# A MODAL PUSHOVER ANALYSIS PROCEDURE TO ESTIMATE SEISMIC DEMANDS FOR BUILDINGS: THEORY AND PRELIMINARY EVALUATION ANIL K.

CHOPRA • RAKESH K. GOEL PEER 2001/03 JAN. 2001

Pacific Earthquake Engineering Research Center

**A Modal Pushover Analysis Procedure to Estimate Seismic Demands for Buildings:
**

Theory and Preliminary Evaluation

Anil K. Chopra University of California Berkeley Rakesh K. Goel California Polytechnic State University San Luis Obispo

A report on research conducted under grant no. CMS-9812531 from the National Science Foundation: U.S.-Japan Cooperative Research in Urban Earthquake Disaster Mitigation

PEER 2001/03

A Modal Pushover Analysis Procedure to Estimate Seismic Demands for Buildings: Theory and Preliminary Evaluation

Anil K. Chopra University of California Berkeley

Rakesh K. Goel California Polytechnic State University San Luis Obispo

A report on research conducted under Grant No. CMS-9812531 from the National Science Foundation: U.S. Japan Cooperative Research in Urban Earthquake Disaster Mitigation

PEER Report 2001/03 Pacific Earthquake Engineering Research Center College of Engineering University of California Berkeley January 2001 i

ii
.

and the structure is pushed to the roof displacement determined from the peak deformation Dn of the nth-mode elastic SDF system.n t = . at which the seismic response. p eff . is determined by pushover analysis.mι ug t .sn ug t . The MPA procedure is extended to estimate the seismic demands for inelastic systems: First. the total demand. show that MPA estimates the response of buildings responding well into the inelastic range to a similar degree of accuracy when compared to standard RSA for estimating peak response of elastic systems. which retains the conceptual simplicity and computational attractiveness of current procedures with invariant force distribution. p eff t = . This pushover curve is idealized as bilinear n and converted to the force-deformation relation for the nth-“mode” inelastic SDF system. Thus the present trend of comparing computed hinge plastic rotations against rotation limits established in FEMA-273 to judge structural performance does not seem prudent.ABSTRACT The principal objective of this investigation is to develop a pushover analysis procedure based on structural dynamics theory. MPA.urn curve is developed from a pushover analysis for force distribution s* . but provides superior accuracy in estimating seismic demands on buildings.… according to an appropriate modal combination rule. rno .
bg
bg
bg
bg
b
g
b
g
Comparing the peak inelastic response of a 9-story SAC building determined by the approximate MPA procedure with rigorous nonlinear response history analysis (RHA) demonstrates that the approximate procedure provides good estimates of floor displacements and story drifts.25 to 3. Combining these peak modal responses by modal combination rule leads to the MPA procedure. The base shear-roof displacement Vbn . 2. The standard response spectrum analysis (RSA) for elastic buildings is reformulated as a Modal Pushover Analysis (MPA). all pushover analysis procedures considered do not seem to compute to acceptable accuracy local response quantities. in the modal expansion of the effective earthquake forces. The peak response of the elastic structure due to its nth vibration mode can be exactly determined by pushover analysis of the structure subjected to lateral forces distributed over the height of the building according to s* = mφ n . a pushover analysis determines the peak response rno of the inelastic MDF system to individual modal terms. Thus the MPA procedure is accurate enough for practical application in building evaluation and design. plastic hinge rotations are less accurate. and nonlinear RHA. Comparing the earthquake-induced demands for the selected 9-story building determined by pushover analysis using three force distributions in FEMA-273. The peak deformation of this SDF system is used to determine the roof displacement. and identifies locations of most plastic hinges. ro . iii
. is determined by combining the rno n = 1. where m is the n mass matrix and φ n its nth-mode.0. However. it is demonstrated that the FEMA force distributions greatly underestimate the story drift demands. such as hinge plastic rotations. and the MPA procedure is more accurate than all the FEMA force distributions methods in estimating seismic demands. Instead. The results presented for El Centro ground motion scaled by factors varying from 0. Second. structural performance evaluation should be based on story drifts known to be closely related to damage and can be estimated to a higher degree of accuracy by pushover analyses.

iv
.

v
.ACKNOWLEDGMENT This research investigation is funded by the National Science Foundation under Grant CMS9812531. Japan Cooperative Research in Urban Earthquake Disaster Mitigation.S. a part of the U. This financial support is gratefully acknowledged.

vi
.

..........................................4.............................................13 3...........................................4
5............................................CONTENTS Abstract ......................1 Summary...............................................................................................................................................47
4......................................1 5.........................................................................................................6 Pushover Analysis ......... Equation of Motion ........6
Elastic Multistory Buildings .................................................................................................
Inelastic Multistory Buildings .............................3 Modal Pushover Analysis with Gravity Loads .....................4..................................4 3........................................1 One-Story Systems ...............................................................2 Response History Analysis .............36 4..................................... 2..................4....1 4............................................3 3....... iii Acknowledgment .................................................................................37 Comparative Evaluation of Analysis Procedures .................4.........................................................27 Uncoupled Modal Response History Analysis .................................................................2 Modal Pushover Analysis ............................................................................................................................................................................................................3 System and Excitation Considered .1 3.......................................................................3 4..........33 4...3 2..................2 Response History Analysis ........2.....................................................5 Response History Analysis ........................................................2 2................................................2 3............................................................4 Modal Response History Analysis...................55 Comparative Evaluation...............38 4.............55 5...............................................................41 4..........................2................................................3 Modal Pushover Analysis ...................4....3 2....27 4......................................................2............3.................................................55 vii
.........1 System and Excitation Considered ..............................................................2 FEMA-273 Pushover Analysis .............................................................................................1 Uncoupled Modal Response History Analysis .....4....12 Modal Pushover Analysis ...........................................36 Modal Pushover Analysis .17 3............................................................
Comparison of Modal and FEMA Pushover Analyses........34 4.........9 3....................................................................1 Underlying Assumptions and Accuracy .............................................................................................................................38 4. vii 1....1 2..............................................................................12 Comparative Evaluation of Analysis Procedures ........13 3..................................................28 4..........3 Summary...................................................................9 Modal Response Spectrum Analysis .....2 Properties of nth-“mode” Inelastic SDF System ........22
4............................................................ Introduction.....................v Table of Contents................

..............................................................................................................................................69
Appendices A: Uncoupled Modal Response History Analysis .....................71 B: Modal Pushover Analysis .65 References..............................
Conclusions................................................................................ 7...85
viii
.................................6..........81 C: FEMA Force Distribution Calculation....................................................................................................................................................................................................

Skokan and Hart. 1999. 1997]. 1988. 1998. several researchers have proposed adaptive force distributions that attempt to follow more closely the time-variant distributions of inertia forces [Fajfar and Fischinger. While these adaptive force distributions may provide better estimates of seismic demands [Gupta and Kunnath. None of the invariant force distributions can account for the contributions of higher modes to response. or for a redistribution of inertia forces because of structural yielding and the associated changes in the vibration properties of the structure. Gupta and Kunnath. Gupta and Krawinkler. 1998. While nonlinear response history analysis (RHA) is the most rigorous procedure to compute seismic demands. 1994.1
Introduction
Estimating seismic demands at low performance levels. 1997. current structural engineering practice uses the nonlinear static procedure (NSP) or pushover analysis in FEMA-273 [Building Seismic Safety Council. 1999. 1999]. 1988. Kim and D’Amore. 2000. 2000] have led to good estimates of seismic demands. 1999.. Both the force distribution and target displacement are based on the assumption that the response is controlled by the fundamental mode and that the mode shape remains unchanged after the structure yields. The seismic demands are computed by nonlinear static analysis of the structure subjected to monotonically increasing lateral forces with an invariant height-wise distribution until a predetermined target displacement is reached. Obviously. Miranda. 2000]. such satisfactory predictions of seismic demands are mostly restricted to low.and medium-rise structures in which inelastic action is distributed throughout the height of the structure [Krawinkler and Seneviratna. after the structure yields both assumptions are approximate. To overcome these limitations. Gupta and Krawinkler.. requires explicit consideration of inelastic behavior of the structure. Attempts 1
. 2000]. 1981. Bracci et al. Krawinkler and Seneviratna. 1991. such as life safety and collapse prevention. Maison and Bonowitz. Lawson et al. but investigations [Saiidi and Sozen. they are conceptually complicated and computationally demanding for routine application in structural engineering practice. However. Fajfar and Fischinger.

First. 2000]. The MPA procedure is then extended to inelastic buildings. Finally.. Next we develop a modal pushover analysis (MPA) procedure for linearly elastic buildings and demonstrate that it is equivalent to the well-known response spectrum analysis (RSA) procedure.
2
. but provides superior accuracy in estimating seismic demands on buildings. 1996.. the seismic demands determined by pushover analysis using three force distributions in FEMA-273 are compared against the MPA and nonlinear RHA procedures. Gupta and Kunnath. 1998. and the errors in the procedure relative to a rigorous nonlinear RHA are documented. Kunnath and Gupta.have also been made to consider more than the fundamental vibration mode in pushover analysis [Paret et al. 2000. Matsumori et al. we show that pushover analysis of a one-story system predicts perfectly peak seismic demands. Sasaki et al. the underlying assumptions and approximations are identified. The principal objective of this investigation is to develop an improved pushover analysis procedure based on structural dynamics theory that retains the conceptual simplicity and computational attractiveness of the procedure with invariant force distribution.. 2000.

2)
. where 0 < a << 1. R y .1
One-Story Systems
EQUATION OF MOTION
Consider the idealized one-story structure shown in Fig. On initial loading. It is the familiar bilinear hysteretic relationship. reloading from a minimum deformation takes place along a path parallel to the initial elastic branch. the yield deformation. the yield strength. Unloading from a maximum deformation takes place along a path parallel to the initial elastic branch. a massless frame that provides stiffness to the system. The yield strength is related to f o .1a. Yielding begins when the force reaches f y and the deformation reaches u y . the strength required for the structure to remain elastic during the ground motion. sign u = .2
2. this system is linearly elastic with stiffness k as long as the force does not exceed f y . through the yield strength reduction factor. The yield strength is the same in the two directions of deformation. It consists of a mass m (or weight w) concentrated at the roof level. The hysteretic relation between the lateral force f s and lateral displacement u of the mass relative to the base of the frame is denoted by f s u. 2.1b. During yielding the stiffness of the frame is a k . This lateral-force displacement relation is idealized as shown in Fig. 2. sign u .mug t
bg
The governing equation for this inelastic system subjected to horizontal ground acceleration
b
g
bg
3
(2. defined by f Ry = o fy (2. Similarly. and a linear viscous damper with damping coefficient c.1)
b
g
ug t is mu + cu + f s u.

3).1. Section 7.. deformation is denoted by um . 2. in addition to the form of the force-deformation relation. the deformation u t depends on three systems parameters: w n .3)
ωn =
k m
ζ =
c 2mω n
f fs = s fy
(2. (a) Idealized one-story structure.e. or absolute (without regard to algebraic sign) maximum. The peak.ug t n where
bg
bg
b
g
bg
(2. (2. m depends on three system parameters: w n . and R y (Chopra . and (b) bilinear hysteretic forcedeformation relation
For a given excitation u g t .5)
For a given u g t . This becomes evident if Eq. z y . and the ductility factor is u m= m uy (2. sign u = .1) is divided by m to obtain ~ u + 2zw n u + w 2 u y f s u. 2001. z . Tn = 2p w n is the natural vibration period.fS
m fs c
ut u
fy
1
αk
k
1
uy ug -fy k k um
u
1
1
(a)
(b)
Fig. u £ u y ). and u y . and z is the damping ratio of the system vibrating within its linear elastic range (i.
bg
4
.4)
and w n is the natural vibration frequency.

26 kN (8. Pushover curve for structure shown in Fig.2 kips) . f o w = 184 .26 kN.2: the dimensions and flexural rigidity of the structural elements are noted.
m
h = 3. Vb (kN)
30
20
10
0 0
2
4 6 Roof Displacement.2311 .2
SYSTEM AND EXCITATION CONSIDERED
Consider the one-story system in Fig. 2. The yield strength of the inelastic system.1. 2.5 sec and z = 5% subjected to the north-south component of the El Centro (1940) ground motion (Fig.826 kips)
for
w = 169. 2. with Tn = 0. α = 0. 2001) and scaled up by a factor of 2.66 m
•
EIb
EIc
EIc
•
•
L = 7. u (cm)
8
Fig.3.9 kN (38. One-story.2
5
. this system and excitation. 2. 6. one-bay frame
50
uy = 1.
is
f y / w = ( f o / w ) ÷ 8 = 0.2. based on
Ry = 8 . Vby = 39.2.32 m
•
Fig.04
40
Base Shear. For
.
and
f y = 39.376 cm.4 in Chopra.

0217 rad . the joint rotation and beam hinge rotation are identical to values q m and q pm determined from RHA. the peak value determined from RHA.
2. The resulting pushover curve is shown in Fig. 2.1 kip-in. e.36 cm .4edetermined from RHAis superimposed.4 ).) and 50. Determined from pushover analysis at the exact peak deformation.077 × 107 mm 4 (146 in. and q pm = 0. (d) shows the rotation q p bt g of the plastic hinges at the
bg
bg
beam ends.3
RESPONSE HISTORY ANALYSIS
Figure 2.4 ).65 kN-m (191.4f. (c) shows the joint rotation q bt g .4 shows the earthquake response of the system described in the preceding section determined by response history analysis (RHA). 2.36 cm. 2. as apparent in Fig.017 rad . (b) shows the lateral force f s t or base shear Vb t normalized relative to the
bg weight. the energy dissipated in 6
. The peak values of the various response quantities are as follows: um = 7. and E = 2 × 108 kPa (29×103 ksi) gives yield moments of 21. and (e) shows the force-deformation relation. Observe that the pushover curve matches the initial loading path of the hysteretic system.
2. pushover analysis cannot provide any cumulative measure of response.134 × 107 mm 4 (75. This pushover curve turns out to be bilinear because the beam and the columns are designed to yield simultaneously when f s reaches f y .3.The yield moments in the beam and columns are defined as the bending moments due to the lateral force f y .4e. However. q m = 0. The system is excited well beyond the yield deformation.6 kip-in. respectively..g.4 PUSHOVER ANALYSIS
Static analysis of the one-story nonlinear system subjected to lateral force that increases in small increments is implemented until the lateral displacement reaches um = 7. the ductility factor
m = 5. 2. A static pushover analysis of this one-story system leads to the force-displacement relationship shown in Fig. The yielding stiffness of each structural element is defined as 3% of its initial stiffness. Implementing this analysis with I c = 6.35. wherein the hysteretic force-deformation history of Fig.
I b = 3.18 kN-m (444.) for the beam and columns.3 in. It is organized in five parts: (a) shows the deformation u t .

(d) plastic hinge rotation. 2.04 0.2311
(a)
Vb / w
0 f / w = −0.0217 •
θ (rad)
0
(c)
−0.2
−0. This represents an inherent limitation of pushover analyses.2
0
(e)
0
(f)
−0.yielding during the ground motion.5 0.2
−0.4 −10
−5
0 u (cm)
5
10
−0. Response of one-story system to El Centro ground motion: (a) deformation.04 0. (e) force-deformation relation.04 0
5
10
15 Time (sec)
20
25
30
0.
15
u (cm)
0 • 7.017 • 0
(d)
θp (rad)
−0.2
Vb / w
0.5 fy / w = 0.4 −10
−5
0 u (cm)
5
10
Fig.4. and (f) pushover curve
7
.2311
y
(b)
−0.4
0.36 −15 0. (c) joint rotation. (b) base shear.4
0.04 0. or the cumulative rotation at a plastic hinge.

8
.

2)
The spatial distribution of these “forces” over the height of the building is defined by the vector s = m ι and their time variation by u g t . The right side of Eq. c.3.m ι u g t
bg
bg
(3.12):
mι =
n =1
Â sn =
N
n =1
Â Γ nmφ n
N
(3. This force distribution can be expanded as a summation of modal inertia force distributions s n (Chopra. Elastic Multistory Buildings
3.n t = Â . and k are the mass. m. 2001: Section 13.snug t
n= n =1
The effective earthquake forces can then be expressed as
af
N
af
N
af
(3.mι ug t
bg
bg bg
(3. each element of the influence vector ι is equal to unity.3)
where φ n is the nth natural vibration mode of the structure.1 MODAL RESPONSE HISTORY ANALYSIS
The differential equations governing the response of a multistory building to horizontal earthquake ground motion u g t are as follows: mu + cu + ku = . (3. and
p eff t = Â peff. and lateral stiffness matrices of the system.1) can be interpreted as effective earthquake forces: p eff t = .1)
where u is the vector of N lateral floor displacements relative to the ground.4)
9
. classical damping.

The equations governing the response of the system are mu + cu + ku = .n t is entirely in the nth-mode. with no contribution from other modes.5)
The contribution of the nth mode to s and to p eff t are: sn = Gn mφ n respectively.9)
in which w n is the natural vibration frequency and z n is the damping ratio for the nth mode.Gn ug t
(3. (3.peff (t ) =
n =1
∑ peff . subjected to u g t : Dn + 2z nw n Dn + w 2 Dn = . (3.8)
where the modal coordinate qn t is governed by
2 qn + 2z nw n qn + w n qn = .sn ug t
bg
(3.9) to the equation of motion for the nth-mode elastic SDF system.n (t ) =
N
n =1
∑ −sn u g (t )
N
(3. p eff .10) gives qn t = Gn Dn t
bg
bg
bg
(3. it can be demonstrated that none of the modes other than the nth mode contribute to the response.12)
.n t = . (3.u g t n Comparing Eqs. Then the floor displacements are u n t = φ n qn t
bg
bg bg bg
(3.sn ug t
bg
bg
bg bg
(3.7)
By utilizing the orthogonality property of modes. an SDF system with vibration propertiesnatural frequency w n and damping ratio z n of the nth-mode of the MDF system. we will outline that the response of the MDF system to p eff .6)
Next.10)
bg
bg bg
10
(3. The solution qn t can readily be obtained by comparing Eq.8) gives the floor displacements u n t = Gnφ n Dn t
bg
(3.11)
and substituting in Eq.9) and (3.

these standard equations have been derived in an unconventional way.13) represent the response of the MDF system to p eff.15)
rt =
bg
n =1
Â rn t =
bg
n =1
Â rnst An bt g
N
(3.Forces sn
An(t )
st rn
ωn.can be expressed by
st rn t = rn An t
bg
bg
bg bg bg
(3. 2001. However. Therefore.16)
This is the classical modal RHA procedure wherein Eq.1). Eqs. Section 13. Conceptual explanation of modal response history analysis of elastic MDF systems
Any response quantity r t story drifts.
Equations (3. 3.13)
st where rn denotes the modal static response.15) and (3. ζn ¨ g(t ) u
(a) Static Analysis of Structure
(b) Dynamic Analysis of SDF System
Fig.16) reflect combining the response contributions of all modes.n t .12) and (3.9) is the standard modal equation governing qn t .1. In contrast to the classical 11
bg
. internal element forces. (3. the response of the system to the total excitation p eff t is ut =
bg
bg
bg
n =1 N
Â un bt g = Â Gnf n Dn bt g
n =1
N
N
(3. and
An t = w 2 Dn t n
bg
(3.13) define the contribution of the nth-mode to the response.12) and (3. 3. and Eqs. the static value of r due to external forces sn . etc.
st The two analyses that lead to rn and An t are shown schematically in Fig. (3.1. (3.14)
is the pseudo-acceleration response of the nth-mode SDF system (Chopra.

19)
will provide the same value of rno . which is valid for structures with well-separated natural frequencies such as multistory buildings with symmetric plan.. 2001. z n of the pseudo-acceleration response (or design) spectrum for the nth-mode SDF system.9) .18)
To develop a pushover analysis procedure consistent with RSA.(3.1). (3.17)
where An is the ordinate A Tn .4 and 13. 3. (3. Sections 12.17) (Chopra. we note that static analysis of the structure subjected to lateral forces f no = Γ n mφ n An (3.g. this response value can be obtained by static analysis of the structure subjected to lateral forces distributed over the building height according to 12
. provides an estimate of the peak value of the total response:
b
g
F N r 2 I 1/ 2 ro ª G Â no J H n =1 K
3.3 MODAL PUSHOVER ANALYSIS
(3. Section 13. the peak value rno of the nth-mode contribution rn t to response r t is determined from
st rno = rn An
bg
bg
bg
(3.2 MODAL RESPONSE SPECTRUM ANALYSIS
The peak value ro of the total response r t can be estimated directly from the response spectrum for the ground motion without carrying out the response history analysis (RHA) implied in Eqs.3).16). The SRSS rule.1. and Tn = 2p w n is the natural vibration period of the nth-mode of the MDF system. 2001. Alternatively.8. In such a response spectrum analysis (RSA). The peak modal responses are combined according to the Square-Root-of-Sum-ofSquares (SRSS) or the Complete Quadratic Combination (CQC) rules. Chopra. we used the modal expansion of the spatial distribution of the effective earthquake forces.derivation found in textbooks (e. the peak nth-mode response as in Eq. This concept provides a rational basis for the modal pushover analysis procedure developed later.

The columns are 345 MPa (50 ksi) steel wideflange sections. can be combined according to Eq.1 System and Excitation Considered The 9-story structure. SAC Steel Project Technical Office.73 m (150 ft) in plan.
13
.18) to obtain an estimate of the peak value ro of the total response. The building’s lateral load-resisting system is composed of steel perimeter moment-resisting frames (MRFS) with simple framing on the farthest south E-W frame.73 m (150 ft) by 45. The interior bays of the structure contain simple framing with composite floors. rno . was designed by Brandow & Johnston Associates1 for the SAC2 Phase II Steel Project. the peak value of the roof displacement due to the nth-mode. California.21)
2 where Dn = An ω n . 46th Street. 3. The bays are 9. 1301 S. SAC is a joint venture of three non-profit organizations: The Structural Engineers Association of California (SEAOC). the Applied Technology Council (ATC). The building has a basement level. shown in Fig.s* = mφ n n
(3. Although not actually constructed. This modal pushover analysis (MPA) for linearly elastic systems is equivalent to the well-known RSA procedure (Section 3. which from Eq.4. A benchmark structure for the SAC project. with the ninth level being the roof. 3.2).19 m (122 ft) in elevation. (3. Obviously Dn and An are available from the response (or design)
spectrum. in both directions. CA 90017. Los Angeles. and 37. this structure meets seismic code and represents typical medium-rise buildings designed for the Los Angeles. with five bays each in the north-south (N-S) and east-west (E-W) directions. (3. and California Universities for Research in Earthquake Engineering (CUREE). Typical floor-to-floor heights (for analysis purposes measured from center-of-beam to center1 2
Brandow & Johnston Associates. The peak modal responses. CA 94804-4698. this building is 45. denoted B1. Third St.. The levels of the 9-story building are numbered with respect to the ground level (see Fig. each determined by one pushover analysis.20)
and the structure is pushed to the roof displacement. 3. 1660 W. Richmond. Consulting Structural Engineers. region.4 COMPARATIVE EVALUATION OF ANALYSIS PROCEDURES
3.15 m (30 ft) on center.2).2. urno .12) is
urno = Gnf rn Dn
(3.

01×106 kg (69.49 m (18 ft). Each frame resists one half of the seismic mass associated with the entire structure. dimension.27.7 kips-sec2/ft). The seismic mass of the ground level is 9. including the steel framing. The strength.2 kips-sec2/ft).65×105 kg (66. The floor-to-floor height of the basement level is 3. These force n distributions will be used in the pushover analysis to be presented later. Concrete foundation walls and surrounding soil are assumed to restrain the structure at the ground level from horizontal displacement. s* (Eq.49 sec.96 m (13 ft).89×105 kg (67. 3. 3. the vibration periods are 2. for the first three modes are shown in Fig. respectively. for the first level is 1. 3.sec2/ft). which are seismic (tension) splices to carry bending and uplift forces. and 0.83 m (6 ft) above the center-line of the beam to column joint. roofing and a penthouse located on the roof. i. The seismic mass of the structure is due to various components of the structure. The force distributions. Column splices. The first three vibration modes and periods of the building for linearly elastic vibration are shown in Fig.00×106 kg (616 kips.2. if desired more complex models. The simple model adopted here is sufficient for the objectives of this study. The column lines employ two-tier construction. mechanical/electrical. for the second through eighth levels is 9.. such as those described in Gupta and Krawinkler (1999) can be used. The column bases are modeled as pinned and secured to the ground (at the B-1 level). we select a weak ground motion: the northsouth component of the El Centro (1940) ground motion scaled down by a factor of 4. are located on the first.4. 1988) using the M1 model developed by Krawinkler and Gupta (1998). and for the ninth level is 1. floor slabs. The seismic mass of the above ground levels of the entire structure is 9. This model is based on centerline dimensions of the bare frame in which beams and columns extend from centerline to centerline.07×106 kg (73.e. monolithic column pieces are connected every two levels beginning with the first level. 14
.0 kips-sec2/ft). To ensure that this structure remains elastic. third.85.3. and shear distortion of panel zones are neglected but large deformation (P∆) effects are included.of-beam) are 3.20). The floor system is composed of 248 MPa (36 ksi) steel wide-flange beams in acting composite action with the floor slab. fifth. 0. The building is modeled in DRAIN-2DX (Allahabadi and Powell. partitions. 3. The 9-story N-S MRF is depicted in Fig.65 m (12 ft) and for the first floor is 5. and seventh levels at 1.0 kips-sec2/ft). ceiling/flooring.

2.Fig.. Nine-story building [adapted from Ohtori et al. 2000]
15
. 3.

12 0. (c) shows the joint rotation q n t of an external joint at 16
bg
bg
bg
. Force distributions s* = mφn .27 sec
Ground −1. and 3 .5
Fig. 3. Each figure is organized in four parts: (a) shows the roof displacement urn t .4.12) and (3.04 1.51 0.4.49 sec 3 T = 0.9th 8th 7th 6th
T = 0. 2.39 3.85 sec 2
Floor
5th 4th 3rd 2nd 1st
T1 = 2. 3.1 3.75 1. First three natural-vibration periods and modes of the 9story building
3. 3.2
Response History Analysis
The structural response due to individual vibration modes.7.05 1. is shown in Figs.1 −2. and 3 n
3.05
s
* 1
s
* 2
s
* 3
Fig.03 −1.728 2.44 1.8 −2.796 0.05 2.94 2.67 −1. 2.13)].5 0 0.5 Mode Shape Component
1
1.487 −1.31 −0.0272 −2. (3. (b) shows the base shear Vbn t normalized relative to the weight W of the building.6.5.37 2.72 −2.61 2. n = 1. determined by RHA [Eqs. respectively.3.93 −1.13 −1.33 2. and 3.5
−1
−0.38 0. 3. n = 1.

Also included are the combined response due to one. which are listed in Tables 3. the results for the roof displacement and top-story drift are shown in Fig. with the first three modes provide essentially the exact response.8. The peak values of displacements of all floors. the exact response considering all modes. As expected. consistent with the increasing significance of the higher mode response among these three sets of response quantities.
17
. The peak values of the various response quantities are noted in these figures. and the percentage errors due to truncation of higher modal contributions.urn relation. and 3.9. where the second and third modal responses are a larger percentage of the top story drift compared to roof displacement.8. and 3. ur 2o = 2. or all modes. 3. respectively. respectively.16)]. and ur 3o = 0. especially in estimating the story drifts.the roof level. The peak values of floor displacements and story drifts determined by RHA.15) and (3.3. 3.1.2.1. two. 3. and (d) shows the Vbn .3.422 cm. drifts in all stories. and three vibration modes. The same method was used to determine the peak values of many response quantities. and rotations of external joints with moment connections are presented in Tables 3. The linear relationship between the base shear and roof displacement for each mode implies that the structure did not yield. larger in story drifts. The errors due to truncating the contributions of vibration modes beyond the third mode are negligible. and even larger in joint rotations.23 cm. errors are smallest in floor displacements. in particular. This is illustrated in Fig. It is apparent that the first mode alone is inadequate.12 cm. errors generally decrease as response contributions of more modes are included. including one. Combining the modal response histories for all modes gives the total response [Eqs. two. the peak roof displacement due to each of three modes is ur1o = 9. 3. (3.2. are presented in Fig. 3. but three modes—perhaps even two modes—are sufficient. three. For a fixed number of modes included.

24E-03 2.282 0.29E-03 2.72E-03 3.235 0.14E-03 2.8 -10.4 -10.3 -8.76E-03 1.024 -0.038 0.069 0.7 -19.199 0.89E-03 1.295 -0.152 0.011
0.006 0.407 0.225 0.259 0.325 0.6 -1.311 0.4 0.08E-03 2.50E-03 4.91E-04 1.055 0.1
Building height is measured from the ground floor to the 9th floor.1 3.56E-03 2.4 -0.4 -7.0 -2.03E-03
3 Modes
2.31E-03
Mode 2
-1.9 -22.202 0.245
0.399 0.3 -3.03E-03 1.3
Table 3.042 0.1 4.03E-03 1.8
-1.65E-03 2.231 -0.1 -19.9 8.6 4.2 -2.4 -22.466 0.78E-04 -3.282 0.333 0.64E-03 3.6 11.2
9.6 1.7 3.28E-04 1.179 0.058 -0.22E-03 2.74E-04 6.1 -14.45E-03 3.15E-03 4.00E-03 2.406 0.264
Error (%)
1 Mode 2 Modes 3 Modes
-23.177 0.364 0.336 0.38E-04
2 Modes
2.260 0.266 0.060 -0.09E-03
Error (%)
1 Mode
-23.7 2.008 -0.11E-03 1.4 -41.8 1.245
0.121
0.9 -15.74E-03 1.307 0.097 0.080 0.2 Peak values of story drift ratios (as % of story height) from RHA for 0.227 0.205 0.125 0.6 9.378
0.157 0.012 0.3 -0.1 -2.062 -0.266 0.253 0.275 0.202 0.42E-04 1.229 0.035 0.156 0.042 0.265
RHA (all modes) 0.2 1.13E-04 9.7 -50.1
Table 3.227 0.097 0.226 0.054 0.202 0.7
2 Modes
-3.50E-03 2.015 -0.183 0.09E-03 1.01E-04 3.99E-03
Mode 3
3.117 0.1 -2.062 0.002 -0.4
-1.26E-04 -5.9 -24.1 -11.6 -1.00E-03 1.73E-03 3.38E-03 2.9
9.173
0.74E-04 9.0 -0.
18
.253 0.5 -2.Table 3.156 0.125 0.060 0.2 0.235 0.5 18.300 0.3 -33.133 -0.401
Error (%)
1 Mode 2 Modes 3 Modes
-23.310 0.01E-04 -2.5 0.260 0.88E-03 2.11E-04 -5.9 -16.043 0.181 0.88E-03 2.5 -1.13E-03 2.413
RHA (all modes) 0.370 0.0 -2.7 4.237 0.63E-03 2.63E-03 2.66E-05 -3.097 0.023 0.130 0.89E-03 1.069 0.7 7.85E-03 3.090 0.2 -20.1 -2.203 0.259 0.90E-03 3.226 0.022 0.001 -0.47E-03 1.177 0.0 -46.0 7.453 0.266 0.8 -5.1 Peak values of floor displacements (as % of building height = 37.173
0.197 0.9 2.9 3.303 0.124 0.3 19.6 0.152 0.6 -0.3 -0.4 -6.76E-03 1.003 0.229 0.265 0.9 -23.258
0.8 -56.25 × El Centro ground motion Story 1st 2nd 3rd 4th 5th 6th 7th 8th 9th Modal Response
Mode 1 Mode 2
Drift Ratio (%) Combined (RHA)
1 Mode 2 Modes 3 Modes
Mode 3
0.321 0.159 0.4 -1.124 0.09E-03 2.44E-03 1.00E-03 1.010 -0.94E-03 2.14 m3) from RHA for 0.060
-0.3 Peak values of joint rotations (radians) from RHA for 0.44E-03 3.088 -0.350 0.25 × El Centro ground motion Displacement /Height (%) Modal Response Combined (RHA)
Mode 1 Mode 2 Mode 3 1 Mode 2 Modes 3 Modes
Floor 1st 2nd 3rd 4th 5th 6th 7th 8th 9th
0.237 0.2 -57.192 0.1 -0.012 -0.032 -0.4 -3.026 0.33E-03 2.0
3 Modes
-5.8 1.74E-03 1.2 -1.263 0.261
-0.400 0.25 × El Centro ground motion Floor
1st 2nd 3rd 4th 5th 6th 7th 8th 9th
3
Modal Response
Mode 1
2.089 0.6 0.65E-03
RHA (all modes)
2.31E-03
Joint Rotation (rad) Combined (RHA)
1 Mode
2.071 0.2 -4.9 1.045 0.0 -10.8 -15.99E-03 2.311 0.4 -53.09E-03 1.475 0.03E-03 -6.009 -0.011 0.

0435
9.12 • 0
(a)
r1
−15 0.1 −15
−7.5
0 u (cm)
r1
7. 3. (b) base shear.0435
Vb1 / W
0. (c) joint rotation.5
9.25 × El Centro ground motion
19
.0435 • 0
(b)
V
b1
−0. (d) force-deformation history.15
u (cm)
9. and (e) pushover curve.1
0.5
15
−0.12
−0. Response due to first mode: (a) roof displacement.00174 • 0
(c)
r1
−5 0
5
10
15 Time (sec)
20
25
30
0.5
0 u (cm)
r1
7.1
0.1 5 x 10
−3
θ (rad)
0.12
0
(d)
0
(e)
15
Fig.5. Excitation is 0.1
/W
0.1 −15
−7.

23
0.05 5 x 10
−3
b2
θ (rad)
0.5
5
−0.1 −5
−2.05
/W
0. and (e) pushover curve.1
0. (d) force-deformation history. Response due to second mode: (a) roof displacement. 3.5
u (cm)
r2
0 • 2.6.0252
−2. Excitation is 0.23 −5
(a)
0. (b) base shear. (c) joint rotation.25 × El Centro ground motion
20
.00244 • 0
(c)
r2
−5 0
5
10
15 Time (sec)
20
25
30
0.0252 • 0
(b)
V
−0.1
Vb2 / W
0.1 −5
−2.23 −2.5
5
Fig.5
0 u (cm)
r2
2.5
0 u (cm)
r2
2.0252
(d)
0
0
(e)
−0.

3.5
0 u (cm)
r3
0. Response due to third mode: (a) roof displacement.422
(e)
−0. and (e) pushover curve.0108
−0.02 −1
−0. (b) base shear. (c) joint rotation.5
1
−0.02
−0.02
/W
V
b3
0 • 0.02
0.02 −1
−0.5
1
Fig.1
u (cm)
r3
0 • 0.422
/W
V
b3
0
(d)
0
−0. Excitation is 0.7.0108 x 10
−3
(b)
−0.02 2
θ (rad)
0.0108
−0.422 −1
(a)
0. (d) force-deformation history.25 × El Centro ground motion
21
.000913 • 0
(c)
r3
−2 0
5
10
15 Time (sec)
20
25
30
0.5
0 u (cm)
r2
0.

3. and it is apparent in the example considered that three modes provide most of the response (Fig. 3.1.5d). implying that the modal truncation errors are small if at 22
. story drifts.3).3
Modal Pushover Analysis
Implementing MPA for the fundamental vibration mode.21). and ur 3o = 0. although both analyses led to identical peak values of the individual modal responses.6d and 3. using the force distribution of Eq. 3.6 present estimates of the combined response according to Eq. considering one.10 and Tables 3.4.18). i. (3. Implementing pushover analysis for the second and third modes.12 cm.4) to roof displacement ur1o = 9. the floor displacements are proportional to the mode shape φ 1 because the structure remains elastic. the value determined by RHA (Fig.4.9). this value would be determined directly from the response (or design) spectrum.5.8) leads to the pushover curve shown in Fig. (3.6e and 3.5. 3. 3.7e and to the floor displacements.ur relations determined by RHA (Figs. These values of the response quantities are identical to the peak response values determined from RHA (Tables 3. i. Figure 3..6.3. and 3. and 3. In RHA the errors arise only from truncating the responses due to higher modes. Observe that the target roof displacement in each pushover analysis is identical to its exact value determined by RHA.422 cm .1. respectively. pushing the structure using the force distribution of Eq. and 3. (3.20) with n = 1 (Fig. This pushover curve is consistent with the relationship between the base shear and roof displacement determined by RHA (Fig. respectively.
3.e.2. 3. these pushover curves are consistent with the Vb . For a fixed number of modes included considered.7d). (3. and the computed response values are identical to the peak response values determined from RHA (Tables 3. 3. and external joint rotations in Tables 3. which would provide the Dn value to be substituted in Eq.12).6.4. leads to the pushover curves shown in Figs. 3.9 and Tables 3. As for the first mode. pushing the structure.4. or three vibration modes. 3.3). The floor displacements. (3.1 through 3. As suggested by Eq.2. and external joint rotations computed by pushover analysis are presented in Tables 3. and 3. 3.5. two. In practical application. the errors in the MPA results are generally larger than in RHA (Fig. and 3. respectively.20) with n = 2 and 3 up to roof displacements
ur 2 o = 2..23 cm .5e. 3. 3. except for the algebraic sign associated with Γ1 and minor round-off errors. confirming that MPA gives the exact values of the individual modal responses. story drifts.e.3). and the errors in these estimates relative to the exact response from RHA considering all modes.

4 Story Drift Ratio (%)
0.25 × El Centro ground motion: first three modal responses and total (all modes) response
(a) Floor Displacements 9th 8th 7th 6th 9th 8th 7th 6th (b) Story Drift Ratios
Floor
4th 3rd 2nd 1st
Floor
5th RHA All Modes 1 Mode 2 Modes 3 Modes
5th 4th 3rd 2nd 1st RHA All Modes 1 Mode 2 Modes 3 Modes
fig3_9a.685 •
−3 3
Mode 2
∆ (cm)
u (cm)
• 2.5
0. Response histories of roof displacement and top-story drift from RHA for 0.23
r2
−15 15 0 • 0.1
0. Additional errors are introduced in pushover analysis due to the approximation inherent in modal combination rules.9.8.4 Displacement/Height (%)
0.3 0.eps
fig3_9b.least three modes are included.422
r2
0
• 1.eps
Ground 0
0.03
−3 3
Mode 3
∆ (cm)
u (cm)
r3
−15 15 9.25 × El Centro ground motion
23
. 3.2 0.1
0.12 •
r1
−15 15 0
r1
0
0
0.3 0. sec
25
30
r
5
10 15 20 Time. Heightwise variation of floor displacements and story drifts from RHA for 0. 3.2 0.
(a) Roof Displacement 15 (b) Top Story Drift 3
Mode 1
∆ (cm)
u (cm)
9.83 •
r3
0
• 0.eps
r
5
10 15 20 Time.48
−3 3 1.6
Fig.5
Ground 0
0. sec
25
30
Fig.59 •
Total (All Modes)
0 −15 0
∆ (cm)
u (cm)
0 −3 0
fig3_8.

73E-05 3.274 0.7 -19.260 -0.03E-03 -1.1 -0.31E-03
Error (%) RHA (all modes)
2.336
RHA (all modes) 0.152 0.374 0.253
RHA (All Modes) 0.9 -14.267 0.2 -0.74E-03 -1.89E-03 1.328 0.89E-03 -1.157 0.9 -15.9 -24.133 -0.2 -16.00E-03 2.8 -14.9 -18.282 0.2 -11.4 0.227 0.270 0.7 -50.4 Peak values of floor displacements (as % of building height = 37.4 -2.8 -22.230 0.253 0.7 -21.331 0.4 -9.4 -9.4 -5.76E-03 -1.133 0.03E-03 6.4 -19.9
3 Modes
-12.2
Table 3.230 -0.023 -0.14 m) from MPA for 0.152 0.57E-03
1 Mode
-23.9 -14.53E-04 -9.125 0.0 -18.24E-03 -2.466 0.270 0.05E-03 3.261
-0.08E-03 2.229 0.010 0.15E-03 1.181 0.245
0.40E-04 5.00E-03 -1.237 0.002 0.080 0.285 0.156 0.03E-03 1.3 -12.253
0.8 -23.31E-03
Joint Rotation (rad) Combined (MPA)
1 Mode
2.133 0.036 -0.235 -0.2 1.22E-04 2.276 0.12E-03 1.4 -14.25 × El Centro ground motion Story 1st 2nd 3rd 4th 5th 6th 7th 8th 9th Modal Response
Mode 1 Mode 2
Drift Ratio (%) Combined (MPA)
1 Mode 2 Modes 3 Modes
Mode 3
-0.1 -18.090 0.3 -2.177 0.Table 3.0 -10.125 0.1 -11.227 0.99E-03
Mode 3
-3.253 -0.4 -7.0 -46.296 -0.4 -53.3 -14.203 0.203 0.173
0.001 0.079 0.09E-04 -3.332 0.0 -0.3 -19.179 0.006 -0.6 -16.197 0.3 -13.4 -4.33E-04 5.0 -2.009 0.022 0.38E-03 2.38E-03
3 Modes
2.2 -20.3 1.9
-12.272 0.229 0.203 0.179 0.6 -15.300 0.106 0.15E-03 2.1 -19.407 0.73E-03 3.4 -22.7 2.055 0.042 0.6 -17.4 -22.310 0.1 -11.336 0.3
Table 3.270 0.321 0.9 -8.9 -15.266 0.76E-03 1.069 0.012 -0.04E-03 3.121
0.8 -15.42E-04 -1.237 -0.22E-03 -2.177 0.370 0.8 -15.259 0.9 -13.8 -56.0 -16.9 -16.267 0.63E-03 -2.89E-03 -2.235 0.97E-03 1.267 -0.00E-03 2.03E-03 3.44E-03 -1.097 0.231 -0.3 -9.026 0.6 Peak values of joint rotation (radians) from MPA for 0.062 -0.09E-03
Mode 2
1.4
-11.89E-03 2.078 0.088 -0.313
0.045 -0.5 Peak values of story drift ratios (as % of story height) from MPA for 0.00E-03 2.156 0.5 -16.9
-11.7 -15.90E-04 -9.203 0.011
0.322 0.96E-03 2.9 -15.9 -13.94E-03 2.3 -41.173
0.7
2 Modes
-13.226 0.31E-03 2.2 -57.282 -0.09E-03 1.042 0.3 -33.058 -0.117 0.015 -0.012 0.00E-03 3.264
Error (%) Modal Response
1 Mode 2 Modes 3 Modes
-23.203 0.043 -0.401
Error (%)
1 Mode 2 Modes 3 Modes
-23.9 -15.28E-03 2.008 -0.048 0.80E-04 3.44E-03 3.09E-03 -1.069 0.65E-03 2.259 -0.63E-03 2.097 0.9 -13.7
24
.74E-03 1.2 -4.47E-03 1.245
-0.048 0.038 -0.78E-04
2 Modes
2.060
0.74E-04 -6.011 -0.024 0.92E-04 -1.003 -0.071 0.106 0.062 0.90E-03 1.72E-03 3.2
-12.25 × El Centro ground motion Floor 1st 2nd 3rd 4th 5th 6th 7th 8th 9th Modal Response
Mode 1
-2.032 0.89E-03 2.9 -13.4 1.00E-03 1.157 0.4 -4.6 -19.060 -0.124 0.260 0.260 0.43E-04 -1.9 -14.385 0.25 × El Centro ground motion Floor 1st 2nd 3rd 4th 5th 6th 7th 8th 9th Displacement /Height (%) Modal Response Combined (MPA)
Mode 1 Mode 2 Mode 3 1 Mode 2 Modes 3 Modes
0.286 0.

1
0.4 Displacement/Height (%)
0.1
0.4 Story Drift Ratio (%)
0.6
Fig.2 0. 3.5
Ground 0
0.(a) Floor Displacements 9th 8th 7th 6th 9th 8th 7th 6th
(b) Story Drift Ratios
Floor
4th 3rd 2nd 1st
Floor
5th
RHA All Modes MPA 1 Mode 2 Modes 3 Modes
5th 4th 3rd 2nd 1st
RHA All Modes MPA 1 Mode 2 Modes 3 Modes
fig3_10a.5
0. shading indicates modal combination error
25
.2 0. Heightwise variation of floor displacements and story drifts from MPA for 0.10.3 0.25 × El Centro ground motion.eps
Ground 0
0.3 0.eps
fig3_10b.

26
.

1) is not valid for inelastic systems. Although classical modal analysis (Section 3. Expanding the displacements of the inelastic system in terms of the natural vibration modes of the corresponding linear system we get ut =
bg
n =1
Â φ nqn bt g
27
N
(4.3)
.1
Inelastic Multistory Buildings
RESPONSE HISTORY ANALYSIS
For each structural element of a building. (3.1) becomes mu + cu + f s u. the initial loading curve is idealized as bilinear. (4. Each structural element of this elastic system is defined to have the same stiffness as the initial stiffness of the structural element of the inelastic system. Therefore. leading to the “exact” nonlinear response history analysis (RHA). but depend on the history of the displacements: f s = f s u. it is useful for later reference to transform Eq.2)
The standard approach is to solve directly these coupled equations. sign u
b
g g bg
(4. Both systems have the same mass and damping. the relations between lateral forces f s at the N floor levels and the lateral displacements u are not single valued.4
4. Eq.2) to the modal coordinates of the corresponding linear system. the natural vibration periods and modes of the corresponding linear system are the same as the vibration properties of the inelastic system undergoing small oscillations (within the linear range). Thus. sign u = .1)
With this generalization for inelastic systems. and the unloading and reloading curves differ from the initial loading branch.mι ug t
b
(4.

sn ug t
bg
b
g
bg bg
(4.and classical n damping-orthogonality property of modes gives F qn + 2z nw n qn + sn = .2). sign u n
b
g
b
g bg
(4. 4. premultiplying by φ T .5)
This resisting force depends on all modal coordinates qn t . This approximate RHA procedure is the preliminary step in developing a modal pushover analysis procedure for inelastic systems. give the same results for u t as obtained directly from Eq. sign qn = φ n f s u n . Unlike Eq.Gn ug t Mn
bg
n = 1. (3.2 UNCOUPLED MODAL RESPONSE HISTORY ANALYSIS
bg
Neglecting the coupling of the N equations in modal coordinates [Eq. (4. (4.4) is rarely solved because it offers no particular advantage over Eq.8) because qr t will generally be nonzero for “modes” other than the nth “mode. (3. The spatial distribution s of the effective earthquake forces is expanded into the modal contributions sn according to Eq.” implying that other “modes” will also contribute to the solution.6)
The solution of Eq. 2. (4.n t given by Eq.2).… N
(4. (3. (4.9) involves
T Fsn = Fsn qn .3). (4.3) in Eq.9) for linearly elastic systems.6) for inelastic systems will no longer be described by Eq. For linear systems. (3. (4. and using the mass. implying coupling of modal coordinates because of yielding of the structure. (4. sign u = .4) represents N equations in the modal coordinates qn . however. The equations governing the response of the inelastic system to p eff. in principle. (4.3) will.4)] leads to the uncoupled modal response history analysis (UMRHA) procedure. these equations are coupled for inelastic systems.6b) are mu + cu + f s u.2). However. Equation (4. Eq. Simultaneously solving these coupled equations and using Eq.4)
where the only term that differs from Eq. qr t = 0 for all 28
bg
. (3.Substituting Eq. where φ n are now the modes of the corresponding linear system.

(a) p 15

eff,1

= −s × 0.25 × El Centro

1

(b) p 5

eff,2

= −s × 0.25 × El Centro

2

ur1 (cm)

0

ur1 (cm)

9.112 •

Mode 1

Mode 1

0

−15 15 Mode 2

−5 5 Mode 2

ur2 (cm)

0

ur2 (cm)

0 • 2.226 5 10 15 20 25 30

−15 0 15

5

10

15

20

25

30

Mode 3

−5 0 5

Mode 3

ur3 (cm)

0

ur3 (cm)

5 10 15 20 Time (sec) 25 30

0

−15 0

−5 0

5

10 15 20 Time (sec)

25

30

Fig. 4.1. Modal decomposition of the roof displacement due to (a) p eff,1 t = - s1 ¥ 0.25

**¥ El Centro ground motion; and (b) p eff,2 t = - s 2 ¥ 0.25 ¥ El Centro ground
**

motion

af

af

modes other than the nth-mode; therefore, it is reasonable to expect that the nth “mode” should be dominant even for inelastic systems. These assertions are illustrated numerically in Figs. 4.1 and 4.2 for the selected 9-story building. Equation (4.6) was solved by nonlinear RHA, and the resulting roof displacement history was decomposed into its “modal” components. The modal decomposition of the roof displacement for the first three modes due to 0.25 ¥ El Centro ground motion demonstrates that because the building does not yield during this weak ground motion, the response to excitation p eff,n t is all in the nth-mode (Fig. 4.1). The structure yields when subjected to the strong excitation of 1.5 ¥ El Centro ground motion, and the modes other than the nth-mode contribute to the response. The second and third modes start responding to excitation p eff,1 t at about 29

bg

bg

(a) p 80

eff,1

= −s × 1.5 × El Centro

1

(b) p 20

eff,2

= −s × 1.5 × El Centro

2

Mode 1

Mode 1 • 1.437

u (cm)

r1

0 • 48.24

ur1 (cm)

Mode 2

0

−80 80

−20 20 Mode 2

u (cm)

r2

0

3.37 •

ur2 (cm)

0 • 11.62 5 10 15 20 25 30

−80 0 80

5

10

15

20

25

30

Mode 3 0.4931 •

−20 0 20

Mode 3 • 0.7783

u (cm)

r3

0

ur3 (cm)

25 30

0

−80 0

5

10 15 20 Time (sec)

−20 0

5

10 15 20 Time (sec)

25

30

Fig. 4.2. Modal decomposition of the roof displacement due to: (a) p eff,1 t = - s1 ¥ 15 ¥ . El Centro ground motion; and (b) p eff,2 t = - s2 ¥ 15 ¥ El Centro ground . motion

af

af

5.2 sec, the instant the structure first yields; however, their contributions to the roof displacement are only 7% and 1%, respectively, of the first mode response (Fig. 4.2a). The first and third modes start responding to excitation p eff,2 t at about 4.2 sec, the instant the structure first yields; however, their contributions to the roof displacement are 12% and 7%, respectively, of the second mode response (Fig. 4.2b). Approximating the response of the structure to excitation p eff,n t

bg

bg

by Eq. (3.8),

substituting Eq. (3.8) in Eq. (4.6) and premultiplying by φ T gives Eq. (4.4), except for the n important approximation that Fsn now depends only on one modal coordinate, qn :

T Fsn = Fsn qn , sign qn = φ n f s qn , sign qn

b

g

b

g

(4.7)

30

**With this approximation, solution of Eq. (4.4) can be expressed by Eq. (3.11) where Dn t is governed by F Dn + 2z nw n Dn + sn = - ug t Ln and
**

T Fsn = Fsn Dn , sign Dn = φ n f s Dn , sign Dn

bg

bg c h

(4.8)

c

h

(4.9)

is related to Fsn qn , sign qn because of Eq. (3.11). Equation (4.8) may be interpreted as the governing equation for the nth-“mode” inelastic SDF system, an SDF system with (1) small amplitude vibration properties—natural frequency

b

g

**w n and damping ratio z n —of the nth-mode of the corresponding linear MDF system; (2) unit
**

mass; and (3) Fsn Ln - Dn relation between resisting force Fsn Ln and modal coordinate Dn defined by Eq. (4.9). Although Eq. (4.4) can be solved in its original form, Eq. (4.8) can be solved conveniently by standard software because it is of the same form as the SDF system [Eq. (2.3)], and the peak value of Dn t can be estimated from the inelastic response (or design) spectrum (Chopra, 2001; Sections 7.6 and 7.12.1). Introducing the nth-“mode” inelastic SDF system also permitted extension of the well-established concepts for elastic systems to inelastic systems. Compare Eqs. (4.4) and (4.8) to Eqs. (3.9) and (3.10): note that Eq. (3.11) applies to both systems.4 Solution of the nonlinear Eq. (4.8) formulated in this manner provides Dn t , which substituted into Eq. (3.12) gives the floor displacements of the structure associated with the nth“mode” inelastic SDF system. Any floor displacement, story drift, or another deformation response quantity r t is given by Eqs. (3.13) and (3.14), where An t is now the pseudost acceleration response of the nth-“mode” inelastic SDF system. The two analyses that lead to rn

bg

bg

bg

bg

b g are shown schematically in Fig. 4.3. Equations (3.13) and (3.14) represent the response of the inelastic MDF system to p eff,n bt g , the nth-mode contribution to p eff bt g . Therefore the response of the system to the total excitation p eff bt g is given by Eqs. (3.15) and

and An t (3.16). This is the UMRHA procedure.

4

Equivalent inelastic SDF systems have been defined differently by other researchers (Villaverde, 1996; Han and Wen, 1997).

31

Conceptual explanation of uncoupled modal response history analysis of inelastic MDF systems
(a) Nonlinear RHA 150 n=1 150
(b) UMRHA n=1
ur1 (cm)
• 75.51 •
ur2 (cm)
0
ur2 (cm)
n=3
14.n t = . where u g t = 3.s n u g t .3.Forces sn Unit mass An(t )
st rn
ωn.02
−150 n=2
−150 50 n=2 14. Fsn / Ln ¨ g(t ) u
(a) Static Analysis of Structure
(b) Dynamic Analysis of Inelastic SDF System
Fig. 4.0 ¥ El Centro ground motion: (a) exact solution by NL-RHA. and 3. Roof displacement due to p eff.575 5 10 15 20 Time (sec) 25 30
ur3 (cm)
0 • 5.4. 2.89 50
r1
0
u (cm)
0 • 78.9 •
0
−50 10
−50 10 n=3
ur3 (cm)
0 • 5. 4. n = 1. and (b) approximate solution by UMRHA
bg
af
bg
32
. ζn.112 5 10 15 20 Time (sec) 25 30
−10 0
−10 0
Fig.

2… N ) according to Eq.n t (n = 1.4) and (4. Such comparison for roof displacement and top-story drift is presented in Figs. These errors arise from the following assumptions and approximations: (1) the coupling among modal coordinates qn t arising from yielding of the system [recall Eqs. and (b) approximate solution by UMRHA
af
bg
bg
4.2. and 3. and (3) the Fsn Ln . (4. this intense excitation was chosen to ensure that the structure is excited well beyond its linear elastic limit. Top story drift due to p eff. where u g t = 3. (4. (2) the superposition of responses to p eff.4 and 4. 2.5.15) is strictly valid only for linearly elastic systems. both for 3.(a) Nonlinear RHA 20 n=1 20
(b) UMRHA n=1 • 5.s n u g t . respectively.Dn relation is 33
bg
bg
. (3. but even for this very intense excitation.744 • 0
−20 10 n=3
−20 10 n=3
∆r3 (cm)
0 • 5.0 ¥ El Centro ground motion. 4.33
0
−20 20
−20 20 n=2
∆r2 (cm)
0
∆r2 (cm)
7.817 5 10 15 20 Time (sec) 25 30
−10 0
−10 0
Fig.1
Underlying Assumptions and Accuracy
The approximate solution of Eq.6) by UMRHA is compared with the “exact” solution by nonlinear RHA.0 ¥ El Centro ground motion: (a) exact solution by NL-RHA.855
∆r1 (cm)
0
∆r1 (cm)
n=2
• 6. the errors in either response quantity are only a few percent. The errors are slightly larger in drift than in displacement. 4.5. n = 1.n t = .956 5 10 15 20 Time (sec) 25 30
∆r3 (cm)
0 • 5.008 •
6.5)] is neglected.

Dn relation? The two sets of forces and displacements are related as follows: V Fsn = bn Γn Dn = urn Γ nf rn 34 (4. (4. How to convert this Vbn . and the base shear Vbn .9) should be determined by nonlinear static analysis of the structure as the structure undergoes displacements u = Dnφ n with increasing Dn . (3.approximated by a bilinear curve to facilitate solution of Eq. the only force distribution that produces displacements proportional to f n is given by Eq. At
the yield point.10)
.20) to some
bg
pre-determined roof displacement. (4.urn pushover curve to the Fsn Ln . this distribution seems to be a rational choiceeven for inelastic systemsto determine Fsn in Eq. such nonlinear static analysis provides the so-called pushover curve.4.Dn relation to be determined in Eq. Although several approximations are inherent in this UMRHA procedure. within the linearly elastic range of the structure. 4.2. when specialized for linearly elastic systems it is identical to the RHA procedure of Section 3. The overall errors in the UMRHA procedure are documented in the examples presented in Section 4. A bilinear idealization of this pushover curve for the nth-“mode” is shown in Fig.Dn curve. (4.9).12) for floor displacements. However.8) in UMRHA. Therefore.8) governing Dn t is based on Eq. The structure is pushed using the force distribution of Eq. (3. which is different than the
Fsn Ln .6a.20). the base shear is Vbny and roof displacement is urny .1. (3. is plotted against roof displacement
urn . When implemented by commercially available software.8) before it can be solved? Because Eq. (4. What is an appropriate invariant distribution of lateral forces to determine Fsn ? For an inelastic system no invariant distribution of forces can produce displacements proportional to f n at all displacements or force levels. Therefore we impose this constraint in developing the UMRHA procedure in this section and modal pushover analysis in the next section. Although most commercially available software cannot implement such displacement-controlled analysis. the relationship between lateral forces f s and Dn in Eq.2 Properties of the nth-“mode” Inelastic SDF System How is the Fsn Ln . it can conduct a force-controlled nonlinear static analysis with an invariant distribution of lateral forces. (4.
4.

4. which may differ from the period of the corresponding linear system. should be used in Eq. 4. 4.V
bn
(a) Idealized Pushover Curve
F
sn
/L
n
(b) F
sn
/ L − D Relationship
n n
Idealized Vbny Actual 1
αnkn
V
bny
/M
* n
1
α ω2 n n
kn 1 1
ωn
2
u u
rny
rn
Dny = ur n y / Γn φ r n
D
n
Fig.6.6b is w 2 . 4.12)
implying that the initial slope of the curve in Fig.6a is 2 kn = ω n Ln . The two are
related through Fsny Ln
2 = w n Dny
(4. which is not a meaningful quantity. 2001.2.
35
. Section 13. (4. the initial slope of the pushover curve in Fig.11)
* in which M n = Ln Γn is the effective modal mass (Chopra.13)
This value of Tn . where the yield values of Fsn Ln and Dn are Fsny Ln = Vbny
* Mn
Dny =
urny Γ nf rn
(4.11). In contrast. Properties of the nth-“mode” inelastic SDF system from the pushover curve
Equation 4.6b.5).8). Knowing Fsny Ln and Dny from n Eq.10 enables conversion of the pushover curve to the desired Fsn Ln − Dn relation shown in Fig. (4. the elastic vibration period Tn of the nth-mode SDF system is computed from
F Ln Dny I 1/ 2 Tn = 2p G H Fsny JK
(4.

(3. 6. This value of the roof displacement is given by Eq. Consider a nonlinear static analysis of the structure subjected to lateral forces distributed over the building height according to s* [Eq.20)]. (3. r o . 4. of the total response r t obtained in Step 8. ωn . 4. (3. develop the base-shear – roof-displacement ( Vbn − urn ) pushover curve for the force distribution s* [Eq. Combine the “modal” responses using Eqs. Repeat Steps 2 to 6 for as many modes as required for sufficient accuracy. the peak value of Dn t . Typically. 5.11). 4.20)]. 8. 9.6b) by utilizing Eq.2. Dn (t ) . and modes.13). Calculate the peak value. with the structure is pushed to the roof displacement urno .6b. is now determined by 36
bg
bg
. 2. which is summarized next as a sequence of steps.16) to determine the total response. Compute the natural frequencies. (4.3 MODAL PUSHOVER ANALYSIS
bg
bg
A pushover analysis procedure is presented next to estimate the peak response rno of the inelastic MDF system to effective earthquake forces p eff . and pseudo-acceleration history. of the nth-“mode” inelastic SDF system (Fig.3
Summary
The inelastic response of an N-story building with plan symmetric about two orthogonal axes to earthquake ground motion along an axis of symmetry can be estimated as a function of time by the UMRHA procedure just developed. n (3. 4. φn . Compute the deformation history.12) and (3.6a). Convert the idealized pushover curve to the Fsn Ln − Dn relation (Fig.21) where Dn . For the nth-mode.3b) with force-deformation relation of Fig.15) and (3. 4. n 3. An t . 7. Calculate histories of various responses by Eqs. the first two or three modes will suffice. for linearly-elastic vibration of the building.n t . 4. Idealize the pushover curve as a bilinear curve (Fig. details are available in Appendix A: 1.4. (3.

Dn . Steps 1 to 4 of the MPA are same as those for UMRHA.18). Compute the peak deformation. for elastic systems.2..3. we will refer to rno as the peak “modal” response even in the case of inelastic systems. seems reasonable.3) because. As shown in Sections 3. The peak “modal” responses rno .
4. (3. which is summarized next as a sequence of steps. as mentioned earlier.2. and (2) it provides the exact modal response for elastic systems. as described in Section 4.21). each determined by one pushover analysis.8). or from the inelastic response (or design) spectrum. although somewhat intuitive for inelastic buildings.2 and 3. the lateral force distribution used possesses two properties: (1) it appears to be the most rational choice among all invariant distribution of forces. 5. This application of modal combination rules to inelastic systems obviously lacks a theoretical basis. joint rotations. (3. etc.3.n t . Eq. It provides results for elastic buildings that are identical to the well-known RSA procedure (Section 3. it can be determined from the inelastic response (or design) spectrum (Chopra. At this roof displacement.6b by solving Eq. 6. e.6). 37
. (4.g. it seems reasonable because it provides results for elastic buildings that are identical to the well-known RSA procedure described in Section 3.
rno also represents the exact peak value of the nth-mode contribution rn t to response r t .6 and 7. are combined using an appropriate modal combination rule. of the nth-“mode” inelastic SDF system (Fig. This pushover analysis. plastic hinge rotations. However. Sections 7. 4. (4. (4. the pushover analysis provides an estimate of the peak value rno of any response rn t : floor displacements. story drifts. 4. The response value rno is an estimate of the peak value of the response of the inelastic system to p eff. to obtain an estimate of the peak value ro of the total response.4.1
Summary
The peak inelastic response of a building to earthquake excitation can be estimated by the MPA procedure just developed. alternatively.8).3b) with force-deformation relation of Fig. Calculate the peak roof displacement urno associated with the nth-“mode” inelastic SDF system from Eq.solving Eq. details are available in Appendix B. 2001.
bg
bg
bg
bg
Thus. governed by Eq.12).

respectively.8. the “exact” results.4. the first two or three “modes” will suffice. although the trends are not systematic as when the system remained elastic (Section 3. determined by the UMRHA procedure. 9. Observe that errors tend to decrease as response contributions of more “modes” are included.1). in particular.4 COMPARATIVE EVALUATION OF ANALYSIS PROCEDURES
The response of the 9-story building described earlier is determined by the two approximate methods: UMRHA and MPA. 4. in contrast to modal analysis (Section 3. the UMRHA procedure lacks a rigorous theory. with. the 38
. From the total rotation of a plastic hinge.7 cm. and the “exact” response from nonlinear RHA for the roof displacement and top-story drift. the combined response due to three “modes”. and ur 3o = 2.0. Determine the total response by combining the peak “modal” responses using the SRSS combination rule of Eq.2).1 Uncoupled Modal Response History Analysis
The structural response to 1. and the errors in the approximate results are shown in Fig. the peak roof displacement due to each of the three “modes” is ur1o = 48. Repeat Steps 3 to 8 for as many “modes” as required for sufficient accuracy. two. subtract the yield value of hinge rotation to determine the hinge plastic rotation. This deficiency also implies that. say. two. extract from the pushover database values of other desired responses. The peak values of response are as noted. 4. Typically. and the percentage errors in the approximate results.0 to 3.9.3 cm.7.1 and 4. is presented next. 4. To ensure that this structure responds well into the inelastic range the El Centro ground motion is scaled up by a factor varying from 1.53 cm.5 ¥ the El Centro ground motion including the response contributions associated with three “modal” inelastic SDF systems.7 shows the individual “modal” responses. also included are the combined responses due to one. three “modes” included. rno . The peak values of displacements of all floors and drifts in all stories are presented in Tables 4. and three “modes”. and compared with the results of a rigorous nonlinear RHA using the DRAIN-2DX computer program. At urno .2.4. Figure 4. 4. ur 2o = 11. 8. The peak values of floor displacements and story drifts including one.18). (3. and three modes are compared with the “exact” values in Fig. This is to be expected.

0. 1. we investigate how the errors in the UMRHA vary with the deformation demands imposed by the ground motion.85.response is much less accurate (Tables 4.1 and 4. However. the errors in responses computed by UMRHA including three “modes” relative to the “exact” response were determined. and 3.75.38 6.88
u (cm)
"Mode" 3
−80 80 0 −80 80 0 −80 0 • 44. just as for elastic systems.1 and 3.44 •
"Mode" 2
r1
0 • 48.5 × El Centro ground motion: individual “modal” responses and combined response from UMRHA. the errors in story drifts are larger compared to floor displacements.7.
(a) Roof Displacement (b) Top Story Drift 80 12 0 • 3.53
∆r3 (cm)
12 0 • 2. 0.2) and 1. 2. sec 25 30 • 48.1
−12 12
∆r (cm)
u (cm)
UMRHA 3 "Modes"
0 • 7.3 11. For this purpose the UMRHA and exact analyses were repeated for ground motions of varying intensity.0.2) if the system yields significantly versus if the system remains within the elastic range (Tables 3.0.5. 0.6 5 10 15 20 Time.62 5.5 (Tables 4.1 and 4.2).7 •
−80 80
∆r1 (cm)
u (cm)
"Mode" 1
−12
r2
0
∆r2 (cm)
12 0
u (cm)
−80 80
−12
r3
0
• 2. Response histories of roof displacement and top-story drift due to 1.24 •
NL−RHA
r
−12 12
∆r (cm)
u (cm)
0 5 10 15 20 Time. 4. sec 25 30
r
−12 0
Fig.25 (Tables 3. and total response from NL-RHA
39
.25. For each excitation. in particular. 0.1 and 3. defined as the El Centro ground motion multiplied by 0. 1.5. for a fixed number of “modes” included. Next. the degree to which the system deforms beyond its elastic limit. recall that the computed errors have been presented earlier for ground motion multipliers 0.2).

with the peak displacement of each “modal” SDF system noted for each ground motion multiplier. in each story drift (Fig. and the error envelope for each case. 4.5 × El Centro ground motion
(a) Floor Displacements 9th 8th 7th 6th 9th 8th 7th 6th
(b) Story Drift Ratios
Floor
4th 3rd 2nd 1st UMRHA 1 "Mode" 2 "Modes" 3 "Modes"
Floor
5th
5th 4th 3rd 2nd 1st UMRHA 1 "Mode" 2 "Modes" 3 "Modes"
Ground −60
−40
−20
0 Error (%)
20
40
60
Ground −60
−40
−20
0 Error (%)
20
40
60
Fig.5
Fig. or three “modes” for 1. For this purpose. indicated by a ground motion multiplier. the pushover curves using force distributions s* [Eq. (3.11. Heightwise variation of floor displacements and story drift ratios from UMRHA and NL-RHA for 1. 4.8.5 Story Drift Ratio (%)
2
2.(a) Floor Displacements 9th 8th 7th 6th 9th 8th 7th 6th NL−RHA UMRHA 1 "Mode" 2 "Modes" 3 "Modes"
(b) Story Drift Ratios
Floor
Floor
5th 4th 3rd 2nd 1st
5th 4th 3rd 2nd 1st NL−RHA UMRHA 1 "Mode" 2 "Modes" 3 "Modes"
Ground 0
0. Shown is the error in each floor displacement (Fig. it will be useful to know the deformation of the system relative to its yield deformation.5 ¥ El Centro ground motion
Figure 4. 4. To interpret these results.10a). two. 4. 4.20)] for the first three n modes of the system are shown in Fig.5 1 1. Two versions of the pushover curve are 40
.9. Heightwise variation of error in floor displacements and story drifts estimated by UMRHA including one.5 Displacement/Height (%)
2
Ground 0
0.5
1 1.10b).10 summarizes the error in UMRHA as a function of ground motion intensity.

Perhaps this explains why the errors are large at this excitation intensity. were noted but are not shown here.10) for each “modal” inelastic SDF system (Figs 4. the ductility factor for the first mode system is only 1. 4. the errors in truncating the higher mode contributions are negligible. First. (2) increase rapidly as the ground motion multiplier increases to 1. 4.7).
41
. 3.4. the first reason mentioned above seems to be the primary source for the errors.53 cm.0. The structure is pushed using the force distribution of Eq. 4. respectively. Figure 4.4).1).75. the values determined by RHA of the nth-mode inelastic SDF system (Fig. UMRHA lacks a rigorous theory and is based on several approximations.11a).5 ¥ the El Centro ground motion.2 Modal Pushover Analysis
The results of modal pushover analysis procedure considering the response due to the first three “modes” was implemented for the selected building subjected to 1. and plastic hinge rotations at the external beam end at each floor level (Table 4. (3) maintain roughly similar values for more intense ground motions. 11.included: the actual curve and its idealized bilinear version.3 cm.01 (Fig.0. Second.11). even though the system remains essentially elastic. The system remains elastic up to ground motion multiplier 0. 4. The location of plastic hinges and their rotations. and 2. 2.5).4) to roof displacements urno = 48.4. Each of these three pushover analyses provides the pushover curve (Fig. and 3 (Fig. and (4) are larger in story drift compared to floor displacements.3).2.7 cm. as mentioned in Section 4. determined from “exact” analyses. (4.10 permits the following observations regarding the accuracy of the UMRHA procedure: the errors (1) are small (less than 5%) for ground motion multipliers up to 0. (3.2. the peak values of displacements at all floors (Table 4. Additional errors are introduced in UMRHA of systems responding beyond the linearly elastic limit for at least two reasons. and. The idealized curve for the first “mode” deviates most from the actual curve near the peak displacement corresponding to ground motion multiplier 1. drifts in all stories (Table 4.20) with n = 1. as mentioned in Section 3. For more intense excitations. the pushover curve for each “mode” is idealized by a bilinear curve in solving Eq.75.6 and 4.

5 18.298
-0.044 1.426 -1.663 0.5 10.772
Error (%) NL 3 1 2 3 RHA “Mode” “Modes” “Modes” “Modes” 1.241 -1.070 1.1 Peak values of floor displacements (as % of building height = 37.9 16.6 2.0 2.0 9.1 3.863 1.350 -0.283 1.983 1.9 31.220 -0.4 1.201 -0.372 1.9 12.9 5.298
0.003 -31.226 -0.371 0.1 1.7 14.5 × El Centro ground motion Floor 1st 2nd 3rd 4th 5th 6th 7th 8th 9th Displacement /Height (%) “Modal” Response Combined (UMRHA)
“Mode” 1 “Mode” 2 “Mode” 3 1 “Mode” 2 “Modes”
-0.4 4.072 1.2 6.7 31.256 -1.2 12.072 -1.1 1.8 1.900 -10.010 0.9 31.7
Table 4.0 -9.698 1.668 -23.4 -1.971 1.256 1.3 6.0 11.130 0.371 -0.8 1.033 0.4 -7.942 1.126 0.8 1.079 0.338 1.5 28.009 -0.3 1.540 0.214 -0.575 -41. and (b) story drifts Table 4.136 1.088 10.293 1.2 22.241 1.0 11.5 28.133 1. Errors in UMRHA as a function of ground motion intensity: (a) floor displacements.820 -0.216 1.8 1.5
1
1.10.877 0.0 7.945 -37.938 -1.200 8.373
-0.6 4.852 1. Noted 60 50 40
(b) Story Drifts
Error Envelope Error for Story No.484 0.057 -0.1 8.003 0.914
2.214 0.982 9.727
1.096 0.722 0.202 11.5
3
0 0
0.370 -0.8 14.844 -25.5 9.317 0.5 × El Centro ground motion Story 1st 2nd 3rd 4th 5th 6th 7th 8th 9th Modal Response
“Mode” 1 “Mode” 2
Drift Ratio (%) Combined (UMRHA)
1 “Mode” 2 “Modes”
“Mode” 3
-1.333 0.256 1.366 -0.135 9.068
0.1 0.291 0.856 2.473 -22.2 4.410 -1.938 1.3
42
.9 12.407 -10.472 1.490 -1.379 1.169 0.526 -0.478 0.513 0.018 0.663 -0.3 25.220 0.763 -15.914
-0.554 1.5 9.260 -15.201 1.071 -0.751 1.154 0.3 0.616 -0.8 0.942 -0.2 Peak values of story drift ratios (as % of story height) from UMRHA for 1.2 1.366 0.187 -0.2 1.338 -1. Noted
7 6
Error (%)
30
1
Error (%)
30
5 8
20
6 5 2
20
2 4 9
9
10 0 0
10
3 8 7
3 1 4
0.863 0.235 -0.8 1.806 -0.376 -1.676 0.376 1.120 1.5 2 GM Multiplier
2.819 2. 4.5 2 GM Multiplier
2.055 -0.138 1.372 -1.121 -0.065 0.5 -3.6 1.495 1.513 -0.430 1.0 1.811 1.315
-0.490 1.(a) Floor Displacements 60 50 40 Error Envelope Error for Floor No.707 1.3 8.410 1.049 -0.820 -19.248
Error (%) NL 3 1 2 3 RHA “Mode” “Modes” “Modes” “Modes” 0.5
1
1.5
3
Fig.14 m) from UMRHA for 1.201 -1.071 0.806 0.104 0.

13 and Tables 4. but was more successful when the excitation was more intense (results are not included). Fortuitously. The first “mode” alone is inadequate. Pushover analysis seems to be inherently limited in computing accurately hinge plastic rotations.1 and 4. but the third “mode” contributions do not seem especially important (Fig. One “mode” pushover analysis was unable to identify the plastic hinges in the upper stories where higher mode contributions to response are known to be more significant. significantly smaller than in UMRHA (compare Fig. the hinges identified in beams at the 6th floor were at variance with the “exact” results.12 and 4.18).4. However. this error is not especially significant because the hinge plastic rotation is very small. overestimated by up to 18% in the middle stories. and three “modes.3 and 4.12 and Tables 4.” two “modes. The drifts are underestimated by up to 13% in the lower stories.3 and 4.” and nonlinear RHA.3 and 4.5 present estimates of the combined response according to Eq.2). 4.3 and 4. MPA including three “modes” underestimates the displacements of the lower floors by up to 8% and overestimates the upper floor displacements by up to 14%.12 and Tables 4. the third “mode” does not contribute because this SDF system remains elastic (Fig. (3. the additional errors due to the approximation inherent in modal combination rules tend to cancel out the errors due to the various approximation contained in the UMRHA.Figures 4. in general.14.4 with Tables 4.” and three “modes. and are within a few percent of the exact values for the upper stories. significant improvement is achieved by including response contributions due to the second “mode”. especially in estimating the story drifts (Fig. 4.11c). 4. Observe that the primary contributor to plastic hinge rotations in the lower stories is the first “mode” and the second “mode” in the upper stories.3 through 4. however. 4.” respectively. considering one. the results were not always accurate. for two or three modes included. the errors in the modal pushover results are. and the errors in these estimates relative to the exact response from nonlinear RHA.4). Furthermore.13 with Fig. 4. The locations of plastic hinges shown in Fig. even if three “modes” are included (Fig.14. Obviously.5).13 and Tables 4. As shown in Figs. were determined by four analyses: MPA considering one “mode. The errors are especially large in the hinge plastic rotations estimated by the MPA procedure.9 and Tables 4. 4. MPA failed to identify the plastic hinges at the column bases in Fig. The second “mode” was necessary to identify hinges in the upper stories.4). 4. two.13 and Table 4. 4. 4. For example. although the error is recorded as 100% if MPA estimates zero rotation when the nonlinear RHA computes a non-zero value. 43
.

5 1 0.85 0. V = 7616 kN.5
2
3
4000 2000 0 0
Actual Idealized
0.85 0.0. 0.0. and 3. 1.9 cm.19 y by
3 2 1. V = 5210 kN. α = 0.5 0. α = 0.85
Base Shear (kN)
8000 6000 4000 2000 0 0
0.25.75 0.5.5.13 y by
Base Shear (kN)
8000 6000
1
1.14 y by
Base Shear (kN)
8000 6000 4000
1.25
5
10 15 20 Roof Displacement (cm)
25
(c) "Mode" 3 Pushover Curve 12000 10000 u = 4.25
1
2
4 6 Roof Displacement (cm)
8
10
Fig.5 0. α = 0. 1. 2. “Modal” pushover curves with peak roof displacements identified for 0.5 2
3
Actual Idealized
2000 0 0
0.2 cm.6 cm. 4.(a) "Mode" 1 Pushover Curve 12000 10000 u = 36.11. V = 4952 kN.0 × El Centro ground motion
44
.75 0.25 0.5 0. 0.75.75
Actual Idealized
20 40 60 Roof Displacement (cm) (b) "Mode" 2 Pushover Curve
80
12000 10000 u = 9.

In this case.4 also apply to MPA. Shown is the error in each floor displacement (Fig. UMRHA is essentially exact.10 and 4.15a). However. 4.5 × El Centro ground motion.15 summarizes the error in MPA considering three “modes” as a function of ground motion intensity. indicated by a ground motion multiplier.5 Story Drift Ratio (%)
2
2.5 Displacement/Height (%)
2
Ground 0
0.15b).(a) Floor Displacements 9th 8th 7th 6th 9th 8th 7th 6th NL−RHA MPA 1 "Mode" 2 "Modes" 3 "Modes"
(b) Story Drift Ratios
Floor
Floor
5th 4th 3rd 2nd 1st
5th 4th 3rd 2nd 1st NL−RHA MPA 1 "Mode" 2 "Modes" 3 "Modes"
Ground 0
0. While various sources of errors in UMRHA. The fact that MPA is able to estimate the response of buildings responding well into the inelastic range to a similar degree of accuracy indicates that this procedure is accurate enough for practical application in building retrofit and design.5
1 1. an observation with practical implications.5
Fig. identified in Section 3.3. 4. shading indicates errors in MPA including three “modes”
Figure 4. implying that the modal combination errors contained in these procedures are acceptable. the errors in MPA were larger for ground motion multipliers less than 0.5 1 1. as discussed in Sections 3.4.3 for elastic systems (or weak ground motions). implying excitations weak enough to limit the response in the elastic range of the structure. 4. whereas MPA contains errors inherent in modal combination rules. now standard in engineering practice. The errors are only weakly dependent on ground motion intensity (Fig.0. 4.2 and 3. As mentioned in Section 3.15) for ground multipliers larger than 1.75. 4. the errors in MPA were fortuitously smaller than in UMRHA (compare Figs.15).12 Heightwise variation of floor displacements and story drift ratios from MPA and NL-RHA for 1.4. implying excitations intense enough to cause significant yielding of the structure. each story drift (Fig.
45
. the MPA procedure is equivalent to the RSA procedure. and the error envelope for each case.

13.(a) Floor Displacements 9th 8th 7th 6th
Floor
5th 4th 3rd 2nd 1st MPA 1 "Mode" 2 "Modes" 3 "Modes"
Ground −60
−40
−20
0 Error (%)
20
40
60
(b) Story Drift Ratios 9th 8th 7th 6th
Floor
5th 4th 3rd 2nd 1st MPA 1 "Mode" 2 "Modes" 3 "Modes"
Ground −60
−40
−20
0 Error (%)
20
40
60
(c) Hinge Plastic Rotations 9th 8th 7th 6th
Floor
5th 4th 3rd 2nd 1st MPA 1 "Mode" 2 "Modes" 3 "Modes"
Ground −120
−80
−40
0 Error (%)
40
80
120
Fig. and hinge plastic rotations estimated by MPA including one. and three “modes” for 1. 4. two.5 ¥ El Centro ground motion
46
. story drifts. Errors in floor displacements.

315
0.125 -1.130 0.222 0.10E-02 9.8 1.60E-03 2.5 × El Centro ground motion Hinge Plastic Rotation (rad) “Modal” Response Combined (MPA)
“Mode” 1
7.0 15.00E+00
1 “Mode”
7.8 -4.197 -0.6 7.012 1.9 -4.9 -100.19E-10 3.00E+00 0.007 1.9 -100.5 1.105 0.2 0.8 7.0
3 “Modes”
-32.683 1.5 -6.116 1.00E+00 0.72E-03 7.00E+00 0.3 -3.1 1.3 Peak values of floor displacements (as % of building height = 37.8 -6.429 0.220 1.0 -5.2 -4.3 11.895 1.88E-03 0.76E-03 4.244 0.694 1.6 13.60E-04 7.37E-03 1.2 6.728
1.202 8.00E+00 0.6 -7.2 -3.7 -12.2 -100.1 62.55E-03 3.7 1.176 0.5 10.00E+00 0.00E+00 0.756 0.5 × El Centro ground motion Story 1st 2nd 3rd 4th 5th 6th 7th 8th 9th “Modal” Response
“Mode” 1 “Mode” 2
Drift Ratio (%) Combined (MPA)
1 “Mode” 2 “Modes” 3 “Modes”
Error (%) NL 1 2 3 RHA “Mode” “Modes” “Modes” 1.820 -7.399 0.304 -1.22E-10
NL RHA
1.1 46.980 -0.066 -0.6 13.844 -7.304 1.00E+00 0.8 -6.3 13.705 1.611 0.652 1.14 m) from MPA for 1.945 -49.018 -0.37E-04
Error (%) “Modal” Response
1 “Mode”
-32.003 -16.8
“Mode” 3
-1.00E+00
Floor
1st 2nd 3rd 4th 5th 6th 7th 8th 9th
“Mode” 2
0.311 0.37E-03 1.88E-03 0.0 -50.4 1.3 1.00E+00 0.5
Table 4.00E+00 0.00E+00 0.518 1.00E+00
3 “Modes”
7.266 -0.116 1.015 0.9 2.02E-03 3.8 -29.1 46.9 1.426 15.18E-03 7.980 0.1 62.057 0.053 1.737
1.00E+00 0.6 1.666
Table 4.371 -0.9 0.6 -8.581 0.756 0.260 -14.1 62.895 1.36E-03 6.05E-03 2.5 7.068
0.02E-03 0.007 1.687 0.88E-03 1.053 -1.733 1.0
47
.911 0.233 1.514 -1.503 -1.2 11.222 0.135 -7.331 1.0 1.26E-04 9.071 0.101 -0.9 7.72E-03 7.1 -8.018 -0.00E+00 0.738 1.00E+00 0.5 2.0 -100.435 0.305
-0.4 Peak values of story drift ratios (as % of story height) from MPA for 1.752 1.942 6.154 0.414 1.1 18.9 1.336
Error (%) NL 3 1 2 3 RHA “Mode” “Modes” “Modes” “Modes” 0.088 12.55E-03 3.8 -29.5 Peak values of hinge plastic rotations (radians) from MPA for 1.2 0.351 -0.2 -100.76E-03 4.668 -13.6 -9.209 1.6 -44.5 10.640 1.399 0.503 1.37E-03 1.009 0.00E+00 3.Table 4.76E-03 4.36E-03 6.5 2.705 -1.00E+00 0.0 -100.00E+00 0.250 0.910 1.156 -0.478 0.2 -100.517 1.72E-03 7.116 1.50E-10 3.298
-0.8 17.614 0.8 -12.233 1.5 × El Centro ground motion Displacement /Height (%) “Modal” Response Combined (MPA)
“Mode” 1 “Mode” 2 “Mode” 3 1 “Mode” 2 “Modes”
Floor 1st 2nd 3rd 4th 5th 6th 7th 8th 9th
0.763 -14.667 1.594 -1.8 1.049 -0.0
2 “Modes”
-32.6 -44.641 1.0 -50.015 0.407 -27.575 -53.018 0.071 -0.220 1.516 0.168 -0.2 1.0 -100.3 -3.640 -1.190 -0.804 1.02E-03 0.667 -1.00E+00
2 “Modes”
7.033 -0.53E-03 7.527 -0.498
1.00E+00 0.99E-03 6.473 -15.298
0.372 0.118 0.9 5.200 8.737
0.00E+00 0.00E+00 0.37E-03 1.745 1.1 13.781 0.8 0.4 -8.879 1.8 -29.259 1.055 0.76E-03 4.781 0.72E-03 7.982 13.900 -0.066 -0.36E-03 6.5 2.55E-03 3.581 0.338 1.36E-03 6.00E+00 0.7 1.00E+00
“Mode” 3
0.02E-03 3.1 46.

Figures 4.7).” and the errors in these estimates relative to the exact response from nonlinear RHA.18 and 48
.5).6 and 4.18) considering one. and three “modes. and plastic hinge rotations at selected external beam end at each floor level (Table 4. 4.4. Obviously gravity load effects will influence the seismic demands due to the first “mode. (3.17 and Tables 4. Also presented at these roof displacements are the displacements at all floors (Table 4. 4.3
Modal Pushover Analysis with Gravity Loads
To evaluate the accuracy of the dynamic response of the system. They are now included in the pushover analysis of the structure for the first “mode” only. the structure is pushed using the force distribution of Eq.11a).53 cm.7). The response contributions of the second and third modes are the same as before (Tables 4. 4.” however. (3.7). The story drifts are underestimated by up to 7% in the lower stories. The errors are especially large in the plastic hinge rotations estimated by the MPA procedure even if three “modes” are included (Fig.8 present estimates of the combined response according to Eq. two. 4. MPA including two “modes” underestimates the displacements of lower floors by up to 6% and overestimates the upper floor displacements by up to 22%.20) with n = 1 to a target roof displacement. as are the roof displacements ur 2o = 11. (3.” but not the contributions of higher “modes.6 through 4.0 cm. resulting in the pushover curve shown in Fig.4) to roof displacement ur1o = 52.3 .17 and 4.4.6 and 4. 4.18 and Tables 4. and the structure is pushed using the force distribution of Eq.20) with n = 1 (Fig. drifts in all stories (Table 4.16a—which is slightly different than the one excluding gravity loads (Fig. Static analysis of the structure for gravity loads provides the initial state—forces and deformations—of the structure. but it is inadequate in estimating the story drifts (Fig. Starting with its initial state under gravity loads. 4.8).18 and Tables 4.4. the third “mode” contributions do not seem especially important (Fig. 3.6 and 4.7. Significant improvement is achieved by including response contributions due to the second “mode.7 cm and ur 3o = 2. the results presented so far did not include gravity load effects. 4. As shown in Fig.17 and Tables 4. and overestimated by up to 29% in the middle stories and by up to 13% in the upper stories. The pushover curves for the second and third modes included in Fig. Results of such analyses for the selected building subject to 1.6). The first “mode” alone provides adequate estimates of floor displacements.” these effects will modify the combined modal response as well as the results of nonlinear RHA.5 × El Centro ground motion are presented below.16 are unchanged.

and three “modes” and by NL-RHA for 1.5
1
1.15.5 2 GM Multiplier
2. 4. Noted 60 50 40
(b) Story Drifts Error Envelope Error for Story No.14. 2−"Modes" •• •• •• •• •• •• •• •• •• •• •• •• •• •• • • • • •• • ••
(c) 3−"Modes" • • • • • • • •• •• • •• •• •• •• •• •• •• •• •• •• •• •• •• •• •• •• •• •• • • • • •• • ••
• • • • • • • • •
(d) Nonlinear RHA •• •• • •• •• •• •• • •• •• •• •• •• •• •• • •• •• • •• •• •• •• • • • • • • •• •• ••
Fig. and (b) story drifts
49
.5 2 GM Multiplier
2. Locations of plastic hinges determined by MPA considering one. Errors in MPA as a function of ground motion intensity: (a) floor displacements. Most pushover analysis procedures do not seem to compute to acceptable accuracy plastic hinge rotations.5
3
Fig. 4.
(a) MPA. 1−"Mode" • • • • • • • • •• •• •• •• •• •• •• •• •• •• •• •• •• •• • • •• • •• • • • • • •• •• • •• •• •• •• (b) MPA. two.8).5
3
0.5
1
1.5 × El Centro ground motion
(a) Floor Displacements 60 50 40 Error Envelope Error for Floor No.Table 4. Noted
Error (%)
30 20 10 0 0
2 8 3 4 1 9
Error (%)
30 20
5
4
6
5
7 2
9
10 0 0
1 8
7 3
6
0.

19 were determined by four analyses: MPA considering one “mode.The locations of plastic hinges shown in Fig. and the errors are only weakly dependent on ground motion intensity.20a). 4. MPA provides response values accurate enough for practical application in building retrofit or design. Shown is the error in each floor displacement (Fig. These errors are only slightly larger than those in Fig. Figure 4. this procedure is able to predict plastic hinge locations essentially consistent with nonlinear RHA. 4. and the error envelope for each case. The second “mode” is necessary to identify hinges in the upper stories.15.20 summarizes the error in MPA considering three “modes” as a function of ground motion intensity. and nonlinear RHA. One “mode” pushover analysis is unable to identify the plastic hinges in the upper stories where higher mode contributions to response are known to be more significant. With two modes included in MPA. indicated by a ground motion multiplier.
50
.20b). 4.” two “modes”. 4. each story drift (Fig. excluding gravity load effects.

16. 0. and 3. 1.0. “Modal” pushover curves with gravity loads included.(a) "Mode" 1 Pushover Curve 12000 10000 u = 36.0.0 × El Centro ground motion
51
.75 0.9 cm. 0. α = 0.25
5
10 15 20 Roof Displacement (cm)
25
(c) "Mode" 3 Pushover Curve 12000 10000 u = 4.75
Actual Idealized
20 40 60 Roof Displacement (cm)
80
(b) "Mode" 2 Pushover Curve 12000 10000 u = 9.19 y by
3 2 1. V = 4952 kN.5 1 0.5 0. α = 0.14 y by
Base Shear (kN)
8000 6000 4000
1.6 cm. V = 7433 kN. noted are peak values of roof displacement for 0.25.3 cm.25
1
2
4 6 Roof Displacement (cm)
8
10
Fig.5 0.25 0.85 0.5 2
3
Actual Idealized
2000 0 0
0.85 0.75 0.85
Base Shear (kN)
8000 6000 4000 2000 0 0
0.50.75.5 0.85.13 y by
Base Shear (kN)
8000 6000
1
1. 4. V = 5210 kN. α = 0. 0. 2.5
2
3
4000 2000 0 0
Actual Idealized
0.

two. and hinge plastic rotations estimated by MPA including one.17.(a) Floor Displacements 9th 8th 7th 6th 9th 8th 7th 6th NL−RHA MPA 1 "Mode" 2 "Modes" 3 "Modes"
(b) Story Drifts
Floor
Floor
5th 4th 3rd 2nd 1st
5th 4th 3rd 2nd 1st NL−RHA MPA 1 "Mode" 2 "Modes" 3 "Modes"
Ground 0
0. and
52
. 4. 4. Heightwise variation of floor displacements and story drift ratios from MPA and NL-RHA for 1. Errors in floor displacements.18.5 Displacement/Height (%)
2
Ground 0
0.5 × El Centro ground motion. gravity loads included.5
1 1.5 Story Drift Ratio (%)
2
2. story drifts.5 1 1.5
Fig. shading indicates errors in MPA including three “modes”
(a) Floor Displacements 9th 8th 7th 6th 9th 8th 7th 6th (b) Story Drift Ratios
Floor
4th 3rd 2nd 1st MPA 1 "Mode" 2 "Modes" 3 "Modes"
Floor
5th
5th 4th 3rd 2nd 1st MPA 1 "Mode" 2 "Modes" 3 "Modes"
Ground −60
−40
−20
0 Error (%) 9th 8th 7th 6th
20
40
60
Ground −60
−40
−20
0 Error (%)
20
40
60
(c) Hinge Plastic Rotations
Floor
5th 4th 3rd 2nd 1st MPA 1 "Mode" 2 "Modes" 3 "Modes"
Ground −120
−80
−40
0 Error (%)
40
80
120
Fig.

5
1
1. 3−"Modes"
(d) Nonlinear RHA
• • • • • • •
•• •• •• •• •• •• ••
•• •• •• •• •• •• ••
•• •• •• •• •• •• ••
• • • • •• •• ••
• • • • • • • •
•• •• • •• •• •• •• •
• •• •• • •• •• •• •• •
• •• •• • •• •• •• •• •
• • • • • •• •• ••
Fig.5 2 GM Multiplier
2. gravity loads included
(a) Floor Displacements 60 50 40 Error Envelope Error for Floor No. 4.20.(a) MPA. gravity loads included
53
. 4.5 2 GM Multiplier
2. 1−"Mode" • • • • • • • •• •• •• •• •• •• •• •• •• •• •• •• •• •• •• • • •• •• •• • • • • • •• •• •• •• •• •• ••
(b) MPA. and three “modes” and by NL-RHA for 1. Locations of plastic hinges determined by MPA considering one.5
1
1. and (b) story drifts. Errors in MPA as a function of ground motion intensity: (a) floor displacements.5 × El Centro ground motion. Noted
Error (%)
30 20
6 8 1 5 2 4 3 9 7
Error (%)
30
4 5
20
6
9 7 1
10 0 0
10 0 0
8 2 3
0. 2−"Modes" •• •• •• •• •• •• •• •• •• •• •• •• •• •• • • • • •• •• ••
(c) MPA. two.5
3
Fig. Noted 60 50 40 (a) Story Drifts Error Envelope Error for Story No.19.5
3
0.

gravity loads included Hinge Plastic Rotation (rad) Floor “Modal” Response
“Mode” 1 1st 2nd 3rd 4th 5th 6th 7th 8th 9th
8.983 1.114 -1.466 0.4 1.11E-03 9.00E+00 0.78E-03 1.00E+00
3 “Modes”
8.8 9.00E+00 0.213 1.850 1.933 1.190 -0.23E-02 1.821 -1.9 -4.156 -0.3 -100.516 0.0 -100.057 0.8 -32.1 1.5 × El Centro ground motion.3 -22.257 0.6 1.2 9.3 9.3 -100.728
1.927 1.637 0.921 1.3 13.6 0.836 -0.75E-03 0.23E-03 3.4 20.854 0.454 1.037 -0.102 1.2 1.04E-02 8.429 1.8 -32.11E-03 9.3 1.9 -3.17E-03 9.130 0.830 -12.6 2.7 -2.6 Peak values of floor displacements (as % of building height = 37.098 20.00E+00 0.00E-03 5.8 1.0 -30.530 1.603 1.2 4.14 m) from MPA for 1.00E+00 0.00E+00 0.414 28.102 1.330 1.213 1.783 1.064 -10.213 1.037 0.7 Peak values of story drift ratios (as % of story height) from MPA for 1.237 0.2 12.00E+00 0.2 -2.5
54
.2 2.429 -1.530 1.23E-03 3.00E+00 0.88E-03 0.860 1.176 0.13E-03 5.637 0.00E-03 5.0 -30.938 1.998 1.594 -1.5 21.436 1.55E-03 3.434 0.371 -0.8
Table 4.26E-03 3.315
0.1 4.00E+00 0.5 1.071 -0.5
3 “Modes”
-32.00E+00 0.00E+00 0.00E+00 0.00E+00 0.372 0.686 -7.055 0.55E-03 3.1 21.35E-03 8.9 31.154 0.319 1.0 37.310 1.128 -1.305
-0.754
1.908 -1.490 -11.00E+00
“Mode” 3
0.35E-03 8.3 -100.8 0.00E-03 5.237 0.197 -0.00E+00
“Mode” 2
0.603 -1.2 9.168 -0.399
-0.114 1.00E+00 0.2 0.953 15. gravity loads included Floor 1st 2nd 3rd 4th 5th 6th 7th 8th 9th Displacement /Height (%) “Modal” Response Combined (MPA)
“Mode” 1 “Mode” 2 “Mode” 3 1 “Mode” 2 “Modes”
0.667 0.125 -1.19E-04 5.5 × El Centro ground motion.913 7.478 0.018 -0.996 -0.263 1.7 -2.049 -0.23E-03 0.5 10.983 1.35E-03 8.00E+00 0.908 1.744 1.068
0.2 -3.998 21.7 16.5 × El Centro ground motion.19E-03 1.9 -6.813 1.4 -4.665 0.2 4.0 16.311 0.071 0.6 2.19E-03 1.11E-03 9.821 1.00E+00
Combined (MPA)
1 “Mode”
8.00E+00 0.8 0.2 4.2 -0.2 21.2 16.Table 4.515 -50.55E-03 3.687 0.2 9.0 -100.35E-10 3.8 Peak values of hinge plastic rotations (radians) from MPA for 1.996 0.330 1.353 -23.0 37.6 19.0
2 “Modes”
-32.5 1.319 1.5
“Mode” 3
-1.00E-10
NL RHA
1.888 1.00E+00
2 “Mode”
8.5 -5.5 21.7 1.434 0.00E+00 0.2 1.23E-03 0.351 -0.530 1.00E-03 5.199 16.105 0.19E-03 1.1 13.066 -0.11E-03 9.214 0.207 18.19E-03 1.00E+00 0.00E+00 0.015 0.831 0.831 0.507
1.461 0.04E-10 3.673
Table 4.101 -0.00E+00 3.399
0.00E+00 0.4 1.4 -6.107 1.998 0.0 37.8 28.527 -0.270 -12.754
0.88E-03 0.854 0.434
Error (%) NL 3 1 2 3 RHA “Mode” “Modes” “Modes” “Modes” 0.009 0.3 -22.3 -22.877 -46.822 1.109 0.7 4.88E-03 1.35E-03 8.5 0. gravity loads included Story 1st 2nd 3rd 4th 5th 6th 7th 8th 9th “Modal” Response
“Mode” 1 “Mode” 2
Drift Ratio (%) Combined (MPA)
1 “Mode” 2 “Modes” 3 “Modes”
Error (%) NL 1 2 3 RHA “Mode” “Modes” “Modes” 1.263 0.514 -1.266 -0.7 19.850 -1.4 -4.858 2.066 -0.00E+00
Error (%) “Modal” Response
1 “Mode”
-32.4 26.033 -0.

plastic hinge rotations.” and nonlinear RHA. computed at the target displacement represent the earthquakeinduced demands on the structure.1 FEMA-273 PUSHOVER ANALYSIS
In this investigation we focus on one step in the nonlinear static procedure in the FEMA-273 document [Building Seismic Safety Council. joint rotations. etc. “Uniform” distribution: s* = m j (where the floor number j = 1.2 COMPARATIVE EVALUATION
Compared in this section are the earthquake-induced demands for the selected building determined by five analyses: pushover analysis using the three force distributions in FEMA-273. 2… N ). Specified in FEMA-273 are three distributions for lateral forces: 1. 1997] The pushover curve. and 3.5 sec . and the exponent k = 1 for fundament period T1 £ 0. gravity load effects were included in all 55
. and varies linearly in between. 5. a plot of base shear versus roof displacement. Comparison of Modal and FEMA Pushover Analyses
5. MPA considering three “modes. k = 2 for T1 ≥ 2. story drifts.5 sec . j 2. is determined by nonlinear static analysis of the structure subjected to lateral forces with invariant distribution over height but gradually increasing values until a target value of roof displacement is reached. Equivalent lateral force (ELF) distribution: s* = m j h k where h j is the height of the j j jth floor above the base.5. assumed to be linearly elastic. The floor displacements. SRSS distribution: s* is defined by the lateral forces back-calculated from the story shears determined by response spectrum analysis of the structure..

11 0.
0. (b) ELF.119 0.3a and Table 5.3.1.11 0. and the locations of all plastic hinges in Fig. Figures 5.2. The pushover curves are given in Fig.0981 0.1.5 times the El Centro ground motion. and (c) SRSS
56
.1.3b and Table 5.0702 0.0466 0. wherein the first two are obvious and the third is determined from response spectrum analysis of the building (Appendix C).21 0.0913 0. 5.analyses.4. 5. and the displacements of upper floors are overestimated by 9-22%.0381 0.042 0. and Table 5.3. The first four floor displacements are within 5% of the “exact” value. The “uniform” distribution overestimates all floor displacements by 17-28%.4 and Tables 5.0896
0.0197 0.1 through 5.11 0.11 0.1 demonstrate that the displacement demands are underestimated by the ELF and SRSS force distributions by 12 to 30 % at the lower six floors of the building.3a and 5.2.11 0. pushover analyses are implemented for a target roof displacement of 52.062 0.0446 0.0654 0. 5. Force distributions in FEMA-273: (a) “uniform”. The three FEMA force distributions are presented in Fig.0 cm. plastic hinge rotation demands in Table 5. 5. 5.177 0.165 0.4a.11 0. the floor displacement demands in Fig. with the errors being larger for the SRSS distribution.281 0.367
(a) Uniform
(b) ELF
(c) SRSS
Fig.4.11 0.112
0. Using each of these force distributions.126 0. The MPA procedure is more accurate than all the FEMA force distributions. Also included in these presentations are the MPA results considering three “modes” and the “exact” demands from nonlinear RHA. 5. the story drift demands in Fig. The errors in the FEMA and MPA estimates of seismic demands relative to the “exact” demands are presented in Fig. the value determined from RHA of the first-mode inelastic SDF system for 1. both presented in Section 4.00719 0. 5.3.

gravity loads are included
57
.24 y by
Base Shear (kN)
8000 6000 4000 2000 0 0
Actual Idealized
10
20 30 40 Roof Displacement (cm)
50
60
Fig. V = 7456 kN.(a) Uniform Distribution 12000 10000 u = 33 cm. (b) ELF. α = 0.14 y by
Base Shear (kN)
8000 6000 4000 2000 0 0
Actual Idealized
10
20 30 40 Roof Displacement (cm) (b) ELF Distribution
50
60
12000 10000 u = 38.8 cm. 5.3 cm. V = 8530 kN.18 y by
Base Shear (kN)
8000 6000 4000 2000 0 0
Actual Idealized
10
20 30 40 Roof Displacement (cm) (c) SRSS Distribution
50
60
12000 10000 u = 39. and (c) SRSS.2. α = 0. Pushover curves using three force distributions in FEMA-273: (a) “uniform”. α = 0. V = 6897 kN.

(The 100% error in a 6th floor hinge is ignored as it simply represents that the MPA estimated zero rotation.2 demonstrate that the story drift demands are greatly underestimated by all the FEMA force distributions. with story drifts under estimated by. whereas nonlinear RHA computed an insignificantly small value. For the ELF distribution.
(a) Floor Displacements 9th 8th 7th 6th 9th 8th 7th 6th
(b) Story Drift Ratios
Floor
4th 3rd 2nd 1st NL−RHA MPA.3 demonstrate that the hinge plastic rotations estimated by all three FEMA force distributions contain unacceptably large errors. 3 "Modes" FEMA Uniform ELF SRSS 0.5 1 1.5
Ground 0
Ground 0
Fig. 5. reaching 35%. but it is still inaccurate. errors are largest in the upper stories.4c and Table 5. larger errors are noted in the upper and lower stories. and overestimated by no more than 32%.Figures 5.5 Story Drift Ratio (%) 2 2. In contrast. 3 "Modes" FEMA Uniform ELF SRSS 0. This seems to be an inherent limitation of pushover analysis. at most.5 Displacement/Height (%) 2
Floor
5th
5th 4th 3rd 2nd 1st NL−RHA MPA. such as hinge plastic rotations.5 1 1. Figure 5. reaching 31%. and Table 5. MPA including three “modes.4b.” and NLRHA. 7%. gravity loads included
58
. For the SRSS distribution.3. the errors are largest in the lower stories.) The pushover analysis procedures considered seem incapable of computing accurately local response quantities.3b and 5. the MPA procedure is more accurate than all the FEMA force distributions. Heightwise variation of floor displacements and story drift ratios estimated using FEMA-273 force distributions. reaching 64%. Modal pushover analysis procedure gives estimates better than all the FEMA force distributions. with errors reaching 37% in this example. For the uniform distribution.

4. 5. The locations of plastic hinges shown in Fig. and the three FEMA analyses. it appears that structural performance evaluation should be based on story drifts that are known to be closely related to damage and can be estimated to a higher degree of accuracy by pushover analyses.” nonlinear RHA (“exact”). The locations of the plastic hinges are not identified correctly by the FEMA force distributions. gravity loads included
The structural engineering profession is now comparing these hinge plastic rotations against rotation limits established in FEMA-273 to judge structural component performance. story drifts. they are not good indicators of damage. While pushover estimates for floor displacements are more accurate. and the ELF distribution fails 59
. the SRSS distribution fails to identify yielding of beams in the middle floors. the “uniform” distribution fails to identify yielding of the beams above the fourth floor.5 were determined by five analyses: MPA considering the three “modes. 5. Based on the results presented here.(a) Floor Displacements 9th 8th 7th 6th MPA FEMA Uniform ELF SRSS 9th 8th 7th 6th
(b) Story Drift Ratios MPA FEMA Uniform ELF SRSS
Floor
4th 3rd 2nd 1st
Floor
−40 −20 0 Error(%) 9th 8th 7th 6th 20 40 60
5th
5th 4th 3rd 2nd 1st
Ground −60
Ground −80
−60
−40
−20
0 20 Error (%)
40
60
80
(c) Hinge Plastic Rotations MPA FEMA Uniform ELF SRSS
Floor
5th 4th 3rd 2nd 1st
Ground −200 −150 −100 −50
0 50 Error (%)
100
150
200
Fig. Errors in floor displacements. and hinge plastic rotations estimated using FEMA-273 force distributions and MPA (including three “modes”).

but fails to predict yielding in a few locations. The MPA procedure identifies yielding in most locations predicted by “exact” analysis.6 and 5.to identify yielding in some locations. Figures 5. and the error envelope for each case.
60
.7 summarize the error in FEMA analyses relative to the “exact” demands as a function of ground motion intensity indicated by a ground motion multiplier. Shown is the error in each floor displacement and each story drift.” The MPA procedure provides estimates of earthquake demands that are significantly more accurate than all FEMA273 analyses. especially in estimating story drifts. Included for comparison is the error in MPA with three “modes. The MPA procedure is superior to FEMA-273 analyses over the entire range of ground motion intensities considered.

84 1.00E+00 2.10E-02 7.6 -17.11E-03 9.65E-03 7.50E-03 0.35E-10 3.19E-04 5.2 4.984 1.927 1.2 6.52E-04 1.7 -29.098 1.154 1.4 16.061 1.015 0.5 -27.2 -12.860 1.45E-03 3.9 0.708 0.19E-03 1.1 10.839 0.372 1.00E+00 0.6 14.344 0.00E+00 0.736 0.1 -3.04E-02 8.5 26.7 19.1.4 0.673
Error (%)
NL RHA 1.5 15.00E+00 0.6 22.9 -77.5 -29.623 1.353 1.5 -3.560 1.8
Floor 1st 2nd 3rd 4th 5th 6th 7th 8th 9th
Uniform 0.0 -59.94E-03 2.1 163.270 0.953 0.436
Error (%)
NL Uniform RHA 0.7 MPA -2.064 1.2 -100. Peak values of floor displacements (as % of building height = 37.4 -12.88E-03 1.007 1.809 0.6 -11.3 -14.3 23.998 1.5 22.4 MPA -2.7 21.14 m) from FEMA force distributions and MPA.1 0.5 -33.417 1.314 1.724 0.3 29.8 2.00E+00
NL RHA
1.6 21.5
61
.318 2.4 -55.58E-04 6.858 1.2 16.2 -4.262 1.414 1.562 1.5 -33.00E+00 0.530 1.78E-03 1.341 1.0 -30.263 0.3 31.487 0.992 1.367 1.667 0.7 -15.836 0.524 0.13E-03 5.4 -50.399
-27.16E-03 0.7 -60.3 -100.3 Peak values of hinge plastic rotations (radians) from FEMA force distributions and MPA Hinge Plastic Rotation (rad) MPA FEMA ELF SRSS
4.067 0.789 0.00E+00 0.7
Table 5.2 9.23E-03 3.8 -2.875 0.128 1.547
-27.26E-03 3.00E+00 4.0 -100.0
Table 5.9 16.462 1.566 1.611 0.0 -71.6 4.7 -28.7 FEMA ELF SRSS -22.93E-03 1.466 0.6 17.5 10.195 0. gravity loads included Displacement /Height (%) FEMA ELF SRSS MPA 1.00E+00 8.55E-03 3.6 -25.23E-02 1.9 15.323 1.8 -35.00E+00
Error (%) FEMA Uniform ELF SRSS MPA
24.3 -11.0 -100.0 -63.178 1.5 -29.4 21.310 1.00E-10
Floor 1st 2nd 3rd 4th 5th 6th 7th 8th 9th
Uniform
1.199 27.62E-03 0.597 0.515 FEMA Uniform ELF 27.3 -22.4 16.09E-03 4.399 1.998 1.75E-03 0.78E-03 0.168 1.0 37.9 11.672 1.03E-03 5.34E-03 2.53E-02 1.8 -100.9 SRSS -22.8 -63.8 -32.214 1.2 19.00E+00 0.089 1.9 -70.4 -23.9 28.59E-03 5.8 17.8 44.9 16.530 1.6 -4.335 2.Table 5.17E-03 9.209 0.877 1.888 0.35E-03 8.51E-03 4.6 -6.5 -32.351 0.0 -57.2 9.294 1.8 7.109 1.6 -41.2 13.8 16.490 0.207 1.938 1.0 -100.830 2.355 0.7 26.1 -26.975 1. gravity loads included Displacement /Height (%) FEMA ELF SRSS MPA 0.011 1.399 1.854 0.00E-03 5.306 1.0 -100.330 1.3 -4.730 1.7 7.2 10.221 1.913 0.234 1.783 1.083 1.686 0.9 -20.5
Story 1st 2nd 3rd 4th 5th 6th 7th 8th 9th
Uniform 2.2 16.8 -100.6 -73.2 5.2 Peak values of story drift ratios (as % of story height) from FEMA force distributions and MPA.

MPA including three “modes” and NL-RHA for 1.5 × El Centro ground motion. Locations of plastic hinges determined from three force distributions in FEMA-273. gravity loads included
62
.5. 5.(a) FEMA: Uniform • • • • • • • • •• •• •• •• •• ••
(b) FEMA: ELF • •• •• •• •• •• •• • •• • •• •• •• •• •• • • • • • •• • •
• • • • •
•• •• •• •• •
•• •• •• •• •
•• •• •• •• •
• •• •• •• •
• • •
• •• ••
(c) FEMA: SRSS • • •• •• •• •• • • •• •• •• • • •• •• ••
• • •
• •• •• ••
• • • • • • • • •
(d) MPA with 3 "Modes" • • • • • • • •• •• •• •• •• •• •• •• •• •• •• •• •• •• •• •• •• •• •• •• •• • • • • •• •• ••
(e) Nonlinear RHA
• • • • • • • •
•• •• • •• •• •• •• •
• •• •• • •• •• •• •• •
• •• •• • •• •• •• •• •
• • • • • •• •• ••
Fig.

5
3
(c) FEMA: SRSS 60 50 40 60 50 40
2 3 1 5 4
(d) MPA (3 "Modes")
Error (%)
30 20 10 0 0
8
Error (%)
30 20
6 8 1 5 2 4 3 9 7
6 7 9
10 0 0
0.5 2 GM Multiplier
2.5 2 GM Multiplier
2. Noted
2
Error (%)
Error (%)
1 2
30 20 10 0 0
1 4 6 7
30 20 10
8
3
5 8
4 9 7
9
6
5
0. Errors in floor displacements from three force distributions in FEMA-273 and from MPA including three “modes”.(a) FEMA: Uniform 60 50 40 60 50 40
3
(b) FEMA: ELF Error Envelope Error for Floor No.6.5
3
Fig.5 2 GM Multiplier
2. 5.5 2 GM Multiplier
2.5
1
1. gravity loads included
63
.5
3
0 0
0.5
1
1.5
1
1.5
1
1.5
3
0.

5.7.5 2 GM Multiplier
2.5
3
0 0
0.5
1
1. Noted
80
40
9 1
Error (%)
7
Error (%)
60
30 20 10
2 5 6 4 8
40
1 6
20
2 5 3 4
3 7
0 0
0.5
1
1.5
3
(c) FEMA: SRSS 60 50 40 60 50 40
3 2 1
(d) MPA (3 "Modes")
Error (%)
30 20
Error (%)
30
5 4
8 9
20
9 6 7 7
10 0 0
4
10
2 3 1 8
5
6
0.5
1
1. gravity loads included
64
.(a) FEMA: Uniform 100 60 50
9 8
(b) FEMA: ELF Error Envelope Error for Story No.5
3
0 0
0.5 2 GM Multiplier
2.5 2 GM Multiplier
2.5 2 GM Multiplier
2.5
1
1.5
3
Fig. Error in story drifts from three force distributions in FEMA-273 and from MPA including three “modes”.

The peak response of an elastic multistory building due to its nth vibration mode can be exactly determined by static analysis of the structure subjected to lateral forces distributed over the building height according to s* = mφ n . which retains the conceptual simplicity and computational attractiveness of the procedure with invariant force distribution. pushover analysis is inherently limited in the sense that it cannot provide any cumulative measure of response. SRSS) leads to the modal pushover analysis (MPA) procedure. hinge plastic rotation. Pushover analysis of a one-story inelastic system predicts perfectly peak seismic demands: deformation. the energy dissipated in yielding or the cumulative rotation of a plastic hinge. This MPA procedure for elastic buildings is shown to be equivalent to the standard response spectrum analysis (RSA) procedure. This system has vibration properties—natural frequency w n and damping ratio..g. etc. ζ n —of the nth-mode of the MDF system. joint rotations. However.g.
65
. 2. and the structure is pushed to the roof displacement determined from the peak deformation Dn of the nth-mode elastic SDF system. now common in structural engineering practice. Combining these peak modal responses by an appropriate modal combination rule (e.6
Conclusions
This investigation was aimed toward developing an improved pushover analysis procedure based on structural dynamics theory.. where the nature and magnitude of errors arising from approximate modal combination rules are well understood. where m is the mass matrix n of the building and φ n its nth mode. For this system. e. Dn is available from the elastic response (or design) spectrum. 3. It has led to the following conclusions: 1.

(ii) Idealize the pushover curve as a bilinear curve and convert it to the bilinear force-deformation relation for the nth-“mode” inelastic SDF system.n t = . n where φ n is now the nth natural mode for small-amplitude linear vibration. with vibration properties in the linear range same as those of the nth-mode elastic SDF system. …) according to an appropriate combination rule (e.
66
.. The response value rno is determined by implementing the following steps: (i) Develop the base-shearroof-displacement Vbn .g. To enable systematic extension of the MPA procedure to inelastic systems.urn
bg
bg
bg
bg
b
g
curve from a
pushover analysis of the structure for the force distribution s* = mφ n . an uncoupled modal response history analysis (UMRHA) is developed by (1) neglecting the coupling among modal coordinates arising from yielding of the system. The MPA procedure developed to estimate the seismic demands for inelastic buildings consists of two phases: (i) the peak response rno of the inelastic MDF system to effective earthquake forces p eff .4. 5. the SRSS rule). p eff (t ) = . and (ii) the total response ro is determined by combining the rno (n = 1. and (2) superposing responses of the inelastic MDF system to individual terms. 2. p eff . These underlying assumptions and approximations are evaluated and the errors in the UMRHA procedure relative to the “exact” nonlinear response history analysis are documented. 6.n t is determined by pushover analysis.sn ug t —n denotes mode number—in the modal expansion of the effective earthquake forces. (iii) Compute the peak deformation Dn of this system with unit mass by nonlinear response history analysis or from the inelastic response (or design) spectrum.mι ug t .

8. 7. This implies that MPA is able to estimate the response of buildings responding well into the inelastic range to a similar degree of accuracy when compared to standard RSA for estimating the peak response of elastic systems. and lead to unacceptably large errors in the hinge plastic rotations. demonstrates that the FEMA force distributions greatly underestimate the story drift demands. Comparing the peak inelastic response of a 9-story SAC building determined by the approximate MPA procedure—including only the first two or three rno terms—with rigorous nonlinear RHA demonstrates that the approximate procedure while providing good estimates of floor displacements and story drifts. and identifying locations of most plastic hinges. and hinge plastic rotations. the errors in the MPA procedure are shown to be only weakly dependent on ground motion intensity. all pushover analysis procedures do not seem to compute with acceptable accuracy local response quantities. Pushover analyses seem to be inherently limited in computing accurately hinge plastic rotations.25 to 3. MPA. The MPA procedure is more accurate than all the FEMA force distributions in estimating floor displacements. joint rotations. such as hinge plastic rotations. story drifts. Thus the MPA procedure is accurate enough for practical application in building evaluation and design. the pushover analysis provides the peak value rno of any response quantity: floor displacements.0. etc. plastic hinge rotations. and nonlinear RHA. However.
67
.(iv) At the roof displacement determined from Dn . The initial state—forces and deformations—of the structure can be considered in the MPA procedure by including these gravity load effects in pushover analysis of the structure only for the first “mode. story drifts.” 10. it fails to compute with acceptable accuracy plastic rotations of the hinges. 9. Comparing the earthquake-induced demands for the selected 9-story building determined by pushover analysis using three force distributions in FEMA-273. Based on results presented for El Centro ground motion scaled by factors varying from 0.

While pushover estimates for floor displacements are even more accurate. with the excitation scaled to cover a wide range of ground motion intensities and building response. structural performance evaluation should be based on story drifts that are known to be closely related to damage and can be estimated to a higher degree of accuracy by pushover analyses. The present trend in the structural engineering profession of comparing computed hinge plastic rotations against rotation limits established in FEMA-273 to judge structural performance does not seem prudent. This report has focused on development of the MPA procedure and its initial evaluation in estimating the seismic demands on a building imposed by a selected ground motion. This new method for estimating seismic demands at low performance levels. should obviously be evaluated for a wide range of buildings and ground motion ensembles.
68
. Instead.11. such as life safety and collapse prevention. they are not good indicators of damage.

. Kim. Spectra. (1997). Krawinkler. E. H. 29:1287-1305. and Kunnath. 20(4-6):452-464. S. Earthq. (1997). S.. Dyn. (1998).. H. Nat.. and Gupta. 5:111-116. Adaptive spectra-based pushover procedure for seismic evaluation of structures. 6th U. Engrg. J. (1988). John A. Stanford. FEMA-273. Method of reliability-based seismic design. Calif. Washington.K. Japan. Bracci. and Powell. Struct. and Fischinger. Berkeley.. Conf. and Krawinkler. Prentice Hall: New Jersey.M. Proc. N2—a method for nonlinear seismic analysis of regular structures. 123:256-265. Blume Earthquake Engineering Center. Proc. Earthq.. Engrg. Seismic demands for performance evaluation of steel moment resisting frame structures (SAC Task 5. (2000). D. Seismic performance and retrofit evaluation for reinforced concrete structures. P.M. (1988). 132. Washington. UCB/EERC-88/06. 69
. Earthquake Engineering Research Center.D. Struc.K. Engrg.3). A. DRAIN-2DX user guide. J.P. J. Fajfar. ASCE. and Reinhorn. University of California. ASCE 123(1):3-10..4. A. J.H. A. Pushover analysis procedure in earthquake engineering. Pros and cons of a pushover analysis of seismic performance evaluation.K. and Seneviratna. Seattle. G. Y. and Krawinkler. (1999).S. Federal Emergency Management Agency.C.W. A. Estimation of seismic drift demands for frame structures. Stanford University. Building Seismic Safety Council (1997).. on Earthq. S. Struct.. Earthq. R.K. (2001). Earthq. H.K. G. Engrg. Report No. Story drift demands for steel moment frame structures in different seismic regions. H. Krawinkler. NEHRP Guidelines for the Seismic Rehabilitation of Buildings. Report No. Engng. 15:417-434. B. Gupta.7
References
Allahabadi. 16(2):367-392 Han.. Chopra. A. and D’Amore. 9th World Conf. Gupta. Struct. (1998). Gupta. S. Kunnath. I: Equivalent nonlinear system. M. Calif. Spectra. Engrg. Dynamics of Structures: Theory and Applications to Earthquake Engineering. (2000). Tokyo-Kyoto. and Wen. (1999).

ASCE. 11th World Conf. Earthq. Conf. 5th U. (2000). Benchmark Control Problems for Seismically Excited Nonlinear Buildings.. Nat.. pp. B.E. New Zealand. Proc. (1981).nd. and Paret. 107(ST5):937-951.. Structural Engineers of California. D.M.. How safe are pre-Northridge WSMFs? A case study of the SAC Los Angeles Nine-Story Building. D. Simple nonlinear seismic analysis of R/C structures. Villaverde. M. Spectra.S. G. Matsumori. Hawaii Miranda. U. Sapporo.S.J. Calif. F. (1991). Jr. R.A.
70
. U. (1999)...A. Lawson. R. S. Maui.C. (1998).. T. Paper No. H. Conf. Earthq. Engrg. Skokan. and Dyke. (1998). Seismic evaluation and upgrading of existing buildings. Engrg. Approximate inelastic procedures to identify failure mechanisms from higher mode effects.. Proc. J.. Struct. Proc. and Freeman. R. 6th U. 122:282-285.. Spencer. Earthq... 1:283-292. Earthquake member deformation demands in reinforced concrete frame structures. S.Kunnath. Naiem. Common pitfalls in pushover analysis.. Japan.F.F. Proc. and Hart..K. Simplified response spectrum seismic analysis of nonlinear structures. Validity of deformation demand estimates using nonlinear static procedures. Dept.edu/~quake/... Structures. R. Ph. Nonlinear static pushover analysis—why. 15(4):765-789.S. (1999). and Bonowitz. Sasaki. D. V. Y. Hokkaido. Ohtori.F. J. Univ. of Civil Engrg. and Krawinkler. Methodology for R/C Bldg.. S. Engrg. Christenson. http://www. Mexico. B. 79-94. Seattle. and Kabeyasawa. Proc. 12th World Conf. (1996). and Gupta. Otani. and Lobo. (1994). T. Berkeley. S.. 1972. M. Div. Eilbekc.K..H.S. Maison. Vance. Washington..A. (2000). 966. Paper No. Engrg. Japan Workshop on Performance-Based Engineering for Reinforced Concrete Building Structures.. Notre Dame University.J. Sasaki. SEAOC 1998 Convention. Dissertation. Proc. K. H. and Sozen. T. Paret.. (2000). Shiohara.. Acapulco. when and how?.-Japan Workshop on Performance-Based Earthq. S. Reliability of nonlinear static methods for the seismic performance prediction of steel frame buildings. Auckland. Engrg. T. on Earthq. Multimode pushover procedure (MMP)—A method to identify the effects of higher modes in a pushover analysis. Mech.. Engrg. Saiidi. Proc. E. (1996). M. ASCE. Indiana. Freeman. B. Earthq.S.K. K. of Calif.

Define the force distribution s* from Eq. using any numerical integration method. for linear-elastic vibration of the building. develop the base-shear – roof-displacement (Vbn . A. 1.Appendix A Uncoupled Modal Response History Analysis
A. Let the roof displacement and base shear at the anchor point be urno and Vbno .. Compute natural frequencies. φn . This step can be conveniently implemented in any commercially available software.1.2. B. Since the target roof displacement may not be known at the start of the procedure. 1977).1) using the FEMA-273 procedure (Building Seismic Safety Council. Apply force distribution of Step 2.20): s* = mφn n n 2. Define the anchor point. iterations may be necessary. Calculate the area under the actual pushover curve.1 STEP-BY-STEP PROCEDURE
A detailed step-by-step implementation of the uncoupled modal response history analysis (UMRHA) procedure is presented in this section and illustrated by an example in the next section.1 incrementally and record the base shears and associated roof displacements. ω n . and modes.. respectively.g. Apn . trapezoidal rule. 1988). 3. 3. e. (3. DRAIN2DX (Allahabadi and Powell. of the bilinear curve at the target roof displacement. For the nth-“mode”. e.1. The structure should be pushed just beyond the target (or expected) roof displacement in the selected mode. 71
.g. Idealize the pushover curve as a bilinear curve (Fig. 2. 3.urn ) pushover curve for the force distribution s* : n 2.2.

Draw the curve OAB by connecting the three points O. urn. 3. This value. and B with straight-line segments to obtain the idealized bilinear curve.6. kn = 0. From the pushover data.
i 3.
i i i αn = Vbno Vbny − 1 urno urny − 1 i 3.4 to 3. If desired. 3.5. This step gives the secant stiffness at a base shear equal to 60% of the yield
base shear. Vbny . Calculate the slope.9. determine the roof displacement. Abn . 3. urny = Vbny kn .6 .6 . Estimate the yield base shear.2. Let the point with base shear = Vbny and roof displacement = urny be
denoted as A. Calculate area under the bilinear curve OAB.1. obtained by judgment.6 × Vbny . Calculate the post-yielding strain-hardening ratio. i i 3. iterations are necessary.4.3. Calculate Vbny = Vbny × Apn Abn . i 3. and a point on the actual pushover curve with base shear equal to
i 0. Calculate the error = 100 × Abn − Apn
(
)
(
)
(
) )
Apn . Calculate the yield displacement.i 3.1.
i i i 3. O. kn . 72
. will be refined by
an iterative procedure that seeks to equate areas under the actual and the idealized pushover curves.0.7.
i 3.8.0.8. A. Replace i+1 with i and repeat Steps 3.
i +1 i i 3.9. corresponding to the estimated yield i i i base shear.4. Calculate initial slope of the idealized bilinear curve.6 × Vbny
(
)
i urn .2. other appropriate methods can be
(
used. If the error exceeds some pre-specified
tolerance. at base shear i equal to 0. Vbny .6 × Vbny .4.9. by connecting a straight line
between origin.

0.10a and 4.16).1. A.11b). 6. 4. of the nth“mode” inelastic SDF system (Fig. Calculate peak values. Dn (t ) . An (t ) . Repeat Steps 3 to 6 for as many modes as required for sufficient accuracy. 9.10b and 4.13).
* 4. (3. Develop the Fsn Ln − Dn relation (Fig. In general first two or three modes will suffice.11a). 8.
5. Compute the Ln and Γ n from Eq.3b) with unit-mass and force-deformation relation of Fig. Compute deformation history. 7.6 × Vbny A •
Idealized αnkn B •
1 Actual
k 1 • O
n
figA_1. of the combined responses obtained in Step 8. Calculate histories of various responses using Eqs.eps
ur n. and pseudo-acceleration history.
Vbn Idealization of Pushover Curve
Vbno Vbny 0.4. 4.12) and (3. A. Scale the horizontal axis by Γnφrn to obtain Dno = urno Γ nφrn and Dny = urny Γ n φrn (Eqs.2.
4. Combine the “modal” responses using Eqs.2.15) and (3.2). 4.4) and effective modal mass from M n = Ln Γn . A.6
ur n y
ur n o
ur n
Fig. r o . (3.1 Idealization of nth-“mode” pushover curve
73
. 4.3. Scale the vertical axis with
* Mn
to
obtain
* Fsno Ln = Vbno M n
and
* Fsny Ln = Vbny M n (Eqs. (3.

Fsn / Ln
Fsn / Ln − Dn Relationship
Vbno / M* n Vbny / Mn
*
1
α ω2
n n
ω2 1
n
figA_2.1. The force distributions. At this point.6 kN.5 cm (25 in.2.2
EXAMPLE
The UMRHA procedure is implemented to calculate the response of the 9-story building described in Section 3. The base-shear – roof-displacement (Vbn .4. s* . Following is step-by-step implementation of the procedure described in Section A. second.). 2. The anchor point. A. Idealized bilinear curves for each of the three modes are included in Fig.eps
Dny = ur n y / Γnφr n Dno = ur n o / Γnφr n
Dn
Fig.3. and 12.urn ) pushover curve for the force distribution s* : n 2.1. The pushover curves for the first three modes. Ap1 = 360777 kN-cm. First three mode shapes and frequencies of the selected building were computed and are shown in Fig.3.). computed for the first three modes are shown in Fig.7 cm (5 in. B. 1.5. Target displacements used to generate these pushover curves are 63.).2 Properties of the nth-“mode” inelastic SDF system
A. 3. ur1o = 63. 74
. 3. A. are shown in Fig. 25.4.5 cm and Vb1o = 8729. is defined at the target roof displacement. The following steps illustrate the procedure to develop the idealize curve for the first “mode”. for the first.4 cm (10 in. n 2. respectively. to the north-south component of the El Centro (1940) ground motion scaled up by a factor of 1. A. 3. 3.2. generated using DRAIN-2DX. and third mode. 3. Area under the actual pushover curve.1.3.

6 kN.4. (4. ur1.9 kN. The curve OAB obtained by connecting the three points O.9. and
α1 = 0.2.18 = 38. The yield displacement. The next estimate of the yield shear is Vbny = 8006.
i i 3.4.4 kN.
i +1 3. and M1 = 2736789 × 1. The initial slope of the idealized bilinear curve.13). The results of the iterative procedure are summarized in Table A.5 38.6 8006.1. Error = 100 × (365100 − 360777 ) 360777 = 1. 75
.8 22.0. ur1 y = Vb1 y k1 = 8006.i 3.198%.1.9.7.4.
i 3.4 210. Area under the bilinear curve OAB.18 kN/cm.
3.86 cm at 0. The post-yielding strain-hardening ratio. k1 . The point A on the i i bilinear curve is defined by ur1 y = 38.4 kN.2.194.
3. The Fs1 L1 − D1 relation for the first “mode” is developed as follows.135. k1 = 0.
i i 3.
i i i 3. The procedure converged after nineteen cycles to give ur1 y = 36. Γ1 = 1. Therefore.1.09 cm. iterations are necessary.
i i i 3.0.6 = 4803. Determined from the pushover database.6 × Vb1 y =
4803.8.4 ) − 1 (63.9.2.
i 3. Vb1 y = 7615.86 = 210.
4.3. α1 = Vb1o Vb1 y − 1 ur1o ur1 y − 1 =
(
)
(
)
(8729. Ab1 = 365100 kN-cm. The first estimate of the yield base shear Vb1 y = 8006.1.3666.23 cm.6 = 22. and B with straight-line segments gives the idealized bilinear curve.6.09 ) − 1 = 0. Also included are the modal damping ratios and the periods calculated from Eq.4 × (360777 365100 ) =
7911.
* 4.09 cm and Vb1 y = 8006. A. L1 = 2736789 kg.3666 = 3740189 kg.01%.5. is calculated as follows. 3. This value exceeds the prespecified tolerance of 0.6 × Vb1 y
(
)
i ur1. The results for other modes are summarized in Table A.8 kN.

4.7.7. The results were generated for first three “modes’ and are included in Fig.62 (cm/sec2).3. 4. 8. The peak values are computed and are summarized in Tables 4.2. 9.2. 4. The combined modal responses are presented in Fig. Scaling the horizontal axis by Γ1φr1 gives D1o = 46. 7.7.4. Histories of roof displacement and top story drifts for the first “mode” are computed and presented in Fig. Deformation and pseudo-acceleration histories of the inelastic SDF systems for the first “mode” with unit mass and force-deformation relation developed in Step 4 are plotted in Fig. 6.51 cm.
* M1
gives
Fs1o L1 =
233.40 (cm/sec2) and
76
.1 and 4. The peak values are also plotted in Fig.8. 4. 5. 4. 4.46 cm and D1y = 26. A. Scaling the vertical axis by Fs1 y L1 = 203.

eps
0 0
2
4 6 8 10 Roof Displacement (cm)
12
Fig.14 y by 6000
Base Shear (kN)
4000
2000
Actual Idealized
figA_3c.3.eps
0 0
5
10 15 20 Roof Displacement (cm) (c) "Mode" 3 Pushover Curve
25
8000 u = 4.2 cm.(a) "Mode" 1 Pushover Curve 12000 10000 u = 36.19 y by
Base Shear (kN)
8000 6000 4000 2000 0 0
Actual Idealized
figA_3a.13
6000
Base Shear (kN)
4000
2000
Actual Idealized
figA_3b. α = 0. V = 7616 kN.6 cm. “Modal” pushover curves for the example building
77
.eps
10
20 30 40 50 Roof Displacement (cm) (b) "Mode" 2 Pushover Curve
60
8000 u = 9. α = 0. α = 0. V
y by
= 4952 kN. V = 5210 kN.9 cm. A.

223
−80 80 n=3
−3 3 n=3
D3 (cm)
0
A3 (g)
• 10.4.5 × El Centro ground motion for the first “mode. A. Histories of deformation and pseudo-acceleration due to 1.06
0 • 1.(a) Deformation 80 n=1 3
(b) Pseuso−acceleration n=1 • 0.2769
D1 (cm)
0 • 35.
78
.m
−80 0
−3 0
25
30
Fig.52 5 10 15 20 Time (sec) 25 30
0 • 1.33
A1 (g)
n=2
0
−80 80
−3 3 n=2
D2 (cm)
0
A1 (g)
• 22.747 5 10 15 20 Time (sec)
figa_4.” second “mode.” and third “mode” inelastic SDF systems.

18 210.26 36.237 0.4 7714.193 0.191 0.83 21.50 36.6 × Vb1 y
i ur1. Properties of “modal” inelastic SDF systems
Properties Ln (kg) Γn
* M n (kg)
“Mode” 1 2736789 1.4 7647.022 0.25 36.037 0.6 4583.11 22.12 3876.18 210.6 7840.05 36.194 0.09 37.184 0.24 36.4 7911.1.151 0.107 0.59 36.51
2
“Mode” 2 -920860 -0.404 0.23 22.64 37.02 21.90 21.4 7672.0 4588.8 7618.75 21.2 4628.5
(cm) 22.76 21.3 4603.9 4570.81 21.193 0.0 7619.70 36.18 210.9
(kN) 4803.2.74 21.309 0. No.79 0.162 0.25 36.5309 488839.193 0.18
(cm) 38.9 7615.529 0.188 0.23 0.8 4647.44 36.28 36.18 210.010
(kN) 8006.86 22.194 0.082 0.5 7624.5 3109.8 7622.017 0.9 4573.Table A.77 21.65 1226.18 210.1 4569.0 4704.29 36.18 210.46 2.1 4574.135 0.39 36.013 0.18 210.2 7690.029 0.198 0.048 0.18 210.09 18.40 46.4927 1.103
“Mode” 3 696400 0.3 7628.18 210.3 7786.74
(kN/cm) 210.948
ζ n (%)
79
.18 210.18 210.8525 1.18 210.18 210.186 0.32 36.18 210.139 0.6
i k1
i ur1 y
i α1
i Ab1
Error (%) 1.3 7658.56 47.180 0.75 21.0 4577. Results of iterative procedure to develop the idealized bilinear curve for the first “mode” inelastic SDF system
Itr.9 4570.176 0.18 210.78 21.182 0.2 4671.5 7633.136
Fsny Ln (cm/sec2) Dny (cm) Fsno Ln (cm/sec ) Dno (cm) Tn (sec)
233.2406 167531.18 210.56 19.7 4580.0. i 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19
i Vb1 y
i 0.95 21.79 21.1 7616.30 37.35 36.62 26.85 36.194 0.170 0.693 0.5 4614.38 22.3666 3740189 203.190 0.192 0.4 4595.86 21.18 210.1 1013.18 210.3 7745.2671 1.2 4571.85 0.063 0.194
(kN-cm) 365100 364060 363276 362684 362235 361892 361631 361432 361279 361162 361073 361004 360951 360911 360880 360856 360838 360824 360813
Table A.910 0.8 4747.59 22.05 52.7 7639.

80
.

or from the inelastic response (or design) spectrum. This is the same example as solved in Appendix A. 12.Appendix B Modal Pushover Analysis
B. of the nth-“mode” inelastic SDF system (Fig. From the total hinge rotation.1 STEP-BY-STEP PROCEDURE
A detailed step-by-step implementation of the modal pushover analysis (MPA) procedure is presented in this section and illustrated by an example in the next section.8) for the peak deformation of the first-“mode” inelastic SDF system with unit mass and force-deformation relation developed in Step 4 gives D1 = 35. Solving Eq.6b by solving Eq. 13.33 cm.18).4.5. Steps 1 to 4 of the MPA are the same as those for UMRHA presented in Appendix A. Dn .8). 10. (4. 4. rno . 81
.1. 4. extract from the pushover database values of other desired responses.3b) with unit mass and force-deformation relation of Fig. (4. B. Following is step-by-step implementation of the procedure described in Section B. 10. the first two or three “modes” will suffice. (3. Compute the peak deformation. (3. Repeat Steps 3 to 7 for as many “modes” as required for sufficient accuracy.21). At urno .1 to the north-south component of the El Centro (1940) ground motion scaled up by a factor of 1. 6. 11. Calculate the peak roof displacement urno associated with the nth-“mode” inelastic SDF system from Eq.2 EXAMPLE
The MPA procedure is implemented to calculate the response of the 9-story building described in Section 3. Determine the total response by combining the peak “modal” responses using the SRSS combination rule of Eq. Typically. subtract the yield hinge rotation to determine the plastic hinge rotation.

1. values of floor displacements and story drifts are extracted.11. 12.3 for the floor displacements normalized by the building height and in Table 4.” and the results are included in Tables 4. The values are summarized in Table 4. Results of Steps 5 and 6 are summarized in Table B. 9.366 × 1 × 35. Peak roof displacement ur1o = Γ1 × φr1 × D1 = 1.28 cm. At ur1o = 48. Also included in Table 4.4 for the story drifts normalized by the story height.4.3 and 4.18) are also included in Tables 4. The total response computed by combining the peak “modal” responses using the SRSS combination rule of Eq.5 are the plastic hinge rotations computed by subtracting the yield hinge rotation from the total hinge rotations.28 cm.33 = 48.4. 8.3 and 4. (3. where results for other ground motion intensities are also included.
82
. Steps 3 to 7 are repeated for first three “modes.

535 14.03 0.71 1.275 1.504 18.312 1.367 1.8451 5.856 31.969 0.748 63.5
µ = Dn Dny
urno = Γnφrn Dn (cm) Dn (cm)
0.27 0.35 1.1.154 78.37 57.184 0.691 0.52
“Mode” 3 1.225 2.82 1.03 26.185 11.35 0.007 36.467 14.18 27.332 13.229 8.Table B.50 35.660 14.513 0.577 16.4222 3.252 9.36 1.913 22.07
“Mode” 2 4.678 0.755 0.13 2.267 5.73 24.70 0.735 3.37 1.457 12.79 0.766 7.755 27.0
µ = Dn Dny
urno = Γnφrn Dn (cm) Dn (cm)
1.023 0.06 1.5
µ = Dn Dny
urno = Γnφrn Dn (cm) Dn (cm)
2.551 2. Calculation of roof displacements urno from peak deformation of inelastic SDF systems
Ground Motion Multiplier 0.332 48.05 0.901 8.395 0.38 22.28 46.200 0.0
µ = Dn Dny
urno = Γnφrn Dn (cm) Dn (cm)
3.690 10.676 6.268 0.117 5.126 13.52 0.59 0.75
µ = Dn Dny
urno = Γnφrn Dn (cm) Dn (cm)
0.25 20.25
Quantity Dn (cm)
“Mode” 1 6.85
µ = Dn Dny
urno = Γnφrn Dn (cm) Dn (cm)
1.436 7.33 1.139
µ = Dn Dny
urno = Γnφrn Dn (cm) Dn (cm)
0.379 21.0
µ = Dn Dny
urno = Γnφrn Dn (cm)
83
.450 4.

84

**Appendix C FEMA Force Distribution Calculations
**

Presented in this appendix are the calculations leading to the FEMA-273 force distributions (Fig. 5.1) used in developing the pushover curves (Fig. 5.2). These distributions were described in Section 5.1 C.1 “UNIFORM” DISTRIBUTION

The lateral force at a floor is equal to the mass at that floor, i.e., s* = m j . For convenience, the j floor forces are normalized with the base shear. The results are summarized in Table C.1.

Table C.1. FEMA273 “uniform” lateral force distribution

Floor, j

mj (10-3×kg)

s* = j

∑i mi

0.112 0.110 0.110 0.110 0.110 0.110 0.110 0.110 0.119

mj

1 2 3 4 5 6 7 8 9

503.5 494.7 494.7 494.7 494.7 494.7 494.7 494.7 534.1

C.2

EQUIVALENT LATERAL FORCE (ELF) DISTRIBUTION 85

The lateral force at a floor is computed from s* = m j h k where m j is the mass, h j is the height j j of the jth floor above the base, and the exponent k = 1 for fundamental period T1 ≤ 0.5 sec, k = 2 for fundamental period T1 > 2.5 sec; and varies linearly in between. For the selected building, T1 = 2.27 and k = 1.885. The resulting lateral forces are summarized in Table C.1. For convenience, the floor forces are normalized with the base shear.

Table C.2. FEMA273 equivalent lateral force (ELF) distribution

Floor j 1 2 3 4 5 6 7 8 9

**m jhk j (10 ×kg-m ) 371.0 1015.0 1963.2 3196.8 4707.4 6488.3 8534.2 10840.6 14471.1
**

-3 k

s* j

=

m jhk j

∑i mihik

0.007 0.020 0.038 0.062 0.091 0.126 0.165 0.210 0.281

C.3

SRSS DISTRIBUTION

The calculation of the SRSS distribution is summarized as a series of steps as follows: 1. For the nth-mode calculate the lateral forces, f jn = Γn m jφ jn An in which j denotes the floor number and An is the pseudo-acceleration of the nth-mode SDF elastic system, leading to columns 2 to 4 of Table C.3. 2. Calculate the story shears, V jn = ∑ i = j fin where j is now the story number. Implementing this step gives columns 5 to 7 of Table C.3. 3. Combine the modal story shears using SRSS rule, V j = Table C.3. 4. Back calculate the lateral forces at the floor levels from the combined story shears V j to obtain column 9 of Table C.3. 86

N

∑ n (V jn )

2

to get column 8 of

8 374.367
87
.0 1231. the lateral forces are normalized by the base shear to obtain column 10 in Table C.0 980.5 -320.2
Lateral Force fj (kN) (9) 203.9 366.098 0.8 1381.2 285.6 -352.5 215.177 0.0 354.2 105.7 355.9 -166.9 446.4 250.4 1842.9 1683.1 -438.0 1476.1 -525.4 1759.3 -6.2 200.
Table C.0 176.7 1578.2 277.0 381.7 234.5 -973.9 832.5 320.7 525.1
Story Shears V j2 (kN) (6) 1114.9 1446.090 0.3 -646.7 95.6 319.9 -153.065 0.070 0.5 -350.4 400.7 101.042 0.1 -967.7 2065.8 -326.3 240.7 Vj (kN) (8) 2268.1 -646.1 832.1 87. FEMA 273 SRSS force distribution
Lateral Forces Floor j (1) 1 2 3 4 5 6 7 8 9 f j1 (kN) f j2 (kN) f j3 (kN) V j1 (kN) (5) 1917.047 0.8 430.0 -5.7 1622.1 1857.2 148.3 222.7 374.6 1233.For convenience.3.6 286.5 159.3.7 694.2 97.6 -359.9 880.6 366.6 V j3 (kN) (7) 478.6 -732.2 fj =
∑ fi
i
fj
(2) (3) (4) 59.7 -46.045 0.2 95.7
(10) 0.0 136.