Experimental assessment of confined masonry walls retrofitted with SRG under lateral cyclic loads Jhair Yacilaa,∗, Jhoselyn Salsavilcaa, Nicola Tarquea, Guido Camatab aPontifical Catholic University of Peru, Civil Engineering Division, Lima, Peru bUniversità degli Studi Gabriele D’Annunzio, Department of Engineering and Geology, Pescara, Italy Abstract Around the world, many informal masonry buildings have collapsed due to the failure of their bearing walls under lateral seismic loads. This is related to the many involved factors, such the quality of the materials, the quality of workmanship, the lack of tech- nical intervention, and the high seismicity of the zone, among others. However, the fact is that these constructions need to be retrofitted in order to upgrade their ultimate strength and allow them to properly absorb inelastic deformations. Currenly, fiber rein- forced polymer (FRP) has been widely studied as a retrofitting technique. However, it has some technical and economic disadvantages that are remedied by fiber reinforced mortar (FRM). In this paper, a variant of FRM known as steel reinforced grout (SRG) is studied as a seismic retrofitting technique for cracked confined masonry walls (CMW). For this purpose, three full-scale cracked walls were repaired, retrofitted with SRG strips, and tested under in-plane cyclic loads at the Pontifical Catholic University of Peru (PUCP). The experimental results show the benefits of SRG in improving the lateral displacement ductility, energy dissipation, and stiffness degradation of CMWs. Keywords: confined masonry wall, seismic retrofit, SRG, seismic vulnerability 1. Introduction Confined masonry is a type of construction widely diffused in Peru due to its easy and fast construction. According to [1], masonry dwellings represent 84% of the total ∗Corresponding author Email address: jhair.yacila@pucp.edu.pe (Jhair Yacila) Preprint submitted to Journal of LATEX Templates August 4, 2019 buildings in Peru and 60% of them were built informally. In the case of informal dwellings, seismic events have evidenced their high vulnerability, which has led to5 human and material loss (i.e. Lima, 1746; Arequipa, 2001; Pisco, 2007). Therefore, there is a necessity to reinforce a huge quantity of confined masonry buildings in order to improve their seismic performance. The main aim of seismic retrofitting is to upgrade the ultimate strength of the build- ing by improving the structure’s ability to absorb inelastic deformations [2]. In this10 way, external reinforcement by using composite materials has arisen as an efficient method due to its advantages, such as its facility of application, high stress/weight ra- tio, and versatility, which means it is applicable to different types of substrates. In this way, one of the most commercial composites is the well-known fiber reinforced polymer (FRP). This composite consists of different kinds of fibers (e.g. carbon, glass,15 basalt, and others), or high strength textiles and an organic binder (e.g. epoxy resin). Different studies have been conducted to assess the effectiveness of the FRP when retrofitting masonry walls. For instance, [2, 3, 4] demonstrated that FRP can signifi- cantly increase the strength, ductility and energy absorption capacity of the masonry walls. In addition, FRP are cataloged as convenient because it does not add mass to20 the structure, it is easy to handle, flexible, quick to install, and have an excellent per- formance in terms of tensile strength and durability [5, 6]. Nevertheless, FRP has also some disadvantages which are related to its inapplicability on wet surfaces, its poor performance at high temperatures and in alkaline environments, possible hazards for workers, the incompatibility of the resin with the substrate materials, and the lack of25 water vapor permeability [7, 8, 9]. On the other hand, reinforcement systems based on cement (fiber reinforced cementitious matrix, FRCM) and mortar (fiber reinforced mortar, FRM, or textile reinforced mortar, TRM), have arisen to overcome these draw- backs. Furthermore, they are reversible (i.e. they can be removed from the surface without major damage) and are not architecturally invasive since the thickness of the30 intervention can be 10 mm or less. A TRM composite can comprise different types of fiber: basalt, glass, carbon, polyparaphenylene benzobisoxazole (PBO), aramid, among others. A particular case of TRM is given when the steel fiber or textile is used into the composite because it 2 is more commonly known as steel reinforced grout (SRG), which has already been35 experimentally studied as a retrofitting technique for concrete and masonry structures [10, 11, 12]. This composite is of great interest due to its mechanical properties and bond efficiency. For instance, steel fibers have a high tensile strength, a higher stiffness than basalt or glass, and less thickness than carbon and aramid. These particular fibers are less fragile than the others because a ductile behaviour is exhibited before tensile40 failure. Furthermore, due to the zinc coating, steel textiles do not get rusty and are protected from sulphate attacks [13]. However, their application is recent compared to other fibers. Hence, knowledge about their design, construction and modeling is still limited. Regarding the application of SRG as retrofitting technique of confined masonry45 walls, it has not yet been studied, which represents a gap in the search of economic alternatives which can help to reduce the seismic vulnerability of confined masonry buildings. This paper presents a criterion for designing, as well as the application process for retrofitting, confined masonry walls with SRG to support in-plane shear loads. Hopefully this will contribute to the formation of guidelines for their design and50 application. For the design process, the CNR and AC434 guidelines have been taken into account [14, 15]. Although the CNR guideline is focused on FRP, it has been assumed that it is also applicable to SRG since similar design concepts were applied. The construction process for retrofitting confined masonry walls with SRG, as well as the considerations to be taken into account during the experimental campaign, are55 also discussed. The experimental results are presented and discussed in terms of lateral deformation, energy dissipation, hysteresis damping, and stiffness degradation. 2. Previous Work 2.1. Description of the tested walls In the experimental campaign conducted by [16, 17], a total of 9 full-scale confined60 masonry walls were built and tested under cyclic lateral loads at the PUCP. These walls were built with king kong bricks of 18 holes with a net area less than 60% of the gross area. These are industrial bricks, with dimensions 230 x 130 x 90 mm, and are 3 commonly used for bearing walls in Peru, although prohibited by the Peruvian Seismic Code in the coastal area [18]. The construction process of a typical confined masonry65 wall implies that steel reinforcing is located in its final position before constructing the masonry panel. It is not recommended to lay the bricks more than 1.3 m high per working day, in order to avoid crushing the mortar in the lower part of the panel. The typical mortar thickness, either horizontal or vertical, varies between 10 and 15 mm and has a typical volumetric ratio cement/sand of 1/4. Moreover, an intentional70 toothed finish is left for a subsequent concrete casting. In this way, it is intended to guarantee a monolithic union between concrete and masonry. The dimensions of a typical confined masonry wall as well as the reinforcing detail are shown in Fig. 1. Figure 1: Geometry and reinforcement detail of a typical wall (dimensions in millimeters) Regarding the detail of reinforcement of the confining frames, it should be taken into account that in confined masonry construction, these elements are not designed to75 act as moment-resisting frames. As a result, detailing of reinforcement is simple [19]. In fact, a common practice in Peruvian Constructions is to use corrugated steel rods of φ1/2” as longitudinal reinforcement and φ1/4” as transverse stirrups. Fig. 1 shows the dimensions and reinforcing detail for a typical tested wall. Regarding the experiments, in order to obtain reliable outcomes regarding the en-80 ergy dissipation, strength and stiffness degradation of masonry walls, the cyclic char- 4 acter of seismic loads should be evaluated. Alternatively, the experiments could be conducted by using a shaking table where any characteristic seismic signal register can be imposed, otherwise, a cyclic test where incremental lateral displacements are im- posed slowly can be used as it was demonstrated by the various experimental studies85 conducted by [20]. Regarding the tests conducted by [16, 17], three walls were tested under only lateral loads, until a repairable limit state was reached, i.e. equivalent to a drift of 0.125% according to the Peruvian Code [18]. The next three, unlike the previous ones, were tested until a collapsed state was reached, which had a drift of 0.833%. Finally, the last90 three had a constant vertical load of 170 kN, which represents the vertical load on a first floor wall from a total of three, and were tested also until reaching a collapsed state, with an associated drift of 0.625%. In this research, a total of three cracked walls were selected from the previous research to be repaired and tested under lateral cyclic loads again. Within this selection, two walls were selected from the group of walls which95 were led up to a collapse state (drift = 0.833%), whereas the last one was selected from the group of walls which were tested with a vertical load. It should pointed out that the walls were randomly selected for representing any cracked wall which could be located at the first floor of a masonry building. 2.2. Characterization of the materials100 In order to characterize the properties of the materials involved in the walls, control tests were carried out at the PUCP [16, 17]. To characterize the mortar employed for the masonry panel, 12 cubic samples with dimensions 50 x 50 x 50 mm were extracted and tested under uni-axial compression. To assess the compressive behaviour and elas- tic modulus of the masonry, 4 masonry prisms with dimensions 230 x 130 x 600 mm105 were made and tested under uni-axial compression parallel to the largest dimension. In the case of a tensile behaviour of the masonry, 4 masonry walls with dimensions of 600 x 600 x 130 mm were made and tested under uni-axial compression parallel to the diagonal of each square sample (diagonal compression test). The concrete compressive strength of the foundation and confining elements was evaluated through compression110 tests of cylindrical specimens 150 mm wide and 300 mm high. For this job, 4 samples 5 (a) With vertical load W-01 (b) without vertical load W-02 (c) without vertical load W-03 Figure 2: Selected walls for retrofitting 6 were extracted from each concrete element. All specimens were properly cured for 28 days before testing. Table 1 shows the average results obtained from this control cam- paign. It is worth commenting that the concrete tensile strength and Young’s modulus were computed in accordance with the CEB-FIP model code [21].115 Table 1: Material properties involved in the walls Material Compressive Tensile Elastic strength strength modulus [MPa] [MPa] [MPa] Mortar 17.50 - - Concrete foundation 27.50 2.18 25900 Concrete columns 19.00 1.49 22500 Concrete beam 28.00 2.22 26200 Masonry 10.00 1.40 5700 3. Steel Reinforced Grout (SRG) In case of multi-storey confined masonry buildings, past earthquakes and findings of experimental studies have demonstrated the critical demand of lateral forces induced to the ground floor level, which cause significant shear cracking, and which in turn may caused the collapse of the building due to a soft story effect [22]. For this reason, in this120 paper, the performance of the confined masonry walls under lateral loads is intended to be improved by means of a novel retrofitting technique called SRG. SRG is composed of ultra-high tensile strength steel fibers 100 mm wide, 0.084 mm thick, and a natural lime mortar 100 mm wide and 10 mm thick. These fibers are uni-directional since the they result from twisting two wires around three straight125 wires. However, they are connected by perpendicular glass fiber filaments, therefore they can also be considered as textiles. As a previous step, the steel wires were coated with zinc before twisting, to protect them against corrosion [23]. The natural lime mortar employed had a M15 resistance class according to EN 998-2 and R1 according to EN 1504-3, as technical specification [24]. Furthermore, it is highly breathable, it130 is made strictly from natural and recycled minerals, and its manufacture produces very 7 low emissions of CO2 and other volatile organic substances. All these properties make it part of the innovative GreenBuilding technology. 3.1. Control tests In order to characterize the material properties involved in SRG, control tests were135 carried out. For instance, to characterize the natural lime mortar which serve as a binder for the steel galvanized fiber, 6 samples of 50 x 50 x 50 mm were made and tested after 28 days of curing, under uni-axial compression (Fig. 3a). Table 2 shows the experimental results from these tests, where the length and width resulted from averaging parallel dimensions from the face subjected to an axial load.140 In the case of the galvanized steel fiber, 5 samples comprised of steel textile and steel plates which were joined by an epoxy resin, were made and tested, once the epoxy resin was totally dry (after 1 day), under uni-axial tension (Fig. 3b). In these tests, two main failure modes were recognized: one was related to the failure at the union between the textile and steel plates (U), whereas the other one was related to the145 middle part of the textile (M). It is worth noting that in these tests, the second failure mode is expected, in order to obtain a representative strength of the textile, since the first mode is linked to the participation of the steel plates in the failure. Table 3 shows the experimental results related to these tests, where E f is the Young’s modulus, f f is the maximum strength, and ε f u is the maximum strain of the steel mesh.150 Finally, the interaction between SRG and masonry substrate under shear loads was explored through 5 debonding tests, as shown in Fig. 3c, which were carried out after 28 days of curing. In these tests, 5 main failure modes were recognized: (1) rupture of the masonry substrate (2) debonding at the mortar-to-substrate interface, (3) debonding at the textile-to-mortar interface, (4) premature cracking of the outer mortar layer, and155 (5) rupture of the textile. Table 4 shows the experimental results related to these tests, where fb is the maximum tensile stress developed by the steel mesh, τ is the stress computed as the relation between the maximum force and the cross sectional area of the SRG, and Slip is the relative displacement of the SRG prior to total failure. For computing the Slip, two control points were located and connected by means of an160 LVDT into the specimens as is shown in Fig. 3. This LVDT measured a relative 8 displacement, δ, between the two control points, which was used to compute the Slip by means of the difference of δ and the elastic deformation of the mesh steel, Slip = δ−ε f L. In this expression, ε f is computed as the relation of the applied load prior total failure of the SRG, P, and the cross area of the steel mesh, A f , multiplied by the elastic165 modulus of the steel mesh, E f , namely, ε f = P/(A f Es). A major detail of the control tests herein discussed is presented by [25]. Table 2: Experimental results from compressive tests Specimen Length Width Maximum Load Stress [mm] [mm] [kN] [MPa] M-01 50.55 50.90 57.75 22.44 M-02 51.35 50.97 60.79 23.22 M-03 51.22 50.95 58.98 22.60 M-04 51.40 51.07 62.62 23.85 M-05 51.32 50.75 57.59 22.11 M-06 51.22 50.80 60.71 23.33 Average 51.18 50.91 59.74 22.93 CV [%] 0.62 0.23 3.31 2.83 Table 3: Experimental results from tensile tests Specimen E f f f ε f u Failure [GPa] [MPa] [%] mode F-01 160 2786 2.18 U F-02 161 2893 2.55 M F-03 155 2879 2.50 M F-04 155 2859 2.50 U F-05 153 2886 2.45 M Average 157 2861 2.44 - CV [%] 2.23 1.52 6.05 - 9 (a) Compressive test (b) Tensile test (c) Debonding test Figure 3: Testing setup for control specimens (dimensions in millimeters) 10 Table 4: Experimental results from debonding tests. Failure modes: (1) masonry substrate, (2) mortar-to- substrate interface, (3) textile-to-mortar interface, (4) outer mortar layer, and (5) textile Specimen fb τ Slip Failure [MPa] [MPa] [mm] mode D-01 2068 0.66 3.50 2,4,5 D-02 1793 0.57 1.20 1 D-03 1612 0.51 2.40 3 D-04 1191 0.38 1.90 3 D-05 2023 0.64 2.20 3,4 Average 1737 0.55 2.24 - CV [%] 20.52 20.47 37.43 - 3.2. Design of SRG reinforcement for shear behaviour enhancement Before retrofitting, it is necessary to properly design the reinforcement in order to minimize the costs in materials and workmanship. For this purpose, some design170 concepts were extracted from the Peruvian Code, CNR-DT and AC434 [18, 14, 15], as explained below. Regarding the adopted reinforcement scheme, CNR-DT recommends horizontal strips when a shear reinforcement is required. The nominal shear resistance of a retrofitted confined masonry wall can be evaluated as the sum of the contributions from the masonry wall and the reinforcement: φvVn = φv (Vm +Vf ) (1) where φv is the strength reduction factor for Load and Resistance Factor Design method (LRFD), taken as 0.75 for shear loads; Vm is the shear contribution of the masonry; and Vf is the shear contribution of the reinforcement. Regarding Vm, it should be evaluated according to local code. For instance, in this work it was evaluated according to the Peruvian Code: Vm = 0.5 ·ν′m ·α · t ·L+0.23 ·Pg (2) where ν′m is the characteristic shear strength of the masonry, α is a wall slenderness factor correction, t is the wall’s thickness, L is the wall’s length, and Pg is the contri- bution of the vertical load to the shear resistance. It is worth noting that Eq. 2 refers to175 11 the shear resistance of a new wall, therefore, an appropriate reduction factor must be employed to take into account the reduced contribution of a damaged wall. Regarding the shear contribution of one reinforcing strip, it can be evaluated as Vf = 1 γ ·0.6 ·d · f f v ·2 · A f v p f v (3) where γ is a partial factor, taken as 1.2 for shear loads, d is the distance between the end of the fiber in compression and the centroid of the opposite confinement column, f f v is the SRG design tensile strength, which can be calculated as f f v = E f ε f v, E f is the tensile modulus of elasticity of the cracked SRG, ε f v is the SRG tensile design strain, which can be taken equal to the ultimate strain of steel textile but not greater than 0.004, ε f v = ε f u ≤ 0.004, A f v is the area of one steel textile branch, and p f v is the separation between strips. Finally, the amount of strips required can be calculated as n = 1 Vf ( Vu φv −Vm ) (4) In the present work, for design purposes, Vu was considered as the maximum shear strength recorded by the hysteretic curves, which meant to assume that the SRG was able to recover the maximum strength in conjunction with the contribution of the re-180 paired walls. According to Peruvian Code [18], Vm is a theoretical value which repre- sent the shear force needed in the wall to produce the first shear crack in the masonry panel. In case of the original walls, Vm was computed using Eq. 2 as 250 kN and 211 kN for the walls with and without vertical load, respectively. However, taking into account that not all the cracks were repaired but only some of them, it was necessary185 to consider that the shear contribution of masonry would be a reduce part of Vm. Ini- tially, 75%Vm was taken as the shear contribution of the repaired masonry for design purposes. Nevertheless, an additional SRG strip was provided for considering that the assume percentage of 75% could be less. Experimentally, the hysteretic behaviour of the walls and what was observed during their tests showed that the lateral force which190 produced the first shear crack in the masonry panel of the retrofitted walls was approx- imately 50% of that recorded by the original walls. Therefore, for design purposes, an average value of 50%Vm could be considered as adequate for the shear contribution of the repaired masonry, where Vm is computed with the Eq. 2. 12 From the geometry of a typical tested wall (Fig. 1), it was possible to deduce195 d = 2500 mm. Regarding SRG’s properties, it was assumed that once the SRG is cracked, the tensile behaviour of the composite is governed by the steel textile. For this reason, E f was taken to be the Young’s modulus of the steel mesh: E f = 150 GPa (Ta- ble 3). Due to the fact that ε f u was greater than 0.004 in all tests (Table 3), ε f v = 0.004 was adopted. Regarding A f v, it was assumed that the strips were 100 mm wide and that200 there were 16 steel cords, which resulted in A f v = 8.6 mm2 according to the manufac- turer’s data sheet [23]. Finally, 450 mm was assumed as the separation between strips for design purposes. Fig. 4 shows the reinforcing scheme adopted for the present work. Figure 4: Details of the reinforcement for the repaired walls (dimensions in millimeters) It should be noted that in Fig. 4, less separation between the strips was assumed at the mid-height of the walls. This assumption was related to the fact that the largest205 number of cracks were concentrated in the mid-height of the walls. 3.3. Procedure for retrofitting confined masonry walls with SRG Before retrofitting, it is necessary to make a proper repair of the cracks since a good repair improves the recovery of the initial stiffness. In this study, cracks greater than 8 mm were opened using hand tools in order to avoid excessive out-of-plane effects. In210 the case of crushed bricks, it is recommended to replace them by new ones (Fig. 5b). Thereafter, the openings were filled with reparation mortar based on Portland cement with a volumetric ratio cement/sand = 1/3. After being repaired, the walls should be 13 properly cured for at least 28 days. However, taking into account what would be done in a massive application, they were moistened three times a day for seven days. In this215 way, it was hoped to guarantee a reasonable resistance of the reparation mortar. Fig. 5 shows the main steps involved in repairing CMW. (a) Cracks opening (b) Filling of openings (c) Curing process Figure 5: Main steps for repairing CMW (W-01) Regarding retrofitting, there are two previous jobs needed for the proper prepa- ration of the zone of intervention. The first one is related to the fact that additional roughness can be provided by punching the bricks lightly by means of pointed tools.220 The second one consists in delimiting the intervention zone by means of Scotch tape. Although these jobs are not obligatory, they allow providing better adhesion between the SRG and the masonry substrate, as well as saving on material by using only what is necessary. 14 The retrofitting process started by moistening the intervention zone in order to avoid225 the absorption of the SRG’s water by the masonry. Then, a first layer of mortar, 5 mm thick, was laid upon the masonry within the area delimited by the Scotch tape. Subsequently, the steel mesh was embedded lightly inside the first layer of mortar. Thereafter, a second layer of mortar 5 mm thick was laid in order to finish covering the embedded steel mesh. Finally, once all the SRG strips were finished, the Scotch tape230 was removed to start the curing process. Fig. 6 shows the main stages of the retrofitting process as explained above. It is worth noting that, in this work, it was possible to anchor the steel mesh by overlapping them 250 mm interspersed at each column’s ending, because it only had to reinforce isolated walls. However, for other applications, a proper anchor for the steel mesh must be previously studied or applied according to235 the manufacturer’s recommendations, in order to guarantee a good transmission of the stresses from the masonry to the SRG. Like reparation mortar, an SRG composite needs a proper curing process of at least 28 days. However, again taking into account what would be done in a massive application, the SRG composite was moistened for 14 days to guarantee a good mortar strength before testing.240 3.4. Boundary conditions and instrumentation for the cyclic tests Before testing, each foundation end was fixed to a reaction slab by means of hy- draulic jacks to restrict them vertically. Another hydraulic jack and a reaction frame were used as rigid horizontal stops also for the foundation ends. A vertical load of 170 kN was applied by another hydraulic jack through two rigid steel beams in order245 to distribute the vertical load along the confinement beam. Regarding the horizontal cyclic loads, they were applied at the top of the wall by means of a dynamic actuator which was controlled by a computer. Regarding the instrumentation, linear variable differential transformers (LVDTs) were placed, as shown in Fig. 7. Two LVDTs (LVDT 1 and 2) were placed along250 the diagonals of the masonry panel to measure their deformations and thus to have enough data in case an idealized strut-and-tie model is carried out. Another two LVDTs (LVDT 3 and 4) were placed at the confinement columns to measure their deformations due to vertical loads and bending effects during the cyclic test. One LVDT (LVDT 5) 15 (a) First layer of mortar (b) Placement of steel fiber mesh (c) Second layer of mortar (d) Retrofitted wall Figure 6: Stages for strengthening CMW with SRG (W-02) 16 was placed between the geometric centre of the confining beam and a reaction frame,255 which was assumed to be static. Fig. 7 shows the general testing setup as well as the instrumentation scheme for each cyclic test. Figure 7: Setup and instrumentation for cycling test The cyclic loading was controlled by displacements, which means that the dynamic actuators applied displacements instead of forces. However, this had an internal load cell that allowed registering the load related to each displacement, thus it was possible260 to plot the corresponding hysteretic behaviour. In order to avoid kinematic effects, a quasi-static test was intended to be carried out by applying an average velocity of 0.25 cycles/minute. Regarding the applied history of displacements, it was defined accord- ing to FEMA 461 [26]. Thereby, each level of displacement resulted from increasing the previous level of displacement by a factor of 1.4. In addition, two cycles were also265 defined for each displacement level. It is worth highlighting that in the previous work [16, 17], only 11 displacement phases were considered, with a maximum displacement level of 20 mm, whereas in this work, 12 displacement phases have been taken into account, with a maximum displacement level of 30 mm, as shown in Fig. 8. 17 0.5 1.0 1.4 2.0 2.8 3.9 5.5 7.7 10.8 15.0 20.0 30.0 -35 -25 -15 -5 5 15 25 35 0 10 20 30 40 50 60 70 80 90 100 D is pl ac em en t, [m m ] Time, [min] Phase 1 Phase 2 Phase 3 Phase 4 Phase 5 Phase 6 Phase 7 Phase 8 Phase 9 Phase 10 Phase 11 Phase 12 Figure 8: History of displacements 4. Discussion of the results270 4.1. Cracking pattern Within the category of confined masonry buildings, masonry walls act as bearing elements. Therefore, a premature failure of these walls would result in the building’s collapse. In performance design, it is intended that the building has a specific perfor- mance level, which in turn is related to the post-earthquake disposition of the building.275 These performance levels can be roughly classified and assigned to certain levels of drift: (1) Immediate Occupancy (IO) – drift = 0.3%, (2) Life Safety (LS) – drift = 0.6%, and (3) Collapse Prevention (CP) – drift = 1.0%, as it was established by FEMA 356 after evaluating a huge quantity of experimental studies referred to masonry walls [27]. Fig. 9 shows the final cracking pattern for all the tested walls, whereas Table 6280 shows the evolution of cracking according to each performance level. In general, all the retrofitted walls showed bending cracks at the columns’ feet prior to a drift of 0.12%, which corresponds to the fifth loading phase. Subsequently, the cracks that were previously repaired started opening. However, it should be noted that not only the repaired cracks were opened during the tests, on the contrary, additional285 cracks took place. This effect can be noted in Table 6 by comparing the cracking pattern of retrofitted walls with those un-retrofitted. In addition, it is worth also noting that the tested walls suffered a mixed failure mode, namely, they started having cracks due to bending effects but finished having cracks also due to shear effects. 18 (a) RW-01 (b) RW-02 (c) RW-03 Figure 9: Cracking pattern of tested walls 19 Concerning the in-plane seismic behavior of the confined masonry walls, initially,290 the masonry panel resists by itself all the effects caused by the lateral forces, whereas the confining elements do not play a significant role. Nevertheless, once the masonry panel is cracked, the vertical reinforcement in confining columns start engaging in resisting tension and compression stresses [19]. This fact explained why and how con- fining elements play an important role into the lateral displacement ductility capacity295 of the confined masonry walls. Therefore, it is also important to assess the cracking evolution of these elements. 4.1.1. RW-01 During the fifth loading phase (drift = 0.12%), the first visible cracks occurred at the columns’ feet because of bending effects. Thereafter, progressive bending cracks began300 to appear along the column’s height. These cracks were produced by the controlled elongation of the confining columns, given by the bond-slip effect present in the RC frames, during each loading phase [28]. It should be taken into account that whereas the steel reinforcement is able to carry tensile stresses, the concrete and masonry are able to carry compressive stresses and the wall is still stable, the confining frames305 will provide lateral displacement ductility to the wall. By comparing the reinforced wall with the original one, even when the cracks in the columns were not repaired, it could be observed that no extra cracks were significantly produced in these elements (Table 6). This is related to the fact that during the test, the confining elements acted as cracked elements, therefore, the cracks opened during the test were almost the same as310 those that were not repaired. Regarding the masonry panel, the first diagonal crack produced by shear effects took place at the seventh loading phase (drift = 0.23%). It should be noted that as the displacements increased, additional cracks due to shear effects occurred. Instead of RW-02 and 03, in this wall it was possible to observe the rupture of two SRG strips at315 the mid-height of the wall, which demonstrated that all the strength of the SRG could be developed. Nevertheless, it is worth mentioning that the ruptures were characterized by the breakage of the steel meshes and not by debonding failure, which demonstrated a perfect adhesion among the SRG’s mortar and masonry substrate. Similarly, other 20 SRG strips showed elongations of the steel mesh, which could be observed because of320 the detachment of the external layer of the mortar. 4.1.2. RW-02 & RW-03 Like RW-01, the first visible cracks in these walls occurred at the columns’ feet due to bending effects. However, unlike RW-01, they took place at the fourth loading phase (drift = 0.083%). With respect to the cracking of the confining columns, like RW-325 01, as the displacements increased, the number of cracks in height increased as well. Furthermore, it could also be noted that no extra cracks were significantly produced in the confining columns (Table 6). Regarding the masonry panel, the first diagonal crack produced by shear effects took place at the eleventh (drift = 0.833%) and ninth (drift = 0.45%) loading phase,330 respectively. In addition, it should be noted that as the displacements increased, addi- tional cracks due to shear effects took place. Unfortunately, in these walls it was not possible to develop the total tensile strength of the SRG strips. Nevertheless, it should be highlighted that this means they were prepared to withstand more tensile stress than they were subjected to. In addition, it is worth noting that a horizontal crack took place335 in both cases at the base of the wall, to have a sort of rocking effect, as is shown by the shape of the hysteresis loops in Fig. 10. The results associated to the first cracking and maximum load capacity are summarized in Table 5. In all cases, the collapse state was governed by the instability of the walls or the abrupt loss in load capacity either within the first or second hysteresis loop of the340 corresponding loading phase. For instance, Fig. 10 (a) and (b) show only one cycle in the last loading phase, which was associated to the instability of the walls, whereas Fig. 10 (c) shows an abrupt loading loss in the second cycle of the last loading phase. Regarding the performance of the retrofitted walls, it is worth highlighting that only the retrofitted walls managed to prevent a collapse. Namely, the original walls345 were not able to withstand the drift level associated with collapse prevention (drift = 1.00%), whereas once repaired and retrofitted with SRG they were able to attain that level of performance (Fig. 11). In fact, this helps to reduce the risk of life-threatening injury, which is of great interest in seismic areas. 21 -400 -300 -200 -100 0 100 200 300 400 -40 -30 -20 -10 0 10 20 30 40 ] Nk[ ,daol laretaL Top Displacement, [mm] W-01 RW-01 (a) -400 -300 -200 -100 0 100 200 300 400 -40 -30 -20 -10 0 10 20 30 40 ] Nk[ ,daol laretaL Top Displacement, [mm] W-02 RW-02 (b) -400 -300 -200 -100 0 100 200 300 400 -40 -30 -20 -10 0 10 20 30 40 ] Nk[ ,daol laretaL Top Displacement, [mm] W-03 RW-03 (c) Figure 10: Hysteresis curves of tested walls 22 IO LS PCIOLSPC -1.25 -0.75 -0.25 0.25 0.75 1.25 0.00 0.20 0.40 0.60 0.80 -400 -300 -200 -100 0 100 200 300 400 -30.00 -20.00 -10.00 0.00 10.00 20.00 30.00 Drift, [%] S he ar s tr es s, [ M P a] L at er al f or ce , [ kN ] Lateral Displacement, [mm] W-01 RW-01 (a) IO LS PC IOLSPC -1.25 -0.75 -0.25 0.25 0.75 1.25 0.00 0.20 0.40 0.60 0.80 -400 -300 -200 -100 0 100 200 300 400 -30.00 -20.00 -10.00 0.00 10.00 20.00 30.00 Drift, [%] S he ar s tr es s, [ M P a] L at er al f or ce , [ kN ] Lateral Displacement, [mm] W-02 RW-02 (b) IO LS PC IOLSPC -1.25 -0.75 -0.25 0.25 0.75 1.25 0.00 0.20 0.40 0.60 0.80 -400 -300 -200 -100 0 100 200 300 400 -30.00 -20.00 -10.00 0.00 10.00 20.00 30.00 Drift, [%] Sh ea r st re ss , [ M Pa ] L at er al f or ce , [ kN ] Lateral Displacement, [mm] W-03 RW-03 (c) Figure 11: Envelope curves of tested walls 23 Table 5: Experimental results of tested walls Specimen Direction First cracking Maximum load Load [kN] Drift [%] Load [kN] Drift [%] RW-01 Push 195 0.116 340 0.833 Pull −180 −0.116 −340 −0.833 RW-02 Push 90 0.083 235 0.833 Pull −120 −0.116 −285 −0.833 RW-03 Push 95 0.083 255 0.625 Pull −125 −0.116 −205 −0.625 4.2. Lateral displacement ductility350 As mentioned above, the original walls were not able to reach the performance level of collapse prevention, which means they did not have enough ductility to resist lateral forces while maintaining their stability. Taking into account the final lateral displace- ment reached by each wall, δu, and the displacement related to the first cracking of the masonry panel, δy, the ductility was evaluated according to Eq. 5. In order to evaluate355 δy, backbone curves were traced from the envelope curves shown in Fig. 11, by follow- ing the recommendations of [29]. These backbone curves were drawn by highlighting three main point: (1) first cracking, (2) maximum strength, and (3) ultimate state, as is shown in Fig. 12. µ = δu δy (5) Table 7 shows a summary of the calculation of the ductility developed for each360 tested wall. It has to be noted that in the case of RW-01 the increment in ductility was about 100%, whereas in the rest it was about 50%. It also has to be pointed out that the original walls withstood higher forces than the retrofitted ones in the first performance level (IO), as can be seen in Fig. 11. However, for the second performance level (LS), both the original and retrofitted walls showed almost the same strength. Finally,365 the retrofitted walls, besides being the only ones which could reach the last desired performance level (PC), were able to withstand almost the same level of lateral forces 24 Table 6: Cracking evolution of tested walls (a) Final state [W-01] (b) IO [RW-01] (c) LS [RW-01] (d) CP [RW-01] (e) Final state [W-02] (f) IO [RW-02] (g) LS [RW-02] (h) CP [RW-02] (i) Final state [W-03] (j) IO [RW-03] (k) LS [RW-03] (l) CP [RW-03] 25 as the LS. This means that after the LS performance level, the retrofitted walls could continue withstanding forces by maintaining their stability. -1.25 -0.75 -0.25 0.25 0.75 1.25 0.00 0.20 0.40 0.60 0.80 -400 -300 -200 -100 0 100 200 300 400 -30.00 -20.00 -10.00 0.00 10.00 20.00 30.00 Drift, [%] S he ar s tr es s, [ M P a] L at er al f or ce , [ kN ] Lateral Displacement, [mm] W-01 RW-01 (a) with vertical load -1.25 -0.75 -0.25 0.25 0.75 1.25 0.00 0.20 0.40 0.60 0.80 -400 -300 -200 -100 0 100 200 300 400 -30.00 -20.00 -10.00 0.00 10.00 20.00 30.00 Drift, [%] Sh ea r st re ss , [ M P a] L at er al f or ce , [ kN ] Lateral Displacement, [mm] W-02 RW-02 W-03 RW-03 (b) without vertical load Figure 12: Backbone curves of tested walls 4.3. Energy dissipation and damping ratio370 The Ultimate Limit State (ULS) design approach considers the maximum strength that could be withstood by the structural elements. This is considering both linear and non-linear behaviour of the elements, either due to material or geometric non-linearity. Whenever non-linear behaviour takes place, inelastic strains are experienced, which produce damage in the structural elements. It should be noted that the more inelastic375 strains there are, the more structural damage will take place. During this process, a 26 Table 7: Improved ductility calculation of tested walls Specimen δy W RW Increment W / RW [mm] δu [mm] µ δu [mm] µ [%] 01 2.8 15.0 5.35 30.0 10.7 100 02 2.8 20.0 7.14 30.0 10.7 50 03 2.8 20.0 7.14 30.0 10.7 50 certain amount of energy absorption and dissipation is involved. In this section, the energy dissipation, Ed , will be evaluated by the area within each hysteretic loop, as shown in Fig. 13. Additionally, the equivalent hysteretic viscous damping, ξhyst , is evaluated as the ratio between the dissipated energy and the elastic strain energy, as380 shown in Fig. 13. Figure 13: Calculation of energy dissipation and damping ratio Fig. 14 shows the cumulative energy dissipated for each loading phase. It has to be noted that the SRG gave the original walls the ability to dissipate more energy by means of greater inelastic displacements. In general terms, both the original and retrofitted walls had almost the same cumulative energy dissipation until the maximum385 displacement of the original walls. However, a freak tendency could be observed in the last loading phase of W-01, where an abrupt increment of energy dissipated was captured, as is shown in Fig. 14. This is related to the fact that in this loading phase an abrupt loss of capacity load (Fig. 10a) was registered, which in turn resulted in a quite large hysteretic loop. A similar phenomenon was registered for RW-01. However, this390 took place for a displacement that corresponded to the double of its original wall and 27 this occurred only in the pulling branch of the hysteresis loop (Fig. 10a). During the cyclic tests, it was possible to note that there were areas enclosed by the loops of the first two loading phases when they were expected to have no areas for corresponding to a linear-elastic behaviour. This anomalous behaviour was related to395 the fact that, like any mechanical equipment, the dynamic actuator needed certain small displacements for being calibrated. Therefore, to take into account this assumption, the first two loading phases were excluded from the calculation of the average hysteresis damping. Fig. 15 shows the variation of the hysteresis damping along the incremental loading phases for each tested wall. It is worth noting that the retrofitted walls had400 greater values of hysteresis damping throughout the tests. Nevertheless, the freak en- ergy dissipation just mentioned also affected the calculation of the hysteresis damping in the last loading phase of W-01. For this reason, this value was also excluded from the computation of the average hysteresis damping. Once this assumption is made, it is possible to note that RW-01 showed an average hysteresis damping of 9.65% against405 the 7.90% of W-01, which means an increment of 20%. RW-02 showed an average hysteresis damping of 12.00% against the 10.50% of W-02, which means an increment of 14%. Finally, RW-03 showed an average hysteresis damping of 12.45% against the 9.90% of W-03, which mean an increment of 26%. 4.4. Initial stiffness and stiffness degradation410 Fig. 16 shows the stiffness degradation of the tested walls. The same effect of the first two hysteresis loops is shown when computing the initial stiffness. Therefore, the initial stiffness was considered as that related to the third loop. It is important to highlight that the recovery of the initial stiffness will be as good as the goodness of the repair. On the other hand, since one aim of this work is to show that repairing415 and retrofitting with SRG can be done in an easy and massive way, this work tried to reproduce an effective and economic repair job, as explained above. This resulted in a recovery of 75% of the original wall’s initial stiffness for RW-01 and 50% for RW-02 and 03. Regarding stiffness degradation, it is known that it can be the result of cracking,420 crushing, rebar buckling, cracks opening and closing, among other factors. Likewise, 28 IO LS CP 0.00 0.25 0.50 0.75 1.00 1.25 0 5 10 15 20 25 30 35 40 45 50 55 0 3 6 9 12 15 18 21 24 27 30 Drift, [%] C um ul at iv e en er gy d is si pa ti on , [ kN -m ] Lateral Displacement, [mm] W-01 RW-01 (a) IO LS CP 0.00 0.25 0.50 0.75 1.00 1.25 0 5 10 15 20 25 30 35 40 45 50 55 0 3 6 9 12 15 18 21 24 27 30 Drift, [%] C um ul at iv e en er gy d is si pa ti on , [ kN -m ] Lateral Displacement, [mm] W-02 RW-02 (b) IO LS CP 0.00 0.25 0.50 0.75 1.00 1.25 0 5 10 15 20 25 30 35 40 45 50 55 0 3 6 9 12 15 18 21 24 27 30 Drift, [%] C um ul at iv e en er gy d is si pa tio n, [ kN -m ] Lateral Displacement, [mm] W-03 RW-03 (c) Figure 14: Cumulative energy dissipation for tested walls 29 IO LS CP 0.00 0.25 0.50 0.75 1.00 1.25 0 5 10 15 20 25 30 35 40 45 50 0 3 6 9 12 15 18 21 24 27 30 Drift, [%] H ys te re si s da m pi ng , [ % ] Lateral Displacement, [mm] W-01 RW-01 (a) IO LS CP 0.00 0.25 0.50 0.75 1.00 1.25 0 5 10 15 20 25 30 35 40 45 50 0 3 6 9 12 15 18 21 24 27 30 Drift, [%] H ys te re si s da m pi ng , [ % ] Lateral Displacement, [mm] W-02 RW-02 (b) IO LS CP 0.00 0.25 0.50 0.75 1.00 1.25 0 5 10 15 20 25 30 35 40 45 50 0 3 6 9 12 15 18 21 24 27 30 Drift, [%] H ys te re si s da m pi ng , [ % ] Lateral Displacement, [mm] W-03 RW-03 (c) Figure 15: Hysteretic damping ratio for tested walls 30 IO LS CP 0.00 0.25 0.50 0.75 1.00 0 20 40 60 80 100 120 140 160 180 200 0 3 6 9 12 15 18 21 24 27 30 Drift, [%] L at er al s tif fn es s, [ kN /m m ] Lateral Displacement, [mm] W-01 (+) RW-01 (+) (a) IO LS CP 0.00 0.25 0.50 0.75 1.00 1.25 0 20 40 60 80 100 120 140 160 180 200 0 3 6 9 12 15 18 21 24 27 30 Drift, [%] L at er al s tif fn es s, [ kN /m m ] Lateral Displacement, [mm] W-01 (-) RW-01 (-) (b) IO LS CP 0.00 0.25 0.50 0.75 1.00 1.25 1.50 0 20 40 60 80 100 120 140 160 180 200 0 3 6 9 12 15 18 21 24 27 30 Drift, [%] L at er al s tif fn es s, [ kN /m m ] Lateral Displacement, [mm] W-02 (+) RW-02 (+) (c) IO LS CP 0.00 0.25 0.50 0.75 1.00 1.25 1.50 0 20 40 60 80 100 120 140 160 180 200 0 3 6 9 12 15 18 21 24 27 30 Drift, [%] L at er al s tif fn es s, [ kN /m m ] Lateral Displacement, [mm] W-02 (-) RW-02 (-) (d) IO LS CP 0.00 0.25 0.50 0.75 1.00 1.25 1.50 0 20 40 60 80 100 120 140 160 180 200 0 3 6 9 12 15 18 21 24 27 30 Drift, [%] L at er al s tif fn es s, [ kN /m m ] Lateral Displacement, [mm] W-03 (+) RW-03 (+) (e) IO LS CP 0.00 0.25 0.50 0.75 1.00 1.25 1.50 0 20 40 60 80 100 120 140 160 180 200 0 3 6 9 12 15 18 21 24 27 30 Drift, [%] L at er al s tif fn es s, [ kN /m m ] Lateral Displacement, [mm] W-03 (-) RW-03 (-) (f) Figure 16: Stiffness degradation for tested walls 31 the level of stiffness degradation is related to the features of the structure (e.g. ma- terial properties, geometry, connection types), as well as to the loading history (e.g. displacement level for each loading phase, number of cycles per phase, increment ra- tio of displacements) [30]. The stiffness degradation is very helpful for design codes425 since it allows them to define the drifts according to the expected performance levels. In this case, following the three main desired performance levels (IO, LS and CP), the stiffness decay of the initial stiffness of the retrofitted walls was evaluated for each of these states, as described in Table 8. Table 8: Percentages of stiffness at performance levels regarding initial stiffness Wall Immediate Occupancy Life Safety Collapse Prevention (IO) (LS) (CP) W-01 35% 12% - RW-01 40% 25% 15% W-02 32% 16% - RW-02 47% 32% 20% W-03 32% 16% - RW-03 48% 30% 16% The vertical load applied to W/RW-01 gives them more stiffness which is clearly430 evidenced in the first loading phases. However, at the same time, this makes them more brittle, which means they lose stiffness more quickly than W/RW-02 and 03, as new cracks take place or existing cracks become enlarged. Moreover, taking into account the fact that before the retrofitting, the walls were totally failed, one can be sure that the SRG had an impact on reducing the brittle behaviour of confined masonry walls.435 Indeed, Fig. 16 shows that regardless of the walls, the retrofitted ones showed a lesser stiffness degradation than the original walls, which in turn was related to the stability of the walls. 5. Conclusions The suitability of SRG as a seismic retrofitting technique was evaluated by applying440 it externally to three confined masonry walls and testing them under cyclic in-plane 32 loads. Prior to the retrofitting, the walls were subjected to cyclic loads and were led to their ultimate limit state. It should be noted that of the three tested walls, one had a vertical load of 170 kN during the testing and the rest were only subjected to lateral loads.445 A design procedure was developed by taking some concepts from the Peruvian Code, CNR-DT and AC434. Subsequently, the retrofitting system consisted of 5 SRG strips, each one with a thickness of 10 mm and a width of 100 mm. Within each strip, there was embedded a mesh of 0.084-mm thick galvanized steel fiber. The retrofitting process was also given in detail, showing its easy maneuverability and applicability of450 the materials involved. The experimental results showed that there was a considerable improvement of the seismic performance of the retrofitted confined masonry walls in comparison with the original ones. In terms of ductility, SRG showed a substantial increment of the lateral deformation capacity by 100% in one wall and 50% in the rest. Parallel to the im-455 provement in ductility, it should be highlighted that the retrofitted walls were able to perform correctly even after the performance level of collapse prevention (drift=1%) while maintaining their stability. In terms of energy dissipation, the retrofitted walls showed they were able to dissipated more energy than the original walls. Likewise, greater average values of hysteresis viscous damping were registered during the incre-460 mental loading phases. Finally, taking into account that the retrofitting was applied to failed walls, it was possible to observe that the SRG allowed the walls to enjoy a slighter degradation of their stiffness than the original walls. In this way, the brittle be- haviour was improved and also the integrity and stability of the walls were guaranteed. Acknowledgment465 The authors are grateful for the financial support provided by CONCYTEC within the framework of the N◦ 232-2015-FONDECYT Agreement, as well as to the industrial company KERAKOLL for providing the necessary materials for the retrofitting. 33 References [1] INEI, Censos nacionales 2014: población y vivienda, Instituto nacional de es-470 tadı́stica e informática. [2] N. Kassem, A. Atta, E. Etman, Structural Behavior of Strengthening Masonry In-filled Frames Subjected to Lateral Load using Bonded and Un-bonded CFRP, KSCE Journal of Civil Engineering 21 (3) (2016) 818–828. [3] M. El-Diasity, H. Okail, O. Kamal, M. Said, Structural performance of confined475 masonry walls retrofitted using ferrocement and GFRP under in-plane cyclic load- ing (2015) 54–69. [4] B. Ozsayin, E. Yilmaz, M. Ispir, H. Ozkaynak, E. Yuksel, A. Ilki, Characteris- tics of CFRP retrofitted hollow brick infill walls of reinforced concrete frames, Construction and Building Materials 25 (10) (2011) 4017–4024.480 [5] E. Quagliarini, F. Monni, F. Greco, S. Lenci, Flexible repointing of historical facing-masonry column-type specimens with basalt fibers: A first insight, Journal of Cultural Heritage. [6] N. Gattesco, I. Boem, Characterization tests of GFRM coating as a strengthening technique for masonry buildings, Composite Structures 165 (2017) 209–222.485 [7] T. C. Triantafillou, C. G. Papanicolaou, Textile Reinforced Mortars (TRM) versus Fiber Reinforced Polymers (FRP) as Strengthening Materials of Concrete Struc- tures, Special Publication 230 (2005) 99–118. [8] C. G. Papanicolaou, T. C. Triantafillou, M. Papathanasiou, K. Karlos, Textile reinforced mortar (TRM) versus FRP as strengthening material of URM walls:490 out-of-plane cyclic loading, Materials and structures 41 (1) (2008) 143–157. [9] A. Razavizadeh, B. Ghiassi, D. V. Oliveira, Bond behavior of SRG-strengthened masonry units: Testing and numerical modeling, Construction and Building Ma- terials 64 (2014) 387–397. 34 [10] A. Borri, P. Casadei, G. Castori, J. Hammond, Strengthening of brick masonry495 arches with externally bonded steel reinforced composites, Journal of composites for construction 13 (6) (2009) 468–475. [11] E. Wobbe, P. Silva, B. Barton, L. Dharani, V. Birman, A. Nanni, T. Alkhrdaji, J. Thomas, T. Tunis, Flexural capacity of RC beams externally bonded with SRP and SRG, in: Proceedings of Society for the Advancement of material and Process500 Engineering 2004 Symposium, 2004, pp. 16–20. [12] B. Barton, E. Wobbe, L. R. Dharani, P. Silva, V. Birman, A. Nanni, T. Alkhrdaji, J. Thomas, G. Tunis, Characterization of reinforced concrete beams strengthened by steel reinforced polymer and grout (SRP and SRG) composites, Materials Sci- ence and Engineering: A 412 (1-2) (2005) 129–136.505 [13] S. De Santis, F. Ceroni, G. De Felice, M. Fagone, B. Ghiassi, A. Kwiecie, G. Lig- nola, M. Morganti, M. Santandrea, M. Valluzzi, A. Viskovic, Round robin test on tensile and bond behaviour of steel reinforced grout systems, Composites Part B. [14] CNR-DT-200-R1, Guide for the design and construction of externally bonded FRP systems for strengthening existing structures, Advisory Committee on Tech-510 nical Recommendations for Construction. [15] AC-434, Acceptance criteria for masonry and concrete strengthening using Fabric–Reinforced Cementitious Matrix (FRCM) and Steel Reinforced Grout (SRG) composite systems, International Code Council - Evaluation Service. [16] J. Manchego, S. Pari, Análisis experimental de muros de albañilerı́a confinada en515 viviendas de baja altura en Lima, Perú, Master’s thesis, Pontificia Universidad Católica del Perú (2016). [17] M. Luján, Refuerzo de muros de albañilerı́a confinada con mallas de acero. [18] E070, Reglamento nacional de edificaciones: Albañilerı́a, Ministerio de Vivienda, Construcción y Seneamiento - SENCICO.520 35 [19] S. Brzev, Earthquake-Resistant Confined Masonry Construction, NICEE, Na- tional Information Center of Earthquake Engineering, Indian ?, 2007. [20] M. Tomaževič, M. Lutman, L. Petković, Seismic behavior of masonry walls: Ex- perimental simulation, Journal of Structural Engineering 122 (9) (1996) 1040– 1047.525 [21] CEB-FIP, Model code 2010, London: Thomas Telford. [22] S. Brzev, M. Astroza, O. Moroni, Performance of confined masonry buildings in the February 27, 2010 Chile earthquake, Earthquake Engineering Research Institute, Oakland, California (www. confinedmasonry. org). [23] Kerakoll, GeoSteel G600, Tech. rep., Kerakoll - The GreenBuilding Company530 (2013). [24] Kerakoll, GeoCalce F Antisismico, Tech. rep., Kerakoll - The GreenBuilding Company (2017). [25] J. Salsavilca, J. Yacila, N. Tarque, G. Camata, Experimental and analytical bond behavior of masonry strengthened with steel reinforced grout (SRG), submitted535 manuscript. [26] A. T. Council, M.-A. E. Center, M. C. for Earthquake Engineering Research (US), P. E. E. R. Center, N. E. H. R. P. (US), Interim Testing Protocols for Determining the Seismic Performance Characteristics of Structural and Nonstructural Compo- nents, Federal Emergency Management Agency, 2007.540 [27] P. FEMA, Commentary for the seismic rehabilitation of buildings, FEMA-356, Federal Emergency Management Agency, Washington, DC. [28] M. Dehestani, S. Mousavi, Modified steel bar model incorporating bond-slip effects for embedded element method, Construction and Building Materials 81 (2015) 284–290.545 [29] A. S. of Civil Engineers, Seismic evaluation and retrofit of existing buildings: Asce/sei, 41-17, American Society of Civil Engineers, 2017. 36 [30] A. T. Council, Effects of strength and stiffness degradation on seismic response, FEMA P440A. 37 Introduction Previous Work Description of the tested walls Characterization of the materials Steel Reinforced Grout (SRG) Control tests Design of SRG reinforcement for shear behaviour enhancement Procedure for retrofitting confined masonry walls with SRG Boundary conditions and instrumentation for the cyclic tests Discussion of the results Cracking pattern RW-01 RW-02 & RW-03 Lateral displacement ductility Energy dissipation and damping ratio Initial stiffness and stiffness degradation Conclusions