Seismic Retrofit of Masonry

PHDDissertation
Seismic retrofit of masonry

with innovative materials for
strengthening and repair
Doctoral Thesis by:

Larisa Garcia Ramonda
Supervisors:
Prof. Pere Roca Fabregat
Prof. Luca Pelà
Barcelona, November 2020
Universitat Politècnica de Catalunya, Barcelona Tech
Doctorat en Enginyeria de la Construcció
Departament d´Enginyeria Civil i Ambiental

Seismic retrofit of masonry with
innovative materials for
strengthening and repair
Doctoral Thesis submitted in fulfilment of the requirements for the

Degree of Doctor of Philosophy in Construction Engineering
by
Larisa Garcia-Ramonda
Thesis Supervisors:
Prof. Pere Roca Fabregat
Prof. Luca Pelà
Barcelona, November 2020
Universitat Politècnica de Catalunya

Departament d’Enginyeria Civil i Ambiental
Programa de Doctorat en Enginyeria de la Construcció
To my family
Acknowledgements
This thesis would have not been possible without the predoctoral grant provided by Secretaria
d’Universitats i Investigació de la Generalitat de Catalunya and the financial support from the Min-
istry of Economy and Competitiveness and from the Ministry of Science, Innovation and Universities
of the Spanish Government, as well as that of the ERDF (European Regional Development Fund)
through projects MULTIMAS (Multiscale techniques for the experimental and numerical analysis of
the reliability of masonry structures, ref. num. BIA2015-63882-P) and SEVERUS (Multilevel eval-
uation of seismic vulnerability and risk mitigation of masonry buildings in resilient historical urban
centres, ref. num. RTI2018-099589-B-I00) and the support of Kerakoll Spa through the RTD project
“Diagonal Compression Tests on Unreinforced and TRM Reinforced Masonry Walls” (ref. num. A-
01217).
I would like to gratefully acknowledge my supervisors Prof. Pere Roca and Prof. Luca Pelà for
giving me this opportunity, for their invaluable help, support and endless encouragement during all
these years. Thank you for sharing with me your passion, enthusiasm, and motivation.
The experimental programme was carried out at the Laboratory of Technology of Structures and
Construction Materials (LATEM) at the Technical University of Catalonia (UPC-BarcelonaTech). I
would like to thanks Jordi Cabrerizo, for his dedication and all the hours invested in the laboratory
teaching me. I would also like to thanks Carlos Hurtado, Robert McAloon and Tomas Garcı́a for their
patience and for being by side during this extensive experimental campaign. Thanks you for making
it possible.
I would like to sincerely thank Jorge and Nirvan, for being there every step of the way, for their
constant help and advice, and most importantly for being my family in Barcelona. We started this
adventure together and even though I am thrilled that we are finishing it together, I already know
how much I am going to miss you. A special thanks for Belen, Philip, Savvas, Sara, Camilla, Janill,
Francesco and Albert, for all the experiences and memories shared during all these years. Thank you
for your love and joy. This journey would not have been the same without all of you.
All my gratitude and love is for my family in Argentina, my parents Susana and Norberto and my
sisters Juliana and Magali for their invaluable support despite the distance, for encouraging me to be
better and pursuit my dreams.
Finally, I would like to thanks Ulric for his unconditional love, support, and constant encourage-
ment. Thank you for being there for me. Thank you for believing in me even when I don’t.
Barcelona, 26 de Noviembre 2020

Larisa Garcia-Ramonda
i
Abstract
As one of the main historical construction materials, masonry is abundant among the architectural
heritage of earthquake-prone areas of the Mediterranean countries. Earthquake mitigation approaches
are now focusing on strengthening solutions based on compatible and environmentally friendly repair
materials, such as Textile Reinforced Mortar (TRM). These solutions have recently evidenced their
efficiency to improve the in-plane lateral strength and displacement capacity of masonry buildings,
which are the most significant parameters considered in the seismic assessment. Notwithstanding the
increasing interest risen by the potentiality of TRM strengthening systems, limited research has been
done about the structural response of TRM applied on historical masonry as a seismic retrofitting or
post-earthquake repair.
This thesis presents an extensive experimental programme on masonry walls composed of hand-
made clay bricks and low strength lime-based mortar, a recurrent typology for historical buildings.
The walls were strengthened with four different TRM strengthening solutions, and one joint repoint-
ing solution with Near Surface Mounted (NSM) helical rebars. The main aim of the experimental
campaign was to deepen the understanding of the in-plane behaviour of strengthened masonry and
to accurately capture the global and local response of strengthened masonry walls, emphasising on
the identification of damage levels and the influence of different configurations of TRM strengthening
solutions on the in-plane response.
Diagonal Compression Tests (DCT) and quasi-static Shear Compression Tests (SCT) were per-
formed to investigate the in-plane response of strengthened masonry and validate TRM as a seismic
retrofitting technique. In addition, a post-earthquake repair methodology has been also investigated.
The methodology consisted on repairing damaged unreinforced masonry walls with different tech-
niques, such as crack repointing and ”scuci-cuci”, and strengthened them with basalt TRM strength-
ening system. The experimental results show the suitability of the proposed solutions and highlight the
effectiveness, to different extents, of the investigated systems in increasing the resistance and ductility
of unreinforced brick masonry. The results also show that the presence of TRM strengthening systems
has a significant impact on the failure mode of the walls and hence, the displacement capacity, energy
dissipation, strength and stiffness degradation of the global response. In turn, TRM strengthening
systems have not significantly influenced the initial stiffness and damping. The experimental results
were compared to the analytical ones assessed with the available formulations provided by the Ital-
ian guideline CNR-DT 200 (2013). The comparison evidenced good agreement; however, particular
attention has to be paid, when defining the parameters, to the failure mode associated to the TRM
strengthening systems.
Finally, a methodology is proposed that pursuits the correlation between both testing methodolo-
gies, DCT and SCT, to effectively assess the in-plane parameters that govern the masonry’s behaviour
in both unreinforced and strengthened configurations. From the correlations investigated, it became
clear that the in-plane response can be characterized by means of an equivalent tensile strength that
takes into account the joint contribution of the masonry and reinforcement. This correlation shed
light on the interpretation of DCT while validating it as an affordable testing methodology to evaluate
the tensile strength of masonry.
Keywords: Masonry, TRM, Diagonal Compression Test, Shear Compression Test, post-earthquake
repair, seismic retrofitting, correlation study, analytical study, equivalent tensile strength.
iii
Resumen
La mamposterı́a de ladrillo es abundante entre el patrimonio arquitectónico ubicado en las zonas
sı́smicas de muchos paı́ses mediterráneos. Las nuevas técnicas de refuerzos sı́smico se centran en
soluciones basadas en materiales compatibles con la mamposterı́a y respetuosos con el medio ambiente,
como el mortero reforzado con textiles (TRM). Estas soluciones han demostrado su eficiencia en la
mejora de los parámetros mas significativos en la evaluación sı́smica: la resistencia a fuerzas en el
plano y la capacidad de deformación de los elementos de mamposterı́a. A pesar del creciente interés
suscitado por la potencialidad de los sistemas de refuerzos basados en el TRM, poca investigación se
ha realizado sobre la respuesta estructural de elementos de mamposterı́a histórica reforzada.
Esta tesis presenta un extenso programa experimental realizado en muros de mamposterı́a de
ladrillos de arcilla hechos a mano y mortero de cal de baja resistencia, una tipologı́a recurrente
en edificios históricos. Los muros fueron reforzados con cuatro soluciones diferentes, tres de ellas
basadas en TRM y un refuerzo de juntas de mortero con barras helicoidales. El objetivo principal
de la campaña experimental es profundizar la comprensión del comportamiento de la mamposterı́a
reforzada sometida a cargas laterales y capturar la respuesta global y local de los muros de mamposterı́a
reforzados, haciendo hincapié en la identificación de los distintos niveles de daño y la influencia de las
diferentes configuraciones de refuerzo TRM.
La campaña experimental involucró ensayos de compresión diagonal (DCT) y de corte compresión
(SCT) para investigar la respuesta en el plano de la mamposterı́a reforzada y validar las soluciones con
TRM como una técnica de refuerzo frente a acciones sı́smicas. Además, se ha investigado al TRM como
una técnica de reparación de daños en muros de mamposterı́a no reforzada con diferentes técnicas de
reparación, como el rellenado de fisuras y la técnica “scuci-cuci”. Posteriormente a la reparación, se le
aplica un refuerzo TRM con textil de basalto. Los resultados experimentales demuestran la idoneidad
de las soluciones propuestas y destacan la efectividad de los distintos sistemas de refuerzos investigados
con el fin de aumentar la resistencia a acciones laterales, ası́ como la ductilidad de la mamposterı́a
no reforzada. Los resultados también demuestran que la presencia de refuerzo TRM tienen un gran
impacto en el modo de falla de los muros, lo que también implica una gran influencia en la capacidad
de desplazamiento, de disipación de energı́a, de degradación de la resistencia y rigidez de la respuesta
global. Al mismo tiempo, los sistemas de refuerzos TRM no han demostrado influencia en la rigidez
inicial y el amortiguamiento de los muros. Los resultados experimentales fueron comparados con los
resultados analı́ticos obtenidos al aplicar las formulaciones disponibles en la directriz italiana CNR-DT
200 (2013). La comparación presenta buenos resultados; sin embargo, al definir los parámetros de la
formulación se debe prestar especial atención al modo de falla asociado a cada sistema de refuerzo
TRM.
Finalmente, se propone una metodologı́a que tiene por fin la correlación entre ambas metodologı́as
de ensayo, DCT y SCT, para evaluar de manera efectiva los parámetros que gobiernan el compor-
tamiento en el plano de la mamposterı́a tanto en configuración reforzada como no reforzada. De
la correlación queda claro que la respuesta a acciones en el plano de la mamposterı́a reforzada esta
definida por la resistencia a tracción equivalente que tiene en cuenta la acción conjunta de la mam-
posterı́a y del TRM. Esta correlación ayuda a mejorar la interpretación del ensayo DCT al mismo
tiempo que lo valida como una metodologı́a de ensayo simple y económica para evaluar la resistencia
a tracción de la mamposterı́a.
Palabras claves: Mamposterı́a, TRM, ensayo de compresión diagonal, ensayo corte compresión,
reparación post-sismo, refuerzo sı́smico, estudio de correlación, estudio analı́tico, resistencia a tracción
equivalente.
v
Contents
Acknowledgements i
Abstract iii
Resumen vi
Table of Contents vii
List of Figures xi
List of Tables xvii
List of Symbols xxiv
1 Introduction 1
1.1 Background and Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Scope and Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.3 Assumptions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.4 Research Dissemination . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.5 Outline . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2 Literature Review 9
2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.2 In-plane shear behaviour of unreinforced masonry . . . . . . . . . . . . . . . . . . . . . 10
2.2.1 Shear Resistance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.2.2 Flexural Resistance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2.3 Textile reinforced mortar . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.3.1 Behaviour of TRM under tensile stress . . . . . . . . . . . . . . . . . . . . . . . 16
2.3.2 Bond strength of TRM on Masonry substrate . . . . . . . . . . . . . . . . . . . 20
2.4 In-plane shear behaviour of strengthened masonry . . . . . . . . . . . . . . . . . . . . 28
2.4.1 Diagonal compression test (DCT) . . . . . . . . . . . . . . . . . . . . . . . . . . 29
2.4.2 Shear Compression Test (SCT): Cyclic in-plane loading . . . . . . . . . . . . . 37
2.5 Experimental design assessment of specimens retrofitted with TRM . . . . . . . . . . . 41
2.6 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
vii
3 Experimental Programme 47
3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
3.2 Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
3.2.1 Brick characterisation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
3.2.2 Mortar characterisation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
3.2.3 TRM characterisation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
3.2.4 Bond behaviour of TRM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
3.3 Construction of specimens . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
3.4 Repair and retrofitting of specimens . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
3.4.1 Strengthening . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
3.4.2 Scuci-cuci . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
3.5 Test summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
3.6 Diagonal Compression Test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
3.6.1 Set-up . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
3.6.2 Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
3.7 Shear Compression Test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
3.7.1 Set-up . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
3.7.2 Test procedure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
3.7.3 Loading Protocol . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
3.8 Digital image correlation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
4 Experimental Results: Diagonal Compression Test 77

4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
4.2 Experimental results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
4.2.1 Unreinforced samples and specimens retrofitted with basalt TRM . . . . . . . . 78
4.2.2 Specimens retrofitted with steel TRM . . . . . . . . . . . . . . . . . . . . . . . 82
4.2.3 Asymmetrically reinforced specimens with basalt TRM and bed joints repointing 83
4.3 Comparison of solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
4.4 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
5 Experimental Results: Shear Compression Test 91

5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
5.2 Experimental Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
5.2.1 Failure modes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
5.2.2 Force-displacement curves . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
5.2.3 Bilinear modelling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
5.2.4 Ductility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
5.2.5 Secant Stiffness . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
5.2.6 Stiffness degradation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
5.2.7 Limit States . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
5.2.8 Energy Dissipation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119
5.2.9 Damping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
viii
5.3 Comparison of solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125
5.4 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
6 Correlation between Diagonal Compression Test and Shear Compression Test 131
6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131
6.2 Interpretation of testing methodologies . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
6.3 Proposed integrated methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
6.3.1 Correlation between Diagonal Compression Test and Shear Compression Test . 135
6.3.2 Correlation between shear response and tensile strength . . . . . . . . . . . . . 143
6.4 Procedure for the application of the proposed methodology . . . . . . . . . . . . . . . 147
6.5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148
7 Analytical Study 151

7.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151
7.2 Analytical Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152
7.2.1 Available TRM guidelines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152
7.2.2 Analytical model for NSM joint repointing . . . . . . . . . . . . . . . . . . . . 153
7.3 Analytical Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154
7.3.1 Shear contribution of masonry . . . . . . . . . . . . . . . . . . . . . . . . . . . 154
7.3.2 Shear contribution of TRM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156
7.3.3 Efficiency of the TRM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164
7.4 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166
8 Conclusions 167
8.1 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167
8.2 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168
8.2.1 Experimental Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168
8.2.2 Correlation between Diagonal Compression Test and Shear Compression Test . 170
8.2.3 Investigation on analytical approaches . . . . . . . . . . . . . . . . . . . . . . . 171
8.3 Future work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172
A Use of DIC in experimental testing of masonry walls 175
Bibliography 179
ix
x
List of Figures
2.1 Typical failure modes of masonry walls, subjected to in-plane loading according to
Tomaževič (1999) a) Shear failure mechanism: shear sliding along bed joint, b) Shear
failure mechanism: shear failure due to diagonal cracking, c) Flexural failure mechanism:
rocking and crushing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.2 Stress state on the brick due to shear-compression loads as proposed by Mann and
Müller (1980). (Graphical representation taken from Calderini et al. (2010)) . . . . . . 13
2.3 Rigid block model for the flexural mechanism due to rocking according to Magenes and
Calvi (1997) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2.4 Tests set-up for the mechanical characterisation of TRM strengthening systems accord-
ing to de Felice et al. (2020): a) Direct tensile test on TRM composite specimens, b)
shear bond test on brick masonry substrate . . . . . . . . . . . . . . . . . . . . . . . . 16
2.5 Tensile response of TRM coupons subjected to direct tensile test according to De Santis
et al. (2017a) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.6 Overview of test setups for TRM tensile tests: a) The clamping method used by De
Santis and De Felice (2015) was aluminium tabs glued by means of FRP, b) De Santis
et al. (2017b) used a direct clamping method to the TRM coupon, c) Caggegi et al.
(2017) used a biaxial machine and directly clamped the TRM coupon . . . . . . . . . 19
2.7 Failure modes in shear bond tests on TRM/FRCM systems defined by the TC RILEM
250 CSM De Santis et al. (2017a): (A) debonding with cohesive failure in the substrate,
(B) debonding at the TRM-substrate interface, (C) debonding at textile-to-matrix in-
terface, (D) sliding of the textile within the reinforcement thickness, (E) tensile rupture
of the textile, (F) tensile rupture of the textile within the mortar matrix (image repro-
duced from Ascione et al. (2015)) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
2.8 Failure modes observed on the shear bond tests perfomed by Ascione et al. (2015): a)
Basalt TRM was characterised by fibre rupture (E), b) SRG experienced debonding
failure at textile-matrix interface (C) . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
2.9 Failure modes observed in the shear bond test performed in De Santis and de Felice
(2015): debonding at the mortar-to substrate interface (mode B) and shear failure at
the textile-to-matrix interface (mode C) . . . . . . . . . . . . . . . . . . . . . . . . . . 24
2.10 a) Set-up used by Santandrea et al. (2020) to performed the shear bond test on stack
bond prisms, b) set-up used by Ascione et al. (2015), De Santis and de Felice (2015) to
performed the shear bond test on single bricks . . . . . . . . . . . . . . . . . . . . . . 25
xi
2.11 Failure modes according to Santandrea et al. (2016): Low density basalt fibre failure
(on the left), and debonding mechanism for high density basalt (on the right). . . . . . 26
2.12 Single shear bond test set-up used by Barducci et al. (2020) . . . . . . . . . . . . . . . 28
2.13 Overview of the different set-ups used by the authors aforementioned: a) Prota et al.
(2006), b) Balsamo et al. (2011), c) Parisi et al. (2013), d) Ismail and Ingham (2013),
e) Babaeidarabad et al. (2014) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
2.14 a-b) Maximum and minimum principal stress results from FEM analysis performed by
Basili et al. (2019), c) Crack pattern at the end of the test of the specimen symmetrically
reinforced . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
2.15 Overview of results of DCT in terms of crack patterns and strain-stress experimental
curves: a) Del Zoppo et al. (2019), b) Wang et al. (2018), c) Casacci et al. (2019) . . . 36
2.16 Three approaches followed to simulate the double-fixed boundary conditions: a) Petry
and Beyer (2015), b) Magenes and Calvi (1992), c) Magenes et al. (2008) . . . . . . . 37
2.17 Shear compression test set-up used to analyse the in-plane behaviour of strengthened
stone masonry walls: a) Gattesco et al. (2015a), b) Godio et al. (2019) . . . . . . . . . 39
2.18 Test set-up for in-plane loading shear walls in Papanicolaou et al. (2011) . . . . . . . . 40
3.1 Elaboration of standardised mortar samples: a) pouring of the material inside the
moulds and compact by stokes of a tamper, b) storing of the moulds after skimming off
the excess of material, c) final look of the sample . . . . . . . . . . . . . . . . . . . . . 49
3.2 Test carried out on binding mortar samples: a) three point bending test, b) compressive
test, c) double punch test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
3.3 Types of textiles used for the wall’s strengthening: a) Bidirectional low density basalt
textile (LDB), b) Helical stainless steel rebar (JR), c) Unidirectional low density steel
textile (LDS), d) Unidirectional medium density steel textile (MDS) . . . . . . . . . . 52
3.4 Failure modes evidenced in the single-lap shear bond test on SRG composite for different
bonded lengths . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
3.5 Identification of SGR effective bonded length. Evolution of the tensile capacity of SGR
composite with increasing bonded length. Test results corresponding to the experimen-
tal campaigns performed by Ascione et al. (2015), Santandrea et al. (2016, 2017, 2020),
De Santis and de Felice (2015), De Santis et al. (2017b), Wang et al. (2020) . . . . . . 55
3.6 Construction procedure: a) brick wet by immersion, b) mixture of the binding mortar,
c) placing of the rulers for the construction, d-e) construction of the wall in Flemish
bond, f) specimen finished for diagonal compression test, g) rebars placing for the
concrete beam located on top of the specimen for Shear compression testing, h) cast of
the concrete beam, i) specimen finished for shear compression test . . . . . . . . . . . 58
3.7 URM specimen built with a defect in the central brick to induce a regular diagonal
crack when subjected to diagonal compression . . . . . . . . . . . . . . . . . . . . . . . 59
xii
3.8 Procedure for the application of the TRM systems: a) creation of grooves along the
mortar joints, b) application of the first layer of mortar, c) set of the fibre net, d) finished
look of the wall retrofitted with basalt TRM, e) application of the first layer of vertical
mortar strips, f) finished look of the wall retrofitted with steel TRM, g) insertion of the
helical stainless steel rebar, h) wrapping of the specimens with wet sackcloth fabric for
curing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
3.9 Procedure of application of double layer strip configuration of LDS: a) application of
the first horizontal layer of LDS, b) application of the first vertical layer of LDS, c)
application of mortar material in the areas with no intersection, d-e) application of the
second horizontal layer of LDS, f) application of mortar material to get an even finished
surface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
3.10 Scuci-cuci intervention on the URM tested walls. Cracks opening with a hand tool and
filling of openings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
3.11 Scuci-cuci intervention on the URM tested walls. Removing and replacing of broken
units, filling of openings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
3.12 Scuci-cuci intervention on the URM tested walls. Removing and replacing of crushed
units. Filling of cracks with lime based strengthening mortar. . . . . . . . . . . . . . . 63
3.13 Setup of the Diagonal Compression Test . . . . . . . . . . . . . . . . . . . . . . . . . . 65
3.14 Evaluation of the ductility according to the approach proposed by Gattesco et al.
(2015b), a) evaluation of γy given by the ratio between the peak strength and the
shear modulus G, b) evaluation of γu which represents the ultimate shear strain corre-
sponding to 20% shear strength reduction . . . . . . . . . . . . . . . . . . . . . . . . . 67
3.15 Set-up of the Shear Compression Test . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
3.16 Set-up of the Shear Compression Test . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
3.17 Loading history for the shear compression testing of the wall specimens . . . . . . . . 73
3.18 Surface treatment and stochastic pattern for DIC analysis . . . . . . . . . . . . . . . . 75
4.1 Crack patterns of URM specimens a) URM 1, b) URM 2 . . . . . . . . . . . . . . . . 78

4.2 Crack patterns of damaged URM specimens repaired and retrofitted with basalt TRM
and zoom on the failure of some yarns at the end of the test. a) URM1 R, b) URM2 R 80
4.3 Crack patterns of specimens retrofitted with basalt TRM a) LDB 1, b) LDB 2 . . . . 81
4.4 Comparison of the diagonal compression load vs. opening displacement curves of un-
reinforced specimens (URM), repaired specimens and retrofitted specimens with basalt
TRM (LDB) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
4.5 Comparison of the diagonal compression load vs. opening displacement curves of un-
reinforced specimens (URM) and specimens retrofitted with steel TRM consisting of
unidirectional strips (LDS and MDS) . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
4.6 Crack patterns of specimens retrofitted with steel TRM and zoom of the debonding of
LDS strips and delamination within the matrix of MDS, a) LDS 1, b) LDS 2 c) MDS 1,
d) MDS 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
xiii
4.7 Diagonal compression load vs. opening displacement curves of unreinforced specimens
(URM) and specimens retrofitted with Basalt TRM in one side and joint repointing in
the other (LDB-JR) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
4.8 Crack patterns of specimens retrofitted with Basalt TRM in one side (left) and joint
repointing in the other side (right): a) LDB-JR 1, b) LDB-JR 2 . . . . . . . . . . . . . 87
4.9 Rate of enhancement of ductility and peak load for all the retrofitted specimens . . . . 89
5.1 Crack pattern at the end of the test: a) unreinforced wall URM 1, b) unreinforced wall
URM 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
5.2 Crack patterns at the end of the test (left: front face, right: back face): a) Repaired
and retrofitted wall URM1 R, b) Repaired and retrofitted wall URM2 R . . . . . . . . 94
5.3 Crack pattern at the end of the test (left: front face, right: back face): a) Retrofitted
wall with Basalt grid LDB 1, b) Retrofitted wall with Basalt grid LDB 2 . . . . . . . . 96
5.4 Detail showing the bed joint sliding surfaces on specimen LDB 2 . . . . . . . . . . . . 97
wall with steel textile LDS 1, b) Retrofitted wall with steel textile LDS 2 . . . . . . . 98
5.6 Toe-crushing observed at the end of the test on specimen LDS 2 . . . . . . . . . . . . 99
wall with two layers of steel textile LDS-DL 1, b)LDS-DL 2 . . . . . . . . . . . . . . . 100
5.8 Force-displacement hysteresis curves and envelope curves of tested walls: a) unrein-
forced wall URM 1, b) unreinforced wall URM 2, c) wall repaired and retrofitted with
LDB, URM1 R, d) wall repaired and retrofitted with LDB, URM2 R, e) wall retrofitted
with LDB (LDB 1), f) wall retrofitted with LDB (LDB 2) . . . . . . . . . . . . . . . . 102
5.9 Force-displacement hysteresis curves and envelope curves of tested walls: a) wall retrofitted
with one vertical and horizontal layer of LDS (LDS 1), b) wall retrofitted with one ver-
tical and horizontal layer of LDS (LDS 2), c) wall retrofitted with two vertical and
horizontal layers of LDS (LDS-DL 1), d) wall retrofitted with two vertical and horizon-
tal layers of LDS (LDS-DL 2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
5.10 Crack pattern evolution in the pushing direction of specimen LDB 2 . . . . . . . . . . 105
5.11 Example of bilinear idealization of an experimental horizontal load vs. displacement
curve . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
5.12 Bilinear idealization of the experimental envelope curves of tested walls: a) unreinforced
wall URM 1, b) unreinforced wall URM 2, c) wall repaired and retrofitted with LDB,
URM1 R, d) wall repaired and retrofitted with LDB, URM2 R, e) wall retrofitted with
LDB (LDB 1), f) wall retrofitted with LDB (LDB 2) . . . . . . . . . . . . . . . . . . . 109
5.13 Bilinear idealization of the experimental envelope curves of tested walls: a) wall retrofitted
with one vertical and horizontal layer of LDS (LDS 1), b) wall retrofitted with one ver-
tical and horizontal layer of LDS (LDS 2), c) wall retrofitted with two vertical and
horizontal layers of LDS (LDS-DL 1), d) wall retrofitted with two vertical and horizon-
tal layers of LDS (LDS-DL 2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
xiv
5.14 Degradation of the lateral stiffness suffered by the specimens on SCT with increasing
lateral displacement: a) specimens retrofitted with LDB compared with URMs, b)
specimens retrofitted with LDS compared to URMs . . . . . . . . . . . . . . . . . . . . 115
5.15 a) Drift capacity of URM specimens compared with the SD and NC limit states set
by EC8 (2004), b) drift capacity of retrofitted specimens according to the limit states
defined by Vanin et al. (2017) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
5.16 Computation of dissipated and strain energy of each displacement amplitude cycle . . 120
5.17 Cumulative dissipated energy of tested specimens . . . . . . . . . . . . . . . . . . . . . 121
5.18 a) Equivalent viscous damping ξeq of the tested specimens, b) Zoom of the damping
coefficient up to cracking (δcr ) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124
5.19 a) Correlation between lateral capacity increment (∆H) and reinforcement ratio (ρ),
b) correlation between ductility increment (∆µ) and reinforcement ratio (ρ) . . . . . . 127
6.1 a) Interpretation of SCT set-up according to Turnsek and Cacovic (1971), b) Interpreta-
tion of DCT set-up by Frocht (1931), c) Mohr’s circle representation of DCT according
to standards (dashed line), according to Frocht’s theory (dash-dotted line) and accord-
ing to the model proposed in the present research (solid line). Each representation is
associated with their corresponding coefficients for the computation of the maximum
and minimum principal stresses σI and σII and the maximum shear τmax . . . . . . . 134
6.2 Horizontal and vertical load at failure for each specimen obtained by SCT compared
with ideal curves governed by Equation (6.7) for different tensile strengths ft,SCT . . . 138
6.3 Comparison of stress state at the centre of the wall, being θ the rotation angle between
two different failure planes and θp the rotation angle that defines the orientation of the
principal planes: a) Mohr’s Circle for DCT, stress plane shown in solid line, b) Mohr’s
Circle for SCT, stress plane shown in dashed line. . . . . . . . . . . . . . . . . . . . . . 141
6.4 Correlation between the results obtained from the proposed methodology: a) Com-
parison between ft,SCT and ft,DCT for all the specimens tested in both unreinforced
and strengthened configurations, b) comparison of the shear parameter τ0,m,SCT and
τ0,m,DCT for all the specimens tested in both unreinforced and strengthened configurations143
6.5 Linear correlation between ft and τ0,m for all specimens tested on both testing method-
ologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145
6.6 Comparison between the proposed expression for the correlation between ft and τ0,m
for all specimens tested: a) Results from DCT testing methodology, b) Results from
SCT testing methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146
7.1 a) Distribution of the stress along the helical rebar embedded in the mortar joints, b)
crack pattern at the end of the test of a specimen retrofitted with NSM bars. Li is the
effective bond length of the i-th bar intersecting the diagonal crack . . . . . . . . . . . 155
xv
7.2 Results from SCT: a) Stress-Strain experimental curve of specimen repaired and retrofitted
with LDB-TRM, b) Stress-Displacement response of specimen repaired and retrofitted
with LDB-TRM, c) Stress-Strain experimental curve of specimen just retrofitted with
LDB-TRM, d) Stress-Displacement response of specimen just retrofitted with LDB-
TRM, e) Stress-Strain experimental curve of specimen retrofitted with LDS-TRM, f)
Stress-Displacement response of specimen retrofitted with LDS-TRM . . . . . . . . . . 158
7.3 Stress-Strain experimental curve of MDS-TRM under DCT configuration . . . . . . . 159
7.4 Digital Image Correlation (DIC) analysis: a) Readings of virtual extensometers located
on the strips of specimen LDS 2 before its failure in SCT configuration, b) Reading of
virtual extensometers located on the strips of specimen LDS 2 during its failure in SCT
configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 160
7.5 Stress-Displacement curve of the average measurement of the LVDT in tension of spec-
imens retrofitted with two layers of LDS tested in SCT configuration . . . . . . . . . . 162
7.6 Stress-strain experimental curves of the specimen asymmetrically retrofitted on one side
with LDB and on the other with NSM joint repointing tested in the DCT configuration 164
7.7 Correlation between the reinforcement ratio ρ and the exploitation ratio ff d /σu,f of the
textiles studied . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165
A.1 Specimens subjected to DCT: a) cold white lighting condition provides good reference
points to build the facets of the surface, b) yellowish lighting condition provides poor
and sparse reference points that do not allow the proper connection between the facets
to build the surface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 176
A.2 Two examples of the error ascribe to the obstruction of the conventional sensors . . . 177
A.3 a) bad quality of the stochastic pattern characterised by speckles of the same size, b)
good quality of the stochastic pattern characterised by different sizes of speckles . . . 177
xvi
List of Tables
3.1 Mechanical properties of bricks used for he construction of walls . . . . . . . . . . . . 48

3.2 Mechanical properties of binding mortar of each wall specimen tested under Diagonal
Compression Test(DCT) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
3.3 Mechanical Properties of binding mortar of each wall specimen tested under Shear
Compression Test(SCT) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
3.4 Average static (Em,st ) and dynamic (Em,dy ) modulus of the binding mortar . . . . . . 51
3.5 Mechanical properties of the products used for the reinforcement of the walls as provided
by the manufacturer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
3.6 Mechanical Properties of the mortar used as reinforcement in each wall specimen tested
under Diagonal Compression Test (DCT) . . . . . . . . . . . . . . . . . . . . . . . . . 53
3.7 Mechanical Properties of the mortar used as reinforcement in each wall specimen tested
under Shear Compression Test (SCT) . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
3.8 Type of test, ID and features of the specimens built and reinforced for the experimental
programme . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
4.1 Summary of the main results of URM and TRM retrofitted specimens . . . . . . . . . 86
5.1 Comparison of peak loads and corresponding displacements of specimens tested and
average increment in comparison with URM expressed in percentage . . . . . . . . . . 107
5.2 Experimental cracking and peak loads, ultimate load computed for the bilinear ideal-
ization and the ratio between the main bilinear and experimental parameters . . . . . 111
5.3 Bilinear parameter to compute the ductility on each tested specimen . . . . . . . . . . 112
5.4 Initial and effective stiffness of tested walls . . . . . . . . . . . . . . . . . . . . . . . . 113
5.5 Experimental displacements corresponding to different damage levels . . . . . . . . . . 118
5.6 Experimental drifts corresponding to different damage levels . . . . . . . . . . . . . . . 118
5.7 Proposed ranges of drift for strengthened masonry according to the failure mode . . . 119
5.8 Uncracked and cracked equivalent viscous damping . . . . . . . . . . . . . . . . . . . . 123
5.9 Reinforcement ratio ρ of studied TRM reinforcement systems . . . . . . . . . . . . . . 126
6.1 Assessment of the equivalent tensile strength ft,SCT obtained from the results of the
SCT according to Turnsek and Cacovic (1971). Computation of the average ft,SCT
associated to each defined group and its corresponding coefficient of variation (C.o.V) 137
6.2 Computation of coefficient α from DCT and SCT experimental results following Equa-
tion (6.2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
xvii
6.3 Comparison between the tensile strength obtained with the SCT results ft,SCT and with
the DCT results ft,DCT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
6.4 Comparison between the values τ0,m,SCT and τ0,m,DCT yielded from the proposed method-
ology for the correlation between SCT and DCT . . . . . . . . . . . . . . . . . . . . . 142
6.5 Equivalent tensile strength ft,DCT and the shear parameter τ0,m,DCT of specimens tested
under DCT configuration, and the ratio between both parameters . . . . . . . . . . . . 144
6.6 Equivalent tensile strength ft,SCT and the shear parameter τ0,m,SCT of specimens tested
under SCT configuration, and the ratio between both parameters . . . . . . . . . . . . 144
7.1 Comparison between the experimental values obtained from DCT and the analytical
values computed according to the Italian guideline CNR-DT 200 (2013) . . . . . . . . 157
7.2 Comparison between the experimental values obtained from SCT and the analytical
values computed according to the Italian guideline CNR-DT 200 (2013) . . . . . . . . 157
7.3 Comparison between the experimental values obtained for repaired and retrofitted spec-
imens and the analytical values computed according to the Italian guideline CNR-DT
200 (2013) and considering a reduced URM capacity . . . . . . . . . . . . . . . . . . . 160
7.4 Comparison between the experimental values obtained from SCT for specimen retrofitted
with two layers of LDS and the analytical values, computed according to the Italian
guideline CNR-DT 200 (2013), considering two different values for ff d , i.e. the one
obtained from the mechanical characterisation and the one recorded during SCT . . . 161
7.5 Comparison between the results provided by the analytical formulation of the guidelines
CNR-DT 215 (2018) and the experimental values obtained from each TRM strength-
ening system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163
7.6 Comparison between the experimental values obtained from DCT on specimen asym-
metrically retrofitted and the analytical values computed according to Li et al. (2005) 163
7.7 Efficiency of TRM in term of exploitation ratio computed between the tensile capacity
of the reinforcement ff d and the ultimate tensile strength of the textile σu,f . . . . . . 164
xviii
List of Symbols
2θ Coordinate rotatio angle in Mohr’s Space
α Coefficient dependent on masonry typology
αt Coefficient of reduction due to shear
β Coefficient of correlation between τ0 and ft
βf d Coefficient dependent on the type of failure
∆x Height of the brick unit
∆y Width of the brick unit
δ Displacement
∆0 Smallest deformation amplitude at which damage is first observed
δcr Lateral displacement corresponding to the instance when the first cracks are visible
δc Displacement at collapse corresponding to a drop of 50% of the lateral load
δHmax Displacement at maximum lateral load
δi+ Displacement in the positive direction for cycle i
δi− Displacement in the negative direction for cycle i
∆m Deformation at which the most severe damade is expected
∆m Deformation at which the most severe damage is expected
δu Ultimate displacement corresponding to the moment the strength has dropped 80%
δy Displacement at yielding correponding to ultimate load
ηa Conversion factor
γ Shear strain at the centre of the wall according to ASTM standard
γf Partial factor
xix
γRd Partial safety factor according to CNR-DT 200
γu Ultimate shear strain
γy Yield shear strain
κ Ratio between the shear stress at the centre of the panel and the mean shear stress
µ Ductility
µ0 Friction coefficient of the mortar joint
ω Frequency of vibration
ωn Natural frequency of vibration of a structure
φ Phase angle
ρ Reinforcement ratio
σ0 Vertical compression stress
σII Minimum principal stress
σI Maximum principal stress
σI Tensile stress at the end of Stage I of tensile test on TRM composite coupon
σsl,t Tensile capacity of the textile from the shear bond test
σt Tensile stress at the attaiment of peak load of tensile test on TRM composite coupon
σu,f Ultimate tensile strength of the textile fibre
σv Designed compression stress of SCT
σx Horizontal stress
σy Vertical stress
τ0,m,DCT Shear parameter of masonry obtained from DCT results
τ0,m,SCT Shear parameter of masonry obtained from SCT results
τ0 Cohesion of the mortar joint
τAST M,RILEM Shear stress at the centre of the wall according to ASTM and RILEM standard
τb Bond strength between the masonry and the mortar matrix
τelastic Shear stress at the centre of the wall according to Frocht’s theory
τmax Maximum shear stress attained during DCT
xx
τxy Shear stress
τy Shear stress corresponding to the yield shear strain
θcr Drift at onset of cracking
θc Drift at collapse
θHmax Drift when the maximum lateral load is attained
θN C Drift for EC8 NC limit state
θp Principal angle in Mohr’s Space
θSD Drift for EC8 SD limit state
θSLU Drift for NTC 2018 SLU limit state
θu Ultimate drift corresponding to the bilienar idealization
θy Drift at yielding
εσt Tensile strain at peak load of tensile test on TRM composite coupon
εc Strains along the shortening (compressed) diagonal of the wall
εf dd Maximum strain of reinforcement at which the debonding takes place
εf d Strain corresponding to the reinforcement tensile capacity
εf k Strain at failure of the reinforcement
εI Tensile strain at the end of Stage I of tensile test on TRM composite coupon
εt Strain along the elongating (tensioned) diagonal of the wall
εu,f Strain at failure of the textile fibre
ϕ Geometric parameter representing the interlocking of masonry pattern
ξ Equivalent viscous damping
AT RM Transversal area of the TRM reinforcement systems
Ab Shear contribution of NSM rebars
Acord Area of a cord of the textile fibre
Aenvelope Area below the experimental envelope curve up to δu
Af ilo Area of a single yarn of the textile fibre
Af Cross-section area of the helical rebars
xxi
ai+1 Amplitud at step i+1 of the loading protocol of SCT
ai Amplitud at step i of the loading protocol of SCT
An Net area of the panel
b Geometrical coefficient
bf Width of TRM strip
D Depth of the groove
Eb,dy Dynamic Young’s Modulus of the brick
Eb,st Static Young’s Modulus of the brick
Edi Energy dissipated at the first complete cycle i
ED Cumulative dissipated energy
Ef Young’s modulus of the textile
EIII Young’s Modulus of TRM composite coupon in stage III of tensile test
EII Young’s Modulus of TRM composite coupon in stage II of tensile test
EI Young’s Modulus of TRM composite coupon in stage I of tensile test
Em,dy Dynamic Young’s modulus of the biding mortar
Em,st Static Young’s modulus of the biding mortar
Em Young’s modulus of masonry
Erm,st Static Young’s modulus of mortar matrix
ESo Strain energy
fb,c Normalized compressive strength of the brick
fb,f Flexural strength of the brick
fb,t Brick tensile strength parallel to the bed joint
fc Compressive strength of masonry
ff d Tensile capacity of the reinforcement
fm,c Compressive strength of the binding mortar
fm,DP T Double punch test compressive strength of the binding mortar
fm,f Flexural strength of the binding mortar
xxii
frm,c Compressive strength of the mortar matrix
frm,f Flexural strength of the mortar matrix
ft,DCT Equivalent tensile strength of masonry computed from DCT results
ft,SCT Equivalent tensile strength of masonry computed from SCT results
ftk Characteristic tensile strength of masonry
ft Tensile strength of masonry
fu Maximum tensile strength of the helical rebars
h Height of the wall
h0 Effective heigth of the walls
Hcr Cracking load recorded when the first cracks became visible
Hi+ Maximum force in the positive direction for cycle i
Hi− Maximum force in the negative direction for cycle i
Hmax Maximum lateral force corresponding to the SCT
Hu Ultimate load of the idealised envelope
K Experiemntal secant stiffness
Kei Effective stiffness for cycle i
Ke Effective stiffness
Kinitial Initial stiffness
Lef f Effective bonded length of the composite on the substrate
Mu Resistant bending moment of un unreinforced masonry wall
pf Centre-to-centre spacing between TRM strips
tf Equivalent thickness of the reinforcement parallel to the applied shear force
tm Thickness of the mortar joint
u20 Theoretical target displacement amplitude in one cycle
Vc Shear capacity due to toe-crushing failure
VEXP Experimental Load obtained from DCT for reinforced specimen
Vmred Reduced URM shear capacity
xxiii
Vm Analytical shear contribution of masonry
Vs,dt Shear capacity due to tensile diagonal failure
Vt,f,EXP Experimental peak force sustained by the TRM strengthening system
Vt,f Analytical shear contribution of the reinforcement
Vt,LDB Shear contribution of LDB retrofit
Vt,N SM Shear contribution of NSM rebars
Vt,R Nominal shear capacity computed according to CNR-DT 200
GAST M,RILEM Shear modulus of elasticity according to ASTM and RILEM standard
d Length of the wall in the direction of the applied shear force
H Horizontal shear force applied SCT
N Axial load
P Applied load during DCT
V Vertical force applied SCT
xxiv
Chapter 1
Introduction
1.1 Background and Motivation
Due to the availability of the component materials, masonry is one of the oldest construction techniques
used worldwide. It is characterised by the assembly of the available units laid dry in courses or bound
together with mortar, which makes it extremely diverse. Its profusely use lays not only on its ease of
construction but also on its efficiency to withstand vertical compressive load, i.e. gravitational loads.
Consequently, when masonry buildings are affected by in-plane lateral loading, such as earthquake
action, they show large vulnerability due to the fact that masonry has almost null tensile strength.
As one of the main historical construction materials, masonry is abundant in the built cultural
heritage located in many earthquake-prone regions of the world. Therefore, the evaluation of their
shear behaviour is absolutely essential. Earthquakes are one of the most dangerous natural hazards,
not only due to the significant amount of human and direct economic losses, but also because they
cause enormous damage, including collapse, to the building stock. In many cases, such building stock
includes unreinforced masonry buildings and particular historical masonry buildings of high cultural
value. Society demands the conservation of our built heritage, as a way to preserve our cultural
identity. Consequently for the sustainability of heritage, it is of paramount importance an accurate
safety assessment of the structural capacity of existing masonry structures for the design of effective
strengthening solutions, as a result of the structural analysis. In addition, there is a necessity to
validate any repairing solutions already present for the sake of restoration and conservation of the
structure.
After recent seismic events, such as L’Aquila 2009 (Italy), Canterbury 2010 (New Zealand), Emilia
2012 and Amatrice-Norcia-Visso 2016 (Italy) earthquakes, composites materials for post-earthquake
repair and seismic retrofitting of masonry structures have received a growing interest from the scientific
community. To improve the deficiencies related to the poor structural performance of URM shear walls
under seismic actions, different strengthening techniques have been developed and applied. The main
aim of the strengthening techniques is to increase low parameters of masonry such as tensile and shear
1
CHAPTER 1. INTRODUCTION
strength.
Among the proposal of different strengthening techniques involving composite materials, one that
reached worldwide consensus at the early stages of its development was Fibre Reinforced Polymers
(FRP). FRP mainly comprising carbon, glass and aramid fibre, have been profusely utilized due to
their high tensile strength, lightweight, relative ease of installation and resistance to corrosion. The
research carried out on the use of FRP has shown its ability to successfully enhance the in-plane
strength of masonry walls as evidenced in Valluzzi et al. (2002), Marcari et al. (2007), and Alecci et al.
(2019). However, FRP systems have shown meaningful limitations precluding their use in several cases.
For instance, high or low temperatures might compromise the efficiency of FRP systems, and wet lay-
up FRP applications are not possible either on moist surfaces or at low temperatures. In addition,
FRP systems typically act as a vapour barrier and therefore, cannot be used when permeability is
required, as in the case of existing masonry structures. These drawbacks stem mainly from the epoxy
matrix, which acts both as the binder of the fibres and the bonding agent between the composite and
the substrate. In addition, the epoxy matrix is also the reason for FRPs irreversibility and possible
early debonding from a weak substrate, which in historical masonry, including clay or tuff brick units,
is detrimental for the substrate material.
According to the ISCARSAH Principles (2003), repair or strengthening solutions applied to ma-
sonry are recommended to be preferable non-invasive and compatible bearing in mind safety and
durability requirements. It has to be assured the compatibility between the existing materials and the
characteristics of the materials used in restoration works, particular attention should be paid when
new materials are used. This compatibility should not only be at a structural level, considering the
technique and historical value of the original structure, but also considering aspect such as the aes-
thetics, the physical, the chemical, and the mechanical properties. Any intervention should weight the
different aspects of architecture, structure, installations and functionality. From a general perspective,
the new composite materials should merely reinforce and not alter the existing structural scheme or
act on the historic materials in a negative and harmful way.
Being clay brick masonry with low strength lime-based mortar abundant in the built heritage of the
Mediterranean, it becomes clear that the alternative solution to FRP systems consists in replacing the
epoxy resins by inorganic matrices, among which the most common ones are lime-based or cement-
based mortars. Textile reinforced mortar (TRM) has been receiving increased attention due to its
mechanical efficiency and satisfactory compatibility with the masonry substrate, overcoming most
of FRP’s drawbacks. One of the major differences between TRM and FRP, and the reason behind
its compatibility with masonry, is the bond properties between TRM and the substrate. Unlike the
epoxy of FRP, the mortar matrix of TRM has evidenced to homogeneously distribute strain and
stress throughout the reinforcement hindering failure modes involving the substrate, i.e. debonding or
detachment, and promoting failure modes involving the rupture of the textile’s fibres by exhausting
its tensile strength. These results are also consequence of the properties of the composites materials.
Mechanical compatibility between the mortar and the textile is of paramount importance to assure
2
similar strain development and gradual stress transfer. Such compatibility is achieved by mortars
having moderate compressive strength and textiles with limited Young’s modulus.
A significant research effort has been undertaken during the last decade in order to validate efficient
masonry seismic strengthening techniques, such as TRM, to improve masonry’s in-plane load-bearing
and displacement capacity. And to a minor extent, research has also been done to evaluate the capa-
bility of the TRM systems for post-earthquake repair. However, there is a scarcity of design provisions
to help practitioners with the safety assessment of the repairing or strengthening solutions. Current
guidelines (AC549 (2013) and CNR DT215 (2018)) base their design procedure on the strength, fo-
cussing in the safety check by comparing the global capacity of the building against its demand. The
approach based on limit state design is missing since no experimental data is available. Serviceability
limits related to the cracking of the mortar matrix and ultimate limit states concerning severe damage,
related to particular failure mechanism of TRM, and event of collapse are yet to be defined.
There is a necessity to better understand the mechanical behaviour and the evolution of local and
global failure mechanisms of TRM strengthening systems retrofitting masonry walls. The fundamental
step towards the accomplishment of these objectives is the performance of experimental campaigns
thoroughly designed so as to identify the limit states associated with different levels of damage. Among
which, the ultimate limit state is of paramount importance for assessing the performance of strength-
ened masonry structures under loading conditions such as those induced by earthquakes.
Nowadays two experimental testing methodologies can be identifies for the global in-plane assess-
ment of unreinforced masonry walls, namely Diagonal Compression Test (DCT) and Shear Compres-
sion Test (SCT). Both the Eurocode 8 (2004) and the Italian Guideline NTC (2018) both recommend
the DCT for the evaluation of the shear strength in unreinforced masonry (URM) walls under in-
plane actions. In turn, the testing protocols provided by FEMA 461(2007) suggests the quasi-static
cyclic testing controlled by imposed displacements. The DCT has been successfully carried out on
different masonry typologies strengthened by means of FRP and TRM. Notwithstanding the valuable
information provided by DCT experimental results regarding the load-bearing capacity, the accurate
definition of damages levels is of utmost importance when evaluating a repair or designing a strength-
ening solution. Damage levels can be quantified from experimental drift results. However, in the case
of monotonic DCT the drift results are not representative of the real behaviour of the structure under
seismic action. As evidenced by Vanin et al. (2017) monotonic test lead to larger drift capacities.
Therefore, cyclic SCT complements the characterisation by providing experimental results regarding
the displacement capacity of strengthened masonry walls. Another advantage of these testing method-
ologies lays on the assessment of the tensile strength of unreinforced and strengthened masonry walls.
Researchers have faced major difficulty developing feasible testing techniques to accurately determine
it. Both DCT and SCT experimental results can be used to assess such property. However, there
are different interpretative models of the DCT and therefore, its results may raise some significant
uncertainties. For this reason, the pursuit of a methodology that combines and validates both test-
ing techniques for the in-plane assessment of masonry walls, unreinforced or strengthened, is still a
3
challenge.
Finally, considering all the aforementioned, three major issues are identified. First, there is the ne-
cessity to improve the knowledge of the in-plane seismic response of repaired or strengthened masonry
walls, not only in terms of load-bearing and displacement capacities but more importantly to under-
stand the failure mechanisms that trigger their collapse. The second one involves the identification
of the meaningful damage levels with the aim of developing a limit state approach for the design of
strengthening solution or the evaluation of existing ones. Finally, an integrated methodology process
for the accurate assessment of the in-plane behaviour strengthened masonry is missing.
1.2 Scope and Objectives
The principal scope of this research is to contribute to the in-depth understanding of the in-plane
structural response of strengthened masonry while validating TRM strengthening systems as an in-
plane behaviour enhancement technique. More specifically, the research focuses in historical masonry
comprising clay bricks and low strength lime-based binding mortar strengthened with TRM systems
based on basalt and steel textiles embedded in lime-based mortar matrix. The outcome of the research
is intended to provide suitable design information for the formulation of code provisions and to allow
prediction of the strengthened wall’s capacities.
Motivated by the necessity to improve the knowledge on in-plane enhancement of TRM strength-
ened masonry structures under seismic actions, the following specific objectives have been defined:
To compile comprehensive state-of-the-art information concerning the bond properties between

TRM strengthening system and masonry substrate; regarding failure modes, bonded length,
stress transfer and tensile capacity and failure mechanism of strengthened masonry subjected to
in-plane loading.
To improve the experimental insight on the different testing methodologies to characterise the
in-plane properties of strengthened masonry by designing an integrated experimental campaign
involving full scale specimens subjected to the two main in-plane testing methodologies, namely
Diagonal Compression Test (DCT) and Shear Compression Test (SCT). This objective includes
the following aims:
– Define the specimens features to be tested and the TRM strengthening configuration re-
garding materials, textile thickness, yarn density and number of layers.
– Design both tests set-up to accurately test the specimens without uncertainties about the
boundary conditions, load configuration and loading protocols.
– Identify the properties of the in-plane behaviour of unreinforced and strengthened masonry.
– Identify the different failure mechanisms that can lead to the collapse of the masonry on its
two configurations, unreinforced and strengthened, and define their corresponding governing
parameters.
4
– Evaluate the lateral load bearing capacity, shear modulus displacement capacity, strength
and stiffness degradation, damping, and energy dissipation, among other parameters, for
both unreinforced and strengthened masonry. In addition, assess the enhancement provided
by the TRM strengthening systems when compared with the unreinforced masonry walls.
– Compute the effectiveness of the different proposed TRM strengthening solutions regarding
their effectiveness in improving the in load-bearing capacity, ductility and displacement
capacity, among other parameters. Moreover, evaluate the influence of the number of
reinforcement layers and textile yarn density on the enhancement provided.
– Provide a feasible correlation between the results of both testing methodologies to charac-
terise the in-plane behaviour of unreinforced and strengthened masonry.
Identify damage propagation and crack patterns to shed light on the failure mechanisms that
trigger collapse of masonry walls on both configuration, unreinforced and strengthened. In the
case of the first configuration, this understanding will be beneficial for choosing the most ap-
propriate consolidation intervention. In the case of the second, it will provide the identification
of different levels of damage, which will allow a better assessment of masonry walls under seis-
mic actions and to establish ranges of value for the meaningful limits states for the design of
retrofitting solutions.
Validate TRM strengthening systems as a seismic retrofitting solution, but also as a post-
earthquake repairing solution in combination with other repairing techniques such as mortar
joint repointing and ”scuci-cuci”.
Review the current formulation used in the codes to calculate the lateral load-bearing capacity
of unreinforced masonry walls. Validate the available formulation according to the experimental
response observed and its suitability to estimate the lateral load-bearing capacity of strengthened
masonry.
Validate the different testing methodologies to assess the shear behaviour of masonry and to
compute the tensile strength. Provide a methodology to correlate the experimental results
to accurately compute the most important properties of strengthened masonry under in-plane
loading, i.e. tensile strength and shear capacity.
Review current analytical formulations for the design of strengthened solutions in terms of load
capacity. Propose a drift limit state, based on the experimental results, to fulfil the required
safety assessment of strengthening solutions on masonry.
The aim of this thesis encompasses a broad subject that is the study of the in-plane behaviour of
masonry strengthened with TRM strengthening solutions. Only one masonry typology was studied, i.e.
handmade clay bricks masonry with low strength lime based binding mortar. This typology was chosen
since it is recurrent historical an existing masonry structures worldwide. The TRM strengthening
solutions tested are commercially available; however, and despite the attempt to embrace as many
potential solutions as possible, only two different reinforcing materials, namely basalt and steel, were
5
studied. To determine the influence of the bond properties in the final outcome different yarn densities
and applied number of yarns were chosen. It is also worth mentioning that only one loading history,
consisting of repeated cycles of step-wise increasing deformation amplitudes, was applied, since it is
more representative of the real seismic action and consequently the structural response would be more
trustworthy.
1.3 Assumptions
One of the main assumptions of this research is that Textile Reinforced Mortar (TRM) is an efficient,
relatively low cost and compatible strengthenign solution for the enhancement of the in-plane response
of masonry walls.
The other assumptions is that the in-plane shear behaviour of strengthened masonry can be char-
acterised by the same experimental testing as URM, and that the response of strengthened masonry is
the results of the joint action between both masonry and the TRM strengthening systems. Therefore,
the response of strengthened masonry can be assessed with the same approaches and methodologies
than the response of URM.
1.4 Research Dissemination
The research conducted in this manuscript has resulted in the following scientific publications at the
moment of submission of the PhD thesis:
Articles in peer-reviewed international journals:
- Garcia-Ramonda, L., Pela, L., Roca, P., and Camata, G. (2020) In-plane shear behaviour by
diagonal compression testing of brick masonry walls strengthened with basalt and steel textile re-
inforced mortars. Constr. Build. Mater., Vol 240. doi.org/10.1016/j.conbuildmat.2019.117905
Articles and presentations in international conferences:
- Garcia-Ramonda, L., Pela, L., Roca, P., and Camata, G. (2020) Experimental study of in-plane
shear behaviour of brick masonry retrofitted with basalt and steel reinforced mortar. Proceeding
of 8th REHABEND 2020 – Construction Pathology, Rehabilitation Technology and Heritage
Management, Granada, Spain, 28-30 September.
- Garcia-Ramonda, L., Pela, L., Roca, P., and Camata, G. (2021) Experimental assessment of
shear response of brick masonry retrofitted with TRM. Accepted for presentation in 12th Inter-
national conference on Structural Analysis of Historical Constructions, Barcelona, Spain
In addition, two manuscript are currently in preparation for submission to peer-reviewed interna-
tional journals for consideration.
6
1.5 Outline
This work is organised in eight chapters. The first chapter introduces the motivation, the scope and
the objectives of the present thesis. The remaining seven chapters are organised in the following way.
Chapter 2: Literature Review
The chapter presents the literature review on the most significant work carried out on the three
principal topic of this research. The first one is related to the characterisation of TRM strengthening
systems, with focus on the bond properties and failure modes. The second one concerns the different
experimental approaches used to characterise the in-plane response of masonry strengthened with
different TRM strengthening solutions. Finally, the third topic regards the available formulations and
code provisions for the evaluation of in-plane capacity of unreinforced masonry and the design and
assessment of TRM strengthening solutions. The discussion includes the identified gaps that motivates
this research.
Chapter 3: Experimental Programme
This chapter details the mechanical characterisation of the component materials of the specimens
and its reinforcement, the preparation of the specimen, its repairing and strengthening. Within the
preparation tasks are included the wall’s construction and the mortar samples construction to control
their resistance. Special attention is given to the TRM application procedure and the repairing
techniques. Finally, the chosen set-ups for testing are described along with the testing procedure,
including loading protocols and methodology for the analysis of the results.
Chapter 4: Experimental Results: Diagonal Compression Test
This chapter gathers the experimental results from the twelve DCT performed in terms of crack-
ing patterns, failure modes, Force-Displacement curves (F-δ), shear strength, stiffness, load bearing
capacity and ductility. The chapter also discusses the compatibility of the different TRM strengthen-
ing solutions along with the improvements, in percentage, of the ductility and load bearing capacity.
Special attention is given to the performance of the unidirectional textiles with different yarn densities
to study the influence that it has on the in-plane response of strengthened masonry.
Chapter 5: Experimental Results: Shear Compression Test
This chapter contains a detailed analysis on the experimental results from the ten SCTs performed
in terms of lateral load-bearing capacity, displacement capacity, limit state, ductility, energy dissipation
and damping coefficient, among other parameters. This analysis helps to better understand the failure
mechanisms that lead the walls to collapse, and to set ground rules regarding drift capacities for better
design strengthening solutions. The discussion is also focused on the influence of reinforcement ratio on
two of the main parameters that can define the in-plane enhancement provided by TRM strengthening
solutions, namely the load bearing capacity and the ductility. The discussion includes the identification
of an optimal reinforcement ratio.
Chapter 6: Correlation between results from Diagonal Compression Test and Shear Compression Test
This chapter presents a discussion regarding the correlation between both testing approaches. The
7
proposed methodology for the accurate assessment of the in-plane behaviour is presented, in which the
failure mechanism triggered by both testing configurations is identified and associated to the Turnsek
and Cacovic (1971) failure criterion. The experimental results from both tests on the specimens, with
the same TRM strengthening solutions, are compared and correlated. The proposed methodology is
validated with specimens only subjected to one of the test configurations and a correlation between
the main parameters that governed the behaviour, tensile strength and shear capacity, is obtained
while, in turn, a new shear parameter is presented. In addition, the proposed methodology allows the
validation of the interpretation proposed by Frocht (1931) and Brignola et al. (2008) and provides a
more accurate determination of a coefficient α used in the calculation of the tensile strength for the
masonry typology under study.
Chapter 7: Analytical Study
This chapter explores the analytical formulations currently available for the estimation of the in-
plane shear capacity contribution of the different TRM strengthening systems studied in the present
work. Their validity is investigated through the comparison between the experimental and analytical
results. In addition, the parameters considered by each formulation to determine their influence on
the efficiency of the strengthening solution are investigated. The exploitation ratio for each type of
TRM is defined to determine the efficiency of each solution. In addition, the correlation between the
reinforcement ratio and the exploitation ratio is investigated in order to determine the influence of the
yarn density and the number of layers on the TRM efficiency.
Chapter 8: Conclusions
The closing chapter of the thesis recapitulates all the research performed and summarises the main
conclusions reached in each of the aforementioned chapters. Finally, it presents suggestions for future
work.
8
Chapter 2
Literature Review
2.1 Introduction
The evaluation and prediction of the in-plane strength of unreinforced masonry walls plays a crucial
role in their seismic assessment and hence in the design of possible retrofitting solutions. A direct
approach to estimate the lateral strength consists of performing experimental test able to simulate
reality, as close as possible, in terms of boundary conditions and acting forces. Current codes, namely
EC8 (2004) and NTC (2018), describe specific failures modes which may occur to the masonry walls,
such as rocking, crushing, sliding through bed joint and diagonal cracking. According to the failure
modes experienced by the URM, the standards recommend different models to assess the load and drift
capacity of the structure. The models are based on the results of extensive experimental programmes
on different masonry typologies.
In the last two decades a considerable effort on the research of the in-plane seismic response of
strengthened shear walls has been made since they are key vertical components to bear the seismic
loading. Among the possible strengthening solutions different techniques based on the availability and
fashion of material were investigated. Recent research has evidenced that TRM retrofitting solutions
improve the resilience of shear masonry walls to natural hazards, such as earthquakes. However,
for masonry application, additional requirements exist for TRM strengthening interventions, namely
materials’ compatibility to the substrate (chemical, physical, mechanical), durability (especially in
cases of historical structures) and reversibility. Inorganic matrices, mainly lime based, succeed in
addressing these challenges as they can be matched with the masonry mortar achieving chemical
compatibility and good adhesive features in respect to the substrate while insuring breathability.
Despite the effectiveness shown by the TRM strengthening system, extensive experimental research
is still needed to define the in-plane behaviour of strengthened masonry elements and to provide insight
into the enhancement of load and drift capacity in order to better design retrofitting or repairing
solutions.
9
CHAPTER 2. LITERATURE REVIEW
2.2 In-plane shear behaviour of unreinforced masonry

According to Tomaževič (1999), masonry shear walls are classified intro three main categories consid-
ering their structural configuration.
Cantilever walls, where the storeys are connected together with floor slabs. The seismic forces
are carried by the walls. Floor slabs, rigid in their plane but flexible in the orthogonal direction
to the plane, distribute the lateral loads onto the walls in proportion to their stiffness. However,
they do not transfer any moments resulting from the bending walls.
Piers weaker than spandrels, which is most often the case in traditional unreinforced masonry
construction. Damage will first occur to the piers. The spandrels are rigid so that the piers may
be considered as fixed above and below. Depending on the geometry and quality of masonry
materials, the piers will either fail in shear due to diagonal compression or rock until crushing
of masonry occurs at the compressed zone. Shear failure is the most frequent failure mechanism
of unreinforced masonry.
Spandrels weaker than piers. The spandrels behave as coupling beams, which connect the walls
together, developing a framing action and transferring bending moment. The coupling elements
influence the rotational restraint at the top of the wall and therefore, the bending moment profile.
The different structural configurations influence the seismic behaviour and hence the failure mech-
anisms of masonry shear walls. According to Magenes and Calvi (1997) and Tomaževič (1999) various
types of failure mechanism define the seismic behaviour of structural masonry walls when subjected to
in-plane loading. The different mechanisms strongly depend on the vertical loads and the mechanical
characteristics of the mortar and the units that compose the masonry, as also evidenced by Preciado
and Ramirez-Gaytan (2019). Also, they depend on the geometry of the wall, specifically height/width
ratio and the boundary restraints as observed in Calderini et al. (2010). The main two types of failure
mechanism are shown in Figure 2.1.
- Shear Failure: it is possible to distinguish two types of failure modes within the shear failure
mechanism: failure due to diagonal cracking, usually associated with the attainment of the
tensile strength of masonry, and failure due to sliding along the bed joints.
- Flexural failure: shear is carried by the compressed masonry and the final failure can be
due to the overturning of the wall with simultaneous crushing of the compressed corner, better
known as rocking, or the crushing of the corner due to the attainment of compression strength
of masonry, also called toe-crushing. This type of failure mechanism is the most ductile and
the least harmful since the wall can withstand large imposed displacement without significant
strength loss.
From the listed categories of shear walls, the ones with cantilever configuration are prone to fail
due to flexural failure mechanisms rather than shear failure mechanisms, presenting a challenge when
it comes to design efficient strengthening solutions.
10
Figure 2.1: Typical failure modes of masonry walls, subjected to in-plane loading according to Tomaževič
(1999) a) Shear failure mechanism: shear sliding along bed joint, b) Shear failure mechanism: shear failure due
to diagonal cracking, c) Flexural failure mechanism: rocking and crushing
2.2.1 Shear Resistance
When masonry walls fail under in-plane loading due to a shear failure mechanism, the mechanical
property that defines its lateral resistance depends on the the model describing the failure mode. As
aforementioned, there are two failure modes that lead the wall to its collapse due to shear.
- Shear Sliding characterised with low vertical load and poor quality bed joint mortar which is
prone to develop horizontal cracks causing the sliding of the upper part of the wall on one of the
horizontal mortar joints or stepped crack failure alternating from head joints to bed joints.
- Tensile Cracking takes place when the principal tensile stresses exceeds the tensile strength
of masonry materials. Is characterised by diagonal cracks developing in the wall. The cracks
alternates from passing through mortar joints to passing through the masonry units, depending
on the relative strength of mortar joints, brick-mortar interface, and bricks. This failure mode
is not only the most common but also the most unfavourable type of wall failure since it is
normally brittle and sudden due to the almost null tensile strength of masonry.
Depending on the expected failure mode the shear capacity Vm of unreinforced masonry subjected
to in-plane loading is usually evaluated on the basis of the Mohr-Coulomb (1900) criterion or in the
basis of the Turnsek and Cacovic (1971) formulation.
The Mohr-Coulomb criterion is used when the shear failure is associated with the sliding on a bed-
joint and characterised by the formation of horizontal cracks. It assumes that the failure shear stress
τ is given by Equation 2.1, where τ0 is the shear strength under zero compression stress, also defined
as the cohesion of the mortar joint and µ0 is its friction coefficient. The EC6 (2006) recommends the
evaluation of the shear strength of masonry according to the Mohr-Coulomb criterion. According to
Tomaževič (2008), this criterion can be accepted in the case of masonry walls built of strong masonry
units and poor quality mortar, in which diagonally oriented cracks passing through the mortar joints
are more likely to occur. The standard also recommends to set the parameter µ0 equal to 0.4.
11
τ = τ0 + µ0 σ0 (2.1)
The formulation proposed by Turnsek and Cacovic (1971) is recommended when the shear failure
is featured by diagonal cracking which cannot be explained using the frictional theory. Under the
hypothesis of masonry as a homogeneous and isotropic material, the failure takes place when the
principal tensile stress σI at the centre of the panel attains a value equal to the tensile strength ft of
the masonry. Considering the theory of elasticity and the interpretation of this model on the Mohr’s
circle, the principal stresses in the middle section of the wall can be computed following Equation
2.2 and 2.3 by neglecting the horizontal normal stress, where b is the shear stress distribution factor.
The initial authors of the criterion defined the parameter b equal to 1.5. Later Turnsek and Sheppard
(1980) and Tomaževič (1999) redefined it as dependent on the shape and geometry of the wall, namely
as a height-width ratio H/B.
r
σ0 2 σ0
σI = ( ) + (bτ )2 − (2.2)
2 2
r
σ0 2 σ0
σII = ( ) + (bτ )2 + (2.3)
2 2
The shear resistance of the wall characterised by diagonal cracking can be obtained from rearrang-
ing Equation 2.2. From Equation 2.4 it is observed that the formulation computes the shear capacity
using only one global parameter, the tensile strength of masonry ft , also called ”referential tensile
strength”.
r
ft σ0
τ= +1 (2.4)
b ft
Another approach that considered the development of the diagonal cracks is the formulation pro-
posed by Mann and Müller (1980). This model highlights that masonry is not homogeneous but
affected by the joints and since it is made up of numerous elements, the failure could be ascribed to
different reasons. The model is based on the hypothesis that no shear stress τ can be transferred in
the vertical joints between the bricks and also assumes a horizontal compressive stress negligible and
therefore the horizontal axial loads cannot produce vertically oriented friction. However, the shear
stress τ acts in the horizontal joints producing torque in the individual brick. Equilibrium can only
be attained via a vertical force couple, which influences the vertical stress σy in such a manner that
one brick half is subjected to larger stress and the other half to a smaller one, as shown in Figure 2.2.
2∆x
σ1,2 = σy ± τ (2.5)
∆y
If the failure occurs due to friction, then the brick subjected to lesser stress is the one failing
earlier and the shear stress is governed by Equation 2.6, where the authors introduced the concept
of reduced cohesion and reduced friction coefficient that are function of the parameter ϕ representing
12
Figure 2.2: Stress state on the brick due to shear-compression loads as proposed by Mann and Müller (1980).
(Graphical representation taken from Calderini et al. (2010))
the interlocking of masonry pattern (ϕ = 2h/b).
c µ0
τ= + σy (2.6)
1 + µ0 ϕ 1 + µ0 ϕ
The case of bricks failing prematurely is associated to an increment of the vertical load σy that
precludes into the bricks cracking since no shear stress can be transferred in the vertical joints. There-
fore the bricks must transfer the double shear force. The brick cracks as the principal tensile stress
σI in the brick exceeds the brick tensile strength fb,t .
fb,t σy
r
τ= 1+ (2.7)
κ fb,t
Despite the similarities between these two last models, the strength domain of each model is defined
through different mechanical parameters. The model proposed by Turnsek and Cacovic (1971) defines
its strength domain through a single global parameter of the material i.e: the tensile strength, which
according to the Italian standard NTC (2018) can be determined by diagonal compression tests.
Conversely, the model proposed by Mann and Müller (1980) defines its strength domain on the basis
of local parameters related to the constituents of the material, i.e: cohesion and friction coefficient of
joints, evaluated through triplet tests. Calderini et al. (2010) concluded that the models may provide
different predictions of the shear strength and their results would rely on the degree of anisotropy of
the type of masonry studied. Coherently with the hypothesis adopted by each model, the Turnsek and
Cacovic (1971) criterion is more suitable for the assessment of those type of masonries that behave
as homogeneous, isotropic materials, while Mann and Müller (1980) criterion is more appropriate for
those that are characterised by their anisotropy.
13
2.2.2 Flexural Resistance
According to Tomaževič (1999), the damage attributed to predominant flexural behaviour is rarely
observed in masonry building, not only due to the low ratio between tensile and compressive strength
of masonry but mainly due to the low moment to shear ratio.
Figure 2.3: Rigid block model for the flexural mechanism due to rocking according to Magenes and Calvi
(1997)
The flexural failure mechanism can be explained through Figure 2.3. By neglecting the tensile
strength of the bed joint, the increasing lateral deformation leads to an increment of the length of the
crack at the tensioned side. Consequently, the effective compressed area of the wall’s cross-section is
reduced. As a result, the equilibrium is maintained if the resultant of the vertical load moves towards
the compressed border of the pier along the diagonal. Ultimately, the lateral resistance is increased
until the stress at the corner, opposite to the load application point, attains the compressive strength
of masonry fc .
Equation 2.8 evaluates the maximum horizontal shear resisted by the pier failing due to rocking.
Mu represents the resistant bending moment of un unreinforced masonry wall, h0 is the effective wall
height defined by its boundary conditions, σ0 is the vertical compressive stress applied and 0.85fc
is the equivalent rectangular stress block that takes into account the vertical stress distribution at
the compressed toe. Rocking and toe-crushing are considered two types of failure modes within the
flexural failure mechanism. However, it is interesting to note that failure defined by overturning is
simultaneously linked to crushing of the compressed corner of the wall.
Mu σ0 D2 t σ0
= (1 − ) (2.8)
h0 2 0.85fc
14
2.3 Textile reinforced mortar
During the last decade, in an effort to overcome certain drawbacks associated to the polymer-based
composites, inorganic matrix composites, also called TRM, have been centred in the spotlight of
the research of the scientific community. These composite systems are also denominated Fibre/Fabric
Reinforced Cementitious Matrix/Mortar (FRCM), Inorganic Matrix-Grid (IMG), and Steel Reinforced
Grout (SRG) if steel cords are embedded in the mortar matrix. Inorganic matrices exhibit lower
bond strength than FRPs due to the possible occurrence of failure modes within the reinforcement
rather than within the TRM-substrate interface. However, TRM advantages in terms of compatibility,
chemical and mechanical, transpirability and reversibility made them appropriate for the retrofitting
of masonry structures. When it comes to masonry substrates, there is also an additional requirement
to be met, the low invasiveness. TRM strengthening systems enable minimal and reversible retrofitting
interventions without altering the permeability of the reinforced elements.
Theoretical and analytical studies have been carried out by different research groups to better
understand the behaviour of TRM strengthening systems. The results have shown the major role
played by the tensile properties of TRM and the bond between the TRM and the masonry substrate.
The fibres comprising the textile mostly provide tensile strength and stiffness, whereas the matrix
allows the stress distribution among the fibres while providing also protection against environmental
actions. Large experimental campaigns carried out by different authors have demonstrated that, unlike
in FRP solutions, the most common failure of TRM strengthening systems is debonding at the matrix-
textile interface. Such failure can strongly affect the tensile behaviour of TRM and consequently the
good adhesion between the textile’s fibres and the mortar is essential to ensure the effectiveness of the
Driven by the potentialities of TRM, significant research have been done, in response to the need
of innovative technologies for improving the safety level of the historical building stock. However,
there is not yet a commonly accepted testing procedure for characterising the tensile and the shear
bond behaviour of TRM strengthening systems. One of the reasons is the major technical issues
related to the different gripping systems used for anchoring the specimen to the testing machine.
Furthermore, there are no available standards, with the exception of the ACI 549 (2013), to provide
either instructions on the TRM testing or design rules for strengthening solutions. As a result, several
test procedures have been developed to characterise the stress-strain behaviour of the TRM under
direct tensile test on TRM coupon specimens. These specimens comprised a single textile layer
sandwiched between two layers of mortar of equal thickness. In addition, different set-ups have been
developed to conduct shear bond tests on specimens strengthened with TRM, with the aim of providing
deeper knowledge of the load transfer capacity between substrate and TRM and their corresponding
failure modes.
With the aim of contributing to the current knowledge of the mechanical properties of TRM
strengthening systems and to the development of standardised experimental procedures, a Round
15
Robin Test (RRT) was organized by the Rilem Technical Committee 250-CSM and compiled in a
number of scientific papers (De Santis et al. (2017a), de Felice et al. (2020)). RRT investigated TRM
composites comprising, among others, basalt and steel textiles embedded in different typologies of
mortar matrices. Both tensile test and shear bond tests, as shown in Figure 2.4, were performed in a
large number of specimens and the most relevant results are shown in the following sections.
Figure 2.4: Tests set-up for the mechanical characterisation of TRM strengthening systems according to
de Felice et al. (2020): a) Direct tensile test on TRM composite specimens, b) shear bond test on brick masonry
substrate
2.3.1 Behaviour of TRM under tensile stress
The resulting behaviour of TRM tested under uniaxial tensile loading is commonly idealized as a
trilinear curve representing three physical states of the TRM composite. The tensile test provides an
insight on the composites Young’s modulus, which is key in the matrix-to-textile bond property since
it affects the cracking of the composite and thus influences the adhesion to the substrate. The tensile
stress-strain behaviour is generally characterised by the three stages show in Figure 2.5:
Stage I: is characterised by a linear elastic behaviour, this stage represents the un-cracked state
of the TRM composite. In this stage the TRM behaves as a composite material in which the
stiffness is dominated by the elastic properties of the mortar matrix. Hence, the slope of the
stress-strain curve reflects the elastic modulus of the mortar matrix EI . The transition to stage
II is defined by the tensile stress σI and the tensile strain εI .
Stage II: the stress-strain curve reflects the formation of transverse cracks within the matrix.
These cracks can propagate along the specimen or be concentrated in a specific location, depend-
16
ing on the tensile properties of the mortar matrix. This stage is characterised by a significant
decrease of stiffness due to the loss of the resisting area (EII ). By the end of this stage the crack
pattern is completely developed.
Stage III: the cracks widen with increasing loading up to the final failure without developing
new cracks. In this stage the textile’s fibres withstand most of the tensile load and consequently
the curve reflects the Young’s modulus of the dry fabric (EIII = Ef ). Finally, the failure is
attained. There are three failure modes: the rupture of the fibre due to the attainment of
its tensile strength, the slippage of the fabric from the matrix or a combination of both. The
mortar matrix although cracked may still provide transversal load redistribution. Generally, the
contribution of the matrix is detected as a higher Young’s modulus and a lower strain associated
to the peak load (εσt ).
Figure 2.5: Tensile response of TRM coupons subjected to direct tensile test according to De Santis et al.
(2017a)
The parameters of stages I and II may have limited importance for the design of strengthening
solutions with rehabilitation purposes. However, they indeed provide information on the textile-to-
matrix stress transfer capacity and on the contribution of the mortar to the mechanical behaviour of
the composite. On the other hand, parameters related to the last stage, more precisely the parameter
associated to the attainment of the peak load (εσt , σt ), are related to the maximum attainable load
by the composite. These parameters are fundamental to determine the fully exploitation of the tensile
strength of the textile.
De Santis and De Felice (2015) carried out tensile test on two types of fabric, glass and steel,
17
to derive their stress-strain curves and to define the mechanical behaviour at the three stages. As
aforementioned, one of the main drawbacks of the direct tensile test is the influence of the clamping
method on the measured tensile strength. The authors approached five different methods to identify
which was more suitable to avoid local stress concentration in the vicinity of the clamp wedge. From
the techniques proposed, the most suitable ones turn out to be the aluminium tabs glued, by means
of FRP, to the bare textile and the direct clamping of the composite coupon, as shown in Figure
2.6-a. The performed tensile tests suggested that the ultimate strength was mainly governed by the
properties of the textile. Nevertheless, due to the stiffening effect of the mortar, the peak strain
was often lower than that of the dry textile. The contribution of the mortar in the un-cracked stage
was more evident on composites with weaker and more deformable textiles, i.e. glass, while almost
negligible on composites with stiffer textile such as steel. In addition, cyclic tensile tests were also
performed to derive residual crack width and hysteretic response. Their outcome evidenced that stiffer
matrices provide worse load transfer leading to wider cracks, which resulted in larger exposition of
the textile to the external aggression. Steel-based composites exhibited better interlocking evidencing
smaller and more distributed cracks development on stage III. However, the denser the steel textile the
larger the crack spacing what led to a splitting failure characterised by the separation of the specimen
into halves and expulsion of large portions of mortar.
Based on the previous results, De Santis and de Felice (2015) continued their research on Steel
Reinforced Grout (SRG) composite as a strengthening solution for masonry structures. The aim
was the identification of the design parameters provided by the mechanical characterisation of the
composite. Tensile tests were performed on dry steel textile specimens and on composite coupons
comprising steel textiles with different densities. The aim of the test was to characterise the entire
response under tension of unidirectional steel textile embedded in lime-based mortar matrix. The
tensile response of the composite coupons was governed by the textile (stage III). Consequently,
mortar cracked in an early stage, the cracks widened with increasing load and finally, the failure
occurred due to tensile rupture, with values of tensile strength and Young’s modulus similar to the
ones obtained by the dry steel. However, this was not the case of the denser steel textile, which showed
a strength reduction due to a premature failure as a result of poor cord interlocking. This feature
was also confirmed with a larger spacing between cracks, when compared to sparser steel textiles.
The phenomenon is related to a low textile-mortar stress transfer capacity which is consequence of a
difficult protrusion of the mortar through the voids of the denser textile.
The aforementioned results are endorsed by the conclusions drawn by a series of direct tensile
tests performed on SRG composites by De Santis et al. (2017b), with the set-up shown in Figure
2.6-b. The outcome evidenced that the lower is the strength of the mortar the lower is its contribution
to the response in the un-cracked and crack development stages, which can be explained in the low
cord-matrix interlocking plus the low mechanical properties of the mortar.
The experimental study on basalt TRM strengthening systems conducted by Caggegi et al. (2017)
involved twenty-two tensile tests on composite coupons. Three different basalt textile reinforcing
18
ratios embedded in two different types of mortar were considered. The results evidenced that the
overlapping of fabric grids, which is translated in a higher reinforcement ratio, reduces the adhesion
between the textile and the mortar promoting premature slippage of the fabric before reaching the
tensile strength of the textile. Conversely, lower reinforcement ratio yielded the higher stress nearly
equal to the tensile strength of the bare basalt fibre. In addition, the failure mode of these specimens
exhibited, in most of the cases, tensile failure of the roving.
Figure 2.6: Overview of test setups for TRM tensile tests: a) The clamping method used by De Santis and De
Felice (2015) was aluminium tabs glued by means of FRP, b) De Santis et al. (2017b) used a direct clamping
method to the TRM coupon, c) Caggegi et al. (2017) used a biaxial machine and directly clamped the TRM
coupon
The literature evidenced the importance of the textile-matrix interlocking and the materials com-
patibility. Considering the same mortar, with low mechanical properties, the stiffer the textile the
lesser the contribution of the mortar to the response not only in the cracked stage (Stage III), but also
in the earliest un-cracked stage (Stage I). Another interesting indicator of compatibility is the crack
distribution, it was demonstrated that the denser the textile the more unfavourable the interlocking.
19
Good interlocking allowed a better crack distribution and hence a better exploitation of the tensile
strength.
Notwithstanding the importance of tensile characterisation of TRM coupons, the literature em-
phasized the influence of the substrate, on which the reinforcement is applied, to the ultimate capacity.
Therefore, the tensile behaviour of TRM coupons has to be combined with the study of TRM-substrate
bond behaviour to better understand the strengthening solutions. In addition, the failure modes ob-
served during tensile test, may change during shear bond test as a consequence of the substrate
presence, resulting into the detriment of the tensile strength efficiency.
2.3.2 Bond strength of TRM on Masonry substrate
The characterisation of the bond between TRM and the substrate is crucial to assess the contribution
of TRM to the shear capacity of masonry walls. It is also of interest to study the failure mechanism of
TRM, when subjected to a parallel force, to better understand the behaviour of strengthened masonry
walls. Many factors influence the bond strength of the TRM strengthening systems, i.e the mechanical
properties of the textile and the matrix, the mortar capability to penetrate into the grid, the friction
between the textile and the mortar and finally, the bonded length of the reinforcement.
The shear bond strength is evaluated through the single lap or double-lap shear test, being the
former the most common one. The single-lap shear bond test is performed on masonry prisms on
which the reinforcement is applied, in the direction orthogonal to the mortar bed joints, to one side of
the substrate, as shown in Figure 2.4-b. The masonry prism is kept fixed by means of a loading frame
and a portion of the textile is left unbounded, clamped and pulled by means of a testing machine.
In the double-lap shear bond tests, a U-shaped textile strip is bonded to the two sides of a substrate
block/prism, which is kept fixed. A saddle or a cylinder is used to pull the textile. These tests evaluate
the effectiveness of the interface TRM-substrate to transfer the shear load activating the weakest failure
mechanism. Along with the peak force, the relative displacement between reinforcement and substrate
at the loaded end are recorded during the testing. The different failure modes that may occur are
shown in Figure 2.7 and listed in the following:
(A) debonding with cohesive failure in the substrate
(B) debonding at the TRM-substrate interface
(C) debonding at textile-to-matrix interface
(D) sliding of the textile within the mortar matrix
(E) tensile rupture of the textile
(F) tensile rupture of the textile within the mortar matrix
Failure modes A to C generally give rise to brittle failure, while D is generally related to a soft load
decrease due to the progressive loss of friction of the textile sliding within the matrix. Conversely,
20
Figure 2.7: Failure modes in shear bond tests on TRM/FRCM systems defined by the TC RILEM 250
CSM De Santis et al. (2017a): (A) debonding with cohesive failure in the substrate, (B) debonding at the
TRM-substrate interface, (C) debonding at textile-to-matrix interface, (D) sliding of the textile within the
reinforcement thickness, (E) tensile rupture of the textile, (F) tensile rupture of the textile within the mortar
matrix (image reproduced from Ascione et al. (2015))
failure mode E is associated to instantaneous load reduction at tensile failure. Finally, failure mode
F is activated by telescopic rupture of the bundles followed by the slipping of the textile out of the
mortar.
The outcome of a wide experimental programme framed in a RRT, allowed De Santis et al. (2017a)
to concluded that cohesive debonding (A) may occur when a strong matrix is bonded to a relatively
weak substrate while detachment at the TRM-substrate interface (B) may take place on smooth
surfaces. Failure mode (B) can also be caused by inaccurate installation, since masonry substrate
tends to be uneven in most cases, hindering a good grip of the mortar. Detachment at textile-matrix
interface (C) may occur with dense textiles, where proper protrusion of matrix though the voids
is not assured. However, in the cases of Steel Reinforced Grout (SRG) composites this failure was
observed in combination with tensile rupture of the cords. TRM strengthening systems with relative
weak interlocking often exhibited slippage failure (D) while TRM strengthening systems, composed
by sparser meshes of relatively weak textiles such as basalt, failed by textile rupture (E). This last
failure mode, normally is the one that allows the full exploitation of the textile strength.
Current literature, focused on the bond characterisation between TRM and the substrate, have
highlighted that the weakness of the TRM strengthening systems is actually at the textile-matrix
interface rather than the TRM-substrate interface, as it was on FRP solutions. Particular attention
has been paid to basalt and SRG in the last couple of years. In the case of the basalt TRM the
attention is focused on its natural origins, featuring it as an eco-friendly material. In the case of SRG
its low cost, compared with other textiles such as carbon or aramid, makes it appealing as a new
strengthening system.
21
The shear bond test on TRM strengthening systems applied on masonry substrate was first con-
ducted by D’Ambrisi et al. (2013), who analysed the effectiveness of the bond between carbon TRM
applied on clay brick masonry. Eight specimens, with three different bond length 110, 230 and 350 mm,
were tested under double-lap shear bond test. The research aimed to identify the influence of bond
length in the debonding force and ultimately to define an effective anchorage length. The experimen-
tal analysis has evidenced that debonding occurred at the textile-matrix interface and no detachment
from the substrate is observed. The debonding mechanism consisted mainly on the gradual loss of
bond between the textile’s fibres and the cementitious matrix. The maximum force transferred from
the substrate to the textile was practically independent of the bond length studied, hence, the effective
anchorage length was set equal to the lower bonded length 110 mm.
In order to broad the knowledge on the mechanical performance of reinforced systems embedded
in inorganic matrices a joint effort was done by different laboratories and presented by de Felice et al.
(2014). The experimental programme comprised the tensile characterisation of the coupons, through
direct unidirectional tensile test, and the TRM-substrate bond performance through single or double
lap shear bond tests, of three strengthening systems. The systems involved SRG, carbon and basalt
textile reinforced mortar (CTRM and BTRM), all three embedded in cement-based and lime-based
mortar matrices. The shear bond tests aimed to investigate the anchorage length and its effect on the
ultimate load. For this reason, different anchorage length, ranging between 55 to 440 mm, were tested.
The experimental programme made possible the identification of the shear bond failure modes of the
TRM systems. It was clearly evidenced the change on failure mode undergo by the mortar based
systems, when compared to the polymeric based ones. The prevalent failure mode was within the
reinforcement and without involving the substrate. The exploitation of the textile strength, defined
as the ratio between the debonding load and the tensile failure load, was found to be higher as the
anchorage length increased. On the base of the bond tests results, the anchorage length for each type
of TRM strengthening systems were derived. It was estimated to be around 220 mm in the case of
SRG and 165 mm in the case on CTRM, while for BTRM was not possible the derivation of the length
due to the absence of a reliable strain profile.
Due to the lack of standards that provides practitioners with recommendations for the testing,
design and control of mortar-based systems, (as only the guidelines issued by the American ACI and
Italian CNR are available), Ascione et al. (2015) proposed a procedure for the qualification of TRM
reinforcement by combining the results from the direct tensile test and the shear bond test to provide
the mechanical parameters for the design of TRM solutions. According to the authors, the evaluation of
TRM-substrate shear bond performance is crucial for applications in which in-plane failure is expected
at the TRM-substrate interface. This is the case of retrofitted masonry walls subjected to in-plane
actions, and in which the reinforcement remains effective as long as the mortar ensures the stress
transfer from the structure to the textile, which ultimately is the responsible of carrying the load.
Ascione et al. (2015) analysed five different TRM strengthening systems, among these systems were
basalt grid and unidirectional steel textile embedded in lime-based mortar and applied on historical
22
clay brick and tuff brick masonry substrates. The results of shear bond tests, shown in Figure 2.8,
on specimens with basalt evidenced no debonding neither at the substrate nor at the textile-matrix
interface. The failure was mainly related to the rupture of the textile (E). It was observed a telescopic
failure of the fibre as a result of a non-uniform stress distribution across the sections. Conversely, steel
textile experienced debonding failure at textile-matrix interface (C). It is interesting to note that in
the tensile tests of the TRM coupons of steel textile embedded on lime based mortar, the contribution
of the latter was nearly negligible even before cracking. The authors reached to the conclusion that
in some cases, the mortar can provide effective load transfer between substrate and textile even if
cracked and in the cases of failure due to rupture of the fibre, the maximum tensile stress reached on
the shear bond test was closed to the tensile strength obtained on the tensile test.
Figure 2.8: Failure modes observed on the shear bond tests perfomed by Ascione et al. (2015): a) Basalt TRM
was characterised by fibre rupture (E), b) SRG experienced debonding failure at textile-matrix interface (C)
The overcoming of some of the testing drawbacks, steam mainly from the gripping systems and the
alignment of the strengthening strip, and with a new procedure to evaluate and assess the results, the
interest of the scientific community have grown around the potentiality of new TRM strengthening
systems, mainly on Steel Reinforced Grout (SRG). A rather large number of campaigns were designed
and conducted and the results have shed some light on the behaviour of TRM.
De Santis and de Felice (2015) characterised SRG system by carrying out shear bond test on
unidirectional steel textile of different densities embedded in lime-based mortar matrix and applied
on weak clay brick and historical clay brick substrates. Good cord-mortar interlocking impeded the
sliding of the textile within the matrix and the high tensile strength of the steel cord avoided the
tensile rupture. Figure 2.9 shows the most prevalent failure modes. The SRG system with less dense
23
textile and lime-based mortar mainly failed by detachment at textile-matrix interface (C), indicating
that the weaker is the matrix the lower is the load transfer performance with the textile. Also, the
minimum preparation on the substrate, which is the case of historical brick, led to a failure at the
matrix-substrate interface (B). The higher density reduced the transfer capacity with the mortar and
caused lower bond strength. It was concluded that the bond performance of SRG systems relies on
the continuity of the mortar matrix in the cross section of the reinforcement which in turn depends
on the protrusion of the mortar through the voids of the textile. Therefore, higher exploitation ratios
of the tensile strength of the textile were evidenced in less dense steel textiles.
Figure 2.9: Failure modes observed in the shear bond test performed in De Santis and de Felice (2015):
debonding at the mortar-to substrate interface (mode B) and shear failure at the textile-to-matrix interface
(mode C)
The RRT carried out by De Santis et al. (2017b) on bond behaviour of SRG systems indicates
that the bond strength and the failure modes of the SRG systems are ascribed to several factors, the
strength of the textile, the cord-matrix interlocking and the mechanical properties of the matrix. Poor
interlocking may encourage sliding of the textile from the matrix. Once again is asserted that a stiffer
matrix leads to a failure mode involving the debonding from the substrate, while more deformable
mortar allows a uniformed distribution of the stress at the TRM-substrate interface and the attainment
of higher bond strength. From the results can be concluded that failure modes involving SRG-substrate
interface and textile-matrix interface were the most recurrent ones and were associated with good
exploitation ratios of tensile strength of the textile.
Santandrea et al. (2020) designed and performed an extensive experimental campaign to study
the bond performance between SRG systems and masonry joints by means of single-lap shear test.
It must be highlighted that most of the previous research on SRG systems was performed on a
single brick as substrate, neglecting masonry mortar joints, as evidenced in Figure. In addition to
24
the shear bond performance, different bonded length ranging from 75 mm to 345 mm were also
tested. Unidirectional high strength steel fibre strips embedded in a lime-based hydraulic mortar
were applied on masonry substrate and a total of seventy-eight samples were tested. The result
evidenced that for bonded length up to 100 mm the failure mode was always due to debonding
from the substrate (B). Whereas, specimens with bonded length greater than 200 mm were mainly
characterised by failure at textile-matrix interface (C). Regarding the effective bonded length, the
outcome of the experimental programme asserts the previous findings on Santandrea et al. (2017), in
which the effective bonded length was set equal or greater than 200 mm for SRG systems applied on
substrate comprising stacked masonry prism. Wang et al. (2020) carried out single-lap shear bond
Figure 2.10: a) Set-up used by Santandrea et al. (2020) to performed the shear bond test on stack bond
prisms, b) set-up used by Ascione et al. (2015), De Santis and de Felice (2015) to performed the shear bond
test on single bricks
test on SRG systems to characterise the bond behaviour of the composite applied on grey clay bricks
masonry prism. The steel textile was investigated in two densities, low and medium density, and was
embedded in lime-based matrix. The two densities demonstrated not only a different exploitation
ratio of the tensile strength of the textile but also led to a shift in the failure mode. Such shift was
also observed with the bonded lengths. The medium density steel textile evidenced significantly lower
efficiency (16%) followed by failure at the TRM-substrate interface (B). While the low-density steel
textile failed at the textile-matrix interface (C) with a higher utilization of tensile strength of the steel
cord (62%).
The experimental results highlight the benefits of using SRG with less dense steel textile embedded
in a low strength lime-based mortar, as strengthening solutions for clay brick masonry. The compati-
bility between the weak mortar matrix and the weak masonry substrate, allows a good bond at that
interface, moving the failure towards the textile-matrix interface. The sparser is the textile the better
interlocking at textile-matrix level, which avoids slippage of the textile and encourages the debonding
25
within the matrix. This failure allows the maximum exploitation of the tensile strength of the steel
textile.
Notwithstanding the promising outcome of SRG as a strengthening system, the failure mode as-
sociated to it, when applied to a weak substrate such as masonry, is still brittle and sudden. For
its application to existing masonry members, the stiffness of the strengthening systems should be
sufficiently limited as to avoid excessive overall stiffness increments that could affect the overall be-
haviour of the strengthened structures. As aforementioned, a more deformable textile such as basalt,
would allow not only a better compatibility with the substrate, which would avoid any detachment or
debonding, but also would promote failure due to rupture of the fibre. Consequently, the exploitation
ratio of the tensile strength of the textile would be maximised.
An experimental campaign was designed by Santandrea et al. (2016) to study the stress transfer
between TRM strengthening systems and masonry substrate, and their corresponding failure mecha-
nisms. To this aim, thirty-two masonry prisms, comprised of seven stacked half bricks, were bonded
with basalt TRM of two different densities, low and high density. Specimens with low density basalt
fibres showed a failure due to rupture of fibres. However, the maximum tensile stress attained was
lower than the one provided by the dry fibre. On the contrary, high density basalt fibres experienced
a failure consisted in the detachment of the top mortar layer from the bottom one, mainly due to the
reduced area of mortar in between fibres bundles what led to a poor textile-matrix interlocking, as
evidenced in Figure 2.11. The phenomenon associated to poor interlocking due to higher density of
textile, was also evidenced in the experimental research of Santandrea et al. (2017), were the higher
density of steel fibres evidenced a reduction of the load carrying capacity of the textile-matrix interface.
Figure 2.11: Failure modes according to Santandrea et al. (2016): Low density basalt fibre failure (on the
left), and debonding mechanism for high density basalt (on the right).
Fifty-two single lap shear tests, involving the pulling of 260 mm basalt TRM (BTRM) strips
bonded on one side of masonry prisms comprising five stacked bricks, were performed by Lignola et al.
26
(2017) in the framework of RRT. The main aim of the RRT was assessing the bond strength of the
strips in terms of exploitation ratio of the strip’s tensile capacity. As a general result, the debonding at
the matrix-substrate interface was an infrequent failure mode, inferring that 260 mm was a sufficiently
bonded length for the exploitation of the textile’s tensile strength. The main conclusions drawn from
the experimental campaign suggest that when bond performance is good, the failure mode tends to
fibre rupture (E) and consequently, the tensile stress attained at single laps shear test is similar to the
tensile strength of the bare textile.
Encouraged by the results obtained by Caggegi et al. (2017) on tensile tests of coupons of BTRM,
Ferrara et al. (2019) aimed at shedding some light on the influence of the reinforcing ratio on the
shear load capacity when subjected to shear bond test. The BTRM systems were applied, with a
bonded length of 260 mm on masonry substrate characterised by five clay bricks and four mortar
joints and tested with single-lap shear test. The test aimed to reproduced an intermediate crack-
induced failure mode, characterised only by shear stress distributed on the interfacial surface between
the substrate and the TRM strengthening system. The results led to the conclusion that the increasing
reinforcement ratio caused a significant reduction of the space between the threads and consequently
a less performing strengthening systems, which also resulted in a lower tensile stress achieved by the
BTRM. The BTRM with lower reinforcement ratio attained higher loads and the failure mode that
characterised it was governed by the rupture of the fibre within the mortar matrix (F). Finally, when
the results were compared with the ones of Caggegi et al. (2017) it was highlighted the importance of
performing both test, tensile and shear bond tests, to characterised the mechanical behaviour of TRM
The experimental programme carried out by Barducci et al. (2020), envisaged the evaluation of the
bond behaviour of basalt TRM applied on clay brick substrate, as shown in Figure 2.12. To this aim,
six single-lap shear bond tests were performed. The basalt grid embedded in lime-based mortar was
laid on clay bricks, duly moistened, with a bonded length of 220 mm. The results evidenced tensile
failure due to rupture of the fibre (E), which showed good bond capacity not only in the textile-matrix
interface, but also in the matrix-substrate interface since no debonding was observed. Moreover full
exploitation of the strength of the textile was detected. The fibre rupture was accompanied by the
typical telescopic behaviour.
Finally, all the results obtained by the Round Robin test organized by RILEM TC 250-CSM, and
presented in Lignola et al. (2017), De Santis et al. (2017b), Caggegi et al. (2017), were analysed all
together by de Felice et al. (2020). The following conclusions were drawn from the responses recorded
for each type of strengthening system. Basalt textile is among the ones that presents relatively low
strength and Young’s Modulus. These features make basalt suitable for applications where large
surfaces of masonry walls need to be strengthened. In addition, the bidirectionality of the grid hinders
cracking in multiple directions. Steel textiles on the other side have high strength but yet intermediate
Young’s Modulus, being appropriate for applications where load application prevails in one direction.
The failure of composites with basalt textile is manly driven by tensile rupture (E). Conversely, in
27
Figure 2.12: Single shear bond test set-up used by Barducci et al. (2020)
system with unidirectional steel fibres the bond capacity relies on the interlocking between the textile
and the matrix. Despite the type of textile under investigation, namely basalt or steel, the tensile
strength of the composite is proportional to the density of the yarns, while the shear strength is
promoted by the protrusion of the matrix in the spaces between adjacent yarns. Therefore, sparser
yarn spacing leads to a better textile-matrix interlocking which would benefit greatly the shear stress
transfer capacity of the reinforcement.
What is more, from the test performed on homogeneous brick substrates and the ones conducted
on brickwork prisms, mainly stacked masonry prisms, the results obtained from the latter assert that
the presence of mortar joints provides higher bond strength and the masonry mortar joints are not
negligible since they provide good gripping.
Despite the knowledge developed so far on the mechanical behaviour, the qualification and accep-
tance of mortar-based composite is still an open issue, since there is only one available guideline, and
the retrofitting design and evaluation are not included in any building code regulation. Therefore,
research on large specimens strengthened with different TRM strengthening systems is still needed.
2.4 In-plane shear behaviour of strengthened masonry
In recent years there has been a marked increase in experimental research oriented towards the in-plane
behaviour of masonry walls strengthened with TRM strengthening systems with the aim of determine
their efficiency for seismic retrofitting techniques. In the available literature different types of shear
test are performed on large specimens of masonry to characterised their in-plane behaviour, namely
Diagonal Compression Test (DCT) and Shear Compression Test (SCT). However, the considerable
portion of this research has been carried out with DCT configuration, lacking further literature on
SCT configuration. Despite the good results obtained from DCT, it is well known that SCT is more
28
representative, in terms of boundary conditions and loading conditions, of the real in-plane behaviour
of masonry elements such as shear walls in masonry buildings.
The results provided by the literature, hereby presented, could be of great interest since it can
provide a more specific insight on the behaviour of in-plane strengthened masonry walls for improving
externally bonded rehabilitation techniques for practitioner.
2.4.1 Diagonal compression test (DCT)
The standards ASTM E-519 (2000) and RILEM TC-76 (1994) define the standard test method for
determining diagonal tensile strength in unreinforced masonry walls. Later on, this testing method-
ology was extended to assess the tensile strength of retrofitted masonry. Diagonal compression test,
even though it is an indirect shear test, is preferred over the shear compression test because it is less
expensive and the set up preparation is less time consuming both in the laboratory and on-site.
Prota et al. (2006) performed DCT on eight tuff masonry panels retrofitted with a strengthening
system consisting of a glass coated grid embedded in a cementitious matrix (CMG). Four different
configurations were tested, among which asymmetrical and symmetrical configurations with one and
two plies have been assessed. From the results it can be evidenced that the CMG system reduces the
high anisotropy inherent to tuff masonry. Moreover, the outcome confirmed the effectiveness of the
strengthening improving the strength and ductility of strengthened panels. The increment was most
notorious on the symmetrical configuration with two plies of reinforcement, which was accompanied by
a moderate increment of stiffness. On the other hand, the asymmetrical strengthening evidenced out-of
plane deflections due to the eccentricities generated on the composite grid by the axial forces, which led
to a brittle failure and consequently lower strength increase. Balsamo et al. (2011) carried out a similar
experimental programme, testing four tuff masonry panels in two different configurations, the first one
involving a glass grid embedded in lime-based mortar and the other a basalt grid also embedded in
lime-based mortar matrix, both configurations were symmetrical. Specimens strengthened with glass
grid evidenced lower increment of strength when compared to the results presented by Prota et al.
(2006). However, the increment (between 150-200%) was followed by a better post-peak response
when compared to the cement-based mortar matrix. Similarly, the basalt grid showed more ductile
behaviour than the glass grid.
Parisi et al. (2013) assessed the effectiveness of the inorganic matrix-grid (IMG) strengthening
systems to improve the shear behaviour of tuff masonry located in seismic prone areas. The double-
sided configuration, with and without transverse connectors, and the single-sided configuration were
assessed. The strengthening systems, including glass grid IMG, contributed to the inelastic redistri-
bution of seismic demand throughout the structure, evidencing that higher ductility lead to larger
displacement and energy dissipation capacity. The increment of ductility was specially remarkable
in the cases of specimens with single-sided configuration with transverse connections of steel fibre-
reinforced polymer (SFRP) ties. The ties prevented the out-of-plane bending previously evidenced in
the specimen lacking such connectors. Particularly, it was observed that in the case of tuff masonry
29
retrofitted with glass grid, the double-sided configuration specimens and the ones with single-sided
configuration and SFRP transverse connector yielded the same shear strength increase. Moreover, the
absence of premature debonding demonstrated the good bond performance of the IMG systems as a
result of the good compatibility given by the mechanical anchorage of the lime-based mortar.
Ismail and Ingham (2013) are among the first ones in performing a research on the efficiency of
TRM strengthening systems for retrofitting two leaf clay masonry walls by means of DCT. The ex-
perimental programme involved the testing of ten wallettes retrofitted with either steel mesh or fibre
glass grid in the symmetrical and single-sided configuration. Interestingly, the mortar matrix was
mixed with polymeric adhesive and as a result some specimens evidenced sudden brittle failure while
others exhibited debonding from the substrate. Similar results were observed in Babaeidarabad et al.
(2014) who tested carbon fibre TRM, which involved several plies of fibre embedded in cementitious
mortar reinforced with short fibre, applied on clay brick masonry walls with running bond pattern.
In this latter case, the excessive reinforcement level, four plies of textile, led to masonry crushing,
while the specimen with only one ply of reinforcement failed due to fabric slippage within the matrix.
However the undesired failure modes, an increment of ductility and shear strength was observed in
most of the specimens. Most importantly from the specimens response it was evidenced the ineffec-
tiveness of increasing the reinforcement level to increase the shear capacity. In addition, Ismail and
Ingham (2013) indicated the lack of design equations for TRM and therefore they carried out their
design following manufacturer recommendations. Conversely, Babaeidarabad et al. (2014) applied the
analytical formulations provided by ACI 549 (2013) and concluded that the predictive formulation
overestimates the shear capacity of the strengthened specimens.
Specimens strengthened with Basalt TRM
The environmental awareness has recently shifted the attention towards new low environmental impact,
eco-friendly and natural materials for construction and repair. As a result, reinforcing textiles from
natural fibres are now under the spotlight, including hemp fibre and basalt fibres. Menna et al.
(2015) developed a strengthening system comprising dry hemp cords, arranged in a bidirectional grid,
impregnated with epoxy resin and embedded in lime-based mortar matrix. This strengthening system
was applied on both tuff masonry walls and clay brick masonry walls and later subjected to DCT.
Different thicknesses of the matrix layer were tested, namely 15 and 40 mm, evidencing that higher
thickness of the matrix layer presents higher stiffness in the elastic phase and less wide softening branch.
The failure of the retrofitted specimens was characterised by diagonal cracking on the compressed strut
and the hemp grid reached its ultimate strain failing in a brittle fashion. The failure did not evidence
premature debonding between the grid and the mortar revealing that the strengthening system was
not only compatible with the substrate but also that the tensile load was successfully transfer to the
hemp cords up to their failure. Finally, the experiments showed an enhancement of the mechanical
properties of the masonry panels in terms of maximum shear strength (up to 325%). In addition, a
ductile behaviour was facilitated by the low stiffness of the hemp cord.
30
Figure 2.13: Overview of the different set-ups used by the authors aforementioned: a) Prota et al. (2006), b)
Balsamo et al. (2011), c) Parisi et al. (2013), d) Ismail and Ingham (2013), e) Babaeidarabad et al. (2014)
Regarding the use of basalt bidirectional grid, the literature shows only limited experimental re-
sults on the application of basalt TRM (BTRM) to clay masonry walls, since most of the research
was mainly developed for tuff masonry. In order to improve the knowledge on BTRM strengthening
systems Marcari et al. (2017) discussed the response of tuff masonry walls retrofitted with basalt tex-
tile embedded into hydraulic-lime mortar matrix. The experimental programme involved three tuff
masonry panels retrofitted with BTRM, two of them in double-sided configuration and the remain-
ing one in the single-sided configuration. Unlike glass fibre or carbon fibre, the BTRM reached its
tensile strength before undergoing debonding or delamination, which suggests a better mechanical
compatibility. In addition, the authors were the first ones to highlight that the BTRM symmetrically
applied, this is on both sides of the wall, evidenced a clear reduction of the variability of the overall
shear-deformation response of the panels.
Del Zoppo et al. (2019) assessed the in-plane shear capacity of nine double leaf solid clay brick
masonry walls strengthened with BTRM. The strengthening system consisted of a basalt grid em-
bedded in a lime-based matrix reinforced with a content of short glass fibres. The author pointed
out that the selection of basalt was due to its better performance compared to glass and its greater
compatibility with masonry mechanical properties with respect to carbon and steel, which have higher
tensile strength and elastic modulus, which was also highlighted by Lignola et al. (2017). The good
compatibility was reflected in no debonding phenomena observed at the end of the testing, as shown in
31
Figure 2.15-a. An interesting observation is that the pattern of the double leaf masonry had poor con-
nection between consecutive brick leaves since headers were only put at the end of a single row. This
pattern affected the diagonal compression capacity of the URM specimen even after the retrofitting
procedure since the headers’ position at the edge of the panel did not influence on the diagonal com-
pression capacity of the URM. Finally, the experimental results were compared with the prediction of
the current codes provisions for FRCM strengthened systems ACI 549 (2013). The scattering between
results was attributed to the limitation of tensile strain to 0.004, which the codes limits to avoid large
cracks in the matrix.
Following the results obtained by Marcari et al. (2017), Basili et al. (2019) designed an experi-
mental programme involving six clay brick masonry walls laid on running bond and retrofitted with
BTRM. Three specimens were retrofitted on one side only while on the other three BTRM was applied
symmetrically. Unlike the specimens tested in Marcari et al. (2017), the retrofitted specimens did not
have any type of anchorage; however, the BTRM layer remained perfectly bonded to the masonry sur-
face, evidencing good compatibility, increasing load carrying capacity (between 30-60%) and showing
higher deformation capacity. Comparing experimental results obtained by applying BTRM on differ-
ent masonry typologies, namely tuff and clay brick, it is asserted the similarities on the increment of
shear strength, ductility and also the qualitatively similar increase of initial stiffness even if the results
appeared more dispersed. Besides the assessment of the efficiency of BTRM applied on clay brick
masonry, the authors performed a numerical simulation of experimental test on reinforced panels with
MIDAS-FEA code as shown in Figure 2.14. The BTRM reinforcement was described as isotropic con-
tinuum material, characterizing it by suitable constitutive relationships described by few mechanical
parameters obtained from the experimental data from DCT. The analysis concluded that despite the
simplicity of the model and the limited number of material parameters utilized, the numerical model
accurately predicted the in-plane shear behaviour of BTRM reinforced masonry walls.
Figure 2.14: a-b) Maximum and minimum principal stress results from FEM analysis performed by Basili
et al. (2019), c) Crack pattern at the end of the test of the specimen symmetrically reinforced
Pre-damaged specimens repaired with TRM
After small or moderate seismic events, diagonal shear failure is observed in masonry structural mem-
bers, namely load bearing walls. In most cases the damage is not repaired, only covered. However,
the load bearing walls remained vulnerable against in-plane actions. In this context, in the last couple
32
of years, TRM strengthening system was not only evaluated as a seismic retrofitting solution but also
as a post-earthquake repairing technique. There is a current need to assess the mechanical behaviour
of repaired walls, but little research has been carried out on pre-damaged masonry walls. Incerti
et al. (2018) performed DCT on four fired-clay bricks and lime-based mortar masonry walls, after
testing the panels were reinforced with bidirectional grid BTRM embedded into lime-based mortar
matrix, without performing any previous repairing intervention. Several cracks were observed within
the matrix, corresponding to the pre-existing cracks, while further damage was limited to small por-
tions of bricks along the main facture line. Finally, the failure was driven by the rupture of fibres.
The TRM reinforcement worked during the whole testing, at the beginning the matrix was effectively
collaborating with the masonry substrate suggesting that during the reinforcement installation the
first layer of mortar cover, and partially restored, the pre-existing cracks. After this point the BTRM
grid started working carrying the tensile load that the substrate was no more able to sustain. Due to
the lack of a proper repair before retrofitting the damaged specimens, the samples were not able to
reach the peak load corresponding to the URM specimens. Mustafaraj and Yardim (2019) designed an
experimental programme involving fifteen two-leaf English bond pre-cracked masonry walls repaired
with three different strengthening repairing techniques: polypropylene fibre reinforced mortar repair,
ferrocement jacketing repair and carbon fibre reinforced polymer wraps repair. From the results, it
was observed that all the three repair methods were effective. Some of the specimens retrofitted with
polypropylene evidenced little increase of both shear strength and strain. From all the repaired panels
tested, ferrocement jacketing technique achieved the best results in terms of ductility. However, due
to the high stiffness of the fibres, the failure modes were mainly characterised by detachment and
debonding from the substrate, evidencing the incompatibility with the masonry under study.
Specimens strengthened with steel reinforced grout (SRG)
Among the various typologies of TRM strengthening systems, steel reinforced grout (SRG) has demon-
strated good mechanical performance when applied on clay brick masonry walls. After the detailed
mechanical characterisation provided by De Santis and De Felice (2015), an experimental campaign
was design by Wang et al. (2018) in order to assess the effectiveness of SRG composites applied on six
grey clay brick masonry walls. The authors investigated two different steel cord densities, namely low
and high density, applied in three configurations: strips in single direction, vertical or horizontal and
strips together in vertical and horizontal directions. The results evidenced the low enhancement effi-
ciency of the specimens retrofitted with only vertical strips in terms of shear strength and stiffness, as
shown in Figure 2.15-b. Nevertheless, the vertical configuration evidenced better enhancement of the
ductility compared to the horizontal one. The failure modes were mainly characterised by debonding
and delamination and no rupture of the steel cord was observed. The specimens retrofitted with a single
direction of SRG undergo large stress concentration at the edges of the strips leading to the detach-
ment of the SRG layers from the substrate. These results indicate that steel cords provide only tensile
stress to increase the shear resistance of the masonry walls and in turn, the strengthening efficiency
33
is influenced by the orientation of the SRG strips, namely vertical or horizontal. The experimental
results were compared with the analytical methodologies available to predict the shear behaviour of
strengthened walls. Due to the strips configuration, the contribution of the SRG was computed as the
sum of the contribution of the strips in each direction. The contribution of the horizontal strips was
computed following the analytical methodology proposed by EC6 (2006) for embedded reinforcement,
namely FRP. Although some methodologies disregard the contribution of the vertical strips, Silva
et al. (2008) proposed the computation of their contribution proportionally to the contribution of the
horizontal strips. Unfortunately, the followed analytical methodologies showed large overestimation
of the shear contribution of the SRG system, as a result of considering the ultimate tensile strength
of the steel cord even though it was never reached. Conversely, the analytical methodology presented
in ACI 549 (2013) predicted the shear contribution of the horizontal reinforcement with fairly good
accuracy since the debonding failure of SRG was taking into account in the computation.
Specimens strengthened with Joint repointing Near Surface Mounted (NSM)
In the cases of masonry façades with exposed bricks, the use of externally bonded strengthening solu-
tions, namely TRM with grid textile or unidirectional SRG strips, may not represent a feasible solution
because it may violate aesthetic and conservation requirements. As a result, repointing techniques
has been developed as they evidenced to be minimally invasive, while respecting the aesthetic of the
exposed bricks façades. The joint repointing technique, also called Near Surface Mounted (NSM)
reinforcement, involves the application of a lineal shaped material with high tensile strength, namely
steel bars or carbon wires, in the horizontal bed joint of the wall to reduce the in-plane vulnerability
of masonry structures.
Ismail et al. (2011) designed a testing programme involving twisted steel reinforcement bars in-
troduced into URM wallettes using NSM techniques, where thin slots were cut into the surface of
the masonry and the reinforcement was bonded into the slot using high bond strength cementitious
grout. The main advantages of the chosen material were the ease of application of the grout and its
compatibility with the porous heritage URM material, when compared with epoxy resins, and the
high corrosion resistance that characterise the twisted stainless steel bars. Fourteen solid clay brick
masonry wallettes were strengthened with different orientations of stainless steel NSM reinforcement
and tested under DCT to investigate their in-plane performance. Only three out of fourteen speci-
mens were reinforced embedding the steel bars in the mortar bed joints by employing a cementitious
grout. Even though this horizontal reinforcement scheme was effective in bridging diagonal cracks and
allowing the wallettes to deform further until ultimate failure occurred, its behaviour resulted in a
smaller increase in shear strength when compared to other reinforcement schemes, but a large increase
in ductility was observed. Finally, the helical profile of the stainless steel reinforcing bars resulted in
excellent mechanical bond between the reinforcement and masonry over short bonded lengths.
Casacci et al. (2019) performed DCT on six clay brick masonry wallettes reinforced with basalt
bars to assess their efficiency to enhance the in-plane response. The basalt bars were applied in the
34
symmetrical and asymmetrical configurations. Unlike the other TRM strengthening systems, NSM
reinforcement did not change the failure mode of URM from stair-stepped to tensile diagonal cracking
and the load carrying capacity increment was not substantial. However, the NSM showed to provide
additional ductility and energy absorption capacity to the masonry, as shown in Figure 2.15-c. The
most remarkable change highlighted in the specimens retrofitted with basalt bars with respect to URM
panels was the increment in displacement capacity. Finally, the analytical methodology developed by
Li et al. (2005) was applied to predict the shear contribution of the NSM reinforcement. For specimens
reinforced on one or both sides with the repointing technique, analytical results overestimated the
strength. Particularly for the specimen asymmetrically retrofitted the scattering between experimental
and analytical results was due to an overestimate of the contribution of the bars. However, the
discrepancies between experimental and analytical values in the case of NSM reinforcement can be
ascribed to different factors since the methodology is based on several hypothesis. Among these factors,
the bond between the grout and the bar and the uniform distribution of the shear stress along the
embedded length of the bar are considered.
Preliminary remarks
The application of TRM strengthening system has evolved in the last years and its effectiveness as
in-plane enhancement of a variety of masonry typologies have been proved by means of Diagonal
Compression Test (DCT). The compatibility of inorganic lime-based mortar matrix with masonry
substrates has been proved over the compatibility provided by cementitious-based one, minimizing the
failure due to detachment or debonding from the substrate. The strengthening systems, involving low
stiffness fibre grids embedded in lime-based mortar matrices, have shown to provide tensile strength
to masonry walls after cracking. As a result, retrofitted specimens have evidenced a shift in their
failure mode, from brittle stair-stepped or joint sliding to tensile diagonal cracking feature by smeared
cracks along the compressed strut of the wall. The application of several plies of reinforcement have
evidenced to be detrimental for the in-plane response of masonry walls, since it may lead to failure
due to masonry crushing. Aside from the typology of TRM and the level of strengthening, the
symmetrical configuration has shown to be the most effective one; however, when this case turns out
to be unfeasible, single-sided configuration with connectors to the unreinforced side has evidenced to
hinder the out of plane failure.
Finally, the good results obtained by all the experimental programmes aforementioned suggests
that DCT is a simple testing methodology to characterise retrofitted masonry walls. However, there
are still the uncertainties raised from the interpretation of the tests results, derived from the testing of
URM under DCT, that might question the real enhancement of the TRM in terms of shear strength
and stiffness. To overcome this drawback, a number of researchers focused their attention on the
Shear Compression Test, to assert the efficiency of TRM strengthening systems as seismic retrofitting
technique.
35
Figure 2.15: Overview of results of DCT in terms of crack patterns and strain-stress experimental curves: a)
Del Zoppo et al. (2019), b) Wang et al. (2018), c) Casacci et al. (2019)
36
2.4.2 Shear Compression Test (SCT): Cyclic in-plane loading
Shear compression tests (SCT) are typically used to characterise the force and drift capacity of in-plane
loaded unreinforced and retrofitted masonry walls. During testing, the masonry wall is subjected to
horizontal in-plane displacements while being vertically loaded at the same time. According to Wilding
et al. (2018) when conducting a SCT, there are two decision that needs to be taken regarding the three
in-plane degrees of freedom at the top of the wall. The first one concerns the horizontal displacement,
which can be monotonically or cyclically applied. The second one concerns the control of the axial
force and the in-plane moment on top of the wall.
Commonly SCT is simulated either in cantilever or double-fixed boundary condition configurations.
In turn, the latter configuration can be applied in three different ways, which differ regarding the
control of the degrees of freedom on the top of the wall, since the loading protocol controlling the
horizontal actuator is operated in displacement controlled mode. There are three approaches to control
the two remaining degrees of freedom, axial displacement and top rotation, corresponding to:
- Force control of the vertical actuator : the total normal force applied remains constant during
testing, while maintaining the shear spam h0 equal to half the wall height by imposing a zero
bending moment condition at mid-height of the wall. This approach was used in some specimens
by Magenes et al. (2008), Petry and Beyer (2015) and by Ferretti et al. (2019);
- Displacement control of the vertical actuator : the two actuators on top restrain the change in
the wall height and top rotation, as a result the normal force in the wall changes as soon as the
wall cracks. This approach was used by Magenes and Calvi (1992) and Frumento et al. (2009);
- Mixed control alternating displacement and force control : the two actuators on top impose an
equal vertical displacement at both wall ends resulting in zero rotation of the wall while keeping
the total normal force constant. This approach was used in some of the tested specimens by
Magenes et al. (2008) and Petry and Beyer (2014, 2015).
Figure 2.16: Three approaches followed to simulate the double-fixed boundary conditions: a) Petry and Beyer
(2015), b) Magenes and Calvi (1992), c) Magenes et al. (2008)
37
The various test configurations might lead to differences in terms of strength and displacement ca-
pacity. However, the possible influence on deriving empirical drift capacity models is not yet accounted
for.
Regarding the different boundary configuration mentioned, a complete study was carried out by
Petry and Beyer (2014) to investigate their influence on the force-deformation response of URM walls.
In the experimental campaign designed by Petry and Beyer (2014) six quasi-static cyclic tests were
performed and the boundary conditions were simulating by varying the axial-load ratio σ0 /fc , where σ0
is the applied normal stress and fc the compressive strength of the masonry, and top-bottom moment
ratio applied to the wall. The fixed-fixed configuration was obtained by controlling the rotation on
the top beam to be zero, while the total normal force was kept constant. To do so, the axial force
and the shear force were introduced with three servo-hydraulic actuators which were controlled in a
fully coupled mode. The authors have also investigated the influence of the axial ratio on the drift
capacity. The results showed that the increment in the axial ratio σ0 /fc leads to a reduction of the
drift capacity. For larger axial ratios σ0 /fc , shear cracks tends to pass through bricks rather than
joints leading to a faster strength degradation. Finally, the findings were compared to the results
of a dataset comprising 64 walls tests to discussed the relationship between the axial stress and the
displacement capacity. The results showed that the drift capacity is strongly dependent on the height
of the test unit, and therefore there is a strong influence of the wall’s aspect ratio H/B on the drift
capacity of URM walls, which leads to larger drift capacities for walls of smaller heights.
There is a considerable amount of literature reviewing the in-plane behaviour of unreinforced clay
masonry bricks in terms of shear capacity, drift and energy dissipation. Beside the aforementioned,
it is worth mentioning Turnsek and Cacovic (1971), Magenes and Calvi (1992), Tomaževič and Weiss
(2012), Knox (2012), Petry and Beyer (2015). However, the available literature is more limited when
it comes to reviewing SCT on strengthened masonry. There is a moderate amount of literature
regarding the in-plane enhancement of Confined Masonry (CM) strengthened with TRM, among
which da Porto et al. (2011), Akhoundi et al. (2018), and Yacila et al. (2019) can be mentioned. On
the other hand, Bui et al. (2015) and Reboul et al. (2018) focused on quasi-static in-plane response
of hollow concrete block masonry walls, strengthened only by vertical Textile Reinforced Concrete
(TRC) sheets. The strengthening system comprised glass fibres embedded in a cement-based matrix
with fibre reinforcement, which led to a moderate increment of the lateral strength capacity but to a
significantly increase of the ductility capacity. The results showed that the dissipation mechanism is
characterised by micro-cracks in the TRC composite, which lead to an improvement of the dissipation
capacity of the walls.
Gattesco et al. (2015a), Godio et al. (2019) analysed the in-plane behaviour of strengthened stone
masonry walls with diverse strengthening techniques. Gattesco et al. (2015a) assessed the structural
efficiency of glass fibre reinforced mortar (GFRM) jacketing involving glass mesh embedded into
conventional mortar. On the other hand, Godio et al. (2019) tested asymmetrically strengthened
stone masonry walls to assess the drift limits corresponding to the damage of the artistic assets
38
that are connected to the structural elements. In order to obtain drift limits corresponding to the
onset of damage to the plaster, double-leaf stone masonry walls coated with traditional plaster were
tested under shear compression, as shown in Figure 2.17. In addition to the drift capacity, the
authors studied the influence of the axial ratio, defined as σ0 /fc where σ0 is the applied normal stress
and fc the compressive strength of the masonry, and the load history. Results showed that with
increasing axial ratio, initial stiffness and force capacity increases whereas ultimate drift decreases.
Load history has negligible influence on the stiffness and the force capacity while, on the contrary, it
has a significant influence on the ultimate drift which is larger when subjected to monotonic loading
than when subjected to cyclic loading.
Figure 2.17: Shear compression test set-up used to analyse the in-plane behaviour of strengthened stone
masonry walls: a) Gattesco et al. (2015a), b) Godio et al. (2019)
It can be seen that there is a dearth of literature with experimental programmes involving SCT on
clay brick masonry walls strengthened with TRM strengthening systems based on lime-based matrix.
An extensive analysis is lacking in terms of stiffness degradation drift capacity, energy dissipation
and damping. Such analysis would be of great importance since at present there is no building code
nor guideline that establishes the referential values when it comes to evaluate or design strengthening
solutions.
Papanicolaou et al. (2007) investigated the application of TRM as a way to increase the load
carrying and deformation capacity of URM walls subjected to cyclic in-plane loading. This study
was a pioneer one on the possibility of using TRM as an alternative to FRP. The authors carried out
SCT on twenty-two single-wythe perforated clay brick and cement mortar masonry walls and studied
the influence of matrix material, number of textile layers and compressive stress level applied. The
specimens were symmetrically strengthened with one or two layers of carbon fibre textile (CFRP)
bonded with a commercial polymer-modified cement mortar or a two-part epoxy adhesive. Other
39
specimens were strengthened with NSM CFRP strips per side, placed along slots formed in every
third bed joint. The results exhibited that increasing the number of layers lead to an increment
on the strength but only a slight reduction on the deformation capacity. Regarding the comparison
between mortar embedded and resin impregnated, the former resulted in lower effectiveness in terms
of strength but higher effectiveness in terms of deformability, which is of higher importance in seismic
design. Finally, it is concluded that the energy dissipation capacity of the TRM-based strengthening
configuration is comparable to the FRP-based one. The authors continued their research switching the
epoxy-based mortar matrix to inorganic lime-based mortar matrix. Papanicolaou et al. (2011) carried
out an investigation on five single-wythe, fire clay brick and stone blocks masonry walls strengthened
with one or two layers of carbon or basalt fibre TRM embedded in two types of mortar, namely fibre
reinforced cement-based mortar matrix and low strength lime-based mortar matrix. The specimens
were subjected to in-plane loading with axial force to simulate the response of shear walls under lateral
loading, as shown in Figure 2.18. The main parameter studied was the number of reinforcement layers
used on the jacketing. In addition, different compressive stress levels were applied to study the
influence on the seismic response. Among the results it is interesting to note that specimens receiving
double-layer overlays exhibited a stiffer behaviour than the single-layer and after the formation of
flexural cracking at the bed joints the specimens failed due to toe crushing or jacket buckling without
evidencing fibre rupture. Therefore, textile did not contribute significantly to the lateral load-bearing
capacity when the failure mode was due to rocking. Finally, in terms of deformation capacity, which
is of crucial importance in seismic retrofitting of unreinforced masonry walls, TRM jacketing proved
to be more effective than FRP, increasing the effectiveness of the shear walls on about 15–30%. As
a general rule, the strength increased with the number of layers and the increase of axial load at the
expense of deformation capacity.
Figure 2.18: Test set-up for in-plane loading shear walls in Papanicolaou et al. (2011)
Hračov et al. (2016) described the employment of polypropylene geo-nets as a strengthening tech-
nique for both adobe bricks and unfired classical brick masonry walls built in Flemish bond. The
main interest of the experimental campaign was related to the improvement of the in-plane behaviour
of the walls by the application of the strengthening method. The walls were strengthened on both
40
sides with the geo-net and anchored by means of steel staples. In the adobe walls the geo-nets were
covered with an adobe plaster after the mechanical fastening of the geo-net. In the brick masonry walls
the geo-net was exposed in order to determine the influence of the plaster on the overall mechanical
behaviour. The geo-net was also applied to a previously damaged unreinforced adobe wall which was
tested for the purpose of repairing retrofit. All specimens were tested by SCT with a sequence of
cyclical horizontal loading.The results showed that the geo-net ensured the integrity and the lateral
resistance of the partly damaged earth brick wall. On unfired clay brick walls the strengthening using
geo-nets led to an increase of the effective ultimate horizontal force and a moderate increase of the
effective stiffness. Both types of masonry, adobe and clay brick, evidenced almost the same ultimate
horizontal load, meaning that geo-nets are the main factor determining the limits of the horizontal
force, neutralising the differences between the different walling materials and therefore the additional
plaster had a minor influence.
As exposed, there is a dearth of experimental results regarding SCT performed on double-leaf clay
brick masonry, but more importantly there is a dearth of experimental results regarding SCT performed
on strengthened masonry with TRM strengthening systems. Data on the hysteretic behaviour of the
strengthened walls, such as strength and stiffness degradation and energy dissipation capacity, can
be obtained only by experimental simulation of seismic behaviour of strengthened masonry walls to
better understand their response and improve their design.
2.5 Experimental design assessment of specimens retrofitted with

TRM
In the absence of experimental evidence, one of the main aims of performing SCT on strengthened
masonry walls is the determination of characteristic drift limit states to identify the different levels
of damage. Building codes such as EC8 (2004) and NTC (2018) define two drift limit states, the
significant damage state and the collapse state, which are estimated as a function of the boundary
conditions, the failure mode and the aspect ratio of the walls. These two drift limit states have to be
verified in the assessment and design of masonry structures located in earthquake prone areas.
The available guidelines, namely ACI 434 (2017), ACI 549 (2013) and CNR-DT 215 (2018), provide
criteria for the evaluation of strengthening solutions for masonry structural elements. These guidelines
provide requirements for the determination of TRM capacity, and will be discussed further on the
present work; however, the guidelines failed to provide the limit states that needs to be verified in
order to meet the requirements of serviceability and ultimate state of the strengthening solution in
earthquake prone areas. The limit states should be defined so the designed strengthening solutions
would withstand the inelastic deformation induced by cyclic horizontal actions, such as seismic events,
and dissipate energy to prevent the total collapse.
Vanin et al. (2017) have highlighted the importance of the accurate estimation of the drift capacity
of URM walls as an essential input parameter in the deformation-based seismic assessment methods
41
for URM structural elements. The deformation capacity of shear masonry walls is of paramount
importance when assessing the ultimate limit state of masonry buildings. Consequently, Vanin et al.
(2017) collected the results of 123 in-situ and laboratory SCT on stone masonry walls and evaluated
their stiffness, strength and drift capacities. The analysis of the results showed that monotonic tests
lead to significantly larger drift capacities than cyclic test, as also evidenced for clay block masonry
walls in Petry and Beyer (2014). Monotonic tests were disregarded since they are not representative
of the real behaviour of the structure under seismic action.
Vanin et al. (2017) proposed two simple drift capacity models for stone masonry suitable for im-
plementation in engineering practice. From a total of sixty-seven tests, five drift limit states were
evaluated. Such drifts were: drift at cracking δcr , at yield δy , at maximum force δHmax , at ultimate
limit state δu and at collapse δc . With the exception of the drift at the onset of cracking δcr , the drift
capacities were determined from the force–displacement envelope and its bilinear approximation as
suggested by Tomaževič (1999). Among the two models proposed, the first one relates the drift to
failure modes and masonry typology while the second one relates the drift to axial load ratio, slender-
ness ratio and masonry typology. Nevertheless, for both models the drift capacities were determined
from the ultimate limit state δu , defined from Equation 2.9. The equation was proposed by different
stone masonry typologies and the value X on the following Equation has to be replaced by the values
0.3, 2.25 6 or 0.45% depending on the masonry typology under analysis. The drift at onset of cracking
δcr is set equal to 0.20% as a general rule whereas the remaining drift limits are expressed as fraction
of δu .
σ0 H0
δu = max{1.5% − 4% ; X%} (2.9)
fc min{L; H}
δy = 0.25δu (2.10)
δmax = 0.70δu (2.11)
δc = 1.15δu (2.12)
As aforementioned, Godio et al. (2019) performed SCT on single-sided strengthened stone masonry
walls and concluded that the experimental drift values were underestimated by the above expressions
when it comes to strengthened specimens.
Similarly to the work done by Vanin et al. (2017), Morandi et al. (2018) assembled a dataset
collecting the results of in-plane cyclic tests on unreinforced masonry walls. The dataset was designed
to be a useful tool to improve the insight of the main seismic parameters that govern the lateral response
of brick masonry walls, since it may be of reference for revision of the current code recommendations.
The dataset gathered the results of 188 in-plane cyclic tests of unreinforced masonry wall built with
different construction masonry technologies with bricks, namely clay, calcium and silicate. Among
others, one of the aims of the dataset was to shed light on brick URM displacement capacity. The
42
dataset gathered the drift limits as imposed by the Italian standard NTC (2018) and the EC8-Part 3
(2004) and were classified according to their failure modes, shear and flexure, and corresponding to the
denomination of Damage Limitation (DL), Severe Damage (SD) and Near Collapse (NC) limit state.
The values of drift at peak force of the tests were assumed to be the DL limit state. The SD limit
state θu , commonly related to the life preservation, differ significantly as a function of the different
experimental failure modes and masonry material. In the case of specimens with shear failure the drifts
range from 0.15% to 0.60%. Conversely, walls characterized by flexural/rocking mechanisms provided
much higher values of drift θu , on average, overall larger than 1.10%. Specimens with hybrid modes
obtained intermediate values of drift θu between pure shear and hybrid failure modes. The overall
mean value of drift was settled around 0.70% and it was almost equal for all the materials. These
results evidenced that the drift limits at SD limit state proposed in the previous NTC (2018) and in
the EC8 (2004) seem to be adequate for flexural modes but in general overestimates the displacement
capacity in the case of shear failure. The drift values obtained at the end of the tests were related
to NC limit state and also were influenced by the failure mode and masonry typology. In the case
of pure shear mechanism the overall mean value of θmax was 0.41%. Conversely, for flexural failure
mode the mean value was 1.35%. Finally, the overall mean of θmax for hybrid modes was 0.80%. The
analysis of the results collected in the dataset allowed to conclude that for brick masonry walls, the
drift capacity is mainly influenced by the failure mode experienced by the specimens, which in turn is
influenced by the axial ratio σ0 /fc .
The literature review provided on the previous section on DCT and SCT on strengthened walls
has showed that the increment of ductility provided by TRM reinforcement, ranges between 70% to
300%, depending on the combination of materials used in the TRM strengthening systems. Hence, the
empirical functions provided by Vanin et al. (2017) and Morandi et al. (2018) failed to meet the real
limit states that the reinforcements can provide to the URM walls in terms of inelastic deformations.
There is the need to better understand the drift capacities of strengthened masonry walls to improve
their design in terms of displacement demand. In particular, there is a lack of research providing
insight on the drift limit state of clay brick masonry retrofitted with TRM strengthening systems.
2.6 Conclusions
An extensive literature review was presented to establish the current state of the art relating to
the in-plane performance of strengthened masonry walls with different TRM strengthening systems.
In addition, a review was also presented on mechanical characterisation tests on tensile and bond
strength performed on different TRM typologies, namely basalt, carbon or glass textiles embedded in
lime-based mortar or cement-based mortar matrices and SRG. The results showed the failure modes
and the level of compatibility with different masonry substrates. Moreover, the review on the bond
behaviour between TRM and different masonry substrate highlights the influence that the substrate
has on the failure mode experienced by the TRM and in turn, the contribution that TRM has on the
43
lateral capacity of the masonry typology being reinforced.

In-plane enhancement of brick masonry walls with TRM strengthening systems has been also
comprehensively studied through DCT. Research has shown that masonry walls strengthened with
double sided configuration TRM experienced an increment of their in-plane strength in a range between
40% to 300%, depending on the masonry typology and the material used for the TRM strengthening
systems. In addition, the strengthened specimens showed an important deformation capacity when
compared to URM specimens without significantly increase of their stiffness. The latter is one on the
main advantages of using TRM as strengthening solution since they do not add significant mass to
the structure. However, among the different masonry typologies that have been studied, research on
strengthened tuff masonry walls has outnumbered the research on strengthened clay brick masonry
walls. Research on strengthened brick masonry walls has been mostly focused on single-leaf clay
masonry walls built on running bond pattern. Investigation on experimental results of DCT on
retrofitted masonry typologies expose a dearth of research on the in-plane response of one of the most
frequent material combinations in historical masonry, namely solid clay brick and lime mortar joints.
The literature review has also evidenced that most of the research is mainly focused on the undam-
aged state of the wall panels, considering only two type of specimens: URM, used as control specimens
to determine the level of enhancement provided by the strengthening technique, and the strengthened
ones. Little attention has been given to the application of TRM on damaged masonry walls to deter-
mine TRM efficiency as a post-earthquake repairing technique. Consequently, there is need for deeper
knowledge on the mechanical characterisation of in-plane response of damaged masonry repaired with
TRM.
In-plane enhancement of brick masonry walls with TRM has been barely studied through SCT.
Even though SCT can provide trustworthy results on the in-plane behaviour of retrofitted masonry
walls, the complexity of its set-up hinders its profuse use. However, the analysis of SCT results, can
provide a comprehensive characterisation in terms of displacement capacity, damping coefficient and
energy dissipation.
The experimental results from both testing methodologies show that the application of TRM
strengthening systems encourage a more uniform distribution of both stresses and cracks, providing
higher shear strength and inelastic deformation capacity. In addition, the results evidenced the shift of
failure mode from brittle and sudden stair-stepped to a diffuse diagonal cracking along the compressed
strut. This last failure mode allows the fully exploitation of the tensile strength of the textile, showing
the potentiality of TRM as a retrofitting technique.
Current TRM guidelines, assessing the effectiveness of retrofitting solutions, are fed from experi-
mental results. These results are mainly focused on the evaluation of the load bearing capacity and
little attention has been given to the displacement capacity provided by the strengthening solution.
The analytical formulations provided by the guidelines have been tested with rather good agreement;
however, further research is needed to provide more insight on the mechanical behaviour of retrofitted
masonry. Finally, drift capacity models experimentally obtained for URM walls can be used as lower
44
bound limits for the determination of drift limits states of retrofitted walls. Nevertheless, additional
experimental research using SCT on TRM strengthened masonry walls must be done to better calibrate
their displacement capacity.
Using the specific knowledge gaps identified in this review on the background and the current
understanding of the failure modes and behaviour of strengthened masonry walls, an integrated ex-
perimental programme was designed and conducted in this PhD thesis, with the aim of adding more
results to the experimental knowledge. Compared with the number of experimental programmes ded-
icated to the investigation of retrofitted specimen with DCT and SCT, little attention has been given
to the relationship between both testing methodologies. Finally, the analysis of experimental results
by means of a holist approach, involving both DCT and SCT testing methodologies, should be con-
ducted in order to define the mechanical parameters that are key to the reliable analysis of the in-plane
behaviour of strengthened masonry, namely drift limit capacity, stiffness and strength degradation,
damping coefficient and energy dissipation.
45
Chapter 3
Experimental Programme
3.1 Introduction
One of the main aims of the experimental programme is to evaluate the in-plane structural behaviour
of strengthened masonry walls. Parameters such as tensile strength of masonry, shear capacity of
masonry and ductility capacity, among others, are of utmost importance to better understand the
behaviour of unreinforced and strengthened masonry when subjected to a cyclic excitation as could
be a seismic event.
There are several experimental test methods that allow the study of the aforementioned parameters
in masonry assemblages. On one hand there is the diagonal compression test (DCT) which covers the
determination of the diagonal tensile strength, and is standardised by ASTM (2000). On the other
hand, there is the reversed cyclic shear compression test (SCT) to determine the force-displacement
properties and hysteretic data of the structure, i.e ductility capacity, damping and energy dissipation.
However, both tests present shortcomings due to some complexities. The former has two possible
interpretations (Brignola et al. (2008), Calderini et al. (2010)) and the latter presents inconveniences
from the technical point of view since the desired boundary conditions can not always be achieved.
This is in fact due to the need of perfect constraints that hinder the wall from translating and rotating
(Tomaževič (2008)).
Despite this fact, the experimental programme involves the performance of both test methods
on one of the most frequent material combinations in historical masonry, i.e. solid clay bricks and
lime mortar joints in two different configurations, unreinforced and strengthened. For this purpose,
handmade solid clay bricks, fired with traditional procedures, and a low mechanical performance lime
mortar were used. For the retrofitting four different reinforcement systems including Textile Reinforced
Mortar (TRM) and Nearly surfaced mounted (NSM) repointing were applied.
This chapter presents the mechanical characterisation of the component materials of the specimens
and its reinforcement, the preparation of the specimen, its repairing and strengthening and the testing
procedure.
47
CHAPTER 3. EXPERIMENTAL PROGRAMME
3.2 Materials
3.2.1 Brick characterisation
The units were handmade solid clay bricks, provided by the company TerraCuita. The bricks pre-
sented rough and irregular surfaces, and slightly variable dimensions due to their traditional way of
manufacturing. The average dimensions were 310 × 145 × 45 mm3 .
Twenty prismatic brick samples with dimension 100 × 100 × 40 mm3 were cut from the units to
evaluate the compressive strength. The samples were tested in compression according to EN 772-
1:2011 CEN (2011) by using a load cell of 3000 kN under load control with a loading rate of 0.15
MPa/s. The compressive strength was corrected by a shape factor of 0.70 in order to obtain the
normalized compressive strength of the brick fb,c CEN (2011). The flexural strength of the brick fb,f
was determined by three-point bending test on 10 units following the EN 772-6:2001 CEN (2002) and
the EN 1015-11:1999 CEN (1999).
The EN 12390-13:2014 CEN (2014) was used as reference for the determination of the static
Young’s modulus of the bricks across the length Eb,st . Twenty-one prisms with dimensions 40 × 40 ×
80 mm3 were cut from seven units. Three loading-unloading compressive cycles, from 10 % to 30 % of
the estimated maximum load, were applied to the specimens. Eb,st was evaluated as the slope of the
third loading cycle. In addition, the expression proposed by Makoond et al. (2020) was used to verify
the value of Eb,st using the dynamic Young modulus of the brick Eb,dy obtained with the procedure
proposed in Makoond et al. (2019). Table 3.1 summarizes the experimental results that were obtained
for each test in terms of average values and coefficient of variation (C.o.V).
Table 3.1: Mechanical properties of bricks used for he construction of walls
Bricks fb,c [MPa] fb,f [MPa] Eb,st [MPa] Eb,dy [MPa]

Average 17.99 2.44 6287 7882
Number of
20 10 21 7
samples
C.o.V 8.3% 20.0% 27.0% 15.0%
3.2.2 Mortar characterisation
The mortar used to bind the units was based on a commercial premixed hydraulic lime mortar Kerakoll
(2016) classified as M5 according to EN 998-2:2010 CEN (2010). Limestone filler was added to the
premixed mortar to reduce its compressive strength in order to replicate a lower strength historical
material as porposed by Segura et al. (2020). Following the EN 1015-11:1999, prismatic samples with
dimensions 160 × 40 × 40 mm3 were prepared during the construction of each wall, to evaluate the
strength of the mortar as shown in Figure 3.1. For each wall built, nine prismatic samples were cast
to evaluate the flexural strength fm,f . The compressive strength fm,c was assessed on the eighteen
48
halves produced by the splitting of the samples under flexure. The mortar samples were tested using
a load cell of 10 kN under load control.
Figure 3.1: Elaboration of standardised mortar samples: a) pouring of the material inside the moulds and
compact by stokes of a tamper, b) storing of the moulds after skimming off the excess of material, c) final look
of the sample
Figure 3.2: Test carried out on binding mortar samples: a) three point bending test, b) compressive test, c)
double punch test
After the test of each wall, under DCT, the remaining masonry was disassembled with the aim of
extracting mortar samples from the bed joints. Mortar samples with dimensions 50 × 50 × 15 mm3
were cut from the joints extracted and subjected to the Double Punch Test (DPT) according to DIN
18555-9:1999 Deutsche Norm (1999) for the determination of the compressive strength fm,DP T . The
samples were tested between 20 mm diameter loading plates by using a 10 kN capacity load cell. The
irregular surface of the mortar was regularized using gypsum powder in order to assure a homogeneous
loading of the sample, following the indication of Pelà et al. (2017, 2018). Figure 3.2 shows the layout
of the three tests.
Table 3.2 summarizes the experimental results of each test for the binding mortar samples associ-
49
ated to the specimens tested in diagonal compression. The outcome is given in terms of average values
and coefficient of variations (C.o.V).
It must be noted the high scattering between fm,c obtained with the standardised test and fm,DP T
from the double punch test. This difference arises from the heterogeneity of the mortar but mostly
from the fact that the DPT test is executed on a small portion of mortar layer with small thickness
when compared to the standardised dimension of the prismatic sample. Thus, the DPT can hardly be
considered a simple uniaxial compression test but more representative of a triaxial stress state, what
is in line with the finding of Marastoni et al. (2016).
Table 3.3 shows the experimental results of each test for the binding mortar samples associated
to the specimens tested in shear compression. The outcome is given in terms of average values and
coefficient of variations. In this case, no DPT tests were performed due to the inaccessibility of the
mortar sample as a consequence of the set-up configuration.
Similarly, Table 3.4 shows the static and dynamic Young’s moduli of the binding mortar Em,st and
Em,dy following the procedures proposed by Makoond et al. (2019, 2020).
Table 3.2: Mechanical properties of binding mortar of each wall specimen tested under Diagonal Compression
Test(DCT)
Specimen fm,c [MPa] C.o.V [%] fm,f [MPa] C.o.V [%] fm,DP T [MPa] C.o.V [%]
URM 1 3.59 5.7 0.96 18.77 4.90 18.8
URM 2 1.98 11.54 0.54 14.11 5.36 24.58
LDB 1 2.48 20.30 0.61 20.35 6.46 27.48
LDB 2 3.10 15.66 0.79 18.94 5.50 25.36
LDS 1 2.17 15.80 0.76 22.07 3.99 26.79
LDS 2 3.28 31.70 0.78 31.31 4.23 23.80
MDS 1 2.37 12.66 0.56 30.11 4.40 22.24
MDS 2 2.22 15.96 0.61 20.43 3.68 27.35
LDB-JR 1 2.10 18.26 0.46 39.65 4.28 28.84
LDB-JR 2 1.77 21.89 0.51 22.33 5.55 27.68
Average [MPa] 2.51 - 0.66 - 4.69 -
C.o.V [%] 24.3 - 23.9 - 31.62 -
Number of Samples 176 - 88 - 496 -
3.2.3 TRM characterisation
Textile
Four different products were used for the strengthening and repair of the masonry walls. Table 3.5
reports their classification and relevant properties as provided by the manufacturer, Kerakoll:
GeoGrid G200 ® (LDB): bidirectional grid of low density basalt fibres with steel micro-cords.
Grid dimension 17 Ö 17 mm2 (see Figure 3.3-a).
GeoSteel G600 ® (LDS): low density steel fibre, consisting in unidirectional sheet made of
strength galvanized steel micro-cords(see Figure 3.3-c). Each fibre comprises 5 cords. Two
50
Table 3.3: Mechanical Properties of binding mortar of each wall specimen tested under Shear Compression
Test(SCT)
Specimen fm,c [MPa] C.o.V [%] fm,f [MPa] C.o.V [%]

URM 1 1.85 21.30 0.34 2.62
URM 2 2.16 20.28 0.62 16.92
LDB 1 3.10 13.60 0.66 28.67
LDB 2 2.28 12.76 0.51 19.70
LDS 1 3.47 20.11 0.77 29.13
LDS 2 2.43 16.07 0.63 17.06
LDS-DL 1 3.04 18.07 0.59 27.11
LDS-DL 2 2.31 13.14 0.48 13.04
Average [MPa] 2.58 - 0.57 -
C.o.V [%] 21.6 - 22.5 -
Number of Samples 134 - 60 -
Table 3.4: Average static (Em,st ) and dynamic (Em,dy ) modulus of the binding mortar
Specimen Em,st [MPa] C.o.V [%] Em,dy [MPa] C.o.V [%]

Binding Mortar 2696 8.0 3987 10.0
cords are twisted around three straight cords to ensure an effective interlocking. The fibres are
fixed to a fiberglass mesh.
GeoSteel G1200 ® (MDS): medium density steel fibre, unidirectional textile of strength galva-
nized steel cords. Each fibre comprises five cords. Two cords are twisted around three straight
cords to ensure an effective interlocking. The fibres are fixed to a fiberglass mesh (see Figure
3.3-d ).
Helibar 6 ® (JR): helical bar made of stainless steel with high mechanical performance for the
repointing of the joints (see Figure 3.3-b).
For the sake of simplicity each type of textile are not identified by their commercial names but by the
name appearing in brackets in the previous listing.
Table 3.5: Mechanical properties of the products used for the reinforcement of the walls as provided by the
manufacturer
Tensile Strength
Young’s modulus Tensile Strength Strain at failure Thickness
Product from Shear bond test
E [GPa] σu,f [MPa] εu,f tf [mm] σsl,t [MPa]
LDB 90 1700 0.019 0.032 1049
LDS 190 2800 0.015 0.084 2096
MDS 190 3000 0.02 0.169 687
JR 160 1250 0.055 - -
51
Figure 3.3: Types of textiles used for the wall’s strengthening: a) Bidirectional low density basalt textile
(LDB), b) Helical stainless steel rebar (JR), c) Unidirectional low density steel textile (LDS), d) Unidirectional
medium density steel textile (MDS)
Mortar Matrix
The mortar matrix used for the application of the textile fabric is Kerakoll GeoCalce F Antisismico ®.
The product is a premixed NHL 3.5 natural hydraulic lime of M15 class according to EN 998-2:2010.
Mortar samples were tested after the application of the reinforcement in each strengthened wall in order
to control its resistance. The samples were cast into metallic moulds after the preparation of every
batches of mortar. Prismatic samples with dimensions 160 × 40 × 40 mm3 were prepared to evaluate
the strength of the mortar in each TRM retrofitted wall, according to the EN 1015-11:1999. Flexural
strength frm,f , was evaluated on six prismatic specimens for each wall built, while the compressive
strength frm,c was assessed on the twelve halves produced by the splitting of the samples under flexure.
The mortar samples were tested by using a load cell of 10 kN capacity for the flexural test and a load
cell of 200 kN capacity for the compressive test. Both the tests were carried out in load control. The
loading rate was constant during the test with values of 20 N/s and 75 N/s respectively. These loading
rates were selected according to the EN 1015-11:1999 CEN (1999) in order to reach the failure within
30 to 90 seconds. The Young’s modulus of the mortar matrix, Erm,st is provided by the manufacturer
and equal to 9 GPa.
Table 3.6 and 3.7 summarize the experimental results of each test for the mortar matrix samples
associated to the specimens tested in diagonal compression (DCT) and shear compression (SCT)
respectively. The results are given in terms of average values and coefficient of variations (C.o.V).
52
Table 3.6: Mechanical Properties of the mortar used as reinforcement in each wall specimen tested under
Diagonal Compression Test (DCT)
Specimen frm,c [MPa] C.o.V [%] frm,f [MPa] C.o.V [%]

URM1 R 13.72 9.20 5.36 8.78
URM2 R 13.87 10.8 5.19 8.59
LDB 1 13.55 14.87 3.81 9.88
LDB 2 17.57 12.36 5.17 16.85
LDS 1 13.93 14.06 4.17 13.83
LDS 2 13.44 8.02 3.59 11.54
MDS 1 12.78 9.05 3.72 10.23
MDS 2 13.46 4.92 3.69 6.11
LDB-JR 1 14.59 6.67 4.41 9.05
LDB-JR 2 14.13 13.41 3.81 10.06
Average [MPa] 14.10 - 9.3 -
C.o.V [%] 4.29 - 16.3 -
Table 3.7: Mechanical Properties of the mortar used as reinforcement in each wall specimen tested under
Shear Compression Test (SCT)
Specimen frm,c [MPa] C.o.V [%] frm,f [MPa] C.o.V [%]

URM1 R 11.53 18.46 3.27 15.53
URM2 R 12.94 16.17 3.18 17.24
LDB 1 14.59 16.29 3.35 11.96
LDB 2 12.94 7.60 3.92 8.70
LDS 1 14.24 5.78 3.81 5.73
LDS 2 12.12 2.80 3.39 6.81
LDS-DL 1 12.71 4.11 3.78 6.54
LDS-DL 2 12.52 13.93 3.37 17.71
Average [MPa] 12.95 - 3.51 -
C.o.V [%] 7.9 - 8.1 -
3.2.4 Bond behaviour of TRM
The adhesion between the textile and the mortar matrix plays a key role in the mechanical behaviour
of the reinforcement systems. A good mortar-textile interlocking leads to a good exploitation of
the tensile strength of the textile, precluding a failure mode involving the reinforcement-substrate
interface. However, to assure that such failure mode does not occur, there is another parameter that
has to be taken into consideration and it is the effective bonded length Lef f of the composite on
the substrate. Lef f represents the minimum length needed to fully establish the shear stress transfer
from the substrate to the TRM strengthening system so the textile can carry all the tensile stress
(Santandrea et al. (2020)).
53
This section collects the results of the single lap shear bond tests performed on TRM with basalt
textile (BTRM) and steel reinforced grout (SRG) applied on clay brick masonry substrate in order
to identify the Lef f and determine the tensile strength, σsl,t , of the different TRM under study in
the current research. The dataset contains a total of 154 test results corresponding to 11 different
experimental campaigns (Ascione et al. (2015), Santandrea et al. (2016, 2017, 2020), De Santis and
de Felice (2015), De Santis et al. (2017b), Lignola et al. (2017), Ferrara et al. (2019), Bellini et al.
(2019), Barducci et al. (2020), Wang et al. (2020)) analysed in the literature review. From the total, 123
tests were performed on SRG of low density, 21 tests were performed on BTRM and the remaining 10
on SRG of medium density. The dataset contains information on bonded length, width and thickness
of the TRM strip, global slip, failure mode (Figure 2.7) and peak load attained by each specimens.
The bonded length ranges from 50 mm to 345 mm in the case of SRG of low density, whereas for
BTRM only three lengths were studied 220, 260 and 345 mm.
In order to establish a correlation between the bonded length and the failure modes evidenced on
SRG composites when subjected to shear bond test, each type of failure mode was plotted against a
bonded length range, as shown in Figure 3.4. The less desired failure modes, the ones involving the
matrix-substrate interface (A and B), were evidenced in correspondence with low values of bonded
length. Conversely, the other four failure modes (C to F) were associated with higher ranges of bonded
length. Even though there was a certain scattering on failure mode C regarding the bonded length
associated to it, it was the most frequent failure mode observed for this type of TRM composite.
Therefore, the matrix-textile interlocking governs the bond behaviour of SRG and the difference in
stiffness between both materials led to a premature failure of the TRM systems.
Figure 3.4: Failure modes evidenced in the single-lap shear bond test on SRG composite for different bonded
lengths
54
In addition, a correlation between the tensile strength σsl,t and the bonded length was studied.
Figure 3.5 displays the 123 test results on SRG, including all types of failure mode and all bonded
length tested. The red-dotted line in Figure 3.5 demonstrates that, there is an evident increment of
capacity as the bonded length increases. However, once the bonded length reaches Lef f equal to 250
mm the tensile strength σsl,t exhibits almost a constant trend. Therefore, it can be said that the
tensile strength attain by the TRM composite is kept constant independently from the increment of
bonded length once the effective bonded length is granted.
Figure 3.5: Identification of SGR effective bonded length. Evolution of the tensile capacity of SGR composite
with increasing bonded length. Test results corresponding to the experimental campaigns performed by Ascione
et al. (2015), Santandrea et al. (2016, 2017, 2020), De Santis and de Felice (2015), De Santis et al. (2017b),
Wang et al. (2020)
The tensile strength of the reinforcement σsl,t was determined from the shear bond test. The
tensile stress attained by each specimen was derived as the peak load divided by the cross section area
of the dry textile. As described before, such tensile strength was in close correlation with the bonded
length and the failure mode. Therefore, to evaluate the σsl,t of SGR composite, the shear bond test
results were filtered considering only those specimens with bonded length equal and greater than 250
mm and with failure modes type C. Consequently, 48 out of the 123 specimens tested of SGR with
low density textile were considered for the computation of the tensile strength. The same procedure
was followed to compute the tensile strength of SRG with medium density, where only 5 specimens
were accounted for. The average values for each TRM are shown in Table 3.5.
55
It is of interest to highlight the influence of the textile density on the final capacity achieved by
each textile. The sparser textile (LDS) has higher σsl,t , which means that a better mortar-textile
interlocking allows a good stress transfer before textile debonding, maximizing the exploitation ratio
(74%) of tensile strength of the steel textile, while the denser textile (MDS) has poorer interlocking
since the mortar is not able to protrude through the voids of the textile and, consequently, achieves
less tensile capacity due to premature debonding at mortar-matrix interface. As a result, there is an
evident reduction of the load-carrying capacity of the SRG matrix-fibre interface as the fibre density
increases.
The test results of the BTRM specimens, evidenced less scattering in terms of failure mode. 20
out of 21 specimen tested failed due to rupture of the basalt fibres independently from the bonded
length studied. Consequently, it could be deemed that the effective bonded length for this type of
TRM composite is greater than 200 mm. Only one specimen underwent textile slippage within the
mortar matrix (D) as a consequence of an uneven distribution of stresses across the textile section.
The rupture of the fibre asserts basalt as a more deformable textile, when compared with SRG,
making basalt much more compatible with clay brick substrate avoiding any premature debonding
and promoting a high exploitation of tensile strength of the textile.
3.3 Construction of specimens
Eighteen double leaf masonry walls with nominal dimensions 1270 × 1270 × 310 mm3 were built in
the laboratory by professional masons. Of these, ten were tested under diagonal compression (DCT)
and the remaining eight under shear compression (SCT).
The construction process of the specimens considered the procedure shown in Figure 3.6 and
described in the following:
The bricks were wet by immersing them in a bucket of water for one minute, in order to avoid
the absorption of the water of the mortar during its construction, as shown in Figure 3.6-a.
The mortar was mixed according to the following proportions for each batch: 9.35 l of water,
®
16.2 kg of filler and 25 kg of Biocalce MuroSano , see Figure 3.6-b.
All specimens were set on a metallic C-profile. For DCT specimens, a sliding interface between
the base of the masonry wall and the surface of the metallic profile was generated by inserting a
3 mm thick Teflon sheet and a 3 mm thick PVC sheet. Sliding interface was necessary in order
to allow the horizontal sliding of the base of the wall during the testing. The Teflon and PVC
sheets were clamped to the metallic profile. While the SCT specimens were laid on the metallic
C-profile filled with concrete. This profile allowed the sliding shear failure to potentially occur
in the bed joints of the lower course of bricks.
Four metallic tube-shape rulers were placed at the corners of the steel profiles prior to con-
struction of the wall. Each ruler was marked to define the height of each course of the wall,
considering the height of both the unit (45 mm) and the mortar joint (around 15 mm), as shown
56
in Figure 3.6-c. Cords were used to ensure the horizontality of the courses and their alignments
to the marks on the rulers.
All specimens were built in Flemish bond with 21 courses and 15 mm mortar joints. Each wall
specimen required the preparation of 6 batches of mortar and the use of 170 bricks. At the end
of the construction of each row, its level was measured to ensure the perfect horizontality and
verticality of the specimens, as shown in Figure 3.6 d-e.
The finished walls were stored under laboratory conditions during the 28 required for the curing
of mortar, see Figure 3.6-f and Figure 3.6-i.
The concrete beam was cast on top of the SCT specimens as shown in Figure 3.6 g-h.
The two URM walls corresponding to DCT were built with a defect in the central brick of the 11th
course, intentionally created to induce a regular diagonal crack pattern. The brick in the centre of the
panel was cut in the middle in order to create a 4 mm thick notch, not filled with binding mortar, as
shown in Figure 3.7.
Finally, to assess the compressive behaviour of masonry, seven stack bond prisms of five bricks
and four running bond walls were tested in compression following EN 1052-1 European Committee
for standardization (CEN) (1999). An average compressive strength of 6.50 MPa (C.O.V 9%) was
obtained on the experimental campaign carried out in Segura et al. (2018). It must be noted that even
though the constituents of masonry presented high scattering on their properties characterisation, is
not reflected on the compression strength of masonry as a whole. This can be attributed to the failure
mode observed in the test, which mainly involved the crushing of masonry. Thus, the heterogeneity of
the materials that comprise the masonry will have a diverse impact on it behaviour depending upon
the failure mode undergo by the specimens when subjected to different test configurations.
3.4 Repair and retrofitting of specimens
3.4.1 Strengthening
The wall specimens were strengthened twenty-eight days after their construction by professional work-
ers from the manufacturer company. From the eighteen specimens built, four were left unreinforced.
The remaining fourteen were retrofitting with different TRM systems and NSM configurations. The
four unreinforced specimens were tested for the two set-ups (DCT and SCT), and as a result four
damaged specimens were later repaired and retrofitted.
The materials for strengthening and repair of the masonry substrate included three different textiles
embedded in mortar matrix, and a NSM helical stainless steel rebars for joint repointing (JR). As
aforementioned (see Section 3.2.3), the first type of textile consisted in a bidirectional grid of low
density basalt (LDB) fibres with steel micro-cords and 17 × 17 mm2 grid spacing. The second and
third type of textiles consisted in unidirectional sheets of galvanized steel micro-cords. Each fibre
comprises five cords, two of which are twisted around three straight cords to ensure an effective
57
Figure 3.6: Construction procedure: a) brick wet by immersion, b) mixture of the binding mortar, c) placing
of the rulers for the construction, d-e) construction of the wall in Flemish bond, f) specimen finished for diagonal
compression test, g) rebars placing for the concrete beam located on top of the specimen for Shear compression
testing, h) cast of the concrete beam, i) specimen finished for shear compression test
58
Figure 3.7: URM specimen built with a defect in the central brick to induce a regular diagonal crack when
subjected to diagonal compression
interlocking. The difference between the two textiles lies in the steel density, defined as the number
of steel yarns per unit width, which is either 1.57 yarn/cm in the case of low density steel (LDS) or
3.14 yarn/cm for the medium density steel (MDS).
Figure 3.8 shows the procedure followed for strengthening the wall specimens. The necessary steps
are listed in the following:
The surfaces of the walls were prepared by removing the dust with a vacuum and by creating
grooves along the mortar joints in order to generate the necessary grip between the wall’s surface
and the mortar matrix of the TRM, see Figure 3.8-a.
The specimens were wet with abundant water to prevent masonry from absorbing the water
during the application of the composite. The first layer of mortar matrix was applied on the
surface of the specimen, shown in Figure 3.8 b,e. The mortar matrix was mixed with a proportion
of 4.6 litres of water every 25 kg of GeoCalce F Antisismico ®.
Then the textile was embedded in the matrix by applying a light pressure on the textile to
guarantee the right adherence to the support and to fill all the voids of the mesh, see Figure 3.8
c,e. The sheets of LDB grids had a width of 800 mm. Two sheets were applied on each side of
the wall with an overlap of 300 mm in the centre of the panel. The overlapping length was based
on the previous analysis performed on the bonded length results obtained from the literature
review and was designed to assure a satisfactory stress transfer between the substrate and the
TRM. The strips of LDS and MDS had a width of 100 mm.
59
A second layer of mortar matrix was applied to cover completely the fibres, as shown in Figure
3.8 d,f. The final thickness of the TRM reinforcement varied between 8 to 10 mm. The procedure
was repeated at both faces of the specimen.
Once the hardening of the mortar had begun, the faces were wet to favour the curing and then
were wrapped with sackcloth fabric, which was kept wet for the following 7 days, see Figure
3.8-h. Once the fabric was wet, it was wrapped with plastic sheets to preserve the humidity of
the specimen.
In the case of joint repointing with NSM helical stainless steel rebars, the application procedure
was similar but the grooves were 30 to 40 mm deep and with a vertical spacing of three courses,
as shown in Figure 3.8-g. The grooves were created, after the curing of the mortar, with a rotary
hammerdrill accessorized with a 20 mm width flat chisel. The curing time of all the walls was
28 days under laboratory conditions.
Figure 3.8: Procedure for the application of the TRM systems: a) creation of grooves along the mortar joints,
b) application of the first layer of mortar, c) set of the fibre net, d) finished look of the wall retrofitted with
basalt TRM, e) application of the first layer of vertical mortar strips, f) finished look of the wall retrofitted with
steel TRM, g) insertion of the helical stainless steel rebar, h) wrapping of the specimens with wet sackcloth
fabric for curing
60
The outcome of the specimens retrofitted with MDS subjected to DCT had been deemed not to
be completely satisfactory due to its geometric features. Consequently for the second test (SCT)
this reinforcement was not applied. Instead, it has been studied the influence of the application of
subsequent layer of LDS with the same 100 mm width strip configuration (LDS-DL). The procedure
followed was the same as the one described before. However, two layers of vertical and horizontal
100 mm width strips were interspersed as shown in Figure 3.9. Although the procedure was similar,
special attention was given to the intersection of the multiple layers of vertical and horizontal strips
to avoid a bulky finished. Thus, where the strips intersect only two layer of mortar were applied and
the four layers of strips were slightly pressed into the mortar matrix. At the same time the remaining
locations, where there were no intersections but just overlapping of layers, were filled with mortar
material to get an even surface. As in all the cases, the curing time of all the mortar was 28 days
under laboratory conditions.
The four URM specimens were tested twenty-eight days after their construction. After the DCT
and SCT tests, they were repaired and retrofitted with LDB, and then tested again twenty-eight days
after the repair.
Figure 3.9: Procedure of application of double layer strip configuration of LDS: a) application of the first
horizontal layer of LDS, b) application of the first vertical layer of LDS, c) application of mortar material in
the areas with no intersection, d-e) application of the second horizontal layer of LDS, f) application of mortar
material to get an even finished surface
61
3.4.2 Scuci-cuci
The two unreinforced specimens after testing under DCT were repaired using a repointing tech-
nique. The cracks developed during testing were filled with the same lime based mortar used for
the retrofitting of the wall, whereas the URM specimens tested under SCT were repaired using a com-
bination of the Scuci-cuci technique and lime mortar repointing, since the latter were more severely
damaged that the former ones.
The aim of the repair was to restore the wall’s structural continuity along the cracks and to recover
the initial stiffness of the wall, lost at the end of the test. The restoration was achieved by substituting
the fractured units with new ones and recovering the locally heavily damaged parts of the masonry.
The material used for the restoration was the same one of the construction in order to assure a good
compatibility in terms of stiffness and resistances, which is key to a good collaboration among new
and existing elements. Therefore, the cracks of the specimens were opened using a rotary hammerdrill
accessorized with a 20 mm width flat chisel and a hand tool 30 mm width flat chisel, Figure 3.10. The
fracture bricks were replaced by new ones as shown in Figure 3.11 and 3.12. Afterwards, the cracks
were filled with the same lime based mortar used for the retrofitting of the walls.
Figure 3.10: Scuci-cuci intervention on the URM tested walls. Cracks opening with a hand tool and filling
of openings
Figure 3.11: Scuci-cuci intervention on the URM tested walls. Removing and replacing of broken units, filling
of openings
62
Figure 3.12: Scuci-cuci intervention on the URM tested walls. Removing and replacing of crushed units.
Filling of cracks with lime based strengthening mortar.
3.5 Test summary

For the sake of simplicity the specimens have the same name for both types of test DCT and SCT.
The specimens were labelled with an alphanumeric identifier using the notation X #, where “X” is
the tag denoting unreinforced masonry (URM) or one of the reinforcement systems LDB, LDS, MDS,
LDB-JR, LDS-DL. The final digit “#” is a number (1 or 2) used to identify each specimen since they
were tested in pairs to check the repeatability of the results. The URM specimens that were later
repaired and retrofitted with LDB are denoted by the “R” tag. Table 3.8 shows the specimen ID
with a description of its main features and the first column specifies to which tests the specimen was
subjected to.
3.6 Diagonal Compression Test
3.6.1 Set-up
The standard ASTM E519M ASTM (2000) was used as reference for the execution of the DCT.
However, a different setup that those suggested by the aforementioned standard was designed to allow
the application of the diagonal compression load without requiring the 45 degree rotation of the walls.
This modification was necessary because the specimens could experience damage during the rotation
operation due to their low strength binding mortar.
The specimens were set on a metallic bench consisting of two parallel H-Shape beams anchored
to the strong floor of the laboratory. Each metallic profile, supporting the specimens, was bolted on
top of the bench in order to avoid its displacement during the execution of the test. Two steel wedges
were placed at two diagonally opposite corners of the specimen. Each wedge was welded to a robust
beam consisting of 2 C Channels placed back to back and stiffened with ribs. The beams at opposite
corners were connected with two dywidag bars. The gap between the steel wedges and the corners
63
Table 3.8: Type of test, ID and features of the specimens built and reinforced for the experimental programme
Type of sample and

Test Specimen ID View of Sample
reinforcement characteristics
DCT URM 1 Unreinforced masonry walls.

SCT URM 2
Damaged masonry walls repaired and

DCT URM1 R retrofitted with bidirectional basalt
SCT URM2 R grid 17 × 17 mm2 , double-sided.
DCT LDB 1 Masonry walls reinforced with bidirectional

SCT LDB 2 basalt grid 17 × 17 mm2 , double-sided.
Masonry walls reinforced with unidirectional

DCT LDS 1 100 mm wide strips of low density steel
SCT LDS 2 fibre, double-sided.
Masonry walls reinforced with unidirectional

DCT MDS 1 100 mm wide strips of medium density steel
MDS 2 fibre, double-sided.
Masonry walls reinforced with two plies of unidirectional

SCT LDS DL 1 100 mm wide strips of low density steel
LDS DL 2 fibre, double-sided.
Masonry walls asymmetrically reinforced with

LDB-JR 1 bidirectional basalt grid 17 × 17 mm2 on one side
DCT
LDB-JR 2 and joint repointing with helical stainless steel
rebar of 6 mm diameter on the other.
64
of the masonry specimens was filled with a layer of epoxy resin and a sheet of compressed wood to
smooth the loading surface. The load was applied by using two hydraulic actuators which provided
the diagonal force by pulling the dywidag bars, as shown in Figure 3.13.
Figure 3.13: Setup of the Diagonal Compression Test
The hydraulic actuators, with load capacity of 600 kN, were controlled through the oil pressure
of the central pump, which was measured with a pressure transducer. The tests were all performed
under displacement control. At the beginning of each test, three cycles were executed, in the range
from 10 kN to 50 kN, and then the load was monotonically increased until failure. The displacement
was applied at a constant rate of approximately 0.5 mm/min. The tests were stopped when the
reduction in strength with respect to the peak load was about 50%, in order to capture correctly the
post-peak softening behaviour. The specimens were instrumented with four linear variable differential
transducers (LVDT), having a displacement range of ± 5 mm and a precision of 5 µm, and two wire
sensors of 1000 mm and a precision of 0.01 mm, as shown in Figure 3.13. The LVDTs were mounted
along the diagonals on both sides of the specimen, in order to measure the shortening of the closing
diagonal (under compression) and the elongation of the opening diagonal (under tension). The wire
sensors were also mounted along the opening diagonal with a gage length of 900 mm.
3.6.2 Methodology
The methodology considered the computation of shear strength, stiffness and ductility capacity of
the DCT test results. The behaviour of the specimens is presented in terms of cracking patterns and
experimental curves P − δ (Load – displacement). The displacement is calculated from the average
readings of the LVDTs located on both sides of the wall. Table 4.1 summarizes the main parameters
obtained.
65
Shear strength and shear stiffness
Two standards, ASTM-E519 (2000) and RILEM TC 76-LUM (1994), provide criteria on how to
evaluate the main results of the DCT. The ASTM and the RILEM standards use Equation (3.1) to
evaluate the shear stress at the centre of the wall, being P the applied load and An the net area of
the panel:
P
τAST M,RILEM = 0.707 × (3.1)
An
Both standards (ASTM (2000) and RILEM TC 76-LUM (1994)) assume an isotropic linearly
elastic model with a pure shear stress state in the centre of the panel. Consequently, the Mohr’s
Circle is centred in the τ − σ plane and the value of the shear strength is computed as Equation (3.1).
However, there is also another interpretation based on the theory of Frocht (1931). According to
Frocht’s formulation, a non-uniform shear stress takes place along the loaded diagonal of the specimen
subjected to diagonal compression. Therefore the Mohr’s circle, corresponding to the centre of the
panel, is not centred in the τ − σ plane and the shear stress is computed according to Equation (3.2).
P
τelastic = 1.05 × (3.2)
An
In addition, the shear strain γ and the shear modulus of elasticity G can be calculated as follows:
τAST M,RILEM
GAST M,RILEM = (3.3)
γ
γ = εc + εt (3.4)
where εc and εt in Equation (3.4) are the strains along the shortening (compressed) diagonal and the
elongating (tensioned) diagonal of the panel, obtained from the average readings of the LVDTs located
on both sides of the wall. In turn, the initial shear stiffness modulus G is calculated as the secant
modulus between the origin and the shear stress at the first shift of the slope in the τ − γ curve, which
corresponds, on average, to the 30% of the maximum stress.
Ductility
The ductility is the ability that the structure has to sustain large deformations in the inelastic domain
of the response. The ductility factor is calculated as µ = γu /γy , is considered to characterise the post-
peak performance of the shear response. The ultimate shear strain γu is calculated as the post-peak
strain for which the corresponding stress reaches a reduction of 20% with respect to the peak one,
following Prota et al. (2006), Parisi et al. (2013), Marcari et al. (2017) and Balsamo et al. (2011).
The available approaches in the literature evaluate the yield strain γy according to different criteria.
Marcari et al. (2017) idealizes the experimental stress-strain curve with a bilinear law and calculates
the yield strain γy by using the shear secant modulus of the τ − γ initial branch at 70% of τmax .
66
The values of τy and γy are determined with an energy equivalence by equating the areas below the
experimental curve and the bilinear idealization. Gattesco et al. (2015b) define γy as the elastic shear
deformation corresponding to a value of the load equal to the peak load. The shear strain γy is given
by the ratio between the peak strength and the shear modulus G, as shown in Figure 3.14.
Figure 3.14: Evaluation of the ductility according to the approach proposed by Gattesco et al. (2015b), a)
evaluation of γy given by the ratio between the peak strength and the shear modulus G, b) evaluation of γu
which represents the ultimate shear strain corresponding to 20% shear strength reduction
The experimental tests carried out in the present research exhibited pronounced hardening be-
haviours after the initial linear loading branch and before reaching the peak resistance. Due to this
phenomenon, it was of paramount importance to select an appropriate approach to evaluate the duc-
tility factor. Since the methods presented in Prota et al. (2006), Balsamo et al. (2011), Parisi et al.
(2013), Marcari et al. (2017) would underestimate drastically the ductility factor due to the existence
of such hardening, this research considers the approach proposed by Gattesco et al. (2015b) to obtain
more consistent and realistic evaluations of the yielding strain.
67
3.7 Shear Compression Test

To investigate the performance, and therefore the efficiency, of TRM as post-earthquake repair or
seismic retrofitting the protocol provided by FEMA 461 (2007) was followed. The quasi-static cyclic
testing protocol replicates the seismic action by slow application of cyclic displacements. The response
of the structure subjected to slow rate displacement loading, can be considered rate-independent
and therefore can be analysed as a vis.cous single-degree-of-freedom (SDoF) systems under harmonic
vibration of steady-state motion. Such analysis is based on the fact that the frequency of vibration
ω of the test is equal to the natural frequency of vibration ωn of the structure. According to Chopra
(2001) the time required for a SDoF systems to complete a cycle of vibration when subjected to an
earthquake ground motion is very close to the natural period of the system Tn = 1/ωn . Consequently
the response of the specimen can be analysed as linearly elastic systems allowing the evaluation of key
parameter such as effective stiffness Ke , dissipated energy Ed , equivalent viscous damping ξeq with a
satisfactory approximation to the real seismic behaviour.
In addition, pseudo-static cyclic loading testes are used to study the hysteretic behaviour and
degradation of strength and stiffness under repeated loading.
3.7.1 Set-up
The samples were laid on a metallic C profile filled with concrete which prevented the sliding along
the artificially smooth surface when the horizontal load was applied, but allowed the sliding shear
failure to potentially occur in the bed joints of the first course of bricks as evidenced in Knox (2012).
In addition, the base was constrained at both ends by two T-shape devices, both the base and the
end-devices were fixed to the strong floor of the laboratory by means of post-tensioned steel bars.
On top of the wall a reinforced concrete beam was placed. The element had the double function
of providing height to uniformly distribute the vertical load and allocating the plates receiving the
horizontal cyclic loading by the actuator. On top of the RC beam laid a stiff metallic H profile stiffened
with ribs, where the vertical load was applied with two jacks of 1000 kN capacity each. The jacks
reacted against a stiff frame anchored to the strong floor. Between the RC beam and metallic profile,
a 3 mm thick Teflon sheet and a 3 mm thick PVC sheet was inserted to provide a smooth horizontal
surface to reduce the friction transfer of the horizontal load to the RC beam. Between the RC beam
and the PVC sheet a layer of cement based mortar of thickness 5 to 10 mm was introduced in order
to level the end surface of the beam and guarantee a uniform vertical load transfer.
3.7.2 Test procedure
The shear compression tests were performed in two steps. First, the vertical force V was gradually
applied under force control. The valves of the jacks were closed once the designed compression stress
was reached (σv =0.3 MPa). With the valves closed, no displacement or rotation of the top of the
wall was possible and applying horizontal load induced a double bending condition (Magenes and
68
Calvi (1992)). As a consequence of the testing mode, the vertical load increased together with the
horizontal load. The increase of the vertical load depended upon the stiffness degradation of the wall.
Secondly, the horizontal shear force H was applied with a hydraulic actuator anchored to a reaction
wall. The actuator had a pushing and pulling capacity of 350 kN and 250 kN respectively. Two steel
plates (530 × 300 × 30 mm3 ) connected by 4 steel rods of 40 mm diameter were mounted aligned with
the horizontal actuator. One of the plates was connected to the horizontal actuator by means of a
hinge, enabling the application of reversed cyclic loading in the horizontal direction, see Figure 3.16.
Wilding et al. (2018) investigated the influence of the different test set-ups for imposing the boundary
conditions in SCT. In addition to the one used for this experimental programme previously described,
some authors like Petry and Beyer (2015), Magenes et al. (2008) controlled the applied vertical load
with actuators in such way that the axial force remained constant during testing, consequently the
rotation of the top of the wall was restrained while the vertical displacement of the actuator remained
equal at both wall ends.
The walls were instrumented with 34 sensors, see Figure 3.15 to measure the following quantities:
- The force applied by the vertical jacks were measured by four pressure transducer and the force
applied by the horizontal actuator was measured by the load cell inside the actuator.
- The horizontal displacement at the top of the wall was measured by two laser sensor of 50 mm
range. The sensors were located on the thickness of the wall one on each direction of the applied
horizontal displacement.
- Two potentiometers and two LVDTs were located on the diagonals on each side of the wall to
measure de diagonal displacement.
- Two potentiometers on each side of the wall, located at the edges, to measure the vertical
displacement.
- Two LDVTs, one on each side of the wall, to quantify the slip between the RC beam and the
wall.
- Two LDVTs, one on each side of the wall, to quantify the slip between the wall and the metallic
base.
- One LDVT, located on one side of the wall to measure the uplift of the wall from the metallic
base and another LDVT on the other side of the wall to quantify the uplift of the metallic base
from the laboratory floor.
- Three wire sensors with 1000 mm range were located to measure the vertical displacement of
the stiff metallic beam.
- An inclinometer to control de movement of the metallic beam.
3.7.3 Loading Protocol
The standard provided by FEMA 461 (2007), recognised that for masonry structures damage is best
predicted by imposed deformation. Therefore, the protocol was designed to replicate a seismic action
by slow cyclic application of displacements. The adopted approach aimed to identify and capture the
69
Figure 3.15: Set-up of the Shear Compression Test
70
Figure 3.16: Set-up of the Shear Compression Test
accumulated damage, to enable the detection of the failure mechanism and to determine the force-
displacement properties. Such information is needed for the structural analysis and the assessment of
the masonry structure. As added value the cyclic testing was intended to provide better understanding
on the strength and stiffness degradation of the structure. As a result, the in-plane quasi-static cyclic
test was performed in displacement control, since it can be correlated to a building deformation
parameters such as drift, following the guidelines provided by FEMA 461 (2007).
The loading history consists of repeated cycles of step-wise increasing deformation amplitudes.
Three cycles at each amplitude were completed. The cycle is defined as the time needed to impose
the target displacement in the positive and negative loading direction up to returning to the original
position of the wall.
The loading history is defined by two parameters. The first one is the smallest deformation
amplitude ∆0 , which has to be smaller than the amplitude at which the lowest damage state is first
observed. The second parameter is the targeted maximum deformation amplitude of the loading
history ∆m . It is an estimated value of the imposed deformation at which the most severe damage
level is expected to initiate. The number of cycles between the two deformation amplitudes has to
be more than 10. Each following amplitude, starting from ∆0 increases its magnitude in 1.4 times,
meaning that the amplitude ai+1 of the step i + 1 is given by equation (3.5)
ai+1 = 1.4 × ai (3.5)
71
The value of the key parameters to design the loading history were determined taking into account
the research carried out by Morandi et al. (2018). The authors assembled a dataset with the results of
188 in-plane cyclic tests of unreinforced masonry wall. The walls present different masonry materials,
bed and head joint typologies, dimensions and boundary conditions. Only the tests performed on spec-
imens with similar mechanical characteristics to the specimens built for this experimental programme
were considered for this analysis.
The dataset gathers the main aspects related to the in-plane lateral strength, failure modes and
displacement capacity at different limit states. According to the conclusions of Morandi et al. (2018)
the drift at Significant Damage limit state (SD) and at Near Collapse limit state (NC) are the ones in
correspondence with the drift experienced at peak lateral load θHmax and at the end of the test θmax
respectively, as also established in EC8 (2004). Moreover, it is concluded that the drift at SD limit is
independent from the final failure mode, while the drift at NC significantly differs with the different
failure modes.
EC8 sets the limit drifts corresponding to the SD and NC states to θSD =0.4% and θN C =0.5%
respectively, for unreinforced masonry walls failing in shear mode. Similar results were obtained from
the dataset, in which the mean value for θHmax and θmax were equal to 0.42% and 0.46%.
As aforementioned, ∆m is the most important parameter when designing the loading protocol.
According to Magenes and Calvi (1997), the diagonal cracking load corresponds to 85% of the peak
shear force. Consequently the drift corresponding to this level of load is 0.39% with a displacement
value of approximately 5.0 mm. Once ∆m is defined, ∆0 is chosen so that there are between the two
displacements at least six amplitudes values.
It must be highlighted that the previous analysis considers only unreinforced specimens and thus,
the values might not be enough to initiate the damage expected in strengthened walls. On the other
hand, there is a lack of guidelines to define the drift at different limit states for retrofitted elements.
However, if it is considered that the strengthened specimen subjected to DCT doubled the displacement
capacity attained by unreinforced ones, the target parameter ∆m can be doubled and then considered
equal to 10.0 mm. This way the same loading protocol could be used for all the specimens allowing a
better comparison across them.
In any case, after the value of ∆m is reached the loading history will be continued using further
increments of amplitude of 0.3×∆m up to 50.0 mm. Which will give a safety margin to allow damage to
occur in the strengthened specimens since it was not reached by any specimen in similar experimental
programmes (Reboul et al. (2018), Hračov et al. (2016)).
The loading history adopted for all the SCT test is shown in Figure 3.17. In summary, the loading
protocol includes 29 displacement amplitude steps and a total number of 87 cycles.
72
Figure 3.17: Loading history for the shear compression testing of the wall specimens
73
3.8 Digital image correlation
Digital Image Correlation (DIC) is an innovative non-contact full-field optical method for measuring
strains and displacements. DIC works by comparing digital images of the specimen under testing
at different stages of deformation. By tracking the pixels from a spotted pattern throughout all the
available images, the systems can measure the displacements at the surface of the specimen and build
a 2D displacement/strain map. For an effective result, the pattern needs to be random and unique
with a range of contrast and intensity levels. The pattern can be achieved by different means, being
the black spray the most popular to create the stochastic pattern.
Due to the advances on the optical field, there is a number of software techniques available that were
developed to obtain sub-pixel resolutions, allowing high resolution measurements with commercially
available digital cameras. Consequently, DIC is now a non expensive measuring technique.
One of the main advantages of DIC is not only the detection of cracks initiations but also that allows
the measurement of displacement after cracking, when traditional sensor may fail due to detachment.
Therefore, if combined with traditional sensors, such as LVDT or potentiometers, during laboratory
testing, it may improve greatly the understanding of the mechanical behaviour of the specimen under
testing. As evidenced by the scarce literature on DIC application to characterize masonry behaviour
(Torres et al. (2020)), DIC makes possible the identification of strain concentration and consequently
the detection of cracks appearance, even when it has not been detected by the sensors yet, allowing
an accurate characterisation of the crack pattern.
One of the main limitations; however, is the resolution of the DIC, which depends on the pixel den-
sity of the camera, the dimension of the surface under study and the quality of the stochastic pattern.
The accuracy of the measurement will depend of the the camera resolution, but most importantly the
relation between the scale of the pattern and the magnitude of the measurement that is intended to
be measured. Therefore, particular attention must be paid in the specimens preparation. First, the
stochastic pattern realized on the surface of the specimen must assure high contrasts and black dots
distributed over white background is strongly suggested. Second, the sizes of the spots influence the
accuracy of the measurement, since error of measured displacements are closely related to the quality
of the pattern (Pan et al. (2010)), so it is important to identify the right scale between the pixels of
the camera and the millimetres measured on specimen. This could be done by predicting beforehand
the expected deformation of the specimen and dividing it by the number of pixels available on the
camera. According to this, the errors could be minimized by deciding whether the DIC will measure
the whole surface or just a region of interest. Finally, parallelism between the specimen’s surface and
the camera and lighting conditions should be kept stable during test execution.
For the current research, the DIC set-up consisted of a commercial reflex camera, model NIKON
D5600 with 24 megapixels accessorized with a target lens Nikon AF-S NIKKOR 18-140 mm, supported
on a fixed tripod. The light sensitivity ISO parameter was manually set to 100, since artificial lighting
was assured by external white light lamps. In addition, low ISO value gives the potential to produce
74
higher image quality. As a result of good illumination, it is possible to set short exposure time and
low aperture number. Considering the artificial lighting, the best fitting for the tests exposure time
was set to 1/20 sec and aperture number to f1/8. The camera’s features were kept constant for all the
pictures, which were taken at a 5 seconds time interval. The focal length (zoom) was set between 20-24
mm depending on the test in order to assure a real scale resolution in the range of 0.2-0.4 mm/pixel.
The surface of the specimen was treated before testing. After cleaning it, it was painted with mate
white paint for indoor walls. Afterwards, it was covered with an stochastic pattern by using a mate
black spray. The final result can be appreciated in Figure 3.18.
Figure 3.18: Surface treatment and stochastic pattern for DIC analysis
The image post-processing was carried out with a commercial software, Aramis GOM Correlate.
The software was chosen since it has included a feature to assess the quality of the stochastic pattern
once applied on the specimen and prior to testing. This allowed any needed enhancement to minimized
then possible errors.
Since the analysis is made from the stochastic pattern, the image is divided in small areas called
facets. The image information contained in the facets is used to identify the facets in all the following
images. Therefore, the pictures taken before and after deformation can be correlated. At the reference
stage, this is normally the first image taken, the facets are square of N × N pixels centred in a chosen
point that will be later used to determine its corresponding location in the deformed image. For all
the tests the facet size was set equal to 19 x 19 pixels, with a distance between facets centroids of 16
pixels. This way the facets are overlapped and no image information is missing.
DIC analysis allowed the derivation of the force-displacement response curves of the different tests,
and the comparison with the displacement measurements recorded with traditional sensors in order
75
to validate the results. Most importantly, the analysis allowed the identification of starting moment
of cracking it source point and the validation of the final crack patterns.
Finally, Annex A presents more information about the different issues related with the use of DIC
in the present research.
76
Chapter 4
Experimental Results: Diagonal

Compression Test
4.1 Introduction
Seismic performance of shear walls depends on their in-plane capacity to safely transfer the lateral
loads to the foundations and hence avoid the collapse of the structure. Simulation of this behaviour
can be achieved by inducing a diagonal compression force on a representative square masonry wall, as
proposed by EC8 (2004) and the Italian guideline NCT (2018). This testing methodology has been
widely used for the determination of the shear parameters of masonry due to its simplicity.
As already mentioned in Chapter 3, the experimental programme involves three different TRM
configurations and a joint repointing. Among the TRM configurations there is a continuous bidirec-
tional grid of basalt fibres and a textile of unidirectional steel fibres, both of them embedded in a
lime mortar matrix. For the steel TRM, two different yarn spacings are investigated to evaluate their
influence on the shear response. Joint repointing solution consists in an asymmetric layout with basalt
TRM on the inner face, and bed joints structural repointing with Near Surface Mounted (NSM) helical
stainless steel bars on the outer face. One of the novelties of the experimental campaign consists in
the testing on walls previously weakened with a small intentional defect localized in the centre of the
panel. Such defect was created with the purpose of inducing a more regular crack pattern and thus
less scattered experimental results.
All the reinforcement systems were applied on walls composed of solid handmade clay bricks and
low strength lime mortar joints. The experimental programme consisted in the execution of diagonal
compression tests (DCT) in order to assess the efficiency of the TRM systems for post-earthquake
repair and seismic retrofit, in terms of stiffness, load bearing capacity and ductility. The experimental
results are compared in terms of crack patterns, failure modes, shear strength, stiffness and ductility.
The evaluation of the ductility of the structural members required the selection of a proper model to
quantify and interpret correctly this parameter of paramount importance in the field of seismic design.
77
CHAPTER 4. EXPERIMENTAL RESULTS: DIAGONAL COMPRESSION TEST
4.2 Experimental results
The behaviour of the specimens is presented int he following sections in terms of cracking patterns
and experimental curves F − δ (Load – displacement). The displacement δ is calculated from the
average readings of the LVDTs located on both sides of the wall. Following the methodology proposed
in Chapter 3, Section 3.6.2, the shear stress τ and the shear stain γ were computed for every specimen
according to Equation 3.1 and 3.4.
4.2.1 Unreinforced samples and specimens retrofitted with basalt TRM
The unreinforced specimens, URM 1 and URM 2, exhibited qualitatively a similar behaviour in the
elastic range. The URM specimens showed a sudden drop in resistance shortly after the appearance
of the diagonal crack. The crack started from the induced defect and propagated towards the opposite
corners of the specimen. The failure of both specimens was characterised by a stair-stepped diagonal
crack pattern through the bed joints and opening head joints, as well as tensile splitting in the bricks,
especially in URM 1, as shown in Figure 4.1.
Figure 4.1: Crack patterns of URM specimens a) URM 1, b) URM 2
78
The peak load was 179 kN in URM 1 (δ=0.15 mm, γ=0.047%), while in URM 2 the peak load was
115 kN (δ=0.25 mm, γ=0.075%). Specimen URM 2 exhibited a rather stable load carrying capacity
in the softening branch, while URM 1 showed a more pronounced softening behaviour. Although the
intentional defect in the centre of the wall induced controlled crack patterns and failure mechanisms,
the shear capacity resulted scattered in the two specimens probably due to the variation of the me-
chanical properties of the masonry components. Such variation is also evidenced in the prevalent
tensile splitting of units along the diagonal crack of wall URM 1, as a consequence of more resistant
mortar, that causes higher shear capacity, as also found by other authors Prota et al. (2006), Giaretton
et al. (2018), Gattesco et al. (2015a).
These damaged walls were subsequently repaired by filling the cracks with the same lime based
mortar used for retrofitting. After being repaired they were retrofitted with LDB and tested again.
The purpose was, as already mentioned in Section 4.1, to assess the behaviour and effectiveness of
the TRM as a post-earthquake repair system. Both specimens, URM1 R and URM2 R, exhibited
qualitatively similar behaviour. An initial linear elastic behaviour was observed until the reopening
of the repaired cracks in the masonry. The first crack appeared around 60-80% of the maximum load,
causing a momentary drop of the load resisted and a reduction of the stiffness. This phenomenon may
be associated to load transmission from the masonry to the reinforcement system. After this point,
a progressive recovery of the load and the stiffness was observed, while a major number of small and
diffused cracks appeared along the diagonal and parallel to the first one. These cracks propagated
gradually towards the edges of the specimens as the load increased. The cracks widened due to the
progressive deformation until the end of the test.
The peak load was 272 kN for URM1 R (δ=3.38 mm, γ=0.98%) while for URM2 R was 241 kN
(δ=4.12 mm, γ=1.20%). The post-peak branch was characterised by a significant residual resistance,
which was due to a soft decrease of the shear resistance as a consequence of the progressive redistri-
bution of stresses along the bidirectional grid. The gradual deformation of the specimen led to the
failure of some of the yarns of the grid, see Figure 4.2. The redistribution of stress throughout the
basalt grid can be recognized in the fact that the diagonal crack pattern is distributed over a wider
area.
The repaired specimens exhibited an elastic stiffness very similar to that of the URM specimens,
i.e. a full recovery of the undamaged stiffness after the repair intervention. The basalt grid reinforce-
ment homogenized the response of the repaired specimens producing a more similar ultimate capacity
compared to the URM ones. In terms of deformability, the ultimate shear strain γu , associated with
the drop of 20% of the maximum shear stress, was 1.71% for the specimen URM1 R and 1.95% for
the specimen URM2 R. These values were of 0.26% for URM 1 and 1.06% for URM 2. .
The LDB was also applied as reinforcement to undamaged masonry specimens. Three different
phenomena were recognized after analysing the response of these strengthened panels. First, the
cracking of masonry, second the cracking of the mortar matrix and third the failure of the yarns of
the basalt textile. This sequence is in agreement with the response revealed by previous studies on
79
Figure 4.2: Crack patterns of damaged URM specimens repaired and retrofitted with basalt TRM and zoom
on the failure of some yarns at the end of the test. a) URM1 R, b) URM2 R
composites subjected to tensile test de Felice et al. (2014), De Santis and De Felice (2015).
Specimens LDB 1 and LDB 2 showed similar linear trends up to 70% of the peak load. Up to
this point, no damage was observed on the mortar coating, even though a decrease of the slope of
the experimental curve F − δ was recorded by the instruments. The first crack became visible, above
the compressed diagonal, at almost 90% of the peak load. After the appearance of these first hairline
cracks, a large number of thin parallel cracks developed in the centre and started propagating towards
the loaded edges along the compressed diagonal. As soon as the peak load was reached, several parallel
cracks developed and crossed almost completely the diagonal of the specimens over a diffused width.
At this stage, the specimen LDB 1 presented a peak load of 310 kN (δ=0.44 mm, γ=0.13%) while the
specimen LDB 2 reached 279 kN (δ=0.58 mm, γ=0.21%). After the peak load, a progressive reduction
of the resistance of the specimens was observed. The softening branch of both specimens was due to
a gradual widening of the cracks, spalling of the matrix cover and consequent exposition of the bare
textile. The progressive failure of some of the yarns generated a drop of the resistance in the post-peak
branch, followed again by a gradual decrease of the resistance. However, the overlapping of the net
of 300 mm in the centre of the walls did not undergo any detachment from the surface, showing the
good compatibility between the LDB and the masonry substrate. This outcome confirms the evidence
from the results of previous studies carried out by Santandrea et al. (2017), D’Ambrisi et al. (2013),
in which the effective bond length was greater than 200 mm. Figure 4.3 shows the crack patterns
of the specimens at the end of the test. Figure 4.4 presents the experimental curves F − δ (Load –
80
displacement) of unreinforced specimens (URM), repaired specimens and retrofitted specimens with
basalt TRM.
The repaired and retrofitted specimens exhibited an average increase of 177% of the ultimate shear
deformation. Similarly, the basalt reinforcement showed significant impact in the post-peak behaviour
of the retrofitted specimens by providing tensile strength after the masonry cracked, and yielding a
remarkable ductility compared with the unreinforced configuration. Thus the ultimate shear strain γu
of specimen LDB 1 was 0.73%, whereas that of specimen LDB 2 was 1.57%. This scatter was caused
by the earlier rupture of some of the yarns of the textile in the first specimen. Nevertheless, the
average increase was of 74%. For both the cases of repaired and retrofitted specimens, the presence of
the basalt grid homogenized the behaviour, reduced the scatter of the results and increased the peak
load and ductility.
Figure 4.3: Crack patterns of specimens retrofitted with basalt TRM a) LDB 1, b) LDB 2
81
Figure 4.4: Comparison of the diagonal compression load vs. opening displacement curves of unreinforced
specimens (URM), repaired specimens and retrofitted specimens with basalt TRM (LDB)
4.2.2 Specimens retrofitted with steel TRM
The LDS and MDS unidirectional textiles were installed with a layout composed of four horizontal
and four vertical 100 mm wide strips.
Figure 4.5 shows the F − δ experimental curves of the four specimens, LDS 1, LDS 2, MDS 1 and
MDS 2. Each pair of specimens presented, qualitatively, similar responses with a linear trend up to
50-60% of the peak load. After this point, a decrease of the slope of the F − δ experimental curves
was detected, indicating that masonry was cracking even though no visible cracks could be observed
from the exterior. In the cases of LDS specimens, the first cracks appeared, above the compressed
diagonal, at 80-90% of the maximum load. For the MDS specimens, the first diagonal cracks were
detected after the peak load mainly in the centre of the panel along the compressed diagonal. The
cracks in all the cases were diffused, and mainly located in the mortar coating of the strips, as shown
in Figure 4.6.
Specimen LDS 1 reached a peak load of 320 kN (δ=0.27 mm, γ=0.077%),whereas LDS 2 reached
237 kN (δ=1.36 mm, γ=0.41%). Specimen MDS 1 reached a peak load of 233 kN (δ=1.71 mm,
γ=0.54%) and specimen MDS 2 attained a peak load of 222 kN (δ=1.16 mm, γ=0.34%).
The two types of specimens showed some specific features in their post-peak branch. LDS 1
presented a noticeable constant decrease of the resistance until a diagonal displacement of 3.45 mm
(shear deformation of 1.0%), where the load became almost stable. LDS 2 showed an increasing
deformation under almost constant load until 3.5 mm (shear deformation of 1.06%). Despite this
difference, the post-peak softening of LDS specimens presented several drops in the resistance, as
a result of subsequent local debonding phenomena of some portions of the strips. The specimens
retrofitted with MDS showed an almost horizontal branch immediately after the peak load, until a
82
diagonal displacement of 3.0 mm (shear strain of 0.85%) was reached. After this stage, the specimen
MDS 1 experienced a drop of its resistance caused by the delamination of the horizontal strips, followed
by a series of local debonding phenomena, which can be clearly identified on the experimental curve.
The specimen MDS 2 evidenced a more gradual and progressive decrease of its resistance.
Finally, no failure of steel fibres was observed in any of the samples, as also evidenced in another
previous experimental programme with steel TRM Wang et al. (2018). However, the LDS samples
experienced debonding from the masonry substrate at the end of the strips near the edges of the
specimens, while MDS samples were characterised by the delamination of the textile within the matrix
rather than the debonding from the substrate. The possible cause of this phenomenon may be the
lower spacing between cords in MDS-TRM which provided lower interlocking to the mortar matrix.
Figure 4.6 shows the crack patterns as well as the debonding and delamination detected on the four
specimens at the end of the test.
Figure 4.5: Comparison of the diagonal compression load vs. opening displacement curves of unreinforced
specimens (URM) and specimens retrofitted with steel TRM consisting of unidirectional strips (LDS and MDS)
4.2.3 Asymmetrically reinforced specimens with basalt TRM and bed joints re-
pointing
The asymmetric system was composed of a bidirectional LDB grid on one face of the wall, and NSM
stainless steel rebars on the other face. Helical rebars were inserted with vertical spacing equal to three
courses in order to balance the contribution of the basalt mesh and thus to provide similar in-plane
strength on both faces of the walls.
Even though the slope (G) of the initial linear behaviour was different in the specimens, the global
behaviour of the two specimens with asymmetrical strengthening showed similar features. Both showed
linear trends up to 70% of the peak load. Up to this point, even if no damage was observed on the
83
Figure 4.6: Crack patterns of specimens retrofitted with steel TRM and zoom of the debonding of LDS strips
and delamination within the matrix of MDS, a) LDS 1, b) LDS 2 c) MDS 1, d) MDS 2
84
exterior mortar coating, the instruments detected a slight change in the stiffness (see Figure 4.7). The
first cracks, on the side of the basalt grid, were visible at almost 95% of the peak load. These cracks
were diffused throughout the width and along the compressed diagonal. At the same time, the other
side with the joint repointing experienced the formation of a series of stair-stepped cracks through the
bed and head joints, spreading from the centre towards the corners. Specimen LDB-JR 2 exhibited
more splitting failures of bricks than specimen LDB-JR 1, which might explain its higher peak load.
After these first thin cracks appeared on the mortar coating of the basalt grid, a major number of
thin parallel cracks developed and started propagating towards the loaded edges along the compressed
diagonal. After the peak load, widening of the cracks under progressive compressive displacement was
observed at both sides.
The post-peak branches of the specimens were characterised by a gradual decrease of the resistance.
This progressive reduction was caused by the complex cooperation between the strengthening effect of
the basalt grid at one side and that of the NSM helical rebars at the other. The rupture of some of the
yarns of the basalt grid occurred at almost the end of the test followed by a reduction of the marginal
resistance, as observed in the experimental curves. No rupture of the helical rebars was registered.
Figure 4.8 shows the final crack pattern observed in the specimens.
The specimen LDB-JR 1 reached a peak load of 185 kN (δ=0.45 mm, γ=0.14%) whereas the
specimen LDB-JR 2 reached a peak load value of 199 kN (δ=0.73 mm, γ=0.23%). The basalt grid
and the helical rebars had a great impact on the post-peak behaviour, providing tensile strength after
the cracking of masonry and allowing the specimen to develop a more ductile behaviour. The ultimate
shear strain γu of specimen LDB-JR 1 was 1.67%, whereas for specimen LDB-JR 2 it was 1.42%.
Figure 4.7: Diagonal compression load vs. opening displacement curves of unreinforced specimens (URM)
and specimens retrofitted with Basalt TRM in one side and joint repointing in the other (LDB-JR)
85
Table 4.1: Summary of the main results of URM and TRM retrofitted specimens
ID of the Wall F [kN] ∆F [%] τAST M,RILEM [MPa] τelastic [MPa] ∆τ [%] GAST M,RILEM [MPa] Gelastic [MPa] ∆G [%] γy [%] γu [%] µ ∆µ [%]
URM 1 179 - 0.32 0.47 - 1415 2102 - 0.022 0.26 11.6 -
URM 2 115 - 0.21 0.31 - 757 1124 - 0.027 1.06 39.2 -
Average 147 - 0.26 0.39 - 1086 1613 - - - 25.4 -
URM1 R 272 85% 0.45 0.67 73% 1170 1561 -3% 0.043 1.71 39.8 56%
URM2 R 241 64% 0.43 0.64 64% 1060 1554 -4% 0.041 1.95 47.7 88%
Average 256 74% 0.44 0.65 69% 1115 1558 -3% - - 43.7 72%
LDB 1 310 111% 0.53 0.79 105% 1600 2376 47% 0.033 0.73 21.7 -15%
LDB 2 279 90% 0.48 0.71 84% 1615 2398 49% 0.030 1.57 52.6 107%
Average 295 100% 0.51 0.75 94% 1608 2387 48% - - 37.2 46%
LDS 1 320 118% 0.59 0.87 125% 1296 1925 19% 0.045 1.80 39.8 57%
LDS 2 237 61% 0.43 0.64 64% 1011 1502 -7% 0.042 1.65 39.0 54%
Average 279 89% 0.51 0.75 94% 1154 1714 6% - - 39.4 55%
MDS 1 233 58% 0.42 0.63 63% 716 1063 -34% 0.059 1.35 22.9 -10%
MDS 2 222 51% 0.40 0.60 55% 840 1248 -23% 0.048 1.79 37.3 47%
Average 227 54% 0.41 0.62 59% 778 1156 -28% - - 30.1 18%
LDB-JR 1 185 26% 0.33 0.49 26% 1387 2060 28% 0.024 1.67 70.7 178%
LDB-JR 2 199 35% 0.36 0.53 37% 956 1420 -12% 0.037 1.42 38.1 50%
Average 192 31% 0.34 0.51 31% 1172 1740 8% - - 54.4 114%
86
Figure 4.8: Crack patterns of specimens retrofitted with Basalt TRM in one side (left) and joint repointing
in the other side (right): a) LDB-JR 1, b) LDB-JR 2
4.3 Comparison of solutions
This section presents a comparative analysis of the experimental results described in the previous
sections. The responses of the different TRM strengthening solutions are compared with the average
value of the URM walls in order to evaluate the gain in structural performance. Table 4.1 summarizes
the experimental results for all tested configurations. The following nomenclature is used: F is the
maximum load registered during the test, τAST M,RILEM and τelastic are the maximum shear stresses at
the centre of the panel according to the standards computed following Equations 3.1 and 3.2 presented
in Chapter 3, ASTM (2000) and RILEM TC 76-LUM (1994), and the elastic theory proposed by
Frocht (1931) respectively, GAST M,RILEM and Gelastic are the shear modulus of elasticity according to
the aforementioned nomenclature and finally γy , γu and µ are the yield strain, ultimate shear strain,
and ductility factor computed according to Section 3.6.2.
The repaired and retrofitted specimens (URM1 R and URM2 R) exhibited in all cases a higher
strength capacity and ductility than the URM specimens. In terms of stiffness, they showed significant
87
scattering probably due to the variability of masonry properties and the influence of the TRM layer at
the beginning of the test. In fact, the contribution of the reinforcement in the initial elastic range may
depend on the level of damage attained during the previous test and the effectiveness of the repair. In
spite of it, these specimens provided satisfactory results in terms of deformability, as they were able to
recover the initial stiffness of the URM original condition without experiencing a significant gain, which
would be undesirable from a seismic resistant point of view, as also highlighted by Benedetti (2019),
Babaeidarabad et al. (2014). The specimens retrofitted with LDB achieved a significant increase of
load bearing capacity and almost doubled the resistance of the URM specimens. In turn, the specimen
repaired and retrofitted with LDB showed a capacity increment of 74%. The application of the basalt
grid continuous reinforcement reduced the variability of the overall shear-deformation response of the
four walls investigated (both repaired-retrofitted and just retrofitted). The percentage increases of load
bearing capacity are respectively 85%, 64%, 111% and 90% in specimens URM1 R, URM2 R, LDB 1,
LDB 2 compared with the mean strength of the URM walls. In addition, the behaviour of the LDB
specimens was characterised by a significant residual resistance owing to the higher ductility provided
by the basalt grid. In fact the ductility increased 72% in the repair-retrofitted specimens (URM1 R
and URM2 R) and 46% in those just retrofitted (LDB 1 and LDB 2). No premature debonding from
the masonry substrate was observed in any of the tests performed on walls retrofitted with LDB. The
failure mode of the LDB was always rupture of the basalt yarns. The strengthening system revealed
to be compatible with the original URM material since the surface did not undergo any detachment
and consequently the stress transfer from the masonry substrate to the textile was achieved to the
extent of allowing the fibres to reach their ultimate tensile capacity.
The four specimens reinforced with LDS and MDS textile presented similar behaviour, despite the
difference in peak load detected in the first series (320 kN and 237 kN). The specimens retrofitted with
LDS textile showed an average increase of 89% in terms of peak load compared to URM, whereas the
increment was 54% for MDS textile. The lower spacing of yarns in MDS might explain this difference
rates due the lower textile-matrix interlocking compared to the LDS. This behaviour is in agreement
with the findings in Lignola et al. (2017), in which the reduction of the grid spacing of the textile
led to a lower performance of the reinforcement. The different yarn spacing between LDS and MDS
textiles, and therefore the different adherence of the textile to the matrix, caused different failure
modes. Delamination within the matrix and debonding from substrate were observed in the post-
peak stage of the MDS and LDS respectively, causing sudden drops of the capacity. Previous studies
by Santandrea et al. (2016), De Santis and De Felice (2015) highlighted the importance of an effective
textile-matrix interlocking in order to allow their proper cooperation and their combined debonding
from the masonry substrate. For this reason, it seems that LDS performs better than MDS, since the
latter exhibited premature textile-to-mortar interface failure. It is important to highlight that failures
of the yarns were observed neither in LDS nor in MDS, due to the high tensile strength of the steel
cords.
The strengthening system with asymmetrical layout, i.e. continuous LDB in one side and JR with
88
NSM rebars in the other, presented a moderate improvement of 31% in terms of peak load, with almost
no increment of the initial shear stiffness. However, the most important outcome was the remarkable
gain in ductility. After the cracking of the mortar joints, the original fragile behaviour of the URM
material was turned into a ductile response by the combined LDB-JR system allowing an increase
of ductility of more than 100%. The overall behaviour of these specimens confirms that this novel
solution could be useful to enhance the ductility of masonry façades with exposed bricks, in which the
application of continuous TRM is feasible only on the inner face of the wall. The use of NSM rebars
represents a minimally invasive reinforcement technique as highlighted by Casacci et al. (2019).
Figure 4.9 shows the rate of increment of each type of reinforcement configuration in terms of duc-
tility and peak load. The graph considers the average results of each pair of specimens. The displayed
histograms show visually the enhancement of capacity and ductility of the retrofitted specimens in
comparison with the reference URM walls. TRM systems with LDB and LDS show to provide the
best compromise between increase of resistance and ductility, meaning that a gain in strength is fol-
lowed by a consistent improvement in the ductile behaviour of the specimen. In turn, MDS and the
asymmetric LDB-JR systems yield much higher increase of one single property compared to the other
one (much higher strength in MDS, much higher ductility in LDB-JR). In addition, the LDB textile
have showed to be a very good retrofitting solution for damaged structures, as it increases both the
structural capacity and the ductile behaviour of the specimens.
Figure 4.9: Rate of enhancement of ductility and peak load for all the retrofitted specimens
89
4.4 Conclusions
This research has investigated the experimental shear behaviour of masonry walls retrofitted with
TRM systems. The experimental programme has comprised diagonal compression tests of ten ma-
sonry samples reinforced in the laboratory with three different TRM systems, based respectively on
continuous bidirectional grids of basalt TRM (LDB), discrete bands of unidirectional steel TRM (LDS
and MDS, i.e. with low and medium yarn densities), and a novel asymmetric layout combining a basalt
grid TRM on the wall’s inner face and bed joints structural repointing with NSM helical stainless steel
bars on the wall’s outer face (LDB-JR). The main conclusions of the research can be summarized as
follows:
All the adopted TRM solutions have demonstrated to be fully compatible with the masonry
substrate composed of solid clay bricks and lime mortar joints. No premature debonding failure
occurred before the peak resistance in the TRM retrofitted walls.
The application of TRM systems has improved the strength of the walls compared to the original
URM material. The highest rates of enhancement have been found in LDB and LDS systems.
The application of TRM systems has also remarkably improved the ductility of the walls com-
pared to the original URM material, without altering the initial stiffness. These outcomes
become of paramount importance in the seismic retrofit of existing masonry structures.
The repair of cracks with a M15 lime based mortar together with continuous LDB grid applied
symmetrically on damaged specimens has allowed a full recovery of the undamaged stiffness of
the walls. This result shows the capability of the investigated technique for post-earthquake
repair of existing masonry structures.
The application of continuous LDB textile has reduced the scattering of the experimental re-
sults in comparison with the URM walls. This means that the TRM application can mitigate
the possible influence of the variability of masonry properties and provide more homogeneous
structural response.
Unidirectional textiles of LDS and MDS fibres have provided qualitatively similar results, but
the former has exhibited better performance in quantitative terms due to the better interlocking
between the textile and the matrix. The increase of the yarn density does not necessarily lead
to an improvement of the structural performance.
All the adopted strengthening solutions have shown to be efficiently and easily implementable.
Among the four solutions, the application of LDB grid can be considered as the less time con-
suming, due to fact that applying a single surface layer is faster than applying a multiple strip
configuration. The NSM system turns out to be the less invasive and the most reversible tech-
nique among the ones tested.
90
Chapter 5
Experimental Results: Shear

Compression Test
5.1 Introduction
A huge effort has been devoted to the experimental evaluation of in-plane cyclic response of URM
piers, as highlighted by Morandi et al. (2018), yet a complete understanding of the in-plane cyclic
behaviour of strengthened masonry walls is still lacking.
For this reason, an experimental campaign involving three different TRM strengthening solutions
was designed. In addition, a post-earthquake repair solution was assessed. Eight solid clay brick ma-
sonry walls were built, of which two of them were repaired and strengthened after shear compression
test (SCT). The remaining six were retrofitted with a bidirectional basalt grid (LDB) and unidirec-
tional steel strips (LDS) prior SCT. Among the four specimens strengthened with LDS, two had one
layer of the textile, while the other two had two layers (LDS-DL).
A thorough analysis was carried out in terms of lateral load-bearing capacity, displacement capac-
ity, limit state, ductility, energy dissipation and damping coefficient. The analysis has helped to better
understand the failure mechanisms that led the piers to collapse, and to design better strengthening
solutions to prevent them. Hence, a rational decision on any treatment, repair or intervention can be
adopted.
5.2 Experimental Results
5.2.1 Failure modes
Figure 5.1 shows the final damage of the unreinforced specimens, characterised by a diagonal stair-
stepped cracking mainly developed through the mortar joints and by tensile splitting of some units.
The general behaviour of the URM specimens evidenced that, with increasing displacement amplitudes,
the cracks developed until the formation of one wide crack on each diagonal of the walls leading them
91
CHAPTER 5. EXPERIMENTAL RESULTS: SHEAR COMPRESSION TEST
into global failure. For both the URM specimens the first cracks were visible at the centre of the panel
at a drift θcr = 0.4% (displacement equal to δcr =5.5 mm). In turn, shortly after the concentration of
cracks on one diagonal, the maximum load was attained at a drift θHmax = 0.9% (displacement equal
to δHmax =10.9 mm). In specimen URM 1 the first cracks appeared during the loading in the pushing
direction, while in specimen URM 2 appeared during the pulling direction and later on the pushing
direction. The crack pattern of specimen URM 1 only shows one diagonal crack because the test was
not continued after the development of this first crack. In the case of specimen URM 2, the response
on both directions was symmetrical regarding the drift value at cracking onset and failure. After the
peak, the degradation of the specimen was sudden and characterised by the shearing of the corners
and crushing of the bricks. Horizontal cracking on the base of URM 1 was visible at drift 0.44% in the
pushing direction. At the end of the test these flexural cracks were visible on both opposite corners of
the specimens (see Figure 5.1-a). The pulling direction did not evidence any flexural cracks. Similarly,
cracking on the base of URM 2, on both directions, were visible at drift 0.61%. In this case, three of
the four corners were affected with flexural cracks by the end of the test. However, the cracks were
wider in the pushing direction rather than in the pulling direction. In spite of this, the cracks fully
developed onto shear failure due to tensile diagonal cracking.
The repaired and retrofitted specimens and the just retrofitted specimens showed also diagonal
cracks developed from the centre and upper part of the panel, with severe damage occurring in the
region where the two cracks intersected, see Figure 5.2 and 5.3. After the onset of cracking, the width
of the existing cracks increased at each following displacement amplitude, enlarging the damage.
On average, specimen URM1 R experienced its first visible cracks at drift θcr = 0.4% (δcr =5.16
mm). However, the cracks on the pushing direction appeared in an earlier stage θcr = 0.3% (δcr =4.08
mm), while in the pulling direction the cracks were visible in the following amplitude step (θcr = 0.5%,
δcr =6.3 mm). This can be explained considering the crack pattern observed at the end of the test of
URM 1. This specimen only developed a crack in the pushing direction which was first repaired and
later retrofitted. Similarly, specimen URM2 R exhibited the first cracks at θcr = 0.5%, but in an early
stage (θcr = 0.4%) the cracks were visible in the pushing direction. Conversely, the maximum load
was attained at drift θHmax = 1.1% (displacement equal to δHmax =14.3 mm) in both directions and for
both specimens.
Specimen URM 2 was heavily damaged and therefore repaired, thoroughly mainly in the corners
where the bricks were fractured. Those bricks were replaced by new ones, of the same typology, i.e.
handmade solid clay bricks, and bound with strengthening mortar. For this reason, the flexural cracks
on top and bottom arose in the second last joints after testing the URM2 R specimen. These flexural
cracks were visible mainly on one side of the specimen (see Figure 5.2 b right) on an early stage of
the test at drift 0.23%. The damage, visible as a horizontal crack, appeared also on the TRM at the
onset of cracking of the specimen.
The repaired and retrofitted specimens showed a more diffuse crack pattern in the centre of the
panel compared to the just retrofitted one. The LDB grid in the former case had to withstand more
92
Figure 5.1: Crack pattern at the end of the test: a) unreinforced wall URM 1, b) unreinforced wall URM 2
93
Figure 5.2: Crack patterns at the end of the test (left: front face, right: back face): a) Repaired and retrofitted
wall URM1 R, b) Repaired and retrofitted wall URM2 R
94
displacement cycles, starting at early stages of the test, because of the re-opening of the repaired
cracks. The cracks re-opening transferred the tensile stress to the LDB grid, causing the spalling of
the mortar and thus the diffuse crack pattern. It should be noted that not only the repaired cracks
were open during the test. As additional cracks took place on both diagonals, as a result of the stress
redistribution provided by the TRM strengthening system.
The case of the just retrofitted specimens presents a rather heterogeneous response when compared
with the repaired ones. Specimen LDB 1 evidenced a behaviour similar to the repaired specimens,
with its first visible cracks at drift θcr = 0.4% (δcr =5.61 mm) for the pushing direction and θcr =
0.6% (δcr =7.81 mm) in the pulling direction. However, specimen LDB 2 presented an early cracking
formation when the drift was equal to θcr = 0.3% (δcr =4.05 mm). The scattering can be attributed to
two possible factors, closely connected. First, the strength of the binding mortar of specimen LDB 2
was significantly lower than its pair (LDB 1), see Table 3.3, which could lead to an early failure of
the mortar joint and consequently masonry failure. Second, the failure mechanisms differed between
both specimens. In the case of specimen LDB 2 the failure mode evidenced a mixed mechanism of
shear failure, associated to diagonal tensile failure, and shear sliding failure on two central bed joints,
as shown in Figure 5.4. On the same line, each specimen attained its maximum load at different drift,
LDB 1 reached it at a drift θHmax = 1.1% (δHmax =14.3 mm) while LDB 2 at a drift θHmax = 0.9%
(δHmax =10.9 mm). The sliding in the pulling direction was also confirmed by the lack of flexural crack
on the corner corresponding to the loading in that direction.
Despite the aforementioned particular features, all the specimens retrofitted with LDB (repaired
and retrofitted, and just retrofitted) evidenced a homogeneous response in terms of crack pattern,
failure mechanism and lateral deformation capacity. The enhancement provided by the reinforcement
is appreciated at the displacement corresponding to the attainment of the maximum load. No toe-
crushing in compression was observed in any of the specimens during the test and up to the largest
displacement amplitude attained. However, once the corners started shearing off, it was observed a
strong degradation of the specimen. As a result, debonding was detected close to the corners of the
specimens where the load was applied. Such debonding, however, was a local phenomenon induced
by the metallic plates through which the cyclic load was applied and did not interfere on the attained
load or deformation capacities. Finally, it can be stated that all specimens underwent shear failure.
Figure 5.5 shows the crack patterns at the end of the tests of specimens retrofitted with the
strip configuration of LDS. It can be noted that the failure mechanisms that led to the failure mode
differ on some aspects. The difference lays on the fact that specimen LDS 1 underwent a mixed
failure mechanism and therefore only one diagonal crack, corresponding to the pushing direction, was
developed. However, flexural cracks at brick-bed joint interface were detected on the corners of both
specimens during the initial stages of loading, at low amplitude steps. In the following cycles, the
cracks on specimen LDS 2 evolved to a diagonal shear cracking from the centre of the panel towards
the corners, on both directions. On the other hand, specimen LDS 1 evidenced a number of sliding
interfaces. Such interfaces where located on the second and third mortar joint, and on two consecutive
95
Figure 5.3: Crack pattern at the end of the test (left: front face, right: back face): a) Retrofitted wall with
Basalt grid LDB 1, b) Retrofitted wall with Basalt grid LDB 2
96
Figure 5.4: Detail showing the bed joint sliding surfaces on specimen LDB 2
bed joints in the centre of the panel. As a result, horizontal cracks were formed on which the specimen
was able to slide mainly during the pulling direction. Therefore, no diagonal crack on that direction
was developed. This mechanism explains the asymmetrical response observed in the specimen.
Despite the scattering on the failure mechanisms, after reaching the maximum load the specimens
experienced delamination within the matrix and, consequently, spalling of the first layer of mortar in
the areas heavily cracked. The following levels of displacement were characterised by a progressive
debonding from the substrate of the horizontal and vertical strips. Finally, at the largest displacement
amplitude the specimen underwent toe-crushing on both directions as also crushing on the top corner
due to the metallic plate, see Figure 5.6. This mechanism did not interfere in the attainment of
the peak load or deformation capacity. Similarly to the cases retrofitted with LDB, specimen LDS 2
evidenced heavy damage at the centre of the wall where both diagonals intersected.
On average, specimens retrofitted with LDS presented the onset of cracking at a drift θcr = 0.6%
(δcr =7.84 mm), which represents an increment of 50% compared with URM specimens. The result
evidences the useful function of the horizontal strips as cracks arrestors. The drift corresponding to
the attainment of the peak load was θHmax = 1.1% (δHmax =14.3 mm) in wall LDS 2 for both directions.
Wall LDS 1 yielded a higher displacement capacity θHmax = 1.9% (δHmax =24.1 mm) due to the sliding
interfaces aforementioned.
Figure 5.7 shows the crack patterns at the end of the tests of specimens retrofitted with a double
layer of LDS strips. The failure mode was toe-crushing on both directions (push and pull) for both
specimens. However, it can be appreciated the diagonal transfer of the loading from its application
point on top of the specimen up to the bottom opposite corner. The general behaviour of the specimens
was symmetrical for both directions. The first cracks were visible at a drift θcr = 0.6% (δcr =7.84 mm),
97
steel textile LDS 1, b) Retrofitted wall with steel textile LDS 2
98
Figure 5.6: Toe-crushing observed at the end of the test on specimen LDS 2
while the peak load was attained at a drift θHmax = 1.4% (δHmax =17.5 mm) with a concentration of
wide cracks on the bottom corners. Shortly after, in the following displacement cycle, the specimens
experienced a complete crushing of the brickwork under compression, and local buckling of the strip
leading to debonding. As a result, it was evidenced a sudden degradation of axial and lateral strength
mainly in specimen LDS-DL 1. Before the toe-crushing took place, the damage experienced on the
corners of the specimens led to the appearance of horizontal flexural cracks in a number of mortar
bed joints over the base of the walls. These cracks allowed the uplift of the walls in the tension part,
with a consequent increase of the compression stress on the opposite compressive zone, leading to a
more severe crushing of the masonry. It should be noted that the increment of the number of layers
of the textile changed the failure mode from shear failure to toe-crushing. This change indicates that
the level of strengthening was excessive for the typology of masonry under study.
99
two layers of steel textile LDS-DL 1, b)LDS-DL 2
100
5.2.2 Force-displacement curves
Figure 5.8 and 5.9 show the experimental force-displacement curves under cyclic in-plane loading,
together with their envelope curves. These force-displacement envelopes are derived from the ex-
perimental hysteretic curves and constructed by connecting the peak force at the first cycle of each
displacement amplitude. The positive direction is the direction in which the horizontal hydraulic ac-
tuator pulls the specimen whereas the negative one is the direction in which the actuator pushes the
specimen (Figure 3.16). The displacement was measured at the top of the wall and its correspond-
ing drifts represents the lateral displacement over the total height of the specimen, expressed as a
percentage.
In general, all the experimental curves are characterised by an initial linear branch corresponding to
the undamaged behaviour of the structure. At the onset of cracking, and with the development of dam-
age and energy dissipation, the response becomes non-linear, as evidenced by the force-displacement
envelope, while at the same time wider hysteretic loops are observed. The hysteretic loops widen
progressively as the residual strength decreases until failure. This behaviour can be associated to a
typical shear failure mechanism, as also evidenced by Tomaževič (1999).
Among the unreinforced specimens, the attainment of the peak load was followed by the sudden
brittle failure due to masonry’s very limited tensile strength, as shown in Figure 5.8 a-b. Therefore,
the walls were unable to withstand inelastic deformation once the maximum lateral force was attained.
The evolution of the damage can be quantified with the development of the width of the hysteretic loops
for the increasing displacement amplitude. The wider the loop, the larger the damage experienced by
the specimen. URM 2 evidenced wider hysteretic loops (see Figure 5.8-b) which is in correspondence
with the more severe damage observed during and after the test, see Figure 5.1-b. The peak load of
URM 1 and URM 2 were -192.38 kN and -162.29 kN in the pushing direction and 150.7 kN and 151.56
kN in the pulling direction, respectively. The load capacity of the URM specimens shows significant
scattering, mainly on the pushing direction, probably due to the variation of the mechanical properties
of the masonry components.
The repaired and retrofitted walls showed a moderate improvement in lateral load-bearing capacity,
on average only 10% higher than URM walls, as shown in Table 5.1. However, the peak load was
reached at a larger displacement compared with the URM walls. The increment of displacement
was of 31%. Consequently, their global behaviour showed larger deformation capacity and ductility
compared to the URM specimens, as seen in Figure 5.8 c-d. However, specimen URM2 R showed
a more ductile and progressive damage than its pair URM1 R. This ductile behaviour, on the post-
peak, was achieved through energy dissipation translated into larger and wider hysteretic loops, while
being able to withstand damage at higher values of lateral force, before the strength degradation took
place. Each specimen showed a symmetrical force-displacement response in both directions. Specimen
URM1 R failed at a peak load of -175.09 kN and 180.08 kN on the pushing and pulling directions
respectively. In turn URM2 R attained the peak load at -194.56 kN and 167.88 kN in the pushing
and pulling directions, respectively. As a result, the response of these specimens can be deemed to be
101
homogeneous.
Figure 5.8: Force-displacement hysteresis curves and envelope curves of tested walls: a) unreinforced wall
URM 1, b) unreinforced wall URM 2, c) wall repaired and retrofitted with LDB, URM1 R, d) wall repaired and
retrofitted with LDB, URM2 R, e) wall retrofitted with LDB (LDB 1), f) wall retrofitted with LDB (LDB 2)
102
Figure 5.9: Force-displacement hysteresis curves and envelope curves of tested walls: a) wall retrofitted with
one vertical and horizontal layer of LDS (LDS 1), b) wall retrofitted with one vertical and horizontal layer of
LDS (LDS 2), c) wall retrofitted with two vertical and horizontal layers of LDS (LDS-DL 1), d) wall retrofitted
with two vertical and horizontal layers of LDS (LDS-DL 2)
The just retrofitted specimens showed a significant increment in both lateral-load bearing and
deformation capacity, 31% and 30% respectively, evidencing the effectiveness of the reinforcement
system. Figure 5.8 e-f shows the different failure mechanisms experienced by each specimen, as
mentioned in the previous section (Section 5.2.1). Such failure modes were due to shear in specimen
LDB 1, and a combination of sliding shear and tensile diagonal cracking in specimen LDB 2. This
difference between specimens can be also observed in the experimental envelope curves, as well as in
the shape and evolution of the hysteretic loops. The curve of LDB 2 is not as wide as its pair and, after
the post peak, evidences a rapid strength degradation. Regarding the shape of the hysteretic loops,
they correspond to the sliding failure mechanism, which allows higher dissipation of energy, i.e. is
wider loops. The heterogeneity of the failure modes had noticeable consequences on the displacements
attained at different damage stages. This can skew the overall results when evaluating the enhancement
of the reinforcement, underestimating the ductile behaviour. For this reason, in Table 5.1 the results
in terms of displacement of specimen LDB 2 are not considered in the average. However, the peak
103
loads are deemed to be sufficiently representative of the strength of the reinforcement, and have not
been disregarded in the calculation of the corresponding average. Specimen LDB 1 had a symmetrical
response and failed at a peak load of -213.10 kN and 207.16 kN on the pushing and pulling directions
respectively. LDB 2 exhibited a rather asymmetrical attainment of the peak load at -244.68 kN and
196.71 kN in the pushing and pulling directions, respectively. The lateral capacity measured in the
negative direction was about 7% larger than that in the positive direction, as a consequence of the
sliding failures previously described.
The specimens URM1 R, URM2 R and LDB 1 exhibited shear failure with progressive strength
and stiffness degradation after the peak load until the complete loss of lateral load bearing capacity.
Small strength losses are appreciated between the first and second cycle of the same displacement
amplitude, indicating that repeatedly imposing of the same displacement has a mild effect on the
TRM reinforcement.
The differences observed in the crack patterns of specimens retrofitted with LDS, can be also
spotted in the force-displacement curves in Figure 5.9 a-b. However, it is worth highlighting that the
post peak presents a particular feature in both specimens. The load was kept almost constant for sev-
eral displacement amplitudes before the degradation of the lateral strength. The phenomenon could
be due to a redistribution of the stresses throughout the strips of textile until debonding from ma-
sonry substrate. Figure 5.10 shows the crack pattern evolution on specimen LDS 2 along the diagonal
corresponding to the pushing direction. The first three images correspond to different displacement
amplitudes in which the load was almost invariable. The cracks evolved gradually until the displace-
ment amplitude of δ = 20.7mm, in which delamination within the matrix and debonding began,
causing sudden strength loss. With the following displacement steps, the damage caused by debond-
ing increased until reaching the final crack pattern (see Figure 5.5). Similarly, LDS 1 experienced the
gradual evolution of the diagonal crack but, due to the sliding surface, the post-peak response was not
reached.
Specimen LDS 1, had an asymmetrical response and failed at a peak load of -224.05 kN and 153.65
kN in the pushing and pulling directions respectively. The lateral capacity measured in the negative
direction was about 46% larger than the one in the positive direction. In the positive direction the
specimen displayed an evident premature loss of stiffness. The asymmetry was consequence of different
failure mechanisms experienced by each direction of loading. LDS 2 exhibited a more symmetrical
attainment of the peak load at -247.85 kN and 217.71 kN in the pushing and pulling directions,
respectively.
As aforementioned, specimen LDS 1 underwent a mixed failure mode and its results cannot be
compared either with those referred to its pair or to the URM specimens. For this reason, the results
from testing LDS 1 have been omitted when calculating average peak load and its corresponding
displacement. The LDS system presents the same enhancement that LDB in terms of displacement
capacity, i.e. 30% higher than URM, while it yields a 42% increment of lateral-load bearing capacity.
104
Figure 5.10: Crack pattern evolution in the pushing direction of specimen LDB 2
Specimens retrofitted with two layers of LDS strips, unlike the specimens retrofitted with one
layer, did not evidence large damage due to debonding since they underwent toe crushing at the
attainment of their peak load. Afterwards, specimen LDS-DL 1 experienced a significant degradation
of the vertical load-bearing capacity, as a result of loss of material on the compressed zone, which led
to the ending of the test. However, the lateral force capacity did not decrease immediately, as shown
in Figure 5.9-c. Along with the crushing of the compressed area, the vertical strips’ end suffered from
debonding from the masonry substrate. The specimen exhibited a symmetrical behaviour on both
directions, and the attainment of the peak load in the pushing and pulling directions were -265.20 kN
and 245.44 kN respectively. In the case of specimen LDS-DL 2, the degradation of the vertical load-
bearing capacity was not observed due to the formation of flexural cracks on the bed joints located on
top of the damaged corners. These cracks allowed the specimen to continue withstanding higher levels
of displacement without losing its lateral capacity, see Figure 5.9-d. The large material loss exhibited
on the corners was the reason why the test was stopped, since it was believed that a sudden failure due
105
to the degradation of the axial load-bearing capacity was prompt to occur. The specimen exhibited a
symmetrical behaviour on both directions, and the attainment of the peak load on the pushing and
pulling directions were -273.59 kN and 258.73 kN respectively.
On average the enhancement of the lateral force capacity represents only a 13% when compared
to the one layer LDS strip configuration. However, it represents an increment of 59% of load-bearing
capacity compared to URM. Regarding displacement capacity, the enhancement was of 26% and 64%
when compared to the one-layer LDS configuration and URM, respectively.
It is worth noticing that the shapes of the experimental hysteretic and envelope curves present
singularities regarding each type of experimental failure mechanism:
- Shear Failure: This mechanism was experienced by specimens URM, URM R,LDB 1 and
LDS 2. Figure 5.8 a-e and Figure 5.9-b, represent the typical response of shear failure. On early
stages, the width of the hysteretic loops increase gradually with each increasing displacement
amplitude evidencing a pinch when the reversal displacement goes through zero force. On
the onset of cracking, the width of the loops increase at a higher rate and the pinch at zero
displacement gets wider. Finally, after the peak load is reached, on both directions, the envelope
shows that the loss of strength is progressive but gradual, evidencing a great ductility. The slope
of the loops gets flatter as the displacement reaches the ultimate one.
- Shear Sliding Failure: This mechanism was experienced by specimens LDB 2 and LDS 1.
Figure 5.8-f and Figure 5.9-a show that on early stages the behaviour of the hysteretic loops
are similar to the shear failure one (described before), but the pinch is no longer at zero load.
Conversely, the intersection of loading and unloading branches with the load axis is located
symmetrically from the displacement axis and therefore there is no pinch, but a ”flag shape”
instead. At the onset of cracking the hysteretic loops get larger and wider. It must be highlighted
that such width is kept almost constant along the cycle. After the peak load, the strength rapidly
decreases, which is evidenced in the envelope curve, while the loops get flatter as they reach the
ultimate displacement.
- Toe-crushing failure: This mechanism was experienced by both specimens retrofitted with
two layers of LDS (LDS-DL). Figure 5.9 c and Figure 5.9 d present the curves of the toe crushing
failure mechanism. Unlike in the other mechanisms, this hysteretic curve is not characterised
by a pinch at zero force, but again, the intersection of the force axis is located symmetrically
opposed to the displacement axis. However, the width of the loops increases with each increasing
displacement amplitude, until the ultimate displacement. The slope of the envelope curve is
much steeper than the one observed in the curves corresponding to the other failure mechanisms.
Finally, a sudden loss of strength in the post peak response is observed.
106
Table 5.1: Comparison of peak loads and corresponding displacements of specimens tested and average incre-
ment in comparison with URM expressed in percentage
Maximum Displacement
∆Hmax ∆δHmax
Load at peak load Type of
Specimen [%] [%]
[kN] [mm] Failure
Hmax (-) Hmax (+) (-) (+) δHmax (-) δHmax (+) (-) (+)
URM Average -177.33 151.13 - - -10.90 10.91 - - Shear
URM1 R -175.09 180.08 -1.3% 19% -14.26 14.25 31% 31% Shear
URM2 R -194.56 167.88 9.7% 11% -14.25 14.23 31% 30 % Shear
Average 10% 31%
LDB 1 -213.10 207.16 20% 37% -14.22 14.17 30% 30% Shear
LDB 2 (*) -244.68 196.71 38% 30% -10.97 10.90 1% 0% Sliding Shear
Average 31% 30%
LDS 1 (**) -224.05 166.97 26.3% 2% -24.10 - 121% - Sliding Shear
LDS 2 -247.85 217.71 40% 44% -14.21 14.19 30% 30% Shear
Average 42% 30%
LDS-DL 1 -265.20 245.44 50% 62% -17.49 17.46 60% 60% Toe-Crushing
LDS-DL 2 -273.59 258.73 54% 71% -20.77 15.97 91% 46% Toe-Crushing
Average 59% 64%
(*) Specimen not considered in the displacement average.
(**) Specimen not considered in the average.
5.2.3 Bilinear modelling
To evaluate the performance of the walls in terms of unbiased parameters and to enable their compari-
son, the experimental load-displacement curves were idealized according to the bilinear model proposed
by Tomaževič (1999). The bilinear idealization is a simplified approach used in the study of the in-
plane seismic response of masonry walls. The idealization permits the determination of the response
spectra in a way similar to linearly elastic systems as also proposed by EC8 (2004) and the Italian
standard NTC (2018), idealising the force-displacement envelope curves as an elastic-perfectly plastic
relationship. In addition, the bilinear idealization enables the computation of ultimate displacement
δu , the ductility µ and the effective stiffness Ke as proposed by Magenes and Calvi (1997).
The bilinear approximation of the experimental envelope curves was obtained as described in Figure
5.11. First, the cracking point (δcr ,Hcr ), which is the moment when the first cracks became visible, was
identified along with the peak load (δHmax ,Hmax ). The identification of the cracking point was done
by finding the change of slope in the envelope curve. Such change meant a variation in the stiffness
due to cracking. Afterwards, the cracking point was validated with Digital Image Correlation. The
ultimate displacement δu was defined as the displacement at which the lateral strength had dropped to
80% of Hmax . If such a large drop was not attained, the largest displacement reached during the test
was taken as δu . The ultimate strength Hu is defined as the maximum load of the bilinear idealization
and is calculated taking into account the energy dissipation capacity of the experimental envelope
107
curve. The ultimate strength is determined so as to produce a bilinear curve enveloping the same
area below as the experimental envelope curve up to δu , defined as Aenvelope . According to Tomaževič
(1999) this ultimate strength of the bilinear curve can be obtained with Equation (5.1). The yielding
displacement δy , which is the displacement at the idealized elastic limit, is then defined as the ratio
between the ultimate strength Hu and the effective stiffness Ke . For the present application, except
for specimen URM 1, the bilinear idealization was calculated for both positive and negative directions.
The case of URM 1 was handled differently; in this case, the bilinear idealization was performed for
the negative direction and mirrored in the positive direction.
Figure 5.11: Example of bilinear idealization of an experimental horizontal load vs. displacement curve
s !
2Aenvelope
Hu = Ke δu − δu2 − (5.1)
Ke
Table 5.2 shows the ratio Hcr /Hmax which on average is equal to 0.69 (CV 10.30%). This value
is consistent with Eurocode 8 and the proposal of Tomaževič (1999) and Magenes and Calvi (1997)
for the bilinear idealization of the hysteretic behaviour of the wall, which propose a Hcr /Hmax value
equal to 0.7. However, the ratio Hu /Hmax yields a higher value than 0.9, which is the value proposed
by the aforementioned authors. The value of the ratio for all the tests is on average equal to 0.95
(CV 3.46%), which means that the idealized maximum for the different reinforcements tested is fairly
similar to the experimental maximum attained.
Figure 5.12 and Figure 5.13 show the bilinear idealization for each specimen tested. The dashed
lines display the main parameters that characterised the bilinear curve for both directions, as shown
in the example Figure 5.11.
108
Figure 5.12: Bilinear idealization of the experimental envelope curves of tested walls: a) unreinforced wall
URM 1, b) unreinforced wall URM 2, c) wall repaired and retrofitted with LDB, URM1 R, d) wall repaired and
retrofitted with LDB, URM2 R, e) wall retrofitted with LDB (LDB 1), f) wall retrofitted with LDB (LDB 2)
109
Figure 5.13: Bilinear idealization of the experimental envelope curves of tested walls: a) wall retrofitted with
one vertical and horizontal layer of LDS (LDS 1), b) wall retrofitted with one vertical and horizontal layer of
LDS (LDS 2), c) wall retrofitted with two vertical and horizontal layers of LDS (LDS-DL 1), d) wall retrofitted
with two vertical and horizontal layers of LDS (LDS-DL 2)
110
Table 5.2: Experimental cracking and peak loads, ultimate load computed for the bilinear idealization and
the ratio between the main bilinear and experimental parameters
Cracking Peak Ultimate

Hcr /Hmax Hu /Hmax
Specimen Load [kN] Load [kN] Load [kN]
Hcr (-) Hcr (+) Hmax (-) Hmax (+) Hu (-) Hu (+) (-) (+) (-) (+)
URM 1 -114.46 - -192.38 150.7 -176.02 - 0.59 - 0.91 -
URM 2 -102.31 104.96 -162.29 151.56 -154.22 146.37 0.63 0.69 0.95 0.97
URM1 R -120.70 126.05 -175.09 180.08 -161.53 174.42 0.69 0.70 0.92 0.97
URM2 R -142.20 132.44 -194.56 167.88 -181.99 162.87 0.73 0.79 0.94 0.97
LDB 1 -135.84 146.19 -213.10 207.16 -195.95 202.80 0.64 0.71 0.92 0.98
LDB 2 -152.48 122.12 -244.68 196.71 -221.36 175.27 0.62 0.62 0.90 0.89
LDS 1 -197.136 118.53 -224.05 153.65 -222.05 147.73 0.88 0.77 0.99 0.96
LDS 2 -213.57 152.76 -247.85 217.71 -242.51 213.33 0.86 0.70 0.98 0.98
LDS-DL 1 -199.18 177.09 -265.20 245.44 -253.58 240.83 0.75 0.72 0.96 0.98
LDS-DL 2 -161.2 178.28 -273.59 258.3 -23.18 257.43 0.59 0.69 1.00 0.99
Average 0.69 0.95
Deviation 0.08 0.03
C.o.V 10.30% 3.46%
5.2.4 Ductility
Paulay and Priestley (1992) define ductility as the ability of the structure to offer resistance in the
inelastic domain of the response. It includes the ability to sustain large deformations and the capacity
to absorb energy by hysteretic behaviour. From the parameters defined on the bilinear curve, it is
possible to compute the ductility factor µ of each specimen (Equation (5.2)). It is defined as the ratio
between the ultimate displacement δu of the bilinear idealization and the displacement related to the
appearance of the first cracks, the yielding displacement δy . The ultimate displacement computed on
the bilinear idealization response is the displacement at which the lateral strength has dropped to 80%
of Hmax . However, it does not mean that the failure was attained. In some cases, additional inelastic
displacement was still possible without reaching the collapse of the walls.
δu
µ= (5.2)
δy
Table 5.3 shows the key parameters to compute ductility for each specimen tested on both direc-
tions. URM specimens were characterised by a sudden and brittle failure and therefore δu represents
almost the failure of the specimen. Conversely, for the retrofitted specimens, the response was much
more ductile and there was still additional displacement capacity after the definition of δu . This is
reflected on an increment between 20% to 60% of the ductility compared to the URM. Such increment
points out the effectiveness of the reinforcement systems. The variability present in the increment
highly depends on the loading direction and the damaged mechanisms involved. For this reason, the
results of specimen LDS 1 were not considered.
The case of specimens repaired and retrofitted with LDB and the cases retrofitted with one layer
111
Table 5.3: Bilinear parameter to compute the ductility on each tested specimen
Yielding Ultimate
Displacement Displacement Ductility ∆µ
Specimen
[mm] [mm]
δy (-) δy (+) δu (-) δu (+) µ (-) µ (+) [%]
URM 1 -8.53 - -14.16 - 1.66 -

URM 2 -8.38 7.68 -13.95 12.26 1.66 1.60
Average 8.07 13.16 1.64
URM1 R -5.45 8.65 -15.99 17.07 2.93 1.97 76% 24%
URM2 R -7.19 9.62 -21.71 17.86 3.02 1.86 82% 16%
Average 8.20 17.93 2.45 50%
LDB 1 -8.09 10.83 -17.89 19.10 2.21 1.76 33% 10%
LDB 2 -5.88 5.80 -14.92 13.18 2.54 2.27 53% 42%
Average 7.87 16.23 2.14 32%
LDS 1 (*) -8.85 4.97 -24.10 9.63 2.72 1.94 64% 21%
LDS 2 -8.90 10.91 -25.05 25.26 2.81 2.32 69% 45%
Average 9.91 25.16 2.56 57%
LDS-DL 1 -9.98 10.61 -20.76 19.77 2.08 1.86 25% 17%
LDS-DL 2 -13.27 11.31 -24.06 22.48 1.81 1.99 9% 25%
Average 11.18 21.55 1.94 19%
(*) Specimen not considered in the average.
of LDS evidenced the highest ductility. In the latter case it was due to a minimal strength degradation
with increasing displacement during the attainment of the peak load, as mentioned in the previous
section.
5.2.5 Secant Stiffness
The slope of the idealized envelope is defined by Tomaževič (1999) as the effective stiffness Ke , or
elastic stiffness according to other authors such as Frumento et al. (2009), Vanin et al. (2017), Morandi
et al. (2018), Mercedes Cedeño (2019). The effective stiffness is defined as the secant stiffness to the
experimental envelope curve at the onset of cracking. This is, the ratio of the load at crack formation
to its associated displacement, Hcr /δcr . Some authors such as Messali et al. (2017), Godio et al.
(2019) and Vanin et al. (2017) proposed also the computation of the initial stiffness Kinitial . This
parameter is defined as the secant stiffness at 15% Hmax of the experimental envelope curve. Kinitial
can be considered as the stiffness of the uncracked section of the specimen. Table 5.4 shows both
stiffness computed for each specimen. The averages of the stiffness are derived from the positive and
negative branches of the experimental envelope curves. It is of interest to note that the initial stiffness
is marginally affected by reinforcement and number of layers. This can stem from the low thickness
112
of the strengthening systems applied. The repaired and retrofitted specimens recovered the stiffness
of the URM, with only a 30% increment, what reflects the goodness of the repair treatment.
Table 5.4: Initial and effective stiffness of tested walls
Initial Stiffness Effective Stiffness

Ke / Kinitial
Specimen [kN/mm] [kN/mm]
Kinitial (-) Kinitial (+) Ke (-) Ke (+) (-) (+)
URM 1 26 29 21 - 0.78 -
URM 2 34 37 18 19 0.54 0.52
Average 31 19 0.62
URM1 R 48 41 30 20 0.62 0.49
URM2 R 38 33 25 17 0.66 0.51
Average 40 23 0.57
LDB 1 33 37 24 19 0.73 0.50
LDB 2 59 46 38 30 0.64 0.65
Average 44 28 0.63
LDS 1 44 40 25 30 0.57 0.74
LDS 2 52 56 27 20 0.52 0.35
Average 48 25 0.55
LDS-DL 1 52 43 25 23 0.49 0.53
LDS-DL 2 39 39 20 20 0.51 0.51
Average 43 22 0.51
Average 0.57
Deviation 0.05
C.o.V [%] 8.71
The EC8 (2004) recommends the use of the cracked stiffness when assessing the behaviour of
masonry buildings under seismic actions. According to the standard, such stiffness, should be taken as
one-half of their respective uncracked value. The average value of the ratio Ke /Kinitial experimentally
obtained is equal to 0.57 (C.o.V= 8.71%) which is in line with the value given by the standard.
The recommendation is made for unreinforced masonry and there are no guidelines when it comes to
retrofitted masonry. However, the specimens retrofitted with different TRM systems are also in good
agreement with the EC8’s recommendations, as shown in Table 5.4.
5.2.6 Stiffness degradation
Due to the complexity of the damage mechanisms experienced by the wall specimens, the degree of
damage is not always easy to define. According to Reboul et al. (2018), a first approach to quantify
such damage consists of computing the stiffness degradation during the cycles. One of the main
113
aims of the experimental envelope curve is to allow a better estimation of the initial stiffness and its
degradation. Consequently, the evolution of the lateral stiffness at each level of lateral displacement
is computed in the first cycle of each amplitude of the in-plane loading protocol. The stiffness at
each level of displacement is defined as the secant stiffness to the peak force in each cycle considering
positive and negative directions, as proposed by Knox (2012), following Equation (5.3).
Hi+ − Hi−
Kei = (5.3)
δi+ − δi−
Where Hi+ and Hi− are the maximum forces in the positive and negative directions, respectively,
for cycle i with their corresponding displacements δi+ and δi− .
The hysteretic loops (see Figure 5.8 and 5.9) show that the secant stiffness decreases as the lateral
displacement increases, since the slope angle of the loops decreases with increasing displacement. This
is validated by Figure 5.14 which shows the degradation of the lateral stiffness suffered by the tested
specimens. The shape of the stiffness degradation is quite similar for unreinforced and strengthened
masonry walls, as stated by Tomaževič (1999).
The URM specimens evidenced their brittle response, since they lost stiffness more rapidly than the
other specimens. They halved their stiffness prematurely. Figure 5.14-a shows that at the moment
in which the cracking displacement (around 5 mm) is reached, the stiffness degradation rate was
already lower. This can be explained by the internal micro-cracking formation in the mortar-brick
interface, neither visible to the naked eye nor possible to be recorded by the applied instrumentation,
but detected by DIC.
The stiffness degradation among the specimens retrofitted with the LDB shows a reasonably con-
sistent trend for those that failed due to tensile diagonal cracking, while the specimens that exhibited
a mixed failure mode with shear sliding (LDB 2) evidenced the same shape but with a higher initial
stiffness. Similarly, specimens retrofitted with LDS evidenced a fairly similar trend despite the fact
that they presented different failure modes. However, the initial stiffness of the LDS 1 was considerably
lower.
From Figure 5.14 it is evident that high stiffness degradation took place at lower displacements
before any significant cracking was visible. At onset of cracking (between 5 to 10 mm of displacement),
the stiffness had almost halved (Table 5.4 shows that the ratio between the cracked stiffness and
uncracked stiffness is on average equal to 0.57) and it reached a phase in which the rate of degradation
was significantly reduced. Such reduction might be due to the deforming capability of the TRM
systems, when the cracks were opened. Once the peak load was attained, the degradation rate increased
once again until reaching the collapse of the specimens where the almost null stiffness was kept
constant.
The presence of the TRM reduced the brittle behaviour of the URM showing less degradation (for
the same displacement level). The fact that the strengthened wall attained their maximum lateral
force with less stiffness degradation is what makes possible the progressive and ductile post peak
114
Figure 5.14: Degradation of the lateral stiffness suffered by the specimens on SCT with increasing lateral
displacement: a) specimens retrofitted with LDB compared with URMs, b) specimens retrofitted with LDS
compared to URMs
115
response.
5.2.7 Limit States
The observed damage propagation and the crack pattern obtained were chosen as parameters for iden-
tifying different levels of damage throughout the testing. These levels are also linked to characteristic
points on the experimental force-displacement curves obtained for the tested walls. The definition of
these levels of damage allows a better assessment of masonry walls under seismic actions and estab-
lishes state limits for the design of retrofitting solutions. According to Vanin et al. (2017), Godio et al.
(2019) five limit states can be defined in terms of displacement and drift:
- Drift at cracking θcr : Is the drift for which the first cracks in the masonry are visible. These
cracks not always are identified by the naked eye, but as a change of the slope in the envelope
curve. In this research, the most of the cases could be identified by DIC. As a result of this
experimental programme, it corresponds to the displacement for which a load equal to 70% Hmax
is attained.
- Drift at yielding θy : Results from the bilinear approximation of the force-displacement re-
sponse. It is associated with the ultimate load Hu .
- Drift at maximum force θHmax : Measured when the peak force is reached. It marks the onset
of damage concentration in few cracks. It is also associated to the SD limit state defined by the
codes.
- Ultimate drift θu : Corresponds to the drift for which the reduction of the lateral strength is
equal to 20% of the peak strength, and is associated to the NC limit state introduced by the
codes.
- Drift at collapse θc : Ideally is related to the loss of axial load bearing capacity. The tests were
stopped before the vertical residual bearing capacity vanished. As an approximation for θc , the
drift at 50% drop of the lateral strength was taken.
EC8 defines (2004) two main limit states corresponding to Significant Damage (SD) and an ultimate
limit state associated with Near Collapse (NC). The limits are defined as a function of the failure mode.
For unreinforced masonry walls failing in shear mode, the limit drifts corresponding to the SD and NC
states are set to θSD =0.4% and θN C =0.5% respectively. The Italian standard NTC (2018) defines a
Ultimate Limit State (SLU), equivalent to EC8’s NC state, for which the drift limitation is also equal
to θSLU =0.5%. Table 5.5 and 5.6 show the limit states in terms of displacement and drift, respectively,
for all tested specimens. When compared with the limit drifts defined by EC8 and the Italian code, it is
observed that both standards underestimate the deformation capacity of the unreinforced tested walls
by 50%, see Figure 5.15-a. The average value of drift obtained in the experimental test, considering
both directions, for the URM specimens for the SD and NC limit states are 0.8% and 1.1% respectively.
Figure 5.15-b shows the different limit states computed for all the retrofitted specimens. There is
a lack of codes and guidelines regarding the limit states for retrofitted masonry. However, a research
116
Figure 5.15: a) Drift capacity of URM specimens compared with the SD and NC limit states set by EC8
(2004), b) drift capacity of retrofitted specimens according to the limit states defined by Vanin et al. (2017)
117
Table 5.5: Experimental displacements corresponding to different damage levels
Cracking Yielding Max Ultimate Collapse

Specimen Displ. [mm] Displ. [mm] Displ. [mm] Displ. [mm] Displ. [mm]
δcr (-) δcr (+) δy (-) δy (+) δHmax (-) δHmax (+) δu (-) δu (+) δc (-) δc (+)
URM 1 -5.55 - -8.53 - -10.89 8.94 -14.16 - -14.16 -
URM 2 -5.56 5.51 -8.38 7.68 -10.90 10.91 -13.95 12.26 -14.18 14.17
URM1 R -4.08 6.25 -5.45 8.65 -14.26 14.25 -15.99 17.07 -18.54 19.34
URM2 R -5.62 7.82 -7.19 9.62 -14.25 14.23 -21.71 17.86 -26.59 21.75
LDB 1 -5.61 7.81 -8.09 10.83 -14.22 14.17 -17.89 19.1 -20.46 21.75
LDB 2 -4.05 4.04 -5.88 5.80 -10.97 10.90 -14.92 13.18 -16.63 15.31
LDS 1 -7.85 3.99 -8.85 5.26 -24.10 7.77 -24.1 9.38 -24.10 10.90
LDS 2 -7.84 7.81 -8.90 10.91 -14.21 14.19 -25.05 25.26 -27.33 27.30
LDS-DL 1 -7.84 7.80 -9.98 10.61 -17.49 17.46 -20.76 19.77 -20.76 20.74
LDS-DL 2 -7.83 7.83 -13.27 11.31 -20.77 15.97 -24.06 22.48 -24.06 22.48
Table 5.6: Experimental drifts corresponding to different damage levels
Cracking Yielding Maximum Ultimate Collapse

Drift Drift Drift Drift Drift
Specimen
[%] [%] [%] [%] [%]
θcr θy θHmax θu θc
URM 1 0.4% 0.7% 0.8% 1.1% 1.1%
URM 2 0.4% 0.6% 0.9% 1.0% 1.1%
URM1 R 0.4% 0.6% 1.1% 1.3% 1.5%
URM2 R 0.5% 0.7% 1.0% 1.6% 1.9%
LDB 1 0.5% 0.7% 1.1% 1.5% 1.7%
LDB 2 0.3% 0.5% 0.9% 1.1% 1.3%
LDS 1 0.5% 0.6% 1.3% 1.3% 1.4%
LDS 2 0.6% 0.8% 1.1% 2.0% 2.2%
LDS-DL 1 0.6% 0.8% 1.4% 1.6% 1.6%
LDS-DL 2 0.6% 1.0% 1.4% 1.8% 1.8%
118
effort has been dedicated to better understand them. The experimental programme carried out on
Hračov et al. (2016) studied the in-plane behaviour of geo-net and wire ropes strengthening techniques
on adobe and clay brick masonry walls. The results detected visible cracks at drifts equal to 0.35%.
In turn, the ultimate drift (θu ) turned out to be, on average, equal to 1.0%. All the tested specimens
were characterised by shear cracks on each diagonal. Papanicolaou et al. (2007, 2011) investigated the
effectiveness of carbon TRM as strengthening material for brick masonry walls. The specimens were
subjected to in-plane cyclic test and even though most of the retrofitted specimen were characterised
by toe-crushing failure the ultimate drift was on average equal to 0.7%. Reboul et al. (2018) carried
out quasi-static loading tests on masonry walls strengthened with carbon and glass TRC, at the end
of the tests the specimens evidenced shear cracks along their diagonals. For these cases, the ultimate
drifts were equal to 0.61% in the case of C-TRC and 0.97% in the case of G-TRC. Bui et al. (2015) also
researched on the efficiency of textile reinforced concrete (TRC) composites as strengthening solutions
for masonry structures. In this case, the in plane load was monotonically applied yielding higher
values of maximum and ultimate drifts (θHmax and θu ), 0.52% and 1.03% respectively. Faella (2014) it
is also reported the values of maximum drift θHmax and ultimate drift θu for walls strengthened with
horizontal strips of LDS, which is the same material used in the present research. The drifts yielded
the values of 1.3% and 1.5% for the specimens failing due to shear cracks.
Despite the considerable effort put on researching the effectiveness of different strengthening sys-
tems for masonry elements, little effort has been directed towards defining the drift capacity for the
assessment of in-plane behaviour of retrofitted masonry walls. Figure 5.15 shows that the values of
drifts are considerably higher than the ones proposed by any of the standards for URM walls (EC8
(2004) and NTC (2018)) and no limits are defined when it comes to retrofitted masonry. Therefore,
from the results obtained in the present research, it is suggested a range for drift values for different
failure modes. Although these values are based on a still insufficient number of experimental tests for
one complete validation, they take into consideration the different results present in the state of the
art. Table 5.7 shows the ranges proposed.
Table 5.7: Proposed ranges of drift for strengthened masonry according to the failure mode
θcr θy θHmax θu θc
Type of Failure
[%] [%] [%] [%] [%]
Shear Failure 0.4-0.6 0.6-0.8 1.0-1.3 1.3-1.6 1.6-2.2
Mixed Failure 0.3-0.5 0.5-0.6 0.9-1.3 1.1-1.3 1.3-1.4
Toe-crushing Failure 0.6 0.8-1.0 1.4 1.6-1.8 -
5.2.8 Energy Dissipation
Another approach to determine the degree of damage experienced by the specimens is to correlate
the damage with the amount of dissipated energy. The energy dissipation is an important parameter
119
of the seismic behaviour not only for the assessment of existing structures but also for the design of
retrofitting solutions.
During the test, as the displacement amplitude continues to increase, a non-linear behaviour takes
place. The non-linearity is reflected in structural damage due to inelastic strains and characterised
by energy dissipation. The amount of dissipated energy at the ith cycle, Edi , is calculated as the area
within a hysteretic loop of a complete loading cycle, as shown in Figure 5.16. The bigger the area
inside the loop, the larger the amount of energy dissipated and consequently, larger damage.
The amount of dissipated energy Edi was evaluated on every displacement amplitude for which the
first loading cycle was completed. The cumulative dissipated energy ED was defined as the sum of the
dissipated energies of all the first loading cycles (for displacement amplitude δi ), from the beginning
of the test until the last completed cycles. Figure 5.17 shows the cumulative energy dissipation for
all the specimens tested. The repaired and retrofitted walls provided higher deformation capacity to
the URM walls, and therefore the ability to dissipate more energy by reaching greater displacement
amplitudes in the inelastic domain, after the attainment of the peak load.
Figure 5.16: Computation of dissipated and strain energy of each displacement amplitude cycle
In general terms, all the tested specimens exhibited a similar trend. The value of cumulative energy
dissipation was almost the same for all the specimens until the cracking displacement (δcr ) of URM.
After the cracking displacement, the retrofitted specimens displayed a noticeably high dissipation
capacity mainly due to the ductile response given by the TRM after cracking.
It should be noted that the different failure modes, previously described in Section 5.2.1, can be
120
Figure 5.17: Cumulative dissipated energy of tested specimens
also evidenced in the energy dissipation parameter. The specimens that were governed by a pure
shear failure, i.e. both URM, both URM R, LDB 1 and LDS 2, achieved fairly similar cumulative
energy dissipation values with increasing displacement amplitudes. Once the ultimate displacement
(δu ) was reached, the dissipation capacity of the following cycles decreased and the cumulative energy
dissipation curve was flattened. The meaning of this behaviour arises in the fact that the extension of
the damage suffered by the wall was so large that was no longer capable of dissipate any more energy.
However, such behaviour was not observed in all the cases. Specimen LDS 2, evidenced the same trend
as the others but did not flatten after δu since the specimen underwent a combination of toe-crushing
and strip debonding after the ultimate displacement rather than a widening of its principal diagonal
cracks. The same happened with the specimens governed by sliding shear (LDB 2 and LDS 1). These
two specimens had the same cumulative energy dissipation value until the failure of LDB 2. However,
there was no such loss of dissipation capacity after δu as in the previous specimens. The values of ED
were higher for sliding shear failure (LDB 2 and LDS 1) rather than pure shear failure.
The specimen LDS-DL 1 that failed due to toe-crushing evidenced higher dissipation capacity
than those failing in shear but lesser dissipation capacity than those failing in sliding shear. Even
121
though specimen LDS-DL 2 failed due to toe-crushing, it evidenced higher energy dissipation capacity
for lower levels of displacement and up to the cracking point. Afterwards, its dissipation capacity
was more similar to the specimens failing in shear rather than its pair failing due to toe-crushing.
The explanation may lie on the severe damage observed in specimen LDS-DL 1 when compared to
LDS-DL 2, which would allow a much more progressive and ductile post-cracking behaviour.
5.2.9 Damping
The damping of a structure of a single degree of freedom can be idealized by a linear viscous damper
when subjected to a steady-state motion, such as the quasi-static loading history. This idealization is
called equivalent viscous damping and it is the simplest form of damping since the governing differential
equation of motion is linear. Unlike the stiffness of a structure, this damping coefficient cannot be
calculated from the dimensions of the structures. Chopra (2001) presents the most common method
for defining it, equating the energy dissipated in a vibration cycle of the actual structure and an
equivalent viscous damping system.
In a structure subjected to a experimental cyclic loading, the dissipated energy Edi is given by
the area enclosed by the experimental force-displacement hysteresis loop. Chopra (2001) defines the
energy dissipated by viscous damping in one cycle of harmonic vibration for a steady state motion of
a SDoF systems as shown in Equation (5.4).
Z u Z 2π/ω Z 2π/ω
ω
Ed (t) = fD du = cu̇2 dt = c [ωu0 cos(ωt − φ)2 ]dt = 2πξ ku2 (5.4)
0 0 0 ωn 0
When both terms Edi and Ed are equated, the equivalent viscous damping ξeq for one loading cycle
in a steady state motion SDoF system, this is ω = ωn (Section 3), is
1 Ed 1 Ed
ξeq = = (5.5)
2π Ku20 4π ESo
In Equation (5.5), the strain energy ESo is calculated according to Equation (5.6) and Figure 5.16
considering u0 as the average between the displacement amplitude δi in the positive and negative
directions, and K is the secant stiffens experimentally determined.
Ku20
ESo = (5.6)
2
The equivalent viscous damping is a good indicator of the energy dissipation capacity and is a
function of failure mode, as also observed in Knox (2012). As aforementioned in Section 5.2.2, the
shape of the hysteretic loop is in close relation with the failure mode underwent by the specimen, and
the equivalent viscous damping follows the energy dissipation in one completed cycle, therefore is also
influenced by the failure.
It is important to note that equivalent viscous damping is intended to model the energy dissipa-
tion at deformation amplitudes within the linear elastic of the overall structure. Over this range of
122
deformations, the damping coefficient may vary with the deformation amplitude. The non-linearity
observed on the damping coefficient due to increasing deformation can be handled by selecting a
damping coefficient to characterise the wall associated with the linearly elastic limit of the structure.
Figure 5.18-a shows the equivalent viscous damping of all the tested specimens. All specimens
evidenced a similar trend with increasing displacement amplitude. The damping coefficient was fairly
constant until the first cracks were visible, i.e. δcr , see Figure 5.18-b, which is line with the findings of
other authors such as da Porto et al. (2011), Hračov et al. (2016), Yacila et al. (2019). Subsequently,
the damping increased due to the evolution of damage.
In the cases of specimens characterised by shear failure it was evidenced a second level of fairly
constant damping coefficient until failure, δHmax . Beyond this point the equivalent viscous damping
coefficient ξeq was considerable higher due to large energy dissipation as a result of severe damage.
In the cases of specimens characterised by other type of failure, such as sliding, the damping largely
increased after the cracking point. The increment is due to the fact that a sliding cyclic response
is very close to an indefinitely elastic perfectly plastic response, as defined by Magenes and Calvi
(1997). Therefore the evaluation of the damping coefficient would lead to very high values which
are unconservative and not representative of the real dissipation mechanism. Specimen LDS-DL -
1 characterised by toe-crushing did not exhibit any constant trend at any point, as it constantly
increased. While specimen LDS-DL 2 showed a sharp increment of the pre-cracking damping. The
phenomena is related with an increment of hysteretic loop width, as a result of some internal damage
suffered by masonry but that was not reflected on the surface of the specimen. For this reason, this
value was not taken into consideration for the average value of pre-cracking damping, since its deemed
not representative of the behaviour of the specimen. In addition, and unlike its pair, the specimen
evidenced a fairly constant post-cracking damping coefficient.
Table 5.8: Uncracked and cracked equivalent viscous damping
Pre-cracking Post-cracking ∆ξeq

Specimen
ξeq C.o.V ξeq C.o.V [%]
URM 1 3.39 21% 5.3 12% -

URM 2 3.24 27% 7.5 48% -
URM1 R 3.28 16% 11.8 41% 84%
URM2 R 3.49 26% 11.5 17% 80%
LDB 1 4.38 40% 16.5 26% 158%
LDB 2 † 3.27 46% 10.0 12% 55%
LDS 1 † 3.96 13% 19.9 18% 209%
LDS 2 3.80 9% 16.7 63% 161%
LDS-DL 1 * 4.44 36% 12.2 24% 90%
LDS-DL 2 * 4.54 32% 8.55 20% 33%
* Specimen characterised by toe-crushing failure.
† Specimens characterised by mixed shear sliding failure.
Table 5.8 reports the equivalent viscous damping ξeq calculated for each tested specimen before
123
Figure 5.18: a) Equivalent viscous damping ξeq of the tested specimens, b) Zoom of the damping coefficient
up to cracking (δcr )
124
and after the formation of shear cracks. It must be noted, as mentioned before, that pre-cracking
damping coefficient is similar in all the specimens independently of their failure mode. In addition,
the average value of the pre-cracking damping coefficient was equal to 3.8 (C.o.V 13.8%). This value
is closed to the 5% which is the default viscous damping used for the linear response analysis proposed
by EC8 (2004). The equivalent viscous damping coefficient computed after cracking is larger than the
linear elastic one and, to some extent, dependent on the type of failure. The repaired and retrofitted
specimens showed an average increment of the damping coefficient of 82% when compared with the
URM ones. In turn, specimens retrofitted with LDB and LDS, failing in shear, exhibited an increment
of 158% and 62%, respectively. Interestingly, the specimens retrofitted with double layer of LDS failing
due to toe-crushing also evidenced a 61% increment when compared to the URM configuration.
On average the post-cracking equivalent viscous damping associated to retrofitted specimens with
shear failure is 12.6% (C.o.V 21.6%), which is similar to the values obtained by Magenes and Calvi
(1997) for specimens failing due to shear under cyclic testing.
5.3 Comparison of solutions
The use of reinforcement systems mainly aims to the improvement of some seismic resistance param-
eters. The most important ones are the restoration of initial stiffness, which is closely related to the
gain of lateral load-bearing capacity and the improvement of the ductility, allowing the structure the
dissipation of larger amounts of energy.
In order to evaluate and compared the test results described in the previous sections, a correlation
study was carried out in which a reinforcement ratio ρ was introduced as an explanatory variable.
The reinforcement ratio, in Equation (5.7), represents the ratio between the axial stiffness of the TRM
reinforcing systems and that of the masonry, as defined by Babaeidarabad et al. (2014), where AT RM
is the transversal area of the TRM reinforcement systems, An is the net transversal area of the wall,
Ef is the elastic modulus of the textile fibre of the TRM reinforcement systems, and Em is Young’s
modulus of masonry. Table 5.9 shows the reinforcement ratio computed for each TRM reinforcement
system used in the current research.
AT RM Ef
ρ= 100 (5.7)
An Em
To better understand the influence that the different TRM systems may entail in the final strength
capacity and ductility, the increment of both parameters were correlated with the incremental ratio ρ
for each TRM system.
Figure 5.19-a illustrates that the percentage of enhancement in terms of lateral load-bearing capac-
ity of the strengthened walls, when compared to URM, increases proportionally with the reinforcement
ratio. However, the experimental evidence showed that walls strengthened with higher reinforcement
ratios, as for instance 2.80, failed due to toe-crushing of masonry. Therefore, despite the strong linear
correlation between the increment on lateral capacity (∆H) and the reinforcement ratio, ρ should be
controlled in order to hinder this typology of failure, so the TRM reinforcement systems could be fully
125
Table 5.9: Reinforcement ratio ρ of studied TRM reinforcement systems
An [mm] Em [MPa]
Masonry 393700 2318
TRM System AT RM [mm] Ef [MPa] ρ

LDB 81.28 90 0.8
LDS 67.2 190 1.4
LDS-DL 134.4 190 2.8
exploited. In addition, Figure 5.19-b displays the relation between the reinforcement ratio and the
ductility increment (∆µ). This increment was obtained comparing the ductility of the strengthened
walls to the one referred to the URM walls. Interestingly, a high ratio ρ of reinforcement may lead to
a better performance of the strengthening system in terms of lateral load-bearing capacity. However,
such enhancement is achieved at the expense of the ductility. The quadratic correlation illustrated in
Figure 5.19-b emphasizes the concept of a optimal reinforcement ratio for which the ductility incre-
ment can be maximized. The optimal ρ corresponding to an optimum compromise between ductility
and strength is 1.63. For this ratio, the ductility increment is equal to 63%, while yielding a lateral
capacity increment of 43%. Following the design and features of the textiles studied, having a ratio ρ
equal to 1.63 would mean a steel textile with a equivalent thickness of 0.093 mm and a yarn spacing
of 3.00 yarns/cm or a bidirectional basalt grid with an equivalent thickness of 0.065 mm and a grid
spacing of 15 × 15mm.
Although there is a clear correlation between ρ and the two main parameters under study, µ and
lateral capacity, none of the explored combinations with the remaining parameters evaluated, ED and
ξeq , evidenced any strong relation with ρ.
The suitability of each type of TRM systems would depend on the particular requirements of the
strengthening intervention. Basalts grids have shown to moderately increase the capacity and ductility
of the walls mainly achieved by the grid’s capacity of stress redistribution. The grid geometric feature
enables the cracks stitching independently from the cracks’ direction. Therefore, the damage is much
more severe, since it covers more surface, but most importantly is controlled. In addition, the good
compatibility, both physical and mechanical, between masonry and LDB TRM is achieved thanks to
the lower stiffness of the textile. Hence the exploitation of the textile is almost fully, since the basalt
yarn reached the failure at the end of the test.
Unidirectional steel textiles offer a fairly similar enhancement of the lateral load-bearing capacity,
like LDB TRM, when only one layer of textile is installed. However, due to the inherent nature of steel,
its contribution to the ductility enhancement is higher. The case with added layers is different, which
should be left for strengthening interventions where load capacity must be improved over ductility.
However, unlike LDB, compatibility between the LDS TRM and masonry is much more complex. The
126
Figure 5.19: a) Correlation between lateral capacity increment (∆H) and reinforcement ratio (ρ), b) correla-
tion between ductility increment (∆µ) and reinforcement ratio (ρ)
127
debonding occurs suddenly, releasing large amounts of energy, which are usually followed by stiffness
and strength degradation.
5.4 Conclusions
This research has investigated the experimental in-plane cyclic behaviour of masonry walls retrofitted
with different TRM systems. The experimental programme has comprised quasi-static cyclic shear
compression tests of eigth masonry samples reinforced in the laboratory with three different TRM
systems, based respectively on continuous bidirectional grids of basalt TRM (LDB), discrete bands of
unidirectional low density steel TRM (LDS) with one and two layers. The main conclusions of the
research can be summarized as follows:
Comparing the crack patterns of the URM walls to the strengthened ones, the latter experienced
a more widespread damage involving both masonry and matrix. The presence of TRM allows
a proper redistribution of stress, and therefore, specimens can withstand larger imposed loads
and displacements, proving the efficiency of the strengthening technique.
Specimens with URM configuration and those retrofitted with LDB and one layer of LDS under-
went shear failure, and consequently could exploit, to different extent, the TRM tensile capacity.
The specimen retrofitted with two layers of LDS led to masonry crushing and therefore indicated
an excessive strengthening solution.
The shape of the experimental force-displacement curves, along with their corresponding ex-
perimental envelope curves, can be associated with different failure modes. Consequently, the
hysteretic shape is therefore dependent on the failure mode undergone by the specimen.
All the TRM reinforcement systems not only enhanced the global in-plane shear behaviour of
the specimens but also homogenized their structural response, reducing the inherent effect of
the influence of scattering of the masonry constituents.
The application of TRM systems has improved the strength and displacement capacity of the
walls compared to the original URM material. Lateral load-bearing capacity was increased in a
range between 10%-60%, while ductility was increased in a range between 20%-60% compared to
the one observed in the URM configuration. Ductility was also evidenced in a greater capacity
of energy dissipation after the cracking point of the specimens.
The repairing technique ”scuci-cuci” used on the damaged walls has allowed a full recovery
of the undamaged stiffness of the walls. The initial stiffness Kinitial is marginally affected by
reinforcement due to the low thickness of the strengthening systems applied. The effective
stiffness Kef f ective can be taken as the cracked stiffness when assessing the behaviour of masonry
building under seismic actions.
128
Due to the lack of codes and guidelines on limit states of strengthened structures, a range for
drift values for different failure modes were suggested in line with the findings of the present
research.
The effect of the different failure modes undergone by the specimens was recognised in the energy
dissipation capacity and therefore in the equivalent damping coefficient, since both resulted in
close relation with the shape of the hysteretic loops. However, pre-cracking ξeq coefficient was
independent of the failure mode since all specimen yielded similar values.
129
Chapter 6
Correlation between Diagonal

Compression Test and Shear
Compression Test
6.1 Introduction
Nowadays, there are different testing methodologies available to assess the shear behaviour of masonry.
To name a few, the direct shear tests of masonry couplets or triplets masonry specimen, the diagonal
compression test (DCT) with or without edge load application, and the shear compression test (SCT).
However, the evaluation of the shear strength of masonry does not only depend on the material
properties but also on the boundary conditions featuring the set-up configuration, making it difficult
to assess the in-plane behaviour of masonry in an unbiased way.
The direct shear test on small triplets is the simplest and most inexpensive test, but the parameters
measured, τ0 and µ, are representative of the behaviour of the bed joint rather that of the masonry
assemblage, as stated by Crisafulli (1997). Therefore, its validity is restricted to the failure of the brick-
mortar interface as a result of sliding or detachment, normally associated to poor quality masonry or
low normal stress levels.
DCT and SCT testing configurations make use of larger specimens to investigate the in-plane
behaviour of masonry panels in a more realistic way, so that the results can be extrapolated for full
size masonry members. Turnsek and Cacovic (1971), Mann and Müller (1980) and Hendry and Khalaf
(2001) identified the role of tensile strength in the definition of the in-plane behaviour of masonry,
but unfortunately its direct experimental characterisation is not always feasible. Unlike couplets or
triplets, large specimens cannot be tested through direct tensile tests, and masonry’s tensile strength
needs to be obtained from an indirect testing methodology. As a result, tensile strength has to
be assessed from DCT and SCT when the failure mode of masonry is featured by tensile diagonal
cracking. Normally, this mechanism is triggered as a result of medium to high normal stress combined
131
CHAPTER 6. CORRELATION BETWEEN DIAGONAL COMPRESSION TEST AND SHEAR
COMPRESSION TEST
with lateral loading, which is the most representative case of existing multi-stories masonry buildings.
However, there are different theories in the available scientific literature about the interpretation of the
experimental results derived from the tests, mainly differing in the definition of the stress state of the
specimen. Nevertheless, since both tests can be used to characterise the parameters that govern the
in-plane behaviour of masonry, an integrated methodology considering the redundant results derived
from both tests can be developed in order to achieve a more reliable characterisation of the in-plane
mechanical properties of masonry walls.
The main novelty of the present work resides in the correlation between the tensile strength and
the shear behaviour of masonry walls in the unreinforced and retrofitted configuration as derived from
DCT and SCT. To the best of the author’s knowledge, such correlation has only been investigated
for small specimens through direct shear and tensile test for the characterisation of the brick-mortar
joint interface of unreinforced masonry by Sinha (1967), Samarasinghe (1980) and Page (1981), among
others. However, there are no references when it comes to retrofitted masonry.
6.2 Interpretation of testing methodologies
The evaluation of the shear load-bearing capacity of masonry is of paramount importance to assess
the seismic vulnerability of existing buildings and to design possible retrofitting solutions. In order to
identify the parameters that characterise the in-plane behaviour of masonry, both the SCT and the
DCT can be performed.
As aforementioned, diagonal compression tests could be carried out to determine the tensile
strength of masonry when the failure is produced by tensile diagonal cracking. However, the in-
terpretation of the test result is not unique and still involves some significant uncertainties. In the
available standards, such as ASTM and RILEM, the tensile strength of masonry ft is obtained by
assuming that the wall fails when the principal tensile stress σI attains its maximum value at the
centre of the panel. According to these standards, the stress field inside the panel, when subjected
to a diagonal compression, is approximated to pure shear and the principal stress directions coincide
with the two diagonals of the wall. Therefore, the tensile strength and the shear strength of the wall
are the same and can be computed using Equation (6.1), where P is peak load attained and An = Bt
is the transversal area of the panel in which B is the width of the wall and t its thickness.
P
τ = σI = 0.707 (6.1)
An
Frocht (1931) investigated theoretically and by means of photoelasticity the stress distribution in a
square plate subjected to diagonal compression. The author demonstrated that, under the hypothesis
of an elastic homogeneous isotropic continuum, the stress field is not uniform and the stress state at
the centre of the plate is not a pure shear, as shown in Figure 6.1-b. The author proved that axial
stress components are present along the horizontal and vertical directions. Therefore, the centre of
the Mohr’s circle is not in the origin, see Figure 6.1-c, and the principal stresses can be computed as
132
COMPRESSION TEST
follows:
P
σI = 0.52 (6.2)
An
P
σII = −1.68 (6.3)
An
Consequently, the stress state at the centre of the panel is characterised by the following stress
components in the global x-y coordinate system (see Figure 6.1 c):
P
σx = σy = −0.58 (6.4)
An
P
τxy = 1.05 (6.5)
An
This interpretation was validated by Brignola et al. (2008). The authors have reproduced the
problem studied by Frocht with the finite element method. The elastic linear solution of an isotropic
continuum loaded diagonally, under the hypothesis of plane stress, validated the stress state that
follows Equations (6.4) and (6.5). Consequently, the principal stresses were the ones presented in
Equations (6.2) and (6.3). In addition, Brignola et al. (2008) assessed the stress distribution in the
non-linear range, which did not significantly affect the value of σI computed by the elastic isotropic
solution. They also proposed a methodology for the evaluation of the tensile strength of masonry
under diagonal compression with Equation (6.6). In this equation, P is the load at failure attained in
the DCT, An is the transversal area of the wall and α is a coefficient dependent on masonry typology.
For the case of regular masonry, the authors recommended the value of 0.5, which is in line with the
findings of Frocht and is also the one recommended by the Italian building code NTC (2018).
P
ft = α (6.6)
An
The SCT allows the evaluation of the shear strength as the average shear stress in a panel subjected
to vertical compression and horizontal in-plane loading. The advantage of this test is that both the
boundary conditions and the loading method provide a stress state at the centre of the panel that
closely resembles the one experienced by the wall when subjected to seismic action, as shown in Figure
6.1-a.
To predict the strength of a masonry wall failing due to tensile diagonal cracking, which is one of
the most common failure modes, Turnsek and Cacovic (1971) proposed a model describing masonry
as an equivalent isotropic material. Despite the fact that masonry is far from being an isotropic
material, the model presents the advantage of characterising the in-plane behaviour of masonry by
only one parameter, the tensile strength of masonry ft . As a result, the shear strength of the wall is
evaluated with Equation (6.7).
133
COMPRESSION TEST
r
ft σ0
τmax = 1+ (6.7)
b ft
Where σ0 is the vertical compressive stress and b is a geometrical coefficient that takes into account
the distributions of the shear stress at the centre of the wall. Turnsek and Sheppard (1980) defined the
coefficient as dependent of the shape and geometry of the wall and computed it as the ratio between
the dimensions, height H and length L, of the wall (b=H/L).
Figure 6.1: a) Interpretation of SCT set-up according to Turnsek and Cacovic (1971), b) Interpretation of
DCT set-up by Frocht (1931), c) Mohr’s circle representation of DCT according to standards (dashed line),
according to Frocht’s theory (dash-dotted line) and according to the model proposed in the present research
(solid line). Each representation is associated with their corresponding coefficients for the computation of the
maximum and minimum principal stresses σI and σII and the maximum shear τmax
134
COMPRESSION TEST
6.3 Proposed integrated methodology
6.3.1 Correlation between Diagonal Compression Test and Shear Compression

Test
A research effort has been dedicated to better understand the in-plane behaviour of masonry by means
of the DCT and SCT configurations. Despite the relevance of the correlation between both tests, there
is a scarcity of available references regarding the correlation between them when it comes to assess
the in-plane behaviour of masonry walls.
Borri et al. (2015) studied the correlation between DCT and SCT for a total number of thirty-five
specimens, among which nineteen were manufactured in the laboratory in the unreinforced configura-
tion and sixteen were cut from six existing buildings. In total, fourteen specimens were tested under
DCT and the remaining twenty-one were subjected to SCT. The study aimed to validate the two test
methods and to discuss and compare the results in terms of shear strength for similar wall panels.
The shear strength from DCT, τ0D , was computed according to Equation (6.8) and the shear strength
from SCT, τ0T , was computed according to Equation (6.9), using the Turnsek and Cacovic (1971)
formulation.
ft P
τ0D = ft = 0.5 (6.8)
1.5 A
rn
ft T ft σ0
τ0T = τmax = = 1+ (6.9)
1.5 An b ft
where ft is the tensile strength of masonry, P is the peak load attained at DCT while T is the
peak horizontal load attained at SCT, An is the transversal cross-section of the wall and σ0 is the
vertical compressive stress experienced by the wall during the testing.
From the eight couples of specimens tested with DCT and SCT, the authors defined the ratio
r = τ0T /τ0D , which varied in a range from 1.13 to 3.10 depending on the masonry typology, which
included double-leaf rubble stone masonry, and double-leaf and single-leaf solid brick masonry. The
results evidenced not plausible correlation between the shear strength from both test configurations.
Ferretti et al. (2019) presented an extensive in-situ experimental campaign on rural masonries, on
which minor destructive tests (MDT), and destructive tests, such as DCT and SCT, were conducted.
The results gave the authors the opportunity of verifying the correspondence with the existing failure
criteria: Mohr-Coulomb, Turnšek and Čačovič, and Mann and Müller. From the results, it was
possible to calibrate the most suitable criterion to describe the shear behaviour of the different masonry
typologies investigated. The outcome was characterized by sliding failure mode due to poor quality
masonry and low compressive axial stress. As a result, the Mohr-Coulomb’s criterion was found to be
the most appropriate failure criteria to predict the shear behaviour of the masonry under study and
to verify the correspondence of the results obtained.
135
COMPRESSION TEST
Both researches, even though of relevance to the field, have focused their attention on unreinforced
masonry. As a result, no correlation between DCT and SCT has been investigated for masonry walls
strengthened with TRM strengthening systems.
In this context, this section pursuits the development of a methodology that can effectively be
used for the analysis of the shear parameters of masonry walls in both configurations, unreinforced
and retrofitted. The approach aims, to validate DCT as a simpler standardised test to evaluate the
tensile strength of masonry by providing an accurate value of the coefficient α. Another objective is to
compare the results of both test configurations, DCT and SCT, in terms of the predicted shear capacity
of masonry and correlate such parameter with the tensile strength of masonry. To this extent, the
results of the eight pairs of samples, including unreinforced (URM), repaired and retrofitted (UMR R),
and just retrofitted (LDB and LDS), tested under both configurations are analysed and compared.
Since all the samples tested under SCT mainly failed due to tensile diagonal cracking, the criterion
proposed by Turnsek and Cacovic (1971) was chosen to describe their in-plane behaviour. As afore-
mentioned, this formulation was developed under the hypothesis of masonry as a homogeneous and
isotropic material. In order to fit the retrofitted specimens into this criterion, strengthened masonry is
hereby considered as a complex material, comprising both clay brick masonry and the TRM strength-
ening system, working as an equivalent homogenous material that can be modelled by the Turnsek
and Cacovic (1971) formulation. Equation (6.7) can be used to assess the lateral capacity of the walls,
independently of their configuration -unreinforced or retrofitted- and to determine the property that
governs the in-plane behaviour of the walls, i.e. the tensile strength ft . It must be highlighted that in
the cases of retrofitted masonry the parameter ft represents an equivalent tensile strength that takes
into account the joint action of both the masonry wall and the TRM strengthening system. For the
sake of clarity, from now on ft would be considered as an equivalent tensile strength independently of
the specimens configuration.
This hypothesis is validated through comparison of the analytical results, obtained from the appli-
cation of Equation (6.7), with the experimental results from Chapter 5. For each specimen subjected
to SCT, the average experimental peak load Hmax , for both directions pushing and pulling involved
in the cyclic test, is computed from Table 5.2. As a result of the phenomenon experienced by spec-
imen LDS 1, which was previously described in Chapter 5, only the pushing direction is taken into
account for this analysis. Each peak load Hmax is then equated to the right side of Equation (6.7)
evaluated on the cross-section of the wall An as shown in Equation (6.10). It should be noted that
in this equation the parameter σ0 is adopted as the current level of compression (N ) experienced by
the specimen during the testing. By solving the non-linear problem, from Equation (6.10), it can be
obtained the equivalent tensile strength ft for each specimen tested under SCT. Table 6.1 includes
the results. Due to the similarity in value of ft yielded for some pairs of specimens, they are clustered
into two experimental groups comprising four specimens each. On one hand, specimens URM and
URM R and, on the other, strengthened specimens LDB and LDS. Strengthened specimen LDS-DL
are also included as an extra group consisting only of two specimens.
136
COMPRESSION TEST
Finally, the average value of ft that describes each group of specimens is determined using the
least square method to best fit the result of each specimen within the same group. The results are
included in Table 6.1, and Figure 6.2 shows the family curves for the ft obtained for each group. In
addition, to better identify them, Figure 6.2 contains the results of SCT, in terms of peak load and
the corresponding vertical load N at the moment of failure, represented as discrete points.
ft,SCT σ0
r
Hmax = τmax An = 1+ An (6.10)
b ft,SCT
Table 6.1 includes the coefficient of variation (C.o.V) for each group. This coefficient provides
an indication of the level of heterogeneity within each group. Among the first group, which includes
URM and URM R specimens, there is a greater variation as a result of relative scattering between the
URM individual specimens. This is to be expected due to the inherent heterogeneity of the component
materials of URM walls. In the remaining two groups the homogeneity is confirmed by the significantly
lower C.o.V, which can be attributed to the presence of the TRM strengthening system. From a general
perspective, it is clear that the specimens strengthened with TRM show smaller C.o.V. Consequently,
the hypothesis of assuming strengthened masonry as an equivalent homogeneous isotropic material
seems plausible, and the behaviour can be described by the Turnsek and Cacovic (1971) formulation
when the failure is due to diagonal cracking as a result of tensile strength exhaustion.
Figure 6.2 and Table 6.1 also show the results corresponding to the specimens strengthened with
two layers of LDS (LDS-DL). Even though this reinforcement configuration was not tested under DCT,
it is interesting to note that, as expected, ft increases with increasing reinforcement ratio.
Table 6.1: Assessment of the equivalent tensile strength ft,SCT obtained from the results of the SCT according
to Turnsek and Cacovic (1971). Computation of the average ft,SCT associated to each defined group and its
corresponding coefficient of variation (C.o.V)
Specimen Hmax [kN] N [kN] ft,SCT [MPa] Average ft,SCT [MPa] C.o.V
URM 1 172 185 0.26
URM 2 157 203 0.22
0.25 8.0%
URM1 R 178 215 0.26
URM2 R 181 238 0.25
LDB 1 210 247 0.31
LDB 2 221 272 0.31
0.31 2.9%
LDS 1 224 278 0.32
LDS 2 233 293 0.32
LDS-DL 1 255 331 0.35
0.36 3.6%
LDS-DL 2 266 340 0.37
The failure mode observed in the specimens tested both under DCT and SCT was due to tensile
diagonal cracking. Consequently, it can be stated that ft governs the behaviour of the masonry wall
under both testing configurations. As it was mentioned before, ft is an inherent property of the
masonry and hence the experimental response of both tests should yield similar value.
137
COMPRESSION TEST
Figure 6.2: Horizontal and vertical load at failure for each specimen obtained by SCT compared with ideal
curves governed by Equation (6.7) for different tensile strengths ft,SCT
The hypothesis that assumes strengthened masonry as an equivalent homogenous material for
extrapolating the Turnsek and Cacovic (1971) formulation can also be extended to the theory proposed
by Brignola et al. (2008). Equation (6.6) can be used to assess the parameter ft from the response
of DCT for each pair of specimens tested on both configurations, unreinforced and strengthened. To
achieve this, the left side of Equation (6.6) is set equal to the average value of ft obtained for the same
pair of specimens tested under SCT computed in Table 6.1. Equation (6.11) proposes a correlation
between the two testing configurations, DCT and SCT, where ft,SCT is the equivalent tensile strength
obtained from the SCT (see Table 6.1), PDCT is the peak load attained in the DCT (see Table 4.1),
by the same pair of specimens, and An is the cross-section of the wall. From this equation, it can be
obtained an accurate value of the coefficient α, as proposed by Brignola et al. (2008).
PDCT
ft,SCT = α (6.11)
An
Table 6.2 presents the computation of α, from Equation (6.11), for each pair of specimens and its
corresponding average according to the grouping observed in Figure 6.2. When applying Equation
(6.11) on the results of specimens URM and URM R, α yields an average value equal to 0.44. In turn,
when applying it to the specimens retrofitted with LDB and LDS, α yields a practically identical
average value, equal to 0.43. Interestingly, a previous research carried out by Turnsek and Sheppard
(1980) on four different types of masonry walls subjected to DCT yielded a value of α equal to 0.45
138
COMPRESSION TEST
when validating Equation (6.6). It is also consistent with the results reported by Segura (2020) based
on the finite elements study performed on the same clay brick masonry with low strength mortar
under diagonal compression. The coefficient α for this numerical study yielded a value equal to 0.40.
In addition, the value of α, assessed by Equation (6.11), is within the range proposed by Brignola
et al. (2008), which varies from 0.35 to 0.56 depending on different masonry typologies.
Table 6.2: Computation of coefficient α from DCT and SCT experimental results following Equation (6.2)
Specimen PDCT [kN] ft,SCT [MPa] α

URM 184* 0.24 0.50
0.44
URM R 256 0.25 0.39
LDB 295 0.31 0.41
0.43
LDS 279 0.32 0.45
Average 0.43
C.o.V 10.1%
* Average considering the specimens tested in Segura (2020) and failing in the same fashion.
Table 6.3 shows the results of ft obtained from each test configuration. The parameter reports
in its label the acronym referring to the tests from which it was computed, i.e. ft,DCT and ft,SCT .
Equation (6.6) is used for the computation of ft,DCT imposing the value of α equal to 0.43 (C.o.V
=10.1%) previously deduced and shown in Table 6.2. The values of the equivalent tensile strength
evidenced some scattering among tests, mainly in the URM configuration, which can be fully explained
by the variability of the materials. However, the retrofitted configurations, including the repaired ones,
evidenced more homogeneity. This asserts the validity of extrapolating the formulation proposed for
URM walls by Frocht (1931), Turnsek and Cacovic (1971), and Brignola et al. (2008) for strengthened
masonry walls, under the hypothesis of an equivalent homogeneous material.
Table 6.3: Comparison between the tensile strength obtained with the SCT results ft,SCT and with the DCT
results ft,DCT
ft,DCT −ft,SCT
Specimen ft,SCT [MPa] ft,DCT [MPa] ∆= ft,SCT ft,SCT /ft,DCT
URM 0.24 0.20 -17% 1.20

URM R 0.25 0.27 6% 0.94
LDB 0.31 0.31 0% 1.00
LDS 0.32 0.31 -5% 1.05
LDS-DL 0.36 - - -
Average 1.05
C.o.V 11%
Under the hypothesis that strengthened masonry with TRM strengthening systems can be consid-
ered as an equivalent homogeneous material, the correlation between both test configurations can be
139
COMPRESSION TEST
analysed by means of a graphical representation. The reference point, as defined by Calderini et al.
(2010), is established at the centre of the wall to evaluate the stress state of the samples on both
testing configurations.
The graphical interpretation of the behaviour of the reference point, requires the drawing of Mohr’s
circles representing the stress state when the failure of the masonry specimen occurs. The Mohr’s circle
for each type of test was drawn on the σ − τ plane, as shown in Figure 6.1. The stress state of DCT
was defined considering the value of α obtained from the expression 6.11 correlating DCT and SCT to
assess the principal tensile stress σI . Considering the ratio σI /σII = −0.3, as demonstrated by Frocht
(1931) and later by Brignola et al. (2008) for loading direction at 45°, the radius of the Mohr’s circle
can be computed along with other meaningful stress states, such as the maximum shear strength τmax .
Equation (6.12) and (6.13) enable the computation of σI and τmax for the Mohr’s circle corresponding
to DCT. If we consider the aforementioned failure criterion of the critical tensile stress, σI can be
assumed equal to ft .
P
σI = 0.43 (6.12)
An
P
τmax = 0.93 (6.13)
An
The Mohr’s circle of SCT is drawn considering the shear stress and the normal stress at failure,
acting on the cross section of the specimen. The stress state, evidenced in Figure 6.1-a, is evaluated
as τ = H/An and σ = V /An following Turnsek and Sheppard (1980), Magenes and Calvi (1997),
Calderini et al. (2010) and Ferretti et al. (2019).
Figure 6.3 shows the graphical representations of the stress state of the masonry wall subjected
to DCT and SCT. The representation differs on the Mohr’s plane in which the tests are analysed. In
Figure 6.3 the stress plane for DCT is represented as a solid line while the stress plane for SCT is
a dashed line. Another aim of the methodology is to defined a shear parameter τ0,m to characterise
the in-plane behaviour of the masonry wall in any configuration -unreinforced and strengthened-
independently from the compressive level, which in the case of SCT depends on the type of control
chosen for the set-up. The parameter τ0,m is defined as the value at which the Mohr’s circle intersects
the shear strength axis (τ ). In the case of SCT τ0,m,SCT , the value corresponds to the stress state
observed during the testing. But in the case of DCT, the stress state τ0,m,DCT needs to be graphically
obtained from the circle. As a result, the stress state of the reference point experiences a rotation and
consequently the angle between the mortar joints and the applied load that generates the new stress
state also changes.
It is well known that masonry exhibits distinct directional properties due to the influence of mortar
joints. Depending upon the orientation of the joints to the applied stress, the failure can occur in
the joints alone or in some form of combined mechanism involving the mortar and the masonry units.
140
COMPRESSION TEST
Figure 6.3: Comparison of stress state at the centre of the wall, being θ the rotation angle between two
different failure planes and θp the rotation angle that defines the orientation of the principal planes: a) Mohr’s
Circle for DCT, stress plane shown in solid line, b) Mohr’s Circle for SCT, stress plane shown in dashed line.
141
COMPRESSION TEST
Samarasinghe (1980) studied the influence of the orientation of bed mortar joint on the brickwork
strength in a state of biaxial stress by testing masonry squared samples with different bed joint lay-up
angles. The lay-up angles, defined as the angle between the direction of the bed mortar joints and the
horizontal plane, were set equal to 0°, 22.5°, 45° and 67.5°. The experimental results concluded that
for angles equal to or less than 45° the failure was initiated at head joints and the crack propagated
through brick units and bed joints, while for angles greater than 45°, the failure was characterized by
bed joint sliding.
The rotation angle 2θ between τmax (Equation (6.13)) and τ0,m,DCT can be assessed by Equation
(6.14). If the rotation angle 2θ is sufficiently small, then the joint sliding failure is precluded and
at the same time its influence on the ft of the samples is expected to be small. Hence, the stress
state τ0,m can be evaluated on both Mohr’s circles. In the case of DCT, Equation (6.15) enables
the computation of τ0,m,DCT . Table 6.4 compares the value of τ0,m,SCT and τ0,m,SCT and Figure 6.4
shows the remarkable agreement between the parameters computed with the results of the two testing
methodologies.

τmax − σI
2θ = arcsin (6.14)
τmax
τ0,m,DCT = τmax cos(2θ) (6.15)
The above procedure has shown its capability in combining the results from the two testing method-
ologies to validate DCT for computing ft by providing a more accurate value of α for a masonry typol-
ogy featured by handmade clay bricks units and low strength lime mortar joints. This is a recurrent
typology in the historical building stock and deeper insight on the evaluation of ft is still missing.
The approach has also included the graphical interpretation of both testing methodologies that al-
lowed the obtention of a shear parameter τ0,m which is independent from the testing configuration, in
terms of boundary constrains and loading conditions. The good agreement between the values of the
parameters ft and τ0,m obtained from both configurations, validates the proposed methodology (see
Figure 6.4).
Table 6.4: Comparison between the values τ0,m,SCT and τ0,m,DCT yielded from the proposed methodology for
the correlation between SCT and DCT
τ0,m,DCT −τ0,m,SCT
Specimen τ0,m,SCT [MPa] τ0,m,DCT [MPa] ∆= τ0,m,SCT
URM 0.42 0.36 -14%

URM R 0.44 0.49 -11%
LDB 0.54 0.56 5%
LDS 0.58 0.56 -3%
142
COMPRESSION TEST
Figure 6.4: Correlation between the results obtained from the proposed methodology: a) Comparison between
ft,SCT and ft,DCT for all the specimens tested in both unreinforced and strengthened configurations, b) com-
parison of the shear parameter τ0,m,SCT and τ0,m,DCT for all the specimens tested in both unreinforced and
strengthened configurations
6.3.2 Correlation between shear response and tensile strength
From the results obtained in the previous section, it can be concluded that the in-plane behaviour
of masonry, unreinforced or strengthened, failing due to tensile diagonal cracking is described by
the equivalent tensile strength ft . In addition, it was identified the shear parameter τ0,m , which
characterises the shear of masonry independently from the boundary and loading conditions of the
testing configurations. This section presents the study on the correlation between both parameters ft
and τ0,m . The aim of the correlation is to provide an expression that can be used to estimate τ0,m ,
for unreinforced and strengthened brick masonry, from the equivalent tensile strength computed with
experimental results of DCT.
Several authors have attempted to develop empirical formulations to describe possible relationships
between the shear and tensile strength of masonry. Due to the impractical manipulation of large
specimens such relationship has been only provided for the mortar joint-brick interface. Among other
authors, Sinha (1967) and Samarasinghe (1980) have studied, through shear and direct tensile tests on
couplet specimens, the effect of the tensile bond strength ftk on the shear bond strength of masonry
bed joints τ0 . It must be noted that this value refers specifically to the shear strength of the mortar
joint and must not be confuse with the average shear strength of the masonry. The experimental results
from Sinha (1967) suggested the relationship following Equation (6.16). The research developed by
Page (1981), suggested another simple empirical relationship following Equation (6.17), in which β is
a coefficient of proportionality that varies between 1.7 to 2.3 depending on the masonry under study.
143
COMPRESSION TEST
0.56
τ0 = 0.8ftk (6.16)
τ0 = βftk (6.17)
Similar to the relationship proposed for the interface between bricks and mortar joints, a correlation
study has been carried out between the parameters obtained in the previous section τ0,m and ft . Table
6.5 and 6.6 present the values of τ0,m,DCT and ft,DCT , and τ0,m,SCT and ft,SCT which were computed
from both testing methodologies. To show the robustness of the methodology proposed in the previous
section, for the correlation between DCT and SCT, to compute ft and τ0,m , Table 6.5 also shows the
results for specimens MDS and LDB-JR tested only under DCT configuration and Table 6.6 shows
specimen LDS-DL only tested under SCT configuration.
Table 6.5: Equivalent tensile strength ft,DCT and the shear parameter τ0,m,DCT of specimens tested under
DCT configuration, and the ratio between both parameters
Specimen τ0,m,DCT [MPa] ft,DCT [MPa] τ0,m,DCT /ft,DCT [MPa]

URM 0.36 0.20 1.82
URM R 0.49 0.27 1.82
LDB 0.56 0.31 1.82
LDS 0.56 0.31 1.82
LDS-DL† - - -
MDS * 0.46 0.25 1.82
LDB-JR * 0.38 0.21 1.82
Average 1.82
C.o.V 0.0%
† Specimen tested only in the SCT configuration
* Specimen tested only in the DCT configuration
Table 6.6: Equivalent tensile strength ft,SCT and the shear parameter τ0,m,SCT of specimens tested under
SCT configuration, and the ratio between both parameters
Specimen τ0,m,SCT [MPa] ft,SCT [MPa] τ0,m,SCT /ft,SCT [MPa]

URM 0.42 0.24 1.76
URM R 0.44 0.25 1.74
LDB 0.54 0.31 1.73
LDS 0.58 0.32 1.80
LDS-DL† 0.67 0.36 1.85
MDS * - - -
LDB-JR * - - -
Average 1.78
C.o.V 2.7%
† Specimen tested only in the SCT configuration
* Specimen tested only in the DCT configuration
144
COMPRESSION TEST
Figure 6.5 evidenced the strong linear correlation between the two parameters, ft and τ0,m from
both testing methodologies, with a coefficient of determination R2 of 0.99. Such correlation conse-
quently confirms that the in-plane behaviour of masonry, failing due to diagonal cracking, is controlled
by its tensile strength, meaning that τ0,m can be described as the ratio to ft . Table 6.5 and 6.6 also
include the ratio between both parameters. The ratios τ0,m /ft yielded for each testing methodology
show only a small scattering and is consistent with the range of values suggested by Page (1981), which
varied form 1.7 to 2.3. However, those results were obtained from small specimens and according to
Crisafulli (1997) may not be representative of the real shear strength of masonry walls.
As a result of a simpler linear regression, and considering the ratios yielded from both testing
methodologies, the following experimental expression is proposed to correlate both parameters inde-
pendently from the testing configuration.
τ0,m = 1.80ft (6.18)
Despite the simplistic nature of the proposed expression the measured data are in very good
agreement with its prediction as shown in Figure 6.6.
Figure 6.5: Linear correlation between ft and τ0,m for all specimens tested on both testing methodologies
In the absence of experimental evidenced obtained through standardised tests, such as DCT, the
Italian building code NTC (2018) suggests the use of the expression from Equation (6.19) for the
determination of the tensile strength of masonry, where τ0 is defined as the shear strength at zero
axial compressive level and for which the building code provides a range of referential values depending
145
COMPRESSION TEST
Figure 6.6: Comparison between the proposed expression for the correlation between ft and τ0,m for all
specimens tested: a) Results from DCT testing methodology, b) Results from SCT testing methodology
146
COMPRESSION TEST
on the masonry typology (see Table C.8.5.I Circ. 21 gennaio 2019).
ft = 1.5τ0 (6.19)
As aforementioned, following this recommendation Borri et al. (2015) carried out a correlation
study between DCT and SCT and computed the shear stress from the results of DCT, τ0D and SCT,
τ0T . In all the cases studied the ratio τ0T /τ0D was not only greater than 1 but close to 2 or 3 in some
cases. It can be seen that the proposed correlation from Equation (6.19) between shear and tensile
strength has led to an underestimation of the shear capacity of the masonry mainly when assessed from
the DCT results. The underestimation is based on the fact that the shear capacity computed in Borri
et al. (2015) is taken as a fraction of the tensile strength, as evidenced from Equation (6.19). From the
experimental results of the present research, it is clearly evident that the tensile strength always yields
lower values than the shear capacity. The expression proposed in Equation (6.18), however, deals
with two parameters yielded from the combination of the test results from two testing methodologies
without previous correlation assumptions. The only initial hypothesis is the consideration of masonry,
unreinforced or strengthened, as a complex but equivalent homogeneous material failing due to tensile
diagonal cracking.
Finally, the proposed simple empirical expression in Equation (6.18) provides a useful tool for
describing the correlation between ft and τ0,m . Such correlations can be applied to describe the
response of URM or strengthened masonry subjected to in plane actions and failing due to tensile
diagonal cracking. Therefore, a simpler test, such as DCT, can be performed to study the in-plane
response of strengthened masonry.
6.4 Procedure for the application of the proposed methodology
Based on the experimental methodology proposed and describe in Section 6.3.1 and 6.3.2, a summary of
the recommended analysis procedure for obtaining reliable estimates of the tensile and shear strength
for masonry strengthened with TRM systems, is presented.
The first step of the analysis process involves the computation of ft from the results of the available
tests, i.e. SCT or DCT. If both tests were conducted, ft is computed following the procedure described
in Section 6.3.1. The subsequent analysis also provides the identification of a more accurate value of
the coefficient α for the investigated masonry typology. If only one test was conducted, the value of
ft is either assessed from Equation (6.10) in the case of having the results of SCT or from Equation
(6.6) in the case of having the results of DCT. If the latter is the case, the value of α can be adopted
from the range suggested by Brignola et al. (2008) according to the masonry typology, in which the
value, hereby yielded for double-leaf masonry typology consisting of clay bricks and low strength
mortar joints, should be also be taken into consideration. Once this value of ft is obtained, a good
estimation of the shear in-plane capacity can be provided by applying the empirical correlation of
Equation (6.18). For unreinforced masonry, one of the advantages of assessing τ0,m is that the value
147
COMPRESSION TEST
may be taken as the shear contribution of the masonry, Vm , for the design of possible strengthening
solutions, since the value is not biased by any imposed compressive level. In the case of strengthened
masonry, τ0,m represent the first estimation of its shear capacity for the verification of limit states of
masonry structures for in-plane actions, such as wind or seismic action.
The proposed methodology can be applied to strengthened masonry as well as unreinforced one
for the estimation of the equivalent tensile and shear strength. However, in the case of URM, the
methodology only applies if the masonry wall is prone to experience failure due to tensile diagonal
cracking, rather than shear sliding or crushing.
6.5 Conclusions
The present research has proposed a robust procedure based on the synergy between two testing
methodologies, Diagonal Compression Test (DCT) and Shear Compression Test (SCT), for the de-
termination of the equivalent tensile strength that describes the in-plane behaviour of historical clay
brick masonry in the unreinforced configuration and strengthened with TRM systems, failing due to
tensile diagonal cracking. The main hypothesis assumed for the assessment is the definition of unrein-
forced or strengthened masonry as an equivalent isotropic homogeneous material. Moreover, in order
to derive meaningful ranges of results, the methodology has been applied to masonry strengthened
with different TRM strengthening systems.
On the basis of the results obtained from the correlation analysis carried out, the following con-
clusion can be drawn:
The same failure mechanism was experienced by the specimens tested under both testing con-
figurations, i.e. tensile diagonal cracking. The failure is accurately described by the simplified
model proposed by Turnsek and Cacovic (1971), governed by only one parameter, in this case the
equivalent tensile strength ft , featuring the behaviour of an equivalent isotropic homogeneous
material. Therefore, the method discussed in this chapter cannot be applied to URM structures
failing in a different fashion.
It has been confirmed that the presence of TRM strengthening systems reduces the inherent
heterogeneity of the component materials of URM, by reducing the coefficient of variation of the
parameters assessed from the experimental outcomes. Consequently, the formulations developed
for isotropic homogeneous material such as Frocht (1931), Turnsek and Cacovic (1971), and
Brignola et al. (2008) can be extrapolated to assess the in-plane behaviour of strengthened
masonry.
The proposed methodology validates the DCT as a simpler standardised test to evaluate the
equivalent tensile strength of masonry walls in both configuration, unreinforced and strength-
ened. In addition, it provides a simple procedure to accurately assess the value of the coefficient
α, as proposed by Brignola et al. (2008). For the material combination hereby investigated, i.e.
148
COMPRESSION TEST
double-leaf historical clay brick masonry with low strength mortar joints, α yielded a value equal
to 0.43.
The proposed methodology allows the definition of a shear parameter τ0,m to characterise the
in-plane behaviour of masonry walls. The novelty of the parameter is on its independence from
the loading and boundary conditions imposed by the testing configurations. This feature makes
τ0,m a first unbiased estimation of the in-plane capacity of the masonry wall, which can be proved
useful for design purposes since it allows the assessment of strengthening solutions.
One of the most useful applications of the proposed methodology, is the provision of the lin-
ear correlation between τ0,m and ft . Such correlation confirms that the in-plane behaviour
of strengthened masonry is defined by the equivalent tensile strength provided by the TRM
strengthening system. Moreover, the correlation shows that the analytical formulation from
Equation (6.19) recommended by the Italian Building Code for the estimation of the shear
strength cannot be applied to interpret DCT or SCT results. This correlation leads to an
underestimation of the shear capacity of the masonry, since the two parameters acting in the
correlation address two different failure mechanisms.
Finally, one of the most important implications of this work is that provides experimental insight
to better understand the interpretation of the stress state of masonry walls when subjected to
DCT, validating the formulation proposed by Brignola et al. (2008)
149
Chapter 7
Analytical Study
7.1 Introduction
In the last decade, numerous techniques have been developed in order to rehabilitate and strengthen
URM structures. Along with this development, the research on innovative materials to be used in
strengthening solutions has also increased. However, the update of the building codes, such as EC6
(2006) or NTC (2018), is not up to speed with these fast developments and still requires additional
provisions for the design of strengthening in masonry structures. Among the available guidelines
providing guidance to practitioners there are the Italian ones for FRP design called CNR-DT 200
(2013), frequently used until the development of the new guideline CNR-DT 215 (2018) based on the
latest research on TRM, as well as the American guideline AC 549 (2013).
According to the mentioned guidelines, the nominal shear capacity of strengthened masonry Vt,R
is evaluated as the sum of the contribution of the unreinforced masonry Vm , and the contribution of
the reinforcement, Vt,f , as shown in Equation (7.1).
Vt,R = Vm + Vt,f (7.1)
The main aim of this chapter is to investigate the validity of the formulations proposed by the
guidelines for the estimation of the in-plane shear capacity of the TRM Vt,f to the global capacity
of the strengthened masonry. The following sections explain the calculation methods proposed by
each guideline to assess Vt,f . The experimental results are later compared with the results of the
analytical formulations to evaluate not only the accuracy of the formulations but also the efficiency
of the different types of strengthening systems investigated in this research. In addition, according to
the failure mode experimentally observed, an insight on the parameters of the formulation is provided
in order to determine the influence on the efficiency of the strengthening solution.
151
CHAPTER 7. ANALYTICAL STUDY
7.2 Analytical Models
7.2.1 Available TRM guidelines
The American guidelines AC549 (2013) were pioneer in the provision of instructions for both testing
and design of FRCM/TRM strengthening solutions. The guideline bases the TRM design on the limit-
state principles and assumes a bilinear behaviour of the TRM composite in uniaxial tension. The first
branch of the bilinear stress-strain diagram is associated with an initial elastic stage up to the cracking
of the matrix, whereas the second branch simulates the cracked behaviour of the composite until the
failure of the textile. The design tensile strength ff v , evaluated with Equation (7.2), is computed
assuming Ef as the tensile Young’s modulus of the fibre, which also represents the Young’s modulus
of the cracked TRM. A main drawback of this capacity model is that the design tensile strain of the
TRM εf v , which is experimentally obtained from the direct tensile test, is restricted to 0.004, to avoid
large cracks in the matrix. The ultimate tensile strain of TRM εu,f , defined in Table 3.5, ranges
between 0.015 to 0.020, hence, limiting εf v to 0.004 by using only 20% of the strain capacity of the
TRM for its design. The shear strength contribution of the TRM is computed according to Equation
(7.3), where Af is the area of the mesh reinforcement by unit width, n is the number of layers of mesh
reinforcement and L is the length of wall in the direction of the applied shear force.
ff v = Ef εf v (7.2)
Vf = 2nAf Lff v (7.3)
In parallel with the development of AC 549 (2013), the Italian guideline CNR-DT 200 (2013) was
updated to address the structural application of composite materials such as FRP. The guideline, was
later taken as the starting point for the development of a new guideline focused on a strengthening
solution much more compatible with masonry substrates, CNR-DT 215 (2018). This new guideline
was oriented towards the application of TRM to rehabilitation purposes. Moreover, CNR-DT 215
(2018) provided guidance for performing standardised experimental tests to characterise the mechan-
ical properties of the composite, including the bond strength between TRM and the substrate. Until
this point, the American document AC 434 (2017) had been the only one providing recommendations
to perform direct tensile tests for TRM characterisation. The development of standardised testing
improved the reliability of strengthening solutions with TRM since it was possible to control the effec-
tiveness of the solutions promoting their safe use in the rehabilitation upgrading of existing structures
(De Santis et al. (2017a), de Felice et al. (2020)).
Despite the advances brought by CNR-DT 215 (2018) in the field of material characterisation, the
formulation that the guidelines provides for the assessment of the shear capacity of the strengthening
solution is, unlike CNR-DT 200 (2013), mostly oriented to full surface coverage TRM solutions. CNR-
DT 215 (2018) disregards the assessment of strip configuration, which is much more common for FRP
152
solutions and was profusely used in the present research. For this reason, the experimental results
hereby exposed are evaluated and validated according to the formulation provided by CNR-DT 200
(2013). According to this standard, the shear contribution of the TRM strengthening solutions is
computed following to Equation (7.4) and (7.5).
1 bf
Vt,f = · 0.6 · d · 2 · tf · εf d · Ef εf d · Ef = ff d (7.4)
γRd pf
εf k
εf d = min{ηa · , εf dd } (7.5)
γf
The following notation is used in the previous equations: γRd is the partial safety factor, d is the
length of the wall in the direction of the applied shear force, tf is the equivalent thickness of the
reinforcement parallel to the applied shear force, εf d is the strain corresponding to the reinforcement
tensile capacity, Ef is the Young’s modulus of the textile, ff d is the reinforcement tensile capacity,
εf k represents the reinforcement strain at failure, εf dd is the maximum strain at which the debonding
takes place and bf and pf are the width and centre-to-centre spacing of the TRM strips, respectively.
In the case of full surface coverage the ratio bf /pf is equal to 1. The conversion factor ηa and the
partial factor γf are defined from CNR-DT 200 (2013).
The design reinforcement strain εf d is chosen according to the type of failure observed in the
experimental response of the specimens. According to CNR-DT 215 (2018) the failure could take
place either in the intermediate area of the specimen or experienced end debonding failure caused by
the lack of enough bonded length. Depending on the type of failure εf d is multiplied by a coefficient
βf d equal to 1.5 in the case of failure in the intermediate area and equal to 1 in the case of end
debonding failure. Similarly, the tensile capacity ff d takes the value equal to σu,f when the failure
of the reinforcement is due to fibre rupture in tension, and the value σsl,t when the failure is due to
debonding (see Table 3.5).
With the aim of performing a comparison with the experimental results, the partial safety factor
γRd is equated to 1.
7.2.2 Analytical model for NSM joint repointing
The joint repointing technique with Near Surface Mounted (NSM) rebars is a relatively new solution
for strengthening and there is not yet enough insight to describe its mechanical behaviour. The
efficiency of the technique has been proven by Li et al. (2005), Ismail et al. (2011) and Casacci et al.
(2019), but an extensive laboratory characterisation it is still required to assess existing analytical
procedures and to address new design provisions.
Li et al. (2005) developed an analytical model to predict the shear contribution of NSM comprising
embedded FRP bar in the mortar joints of masonry concrete walls. Later, Casacci et al. (2019) applied
this approach to predict the shear contribution of NSM basalt bars on small masonry specimens made
of fire-clay bricks and lime based mortar. The results evidenced the effectiveness of the NSM technique
153
to increase the shear capacity of the masonry, as well as the correctness of the analytical model to
predict the shear capacity observed.
The analytical model hereby used to determine the contribution of helical stainless steel rebars is
presented in Li et al. (2005). The analytical model is based on several hypotheses: first, it is assumed
a constant shear crack inclination angle of 45 degrees, second, each bar intersected by the crack is
divided into two parts at the two sides of the crack. Therefore, the shear contribution provided by
the helical rebars Vt,N SM is computed according to Equation (7.6) as the sum of the forces resisted
by the bars intersecting the diagonal crack.
Neither debonding of the helical rebars from the mortar nor tensile failure of the rebars were
observed during the test of the present experimental programme, hence, the capacity of the NSM
reinforcement is considered to be limited by the bond failure between the mortar matrix used for joint
repointing and the masonry. Therefore, the force carried by each rebar is equal to the product of the
bond strength τb , the bonded area Ab between the masonry and the mortar of the joint.
n
X n
X
Vt,N SM = Ab ft,i = (2D + tm )τb Li , Li ≤ Le (7.6)
i=1 i=1
fu Af
Le = (7.7)
(2D + tm )τb
The following notation is used in the previous equations: tm and D are the mortar thickness and
the depth of the groove, respectively, Le is the effective length defined, by Equation (7.7), as the
minimum length at which the maximum stress of the helical stainless steel rebar can be achieved, fu
and Af are the maximum tensile strength of the helical rebar and its cross-section area and finally
τb is the bond strength assumed to be uniform along the Le . Li is the effective bond length of the
i-th bar intersecting the diagonal crack, see Figure 7.1, and is always the shortest part of the rebar
intersected by the diagonal crack.
In the absence of experimental evidence τb might be estimated as 3% of the compressive strength
of the mortar matrix. This value was suggested by Paulay and Priestley (1992) and Li et al. (2005)
to assess the bond strength of the binding mortar and can be extended to the mortar matrix.
7.3 Analytical Results
7.3.1 Shear contribution of masonry
The prediction of the shear capacity of an unreinforced masonry wall (URM) is related to the failure
mechanism experienced when subjected to an in-plane action. As mentioned before, there are four
types of failure mechanisms that were identified with an extensive number of experimental programmes.
(a) Sliding along a straight crack at horizontal bed joints. A particular case of this type of failure is
stepped cracks, alternating from head joint to bed joint, associated mainly to URM walls with
154
Figure 7.1: a) Distribution of the stress along the helical rebar embedded in the mortar joints, b) crack pattern
at the end of the test of a specimen retrofitted with NSM bars. Li is the effective bond length of the i-th bar
intersecting the diagonal crack
poor mortar joints.
(b) Diagonal tension cracks through the mortar and masonry units.
(c) Flexural failure due to rocking of the wall.
(d) Crushing of the compressed zone of the masonry wall.
The present experimental campaign evidenced mainly tensile diagonal cracking and crushing of
masonry due to high level of strengthening.
The EC6 (2006) and EC8 (2004) propose the formula to calculate the shear strength of masonry
based on the Mohr-Coulomb failure criterion. Both the Italian standard NTC 2018 (2018) and the
American Standard ASCE 41-17 (2017) define the shear capacity in accordance with Equation (7.8).
Evidently, both standards follow the model proposed by Turnsek and Cacovic (1971), where An is the
net area of the wall, ft is the tensile strength of the masonry, σ0 is the design axial compression stress
and b is a shape coefficient depending on the geometry of the wall, height h and length l, and equal
to h/l which ranges between 0.67 and 1.5.
r
ft σ0
Vs,dt = An 1+ (7.8)
b ft
For the present experimental programme, σ0 was set equal to 0.30 MPa in Equation (7.8), namely
the design axial compressive stress (refer to Chapter 3), and ft was set equal to 0.24 MPa. This latter
value was derived from the comparative analysis performed in Chapter 6. As a result, Vs,dt yields a
value equal to 142 kN.
155
From the correlation analysis between DCT and SCT, presented in Chapter 6 Section 6.3.1, the
parameter τ0,m,Av for the masonry panel in the unreinforced configuration has been assessed. When
this parameter is evaluated on the cross-section of the wall An , the shear contribution of unreinforced
masonry can be computed as Vm = τ0,m,Av An yielding a value equal to 140 kN. Such value is in good
agreement with prediction Vs,dt obtained by analytical formulations. For the sake of just assessing
the validity of the analytical formulation to predict the shear contribution of the TRM strengthening
system, the value of Vm , in Equation (7.1) is taken equal to the capacity experimentally obtained,
namely 140 kN.
In addition to the tensile diagonal failure, another failure mechanism was observed in the exper-
imental response of retrofitted specimens, i.e. masonry crushing due to compression. According to
the available standards, EC8 (2004) and ASCE 41-17 (2017), the capacity of URM wall controlled by
flexure under axial load N can be assessed with Equation (7.9).
LN N
Vc = (1 − 1.15 ) (7.9)
2h0 Ltfc
where L is the in-plane horizontal dimension of the wall, h0 is the distance between the section
where the flexural capacity is attained and the contra-flexure point, t is the thickness of the wall and
fc is the compressive strength of masonry.
There are two specimens, among the ten tested, that failed due to toe-crushing as a consequence
of excessive strengthening, namely LDS-DL specimens. To evaluate Equation (7.9), N was set equal
to 355 kN. Such value was recorded during the testing at the moment in which the peak load was
attained, h0 was set equal to 0.5 h, and fc was set equal to 6.50 MPa. As a result, Vc yielded a value
equal to 285 kN.
7.3.2 Shear contribution of TRM
The experimental results have been compared with the analytical formulations for the prediction of
the shear capacity of reinforced walls. Table 7.1 and 7.2 include the experimental values obtained
from the DCT and SCT for the reinforced and unreinforced walls (VEXP and Vm respectively) and the
analytical values Vt,f computed using Equation (7.4). The error is evaluated between the analytical
value Vt,f and the experimental peak force sustained by the TRM strengthening system, Vt,f,EXP . The
latter value is computed as the difference between the average shear force reached by the strengthened
specimens, VEXP , and the average shear force carried by the URM specimens, Vm .
The readings of the LVDTs, located on the centre of the walls during SCT, were analysed with the
aim of asserting the chosen values of ff d , which were obtained from the mechanical characterisation
study presented in Chapter 3, Section 3.2.3-Table 3.5. LVDT readings were divided by their reference
length, measured before the test, and averaged to obtained the strain in tension of the different TRM
strengthening solutions. In some cases, due to anomalous readings resulting from sensor detachment,
a LVDT was omitted. For the specimens tested only under the DCT configuration, namely MDS and
156
Table 7.1: Comparison between the experimental values obtained from DCT and the analytical values com-
puted according to the Italian guideline CNR-DT 200 (2013)
ID VEXP Vm Vt,f,EXP Geometry [mm] ff d Vt,f Error

Specimen [kN] [kN] [kN] d tf bf pf [MPa] [kN] [%]
LDB 295 140 155 1270 0.032 - - 1700.00 117 -24
LDB R 256 140 116 1270 0.032 - - 1700.00 117 0.6
LDS 279 140 139 1270 0.084 100 250 2096.00 152 9.5
MDS 227 140 87 1270 0.169 100 250 687.00 100 14
Table 7.2: Comparison between the experimental values obtained from SCT and the analytical values computed
according to the Italian guideline CNR-DT 200 (2013)

LDB 216 140 76 1270 0.032 - - 1700.00 83 9.7
LDB R 180 140 40 1270 0.032 - - 1700.00 83 110
LDS 229 140 89 1270 0.084 100 250 2096.00 107 21
LDB-JR, the same procedure was followed with the measurements recorded by the LVDTs located in
the centre of the wall.
Figure 7.2-a,c,e show the experimental stress-strain curve of LDB-TRM and LDS-TRM from the
cracking point of masonry onwards. At this stage, the masonry substrate begins to transfer the carried
load to the TRM strengthening solution until TRM finally fails either due to rupture or debonding.
Specimens of both types of strengthening solutions, LDB and LDS, evidenced similar tensile response.
However, in the specimens just retrofitted with LDB-TRM, the composite experienced textile failure
before reaching the expected tensile strength of the bare textile σu,f . The result may be consequence
of a non-uniform loading distribution as the cracking developed during testing which might hinder the
attainment of σu,f and consequently the yielded capacity was lower than the one expected. Conversely,
both LDB-TRM applied on repaired specimens and LDS-TRM reached the stress expected. In the
first case the composite attained σu,f while in the second case LDS-TRM experienced delamination
within the matrix followed by debonding from the substrate without evidencing textile rupture; there-
fore, LDS-TRM attained the ff d obtained by experimental testing with single-lap shear bond test as
described in Chapter 3.
Figure 7.3 shows the stress-strain experimental curve of MDS-TRM until the textile experienced
delamination after the peak load was attained for DCT. It is of interest to note that the stress recorded
is slightly higher than the one chosen as ff d obtained from the characterisation of the material (refer
to Chapter 3), which could be consequence of the localized LVDT reading.
Table 7.1 and 7.2 compute the error between the analytical predictions and the experimental
157
Figure 7.2: Results from SCT: a) Stress-Strain experimental curve of specimen repaired and retrofitted with
LDB-TRM, b) Stress-Displacement response of specimen repaired and retrofitted with LDB-TRM, c) Stress-
Strain experimental curve of specimen just retrofitted with LDB-TRM, d) Stress-Displacement response of
specimen just retrofitted with LDB-TRM, e) Stress-Strain experimental curve of specimen retrofitted with
LDS-TRM, f) Stress-Displacement response of specimen retrofitted with LDS-TRM
158
Figure 7.3: Stress-Strain experimental curve of MDS-TRM under DCT configuration
results from DCT and SCT, respectively. Even though the error in the specimens just retrofitted
with LDB-TRM is not large, it can be easily explained since is pertained to the textile failure at a
lower stress. However, the results of the analytical solution for the specimens repaired and retrofitted
with LDB-TRM greatly differ from one test to the other. The analytical prediction for DCT is highly
accurate, with an error beneath 1%, while for SCT is over 100%. As evidenced by Figure 7.2-a, the
LDB textile was efficient on providing tensile strength to increase the in-plane capacity of masonry,
hence, the focus should be shifted towards the contribution of masonry to explain the scattering. The
URM specimens of SCT were more severely damaged when compared to the ones of DCT. Therefore,
the repairing was more extensive in the first case. As a result of the restoration, the URM R specimens
of SCT had more weakened interfaces that may have led to a lower in-plane capacity of the URM.
Figure 7.2-b and d display, for specimens repaired and retrofitted and just retrofitted, the evolution
of the stress on the TRM with the increasing displacement amplitudes applied during the SCT.
The response differs between the repaired walls and the just retrofitted ones mainly because TRM
deformations are registered on early stages of the test, and hence, the masonry transferred the tensile
stress to the LDB grid at low displacement amplitudes. This early stage tensile stress transfer from
the substrate to the TRM asserts the fact that the in-plane capacity of the masonry had to be lower
than the one predicted Vm due to the damage suffered, as also evidenced by Alecci et al. (2019). On
the stage before the transfer began, the specimens were carrying, on average, a horizontal load equal
to 100 kN, which represents 70% of the capacity of the unreinforced and undamaged masonry wall
Vm . Table 7.3 shows the prediction of the shear capacity of the repaired and retrofitted specimens
when the reduced URM capacity due to the damage Vmred is used. It should be noted that the error
is reduced and beneath 5%.
159
Table 7.3: Comparison between the experimental values obtained for repaired and retrofitted specimens and
the analytical values computed according to the Italian guideline CNR-DT 200 (2013) and considering a reduced
URM capacity
ID VEXP Vmred Vt,f,EXP Geometry [mm] ff d Vt,f Error

URM R 180 100 80 1270 0.032 - - 1700 83 4.3
Figure 7.4: Digital Image Correlation (DIC) analysis: a) Readings of virtual extensometers located on the
strips of specimen LDS 2 before its failure in SCT configuration, b) Reading of virtual extensometers located
on the strips of specimen LDS 2 during its failure in SCT configuration
The case of LDS-TRM is different, as the strengthening solution reached the expected stress and
yet the error, even though acceptable, is unexpected. To better understand this phenomenon the
response of the specimens retrofitted with LDS was analysed by means of Digital Image Correlation
(DIC). To this extend, high-resolution images were taken every 5 seconds and processed by GOM
correlate software. The software allows a detailed representation of the test results as displacement
and strain fields. These fields do not only enable the accurate characterisation of the crack pattern,
but also the construction of virtual extensometers to replicate the physical ones. This feature of the
software allows the display of one extensometer on each horizontal strip applied on the specimens
retrofitted with LDS. Figure 7.4 shows the horizontal displacement experienced by each strip before
the attainment of the peak load and when the specimen reached its failure and it can be evidenced
that the strip located in the centre of the panel is carrying more load, evidencing higher displacements,
than the ones near the edges.
The predictive formulation provided by the guideline CNR-DT 200 (2013) considers a uniform
distribution of the stress along the reinforcement. Hence, in the case of LDS the contribution of each
160
strip is considered to be equal, which results in an overestimation of the TRM contribution to the
shear capacity of the retrofitted wall.
Figure 7.5 shows the experimental stress-strain curve of LDS-TRM, applied with two layers, along
with the response of the LDS-DL-TRM with increasing displacement amplitudes. From the figure
on the left, it can be noted that the strengthening solution with two layer of LDS did not achieve
the expected tensile stress, namely the ff d of LDS-TRM. Conversely, the LDS-DL-TRM reached over
half of its expected tensile strength before the masonry panel failed as a consequence of toe-crushing.
The guidelines CNR-DT 200 (2013) and CNR-DT 215 (2018) recommend that the contribution of the
strengthening solution does not lead to the failure of the compressed strut, namely the design shear
capacity of the retrofitted masonry Vt,R should not exceed the compressive strength of masonry on
the edge of the wall.
Table 7.4 includes the computation of the shear contribution of the two layers of LDS-TRM con-
sidering both values of ff d , the one experimentally obtained by characterisation of the composite (one
layer) and the one experimentally recorded during SCT. The first option yields a shear contribution of
the composite Vt,f with an error of 79%. Not only the error is substantial, but also the total shear ca-
pacity of the retrofitted wall yields a value of Vt,R equal to 355. Equation (7.9), which is recommended
by EC8 (2004) to control the crushing of masonry, yields Vc equal to 285 kN. Consequently, Vt,R equal
to 355 kN exceeds 25% the maximum capacity of the compressive strut. The second option, considers
the reduced value of ff d , according to the findings with SCT, and the composites contribution Vt,f
to the shear capacity only evidenced an acceptable error of 21%. The total shear capacity of the
retrofitted wall yields a value of Vt,R equal to 285 kN. Such value is in complete agreement with Vc .
Hence, it can be concluded that the computation should be done with the reduced value of ff d since
the level of strengthening, provided by two layers of LDS-TRM, encourages masonry crushing before
reaching the tensile capacity due to debonding of TRM σsl,t .
Table 7.4: Comparison between the experimental values obtained from SCT for specimen retrofitted with
two layers of LDS and the analytical values, computed according to the Italian guideline CNR-DT 200 (2013),
considering two different values for ff d , i.e. the one obtained from the mechanical characterisation and the one
recorded during SCT

LDS-DL 260 140 120 1270 0.169 100 250 2096 215 79
LDS-DL 260 140 120 1270 0.169 100 250 1420 145 21
The error between the analytical predictions and the experimental results of specimens retrofitted
with two layers of LDS-TRM, is in line with the error obtained for specimens retrofitted with one
layer of LDS-TRM. Consequently, the error is also related with a non-uniform distribution of stress
throughout the strips applied on the specimens.
For the sake of comparison, the analytical formulation proposed by CNR-DT 215 (2018) was inves-
161
Figure 7.5: Stress-Displacement curve of the average measurement of the LVDT in tension of specimens
retrofitted with two layers of LDS tested in SCT configuration
tigated to assess the shear capacity of each type of TRM strengthening solution Vt,f . The guidelines
compute Vt,f following Equation (7.10).
Vt,f = 0.5 · nf · tf · bf · αt · εf d · Ef εf d · Ef = ff d (7.10)
where nf is the number of layers applied, tf is the equivalent thickness of the reinforcement parallel
to the applied shear force, bf is the length of the reinforcement measured in the orthogonal direction
to the applied force, εf d is the strain corresponding to the reinforcement tensile capacity, Ef is the
Young’s modulus of the textile, ff d is the reinforcement tensile capacity, and αt is a coefficient of
reduction due to shear that in absence of experimental results the guideline suggests to be taken equal
to 0.80.
Table 7.5 compares the results yielded from the analytical formulations, provided by the guideline
following Equation (7.10), for each type of TRM strengthening system investigated. The error is
evaluated between the analytical value Vt,f,CN R215 and the experimental peak force sustained by the
TRM strengthening system, Vt,f,EXP . Only in the case of MDS the experimental value was taken
from the DCT results, for the remaining TRM systems the experimental values were taken from the
SCT results. The computed errors are higher than the ones yielded by the comparison between the
experimental values and the analytical values computed with Equation (7.4) provided by CNR-DT
200 (2013), as shown in Table 7.1, Table 7.2, and Table 7.4.
The errors yielded by the analytical formulation proposed by CNR-DT 215 (2018) can be ascribed
to several factors. First, as previously mentioned, this guideline disregards the assessment of strip
configurations which was profusely used in this research. Second, more experimental insight is needed
in order to properly define the coefficient of reduction αt .
Table 7.6 presents the comparison between the experimental values obtained from the DCT for
the strengthened wall in the asymmetrical configuration with LDB on one face and NSM rebars in the
162
Table 7.5: Comparison between the results provided by the analytical formulation of the guidelines CNR-DT
215 (2018) and the experimental values obtained from each TRM strengthening system
TRM System ff d [MPa] bf [mm] tf [mm] Vt,f,CN R215 [kN] Vt,f,EXP [kN] Error [%]
LDB 1700 1270 0.032 55 76 -27%
LDS 2096 800 0.084 113 79 43%
MDS 687 800 0.169 74 87 -15%
LDS DL 1420 800 0.168 153 120 27%
other. The analytical value Vt,f was computed as the sum of the values obtained from Equation (7.4)
and Equation (7.6). The former takes into account the side retrofitted with LDB Vf,LDB while the
latter takes into account the side retrofitted with joint repointing NSM helical rebars Vf,N SM . The
error is evaluated between the analytical value Vt,f , and the experimental peak force sustained by the
TRM strengthening system Vt,f,EXP .
Table 7.6: Comparison between the experimental values obtained from DCT on specimen asymmetrically
retrofitted and the analytical values computed according to Li et al. (2005)
ID VEXP Vm Vt,f,EXP D tm τb Vt,LDB Vt,N SM Vt,f Error

Specimen [kN] [kN] [kN] [mm] [mm] [MPa] [kN] [kN] [kN] [%]
LDB-JR 192 140 52 30 15 0.40 41 35 76 13
From the computed error it can be observed that the analytical formulation predicts fairly ac-
curately the shear contribution of the strengthening, namely LDB and NSM joint repointing. The
overestimation of the capacity provided by the strengthening solutions might be ascribed to several
factors. First, in the absence of experimental values the bond strength τb was taken equal to 3% of
frm,c . Second, parallel to the observed response of LDB in the other specimens, Figure 7.6 evidenced
that LDB-TRM did not achieve its expected tensile strength σu,f . Finally, in the analytical approach
proposed by Li et al. (2005) a uniform distribution of the shear stress, along the embedded length
of the rebars, is considered. Moreover, the contribution of all the rebars are equally considered. Un-
der the hypothesis that the rebars may behave similarly to the LDS strips, the stress may have been
non-uniformly distributed throughout the rebars. Unfortunately, DIC analysis did not provide enough
insight on this matter, since only a small portion of the wall, namely the centre, was being recorded
during testing.
Despite there is possible room for improvement and validation, the scattering observed, namely
the error computed, is moderate and consistent with the findings of other researchers, such as Li et al.
(2005) and Casacci et al. (2019).
163
Figure 7.6: Stress-strain experimental curves of the specimen asymmetrically retrofitted on one side with LDB
and on the other with NSM joint repointing tested in the DCT configuration
7.3.3 Efficiency of the TRM
The exploitation ratio is a useful parameter to evaluate the most effective strengthening solution
and contributes to decision-making regarding the enhancement of the seismic performance of existing
masonry structures.
The exploitation ratio, which accounts for the percentage of the textile’s usable tensile strength, is
computed as the ratio between the tensile capacity of the reinforcement ff d , as defined in the previous
section, and the ultimate tensile strength of the textile σu,f presented in Table 3.5. Table 7.7 presents
the calculation of the exploitation ratio for each type of reinforcement investigated in the present
research.
Table 7.7: Efficiency of TRM in term of exploitation ratio computed between the tensile capacity of the
reinforcement ff d and the ultimate tensile strength of the textile σu,f
ID ff d σu,f Exploitation Ratio

Specimen [MPa] [MPa] ff d /σu,f
LDB 1700 1700 1.0

LDB R 1700 1700 1.0
LDS 2096 2800 0.75
MDS 687 3000 0.23
LDS-DL 1420 2800 0.51
Given the type of failure experienced by the specimens retrofitted with LDB, namely fibre rupture,
the exploitation ratio was equal to one. When the exploitation ratio was lower than one the textiles,
164
LDS and MDS, never reached their failure due to rupture. Instead, the failure was always due to
delamination or debonding. The exploitation ratio obtained for specimens retrofitted with one layer
of LDS was equal to 0.75 while for specimens retrofitted with two layers of LDS was equal to 0.51. In the
latter case, the low exploitation ratio was consequence of a high level of strengthening which led to the
crushing of the masonry before the masonry was able to fully transfer the tensile stress to the textile.
For the specimens retrofitted with MDS, which also failed due to debonding, moderate disagreement
was observed between the experimental and analytical results, as they yielded a significantly low
exploitation ratio of only 0.23. This low value might be attributed to deficient bonding with the
substrate induced by the higher yarn density. For this reason, MDS configuration was not investigated
further under SCT configuration.
Figure 7.7 illustrates the linear correlation between the reinforcement ratio ρ, defined as the ratio
between the axial stiffness of the TRM reinforcing systems and that of the masonry (see Equation
(5.7)), and the exploitation ratio ff d /σu,f .
Figure 7.7: Correlation between the reinforcement ratio ρ and the exploitation ratio ff d /σu,f of the textiles
studied
It is evident that lower ρ yields higher efficiency with a failure due to textile rupture rather
than failure involving the textile-matrix interface. The textile with larger grid spacing, namely LDS,
exhibited higher exploitation ratio rather than those with smaller spacing such as MDS. This is
consequence of the matrix-textile load transfer provided by larger amount of mortar passing thought
the voids of fabric allowing higher bond efficiency.
The benefits of sparser textiles are also evidenced in the solution with two layers of LDS. The
reinforcement ratio ρ of MDS is equal to the one yielded by LDS-DL. However, the former is achieved
165
by its yarn density while the latter is achieved by applying two layers of a sparser textile LDS.
Notwithstanding the same value of ρ, the textile with larger grid spacing yielded higher exploitation
ratio compared to the denser one.
7.4 Conclusions
This chapter has presented the analytical formulations and approaches currently available to predict
the in-plane shear contribution of the TRM strengthening solutions for masonry structures. The
analysis has addressed the analytical prediction of the shear capacity of masonry walls retrofitted with
different TRM solutions and the comparison with the experimental results. The main conclusions of
the analysis can be summarized as follows:
The results provided by the analytical approaches are in good agreement with the experimental
ones. The predictive analytical formulations are deemed to be fairly accurate.
The shear capacity of repaired and retrofitted specimens URM R can be accurately estimated
with the analytical formulation for TRM as suggested by CNR-DT 200 (2013) by adopting a
reduced in-plane shear contribution of the unreinforced masonry. Consequently, the degree of
damage observed in the masonry wall should be taken into consideration when evaluating a
repairing solution.
The analytical outcomes of specimens retrofitted with unidirectional strips, namely LDS, MDS
and LDS-DL, show that the tensile stress is not uniformly distributed throughout the strips and
hence their contribution to the shear capacity decreases as the strip gets further from the centre
of the wall. The DIC also validates this result.
The analytical approach proposed by Li et al. (2005) predicts fairly accurately the shear contri-
bution of near surface mounted (NSM) joint repointing with helical stainless steel rebars.
The exploitation ratio is a useful parameter to evaluate the efficiency of the strengthening solu-
tion. There is an inverse linear correlation between the reinforcement ratio ρ and the exploitation
ratio ff d /σu,f . An increment of ρ leads to a decrement of the exploitation ratio. However, it
must be noted that this increment of ρ is a consequence of increasing the yarn density of the
textile rather than increasing the equivalent thickness of the textile tf . Therefore, the decrease
in efficiency of the strengthening solution is linked to the difficult penetration of the mortar
through the voids of the denser textile.
Considering the efficiency of the textile, it is better to apply two layers of sparser unidirectional
textile, namely LDS, rather than a single layer of a denser textile, namely MDS. Even though
the former may lead to masonry crushing, as a consequence of high level of strengthening, the
latter is misspent since only a quarter of its tensile strength is used.
166
Chapter 8
Conclusions
8.1 Summary
Experimental testing is one of the most useful tools for the reliable evaluation of masonry’s main me-
chanical properties. The present work has focused on the investigation of two of the main experimental
testing methodologies available to assess the in-plane response of masonry, Diagonal Compression Test
(DCT) and Shear Compression Test (SCT). These tests were performed on unreinforced masonry
(URM) walls as well as on masonry strengthened with different Textile Reinforced Mortar (TRM)
strengthening solutions. The analysis of the experimental response provided a meaningful insight on
the failure mechanisms that triggered the collapse of strengthened masonry.
The failure mechanisms were associated to the specific property of the material that governs the
in-plane behaviour, i.e. the tensile strength. The results of the tests helped reducing the uncertainty
involved in the evaluation of inherent masonry properties.
Chapter 3 presents, in addition to the mechanical characterisation of the materials involved in
this research, the technical information regarding the set-up configuration of both tests in terms of
boundary conditions, load application and loading protocols. Chapter 4 and 5 present the results
observed during the experimental programme involving twenty-two specimens, in both configurations
unreinforced and strengthened, tested under monotonically diagonal compression and cyclical shear
compression. Among the TRM strengthening solutions there are four externally bonded solutions and
one based on joint repointing with Near Surface Mounted (NSM) helical rebars. The externally bonded
solutions comprised continuous bidirectional grids of basalt (LDB) and discrete bands of unidirectional
steel of two different yarn densities (LDS and MDS). The sparser one was also applied in one and two
layer configurations (LDS DL). All the mentioned solutions were embedded in a lime-based mortar
matrix and were symmetrically applied on the specimens, except for the joint repointing NSM that
was asymmetrically applied in combination with LDB (LDB-JR).
The experimental response observed on strengthened masonry for both testing methodologies
showed that the most likely failure mode is diagonal cracking as a result of reaching the material’s
167
CHAPTER 8. CONCLUSIONS
tensile strength. The specimens strengthened with two layers of TRM showed to be prone to masonry
crushing due to excessive strengthening level. Due to the similarities detected in terms of crack pat-
tern, failure mechanism, shear strength and displacement capacities, it becomes evident that there is a
correlation between both testing methodologies, DCT and SCT. Notably, there is not yet a methodol-
ogy that correlates them for a more reliable assessment of the in-plane behaviour of masonry. To this
aim, Chapter 6 presents a correlation study to pursuit the development of a methodology that can
effectively be used for the analysis of the shear parameters of masonry walls in both configurations,
unreinforced and retrofitted. From the correlation a new shear parameter τ0,m is presented as a first
unbiased estimation of the in-plane capacity of the masonry wall. In Chapter 7 the parameter τ0,m is
used to compute the shear contribution of masonry, while the shear contribution of TRM strengthen-
ing systems is assessed with the available formulations provided by the Italian guideline CNR-DT 200
(2013).
8.2 Conclusions
This section summarises the main specific conclusions obtained for the experimental programme, the
correlation study and the analytical approach.
8.2.1 Experimental Results
All the adopted Textile Reinforced Mortar (TRM) solutions have demonstrated to be fully
compatible with the masonry substrate composed of solid clay bricks and lime mortar joints, since
no premature debonding failure occurred before the peak resistance was attained. They have
also shown to be efficient and easily implementable. Among the five solutions, the application
of the bidirectional basalt grid (LDB) can be considered as the less time consuming, due to fact
that applying a single surface layer is faster than applying a multiple strip configuration. The
Near Surface Mounted (NSM) system turns out to be the less invasive and the most reversible
technique among the ones tested.
All the TRM reinforcement systems enhanced the global in-plane behaviour of the specimens
and also provided a more homogeneous response in comparison with the URM walls, reducing
the inherent scattering of the masonry constituents. This means that the TRM strengthening
systems application can mitigate the possible influence of the variability of masonry properties
by providing more homogeneous structural response.
The application of TRM strengthening systems improved the strength of the walls compared
to the original URM. From the results of the DCT it is evidenced that the highest rates of
enhancement have been obtained for the LDB and the low density steel (LDS) systems, with
an increment of 100% and 89%, respectively, in terms of load-bearing capacity. In the case of
SCT, the results have shown that the lateral load-bearing capacity was increased in a range
between 30%-60%. The specimens repaired and retrofitted (URM R) evidenced a load-bearing
168
capacity increment of 70% in the cases tested with DCT and 10% in the cases tested with SCT.
This difference in enhancement evidenced by each testing methodology might be ascribed to one
reason, the DCT were performed monotonically while SCT cyclically, asserting that experimental
results obtained from monotonic test yield higher values. Notwithstanding the different results
in terms of load bearing capacity yielded by each pair of specimens tested under both testing
methodologies, they have also evidenced a remarkable improvement in terms of displacement
capacity and thus, in ductility. The application of TRM strengthening systems has not altered
the initial stiffness. These outcomes become of paramount importance in the seismic retrofit of
existing masonry structures. In specimens tested under SCT, ductility evidenced an increment
in the range of 20%-60% , which was also translated in a greater capacity of energy dissipation
after the cracking point of the specimens. In specimens tested under DCT the increment in
ductility evidenced a range of 18%-114%. However for the specimens strengthened with the
same TRM systems that those tested under SCT, the increment in ductility yielded from DCT
was 46%-72%. Similarly to the load-bearing capacity, the monotonic DCT led to higher values
of ductility.
Both repairing techniques, crack repointing with a M15 lime-based mortar and the ”scuci-cuci”
technique, used together with the continuous LDB grid symmetrically applied on the damaged
walls, allowed the full recovery of the undamaged stiffness of the walls. The initial stiffness
was marginally affected by reinforcement due to the low thickness of the strengthening systems
applied. These results show the capability of the investigated technique for post-earthquake
repair of existing masonry structures.
Specimens strengthened with unidirectional textiles of LDS and MDS fibres, subjected to DCT,
provided qualitatively similar results in terms of crack pattern, but specimens strengthened
with LDB exhibited better performance in quantitative terms due to the better interlocking
between the textile fibre and the matrix. The specimens strengthened with one and two layers
of LDS subjected to SCT were characterised by different failure mechanisms. The specimens
with one layer of LDS underwent shear failure, and consequently exploited, to a certain extent,
the textile’s tensile capacity, while the specimens with two layer of LDS led to masonry crushing.
Yarn density showed to have a significant influence on the shear capacity provided to the URM.
The sparser the yarns, the better the textile-mortar interlocking and hence, the better the in-
plane structural performance. Considering the same yarn density, the increment in the number
of layers resulted in the detriment of the vertical integrity of the structure. Despite the fact
that two layers of LDS evidenced a significant increment in the lateral load-bearing capacity
when compared to the URM material, the collapse due to crushing of the masonry indicates an
excessive level of strengthening.
The most remarkable conclusions regarding the experimental results obtained from the cyclic SCT
are:
169
Comparing the crack patterns of the URM walls with those of the strengthened ones, the latter
experienced a more widespread damage involving both masonry and matrix. The presence of
TRM strengthening systems allow a proper redistribution of stress, and therefore, specimens can
withstand larger imposed displacements, proving the efficiency of the strengthening technique.
One important outcome from the SCT results is the correlation between the shape of the experi-
mental force-displacement curves, along with their corresponding experimental envelope curves,
and the different failure modes experienced by each specimen. The hysteretic shape is there-
fore dependent on the failure mode undergone by the specimen. Along with the shape, it was
recognised that also the energy dissipation capacity and the equivalent damping coefficient, are
dependent on the failure mode, since both are in close relation with the shape of the hysteretic
loops. Conversely, pre-cracking equivalent damping coefficient is independent of the failure mode
since all specimens yielded similar values.
EC8 (2004) recommends the use of the cracked stiffness when assessing the behaviour of masonry
buildings under seismic actions. According to the standards, such stiffness, should be taken as
one-half of their respective uncracked value, i.e. the initial stiffness computed at the 15% of
Hmax . However, when it comes to strengthened masonry there is insufficient guidance. The
experimental results from SCT shed light on this matter and validates the computation of the
cracked stiffness of strengthened masonry as the slope of the idealized envelope at the onset of
cracking. At the same time the results showed that, for strengthened masonry, the cracked stiff-
ness, also defined as effective stiffness, can be computed as the half of their respective uncracked
stiffness.
To design strengthening solutions or to assess strengthened structures, there is a lack of codes and
guidelines based on the limit state approach. Therefore, a range of drift values for different failure
modes was suggested based on the results obtained in the present research for the displacement
capacity of strengthened masonry.
8.2.2 Correlation between Diagonal Compression Test and Shear Compression

Test
The method discussed in the present work can only be applied to masonry structures, unre-
inforced or strengthened, failing in a tensile diagonal cracking mode. This failure mode was
experimentally observed on the specimens tested under both testing configurations, DCT and
SCT. The failure was accurately described by the simplified model proposed by Turnsek and Ca-
covic (1971), governed by only one parameter, corresponding to the equivalent tensile strength ft ,
featuring the behaviour of an equivalent isotropic homogeneous material. The formulations de-
veloped for isotropic homogeneous material such as Frocht (1931), Turnsek and Cacovic (1971),
and Brignola et al. (2008) could be utilized to assess the in-plane behaviour of strengthened
masonry.
170
The experimental results yielded from both testing configurations, DCT and SCT, were used to
propose a methodology that can be effectively used for the analysis of the shear parameters of
masonry walls in both configurations, unreinforced and retrofitted. The proposed methodology
validates the DCT as a simpler standardised test to evaluate the equivalent tensile strength of
masonry walls failing due to tensile diagonal cracking in both configurations. In addition, it
provides a simple procedure to accurately assess the value of the coefficient α, as proposed by
Brignola et al. (2008). The coefficient α depends on the masonry typology and allows the compu-
tation of its tensile strength. For the material combination hereby investigated, i.e. double-leaf
historical clay brick masonry with low strength mortar joints, α yielded a value equal to 0.43.
Finally, one of the most important implications of this study is that it provides experimental in-
sight to better understand the interpretation of the stress state of masonry walls when subjected
to DCT, validating the formulation proposed by Brignola et al. (2008)
The proposed methodology also allows the definition of a shear parameter τ0,m to characterise the
in-plane behaviour of masonry walls. The parameter is graphically defined as the value at which
the Mohr’s circle intersects the shear strength axis (τ ). The novelty of the parameter lays on its
independence from the loading and boundary conditions imposed by the testing configurations.
The proposed methodology provides the linear correlation between the shear parameter τ0,m
and the equivalent tensile strength ft . Such correlation confirms that the in-plane behaviour of
strengthened masonry can be defined by an equivalent tensile strength provided by the TRM
strengthening system.
8.2.3 Investigation on analytical approaches
The results provided by the reviewed analytical approaches are in good agreement with the
experimental ones. The predictive analytical formulations are deemed to be fairly accurate,
while being simple to implement. The formulation provided by CNR-DT 200 (2013) yielded
significantly accurate results when the tensile capacity ff d was taken accordingly to the failure
mode associated to the TRM strengthening systems, i.e. rupture of the textile or debonding
within the matrix. Similarly, the analytical approach proposed by Li et al. (2005) predicts fairly
accurately the shear contribution of near surface mounted (NSM) joint repointing with helical
stainless steel rebars.
The shear capacity of the repaired and strengthened masonry walls Vt,R , in which the TRM
strengthening systems is applied as post-earthquakes repairing technique can be accurately es-
timated with the analytical formulation provided by CNR-DT 200 (2013). However, in order to
predict Vt,R , the shear contribution of masonry Vm has to be adopted with a reduced value due to
the damaged condition of the masonry. The experimental results of the present research showed
that for slightly damaged masonry integrity can be restored by crack repointing techniques,
and its in-plane shear contribution Vm can be considered equal to the undamaged one, whereas
for severely damaged masonry, for which more invasive repairing techniques are required, i.e.
171
”scuci-cuci”, its in-plane shear contribution must be reduced when assessing the strengthening
solutions. From the experimental results obtained in this work, the reduction is suggested to be
of 30%.
The moderate scattering evidenced between the analytical results, using the formulation provided
by CNR-DT 200 (2013) and the experimental results of specimens retrofitted with unidirectional
strips, namely LDS, MDS and LDS-DL, showed that the tensile stress is not uniformly distributed
throughout the strips. Digital image correlation (DIC) validated that the strip’s contribution to
the shear capacity, by providing tensile stress, decreases as the strip gets further away from the
centre of the wall.
The exploitation ratio is a useful parameter to evaluate the effectiveness of the strengthening
solution. There is an inverse linear correlation between the reinforcement ratio ρ, defined as the
ratio between the axial stiffness of the TRM reinforcing systems and that of the masonry, and
the exploitation ratio ff d /σu,f defined as the ratio between the debonding load and the tensile
failure load. An increment of ρ leads to a decrement of the exploitation ratio. However, it must
be noted that in the present research this increment of ρ is a consequence of increasing the yarn
density of the textile or increasing the number of layer of the textile. Therefore, the decrease
in efficiency of the strengthening solution is linked in the first case to the difficult penetration
of the mortar through the voids of the denser textile. In the second case is a consequence of
the high level of strengthening provided by the increasing number of layers which leads to the
crushing of masonry before the TRM can start carrying the tensile load.
8.3 Future work

Considering the knowledge gaps identified in the present work, the following future works are suggested:
For the SCT experimental programme herein presented, a quasi-static cyclic testing protocol with
only one axial ratio σ0 /fc was adopted with the aim of describing the in-plane behaviour of a
masonry wall within a low rise multi-storey building. However, several researchers have evidenced
the influence of the axial ratio σ0 /fc on the in-plane response of URM specimens. The peak load
and initial stiffness increase linearly with the axial load, while the displacement capacity sensibly
decreases with increasing axial ratio. Performing SCT with different axial ratios would provide
valuable information regarding the response of strengthened masonry and regarding the influence
of increasing axial ratio on the enhancement provided by the TRM strengthening systems. In
addition, having experimental results with different compressive levels would provide further
insight on the failure surface of shear walls. Finally, dynamic testing would allow for assessment
of the dynamic characteristics of the strengthened wall.
The present research suggests a range of drift values for different failure modes. However, these
values are based on a still insufficient number of experimental tests for a complete validation.
Therefore, further experimental research on different masonry typologies strengthened with dif-
172
ferent TRM strengthening systems is required to improve the knowledge on drift capacities of
strengthened masonry. A larger data base would allow the formulation of drift capacity models,
proposing equations to predict the drift capacities (corresponding to different levels of damage)
based not only on the failure mode but also on the boundary conditions, the axial ratio and the
aspect ratio.
It would be interesting to perform an analytical study using finite element models (FEM) to vali-
date the experimental results obtained from both testing methodologies, DCT and SCT, for both
specimens configurations, unreinforced and strengthened. It should be noted that the constitu-
tive model in tension must reflect the deformation capacity provided by the TRM strengthening
system. Since experimental campaigns of the magnitude hereby presented are costly, numerical
models are a good alternative to analyse the in-plane response of strengthened masonry with
different configurations, i.e. aspect ratio H/B, axial ratios σ0 /fc , masonry typologies or TRM
strengthening systems. Such complementary numerical study would also allow the analysis of
the seismic response of multi-storey masonry building in the strengthened configuration. Par-
ticular interest should be paid to the modelling and validation of repaired masonry to analyse
the TRM strengthening systems as a post-earthquake solution.
Regarding the proposed methodology it would be of interest to expand its applicability range
to other masonry typologies, including different materials but also different patterns, to validate
the accurate computation of the coefficient α and the equivalent tensile strength ft . In addi-
tion, further research on different TRM strengthening solutions could be applied to validate the
empirical linear correlation proposed between the equivalent tensile strength ft and the shear
parameter τ0,m .
To improve the validation of the analytical formulations provided by the available guidelines and
approaches, it is of utmost importance to improve the knowledge of the mechanical properties of
the TRM strengthening components. Therefore, additional research should be carried in order
to analyse, among other aspects, the bond properties between helical rebars and the mortar
matrix, the bond properties between the two layers of TRM and between the doubled layer
TRM and the substrates.
173
Appendix A
Use of DIC in experimental testing of

masonry walls
The use of Digital Image Correlation (DIC) in the present work had two main aims. The first one
was to control the strains and displacements observed during testing. The second one was to register
the crack evolution during testing and to identify the final crack pattern. The DIC analysis performed
in this research allowed the identification of some experimental difficulties ascribed to the limitations
presented by the DIC methodology mainly related to the first aim. The displacement measurements
obtained by DIC were compared with the ones obtained from the conventional sensors and evidenced
significant scattering. However, DIC provided valuable qualitative information regarding the crack
formation and its evolution, which allowed a better understanding of the failure mechanism triggered
by in-plane actions.
The implementation of DIC technique is based on the tracking of pixels from a stochastic pattern.
This pattern, generated by black spray, has to be unique and random comprising a spotted pattern
characterised by different speckles sizes and a range of contrast and intensity levels. The randomness
of the pattern allows the clear identification of the facets throughout all the available pictures. The
displacement is then measured by applying a correlation algorithm between this facets.
The effectiveness of the DIC methodology relies, among others parameters, on the lighting condi-
tions and the stochastic pattern, specifically on its scale resolution. The lighting conditions strongly
influence the quality of the reference points to build the facets of the surface in the GOM Correlate
software, as shown in Figure A.1. The light must be cold white and fixed so to assure a homogenous
distribution. The lighting condition in the case of SCT was assured since the lamps were fixed to the
testing steel frame. This provided a homogenous and constant lighting condition. However, this was
not the case of Diagonal Compression Tests (DCT), where the lighting was provided by individual
standing lamps, which were, to certain extent, influenced by the environmental light fluctuation.
Notwithstanding the importance of good lighting conditions, the major difficulties were encoun-
tered in the definition of the scale resolution. This scale, defined as the dimension of the pixels on
the picture, has to be in agreement with the measures that it is desired to quantify. The current
research used a commercial reflex camera with 4000 x 6000 pixels and the specimens were square
175
APPENDIX A. USE OF DIC IN EXPERIMENTAL TESTING OF MASONRY WALLS
Figure A.1: Specimens subjected to DCT: a) cold white lighting condition provides good reference points to
build the facets of the surface, b) yellowish lighting condition provides poor and sparse reference points that do
not allow the proper connection between the facets to build the surface
with a dimension of 1300 mm per side. Therefore, the best resolution achievable of the full surface
was 0.3 mm/pixel. Unfortunately, this scale was in the order of magnitude of masonry displacement
in compression. Therefore, in the case of DCT it was decided not to track the full surface but only
a region of interest located at the centre on the wall. The scale resolution was on average equal
to 0.15 mm/pixel. In spite of representing an improvement, in those specimens in which the light-
ing conditions were not an issue, only the displacements in tension were measured with a moderate
agreement. In all the cases, all the displacements smaller than the scale resolution evidenced high
scattering and random variations, associated to noise and interference. Qualitatively the DIC force-
displacement curves showed some resemblance, but quantitatively the displacement evidenced high
errors in the smaller displacements when compared to the readings of the LDVT sensors. This error
experienced an exponential reduction as the displacements increased. However, meaningful parameter
such as shear stiffness cannot be computed from this results. At this point, it is of interest to note
that according to Criado Ibáñez (2015) and Torres et al. (2020), among other authors, the accuracy
of DIC measurements can be equal to a quarter of the scale resolution. However, in this research this
enhancement in accuracy was not appreciated on either of the scale resolutions analysed.
In the case of Shear Compression Test (SCT), the imposed displacements were several orders of
magnitude bigger than the scale resolution. However, the accuracy of the displacements measured by
DIC was significantly poor. The errors were ascribed to different reasons, such as software problems,
elements involved in the testing and the specimen itself. The first error was associated to the soft-
176
ware having problems to properly correlate the facets between pictures. This mismatched correlation
provided inaccurate measurements resulting in false information. The second error was associated to
the conventional sensors placed on the surface of the specimen which, along with their cables and
accessories, obstructed in most of the cases the DIC analysis, as evidenced in the examples shown
in Figure A.2. As a result, the facet surface built by GOM software experienced interferences and
missed parts of the surface information that were of great value for the analysis. Finally, in the cases
of specimens retrofitted with TRM, the deformations experienced by the strengthening systems as
consequence of the testing were in detriment of the DIC measurement. This deformations, i.e. mortar
spalling or detachment, generated movement outside the plane monitored by DIC, making the pattern
no longer visible for the camera. Therefore, the information provided by the area was no longer of use
for the analysis.
Figure A.2: Two examples of the error ascribe to the obstruction of the conventional sensors
In addition to the scale resolution, there is another feature that needs to be taken into consideration
when generating the stochastic pattern, namely the speckles sizes. In the present research the speckles
were created using a spray, which is relatively easy to master but does not provide a good control over
the size of the speckles, as evidenced in Figure A.3. In the lack of expertise, this issue could represent
a great problem when creating the pattern on large specimens.
Figure A.3: a) bad quality of the stochastic pattern characterised by speckles of the same size, b) good quality
of the stochastic pattern characterised by different sizes of speckles
177
From the results obtained in the present experimental campaign, the use of DIC is highly recom-
mended as a complementary tool for the evaluation of masonry’s structural behaviour. From a general
perspective, DIC analysis can provide a meaningful and detailed information on the strain develop-
ment and crack evolution experience by the tested specimen. However, to control displacement and
strain measurements further research needs to be carried out. Nevertheless, there are some considera-
tions that must be taken from this experience. First, due to the difference between masonry’s tensile
and compressive strengths, DIC analysis cannot target both behaviours with the same scale resolu-
tion. For this reason, the displacement that will be controlled by DIC must be decided in advanced.
Consequently the design of the stochastic patter, speckles size and scale resolution should be designed
taking into account the mentioned difficulties. Second, in order to obtain as much information as
possible, the region of interest under analysis should be, as far as possible, empty of sensors or any
other testing hardware that can compromise the visibility of the surface. Therefore, it is necessary to
ensure a correct building of the facets of the surface and its correlation between pictures. Finally, if
all issues are addressed, DIC analysis can provide a large amount of information, since practically any
type of sensor can be virtually built.
178
Bibliography
Akhoundi, F., Vasconcelos, G., Lourenço, P., Silva, L. M., Cunha, F., and Fangueiro, R. (2018).
In-plane behavior of cavity masonry infills and strengthening with textile reinforced mortar. Eng.
Struct., 156:145–160.
Alecci, V., Barducci, S., D’Ambrisi, A., De Stefano, M., Focacci, F., Luciano, R., and Penna, R.
(2019). Shear capacity of masonry panels repaired with composite materials: Experimental and
analytical investigations. Compos. Part B Eng., 171(March):61–69.
American Concrete Institute (2013). ACI 549.4R-13 - Guide to Design and Construction of Externally
Bonded Fabric-Reinforced Cementitious Matrix (FRCM) Systems for Repair and Strengthening
Concrete and Masonry Structures.
Applied Technology Council (2007). Interim Testing Protocols for Determining the Seismic Perfor-
mance Characteristics of Structural and Nonstructural Components - FEMA 461.
Ascione, L., De Felice, G., and De Santis, S. (2015). A qualification method for externally bonded Fibre
Reinforced Cementitious Matrix (FRCM) strengthening systems. Compos. Part B Eng., 78:497–506.
ASTM (2000). Standard Test Method for Diagonal Tension ( Shear ) in Masonry Assemblages.
Babaeidarabad, S., C, D., and Nanni, A. (2014). URM walls strengthened with fabric-reinforced
cementitious matrix (FRCM) subjected to in-plane and out-of-plane load. In J. Compos. Constr.,
volume 18.
Balsamo, A., Di Ludovico, M., Prota, A., and Manfredi, G. (2011). Masonry walls strengthened with
innovative composites. Am. Concr. Institute, ACI Spec. Publ., 275:1–18.
Barducci, S., Alecci, V., De Stefano, M., Misseri, G., Rovero, L., and Stipo, G. (2020). Experimental
and Analytical Investigations on Bond Behavior of Basalt-FRCM Systems. J. Compos. Constr.,
24(1).
Basili, M., Vestroni, F., and Marcari, G. (2019). Brick masonry panels strengthened with textile
reinforced mortar: experimentation and numerical analysis. Constr. Build. Mater., 227.
179
BIBLIOGRAPHY
Bellini, A., Shahreza, S. K., and Mazzotti, C. (2019). Cyclic bond behavior of FRCM composites
applied on masonry substrate. Compos. Part B Eng., 169(April):189–199.
Benedetti, A. (2019). Diagonal Compression Behaviour of Masonry Walls Reinforced with FRM
Coatings. Struct. Anal. Hist. Constr. An Interdiscip. Approach Rilem Bookseries, 18:474–483.
Borri, A., Castori, G., and Corradi, M. (2015). Determination of Shear Strength of Masonry Panels
Through Different Tests. Int. J. Archit. Herit., 9(8):913–927.
Brignola, A., Frumento, S., Lagomarsino, S., and Podestà, S. (2008). Identification of Shear Parameters
of Masonry Panels Through the In-Situ Diagonal Compression Test. Int. J. Archit. Herit., 3(1):52–
73.
Bui, T.-l., Si Larbi, A., Reboul, N., and Ferrier, E. (2015). Shear behaviour of masonry walls strength-
ened by external bonded FRP and TRC. Compos. Struct., 132:923–932.
Caggegi, C., Lanoye, E., Djama, K., Bassil, A., and Gabor, A. (2017). Tensile behaviour of a basalt
TRM strengthening system: Influence of mortar and reinforcing textile ratios. Compos. Part B
Eng., 130:90–102.
Calderini, C., Cattari, S., and Lagomarsino, S. (2010). The use of the diagonal compression test to
identify the shear mechanical parameters of masonry. Constr. Build. Mater., 24(5):677–685.
Casacci, S., Gentilini, C., Di Tommaso, A., and Oliveira, D. V. (2019). Shear strengthening of
masonry wallettes resorting to structural repointing and FRCM composites. Constr. Build. Mater.,
206:19–34.
CEN (1999). EN 1015-11 - Methods of test for mortar for masonry - Part 11: Determination of flexural
and compressive strength of hardened mortar.
CEN (2002). EN 772-6, Methods of test for masonry units. Part 6: Determination of bending tensile
strength of aggregate concrete masonry units.
CEN (2010). EN 998-2: Specification for mortar for masonry - Part 2: Masonry Mortar.
CEN (2011). EN 772-1, Methods of test for masonry units. Part 1: Determination of compressive
strength.
CEN (2014). EN 12390-13: Testing hardened concrete. Part 13: Determination of secant modulus of
elasticity in compression.
Chopra, A. K. (2001). Dynamics of structures: Theory and applications to earthquake engineering.
CNR - Consiglio Nazionale delle Ricerche (2018). DT 215/2018 - Istruzioni per la progettazione,
l’esecuzione ed il controllo di interventi di consolidamento statico mediante l’utilizzo di compositi
fibrorinforzati a matrice inorganica (in Italian).
180
BIBLIOGRAPHY
CNR-DT 200 R1/2013 (2013). Guide for the design and construction of externally bonded FRP
systems for strengthening existing structures.
Criado Ibáñez, L. E. (2015). Aplicación de técnicas de fotogrametrı́a a la medida de deformaciones de

probetas de suelo en laboratorio. PhD thesis, Universitat Politècnica de Catalunya.
Crisafulli, F. J. (1997). Seismic behaviour of reinforced concrete structures with masonry infills. PhD
thesis, University of Canterbury.
da Porto, F., Mosele, F., and Modena, C. (2011). In-plane cyclic behaviour of a new reinforced
masonry system: Experimental results. Eng. Struct., 33(9):2584–2596.
D’Ambrisi, A., Feo, L., and Focacci, F. (2013). Experimental and analytical investigation on bond
between Carbon-FRCM materials and masonry. Compos. Part B Eng., 46:15–20.
de Felice, G., D’Antino, T., De Santis, S., Meriggi, P., and Roscini, F. (2020). Lessons Learned on the
Tensile and Bond Behavior of Fabric Reinforced Cementitious Matrix (FRCM) Composites. Front.
Built Environ., 6(5).
de Felice, G., De Santis, S., Garmendia, L., Ghiassi, B., Larrinaga, P., Lourenço, P. B., Oliveira,
D. V., Paolacci, F., and Papanicolaou, C. G. (2014). Mortar-based systems for externally bonded
strengthening of masonry. Mater. Struct., 47:2021–2037.
De Santis, S., Carozzi, F. G., de Felice, G., and Poggi, C. (2017a). Test methods for Textile Reinforced
Mortar systems. Compos. Part B Eng., 127:121–132.
De Santis, S., Ceroni, F., de Felice, G., Fagone, M., Ghiassi, B., Kwiecień, A., Lignola, G. P., Morganti,
M., Santandrea, M., Valluzzi, M. R., and Viskovic, A. (2017b). Round Robin Test on tensile and
bond behaviour of Steel Reinforced Grout systems. Compos. Part B Eng., 127:100–120.
De Santis, S. and de Felice, G. (2015). Steel reinforced grout systems for the strengthening of masonry
structures. Compos. Struct., 134:533–548.
De Santis, S. and De Felice, G. (2015). Tensile behaviour of mortar-based composites for externally
bonded reinforcement systems. Compos. Part B Eng., 68:401–413.
Del Zoppo, M., Di Ludovico, M., Balsamo, A., and Prota, A. (2019). In-plane shear capacity of
tuff masonry walls with traditional and innovative Composite Reinforced Mortars (CRM). Constr.
Build. Mater., 210:289–300.
Deutsche Norm (1999). DIN 18555-9: Testing of mortar containing mineral binders - Part 9: Deter-
mining the compressive strength of hardened mortar.
European Committee for standardization (CEN) (1999). EN 1052-1:Methods of test for masonry -
Part 1: Determination of compressive strength.
181
BIBLIOGRAPHY
European Standard (2004). Eurocode 8 : Design of structures for earthquake resistance — Part 3:
Assessment and retrofitting of buildings.
European Standard (2006). Eurocode 6 - Part 3: Simplified calculation methods for unreinforced
masonry structures.
Faella, C. (2014). Report Prove Cicliche di Taglio Compressione su Pannelli Murari Riparati e Rin-
forzati (in Italian). Technical report.
Ferrara, G., Caggegi, C., Gabor, A., and Martinelli, E. (2019). Experimental study on the adhesion of
basalt Textile Reinforced Mortars (TRM) to Clay Brick Masonry: The influence of textile density.
Fibers, 7(103).
Ferretti, F., Ferracuti, B., Mazzotti, C., and Savoia, M. (2019). Destructive and minor destructive
tests on masonry buildings: Experimental results and comparison between shear failure criteria.
Constr. Build. Mater., 199:12–29.
Frocht, M. M. (1931). Recent advances in photoelasticity and an investigation of the stress distribution
in square blocks subjected to diagonal compression. Trans. ASME, Appl. Mech. Div., 53(15).
Frumento, S., Magenes, G., Morandi, P., and Calvi, G. M. (2009). Interpretation of experimental shear
tests on clay brick masonry walls and evaluation of q-factors for seismic design. Number May.
Gattesco, N., Amadio, C., and Bedon, C. (2015a). Experimental and numerical study on the shear
behavior of stone masonry walls strengthened with GFRP reinforced mortar coating and steel-cord
reinforced repointing. Eng. Struct., 90:143–157.
Gattesco, N., Boem, I., and Dudine, A. (2015b). Diagonal compression tests on masonry walls strength-
ened with a GFRP mesh reinforced mortar coating. Bull. Earthq. Eng., 13:1703–1726.
Giaretton, M., Dizhur, D., Garbin, E., Ingham, J., and Da Porto, F. (2018). In-plane strengthening of
clay brick and block masonry walls using textile reinforced mortar. J. Compos. Constr., 22(5):1–10.
Godio, M., Vanin, F., Zhang, S., and Beyer, K. (2019). Quasi-static shear-compression tests on
stone masonry walls with plaster: Influence of load history and axial load ratio. Eng. Struct.,
192(May):264–278.
Hendry, A. and Khalaf, F. (2001). Masonry wall construction, volume 39. London.
Hračov, S., Pospı́šil, S., Garofano, A., and Urushadze, S. (2016). In-plane cyclic behaviour of unfired
clay and earth brick walls in both unstrengthened and strengthened conditions. Mater. Struct.
Constr., 49(8):3293–3308.
ICC Evaluation Service (2017). AC434 - Acceptance criteria for masonry and concrete strengthening
using fabric-reinforced grout (SRG) composite systems.
182
BIBLIOGRAPHY
ICOMOS (2003). Icomos Charter- Principles for the Analysis , Conservation and Structural Restora-
tion of Architectural Heritage.
Incerti, A., Ferretti, F., and Mazzotti, C. (2018). Influence of FRCM retrofitting systems on the shear
behaviour of pre-damaged masonry panels. In Milano, G., Talercio, A., and Garrity, S., editors,
Proc. 10th Int. Mason. Soc. Conf., pages 2240–2249, Milan.
Ismail, N. and Ingham, J. M. (2013). Polymer textiles as a retrofit material for masonry walls. Proc.
Inst. Civ. Eng. - Struct. Build., 166(1):1–11.
Ismail, N., Petersen, R. B., Masia, M. J., and Ingham, J. M. (2011). Diagonal shear behaviour
of unreinforced masonry wallettes strengthened using twisted steel bars. Constr. Build. Mater.,
25:4386–4393.
Kerakoll (2016). Technical data sheet Biocalce MuroSano.
Knox, C. (2012). Assessment of Perforated Unreinforced Masonry Walls Responding In Plane. PhD
thesis, The university of Auckland.
Li, T., Galati, N., Tumialan, J. G., and Nanni, A. (2005). Analysis of unreinforced masonry concrete
walls strengthened with glass fiber-reinforced polymer bars. ACI Struct. J., 102(4):569–577.
Lignola, G. P., Caggegi, C., Ceroni, F., De Santis, S., Krajewski, P., Lourenço, P. B., Morganti, M.,
Papanicolaou, C. C., Pellegrino, C., Prota, A., and Zuccarino, L. (2017). Performance assessment
of basalt FRCM for retrofit applications on masonry. Compos. Part B Eng., 128:1–18.
Magenes, G. and Calvi, G. M. (1992). Cyclic behaviour of brick masonry walls. In Balkema, editor,
Proc. Tenth World Conf. Earthq. Eng. 19-24 July 1992 Madrid, Spain, pages 3517–3522.
Magenes, G. and Calvi, G. M. (1997). In-plane seismic response of brick masonry walls. Earthq. Eng.
Struct. Dyn., 26:1091–1112.
Magenes, G., Morandi, P., and Penna, A. (2008). D7.1 c Test results on the behaviour of masonry
under static cyclic in plane lateral loads. Technical report, EURCENTRE, Pavia.
Makoond, N., Cabané, A., Pelà, L., and Molins, C. (2020). Relationship between the static and
dynamic elastic modulus of brick masonry constituents. Constr. Build. Mater., 259.
Makoond, N., Pelà, L., and Molins, C. (2019). Dynamic elastic properties of brick masonry con-
stituents. Constr. Build. Mater., 199:756–770.
Mann, W. and Müller, H. (1980). Failure of Shear-Stressed Masonry - An Enlarged Theory, Tests and
Application to Shear Walls. In Proc. Int. Symp. load Bear. brickwork, pages 223–236, London.
183
BIBLIOGRAPHY
Marastoni, D., Pelà, L., Benedetti, A., and Roca, P. (2016). Combining Brazilian tests on masonry
cores and double punch tests for the mechanical characterization of historical mortars. Constr.
Build. Mater., 112:112–127.
Marcari, G., Basili, M., and Vestroni, F. (2017). Experimental investigation of tuff masonry panels
reinforced with surface bonded basalt textile-reinforced mortar. Compos. Part B, 108:131–142.
Marcari, G., Manfredi, G., Prota, A., and Pecce, M. (2007). In-plane shear performance of masonry
panels strengthened with FRP. Compos. Part B Eng., 38:887–901.
Menna, C., Asprone, D., Durante, M., Zinno, A., Balsamo, A., and Prota, A. (2015). Structural
behaviour of masonry panels strengthened with an innovative hemp fibre composite grid. Constr.
Build. Mater., 100:111–121.
Mercedes Cedeño, L. E. (2019). Análisis del comportamiento frente acciones cı́clicas de muros de
mamposterı́a refrozados con materiales compuestos de matriz inorganica y tejidos de fibra vegetales.
PhD thesis, Universitat Politècnica de Catalunya.
Messali, F., Metelli, G., and Plizzari, G. (2017). Experimental results on the retrofitting of hollow
brick masonry walls with reinforced high performance mortar coatings. Constr. Build. Mater.,
141:619–630.
Ministero delle Infrastrutture e dei Trasporti (2018). DM 17/01/2018 - Aggiornamento delle ”Norme
Tecniche per le Costruzioni” (in italian). pages 1–198.
Morandi, P., Albanesi, L., Graziotti, F., Li Piani, T., Penna, A., and Magenes, G. (2018). Development
of a dataset on the in-plane experimental response of URM piers with bricks and blocks. Constr.
Build. Mater., 190(November):593–611.
Mustafaraj, E. and Yardim, Y. (2019). Retrofitting damaged unreinforced masonry using external
shear strengthening techniques. J. Build. Eng., 26(August):100913.
null, N. (2017). Seismic Evaluation and Retrofit of Existing Buildings. American Society of Civil
Engineers, asce/sei 4 edition.
Page, A. W. (1981). Biaxial Compressive Strength of Brick Masonry. Proc. Inst. Civ. Eng. (London).
Part 1 - Des. Constr., 71:893–906.
Pan, B., Lu, Z., and Xie, H. (2010). Mean intensity gradient: An effective global parameter for
quality assessment of the speckle patterns used in digital image correlation. Opt. Lasers Eng.,
48(4):469–477.
Papanicolaou, C., Triantafillou, T., and Lekka, M. (2011). Externally bonded grids as strengthening
and seismic retrofitting materials of masonry panels. Constr. Build. Mater., 25(2):504–514.
184
BIBLIOGRAPHY
Papanicolaou, C. G., Triantafillou, T. C., Karlos, K., and Papathanasiou, M. (2007). Textile-reinforced
mortar (TRM) versus FRP as strengthening material of URM walls : in-plane cyclic loading. Mater.
Struct., 40:1081–1097.
Parisi, F., Iovinella, I., Balsamo, A., Augenti, N., and Prota, A. (2013). In-plane behaviour of tuff
masonry strengthened with inorganic matrix-grid composites. Compos. Part B Eng., 45:1657–1666.
Paulay, T. and Priestley, M. (1992). Seismic design of reinforced concrete and masonry buildings.
Pelà, L., Kasioumi, K., and Roca, P. (2017). Experimental evaluation of the shear strength of aerial
lime mortar brickwork by standard tests on triplets and non-standard tests on core samples. Eng.
Struct., 136:441–453.
Pelà, L., Roca, P., and Aprile, A. (2018). Combined In-Situ and Laboratory Minor Destructive Testing
of Historical Mortars. Int. J. Archit. Herit., 12(3):334–349.
Petry, S. and Beyer, K. (2014). Influence of boundary conditions and size effect on the drift capacity
of URM walls. Eng. Struct., 65:76–88.
Petry, S. and Beyer, K. (2015). Cyclic test data of six unreinforced masonry walls with different
boundary conditions. Earthq. Spectra, 31(4):2459–2484.
Preciado, A. and Ramirez-Gaytan, A. (2019). Numerical Seismic Reponse and Failure MOdes of Old
URM walls under lateral loading. In Struct. Anal. Hist. Constr., volume 18, pages 1293–1300.
Springer International Publishing.
Prota, A., Marcari, G., Fabbrocino, G., and Manfredi, G. (2006). Experimental In-Plane Behavior of
Tuff Masonry strengthened with cementitious matrix-grid composites. Compos. Constr., 10:223–233.
Reboul, N., Mesticou, Z., Si Larbi, A., and Ferrier, E. (2018). Experimental study of the in-plane
cyclic behaviour of masonry walls strengthened by composite materials. Constr. Build. Mater.,
164:70–83.
RILEM Recommendation for Testing and Use of Construction Material (1994). RILEM TC 76-LUM.
Diagonal tensile strength of small wall specimens.
Samarasinghe, W. (1980). The in-plane failure of brickwork. PhD thesis, University of Edinburgh.
Santandrea, M., Daissè, G., Mazzotti, C., and Carloni, C. (2017). An Investigation of the Debonding
Mechanism between FRCM Composites and a Masonry Substrate. Key Eng. Mater., 747:382–389.
Santandrea, M., Focacci, F., Mazzotti, C., Ubertini, F., and Carloni, C. (2020). Determination of the
interfacial cohesive material law for SRG composites bonded to a masonry substrate. Eng. Fail.
Anal., 111(March 2019).
185
BIBLIOGRAPHY
Santandrea, M., Imohamed, I., Carloni, C., Mazzotti, C., de Miranda, S., and Ubertini, F. (2016). A
study of the debonding mechanism in steel and basalt FRCM masonry joints. Brick Block Mason.
- Trends, Innov. Challenges, pages 433–440.
Segura, J. (2020). Laboratory experimental procedures for the compression and shear characterisation
of historical brick masonry. PhD thesis, Universitat Politècnica de Catalunya.
Segura, J., Aponte, D., Pelà, L., and Roca, P. (2020). Influence of recycled limestone filler additions
on the mechanical behaviour of commercial premixed hydraulic lime based mortars. Constr. Build.
Mater., 238.
Segura, J., Pelà, L., and Roca, P. (2018). Monotonic and cyclic testing of clay brick and lime mortar
masonry in compression. Constr. Build. Mater., 193:453–466.
Silva, P. F., Yu, P., and Nanni, A. (2008). Monte Carlo Simulation of Shear Capacity of URM Walls
Retrofitted by Polyurea Reinforced GFRP Grids. J. Compos. Constr., 12(4):405–415.
Sinha, B. (1967). Model Studies Related To Load-Bearing Brickwork. PhD thesis.
Tomaževič, M. (1999). Earthquake-Resistant Design of Masonry Buildings, volume 1. Imperial Collage

Press.
Tomaževič, M. (2008). Shear resistance of masonry walls and Eurocode 6 : shear versus tensile strength
of masonry. Mater. Struct.
Tomaževič, M. and Weiss, P. (2012). Robustness as a criterion for use of hollow clay masonry units
in seismic zones: An attempt to propose the measure. Mater. Struct. Constr., 45(4):541–559.
Torres, B., Varona, F. B., Baeza, F. J., Bru, D., and Ivorra, S. (2020). Study on retrofitted masonry
elements under shear using digital image correlation. Sensors (Switzerland), 20(7).
Turnsek, V. and Cacovic, F. (1971). Some experimental results on the Strength of Brick Masonry
Walls. Proc. 2nd Int. Brick Mason. Conf.
Turnsek, V. and Sheppard, P. (1980). The shear and flexural resistance of masonry walls. In Proc.
Int. Res. Conf. Earthq. Eng., pages 517–573, Skopje, Macedonia [Yugoslavia].
Valluzzi, M. R., Tinazzi, D., and Modena, C. (2002). Shear behavior of masonry panels strengthened
by FRP laminates. Constr. Build. Mater., 16:409–416.
Vanin, F., Zaganelli, D., Penna, A., and Beyer, K. (2017). Estimates for the stiffness, strength and
drift capacity of stone masonry walls based on 123 quasi-static cyclic tests reported in the literature.
Bull. Earthq. Eng., 15(12):5435–5479.
Wang, X., Lam, C. C., and Iu, V. P. (2018). Experimental investigation of in-plane shear behaviour
of grey clay brick masonry panels strengthened with SRG. Eng. Struct., 162:84–96.
186
BIBLIOGRAPHY
Wang, X., Lam, C. C., and Iu, V. P. (2020). Bond behaviour of steel-TRM composites for strengthening
masonry elements: Experimental testing and numerical modelling. Constr. Build. Mater., 253.
Wilding, B. V., Dolatshahi, K. M., and Beyer, K. (2018). Shear-compression tests of URM walls:
Various setups and their influence on experimental results. Eng. Struct., 156:472–479.
Yacila, J., Salsavilca, J., Tarque, N., and Camata, G. (2019). Experimental assessment of confined
masonry walls retrofitted with SRG under lateral cyclic loads. Eng. Struct., 199(April).
187

Seismic Retrofit of Masonry

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

Seismic Retrofit of Masonry

Uploaded by

Copyright:

Available Formats

PHDDissertation

Seismic retrofit of masonry

Doctoral Thesis by:

Barcelona, November 2020

Universitat Politècnica de Catalunya, Barcelona Tech

Doctorat en Enginyeria de la Construcció

Departament d´Enginyeria Civil i Ambiental

Doctoral Thesis submitted in fulfilment of the requirements for the

Barcelona, November 2020

Universitat Politècnica de Catalunya

Barcelona, 26 de Noviembre 2020

Table of Contents vii

List of Tables xvii

List of Symbols xxiv

4 Experimental Results: Diagonal Compression Test 77

5 Experimental Results: Shear Compression Test 91

7 Analytical Study 151

A Use of DIC in experimental testing of masonry walls 175

4.1 Crack patterns of URM specimens a) URM 1, b) URM 2 . . . . . . . . . . . . . . . . 78

5.17 Cumulative dissipated energy of tested specimens . . . . . . . . . . . . . . . . . . . . . 121

3.1 Mechanical properties of bricks used for he construction of walls . . . . . . . . . . . . 48

2θ Coordinate rotatio angle in Mohr’s Space

α Coefficient dependent on masonry typology

αt Coefficient of reduction due to shear

β Coefficient of correlation between τ0 and ft

βf d Coefficient dependent on the type of failure

∆x Height of the brick unit

∆y Width of the brick unit

∆0 Smallest deformation amplitude at which damage is first observed

δc Displacement at collapse corresponding to a drop of 50% of the lateral load

δHmax Displacement at maximum lateral load

δi+ Displacement in the positive direction for cycle i

δi− Displacement in the negative direction for cycle i

∆m Deformation at which the most severe damade is expected

∆m Deformation at which the most severe damage is expected

δy Displacement at yielding correponding to ultimate load

γ Shear strain at the centre of the wall according to ASTM standard

γu Ultimate shear strain

γy Yield shear strain

µ0 Friction coefficient of the mortar joint

ωn Natural frequency of vibration of a structure

σ0 Vertical compression stress

σII Minimum principal stress

σI Maximum principal stress

σu,f Ultimate tensile strength of the textile fibre

σv Designed compression stress of SCT

τ0,m,DCT Shear parameter of masonry obtained from DCT results

τ0,m,SCT Shear parameter of masonry obtained from SCT results

τ0 Cohesion of the mortar joint

τb Bond strength between the masonry and the mortar matrix

τmax Maximum shear stress attained during DCT

τy Shear stress corresponding to the yield shear strain

θcr Drift at onset of cracking

θHmax Drift when the maximum lateral load is attained

θN C Drift for EC8 NC limit state

θp Principal angle in Mohr’s Space

θSD Drift for EC8 SD limit state

θSLU Drift for NTC 2018 SLU limit state

θu Ultimate drift corresponding to the bilienar idealization

εc Strains along the shortening (compressed) diagonal of the wall

εf dd Maximum strain of reinforcement at which the debonding takes place

εf d Strain corresponding to the reinforcement tensile capacity

εf k Strain at failure of the reinforcement