous t'wo chapters; the two walls are simply termed 53-1 and S3-2. Fig. 7.1
shows the wall dimensions and prestressing layout. The effective wall height, h., of 5,250 mm
was determined by the maximum clear height in the Civil Test Hall. A storey height
of
1,950
nrm was chosen for the two lower levels, corresponding to approximately 670/o scale model
a
of
typical 3,000 mm storey height. Assuming that the effective wall height (equivalent height of
the total seismic mass) amounts to about 213 of the total height, the test units correspond to
67Yo scale models of a 4 storey
TABLE
7.l-Wall
building of l2 m height. The wall length of 2,400 mm was cho-
specifications
thick wall self external building masonry number
of
weight weight weight strength
f,nu
NN.*
N
b*
strands
height length ness
Wall
l*
s3-l 5,250 2,400
t40
s3-2
5,250 2,400
mm
mm
64
356
420
140
64
142
206
rnm
KN
KN
KN
*initial. "unconfined strength, halfheight units, -f-: (P+N)/(l*b*)
-t46-
initial
axial loati
prestress
ratio
Plf""
948
0.136
427
l0l7
0.1 39
KN
MPa
17.9
398
3.6
MPa
f./ft. -
1"",
opplicofion
o{
lolerol lood
2x
HD1O
ironsverse
relnforcemenl
3 Unbonded
'Preslressing
slronds
15 Serles
grouied
concrefe
mosonry
Holf height
ocks
Wosh-oul
Fig.7.l-Wall
geometry for Series 3 walls
sen to reflect a realistic
wall aspect ratio (heighVlength) of approximately 3.4. Using 15 series
concrete masonry blocks with nominal thickness
67yo scale models
of 140 mm, the wall
test units represented
of ZOg mm thick walls. This appears to be a realistic wall thickness that
could be achieved with readîly available 20 or 25 series concrete masonry blocks. The wall
dimensions and layout could also represent full scale walls of 5.25 m heiglrt using 140 mm
masonry.
Ordinary 15 series concrete masoffy blocks, Trade MortaflM and 17.5 MPa grout were used
for construction. Confining plates were embedded in all bed joints in the lower 1,000 mm of
-t47-
Bed joint conf inmenl ploles

25mm x 3.2mm flol mild sleel
Gr. 250, full penelrolion welded
Fig. 7.2-Dimensions and locafion of confining plates

the walls and extended approximately 600-800 mm from the wall ends. Fig. 7.2 shows the
dimensions and locations of the embedded confining plates, referred to as CPl00 masonry
when embedded in bed joints between half height masonry units. The plates were welded from
25 mm by 3 mm mild steel flat and proportioned such that grout continuity was not impeded.
Research presented in Chapter 3 and Chapter 6 has shown that such a system confines the
grouted concrete masonry, increasing the axial strength and strain capacity.
Both walls were prestressed with three 15.2 mm high strength strands placed at +400 mm and
0 mm from the wall centre line. The prestressing tendons remained unbonded over the entire
wall height and had a typical unbonded lengh of 6,400 mm, including unbonded lengths in the
loading beam (300 mm), foundation (300 mm), strong floor (400 mm) and load cells underneath the strong floor (300 mm).
The initial prestressing force for both walls was determined based on a required wall drift
2Tobefore tendon yield assuming that the wall simply rocks around the lower corners. 2%
of
dift
is a typical code defined maximum and corresponds to a wall displacement of 105 mm at the
equivalent height of
h":
5.25 m. This resulted in a target initial prestressing steel force of 400
kN or approximately 0.55f0,. The achieved wall initial prestressing force and axial force ratios
(due to prestressing and building weight) are given in Table 7.1-
-148-
Horizontal shear reinforcing was provided at 400 mm vertical spacing. Each level consisted
of
two HDl0 deformed bars (10 mm deformed bars,430 MPa yield shength) placed one on each
side of the prestressing ducts.
For S3-1, self-weight and tributary loads from supported floors was stipulated to be 1.0 MPa
masoffy stress or 336 kN at the wall base, corresponding to a total building weight of 400 kI.I.
As the wall self-weight of the tested assembly (masonry and RC concrete) was approximately
64 ld{, an additional externally applied axial force of 336
k}.,1
was required (refer to external
axial load system discussed in section 7.3.1). Lighter axial loading was assumed for wall S3-2
such that the total building weight amounted to 200 kN (0.5 MPa masonry stress), requiring
136
kN of additional axial force to be applied. This corresponded to the wall strength (restor-
ing moment) due to prestressing (P) and building weight (N) being provided on a 50Yo:50o/o
basis for
wall S3-l and on
a 67o/o:33Yo basis
for S3-2.

The walls were constructed in running bond by experienced blocklayers under supervision,
using standard grey precast concrete masonry units and DRICONTM trade mortar. Open-ended
concrete masonry units were used to avoid having to thread the masonry units onto the ovalised
prestressing ducting, see Fig. 7.3(a); note the HDl0 (deformed l0 mm bar, &
:430 MPa) shear
reinforcing bars placed on each side of the oval duct. Confining plates, as shown in Figs. 7.2
and 7.3(b), were embedded in the zone of half height masonry units with mortar on both sides
for best integrity. Grouting with 17.5 MPa target strength ready-mixed regular block fill, containing SIKA CAVEXTM expansion agent, was performed the subsequent day. Vibration of the
grout was carried out for all grout pours. Wash-out ports were provided on the north face in the
lowest halfheight block in each corner.
Construction of wall S3-1 proceeded in four phases:

1. The first storey was constructed on the wall foundation situated in the testing rig. After
grouting and curing of the first storey, a RC floor slab was placed on top of this storey on a
mortar pack.
2. The second storey was next built in the test setup on top of the first storey RC slab. Meanwhile, the third storey was constructed on the second storey RC floor slab temporarily
placed on the laboratory floor.
-r49-
Fig.7.fia)
Prestressing ducting and shear reinforcing and (b) confining plate
3. After grouting and curing of the second and third storeys, the third storey assembly was
placed on top of the first and second storeys in the test setup. Vertical alignment of the three
stories was simply ensured using a plumb bulb.
4. Placing of the reinforced concrete loading beam on a mortar pack, and installation and
stressing of the prestressing strands.
Construction of wall S3-2 proceeded differently. Only the bottom part of the first storey wall
of
S3-l had been damaged during testing, therefore, only the bottom storey wall was replaced for
S3-2 in the following steps:
l.
2od and 3'd storey
walls, including RC slabs and loading beam, were lifted offthe first storey
wall using chain blocks attached to the testing rig.
2. The damaged first storey wall

3. The first
was removed.
storey was reconstnrcted on the wall foundation and grouted.
4. Upon curing, the
second and third storey assembly was repositioned on top of the
first sto-
rey wall on a mortarpack.
The reinforced concrete foundationo floor slabs and loading bearn were manufactued of 30
MPa target strength concrete, prior to masonry constnrction. Fig. 7.1 shows the concrete
dimensions and Fig. 7.4 shows the reinforcing layout.
-150-
Rl6 SllrupE Sl O
100
Plon Vlcw
Foundolion reinforcement
Plon Vlew
Loqding beom reinf orcemeni
oo
!lD'12
o
N
O l4O opprox.
o
r
Plon View
tl
@l
-*Flg
40O LoD
Elevolion
Floor slob reinforcement

Fig. 7.4-Reinforcing layout for RC elements

Average masonry material properties were determined by material testing, typically in samples
of three. The masonry crushing strength

400 mm
f'-,
masonry elastic modulus E,o (masonry: 600 mm x
140 mm prism), grout crushing strength
mortar: 200 mm high
"
f'r, mortar crushing
100 mm diam. cylinder), and average age at day
strength
2nd
(gout
and
of testing are given in
Table 7.2. Average strengths of concrete were 38 MPa for foundation and
MPa for
fi
l't floor
slab and 49
floor slab and loading beam.
Seven-wire prestressing strand of the specification 'AS 13ll-1987, Super, Low Relaxation'
was used for both wall tests. Nominal properties of the 15.2 mm diam. prestressing stand were
-151-
tensile sffength (rupture strength) of 250 klrf
(fpr:
1785 MPa), yield strength of 213
1520 MPa), minimum ultimate elongation capacity of 3.5Yo, elastic modulus
Eo.:
kl.l
(f;r:
190 GPa,
maximum relaxation after 1000 hours of 2.5% and tendon area of 140 mm2. Proof testing of a
ASl3l l-1987 [7-l] confirmed the above properties with

the following results: fr, = 252-257 kN, fp"-- 271-274 ld{, elongation of 7.0-7 .7% and Eo. :
strand sample by the manufacturer to
l9l-192
GPa.
7.3 TESTING DETAILS

7.3.1 Test setup
Wall testing reported in this chapter was conducted using the setup shown in Fig 7.5, consisting
in principle of the test unit, a foundation beam, a loading beam, a horizontal hydraulic actuator
and a reaction wall. The foundation beam was clamped to the floor with stress bar and the
loading beam clamped to the top of the wall by the three vertical prestressing strands. The
hydraulic actuator applied lateral load at a nominal height of 5,250 mm above the foundation
beam, and was prestressed to the loading beam by means of two 23 mm VSL CT stress bars
that spanned the full length of the loading beam.
Addition axial load, modelling gravity and live loads from suspended floors, was applied by
means of the external axial load system shown schematically in Fig. 7.6. The system consisted
of trvo vertical 23 mm VSL CT stress bars spanning between the foundation beam and the steel
cross-beam. Hydraulic hollow jacks applied axial force in the bars. The jack shown on the left
hand side (north face) was connected to a hydraulic pump providing constant pressure thus
constant force. The cross beam was free to pivot about a horizontal axis parallel to the plane
of
the wall, thus equalising the force in the two external bars. The jack on the right hand side
(south face) was passive and was used solely to adjust the pivot angle of the cross-beam. As the
wall was displaced laterally and rocking action extended the distance between the foundation
TABLE 7 .Z-Material properties
I't storey
Wall
fs
fn,
uloo I uzoo lcproo
s3-l
t7.9
s3-2
r3.6
MPa
t2.7
18.I
MPa
Cavex
fs I t, lAee at
No Cavexl mortar I DOT
l 1.3
25
t3.7
7.6
12.6
23.9
2l
MPa
MPa
MPa
MPa
davs
-152-
2nd
and
f-
td
store ys
fsl Age at
u200
Cavex
r6.0
8.3
30
MPa
MPa
days
DOT
400
kN
Actuotor
iiiiiiiiiiiiiiiiiiiiiiittiiiiiitiiiillllil
Strong Floor
Fig.7.!-Testing
setup
and the loading beam, the left hand side jack length automatically adjusted accordingly so that
the external axial load was kept constant.
Fig. 7 .7 shows a photograph of the south face of wall 33- I in the testing setup, ready for test-
ing. Lateral restraint was provided to the wall at RC floor slab and loading beam levels by
means of horizontal steel sections on both sides of the wall, spanning between the black col-
umn in the foreground and the strong wall extension in the background. Steel rollers were
attached to the RC slabs, two on each side, and bore against the horizontal beams. Lateral
restraint to the loading beam was provided by low friction Teflon to stainless steel interfaces,
two on each side.
7,3.2 Instrumentation
Fig. 7.8 schematically shows the wall instnrmentation. Lateral displacement was measured at
floor and loading beam levels as shown in Fig. 7.8(a). Instrumentation was tripled at each level,
consisting of portal gauges (PG for small displacements), turn potentiometers (TP for large dis-
placements) and digital display extensiometers (DG for verification of PG and TP). These
-153-
Block-oul
in
500
RC slob
kN
llow jock
sfonl
moinloined
by hydroulic Pump
for
odjuslmenl of exlarnol
bor lenglh
Fig. 7.6-External axial load system

instnrments were mounted on an independent instrument column which is seen between wall
and strong wall extension in Fig.7.7. Horizontal load was measured by the load cell LCH posi-
tioned in series with the lateral hydraulic actuator, see Fig. 7.8(b). Force in the prestressing
strand was measured by load cells
LCPl,LCP?
and LCP3
lear the anchorage point in the test
hall basement. Applied extemal axial load was measured by load cells LCAN and LCAS.
Fig. 7.8(c) shows flexural instrumentation FN (north face) and FS (south face) of the portal
gauge type used for measuring vertical strain throughout the wall. Portal gauges FN0l-05,
FN08-27, FN30, FN38, FS01-05, FS08-27, FS30 and FS31 had a nominal gauge length of 200
mm; all other FN and FS portal gauges had a nominal gauge length of 400 mm, All FN and FS
portal gauges spanned between bed joints. Sliding displacement was monitored by the instrumentation SL shown in Fig. 7.8(d).

The walls were subjected to cyclic in-plane loading applied to the loading beam located
approximately 5,250 mm above the wall base (see Fig. 7.5). The testing procedure adopted for
-t54-
Fig.7.7-Photograph of S3-1 in the testing setup ready to be tested.

these tests was based on that adopted by Rahman and Restepo 17-21and Holden
[7-3], and
consisted of t'wo loading regimes:
(i) Force controlled load cycles for serviceability loads with forces limited by the nominal flexural strength. These force cycles consisted of one cycle (one push and one pull excursion) to
the wall cracking strength, to the wall maximum serviceability strength (maximum allowed
serviceability limit state concrete masonry stress), and to 75% and 100% of nominal flexural
strength.
(ii) Displacement-controlled cyclic loading, where the wall was cycled twice to the drift target,
followed by one cycle to the previous drift target. ln this chapter drift is defined as the ratio
the top lateral displacement d to the wall height
4: y:
of
d/h". Displacement to drift levels rather
than displacement ductility was chosen in the spirit of performance based design, where the
drift ratio appears to be a suitable design parameter.
-155-
TP,SLS
= Tum Pol.nllom!1.,
PG.FI{,6.S1-SLI{ - Porlol Goug.

OC . olgBcl Erl ndom|.r
South foce
(o) In-plone Loterol
Push
+
o
C'
o
5
!o
z.
displqcemeni
Pull
o
.c
l
UI
(b) Lood Cells

Fig. 7.8-Instrumentation, (a) in-plane lateral displacement and (b) load cells
7.3.4 Flexural strength prediction

For a symmetric distribution of post-tensioning and gravity actions, first cracking strength, V",,
maximum serviceability strength, %, and the nominal flexural strength, V6, ore defined by
Eqns. 8.2, 8.6 and 8.17, respectively. The base shears,
V", V* and Vs, correspond to the calcu-
lated flexural moment at the wall base. For nominal flexural strength, the tendon force increase
AP due to wall deformation was calculated using Eqns. 8.25 and 8.26. The equivalent rectangu-
156 -
Pull
+.
Lcvrl
Push
Push
Pull
Norlh foce
South foce
(c) Verticol
deformotion
Push
Push
Pull
South foce
North f oce
(d) Sliding
displocement
Fig.7.&-Instrumentation (Cont.), (c) vertical deformation and (d) sliding displ.

lar compression zone length, a, defined by Eqn. 8.18 was calculated using
cr:
0.96 (confined
grouted concrete masonry). Table 7.3 reflects the calculated strengths.

The predicted concrete masonry shear strength, V,,, (force), was calculated according to Eqn.
5.2. Shear strength due to shear reinforcing, Vr,, was calculated using Eqn. 6.1. The predicted
-t57-
wall shear strengths V,: V^
* V,, are listed in Table 7.3. The initial
prestress force was used in
this calculation.
7.4 TEST RESULTS

This section describes the observed behaviours of S3-l and S3-2 and presents the measured
masonry compression toe stains at various wall states. General wall behaviour is summarised
in Thble 7.4, where V,n* is the maximum lateral force recorded and du.u* is the corresponding
displacement. The ultimate displacement capacity, d", is defined as the point at which the lateral wall strength had degraded below 80% of Vn.,",.. The ultimate drift capacity is defined as yu
= 4/h" with tq
:5.25
m.
Rocking response was recorded for both wall tests, with a single large crack opening up along
the wall-foundation construction joint. No distributed flexural cracks were observed. Addition-
ally, singular horizontal cracks developed at the interface between the bottom storey and the
level2 floor slab for both walls as a result of flexural action. These cracks developed at loading
level Vs for
S3-l
and 0.25% drift for S3-2, extending 400-600 mm from the wall ends, and
closed fully upon unloading. At force levels near nominal flexural strength,
Vs: Mo/h*
wall
TABLE 7.3-Strength predictions
\ilall
v",
vf
AP
s3-1
420
398
62
r26
165
2l
673
s3-2
206
427
48
98
t29
2l
647
KN
KN
KN
KN
KN
KN
KN
comments
IABLE 7.4-Predictions and results

Wall
vf
s3-l
r65
s3-2
direction
Vt"*
d"."*
T"."t
du
Y"
Push
208
49.5
0.94
79.2
l.5l
Pull
-t92
-36.6
0.70
-67.2
1.28
Push
t74
65
1.24
79.8
1.52
Pull
-156
-54.4
1.04
-93.s
1.78
KN
fitm
o/o
mm
r29
KN
-158-
Masonry crushing,
gradual flexural failure
Masonry crushing,
gradual flexural failure
softening occurred due to extensive opening of the base crack and the initiation of rocking.
Tendon yielding was not experienced.
regions, resulting in gradual strength degradation. Visual inspection showed that using CPl00,
inelastic vertical masoffy strain had penetrated into the third course above the foundation (0-
300 mm). The cycles to drifts of 0.25% to 0.5oh revealed little evidence of distress of the
masonry compression zones. First visual indication of fine vertical splitting cracks in the face
shells of the first masonry course was observed at 0.5% drift. During cycles to drifts of 0.75o/o
and 1.0% some spalling of mortar in the lower bed joints was obseryed and vertical cracks
occurred in the masonry face shell at each wall end in the first course (100 mm). No strenglh
degradation was observed at this stage.
Negligible sliding displacements were measured and there was no indication of friction created
by the lateral restraint devices due to out-of-plane wall displacement.

Fig.7.9, depicting the force-displacement response for 53-1 and S3-2, demonstrates the nearly
non-linear elastic behaviour exhibited by the tested unbonded post-tensioned walls prior to toe
crushing. It is seen that the walls returned to their original alignment, even after large displacement excursions. This type of response may be categorised as 'origin oriented'. The individual
It
is
noted that the term 'elastic' in a rigorous sense indicates reversibility and no accumulation
of
curves for each excursion appear pinched, implying little hysteretic energy dissipation.
damage. In reality the walls did accumulate damage from the onset of masonry crushing, caus-
ing a reduction in wall stiffiress and some hysteretic energy dissipation. Thus the behaviour
was temred 'nearly' non-linear elastic.
Significant strength degradation for both walls occurred in the first excursion cycle to +I.5o/o
drift. With the intent of displacing S3-2 to -l.SYo drift, the wall was accidentally displaced
much further to -z.l%o drift, resulting in extensive crushing of the compression zone and subsequent severe strength loss for the pull direction.
Notably, neither wall exhibited symmetrical response in the push and pull directions, with the
maximum push strength being 8-10% higher than the maximum pull strength. This discrepancy was recorded despite symmetric strengthening of the wall comers as indicatednFig.T.2.
-159-
250
1200
V=165kN
200
zv
o
o
900
150
600
100
300
50
.c.
o
tl,
-50
-300
-100
-600
U)
G
dl
-150
-900
-200
-250
V. = 165 kN
-80-70-60-5040-30-20-10
10 20 30 40 50 60 70
zgE
co
o
@
(g
co
-1200
80

(a) S3-l
Drift (%)
200
150
z.Y
-2.O
-1.5
V = 129
-0.5
0.0
1.0
1.5
kttl
600
100
300
50
-g
o
U'
o
(n
-50
U)
-3oo
(t)
-100
-600 =
s
-150
-200
-120 -100
5
#
ogco
e
z
V=129kN
-80 -60 -40 -20 0
20 40 60

(b) s3-2
Fig. 7.9-Force.displacement histories
-160-
80
-900
1.5
,1 i
0.5
nA^/\/\^
;o
-0.5
I/
/\ A
',- r- rvoVrVt\/t\/eVs ',\ /,,\
I
't2
r6
'ol
YY
-1
-1.5
-2
Cycle
(a) S3-l
1.5
ll
,|
tt
0.5
:g
Eo
o
AA/\/\ A/ /\ A/
t\t
/'V'v'
'1
,n
1 2 .-oVuVu
-0.5
,'\l,v',
IY
It '1
1t
It
tl
-1
-1.5
-2
Cyclo
(b) s3-2
Fig.
7.I0-Cyclic loading histories
It is seen in Fig. 7.9 that both S3-l and S3-2 developed significantly higher strength than predicted (V1), specifi cally 2lVo and.28o/o more than Vl for S3-l and S3-2, respectively.
Fig. 7.10 shows the loading histories for the test units. The two walls were taken through similar cyclic displacement patterns with a few exceptions, most notably that S3-2 additionally was
taken through two cycles to +1.25% drift.
-16l-

Gradual strength degradation was observed for both walls and was atfibuted to spalling of face
shells and crushing of the grout core. Rupture of the confining plates was not observed but
some distortion of the shape occurred in the bed joints in the crushed comers. Large displace-
ment capacity was observed with reliable wall resistance for drifts, yu, ranging from 1.3% to
l.5o/o for
S3-l and from I .5%oto l.8Vo for S3-2, refer to Thble 7.4.Damage patterns after fail-
ure are depicted by the photographs in Fig. 7.1 1, showing both sides of the lower storey of both
walls. The photographs indicate that the extent of damage was confined to the lowest three
masonry courses (300 mm) in the toe regions, with the majority of damage concentrated in the
lowest two courses (200 mm). The damage to the masonry units immediately above the third
course (300 mm and further above base) generally was limited to vertical splitting cracking
near the wall end. No cracking, whatsoever, was observed in the upper two stories.
7.4.3 Sliding
Fig. 7.12 presents the recorded sliding displacement for the two wall tests. It appears from
Figs. 7.12(a) and (b) that sliding increased gradually with lateral displacement. The maximum
sliding displacement was of the order of 1.5 mm to 2.5 mm, or less than 3% of the total lateral
displacement measured at any excursion peak. It is therefore concluded that base sliding displacements were insignificant in comparison with the total lateral displacement and that the
shear friction between the intentionally roughened foundation and the masonry (cleaned for
mortar droppings before grouting) was adequate to resist sliding.
7.4.4 Prestressing force

The prestressing tendon force histories (sum of force of all tendons) for both walls are plotted
against the lateral displacement in Figs- 7-13(a) and
ft).
For reference, it is noted that the nom-
inal yield strength for one strand is about 213 kN.

These figures reveal that the prestressing force at all excursion peaks exceeded the initial prestressing force and that little tendon force loss occured until the detrimental excursions to
*I.5o/o drift. Recorded data showed that all prestressing strands remained elastic throughout
testing. After failure prestress losses of
8o/o
and l0% were recorded for S3-1 and S3-2, respec-
tivelv.
162 -
(a) S3-l Southface
(b) S3-1 North face -West comer
Fig. 7.f
f{a)
and O), S3-1 Damage accumulation at failure
-163-
(c) S3-2 South face
i
I
i'-
(d) 53-2 North face
Fig. 7.11-(c) and (d), S3-2 Damage accumulation at lailure
-164-
Drift (%)
-1.5
2.5
-1.0
-0.5
1.0
0.0
1.5
2.0
1.5
E
E
g)
.E
1.0
0.5
0.0
U'
-0.5
(D
(E
-1.0
m -1.5
-2.O
-2.5
-80-70-60-50-40-30-20-10
10 20 30 40 50 60 70 80

(a) S3-1
Drift (%)
2.5
-2.O -1.5 -1.0 -0.5
1.0
0.0
1.5
2.O
1.5
E
E
1.0
o)
0.5
0.0
-0.5
.c
o
o
(u
dl
-1.0
-1.5
-2.O
-2.5
-'120 -100
-80 -60 -40 -20 0
20 40 60

(b) s3-2
Fig. 7.l2-Sliding displacement
-165-
80
525
zL
o
Io
II
o
(r,
o
o
E
(L
6
o
s00
475
450
425
400
375
350
InitialPrestress 398
-80-70-60-5040-30-20-10
10 20 30 40 50 60 70 80

(a) S3-l
Drift (%)
-1.5 -1.0 -0.5 0.0
550
z-g
0.5
[
[l
500
[
475
525
o
e
o
4
ll- 450
(t,
@
425
a 400
E
TL
375
(E
350
r
I
lnitial Prestress 427 kN
l-
[
[
300 L325
-120 -100
-80 -60 40 -20 0 20 40

(b) s3-2
Fig. 7.l3-Prestressing force histories
-166-
60
7.4.5 External axial force

Figs. 7.14(a) and (b) show the variation of the externally applied axial force for S3-l and S3-2.
The target forces and typical excursion peak forces are shown in the figures and indicate that
approximately 30 kN more than the target force typically was applied for both walls. This corresponds to a total axial load increase (N+P) of approximately 3.8%o and 5.0% for
S3-l and 53-
2, respectively. The external axial force variation was attributed to friction in the hydraulic
jack.

Vertical sffain along the bottom of the walls recorded during testing is shown in Figs. 7.15 and
7.16. Each figure (plots a through h) shows strain plots at displacement peaks for the first cycle
to a given lateral drift for instrument levels 1 through 4 above the base, i.e. average strain for 0200 mm,200-400 mm,400-600 mm and 600-800 mm above the foundation. As flexural instnr-
mentation was mounted on both sides of the wall, the average strain at each location was used.
The sign convention in these figures defines compression strain as negative. The vertical thick
lines in the figures indicate the wall extremities, i.e. the exteme masonry fibres at+1.2 m from
the wall centre line. Plots (a), (c), (e) and (g) relate to low displacement response, excursions to
V"o V", Vs and approximately 0.3o/o drift, and plots (b), (d),
(0
and (h) relate to response
beyond 0.3% drift. Strain measurements for these walls were recorded between bed joints
of
two courses of CPl00.
No plots of vertical strain above level 4 have been included in this report because the recorded
vertical strain indicated purely elastic response (vertical strain significantly lower than 0.001).
This attests to the vertical extent of the compression zone undergoing plastic deformation
remained rather short. Furthermore, the resolution
of the utilised
instruments
of perhaps
0.0002 strain (using a gauge length of 200 mm) excluded meaningful interpretation of data
recorded above level4. It is remarked that the data plotted in Figs. 7.15 and 7.16 is associated
with some degree of uncertainty, especially at high drift levels where deterioration of the
masoffy unit face shells may have affected instrument reading.
Figs. 7.15(a)+(b) and 7.16(a)+(b) reveal that vertical strain varied nearly linearly along the
length of the walls base for all displacement levels.
-167-
400
zJ
o
g
o
lr
Typical load at excursion peak

375
350
(r,
325
[L
300
275
-80-70-60-5040-30-20-10
10 20 30 40 50 60 70 80

(a) S3-l
Drift (%)
-2.0 -1.5 -1.0 -0.5
190
z.V
o
e
o
TL
o
a
E
U,
E
(L
180
1.0
Typical loVd at excursion peak 17(
170
--r,--=G--
160
150
140
130
120
0.0
110
-120 -100
Target 140 kN
-80 -60 -40 -20 0
20 40

(b) s3-2
Fig. 7.l4-External axial load histories
-168-
60
1.5
0.0200
0.0050
L.vcl l (rvrnga
0.0175
0.00.15
0.0150
0.0040
0.01
25
0.0100
0.00 35
0.0075
0.0030
0,0050
0.0025
0.0025
0.0000
0_0020
-0.0025
0.0015
0.0010
;
T
-0.0050
-0.0075
-0.0100
-0.0125
0_0005
-0.0150
0,0000
.0.0 r 75
.0.0005
-0.0200
-0.0225
-0.00r0
-0.0250
.0.0275
-0.00'15
-0.0300
.0.0020
-0.0325
-0.0350
-0.0025
-1.0 -0.8 -0.6 -0." -0.2 0.0 0.2 0.4 0.6

Loc.tlon !|otr9 wrll lcngth (mml
0.8
tn
-1.2 -1.0 -0.8 -0.8 -0.i1 -0.2 0.0 0.2 0.'l 0.6
Loc.don llong f,rll l.nglh (mn)
1.2
0,0002
0.0001
0_0000
.0.0001
w F=
L.Y.l
1.2
(b) Above 0.250 ,level0-200 mm
(a) V.,-0.257o,level 0-200 mm

0.0003
0.6 1,0
(rvrrag. 200{00nn
L!vol
0.0003
(.v.rugt 200-'l00nn
0.0000
-0,0003
-0,0006
-0.0009
.0.0002
{.0012
.0.0003
c
@
E -0.0015
I?
.0,000,1
.0.0005
-0.0018
-0.0021
-0.0006
-0.002,1
-0.0007
-0.0027
.0.0008
+.1.3rt
-0.0009
-0.0010
-0.0033
+o,at*
-0.001'l
-'t.2
-1-O -0.8 -0.6 -0.4
-0.2 0.0 0.2 0.4 0.6 0.8 1.0
o/o,
+0,62!a
-{F0.s6ta +1.23*
-0.0036
1.2
-r.2 -1.0.0.8 -0.6 -0.t -0.2 0.0 0.2 0.4 0.8 0.8 1.0
Loc.llon rloFg wall letrgth (nG)
Loc.llor slong r.ll hnglh (mml
(c) V",-0.2
-l-.t.ot!a
+-0.71% +-0.53!6
-{r.0030
(d) Above 0.250 ,level200-400 mm
level 200400 mm
Fig. 7.15-53-1, Ilorizontal strain profiles,levels 0-200 mm and 200-400 mm
-t69-
'1.2
0.0002
0.0002
0.000
0-000'1
0.0000
0.0000
-0.0001
{.0002
-0.000'1
.0.0003
-0.0002
f;
a
-o.ooor
-0.0004
-0.0004
c!
IE
t
S
-0.0005
-o,oooo
-o.oooz
-0.000s
-0.0005
-0.0009
-0.000e
-0.00r0
-0.001
-0.0007
-0_0012
.0.0008
.0.0013
-0.0014
.0.8 -0.6 -0.i| -0-2 0.0 0.2 0.4 0,8

Locrtlon .long wrll langrh (mm)
0.8 1.0
-1.2 -1.0 -0_8 -0.6
1.2
-O_1 -O-2
0.0 0,2 0.4 0.8 0,8 1.0
1.2
Locltlon .lon9 wrll langth (n ml
(f) Above 0.25o/o,level400-600 mm
(e) V".-0.25Vo, level 400-600 mm
0.0002
0.0002
0.000
0.000r
0_0000
0.0000
-0.000r
-0.000
-0,0002
.0.0002
-0.0003
;o
-0.0003
-0.0004
-0.0005
-0.0004
-0_0006
-0.0005
-0.0007
.0.0006
-0-0008
-0.0007
-0.0009
-0.0010
-1
.0 -0.8 -0.6 -0.'l -0.2
0.0 0.2 0.4 0.6 0.8
Loc.alon rlong
(d
Fig.
1.0
rell length (mml
.2
-1.O -0.8 -0.6
-0.1 -0,? 0.0 0.2 0.4
0.6 0.8
1 .0
Loc.alon.long wlll length (mml
(h) Above O.2s%o,level 600-800 mm
%,-0.25olo, level 600-800 mm
7.lf(Cont.)
-1
S3-1, Horiz. strain profiles,levels 400-600 mm and 600-800 mm
-170-
1,2
0.0050
0.0200
0.0045
0.0175
Lowl t (rv.ngo
0.0150
0-00,+0
0.0125
0.0035
0.0100
0.00t0
0.0075
0.0025
0-0050
0.0020
0.0025
G
?
6
T
0.00 r5
t
a
It
0.00r0
0.0000
-0.0025
0.0005
-0.0050
0.0000
{.0075
-0.0005
4.0100
-0.0010
-0.0 I 25
-0_001 5
-0.0'150
-0.0020
-0.0175
-0.0025
-0.0200
4.0030
.0.0225
-1-2 -1.0 -0.8 -0.8 -0..1 -0.2
0-0 0.2 0.1 0.8
Locrtlon.lon! wrll lcngth
0.8
t.0
-r.2 -1.0 {.8 4.6
1.2
-0.,1
{.2 0.0 0.2 0..f 0.6 0.8

rrll l.nglh (mnl
1.0
12
Loc.llon tlong
(mm)
O) Above 0.3%o,lwel 0-200 mm
(a) V..-0.3%, level 0-200 mm
0.0002
0,0002
0.0000
0.0001
-0.0002
0.0000
-0.0004
-0.0001
-0.0004
-0.0002
-{,.0008
-0.0003
c
E
6
-0.001 0
-o.ooo/.
i? -o.ooos
-0.0012
C -0.0014
-0.0018
-0.0006
-0.00 18
-0.0007
-0.0020
.0.0008
.0.0022
-0.000s
-0.002if
-0.0010
-0,0028
-0,0011
-0.0028
-.t.2.1.0,0.8.0.6 -0.4 -0.2 0.0 0.2 0.,1 0,6 0.8 1.0 1.2
Loc.llon .long
r.ll
hngth
(c) V",-0.3%,level 200400
(mml
mm
+-1-27 -4.--1.02t{
+.0.75* --..-.{t.sfi(
+o,aa* +0,8t*
-+F0_9t!( +1.22*
-1.2 -t.0 -0.8 -0.6 -0.4 -O.2 0.0 0.2 0.,1 0.8
LoEltlon rlong wlll l.ngrh (nml
0.8 1.0
(d) Above O.3Yo,level200-400 mm
Fig. 7.16-S3-2,Horizontal strain profiles,levels 0-200 mm and 200-400 mm
-t7t-
1.2
0.0002
0.0002
L!v.l t (rvorrlc a00-600mn,
0.000 I
0.0000
0_0000
.0.0002
-0.0001
-0.0004
-0.0002
{.0003
{}.0006
a -0.0004
-0.0008
-0.0005
-0.0010
-0.0004
-0.0007
-0.00 r 2
-0.0008
-0.0014
-0.0009
-0.00 r 6
-0-001 0
-0.00r
.r.2 -1_0.0,8 -0.6 -0.a -o.2 0.0 0.2 0.4 0_6
-0.0018
0,8
-1.2 -1.0 -0.8 -0.6 -0.4 -0.2 0.0

Loc.llqn .long {!ll
1.0
Lo3.llon .long rall lcngth (mm)
0.2 0.4 0.6 0.8 1.0
1.2
16notlt (mm)
(f) Above 0.3Yo,level400-600 mm
(e) V",-0.3%, level400-600 mm
0.0002
0.0002
Lrv.l
(rv!.rgr 600-800nm)
0.0001
0.0001
0.0000
0.0000
,0.0001
-0.0002
-0.0001
-0.0003
-0,0002
-0.0004
-0.0005
,0.0003
F
-0,0004
E -0,0007
-0.0006
.0.0005
-0,0008
-0.0009
-0,0006
-0.001 0
-0,0007
-0.001
-0.0012
-0.0008
-0.0013
.0,0009
-0.001,1
-0.0015
-0.0010
-1,2 -r.0
-0.8
{.5.0.,1 -0.2 0.0 0.2 0.4 0.6
0.8
'|.0
1.2
Locrllon rlong wrll llngth (mm)
-1.2 -1.0 -0.8 -0.6 -{t.ia -0.2 0.0 0.2 0.4 0.6
Locatlon llong wrll lgngth (mml
0.8
1.0
(h) Above 0.3o ,level 600-800 mm
(g) V",-0.3%,level 600-800 mm
Fig.7.16-(Cont.) S3-2, Horiz. strain profiles,levels 400-600 mm and 600-800 mm
-172-
1.2
The general tends for masonry strain for low level response (plots (a)) were extreme fibre
stains of 0.0010 to 0.0020 for readings up to Vr and approximately 0.0020-0.0030 for 0.25o/o0.3% drift. The exffeme fibre strain for large displacement response ranged from approximately 0.02 to 0.04. Extreme strains at levels 2 to 3 exceeded 0.0015 strain indicating the
masonry responded inelastically. Some inelastic response (strain higher then 0.0015) was also
recorded for S3-2 level 4. Note that the above considerations of extreme strain at wall ends
(t I .2 m) were based on exfapolation of recorded strain
at lateral locations +0.9m and + I . I m.
Table 7.5 presents a summary of vertical strain measurements and vertical dimension of the
plastic region for the extreme masonry fibre, related to drift level peaks. The extreme masonry
sfrain was foundby extrapolation of the results presented in Figs.7.l5 and 7.16to the wall
extreme ends at +1.2 m. Determination of the vertical extent ho of inelastic masonry response
was based on a reference strain of 0.0015, such that a masonry instrument level with average
extreme fibre strain measurement exceeding the reference strain was deemed inelastic. The
TABLE 7.S-Vertical strain

Push
Push
WaIl
s3-l
s3-2
drift | ,o"io I n'
levent
drift | ,o"in I rt"
levent
vl
0.0010
l/f
0.00t4
0.26%
0.0020
200
0.24%
0.0024
200
0.53%
0.0037
400
0.47%
0.0032
400
0.7r%
0.0057
400
0.62%
0.0035
600
r.06%
0.0120
400
0.96%
0.0056
600
t3t%
0.0226
400
r.23%
0.0126
400
152%
0.0362
400
t5l%
0.0193
400
vf
0.0022
200
vJ'
0.00r2
0.3r%
0.0033
200
0.30%
0.0024
200
0.5r%
0.0036
400
4.44%
0.002s
600
0.7s%
0.0043
400
0.68%
0.0038
600
r.02%
0.0066
400
0.99%
0.0057
600
1.27%
0.0099
400
r.22%
0.0079
600
Vr-
2.r2%
0.0437
400
|50%
0.0242
400
mm
V**
Tu
V-*
-t73-
mm
Vrn"t
extreme strain presented in Thble 7.5 reflects the average of the strains measured by the instruments within the vertical extent of ho. The following was observed from Thble 7.5: (1) vertical
strain at nominal flexural strength Vs indicates elastic masonry response (in italic font) at this
load level (with the exception of S3-2 pull), (2) shain at
V,'*
ranged from 0.0056 to 0.0079,
(3) sfrain at yu ranged from I .9o/o to 2.4o/o, (a) ho at V,no measured 400 mm for pull excursions
and 600 mm for push excursions and (5)
lb
at % measured 400 mm
for both walls.
7.4.7 Lateral wall displacement profile

Figs. 7. l7 and 7.18 show the lateral displacements at the RC floor slab and loading beam levels
recorded at peaks of first excursions to the specified drift levels.
Studying the low level response plots (a),
it is seen that both walls exhibited distinct flexural
action for loading between V", and Vr. These displacement profiles are 'curved' with distinct
'kinks'at the first RC floor slab level, suggesting that a significant amount of the lateral displacement arose from flexural action throughout the wall. The curvature became less distinct
as the
walls were loaded to higher force.
Looking at the plots of high drift level lateral displacement profi.les in Figs.7.17(b) and
7.18(b), it is clear that rocking action at the base became the dominant source of lateral displacement. The lateral displacement profiles became straight lines and the 'kink' previously
observed at the first RC floor slab level disappeared.
7.5 DISCUSSION
7.5.1 Flexural response
Fig. 7.9 indicates that the developed flexural strength V*u* (base shear corresponding to the
maximum moment) exceeded the predicted flexural strength Vl in both wall tesl specifically
53-1 developed2lo/o higher strength (average of push and pull) than V6, while S3-2 developed
28% higher strength than Vs. These measured strengths were surprisingly high because calcu-
lating the wall strength (base shear corresponding to the restoring moment due to P+N) using
the measured peak values of P and N and cr
0.96 could not account for the measured strength
in either of the wall tests. The author is confident that the lateral load (LCH) was measured
within
5olo
of accuracy and that little horizontal resistance was caused by the lateral restraining
devices. The load cells measuring the additional axial force (LCAN and LCAS) recorded simi-
n4-
6S0
sdm
4!100
E
E
c
.9
{CIe
3500
E
t
e
g
ooo
2roo
5q
,o
't
looo
tcoo
o||Ctaf
fm0
+-w
500
0
.C{-7{.5.aa.2.l0lzE
Labral Dl.Fl.tsrmnl (mtn)
(a) Vo-V1 (digfial ortensiometers)
!!00
lo0
100
tooo
g loo
I
E
f
e
tI
6
rgoo
2500
lo00
It
o
E
rroo
10oo
!00
0
{o -70 {0 s
{o .to -n 'ro 0 10 N,
l-alrral Dhptacament (mml
(b) All drift lwels (Pc and TP)
Fig. 7.17-rS&1, Lateral displacement
profle
- 175-
a0
Bdeodomo@
-"o..W
+--v,
.-6...Vo
+'V.cr
"Q .' Ver
55oo
5000
tE00
1000
F
g
3000
E
t
aom
estto
E,
z00o
.o
c
toil
r
Io' ttoo
Otgitd EensioroBtars
t0m
0:
42.10{.t{.20
Lalerel lDlrplacermnt
(nml
(a) V*-% (digital exronsiometo,ts)
t*r
,*1
1600
1
o
c
E
c
e=o
t
.8
c
E
E
ld
E
..{'-"vf
-+--vs
"$..v6
+-v9f
!00
-+jvl
..,
*JI
'*l
*1
"-ll +4.1W
'*l +-1.03%
-e-{7596,
+4-5ilt
---r--1.27%
"*ll +4.3ti6
-+-.w
ill
-F-llc
->-ll0r
4m ll|rl0 rl0o
.-r'-1,5"6
+ - - | ft|o$
-:t-.:0,9&16
--
r:c:r0;68!i6
.-r--0,"f4%
-.s--0r396
,,w
-{-.Ve
.-,f
..o-"Vq'
.t0
d10
-m
160
4n .10 .t0
-20
-tO 0 d0'
Leleral Dlsolacadtltnf trlDtl

(
O) All drift lwels FG and TP)
Flg- ?.1E-53'2r Lateral dfuplacement profile
-176-
gr
..-
u,- \fcf
Drift (%)
-2.0
250
-1.5
-1.0 -0.5
1.5
1.0
0.0
200
zlz
150
100
50
.E
0)
-50
a
.h
(u
-100
-150
II
\\-------
JII
/'
-200
-250
-120 -100
S3-1
,*,t'-t,
-80 -60 -40 -20 0 20 40
60
Fig. 7.19-Force-displacement envelopes
lar force throughout the tests, suggesting that their measurements were reliable. Apart from
constnrction inaccuracies, e.g. location of prestressing strand and external axial force device,
the load
the last source of error could be attributed to measwement of prestressing force. While
cell readings (LCpl, LCp2 and LcP3) were considered reliable, the actual prestressing force
between the bottom of the walls and the loading beam may have been higher than measured
because of friction generated between the strands and the holes in the foundation beam and
strong floor due to misalignment. This source of error was presumably not large because fric-
tion would have caused hysteresis that would have shown on the prestressing force histories in
Fig.7.l2
as increased area between loading and unloading curves.
Comparing the force-displacement envelopes (peak response of first excursion cycle for each
displacement level) for the walls, shown in Fig. 7 .19, it is seen that S3-l developed Zl%olttgher
strength than S3-2 (average of loading directions). This was a direct result of the total axial
force on S3-l being approximately 29Yo higher than that of S3-2.

Comparing the response envelopes in Fig. 7.19 and the results presented in Table 7.4,it is clear
that S3-2 developed the largest drift capacity; S3-1 sustained 1.40% drift (average of loading
-r77
directions) while S3-2 sustained l.65Vo drift. This was expected because of the lighter axial
force applied to S3-2, despite the fact that the S3-2 masonry crushing strength was significantlv lower than that of 53-
l.
7,5.2 Tendon behaviour

Tendon yielding was not expected and did not occur, as is illustrated in Fig.7.l3. At the peaks
of excursions both walls developed total tendon force higher that the initial tendon force. It
appears from the recorded tendon force data that
(single tendon)
of
S3-l developed a maximum tendon force
198 kN while S3-2 developed a maximum tendon force of 209 kN. The
nominal tendon yield force
of 213 kN,
reached during testing of S3-2.

based on the assumptions
specified by the strand manufacturer, was nearly
It is concluded that determining the initial prestressing force
of rigid body rocking behaviour and tendon yielding at 2% dirft.
frrnctioned as intended for both walls, maximising the tendon force (economy) while protecting the tendon against inelastic stress.

Table 7.5 shows the vertical strain recorded in the walls at various stages of the tests. It is seen
that the calculated masoffy strains corresponding to the nominal flexural condition, V1, ranged
from 0.001 to 0.0022. These results do not conform with the code defined nominal flexural
strength condition set out by NZS 4230 for confined masonry, that specifu a useable extreme
fibre masonry strain at nominal flexural strength of 0.008.

The masonry strain in the plastic deformation zone (ho) at
a small variation in vertical strain. The lack
V**
ranged from 0.0056 to 0.0079,
of visible damage at this limit state impelled that
the masonry fibres in the extreme ends of the walls still provided some axial stength resisting
the overturning moment. Stains at this limit state compared favourably with the code defined
useable strain of 0.008.
Table 7.5 indicates that masoffy extreme fibre strains ranged from 0.019 to 0.024 at the ultimate drift, yu, with values again showing little scatter considering the circumstances of measurement. These values are of course of theoretical character as the toe regions were damaged at
this stage, so that those strains should not be regarded as the reliable ultimate strain capacity
of
concrete masonry. The results do nevertheless suggest that masonry strains far beyond those
related to nominal flexural strength can be expected for rocking wall systems, while still pro-
-178-
wall
viding significant axial strength. At this limit state the effective flexural neutral axis at the
is found in Figs.
base had migrated towards the middle of the wall. Clear evidence of this
7.150) and 7.16(b), showing that the crossing of the strain profiles with the zero strain axis
moved closer to the wall centre as the walls were displaced to larger and larger displacements.
in
The height of the plastic deforrration zone was tracked throughout testing and is presented
V66,*,
Table 7.5. Inelastic deformation occurred over a height of approximately 400-600 mm at
while measgrements showed that this height had shortened at ultimate drift. Some unloading of
inethe masonry above 400 mm occurred at yu due to crushing of masonry below Nevertheless
lastic deformation had occurred over 400-600 mm-
7.5.4 Hysteretic enerry dissipation

Figs. 7.9(a) and (b) illustrate that little hysteretic energy dissipation occurred in the two wall
tests,
until reaching the 1.5% drift level. Nevertheless, the energy dissipated in each excursion
(half cycle) due to hysteresis can be quantified by integration of the area enclosed by the loading and unloading curves. Results of such integration applied on to the force-displacement
relationships for S3-l and S3-2 are illustrated in Fig. 7.20, showing the cumulated energy dissipation as a function of the wall lateral displacement history. As discussed above, the exter-
nally applied axial load varied during testing and therefore caused 'parasitic'
energy
dissipation. Calculations showed that the parasitic component of the energy dissipation
53amounted to less than l}vo and LlVo of the total hysteretic energy dissipation for 33-l and
2, respectively. plots (a) and O) in Fig. 7.20 show the 'true' hysteretic energy dissipated by
wall flexural action, calculated by subtraction of parasitic energy dissipation from total calculated energy dissipation.
It is seen in Fig. 7 .20
that similar amounts of energy were dissipated by the two walls and that
gs-g1%of the energy that was exerted by the horizontal actuator to displace a wall from zero
displacement to a displacement peak was typically released upon unloading, thus 10-15% of
energy was lost in a single excursion. This can be compared to the elasto-plastic hysteretic
behaviour of reinforced concrete masonry walls that dissipate energy in both the excursion segment to a displacement peak and the excursion segment that returns the wall to its original
alignment. positive work is required in both the loading and unloading branches of the excur-
sion. Clearly PCM dissipates far less energy than reinforced concrete masonry for walls of
similar characteristics.
r79 -
40
-J
P
o
C
35
30
UJ
o
o
o
25
15
20
q,
E0)
10
(U
=
E
()=
5
0
-80 -60 40 -20 0
20 40
60

(a) S3-l
Drift (%)
-lz
P
o
C
40
-2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0
35
30
LU
-9
25
q)
20
o
o
E
o
(E
15
10
=E
o=
-120 -100
-80 -60 40 -20 0
20 40

(b) s3-2
F
ig. 7.2G-Cumulated hysteretic energy dissipation
-180-
60
1.5
Drift (%)
-0.03 -0.02 -0.01 0.00 0.01
0.03
0.02
60
40
2
gzo
eo
a
(U
S -2o
(u
---+
-40
S3-1
S3-2
-60
I
-2.O
-1.5
-1.0 -0.5 0.0 0.5
1.0
1.5
2.0
Fig.
.Zl-Ilitial wall stiffness
7.5.5 Sliding proPensitY

Significant wall sliding was not measured as documented by Fig.7.l2. This phenomenon was
overcome by adequate shear friction befiveen wall and foundation provided by intentional
roughening of the wall to foundation interface and cleaning of the wall cavities in the compression zone for mortar droppings. Refer to chapter 8 for an in-depth discussion.

Estimation of initial stiffiress is of considerable interest for serviceability and ultimate loading
in relation with vibration characteristics and earthquake loading. Fig. 7.21 shows curves for
both walls relating to the wall secant stiffness at the theoretical first cracking limit state. It is
seen in the figure that the
initial stiftress varied between the walls, however the average initial
stiffiress of both wall were similar, 4l
I kN/mm for 53- I
and 4 I .0 kN/mm for S3-2. This is not
surprising because the wall initial stiffness theoretically only depends on the wall dimensions
and the masonry elastic properties, and these were nearly identical for both tests, with the
exception of the lowest storey where the elastic modulus of the masonry for S3-2 was some-
t8l
what lower than that of S3-1. The average initial stiffiress for the walls shown in Fig. 7 .21
of
41.1 kN/mm can be compared with the theoretical elastic stiftress of 44.1 kN/mm (based on
Eqns. 8.2 and 8.4 using
E^: 14.4 GPa and v :
0.2). The measured average wall initial stiffness
was thus 7%olower than the theoretical one, indeed a close correlation that confirms using ordi-
nary flexural theory for estimation of the wall initial stiffiress for walls with aspect ratios (hJl,")
ofthe order of2or higher.
7.6 COMPARISON WITH RCM WALL TESTING

This section compares the perfonnance of the two Series 3 walls with reinforced concrete
RCM) wall testing by Priestley and Elder [7-5]. Series 3 walls (S3-l and S3-2) and
the Priestley/Elder walls (PEl, PEZ and PE3) were of similar plane dimensions. The wall
masonry
heights were comparable with 5.25 m for the 53 walls and 6.0 m for the PE walls. There were
significant differences in the material properties and vertical reinforcement ratios. Nevertheless, meaningful comparison can be done after careful interpretation of results.
7.6.1 Priestley and Elder walls

The general dimensions and reinforcement of the walls tested by Priestley and Elder are shown
in Fig. 7.22. The wall thickness was 140 mm and the wall length was 2400 mm. The wall
height (top of foundation to point of lateral load application) was 6.0 m, thus l.14 times that
of
the 53 walls. The longitudinal reinforcement at the wall base consisted of DHl6 (16 mm
deformed bar with nominal yield stength of 430 MPa) at 200 mm centres and was lap spliced
at the wall base over 1000 mm for PEI and PE2 and 1300 mm for PE3. Externally applied
axial load and wall self-weight amounted to a total of 640 kN for PEI and PE2, and 250 kN for
PE3. PE2 had 600 mm long stainless steel confining plates embedded in the plastic hinge zone.
It was
estimated that the masonry crushing strength
of the PE walls was approximately
25
MPa.
The primary intent with the PE wall testing was to evaluate the available ductility of tall (and
slender) RCM walls as a function of axial load, lap splice length and bed joint confinement.
The main conclusions from the PE tests were that the testing results gave reasonable confirma-
tion of theoretical ductility capacity charts developed by the authors, however the theoretical
predictions were found to be conservative for the lightly loaded wall (PE3). It was found that
lap splicing in the plastic hinge zone clearly was a problem and should be avoided if possible.
-182-
!
I
ol
oi
o
b
o
a
O'
|/'
!
c
Ir
I
I
ZA
WALL EIVATIO{
B
RI}IFORCE'IENT
SECtlOf.l A'A
DE TAILS
ig. 7 .Z2-PriestleylElder wall specifi c ation, from [7-51
It was confirmed that ductility capacity of RCM walls is more of
a problem for
tall walls than
squat walls, but that the wall slendemess (height to thickness ratio) used in these tests was
little concern. It was finally concluded that the
use of confining plates
of
within the critical mortar
joints of the potential plastic hinge will substantially improve response to severe seismic
attack.
7.6.2 Stratery for comparison of results

In relation with evaluation of the present PCM wall testing it is irueresting to draw parallels to
the RCM wall performance discussed above. As suggested above this is not directly possible
because
of significant dissimilarities in wall height, masonry strength, applied axial load and
design strength. The results of these differences are presented in Table 7.6 where the applied
axial load, N, the theoretical nominal strength,
Mn: Vr\,
the experimental ultimate displace-
ment capacity, 4,.*0, (average of both directions of testing) and drift capacity, %,
il
given.
The axial load ratio due to self-weight is thr and the wall gross section area is given as
b*l*
Ar:
The theoretical strengths are based on [7-5] for the PE walls and Table 7.3 for the 53
walls.
183 -
The primary indicators of performance are:
L Strength
2. Displacement capacity
3. Strength degradation
4. Energy dissipation
5. Damage/Repairability
Clearly items
I and 2 above directly depend on applied axial load, N, vertical reinforcement
properties, masonry strength and wall dimensions. Therefore these perforrnance indicators
cannot be directly compared. Comparison of strength and displacement capacity consequently
is based on extrapolation of the PCM wall results. This is done by application of the prediction
procedure presented in Chapter 8 which was verified and calibrated against the Series 3 wall
testing. Specifically, two predictions have been carried out for f'
0.14 m,
tr":6.0 m: (PCM-I)
ing steel area of Ao, :
N: 640 kN and Yn:272
l115 mm2, and (PCM-2) N
^:
25 MPa,
l*
: 2.4 m, br:
kN which was achieved by a prestress250 kN and
Yn: 222 kN which was
: 1075 mm2. The initial prestress was taken Bs t" :

956 MPa or 630/o of the yield strength, E-: 20000 MPa, Eo, : 190 GP4 \: 0.456 ffi, t ,:
0.013 and K: 1.06. Results of the prediction are presented in Table 7.7. Figs. 7.23 to7.25
achieved by a prestressing steel area of
A*
show the predictions overlaid on the experimental force-displacement curves by Priestley and
Elder. It is noted that both PCM-1 and PCM-2 predictions were based on confined half height
masonry blocks, CPl00. Only PE2 had confinement plates embedded, in this case
in full
height masonry blocks, CP200. Thus comparisons and conclusions from Figs. 7.23 to 7.25
TABLE 7.6-Theoretical and experimental results for PE and 53 walls

Wall
PEI
lN:N/(f'mAg)
vn
Mn
Au,"*p
640
0.076
273
1632
50
Y,r,"*p
0.E3
PE2
640
o.o76
273
1632
77
1.28
PE3
250
0.030
222
1332
89
r.48
s3-l
420
0.070
165
866
73
1.40
s3-2
206
0.045
t29
677
87
r.65
KN
kllm
mm
o/o
kfl
TABLE 7.7-Theoretical performance of PCM walls for comparison

Wall
PCM.I
PCM-2
lp=N/(P.Ag)
vn
Mn
Artn
Tu,*t
640
0.076
273
1632
58
0.97
250
0.030
222
1632
7l
l.19
KN
klrlm
mm
Vo
KN
-184-
r0
60
DEFLEUII0.I ln nl
.m
-2m
+PCM-I
l^l
,r .l3l
-m
I
?&
t(2
Prediction
Ultimate Displacement Capacity
o.7r
Fig.1.23-.PEl experimental result and PCM-I prediction envelope
0rr
ll
u2
.+PCM-l
lllll
.5.50
A
a.li
l.!a
t.LZ
Prediction
Ultimate Displacement CaPaclty
0'n
Fig.7.2EPE2 experimental result and PCM-I prediction
-185-
envelope
on
rl
l.m
{-PCM-2 ftediction
ll
160
Ultimate Displacement Capacity
0m
Fig.7.2*PE3 experimental result
and PCM-2 prediction envelope
strictly only apply to the specific construction methods applied, and are only suggestive in
tenns of comparison of CPl00 and CP200.
Items 3 through 5 can be discussed based on observations from the actual tests because the
trends for these items are controlled by the nature of PCM or RCM construction rather than by
actual applied force and strength.
7.6.3 Comparison of results

Comparing PEI and PCM-1 of similar nominal strength in Fig. 7.23,it is found that the maximum sfength of PEI exceeded that of PCM-1. This is because of considerable strain harden-
ing in the PEI mild reinforcement. Only modest prestressing steel stress increase is expected
for PCM-l as a result of to lateral deformation. It is also observed that the displacement capac-
ity of PCM-l is expected to exceed the displacement capacity measured for PE1.
Testing of PE2 with confining plates improved the RCM wall displacement capacity considerably. The maximum strength remained the same in comparison with PE1. Fig. 7.24 reveals that
the expected displacement capacity of PCM-I is lower than that measured for PE2. It is thus
suggested that a RCM wall confined with CP200 and under high permanent axial load provides
186 -
higher displacement capacity than a PCM wall of similar proportions confined with CPl00 or
practical
Cp200. The actual measured drift capacity for PE2 of 1.28% may however be of little
use because
it may exceed the drift limitations set out in the structural codes.
Comparing the result from testing of PE3 with predictions for PCM-2 in Fig. 7.25, both walls
subjected to relatively light permanent axial load, suggests that RCM walls provide larger displacement capacity than pCM walls, whether the RCM wall is confined or not. The ultimate
drift ratios achieved by pE3 (test) and PCM-I (prediction) of 1.48% and Ll9olo, respectively,
given by
are high and could both be considered excessive in comparison with drift limitations
structural codes.
The higher displacement capacity for RCM walls when compared to PCM walls is primarily
caused by considerable difference in the vertical extent of the plastic hinge.
It was found in the
priestley/Elder testing that the RCM plastic hinge typically extended vertically about 0.51* or
approximately l.Zm. pCM Series 3 testing suggested a vertical extent of the plastic deformation zone of about 0.0Stt or 0.48 m. Evidently higher displacement capacity can be expected
from RCM walls for a given masonry extreme fibre maximum strain.
Comparing force-displacement histories for Series 3 walls (Fig. 7.9) and PE walls (Figs. 7.23
to7.Z1)does not conclusively reveal whether strength degradation occurs more rapidly for one
wall type than the other. Reinforced concrete masonry exhibits high hysteretic energy dissipa-
tion in comparison with pCM because of compression and tension straining of the flexural
reinforcement. The prestressing steel in PCM walls dissipates little energy because little straining beyond the elastic limit as a result of unbonding of the preshessing is expected.
perhaps the greatest behavioural difference between PCM and RCM is found in the category
of
Damage/Repairability. Fig. 7.1I shows the Series 3 PCM wall damage after nominal failure.
As discussed in section7.4.2,damage was limited to the lowest three masonry courses. There
was no suggestion at any stage of testing that the ability to support vertical load was at risk.
It
is felt that damage could easily be repair, reinstating the full wall strength and integriry. On the
contrary, substantial damage was sustained in the Priestley/Elder wall testing. After nominal
failure, the first storey of all walls were cracked throughout, and face shell spalling and grout
core damage was found in lowest 5-7 masonry courses (1.0 m to 1.4 m). It was suggested that
the damage of the unconfined walls was impossible to repair while repair of the confined wall,
pE2, possibly could be carried out successfully, despite the extent of damage and flexural
-t87-
cracking. Clearly post-earthquake damage repair is much simpler for PCM because of the limited damage extent and virtually non-existing flexural cracking.
7.6.4 Conclusions from comparison of PCM with RCM testing

Conclusions of the PCM vs. RCM comparison can for masonry walls of similar dimensions,
masonry material properties and strength be summarised in:
(l)
RCM walls are expected to
exhibit larger displacement capacity than PCM walls because of longer plastic hinge length.
(2) PCM walls are expected to sustain much less and more localised damage during strong
ground motion than RCM walls, thus simpler and more economical post-earthquake repair. (3)
Both PCM and RCM walls with confining plates are likely to exhibit drift capacities in excess
of drift limitations for masonry walls given by structural codes.
It is reiterated that the above conclusions are based on experiments with significant differences
in the material properties and structural dimensions, underlining the suggestive nature of the
above comparison.
7.7 CONCLUSIONS
It is concluded that PCM walls of realistic proportions,
strengthened in the flexural compres-
sion zones with confining plates, can successfully withstand severe cyclic loading imposed by
an earthquake. Ductile response was measured with reliable drift capacity
of I.5%.
Only localised damage occurred, as shown in Fig. 7.11, making earthquake damage simple to
repair.
All visual damage occurred to the lowest 300400 mm in the flexural compression
zone.
This is supportedbythevertical strainplots shown in Figs.7.l5 and 7.16and Table 7-5,that

suggest inelastic masoffy response in the lowest 400-600 mm of the walls.
Relatively little energy dissipation was observed during cycling of the walls.
Tendon force loss was not recorded.
Measurements at nominal flexural strength suggest an extreme masonry fibre strain
in the
order of 0.0010 to 0.0020, values substantially lower than 0.008 stipulated by NZS 4230:1990
lT4lfor
confined masonry. At maximum wall strength, average strains of 0.0066-0.0080 were
measured.
At ultimate displacement, average measured strains in the extreme fibre were
188 -
as
high as 0.019-0.024. At this stage the exheme masonry fibres no longer carried axial load as
face shetl spalling had occurred.
It was concluded that
adequate shear friction between walt and foundation was provided by
intentional roughening of the wall to foundation interface'

Comparison between pCM and RCM walls suggested that RCM walls are expected to exhibit
larger displacement capacity than PCM walls because of longer plastic hinge length, that PCM
are expected to sustain much less and more localised damage during strong ground
walls
motion than RCM walls, thus simpler and cheaper post-earthquake repalr, and finally that bottl
pCM and RCM walls with confining plates are likely to exhibit large drift capacities which
may be in excess of drift limitations for masonry walls given by snuctural codes.
- 189-
7.8 REFERENCES
[7-l] AS 1311-1987, Steel tendons for prestressed concrete -
7-wire stress-relieved steel
strandfor tendons in prestressed concrete, Standards Aushalia, Homebush, NSW, Australia.
[7-2] Rahman, A.M. and Restepo, J.I. (2000), Earthquake Resistant Concrete Buildings:
Seismic Performance of Cantilever Walls Prestressed Using Unbonded Tendons,
Research Report 2000-05, Department of Civil Engineering, University of Canterbury,
Christchurch, New Zealand.
[7-3] Holden,
T.J.,
A Comparison of the Seismic Performance of Precast Wall Construction:
Emulation and Hybrid Approaches, Research Report 2001-04, Deparhnent of Civil

Engineering, University of Canterbury, Christchurch, New Zealand,.
[74] NZS 4230:1990, Code of Practice for
Association of New Zealand,1990, Wellington.
[7-5] Priestley, M.J.N. and Elder, D. McG., Cyclic Loading Tests of Slender Concrete
Masonry Shear Walls, Bulletin of the New Zealand Society for Earthquake Engineering,
Vol. 15, No. l, March 1982,pp.3-21.
-190-
Chapter
ANALYSIS AND DESIGN
8. PREDICTION OF WALL IN-PLAIIE
BEHAVIOUR
This chapter considers the in-plane flexural response of post-tensioned concrete masonry cantilever walls with un-bonded prestressing tendons, where the lateral force is assumed to be acting at the top of the wall or at some effective height tr". For other structural shapes and loading
configurations, the formulae should be mod"ified accordingly. Note that the term 'tendon' in the
following refers to both prestressing strands and bars.

The present chapter is concerned with analytical methods for prediction of the in-plane forcedisplacement relationship for unbonded post-tensioned concrete masonry. The scope of such
methods is to describe the wall behaviour throughout the entire loading range, including non-
linear large displacement response that can be anticipated in a seismic event.

Two approaches of various complexities have been investigated:
(I
) Flexure/Rocking model
(2) Finite Element model
Method
(l)
represents a simple approach based on ordinary flexural theory @lane sections)
where integration of curvature along the wall height can be applied to determine the wall lateral displacement and vertical deformation. In the large displacement range a rigid body rock-
ing analogy is used to describe the behaviour. This method largely focuses on establishing
analytical expressions based on equilibrium and first engineering principles that can capture
pCM response with reasonable accuracy. The prediction formulae can be used for limit state
design as well as for establishing the most likely force-displacement relationship (continuous)
for use in non-linear numerical analysis. This approach is the primary focus in this chapter.
Method (2): The Finite Element Model (FEM) technique supplements method
(l)
and is partic-
ularly useful for modelling of the wall internal force flow in the regions of force concenfra-
l9l -
tions, notably in the wall compression toe. The suitability of using FEM fibre elements for
modelling of PCM in-plane response is investigated in Chapter 9 in conjunction with dynamic
analysis.
A third method is the strut and tie methodology, with which the wall intemal force flow can be
analysed. This method is in particular useful for analysis of walls with openings and walls
unusual shape where the force flow cannot easily be described by method
of
(l). As all walls
investigated in this document were rectangular and without openings, pursuing this method for
analysis was not considered merited in this document.
8.1 FLEXURAL
RESPONSE OF CANTILEVER WALLS
It is assumed for flexural calculations that plane sections remain plane, i.e. there is a linear
strain distribution across the wall lenglh. This assumption enables, using first engineering prin-
ciples, calculation of strength, stiffrress and displacement, and implies distributed cracking up
the wall height. From laboratory wall tests (Chapters 5, 6 and 7) it was observed that the PCM
wall flexural response was primarily due to rocking where a crack opened at the base, and that
distributed flexural cracking was non-existing. This type of rocking behaviour is attributed to
prestressing with unbonded tendons. Despite this discrepancy between theory and observation,
it
appears that the assumption
of plane section response with distributed wall cracks provides
sufficiently accurate design rules.
8.1.1 Limit states

The flexural design procedure should be based on Limit State Design, as outlined by NZS
4203:1992 [8-1], which identifies nvo limit states, namely the Serviceability Limit State (SLS)
and the Ultimate
Limit State (ULS). The serviceability limit state generally deals with function
of the structure subjected to normal occupancy loads (service loads). It enswes durability and
comfonable occupancy of the structure, by e.g. crack limitation or limitation of deflection- The
criteria for the ultimate limit state relate to the strength and stability of a structure. This limit
state is reached when the structure or a part
of it reaches its ultimate capacity and fails to sup-
port further loading. These limit states generally need to be evaluated immediately after prestressing transfer and after long term losses (refer to Chapter 4).
Moreover, structures designed to withstand earthquake loading in a ductile manner should be

designed according to the capacity design principle, an extension of ultimate limit state design.
-r92-
The principle is based on identiffing critical sections which will sustain inelastic deformation
during a seismic attack. The strength of the critical sections are typically higher than the nominal capacity derived directly from codes and is termed the flexural overstrength. The rest of the
sfucture is subsequently designed to withstand the forces arising in conjunction with development of the maximum credible strength (oversfiength) at the critical locations.
Serviceability Limit State
The flexural serviceability limit state for prestressed masonry is concerned with flexural
strength, stiftress and deflections. The following flexural states represent the limiting flexural
moments for a wall to remain elastic for un-cracked and cracked sections'
First Cracking: This limit state corresponds to the state when the extreme fibre of the wall
decompresses (the tensile strength of concrete masonry is disregarded).
Maximum Serviceability momenf..At this cracked section state, the compressive stress in the
extreme compression fibre has reached its elastic limit set out by the code as a sffess limitation.
Reinforcement and concrete masonry remain elastic in this state.
Ultimate Limit State:
The flexural ultimate limit state for prestressed masonry is primarily concemed with flexural
strength. Additionally for ductility purposes, overstrength, stiffitess and deflections should be
considered:
Nominal strength: The nominal strength according to NZS 4230;1990 [8-2J is per definition
(unconachieved when the concrete masonry strain in the extreme fibre, e*u, equals 0.0025
fined concrete masonry). Notable, this value of
rn.,u
may be altered in the future to provide
greater alignment with other national masonry design codes (see Chapter 3)'
Overstrengf1: This strength corresponds to the maximum moment strength developed by the
wall, taking into account shess increase and yielding of the prestressing tendons. At this stage,
large deformations are expected and the maximum concrete masonry strain is likely to have
surpassed 0.0025. past the maximum wall strength, the wall resistance gradually degrades
until
failure.
The applied forces and loads represented by the symbols V*, M*,
N' and P'used in the follow-
ing equations are all factored loads calculated according to the applicable limit state as defined
-r93-
,dl
7
t/
\
\
t\
thickness
bw
t[,
Moment
Cu
rvoiu re
V
I
Def
lection
Fig. 8.I-Definition of cantilever wall dimensions, forces and deformations

in the New Zealand loading standard NZS 4203:1992.The axial force
N'is
due to dead and
live loads, P' is the prestressing force (initial force after anchor lock-offor force after all long
term losses), and
V'
is the applied lateral load due to lateral actions.
It is assumed that moment
M. only arises from lateral forces V', i.e. pennanent loads and prestressing do not introduce
permanent moment in the wall. It is also assumed that the prestressing tendons are placed sym-
metrically in the wall with respect to the wall centre line such that the resulting total prestress-
ing force is applied at the wall centre line. Fig 8.1 shows the various definitions of wall
dimensions and forces.
8.1.2 First cracking

The moment corresponding to first cracking N4, may be evaluated by Eqn. 8.1. The formula is
based on the flexural state at which one
wall end decompresses and the other end compresses
to a stress of nvice the average masonry stress
=M
6
t,
where
as shown
f*:
in Fig. 8.2.
(P+ /\r)
l*b.
(8.1)
(8.2)
where b* is the wall thickness, 1,, is the wall length, V., is the applied force at the top of the
wall corresponding to the l't cracking moment M", and tq is the effective wall height (refer to
Chapter 9 for definition of h.).
-r94-
o)
Mosonry woll
b) Stress dislribulion
Fig. 8.2-First cracking
The curvature at I't cracking can be calculated as follows:
r _ 2(P+1\r) _
Ycr)Dl
E^l;b*
2f^
(8.3)
Lm'w
The lateral displacement of the top of the wall d", at V., should be based on the concrete
masoffy wall elastic properties (see Chapter 3) and consists of a component due to shear defor-
in
mation 4*n and a component due to flexure d..n. Integration of the wall curvature is applied
order to evaluate the lateral displacement as shown schematically in Fig' 8'1'
d,, : d,,lt* d",,h = 3
+
* 2(r +:)!r ro
i4:;!
rpjU* 5
(8.4)
E^b*
where poisson's ratio typically is taken as v
noted that the shear deformation component
d",r1,
0.2 for concrete materials t8-31-
It
should be
can be of significant magnitude for squat
walls under serviceability loads, whereas for the ultimate limit state it becomes increasingly
insignificant.
8.1.3 Maximum serviceability moment

At the serviceability limit
state (MSM), the applied lateral force has surpassed that necessary
to initiate cracking at the base of the wall. The serviceability moment is limited by M" which
occgrs when the stress in the extreme compression fibre at the base of the wall has reached
-t95-
o)
Mosonry woll
b) Slress distribution ond crock

Fig.
kf*
Lengfh
8.}-Maximum serviceability moment
as shown in Fig. 8.3, where
f,n is concrete masonry crushing stength. For
prestressed
concrete, k (symbol adopted in this paper) is set out by NZS 3101:1995 [8-3], Table 16.l and
typically ranges between 0.45 and 0.55, dependent on load category. These values may also be
assumed for prestressed concrete masonry.
It is noted that Eqn. 8.5 must be satisfied before
use of the equations relating to the maximum
serviceability moment. This requirement is generally fulfilled.
kf;
>2f^
(8.5)
It may be shown that the maximum serviceability moment can be calculated from force equilibrium as follows, adopting
M"
k:
0.55 from load category tV (infrequent transient loads):
:+(t -#),ir- : r^(os -r ufi)tlb* :
rt"h"
(8.6)
where V" is the corresponding lateral force.
M,:
(t-#)r",
(8.7)
Eqn. 8.7 relates the MSM moment to the first cracking moment. The masonry is assumed to
remain linear elastic, hence the extreme masomy strain
from:
196 -
me
corresponding to kf',n can be found
Mo me
nt
Cu
rvotu re
Fig. 8,4-Curvature distribution st maximum serviceability moment
kf;
!::-
-me
(8.8)
"m
The corresponding curvature at the wall base, $", is defined by:
A" =
where c" =
z(N+P) _2f^lo _
kf;
b*
kf;
3.6fnlw
(8.e)
JM
q"=#,r-:o-^h
(8.10)
Fig. g.4 shows the variation of moment and curvature along the height of the wall at the malcimum serviceability moment, assuming plane section response. The curr'"ature varies from $u at
the base to Q", at the height, h.o at which the lst cracking occurs, Between the heights
the curvature varies linearly between 0.,
md zero. Fig. 8.3 and Eqn.
8.1
h*
and
I define the nondimen-
sional crack length at the base of the wall at the maximum serviceabilit5r moment, again
assuming
k:0.55:
T":r'#=l-t*
(8.11)
and Eqn. 8.12 defines the resulting cracked wall height.
h",:
o"W)= r"(r
)
-t97 -
(8.12)
The total displacement d" of the top of the wall due to flexure can then be calculated with:
d"
: d"It* d"rh
(8.13)
The displacement 4n of the top of the wall due to flexure can then be calculated by double
integration of curvature along the wall height (see Fig. 8.1) with the following result:
, :
"rI
2f^h"
#l*
"
?(h,. r,,lr -rJ)] . Y,o" - h",)'
")(+").
which may be approximated with, assuming
(8.14)
k:0.55:
d"rt:(o.m -o.ozs*Y#,
(8. l 5)
The shear deformation is defined by:
d"rh:'ffi'"
(8.16)
At this flexural state, it is assumed that the relatively small deformations of the wall do not
result in significant tendon force increase or migration of the tendon force eccentricity.
8.1.4 Nominal flexural strength

At this ultimate limit state, an equivalent rectangular stess block is assumed with a stress of
0.85fn
(o:
0.85) and an extreme fibre strain of e,ou: 0.0025, corresponding to the definition
of nominal strength in NZS
4230:1990
for unconfined concrete masonry. For confined
masonry NZS 4230:1990 recommends using an average stess of

based on unconfined prism strength) and e-u
0.9Kf. (a : 0.9K with
f,
:0.008. K is defined in Eqn. 3.15. The corre-
sponding moment Mn and lateral force Vl can be evaluated by simple equilibrium, as shown in
Fig. 8.5, with the following equation:
Mn
: (P +
^P)(?
",-;). -(+ -|)
: ,ro"
(8.17)
where a is the length of the equivalent ultimate compression block given by:
P+AP+N
a:ffi
(8.1 8)
-198-
P+AP
o
c
o
to
o
o
o)
c
o
o
-9
E
o
ot
:(t
o
o
!E
F
=
df i"
N+P+AP
Fig. 8.5-Wall equilibrium at nominal flexural strength
In these equations, Ap accounts for the increase in tendon force that arises from the flexural
deformation and e, accounts for the associated tendon force eccentricity. Both AP and er may
initially be assumed to equal zero for simple use. This approach is similar to the method used
in NZS 3101:1995. A better estimate of the nominal strength may be obtained from Eqn. 8.17
when taking into account the tendon force increase AP and the associated tendon force eccen-
tricity
e.
(further on that issue below).
It is observed from Fig. 8.5 that
there is moment reversal near the top of the wall due to e,
which results in reversal of curvature. This effect is not taken into account when calculating
wall deformations in the following because it has a negligible effect on the predicted wall
behaviour at nominal flexural strength.
The total lateral displacement d,, is given by the sum of the flexural displacement, doq, and
shear displacemeot, 4rr,, corresponding to Mo, and may be evaluated using Eqn' 8'19:
d, : dnfl* drrh
unconfine
Confined:
d: dn,
rlt
(8.1e)
where
Q3oEl-
1.38,,
+0.856#
: Q.$e1- 5.40q,, * r.unff

-199-
(8.20)
(8.21)
u.
I
I
I
ormed
shope
def
Fig.8.fWail
I2(l + v)h"
d rrh
>n
a:-
deformation at nominal flexural strength
5Emlwb*
(8.22)
v1.
P+AP+N
(8.23)
f;l*b*
Eqns. 8.20 and 8.21 were developed using numerical integration and curve fitting, and are thus
of an approximate nature, and are valid for axial load ratios,
L, of 0.05 to 0.025. The extreme
fibre strain was taken 8s enu : 0.003 for unconfined concrete masonry (consistent with interna-
tional codes, but not NZS 4230) and 0.008 for confined concrete masonry. Detailed information on derivation of these equations may be found in Appendix A.
The total tendon force increase AP at tmu of 0.003 (or 0.008) is difftcult to evaluate for prestressed walls
with unbonded tendons because the tendon stress increase depends on the defor-
mation of the entire wall between points of anchorage. However, the force increase (or
decrease) in each tendon in the wall cross section may be evaluated based on the estimated wall
end elongation, u", (tension end) and shortening (compression end), us, assuming a linear vari-
ation of vertical deformation across the wall top as shown in Fig. 8.6. The following equations
were established for unconfrned and confined concrete masonry (refer to Appendix A):
-200-
Unconfined
I u"
(4.u81-2.37\n+
0.$r*
(8.24)
: (3.36qf,-z.tz1^- o.oofft:"
u,
confined: u" : (22.51f,-rc.+1^*
r.rr*
(8.2s)
u,
r.67E1-r.64e,-O.r+21&
"m
In these equations, elongation is positive and shortening is negative. It is clear that the tendon
force increase due to vertical deformation will increase the axial load ratio. Iteration using
Eqns. g.24 or g.25 is therefore needed to find AP
!lr,
such that the calculated axial force
ratio at nominal flexural strength, (n, injected in the equations on the right hand side in fact
corresponds to the calculated tendon force increase on the left hand side of the equations.
The effective total tendon force eccentricity relative to the wall centre line can be evaluated by:
t=
Ztr,+
a,P)ri
Ltri+
p1 and
where L,P,
L'P)
U;
(8.26)
tno,t
Api are the initial tendon force and tendon force increase of the
horizontal location of the
j'th
j'th
tendon, and yj is the
tendon with respect to the wall centre line taken as positive
towards the tension end of the wall. The tendon vertical extensiorL q, is defined in Fig. 8.6 and
is the tendon length (approximately the height of the wall
than h. for multi-storey building). Ap.; is the area of the
h* which is significantly
longer
j'th tendon and Eo, is the elastic modu-
lus of the prestressing steel. It must be ensured that P1+AP; does not exceed the tendon yield
strength.
Iteration process for calculation of Mo

1. calculate (n using Eqn. 8.23 using
3n1!-gl,,;
AP:0-
2. calculate u" and u, using Eqns. 8.24 or 8-25.

3. calculate AP:
!ar,
using Eqn. 8.26-
4. calculate (n using Eqn.

5. repeat
steps (2) to
8.23 using AP from (3).
(4)until convergence ofe".
-20t -
6. calculate Mn using Eqn. 8.17 and
do
using Eqn. 8.19.
The masonry design codes BS 5628:1995 [8-4] and AS 3700:1998 [8-5] present formulae for
calculating the tendon stress increase, but are not applicable for in-plane wall bending because
they were developed for out-of-plane response. NZS
3 I 01 : I
995 recognises that the design ten-
don force for unbonded tendons will exceed the tendon force following losses. Using the nota-
tion presented here, the increase in tendon force is given by:
L,P : Ao,
(8.27)
where
fr": t, ,
fpr<.fpy and forlfru* 400MPa
(8.28)
where Ao, is the total prestressing tendon area, S, is the resulting average tendon stress conesponding to P*AP,
$,
is the tendon yield stress, defined in section 2, ard,
f"
is the tendon stress
corresponding to P. This equation seems to provide reasonable results but has not been validated for all wall configurations.
Yz-3A
It would be prudent to assume a total tendon force increase of
times the result calculated by Eqn. 8.27 when the presfessing tendons are approximately
evenly distributed along the length of the wall. Eqn. 8.29 evaluates the resulting tendon eccen-
tricity, e,, due to the total tendon force increase AP acting at an eccentricity of lo./6, assuming
that tendons of equal area are evenly distributed across the wall.
o:-
lwLP
(8.2e)
6(P + LP\
Having calculated AP and e,, the nominal flexural strength, Mn, and corresponding displacement,
d.,.,,
can then be evaluated using Eqns. 8.17 and 8.19.
8.1.5 Yield strength

Contrary to reinforced concrete walls, the yield strength for unbonded prestressed walls is typ-
ically found at displacements beyond the displacement at nominal flexural strength. Structural
testing has consistently shown that the behaviour of unbonded prestressed walls loaded beyond
the nominal strength is dominated by rocking as illustrated in Fig. 8.7. Even for walls without
specially placed confinement plates, experimental observations consistently demonstrate that
the wall is able to support compression strains far beyond 0.003. In Fig. 8.7, the wall has
rocked over by a displacement
d*
corresponding to a rotation 0. At this state,
-202 -
it is assumed
Fig. 8.7-Wall rocking response

that the extreme tendon at the tension side of the wall yields, resulting in a tendon strain
increase of:
a^ - Vor-fpr)
u.py_
(8.30)
where Eo, is the modulus of elasticity for the tendon steel as defined in Chapter 3 and fn, is
latertaken as the tendon stress in the extreme tendon at nominal strength. If a wall is displaced
ally beyond
d*
some reduction of prestress should be anticipated upon unloading. Notably'
this does not mean that wall strength is permanently reduced because the tendons can be fully
activated by subsequent load"ing excursions. The wall rotation 0 can be related to the wall displacement increase at first tendon yield do and the tendon shain increase Aeo, in the following
way:
t"-
thus
d,r:0h"=#:W t"h?
(8.31)
pc, and it is assumed F : 0.95 for unconfined masonry and B = 0.96 for confined
masonry tg-2]. ln this equation, eo, is the eccentricity of the extreme tendon at the wall tension
side with respect to the compressive end of the wall. The length of the compression zone, c, is
where a
calculated at the nominal sfiengthbased on Eqns. 8.18, thus assuming that the wall rocks about
an axis at the distance, c, from the extreme compression fibre in the wall. As do is considered
the displacement increment beyond do, the stress state in the exfreme tendon should rigorously
-203 -
be taken as fo,, however using
f,. (initial tendon stress in unloaded state) instead of
{0,
in Egn.
8.31 generally result in little enor.
Given 0, the force increase in the individual tendons can be calculated as:
M,rj:ryEp,Ap,j=
LP,
(fpy-fpr)
,t-
Aorih
(8.32)
-te
: }LP,ri
(8.33)
where e,i is the location of the j'th tendon with respect to the compression end of the wall, Apsl
is the area of the j'th tendon and AP, is the total tendon force increase above that at Mn. Note
that Eqn. 8.32 assumes linear variation of the tendon force increase with respect to the lateral
location of the tendons. The resulting moment increase M* is then given by:
Mty
Z * *,(" u-7)
j:l
: Z LP,ri",i -7or,
(8.34)
j=r
where n is the total number of tendons along the length of the wall and the compression zone
length at frst tendon yield may be calculated as:
P+LPu+N
or:W
Finally the yield moment M,
M,
du
(8.3s)
*d
displacement d" can be evaluated as:
: (N+ P . *r(+ -?*
Mry
: vrh,
: dr+ dr,
(8.36)
(8.37)
8.1.6 Flexural overstrength

The maximum credible strength of an unbonded prestressed wall may be evaluated by assum-
ing that all tendons have reached their yield strength. Consequently, the flexural overstrength,
Mo, maybe evaluated as:
Mo =
w+P)(+-?:
voh,
(8.38)
where ao is the length of the equivalent ultimate compression block and P" is the total tendon
force when all tendons are yielding given by:
-204 -
o, =
ffi*
and
Pr:
(8.3e)
Ai,,fpy
At this state, it is assumed that the tendon closest to the flexural compression zone has reached
its yield stress. The resulting displacement can then be evaluated using the following equation
which is similar to Eqn. 8.31:
) _ ) -for-f^. o: ,=
do:dn_fiffi
In this equation
e," is the distance
(s.40)
from the compression end of the wall to the closest tendon
and fp, is the tendon stress in the same tendon at nominal strength.
It is noted that the Eqn. 8.40 is not appropriate if the closest tendon is located within the flexural compression zone, i.e. e,. ( c, and that if the tendon closest to the compression zone is near
to the location of the flexural neutral axis, unrealistically large values of
do are calculated.
When all tendons are located near the wall cenfreline, the wall yield strength coincides with the
wall overstrength. It can be argued for conservatism that the tendon yield stress, $r, in Eqn.
g.39 should be replaced with the tendon ultimate strength, f,u,
it
order to establish the maxi-
mum credible wall flexural strength. It is, however, unnecessary to modify Eqn. 8.40 accordingly because the tendon strain at ultimate strength is of the order of 5Yo and therefore not
attainable in reality for walls of any geometry. Use of fou is only necessary if the prestressing is
unbonded over a height significantly shorter than the wall height.
8.1.7 Ultimate displacement capacity

The ultimate displacement is limited by the strain capacrty of the tendons as well as the crush-
ing strain of the masonry. Generally, the tendon ultimate strain is of the order of
5Yo,
which
would require unrealistically high lateral displacement to govern. As a result concrete masonry
compression failure is expected. Confinement by the foundation is likely to increase the failure
masonry strain beyond 0.003. As the extreme concrete masonry fibres fail, there is a tendency
for the compression zone to migrate towards the centre of the wall, reducing the wall strength
gradually. Experiments conducted at the University of Auckland and presented in Chapters 5,6
and 7 have shown drift ratio capacities
of l% - 2% for prestressed concrete masonry walls of
various aspect ratios, suggesting high displacement capacity. It is noted thatthis limit state may
occur before tendon yielding, depending on the wall aspect ratio, the prestressing steel area and
the initial tendon stress
f".
-205
du
l/
"ll^,wr,r--1
ploslic
deformofion
he
Fig. 8.8-Vertical strain evaluation at ultimate displacement capacity
The drift ratio or the drift angle is in this document defined as the ultimate displacement
d'.,
divided by the effective height, and is thus termed 'drift capacity':
tu -d"he
^t
(8.41)
Evaluation of the extreme masonry strain at displacements beyond nominal flexural strength
necessitates definition of a plastic hinging zone at the bottom of the wall. Assuming that all lat-
eral displacement at the top of the wall is due to rotation, 0, of the plastic hinge as shown in
Fig. 8.8, the masonry extreme fibre strain, tmu, can be related to the wall lateral displacement,
d":
rh-l'e
d,: \lh"-t)
and g : Qho: #oo
(8.42)
,du : oo(0"-+)
. , : a P+^P+N: \ut*
Eû wnere
1_
p: WF
(8.43)
Ep__
tn this equation, AP should correspond to the actual tendon stress state at the displacement
For consistency with the calculations at Mo, My and Mo, it is assumed c[ : 0.85 and B
unconfined concrete masoffy (U200) and
cr: 0.9K and B:0.96
dr,.
: 0.85 for
for confined masonry (CP200
an CP100). Refer to Chapter 3 for definition of U200, CP200 and CP100.
It is accentuated that
Eqn. 8.43 is of idealised nature and simply attempts to relate the lateral displacement to the
masonry strain state in the compression toe region at the wall state where initiation of strength
degradation due to masonry crushing is anticipated to commence. Eqn. 8.42 assumes that the
-206-
TABLE 8.I-Basic
Concrete
Masonry
deformation zone
v max
t'-u
0.8vmax
meters
Vr.,
and 0.8vm'*
t'to
h/hu
u200**
cP200*"
0.016
0.020*
0.071'i
0.016'
0.020
0.071|
cPl00**
0.00E
0.013
0.076
sfrain
strain
gqn. 8.a3, * based on uisual inspection **refer to Chi rpter 3

ng Eqn
usln-lne
total rotation 0 occurs at a height of h/2 above the wall base. This is in line with the crurent
thinking for plastic hinge zone rotation for reinforced concrete masoffy walls [8-6]- For evaluation of
{, it is acceptable to interpolate
between the axial forces calculated at nominal flexu-
as
ral shength, first tendon yield and overstrength relative to the displacements do, { -d 4,
based on
applicable (with a maximum of N+Pr). The base shear corresponding to d,., can be
Eqn. 8.17 using the appropriate axial force or on interpolation between V6, V, and
V' with
maximum of V..
is parallel
The approach used in this document for calculating maximum displacement capacity
to
to that used for reinforced concrete masonry (RCM). Eqn. 8.43 relates the total displacement
the
the total sfain in the plastic deformation zone, and disregards any elastic deformation of
wall above the plastic deformation zone. For RCM, the displacement due to plastic hingtng
du is calconstitutes a displacement increment dp due to a plastic strain increment of &np, where
the wall disculated as d, + do and o as &nu - s-v t8-61. The quantities d, and Qnv represent
placement and extreme fibre strain at fust yielding of the extreme mild steel reinforcing barthe limit
The reason for using the simple approximate relationship defined by Eqn. 8.43 is that
and that ernu for
state defined by q and env does not exist for PCM (in the conventional sense)
pCM proves to be very large relative to typical values of g (normally elastic concrete masonry
strain) for RCM.
far
The ultimate masonry strain capaciry 4nu, is associated with flexural action and exceeds by
the unaxial strain capacity determined by prism testing (refer to Chapter 3). Experimentally
of
determined values of ,n.,o and ho/tt are given in Table 8.1 and were based on a compilation
results from the structural testing
of
unbonded post-tensioned concrete masonry walls
described in Chapters 5 through 7. Given the small number of tests and unknown repeatability
of these, it would be prudent to reduce the strain values given in Thble 8.1 for design. The
-207 -
design value of the ultimate strain capacity should be taken tts enu
1,3
to
: t'*u/y, where yis taken as
1.5.
Normalisation of
indicated that
ho
with t5 was based on comparison of results from the three test series which
was likely to be related to the effective wall height, h", rather than the wall
length, 1,. The values of
for U200 and CP200 (wall test Series I and 2) were based on visual
inspection of the wall toe regions when strength degradation of

strain (e* > 0.0015) in the masonry above the
of the Series
0.8Vr*
had occurred. Inelastic
frst course (200 mm) was not measured in any
I and 2 U200 and CP200 tests, thus \
200 mm was assumed. Series
and 2
testing results do not confirm that U200 or CP200 concrete masonry are capable of developing
inelastic strain above the first course, therefore the results for U200 and CP200 should for
taller walls be extrapolated with caution.
The plastic deformation zone lenglh for CP100 masonry was based on the results from testing
Series 3, which reflected realistic wall dimensions in multi-storey buildings. Extensive and
reliable measurement readings revealed that inelastic strain (e* > 0.0015) occurred over
height of 400-600 mm at both V,n* and 0.8V.*. This confirms that CPl00 is capable of developing inelastic strain over a height of four or more courses of concrete masoffy. The longer
for the 5.25 m high Series 3 walls also explains why the average strain (theoretical) in ttre plas-
tic deformation zone for Vn,"*, given in Table 8.1, for CPl00 is lower than that of CP200. The
ho
value given in Table 8.1 for CPl00 was based on actual measured values, while ho values
given in Table 8.1 for CP200 were based on visual inspection. For design pu{poses
\ftr"
for
CP100 may be taken as 0.08 which conforms well with that typically used for reinforced concrete walls [8-6].
It may
seem counter-intuitive that the values for '-u given
in Table 8.1 for U200 and CP200
are higher than those for CPl00. However as suggested above, the results for U200 and CP200
walls cannot directly be compared with the results for CP100 walls because the modular nature
of the masonry blocks determined that ho:200 mm for the squat walls U200 and CP200.
It was found by strucfural testing when using CP200 and CPl00, that the wall strength persisted at levels close to V.* for displacements beyond d.,** (corresponding to V.*), until displacements near do.su'nu* (corresponding to 0.8V.u*) were reached. It can therefore be argued
that dependable wall displacement capacity, d", with little strength degradation may be calcu-
lated using
n,u
related to 0.8Vr"*. This is generally not the case for U200, for which testing
-208-
proved that significant strength degradation was imminent after d.,n,*. Measurements relating
to 0.gVn'o* can be used for U200
if it is assumed
that reliable wall performance persists until
the strenglh had degraded to 0.8V'u*.
It
appears from Table 8.1 that the plastic zone heights are of the order of 7-8o/o of the wall
height, or 200 mm to 400 mm for the tested wall configurations. This is a short length when
compared with typical plastic hinge lengths for reinforced concrete masonry walls of the order
of 0.5x1* [8-6], or 900 mm to 1,500 mm when applied to the experimental wall lengths.
Embedding of confining plates can consequently be limited to a small region in each end of the
wall, say over a vertical distance of 1.5 to 2 times
ho and a
horizontal extent of 1-5 times
(l*
Vertical sfrain measurements from testing Series | , 2 and3, as discussed in Chapters 5, 6 andT
confirm the magnitude of strain and plastic zone length proposed in Table 8.1. Nevertheless, it
is emphasised that the results in Table 8.1 are of preliminary nature and should be used with
caution due to considerable scatter of results, and as stated above, that these values are of theo-
retical and empirical character. Additionally, the proposed ultimate strains should not be
regarded as the reliable ultimate uniaxial strain capacity of concrete masomy.
8.1.8 Calculation examPle

Consider the wall shown in Fig. 8.9(a). It is assumed that the five storey wall is 15 m high, 3-6
m long, 190 mm thick and prestressed with five 15.2 mm high strength prestressing strands
(4,: : 140 mm). Half height concrete masonry blocks from the 20 series (100 mm high) are
used in the plastic deformation zone; regular 20 series blocks elsewhere. The
wall self weight
is calculated to be 225 kN and the dead load of the floors and roof amounts to 0.5 MPa at the
base of the wall, resulting in a total dead load of 567
kN'
The calculation is performed on the equivalent structure shown in Fig. 8.9O) with an effective
height, lL =
2/3xh*:
10 m (refer to Chapter 9 for further considerations on h"), at which the
wall lateral load is assumed to be applied. The tendons are placed symmetrically about the wall
(the five
centre line at zero, L200 mm and +400 mm eccentricities from the wall centre line
stands are represented with one line in Fig. S.9). ln the calculation, the tendon elastic elongation capacity is based on the actual tendon length, approximated with h* using the effective
tendon elastic modulus of Eorxh/ho An initial tendon stress of 0.67fou is selected, based on an
estimated first tendon yield at a lateral drift of about 1.57o assuming that the wall rocks as a
-209 -
tl
r;
o
ll
tl
oE
o
ll
o)
b) Equivolonf sfructure
Proiolype
Fig. 8.9-Post-tensioned concrete masonry cantilever wall
TABLE 8.2-Predicted force and displacement

First
Maximum
Nominal
First ten-
Ultimate
cracking
serviceability moment
strength
don yield
displacement
capacity
83.9
182
227
248
249
253
kN
2.9
10.8
4t.2
158
t92
310
mm
AP
34
I40
160
231
KN
Well overstrength
rigid body around the lower corners. A total prestressing force of 831 kN is found, resulting in
an
initial axial load ratio of
E:
0.114.
Confinement plates are imagined embedded in the horizontal bed joints in the wall corners by
the base over a height of
2"4 :2x0.076x10
1.5
m and K
1.08 is assumed (refer to
Chapter 3). The confinement plate length is taken as 2*l* or about 800 mm- It is assumed that
the height of the plastic hinge zone is 0.076xh.
0.76 m and the ultimate flexural strain is
0.013, as recommended in Table 8.1.

Table 8.2 and Fig. 8.10 present the predicted wall in-plane response with the base shear
Y lat-
eral displacement d and tendon force increase AP related to the equivalent structure shown in
Fig 8.9(b). Material properties and wall dimensions are specified in Fig. 8.10. Specific details
on the calculation example may be found in Appendix B. It is seen in Fig. 8.10 that wall sof-
-2r0-
First
Overstrength
tendon Yield
hw=l5m
ho=l0m
lw=3.6m
b* - 0.19 m
Nominal strength
f.
ultimate
z,'
Y
o
E.
displacement
Maximum
seMceability moment
capacity
Driftratioy=
rso
1.94%
o
6
Aro = 700 mmz

Epo = 190 GPa
fpv = 1517 MPa
N=567kN
P = 83'l kN
lt
E'0114
cl
= 18 MPa
= 80of'm
fp
= 1187 MPa
fm = 2.0,1 MPa
100
First
cracking
rh = 0.76 m
sru = 0'013
K = 1.08
cr=0.9K=0.972
F=0e6
+Prediction
50
100
ultimate Dlsplacemnt
(,nt)
o'to'uluto-e'r
200
250
300
Fig. 8.l0-Predicted in-plane wall response
tening initiates between the maximum serviceability moment and the nominal strength limit
length at the wall
states. Between the two formerly mentioned limit states, the theoretical crack
is reached 34
base increased from 0.60x1* to 0.88x1* The wall ultimate displacement capacity
to a
mm after the first tendon yield limit state. It is expected that the tendons remain elastic up
drift of l.5g% (displacement of 158 mm) which is slightly higher than the initially considered
drift of 1.5% before tendon yield. The displacement at wall overstrength is, in this case, only of
theoretical interest. For this particular wall, the ratio of overstrength to nominal strength
amounted to
l.l l.
8.2 DESIGN CONSIDERATIONS

8.2.1 Required prestressing force and area
the ServThe required prestressing force, B may be determined by the applied actions at either
iceability Limit state (sLS) or at the Ultimate Limit state (uLS). For prestressed concrete with
bonded tendons,
it is normally SLS that governs
the required prestressing force, however for
prestressed concrete masonry with unbonded tendons,
it
is not clear whether SLS or ULS loads
govern. This is largely because the wall ultimate strength (nominal flexural strength) does not
benefit from significant tendon stress increase due to flexural deformation. It appears for PCM
walls with unbonded tendons that the dependable nominal strength QsM,, rarely is more than
20o/olargerthan the maximum allowable elastic service moment, M., given by Eqn. 8.6-
-ztt
Supposing that for wind actions the ultimate flexural load Mu* is roughly equal to 1.4 times the
service load M,*, it is found that the ULS requirement is likely to govem the design axial prestress. For seismic loads determined for limited ductility demands, the strength requirement at
ULS again appears to govern the design axial prestress. Current studies indicate that the ULS
requirement given by Eqn. 8.44 generally is satisfied, if the required design axial prestress has
been determined based on un-cracked section analysis at SLS in accordance
Qr
Mr'-
with Eqn. 8.45.

(8.44)
M,,
*
(8.4s)
M"r2 M,
Prior to determination of the required prestressing force, P, it needs to be decided whether the
wall must remain un-cracked at SLS or not. In that regard it is reiterated, that only a single
crack at the base is expected to open when the
l't cracking moment
is exceeded at the base, and
that this crack will close immediately upon unloading of the wall, unlike the behaviour for con-
ventionally reinforced concrete consffuction. Consequently PCM wall cracking at SLS loads is
not a durability issue but rather an issue related to function of the structure. Seen on the basis
of the expected crack paffern, it could be argued to require that PCM walls remain un-cracked
for all SLS load combinations, except for earthquake load combinations.
Altematively, the required prestressing force could be determined by shear strength considerations at ULS.
When the required prestressing force P has been determined, the prestressing steel amount is
calculated, assuming that P needs to be present after all losses (refer to Chapter 4), and at the
same time respecting the allowable prestressing steel stress before and immediately after prestress transfer.
The anchorage region, to which the concentrated prestressing load is applied, shall be designed
accordingtoNZS3l0l:1995,section 16.3.9,inordertoavoidbursting,splittingandspalling.
It is conceivable that the prestressing force needs to be applied in two stages, depending on the
early strength of the masomy and the magnitude of the anticipated prestressing loss. Initially, a
lower prestressing force could be applied to resist construction loads. The final prestressing
force is then applied at a later date.
-2r2-
ô
t] '
"E
0.04
o.o3
,:
E o.oz
[" fraction of
Fig. E.l
l-Cracked
f'
section analysis
8.2.2 Un-cracked section analysis

prestressing
To ensure un-cracked response at SLS Eqn 8.45 needs to be satisfied. The required
force is then given by:
6M^ *
P>+-N"
(8.46)
t yt,
8.23
Cracked section analYsis
(see
ln this case, the serviceability moment may be equated with the morimum elastic moment
Eqn. 8.6), resulting in the required average prestress on the wall being given by:
(") M:
ot
'*('-#),iu"
i.
Fl
-+)
frlu_=:('
where (c)
(8.47)
r,:v;
plotting the left hand side of Eqn. 8.47(b) for realistic values of ( between 0 and 0.4 and values
of k between 0.4 and 0.6, the curves shown in Fig. 8.11 materialise. Given the right hand side
of Eqn. 8.470) and the allowable value of k, ( may be deterrnined graphically. The average
axial concrete masonry stress
f,
may then be deterrrined by Eqn. 8.a7(c) and the required pre-
stress force, R may be determined bY:
-2t3
P>f*l*bw-
(8.48)
Ns
8.2.4 Ultimate limit state

8.2.4.1 Strength design
As the ultimate limit state flexural actions may govern the design presfressing force P, it must
be verified that the dependable flexural strength exceeds the ultimate flexural actions
as
defined by Eqn. 8.44, using Qs:0.85.
This can be done by evaluating the required prestressing force, P, which easily can be derived
by trial-and-error from the equations presented in section
Ll.4 by initially
assuming that AP
0. First approximation of the required total prestressing steel area Ap, may then be calculated
with the assumption that the effective presftessing steel stress after all losses, f,., amounts to
about 0.50fp" (refer to Chapter 4).
closer approximation can then be obtained when AP
(Eqns. 8.24 or 8.25) is included in the calculation of the nominal strength I\d. Depending on
the required accuracy, P can be determined after several iterations.
8.2.4.2 Capacity design

Calculation of the wall overstrength, Mo, is of considerable significance for ductile seismic
design. The principle of capacity design states that the forces generated throughout the structure, caused by development of the maximum strength of the ductile yielding components cre-
ating a mechanismn must be exceeded by the dependable capacity
of all
other joining
components. This will ensure, in the event of a major earthquake, that yielding will occur at the
specially selected locations (plastic hinges) and that all other types of potential failure modes
are eliminated. This means for a cantilever wall, that the
wall shear strength, Vi, must exceed
the base shear, Vo, that develops the maximum flexural strength of the wall (overstrength Mo)
and that the wall flexural strength outside the plastic hinge must exceed the wall moment distri-
bution corresponding to development of the wall overstrength, Mo. Additionally, the moment
distribution (demand) arising from development of flexural overstrength may have to be ampli-
fied to take into account higher mode effects. Refer to Paulay and Priestley [8-6] for further
discussion.
-2r4-
8.2.4.3 Displacement caPacitY

yield
For ductile seismic design, it is necessary to ensure that the nominal flexural sfrength and
strength are suffrciently large to limit excessive displacements (storey drift) during seismic
attack. This concept
is similar to reducing the ductility
demand
for reinforced
concrete
masonry by increasing the flexural strength (yietd srength). A prestressing steel area increase
is needed if the wall strength is insufficient to ensure reasonable limitation to the lateral displacements.
It is desirable to ensure
that the prestressing steel in an unbonded post-tensioned concrete
masoffy wall responds elastically to seismic excitation for several reasons: the wall retains (l)
the initial stiftress, (2) the strength, (3) the self centring capability, and (a) restressing of the
tendons after the seismic event is not necessary. A low initial tendon shess relative to the tendon yield strength allows for larger wall displacements before first tendon yield at the displace-
ment of
d"
Conversely, a relatively high initial tendon stress is desirable because
it
reduces
time dependent prestress loss and reduces the required prestressing steel area (better economy). Ideally, the seismic displacement demand, dE*, should be exceeded by both the tendon
*d
yield displacement,4
the displacement capacity, du'
The displacement demand imposed on an unbonded wall stnrcture in a seismic event, ds*, cannot be predicted using existing loadings codes and acceleration spectra. However, using
numerical non-linear time-history analysis, the displacement demand can be estimated as

described in Chapter 9.
8.2.5 Base sliding

preBase sliding response should be avoided, however such behaviour is not anticipated for
of normal wall aspect ratios and prestressing lwels.

Base sliding for pCM is conholled by the available sliding resistance, which is provided by
folshear friction by means of aggregate interlock mechanisms. A check at ultimate using the
stressed concrete masonry walls @CM)
lowing formula is recommended:

Vbo,"3 0p(P +
(8.4e)
/f)
where V6u," is the base shear due to lateral forces, p is the dependable coefficient of friction for
generally taken
concrete to concrete contact and 0 is the appropriate strength reduction factor
as 0.85, except when
V6*" is derived from capacity design considerations where 0: 1.0' The
-215 -
recorlmended value of the shear friction coefficient (Coulomb friction), p, depends on the concrete roughness amplitude (average distance between 'valley' and 'peak'), ru. The following
values are applicable to concrete [8-6] but may be used for concrete masonry as no values spe-
cifically for concrete masonry are available:
(a) p:1.4 for rn)5mm

(b) p: 1.0 for 2mm(ro35
(c) lL: 03 otherwise
(8.50)
mm
Eqns. 8.50 (a) and (b) both relate to concrete surfaces that have purposely been roughened, and
all equations (a, b and c) are based on concrete placed against a clean surface.
For concrete masonry walls, the foundation and above floor slabs (whether intentionally
roughened or not) must be clean before the first course of concrete blocks is laid. 'Wash-outs'
(clean-out ports) may be necessary for grouted construction to ensure that mortar dropping
inside the flues does not prohibit effective aggregate interlock between grout and the founda-
tion/floor slab. New
tains
7 mm
Zealand, standard grout
with specified strength of 17.5 MPa typically con-
to Collins [8-7], aggregate interlock increases with
aggregate. According
increasing aggregate size. Therefore the fine aggregate ofthe grout is unfavourable to larger
aggregate in concrete. Furthermore, it should be kept in mind that the largest particle type in
mortar is sand and that the mortar bed joints typically occupy 25-50% of the wall cross section.
It can be argued that the maximum friction factor of

achievable. The minimum value of F
1.4 given in Eqn. 8.50,
in reality, is not
0.7 was violated by some of the later tests in Series I
and2. This is of little concem because base sliding for these walls was highly affected by mul-
tiple use of the same foundation. The foundation top surface had after some tests effectively
been smoothened to a very small roughness amplitude that would not apply to real construc-
tion.
The propensity for sliding is highly dependent on the wall aspect ratio:
aspect ratio results in an increase of the base shear needed to develop the
tL/l* A
reduction of
wall flexwal strength.
A high aspect ratio is therefore preferred in terms of reducing the propensity for
base sliding.
The minimum aspect ratio permitted can be estimated in the following way. Consider the nom-
inal flexural strength of a cantilever wall assuming no tendon force increase:
M,: w.
n(?-;)
= vfh"
-2t6
where
e\ anoa? N+P
T ;:7ffi
The maximum base shear can be expressed by:
Vô":
(8.s 1)
l.33Vtau
where the 1.33 factor accounts for maximum conceivable wall strength (overstrength) and 0L
due to higher
the seismic dynamic amplification factor, accounting for base shear amplification
Mn
mode effects (see section 9.3.2.4). Substitution of the equation for a into the equation for
and solving for Vlgives:
vr:
ltw+
";;(
(8.52)
-h)
Injecting Eqn. g.52 into Eqn. 8.51 results in the following requirement for the wall aspect ratio:
L,
l*-
oz(r
l)'.
(8.53)
A suggested minimum aspect ratio can be estimated using 6:0.1, cr:0'96, Q: 2'5, 0: l'0
2.1
and lr :0.7 .Using these values in Eqn. 8.53, a minimum wall aspect ratio of approximately
is found. Assuming that the effective wall height amounts to
minimum absolute aspect ratio of
h*/|*:3.2
2/3
of the actual wall heigh!
is found. The value of o)o: 2.5 used above origi-
nated from section9.3.2.4. A more detailed calculation reflecting the actual wall dimensions,
overstrength, roughness amplitude and base shear amplification may
justiff a lower
aspect
ratio.
It should
be kept in mind that there is always an effective concrete masonry compressive force
yielding of the
at the base of the wall (in excess of IrI) due to flexure, even after initiation of
tendons. This differs from the typical response of conventionally reinforced walls cyclically
loaded beyond yield, where the wall effectively is supported laterally at the base by dowel
action alone. Finally, if significant degradation of the concrete masonry in the wall comers is
bars,
expected to occur, other means of inhibiting base sliding may be required, such as dowel
sliding blocking devices or recessing of the wall into the foundation.
-217 -
8.3 \IERIFICATION OF PREDICTION METIIOD

Three structural testing series were carried out in the Civil Test Hall at the University of Auckland. Test Series
I focused on single storey walls built with ordinary concrete masonry
(uncon=
fined) and is described in Chapter 5. Test Series 2, reported in Chapter 6, focused on single
storey walls with enhanced wall corners such that the masonry strength and vertical strain
capacity was increased, with the overall scope of increasing the wall displacement capacity.
Test Series 3 is described in Chapter 7 and deals with testing of two three-storey walls of real-
istic proportion. This testing series consisted of nvo walls.
A detailed comparison
between predictions based on the flexure/rocking model and experi-
mentally obtained results may be found in Appendix C. The following conclusions were
drawn:
Initial stiffiress:
For walls of low aspect ratio (say below 1.0), the predicted initial stiffiress based on V", and d",
overestimates the actual wall stiffness by a factor of the order of 1.5. For aspect ratios above 2,
accurate prediction is expected.
Global response:
Using the predicted base shear and displacement at maximum serviceability moment, at nominal strength, at first tendon yield and at overstrength, accurate and consistent estimation of the
force-displacement envelope (skeleton curve) can be achieved.
Displacement capacity
The predicted displacement capacity, du, was captured with reasonable accuracy using the values for exfreme fibre strain and plastic zone length suggested in Table 8.1. Based on structural
testing of 2.8 m high walls, enu : 0.020 and
\/h":
0.071 are adopted for unconfined and con-
fined masonry of 200 mm block height (U200 and CP200). Predictions for the walls using confined masonry of 100 mm block height (CPl00) were based results from testingof 5.25 m high
walls: .u
0.013 and hoftr" -- 0.076. The strength associated with do for confined concrete
masonry (CP200/CP100) should be based on interpolation of base shear between the nominal
strength, first yield strength and overstrength limit states, as a function of displacement, as
applicable. The maximum strength achieved at du should be limited by M". For unconfined
concrete masonry it appears prudent to calculate the wall capacity at d" using 0.8Vr.
-2t8-
8.4 REFERENCES
tg-ll NZS 4203: lgg2, Code of Practicefor General Stntcural
Destgn and Design Loadings
for Buildings, Standards Association of New 7naland,l992, Wellington'

Stntcturcs, Standards
tS-21 NZS 4230:1990, Code of Practice for the Design of Masonry
Association of New Zealand, I 990, Wellington.
t8-31 NZS 3101: Lggs, Concrete Stntctures Standard,
Standards Association of New Zealan4
1995, Wellinglon.
tg4l
BS 5628: lgg5, Part 2: Code of Practice
for
use of Masonry. Structural Use of Rein'
Institution, London.
forced and Prestressed Masonry, British Standards
Ausfialia, Homebustt"
t8-51 AS 3700-1998, Masonry Stntcturzs, Standards Association of
NSW Australia.
Reinforced Concrcte and Masonry
18-61 paulay, T., and Priestley, M.J.N., Seismic Design of
Buildings,John Wiley & Sons [nc.,I992,744p.
t8-7]
Collins, M.P. and Kuching, D., How Safe are Larye Lightly Reinforced Concrete Beams,
pp- 482Slabs and Footings?, ACI Structural Journal, July-Aug. 1999, Vol. 96, No. 4,
490.
-2t9 -
-220 -
Chapter 9
DYNAMIC ANALYSIS
9. INTRODUCTION
This chapter is concemed with non-linear time-history analysis of unbonded post-tensioned
concrete masonry (pCM) cantilever walls. Analyses were carried out in order to clarify the
relationship between wall stiftress and stength, and the wall displacement demand due to
dynamic seismic excitation. The effects of the PCM wall force-displacement characteristics
and structural damping were investigated for a selected building period range- Results from
multi-degree of freedom FEM modelling were compared with results from single degree of
freedom equivalent structure modelling. Pseudo displacement spectra
(Si
were developed for
design use, and a displacement focused design procedure is outlined.
Following research presented in the previous chapters, it can be argued that the hysteretic form
of pCM for in-plane static loading is well established. However, the dynamic response of such
walls, when subjected to seismic excitation, is uncertain and remains a controversial issue
associated
because many researchers are concerned that the inherent low energy dissipation
with unbonded tendons may lead to excessive displacement demand.
g.I
MIJLTI-DEGREE OF FREEDOM STRUCTURE
This section defines a prototype structure and outlines appropriate modelling using the finite
element method (FEM).
9.f
.1 Prototype
structures
Fig. 9.1(a) shows a typical wall suitable for a five-storey apartment/office building' The wall is
3.6 m long, 15 m high (3 m storey height) and 190 mm thick. It is assumed that the lateral
forces arising from ground shaking are distributed evenly between the walls carrying the lateral
force in a given direction, and that there is negligible coupling between these walls because
-221 -
of
tr
Elostlc
tl
lr)
oE
ll
'
r;
E
o
ol
*ll
tJ)
Tr
b)FEM model
c)MDOF
model
=0.354
sec
T=0.559 sec
d)1st mode e)Equivolonl
shop
SDOF model
Fig. 9.l-S-Storey PCM wall prototype and modelling

relatively flexible floor slabs. Thus the building response is represented by a wall analysed as a
free-standing cantilever The wall is post-tensioned with six 15.2 mm high strength prestress-
ing strands (elastic modulus: Ep, : 190 GPa, single tendon area:
strength:
fpu:
,\s
140 mm2, ultimate
1785 MPa) located in pairs at -800 mm, zero and 800 mm eccentricities as illus-
trated in Fig. 9.1(b). The initial prestressing steel stress is 0.6t", resulting in a total initial prestressing force of P
892 kN. The axial force due to wall self weight and factored dead and
live loads from the floors and roof is 5781d{. A grouted concrete masonry crushing strength
fn,n
18 MPa is assumed, based on typical values encountered in New Zealand. Confining
of
of
the compression toe regions is achieved by embedding steel confining plates in the horizontal
mortar joints, refer to Chapter 3. An initial elastic modulus of
E:
800f'.
14400 MPa is
assumed.
In order to investigate taller structures with longer fundamental period, an 8-storey PCM wall
with the following dimensions was also investigated. The wall is 5.0 m long,24 m high (3 m
storey height) and 240 mm thick, and post-tensioned with eight 15.2 mm high stength prestressing strands located in pairs at -1000 mm, -400 mm,400 mm and 1000 mm eccentricities.
The initial prestressing steel shess is 0.68f0,,, resulting in a total initial prestressing force of P
1360 kN. The axial force due to wall self-weight and factored dead and live loads from the
floors and roof is I 170 kN. Material properties of the prestressing steel and concrete masonry
were similar to those of the 5-storey wall.
-222
9.1.2 Finite element modelling

The prototype walls were modelled in DRAIN-2DX [9-1] using fibre elements, elastic beam
elements, and inelastic truss elements. Extensive evaluation of the DRAIN-2DX fibre element
(element 15) for modelling of unbonded post-tensioned concrete has previously been carried
out successfully 1g-Zl.Because of the similarity between concrete and grouted concrete
masonry material behaviour, such evaluation for PCM was deemed unnecessary.
The elements were defined according to the wall dimensions and materials. The vertical distri-
bution of the fibre and beam elements is shown schematically in Fig. 9.1(b) for the 5-storey
wall. Fibre elements were used at all construction joints, i.e. at the foundation level and at each
intermediate floor levels, allowing for crack opening at these discrete locations. Beam eleelements were used elsewhere. Confined masonry material properties were used for the fibre
ment at the base (specially confined region) and uncon{ined masonry properties were used for
all other fibre end beam elements (refer to Chapter 3). Fig. 9.2(a) shows the constitutive relationships assumed for the masonry.
Compression unloading properties for the individual fibres in the DRAIN-2DX fibre element
path, i.e- for
are outlined in Fig. 9.2(b). ln the elastic range the fibre unloads along the loading
any strain excursion before the fibre strain has once has surpassed the strain em colresponding
path is
to the fibre maximum strength, f'r.For fibre strain excursions beyond e-, the unloading
controlled by the stiftress degradation parameter FU (range of 0 to l). Using FU 0 results in

unloading along a line pointing directly towards the origin (o). Using FU I results in unload-
ing along a line with a slope corresponding to the initial stiffness (defined by the line (o)-(i) in
Fig. 9.2(b)). Using FU in the range between 0 and I results in unloading along a path with a
slope based on linear interpolation between the slopes for
FU:
0 and FU
1. The
lower bound
of stress is given by the basis line which is defined by the origin (o) and the sfiess plateau G) as
the
shown in Fig. 9.2(b). Reloading occurs along the previous unloading path and is bound by
envelope. Kurame et. al. [9-2] recommended using FU
: 0.5 for concrete based on research by
Sinha et. al. t9-3]. [n absence of relevant information on unloading properties specific to concrete masonry
FU:0.5
has been assumed to be applicable'
The lateral seismic mass of the structure, assumed lumped at each floor level, is given in Fig.
9.1(c). No attempt has been made to model the vertical and rotational seismic mass of the
^223 -
-Unconfined,
fm = lE MPa
Confined, fm = 18 MPa
/,
6'
A
f ql
\s/
! rv
E
g,
\
0.000 0.002 0.0(N 0.006 0.008 0.010 0.012
0.0
t4 0.016 0.018 0.020
Straln
f^
-Envelope
Unbading - rclradirq
(b)
(D
o
o
e
I ft
g.
'-_,:l---(o)
--
*-"_""---in)f
-'-.--{*",",,n.
rcsldual
plaleau
Fig. 9.2-Masonry constitutive relationships
sfucture. Modal analysis based on the wall initial stiffiress EI resulted in a first mode natural
period of T1 : 0.364 sec and mode shape {0r } for the S-storey wall, as shown in Fig. 9.1(d),
and Tr -- 0.621sec for the 8-storey wall.
9.fJ
Push-over analysis
Fig. 9.3 shows the result of a cyclic push-over analysis superimposed on the experimentally
determined force-displacement response of Wall
I of Series 3 (Chapter 7) using FEM fibre
modelling. The figure reveals an experimental force-displacement response characterised with

a distinct softening that occurs upon opening of a single crack at the base.
-224
It is noted that
the
200
150
(E
o
E
(')
o
o
o
100
50
0
0.01
0.00
o.o2 0.03 0.04
0.06
0.05
DisPlacement (m)
Fig. 9.3-Cyclic push{ver response
250
r;
z
150
.E
u,
S too
ot!
50
model, 1st elastic mode
-Fibre
-+ ATC40 Bi-linear APProximation
-+ Analytical Bi-linear Approximation
0.00
0.10
0.05
Dlsplacement at h' (m)
0.15
Fig. 9.4-Force-displacement relationships
wall returned to its original position and the initial stiftress was reinstated upon unloading, and
modthat little hysteretic damping occurred. The FEM fibre model response shown in Fig. 9.3
elled the experimental curve reasonably accurately in terms of initial stiftress and softening
0'M0 m.
and the hysteretic shape was captured well for displacement up to approximately
Monotonic push-over analysis was performed on the multi degree of freedom (MDOF) strucat
ture (fibre model) shown in Fig. 9.1(b) by applying a lateral force vector with components
eachfloorlevelproportionalto tqr). RefertoFig. 9.4for theresultingenvelopecurve.
-22s
9.2 EQUTVALENT SINGLE.DEGREE OF FREEDOM STRUCTURE

It is desirable to tansform the MDOF fibre model into a single
degree of freedom (SDOF)
structure. This allows for reduction of analysis effort and development of a simplified design
approach. In the transformation to a SDOF structure,
it is important to retain key structural
properties observed in the MDOF model: base shear, base moment and structural period" so
that the maximum displacement response
d**
at some equivalent height h* remains compara-
ble between the MDOF and SDOF models.
9.2.1 Tbansformation
It can be shown that the nh mode response of
a linear elastic
MDOF structure can be modelled
exactly by an equivalent SDOF structure with a modal mass of Mn*, an equivalent height of
h*
and a uniform elastic stiftress of EI. Expressions for these quantities are [9-4]:
M-*
-.-n
:9
AI
12
where
Mo
/V
L:: >*jLj,
and Mo:
j=r
nr
ro
where Lsr= >hj*i[in
hn* :3
Ln
'm2
{0r} :
(g.t)
j=r
Q.z)
j=l
ml
{m} :
\m/\i)t
t8.2
18.2
m3
18.2
m4
18.2
m5
t8.2
Qrs
1.000
Qr+
0.718
0n
0.451
0rz
0.222
Qrz
(t)
(e.3)
(e.4)
0.061
-226 -
hr
{h} :
ln Eqns. 9.1
and
15
h2
t2
h3
h4
hs
9.2,\,01"
and
(e.s)
(m)
correspond to ttre mass, modal displacement of mode n and
the height above the base of floor level j. The values of m,
Q1
and h are provided in Eqns. 9.3 to
SDOF
9.5 for the MDOF structure shown in Fig. 9.1: From this the properties of an equivalent
structure are (Fig. 9. I (e)):
r!:
++.ast,rf :
531.8
trr^,Mr:32'28t,I'Ir:
61'8
t,h;:11'90m
and
nrln* = 0.793
not
From experience lg4l, it was anticipated that higher mode effects (n:2,3,4,"') would
contribute significantly to the wall base moment. This implies that the maximum displacement
of the wall could be modelled reasonably accurately using only the l*t mode SDOF structure
for analysis, because the lateral displacement for cantilever wall ductile response arises nearly
exclusively from flexural action adjacent to the basen thus
Thbles 9.1 and 9.2 show values of h*/h*,
h*:
hl*
and
M*
- Mt*'
M*/M md 0r calculated for walls with
a range
of
assumpnumber of stories N, where ho, is the wall height and M, is the total seismic mass. The
tions were: all stories of identical height and constant flexural rigidity EI over the entire wall
floor
height. Results in Table 9.1 represent walls with seismic mass distributed uniformly at all
mass m at
levels and the roof while results in Table 9-2 represent walls with identical seismic
floor levels and mJ2 at the roof. It is observed that the results are independent of Mu h*
all
and EI.
9.2.2 Definition of SDOF structure

The next step in the process is to define the force-displacement relationship for the equivalent
SDOF structure. It is clear that this relationship must be non-linear in order to capture the
the SDOF
expected ductile behaviour. This conflicts with the assumption of using h* to define
-227 -
TABLE 9.l-Equivalent SDOF properties, m at each floor and roof

J
4
2
t
5
7
6
N
8
Num. stories
M*/M U.b)J 0.659 0.667 0.679 0.696 0.721 0.791 1.000
Mass particip.
Ettbctrve hetght ll*/llw 0.'l'lt) U./IJ U. /dJ u. t94 U.6UY 0.E33 U.U /Y I.UUU
0.025 0.032 0.043 0.061 0.093 0.156 0.320 1.000
I
li
oo
VGv
2
0.095 0.r21 0.160 0.222 0.328 0.s32 1.000
_H
a
-'v
F
J
0.200 0.254 0.33 r 0.451 0.647 1.000
LA
(J
d
I
Qtrc)
4
0.334 0.418 0.538 0.718 1.000
EO0)
Itro)0)
AA.o
_=tht
lJ-
'-
0.487 0.603 0.765 1.000

0.652 0.799 1.000
0.825 1.000
1.000
TABLE 9.2-Equivalent SDOF properties, m at each floor, m/2 at roof

8
Num. stories
N
particip.
M*/IV[, 0.650
Mass
Effective height h./h,, 0.731
0.026
oo Lr
dv^v
2
0.097
a
E L)
J
0.204
d9P()
!.t
4
0.339
E()() tr {.)
0.492
5
troo)
(so.()
0.657
'ct cd c!
6
0.827
*=chx
8
L000
6=
0.655 0.662 0.67r 0.684 0.706 0.752 1.000

0.733 0.735 0.739 0.746 0.760 0.802 1.000
0.033
0.t24
0.2s9
0.424
0.609
0.803
0.04s 0.063
0.r65 0.228
0.338 0.459
0.545 0.723
0.770 1.000
1.000
0.096 0.t62 0.327 r.000

0.336 0.540 1.000
0.654 r.000
1.000
1.000
structure because of the inherent linear elastic assumption. Nevertheless, the assumption also
holds for ductile PCM wall response as will appear from the numerical analysis below.
The expected PCM wall behaviour represented by the MDOF fibre model is shown in Fig. 9.4
(envelope curve), along with two bi-linear curves. The curve denoted 'AIC40' was determined
from the fibre element push-over curve according to the procedure outlined in the ATC40 document [9-5]. The procedure essentially equates the area under the push-over curve with the
area under the bi-linear approximation, given a particular target point P on the push-over curve
and that the steep branch of the bi-linear curve coincides with the push-over curve at 0.6Vr'.
The curve denoted 'Analytical' originated from the simple analysis procedure proposed in
Chapter 8, based on a wall initial stifkress of K1 = V./d"", a wall 'post-yield' stiffiress of
(Vr, - V")/(d,, - d") and using h. :
&:
h*, refer to Fig. 9.5. It is seen in Fig. 9.4 that there is negligi-
ble difference between the two approximate curves. Because of the limited hysteretic damping
produced by PCM walls, it is appropriate to use a bi-linear elastic model (loading and unload-
- 228
250
Vr, dn
First
tendon yield
vv', dv
/
200
v,y, d,y
Nominal strength
fv=15m
tte= 11.9m
rcn
L=3.6m
b. = 0.19 m
Maximum
serviceability moment
fn
t1t,
Ap,
fr
840 mm2
Eo' = '190 GPa

fp, = '1785 MPa
q,
to
='18 MPa
E- = 800fm
V.,4
0,
100
N=578kN
P=89?kN
First
0.120
0.9o m
em = 0.013
cracKing
E=
hp
50
V- do
Predictlon
+FAnalytical Bi-linear
40
20
60
B0
100 "120 140 160 180
DisPlacement at
200
h' (mm)
Fig.9.5-5-storey wall, Analytical prediction model based on Chapter
vy
o
(,
L
l!
looding ond
unlooding olong
some curve
Displocement
Fig. 9.6-Bi-linear elastic model

ing along the same curve) for the SDOF structure. Refer to Fig. 9.6 for definition of the bi-linear elastic model.
9.3 DYNAMIC ANALYSIS

Time-history analyses were performed with the fibre MDOF and bi-linear elastic SDOF modthe
els defined above using NZS 4203:199219-6lcompliant earthquake records. Fig.9.7 shows
NZS 4203 Uniform Hazard Spectra (elastic acceleration spectrum) for intermediate soils and a
1.2 (zone of highest seismicity) with 3% and sYo viscous damping,
seismic zone factor of
Z:
-229 -
1.6
4203(Z=1.2)5o/o
-NZS 4203 (Z=1.2) 3Yo

--ltlZs
1.4
't.2
----.
il
.......\.
UBC(Zone4\5%
,'/
:0.8
u,
:"-....
0.6
0.2
0.'l
o.2
0.4
0.3
0.5
0.6
o.7
0.8
0.9
Period (sec.)
Fig. 9.7-Uniform Earthquake Hazard Spectra

specified for inelastic time-history analysis. The 5% damped spectrum is specifically given by
NZS 4203 while the 3Vo spectrum was derived by scaling of the 5% specfum as recommended
by Kawashima and Aizawalg-71: S."(T,E)
S".(T,:0.05)K1.5(408+l)F0.51. For reference,
the UBC (1997) [9-8] elastic spectrum for seismic zone 4 and soil category Sp with
5olo
viscous
damping is also shown in Fig. 9.7, suggesting a similar level of seismicity. Both the NZS 4203
and the UBC spectra are associated with an earthquake return period of approximately 475
years.
As indicated, two viscous damping ratios 8,3% and sYo, were considered for analysis. The 5%
spectrum shown in Fig. 9.7 represents the expected behaviour of nearly all 'ordinary' structures made of reinforced concrete, reinforced concrete masonry and structural steel. The origin
of the '5o/o' remains unclear. It has for example been suggested that 5Yo was selected in the
'early days' of electonic calculation because it resulted in numerical stability. Paulay and
Priestley [9-9] stated that damping ratios for ductile reinforced concrete response generally
range from 2Vo to 7Yo, such that 5% is readily justifiable. Kurama et al. [9-2] argued that
3%o
viscous damping was appropriate for dynamic analysis of unbonded post-tensioned concrete
walls. Given the above statements
it is recommended to adopt 37o viscous damping for
unbonded PCM walls until further research confirms/rejects the appropriateness of using this
viscous damping ratio.
-230-
It is clarified that
( in this study is completely
hysteretic damping.
dissociated from being a means of modelling
of
is attributed to energy dissipation throughout the structure due to effects
and defonnasuch as concrete micro cracking, friction at molecular level (heat development)
well
tion of non-structural elements, and occurs in the structure at any response level (elastic as
mainly attributed to ineas inelastic). Hysteretic damping is, for reinforced concrete materials,
in
lastic deformation of reinforcing steel, and solely is a feature of localised inelastic action
plastic hinges.
Viscous damping proportional to initial elastic stiffiress was assigned to the models, with 3%
assigned to modes I
mass proportional damping for the SDOF model andSo/oviscous darnping
and 2 for the MDOF model (Rayleigh damping).
It is noted that no viscous
damping was
large deformation
assigned to the base fibre element in the MDOF model because the expected
(critical
response (rotation) would result in superficially high viscous damping in this element
for elements with significant stiffness degradation). Some amount of hysteretic energy can be
deformation.
dissipated in the fibre elements as suggested in Fig. 9.3 in case of inelastic fibre
strainThis effect provides only a small amount of energy dissipation in comparison with steel
ing. The SDOF model did not have hysteretic energy dissipation ability.
9.3,1 Earthquake records

NS,
Four strong motion earthquake accelerograms were selected for analysis: El Centro 1940
Taft Lg52 S-6gE, Hachinohe 196g N-s and Matahina Dam 1987 N838. Each accelerogram
the
was nornalised (frequency content modified while preserving phase angle) to comply with
NZS 4203 elastic 5%o damped spectrum (using
Z: 1.0) and termed 'Nl, N2, N3 and N4'. Plots
of ground acceleration a, (fraction of g) vs. time of the normalised records (Z = | .2) are shown
in Fig. 9.8. It is seen that the maximum acceleratioll Bg,max ranged between 0.439 and 0.51g'
and that the duration of the srong ground motion records varied considerably' The strongest
(in terms of a) occurred at the beginning of the records for Nl and N2 while N3
intensity
by a
showed relatively high intensity throughout the duration of the record. N4 is characterised
short duration of strong intensity over approximately 8 seconds.
The 3% and
spectra.
5o/o
damped spectral responses of
It is seen that there is good
Nl-N4 are shown in Fig. 9.9 with the hazard
agreement between the response of the individual records
and the hazard spectra, in particular for the 5% damped response. The average of the records
-23t
0.50
0.40
(a)
Nl
0.30
0.20
o.1o
^s, o.oo
"
-0,10
-0.20
-0.30
-0.40
ll
,ll
vtu
lu"''
I l'
t\|
'l
ll ll1
til
fII
|/
lil
'tl'
fl'q
81012
lfl lfiy
NZS4203,Z=1.2
-N1,
-0.50
uf
UII'I
20
16
14
Tlme (sec)
(b) N2
0.50
0.40
0.30
0.20
0.10
o.*
? -0.10
"
-0.20
llLr
lllll.lltlll llilrl .rl ll..l,..l

tl rllll
ttll
tl
-0.30
-0.40
-0.50
.,,1,,..t
l,
'l
l
I
v.-./ as*
= 0'5Og
10
20
15
25
30
35
40
Tlme (sec)
0.50
0.40
0.30
(c) N3
arE
0.20
^ 0.10
*t o.*
-0.10
-0.20
= 0.489
tlllllr . lll
1r[
lllh
ltilll'I tll I
,lr' ,lllllh
rqllt ltTlllt14
I'
t lh.ll
t
I'
'l
ltrlil'f
-0.30
-0.rrc
-0.50
10
20
15
,l
25
["1
ffi
-N3,
NZS4203,Z=1.2
30
35
40
Time (sec)
(d) N4
^
{
"
0.50
0.40
0.30
0.20
0.10
o.oo
-0.10
-0.20
-0.30
-0.40
-0.50
-ar*=0.519
., .II tillt.
rl ,Illlllflltl,Il
'$tnll;lt'lt
,rltl
lll
lilt
ll
\'lh^. n.*
If llJl^/ v'
t,
v'rAr^- -*v
".n
-N4,
81012
Time (sec)
Fig. 9.8-Normalised earthquake records Nl :N4
-232 -
NZS4203,Z=1.2
18
,Z=1.2,3To
1.6
-MS4203
---N2,3%
-N1,3%
-N4,3%
--..*-*Avorage
-N3,3%
N1:M 3%
1'o
1.2
E1
c
o
E
g
o.e
o.o
.9
o.+
(a)3% damping
t,o
0.2
0
1.2
-N2S4203,2=1.2,1Yo
---N2,5%
-N1,5%
50/6
-N4,5%
--N3.
- *- Average Nl :1,14 5%
a
co
=1
o
I6
o.e
tr
gE 0.6
o
o
a
t o.r
'E
g
o.'
O.z
o'1
o'2
o'3
o'4
Fig.9.9-Spectral response of N1:N4,

Nl:N4
o'7
t t*"a
0'8
0'9
n"t"rt
3o/o and 57o
damping
(shaded lines) show close correlation with both the 3o/o and
5o/o
damped hazard specta.
The motivation for using these normalised (artificial) records was to reduce variation of
ensure code comresponse of the models as a result of variability in earthquake records and to
pliance.
9.3.2 Time-History response: MDOF vs. SDOF'

Fig. 9.10 shows the time-history response of the 5-storey MDOF and equivalent SDOF strucrelationtures in terms of lateral displacement at h* due to Nl through N4. The SDOF bi-linear
-233 -
0.16
;
lE
(a)
Nl
o.12
0.04
6
E
o.oo
E
o
n
A
\^ nAn illlnaAllJt
UV' v vwvvv u
./l
vl
-0.04
-0.08
-o.12
.9
-d.*=0..t50m
fi
o.oe
\rrl\A^^^ n^^-rrA
yl
NV
Nl,3%
-MDOF,
Nl,
-SDOF,
-0.16
81012
37O
18
14
Time (sec)
0.12
o.oo
-MDOF,
N2,3%
l.
(b) N2
a!
E
E
o.o+
o.oo
0,
H -0.04
N2, 3%
-SDOF,
/Il r
ln.
ll
/1
-^{frfi{ Hffi
ftduq ru ffi lltw
I
CL
f; o.oe
Yv
@-o*=o.tot
ffim-s
il
-o.12
15
10
20
25
30
Time (sec)
o.12
E
(c) N3
;o
b
o.oe
-MDOF,
N3,3olo
-SDOF,
!c -0.04
If
f; -o.oe
It
r rl. -l I
llll,r ll tll t lh ll ll llll, -
hr{
0.00
3olo
0,04
N3,
ryTrll l|l ll!|ll|ll
'l
il
lt
J llllllfl dllhl tlMlh,.

- "..""llullll[Yil,ulllllftll|F'
ll["-
ll
ll
!'r'l
(!)-dffi.0.0e5m
ô.12
15
10
20
25
30
35
40
Time (sec)
0,12
(d) N4
o.oe
o.*
0.00
-MDOF,
N4,37o
-SDOF,
N4' 37o
P--d.-=o.1oe.
-^^^^^ al
^-,,.'.'VU
/l
H -0.04
^/ILm/l
JVw v'^
-rL
-o.oa
l,|v
-0.12
0246
810121416
Time (sec)
Fig. 9.10-5-storey, Displacement response MDOF vs. SDOE, Nl:N4
-234-
20
ship was determined by the 'Analytical' procedure discussed above. Convergence of the time-
history response was confirmed by analyses with varying time step length.
It is noted that the maximum
response to
Nl
and N4 occurred at the same time for the MDOF
and SDOF models and that the maximum displacements virtually were the same. The maximum SDOF response to N2 and N3 did not occur simultaneously with that of the MDOR and
the SDOF resporxie was larger than the MDOF response.
Fig. 9.I
shows similar plots of time-history response for the 8-storey sffuctures (3% damped)'
It is seen that the maximum
response to N1:N4 occurred at the same time for the MDOF and
sDoF models and that the maximum displacements were similar.
9.3.2.1 Displacement resPonse

Tables 9.3 and 9.4 summarise results for both prototype structures and
earthquake records.
It is
seen that the maximum displacement response
all four normalised
(d.*)
varied signifi-
cantly between the earthquake records, with the Nl record causing the largest displacement
the
response for the S-storey wall, the N4 record causing the largest displacement response for
g-storey wall and the N2 record causing the least displacement demand for both walls- Com-
TABLE 9.3-Analysis results for S-storey wall models, maximum values

SDOF Bi-linear elestic model
MDOF Fibre model
d..r(h*)
N1
N2
N3
N4
Average
St.
Dev.
StDev/Av
ffi
0.078 0.101 398

0.079 0.101 4r7
0.109 0.141 430
0.
t03
0.028
0.28
0./,33
0.036
0.27
o.ol24 0.0125 o.ol33

2587 0.0065 0.0067 0.0073
2583 0.0066 0.0067 0.0075
2606 0.0092 0.0094 0.0099
2614 0.0087 0.0088 0.0095
432
31 39 0.0024 0.0024 0.0024
0.07 0.01 0.28 0.27 0.25
Mr",
205
203
206
2584
2435
2417
2449
2471
2t7
0.151
0.101
0.095
0.106
0.113
0.022
0.20
66
0,0127
0.0085
0.0080
0.0089
0.0095
0.0019
0.20
0.03 0.03
TABLE 9.4-Anatysis results for 8-storey wall models, maximum values

SDOF Bi-linear elastic model
MDOF Fibre model

EQ
NI
N2
N3
N4
Average
St. Dev.
StDev/Av
*)v
*) d."r(h*
0.n6 0.299 660 6289 0.0122 0.0124 0.0132

0.132 0.175 850 6374 0.0071 0.0073 0.0082
0.233 0.308 822 6s02 0.0126 0.0128 0.0138
0.217 0.286 9s7 6236 0.0117 0.0119 0.0129
0.202 0.267 822 6350 0.0109 0.0111 0.0120
0.041 0.054 107 100 0.0022 0.0022 0.0022
0.20 0.20 0.I3 0.02 0.20 0.20 0.19
-235-
0.2t9
0.r32
0.2t7
321
5930 0.01l9
3t2
57& 0.0072
32r
5927
5962 0.0t27
5896 0.0109
77
0.0022
0.20
0.01
0.236
323
0.201
319
0.040
0.20
0.01
0.01l7
i
(a) Nr
!5
0.25
0.20
G}
o.rs
o.1o
o.os
o.oo
/l
E +.os
nnA"/\nn
ln ,l
-1
IV
-E -o.ro
.fr -o,rs
=O.?2Gm
-0.20
N1,306
-MDOF,
-0.25
-t
vvvIt\l/,\
rlvfvlv,v
V Y/ V \/ V
10
14
12
N1,3%'
-SDOF,
20
16
Time (sec)
0.25
^.9 0.20
o.rs
(b) N2
i o.ro
!5 o.os
o.oo
-o.os
o.ro
MDOF, N2,
37o
- SDOF,
A ^r
/i ^ l\
*-+llfl#fh
v v lf'vu
N2,
3%
q*
\_.=
/t
n (lll
A-a
vl
= 0132 m
"
ra
vwv'
-t?
fl
-1,
rl
AAr.
ll lf Yll lf lrvv"
vv
-v
.6 -o.rs
o -0.20
-0.25
15
30
25
Time (sec)
^
(c) N3
o.25
0.20
o.rs
i o.ro
!6 o.os
o.oo
F
-o.os
t -o.ro
.F -o.rs
o -0.20
-MDOF,
-0.25
15
N3,370
N3,370
-SDOF,
20
Time (sec)
^
5.
(d) N4
o.25
0.20
o.rs
-MDOF,
N4,3%
-SDOF,
o.ro
/[\nA
I
ll,/ \v/"
ra
o.oo
ffi
/\
n
!5 o.os
E
E
o
N4, 370
VI
-o.os
-o.ro
-o.rs
vv
-o.zo
-0.25
12
0246810
II III VV
il\l
II
14'16
Time (sec)
Fig. 9.11--8-Storey, Displacement response MDOF vs. SDOE, N1:N4
-236-
18
paring the response of the MDOF and SDOF structures for a given earthquake, it is seen that
reasonably consistent displacement demand arose between the SDOF and MDOF structures.
9.3.2.2 Drift demand

In the context of displacement focused design, it is important to evaluate the wall drift demand
(also termed drift angle). Structural codes generally stipulate 'interstorey' drift limitations,
given time t during the earthquake, conforrt
Ti,ro, so that the structural response must, at any
to:
(e.6)
where d; and di-r slmbolise the lateral displacements at floors
between these floors. Limitation of the interstorey drift demand
and
j-1, and
is the height
1 is intended to limit
lateral
to
deformation to avoid excessive floor slab and beam deforrnation, to avoid excessive damage
partitions, and to
nonstructural elements such as windows, exterior wall cladding and interior
limit p-A effects. The largest value for the right hand side (RHS) of Eqn. 9.6 is normally found
near the toP storey.
,Roof
drif
is a measure of the global lateral deformation and is defined as follows:
d,
(e.7)
T-: n
is the maximum lateral displacement at roof height and h* is the wall

moderate stoheight. It is found that y1 always is larger ft* T* For ductile cantilever walls of
the wall
rey height it is typically found that y; is only slightly higher than yu, because most of
In this equation,
deformation results from development of a plastic hinge at the base'
It is interesting to calculate the drift demand at the equivalent height h*
because
it is available
from both MDOF and SDOF analYsis:
dt*
(e.8)
lp: fi
ln this equation {,. is the displacement demand at the height h*.

on the results
Comparison of interstorey, roof and equivalent height drifts can be founded
md y, for the MDOF structures it is revealed that
shown in Tables 9.3 and 9.4. Comparing
Tr'.
-237 -
1* is larger than y5* in all cases, as expected, and that the difference is minimal. Thus for ductile
unbonded PCM walls,
fi', is a good estimate for y* Comparing
y1,*
for the MDOF and SDOF
structures it is seen that y1r1se6ry is a good approximation for {,.6,1oory. Maximum interstorey
drift was calculated for the fibre models and are shown in Tables 9.3 and 9.4. The results suggest that Ti on average is 8.2Vo larger than {,*iyooo,. Given the above statements
it is proposed
that the interstorey drift demand on a ductile unbonded PCM wall can be approximately estimated by means of the wall drift at the equivalent height of the SDOF structure as follows:
t :
(e.e)
1.1 xTr,*(sDOF)
9.3.2.3 Base moment response

As expected, the maximum moment (M."J demand did not vary significantly between the
models when considering different earthquake records, simply because of the distinct base
moment strength and low strain hardening (of the order
of
L%).
h appears that higher
mode
effects in the MDOF structure on average resulted in approximately 6-7Vo higher base moment,
compared to that of the SDOF structure.
9.3.2.4 Base shear response

The base shear response was drastically different between the MDOF and SDOF models. Figs.
9.12 and 9.13 show the MDOF maximum storey shear (Vr,aoor) normalised with SDOF maxi-
mum shear (Vsoor) for the 5 and 8 storey structures and illustrate the large difference in shear
demand between the MDOF and SDOF models. For the 5-storey wall, it was on average found
that the MDOF maximum base shear (Vu",") amounted to approximately 210%o of that of the
SDOF structure, and for the 8-storey wall the MDOF/SDOF maximum shear force ratio
amounted to approximately 260%, clearly indicating a very strong influence of higher mode
effects. It is also seen from Figs. 9.12 and 9.13 that V5o6r was exceeded in all storeys, even in
the top one.
Paulay and Priestley [9-9] explain that mode shapes in 2nd and 3'd modes of vibration of cantilevers with fixed or hinged bases are very similar, and that this suggests that the formation of a
plastic hinge at the wall base may not significantly affect the response in the 2nd and 3'd modes.
While a base plastic hinge will greatly reduce the wall actions associated with first mode
response, it can be expected that those resulting from higher mode response of an inelastic can-
tilever will be comparable with elastic response actions. In reality, it is improbable that higher
-238
_3
8
E
2
0.0
l'0
0.5
Storoy
1'5
Shsr AnPlllltld
VIs'V@t
Fig. 9.12-5-storey wall, storey shear
5
T
E4
3
0 .{-
1.5
0.0
StotEY Shr Amplmcrdon
2.0
2.5
VFoCVmr
Fig. 9.13-8-storey wall, storey shear

mode effects
will contribute significantly to the moment distribution in the wall
because of the
of the
short shear spans associated with the displacement patterns of higher modes and because
distinct base moment strength (normally substantially lower than the strength required for
strictly elastic response). ln contrast, the I't mode base shear is reduced by plastic hinging,
while the higher mode shear contribution may still be significant. The storey shear amplification due to higher mode effects are undoubtedly dependent on the building height (or fundamental period T). paulay and Priestley recommend a dynamic shear amplification factor,0)"
1.4 for 5-storey walls to 1.6 for 8-storey
l0] on reinforced
of
walls. More recent studies by Priestley and Amaris [9-
concrete cantilever walls suggest much higher dynamic shear amplification.
-239 -
Values of c.\
I .8
for 4-storey walls and
ou:2.5 for
8-storey walls were found when compar-
ing results from SDOF displacement based analysis with MDOF ductile time-history analysis.
Research by Kurame t9-11] on precast unbonded concrete cantilever walls also confirm shear
amplification of such magnitude.

The observations by Priestley and Amaris and by Kurame correspond well with the 5-storey
and 8-storey analyses results given above. This holds despite the fact that unbonded PCM
shows of significantly lower hysteretic damping and lower viscous damping (3% assumed)
than RC and RCM, that in conjunction are likely to less effectively suppress higher mode
will
effects. As
be shown in section 9.5, the achieved base moment reduction due to rocking
(hingrng adjacent to the base) relative to purely elastic response was 4 for the 5-storey wall and
5 for the 8-storey wall, thus the SDOF base shear was reduced 4-5 fold but the MDOF (real)
base shear reduced
with only 4/2.1
1.9 for the S-storey
wall and 5/2.6:
1.9
for the 8-storey
wall - clearly a significant contribution to the base shear from higher mode effects. It is finally
noted that more accurate estimation of c4 could be achieved by more accurate modelling of the
actual wall shear stiffness, however this was beyond the scope of this study.
Given these results, it was concluded that the SDOF model, defined above, predicted the wall
displacement response at h* and the base moment with reasonable accuracy, and that the base
shear demand could suitably be predicted
by estimating the amplification due to higher mode
effects, e.g. using interpolation between oJr:2.1 for the S-storey wall (T
2.6 for the 8-storey wall
(T:
0.364 sec) and rou :
0.621 sec) based on the building fundamental period T (extrapo-
lation may be used for buildings of period lower than T

aspect ratio
A,: h*/l* may simply be used for interpolation
0.364 sec). Altematively, the wall

to find
CIu
because this parameter
is indicative of the building period and dynamic properties. It should be kept in mind that the
proposed values are valid for force reductions R of 4 to 5 (see next section for definition of R).
Using
R:
(purely elastic response) will result in cl, slightly higher than 1.0 because the
I't mode actions (base moment and base shear) that dominate purely elastic
response
'full'
will
be
developed. Therefore, both the wall aspect ratio, A,, and the achieved force reduction, R,
should be considered when estimating
ton
9.4 SPECTRAL ANALYSIS

Spectral analysis consists of calculating the nominal structural response as a function of the
structural period T for given structural characteristics. In the present case, the characteristics
-240 -
were the strain hardening ratio cr : KzA(r and the viscous damping ratio (. A parametric analysis has shown that the ratio
c is typically of the order of 0.01 to 0.02. Calculations were per-
formed for two viscous damping ratios: 3% and SVo.

The bi-linear elastic SDOF model, as illusfrated in Fig. 9.6 was characterised by the following
parameters: initial stiftress K1, post-yield stiffiress K2, yield strength
M*. It
is noted that the yield strength signifies the
due to opening of the base crack;
Vr' and seismic
mass
wall strength at which wall response softens
it is not associated with yielding of the prestressing steel.
9.4.1 Elastic response
V, required to ensure purely elastic wall response (a value

:
was
strictly associated with developing the base moment M, Vrh* for the SDOF structure)
determined as the maximum wall strength demand for given T, Kr, M*, e and earthquake
The minimum wall yield strength
record
(Ee.
The elastic spectral acceleration S"" was defined as Vy/gM*.
shown that Su": fr(T, E, EQ); noting that Su" is independent of choice of
It
can easily be
Kt'
merely can be
The reader is reminded that the spectral acceleration for the elastic structure, S"",
of the
read off the Uniform Earttrquake Hazard Spectra in Fig. 9.7 because of normalisation
earthquake records.
9.4.2 Ductile resPonse

the
The wall ductile response was calculated for a range of force reduction values R, defining
: Vr/R' where v, was determined from the
yield strength of the ductile sDoF structure as
\'
elastic response as explained above. The definition of
Vr', Kr and K, is shown in Fig' 9'6' A
for
realistic range of R values,2 - 6,was selected, reflecting typically achieved force reduction
ductile reinforced concrete masonry cantilever walls. The spectral displacement 56 frI(T' &
Ee) was defined as the maximum recorded displacement for a given calculation and has
the unit of lenglh. It is observed that 56 is independent of the choice of K1 and K2. Running
cL e,
the
multiple time-history SDOF analyses with various K, (but constant M*/K, ratio) confirmed
independence of the choice of K1 and K2.
9.4.3 Spectral disPlacement

Fig.
9. 14 shows the spectral displacement
for a
lo/o and q
3Yo and
5%. A period range
of
from the four

0.1 sec to 1.0 sec was investigated, R was varied between 2 and6, and responses
-24r -
0.35
aE
0.30
o.zs
E,=3Yo, g = 0.01
-R=1
-+-R=2
-r-R=3
-.-R=4
-*-R = 5
-x-R= 6
(! 0.20
CL
.9
o.rs
0.10
lt
u, 0.05
0.4 0.5 0.6
0.2
0.1
0.7
Period T (sec)
0.35
E=5o/o,o=0.01
atr
-R=
-e-R=2
-r-R = 3
0.30
-+-R = 4
-*-R = 5
-*-R = 6
5 o.zs
E
c
TL
0.20
.9
3o
o.ts
(,
o&
0.10
(,'E 0.05
o'2 o'3 o o
Fig. 9.l4-Displacement spectra,
cx
,"?r3o
r?;1"r
o't o'8 o'e
: 0.01
earthquake records were averaged. Defining the characteristic response value of 36 as the average response is in line
with the requirements for inelastic time-history analysis given in inter-
national documents and codes 19-5,9-6,9-81. NZS 4203 specifically requires the use of
minimum of three earthquake records which must conform with the earthquake hazard spectrum (Fig. 9.7) in the period range of interest (a narrow period band around the structural fundamental period). UBC 1997 [9-8] requires that seven or more time-history analyses are
performed before the average value of the response of the parameter of interest may be used for
design. It was however felt for the present case, that averaging time-history response from the
four normalised earthquake records was suitable because of expectation of more consistent
response from such records than scaled natural records. As discussed above, four earthquake
-242 -
0.35
E,=3Vo,a=0.02
-R=1
-+-R=2
-+R=3
+-R=4
+-R=5
a0.30
E
E o.zs
-x-R = 6
0.20
CL
.t2
3IE o.ts
o
o90.10
0' 0.05
0.00
0.1
0.4 0.5 0.6
0.2
0.7
Period T (sec)
0.35
1=So/o,a=0.02
-R=1
-e-R = 2
-rR=3
=0.30
-+R=4
-*-R=5
E'o ru
-x-R=6
! o.zo
EL
.9
3o o.rs
t
go.1o
tt
(/,! 0.05
0.00
o'1 o'2 o'3
Fig. 9.I5-Displacement
o't
oo
,*3or?;l"l
spectra, ct, : 0.02
for all perirecords were used for analysis and each record conformed to the hazard spectrum
ods investigated.
It is seen from Fig. 9.14 that the displacement demand
increased with increasing stnrctural
period, and increased with increasing force reduction R. Comparing the rezults fot
damping,
it is seen that
56(3%) generally was about 20% lngher than Sd(5%)'
essential to estimate the available damping,
:2vo
(, realistically.
3o/o
and5/o
It is therefore
Results of similar analyses using
influand plotted in Fig. 9.15, showed only negligible reduction in 56, suggesting little
ence of cr on the displacement response for cr in the range

case of
cr:0.01 is considered.
-243 -
of l%-2Yo.In the following only the
0.35
E=
atr
3o/o, cr
= 0.01
0.30
6 o.zs
E
o
()
-G 0.20
CL
.9,
=IE
0.15
(,
0.10
(/,E 0.05
0.00
0.1
0.2
0.3
0.4 0.5 0.6
a.7
0.8
0.9
Period T (sec)
0.35
Z=
SVo, cr
= 0.01
0.30
E o.zs
E
o
(J
EL
0.20
.!)
=IE
0.15
0.10
(,
U'
E
u, 0.05
0.00
0.1 0.2 0.3 0.4 0.5 0.6
0.7
0.8
0.9
Period T (sec)
Fig. 9.16-Design displacement Spectra, cr : 0.01
In order to better exploit the information presented in Fig. 9.14, the curves were fitted to power
expressions that are plotted in Fig. 9.16. Note that this curve fitting is only valid in the 0.1 sec
to 1.0 sec period range and should not be used for extrapolation beyond periods of 1.0 sec.
These curves, referred to as design displacement spectra, represent a further abstraction of the
analysis results. Nevertheless, these are regarded as realistic because there is no apparent reason for the curves not being smooth (similar to the hazard spectra given in Fig. 9.7).
-244-
30
t,=3T", a = 0.01
Itc
F25
E
o
Bzo
t
3rs
E
o
g8to
CL
.9
o5
!
0.1 0.2 0.3
o.o
,J;3o r?;?"r
o't
0.8
0'e
Fig. 9.l7-Displacement ductility demand p4.
9.4.4 Displacement ductility demand

calculated by:
The displacement ductility demand, Fo, on the idealised SDOF structure can be
dô"
,Sd
S6R(2n\2
lro-T:T7n W
where
d,r:
(e.10)
d./R is the yield displacement of the ductile SDOF structure, and E
JV/t<t
ttt
Fig.
displacement corresponding to the required strength of the linear elastic SDOF stmcture.
:0'0 1 and 1: 3o/o' It is seen
shows p6 (average) calculated for R values of l , 2, 4 and 6, d
9. 17
period
from the figure that very large ductility demands arose in the short period range. For a
strucof 0.4 sec and R: 4, a displacement ductility demand of l0 was found for a ductile PCM
ture. This can be compared to a typical value of
ture of similar period and force reduction.
lta:4
for
reinforced concrete masonry stuc-
It is emphasised
that unbonded PCM walls can,
without difficulty, be designed to accommodate high ductility demand, because of their inherent behaviour characteristics: high displacement capacity and relatively high initial stiftress
(seeChapters5-7).
-245-
9.5
USE OF DISPLACEMENT SPECTRA
It is useful to regard a displacement, 4g, as the design target rather than a specified structural
performance parameter in terms of R or p6 for reinforced masoffy and concrete. Such an
approach can be termed 'displacement focused design' and could proceed as follows:
Displacement focused desi gn
l.
Given wall gross dimensions and tributary lateral seismic mass, fmd T,
2. Establish appropriate
[ (0.03 recommended),
3. Determine Su", given (Fig. 9.7 for example),
4. Calculate the elastic design base shear V,
: SugM* (SDOF only) (Tables 9.1 and 9.2),
5. Determine 4c as the smaller of the estimated wall displacement capacity d (section 8.1.7)
and code displacement
6. Enter the Sa chart
(drift) limitation (e.g. [9-6,9-8]),
(assume cr) at T,
4,
and find the appropriate R value (interpolation) (Fig.
9.16),
7. Design the wall for the base moment Mv' : (Vr/R)h*,

8. Calculate
and
veriff the assumed values of
du, and cr. Some
iteration may be necessary
9. Design for (MDOF) base shear (section 9.3.2.4) to ensure sufficient shear strength and no
sliding of the wall.
With regard to the structure shown in Fig. 9.1(c): T

found from Fig. 9.7 that
Su"
L25, thus
Vy:
Mr'
0.364 sec,
s:
0.01 and (,
:3%, it was
759 kN. A wall drift limitation of 1.0% or 4e
0.112 m was assumed, for which Fig. 9.16 produced

base moment would be
R:4
(approximately), so that the design
2125 klt{m and the peak design interstorey shear force is
t&Vr' --
2.1*75914: 398 kN. For comparison, the actual capacity was found to be M :2240 kNm (or
approximately
V:200
kN, see Fig. 9.a). The actual shear strength depends on the wall dimen-
sions and shear reinforcement ratio and the base shear friction capacity depends on construction joint details. Thus the design objective of d,, was achieved with the given design.
p6:
10.6
was calculated.
With regard to the 8-storey wall: T :0.621 sec,

9.7 that
Su"
:0.99, thus V,
1572 kN.
cr: 0.01 and ( :3%, it was found from Fig.
A wall lateral displacement limitation of 0.2 m at h* or
4r: 0.200 m was assumed, for which Fig. 9.16 produced R : 5 (approximately), so that the
design base moment would be Mr' : 5810 kNm and the peak design interstorey shear force is
r&Vu' :2.6*157215: 817 kN. For comparison, the actual capacity was found to be M: 5550
-246-
kNm (or approximately V
300
lN).
Thus the design objective of
sonable accuracy (M 5% lower than Mr') with the given design.
d, was achieved with rea-
p6:
10.5 was calculated.
9.6 NOTES
Higher order effects must be taken into account in the wall design, in particular for detemrining
the actual (MDOF) base shear demand, which may amount Io 2 to 3 times V.o from SDOF
analysis. The maximum SDOF base shear Vn.o and base moment
Mr', respectively,
M'"*
are higher than
Vr'
and
because of strain hardening. Strain rate dependency of the concrete masonry
was not considered in this analysis. Detailed design of the entire wall should be conducted
according to capacity design principles, e.g. [9-9]'
9.7 CONCLUSIONS
It
elastic force-displacehas been shown that a SDOF equivalent structure, based on a bi-linear
of multi-storey
ment relationship, can be used to accurately estimate the displacement response
unbonded post-tensioned concrete masonry cantilever walls.
It was
Displacement spectra were developed for PCM walls of realistic structural dimensions.
strucfound that the wall displacement demand was strongly dependent on the fundamental
(E). Response was
tural period (T), the force reduction factor (R) and the viscous damping ratio
not sensitive to variation of the strain hardening ratio (a) between lVo and2o/o.
than that for
The displacement ductility demand on PCM walls was found to be much higher
reinforced concrete masonry walls of similar proportions'
With reference to the results from the detailed MDOF models, it was concluded that the wall
howwer, the base
base moment was determined with reasonable accuracy by the SDOF model,
shear was underestimated by a factor of
2.I to 2.6. The base shear amplification due to higher
investigamode effects is subject to further investigation given the preliminary nature of this
tion.
It
was concluded that the proposed analysis method is suitable for displacement focused
design.
-247 -
9.8 REFERENCES
l9-l]
Prakash, V., Powell, G. and Campbell, 5.,
DMIN-2DX
Base Prcgram Description and
User Guide; Version 1./0, Report No. UCB/SEMM-93117, Department of Civil Engineering, University of California, Berkeley, Califomia, 1993.
t9-21 Kurame, Y., Pessiki, P., Sause, R. and Lu, L.W., Seismic Behavior and Design of
Unbonded Post-knsioned Precast Concrete Walls, PCI Journal, May-June 1999, pp.
72-89.
t9-31 Sinha, B.P., Gerstle, K.H. and Tulin, L.G., Stress-strain Behaviour for Concrete under
Cyclic Loading, ACI Journal, Vol. 61, No. 2, February 1964,pp.195-211.
[94]
Chopra, A.K., Dynamics of Structures: Theory and Applications to Earthquake Engineering, Prentice Hall, 1995.
t9-5]
ATC40, Seismic Evaluation and Renoft of Concrete Buildings, Vol. 1, Applied Technology Council, California Seismic Safety Commission, Report SSC 96-01.
[9-6]
NZS 4203:1992, Code of Practicefor General Stntctural Design and Design Loadings
for Buildings,
Standards Association of New Zealand, Wellington.
19-71 Kawashima K., and Aizawa, K., Modification of Earthquake Response Specta with
Respect to Damping Ratio,3'd U.S. National Conference on Earthquake Engineering,
Charleston, South Carolina, Aug. 1986.
t9-81 Untfurm Building Code 1997, International Conference of Building Officials, Whittier;
California, USA.
t9-91 Paulay, T., and Priestley, M.J.N., Seismic Design of Reinforced Concrete and Masonry
Buildings, John Wiley & Sons lnc.,1992,744 p.
[9-10] Priestley, M.J.N and Amaris, A.D., Dynamic Amplification of Seismic Moments and
Shear Forces in Cantilever Walls, European School for Advanced Studies in Reduction
of Seismic Risk, Research Report No. ROSE-200210I, Pavia, Italy.
[9-lU Kurame, Y., Seismic Behavior and Design of Unbonded Post-Tbnsioned Precast Concrete Walls, PhD Dissertation, Departnent of Civil and Environmental Engineering,
Lehigh University, Bethlehem, Pennsylvanea, USA, May 1997.
-248-
Chapter 10
SUMMARY OF CONCLUSIONS
10. CONCLUSIONS AND FUTURE RESEARCH
complete analysis
The general scope set out in Chapter L: 'to provide detailed guidelines for
achiwed' Analand design of ductile post-tensioned concrete masonry walls' was successfully
proposed in
ysis and design of pCM walls can readily be conducted using the procedures
investigations,
Chapters g and 9. The analytical procedures were supported by experimental
and international research and regulations.
properties
This thesis investigated a broad array of PCM wall features, ranging from material
the investigated elements
and large scale structural testing to theoretical considerations. While
fell outside the

were examined thoroughly, there are many more issues to be explored that
scope of this thesis. some of these issues are listed in section 10.6.
concrete
Three series of wall tests confirmed that fully grouted unbonded post-tensioned
The PCM walls
masonry is a competent construction form for ductile structural wall systems.
by rocking response.
reported herein exhibited a nearly non-linear elastic behaviour dominated
capacity was measured: up to 1olo was measured for u200 unconfined concrete
Large drift
and 1'5olo was

masoffy, more than l7o was measured for CP200 confined concrete masonry
a realistic 4measwed for Cpl00 confined concrete masoffy. Testing Series 3 that reflected
storey cantilever wall was particularly successful.
All fully grouted wall
tests showed that only
therefore be
localised damage in the lower wall corners occurred. Earthquake damage would
easy to repair.
partially and ungrouted PCM walls wore capable of developing significant strength, exceeding
the predicted nominal flexural strength. Significant displacement capacity was recorded
despite their brittle failure mode.
Series 2
Limited energy dissipation was observed during wall cycling. One wall from testing
dissipation.
incorporated energy dissipators that successfully increased the hysteretic energy
-249-
Theoretical procedures for determining PCM wall in-plane force displacement-relationship

successfully predicted the observed experimental behaviour. lnelastic time-history analysis
using strong motion earthquake records revealed that PCM walls are likely to undergo signifi-
cantly larger lateral displacements than reinforced concrete masonry walls of similar dimensions and strength. However,
it was found, for a S-storey PCM prototype building, that the
expected drift demand was less than |o/o and therefore lower than the typical code drift limita-
tion of l.syo. The 5-storey prototype model did not rely on any external energy dissipation.
Based on two creep and shrinkage testing series conducted at the University of Auckland and
on international research, it was concluded that long term prestress losses are considerable in
both grouted and ungrouted concrete masonry and must be taken into account in design. Use
of
high strength prestressing steel is required to minimise prestress loss.

Detailed conclusions for the individual chapters are presented below
1O.I MATERIAL PROPERTTES - CHAPTER

Accurate prediction of
f'.
and
is necessary for design of PCM walls because of the signifi-
cant influence of these parameters on the wall strength and displacement capacity. Unfortunately, the current New Zealand masonry code specifications poorly reflect actual concrete
masonry properties (in particular strength and elastic modulus). For the time being it is recommended to base the concrete masonry properties on the rational methods presented in this document.
Theoretical stress-strain curves for unconfined and confined concrete masonry may be established using the Priestley-Elder formulation. The required information to do so is limited to fn,
and, in case of confined masonry, the confining steel dimensions and yield strength. Using con-
finement plates in the bed joints greatly improves the masonry strain capacity.
A uniaxial
dependable strain capacity of more than lVo can readily be achieved.
IO.2 TIME DEPENDENT EFFECTS - CHAPTER
For grouted concrete masonry is was recommended to use the creep coefficient C.
3.0,
value in agreement with the values stipulated by BS 5628 and AS 3700. For ungrouted concrete masonry C"
: 3.0 is similarly recommended for design purposes. Values predicted by BS
5628 and AS 3700 appear to be on the conservative side. The following shrinkage strains were
-250-
er6:700 pe and for

recommended for design underNew Zealand conditions: for grouted walls
un-grouted walls
tr6:400
pa, values in agreement with international research.
be based on
Creep and shrinkage properties for partially grouted concrete masonry should
according to the
interpolation between the results for grouted and un-gtouted concrete masonry
ratio of grouted cells to the total number of cells'
ratio of 0.15, about
Using high strength strand with an initial stress of 0.7f* and an axial load
The use ofhigh
20% loss ofprestress can be expected after all long term losses have occurred.
in about 30% prestrength bar instead of strand, under similar conditions, is expected to result
In both grouted and
stress loss. Less loss is anticipated using ungrouted concrete masonry.
account in design'
ungrouted concrete masonry, losses are considerable and must be taken into
prestressing steel at some
However, high prestress loss can be countered by restressing of the
prestress loss by about
later time after the initial stressing. It is, in theory possible to reduce the
50%by restressing about 3 months after the initial stressing.
10.3 STRUCTURAL TESTING - CHAPTERS 5,60 AND 7

Series
l:
It was concluded that fully grouted unbonded post-tensioned concrete masonry is a competent
reported herein
material combination for ductile structural wall systems. The PCM walls
Large drift
exhibited a nearly non-linear elastic behaviour dominated by rocking response.
capacity of more than 1.0o4 was measured. Only localised damage occurred'
(squat) walls at relatively low
Tendon force loss due to yielding should be expected for short
Even after tendon yielding, reliable and self-centring wall behaviour is
wall drift ratios.
wall cycling'
expected. Before tendon yielding, little energy dissipation was observed during
partially and ungrouted pCM walls were capable of developing significant strength, exceeding
displacethe predicted nominal flexural strength. Despite their brittle failure mode, significant
ment capacity was recorded'
strength. The
Additional axial load from the prestressing enhanced the concrete masonry shear
paulay and priestley [5-4] provisions for masonry shear strength was found to most accurately
provision is highly conreflect experimental performance. The NZS 4230:1990 shear strength
servative.
-25r
Series 2:
Strengthening of the flexural compression zones
of fully grouted unbonded post-tensioned
concrete masoffy walls improved the wall displacement capacity and delayed the onset
of
strength degradation in comparison with PCM walls made with unconfined masonry. The PCM
walls from Series 2 behaved similarly to the behaviour observed for testing Series
l;
notably
they exhibited a nearly nonJinear elastic behaviour dominated by rocking response. The
exception was P2-CP-CA-ED that exhibited rocking response with non-linear elasto-plastic
behaviour due to the presence of energy dissipation devices.
The maximum wall strengths were insensitive to strengthening of the compression zone. Drift
capacities of 0.7o/o to 2o/o were measured using CP200, l.Io/oto 2.6% with CP100 and
l.l%
to
1.4% with high strength block corners. The highest performing solution was CPl00.
Only localised damage occurred. Relatively little energy dissipation was observed during
cycling of all walls, except for P2-CP-CA-ED that had 'dog-bone'energy dissipation devices
embedded. This simple device proved successful and provided a 33o/o wall strength increase
and a threefold increase of hysteretic damping relative to P2-CP-CA.
Exteme masoffy fibre sfiain at nominal flexural strength in the order of 0.0035 was found for
CP200 and 0.0020 was found for CPl00. Both values are substantially lower than 0.008 stipulated by NZS 4230:1990 for confined masonry. At ultimate displacement, extreme fibre strains
as high as 0.046-0.050 were measured
for both CP200 and CP100.
Series 3:
This testing series provided final confirmation that PCM walls of realistic proportions,
strengthened in the flexural compression zones with confining plates (CPl00), successfully
can withstand severe cyclic loading as imposed by major earthquakes. Ductile response was
measured, with reliable drift capacity
of l.5o/o for both walls tested. Relatively little energy dis-
sipation was observed during cycling of the walls. Adequate shear friction between wall and
foundation was achieved by intentional roughening of the wall to foundation interface.
All visual
damage occuned to the lowest 300400 mm
in the flexural compression zone. Ine-
lastic masonry response was measured in the lowest 400-600 mm of the walls. Exffeme
masoffy fibre strain in the order of 0.0010 to 0.0020 were measured at nominal flexural
strength. These values were substantially lower than 0.008 as stipulated by NZS 4230:1990 for
-252 -
were measured
confrned m4sonry. Average strains in the extreme fibre as high as 0.019-0.024
at ultimate disPlacement.
to exhibit
Comparison between pCM and RCM walls suggested that RCM walls are expected
plastic hinge
larger displacement capacity than PCM walls primarely because of a longer
strong
lenglh. pCM walls are expected to sustain much less and more localised damage during
ground motion than RCM walls, resulting in simpler and cheaper post-earthquake repair.
10.4 AFIALYSIS AND DESIGN. CHAPTER
the
For walls of low aspect ratio (say below 1.0), the predicted initial stiffiress overestimates
actual wall stiftress. For aspect ratios above 2, acaxate prediction is expected.
(skeleton curve) can be
Accurate and consistent estimation of the force-displacement envelope
at maximum serviceachieved, using the predicted base shear and corresponding displacement
ability moment, at nominal strength, at fust tendon yield and at overstrength.

consistThe predicted displacement capacity, du, was captured with reasonable accuracy and
by the
ency (it is acknowledged that the prediction method in the first place was calibrated
same tests). The strength
associated with d" for confined concrete masonry (CP200/CP100)
(U200) wall capacity at

was captured reasonably accurately. The unconfined concrete masoffy
d'.,
should be taken as 0.8Vr.
10.5 DYNAMIC ANALYSIS. CHAPTER 9:

A SDOF equivalent structure can be used to accurately estimate the displacement response of
of
multi-storey unbonded post-tensioned concrete masonry cantilever walls. From an array
on the funanalyses, it was found that the wall displacement demand was strongly dependent
ratio (().
damental structural period (T), the force reduction factor (R) and the viscous damping
(ct) between l% and2%Response was not sensitive to variation of the strain hardening ratio
that for
The displacement ductility demand on PCM walls was found to be much higher than
amplificareinforced concrete masonry walls of similar proportions. Likewise, the base shear
tion factor of Z.I to 2.6 recorrmended for PCM walls to account for higher mode effects, was
much higher than that recommended for reinforced concrete masonry'
-2s3 -
Displacement limitation is the primary focus of the proposed method. This makes the method
suitable for displacement focused design where emphasis is put on displacement limitation
rather than strength.
10.6 FUTTJRE RESEARCH

Experimental research:
Prediction methods described in Chapter 8 correlated well with the experimental results presented in Chapters 5, 6 and 7. Unfornmately, no experimental results were available to confirm
that the predicted dynamic response, as presented in Chapter 9, reflects reality.
It is therefore
recommended to conduct dynamic shake table (scale model) testing. The wall displacement
demand and base shear amplification are of particular interest in such investigation.
It is recommended to conduct fuither creep and shrinkage experiments. The current
database
was established under laboratory conditions and does not directly reflect the variability of the
New Zealand exterior conditions- The influence of wall weather proofing on creep and shrinkage should also be addressed.
Only one wall had energy dissipators incorporated. While the wall test wuls successful, there
remains further need to investigate this issue with regards to energy dissipator type, location
and strength.
Theoretical work:
The displacement spectra developed in Chapter 9 covered a limited number of earthquakes and
were only specific to medium soil conditions and a seismic zone factor of
Z:
1.2. To expand
the understanding and statistical reliability of the predicted PCM wall dynamic behaviour, a
larger array of earthquake records, earthquake intensities, and soil types should be investigated.
Displacement spectra reflecting additional external damping should also be developed system-
atically.
Energy dissipation arising from deformation of building floors should be investigated in detail.
Damping from that source is normally disregarded for design of reinforced concrete masonry
because of it's low magnitude in comparison with the hysteretic damping occurring in the plas-
tic hinge zone. For PCM walls with unbonded tendons that exhibit little hysteretic damping,
the energy dissipation from yielding of floor slabs may have significant impact on the seismic
displacement demand.
-254-
that the
With reference to Chapter 9 results from the detailed MDOF models, it was concluded
however, the
wall base moment was deterrrined with reasonable accurac,y by the SDOF model,
due to
base shear was wrderestimate by a factor af 2.1 to 2.6. The base shear amplification
investigahigher mode effects should be explored further given the preliminary nature of this
tion
Given the
The theoretical concepts developed in this thesis were specific to concrete masonry.
of the proposed anatysis and design method, it would be obvious to port the method
simplicity
to analysis and design of unbonded post-tensioned concrete walls.
-255 -
-256-
Appendix A
A.
DEFORMATION AT NOMINAL FLEXURAL STRENGTII
displacement of the
The structural deformation at nominal flexural strength, Mn, notably lateral
be evaluated by intetop of the wall and vertical extension and shortening of the wall ends, can
gration of wall curvature as illustrated by Fig
A.l.
The moment varies linearly along the wall
For a given axial load

height when the lateral load, Vs, is assumed applied at the top of the wall.
(p+N) or axial load ratio, 6, th" moment-curvature relationship for wall flexure can be estabstate is defined
lished, assuming that plane sections remain plane. The nominal strength limit
by the maximum usable strain in the extreme masonry fibre at the wall base, in this study
concrete
assumed to be 0.003 for unconfined concrete masonry and 0.008 for confined
in the folmasonry (refer to Chapter 3). Numerical integration of the curvature, as described
the
lowing, allows for calculation of the wall lateral and vertical displacements at the top of
wall. Evaluation of the vertical deformation enables evaluation of the tendon force increase,
possible to present the result
Ap, due to structural deformation at nominal flexural strength. It is
in a non-dimensional fashion so that any wall shape can be analysed.
It is acknowledged that a linear strain distribution inherently is associated with distributed flexural cracking, as suggested in Fig. A.2, which is consistently contradicted by testing showing
dn
t\
Mn
Curvoture
Moment
Fig.
A.l-Definition of cantilever wall dimensions,

-2s7
Def
lection
forces and deformations
Flexurol
Crocking
Rocking
Fig. A.2-Flexure / rocking cracking patterns

that only a single crack opens up at the base of the wall. Therefore one may regard the base
crack as a'summation' of the distributed cracking.
Method:
The approach consists of the following steps:
1. Definition of the stress-strain relationship for concrete masonry (o-e)
2. Selection of axial force ratio E' = :+-yJ; t*b*

3. Calculation of moment curvature relationship
(M-0)
4. Numerical integration of curvature over the height of the wall to determine the lateral displacement,
4,
and the vertical
wall end extension and shortening, u" and u, (refer to defini-
tion in Fig. A.3)

Steps 2 through 4 are repeated for a realistic ftmge of axial load ratios,
(:
0.05-0.25. Using the
above procedure for both confined and unconfined concrete masonry (different o-e relationships), a series of normalised approximate equations can be generated.
Stress-sffain relationship
The stress-strain relationships used for the derivation is shown in Fig. 3.7 for both unconfined
and confined concrete masonry, based on the Priestley-Elder relationship. It was assumed that
%rn
is limited to 0.004 for unconfined masonry. The confinement plate properties were p,
0.00577 and tr,
:240
MPa. Further detail may be found in Chapter 3.
-258-
Fig.A3-Walldeformationatnominalflexuralstrength
Moment curvature relationship
plane sections remain

The moment-curvature relationship, M-0, was calculated assuming that
relaplane and that the masonry had no tensile strength. Fig. A.4 shows the moment-curvature
the shown dimentionships for two axial load ratios, assuming confined concrete masonry and
were modified
sions. The 'Original' cuves, originating from moment curyature analysis,
was
slightly for numerical integration purposes and are denoted 'Modified'. The modification
the moment
based on a limiting moment taken as the average of the maximum moment and
for a range of
corresponding to the ultimate masonry strain. Similar curves were generated
was
axial load ratios, for both confined and unconfined masonry. The maximum curvature
basedonauseablestrainofsn'u:0.003forunconfinedconcretemasonryand0.003forconfined concrete masonrY.

Lateral displacement:
The lateral displacement was evaluated numerically, based on integration of curvature:
h"
dn:
(A.l)
ot')tl" - x\dx
in
where the curvature at any grven height, Q(x), corresponds to the moment M(x) as shown
Fig.
A.l.
-259 -
1.8
1.(D7Kf' = 17.8 MPa' Pdestley-Eldr

shasstrain rladonshlp tor confhed masonry
L = 3.0. b" = 0.1,1
1.6
E=fJf,=0.d
1.4
Ultlmrlo msimry
sttaln qil = 0,008
1.2
zx 1.0
=Ca
E
-o
E=t t?n=o,lo
0.8
0.6
0;4
o.2
0.0
0.0000
0.0100
0-(x)50
0.0150
0.0200
0.0250
Curr.rlur. (lrrrd)
Fig. A.4-Moment-curvature relationships

Vertical displacement:
The vertical displacement of each end of the wall measured at the top of the wall, as shown on
Fig. A.3, can be evaluated by integration of strain along the wall ends:
h"
t,rf
hs
ur: Je"(x)dx and ur: Jer(x)dx
(A.2)
The distribution of vertical strain along the wall ends, e"(x) and e*(x), is easily derived in con-
junction with the moment-curyature analysis.

Norrnalisation:
It
can be shown that the
ised
wall lateral displacement at nominal flexural strength can be nomral-
with
f;
"m,
h?"
(A'3)
and the vertical extensior/shortening of the wall ends
f;
h"
with
(A.4)
Eum
Only the influence of the masonry elastic modulus, En,, appears not to be entirely linear
because the deformation of the
wall section adjacent to the base for moment near Mo is rather

-260 -
csntslled by the eEfreme maEonry stain, e*,,

hased on
Rs
How.ever, ilf
wall deformafion pred[otions are
E- ln the vioinity of l000fnn Httle error'Should regult.
del
Followi4g
ane
ttB rasule of'the
rlna,V,etg, presomtsd
in gnphieal fonn in ftgs. A5 and A'6' The
pol'pomial
thick linesl.epresent iho edsulaEd rqspons aod the,'ttdn linos rc,present 2d order
approxi tions. The pqtynomial appmximatiom givenin Eqns. 8'20' 821,8'24and 8'25 were
based onpol1rotm,ial ourve
fittiryli
-261 -
qo
L4|al dltdaqenett sl qFll

Unodrodloontndarrnsqny
,0:78
db
(Ue0)
=y'(ffiLe)4F{!!li)
:'o/0
0i74
o12
O,7
i1.68
rq06
9.e{
ods o.ffi 0'gr,6 0.f
o.Jt25 0J6 017'5 0.! oa25 0!5

Affrlrirlr$Es ilf-
{I0r
!lii-y.'Fin|l,E'o50
g
1.0:56
iu
E+',
E
a;
Eo.oe
{:70
475
o.os 0.06 0.075 0.r
0.125
olq o1?5 0.t o.n6
ldrlhdntbEr
0.z5
1"J'..
,0.f0
o36
tr
o
C
F O30
tt
zo
Ior.
E
o:
9.ffi
0,1r5
0;025 0.05 0l0W
0.l
|1125
ilrlel lcr{
0.t5 0.f76 aA a'z?E 035

t{lirEr hlf'
Fig- A,S-W-all deforoa-dol at lU" for nnconffned concretemasorry
-?62 -
trEldlokflfitdwll
Efrisdhct*marunYFEmo)
4.y.
-:'-]
(fr.tL.y(Enla)
l"l
l,:l
:L
0.0415
.o6
o.oE
o.@ 0.1
E:&nslon
Oonfr
ofwil
o,'135i 0"15 0.175 O"2

rE lh.d.ffilqEr &,?i
o%
0:5
dtl4
nsd aon(rltb
m*mty (cruol
{.7
-t.l
-1.3
0,1
0.90
o.45-
oJo.
030
030
02t
o20
o.16
0.lo
G16
Fig.
Aj-lYall
defonnati6ri at l[{rfor oonfilGd ooncrctG mflsonrT
-263 -
-264 -
Appendix B
B. FORCE-DISPLACEMENT PREDICTION - CALCULATION

EXAMPLE
prediction example
This appendix presents the detailed calculations for the force-displacement
are given' Wall dimensions,
discussed in section 8.1.8. Reference to the applied equations
and 8'10' Force and
material properties, and initial stress and force are given in Figs' 8'9
area in mm2'
moment is given in kN and kNm, distance in m, stress in MPa and steel
First cracking limit state:
Eqn.8.l
, M",-
Eqn.
8.2:
Eqn.
8.4: d,,
V",
(567+931)3'6
832
lo
839kNm
83.9frN
:3ffi
.3ffi
: o.oolem
Maximum serviceability moment:

Eqn.
8.6: M" :
z.o+(o.s
,": # :
Eqn.8.r3:
- r.zr$)l.oto.19
16/
r820kNm
r82kN
a": (o.t-0.02e#)#.Tffi
t82
:0'0108n
Nominal strength:
First iteration using (n : 0.114:
Eqn. 8.25: u":0.0117 m and
ur:
-0.00384 m
Eqn.8.26:APr:10-lkN,APz:8'5kNI,AP3:7'0kN,AP4:5'5kN'AP5:3'9kN
and AP
: 35.0 kN, e, : 0.004 6 --) f^: 0.1 16
Second iteration using
(n:
0.116:
Eqn. 8.25: u.:0.0115 m and
ur:
-0.00387 m
-26s -
Eqn. 8.26: AP1
:9.8
kI\,[,
APr:8.3
andAP :34.CI kI.[*
nall3,1+
Eqn' 8'18:
Eqn 8.17:
Mo:
7r't',
\: ff
= 6.8 kh[, AP4 = 5.3 ldr|, APj
:3.8
k]-.[
q:0.-0M m --> Eo = 0.116 (OK)
197=
u=glz. rg . o.l9
(83r +
kf'.I, AP3
: o.4irm
34t(T+0.004-ry)+tt(T-ry) :
2272kNm
= Z27kN
Eqn 8.19:
dn
(7.63.0.1162-s.40.0.116+
Stness in tendon
(t +o'z)to r' :
r.eey$+9
-''14/100'3'6 s ia'+=ffi=227
firthest away from compressi,on end of wall:
(pr+Mr) _ g3l,/s+qR
4or:
______.
tr:tLs7Iu{Pa
Stress in terrdon closesttor conrpression end of,wall:
fpss
(ps + Ap5) g3l,/5 +
ff:
T:
3_B
tZI4MPe
First tendon yield:
c,: a/F:0,431/0,96 :0.449 m

Eqn 8.3r: d,y
-#ffi
where 151h*
Eqn.8.32:
m:
10/15 rnodifies E,*
o.l r72m
reflect the actual tendon length.
M,r, = (1517 -t2s7)t4o tffi=

LPffi = 32'2kN
28.lhN
M*o:
23'9kN
Mro,
-266 -
i6.4kN
o'Mr2m
LPrrr:
Eqn.8.33: LP,
19.8&N
l40.4kN
Eqn. 8.35:
*rt + 140 +.591

o, = 6.972.tg. 0.19 :
Eqn. 8.34:
M,,
Eqn. 8.36:
M, :
36.4(3-6/2+ 0.4) ...- 19.8(3.6
(831 + 34 +
vr: 'ry
=ro
Eqn. 8.37:
Eqn. 8.39:
P,
56?)(+
/2-O'4)-ry 40 :
-ry)*
zzs
229kNm
: 2475kNm
:248kN
0.0+t + 0j172 = 0'I58 m
4:
Overstrength:
0.463m
: 5' 140 ' l5l7 :
l062kN
1062+567
f,
-o _ 0.972. 18.0.19. =o.4gom
: ( 1062 + s6T(+- ry)
Eqn. 8.38: Mo
,,533 :
vo:rro
Eqn. 8.40:
a"
2533kNm
253kN
: ffi
#=
o.3lon
Ultimate displacement capacity:
Assume:
":t
o'4eo)
\{ar+ ao)
lto'out+
:tT:0.496nt
states is minimal
The variation of c between the first tendon yield and overstrength limit
because of
little axial force variation.
ozofro-ry)
Eqn.8.43:
du: #0.013
A]92m
-267 -
The wall strength at du is found by interpolation between first tendon yield and overstrength
limit states with respect to displacemsnt:
vu: vy.ffirdu-dy) = 248.0ffi
-268
Q.r)2-0.158)
249kN
Appendix C
C. COMPARISON OF PREDICTED VS. EXPERIMENTAL

BEHAVIOUR
C.l FLEXIJRE/ROCKING
PREDICTION METHOD
present comparison'
Only fully grouted walls that failed in flexure are incorporated in the
C.1.1 AssumPtions
f''' of 16 MPa'
Following assumptions were made: minimum unconfined masonry stength,
of 200 mm for the single storey walls and 400 mm for the three storey
plastic zone height,
\,
and
walls, ultimate masonry compression strain, tmu, of 0.020 for U200 and CP200
5,6 andT '
0.013 for CPl00 masonry. All other wall details may be found in Chapters
e,,,u
of
predictions of the wall force-displacement characteristics were carried out according to the forfor FG-L3.0-W15-P2mulae presented Chapter 8. The only deviating assumption was made
Mn was based on
CPC that had constant force applied to the preshessing barso consequently
were not calcuN+p and the base shear and displacement at first tendon yield and overstrength
(K: 1.08),
lated. All walls in Series 2 and 3 were confined with CP200 (K: 1.04) or CP100
blocks incorporated in
except for FG-L3.0-P2C-HB from series 2 which had one high strength
masonry properties.
each wall corner and was calculated using CP200 confined concrete
C.1.2 Comparison
and response for loadEach figure in the following shows comparisons of the global response
describes the
ing up to nominal shength. In these curyes, the ttrick line connecting the markers
the following
predicted response (envelope/back bone curve) with the markers representing
first tendon yield, diamond: nominal strength'
overstrength, triangle
states: square
limit
solid
circle: maximum serviceability, hexagon: first cracking and circle: ultimate
displace-
are shown in thin

ment capacity. The experimentally determined force displacement curves
line.
-269 -
C.l.z.l
Series
I walls
Figs. C.l-C.5 show comparison of predictions vs. recorded force-displacement behaviour for
Series
I testing of unconfined
single storey walls.
In terms of global measured response vs. prediction, it is clear from Figs. C.1-C.5 that there is
good correlation between experiments and predictions at the maximum serviceability, nominal
strength and tendon yield limit states.
In terms of wall response beyond tendon yield, it is seen that the predicted strength at the predicted wall displacement capacity (circle located on the prediction curve) was only achieved
for one wall, notably FG-L3.0-W20-P3, which had a relatively low axial load applied to it. A
lower bound strength at wall displacement capacity, do, was estimated using
1.0
in for calculation of the wall flexural compression zone length
o:
cu (centroid
0.5 and F
of compressive
stress located at a distance of 2/3cu away from the extreme compression fibre). These results are
shown with the circular markers slightly offset vertically from the circles on the prediction
curves.
ln terms of the predicted wall displacement capaciry it is seen from Figs. C.l-C.5 that initiation of strength degradation generally is captures well. Only wall FG-L3.0-W20-P3 showed
sigrrrifisanlly large displacement capacity than predicted. This wall differed from the other
walls in series 1 in the sense that the maximum wall axial force due to two yielding tendons
(third tendon located in the wall compression zone and had nearly lost all prestress at that
stage) was applied to a larger wall section, thus resulting in a shorter compression zone and a
proportionally lower extreme masonry fibre strain.

It is seen from the low level plots in Figs. C.1-C.5 that the predicted wall initial stiffness represented by the
'first cracking' limit state (hexagonal marker) in all
cases exceeded the measured
initial stiffness. This discrepancy is mainly attributed to inadequacy of the plane section
assumption to describe deformation of walls of low aspect ratio. This observation is confirmed
by studying the low level response of the wall withthe highest aspect ratio, FG-1.8-Wl5-P2, in
Fig. C.5 which indicates an accurate prediction of the measured initial stiffiress.
-270 -
Drift (%)
-0.5
z
j
0.0
600
1500
400
1000
200
500 ;c
og o
(U
.c,
U)
o
6
o
dl
tl<
-500
-200
-1000
400
dl
-1500
-600
-50
40
-30 -20 -10 0
10 20 30 40
50
Lateral DisPlacement (mm)

Global resPonse
Drift (%)
-0.1
0.0
0.1
o.2
400
1000
500
200
Y
$
go
o
o
-500
-200
co)
E
o
o
o
(u
[D
-1000
-400
Irtlll
56
-7-6-54-3-2-101234
Low level response
Fig.
z.v,E
C.l-series 1 FG-L3.0-W20-P3: Prediction

-27t -
vs. experimental behaviour
600
1500
4oo [-
z.Y
1000
E
z
500 5
200
.f
o
o
.c,
U)
oEco
fl
o
o
-roo
(E
(D
-5oo
I
400
-1000
|-
-600
fi=
-1500
-20 -10
40
10
20
30

Global response
Drift (%)
600
-o.2
-0.1
0.0
0.1
0.2
1500
400
z
5
1000
500 5
200
fl
go
(E
E
dl
ofl o
il,
U)
e
z
-qnn
-200
c)
u,
-1ooo S
-400
-1500
-600
-7-6-54-3-2-101234
67

Low level response
Fig. C.2-Series I FG-L3.0-W15-P3: Prediction vs. experimental behaviour
-272-
Drift (%)
-0.s 0.0
0.s
1200
400
900
300
zv
600
200
300
100
o
o
o
a
o
dl
-100
-300
-200
-600
-300
-900
400
-1200
-50
-40
-30 -20 -10 o
10 20 30
40
z.YE
C
o
E
o
o
U,
o
dl
50

Global response
Drift (%)
-o.2
-0.1
0.0
0.1
o.2
1200
400
900
300
?
-v
Y
600
2oo
300
1oo
go
(U
a
o
o
-100
-300
-2oo
-600
-300
-900
400
t-
-7 -6 -5 4 -3 -2 -1 0'l
2 3 4 5 6
z-vE
co
E
o
o
(s
dl
-1200
7

Low level response
Fig.
c.3-series I FG-L3.0-W15-P2C: Prediction

-273 -
Drift (%)
-1.2
-0.8
-0.4 0.0
0.4
1200
400
300
zl.
(E
|_
900
[
200
600
l-
300
100
or
E
!
z
c
o
E
o
U)
-100 lo
a
(U
dl -roo
-300
-600
[
400
-900
-300
o
.n
o
(D
-1200
-20 -10 0
40
10
20

Global response
Drift (%)
-o.2
-0.1
0.0
0.1
0.2
1200
400
2
.Y
;
300
900
2oo
600
1oo
300
go
(u
z.YE
c
o
E
o
U)
o
o
-100
-300
-2oo
-600
-300
-900
400
rrtrll
t-
-7-6-54-3-2-101234
o
o
(g
dl
-1200
67

Low level response
Fig. C.4-Series 1 FG-L3,0-W15-P2E: Prediction vs. experimental behaviour
-274-
200
450
150
300
2
.Y
;so
1oo
150
-50
-150
-1oo
-1
;c.
ogo
eo
G
o
o
zl<E
-3oo
50
o
o
fi
-450
-200
-50 40 -30 -20 -10 0
10 20 30
40

Global response
Drift (%)
-o.2
-0.1
0.0
0.1
o.2
200
450
150
300
100
2
lz
;50
o
150
-50
-150
S -1oo
-3oo
-150
-450
-200
v
;c
oflo
eo
o
a
tttttl
tttttl
-7-6-54-3-2-101234
67

Low level response
Fig. C.S-Series I FG-L1.8-W15-P2: Prediction vs. experimental behaviour
-27s -
o
U'
C.1.2.2 Series 2 walls

Figs. C.6-C.8 show experimental results and predictions for three (four) single storey walls
with strengthened corners.
With regards to global measured response vs. prediction, it appears from Figs. C.6-C.8 that
there is good correlation between experiments and predictions at the maximum serviceability,
nominal strength, overstrength and tendon yield limit states.
In terms of wall response beyond the overstrength limit state, it is seen that the predicted
strength at the predicted wall displacement capacity, d.,, (circle located on the prediction curve)
was achieved for all walls in this test series. The lower bound strength at wall displacement
capacity, du, estimated as 0.8Vo, appears to significantly underestimate the wall shength.
In terms of the predicted wall displacement capacity, it is seen from Figs. C.6-C.8 that initiation of strength degradation generally is captures reasonably consistently.
It is seen from the low level plots in Figs. C.6-C.8 that the predicted wall initial stiffiress represented by the
'first cracking' limit state (hexagonal marker), also for the confned walls, in all
cases exceeded the measured initial stiffiress. As was explained above, this discrepancy is
attributed to inadequacy of the plane section assumption to describe the deformation of walls
of low aspect ratio.
C.1.2.3 Series 3 walls

Figs. C.9 and C.l0 show experimental results and predictions for the two 3-storey walls with
strengthened corners.
With regards to global measured response vs. prediction, it appears from Figs. C.9 and C.10
that there is good correlation between experiments and predictions at the maximum servicea-
bility, nominal strength and tendon yield limit states. It is observed in these figures that the
measured strengths were considerably higher than the predicted strength. This was attributed to
friction in the testing setup. The overstrength limit state was not reached in any of the Series
wall tests.
-276-
Drift (%)
-0.5 0.0
0.5
1200
400
900
300
z-g
(g
o
E
U)
o
(t,
$
dl
600
200
300
100
0
-100
-300
-200
-600
-300
-900
-400
-1200
-50
-40
-30 -20 -10 0
10 20
30
40
zl4E
co
E
o
o)
(t
$
dl
50

Global response
Drift (%)
-o.2
-0.1
0.0
0.1
o.2
1200
400
900
300
zoo
2
lz
Y
o
go
600
U)
-100
-300
-2oo
-600
-300
-900
300
1oo
o
o
fi
z.YE
o
E
o
o
o
(g
@
400
t:
-7 -6 -5 4 -3 -2 -1 0 1 2 3 4 5 6
-1200
7

Low level response
behaviour
Fig. c.6-series 2 FG-L3.0-W15-P2-CP: Prediction vs. experimental
-277 -
Drift (%)
0.0 0.5 1.0 1.5 2.0
-2.5 -2.0 -'1.5 -1.0 -0.5
2.5
900
300
z.Y
t-
200
600
100
300
o
E
-100
(n
U)
(E
z.YE
-300
o
E
o
o
@
tU
(D
-200
-600
-300
-900
-80 -70 -60 -50
40 -30 -20 -10
10 20 30 40 50 60 70
80

Global response
Drift (7o)
-0.1
0.0
0.1
900
300
.:C
o
o
E
U)
o
(t
200
600
100
300
0
-100
-300
-200
-600
-300
-900
(U
(D
zJE
co
o
o
o
o
dl
-7-6-54-3-2-101234
Low level response
Fig. C.7-Series 2 FG-L3.0-W15-P2-CP-CA: Prediction vs. experimental behav.
-278-
Drift (%)
-1.5
-1.0
-0.5 0.0
1,0
0.5
1.5
1200
400
900
300
?
v
;o
600
2oo
300
1oo
eo
U'
o
o
-100
-300
-zoo
-600
-300
-900
-400
-1200
z.YE
co
o
o
a
o
trl
-50 40 -30 -20 -10 0
10 20
30
40
50

Global response
Drift (%)
-o.2
-0.1
0.0
0.1
0.2
1200
400
900
300
2
.Y
;o
600
2oo
300
1oo
eo
a
o
o
|/
-100
-300
-2oo
-600
-300
-900
400
t:
ltlll
-7 -6 -5 4 -3 -2 -1 0 1 2 3 4 5 6
z-vE
=
o
E
o
o
a
o
dl
-1200
7

Low level response
Fig. c,g-series 2 FG-L3.0-W15-P2-IIB: Prediction vs. experimental behaviour
-279 -
In terms of wall response beyond the overstrength limit state, it is seen that the predicted
strength at the predicted wall displacement capacity,
d, (circle
located on the prediction curve)
was achieved for both walls in this test series. This was expected because the prediction model
was calibrated with these two wall tests. The lower bound strength at wall displacement capac-
ity,
4,
estimated as 0.8Vu, appears to significantly underestimate the wall strength. The initia-
tion of strength degradation generally was captures well.
It is seen from the low level plots in Figs. C.9 and C.10 that the predicted wall initid stiffiress,
represented by the 'first cracking'
limit state (hexagonal marker), in both cases captured the
actual resonse.
C.13
Conclusion
The following conclusions were drawn from comparison between predictions using the procedures developed in this chapter and experimental results.
Initial stiffiress:
For wall of low aspect ratio (say below 1.0), the predicted initial stiffiress based on V", and d".
overestimates the actual wall stiftress by a factor of the order of 1.5. For aspect ratios above 2,
accurate prediction is expected.
Global response:
Using the predicted base shear and displacement at maximum serviceability moment, nominal
strength, first tendon yield and overstrength, an accurate estimation of the force-displacement
envelope (skeleton curve) can be achieved.
Displacement capacity:
The predicted displacement capacity,
d,.,,
was captured with reasonable accuracy using the val-
ues for extreme fibre strain and plastic zone length given in Thble 8.1. The strength associated
with d" for CPl00 confined concrete masonry can be based on Vu. For CP200 confined
masonry and U200 unconfined masonry,
it
appears prudent to
limit the wall stength (base
shear) at d" to 0.8%. The maximum strength predicted at du should be limited by Mo.
-280 -
Drift (%)
-0.s 0.0
250
0.5
1200
200
900
150
z-v
600
100
300
(0
50
-c
o
a
(u
-50
-300
-100
-600
-150
-900
-200
-1200
dt
z.vE
c,
o
E
o
o
o
o
dl
-250
----90.80.70-60-50.40.30-20-100102030405060708090

Global response
Drift (%)
150
-0.09
-0.06
-0.03 0.00
0.03
0.06
0.09
600
100
950
o
go
z.YE
200
co
-200
o
a
(E
400 (Il
ffi -50
dl
-100
-600
-150
-5
-3-2-10123
Low level response
Fig.
400
c.r-Series 3 s3-l: Prediction
-28r -
Drift (%)
200
-2.O -1.5 -1.0 -0.5
0.0
0.5
1.0
1.5
1200
150
Z
J
900
100
z.YE
600
:50
(E
300
go
tr
o
o
E
o
-50
-300
lo -1oo
-600
(g
-900
-150
-200
o
U'
-1200
-1 20
-100
-80
-60 -40 -20 0 20 40 60
80

Global response
Drift (%)
125
-0.09
-0.06
-0.03 0.00 0.03
0.06
0.09
600
100
75
400
z.
50
200
(u
25
so
o
U)
-25
(u
co
g
y
ogco
-2oo
-50
u)
ct
dl
-400 =
-75
-100
-125
e
z
-600
-3-2-10123
Low level response
Fig. C.l0--Series 3 S3-2: Prediction vs. experimental behaviour
-282

PHD Laursen-Masonry Prestressed Walls

Uploaded by

Copyright:

Available Formats

PHD Laursen-Masonry Prestressed Walls

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

PHD Laursen-Masonry Prestressed Walls

Uploaded by

Copyright:

Available Formats

What is the thesis exploring?

What is the thesis exploring?

What did the literature review find about post-tensioning of masonry?

What did the literature review find about post-tensioning of masonry?

http://researchspace.auckland.ac.

Any use you make of these documents or images must be for

To request permissions please use the Feedback form on our webpage.

Seismic Analysis and Design

Post-tensioned Concrete Masonry Walls

A thesis submitted in partial fulfilment of the requirements

Supervised by Dr. Jason M. Inghant

It was concluded that structural testing of PCM walls

of this construction type.

it is essential to use high strength steel for

reduce long term losses.

severe ground shaking from an earthquake.

suggested that partially and

ungrouted PCM walls may suitably be used in strength design (non-ductile).

ties to this project.

testing setup, and

in testing of t'wo walls reported herein are acknowledged.

Dr. J.M. Ingham, SeniorLecturer (principal advisor)

Mr. E. Lapish, ConsultantEngineer

Assistant Professor C.E. Ventura, Department

of Civil Engineering, University of British

Columbia, Canada (chair)

The Masonry Society, Boulder, Colo-

England, September 2002, paper reference247.

Laursen, P.T. and Ingham, J.M. (2001), Material prcperties

New Zealand concrete

Laursen, P. and Ingham, J. M. (2001), Seismic Resistance of Prestressed Concrete Masonry

of l2th lnternational Brick/Block Masonry Conference, p. 905, Madrid,

of 12ft International Brick/Block Masonry Confer-

Spain, October 2000.

Laursen, P. and Ingham, J. M. (2000), Cunent Research on Post-tensioned Concrete Masonry

land" October 2000.

Laursen, P. and Ingham, J. M. (1999), Ductile response of post-tensioned concrete masonry

Belgium, July 1995,

Simple Wall Confgurations, School of Engineering Research

Tbsting of Dricon Superset

PRINCIPLE OF UNBONDED POS}TENSIONED WALLS

Code rules and

THEORETICAL STRESS-STRAIN RELATIONSHIP......... ....... 33

TIME DEPENDENT EFFECTS

4.5.1 Typical losses - additive approach

STRUCTURAL TESTING. SERIES I

5.5.1.6 Sliding propensity

5.5.2 Partially and ungrouted walls...

CONCLUSIONS AND FUTURE

STRUCTURAL TESTING - SERIES 2

ENHANCED SINGLE-STOREY PCM

STRUCTURAL TESTING. SERIES

7.4.5 External axial force

ANALYSIS AND DESIGN

PREDICTION OF WALL IN-PLANE BEHAVIOUR............ .....191

8.2.2 Un-cracked section analysis.....

EQUIVATENT STNGLE-DEGREE OF FREEDOM STRUCTURE ........... 226

CONCLUSIONS AND FUTURE RESEARCH.............. ...........249

DEFORMATION AT NOMINAL FLEXURAL STRENGTH.....................257