This action might not be possible to undo. Are you sure you want to continue?

# 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 .

ro . This pushover curve is idealized as bilinear n and converted to the force-deformation relation for the nth-“mode” inelastic SDF system. Combining these peak modal responses by modal combination rule leads to the MPA procedure. Instead. is determined by pushover analysis. rno . but provides superior accuracy in estimating seismic demands on buildings. p eff t = .urn curve is developed from a pushover analysis for force distribution s* . in the modal expansion of the effective earthquake forces. Comparing the earthquake-induced demands for the selected 9-story building determined by pushover analysis using three force distributions in FEMA-273.mι ug t . However. Thus the MPA procedure is accurate enough for practical application in building evaluation and design. The peak deformation of this SDF system is used to determine the roof displacement. MPA.0. a pushover analysis determines the peak response rno of the inelastic MDF system to individual modal terms. which retains the conceptual simplicity and computational attractiveness of current procedures with invariant force distribution. 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. The MPA procedure is extended to estimate the seismic demands for inelastic systems: First. The results presented for El Centro ground motion scaled by factors varying from 0. at which the seismic response. iii .n t = . 2. it is demonstrated that the FEMA force distributions greatly underestimate the story drift demands. Second. where m is the n mass matrix and φ n its nth-mode. 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. all pushover analysis procedures considered do not seem to compute to acceptable accuracy local response quantities. plastic hinge rotations are less accurate. and the structure is pushed to the roof displacement determined from the peak deformation Dn of the nth-mode elastic SDF system. the total demand. and nonlinear RHA.ABSTRACT The principal objective of this investigation is to develop a pushover analysis procedure based on structural dynamics theory. and identifies locations of most plastic hinges. 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. The standard response spectrum analysis (RSA) for elastic buildings is reformulated as a Modal Pushover Analysis (MPA). p eff . 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 .… according to an appropriate modal combination rule. such as hinge plastic rotations.25 to 3. is determined by combining the rno n = 1. and the MPA procedure is more accurate than all the FEMA force distributions methods in estimating seismic demands.sn ug t . 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. The base shear-roof displacement Vbn .

iv .

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

vi .

..........................5 Response History Analysis ............................................................1 4.....2 FEMA-273 Pushover Analysis ..............................................55 Comparative Evaluation...........................................................................................................1 System and Excitation Considered .....................................................................4 Modal Response History Analysis....2 Response History Analysis .................................................... Introduction...................................................4............. Inelastic Multistory Buildings ................1 2......34 4............4........................ Equation of Motion .........................12 Modal Pushover Analysis ......................................................1 Summary.........2 Modal Pushover Analysis ...............................................................3 2....27 Uncoupled Modal Response History Analysis ...................... 2.9 3.....17 3..36 4...............................47 4.4.......v Table of Contents.....................................6 Elastic Multistory Buildings ...................................................................................................................................3 Summary................. vii 1.................................................1 5..................................................................3 4....................................................3..................................................................................................................................................................33 4.................1 Uncoupled Modal Response History Analysis .............................3 2...............................................................3 Modal Pushover Analysis with Gravity Loads ..............................CONTENTS Abstract ...........................................4...........................................................................................................38 4................................2...................................................................................... iii Acknowledgment ........................................2..............3 Modal Pushover Analysis .........................28 4.........................1 Underlying Assumptions and Accuracy ...............12 Comparative Evaluation of Analysis Procedures ...27 4.................9 Modal Response Spectrum Analysis ....................................................................................................36 Modal Pushover Analysis .4......................................................4 5...............................................2....................................13 3. Comparison of Modal and FEMA Pushover Analyses................................................1 3..................55 5..................................................13 3.......................................2 3...............................................3 3................55 vii ................4....41 4.............................................38 4...........................3 System and Excitation Considered .............................................................................................................................................................................................2 Properties of nth-“mode” Inelastic SDF System ...............................4 3..............6 Pushover Analysis ....................................................................................................22 4................................37 Comparative Evaluation of Analysis Procedures ........................................2 Response History Analysis ......................................................2 2.......1 One-Story Systems ........

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

1999. 1988. Fajfar and Fischinger. 2000. after the structure yields both assumptions are approximate. 2000]. Maison and Bonowitz. Gupta and Kunnath. 1999]. None of the invariant force distributions can account for the contributions of higher modes to response. 1997. such as life safety and collapse prevention. However. current structural engineering practice uses the nonlinear static procedure (NSP) or pushover analysis in FEMA-273 [Building Seismic Safety Council. Gupta and Krawinkler. To overcome these limitations. 1988. Bracci et al.. Krawinkler and Seneviratna. While these adaptive force distributions may provide better estimates of seismic demands [Gupta and Kunnath. 2000]. but investigations [Saiidi and Sozen. Lawson et al. 1998. Attempts 1 .. 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. Miranda. 1981. 1997]. 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. 1994. or for a redistribution of inertia forces because of structural yielding and the associated changes in the vibration properties of the structure. 2000] have led to good estimates of seismic demands. they are conceptually complicated and computationally demanding for routine application in structural engineering practice. requires explicit consideration of inelastic behavior of the structure. While nonlinear response history analysis (RHA) is the most rigorous procedure to compute seismic demands. Obviously.and medium-rise structures in which inelastic action is distributed throughout the height of the structure [Krawinkler and Seneviratna. 1999. 1991. Skokan and Hart. such satisfactory predictions of seismic demands are mostly restricted to low. Gupta and Krawinkler.1 Introduction Estimating seismic demands at low performance levels. 1998. several researchers have proposed adaptive force distributions that attempt to follow more closely the time-variant distributions of inertia forces [Fajfar and Fischinger. Kim and D’Amore. 1999.

. First. 2 . the underlying assumptions and approximations are identified. 1998.have also been made to consider more than the fundamental vibration mode in pushover analysis [Paret et al... 1996. 2000]. but provides superior accuracy in estimating seismic demands on buildings. 2000. Sasaki et al. Gupta and Kunnath. the seismic demands determined by pushover analysis using three force distributions in FEMA-273 are compared against the MPA and nonlinear RHA procedures. Kunnath and Gupta. 2000. Finally. The MPA procedure is then extended to inelastic buildings. Matsumori et al. 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. 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. we show that pushover analysis of a one-story system predicts perfectly peak seismic demands. and the errors in the procedure relative to a rigorous nonlinear RHA are documented.

Unloading from a maximum deformation takes place along a path parallel to the initial elastic branch. and a linear viscous damper with damping coefficient c. R y .1) b g ug t is mu + cu + f s u. The yield strength is the same in the two directions of deformation. this system is linearly elastic with stiffness k as long as the force does not exceed f y . During yielding the stiffness of the frame is a k .1b. through the yield strength reduction factor. the strength required for the structure to remain elastic during the ground motion. the yield deformation. 2.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 (2.1 One-Story Systems EQUATION OF MOTION Consider the idealized one-story structure shown in Fig. This lateral-force displacement relation is idealized as shown in Fig. reloading from a minimum deformation takes place along a path parallel to the initial elastic branch. The yield strength is related to f o . On initial loading. 2.2) . It is the familiar bilinear hysteretic relationship. the yield strength. a massless frame that provides stiffness to the system. sign u . where 0 < a << 1. Yielding begins when the force reaches f y and the deformation reaches u y .1a. Similarly. 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. sign u = .2 2.

in addition to the form of the force-deformation relation.3). u £ u y ).e. (a) Idealized one-story structure. sign u = . and u y .fS m fs c ut u fy 1 αk k 1 uy ug -fy k k um u 1 1 (a) (b) Fig. The peak. and (b) bilinear hysteretic forcedeformation relation For a given excitation u g t . 2. This becomes evident if Eq.4) and w n is the natural vibration frequency.1) is divided by m to obtain ~ u + 2zw n u + w 2 u y f s u. m depends on three system parameters: w n . and the ductility factor is u m= m uy (2.5) For a given u g t . 2001. z . or absolute (without regard to algebraic sign) maximum. the deformation u t depends on three systems parameters: w n .ug t n where bg bg b g bg (2.3) ωn = k m ζ = c 2mω n f fs = s fy (2.1. and R y (Chopra . and z is the damping ratio of the system vibrating within its linear elastic range (i. z y .. bg 4 . Tn = 2p w n is the natural vibration period. deformation is denoted by um . Section 7. (2.

826 kips) for w = 169. One-story. is f y / w = ( f o / w ) ÷ 8 = 0.4 in Chopra. and f y = 39. 2.2. α = 0.9 kN (38. m h = 3.5 sec and z = 5% subjected to the north-south component of the El Centro (1940) ground motion (Fig.2311 . based on Ry = 8 .2 kips) . 2. For . u (cm) 8 Fig. f o w = 184 .26 kN.2 5 . Pushover curve for structure shown in Fig. one-bay frame 50 uy = 1.66 m • EIb EIc EIc • • L = 7. 2001) and scaled up by a factor of 2. Vby = 39. Vb (kN) 30 20 10 0 0 2 4 6 Roof Displacement.2 SYSTEM AND EXCITATION CONSIDERED Consider the one-story system in Fig.2: the dimensions and flexural rigidity of the structural elements are noted. 2. The yield strength of the inelastic system.3.2.32 m • Fig.26 kN (8. this system and excitation. with Tn = 0.376 cm.04 40 Base Shear.1. 6. 2.

18 kN-m (444. The resulting pushover curve is shown in Fig. and q pm = 0.The yield moments in the beam and columns are defined as the bending moments due to the lateral force f y .) and 50. Implementing this analysis with I c = 6.4 ).36 cm.6 kip-in. as apparent in Fig. e.3.3 RESPONSE HISTORY ANALYSIS Figure 2. 2. (c) shows the joint rotation q bt g . and E = 2 × 108 kPa (29×103 ksi) gives yield moments of 21.4 ). 2.65 kN-m (191.36 cm .g. 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 . 2. I b = 3.017 rad .. the peak value determined from RHA. respectively. 2. Observe that the pushover curve matches the initial loading path of the hysteretic system.4e.077 × 107 mm 4 (146 in.134 × 107 mm 4 (75. (d) shows the rotation q p bt g of the plastic hinges at the bg bg beam ends. The yielding stiffness of each structural element is defined as 3% of its initial stiffness. A static pushover analysis of this one-story system leads to the force-displacement relationship shown in Fig. It is organized in five parts: (a) shows the deformation u t . the ductility factor m = 5.4 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 joint rotation and beam hinge rotation are identical to values q m and q pm determined from RHA. 2. pushover analysis cannot provide any cumulative measure of response. 2. The system is excited well beyond the yield deformation. (b) shows the lateral force f s t or base shear Vb t normalized relative to the bg weight.0217 rad . wherein the hysteretic force-deformation history of Fig. the energy dissipated in 6 .4edetermined from RHAis superimposed.3 in. and (e) shows the force-deformation relation.4 shows the earthquake response of the system described in the preceding section determined by response history analysis (RHA).35. Determined from pushover analysis at the exact peak deformation.) for the beam and columns. The peak values of the various response quantities are as follows: um = 7.1 kip-in. q m = 0. However.4f.

4 0.017 • 0 (d) θp (rad) −0.5 fy / w = 0. (d) plastic hinge rotation. (e) force-deformation relation.0217 • θ (rad) 0 (c) −0. This represents an inherent limitation of pushover analyses.04 0.4 −10 −5 0 u (cm) 5 10 −0.4 0. Response of one-story system to El Centro ground motion: (a) deformation.2 −0.4 −10 −5 0 u (cm) 5 10 Fig. 15 u (cm) 0 • 7. or the cumulative rotation at a plastic hinge. and (f) pushover curve 7 .04 0 5 10 15 Time (sec) 20 25 30 0.04 0.yielding during the ground motion.36 −15 0.2 Vb / w 0. (b) base shear.2 0 (e) 0 (f) −0.2311 y (b) −0.04 0. (c) joint rotation.2311 (a) Vb / w 0 f / w = −0. 2.4.2 −0.5 0.

8 .

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 = .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 .snug t n= n =1 The effective earthquake forces can then be expressed as af N af N af (3. m.1) where u is the vector of N lateral floor displacements relative to the ground. and k are the mass. classical damping.1) can be interpreted as effective earthquake forces: p eff t = . 2001: Section 13. and lateral stiffness matrices of the system.3) where φ n is the nth natural vibration mode of the structure. This force distribution can be expanded as a summation of modal inertia force distributions s n (Chopra.12): mι = n =1 Â sn = N n =1 Â Γ nmφ n N (3. and p eff t = Â peff. The right side of Eq.mι ug t bg bg bg (3. Elastic Multistory Buildings 3.m ι u g t bg bg (3.4) 9 .n t = Â .3. (3. each element of the influence vector ι is equal to unity. c.

sn ug t bg bg bg bg (3.9) and (3.5) The contribution of the nth mode to s and to p eff t are: sn = Gn mφ n respectively.9) to the equation of motion for the nth-mode elastic SDF system.n t is entirely in the nth-mode.8) where the modal coordinate qn t is governed by 2 qn + 2z nw n qn + w n qn = .8) gives the floor displacements u n t = Gnφ n Dn t bg (3.n (t ) = N n =1 ∑ −sn u g (t ) N (3.sn ug t bg (3.9) in which w n is the natural vibration frequency and z n is the damping ratio for the nth mode.7) By utilizing the orthogonality property of modes. (3. The equations governing the response of the system are mu + cu + ku = . p eff .11) and substituting in Eq. with no contribution from other modes.12) . an SDF system with vibration propertiesnatural frequency w n and damping ratio z n of the nth-mode of the MDF system.10) bg bg bg 10 (3. (3. (3.6) Next.10) gives qn t = Gn Dn t bg bg bg (3.peff (t ) = n =1 ∑ peff . subjected to u g t : Dn + 2z nw n Dn + w 2 Dn = . we will outline that the response of the MDF system to p eff . Then the floor displacements are u n t = φ n qn t bg bg bg bg (3.u g t n Comparing Eqs.Gn ug t (3. it can be demonstrated that none of the modes other than the nth mode contribute to the response.n t = . The solution qn t can readily be obtained by comparing Eq.

Eqs.1). etc. and An t = w 2 Dn t n bg (3.14) is the pseudo-acceleration response of the nth-mode SDF system (Chopra.n t . 2001. (3.1.12) and (3.1. (3.13) define the contribution of the nth-mode to the response. 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 (3.9) is the standard modal equation governing qn t .15) rt = bg n =1 Â rn t = bg n =1 Â rnst An bt g N (3.16) reflect combining the response contributions of all modes.Forces sn An(t ) st rn ωn.12) and (3. (3. internal element forces. Conceptual explanation of modal response history analysis of elastic MDF systems Any response quantity r t story drifts. However. Section 13. Therefore. the static value of r due to external forces sn .13) represent the response of the MDF system to p eff. st The two analyses that lead to rn and An t are shown schematically in Fig. 3.16) This is the classical modal RHA procedure wherein Eq. Equations (3.13) st where rn denotes the modal static response. and Eqs.15) and (3. these standard equations have been derived in an unconventional way. 3. In contrast to the classical 11 bg . ζn ¨ g(t ) u (a) Static Analysis of Structure (b) Dynamic Analysis of SDF System Fig.can be expressed by st rn t = rn An t bg bg bg bg bg (3.

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. Section 13.9) .1. The peak modal responses are combined according to the Square-Root-of-Sum-ofSquares (SRSS) or the Complete Quadratic Combination (CQC) rules.8.1). we note that static analysis of the structure subjected to lateral forces f no = Γ n mφ n An (3..19) will provide the same value of rno . Alternatively. This concept provides a rational basis for the modal pushover analysis procedure developed later. z n of the pseudo-acceleration response (or design) spectrum for the nth-mode SDF system.18) To develop a pushover analysis procedure consistent with RSA. (3. 2001. 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. 2001.17) (Chopra.4 and 13. (3. The SRSS rule. In such a response spectrum analysis (RSA). 3. which is valid for structures with well-separated natural frequencies such as multistory buildings with symmetric plan. we used the modal expansion of the spatial distribution of the effective earthquake forces.3 MODAL PUSHOVER ANALYSIS (3.derivation found in textbooks (e.16). and Tn = 2p w n is the natural vibration period of the nth-mode of the MDF system.17) where An is the ordinate A Tn .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.g. Chopra. Sections 12. the peak nth-mode response as in Eq.3).

this building is 45.1 System and Excitation Considered The 9-story structure. can be combined according to Eq. Richmond. Los Angeles. The columns are 345 MPa (50 ksi) steel wideflange sections. in both directions.2. region. and 37. shown in Fig. 1301 S. 13 . denoted B1.21) 2 where Dn = An ω n . which from Eq. rno .15 m (30 ft) on center. 3.12) is urno = Gnf rn Dn (3. Third St. 46th Street.s* = mφ n n (3.19 m (122 ft) in elevation.20) and the structure is pushed to the roof displacement. Typical floor-to-floor heights (for analysis purposes measured from center-of-beam to center1 2 Brandow & Johnston Associates. the peak value of the roof displacement due to the nth-mode.4 COMPARATIVE EVALUATION OF ANALYSIS PROCEDURES 3.. CA 94804-4698. with the ninth level being the roof. the Applied Technology Council (ATC). The building has a basement level. Consulting Structural Engineers. urno . each determined by one pushover analysis. with five bays each in the north-south (N-S) and east-west (E-W) directions. The peak modal responses. 1660 W.18) to obtain an estimate of the peak value ro of the total response. and California Universities for Research in Earthquake Engineering (CUREE).4. This modal pushover analysis (MPA) for linearly elastic systems is equivalent to the well-known RSA procedure (Section 3. The levels of the 9-story building are numbered with respect to the ground level (see Fig. 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. (3. 3. (3.2). SAC Steel Project Technical Office. SAC is a joint venture of three non-profit organizations: The Structural Engineers Association of California (SEAOC). this structure meets seismic code and represents typical medium-rise buildings designed for the Los Angeles. The interior bays of the structure contain simple framing with composite floors.2). 3.73 m (150 ft) in plan. was designed by Brandow & Johnston Associates1 for the SAC2 Phase II Steel Project. Although not actually constructed.73 m (150 ft) by 45. A benchmark structure for the SAC project. Obviously Dn and An are available from the response (or design) spectrum. California. CA 90017. The bays are 9.

3. These force n distributions will be used in the pushover analysis to be presented later. The floor system is composed of 248 MPa (36 ksi) steel wide-flange beams in acting composite action with the floor slab.e.49 sec. The strength.65 m (12 ft) and for the first floor is 5.0 kips-sec2/ft). The force distributions. Each frame resists one half of the seismic mass associated with the entire structure.2 kips-sec2/ft). third. for the second through eighth levels is 9.3. we select a weak ground motion: the northsouth component of the El Centro (1940) ground motion scaled down by a factor of 4. The column bases are modeled as pinned and secured to the ground (at the B-1 level). The seismic mass of the structure is due to various components of the structure. 1988) using the M1 model developed by Krawinkler and Gupta (1998).7 kips-sec2/ft). The first three vibration modes and periods of the building for linearly elastic vibration are shown in Fig.96 m (13 ft).sec2/ft). i. respectively. ceiling/flooring.85. for the first level is 1.2.of-beam) are 3.83 m (6 ft) above the center-line of the beam to column joint. and 0. 14 . for the first three modes are shown in Fig. To ensure that this structure remains elastic. The column lines employ two-tier construction.49 m (18 ft). and for the ninth level is 1. are located on the first. Column splices. roofing and a penthouse located on the roof.07×106 kg (73. which are seismic (tension) splices to carry bending and uplift forces. The 9-story N-S MRF is depicted in Fig.4.01×106 kg (69. This model is based on centerline dimensions of the bare frame in which beams and columns extend from centerline to centerline. s* (Eq.00×106 kg (616 kips. The simple model adopted here is sufficient for the objectives of this study. partitions. if desired more complex models. mechanical/electrical. The floor-to-floor height of the basement level is 3. The seismic mass of the ground level is 9. The building is modeled in DRAIN-2DX (Allahabadi and Powell.89×105 kg (67. and seventh levels at 1. fifth. 3. floor slabs. Concrete foundation walls and surrounding soil are assumed to restrain the structure at the ground level from horizontal displacement.20). such as those described in Gupta and Krawinkler (1999) can be used.65×105 kg (66. 3. monolithic column pieces are connected every two levels beginning with the first level. dimension. and shear distortion of panel zones are neglected but large deformation (P∆) effects are included. including the steel framing. The seismic mass of the above ground levels of the entire structure is 9.0 kips-sec2/ft). the vibration periods are 2.27. 0. 3..

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

(3.67 −1.0272 −2.05 s * 1 s * 2 s * 3 Fig.5.33 2. 2.8 −2.9th 8th 7th 6th T = 0. respectively.487 −1.05 2.2 Response History Analysis The structural response due to individual vibration modes. Force distributions s* = mφn . n = 1. Each figure is organized in four parts: (a) shows the roof displacement urn t .93 −1.4.75 1.31 −0. n = 1.04 1.5 −1 −0.38 0.05 1.5 Fig.13 −1.728 2.39 3. 3. (c) shows the joint rotation q n t of an external joint at 16 bg bg bg . 3. 2. 3.4.7.44 1.6.3. 3. (b) shows the base shear Vbn t normalized relative to the weight W of the building. and 3 n 3.37 2.796 0.1 3. and 3.49 sec 3 T = 0.51 0.61 2.94 2. is shown in Figs.1 −2. and 3 .12 0. First three natural-vibration periods and modes of the 9story building 3.12) and (3.03 −1. determined by RHA [Eqs.5 0 0.85 sec 2 Floor 5th 4th 3rd 2nd 1st T1 = 2.5 Mode Shape Component 1 1.27 sec Ground −1.72 −2.13)].

8. 3. and 3. The peak values of displacements of all floors. consistent with the increasing significance of the higher mode response among these three sets of response quantities. and 3.3. respectively. errors are smallest in floor displacements.16)]. 3. three. and even larger in joint rotations. and ur 3o = 0.12 cm.422 cm. where the second and third modal responses are a larger percentage of the top story drift compared to roof displacement. 3. two. 17 . are presented in Fig.the roof level. Combining the modal response histories for all modes gives the total response [Eqs. the peak roof displacement due to each of three modes is ur1o = 9. and rotations of external joints with moment connections are presented in Tables 3. As expected. The errors due to truncating the contributions of vibration modes beyond the third mode are negligible. with the first three modes provide essentially the exact response. two.2. especially in estimating the story drifts. 3. It is apparent that the first mode alone is inadequate. The same method was used to determine the peak values of many response quantities.23 cm.urn relation. the exact response considering all modes.2. or all modes. the results for the roof displacement and top-story drift are shown in Fig. errors generally decrease as response contributions of more modes are included. respectively. larger in story drifts. which are listed in Tables 3.3. For a fixed number of modes included. and three vibration modes.9. (3. and the percentage errors due to truncation of higher modal contributions. ur 2o = 2. Also included are the combined response due to one.8. 3. and (d) shows the Vbn . but three modes—perhaps even two modes—are sufficient. The peak values of the various response quantities are noted in these figures.1. The peak values of floor displacements and story drifts determined by RHA. This is illustrated in Fig. The linear relationship between the base shear and roof displacement for each mode implies that the structure did not yield. including one.1.15) and (3. in particular. drifts in all stories.

307 0.99E-03 2.333 0.062 -0.9 8.258 0.44E-03 3.3 Peak values of joint rotations (radians) from RHA for 0.8 1.045 0.1 Table 3.4 -22.6 0.13E-03 2.2 -1.8 1.121 0.Table 3.4 -41.63E-03 2.133 -0.237 0.303 0.350 0.2 9.3 Table 3.6 0.65E-03 RHA (all modes) 2.229 0.038 0.097 0.364 0.413 RHA (all modes) 0.1 -2.3 -33.13E-04 9.14 m3) from RHA for 0.00E-03 2.0 -0.266 0.25 × El Centro ground motion Floor 1st 2nd 3rd 4th 5th 6th 7th 8th 9th 3 Modal Response Mode 1 2.74E-04 6.90E-03 3.15E-03 4.8 -1.1 Building height is measured from the ground floor to the 9th floor.03E-03 3 Modes 2.4 -1.66E-05 -3.042 0.5 18.229 0.071 0.26E-04 -5.09E-03 2.7 -50.054 0.310 0.38E-04 2 Modes 2.9 2.74E-04 9.09E-03 Error (%) 1 Mode -23.1 -2.125 0.5 -2.260 0.4 -0.6 -1.253 0.1 3.152 0.5 0.76E-03 1.6 4.89E-03 1.265 RHA (all modes) 0.38E-03 2.012 -0.225 0.090 0.0 -10.11E-03 1.263 0.72E-03 3.325 0.192 0.9 3.156 0.01E-04 -2.4 0.062 0.089 0.45E-03 3.097 0.24E-03 2.260 0.124 0.264 Error (%) 1 Mode 2 Modes 3 Modes -23.74E-03 1.226 0.157 0.9 -24.237 0.311 0.76E-03 1.275 0.3 -8.026 0.003 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 0.295 -0.73E-03 3.44E-03 1.008 -0.012 0.227 0.31E-03 Joint Rotation (rad) Combined (RHA) 1 Mode 2.181 0.235 0.197 0.253 0.03E-03 -6.03E-03 1.466 0.6 1.010 -0.226 0.023 0.9 9.0 3 Modes -5.78E-04 -3.259 0.300 0.261 -0.4 -1.080 0.022 0.235 0.097 0.117 0.069 0.33E-03 2.399 0.202 0.401 Error (%) 1 Mode 2 Modes 3 Modes -23.31E-03 Mode 2 -1.7 2.8 -5.0 -2.177 0.3 19.8 -10.265 0.2 -2.2 Peak values of story drift ratios (as % of story height) from RHA for 0.88E-03 2.8 -15.14E-03 2.202 0.1 -14.2 1.002 -0.124 0.203 0.060 0.282 0.88E-03 2.3 -3.231 -0.1 Peak values of floor displacements (as % of building height = 37.2 -20.85E-03 3.4 -6.1 -0.47E-03 1.2 0.7 2 Modes -3.4 -53.29E-03 2.6 -0.199 0.6 11.28E-04 1.7 4.63E-03 2.173 0.8 -56.035 0.321 0.94E-03 2.22E-03 2.179 0.99E-03 Mode 3 3.453 0.400 0.202 0.08E-03 2.125 0.3 -0.50E-03 2.7 3.011 0.311 0.088 -0.060 -0.0 -46.011 0.024 -0.407 0.183 0.89E-03 1.227 0.65E-03 2.01E-04 3.5 -1.245 0. 18 .006 0.2 -57.6 -1.205 0.64E-03 3.159 0.042 0.009 -0.173 0.058 -0.0 -2.043 0.50E-03 4.130 0.25 × 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 0.03E-03 1.7 7.055 0.001 -0.069 0.4 -7.3 -0.1 -19.7 -19.032 -0.378 0.370 0.475 0.266 0.74E-03 1.406 0.0 7.245 0.1 4.259 0.9 1.4 -10.336 0.9 -23.42E-04 1.177 0.00E-03 1.060 -0.09E-03 1.1 -11.266 0.00E-03 1.1 -2.2 -4.282 0.91E-04 1.11E-04 -5.156 0.09E-03 1.6 9.015 -0.4 -3.9 -22.9 -15.152 0.9 -16.56E-03 2.

5.25 × El Centro ground motion 19 . (b) base shear.0435 • 0 (b) V b1 −0.5 15 −0.12 −0.5 9.12 0 (d) 0 (e) 15 Fig.1 −15 −7.00174 • 0 (c) r1 −5 0 5 10 15 Time (sec) 20 25 30 0. (c) joint rotation.1 −15 −7.0435 Vb1 / W 0.15 u (cm) 9.1 5 x 10 −3 θ (rad) 0.5 0 u (cm) r1 7. Response due to first mode: (a) roof displacement. 3.1 0.5 0 u (cm) r1 7.1 /W 0.1 0. (d) force-deformation history.0435 9. and (e) pushover curve.12 • 0 (a) r1 −15 0. Excitation is 0.

1 0. 3.5 5 Fig.6.00244 • 0 (c) r2 −5 0 5 10 15 Time (sec) 20 25 30 0. (b) base shear.5 5 −0.23 −2.0252 −2.5 u (cm) r2 0 • 2.1 −5 −2.5 0 u (cm) r2 2.23 0.23 −5 (a) 0. (d) force-deformation history.0252 • 0 (b) V −0. Excitation is 0.05 5 x 10 −3 b2 θ (rad) 0. (c) joint rotation.25 × El Centro ground motion 20 . and (e) pushover curve.5 0 u (cm) r2 2.05 /W 0.0252 (d) 0 0 (e) −0.1 −5 −2.1 Vb2 / W 0. Response due to second mode: (a) roof displacement.

000913 • 0 (c) r3 −2 0 5 10 15 Time (sec) 20 25 30 0.422 (e) −0. Response due to third mode: (a) roof displacement.02 0.0108 −0.0108 −0.02 /W V b3 0 • 0. 3. and (e) pushover curve.422 −1 (a) 0.5 0 u (cm) r2 0.1 u (cm) r3 0 • 0. (b) base shear. (d) force-deformation history.5 1 Fig.5 1 −0.5 0 u (cm) r3 0. Excitation is 0.02 −1 −0.0108 x 10 −3 (b) −0.02 −1 −0.02 2 θ (rad) 0.02 −0.25 × El Centro ground motion 21 .422 /W V b3 0 (d) 0 −0.7. (c) joint rotation.

3.12).20) with n = 2 and 3 up to roof displacements ur 2 o = 2.4. and the errors in these estimates relative to the exact response from RHA considering all modes. In practical application.6 present estimates of the combined response according to Eq. Observe that the target roof displacement in each pushover analysis is identical to its exact value determined by RHA..6e and 3. or three vibration modes. Figure 3.1 through 3.4.3). 3. 3.1. these pushover curves are consistent with the Vb . 3. 3.5d). These values of the response quantities are identical to the peak response values determined from RHA (Tables 3. and 3. which would provide the Dn value to be substituted in Eq.4. implying that the modal truncation errors are small if at 22 .3.5. respectively. pushing the structure using the force distribution of Eq.4. story drifts.6d and 3. For a fixed number of modes included considered.12 cm.3). this value would be determined directly from the response (or design) spectrum.9 and Tables 3. leads to the pushover curves shown in Figs. and 3. and 3. respectively. the value determined by RHA (Fig.6. Implementing pushover analysis for the second and third modes.18). 3. pushing the structure.. 3. 3. and the computed response values are identical to the peak response values determined from RHA (Tables 3. 3.20) with n = 1 (Fig. and external joint rotations in Tables 3. 3.7d).5. the floor displacements are proportional to the mode shape φ 1 because the structure remains elastic. although both analyses led to identical peak values of the individual modal responses.9).8) leads to the pushover curve shown in Fig. (3.5e. and ur 3o = 0.2.2. As suggested by Eq. In RHA the errors arise only from truncating the responses due to higher modes.3 Modal Pushover Analysis Implementing MPA for the fundamental vibration mode.5. and external joint rotations computed by pushover analysis are presented in Tables 3.422 cm . i. and 3. except for the algebraic sign associated with Γ1 and minor round-off errors.ur relations determined by RHA (Figs.10 and Tables 3. (3. using the force distribution of Eq. As for the first mode. and it is apparent in the example considered that three modes provide most of the response (Fig.3).e. 3. (3. 3. considering one. respectively. two.4) to roof displacement ur1o = 9. story drifts.6.23 cm . (3. This pushover curve is consistent with the relationship between the base shear and roof displacement determined by RHA (Fig. (3. the errors in the MPA results are generally larger than in RHA (Fig.e. and 3. The floor displacements. i.7e and to the floor displacements.21). 3.1. confirming that MPA gives the exact values of the individual modal responses.

least three modes are included. Additional errors are introduced in pushover analysis due to the approximation inherent in modal combination rules.1 0.eps fig3_9b.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. (a) Roof Displacement 15 (b) Top Story Drift 3 Mode 1 ∆ (cm) u (cm) 9. sec 25 30 Fig.59 • Total (All Modes) 0 −15 0 ∆ (cm) u (cm) 0 −3 0 fig3_8.5 Ground 0 0.48 −3 3 1. sec 25 30 r 5 10 15 20 Time.5 0.25 × El Centro ground motion 23 .03 −3 3 Mode 3 ∆ (cm) u (cm) r3 −15 15 9.685 • −3 3 Mode 2 ∆ (cm) u (cm) • 2.3 0.6 Fig.8.4 Displacement/Height (%) 0.1 0.eps r 5 10 15 20 Time. Response histories of roof displacement and top-story drift from RHA for 0. 3.eps Ground 0 0.2 0.12 • r1 −15 15 0 r1 0 0 0.9.2 0. 3.83 • r3 0 • 0.3 0. Heightwise variation of floor displacements and story drifts from RHA for 0.4 Story Drift Ratio (%) 0.422 r2 0 • 1.23 r2 −15 15 0 • 0.

276 0.73E-05 3.76E-03 -1.73E-03 3.282 -0.069 0.264 Error (%) Modal Response 1 Mode 2 Modes 3 Modes -23.90E-03 1.7 24 .4 Peak values of floor displacements (as % of building height = 37.72E-03 3.2 -11.310 0.260 0.203 0.179 0.7 -15.8 -56.9 -24.062 0.63E-03 2.173 0.7 -50.124 0.032 0.53E-04 -9.33E-04 5.4 -4.4 -5.03E-03 -1.14 m) from MPA for 0.003 -0.133 0.058 -0.28E-03 2.89E-03 1.336 0.009 0.3 -33.097 0.296 -0.7 2 Modes -13.4 -14.4 0.253 0.4 1.9 -15.010 0.4 -2.088 -0.060 -0.012 -0.015 -0.09E-03 Mode 2 1.022 0.80E-04 3.1 -11.012 0.055 0.230 -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 0.332 0.106 0.267 0.235 0.15E-03 2.9 -14.00E-03 2.9 -12.9 -13.04E-03 3.9 -14.40E-04 5.2 -4.43E-04 -1.00E-03 -1.4 -22.1 -19.00E-03 2.062 -0.7 2.267 -0.4 -19.152 0.060 0.63E-03 -2.31E-03 Error (%) RHA (all modes) 2.0 -46.407 0.25 × 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 -0.043 -0.230 0.177 0.157 0.3 1.272 0.0 -16.253 RHA (All Modes) 0.00E-03 2.9 3 Modes -12.237 -0.7 -19.374 0.Table 3.3 -9.045 -0.008 -0.336 RHA (all modes) 0.069 0.001 0.078 0.300 0.023 -0.011 0.197 0.7 -21.042 0.0 -2.38E-03 3 Modes 2.92E-04 -1.156 0.125 0.9 -15.226 0.89E-03 -2.237 0.2 -57.09E-03 1.259 0.038 -0.3 Table 3.0 -18.03E-03 3.4 -9.5 -16.024 0.117 0.09E-04 -3.15E-03 1.74E-03 -1.05E-03 3.2 -0.3 -13.235 -0.227 0.261 -0.74E-04 -6.03E-03 1.125 0.6 -17.4 -53.2 1.44E-03 -1.036 -0.9 -13.22E-04 2.090 0.173 0.286 0.00E-03 3.370 0.8 -22.401 Error (%) 1 Mode 2 Modes 3 Modes -23.328 0.048 0.179 0.121 0.90E-04 -9.260 0.9 -11.94E-03 2.24E-03 -2.282 0.9 -13.321 0.253 -0.4 -11.25 × El Centro ground motion Floor 1st 2nd 3rd 4th 5th 6th 7th 8th 9th Modal Response Mode 1 -2.253 0.270 0.313 0.331 0.38E-03 2.6 -19.285 0.4 -9.4 -7.6 Peak values of joint rotation (radians) from MPA for 0.9 -14.274 0.3 -2.229 0.3 -41.31E-03 2.227 0.079 0.08E-03 2.133 0.3 -19.00E-03 1.42E-04 -1.466 0.080 0.3 -14.203 0.106 0.2 -16.203 0.09E-03 -1.74E-03 1.65E-03 2.0 -10.9 -13.89E-03 2.3 -12.22E-03 -2.071 0.245 0.03E-03 6.8 -23.267 0.9 -16.270 0.2 -12.8 -15.6 -15.0 -0.57E-03 1 Mode -23.4 -4.260 -0.1 -11.322 0.1 -18.12E-03 1.177 0.203 0.78E-04 2 Modes 2.006 -0.156 0.002 0.9 -8.026 0.99E-03 Mode 3 -3.2 Table 3.203 0.76E-03 1.31E-03 Joint Rotation (rad) Combined (MPA) 1 Mode 2.96E-03 2.97E-03 1.133 -0.47E-03 1.1 -0.6 -16.9 -15.5 Peak values of story drift ratios (as % of story height) from MPA for 0.89E-03 2.89E-03 -1.181 0.270 0.229 0.011 -0.9 -15.8 -15.231 -0.157 0.2 -20.8 -14.245 -0.4 -22.097 0.042 0.385 0.9 -18.048 0.266 0.152 0.259 -0.44E-03 3.

5 Ground 0 0.4 Story Drift Ratio (%) 0. Heightwise variation of floor displacements and story drifts from MPA for 0.2 0.4 Displacement/Height (%) 0. shading indicates modal combination error 25 .eps fig3_10b.10.eps Ground 0 0. 3.3 0.1 0.5 0.6 Fig.(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.1 0.25 × El Centro ground motion.2 0.3 0.

26 .

Therefore. Eq.1) With this generalization for inelastic systems. sign u = . Both systems have the same mass and damping.4 4. but depend on the history of the displacements: f s = f s u. the initial loading curve is idealized as bilinear.1) becomes mu + cu + f s u. the relations between lateral forces f s at the N floor levels and the lateral displacements u are not single valued. Thus.3) . (4.1 Inelastic Multistory Buildings RESPONSE HISTORY ANALYSIS For each structural element of a building. and the unloading and reloading curves differ from the initial loading branch. leading to the “exact” nonlinear response history analysis (RHA).mι ug t b (4.2) to the modal coordinates of the corresponding linear system.1) is not valid for inelastic systems. 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). (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. Although classical modal analysis (Section 3. 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.2) The standard approach is to solve directly these coupled equations. sign u b g g bg (4. it is useful for later reference to transform Eq.

The equations governing the response of the inelastic system to p eff. This approximate RHA procedure is the preliminary step in developing a modal pushover analysis procedure for inelastic systems.” implying that other “modes” will also contribute to the solution.Gn ug t Mn bg n = 1. (4.3) will. (4. (3.… N (4. and using the mass. However. Eq.4) where the only term that differs from Eq. sign u n b g b g bg (4.3) in Eq.2). Simultaneously solving these coupled equations and using Eq. 2. (4.4) represents N equations in the modal coordinates qn .6) The solution of Eq. premultiplying by φ T .sn ug t bg b g bg bg (4. For linear systems. The spatial distribution s of the effective earthquake forces is expanded into the modal contributions sn according to Eq.6b) are mu + cu + f s u. however.Substituting Eq. (3. give the same results for u t as obtained directly from Eq. 4. (3.2). where φ n are now the modes of the corresponding linear system.5) This resisting force depends on all modal coordinates qn t .6) for inelastic systems will no longer be described by Eq.n t given by Eq. these equations are coupled for inelastic systems. Unlike Eq. Equation (4.2). (4. (4. (3. in principle.8) because qr t will generally be nonzero for “modes” other than the nth “mode.4)] leads to the uncoupled modal response history analysis (UMRHA) procedure.2 UNCOUPLED MODAL RESPONSE HISTORY ANALYSIS bg Neglecting the coupling of the N equations in modal coordinates [Eq. implying coupling of modal coordinates because of yielding of the structure.9) involves T Fsn = Fsn qn . (4. qr t = 0 for all 28 bg .and classical n damping-orthogonality property of modes gives F qn + 2z nw n qn + sn = . (4. sign u = .9) for linearly elastic systems. (3.4) is rarely solved because it offers no particular advantage over Eq. sign qn = φ n f s u n .3). (4.

(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

575 5 10 15 20 Time (sec) 25 30 ur3 (cm) 0 • 5. Fsn / Ln ¨ g(t ) u (a) Static Analysis of Structure (b) Dynamic Analysis of Inelastic SDF System Fig. Roof displacement due to p eff.4.9 • 0 −50 10 −50 10 n=3 ur3 (cm) 0 • 5.112 5 10 15 20 Time (sec) 25 30 −10 0 −10 0 Fig.89 50 r1 0 u (cm) 0 • 78. and 3. 4. and (b) approximate solution by UMRHA bg af bg 32 .51 • ur2 (cm) 0 ur2 (cm) n=3 14.Forces sn Unit mass An(t ) st rn ωn.s n u g t . where u g t = 3.02 −150 n=2 −150 50 n=2 14.0 ¥ El Centro ground motion: (a) exact solution by NL-RHA.3.n t = . n = 1. 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. ζn. 4. 2.

008 • 6.817 5 10 15 20 Time (sec) 25 30 −10 0 −10 0 Fig. respectively. 4.956 5 10 15 20 Time (sec) 25 30 ∆r3 (cm) 0 • 5.33 0 −20 20 −20 20 n=2 ∆r2 (cm) 0 ∆r2 (cm) 7. 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.(a) Nonlinear RHA 20 n=1 20 (b) UMRHA n=1 • 5. both for 3. Top story drift due to p eff. but even for this very intense excitation. The errors are slightly larger in drift than in displacement.s n u g t .2.4) and (4. the errors in either response quantity are only a few percent.6) by UMRHA is compared with the “exact” solution by nonlinear RHA.n t (n = 1.4 and 4. 2.0 ¥ El Centro ground motion: (a) exact solution by NL-RHA.2… N ) according to Eq. 4. and (b) approximate solution by UMRHA af bg bg 4. Such comparison for roof displacement and top-story drift is presented in Figs. (4. and (3) the Fsn Ln .5. n = 1.0 ¥ El Centro ground motion.5)] is neglected. and 3.Dn relation is 33 bg bg . where u g t = 3.744 • 0 −20 10 n=3 −20 10 n=3 ∆r3 (cm) 0 • 5. (2) the superposition of responses to p eff. this intense excitation was chosen to ensure that the structure is excited well beyond its linear elastic limit.855 ∆r1 (cm) 0 ∆r1 (cm) n=2 • 6. (3.n t = .1 Underlying Assumptions and Accuracy The approximate solution of Eq. (4.15) is strictly valid only for linearly elastic systems.5.

approximated by a bilinear curve to facilitate solution of Eq. When implemented by commercially available software.Dn relation? The two sets of forces and displacements are related as follows: V Fsn = bn Γn Dn = urn Γ nf rn 34 (4. However. Therefore we impose this constraint in developing the UMRHA procedure in this section and modal pushover analysis in the next section. which is different than the Fsn Ln . within the linearly elastic range of the structure. (4. 4. How to convert this Vbn .20) to some bg pre-determined roof displacement. At the yield point.8) governing Dn t is based on Eq. this distribution seems to be a rational choiceeven for inelastic systemsto determine Fsn in Eq. (3. The overall errors in the UMRHA procedure are documented in the examples presented in Section 4.12) for floor displacements. Therefore.Dn relation to be determined in Eq. (4.1.2 Properties of the nth-“mode” Inelastic SDF System How is the Fsn Ln . such nonlinear static analysis provides the so-called pushover curve.4.urn pushover curve to the Fsn Ln . is plotted against roof displacement urn .8) in UMRHA.Dn curve. (4.9) should be determined by nonlinear static analysis of the structure as the structure undergoes displacements u = Dnφ n with increasing Dn . when specialized for linearly elastic systems it is identical to the RHA procedure of Section 3. Although several approximations are inherent in this UMRHA procedure.6a. the only force distribution that produces displacements proportional to f n is given by Eq. it can conduct a force-controlled nonlinear static analysis with an invariant distribution of lateral forces. (4. (3. A bilinear idealization of this pushover curve for the nth-“mode” is shown in Fig.8) before it can be solved? Because Eq. the base shear is Vbny and roof displacement is urny . (3. 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.9). 4.10) . (4. the relationship between lateral forces f s and Dn in Eq. The structure is pushed using the force distribution of Eq.20). and the base shear Vbn . Although most commercially available software cannot implement such displacement-controlled analysis.2.

35 . Section 13.8).6b is w 2 .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.11) * in which M n = Ln Γn is the effective modal mass (Chopra.11).2. 4. the initial slope of the pushover curve in Fig.6. Properties of the nth-“mode” inelastic SDF system from the pushover curve Equation 4.13) This value of Tn .5).6b. 4.10 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 (4.12) implying that the initial slope of the curve in Fig.6a is 2 kn = ω n Ln . (4. should be used in Eq. In contrast. 4. where the yield values of Fsn Ln and Dn are Fsny Ln = Vbny * Mn Dny = urny Γ nf rn (4. The two are related through Fsny Ln 2 = w n Dny (4. which is not a meaningful quantity. 2001. which may differ from the period of the corresponding linear system. 4. Knowing Fsny Ln and Dny from n Eq.

Consider a nonlinear static analysis of the structure subjected to lateral forces distributed over the building height according to s* [Eq. (3.3 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. Typically. 4. 4. (3.n t .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 . 2. r o . Idealize the pushover curve as a bilinear curve (Fig. ωn .2. Calculate the peak value. 7. Combine the “modal” responses using Eqs.6b) by utilizing Eq. Compute the deformation history. 6. An t . the peak value of Dn t . with the structure is pushed to the roof displacement urno . (3. φn .20)]. which is summarized next as a sequence of steps. This value of the roof displacement is given by Eq. (3. is now determined by 36 bg bg .21) where Dn .4.20)].13).15) and (3.6b. 8.11). of the total response r t obtained in Step 8. Compute the natural frequencies. of the nth-“mode” inelastic SDF system (Fig. Convert the idealized pushover curve to the Fsn Ln − Dn relation (Fig. Dn (t ) . Calculate histories of various responses by Eqs. and pseudo-acceleration history. for linearly-elastic vibration of the building. and modes. 4. 4. details are available in Appendix A: 1.12) and (3. 4. the first two or three modes will suffice.3b) with force-deformation relation of Fig. For the nth-mode. 5.16) to determine the total response. Repeat Steps 2 to 6 for as many modes as required for sufficient accuracy. n (3. develop the base-shear – roof-displacement ( Vbn − urn ) pushover curve for the force distribution s* [Eq. 9. (4.6a).

. However. as mentioned earlier.3) because.6b by solving Eq. 2001.2. Calculate the peak roof displacement urno associated with the nth-“mode” inelastic SDF system from Eq. This pushover analysis. although somewhat intuitive for inelastic buildings.18). At this roof displacement. 5. it can be determined from the inelastic response (or design) spectrum (Chopra.12). or from the inelastic response (or design) spectrum. Sections 7. bg bg bg bg Thus. for elastic systems.21). Dn . to obtain an estimate of the peak value ro of the total response. details are available in Appendix B. governed by Eq. As shown in Sections 3. (3.solving Eq. (4.2. we will refer to rno as the peak “modal” response even in the case of inelastic systems. plastic hinge rotations.3. are combined using an appropriate modal combination rule. alternatively.8). of the nth-“mode” inelastic SDF system (Fig. etc. (4. the lateral force distribution used possesses two properties: (1) it appears to be the most rational choice among all invariant distribution of forces.g.4. rno also represents the exact peak value of the nth-mode contribution rn t to response r t . Compute the peak deformation.6 and 7.3b) with force-deformation relation of Fig. This application of modal combination rules to inelastic systems obviously lacks a theoretical basis. 4. and (2) it provides the exact modal response for elastic systems. each determined by one pushover analysis. Steps 1 to 4 of the MPA are same as those for UMRHA. e.1 Summary The peak inelastic response of a building to earthquake excitation can be estimated by the MPA procedure just developed. It provides results for elastic buildings that are identical to the well-known RSA procedure (Section 3. the pushover analysis provides an estimate of the peak value rno of any response rn t : floor displacements.3.2 and 3. which is summarized next as a sequence of steps. 4. The peak “modal” responses rno . 6. Eq. The response value rno is an estimate of the peak value of the response of the inelastic system to p eff. it seems reasonable because it provides results for elastic buildings that are identical to the well-known RSA procedure described in Section 3.6). joint rotations. story drifts. (4.8). 37 . seems reasonable. (3. 4.n t . as described in Section 4.

4. is presented next. also included are the combined responses due to one.4. the peak roof displacement due to each of the three “modes” is ur1o = 48. subtract the yield value of hinge rotation to determine the hinge plastic rotation. and compared with the results of a rigorous nonlinear RHA using the DRAIN-2DX computer program.18).5 ¥ the El Centro ground motion including the response contributions associated with three “modal” inelastic SDF systems. and three “modes”.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.0. two.0 to 3. the combined response due to three “modes”. This deficiency also implies that. 4. and three modes are compared with the “exact” values in Fig. in contrast to modal analysis (Section 3. although the trends are not systematic as when the system remained elastic (Section 3. To ensure that this structure responds well into the inelastic range the El Centro ground motion is scaled up by a factor varying from 1.7 cm. the 38 . and the errors in the approximate results are shown in Fig. Typically. The peak values of floor displacements and story drifts including one.1 Uncoupled Modal Response History Analysis The structural response to 1. the UMRHA procedure lacks a rigorous theory. (3. and the “exact” response from nonlinear RHA for the roof displacement and top-story drift.53 cm. respectively. three “modes” included. extract from the pushover database values of other desired responses. This is to be expected. The peak values of response are as noted.3 cm. Figure 4. the “exact” results.1).9. 4. the first two or three “modes” will suffice. say. Observe that errors tend to decrease as response contributions of more “modes” are included. in particular. 8. two. From the total rotation of a plastic hinge.7 shows the individual “modal” responses. Repeat Steps 3 to 8 for as many “modes” as required for sufficient accuracy. and the percentage errors in the approximate results. 9.2. ur 2o = 11. and ur 3o = 2.4. with. At urno . Determine the total response by combining the peak “modal” responses using the SRSS combination rule of Eq.1 and 4.2). determined by the UMRHA procedure. rno .8. 4.7. The peak values of displacements of all floors and drifts in all stories are presented in Tables 4.

1 and 4.7 • −80 80 ∆r1 (cm) u (cm) "Mode" 1 −12 r2 0 ∆r2 (cm) 12 0 u (cm) −80 80 −12 r3 0 • 2.2). the errors in story drifts are larger compared to floor displacements.2) and 1. (a) Roof Displacement (b) Top Story Drift 80 12 0 • 3. 4.1 and 3.2).24 • NL−RHA r −12 12 ∆r (cm) u (cm) 0 5 10 15 20 Time. in particular.88 u (cm) "Mode" 3 −80 80 0 −80 80 0 −80 0 • 44. and total response from NL-RHA 39 . Next. Response histories of roof displacement and top-story drift due to 1. 2. 0. 0.5.85.6 5 10 15 20 Time. sec 25 30 r −12 0 Fig.0.1 and 4.7.5 × El Centro ground motion: individual “modal” responses and combined response from UMRHA. the degree to which the system deforms beyond its elastic limit.2) if the system yields significantly versus if the system remains within the elastic range (Tables 3. the errors in responses computed by UMRHA including three “modes” relative to the “exact” response were determined.response is much less accurate (Tables 4.53 ∆r3 (cm) 12 0 • 2.25. For this purpose the UMRHA and exact analyses were repeated for ground motions of varying intensity.38 6.62 5. sec 25 30 • 48. recall that the computed errors have been presented earlier for ground motion multipliers 0.1 −12 12 ∆r (cm) u (cm) UMRHA 3 "Modes" 0 • 7.0. However. 1.44 • "Mode" 2 r1 0 • 48. For each excitation.1 and 3.75. and 3. we investigate how the errors in the UMRHA vary with the deformation demands imposed by the ground motion. for a fixed number of “modes” included.25 (Tables 3. 0.5.0. 1.5 (Tables 4. defined as the El Centro ground motion multiplied by 0. just as for elastic systems.3 11.

4.20)] for the first three n modes of the system are shown in Fig.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. Heightwise variation of floor displacements and story drift ratios from UMRHA and NL-RHA for 1.10a).(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. 4. (3. Two versions of the pushover curve are 40 . For this purpose. the pushover curves using force distributions s* [Eq.5 Displacement/Height (%) 2 Ground 0 0. it will be useful to know the deformation of the system relative to its yield deformation. Shown is the error in each floor displacement (Fig.10b). in each story drift (Fig.9. Heightwise variation of error in floor displacements and story drifts estimated by UMRHA including one. and the error envelope for each case.5 Story Drift Ratio (%) 2 2. 4. two. indicated by a ground motion multiplier. To interpret these results. or three “modes” for 1. 4.5 ¥ El Centro ground motion Figure 4.10 summarizes the error in UMRHA as a function of ground motion intensity. with the peak displacement of each “modal” SDF system noted for each ground motion multiplier.11.5 1 1.5 1 1.8. 4.5 Fig.

75.11a).3). (3.4.53 cm. Additional errors are introduced in UMRHA of systems responding beyond the linearly elastic limit for at least two reasons. 11. determined from “exact” analyses.2. 41 . The location of plastic hinges and their rotations.2. First. 3. the values determined by RHA of the nth-mode inelastic SDF system (Fig.4) to roof displacements urno = 48. and 2.2 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. (3) maintain roughly similar values for more intense ground motions.5 ¥ the El Centro ground motion. as mentioned in Section 4. Second. and (4) are larger in story drift compared to floor displacements. The idealized curve for the first “mode” deviates most from the actual curve near the peak displacement corresponding to ground motion multiplier 1. UMRHA lacks a rigorous theory and is based on several approximations.11). and.7 cm. the first reason mentioned above seems to be the primary source for the errors.7). even though the system remains essentially elastic.10) for each “modal” inelastic SDF system (Figs 4. were noted but are not shown here.included: the actual curve and its idealized bilinear version.3 cm. For more intense excitations.4). The structure is pushed using the force distribution of Eq. and plastic hinge rotations at the external beam end at each floor level (Table 4. as mentioned in Section 3.0. 2. the peak values of displacements at all floors (Table 4. drifts in all stories (Table 4. and 3 (Fig.01 (Fig. the errors in truncating the higher mode contributions are negligible. Perhaps this explains why the errors are large at this excitation intensity.0.4. the pushover curve for each “mode” is idealized by a bilinear curve in solving Eq. (2) increase rapidly as the ground motion multiplier increases to 1. 4.1).20) with n = 1. 4. Each of these three pushover analyses provides the pushover curve (Fig.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. (4. respectively. The system remains elastic up to ground motion multiplier 0. 4.75. the ductility factor for the first mode system is only 1. 4.6 and 4.5). Figure 4.

426 -1.863 0.983 1.5 3 Fig.14 m) from UMRHA for 1.248 Error (%) NL 3 1 2 3 RHA “Mode” “Modes” “Modes” “Modes” 0.202 11.806 0.096 0.490 -1.5 28.104 0.707 1.478 0.819 2.938 1.914 -0.187 -0.856 2.914 2.376 1.5 3 0 0 0.379 1.945 -37.844 -25.3 25.370 -0.9 12.7 31.138 1.473 -22.9 31.071 -0.216 1.372 -1.226 -0.5 10.5 9. 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 0.820 -19.2 12.338 -1.2 Peak values of story drift ratios (as % of story height) from UMRHA for 1.2 4.120 1.971 1.003 0.235 -0.(a) Floor Displacements 60 50 40 Error Envelope Error for Floor No.2 1.133 1.1 Peak values of floor displacements (as % of building height = 37.256 1.169 0.130 0.616 -0.256 1.10.293 1.5 9.5 × 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 -1.575 -41.490 1.5 1 1.201 1.554 1.2 6.079 0. and (b) story drifts Table 4.5 -3.366 -0.8 1.751 1.5 18.9 5.900 -10.8 0. 4.3 0.495 1.018 0.4 -7.410 -1.033 0.8 1.376 -1.214 0.373 -0.811 1.430 1.938 -1. Errors in UMRHA as a function of ground motion intensity: (a) floor displacements.9 16.298 0.0 2.283 1.4 1.2 1.049 -0.6 2.5 1 1.260 -15.371 -0.220 -0.214 -0.2 22.982 9.513 -0.763 -15.1 0.338 1.3 6.0 7.057 -0.072 -1.5 × 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” -0.8 1.806 -0.1 3.0 9.055 -0.8 1.540 0.4 -1.088 10.068 0.072 1.135 9.698 1.942 -0.9 31.4 4.5 2 GM Multiplier 2.200 8.009 -0.1 1.3 42 .372 1.0 1.220 0.1 8.003 -31.121 -0.5 2 GM Multiplier 2.663 0.7 14.1 1.366 0.407 -10.9 12.3 8.154 0.6 4.044 1.820 -0.5 28.8 1.772 Error (%) NL 3 1 2 3 RHA “Mode” “Modes” “Modes” “Modes” 1.201 -0.0 -9.863 1.065 0.722 0.8 14.410 1.472 1.126 0.291 0.513 0.010 0.3 1.333 0.298 -0.315 -0.241 -1.676 0.350 -0.852 1.070 1.0 11.136 1. Noted 60 50 40 (b) Story Drifts Error Envelope Error for Story No.668 -23.526 -0.317 0.201 -1.877 0.663 -0.0 11.371 0.7 Table 4.727 1.071 0.256 -1.241 1.6 1.942 1.484 0.

overestimated by up to 18% in the middle stories. 4.3 and 4. considering one. Furthermore.5 present estimates of the combined response according to Eq.” and nonlinear RHA. The first “mode” alone is inadequate. Pushover analysis seems to be inherently limited in computing accurately hinge plastic rotations. 4.18). but was more successful when the excitation was more intense (results are not included). 4. 4. 4.4). MPA failed to identify the plastic hinges at the column bases in Fig. significant improvement is achieved by including response contributions due to the second “mode”. The errors are especially large in the hinge plastic rotations estimated by the MPA procedure.9 and Tables 4.Figures 4.12 and Tables 4.4 with Tables 4.3 through 4.13 and Table 4. the hinges identified in beams at the 6th floor were at variance with the “exact” results. were determined by four analyses: MPA considering one “mode. significantly smaller than in UMRHA (compare Fig. the errors in the modal pushover results are.2). As shown in Figs.3 and 4. The drifts are underestimated by up to 13% in the lower stories. 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. especially in estimating the story drifts (Fig.14. 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. even if three “modes” are included (Fig.14.3 and 4.” two “modes.3 and 4.1 and 4.13 with Fig. the results were not always accurate. and the errors in these estimates relative to the exact response from nonlinear RHA. For example.” and three “modes. (3. for two or three modes included. 4. 4. and three “modes. this error is not especially significant because the hinge plastic rotation is very small. However. two. the third “mode” does not contribute because this SDF system remains elastic (Fig. Obviously.13 and Tables 4.4).11c). The second “mode” was necessary to identify hinges in the upper stories.12 and Tables 4. but the third “mode” contributions do not seem especially important (Fig.” respectively. 4.5). and are within a few percent of the exact values for the upper stories. 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. however. Fortuitously.12 and 4.4. in general. 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%. The locations of plastic hinges shown in Fig.13 and Tables 4. although the error is recorded as 100% if MPA estimates zero rotation when the nonlinear RHA computes a non-zero value. 4. 43 .

5 0. 1.5 1 0. 1.5.0 × El Centro ground motion 44 . 2. V = 4952 kN. α = 0. 0. “Modal” pushover curves with peak roof displacements identified for 0.85 0.9 cm.13 y by Base Shear (kN) 8000 6000 1 1.75 0.14 y by Base Shear (kN) 8000 6000 4000 1. and 3.75. α = 0.25 5 10 15 20 Roof Displacement (cm) 25 (c) "Mode" 3 Pushover Curve 12000 10000 u = 4. α = 0.0. 4.85 Base Shear (kN) 8000 6000 4000 2000 0 0 0.5 2 3 4000 2000 0 0 Actual Idealized 0.25.19 y by 3 2 1.75 Actual Idealized 20 40 60 Roof Displacement (cm) (b) "Mode" 2 Pushover Curve 80 12000 10000 u = 9.5 0.6 cm.25 0.(a) "Mode" 1 Pushover Curve 12000 10000 u = 36.5 0.0.85 0.11.5 2 3 Actual Idealized 2000 0 0 0. V = 7616 kN.75 0.2 cm.25 1 2 4 6 Roof Displacement (cm) 8 10 Fig.5. V = 5210 kN. 0.

identified in Section 3.15) for ground multipliers larger than 1. 4. the errors in MPA were larger for ground motion multipliers less than 0. In this case. the MPA procedure is equivalent to the RSA procedure. and the error envelope for each case. 4. 45 .2 and 3. implying excitations intense enough to cause significant yielding of the structure. as discussed in Sections 3.5 1 1.0. whereas MPA contains errors inherent in modal combination rules. each story drift (Fig.10 and 4. Shown is the error in each floor displacement (Fig. implying that the modal combination errors contained in these procedures are acceptable.4 also apply to MPA. now standard in engineering practice. However. While various sources of errors in UMRHA.5 Story Drift Ratio (%) 2 2.5 Fig. 4.15b).3 for elastic systems (or weak ground motions). 4. implying excitations weak enough to limit the response in the elastic range of the structure. the errors in MPA were fortuitously smaller than in UMRHA (compare Figs.12 Heightwise variation of floor displacements and story drift ratios from MPA and NL-RHA for 1.4.15). The errors are only weakly dependent on ground motion intensity (Fig.3. UMRHA is essentially exact. 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.(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 0.5 1 1. indicated by a ground motion multiplier.75.4. shading indicates errors in MPA including three “modes” Figure 4. As mentioned in Section 3.15 summarizes the error in MPA considering three “modes” as a function of ground motion intensity. an observation with practical implications.5 Displacement/Height (%) 2 Ground 0 0. 4.15a).5 × El Centro ground motion.

(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 three “modes” for 1. 4. two. story drifts.13. and hinge plastic rotations estimated by MPA including one. Errors in floor displacements.5 ¥ El Centro ground motion 46 .

980 -0.3 11.00E+00 0.8 1.72E-03 7.8 -29.244 0.02E-03 3.053 -1.8 0.640 -1.298 -0.222 0.581 0.36E-03 6.009 0.02E-03 0.407 -27.6 13.1 46.76E-03 4.00E+00 0.154 0.053 1.14 m) from MPA for 1.8 1.00E+00 0.5 1.478 0.9 -100.5 × El Centro ground motion Hinge Plastic Rotation (rad) “Modal” Response Combined (MPA) “Mode” 1 7.55E-03 3.733 1.8 -12.575 -53.220 1.1 1.55E-03 3.00E+00 0.233 1.763 -14.101 -0.435 0.1 -8.503 1.756 0.50E-10 3.5 10.652 1.266 -0.705 -1.02E-03 3.756 0.5 2.60E-03 2.156 -0.88E-03 0.10E-02 9.5 × 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 0.088 12.298 0.00E+00 0.0 -100.7 -12.911 0.426 15.1 62.6 -44.3 1.8 -6.414 1.518 1.8 -6.614 0.2 1.728 1.1 62.3 Peak values of floor displacements (as % of building height = 37.057 0.066 -0.9 -4.071 -0.517 1.1 13.00E+00 0.399 0.2 -100.7 1.105 0.200 8.4 Peak values of story drift ratios (as % of story height) from MPA for 1.202 8.351 -0.331 1.745 1.9 7.135 -7.0 -5.033 -0.9 5.737 0.4 1.527 -0.00E+00 3.514 -1.36E-03 6.1 46.00E+00 0.5 2.5 -6.9 0.5 Table 4.8 -29.900 -0.304 -1.018 -0.694 1.2 -100.72E-03 7.820 -7.8 -29.429 0.0 15.007 1.2 0.594 -1.666 Table 4.516 0.9 -100.5 Peak values of hinge plastic rotations (radians) from MPA for 1.00E+00 0.338 1.18E-03 7.071 0.6 13.3 13.6 -7.118 0.8 17.015 0.0 -50.00E+00 0.130 0.37E-04 Error (%) “Modal” Response 1 “Mode” -32.473 -15.007 1.168 -0.233 1.26E-04 9.00E+00 0.372 0.00E+00 0.00E+00 2 “Modes” 7.00E+00 1 “Mode” 7.76E-03 4.305 -0.640 1.2 -100.0 -100.60E-04 7.2 0.209 1.304 1.0 -50.015 0.00E+00 0.982 13.6 -8.315 0.879 1.250 0.687 0.00E+00 0.844 -7.9 2.667 -1.498 1.2 6.88E-03 1.116 1.00E+00 0.5 7.683 1.336 Error (%) NL 3 1 2 3 RHA “Mode” “Modes” “Modes” “Modes” 0.781 0.222 0.003 -16.55E-03 3.99E-03 6.260 -14.781 0.36E-03 6.0 47 .2 -3.5 10.018 0.00E+00 0.6 -44.668 -13.36E-03 6.8 -4.066 -0.02E-03 0.190 -0.6 7.055 0.0 1.72E-03 7.1 46.9 1.1 18.00E+00 Floor 1st 2nd 3rd 4th 5th 6th 7th 8th 9th “Mode” 2 0.068 0.6 -9.116 1.5 2.0 3 “Modes” -32.37E-03 1.76E-03 4.581 0.6 1.3 -3.0 -100.259 1.2 -4.2 11.049 -0.Table 4.737 1.503 -1.8 7.399 0.05E-03 2.197 -0.018 -0.00E+00 0.910 1.5 × 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” 1.37E-03 1.00E+00 0.220 1.945 -49.738 1.0 2 “Modes” -32.895 1.752 1.00E+00 3 “Modes” 7.942 6.311 0.4 -8.00E+00 “Mode” 3 0.76E-03 4.012 1.00E+00 0.371 -0.641 1.1 62.125 -1.53E-03 7.37E-03 1.980 0.00E+00 0.22E-10 NL RHA 1.37E-03 1.9 1.3 -3.705 1.176 0.667 1.611 0.7 1.72E-03 7.804 1.19E-10 3.895 1.116 1.8 “Mode” 3 -1.88E-03 0.

4. and three “modes.17 and 4. Figures 4.0 cm.20) with n = 1 (Fig. Results of such analyses for the selected building subject to 1.6 and 4.7). 4. 4.7 cm and ur 3o = 2.18) considering one.6 through 4.18 and Tables 4.4.” these effects will modify the combined modal response as well as the results of nonlinear RHA.18 and Tables 4.” however.” and the errors in these estimates relative to the exact response from nonlinear RHA. 4.3 Modal Pushover Analysis with Gravity Loads To evaluate the accuracy of the dynamic response of the system. Starting with its initial state under gravity loads. Significant improvement is achieved by including response contributions due to the second “mode.53 cm.17 and Tables 4.18 and 48 . the third “mode” contributions do not seem especially important (Fig.5). (3.” but not the contributions of higher “modes. The response contributions of the second and third modes are the same as before (Tables 4.7).4.4) to roof displacement ur1o = 52. The story drifts are underestimated by up to 7% in the lower stories.7. drifts in all stories (Table 4.11a).7).8). The pushover curves for the second and third modes included in Fig. 4. but it is inadequate in estimating the story drifts (Fig.16 are unchanged. They are now included in the pushover analysis of the structure for the first “mode” only. 4. two. (3. as are the roof displacements ur 2o = 11. 4.5 × El Centro ground motion are presented below.16a—which is slightly different than the one excluding gravity loads (Fig. The errors are especially large in the plastic hinge rotations estimated by the MPA procedure even if three “modes” are included (Fig. and overestimated by up to 29% in the middle stories and by up to 13% in the upper stories. the results presented so far did not include gravity load effects.6 and 4.8 present estimates of the combined response according to Eq.6 and 4. The first “mode” alone provides adequate estimates of floor displacements. 4. the structure is pushed using the force distribution of Eq. 3. MPA including two “modes” underestimates the displacements of lower floors by up to 6% and overestimates the upper floor displacements by up to 22%. Also presented at these roof displacements are the displacements at all floors (Table 4. resulting in the pushover curve shown in Fig. As shown in Fig. (3.17 and Tables 4. and plastic hinge rotations at selected external beam end at each floor level (Table 4.20) with n = 1 to a target roof displacement. Static analysis of the structure for gravity loads provides the initial state—forces and deformations—of the structure. and the structure is pushed using the force distribution of Eq. Obviously gravity load effects will influence the seismic demands due to the first “mode.6).3 .

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.Table 4. and (b) story drifts 49 .5 2 GM Multiplier 2.5 1 1.5 2 GM Multiplier 2. and three “modes” and by NL-RHA for 1.5 3 Fig. two. Locations of plastic hinges determined by MPA considering one. 4. (a) MPA. Most pushover analysis procedures do not seem to compute to acceptable accuracy plastic hinge rotations.15.5 3 0.5 × El Centro ground motion (a) Floor Displacements 60 50 40 Error Envelope Error for Floor No. Noted 60 50 40 (b) Story Drifts Error Envelope Error for Story No. 1−"Mode" • • • • • • • • •• •• •• •• •• •• •• •• •• •• •• •• •• •• • • •• • •• • • • • • •• •• • •• •• •• •• (b) MPA.5 1 1.14. 2−"Modes" •• •• •• •• •• •• •• •• •• •• •• •• •• •• • • • • •• • •• (c) 3−"Modes" • • • • • • • •• •• • •• •• •• •• •• •• •• •• •• •• •• •• •• •• •• •• •• •• • • • • •• • •• • • • • • • • • • (d) Nonlinear RHA •• •• • •• •• •• •• • •• •• •• •• •• •• •• • •• •• • •• •• •• •• • • • • • • •• •• •• Fig.8). 4. Errors in MPA as a function of ground motion intensity: (a) floor displacements.

each story drift (Fig. MPA provides response values accurate enough for practical application in building retrofit or design. Figure 4. Shown is the error in each floor displacement (Fig. 4. this procedure is able to predict plastic hinge locations essentially consistent with nonlinear RHA. 50 . and the errors are only weakly dependent on ground motion intensity.” two “modes”. The second “mode” is necessary to identify hinges in the upper stories.20b). With two modes included in MPA.The locations of plastic hinges shown in Fig. indicated by a ground motion multiplier. 4. 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.20 summarizes the error in MPA considering three “modes” as a function of ground motion intensity. and the error envelope for each case. excluding gravity load effects. 4.15. These errors are only slightly larger than those in Fig. 4.20a).19 were determined by four analyses: MPA considering one “mode.

0.75 Actual Idealized 20 40 60 Roof Displacement (cm) 80 (b) "Mode" 2 Pushover Curve 12000 10000 u = 9.75 0.6 cm. 2.5 2 3 Actual Idealized 2000 0 0 0. V = 5210 kN. α = 0.5 2 3 4000 2000 0 0 Actual Idealized 0.5 1 0.(a) "Mode" 1 Pushover Curve 12000 10000 u = 36. “Modal” pushover curves with gravity loads included.5 0.50.25 1 2 4 6 Roof Displacement (cm) 8 10 Fig.25 0. 0. 0.25.85.85 0.75. noted are peak values of roof displacement for 0.0 × El Centro ground motion 51 . V = 4952 kN.75 0. α = 0.5 0.13 y by Base Shear (kN) 8000 6000 1 1. V = 7433 kN. 4.85 Base Shear (kN) 8000 6000 4000 2000 0 0 0. and 3.0. α = 0.14 y by Base Shear (kN) 8000 6000 4000 1.85 0. 1.9 cm.25 5 10 15 20 Roof Displacement (cm) 25 (c) "Mode" 3 Pushover Curve 12000 10000 u = 4.3 cm.5 0.19 y by 3 2 1.16.0.

two. Errors in floor displacements. and 52 . story drifts.18. 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.(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. gravity loads included. and hinge plastic rotations estimated by MPA including one.5 1 1. Heightwise variation of floor displacements and story drift ratios from MPA and NL-RHA for 1.5 Story Drift Ratio (%) 2 2. 4.5 1 1. 4.5 Displacement/Height (%) 2 Ground 0 0.17.5 × El Centro ground motion.5 Fig.

Locations of plastic hinges determined by MPA considering one. 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. 3−"Modes" (d) Nonlinear RHA • • • • • • • •• •• •• •• •• •• •• •• •• •• •• •• •• •• •• •• •• •• •• •• •• • • • • •• •• •• • • • • • • • • •• •• • •• •• •• •• • • •• •• • •• •• •• •• • • •• •• • •• •• •• •• • • • • • • •• •• •• Fig. gravity loads included 53 .5 2 GM Multiplier 2.5 1 1. Errors in MPA as a function of ground motion intensity: (a) floor displacements.5 2 GM Multiplier 2.5 1 1.5 × El Centro ground motion. 1−"Mode" • • • • • • • •• •• •• •• •• •• •• •• •• •• •• •• •• •• •• • • •• •• •• • • • • • •• •• •• •• •• •• •• (b) MPA. and three “modes” and by NL-RHA for 1. Noted 60 50 40 (a) Story Drifts Error Envelope Error for Story No.19. gravity loads included (a) Floor Displacements 60 50 40 Error Envelope Error for Floor No. and (b) story drifts. 4. two.(a) MPA. 4.20.5 3 Fig.5 3 0.

071 0.996 0.2 4.78E-03 1.9 31.19E-03 1.938 1.667 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” 1.319 1.908 1.00E+00 0.2 12.2 4.107 1.927 1.88E-03 1.429 -1.066 -0.214 0.399 -0.4 26.105 0.3 13.754 1.064 -10.2 9.19E-03 1.4 -4.6 19.434 Error (%) NL 3 1 2 3 RHA “Mode” “Modes” “Modes” “Modes” 0.0 -100.4 20.850 -1.783 1.009 0.434 0.257 0.00E+00 0.125 -1.2 1.5 3 “Modes” -32.14 m) from MPA for 1.594 -1.00E+00 “Mode” 3 0.310 1.858 2.813 1.55E-03 3.0 -30.23E-03 0.0 37.913 7.7 Peak values of story drift ratios (as % of story height) from MPA for 1.353 -23.11E-03 9.461 0.311 0.9 -4.6 0.237 0.00E+00 0.8 9.3 1.00E+00 0.4 1.8 0.114 1.5 × El Centro ground motion.673 Table 4.2 21.00E+00 0.88E-03 0.00E+00 0.23E-03 0.00E+00 0.8 28.998 0.5 “Mode” 3 -1.830 -12.478 0.6 2.998 1.00E+00 0.516 0.17E-03 9.8 1.75E-03 0.0 -30.429 1.23E-03 3.6 Peak values of floor displacements (as % of building height = 37.19E-03 1.109 0.1 1.266 -0.190 -0.2 -3.068 0.3 -22.3 -22.071 -0.04E-10 3.2 9.7 19.00E+00 2 “Mode” 8.5 -5.414 28.00E-03 5.037 0.5 1.3 9.033 -0.0 16.0 2 “Modes” -32.11E-03 9.2 -0.23E-02 1.263 0.00E+00 0.996 -0.11E-03 9.3 -22.983 1.4 -4.850 1.00E+00 0.270 -12.888 1.207 18.00E+00 “Mode” 2 0.0 37.1 4.983 1.5 0.831 0.921 1.130 0.7 4.55E-03 3.665 0.00E-03 5.454 1.603 -1.7 -2.821 1.5 10.35E-03 8.199 16.7 1.213 1.514 -1.55E-03 3.00E+00 0.1 21.4 1.00E-03 5.319 1.00E+00 3.728 1.237 0.263 1.049 -0.7 -2.213 1.4 -6.176 0.128 -1.23E-03 3.434 0.35E-03 8.637 0.8 -32.057 0.0 -100.2 2.00E+00 0.5 1.686 -7.908 -1.5 21.3 -100.860 1.213 1.00E+00 0.00E-10 NL RHA 1.330 1.037 -0.055 0.836 -0.3 -100.015 0.26E-03 3.19E-03 1.8 Table 4.00E+00 Error (%) “Modal” Response 1 “Mode” -32.00E+00 0.854 0.637 0.515 -50.101 -0.11E-03 9.00E-03 5.530 1.315 0.19E-04 5.877 -46.098 20.102 1.04E-02 8.744 1.018 -0.00E+00 0.2 16.490 -11.330 1.933 1.8 -32.822 1.527 -0.114 -1.102 1.7 16.603 1.2 4.00E+00 0.8 0.998 21.687 0.2 9.168 -0.953 15.3 -100.530 1.2 0.6 2.154 0.2 -2.Table 4.00E+00 0.5 21.156 -0.0 37. gravity loads included Hinge Plastic Rotation (rad) Floor “Modal” Response “Mode” 1 1st 2nd 3rd 4th 5th 6th 7th 8th 9th 8.5 54 .197 -0.9 -6.854 0.530 1.066 -0.371 -0.1 13.754 0.466 0.831 0.305 -0.00E+00 Combined (MPA) 1 “Mode” 8.00E+00 0.507 1.372 0.35E-10 3.13E-03 5.35E-03 8.35E-03 8.5 × El Centro ground motion.8 Peak values of hinge plastic rotations (radians) from MPA for 1.436 1.88E-03 0. 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” 0.5 × El Centro ground motion.821 -1.351 -0.00E+00 0.6 1.399 0.9 -3.00E+00 3 “Modes” 8.2 1.

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. and 3. and the exponent k = 1 for fundament period T1 £ 0.1 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. 2… N ). 5.5 sec . SRSS distribution: s* is defined by the lateral forces back-calculated from the story shears determined by response spectrum analysis of the structure. Comparison of Modal and FEMA Pushover Analyses 5. assumed to be linearly elastic. 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. The floor displacements.5. Specified in FEMA-273 are three distributions for lateral forces: 1. a plot of base shear versus roof displacement.” and nonlinear RHA. k = 2 for T1 ≥ 2. computed at the target displacement represent the earthquakeinduced demands on the structure.5 sec .. story drifts.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. “Uniform” distribution: s* = m j (where the floor number j = 1. MPA considering three “modes. j 2. plastic hinge rotations. gravity load effects were included in all 55 . etc. joint rotations. 1997] The pushover curve. and varies linearly in between.

00719 0. The MPA procedure is more accurate than all the FEMA force distributions.0896 0. both presented in Section 4.177 0.0913 0.11 0. 0. and (c) SRSS 56 .3.11 0.0654 0.3a and 5.0466 0.0702 0. Using each of these force distributions.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. 5. and the locations of all plastic hinges in Fig. Also included in these presentations are the MPA results considering three “modes” and the “exact” demands from nonlinear RHA.11 0. and the displacements of upper floors are overestimated by 9-22%.4 and Tables 5.3a and Table 5.4.analyses.1.062 0. The “uniform” distribution overestimates all floor displacements by 17-28%.11 0. 5. wherein the first two are obvious and the third is determined from response spectrum analysis of the building (Appendix C). the story drift demands in Fig. The three FEMA force distributions are presented in Fig.3. the value determined from RHA of the first-mode inelastic SDF system for 1. The pushover curves are given in Fig. the floor displacement demands in Fig.042 0.0981 0.11 0.0446 0.4a.4.3. The first four floor displacements are within 5% of the “exact” value.0197 0.112 0.1 through 5. 5. The errors in the FEMA and MPA estimates of seismic demands relative to the “exact” demands are presented in Fig.165 0.3b and Table 5. 5. 5. 5.2. Force distributions in FEMA-273: (a) “uniform”.11 0.11 0.1.119 0.1.0 cm. with the errors being larger for the SRSS distribution. and Table 5.2.21 0.281 0.5 times the El Centro ground motion.126 0.0381 0. 5.367 (a) Uniform (b) ELF (c) SRSS Fig. (b) ELF. Figures 5. pushover analyses are implemented for a target roof displacement of 52. plastic hinge rotation demands in Table 5.

3 cm.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 = 6897 kN. (b) ELF. gravity loads are included 57 . Pushover curves using three force distributions in FEMA-273: (a) “uniform”. 5. V = 7456 kN.(a) Uniform Distribution 12000 10000 u = 33 cm. 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. 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 = 39.8 cm. α = 0.2. and (c) SRSS. α = 0.

and overestimated by no more than 32%. In contrast. reaching 31%. 3 "Modes" FEMA Uniform ELF SRSS 0. reaching 35%.” and NLRHA. reaching 64%. the MPA procedure is more accurate than all the FEMA force distributions.5 Ground 0 Ground 0 Fig.5 Story Drift Ratio (%) 2 2.4c and Table 5.3.) The pushover analysis procedures considered seem incapable of computing accurately local response quantities. For the uniform distribution. larger errors are noted in the upper and lower stories.5 Displacement/Height (%) 2 Floor 5th 5th 4th 3rd 2nd 1st NL−RHA MPA. For the SRSS distribution.3 demonstrate that the hinge plastic rotations estimated by all three FEMA force distributions contain unacceptably large errors.5 1 1. such as hinge plastic rotations. with story drifts under estimated by. at most. with errors reaching 37% in this example. (a) Floor Displacements 9th 8th 7th 6th 9th 8th 7th 6th (b) Story Drift Ratios Floor 4th 3rd 2nd 1st NL−RHA MPA. Figure 5. For the ELF distribution.2 demonstrate that the story drift demands are greatly underestimated by all the FEMA force distributions.5 1 1. Modal pushover analysis procedure gives estimates better than all the FEMA force distributions. 7%. the errors are largest in the lower stories. 3 "Modes" FEMA Uniform ELF SRSS 0. MPA including three “modes. (The 100% error in a 6th floor hinge is ignored as it simply represents that the MPA estimated zero rotation. and Table 5. Heightwise variation of floor displacements and story drift ratios estimated using FEMA-273 force distributions. This seems to be an inherent limitation of pushover analysis.4b.Figures 5. gravity loads included 58 . 5. whereas nonlinear RHA computed an insignificantly small value.3b and 5. but it is still inaccurate. errors are largest in the upper stories.

5 were determined by five analyses: MPA considering the three “modes. 5. and hinge plastic rotations estimated using FEMA-273 force distributions and MPA (including three “modes”). 5.(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. and the three FEMA analyses. and the ELF distribution fails 59 .” nonlinear RHA (“exact”). they are not good indicators of damage. Based on the results presented here. While pushover estimates for floor displacements are more accurate. the SRSS distribution fails to identify yielding of beams in the middle floors. story drifts. the “uniform” distribution fails to identify yielding of the beams above the fourth floor. 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. The locations of plastic hinges shown in Fig.4. 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. The locations of the plastic hinges are not identified correctly by the FEMA force distributions. Errors in floor displacements.

Figures 5.7 summarize the error in FEMA analyses relative to the “exact” demands as a function of ground motion intensity indicated by a ground motion multiplier.to identify yielding in some locations.6 and 5. The MPA procedure is superior to FEMA-273 analyses over the entire range of ground motion intensities considered. and the error envelope for each case. especially in estimating story drifts.” The MPA procedure provides estimates of earthquake demands that are significantly more accurate than all FEMA273 analyses. 60 . The MPA procedure identifies yielding in most locations predicted by “exact” analysis. Included for comparison is the error in MPA with three “modes. but fails to predict yielding in a few locations. Shown is the error in each floor displacement and each story drift.

59E-03 5.938 1.234 1.50E-03 0.8 44.783 1.3 23.00E+00 2.00E-03 5.462 1.00E+00 Error (%) FEMA Uniform ELF SRSS MPA 24.530 1.4 MPA -2.335 2.330 1.323 1.3 31.154 1.4 0.8 Floor 1st 2nd 3rd 4th 5th 6th 7th 8th 9th Uniform 0.23E-02 1.351 0.9 0.209 0.007 1.88E-03 1.6 21.58E-04 6. Peak values of floor displacements (as % of building height = 37.067 0.00E+00 4.417 1.0 -57.310 1.8 16.26E-03 3.560 1.5 -29.4 16.5 22.466 0.5 -33.8 -2.Table 5.399 1.3 -4.984 1.5 -29.6 14.0 -59.927 1.7 -29.436 Error (%) NL Uniform RHA 0.270 0.1 0.888 0.5 15.064 1.611 0.7 FEMA ELF SRSS -22.2 -4.84 1.2 9.13E-03 5.736 0.789 0.830 2.9 16.344 0.5 10.23E-03 3.2 Peak values of story drift ratios (as % of story height) from FEMA force distributions and MPA.4 -55.5 26.858 1.1.2 19.207 1.3 -22.7 -28.7 MPA -2.00E+00 0.975 1.730 1.2 10.9 -77.3 -11.0 -100.09E-03 4.00E+00 0.34E-03 2.2 6.3 -100.55E-03 3.9 -70.53E-02 1.992 1.860 1.168 1.0 -100.00E+00 0.6 17.089 1.00E+00 NL RHA 1.6 22.178 1.3 29.62E-03 0.8 -35.0 -30.4 16.854 0.524 0.6 -25.214 1.809 0.8 -63.03E-03 5.623 1.0 -100.0 -71.530 1.75E-03 0.262 1.998 1.4 -12.913 0.672 1.14 m) from FEMA force distributions and MPA.5 61 .0 37.998 1.836 0.353 1.128 1.0 Table 5.318 2.1 10.515 FEMA Uniform ELF 27.45E-03 3.2 13. gravity loads included Displacement /Height (%) FEMA ELF SRSS MPA 0.195 0.0 -100.9 -20.65E-03 7.083 1.7 Table 5.490 0.839 0.1 -3.566 1.314 1.2 5.015 0.9 11.7 7.098 1.9 15.414 1.00E+00 0. gravity loads included Displacement /Height (%) FEMA ELF SRSS MPA 1.2 4.294 1.8 17.35E-03 8.399 1.1 -26.6 -41.2 9.3 -14.399 -27.708 0.51E-03 4.8 -100.6 4.372 1.367 1.7 -15.597 0.8 -100.8 2.953 0.17E-03 9.52E-04 1.221 1.686 0.7 21.199 27.19E-04 5.5 Story 1st 2nd 3rd 4th 5th 6th 7th 8th 9th Uniform 2.7 26.6 -17.306 1.04E-02 8.7 19.0 -63.487 0.2 16.94E-03 2.00E+00 0.8 -32.35E-10 3.061 1.724 0.78E-03 0.5 -3.6 -11.7 -60.8 7.16E-03 0.9 SRSS -22.011 1.109 1.6 -6.6 -73.673 Error (%) NL RHA 1.19E-03 1.78E-03 1.4 21.11E-03 9.9 28.93E-03 1.9 16.00E-10 Floor 1st 2nd 3rd 4th 5th 6th 7th 8th 9th Uniform 1.5 -33.00E+00 0.263 0.2 16.4 -50.5 -32.667 0.562 1.547 -27.00E+00 8.5 -27.1 163.2 -12.875 0.877 1.6 -4.2 -100.341 1.355 0.3 Peak values of hinge plastic rotations (radians) from FEMA force distributions and MPA Hinge Plastic Rotation (rad) MPA FEMA ELF SRSS 4.10E-02 7.4 -23.

5. Locations of plastic hinges determined from three force distributions in FEMA-273.(a) FEMA: Uniform • • • • • • • • •• •• •• •• •• •• (b) FEMA: ELF • •• •• •• •• •• •• • •• • •• •• •• •• •• • • • • • •• • • • • • • • •• •• •• •• • •• •• •• •• • •• •• •• •• • • •• •• •• • • • • • •• •• (c) FEMA: SRSS • • •• •• •• •• • • •• •• •• • • •• •• •• • • • • •• •• •• • • • • • • • • • (d) MPA with 3 "Modes" • • • • • • • •• •• •• •• •• •• •• •• •• •• •• •• •• •• •• •• •• •• •• •• •• • • • • •• •• •• (e) Nonlinear RHA • • • • • • • • •• •• • •• •• •• •• • • •• •• • •• •• •• •• • • •• •• • •• •• •• •• • • • • • • •• •• •• Fig. MPA including three “modes” and NL-RHA for 1.5.5 × El Centro ground motion. gravity loads included 62 .

5 2 GM Multiplier 2.5 1 1.(a) FEMA: Uniform 60 50 40 60 50 40 3 (b) FEMA: ELF Error Envelope Error for Floor No.5 3 0. 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 0.5 2 GM Multiplier 2. Errors in floor displacements from three force distributions in FEMA-273 and from MPA including three “modes”.5 2 GM Multiplier 2.5 1 1.5 1 1.5 3 Fig.6. gravity loads included 63 . 5.5 3 0 0 0.5 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 1 1.5 2 GM Multiplier 2.

5 2 GM Multiplier 2.(a) FEMA: Uniform 100 60 50 9 8 (b) FEMA: ELF Error Envelope Error for Story No. 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 3 0 0 0. 5.5 3 0 0 0.5 2 GM Multiplier 2.5 2 GM Multiplier 2.5 3 Fig.5 1 1.5 1 1.5 2 GM Multiplier 2.5 1 1. gravity loads included 64 .7.5 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. Error in story drifts from three force distributions in FEMA-273 and from MPA including three “modes”.

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

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

and identifying locations of most plastic hinges. MPA. joint rotations.25 to 3. Comparing the earthquake-induced demands for the selected 9-story building determined by pushover analysis using three force distributions in FEMA-273. the pushover analysis provides the peak value rno of any response quantity: floor displacements. However. and lead to unacceptably large errors in the hinge plastic rotations. it fails to compute with acceptable accuracy plastic rotations of the hinges. and hinge plastic rotations. 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. etc. all pushover analysis procedures do not seem to compute with acceptable accuracy local response quantities. demonstrates that the FEMA force distributions greatly underestimate the story drift demands.0. story drifts. story drifts. such as hinge plastic rotations. plastic hinge rotations.” 10. The MPA procedure is more accurate than all the FEMA force distributions in estimating floor displacements. 8. 7. Thus the MPA procedure is accurate enough for practical application in building evaluation and design.(iv) At the roof displacement determined from Dn . 9. Based on results presented for El Centro ground motion scaled by factors varying from 0. Pushover analyses seem to be inherently limited in computing accurately hinge plastic rotations. and nonlinear RHA. 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. 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. 67 . the errors in the MPA procedure are shown to be only weakly dependent on ground motion intensity.

While pushover estimates for floor displacements are even more accurate. Instead. such as life safety and collapse prevention. 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. should obviously be evaluated for a wide range of buildings and ground motion ensembles. with the excitation scaled to cover a wide range of ground motion intensities and building response. 68 . This new method for estimating seismic demands at low performance levels. they are not good indicators of damage.11. 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. 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.

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

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

. 71 .2.2.. 1977). B. Idealize the pushover curve as a bilinear curve (Fig.g. Apn . φn . The structure should be pushed just beyond the target (or expected) roof displacement in the selected mode. e. Define the force distribution s* from Eq. Apply force distribution of Step 2.20): s* = mφn n n 2.1. respectively. e. 1988). Define the anchor point. iterations may be necessary.Appendix A Uncoupled Modal Response History Analysis A. Let the roof displacement and base shear at the anchor point be urno and Vbno .urn ) pushover curve for the force distribution s* : n 2. using any numerical integration method. A. For the nth-“mode”. DRAIN2DX (Allahabadi and Powell. of the bilinear curve at the target roof displacement.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. ω n .g. for linear-elastic vibration of the building. trapezoidal rule. 1. Compute natural frequencies. (3. This step can be conveniently implemented in any commercially available software. 3. 3.1 incrementally and record the base shears and associated roof displacements. 2. Calculate the area under the actual pushover curve. Since the target roof displacement may not be known at the start of the procedure. 3.1.1) using the FEMA-273 procedure (Building Seismic Safety Council. and modes. develop the base-shear – roof-displacement (Vbn .

This value. Calculate the error = 100 × Abn − Apn ( ) ( ) ( ) ) Apn . kn .2.4. at base shear i equal to 0. other appropriate methods can be ( used.6 × Vbny ( ) i urn .9. kn = 0. Calculate Vbny = Vbny × Apn Abn . Let the point with base shear = Vbny and roof displacement = urny be denoted as A.6. i i i 3. determine the roof displacement.9. 3. 72 . iterations are necessary. and B with straight-line segments to obtain the idealized bilinear curve. i +1 i i 3.0.7. Replace i+1 with i and repeat Steps 3.1. Calculate initial slope of the idealized bilinear curve. If the error exceeds some pre-specified tolerance. Calculate area under the bilinear curve OAB.8. 3. Abn .6 × Vbny . A. i 3.3. Vbny .4 to 3. i i 3. Vbny .6 × Vbny .4.0. If desired. Calculate the slope.8. Calculate the yield displacement.2. by connecting a straight line between origin.4. i 3. urny = Vbny kn . corresponding to the estimated yield i i i base shear.i 3. 3. i i i αn = Vbno Vbny − 1 urno urny − 1 i 3.9. Draw the curve OAB by connecting the three points O. i 3.5. From the pushover data. Calculate the post-yielding strain-hardening ratio. This step gives the secant stiffness at a base shear equal to 60% of the yield base shear. urn. O. will be refined by an iterative procedure that seeks to equate areas under the actual and the idealized pushover curves. obtained by judgment. and a point on the actual pushover curve with base shear equal to i 0. Estimate the yield base shear.1.6 .6 .

4. Calculate histories of various responses using Eqs. An (t ) . and pseudo-acceleration history. 4. A. 5.10a and 4. A.1.3b) with unit-mass and force-deformation relation of Fig. Scale the vertical axis with * Mn to obtain * Fsno Ln = Vbno M n and * Fsny Ln = Vbny M n (Eqs. (3.3. Dn (t ) . Compute deformation history. Compute the Ln and Γ n from Eq. 8. 9. of the nth“mode” inelastic SDF system (Fig.0.16). * 4.4) and effective modal mass from M n = Ln Γn .1 Idealization of nth-“mode” pushover curve 73 . Develop the Fsn Ln − Dn relation (Fig.2. Combine the “modal” responses using Eqs. Vbn Idealization of Pushover Curve Vbno Vbny 0. 4.6 × Vbny A • Idealized αnkn B • 1 Actual k 1 • O n figA_1. of the combined responses obtained in Step 8. r o . (3. 6.6 ur n y ur n o ur n Fig. Calculate peak values. (3.10b and 4.2. 4.15) and (3.2). A. Repeat Steps 3 to 6 for as many modes as required for sufficient accuracy.eps ur n. In general first two or three modes will suffice. 4.11b).4. Scale the horizontal axis by Γnφrn to obtain Dno = urno Γ nφrn and Dny = urny Γ n φrn (Eqs. 7.12) and (3.11a).13).

B. are shown in Fig. for the first.4 cm (10 in.2 EXAMPLE The UMRHA procedure is implemented to calculate the response of the 9-story building described in Section 3. Area under the actual pushover curve. ur1o = 63.4.5 cm and Vb1o = 8729.2 Properties of the nth-“mode” inelastic SDF system A. respectively.urn ) pushover curve for the force distribution s* : n 2. Ap1 = 360777 kN-cm. is defined at the target roof displacement.1.6 kN.5.1. The pushover curves for the first three modes. A.5 cm (25 in.). 3. At this point. and 12.eps Dny = ur n y / Γnφr n Dno = ur n o / Γnφr n Dn Fig. Target displacements used to generate these pushover curves are 63. 74 .7 cm (5 in.2. The force distributions. Following is step-by-step implementation of the procedure described in Section A. 2. 3. 3.1.4.3. second. The anchor point. Idealized bilinear curves for each of the three modes are included in Fig. 1. 3.). 25.Fsn / Ln Fsn / Ln − Dn Relationship Vbno / M* n Vbny / Mn * 1 α ω2 n n ω2 1 n figA_2. The base-shear – roof-displacement (Vbn .). First three mode shapes and frequencies of the selected building were computed and are shown in Fig. A. to the north-south component of the El Centro (1940) ground motion scaled up by a factor of 1.2. generated using DRAIN-2DX. and third mode. s* . The following steps illustrate the procedure to develop the idealize curve for the first “mode”. n 2. computed for the first three modes are shown in Fig. 3. A.3.3.

Error = 100 × (365100 − 360777 ) 360777 = 1.4 kN.0.1. L1 = 2736789 kg.9. The point A on the i i bilinear curve is defined by ur1 y = 38.194.9. The next estimate of the yield shear is Vbny = 8006.3. iterations are necessary. Also included are the modal damping ratios and the periods calculated from Eq. 75 .8 22. i 3. i i i 3.3666 = 3740189 kg. The procedure converged after nineteen cycles to give ur1 y = 36.01%.4 × (360777 365100 ) = 7911.7.18 = 38.86 cm at 0.5 38. i i 3. Γ1 = 1.4 ) − 1 (63. 4.i 3. 3.9. The curve OAB obtained by connecting the three points O. i i 3.4 210.6 × Vb1 y ( ) i ur1.4.13). (4.8 kN.2. Vb1 y = 7615.6 kN.86 = 210.18 kN/cm.6 = 4803.6 8006. * 4.2. Therefore. The results for other modes are summarized in Table A.6 = 22.6.1. k1 = 0. A.4.09 cm.9 kN.1.23 cm.0.09 ) − 1 = 0.198%. 3.8.4 kN. ur1.5.135. k1 . The Fs1 L1 − D1 relation for the first “mode” is developed as follows.1. is calculated as follows. and B with straight-line segments gives the idealized bilinear curve. i 3. Area under the bilinear curve OAB. The post-yielding strain-hardening ratio. i +1 3. The first estimate of the yield base shear Vb1 y = 8006. The yield displacement.09 cm and Vb1 y = 8006. i i i 3. and α1 = 0. α1 = Vb1o Vb1 y − 1 ur1o ur1 y − 1 = ( ) ( ) (8729.2. The initial slope of the idealized bilinear curve.3666.6 × Vb1 y = 4803. and M1 = 2736789 × 1. This value exceeds the prespecified tolerance of 0. The results of the iterative procedure are summarized in Table A.4. ur1 y = Vb1 y k1 = 8006. 3. Ab1 = 365100 kN-cm. Determined from the pushover database.

6.4.4. Scaling the horizontal axis by Γ1φr1 gives D1o = 46. 4. The peak values are computed and are summarized in Tables 4.3.2. 5. The results were generated for first three “modes’ and are included in Fig.2.8. A. 4. 9.62 (cm/sec2). Histories of roof displacement and top story drifts for the first “mode” are computed and presented in Fig.46 cm and D1y = 26.40 (cm/sec2) and 76 . The peak values are also plotted in Fig. 4.7. 7. 4.1 and 4.51 cm. 4. * M1 gives Fs1o L1 = 233. Scaling the vertical axis by Fs1 y L1 = 203. 8. The combined modal responses are presented in Fig.7.7. 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.

eps 10 20 30 40 50 Roof Displacement (cm) (b) "Mode" 2 Pushover Curve 60 8000 u = 9.9 cm.2 cm.14 y by 6000 Base Shear (kN) 4000 2000 Actual Idealized figA_3c.eps 0 0 5 10 15 20 Roof Displacement (cm) (c) "Mode" 3 Pushover Curve 25 8000 u = 4. “Modal” pushover curves for the example building 77 . V = 5210 kN. α = 0. V y by = 4952 kN. A.(a) "Mode" 1 Pushover Curve 12000 10000 u = 36.3. V = 7616 kN. α = 0.6 cm.19 y by Base Shear (kN) 8000 6000 4000 2000 0 0 Actual Idealized figA_3a.eps 0 0 2 4 6 8 10 Roof Displacement (cm) 12 Fig. α = 0.13 6000 Base Shear (kN) 4000 2000 Actual Idealized figA_3b.

06 0 • 1.” and third “mode” inelastic SDF systems.(a) Deformation 80 n=1 3 (b) Pseuso−acceleration n=1 • 0.33 A1 (g) n=2 0 −80 80 −3 3 n=2 D2 (cm) 0 A1 (g) • 22. A.747 5 10 15 20 Time (sec) figa_4.” second “mode.223 −80 80 n=3 −3 3 n=3 D3 (cm) 0 A3 (g) • 10.5 × El Centro ground motion for the first “mode. Histories of deformation and pseudo-acceleration due to 1. 78 .2769 D1 (cm) 0 • 35.52 5 10 15 20 Time (sec) 25 30 0 • 1.4.m −80 0 −3 0 25 30 Fig.

198 0.191 0.9 7615.38 22.7 7639.103 “Mode” 3 696400 0.8 4747.59 36.18 (cm) 38.5 3109.162 0.037 0.2671 1.8 7622.85 36.18 210.1 4569.194 0.24 36.176 0.18 210.18 210.1.3 7786.50 36.0 4588.529 0.56 19.5 7633.18 210.75 21.190 0.3 4603.18 210.2.4 7714.9 4573.188 0.404 0.4 4595.9 4570.95 21.082 0.23 0.3 7745.3666 3740189 203.8525 1.107 0.8 4647.193 0.0.193 0.3 7628.18 210.4 7647.6 × Vb1 y i ur1.74 21.5 7624.51 2 “Mode” 2 -920860 -0.6 i k1 i ur1 y i α1 i Ab1 Error (%) 1.1 1013.193 0.28 36.0 4704.79 0.136 Fsny Ln (cm/sec2) Dny (cm) Fsno Ln (cm/sec ) Dno (cm) Tn (sec) 233.02 21.063 0.4 7672.2 4571.44 36.29 36.18 210.75 21.022 0.18 210.9 4570.64 37.048 0.85 0.017 0.25 36.12 3876.194 0.6 4583.79 21.59 22.32 36.05 52.184 0.5309 488839.013 0.5 (cm) 22.26 36.139 0.4927 1. i 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 i Vb1 y i 0.3 7658.0 4577.Table A.2406 167531.56 47.7 4580.18 210.170 0.135 0.151 0.76 21.18 210.4 7911.05 36.35 36.18 210.693 0.83 21. No.18 210.5 4614.192 0.180 0. Results of iterative procedure to develop the idealized bilinear curve for the first “mode” inelastic SDF system Itr.6 7840.18 210.09 37.18 210.18 210.182 0.1 4574.237 0.8 7618.029 0.78 21. Properties of “modal” inelastic SDF systems Properties Ln (kg) Γn * M n (kg) “Mode” 1 2736789 1.2 7690.2 4671.23 22.1 7616.18 210.18 210.77 21.70 36.25 36.9 (kN) 4803.194 0.309 0.86 22.910 0.11 22.18 210.2 4628.62 26.010 (kN) 8006.90 21.65 1226.186 0.948 ζ n (%) 79 .86 21.09 18.74 (kN/cm) 210.30 37.40 46.194 (kN-cm) 365100 364060 363276 362684 362235 361892 361631 361432 361279 361162 361073 361004 360951 360911 360880 360856 360838 360824 360813 Table A.81 21.39 36.46 2.0 7619.

80 .

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. 10. or from the inelastic response (or design) spectrum. 4. Repeat Steps 3 to 7 for as many “modes” as required for sufficient accuracy. 6.4. Following is step-by-step implementation of the procedure described in Section B. the first two or three “modes” will suffice. 13. rno . 10.5. 12. At urno . Solving Eq. 4. Dn . extract from the pushover database values of other desired responses.18). This is the same example as solved in Appendix A.3b) with unit mass and force-deformation relation of Fig. B.8). From the total hinge rotation. (3.8) 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. subtract the yield hinge rotation to determine the plastic hinge rotation. 11. 81 . Compute the peak deformation. Typically. Steps 1 to 4 of the MPA are the same as those for UMRHA presented in Appendix A. Determine the total response by combining the peak “modal” responses using the SRSS combination rule of Eq.1 to the north-south component of the El Centro (1940) ground motion scaled up by a factor of 1.6b by solving Eq.21). (4.1. (4. Calculate the peak roof displacement urno associated with the nth-“mode” inelastic SDF system from Eq.Appendix B Modal Pushover Analysis B. (3.33 cm.2 EXAMPLE The MPA procedure is implemented to calculate the response of the 9-story building described in Section 3. of the nth-“mode” inelastic SDF system (Fig.

28 cm.366 × 1 × 35. Steps 3 to 7 are repeated for first three “modes.3 for the floor displacements normalized by the building height and in Table 4.3 and 4.4. Peak roof displacement ur1o = Γ1 × φr1 × D1 = 1.5 are the plastic hinge rotations computed by subtracting the yield hinge rotation from the total hinge rotations. (3. At ur1o = 48.11. 82 .” and the results are included in Tables 4. 8. The values are summarized in Table 4.28 cm.1.33 = 48.4 for the story drifts normalized by the story height. values of floor displacements and story drifts are extracted. 9.4.3 and 4. 12. where results for other ground motion intensities are also included.18) are also included in Tables 4. The total response computed by combining the peak “modal” responses using the SRSS combination rule of Eq. Results of Steps 5 and 6 are summarized in Table B. Also included in Table 4.

504 18.35 0.367 1.535 14.79 0.268 0.06 1.52 0.37 1.748 63.13 2.660 14.312 1.901 8.690 10.577 16.766 7.07 “Mode” 2 4.18 27.332 13.551 2.379 21.154 78.03 26. Calculation of roof displacements urno from peak deformation of inelastic SDF systems Ground Motion Multiplier 0.25 20.38 22.05 0.59 0.856 31.450 4.185 11.Table B.267 5.5 µ = Dn Dny urno = Γnφrn Dn (cm) Dn (cm) 0.85 µ = Dn Dny urno = Γnφrn Dn (cm) Dn (cm) 1.755 0.75 µ = Dn Dny urno = Γnφrn Dn (cm) Dn (cm) 0.50 35.36 1.117 5.27 0.1.126 13.436 7.913 22.33 1.8451 5.200 0.007 36.71 1.755 27.969 0.139 µ = Dn Dny urno = Γnφrn Dn (cm) Dn (cm) 0.467 14.4222 3.229 8.252 9.332 48.0 µ = Dn Dny urno = Γnφrn Dn (cm) 83 .25 Quantity Dn (cm) “Mode” 1 6.513 0.0 µ = Dn Dny urno = Γnφrn Dn (cm) Dn (cm) 3.73 24.225 2.184 0.023 0.28 46.5 µ = Dn Dny urno = Γnφrn Dn (cm) Dn (cm) 2.457 12.0 µ = Dn Dny urno = Γnφrn Dn (cm) Dn (cm) 1.37 57.676 6.395 0.03 0.70 0.735 3.691 0.275 1.52 “Mode” 3 1.35 1.678 0.82 1.

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

5 -350.6 1233.367 87 .7 694.0 1476.7 1622.8 374.047 0.042 0.1 -525.2 95.3 -6.0 136.7 -46.2 Lateral Force fj (kN) (9) 203.070 0.7 234.1 Story Shears V j2 (kN) (6) 1114.3.0 980.6 286.2 148.4 1759.4 400.6 -732.8 1381.177 0.9 366.090 0.9 -166.3 240.4 250.4 1842.5 -320.7 374.0 -5.6 319.7 1578.9 1446. the lateral forces are normalized by the base shear to obtain column 10 in Table C.1 -967.6 V j3 (kN) (7) 478.3.7 Vj (kN) (8) 2268.1 1857.5 -973.9 880.3 222.6 -352.2 285.7 101.2 105.7 525.1 87.6 -359.7 (10) 0.8 430.6 366.7 95.8 -326.1 832.0 1231.2 200.7 355.5 215.2 277.1 -438.1 -646.065 0.2 97.5 159.For convenience. 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) 1917.045 0.9 446.9 -153.3 -646.9 832.0 381.7 2065. Table C.9 1683.098 0.2 fj = ∑ fi i fj (2) (3) (4) 59.0 176.0 354.5 320.