CYCLIC PERFORMANCE OF PERFORATED WOOD SHEAR WALLS WITH OVERSIZE OSB PANELS By Ming He,'Henrik Magnusson,2 Frank Lam,3 Member,ASCE,and Helmut G.L.Prion' n failure mode,especially for walls with oversi A n INTRODUCTION fiths 1996)it was observed that under monotonic loading con- The shear walls n tal985) ea tica culate eismic loading Its have show for the latter. s ( alls with open en tion provides acon wall without ope ith s is inereased.the gt)(Patton-Mallo ntegrit of od frame build y in mult create veak story Many models h ddes a sub the her linear finite-elemen models w re dev ut or with pen rack osek 1976,Fo shear wall system i and 1984:Gupta and Ku 1985) of the 1996 Line and Dougl nite-elem tprograms to per n time lls24× 12 of s with I pe liatrault 199).The aspect ratios rge pan els4.8×4.8 m)(Enjily Wood Sci..Univ.of British Columbia wa 118S-20P uct.Engrg,Lund Inst.of Technol.,Bo (OSB) Nood Sci..Univ.of Britis mbia.Vancou- sses with sheathir of Wood Sci.Univ.of British Colu nbia.Vancou shown that the racking strength of wall ver.BC hits for man 997.Th paper on A f the 15973-94 nd existing ruls has night een investiga case 10/JOURNAL OF STRUCTURAL ENGINEERING/JANUARY 199 d 05 Jan 2009 to 222.55.175.206.Rodistribution sub ct to ASCE liconse or copyright;soo http:pubs asco.orgfcopyright
10 / JOURNAL OF STRUCTURAL ENGINEERING / JANUARY 1999 CYCLIC PERFORMANCE OF PERFORATED WOOD SHEAR WALLS WITH OVERSIZE OSB PANELS By Ming He,1 Henrik Magnusson,2 Frank Lam,3 Member, ASCE, and Helmut G. L. Prion4 ABSTRACT: This paper reports the test results from a study investigating the influence of openings on the lateral resistance of wood-based shear walls built with both standard and oversize oriented strand board panels under monotonic and cyclic loading conditions. Test results showed that the application of nonstandard oversize panels significantly improved the performance of the perforated shear walls compared with standard 1.2 3 2.4 m panels. Door and window openings caused a significant decrease in the strength and stiffness of the walls and precipitated a change in failure mode, especially for walls with oversize panels. Although nail failure modes were commonly observed in walls without openings, a combination of nail and panel failures were observed in shear walls with openings. A newly proposed cyclic test protocol was used that consisted of fewer but more severe displacement excursions, compared with many other test protocols. This was believed to better reflect typical earthquake excitation and avoid low cycle nail fatigue failures, which were observed previously with long sequence cyclic test protocols. INTRODUCTION Wood frame construction using shear wall and diaphragm systems has been shown to be a very cost-effective means of building single- and multifamily residences. The shear walls provide buildings with a very efficient load resisting system to carry three major load components, namely (1) vertical loads, (2) transverse wind loads, and (3) in-plane lateral forces imposed by wind and seismic loading. For the latter, wood frame construction has a history of excellent performance, especially in residential applications with evenly distributed and relatively small openings and no excessive overhangs. For example, in the 1995 Kobe earthquake in Japan, it was evident that residences built with North American wood-based shear wall techniques (2 3 4 platform construction) survived extreme seismic forces with relatively little damage, even in areas where liquefaction of the supporting ground caused widespread building collapses. As was shown in the 1994 Northridge Earthquake and Hurricane Andrew in Florida in 1993, however, the structural integrity of wood frame buildings under the action of natural hazards such as earthquakes and wind is not necessarily guaranteed, especially in multistory buildings where large openings often create weak story effects. Over the past few decades a substantial amount of experimental work has been done on the structural behavior of wood-based shear wall systems without or with openings. Following studies on full-size shear walls or even full scale buildings under static lateral loads, perforated shear wall systems were studied under static and dynamic loading conditions (Patton-Mallory et al. 1985; Falk and Itani 1987; Sugiyama 1994; White and Dolan 1995, 1996; Line and Douglas 1996; Rose and Keith 1996). From tests on long shear walls (2.4 3 12 m) with openings (Johnson and Dolan 1996) and perforated shear walls with large panels (4.8 3 4.8 m) (Enjily and Grif- 1 Grad. Res. Asst., Dept. of Wood Sci., Univ. of British Columbia, Vancouver, BC Canada V6T 1Z4. 2 Visiting Student, Dept. of Struct. Engrg., Lund Inst. of Technol., Box 118, S-221 00 Lund, Sweden. 3 Asst. Prof., Dept. of Wood Sci., Univ. of British Columbia, Vancouver, BC Canada V6T 1Z4. 4 Asst. Prof., Dept. of Wood Sci., Univ. of British Columbia, Vancouver, BC Canada V6T 1Z4. Note. Associate Editor: William M. Bulleit. Discussion open until June 1, 1999. To extend the closing date one month, a written request must be filed with the ASCE Manager of Journals. The manuscript for this paper was submitted for review and possible publication on August 8, 1997. This paper is part of the Journal of Structural Engineering, Vol. 125, No. 1, January, 1999. qASCE, ISSN 0733-9445/99/0001-0010– 0018/$8.00 1 $.50 per page. Paper No. 16392. fiths 1996) it was observed that under monotonic loading conditions the reduction in racking strength and stiffness of perforated shear wall was directly related to the proportional area of the openings. Based on their research work, empirical equations were introduced by Patton-Mallory et al. (1985) (based on the ratio of effective wall length) and by Sugiyama (1994) (using the sheathing area ratio) to calculate the strength loss due to openings. Experimental results have shown, however, that Patton-Mallory’s equation tends to overestimate the initial stiffness and ultimate load capacity of shear walls with openings (Patton-Mallory et al. 1985), whereas Sugiyama’s equation provides a conservative estimate of the shear load ratio (the ratio of the ultimate lateral load of a shear wall with openings to that of a wall without openings) at ultimate capacity (Johnson and Dolan 1996; Rose and Keith 1996). The results also indicated that as the area of the openings is increased, the governing design criterion may more likely be serviceability (stiffness) rather than ultimate load (strength) (Patton-Mallory et al. 1985; Enjily and Griffiths 1996). Under dynamic loading, shear walls with openings generally were found to exhibit lower damping ratios than similar walls without openings (Falk and Itani 1987). Many models have been proposed to analyze and predict the performance of wood-based shear walls subjected to lateral loads. Linear and nonlinear finite-element models were developed to simulate the load-deformation relations of shear walls under monotonic lateral racking loads (Polensek 1976; Foschi 1977, 1990; Tuomi and McCutcheon 1978; Easley et al. 1982; Itani and Cheung 1984; Gupta and Kuo 1985). The hysteretic behavior of the wall or its components under cyclic and dynamic loading was then studied and modeled in nonlinear fi- nite-element programs to perform time-history analysis and predict the dynamic response of shear walls (Dolan 1989; Filiatrault 1989). The effects of openings and aspect ratios were also modeled (White and Dolan 1995). The development of these analytical models contributed greatly to a better understanding of the structural performance of wood-based shear wall systems. Oriented strand board (OSB) panels are often produced in large presses with sheathing sizes of up to 3.3 3 7.3 m, before they are cut into the standard 1.2 3 2.4 m panels. It has been shown that the racking strength of walls can be substantially increased (Lam et al. 1997) by using full-size panels instead of the standard panels, which could have significant economic benefits for prefabricated house construction or industrial building applications. The effect of openings in walls with large-size panels has not been investigated before, however, and existing rules might not apply to the continuous panel case. Downloaded 05 Jan 2009 to 222.66.175.206. Redistribution subject to ASCE license or copyright; see http://pubs.asce.org/copyright
The study described in this paper addre had kN/mm under lateral and c on the A tests.Furthermore all of the sizing rest t impro hen ful a refe rence to the foundation,s ate the rial Testing Function Gen m or nonstandard under monoondon dimen ached to ng 12.7 mm di iameter clas EXPERIMENTAL STUDIES Test Facilities load distributior beam connected to the actuator a tt甘 n tachdo-mg hreaded s (diameter 27 of the wall thre ugh a mm)to the p of the he jacks were with an inline load cell was mounted onto the re tion fram 24 m.and r providedas cal load d on p in lin vith th The cal load ented an alent oc suppo era load ac sories over an area of3 386/ 16 me was s fixed to the quisition softw ed to regular time intervals The data t-pli 24 244 1250 (b) FIG.1.General Shear Wall Test Assembly(Showing Wall without Openings JOURNAL OF STRUCTURAL ENGINEERING/JANUARY 1999/11 5Jan2009o222.65175206.R0d copyrigh
JOURNAL OF STRUCTURAL ENGINEERING / JANUARY 1999 / 11 FIG. 1. General Shear Wall Test Assembly (Showing Wall without Openings) The study described in this paper addressed these issues. It was part of a comprehensive study on the lateral resistance of wood-based shear walls built with conventional and oversize OSB panels. The previous phase of the study focused on the performance of walls without openings and produced encouraging results, showing significant improvements when fullsized panels were used. The objective of this study was to investigate the structural performance of perforated woodbased shear wall systems built with either standard 1.2 3 2.4 m or nonstandard large dimension 2.4 3 7.3 m OSB panels under monotonic and cyclic lateral loading conditions, and verify existing design rules and computer predictions. EXPERIMENTAL STUDIES Test Facilities Fig. 1 shows the schematics of a typical shear wall test specimen with conventional size sheathing panels. The test specimen was mounted in the load frame with variable inplane lateral load applied by a 222 kN servocontrolled hydraulic actuator along the top of the wall through a steel load distribution beam that was bolted to the wall. The actuator with an inline load cell was mounted onto the reaction frame through a pin connection. The reaction frame was made from three bolt-connected members and was set up in line with the long axis of the wall specimen to support the lateral load actuator. One W310 3 143 steel I beam, 3.7 m in height, was used as the reaction column, and two HSS 101.6 OD 3 6.35 mm hollow structural tubes were used as bracing to stiffen the reaction frame. The frame was fixed to the ground by six 50.8 mmf bolts and welded to the base beam at one end. The reaction frame had a stiffness of 66.7 kN/mm under lateral load, which was adequate to ensure that the lateral deflection of the frame could be neglected without causing any signifi- cant error in the tests. Furthermore, all of the displacement values (Du and Dyield) were measured directly from the top of the walls with a reference to the foundation, so that the effect of the flexibility of the reaction frame was eliminated. An MTS Micro-controller (458.10) and Material Testing Function Generator (Exact-340) were used to drive the actuator in a displacement control mode to develop the required lateral loading pattern. The wall specimen was attached to the W150 3 22 steel load distribution and base beams using 12.7 mm diameter class 5 steel bolts at a spacing of 0.4 m. The assembly was then erected with the base beam attached to the test floor and the load distribution beam connected to the hydraulic actuator at one end. Out-of-plane supports were provided at the two ends of the load distribution beam. Six hydraulic jacks were mounted to the test floor and attached through 2.1-m-long threaded steel rods (diameter 12.7 mm) to the top of the load distribution beam. The jacks were positioned symmetrically about the center of the wall with a spacing of 2.4 m, and provided a static distributed vertical load of 66.7 kN on the wall through the load distribution beam. The vertical load represented an equivalent occupancy floor load of 5.0 kPa from two stories over an area of 7.3 3 7.3 m, supported by four walls. A 386/25 personal computer data acquisition system with 16 channels and Lab-Tech (Tempe, Ariz.) Notebook data acquisition software was used to collect the experimental data at regular time intervals. The data collected included the lateral Downloaded 05 Jan 2009 to 222.66.175.206. Redistribution subject to ASCE license or copyright; see http://pubs.asce.org/copyright
terior nail spac Test Wall Configurations pane ar wall re constructed in ed:a 7 in tests that the nails aong the perimeter of a sh wall 、 Loading Schemes 7.8m nbased on recommendations in the ASTM Stan da to walls n cyelic te and dimensions of the door and window.as shown in Fig.2. TABLE 1.Shea r Wall Test Program FCC Protocol (Karacabeyli 1995) The FCC protocol consists of a sequence of sinusoidal cycle anel Lateral load 3 4) (5) the nominal yield slip and Table 2)The efined as the disp ace Cyelic (FCC) otonic test.The maximum 24×73 Note:N/A-nt available,FCC-Forintek Canada Corp. ayou the cycle frequer New Cyclic Loading Protocol which is typically obtained from the maximum load point of 20% B009 549 40 0 20 240 280 447 FIG.2.Layout of Shear Wall with Door and Window Openings FIG.3.FCC-Forintek Protocol 12/JOURNAL OF STRUCTURAL ENGINEERING/JANUARY 199 d 05 Jan 2009 to 222.66.175.206.Rodistribution sub t to ASCE ght:soo http:/ipubs.asco.org/copyright
12 / JOURNAL OF STRUCTURAL ENGINEERING / JANUARY 1999 FIG. 2. Layout of Shear Wall with Door and Window Openings FIG. 3. FCC-Forintek Protocol TABLE 1. Shear Wall Test Program Wall number (1) Panel size (m) (2) Openings (3) Panel layouta (4) Lateral load types (5) Nail spacing (mm) (6) Number of nails (7) 1 1.2 3 2.4 N/A A Monotonic 152 424 2 2.4 3 7.3 N/A B Monotonic 76 409 3 1.2 3 2.4 YES C Monotonic 152 395 4 2.4 3 7.3 YES D Monotonic 102 391 5 1.2 3 2.4 N/A A Cyclic (FCC) 152 424 6 2.4 3 7.3 N/A B Cyclic (new) 76 409 7 1.2 3 2.4 YES C Cyclic (new) 152 395 8 2.4 3 7.3 YES D Cyclic (new) 102 391 Note: N/A = not available; FCC = Forintek Canada Corp. a Panel layout: load, vertical load, actuator displacements, lateral displacements at the top of the shear wall, uplift at the base of the shear wall, and relative movements between panel and frame at various locations. The sampling frequency was 10 Hz for cyclic load conditions. Test Wall Configurations The test shear walls, 2.44 3 7.32 m, were constructed in the laboratory employing standard wood frame construction details and using No. 2 and Better Spruce Pine Fir 38 3 89 mm lumber as framing members. A Bostitch-N80CB pneumatically driven nail gun with coil-fed 3.11f 3 76 mm (10d) common nails was used to connect the members. The stud members (2.4 m in length) were spaced 400 mm apart. The top plate and the end studs consisted of double members, whereas the bottom plate and the interior studs consisted of single members. Performance Rated W24 OSBs (CSA-0325.0- M88), 9.5 mm thick, were used as sheathing panels. The sheathing panels were connected to the framing members with 2.66f 3 50 mm (6d) pneumatically driven spiral nails. Eight shear walls were tested with one replicate per test type (Table 1). In tests 1, 3, 5, and 7, conventional 1.2 3 2.4 m panels were used with continuous blocking installed at the midheight of the walls. The conventional size panels were staggered and oriented with their long axes parallel to the length of the wall, as shown in Table 1, layout A. In all other tests, 2.4 3 7.3 m oversize panels were used (Table 1, layout B). Door and window openings were introduced to walls 3, 4, 7, and 8 as shown in layouts C and D in Table 1. The locations and dimensions of the door and window, as shown in Fig. 2, were selected in accordance with the Canadian Wood-Frame House Construction guide (Canada Mortgage and Housing Corporation 1985). At the vertical edges of openings, studs were doubled with the inner studs cut to reach the lintels which were end-nailed through the outer studs. The interior nail spacing in all the tests was 305 mm. In shear walls with conventional 1.2 3 2.4 m panels, a standard nail spacing of 152 mm around the perimeter of the panels was used. In shear walls sheathed with oversize 2.4 3 7.3 m panels, two nail spacing patterns along the perimeter of the shear wall were used: a 76 mm nail spacing in tests 2 and 6, and a 102 mm nail spacing in tests 4 and 8. These nail spacing patterns were chosen to keep the total number of nails in the various walls similar. Considering that the nails along the perimeter of a shear wall have a significant influence on the performance of the wall, the nail distribution patterns were studied and reported in a previous article (Lam et al. 1997), whereas in the present phase of shear wall study, more attention was paid to study the failure modes of nails and sheathing panels in shear walls with openings. Loading Schemes Two types of loading scheme were used: (1) monotonic loading for walls 1–4, and (2) cyclic loading for walls 5–8. The loading rate for the monotonic tests was approximately 7.8 mm/min based on recommendations in the ASTM Standard E564-73. In cyclic tests two protocols were used: (1) a Forintek Canada Corp. (FCC) cyclic test protocol in test 5; and (2) a newly proposed protocol in tests 6–8. FCC Protocol (Karacabeyli 1995) The FCC protocol consists of a sequence of sinusoidal cycle groups, each of which contains three identical cycles (Fig. 3). The amplitude of each cycle group is taken as a percentage of the nominal yield slip (Dyield) and is increased stepwise, with interspaced decreasing cycles, until specimen failure (Fig. 3 and Table 2). The nominal yield slip is defined as the displacement at one-half of the maximum load obtained during a monotonic test. The maximum load (Pmax), yield point (Dyield) from the monotonic tests in wall 1, and cycle frequencies for shear wall 5 are listed in Table 3. The amplitudes of cycle groups as a percentage of the yield point for wall 5 are listed in Table 2. Beyond cycle group 18 both the amplitude schedules and the cycle frequencies were adjusted because of limited data storage capacity. At this cycle group, the maximum load had been well exceeded. New Cyclic Loading Protocol The FCC test protocol is a particularly long test sequence, designed to capture the full stabilized load-deformation curve, which is typically obtained from the maximum load point of Downloaded 05 Jan 2009 to 222.66.175.206. Redistribution subject to ASCE license or copyright; see http://pubs.asce.org/copyright
Ampltud Schedule Percentage 890123456789012 20300500900000物 Time (sec 0咖 40 TABLE 3.Parameters in Cyclic Tests Using FCC Protocol Loading Rate Cycle Cycle (2 .08 were cited fro iotproocgYorisraetaTostProtocoltormas时 TEST RESULTS AND DISCUSSION The following values.summarized in Table 5.were obtained failure mo. Cyem (puvr)t the ae uni veles)(N). 、6 wall at maximun and v correspon (m)m er TABLE 5.Shear W ments in the first cycle group were very all and dic (5 walls.This cycle group was eliminated from subsequent tests. TABLE 4.Parameters in Cyclic Tests Using New Protocol v(mm) (mm) (mm) Loading 12345678 168 Defin rate +68.74.-48.65 391 (2)3) 678 cls and f the hydraulic cylinder action of the hy draulic cyli JOURNAL OF STRUCTURAL ENGINEERING/JANUARY 1999/13 2009 ASCE
JOURNAL OF STRUCTURAL ENGINEERING / JANUARY 1999 / 13 TABLE 2. Displacement Amplitude Schedule as Percentage of Dyield in FCC Protocol Cycle group (1) Wall number 5 (%) (2) 1 25 2 50 3 25 4 100 5 50 6 150 7 100 8 200 9 150 10 250 11 200 12 300 13 250 14 350 15 300 16 400 17 350 18 450 19 500 20 600 21 700 22 800 TABLE 3. Parameters in Cyclic Tests Using FCC Protocol Wall (1) Pmax a (kN) (2) Dyield a (mm) (3) Loading Rate Cycle group (4) Cycle frequency (Hz) (5) Cycle group (6) Cycle frequency (Hz) (7) 5 62.8 10.0 1–19 0.25 20–22 0.083 a Values were cited from monotonic tests. FIG. 4. New Cyclic Test Protocol: (a) Test Protocol for Wall 6; (b) Test Protocol for Walls 7 and 8 TABLE 5. Shear Wall Summary Test Results Wall (1) Pmax a (kN) (2) Du (mm) (3) Dyield (mm) (4) Su (kN/m) (5) G9 (MN/m) (6) D (7) N (8) 1 62.77 54.04 9.82 8.58 1.07 5.5 424 2 125.21 38.07 7.40 17.12 2.81 5.1 409 3 45.07 42.93 8.25 6.16 0.91 5.2 395 4 63.53 24.89 5.18 8.69 2.04 4.8 391 5 159.29, 251.69 31.93 12.62 — — — 424 6 1115.33, 2110.01 31.95 6.16 — — — 409 7 146.15, 233.01 49.76 10.35 — — — 395 8 168.74, 248.65 21.24 4.77 — — — 391 Note: Su = ultimate shear strength of shear wall; G9 = shear stiffness of shear wall; D = ductility ratio; N = total number of nail connectors between panels and firms. a 1 = extension of the hydraulic cylinder; 2 = contraction of the hydraulic cylinder. TABLE 4. Parameters in Cyclic Tests Using New Protocol Wall number (1) Pmax (kN) (2) v1 (mm) Value (3) Definition (4) v2 (mm) Value (5) Definition (6) v3 (mm) Value (7) Definition (8) Loading rate (mm/sec) (9) 6 125.2 4.1 D0.35Pmax 7.5 D0.5Pmax 18.2 D0.8Pmax 0.4 7 45.1 9.3 D0.5Pmax 21.0 D0.8Pmax — — 0.2 8 63.5 6.9 D0.5Pmax 13.1 D0.8Pmax — — 0.2 the third cycle in each group. Because of the extensive yielding induced in nails during the preultimate part of the test, it was commonly observed that nails failed due to low cycle fatigue and the wall would not be able to reach its maximum load carrying capacity. Such failures were rarely observed under real or simulated seismic conditions, however, and a new cyclic loading protocol was proposed to assure a more realistic failure mode. This proposed loading scheme, as used in shear wall tests 6, 7, and 8, consisted of two or three groups of cycles, three identical cycles in each group, and one final unidirectional loading (pushover) until the wall failed. The amplitudes of these cycle groups, v1, v2, and v3 corresponded to the displacement at a certain percentage of the maximum load, obtained from the monotonic shear wall tests. All parameters and cyclic amplitude schedules are listed in Table 4 and illustrated in Fig. 4. The number of cycle groups in tests 7 and 8 was reduced to two from three in test 6. It was observed that the displacements in the first cycle group were very small and did not contribute in any way to determine the cyclic behavior of the walls. This cycle group was eliminated from subsequent tests. TEST RESULTS AND DISCUSSION The following values, summarized in Table 5, were obtained from shear wall tests: Pmax = maximum load carrying capacity of shear wall (in cyclic tests for both positive and negative cycles) (kN); Du = displacement of shear wall at maximum load (mm); Dyield = yield slip defined as the top displacement of shear wall at one-half of maximum load (mm); Su = ultimate shear strength of shear wall (kN/m). Downloaded 05 Jan 2009 to 222.66.175.206. Redistribution subject to ASCE license or copyright; see http://pubs.asce.org/copyright
s.P () panelss more abrupt in tems of ate of egrada L=7.315 m,the length of shear wall measured paralle Figs.6-9 show the hysteresis curves from the cyclic tests ectively.n each a mirror image of品 fness of shear wall (MN/m) dis G=2,0 the shear s fo hrthe height of shear wall. ilar to those in ic test all D.wall 5 c Compared with the mor N at D= ated N=total number of nail connectors between panels and frame. (Fig4).All three walls reached their ultimate Influence of Openings on Performance of Shear Walls 水 wal wall n)had a 28% and a 15% on ir sons bet no ing) ing the stat test r )80600403020.100102030405007000 el with Top Wall Displacement (mm) with opening wall 4 had a much highe 55 (0 rea hed its pea .e ling at the comer of the nin Therea r,th orm of b e of load arate sma and at即 shear walls similar to the wall ath (Fg nt to tha although 70805时 3020.100102030405060700 the Top Wall Displacement (mm) the 120.1000-60 0 80100 Top V Top Wall Displacement (mm) ment (m FIG.5.Load Displacement Curves for Shear Walls 1-4 Curve for Wall 7with Mono 14/JOURNAL OF STRUCTURAL ENGINEERING/JANUARY 199 ded 05 Jan 2009 to 222.65.175.206.Rodistribution s to ASCE
14 / JOURNAL OF STRUCTURAL ENGINEERING / JANUARY 1999 FIG. 8. Cyclic Load-Displacement Curve for Wall 7 with Monotonic Envelope of Wall 3 FIG. 7. Cyclic Load-Displacement Curve for Wall 6 with Monotonic Envelope of Wall 2 FIG. 6. Cyclic Load-Displacement Curve for Wall 5 with Monotonic Envelope of Wall 1 FIG. 5. Load Displacement Curves for Shear Walls 1–4 Pmax Su = (1) L where L = 7.315 m, the length of shear wall measured parallel to the loading direction. G9 = shear stiffness of shear wall (MN/m). P H max G9 = (2) 2DyieldL where H = 2.438 m, the height of shear wall. D = ductility factor. Du D = (3) Dyield N = total number of nail connectors between panels and frame. Influence of Openings on Performance of Shear Walls The monotonic test results (Table 5 and Fig. 5) show that openings significantly decreased the strength (Su) and stiffness (G9) of shear walls. With a 152 mm conventional nail spacing and built with conventional sized panels, wall 3 (with openings) had a 28% drop in strength and a 15% reduction in stiffness compared with wall 1 (without opening). Comparisons between wall 2 (large panel, no opening) and wall 4 (large panel, with openings) indicate a 50% drop in strength and a 27% reduction in stiffness. The static test results also show differences in the performance of shear walls built with oversize and regular panels. By comparing the static test results between wall 4 (oversize panel with openings) and wall 3 (conventional panel with openings) an increase of stiffness and strength of 124 and 41%, respectively, was observed. More significantly, compared with wall 1 (conventional panels without opening), wall 4 had a much higher stiffness (91%) and reached a slightly higher peak load level with fewer nails. It was also noticed that wall 4 reached its peak load level of 64 kN when the panel started to fail in compression buckling at the corner of the door opening. Thereafter, the failure of the oversize panel progressed in the form of buckling and tearing at the corners of the door and window openings, resulting in a gradual reduction of load carrying capacity. Upon continued loading, the sheathing panel was torn into three separate smaller panels, and at approximately 45 mm wall displacement the wall consisted essentially of three individual shear walls, similar to the wall with conventional panels. From this stage onward, both walls followed a similar load-deformation path (Fig. 5). It is important to note that although the strength and stiffness of the walls increased significantly with the larger sheathing panels, the displacement at ultimate load was lower than for the walls sheathed with conventional panels. Furthermore, the postpeak behavior of the walls with large panels was more abrupt in terms of rate of strength degradation. Figs. 6–9 show the hysteresis curves from the cyclic tests together with the corresponding static envelope curves for walls 5–8, respectively. In each case, a mirror image of the monotonic curve was drawn in the lower left quadrant to compare with the cyclic curves during the negative displacement phases. With the exception of wall 5, the peak load values for the shear walls under cyclic loading conditions were very similar to those in the monotonic tests. Compared with the monotonic test (wall 1), wall 5 (Fig. 6) experienced an early peak load of 59 kN at a much smaller displacement followed by a substantial drop in the load carrying capacity. This was caused by low cycle fatigue failure of the nails as a result of repeated reversed bending. To avoid nail fatigue failure, the newly proposed cyclic test protocol with fewer load cycles was used in tests 6, 7, and 8 (Fig. 4). All three walls reached their ultimate Downloaded 05 Jan 2009 to 222.66.175.206. Redistribution subject to ASCE license or copyright; see http://pubs.asce.org/copyright
(Nx)peo L 120 oment Curve for Wall s with Mono and ductility.During these testsno (b) 如a aninstegc and an un when the wall was subjec to a loading push-forward s with c rsize panel.which the wall in the pull 5997 dy of shea Failure Modes and frame. h The failur the wa nce of per the in buildings subected to earthquake d decrease afte serve ken the re lo-deecve of Shear Wall Predicted by dictior opening curve of the all veie e results The Panon-Mallory Equation (1985 To determine the effect of door and windo The ckling and tearing staried to wal ∑L rp-M= Bec of the ince of the wal full height sh l failures on the m益 thed h.For the JOURNAL OF STRUCTURAL ENGINEERING/JANUARY 1999/15 5 n2009to222.5 CE
JOURNAL OF STRUCTURAL ENGINEERING / JANUARY 1999 / 15 FIG. 9. Cyclic Load-Displacement Curve for Wall 8 with Monotonic Envelope of Wall 4 FIG. 10. Locations of Panel Failures with Failure Load Levels: (a) Shear Wall with Regular Panels; (b) Shear Wall with Oversize Panel load in the pushover phase and exhibited very similar envelope curves for the monotonic and cyclic tests with slight variances in maximum load capacity and ductility. During these tests no nail fatigue failure was observed. Note that in test 8 an error in the controlling system caused an instant push-forward of the hydraulic cylinder and an unexpected cycle had been added to the wall before the stipulated loading scheme was applied. Degradations of both stiffness and strength in wall 8 could be observed in the cyclic test when the wall was subjected to a loading in the push-forward direction (Fig. 9, upper right). It seemed that this impact load did not reduce the load capacity and stiffness of the wall in the pull-back loading condition (Fig. 9, lower left). Failure Modes The failure modes in walls with openings differed signifi- cantly from those without openings. In the latter, almost all the deformations occurred in the nails along the panel edges, whereas panel failures were observed in the walls with openings. The openings in shear walls resulted in stress concentrations around openings making perforated shear walls more prone to panel failures. The panel failures were particularly significant for the walls with one large panel and resulted in a relatively steep load decrease after ultimate. Because the large panels tried to displace as a single unit, very high shear and moment forces had to be transmitted through the narrow strips above and below openings. Once these links were broken, the remaining wall system behaved similarly to the one with conventional sized panels. In the monotonic test on wall 4, for example, the panel failure, which was first observed in the form of panel crushing, followed by buckling at the top corner of the door opening at a load level of 60 kN [Fig. 10(b)], occurred before significant nail failure were observed. After the wall reached its peak load of 64 kN, the panel gradually buckled at the top corner of the window opening and tore at the top corners of both door and window openings. The occurrence of panel failures eventually separated the oversize panel into three pieces. The eventual overlapping of the loaddisplacement curve of the wall with oversize panel with that of the wall with conventional panels indicated this panel breakdown (Fig. 5), which removed the benefit of the oversize panel. The remaining shear strength appeared to result solely from these smaller solid pieces of panels. Similar panel failures of buckling and tearing started to occur in the wall sheathed with conventional panels at a load level of approximately 37 kN, prior to its peak value of 45 kN. Because of the smaller original dimensions of the conventional panels, the influence of the panel failures on the performance of the wall was not as significant as that on the wall with oversize panel. Nail withdrawal, the most common nail failure mode, occurred mainly along the edges of the panels at the midheight of the wall sheathed with conventional panel and along the bottom edge of the panel in the shear walls with oversize panel, which was the same as in the shear wall tests without opening. As concluded in the previous experimental study of shear walls without opening (Lam et al. 1997), the performance of shear walls is mainly governed by the performance of the fasteners between panel and frame. These test results on perforated shear walls indicate, however, that both nail failure and panel failure contribute to strength reduction of the walls with openings. Consequently, nail reinforcements will not signifi- cantly improve the performance of perforated shear walls. As discussed before, nail low cycle fatigue failures limited the load carrying capacity of wall 5. Since this type of failure, which is caused by excessive cyclic yielding, was rarely found in buildings subjected to earthquake loads, a more realistic test protocol was introduced. Under the new cyclic test procedures, as used in walls 6, 7, and 8, no nail fatigue failures occurred and more realistic degradation characteristics were observed (He et al. 1997). Both monotonic and cyclic test protocols produced similar load-deflection envelope curves and failure modes. Lateral Resistance of Shear Wall Predicted by Empirical Equations Predictions were made of the lateral resistance of a perforated shear wall under monotonic loading condition by using empirical equations of Patton-Mallory et al. (1985) and Sugiyama (1994). These values are compared with experimental results. Patton-Mallory Equation (1985) To determine the effect of door and window openings, the ratio, rP2M, of opening-free or effective wall length is calculated as follows: O Li rP2M = (4) L where Li = length of opening-free or full height sheathed wall segment, whereas L = total wall length. For the walls tested, L1 = 1.655 m, L2 = 1.835 m, L3 = 1.296 m, and L = 7.315 m (Figs. 2 and 11). Downloaded 05 Jan 2009 to 222.66.175.206. Redistribution subject to ASCE license or copyright; see http://pubs.asce.org/copyright
showing that these ot be imum capacity of a shear wall without openings.P =-MPw Sugivama Equation (1994) m a good ing area first n 1+ (6) because oversize panels without op have a c H∑L 三 )that th F-3m corol which is defined as the ratio of the ultimate lateral load of a and nail s.whaalernlsoridtationanddime f she ings is then better pr (8) Numerical Analysis ragm For the shear wall with panels(wall 3).both ance of perl ar wal 89 del in this of fou tural g panel,fram Patton-Ma connectio umd that the panelis biect to A2 L1 2L3. n DaP for built with conventional 1 obtained from x FIG.11.Variables Used in Empirical Equations then be are 16 and t31 4 but Regula ive, (kN) 07 The (kN ificantly lower than the test values causing a sepa ation Patton-Mallory ne highest shea 303 13 previously) 16/JOURNAL OF STRUCTURAL ENGINEERING/JANUARY 1999 ed 05 Jan 2009 to 222.65.175.206.Redistrib o ASCE lic right:soo http:/pubs.asco.org/copyright
16 / JOURNAL OF STRUCTURAL ENGINEERING / JANUARY 1999 TABLE 6. Results from Empirical Equations Compared with Experimental Data Equations (1) Wall (2) 3 (3) 4 (4) Panel Regular panels Oversize panel Pmax (kN) 45.07 63.53 Pmax(W.O.) (kN) 62.77 125.21 Patton-Mallory equation rP-M Ppredict (kN) Error (%) 0.65 41.07 28.9 0.65 81.92 129.0 Sugiyama equation rs F Ppredict (kN) Error (%) 0.75 0.50 31.43 230.3 0.75 0.50 62.69 21.3 FIG. 11. Variables Used in Empirical Equations The ultimate lateral resistance, Ppredict, of a shear wall with openings is then predicted by multiplying rP2M with the maximum capacity of a shear wall without openings, Pmax(W.O.) Ppredict P = r P 2M max(W.O.) (5) Sugiyama Equation (1994) The Sugiyama equation determines the sheathing area ratio, rs, a function of the ratio of opening-free area to total sheathing area first 1 rs = (6) O Ai 1 1 H L O i where Ai = area of opening; and H = height of wall. For the walls tested, A1 = 1.989 m2 (door) and A2 = 1.890 m2 (window) (Figs. 2 and 11), whereas H = 2.438 m and Li is as defined above. The ratio, rs, is used to obtain the shear load ratio (F) for the apparent shear deformation angle of 1/100 radi and for ultimate stage rs F = (7) 3 2 2rs which is defined as the ratio of the ultimate lateral load of a shear wall with openings to that of a wall without openings. The predicted ultimate load capacity of a shear wall with openings is then Ppredict max(W.O.) = FP (8) The predicted load carrying capacities of shear walls with openings combined with experimental results are listed in Table 6. For the shear wall with conventional panels (wall 3), both equations conservatively underestimate the maximum capacity (8.9% by Patton-Mallory equation and 30.3% by Sugiyama equation). For the shear wall with oversize panel (wall 4), the Patton-Mallory equation overestimates the load carrying capacity by 29%, whereas the Sugiyama equation’s result is very close to the test value. It is evident that errors vary widely, showing that these simple equations may not be adequate to predict the performance of shear walls accurately. This is, however, understandable, because these are highly simplified equations and are intended for practical design applications using standard wood frame construction practice, which make them a good tool to provide a reasonable, and more importantly, conservative approximation of the strength of perforated shear walls. Furthermore, the predicted results for shear walls with different sheathing patterns are not consistent. For shear walls with regular panels, the predictions are conservative, whereas the predictions for shear walls with oversize panels tend to be overestimated. The approximate equations should thus not be used in this case. The discrepancy arises because oversize panels without openings have a comparatively high stiffness and strength level that can be lost to a large extent in the case of large openings. Another possible reason why the equations did not provide better prediction is that the empirical equations were developed based on test results of shear walls sheathed vertically with conventional panel and wall dimensions, while in present shear wall tests the orientation and dimension of sheathing panels were different from previous tests. It was also indicated in previous studies (Johnson and Dolan 1996; Line and Douglas 1996) that the error of prediction by the Sugiyama equation increases with increasing opening area ratios. Considering the variations in wall configurations (such as panel orientation and dimensions, and nail spacings), wall materials and test conditions, these equations may not be adequate in the analysis of shear walls in buildings other than standard 1-2 story residences. For a better prediction of the performance of shear walls, a more appropriate analytical methodology, such as numerical simulations, should be adopted. Numerical Analysis A nonlinear finite-element program, shear wall/diaphragm analysis program (DAP), developed by Foschi (1990), was used to predict the performance of perforated shear walls. The shear wall model in this program consists of four basic structural components: sheathing panel, frame, connections between frame members, and connections between panel and frame. It is assumed that the panel is subject to a plane stress field and has elastic orthotropic material properties. The frame members are considered as beam elements and assumed to be linear elastic. Nonlinear spring elements are used to model two types of nail connectors either between frame members or between panel and frame. The nonlinearity of the shear wall assembly is therefore solely determined by the behavior of the nail connectors. The numerical results from DAP for perforated shear walls built with conventional and oversize panels combined with experimental curves are shown in Fig. 12. The properties for the components used in the model were obtained from experimental tests. It can be seen that these numerical curves reveal convincingly the nonlinear behavior of perforated shear walls. The predicted lateral load carrying capacities for shear walls 3 (conventional panels with openings) and 4 (oversize panel with openings) are 38 kN and 62 kN. They are 16 and 3% lower than the results from tests, respectively, showing conservative, but reasonable estimates in shear wall strength. The predicted stiffness values for both walls, however, are significantly lower than the test values, causing a separation of numerical curves from experimental curves from the very beginning. After examining the distribution of in-plane shear stresses calculated by DAP, it is observed that the highest shear stresses occur around the corners of door and window openings, presenting a good prediction in panel failure modes which matched observations in the shear wall tests (discussed previously). Downloaded 05 Jan 2009 to 222.66.175.206. Redistribution subject to ASCE license or copyright; see http://pubs.asce.org/copyright
wic a a the sc ight of th were th same 8 op Wau Mallor ate the d by Pat 四 and overestimate that of she of his still the -FEM Resut ded the 6 ACKNOWLEDGMENTS FIG.12.Compa n between Test Results a ) Panel with Openings sh Col Re Since the program (DAP)considers only in-pla ane linear ons from DAP from Paul Symons in with thank pre APPENDIX I. REFERENCES C(5) CONCLUSIONS Openings in wood-based shear walls caused a ,Dodds,R.H for wood ngthen 2478 oversize panels aaona hou )Part he6e鸣r Fosch found to be slightly lower o in the He. by the e b panel failures.Pa failure D. JOURNAL OF STRUCTURAL ENGINEERING/JANUARY 1999/17 2009 SCE
JOURNAL OF STRUCTURAL ENGINEERING / JANUARY 1999 / 17 FIG. 12. Comparison between Test Results and Numerical Results: (a) Shear Wall 3, Conventional Panels with Openings; (b) Shear Wall 4, Oversize Panel with Openings Since the program (DAP) considers only in-plane linearelastic deformation of the panels, the observed highly nonlinear panel failure modes such as out-of-plane buckling and tearing are not represented. Although the predictions from DAP and the experimental results agreed reasonably well, the program does not provide a true representation of the failure mechanism of the physical system under the larger loads. CONCLUSIONS Openings in wood-based shear walls caused a significant decrease in shear strength and stiffness of shear walls, because of the reduction in the effective sheathing area. To strengthen the racking resistance of shear walls with openings, large nonstandard panels were used. When shear walls with openings were built with these oversize panels, a significant improvement in shear wall performance resulted, obtaining even higher shear strength and stiffness than walls without openings built with conventional panels. The results also indicate that the impact of openings on a shear wall with oversize panel is more significant in contrast to corresponding reductions in shear walls with conventional panel. In addition, the ductility ratios of the walls with openings were found to be slightly lower than those without opening. The failure modes in shear walls with openings differed from those shown in the shear walls without opening. The performance of shear walls was not only governed by the behavior of the nail connectors between panel and frame, but a combination of nail and panel failures. Panel failure modes were in the form of panel crushing and buckling and panel tearing around the corners of openings. Nail withdrawal was the dominant nail failure mode in the tests (except for wall 5, in which nail low cycle fatigue occurred), which occurred mainly along the edges of the panels at the midheight of the wall with conventional panels and along the bottom edge of the panel in the shear wall with oversize panel. Both panel and nail failure modes and locations were the same under monotonic and cyclic loading conditions. Considering the fact that panel failure locations were mainly concentrated in the corner regions of openings, localized reinforcement methods should be further investigated. The selection of a suitable cyclic loading protocol proved to be an important issue. A new protocol was used that seemed to better represent the amount of energy dissipation expected in an earthquake while avoiding unrealistic failure modes. Empirical equations developed by Patton-Mallory and Sugiyama tend to underestimate the lateral load capacity of shear walls with conventional panels and overestimate that of shear walls with oversize panels. For the small number of walls tested, the errors vary in a wide range and no decisive conclusions can be made. Considering the variations in shear wall configurations and test conditions, it was found that the equations may not predict the performance of shear wall universally and should be used with caution in shear wall analysis. As a contrast, a numerical analysis of the shear walls provides a better simulation of shear wall behavior not only in strength and stiffness values, but also in predicting areas of high stresses that coincided with observed failure modes. A significant discrepancy still exists between the predicted data by numerical method and the test results and fine tuning of the program is needed to provide more consistent results in both real and simulated structural systems. ACKNOWLEDGMENTS The writers gratefully acknowledge financial and in-kind support from the Structural Board Association, the National Research Council-Industry Research Assistant Program, and the Departments of Wood Science and Civil Engineering of the University of British Columbia. Forintek Canada Corp. and the Natural Sciences and Engineering Research Council are acknowledged for providing funding support to the research program of the third author (FSP0166869). Ainsworth Lumber Co. Ltd. is also thanked for providing the testing material. The technical contribution from Paul Symons in the research program is acknowledged with thanks. APPENDIX I. REFERENCES Canada Mortgage and Housing Corporation. (1985). Canadian woodframe house construction, CMHC, Ottawa. Dolan, J. D. (1989). ‘‘The dynamic response of timber shear walls.’’ PhD dissertation, Dept. of Civ. Engrg., Univ. of British Columbia, Vancouver, BC Canada. Easley, J. T., Foomani, M., and Dodds, R. H. (1982). ‘‘Formulas for wood shear walls.’’ J. Struct. Engrg., ASCE, 108(11), 2460–2478. Enjily, V., and Griffiths, R. D. (1996). ‘‘The racking resistance of large wall panels.’’ Proc., Int. Wood Engrg. Conf., Vol. 2, Omnipress, Madison, Wisc., 321–328. Falk, R. H., and Itani, R. Y. (1987). ‘‘Dynamic characteristics of wood and gypsum diaphragms.’’ J. Struct. Engrg., ASCE, 113(6), 1357– 1370. Filiatrault, A. (1990). ‘‘Static and dynamic analysis of timber shear walls.’’ Can. J. Civ. Engrg., Ottawa, 17, 643–651. Foschi, R. O. (1977). ‘‘Analysis of wood diaphragms and trusses, Part one: Diaphragms.’’ Can. J. Civ. Engrg., Ottawa, 4(3), 345–352. Foschi, R. O. (1990). ‘‘Diaphragm analysis program-user’s manual (Version 1.1).’’ University of British Columbia, Vancouver, BC, Canada. Gupta, A. K., and Kuo, G. P. (1985). ‘‘Behavior of wood-framed shear walls.’’ J. Struct. Engrg., ASCE, 111(8), 1722–1733. He, M., Lam, F., and Prion, H. G. L. (1998). ‘‘Influence of cyclic test protocols on performance of wood based shear walls.’’ Can. J. Civ. Engrg., Ottawa, 25, 539–550. Itani, R. Y., and Cheung, C. K. (1984). ‘‘Nonlinear analysis of sheathed wood diaphragms.’’ J. Struct. Engrg., ASCE, 110(9), 2137–2147. Johnson, A. C., and Dolan, J. D. (1996). ‘‘Performance of long shear walls with openings.’’ Proc., Int. Wood Engrg. Conf., Vol. 2, Omnipress, Madison, Wisc., 337–344. Downloaded 05 Jan 2009 to 222.66.175.206. Redistribution subject to ASCE license or copyright; see http://pubs.asce.org/copyright
Co Vol l,Hoom.Den am. APPENDIX II.NOTATION The following symbols are used in this paper area of opening (m) 1985 110.227 1 二of (/m) ram mec管8表47Khp P 10 P=predicted maximum load carrying capacity of shear ing-free ratio or sh on,W..(1978)."R 04(7)1- f(Nm). wall analysis.' wall at maximum load (mm) sponse of timber shear =nominal yield displacement (mm) 18/JOURNAL OF STRUCTURAL ENGINEERING/JANUARY 1999 oaded 05 Jan 2009 to 222.66.175.206.Redistribution subiect to ASCE B
18 / JOURNAL OF STRUCTURAL ENGINEERING / JANUARY 1999 Karacabeyli, E. (1995). ‘‘Lateral resistance of nailed shear walls subjected to static and cyclic displacements.’’ Res. Rep. Presented at FPS 49th Annu. Meeting, Forintek Canada, Vancouver. Lam, F., Prion, H. G. L., and He, M. (1997). ‘‘Lateral resistance of wood shear walls with large sheathing panels.’’ J. Struct. Engrg., ASCE, 123(12), 1666–1673. Line, P., and Douglas, B. K. (1996). ‘‘Perforated shearwall design method.’’ Proc., Int. Wood Engrg. Conf., Vol. 2., Omnipress, Madison, Wisc., 345–349. Patton-Mallory, M., Wolfe, R. W., Soltis, L. A., and Gutkowski, R. M. (1985). ‘‘Light-frame shear wall length and opening effects.’’ J. Struct. Engrg., ASCE, 111(10), 2227–2239. Polensek, A. (1976). ‘‘Finite element analysis of wood-stud walls.’’ J. Struct. Engrg., ASCE, 102(7), 1317–1335. Rose, J. D., and Keith, E. L. (1996). ‘‘Wood structural panel shear walls with gypsum wallboard and window/door openings.’’ Res. Rep. 157, American Plywood Association, Tacoma, Wash. ‘‘Standard method of static load test for shear resistance of framed walls for buildings. ASTM standard E 564-76, 04.07, ASTM Philadelphia, Pa. (1991). Sugiyama, H. (1994). ‘‘Empirical equations for the estimation of racking strength of a plywood-sheathed shear wall with openings.’’ Summaries of Tech. Papers of Annu. Meeting, Architectural Institute of Japan, Japan. Tuomi, R. L., and McCutcheon, W. J. (1978). ‘‘Racking strength of lightframe nailed walls.’’ J. Struct. Engrg., ASCE, 104(7), 1131–1140. White, M. W., and Dolan, J. D. (1995). ‘‘Nonlinear shear-wall analysis.’’ J. Struct. Engrg., ASCE, 121(11), 1629–1635. White, M. W., and Dolan, J. D. (1996). ‘‘Seismic response of timber shear walls.’’ Proc., Int. Wood Engrg. Conf., Vol. 2, SBI, Horsholm, Denmark, 329–336. APPENDIX II. NOTATION The following symbols are used in this paper: Ai = area of opening (m2 ); D = ductility ratio; F = shear load ratio; G9 = shear stiffness of shear wall (N/m); H = height of shear wall (m); L = length of shear wall (m); Li = length of opening-free wall segment (m); N = total number of nail connectors between panels and frame; Pmax = maximum load carrying capacity of shear wall (kN); Pmax(W.O.) = maximum capacity of a shear wall without openings (kN); Ppredict = predicted maximum load carrying capacity of shear wall with openings (kN); r = opening-free ratio or sheathing area ratio; Su = ultimate shear strength of shear wall (N/m); v = amplitude of cycle (mm); Du = displacement of shear wall at maximum load (mm); and Dyield = nominal yield displacement (mm). Downloaded 05 Jan 2009 to 222.66.175.206. Redistribution subject to ASCE license or copyright; see http://pubs.asce.org/copyright
