13

th
World Conference on Earthquake Engineering
Vancouver, B.C., Canada
August 1-6, 2004
Paper No. 1476


A SUMMARY OF
FEMA 440: IMPROVEMENT OF NONLINEAR STATIC SEISMIC
ANALYSIS PROCEDURES


Craig D. COMARTIN
1
, Mark ASCHHEIM
2
, Andrew GUYADER
3
, Ronald
HAMBURGER
4
, Robert HANSON
5
, William HOLMES
6
, Wilfred IWAN
7
,
Michael MAHONEY
8
, Eduardo MIRANDA
9
, Jack MOEHLE
10
, Christopher
ROJAHN
11
, Jonathan STEWART
12



SUMMARY

The Applied Technology Council (ATC), with primary funding provided by the Federal
Emergency Management Agency (FEMA) and supplemental support from the Pacific
Earthquake Engineering Research Center (PEER), is in the final stages of a project (ATC 55) to
evaluate and improve the application of inelastic analysis procedures for use with performance-
based engineering methods for seismic design, evaluation, and rehabilitation of buildings. The
project will culminate with the publication of FEMA 440: Improvement of Nonlinear Static
Seismic Analysis Procedures. This paper provides a preview of the key conclusions of the work.
The focus is on anticipated recommendations to improve inelastic analysis procedures as
currently documented in FEMA 356 [1] and ATC 40 [2]. General categories of improvements
include:
Displacement modification procedures (Coefficient Method)
Equivalent linearization procedures (Capacity Spectrum Method)
Multi-degree-of-freedom effects
Soil-structure interaction effects

1
CDComartin,Inc, ccomartin@comartin.net
2
Santa Clara University, maschheim@scu.edu
3
California Institute of Technology, guyader@caltech.edu
4
Simpson Gumpertz and Heger, ROHamburger@sgh.com
5
University of Michigan, Robert.Hanson@associates.dhs.gov
6
Rutherford&Chekene, wholmes@ruthchek.com
7
California Institute of Technology, wdiwan@cco.caltech.edu
8
FEMA, mike.mahoney@dhs.gov
9
Stanford University, miranda@ce.stanford.edu
10
University of California, Berkeley, moehle@peer.berkeley.edu
11
Applied Technology Council, crojahn@atcouncil.org
12
University of California, Los Angeles, jstewart@seas.ucla.edu
2

The publication FEMA 440: The Improvement of Nonlinear Static Seismic Analysis Procedures
will be the product of the project. This document will provide a review and discussion of
simplified inelastic seismic analysis of new and existing buildings. It will contain guidelines for
applications of selected procedures including their individual strengths, weaknesses and
limitations. The document will also contain illustrative examples and expert commentary on key
issues.

EVALUATION OF CURRENT PROCEDURES

Current nonlinear static procedures estimate the global displacement response of a building or
structure utilizing a single-degree-of freedom representation of behavior based on a nonlinear
force-displacement relationship (pushover curve) based on the monotonic static response to a
lateral load vector. The efficacy of procedures to predict SDOF responses of a nonlinear
oscillator can be investigated by comparing the estimates to actual results from multiple
nonlinear responses history analyses in a statistical format. Parameters affecting the maximum
displacement of a SDOF oscillator and the variations assumed for the evaluation study are
summarized as follows [3,4] (see Figure 1):

Predominant hysteretic behavior


Figure 1: Basic hysteretic models used in the evaluation of current procedures.



The elastic-perfectly plastic (EPP) model is used as a reference model. This model has been
widely used in previous investigations. This is a reasonable hysteretic model for steel beams
which do not experience lateral or local buckling or connection failure. It is also a good model of
Elast oplasti c Perfectly Plasti c Model
-250.00
-200.00
-150.00
-100.00
-50.00
0.00
50.00
100.00
150.00
200.00
250.00
-30.00 -20.00 -10.00 0.00 10.00 20.00 30.00
Displacement
F
o
r
c
e
Elastoplastic Perfectly Plastic
EPP
Elast oplasti c Perfectly Plasti c Model
-250.00
-200.00
-150.00
-100.00
-50.00
0.00
50.00
100.00
150.00
200.00
250.00
-30.00 -20.00 -10.00 0.00 10.00 20.00 30.00
Displacement
F
o
r
c
e
Elastoplastic Perfectly Plastic
EPP
Modifi ed Clough - St iffness Degrading Model
-250.00
-200.00
-150.00
-100.00
- 50.00
0.00
50.00
100.00
150.00
200.00
250.00
- 30.00 - 20.00 - 10.00 0.00 10.00 20.00 30.00
Displacement
F
o
r
c
e
Modified Clough
SD
Modifi ed Clough - St iffness Degrading Model
-250.00
-200.00
-150.00
-100.00
- 50.00
0.00
50.00
100.00
150.00
200.00
250.00
- 30.00 - 20.00 - 10.00 0.00 10.00 20.00 30.00
Displacement
F
o
r
c
e
Modified Clough
SD
Strength and stiffness degrading model
-600
-400
-200
0
200
400
600
-400 -300 -200 -100 0 10 0 200 300 400
Displacement
F
o
r
c
e
Strength and Stiffness Degrading
SSD
Strength and stiffness degrading model
-600
-400
-200
0
200
400
600
-400 -300 -200 -100 0 10 0 200 300 400
Displacement
F
o
r
c
e
Strength and Stiffness Degrading
SSD
Elastoplastic Perfectly Plasti c Model
-250.00
-200.00
-150.00
-100.00
-50.00
0.00
50.00
100.00
150.00
200.00
250.00
-30.00 -20.00 -10.00 0.00 10.00 20.00 30.00
Displacement
F
o
r
c
e
Nonlinear Elastic
NE
Elastoplastic Perfectly Plasti c Model
-250.00
-200.00
-150.00
-100.00
-50.00
0.00
50.00
100.00
150.00
200.00
250.00
-30.00 -20.00 -10.00 0.00 10.00 20.00 30.00
Displacement
F
o
r
c
e
Nonlinear Elastic
NE
3
the behavior of other highly ductile systems including buckling restrained braced frames (BRBF)
and eccentric braced frames (EBF).
The stiffness-degrading (SD) model is representative of well-detailed and flexurally-controlled
reinforced concrete structures whose lateral stiffness decreases as the level of lateral deformation
increases. In this model the unloading stiffness is always the same as the initial stiffness.
The strength and stiffness-degrading (SSD) model is aimed at approximately reproducing the
hysteretic behavior of structures whose lateral stiffness and lateral strength decreases when
subjected to cyclic reversals. This model does not represent systems that loose strength in the
same cycle as yielding or experience P-delta effects. This distinction in the type of strength
degradation is discussed below.
The nonlinear elastic (NE) model unloads on the same branch as the loading curve and therefore
exhibits no hysteretic energy dissipation. This model approximately reproduces the behavior of
pure rocking structures.

Basic global strength (R)
In this study the lateral strength is normalized by the strength ratio R, which is defined as
y
a
F
S m
R

= (Eqn. 1)
where m is the mass of the system, S
a
is the acceleration spectral ordinate corresponding to the initial
period of the system and F
y
is the lateral yielding strength of the system (see Figure 2). The numerator in
(1) represents the lateral strength required to maintain the system elastic, which sometimes is also referred
to as the elastic strength demand. Nine levels of normalized lateral strength were considered
corresponding to R=1, 1.5, 2, 3, 4, 5, 6, 7 and 8.



Figure 2: Global strength parameter, R

Period (T)
Single-degree-of-freedom (SDOF) systems with periods of vibration between 0.05s and 3.0s were used in
this investigation. A total of 50 periods of vibration were considered (40 between 0.05s and 2.0s equally
spaced at 0.05s and 10 periods between 2.0s and 3.0s equally spaced at 0.1s). The initial damping ratio,
, was assumed to be equal to 5% for all systems.

Ground motion
A total of 100 earthquake ground motions recorded on different site conditions were used in this study.
Ground motions were divided into five groups with 20 accelerograms in each group. The first group
(NEHRP [5] site class B) consisted of earthquake ground motions recorded on stations located on rock
with average shear wave velocities between 760 m/s (2,500 ft/s) and 1,525 m/s (5,000 ft/s). The second
group (NEHRP site class C) consisted of records obtained on stations on very dense soil or soft rock with
y
a
F
S m
R

=
mS
a
S
d
F
y
T
0
Demand spectrum
Capacity curve y
a
F
S m
R

=
mS
a
S
d
F
y
T
0
Demand spectrum
Capacity curve
4
average shear wave velocities between 360 m/s (1,200 ft/s) and 760 m/s (2,500 ft/s). The third group
(NEHRP site class D) consisted of ground motions recorded on stations on stiff soil with average shear
wave velocities between 180 m/s (600 ft/s) and 360 m/s (1,200 ft/s). The fourth group corresponds to
ground motions recorded on very soft soil conditions with shear wave velocities smaller than 180 m/s that
can be classified as type E. Finally the fifth group corresponds to 20 ground motions influenced by
forward-directivity effects.

The result of the variation in these basic parameters was a database of 180,000 nonlinear response history
analyses representing the maximum displacement response of a SDOF oscillator subject to earthquake
motions. The accuracy of the approximate nonlinear static procedures was determined by the comparing
the predictions to actual response histories as a benchmark. Detailed results of the evaluation will be
published in FEMA 440. Selected illustrative results are included in this paper in the subsequent sections
in conjunction with proposed improvements to the two NSP procedures.

STRENGTH DEGRADATION

It is important to distinguish between two different types of strength degradation. Consider the hysteretic
response of two oscillators shown in Figure 3. While both exhibit inelastic strength degradation, note that
the first (cyclic strength degrading) loses strength only in the cycles subsequent to that in which it yields.
The slope of the post-elastic portion of the curve is not negative. The post-elastic stiffness, , in any cycle
is zero or positive. In contrast, the other oscillator (in-cycle strength degrading) has a negative in the
cycle in which the yielding occurs.


Strength and stiffness degrading model
-600
-400
-200
0
200
400
600
-400 -300 -200 -100 0 100 200 300 400
Displacement
F
o
r
c
e
Cyclic strength degradation In-cyclic strength degradation
Strength loss occurs in subsequent cycles;
not in the same cycle as yield.
Strength loss occurs
in same cycle as yield.
Strength and stiffness degrading model
-600
-400
-200
0
200
400
600
-400 -300 -200 -100 0 100 200 300 400
Displacement
F
o
r
c
e
Cyclic strength degradation In-cyclic strength degradation
Strength loss occurs in subsequent cycles;
not in the same cycle as yield.
Strength loss occurs
in same cycle as yield.

= 0
Strength and stiffness degrading model
-600
-400
-200
0
200
400
600
-400 -300 -200 -100 0 100 200 300 400
Displacement
F
o
r
c
e
Cyclic strength degradation In-cyclic strength degradation
Strength loss occurs in subsequent cycles;
not in the same cycle as yield.
Strength loss occurs
in same cycle as yield.
Strength and stiffness degrading model
-600
-400
-200
0
200
400
600
-400 -300 -200 -100 0 100 200 300 400
Displacement
F
o
r
c
e
Cyclic strength degradation In-cyclic strength degradation
Strength loss occurs in subsequent cycles;
not in the same cycle as yield.
Strength loss occurs
in same cycle as yield.
Strength and stiffness degrading model
-600
-400
-200
0
200
400
600
-400 -300 -200 -100 0 100 200 300 400
Displacement
F
o
r
c
e
Strength and stiffness degrading model
-600
-400
-200
0
200
400
600
-400 -300 -200 -100 0 100 200 300 400
Displacement
F
o
r
c
e
Strength and stiffness degrading model
-600
-400
-200
0
200
400
600
-400 -300 -200 -100 0 100 200 300 400
Displacement
F
o
r
c
e
Cyclic strength degradation In-cyclic strength degradation
Strength loss occurs in subsequent cycles;
not in the same cycle as yield.
Strength loss occurs
in same cycle as yield.
Strength and stiffness degrading model
-600
-400
-200
0
200
400
600
-400 -300 -200 -100 0 100 200 300 400
Displacement
F
o
r
c
e
Strength and stiffness degrading model
-600
-400
-200
0
200
400
600
-400 -300 -200 -100 0 100 200 300 400
Displacement
F
o
r
c
e
Strength and stiffness degrading model
-600
-400
-200
0
200
400
600
-400 -300 -200 -100 0 100 200 300 400
Displacement
F
o
r
c
e
Cyclic strength degradation In-cyclic strength degradation
Strength loss occurs in subsequent cycles;
not in the same cycle as yield.
Strength loss occurs
in same cycle as yield.

= 0

Figure 3: Two types of strength degradation


The distinction between these two types of behaviors is important because the dynamic response of the
two oscillators subject to earthquake motions can be radically different. The results of the ATC 55
evaluation study, as well as other previous research, demonstrate that cyclic-strength-degrading oscillators
(SSD in Figure 1) often exhibit maximum inelastic displacements that are about the same or even less than
those that do not lose strength. The in-cycle strength-degrading counterpart, in contrast, can be prone to
dynamic instability particularly when subject to ground motions that include large velocity pulses often
associated with near field records. If one were to generate a pushover curve for each oscillator in Figure 3
5
using the second-cycle backbone procedure of FEMA 356 there might be very little difference, if any,
between the two. Additional studies were conducted as a part of the ATC 55 project to confirm and
illustrate this difference using an in-cycle strength-degrading oscillator with characteristics as illustrated
in the right side of Figure 3. These studies confirm the potential for dynamic instability and imply that
structures with significant negative post-elastic stiffness must have a critical minimum strength to avoid
collapse. This is discussed further in subsequent sections.

DISPLACEMENT MODIFICATION

The FEMA 440 document will propose several improvements to the basic displacement modification
procedure in FEMA 356 [1]. These relate to the coefficient method equation for the target displacement,

t
for estimating the maximum inelastic global deformation demands on buildings for earthquake ground
motions
g
T
S C C C C
e
a t
2
2
3 2 1 0
4
= (Eqn. 2)
where the coefficients are currently defined as follows:
C
o
= modification factor to relate spectral displacement of an equivalent SDOF system to the roof
displacement of the building MDOF system.
C
1
= modification factor to relate the expected maximum inelastic displacements to displacements
calculated for linear elastic response.
C
2
= Modification factor to represent the effect of pinched hysteretic shape, stiffness degradation and
strength deterioration on the maximum displacement response.
C
3
= Modification factor to represent increased displacements due to dynamic P- effects.

Based on the analyses of the current procedures two alternatives for improvement of the factor C
1
are were
initially considered:
ALTERNATIVE 1: ) R (
c ) T / T ( a
C
b
g e
1
1 1
1
1

+ = (Eqn. 3)
SOIL
PROFILE
a b c T
g
(s)
B 42 1.60 45 0.75
C 48 1.80 50 0.85
D 57 1.85 60 1.05
ALTERNATIVE 2: ) R (
) T / T ( a
C
b
g e
1
1
1
1

+ = (Eqn. 4)
SOIL
PROFILE
a b T
g
(s)
B 151 1.60 1.60
C 199 1.83 1.75
D 203 1.91 1.85
These are both compared to the current definition in Figure 4. Figure 5 illustrates the substantial
improvement in error reduction with either alternative
6

Figure 4: Comparison of current and potential C
1
coefficients

A simplified version of these alternatives is also under consideration for inclusion in the final
recommendations as follows:
SIMPLIFIED ALTERNATIVE:
1 2
1
1
90
R
C
T

= + (Eqn. 5)


Figure 5: Comparison of mean errors for C
1
coefficients for site class C
The current definitions C
2
and C
3
are not clearly independent of one another. C
2
is intended to represent
changes in hysteretic behavior due to pinching, stiffness degradation, and strength degradation. However,
strength and stiffness degradation due to P- effects are supposedly addressed by C
3
as well. FEMA 440
will recommend that C
2
be used to modify displacements for purely cyclic strength losses. In this case,
example results indicate that the current specification over estimates the actual effect of cyclic degradation
(see Figure 6).
WITHOUT CAPPING
0.0
0.5
1.0
1.5
2.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
PERIOD [s]
E[( ?
i
)
app
/( ?
i
)
ex
]
R=6.0
R=4.0
R=3.0
R=2.0
R=1.5
SITE CLASS C
T
s
= 0.55 s
WITH CAPPING
0.0
0.5
1.0
1.5
2.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
PERIOD [s]
E[( ? i)app /( ? i )ex ]
R=6.0
R=4.0
R=3.0
R=1.5
R=1.5
SITE CLASS C
T
s
= 0.55 s
ALTERNATIVE I
0.0
0.5
1.0
1.5
2.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
PERIOD [s]
E[( ?
i
)
app
/(?
i
)
ex
]
R = 6.0
R = 4.0
R = 3.0
R = 2.0
R = 1.5
SITE CLASS C
T
g
= 0.85 s
ALTERNATIVE II
0.0
0.5
1.0
1.5
2.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
PERIOD [s]
E[( ? i)app /(? i)ex ]
R = 6.0
R = 4.0
R = 3.0
R = 2.0
R = 1.5
SITE CLASS C
T
g
= 1.75 s
Current
Proposed
WITHOUT CAPPING
0.0
0.5
1.0
1.5
2.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
PERIOD [s]
E[( ?
i
)
app
/( ?
i
)
ex
]
R=6.0
R=4.0
R=3.0
R=2.0
R=1.5
SITE CLASS C
T
s
= 0.55 s
WITH CAPPING
0.0
0.5
1.0
1.5
2.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
PERIOD [s]
E[( ? i)app /( ? i )ex ]
R=6.0
R=4.0
R=3.0
R=1.5
R=1.5
SITE CLASS C
T
s
= 0.55 s
WITHOUT CAPPING
0.0
0.5
1.0
1.5
2.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
PERIOD [s]
E[( ?
i
)
app
/( ?
i
)
ex
]
R=6.0
R=4.0
R=3.0
R=2.0
R=1.5
SITE CLASS C
T
s
= 0.55 s
WITHOUT CAPPING
0.0
0.5
1.0
1.5
2.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
PERIOD [s]
E[( ?
i
)
app
/( ?
i
)
ex
]
R=6.0
R=4.0
R=3.0
R=2.0
R=1.5
SITE CLASS C
T
s
= 0.55 s
WITH CAPPING
0.0
0.5
1.0
1.5
2.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
PERIOD [s]
E[( ? i)app /( ? i )ex ]
R=6.0
R=4.0
R=3.0
R=1.5
R=1.5
SITE CLASS C
T
s
= 0.55 s
WITH CAPPING
0.0
0.5
1.0
1.5
2.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
PERIOD [s]
E[( ? i)app /( ? i )ex ]
R=6.0
R=4.0
R=3.0
R=1.5
R=1.5
SITE CLASS C
T
s
= 0.55 s
ALTERNATIVE I
0.0
0.5
1.0
1.5
2.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
PERIOD [s]
E[( ?
i
)
app
/(?
i
)
ex
]
R = 6.0
R = 4.0
R = 3.0
R = 2.0
R = 1.5
SITE CLASS C
T
g
= 0.85 s
ALTERNATIVE II
0.0
0.5
1.0
1.5
2.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
PERIOD [s]
E[( ? i)app /(? i)ex ]
R = 6.0
R = 4.0
R = 3.0
R = 2.0
R = 1.5
SITE CLASS C
T
g
= 1.75 s
ALTERNATIVE I
0.0
0.5
1.0
1.5
2.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
PERIOD [s]
E[( ?
i
)
app
/(?
i
)
ex
]
R = 6.0
R = 4.0
R = 3.0
R = 2.0
R = 1.5
SITE CLASS C
T
g
= 0.85 s
ALTERNATIVE I
0.0
0.5
1.0
1.5
2.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
PERIOD [s]
E[( ?
i
)
app
/(?
i
)
ex
]
R = 6.0
R = 4.0
R = 3.0
R = 2.0
R = 1.5
SITE CLASS C
T
g
= 0.85 s
ALTERNATIVE II
0.0
0.5
1.0
1.5
2.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
PERIOD [s]
E[( ? i)app /(? i)ex ]
R = 6.0
R = 4.0
R = 3.0
R = 2.0
R = 1.5
SITE CLASS C
T
g
= 1.75 s
ALTERNATIVE II
0.0
0.5
1.0
1.5
2.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
PERIOD [s]
E[( ? i)app /(? i)ex ]
R = 6.0
R = 4.0
R = 3.0
R = 2.0
R = 1.5
SITE CLASS C
T
g
= 1.75 s
Current
Proposed
SOIL PROFILE: C
R = 5
0.0
0.5
1.0
1.5
2.0
2.5
0.0 0.3 0. 5 0. 8 1.0
PERIOD [s]
C
1
Alternative 1
Alternative 2
FEMA-356
7


Figure 6: Comparison of current FEMA 356 C
2
coefficients to actual example results for
cyclic-strength-degrading oscillators

The following specification is being investigated as an improvement for FEMA 440:
2
2
1
800
1
1

+ =
T
R
C (Eqn. 6)
The plot in Figure 7 shows that this equation does not fall below a value of 1.0. This bias is likely to be
recommended since the studies do not include the effects of the duration of shaking that may be important
for structures subject to cyclic strength degradation.


Figure 7: An improved coefficient C
2
as a function of period and initial strength for cyclic-
strength-degrading behavior

Figure 8 illustrates a comparison between the current C
3
specified in FEMA 356 with actual results of
response history analyses using oscillators that exhibit in-cycle loss of strength and/or P-delta effects
result in in a negative post-elastic stiffness, . It is clear that negative post elastic stiffness can result in
dynamic instability and collapse depending on the magnitude of , as well as initial period and initial
strength. Consequently FEMA 440 will recommend elimination of the C
3
coefficient and introduce a
minimum initial strength (maximum R) requirement for systems with in-cycle strength degradation and
0
.
1
0
.
3
0
.
4
0
.
6
0
.
7
0
.
9
1
.
0
1
.
2
1
.
3
1
.
5
1
.
6
1
.
8
1
.
9
2
.
1
2
.
2
2
.
4
2
.
5
2
.
7
2
.
8
3
.
0
1
1.6
2.2
2.8
3.4
4
4.6
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
2
C
2
T
R
2
2
1
800
1
1

+ =
T
R
C
R
T
C
2
0
.
1
0
.
3
0
.
4
0
.
6
0
.
7
0
.
9
1
.
0
1
.
2
1
.
3
1
.
5
1
.
6
1
.
8
1
.
9
2
.
1
2
.
2
2
.
4
2
.
5
2
.
7
2
.
8
3
.
0
1
1.6
2.2
2.8
3.4
4
4.6
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
2
C
2
T
R
2
2
1
800
1
1

+ =
T
R
C
R
T
C
2
SITE CLASS B
0.0
1.0
2.0
3.0
0.0 0.5 1.0 1.5
PERIOD [s]
C
2
Collapse Prevention
Life Safety
Immediate Occupancy
FRAMING TYPE 1
SITE CLASSES B
(mean of 20 ground moti ons)
0.2
0.6
1.0
1.4
1.8
0.0 0.5 1.0 1.5 2.0 2.5 3.0
PERIOD[s]
C2=C1,SD/C1,EPP
R = 6.0
R = 5.0
R = 4.0
R = 3.0
R = 2.0
R = 1.5
Current (FEMA 356) Actual
SITE CLASS B
0.0
1.0
2.0
3.0
0.0 0.5 1.0 1.5
PERIOD [s]
C
2
Collapse Prevention
Life Safety
Immediate Occupancy
FRAMING TYPE 1
SITE CLASSES B
(mean of 20 ground moti ons)
0.2
0.6
1.0
1.4
1.8
0.0 0.5 1.0 1.5 2.0 2.5 3.0
PERIOD[s]
C2=C1,SD/C1,EPP
R = 6.0
R = 5.0
R = 4.0
R = 3.0
R = 2.0
R = 1.5
Current (FEMA 356) Actual
8
significant P-delta effects. This minimum strength requirement would also apply to solutions from
equivalent linearization procedures.

Current (FEMA 356)
T = 1.0s
0
1
2
3
4
5
6
7
8
0 1 2 3 4 5
R
C3
= 0.21
= 0.06
Actual
6
T = 1.0s
0
1
2
3
4
5
6
7
8
0 1 2 3 4 5 6
R
i / e
= 0.21
= 0.06
1984 Morgan Hill, California Earthquake
Gilroy #3, Sewage Treatment Plant, Comp. 0
Current (FEMA 356)
T = 1.0s
0
1
2
3
4
5
6
7
8
0 1 2 3 4 5
R
C3
= 0.21
= 0.06
Actual
6
T = 1.0s
0
1
2
3
4
5
6
7
8
0 1 2 3 4 5 6
R
i / e
= 0.21
= 0.06
1984 Morgan Hill, California Earthquake
Gilroy #3, Sewage Treatment Plant, Comp. 0

Figure 8: Comparison of current FEMA 356 C
3


coefficients to actual example results for in-
cycle strength-degrading oscillators

EQUIVALENT LINEARIZATION

The capacity spectrum method documented in ATC 40 [2] is a form of equivalent linearization based on
two fundamental assumptions. The period of the equivalent linear system is assumed to the secant period
and the equivalent damping is related to the area under the capacity curve associated with the inelastic
displacement demand. ATC 40 also limits damping for systems that exhibit strength and stiffness
degrading behavior. The average errors associated with the procedures are illustrated in Figure 9. For
non-degrading structures the current method underestimates displacements, but generally over estimates
for structures with degrading behavior.


Figure 9: Errors associated with current ATC 40 nonlinear static procedure

The focus of the ATC 55 effort to improve equivalent linearization has been to develop better procedures
to estimate equivalent period and equivalent damping. This is an extension of previous work [6,7] in
which both parameters are expressed as functions of ductility. These relationships are based on an
optimization process whereby the error between the displacement predicted using the equivalent linear
oscillator and using nonlinear response history analysis is minimized [8,9]. Conventionally, the
measurement of error has been the mean of the absolute difference between the displacements. Although
this seems logical, it might not lead to particularly good results from an engineering standpoint. This is
SITE CLASS C
0. 0
0. 5
1. 0
1. 5
2. 0
2. 5
3. 0
0. 0 0.5 1. 0 1. 5 2. 0 2. 5 3.0
PERIOD [s]
E[( i )app/ ( i )e x]
R = 8. 0
R = 6. 0
R = 4. 0
R = 3. 0
R = 2. 0
R = 1. 5
APPROXIMATE: ATC40 - TYPE C
EXACT: STRENGTH AND STIFFNESS DEGRADING
Type C structures (severe degradation)
SITE CLASS C
0. 0
0. 5
1. 0
1. 5
2. 0
2. 5
3. 0
0. 0 0.5 1. 0 1. 5 2. 0 2. 5 3.0
PERIOD [s]
E[( i )app/ ( i )e x]
R = 8. 0
R = 6. 0
R = 4. 0
R = 3. 0
R = 2. 0
R = 1. 5
APPROXIMATE: ATC40 - TYPE C
EXACT: STRENGTH AND STIFFNESS DEGRADING
Type C structures (severe degradation)
SITE CLASS C
0. 0
0. 5
1. 0
1. 5
2. 0
2. 5
3. 0
0. 0 0.5 1. 0 1. 5 2. 0 2. 5 3.0
PERIOD [s]
E[( i )a pp/ ( i )ex]
R = 8. 0
R = 6. 0
R = 4. 0
R = 3. 0
R = 2. 0
R = 1. 5
APPROXIMATE: ATC40 - TYPE A
EXACT: ELASTO PLASTIC
Type A structures (non-degrading)
9
illustrated in Figure 10. It is possible to select linear parameters for which the mean error is zero as for the
broad, flat distribution. However, the narrower curve might represent equivalent linear parameters that
provide better results from an engineering standpoint, since the chance of errors outside a 20% to +10%
range, for example, are much lower. This is owing to the smaller standard deviation in spite of the 5%
mean error.


Figure 10: Illustration of probability density function of displacement error for a Gaussian
distribution

This general strategy has been applied to a series of elasto-plastic, stiffness degrading, and strength-and-
stiffness-degrading hysteretic models generate optimal equivalent linear effective periods and damping for
a range on periods and ductilities as illustrated in Figure 11. Also, shown in Figure 11 are the current
CSM specifications in ATC 40 [2].


Figure 11: New optimal effective (equivalent) linear parameters for elastoplastic system.T
0
=0.1-
2.0

Using the results for discrete values of ductility, a curve fitting process has leads to empirical expressions
relating effective period, T
eff
, and effective damping,
eff
, to ductility, . Generally these expressions are
10
dependent on hysteretic type. However, reasonably good results can be obtained for all types of behavior
with the following simplified expressions:

For 4.0 < : ( ) ( )
2 3
4.85 1 1.08 1 5
eff
= + (Eqn. 7)
( ) ( )
2 3
/ 1 0.167 1 0.0310 1
eff o
T T = (Eqn. 8)
For 4.0 6.5 : ( ) 13.6 0.318 1 5
eff
= + + (Eqn. 9)
( ) / 1 0.283 0.129 1
eff o
T T = + (Eqn. 10)
For 6.5 > :
2
2
0
0.64( 1) 1
19.01 5
0.64( 1)
eff
eff
T
T

| `
= +

\ .
(Eqn. 11)

( )
( )
0.5
1
/ 1 0.89 1
1 0.5 1 1
eff o
T T

| !

=


+
\ .

(Eqn. 12)

The solution to the two equations for effective period and damping are illustrated in acceleration-
displacement response spectrum (ADRS) format in Figure 12. Note that the maximum acceleration does
not fall on the ADRS demand curve for the optimal effective damping. As a part of the proposed
improvements a numerical transformation will be included to generate a modified ADRS (MADRS) that
will represent correct values on the acceleration axis. This transformation facilitates the development of
several solution procedures that are very similar to those for the capacity spectrum method in ATC 40. It
should be noted that the recommendations will include a limit on minimum strength as discussed in the
previous section.


Figure 12: Description of the Modified ADRS (MADRS) and its use (from Iwan 2002).

11
MULTI-DEGREE-OF-FREEDOM EFFECTS

In order to compare and illustrate techniques for improving the results of nonlinear static procedures
related to the effects of higher modes, five example buildings have been analyzed. The objective has been
to compare estimates made using simplified inelastic procedures with results obtained by nonlinear
response history analysis. The basic outline of this effort was as follows:

EXAMPLE BUILDINGS
3-Story Steel Frame (SAC LA Pre-Northridge
M1 Model)
3-Story Weak Story Frame (lowest story at 50%
of strength)
8-Story Shear Wall (Escondido Village)
9-Story Steel Frame (SAC LA Pre-Northridge
M1 Model)
9-Story Weak Story Frame (lowest story at 50%
of strength)
GROUND MOTIONS
11 Site Class C Motions, 8-20 km, 5 events
4 Near Field Motions:
GLOBAL DRIFT LEVELS
Ordinary Motions (scaled to result in specified
global drift)
0.5, 2, 4% for frames
0.2, 1, 2% for wall
Near-Field (unscaled)
1.8 to 5.0% for 3-story frames
1.7-2.1% for 9-story frames
0.6 2.1% for wall

LOAD VECTORS/METHODS ILLUSTATED
First Mode
Inverted Triangular
Rectangular (Uniform)
Code
Adaptive
SRSS
Multimode Pushover (MPA)
RESPONSE QUANTITIES
(Peak values generally occur at different
instants in time)
Floor and roof displacements
Interstory Drifts
Story Shears
Overturning Moment
ERRORS
Mean over all floors
Maximum over all floors




The results of the illustrative examples are consistent with previously published observations by
researchers. It is apparent that the approximate procedures can generally predict maximum displacements
reasonably well. Multi-mode pushover can also provide good estimates of maximum inter-story drifts for
some cases. But beyond that, the nonlinear static procedures cannot provide reliable estimates of MDOF
effects. FEMA 440 will recommend nonlinear response history analysis to determine these effects.

As a part of the MDOF an interesting and potentially promising observation for future development has
been made [10]. To generate the example results, ground motions were scaled to give pre-determined
constant roof displacements for each case. This effectively normalized the MDOF effects to the roof
displacement. In general it was noted that any single response history analysis provided a better estimate
of MDOF effects than any of the approximate methods (see Figure 13). This suggests that seismic hazard
could be characterized by the maximum inelastic displacement at the roof level. This could be determined
for a structure with nonlinear static procedures using the NEHRP [5] maps, for example. When necessary,
nonlinear response history analysis could be used to investigate MDOF effects by using a small number of
records scaled to give the same roof displacement. This procedure could avoid the both the necessity of
generating a series of spectrum-compatible records and the difficulty of combining results of the analyses
for practical use.

12
Weak2 % Weak4 %
0 50000 100000 150000 200000
1st
2nd
3rd
4th
5th
6th
7th
8th
9th
Floor
Overturning Moment (kips-ft)
2% Drift
0 50000 100000 150000 200000
1st
2nd
3rd
4th
5th
6th
7th
8th
9th
Floor
Overturning Moment (kips-ft)
2% Drift
0 50000 100000 150000 200000
1st
2nd
3rd
4th
5th
6th
7th
8th
9th
Floor
Overturning Moment (kips-ft)
4% Drift
0 50000 100000 150000 200000
1st
2nd
3rd
4th
5th
6th
7th
8th
9th
Floor
Overturning Moment (kips-ft)
4% Drift
Median
Code SRSS
Adaptive
Multimode
Rectangular
Inverted Triangular
First Mode
Min Max
Mean
SD SD
Overturning MomentsWeak-story 9-story frame


Figure 13:Selected results from the MDOF illustrative examples (from Aschheim 2002)


SOIL-STRUCTURE INTERACTION EFFECTS

FEMA 356 [1] currently contains limitations (caps) on the maximum value of the coefficient C
1
, the ratio
of the maximum inelastic displacement of a single degree of freedom elasto-plastic oscillator to the
maximum response of the fully elastic oscillator. FEMA 356 includes the capping limitations for two
related reasons. First, there is a belief in the practicing engineering community that short stiff buildings
simply do not respond to seismic shaking as adversely as might be predicted analytically. Secondly, it was
felt that the required use of the empirical equation without out relief in the short period range would
motivate practitioners to revert to the more traditional, and apparently less conservative, linear procedures.
The current limitations are not founded directly on theoretical principles or empirical data. Much of the
reduction in response of short period structures is due to soil-structure interaction effects. In lieu of the
capping of C
1
, FEMA 440 will introduce adjustments to seismic demand intended to address soil-structure
interaction effects inelastic analyses [11]. These are fundamentally similar to those in the NEHRP [5]
intended for linear analyses.

Short, stiff buildings generally are more sensitive to interaction between soil material strength and
stiffness with that of the structure and its foundations than are longer period structures. This is patially
accounted for by modeling the stiffness and strength of foundation and supporting soils in the structural
analysis as outlined in FEMA 356 and ATC 40. FEMA 356 will include procedures to reduce spectral
ordinates for the kinematic effects of base slab averaging and embedment of the structure (see Figure 14).
Additionally, the document will include procedures to modify the damping of the overall systems to
account for the inertia effects of foundation damping (see Figure 15). These improvements will rationally
result in reductions in estimated response for short period structures.

13

Figure 14:Reduction in free-field motion due to kinematic effects of base slab averaging and
embedment. (from Stewart 2003)





Figure 15:Foundation damping



REFERENCES

1. BSSC, A Prestandard And Commentary For The Seismic Rehabilitation Of Buildings, prepared
by the Building Seismic Safety Council; published by the Federal Emergency Management
Agency, FEMA 356 Report, 2000, Washington, DC.
2. ATC, The Seismic Evaluation and Retrofit of Concrete Buildings, Volume 1 and 2, ATC-40
Report, 1996, Applied Technology Council, Redwood City, California.
3. Miranda, E. and Akkar S., Evaluation of approximate methods to estimate target displacements
in nonlinear static procedures, PEER-2002/21, Proc. Fourth U.S.-Japan Workshop on
Performance-Based Earthquake Engineering Methodology for Reinforced Concrete Building
1 1.5 2
Period Lengthening, T
eq
/T
eq
0
10
20
30
F
o
u
n
d
a
t
i
o
n

D
a
m
p
i
n
g
,

f

(
%
)
e/r
u
= 0
PGA > 0.2g
PGA < 0.1g
h/r

= 0.5
1.0
2.0

0 0.2 0.4 0.6 0.8 1 1.2

Period (s)
0.4
0.5
0.6
0.7
0.8
0.9
1
F
o
u
n
d
a
t
i
o
n
/
f
r
e
e
-
f
i
e
l
d

R
R
S

f
r
o
m

b
a
s
e

s
l
a
b

a
v
e
r
a
g
i
n
g

(
R
R
S
b
s
a
)
Simplified Model
b
e
= 65 ft
b
e
= 130 ft
b
e
= 200 ft
b
e
= 330 ft
0 0.4 0.8 1.2 1.6 2
Period (s)
0
0.2
0.4
0.6
0.8
1
1.2
F
o
u
n
d
a
t
i
o
n
/
f
r
e
e
-
f
i
e
l
d

R
R
S

f
r
o
m

e
m
b
e
d
m
e
n
t

e
f
f
e
c
t
s

(
R
R
S
e
)
Site Classes C and D
e = 10 ft
e = 20 ft
e = 30 ft
C
D
14
Structures, 22-24 October 2002, Toba, Japan, Pacific Earthquake Engineering Research Center,
University of California, Berkeley, Dec. 2002, pages 75-86.
4. Ruiz-Garcia, J. and Miranda, E. Inelastic displacement ratio for evaluation of existing
structures, Earthquake Engineering and Structural Dynamics. 32(8), 1237-1258, 2003.
5. Building Seismic Safety Council, BSSC. NEHRP Recommended Provisions for Seismic
Regulations for New Buildings and Other Structures, Part 1 Provisions and Part 2
Commentary, Federal Emergency Management Agency, Washington D.C., February, 2001.
6. Iwan, W. D. Estimating inelastic response spectra from elastic spectra. Intl. J. Earthq. Engng.
and Struc. Dyn., 1980, 8, 375-388.
7. Iwan, W. D. and Gates, N.C., The effective period and damping of a class of hysteretic
structures, Intl. J. Earthq. Engng and Struc. Dyn., 1978, 7, 199-211.
8. Iwan, W. D., and Guyader, A.C., "An improved capacity spectrum method employing statistically
optimized linear parameters," Paper No. 3020, 13th World Conference on Earthquake
Engineering, Vancouver, B.C., Canada, August 1-6, 2004
9. Guyader, A.C., and, Iwan, W.D., "A Statistical Approach to Equivalent Linearization with
Application to Performance-Based Engineering, California Institute of Technology, EERL
Report No. 2004-04, 2004.
10. Aschheim, M., Tjhin, T., Comartin, C., Hamburger, R., and Inel, M., "The scaled nonlinear
dynamic procedure," ASCE Structures Congress, Nashville, TN, May 22-26, 2004.
11. Stewart, J.P., Comartin, C.D., and Moehle, J.P., Implementation of soil-structure interaction
models in performance based design procedures, Paper No. 1546, 13th World Conference on
Earthquake Engineering, Vancouver, B.C., Canada, August 1-6, 2004