Engineering Practice Paper

Earthquake Spectra
Implementing the 2021, Vol. 37(3) 2150–2173
Ó The Author(s) 2021
performance-based seismic Article reuse guidelines:
sagepub.com/journals-permissions

design for new reinforced DOI: 10.1177/8755293020981968

journals.sagepub.com/home/eqs

concrete structures:
Comparison among ASCE/
SEI 41, TBI, and LATBSDC

Siamak Sattar, M.EERI1, Anne Hulsey, M.EERI2,

Garrett Hagen3, Farzad Naeim, M.EERI4, and
Steven McCabe, M.EERI1

Abstract
Performance-based seismic design (PBSD) has been recognized as a framework for
designing new buildings in the United States in recent years. Various guidelines and
standards have been developed to codify and document the implementation of
PBSD, including ‘‘Seismic Evaluation and Retrofit of Existing Buildings’’ (ASCE 41-17), the
Tall Buildings Initiative’s Guidelines for Performance-Based Seismic Design of Tall Buildings
(TBI Guidelines), and the Los Angeles Tall Buildings Structural Design Council’s An
Alternative Procedure for Seismic Analysis and Design of Tall Buildings Located in the Los
Angeles Region (LATBSDC Procedure). The main goal of these documents is to regular-
ize the implementation of PBSD for practicing engineers. These documents were
developed independently with experts from varying backgrounds and organizations
and consequently have differences in several degrees from basic intent to the details
of the implementation. As the main objective of PBSD is to ensure a specified build-
ing performance, these documents would be expected to provide similar recommen-
dations for achieving a given performance objective for new buildings. This article
provides a detailed comparison among each document’s implementation of PBSD for
reinforced concrete buildings, with the goal of highlighting the differences among
these documents and identifying provisions in which the designed building may
achieve varied performance depending on the chosen standard/guideline. This com-
parison can help committees developing these documents to be aware of their

1
National Institute of Standards and Technology, Gaithersburg, MD, USA
2
Stanford University, Stanford, CA, USA
3
Degenkolb Engineers, Los Angeles, CA, USA
4
Farzad Naeim Inc., Irvine, CA, USA

Corresponding author:
Siamak Sattar, National Institute of Standards and Technology, 100 Bureau Drive, Gaithersburg, MD 20899, USA.
Email: siamak.sattar@nist.gov
Sattar et al. 2151

differences, investigate the sources of their divergence, and bring these documents
closer to common ground in future cycles.

Keywords
Performance-based seismic design, ASCE 41, TBI, LATBSDC, reinforced concrete
structures
Date received: 18 September 2019; accepted: 27 October 2020

Introduction
Performance-based seismic design (PBSD) has been recognized as a suitable tool for the
design and evaluation of buildings because of its capability to evaluate the response of a
building with respect to an explicit target performance, unlike the tightly regulated, pre-
scriptive methods of traditional design approaches; the prescriptive design approaches are
intended to provide a minimum safety standard and do not necessarily identify a specific
performance objective. Various standards and guidelines have been developed within the
past 10–15 years to transform PBSD from a research practice to the state-of-practice and
standardize the implementation of PBSD for practicing engineers. Three widely used docu-
ments developed for the implementation of PBSD are (1) ASCE 41-17 (ASCE/SEI, 2017)
Seismic Evaluation and Retrofit of Existing Buildings (hereafter referred to as ASCE 41), (2)
the Pacific Earthquake Engineering Research Center, Tall Buildings Initiative’s Guidelines
for Performance-Based Seismic Design of Tall Buildings (TBI Guidelines, version 2.03) (TBI,
2017), and (3) the Los Angeles Tall Buildings Structural Design Council’s An Alternative
Procedure for Seismic Analysis and Design of Tall Buildings Located in the Los Angeles
Region (LATBSDC, 2020). These documents were developed for slightly different purposes.
One of the earliest documents in the progression of codifying PBSD for implementation in
practice was FEMA 273 (Federal Emergency Management Agency (FEMA), 1997). In 2000,
FEMA and American Society of Civil Engineers (ASCE) updated FEMA 273 to adapt it into
a national pre-standard and published it as FEMA 356 (FEMA, 2000). In 2006, ASCE pub-
lished ASCE/SEI 41-06 (ASCE/SEI, 2006) as the standard for applying the PBSD concepts
in seismic evaluation of existing buildings. The most recent version of this standard is ASCE/
SEI 41-17 (ASCE/SEI, 2017). This standard was originally developed to be used in the eva-
luation of existing buildings. However, in recent years, ASCE 41 has also been widely used in
the design of new buildings.
Another document developed for the implementation of the PBSD is TBI. In 2010, the
Pacific Earthquake Engineering Research (PEER) center published the the first version of
TBI (TBI, 2010) as a PBSD design approach alternative to ASCE 7 and the International
Building Code (IBC) (International Code Council, 2009). The most recent version of TBI
Guidelines was published in 2017 (TBI, 2017). The performance of the buildings designed
per these guidelines is intended to satisfy the ASCE 7 target performance for a Risk
Category II building, which is the focus of this article. TBI was developed considering tall
buildings with a long fundamental period to be constructed in high seismic regions. These
buildings tend to have periods significantly larger than 1.0 s with slender aspect ratios and
significant mass participation and response in higher modes.
The third document considered in this article is the LATBSDC Procedure. In 2005, the
Los Angeles Tall Buildings Seismic Design Council first published a procedure document
as alternative procedures for seismic analysis and the design of tall buildings located in the
2152 Earthquake Spectra 37(3)

Los Angeles region. The LATBSDC has updated this document every three years since
2005. The 2020 version of this document was published with the intent of designing build-
ings that meet optimal safety performance and are likely to be repairable after major earth-
quakes. The LATBSDC Procedure states that although the document was developed for
the design of tall buildings, it can also be used for the design of other building types, and
the City of Los Angeles permits use of this document for the analysis and design of other
building types.
Despite the differences in the intent for developing these three documents, as well as the
differences in the jurisdictions that adopt them, these guidelines/standards aim to standar-
dize the implementation of PBSD. As such, one may expect that these guidelines would be
similar. Although there are similarities among these documents, especially between TBI
and LATBSDC, the three documents were developed independently and have differences
in various degrees in how they address the PBSD approach. The differences range from
philosophical differences to detailed calculations of structural demands and capacities as
well as analysis of the building models. These differences can potentially lead to different
building designs and performances. Buniya et al. (2020) studied the implication of using
ASCE 41 (ASCE/SEI, 2017), TBI (TBI, 2010), and LATBSDC (LATBSDC, 2014) on the
design and assessment of a 12-story reinforced concrete special moment frame, and showed
that TBI and LATBSDC can lead to stiffer and stronger design of RC moment frames.
Their study highlighted the importance of understanding the differences among the ASCE
41, TBI, and LATBSDC. This article provides a detailed comparison and highlights the
differences among these three documents on the implementation of PBSD with the focus
on reinforced concrete (RC) buildings. The comparison is conducted in three levels. The
article first discusses the philosophical differences among the documents, then reviews the
differences in the modeling and analysis recommendations, and finally provides a discus-
sion on the acceptance criteria proposed in the three documents.

Philosophical differences
Any comparison made between various standards must begin with the purpose and scope of
the respective documents, which can give context to understand the differences and similari-
ties. As the title of ASCE 41 suggests, this document was developed primarily for evaluating
existing buildings. As such, it is left to remain as broad as possible to capture a wide range of
building types from various time periods and construction practices. By employing and exten-
sively codifying a PBSD approach, ASCE 41 became a ‘‘first-generation’’ document for
implementing PBSD in 2006 and has since been used in the design of new buildings, though
this was not the intended purpose. The availability of such a document allowed more design
professionals to take advantage of ASCE 7’s recognition of ‘‘alternative procedures’’ (ASCE
7 Section 1.3.1) in cases for which the prescriptive approach is not permitted, the performance
objectives need to be enhanced, or when PBSD simply results in a more efficient design.
The TBI Guidelines, on the contrary, were developed specifically for tall building design as
an alternative to the prescriptive provisions available. In cases where ASCE 7 and other
material reference standards place restrictions on height or options for seismic force–resisting
structural systems, engineers have used capacity-based design procedures for determining
member proportions and then analyzing the performance at various shaking levels using non-
linear structural analysis. PEER launched the Tall Building Initiative to develop the TBI
Guidelines for the capacity-based design approach with the purpose of both providing a
framework for seismic design and laying the groundwork for future Building Code provi-
sions. In preparation for the TBI Guidelines, the Tall Building Initiative also worked with the
Sattar et al. 2153

Applied Technology Council to produce PEER/ATC 72-1 Modeling and Acceptance Criteria
for Seismic Design and Analysis of Tall Buildings (hereafter referred to as ATC 72) (PEER/
ATC, 2010), which addressed concerns regarding the reliability of performance-based analyti-
cal methods for predicting tall building performance and issues related to modeling, analysis,
and acceptance criteria for various structural systems. TBI relied heavily on ATC 72 as it
crafted the Guidelines. TBI also took the ASCE 7–defined Risk Category II and a seismic
hazard similar to that found in the Western United States as its starting point. The TBI
Guidelines also include provisions for other risk categories or seismic hazards, where the
methodology may require a few modifications. In addition, the TBI Guidelines are only appli-
cable to structures that behave like tall buildings, namely those that have a fundamental
period much greater than 1 s, are subject to higher modes, and experience significant lateral
drift due to axial rather than shear deformations (TBI, Section 1.2).
Finally, the LATBSDC Procedure was developed for similar purposes as the TBI
Guidelines, specifically to identify demands experienced by tall buildings (LATBSDC
Section 2.1). While not explicitly meant to be a model for future building code provisions,
as stated in the TBI Guidelines, LATBSDC Procedure does present additional considera-
tions with very thorough, transparent commentary on how and why certain portions differ
from ASCE 7, ASCE 41, and/or the TBI Guidelines. Again, unlike ASCE 41, the
LATBSDC Procedure is focused specifically on tall buildings, but this distinction is quanti-
fied as a height greater than 160 feet (49 m) above the adjacent ground surface (LATBSDC
Section 2.2), rather than on more general structural behavior as in the TBI Guidelines. The
LATBSDC Procedure also includes provisions for Risk Categories other than Risk
Category II.

Performance objectives
The justification for the use of PBSD in new design is based on ASCE 7 Section 1.3.1.3,
which allows alternative procedures, as long as they result in designs that are at least as struc-
turally reliable as those produced by the prescriptive methods. Therefore, the performance
objectives should be tied to ASCE 7’s expected performance. The explicit performance objec-
tive for a Risk Category II building in ASCE 7 is Collapse Prevention (CP) performance
under the risk-targeted maximum considered earthquake (MCER), which in turn is expected
to satisfy Life Safety (LS) for the Design Earthquake (taken as 2/3 MCER) and Immediate
Occupancy (IO) for frequent shaking. The loads in ASCE 7 are calculated based on the
Design Earthquake, while the large margins of safety (incorporated into the R-factors as
approximately a 1.5 margin) ensure that the CP performance objective at MCER is met.
Unlike ASCE 7, the PBSD procedures developed in the three documents provide a
more explicit assessment of the behavior, such as the deformations in each member, and
can do so at multiple hazard levels. ASCE 41 provides a range of performance objectives,
such as basic and enhanced, to be applied on a case-by-case basis at the discretion of the
engineer, building owner, or governing jurisdiction. In reference to the new building
design, ASCE 41 includes an objective named the Basic Performance Objective Equivalent
to New Building Standards (BPON), which is intended to correspond to the same perfor-
mance objective targeted as a minimum in ASCE 7 provisions for new buildings. In ASCE
41, the identified performance level for a Risk Category II building under the BSE-2N-
level earthquake hazard (equivalent to the ASCE 7 MCER) is CP, just as in ASCE 7. For
the ASCE 7–defined Design Earthquake, or BSE-1N in ASCE 41, there are two
2154 Earthquake Spectra 37(3)

performance levels specified for structural and nonstructural elements: Life Safety and
Position Retention for structural and nonstructural performance, respectively.
The TBI Guidelines and LATBSDC Procedure handle the equivalent performance objec-
tive reliability requirement through alternate approaches. Where ASCE 41 uses the two
seismic hazards defined in ASCE 7 as its starting point, the TBI and LATBSDC perfor-
mance objectives are based on the underlying expectation that a well-designed building
would meet IO for frequent earthquakes, LS for the Design Earthquake, and CP for the
MCER. LATBSDC originally included all three objectives but removed the intermediate
Life Safety at the Design Earthquake objective in a 2008 revision. The rationale for the
removal of LS was that the IO and CP performance levels are claimed to be more tangible
than Life Safety. It is argued that the objective of Life Safety at the Design Earthquake is
not well defined and does not provide much additional information.
The two performance objectives used by both TBI and LATBSDC are as follows: ser-
viceable behavior (essentially elastic response and limited damage) under service-level
earthquake (SLE) (having a 50% probability of exceedance in 30 years or a 43-year return
period) and a low probability of collapse under extreme shaking as quantified by the
ASCE 7 MCER. Although LATBSDC Procedure explicitly uses the term ‘‘low probability
of collapse,’’ TBI Guidelines point out that the computational capabilities required for cal-
culating the probability of collapse as a function of ground motion intensity may not be
currently available for tall buildings, as they are for other classes of structural systems.
Therefore, TBI Guidelines articulate the CP at MCER objective as ‘‘providing adequate
safety against collapse’’ and achieves this implicitly by satisfying the following: (1) accepta-
ble response of elements; (2) sufficiently small strength demand on force-controlled actions
or elements; (3) deformation demands in deformation-controlled actions are within the
experimentally verified limits; and (4) acceptable residual drift. The LATBSDC Procedure
uses a similar approach to ensure that the probability of collapse is low at the MCER
hazard level. In addition, LATBSDC Procedure mentions that application of residual
story drift ratios and reduction factors on capacity of force-controlled actions, as well as
limiting the deformation of deformation-controlled actions to values less than or equal to
the capping point are intended to lead to design of buildings that are repairable after
major earthquakes.
Unlike ASCE 41, TBI and LATBSDC do not have a separate, parallel evaluation for
nonstructural performance objectives. That said, their acceptance criteria for both hazard lev-
els do include story drift limits, which can act as a proxy for assessing some aspects of non-
structural performance for drift-sensitive elements. Another benefit of drift limits is the
ability to assess the performance of a building as a whole, rather than each individual compo-
nent. TBI and LATBSDC acceptance criteria include both component response and story
drift limits. ASCE 41, on the contrary, only includes component response for the structural
acceptance criteria, and only the nonstructural acceptance criteria include story drift limits.
Table 3 in Appendix 1 provides a high-level comparison among the three Performance-Based
Earthquake Engineering (PBEE) standards and Guidelines for new buildings.

Modeling recommendations in ASCE 41, TBI, and LATBSDC

One of the main steps in the PBSD process is creating a building model. Many of the mod-
eling techniques and requirements are consistent across the standards, yet there are subtle
differences. These differences are outlined below.
Sattar et al. 2155

Table 1. Factors to convert the nominal to expected material strengths

Material ASCE 41 TBI LATBSDC Ratio w.r.t. ASCE 41

Concrete (ultimate compressive) 1.5 1.3 1.3 0.87

Reinforcing Steel 1.25 1.17 1.17 0.94
ASCE: American Society of Civil Engineers; TBI: Tall Buildings Initiative; LATBSDC: Los Angeles Tall Buildings
Structural Design Council.

Strength and stiffness properties

PBSD is primarily based on modeling the expected behavior of a building, unlike ASCE 7.
In PBSD, usually nominal material strengths are multiplied by factors to represent expected
values for members expected to be controlled by deformation-controlled response, meaning
sudden strength loss is not expected in the member. While the TBI Guidelines and the
LATBSDC Procedure use identical factors, they differ from the factors used in ASCE 41.
Table 1 summarizes the respective values for each standard, along with the ratio of the TBI
and LATBSDC factors with respect to ASCE 41; the ratios show that the expected strength
in TBI and LATBSDC can vary from 87% to 1.43% of ASCE 41 depending on the mate-
rial type. Note that in some cases, ASCE 41 requires material testing for which the expected
material strengths would be associated with the mean tested material property values.
The component stiffnesses are also calculated differently among the three documents,
most notably for concrete components. The comparisons between flexural stiffness of vari-
ous components are summarized in Table 2. As the tables in each standard include a
slightly different list of components, some of the numerical comparisons are based on the
most similar components from ASCE 41, as indicated in the table footnotes.
In addition to discrepancies in the factors used, there are a few conceptual differences that
cause variation in the modeled stiffnesses and results in computed forces and displacements of
the building. The effective stiffness for concrete components includes the modulus of elasticity,
Ec, which is a function of concrete strength, fc0 . For the stiffness comparison between the three
documents, the expected fc0 is used; however, ASCE 41 does not provide explicit instructions on
how to calculate Ec although ACI 318 is referred to for strengths and other properties through-
out Chapter 10 of the document. The LATBSDC Procedure is the only guideline that provides
a different equation to calculate the Ec of high-strength concrete. This is because concrete tests
conducted during construction of various projects in the LA region showed that the Ec of con-
crete generally achieved in southern California may be less than those suggested by ACI 318
equation. Therefore, LATBSDC Procedure adopted a revised formula to limit drifts and slab
deflections.
Another conceptual difference regarding effective stiffness is the TBI and LATBSDC
Procedures’ use of separate stiffnesses for the lower hazard level (50%/30 years) perfor-
mance objective versus the MCER-level objective. This distinction presumably addresses
the expected damage state for the performance and reduced cracking associated with the
lower service-level hazard relative to the MCER. ASCE 41, on the contrary, consistently
defines cracked stiffness as the secant stiffness to component yield deformations indepen-
dent of hazard level. Table 2 shows examples of flexural stiffness, in which the component
stiffness calculated based on TBI Guidelines and LATBSDC Procedure can be 0.93 to 1.55
and 0.57 to 2 times that of ASCE 41, respectively. The ratios for comparison in Table 2
assumed a nominal strength of 5000 psi (34.5 MPa).
2156 Earthquake Spectra 37(3)

Table 2. Comparison of recommended effective flexural stiffnesses between ASCE 41, TBI, and
LATBSDC
Component type ASCE 41 TBI and LATBSDCa Ratio w.r.t. ASCE 41b
MCER 50%/30 TBI or LA TBI or LA
MCER 50%=30
ASCE 41 ASCE 41

RC beamsc 0.3 EcIg 0.3 EcIg 0.5 EcIg 1.0 1.67

Coupling beam 0.3 EcIgd 0:07 hl Ec Ig 0:07 hl Ec Ig <1.0 <1.0e
<0:3Ec Ig <0:3Ec Ig
RC columns 0.3–0.7 0.7 EcIg 0.7 EcIg 0.86–2.0 0.86–2.0
EcIgf
Flat slab (non- 0.33 EcIg – – – –
prestressed out-of- or 4c1/l1ø1/3
plane)
Flat slab (prestressed 0.5 EcIg – – – –
out-of-plane)g
Non-PT diaphragm 1.0 EcIg 0.25 EcIg 0.5 EcIg 0.25 0.5
(in-plane)
PT diaphragm (in- 1.0 EcIg 0.5 EcIg 0.8 EcIg 0.5 0.8
plane)
Structural walls (in- 0.35 EcIg 0.35 EcIgh 0.75 EcIg 1.0 2.14
plane)
Structural walls (out- – 0.25 EcIg 0.25 EcIg – –
of-plane)
Basement wallsi (in- j
0.8 EcIg 1.0 EcIg – –
plane)
Basement wallsi (out- – 0.25 EcIg 0.25 EcIg – –
of-plane)
ASCE: American Society of Civil Engineers; TBI: Tall Buildings Initiative; LATBSDC: Los Angeles Tall Buildings
Structural Design Council; RC: reinforced concrete; PT: post-tensioned; MCER: risk-targeted maximum considered
earthquake.
a
LATBSDC requires a different equation for fc0 .6000 psi.
b
Ratios include the variation in Ec and expected strength factors for the three standards, assuming nominal
fc0 = 5000 psi.
c
Note that ASCE 41 allows the use of two times Ig of the beam web alone to account for the stiffness contribution of
the effective slab width unless Section 10.3.1.3 is used to calculate the effects of the T-beam explicitly.
d
ASCE 41 10.7.2.2 allows the use of RC beams (non-pressed) stiffness value in the absence of more detailed analysis
e
This value is computed assuming the upper bound stiffness of 0:3Ec Ig .
f
ASCE 41: RC column stiffness is interpolated based on axial compression between 0.1 and 0.5 Ag fc0 .
g
Compared to ASCE 41 flat slab (non-prestressed) for ratio. The 0.5 coefficient is also multiplied by the effective slab
width in ASCE 41, thus reducing the effective stiffness of the gross two-way slab properties further.
h
Use nonlinear fiber elements to explicitly account for cracking of concrete. Note that ASCE 41 allows fiber elements
as an alternative to the permitted effective stiffness values and also provides commentary in Section C10.7.2.2 with
equations that align more closely with the fiber analysis.
i
Compared to ASCE 41 walls for ratio.
j
ASCE 41 commentary C10.7.2.2 notes that alternate values based on analysis are permitted where little to no
cracking is expected to occur. It is noted that ASCE 41-06 previously had 0.8 for uncracked walls.

The comparison between the effective shear stiffness among the three documents shows that
the shear stiffnesses recommended by TBI and LATBSDC for the linear models at the SLE
shaking are the same as those for ASCE 41, while the values suggested by TBI and LATBSDC
for the nonlinear models at the MCER hazard level are taken as 50% of gross for walls and
post-tensioned (PT) diaphragms and 25% for non-PT diaphragms. Table 4 in Appendix 1 pro-
vides comparison of shear stiffness of various components in the three documents. The TBI
Sattar et al. 2157

Guidelines and LATBSDC Procedure also recommend considering lower and upper bound
stiffnesses for considering the backstay effects for the MCER level analysis. LATBSDC
Procedures provides a set of lower and upper bound stiffnesses for diaphragms at the podium
level or below, while TBI Guidelines refer to ATC 72 and NIST (2016) for further guidance on
design and modeling considerations to address backstay effects.

Modeling considerations
The TBI Guidelines and the LATBSDC Procedure include specific guidelines regarding
various modeling issues, five of which will be addressed here and compared to the corre-
sponding portions of ASCE 41. Some of the applications depend on the hazard level or
the analysis type (linear vs nonlinear).

Model type. ASCE 41 allows a simplified two-dimensional (2D) model if certain diaphragm
and torsional requirements are met. Both TBI and LATBSDC require a full three-
dimensional (3D) model.

Effective viscous damping ratio. ASCE 41 discusses effective viscous damping ratios for linear
static procedure (LSP), linear dynamic procedure (LDP), and nonlinear static procedure
(NSP) in the context of the damped response spectrum. This ratio is specified as 5%,
except for a few listed building types. There is no distinction between the damping ratios
at different hazard levels or performance objectives. The damping ratio b is capped at 3%
for the nonlinear dynamic performance (NDP). This reduction is meant to prevent double
counting of the hysteretic damping that is captured explicitly during the analysis.
The TBI Guidelines and the LATBSDC Procedure require considering a small amount
of viscous damping that is not explicitly considered in the simulation model. The modal or
Rayleigh damping for the primary modes of response for serviceability evaluation is not
permitted to exceed a fraction of critical damping computed as the minimum of
pffiffiffiffiffiffiffiffiffiffiffiffiffi
0:36= height and 0.05. The LATBSDC Procedure also allows the use of a combination of
modal damping and stiffness-proportioned Rayleigh damping or linearly increasing damp-
ing from 0.2T to the period of 0.0 seconds to prevent overestimation of floor accelerations.
pffiffiffiffiffiffiffiffiffiffiffiffiffi
For the MCER evaluation, the 0:36= height equation can be used, but the damping does
not need to be taken in less than 2.5%.
The LATBSDC Procedure also recommend scaling the ground motion time histories to
a 2.5% damped response spectrum for serviceability evaluation when nonlinear dynamic
response analysis is employed. The TBI Guidelines allow the use of 5% damped response
spectra, and if a different damping ratio is desired, adjustment factors from Rezaeian et al.
(2014) are suggested.

Diaphragms. ASCE 41 classifies diaphragms as fully rigid or flexible (as opposed to semi-rigid)
when the deformations of the diaphragm relative to the vertical elements meet specified limits
or as semi-rigid otherwise. ASCE 41 also notes that chords, collectors, and ties are required in
the structure, and the axial actions must be treated as force-controlled behaviors. The docu-
ment discusses modeling of diaphragms quite generally and does not have specific require-
ments for modeling the flow of forces through the diaphragms at vertical discontinuities.
The TBI Guidelines require diaphragm modeling based on the anticipated level of crack-
ing in the concrete. The TBI Guidelines have a set of requirements for diaphragms to be
modeled as rigid in-plane elements, but warn that modeling diaphragms as rigid elements
2158 Earthquake Spectra 37(3)

may lead to unrealistically large transfer forces at locations with significant vertical discon-
tinuities in the seismic force–resisting systems (e.g. podium levels and other setback levels).
In those cases, TBI Guidelines recommend including the flexibility of the diaphragms in
the finite element model. In addition, the transfer and dissipation of chord forces, com-
bined with other forces, are required to be evaluated at diaphragm discontinuities as well
as the levels with significant vertical discontinuities. The LATBSDC Procedure uses a simi-
lar text as the TBI Guidelines regarding modeling diaphragms.

Beam-column joint flexibility. The first mention of modeling connections in ASCE 41 is a

requirement that they must be explicitly included in nonlinear models if their behavior is a
limiting factor for strength or ductility, or if the connection flexibility contributes significantly
to the relative deformation of surrounding components. All the subsequent guidance on joint
modeling is found in material-specific chapters. Concrete moment frames that adhere to the
strong-column–weak-beam requirement should use the beam centerline-to-centerline dimen-
sions with rigid column offsets. For nonlinear models, ASCE 41 provides tables with model-
ing parameters to define the appropriate load–deformation response of the connection.
The TBI Guidelines require modeling the shear behavior of beam-column joints as elastic,
force-controlled actions using guidelines in Moehle (2014) and NIST (2017a), or ASCE 41.
The LATBSDC Procedure states that in the absence of explicit modeling of panel zones, the
required joint flexibility can be modeled using the centerline-to-centerline beam dimensions.

Inclusion of components in model. ASCE 41 classifies all structural components that either con-
tribute to the structure’s lateral stiffness or are loaded due to lateral deformations as either a
primary or a secondary component. While all primary and secondary components must be
able to accommodate deformations, if a component is not necessary to resist the lateral forces
associated with a given Performance Level, it is classified as secondary. Due to the limitations
of linear analysis in accounting for strength and stiffness degradation, ASCE 41 conserva-
tively excludes secondary components from linear models. However, if the total stiffness of
secondary components exceeds 25% of the total initial lateral stiffness, some of the secondary
components must be reclassified as primary. Nonlinear analysis is better able to address the
degradation and associated deformations during the analysis, so both primary and secondary
components are included in the model. As with the potential reclassification of secondary ele-
ments in linear models, if the nonstructural components exceed 10% of the stiffness or
strength of a given story, they must be reclassified as structural components.
The TBI Guidelines require linear structural models to include both the intended lateral
force–resisting system and any components of the gravity system or nonstructural compo-
nents that also contribute to the lateral stiffness and strength under the anticipated defor-
mations. TBI Guidelines note that while linear models usually do not include elements that
are not part of the intended lateral resistance, in tall buildings such elements can cumula-
tively contribute significant stiffness, so they need to be included to properly model the
expected behavior. For the MCER level evaluation, the TBI Guidelines also require a final
compatibility check to ensure that any elements that were not included in the model could
withstand the maximum drifts that were reached during analysis.
The LATBSDC Procedure is not explicit about what components should be included in
the linear model for the lower hazard level, except for a general comment that the model’s
stiffness and strength should adequately represent the actual structure’s behavior. For the
MCER level hazard, the model should include the lateral force–resisting system and other
structural members that contribute to the lateral stiffness. The LATBSDC Procedure
Sattar et al. 2159

requires claddings and their connections to the structure to accommodate MCER displace-
ments without failure.

Nonlinear modeling
In some ways, nonlinear modeling is the heart of PBSD framework. To realistically ana-
lyze the expected performance, the model should represent the key characteristics of the
building components including cyclic and monotonic response characteristics. The follow-
ing sections describe how each standard addresses component degradation and a few com-
ponent types that are handled differently in each document.

Modeling component degradation. Nonlinear models include load–deformation curves for each
component. In addition to the monotonic backbone curve that includes elastic stiffness,
yielding, hardening, ultimate strength, post-peak response, and residual strength, a dynamic
analysis must also address hysteresis and cyclic degradation that occur between cycles. The
combination of these degradation types can be addressed using various conceptual models.
The typical force–deformation backbone curve in ASCE 41 is shown in Figure 1. The
material-specific chapters in ASCE 41 (Chapters 8–12) include tables with modeling para-
meters (lowercase letters in Figure 1) for various types of components. The parameters pro-
vide partial information, such as plastic deformation (a) or the ultimate deformation ratio
(point E). In the absence of parameters for a particular component type, ASCE 41 provides
guidance on how to derive a similar curve based on cyclic experimental testing, such that the
curve already includes the effects of cyclic degradation in addition to the monotonic response.
ASCE 41 is explicit in allowing the load–deformation relations described above to be
used as an envelope for hysteretic response. This approach assumes that the cyclic degra-
dation is included in the backbone curve.
Both the TBI Guidelines and the LATBSDC Procedure state that the applicable ASCE
41 curves can be used as a modified backbone curve. If so, the steep slope between points
C and D must be altered, although there is no guidance on how they should be altered. In
lieu of adapting the ASCE 41 curves, both standards include modeling options from ATC
72 with minor modifications.

Figure 1. Typical load–deformation relations for component modeling (ASCE 41).

2160 Earthquake Spectra 37(3)

Modeling of different component types. This section summarizes the modeling recommenda-
tions in the three standards/guidelines for various components of the RC structures.

Beams and columns. The TBI Guidelines allow the ASCE 41, NIST (2017b), or ATC 72 meth-
ods for modeling flexure in beams and columns. The LATBSDC Procedure does not specifi-
cally discuss the lumped plasticity modeling approach for beams and columns. However,
LATBSDC Procedure states that if the fiber modeling approach is employed, the concrete
stress–strain behavior shall comply with ASCE 41 backbone curves or be based on labora-
tory test data.

Slabs. The TBI Guidelines refer to both ASCE 41 and ATC 72 for modeling slabs in slab-
column frames. Unlike beam and column modeling, however, the two documents discuss
the same two models: effective beam width and equivalent frame. The TBI Guidelines and
LATBSDC Procedure specify that when the slab acts as a framing element to engage grav-
ity columns, out-of-plane modeling may be required.

Shear walls. ASCE 41 allows slender shear walls to be modeled with equivalent beam-
column elements. For short and intermediate walls that may experience inelastic shear, the
modified beam-column analogy or a method including multiple nodes and springs is
required. The nonlinear backbone curve is defined according to one of two ASCE 41 tables
for modeling parameters, depending on whether flexure or shear governs the response. For
nonlinear analysis, the effective stiffness for the flexural backbone curve is calculated
through moment-curvature analysis, fiber analysis, or 0.35 EcIg value provided in the effec-
tive stiffness table. ASCE 41-17 also introduced commentary for adding the flexibility con-
tributions of bar-slip to fiber analysis or moment-curvature analysis. When shear governs,
the backbone curve includes the effects of flexural, shear, and bar-slip deformations. A tri-
linear backbone for shear is used, identifying the cracking and yield points of shear-
governed deformations. The member strengths are determined based on ACI 318 principles
(ACI, 2014), along with specific requirements as to what material to include in the analysis.
The TBI Guidelines specify that either fiber or moment-curvature models are appropri-
ate for modeling shear walls and refer to ATC 72 for guidance. For fiber models, ATC 72
gives an overview of the influence of material models and element size, with general recom-
mendations. In the absence of a model titled moment-curvature in ATC 72, TBI Guidelines
are likely referring to the equivalent beam-column model. ATC 72 indicates that a flexural
stiffness of 0.4–0.5 EcIg is typically appropriate for axial loads around 0:1Ag fc0 , along with
an equation for lower loads, although moment-curvature and bar-slip should be used to
validate any effective stiffness values. The TBI Guidelines also note that higher modes in tall
buildings may induce nonlinear behavior in upper stories, requiring verification of elastic
behavior or the use of nonlinear elements throughout the full height of the wall.
The commentary of the LATBSDC Procedure notes that fiber modeling is the common
practice to represent axial and bending responses of shear walls in nonlinear analysis, with
shear response represented by a linear stiffness. The document states that instead of using
the tabulated stiffness values, fiber model shall be used directly to represent axial and flex-
ural stiffness. The LATBSDC Procedure specifies a few general requirements for the mate-
rial models and provides equations for the plastic hinge length, to be used as the dimension
of the fiber element. The effective stiffness table in the LATBSDC Procedure distinguishes
between structural and basement perimeter walls which serve the function of retaining
Sattar et al. 2161

walls as well, with the impression that basement walls can be assigned an effective flexural
stiffness of 0.8 EcIg and, therefore, do not need to use fiber models.

Coupling beams. The TBI Guidelines refer to NIST (2017a) or ASCE 41 for flexure-
controlled conventionally reinforced coupling beams. For diagonally reinforced coupling
beams, the strength of the beam shall be modified to incorporate the contribution from
diagonally placed reinforcement and the slab. The current discussion in ASCE 41 is lim-
ited to conventionally reinforced and/or diagonally reinforced coupling beams; conven-
tionally reinforced beams would not be appropriate in the new design. In the absence of
further analysis, ASCE 41 allows the effective stiffness as that of a non-prestressed beam
to be taken, or 0.3 EcIg for flexure and 0.4 EcAw for shear. These stiffness values are differ-
ent than values recommended by TBI and LATBSDC, as shown in Table 2. The TBI
Guidelines is the only document that discusses the need for consideration of the ‘‘grow’’ on
the tension side of the wall, which in turn will increase the coupling beams’ rotation
demand. However, this document does not provide guidance on the implementation of
this phenomenon in numerical models. Another difference between the three documents is
that coupling beam and shear wall actions related to shear demands are deformation-
controlled in ASCE 41, but force-controlled in TBI and LATBSDC.

Analysis recommendations in ASCE 41, TBI, and LATBSDC

In addition to differences in the modeling approaches, there are various alternatives for
the analytical procedures in the three documents. Here, analysis refers to the way in which
the model is subjected to representative earthquake loads. There are four basic procedures
used in the standards implementing PBSD, differing in whether they assume linear or non-
linear behavior and in static versus dynamic loading types. These four procedures include
linear static, linear dynamic, nonlinear static, and nonlinear dynamic procedures.
All four procedures are outlined in ASCE 41. However, because ASCE 41 is meant to
cover a large variety of building types, there are restrictions on the use of the procedures.
Neither linear procedure is allowed for buildings with irregularities such as discontinuities,
weak story, or torsional issues. While tall buildings often include discontinuities, the more
inherent analytical concern is higher mode participation that would invalidate static analy-
sis procedures. Therefore, ASCE 41 does not permit the use of solely static procedures if
higher mode effects are significant. One way to account for both the nonlinear and
dynamic behavior is to supplement an NSP analysis with LDP. In such a case, the LDP
deformation-controlled acceptance criteria can be multiplied by 1.33, whereas the NSP cri-
teria remain the same. Based on these restrictions of ASCE 41, the higher mode effects
and discontinuities associated with most tall buildings would require either a combination
of NSP and LDP or NDP for any hazard level considered.
The TBI Guidelines and LATBSDC Procedure also stipulate the use of dynamic proce-
dures. The MCER-level analysis in TBI and LATBSDC always requires NDP. However,
the analysis for the SLE shaking (50%/30 years) allows LDP even in the presence of irre-
gularities. This is presumably due to the fact that the IO acceptance criteria limit the
acceptable response to linear behavior, making the linear analysis representative of true
behavior. The modal combination method is limited to complete quadratic combination
(CQC) in TBI and LATBSDC, whereas ASCE 41 also allows square root of the sum of
the squares (SRSS). Though not explained by either document, this likely reflects the fact
that SRSS can underestimate forces when combining similar frequencies (Wilson et al.,
1981). Unlike ASCE 41, TBI and LATBSDC require a 3D model, so the orthogonal
2162 Earthquake Spectra 37(3)

frequencies will be closer together than the sequentially higher modes of a single 2D plane.
Therefore, CQC is a more appropriate method.

Accidental torsion
Accidental torsion is also treated differently among the three documents. Because ASCE
41 allows 2D linear models, which do not inherently capture the torsional behavior, it has
the most stringent requirements for accidental torsion. However, as long as nonlinear pro-
cedures and a 3D model are used in ASCE 41, accidental torsion can be ignored for lower
seismic hazard levels as long as they are captured in worst-case higher hazard evaluations.
Considering accidental torsion only for the highest hazard-level analysis is appropriate due
to the fact that the effects of accidental torsion would affect the building response the most
at the CP level.
For ASCE 41, accidental torsion must be included if the accidental torsional moment is
greater than 25% of the actual torsional moment and the addition of the accidental torsion
increases the torsional amplification multiplier for displacements, h, by more than 10% for
any level. If h at any level exceeds 1.5, the 2D model is not permitted. When using a 3D
nonlinear model, ASCE 41 allows for a simpler consideration of accidental torsion by using
amplification factors with the center-of-mass analysis, rather than running a total of four
separate analyses to represent each eccentric case. The amplification factors must come
from a parametric sensitivity study that considers all forces, drifts, and deformations.
The TBI Guidelines, on the contrary, do not require consideration of accidental torsion for
serviceability analysis. The previous edition of TBI Guidelines state that the 3D model can
detect torsional instability in the design, while the accidental eccentricity caused by random
variability at any given level will approach zero over the many stories of a tall building.
As with ASCE 41, LATBSDC Procedure does not require accidental torsion for the
lower level seismic hazard analysis. However, the amplification factor, Ax = (h=1:2)2 , is cal-
culated for this lower level as an indication of whether accidental torsion should be included
at higher seismic hazard. This check occurs at the 30%/50-year level when using LDP
because the computational burden is less than at the MCER level using NDP. If Ax exceeds
1.5 for any floor where the maximum average MCER drift ratio at any location of the floor
exceeds 0.010, accidental torsion is required for the MCER analysis. The Ax of 1.5 corre-
sponds to h = 1:47, which is close to ASCE 41’s h cutoff of 1.5 for 2D models. If the 30%/
50-year analysis flag does indicate that accidental torsion is required, it can be handled in
one of three ways: (1) a single pair of ground motions that produces demand values higher
than the suite mean for critical actions can be reapplied in the original location of center of
mass as well as new locations corresponding to minimum accidental eccentricity. The ratio
of the maximum demands with accidental eccentricity to those without, g, is recorded for
various actions. For any actions where g.1:2, the permissible force and deformation limits
are divided by g; (2) rather than reducing the limits by g, all the ground motion pairs (not
just one) can be reapplied at the minimum eccentricity in addition to the original analysis;
(3) the material properties and configurations at relevant locations within the building can
be varied to demonstrate the vitality of the building via a sensitivity analysis.

Ground motion selection and application

Although all three standards require a suite of ground motions for the NDP, the methods
for compiling and applying the ground motions vary in terms of identifying the periods of
interest, damping of the building, and number and orientation of ground motions.
Sattar et al. 2163

Periods of interest. ASCE 41 allows the use of amplitude scaling or spectral matching for
selection and scaling of ground motions, as do TBI and LATBSDC. Scaling requirements
in ASCE 41 states that the suite average of the SRSS spectrum should be above the target
response spectrum between 0:2T to 1:5T , where T is the fundamental period of the build-
ing. The use of other selection and scaling methods would be permitted; however, a 10%
penalty is incurred for Spectral Matching, and dispersion is required to be maintained
when scaling to two or more periods (Method 2 in ASCE 7 Chapter 16).
The TBI Guidelines refer to Chapter 16 of ASCE 7 for development of a horizontal target
spectrum for the MCER hazard level. Although TBI Guidelines allow both methods identi-
fied in ASCE 7 for calculating the target spectrum (i.e. MCER response spectrum and site-
specific procedure), it prefers Method 2 (site-specific procedure) that requires two or more
site-specific target spectra. The TBI Guidelines require a sufficient number of matching peri-
ods to provide 90% participation mass in each horizontal direction of the structure.
The LATBSDC Procedure requires the use of site-specific seismic hazard analysis to
compute the response spectrum. The MCER target spectrum derived from site-specific
probabilistic seismic hazard analysis shall not fall below 80% of that calculated per
Chapter 11 of ASCE 7-16. If this requirement is not satisfied, a peer-review approval shall
be obtained. In addition, LATBSDC Procedure requires the 10% penalty for Spectral
Matching only if CMS ground motions are matched.

Number of ground motions required. All three documents require a suite of minimum of 11
ground motion pairs for each target spectrum for MCER analysis. The TBI Guidelines and
LATBSDC Procedure require the incorporation of vertical component of the ground motion
in the MCER evaluation where the impact is significant, that is, where significant discontinu-
ities exist in the vertical load-carrying components. ASCE 41 requires consideration of the
vertical component of the ground motion for horizontal cantilever components that support
gravity load, horizontal prestressed components, and components in which demands due to
gravity exceed 80% of the nominal capacity of the component. Unlike LATBSDC and TBI,
ASCE 41 does not require the simultaneous consideration of vertical seismic effects with the
horizontal response. Moreover, when spectrum matching is employed for the ground motion
selection, LATBSDC Procedure does not require the increase in the target spectrum by
110% if it is based on the uniform hazard or when peer review approved hybrid matching
and scaling approaches are utilized. ASCE 41 refers to ASCE 7 which required a minimum
of 11 ground motions, and the components must be rotated 90° in near-field sites.

Ground motion record orientation. For sites that are less than 6 miles from an active fault,
ASCE 41 requires each ground motion pair to be applied with the fault-normal compo-
nent aligned with respect to the orientation of the governing fault and the principal axes
of the building. Then the pair is rotated 90°, and the results of the two orientations are
averaged for that ground motion pair.
The TBI Guidelines and LATBSDC Procedure refer to Chapter 16 of ASCE 7-16 for
ground motion orientation. TBI and LATBSDC specifically require consideration of the rup-
ture directivity effects on target response spectral acceleration used for ground motion selec-
tion of a near-field site. In addition, TBI and LATBSDC recommend consideration of fling-
step effects to the slip-parallel direction of selected ground motion time-series when the fling-
step effects are anticipated to be important to the structural response in a near-field site. The
commentary of both documents reference relatively recent research papers by Burks and
Baker (2016) and Kamai et al. (2014) for implementation of fling-step effects.
2164 Earthquake Spectra 37(3)

Acceptance criteria recommendations in ASCE 41, TBI, and

LATBSDC
After building a model and retrieving the demands from analysis, the demands are com-
pared against the acceptance criteria to assess the performance of the building compo-
nents. While in regard to the philosophy, modeling, and analysis the TBI Guidelines and
the LATBSDC Procedure tend to deviate from ASCE 41 in similar ways, there seems to
be more divergence between them concerning the acceptance criteria. One area of conten-
tion is the serviceability criteria and how heavily they should control building design in
comparison with the MCER criteria.
Rather than comparing the acceptance criteria for the performance objectives of each
standard (which would compare ASCE 41’s Life Safety to TBI and LATBSDC’s IO), the
following sections will level the playing field with a simple side-by-side comparison of the
criteria for IO and CP. The CP acceptance criteria section includes two additional simplifi-
cations for the purpose of comparing the standards. The discussion of ASCE 41 only
includes the acceptance criteria for NDP, even though ASCE 41 would also permit a com-
bination of LDP and NSP. In addition, for the LATBSDC Procedure, all the MCER cri-
teria include a risk reduction factor, k, to adjust the performance for various risk
categories. k is equal to 1.0 for Risk Category II (the only category that the TBI Guidelines
apply to).

Serviceability acceptance criteria

The serviceability requirement ensures that a building will not lose its functionality in the
event of SLE shaking. This means that the demands identified in the analysis should not
exceed the elastic response of the components or, in cases where they do, must be low
enough that the building could still survive a future MCER earthquake, even without
repairs. The LATBSDC Procedure states that for IO any cosmetic or nonstructural repairs
should be small enough that they do not substantially affect the functionality of the build-
ing. The TBI Guidelines require the nonstructural damage to be below a threshold that
may limit the occupancy of the buildings after the event. For ASCE 41, position retention
is considered as the target performance at IO.
The acceptance criteria are broken into story drift limits and component response lim-
its. The component response compares the analysis demand to the component capacity.
The criteria for the demand to capacity comparison differ for LDP and NDP analyses.
The components are also distinguished based on their force–deformation curves, that is,
deformation-controlled and force-controlled.

Story drift limits. To quantify whether the IO performance level is exceeded, both the TBI
Guidelines and the LATBSDC Procedure include a limit of 0.5% for the story drift at any
level. ASCE 41 does not include drift limits for the structural performance. It does include
story drift for nonstructural performance, yet the limits are based on the type of compo-
nent rather than having a single limit.

Linear dynamic (response spectrum) analysis. ASCE 41 does not permit a stand-alone LDP
analysis for a tall building with discontinuities and would require an additional NSP anal-
ysis. However, the TBI Guidelines and the LATBSDC Procedure permit LDP for the servi-
ceability level because the behavior is meant to stay close to the linear range such that
Sattar et al. 2165

LDP is representative of the true behavior and because the CP behavior of the building
which is most indicative of ultimate performance is evaluated using a 3D nonlinear
dynamic procedure.

Deformation-controlled actions. Each deformation-controlled component is assessed using

the following combination of gravity and earthquake demands, QUD = QG + QE . For ASCE
41, QG is 1:1(QD + QL + QS ) (dead, live, and snow loads) or 0:9QD for the linear analysis
procedure. Note that in ASCE 41 live load is defined as the larger of 25% of the unre-
duced live load obtained in accordance with ASCE 7 and the actual live load. The TBI
Guidelines and the LATBSDC Procedure simply take QG as 1:0D + Lexp . QE is the same for
the three documents, combining the lateral demands from two orthogonal directions,
1:0Ex or y + 0:3Ey or x
The deformation-controlled component capacities are calculated using the typical
methods and equations from material codes (ACI 318 or AISC 341, 360). However, the
three standards differ in the values used for material strength and ffactors. Both ASCE
41 and the LATBSDC Procedure use expected strengths and ffactors of 1.0. The TBI
Guidelines, on the contrary, use nominal strengths and the ffactor of 1.0. The commen-
tary of TBI Guidelines explains that since the expected strength will be higher than nom-
inal, the demand to capacity ratio of 1.5 is expected to result in minor damage. As
discussed in the analysis section, the factors (known as m-factors in ASCE 41) account for
the disparity between the demands from a linear analysis and the demands that would
actually be induced in the components. At the serviceability level, these factors are low so
that the level of inelasticity will not jeopardize the reliability of the structure. ASCE 41
Chapter 10 provides m-factor tables that define the performance level limits for each com-
ponent type. The m-factor values vary between 1 and 2 for IO performance level and
between 2 and 7 for other performance levels, depending on the component type and char-
acteristic as well as the performance level. The TBI Guidelines and the LATBSDC
Procedure have a single factor of 1.5 for all materials and component types in a Risk
Category II building. Table 5 in Appendix 1 summarizes the ffactors and material prop-
erties recommended in each document to calculate the capacity for deformation-controlled
actions at serviceability performance level.

Force-controlled actions. ASCE 41 distinguishes between LDP demand calculations for

deformation- vs force-controlled actions, although this distinction is absent in the TBI
Guidelines and the LATBSDC Procedure. ASCE 41 suggests two methods to calculate a
force-controlled action in a component. The first method is computing the action using the
limit-state analysis. This method considers the maximum action that can be delivered to a
component, based on the expected strengths of the surrounding components. However,
this method is not sensitive to the hazard level under consideration and may be unrealistic
for the IO at the 50%/30-year hazard level. The alternative method is to compute the
action as the summation of the action caused by gravity and earthquake forces, where the
earthquake portion is reduced by C1 C2 J and amplified by x depending on the performance
level, that is, QUF = QG + (xQE =C1 C2 J ) (note, this equation applies for linear procedures,
and an additional g-factor is applied depending on the criticalness of the member in non-
linear procedures). For buildings with T . 1.0 s, C1 C2 = 1:0, and for IO, J = 1. x is equal
to 1.0 for lower than Life Safety performance or 1.3 otherwise. Therefore, for the purposes
of this comparison, the force-controlled demand is again the same among the three docu-
ments at the higher hazard of evaluation.
2166 Earthquake Spectra 37(3)

When calculating force-controlled capacities, ASCE 41 continues to use a ffactor of

1.0 but uses the nominal, or lower bound (mean minus 1 standard deviation), rather than
expected strength. The TBI Guidelines make no distinction between deformation- and
force-controlled actions and continue to use nominal strengths and ffactors from the
material codes. The LATBSDC Procedure also keeps the same capacities as for deforma-
tion-controlled, using expected strengths and f = 1:0.
ASCE 41 does not apply m-factors for the force-controlled criteria. Without an m-fac-
tor, the requirement is simply that the demand must not exceed the capacity. The TBI
Guidelines use a factor of 1.0. Finally, the LATBSDC Procedure reduces the capacity by a
factor of 0.7. The comparison between the capacity of the component modified by the
capacity factors calculated using the three documents depends on the action type and the
corresponding ffactor. For instance, for the axial force in a rectangular concrete col-
umn, the LATBSDC Procedure computes the highest capacity, and the capacity computed
per TBI Guidelines is marginally lower than ASCE 41. Table 5 in Appendix 1 summarizes
the ffactors and material properties recommended in each document to calculate the
capacity for force-controlled actions at serviceability performance level.

Nonlinear dynamic response history analysis. In cases where the LDP acceptance criteria are
not satisfied, the designer can make a second assessment using NDP before modifying the
design. It should be noted that the use of NDP for serviceability assessment is currently
not common in practice. For NDP analysis, the component demands are determined in
the same manner for both deformation- and force-controlled actions. As discussed in the
analysis section, the demand is either the maximum or the mean response achieved from
the ground motion suite, depending on the number of ground motions used.

Deformation-controlled actions. The NDP acceptance criteria for deformation-controlled

actions are similar among the three documents. ASCE 41 provides tables for nonlinear
modeling and acceptance criteria based on the component type and the desired perfor-
mance level. The TBI Guidelines and the LATBSDC Procedure both discuss using testing
data to define deformation limits but permit the use of the IO criteria from the ASCE 41
tables.

Force-controlled actions. ASCE 41 uses the same capacity calculations for force-controlled
actions, independent of the analysis procedure used. As with LDP, the capacity is based
on nominal strengths and f = 1:0. When using NDP, the TBI Guidelines suggest basing the
expected capacity on laboratory testing data. If not using laboratory testing, the capacity
can be increased from the LDP assessment by using expected strength and the correspond-
ing ffactor. Appendix B of the LATBSDC Procedure defines a set of factors to account
for conservatism in nominal strength of various actions in different structural components.
Table 5 in Appendix 1 summarizes the ffactors and material properties recommended
in each document to calculate the component’s capacity for force- and deformation-
controlled actions using NDP at serviceability performance level.

Collapse prevention acceptance criteria

The performance objective for a new Risk Category II building at the MCER hazard level
is to satisfy the CP. The analysis requires NDP, and the demands are determined from the
Sattar et al. 2167

response history analyses of the ground motion suite, as discussed in the analysis section.
The ASCE 41 acceptance criteria only consider the component-level analysis. The TBI
Guidelines and the LATBSDC Procedure include additional global acceptance criteria to
limit permanent drift and structural instability. In addition, the triggers for identifying an
unacceptable response differ among the three documents. According to ASCE 41, an
unacceptable response occurs if (1) analysis fails to converge, or (2) estimated demand of a
deformation-controlled action exceeds the valid range of modeling, or (3) estimated
demand of a force-controlled critical action model elastically exceeds expected capacity, or
(4) estimated deformation demand of components not explicitly modeled exceeds the
deformation limit at the loss of gravity load capacity. TBI Guidelines have six triggers for
an unacceptable response. The first four match ASCE 41’s triggers with a minor modifica-
tion that TBI Guidelines add the ordinary actions to the third trigger in ASCE 41 and also
do not specify how the force-controlled actions should be modeled. In addition to the four
triggers identified in ASCE 41, TBI Guidelines have two global criteria that state that if
peak transient or residual interstory drift ratios in any story exceed 4.5% and 1.5%,
respectively, the response would be unacceptable. The LATBSDC Procedure has three
triggers, including the first and fourth triggers identified by ASCE 41. The second and
third triggers are combined into one, stating that if the estimated demand in deformation-
controlled or force-controlled actions exceeds valid range of modeling, the response would
be unacceptable. It should be noted that, unlike the TBI Guidelines, most of the accep-
tance criteria in the LATBSDC Procedure (except the global acceptance criteria) are evalu-
ated either for the maximum average response of all ground motions or a factor times the
maximum average response.

Global acceptance criteria. As in the serviceability acceptance criteria, the TBI Guidelines and
the LATBSDC Procedure have the same limits on story drift. There is a limit on both the
mean and maximum from the response history analyses of the ground motion suite. The
mean peak transient story drift must be \0.03, whereas the maximum must be \0.045,
that is, unacceptable response. The peak limit ensures that the system and detailing will
perform as intended, nonstructural components will not threaten life safety, the gravity
system is protected, and P-Delta effects are controlled for vertical loads. The residual drift
is also limited so that the building will not be condemned after an earthquake. The mean
and maximum residual drift should be \0.01 and \0.015, respectively.

Deformation-controlled actions. ASCE 41 uses the same process for the NDP acceptance cri-
teria at CP as it does for IO, but with larger allowable deformation limits. The deformation
demand determined from the ground motion suite analysis is compared to the CP accep-
tance criteria found in the tables in Chapter 10. If a component demand exceeds the accep-
tance criteria, the component design would need to be revised. In a similar situation, the
TBI Guidelines state that if the deformation demand in a component for any response his-
tory analyses exceeds the valid range of modeling, the stability of the structure should be
checked by assuming a negligible strength for the component for the remaining of the anal-
ysis or considering the analysis to have an unacceptable response. One of the key differ-
ences between TBI Guidelines and the other two documents is that TBI allows the use of
the ultimate deformation capacity.
The acceptance criteria in LATBSDC Procedure are defined as the plastic or total rota-
tion capacity of the component divided by the importance factor. However, larger defor-
mations can be permitted if supported by appropriate laboratory tests or approved by the
peer-review process. The LATBSDC Procedure also recommends a set of deformation
2168 Earthquake Spectra 37(3)

limits (in terms of strain or rotation) for reinforced concrete walls, coupling beams, and
slab outrigger beams. The LATBSDC Procedure does not specifically discuss checking the
stability of the structure as no unacceptable response is allowed in the document.

Force-controlled actions. Another difference among the three documents is the classification
of force-controlled actions. The NDP force-controlled acceptance criteria for ASCE 41 are
similar to those of LDP, except for the added g for critical elements. The demand from the
response history analyses must not exceed the capacity, as calculated using the nominal
strength and a ffactor of 1.0. The TBI Guidelines, however, delineate among critical,
ordinary, and noncritical force-controlled actions. If failure of a force-controlled action
would likely lead to partial or total structural collapse, it is considered critical. If the fail-
ure might lead to local collapse, it is ordinary, and if it is unlikely that the failure of a com-
ponent leads to structural collapse, the action is considered as noncritical. The LATBSDC
Procedure classifies the force-controlled actions into critical not sensitive to vertical accel-
erations, critical sensitive to vertical accelerations, and ordinary.
The TBI Guidelines and the LATBSDC Procedure address the more severe consequences
of a critical action by inflating the mean seismic demand. The TBI Guidelines suggest a 1.3
factor for this increase and the LATBSDC Procedure suggests 1.3 and 1.5, depending on
whether nominal or expected material properties are used for strength calculations. The
TBI Guidelines and LATBSDC Procedure use f-factors from the material codes for critical
force-controlled components. Both documents use f = 0:9 for ordinary actions. The TBI
Guidelines use f = 1:0 for noncritical actions. The TBI Guidelines and the LATBSDC
Procedure are similar in how they handle ordinary actions. Table 6 in Appendix 1 sum-
marizes the f-factor and material properties recommended in each document to calculate
the capacity of force- and deformation-controlled actions at CP performance level. The
three documents have different equations for calculating the acceptance criteria for force-
controlled actions. The TBI Guidelines and LATBSDC Procedure consider bias factors
when computing the acceptance criteria of force-controlled actions accounting for conser-
vatism in nominal resistance calculated in accordance with the applicable material stan-
dards. The LATBSDC Procedure is the only document that provides a set of bias factor
values for different components and actions.

Conclusion
This article provides a detailed comparison of the three standards/guidelines on the imple-
mentation of the PBSD framework, including ASCE 41-17, TBI, and LATBSDC for Risk
Category II RC structures. This study compares the three documents in terms of differ-
ences in the intent of the development, modeling recommendations, and analysis and
acceptance criteria for each document.
All three documents achieve the performance objective through the assessment of the
response at an MCER and at more frequent hazard levels, that is, 475-year return period in
ASCE 41 and 72-year return period in TBI and LATBSDC. This article shows that modeling
recommendations among the three documents are different in terms of expected material
properties, component stiffness, damping ratio, and inclusion of components in the model.
The TBI Guidelines and LATBSDC Procedure allow the use of force–deformation backbone
curves in ASCE 41 for various components, but also provide more modeling options per rec-
ommendations in ATC 72. The TBI Guidelines also allow the use of the force–deformation
backbones in NIST (2017a). In general, the deformation capacity in ATC 72 and the NIST
guidelines is, in many cases, higher than ASCE 41. In addition to the modeling requirements,
Sattar et al. 2169

the analysis and acceptance criteria of three documents do not match. In TBI and
LATBSDC’s ground motion selection processes, the average or maximum response of the
components is compared against the acceptance criteria, depending on the number of ground
motions used in the analysis for serviceability evaluation, while for the CP evaluation the
average response is compared against the acceptance criteria. When ASCE 41 is employed as
the evaluation standard, the average response of 11 ground motions needs to be compared
against the acceptance criteria. ASCE 41 allows one unacceptable response out of 11 for life
safety performance level and lower; however, no unacceptable response is permitted for dam-
age control performance level and higher. The TBI Guidelines allow one unacceptable
response for Risk Category II buildings. Both ASCE 41 and TBI require 120% of the med-
ian response (not less than average) to be used when a ground motion produces unacceptable
response. The LATBSDC Procedure does not permit any unacceptable response.
The computation of component demand using linear procedure at the 72-year hazard level
among the three documents is similar. However, there are multiple discrepancies in the calcu-
lation of capacity. In the calculation of the capacity for force-controlled actions using response
spectrum analysis, TBI Guidelines use nominal material properties and ffactor from mate-
rial codes, and capacity factor of 1.0 in force-controlled actions. However, LATBSDC
Procedure uses expected material properties with f = 1 and reduces the strength a 0.7 factor,
whereas ASCE 41 uses nominal material properties and f = 1. The same discrepancy in the
use of ffactor and material properties exists for capacity calculation of deformation-
controlled actions using response spectrum analysis. Moreover, ASCE 41 uses m-factors as a
capacity modifier that ranges between 1.0 and 2 for IO performance level, whereas TBI and
LATBSDC use a constant factor of 1.5 for Risk Category I and II buildings.
The demand calculation for force- and deformation-controlled actions using nonlinear
dynamic analysis is similar among the three documents at the 72-year hazard level. The
capacity calculation for deformation-controlled actions is the same, but for force-
controlled actions is different in terms of ffactors and material properties.
At the MCER hazard level, the computation of demand in the deformation-controlled
actions is different among the documents in terms of the use of mean versus maximum
response. Similarly, there are differences in the capacity calculations. For instance,
LATBSDC Procedure defines the acceptance criteria of deformation-controlled actions as
the plastic or total rotation capacity of the component divided by the importance factor of
the building. The TBI Guidelines allow the use of capacity deformation prescribed in ASCE
41. The TBI Guidelines define the ultimate component deformation based on recommenda-
tion in ASCE 41, ATC 72, and NIST (2017a), but does not require stopping the analysis if
one component exceeds the ultimate deformation. For the force-controlled actions, multiple
discrepancies exist in the calculation of both demand and capacity. In addition to the
component-level acceptance criteria, TBI and LATBSDC have global acceptance criteria in
terms of the peak transient drift, residual drift, and loss in the story strength. These criteria
are not present in ASCE 41.
This study reveals that, despite the similarities that exist between ASCE 41, TBI, and
LATBSDC—especially between TBI and LATBSDC—the three standards/guidelines dif-
fer at multiple levels, from developing the simulation models, calculating component
demand and capacity, and defining the acceptability of the building performance.
Although the three documents are developed with different intentions and may target dif-
ferent building types, they are being used in practice to design new buildings. The differ-
ences between these standards and guidelines can potentially lead to different building
2170 Earthquake Spectra 37(3)

designs and different resiliency of buildings designed per these standards/guidelines as

shown by Buniya et al. (2020) for a 12-story reinforced concrete special moment frame.
Future studies should investigate the impacts of implementing different PBSD guidelines
and standards in the design of other structural systems. The inconsistency among the
PBSD guidelines may imply a need to investigate the sources of the differences in the
implementation of PBSD and ideally to unify these differences to the extent possible if the
same resiliency is intended regardless of the location and standard.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/
or publication of this article.

Funding
The author(s) received no financial support for the research, authorship, and/or publication of this
article.

References
ACI (2014) ACI 318: Building Code Requirements for Reinforced Concrete. Farmington Hills, MI:
ACI.
ASCE/SEI (2006) Seismic Rehabilitation of Existing Buildings. ASCE/SEI 41-06. American Society
of Civil Engineers, Reston, VA.
ASCE/SEI (2017) Seismic rehabilitation of existing buildings. ASCE/SEI 41-17. Reston, VA:
American Society of Civil Engineers.
Buniya M, Barbosa A and Sattar S (2020) Assessment of a 12-story code-compliant reinforced
concrete special moment frame using performance-based seismic engineering standards and
guidelines: ASCE 41, TBI, and LATBSDC. ACI SP-339. 3rd January 2020. Farmington Hills,
MI: American Concrete Institute, pp. 134–154.
Burks LS and Baker JW (2016) A predictive model for fling-step in near-fault ground motions based
on recordings and simulations. Soil Dynamics and Earthquake Engineering 80: 119–126.
FEMA (1997) NEHRP guidelines for the seismic rehabilitation of buildings. FEMA 273, October.
Washington, DC: Federal Emergency Management Agency.
FEMA (2000) Prestandard and commentary for the seismic rehabilitation of buildings. FEMA 356,
November. Washington, DC: Federal Emergency Management Agency.
ICC, 2009, International Building Code. International Code Council, Inc. Country Club Hills, Illinois.
Kamai R, Abrahamson N and Graves R (2014) Adding fling effects to processed ground-motion
time histories adding fling effects to processed ground-motion time histories. Bulletin of the
Seismological Society of America 104: 1914–1929.
LATBSDC (2014) An Alternative Procedure for Seismic Analysis and Design of Tall Buildings Located in
the Los Angeles Region. Los Angeles Tall Buildings Structural Design Council, Los Angeles, CA.
LATBSDC (2020) An Alternative Procedure for Seismic Analysis and Design of Tall Buildings Located
in the Los Angeles Region. Los Angeles, CA: Los Angeles Tall Buildings Structural Design
Council.
Moehle JP (2014) Seismic Design of Reinforced Concrete Buildings. New York: McGraw-Hill
Education.
NIST (2016). Seismic design of cast-in-place concrete diaphragms chords, and collectors: A guide for
practicing engineers, 2nd ed. GCR 16-917-42, NEHRP Seismic Design Technical Brief No. 3,
Applied Technology Council for the National Institute of Standards and Technology,
Gaithersburg, MD.
NIST (2017a) Guidelines for nonlinear structural analysis for design of buildings, Part IIb—Concrete
Moment Frames. No. NIST GCR 17-917-46v3, April. Gaithersburg, MD: National Institute of
Standards and Technology.
Sattar et al. 2171

NIST (2017b) Recommended modeling parameters and acceptance criteria for nonlinear analysis in
support of seismic evaluation, retrofit, and design. No. NIST GCR 17-917-45, April. Gaithersburg,
MD: National Institute of Standards and Technology.
PEER/ATC (2010) Modeling and acceptance criteria for seismic design and analysis of tall buildings.
No.PEER 2010/111, October. Redwood City, CA: Applied Technology Council.
Rezaeian S, Bozorgnia Y, Idriss IM, Abrahamson N, Campbell K and Silva W (2014) Damping
scaling factors for elastic response spectra for shallow crustal earthquakes in active tectonic
regions: ‘‘average’’ horizontal component. Earthquake Spectra 30: 939–963.
TBI (2010) Guidelines for Performance-based Seismic Design of Tall Buildings. Pacific Earthquake
Engineering Research Center, Berkeley, CA.
TBI (2017) Guidelines for Performance-Based Seismic Design of Tall Buildings. Berkeley, CA: Pacific
Earthquake Engineering Research Center.
Wilson EL, DerKiureghian A and Bayo EP (1981) A replacement for the SRSS method in seismic
analysis. Earthquake Engineering & Structural Dynamics 9: 187–192.

Appendix 1

Table 3. High-level comparison of PBEE standards/guidelines for new buildings

ASCE 41 TBI LATBSDC

Scope Developed for evaluation Developed for design of Developed for design of
of existing buildings structures with: buildings with hn . 160 ft
All building types  T .. 1 s Includes modifier, ki , for
Includes nonstructural  Significant mass other Risk Categories
components participation and Could be used for other
lateral response at building types
higher modes Considers nonstructural
 Slender aspect ratio components via
for seismic force– referenced Building Code
resisting system prescriptions and story
May be applicable, with drift limits
modifications, to
structures that do not
exhibit significant inelastic
response or in other
seismic regions
Level of analysis Component-level Component-level Component-level
acceptance criteria acceptance criteria with acceptance criteria with
Structural-level story additional drift limits and additional drift limits and
drifts for nonstructural stability checks for base shear checks for
acceptance criteria structural-level analysis structural-level analysis
Hazard levels  For BPON: MCER and  50%/30 (43-yearr  50%/30 (43-year
2/3 MCER return period) return period)
 MCER per ASCE 7-10  MCER per ASCE 7-10
Performance  BPON: Life Safety and  Serviceability  Serviceability
objectives nonstructural Position (essentially elastic) (essentially elastic)
Retention for 2/3  Collapse Prevention  Collapse Prevention
MCER, Collapse
Prevention for
MCE_R
ASCE: American Society of Civil Engineers; TBI: Tall Buildings Initiative; BPON: Basic Performance Objective
Equivalent to New Building Standards; LATBSDC: Los Angeles Tall Buildings Structural Design Council;
MCER: risk-targeted maximum considered earthquake.
2172 Earthquake Spectra 37(3)

Table 4. Comparison of recommended shear effective stiffnesses

Component type ASCE 41 TBI and LATBSDCa Ratio w.r.t. ASCE 41b
MCER 50%/30 MCER 50%=30
ASCE 41 ASCE 41

RC beams (non-prestressed) 0:4Eexp

c Aw 0:4Eexp
c Aw 0:4Eexp
c Aw
A2g Ag 2
Aw Aw
RC beam (prestressed) 0:4Eexp
c Aw – – – –
Coupling beam 0:4Eexp
c Aw 0:4Eexp
c Aw 0:4Eexp
c Aw
A2g A2g
Aw Aw
RC columns 0:4Eexp
c Aw 0:4Eexp
c Aw 0:4Eexp
c Aw
A2g A2g
Aw Aw
Flat slab (non-prestressed) 0:4Eexp
c Aw – – – –
Flat slab (prestressed) 0:4Eexp
c Aw – – – –
Non-PT transfer diaphragm (in-plane)c – 0:4Eexp
c Aw 0:4Eexp
c Aw – –
PT transfer diaphragm (in-plane)c – 0:2Eexp
c Aw 0:4Eexp
c Aw – –
Structural walls (in-plane) 0:4Eexp
c Aw 0:2Eexp
c Aw 0:4Eexp
c Aw
A2
0:5 Awg 1:0 Awg
A2

Basement walls – 0:2Eexp

c Aw 0:4Eexp
c Aw – –
ASCE: American Society of Civil Engineers; TBI: Tall Buildings Initiative; RC: reinforced concrete; PT: post-tensioned;
LATBSDC: Los Angeles Tall Buildings Structural Design Council; MCER: risk-targeted maximum considered
earthquake.
a
LATBSDC requires a different equation for fc0 ø6000 psi. Ratios are computed assuming nominal fc0 = 5000 psi.
b
For beams, coupling beams, walls, and columns, ASCE 41 uses Aw for gross area of cross section, whereas TBI and
LATBSDC use Ag. Ag =Aw can be considered as 1.0.
c
TBI and LATBSDC recommend considering lower and upper bound stiffnesses for considering the backstay effects
for the MCER level analysis.

Table 5. Serviceability acceptance criteria

ASCE 41 TBI LATBSDC

Global Limits None Story drift \0.5% Story drift \0.5%

Response spectrum analysis

(a) Deformation-controlled actions
Capacity  Strength: expected  Strength: nominal  Strength: expected
 f = 1:0  f = 1:0  f = 1:0
m-factors  See ASCE 41 Chapter 10  1.5  1.5
For concrete, 1 \ m \ 7
(b) Force-controlled actions
Capacity  Strength: lower bound  Strength: nominal  Strength: expected
 f = 1:0  f from material  f = 1:0
codes
Capacity factors  none  1.0  0.7

Nonlinear dynamic response analysis

(a) Deformation-controlled actions
Capacity  See ASCE 41  ASCE 41 IO  ASCE 41 IO
for IO criteria criteria criteria
(b) Force-controlled actions
Capacity  Strength: nominal  Strength: based on  LATBSDC
 f = 1:0 laboratory testing Procedure does not
data or provide specific
 Strength: expected guidancea
 f = 1:0
ASCE: American Society of Civil Engineers; TBI: Tall Buildings Initiative.; LATBSDC: Los Angeles Tall Buildings
Structural Design Council; IO: Immediate Occupancy.
a
It is common to use the serviceability acceptance criteria from the elastic response spectrum analysis.
Sattar et al. 2173

Table 6. Collapse prevention acceptance criteria

ASCE 41 TBI LATBSDC

(a) Deformation-controlled actions

Capacity  See CP criteria from  ASCE 41 CP criteria  Plastic or total rotation
Tables in ASCE or valid range of modeling capacities from ASCE 41
41 Chapter 10 divided by importance
factor
(b) Force-controlled actions
Critical actions
Capacity  Strength: nominal  Strength: expected  Strength: expected
 f = 1:0  f from material codes  f from material codes
g-factor
Non-critical actions
Capacity  Strength: nominal  Strength: expected  N/A
 f = 1:0  f = 1:0
Ordinary actions
Capacity  Strength: nominal  Strength: expected  Strength: expected
 f = 1:0  f = 0:9  f = 0:9
ASCE: American Society of Civil Engineers; TBI: Tall Buildings Initiative; LATBSDC: Los Angeles Tall Buildings
Structural Design Council; CP: Collapse Prevention.