Development of a Wood-Frame Shear Wall Model

in ABAQUS
Jian Xu1 and J. Daniel Dolan, F.ASCE2
Downloaded from ascelibrary.org by California State Polytechnic University, Pomona on 08/25/13. Copyright ASCE. For personal use only; all rights reserved.

Abstract: After modifications, the Bouc-Wen-Barber-Wen 共BWBN兲 model was used in simulating the hysteretic behavior of wood
light-frame shear walls. A complete shear wall model is composed of boundary framing members and two-diagonal modified BWBN
hysteretic springs. The parameters of the springs can be estimated based on the shear wall test data or detailed shear wall simulation
results. To verify the accuracy and efficiency of this shear wall model, a two-story residential building was simulated and a three-
dimensional dynamic analysis was conducted on it. The resultant structural responses agreed with the test results well.
DOI: 10.1061/共ASCE兲ST.1943-541X.0000031
CE Database subject headings: Finite element method; Shear walls; Frames; Seismic effects; Earthquakes.

Introduction used. This is not computationally efficient to model every nail

connector independently. A macro-shear wall model is needed,
A typical wood-frame shear wall is built using wood framing which represents the hysteretic behavior of a whole shear wall
members 共studs, sill plates, and top plates兲 and sheathing panels when subjected to lateral loads.
共plywood or oriented strandboard 共OSB兲 panels, etc.兲. The wood
framing members form a stand on which the sheathing panels are
attached by nails or other types of discrete fasteners. The framing Previous Models for Wood-Frame Shear Walls
members are used to resist vertical and out-of-plane loading. The
in-plane lateral loads are resisted by the racking of the sheathing To understand the characteristics of shear wall performance bet-
panels. Tests have shown that the most common failure modes of ter, many simulation studies were conducted based on the findings
a shear wall under lateral loads are the tearing of the sheathing in the corresponding testing. Some of the more recent simulation
and pullout of the sheathing fasteners 共Dolan 1989; He et al. studies are described here.
1999; Salenikovich 2000; Durham et al. 2001兲. On the other hand, Tarabia 共1994兲 and Tarabia and Itani 共1997a,b兲 accomplished
the sheathing fasteners are also the source of ductility for the modeling a whole 3D light-frame building using the finite-
shear wall. Basically, the performance of the sheathing fasteners element method 共FEM兲. Five elements were used to represent
controls the shear wall behavior. shear walls. A two-node linear element with two translational
Since shear walls are the most important component within the degrees-of-freedom 共DOF兲 at each end was used for the frame.
light-frame building system, accurate shear wall modeling is the Two-dimensional 共2D兲 plane elements were employed for sheath-
key in successful simulation of a whole system. In the writers’ ing. Sheathing interface elements were used to prevent overlap-
opinions, the accuracy and efficiency of the shear wall models ping of adjacent sheathing panels. Linear springs with different
available are not satisfactory because they are unable to capture values in tension, compression, and shear were used for framing
some important features of real structural behavior 共e.g., the connectors. Sheathing frame fasteners were modeled as a two-
strength degradation兲 and cannot accurately capture the energy perpendicular decoupled nonlinear spring system connecting
dissipation characteristics because of their simple representation sheathing and framing elements. A lumping technique was used to
of load-displacement relationship. These problems will become evaluate the stiffness matrix of each group of nails located on one
significant as larger structures with contributions from higher line as a single element. The hysteretic model of Kivell et al.
modes begin to be modeled.
共1981兲 was used to represent the hysteretic performance of nail
In Xu and Dolan 共2009兲, a detailed wood-frame shear wall
connections and intercomponent connections.
model has been developed in ABAQUS 共2005兲 successfully.
Folz and Filiatrault 共2001, 2004a,b兲 formulated a discrete
However, for a whole three-dimensional 共3D兲 building model,
mass-stiffness dynamic FEM model derived based on the princi-
hundreds or even thousands of nail connector elements must be
pal of virtual work to predict the hysteretic performance of light-
frame shear walls and buildings. This model has been
1
2
Design Engineer, KPFF Engineering, Seattle, WA 98101. incorporated into the computer program SAWS 共seismic analysis
Professor, Dept. of Civil and Environmental Engineering, of wood-frame structures兲. In this model, the connector and shear
Washington State Univ., Pullman, WA 99164 共corresponding author兲. wall hysteresis loops were composed of a backbone curve and
E-mail: jddolan@wsu.edu
piecewise, predefined straight line path segments between maxi-
Note. This manuscript was submitted on July 4, 2008; approved on
February 9, 2009; published online on March 2, 2009. Discussion period mum displacement and minimum displacement. Parameters of
open until January 1, 2010; separate discussions must be submitted for nail connectors were obtained from test data and the shear wall
individual papers. This paper is part of the Journal of Structural Engi- spring’s parameters were based on the cyclic analysis of detailed
neering, Vol. 135, No. 8, August 1, 2009. ©ASCE, ISSN 0733-9445/ shear wall models. The same backbone was used for both mono-
2009/8-977–984/$25.00. tonic and hysteresis curves. This light-frame structure model sim-

JOURNAL OF STRUCTURAL ENGINEERING © ASCE / AUGUST 2009 / 977

J. Struct. Eng. 2009.135:977-984.

 this shear wall model is the decoupling of in-plane and out-of-
plane responses. The other limitation is that it cannot account for
the interaction between intersecting walls when no tie-down an-
chors are used between studs and foundation.
Paevere 共2002兲 developed a differential hysteresis model that
was based on the phenomenological hysteresis Bouc-Wen-Barber-
Wen 共BWBN兲 model. Paevere 共2002兲 conducted a sensitivity
analysis of hysteretic modeling and provided a system identifica-
tion methodology for hysteretic modeling. He validated his model
with test results from a one-story building and used the model to
Downloaded from ascelibrary.org by California State Polytechnic University, Pomona on 08/25/13. Copyright ASCE. For personal use only; all rights reserved.

investigate the displacement demands placed on structures by dif-

ferent seismic spectra.

Rationale of This Study

Two simulation methods were considered in this study. One con-

sists of the framing members and a single hysteretic spring
connecting the top plate and the sill plate 关Fig. 1共a兲兴, which is
similar to the model developed by Folz and Filiatrault 共2001,
2004a,b兲. It is simple but can only represent racking behavior of
Fig. 1. Super shear wall model
the shear wall. The other uses a pair of diagonal hysteretic springs
关Fig. 1共b兲兴. This method was originally developed by Kasal and
Leichti 共1992兲 and Kasal et. al. 共1994兲 and used by Du 共2003兲 and
Collins et al. 共2005a,b兲. It is relatively simple and is able to ac-
plified the 3D structure into a 2D planar model composed of
count for uplift effect by introducing vertical springs along the sill
zero-height shear wall spring elements that connect the floor and
plate.
roof diaphragms together or to the foundation 共pancake model兲.
In this study, the first model is used only for spring parameter
All the horizontal diaphragms were assumed to be in-plane rigid.
estimation purposes, and the second is used for final simulation.
This model is simple and efficient. However, it cannot show the
influence of diaphragm rigidity on structural responses and cannot
represent the structural overturning and flexural response of shear
walls. Model Development
Collins et al. 共2005a,b兲 developed a light-frame structure
model in ANSYS. In this study, the nail connector model is based The modified BWBN model used for a single nail connector, as
on a phenomenological model presented by Dolan 共1989兲 and described in Xu and Dolan 共2009兲, was employed for the super
Kasal and Xu 共1997兲, in which the coupling of nail connections shear wall model with some further modifications
was not taken into account. In the shear wall model, a pair of Eq. 共11兲 was changed to
diagonal hysteretic nonlinear springs was used to represent in- ␣ = ␣0 ⫻ e共−0.01⫻M_Dis兲 共1兲
plane action of a shear wall. Plate elements were used to represent
the sheathing and beam elements for framing, which provide the And Eq. 共14兲 was changed to
out-of-plane resistance of wall assembly. The frame intersections 2/␨1.5兲兴
h共z兲 = 1.0 − ␨1 ⫻ e关−共z⫻sgn共u̇共t兲 − ratio ⫻ q ⫻ zu兲 2 共2兲
are modeled as pinned connections. This model is relatively de-
tailed, and it could reflect the out-of-plane action of shear walls The other equations as derived in Xu and Dolan 共2009兲 remain
and accurately represent the distribution of the mass, which will unchanged. These changes were necessary to allow the model to
give more accurate results for seismic analysis. One limitation of accurately account for the significantly larger displacement ranges

Fig. 2. 4 ⫻ 8 ft solid shear wall validation based on test results

978 / JOURNAL OF STRUCTURAL ENGINEERING © ASCE / AUGUST 2009

J. Struct. Eng. 2009.135:977-984.

Downloaded from ascelibrary.org by California State Polytechnic University, Pomona on 08/25/13. Copyright ASCE. For personal use only; all rights reserved.

Fig. 3. 4 ⫻ 8 ft solid shear wall validation based on detailed shear wall model analysis results

associated with walls when compared to individual nail connec- and shortening of the frame members are very small compared
tions. A complete list of variables with their associated units is with the racking displacement of the shear wall. Additional limi-
provided in Xu and Dolan 共2009兲, with the exception that ␺0 units tations of the model are that out-of-plane loading is not included
change to mm4/3 and ␦␺ units change to mm1/3 due to the change and regular beam elements or shell elements would have to be
in magnitude associated with walls. used to simulate this response, and effects of other conditions
It is easy to estimate the parameters for the hysteretic spring in such as moisture or decay are not included.
the single spring model 关Fig. 1共a兲兴 based on the experimental
hysteresis or detailed shear wall model analysis results because
the relative displacement and corresponding force in the single Model Validation
spring are the same as those of the global shear wall. The same
genetic algorithm program used for a single nail connector ele- Experimental and detailed model analysis results of the shear wall
ment was used for the parameter estimation purpose. The stiffness configurations mentioned in Xu and Dolan 共2009兲 were used here
and force at time, t, in the spring are K共t兲 and r共t兲, respectively. for the macroshear wall model validation. Figs. 2 and 3 show the
The parameters estimated for the single-spring shear wall comparison of the hysteretic loops and hysteretic energy between
model can be employed for each diagonal hysteretic spring in the the test results or detailed shear wall model analysis results and
diagonal-spring shear wall model 关Fig. 1共b兲兴 directly. However, the macroshear wall model results. The suitable parameter sets for
the stiffness and resistant force in each diagonal spring should be each macroshear wall spring are also presented in the relevant
modified from the single-spring model considering the geometry figures. The parameter estimations shown in Fig. 2 are in accor-
relations presented in Eqs. 共3兲 and 共4兲. The racking performance
and the force distribution in springs are shown in Fig. 1共b兲. In the
solver for each diagonal spring, the shear wall drift u共t兲 is used as
the input, and the resultant force and stiffness of the spring are
divided by 共2 ⫻ 兩cos共␪SW兲兩兲, which means the resistant stiffness
and the resistant force within each diagonal spring are
K共t兲
K⬘共t兲 = 共3兲
2 ⫻ 兩cos共␪SW兲兩

r共t兲
r⬘共t兲 共4兲
2 ⫻ 兩cos共␪SW兲兩
According to the geometry conditions, the stiffness and the
force of the global shear wall will be K共t兲 and r共t兲, respectively.
Two assumptions were made as the basis of the modifications:
共1兲 the racking displacement amount of the shear wall is far less
than the height and the length of the walls; 共2兲 the lengthening

Table 1. Quantification of Model Accuracy

4 ⫻ 8 ft wall
From From detailed
Case test data shear wall analysis
Error in peak force ⫺3.71% 1.72%
Error in peak hysteretic energy ⫺1.94% ⫺5.54% Fig. 4. Elevations of Phase 9 test structure showing major structural
Correlation coefficient 0.997 0.999 components 共Fischer et al. 2001兲

JOURNAL OF STRUCTURAL ENGINEERING © ASCE / AUGUST 2009 / 979

J. Struct. Eng. 2009.135:977-984.

 Table 2. Macroshear Wall Parameter Estimation Results
1st-story 1st-story 2nd-story 1st and 2nd-story
Parameters E wall W wall E and W walls N and S walls
␣ 0.06341 0.06363 0.10502 0.03941
␤ 0.02368 0.15735 0.07455 0.16259
␻ 1.61484 1.90654 1.6037 2.20759
z0 0.89626 0.92865 0.92814 0.94225
N 1.17407 1.01296 1.20403 1.04636
Y0 0.69376 0.85156 0.8543 0.71604
Downloaded from ascelibrary.org by California State Polytechnic University, Pomona on 08/25/13. Copyright ASCE. For personal use only; all rights reserved.

dy 0.18658 0.14918 0.13323 0.12251

dn 0.00001 0.00001 0.00002 0.00001
X 0.00001 0.00001 0.00002 0.00002
Fig. 5. Parameter estimation of single nail joint
G ⫺0.00420 ⫺0.1002 ⫺0.0414 ⫺0.0846
dh 0.00002 0.00003 0.00004 0.00003
P 0.11352 0.16168 0.18282 0.19105
dance with the test data and the parameter estimations shown in Q 0.12099 0.11185 0.14631 0.14520
Fig. 3 are in accordance with the detailed shear wall analysis
results.
The accuracy of the model was quantified through the com- objective. This test structure represents a simplified full-scale
parison of peak load, peak hysteretic energy, and the correlation two-story single-family house incorporating several characteris-
coefficient between the experimental and simulation results. The tics of recent California residential construction.
comparisons are presented in Table 1. The comparisons showed a This two-story single-family wood-frame house was tested
satisfactory performance of the super shear wall model. using the UC San Diego uniaxial earthquake simulation system.
The main objectives of the study were to determine the dynamic
characteristics and the seismic performance of the test structure
3D Wood-Frame Structure Simulation under various levels of seismic shaking and structural configura-
tions 共Fischer et al. 2001兲. The structure was tested during 10
To verify the accuracy and efficiency of the macroshear wall phases of construction to determine the performance of the struc-
model, a 3D light-frame structure was simulated, and the analysis ture with fully sheathed shear walls, symmetrical and unsym-
results were compared with the test results. The two-story single- metrical door and window openings, perforated shear wall
family house tested in Task 1.1.1 of the CUREE-Caltech Wood- construction, conventional construction, and with and without
frame Project 共Fischer et al. 2001兲 was chosen as the modeling nonstructural wall finish materials. The total weight of the struc-
ture was 109.33 kN with 61.52 kN at the floor diaphragm level
and 47.82 kN at the roof diaphragm level.
Phase 9 is the final configuration without finish material ap-
plied. The torsional irregularity is the most significant compared
to the other phases without finish material applied. So the con-
figuration in Phase 9 is ideal for the 3D model validation in this
study. The architectural and structural plan for Phase 9 configu-
ration of the test structure are shown in Fig. 4.

Shear Wall Simulation

The studs and the plates used in this test were Douglas Fir, “stud”
grade. From Table 4A in NDS 共2001兲, the design modulus of
elasticity is 1.4⫻ 106 psi 共9.65 kN/ mm2兲, and Poisson’s ratio
was assumed to be 0.3. Mechanical properties for OSB were
based on those for the structural panels with the span rating of

Table 3. Properties of Frames and Diaphragms

X-sectional
MOE area Thickness Poisson’s
Component Element 共kN/ mm2兲 共mm2兲 共mm兲 ratio
Roof
diaphragm 4-node shell 11.72 N/A 2540 0.3
Floor
diaphragm 4-node shell 11.72 N/A 4.7 0.3
Fig. 6. 2nd-story east and west walls configuration and detailed Frames Truss 11.72 1 ⫻ 106 N/A N/A
model Note: N/A⫽not available.

980 / JOURNAL OF STRUCTURAL ENGINEERING © ASCE / AUGUST 2009

J. Struct. Eng. 2009.135:977-984.

Downloaded from ascelibrary.org by California State Polytechnic University, Pomona on 08/25/13. Copyright ASCE. For personal use only; all rights reserved.

Fig. 7. Screen capture of model used in ABAQUS for two-story building simulation

24/16 共Table 3.2 2001 NDS Supplements兲. The MOE was as- displacement was restricted along the centerline of the 2nd-story
sumed to be 4.99 and 3.81 kN/ mm2 in strong axis and weak axis, east and west walls to account for the corresponding boundary
respectively. conditions 共Fig. 6兲. The load values obtained from the hysteretic
The nail joints used in this building were 8d box and 3/8-in. loops were doubled to account for the simplification. 共3兲 The
OSB sheathing panels. The investigators conducted monotonic shake table is uniaxial along the north–south direction, so the
and hysteretic testing on the joints with an average ultimate ca- analysis results will not be sensitive to changes in the south and
pacity of 0.996 kN 共224 lb兲 共Fischer et al. 2001兲. The test results north wall lateral modeling. Four pieces of solid shear wall were
of the connection constructed with 8 d box nail and 19/32-in. assumed to form each north and south wall. The load values in the
OSB sheathing panels in the CUREE project completed by Dolan 1/4 wall hysteretic results were quadrupled to form the final hys-
and Carradine 共2003兲 were used for the hysteretic connection re- teretic loops. 共4兲 Nailing along three edges of the 1st-story east
sponse. The tested connection hysteretic loops and the model pa- and west walls were staggered in two rows. Nails were assumed
rameters are shown in Fig. 5. The ultimate capacity of hysteretic to be in one row with 1/2 of the real spacing in the detailed
data was 1.38 kN 共310 lb兲. The ratio between the two connection model. 共5兲 Hold-down connectors were used as the tie-down an-
capacities is 0.72. Therefore, the initial stiffness and the ultimate chors. Since the hysteretic data of this connection is not available,
strength of the joint were reduced by 28%. The other parameters this connection was regarded as an elastic spring with a stiffness
remained the same since the basic hysteretic loop shapes of both equal to the ratio of the ultimate capacity to the reference dis-
connection configurations were similar. placement, which was 26.5 kN/mm. 共6兲 Bearing springs were
Some assumptions were made in the modeling process: 共1兲 modeled between sheathing panels to prevent overlapping but
since cracks at the corners between the wall segment and header have no resistance to separation.
are going to appear and develop rapidly with the loading process, The parameter estimation process was conducted on each
the stiffness and capacity contributions from the headers were shear wall based on the detailed shear wall simulation. The suit-
ignored. 共2兲 Because the configurations of the east and west walls able parameters for each super shear wall model were presented
were symmetric, only half of each wall was simulated. Vertical in Table 2.

JOURNAL OF STRUCTURAL ENGINEERING © ASCE / AUGUST 2009 / 981

J. Struct. Eng. 2009.135:977-984.

 3D Dynamic Model

Some assumptions were made to simplify the simulation: 共1兲 the

diaphragm acts elastically; 共2兲 the shear walls are connected to
the diaphragm through horizontal framing members, and the in-
terstory sliding is negligible; and 共3兲 the mass distributes uni-
formly in the floor and roof diaphragms. The total floor level
weight was assumed to distribute evenly to the nodes of the floor
diaphragm, as was the roof weight. While recent tests conducted
at SUNY Buffalo by Christovasilis et al. 共2007兲 showed that in-
Downloaded from ascelibrary.org by California State Polytechnic University, Pomona on 08/25/13. Copyright ASCE. For personal use only; all rights reserved.

terstory slip can provide a significant portion of the overall drift

for conventional construction, the tests used to validate this model
showed minimal interstory slip due to mechanical anchorages.
Therefore, the model was simplified by not adding nonlinear con-
nectors between the wall and the floor diaphragm to simulate the
slip potential.
According to the quasistatic test results reported by Fischer et
al. 共2001兲, the equivalent in-plane stiffness of the floor diaphragm
was 38 kN/mm, and the equivalent viscous damping ratio is 8.3%.
The floor diaphragm was modeled with 4 ⫻ 6 shell elements. The
in-plane stiffness of the floor diaphragm was calibrated using the
quasistatic test results given above. The constant elastic modulus
of OSB element, 11.72 kN/ mm2, was assigned to these shell
elements. The two supporting edges were pinned, and a unit de-
flection was applied at the midspan. The thickness of the floor
diaphragm elements 共4.7 mm兲 was adjusted such that the total
reaction at the two supporting edges in the numerical model
equaled to 38 kN when the displacement at the middle span was
1 mm. Since the supporting edges were pinned, the in-plane flex-
ural deflection of the floor diaphragm was eliminated. All the Fig. 8. Comparison of experimental and numerical mode shapes
flexural and shear deflections in the diaphragm were represented
as equivalent shear deflection.
The mass of the floor was 0.0063 kN/ 共mm/ s2兲. Based on the c = ␰ ⫻ 共2m ⫻ ␻1兲 共6兲
first vibration mode, the equivalent damping coefficient was c where m = total mass, ␻1 共different from the stiffness parameter
= ␰ ⫻ 共2m ⫻ ␻兲 = 0.081 kN⫻ s / mm. The related Rayleigh damp- in the modified BWBN model兲 is the circular frequency of the
ing factors ␣, ␤ were set to be 0.726 and 0.00202, respectively, fundamental vibration mode of the structure, which was
which make the equation 25.12 s−1 from the test results, and ␰ = damping ratio, which is
set to be 1% here. The total damping coefficient then was dis-
tributed to each wall evenly. The damping coefficient was
c=␣⫻m+␤⫻k 共5兲
0.00068 kN/共mm/s兲 for the east and the west walls. Since there
The roof diaphragm is very stiff compared to the floor dia- was no data that were able to supply the fundamental vibration
phragm because of the contribution of the trusses, and was mod- mode information in the east-west direction, and the properties of
eled by 4 ⫻ 6 shell elements with very high in-plane stiffness to the north and the south walls will not influence the results signifi-
simulate rigidity. cantly, the same damping coefficient was applied to the north and
Vertical and horizontal framing members were simulated with the south walls.
truss elements 共rotational DOFs at two ends are released兲 because The recorded acceleration time histories of the shake table for
all shear wall in-plane capacity is assumed from the two diagonal test Phase 9 were used as the input ground accelerations for the
hysteretic springs. The axial stiffness of the truss elements was numerical model. Since the structure was tested under 5 seismic
assumed very high to eliminate the stretch and compression in the levels 共6 if the repeat of Level 3 was counted兲, and there was no
framing members, which may cause unreal tension or compres- retrofit in the test series, the 5 共or 6兲 levels of earthquake histories
sion in the hysteretic springs. The diaphragm shared the four cor- were combined into a train of ground motions for the dynamic
ner nodes with the horizontal truss elements. This can eliminate time-history analysis. This allows the accumulated earthquake
the flexural deformation of the diaphragm, while the shear defor- damage to be considered.
mation of the diaphragm was not restrained at all. The properties The model for the simulation is presented in Fig. 7.
of the frames and diaphragms are shown in Table 3.
Each of the eight shear walls was represented with a pair of
super shear wall hysteretic spring elements. The parameters of Result Comparisons
each super shear wall hysteretic spring were imported to
ABAQUS 共2005兲 through an input file. The experimental fundamental period is 0.253 s, and the numeri-
A pair of dashpots with a damping ratio of 1% was placed in cal result is 0.287 s. Therefore, the numerical underestimates the
each shear wall to take the elastic damping effect into account. initial structural stiffness. The comparison of initial experimental
The total damping coefficient for all the walls in each direction is and numerical fundamental mode shapes is shown in Fig. 8. Since
derived from the equation the roof diaphragm was assumed to be rigid and flat, the ridge and

982 / JOURNAL OF STRUCTURAL ENGINEERING © ASCE / AUGUST 2009

J. Struct. Eng. 2009.135:977-984.

Downloaded from ascelibrary.org by California State Polytechnic University, Pomona on 08/25/13. Copyright ASCE. For personal use only; all rights reserved.

Fig. 10. Global hysteresis comparison for Level-5 earthquake input

Fig. 9. Relative displacement comparisons between model prediction
and shake table tests

developed. Satisfying performance was validated through a veri-

eave had the same displacements in the east and west elevations. fication procedure. A simple two-story 3D light-frame structure
From the comparison of the mode shapes, the stiffness of the was simulated with this shear wall model. According to the com-
1st-story west wall was underestimated. The underestimation is parisons, the simple 3D numerical model can predict the nonlin-
partly due to the secant stiffness of the tie-down anchors being ear wood-frame structural response accurately. The magnitudes of
taken as the elastic stiffness. errors are acceptable, considering the variations of material and
After the time history analysis, the relative displacement his- nailing details.
tories in Level 4 共design level兲 and Level 5 for the 1st-story east
wall and for the global response 共midspan of the roof兲 were com-
pared with the test results and presented in Figs. 9共a–d兲. The Table 4. Comparison of Ultimate Drifts and Reactions
global hysteretic responses for Level 5 shaking were also com-
Test Model Error
pared with the test results and are shown in Figs. 10共a and b兲. Items results results 共%兲
These results show good agreement between the numerical and
the experimental results when the structure was subjected to the Level 4 Maximum global response 共mm兲 69.73 71.29 2.2
high-amplitude ground motions. Minimum global response 共mm兲 ⫺47.61 ⫺44.24 ⫺7.1
The comparison of peak drifts and imposed loads between the Maximum base shear 共kN兲 97.96 100.52 2.6
numerical and the experimental results are shown in Table 4. Minimum base shear 共kN兲 ⫺76.4 ⫺79.92 4.6
While some of the error terms for displacement seem high 共17% Maximum 1st-story 40.75 43.05 5.6
maximum兲, in applied terms the error is only 7 mm 共0.27 in.兲 East wall drift 共mm兲
of displacement. Minimum 1st-story ⫺26.36 ⫺25.67 ⫺2.6
The simulation of the two-story structure for an 89 s record East wall drift 共mm兲
required 22 min. of computation time for a 64⫻ 2 dual core pro- Level 5 Maximum global response 共mm兲 68.57 78.5 14.5
cessor running at 2 GHz with 960 MB of ram under Windows XP Minimum global response 共mm兲 ⫺109.65 ⫺99.7 ⫺9.1
operating system. Maximum base shear 共kN兲 97.38 95.53 ⫺1.9
Minimum base shear 共kN兲 ⫺122.79 ⫺119.62 ⫺2.6
Maximum 1st-story 40.33 47.24 17.1
Conclusions East wall drift 共mm兲
Minimum 1st-story ⫺67.91 ⫺62.82 ⫺7.5
With some modifications being made to the nail joint model pre-
East wall drift 共mm兲
sented in Xu and Dolan 共2009兲, a super shear wall model was

JOURNAL OF STRUCTURAL ENGINEERING © ASCE / AUGUST 2009 / 983

J. Struct. Eng. 2009.135:977-984.

 References 130共9兲, 1361–1370.
He, M., Magnusson, H., Lam, F., and Prion, H. G. L. 共1999兲. “Cyclic
ABAQUS. 共2005兲. ABAQUS analysis user’s manual (Version 6.5), Hib- performance of perforated wood shear walls with oversize OSB pan-
bitt, Karlsson, and Sorenson, Pawtucket, R.I. els.” J. Struct. Eng., 125共1兲, 10–18.
Christovasilis, I. P., Filiatrault, A., and Wanitkorkul, A. 共2007兲. “Seismic Kasal, B., and Leichti, R. J. 共1992兲. “Nonlinear finite-element model for
test of a full-scale two-story light-frame wood building: NEESWood light-frame stud walls.” J. Struct. Eng., 118共11兲, 3122–3135.
benchmark test.” NEESWood Project Test Rep. No. NW01. Kasal, B., Leichti, R. J., and Itani, R. Y. 共1994兲. “Nonlinear finite-element
Collins, M., Kasal, B., Paevere, P., and Foliente, G. C. 共2005a兲. “Three- model of complete light-frame wood structures.” J. Struct. Eng.,
dimensional model of light-frame wood buildings. I: Model descrip- 120共1兲, 100–119.
tion.” J. Struct. Eng., 131共4兲, 676–683. Kasal, B., and Xu, H. 共1997兲. “A mathematical model of connections with
Downloaded from ascelibrary.org by California State Polytechnic University, Pomona on 08/25/13. Copyright ASCE. For personal use only; all rights reserved.

Collins, M., Kasal, B., Paevere, P., and Foliente, G. C. 共2005b兲. “Three- nonlinear nonconservative hysteretic behavior.” Earthquake perfor-
dimensional model of light-frame wood buildings. II: Experimental mance and safety of timber structures, G. C. Foliente, ed., Forest
investigation and validation of analytical model.” J. Struct. Eng., Products Society, Madison, Wis., 105–107.
131共4兲, 684–692. Kivell, B. T., Moss, P. J., and Carr, A. J. 共1981兲. “Hysteretic modeling of
Dolan, J. D. 共1989兲. “The dynamic response of timber shear walls.” Ph.D.
moment resisting nailed timber joints.” New Zealand Nat. Soc. Earth-
dissertation, Dept. of Civil Engineering. Univ. of British Columbia,
quake Eng. Bull, 14共4兲, 233–245.
Vancouver, BC, Canada.
National Design Specification for Engineered Wood Construction. 共2001兲.
Dolan, J. D., and Carradine, D. M. 共2003兲. “Nail fastener connections.”
American Forest and Paper Association, Washington, D.C.
Rep. prepared for CUREE Nail Fastener Connections, Washington
Paevere, P. J. 共2002兲. “Full-scale testing, modelling and analysis of light-
State Univ., Pullman, Wash.
Du, Y. 共2003兲. “The development and use of a novel finite element for the frame structures under lateral loading.” Ph.D. thesis, Dept. of Civil
evaluation of embedded fluid dampers within light-frame timber and Environmental Engineering, Univ. of Melbourne, Parkville, Vic-
structures with seismic loading.” Ph.D. dissertation, Dept. of Civil and toria, Australia.
Environmental Engineering. Washington State Univ., Pullman, Wash. Salenikovich, A. J. 共2000兲. “The racking performance of light-frame
Durham, J., Lam, F., and Prion, H. G. L. 共2001兲. “Seismic resistance of shear walls.” Ph.D. dissertation, Dept. of Wood Science and Forest
wood shear walls with large OSB panels.” J. Struct. Eng., 127共12兲, Products, Virginia Polytechnic Institute and State Univ., Blacksburg,
1460–1466. Va.
Fischer, D., Filiatrault, A., Folz, B., Uang, C.-M., and Seible, F. 共2001兲. Tarabia, A. M. 共1994兲. “Response of light-frame buildings due to earth-
“Shake table tests of a two-story house.” CUREE Publication No. quake loadings.” Ph.D. dissertation, Dept. of Civil and Environmental
W-06, Richmond, Calif. Engineering, Washington State Univ., Pullman, Wash.
Folz, B., and Filiatrault, A. 共2001兲. “Cyclic analysis of wood shear Tarabia, A. M., and Itani, R. Y. 共1997a兲. “Static and dynamic modeling of
walls.” J. Struct. Eng., 127共4兲, 433–441. light-frame wood buildings.” Comput. Struc., 63共2兲, 319–334.
Folz, B., and Filiatrault, A. 共2004a兲. “Seismic analysis of woodframe Tarabia, A. M., and Itani, R. Y. 共1997b兲. “Seismic response of light-frame
structures. I: Model formulation.” J. Struct. Eng., 130共9兲, 1353–1360. wood buildings.” J. Struct. Eng., 123共11兲, 1470–1477.
Folz, B., and Filiatrault, A. 共2004b兲. “Seismic analysis of woodframe Xu, J., and Dolan, J. D. 共2009兲. “Development of nailed wood joint
structures. II: Model implementation and verification.” J. Struct. Eng., element in ABAQUS.” J. Struct. Eng., 135共8兲, 968-976

984 / JOURNAL OF STRUCTURAL ENGINEERING © ASCE / AUGUST 2009

J. Struct. Eng. 2009.135:977-984.