Assessment of Seismic Capacity for Reinforced Concrete Frames with Perforated Unreinforced Brick Masonry Infill Wall

Infill walls increase the strength and stiffness of the reinforced concrete frames, but they usually are not considering in design. However, when the infills are considered in the design, the opening for doors/windows necessitates investigation as well. This research work aims to investigate the effect of perforations (openings) in the infill walls on the performance of infilled RC frames, in other words, this research investigates the number of infill walls in infilled RC frames. Based on the current construction practices in Pakistan, two full scales perforated infilled RC frames were constructed in the laboratory. One infilled RC frame has an eccentric door and window (specimen-1) while the other has only window at its centre (specimen-2). Both the specimens were tested against reverse cyclic loading (quasi-static test). From the experimental testing, it was found that infilled RC frame having less amount of opening in infill wall has more resistance to lateral loads, have more stiffness and dissipated higher energy as compared to infilled RC frame having a significant size of the opening in infill wall. Similarly, displacement ductility (μD) and Response modification factor (R) also depend on the quantity of opening in infill wall in infilled RC frame.


Introduction
Infilled walls are provided in Reinforced Concrete (RC) frames around the globe for insulation against temperature, moisture, noise and fire [1]. These are considered as nonstructural components and are often, therefore, usually neglected in structural analysis and design [2]. If separation joints are not provided between the infilled walls and frames, under seismic excitations, these walls might contribute to the load resisting mechanism and failure pattern of RC frames [3]. During an earthquake, both the positive and negative role of infilled walls on RC frames have been observed. In some cases, the infilled wall increased the strength and stiffness of RC frames [4] while in other cases, these might led to shear failure of RC frame columns [5].
It is worthy of mentioning that the absence of infilled wall in the bottom of the middle storeys may result in forming soft storey mechanism [6]. Similarly, the partial infilled wall can create short column effect [7], and significant damage of infilled wall creates falling hazard [8]. Different damage mechanisms of infilled frames are possible, depends on the relative stiffness, strength of the infill and the bounding frame [7,9]. The strength of an infilled frame depends on the interaction between these two and possible to change their load resisting mechanism and failure pattern [10]. At lower drift levels, an infilled frame performs like a monolithic composite wall; at larger drift levels (when infill wall separate from the bounding frame) compressive contact stresses develop between the frame and the wall [11,12].
Opening in infilled frames reduces strength, stiffness and energy dissipation of the infilled frames. Many researchers have investigated the effect of opening on the performance of RC frames through the experimental setup [3,4,13,14]. For damage evaluation, FEMA 306 [14] and FEMA 307 [15] identifies that openings within the infill panels are the most significant parameter affecting seismic behavior of infilled systems. Doors and windows are the two most prevalent opening types. Infilled panels with openings are best viewed as assemblies of subcomponents of the appropriate material. These subcomponents interact with the surrounding frame and can alter the frame response. Principal types of interaction that can occur are strong columns and strong piers inducing shear failure in the beams strong spandrel components reducing the ductility by causing short-column effects and infill inducing tension yielding or bar splice failures in the column. Therefore in this research, two Infilled RC frame specimen, in which in one specimen only widow opening was provided while in other both door, and window openings were provided were investigated in quasi-static loading [17][18][19][20][21].
This paper is organized mainly in different sections. The introduction portion is followed by the methodology, the considered specimens and material properties. Then the instrument setup is reported, which is followed by the results that incorporate failure mechanisms, backbone curves, hysteresis loops and bi-linear idealizations. The performance levels are also calculated for these specimens, and finally, conclusions and recommendations are provided.

Methodology and Specimens
In the presented experimental activity, the effect of perforations (openings) in the infill walls on the lateral strength and stiffness of (RC) frames was carried out. For this purpose, typical reinforced concrete full-scale frames were tested under in-plane lateral loading. Two full-scale one-storey, one-bay RC frames having openings at different positions were fabricated. The infill wall of one RC frame was constructed having window while the other infilled frame was provided with both door and window. Each infilled frame was constructed on a RC pad having 3.7×0.61×0.35 m (12×2×1 ft) dimensions. The dimensions of both beam and columns were 0.30×0.30 m (1×1 ft). The flow chart of the methodology is shown in Figure 1.

Figure 1. Flow chart of the methodology
A reinforced concrete pad was fixed to the floor using 25 mm (1 inch) diameter anchors to prevent uplifting of the wall during testing. The infill wall was built in English bond pattern having a thickness of 228.6 mm (9 inches). The elevation detail and infilled frame model with window and door opening constructed in the laboratory are shown in Figure 2. Elevation details and actual model after the construction of the window opening constructed in the laboratory is shown in Figure 3. The section of the frame is shown in Figure 4 with the details of the section AA, section BB and section CC in Figure 5. Reinforcement detail of RC foundation is shown in Figure 6.

Material Properties
This section consists of the material collection that has been required for construction of the infill wall. One of the primary materials of this research work was brick. Therefore, after conducting various tests on bricks, first-class brick from locally available resource was selected. The reinforcement selected was of grade 40, having a yield strength of 276 MPa (40ksi). Type-1 cement from the local cement factory, i.e. Kohat factory of cement was used as a binding material. The coarse aggregate from Basai quarry was selected because of its irregular shape as it enhances the strength of standard excellent fineness modulus. ASTM C136 [22] was used for the sieve analysis, and the result of the sieve analysis of the coarse aggregate is shown in Figure 7. Similarly, ASTM C136 was used for the sieve analysis of the fine aggregate, which was collected from the Nizampur quarry; the result is shown in Figure 8. After collection of different material, the second phase of the research work was to conduct different tests on the material selected and to find out its properties, i.e. compressive strength, modulus of elasticity, diagonal compression and water absorption. ASTM C-109 specification [23] was used to find the compressive strength of mortar. The cement to the sand ratio of mortar consisted of one part of cement and six parts of sand with water to cement ratio of 0.50. The mortar cubes were prepared in standard size moulds of 51×51×51 mm (2×2×2 in) and were removed from the moulds after 24 hours. The casted concrete cubes were placed in water for 28-days. The compressive strength of mortars was determined in Universal Testing Machine (UTM) after 28-days of curing. The tests performed in the laboratory to determine the properties of the material are summarized in Table 1. The compressive strength of masonry prisms was determined according to ASTM C-1314 [24]. The dimensions of the prepared masonry prism were 406 mm (16 inches) high, 406 mm (16 inches) length and a thickness of 229 mm (9 inches). The masonry prism was prepared in English bond with 0.01 m (3/8 in) mortar joints. After that, the compressive strength of masonry prism was then determined by dividing peak failure load into areas of the loaded surface. The compressive strength of bricks prism came out to be 6.6 MPa (960 psi). The stress-strain curve of the brick prism is shown in Figure 9. Modulus of elasticity (Em) was determined according to ASTM C-1314 specification. Two dial gauges were used to measure deformation in the specimen with applied uniaxial loading, Equation 1 provides modulus of elasticity of prism.

Modulus of elasticity = (longitudinally stress)/(longitudinally strain)
( Similarly, masonry prisms were also tested in a diagonal direction to determine the shear strength and modulus of rigidity. The shear stress-strain curve of the brick prism is shown in Figure 10, and the modulus of rigidity was calculated from the shear stress-shear strain curve as the secant modulus between two points corresponding to 1/20 and 1/3 of peak shear strength. The compressive strength of bricks was determined by using ASTM standard specification C-67 [25]. Nine bricks were tested in Universal Testing Machine (UTM) under uniaxial loading. Before testing the bricks; steel plates were placed on top and bottom side of the brick so that the load is equally distributed. After testing the brick samples, the average compressive strength obtained was 12.7 MPa (1.8 ksi). Initial rate of absorption of the brick unit was measured by adopting ASTM C-67 specification and was calculated by using Equation 2.
Initial rate of absorption = ((Ww -Wd)30)/AC) (gm/min/0.02 m 2 (gm/min/30 in 2 )) ( After performing the IRA on different samples, the average value IRA test was 65g/min/0.019m 2 (0.14Ibs/min/30in 2 ). Water absorption of bricks sample was determined by picking the oven-dried sample and was placed in water for 24 hours-the water absorption of bricks masonry determined by using Equation 3.
In the end, average water absorption value was recorded as 20%. The diagonal tensile strength of bricks was determined according to ASTM E 519-02 [26]. Brick masonry prisms of dimension 0.41 m (16 inches) high, 0.41m (16 inches) long and 0.23m (9 inches) thick were subjected to tests. ASTM E 519-02 specification was used to determine shear stress in brick masonry prism. Shear stress is determined by using Equation 4.
RILEM specification was used to determine the tensile strength of prism from diagonal compression testing, [21]. As there are no ASTM, specification equation (5) is used to compute the tensile strength of brick masonry prism. ftμ = (0.5 Pμ / An) The modulus of rigidity "Gm" was determined after plotting shear stress vs. shear strain curve. In the end, the tensile strength of bricks masonry was 51.71 kPa (7.5 psi). The compression strength of the concrete cylinder was determined by adopting ASTM C-39 [27] specification. Concrete cylinders were prepared in 1:2:4 mix proportion with water to binder ratio of 0.45. Before testing, the concrete surface was capped in gypsum.

Instrument Set-up
Quasi-static loading has been used to determine the lateral stiffness of the RC frame structure. The quasi-static test is of two conditions i) Force controlled ii) Displacement controlled. In the displacement control test, the displacement is changed while in the force-controlled test, the force is under our control and is increased incrementally. Therefore, this method of testing is used to find the failure-mechanism of cracks and energy dissipation of structures. In this method, gauges are set up at pre-determined locations, and structural stiffness is continuously monitored. The arrangement of loading is shown in Figure 11.

Figure 11. Schematic layout of loading
The test setup consists of two hydraulic jacks fixed with steel frame having zero deformation (strong frame). One jack applies vertical loads while the other applies horizontal loads. These jacks are controlled through hydraulic machines. In instrumentation, two load cells of 50tons capacity were incorporated to measure horizontal and vertical loads on structures. Table 3 is showing instrumentation layout of dial gauges incorporated in frames with door and window opening.  Figure 12 shows location and designation of dial gauges incorporated in frame with door and window opening. Moreover, Table 3 shows instrumentation layout of dial gauge incorporated in frame with window opening. Figure 13 shows the location and designation of dial gauges used in frame with window opening.  Before application of vertical and horizontal loading on RC-frames, loading setup is shown in Figure 14 for specimen-1. Moreover, loading setup is shown in Figure 15 for specimen-2. The displacement test protocol has been shown in Figure 16. For both specimens, the same loading procedure was adopted, which consist of an increasing amount of in-plane quasi-static loading cycles in reverse nature. Eight (8) tons of pre-compressive loading was applied on both frames. Displacement cycle consists of loading the structure to a displacement of the desired amount. The load was removed after achieving the desired level of displacement in structure. After unloading, reloading stage was started to the same range of specified displacement. The crack that appears due to loading and unloading was adequately monitored and recorded. The models were laterally loaded with increments until the maximum capacity of RC frames was consumed.

Failure-Mechanism of Specimen 1
The results obtained from the experimental data are deliberated. The mechanism of cracks in each displacement cycle is also discussed. The quasi-static test for specimen-1 was started from 0.25 mm (0.01 inch) displacement cycle and ended on 100 mm (3.9 inch). Each displacement cycle was inspected for cracks mechanism. The Failure-Mechanism of cracks from 0.25-1.0 mm (0.01-0.04 inch) displacement cycle was the first recorded cycle, i.e. up to 1mm and no cracks appeared in both the infill wall as well as in the RC frame. This shows that the structure can withstand the external loads under the above-applied displacement and loading conditions. At a 2 mm (0.078 inches) displacement cycle, a vertical crack appeared at the left side of the column wall joint, (See Figure17-a). At 4 mm (0.16 inch) displacement cycle, diagonal cracks were originated from the window corners at the sill-level, as shown in Figure17-b. A horizontal crack at the centre pier was seen due to bed joint sliding. A vertical crack also appeared at the right wall-column joint. At 6 mm (0.24 inch) displacement cycle diagonal crack was seen at the right pier of the door corner below the lintel beam (refer Figure 17-c). A horizontal crack below the centre pier was also noticed in this cycle. In this cycle, new cracks did not appear, but a small shear crack was seen at the right corner. The already appeared crack propagated is shown in Figure 17 After the collapse of the central pier, the load has been distributed to another uncracked part of the structure. Applied displacement and the corresponding cracks at displacement cycle of 60 mm (2.36 inch) (See Figure 11-q) several crossed diagonal cracks propagated at the spandrel masonry, as observed fromFigure11-r. At 84 mm cycle corner compression was observed (Figure17-s). No more cracks were observed, and the test was stopped at a displacement cycle of 100 mm (3.94

Backbone Curve and Hysteresis Loops of Specimen 1
The Hysteresis loop and the capacity curve of the tested specimen 1 are shown in Figures 18 and 19, respectively. The maximum resistance of the structure was 137.73 kN (30.96 Kip) at a drift of 1.42% and after the peak load degradation started. At 3.51% drift, peak load/strength was degraded by 28%. Since the frame was degraded by more than 20% (collapse prevention) at this phase, therefore the test was terminated. Many researchers have used various curves to idealize the backbone curve; the model adopted for this work is shown in Figure 20 and used by Dehghani et al. (2015) [28], which is based on equal energy principle. The total resistance of the structure is Vmax according to the design proposed by Dehghani. Vs denote the point on the actual curve where the structural response dramatically deviates from the elastic response and where the first actual significant yield occurs. Ultimate displacement (Δu) is the displacement point on the curve of the envelope corresponding to 80 per cent of the peak resistance, i.e. 0.8Vmax. Δy is the yield displacement on bilinear idealization curve and the load (Vy) corresponding to yield displacement. The first line of the bilinear curve must intersect with the actual curve at a lateral force of 0.75Vy to obtain equivalent areas under the capacity and bilinear curve. The slope of the first line is the effective stiffness. Displacement ductility is the ratio of ultimate displacement to yield displacement (μ=∆u/∆y). is the over-strength factor, and the relationship was calculated by = / . The structure's capacity for energy dissipation is represented by a reduction in ductility (Rμ) and was calculated by = √(2 − 1). The response modification factor ( ) was derived as = Rs×Rµ (product of over-strength factor and ductility).

Bi-linear Idealization Curve of Specimen 1
Stiffness, Disp. Ductility and Response Modification Factor for specimen 1 are calculated from the bilinear and the obtained capacity curve, as shown in Figure 21. The yield Stiffness, Displacement ductility and Response Modification Factor for specimen 1 are shown in Table 4.   The test for frame 2 was started at 0.25 mm (0.01 inch) and ends at a displacement cycle of 60 mm (2.36 inch). Failure-Mechanism of cracks from 0.25 mm (0.01 inch) to 0.75 mm (0.029 inch). Testing of the specimen was started at a 0.25 mm (0.01 inch) cycle. Up to 0.50 mm (0.02 inch), the structure was strong enough to show resistance to external loadings due to which no cracks were observed. At 0.75 mm (0.029 inch) displacement cycle small diagonal crack appeared at the right masonry pier of the infill wall. Crack seen at 0.75 mm (0.029 inch) cycle extended towards the window corner in this cycle, depicted in Figure 22-a. Vertical crack a wall column joint was observed at the right masonry pier of the infill wall, depicted in Figure 22-c. At a displacement cycle of 4 mm (0.157 inches), diagonal cracks arise from each corner of the window, depicted in Figure 22-d. At 6 mm (0.23 inch) displacement cycle bond tensile failure occurred at the sill masonry, depicted in Figure 22-e. At 8 mm (0.31 inch) cycle sliding of beam-wall joint occurred, depicted in Figure 22-f. At 10 mm (0.39 inch) cycle, diagonal cracks were originated in the spandrel masonry, depicted in Figure 22-g. The Failure-Mechanism of cracks at 12 mm (0.47 inch) displacement cycle was such that bond shear failure at bed joints, bond tensile failure occurred at left and right of masonry piers, depicted in Figure 22-h.
In this displacement, several diagonal cracks were originated at the left masonry pier of the infill wall, depicted in Figure 22-i. The diagonal cracks that were originated in the previous cycles widen in this cycle, depicted in Figure 22j. A series of cracks originated were observed at the whole infill wall, i.e. spandrel masonry left and right masonry pier and at the sill masonry and is depicted in Figure 22-k. Spalling of mortar from the beam-wall joint and the cracks at the sill masonry widen, which is depicted in Figure 22 (e) 6.0mm (f) 8.0mm (g) 10mm (h) 12mm (q) 60mm

Backbone Curve and Hysteresis Loops of Specimen 2
The Hysteresis loop and the capacity curve of the tested specimen 2 are shown in Figures 23 and 24, respectively. Specimen-1 resisted maximum load of 137.73 kN (30.96 Kip) at a drift of 1.42% whereas specimen 2 resisted maximum load of 226.11 kN (50.83Kip) at a drift of 1.74%, therefore strength was increased by 67% in the case of specimen-2 in which only window opening was provided instead of both door and window openings. Similarly, stiffness was also increased by 75% in the case of specimen-2.  providing both door and window openings in the infilled frame instead of the only window opening, ductility increased by 14.04%, it also increased the response modification factor of the frame. The yield Stiffness, Displacement ductility and Response Modification Factor for specimen-2 are shown in Table 5.  Comparison of Seismic parameters for specimen-1 and specimen-2 are given in Table 6.

Performance Levels
There are three performance levels, according to ASCE/SEI 41-17 [29], Immediate Occupancy Level (IO); IO is the point on the backbone curve corresponding to yield drift. Yield drift was determined by 0.9 Vmax divide by dy=0.9Vmax/dy. The IO performance level was at a drift of 0.36%. At a drift of 1.76% drift LS performance level was noticed; while the CP performance level, which is at 20% degradation of the load was at 2.35% after an extrapolation.
The Performance Levels of Specimen-2 are shown in Figure 27, with IO = 0.34% drift, LS = 1.62% drift and CP = 2.16% drift. The LS performance level of specimen-2 as compare to specimen-1 was decreased by 7.95 % while CP performance of was decreased by 8.08 %, therefore specimen 1 is more venerable to lateral loading as compare to specimen-2.

Conclusions and Recommendations
Pakistan is one of the seismically dynamic countries and has encountered some overwhelming earthquakes in the past time, e.g., 2005 earthquake of Kashmir (7.6 on Richter scale) which resulted in substantial economic, social and human losses. Approximately 0.4 million houses were wholly and somewhat damaged. Dominant parts of the damaged structures were infill walls. These infill walls are sometimes provided with openings, i.e. door and windows for entrance and air etc. to find out the effect of the openings on the infill wall. Two full scales perforated reinforced concrete frames were constructed in the laboratory-based on the current construction practices in Pakistan. One RC frame having eccentric door and window, while the other has a window at its centre. Both the frames were tested against reverse cyclic loading (quasi-static test). After performing the test, it was concluded that the door window junction being weak undergoes initial cracking. The maximum load taken by RC frame with door and window (specimen-1) was 148.65 kN (33.42 kips) while RC frame with a window only (specimen-2) reached to a maximum load of 252.16 kN (56.69 kips). The effective stiffness of specimen-1 and specimen-2 were 6.61 kN/m (58.50kip/in) and 12.7142 kN/m (112.53 kip/in) respectively. From the hysteresis loop, it is concluded that greater energy is dissipated in specimen 2 (infilled frame with window opening only) because of the internal friction between the infilled wall and the frame, and a higher number of cracks within the infilled wall. The displacement ductility (µD) and Response modification factor (R) of specimen-1 was 7.05 and 6.95, whereas that of specimen-2 was found 6.06 and 7.09, respectively. Response modification factor is sensitive to the geometric configuration as compared to the single value as suggested in the available codes of practice. Immediate Occupancy (IO), Life Safety (LS) and Collapse Prevention (CP) were 0.36, 1.76 and 2.35% drift respectively whereas IO, LS and CP were found 0.34%, 1.62% and 2.16% of drift respectively for specimen-2. The quantity of infilled wall decreased the performance levels along with the response modification factor.
It is worthy to note that proper attention should be provided to strengthen the door-window junction during construction as this is a weak zone, and therefore it can undergo initial cracking if special care is not provided. Since first-class bricks comprise of better mechanical properties such as strength, therefore these should be strictly used as infill material to avoid any catastrophe. Initial on-site testing, as well as laboratory testing, is required before using in the construction. The corner of windows should be filled with a material so that stress concentration at the corner is minimized.
For the future study, the following are recommendations:  The same frames should be studied again with the opening fastened with a metallic frame;  The effect of black masonry infilled RC with openings should be studied;  The location of the window should be changed and then again should be compared with the specimen 2.

Funding
This work was sponsored by Higher Education Commission, Government of Pakistan within the funded projects.

Acknowledgements
The Authors highly acknowledges the support of laboratories management the testing facilities management at the University of Engineering and Technology, Peshawar, Pakistan for allowing to utilize the facilities for the experimental testing. The Authors also acknowledges the reviewer suggestions made during the review phase of the publication.

Conflicts of Interest
The authors declare no conflict of interest.