Eleventh U.S.

National Conference on Earthquake Engineering

Integrating Science, Engineering & Policy
June 25-29, 2018
Los Angeles, California

USING FEMA P695 TO INTERPRET ASCE

41 SEISMIC PERFORMANCE OF SPECIAL
MOMENT FRAMES
M.S. Speicher 1, J. Dukes2, and K.K.F. Wong3

ABSTRACT

ASCE 41 contains methodologies used by practicing engineers for the assessment of existing
buildings and the design of new buildings. The National Institute of Standards and Technology
recently completed a study investigating the relationship between ASCE 41 and traditional design
standards (e.g., ASCE 7 and AISC 341). The results indicated there are many inconsistencies
between the two approaches, with some unwarranted conservatism in ASCE 41. To further
investigate this relationship, this paper presents the results of a collapse assessment of a set of four
steel special moment frames. The goal is to verify that the design intent is being met (i.e., no greater
than 10 % probability of collapse given a risk-targeted maximum considered earthquake). The
effects of modeling assumptions, such as using default ASCE 41 backbone curves versus
experimentally-derived backbone curves, is discussed. The results are used to scrutinize the
performance indicated by the ASCE 41 assessment. In general, the performance indicated by an
ASCE 41 assessment is shown to be conservative relative to the collapse performance indicated
by a FEMA P695 assessment.

1
Research Structural Engineer, National Institute of Standards and Technology (email: matthew.speicher@nist.gov)
2
Research Structural Engineer, National Institute of Standards and Technology (email: jazalyn.dukes@nist.gov)
3
Research Structural Engineer, National Institute of Standards and Technology (email: kevin.wong@nist.gov)

Speicher MS, Dukes J, Wong KKF. Using FEMA P695 to Interpret ASCE 41 Seismic Performance of Special
Moment Frames. Proceedings of the 11th National Conference in Earthquake Engineering, Earthquake Engineering
Research Institute, Los Angeles, CA. 2018.
 Eleventh U.S. National Conference on Earthquake Engineering
Integrating Science, Engineering & Policy
June 25-29, 2018
Los Angeles, California

Using FEMA P695 to Interpret ASCE 41 Seismic Performance of

Special Moment Frames

M.S. Speicher 1, J. Dukes2, and K.K.F. Wong3

ABSTRACT

ASCE 41 contains methodologies used by practicing engineers for the assessment of existing
buildings and the design of new buildings. The National Institute of Standards and Technology
recently completed a study investigating the relationship between ASCE 41 and traditional design
standards (e.g., ASCE 7 and AISC 341). The results indicated there are many inconsistencies
between the two approaches, with some unwarranted conservatism in ASCE 41. To further
investigate this relationship, this paper presents the results of a collapse assessment of a set of four
steel special moment frames. The goal is to verify that the design intent is being met (i.e., no greater
than 10 % probability of collapse given a risk-targeted maximum considered earthquake). The
effects of modeling assumptions, such as using default ASCE 41 backbone curves versus
experimentally-derived backbone curves, is discussed. The results are used to scrutinize the
performance indicated by the ASCE 41 assessment. In general, the performance indicated by an
ASCE 41 assessment is shown to be conservative relative to the collapse performance indicated by
a FEMA P695 assessment.

Introduction

Performance-based seismic design is intended to let the structural engineer develop earthquake-
resistant building design solutions that are more efficient and cost-effective than those obtained
using the prescriptive building code requirements found in ASCE 7 [1]. ASCE 41 [2] contains
performance-based methodologies used by some practicing engineers for both the assessment of
existing buildings and the design of new buildings. The National Institute of Standards and
Technology (NIST) recently completed a study [3-5] investigating the relationship between ASCE
41 and traditional design standards (e.g., ASCE 7 and AISC 341 [6]) for a variety of framing
systems, including steel special moment frames. In general, the study showed inconsistencies
between the two approaches, with some potentially unwarranted conservatism in ASCE 41. It was
recognized that to more fully contextualize the results and to see if the intended performance

1
Research Structural Engineer, National Institute of Standards and Technology (email: matthew.speicher@nist.gov)
2
Research Structural Engineer, National Institute of Standards and Technology (email: jazalyn.dukes@nist.gov)
3
Research Structural Engineer, National Institute of Standards and Technology (email: kevin.wong@nist.gov)

Speicher MS, Dukes J, Wong KKF. Using FEMA P695 to Interpret ASCE 41 Seismic Performance of Special
Moment Frames. Proceedings of the 11th National Conference in Earthquake Engineering, Earthquake Engineering
Research Institute, Los Angeles, CA. 2018.
objectives of ASCE 7 is being met, the collapse likelihood of each building subjected to strong
earthquake ground motions needs to be determined.

This paper presents the results of the collapse assessment of four special moment frames originally
investigated in the NIST study. First the methodology is discussed in terms of setting up the models
and determination of collapse fragilities. The buildings are modeled in OpenSees [7] using
commonly accepted approaches. FEMA P695 [8] is used as a guide in conducting analysis to
determine the probability of collapse in a “standardized” manner. The results indicate that these
ASCE 7-designed frames meet the intent of the code. This helps contextualize the observations
made in Harris and Speicher [3], which showed that, in many cases, ASCE 41 is overly
conservative.

Methodology

Four steel special moment framed buildings are assessed in this paper. The modeling details are
given in Fig. 1 and the member sizes are given in Table 1. Harris and Speicher [3] conducted an
ASCE 41 performance assessment using 3D models in Perform-3D. Given the cost of running
incremental dynamic analysis, 2D models were created in OpenSees. It was found that the 2D and
3D models matched well, which is not surprising given the buildings are symmetric and the lateral
systems in each direction are non-intersecting.
B4 B4 B4
Legend
(c)
4.27 m

RBS
E4

C4

C4

E4
Panel Zone
dc/2
Plastic Hinge a b B4 B4 B4
of Beam
4.27 m

E4

C4

C4

E4
Plastic Hinge
of Column db
B3 B3 B3
Pinned
Connection
4.27 m

c
E3

C3

C3

E3
Leaning
Column 2b/3 B3 B3 B3
4.27 m

E3

C3

C3

E3

dc
B4 B4 B2 B2 B2
B4
4.27 m

4.27 m
E2

C2

C2

E2

E2

C2

C2

E2

B3 B3 B2 B2 B2
B3
4.27 m

4.27 m
E2

C2

C2

E2

E2

C2

C2

E2

B2 B2 B2 B1 B1 B1
C1
4.27 m

4.27 m
E1

E1

E1

C1

C1

E1
C1

subassembly
B1 B1 B1 B1 B1 B1
5.49 m

5.49 m
E1

C1

E1

E1

C1

C1

E1
C1

(a) (b)

9.14 m 9.14 m 9.14 m 9.14 m 9.14 m 9.14 m 9.14 m 9.14 m

Figure 1. OpenSees modeling details for the 4- and 8-story special moment frames.
The moment frames have reduced beam section (RBS) connections. The RBS stiffness was
modeled using a prismatic cross-section over the length of the RBS (denoted as b). The width of
the RBS was assumed to be equivalent to the actual RBS width at b/3 away from the center of the
RBS. The nonlinear behavior was captured by a nonlinear rotational spring placed at the RBS
center. The nonlinear spring was assigned a stiffness of 10 times that of the unreduced beam. Since
this spring is in series with the elastic beam elements, the elastic beam stiffness was increased to
give an overall beam (including the RBS and nonlinear spring) stiffness equal to that without the
nonlinear spring (see [9] for more discussion on this approach).

Table 1. Member sizes for 4-story and 8-story frames.

Beam ELF RSA
Size a (mm) b (mm) c (mm) Size a (mm) b (mm) c (mm)
4-story
B1 W24×84 121 406 57 W24×55 95 406 44
B2 W24×84 121 406 57 W24×55 95 406 44
B3 W24×76 114 406 57 W24×55 95 406 44
B4 W24×55 95 406 44 W24×55 95 406 44
8-story
B1 W27×114 133 457 64 W24×84 121 406 57
B2 W27×114 133 457 64 W24×76 114 406 57
B3 W27×94 127 457 64 W24×55 95 406 44
B4 W24×55 95 406 44 W21×44 83 356 38

Column ELF RSA Column ELF RSA

4-story
E1 W14×159 W14×132 C1 W14×257 W14×193
E2 W14×145 W14×132 C2 W14×233 W14×193
8-story
E1 W18×175 W18×175 C1 W18×258 W18×192
E2 W18×158 W18×106 C2 W18×258 W18×192
E3 W18×130 W18×71 C3 W18×234 W18×143
E4 W18×71 W18×55 C4 W18×143 W18×119
Note: ELF = equivalent lateral force, RSA = response spectrum analysis. ELF and RSA were created for each building height.

OpenSees can capture cyclic and in-cycle degradation using the Ibarra Modified Krawinkler
(IMK) model with the Bilin material. The force-deformation parameters for the RBS followed the
recommendations made by Lignos [10], which were derived using multivariate regression analysis
of a suite of experimental results. These parameters include the plastic rotation capacity, θp, the
post-capping rotation capacity, θpc, the yield strength, Μy, the capping strength, Mc, the ultimate
rotation, θu, the residual strength ratio, κ, and the reference cumulative plastic rotation parameter,
Λ.

Nonlinear springs were added to the columns at d/2 away from the face of the column (see Fig. 1),
where d is the depth of the column. These nonlinear springs followed the same approach (10 times
stiffness) as that discussed for the beams. The force-deformation parameters for the column hinges
follow the recommendations produced by ATC project 114 [11] using the monotonic backbone
(since our model captures degradation). The panel zones were modeled explicitly using the
approach outlined by Krawinkler [12]. A series of “rigid” elements with pinned connections was
placed at the panel zone region with one corner being tied together with nonlinear rotational spring.
The spring parameters were based on fundamental mechanics. Additionally, though the column
splice was designed at 1.2 m above the beam-to-column joint, this was ignored in the model.

To illustrated the difference in force-deformation behavior between the OpenSees model (used in
this study) and the Perform-3D model (used in [3]), a subassembly model was created in both
platforms as shown in Fig. 2. We assumed doubly-symmetric bending for both the beams and
columns, thus zero moment at the mid-spans. As shown in Fig. 3, assuming the columns remain
elastic, the (a) total drift can be broken down into the following actions: (b) elastic deformation of
the beams and columns, (c) inelastic behavior of the beam hinges, and (b) joint panel zone
deformation. The results of a displacement-controlled reversed cyclic simulation are shown in Fig.
4. The member properties and RBS hinge properties are those of the 4-story ELF design and are
from the locations as indicated in Fig. 1.

Perform-3D Model OpenSees Model

F F
W14x257

W14x257
2.13 m

2.13 m
Column Plastic Hinge a Column Plastic Hinge a
Beam Plastic Hinge Beam Plastic Hinge
b b
W24x84 dc/2 W24x84 W24x84 W24x84
W14x257

W14x257
RBS RBS
2.74 m

2.74 m

2.5x End Zone Reduced Beam Panel Zone Reduced Beam

Stiffness Moment of Inertia Moment of Inertia
a = 121 mm a = 121 mm
b = 406 mm b = 406 mm
c = 57 mm c = 57 mm

(a) (b)
4.57 m 4.57 m 4.57 m 4.57 m
Figure 2. Subassembly models used to compare (a) Perform-3D and (b) OpenSees.

The force-displacement behavior reasonably matches for drift levels up to approximately 5 %.

Beyond this level, the Perform-3D model loses all strength and the response travels along the zero-
force level. The Perform-3D model’s force-deformation curve was created using the default a, b
and c parameters given in ASCE 41-06 Table 5-6 [13]. The endzone stiffness parameter used in
Perform-3D was set to 2.5, which accounts for a portion the stiffness differences between the two
models. Perform-3D also includes elastic shear deformations and the OpenSees model does not,
which contributes to the stiffness difference.

Note that the Perform-3D force-displacement response at the top of the subassembly has a sudden
post-capping strength decrease due to the RBS hinge (even though the RBS inelastic spring has a
more gradual decrease). This can be explained by examining the drift contributions as conceptually
illustrated in Fig 5. At step k, the force is shown reaching the capping (peak) level. Once the
strength starts to drop on the subassembly, say at step k+1, the remaining elastic elements have a
reduced contribution to the drift thus the effective drop is sudden. Note, we assume displacement
∆k+1 is only slightly larger than ∆k, though Fig. 5 shows this exaggerated. Figure 4 shows the
actual behavior for the Perform-3D model with the sudden drop.

In contrast, the OpenSees subassembly model has a gradual post-capping strength deterioration
behavior. Lignos and Krawinkler [14] reported that the slope of this degradation is one of the most
important parameters that influences the nonlinear response. The slope is established by θpc. This
descending branch continues until the force-deformation reaches the residual strength or the
ultimate rotation. For this paper, θu was taken as 0.20 radians for all hinges. Lignos and Krawinkler
[10] reported θu values of approximately 0.07 radians for fully-reversed cyclic tests, but monotonic
tests may be up to 3 times larger. The residual strength factor, κ, was taken as 0.4, which is double
the value given in ASCE 41 and used in Harris and Speicher [3]. Note, the values of θu and κ
have been reported to be non-critical in studies where collapse capacities are being determined
[10].

(a) ∆ (b) ∆e (c) ∆i (d) ∆p

F F F F

= + +

Elastic Inelastic Panel Zone

Total Displacement Displacement Displacement Displacement
Figure 3. Drift contributions for a moment frame subassembly.
600

400
Base Shear (kN)

200

-200

-400 OpenSees
Perform-3D
-600
-0.1 -0.05 0 0.05 0.1
Drift Ratio
Figure 4. Comparison of the response of subassembly to cyclic loading for OpenSees and
Perform-3D (model used in Harris and Speicher [3]).

(a) F F OpenSees
F F
and OpenSees
Fk Fk Perform-3D Fk OpenSees Fk
OpenSees
= + + Perform-3D
Perform-3D Perform-3D
∆ ∆e ∆i ∆p
∆k ∆e,k ∆i,k ∆p,k

(b) F F OpenSees
F F
and OpenSees
OpenSees Perform-3D OpenSees
Fk+1 Fk+1 Fk+1 Fk+1
= + + Perform-3D
Perform-3D ∆ ∆e Perform-3D ∆i ∆p
Fk+1 Fk+1 Fk+1 Fk+1
∆k+1 ∆e,k+1 ∆i,k+1 ∆p,k+1

Figure 5. Explanation of drift between (a) some point, k, at the capping moment and (b) some
point, k+1, just beyond the capping moment.
 Fragility Curve Development

To estimate each building’s collapse fragility, incremental dynamic analysis (IDA) was carried
out. IDA involves analyzing a structural model to multiple ground motion records that are scaled
over a range of levels to produce a response curve [15]. The suite of ground motions used for the
IDA is referred to as the “far-field” set in FEMA P695. This set consists of 22 record pairs from
sites that were more than 10 km from the fault rupture. Given our 2D model, we applied each pair
independently, resulting in 44 earthquake runs at each intensity level. The ground motions were
first normalized per FEMA P695 using the FEMA 695 Toolkit [16] and then scaled up to the
risked-targeted maximum considered earthquake (MCER).

The analyses were performed using the parallel version of OpenSees, OpenSeesMP, on the
Extreme Science and Engineering Discovery Environment (XSEDE) platform [17]. The initial
number of dynamic analyses for each building type was 44 earthquakes times 20 scale factors (0.1
to 2.0). If the structure did not reach the collapse criteria for any earthquake by the last scale factor,
additional analyses were performed at higher scale factors until collapse was achieved.

The ground motion intensity measure (IM) used in this paper is the median spectral acceleration
of the suite of ground motions at the fundamental period (CuTa) of the building, ST. Collapse was
defined as reaching 12.5 % interstory drift in any story. The IM at which the structure first reaches
12.5 % interstory drift is used in the calculation of the fragility curve. To develop the fragility
curve, the IMs at which the structure collapses for each ground motion is aggregated across the
ground motion suite and then is used to determine the fragility curve. The fragility function was
calculated using “Method A” as described by Porter et al. [18], where the formulation is based on
scaling all the ground motions until all specimens have collapsed. Equation 1 shows the calculation
of the fragility function, where P(C | IM = x) is the probability of collapse of the structure at IM x.

 ln( x / θ ) 
P(C | IM = x) = Φ  (1)
 β 

The normal cumulative distribution function is denoted by Φ, θ represents the median of the
fragility function, and β is the standard deviation of ln(IM). Equations 2 and 3 show the formulation
of the fragility function estimators ( θˆ , β̂ ) over n number of earthquakes, which are method of
moments estimators of a normal distribution (Baker, 2015).

1 n
ln θ = ∑ ln IM i
ˆ (2)
n i =1

1 n
βˆ = ∑ (ln(IM i / θˆ)) 2
n − 1 i =1
(3)
 Results

The results are presented in the form of IDA curves and associated fragility curves. Fig. 6 shows
the IDA curves for the 4 and 8-story ELF and RSA models. The IDA curves illustrate the
progression of damage as intensities are increased. After approximately 7.5 % interstory drift, a
small increase in intensity generally results in a large increase in interstory drift. Therefore, the
choice of collapse being defined as 12.5 % interstory drift is deemed sufficient. Figure 7 shows
the fragility curves of the 4- and 8-story ELF and RSA designed buildings. From these curves, the
probability of collapse at the MCER can be estimated. The MCER spectral accelerations at CuTa for
the 4-story and 8-story buildings are 0.87 g and 0.51 g, respectively.

5 5

4 4

3 3
(g)

(g)
T

T
2 2
S

S
1 1

0 0
0.025 0.05 0.075 0.1 0.025 0.05 0.075 0.1

Max Story Drift Ratio Max Story Drift Ratio

(a) (b)
2.5 2.5

2 2

1.5 1.5
(g)

(g)
T

1 1
S

0.5 0.5

0 0
0.025 0.05 0.075 0.1 0.025 0.05 0.075 0.1

Max Story Drift Ratio Max Story Drift Ratio

(c) (d)
Figure 4. Incremental dynamic analysis curves for the (a) 4-story ELF-designed, (b) 4-story
RSA-designed, (c) 8-story ELF-designed, and (d) 8-story RSA-designed frames.

Discussion

To understand the implications of the fragility curves, Table 2 shows a summary of results reported
by Harris and Speicher [3]. For both the nonlinear static and dynamic procedures, at least one RBS
connection fails the assessment in one out of four frames. For the nonlinear dynamic procedure,
the assessment indicates worse performance with at least one RBS connection failing the criteria
in three out of four frames. The column and panel zones also show failures, but the discussion
presented here is focused on the RBS connection because the newest version of ASCE 41 (i.e.,
ASCE 41-17) has updated modeling assessment criteria that may correct the problems seen in the
columns. Recall that the design intent given in ASCE 7 is that a building will have a probability
of collapse given an MCER of less than or equal to 10 %. However, the FEMA P695 methodology
is intended to be for a suite of archetype buildings. If each of these buildings are taken as
individuals, FEMA P695 suggests that a code-complying building may be acceptable if the
probability of collapse given an MCER reaches as high as 20 %. The IDA and the resulting fragility
curves shown in Fig. 7 suggest that the 4-story buildings and the 8-story ELF-designed building
are adequate in terms of collapse probability. In contrast, RSA-designed 8-story frames has a
probability of collapse of approximately 19 %. Though this frame does not meet the 10 % goal of
a code-designed building as determined by the methodology described herein, it does fall below
the 20 % limit mentioned above. Further investigation is being carried out to verify these results
and determine any deficiencies that could be addressed to help reduce the RSA-design collapse
probabilities.

1 1
0.9 0.9
0.8 0.8
0.7 0.7
Probability of collapse

Probability of collapse
0.6 0.6
0.5 0.5
0.4 Data 0.4 Data
0.3 Fragility 0.3 Fragility
0.2 MCE 0.2 MCE
R R
0.1 0.1

0 1 2 3 4 5 0 1 2 3 4 5
S (g) S (g)
T T
(a) (b)
1 1
0.9 0.9
0.8 0.8
0.7 0.7
Probability of collapse

Probability of collapse

0.6 0.6
0.5 0.5
0.4 Data 0.4 Data
0.3 Fragility 0.3 Fragility
0.2 MCE 0.2 MCE
R R
0.1 0.1

0 0.5 1 1.5 2 2.5 0 0.5 1 1.5 2 2.5

S (g) S (g)
T T
(c) (d)
Figure 5. Fitted fragility curves for the (a) 4-story ELF-designed, (b) 4-story RSA-designed,
(c) 8-story ELF-designed, and (d) 8-story RSA-designed frames.

Given these findings, it appears ASCE 41 is overly-conservative at identifying performance issues

in the set of frames, though the trends between frames are similar. One issue that needs further
investigation is the conservative nature of the Perform-3D model which used modeling parameters
from ASCE 41. This model has a backbone curve with a relatively steep descending branch beyond
the capping point, which results in an almost vertical drop in stiffness for a subassembly. As
mentioned before, the nonlinear response has been shown to be sensitive to this slope. ASCE 41
does not give clear guidance on the post-capping slope, and in this case the approach used in [3]
may be conservative and result in unfavorable performance, especially for the nonlinear dynamic
procedures (as opposed to the nonlinear static procedures). Other reasons for conservatism are
explored in a companion study looking at other steel framing systems [19-21], which includes (1)
the method used for selecting and scaling ground motions and (2) ASCE 41 modeling parameters,
derived from fully-reversed cyclic loading protocols, which may not be a true representation of
buildings at MCER or greater level earthquakes [22]. Regardless, the results presented in this paper
provides further context to the results seen in Harris and Speicher [3] as summarized in Table 2.

Table 2. Summary of performance predicted using the nonlinear procedures in ASCE 41 for
the 4- and 8-story buildings (the performance is for Collapse Prevention under Basic
Safety Earthquake-2 hazard level) [3].
Design Nonlinear Static Procedure Nonlinear Dynamic Procedure
Building
Approach RBS connection Column Panel Zone RBS connection Column Panel Zone
ELF Pass Pass Pass Pass Pass Pass
4-Story
RSA × Pass Pass × Pass Pass
ELF Pass × Pass × × Pass
8-Story
RSA Pass × Pass × × ×

In contrast to the Perform-3D model, the OpenSees model uses a more sophisticated and less
conservative approach. The difference in backbone curves can be seen in the subassembly
discussion presented in the Methodology section. Both in-cycle and cyclic degradation was
accounted for therefore a monotonic backbone curve was used. This resulted in more energy being
dissipated in the hysteresis and a more gradual strength loss beyond the capping point. Further
investigation into the sensitivity of the modeling parameters and analysis approaches may be
beneficial to fully vet these results.

Conclusion

A set of buildings designed using ASCE 7 and assessed using ASCE 41 were shown to be deficient
(per ASCE 41) in a previous study by NIST. The results appeared to be conservative for the
nonlinear dynamic procedure in ASCE 41. This paper presents a preliminary study of the collapse
probability of four special moment frames using incremental dynamic analysis. The resulting
fragility curves suggest three out of four frames has a less than or equal to 10 % probability of
collapse given a risk-targeted maximum considered earthquake, therefore meeting the intent of
ASCE 7. The remaining frame has a greater than 10 % probability of collapse, thus not satisfying
the intent of the code (though the 8-story RSA-designed frame does not exceed 20 %, which can
be considered the limit for an individual building). We note the high-level similarities and the
differences in the results for the ASCE 41 assessment and the IDA, and suggest that implementing
the ASCE 41 modeling parameters in a Perform-3D model may result in a conservative assessment
outcome.

Disclaimer and Acknowledgements

Commercial software may have been used in the preparation of information contributing to this
paper. Identification in this paper is not intended to imply recommendation or endorsement by
NIST, nor is it intended to imply that such software is necessarily the best available for the purpose.
No formal investigation of uncertainty or error is included in this study. This work used the
Extreme Science and Engineering Discovery Environment (XSEDE) through allocation TG-
MSS170023 supported by NSF grant number ACI-1548562.

References

1. ASCE, Minimum Design Loads for Buildings and Other Structures, in ASCE/SEI 7-13. 2013, American
Society of Civil Engineers.
2. ASCE, Seismic Evaluation and Retrofit of Existing Buildings, in ASCE/SEI 41-13. 2013, American Society
of Civil Engineers: Reston, VA.
3. Harris, J.L. and M.S. Speicher, Assessment of First Generation Performance-Based Seismic Design
Methods for New Steel Buildings Volume 1: Special Moment Frames, National Institute of Standards and
Technology: Gaithersburg, MD. 2015.
4. Harris, J.L. and M.S. Speicher, Assessment of First Generation Performance-Based Seismic Design
Methods for New Steel Buildings Volume 2: Special Concentrically Braced Frames, National Institute of
Standards and Technology: Gaithersburg, MD. 2015.
5. Harris, J.L. and M.S. Speicher, Assessment of First Generation Performance-Based Seismic Design
Methods for New Steel Buildings Volume 3: Eccentrically Braced Frames, National Institute of Standards
and Technology: Gaithersburg, MD. 2015.
6. AISC, Seismic Provisions for Structural Steel Buildings, in ANSI/AISC 341-10. 2010, American Institute of
Steel Construction: Chicago, IL.
7. McKenna, F. and G.L. Fenves, OpenSees command language manual. Version 2.5.0. 2016, Pacific
Earthquake Engineering Research Center: Berkeley, CA.
8. FEMA, Quantification of Building Seismic Performance Factors, in FEMA P695. 2009, Federal
Emergency Managament Agency: Washington, D.C.
9. Ibarra, L.F., R.A. Medina, and H. Krawinkler, Hysteretic models that incorporate strength and stiffness
deterioration. Earthquake Engineering and Structural Dynamics, 2005. 34(12): p. 1489-1511.
10. Lignos, D.G. and H. Krawinkler, Deterioration modeling of steel components in support of collapse
prediction of steel moment frames under earthquake loading. Journal of Structural Engineering, 2011.
137(11): p. 1291-1302.
11. NIST, Guidelines for Nonlinear Structural Analysis for Design of Buildings, Part IIa – Steel Moment
Frames: Gaithersburg, MD. 2017.
12. Krawinkler, H., Shear in Beam-Column Joints in Seismic Design of Steel Frames. Engineering Journal,
1978. 15(3): p. 82-91.
13. ASCE, Seismic Rehabilitation of Existing Buildings, in ASCE/SEI 41-06. 2006, American Society of Civil
Engineers: Reston, VA.
14. Lignos, D.G., H. Krawinkler, and A.S. Whittaker, Prediction and validation of sidesway collapse of two
scale models of a 4-story steel moment frame. Earthquake Engineering and Structural Dynamics, 2011.
40(7): p. 807-825.
15. Vamvatsikos, D. and C. Allin Cornell, Incremental Dynamic Analysis. Earthquake Engineering and
Structural Dynamics, 2002. 31(3): p. 491-514.
16. Hardyniec, A. and F. Charney, A Toolkit for Performing FEMA P-695 Evaluations. Earthquake Spectra,
2016. 32(2): p. 653-676.
17. Towns, J., et al., XSEDE: Accelerating Scientific Discovery. Computing in Science & Engineering,
2014. 16(5): p. 62-74.
18. Porter, K., R. Kennedy, and R. Bachman, Creating fragility functions for performance-based earthquake
engineering. Earthquake Spectra, 2007. 23(2): p. 471-89.
19. Speicher, M.S. and J.L. Harris, Collapse Prevention Seismic Performance Assessment of New Eccentrically
Braced Frames using ASCE 41. Engineering Structures, 2016. 117: p. 344-357.
20. Speicher, M.S. and J.L. Harris, Collapse prevention seismic performance assessment of new special
concentrically braced frames using ASCE 41. Engineering Structures, 2016. 126: p. 652-666.
21. Speicher, M.S. and J.L. Harris, Collapse Prevention seismic performance assessment of new buckling-
restrained braced frames using ASCE 41. Engineering Structures, 2018. 164: p. 274-289.
22. Maison, B.F. and M.S. Speicher, Loading Protocols for ASCE 41 Backbone Curves. Earthquake Spectra,
2016. 32(4): p. 1-20.