See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/238072956

The racking performance of shear walls with various aspect ratios. Part 1.
Monotonic tests of fully anchored walls

Article in Forest Products Journal · October 2003

CITATIONS READS

12 1,289

2 authors:

Alexander Salenikovich J. Daniel Dolan

Université Laval Washington State University
56 PUBLICATIONS 745 CITATIONS 87 PUBLICATIONS 1,746 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by J. Daniel Dolan on 09 April 2015.

The user has requested enhancement of the downloaded file.

 The racking performance of shear walls
with various aspect ratios. Part 1.
Monotonic tests of fully anchored walls

✳
Alexander J. Salenikovich
✳
J. Daniel Dolan

Abstract
Current design values for light-frame timber shear walls are based on results of standard monotonic tests of 2.4-m (8-ft.) square
walls restrained against overturning. The shear resistance of walls is calculated in terms of load per unit length, assuming that shear
forces are distributed uniformly throughout the wall of any size. In past earthquakes, structural failures occurred near large open-
ings where the lateral forces were transmitted through narrow wall segments. To improve our understanding of shear wall racking
performance and to facilitate further development of seismic design methodology, a comprehensive study has been conducted that
combines experimental and numerical analyses of shear walls of various configurations. This study included static monotonic and
cyclic tests of full-size shear walls with height-to-length ratios of 4:1, 2:1, 1:1, and 2:3. Discussed in this paper is the static
monotonic response of shear walls with overturning restraint representative of segmented wall construction practices. These walls
were attached to the foundation by means of hold-down anchors and shear bolts. Test results revealed that the performance of seg-
mented walls did not depend on the aspect ratio with the exception of narrow (4:1) walls, which exhibited 50 percent lower stiffness
per unit length relative to the other walls tested. Differences in failure patterns and reduction in deformation capacity were ob-
served when low-density studs were used at the wall ends. Traditional methods of analyzing segmented shear walls with hold-down
anchors were shown to be sufficiently conservative.

alls as components of the lateral bers are called struts. Struts collect the resist the overturning moment created by
force resisting system of a building are horizontal forces from the upper parts of the shear forces. Segmented wall design
defined as shear walls. When resisting the building through framing fasteners practice usually requires that the chords
wind or an earthquake, shear walls act as and transfer the load to the sheathing be attached to the lower structures (foun-
vertical cantilevers transferring the lat- through the sheathing fasteners. dation or story below) through hold-
eral forces from the upper parts of the Sheathing panels provide racking resis- down anchors and anchor bolts to re-
building to the foundation. In light-frame tance, and transfer the load to the chords strain shear walls from overturning.
buildings, shear walls typically consist of through the sheathing fasteners. Chords Bolts attaching the struts to the upper and
lumber framing and exterior and interior
panel sheathing attached with dowel- The authors are, respectively, Assistant Professor, Dept. of Wood and Forest Sciences, Fac-
type fasteners (nails, staples, or screws). ulty of Forestry and Geomatics, Laval Univ., Quebec, Canada (formerly Post Doctoral Asso-
Exterior walls often represent a combi- ciate, Forest Prod. Lab., Mississippi State Univ.); and Professor, Dept. of Civil and Environ-
nation of fully-sheathed segments of var- mental Engineering, Washington State Univ., Pullman, WA 99164-2910. The authors
respectfully acknowledge the financial support given to the research project by the
ious lengths interrupted by windows and USDA/CSREES NRI Competitive Grant Program through award number 95-37103-2102.
doors of various sizes. In shear wall anal- Approved as Journal Article No. FP-231 of the Forest & Wildlife Research Center, Missis-
ysis, the vertical framing members at the sippi State Univ. This paper was received for publication in June 2002. Article No. 9503.
✳Forest Products Society Member.
ends of the wall segment are called ©Forest Products Society 2003.
chords and the horizontal framing mem- Forest Prod. J. 53(10):65-73.

FOREST PRODUCTS JOURNAL Vol. 53, No. 10 65

 Tech that combines experimental and
numerical analyses of shear walls of var-
ious configurations, including seg-
mented and conventional walls
(Salenikovich 2000). Presented in Part 1
of this study are results of monotonic
tests on shear walls with hold-down an-
chorage representative of segmented
wall design practices. The objective was
to analyze the effects of aspect ratio on
shear wall strength and deformation
characteristics under static monotonic
load. Observations of failure patterns
and measurements of sheathing and
framing displacements and forces in an-
chor bolts were conducted to describe
Figure 1. — Geometry and notation of wall specimens. the shear wall racking performance. Cy-
clic test results are presented in Part 2 of
lower structures to prevent wall sliding the stiffness and strength of long walls the study (Salenikovich and Dolan
are called shear bolts. As opposed to the were in linear proportion to their length 2003). The tests on conventional shear
segmented wall construction, conven- (Patton-Mallory et al. 1984), while other walls will be evaluated in subsequent ar-
tionally built walls are secured to under- observations did not confirm this (Wolfe ticles.
lying structures by nails or shear bolts 1983). Stiffness and strength of narrow
walls were found greatly reduced, be-
Experimental
only; therefore, the overturning is re-
sisted only by the weight of the building cause any hold-down movement was Specimens
and the sheathing nails along the bottom magnified by the high aspect ratio
All shear wall specimens were 2.4-m
of the wall. (Commins and Gregg 1994). Evaluation
(8 ft.) tall. Four aspect ratios were con-
of single-family residential buildings in
The height-to-length ratio of a shear sidered (Fig. 1):
the 1994 earthquakes in Northridge,
wall is called the aspect ratio. Shear wall
design traditionally uses the shear resis-
California (Andreason and Rose 1994) • 4:1 – the size often accommodating
revealed that damage occurred at the garage doors and wide “view” win-
tance calculated in terms of load per unit narrow wall segments near large open- dows.
length without regard to the aspect ratio, ings such as wide windows and garage
assuming that the lateral force is distrib- doors and along the top of the founda- • 2:1 – the minimum width of the tradi-
uted uniformly throughout the total tion. A thorough wall bracing, together tional wall allowed in U.S. model
length of all shear panels without open- with intensive fastening of sheathing to codes to resist high seismic loads. It is
ings (Breyer et al. 1999). However, there framing and framing to foundation, was the typical width of fully sheathed
is an opinion that the shear distribution recommended for narrow walls to de- wall segments filling the space be-
is not uniform (Diekmann 1997). Cur- velop required shear resistance (Tissel tween windows and doors in residen-
rent design values for shear walls were and Rose 1994). tial buildings.
proposed by the American Plywood As-
sociation (APA) based on static This information suggests that wall • 1:1 – the ASTM standard specimen
monotonic tests on 2.4-m (8-ft.) square aspect ratio, hold-down restraint, and size serving as a reference point for
walls fully restrained against overturn- quality of sheathing attachment deter- comparison with other tests.
ing (Tissel 1990). Tests were performed mine the stiffness and strength of shear
using static monotonic (non-reversed) walls. In the transition to perfor-
• 2:3 – the size for investigating if lon-
ger walls have the same performance
loading applied in several stages at a mance-based design philosophy, all characteristics as square walls.
uniform rate of displacement according these factors may be important to con-
sider when establishing multiple perfor- Throughout the study, materials and
to ASTM standards E 72 (ASTM framing techniques were kept constant.
1995a) and E 564 (ASTM 1995b). mance levels for the structures. In light
of the ongoing discussions, it is neces- Acronymic notation of individual test
Very few investigations in shear wall sary to investigate strength and stiffness specimens was used to indicate the wall
performance have focused on aspect ra- of shear walls in a wide range of aspect length in feet (02, 04, 08, or 12), the type
tio effects. White and Dolan (1994) in- ratios using uniform manufacturing and of overturning restraint (FA = full an-
vestigated the effects of aspect ratio on testing procedures. Information ob- chorage), the load regime (m =
shear wall response using numerical tained in such a study can improve our monotonic), and replication number (1,
methods, and validated the model for understanding of the racking perfor- 2, etc.). ASTM E 564 standard practice
shear walls with a 1:1 aspect ratio. Dif- mance of shear walls and will contribute requires testing a minimum of two shear
ferent materials, fabrication techniques, to further development of seismic de- wall specimens of a given configuration.
and test procedures used in most sign methodology. A third specimen is tested if the differ-
previous studies prohibit direct compar- ence between the strength or shear stiff-
ison of the results. Some research results This study is part of a comprehensive ness of the two specimens exceeds 15
supported the common assumption that research program conducted at Virginia percent.

66 OCTOBER 2003
 the vertical edges. As a means of over-
turning restraint, Simpson Strong-Tie®
HTT22 connectors were attached on the
inside of the chords by thirty-two 16d
(‡3.8- by 82.6-mm) sinker nails.
Test setup
The specimens were stored in the lab-
oratory ambient conditions for at least 2
weeks after fabrication to allow for
wood relaxation around the nails. Each
specimen was tested in a horizontal po-
sition (Fig. 2). No dead load was applied
in the plane of the wall, which conserva-
tively represented a wall parallel to floor
joists. The wall was supported by two
76- by 127-mm (3- by 5-in.) steel beams
attached to the wall top and bottom
Figure 2. — Test setup. plates with ‡15.9-mm (5/8-in.) bolts
spaced 0.6 m (24 in.) on centers. These
bolts were secured by nuts with the use
of 64- by 64-mm- (2.5- by 2.5-in.-) wide
and 6-mm- (0.25-in.-) thick steel plate
washers. In 0.6-m (2-ft.) walls, the dis-
tance between the shear bolts was re-
duced as shown in Figure 3.
To eliminate interference of the sup-
port with the sheathing displacements,
the narrow face of the support beams
was oriented toward the plates. To re-
duce the amount of the wall slip along
the support during the test, the oversize
of holes for the shear bolts was mini-
mized. Holes in the supporting beams
were only 0.8 mm (1/32 in.) larger than
the bolt diameter; holes in the top and
bottom plates were drilled without over-
size. At the ends, the wall was anchored
to the supporting beam by ‡15.9-mm
(5/8-in.) bolts through the hold-down
connectors. Anchor bolts were instru-
mented with strain gages to measure the
tension forces transferred to the chords
through the hold-down anchors. The
bolts were tightened to approximately
Figure 3. — Anchorage and effective length of 0.6-m (2-ft.) wall. 18 kN (4 kips) of initial tension. The
holes for the anchor bolts were oversized
Prior to wall assembly, modulus of common nails every 0.60 m (2 ft.). All by 13 mm (1/2 in.) to minimize the base
elasticity and density of each framing studs were attached to the single bottom shear effects on the tension force mea-
member were measured using a plate and the double top plate, with two surements in the instrumented bolts.
Metriguard® Model 340 transverse vi- 16d common nails at each end. A single The steel beam at the bottom plate
bration E-computer (Metriguard 1994); layer of OSB sheathing, 11 mm (7/16 was secured to the reinforced-concrete
moisture content was measured using a in.) thick, was attached to one wall side reaction wall. The steel beam at the top
Delmhorst® Model J-3 moisture meter. by power-driven 8d (‡3.3- by 63.5-mm) plate distributed the racking load from a
The wall frame was assembled of 38- by common SENCO® nails at 0.15 m (6 programmable hydraulic actuator. The
89-mm (2- by 4-in.-nominal) spruce- in.) on centers along the edges and 0.30 actuator, with a displacement range of ±
pine-fir (SPF) stud grade members m (12 in.) on centers along intermediate 152 mm (6 in.) and a capacity of 245 kN
spaced 0.41 m (16 in.) on centers, except studs. The long dimension of the sheath- (55 kips), was secured between the
for the walls with the aspect ratio of 4:1 ing was oriented parallel to the studs and support and the distribution beam by
with studs at the wall ends only (Fig. 1). fastened to the framing with a 19-mm means of the hinged connections shown
The chords consisted of two studs fas- (3/4-in.) edge distance along the top and in Figure 2. If these hinges were omit-
tened by two 16d ( 4.1- by 89-mm) bottom plates, and 10 mm (3/8 in.) along ted, the separation of the wall framing

FOREST PRODUCTS JOURNAL Vol. 53, No. 10 67

 chords equals the overall wall length
(i.e., the total width of sheathing). The
thickness of the end studs and the loca-
tion of the anchor bolts are neglected. If
the anchors are located inside the wall,
this approach overestimates the actual
distance between the vertical reactions,
and, therefore, underestimates the forces
acting in the wall (Commins and Gregg
1994). To eliminate this error in our
analysis, the effective wall length (L0)
was introduced, measuring the distance
between the vertical reaction forces as
shown in Figure 3. It was assumed that
the compression reaction went through
the overturning point at the centerline of
the compression chord, and the tensile
reaction went through the center of the
anchor bolt on the tension side of the
wall. Since the anchor bolts were located
Figure 4. — Instrumentation of shear wall test specimen. inside the wall, L0 was shorter than the
overall length (L) by some 0.15 m (6 in.).
during the test would be restrained by ments were of limited use. Pot 5 (at- Although for long walls this difference
the weight of the equipment, and the tached to a rigid foundation) measured is insignificant, for narrow walls it is a
load measurement would be biased by lateral translation of the top of the right considerable change, increasing the ef-
the induced moment. Two casters at- end stud. The difference between read- fective aspect ratios in 1.2-m (4-ft.) and
tached to the load distribution beam al- ings of channels 1 and 5 illustrated the 0.6-m (2-ft.) walls by 14 percent and 33
lowed free movement of the specimen separation of the stud from the top plate percent, respectively. To illustrate the
top, parallel to the direction of the ap- and slip between the top plate and load racking resistance of shear walls with
plied load. The casters rolled along the distribution beam during the test. Pot 6 the effective aspect ratios, the racking
greased surface of plastic pads laid on recorded the horizontal slip of the wall loads (F) were normalized by the effec-
the concrete floor to reduce friction in- relative to the supporting steel beam. tive wall length, and the subsequent
duced by the wall weight. The load was Channels 7 and 8 represented anchor analysis was based on unit shear load (v)
applied at a constant rate of 15 mm/min. bolts instrumented with strain gages to (kN/m (kip/ft.)):
(0.6 in./min.) in a single stroke. Each measure tension forces in the anchors. v = F/L0 [1]
test was stopped when the specimen LVDTs (channels 13 and 14) aligned
fully exhausted its ability to resist load. Similarly, tension forces in anchor
with the wall edges were mounted on the
bolts were predicted as follows:
supporting steel beam to measure uplift
Instrumentation
displacement of the end studs. Walls T = F h/L0 [2]
and measurements with aspect ratios of 1:1 and 2:3 accom-
A data-acquisition system was equip- where:
modated additional LVDTs (15) and (15
ped with 16 channels. The monitored re- and 16), respectively, to measure uplift h = 2.44 m (96 in.), the height
sponses are indicated in Figure 4. All of intermediate studs. These observa- of shear wall
data were recorded at a frequency 15 tions helped in estimating the wall rota- Usually, shear wall deflections are
times per second. The hydraulic actuator tion. To measure the displacement of characterized by story drift after deduct-
contained the internal linear variable sheathing relative to the frame, one ing slippage of the wall between the sup-
differential transducer, LVDT (channel sheathing panel in each wall accommo- porting structures. In this study, mea-
1), and the load cell (channel 2) that sup- dated four LVDTs (9 to 12) near the surements on channel 6 showed that the
plied information on the applied dis- panel corners. The LVDT probes rested slip displacements were negligible, less
placement and force that was used for against polished steel plates attached to than 0.25 mm (0.01 in), assuming that
the load-deflection analysis of the tests. the studs to reduce friction when the the steel beam did not slip relative to the
Resistance potentiometers (pots) 3 probe moved along the stud. It was as- concrete wall. Therefore, the load-de-
and 4 measured diagonal elongation of sumed that channels 9 and 11 measured flection curves and parameters, such as
the wall between the top and bottom only vertical translation components shown in Figure 5, were obtained using
plates. The diagonal measurements and channels 10 and 12 measured the data from channels 1 and 2.
were taken to obtain information on horizontal components. The maximum load point determined
shear deformation of the wall assuming the wall maximum shear strength (vpeak)
the specimen distorted as a parallelo- Load-deflection parameters and deflection capacity ('peak). The fail-
gram (ASTM E 564). This assumption In traditional shear wall analysis ure point (vfailure, 'failure) was consid-
was valid only until separation of the (Breyer et al. 1999), the unit shears and ered to occur at 0.8 vpeak (i.e., when a
studs from top and/or bottom plates overturning moments are estimated as- 20% decrease of vpeak occurred). The
started; therefore, the diagonal measure- suming that the distance between the area under the load-deflection curve

68 OCTOBER 2003
 The bilinear EEEP curves depict how
an ideal perfectly elastic-plastic wall
would perform, dissipating an equiva-
lent amount of energy, and allow com-
parison of the nonlinear performance of
different walls on the equivalent energy
basis. The elastic shear modulus, G, was
obtained as follows:
G = ke h [4]

General observations
All walls exhibited a significant
amount of racking. Figure 6 shows a
graph of sheathing displacements rela-
tive to the framing near the corners of
the first panel in the 12FAm1 wall. Sim-
ilar displacements were observed in
08FAm walls. At peak loads, the hori-
zontal displacements reached 5 mm (0.2
Figure 5. — Performance parameters of shear walls. in.) and vertical displacements reached
10 mm (0.4 in.). Bearing and friction
forces between adjacent edges of
Table 1. — Performance parameters of fully anchored shear walls under monotonic sheathing panels might have provided
load. a for a relatively uniform distribution of
Units 12FAm1 12FAm2 08FAm1 08FAm2 04FAm1 04FAm2 02FAm1 02FAm2
the sheathing displacements relative to
vpeak
the framing in these walls (There was
kN/m 10.33 9.33 10.82 10.40 9.76 10.11 10.17 9.51
less than a 50% difference between cor-
kip/ft. 0.708 0.639 0.741 0.713 0.669 0.693 0.697 0.652 responding displacements at the top and
'peak mm 79.9 53.5 81.2 65.6 49.0 65.2 150.8 112.5 the bottom). In walls with one sheathing
in. 3.15 2.11 3.20 2.58 1.93 2.57 5.94 4.43 panel (04FAm and 02FAm), displace-
vyield kN/m 9.15 8.36 9.47 9.28 8.48 9.06 8.56 8.29 ments at the top were less than a half of
kip/ft. 0.627 0.573 0.649 0.636 0.581 0.621 0.587 0.568 those at the bottom, and horizontal dis-
'yield mm 13.7 12.5 14.2 12.4 10.3 13.2 20.8 24.6 placements were less than a half of verti-
in. 0.54 0.49 0.56 0.49 0.41 0.52 0.82 0.97 cal displacements. The displacements
'e mm 6.2 5.6 6.5 5.6 4.8 5.9 9.5 11.1
might be more uniform if the chords
were anchored at the top as well as at the
in. 0.24 0.22 0.25 0.22 0.19 0.23 0.38 0.44
b b
bottom.
'failure mm 116.8 87.2 107.2 107.2 62.4 105.4 152.4 152.4
Major performance characteristics of
in. 4.60 3.43 4.22 4.22 2.46 4.15 6.0 6.0
fully-anchored shear walls under static
G kN/mm 1.63 1.63 1.63 1.83 2.00 1.68 1.04 0.83 monotonic loads are shown in Table 1.
kip/in. 9.32 9.31 9.30 10.43 11.42 9.58 5.94 4.76 Eight walls of different sizes developed
wfailure kN˜m/m 1.01 0.68 0.95 0.94 0.49 0.90 1.26
a
1.19
a
the average vpeak of 10.1 kN/m (0.69
kip˜ft./ft. 0.226 0.152 0.213 0.211 0.109 0.201 0.283 0.268 kip/ft.) with a remarkably low coeffi-
aSeeFigure 1 for shear wall notation.
cient of variation (5%). These data indi-
bThe test was stopped before failure was observed. cate that narrow walls can develop the
same unit shears as long walls if calcula-
tions are based on a proper effective
limited by the failure point approxi- ness (ke = 0.4vpeak/'e, where 'e = de-
length. Note however, if the loads were
mated unit work to failure (wfailure) (i.e., flection at 0.4vpeak).1 To find the yield
normalized by the overall wall length
the energy dissipated by the wall of unit load (vyield), the following equation was
(width of shear panel), the shear forces
length). Using these data, the equivalent derived by equating the areas under the
in the narrow walls would be signifi-
energy elastic-plastic (EEEP) curve was observed and EEEP curves:
cantly underestimated.
derived as shown in Figure 5. The initial
slope of the EEEP curve, drawn through v yield Typical load-deflection curves of the
0.4 vpeak on the load-deflection curve, tested walls are presented in Figure 7.
2w failure
determined the unit secant elastic stiff- ( )2 k e [3] These graphs do not include walls
failure failure
ke 04FAm1 and 12FAm2, which are dis-
cussed in the next section. Graphs
1The value of 0.4v 02FAm and 08FAm represent the aver-
peak is a value that is used in several
national and international test standards for deter- where the expression in parentheses de- age curves from two matching tests. Es-
mining the initial stiffness of non-linear responses. termined the yield deflection ('yield) sentially, walls with aspect ratios d 2:1

FOREST PRODUCTS JOURNAL Vol. 53, No. 10 69

 the measured vertical displacement was
caused by the stud rotation about the
hold-down anchor; the nailslip and de-
formation of the hold-down anchor con-
tributed the remaining one-third of the
uplift. At typical failure, the right end
stud separated from the top plate and the
sheathing panels gradually unzipped
along the top and/or bottom plates with
the sheathing nails tearing through the
panel edges. Usually, the sheathing un-
zipped along one of the studs and bottom
plate with the nails pulling heads through
the sheathing as shown in Figure 6.
Specific gravity effects
Walls 04FAm1 and 12FAm2 are dis-
cussed separately because their perfor-
mance deviated from the other walls.
The response curves are shown in com-
Figure 6. — Typical sheathing displacements in a long fully anchored wall. parison with the matching walls in
Figure 10. Although the elastic stiffness
and the load capacity of the matching
walls differed less than 10 percent, there
was at least a 35 percent reduction in de-
flections at the peak and failure loads.
These walls dissipated approximately 40
percent less energy per unit length than
the other walls. In other words, the duc-
tility and toughness of these two walls
were significantly reduced when com-
pared to the matching walls.
Figure 11 illustrates the predominant
failure mode of wall 04FAm1: the
sheathing nails along the right end stud
were pulled out. Similar failure was ex-
perienced by wall 12FAm2. Information
on the specific gravity and location of
each framing member recorded during
wall manufacture suggested that these
failure modes and the lower perfor-
mance were associated with the lower
density of the right end studs (Table 2).
The right end studs in walls 04FAm1
Figure 7.— Load-deflection curves of fully anchored walls during monotonic tests. and 12FAm2 had significantly lower
specific gravity (0.42 and 0.37, respec-
developed identical load-deformation mm (6 in.), more than 10 times 'yield, tively) than the rest of the walls. It is,
patterns, reaching the maximum resis- without noticeable strength degradation, therefore, rational that the entire row of
tance at deflections beyond 64 mm (2.5 because there was small displacement nails along the low-density studs pulled
in.), and then gradually degrading. A 20 demand on sheathing-to-framing con- out and started the early failure mecha-
percent decrease in resistance occurred nections. The large drifts were due to a nism.
past 105-mm (4.1-in.) deflections, ap- greater effect of rigid body rotation: the Figure 12 shows the sheathing dis-
proximately seven times 'yield. uplift displacements of the end studs placements of wall 04FAm1. The verti-
Narrow (4:1) walls were approxi- contributed horizontal deflections in cal displacements at the bottom were
mately half as stiff relative to the longer proportion to the aspect ratio. three to four times larger than at the top.
walls, which explains their poor service Figure 8 shows typical vertical dis- A similar graph of vertical displace-
record during earthquakes. At a deflec- placements of the end studs. Since the ments was recorded for wall 12FAm2.
tion of 64 mm (2.5 in.), they resisted end studs were anchored to the founda- This information indicates that most of
only 8.6 kN/m (0.59 kip/ft.), approxi- tion, uplift displacements did not exceed the work was done by the nails at the
mately 15 percent less than the other 5 mm (0.2 in.) at peak load. Analyzing bottom plate. The lack of connection re-
walls. Nevertheless, these narrow walls the wall distortion pattern, it can be con- sistance along the end stud due to low
developed deflections exceeding 152 cluded that approximately two-thirds of wood density overloaded the nails at the

70 OCTOBER 2003
Table 2. — Specific gravity and strength parameters of shear walls under monotonic bottom plate and caused the early
load. strength degradation of the entire shear
12FAm1 12FAm2 08FAm1 08FAm2 04FAm1 04FAm2 02FAm1 02FAm2
wall.
SG averagea 0.49 0.46 0.53 0.53 0.46 0.46 0.47 0.47 Design implications
SG right end studb 0.53 0.37 0.56 0.52 0.42 0.49 0.44 0.50 According to the International Build-
Fpeak (kips)c 8.14 7.35 5.56 5.34 2.34 2.43 1.04 0.98 ing Code (ICC 2000), the allowable
vpeak/vseismic 2.96 2.67 3.10 2.98 2.80 2.90 2.91 2.73 shear for walls of the tested configura-
Fpeak/Fseismic 2.84 2.56 2.90 2.79 2.45 2.53 2.18 2.04 tion (11-mm [7/16-in.] vertical sheath-
Tpeak (kips) 5.48 4.68 4.23 4.95 4.65 5.15 7.23 4.02 ing attached with 8d nails with 150-mm
aSpecific gravity of all framing members based on the ovendry volume. [6-in.] spacing at panel edges, and with
bSpecific gravity of the right end stud based on the ovendry volume. stud spacing 410 mm [16-in.] on cen-
c1 kip = 1000 lbf = 4.45 kN.
ters) with framing of Douglas-fir-larch
or southern pine equals 3.79 kN/m
(0.260 kip/ft.). Allowable shears are ad-
justed for specific gravity of the framing
lumber using the specific gravity adjust-
ment factor = [1  (0.5  SG)] d 1, where
SG is specific gravity for the species of
lumber in the National Design Specifi-
cation for Wood Construction (AF&PA
1997). For SPF species, SG = 0.42.
Therefore, in seismic design, the allow-
able shear for the tested walls equals
vseismic = 3.79 u 0.92 = 3.49 kN/m
(0.239 kip/ft.). In wind design, a 40 per-
cent increase of design capacities is per-
mitted (ICC 2000): vwind = 3.49 u 1.40 =
4.89 kN/m (0.335 kip/ft.). Corre-
sponding design load capacities
(Fseismic and Fwind) can be estimated by
multiplying the allowable shears and the
overall wall length (L).
To compare the design values with
test results, Table 2 shows the ultimate
Figure 8. — Vertical displacements of studs. load (Fpeak) resisted by each shear wall
tested and strength ratios: vpeak/vseismic
and Fpeak/Fseismic. The corresponding
strength ratios for wind design (not
shown in the table) are 40 percent lower.
The comparison shows that design val-
ues for shear walls with the aspect ratios
d 1:1 are sufficiently conservative.
Strength ratios exceeded 2.8 for seismic
design (2.0 for wind design) with the ex-
ception of wall 12FAm2, which had a
very low-density right end stud. Walls
with the aspect ratio 2:1 developed simi-
lar unit shears as the longer walls; how-
ever, their strength ratios based on Fpeak
were less conservative, because the de-
sign capacity included the overall
length. To equalize the safety levels of
narrow walls and long walls the load ca-
pacities should be calculated using the
effective length (L0). Alternatively, ap-
propriate adjustment factors should be
recommended for narrow (2:1) walls,
depending on the anchorage conditions.
Although shear walls with the aspect
ratio 4:1 are not permitted in engineered
Figure 9. — Typical view of the wall bottom after failure. design (ICC 2000), Table 2 shows that

FOREST PRODUCTS JOURNAL Vol. 53, No. 10 71

 their shear capacity could be predicted
quite safely using current allowable val-
ues and the effective wall length. How-
ever, a reduction factor between 0.5 and
0.85 would be recommended to account
for the lower stiffness of these walls as
follows from previous discussion of
Figure 7. The amount of reduction
would depend on location and stiffness
of hold-down anchorage.
Table 2 shows the ultimate forces
(Tpeak) measured in the anchor bolt on
the tension side of each tested wall. On
the average, Tpeak measured 22.7 kN
(5.05 kips), while the allowable tension
load for the HTT22 hold-down anchor is
23.4 kN (5.26 kips) (Simpson 1995).
Figure 10. — Load-deflection curves of walls 12FAm2 and 04FAm1. None of the hold-down anchors or an-
chor bolts had any visible sign of dam-
age after the tests. These data indicate
that the anchors had sufficient over-
strength to provide a desirable failure
mechanism (i.e., yielding of sheath-
ing-to-framing connections). Figure 13
shows an example graph of measured
and predicted tension forces in anchor
bolts as a function of shear wall defor-
mation. The forces predicted by Eq. [2]
showed good correlation with the mea-
surements. These observations provide
experimental evidence that the use of the
effective wall length in shear wall analy-
sis offers accurate force predictions.
Conclusions
Testing walls in a wide range of aspect
ratios allowed direct comparisons
among the walls of different sizes. It was
found that the maximum shear strength
of fully anchored walls did not depend
Figure 11. — Typical failure mode of walls with low-density end studs. on the aspect ratio, and could be accu-
rately predicted assuming a proper ef-
fective wall length. Walls with aspect ra-
tios d 2:1 were equally stiff, on a unit
length basis, while narrow (4:1) walls
were half as stiff, because their deflec-
tions were magnified by rigid body rota-
tion in proportion with the aspect ratio.
The shear strength of narrow walls did
not degrade at high deflections due to
small displacement demand on sheath-
ing-to-framing connections.
Typically, uniform wood density al-
lowed better distribution of sheathing
displacements relative to the perimeter
framing and the development of full
load capacity of the sheathing-to-fram-
ing connections. However, when sheath-
ing nails withdrew from the wood be-
cause of low wood density, the shear
wall ductility and energy dissipation was
Figure 12. — Sheathing displacements of walls with low-density end studs. reduced up to 50 percent. The designer

72 OCTOBER 2003
 Diekmann, E.F. 1997. Diaphragms and
shearwalls. In: Wood Engineering and Con-
struction. 3rd ed. K.F. Faherty and T.G. Wil-
liamson, eds. McGraw-Hill, New York. pp.
8.47-8.79
Federal Emergency Management Agency
(FEMA). 1997. NEHRP Guidelines for the
Seismic Rehabilitation of Buildings (FEMA
Publication 273). FEMA, Washington, DC.
International Code Council (ICC). 2000. Inter-
national Building Code 2000. ICC, Falls
Church, VA.
Metriguard. 1994. Precision testing equipment
for wood. Catalog 21-1. Metriguard, Inc.,
Pullman, WA. pp.13-16.
Patton-Mallory, M., R.M. Gutkowski, and
L.A. Soltis. 1984. Racking performance of
light-frame walls sheathed on two sides. Res.
Paper FPL 448. USDA Forest Serv., Forest
Prod. Lab., Madison, WI.
Salenikovich, A.J. 2000. The racking perfor-
mance of light-frame shear walls. PhD diss.
Figure 13. — Tension forces in anchor bolts. Virginia Polytechnic Inst. and State Univ.,
Blacksburg, VA.
should use higher grade lumber or dense tion (NDS) for wood construction. AF&PA, __________ and J.D. Dolan. 2003. The packing
Washington, DC. performance of shear walls with various as-
wood species for framing members in pect ratios. Part 2. Cyclic tests of fully-an-
American Society for Testing and Materials
narrow shear wall segments when they (ASTM). 1995a. Standard test methods of chored walls. Forest Prod. J. (in press).
are critical components in the lateral conducting strength tests of panels for build- Simpson Strong-Tie Company. 1995. Wood
force resisting system (such as ground- ing construction. ASTM E 72-95. ASTM, construction connectors. Simpson Strong-Tie
floor shear walls with large openings) to West Conshohocken, PA. Company, Pleasanton, CA. 76 pp.
ensure acceptable performance. __________. 1995b. Standard practice for Tissel, J.R. 1990. Structural panel shear walls.
static load test for shear resistance of framed Report 154. American Plywood Assoc., Ta-
Comparison of monotonic test results walls for buildings. ASTM E 564-95. ASTM, coma, WA. 19 pp.
with published design values proved the West Conshohocken, PA.
Andreason, K.R. and J.D. Rose. 1994. __________ and J.D. Rose. 1994. Wood struc-
traditional design practice for seg- tural panel sheathing for narrow-width wall
mented shear walls to be sufficiently Northridge, California earthquake. Structural
performance of buildings in San Fernando bracing. Research Report 156. American Ply-
conservative. However, using the effec- Valley, California (January 17, 1994). APA wood Assoc., Tacoma, WA. pp. 20
tive wall length in shear wall analysis Rept. T94-5. American Plywood Assoc., Ta- White, M.W. and J.D. Dolan. 1994. Effect of
would enhance the accuracy and would coma, WA. 34 pp. openings and aspect ratio on the dynamic re-
equalize the safety levels of walls with Breyer, D.E., K.J. Fridley, and K.E. Cobeen. sponse of timber shear walls. Presented at the
1999. Design of Wood Structures ASD. 4th Forest Products Society 48th annual meeting.
various aspect ratios. Forest Prod. Soc., Madison, WI.
ed. McGraw-Hill, New York.
Commins, A.D. and R.C. Gregg. 1994. Cyclic Wolfe, R.W. 1983. Contribution of gypsum
Literature cited performance of tall-narrow shearwall assem- wallboard to racking resistance of light-frame
American Forest and Paper Association blies. Simpson Strong-Tie Co., Pleasanton, walls. Res. Paper FPL 439. USDA Forest Ser-
(AF&PA). 1997. National design specifica- CA. 12 pp. vice, Forest Prod. Lab., Madison, WI.

FOREST PRODUCTS JOURNAL Vol. 53, No. 10 73

View publication stats