ATC-19

Structural Response Modification Factors

by
APPLIED TECHNOLOGY COUNCn..
555 Twin Dolphin Drive, Suite 550
Redwood City, California 94065

Funded by
NATIONAL SCIENCE FOUNDATION
Grant No. ECE-8600721
and
NATIONAL CENTER FOR EARTHQUAKE ENGINEERING RESEARCH
NCEER Project No. 92-4601

PRINCIPAL INVESTIGATOR
Christopher Rojahn

PROJECT CONSULTANTS
Andrew Whittaker
Gary Hart

PROJECT ENGINEERING PANEL

Vitelmo Bertero
Gregg Brandow
. Sigmund Freeman
William Hall
Lawrence Reaveley*
*ATC Board Representative

1995

__
---_.-. .. _ .
Preface
In 1986, the Applied Technology Council (ATC) was
awarded a grant from the National Science Founda-
tion (NSF) to evaluate structural response modifica-
tion factors (R factors). R factors are used in current
seismic building codes to reduce ground motions as-
sociated with design level earthquakes to design force
levels. The initial objectives of the project (known as
ATC-19) were to: (1) document the basis for the val-
ues assigned to R factors in model seismic codes in
the United States, (2) review the role played by R fac-
tors in seismic design practice throughout the United
States; (3) present state-of-knowledge on R factors;
and (4) propose procedures for improving the reliabil-
ity of values assigned to R.
In 1991, the scope of the effort was expanded with
funding from the National Center for Earthquake En-
gineering Research (NCEER) to address and/or docu-
ment (1) how response modification factors are used
for seismic design in other countries; (2) a rational
means for decomposing R into key components using
state-of-the-knowledge information; (3) a framework
(and methods) for evaluating the key components of
R; and (4) the research necessary to improve the reli-
ability of engineered construction designed using R
factors. The results from the original and expanded
objectives described above are documented in this re-
port.
The primary ATC-19 project consultants, who pre-
pared the major portions of this report, were Gary
Hart and Andrew Whittaker, senior-level earthquake
engineering researchers from southern and northern
California, respectively. Their work was overviewed
and guided by an advisory "blue-ribbon" Project En-
ATC-19 Preface
gineering Panel (PEP) consisting ofVitelmo Bertero,
Gregg Brandow, Sigmund Freeman, William Hall,
and Lawrence Reaveley (ATC Board Representa-
tive). Nancy Sauer and Peter Mork provided editorial
and publication preparation assistance. The affilia-
tions of these individuals are provided in the Project
Participants list.
ATC gratefully acknowledges the valuable support
and patience of the NSF Project Officer, S. C. Liu.
ATC also gratefully acknowledges the valuable input
of participants in the companion NCEER-funded
ATC-34 Project: The late Peter Gergely (Cornell
University), who served on the NCEER Research
Committee and played a key role in acquiring
NCEER support for this investigation; Project Direc-
tor Andrew Whittaker (University of California at
Berkeley); PEP members Vitelmo Bertero (Universi-
ty of California at Berkeley), Ian Buckle (NCEER),
Sigmund Freeman (Wiss, Janney, Elstner Assoc.,
Inc.), Gary Hart (University of California at Los An-
geles), Helmut Krawinkler (Stanford University),
Ronald Mayes (Dynamic Isolation Systems), Andrew
Merovich (Andrew Merovich & Assoc.), Joseph
Nicoletti (URSlBlume), Guy Nordenson (Ove Arup
& Partners), Masanobu Shinozuka (University of
Southern California), and John Theiss (ATC Board
Representative); and consultants Howard Hwang
(Memphis State University), Onder Kustu (OAK En-
gineering), and Yi-Kwei Wen (University ofIl1inois).
Christopher Rojahn
ATC Executive Director &
ATC-19 Principal Investigator
III
._---_.-._ __ _-- .. .. ......
 Table of Contents
Preface iii

1. Introduction 1
1.1 Background 1
1.2 Objectives of the Report 2
1.3 Organization of the Report : 2

2. History of Response Modification Factors 5

2.1 Introduction 5
2.2 R Factor Development 5
2.3 Rw Factor Development 8
2.4 Comparison of K, R, and Rw 8

3. Use of Response Modification Factors 11

3.1 Introduction '" 11
3.2 R Factors in SeisnUc Building Codes 11
3.2.1 Europe 11
3.2.2 Japan 11
3.2.3 Mexico 13
3.2.4 Summary 14
3.3 Use of R Factor Equivalents for Bridge Design 15

4. Components of Response Modification Factors 17

4.1 Introduction 17
4.2 Impact of the R Factor on Design 17
4.3 Force-Displacement Response of Buildings 18
4.4 Experimental Evaluation of Force-Displacement Relationships 20
4.5 Key Components of R 21
4.5.1 Strength Factor 22
4.5.2 Ductility Factor 23
4.5.3 Redundancy Factor 27
4.5.4 DaIllping Factor 29
4.6 Systematic Evaluation of R Factors 31
4.7 Reliability of Values for R 32

5. Conclusions and Recommendations 33

5.1 Summary and Concluding Remarks 33
5.2 Recommendations ; 34

Appendix A: Evaluation of Building Strength and Ductility 35

Appendix B: Glossary of Terms 49

References 51

Project Participants 53

Applied Technology Council Projects and Report Information 55

ATC-19 Table of Contents v

-------_.----
1. Introduction
1.1 Background
The seismic design of buildings in the United States
is based on proportioning members of the seismic
framing system for actions determined from a linear
analysis using prescribed lateral forces. Lateral force
values are prescribed at either the allowable (work-
ing) stress or the strength level. The Uniform Build-
ing Code (lCBO, 1991) prescribes forces at the
al10wable stress level and the NEHRP Recommended
Provisions for the Development ofSeismic Regula-
tionsfor New Buildings, hereafter denoted as the
NEHRP Provisions (BSSC, 1991) prescribes forces
at the strength level. The seismic force values used in
the design of buildings are calculated by dividing
forces that would be associated with elastic response
by a response modification factor, often symbolized
asR.
Response modification factors were first proposed by
the Applied Technology Council (ATC) in the ATC-
3-06 report published in 1978. The NEHRP Provi-
sions, first published in 1985, are based on the seis-
mic design provisions set forth in ATC-3-06. Similar
factors, modified to reflect the allowable stress
design approach, were adopted in the Uniform Build-
ing Code (UBC) a decade later in 1988.
The concept of a response modification factor was
proposed based on the premise that well-detailed
seismic framing systems could sustain large inelastic
deformations without collapse (ductile behavior) and
develop lateral strengths in excess of their design
strength (often termed reserve strength). The R factor
was assumed to represent the ratio of the forces that
would develop under the specified ground motion if
the framing system were to behave entirely elasti-
cally (termed hereafter as elastic design) to the pre-
scribed design forces at the strength level (assumed
equal to the significant yield level).
In the UBC, gravity (dead, live, and snow) and envi-
ronmental (wind, seismic) loads are prescribed at the
service level. Until the recent advent of Load and
Resistance Factor Design (LRFD), an ultimate
strength approach that is just beginning to be used in
practice, steel framing systems have typically been
ATC-19 1: Introduction
designed for service-level actions using allowable
stresses. Reinforced concrete framing systems are
designed for ultimate strength-level actions, which
are calculated by multiplying the service-level
actions by load factors. Prescribed seismic forces are
calculated in the UBC by dividing the elastic spectral
forces by a response modification factor (Rw ): values
for Rw range between 4 and 12.
In the NEHRP Provisions, loads are prescribed at the
strength level. In practice, steel framing systems are
designed for ultimate-level actions by using al1ow-
able stress values multiplied by 1.7; reinforced con-
crete framing systems are designed at the strength
level for ultimate actions. Prescribed seismic forces
are calculated in the NEHRP Provisions by dividing
the elastic spectral forces by a response modification
factor, R. Values for R range between 1.25 and 8. The
relationship between the response modification fac-
tors in the NEHRP Provisions (R) and the UBC (Rw )
is presented later in this report.
When using response modification factors substan-
tially greater than one, the designer makes a signifi-
cant assumption; that is, that linear analysis tools can
be used to obtain reasonable estimates of nonlinear
response quantities. This assumption has recently
been questioned and is discussed in detail in ATC-34
(ATC, 1995).
Use of large response modification factors underlies
a second common assumption of seismic design; that
is, that significant nonlinear response and hence sig-
nificant damage is expected if the design earthquake
occurs. This assumption is of course a direct result of
using design forces that are significantly less than the
elastic spectral forces. The consequences of this
assumption are considered in detail in this report.
The R factors for the various framing systems
included in the ATC-3-06 report were selected
through committee consensus on the basis of (a) the
general observed performance of like buildings dur-
ing past earthquakes, (b) estimates of general system
toughness, and (c) estimates of the amount of damp-
ing present during inelastic response. Thus, there is
little technical basis for the values of R proposed in
1
-----.---_.__ - .. .
ATC-3-06. The commentary to ATC-3-06 notes
that " ... values of R must be chosen and used with
judgement" and that " ... lower values must be used
for structures possessing a low degree of redundancy
wherein all the plastic hinges required for the forma-
tion of a mechanism may be formed essentially
simultaneously and at a force level close to the speci-
fied design strength." To further underscore the
uncertainties associated with the values assigned to R
for different seismic framing systems, a footnote to
the table listing the response modification coeffi-
cients states, "These (values for R) are based on best
judgment and data available at time of writing and
need to be reviewed periodically."
Given the fiscal and social consequences of wide-
spread building failure that could occur in an earth-
quake if poor choices for values of R are used in
design, it is evident to enlightened design profession-
als that the values assigned to R in current seismic
regulations should reflect the most current knowl-
edge in earthquake engineering and construction
practice in the United States. Nearly twenty years
have passed since R factors were first introduced in
the United States. In this space oftime, much has
been learned about the likely performance of seismic
framing systems in moderate-to-severe earthquakes,
especially following the 1989 Lorna Prieta and 1994
Northridge earthquakes. This new knowledge, com-
bined with changing public expectations of accept-
able levels of earthquake-induced damage and
changes in construction practice, makes 1995 an
appropriate year in which to publish a formal review
of response modification factors and the ways in
which the factors are used (and misused) in current
design practice.
1.2 Objectives of the Report
This report has several key objectives.
1. To document the basis for the values assigned
to R in current seismic codes in the United
States.
2. To review the role played by R factors in
seismic design practice in the United States.
3. To describe how response modification fac-
tors are used for seismic design in other
countries.
4. To present up-to-date information on R fac-
tors.
2 1: Introduction
_---------,---
.
5. To develop a rational means of decomposing
R into key components.
6. To propose a framework (and methods) for
evaluating the key components of R.
7. To recommend research necessary to
improve the reliability of engineered con-
struction designed using R factors.
The primary audience for this report is licensed pro-
fessional engineers familiar with both current build-
ing seismic design criteria and structural dynamics.
However, the report has been written to be under-
standable to a broad audience, with the intent ofhav-
ing a strong impact on the design professionals and
the code-change process. The secondary audience for
the report is the academic/research community.
1.3 Organization of the Report
Chapter 2 provides an historical perspective on how
the values of R in use today were developed. The
relationship between K factors introduced in the late
1950 s, R factors introduced in ATC-3-06, and R w
factors introduced into the 1988 UBC (ICBO, 1988)
is established, and the shortcomings of seismic
design using R factors are enumerated.
Chapter 3 discusses the use of response modification
factors for the seismic design of new buildings out-
side the United States and for the seismic design of
new bridges in the United States, to provide perspec-
tive on the conclusions drawn in this report. The fac-
tors used in three common framing systems, the
European, Japanese, and Mexican codes, are com-
pared with the corresponding values in the 1991
NEHRP Provisions (BSSC, 1991). This chapter
includes some conclusions about the likely behavior
ofcode-compliant buildings in the United States dur-
ing severe earthquake shaking.
Chapter 4 discusses the impact of R factors on the
seismic design process in the United States, experi-
mental estimates of R for two steel-braced framing
systems, and proposes a new formulation for R.
Unresolved issues associated with the proposed for-
mulation for R are described, and strategies for
resolving these issues are proposed.
In Chapter 5, the significant issues raised in this
report are summarized, and key conclusions are
drawn. Recommendations for further study complete
this chapter. A list of references follows Chapter 5.
ATC-19
,-----_.-._---_..........
 Reliable values for R will likely be proposed on the
basis ofthe statistical evaluation of reserve strengths
and system ductility values. Reserve strength and
ductility can be estimated using nonlinear static anal-
ysis. Appendix A provides an overview of nonlinear
static analysis and presents the results of such an
analysis of a nonductile reinforced concrete moment
frame building. This analysis was performed as part
of the ongoing FEMA-funded Building Seismic
Safety Council (BSSC) project to develop guidelines
and commentary for the seismic rehabilitation of
ATC-19 1: Introduction
-,-r---,----------'-.-----
buildings (ATC' s portion of this project is known as
ATC-33). The results of the nonlinear static analysis
presented in Appendix A are used to calculate draft
strength and ductility factors.
Appendix B contains a comprehensive glossary of
tenns used in this report. Following Appendix B are
references, a list of the individuals who have contrib-
uted to the preparation of this report, and infonnation
on other available ATC reports, including companion
reports and other resource documents.
3
,----_._,_.. .......
_,
_L--
2. History of Response Modification Factors
2.1 Introduction
In 1957, a committee of the Structural Engineers
Association of California (SEAOC) began develop-
ment of a seismic code for California. This effort
resulted in the SEAOC Recommended Lateral Force
Requirements (also known as the SEAOC Blue Book)
being published in 1959 (SEAOC, 1959). Commen-
tary to the requirements was first issued in 1960.
These recommendations represented the profession's
state-of-the-art knowledge in the field of earthquake
engineering; the seismic design requirements in the
1959 Blue Book were significantly different from
previous seismic codes in the United States. For the
first time the calculation of the minimum design base
shear explicitly considered the structural system
type. The equation given for base shear was
v= KCW (2-1)
where K was a horizontal force factor (the predeces-
sor of R and Rw ); C was a function of the fundamen-
tal period of the building; and W was the total dead
load. The K factor was assigned values of 1.33 for a
bearing wall building, 0.80 for dual systems, 0.67 for
moment-resisting frames, and 1.00 for framing sys-
tems not previously classified. The term C defined
the shape ofthe design response spectrum, and was
calculated as follows:
C = 0.05 (2-2)
3..11'
where T was the fundamental period of vibration in
the direction under consideration. The Blue Book was
developed as a seismic design code for California
alone. California was assumed to have uniform seis-
micity, and a seismic zone factor was not required in
Equation 2-1. The intent of the Blue Book was to:
"... provide minimum standards to assure
public safety. Requirements contained in
such codes are intended to safeguard
against major structural failures and to
provide protection against loss of life
and personal injury... The "Recom-
ATC-19 2: History of Response Modification Factors
mended Lateral Force Requirements"
are intended to provide this protection in
the event of an earthquake of intensity or
severity of the strongest of those which
California has recorded ... The code does
not assure protection against non-struc-
tural damage ... Neither does it assure
protection against structural damage ..."
The seismic provisions in the 1961 UBC OCBO,
1961) were adopted from the 1959 Bluebook. Seis-
mic zonation was considered through the use of a Z
factor which was set equal to 1.0 in zone 3 (the
region of highest seismicity), 0.50 in zone 2, and
0.25 in zone 1. The minimum design base shear in
the 1961 UBC was calculated as:
v= ZKCW (2-3)
where all terms were defined as above.
2.2 R Factor Development
The development of response modification (R) fac-
tors, first introduced in ATC-3-06 (ATC, 1978), can
be traced back to the horizontal force factor (Rojahn,
1988, and Rojahn and Hart, 1988). This section sum-
marizes the development process.
The publication of ATC-3-06 defined a benchmark in
seismic engineering in the United States. ATC-3-06
constituted a significant departure from previous
seismic codes and embodied several new concepts
that included: (a) classification of building use-group
categories into seismic hazard exposure groups, (b)
national seismic hazard maps, (c) tools for elastic
dynamic analysis, (d) use of response modification
(R) factors in lieu of K factors, (e) explicit drift lim-
its, (f) discussion of the influence of orthogonal exci-
tation effects, (g) materials design based on strength
methods instead of allowable stress, (h) provisions
for soil-structure interaction, and (i) detailed seismic
design requirements for architectural, electrical, and
mechanical systems and components. In regard to
response modification factors, ATC-3-06 noted that:
5
._-----.-._-_._~._" ... -_._ .
 1. . R factors were intended to reflect reductions in
design force values that were justified on the
basis of risk assessment, economics, and nonlin-
ear behavior.
2. The intent was to develop R factors that could be
used to reduce expected ground motions pre-
sented in the fonn of elastic response spectra to
lower design levels by bringing modem struc-
tural dynamics into the design process. Figure
2.1 illustrates the use of R factors to reduce elas-
tic spectral demands to design force levels.
Given that R was to be a response reduction factor, it
was decided to place R in the denominator of the
base shear equation. The end result was that R factors
were inversely proportional to the K factors used in
previous codes. The base shear equation for struc-
tures for which the period of vibration ofthe building
T was not calculated took the following form:
2.5Aa
v= -yw (2-4)
In this expression, V is the seismic base shear force,
Aa is the effective peak. acceleration of the design
ground motion (expressed as a fraction of g), R is the
response modification factor, and W the total reactive
weight The factor of2.5 is a dynamic amplification
factor that represents the tendency for a building to
amplify accelerations applied at the base.
Only horizontal seismic forces were considered in
ATC-3-06 for two reasons. First, buildings had
always been designed to withstand vertical forces
greater than those produced by mean (unfactored)
gravity loads, thereby providing assumed reserve
capacity for vertical seismic motions, and second,
because the analysis and design tools needed to
account for vertical ground motion effects were not
routinely available in the 1970s. Furthermore, ground
motion data available at the time suggested that peak
vertical motions were normally less than 2/3 of
peak horizontal motions, leading to the conclusion
that the responses caused by vertical motions should
be less severe than those caused by horizontal
motions.
For structures for which the fundamental building
period was calculated, the base shear equation in

--
.....
C)
ATC 3-06 elastic response spectrum for a
c:
0 rock site and 5% damping
r<l
CD
CD
0
0
r<l

-...
Cii 1
0
CD
Q.
Design spectrum for a special
moment-resisting space frame
UI (R =8)
"tI
CD
.!:!
Cii
...0E ;-----------
Z / ------ --------------_ ..
0
Period (seconds)

Figure 2. 1: Use of R factors to reduce elastic spectral demands to the design force level. Each point on the
elastic response spectrum for a rock site (top curve) is divided by R to produce the design spec-
trum (bottom curve) for a given structure type, in this case a special moment-resisting space frame.
where R= 8.

6 2: History of Response Modification Factors ATC-19

_.-....---------_._--' ---.----r--------.-.--" .
ATC-3-06 was given as:
1.2A y S
V = Ry{).67 w (2-5)
In this expression, A y is the effective peak velocity-
related acceleration, S is a soil profile coefficient, and
T is the fundamental period of the building. The soil
profile coefficient is used to account for soil proper-
ties that could amplify the bedrock motion; its val-
ues, as defined in ATC-3-06, range from 1.0 to 1.5.
The base shear of Equation 2-4 provides an upper
limit on the base shear calculated using Equation 2-5.
1.67) ( V) (2-6)
Vw ( 1.33 = 0.9
where Vw was the allowable-stress design lateral
seismic base shear (1976 UBC) and V was the
strength- design lateral seismic base shear (ATC 3-
06). The numerical factors in Equation 2-6 accounted
for differences between the allowable-stress design
and strength design methods: 1.67 represented the
margin of safety in allowable-stress design, 1.33 rep-
resented the pennissible increase in allowable-stress
design, and 0.9 was the capacity reduction factor for
flexure in strength design.

Individuals who participated in the ATC-3-06 R fac- Using the expression for Vw as specified in the 1976
tor development process (ATC, 1978, page 8, Struc- UBC (leBO, 1976), it followed that:
tural Design, Details, and Quality Assurance
Committee) have indicated that committee members ZIKCS. w(1.67)
I 1.33
= 0.9Ry{).67
(1.2AS) W (2-7)
first independently developed R values for each
structural system type based on their own experience. where Z was a zone factor, I was an importance fac-
The values of R selected for inclusion in ATC-3-06 tor, K was a horizontal force factor, C defined the
represented the consensus opinion of the experts spectral shape ( 1/( 15 Jh
,and Si was a site
involved in the development effort. coefficient.
The first step in assigning consensus R values was to Substituting Z = 1= T= 1.0, Si = 1.5, Ay = 0.4, and S
set a maximum value of R for the structure types con- = 1.2 in Equation 2-7:
sidered to provide the best seismic perfonnance; that
is, those with the highest reserve strength or ductility.
(1.0)(1.0)K(0.067)(1.5) ( 1.67)
1.33 (2-8)
This category included special moment frames and
dual systems composed of reinforced concrete shear (1.2 x 0.4 x 1.2)
= "--~~.-:-;:~....;..
wall structures with special moment frames capable 0.9R(1.0)
of resisting at least 25 percent of the prescribed seis-
resulted in
.mic forces.
C.W. Pinkham (personal communication), a member (0.1256)K = 0~4 (2-9)
of the team that developed R factors, described the
procedure used to calculate R for special steel
yielding
moment frames.
The maximum value of R was determined by equat- R = 5.1 (2-10)
ing Vw computed for allowable stress design per the
K
1976 UBC (equivalent to the 1974 Blue Book In the 1976 UBC, K was set equal to 0.67 for
(SEAOC, 1974) to V computed for strength design moment resisting frame systems. The corresponding
in ATC-3-06. Implicit in this undertaking was the value of R in ATC-3-06 was thus computed as:
decision not to increase the design base shear to
improve seismic performance, but rather to achieve 5.1 8
R = 0.67= . (2-11)
improved seismic performance by requiring better
detailing.
The response modification factor for reinforced con-
For special steel moment frames, the maximum value crete shear-wall structures with special moment
of R was computed at a fundamental period equal to frames was also assigned the maximum value of
1.0 second: eight. Values of R for other framing systems were

ATC-19 2: History of Response Modification Factors 7

-----------------------_.._._---_ -_ .. .......
generally assigned on the basis of Equation 2-10, where S is a site coefficient and T is the fundamental
then adjusted in accordance with the consensus opin- period of vibration.
ion of the committee. Framing systems not consid-
ered in the 1976 UBC were assigned R values by If it is assumed that CS = 0.14 and Z = 1 in Equation
consensus opinion of the committee. 2-12, and that C = 2.75 and Z= 0.4 in Equation 2-13,
it follows that
2.3 Rw Factor Development
K(0.14)C = (2.75)0.4 (2-15)
Values for structural response modification factors Rw
for allowable-stress design (RwJ were determined by
and that
the Seismology Committee of the Structural Engi-
neers Association of California (SEAOC) and pub-
lished in the 1988 Blue Book (SEAOC, 1988). (2-16)
SEAOC elected to introduce Rwo rather than R, to
ease the eventual transition from allowable-stress Substituting Equation 2-1 o into Equation 2-16 yields
design to strength design. the following relationship between Rw and R
Similar to R, Rw is inversely proportional to K. The
relationship between values of K in the 1985 UBC
Rw = 7.86
5.1
R = 1.54R
. . (2-17)
and values of R w in the 1988 UBC can be demon-
strated as follows. Table 2.1 displays the values of K (/985 UBCJ and
Rw (1988 UBC) for several framing systems.
The equation given in the 1985 UBC (lCBO, 1985)
for calculating the design base shear at the allowable
stress level (VD ) is: Table 2.1 Relationship of K and Rw

Framing
VD = (ZIKC5)W (2-12) System 1985 USC 1988 USC

The parameters Z and I are used to quantify the seis-

Bearing wall K = 1.33 Rw= 6
mic zone and the importance of the building occu- Dual steel and
concrete
K = 0.80 Rw = 10
pancy, respectively. The parameter S is used to
account for site characteristics, and C is a numerical Ductile steel
coefficient that is a function of the fundamental and concrete
K = 0.67 Rw = 12
period of vibration of the building and the defmed = 1.00
spectral shape. The maximum value of C is set equal
Other K Rw= 8
to 0.12; the maximum value of CS is set at 0.14. K is
a numerical coefficient referred to as the horizontal
2.4 Comparison of K, R. and R w
force factor. With few exceptions, the R factors tabulated in ATC-
The 1988 Blue Book (SEAOC, 1988) and 1994 UBC 3-06 are the same as those in the 1991 NEHRP Pro-
(lCBO, 1994) use an alternative equation for calcu- visions. The exceptions include an increased value of
lating VD , namely: R in the NEHRP Provisions for special concrete
moment-resisting space frames, and the addition of R
factors for concrete intermediate moment-resisting
(2-13)
space frames. The values assigned to R w in the 1994
UBC are the same as those listed in the 1988 Blue
where Z and I are the seismic zone and importance Book. For reference, values ofR in ATC-3-06 and the
factors, respectively. For this example, let! = 1 in 1991 NEHRP Provisions, and R w in the 1994 VBC
Equations 2-12 and 2-13. The factor C in Equation 2- are listed in Table 2-2, for framing systems grouped
13 has a maximum value of2.75 and is defmed as: according to K value.

C = 1.255 (2-14) The link between K, R, and Rw was established in the

ro 67 previous section. Values of the horizontal force fac-

8 2: History of Response Modification Factors ATC-19

----_.. _-----~"'-" ..
tor K represented the consensus opinion of expert ratio cannot be used to uniformly reduce elastic
design professionals and academicians in the late spectral demands to design (inelastic) spectral
1950s. Despite a many-fold increase in knowledge demands (measured typically as base shear), R
and the advent of powerful analysis tools, no sub- must be period-dependent. This dependence is
stantive review of, or changes to, response reduction recognized in the Eurocode and the Mexican
factors have been made since the 1950s. Code (see Chapter 3 for further discussion).
Recent studies by researchers (e.g., Bertero, 1986)
4. The reserve strength (strength in excess of the
and design professionals, including Project ATC-34,
design strength) of buildings designed in differ-
have identified major shortcomings in the values and
ent seismic regions will likely vary substantially.
formulation of the response modification factors used
Given that reserve strength is a key component
in seismic codes in the United States. These short-
of R (see Chapter 4), R should be dependent on
comings include the following:
either the seismic zone or some ratio of gravity
loads to seismic loads.
1. A single value of R for all buildings of a given
framing type, irrespective of building height,
5. Seismic design using the response modification
plan geometry, and framing layout, cannot be
factors listed in seismic codes and guidelines in
jus,tified.
the United States will most probably not result in
a uniform level of risk for all seismic framing
2. The use of the values assigned to R for some
systems.
framing systems will likely not produce the
desired performance in the design earthquake. These shortcomings and other related issues are
addressed in the remainder of this report.
3. The response modification factor is intended, in
part, to account for the ductility of the framing
system. Recognizing that a constant ductility

ATC-19 2: History of Response Modification Factors 9

--------,-----------1----
Table 2.2 Tabulated Values for K, Rand Rw
Basic Structural System (K factor) R R Rw
V\TC, 1978) rosse, 1991) (leBO, 1994)
Bearing Wall System (K=1.33)
1. Light Framed Walls with Shear Panels 6.5 6.5
a. Plywood walls, 3 stories or less 8
b. All other light framed walls 6
2. Shear walls
a. Concrete 4.5 4.5 6.0
b. Masonry 3.5 3.5 6.0
3. Braced Frames Carrying Gravity Loads 4.0 4.0
a. Steel 6
b. Concrete 4
Building Frame System (K=1.00)
1. Steel Eccentric Braced Frames (EBF) 7.0-8.0 10.0
2. Concentric Braced Frames 7.0
3. Shear Walls
a. Concrete 5.5 5.5 8.0
b. Masonry 4.5 4.5 8.0
Dual System (K=O.80)
1. Shear Walls
a. Concrete with Special Moment Resisting Space
Frame (SMRSF) 8.0 8.0 12.0
b. Concrete with Concrete Intermediate Moment
Resisting Space Frame (IMRSF) 6.0 9.0
c. Masonry with Concrete SMRSF 6.5 6.5 8.0
d. Masonry with Concrete IMRSF 5.0 7.0
2. Steel EBF with Steel SMRSF 7.0-8.0 12.0
3. Concentric Braced Frames
a. Steel with SMRSF 6.0 6.0 10.0
b. Concrete with Concrete SMRSF 6.0 6.0 9.0
c. Concrete with Concrete IMRSF 5.0 6.0
Moment Resisting Frame System (K=O.67)
1. Special Moment Resisting Space Frames (SMRSF)
a. Steel 8.0 8.0 12.0
b. Concrete 7.0 8.0 12.0
2. Concrete Intermediate Moment Resisting Space
4.0 8.0
Frames (IMRSF)
3. Ordinary Moment Resisting Space Frames
a. Steel 4.2 4.5 6.0
b. Concrete 2.0 2.0 5.0

10 2: History of Response Modification Factors ATC-19

3. Use of Response Modification Factors
3.1 Introduction
The use of response modification factors is not
restricted to the seismic design of buildings in the
United States. R factors, or their equivalents, are used
for the seismic design of buildings in Europe, Japan,
Mexico, and other countries, and for the seismic
design of bridges in the United States. This chapter
reviews the use of R factors in seismic building codes
in Europe, Japan, and Mexico, and seismic bridge
codes in the United States, in order to place seismic
design practice for buildings in the United States in
perspective.
3.2 R Factors in Seismic Building Codes
3.2.1 Europe
The seismic design procedure in the 1988 Eurocode
(CEC, 1988) is a single-level design procedure that
reduces elastic spectral demands to the strength
design level through the use of a period-dependent,
response reduction factor (q) as follows:
for T < T} T steel allowable stress equal to the yield stress.
1 + l1(Tl~O -1)
q = (3-1)
T and explicit evaluation of strength and ductility, and
1 + Tl (Tl ~O/ q - 1)
for T> T}
(3-2)
q=q
where T is the fundamental period ofthe building; Tl
is a characteristic period of the design spectrum
(lower-bound period to the constant-acceleration por-
tion of the spectrum); 11 is a factor related to the sys-
tem equivalent viscous damping ~ and equal to 1.0
for ~ equal to five percent of critical; ~O is a
pseudo-acceleration spectrum amplification factor
(set equal to 2.5); and q is a system behavior factor
that varies as a function of material type, building
strength and stiffness regularity. Values for q range
between one and five for reinforced concrete framing
systems. Recognizing that ductility cannot be used to
ATC-19 3: Use of Response Modification Factors
reduce substantially elastic force demands in the
short-period range from 0 to Tl , equation 3.1 shows
how q varies from q = q at T= Tl' to q=1.0 at
T=O.O.
Inelastic displacement values (ds ) are estimated in
the 1988 Eurocode as the product of the displacement
values (de) computed using the reduced (design)
seismic forces and the behavior factor q. For T less
than Tl' the ratio q/q exceeds 1.0 and the inelastic
displacement values exceed the elastic displacement
values; for T greater than T}, the ratio q/q equals
1.0, and the inelastic displacement values equal the
elastic displacement values.
3.2.2 Japan
The Japanese 1981 Building Standard Law (IAEE,
1992) includes a two-phase or two-level procedure
for the seismic design of buildings. The first phase
(Level I) design follows an approach similar to that
adopted in the NEHRP Provisions (BSSC, 1991).
Steel structures are designed at the strength level,
based on allowable stress design procedures with the
Strength design is used for reinforced concrete struc-
tures. The second phase (Level II) design is a direct
may be regarded as a check of whether these are suf-
ficient for severe ground motions. Timber structures
and low-rise structures satisfying rigidity, eccentric-
ity, and detailing limitations do not require Level II
design. Other structures, including all structures
between 31 and 60 meters high, are subject to both
Level I and Level II design. Normative practice is for
the seismic framing system to be designed using the
Level I procedure and for the Level I design to be
checked (and modified as necessary) using the Level
II procedure. Structures over 60 meters high are sub-
ject to special approval by the Ministry of Construc-
tion.
In the Level I design, the seismic coefficient at each
story ( Cj ) is determined as the product of four vari-
ables:
(3-3)
11
._----_._-_._ - .
where Z represents the seismic zone, Rt defmes the
spectral shape that varies as a function of soil type,
Ai defines the vertical distribution of seismic force
in the building, and Co represents the peak ground
acceleration. In regions of high seismicity, Z is equal
to 1.0. Except for wood structures on soft subsoil,
Co is set equal to 0.2. The seismic design shear force
in the i-th story (Qi ) is calculated as:
(3-4)
where W is the reactive weight above the i-th story.
For Level I design, seismic actions are computed
using unreduced seismic forces. Interstory drift is
limited to 0.5 percent of the story height for the pre-
scribed seismic forces unless it can be demonstrated
that greater drift can be tolerated by the nonstructural
components, in which case the drift limit can be
increased to 0.8 percent of the story height.
In Level II design, the engineer checks plan eccen-
tricity, distribution oflateral stiffness, minimum code
requirements (in some cases), and ultimate lateral-
load-carrying capacity of each story. The ultimate
lateral load capacity is computed using plastic analy-
sis and ultimate seismic demands are estimated as:
(3-5)
where Qud is the lateral seismic shear for severe
earthquake motions, calculated according to Equa-
tion 3-4 using Co equal to 1.0, Ds is a framing sys-
tem-dependent ductility factor (less than 1.0), and
Table 3.1 Coefficient Ds for Steel Framed Buildings in Japan's 1981 Building Standard law
Behavior of Members (1) Ductile moment
frame
A. Members with
0.25
excellent ductility
B. Members with
0.30
good ductility
C. Members with fair 0.35
ductility
D. Members with 0.40
poor ductility
12 3: Use of Response Modification Factors
Fes is a measure of the regularity of the building.
There is no displacement check in the Level II
design.
The regularity factor (Fes ) is calculated as:
(3-6)
where Fe is a measure of the plan irregularity of the
building, and Fs reflects the unifonnity of the distri-
bution of lateral stiffness over the height ofthe build-
ing. For reference, Fe and Fs range in value between
1.0 (regular) and 1.5 (most irregular). The design
penalties associated with selecting a highly irregular
seismic framing system are clearly evident.
The ductility factor (Ds ) varies as a function ofstruc-
tural material, type of framing system, and key
response parameters. Materials are identified as
either steel or reinforced concrete; steel-reinforced
concrete is included under the heading of reinforced
concrete. Table 3-1 displays values of Ds for steel
seismic framing systems from the 1981 Building
Standard Law (BSL). These values range between
0.25 and 0.50. The "behavior of members" rating in
the first column is based on the proportioning of the
structural members. For example, members in ductile
moment frames with excellent ductility have smaller
width-to-thickness (or depth-to-thickness) ratios than
members in ductile moment frames with/air ductility
or poor ductility. Stocky bracing members in braced
frames are associated with excellent ductility and
slender braces are associated with/air ductility.
Type of Frame
(2) Concentrically (3) Frames other than
braced frame (1) and (2)
0.30
0.35
0.40 0.35
0.45 0.40
0.50 0.45
ATC-19
,-----_._-- -
Table 3.2 Coefficient D s for Reinforced Concrete Frame Buildings in Japan's 1981 Building Standard Law
Behavior of Members (1) Ductile moment
frame
A. Members with
0.30
excellent ductility
B. Members with
0.35
good ductility
C. Members with fair
0.40
ductility
D. Members with
0.45
poor ductility
For reinforced concrete construction, values for Ds
vary between 0.3 and 0.55, as shown below in Table
3.2. For steel-reinforced concrete construction
(termed composite construction in the United States),
values for D s are reduced from those in the table by
0.05. For a reinforced concrete ductile moment frame
to be assigned excellent ductility, columns have to be
designed to be flexure-critical, have a longitudinal
reinforcement ratio less than 0.8 percent, and have
low axial 0.35fc) and shear O.lfc') stresses at
the fonnation of the mechanism. The limiting shear
stress in beams in an excellent ductility frame is
0.15fc' . Poor ductility would be assigned to a
moment frame in which the axial and shear stress
values in the columns are much higher than the limits
noted above, and for frames incorporating shear-crit-
ical beams or columns. For a shear wall to posses
excellent ductility, the wall has to be flexure-critical
and have a low shear stress O.lfc ) at the formation
of the mechanism. The reader is referred to the
Tables Cl through C4 (reinforced concrete construc-
tion), and 01 through 04 (steel construction) in the
1981 Building Standard Law for more detailed infor-
mation on frame and ductility classifications.
3.2.3 Mexico
The 1987 Mexico City Building Code uses a reduc-
tion factor to reduce elastic spectral demands to a
strength design level. The response reduction factor
(Q' ) is period-dependent and is calculated as fol-
lows:
ATC-19 3: Use of Response Modification Factors
Type of Frame
(3) Frames other than
(2) Shear walls
(1) and (2)
0.40 0.35
0.45 0.40
0.50 0.45
0.55 0.50
for T< TA:
(3-7)
Q' = Q (3-8)
where T is the fundamental period of the building;
TA is a characteristic period of the design spectrum
(lower bound period to the constant acceleration por-
tion of the spectrum); and Q is a system behavior fac-
tor that varies as a function of material type, building
strength and stiffness regularity (Gomez and Garcia-
Ranz, 1988). Values for Q range between 1.0 and
4.0. The Mexico City Building Code recognizes that
ductility cannot be used to reduce elastic force
demands substantially in the short-period range from
oto TA by reducing Q' from Q at period TA , to 1.0
at a period of 0.0 second.
Inelastic displacement values are estimated in the
Mexico City Building Code as the product of the
response reduction factor Q and the displacement
values computed using the reduced seismic forces.
When T is less than TA , the ratio Q/ Q' exceeds 1.0
and the inelastic displacement values exceed the elas-
tic displacement values; when T is greater than TA ,
the ratio Q/Q' equals 1.0, and the inelastic displace-
ment values equal the elastic displacement values.
13
Table 3.3 Response Modification Factor Comparison for Rock Sites
Structural System Period Europe a
RC Structural Wall T= 0.1 sec. 2.0
T= 1.0 sec. 3.5
RC Moment Frame T = 0.1 sec. 2.3
T = 1.0 sec. 5.0
Steel Moment T = 0.1 sec. 2.5
Frame
T = 1.0 sec. 6.0
a. T 1 equal to 0.2 second, TI equal to 1.0, ~o
b. Inverse of D s '
c. TA equal to 0.2 second.
3.2.4 Summary
The application of response modification factors (or
their equivalents) to the seismic design of buildings
in Europe (1988 Eurocode), Japan (1981 Building
Standard Law), and Mexico (1987 Mexico City
Building Code) has been reviewed. In order to draw
broad conclusions about seismic design practice in
the United States from this information, consider
three ductile framing systems of regular configura-
tion, all located on rock sites: (l) a reinforced con-
crete structural wall, (2) a reinforced concrete
moment-resisting space frame, and (3) a steel
moment-resisting space frame. Assume that analysis
at fundamental periods of 0.1 second and 1.0 second
is sufficient for the purpose of comparison. The
response modification factors determined from each
code are presented in Table 3.3 together with values
of R from the 1991 NEHRP Provisions. Of the seis-
mic codes being compared, only the 1981 BSL does
not use a response modification factor to reduce elas-
tic spectral demands to a strength (first significant
yield) design level. Therefore, Table 3-3 lists values
for the inverse of Ds .
Table 3-3 shows that the response modification fac-
tors used in the NEHRP Provisions are substantially
greater than the corresponding values in the Eur<r
pean and Mexican codes. Given that construction
standards in Europe and Mexico are likely compara-
ble with those in the United States, substantial
numerical differences between the response modifi-
cation factors, and assuming similar standards of
quality control and structural details, the substantial
14 3: Use of Response Modification Factors
Japan b Mexico c United States
2.5 2.5 5.5
2.5 4.0 5.5
3.3 2.5 8.0
3.3 4.0 8.0
4.0 2.5 8.0
4.0 4.0 8.0
equal to 2.5.
differences in the response modification factors sug-
gest that a building designed according to the
NEHRP Provisions will likely suffer more damage in
the design earthquake than similar buildings
designed for the requirements of either the Eurocode
or the Mexico City Building Code.
It is interesting to note that the response modification
factors in the European and the Mexican codes do not
account for reserve strength; that is, the factors in
these two codes are intended to be a measure of duc-
tility alone. This is in contrast to the NEHRP Provi-
sions wherein the values assigned to R are intended
to account for reserve strength and ductility.
A direct comparison of the values assigned to
response modification factors in the NEHRP Provi-
sions (R) and the Japanese BSL (1/ Ds) is difficult
because the factors are used differently. In the
NEHRP Provisions, R is used to reduce elastic forces
to the strength (first significant yielding) level for
design. In the BSL, the factor is used in the Level II
procedure to reduce elastic forces for comparison
with the maximum strength of a framing system
(often designed using the Level I procedure). This
maximum strength, computed using either nonlinear
static analysis or plastic analysis, may exceed the
design strength at first significant yielding by
upwards of 100 percent (see Chapter 4 for additional
information). Assuming that the maximum strength
of most framing systems in the U.S. is two to three
times the design strength, and similar elastic spectral
demands, and recognizing that values of R exceed
those of 1/Ds by a factor of between two and three,
ATC-19
---------------------_ __ .
the framing systems resulting from U.S. and Japa-
nese practice will likely be similar.
Inelastic displacement values are calculated in the
NEHRP Provisions as the product of the elastic dis-
placement values computed using the reduced seis-
mic forces and a displacement amplification factor
that is numerically smaller than the response modifi-
cation factor. The calculated inelastic displacement
values are thus smaller than the elastic displacement
values computed using the unreduced seismic forces.
A different procedure is used by the European and
Mexican codes wherein inelastic displacement values
are calculated as the product of the displacement val-
ues computed using the reduced seismic forces and a
displacement amplification factor equal to or larger
than the response reduction factor. The resulting
inelastic displacement values are equal to or greater
than the elastic displacement values computed using
the unreduced seismic forces. The European and
Mexican procedures for computing inelastic dis-
placements are more consistent with the results of
recent research (Miranda and Bertero, 1994) than the
procedure adopted in the NEHRP Provisions. The
reader is referred to Report ATC-34 (ATC, 1995) for
additional information on the calculation of inelastic
displacements.
3.3 Use 01 R Factor Equivalents for
Bridge Design
In 1982, ATC published the ATC-6 (ATC, 1982a)
report entitled Seismic Design Guidelines for High-
way Bridges. These guidelines were developed by a
team of nationally recognized bridge engineering
experts. The format of ATC-6 paralleled that of
ATC-3. In particular, it introduced R factors to
reduce elastic spectral demands to a strength design
level. The ATC-6 report recommended different val-
ues of R for framing elements and connections; the
values for R being smaller for connections to pro-
mote plastic hinging in the framing elements and to
preclude inelastic behavior in the connections. As
such, the ATC-6 design methodology for bridges dif-
fered from the ATC-3 design methodology for build-
ings in which one value for R was used for the entire
building.
The Caltrans Bridge Design Specification (Caltrans,
1990) makes use of a period-dependent response
reduction factor, which is termed an adjustment fac-
tor for ductility and risk assessment and denoted as Z.
The Z factors are used to reduce elastic spectral
demands to strength-design actions, so Z performs a
similar role to R. Figure 3.1 presents Caltrans Z fac-
tors as a function of period and structure/component
type. The reduction in values of Z with increasing
period is based in part on the increase in spectral dis-
placements with increasing period. For slender col-
umns, large displacements may result in significant
second-order (or P-b.) effects.

8
"~
..... ~
...
.....
N 6 ......
r... I - Well-confined ductile multi-column bents

~t'<
0
U

..
ell
II.
c 4
............... Well-confined ductile single column bents
..
Gl
E
III "'r -
:::l Piers, abutment walls, and wingwalls
:c
c(
2
I - Hinge restrainer cables (Z=1.0)
r
\ Well-reinforced concrete shear keys (Z=O.8)
o
o 1.0 2.0 3.0
Period (seconds)

Figure 3.1: Caltrans Z factors.

ATC-19 3: Use of Response Modification Factors 15

,-----_.-_..-_ .
The Z factors for single- and multi-column bents are
constant for periods less than 0.6 second, and
decrease linearly between periods of 0.6 and 3.0 sec-
onds. The period-dependent trends for Z in the short-
period range are not supported by analytical studies
(Miranda and Bertero, 1994). In particular, although
Z tends to decrease with increasing period, strength
reduction due to inelastic behavior is minimal for
very stiff structures, and tends to increase with
increasing period.
Table 3.4 shows the values of Caltrans Z factors and
ATC-6 R factors (for a period of 0.3 second) for
bridges founded on rock. Caltrans defines the seismic
hazard at a bridge site in terms of the maximum cred-
ible earthquake whereas AASHTO defines the seis-
mic hazard using probabilistic techniques based on a
1O-percent probability of being exceeded in 50 years.
This difference in the definition of the design earth-
quake is responsible for the larger response reduction
factors used in the Caltrans procedure, because use of
either the Caltrans or the AASHTO procedure is
intended to produce columns ofa similar size (Ian
Buckle, personal communication).
Project ATC-32 is currently reviewing Caltrans' seis-
mic design procedures for bridges. Improved Z fac-
tors are being developed that depend on (a) bridge
importance, (b) structure-to-site period ratio, and (c)
element type (column, pier, or connection). Values
for Z factors for ordinary and important bridge struc-
tures are reported in Table 3.4. These improved Z
factors are intended to be used with elastic analysis
results that consider the stiffness degradation that
will occur during a major seismic event and flexural
capacities that consider probable rather than nominal
material strengths. The net result of the proposed
ATC-32 design procedures for ductile components is
that most design quantities will often be only nomi-
nally different than those for current Caltrans
designs. The ATC-32 recommendations have not
been formally adopted to date by Caltrans.

Table 3.4 Bridge Response Modification Factors

ATC-3?
Ordinary Important
Frame type Caltrans Z ATC-6 R Bridges Bridges
Single-column bent 6 3 :5>3
Multiple-column bent ::;3

a. Proposed, not yet adopted

16 3: Use of Response Modification Factors ATC-19

------,--------- ------_._.---_ .
4. Components of Response Modification Factors

4.1 Introduction 4.2 Impact of the R Factor on Design

The commentary to the 1988 NEHRP Provisions
(BSSC, 1988) defines the R factor as " ... an empiri-
cal response modification (reduction) factor intended
to account for both damping and ductility inherent in
a structural system at displacements great enough to
approach the maxim um displacement ofthe system."
This definition provides some insight into the devel-
opers' qualitative understanding of the seismic
response of buildings and the expected behavior of a
code-compliant building in the design earthquake.
The components of R can be dermed in several ways,
each dependent on the performance level under con-
sideration. In this report, only the life-safety perfor-
mance level is considered explicitly. Section 4.2
provides a framework for a discussion on the disag-
gregation of R into its primary components by dis-
cussing how R is used to link elastic and inelastic
response. Section 4.3 introduces some key issues
associated with describing the force-displacement
response of a building (expanded on by example in
Appendix A). Finally, sections 4.4 and 4.5 address
the disaggregation of R into its key components.
The key parameters influencing the response of an
elastic single-degree-of-freedom (SDOF) system are
indicated in Figure 4.1, which illustrates a single-
story moment frame with massless columns. The
floor mass m is attached to the ground by two elastic
columns (springs) of combined lateral stiffness k.
Damping c is introduced by a dashpot linking the
floor and the ground. The SDOF in this model is the
horizontal translation of the floor with respect to the
ground. The inertial force developed by the floor
mass during earthquake shaking is a function of the
properties of the SDOF system (m, k, and c) and the
characteristics of the earthquake ground motion. For
an elastic SDOF system, seismic actions and dis-
placements can be determined using an earthquake
response spectrum - the envelope of the maximum
responses of SDOF oscillators to one earthquake
ground motion. Response spectra vary widely in fre-
quency content and amplitude. For reference, the
pseudo-acceleration spectra (Clough and Penzien,
1993) corresponding to the 1 Centro, SCT, Sylmar,
and JMA earthquake ground motions are presented in
Figure 4.2. These earthquake histories were recorded
during the 1940 Imperial Valley, 1985 Mexico City,
1994 Northridge, and 1995 Hyogoken-Nanbu earth-
quakes, respectively.

Floor mass, m

Column stiffness, Dashpot. c Column stiffness,

kJ2 k/2

Figure 4.1: Single degree of freedom system.

ATC-19 4: Components of Response Modification Factors 17

 3r------.---...,.------,.----.---,....-----.----r-----,
\
EI Centro 1940
't \
2.5
,.\\ Sylmar 1994
,I.
'
, \'J\' . ------ \orMA 1995
: '
. \:
' .: SCT 1985
c 2 ... i}' t ~:~;.:. ~ ~"""""""':"""""""

.
.Q
~ f f" \: : :
".~ : :
Q)
a;
8 1.5 I \ \
co J \:
I
o I . \: \ .
-0
~ 1 I ,J \.. ., ..\ :
t
: .
I ....: ..... :;.~._ :: :
, ,.
lI)
a.. . '~: ..... :
"
..... .', ~........ -.......... ..
.... .. - : '0_;-. ~ ...... _
I::~..~.~..L..-_~i__:=======~==.=-:::.-:.~.'~-~.~.~-;-~.-~.~-
OL
o 0.5 1 1.5 2 2.5 3 3.5
. -.
4
..
Period (seconds)

Figure 4.2: Sample elastic pseudo-acceleration spectra for 5% damping.

The impact of R on the seismic design of buildings is
clearly seen by comparing the equations for the
design base shear for inelastic response (Equation 2-
4) and the base shear for elastic response (Ve):
(4-1)
where Se,5 is the elastic 5-percent damped pseudo-
acceleration response spectral ordinate calculated at
the fundamental period of the building; and W is the
reactive weight, equal to Mg for the simple structure
depicted in Figure 4.1. Note that W in Equation 4-1 is
the total reactive weight and not the reactive weight
in the fundamental mode. For seismic design in the
United States, the spectrum has in the past generally
corresponded to an earthquake ground motion with a
10-percent probability of being exceeded in 50
years 1, which is often tenned the design earthquake.
The elastic spectral ordinate in Equation 4-1 is equiv-
alent to the term 2.5A a in Equation 2-4. Equations 2-
1. It is anticipated that the 1997 NEHRP Recommended Provi-
sions for the Seismic Regulation ofNew Buildings, under
development by the Building Seismic Safety Council (BSSC),
and the NEHRP Guidelinesfor Seismic Rehabilitation of
Buildings, under development by ATC (ATC-33 project) for
BSSC, will incorporate seismic hazard maps that reflect longer
recurrence intervals (i.e., ground motions have a 2% probabil-
ity of being exceeded in 50 years).
4 and 4-1 are thus identical if the response modifica-
tion factor R in Equation 2-4 is equal to 1.0. In prac-
tice, the design base shear (for inelastic response) is
calculated by dividing the base shear for elastic
response by the response modification factor R, a
value that generally varies between 4.0 and 8.0. The
substantial difference between the ordinates of elastic
and design base shear spectra is clearly seen in Fig-
ure 2-1.
4.3 Force-Displacement Response of
Buildings
A typical force-displacement relationship for a build-
ing frame is shown in Figure 4.3. This relationship
describes the response of the building frame sub-
jected to monotonically increasing displacements.
For the purposes of design, this nonlinear relation-
ship is often approximated by an idealized bilinear
relationship. Two bilinear approximations are widely
used and these methods are described below. Either
of these methods can be used to estimate yield forces
and yield displacements; the two methods will gener-
ally produce similar results for most ductile framing
systems.
The first approximation, developed for characteriz-
ing the load-displacement relation for reinforced con-
crete elements (pauley and Priestley, 1992), assumes

18 4: Components of Response Modification Factors ATC-19

_--------_._--
.
._---_.----------------.---_._.- - _ .
 Ductile response
->
' y - Brittle response

f--
, ,
1---''
r
~
""'--
v

Roof displacement (d)

Figure 4.3: Sample base shear force versus roof displacement relationship.

Base shear force (V) Base shear force (V)

Actual response

Vy Vy

.75Vy

ti.y 6m , 6m
Displacement (~) 0 Isplacement (d)
(a) Paulay & Priestley (b) Equal Energy

Figure 4.4: Bilinear approximations to a force-displacement relationship.

a priori knowledge of the yield strength (Vy) of the
frame. The elastic stiffness is based on the secant
stiffness of the frame calculated from the force-dis-
placement curve at the force corresponding to
O.75Vy The determination of the elastic stiffness (K)
is shown in Figure 4.4a.
The second method used to approximate the force-
displacement relation of a frame is commonly termed
the equal-energy method. This method assumes that
the area enclosed by the curve above the bilinear
approximation is equal to the area enclosed by the
curve below the bilinear approximation. This bilinear
approximation is illustrated in Figure 4.4b.
The nonlinear relationships presented in Figure 4.4
are described by the yield force (Vy ), yield displace-
ment (~y), maximum force (Va), displacement corre-
sponding to a limit state (~m)' and the displacement
immediately prior to failure (~u), Displacements ~m
and d u are well beyond the yield displacement for
ductile framing systems. The elastic stiffness (Ko) is
calculated by dividing the yield force by the yield
displacement. The post-yield stiffness (K 1) is com-
monly defmed as a fraction (a) ofthe elastic stiffness
as follows:

ATC.19 4: Components of Response Modification Factors 19

,--------.,-----------'-_ _.- .. ...., ..... .,._ ..


(4-2)
where all tenns in this equation are defined above.
The ability of a building frame to be displaced
beyond the elastic limit, while resisting significant
. force and absorbing energy by inelastic behavior, is
tenned ductility. Displacement ductility is defmed as
the difference between t:.m and t:.y- The maximum
displacement ductility is the difference between t:. u
and t:.y . The displacement ductility ratio is generally
defined as the ratio of t:.m to t:.Y' namely:
t:. m
~~ = (4-3)
t:. y
where t:.m is always greater than t:.y- Brittle failures
are characterized by negligible ductility. Brittle fail-
ures of this type are common in older construction
built before the advent of ductile detailing in the mid-
1960s.
The force-displacement relationship for a building
can be determined either experimentally or analyti-
cally. Experimental evaluation is difficult, extremely
costly, and therefore rare. Pseudo-dynamic testing of
full-scale buildings, and earthquake simulator testing
of small- to moderate-scale models of buildings,
have provided force-displacement relations for build-
ings of different scales. The use of earthquake simu-
lator test data to evaluate the force-displacement
response of a building is described in Section 4.4
below.
Nonlinear finite element analysis software is a cost-
effective analytical tool used by academicians and
design professionals to estimate force-displacement
relationships for buildings. Nonlinear static (or
pushover) analysis, somewhat routine in the larger
architectural/engineering firms in Japan since the
early 1980s, is now being promoted in the ATC-33
project as the preferred analysis method for seismic
rehabilitation projects. For reference, a description of
nonlinear static analysis, together with a sample anal-
ysis of a seven-story reinforced concrete building, is
presented in Appendix A.
20 4: Components of Response Modification Factors
4.4 Experimental EValuation of Force-
Displacement Relationships
In the mid-1980s, data from experimental research at
the University of California at Berkeley were used to
develop base shear-roof displacement relationships
for steel braced frames and a draft fonnulation for the
response modification factor.
The base shear-roof displacement relationships were
established using data acquired from the testing of
two code-compliant braced steel frames; one concen-
trically braced (Uang and Bertero, 1986) and one
eccentrically braced (Whittaker et aI., 1987). The
force-displacement curves were developed by plot-
ting the roof displacement at the time corresponding
to the maximum base shear force for each earthquake
simulation and each model.
Using these data, the Berkeley researchers proposed
splitting R into three factors that account for contri-
butions from reserve strength, ductility, and viscous
damping, as follows:
(4-4)
In this expression Rs is the strength factor, RIJ. is the
ductility factor, and l?f. is the damping factor. Using
data from the most severe earthquake simulation test,
the strength factor was calculated as the maximum
base shear force divided by the design base shear
force at the strength level. The ductility factor was
calculated as the base shear for elastic response
divided by the maximum base shear force and the
damping factor was set equal to 1.0.
The experimentally determined values of R for the
concentrically braced frame was 4.5 and that for the
eccentrically braced frame was 6.0. These values
were significantly less than the values of 6.0 and 8.0
adopted in the 1991 NEHRP Provisions. The experi-
mentally determined values for the strength, ductil-
ity, and response modification factors are listed in
Table 4.1. The method used to calculate values for
Table 4.1 Experimental Reduction Factors for Steel
Frames
System R
Concentricallybraced 2.43 1.85 4.5
Eccentrically-braced 2.85 2.12 6.0
ATC-19

--en
en
ca
0.
1---7'--,.'+
~-- Period range of
Figure 4.5: Experimental evaluation of strength and ductility factors.
the strength and ductility factors is depicted in Figure
4.5. The reader is referred to Uang and Bertero
(1986) and Whittaker et al. (1987) for additional
infonnation.
4.5 Key Components of R
Much research (ATC, 1982b; Freeman, 1990; ATC,
1995) has been completed since the first fonnulation
for R (Equation 4-4) was proposed. Recent studies,
including those in the companion Project ATC-34,
support a new fonnulation for R; that is, a fonnula-
tion in which R is expressed as the product of three
factors:
where Rs is the period-dependent strength factor, Ril
is the period-dependent ductility factor, and RR is the
redundancy factor. This formulation, with the excep-
tion of the redundancy factor, is similar to those pro-
posed by the Berkeley researchers (see Section 4.4)
and Freeman (1990). The Freeman formulation,
which was developed independently of the Berkeley
fonnulation, described the response reduction factor
as the product of a strength-type factor and a ductil-
ity-type factor.
The redundancy factor, developed as part of Project
ATC-34, is proposed in this report for the flTSt time
ATC-19 4: Components of Response Modification Factors
Elastic Spectrum (; =5%)
--tt-------- Required elastic
strength = R\!Rs V d
----Lj--fl---- Maximum strength
(RsVd)
Design strength( V d )
model
Period (seconds)
in the literature. The function ofthis factor is to
quantify the improved reliability of seismic framing
systems that use multiple lines of vertical seismic
framing in each principal direction of a building.
A fourth factor, the viscous damping factor (R~), was
considered for inclusion in the new fonnulation -
primarily to account for response reduction provided
by supplemental viscous damping devices. Such a
viscous damping factor could be used to reduce dis-
placements in a nonlinear framing system, but cannot
be used to proportionally reduce force demands,
especially for highly-damped frames. Recognizing
that seismic design using response modification fac-
tors will remain force-based in the near tenn, the
(4-5) damping factor was excluded from the new formula-
tion.
One objective of this report is to provide the reader
with information regarding the key components (or
factors) that influence the numerical values assigned
to R in the United States. The fonnulation of R in
Equation 4-5 was put forth to provide a framework
for the rational evaluation of these parameters. Any
evaluation of the key components of R must address
the fact that the components are not independent of
one another. The background infonnation and
research data presented in the following subsections
are intended to provide the reader with insight into
the four key components (i.e., reserve strength, duc-
21
---------r------------.---"----------
tHity, damping, and redundancy) as well as the rela-
tionships between these four components. No
relative importance should be inferred from the order
in which the material is presented.
The proposed formulation does not specifically
address the effects of plan and vertical irregularity in
framing systems. Irregularity could be addressed by
reducing the response modification factor by a regu-
larity factor similar to that prescribed for the Level II
seismic design procedure in the Japanese 1981 Build-
ing Standard Law (see Section 3.2.2 for details). Sig-
nificant force-based penalties (higher design base
shears) for the design of irregular framing systems
would both discourage the use of irregular framing
and reduce the uncertainties associated with the non-
linear response of irregularly framed buildings. For
additional information, the reader is referred to the
ATC-34 document.
4.5.1 Strength Factor
The maximum lateral strength of a building will gen-
erally exceed its design strength. Merovich (unpub-
lished) notes that:
..... In general, members are designed
with capacities equal to, or in excess of
their design loads. While the degree to
which their capacities exceed the design
requirements is a measure of the design
efficiency, all properly executed designs
contain some degree of overstrength or
excess capacity as a consequence of the
design procedure. In some instances, .
geometry or other detail code provisions
will dictate larger member sizes and
hence greater capacities than those
solely based upon conformity to stress/
strength provisions. In other instances,
design provisions related to displace-
ment parameters will produce larger
member sizes than those dictated by
stress/strength provisions. For members
that are sized to resist significant gravity
loads, a substantial percentage of the
overall capacity may be available since
actual loads are probably at levels far
below the design value at the time of the
earthquake..."
The strength factor will likely depend on many
parameters not immediately obvious to many design
22 4: Components of Response Modification Factors
professionals. For example, code-mandated limits on
interstory drift may require the use of member sizes
in flexible (long-period) framing systems that are
greater than those required for strength alone - giv-
ing rise to period-dependent strength factors for drift-
limited framing systems. Also, buildings located in
lower seismic zones will likely have different reserve
strength values than those in higher seismic zones
because the ratio of gravity loads to seismic loads
will differ - resulting in zone-dependent values for
the strength factor. Differences in regional construc-
tion practices and differences between actual and
nominal material strength will also affect the value of
the strength factor, but in less predictable ways.
Osteraas and Krawinkler (1990) made some qualita-
tive observations regarding the likely reserve
strength of buildings as follows.
"... Small, low-rise (buildings) with non-
structural partitions and architectural
elements whose design is controlled by
loading conditions other than seismic
will have high (reserve strength) '" The
effect of nonstructural partitions ... will
decrease with increasing height, as the
'scale' of the nonstructural elements
becomes small compared to that of the
structural elements and as the seismic
loading condition (controls the member
proportions)..."
A method for evaluating the reserve strength of a
building follows. Sample values of Rs calculated by
different researchers are also included.
Evaluation of Strength Factors
Nonlinear static analysis (also termed pushover anal-
ysis) can be used to estimate the strength of a build-
ing or framing system (ATC, 1982b; Bertero, 1986;
Freeman, 1990; Hwang and Shinozuka, 1994; Uang
and Bertero, 1986; Whittaker et aI., 1990). The pro-
cedure used to estimate the strength of a building is
straightforward, but requires the analyst to select a
limiting state of response. Typical limiting responses
include maximum interstory drift and maximum
plastic hinge rotation. The steps in the procedure are
as follows:
1. Using nonlinear static analysis, construct the
base shear-roof displacement relationship for the
building.
ATC-19
2. At the roof displacement corresponding to the
limiting state of response, calculate the base
shear force Vo in the building. The reserve
strength of the building is equal to the difference
between the design base shear (Vd ) and Vo'
3. Calculate the strength factor using the foHowing
expression:
R = VVdo
(4-6)
S and 1.9, respectively.
This method of evaluating the strength factor was
used to create the estimates of strength factors given
below. Appendix A demonstrates the use of non lin-
ear static analysis to construct the base shear-roof
displacement relation for a building and evaluate the
strength factor for that building.
Estimates of Strength Factors
The reserve strength in common seismic framing sys-
tems has been studied by a number of researchers
using nonlinear static analysis. The results of some of
these studies are summarized below.
Freeman (1990) reported strength factors for three
three-story steel moment frames, two constructed in
seismic zone 4 and one in seismic zone 3. The
strength factors, after modification to reflect strength
design, were reported as 1.9,3.6, and 3.3, respec-
tively. Earlier studies by Freeman (ATC, 1982b) esti-
mated strength factors, after modification to reflect
strength design, of approximately 2.8 and 4.8, for
four-story and seven-story reinforced concrete
moment frames, respectively.
Osteraas and Krawinkler (1990) conducted a detailed
study of reserve strength and ductility in three struc-
tural systems: distributed moment frames, perimeter
moment frames, and concentric braced frames. The
framing systems were designed assuming (a) seismic
loads per UBC seismic zone 4 and soil type $2, (b)
dead loads of 70 psf, (c) Jive loads of30 psf, (d) a 3-
bay by 5-bay building plan using 24 square foot bays,
and (e) an elastic period computed using a simplified
relation related to building height. Osteraas and
Krawinkler reported strength factors ranging from
1.8 to 6.5 for the three framing systems. For distrib-
uted moment frames, the strength factor ranged
between 6.5 in the short-period range to 2.1 at a
period of 4.0 seconds. For perimeter moment frames,
the strength factor ranged between 3.5 in the short-
ATC-19 4: Components of Response Modification Factors
period range to 1.8 at a period of 4.0 seconds, and for
concentric braced frames, the strength factor ranged
between 2.8 at 0.1 second to 2.2 at 0.9 second.
Uang and Maarouf(l993) analyzed two buildings
shaken by the 1989 Lorna Prieta earthquake: a 13-
story steel frame building and a six-story reinforced
concrete perimeter moment frame building. The
strength factors reported for these two buildings,
after modification to reflect strength design, were 4.0
Hwang and Shinozuka (1994) studied a four-story,
reinforced concrete, intennediate moment frame
building located in UBC seismic zone 2. The design
base shear for this building was 0.09W. The maxi-
mum lateral resistance of the building was calculated
to be 0.26W, resulting in a strength factor of2.2 ifno
limits are placed on the damage to the framing sys-
tem. (lfthe perfonnance level selected for the design
earthquake were no damage to the structural frame,
the strength factor would have been approximately
1.6).
The scatter in the reported values for the strength fac-
tor is significant - and too large to be of much use
to the design professional community. It is clear that
coordinated and systematic studies are needed to
develop strength factors of sufficient reliability to be
included in seismic design codes. These studies
ought be conducted at the national level to effectively
address the issues identified earlier in this section.
4.5.2 Ductility Factor
The seismic response parameters of displacement
capacity, ductility, and ductility ratio are closely
inter-related, but often confused. For example, a
frame with a large displacement capacity might have
smaH ductility and a small ductility ratio, and a frame
with a small displacement capacity might have small
ductility but a large ductility ratio.
Consider the force-displacement relationships for
two one-story building frames shown in Figure 4.6.
The nonnalized force-displacement relationships are
idealized as elastic-plastic, the yield drift ratios are
assumed to be 0.2 percent (Frame A) and 1.0 percent
(Frame B), and the maximum interstory drift ratios
are assumed to be 1.2 percent (Frame A) and 3.0 per-
cent (Frame B). The key seismic response parameters
are listed in Table 4.2 below. The values of the
response parameters are constrained by the interstory
drift limit of 1.5 percent, which is consistent with the
23
 Ductility of A
r - Code maximum drift
,. of 1.5%

Displacement
capacity of A
CI)
Ductili of B
f:!
o
LL

---- -------------~--- ...\

A
\.8 +

1.0% 2.0% 3.0%

Drift (% of story height)

Figure 4.6: Definition of terms for two example one-story frames.

ratio, and rotation ductility ratio. For the purposes of

Table 4.2 Seismic Response Parameters for Two this discussion, displacement ductility ratio at the
Example One-Story Frames
system level is used to determine the ductility factor.
Response Parameter Frame A Frame B The calculation of displacement ductility ratio for a
building is demonstrated by example in Appendix A.
Yield drift 0.2% 1.0%
It must be recognized that the ductility factor is a
Drift Capacity 1.2% 1.5% measure of the nonlinear response of the complete
Displacement Ductility 1.0% 0.5% framing system and not components of the framing
system, regardless of which ductility parameter is
Displacement Ductility 6 1.5 used.
Ratio
Assuming that reliable estimates ofdisplacement
ductility are available, the next step in estimating the
drift limits set forth in the UBC. The drift values in
ductility factor is to derive a relationship between
Table 4.2 are expressed as a percentage of the story
displacement ductility and the ductility factor. This
height. These data illustrate the importance of defin-
step has been the subject of much research in recent
ing response parameters with respect to specific limit
years. The relationships developed by Newmark and
states. By limiting drifts to 1.5 percent, the stiffer
Hall (1982), Krawinkler and Nassar (1992), and
frame (Frame A) is more ductile and has a higher
Miranda and Bertero (1994) are paraphrased below
ductility ratio than the more flexible frame (Frame
as background information for the reader. Although
B). However, if the drift limit state is removed, the broad consensus has not yet been reached on the use
more flexible frame has substantially more ductility of one of the suites of relationships outlined below,
(equal to 2 percent) than the stiffer frame. the latter two sets of relationships best fit the avail-
The ductility ratios (J.1.) can be computed at the sys- able data.
tem, story, and element levels. At the system and
Newmark and Hall Research
story levels, the ductility ratio is normally expressed
in terms of the displacement ductility ratio. At the Newmark and Hall (1982) provided relationships that
element level, ductility ratio can be expressed in can be used to estimate the ductility factor (R~ for
terms of strain ductility ratio, curvature ductility elasto-plastic SDOF systems as follows:

24 4: Components of Response Modification Factors ATC-19

._-------------_._ - .. ..
 8.---.---.---,..----,---,---,..-----r--....,

... 6 ;. . . _ _;_._.:
;:-._.-: . :J.1=6.
_._:_._ :. _._
o ..
U /.
~ : /: . . : J.1=4 : .
~4 j ' ';";< - ' - ' ; -'-': - ' - ' ; -'_'; _'_'; _._._
13 i-'-:' ""
:J .-_/ :J.1=2
a 2 1...... :..
tY'
OL..--.....L-.--"""----L-_--J...._ _~_ _.l..--_--..I._ _--'
o 0.5 1 1.5 2 2.5 3 3.5 4
Period (seconds)

Figure 4.7: Newmark and Hall R!J. - Jl - T relationship.

for frequencies above 33 Hz (periods below 0.03 sec- R~ = [c(Jl- 1) + 1] lie (4-10)
ond):
where:
RlJ. = 1.0 (4-7)
Ta b
c(T, a) = - - + - . (4-11)
for frequencies between 2 Hz and 8 Hz (periods 1+T a T
between 0.12 second and 0.5 second):
The regression parameters a and b were obtained for
different strain-hardening ratios (tenned a in Figure
4.9) as follows:

for frequencies less than 1 Hz (periods exceeding 1.0

second):
(4-9)
Figure 4.7 illustrates the Newmark and Hall relation-
ships for displacement ductility ratios of2, 4 and 6.
Estimates for R between 0.03 second and 0.12 sec-
ond, and 0.5 se60nd and 1.0 second can be interpo-
lated between the limiting values given by Equations
4-7,4-8, and 4-9.
Krawinkler and Nassar Research
Krawinkler and Nassar (1992) developed a
RI1 - J.1- T relationship for SDOF systems on ~o~k
or stiff soil sites. They used the results of a statistical
study based on 15 western United States ground
motion records from earthquakes ranging in magni-
tude from 5.7 to 7.7. Developed assuming damping
equal to 5 percent of critical, the Krawinkler and
Nassar equation is:
a =0%: a = 1.00 b = 0.42
a =2%: a = 1.00 b = 0.37
a =10%: a =1.00 b = 0.29
Note that a equal to 0% corresponds to an elastic
plastic system. The relationships between R!J. and T
for displacement ductility ratios of 2, 4, and 6 are
presented in Figure 4.8 for a strain-hardening ratio of
10 percent.
Krawinkler and Nassar also studied the implications
of extending their R!J. - J.1- T relationships to multi-
ple-degree-of freedom (MDOF) systems. Three
model types were analyzed for target ductility ratios
between 2 and 8: strong column-weak beam, weak
column-strong beam, and weak first story. Given that
the latter two failure modes are discouraged by the
model building codes and the NEHRP Provisions.
only the strong column-weak beam results are
reported here. The objective of this study was to
develop a procedure whereby the maximum story
displacement ductility ratio in a MDOF system could
be limited to the corresponding ductility ratio in the

ATC-19 4: Components of Response Modification Factors 25

 8 .--------r---,-.-----.---.- - T .- - - - r - - - r - - - , . . - - - - - - ,
.( / -':-'-' ~._.- .:-. _.~ '~-:'6'[-' -. ~. -.-
... 6 /.: : : : : : : .
,g ': .
~
u. I: -:
_' '_ _- ';-
: -
: ~ =4- :.:-
- :- - - - : _
~ 4 ../"; "": ": : , : : :" , ': .

o
g II
., :~=2
2-V
O ' - - - - " - -..........- -.......- -.........--L.-----'----I..---..I
o 0.5 1 1.5 2 2.5 3 3.5 4
Period (seconds)

Figure 4.8: Krawinkler and Nassar R~ - ~ - T relationship.

...
~
m 2 r-----..,....------.------r----~--------,
(J)
Q)
(J)
co
1.5 ............ : : ~ . ."" +---: . .. . .
..0
u..
oo
'.
.......
-- .
_' 0

fg 1
co
Q)
~
(J)
lIE lIE ~=4;a= 0%
ct- - -0 ~=4; a= 10%
0.5 ..........................
Q)
~ . .- +- - + ~ = 8; a = 0%
..0 CJ 0 ~ = 8; a = 10%
u..
oo 0 '-- ---' ........ ......... ..I-. ----l

::E 0 0.5 1 1.5 2 2.5

Period (seconds)

Figure 4.9: MDOF modification factors.

SDOF system. Krawinkler and Nassar concluded that
the strength demands for SDOF systems must gener-
ally be increased to be applicable for MDOF frame
structures. The modification factor, defined as the
required base shear strength of the MDOF system
divided by the inelastic strength demand of the corre-
sponding first-mode SDOF system, limits the story
ductility ratio in the MDOF system to the target duc-
tility ratio. Modification factors for target ductility
ratios of four and eight, and strain hardening ratios of
o percent and 10 percent are presented in Figure 4.9.
The reader is referred to Nassar and Krawinkler
(1991) for additional information.
system is approximately equal to the corresponding
SDOF system strength demand, suggesting that
higher-mode effects need not be considered in this
period range. For buildings with fundamental periods
exceeding 0.75 second, higher-mode effects will
necessitate an increase in the design lateral strength if
target ductility ratios are to be satisfied. In general,
the modification factor increases with increasing tar-
get ductility ratio and decreases with increasing
strain hardening. MDOF systems without strain hard-
ening drift more than the corresponding SDOF sys-
tem, and increased lateral strength is required to limit
the story ductility ratio to the target value.

For buildings with fundamental periods less than

0.75 second, the base shear demand on the MDOF

26 4: Components of Response Modification Factors ATC-19

 8.--------r---,.---...,.----,.--.....,--.------,--~

~6
o
U
rn
u.
~4
;::
(.J
::s . : Il = 2 :
o ..- . - .:---- .::-. - - '- -
2
Nassar & Krawinkler
- - - - Miranda - Rock
.- .- .- Miranda - Alluvium
oL-_---1..._ _"""--_--J._ _.....L..-_----''--_--'-_ _''---_---'
o 0.5 1 1.5 2 2.5 3 3.5 4
Period (seconds)

Figure 4.10: Ductility factor comparison.

Miranda and Bertero Research
Miranda and Bertero (1994) summarized and
reworked the RI,J. - Jl. - T relationships developed by a
number of researchers including Newmark and Hall
(1982), Riddell and Newmark (1979), and Krawin-
kler and Nassar (1992), in addition to developing
general RI,J. - Jl. - T equations for rock, alluviu~, and
soft soil sites. The Miranda and Bertero equatIons
presented below were developed using 124 ground
motions recorded on a wide range of soil conditions,
and assumed five percent of critical damping. Their
equation for the ductility factor is
Jl. - 1
RIl =--+1
<1>
where:
for rock sites:
the differences between these relationships are rela-
tively small, they can be ignored for engineering pur-
poses.
4.5.3 Redundancy Factor
A redundant seismic framing system should be com-
posed of multiple vertical lines of framing, each
designed and detailed to transfer seismic-induced
inertial forces to the foundation. Although redun-
dancy is encouraged for lateral force-resisting sys-
tems designed in the United States, the trend in
California in recent years has been to construct seis-
mic framing systems composed of only a small num-
ber of vertical lines of seismic framing; that is,
(4-12) framing systems with minimal redundancy. This
trend in California is likely a result of poor under-
standing by earthquake engineers of the important
role played by redundancy in the response ofseismic
framing systems to severe earthquake shaking.

<I> = 1+ 1 _ ~e-1.5(1n(n-o.6f Few studies have examined the effect of redundant

lOT - Jl.T 2T seismic framing in a quantitative way. However such
studies have been carried out for wind framing sys-
for alluvium sites: 2 tems by Moses (1974). In this study, it was noted that
<I> = 1 + 1 _ 2:.. e -2(ln(n- o.2 ) safety margins for wind framing system collapse
12T - Jl.T 5T modes depend on the sum of several strength and
load variables. Therefore, the reliability of the fram-
.for soft soil sites: ing system will be higher than the reliability of indi-
Tg 3Tg -3(1n(TIT,)-O.25)2 vidual members of the framing system. Moses
<1>= 1 +3T-4Te
concluded that a partial safety factor less than or
and T is the predominant period of the ground equal to one was appropriate for a redundant system,
motio~ (see Miranda and Bertero (1994) for details). and that the required mean member strength could be
reduced in such a system below that necessary for
A comparison of the Nassar and Krawinkler and nonredundant or determinant systems. A mean
Miranda and Bertero R Il - Jl. - T relationships for rock strength reduction factor inversely proportional to the
and alluvium sites is presented in Figure 4.10. Since

ATC-19 4: Components of Response Modification Factors 27

._.._---_.....\
------l
 denotes a moment
connection
(and plastic ~J
hinge)
r
- ,__
denotes an
idealized pin

Mp = 100 k'

Seismic Seismic
framing framing

Frame A Frame B

Figure 4.11: Redundancy in moment frame systems.

square root of the number of independent strength
terms (plastic hinges in the sway mechanism) in the
redundant wind framing system was proposed. As
illustrated below, similar logic can likely be applied
to seismic framing systems.
Consider the two framing systems with identical
geometry depicted in Figure 4.11.Frame A is com-
posed of one bay of seismic framing with each beam
member capable of developing a nominal plastic
moment of 200 units. Frame B is composed of two
bays of seismic framing with each beam member
capable of developing a nominal plastic moment of
100 units. Both limit analysis and nonlinear static
analysis would assign both frames the same maxi-.
mum lateral strength. However, using a methodology
similar to that proposed by Moses (1974) for wind
framing systems, the ratio of the nominal moment
strength (Mp ) of the beams in Frame A (eight total
plastic hinges) and Frame B (sixteen total plastic
hinges) should be:
MA In
-!...B = ~=1.4 (4-13)
Mp 11./16
To achieve a similar level of reliability, the design
lateral strength of Frame A should be 40 percent
higher than that of Frame B.
As another example, consider the two framing sys-
tems depicted in Figure 4.12. Frame C is composed

denotes an
I~ Id.a1~"" plo

Frame C Frame D

Figure 4.12: Redundancy in shear wall systems.

28 4: Components of Response Modification Factors ATC-19

r--.-------,--",----
of three bays of fram ing including one flexural wall
capable of developing a nominal plastic moment of
1000 units. Frame D is composed of three bays of
framing including two flexural walls, each capable of
developing a nominal plastic moment of 500 units.
Limit analysis would assign both framing systems
the same lateral strength. Using the methodology of
Moses (1974), the ratio of the design lateral strength
of the shear walls in Frame C(one plastic hinge - at
the base of the wall) and Frame D (two plastic hinges
- one hinge at the base of each wall) should be 1.4
to achieve a similar level of reliability.
Four lines of strength- and defonnation-compatible
vertical seismic framing in each principal direction of
a building have been recommended as the minimum
necessary for adequate redundancy (Bertero, 1986;
Whittaker et aI., 1990). It could be possible to penal-
ize less redundant designs by requiring that higher
design forces be used for such framing systems. For
example, if it is assumed that four lines of strength-
and deformation-compatible vertical seismic framing
should form the basis of the response reduction fac-
tors in the UBC and NEHRP Provisions, redundancy
could be explicitly accounted for by modifying the R
factor in a manner similar to that suggested in Table
4.3.
The values shown in Table 4.3 are proposed to dem-
onstrate a likely trend, stimulate discussion among
design professionals and researchers, and to promote
research and study. These draft values for the redun-
dancy factor have no technical basis and are not
intended for implementation in seismic codes or
guidelines.
Table 4.3 Draft Redundancy factorsa
Lines of Vertical Draft
Seismic Framing Redundancy Factor
2 0.71
3 0.86
4 1.00
a. Values not intended for use in design or
codes
The use of strength- and defonnation-compatib1e
framing was emphasized in the previous paragraph.
To illustrate the importance of setting limits on the
ATC-19 4: Components of Response Modification Factors
------,----_----:---_
relative strength and stiffness of the lines of vertical
seismic framing in a redundant framing system, con-
sider the base shear force-roof displacement response
of the reinforced concrete shear wall-steel moment
frame dual system shown in Figure 4.13. This dual
system was chosen because many design profession-
als consider the dual system to be a redundant seis-
mic framing system.
The 1991 NEHRP Provisions state that ..... the
moment frame (in the dual system) shall be capable
of resisting at least 25 percent of the design forces.
The total shear force resistance is to be provided by
the combination of the moment frame and the shear
walls or braced frames in proportion to their rigidi-
ties." For the purpose of this discussion, the shear
walls are assumed to be ten times stiffer than the
moment frames. The design base shear values for the
walls and moment frames are therefore 91 percent
and 25 percent of the system design base shear (V in
Figure 4.13), respectively. Assuming that the shear
walls yield at a roof-drift ratio' (calculated by divid-
ing the roof displacement by the building height) of
0.2 percent, and fail at a roof-drift ratio of 1.0 per-
cent, it is evident from Figure 4.13 that the moment
frames (also termed the back-up frames) neither con-
tribute substantially to the force-displacement
response of the building nor dissipate significant
energy at a roof displacement corresponding to the
displacement capacity of the shear walls. For the
moment fr:unes to contricute sigr.iticantly to the
response of a dual system, their stiffness and strength
should be similar to that of the shear walls.
The need to use elements of similar strength and
stiffness applies to all lines of vertical seismic fram-
ing in a building. It is not sufficient to provide multi-
ple lines of vertical seismic framing in a building -
the multiple lines of framing must be strength- and
deformation-compatible to be capable of good
response in a design earthquake. Seismic frames not
meeting these conditions should probably not be con-
sidered redundant systems.
4.5.4 Damping Factor
Damping is the general tenn often used to character-
ize energy dissipation in a building frame, irrespec-
tive of whether the energy is dissipated by hysteretic
behavior or by viscous damping.
Damping accomplished by hysteretic behavior in a
building responding in the elastic range is generally
29
__
.."_. .".__.. ..
",, " ......
 Base Shear r Dual system

,,---------f-\ - Dual system "fails at

defonnation capacity
of shear wall
0.91V #

0.5V
( Back-up moment frame

1.0%
Roof Drift Index (%)

Figure 4.13: Force-displacement relationships for a dual system.

termed equivalent viscous damping and is assigned a follows.

value equal to five percent of critical. The use of five
percent equivalent viscous damping is reasonable - Consider again the elastic SOOF system shown in
values of equivalent viscous damping ranging Figure 4.1. The equation of motion for this frame is
between five percent (steel frames) and seven percent given as either
(shear walls) (ATC, 1974), and five percent (steel
frames) and eight percent (shear walls) (DOD, 1986), mvt(t) + ev(t) + kv(t) = 0 (4-14)
have been reported. However, given that such damp-
ing is probably heavily dependent on the type and or
arrangement of interior and exterior nonstructural
elements, there is no compelling reason to reduce mv(t) + ev(t) + kv(t) = -mvg(t) (4-15)
seismic demands on selected framing systems to
reflect marginal increases of two to three percent in
building damping. where m, e, and k, are defined in Figure 4.1; yt(t),
v(t), and v(t) are the total acceleration, relative veloc-
The damping factor as discussed in Sections 4.4 and ity, and relative displacement of the mass, respec-
4.5 is intended to account for the influence of supple- tively; and Yg(t)is the ground acceleration.
mental viscous damping devices on the force and dis-
placement response of buildings. This influence has Equation 4-14 can be rewritten as:
been studied by a number of researchers (Riddell and
Newmark, 1979; Wu and Hanson, 1989). These stud- (4-16)
ies have focused on displacement response only, and
data from the Riddell and Newmark studies have and then simplified to read:
been implemented in the 1994 UBC for the design of
seismic isolation systems. vt(t) = -cJv(t)-2co~v(t) (4-17)
However, current seismic design procedures using R
where co = kim and 2co~ = elm
2
factors are force-based procedures. The addition of
viscous damping to a building frame will always
In this equation, cJv(t) is the hysteretic (or spring)
serve to reduce displacements, but may increase the
force per unit mass developed in the columns of the
inertial forces if the viscous forces become substan-
SDOF system, and 2co~v(t) is the damping force per
tial. This relationship can be demonstrated simply as

30 4: Components of Response Modification Factors ATC-19

---------_.----
 unit mass developed in the dashpot of the SDOF sys-
tem. The solution to the equation of motion for the
SDOF system is
I Viscous
v(t) = _(_1_)
roD 0
Iv ('t)e-~CJ){t-t)sinCOD(t - 't)d't(4-18)
g
where is the angular frequency; and roD is the
damped angular frequency (Clough and Penzien,
1993). The maximum value ofv(t) is termed the
spectral displacement; the spectral displacement is
equal to the pseudo-displacement. From Equation 4-
18, it is clear that an increase in damping (I;) will lead
to a reduction in the displacement response.
The maximum value of jit(t) is termed the spectral
acceleration; the maximum value of co 2 v(t) is the
pseudo-acceleration which is typically used in seis-
mic design to represent the maximum acceleration of
the mass in the SDOF system. From Equation 4-17, it
can be seen that the spectral acceleration is approxi-
mately equal to the pseudo-acceleration if the damp-
ing forces (equal to cv in Equation 4-15) are small.
If the damping forces are large, the contribution of
the damping force to the inertial force will become
significant. The foregoing discussion is limited to the
elastic response of a SDOF system. The relation
between spring force (= lev in the elastic range) and
damping force (cv ) will be further exacerbated if the
frame is designed using a large response modifica-
tion factor, because the value of the spring force will
be limited to the inelastic strength of the frame (= Vo
in Section 4-3) whereas the damping force can con-
tinue to increase as the velocity increases.
Table 4-4 lists values of the damping factor R~ for
reducing displacement response only, taken from the
1994 UBC provisions for seismic isolation systems
and from the work of Wu and Hanson (1989) for dif-
ferent levels of viscous damping. These factors
should not be used to reduce hysteretic force
demands unless the forces developed in the viscous
elements are explicitly accounted for in the design
process.
4.6 Systematic Evaluation of R Factors
Response modification factors playa key, but contro-
versial, role in the seismic design process in the
United States. No other parameter in the design base
shear equation (Equations 2-4 and 2-5) impacts the
design actions in a seismic framing system as does
ATC19 4: Components of Response Modification Factors
r'--'
,-----------------_.,,_.._-_
Table 4.4 Damping Factor as a Function of Viscous
Damping
1994 Wuand
damping
(% critical)
UBC
(l?f.)
Hanson
(Rc,)
2 0.80
5 1.00 1.00
7
10 1.20 1.19
12
15 1.39
20 1.50 1.56
the value assigned to R. Despite the profound influ-
ence of R on the seismic design process, and ulti-
mately on the seismic performance of buildings in
the United States, no sound technical basis exists for
the values of R tabulated in seismic design codes in
the United States. There is an obvious and pressing
need to develop a rational technical basis for R fac-
tors if equivalent lateral force design procedures are
to be retained for seismic design.
Ifthe new formulation for R presented in this chapter
is to be implemented in building codes in the United
States, systematic and coordinated studies are
required to support or modify the proposed values for
the strength, ductility, redundancy, and damping fac-
tors. Research is also needed to fully characterize the
interdependence of the four factors. There is no merit
in replacing unsubstantiated R values with three or
four times as many unsubstantiated values for
strength, ductility, redundancy, and damping factors.
Strength and ductility factors for most seismic fram-
ing systems will likely vary between seismic zones
due to differences in the ratios of gravity loads to
seismic loads. Consequently, strength and ductility
factors should be evaluated for each seismic framing
system in each seismic zone using standardized defi-
nitions of reserve strength and ductility. The studies
conducted by Osteraas and Krawinkler (1990) for
three framing systems in seismic zone 4 provide a
good model for how to carry out such studies. For
each seismic framing system considered, multiple
plan and vertical building geometries must be ana-
lyzed in a systematic manner to provide the data nec-
essary to quantify the R factors. The draft procedures
described above for the evaluation of the different
31
.....
 component factors of R could be used for such coor-
dinated studies.
Framing systems with less than four vertical lines of
strength- and deformation-compatible seismic fram-
ing in each principal direction ofthe building, or
those possessing minimal torsional redundancy,
should be penalized through the use of a redundancy
factor. Limits must be placed on the relative strength
and stiffness of the vertical lines of seismic framing
in each principal direction of a building. The numeri-
cal values assigned to the redundancy factor must be
established using reliability theory.
Should seismic design practice in the United States
shift to displacement-based procedures rather than
force-based procedures, it may be appropriate to
include a damping component in the R factor. Before
this can be done, values for the damping factor will
32 4: Components of Response Modification Factors
--.....--------,----- ,----------------------- ._..-.__
have to be established through careful research.
4.7 Rel1ab111ty of Values for R
It is of paramount importance that revised values for
R and the values for the component strength, ductil-
ity, and redundancy factors be reliable, in the sense
that buildings designed using these factors should
meet the assumed performance level in the design
earthquake. Values for the strength and ductility fac-
tors should be evaluated using a consistent methodol-
ogy. It is also important that sufficient numbers and
types of buildings be analyzed to permit statistical
evaluation and interpretation of the responses. The
values assigned to both R and its component factors
should aim to provide either a uniform level of risk
for all framing systems or a level of risk that is con-
sistently less than a yet-to-be-determined threshold.
ATC-19
...............
 5. Conclusions and Recommendations
5.1 Summary and Concluding
Remarks
The objectives of this report were three-fold: (I) to
review the role played by response modification fac-
tors (R factors) in the seismic design of buildings in
the United States, (2) to document the basis of the
values of R (and its working-stress equivalent Rw )
utilized in seismic codes in the United States, and (3)
to propose a method for the systematic and rational
determination of values of R for seismic framing sys-
tems used in the United States, including the determi-
nation of values for the key components of R.
The focus of Chapter 2 was the historical develop-
ment of values for Rand Rw . It was reported that val-
ues for R and Rw could be traced directly to the
horizontal force factor K first introduced into seismic
codes in the United States in 1959, and that the val-
ues for K represented the consensus opinion of expert
structural engineers in California in the late 1950s.
The values assigned to K were based on engineering
judgement, and not on detailed analysis - not sur-
prising since the requisite analytical tools were not
available to the design professional community until
the mid 1970s. This investigation concluded that:
1. There is no mathematical basis for the response
modification (R) factors tabulated in modern
seismic codes in the United States.
2. A single value of R for all buildings of a given
framing type, irrespective of plan and vertical
geometry, cannot be justified.
3. To ensure consistent levels of damage, values for
R should depend on both the fundamental period
of the building and the soil type on which the
building is founded.
4. The values assigned to R for a given framing sys-
tems should vary between seismic zones because
the reserve strength in a framing system will
probably be a function of the ratio of the gravity
loads to the seismic loads. Also, detailing
requirements currently vary by zone.
ATC-19 5: Conclusions and Recommendations
,_._-.,------------- ,-------r---.------"--.----.- .
5. The values currently assigned to R for different
framing system will probably not result in uni-
form levels of risk for all buildings.
The use of response modification factors is not lim-
ited to the seismic design of buildings in the United
States. A similar response reduction factor (Z) is used
by Caltrans and AASHTO for the seismic design of
bridges in the United States. Moreover, response
modification factors or their equivalents are used in a
number ofcountries for seismic design. In Chapter 3,
response modification factors in use in Europe,
Japan, and Mexico were studied, and the numerical
values of R for three framing systems were compared
with those used in the United States. The investiga-
tion showed that larger values of R are used in the
United States than in Europe and Mexico. Given that
the seismic performance objectives are similar in all
cases, the conclusion drawn from the study was that
the values of R in seismic codes in the United States
may not be sufficiently conservative.
Chapter 4 described the impact of R on the seismic
design process, introduced force-displacement rela-
tions for buildings, summarized the experimental
evaluation of R for two steel-braced framing systems,
and proposed a new formulation for R. The results of
an experimental program at the University ofCalifor-
nia at Berkeley were reviewed. These results suggest
that the values of R tabulated in the NEHRP Provi-
sions for concentrically and eccentrically braced
frames are not sufficiently conservative, and that fur-
ther study is needed.
The new formulation for the response modification
factor splits R into factors related to reserve strength,
ductility, and redundancy. The implications of intro-
ducing a damping factor into the new formulation
were also addressed. Further, procedures for evaluat-
ing strength and ductility factors using commercially
available analytical tools were proposed.
The lessons to be learned from the 1994 Northridge
and 1995 Kobe earthquakes are clear. First, moderate
magnitude earthquakes can produce elastic seismic
demands substantially greater than those assumed by
seismic codes in the United States. Second, moderate
33
 magnitude earthquakes in major urban areas can
result in huge social and economic losses. Although
damage and loss of older structures was expected in
these earthquakes, damage to buildings and bridges
designed to modern seismic standards was not. The
degree of loss in both earthquakes was too high -
the seismic risks can and must be mitigated. One key
step in the mitigation effort is to improve the reliabil-
ity of new construction in a timely manner. One
obvious way to improve the reliability of new con-
struction is to improve the reliability of the response
modification factors used in the seismic codes.
5.2 Recommendations
Static lateral force analysis and design procedures are
key components of routine seismic design in practice
in the United States. One key step in the procedure is
the calculation of design forces. These are typically
calculated by dividing the elastic spectral force by a
response modification factor (R). Even with the
advent of powerful nonlinear analysis packages, elas-
tic design procedures are likely to be used for much
seismic analysis and design in the foreseeable future
and such procedures necessarily make use of a
response modification factor in one fonn or another.
34 5: Conclusions and Recommendations
------l.....--------..,-----
r-'
For this reason, and recognizing that there is no
sound basis for the values assigned to R in U.S. seis-
mic codes, it is of paramount importance to establish
appropriate values for R through rigorous research
and study so that the intended perfonnance of the
building stock can be realized with a high degree of
reliability.
This report has proposed a draft fonnulation for the
systematic quantification of R and its component fac-
tors and has recognized the need to quantify the
redundancy factor through reliability analysis. A
coordinated action plan is currently being developed
(ATe, 1995) to systematically evaluate R for differ-
ent framing systems in all regions in the United
States, and to improve the reliability of current seis-
mic design procedures. The execution of this action
plan will require significant funding from federal
agencies. However, the rapid implementation of this
plan will substantially reduce the exposure offederal,
state, and local agencies to substantial loss in future
earthquakes in the United States. Further, the knowl-
edge gained from these studies should prove useful
for the evaluation and rehabilitation of older con-
struction.
ATC-19
 Appendix A: Evaluation of Building Strength and
Ductility
A.I Introduction
A trend in earthquake engineering practice in the
United States, especially in California, is for the
strength and deformation characteristics ofa building
to be evaluated using nonlinear static analysis. This
kind of analysis can estimate the likely maximum
strength and deformation capacity of the building,
identify potential weak and/or soft stories in the
building, and identify poorly proportioned framing
elements that give rise to excessive deformation
demands at the element level.
In order to evaluate the probable force-deformation
response of a building, member proportions must be
established or known a priori. For new construction,
preliminary member proportions are generally estab-
lished using the procedures set forth in a seismic
code; for existing construction, member sizes are
known.
The purpose of this appendix is two-fold: (1) to intro-
duce the reader to key features of nonlinear static
analysis and (2) to demonstrate techniques for evalu-
ating strength and ductility factors. Nonlinear static
analysis is one of four procedures proposed by the
ATC-33 project team (ATC, in progress) for the seis-
mic rehabilitation of existing buildings - the
remaining three procedures being linear static analy-
sis, linear dynamic analysis, and nonlinear response-
history analysis.
The salient features of nonlinear static analysis are
introduced below. We first provide a brief descrip-
tion of nonlinear analysis. This is followed by a
description of a nonproprietary computer code for
nonlinear analysis (DRAIN-2DX). DRAIN-2DX is
then used to analyze a nonductile, seven-story rein-
forced concrete frame building, constructed in 1966,
and to evaluate the base shear-roof displacement
relationship for this building.
The base shear-roof displacement relationship is ana-
lyzed in Section AA to estimate the strength and duc-
tility factors for the subject seven-story building -
with the intent of demonstrating to the reader how to
calculate the strength factor, the different means by
ATC-19 AppendiX A: Evaluation of Building Strength and Ductility
-~--------,,_
1-'---" .._ --
which to calculate the displacement ductility ratio for
the subject building, and how to estimate the ductility
factor.
A.2 Nonlinear Static Analysis
A.2.1 Introduction
The two common methods of nonlinear analysis are
nonlinear static analysis and nonlinear response-his-
tory analysis. For both methods of analysis, a fram-
ing system is modeled and analyzed as an assembly
of elements and components. The data output from
either analysis procedure includes force and defor-
mation demands on elements and components.
Nonlinear static analysis is less demanding in a com-
putational sense than nonlinear response-history
analysis, but more rigorous than linear methods of
analysis. In nonlinear response-history analysis, a
mathematical model of a building is subjected to dig-
itized records of earthquake ground motions. The
analysis is generally terminated at the end of the
earthquake ground motion record - often after more
than 2000 time steps.
In nonlinear static analysis, increasing inertial forces
(or displacements) are imposed on a mathematical
model of a building. The analysis is terminated once
a target displacement is reached - often after fewer
than SO load steps. The target displacement repre-
sents a maximum building displacement during
earthquake shaking; brief comments on the selection
of the target displacement are provided in Section
A.2.7.
The remainder of Section A.2 is devoted to nonlinear
static analysis. The following sections provide the
reader with an introduction to the subject. Much
additional information is available in the literature.
A2.2 Basics of Nonlinear Static Analysis
For the purposes of seismic analysis, a building
should be modeled and analyzed as a three-dimen-
sional assembly of elements and components. Two-
dimensional modeling and analysis of a building will
generally be acceptable if either the torsional effects
are small or the three-dimensional effects can be
35
 accounted for separately.
The mathematical model of a building should include
the following: (a) all elements and components ofthe
seismic and gravity framing systems, (b) nonstruc-
tural components in the building likely to possess
significant stiffness and strength, and (c) elements of
the foundation system (footings, piles, etc.) that are
sufficiently flexible and/or weak to contribute to the
response of the building. The distribution of the
equivalent lateral static loads (see Section A.2.6) in
the mathematical model should be adequate to cap-
ture all key dynamic effects on the seismic and grav-
ity framing system, the nonstructural components,
and the foundation.
Gravity loads should be imposed on the mathemati-
cal model to reflect those loads likely to be present
during earthquake shaking. The initial gravity load-
ing conditions (QG) can be described by one of the
following two equations (ATC, in progress):
QG = 1.1(QD + QL + Qs) (A-I)
QG = O.9QD (A-2)
where QD , QL , and Qs are the dead, live, and snow
loads, respectively. Equations A-I and A-2 are
intended to provide upper- and lower-bound esti-
mates, respectively, on the likely gravity loads on an
element or component. Other load combinations
(BSSC, 1991) can also be considered.
A.2.3 Modeling Elements, Components, and
Connections
The mechanical characteristics (i.e., force-deforma-
tion) of each element and component of the building
should be modeled in sufficient detail that their
important effects on the response of the building are
reasonably represented. In most instances, the
mechanical characteristics estimated for the analysis
will be elastic stiffness, inelastic stiffness, and yield
strength. Failure modes (e.g., shear) that may occur
at deformations smaller than those anticipated in the
analysis should be accounted for in the element or
component model.
Elements and components of buildings should be
modeled using actual rather than nominal geometries
and mechanical properties. For example, the mechan-
ical characteristics of a beam in a reinforced concrete
frame should account for the likely presence of a
36 Appendix A: Evaluation of Building Strength and Ductility
r-----------'--- ,----------
reinforced concrete floor slab. Similarly, the yield
strength of a wide-flange steel beam should be based
on the likely yield stress rather than the nominal
yield stress.
Connections between framing members should be
modeled unless the connection is sufficiently stiff to
prevent relative deformation between the connected
elements or components and the connection is stron-
ger than the connected elements or components.
A.2.4 Nonlinear Static Procedure
The nonlinear static procedure requires an a priori
estimate of the target displacement. The target dis-
placement serves as an estimate of the maximum dis-
placement of a selected point (node) in the subject
building during the design earthquake.The node
associated with the center of mass at the roof level is
often the target point or target node selected for com-
parison with the target displacement.
Nonlinear static analysis is integrated into the four-
step nonlinear static procedure as follows:
1. Develop a two- or three-dimensional mathe-
matical model of the building, as described
in Section A.2.2.
2. Impose constant gravity loads, and then apply
static lateral loads (or displacements) in pat-
terns that approximately capture the relative
inertial forces developed at locations of sub-
stantial mass.
3. Push the structure using the load patterns of
Step 2 to displacements larger than those
associated with the target displacement (Le.,
the displacement of the target node exceeds
the target displacement).
4. Estimate the forces and deformations in each
element at the level of displacement corre-
sponding to the target displacement (Step 4).
The element force and deformation demands of Step
4 are then compared with the element capacities in a
manner similar to that demonstrated in Section A.3.
A.2.5 DRAIN Computer Code
DRAIN-2DX (Prakash et aI., 1992) is a two-dimen-
sional, general-purpose, nonlinear, finite-element
analysis program developed at the University of Cal-
ifornia at Berkeley. The modeling and analysis pro-
cedures incorporated in DRAIN-2DX are
ATC-19
-----------_.. -._--_ - .. .,,"" '
I
summarized below. The DRAIN-2DX computer
code could be used for Steps 2,3, and 4 of the non-
linear static procedure described above.
Building framing systems are modeled as two-
dimensional (X-Y) assemblages of nonlinear ele-
ments connected at nodes. Unless a node is restrained
or slaved to another node, each node has three
degrees of freedom. Elements (and components) are
divided into groups, although alI elements of a given
type (e.g., beam-column) need not be assigned to a
single group. Masses are lumped at nodes, so the
nodal points should be selected to adequately captUre
the inertial response ofthe building. By lumping the
masses at nodes, the mass matrix is diagonal.The
damping matrix can be made proportional to the ele-
ment stiffness values and nodal masses.
Numerous analysis types are available with DRAIN-
2DX, including (a) static gravity analysis for com-
bined element and nodal loads, (b) nonlinear static
analysis for nodal loads, (c) eigen analysis for the
evaluation of mode shapes and periods, (d) response-
spectrum analysis, (e) nonlinear dynamic analysis for
ground motions defined by acceleration records, (f)
nonlinear dynamic analysis for ground motions
defined by displacement records, and (g) nonlinear
dynamic analysis for specified initial nodal velocities
(for shock analysis). The program is sufficiently flex-
ible to allow a building (or structure) to be analyzed
for several analysis segments (or types), thus facili-
tating sequential static and dynamic analysis.
Loads are input as either patterns for static loads or
as records for dynamic loads. Seven different load
types are available with DRAIN-2DX, including (a)
static element load patterns - typically used for
gravity loads, (b) static nodal load patterns consisting
of vertical, lateral. and rotational loads applied on
nodes for gravity and static analysis segments, (c)
ground acceleration records, (d) ground displacement
records, including an allowance for multiple support
excitation, and (e) response spectra.
DRAIN-2DX can perform both static and dynamic
analysis. In static analysis, the load is typically
applied in a number of steps. The program selects
load substep sizes within each step by projecting the
next stiffness change (known as an event) and termi-
nating the substep at that event. The structure stiff-
ness is then changed at the end of each substep, and
. the analysis is continued for the following substep.
ATC-19 AppendiX A: Evaluation of Building Strength and Ductility
The static analysis segment is complete once either
the entire load has been applied or a target displace-
ment value is reached. In dynamic analysis, the time
step can be selected to be constant or variable. Fur-
ther options for dynamic analysis include event cal-
culations within time steps and corrections at the end
of each time step to improve the energy balance or
equilibrium.
Second-order (or P - A) effects can be modeled in
DRAIN-2DX by considering geometric stiffness for
each element, and including second-order forces in
the calculation ofthe resisting forces. For static anal-
ysis, the geometric stiffness is modified at each
event. For dynamic analysis, the geometric stiffness
can be kept constant or allowed to vary.
Six element types are currently available in the
DRAIN-2DX element library: (I) a truss element,
Type 01; (2) a beam-column element, Type 02; (3) a
connection element, Type 04; (4) a panel element,
Type 06; (5) a link element, Type 09; and (6) a fiber
beam-column element, Type IS. Of these six ele-
ments, the most commonly used are the truss, the
beam-column and the connection elements. Some
introductory remarks on these three elements folIow.
The reader is referred to Prakash et al. (1992) for
additional information.
Truss elements transmit axial loads only and can be
arbitrarily oriented in the X-Yplane. The inelastic
response of these elements can be specified as either
yielding in tension and elastic buckling in compres-
sion or yielding in both tension and compression. A
two-component parallel model (an element consist-
ing of elastic and inelastic components in parallel) is
used to capture strain-hardening effects.
Beam-column elements possess axial and flexural
stiffness and can be arbitrarily oriented in the X-Y
plane. Shear deformations and rigid-end offsets can
be accounted for in the beam-column element. Yield-
ing is concentrated in the plastic hinges at the ele-
ment ends, and strain-hardening is approximated by a
two-component parallel model. Different yield
moments can be specified at the two element ends as
well as for positive and negative flexure - two fea-
tures necessary to model reinforced concrete col-
umns and beams. Gravity and other static loads
applied to an element can be captured by specifying
end clamping or fixed-end forces. Second-order
effects can be included by introducing equilibrium
37
-----------------"--_._..- .....
 correction and geometric stiffness as noted above.
Three modes of deformation are available to beam-
column elements - axial deformation, flexural rota-
tion at element end 1, and flexural deformation at
element end 2. A plastic hinge forms when the
moment in the element reaches the yield moment.
Inelastic axial deformations are assumed not to
occur; that is, a beam-column cannot yield in axial
tension or compression.
The connection element connects two nodes with
identical coordinates in theX-Yplane. This element
can connect either rotational displacements of the
nodes or the translational displacements ofthe nodes,
and it can be specified to achieve complex inelastic
behaviors. A common applica~ion for this element is
the modeling of beam-column panel zones in steel
frames.
A.2.6 Lateral Load Profiles for Analysis
Lateral loads should be applied in patterns that both
approximately capture the vertical distribution of
inertial forces expected in the design earthquake and
account for the horizontal distribution of inertial
forces in the plane of each floor diaphragm. Load
patterns that bound the plausible distributions ofiner-
tial force should be considered for design.
Two vertical distributions of inertial force commonly
used for nonlinear static analysis are the distribution
defined by the first-mode shape ordinates ofthe
building and the constant acceleration distribution,
which corresponds to the formation of a weak first
story. For flexible buildings, a vertical distribution of
seismic force that reflects the likely contributions of
higher modes should be considered.
A.2.7 Target Displacement Calculation
The method most commonly used to evaluate the tar-
get displacement is based on the assumption that
elastic and inelastic displacements are equal; that is,
the inelastic displacement of a SDOF oscillator with
initial (elastic) period T is equal to the elastic spectral
displacement calculated using period T. This
assumption is based primarily on the work of
Miranda and Bertero (1994) who demonstrated by
exhaustive analysis that for periods greater than 0.5
second (for a rock site), mean elastic displacements
were approximately equal to mean inelastic displace-
ments. This assumption should be carefully reviewed
by the design professional calculating a target dis-
placement, because the scatter in the ratio of elastic
38 AppendiX A: Evaluation of Building Strength and Ductility
,_._--.------------ ._--------------_._-_..---_._ _ _
to inelastic displacements is large - for any given
ground motion the ratio of elastic to inelastic dis-
placements could range between 0.5 and 2.0. A con-
servative approach to calculating the target
displacement, in the absence of additional informa-
tion, would be to increase the target displacement by
between 50 percent and 100 percent; that is, to
assume that the inelastic displacement is equal to 1.5
to 2.0 times the elastic displacement.
Nonlinear static analysis makes use offorce-defor-
mation relationships for beams and columns that are
generally based on monotonic force-deformation
analysis. This assumption will likely be adequate for
buildings designed to experience less than three dis-
placement cycles to between 80 percent and 100 per-
cent ofthe target displacement. On the other hand,
consider a building in the near-source zone with a
fundamental period of 0.5 second subject to a Richer
magnitude 7.5 event - this building may be sub-
jected to 10 to 20 displacement cycles to between 80
percent and 100 percent of the target displacement.
The strength and stiffness of the structural compo-
nents and elements in this building will most likely
degrade substantially over the course of the 10 to 20
displacement cycles. The question thus arises as to
how the design professional should account for the
effects ofcumulative damage. At present, there are
no definitive answers for building framing systems
(Reinhom, private communication). In the absence of
definitive data, the design professional should reduce
the monotonic deformation capacity of structural
framing elements and components to indirectly
account for the deleterious effects of prolonged
strong ground shaking.
A.3 Seismic Evaluation of an Example
Building
A.3.1 Description of Building
The building selected for sample analysis is a seven-
story reinforced concrete building located in Los
Angeles, approximately 13 miles south ofthe epicen-
ter of the 1971 San Fernando earthquake. This build-
ing was damaged in both the 1971 earthquake and the
1994 Northridge earthquake.
This building was the subject of detailed analysis fol-
lowing both the 1971 earthquake (DOC, 1973) and
the 1994 earthquake (Lynn, private communication).
The latter analysis effort was funded by the Federal
Emergency Management Agency to verify the non-
ATC-19
.. .
linear static analysis procedures being developed for
the ATC-33 project (in progress). The analysis
results presented below are an extension of the
FEMA study. The results of this study are contained
in a background report to ATC-33 Guidelines and
Commentary for the Seismic Rehabilitation ofBuild-
ings.
The 63,000-square-foot building, designed in 1965,
is approximately 62 feet by 160 feet in plan. The typ-
ical framing consists of columns on a 20-foot (trans-
verse) by 19-foot (longitudinal) grid. Spandrel beams
are located on the perimeter frames. The floor system
is a reinforced concrete flat slab, 10 inches thick at
the second floor, 8.5 inches thick at the third to sev-
enth floors, and 8 inches thick at the roof. The
ground floor is a four inch thick slab-on-grade, and
the foundation is piled. A typical floor framing plan
is presented in Figure A-I. Atypical transverse sec-
tion and typical beam and column details are pre-
sented in Figure A-2. Interior columns are 18 inches
square and exterior columns are 14 inches by 20
inches in plan. Spandrel beam sizes are shown in Fig-
ure A-2.
The seismic framing system is composed of interior
slab-column moment frames and perimeter beam-
column moment frames. The design base shear force
at the working stress level was (DOC, 1973):
v = ZKCW = 1.0xO.67xO.057xW=O.04W
(A-3)
The north face of the building, along column line D,
has four bays of masonry infill between the ground
and second floor level, all at the eastern end of the
structure, between column lines 5 and 9. For simplic-
ity, these infill walls were not included in the mathe-
matical model described below. (Were this
evaluation to be used for the purpose of seismic reha-
bilitation, the infill walls would have been included
in the mathematical model.)
The reader is referred to the Department of Com-
merce report (DOC, 1973, pp. 359-393) for addi-
tional information regarding the design and
construction of this sample building.
A3.2 Modeling of the Building
The first question confronting the engineer charged
with evaluating the building is what to include in the
ATC19 Appendix A: Evaluation of Building Strength and Ductillty
,------r---.-----,.,,----.-..---...
mathematical model of the building. Although some
design professionals would choose to exclude the
interior slab-column framing from the mathematical
model of this framing system, it is inappropriate to
do so in this instance, as is demonstrated below.
The second question to be answered is whether the
building can be represented using two-dimensional
mathematical models; that is, uncoupling the three-
dimensional independent framing systems along each
principal axis of the building. In the sample building,
the torsional response is small- especially so after
the infill masonry walls are removed. As such, Lynn
(private communication) modeled the three-dimen-
sional building with two two-dimensional framing
systems - one per building axis.
Since the purpose of this appendix is to demonstrate
the use of nonlinear static analysis, only the results of
the analysis of the longitudinal framing are summa-
rized. The reader is referred to the aforementioned
background report for additional information.
Two exterior frames and two interior frames were
included in the mathematical model of the longitudi-
nal framing. The mathematical model of one-half of
the framing system is presented in Figure A-3. The
mathematical models of the interior and exterior
frames were linked together with rigid struts to simu-
late the assumed rigid floor diaphragm. The reactive
weights assigned to the seven suspended floors are
presented in Table A-I.
A3.3 Modeling of Key Elements
The mathematical model of the frames was com-
posed of columns and beams. Beam-column joints
were not included in the model.
The reinforced concrete columns were modeled
using their gross-section stiffness. The axial force-
moment yield surfaces were established using stan-
dard interaction curves, with capacity reduction fac-
tors (c\> ) set equal to 1. A strain-hardening ratio of
five percent was assumed for all columns.
The exterior reinforced concrete beams were mod-
eled as L-beams. The slab width assumed to contrib-
ute to the strength and stiffuess of the edge beams
was set equal to 30 percent of the perpendicular span
(often termed 12 ), The strength and stiffness values
of the interior slab-beams were calculated using a
slab width equal to 60 percent of the perpendicular
span. Beam and slab beam stiffness values were esti-
39
 0

r
0

:'

il
~
&
J
1

. .o~ ... ... ! .....

".....
eb u
~ .
Figure A-1 Typical floor plan of sample building (DOC, 1973, p. 363).

40 Appendix A: Evaluation of Building Strength and Ductility ATC-19

1--'----.:---------- ,-----------_._------_.-_.__ __ _ . .... ...

 I i~
I .41-.8 I .h~.l1
I ~.- .. I .'.-.. I .41-.8 I ..8-,1:1

I~
o '.' '.;: .'
.. ' .'

~
ti t! ~. ~ ~. l r 1

'.

'.

o z:

.......
;
2
u
w

w
..
~
....
z:
<

..
o~... .- . . "*
<
......
~
":f,- .1- ~ ,.... ~
. .f- ~f,-
. :f- '- ~
.0, - ..
~ I-

3
-
.'.
r ~
0- '-./
");1

1'_ .-,.
~OJa"" CP18
ii!tt-1r Tr-P.: I !
':' 1
i
11 . .r

o
I

~
..., : . ;:u ..I
i,n
~
. ;1
'2
J I:
..
1 =: i!
!

I ,.~(r::1
J
]
:
!: ii

a "71
J
:a
1
.."
: n
11 :

!: ~s ..<=
: <
< ':~"1<:' "';Y<::::',.'.:-:,:.,.:: ;'(I:I~-L( ::.<:f' . Cl

"2
" : ' ' '," . ', '0 . . . . . ' _

~
" , . . '. : : .' : "
."
0
I~ ~ 4(
.....fM
_..,..

..."..........
.....
__ .... ': . .
u
..
~_~_i:_~- " '.' ' '-.t...t----~A
" <
l....,.. ~.=<t';j.:.Ao&
1.,- ....::-l'::=:;:;-;;:r. u

..
0:
....

Figure A-2 Transverse section and typical details of sample building (DOC, 1973, p. 365).

ATC-19 Appendix A: Evaluation of Building Strength and Ductility 41

,------------------------_._._--_ _ .
 \,
Extenor Frame r Rigid Links

\,
Interior Frame

....,
....,
--<- 1--.[ ....,
.....

Figure A-3 Mathematical model of the sample building longitudinal framing.

shapes are shown in Figure A-4.

Table A-l Sample Building Reactive Weights
A.3.5 Nonlinear Static Analysis Results
Floor level Weight
(kips) The key results of three analyses of the sample build-
ing are presented below. The data from the first two
Roof 1185
analyses are presented to demonstrate differences in
7th 1350 response resulting from the use of different load pro-
6th 1350 files; Analysis 1 uses a triangular load profile, and
Analysis 2 a rectangular profile. The results ofthe
5th 1350
third analysis are presented to demonstrate the
4th 1350 importance of including the interior frame in the
mathematical model. Analysis 3 uses a triangular
3rd 1350
load profile but considers only the response of the
2nd 1548 perimeter (exterior) frame.

mated as one-half of gross stiftbess; flexure yield
surfaces were established using nominal material
properties. A strain-hardening ratio of.five percent
was assumed for all beams and slab-beams.
A.3.4 Eigen Analysis Results
The modal periods and shapes of the building frame
were established using the eigen solver in DRAIN-
2DX. The first three modal periods and the percent-
ages of the total mass in each of these three modes
are presented in Table A-2. The first three mode
The base shear versus roof displacement relations for
Analysis 1 and Analysis 2 are presented in Figure A-
5. The strength of the framing system, calculated
using a rectangular force profile, at a roof displace-
ment of20 inches (2.5 percent roof drift), is approxi-
mately 10 percent larger than that calculated with a
triangular profile. Using the triangular profile
response data, and the equal energy method (see Sec-
tion 4.3), the yield displacement was calculated to be
approximately equal to 4.5 inches, and the yield force
to be approximately equal to 16 percent of the reac-
tive weight of the building.

TableA-2 Dynamic Characteristics in the The locations of plastic hinges in the exterior frames
Longitudinal Direction, Sample at a roof displacement of20 inches are presented in
Building Figure A-6 for Analysis 1 and Analysis 2. The mech-
anisms associated with the two force profiles are dif-
Mode Perbd %o(TotaJ
(sees.) Mass ferent - the triangular profile results in a sway
mechanism involving the lower four stories and the
1 1.33 84 rectangular profile results in concomitant mecha-
2 0.45 11 nisms (i.e., a lower four-story sway mechanism and a
fourth story sway mechanism). Although the exist-
3 0.26 3

42 Appendix A: Evaluation of Building Strength and Ductility ATC-19

._-------,----------------_ ----_ . ..
 ------_
------ ..

Mode 1 Mode 2 Mode 3

Figure A-4 Mode shapes of sample building in the longitudinal direction.

0.25...-------r------.,...-----..----------.

. .
0.2 . , . : .. ~.:.:.~ 0":" ~ "....' _. _ .~.~. ':"'": ,-:-.-:.:-:.
...
-~
-c
.~ 0.15 ...........///
f~
.... : "
~-=_~__

,
__-..... ~,:"":""':7--::::--,_ N._---:_....__.::.... ......_:=.::..;..~.~.:_.:-:::.:;;.:.:: __.~=::.. .....

, .
. .
:E
o
Q) II
I!
. . ..
...
(.)

.......I f:~ ;..... ~ ............ ~ ..............

m
~
0.1 .
rJ)
I Analysis 1: Triangular pattern
Q)
rJ)
a3 I . - - - Analysis 2: Rectangular pattern
CD I ., .
0.05 .. l, . . .
~

OL.- ....l..- ---JL...- .....L.. ----l

o 5 10 15 20
Displacement (inches)

Figure A-S Base shear versus roof displacement relations for sample building analyses 1 and 2.

ATC19 Appendix A: Evaluation of Building Strength and Ductility 43

---------_._------,.. - _-_ _..


(a) Analysis 1
(b) Analysis 2
Figure A-6 Plastic hinge locations, roof displacement of 20 inches, sample bUilding analysis 1 and 2.
ence of two mechanisms may seem counterintuitive,
it should be noted that only the hinges in the perime-
ter frame are shown in Figure A-6 and the force-
deformation relationship for the interior frames plays
a key role in the force-deformation response ofthe
building.
A typical column and beam in the second story of the
perimeter frame (denoted C and B in Figure A-3,
respectively) were each analyzed for the purpose of
demonstrating part of a typical seismic evaluation
procedure.
For the sample column, the maximum rotation capac-
ity of the subject column was calculated to be 0.005
44 Appendix A: Evaluation of Building Strength and Ductility
r'.- ----.....----------",----
denotes plastic
hinge
+
(
radian. This calculation was based on a plastic hinge
length ofO.5d (8 inches) and an axial load equal to
the sum of the plastic beam shear forces and dead
loads above the second story. This maximum rotation
of 0.005 radian was realized at roof displacement
values of 12 inches and 10 inches, for Analysis 1 and
Analysis 2, respectively.
The maximum rotation capacity of the sample beam
was estimated to be 0.03 radian, assuming a plastic
hinge length ofO.5d (14 inches). This maximum
beam rotation was reached at roof displacement val-
ues of 19 inches and 16 inches, for Analysis 1 and
Analysis 2, respectively.
ATC-19
 0.25.--------r-----,--,------.-- .....,
Analysis 1: All frames
Analysis 3: Exterior frames only
0.2 ... ... ... ... ... ... ... ... ... ... ... ... ... ... ............... '"0 ... ... ... ... ... ... ... ... ... ... ... ... .. ... ... ... ... ... ... ...

-~
f

E
.~
(,)
0.15 ....................... ,. ... ... ... ... ... ... . ... . . ... ... ... ... ... ... ... ... ...

~o
...
(,)

m 0.1
or.
... .. .
' . ..'-. . - - -. - . - -
'... ~ ...

--
(/)
__ ... -- I
... ... ... ... ... ...
Q)

~ --'.
CD
0.05 . . . /'" . . . '.. .. . ..
'.' ........ . ~

5 10 15 20
Displacement (inches)

Figure A-7 Base shear versus roof displacement relations for analyses 1 and 3.

This demand-capacity evaluation is simply intended number of columns and a seismic retrofit requiring
to demonstrate the nonlinear static procedure. The the provision of a new seismic framing system. Fur-
sample evaluation is by no means sufficiently rigor- ther, a decision to exclude the interior frame from the
ous for the seismic evaluation of existing construc- analysis model could result in a flexible retrofit solu-
tion. In a full evaluation, all beams, columns, joints, tion incapable of protecting the existing framing sys-
and components should be examined closely. In this tern.
example, the sample column was assumed to be flex-
ure-critical; this column is actually shear-critical, and A.4 Estimation of Strength, Ductility,
it could not accommodate the shear forces associated and R Factors
with a plastic hinge rotation of 0.005 radian.
A4.1 General
To demonstrate the importance of considering all of
The calculation of strength and ductility factors is
the structural framing in the mathematical model,
demonstrated in this section by use of a force-dis-
consider the base shear versus roof displacement
placement relationship established in Section A.3 for
relationships for Analysis 1 and Analysis 3 presented
the seven-story nonductile reinforced concrete
in Figure A-7. The data presented in this figure dem-
moment frame. For the purpose ofthis discussion,
onstrate that the stiffness of the exterior frames and
the results ofAnalysis 1 (triangular load pattern) are
interior frames is similar - that is, the stiffness of
used to derive estimates of the strength and ductility
the interior slab-column frame approaches that ofthe
factors, and the maximum roof displacement is
exterior beam-column frame. It also shows that the
assumed to be eight inches. This roof displacement
strength of the interior frames and exterior frames is
estimate ignores both the likelihood of shear failure
similar. If a designer were to ignore the stiffness and
in the non-ductile columns and the limited deforma-
strength of the interior frames, the fundamental
tion capacity of the interior frame column-slab con-
period of the building would be overestimated by
nections. These assumptions would not be valid were
40% and the target displacement overestimated by a
this an evaluation for the purpose of assessing the
factor approaching two. Such an error in judgment
seismic vulnerability of the building. The displace-
might mean the difference between a ~eismic r~tr~fit
ment capacity of the frame is reduced from 12 inches
involving jacketing and/or strengthemng of a limited

ATC-19 AppendiX A: Evaluation of BUilding Strength and Ductility 45

._---------,,-,--- ,---------_._-_._-,.. _----_ _ .. ,.

(see Section A.3.5) to 8 inches to reflect the likely
degradation in defonnation capacity resulting from
multiple cycles of loading (see Section A.2.7).
A.4.2 Strength Factor
The force-displacement relationship for Analysis I is
reproduced in Figure A-8. The base shear force (Vo)
at a roof displacement of eight inches is approxi-
mately 17 percent of the reactive weight ofthe build-
ing. The design base shear for this building at the
strength level (Vd ), calculated by multiplying the
working-stress design base shear (Equation A-3) by a
seismic load factor of 1.40, is approximately 0.06W.
Using these data, the strength factor is calculated as:
Va 0.17W
Rs = Vd = 0.06W = 2.8
Note that if the interior frame was not included in the
mathematical model (as was probably the case when
the building was designed), the base shear force at a
roof displacement of eight inches is approximately
equal to 0.08W, and the resulting strength factor
would be equal to 1.3.
A.4.3 Ductility Factor
A bilinear approximation to the calculated force-dis-
placement relationship is shown in Figure A-8. This
approximation is based on the equal-energy method
(see Section 4.3 for details) and assumes that the
yield force (Vy) is equal to Vo. The maximum dis-
placement (~m) is eight inches and the yield displace-
ment (~y) is six inches. Using these data, the
displacement ductility ratio is calculated as:
~m 8
J.l = ~y = 6 = 1.3
The ductility factor can be calculated using either the
Nassar and Krawinkler (1991) or Miranda and Bert-
ero (1994) relationships. For the purpose of this dis-
cussion, the Miranda and Bertero equation (Equation
4-12) for a rock site is used to estimate the ductility
factor. For a fundamental period of 1.33 seconds and
a ductility ratio of 1.3, <I> is equal to 0.76, and the
ductility factor is equal to:
J.l- 1 1.3 -1
RIJ. = <i>'""" + 1 = 0.76 + 1 = 1.4
Using these data and a strain-hardening ratio equal to
Appendix A: Evaluation of Building Strength and Ductility
46
_._---,
opercent (consistent with the assumptions made in
constructing the bilinear approximation to the force-
displacement relationship), the ductility factor calcu-
lated using the Nassar and Krawinkler relation
(Equation 4-10) is equal to 1.3.
A.4.4 Summazy
Estimates for the strength and ductility factors for the
sample building have been made in the preceding
two subsections for the purpose of demonstrating
how the results of nonlinear static analysis can be
used to evaluate key components of the response
modification factor. These estimates of 2.8 for the
strength factor and 1.3 for the ductility factor, are
likely upper bounds on the probable values if column
shear strength and joint defonnation capacity are
(A-4)
considered. Note that if the Krawinkler and Nassar
modification factor for MDOF systems is considered
(see Section 4.5.2.3 and Figure 4-9 for details), the
ductility factor will be reduced.
The seismic framing system in the sample building is
composed of a perimeter beam-column moment
frame and an interior slab-column moment space
frame. If the interior space frame is ignored, the
building has good redundancy in the longitudinal
direction, with 16 vertical bays of seismic framing
(eight per face) but only marginal redundancy in the
transverse direction with six vertical bays of seismic
framing. If the interior space frame is included in the
evaluation, and assuming that the stiffness and
strength of the interior framing are similar to those of
the perimeter framing, the subject building has excel-
lent redundancy (RR equal to 1.0) in both the longitu-
dinal and transverse directions.
(A-S)
The response modification factor (R) could be calcu-
lated as the product of the strength-ductility, and
redundancy factors. For this example, R is equal to:
R = R s R IJ. RR = 2.8 x 1.3 x 1.0 = 3.6 (A-7)
For the reasons cited above, this value for R is
approximate only. Similar buildings constructed in
the early 1960s will likely have values of R that vary
between 2 and 4. Much additional study (see Chapter
5 for details) is needed before values of R can be
(A-6) rationally assigned to either new or existing seismic
framing systems.
ATC-19
------....,.---_._------,.. _._---_._, ..... ,""
 0.25 r - - - - - - r - - - - , - - - - - , - - - - - - - ,
- - Analysis 1: All frames
. - - - Bilinear approximation
0.2 ; .

~
' -'
Maximum base shear
C I
coefficient = 0.17
.~ 0.15 .
.2 .
:t:
Q)
o I :
o
... I . . .
~ 0.1 ......... f:::
.s:::
(fl I
~ /
~ / Design base shear coefficient = 0.06
0.05 / : ; ; .
/
'I

o L----"-------_..L- .....L...- .....J

o 5 10 15 20
Displacement (inches)

Figure A-8 Base shear versus roof displacement relation for analysis 1.

ATC-19 Appendix A: Evaluation of BUilding Strength and Ductility 47

- - - 0 _ _- -
-------.,---_._----_.__._--_._ .." ........
Appendix B: Glossary of Terms
Acceleration:
Effective Peak Acceleration: A coefficient rep-
resenting ground motion at a period of about OJ
sec.
Effective Peak Velocity: A coefficient repre-
senting ground motion at a period of about 1.0
sec.
Allowable-Stress Design (ASD): A method of pro-
portioning structures such that the computed elastic
stress does not exceed a specified limiting stress -
also known as Working-Stress Design (WSD).
Base: The level at which the horizontal seismic
ground motions are considered to be imparted to the
building.
Base Shear: Total horizontal force or shear at the
base.
Building: Any structure, fully or partially enclosed,
used or intended for sheltering persons or property.
Critical Damping: The smallest level of damping
for which no oscillation occurs in free vibration
response.
Damping: A measure of energy dissipation. Damp-
ing in a structure is typically defined in terms of per-
cent of critical damping.
Design Action: The force in an element (axial, flex-
ure, shear, torque) for which the element is to be
designed.
Design Basis Earthquake: The earthquake that pro-
duces ground motions at the site under consideration
that have a 10 percent probability of being exceeded
in 50 years.
Diaphragm: A horizontal, or nearly horizontal, sys-
tem designed to transmit seismic forces to the verti-
cal elements of the seismic force-resisting system.
Ductility: A measure of the ability of a material, ele-
ment, or system to deform beyond yield.
Ductility Ratio: The ratio of maximum deformation
to yield deformation - only applies to deformations
larger than yield. Ductility ratio can be introduced at
ATC-19 Appendix B: Glossary of Tenns
,--------_._----_.. _--_._
the material, element, or system level.
Element: One of the major parts that compose the
structural system. A structural wall or a moment
frame is an element. The boundary member of a
structural wall, or the individual column or beam in a
frame is a component of an element.
Failure: The inability to satisfy a predetennined
limit state. For the limit state of collapse, failure is
often defined as the displacement at which the resis-
tance (base shear) of the building falls below 80 per-
cent of the maximum resistance. Other definitions of
failure could include (a) drifts exceeding predeter-
mined limits, and (b) plastic hinge rotations exceed-
ing predetermined values.
Flexure-critical: The failure mode of a component
or element that is governed by flexural response.
Frame:
Braced Frame: An essentially vertical truss, or
its equivalent, of the concentric or eccentric type
that is provided in a building frame or dual sys-
tem to resist seismic forces.
Moment Frame: A frame in which members
and joints are capable of resisting forces by flex-
ure as well as along the axis of the members.
Types of moment frames are: intermediate
moment, ordinary moment, and special moment.
Frame System:
Building Frame System: A structural system
with an essentially complete space frame provid-
ing support for vertical loads. Seismic force
resistance is provided by structural walls or
braced frames.
Dual Frame System: A structural system with
an essentially complete space frame providing
support for vertical loads. Seismic force resis-
tance is provided by a moment resisting frame
and structural walls or braced frames.
Space Frame System: A structural system com-
posed of interconnected members other than
bearing walls that is capable of supporting verti-
cal loads and that also may provide resistance to
49
.
 seismic forces.
Load:
Dead Load (QD): The gravity load due to the
weight of all permanent structural and nonstruc-
tural components of a building such as floors,
roofs, and the operating weight of fixed service
equipment.
Gravity Load (W): The total load and the appli-
cable portions of other loads.
Live Load (QrJ: The load superimposed by the
use and occupancy of the buildings, not includ-
ing the wind load, earthquake load, or dead load.
The live load may be reduced for tributary areas
as permitted by the building code administered
by the regulatory agency.
Load Factor: A factor by which a nominal load
effect is multiplied to account for the uncertainties
inherent in the determination of the load effect.
Partial Safety Factor: The factor by which the ele-
ment safety factor should be modified to account for
its presence in the structural system, so that the over-
all failure probability is similar to that desired for the
element.
Performance Objective: A level of seismic func-
tionality that a building owner or occupant expects of
a structure. Sample performance objectives include
no collapse, preservation of life safety, damage con-
trol, immediate occupancy, and fully functional.
Redundancy: A measure of the number of lines of
vertical seismic framing in a building. The
50 Appendix B: Glossary of Terms
...._ .. _----- . _----------------------
'---..,.------,-_.'----, ----------,-----'------_."----"-_._'--
greater the number of lines of vertical seismic
framing of similar strength and stiffness, the
greater the redundancy (and reliability) of the
seismic framing system.
Reserve Strength: The difference between design
strength and maximum strength.
Resistance: The maximum load-carrying capacity,
as defined by a limit state.
Risk: Exposure to loss. Risk is defined as the proba-
bility of seismically-induced unacceptable per-
formance.
Shear-critical: The failure mode of a component or
element that is governed by shear response.
Story Drift Ratio: The relative displacement of two
adjacent floors, divided by the story height.
Story Shear: The summation of design lateral force
at the level above the story under consideration.
Strength Design: A method of proportioning struc-
tures based on the ultimate strength of critical sec-
tions.
Uniform Risk: A term used to describe equallikeli-
hood of loss or damage.
Wall: A component, usually placed vertically, used
to enclose or divide space.
Wall System, Bearing: A structural system with
bearing walls providing support of all or major por-
tions of the vertical loads. Structural walls or braced
frames provide seismic force resistance.
ATC-19
..--.,,.
 References
ATC, 1974,An Evaluation ofa Response Spectrum
Approach to the Seismic Design ofBuildings,
ATC-2 Report, Applied Technology Council,
Redwood City, California.
ATC, 1978, Tentative Provisions for the Develop-
ment ofSeismic Regulationsfor Buildings, ATe-
3-06 Report, Applied Technology Council, Red-
wood City, California.
ATC, 1982a, Seismic Design Guidelinesfor Highway
Bridges, ATC-6 Report, Applied Technology
Council, Redwood City, California.
ATC, 1982b, An Investigation ofthe Correlation
Between Earthquake Ground Motion and Build-
ing Performance, ATC-IO Report, Applied
Technology Council, Redwood City, California.
ATC, 1995, A Critical Review ofCurrent
Approaches to Earthquake Resistant Design,
ATC-34 Report, Applied Technology Council,
Redwood City, California.
ATC, in progress. Guidelines and Commentary for
the Seismic Rehabilitation ofBuildings, Volumes
I and II, ATC-33.03 Report, Applied Technology
Council, Redwood City, California.
BSSC, 1985, 1988,1991, 1994, NEHRP Recom-
mended Provisions for the Development ofSeis-
mic Regulationsfor New Buildings, Building
Seismic Safety Council, Washington, D.C.
Bertero, V.V., 1986, "Evaluation of response reduc-
tion factors recommended by ATC and
SEAOC," Proceedings ofthe Third U.S.
National Conference on Earthquake Engineer-
ing, Charleston, North Carolina.
Caltrans, 1990, Bridge Design Specifications Man-
ual, California Department of Transportation,
Division of Structures, Sacramento, California.
Clough, R.W. and J. Penzien, 1993. Dynamics of
Structures, McGraw Hill, New York.
CEC, 1988, Structures in Seismic Regions - Design -
Part 1, Eurocode No.8, Commission of the
ATC-19 References
,_._-----
I
--_" - ,----_---:.._-;----------~-_
European Communities, Luxembourg.
DOC, 1973, San Fe mando, California Earthquake of
February 9, 1971, Vol. 1, ed. L.M. Murphy, U.S.
Department of Commerce, Washington, D.C.
DOD, 1986, Seismic Design for Essential Buildings,
TM-5-809-10-1, Departments of the Army,
Navy, and Air Force, Washington, D.C.
Freeman, S.A., 1990, "On the correlation of code
forces to earthquake demands," Proceedings of
the 4th U.S.-Japan Workshop on Improvement of
Building Structural Design and Construction
Practices, ATC-15-3 Report, Applied Technol-
ogy Council, Redwood City, California.
Gomez, R. and F. Garcia-Ranz, 1988, "Complemen-
tary technical norms for earthquake resistant
design", Earthquake Spectra, EERI, 4 (3): 441-
460.
Hwang, H. and M. Shinozuka, 1994, "Effect of large
earthquakes on the design of buildings in eastern
United States," Proceedings ofthe Fifth U.S.
National Conference on Earthquake Engineer-
ing, Chicago, Illinois.
lAEE, 1992, Earthquake Resistant Regulations, A
World List, International Association for Earth-
quake Engineering, Tokyo, Japan.
ICBO, 1961, 1976,1985,1988,1991,1994, Uniform
Building Code, International Conference of
Building Officials, Whittier, California.
Krawinkler, H. and A.A. Nassar, 1992, "Seismic
design based on ductility and cumulative damage
demands and capacities," Nonlinear Seismic
Analysis and Design of Reinforced Concrete
Buildings, Fajfar, Krawinkler, edd, Elsevier
Applied Science, New York.
Miranda, E. and V.V. Bertero, 1994, "Evaluation of
strength reduction factors for earthquake-resis-
tant design," Earthquake Spectra, EERI, 10 (2):
357-379.
Moses, F., 1974, "Reliability of structural systems,"
51
..__._._._--_.... _...
 Journal of the Structural Division, ASCE, 100
(ST9): 1813-1820.
Nassar AA and H. Krawinkler, 1991, "Seismic
Demandsfor SDOF and MDOF Systems," John
A. Blume Earthquake Engineering Center,
Report No. 95, Stanford University, Stanford,
California.
Newmark, N.M. and W.J. Hall, 1982, Earthquake
Spectra and Design, EERI Monograph Series,
EERI, Oakland.
Osteraas, J.D. and H. Krawinkler, 1990, Strength and
Ductility Considerations in Seismic Design, John
A. Blume Earthquake Engineering Center,
Report 90, Stanford University, California.
Paulay, T. and MJ.N. Priestley, 1992, Seismic
Design ofReinforced Concrete and Masonry
Buildings, John Wiley and Sons, New York.
Prakash, V., G.H. Powell, and F.C. Filippou, 1992,
DRAIN-2DX' Base Program User Guide,
Department of Civil Engineering, Report No.
UCB/SEMM-92/29, University of California,
Berkeley, California.
Riddell, R. and N.M. Newmark, 1979, Statistical
Analysis of the Response ofNonlinear Systems
Subject to Earthquakes, Civil Engineering Stud-
ies, Structures Research Series, 468, University
of Illinois, Urbana, Illinois.
Rojahn, C., 1988a, "An investigation of structural
response modification factors," Proceedings of
the Ninth World Conference on Earthquake
Engineering, Kyoto, Japan.
Rojahn, C. and G.C. Hart, 1988b, "U.S. code focus-
ing on R-factor ofUBC, ATC-3, and NEHRP,"
52 References
.----.---------_..._----
Proceedings ofthe 3rd US-Japan Workshop on
Improvement ofStructural Design and Construc-
tion Practices, ATC-15-2 Report, Applied Tech-
nology Council, Redwood City, California.
SEAOC, 1959, 1974, 1985, 1988, 1990, Recom-
mended Lateral Force Requirements and Com-
mentary, Seismology Committee, Structural
Engineers Association of California, Sacra-
mento, California.
Uang, C.M. and V.V. Bertero, 1986, Earthquake
Simulation Tests and Associated Studies ofa 0.3-
Scale Model ofa Six-Story Concentrically
Braced Steel Structure, Earthquake Engineering
Research Center, Report No. UCBIEERC-86110,
University of California, Berkeley, California.
Uang, C.M. and A Maarouf, 1993, "Safety and econ-
omy considerations ofUBC seismic force reduc-
tion factors," Proceedings ofthe 1993 National
Earthquake Conference, Memphis, Tennessee.
Whittaker, AS., C.M. Uang, and V.V. Bertero, 1987,
Earthquake Simulation Tests and Associated
Studies of a O.3-Scale Model ofa Six-Story
Eccentrically Braced Steel Structure, Earthquake
Engineering Research Center, Report No. UCBI
EERC-87/02, University ofCalifornia, Berkeley,
California.
Whittaker, AS., C.M. Uang, and V.V. Bertero, 1990,
An Experimental Study of the Behavior of Dual
Steel Systems, Earthquake Engineering Research
Center, Report No. UCBIEERC-88114, Univer-
sity of California, Berkeley, California
Wu,1. and R.D. Hanson, 1989, "Study of inelastic
spectra with high damping," ASCE Journal of
Structural Engineering, 115 (6): 1412-1431.
ATC-19