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 .

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

iv .

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

vi .

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

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

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

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

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

and the ductility factor is u m= m uy (2. 2001. (2. and R y (Chopra . the deformation u t depends on three systems parameters: w n . This becomes evident if Eq. sign u = . z . deformation is denoted by um .fS m fs c ut u fy 1 αk k 1 uy ug -fy k k um u 1 1 (a) (b) Fig. 2. or absolute (without regard to algebraic sign) maximum. and u y . and (b) bilinear hysteretic forcedeformation relation For a given excitation u g t .1. bg 4 . m depends on three system parameters: w n . u £ u y ). in addition to the form of the force-deformation relation.5) For a given u g t .. (a) Idealized one-story structure.e.1) is divided by m to obtain ~ u + 2zw n u + w 2 u y f s u. and z is the damping ratio of the system vibrating within its linear elastic range (i.4) and w n is the natural vibration frequency. The peak. Tn = 2p w n is the natural vibration period.ug t n where bg bg b g bg (2.3) ωn = k m ζ = c 2mω n f fs = s fy (2.3). Section 7. z y .

Pushover curve for structure shown in Fig. For .2.66 m • EIb EIc EIc • • L = 7. one-bay frame 50 uy = 1.2 SYSTEM AND EXCITATION CONSIDERED Consider the one-story system in Fig.2: the dimensions and flexural rigidity of the structural elements are noted. Vby = 39.376 cm. 2.2 5 . is f y / w = ( f o / w ) ÷ 8 = 0. Vb (kN) 30 20 10 0 0 2 4 6 Roof Displacement.2 kips) . this system and excitation.26 kN.1.9 kN (38.26 kN (8. The yield strength of the inelastic system. m h = 3. and f y = 39. based on Ry = 8 . f o w = 184 . 2001) and scaled up by a factor of 2.04 40 Base Shear. 2.4 in Chopra. 6. α = 0. u (cm) 8 Fig. with Tn = 0. 2. 2.2311 .5 sec and z = 5% subjected to the north-south component of the El Centro (1940) ground motion (Fig. One-story.32 m • Fig.2.826 kips) for w = 169.3.

wherein the hysteretic force-deformation history of Fig.3 RESPONSE HISTORY ANALYSIS Figure 2. 2. and q pm = 0.65 kN-m (191. Implementing this analysis with I c = 6.134 × 107 mm 4 (75. as apparent in Fig.35. 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 . I b = 3.017 rad .36 cm .36 cm. respectively. 2.4f.4 ). 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 . (d) shows the rotation q p bt g of the plastic hinges at the bg bg beam ends.4edetermined from RHAis superimposed.4e.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 = 7.6 kip-in.3 in. pushover analysis cannot provide any cumulative measure of response.0217 rad .4 ).3. and (e) shows the force-deformation relation. the joint rotation and beam hinge rotation are identical to values q m and q pm determined from RHA.4 shows the earthquake response of the system described in the preceding section determined by response history analysis (RHA). 2. the ductility factor m = 5. e.. (c) shows the joint rotation q bt g . (b) shows the lateral force f s t or base shear Vb t normalized relative to the bg weight. 2. The system is excited well beyond the yield deformation. The resulting pushover curve is shown in Fig.The yield moments in the beam and columns are defined as the bending moments due to the lateral force f y .1 kip-in. Determined from pushover analysis at the exact peak deformation.) for the beam and columns. Observe that the pushover curve matches the initial loading path of the hysteretic system. the energy dissipated in 6 .g. The peak values of the various response quantities are as follows: um = 7. and E = 2 × 108 kPa (29×103 ksi) gives yield moments of 21.077 × 107 mm 4 (146 in. However.) and 50. q m = 0. 2. 2. The yielding stiffness of each structural element is defined as 3% of its initial stiffness.18 kN-m (444. the peak value determined from RHA.

04 0.2 −0. 15 u (cm) 0 • 7.017 • 0 (d) θp (rad) −0. (d) plastic hinge rotation.4 0. (b) base shear. Response of one-story system to El Centro ground motion: (a) deformation.2311 y (b) −0.4 0. (c) joint rotation.2 −0. This represents an inherent limitation of pushover analyses.04 0.5 fy / w = 0.36 −15 0.4.2311 (a) Vb / w 0 f / w = −0.0217 • θ (rad) 0 (c) −0.04 0.4 −10 −5 0 u (cm) 5 10 −0. 2. and (f) pushover curve 7 . (e) force-deformation relation. or the cumulative rotation at a plastic hinge.2 0 (e) 0 (f) −0.4 −10 −5 0 u (cm) 5 10 Fig.2 Vb / w 0.04 0 5 10 15 Time (sec) 20 25 30 0.5 0.yielding during the ground motion.

8 .

c.1) can be interpreted as effective earthquake forces: p eff t = . and k are the mass.mι ug t bg bg bg (3. m.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 .3) where φ n is the nth natural vibration mode of the structure. Elastic Multistory Buildings 3.1 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 = .4) 9 . The right side of Eq. This force distribution can be expanded as a summation of modal inertia force distributions s n (Chopra. each element of the influence vector ι is equal to unity. (3. and lateral stiffness matrices of the system.3. and p eff t = Â peff.n t = Â . classical damping.snug t n= n =1 The effective earthquake forces can then be expressed as af N af N af (3.1) where u is the vector of N lateral floor displacements relative to the ground.m ι u g t bg bg (3. 2001: Section 13.12): mι = n =1 Â sn = N n =1 Â Γ nmφ n N (3.

sn ug t bg (3. (3. Then the floor displacements are u n t = φ n qn t bg bg bg bg (3. an SDF system with vibration propertiesnatural frequency w n and damping ratio z n of the nth-mode of the MDF system.10) bg bg bg 10 (3. (3. with no contribution from other modes.9) to the equation of motion for the nth-mode elastic SDF system.Gn ug t (3.u g t n Comparing Eqs. The equations governing the response of the system are mu + cu + ku = .n (t ) = N n =1 ∑ −sn u g (t ) N (3.11) and substituting in Eq.12) . we will outline that the response of the MDF system to p eff .7) By utilizing the orthogonality property of modes.peff (t ) = n =1 ∑ peff . p eff .8) gives the floor displacements u n t = Gnφ n Dn t bg (3.8) where the modal coordinate qn t is governed by 2 qn + 2z nw n qn + w n qn = .9) and (3.10) gives qn t = Gn Dn t bg bg bg (3.n t = . it can be demonstrated that none of the modes other than the nth mode contribute to the response. subjected to u g t : Dn + 2z nw n Dn + w 2 Dn = .5) The contribution of the nth mode to s and to p eff t are: sn = Gn mφ n respectively.6) Next. The solution qn t can readily be obtained by comparing Eq.n t is entirely in the nth-mode.sn ug t bg bg bg bg (3.9) in which w n is the natural vibration frequency and z n is the damping ratio for the nth mode. (3.

15) rt = bg n =1 Â rn t = bg n =1 Â rnst An bt g N (3. 2001.9) is the standard modal equation governing qn t .16) reflect combining the response contributions of all modes. (3. etc. (3. In contrast to the classical 11 bg . internal element forces. the static value of r due to external forces sn .15) and (3. 3.n t .1.Forces sn An(t ) st rn ωn.1.can be expressed by st rn t = rn An t bg bg bg bg bg (3.14) is the pseudo-acceleration response of the nth-mode SDF system (Chopra. and Eqs. 3. Section 13. 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. (3.1). ζn ¨ g(t ) u (a) Static Analysis of Structure (b) Dynamic Analysis of SDF System Fig. Conceptual explanation of modal response history analysis of elastic MDF systems Any response quantity r t story drifts.16) This is the classical modal RHA procedure wherein Eq. Equations (3.13) define the contribution of the nth-mode to the response. and An t = w 2 Dn t n bg (3. However. st The two analyses that lead to rn and An t are shown schematically in Fig.13) represent the response of the MDF system to p eff. Eqs.12) and (3.12) and (3.13) st where rn denotes the modal static response. these standard equations have been derived in an unconventional way. Therefore.

9) . 2001.2 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. which is valid for structures with well-separated natural frequencies such as multistory buildings with symmetric plan. 2001. and Tn = 2p w n is the natural vibration period of the nth-mode of the MDF system. (3.(3.4 and 13.17) (Chopra.3). 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 3.8.derivation found in textbooks (e. Section 13. This concept provides a rational basis for the modal pushover analysis procedure developed later.16). The peak modal responses are combined according to the Square-Root-of-Sum-ofSquares (SRSS) or the Complete Quadratic Combination (CQC) rules.. we note that static analysis of the structure subjected to lateral forces f no = Γ n mφ n An (3.3 MODAL PUSHOVER ANALYSIS (3. Chopra. 3. The SRSS rule. Alternatively. the peak nth-mode response as in Eq.18) To develop a pushover analysis procedure consistent with RSA. (3. z n of the pseudo-acceleration response (or design) spectrum for the nth-mode SDF system. In such a response spectrum analysis (RSA).g. this response value can be obtained by static analysis of the structure subjected to lateral forces distributed over the building height according to 12 .1). Sections 12.1. 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 (3.19) will provide the same value of rno . we used the modal expansion of the spatial distribution of the effective earthquake forces.17) where An is the ordinate A Tn .

urno .12) is urno = Gnf rn Dn (3.1 System and Excitation Considered The 9-story structure. 3. Consulting Structural Engineers.4 COMPARATIVE EVALUATION OF ANALYSIS PROCEDURES 3. the peak value of the roof displacement due to the nth-mode.2)..2. rno . in both directions. (3. which from Eq. can be combined according to Eq. shown in Fig. region. with the ninth level being the roof. The bays are 9. Typical floor-to-floor heights (for analysis purposes measured from center-of-beam to center1 2 Brandow & Johnston Associates. each determined by one pushover analysis. The columns are 345 MPa (50 ksi) steel wideflange sections.s* = mφ n n (3. Third St. Los Angeles. This modal pushover analysis (MPA) for linearly elastic systems is equivalent to the well-known RSA procedure (Section 3.73 m (150 ft) in plan. 46th Street. (3. A benchmark structure for the SAC project.15 m (30 ft) on center. 3. The levels of the 9-story building are numbered with respect to the ground level (see Fig. SAC Steel Project Technical Office. this structure meets seismic code and represents typical medium-rise buildings designed for the Los Angeles. 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. The building has a basement level. and California Universities for Research in Earthquake Engineering (CUREE).19 m (122 ft) in elevation. Obviously Dn and An are available from the response (or design) spectrum. CA 90017. The interior bays of the structure contain simple framing with composite floors. 1301 S.73 m (150 ft) by 45.18) to obtain an estimate of the peak value ro of the total response.20) and the structure is pushed to the roof displacement. 3. 13 . Richmond. CA 94804-4698. was designed by Brandow & Johnston Associates1 for the SAC2 Phase II Steel Project. The peak modal responses. California. the Applied Technology Council (ATC).2). this building is 45. and 37.4.21) 2 where Dn = An ω n . 1660 W. with five bays each in the north-south (N-S) and east-west (E-W) directions. Although not actually constructed. SAC is a joint venture of three non-profit organizations: The Structural Engineers Association of California (SEAOC). denoted B1.

3. third. The seismic mass of the above ground levels of the entire structure is 9.65 m (12 ft) and for the first floor is 5. Each frame resists one half of the seismic mass associated with the entire structure. for the first three modes are shown in Fig. such as those described in Gupta and Krawinkler (1999) can be used. ceiling/flooring. The 9-story N-S MRF is depicted in Fig.. 3. s* (Eq. including the steel framing.20). roofing and a penthouse located on the roof. are located on the first. and 0.27. Concrete foundation walls and surrounding soil are assumed to restrain the structure at the ground level from horizontal displacement. for the first level is 1.7 kips-sec2/ft). The column lines employ two-tier construction. for the second through eighth levels is 9. Column splices. and shear distortion of panel zones are neglected but large deformation (P∆) effects are included. The first three vibration modes and periods of the building for linearly elastic vibration are shown in Fig. To ensure that this structure remains elastic. The floor-to-floor height of the basement level is 3.00×106 kg (616 kips.65×105 kg (66. respectively. The floor system is composed of 248 MPa (36 ksi) steel wide-flange beams in acting composite action with the floor slab. 14 . These force n distributions will be used in the pushover analysis to be presented later. This model is based on centerline dimensions of the bare frame in which beams and columns extend from centerline to centerline. fifth.0 kips-sec2/ft). dimension.01×106 kg (69.89×105 kg (67. The strength. if desired more complex models. mechanical/electrical.07×106 kg (73.49 sec.83 m (6 ft) above the center-line of the beam to column joint. and for the ninth level is 1. 3. The building is modeled in DRAIN-2DX (Allahabadi and Powell. partitions. The simple model adopted here is sufficient for the objectives of this study.3. monolithic column pieces are connected every two levels beginning with the first level. 3. The seismic mass of the structure is due to various components of the structure.96 m (13 ft). 0. The force distributions.0 kips-sec2/ft).of-beam) are 3. we select a weak ground motion: the northsouth component of the El Centro (1940) ground motion scaled down by a factor of 4. i. The column bases are modeled as pinned and secured to the ground (at the B-1 level). which are seismic (tension) splices to carry bending and uplift forces.e.sec2/ft).49 m (18 ft).2.4. The seismic mass of the ground level is 9. floor slabs. 1988) using the M1 model developed by Krawinkler and Gupta (1998). the vibration periods are 2.85.2 kips-sec2/ft). and seventh levels at 1.

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

49 sec 3 T = 0.5 Mode Shape Component 1 1.27 sec Ground −1. and 3. Each figure is organized in four parts: (a) shows the roof displacement urn t .487 −1. 3.37 2.04 1.0272 −2.67 −1.85 sec 2 Floor 5th 4th 3rd 2nd 1st T1 = 2. respectively.33 2.44 1.7.51 0.5 −1 −0. and 3 n 3.1 −2. (c) shows the joint rotation q n t of an external joint at 16 bg bg bg . (b) shows the base shear Vbn t normalized relative to the weight W of the building.4.38 0.5 Fig.13)]. n = 1.5.1 3. 3. 3.05 s * 1 s * 2 s * 3 Fig. First three natural-vibration periods and modes of the 9story building 3.05 2.13 −1.39 3.796 0. is shown in Figs.05 1.12 0.2 Response History Analysis The structural response due to individual vibration modes.12) and (3.5 0 0. 3.03 −1.75 1. (3. and 3 .94 2.8 −2. n = 1.3. determined by RHA [Eqs.31 −0. Force distributions s* = mφn .61 2.93 −1. 2.6.728 2. 2.9th 8th 7th 6th T = 0.4.72 −2.

the results for the roof displacement and top-story drift are shown in Fig. The errors due to truncating the contributions of vibration modes beyond the third mode are negligible. two. The linear relationship between the base shear and roof displacement for each mode implies that the structure did not yield. 3. and ur 3o = 0. are presented in Fig.2. in particular. It is apparent that the first mode alone is inadequate.1. but three modes—perhaps even two modes—are sufficient. especially in estimating the story drifts. and (d) shows the Vbn .8. where the second and third modal responses are a larger percentage of the top story drift compared to roof displacement. This is illustrated in Fig. 3.urn relation. or all modes.12 cm. The peak values of displacements of all floors. three. and the percentage errors due to truncation of higher modal contributions. the peak roof displacement due to each of three modes is ur1o = 9. and even larger in joint rotations. and 3.8. consistent with the increasing significance of the higher mode response among these three sets of response quantities. the exact response considering all modes. The peak values of floor displacements and story drifts determined by RHA. which are listed in Tables 3. and rotations of external joints with moment connections are presented in Tables 3. The same method was used to determine the peak values of many response quantities.3. drifts in all stories. and 3. As expected.the roof level.422 cm. 3. 17 . respectively. For a fixed number of modes included. ur 2o = 2. and three vibration modes. errors are smallest in floor displacements. including one. 3.9. respectively. 3.3. The peak values of the various response quantities are noted in these figures.2. Combining the modal response histories for all modes gives the total response [Eqs. (3. two.15) and (3.16)]. with the first three modes provide essentially the exact response. Also included are the combined response due to one. errors generally decrease as response contributions of more modes are included.1. larger in story drifts.23 cm.

91E-04 1.009 -0.237 0.325 0.4 -22.4 -10.197 0.264 Error (%) 1 Mode 2 Modes 3 Modes -23.76E-03 1.99E-03 2.09E-03 1.026 0.2 Peak values of story drift ratios (as % of story height) from RHA for 0.8 -10.060 -0.259 0.042 0.124 0.089 0.3 -0.266 0.26E-04 -5.4 0.022 0.38E-03 2.080 0.370 0.6 9.227 0.8 -56.245 0.125 0.65E-03 RHA (all modes) 2.121 0.09E-03 Error (%) 1 Mode -23.09E-03 2.13E-03 2.237 0.263 0.3 Peak values of joint rotations (radians) from RHA for 0.226 0.5 -1.4 -3.032 -0.4 -53.2 -20.33E-03 2. 18 .03E-03 -6.202 0.1 Table 3.3 -3.035 0.0 7.023 0.72E-03 3.181 0.012 0.7 4.63E-03 2.22E-03 2.235 0.265 RHA (all modes) 0.3 -33.311 0.8 1.7 3.9 -22.94E-03 2.3 -8.88E-03 2.42E-04 1.261 -0.202 0.307 0.88E-03 2.002 -0.1 -2.5 18.295 -0.253 0.060 0.259 0.130 0.56E-03 2.85E-03 3.7 2.1 -2.378 0.229 0.25 × El Centro ground motion Floor 1st 2nd 3rd 4th 5th 6th 7th 8th 9th 3 Modal Response Mode 1 2.6 0.227 0.300 0.097 0.31E-03 Joint Rotation (rad) Combined (RHA) 1 Mode 2.9 -23.266 0.042 0.4 -0.097 0.00E-03 1.9 -15.5 -2.03E-03 3 Modes 2.179 0.11E-04 -5.1 Peak values of floor displacements (as % of building height = 37.235 0.3 Table 3.159 0.Table 3.258 0.226 0.011 0.6 -1.310 0.1 3.177 0.173 0.73E-03 3.90E-03 3.156 0.01E-04 -2.333 0.0 3 Modes -5.245 0.47E-03 1.177 0.63E-03 2.058 -0.156 0.08E-03 2.6 11.133 -0.74E-03 1.6 1.229 0.282 0.006 0.50E-03 4.1 -19.010 -0.8 -15.071 0.024 -0.9 9.1 Building height is measured from the ground floor to the 9th floor.202 0.6 4.45E-03 3.350 0.6 0.203 0.038 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 0.28E-04 1.9 3.054 0.11E-03 1.15E-03 4.0 -2.265 0.01E-04 3.2 1.364 0.99E-03 Mode 3 3.00E-03 1.253 0.7 7.0 -46.060 -0.125 0.466 0.231 -0.001 -0.015 -0.157 0.9 8.8 -5.069 0.03E-03 1.406 0.74E-03 1.192 0.124 0.13E-04 9.225 0.14E-03 2.045 0.31E-03 Mode 2 -1.413 RHA (all modes) 0.336 0.062 0.89E-03 1.6 -1.44E-03 1.055 0.89E-03 1.0 -10.090 0.2 -2.117 0.008 -0.9 1.401 Error (%) 1 Mode 2 Modes 3 Modes -23.321 0.14 m3) from RHA for 0.00E-03 2.475 0.183 0.1 -0.1 -14.66E-05 -3.29E-03 2.4 -7.74E-04 6.7 2 Modes -3.062 -0.8 1.0 -2.311 0.4 -6.0 -0.1 -2.1 4.03E-03 1.2 0.069 0.3 19.453 0.003 0.3 -0.407 0.76E-03 1.4 -41.78E-04 -3.44E-03 3.303 0.399 0.7 -50.4 -1.2 -4.173 0.9 -16.400 0.282 0.64E-03 3.011 0.5 0.088 -0.275 0.1 -11.2 -57.152 0.4 -1.8 -1.65E-03 2.6 -0.9 2.2 -1.043 0.9 -24.260 0.09E-03 1.25 × 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 0.205 0.38E-04 2 Modes 2.50E-03 2.74E-04 9.012 -0.266 0.24E-03 2.199 0.7 -19.152 0.097 0.2 9.260 0.

5 9.1 /W 0.5 0 u (cm) r1 7.00174 • 0 (c) r1 −5 0 5 10 15 Time (sec) 20 25 30 0.0435 Vb1 / W 0.5 0 u (cm) r1 7.0435 • 0 (b) V b1 −0. and (e) pushover curve.25 × El Centro ground motion 19 . Response due to first mode: (a) roof displacement.0435 9. Excitation is 0.1 −15 −7.5.1 −15 −7.12 0 (d) 0 (e) 15 Fig.5 15 −0. 3. (c) joint rotation. (b) base shear.15 u (cm) 9. (d) force-deformation history.1 5 x 10 −3 θ (rad) 0.1 0.12 • 0 (a) r1 −15 0.12 −0.1 0.

5 0 u (cm) r2 2. Response due to second mode: (a) roof displacement.23 −5 (a) 0.1 Vb2 / W 0. Excitation is 0.05 /W 0.00244 • 0 (c) r2 −5 0 5 10 15 Time (sec) 20 25 30 0. (c) joint rotation.1 −5 −2. (d) force-deformation history.1 0.5 0 u (cm) r2 2.6.5 5 −0. 3. (b) base shear.0252 (d) 0 0 (e) −0.05 5 x 10 −3 b2 θ (rad) 0.25 × El Centro ground motion 20 . and (e) pushover curve.0252 • 0 (b) V −0.0252 −2.23 −2.5 u (cm) r2 0 • 2.23 0.5 5 Fig.1 −5 −2.

(d) force-deformation history.02 2 θ (rad) 0.25 × El Centro ground motion 21 .5 1 −0.422 −1 (a) 0.5 1 Fig. 3.5 0 u (cm) r3 0.7.422 (e) −0.02 −0.02 −1 −0.0108 −0.02 0. Excitation is 0.02 −1 −0. and (e) pushover curve.0108 −0.000913 • 0 (c) r3 −2 0 5 10 15 Time (sec) 20 25 30 0.422 /W V b3 0 (d) 0 −0.5 0 u (cm) r2 0.0108 x 10 −3 (b) −0.1 u (cm) r3 0 • 0. (c) joint rotation.02 /W V b3 0 • 0. Response due to third mode: (a) roof displacement. (b) base shear.

18).5.5.4.6. (3. (3. 3. i. 3. 3. the errors in the MPA results are generally larger than in RHA (Fig.8) leads to the pushover curve shown in Fig.7e and to the floor displacements. respectively. (3.6d and 3.2. 3. respectively. For a fixed number of modes included considered.3). respectively.3. The floor displacements. implying that the modal truncation errors are small if at 22 .422 cm . (3.4. the floor displacements are proportional to the mode shape φ 1 because the structure remains elastic.2. 3.. As for the first mode. confirming that MPA gives the exact values of the individual modal responses.20) with n = 1 (Fig.10 and Tables 3. leads to the pushover curves shown in Figs.3).1. 3.3). or three vibration modes. Implementing pushover analysis for the second and third modes.3 Modal Pushover Analysis Implementing MPA for the fundamental vibration mode. using the force distribution of Eq. and 3. 3. except for the algebraic sign associated with Γ1 and minor round-off errors.4. pushing the structure. story drifts. considering one. two.5.6 present estimates of the combined response according to Eq. which would provide the Dn value to be substituted in Eq. these pushover curves are consistent with the Vb .12). and it is apparent in the example considered that three modes provide most of the response (Fig.5e. pushing the structure using the force distribution of Eq. These values of the response quantities are identical to the peak response values determined from RHA (Tables 3. 3. and external joint rotations in Tables 3. although both analyses led to identical peak values of the individual modal responses. This pushover curve is consistent with the relationship between the base shear and roof displacement determined by RHA (Fig. Figure 3. Observe that the target roof displacement in each pushover analysis is identical to its exact value determined by RHA.4) to roof displacement ur1o = 9. and 3. and external joint rotations computed by pushover analysis are presented in Tables 3. and ur 3o = 0. i.5d). and 3.1.6. and 3. (3. and the computed response values are identical to the peak response values determined from RHA (Tables 3.9). In practical application. story drifts. 3. 3.21).20) with n = 2 and 3 up to roof displacements ur 2 o = 2. this value would be determined directly from the response (or design) spectrum.e.12 cm. and 3. 3. and the errors in these estimates relative to the exact response from RHA considering all modes.23 cm .7d). the value determined by RHA (Fig.e. As suggested by Eq.6e and 3. 3. 3. In RHA the errors arise only from truncating the responses due to higher modes.1 through 3..9 and Tables 3.ur relations determined by RHA (Figs.4.

685 • −3 3 Mode 2 ∆ (cm) u (cm) • 2.5 Ground 0 0. Response histories of roof displacement and top-story drift from RHA for 0. Additional errors are introduced in pushover analysis due to the approximation inherent in modal combination rules. (a) Roof Displacement 15 (b) Top Story Drift 3 Mode 1 ∆ (cm) u (cm) 9. Heightwise variation of floor displacements and story drifts from RHA for 0.48 −3 3 1.9.eps Ground 0 0.59 • Total (All Modes) 0 −15 0 ∆ (cm) u (cm) 0 −3 0 fig3_8.least three modes are included.6 Fig.2 0.25 × El Centro ground motion 23 .83 • r3 0 • 0.5 0.3 0.23 r2 −15 15 0 • 0.03 −3 3 Mode 3 ∆ (cm) u (cm) r3 −15 15 9.12 • r1 −15 15 0 r1 0 0 0.1 0. sec 25 30 r 5 10 15 20 Time. 3. 3. sec 25 30 Fig.422 r2 0 • 1.eps fig3_9b.4 Displacement/Height (%) 0.1 0.2 0.4 Story Drift Ratio (%) 0.8.3 0.25 × El Centro ground motion: first three modal responses and total (all modes) response (a) Floor Displacements 9th 8th 7th 6th 9th 8th 7th 6th (b) Story Drift Ratios Floor 4th 3rd 2nd 1st Floor 5th RHA All Modes 1 Mode 2 Modes 3 Modes 5th 4th 3rd 2nd 1st RHA All Modes 1 Mode 2 Modes 3 Modes fig3_9a.eps r 5 10 15 20 Time.

276 0.125 0.5 -16.0 -18.156 0.069 0.230 0.267 0.173 0.8 -23.2 -20.125 0.09E-03 1.7 -19.203 0.012 0.15E-03 1.14 m) from MPA for 0.177 0.078 0.9 -24.055 0.3 -12.245 0.8 -56.8 -15.63E-03 -2.3 -2.22E-04 2.3 -9.03E-03 -1.152 0.062 0.9 -18.266 0.179 0.097 0.6 Peak values of joint rotation (radians) from MPA for 0.99E-03 Mode 3 -3.5 Peak values of story drift ratios (as % of story height) from MPA for 0.0 -10.009 0.33E-04 5.259 0.72E-03 3.3 -19.9 -13.40E-04 5.038 -0.3 -33.133 -0.1 -11.259 -0.Table 3.1 -0.00E-03 1.9 -15.25 × El Centro ground motion Floor 1st 2nd 3rd 4th 5th 6th 7th 8th 9th Modal Response Mode 1 -2.300 0.9 -8.002 0.401 Error (%) 1 Mode 2 Modes 3 Modes -23.63E-03 2.181 0.015 -0.270 0.024 0.088 -0.282 0.89E-03 2.4 -14.96E-03 2.6 -19.321 0.9 -14.6 -15.09E-04 -3.42E-04 -1.008 -0.4 -19.7 24 .260 -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 -0.1 -18.00E-03 2.00E-03 2.3 1.09E-03 Mode 2 1.106 0.2 Table 3.079 0.89E-03 -1.00E-03 2.023 -0.2 1.03E-03 3.44E-03 -1.0 -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 0.336 RHA (all modes) 0.177 0.270 0.9 -15.260 0.272 0.133 0.4 -5.011 -0.03E-03 6.336 0.4 -9.106 0.53E-04 -9.4 -2.94E-03 2.173 0.90E-04 -9.032 0.4 -22.385 0.03E-03 1.036 -0.285 0.267 0.8 -14.3 -14.73E-03 3.6 -16.001 0.74E-03 1.231 -0.235 -0.253 0.374 0.3 -13.4 1.045 -0.022 0.4 -4.7 2.310 0.89E-03 1.89E-03 2.117 0.058 -0.203 0.152 0.1 -19.060 -0.466 0.12E-03 1.322 0.05E-03 3.74E-04 -6.237 -0.011 0.4 Peak values of floor displacements (as % of building height = 37.38E-03 3 Modes 2.282 -0.203 0.0 -16.4 -11.7 2 Modes -13.0 -46.133 0.253 0.4 -53.253 RHA (All Modes) 0.7 -50.9 -15.006 -0.9 -11.245 -0.76E-03 1.003 -0.2 -11.28E-03 2.0 -2.15E-03 2.6 -17.203 0.260 0.080 0.1 -11.8 -15.4 -4.097 0.124 0.332 0.370 0.328 0.156 0.24E-03 -2.157 0.9 -16.8 -22.09E-03 -1.2 -57.44E-03 3.3 -41.227 0.267 -0.069 0.4 0.274 0.31E-03 2.286 0.048 0.010 0.4 -9.7 -21.00E-03 3.74E-03 -1.071 0.47E-03 1.4 -7.2 -16.9 -14.4 -22.9 -13.9 -13.76E-03 -1.237 0.230 -0.026 0.048 0.80E-04 3.227 0.062 -0.197 0.226 0.203 0.407 0.042 0.04E-03 3.043 -0.2 -12.042 0.9 -12.313 0.92E-04 -1.157 0.31E-03 Joint Rotation (rad) Combined (MPA) 1 Mode 2.43E-04 -1.121 0.179 0.9 -14.090 0.261 -0.3 Table 3.7 -15.296 -0.9 -15.78E-04 2 Modes 2.89E-03 -2.22E-03 -2.90E-03 1.270 0.97E-03 1.08E-03 2.00E-03 -1.57E-03 1 Mode -23.38E-03 2.65E-03 2.012 -0.264 Error (%) Modal Response 1 Mode 2 Modes 3 Modes -23.229 0.229 0.060 0.9 3 Modes -12.9 -13.235 0.331 0.2 -4.253 -0.73E-05 3.2 -0.31E-03 Error (%) RHA (all modes) 2.

1 0.3 0.5 Ground 0 0.25 × El Centro ground motion.10. Heightwise variation of floor displacements and story drifts from MPA for 0.2 0.eps Ground 0 0.3 0.4 Story Drift Ratio (%) 0. 3. shading indicates modal combination error 25 .(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.2 0.4 Displacement/Height (%) 0.1 0.eps fig3_10b.5 0.6 Fig.

26 .

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

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

(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

n t = .9 • 0 −50 10 −50 10 n=3 ur3 (cm) 0 • 5. n = 1. Fsn / Ln ¨ g(t ) u (a) Static Analysis of Structure (b) Dynamic Analysis of Inelastic SDF System Fig. and 3. and (b) approximate solution by UMRHA bg af bg 32 .89 50 r1 0 u (cm) 0 • 78. 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.s n u g t .4.3. where u g t = 3.51 • ur2 (cm) 0 ur2 (cm) n=3 14.02 −150 n=2 −150 50 n=2 14. 2.Forces sn Unit mass An(t ) st rn ωn. Roof displacement due to p eff. 4.0 ¥ El Centro ground motion: (a) exact solution by NL-RHA.575 5 10 15 20 Time (sec) 25 30 ur3 (cm) 0 • 5. ζn.112 5 10 15 20 Time (sec) 25 30 −10 0 −10 0 Fig. 4.

The errors are slightly larger in drift than in displacement. Top story drift due to p eff.(a) Nonlinear RHA 20 n=1 20 (b) UMRHA n=1 • 5.1 Underlying Assumptions and Accuracy The approximate solution of Eq.5. Such comparison for roof displacement and top-story drift is presented in Figs.5.n t (n = 1.4 and 4.2.855 ∆r1 (cm) 0 ∆r1 (cm) n=2 • 6. where u g t = 3.817 5 10 15 20 Time (sec) 25 30 −10 0 −10 0 Fig.0 ¥ El Centro ground motion. 4. and 3.6) by UMRHA is compared with the “exact” solution by nonlinear RHA.15) is strictly valid only for linearly elastic systems.956 5 10 15 20 Time (sec) 25 30 ∆r3 (cm) 0 • 5. but even for this very intense excitation. both for 3. respectively. (4.0 ¥ El Centro ground motion: (a) exact solution by NL-RHA. (3. (4. 4.4) and (4. n = 1.008 • 6.s n u g t . (2) the superposition of responses to p eff.n t = .Dn relation is 33 bg bg . 2. and (3) the Fsn Ln . this intense excitation was chosen to ensure that the structure is excited well beyond its linear elastic limit.2… N ) according to Eq. and (b) approximate solution by UMRHA af bg bg 4. 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. the errors in either response quantity are only a few percent.33 0 −20 20 −20 20 n=2 ∆r2 (cm) 0 ∆r2 (cm) 7.744 • 0 −20 10 n=3 −20 10 n=3 ∆r3 (cm) 0 • 5.5)] is neglected.

8) in UMRHA. which is different than the Fsn Ln . (3. Although most commercially available software cannot implement such displacement-controlled analysis. 4.8) before it can be solved? Because Eq. (4. (3. At the yield point. this distribution seems to be a rational choiceeven for inelastic systemsto determine Fsn in Eq.8) governing Dn t is based on Eq.20) to some bg pre-determined roof displacement. How to convert this Vbn . the relationship between lateral forces f s and Dn in Eq. Therefore. Although several approximations are inherent in this UMRHA procedure. (4.6a. A bilinear idealization of this pushover curve for the nth-“mode” is shown in Fig.Dn relation to be determined in Eq. The structure is pushed using the force distribution of Eq.10) . when specialized for linearly elastic systems it is identical to the RHA procedure of Section 3. Therefore we impose this constraint in developing the UMRHA procedure in this section and modal pushover analysis in the next section.approximated by a bilinear curve to facilitate solution of Eq. (4. (4.Dn curve.urn pushover curve to the Fsn Ln .12) for floor displacements. 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.2 Properties of the nth-“mode” Inelastic SDF System How is the Fsn Ln . the only force distribution that produces displacements proportional to f n is given by Eq.2. (4.9).9) should be determined by nonlinear static analysis of the structure as the structure undergoes displacements u = Dnφ n with increasing Dn . it can conduct a force-controlled nonlinear static analysis with an invariant distribution of lateral forces. and the base shear Vbn .Dn relation? The two sets of forces and displacements are related as follows: V Fsn = bn Γn Dn = urn Γ nf rn 34 (4. (3. is plotted against roof displacement urn . the base shear is Vbny and roof displacement is urny . However. 4. When implemented by commercially available software. such nonlinear static analysis provides the so-called pushover curve.4. within the linearly elastic range of the structure. The overall errors in the UMRHA procedure are documented in the examples presented in Section 4.1.20).

which may differ from the period of the corresponding linear system. Knowing Fsny Ln and Dny from n Eq.11). 4. (4. The two are related through Fsny Ln 2 = w n Dny (4.6a is 2 kn = ω n Ln . 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 (4.2.13) This value of Tn .12) implying that the initial slope of the curve in Fig.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. should be used in Eq. 4.6b. which is not a meaningful quantity.6. 4.11) * in which M n = Ln Γn is the effective modal mass (Chopra. (4. Properties of the nth-“mode” inelastic SDF system from the pushover curve Equation 4. In contrast. 4.10 enables conversion of the pushover curve to the desired Fsn Ln − Dn relation shown in Fig.8). 2001.6b is w 2 . Section 13. 35 .5). where the yield values of Fsn Ln and Dn are Fsny Ln = Vbny * Mn Dny = urny Γ nf rn (4. the initial slope of the pushover curve in Fig.

4. the peak value of Dn t . Idealize the pushover curve as a bilinear curve (Fig. of the total response r t obtained in Step 8.6a). details are available in Appendix A: 1.2. which is summarized next as a sequence of steps.20)]. for linearly-elastic vibration of the building. 6.4. r o . An t . ωn . Repeat Steps 2 to 6 for as many modes as required for sufficient accuracy. is now determined by 36 bg bg .15) and (3.6b) by utilizing Eq. For the nth-mode. Consider a nonlinear static analysis of the structure subjected to lateral forces distributed over the building height according to s* [Eq. 4. (4. (3. of the nth-“mode” inelastic SDF system (Fig. n 3. Compute the natural frequencies.12) and (3.11). 2. develop the base-shear – roof-displacement ( Vbn − urn ) pushover curve for the force distribution s* [Eq. and pseudo-acceleration history. (3.3b) with force-deformation relation of Fig. φn . This value of the roof displacement is given by Eq. 4.3 MODAL PUSHOVER ANALYSIS bg bg A pushover analysis procedure is presented next to estimate the peak response rno of the inelastic MDF system to effective earthquake forces p eff . Calculate histories of various responses by Eqs.n t . (3. Convert the idealized pushover curve to the Fsn Ln − Dn relation (Fig. Dn (t ) . n (3. 4. Combine the “modal” responses using Eqs.21) where Dn .16) to determine the total response. Calculate the peak value. 7. Compute the deformation history.20)]. (3. 4.6b. and modes.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. 9. the first two or three modes will suffice. Typically. 5. with the structure is pushed to the roof displacement urno . 8.13). 4.

However. which is summarized next as a sequence of steps.2. e. joint rotations.2.2 and 3. 4. governed by Eq. (4. At this roof displacement. As shown in Sections 3. bg bg bg bg Thus. as mentioned earlier. (4. the lateral force distribution used possesses two properties: (1) it appears to be the most rational choice among all invariant distribution of forces.6).g.8).6 and 7. story drifts. alternatively. Eq. details are available in Appendix B. and (2) it provides the exact modal response for elastic systems. Calculate the peak roof displacement urno associated with the nth-“mode” inelastic SDF system from Eq. (3. This pushover analysis. for elastic systems. seems reasonable. it can be determined from the inelastic response (or design) spectrum (Chopra.1 Summary The peak inelastic response of a building to earthquake excitation can be estimated by the MPA procedure just developed. are combined using an appropriate modal combination rule. (4. 4.6b by solving Eq.n t . This application of modal combination rules to inelastic systems obviously lacks a theoretical basis. 37 . each determined by one pushover analysis.21). 5.4. Sections 7. It provides results for elastic buildings that are identical to the well-known RSA procedure (Section 3. Steps 1 to 4 of the MPA are same as those for UMRHA. 4. or from the inelastic response (or design) spectrum. we will refer to rno as the peak “modal” response even in the case of inelastic systems. etc. the pushover analysis provides an estimate of the peak value rno of any response rn t : floor displacements.. (3. although somewhat intuitive for inelastic buildings.18). it seems reasonable because it provides results for elastic buildings that are identical to the well-known RSA procedure described in Section 3. to obtain an estimate of the peak value ro of the total response. 2001.3. Compute the peak deformation. of the nth-“mode” inelastic SDF system (Fig. 6.3) because.3. The peak “modal” responses rno .12). as described in Section 4.solving Eq. Dn .3b) with force-deformation relation of Fig. rno also represents the exact peak value of the nth-mode contribution rn t to response r t . plastic hinge rotations.8). The response value rno is an estimate of the peak value of the response of the inelastic system to p eff.

8. From the total rotation of a plastic hinge. 4. Repeat Steps 3 to 8 for as many “modes” as required for sufficient accuracy. 4. The peak values of floor displacements and story drifts including one. and the percentage errors in the approximate results. and the “exact” response from nonlinear RHA for the roof displacement and top-story drift. Determine the total response by combining the peak “modal” responses using the SRSS combination rule of Eq.2). two.9. 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. the 38 . and three modes are compared with the “exact” values in Fig.4.1 Uncoupled Modal Response History Analysis The structural response to 1. two. The peak values of response are as noted. and compared with the results of a rigorous nonlinear RHA using the DRAIN-2DX computer program. This is to be expected. respectively.7 shows the individual “modal” responses. determined by the UMRHA procedure. say.18). the peak roof displacement due to each of the three “modes” is ur1o = 48. Observe that errors tend to decrease as response contributions of more “modes” are included.7.53 cm. ur 2o = 11. At urno . Typically. subtract the yield value of hinge rotation to determine the hinge plastic rotation.4 COMPARATIVE EVALUATION OF ANALYSIS PROCEDURES The response of the 9-story building described earlier is determined by the two approximate methods: UMRHA and MPA.4.2. although the trends are not systematic as when the system remained elastic (Section 3.3 cm.7 cm. is presented next. in contrast to modal analysis (Section 3. 8. and ur 3o = 2. the “exact” results. The peak values of displacements of all floors and drifts in all stories are presented in Tables 4. 4. and the errors in the approximate results are shown in Fig.1 and 4. also included are the combined responses due to one.0. in particular.1). 4. Figure 4. extract from the pushover database values of other desired responses. and three “modes”. the UMRHA procedure lacks a rigorous theory. This deficiency also implies that. with.0 to 3. (3. the combined response due to three “modes”. the first two or three “modes” will suffice.5 ¥ the El Centro ground motion including the response contributions associated with three “modal” inelastic SDF systems. rno . 9. three “modes” included.

0.25. and 3.7.62 5. 4.25 (Tables 3. recall that the computed errors have been presented earlier for ground motion multipliers 0.3 11.7 • −80 80 ∆r1 (cm) u (cm) "Mode" 1 −12 r2 0 ∆r2 (cm) 12 0 u (cm) −80 80 −12 r3 0 • 2. sec 25 30 r −12 0 Fig. the errors in story drifts are larger compared to floor displacements.5.0. For each excitation.6 5 10 15 20 Time. for a fixed number of “modes” included.5.2). 1. defined as the El Centro ground motion multiplied by 0.38 6. For this purpose the UMRHA and exact analyses were repeated for ground motions of varying intensity.1 and 4.44 • "Mode" 2 r1 0 • 48. However. the errors in responses computed by UMRHA including three “modes” relative to the “exact” response were determined.5 × El Centro ground motion: individual “modal” responses and combined response from UMRHA.75.1 and 3. 0. (a) Roof Displacement (b) Top Story Drift 80 12 0 • 3. in particular. the degree to which the system deforms beyond its elastic limit.1 and 3.24 • NL−RHA r −12 12 ∆r (cm) u (cm) 0 5 10 15 20 Time.1 −12 12 ∆r (cm) u (cm) UMRHA 3 "Modes" 0 • 7.0.2) and 1.85.2) if the system yields significantly versus if the system remains within the elastic range (Tables 3. Response histories of roof displacement and top-story drift due to 1. 1.88 u (cm) "Mode" 3 −80 80 0 −80 80 0 −80 0 • 44. Next. sec 25 30 • 48.2). 0.response is much less accurate (Tables 4. 2. we investigate how the errors in the UMRHA vary with the deformation demands imposed by the ground motion. and total response from NL-RHA 39 .5 (Tables 4. 0. just as for elastic systems.53 ∆r3 (cm) 12 0 • 2.1 and 4.

Two versions of the pushover curve are 40 . Heightwise variation of floor displacements and story drift ratios from UMRHA and NL-RHA for 1. To interpret these results. Heightwise variation of error in floor displacements and story drifts estimated by UMRHA including one.11. two. and the error envelope for each case.5 1 1. it will be useful to know the deformation of the system relative to its yield deformation.20)] for the first three n modes of the system are shown in Fig.5 Displacement/Height (%) 2 Ground 0 0. 4.(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 0.10b). the pushover curves using force distributions s* [Eq. Shown is the error in each floor displacement (Fig. indicated by a ground motion multiplier. in each story drift (Fig. 4.5 ¥ El Centro ground motion Figure 4. 4. 4. For this purpose. 4.9. with the peak displacement of each “modal” SDF system noted for each ground motion multiplier.5 Fig.5 1 1. or three “modes” for 1.10a).5 Story Drift Ratio (%) 2 2.5 × 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. (3.10 summarizes the error in UMRHA as a function of ground motion intensity.8.

11). and (4) are larger in story drift compared to floor displacements.53 cm. respectively.75.3 cm. determined from “exact” analyses. Perhaps this explains why the errors are large at this excitation intensity. 41 .5 ¥ the El Centro ground motion. the values determined by RHA of the nth-mode inelastic SDF system (Fig.7 cm. and plastic hinge rotations at the external beam end at each floor level (Table 4. The structure is pushed using the force distribution of Eq.6 and 4. (3.0. as mentioned in Section 4. and 2. as mentioned in Section 3.2. the errors in truncating the higher mode contributions are negligible.7). First. 11. drifts in all stories (Table 4.1). and. (4. the first reason mentioned above seems to be the primary source for the errors. For more intense excitations. even though the system remains essentially elastic. the pushover curve for each “mode” is idealized by a bilinear curve in solving Eq.4. Additional errors are introduced in UMRHA of systems responding beyond the linearly elastic limit for at least two reasons.4. and 3 (Fig. the peak values of displacements at all floors (Table 4. The idealized curve for the first “mode” deviates most from the actual curve near the peak displacement corresponding to ground motion multiplier 1. 4. 4.2. UMRHA lacks a rigorous theory and is based on several approximations. Second. Figure 4. The location of plastic hinges and their rotations.5). 3. 4. 4.0.included: the actual curve and its idealized bilinear version. (3) maintain roughly similar values for more intense ground motions.01 (Fig.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. Each of these three pushover analyses provides the pushover curve (Fig.4).20) with n = 1. 2. (2) increase rapidly as the ground motion multiplier increases to 1. the ductility factor for the first mode system is only 1. were noted but are not shown here. The system remains elastic up to ground motion multiplier 0.10) for each “modal” inelastic SDF system (Figs 4.75.10 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) to roof displacements urno = 48.11a).3).

410 1.104 0.260 -15.8 1.495 1.820 -0.6 2.914 2.5 28.256 -1.722 0.5 3 0 0 0.430 1.5 1 1.698 1.938 1.0 9.9 12.(a) Floor Displacements 60 50 40 Error Envelope Error for Floor No.4 -1.8 0.806 -0.256 1.942 1.187 -0.044 1.5 -3.317 0.003 0.4 4.4 1.863 1.942 -0.072 1.220 -0.138 1.0 1.490 1.065 0.763 -15.096 0.863 0.727 1. 4.315 -0.5 2 GM Multiplier 2.5 3 Fig.5 1 1.5 10.2 6.0 11.407 -10.010 0.478 0.071 -0.526 -0.5 2 GM Multiplier 2.7 Table 4.982 9.068 0.2 4.2 1.844 -25.088 10.241 -1.9 31.5 9.3 6.5 × 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” -0.9 12.1 Peak values of floor displacements (as % of building height = 37.8 1.820 -19.235 -0.136 1.819 2.983 1.513 0.8 1.772 Error (%) NL 3 1 2 3 RHA “Mode” “Modes” “Modes” “Modes” 1.6 4.2 12.410 -1.070 1.133 1. Noted 7 6 Error (%) 30 1 Error (%) 30 5 8 20 6 5 2 20 2 4 9 9 10 0 0 10 3 8 7 3 1 4 0.121 -0.0 11.126 0.338 -1.214 0.900 -10.663 -0.472 1.135 9.370 -0.971 1. Noted 60 50 40 (b) Story Drifts Error Envelope Error for Story No.5 9.009 -0.379 1.663 0.241 1.2 Peak values of story drift ratios (as % of story height) from UMRHA for 1.3 1.3 42 .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 -1.298 -0.071 0.200 8.216 1.473 -22.1 1.426 -1.668 -23.1 1.7 14.072 -1.707 1.220 0.201 1.291 0.540 0.1 3.484 0.8 1.298 0.350 -0.226 -0.366 0.811 1.914 -0.376 -1.293 1.1 0.256 1.490 -1.376 1.049 -0.9 5.9 16.154 0.372 -1.120 1.130 0.018 0.806 0.852 1.7 31.3 8.8 14.14 m) from UMRHA for 1.4 -7.616 -0.201 -1.554 1.333 0.945 -37.371 0.283 1.575 -41.877 0.372 1.003 -31.3 0.10.373 -0.938 -1.2 1.1 8.9 31.079 0.5 18.5 28.055 -0.214 -0. and (b) story drifts Table 4.0 7.033 0.248 Error (%) NL 3 1 2 3 RHA “Mode” “Modes” “Modes” “Modes” 0.057 -0.366 -0.169 0.338 1.513 -0.676 0.0 2.202 11.8 1.856 2. Errors in UMRHA as a function of ground motion intensity: (a) floor displacements.2 22.371 -0.3 25.6 1.751 1.201 -0.0 -9.

significant improvement is achieved by including response contributions due to the second “mode”.3 and 4. 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. 4.4.5 present estimates of the combined response according to Eq. this error is not especially significant because the hinge plastic rotation is very small. The locations of plastic hinges shown in Fig.13 and Table 4. The second “mode” was necessary to identify hinges in the upper stories.13 and Tables 4.14.3 and 4. 4. 4. MPA failed to identify the plastic hinges at the column bases in Fig. 4.1 and 4.18). were determined by four analyses: MPA considering one “mode. and three “modes. For example. although the error is recorded as 100% if MPA estimates zero rotation when the nonlinear RHA computes a non-zero value.2).12 and 4.12 and Tables 4. but the third “mode” contributions do not seem especially important (Fig. the results were not always accurate. the third “mode” does not contribute because this SDF system remains elastic (Fig. However. Fortuitously. 4.3 through 4. significantly smaller than in UMRHA (compare Fig. but was more successful when the excitation was more intense (results are not included).4). the errors in the modal pushover results are. 4.3 and 4. 43 . 4. two. however.4).” and nonlinear RHA. and are within a few percent of the exact values for the upper stories. Obviously.Figures 4.” respectively. The errors are especially large in the hinge plastic rotations estimated by the MPA procedure.12 and Tables 4. even if three “modes” are included (Fig.” two “modes. 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%. in general.9 and Tables 4. overestimated by up to 18% in the middle stories. 4. for two or three modes included. Pushover analysis seems to be inherently limited in computing accurately hinge plastic rotations. Furthermore. considering one.” and three “modes. (3. 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.5). and the errors in these estimates relative to the exact response from nonlinear RHA.3 and 4. 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. 4. especially in estimating the story drifts (Fig. The first “mode” alone is inadequate.4 with Tables 4.13 and Tables 4. the hinges identified in beams at the 6th floor were at variance with the “exact” results.13 with Fig.14.11c). As shown in Figs. The drifts are underestimated by up to 13% in the lower stories.

5 2 3 Actual Idealized 2000 0 0 0.5 2 3 4000 2000 0 0 Actual Idealized 0. α = 0.0.(a) "Mode" 1 Pushover Curve 12000 10000 u = 36.0 × El Centro ground motion 44 .11.5 1 0. V = 7616 kN. V = 4952 kN.5 0. 4. 0.2 cm. 1. and 3.75 Actual Idealized 20 40 60 Roof Displacement (cm) (b) "Mode" 2 Pushover Curve 80 12000 10000 u = 9.85 0.75.5.9 cm.5.85 0.85 Base Shear (kN) 8000 6000 4000 2000 0 0 0.0. “Modal” pushover curves with peak roof displacements identified for 0.75 0.6 cm.25 5 10 15 20 Roof Displacement (cm) 25 (c) "Mode" 3 Pushover Curve 12000 10000 u = 4. V = 5210 kN.75 0. α = 0.13 y by Base Shear (kN) 8000 6000 1 1. 0.25 1 2 4 6 Roof Displacement (cm) 8 10 Fig. 2.14 y by Base Shear (kN) 8000 6000 4000 1.5 0.25 0. 1.19 y by 3 2 1.5 0.25. α = 0.

4.75. 4.15 summarizes the error in MPA considering three “modes” as a function of ground motion intensity.5 Displacement/Height (%) 2 Ground 0 0.3. implying that the modal combination errors contained in these procedures are acceptable.5 1 1.5 Fig. 4. 45 . as discussed in Sections 3.4. While various sources of errors in UMRHA. implying excitations weak enough to limit the response in the elastic range of the structure. and the error envelope for each case.15) for ground multipliers larger than 1. Shown is the error in each floor displacement (Fig. 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. 4. now standard in engineering practice. The errors are only weakly dependent on ground motion intensity (Fig.3 for elastic systems (or weak ground motions). implying excitations intense enough to cause significant yielding of the structure. the MPA procedure is equivalent to the RSA procedure.15).4. an observation with practical implications.(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 0.0. the errors in MPA were larger for ground motion multipliers less than 0. the errors in MPA were fortuitously smaller than in UMRHA (compare Figs.5 Story Drift Ratio (%) 2 2.5 × El Centro ground motion.2 and 3. 4. In this case.12 Heightwise variation of floor displacements and story drift ratios from MPA and NL-RHA for 1.15b). However.10 and 4. indicated by a ground motion multiplier. shading indicates errors in MPA including three “modes” Figure 4.5 1 1. UMRHA is essentially exact. whereas MPA contains errors inherent in modal combination rules. each story drift (Fig. identified in Section 3.4 also apply to MPA.15a). As mentioned in Section 3.

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

066 -0.763 -14.0 -100.503 1.7 -12.0 -50.244 0.00E+00 1 “Mode” 7.687 0.694 1.5 Peak values of hinge plastic rotations (radians) from MPA for 1.2 -100.910 1.8 1.371 -0.18E-03 7.982 13.2 0.250 0.105 0.667 1.1 46.518 1.00E+00 3 “Modes” 7.5 × 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 0.053 1.5 2.202 8.5 2.1 62.426 15.304 -1.900 -0.130 0.233 1.53E-03 7.8 0.00E+00 2 “Modes” 7.737 0.88E-03 0.018 0.3 -3.668 -13.1 1.4 1.76E-03 4.895 1.745 1.2 -100.209 1.05E-03 2.00E+00 0.00E+00 “Mode” 3 0.6 -9.640 1.9 2.99E-03 6.14 m) from MPA for 1.336 Error (%) NL 3 1 2 3 RHA “Mode” “Modes” “Modes” “Modes” 0.37E-03 1.5 10.0 47 .8 -29.9 5.72E-03 7.4 -8.5 × 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” 1.1 -8.737 1.36E-03 6.0 -5.37E-04 Error (%) “Modal” Response 1 “Mode” -32.55E-03 3.8 -29.50E-10 3.6 -8.156 -0.Table 4.0 -100.7 1.259 1.125 -1.1 13.1 18.007 1.6 -44.473 -15.0 -100.8 17.581 0.478 0.116 1.9 7.9 1.666 Table 4.222 0.5 Table 4.55E-03 3.154 0.6 -44.071 -0.19E-10 3.8 7.781 0.372 0.756 0.517 1.705 -1.667 -1.135 -7.8 -4.222 0.752 1.72E-03 7.00E+00 0.018 -0.2 -3.37E-03 1.37E-03 1.36E-03 6.7 1.266 -0.980 0.733 1.012 1.55E-03 3.2 -4.804 1.2 1.351 -0.37E-03 1.22E-10 NL RHA 1.066 -0.00E+00 0.738 1.3 11.399 0.1 62.705 1.0 1.015 0.00E+00 0.8 -12.72E-03 7.575 -53.5 2.00E+00 3.756 0.514 -1.2 11.2 6.516 0.435 0.298 0.168 -0.00E+00 0.4 Peak values of story drift ratios (as % of story height) from MPA for 1.5 × El Centro ground motion Hinge Plastic Rotation (rad) “Modal” Response Combined (MPA) “Mode” 1 7.0 2 “Modes” -32.6 13.033 -0.9 -100.116 1.9 0.36E-03 6.5 -6.527 -0.220 1.315 0.2 0.200 8.00E+00 0.5 10.5 7.6 -7.3 Peak values of floor displacements (as % of building height = 37.233 1.00E+00 0.00E+00 0.88E-03 0.9 -4.60E-04 7.015 0.0 15.003 -16.02E-03 0.781 0.76E-03 4.6 7.304 1.071 0.820 -7.8 -29.0 3 “Modes” -32.298 -0.60E-03 2.895 1.26E-04 9.197 -0.00E+00 0.399 0.3 13.02E-03 0.018 -0.3 -3.00E+00 0.9 1.3 1.72E-03 7.053 -1.414 1.8 1.594 -1.057 0.429 0.945 -49.8 -6.049 -0.8 “Mode” 3 -1.007 1.055 0.652 1.118 0.1 46.331 1.338 1.1 46.00E+00 0.614 0.498 1.305 -0.640 -1.088 12.068 0.10E-02 9.503 -1.76E-03 4.311 0.00E+00 Floor 1st 2nd 3rd 4th 5th 6th 7th 8th 9th “Mode” 2 0.6 1.00E+00 0.980 -0.00E+00 0.728 1.641 1.611 0.36E-03 6.00E+00 0.220 1.879 1.6 13.1 62.00E+00 0.942 6.009 0.8 -6.683 1.844 -7.02E-03 3.76E-03 4.176 0.88E-03 1.00E+00 0.0 -50.00E+00 0.190 -0.407 -27.260 -14.911 0.02E-03 3.00E+00 0.00E+00 0.101 -0.5 1.581 0.116 1.9 -100.2 -100.

and overestimated by up to 29% in the middle stories and by up to 13% in the upper stories.6 through 4.16a—which is slightly different than the one excluding gravity loads (Fig.17 and Tables 4. The first “mode” alone provides adequate estimates of floor displacements. 4. 3. the third “mode” contributions do not seem especially important (Fig.4. Obviously gravity load effects will influence the seismic demands due to the first “mode. the results presented so far did not include gravity load effects.5 × El Centro ground motion are presented below. and the structure is pushed using the force distribution of Eq. They are now included in the pushover analysis of the structure for the first “mode” only.20) with n = 1 to a target roof displacement. MPA including two “modes” underestimates the displacements of lower floors by up to 6% and overestimates the upper floor displacements by up to 22%.18) considering one. 4.7). Figures 4. The errors are especially large in the plastic hinge rotations estimated by the MPA procedure even if three “modes” are included (Fig. 4.5). (3. (3.” however. 4.6 and 4. two. 4.7.8 present estimates of the combined response according to Eq. Static analysis of the structure for gravity loads provides the initial state—forces and deformations—of the structure.53 cm.17 and Tables 4. and plastic hinge rotations at selected external beam end at each floor level (Table 4. The response contributions of the second and third modes are the same as before (Tables 4.11a).4) to roof displacement ur1o = 52.20) with n = 1 (Fig. (3.3 .4. drifts in all stories (Table 4.7 cm and ur 3o = 2.6).17 and 4.8). as are the roof displacements ur 2o = 11. and three “modes.18 and 48 .16 are unchanged. Results of such analyses for the selected building subject to 1.7).6 and 4. resulting in the pushover curve shown in Fig. 4. but it is inadequate in estimating the story drifts (Fig. Significant improvement is achieved by including response contributions due to the second “mode.” and the errors in these estimates relative to the exact response from nonlinear RHA.6 and 4.4.7). the structure is pushed using the force distribution of Eq.18 and Tables 4. Also presented at these roof displacements are the displacements at all floors (Table 4. As shown in Fig. The story drifts are underestimated by up to 7% in the lower stories.18 and Tables 4.” these effects will modify the combined modal response as well as the results of nonlinear RHA. The pushover curves for the second and third modes included in Fig.0 cm. Starting with its initial state under gravity loads.” but not the contributions of higher “modes. 4.3 Modal Pushover Analysis with Gravity Loads To evaluate the accuracy of the dynamic response of the system.

14.5 2 GM Multiplier 2. 4.5 2 GM Multiplier 2.5 1 1. (a) MPA. Noted 60 50 40 (b) Story Drifts Error Envelope Error for Story No. 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 0.5 3 Fig. Most pushover analysis procedures do not seem to compute to acceptable accuracy plastic hinge rotations.5 3 0. Locations of plastic hinges determined by MPA considering one.5 1 1. Errors in MPA as a function of ground motion intensity: (a) floor displacements.8). 4. and (b) story drifts 49 . 1−"Mode" • • • • • • • • •• •• •• •• •• •• •• •• •• •• •• •• •• •• • • •• • •• • • • • • •• •• • •• •• •• •• (b) MPA.15. 2−"Modes" •• •• •• •• •• •• •• •• •• •• •• •• •• •• • • • • •• • •• (c) 3−"Modes" • • • • • • • •• •• • •• •• •• •• •• •• •• •• •• •• •• •• •• •• •• •• •• •• • • • • •• • •• • • • • • • • • • (d) Nonlinear RHA •• •• • •• •• •• •• • •• •• •• •• •• •• •• • •• •• • •• •• •• •• • • • • • • •• •• •• Fig.5 × El Centro ground motion (a) Floor Displacements 60 50 40 Error Envelope Error for Floor No. and three “modes” and by NL-RHA for 1.Table 4. two.

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

5 2 3 4000 2000 0 0 Actual Idealized 0.85 0. and 3.16. 1.25. V = 5210 kN.9 cm. V = 7433 kN.75 Actual Idealized 20 40 60 Roof Displacement (cm) 80 (b) "Mode" 2 Pushover Curve 12000 10000 u = 9.5 0.14 y by Base Shear (kN) 8000 6000 4000 1.85 0.6 cm.75 0. V = 4952 kN.0.85. α = 0.5 1 0. “Modal” pushover curves with gravity loads included. noted are peak values of roof displacement for 0.(a) "Mode" 1 Pushover Curve 12000 10000 u = 36.5 0. α = 0.75 0.5 0.85 Base Shear (kN) 8000 6000 4000 2000 0 0 0. α = 0.19 y by 3 2 1. 0.25 0.0.75.5 2 3 Actual Idealized 2000 0 0 0.13 y by Base Shear (kN) 8000 6000 1 1. 2.3 cm.25 1 2 4 6 Roof Displacement (cm) 8 10 Fig. 0.25 5 10 15 20 Roof Displacement (cm) 25 (c) "Mode" 3 Pushover Curve 12000 10000 u = 4.50. 0.0 × El Centro ground motion 51 . 4.

two. Heightwise variation of floor displacements and story drift ratios from MPA and NL-RHA for 1. 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. gravity loads included.5 Displacement/Height (%) 2 Ground 0 0.5 × El Centro ground motion. story drifts.5 1 1.(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. 4.5 Story Drift Ratio (%) 2 2. and hinge plastic rotations estimated by MPA including one.5 1 1.18. 4.5 Fig. Errors in floor displacements.17. and 52 .

Errors in MPA as a function of ground motion intensity: (a) floor displacements. Noted 60 50 40 (a) Story Drifts Error Envelope Error for Story No. 4.5 1 1.20.5 1 1.5 2 GM Multiplier 2. gravity loads included 53 . 3−"Modes" (d) Nonlinear RHA • • • • • • • •• •• •• •• •• •• •• •• •• •• •• •• •• •• •• •• •• •• •• •• •• • • • • •• •• •• • • • • • • • • •• •• • •• •• •• •• • • •• •• • •• •• •• •• • • •• •• • •• •• •• •• • • • • • • •• •• •• Fig. and three “modes” and by NL-RHA for 1.(a) MPA. two.5 3 0.5 × El Centro ground motion. Noted Error (%) 30 20 6 8 1 5 2 4 3 9 7 Error (%) 30 4 5 20 6 9 7 1 10 0 0 10 0 0 8 2 3 0. 4. Locations of plastic hinges determined by MPA considering one. gravity loads included (a) Floor Displacements 60 50 40 Error Envelope Error for Floor No. 2−"Modes" •• •• •• •• •• •• •• •• •• •• •• •• •• •• • • • • •• •• •• (c) MPA.5 3 Fig. 1−"Mode" • • • • • • • •• •• •• •• •• •• •• •• •• •• •• •• •• •• •• • • •• •• •• • • • • • •• •• •• •• •• •• •• (b) MPA.5 2 GM Multiplier 2. and (b) story drifts.19.

8 0.0 -100.3 -22.998 1.311 0.0 -30.3 9.850 1.8 28.3 13.5 × El Centro ground motion.2 -2.372 0.744 1.908 1.3 -22. 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” 1.04E-02 8.00E+00 Error (%) “Modal” Response 1 “Mode” -32.4 1.19E-03 1.5 21.1 21.8 -32.305 -0.5 -5.1 13.064 -10.1 1.8 0.4 1.35E-10 3.3 -100.00E+00 0.953 15.434 0.88E-03 0.237 0.2 4.665 0.1 4.2 1.2 2.00E+00 0.836 -0.998 21.998 0.831 0.908 -1.156 -0.783 1.9 31.2 -3.319 1.603 1.0 16.00E+00 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” 0.35E-03 8.176 0.5 1.213 1.6 1.4 -6.55E-03 3.2 4.Table 4.23E-03 0.850 -1.00E+00 0.813 1.353 -23.19E-03 1.515 -50.55E-03 3.821 -1.7 16.130 0.00E-03 5.507 1.996 0.066 -0.854 0.754 1.00E+00 0.102 1.310 1.5 × El Centro ground motion.14 m) from MPA for 1.687 0.4 -4.594 -1.527 -0.319 1.7 1.00E+00 0.938 1.009 0.603 -1.9 -3.107 1.315 0.00E+00 0.822 1.0 37.9 -6.037 0.5 × El Centro ground motion.831 0.516 0.125 -1.19E-04 5.04E-10 3.00E+00 0.00E+00 0.5 1.23E-03 0.098 20.6 2.0 2 “Modes” -32.429 -1.11E-03 9.8 1.4 26.237 0.13E-03 5.454 1.933 1.7 -2.19E-03 1.371 -0.5 54 .00E+00 0.6 0.066 -0.037 -0.2 4.2 21.913 7.213 1.983 1.637 0.75E-03 0.197 -0.490 -11.2 1.673 Table 4.6 Peak values of floor displacements (as % of building height = 37.00E-03 5.7 -2.78E-03 1.17E-03 9.00E-03 5.414 28.055 0.071 0.927 1.2 9.154 0.3 1.55E-03 3.466 0.88E-03 1.5 “Mode” 3 -1.23E-02 1.88E-03 0.478 0.530 1.00E+00 0.168 -0.00E+00 2 “Mode” 8.514 -1.637 0.399 -0.109 0.049 -0.821 1. gravity loads included Hinge Plastic Rotation (rad) Floor “Modal” Response “Mode” 1 1st 2nd 3rd 4th 5th 6th 7th 8th 9th 8.35E-03 8.23E-03 3.00E+00 “Mode” 3 0.6 19.5 0.26E-03 3.8 Table 4.35E-03 8.19E-03 1.2 12.128 -1.983 1.436 1.351 -0.114 1.8 Peak values of hinge plastic rotations (radians) from MPA for 1.7 19.888 1.00E+00 Combined (MPA) 1 “Mode” 8.018 -0.2 -0.667 0.2 16.057 0.11E-03 9.4 20.3 -100.0 -100.860 1.996 -0.207 18.4 -4.6 2.2 0.5 21.5 3 “Modes” -32.105 0.00E-03 5.00E+00 0.11E-03 9.00E+00 “Mode” 2 0.8 -32.434 Error (%) NL 3 1 2 3 RHA “Mode” “Modes” “Modes” “Modes” 0.102 1.0 -30.00E+00 0.0 37.830 -12.214 0.7 Peak values of story drift ratios (as % of story height) from MPA for 1.686 -7.263 1.7 4.754 0.330 1.190 -0.071 -0.00E+00 0.263 0.858 2.213 1.35E-03 8.2 9.015 0.854 0.399 0.199 16.23E-03 3.921 1.00E+00 0.00E+00 0.00E+00 3 “Modes” 8.3 -22.9 -4.5 10.3 -100.00E-10 NL RHA 1.00E+00 3.033 -0.270 -12.266 -0.00E+00 0.00E+00 0.8 9.429 1.114 -1.434 0.330 1.2 9.0 37.11E-03 9.877 -46.461 0.728 1.00E+00 0.101 -0.068 0.530 1.257 0.530 1.

etc. and the exponent k = 1 for fundament period T1 £ 0.” and nonlinear RHA. plastic hinge rotations. assumed to be linearly elastic. k = 2 for T1 ≥ 2. and 3.1 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. story drifts. joint rotations.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. j 2. Comparison of Modal and FEMA Pushover Analyses 5. gravity load effects were included in all 55 . and varies linearly in between. Specified in FEMA-273 are three distributions for lateral forces: 1.5. computed at the target displacement represent the earthquakeinduced demands on the structure. 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.5 sec . MPA considering three “modes. SRSS distribution: s* is defined by the lateral forces back-calculated from the story shears determined by response spectrum analysis of the structure. 1997] The pushover curve. a plot of base shear versus roof displacement.. 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. “Uniform” distribution: s* = m j (where the floor number j = 1. 5. 2… N ).5 sec . The floor displacements.

The “uniform” distribution overestimates all floor displacements by 17-28%.4 and Tables 5. The three FEMA force distributions are presented in Fig.11 0.177 0.4. 5. 5. plastic hinge rotation demands in Table 5.0981 0. 5.00719 0.3.4.0 cm.3b and Table 5.0381 0.0702 0.0466 0. and (c) SRSS 56 .11 0. 5.0197 0.126 0.1.21 0.11 0.165 0. The pushover curves are given in Fig.11 0. (b) ELF.0654 0.3a and Table 5. The first four floor displacements are within 5% of the “exact” value. 5. with the errors being larger for the SRSS distribution. the value determined from RHA of the first-mode inelastic SDF system for 1. pushover analyses are implemented for a target roof displacement of 52. Force distributions in FEMA-273: (a) “uniform”.367 (a) Uniform (b) ELF (c) SRSS Fig.0446 0.062 0.11 0. both presented in Section 4.1. and the locations of all plastic hinges in Fig. the floor displacement demands in Fig. Figures 5.11 0. wherein the first two are obvious and the third is determined from response spectrum analysis of the building (Appendix C).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. the story drift demands in Fig. The errors in the FEMA and MPA estimates of seismic demands relative to the “exact” demands are presented in Fig.4a. and Table 5. Also included in these presentations are the MPA results considering three “modes” and the “exact” demands from nonlinear RHA.281 0.5 times the El Centro ground motion.042 0.1. and the displacements of upper floors are overestimated by 9-22%. 5. The MPA procedure is more accurate than all the FEMA force distributions.1 through 5.2. Using each of these force distributions.112 0.3a and 5.analyses.0913 0.119 0.0896 0.11 0. 0.3.3. 5.2.

Pushover curves using three force distributions in FEMA-273: (a) “uniform”. V = 7456 kN. α = 0. V = 8530 kN. and (c) SRSS.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. α = 0. (b) ELF. gravity loads are included 57 . V = 6897 kN.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 = 38.3 cm.8 cm. 5.(a) Uniform Distribution 12000 10000 u = 33 cm.2.24 y by Base Shear (kN) 8000 6000 4000 2000 0 0 Actual Idealized 10 20 30 40 Roof Displacement (cm) 50 60 Fig. α = 0.

3b and 5.2 demonstrate that the story drift demands are greatly underestimated by all the FEMA force distributions. the errors are largest in the lower stories. Figure 5. the MPA procedure is more accurate than all the FEMA force distributions. Modal pushover analysis procedure gives estimates better than all the FEMA force distributions. Heightwise variation of floor displacements and story drift ratios estimated using FEMA-273 force distributions. (a) Floor Displacements 9th 8th 7th 6th 9th 8th 7th 6th (b) Story Drift Ratios Floor 4th 3rd 2nd 1st NL−RHA MPA.5 Ground 0 Ground 0 Fig.5 Displacement/Height (%) 2 Floor 5th 5th 4th 3rd 2nd 1st NL−RHA MPA. For the uniform distribution. (The 100% error in a 6th floor hinge is ignored as it simply represents that the MPA estimated zero rotation.4b. This seems to be an inherent limitation of pushover analysis. but it is still inaccurate.5 1 1. 7%. gravity loads included 58 .” and NLRHA.) The pushover analysis procedures considered seem incapable of computing accurately local response quantities.4c and Table 5. reaching 35%.5 Story Drift Ratio (%) 2 2. MPA including three “modes. 5. with errors reaching 37% in this example.3. 3 "Modes" FEMA Uniform ELF SRSS 0. In contrast. 3 "Modes" FEMA Uniform ELF SRSS 0.5 1 1. larger errors are noted in the upper and lower stories. errors are largest in the upper stories. such as hinge plastic rotations. For the SRSS distribution. with story drifts under estimated by. reaching 64%. reaching 31%. and overestimated by no more than 32%. whereas nonlinear RHA computed an insignificantly small value.3 demonstrate that the hinge plastic rotations estimated by all three FEMA force distributions contain unacceptably large errors. at most.Figures 5. For the ELF distribution. and Table 5.

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

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

9 15.515 FEMA Uniform ELF 27.4 0.50E-03 0.877 1.530 1.09E-03 4.0 -100.10E-02 7.3 -11.064 1.860 1.854 0.341 1.214 1.75E-03 0.270 0.992 1.04E-02 8.6 -41.23E-02 1.88E-03 1.7 7.7 19.462 1.708 0.7 Table 5.2 9.524 0.52E-04 1.00E+00 0.0 -100.Table 5.089 1.314 1.00E+00 4.9 16.083 1.436 Error (%) NL Uniform RHA 0.4 21.0 -71.730 1.209 0.23E-03 3.0 Table 5.335 2.84 1.34E-03 2.00E+00 0.399 1.3 -14.4 16.1 0.8 -100.4 -23.1.888 0.7 FEMA ELF SRSS -22.3 23.6 4.2 -4.306 1.9 0.6 14.466 0.5 Story 1st 2nd 3rd 4th 5th 6th 7th 8th 9th Uniform 2.414 1. gravity loads included Displacement /Height (%) FEMA ELF SRSS MPA 1.234 1.65E-03 7.809 0.5 15.2 5.490 0.5 -32.938 1.195 0.5 61 .35E-03 8.6 -73.9 -70.5 26.17E-03 9.0 -30.067 0.7 -29.0 -59.4 -12.00E+00 0.015 0.294 1.8 -100.1 -26.168 1.55E-03 3.998 1.672 1.6 -25.00E+00 8.673 Error (%) NL RHA 1.00E+00 2.6 -4.263 0.00E+00 NL RHA 1.11E-03 9.5 10.5 -29.323 1.8 2.221 1.975 1.00E-03 5.399 -27.2 -12.53E-02 1.0 -100.178 1.9 28.19E-03 1.623 1.566 1.355 0.03E-03 5.5 22.724 0.372 1.2 13.7 -28.5 -29.7 MPA -2.2 16.913 0.128 1.78E-03 0.9 SRSS -22.487 0.8 -2.199 27.4 16.7 26.839 0.00E+00 0.061 1.011 1.3 -22.530 1.1 10.19E-04 5.2 16.417 1.9 16.94E-03 2.2 Peak values of story drift ratios (as % of story height) from FEMA force distributions and MPA.830 2.51E-03 4.35E-10 3.3 29.560 1.098 1.6 17.789 0.7 -60.59E-03 5.3 -100.2 4.16E-03 0.0 -63.5 -3.2 10.353 1.13E-03 5.783 1.6 21.6 -6. gravity loads included Displacement /Height (%) FEMA ELF SRSS MPA 0.26E-03 3.9 -20.8 44.6 -11.318 2.8 -32.154 1.78E-03 1.14 m) from FEMA force distributions and MPA.998 1.00E+00 0.3 -4.007 1.344 0.858 1.8 7.0 -100.2 6.00E+00 0.1 -3.1 163.5 -33.399 1.8 -35.0 37.310 1.547 -27.7 21.109 1. Peak values of floor displacements (as % of building height = 37.330 1.2 9.00E-10 Floor 1st 2nd 3rd 4th 5th 6th 7th 8th 9th Uniform 1.984 1.3 Peak values of hinge plastic rotations (radians) from FEMA force distributions and MPA Hinge Plastic Rotation (rad) MPA FEMA ELF SRSS 4.4 MPA -2.62E-03 0.686 0.836 0.562 1.367 1.8 -63.667 0.58E-04 6.8 16.4 -55.351 0.5 -27.7 -15.8 17.597 0.8 Floor 1st 2nd 3rd 4th 5th 6th 7th 8th 9th Uniform 0.00E+00 Error (%) FEMA Uniform ELF SRSS MPA 24.6 22.9 -77.875 0.927 1.262 1.611 0.736 0.2 19.6 -17.0 -57.93E-03 1.953 0.3 31.5 -33.207 1.45E-03 3.4 -50.2 -100.9 11.

MPA including three “modes” and NL-RHA for 1. gravity loads included 62 .5 × El Centro ground motion. Locations of plastic hinges determined from three force distributions in FEMA-273.5.(a) FEMA: Uniform • • • • • • • • •• •• •• •• •• •• (b) FEMA: ELF • •• •• •• •• •• •• • •• • •• •• •• •• •• • • • • • •• • • • • • • • •• •• •• •• • •• •• •• •• • •• •• •• •• • • •• •• •• • • • • • •• •• (c) FEMA: SRSS • • •• •• •• •• • • •• •• •• • • •• •• •• • • • • •• •• •• • • • • • • • • • (d) MPA with 3 "Modes" • • • • • • • •• •• •• •• •• •• •• •• •• •• •• •• •• •• •• •• •• •• •• •• •• • • • • •• •• •• (e) Nonlinear RHA • • • • • • • • •• •• • •• •• •• •• • • •• •• • •• •• •• •• • • •• •• • •• •• •• •• • • • • • • •• •• •• Fig. 5.

5 2 GM Multiplier 2.6.5 1 1.5 1 1. gravity loads included 63 .5 2 GM Multiplier 2.(a) FEMA: Uniform 60 50 40 60 50 40 3 (b) FEMA: ELF Error Envelope Error for Floor No.5 3 0.5 3 Fig.5 2 GM Multiplier 2.5 1 1. Errors in floor displacements from three force distributions in FEMA-273 and from MPA including three “modes”.5 3 0 0 0.5 1 1. 5.5 2 GM Multiplier 2.5 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. 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. 5. Error in story drifts from three force distributions in FEMA-273 and from MPA including three “modes”.5 1 1.5 3 0 0 0.5 3 0 0 0.5 1 1.5 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.(a) FEMA: Uniform 100 60 50 9 8 (b) FEMA: ELF Error Envelope Error for Story No. gravity loads included 64 .5 2 GM Multiplier 2.5 2 GM Multiplier 2.7.5 2 GM Multiplier 2. Noted 80 40 9 1 Error (%) 7 Error (%) 60 30 20 10 2 5 6 4 8 40 1 6 20 2 5 3 4 3 7 0 0 0.5 1 1.5 3 Fig.

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

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

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

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

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

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

1) using the FEMA-273 procedure (Building Seismic Safety Council. develop the base-shear – roof-displacement (Vbn . e. respectively. Since the target roof displacement may not be known at the start of the procedure. Define the anchor point. φn . Compute natural frequencies.. 3. Idealize the pushover curve as a bilinear curve (Fig.1. 71 . 1. 3. DRAIN2DX (Allahabadi and Powell. For the nth-“mode”.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.g.2. for linear-elastic vibration of the building. The structure should be pushed just beyond the target (or expected) roof displacement in the selected mode. and modes. This step can be conveniently implemented in any commercially available software. 1977).1. using any numerical integration method.20): s* = mφn n n 2. iterations may be necessary.urn ) pushover curve for the force distribution s* : n 2. B.2. 3. Define the force distribution s* from Eq. Apn . Calculate the area under the actual pushover curve. 1988). of the bilinear curve at the target roof displacement.Appendix A Uncoupled Modal Response History Analysis A. ω n . e. Let the roof displacement and base shear at the anchor point be urno and Vbno . Apply force distribution of Step 2..1 incrementally and record the base shears and associated roof displacements. trapezoidal rule. A. (3. 2.g.

1. other appropriate methods can be ( used. If the error exceeds some pre-specified tolerance. urny = Vbny kn . i 3. iterations are necessary. Calculate the error = 100 × Abn − Apn ( ) ( ) ( ) ) Apn .4.6.4 to 3. Calculate area under the bilinear curve OAB. kn = 0. i i i αn =  Vbno Vbny − 1  urno urny − 1         i 3.8. determine the roof displacement.1. Replace i+1 with i and repeat Steps 3. Calculate Vbny = Vbny × Apn Abn .9.0. Calculate initial slope of the idealized bilinear curve. Draw the curve OAB by connecting the three points O. This step gives the secant stiffness at a base shear equal to 60% of the yield base shear.8. Vbny . kn .6 × Vbny . i +1 i i 3. O. i i 3. corresponding to the estimated yield i i i base shear. by connecting a straight line between origin. Calculate the yield displacement.2.6 . will be refined by an iterative procedure that seeks to equate areas under the actual and the idealized pushover curves.7.i 3.0. From the pushover data. This value. i 3. 3. obtained by judgment.4. If desired. Let the point with base shear = Vbny and roof displacement = urny be denoted as A. Calculate the post-yielding strain-hardening ratio. urn. Abn .6 × Vbny .9. A. and B with straight-line segments to obtain the idealized bilinear curve.6 × Vbny ( ) i urn . 3. i 3.5.9. Estimate the yield base shear.2. i i i 3. at base shear i equal to 0. 3.3. Vbny .4. Calculate the slope. 72 .6 . and a point on the actual pushover curve with base shear equal to i 0.

11b). of the combined responses obtained in Step 8.6 × Vbny A • Idealized αnkn B • 1 Actual k 1 • O n figA_1. r o . An (t ) . Calculate peak values.2.11a). In general first two or three modes will suffice. A. Develop the Fsn Ln − Dn relation (Fig. 4.12) and (3.2). Dn (t ) .15) and (3. Scale the horizontal axis by Γnφrn to obtain Dno = urno Γ nφrn and Dny = urny Γ n φrn (Eqs.10b and 4.4) and effective modal mass from M n = Ln Γn . 9. A. 4.10a and 4. Calculate histories of various responses using Eqs. Combine the “modal” responses using Eqs. 7. Repeat Steps 3 to 6 for as many modes as required for sufficient accuracy. (3. 4.6 ur n y ur n o ur n Fig. A. 4.3. (3.1 Idealization of nth-“mode” pushover curve 73 . Vbn Idealization of Pushover Curve Vbno Vbny 0.16). 4.eps ur n. of the nth“mode” inelastic SDF system (Fig.3b) with unit-mass and force-deformation relation of Fig. Compute the Ln and Γ n from Eq. Compute deformation history.4. and pseudo-acceleration history. Scale the vertical axis with * Mn to obtain * Fsno Ln = Vbno M n and * Fsny Ln = Vbny M n (Eqs. 6.2.13). * 4.0. (3.1. 5. 8.

5 cm and Vb1o = 8729. First three mode shapes and frequencies of the selected building were computed and are shown in Fig. The following steps illustrate the procedure to develop the idealize curve for the first “mode”.eps Dny = ur n y / Γnφr n Dno = ur n o / Γnφr n Dn Fig.). Area under the actual pushover curve.urn ) pushover curve for the force distribution s* : n 2.2 Properties of the nth-“mode” inelastic SDF system A.5. The pushover curves for the first three modes. 3. A. Ap1 = 360777 kN-cm.4 cm (10 in. 3. n 2. 1.3. A.6 kN. 3. The anchor point.3. are shown in Fig. 3. s* .1. and 12.Fsn / Ln Fsn / Ln − Dn Relationship Vbno / M* n Vbny / Mn * 1 α ω2 n n ω2 1 n figA_2.2 EXAMPLE The UMRHA procedure is implemented to calculate the response of the 9-story building described in Section 3. The force distributions. A. 25. 74 .). Following is step-by-step implementation of the procedure described in Section A.1. is defined at the target roof displacement.5 cm (25 in. 3. respectively.2. B. computed for the first three modes are shown in Fig. and third mode. second. Idealized bilinear curves for each of the three modes are included in Fig.1. 2.4.). generated using DRAIN-2DX.3. for the first. The base-shear – roof-displacement (Vbn . ur1o = 63.2.7 cm (5 in. Target displacements used to generate these pushover curves are 63. At this point. to the north-south component of the El Centro (1940) ground motion scaled up by a factor of 1.4.

k1 .3666 = 3740189 kg. The next estimate of the yield shear is Vbny = 8006.4. Error = 100 × (365100 − 360777 ) 360777 = 1.1.2. i i i 3.198%. (4. and M1 = 2736789 × 1. The results for other modes are summarized in Table A. The point A on the i i bilinear curve is defined by ur1 y = 38. The first estimate of the yield base shear Vb1 y = 8006. The results of the iterative procedure are summarized in Table A. Determined from the pushover database. α1 =  Vb1o Vb1 y − 1  ur1o ur1 y − 1 =         ( ) ( ) (8729.9 kN. i i 3.5 38.     i 3.4.6 8006.6 kN.18 = 38.6 = 22. ur1.3666. The post-yielding strain-hardening ratio. * 4. Γ1 = 1.09 cm. ur1 y = Vb1 y k1 = 8006.23 cm.2. L1 = 2736789 kg.194. The curve OAB obtained by connecting the three points O.5.2. i i i 3. 75 . 3.1.0. i 3.0. Area under the bilinear curve OAB.9. is calculated as follows. i i 3. k1 = 0.4 kN.8 22. 3. and α1 = 0.1.6 = 4803.86 cm at 0. The initial slope of the idealized bilinear curve. The Fs1 L1 − D1 relation for the first “mode” is developed as follows.09 ) − 1 = 0.09 cm and Vb1 y = 8006. Therefore. iterations are necessary.7. This value exceeds the prespecified tolerance of 0. i +1 3.4.4 ) − 1 (63.4 × (360777 365100 ) = 7911.6 × Vb1 y ( ) i ur1.6 × Vb1 y = 4803. The procedure converged after nineteen cycles to give ur1 y = 36.4 kN.9.6. Also included are the modal damping ratios and the periods calculated from Eq.3. Ab1 = 365100 kN-cm.86 = 210. Vb1 y = 7615.135.8 kN.01%. and B with straight-line segments gives the idealized bilinear curve. The yield displacement.8.18 kN/cm.13). 4.i 3.4 210.9. A.1. 3.

6. 4.40 (cm/sec2) and 76 .51 cm. 4. 9. * M1 gives Fs1o L1 = 233. 4. The peak values are also plotted in Fig.2.46 cm and D1y = 26.4.2.3.1 and 4.7. 5. The combined modal responses are presented in Fig.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. Scaling the vertical axis by Fs1 y L1 = 203.62 (cm/sec2). 4. A. 4. Scaling the horizontal axis by Γ1φr1 gives D1o = 46. Histories of roof displacement and top story drifts for the first “mode” are computed and presented in Fig. The peak values are computed and are summarized in Tables 4.7. 7. The results were generated for first three “modes’ and are included in Fig.8.4. 8.

2 cm. V = 7616 kN.6 cm. “Modal” pushover curves for the example building 77 .9 cm. α = 0.3.(a) "Mode" 1 Pushover Curve 12000 10000 u = 36. V = 5210 kN.eps 0 0 5 10 15 20 Roof Displacement (cm) (c) "Mode" 3 Pushover Curve 25 8000 u = 4.19 y by Base Shear (kN) 8000 6000 4000 2000 0 0 Actual Idealized figA_3a. α = 0. A.14 y by 6000 Base Shear (kN) 4000 2000 Actual Idealized figA_3c.13 6000 Base Shear (kN) 4000 2000 Actual Idealized figA_3b. V y by = 4952 kN.eps 0 0 2 4 6 8 10 Roof Displacement (cm) 12 Fig. α = 0.eps 10 20 30 40 50 Roof Displacement (cm) (b) "Mode" 2 Pushover Curve 60 8000 u = 9.

” and third “mode” inelastic SDF systems. A.223 −80 80 n=3 −3 3 n=3 D3 (cm) 0 A3 (g) • 10.4.33 A1 (g) n=2 0 −80 80 −3 3 n=2 D2 (cm) 0 A1 (g) • 22.(a) Deformation 80 n=1 3 (b) Pseuso−acceleration n=1 • 0.06 0 • 1.m −80 0 −3 0 25 30 Fig.5 × El Centro ground motion for the first “mode. Histories of deformation and pseudo-acceleration due to 1.” second “mode.747 5 10 15 20 Time (sec) figa_4.52 5 10 15 20 Time (sec) 25 30 0 • 1. 78 .2769 D1 (cm) 0 • 35.

237 0.4927 1.017 0.85 36.3 7658.693 0.32 36.3666 3740189 203.048 0.75 21.9 4570.78 21.75 21.170 0.309 0.59 22.4 7714.3 7628.76 21.182 0.184 0.44 36.2 4628.56 47.8525 1.50 36.136 Fsny Ln (cm/sec2) Dny (cm) Fsno Ln (cm/sec ) Dno (cm) Tn (sec) 233.5 3109.6 × Vb1 y i ur1.8 4747.5 7624.18 210.12 3876.194 0.4 7911.948 ζ n (%) 79 .18 210.162 0.194 0.11 22.86 22.46 2.30 37.6 4583.2 4671.191 0.23 0.107 0.90 21.529 0.8 4647.77 21.05 52.151 0.029 0.74 21.910 0.18 210.25 36.81 21.05 36.09 18.9 7615.18 210.70 36.139 0.18 210.18 210.188 0.7 7639.18 210.4 7672.2 7690.18 210.4 7647.194 0.5 (cm) 22.063 0.6 i k1 i ur1 y i α1 i Ab1 Error (%) 1.2.0.1 1013.0 4577.082 0.103 “Mode” 3 696400 0.18 (cm) 38.193 0.18 210.037 0.18 210.18 210.0 4588.79 0.62 26.5309 488839.35 36.1 4574.39 36.65 1226.18 210.38 22.8 7618.9 4573.3 4603.40 46.6 7840.180 0.51 2 “Mode” 2 -920860 -0.1.95 21.09 37.79 21.24 36.23 22.5 4614.25 36.013 0.18 210.18 210.1 4569.022 0.2406 167531.83 21.4 4595.7 4580.8 7622.0 4704.5 7633. Properties of “modal” inelastic SDF systems Properties Ln (kg) Γn * M n (kg) “Mode” 1 2736789 1.18 210.2 4571.010 (kN) 8006.02 21.9 4570.28 36.9 (kN) 4803.3 7745.1 7616.85 0.194 (kN-cm) 365100 364060 363276 362684 362235 361892 361631 361432 361279 361162 361073 361004 360951 360911 360880 360856 360838 360824 360813 Table A.59 36.Table A.18 210.404 0.29 36.190 0.3 7786.198 0.2671 1.64 37.192 0.26 36.18 210.18 210.74 (kN/cm) 210.186 0. No.193 0.56 19.193 0.135 0.86 21. i 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 i Vb1 y i 0.176 0. Results of iterative procedure to develop the idealized bilinear curve for the first “mode” inelastic SDF system Itr.0 7619.

80 .

Appendix B Modal Pushover Analysis B.3b) with unit mass and force-deformation relation of Fig. 81 . Determine the total response by combining the peak “modal” responses using the SRSS combination rule of Eq. rno . 11.1 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. Compute the peak deformation.21).5. of the nth-“mode” inelastic SDF system (Fig.6b by solving Eq. 12. Dn . 10. 4. 13. Following is step-by-step implementation of the procedure described in Section B. Solving Eq.4.8).1 to the north-south component of the El Centro (1940) ground motion scaled up by a factor of 1. extract from the pushover database values of other desired responses. 6. (4. B. (3. Steps 1 to 4 of the MPA are the same as those for UMRHA presented in Appendix A. From the total hinge rotation. or from the inelastic response (or design) spectrum. Typically. (3. 4. the first two or three “modes” will suffice. Calculate the peak roof displacement urno associated with the nth-“mode” inelastic SDF system from Eq. This is the same example as solved in Appendix A.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. At urno . (4.33 cm. 10.1.2 EXAMPLE The MPA procedure is implemented to calculate the response of the 9-story building described in Section 3.18). Repeat Steps 3 to 7 for as many “modes” as required for sufficient accuracy. subtract the yield hinge rotation to determine the plastic hinge rotation.

12.28 cm. Also included in Table 4.4.18) are also included in Tables 4.” and the results are included in Tables 4.3 and 4.28 cm.5 are the plastic hinge rotations computed by subtracting the yield hinge rotation from the total hinge rotations. values of floor displacements and story drifts are extracted.33 = 48.3 for the floor displacements normalized by the building height and in Table 4. (3.366 × 1 × 35.3 and 4. Peak roof displacement ur1o = Γ1 × φr1 × D1 = 1.4. 9. The values are summarized in Table 4. where results for other ground motion intensities are also included. Results of Steps 5 and 6 are summarized in Table B. The total response computed by combining the peak “modal” responses using the SRSS combination rule of Eq.1. At ur1o = 48. 82 . Steps 3 to 7 are repeated for first three “modes.11.4 for the story drifts normalized by the story height. 8.

36 1.275 1.75 µ = Dn Dny urno = Γnφrn Dn (cm) Dn (cm) 0.367 1.50 35.766 7.25 20.05 0.5 µ = Dn Dny urno = Γnφrn Dn (cm) Dn (cm) 2.139 µ = Dn Dny urno = Γnφrn Dn (cm) Dn (cm) 0.690 10.007 36.154 78.79 0.436 7.13 2.735 3.513 0.535 14.577 16.1.450 4.332 48.678 0.Table B.71 1.268 0.73 24.184 0.37 57.52 “Mode” 3 1.07 “Mode” 2 4.35 1.8451 5.551 2.28 46.225 2.126 13.660 14.03 26.200 0.755 0.332 13.06 1.59 0.25 Quantity Dn (cm) “Mode” 1 6.85 µ = Dn Dny urno = Γnφrn Dn (cm) Dn (cm) 1.4222 3.312 1.856 31.755 27.37 1.38 22.52 0.395 0.252 9.35 0.18 27.03 0.379 21.023 0.117 5.70 0.691 0.913 22.504 18.676 6.33 1.901 8.0 µ = Dn Dny urno = Γnφrn Dn (cm) Dn (cm) 3.969 0. Calculation of roof displacements urno from peak deformation of inelastic SDF systems Ground Motion Multiplier 0.457 12.185 11.5 µ = Dn Dny urno = Γnφrn Dn (cm) Dn (cm) 0.27 0.267 5.748 63.229 8.82 1.0 µ = Dn Dny urno = Γnφrn Dn (cm) Dn (cm) 1.467 14.0 µ = Dn Dny urno = Γnφrn Dn (cm) 83 .

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

045 0.065 0.2 95.9 832.7 1578.3 240.5 -973.6 366.2 277. 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) 1917.0 1476.0 176. Table C. the lateral forces are normalized by the base shear to obtain column 10 in Table C.7 234.047 0.3.1 -967.0 381.5 215.0 354.7 95.3 -646.367 87 .7 2065.2 200.070 0.2 105.7 374.4 400.177 0.1 -646.3 222.7 525.7 -46.2 148.2 285.6 319.6 V j3 (kN) (7) 478.6 1233.1 Story Shears V j2 (kN) (6) 1114.7 101.9 -153.6 286.4 250.8 374.2 97.0 1231.4 1842.2 fj = ∑ fi i fj (2) (3) (4) 59.042 0.7 Vj (kN) (8) 2268.8 430.1 -438.9 1446.098 0.0 980.8 1381.3 -6.7 1622.7 694.5 -320.1 87.7 355.0 136.5 159.6 -359.9 1683.5 320.For convenience.4 1759.1 1857.1 -525.6 -352.7 (10) 0.6 -732.0 -5.9 366.9 -166.1 832.9 880.2 Lateral Force fj (kN) (9) 203.3.9 446.090 0.8 -326.5 -350.

Sign up to vote on this title
UsefulNot useful

Master Your Semester with Scribd & The New York Times

Special offer: Get 4 months of Scribd and The New York Times for just $1.87 per week!

Master Your Semester with a Special Offer from Scribd & The New York Times