CHAPTER 4 PUNCHING SHEAR BEHAVIOUR OF RECYCLED AGGREGATE CONCRETE TWO WAY SLABS

Transcription

1 103 CHAPTER GENERAL PUNCHING SHEAR BEHAVIOUR OF RECYCLED AGGREGATE CONCRETE TWO WAY SLABS The problem of punching shear in reinforced concrete slabs under a concentrated load or around a column has been treated quite extensively in the past. The investigation includes concentrated punching under various boundary conditions of the slab. Relatively, very little attention has been directed towards understanding the punching shear behaviour of Recycled Coarse Aggregate Concrete (RCAC) slab elements. Hence, there is need to understand the behaviour of RCAC slabs in punching shear. To achieve this, RCAC slab specimens are cast and tested under concentrated punching load in this investigation. The details of the study are presented in this chapter. 4.2 EXPERIMENTAL PROGRAM The experiments are focused on the influence of Recycled Concrete Aggregate (RCA) on punching shear behavior of simply supported and fixed two way slabs. The experimental program consisted of testing a series of eighteen two way slabs for simply supported condition and eighteen two way slabs for restrained condition. The details of test program planned in this study are shown in Table 4.1 and are extracted from Table 3.10 of chapter 3. It consists of testing a total number of 36 Recycled aggregate concrete slabs, 18 slabs are tested with all edges simply supported and 18 slabs are tested with fully restrained edges on all sides. In the previous chapter, the

2 104 mix proportions, casting of specimens and curing process has been discussed. The following two edge conditions are important and extreme values (maximum and minimum) are obtained for these two edge conditions. 1) All edges simply supported 2) All edges restrained Accordingly, it is planned to investigate the punching shear behaviour of RCAC slabs for the above mentioned two edge conditions. All the slabs were tested under a centrally applied square patch load. The loading arrangement for simply supported slabs is depicted in Figure 4.1 and that for restrained slabs is depicted in Figure 4.2 and Figure 4.3. This type of loading arrangement has been used earlier by Kuang et al. (1992) and Jahangir Alam et al. (2009) for the investigation of punching shear behaviour of reinforced concrete slabs. Nomenclature Table: 4.1 Experimental program for punching shear Replacement of NCA with RCA No. of Slab specimens * Total *with tension reinforcement at center of slab # with tension reinforcement at center and edges of slab No. of Slab specimens # NCAC-S 0% 3 0 RCAC-20-S 20% 3 0 RCAC-40-S 40% 3 0 RCAC-60-S 60% 3 0 RCAC-80-S 80% 3 0 RCAC-100-S 100% 3 0 NCAC-F 0% 0 3 RCAC-20-F 20% 0 3 RCAC-40-F 40% 0 3 RCAC-60-F 60% 0 3 RCAC-80-F 80% 0 3 RCAC-100-F 100% 0 3

3 EXPERIMENTAL SET-UP AND EQUIPMENT The development of the experimental program required a number of items deemed to supply a reliable method of testing. The reliability was regarded as the possibility to reproduce the experiments under similar conditions, eventually obtained almost similar results. The main aim of the testing equipment is to reproduce as accurately as possible the actual loading situation as experienced in real structures, while the measuring devices give an accurate readings of the phenomena investigated. 4.4 STRUCTURAL LOADING FRAME AND PLATFORM The set-up was especially designed by the author to be able to perform testing on medium scale two way slabs. The loading frame was designed with beam and column elements. The loading platform consists of four welded steel beams of ISLB 150 in square shape and it is supported on four columns of ISLB 150 place at four corners. The loading platform and loading frame should be stiff enough to support the loading without significant deformations. The loads acting vertically from the top, the same way as own weight and dead loads are acting. Detailing and structural design for steel members and connections were done according to the Indian Standards IS-800: LOADING ARRANGEMENT AND TESTING The loading arrangement for all edges simply supported and all edges restrained are as follows.

4 Simply Supported Edge Condition The slab specimen is placed over the loading platform and steel rods of 16 mm diameter have been kept below the slab along the four edges to simulate the simply supported edge condition. Placing of steel rods allow free rotation along the edges thus simulating the simply supported edge condition. A single concentrated patch load was applied at the geometric center of each slab with a rigid bearing plate of 100x100x20 mm with a 3-mm thick plywood packing between the slab and the bearing plate. Over this rigid plate, solid circular rod of 50 mm diameter was kept to distribute the load from the hydraulic jack to the slab specimen. The load was applied through hydraulic jack was measured in increments of 200N which corresponds to one unit of calibrated proving ring with 500 kn capacity, vertical deflections at the geometrical center of the slab specimens were measured by using dial gauge with a least count of 0.01 mm. The load at the first crack and the corresponding deflection at the bottom centre of the slab were recorded. The ultimate punching shear load and corresponding deflection at the centre were also observed and recorded. The overall view of a specimen in position ready for testing is shown in Figure 4.1.

5 107 Fig: 4.1 Overall view of a simply supported slab specimen in position ready for testing Restrained Edge Condition The slabs were restrained on all four sides. A single concentrated patch load was applied at the geometric center of each slab with a rigid bearing plate of 100x100x20 mm with a 3-mm thick plywood packing between the slab and the bearing plate. Over this rigid plate, solid circular rod of 50 mm diameter was kept to distribute the load from the hydraulic jack to the slab specimen. For fixidity the slab edges clamped to the loading platform with the C type clamps, which are tightened by bolt and nuts, the arrangement is shown in Figure.4.2. The load was applied through hydraulic jack and was measured with a calibrated proving ring of 500 kn capacity. The vertical

6 108 deflections were measured by using dial gauge with a least count of 0.01 mm. The vertical deflections were measured at the geometrical centre of the slab specimens. The load has been applied incrementally. The load increment was selected such that there will be as many number of readings as possible. The load was applied in increments of 200 N which corresponds to one unit of proving ring. Deflections have been recorded for each load increment. The load at the first crack and the corresponding deflection at the bottom centre of the slab were recorded. The ultimate punching shear load and corresponding deflection at the centre were also observed and recorded. The overall view of a specimen in position ready for testing is shown in Figure 4.3. Solid plate Packing material Slab Bolt Nut Angular frame supporting beam Fig: 4.2 Edge restraint arrangement of slab

7 109 Fig: 4.3 Overall view of a restrained slab specimen in position ready for testing 4.6 TEST RESULTS AND DISCUSSION General Based on the results obtained from the experimentation, the results and analysis is presented below Slab With Four Edges Simply Supported The load versus central deflection curves for the slabs are presented in Figure 4.4. It may be seen that the typical relationship was linear until flexural cracking, which occurred just below the loading point. This event was characterized by a noticeable reduction in slab stiffness. With increasing load, new cracks were formed and the existing ones kept propagated in the radial direction, predominantly towards the

8 110 corners of the slab. The slope of the load-deflection curve also kept decreasing until punching failure occurred. A sudden drop in the applied load marked this event. At this stage, punching shear failure was clearly visible on the top face but, on the bottom face, only an outline of the truncated failure cone with a much larger perimeter had formed. Also, there was noticeable lifting of the corners of the slab Load (kn) NCAC-S RCAC-20-S RCAC-40-S RCAC-60-S RCAC-80-S RCAC-100-S Deflection (mm) Fig: 4.4 Load Vs deflection curves for simply supported slabs First crack load The first crack loads of the experimental investigation are summarized in Table.4.2. The values presented here represent the average of punching shear strengths, load and deflection obtained for three specimens in each group. From the Table.4.2 and Figure 4.5 it is observed that there is a decrease in first crack load as the replacement of NCA with RCA increases. The first crack load of NCAC-S is 14.6 kn and for the RCAC-20-S to RCAC-100-S is between 14.2 to 12.2 kn. The first crack

9 111 load of RCAC-20-S to RCAC-100-S decreases by 2.74 to 16.44%, when compared with NCAC-S. This shows that replacement of NCA with RCA up to 40% there is a marginal decrement in the first crack load and beyond 40% replacement of NCA with RCA the first crack load decreases rapidly. Table: 4.2 First crack load of simply supported slab specimens Sl. No. Nomenclature of slab specimen First crack Load (kn) % decrease w.r.t. NCAC 1 NCAC-S RCAC-20-S RCAC-40-S RCAC-60-S RCAC-80-S RCAC-100-S First crack load (kn) Percentage Replacement of NCA with RCA Fig: 4.5 Effect of percentage replacement of NCA with RCA on first crack load of simply supported slab specimens

10 Ultimate load The ultimate loads of the experimental investigation are summarised in Table.4.3. The values presented here represent the average of punching shear strengths, load and deflection obtained for three specimens in each group. From the Table 4.3 and Figure 4.6 it is observed that there is a decrease in ultimate load as the replacement of NCA with RCA increases. The ultimate load of NCAC-S is kn and for the RCAC- 20-S to RCAC-100-S is between to kn. The ultimate load of RCAC-20- S to RCAC-100-S decreases in the range of 2.07 to 14.11%, when compared with NCAC-S slab specimens. The first crack occurs first in RCAC than that of NCAC slab specimens, it is due to the presence of adhered mortar over the RCA. From the above discussion it can be concluded that, as the percentage replacement increases the first crack load decreases. This shows that as the replacement of NCA with RCA increases the ultimate load decreases. It should be noted that the reduction in the ultimate loads are less than at material level, i.e. the mechanical properties, this may be due to the contribution of steel reinforcement. Table: 4.3 Ultimate load of simply supported slab specimens Sl. No. Nomenclature of slab specimen Ultimate crack Load (kn) % decrease w.r.t. NCAC 1. NCAC-S RCAC-20-S RCAC-40-S RCAC-60-S RCAC-80-S RCAC-100-S

11 Ultimate load (kn) Percentage Replacement of NCA with RCA Fig: 4.6 Effect of percentage replacement of NCA with RCA on Ultimate load of simply supported slab specimens Load deflection response The central deflections of various slab specimens are tabulated in Table 4.4. The central Load Vs deflection response of various slab specimens is shown in Figure.4.4. From Figure 4.4, it is observed that the central deflections corresponding to first crack load of RCAC-20-S to RCAC-100-S decreases in the range of 3 to 30%, when compared with NCAC-S slab specimens and the central deflections corresponding to ultimate load of RCAC-20-S to RCAC-100-S decreases in the range of 3 to 15%, when compared with NCAC-S specimens. From the above observations it is clear that similar trend is observed at first crack and ultimate failure stages. But, rate of decrease of deflections at first crack stage is more when compared to ultimate stage. At the first cracking stage, rate of decrease of deflections are more due to presence of old mortar over the RCA, but at ultimate stage rate of decrease of deflections are less due to the contribution of steel bars.

12 114 Table: 4.4 Maximum central deflection at first crack load and at ultimate load of simply supported slab specimens Nomenclature of slab specimen First crack Load (kn) Deflection at first crack load (mm) Ultimate Load (kn) Deflection at ultimate load (mm) NCAC-S RCAC-20-S RCAC-40-S RCAC-60-S RCAC-80-S RCAC-100-S Stiffness From the load-deflection curves, two values of the stiffness of the tested slabs have been obtained. The un-cracked stiffness (K i ) at 1mm deflection is indicated by the slope of the line of a value less than the first crack load, and the ultimate stiffness (K u ) is measured at 25mm deflection by the slope of the line at about 90% of the ultimate deflection of the lowest sample. The stiffness values are depicted in Table 4.5 and Figure 4.7. From Table 4.5 and Figure 4.7 it can be concluded that the slope becomes steeper when the percentage replacement of NCA with RCA increases. This indicates that the stiffness decreases as the percentage replacement of NCA with RCA increases. Stiffness degradation is defined as the ratio between the ultimate stiffness and the un-cracked stiffness. As the stiffness degradation decreased, the specimen indicates higher ductility. Out of RCAC slab specimens RCAC-100-S shows 9% decrease in stiffness degradation when compared with NCAC-S. So, RCAC slabs specimens are more ductile than the NCAC slab specimens.

13 115 Table: 4.5 Stiffness of simply supported slab specimens Nomenclature of slab specimen Load before cracking at 1mm deflection Load after cracking at 25mm deflection Initial K i (kn/mm) Stiffness Ultimate K u (kn/mm) degradation K u /K i NCAC-S RCAC-20-S RCAC-40-S RCAC-60-S RCAC-80-S RCAC-100-S Stiffness degradation (K u /K i ) Percentage replacement of NCA with RCA Fig: 4.7 Stiffness degradation curve for simply supported edge condition

14 Energy absorption The energy absorption is defined as the area under the load-deflection curve. The values were determined from test results, and are listed in Table 4.6 and Figure 4.8 (af). RCAC-20-S to RCAC-100-S shows 4.17 to 27.50% decrease in the energy absorption, when compared with NCAC-S. Therefore, it can be concluded that as the replacement percentage of NCA with RCA increases energy absorption decreases. Table: 4.6 Energy absorption of simply supported slab specimens Nomenclature of slab specimen Energy absorption (knm) % decrease w.r.t NCAC NCAC-S RCAC-20-S RCAC-40-S RCAC-60-S RCAC-80-S RCAC-100-S

15 % NCAC-S 50 % RCAC-20-S Load (kn) Load (kn) Deflection (mm) Deflection (mm) (a) NCAC-S (b) RCAC-20-S 50 % RCAC-40-S 50 % RCAC-60-S Load (kn) Load (kn) Deflection (mm) Deflection (mm) (c) RCAC-40-S (d) RCAC-60-S 50 % RCAC-80-S 50 % RCAC-100-S Load (kn) Load (kn) Deflection (mm) Deflection (mm) (e) RCAC-80-S (f) RCAC-100-S Fig: 4.8 (a-f) Energy absorption curves for simply supported edge condition

16 Cracking and Failure Patterns The close up view of slab at top and bottom after failure are presented in Figure.4.9 and Figure 4.10, respectively. The typical top view of the failure slab specimen is presented in Figure The final cracking patterns of the NCAC-S, RCAC-20-S, RCAC-40-S, RCAC-60-S, RCAC-80-S and RCAC-100-S slabs are presented in Figures.4.12 to It is observed that the cracks on the bottom face are radial, running predominantly between the loading point and the corners. A circular punching surrounding the patch load occurred on the top surface, this was reflected on the bottom face with an enlarged area, clearly identifying the truncated cone. Similar failure pattern is reported by Kuang and Morley (1992) with respect to RCC slabs. This general pattern of cracking applies to all the slabs but, depending on the replacement percentage of NCA with RCA some variations in the extent and spacing of cracks and the perimeter of the failure cone at the bottom were noted. A careful observation reveals that the bottom perimeter decreases as the replacement percentage of NCA with RCA increases, although the overall cracking pattern remains identical for all slab specimens. Fig: 4.9 Typical Close up top view of simply supported slab specimens after failure

17 119 Fig: 4.10 Typical Close up bottom view of simply supported slab specimens after failure Fig: 4.11 Typical Top view of simply supported slab specimens after failure

18 120 Fig: 4.12 Cracking pattern of NCAC-S slab Fig: 4.13 Cracking pattern of RCAC-20-S slab

19 121 Fig: 4.14 Cracking pattern of RCAC-40-S slab Fig: 4.15 Cracking pattern of RCAC-60-S slab

20 122 Fig: 4.16 Cracking pattern of RCAC-80-S slab Fig: 4.17 Cracking pattern of RCAC-100-S slab

21 Evaluation of critical perimeter The punching shear failure is generally characterised by the shearing off of a segment of the slab in the shape of a truncated cone or pyramid. The perimeter of the truncated cone on the loaded face is much smaller than that at the bottom. This is due to the formation of inclined shear cracks. To calculate the shear stress associated with punching shear failure, the load at failure is usually divided by the length of the critical perimeter u o of the failure zone times the effective depth d of the slab. u o is determined on the assumption that the critical section is located at a distance kd from the face of the loaded area and has a shape identical to the loaded area. For a square patch load of side c, the length of critical perimeter u o may therefore be expressed as u o 4 c 2kd (4.1) To establish the value of k on the basis of the present tests, the distance of the crack line from the center of the loaded area was measured at eight points along the perimeter, both on the top and bottom faces of the slab. The average values thus obtained for each slab are presented in Table 4.7. From these measurements, k can be estimated for each slab using the following expression which is reported by Mansur (2000) ( rt rb ) 4c k 8d (4.2) Where r t and r b are average radius of the failure cone at the top and bottom faces of the slab, respectively. The values of k and u o calculated for each slab by using equations 4.1 and 4.2 and are presented in Table 4.7. It can be seen that k and u o decreases as the replacement ratio of NCA with RCA increases. According to IS-456, ACI-318, BS and Euro codes the k value is 0.5, 0.5, 1.5 and 2.0, respectively. When compared

22 124 with IS-456, ACI-318 and BS-8110 the k value is highly conservative. When compared to Euro code 2 the values are conservative up to replacement of recycled aggregate RCAC 80-S. Table: 4.7 Evaluation of critical perimeter of simply supported slab specimens Nomenclature of slab specimen Average radius of truncated failure cone Top face Bottom face r t (mm) r b (mm) Coefficient K Critical perimeter u o (mm) NCAC-S RCAC-20-S RCAC-40-S RCAC-60-S RCAC-80-S RCAC-100-S Average k (RCAC 20 to RCAC 100) Prediction of failure mode The design provisions for the shear strength incorporated in the building codes are a direct result of the empirical procedures derived from tests on slab specimens. It is important to examine the existing formulae for predicting the shear strength of RCAC-S slabs. Although punching failure occurs to the concrete shearing in a highly stressed compression zone adjacent to the column, the pre-ultimate deformations depend primarily on the flexural characteristics of the slab. Thus, it is necessary to examine the flexural and shearing capacity of every slab at failure. The ultimate flexural strength P yield should be estimated for the various test slabs, assuming the shear failure will not govern the failure criteria. The flexural ultimate load capacity of two way slabs can be calculated using the concept of Johansen s yield line theory [Johanson (1962), Robert Park and

23 125 William L. Gamble (2000)]. For square slabs supported along all four edges with the corners free to lift, and from the consideration of the virtual work done by the actions on the yield lines, the flexural punching load required for this mechanism is in which m is the ultimate flexural moment per unit width calculated from the flexural equation which is reported by Robert Park and William L. Gamble (2000). L P 8m 0. ( s c) 172 yield (4.3 ) f y m As f y d 0. 59As ( 4.4 ) fc Where, L = side dimensions of square slab; s = side dimension between slab supports; c = side length of a square patch load; A s = area of tensile steel reinforcement per unit width; f y = yield strength of reinforcement; f c= cylindrical compressive strength; and d = effective depth of slab. The ratio of the observed strength at failure P u and computed strength P yield, P u o has often been used to identify slab failure mode. If o Pyield < 1.0 the failure mode is shear failure and o >1.0 the failure mode is flexural. The term o was first introduced by Hognestad (1953). The flexural strengths P yield and the o ratios are calculated based on equations 4.3 and 4.4 given in Table 4.8. From 12th column in Table 4.8, it is observed that o < 1.0 for all slab specimens. Hence, all the slab specimens are failed in punching shear by dominating flexure.

24 126 Table: 4.8 Comparison of test results with the codes of Simply supported slab specimens Ultimate load (kn) Nomenclature of slab specimen Tested Value (P u ) IS 456 (P IS ) ACI 318 (P ACI ) Predicted Euro BS Code (CEB) (P BS ) (P CEB ) Yield line theory (P yield ) P u P IS P u P ACI P u P BS P u P CEB P u P yield NCAC-S RCAC-20-S RCAC-40-S RCAC-60-S RCAC-80-S RCAC-100-S

25 Comparison with design codes The tests results are compared with the values predicted by the punching shear formulations of IS 456, ACI 318, BS 8110 and Euro code 2 and are depicted in Table.4.8. In this analysis, the strength reduction factor and the partial safety factors for materials have been taken as unity. The comparison of the experimental failure loads and the design punching shear strengths is shown in Figure It may be seen from Table.4.8 and Figure.4.18 that the IS and ACI codes gives lesser predictions. But Euro code predictions were closer to the experimental load carrying capacities and are realistic than other codes. The ratio of test to predicted loads by IS code, ACI code, BS code and Euro code ranges from 1.41 to 1.44, 1.19 to 1.22, 0.80 to 0.85 and 1.05 to 1.12 respectively. The tested loads for all replacements of NCA with RCA are higher by 41 to 44% when compared to IS code, 19 to 22% when compared with ACI code and 5 to 12% when compared with Euro code. But when compared with BS code all slab specimens show 15 to 20% lesser value P u P IS P ACI P BS P CEB Load (kn) % Replacement of NCA with RCA Fig: 4.18 Comparison between experimental strength and code predictions for simply supported slab specimens

26 Punching shear strength The ultimate punching shear resistance for various specimens summarized in Table.4.9 along with IS 456 and ACI 318 building code predictions. The results represents the average of three specimens tested and are normalised to u o d f c and u o d f '. u o is the critical perimeter of the critical shear failure surface taken at a c distance 2 d away from the face of the patch load. It is observed in Table.4.9, Figure 4.19 and Figure 4.20 that the normalised ultimate punching shear capacity varies almost linearly with the replacement percentage of NCA with RCA. Table: 4.9 Normalised ultimate punching shear strength of simply supported slab specimens Nomenclature of slab specimen Normalised Ultimate shear stress (with f c ) u o d P u f c Normalised Ultimate shear stress (with f c ) u o d P u f ' c Tested IS 456 Tested ACI 318 NCAC-S RCAC-20-S RCAC-40-S RCAC-60-S RCAC-80-S RCAC-100-S

27 Normalised shear stress (MPa) Experimental IS Best fit % Replacement of NCA with RCA Fig: 4.19 Variation of Normalised ultimate punching shear strength (with f c ) with the replacement of NCA with RCA for simply supported slab specimens Normalised shear stress (MPa) Experimental Best fit ACI % Replacement of NCA with RCA Fig: 4.20 Variation of Normalised ultimate punching shear strength (with f c ) with the replacement of NCA with RCA for simply supported slab specimens

28 Regression model for punching shear stress A simple regression model has been developed from the results of present investigation for predicting the punching shear strength of RCAC slabs. To develop the punching shear strength model, linear regression technique has been adopted. The linear regression is in the form of Y=A+BX where Y is independent variable, X is dependent variable and A and B are called regression coefficients. The A and B are determined from regression analysis in accordance with the principle of least squares method. For predicting the shear stress IS code and ACI code uses cube and cylinder compressive strength, respectively. Hence, the proposed models for punching shear stress with f c and f c are as given below Y= A BX = f c Pu = u d o f c.. (4.5) Y = A BX = f ' c Pu = u d o f ' c.. (4.6) From the results of the present study, a simple regression models has been developed similar to the equations given by IS and ACI codes connecting shear stress with cube compressive strength f c and cylinder compressive strength f c, considering critical perimeter at a distance d/2 from the face of the patch load and are presented as equation 4.7 and 4.8 with a standard deviation of and , respectively. τ = Y f c Y = r There fore, τ = ( r ) f c.. (4.7) And,

29 131 τ = Y f c Y = ( r ) There fore, τ = ( r ) f c.. (4.8) Where, r is replacement ratio of NCA with RCA. A comparison of the ultimate shear stress by regression models (Eq. 4.7 and 4.8), IS code, ACI codes and experimental values are presented in Table 4.10 and Figure From the Table 4.10 and Figure 4.21 it can be observed that the proposed model compared well with the experimental shear stress. Table: 4.10 Performance of Regression model for simply supported slab specimens Ultimate shear stress Ultimate shear stress Nomenclature Experimental (with f c ) (N/mm 2 ) (with f c ) (N/mm 2 ) of slab Ultimate shear Regression Regression specimen stress (N/mm 2 ) model IS 456 model ACI 318 (Eq.4.7) (Eq.4.8) NCAC-S RCAC-20-S RCAC-40-S RCAC-60-S RCAC-80-S RCAC-100-S

30 132 Shear stress (MPa) ACI 318 IS 456 Regression model with fc Regression model with f'c Experimental % Replacement of NCA with RCA Fig: 4.21 Variation of Shear stress with the replacement of NCA with RCA of simply supported slab specimens Slab With Four Edges Restrained The load versus central deflection curves for the slabs are presented in Figure It may be seen that the typical relationship was linear until flexural cracking, which occurred just below the loading point. This event was characterised by a noticeable reduction in slab stiffness. With increasing load, new cracks were formed and the existing ones kept propagated in the radial direction, predominantly towards the corners of the slab. The slope of the load-deflection curve also kept decreasing until punching failure occurred. A sudden drop in the applied load marked this event. At this stage, punching shear failure was clearly visible on the top face but, on the bottom face, only an outline of the truncated failure cone with a much larger perimeter had formed.

31 Load (kn) NCAC-F RCAC-20-F RCAC-40-F RCAC-60-F RCAC-80-F RCAC-100-F Deflection (mm) Fig: 4.22 Load deflection curves of restrained slab specimens First crack load The first crack loads of the experimental investigation are summarized in Table The values presented here represent the average of punching shear strengths, load and deflection obtained for three specimens in each group. From the Table.4.11 and Figure 4.23 it is observed that there is a decrease in first crack load as the replacement of NCA with RCA increases. The first crack load of NCAC-F is kn and for the RCAC-20-F to RCAC-100-F is between 18.2 kn to 15.8 kn. The first crack load of RCAC-20-F to RCAC-100-F decreases in the range of 4.21 to 16.84%, when compared with NCAC-F. This shows that replacement of NCA with RCA up to 40% there is a marginal decrement in the first crack load and beyond 40% replacement of NCA with RCA the first crack load decreases rapidly.

32 134 Table: 4.11 First crack load of restrained slab specimens Sl.No. Nomenclature of slab specimen First crack Load (kn) % decrease w.r.t. NCAC 1. NCAC-F RCAC-20-F RCAC-40-F RCAC-60-F RCAC-80-F RCAC-100-F First crack load (kn) Percentage of Replacement of NCA with RCA Fig: 4.23 Effect of percentage replacement of NCA with RCA on first crack load of restrained slab specimen Ultimate load The ultimate loads of the experimental investigation are presented in Table The values presented here represent the average of punching shear strengths, load and deflection obtained for three specimens in each group. From the Table.4.12 and

33 135 Figure 4.24 it is observed that there is a decrease in ultimate as the replacement of NCA with RCA increases. The ultimate load of NCAC-F is kn and for the RCAC-20-F to RCAC-100-F is between kn to kn. The ultimate load of RCAC-20-F to RCAC-100-F decreases in the range of 1.88 to 12.50%, when compared with NCAC-F. This shows that as the replacement of NCA with RCA increases the ultimate load decreases linearly. The first crack occurs first in RCAC than that of NCAC slab specimens, it is due to the presence of adhered mortar over the RCA. From the above discussion it can be concluded that, as the percentage replacement of NCA with RCA increases the first crack load and ultimate load decreases. It should be noted that the reduction in the ultimate loads are less than at material level, i.e. the mechanical properties, this may be due to the contribution of steel bars. Table: 4.12 Ultimate load of restrained slab specimens Sl. No. Nomenclature of slab specimen Ultimate Load (kn) % decrease w.r.t. NCAC 1. NCAC-F RCAC-20-F RCAC-40-F RCAC-60-F RCAC-80-F RCAC-100-F

34 Ultimate load (kn) Percentage of Replacement of NCA with RCA Fig: 4.24 Effect of percentage replacement of NCA with RCA on Ultimate load of restrained slab specimens Load deflection response Central deflection corresponding to first crack load and ultimate load are presented in Table The central Load Vs deflection response of various slab specimens is shown in Figure From Table 4.13 and Figure 4.22, it is observed that the central deflections corresponding to first crack load of RCAC-20-F to RCAC- 100-F decreases in the range of 3 to 32%, when compared with NCAC-F. The central deflections at ultimate load of RCAC-20-F to RCAC-100-F decreases in the range of 2 to 9%, when compared with NCAC-F. Similar trends are observed in first crack stage and ultimate load stage. But, rate of decrease of deflections at first crack stage is more when compared with ultimate stage. At the first cracking stage, rate of decrease of deflections are more due to presence of old mortar over the RCA, but at ultimate stage rate of decrease of deflections are less due to the contribution of steel reinforcement.

35 137 Table: 4.13 Maximum central deflection at first crack load and at ultimate load of restrained slab specimens Nomenclature of slab specimen First crack Load (kn) Deflection at first crack load (mm) Ultimate Load (kn) Deflection at ultimate load (mm) NCAC-F RCAC-20-F RCAC-40-F RCAC-60-F RCAC-80-F RCAC-100-F Stiffness From the load-deflection curves, two values of the stiffness of the tested slabs have been obtained. The un-cracked stiffness (K i ) at 1mm deflection is indicated by the slope of the line of a value less than the first crack load, and the ultimate stiffness (K u ) is measured at 25mm deflection by the slope of the line at about 90% of the ultimate deflection of the lowest sample. The stiffness values are depicted in Table 4.14 and Figure From Table 4.14 and Figure 4.25 it can be concluded that the slope becomes steeper when the percentage replacement of NCA with RCA increases. This indicates that the stiffness decreases as the percentage replacement of NCA with RCA increases. Stiffness degradation is defined as the ratio between the ultimate stiffness and the un-cracked stiffness. As the stiffness degradation decreased, the

36 138 specimen indicates higher ductility. So, RCAC slabs specimens are more ductile than the NCAC slab specimens. Table: 4.14 Stiffness of restrained slab specimens Nomenclature of slab specimen Load before cracking at 1mm deflection Load after cracking at 25mm deflection Initial K i (kn/mm) Stiffness Ultimate K u (kn/mm) degradation K u /K i NCAC-F RCAC-20-F RCAC-40-F RCAC-60-F RCAC-80-F RCAC-100-F Stiffness degradation (K u /K i ) Percentage replacement of NCA with RCA Fig: 4.25 Stiffness degradation curve for restrained edge condition

37 Energy absorption The energy absorption is defined as the area under the load-deflection curve. The values were determined from test results, and are listed in Table 4.15 and Figure 4.26 (a-f). RCAC-20-F to RCAC-100-F shows 4.70 to 22.15% decrease in the energy absorption, when compared with NCAC-F. Therefore, it can be concluded that as the replacement of NCA with RCA increases energy absorption decreases. Table: 4.15 Energy absorption of restrained slab specimens Nomenclature of slab specimen Energy absorption (knm) % decrease w.r.t NCAC NCAC-F RCAC-20-F RCAC-40-F RCAC-60-F RCAC-80-F RCAC-100-F

38 NCAC-F RCAC-20-F Load (kn) Load (kn) Deflection (mm) Deflection (mm) (a) NCAC-F (b) RCAC-20-F 80 RCAC-40-F 80 RCAC-60-F Load (kn) Load (kn) Deflection (mm) Deflection (mm) (c) RCAC-40-F (d) RCAC-60-F RCAC-80-F RCAC-100-F Load (kn) Load (kn) Deflection (mm) Deflection (mm) (e) RCAC-80-F (f) RCAC-100-F Fig: 4.26 (a-f) Energy absorption curves for restrained edge condition

39 Cracking and Failure Patterns The close up view of slab at top and bottom after failure are presented in Figure.4.27 and Figure 4.28, respectively. The typical top view of the failure slab specimen is presented in Figure The final cracking patterns of the NCAC-F, RCAC-20-F, RCAC-40-F, RCAC-60-F, RCAC-80-F and RCAC-100-F slabs are presented in Figures.4.30, 4.31, 4.32, 4.33, 4.34 and 4.35, respectively. It may be seen that the cracks on the bottom face are radial, running predominantly between the loading point and the corners. A circular punching surrounding the patch load occurred on the top surface, this was reflected on the bottom face with an enlarged area, clearly identifying the truncated cone. Similar failure pattern is reported by Kuang and Morley (1992) with respect to RCC slabs. This general pattern of cracking applies to all the slabs but, depending on the replacement of NCA with RCA some variations in the extent and spacing of cracks and the perimeter of the failure cone at the bottom were noted. A careful observation reveals that the bottom perimeter decreases as the replacement of NCA with RCA increases, although the overall cracking pattern remains identical for all slab specimens. Fig: 4.27 Typical Close up top view of restrained slab specimens after failure

40 142 Fig: 4.28 Typical Close up top view of restrained slab specimens after failure Fig: 4.29 Typical top view of restrained slab specimens after failure

41 143 Fig: 4.30 Cracking pattern of NCAC-F slab Fig: 4.31 Cracking pattern of RCAC-20-F slab

42 144 Fig: 4.32 Cracking pattern of RCAC-40-F slab Fig: 4.33 Cracking pattern of RCAC-60-F slab

43 145 Fig: 4.34 Cracking pattern of RCAC-80-F slab Fig: 4.35 Cracking pattern of RCAC-100-F slab

44 Evaluation of critical perimeter The punching shear failure is generally characterised by the shearing off of a segment of the slab in the shape of a truncated cone or pyramid. The perimeter of the truncated cone on the loaded face is much smaller than that at the bottom. This is due to the formation of inclined shear cracks. To calculate the shear stress associated with punching shear failure, the load at failure is usually divided by the length of the critical perimeter u o of the failure zone times the effective depth d of the slab. u o is determined on the assumption that the critical section is located at a distance kd from the face of the loaded area and has a shape identical to the loaded area. For a square patch load of side c, the length of critical perimeter u o may therefore be expressed as u o 4 c 2kd (4.9) To establish the value of k on the basis of the present tests, the distance of the crack line from the center of the loaded area was measured at eight points along the perimeter, both on the top and bottom faces of the slab. The average values thus obtained for each slab are presented in Table From these measurements, k can be estimated as which is reported by Mansur (2000) ( rt rb ) 4c k 8d (4.10) Where r t and r b are average radius of the failure cone at the top and bottom faces of the slab, respectively. The values of k and u o calculated for each slab by using equation 4.9 and 4.10 and are presented in Table It can be seen that k and u o decreases as the replacement ratio increases. According to IS-456, ACI-318, BS-8110 and Euro codes the k value is 0.5, 0.5, 1.5 and 2.0, respectively. When compared with IS-456, ACI-

45 and BS-8110 the k value is highly conservative. When compared to Euro code 2 the values are conservative up to replacement of recycled aggregate RCAC-40-F. Table: 4.16 Evaluation of critical perimeter of restrained slab specimens Average radius of truncated Critical Nomenclature of slab specimen failure cone Top face Bottom face Coefficient K perimeter u o (mm) r t (mm) r b (mm) NCAC-F RCAC-20-F RCAC-40-F RCAC-60-F RCAC-80-F RCAC-100-F Average k (RCAC 20 to RCAC 100) Prediction of failure mode The design provisions for the shear strength incorporated in the building codes are a direct result of the empirical procedures derived from tests on slab specimens. It is important to examine the existing formulae for predicting the shear strength of RCAC-F slabs. Although punching failure occurs to the concrete shearing in a highly stressed compression zone adjacent to the column, the pre-ultimate deformations depend primarily on the flexural characteristics of the slab. Thus, it is necessary to examine the flexural and shearing capacity of every slab at failure. The ultimate flexural strength P yield should be estimated for the various test slabs, assuming the shear failure will not govern the failure criteria.

46 148 The flexural ultimate load capacity of two way slabs can be calculated using the concept of Johansen s yield line theory [Johanson (1962), Robert Park and William L. Gamble (2000)]. For square slabs supported along all four edges restrained, and from the consideration of the virtual work done by the actions on the yield lines, the flexural punching load required for this mechanism is in which m is the ultimate flexural moment per unit width calculated from the flexural equation which is reported by Robert Park and William L. Gamble (2000). L P fyield 8( m m') (4.11 ) ( s c) f y m As f y d 0. 59As ( 4.12 ) fc f y m ' A' s f y d' 0.59A' s ( 4.32 ) fc Where, L = side dimensions of square slab; s = side dimension between slab supports; c = side length of a square patch load; A s = area of tensile reinforcement at center of the slab; A s = area of tensile reinforcement at edge of the slab; f y = yield strength of reinforcement; f c= cylindrical compressive strength; and d =d = effective depth of slab.

47 149 The ratio of the observed strength at failure P u and computed strength P yield, P u o has often been used to identify slab failure mode. If o Pyield < 1.0 the failure mode is shear failure and o >1.0 the failure mode is flexural. The term o was first introduced by Hogenstad(1953). The flexural strengths P yield and the o ratios are calculated based on equations 4.11 to 4.13 given in Table Table 4.17 summarises the predicted and experimental results. From 12th column in Table 4.17, it is observed that o < 1.0 for all slab specimens. Hence, all the slab specimens are failed in punching shear by dominating flexure.

48 150 Table: 4.17 Comparison of test results with the codes of restrained slab specimens Ultimate load (kn) Nomenclature of slab specimen Tested Value (P u ) IS 456 (P IS ) Predicted BS ACI (P ACI ) (P BS ) Euro Code 2(CEB) (P CEB ) Yield line theory (P yield ) P u P IS P u P ACI P u P BS P u P CEB P u P lyield NCAC-F RCAC-20-F RCAC-40-F RCAC-60-F RCAC-80-F RCAC-100-F

49 Comparison with design codes The tests results are compared with the values predicted by the punching shear formulations of IS 456, ACI 318, BS 8110 and Euro code 2 and are depicted in Table In this analysis, the strength reduction factor and the partial safety factors for materials have been taken as unity. The comparison of the experimental failure loads and the design punching shear strengths is shown in Figure It may be seen from Table.4.17 and Figure 4.36 that the IS, ACI, BS and Euro codes gives conservative predictions. BS code predictions were closer to the experimental load carrying capacities and realistic than other codes. The ratio of test to predicted loads by IS code, ACI code, BS code and Euro code ranges from 1.90 to 1.92, 1.60 to 1.62, 1.08 to 1.13 and 1.42 to 1.49 respectively. The tested loads for all replacements of NCA with RCA are higher by 90 to 92% when compared to IS code, 60 to 62% when compared with ACI code, 8 to 13% when compared with BS code and 42 to 49% when compared with Euro code P u P IS P ACI P BS P CEB Load (kn) % Replacement of NCA with RCA Fig: 4.36 Comparison between test strength and code predictions of restrained slab specimens

50 Punching shear strength The ultimate punching shear resistance for various specimens summarized in Table.4.9 along with IS 456 and ACI 318 building code predictions. The results represents the average of two specimens tested and are normalised to u d o f c and u d o f '. u o is the critical perimeter of the critical shear failure surface taken at a c distance 2 d away from the face of patch load. It is observed in Table.4.18, Figure 4.37 and Figure 4.38 that the normalised ultimate punching shear capacity varies almost linearly with the replacement of NCA with RCA. Table: 4.18 Normalised ultimate punching shear strength of restrained slab specimens Nomenclature of slab specimen Normalised Ultimate shear stress (with f c ) u d o P u f c Normalised Ultimate shear stress (with f c ) u d o P u f ' c Tested IS 456 Tested ACI 318 NCAC-F RCAC-20-F RCAC-40-F RCAC-60-F RCAC-80-F RCAC-100-F

51 Normalised shear stress (MPa) Experimental IS Best fit % Replacement of NCA with RCA Fig: 4.37 Variation of Normalised ultimate punching shear strength (with f c ) with the replacement of NCA with RCA for restrained slab specimens 0.50 Normalised shear stress (MPa) Experimental ACI Best fit % Replacement of NCA with RCA Fig: 4.38 Variation of Normalised ultimate punching shear strength (with f c ) with the replacement of NCA with RCA for restrained slab specimens

52 Regression model for punching shear strength A simple regression model has been developed from the results of present investigation for predicting the punching shear strength of RCAC slabs. To develop the punching shear strength model, linear regression technique has been adopted. The linear regression is in the form of Y=A+BX where Y is independent variable, X is dependent variable and A and B are called regression coefficients. The A and B are determined from regression analysis in accordance with the principle of least squares method. For predicting the shear stress IS code and ACI code uses cube and cylinder compressive strength, respectively. Hence, the proposed models for punching shear stress with f c and f c are as given below Y= A BX = f c Pu = u d o f c.. (4.13) Y = A BX = f ' c Pu = u d o f ' c.. (4.14) From the results of the present study, a simple regression models has been developed similar to the equations given by IS and ACI codes connecting shear stress with cube compressive strength f c and cylinder compressive strength f c, considering critical perimeter at a distance d/2 from the face of the patch load and are presented as equation 4.15 and 4.16 with a standard deviation of and , respectively. τ = Y f c Y = ( r ) There fore, τ = =( r ) f c.. (4.15) And,

53 155 τ' = Y f c Y = ( r) There fore, τ' = ( r ) f c.. (4.16) Where, r is replacement ratio of NCA with RCA. A comparison of the ultimate shear stress by regression models (Eq and 4.16), IS code, ACI codes and experimental values are presented in Table 4.19 and Figure From the Table 4.19 and Figure 4.39 it can be observed that the proposed model compared well with the experimental shear stress. The experimental and regression model values give more conservative predictions. Table: 4.19 Performance of Regression model for retrained slab specimens Ultimate shear stress Ultimate shear stress Nomenclature Experimental (with f c ) (N/mm 2 ) (with f c ) (N/mm 2 ) of slab Ultimate shear Regression Regression specimen stress (N/mm 2 ) model IS 456 model ACI 318 (Eq.4.15) (Eq.4.16) NCAC-F RCAC-20-F RCAC-40-F RCAC-60-F RCAC-80-F RCAC-100-F

54 156 Shear stress (MPa) ACI 318 IS 456 Regression model with fc Regression model with f'c Experimental % Replacement of NCA with RCA Fig: 4.39 Variation of Shear stress with the replacement of NCA with RCA for 4.7 EFFECT OF EDGE CONDITION restrained slab specimens General In the previous sections the behaviour of recycled aggregate concrete two way slabs with percentage of replacement from 0 to 100 with two different edge conditions under punching shear loading have been presented. For each case, failure loads, loaddeflection response, crack pattern and regression models have been discussed in detail. In this section a comparative analysis is presented viz -viz edge conditions. The comparison is presented for the following parameters: 1. First crack and ultimate loads 2. Load deflection response 3. Regression models