The influence of washer reinforced nail connections on the lateral resistance of shear walls with large openings

Transcription

1 The influence of washer reinforced nail connections on the lateral resistance of shear walls with large openings Richard, Nicolas 1, Prion, H.G.L. 2, and Daudeville L 3. ABSTRACT This paper is a summary of an experimental study on the lateral performance of wood frame shear walls with very large openings. Firstly, a study on the lateral resistance of common nail connections with lumber and Oriented Strand Board (OSB) revealed that the sheathing connection experienced predominantly pull-through failure modes. Washers were considered as a means to reinforce the nail connection and reduce pull-through failures. Connection tests on two nail lengths (52 mm and 78 mm) were conducted under monotonic and cyclic lateral loading. Results indicate significant improvements in nail performance with the reinforced 78 mm long nail connections. The second phase of the project was to evaluate the influence of reinforced nailing on the static and dynamic performance of fullsize (2.4 m x 2.4 m) perforated wood shear walls, sheathed with OSB panels. Combinations of parameters such as the number of hold-down, the shape of the OSB panels, the nail distribution, and bracing systems were studied. Comparisons of the dissipated energy per cycle revealed a higher capacity for walls with washer reinforcements. The combination of a bracing system with the reinforced 78 mm long nails gave the best results in terms of initial stiffness, maximum load, and dissipated energy during shake table tests. 1. INTRODUCTION Light frame residential construction with properly designed wood based shear walls have generally performed well in past earthquakes. For example, in the 1995 Kobe earthquake, such buildings survived high seismic forces with little damage. Some bigger multi-storey buildings with large openings and irregular plan layouts, however, have not performed adequately. In some cases this poor performance has led to many fatalities and high financial losses as experienced in the 1994 Northridge earthquake. As a result, researchers have been prompted to further examine the performance of shear walls under extreme seismic loading. Of particular interest are cases where large openings severely weaken the walls and potentially cause soft storey failures. This is, for example, very common for small businesses on narrow building lots with large storefront windows. Whereas much research has been done on conventional shear walls, both statically and dynamically, little information is available on walls that are configured to resemble a portal frame around a large opening. This paper presents some tests results on the static and dynamic performance of such perforated wood based shear wall systems. It also addresses the influence of reinforcement detailing such as washer reinforced nailing, the number of hold-down, the shape of OSB panels, the nail distribution, and bracing systems on the shear wall performance. 2. CONNECTIONS TESTS 2.1. Materials The wall panels were constructed of kiln-dried nominal 38 x 89 mm No. 2 and Better Spruce-Pine-Fir dimension lumber and 9.5 mm thick performance rated W24 OSB panels (CSA M88). All specimens were tested within a few days of manufacturing without further conditioning. Two different lengths of common nails were considered (52 mm and 78 mm). Common washers with an exterior diameter of 11.3 mm were introduced as a means to reinforce the nail connection. 1 Ph-D student, L.M.T. Ecole Normale Superieure (E.N.S.) de Cachan, Paris, France 2 Professor, Dept of Civil Engineering, University of British Columbia (U.B.C.), Vancouver, Canada 3 Professor, Dept of Mechanic, L.M.T. Ecole Normale Superieure (E.N.S.) de Cachan, Paris, France

2 2.2. Monotonic tests on connections Lateral load tests on panel connections were performed with 52 and 78 mm common nails, with and without washers. The main differences between the load deformation behaviour of nail connections with or without washer reinforcementwere the maximum load, the corresponding displacement, and the post peak load response. For both lengths, nails without washer reinforcements generally exhibited a more severe load reduction in the post peak region as a result of pull-through type failures. Nails with washer reinforcements exhibited a more ductile behaviour due to the deformation of the wood caused by the nail shank reacting against the framing member. Eventually this did lead to a pullout type failure. Depending on the density of the framing member and the OSB, some pull-through failures also took place resulting once again in a brittle failure. The static load-slip response of the nail connections with washer reinforcements also tends to be less variable in comparison with the cases without washer reinforcements. Figure 1 represents the different responses of the two populations of 78 mm nails. Similar behaviour can be observed with the 52 mm nails. Figure 1 Lateral load-deformation response of the 78 mm nail connections with and without reinforcement 2.3. Cyclic tests on connections Cyclic test on the same group of nail connections were performed using a modified load schedule (Figure 2) based on a protocol originally proposed by He et al. (1998). This load schedule consists of the following sequence: a) 3 cycles to the displacement at 80% of peak load (static test) with a frequency of Hz; b) 3 cycles to the displacement at peak load (static test) with a frequency of Hz; c) 3 cycles to twice the displacement at peak load (static test) with a frequency of Hz; d) a final ramp until failure. This protocol represents what could be expected in a near fault earthquake where early large impulses will cause extensive plastic behaviour. This mobilizes the energy-dissipating characteristics of the connection without causing fatigue fractures in the nails, as was observed in many other tests with long loading protocols. This modified procedure also has an asymmetric displacement to force the specimen failure in one direction. As shown in Figure 2, there is no detectable difference in the cyclic response of the 52 mm nail connection with or without washer reinforcement. However, the post peak cyclic response of the 75 mm nail connection is improved with reinforcement. This observation is different from the findings of the static tests where significant improvement in post peak response can be observed in both the 52 mm and the 78 mm washer reinforced nail connections. The failure mechanism of the 52 mm washer reinforced nail connection in the cyclic tests is also quite different from that in the static test. Here, all the specimens failed in pullout mode.

3 Figure 2 Cyclic test load schedule Since the cyclic response of the shorter un-reinforced 52 mm nail connection is governed by pullout mode; washer reinforcement cannot improve their cyclic performance. However, the longer 78 mm nail connections have better anchorage; therefore, all the un-reinforced 78 mm nail specimens failed in pull-through (one couldn t even sustain the last set of cycles) and washer reinforcement significantly improved their post peak cyclic performance. Furthermore, the post peak response of the reinforced 78 mm nail connection seems to be very repeatable. Figure 2 also shows the shapes of the deformed nails extracted from the failed specimens. Un-reinforced nails tend to bend with a single plastic hinge whereas washer reinforced nails tend to deform into an S-shape with two plastic hinges. 3. MONOTONIC TESTS ON SHEAR WALLS Six 2.4 m x 2.4 m shear walls, each with a symmetrically located 1.2 m wide by 2 m high door opening, were tested under monotonic quasi-static lateral loading at the University B.C. Earthquake Laboratory. Each wall specimen was mounted in a test frame on a 3.3 m x 3.3 m shake table. The top of the wall was restrained against both in-plane and out of plane movements while the shake table was controlled to move the base of the wall laterally in the direction parallel to the length of the wall. The maximum stroke of the shake table was 150 mm. Displacement transducers and load cells were mounted to monitor the relative displacement between the top of the wall and the shake table. Various wall configurations were tested to evaluate the influence of: the number of hold-downs, the panel shape (rectangular or L-shaped), the nailing density (N.D.), and the use of longer nails with washers as reinforcements (Figure 4). Table 1 summarises the various parameters for each wall. Perimeter Nail Nails used to Number of Description Wall Spacing (1) connect the panels Hold-downs 1 Rectangular panels 150 mm 64 mm 2 2 Brace system with rectangular panels 150 mm 64 mm 2 3 L-shaped panels 150 mm 64 mm 2 4 L-shaped panels 150 mm 64 mm 4 5 Rectangular panels 50 mm 64 mm 2 6 Brace system with 64 mm & 78 mm 150 mm rectangular panels plus washer mm & 78 mm Rectangular panels 150 mm plus washer 2 Table 1. Shear wall monotonic test parameters (1) 300mm spacing for all interior nailing

4 Figure 5 shows the test results of full size shear walls under monotonic loading. Wall 1 was a conventional system made with rectangular panels without any reinforcements. Typically, failure initiated at the lintel to wall connection where it opened during loading. Comparisons of results from Wall 1 (rectangular panels with normal nail spacing) and wall 2 (braced system) indicate that addition of braces did not improve the initial stiffness of the system but slightly increased the maximum load capacity of the system by 16%. The L-shape panels (Walls 3) had the highest initial stiffness but the sheathing was also heavy damaged under high loads (buckled or tension failure). It was judged that this type of walls would not perform favourably under cyclic loading. Results from Walls 3 and 4 show that there was very little difference in response between the L-shape panel system with 2 or 4 hold-downs. However, comparing the results from Walls 1 and 5 (reduced nail spacing) indicate that Wall 5 had higher capacity (+47.3%) but this could not be sustained beyond a displacement of 80 mm because of the failure of one of the hold-downs. This led to up-lift of the wall at one corner and reduced nail spacing no longer governed the behaviour. Figure 4 Figure 5: Monotonic load deformation test results of full size walls. The behaviour of a braced wall with rectangular panels and 78 mm nails reinforced with washer was studied in Wall 6. This type of reinforcement was only used around the perimeter of the wall specimen and normal 63 mm long un-reinforced nails were used elsewhere. In comparisons with Wall 1, results show that Wall 6 had higher stiffness (+60%), high capacity (+40.6%) and good ductility. A final monotonic test was performed on an un-braced wall with 78 mm washer reinforced nails (Wall 8). Unfortunately, the nailing in this wall was poor and some of the studs were too short. As a result only a slight increase in stiffness and little increase in maximum load were observed in comparisons with Wall CYCLIC TESTS ON SHEAR WALLS Two un-braced walls with rectangular panels were tested cyclically: one wall with normal 64 mm nailing (Wall 9) and a second wall with 78 mm washer reinforced nailing (Wall 10). As the monotonic test results from Walls 1 and 8 were similar in terms of maximum load and the displacement corresponding to 50% and 80% of maximum load, identical loading cycles were chosen for Walls 9 and 10. The original UBC protocol (He et al. 1998) was used for this test (3 cycles at the corresponding displacement of 50% of the maximum load, then 3 cycles at 80% of the maximum load, then 1 cycle back to 50% of the maximum load, and a final push over). Unfortunately this last cycle was forgotten for Wall 10. At 150 mm of wall displacement the limit of the shake table was reached; it was decided to add one more cycle to 150 mm after the unloading process.

5 Figure 6.: Cyclic load-deformation test results of full size walls & the dissipated energy per cycle. Comparisons of cyclic versus monotonic responses: The initial stiffnesses are almost identical for all the walls. Each cyclic test followed their respective envelope curve even in term of maximum load. The cyclic response of Wall 10 (washer reinforced) exceeded its envelope curve, but it was already known that the corresponding envelope curve came from monotonic test of Wall 8 which had construction problems. The post-peak behaviour is slightly different between the monotonic and cyclic tests. In general the cyclic results indicate a slightly steeper decline in load-deformation response compared to the monotonic results. These observations on post peak behaviour are similar to the observations made on the connections (with and without washer). Comparisons of the walls with and without washer reinforcements: The main difference between the cyclic response of walls with and without washer reinforcement is the load capacity at high displacements. The wall with washer reinforcement could sustain a much higher load than the other one, despite of the failure of some framing members (studs and bottom plates were severely split). The responses to second and third load cycles were also repeatable and stable. In terms of dissipated energy, an average of 30% drop was observed between the first and second cycle. The total energy dissipated by the wall with and without washer is about 5700kN*mm and 5400kN*mm, respectively. Although little difference was observed under cyclic loading between Walls 9 and 10 in terms of dissipated energy, the advantage of washer reinforced nailing might be more dominant during the post peak response which could show up more prominently in earthquakebased dynamic loading. 5. DYNAMIC TESTS ON SHEAR WALLS Three walls were dynamically tested on the UBC shake table with the Joshua Tree Landers California Earthquake (E-W component). The record was scaled to a peak acceleration of approximately 0.4g. An inertia mass of 3400 kg was attached to the top of the wall. No dead weight was applied. The test configurations included: an unbraced wall with rectangular panels and normal 64 mm nailing (Wall 11); an unbraced wall with rectangular panels and 78 mm washer reinforced nailing (Wall 12); and a braced wall with rectangular panels and with 78 mm washer reinforced nailing (Wall 13). Impact and Sine Sweep tests were performed before the dynamic test to determine the natural frequency of the system. After the dynamic tests, each wall was retested with one more monotonic test cycle: once to the shake table displacement limit of 75 mm (towards the left in Figures 7 to 9) and return to a neutral position; and once more to the maximum shake table displacement of 150mm (towards the right in Figures 7 to 9) from the last position. This procedure provided additional information about the reserve capacity of the walls to sustain high loads after major events. This process involved a sequence of re-bolting the wall specimens on to the Shake Table which led to the discontinuity observed in the Figures. Dynamic test results: The impact and Sine Sweep tests on walls 11 to 13 resulted in a natural frequency of 2 Hz for the wall and frame system. In Figures 7 to 9 only the main hysteresis loops from the dynamic tests are presented.

6 Figure 7 Figure 8 The left graph (Figure7) shows the behaviour of Wall 11 and the difference between the monotonic curve and the dynamic hysteresis loops followed by a pushover test. The wall under dynamic excitation reached a lower shear force compared to the peak load experienced in the monotonic test. In general the behaviour observed during the cyclic tests remains true for this dynamic test. The dynamic tests almost reach the limit of the shake table, which allows a displacement of 75 mm on either side of the zero position. The graphs on the right (Figure 8) show the behaviour of Wall 12 (with 78 mm washer reinforced nailing). A higher shear force was experience by Wall 12 compared to Wall 11. The dynamic behaviour of Wall 12 agreed with the envelope curve. The two final push-over tests on Wall 12 reached the same peak load as for the monotonic tests. The maximum displacements during the earthquake were lower in Wall 12 compared to Wall 11. The first peaks of major excitation took place between 50 mm and 20 mm, the second between 80 mm and 35 mm. The difference between Walls 11 and 12 is also evident from the shape of the loops. Wall 12 seems to have wider hysteresis loops, especially in the second set of cycles. This last point will be highlighted by the comparisons of the dissipated energy between the walls. Walls 11 and 12 also show a non-symmetric behaviour. Finally, the graph in Figure 9 shows the behaviour of Wall 13 (braced wall with 78 mm washer reinforced nailing). Wall 13 had a more symmetric behaviour under dynamic loading as well as a much higher stiffness compared to Walls 11 and 12. Once again the dynamic hysteresis loops and the monotonic envelop curve agreed well. The shapes of the hysteresis loops are much wider compared to those from Wall 11 and 12. From the dynamic hysteresis loops the dissipated energy during the earthquake was calculated.. In the graph in Figure 10, Wall 11 dissipated much less energy than Walls 12 and 13. Wall 13 with bracing and reinforced nailing dissipated the most energy. Wall 11 dissipated only 67% of the energy of Wall 12, and slightly over 50% of the dissipated energy of Wall 13.

7 Figure 9 Figure CONCLUSIONS The aim of this research project in UBC was to study the influence of washers from connections (very simple tests) to shear wall structures, including one major opening, under monotonic, cyclic and dynamic tests. Two other kinds of reinforcement were also considered for the wall: a diagonal brace fixed with a steel plate and L-shaped panels instead of the regular panels for the perforated wall. Results on different nails lengths show that the washer reinforcement was not effective for shorter nails (52 mm) under cyclic load, but a significant improvement for longer nails (78 mm). The improvement of the washers took place relatively late in the load-slip curve. The longer 78 mm nails were able to provide better anchorage, thus taking advantage of the washer reinforcements. When applied to a shear wall structure, very little improvement for both monotonic and cyclic loading were evident; however, under dynamic loads under numerous load cycles, the washers significantly increased the dissipated energy. Wall reinforced with washers dissipated 50% more energy than normal walls. Parameters like the number of hold-downs or the shape of the OSB panels improved the initial stiffness but led to brittle failures as the panels were torn or buckled. No cyclic or dynamic tests were conducted on these types of walls. A higher nail density helped to increase the shear capacity, but that system also behaved in a brittle mode due to hold-down failures. The best combination seems to be the use of brace member with perimeter washer reinforced 78 mm nails. Both initial stiffness and maximum load were improved in the monotonic test; under dynamic excitation the global behaviour was much more symmetric and the dissipated energy was twice the normal wall. LITERATURE Dolan, J.D The Dynamic Response of Timber Shear Walls. Ph.D. Thesis, University of British Columbia, Vancouver, BC. Dolan, J.D Proposed Test Method For Dynamic Properties Of Connections Assembled With Mechanical Fasteners. J. Testing and Evaluation, 22(6), pp Durham, J.; He, M.; Lam, F.; Prion, H.G.L Seismic Resistance Of Wood Shear Walls With Oversize Sheathing Panels. 5 th World Conference on Timber Engineering, Montreux-Lausanne, Switzerland. 1, pp Filiatrault, A Static and Dynamic Analysis of Timber Shear Walls. Can. J. of Civil Engrg., NRC, 17(4), pp Foschi, R.O Load-Slip Characteristics of Nails. Wood Sci., 7(1), pp He, M.; Lam, F.; Prion, H.G.L Influence of Cyclic Test Protocols on Performance of Wood Based Shear Walls. Can. J. of Civil Engrg., NRC, (25), pp

8 He, M.; Magnusson, H.; Lam, F.; Prion, H.G.L Cyclic Performance of Perforated Wood Shear Walls with Oversized Panels. J. of Struct. Engrg., ASCE, 125(1), pp Johnson, A.C.; Dolan, J.D Performance of Long Shear Walls with Openings. Proc., Int. Wood Engrg Conf., Vol. 2, Omnipress, Madison, Wisc., pp Lam, F.; Prion, H.G.L.; He, M Lateral Resistance of Wood Shear Walls with Large Sheathing Panels. J. of Struct. Engrg., ASCE, 123(12), pp