Asian Center for Engineering Computations and Software, AIT

Asian Institute of Technology, Thailand

14th ASEP International Convention, Philippines, May 2009

Modeling of Shear Walls for Nonlinear and

Pushover Analysis of Tall Buildings
Naveed Anwar, D. Eng
Some of the Questions Related to Shear Walls

What is a Shear Wall

How does a Shear Wall behave

What is the normal role of Shear Wall

What role a Shear Wall can play

How to Model and Design the Shear Walls for the
intended role
What is a Shear Wall?

How can we “tell” when a member is a shear wall

Is the definition based on ?

Intended Use

Shape in Cross-section

Geometry in Elevation

Loading Type and Intensity

Behavior and Theory

Location, Direction, Orientation
Shear Wall or Column

Wall Column
Shear Wall or Frame

Shear Wall Shear Wall or Frame ? Frame

Shear Wall or Truss?
Conventional Role of Shear Walls

Provide lateral stiffness to buildings

Reduce Drift Ratio

Reduce wind-induced acceleration

Provide strength against lateral loads

Shift moment and shear away from frame members

Change the deformation and mode of the building

Interact with frames to convert shear and moment to axial
forces through outriggers etc.
Tall Shear Wall Design

Primarily governed by W=1
Flexural Strength, can
be allowed to yield at

H=100 m
W=1
well defined locations

Shear is generally not

H=50 m
the critical factor for tall
shear walls

All walls must and can 25 50 (2 times)
V
remain elastic in shear M 3,333 (4 times)
833
without failure
Seismic Code Development

Actual Elastic Demand Vs Code-mandated Design Forces

The difference is expected be handled by
yielding, ductility, energy dissipation and reduction of demand
Typically handled by “Response reduction Factor”
New Design Approaches

Current building codes do not adequately address many
critical aspects in seismic design of tall buildings

Performance based design provides a desirable
alternative

Reinforced concrete walls are effective to resist lateral
loads while providing good performance

Various approaches exist to predict the reliable nonlinear
and inelastic response of RC walls
Performance Based Design in
International Context

Explicitly stated by local authorities in some countries
such as Japan and China

UBC, IBC and other codes provide little-to-no specific
guidance

Eurocode 8 is not performance based

Much framework for performance based design is in
Vision 2000, ATC40 and FEMA 356

Recently, performance based design of high-rise
buildings issued in LATBSDC 2008 and SEAONC 2007
Basic Vertical Seismic Systems

Moment Resisting Frame

Braced Frame Shear Walls

Typical Multi-Story Structural Systems
Nonlinear Performance Comparisons

Six alternative structural systems compared by pushover plots for specific four-
story building

BF – Braced Frame

SW – Shear Wall

EBF – Eccentric Braced Frame

MF – Steel Moment Frame

MF+Dampers – Steel Moment
Frame with Passive Dampers

BI – Base Isolation
Nonlinear Performance Comparisons
Role of Shear Walls, Outriggers, Dampers.
Ductile Core Wall Structural System

Offer lower costs, faster
construction and flexible
architecture

Seismic forces are
resisted by reinforced
core surrounded by
elevator banks

For buildings 100 m or
taller, core has a minimum
dimension of 10 m in plan Concrete Core Wall Building under Construction, the
Washington Mutual/Seattle Art Museum

and 50cm 90cm thick

Ductile Core Wall System Projects
Ductile Core Wall Structural System

3D View Lateral Force Resisting Plan View

System
One Rincon Hill in San Francisco, California (57-story, 625 feet)
Modeling and Analysis Goals

For static push-over analysis, overall strength should
be calculated correctly and the stiffness along the
curve should be essentially accurate.

For a dynamic analysis, the cyclic behavior and
energy dissipation should be essentially correct

Meaningful deformation demand-capacity values
and usage ratios should be calculated for assessing
performance

The demand-capacity values and deflected shape
should show any concentrations of damage

The Goal is to get results that can be useful for
design, not to get an exact simulation of the behavior
Distinct Parts in a Wall
Vertical cantilever Well-defined vertical and horizontal Staggered openings
Type segments

Yielding of the vertical steel Yielding of longitudinal steel Yielding of vertical ties
Desirable

Shear yield or vertical crushing of Shear yield, severe diagonal cracking or Yielding of horizontal ties and
Undesirable concrete concrete crushing crushing of struts

Strut and Tie Action in Right Part

Openings in Shear Walls

Very Small Openings Medium Openings Very Large Openings

may not alter wall may convert shear may convert the Wall
behavior wall to Pier and to Frame
Spandrel System

Spandrel Beam

Wall Column
Pier Pier
Main Aspects of behavior for Planner walls

In-Plane Behavior : Key Aspects

Unsymmetrical Bending Behavior

As a cantilever bends and concrete cracks, the neutral
axis shifts towards the compression side.
Connecting a Beam to a Shear Wall

If a beam element is connected to a shear wall, a beam
element must be imbedded in the wall
Coupling Beam Behavior - Bending

Elastic Behavior

Curvature varies linearly along length

There may be significant local
deformation in the pier

Actual Behavior

Plastic
zone may form near end

Crack may open because of bond slip
Coupling Beam Behavior - Shear

Elastic Behavior

Compression diagonal shortens

Tension diagonal extends

Beam as a whole does not extend

Actual Behavior with Conventional Reinforcement

Vertical steel yields

Horizontal steel does not yield

Beam as a whole does not extend

Actual Behavior with Diagonal Reinforcement

Tension diagonal yields

Compression diagonal has a much smaller deformation

Beam as a whole must increase in length
Handling Nonlinearity in Shear Walls

Hinging is expected in shear
walls near the base

Difficult to convert a large
shear wall core into an
equivalent column and beam
system

The question remains on how
to effectively models

Another major question is the
length of the hinge zone
Hinge Length for a Wall

Paulay and Priestly (“Seismic Design of Reinforced Concrete and
Masonry Buildings”, Wiley, 1992)
Lp = 0.2 Dw + 0.044 he
Lp = hinge length
Dw = depth of wall cross section
he = effective wall height (height of cantilever wall with a single load at
the top and the same moment and shear at the hinge as in the actual
wall
A larger shear (i.e., a larger bending moment gradient) gives a smaller
hinge length

FEMA 356 recommends a hinge length equal to smaller of (a) one

half the cross section depth (b) the story height.
Nonlinear Modeling of Shear Walls

For Elastic Model

Shell or Membrane model is common

Normal shell model can not handle Nonlinearity or
hinging

A Study carried out to compare various methods in
an attempt to answer the questions
1. Single Column model
2. Fiber or Frame model
3. Strut and Tie model
4. Nonlinear Layered Shell model
The Main Comparative Parameter

The Moment
Curvature of the Wall
Section is used as
the reference for
comparison of the
wall model response

This is reasonable,
as the wall is tall
enough to deform in
flexure
Single Column Model

Simplest model

Equivalent column at the center line of wall section

Rigid links are required to make deformation compatibility

Non-linear axial-flexural hinges at the top and bottom

Optional shear hinges at the mid height

Requires predefined hinge length

Suitable for walls of small proportions

Difficult to handle cellular core walls or walls with openings

Disregards the wall rocking and effect of neutral axis shift

Used as reference model and quick assessment of
performance
Column Model for Planer Walls

• Specially Suitable when H/B

is more than 5
HH
• The shear wall is
represented by a column of
section “B x t”
tt • The beam up to the edge of
BB the wall is modeled as
normal beam
• The “column” is connected to
beam by rigid zones or very
Rigid Zones
large cross-section
Column Models for Cellular Walls

Difficult to extend the concept to
tt Non-planer walls
HH
Core Wall must be converted to
“equivalent” column and
BB appropriate “rigid” elements

Can be used in 2D analysis but
more complicated for 3D analysis

After the core wall is converted to
2t
2t planer wall, the simplified
HH procedure can used for modeling
tt
BB
Single Column Model

Disregards the neutral axis
shift on vertical
displacements

Disregards the rocking of
wall

Computes Response
assuming plane -Section
remain plane

Not suitable for short/squat
walls

Can not capture geometric
changes, openings,

Single Column Experimentally

Model Behavior Observed Behavior
Single Column Model

Moment Hinge, directly using the Moment Curvature

of the Wall Section, multiplied by Hinge Length

Shear Wall Frame Element

Axial Load-Deformation Hinge Property
Fiber or Frame Model

Wall section is discretized by closely spaced
columns

Nonlinear axial load-deformation hinges are used

Different ductility shall be used for unconfined and
confined portion of the wall

Eliminate the predefined hinge length which is
needed in single element models
 Fiber or Frame Model

Shear Wall Section

Discretized into Frame Elements. Each

column acts as a “Fiber” representing
part of the wall
Fiber or Frame Model

Shear link Diaphragm constraint and
Beam constraint

element is
used to Axial hinge

provide shear
Diaphragm constraint and
stiffness Beam constraint

Shear link

Diaphragm constraint and

Beam constraint

Release moment in both ends of fiber

element

Fiber or Frame Model

Axial Hinges
Hinges for Nonlinear Modeling

Upper Portion is assumed or designed to be Elastic

Axial Hinges for Column Fibers

Moment Hinges for the Spandrel
Moment Hinges
Strut and Tie Model

Extensively used for deep beams and shear walls

Nonlinear axial load-deformation hinges are used

Difficult to determine the size and reinforcement in
diagonal elements

Hinges in diagonal struts should be force control to
detect shear failure or may or the diagonals may be
forced to remain elastic
Strut and Tie Model

txt

C
t x 2t

t
B
Strut and Tie Model

Opening

Displacement controlled Axial Hinges

Opening
Force Controlled Axial Hinges
matching shear capacity

Wall with Opening

Full Wall
Nonlinear Layered Shell

This element is not available in many software yet

Nonlinear stress-strain relationship is sampled at
Gauss points

Integration is performed by standard 2x2 Gauss
points

Equivalent to having two fibers in each local 1 & 2
directions

Stresses at locations other than Gauss points are
interpolated or extrapolated
 Nonlinear Layered Shell

Layered Shell Nonlinear stress-strain curve of concrete

Nonlinear stress-strain curve of steel

Practical Shear Wall Model

Membrane behavior of vertical stress in concrete
S22 and rebar stress S11 is taken to be nonlinear

Horizontal rebar is neglected

Out of plane behavior is assumed liner, single
concrete plate layer is used
 Shear Wall Model using Shell Elements

Nonlinear behavior
in vertical rebar

Nonlinear behavior in S22

component of concrete

N = Nonlinear, L = Linear

S22

S12 S11
Comparative Study

Two walls are selected to compare the non linear
pushover curves generated by various modeling
technique

Pushover analysis is performed by displacement
control (top displacement of 5% drift)

Inverted triangular loading is used

Axial hinges are assigned in the mid length of the
member for fiber or frame model and strut and tie
models

For the cracked section models, 50% bending
stiffness and 40% shear stiffness of gross section
are used
Comparative Study

20 Stories @ 3.2 m = 64 m

Wall -01: Planner Wall

Comparative Study

20 Stories @ 3.2 m = 64 m

Wall -02: Core Wall with Opening

Time Period Comparison
Planner Wall
Single Column Full Shell Full Shell
Fiber/ Frame Strut and Tie
Mode (Cracked) (Gross) (Cracked)
sec sec
sec sec sec

1 2.24 1.58 1.59 1.49 1.42

2 0.37 0.26 0.27 0.25 0.25
3 0.14 0.10 0.10 0.09 0.12
4 0.08 0.08 0.08 0.08 0.10

Core Wall
Single Column Full Shell Full Shell
Fiber or Frame Strut and Tie
Mode (Cracked) (Gross) (Cracked)
sec sec
sec sec sec
1 2.85 1.85 1.87 1.83 2.16
2 2.06 1.43 1.45 1.41 1.61
3 0.47 0.31 0.32 0.30 0.38
4 0.35 0.25 0.27 0.24 0.31
Time Period Comparison

The Elastic stiffness should be represented
realistically. This can be checked through time
period comparison.

It is difficult to estimate the level of cracking or the
size of members for Fiber or Strut-Tie models.

Shell Models tend to stiffer than others due to shear
strain contribution and higher in-plane stiffness

Loss of mass in Fiber and Strut and Tie model and
overlapping mass in Column model should be
considered

Time is effected by nonlinear response due to
reduction in stiffness
Moment-Curvature Relationship

Planner Wall
Moment-Curvature (Planner Wall)
4000

3500

3000
Moment (Ton-m)

2500

Single Column
2000
Strut and Tie
Fiber or Frame
1500 Nonlinear Shell

1000

500

0
0 0.005 0.01 0.015 0.02 0.025 0.03 0.035 0.04

Curvature
Moment-Curvature Relationship

Core Wall
Moment-Curvature (Core Wall)
20000

18000

16000

14000
Moment (Ton-m)

12000

10000 Single Column

Fiber or Frame
8000 Strut and Tie

6000

4000

2000

0
0 0.005 0.01 0.015 0.02 0.025 0.03 0.035 0.04 0.045

Curvature
Base Shear Vs. Top Displacement (Ton, m)

Fiber or Frame Model Strut and Tie Model

Planner Wall

Core Wall
Hinge Formation

Fiber/Frame Strut and Tie Fiber/Frame Strut and Tie

Limitations of Pushover Analysis

Static pushover analysis is typically unidirectional,
single pattern load analysis, in which most hinges
will deform monotonically

Higher mode contributions are not considered

Material and section hysterics can not be considered
directly

The hinge properties typically will be based on the
envelop curve from the expected hysteresis curves

The material or section degradation due to cyclic
response is not explicitly considered

The Dynamic effects are not considered
Nonlinear Time History Analysis

For a detailed nonlinear time history analysis, the
effective of material as well as section level hysterics
and degradation for cyclic response needs to be
considered

Although the basic modeling approaches presented
for the static pushover analysis are also suitable for
the NLTH, the hinge properties as well as modeling
should represent the hysteric behavior

The NLTH takes considerably more effort and
understanding, specially for selection and scaling of
Time
Conclusions

The objective of this study was to investigate the various
approaches of nonlinear modeling of shear walls to predict
their nonlinear response by Pushover Analysis

Refined fiber or frame model has the capability to represent
the nonlinear flexural behavior more reliable than strut and tie
model

The fiber model can be used to estimate the extent of yielding
in the shear walls and can be used to determine the hinge
length more realistically than based on single or double story
concept

Both models lack the proper representation nonlinear shear
behavior and shear flexural interaction behavior