FEMA P-2139-3 Masonry 508
FEMA P-2139-3 Masonry 508
FEMA P-2139-3 Masonry 508
Prepared by
APPLIED TECHNOLOGY COUNCIL
201 Redwood Shores Parkway, Suite 240
Redwood City, California 94065
www.ATCouncil.org
Prepared for
FEDERAL EMERGENCY MANAGEMENT AGENCY
Mai (Mike) Tong, Project Officer
Robert D. Hanson, Technical Advisor
Washington, D.C.
Cover photograph – Two-story special reinforced masonry shear wall building (credit: T. Escobar, Masonry Institute
of America).
Foreword
Most buildings in the United States are less than five stories tall. These low-rise
buildings typically possess fundamental periods less than one-half second and
thus are referred to as short-period buildings. Many commonly used analytical
models have predicted that short-period buildings designed to current building
codes are likely to suffer severe damage or collapse during design-level
earthquakes. However, post-earthquake field investigations have not confirmed
these predictions. Since this uncertainty is found across all types of building
structures and construction materials permitted by current building codes and
standards, it decreases confidence in the earthquake resilience of such code-
compliant buildings. This technical resource series provides the findings and
conclusions related to this issue and recommendations for improving seismic
design of short-period buildings.
FEMA is grateful to the Applied Technology Council (ATC) for managing this
sophisticated multi-year project series to a successful completion, to the Project
Technical Committee and the Project Review Panel for their dedicated effort
leading to invaluable technical findings and recommendations. FEMA is also
thankful to the project workshop participants for their scrutiny and valuable
comments. Resolving the uncertainty in short-period building seismic collapse
performance will strengthen confidence in seismic building codes. This project
series will also contribute to improving seismic design and predicting collapse
potential of short-period buildings in high-seismic communities in the nation.
In 2013, the Applied Technology Council (ATC) was awarded the first in a
series of task orders under contracts HSFE60-12-D-0242 and HSFE60-17-D-
0002 with the Federal Emergency Management Agency (FEMA) to
investigate “Solutions to the Issue of Short Period Building Performance,”
designated the ATC-116 Project series. The purpose of this series of projects
was to investigate the response behavior and collapse performance of
different structural systems and to identify causes and develop solutions for
the short-period building seismic performance paradox. Studies investigated
three structural systems: wood light-frame, special reinforced masonry shear
wall, and steel special concentrically braced frame (SCBF) systems.
Jeff Corson and Linda Peters assisted in the development of the design of the
reinforced masonry archetypes, and Jianyu Cheng and Andreas Koutras
assisted in the reinforced masonry numerical modeling. Lisa Star provided
technical guidance on the development of the reinforced masonry soil-
structure interaction and foundation flexibility parametric study. The Project
Review Panel, consisting of Tony Court, Bill Holmes, Onder Kustu, Jim
Malley, Steve Pryor, and Jason Thompson, provided technical review and
advice at key stages of the work.
ATC also gratefully acknowledges Mike Tong (FEMA Project Officer) and
Bob Hanson (FEMA Technical Advisor) for their input and guidance in the
preparation of this report, Scott Schiff who assisted in ATC project
management, and Carrie J. Perna who provided ATC report production
services. The names and affiliations of all who contributed to this report,
including those who participated in the review workshop focused on
reinforced masonry systems, are provided in the list of Project Participants at
the end of this report.
Foreword....................................................................................................... iii
Preface.............................................................................................................v
List of Figures............................................................................................... xi
Figure 3-9 Isometric view of the BOX archetype (BOX1) ............... 3-19
Figure 4-3 Frame models for masonry wall systems ........................... 4-5
Figure 4-9 Steel material model for reinforcing bars ......................... 4-11
Figure 4-11 Monotonic and cyclic behaviors of bond-slip model ....... 4-13
Figure 4-26 Mode shapes for the first three modes of the refined
finite-element model of COM2 ........................................ 4-30
Figure 4-28 Frame model with OpenSees for COM2 and COM5 ....... 4-32
Figure 4-29 Modeling of a first-story wall for COM2 and COM5 ...... 4-33
Figure 4-33 Mode shapes for the simplified model of COM2B .......... 4-37
Figure 4-34 Foundation model for the soft site for SSI analyses ........ 4-40
Figure 4-35 Soil springs and dampers for SSI analyses ...................... 4-40
Figure 4-36 Foundation model for the stiff site for SSI analyses ........ 4-41
Figure 4-39 Collapse rates from IDA and derived collapse fragility
curve for COM2B. ST for the collapse fraction data has
been scaled only by the factor of 1.2 for 3D analyses ..... 4-44
Figure 5-3 Zoomed-in pushover curves for all archetypes .................. 5-3
Figure 5-5 Collapse rates from IDA data and collapse fragility
curves derived for very high-seismic baseline models.
ST for the collapse fraction data has been scaled by 1.2,
whereas that for the collapse probability curve has been
scaled by 1.2×SSF .............................................................. 5-6
Figure 5-9 Collapse rates from IDA data and collapse fragility
curves derived for COM2B, COM2B-DC1, and
COM2B-DC2. ST for the collapse fraction data has
been scaled by 1.2, whereas that for the collapse
probability curve has been scaled by 1.2×SSF ................. 5-10
Figure 5-12 Pushover curves for COM2B with and without soil
springs and dampers ......................................................... 5-16
Figure 5-14 Forces in vertical soil springs for the stiff and soft sites
at the respective maximum base shear ............................. 5-17
Figure 5-15 Moments in foundation slabs and grade beams for the
stiff and soft sites at the respective maximum base
shear ................................................................................. 5-18
Figure A-43 Four-story residential building East elevation: RES2 ..... A-28
Figure C-2 Translational and rotational springs and dampers ............. C-6
Table 4-2 Natural Periods for the First Three Modes of the
Building Archetypes from Refined Finite-Element
Models ............................................................................. 4-29
Table 4-3 Natural Periods for the First Three Modes of the
Baseline Building Archetypes from Simplified
Models ............................................................................. 4-37
Table 5-8 Collapse Rates and Mean Peak First-Story Drift Ratios
of Survivors at 50 Percent-of-MCER and MCER
Ground-Motion Intensities of the Displacement
Capacity Parametric Study Archetype Models ................ 5-15
Table 5-9 Modal and Pushover Analysis Results of the SSI and
Foundation Flexibility Parametric Study Archetype
Models .............................................................................. 5-19
Table 5-12 Collapse Rates and Mean Peak First-Story Drift Ratios
of Survivors at 50 Percent-of-MCER and MCER
Ground-Motion Intensities of the SSI and Foundation
Flexibility Parametric Study Archetype Models .............. 5-22
Table C-3 Spring and Damper Properties Per Node for COM2B
Soft Site..............................................................................C-8
Table C-4 Spring and Damper Properties Per Node for COM2B
Stiff Site .............................................................................C-9
Studies conducted before the start of the ATC-116 Project series have used
the methodology described in FEMA P-695, Quantification of Building
Seismic Performance Factors (FEMA, 2009b), to evaluate the collapse
performance of common code-permitted seismic-force-resisting systems.
For example, one widely cited previous collapse performance study is
described in NIST GCR 12-917-20, Tentative Framework for Development
of Advanced Seismic Design Criteria for New Buildings (NIST, 2012a). This
and other similar studies have shown that many seismic-force-resisting
systems achieve the collapse performance target (i.e., less than a 10 percent
probability of collapse given MCER ground motions). However, these
studies have also found that shorter-period buildings have calculated
probabilities of collapse that exceed those of longer-period buildings, and
generally exceed the 10 percent target for acceptable collapse performance.
Overall, the approach was to: (1) establish benchmarks for the historically
observed performance of short-period buildings; (2) conduct parametric
analytical studies on archetypical short-period buildings using advanced
numerical models and the latest available research and test data; and
(3) identify modeling parameters or building characteristics that provide the
best match between the simulated and benchmark performance.
Archetype designs provided the basis for advanced numerical models. First,
for each building archetype, a refined finite-element model was developed and
subjected to pushover and time-history analyses. Second, the results obtained
were then used to calibrate computationally efficient frame models for
incremental dynamic analysis (IDA). Following FEMA P-695 procedures, the
IDA results provided collapse performance metrics in terms of the conditional
probability of collapse given the MCER ground-motion level.
Due to budget and time constraints, not all reinforced masonry archetype
buildings were designed, modeled, and analyzed. Table 1-2 lists the names
of the archetype models for which parametric studies were completed. A
total of 20 computationally efficient frame models were developed for three
parametric studies (each archetype analyzed for the displacement capacity
parametric study and the SSI and foundation flexibility parametric study had
two variations). Although they were not all analyzed, the designs for all the
baseline configuration archetype buildings were completed, and details for
these designs are provided in the report. The COM occupancy was
prioritized over the RES occupancy because it was anticipated that the latter
typically have so much masonry wall that archetypes would need to have
Table 1-2 Reinforced Masonry Building Archetypes Developed for the Parametric Studies
(3)
Soil-Structure
(1) (2) Interaction and (4) (5)
No. of Baseline Displ. Foundation Slab-Induced Coupling of
Archetype ID Stories Configuration Capacity Flexibility Wall Coupling Flanged Walls
Commercial Buildings: High Seismic
COM1 1 COM1B COM1B-DC - - -
COM2 2 COM2B COM2B-DC COM2B2-SS - -
COM3 4 COM3B COM3B-DC - - -
Commercial Buildings: Very High Seismic
COM4 1 COM4B COM4B-DC - - -
COM5 2 COM5B COM5B-DC - - -
COM6 4 COM6B COMB6B-DC - - -
Chapter 5 summarizes the analytical results for each of the parametric studies
on short-period reinforced masonry buildings.
References and a list of project participants are provided at the end of this
report.
2.1 Introduction
This chapter summarizes the observed damage and collapse of modern short-
period reinforced masonry buildings in past earthquakes, presents key results
from shake-table tests, and describes the performance benchmarks used in
this study for comparison with baseline archetype model results.
The 140 reinforced masonry buildings surveyed after the 1994 Northridge
earthquake represent only a fraction of all reinforced masonry buildings in
the area of strongest ground motions, although the total number of reinforced
masonry buildings is not known. For reference, post-earthquake safety
evaluations of 114,039 potentially damaged buildings included 10,393
commercial and industrial buildings, of which 637 (6.1 percent) were
deemed unsafe and assigned a red tag (Table 4-2, OES, 1995). Those safety
inspections included 3,068 “Class C” structures having exterior walls of
brick, concrete block, or poured-in-place concrete (e.g., tilt-up buildings), of
which 277 (9 percent) were assigned a red tag (Table 4-3, OES, 1995). Here,
“Class C” structures included unreinforced masonry buildings that were
much more vulnerable to earthquake damage than reinforced masonry
buildings. Only a fraction of all buildings was safety inspected, and red tag
percentages based on the number of inspected buildings, rather than the total
number of buildings in the affected area, represent an upper bound on
observed damage.
More meaningful estimates of red tag percentages are shown in Figure 2-1
for wood buildings and for other (non-wood) buildings for each of five MMI
regions (V–IX). Here, the red tag percentage is the ratio of the number of
buildings assigned a red tag (Table 4-3, OES, 1995) to the number of
2.0%
Red Tag - Wood (1,614 Buildings)
1.8%
Red Tag - Non-Wood (313 Buildings)
Percentage of Red Tags by MMI Group
1.6%
1.4%
1.2%
25
1.0%
0.8% 230
0.6% 165
0.4%
109
1,023
0.2%
325
0.0%
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0
Figure 2-1 Percentage of wood and non-wood buildings assigned a red tag
as a function of 0.3-second response spectral acceleration for
five MMI regions (V–IX) based on post-earthquake safety
inspections following the 1994 Northridge earthquake.
As shown in Figure 2-1, red tag percentages are quite low, less than 1
percent, even for the MMI IX region (e.g., average 0.3-second response
spectral acceleration of 1.35g). The number of buildings assigned a red tag is
shown in the figure for MMI regions VII, VIII, and IX. For example, only
25 of the 313 non-wood buildings assigned a red tag are in the MMI IX
As shown in Figure 2-1, non-wood building red tag damage data are sparse
but suggest that reinforced masonry buildings have collapse rates that are no
larger than that of wood buildings, for which data are more reliable for
generating collapse statistics. Hence, for the purposes of this study,
benchmark target collapse rates for modeled performance of reinforced
masonry buildings are assumed to be the following:
• One-Story Buildings. 1 percent (0 percent to 2 percent) probability of
collapse given MCER ground motions of SMS = 1.5g; and
• Two-or-More-Story Buildings. 2.5 percent (0 percent to 5 percent)
probability of collapse given MCER ground motions of SMS = 1.5g.
3.1 Introduction
This chapter describes the rationale for selecting the occupancy, height, and
seismic design criteria for the reinforced masonry archetypes used for this
study. It presents reinforced masonry building characteristics used in other
relevant studies and highlights key aspects of this study’s archetype
configurations and detailed designs.
The information presented in this chapter provides the basis for the numerical
models described in Chapter 4 and used to predict the seismic collapse
performance of the archetypes. Additional design details are provided in
Appendix A.
This section presents factors that were expected to influence the seismic
response and performance of reinforced masonry buildings and thus were
considered in the design of the archetypes.
Masonry design and construction practice varies widely across the United
States and evolved significantly during the twentieth century. The practice
of reinforcing masonry began with early innovators in the 1920s, ultimately
leading to the introduction of reinforced masonry to the building codes in the
1950s. Ongoing research and code development, particularly since the mid-
1970s, has contributed to the steady evolution of design practice. The use of
reinforcement in masonry, which started in the West and progressed
eastward, is now widespread in the country. When modern reinforced
masonry was first developed in the United States, multi-wythe systems—in
which two wythes of solid-unit clay-brick masonry were connected by a
reinforced and grouted space between them—were common. Over time, this
system has been almost entirely replaced in common practice by hollow-unit
masonry with reinforcement placed in the grout-filled cores of the units.
Clay brick masonry is often used as a nonstructural veneer rather than as a
structural material.
Because of this variability, the same building design problem may result in
very different solutions depending on the designer’s location and choice
among several code-sanctioned design methods. This study focused on
seismic-force-resisting systems common in high-seismic areas in the western
United States, specifically: loadbearing, fully grouted, reinforced hollow-unit
concrete masonry with cantilevered shear walls. Masonry walls are generally
identified either as perforated, with regular or irregular openings, or as
cantilevered. All the walls in this study were the cantilever type. Nominal
8-inch thick walls were used in all archetypes wherever possible, although
12-inch walls were required in one case. The chosen design process for the
archetypes was intended to represent typical practice exercised in areas of
significant seismicity using the normal standard of care for the referenced
codes and, inherently, the associated time of practice. The strength design
provisions of the 2013 version of TMS 402/602 were used. Regardless of
Seismic Design Category, all walls were designed as “special reinforced
masonry shear walls” per ASCE/SEI 7-10 with a response modification
coefficient (R) equal to 5.
The foundation system and its connectivity and interaction with the ground
have long been identified as a possible source for explaining performance
under seismic loads that is different than would be predicted by numerical
models that ignore any such flexibility or energy dissipation. Accordingly,
this study included a parametric study regarding the effects of soil-structure
interaction (SSI) and foundation flexibility, comparing building response on
sites with stiff soil and soft soil. Unique designs were prepared for each site
using different allowable soil bearing pressures and moduli of subgrade
reaction. Both sites were characterized as Site Class D.
Details of the soil properties and designs are documented in Section 3.4.4
and Appendix A.
At the most basic level, the behavior of reinforced masonry shear walls under
seismic loads can be characterized in one of two fundamental categories
(NIST, 2014):
• Flexure-dominated walls are walls whose behavior is dominated by
flexure, with reliable ductility and inelastic displacement capacity.
• Shear-dominated walls are walls whose behavior, often for reasons
beyond the control of the structural designer, is dominated by shear, with
limited ductility capacity.
The following limitations were imposed to bound the variables in the NIST
study:
• all walls were cantilevered (not perforated) with rectangular cross
sections,
• flanged-wall behavior was not considered, and
• coupling between walls was not considered.
The archetypes from the NIST study were an important reference, but there
are several differences in the chosen archetypes for this study. For example,
In the NIST study, the retail occupancy building (Figure 3-1) was 204 feet
square. However, only one 24-foot segment of exterior wall centered on
each elevation was considered as a participating structural element and was
optimized for strength relative to the seismic design load. Structural walls
had aspect ratios ranging between 0.5 and 1. Because it was a goal of this
study to consider elements of short-period building construction that may
participate in the seismic-force-resisting system, even if they are not
explicitly necessary, all exterior masonry walls were assumed to be
participating elements as defined by TMS 402/602.
The residential archetype in the NIST study was nearly duplicated in this
study, with an identical plan layout but the exclusion of exterior longitudinal
walls, which were considered unrealistic elements for a residential
occupancy. However, in this study, the design considered flanged walls in
accordance with the recommendations in TMS 402/602, rather than
considering each wall as an independent rectangular cantilever wall in the
design. Coupling between shear walls due to beam or slab bending was not
explicitly included in the design for either study.
The NIST study did not include a structure like the commercial archetype
chosen for this study, with flanged walls on the building perimeter.
Finally, while the NIST archetypes included two-, four-, eight-, and twelve-
story versions, this study was limited to one-, two-, and four-story versions
because of the focus on short-period structures.
The use of reinforced masonry as the primary structural system for single-
family homes is rare, and the occupancy is not included in this study. In
contrast, reinforced masonry is common in multi-family residential
construction, because, in addition to its structural role, the masonry walls can
provide fire and acoustic separation between units and egress corridors
(Figure 3-5). Low- to mid-rise hotel structures using reinforced masonry
party and corridor walls with precast hollow-core concrete diaphragms are
common. Although designs were prepared for this archetype, numerical
analyses were not ultimately conducted as part of this study.
Among the most common reinforced masonry building types in the United
States is the single-story retail or warehouse structure, often referred to as
“big box” retail (Figure 3-6). These structures are characterized by large
plan dimensions with reinforced masonry walls defining the entire exterior
envelope and seismic-force-resisting system. The interior structure is often
composed of steel columns with open-webbed steel joists and joist girders,
with a corrugated metal deck or panelized wood diaphragm. The diaphragm
is flexible relative to the masonry walls. Although designs were prepared for
this archetype, numerical analyses were not ultimately conducted as part of
this study.
This section documents the criteria for designing the archetype reinforced
masonry buildings and the basic configurations that were used for this study.
Detailed documentation of the designs for each archetype are provided in
Appendix A.
Archetype design dead and live loads are summarized in Table 3-2. The
weight of all walls, except those of the exterior, was averaged across the
floor plate, and a portion was assigned to the roof for the determination of
seismic loads. The interior wall dead load value used was 15 pounds per
square foot (psf) distributed uniformly over each story.
Soil properties were assumed to represent typical West Coast sites, which are
described in detail in Appendix C. Assumed site-specific properties of soils
are summarized in Table 3-5. For the SSI and foundation flexibility study
only, a soft soil profile was assumed, with a lower bearing pressure. Both
stiff and soft soil sites were characterized as Site Class D.
Commercial buildings with heights of one story, two stories, and four stories
are common, so the selected reinforced masonry commercial building
archetypes included all three.
The retail archetype, representing a typical “big box” store, was also self-
limiting in that reinforced masonry structures of that type rarely exceed one
story. Therefore, only one height—one story—was designed for the
reinforced masonry retail archetypes.
In this archetype, as with the others, the reinforced masonry walls served as
gravity-load-carrying elements, as well as elements of the seismic-force-
resisting system. Floor beams and girders transferred gravity loads through
bearing plates on masonry pilasters built into the walls. For the purposes of
design, it was assumed, as is common practice, that the floor system was
effectively pin-connected to the walls, meaning that it provided kinematic
compatibility between the shear walls without inducing flexural-coupling
action.
The archetypes also included flanged walls, which are common in practice.
TMS 402/602 provides guidance to account for the behavior of flanged
walls, and this guidance was followed explicitly in the design of the
archetypes. In practice, anecdotal experience suggests that many designers
prefer to ignore the intersecting-wall effects and design the in-plane walls in
each direction independently. Designs that incorporate flanged-wall
behavior can be challenging, considering the need to adhere to maximum
reinforcement limits and competing demands for a single flanged wall
element that must meet code requirements for design forces in two
orthogonal directions. In some cases, provisions of the TMS 402 code that
allow maximum reinforcement requirements to be bypassed under certain
conditions were invoked (NIST, 2014). These conditions are specifically
identified for the commercial archetypes in Appendix A.
The SSI and foundation flexibility study investigated SSI inertial and
kinematic effects and foundation flexibility on response behavior and
collapse performance. The archetypes for the SSI and foundation flexibility
parametric study were based on two designs, one for sites with relatively stiff
soils (allowable bearing pressure = 3,000 psf) and one for sites with soft soils
(allowable bearing pressure = 1,500 psf). The ELF seismic-design forces
were not altered, only the allowable soil bearing pressure, so the resulting
design changed only the footing dimensions and reinforcing.
The plan dimensions were chosen to be double the plan dimensions of the
commercial archetypes, or 96 feet × 182 feet. The building was single story
with a height of 24 feet to the roof plane. In accordance with common
practice for structures of this type in the western United States, the roof was
designed as a panelized wood system with prefabricated panels of plywood
on wood joists supported on bar joists and joist girders supported on columns
and walls (Lawson, 2013). The roof was a low-slope gable form.
The walls were tall and slender, with few openings that were expected to
have little or no influence on the behavior. Control-joint spacing was
approximately 24 feet on center, resulting in shear-wall segments with a
shear span-to-depth ratio of approximately 1.
This building type and occupancy are so common that it was considered
relevant to the study; however, it is recognized that the size and flexibility of
the wood diaphragm would result in a structure that would not fit the
description of a short-period structure, as defined in this study.
4.1 Introduction
A wall may also develop base sliding. In that case, the resistance is provided
by shear friction, as well as the dowel action of the vertical reinforcement,
depending on the magnitude of the axial compressive load on the wall and
the amount of clamping force developed by the vertical reinforcement.
Ideally, a numerical model used to simulate the inelastic behavior and assess
the collapse potential of a reinforced masonry building must be capable of
capturing the aforementioned failure mechanisms and the system effects that
may influence the strength and ultimate behavior of the wall system. In
particular, the model should have sufficient sophistication to accurately
describe the physical mechanisms that govern the inelastic behavior,
Beam-truss models have also been used to analyze reinforced masonry and
concrete shear walls (Moharrami et al., 2015; Lu et al., 2016; Lu and
Panagiotou, 2013; 2016). These models can capture the flexural and shear
behavior of walls. Although they are computationally more efficient than
refined finite-element models, they require more simplifying assumptions
and, therefore, may not be as realistic as finite-element models.
Considering the pros and cons of various modeling methods and the need for
a computationally efficient, yet-accurate method for incremental dynamic
analysis, a two-pronged approach was adopted to assess the collapse
potential of the reinforced masonry building archetypes. First, for each
building archetype, a refined finite-element model was developed. The
model was able to capture both the flexural and the shear behavior of the
walls. Dynamic response-history analyses were performed with the model
using three sets of bi-directional ground motion records that had different
Figure 4-4 shows the refined finite-element model developed in this study for
a two-story commercial building archetype. It is representative of the refined
models developed for all the archetypes. The analyses were conducted with
the commercial software LS-DYNA (LSTC, 2018). Custom material models
were implemented in the software to meet the special modeling needs to
capture the various failure mechanisms. A smeared-crack material law was
implemented in the shell elements to describe the crushing and cracking
behavior of the masonry walls and horizontal concrete diaphragms. For
masonry walls in the bottom story, where significant diagonal shear cracks
might develop, cohesive crack interface elements were introduced to represent
cracks in a discrete fashion, and to model sliding along crack surfaces and
The steel beams in the floor and roof systems and the steel columns in the
gravity frames were represented by beam-column elements with fiber
sections and a bi-linear material law. Zero-length springs were used to
connect the steel beams to the columns and masonry walls.
The cross section of each shell element was discretized into three layers of
masonry or concrete to simulate the out-of-plane bending behavior of a wall
or that of a horizontal diaphragm. A smeared-crack material model was used
to simulate the compressive behavior, as well as tensile cracking of the
material in an efficient fashion. The smeared-crack model adopted a simple
and robust orthotropic material law. As shown in Figure 4-6, before
cracking, the axes of orthotropy were aligned with the directions of the
principal strains. Cracking initiated when the maximum principal stress
reached the tensile strength of the material. Beyond that point, the axes of
orthotropy remained fixed, with directions parallel and perpendicular to the
direction of the first crack, as illustrated in Figure 4-6. Shear strain and shear
stress could develop along the crack since the axes of orthotropy would not
necessarily coincide with the directions of the principal strains after the
development of the crack. After this point, the shear stress was related to the
shear strain with an elastic-perfectly plastic material law, with the initial
yield strength assumed to be 50 percent of the tensile strength, ft. Another
crack could develop in the direction perpendicular to the first crack when the
tensile stress parallel to the first crack exceeded the tensile strength.
The uniaxial stress-strain law for tension and compression in each of the
orthotropic directions is shown in Figure 4-7. After crack initiation, the
stress-strain relation for the tension regime assumed an exponential function
to represent strain softening and a reduced elastic stiffness for unloading and
reloading. The reduced stiffness allowed complete crack closure upon
unloading. In the first compressive loading cycle, the material response was
linearly elastic until it reached the stress level, fo; after that, it was described
by a parabolic function to represent strain hardening and a linear function
with a negative slope to represent post-peak strain softening. Unloading in
compression followed the initial stiffness until reaching zero stress. The
stress remained zero upon the further decrement of the compressive strain
until the strain reached zero. After that point, tensile stress might develop.
The residual plastic strain in compression was denoted by εpl, which was
calculated during each unloading in the compression regime. Reloading in
compression followed the same path, and compressive stress developed when
the compressive strain exceeded εpl. Interaction of the compressive responses
in the two orthotropic directions was modeled in a simple and efficient
manner by assuming that the residual plastic strain, εpl, was isotropic and
equal to the larger of the two developed in these directions. When the
compressive strain exceeded ε0, the tensile strength and shear strength were
reduced in the same proportion as the decrement of the compressive resistance
with respect to the peak. The material was considered completely crushed
when the compressive strain in any of the two directions reached εu. The shell
element was removed when this condition occurred at all sampling points.
This material model was implemented as a user-defined material in LS-
DYNA. Although similar material laws were available in LS-DYNA, they
did not allow the element removal strategy used in this study, which required
the material law to pass information to the adjacent interface elements.
Figure 4-8 shows the cohesive-crack interface model used to simulate shear
and flexural cracks in a wall in a discrete fashion. The interface elements had
zero-thickness and were compatible with the shell elements in that the
interface could have a distributed normal stress developed over the area to
resist the out-of-plane bending moment developed by a shell element. A 3D
plasticity model was used to simulate mixed-mode (mode I, II, and III)
fracture, crack opening and closing, and relative shear sliding in the interface.
Failure was governed by a yield surface that had the shape of a hyperboloid,
as shown in Figure 4-8(b), and damage in shear and tension was modeled with
a set of softening rules that governed the evolution of the yield surface. The
model could simulate reversible joint dilatation due to the wedging action of
the asperities, and irreversible joint compaction due to damage induced by
cyclic shear reversals under compression. The cohesive-crack material law
was based on the work of Kottari (2016), which is an extension of the two-
dimensional cohesive crack law of Koutromanos and Shing (2012). The
material model was implemented as a user-defined material in LS-DYNA.
Each interface element had four Gauss integration points.
The beam elements used to model reinforcing bars had the Hughes-Liu
beam-element formulation (LSTC, 2018). It accounted for geometric
nonlinearity. Although beam elements were computationally more
demanding than truss elements, they could describe bar buckling and dowel
actions. The uniaxial material law of Kim and Koutromanos (2016) was
implemented to model the cyclic behavior of steel. The responses of the
model under a monotonically increasing displacement and cyclic loading are
shown in Figure 4-9. The material model also accounted for bar fracture due
to low-cyclic fatigue. Fracture was triggered when a scalar damage
parameter, which was based on the cumulative plastic work related to the
tensile stress, exceeded a critical value. The model had been calibrated and
used by Moharrami and Koutromanos (2017) for structural collapse
simulations. The beam element was removed as soon as fracture was
detected at any point at a beam cross section.
Figure 4-9 Steel material model for reinforcing bars (Kim and
Koutromanos, 2016).
Bar slip was modeled with the bond-slip law developed by Murcia-Delso and
Shing (2014) for reinforced concrete. The monotonic and cyclic behaviors of
the bond-slip model are shown in Figure 4-11. The bond strength was
specified as a function of the compressive strength of the grout, and the
model accounted for the degradation of bond resistance as the gap between
the bar and the surrounding grout or concrete increased. The masonry
behavior in dowel actions along each of the local coordinates y and z was
represented by a simple uniaxial material law, which is shown in Figure
4-12. The behavior along each direction was assumed to be independent of
the other. Under monotonic loading, the material behavior was initially
elastic-perfectly plastic. Beyond a certain deformation, the model showed a
post-peak softening behavior described by a linear function until a specified
residual strength was reached, as shown in Figure 4-12(a). Unloading
occurred with a stiffness higher than the initial stiffness. After the stress
reached zero, it remained zero upon further decrement in deformation. The
deformation at which the stress reached zero represented a gap. The gap
generated this way could only increase in each loading direction and could be
different in the positive and negative directions. Reloading followed the
initial stiffness once the gap was closed. The initial stiffness and yield
strength were determined with the dowel law proposed by Dulacska (1972)
for reinforced concrete. The calibration was performed with a simple model
consisting of a dowel bar embedded in masonry, which was subjected to
increasing shear displacement along a frictionless joint, as shown in Figure
4-12(b). The slope of the linear softening branch was determined in an ad
hoc manner by matching results of reinforced masonry wall analyses to
quasi-static wall test data, as presented in a following section. Using
Dulacska’s formula, the dowel strength, Pmax, was calculated as a function of
The reduction of the bond and dowel resistances due to the compressive
crushing of the surrounding masonry was modeled with a strength-reduction
factor, k. The factor was equal to one before the onset of crushing of
masonry (i.e., when the maximum compressive strain had not exceeded εo)
and was reduced to a residual value kres when the maximum masonry
compressive strain reached εu, as shown in Figure 4-13. The compressive
strain developed in the surrounding masonry was estimated by assuming that
it was equal to the axial compressive strain developed at the masonry side of
the interface element.
A bi-linear material law was introduced to restrain the spinning of the bar
with respect to the masonry. The interface element was implemented as a
user-defined element in LS-DYNA using the formulation proposed by
Kottari et al. (2017), which allowed the connection of beam elements to
larger shell elements.
The calibrated models were used for the analyses of the building archetypes,
as will be discussed later. For the analyses of the wall tests and the building
archetypes, the values of the material parameters were either kept constant or
specified as functions of the material strengths in a consistent fashion.
Figure 4-14 and Figure 4-15 compare the experimental and numerical results
for flexure-dominated walls tested by Kapoi (2012) and Sherman (2011),
respectively. Both walls had the same dimensions and similar reinforcement
details. The wall in Figure 4-14 had five #4 vertical bars and nine #4
horizontal bars, whereas the wall in Figure 4-15 had five #6 vertical bars and
Figure 4-14 and Figure 4-15 show that the finite-element models were able to
adequately capture the lateral strengths, hysteretic responses, and failure
mechanism of the walls. However, the extent of toe crushing along the
height of the walls was underestimated by the models, as shown by the extent
of element removals in the figures. In the analyses, compressive softening
was localized in a single row of shell elements, leading to unloading in the
elements above, whereas crushing was more spread into the upper courses in
the actual tests. This strain localization was a numerical artifact, as discussed
above.
The two shear-dominated walls modeled here were tested by Voon and
Ingham (2006) and Ahmadi (2012), respectively. One had horizontal
reinforcement and the other did not. As shown in Figure 4-16 and Figure
4-17, the finite-element models were able to reproduce the hysteretic
responses and failure mechanism well.
Archetypes COM1, COM2, and COM3 have one, two, and four stories,
respectively, and were designed for a high-seismic intensity (SMS = 1.5g).
COM4, COM5, and COM6 have the same number of stories as COM1, COM2,
and COM3, respectively, but were designed for a very high-seismic intensity
(SMS = 2.25g). Except for COM4, the latter group had more reinforcement in the
walls, and COM6 had a larger total wall cross-sectional area than COM3. The
walls in COM1 and COM4 had identical designs with the reinforcement
governed by the minimum prescriptive requirements of the code, but COM4 had
a stronger roof diaphragm. The roof system of the archetypes consisted of a
corrugated steel deck on steel joists, and the floor diaphragms were constructed
of concrete slabs cast on corrugated steel pans supported on steel beams and
joists. The gravity loads were carried by the walls and steel columns.
The configurations of the refined finite-element models developed for the six
archetypes are shown in Figure 4-20 (where the wall reinforcement shown
for the two-story case represents that of COM5).
x (Longitudinal)
z (Transverse)
COM1 and COM4 COM2 and COM5
COM3 COM6
Figure 4-20 Finite-element models for commercial building archetypes.
The steel columns and beams in the gravity-frame systems were modeled
with displacement-based beam-column elements having a fiber-section
discretization and a bi-linear material law. Each column had eight elements
per story. Similar discretization was used for the beams. The columns were
assumed to be hinge-connected at their bases. The beams at each floor were
connected with clip angles to the columns and to the webs of the other beams
running in the orthogonal direction. These connections did not provide much
moment or axial-force restraint to the ends of the connected beams. This did
not allow the development of strong composite action between the beams and
the floor slabs at the locations of the connections. To represent this situation,
the beams under the floor slabs in the two- and four-story archetypes were
connected to the supporting beams and the columns with zero-length springs
that did not transmit moments or axial forces but exerted strong translational
restraint in the vertical and lateral directions. Beams were connected to the
walls in the same way. The beams for the roof decks had regular hinge
connections, except for the beam elements representing the joists, which
were connected to the supporting beams without the axial restraint. The
modeling of the horizontal diaphragms and the aforementioned connection
details are further described in the following section.
For the analysis, the weight of each archetype was based on the following
load combination: 1.05D + 0.25L (with D and L being the dead and live
loads, respectively), according to the FEMA P-695 procedure. This resulted
in a total building (excluding the foundation) weight of 274 kips for COM1
and COM4, 870 kips for COM2 and COM5, 2,200 kips for COM3, and
The floor and roof diaphragms were modeled with nonlinear shell and beam
elements. Each floor diaphragm had a concrete slab cast on a corrugated
steel pan with 2-inch ribs running along the transverse direction (z direction)
of the buildings (see Figure 4-20 for the coordinate system). The distance
from the top surface of the concrete slab to the top of the ribs was 3 inches,
and the slab contained temperature and shrinkage steel running perpendicular
to the ribs. To simplify modeling, the ribs were not explicitly represented.
The slab was assumed to have a uniform thickness of 4 inches, including the
concrete topping, and a 0.0474-inch-thick bottom layer representing the steel
pan. The steel layer was set to act only along the direction of the ribs. The
steel beams supporting the slabs were modeled with fiber-section beam
elements with a bi-linear material law. The centroids of the beam and shell
elements were offset vertically to represent the bending stiffness and capacity
of the composite section correctly.
For all the archetypes, the roof diaphragm was a light-corrugated steel deck
with 1.5-inch ribs, as shown in Figure 4-22, running along the transverse
direction (z direction) of the buildings. The deck panels were stitched
together with fasteners and are supported on steel joists. Experimental
studies by Tremblay et al. (2004) showed that the in-plane shear stiffness and
The nonlinear in-plane shear response of the shell elements representing the
roof deck was calibrated with experimental data. The calibration accounted
for both the behavior of the panel and that of the side-lap connectors. The
material law for shear assumed that the panel was elastic up to 40 percent of
the peak strength, had linear hardening afterwards until the peak strength was
reached, and had linear softening after the peak until a constant residual
strength equal to 20 percent of the peak was reached. The initial stiffness
and peak strength were estimated based on the guidelines of the Steel Deck
Institute Diaphragm Design Manual (Luttrell, 1995). The rest of the
parameters were determined with the test data of Tremblay et al. (2004) and
Essa et al. (2003) on roof-deck diaphragms with fasteners similar to those
used for the building archetypes. It was assumed that the peak strength
occurred at a shear angle of 0.008 radians, and the residual strength was
reached at a shear angle of 0.20 radians. The same deformation limits were
used for both diaphragm types based on the data of Essa et al. (2003). Figure
4-22 shows the shear responses of the decks tested by Essa et al. (2003)
under monotonic and cyclic loads with screw fasteners. The roof diaphragms
of the archetypes with screw fasteners were modeled as stronger than the
diaphragms tested by Essa et al. (2003) because the former have a thicker
steel panel (0.0474 inches versus 0.0295 inches) and have panel-to-frame
welds located in every flute rather than every other flute. Figure 4-22 also
shows the response of the model under monotonic and cyclic loading for a
deck panel that had the same configuration as that tested by Essa et al. (2003)
but had the side laps connected with the screw or weld fasteners specified for
the archetypes.
The shell elements representing the roof and floor diaphragms were directly
connected to the shell elements representing the walls. Due to the large
variation in the type and detailing of connectors in practice, the flexibility
The values of the material parameters for the archetype models were
determined with the guidelines developed in the validation studies as
discussed in Section 4.3.8. Expected strengths were used for the masonry
and reinforcing steel. The compressive strength of masonry, f m′ , was
assumed to be 3,125 psi for COM6 and 2,500 psi for the other archetypes.
These values were 1.25 times the respective nominal strengths specified for
the design, as recommended in a prior study (NIST, 2010). The compressive
behavior of masonry was determined in the way consistent with the
validation studies. The ultimate compressive strain, εu, (see Figure 4-7) was
assumed to be -0.047 for the one- and two-story archetypes, and -0.035 for
the four-story archetypes, because the latter had a larger element size (8
inches versus 6 inches). The values of εu were determined according to the
size of the shell elements used in the respective archetype models in order to
have consistent fracture energy dissipation in the expected plastic-hinge
zone, whose height was assumed to be 20 percent of the wall height in the
bottom story. The tensile strength of masonry was assumed to be 10 percent
of f m′ .
For the steel reinforcement, the expected yield and tensile strengths were
taken to be 1.13 times the nominal strengths, based on the data provided by
Nowak et al. (2008), resulting in an expected yield strength of 68 ksi and an
expected tensile strength of 102 ksi. The values of the material parameters
for the cohesive crack interface elements were determined with the
guidelines from the validation studies. For the calculation of the bond and
dowel strengths for the bond-slip/dowel-action interface elements, the
compressive strength of the grout was used. The grout strength was assumed
to be 4,375 psi for COM6 and 3,500 psi for the other archetypes. These
values were based on the data reported by Mavros et al. (2016) and Stavridis
et al. (2016) for masonry with comparable compressive strengths.
Transverse direction
Figure 4-23 Base shear-vs.-story-drift ratio curves for high-seismic
archetypes.
Longitudinal direction
Transverse direction
Figure 4-24 Base shear-vs.-story-drift ratio curves for very high-seismic
archetypes.
Table 4-2 Natural Periods for the First Three Modes of the Building
Archetypes from Refined Finite-Element Models
Natural Period (sec)*
Archetype ID No. of Stories 1st mode 2nd mode 3rd mode
Commercial Buildings: High Seismic
COM1 1 0.164 (T) 0.096 (L) 0.081 (TO)
COM2 2 0.206 (T) 0.190 (T) 0.180 (L)
COM3 4 0.333 (L) 0.294 (T) 0.168 (TO)
Commercial Buildings: Very High Seismic
COM4 1 0.149 (T) 0.091 (L) 0.074 (TO)
COM5 2 0.197 (T) 0.187 (T) 0.177 (L)
COM6 4 0.282 (L) 0.259 (T) 0.166 (TO)
* T: transverse direction; L: longitudinal direction; TO: torsional
3rd Mode
Figure 4-26 Mode shapes for the first three modes of the refined finite-
element model of COM2.
Motion 17 (SUPERST/B-POE)
Figure 4-28 shows the frame model developed for the two-story archetypes
(COM2 and COM5). It is representative of the simplified models developed
for all the building archetypes. The refined finite-element analyses showed
that major damage always concentrated in the first-story walls. For COM2
and COM5, these walls deformed in double curvature with a mixed flexural-
shear behavior that was eventually dominated by shear. Hence, in the
simplified models, the walls in the second stories were modeled with elastic
Figure 4-28 Frame model with OpenSees for COM2 and COM5.
For the calibration of each simplified frame model, two sets of ground
motion records (Motion 1 and Motion 2 shown in Figure 4-27) were used to
determine the parameters of the hysteretic models, and the third set (Motion
17) was used to check the general validity of the calibration. For each set of
records, the ground motions were scaled to two levels, the MCER level and a
higher level. The simplified models calibrated in this way are referred to as
the baseline models in the parametric studies. From here on, they are
identified with a letter B after the respective archetype number. Their
variants considered in the parametric studies are discussed in Section 4.7.
The floor and roof mass was distributed among the nodes of the grillage
model, and the wall mass was lumped at each end of the walls. The mass for
the vertical and horizontal directions was the same. For the response-history
analyses, Rayleigh damping was used with the damping ratio equal to one
percent of the critical for the first and second modes. The initial stiffness of
the model was used to calculate the damping matrix. Even though the
damping ratio was different from that assumed for the refined finite-element
models, it was found to provide a better match of the response histories. In
Appendix B, the response histories and hysteresis curves from the simplified
models are compared to those from the refined models for the three sets of
ground-motion records.
Pushover analyses were performed with the simplified models using the
same load distributions as those used in the refined finite-element analyses.
For these analyses, the grillage model representing the horizontal diaphragms
was assigned a very high in-plane stiffness so that the horizontal
displacement at each floor was uniform. For each building archetype, two
pushover analyses were conducted, one for each direction. In Figure 4-31,
results of the pushover analyses for COM2B are compared to the pushover
curves obtained with the refined finite-element model. The slopes of the
rising branch of the pushover curves were well captured by the simplified
model, but the refined model showed a substantially more gentle post-peak
load degradation. This can be attributed to the fact that the simplified models
were calibrated with the hysteresis curves from the response-history analyses
of the refined models. The hysteretic response included the effects of cyclic
displacement reversals, as well as the bi-axial deformations of the L-shaped
walls, which caused more severe damage and load degradation than the uni-
axial lateral loading in a pushover analysis. As shown in Figure 4-31, in an
average sense, the slopes of the post-peak regime of the pushover curves
from the simplified models match the envelopes of the hysteresis curves from
the response-history analyses of the refined models.
600 600
LS-DYNA - M1 2xMCE LS-DYNA - M1 2xMCE
LS-DYNA - M2 1.8xMCE LS-DYNA - M2 1.8xMCE
400 LS-DYNA - M17 2xMCE 400 LS-DYNA - M17 2xMCE
LS-DYNA - Pushover LS-DYNA - Pushover
OpenSEES - Pushover OpenSEES - Pushover
200 200
Force - X (kips)
Force - X (kips)
0 0
-200 -200
-400 -400
-600 -600
-30 -20 -10 0 10 20 30 -30 -20 -10 0 10 20 30
1
st Roof Drift Ratio - X (%)
Story Drift Ratio - X (%)
Longitudinal direction
800 800
LS-DYNA - M1 2xMCE LS-DYNA - M1 2xMCE
LS-DYNA - M2 1.8xMCE LS-DYNA - M2 1.8xMCE
600 600
LS-DYNA - M17 2xMCE LS-DYNA - M17 2xMCE
LS-DYNA - Pushover LS-DYNA - Pushover
400 OpenSEES - Pushover 400 OpenSEES - Pushover
200 200
Force - Z (kips)
Force - Z (kips)
0 0
-200 -200
-400 -400
-600 -600
-800 -800
-30 -20 -10 0 10 20 30 -30 -20 -10 0 10 20 30
1
st Roof Drift Ratio - Z (%)
Story Drift Ratio - Z (%)
Transverse direction
Figure 4-31 Comparison of pushover analysis results from simplified
(OpenSees) and refined finite-element (LS-DYNA) models
for COM2B.
Eigenvalue analyses were conducted with the simplified frame models. The
natural periods calculated for the first three modes of each archetype are
shown in Table 4-3. Vibration modes dominated by the vertical deflection of
the horizontal diaphragms are ignored. The fundamental periods of the
simplified models matched those of the refined finite element models well.
The mode shapes for the first three modes of COM2B are shown in Figure
4-33. For the one-story and four-story archetypes, the first three modes of
the two models match well. However, for the two-story archetypes, the
second and third mode shapes of the two models do not match.
3rd Mode
Figure 4-33 Mode shapes for the simplified model of COM2B.
Baseline archetypes of one, two, and four stories designed for high-seismic
and very high-seismic ground motions were modeled to investigate response
behavior and collapse performance. The baseline models were calibrated
with results of response-history analyses performed on refined finite-element
models without any modifications. The influence of fundamental period (or
the number of stories), the ratio of the maximum base shear (Vmax) to the
seismic weight (W), and the ratio of Vmax to the design base shear (i.e., the
overstrength factor Ω) on collapse probability were investigated.
The design of a masonry building, such as the number of walls, the geometric
configuration of the walls, and the amount and spacing of the reinforcement,
varies from one building to another depending on the architectural design and
the preference of the engineer. Variations in the design may result in
buildings with different displacement capacities, even though they have the
same number of stories and functional characteristics. To investigate the
influence of these variations on the collapse probability of a building, a
parametric study focusing on displacement capacity was conducted. The
displacement capacity was varied by changing the post-peak slope of the
base shear-vs.-first-story drift curves for each baseline model. In general,
increasing the steepness of the post-peak slope results in a smaller collapse
displacement as determined by incremental dynamic analysis. Two variants
were considered for each baseline archetype model. Variant COMXB-DC1
Different foundation models were developed for the stiff and soft sites, with
the plan view of the foundation model for the soft site shown in Figure 4-34.
The foundation around the perimeter of the building was modeled with
1-foot-long displacement-based beam elements, which had fiber sections that
accounted for material nonlinearity. The foundation underneath each wall
was assumed to be rigid to compensate for the omission of the horizontal
wall dimension, since each wall was represented by a line element. The mass
of the foundation slab and grade beams was included. To capture inertial
interaction, the properties of the soil were modeled with springs and
dampers, which were placed at each node of the foundation model with a
nodal spacing of 1 foot, as shown in Figure 4-35. There were three pairs of
springs and dampers at each node aligned in the x, y, and z directions. The
spring and the damper in each pair were parallel to one another and the
spring had a bi-linear material law with a strain hardening ratio of 15 percent
to account for passive soil pressure. Damping was assumed to be viscous,
and the value of the damping coefficient was calculated from the damping
ratio determined for the site using the formula recommended in NIST GCR
12-917-21, Soil-Structure Interaction for Building Structures (NIST, 2012b).
The vertical springs and dampers did not exert tension by using the model
shown in Figure 4-35. In addition, a rotational spring was introduced at each
node of the exterior foundation to account for the rotational resistance of the
foundation strip about its long axis. This spring had an elastic-perfectly
Figure 4-34 Foundation model for the soft site for SSI analyses.
For the stiff site, the foundation was modeled in a similar way. However,
since failure was expected to occur in the grade beams only, the foundation
slabs underneath the walls and the exterior columns were assumed to be
rigid, and the grade beams were modeled with displacement-based fiber-
section beam elements, as shown in Figure 4-36. The soil springs and
dampers were distributed along the exterior foundation slabs and grade
beams with a 1-foot spacing. The properties of all the soil springs and
dampers are given in Appendix C.
Figure 4-36 Foundation model for the stiff site for SSI analyses.
IDAs were conducted on all archetypes using the 22 sets of far-field ground-
motion records from the FEMA P-695 database. Each set has two orthogonal
horizontal ground-motion components. Each record set was applied to an
archetype model twice, with the orientation of the two components rotated by
90 degrees in the second application. Hence, each archetype model was
analyzed with 2 × 22 record sets. The acceleration response spectra for the
44 records (with 5 percent damping) are shown in Figure 4-37. The spectral
curves shown have been scaled so that the median spectral intensity at the
3
S (g)
a
0
0 0.5 1 1.5 2
Period T (sec)
The records sets were scaled up simultaneously and gradually from a low-
intensity level, with the median spectral intensity at the code-based period (T
in Table 4-1) of the archetype increased in increments of 0.1g until all the
records resulted in collapse. At each increment, the median spectral intensity
(ST) at period T (calculated with 5 percent damping) was plotted against the
maximum first-story drift ratio (taken as the larger of the two orthogonal
directions) induced by each set of ground-motion records. Since the
response of all the archetypes was dominated by the weak first story, the
first-story drift ratio was used for the IDA plots. Each curve represents the
response to one ground-motion record set.
Figure 4-38 shows the IDA curves obtained for COM2B, the MCER spectral
intensity, SMT, for the design earthquake at period T, the median collapse
spectral intensity, SCT,raw, based on the raw data from the IDAs, and the
adjusted median collapse intensity, SCT,factored , which is equal to 1.2 × SSF ×
SCT,raw. The median collapse intensity (SCT,raw) is defined as the median
spectral intensity at which 50 percent of the 44 earthquake record sets results
in collapse. The factor of 1.2 applied to the adjusted collapse intensity is
recommended in the FEMA P-695 methodology if the value of SCT,raw is
obtained from 3D nonlinear analyses using bi-axial ground-motions rather
than 2D analyses. Previous studies have shown that the median collapse
intensity resulting from 3D analyses is on average 20 percent less than that
6.43% S = 3.77g
CT, factorized
5
4 S = 2.37g
CT, raw
(g)
3
T
S
1
S = 1.50g
MT
0
0 5 10 15 20 25
st
1 Story Drift Ratio (%)
Figure 4-38 Incremental dynamic analysis results for COM2B. Red markers
represent the incipient collapse points of individual ground
motions. Vertical and horizontal red dashed lines show the
lognormal fitted median incipient collapse drift ratio and SCT,raw,
respectively.
Incipient collapse was identified as the point (shown as red markers in Figure
4-38) on an IDA curve after which the story drift increased indefinitely with
no further increase in ST (i.e., the point after which the IDA curve goes flat)
or when the first-story drift ratio reached 10 percent, whichever came first.
The latter criterion took into account the rotational limit of the beam-to-
Figure 4-39 shows the collapse rates (collapse fraction) determined from the
IDAs. For the IDA data shown, the spectral intensity was scaled by the
factor of 1.2 to account for the 3D analyses. The collapse fragility curve
shown in the figure is based on a lognormal distribution with the median
collapse intensity equal to SCT,factored and the lognormal standard deviation, β,
assumed to be 0.5. The β value accounts for the dispersion introduced by
uncertainties in the record-to-record variability of the ground motions, design
requirements, test data, and modeling. Prior studies (NIST, 2010) on similar
building archetypes have indicated that 0.5 is a good estimate for the β value.
The difference between the collapse fraction data points and the collapse
fragility curve shown in the Figure 4-39 is largely due to the spectral shape
factor (SSF), which was applied to the modified median spectral intensity
SCT,factored to produce the fragility curve, but not to the collapse fraction data.
The probability of collapse at the MCER ground-motion intensity P[C|MCER]
was determined directly from the collapse fragility curves.
1
0.8
Collapse Fraction or Probability
0.6
(S , 0.5)
CT,factorized
0.4
0.2
Collapse Probability
Collapse Fraction
0
0 2 4 6 8 10 12 14
S (g)
T
Figure 4-39 Collapse rates from IDA and derived collapse fragility
curve for COM2B. ST for the collapse fraction data has
been scaled only by the factor of 1.2 for 3D analyses.
Select response quantities obtained for each intensity stripe (at increments of
0.1g) were archived in a spreadsheet file. Table 4-5 summarizes the archived
response quantities, which include the peak response quantities for each
record set and for each response direction of an archetype model, the
response statistics for the 2 × 22 ground motion record sets, and the collapse
statistics. The file names are catalogued in Appendix D.
5.1 Introduction
This chapter presents results of three parametric studies performed for the
commercial reinforced masonry building archetypes, whose baseline models
are denoted as COM1B through COM6B. The probability of collapse was
calculated for each baseline model and its variants considered in the
parametric studies using the FEMA P-695 methodology, which consisted of
pushover analyses and incremental dynamic analyses (IDAs). The analyses
were conducted with the simplified frame models that were calibrated with
refined finite-element models, as discussed in Chapter 4.
800 800
Force - X (kips)
Force - X (kips)
600 600
400 400
200 200
0 0
0 5 10 15 20 0 5 10 15 20
st
1 Story Drift Ratio - X (%) Roof Drift Ratio - X (%)
Longitudinal direction
1200 1200
COM1B COM1B
COM2B COM2B
1000 1000
COM3B COM3B
800 800
Force - Z (kips)
Force - Z (kips)
600 600
400 400
200 200
0 0
0 5 10 15 20 0 5 10 15 20
st
1 Story Drift Ratio - Z (%) Roof Drift Ratio - Z (%)
Transverse direction
Figure 5-1 Pushover curves for high-seismic baseline models (COM1B,
COM2B, and COM3B).
1200 1200
Force - X (kips)
Force - X (kips)
900 900
600 600
300 300
0 0
0 5 10 15 20 0 5 10 15 20
st
1 Story Drift Ratio - X (%) Roof Drift Ratio - X (%)
Longitudinal direction
1800 1800
COM4B COM4B
COM5B COM5B
1500 1500
COM6B COM6B
1200 1200
Force - Z (kips)
Force - Z (kips)
900 900
600 600
300 300
0 0
0 5 10 15 20 0 5 10 15 20
st
1 Story Drift Ratio - Z (%) Roof Drift Ratio - Z (%)
Transverse direction
Figure 5-2 Pushover curves for very high-seismic baseline models (COM4B,
COM5B, and COM6B).
1800 1800
COM1B COM1B
COM2B COM2B
1500 1500
COM3B COM3B
COM4B COM4B
1200 COM5B 1200 COM5B
COM6B COM6B
Force - X (kips)
Force - Z (kips)
900 900
600 600
300 300
0 0
0 0.5 1 1.5 2 2.5 3 0 0.5 1 1.5 2 2.5 3
Roof Drift Ratio - X (%) Roof Drift Ratio - Z (%)
Figure 5-1 and Figure 5-2 show that COM5B (two-story, very high-seismic
site) has a higher displacement capacity than COM2B (two-story, high-
seismic site) even though the former has more vertical reinforcement.
However, this difference in displacement capacity is less significant in the
pushover analysis results from the refined finite-element models (as shown in
Figure 4-23 and Figure 4-24) probably because the simplified models are
calibrated with the hysteresis curves generated with bi-directional earthquake
loading. If the behavior of a wall is governed by flexure, a higher amount of
vertical reinforcement will lead to a smaller displacement capacity because it
will cause more severe masonry toe crushing. However, this is not
necessarily true for wall behavior dominated by diagonal shear or base
sliding. For these cases, the higher amount of vertical reinforcement
Figure 5-1 and Figure 5-2 also show that COM6B (four-story, very high-
seismic site) has a higher displacement capacity than COM3B (four-story,
high-seismic site). This can be largely attributed to the fact that COM6B has
more wall cross-sectional area than COM3B.
Table 5-1 shows that the ratio of Vmax to W decreases with increasing number
of stories, indicating that lower-rise archetypes have more reserve strength.
The one-story archetypes have by far the largest Vmax/W values because, as a
result of the archetypes having configurations representative of actual
buildings, they have a lot more walls than what is needed to resist code-
specified seismic demands.
Table 5-2 shows the values of other parameters obtained from the pushover
analyses. In this table, Vmax,av is the average of the maximum base shear
strengths obtained for the two orthogonal directions of the archetype model,
Ω is the overstrength factor based on Vmax,av, and µT is the average of the
period-based ductility values calculated for the two orthogonal directions.
Table 5-2 also shows the values of the median collapse intensity SCT
determined from the raw IDA data (equal to SCT,raw defined in Section 4.8)
and the calculated collapse margin ratios. The collapse margin ratio
(CMR3D) is calculated as 1.2 × SCT /SMT, where SMT is the MCER spectral
intensity at the design period of the archetype. The adjusted collapse margin
ratio (ACMR) is calculated as SSF × CMR3D.
The collapse fraction data obtained from the IDAs and the lognormal
collapse fragility curves derived from the IDA data for the high-seismic and
very high-seismic archetypes are presented in Figure 5-4 and Figure 5-5,
respectively. The dispersion factor, β, for these curves is assumed to be 0.5.
The collapse probabilities for the 0.5 × MCER- and MCER-level ground
motions are determined from the collapse fragility curve for each baseline
model. As shown in Table 5-2, all the baseline models have collapse
probabilities less than 10 percent for the MCER-level motion. The one-story
archetypes have significantly lower probabilities of collapse. This applies to
the entire spectral intensity range, as shown by the fragility curves in Figure
5-4 and Figure 5-5.
Figure 5-5 Collapse rates from IDA data and collapse fragility
curves derived for very high-seismic baseline models.
ST for the collapse fraction data has been scaled by 1.2,
whereas that for the collapse probability curve has
been scaled by 1.2×SSF.
Table 5-4 shows the collapse rates and the mean peak first-story drift ratios
of the survivors for the 0.5 × MCER- and MCER-level ground motions based
on the raw IDA data without applying the SSF or the 3D-analysis factor.
In Figure 5-6 through Figure 5-8, the probability of collapse for the MCER-
level ground motions is plotted against the normalized pushover strength
(Vmax,av/W), the overstrength factor (Ω), and the calculated fundamental
period (T1), respectively. These figures show that the probability of collapse
increases when the normalized pushover strength decreases or when the
period increases (due to the increase of the number of stories). The clearest
trend is the influence of the overstrength factor, as shown in Figure 5-7. The
collapse probability increases as the value of Ω decreases. However, the
relation between the collapse probability and Ω is not linear. There is a rapid
increase in collapse probability when Ω drops below 2.5.
0.12
COM3B
0.1 COM6B
0.08
COM5B
]
R
0.06
P[C|MCE
0.04
COM2B
0.02
COM4B
COM1B
0
0.2 0.4 0.6 0.8 1 1.2 1.4
V /W
max,av
COM3B
0.1
COM6B
0.08
]
COM5B
R
0.06
P[C|MCE
0.04
COM2B
0.02
COM4B
COM1B
0
0 2 4 6 8
0.12
COM3B
0.1
COM6B
0.08
COM5B
]
R
0.06
P[C|MCE
0.04
COM2B
0.02
COM4B
COM1B
0
0 0.1 0.2 0.3 0.4
T (sec)
1
The collapse fraction data and the collapse fragility curves derived for
COM2B, COM2B-DC1, and COM2B-DC-2 are shown in Figure 5-9, and the
collapse fragility curves for all the baseline models and their variants are
shown collectively for the high-seismic and very high-seismic cases in
Figure 5-10 and Figure 5-11, respectively.
0.8
Collapse Fraction or Probability
0.6
0.4
Collapse Probability (COM2B)
Collapse Fraction (COM2B)
Collapse Probability (COM2B-DC1)
0.2
Collapse Fraction (COM2B-DC1)
Collapse Probability (COM2B-DC2)
Collapse Fraction (COM2B-DC2)
0
0 2 4 6 8 10 12 14
S (g)
T
Figure 5-9 Collapse rates from IDA data and collapse fragility curves
derived for COM2B, COM2B-DC1, and COM2B-DC2.
ST for the collapse fraction data has been scaled by 1.2,
whereas that for the collapse probability curve has been
scaled by 1.2×SSF.
0.9
0.8
0.7
0.6
Collapse Probability
0.5
COM1B
COM1B-DC1
0.4
COM1B-DC2
0.3
COM2B
COM2B-DC1
0.2 COM2B-DC2
COM3B
0.1 COM3B-DC1
COM3B-DC2
0
0 2 4 6 8 10 12 14 16 18 20
S
T
0.9
0.8
0.7
0.6
Collapse Probability
0.5
COM4B
COM4B-DC1
0.4
COM4B-DC2
0.3
COM5B
COM5B-DC1
0.2 COM5B-DC2
COM6B
0.1 COM6B-DC1
COM6B-DC2
0
0 2 4 6 8 10 12 14 16 18 20
S
T
Table 5-5 summarizes the pushover analysis results obtained for the
displacement capacity parametric study, and Table 5-6 shows the calculated
collapse probabilities. The collapse probabilities vary in an expected
manner, and the variation with respect to the change in displacement capacity
Table 5-8 shows the collapse rates and the mean peak first-story drift ratios
of the survivors for the 0.5 × MCER- and MCER-level ground motions based
on the raw IDA data without applying the SSF or the 3D-analysis factor. In
general, the collapse rate at MCER increases significantly as the displacement
capacity is decreased by 30 percent (as shown by COMXB-DC2).
Furthermore, the mean peak first-story drift ratio of survivors at MCER
increases slightly as the displacement capacity increases.
In Figure 5-12, the pushover analysis curves obtained for the stiff and soft
sites are compared to those for the baseline model. For the stiff site, the
influence of the soil and foundation flexibility on the ascending branch of the
pushover curves is negligible. Nevertheless, the post-peak slope of the
pushover curve for the x direction for the stiff site is significantly more gentle
than that for the baseline model. This is largely attributed to the rotation and
flexural yielding of the foundation slab below the end walls (Wall 1 and
Wall 3 as numbered in Figure 4-36), which resulted in a reduction of the
plastic rotation demand on the bottom spring of the wall elements and thus
delayed the inelastic softening behavior of the walls. The change of the
rotation demand also slightly reduced the peak shear resistance of the walls,
which depended on the moments developed in the top and bottom springs of
each wall element.
600 600
w/o SSI w/o SSI
w/ SSI - Stiff Site w/ SSI - Stiff Site
500 w/ SSI - Soft Site 500 w/ SSI - Soft Site
400 400
Force - X (kips)
Force - Z (kips)
300 300
200 200
100 100
0 0
0 5 10 15 20 0 5 10 15 20
Roof Displacement - X (in) Roof Displacement - Z (in)
Figure 5-12 Pushover curves for COM2B with and without soil springs and
dampers.
Figure 5-13 shows the vertical displacements of the foundations for the stiff
and soft sites at the respective peak base shear. Because the stiff site had a
smaller and weaker foundation, it had more significant foundation rotation
and uplift. The stiffer soil also resulted in a more severe foundation edge
uplift. For the soft site, the foundation rotation and uplift were significantly
smaller because the foundation was larger and stronger. This resulted in only
a small change in the post-peak slopes of the pushover curves.
X Soft Site
1.5 Push Dir. Stiff Site
Displacement - Vert. (in)
0.5
Figure 5-13 Vertical displacements of the foundations for the stiff and soft
sites at the respective maximum base shear.
These results indicate that the pushover response was significantly affected
by the flexural deformation and rotation of the foundation slab and less by
the soil deformation. Figure 5-14 shows the forces developed in the vertical
soil springs (which have a spacing of 1 foot) along the foundation strip in the
longitudinal (x) direction of the building at the respective maximum base
shear for both sites. The stiff site has zero force in the soil springs along a
significant portion of the foundation due to the foundation uplift. For both
the soft and stiff sites, the compressive forces developed in the soil springs
are far below the bearing capacities. Because of the significant loss of soil
contact for the stiff site, the foundation could slide under lateral earthquake
forces. However, the influence of the soil-foundation contact condition on
the horizontal (bearing or friction) resistance of the soil was not accounted
for in the model.
5
X
Push Dir.
0
-5
Spring Force - Vert. (kip)
-10
-15
Figure 5-14 Forces in vertical soil springs for the stiff and soft sites at the
respective maximum base shear.
The moments developed in the foundation slabs and grade beams at the
respective maximum base shear are shown in Figure 5-15. Flexural yielding
occurs in the foundation slabs or grade beams for both sites. To understand
the interaction of the walls with the foundation slab and the underlying soil,
and the resulting lateral load resisting mechanisms of the system, a U-shaped
end segment of the foundation slab underneath Wall 1 and Wall 4 (the end
walls on the windward side) for each site was isolated as a free body, as
500
Yielded
-500
Yielded
Wall 1 Ext Column Wall 2 Ext Column Wall 3
0 10 20 30 40 50 60 70 80 90
Long Strip (ft)
Figure 5-15 Moments in foundation slabs and grade beams for the stiff and
soft sites at the respective maximum base shear.
Table 5-10 compares the fundamental periods and the key parameters of the
pushover curves obtained for COM2B, COM2B-SS1, and COM2B-SS2.
Model COM2B-SS2 has a significantly longer period than COM2B due to
the soft soil.
Table 5-9 Modal and Pushover Analysis Results of the SSI and Foundation
Flexibility Parametric Study Archetype Models
Transverse Direction Longitudinal Direction
Archetype T1 ∆U,80 ∆U,max T1 ∆U,80 ∆U,max
ID (sec) Vmax/W (in/in) (in/in) (sec) Vmax/W (in/in) (in/in)
Commercial Buildings: High Seismic
COM2B 0.20 0.68 0.017 0.051 0.18 0.64 0.027 0.056
COM2B-SS1 0.24 0.67 0.022 - 0.27 0.62 0.041 -
COM2B-SS2 0.35 0.68 0.018 0.053 0.42 0.65 0.029 0.056
Table 5-10 Pushover and Collapse Analysis Results of SSI and Foundation Flexibility Parametric Study
Archetype Models
Pushover Analysis Collapse Analysis
P[CO| P[CO|
Archetype SCT(1) 0.5MCER] MCER]
ID Vmax,av/W Ω µT (g) CMR3D SSF ACMR (%) (%)
Commercial Buildings: High Seismic
COM2B 0.66 3.29 8.516 2.37 1.90 1.33 2.52 0.07 3.22
COM2B-SS1 (2)
0.64 3.22 9.469 - - - - - -
COM2B-SS2
0.67 3.33 9.172 2.35 1.88 1.33 2.50 0.07 3.34
(Unfiltered)
COM2B-SS2
0.67 3.33 9.172 2.36 1.89 1.33 2.51 0.07 3.28
(Filtered)
(1)
Without correction for 3D analysis (raw data).
(2)
No incremental dynamic analysis performed.
3
COM2B
COM2B-SS2-Unfiltered
COM2B-SS2-Filtered
2.5
2
(g)
1.5
T, Median
S
0.5
0
0 5 10 15
st
Median Max 1 Story Drift Ratio (%)
0.8
Collapse Fraction or Probability
0.6
0.4
Collapse Probability (COM2B)
Collapse Fraction (COM2B)
Collapse Probability (COM2B-SS2-Unfiltered)
0.2
Collapse Fraction (COM2B-SS2-Unfiltered)
Collapse Probability (COM2B-SS2-Filtered)
Collapse Fraction (COM2B-SS2-Filtered)
0
0 2 4 6 8 10 12 14
S (g)
T
Figure 5-18 Collapse rates from IDA data and collapse fragility
curves derived for COM2B and COM2B2-SS2. ST
for the collapse fraction data has been scaled by
1.2, whereas that for the collapse probability curve
has been scaled by 1.2×SSF.
Table 5-11 shows the median values of the peak roof drift, peak first-story
drift, and response spectral intensity at incipient collapse, as well as the
values of the lognormal standard deviation of these quantities, for the SSI
and foundation flexibility parametric study. As shown, the inclusion of SSI
and foundation flexibility in the analysis results in very small changes.
Table 5-12 Collapse Rates and Mean Peak First-Story Drift Ratios of Survivors at 50 Percent-of-MCER and
MCER Ground-Motion Intensities of the SSI and Foundation Flexibility Parametric Study
Archetype Models
6.1 Introduction
Key findings of the parametric studies are summarized and discussed in the
following sections. A table is provided with each parametric study,
summarizing key archetype model properties (e.g., number of stories, first-
mode period, T1, overstrength, Ω, and normalized pushover strength, Vmax/W)
and key collapse results (e.g., roof and first-story drift ratios at the point of
incipient collapse, the collapse margin ratio, CMR3D, and the probability of
collapse given MCER ground motions). The target range of benchmark
collapse probabilities are also provided for comparison with the collapse
probabilities of archetype models designed for high-seismic loads (SMS =
Table 6-1 Summary of Key Model Properties and Collapse Results of Baseline Archetype Models
Model Properties Collapse Results
Benchmark
Period Strength Drift Ratio* P[COL| Collapse
Archetype No. of T1 MCER] Probability
ID Stories (sec) Ω Vmax/W Roof First Story CMR3D (%) (%)
High-Seismic (SMS = 1.5g) Baseline Archetype Models
COM1B 1 0.14 5.6 1.11 0.088 0.088 3.35 0.14 0 to 2
COM2B 2 0.19 3.3 0.66 0.035 0.064 1.90 3.2 0 to 5
COM3B 4 0.31 2.4 0.47 0.015 0.058 1.43 9.9 0 to 5
Very High-Seismic (SMS = 2.25g) Baseline Archetype Models
COM4B 1 0.14 4.0 1.19 0.092 0.092 2.42 1.0 NA
COM5B 2 0.18 2.6 0.77 0.048 0.088 1.56 7.2 NA
COM6B 4 0.25 2.3 0.68 0.021 0.084 1.46 9.4 NA
* Median drift ratio at incipient collapse.
14%
12%
ASCE/SEI 7-10
10% 4s 4s
8%
2s
6% ≥ 2-story BM
4%
2s 1-story BM
2%
1s 1s
0%
0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0
Overstrength, Ω
Figure 6-2 MCER collapse probabilities of baseline archetype models
plotted as a function of average archetype model
overstrength (Ω ), and benchmark (BM) values of MCER
collapse probability and the MCER collapse-safety objective
of ASCE/SEI 7-10.
Table 6-2 Summary of Key Model Properties and Collapse Results of Displacement Capacity Parametric
Study Archetype Models with Baseline, Enhanced (DC1) and Reduced (DC2) Post-Capping
Strength
Model Properties Collapse Results
Reinforced- Benchmark
Masonry Building Period Strength Drift Ratio* P[COL| Collapse
Archetype Model No. of T1 MCER] Probability
ID Stories (sec) Ω Vmax/W Roof First Story CMR3D (%) (%)
High-Seismic (SMS = 1.5g) Baseline Archetype Models
COM1B-DC1 1 0.14 5.6 1.11 0.093 0.093 3.54 0.10 0 to 2
COM1B 1 0.14 5.6 1.11 0.088 0.088 3.35 0.14 0 to 2
COM1B-DC2 1 0.14 5.6 1.11 0.072 0.072 2.99 0.29 0 to 2
COM2B-DC1 2 0.19 3.3 0.66 0.042 0.078 2.04 2.3 0 to 5
COM2B 2 0.19 3.3 0.66 0.035 0.064 1.90 3.2 0 to 5
COM2B-DC2 2 0.19 3.3 0.66 0.027 0.048 1.70 5.1 0 to 5
COM3B-DC1 4 0.31 2.4 0.47 0.019 0.073 1.56 7.2 0 to 5
COM3B 4 0.31 2.4 0.47 0.015 0.058 1.43 9.9 0 to 5
COM3B-DC2 4 0.31 2.4 0.47 0.011 0.043 1.27 14.7 0 to 5
Very High-Seismic (SMS = 2.25g) Baseline Archetype Models
COM4B-DC1 1 0.14 4.0 1.19 0.094 0.094 2.51 0.8 NA
COM4B 1 0.14 4.0 1.19 0.092 0.092 2.42 1.0 NA
COM4B-DC2 1 0.14 4.0 1.19 0.080 0.080 2.20 1.6 NA
COM5B-DC1 2 0.18 2.6 0.77 0.050 0.092 1.65 5.8 NA
COM5B 2 0.18 2.6 0.77 0.048 0.088 1.56 7.2 NA
COM5B-DC2 2 0.18 2.6 0.77 0.040 0.073 1.42 10.4 NA
COM6B-DC1 4 0.25 2.3 0.68 0.023 0.089 1.53 7.7 NA
COM6B 4 0.25 2.3 0.68 0.021 0.084 1.46 9.4 NA
COM6B-DC2 4 0.25 2.3 0.68 0.018 0.070 1.33 12.7 NA
* Median drift ratio at incipient collapse.
8% DC1
DC1
6% DC2 DC1
4%
DC1
2%
DC2 DC1
DC2
0%
4.0 5.0 6.0 7.0 8.0 9.0 10.0 11.0
1st-Story Drift Ratio at Incipient Collapse (%)
For the archetype model with nonlinear springs and dampers representing
soft soil site conditions, collapse analyses were performed using earthquake
records modified for kinematic interaction effects (i.e., filtered records), as
well as the standard set of (unmodified) FEMA P-695 earthquake records.
Table 6-3 summarizes key model properties and collapse results for the two-
story baseline archetype model (COMB2B) with a fixed base and variant
archetype models with flexible foundations that incorporate SSI effects.
Table 6-3 Summary of Key Properties and Collapse Results of SSI and Foundation Flexibility Archetype
Models with a Fixed Base and with Flexible Foundations that Incorporate SSI Effects
Model Properties Collapse Results
Reinforced- Benchmark
Masonry Building Period Strength Drift Ratio* P[COL| Collapse
Archetype Model No. of T1 MCER] Probability
ID Stories (sec) Ω Vmax/W Roof First Story CMR3D (%) (%)
High-Seismic (SMS = 1.5g) Archetype Models
COM2B 2 0.19 3.3 0.66 0.035 0.064 1.90 3.2 0 to 5
COM2B-SS1 2 0.26 3.2 0.65
COM2B-SS2
2 0.39 3.3 0.67 0.037 0.067 1.88 3.3 0 to 5
(unfiltered)
COM2B-SS2
2 0.39 3.3 0.67 0.036 0.066 1.89 3.3 0 to 5
(filtered)
* Median drift ratio at incipient collapse.
The overall findings of the SSI and foundation flexibility parametric study
include the following:
1. Nonlinear Static (Pushover) Response. The nonlinear static (pushover)
curves of variant building archetype models representing stiff (SS1) and
soft (SS2) site conditions were found to be qualitatively similar to the
pushover curve of the baseline building archetype model with rigid,
fixed-base foundations (e.g., see Figure 5-12). However, these two
variant models were moderately to significantly more flexible at
pushover displacements up to peak strength, and the soft site model
showed a gentler post-capping slope compared to the fixed-based case.
For the archetype models evaluated in this parametric study, the
foundations had sufficient strength to develop yielding mechanisms and
Table 6-4 Summary of Key Model Properties and Collapse Results of Archetype Models of Prior Studies
of Special Reinforced Masonry Shear Wall Buildings Designed for High-Seismic Loads
Model Properties Collapse Results
The NIST GCR 10-917-8 study evaluated collapse performance for one-story,
two-story, four-story, eight-story and twelve-story heights of special and
ordinary reinforced masonry building archetypes designed for two FEMA
P-695 seismic levels, SDC Dmax (SMS = 1.5g) and SDC Dmin (SMS = 0.75g); only
model properties and collapse results for one-story, two-story and four-story
special reinforced masonry archetypes designed and evaluated for SMS = 1.5g
(i.e., high-seismic design level) are reported in Table 6-4. To determine how
collapse margin ratios vary with design R values, the NIST GCR 12-917-20
study redesigned the one-story, two-story and four-story high-gravity load
archetypes of the NIST GCR 10-917-8 study using R values of 1, 2, 4, and 8.
Model properties and collapse results are reported in Table 6-4 for R = 4 and R
= 6 designs of the NIST GCR 12-917-20 study, which bound the value of R = 5
used for design of the archetypes of this study (as described in Chapter 3).
Two collapse results plotted in Figure 1-1 of this report for short-period
masonry archetypes represent the average values of the collapse probabilities
of performance groups PG-1S and PG-5S of the NIST GCR 10-917-8 study
(i.e., collapse probabilities of 15 percent and 26 percent, respectively, for an
average design period of T = 0.32 seconds). The individual collapse
probabilities for one-story, two-story and four-story archetype models that
range from 6.8 percent and 8.4 percent for four-story archetype models to 79
percent for one-story archetype models are reported in Table 6-4.
The results of the NIST GCR 12-917-20 study show similar trends (although
different values) of collapse probabilities for one-story and two-story
archetype models, but significantly different and much lower collapse
probabilities for the four-story archetype models (i.e., only 1.4 percent) than
those of the NIST GCR 10-917-8 study. The explanation for differences in
collapse probabilities of the two NIST studies and those of this study is
As per the original “Zone 4” approach of FEMA P-695, all of the ground
motions represent “far-field” sites and purposely ignore higher levels of
ground shaking typical of sites closer to the fault(s) governing site seismic
Damage and Loss Estimation. The archetype models, methods, and results
of this study that better represent observed collapse performance of
reinforced masonry buildings may be of interest to practitioners developing
estimates of earthquake damage and loss using, for example, the FEMA P-58
technology (FEMA, 2018), or other methods, such as those of the insurance
industry. The probability of collapse is an important element of building
damage and loss estimation and the improved models and methods of this
study would provide a more reliable characterization of the earthquake
collapse risk of reinforced masonry buildings.
Areas for Future Improvement. The following is a list of specific areas for
future improvement of numerical modeling of reinforced masonry.
• Even though the refined finite-element modeling scheme presented in
Chapter 4 has been extensively verified, the numerical models have
many material parameters to calibrate. It would be desirable to perform
a systematic numerical parametric study to evaluate the sensitivity of the
results to the modeling assumptions and material parameters to provide
A.1 Introduction
A.2.1 Codes
Design loads were based on design requirements of the 2015 IBC and, by
reference, ASCE/SEI 7-10
A.2.2 Materials
Material Properties
Hollow Concrete f’m = 2,000 psi—ASTM C652
Masonry Units f’m = 2,500 psi (COM 9 only)
Unit Sizes 8 in x 8 in x 16 in nominal (minimum),
12 in x 8 in x 16 in nominal (COM9 only)
Bond Pattern Running Bond
Mortar Type S
Average Compressive Strength = 2000 psi—ASTM C270
Steel Reinforcing Fy = 60 ksi—ASTM A615
Soil properties were assumed to represent typical West Coast sites. The
assumed site-specific soil properties are summarized in Table A-2. For the
soil-structure interaction (SSI) and foundation flexibility parametric study
only, a soft soil profile was assumed, with a lower bearing pressure. Both
stiff and soft soil sites were characterized as Site Class D.
Archetype design dead and live loads are summarized in Table A-3. The
weight of all walls, except those of the exterior, was averaged across the
floor plate, and a portion was assigned to the roof for the determination of
seismic loads. The interior wall dead load used was 15 pounds per square
foot (psf) distributed uniformly over each story.
ASCE/SEI 7-10 allows several methods for calculating the period for
masonry shear wall structures. The value of Ta calculated using ASCE/SEI 7
Equation 12.8-7 is used commonly in practice, whereas FEMA P-695 studies
use T = CuTa for design. The archetypes were designed using Ta, but,
because the periods all fall in or very near the constant acceleration region of
the design spectra, T < Ts, the design was unaffected by this choice. Table A-
4 presents both periods of vibration for all archetypes.
Key configuration and design criteria for the reinforced masonry building
archetypes are summarized in Table A-6.
In the following sections, the structural designs for each of the commercial
archetypes are documented, including plans, elevations, and details. The
floor and roof framing are identical for all COM archetypes and are shown
only once for the COM2 archetypes. The foundation designs are unique, but
similar, and are indicated on a single plan (Figure A-3). Masonry
reinforcement for all unique conditions are shown in wall elevation views.
The structural designs for the baseline archetypes (e.g., COM2B) were
identical to the designs for the variant archetypes (e.g., COM2B-DC), with
the one exception being the foundation design for the COM2 archetype on a
soft soil site for the SSI and foundation flexibility parametric study. Hence,
plans, elevations, and details labeled with COMX should be assumed to
apply to all variants, unless otherwise noted.
Selected design results for the individual walls of the high-seismic and very
high-seismic commercial archetypes are tabulated in Table A-7. Design
results for the moderate seismic archetypes are not provided because these
archetypes were not ultimately included in the study. These results include:
• demand-to-capacity ratios (D/C) for strength,
• governing response (flexure or shear),
• comments indicating triggers for relevant code provisions in TMS 402,
• vertical and horizontal reinforcement, and
• the calculated drift ratio divided by the maximum allowed drift ratio.
“EW Straight Wall” Refers to the straight walls oriented to carry shear in
the East-West direction, centered on the North and
South elevations.
“EW Flanged Wall” Refers to the corner flanged wall, specifically the
portion that is oriented to carry shear in the East-West
“NS Flanged Wall” Refers to the corner flanged wall, specifically the
portion that is oriented to carry shear in the North-
South direction, and which appears in the East and
West elevations.
Table A-7 Design Results for COM1, COM2, and COM3 Archetypes
Drift Ratio/
Wall D/C(1) Criteria(2) Comment Reinforcement Max Drift Ratio
COM1: One Story, High Seismic (8" CMU, f′m = 2000 psi)
Vertical: #4 @32"
EW Straight Wall 0.34 Flexure Vn ≥ 2.5Vu(3)
Horizontal: #5 @ 32"
Vertical: #5 @ 48"
NS Flanged Wall 0.14 Flexure Vn ≥ 2.5Vu(3)
Horizontal: #5 @ 32"
COM2: Two Story, High Seismic (8" CMU, f′m = 2000 psi)
Vertical: #5 @ 24"
EW Straight Wall 0.67 Flexure Vn ≥ 2.5Vu(3)
Horizontal: #5 @ 32"
Vertical: #5 @ 24"
EW Flanged Wall 0.69 Flexure Vn ≥ 2.5Vu(3) 0.16
Horizontal: #5 @ 32"
Vertical: #5 @ 32"
NS Flanged Wall 0.71 Flexure Vn ≥ 2.5Vu(3)
Horizontal: #5 @ 32"
COM3: Four Story, High Seismic (8" CMU, f′m = 2000 psi)
Vertical: #6 @ 16"
EW Straight Wall Base 1.02(5) Flexure ρ/ρmax=1.0(4)
Horizontal: #4 @ 24"
Vertical: #6 @ 32"
EW Straight Wall Level 3 0.63 Flexure -
Horizontal: #4 @ 24"
Vertical: #5 @ 16"
EW Flanged Wall Base 1.02(5) Flexure ρ/ρmax=1.34(4)
Horizontal: #4 @ 24"
0.71
Vertical: #5 @ 32"
EW Flanged Wall Level 3 0.67 Flexure -
Horizontal: #4 @ 24"
Vertical: #5 @ 16"
NS Flanged Wall Base 0.97 Flexure -
Horizontal: #4 @ 24"
Vertical: #5 @ 32"
NS Flanged Wall Level 3 0.75 Flexure ρ/ρmax=1.04(4)
Horizontal: #4 @ 24"
(1)
D/C = Demand/Capacity ratio.
(2)
Criteria indicates if the maximum demand/capacity ratio (D/C) resulted from combined flexure (with compression) or shear.
(3)
TMS 402 §7.3.22.6.1.1 states Vn need not exceed 2.5Vu. When this provision is invoked, shear-dominated behavior may
result.
(4)
TMS 402 §9.3.6.5 was invoked to allow exceedance of ρmax by meeting checks for confinement.
(5)
D/C ratio near or slightly > 1 accepted. Load combination including Ev governs.
Vertical: #4 @32"
EW Straight Wall 0.52 Flexure Vn ≥ 2.5Vu(3)
Horizontal: #5 @ 32"
Vertical: #4 @32"
EW Flanged Wall 0.44 Flexure Vn ≥ 2.5Vu(3) < 0.01
Horizontal: #5 @ 32"
Vertical: #5 @ 48”
NS Flanged Wall 0.32 Flexure Vn ≥ 2.5Vu(3)
Horizontal: #5 @ 32"
COM5: Two Story, Very High Seismic (8" CMU, f′m = 2000 psi)
Vertical: #5@16"
EW Straight Wall 0.87 Flexure -
Horizontal: #5 @ 32"
Vertical: #5 @ 16"
EW Flanged Wall 0.84 Flexure - 0.25
Horizontal: #5 @ 32"
Vertical: #5 @ 24”
NS Flanged Wall 0.91 Shear -
Horizontal: #5 @ 32"
COM6: Four Story, Very High Seismic (12" CMU, f′m = 2500 psi)
Vertical: #8 @ 16"
EW Straight Wall Base 0.87 Flexure -
Horizontal: #5 @ 24"
Vertical: #5 @ 24"
EW Flanged Wall Base 0.59 Flexure -
Horizontal: #6 @ 16"
0.89
EW Flanged Wall Level Vertical: #5 @ 24"
0.24 Flexure -
3 Horizontal: #6 @ 16"
Figure A-1 Single-story commercial building foundation plan: COM1, COM4, COM7.
Figure A-2 Single-story commercial building roof framing plan: COM1, COM4, COM7.
Figure A-3 Two-story commercial building foundation plan: COM2, COM5, COM8.
Figure A-4 Two-story commercial building foundation plan: COM2 archetype on soft soil.
Figure A-6 Two-story commercial building roof framing plan: COM2, COM5, COM8.
In the following sections, the structural designs for each of the multi-family
residential and hotel archetypes are documented, including plans, elevations,
and details. The floor and roof framing are identical for all archetypes and
are shown only once for the RES1 archetype. The foundation designs are
unique, but similar, and are tabulated on a single plan. Masonry
reinforcement for all unique conditions are shown in wall elevation views.
A.4.1.1 Four-Story Residential Building Plans (RES1, RES2, RES3)
Figure A-38 Four-story residential building foundation configuration: RES1, RES2, RES3.
Figure A-40 Four-story residential building roof configuration: RES1, RES2, RES3.
In the following sections, the structural designs for each of the retail,
industrial, and warehouse archetypes are documented, including plans,
elevations, and details. The floor and roof framing are identical for all
archetypes and are shown only once for the BOX1 model. The foundation
designs are unique, but similar, and are tabulated on a single plan. Masonry
reinforcement for all unique conditions are shown in wall elevation views.
A.5.1.1 Building Plans (BOX1, BOX2, BOX3)
This appendix presents numerical results generated for the calibration of the
simplified frame models of the six commercial building archetypes, COM1
through COM6, with the refined finite-element models, as discussed in
Chapter 4. The results include the response time histories and base shear-vs.-
first-story drift hysteresis curves obtained with the refined finite-element
models (using LS-DYNA) and simplified frame models (using OpenSees) for
three sets of ground-motion records selected in the calibration process. The
selected motions are Motion 1, Motion 2, and Motion 17 from the FEMA
P-695 database. For these analyses, each record set was scaled to two
intensity levels, the MCER level and a level beyond MCER. Deformed
meshes of the refined finite-element models are also presented to illustrate
the failure mechanisms of the archetypes.
As the following results show, for the multi-story archetypes, the upper-story
drift histories obtained with the refined and the simplified models do not
match well, especially for the four-story archetypes (COM3 and COM6).
This is attributed to the deficiency of the simplified models in capturing the
rigid-body rotation of the upper-story walls, which is the main source of
story drifts occurring in the upper stories. Nevertheless, the good match of
the base-shear time histories, the first-story drift histories, and the base shear-
vs.-first-story drift hysteresis curves obtained with the two modeling
approaches provides assurance that the incremental dynamic analysis results
obtained with the simplified models are reliable. That is because collapse
was always triggered by the story drifts concentrated in the first stories of the
archetypes.
Force - X (kips)
0 0
-1
-200
LS-DYNA
-2 OpenSEES
-400
st
-2 -1 0 1 2
1
0 5 10 15 20 25
st
time (sec) 1 Story Drift Ratio - X (%)
3 400
LS-DYNA
2 OpenSEES
200
Story Drift Ratio - Z (%)
Force - Z (kips)
0 0
-1
-200
LS-DYNA
-2
OpenSEES
-3 -400
st
-4 -2 0 2 4
1
0 5 10 15 20 25
st
time (sec) 1 Story Drift Ratio - Z (%)
Figure B-4 Comparison of response histories for the first story: COM1 –
Motion 1 (1.5×MCER).
25 400
LS-DYNA
20
OpenSEES
200
Story Drift Ratio - X (%)
15
Force - X (kips)
10
0
5
0 -200
LS-DYNA
-5 OpenSEES
-10 -400
st
-10 0 10 20
1
0 2 4 6 8 10 12
st
time (sec) 1 Story Drift Ratio - X (%)
20 400
LS-DYNA
15 OpenSEES
200
Story Drift Ratio - Z (%)
10
Force - Z (kips)
0
5
-200
0 LS-DYNA
OpenSEES
-5 -400
st
-5 0 5 10 15 20
1
0 1 2 3 4 5 6 7 8 9
st
time (sec) 1 Story Drift Ratio - Z (%)
Figure B-5 Comparison of response histories for the first story: COM1 –
Motion 1 (3.0×MCER).
Force - X (kips)
0 0
-200
LS-DYNA
OpenSEES
-0.5 -400
st
-0.5 0 0.5
1
0 5 10 15 20 25
st
time (sec) 1 Story Drift Ratio - X (%)
1 400
LS-DYNA
OpenSEES
0.5 200
Story Drift Ratio - Z (%)
Force - Z (kips)
0 0
-0.5 -200
LS-DYNA
OpenSEES
-1 -400
st
-1 -0.5 0 0.5 1
1
0 5 10 15 20 25
st
time (sec) 1 Story Drift Ratio - Z (%)
Figure B-6 Comparison of response histories for the first story: COM1 –
Motion 2 (1.5×MCER).
5 400
LS-DYNA LS-DYNA
OpenSEES OpenSEES
0 200
Story Drift Ratio - X (%)
Force - X (kips)
-5 0
-10 -200
-15 -400
st
-15 -10 -5 0 5
1
0 5 10 15 20 25
st
time (sec) 1 Story Drift Ratio - X (%)
4 400
LS-DYNA LS-DYNA
2 OpenSEES OpenSEES
200
Story Drift Ratio - Z (%)
0
Force - Z (kips)
0
-2
-200
-4
-6 -400
st
-6 -4 -2 0 2 4
1
0 5 10 15 20 25
st
time (sec) 1 Story Drift Ratio - Z (%)
Figure B-7 Comparison of response histories for the first story: COM1 –
Motion 2 (2.7×MCER).
Force - X (kips)
0 0
-1 -200
LS-DYNA
OpenSEES
-2 -400
st
-2 -1 0 1 2
1
0 5 10 15 20 25
st
time (sec) 1 Story Drift Ratio - X (%)
1 400
LS-DYNA
OpenSEES
0.5 200
Story Drift Ratio - Z (%)
Force - Z (kips)
0 0
-0.5 -200
LS-DYNA
OpenSEES
-1 -400
st
-1 -0.5 0 0.5 1
1
0 5 10 15 20 25
st
time (sec) 1 Story Drift Ratio - Z (%)
Figure B-8 Comparison of response histories for the first story: COM1 –
Motion 17 (1.5×MCER).
5 400
LS-DYNA LS-DYNA
OpenSEES OpenSEES
0 200
Story Drift Ratio - X (%)
Force - X (kips)
-5 0
-10 -200
-15 -400
st
-15 -10 -5 0 5
1
0 5 10 15 20 25
st
time (sec) 1 Story Drift Ratio - X (%)
4 400
LS-DYNA
OpenSEES
2 200
Story Drift Ratio - Z (%)
Force - Z (kips)
0 0
-2 -200
LS-DYNA
OpenSEES
-4 -400
st
-4 -2 0 2 4
1
0 5 10 15 20 25
st
time (sec) 1 Story Drift Ratio - Z (%)
Figure B-9 Comparison of response histories for the first story: COM1 –
Motion 17 (3×MCER).
Figure B-14 Comparison of response histories for the first story: COM2 –
Motion 1 (2×MCER).
600
0
400
Story Drift Ratio - X (%)
200
-10
Force - X (kips)
0
-20 -200
-400
-30 LS-DYNA LS-DYNA
-600
OpenSEES OpenSEES
st
-800
1
-40
0 1 2 3 4 5 6 7 8 -40 -30 -20 -10 0 10
st
time (sec) 1 Story Drift Ratio - X (%)
10 800
LS-DYNA 600
OpenSEES
400
Story Drift Ratio - Z (%)
5
200
Force - Z (kips)
-200
0
-400
LS-DYNA
-600
OpenSEES
st
-800
1
-5
-5 0 5 10
0 1 2 3 4 5 6 7 8
st
time (sec) 1 Story Drift Ratio - Z (%)
Figure B-16 Comparison of response histories for the first story: COM2 –
Motion 2 (2×MCER).
Figure B-18 Comparison of response histories for the first story: COM2 –
Motion 17 (2×MCER).
-0.5
-1
nd
2
-1.5
0 5 10 15 20 25
time (sec)
1.5
LS-DYNA
1 OpenSEES
Story Drift Ratio - Z (%)
0.5
-0.5
nd
2
-1
0 5 10 15 20 25
time (sec)
0
Story Drift Ratio - X (%)
-1
-2
-3
LS-DYNA
-4
OpenSEES
nd
-5
2
0 1 2 3 4 5 6 7 8
time (sec)
4
LS-DYNA
OpenSEES
2
Story Drift Ratio - Z (%)
-2
nd
2
-4
0 1 2 3 4 5 6 7 8
time (sec)
200
Base Shear - X (kips)
-200
-400
-600
-800
0 5 10 15 20 25
time (sec)
800
LS-DYNA
600
OpenSEES
400
200
Base Shear - X (kips)
-200
-400
-600
-800
0 5 10 15 20 25
time (sec)
200
Base Shear - X (kips)
-200
-400
-600
-800
0 1 2 3 4 5 6 7 8
time (sec)
800
LS-DYNA
600
OpenSEES
400
200
Base Shear - X (kips)
-200
-400
-600
-800
0 1 2 3 4 5 6 7 8
time (sec)
6
500
Story Drift Ratio - X (%)
Force - X (kips)
2 0
0
-500
LS-DYNA
-2 LS-DYNA
OpenSEES
OpenSEES
st
-1000
1
-4 -4 -2 0 2 4 6 8
0 5 10 15 20 25
st
time (sec) 1 Story Drift Ratio - X (%)
6 1200
800
4
Story Drift Ratio - Z (%)
400
2
Force - Z (kips)
0
0
-400
-800 LS-DYNA
-2 LS-DYNA
OpenSEES
OpenSEES
st
-1200
1
-4 -4 -2 0 2 4 6
0 5 10 15 20 25 st
1 Story Drift Ratio - Z (%)
time (sec)
Figure B-26 Comparison of response histories for the first story: COM3 –
Motion 1 (MCER).
20 1000
LS-DYNA LS-DYNA
15 OpenSEES OpenSEES
500
Story Drift Ratio - X (%)
10
Force - X (kips)
-500
0
st
-1000
1
-5
-5 0 5 10 15 20
0 1 2 3 4 5 6 7 8 9
st
time (sec) 1 Story Drift Ratio - X (%)
15 1200
LS-DYNA
800
OpenSEES
10
Story Drift Ratio - Z (%)
400
Force - Z (kips)
5 0
-400
0
-800 LS-DYNA
OpenSEES
st
-1200
1
-5
0 1 2 3 4 5 6 7 8 9 -5 0 5 10 15
st
time (sec) 1 Story Drift Ratio - Z (%)
Figure B-27 Comparison of response histories for the first story: COM3 –
Motion 1 (2×MCER).
Force - X (kips)
-2 0
-3
-500
-4 LS-DYNA
OpenSEES
st
-1000
1
-5 -5 -4 -3 -2 -1 0 1
0 5 10 15 20 25
st
time (sec) 1 Story Drift Ratio - X (%)
2 1200
LS-DYNA
800
OpenSEES
1
400
Story Drift Ratio - Z (%)
Force - Z (kips)
0
0
-400
-1
-800 LS-DYNA
OpenSEES
st
-1200
1
-2 -2 -1 0 1 2
0 5 10 15 20 25
st
time (sec) 1 Story Drift Ratio - Z (%)
Figure B-28 Comparison of response histories for the first story: COM3 –
Motion 2 (MCER).
5 1000
LS-DYNA
0 OpenSEES
500
Story Drift Ratio - X (%)
-5
Force - X (kips)
0
-10
-500
-15 LS-DYNA
OpenSEES
st
-1000
1
-20
0 1 2 3 4 5 6 -20 -15 -10 -5 0 5
st
time (sec) 1 Story Drift Ratio - X (%)
4 1200
LS-DYNA
OpenSEES 800
3
Story Drift Ratio - Z (%)
400
2
Force - Z (kips)
1
-400
0 -800 LS-DYNA
OpenSEES
st
-1200
1
-1
0 1 2 3 4 5 6 -1 0 1 2 3 4
st
time (sec) 1 Story Drift Ratio - Z (%)
Figure B-29 Comparison of response histories for the first story: COM3 –
Motion 2 (1.8×MCER).
Force - X (kips)
0
-1
-500
-2
st
-1000
1
-3
0 5 10 15 20 25 -3 -2 -1 0 1 2
st
time (sec) 1 Story Drift Ratio - X (%)
1 1200
LS-DYNA LS-DYNA
OpenSEES 800 OpenSEES
0.5
Story Drift Ratio - Z (%)
400
Force - Z (kips)
0 0
-400
-0.5
-800
st
-1200
1
-1
-1 -0.5 0 0.5 1
0 5 10 15 20 25
st
time (sec) 1 Story Drift Ratio - Z (%)
Figure B-30 Comparison of response histories for the first story: COM3 –
Motion 17 (MCER).
10 1200
LS-DYNA
5 800 OpenSEES
Story Drift Ratio - X (%)
0 400
Force - X (kips)
-5 0
-10 -400
-1200
1
-20
0 2 4 6 8 10 12 14 16 18 20 -20 -15 -10 -5 0 5 10
st
time (sec) 1 Story Drift Ratio - X (%)
6 1200
LS-DYNA
800
4 OpenSEES
Story Drift Ratio - Z (%)
400
2
Force - Z (kips)
0
-400
-2 -800 LS-DYNA
OpenSEES
st
-1200
1
-4 -4 -2 0 2 4 6
0 2 4 6 8 10 12 14 16 18 20
st
time (sec) 1 Story Drift Ratio - Z (%)
Figure B-31 Comparison of response histories for the first story: COM3 –
Motion 17 (2×MCER).
The nearly identical drift ratios for the second, third, and fourth stories of the
LS-DYNA model shown in Figure B-32 and Figure B-33 are due to the rigid-
body rotations of the walls above the first story. This was not captured by
the OpenSees model, as previously discussed.
0 0
-1 -1
nd
nd
2
-2 -2
2
0 5 10 15 20 25 0 5 10 15 20 25
time (sec) time (sec)
2 3
LS-DYNA LS-DYNA
OpenSEES 2 OpenSEES
1
Story Drift Ratio - X (%)
-1
-1
rd
rd
3
3
-2 -2
0 5 10 15 20 25 0 5 10 15 20 25
time (sec) time (sec)
2 3
LS-DYNA LS-DYNA
OpenSEES 2 OpenSEES
1
Story Drift Ratio - X (%)
1
0
0
-1
-1
th
th
4
4
-2 -2
0 5 10 15 20 25 0 5 10 15 20 25
time (sec) time (sec)
Figure B-32 Comparison of story-drift histories for the upper stories: COM3 – Motion 1 (MCER).
2
1
1
0
0
-1
-1
nd
nd
2
-2 -2
2
0 1 2 3 4 5 6 7 8 0 1 2 3 4 5 6 7 8
LS-DYNA LS-DYNA
3 OpenSEES OpenSEES
2
2
1
1
0
0
-1
-1
rd
rd
3
3
-2 -2
0 1 2 3 4 5 6 7 8 0 1 2 3 4 5 6 7 8
time (sec) time (sec)
4 3
LS-DYNA LS-DYNA
3 OpenSEES OpenSEES
2
Story Drift Ratio - X (%)
0
0
-1
-1
th
th
4
-2 -2
0 1 2 3 4 5 6 7 8 0 1 2 3 4 5 6 7 8
time (sec) time (sec)
Figure B-33 Comparison of story-drift histories for the upper stories: COM3 – Motion 1 (2×MCER).
1000
LS-DYNA
OpenSEES
500
Base Shear - X (kips)
-500
-1000
0 5 10 15 20 25
time (sec)
1200
LS-DYNA
800 OpenSEES
400
Base Shear - Z (kips)
-400
-800
-1200
0 5 10 15 20 25
time (sec)
-500
-1000
0 1 2 3 4 5 6 7 8
time (sec)
1200
800
400
Base Shear - Z (kips)
-400
-800 LS-DYNA
OpenSEES
-1200
0 1 2 3 4 5 6 7 8
time (sec)
Force - X (kips)
0 0
-2 -200
LS-DYNA
OpenSEES
-4 -400
st
-4 -2 0 2 4
1
0 5 10 15 20 25
st
time (sec) 1 Story Drift Ratio - X (%)
2 400
LS-DYNA LS-DYNA
OpenSEES OpenSEES
0 200
Story Drift Ratio - Z (%)
Force - Z (kips)
-2 0
-4 -200
-6 -400
st
-6 -4 -2 0 2
1
0 5 10 15 20 25
st
time (sec) 1 Story Drift Ratio - Z (%)
Figure B-39 Comparison of response histories for the first story: COM4 –
Motion 1 (MCER).
10 OpenSEES
200
Force - X (kips)
0
0
-200
-5 LS-DYNA
OpenSEES
-10 -400
st
-10 -5 0 5 10 15
1
0 2 4 6 8 10 12
st
time (sec) 1 Story Drift Ratio - X (%)
20 500
LS-DYNA
15
OpenSEES
250
Story Drift Ratio - Z (%) 10
Force - Z (kips)
5
0
0
-5 -250
LS-DYNA
-10 OpenSEES
-15 -500
st
-10 0 10 20
1
0 1 2 3 4 5 6 7 8 9
st
time (sec) 1 Story Drift Ratio - Z (%)
Figure B-40 Comparison of response histories for the first story: COM4 –
Motion 1 (2×MCER).
0.6 400
LS-DYNA
0.4 OpenSEES
200
Story Drift Ratio - X (%)
0.2
Force - X (kips)
0 0
-0.2
-200
LS-DYNA
-0.4
OpenSEES
-0.6 -400
st
0 5 10 15 20 25
st
time (sec) 1 Story Drift Ratio - X (%)
1 400
LS-DYNA
OpenSEES
0.5 200
Story Drift Ratio - Z (%)
Force - Z (kips)
0 0
-0.5 -200
LS-DYNA
OpenSEES
-1 -400
st
-1 -0.5 0 0.5 1
1
0 5 10 15 20 25
st
time (sec) 1 Story Drift Ratio - Z (%)
Figure B-41 Comparison of response histories for the first story: COM4 –
Motion 2 (MCER).
5 OpenSEES OpenSEES
200
Story Drift Ratio - X (%)
Force - X (kips)
0
-5
-200
-10
-15 -400
st
-15 -10 -5 0 5 10
1
0 5 10 15 20 25
st
time (sec) 1 Story Drift Ratio - X (%)
5 400
LS-DYNA LS-DYNA
OpenSEES OpenSEES
0 200
Story Drift Ratio - Z (%)
Force - Z (kips)
-5 0
-10 -200
-15 -400
st
-15 -10 -5 0 5
1
0 5 10 15 20 25
st
time (sec) 1 Story Drift Ratio - Z (%)
Figure B-42 Comparison of response histories for the first story: COM4 –
Motion 2 (1.8×MCER).
2 400
LS-DYNA
OpenSEES
1 200
Story Drift Ratio - X (%)
Force - X (kips)
0 0
-1 -200
LS-DYNA
OpenSEES
-2 -400
st
-2 -1 0 1 2
1
0 5 10 15 20 25
st
time (sec) 1 Story Drift Ratio - X (%)
1 400
LS-DYNA LS-DYNA
0
Force - Z (kips)
0
-0.5
-200
-1
-1.5 -400
st
0 5 10 15 20 25
st
time (sec) 1 Story Drift Ratio - Z (%)
Figure B-43 Comparison of response histories for the first story: COM4 –
Motion 17 (MCER).
Force - X (kips)
0
-5
-200
-10 -400
st
-10 -5 0 5
1
0 5 10 15 20 25
st
time (sec) 1 Story Drift Ratio - X (%)
4 400
LS-DYNA LS-DYNA
2 OpenSEES OpenSEES
200
Story Drift Ratio - Z (%)
0
Force - Z (kips)
-2 0
-4
-200
-6
-8 -400
st
-8 -6 -4 -2 0 2 4
1
0 5 10 15 20 25
st
time (sec) 1 Story Drift Ratio - Z (%)
Figure B-44 Comparison of response histories for the first story: COM4 –
Motion 17 (2×MCER).
200
Force - X (kips)
5 0
-200
0 -400
LS-DYNA
-600
OpenSEES
st
-800
1
-5
-5 0 5 10 15
0 5 10 15 20 25
st
time (sec) 1 Story Drift Ratio - X (%)
10 800
LS-DYNA 600
8
OpenSEES
6 400
Story Drift Ratio - Z (%)
4 200
Force - Z (kips)
2 0
0 -200
-2 -400
LS-DYNA
-4 -600
OpenSEES
st
-800
1
-6
-6 -4 -2 0 2 4 6 8 10
0 5 10 15 20 25
st
time (sec) 1 Story Drift Ratio - Z (%)
Figure B-48 Comparison of response histories for the first story: COM5 –
Motion 1 (MCER).
25
800
LS-DYNA 600
20
OpenSEES
400
15
Story Drift Ratio - X (%)
200
Force - X (kips)
10
0
5
-200
0 -400
LS-DYNA
-5 -600
OpenSEES
st
-800
1
-10
-10 -5 0 5 10 15 20 25
0 1 2 3 4 5 6 7 8 9 10
st
time (sec) 1 Story Drift Ratio - X (%)
15 800
LS-DYNA 600
10 OpenSEES
400
Story Drift Ratio - Z (%)
200
5
Force - Z (kips)
0 -200
-400
-5 LS-DYNA
-600
OpenSEES
st
-800
1
-10
0 1 2 3 4 5 6 7 8 9 10 -10 -5 0 5 10 15
st
time (sec) 1 Story Drift Ratio - Z (%)
Figure B-49 Comparison of response histories for the first story: COM5 –
Motion 1 (2×MCER).
Force - X (kips)
0
-2 -200
-400
-4
-600
st
-800
1
-6
-6 -4 -2 0 2 4
0 5 10 15 20 25
st
time (sec) 1 Story Drift Ratio - X (%)
4 800
LS-DYNA LS-DYNA
600
2 OpenSEES OpenSEES
400
Story Drift Ratio - Z (%)
200
0
Force - Z (kips)
0
-2 -200
-400
-4
-600
st
-800
1
-6
0 5 10 15 20 25 -6 -4 -2 0 2 4
st
time (sec) 1 Story Drift Ratio - Z (%)
Figure B-50 Comparison of response histories for the first story: COM5 –
Motion 2 (MCER).
10 800
600
0 400
Story Drift Ratio - X (%)
200
Force - X (kips)
-10 0
-200
-20 -400
LS-DYNA LS-DYNA
-600
OpenSEES OpenSEES
st
-800
1
-30
0 1 2 3 4 5 6 -30 -20 -10 0 10
st
time (sec) 1 Story Drift Ratio - X (%)
10 800
LS-DYNA 600
OpenSEES
400
Story Drift Ratio - Z (%)
5 200
Force - Z (kips)
-200
0
-400
LS-DYNA
-600
OpenSEES
st
-800
1
-5
-5 0 5 10
0 1 2 3 4 5 6
st
time (sec) 1 Story Drift Ratio - Z (%)
Figure B-51 Comparison of response histories for the first story: COM5 –
Motion 2 (1.9×MCER).
1
200
Force - X (kips)
0 0
-200
-1
-400
LS-DYNA
-2 -600
OpenSEES
st
-800
1
-3
-3 -2 -1 0 1 2 3
0 5 10 15 20 25
st
time (sec) 1 Story Drift Ratio - X (%)
2 800
LS-DYNA 600
OpenSEES
1 400
Story Drift Ratio - Z (%)
200
Force - Z (kips)
0 0
-200
-1 -400
LS-DYNA
-600
OpenSEES
st
-800
1
-2
-2 -1 0 1 2
0 5 10 15 20 25
st
time (sec) 1 Story Drift Ratio - Z (%)
Figure B-52 Comparison of response histories for the first story: COM5 –
Motion 17 (MCER).
10 800
LS-DYNA 600
OpenSEES
5 400
Story Drift Ratio - X (%)
200
Force - X (kips)
0 0
-200
-5 -400
LS-DYNA
-600
OpenSEES
st
-800
1
-10
0 2 4 6 8 10 12 -10 -5 0 5 10
st
time (sec) 1 Story Drift Ratio - X (%)
5 800
LS-DYNA 600
0
OpenSEES
400
-5
Story Drift Ratio - Z (%)
200
Force - Z (kips)
-10
0
-15
-200
-20 -400
LS-DYNA
-25 -600
OpenSEES
st
-800
1
-30
0 2 4 6 8 10 12 -30 -25 -20 -15 -10 -5 0 5
st
time (sec) 1 Story Drift Ratio - Z (%)
Figure B-53 Comparison of response histories for the first story: COM5 –
Motion 17 (2×MCER).
-1
-2
nd
-3
2
0 5 10 15 20 25
time (sec)
2
LS-DYNA
1
OpenSEES
0
Story Drift Ratio - Z (%)
-1
-2
-3
-4
nd
2
-5
0 5 10 15 20 25
time (sec)
-1
-2
nd
-3
2
0 1 2 3 4 5 6 7 8 9
time (sec)
4
LS-DYNA
OpenSEES
2
Story Drift Ratio - Z (%)
-2
nd
2
-4
0 1 2 3 4 5 6 7 8 9
time (sec)
200
Base Shear - X (kips)
-200
-400
-600
-800
0 5 10 15 20 25
time (sec)
800
LS-DYNA
600
OpenSEES
400
200
Base Shear - X (kips)
-200
-400
-600
-800
0 5 10 15 20 25
time (sec)
200
Base Shear - X (kips)
-200
-400
-600
-800
0 1 2 3 4 5 6 7 8 9
time (sec)
800
LS-DYNA
600
OpenSEES
400
200
Base Shear - X (kips)
-200
-400
-600
-800
0 1 2 3 4 5 6 7 8 9
time (sec)
Force - X (kips)
0
-5
-1000
LS-DYNA
OpenSEES
st
-2000
1
-10
0 2 4 6 8 10 12 -10 -5 0 5
st
time (sec) 1 Story Drift Ratio - X (%)
20 2000
LS-DYNA LS-DYNA
15 OpenSEES OpenSEES
1000
Story Drift Ratio - Z (%)
10
Force - Z (kips)
0
-1000
0
st
-2000
1
-5
-5 0 5 10 15 20
0 2 4 6 8 10 12
st
time (sec) 1 Story Drift Ratio - Z (%)
Figure B-61 Comparison of response histories for the first story: COM6 –
Motion 1 (MCER).
40 2000
LS-DYNA LS-DYNA
30 OpenSEES OpenSEES
1000
Story Drift Ratio - X (%)
20
Force - X (kips)
10
-1000
0
st
-2000
1
-10
0 1 2 3 4 5 6 7 8 9 10 -10 0 10 20 30 40
st
time (sec) 1 Story Drift Ratio - X (%)
25 2000
LS-DYNA LS-DYNA
20
OpenSEES OpenSEES
1000
15
Story Drift Ratio - Z (%)
Force - Z (kips)
10
0
5
0 -1000
-5
st
-2000
1
-10
0 1 2 3 4 5 6 7 8 9 10 -10 -5 0 5 10 15 20 25
st
time (sec) 1 Story Drift Ratio - Z (%)
Figure B-62 Comparison of response histories for the first story: COM6 –
Motion 1 (2×MCER).
Force - X (kips)
-1 0
-2
-1000
-3
st
-2000
1
-4
0 5 10 15 20 25 -4 -2 0 2
st
time (sec) 1 Story Drift Ratio - X (%)
4 2000
LS-DYNA
OpenSEES
Story Drift Ratio - Z (%) 2 1000
Force - Z (kips)
0 0
-2 -1000
LS-DYNA
OpenSEES
st
-2000
1
-4
-4 -2 0 2 4
0 5 10 15 20 25
st
time (sec) 1 Story Drift Ratio - Z (%)
Figure B-63 Comparison of response histories for the first story: COM6 –
Motion 2 (MCER).
10 2000
LS-DYNA
OpenSEES
0
1000
Story Drift Ratio - X (%)
-10
Force - X (kips)
0
-20
-1000
-30 LS-DYNA
OpenSEES
st
-2000
1
-40
0 1 2 3 4 5 6 -40 -30 -20 -10 0 10
st
time (sec) 1 Story Drift Ratio - X (%)
15 2000
LS-DYNA LS-DYNA
OpenSEES OpenSEES
10 1000
Story Drift Ratio - Z (%)
Force - Z (kips)
5 0
0 -1000
st
-2000
1
-5
-5 0 5 10 15
0 1 2 3 4 5 6
st
time (sec) 1 Story Drift Ratio - Z (%)
Figure B-64 Comparison of response histories for the first story: COM6 –
Motion 2 (2×MCER).
Force - X (kips)
-1 0
-2
-1000
-3
st
-2000
1
-4
0 5 10 15 20 25 -4 -2 0 2
st
time (sec) 1 Story Drift Ratio - X (%)
2 2000
LS-DYNA LS-DYNA
1.5
OpenSEES OpenSEES
1000
1
Story Drift Ratio - Z (%)
Force - Z (kips)
0.5
0
0
-0.5 -1000
-1
st
-2000
1
-1.5
-2 -1 0 1 2
0 5 10 15 20 25
st
time (sec) 1 Story Drift Ratio - Z (%)
Figure B-65 Comparison of response histories for the first story: COM6 –
Motion 17 (MCER).
5 2500
LS-DYNA LS-DYNA
-5 500
Force - X (kips)
-10 -500
-15 -1500
st
-2500
1
-20
0 2 4 6 8 10 12 14 -20 -15 -10 -5 0 5
st
time (sec) 1 Story Drift Ratio - X (%)
10 2000
LS-DYNA LS-DYNA
OpenSEES OpenSEES
5 1000
Story Drift Ratio - Z (%)
Force - Z (kips)
0 0
-5 -1000
st
-2000
1
-10
0 2 4 6 8 10 12 14 -10 -5 0 5 10
st
time (sec) 1 Story Drift Ratio - Z (%)
Figure B-66 Comparison of response histories for the first story: COM6 –
Motion 17 (2×MCER).
The nearly identical drift ratios for the second, third, and fourth stories of the
LS-DYNA model shown in Figure B-67 and Figure B-68 are due to the rigid-
body rotations of the walls above the first story. This was not captured by
the OpenSees model, as previously discussed.
-1
0
-2
-3
-2
LS-DYNA
-4
OpenSEES
nd
nd
-4
2
2
-5
0 1 2 3 4 5 6 7 8 9 10 0 1 2 3 4 5 6 7 8 9 10
time (sec) 2 time (sec)
4
LS-DYNA
1
OpenSEES
2 0
-1
0
-2
-3
-2
LS-DYNA
-4
OpenSEES
rd
rd
3
-5
3
-4
0 1 2 3 4 5 6 7 8 9 10 0 1 2 3 4 5 6 7 8 9 10
time (sec) time (sec)
4 2
LS-DYNA
1
OpenSEES
2 0
Story Drift Ratio - Z (%)
Story Drift Ratio - X (%)
-1
0
-2
-3
-2
LS-DYNA
-4
OpenSEES
th
th
-5
4
-4
0 1 2 3 4 5 6 7 8 9 10 0 1 2 3 4 5 6 7 8 9 10
time (sec) time (sec)
Figure B-67 Comparison of story-drift histories for the upper stories: COM6 – Motion 1 (MCER).
0 -5
-2 LS-DYNA
OpenSEES
nd
nd
2
-4
2
-10
0 1 2 3 4 5 6 7 8 9 10 0 1 2 3 4 5 6 7 8 9 10
0 -5
-2 LS-DYNA
OpenSEES
rd
rd
3
3
-4 -10
0 1 2 3 4 5 6 7 8 9 10 0 1 2 3 4 5 6 7 8 9 10
time (sec) time (sec)
8 6
LS-DYNA
4
6 OpenSEES
Story Drift Ratio - Z (%) 2
Story Drift Ratio - X (%)
4
0
2
-2
0
-4
-2 LS-DYNA
-6
OpenSEES
th
th
-8
4
-4
0 1 2 3 4 5 6 7 8 9 10 0 1 2 3 4 5 6 7 8 9 10
time (sec) time (sec)
Figure B-68 Comparison of story-drift histories for the upper stories: COM6 – Motion 1 (2×MCER).
2000
1000
Base Shear - X (kips)
-1000
LS-DYNA
OpenSEES
-2000
0 1 2 3 4 5 6 7 8 9 10
time (sec)
2000
1000
Base Shear - Z (kips)
-1000
LS-DYNA
OpenSEES
-2000
0 1 2 3 4 5 6 7 8 9 10
time (sec)
-1000
-2000
0 1 2 3 4 5 6 7 8 9 10
time (sec)
2000
1000
Base Shear - Z (kips)
-1000
LS-DYNA
OpenSEES
-2000
0 1 2 3 4 5 6 7 8 9 10
time (sec)
C.1 Introduction
The default case for the analysis of structural response is the fixed-base
condition, where the foundation and supporting soil are assumed to be
perfectly rigid. The flexible-base condition includes the deformation of the
FEMA P-2139-3 C: Development of Soil Springs, Soil Dampers, and Frequency- C-1
Modified Ground Motion Records for the SSI Parametric Study
soil and foundation. The soil flexibility increases the overall flexibility of the
building, resulting in a lengthening of the fundamental period. As is well
known from spectral acceleration diagrams, the energy content of an
earthquake varies with frequency. Significant period lengthening will impact
the seismic demand on the structure, as spectral acceleration is often used to
estimate base shear for an elastic response of a structure. In general, for
structures with short fundamental periods, period lengthening will cause an
increase in spectral acceleration, whereas for structures with long periods,
spectral acceleration decreases with period lengthening. The change in the
flexibility of the system can also change the load and deformation
distribution within the structure and foundation, the effects of which are
often complicated.
C-2 C: Development of Soil Springs, Soil Dampers, and Frequency- FEMA P-2139-3
Modified Ground Motion Records for the SSI Parametric Study
component due to natural seismic wave-path complexities. Kinematic
interaction also includes effects due to foundation embedment, which are not
discussed here in detail due to the shallow foundation depths of the building
archetypes selected for this study.
Two soil sites with differing soil conditions were selected for the SSI study.
The sites vary in terms of site class, which was classified based on the shear-
wave velocity in the upper 30 meters (Vs30), per FEMA P-1050. The soft site
was selected to be near the site class D/E boundary, with a target Vs30 = 180
m/s (591 ft/s). The stiff site was selected to be near the site class C/D
boundary, with a target Vs30 = 360 m/s (1181 ft/s). More information about
the conditions at the selected sites is given below.
C.3.1 Soft Site
The soft site was selected to be near the site class D/E boundary, with a
target Vs30 = 180 m/s (591 ft/s). The Imperial Valley Wildlife Liquefaction
(WLA) site was selected as a representative site. The site consists of
unsaturated soft clay near the soil surface. The shear-wave-velocity profile is
documented in Star et al. (2015). The estimated Vs30 at WLA is 183 m/s (600
ft/s). A simplified assumed shear-wave profile of the site is taken to be Vs
= 100 m/s (328 ft/s) from surface to 3 m (10 ft) below the surface, Vs = 140
m/s (459 ft/s) from 3 m to 5 m (10 ft to 16 ft) below the surface, Vs = 170 m/s
(558 ft/s) from 5 m to 12 m (16 ft to 39 ft) below the surface, and Vs = 230
m/s (755 ft/s) more than 12 m (39 ft) below the surface. Undrained shear
strength of the soil was estimated as 32 kPa (668 lbs/ft2). Poisson’s ratio for
the soil was estimated to be 0.45, and the unit weight was estimated to be
16 kN/m3 (103 lbs/ft3). At large strains, the shear modulus of the soil reduces
from the maximum elastic value. The effective reduction factor for shear
modulus was estimated to be 0.6.
C.3.2 Stiff Site
The stiff soil site was selected to be near the site class C/D boundary, with a
target Vs30 = 360 m/s (1181 ft/s). The Cholame Parkfield 12W site was
selected as a representative site. The site consists of dense sands near the soil
surface. The shear-wave-velocity profile is documented in Thompson et al.
FEMA P-2139-3 C: Development of Soil Springs, Soil Dampers, and Frequency- C-3
Modified Ground Motion Records for the SSI Parametric Study
(2010). Based on shear-wave velocity data, Vs30 is estimated to be 354 m/s
(1161 ft/s). A simplified, assumed shear-wave profile of the site is taken to
be Vs = 295 m/s (968 ft/s) from surface to 1 m (3.3 ft) below the surface,
Vs = 310 m/s (1017 ft/s) from 1 m to 2 m (3.3 ft to 6.6 ft) below the surface,
Vs = 315 m/s (1033 ft/s) from 2 m to 10 m (6.6 ft to 33 ft) below the surface,
Vs = 335 m/s (1099 ft/s) from 10 m to 15 m (33 ft to 49 ft) below the surface,
Vs = 370 m/s (1214 ft/s) from 15 m to 20 m (49 ft to 66 ft) below the surface,
and Vs = 415 m/s (1362 ft/s) more than 20 m (66 ft) below the surface. The
friction angle of the soil was estimated to be 40 degrees, Poisson’s ratio was
estimated to be 0.35, and the unit weight of the soil was estimated to be 19
kN/m3 (122 lbs/ft3). At large strains, the shear modulus of the soil reduces
from the maximum elastic value. The effective reduction factor for shear
modulus was estimated to be 0.8.
Figure C-1(b) shows the layout of the footing structures in the COM2B
design for stiff soil. For the stiff site configuration, springs and dampers
were developed for five different foundation types: under the L-shaped
corner foundations, under the rectangular wall foundations, under the
external-wall isolated footings, under the interior isolated footings, and under
the grade beams connecting the external perimeter foundations. The soil-
foundation springs and dampers were not included for the slab-on-grade.
For all foundations located along the exterior of the structure (i.e., the long
and short exterior strip foundations for the soft site, and the L-shaped corner
foundations, rectangular wall foundations, external isolated footings, and the
grade beams for the stiff site), springs and dampers were developed at nodes
located at a 1-foot spacing along the center line of each foundation. Three
springs and parallel dampers control translation in the two horizontal
C-4 C: Development of Soil Springs, Soil Dampers, and Frequency- FEMA P-2139-3
Modified Ground Motion Records for the SSI Parametric Study
(subscripts x, z) direction and one vertical (subscript y) direction. Rotation
of the foundation in the in-plane direction is controlled by the vertical
springs. Rotational springs (notated with the subscripts xx and zz) control
the out-of-plane rotation. Figure C-2 shows a cross section through a single
node of a foundation, with the vertical and horizontal translations and out-of-
plane rotational springs and dampers shown. The spring controlling
translation along the length of the foundation is not shown.
For each interior footing, the foundation spring stiffness and damping values
were developed as lumped impedance values, located at a single node under
the center of the square foundations. Five springs and parallel dampers
control translation in two horizontal (subscripts x, z) directions and one
vertical (subscript y) direction and rotations about the x and z axes (notated
with the subscripts xx and zz).
(a)
(b)
Figure C-1 Layout of COM2B modeled footings for SSI study: (a) soft site
and (b) stiff site.
FEMA P-2139-3 C: Development of Soil Springs, Soil Dampers, and Frequency- C-5
Modified Ground Motion Records for the SSI Parametric Study
Figure C-2 Translational and rotational springs and dampers.
Interior Isolated 9’ x 9’
Interior Isolated 6’ x 6’
The average effective shear-wave velocity of the soil under each foundation
is used, along with the density of the soil, to estimate the small strain shear
modulus, G0. As described in NIST (2012b), the effective depth interval for
calculating the shear-wave velocity depends on the size of the foundation and
the spring orientation. For the translations and vertical springs, the effective
C-6 C: Development of Soil Springs, Soil Dampers, and Frequency- FEMA P-2139-3
Modified Ground Motion Records for the SSI Parametric Study
depth is equal to the half-dimension of an equivalent square foundation with
an area matching the actual foundation. For the rotational springs, it is equal
to the half-dimension of an equivalent square foundation matching the
moment of inertia of the actual foundation. The small-strain shear modulus
is reduced by the strain reduction factors described in appendix Section C.3
to get a modulus appropriate for estimating springs and dampers for each
footing type.
The impedance springs and dampers calculated using Pais and Kausel (1988)
equations, as given in Table 2-2a of NIST (2012b), are frequency dependent.
However due to the limitations of the modeling software, implementing
frequency-dependent springs and dampers was not possible. Following
NIST (2012b), a single frequency value was selected corresponding to the
first-mode, flexible-base period of the system of approximately 0.3 seconds.
FEMA P-2139-3 C: Development of Soil Springs, Soil Dampers, and Frequency- C-7
Modified Ground Motion Records for the SSI Parametric Study
As shown in Figure C-3, the capacity of the foundation in rotation is
controlled primarily by the uplift of the foundation. The out-of-plane
rotational springs capacity, Koop,capacity, can be estimated using static
equilibrium to estimate the maximum moment that the foundation could
support knowing the average static vertical load on each foundation and the
bearing capacity. In order to account for the reduction in vertical loads due
to general rocking of the structure, the rotational capacities were reduced by
50 percent. No strain hardening was included.
Table C-3 Spring and Damper Properties Per Node for COM2B Soft Site
Footing
Type Translation-X Translation-Z Vertical-Y Rotation-ZZ Rotation-XX
Long Strip Kx = 1.5x102 kips/ft Kz = 2.0x102 kips/ft Ky = 3.0x102 kips/ft - Kxx = 3.3 x103 kip-
Kx,capacity = 2.7 kips Kz,capacity = 2.7 kips Ky,capacity = 16.3 kips ft/rad
βx = 17.0% βz = 12.8% βy = 7.9% Kxx,capacity = 2.6 kip-ft
βxx = 0.03%
Short Strip Kx = 2.5x102 kips/ft Kz = 2.0x102 kips/ft Ky = 4.6x102 kips/ft Kzz = 2.8x103 kip- -
Kx,capacity = 2.7 kips Kz,capacity = 2.7 kips Ky,capacity = 16.3 kips ft/rad
βx = 10.5% βz = 13.0% βy = 2.6% Kzz,capacity = 2.6 kip-ft
βzz = 0.03%
Interior Kx = 5.5x103 kips/ft Kz = 5.5x103 kips/ft Ky = 7.7x103 kips/ft Kzz = 1.3x105 kip- Kxx = 1.3x105 kip-
Isolated Kx,capacity = 54 kips Kz,capacity = 54 kips Ky,capacity = 354 kips ft/rad ft/rad
βx = 9.7% βz = 9.7% βy = 17.2% βzz = 0.3% βxx = 0.3%
C-8 C: Development of Soil Springs, Soil Dampers, and Frequency- FEMA P-2139-3
Modified Ground Motion Records for the SSI Parametric Study
Table C-4 Spring and Damper Properties Per Node for COM2B Stiff Site
Footing
Type Translation-X Translation-Z Vertical-Y Rotation-ZZ Rotation-XX
Wall L- Kx = 4.3x103 kips/ft Kz = 4.2x103 kips/ft Ky = 8.2x103 kips/ft - -
Shaped βx = 4.0% βz = 4.1% Ky,capacity = 210 kips
βy = 0.03%
Wall Kx = 4.1x103 kips/ft Kz = 4.6x103 kips/ft Ky = 1.1x104 kips/ft - Kxx = 1.7x104 kips-
Rectangular βx = 2.2% βz = 2.0% Ky,capacity = 118 kips ft/rad
βy = 0.01% Kxx,capacity = 3.7 kip-ft
βxx = 0.00%
Exterior Kx = 8.8x103 kips/ft Kz = 8.8x103 kips/ft Ky = 2.9x104 kips/ft - Kxx = 4.9x104 kips-
Isolated βx = 1.8% βz = 1.8% Ky,capacity = 249 kips ft/rad
(x and z βy = 0.00% Kxx,capacity = 8.8 kip-ft
given for βxx = 0.00%
footings on
long wall.
Reverse for
short wall)
Interior Kx = 4.7x104 kips/ft Kz = 4.7x104 kips/ft Ky = 6.1x104 kips/ft Kzz = 4.7x105 kips- Kxx = 4.7x105 kips-
Isolated βx = 2.3% βz = 2.3% Ky,capacity = 1960 kips ft/rad ft/rad
βy = 3.7% βzz = 0.00% βxx = 0.00%
Under Kx = 2.9x103 kips/ft Kz = 3.5x103 kips/ft Ky = 4.2x103 kips/ft - Kxx = 4.7x105 kips-
Grade βx = 1.6% βz = 1.3% Ky,capacity = 47 kips ft/rad
Beam Kxx,capacity = 0.9 kip-ft
βy = 2.3%
(x and z βxx = 0.0%
given for
footings on
long wall.
Reverse for
short wall)
The soil-foundation spring and damper values are calculated individually for
each foundation configuration at each soil site. In the following calculations,
the COM2B stiff site rectangular wall foundation was selected as an example
case, with calculations necessary to compute the stiffness and damping
parameters presented.
FEMA P-2139-3 C: Development of Soil Springs, Soil Dampers, and Frequency- C-9
Modified Ground Motion Records for the SSI Parametric Study
Horizontal (x,z): zpA = A / 4 = 2.98 ft
Rotation (zz): zp,zzI = 4 I z = 4.23 ft
The product of the density, ρ, and the square of the shear-wave velocity gives
the small strain shear modulus. At large strains, the effective shear modulus,
Geff, of the soil reduces from the small strain value. The effective reduction
factor for shear modulus was estimated to be 0.8.
Geff, horiz = 0.8 × ρ × Vs2,avg ,horiz = 2.82 ×106 lb/ft2
C-10 C: Development of Soil Springs, Soil Dampers, and Frequency- FEMA P-2139-3
Modified Ground Motion Records for the SSI Parametric Study
Ky,static = 6.7×104 kips/ft
Dynamic spring modifiers (α) and damper values (β) are frequency
dependent. First, the dimensionless frequency is calculated following NIST
(2012b) Equation 2-15:
ao ω=
= B 0.03
Vs
A single value of α and β is then calculated for each of the five degrees of
freedom using Pais and Kausel (1988) equations, as given in Table 2-3a of
NIST (2012b). The input parameters for these calculations are the
dimensionless frequency, the foundation dimensions L and B, the static
spring stiffnesses determined in Step 2 of these example calculation, the
shear moduli determined in Step 1 of these example calculations, and
Poisson’s ratio of the soil. The calculated dynamic spring modifiers and
damper values are:
αx = 1.00 βx = 2.24%
αz = 1.00 βz = 1.99%
αzz = 1.00 βzz = 0.007%
αxx = 1.00 βxx = 0%
αy = 1.00 βy = 3.39%
The modifiers are multiplied by the static stiffnesses to get the final stiffness
values:
FEMA P-2139-3 C: Development of Soil Springs, Soil Dampers, and Frequency- C-11
Modified Ground Motion Records for the SSI Parametric Study
4. Determine the spring capacities
Kz,capacity = V × tanϕ′
K y ,capacity=
Qult
= Dγ N q Fqs Fqd +
( 2B ) γ N F F × L × B / FS = 1.4 ×103 kips
γ γs γd
FS 2
The rotational spring capacity was estimated using static equilibrium. The
vertical force on the foundation is assumed to be 61 kips. With the
foundation rotated so that the majority of the footing experiencing uplift and
a sliver at the edge of the footing at the bearing capacity, the moment at static
equilibrium can be calculated as the vertical force, V, times the moment arm,
a, between the vertical force assumed to be at the center of the foundation
and the point of rotation at the edge of the foundation. The value is reduced
by 50 percent to account for redistribution of the vertical load.
The nodes are located every 1 foot along the centerline of the foundation.
That means that there are 12 nodes along the 12-foot-long foundation. The
calculated foundation stiffnesses and capacities are divided by the number of
nodes to get the nodal spring value. The damping ratios do not need to be
reduced based on the number of nodes. The rotational stiffness, Kzz, is taken
as null because the distributed vertical spring stiffness, Ky, will provide the
stiffness against rotation. The vertical nodal spring stiffness and damping are
modified by a stiffness factor, Rk,zz, and a damping factor, Rc,zz, to produce
C-12 C: Development of Soil Springs, Soil Dampers, and Frequency- FEMA P-2139-3
Modified Ground Motion Records for the SSI Parametric Study
values equivalent to the rotational impedances. Equation 2-20 and Equation
2-21 from NIST (2012b) are combined to obtain:
3K zz
Rk ,zz = = 1.88
K y L2
3czz
Rc ,zz = = 0.0019
c y L2 Rk ,zz
Per node, the updated spring stiffness and damping properties are:
Kx = Kx/N = 4.1×103 kip/ft
Kz = Kz/N = 4.6×103 kip/ft
Kzz = —
Kxx = Kxx/N = 1.74×104 kip-ft/rad
Ky = Ky × Rk,zz/N = 1.1×104 kip/ft
Kxx,capacity = Kxx,capacity/N = 3.7 kip-ft
Ky,capacity = Ky,capacity/N = 118 kips
βx = 2.24%
βz = 1.99%
βzz = —
βxx = 0%
βy = βy × Rc,zz = 0.01%
The ratio of FIM and free-field ground motion in the frequency domain is
presented as a transfer function, Hu. The period-dependent transfer function,
Hu, is calculated using semi-empirical equations, as in NIST (2012b) Equation
3-3 and Equation 3-4. It is dependent on a parameter κa, the dimensions of the
foundation, and the shear-wave velocity of the soil. The overall dimensions of
FEMA P-2139-3 C: Development of Soil Springs, Soil Dampers, and Frequency- C-13
Modified Ground Motion Records for the SSI Parametric Study
the structure in plan view (96 feet × 48 feet) were selected for calculating Hu,
and the embedment was assumed to be negligible. The foundation parameter
κa is calculated using the semi-empirical model by Kim and Stewart (2003), as
given in NIST (2012b) Equation 3-5, based on the shear-wave velocity of the
soil at the surface of the site, reduced by the effective reduction factor for shear
modulus, which is estimated to be 0.8 for the stiff site and 0.6 for the soft site.
Figure C-4 shows the calculated transfer functions for the soft and stiff sites.
C-14 C: Development of Soil Springs, Soil Dampers, and Frequency- FEMA P-2139-3
Modified Ground Motion Records for the SSI Parametric Study
Figure C-5 Sample ground motion before (blue) and after (orange)
application of frequency-dependent transfer function.
FEMA P-2139-3 C: Development of Soil Springs, Soil Dampers, and Frequency- C-15
Modified Ground Motion Records for the SSI Parametric Study
Appendix D
Archive of Peak Response
Calculations
Table D-1 Peak Response Archive Files for Short-Period Reinforced Masonry Commercial Building
Archetypes
Displacement Stripe_Statistics_ATC116_COM1B-DC1
COM1B-DC 1 High
Capacity Stripe_Statistics_ATC116_COM1B-DC2
COM2B 2 High Baseline Stripe_Statistics_ATC116_COM2B
Displacement Stripe_Statistics_ATC116_COM2B-DC1
COM2B-DC 2 High
Capacity(1) Stripe_Statistics_ATC116_COM2B-DC2
SSI/Foundation Stripe_Statistics_ATC116_COM2B-SS2-Unfiltered
COM2B-SS 2 High
Flexibility(2) Stripe_Statistics_ATC116_COM2B-SS2-Filtered
COM3B 4 High Baseline Stripe_Statistics_ATC116_COM3B
Displacement Stripe_Statistics_ATC116_COM3B-DC1
COM3B-DC 4 High
Capacity(1) Stripe_Statistics_ATC116_COM3B-DC2
COM4B 1 Very High Baseline Stripe_Statistics_ATC116_COM4B
Displacement Stripe_Statistics_ATC116_COM4B-DC1
COM4B-DC 1 Very High
Capacity(1) Stripe_Statistics_ATC116_COM4B-DC2
COM5B 2 Very High Baseline Stripe_Statistics_ATC116_COM5B
Displacement Stripe_Statistics_ATC116_COM5B-DC1
COM5B-DC 2 Very High
Capacity(1) Stripe_Statistics_ATC116_COM5B-DC2
Displacement Stripe_Statistics_ATC116_COM6B-DC1
COM6B-DC 4 Very High
Capacity(1) Stripe_Statistics_ATC116_COM6B-DC2
(1)
Two drift capacity levels: 70% and 130% of the drift capacity of the baseline model.
(2)
Soft site with unfiltered and filtered ground motions.