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

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

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

Pacific Earthquake Engineering Research Center

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

Theory and Preliminary Evaluation

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

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

PEER 2001/03

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

Anil K. Chopra University of California Berkeley

Rakesh K. Goel California Polytechnic State University San Luis Obispo

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

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

ii .

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

iv .

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

vi .

.................................................4....................................................................................6 Elastic Multistory Buildings ................................................1 Summary....................................................36 Modal Pushover Analysis ................4...........................................................................................................2 Response History Analysis .......................................9 3.......................2 Response History Analysis ................................................27 Uncoupled Modal Response History Analysis ...........................34 4.........4 5..1 Underlying Assumptions and Accuracy .........................................................................2...............................................................................................................55 5.. Introduction............................ Equation of Motion .................................................................... Comparison of Modal and FEMA Pushover Analyses................................................4.................................................................................2 Modal Pushover Analysis .............................................................37 Comparative Evaluation of Analysis Procedures .................................................3 4.3 3..........................................................................3 Modal Pushover Analysis ..........................1 2......................................................................................................................47 4..13 3....................................... 2................... vii 1..............................................................1 Uncoupled Modal Response History Analysis ..........................................4 3................41 4...................28 4.....................................4............3 Modal Pushover Analysis with Gravity Loads ......................................................................................................................v Table of Contents.........3....................................................4......................3 Summary................................2................................13 3.................27 4.......................................9 Modal Response Spectrum Analysis ....................22 4..............................12 Comparative Evaluation of Analysis Procedures .. Inelastic Multistory Buildings .1 System and Excitation Considered ..........................................................3 2..........................12 Modal Pushover Analysis ..................................................4.......2....4 Modal Response History Analysis..................................................5 Response History Analysis ......33 4.............................55 vii ...........................................................................................38 4....................... iii Acknowledgment .................................CONTENTS Abstract ......................2 Properties of nth-“mode” Inelastic SDF System ......................................1 3.....................................................................................1 One-Story Systems .................................................2 3.................55 Comparative Evaluation.......................................1 5...................................................17 3.....................................................................3 System and Excitation Considered .........................................36 4.......3 2........6 Pushover Analysis ..........1 4.................................................2 2........................38 4.............................................2 FEMA-273 Pushover Analysis ..

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

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

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

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

sign u = . 2.1) is divided by m to obtain ~ u + 2zw n u + w 2 u y f s u. This becomes evident if Eq.1.fS m fs c ut u fy 1 αk k 1 uy ug -fy k k um u 1 1 (a) (b) Fig.3). and R y (Chopra . (a) Idealized one-story structure. z .e. bg 4 .5) For a given u g t .. 2001. The peak. or absolute (without regard to algebraic sign) maximum. z y . and z is the damping ratio of the system vibrating within its linear elastic range (i. Tn = 2p w n is the natural vibration period. u £ u y ). the deformation u t depends on three systems parameters: w n .ug t n where bg bg b g bg (2. and (b) bilinear hysteretic forcedeformation relation For a given excitation u g t . (2. Section 7.3) ωn = k m ζ = c 2mω n f fs = s fy (2. and the ductility factor is u m= m uy (2. deformation is denoted by um . in addition to the form of the force-deformation relation.4) and w n is the natural vibration frequency. and u y . m depends on three system parameters: w n .

1. f o w = 184 .2311 . m h = 3.9 kN (38.2: the dimensions and flexural rigidity of the structural elements are noted. this system and excitation.26 kN (8. α = 0.3. 2001) and scaled up by a factor of 2. is f y / w = ( f o / w ) ÷ 8 = 0.2 kips) .4 in Chopra. The yield strength of the inelastic system. 2.826 kips) for w = 169.26 kN. 2. For .376 cm.32 m • Fig. Vb (kN) 30 20 10 0 0 2 4 6 Roof Displacement.66 m • EIb EIc EIc • • L = 7.04 40 Base Shear. u (cm) 8 Fig. Vby = 39.2 5 . 2. based on Ry = 8 . 2. 6.2. one-bay frame 50 uy = 1.5 sec and z = 5% subjected to the north-south component of the El Centro (1940) ground motion (Fig. One-story. with Tn = 0.2 SYSTEM AND EXCITATION CONSIDERED Consider the one-story system in Fig. Pushover curve for structure shown in Fig.2. and f y = 39.

Observe that the pushover curve matches the initial loading path of the hysteretic system. (b) shows the lateral force f s t or base shear Vb t normalized relative to the bg weight. (c) shows the joint rotation q bt g . 2.4 shows the earthquake response of the system described in the preceding section determined by response history analysis (RHA). (d) shows the rotation q p bt g of the plastic hinges at the bg bg beam ends. 2. 2.017 rad .1 kip-in. 2. and q pm = 0. and (e) shows the force-deformation relation.3 RESPONSE HISTORY ANALYSIS Figure 2.3 in. e.4 ).The yield moments in the beam and columns are defined as the bending moments due to the lateral force f y .36 cm.077 × 107 mm 4 (146 in. pushover analysis cannot provide any cumulative measure of response. Determined from pushover analysis at the exact peak deformation. I b = 3. The peak values of the various response quantities are as follows: um = 7. respectively.0217 rad . wherein the hysteretic force-deformation history of Fig.4f.) for the beam and columns.134 × 107 mm 4 (75.35. as apparent in Fig.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.) and 50. and E = 2 × 108 kPa (29×103 ksi) gives yield moments of 21.65 kN-m (191.36 cm .18 kN-m (444. the peak value determined from RHA.3.. 2. The system is excited well beyond the yield deformation. the ductility factor m = 5.4e. the energy dissipated in 6 . It is organized in five parts: (a) shows the deformation u t . q m = 0. The yielding stiffness of each structural element is defined as 3% of its initial stiffness. A static pushover analysis of this one-story system leads to the force-displacement relationship shown in Fig. the joint rotation and beam hinge rotation are identical to values q m and q pm determined from RHA.4edetermined from RHAis superimposed. The resulting pushover curve is shown in Fig. However.4 ).g. This pushover curve turns out to be bilinear because the beam and the columns are designed to yield simultaneously when f s reaches f y . Implementing this analysis with I c = 6.6 kip-in. 2.

2 Vb / w 0.5 fy / w = 0.2 −0.4 0.4 −10 −5 0 u (cm) 5 10 Fig.yielding during the ground motion.0217 • θ (rad) 0 (c) −0.4 0. 15 u (cm) 0 • 7.04 0.2311 (a) Vb / w 0 f / w = −0.4.04 0 5 10 15 Time (sec) 20 25 30 0.017 • 0 (d) θp (rad) −0.2 −0. or the cumulative rotation at a plastic hinge.04 0.2 0 (e) 0 (f) −0. (b) base shear.04 0. 2.4 −10 −5 0 u (cm) 5 10 −0.36 −15 0. and (f) pushover curve 7 .5 0. (e) force-deformation relation. (c) joint rotation.2311 y (b) −0. This represents an inherent limitation of pushover analyses. (d) plastic hinge rotation. Response of one-story system to El Centro ground motion: (a) deformation.

8 .

4) 9 .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 . Elastic Multistory Buildings 3. (3.mι ug t bg bg bg (3. m.3) where φ n is the nth natural vibration mode of the structure. and p eff t = Â peff.1) can be interpreted as effective earthquake forces: p eff t = . The right side of Eq.3.snug t n= n =1 The effective earthquake forces can then be expressed as af N af N af (3. c.n t = Â . 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.1) where u is the vector of N lateral floor displacements relative to the ground.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 = . and lateral stiffness matrices of the system. classical damping.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. and k are the mass.

p eff .8) where the modal coordinate qn t is governed by 2 qn + 2z nw n qn + w n qn = . (3.10) bg bg bg 10 (3.9) to the equation of motion for the nth-mode elastic SDF system. (3.12) .sn ug t bg bg bg bg (3.Gn ug t (3. Then the floor displacements are u n t = φ n qn t bg bg bg bg (3. subjected to u g t : Dn + 2z nw n Dn + w 2 Dn = .n t is entirely in the nth-mode. an SDF system with vibration propertiesnatural frequency w n and damping ratio z n of the nth-mode of the MDF system. it can be demonstrated that none of the modes other than the nth mode contribute to the response.n t = .5) The contribution of the nth mode to s and to p eff t are: sn = Gn mφ n respectively. we will outline that the response of the MDF system to p eff . with no contribution from other modes. The equations governing the response of the system are mu + cu + ku = . The solution qn t can readily be obtained by comparing Eq.7) By utilizing the orthogonality property of modes.9) and (3.11) and substituting in Eq.9) in which w n is the natural vibration frequency and z n is the damping ratio for the nth mode.sn ug t bg (3.u g t n Comparing Eqs.n (t ) = N n =1 ∑ −sn u g (t ) N (3.6) Next.10) gives qn t = Gn Dn t bg bg bg (3.peff (t ) = n =1 ∑ peff . (3.8) gives the floor displacements u n t = Gnφ n Dn t bg (3.

Eqs. Therefore. 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 (3.16) reflect combining the response contributions of all modes. (3. (3. internal element forces.n t . 3.15) rt = bg n =1 Â rn t = bg n =1 Â rnst An bt g N (3.16) This is the classical modal RHA procedure wherein Eq.1. ζn ¨ g(t ) u (a) Static Analysis of Structure (b) Dynamic Analysis of SDF System Fig.13) define the contribution of the nth-mode to the response. 3. the static value of r due to external forces sn .1). (3.13) represent the response of the MDF system to p eff. st The two analyses that lead to rn and An t are shown schematically in Fig. these standard equations have been derived in an unconventional way.12) and (3. Equations (3. and An t = w 2 Dn t n bg (3.15) and (3.9) is the standard modal equation governing qn t . Section 13. Conceptual explanation of modal response history analysis of elastic MDF systems Any response quantity r t story drifts.Forces sn An(t ) st rn ωn. 2001.14) is the pseudo-acceleration response of the nth-mode SDF system (Chopra.can be expressed by st rn t = rn An t bg bg bg bg bg (3. etc. In contrast to the classical 11 bg . However. and Eqs.1.12) and (3.13) st where rn denotes the modal static response.

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.g.3 MODAL PUSHOVER ANALYSIS (3. and Tn = 2p w n is the natural vibration period of the nth-mode of the MDF system. we used the modal expansion of the spatial distribution of the effective earthquake forces. 2001. (3. Alternatively. Chopra.1. This concept provides a rational basis for the modal pushover analysis procedure developed later.19) will provide the same value of rno . Section 13. we note that static analysis of the structure subjected to lateral forces f no = Γ n mφ n An (3.derivation found in textbooks (e. The SRSS rule.8.17) (Chopra.4 and 13. 3. The peak modal responses are combined according to the Square-Root-of-Sum-ofSquares (SRSS) or the Complete Quadratic Combination (CQC) rules. 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 (3.1).18) To develop a pushover analysis procedure consistent with RSA.17) where An is the ordinate A Tn . which is valid for structures with well-separated natural frequencies such as multistory buildings with symmetric plan.3). 2001.16).. 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 3.9) . Sections 12. z n of the pseudo-acceleration response (or design) spectrum for the nth-mode SDF system. (3. this response value can be obtained by static analysis of the structure subjected to lateral forces distributed over the building height according to 12 . the peak nth-mode response as in Eq.(3. In such a response spectrum analysis (RSA).

Obviously Dn and An are available from the response (or design) spectrum. 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.2. The bays are 9. The building has a basement level.73 m (150 ft) in plan. CA 90017.2).. Consulting Structural Engineers. Richmond. in both directions. 46th Street. 3. (3. with five bays each in the north-south (N-S) and east-west (E-W) directions.21) 2 where Dn = An ω n . Third St. can be combined according to Eq. The levels of the 9-story building are numbered with respect to the ground level (see Fig. 3.20) and the structure is pushed to the roof displacement. 3. this structure meets seismic code and represents typical medium-rise buildings designed for the Los Angeles. shown in Fig. SAC is a joint venture of three non-profit organizations: The Structural Engineers Association of California (SEAOC). California.2). Typical floor-to-floor heights (for analysis purposes measured from center-of-beam to center1 2 Brandow & Johnston Associates. CA 94804-4698. was designed by Brandow & Johnston Associates1 for the SAC2 Phase II Steel Project. and California Universities for Research in Earthquake Engineering (CUREE).s* = mφ n n (3. the peak value of the roof displacement due to the nth-mode. The interior bays of the structure contain simple framing with composite floors. region.73 m (150 ft) by 45. rno . (3. with the ninth level being the roof. and 37. A benchmark structure for the SAC project.15 m (30 ft) on center. the Applied Technology Council (ATC). each determined by one pushover analysis. which from Eq.19 m (122 ft) in elevation.4 COMPARATIVE EVALUATION OF ANALYSIS PROCEDURES 3. 1660 W. The columns are 345 MPa (50 ksi) steel wideflange sections.12) is urno = Gnf rn Dn (3. This modal pushover analysis (MPA) for linearly elastic systems is equivalent to the well-known RSA procedure (Section 3. 1301 S.1 System and Excitation Considered The 9-story structure.18) to obtain an estimate of the peak value ro of the total response.4. SAC Steel Project Technical Office. The peak modal responses. 13 . this building is 45. Los Angeles. urno . Although not actually constructed. denoted B1.

The first three vibration modes and periods of the building for linearly elastic vibration are shown in Fig. 3. The building is modeled in DRAIN-2DX (Allahabadi and Powell. if desired more complex models. The column lines employ two-tier construction. This model is based on centerline dimensions of the bare frame in which beams and columns extend from centerline to centerline. ceiling/flooring. for the second through eighth levels is 9. 0. the vibration periods are 2. for the first three modes are shown in Fig. The simple model adopted here is sufficient for the objectives of this study. The floor-to-floor height of the basement level is 3. The seismic mass of the above ground levels of the entire structure is 9.of-beam) are 3.89×105 kg (67.49 sec.27.e.. The force distributions. mechanical/electrical. The seismic mass of the ground level is 9. and shear distortion of panel zones are neglected but large deformation (P∆) effects are included. dimension. including the steel framing.49 m (18 ft). 3. we select a weak ground motion: the northsouth component of the El Centro (1940) ground motion scaled down by a factor of 4. which are seismic (tension) splices to carry bending and uplift forces. roofing and a penthouse located on the roof. fifth.3. Column splices. monolithic column pieces are connected every two levels beginning with the first level. 3.7 kips-sec2/ft). third. respectively.0 kips-sec2/ft).sec2/ft). The floor system is composed of 248 MPa (36 ksi) steel wide-flange beams in acting composite action with the floor slab.2 kips-sec2/ft). Concrete foundation walls and surrounding soil are assumed to restrain the structure at the ground level from horizontal displacement. i. and seventh levels at 1.20).83 m (6 ft) above the center-line of the beam to column joint.85. To ensure that this structure remains elastic.65×105 kg (66.07×106 kg (73. and for the ninth level is 1.96 m (13 ft). The 9-story N-S MRF is depicted in Fig. s* (Eq. floor slabs. partitions. such as those described in Gupta and Krawinkler (1999) can be used. 3. The seismic mass of the structure is due to various components of the structure. The column bases are modeled as pinned and secured to the ground (at the B-1 level). for the first level is 1.00×106 kg (616 kips. The strength. are located on the first.2.0 kips-sec2/ft). Each frame resists one half of the seismic mass associated with the entire structure. 1988) using the M1 model developed by Krawinkler and Gupta (1998).65 m (12 ft) and for the first floor is 5. and 0.01×106 kg (69. These force n distributions will be used in the pushover analysis to be presented later. 14 .4.

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

44 1. (3.33 2.728 2. respectively. 2.5 0 0.796 0.38 0. (c) shows the joint rotation q n t of an external joint at 16 bg bg bg .3.5 −1 −0.8 −2.05 1.94 2.27 sec Ground −1.5.51 0. 3.05 s * 1 s * 2 s * 3 Fig.49 sec 3 T = 0. First three natural-vibration periods and modes of the 9story building 3. 3.13 −1.1 3. determined by RHA [Eqs.93 −1.4.72 −2.12 0.2 Response History Analysis The structural response due to individual vibration modes. (b) shows the base shear Vbn t normalized relative to the weight W of the building.4. n = 1.67 −1.12) and (3. and 3.03 −1.5 Fig.0272 −2.61 2.7.6.85 sec 2 Floor 5th 4th 3rd 2nd 1st T1 = 2. and 3 .1 −2.37 2. and 3 n 3.487 −1. is shown in Figs. Each figure is organized in four parts: (a) shows the roof displacement urn t . 3.13)]. n = 1. Force distributions s* = mφn .9th 8th 7th 6th T = 0. 3.05 2.39 3. 2.75 1.04 1.5 Mode Shape Component 1 1.31 −0.

with the first three modes provide essentially the exact response.1.422 cm. The peak values of floor displacements and story drifts determined by RHA. Combining the modal response histories for all modes gives the total response [Eqs. the results for the roof displacement and top-story drift are shown in Fig. including one. The errors due to truncating the contributions of vibration modes beyond the third mode are negligible. respectively.1. and (d) shows the Vbn .15) and (3. but three modes—perhaps even two modes—are sufficient.12 cm. drifts in all stories. 3. in particular. and ur 3o = 0. errors are smallest in floor displacements. which are listed in Tables 3. and rotations of external joints with moment connections are presented in Tables 3.urn relation. It is apparent that the first mode alone is inadequate. Also included are the combined response due to one. and 3. (3. two. errors generally decrease as response contributions of more modes are included. larger in story drifts.23 cm. The peak values of the various response quantities are noted in these figures. or all modes.2.9. The peak values of displacements of all floors. and even larger in joint rotations.the roof level. 3. the peak roof displacement due to each of three modes is ur1o = 9. The linear relationship between the base shear and roof displacement for each mode implies that the structure did not yield. and 3. respectively.3. 3. two. For a fixed number of modes included.2. This is illustrated in Fig. three. As expected. where the second and third modal responses are a larger percentage of the top story drift compared to roof displacement. 3. the exact response considering all modes. 3.16)]. are presented in Fig.8. and the percentage errors due to truncation of higher modal contributions.3. and three vibration modes. 17 . The same method was used to determine the peak values of many response quantities. consistent with the increasing significance of the higher mode response among these three sets of response quantities. ur 2o = 2. especially in estimating the story drifts.8.

0 -2.13E-04 9.7 2.45E-03 3.72E-03 3.25 × 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 0.055 0.231 -0.11E-03 1.1 Table 3.09E-03 2.060 -0.3 19.282 0.121 0.333 0.173 0.85E-03 3.2 0.6 9.44E-03 3.38E-04 2 Modes 2.192 0.8 -15.7 -19.14 m3) from RHA for 0.024 -0.042 0.088 -0.364 0.009 -0.65E-03 RHA (all modes) 2.28E-04 1.88E-03 2.2 Peak values of story drift ratios (as % of story height) from RHA for 0.76E-03 1.235 0.6 0.156 0.2 -57.002 -0.14E-03 2.295 -0.227 0.229 0.0 -2.73E-03 3.229 0.199 0.76E-03 1.124 0.6 -1.03E-03 1.321 0.045 0.012 0.08E-03 2.205 0.99E-03 2.3 -8.4 -10.8 -10.78E-04 -3.3 -0.152 0.026 0.475 0.130 0.307 0.03E-03 -6.090 0.63E-03 2.90E-03 3.2 -1.1 Building height is measured from the ground floor to the 9th floor.3 -0.012 -0.0 3 Modes -5.237 0.9 8.6 11.6 -0.9 -22.266 0.9 1.7 3.5 -1.080 0.311 0.65E-03 2.56E-03 2.202 0.89E-03 1.9 -16.156 0.2 9.0 -10.062 0.261 -0.9 9.058 -0.406 0.4 -41.91E-04 1.183 0.09E-03 1.089 0.4 -6.42E-04 1.44E-03 1.09E-03 Error (%) 1 Mode -23.015 -0.235 0.01E-04 -2.060 0.336 0.24E-03 2.011 0.125 0.99E-03 Mode 3 3.264 Error (%) 1 Mode 2 Modes 3 Modes -23.74E-04 9.173 0.9 3.042 0.09E-03 1.378 0.7 2 Modes -3.177 0.203 0.1 -19.0 -46.260 0.181 0.325 0.097 0.8 -56.282 0.11E-04 -5.069 0.399 0.300 0.0 7.1 -2.38E-03 2.00E-03 2.1 4.1 -2.413 RHA (all modes) 0.006 0. 18 .407 0.117 0.74E-04 6.22E-03 2.8 -1.263 0.50E-03 2.265 RHA (all modes) 0.4 -53.2 -4.01E-04 3.1 -11.8 1.245 0.124 0.1 -2.3 -3.177 0.237 0.88E-03 2.5 0.370 0.1 Peak values of floor displacements (as % of building height = 37.66E-05 -3.50E-03 4.9 -23.1 -14.202 0.097 0.202 0.25 × El Centro ground motion Floor 1st 2nd 3rd 4th 5th 6th 7th 8th 9th 3 Modal Response Mode 1 2.4 -22.5 18.64E-03 3.010 -0.6 4.001 -0.33E-03 2.7 -50.9 -15.008 -0.060 -0.226 0.245 0.2 -2.022 0.31E-03 Mode 2 -1.038 0.125 0.47E-03 1.74E-03 1.3 -33.152 0.265 0.7 7.400 0.7 4.350 0.1 -0.253 0.29E-03 2.03E-03 3 Modes 2.9 2.03E-03 1.4 -1.303 0.13E-03 2.94E-03 2.4 -1.00E-03 1.258 0.4 0.89E-03 1.8 1.3 Peak values of joint rotations (radians) from RHA for 0.035 0.266 0.2 -20.197 0.00E-03 1.133 -0.259 0.26E-04 -5.227 0.097 0.275 0.8 -5.401 Error (%) 1 Mode 2 Modes 3 Modes -23.5 -2.003 0.6 -1.226 0.310 0.071 0.260 0.159 0.0 -0.74E-03 1.63E-03 2.1 3.15E-03 4.311 0.179 0.157 0.266 0.062 -0.011 0.069 0.253 0.259 0.043 0.032 -0.4 -0.3 Table 3.023 0.6 0.6 1.25 × 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 0.9 -24.Table 3.2 1.453 0.466 0.31E-03 Joint Rotation (rad) Combined (RHA) 1 Mode 2.054 0.4 -3.4 -7.225 0.

12 −0. Response due to first mode: (a) roof displacement.15 u (cm) 9.5.0435 • 0 (b) V b1 −0.1 /W 0.25 × El Centro ground motion 19 .5 0 u (cm) r1 7.1 0. Excitation is 0.12 • 0 (a) r1 −15 0. and (e) pushover curve. 3.5 9.1 0. (d) force-deformation history. (c) joint rotation.12 0 (d) 0 (e) 15 Fig. (b) base shear.0435 Vb1 / W 0.5 0 u (cm) r1 7.0435 9.1 −15 −7.00174 • 0 (c) r1 −5 0 5 10 15 Time (sec) 20 25 30 0.5 15 −0.1 −15 −7.1 5 x 10 −3 θ (rad) 0.

00244 • 0 (c) r2 −5 0 5 10 15 Time (sec) 20 25 30 0.23 −5 (a) 0. 3. (c) joint rotation.23 −2.5 0 u (cm) r2 2.0252 (d) 0 0 (e) −0. and (e) pushover curve.5 5 Fig.05 /W 0.23 0.6. (b) base shear.1 −5 −2.1 0.25 × El Centro ground motion 20 .0252 −2.0252 • 0 (b) V −0. Excitation is 0.5 0 u (cm) r2 2.1 −5 −2.5 u (cm) r2 0 • 2.05 5 x 10 −3 b2 θ (rad) 0.1 Vb2 / W 0. Response due to second mode: (a) roof displacement.5 5 −0. (d) force-deformation history.

000913 • 0 (c) r3 −2 0 5 10 15 Time (sec) 20 25 30 0.5 1 Fig.5 1 −0.1 u (cm) r3 0 • 0.0108 x 10 −3 (b) −0.5 0 u (cm) r2 0.5 0 u (cm) r3 0.0108 −0. Response due to third mode: (a) roof displacement.0108 −0.25 × El Centro ground motion 21 . Excitation is 0. (b) base shear.422 (e) −0. and (e) pushover curve.7.02 0. (c) joint rotation.02 −1 −0.02 −1 −0.02 −0.422 −1 (a) 0.422 /W V b3 0 (d) 0 −0.02 /W V b3 0 • 0.02 2 θ (rad) 0. 3. (d) force-deformation history.

this value would be determined directly from the response (or design) spectrum.6 present estimates of the combined response according to Eq. 3.23 cm . pushing the structure using the force distribution of Eq. 3.3 Modal Pushover Analysis Implementing MPA for the fundamental vibration mode.21). respectively. and the errors in these estimates relative to the exact response from RHA considering all modes.4.2. and 3. respectively. 3.. 3.5e. the errors in the MPA results are generally larger than in RHA (Fig. i. and 3. 3. (3. and it is apparent in the example considered that three modes provide most of the response (Fig. and external joint rotations in Tables 3.3).5d). or three vibration modes. Implementing pushover analysis for the second and third modes.4) to roof displacement ur1o = 9. and the computed response values are identical to the peak response values determined from RHA (Tables 3. In RHA the errors arise only from truncating the responses due to higher modes. leads to the pushover curves shown in Figs. (3. In practical application. and ur 3o = 0.1. 3. 3.1 through 3. these pushover curves are consistent with the Vb .10 and Tables 3.5.20) with n = 2 and 3 up to roof displacements ur 2 o = 2. Figure 3.7e and to the floor displacements.4.ur relations determined by RHA (Figs. 3. respectively. story drifts. and 3.3). These values of the response quantities are identical to the peak response values determined from RHA (Tables 3.20) with n = 1 (Fig. i.6. confirming that MPA gives the exact values of the individual modal responses.4. two.6e and 3.6d and 3.9).422 cm .6. The floor displacements.e.4. implying that the modal truncation errors are small if at 22 . 3.8) leads to the pushover curve shown in Fig. story drifts.2.9 and Tables 3. This pushover curve is consistent with the relationship between the base shear and roof displacement determined by RHA (Fig. Observe that the target roof displacement in each pushover analysis is identical to its exact value determined by RHA. For a fixed number of modes included considered. except for the algebraic sign associated with Γ1 and minor round-off errors.3. considering one. the value determined by RHA (Fig. the floor displacements are proportional to the mode shape φ 1 because the structure remains elastic.12).e. using the force distribution of Eq. although both analyses led to identical peak values of the individual modal responses.5.18). 3.12 cm. 3.1.7d). which would provide the Dn value to be substituted in Eq.3). As for the first mode. and 3.5. and 3. (3. pushing the structure. 3. (3. As suggested by Eq.. 3. (3. and external joint rotations computed by pushover analysis are presented in Tables 3.

4 Story Drift Ratio (%) 0.48 −3 3 1.5 Ground 0 0.685 • −3 3 Mode 2 ∆ (cm) u (cm) • 2.23 r2 −15 15 0 • 0.422 r2 0 • 1. sec 25 30 Fig. Response histories of roof displacement and top-story drift from RHA for 0.8.25 × 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 Ground 0 0.eps fig3_9b.9.2 0.1 0. (a) Roof Displacement 15 (b) Top Story Drift 3 Mode 1 ∆ (cm) u (cm) 9.1 0.6 Fig.2 0.eps r 5 10 15 20 Time. sec 25 30 r 5 10 15 20 Time. 3.5 0.83 • r3 0 • 0.3 0. Additional errors are introduced in pushover analysis due to the approximation inherent in modal combination rules.3 0.12 • r1 −15 15 0 r1 0 0 0.least three modes are included. 3.59 • Total (All Modes) 0 −15 0 ∆ (cm) u (cm) 0 −3 0 fig3_8.4 Displacement/Height (%) 0. Heightwise variation of floor displacements and story drifts from RHA for 0.25 × El Centro ground motion 23 .03 −3 3 Mode 3 ∆ (cm) u (cm) r3 −15 15 9.

133 0.9 -11.006 -0.89E-03 2.9 -15.055 0.230 -0.466 0.173 0.177 0.266 0.73E-03 3.264 Error (%) Modal Response 1 Mode 2 Modes 3 Modes -23.157 0.245 0.125 0.321 0.9 -13.322 0.1 -18.0 -10.63E-03 2.012 -0.65E-03 2.245 -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 0.274 0.011 0.38E-03 3 Modes 2.44E-03 3.3 1.045 -0.179 0.259 0.285 0.261 -0.8 -14.197 0.0 -0.4 -4.179 0.40E-04 5.03E-03 -1.0 -46.227 0.76E-03 1.156 0.310 0.42E-04 -1.026 0.05E-03 3.33E-04 5.4 -53.22E-03 -2.31E-03 2.3 -9.63E-03 -2.7 -15.088 -0.09E-03 1.0 -16.012 0.181 0.3 Table 3.3 -41.203 0.00E-03 1.060 -0.078 0.157 0.9 -18.5 -16.89E-03 1.80E-04 3.260 -0.8 -22.002 0.036 -0.74E-03 1.28E-03 2.6 -17.276 0.270 0.3 -2.31E-03 Joint Rotation (rad) Combined (MPA) 1 Mode 2.22E-04 2.124 0.407 0.3 -12.15E-03 2.5 Peak values of story drift ratios (as % of story height) from MPA for 0.4 -22.57E-03 1 Mode -23.253 0.270 0.106 0.9 -15.2 Table 3.09E-04 -3.1 -11.267 -0.023 -0.038 -0.1 -0.4 -22.069 0.31E-03 Error (%) RHA (all modes) 2.4 0.374 0.90E-03 1.4 -9.022 0.260 0.3 -33.235 0.008 -0.015 -0.313 0.133 -0.2 -20.24E-03 -2.90E-04 -9.7 -21.237 0.133 0.270 0.0 -18.125 0.173 0.253 RHA (All Modes) 0.229 0.062 0.2 -4.048 0.99E-03 Mode 3 -3.03E-03 6.282 0.8 -15.92E-04 -1.9 3 Modes -12.9 -13.9 -14.9 -15.300 0.2 -57.7 -19.237 -0.72E-03 3.08E-03 2.00E-03 2.272 0.7 -50.332 0.058 -0.009 0.00E-03 2.6 -15.7 2.062 -0.6 -19.97E-03 1.010 0.042 0.76E-03 -1.3 -13.00E-03 -1.43E-04 -1.3 -19.203 0.9 -13.331 0.259 -0.267 0.069 0.043 -0.048 0.336 0.042 0.09E-03 -1.4 -19.94E-03 2.9 -8.9 -12.203 0.4 -2.385 0.152 0.060 0.121 0.4 -4.032 0.253 0.080 0.38E-03 2.253 -0.286 0.4 -9.2 -11.7 24 .1 -19.96E-03 2.071 0.203 0.73E-05 3.03E-03 1.079 0.0 -2.4 -7.6 Peak values of joint rotation (radians) from MPA for 0.203 0.00E-03 2.296 -0.2 -16.267 0.370 0.024 0.7 2 Modes -13.226 0.3 -14.89E-03 -1.260 0.04E-03 3.8 -23.89E-03 2.4 -5.9 -24.003 -0.001 0.09E-03 Mode 2 1.6 -16.1 -11.12E-03 1.4 1.2 1.00E-03 3.9 -14.53E-04 -9.74E-03 -1.4 Peak values of floor displacements (as % of building height = 37.106 0.4 -14.328 0.4 -11.78E-04 2 Modes 2.9 -15.89E-03 -2.011 -0.097 0.117 0.2 -0.8 -56.097 0.229 0.2 -12.090 0.9 -16.44E-03 -1.8 -15.336 RHA (all modes) 0.9 -14.156 0.231 -0.152 0.401 Error (%) 1 Mode 2 Modes 3 Modes -23.74E-04 -6.230 0.9 -13.14 m) from MPA for 0.177 0.Table 3.282 -0.227 0.235 -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 -0.15E-03 1.25 × El Centro ground motion Floor 1st 2nd 3rd 4th 5th 6th 7th 8th 9th Modal Response Mode 1 -2.47E-03 1.03E-03 3.

4 Story Drift Ratio (%) 0.2 0.3 0.5 0. Heightwise variation of floor displacements and story drifts from MPA for 0. 3.1 0.4 Displacement/Height (%) 0.eps fig3_10b.5 Ground 0 0.10.6 Fig.2 0.1 0.(a) Floor Displacements 9th 8th 7th 6th 9th 8th 7th 6th (b) Story Drift Ratios Floor 4th 3rd 2nd 1st Floor 5th RHA All Modes MPA 1 Mode 2 Modes 3 Modes 5th 4th 3rd 2nd 1st RHA All Modes MPA 1 Mode 2 Modes 3 Modes fig3_10a.3 0.25 × El Centro ground motion. shading indicates modal combination error 25 .eps Ground 0 0.

26 .

sign u b g g bg (4. 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). Both systems have the same mass and damping. and the unloading and reloading curves differ from the initial loading branch. it is useful for later reference to transform Eq.2) to the modal coordinates of the corresponding linear system.mι ug t b (4. Although classical modal analysis (Section 3. (4. Eq. Thus. 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. the relations between lateral forces f s at the N floor levels and the lateral displacements u are not single valued. the initial loading curve is idealized as bilinear.1) becomes mu + cu + f s u.4 4. leading to the “exact” nonlinear response history analysis (RHA).1) With this generalization for inelastic systems. but depend on the history of the displacements: f s = f s u. Therefore. sign 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 (4.1) is not valid for inelastic systems. (3.1 Inelastic Multistory Buildings RESPONSE HISTORY ANALYSIS For each structural element of a building.2) The standard approach is to solve directly these coupled equations.3) .

4) where the only term that differs from Eq. (4.… N (4. (4. premultiplying by φ T . sign u n b g b g bg (4. Simultaneously solving these coupled equations and using Eq.Substituting Eq.and classical n damping-orthogonality property of modes gives F qn + 2z nw n qn + sn = .3). qr t = 0 for all 28 bg . (3. Unlike Eq.2). these equations are coupled for inelastic systems.9) involves T Fsn = Fsn qn . For linear systems. where φ n are now the modes of the corresponding linear system. in principle. sign u = . give the same results for u t as obtained directly from Eq.6b) are mu + cu + f s u. (4.” implying that other “modes” will also contribute to the solution. (3. Equation (4.8) because qr t will generally be nonzero for “modes” other than the nth “mode. (3.sn ug t bg b g bg bg (4. The equations governing the response of the inelastic system to p eff. Eq.2).4)] leads to the uncoupled modal response history analysis (UMRHA) procedure. 2. However.n t given by Eq. (3.3) in Eq. (4. implying coupling of modal coordinates because of yielding of the structure.6) The solution of Eq. (4. The spatial distribution s of the effective earthquake forces is expanded into the modal contributions sn according to Eq. (3.Gn ug t Mn bg n = 1. and using the mass. however. (4. This approximate RHA procedure is the preliminary step in developing a modal pushover analysis procedure for inelastic systems. (4.6) for inelastic systems will no longer be described by Eq. sign qn = φ n f s u n . 4. (4.4) is rarely solved because it offers no particular advantage over Eq.4) represents N equations in the modal coordinates qn .2).9) for linearly elastic systems.2 UNCOUPLED MODAL RESPONSE HISTORY ANALYSIS bg Neglecting the coupling of the N equations in modal coordinates [Eq.3) will.5) This resisting force depends on all modal coordinates qn t .

(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

51 • ur2 (cm) 0 ur2 (cm) n=3 14. and (b) approximate solution by UMRHA bg af bg 32 . n = 1.9 • 0 −50 10 −50 10 n=3 ur3 (cm) 0 • 5.n t = . Fsn / Ln ¨ g(t ) u (a) Static Analysis of Structure (b) Dynamic Analysis of Inelastic SDF System Fig. 2.89 50 r1 0 u (cm) 0 • 78. ζn. and 3. 4. where u g t = 3.s n u g t . 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) • 75.4. Roof displacement due to p eff. 4.575 5 10 15 20 Time (sec) 25 30 ur3 (cm) 0 • 5.112 5 10 15 20 Time (sec) 25 30 −10 0 −10 0 Fig.Forces sn Unit mass An(t ) st rn ωn.3.0 ¥ El Centro ground motion: (a) exact solution by NL-RHA.02 −150 n=2 −150 50 n=2 14.

4. (2) the superposition of responses to p eff. both for 3. this intense excitation was chosen to ensure that the structure is excited well beyond its linear elastic limit.(a) Nonlinear RHA 20 n=1 20 (b) UMRHA n=1 • 5.817 5 10 15 20 Time (sec) 25 30 −10 0 −10 0 Fig.4 and 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. and (b) approximate solution by UMRHA af bg bg 4. and 3.33 0 −20 20 −20 20 n=2 ∆r2 (cm) 0 ∆r2 (cm) 7. Top story drift due to p eff.s n u g t .956 5 10 15 20 Time (sec) 25 30 ∆r3 (cm) 0 • 5. where u g t = 3.2… N ) according to Eq. The errors are slightly larger in drift than in displacement.n t = . n = 1. the errors in either response quantity are only a few percent.008 • 6. respectively. (3.2. but even for this very intense excitation.0 ¥ El Centro ground motion.744 • 0 −20 10 n=3 −20 10 n=3 ∆r3 (cm) 0 • 5. 4.0 ¥ El Centro ground motion: (a) exact solution by NL-RHA.6) by UMRHA is compared with the “exact” solution by nonlinear RHA.5. (4.1 Underlying Assumptions and Accuracy The approximate solution of Eq.n t (n = 1.855 ∆r1 (cm) 0 ∆r1 (cm) n=2 • 6. 2. and (3) the Fsn Ln .5)] is neglected.15) is strictly valid only for linearly elastic systems.5. (4.Dn relation is 33 bg bg .4) and (4. Such comparison for roof displacement and top-story drift is presented in Figs.

when specialized for linearly elastic systems it is identical to the RHA procedure of Section 3. 4. this distribution seems to be a rational choiceeven for inelastic systemsto determine Fsn in Eq. 4. At the yield point. (4.Dn relation to be determined in Eq. (4.8) governing Dn t is based on Eq. 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. When implemented by commercially available software. Although several approximations are inherent in this UMRHA procedure. is plotted against roof displacement urn . within the linearly elastic range of the structure.approximated by a bilinear curve to facilitate solution of Eq.9) should be determined by nonlinear static analysis of the structure as the structure undergoes displacements u = Dnφ n with increasing Dn .2. (4.Dn relation? The two sets of forces and displacements are related as follows: V Fsn = bn Γn Dn = urn Γ nf rn 34 (4. (3.10) . and the base shear Vbn . which is different than the Fsn Ln . How to convert this Vbn .8) before it can be solved? Because Eq. Therefore we impose this constraint in developing the UMRHA procedure in this section and modal pushover analysis in the next section.1.20) to some bg pre-determined roof displacement.20). Although most commercially available software cannot implement such displacement-controlled analysis.Dn curve. such nonlinear static analysis provides the so-called pushover curve.2 Properties of the nth-“mode” Inelastic SDF System How is the Fsn Ln .8) in UMRHA.urn pushover curve to the Fsn Ln . (4.6a. However. Therefore. (4.4. A bilinear idealization of this pushover curve for the nth-“mode” is shown in Fig. (3. The overall errors in the UMRHA procedure are documented in the examples presented in Section 4.9). the only force distribution that produces displacements proportional to f n is given by Eq. it can conduct a force-controlled nonlinear static analysis with an invariant distribution of lateral forces.12) for floor displacements. The structure is pushed using the force distribution of Eq. the relationship between lateral forces f s and Dn in Eq. (3. the base shear is Vbny and roof displacement is urny .

4. Section 13.12) implying that the initial slope of the curve in Fig. 4.6b.6b is w 2 . 35 . 4.6a is 2 kn = ω n Ln . which may differ from the period of the corresponding linear system. The two are related through Fsny Ln 2 = w n Dny (4. which is not a meaningful quantity. In contrast. 2001.11). where the yield values of Fsn Ln and Dn are Fsny Ln = Vbny * Mn Dny = urny Γ nf rn (4.13) This value of Tn .8). Knowing Fsny Ln and Dny from n Eq. the initial slope of the pushover curve in Fig.5).10 enables conversion of the pushover curve to the desired Fsn Ln − Dn relation shown in Fig. Properties of the nth-“mode” inelastic SDF system from the pushover curve Equation 4. the elastic vibration period Tn of the nth-mode SDF system is computed from F Ln Dny I 1/ 2 Tn = 2p G H Fsny JK (4.6. (4. (4.2. should be used in Eq. 4.V bn (a) Idealized Pushover Curve F sn /L n (b) F sn / L − D Relationship n n Idealized Vbny Actual 1 αnkn V bny /M * n 1 α ω2 n n kn 1 1 ωn 2 u u rny rn Dny = ur n y / Γn φ r n D n Fig.11) * in which M n = Ln Γn is the effective modal mass (Chopra.

with the structure is pushed to the roof displacement urno .6b) by utilizing Eq. develop the base-shear – roof-displacement ( Vbn − urn ) pushover curve for the force distribution s* [Eq. and modes. r o . φn .21) where Dn . (3. Convert the idealized pushover curve to the Fsn Ln − Dn relation (Fig. 4.15) and (3.3 Summary The inelastic response of an N-story building with plan symmetric about two orthogonal axes to earthquake ground motion along an axis of symmetry can be estimated as a function of time by the UMRHA procedure just developed. of the nth-“mode” inelastic SDF system (Fig.16) to determine the total response.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 . 9.20)]. Calculate histories of various responses by Eqs. n (3. ωn . 4. 2. the peak value of Dn t . the first two or three modes will suffice. n 3. (3. 8. This value of the roof displacement is given by Eq. and pseudo-acceleration history. An t .6a). 7. Idealize the pushover curve as a bilinear curve (Fig. Typically. 4. For the nth-mode.3b) with force-deformation relation of Fig.20)]. Consider a nonlinear static analysis of the structure subjected to lateral forces distributed over the building height according to s* [Eq. 4. for linearly-elastic vibration of the building. details are available in Appendix A: 1. Combine the “modal” responses using Eqs. which is summarized next as a sequence of steps. 6.6b. 5. (3. Dn (t ) .11). (3. 4. (4.2.4. Calculate the peak value. Compute the deformation history.13). Compute the natural frequencies. 4. of the total response r t obtained in Step 8. Repeat Steps 2 to 6 for as many modes as required for sufficient accuracy. is now determined by 36 bg bg .12) and (3.n t .

4. as described in Section 4.8).6 and 7.6b by solving Eq. rno also represents the exact peak value of the nth-mode contribution rn t to response r t .3. governed by Eq. 4.g. 2001.3. At this roof displacement.2. we will refer to rno as the peak “modal” response even in the case of inelastic systems. 37 . It provides results for elastic buildings that are identical to the well-known RSA procedure (Section 3. 6. (4. (3. e. the pushover analysis provides an estimate of the peak value rno of any response rn t : floor displacements. Compute the peak deformation. each determined by one pushover analysis.2. as mentioned earlier. which is summarized next as a sequence of steps. it seems reasonable because it provides results for elastic buildings that are identical to the well-known RSA procedure described in Section 3. joint rotations.n t . for elastic systems. of the nth-“mode” inelastic SDF system (Fig. etc. or from the inelastic response (or design) spectrum.8). to obtain an estimate of the peak value ro of the total response. This application of modal combination rules to inelastic systems obviously lacks a theoretical basis.. Eq. and (2) it provides the exact modal response for elastic systems.3b) with force-deformation relation of Fig. plastic hinge rotations. 4. bg bg bg bg Thus. As shown in Sections 3. The response value rno is an estimate of the peak value of the response of the inelastic system to p eff.6). story drifts. (4. 5. Dn . (3. alternatively. Steps 1 to 4 of the MPA are same as those for UMRHA. Sections 7. (4. are combined using an appropriate modal combination rule.18). seems reasonable. This pushover analysis.21). 4.12). The peak “modal” responses rno . it can be determined from the inelastic response (or design) spectrum (Chopra.3) because.1 Summary The peak inelastic response of a building to earthquake excitation can be estimated by the MPA procedure just developed. details are available in Appendix B.solving Eq.2 and 3. the lateral force distribution used possesses two properties: (1) it appears to be the most rational choice among all invariant distribution of forces. However. although somewhat intuitive for inelastic buildings. Calculate the peak roof displacement urno associated with the nth-“mode” inelastic SDF system from Eq.

and compared with the results of a rigorous nonlinear RHA using the DRAIN-2DX computer program. the peak roof displacement due to each of the three “modes” is ur1o = 48. and the errors in the approximate results are shown in Fig. 4. rno . The peak values of response are as noted.7.53 cm. The peak values of floor displacements and story drifts including one. the “exact” results. (3.2. subtract the yield value of hinge rotation to determine the hinge plastic rotation. and three “modes”.8. the 38 .4.1 Uncoupled Modal Response History Analysis The structural response to 1. From the total rotation of a plastic hinge. 4. ur 2o = 11. also included are the combined responses due to one. the combined response due to three “modes”. extract from the pushover database values of other desired responses.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. two.1 and 4. two. the first two or three “modes” will suffice.18). Observe that errors tend to decrease as response contributions of more “modes” are included.7 shows the individual “modal” responses.0. Determine the total response by combining the peak “modal” responses using the SRSS combination rule of Eq. 8. three “modes” included. is presented next. Repeat Steps 3 to 8 for as many “modes” as required for sufficient accuracy. This is to be expected. The peak values of displacements of all floors and drifts in all stories are presented in Tables 4. say.1). determined by the UMRHA procedure. the UMRHA procedure lacks a rigorous theory. Typically. At urno .7 cm. To ensure that this structure responds well into the inelastic range the El Centro ground motion is scaled up by a factor varying from 1.3 cm. respectively. in particular. This deficiency also implies that.4. and the percentage errors in the approximate results.9.5 ¥ the El Centro ground motion including the response contributions associated with three “modal” inelastic SDF systems. 9.2). Figure 4. and three modes are compared with the “exact” values in Fig. in contrast to modal analysis (Section 3. with.0 to 3. although the trends are not systematic as when the system remained elastic (Section 3. 4. and the “exact” response from nonlinear RHA for the roof displacement and top-story drift. and ur 3o = 2. 4.

1. and total response from NL-RHA 39 .44 • "Mode" 2 r1 0 • 48. However.53 ∆r3 (cm) 12 0 • 2. 0.2). 0. we investigate how the errors in the UMRHA vary with the deformation demands imposed by the ground motion. recall that the computed errors have been presented earlier for ground motion multipliers 0. Next. For each excitation.25 (Tables 3.2) and 1.1 and 4. just as for elastic systems.7.0.5. defined as the El Centro ground motion multiplied by 0. the errors in responses computed by UMRHA including three “modes” relative to the “exact” response were determined.7 • −80 80 ∆r1 (cm) u (cm) "Mode" 1 −12 r2 0 ∆r2 (cm) 12 0 u (cm) −80 80 −12 r3 0 • 2.5.24 • NL−RHA r −12 12 ∆r (cm) u (cm) 0 5 10 15 20 Time.0.3 11. in particular. for a fixed number of “modes” included. sec 25 30 r −12 0 Fig.85.1 and 3.response is much less accurate (Tables 4. the errors in story drifts are larger compared to floor displacements.1 −12 12 ∆r (cm) u (cm) UMRHA 3 "Modes" 0 • 7. For this purpose the UMRHA and exact analyses were repeated for ground motions of varying intensity. 1.2) if the system yields significantly versus if the system remains within the elastic range (Tables 3.38 6.0.1 and 4. 4.6 5 10 15 20 Time.75.5 × El Centro ground motion: individual “modal” responses and combined response from UMRHA.2).25. 2. sec 25 30 • 48. Response histories of roof displacement and top-story drift due to 1. the degree to which the system deforms beyond its elastic limit.62 5.5 (Tables 4.1 and 3.88 u (cm) "Mode" 3 −80 80 0 −80 80 0 −80 0 • 44. and 3. 0. (a) Roof Displacement (b) Top Story Drift 80 12 0 • 3.

two. indicated by a ground motion multiplier.5 1 1. Heightwise variation of floor displacements and story drift ratios from UMRHA and NL-RHA for 1. 4.5 Displacement/Height (%) 2 Ground 0 0.(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. it will be useful to know the deformation of the system relative to its yield deformation.5 ¥ El Centro ground motion Figure 4. Shown is the error in each floor displacement (Fig.5 Story Drift Ratio (%) 2 2.10b). Heightwise variation of error in floor displacements and story drifts estimated by UMRHA including one. the pushover curves using force distributions s* [Eq. in each story drift (Fig.5 × El Centro ground motion (a) Floor Displacements 9th 8th 7th 6th 9th 8th 7th 6th (b) Story Drift Ratios Floor 4th 3rd 2nd 1st UMRHA 1 "Mode" 2 "Modes" 3 "Modes" Floor 5th 5th 4th 3rd 2nd 1st UMRHA 1 "Mode" 2 "Modes" 3 "Modes" Ground −60 −40 −20 0 Error (%) 20 40 60 Ground −60 −40 −20 0 Error (%) 20 40 60 Fig.10a). 4. To interpret these results. 4. Two versions of the pushover curve are 40 .11. and the error envelope for each case.8. (3. 4. For this purpose.5 1 1. with the peak displacement of each “modal” SDF system noted for each ground motion multiplier. 4.5 Fig.9. or three “modes” for 1.20)] for the first three n modes of the system are shown in Fig.10 summarizes the error in UMRHA as a function of ground motion intensity.

Second. (4.7).3 cm.5 ¥ the El Centro ground motion. 4. respectively. First.11).75.included: the actual curve and its idealized bilinear version. the values determined by RHA of the nth-mode inelastic SDF system (Fig.53 cm.4) to roof displacements urno = 48. (2) increase rapidly as the ground motion multiplier increases to 1. 4.6 and 4. the errors in truncating the higher mode contributions are negligible. the ductility factor for the first mode system is only 1. and 3 (Fig. UMRHA lacks a rigorous theory and is based on several approximations.3). the first reason mentioned above seems to be the primary source for the errors. 2. (3) maintain roughly similar values for more intense ground motions. the peak values of displacements at all floors (Table 4. 41 . as mentioned in Section 3. determined from “exact” analyses. even though the system remains essentially elastic. were noted but are not shown here. Figure 4.2 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 1.10 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. drifts in all stories (Table 4. Each of these three pushover analyses provides the pushover curve (Fig.11a). and. as mentioned in Section 4. 3.7 cm. (3. The idealized curve for the first “mode” deviates most from the actual curve near the peak displacement corresponding to ground motion multiplier 1. For more intense excitations. Additional errors are introduced in UMRHA of systems responding beyond the linearly elastic limit for at least two reasons.0. The structure is pushed using the force distribution of Eq. 4. The system remains elastic up to ground motion multiplier 0. The location of plastic hinges and their rotations. and plastic hinge rotations at the external beam end at each floor level (Table 4.4). and (4) are larger in story drift compared to floor displacements. 4.01 (Fig.4.10) for each “modal” inelastic SDF system (Figs 4.4.5).20) with n = 1. Perhaps this explains why the errors are large at this excitation intensity.1). and 2.0. the pushover curve for each “mode” is idealized by a bilinear curve in solving Eq.75.2. 11.2.

138 1.554 1.260 -15.3 8.154 0.068 0.410 1.806 -0.852 1.914 2.6 4.5 3 Fig.220 0.526 -0. 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 0.214 0.133 1.136 1.490 -1.5 1 1.772 Error (%) NL 3 1 2 3 RHA “Mode” “Modes” “Modes” “Modes” 1.938 -1.214 -0.315 -0.806 0.298 0.8 1.707 1.9 16.079 0.942 -0.2 6.2 4.0 1.366 -0.2 12.472 1.(a) Floor Displacements 60 50 40 Error Envelope Error for Floor No.698 1.900 -10.0 11.0 9.201 -0.7 31.663 -0.070 1.169 0.945 -37.722 0.575 -41.4 4.088 10.373 -0.9 5.372 -1.3 0.372 1.513 0.201 1.2 22.120 1.877 0.241 1.376 -1.072 -1.3 1.2 1.5 10.126 0.5 18.241 -1.371 -0.298 -0.4 -1.763 -15.226 -0. Errors in UMRHA as a function of ground motion intensity: (a) floor displacements.216 1. 4.0 11.2 1.5 9.7 Table 4.668 -23.484 0.10.727 1.513 -0.130 0.663 0.366 0.220 -0.202 11. Noted 60 50 40 (b) Story Drifts Error Envelope Error for Story No.057 -0.293 1.071 -0.407 -10.5 × 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” -0.072 1.018 0.104 0.256 1.3 25.914 -0.4 -7.2 Peak values of story drift ratios (as % of story height) from UMRHA for 1.5 × El Centro ground motion Story 1st 2nd 3rd 4th 5th 6th 7th 8th 9th Modal Response “Mode” 1 “Mode” 2 Drift Ratio (%) Combined (UMRHA) 1 “Mode” 2 “Modes” “Mode” 3 -1.938 1.096 0.201 -1.135 9.478 0.1 1. and (b) story drifts Table 4.4 1.751 1.5 28.187 -0.14 m) from UMRHA for 1.0 -9.8 1.256 -1.410 -1.540 0.055 -0.8 14.6 1.942 1.1 Peak values of floor displacements (as % of building height = 37.0 7.338 1.490 1.1 8.9 31.3 42 .9 31.844 -25.379 1.338 -1.495 1.200 8.121 -0.0 2.863 0.065 0.9 12.003 -31.248 Error (%) NL 3 1 2 3 RHA “Mode” “Modes” “Modes” “Modes” 0.8 0.5 -3.5 2 GM Multiplier 2.1 0.982 9.7 14.009 -0.983 1.371 0.971 1.071 0.616 -0.820 -19.376 1.370 -0.1 1.819 2.811 1.033 0.820 -0.5 2 GM Multiplier 2.8 1.5 1 1.333 0.044 1.473 -22.6 2.291 0.1 3.856 2.350 -0.049 -0.5 3 0 0 0.283 1.426 -1.003 0.9 12.8 1.010 0.317 0.8 1.676 0.430 1.3 6.256 1.863 1.235 -0.5 9.5 28.

4 with Tables 4.” and nonlinear RHA. However. the results were not always accurate. 4.3 and 4. MPA including three “modes” underestimates the displacements of the lower floors by up to 8% and overestimates the upper floor displacements by up to 14%. and three “modes. Pushover analysis seems to be inherently limited in computing accurately hinge plastic rotations.5). Obviously. even if three “modes” are included (Fig. and the errors in these estimates relative to the exact response from nonlinear RHA. Fortuitously. significant improvement is achieved by including response contributions due to the second “mode”.2). 4. Furthermore. the third “mode” does not contribute because this SDF system remains elastic (Fig. The drifts are underestimated by up to 13% in the lower stories. 4.3 and 4.13 and Tables 4. especially in estimating the story drifts (Fig. this error is not especially significant because the hinge plastic rotation is very small. the hinges identified in beams at the 6th floor were at variance with the “exact” results. and are within a few percent of the exact values for the upper stories. The errors are especially large in the hinge plastic rotations estimated by the MPA procedure. (3. 4. 4. two. For example.1 and 4.3 and 4. considering one.11c).9 and Tables 4.13 with Fig.4.3 and 4.5 present estimates of the combined response according to Eq.13 and Tables 4. for two or three modes included.12 and Tables 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. As shown in Figs.14. however.3 through 4. although the error is recorded as 100% if MPA estimates zero rotation when the nonlinear RHA computes a non-zero value. MPA failed to identify the plastic hinges at the column bases in Fig. The first “mode” alone is inadequate.Figures 4. significantly smaller than in UMRHA (compare Fig. 43 .” two “modes.12 and Tables 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. in general.4).18). The second “mode” was necessary to identify hinges in the upper stories.13 and Table 4. the errors in the modal pushover results are. 4. were determined by four analyses: MPA considering one “mode.14. 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. 4. but was more successful when the excitation was more intense (results are not included). but the third “mode” contributions do not seem especially important (Fig.12 and 4. The locations of plastic hinges shown in Fig. overestimated by up to 18% in the middle stories. 4. 4.4).” and three “modes.” respectively.

5 2 3 Actual Idealized 2000 0 0 0. 0.2 cm. 4. 1.5.5 0.75 0.0. 0. α = 0. V = 7616 kN.0 × El Centro ground motion 44 .75.13 y by Base Shear (kN) 8000 6000 1 1.85 Base Shear (kN) 8000 6000 4000 2000 0 0 0.14 y by Base Shear (kN) 8000 6000 4000 1. 1. “Modal” pushover curves with peak roof displacements identified for 0.5 1 0.25 1 2 4 6 Roof Displacement (cm) 8 10 Fig.6 cm. V = 5210 kN.25.9 cm.19 y by 3 2 1. α = 0.85 0.75 0. α = 0.25 5 10 15 20 Roof Displacement (cm) 25 (c) "Mode" 3 Pushover Curve 12000 10000 u = 4.0. and 3.11.5 0.(a) "Mode" 1 Pushover Curve 12000 10000 u = 36.5 2 3 4000 2000 0 0 Actual Idealized 0.5 0.75 Actual Idealized 20 40 60 Roof Displacement (cm) (b) "Mode" 2 Pushover Curve 80 12000 10000 u = 9.85 0.25 0. 2.5. V = 4952 kN.

5 1 1.5 1 1. Shown is the error in each floor displacement (Fig.5 Displacement/Height (%) 2 Ground 0 0.3 for elastic systems (or weak ground motions). each story drift (Fig. now standard in engineering practice.10 and 4. identified in Section 3.(a) Floor Displacements 9th 8th 7th 6th 9th 8th 7th 6th NL−RHA MPA 1 "Mode" 2 "Modes" 3 "Modes" (b) Story Drift Ratios Floor Floor 5th 4th 3rd 2nd 1st 5th 4th 3rd 2nd 1st NL−RHA MPA 1 "Mode" 2 "Modes" 3 "Modes" Ground 0 0.5 × El Centro ground motion. While various sources of errors in UMRHA. 4. the errors in MPA were larger for ground motion multipliers less than 0. 4.0. implying excitations weak enough to limit the response in the elastic range of the structure.15 summarizes the error in MPA considering three “modes” as a function of ground motion intensity. indicated by a ground motion multiplier. 4.15). and the error envelope for each case. 4. the errors in MPA were fortuitously smaller than in UMRHA (compare Figs. as discussed in Sections 3. whereas MPA contains errors inherent in modal combination rules.3.15) for ground multipliers larger than 1.4.12 Heightwise variation of floor displacements and story drift ratios from MPA and NL-RHA for 1. UMRHA is essentially exact. 4.75.5 Story Drift Ratio (%) 2 2.15b). The errors are only weakly dependent on ground motion intensity (Fig.15a). implying that the modal combination errors contained in these procedures are acceptable. 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. However. the MPA procedure is equivalent to the RSA procedure. an observation with practical implications. As mentioned in Section 3. implying excitations intense enough to cause significant yielding of the structure.5 Fig.4. 45 .4 also apply to MPA.2 and 3. shading indicates errors in MPA including three “modes” Figure 4. In this case.

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

6 -44.0 -100.00E+00 “Mode” 3 0.066 -0.50E-10 3.687 0.980 -0.00E+00 Floor 1st 2nd 3rd 4th 5th 6th 7th 8th 9th “Mode” 2 0.304 -1.19E-10 3.00E+00 0.1 46.88E-03 0.200 8.260 -14.575 -53.473 -15.527 -0.9 2.581 0.55E-03 3.641 1.3 13.5 1.00E+00 0.156 -0.4 1.3 -3.049 -0.220 1.517 1.304 1.00E+00 3 “Modes” 7.8 17.498 1.910 1.233 1.222 0.125 -1.76E-03 4.503 -1.00E+00 0.666 Table 4.503 1.8 -29.6 13.895 1.250 0.105 0.399 0.804 1.640 -1.2 0.244 0.581 0.331 1.9 1.745 1.0 -100.737 0.259 1.738 1.00E+00 0.900 -0.00E+00 0.640 1.2 0.26E-04 9.435 0.667 1.00E+00 0.3 -3.942 6.168 -0.00E+00 3.Table 4.00E+00 0.5 2.057 0.5 7.068 0.737 1.781 0.233 1.015 0.009 0.5 × El Centro ground motion Hinge Plastic Rotation (rad) “Modal” Response Combined (MPA) “Mode” 1 7.1 62.8 0.5 10.00E+00 0.36E-03 6.9 -100.222 0.1 13.72E-03 7.781 0.336 Error (%) NL 3 1 2 3 RHA “Mode” “Modes” “Modes” “Modes” 0.3 Peak values of floor displacements (as % of building height = 37.407 -27.18E-03 7.015 0.8 “Mode” 3 -1.298 0.6 -7.371 -0.5 2.72E-03 7.8 1.101 -0.76E-03 4.652 1.895 1.5 Peak values of hinge plastic rotations (radians) from MPA for 1.02E-03 0.116 1.220 1.00E+00 0.88E-03 1.190 -0.266 -0.752 1.6 -8.071 -0.1 46.0 -5.338 1.00E+00 0.00E+00 0.197 -0.1 18.683 1.0 2 “Modes” -32.399 0.76E-03 4.55E-03 3.315 0.053 -1.135 -7.0 15.00E+00 1 “Mode” 7.7 1.00E+00 0.2 -3.012 1.00E+00 0.36E-03 6.5 Table 4.9 7.76E-03 4.018 -0.478 0.00E+00 0.007 1.72E-03 7.018 0.9 0.60E-03 2.0 3 “Modes” -32.516 0.14 m) from MPA for 1.00E+00 0.945 -49.414 1.88E-03 0.36E-03 6.2 -100.72E-03 7.00E+00 0.176 0.36E-03 6.667 -1.820 -7.911 0.116 1.763 -14.37E-04 Error (%) “Modal” Response 1 “Mode” -32.980 0.066 -0.60E-04 7.2 -100.2 11.1 -8.8 -29.118 0.116 1.7 1.614 0.2 -4.4 -8.8 7.3 1.5 -6.756 0.1 62.0 -50.02E-03 0.6 1.7 -12.0 -100.426 15.007 1.10E-02 9.9 -100.705 1.154 0.4 Peak values of story drift ratios (as % of story height) from MPA for 1.668 -13.55E-03 3.055 0.694 1.130 0.879 1.982 13.05E-03 2.6 -9.033 -0.02E-03 3.6 13.00E+00 0.9 5.305 -0.018 -0.003 -16.705 -1.00E+00 2 “Modes” 7.8 -4.8 -12.611 0.02E-03 3.53E-03 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 0.8 -6.2 -100.1 62.8 -6.756 0.844 -7.311 0.429 0.0 1.372 0.6 -44.1 46.1 1.99E-03 6.594 -1.202 8.728 1.5 10.9 1.5 2.514 -1.209 1.8 1.351 -0.518 1.0 47 .37E-03 1.2 1.053 1.298 -0.0 -50.37E-03 1.088 12.8 -29.22E-10 NL RHA 1.5 × El Centro ground motion Story 1st 2nd 3rd 4th 5th 6th 7th 8th 9th “Modal” Response “Mode” 1 “Mode” 2 Drift Ratio (%) Combined (MPA) 1 “Mode” 2 “Modes” 3 “Modes” Error (%) NL 1 2 3 RHA “Mode” “Modes” “Modes” 1.37E-03 1.00E+00 0.071 0.00E+00 0.2 6.37E-03 1.3 11.9 -4.6 7.733 1.

Static analysis of the structure for gravity loads provides the initial state—forces and deformations—of the structure.8 present estimates of the combined response according to Eq.17 and Tables 4.18 and Tables 4.5).6 and 4. The response contributions of the second and third modes are the same as before (Tables 4. MPA including two “modes” underestimates the displacements of lower floors by up to 6% and overestimates the upper floor displacements by up to 22%. resulting in the pushover curve shown in Fig.7). The errors are especially large in the plastic hinge rotations estimated by the MPA procedure even if three “modes” are included (Fig.6 and 4. Results of such analyses for the selected building subject to 1. They are now included in the pushover analysis of the structure for the first “mode” only. and the structure is pushed using the force distribution of Eq.7).6). two. the third “mode” contributions do not seem especially important (Fig.20) with n = 1 to a target roof displacement.0 cm. 3.4. (3.11a). As shown in Fig. (3.16a—which is slightly different than the one excluding gravity loads (Fig.16 are unchanged. the results presented so far did not include gravity load effects.3 Modal Pushover Analysis with Gravity Loads To evaluate the accuracy of the dynamic response of the system.3 . 4. 4.20) with n = 1 (Fig.” but not the contributions of higher “modes. The story drifts are underestimated by up to 7% in the lower stories. the structure is pushed using the force distribution of Eq. as are the roof displacements ur 2o = 11.7).” these effects will modify the combined modal response as well as the results of nonlinear RHA. (3.6 and 4. and three “modes. The pushover curves for the second and third modes included in Fig.8). Also presented at these roof displacements are the displacements at all floors (Table 4.17 and 4.6 through 4.53 cm.7 cm and ur 3o = 2.” and the errors in these estimates relative to the exact response from nonlinear RHA. but it is inadequate in estimating the story drifts (Fig.18 and Tables 4. The first “mode” alone provides adequate estimates of floor displacements.18 and 48 . drifts in all stories (Table 4. 4.4. Figures 4.4) to roof displacement ur1o = 52. Significant improvement is achieved by including response contributions due to the second “mode.4.18) considering one.” however. and overestimated by up to 29% in the middle stories and by up to 13% in the upper stories. 4. Obviously gravity load effects will influence the seismic demands due to the first “mode.17 and Tables 4. 4. and plastic hinge rotations at selected external beam end at each floor level (Table 4. 4. Starting with its initial state under gravity loads.7.5 × El Centro ground motion are presented below. 4.

5 1 1.5 × 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. 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. Most pushover analysis procedures do not seem to compute to acceptable accuracy plastic hinge rotations.5 2 GM Multiplier 2. and (b) story drifts 49 .5 1 1. Errors in MPA as a function of ground motion intensity: (a) floor displacements. two.Table 4.8).15. 4.5 2 GM Multiplier 2.14. Noted 60 50 40 (b) Story Drifts Error Envelope Error for Story No. 1−"Mode" • • • • • • • • •• •• •• •• •• •• •• •• •• •• •• •• •• •• • • •• • •• • • • • • •• •• • •• •• •• •• (b) MPA.5 3 Fig. (a) MPA. 4. 2−"Modes" •• •• •• •• •• •• •• •• •• •• •• •• •• •• • • • • •• • •• (c) 3−"Modes" • • • • • • • •• •• • •• •• •• •• •• •• •• •• •• •• •• •• •• •• •• •• •• •• • • • • •• • •• • • • • • • • • • (d) Nonlinear RHA •• •• • •• •• •• •• • •• •• •• •• •• •• •• • •• •• • •• •• •• •• • • • • • • •• •• •• Fig.5 3 0. Locations of plastic hinges determined by MPA considering one.

4. Figure 4. indicated by a ground motion multiplier.” two “modes”. Shown is the error in each floor displacement (Fig. With two modes included in MPA. excluding gravity load effects. 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. 4.20b).20 summarizes the error in MPA considering three “modes” as a function of ground motion intensity.20a). 50 .19 were determined by four analyses: MPA considering one “mode. and the error envelope for each case. 4. and nonlinear RHA. each story drift (Fig. and the errors are only weakly dependent on ground motion intensity.The locations of plastic hinges shown in Fig. The second “mode” is necessary to identify hinges in the upper stories.15. MPA provides response values accurate enough for practical application in building retrofit or design. 4. These errors are only slightly larger than those in Fig. this procedure is able to predict plastic hinge locations essentially consistent with nonlinear RHA.

5 2 3 Actual Idealized 2000 0 0 0.(a) "Mode" 1 Pushover Curve 12000 10000 u = 36.9 cm.5 0.25. 1.6 cm. 0.19 y by 3 2 1. V = 5210 kN.5 1 0. 4. V = 7433 kN.0.5 0. V = 4952 kN.25 1 2 4 6 Roof Displacement (cm) 8 10 Fig. “Modal” pushover curves with gravity loads included. and 3.75 Actual Idealized 20 40 60 Roof Displacement (cm) 80 (b) "Mode" 2 Pushover Curve 12000 10000 u = 9. α = 0.75. α = 0.85 0.75 0.50.85. α = 0.5 0.13 y by Base Shear (kN) 8000 6000 1 1.75 0.14 y by Base Shear (kN) 8000 6000 4000 1. noted are peak values of roof displacement for 0.0. 0.16. 0.85 Base Shear (kN) 8000 6000 4000 2000 0 0 0.25 5 10 15 20 Roof Displacement (cm) 25 (c) "Mode" 3 Pushover Curve 12000 10000 u = 4.25 0. 2.5 2 3 4000 2000 0 0 Actual Idealized 0.85 0.3 cm.0 × El Centro ground motion 51 .

5 × El Centro ground motion. two. gravity loads included. story drifts.5 Fig.5 1 1. 4.5 Displacement/Height (%) 2 Ground 0 0. 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. Heightwise variation of floor displacements and story drift ratios from MPA and NL-RHA for 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. and hinge plastic rotations estimated by MPA including one.5 Story Drift Ratio (%) 2 2.17. 4. and 52 .5 1 1.18. Errors in floor displacements.

2−"Modes" •• •• •• •• •• •• •• •• •• •• •• •• •• •• • • • • •• •• •• (c) MPA. Locations of plastic hinges determined by MPA considering one.5 1 1.5 3 0.20.5 1 1.19. gravity loads included 53 . gravity loads included (a) Floor Displacements 60 50 40 Error Envelope Error for Floor No. 4.5 × 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. Noted 60 50 40 (a) Story Drifts Error Envelope Error for Story No. and three “modes” and by NL-RHA for 1.5 2 GM Multiplier 2.5 2 GM Multiplier 2.5 3 Fig. Errors in MPA as a function of ground motion intensity: (a) floor displacements. 3−"Modes" (d) Nonlinear RHA • • • • • • • •• •• •• •• •• •• •• •• •• •• •• •• •• •• •• •• •• •• •• •• •• • • • • •• •• •• • • • • • • • • •• •• • •• •• •• •• • • •• •• • •• •• •• •• • • •• •• • •• •• •• •• • • • • • • •• •• •• Fig.(a) MPA. 1−"Mode" • • • • • • • •• •• •• •• •• •• •• •• •• •• •• •• •• •• •• • • •• •• •• • • • • • •• •• •• •• •• •• •• (b) MPA. two. and (b) story drifts. 4.

168 -0.7 -2.176 0.3 9.00E+00 0.996 0.996 -0.908 -1.2 12.850 -1.888 1.514 -1.2 1.213 1.114 -1.8 Peak values of hinge plastic rotations (radians) from MPA for 1.330 1.11E-03 9.8 0.850 1.4 20.6 Peak values of floor displacements (as % of building height = 37.071 -0.00E+00 0.00E+00 0.983 1.7 Peak values of story drift ratios (as % of story height) from MPA for 1.5 3 “Modes” -32.754 0.214 0.860 1.1 1.213 1.88E-03 0.7 16.114 1.037 0.667 0.107 1.4 1.3 -22.319 1.270 -12.00E+00 0.8 Table 4.071 0. gravity loads included Story 1st 2nd 3rd 4th 5th 6th 7th 8th 9th “Modal” Response “Mode” 1 “Mode” 2 Drift Ratio (%) Combined (MPA) 1 “Mode” 2 “Modes” 3 “Modes” Error (%) NL 1 2 3 RHA “Mode” “Modes” “Modes” 1.35E-10 3.311 0.5 × El Centro ground motion.516 0.744 1.371 -0.822 1.490 -11.530 1.429 -1.066 -0.0 37.19E-04 5.130 0.673 Table 4.953 15.2 9.3 -100.3 -22.9 -3.913 7.353 -23.237 0.125 -1.2 -0.2 9.921 1.938 1.9 -4.8 1.00E+00 0.515 -50.04E-10 3.466 0.372 0.4 -4.3 1.7 1.0 -100.23E-03 0. gravity loads included Hinge Plastic Rotation (rad) Floor “Modal” Response “Mode” 1 1st 2nd 3rd 4th 5th 6th 7th 8th 9th 8.854 0.908 1.836 -0.23E-03 3.478 0.11E-03 9.55E-03 3.018 -0.2 4.4 -6.2 0.877 -46.5 × El Centro ground motion.55E-03 3.983 1.190 -0.033 -0.207 18.066 -0.603 1.266 -0.2 2.13E-03 5.594 -1.3 -100.049 -0.6 2.434 Error (%) NL 3 1 2 3 RHA “Mode” “Modes” “Modes” “Modes” 0.728 1.009 0.1 4.00E+00 0.0 -30.854 0.927 1.858 2.783 1.5 1.305 -0.8 -32. 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” 0.8 28.0 16.7 19.109 0.102 1.637 0.14 m) from MPA for 1.00E+00 0.3 -100.197 -0.00E+00 0.064 -10.434 0.00E-03 5.436 1.399 -0.78E-03 1.35E-03 8.00E-03 5.5 54 .1 21.665 0.821 1.23E-02 1.00E+00 0.00E+00 0.454 1.00E+00 Error (%) “Modal” Response 1 “Mode” -32.88E-03 1.00E+00 3 “Modes” 8.Table 4.00E+00 0.998 1.015 0.6 0.35E-03 8.00E+00 “Mode” 2 0.686 -7.11E-03 9.530 1.23E-03 0.154 0.4 -4.55E-03 3.199 16.101 -0.19E-03 1.2 9.037 -0.6 2.4 1.461 0.330 1.8 0.00E-03 5.00E+00 “Mode” 3 0.00E+00 3.310 1.00E+00 2 “Mode” 8.00E-03 5.35E-03 8.257 0.998 21.998 0.5 21.414 28.5 “Mode” 3 -1.507 1.0 2 “Modes” -32.0 37.19E-03 1.00E+00 Combined (MPA) 1 “Mode” 8.055 0.156 -0.19E-03 1.8 -32.88E-03 0.0 -100.933 1.5 10.603 -1.5 1.102 1.319 1.754 1.75E-03 0.04E-02 8.2 16.821 -1.11E-03 9.351 -0.530 1.813 1.00E+00 0.5 -5.2 1.105 0.6 19.2 4.434 0.00E+00 0.263 1.399 0.0 37.068 0.213 1.831 0.0 -30.7 4.5 21.237 0.00E+00 0.5 × El Centro ground motion.35E-03 8.19E-03 1.00E+00 0.2 4.1 13.2 21.7 -2.831 0.3 13.315 0.6 1.527 -0.2 -2.128 -1.9 -6.5 0.23E-03 3.687 0.098 20.17E-03 9.00E+00 0.8 9.00E+00 0.2 -3.4 26.00E+00 0.637 0.263 0.3 -22.00E-10 NL RHA 1.830 -12.057 0.00E+00 0.9 31.26E-03 3.429 1.

2 COMPARATIVE EVALUATION Compared in this section are the earthquake-induced demands for the selected building determined by five analyses: pushover analysis using the three force distributions in FEMA-273. k = 2 for T1 ≥ 2.5 sec . Comparison of Modal and FEMA Pushover Analyses 5.5. story drifts. 5. computed at the target displacement represent the earthquakeinduced demands on the structure.5 sec . The floor displacements. 1997] The pushover curve. MPA considering three “modes. 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. j 2. SRSS distribution: s* is defined by the lateral forces back-calculated from the story shears determined by response spectrum analysis of the structure. assumed to be linearly elastic. and varies linearly in between. and 3.” and nonlinear RHA. etc. and the exponent k = 1 for fundament period T1 £ 0. joint rotations. Specified in FEMA-273 are three distributions for lateral forces: 1.1 FEMA-273 PUSHOVER ANALYSIS In this investigation we focus on one step in the nonlinear static procedure in the FEMA-273 document [Building Seismic Safety Council. 2… N ). “Uniform” distribution: s* = m j (where the floor number j = 1. Equivalent lateral force (ELF) distribution: s* = m j h k where h j is the height of the j j jth floor above the base. plastic hinge rotations. gravity load effects were included in all 55 .. a plot of base shear versus roof displacement.

0 cm.3.4.0466 0. The errors in the FEMA and MPA estimates of seismic demands relative to the “exact” demands are presented in Fig.11 0. with the errors being larger for the SRSS distribution.367 (a) Uniform (b) ELF (c) SRSS Fig.11 0.042 0.11 0.1 through 5.21 0.11 0. The MPA procedure is more accurate than all the FEMA force distributions. Also included in these presentations are the MPA results considering three “modes” and the “exact” demands from nonlinear RHA. The “uniform” distribution overestimates all floor displacements by 17-28%.4. Force distributions in FEMA-273: (a) “uniform”.11 0.281 0. 0.4a. the story drift demands in Fig. and the displacements of upper floors are overestimated by 9-22%.3b and Table 5.5 times the El Centro ground motion.11 0. Figures 5.0702 0.1 demonstrate that the displacement demands are underestimated by the ELF and SRSS force distributions by 12 to 30 % at the lower six floors of the building. The first four floor displacements are within 5% of the “exact” value. and Table 5.165 0. 5.3. pushover analyses are implemented for a target roof displacement of 52.2.1.112 0. (b) ELF.062 0. 5.3a and 5.0896 0. 5.0381 0.1. 5.0913 0.1. and (c) SRSS 56 . Using each of these force distributions.3. and the locations of all plastic hinges in Fig.3a and Table 5. wherein the first two are obvious and the third is determined from response spectrum analysis of the building (Appendix C).177 0. The pushover curves are given in Fig. the floor displacement demands in Fig. the value determined from RHA of the first-mode inelastic SDF system for 1.126 0.0654 0.0197 0.119 0.analyses. The three FEMA force distributions are presented in Fig. both presented in Section 4. 5.0981 0.11 0. 5. 5.0446 0.2. plastic hinge rotation demands in Table 5.4 and Tables 5.00719 0.

8 cm. and (c) SRSS. gravity loads are included 57 . V = 7456 kN.3 cm. V = 6897 kN. (b) ELF. α = 0. V = 8530 kN.2. 5. α = 0. Pushover curves using three force distributions in FEMA-273: (a) “uniform”.(a) Uniform Distribution 12000 10000 u = 33 cm.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 = 38.18 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 = 0.24 y by Base Shear (kN) 8000 6000 4000 2000 0 0 Actual Idealized 10 20 30 40 Roof Displacement (cm) 50 60 Fig.

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

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

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

14 m) from FEMA force distributions and MPA.84 1.8 -35.10E-02 7.3 -100.23E-02 1.984 1.938 1.04E-02 8.310 1.294 1.3 -4.1.730 1.8 -100.007 1.00E+00 NL RHA 1.35E-03 8.6 -25.462 1.2 4.530 1.34E-03 2.314 1.52E-04 1.783 1.098 1.913 0.466 0.8 -63.8 2.5 -27.4 -12.064 1.011 1.4 -23.4 -55.2 19.59E-03 5.5 26.9 -77.4 MPA -2.5 15.335 2.78E-03 0.341 1.263 0.00E+00 0.00E+00 0.836 0.789 0.487 0.62E-03 0.5 10.6 22.6 -41.1 10.207 1.1 163.Table 5.6 -17.061 1.8 -32.7 21.4 0.00E+00 0.00E+00 0.515 FEMA Uniform ELF 27.178 1.209 0.45E-03 3.330 1.667 0.089 1.7 7.8 -100.0 -100.0 -100.597 0.7 FEMA ELF SRSS -22.4 21.809 0.888 0.51E-03 4.953 0.5 Story 1st 2nd 3rd 4th 5th 6th 7th 8th 9th Uniform 2.547 -27.6 -6.839 0.306 1.562 1.8 7. gravity loads included Displacement /Height (%) FEMA ELF SRSS MPA 0.94E-03 2.2 6.724 0.78E-03 1.8 44.109 1.7 -29.566 1.2 9.686 0.270 0.00E-03 5.083 1.7 Table 5.50E-03 0.2 Peak values of story drift ratios (as % of story height) from FEMA force distributions and MPA.067 0.3 -22.2 13.4 -50.199 27.19E-04 5.3 23.0 -57.3 -11.65E-03 7.7 -28.2 5.00E+00 0.03E-03 5. gravity loads included Displacement /Height (%) FEMA ELF SRSS MPA 1.2 -12.4 16.7 -15.8 Floor 1st 2nd 3rd 4th 5th 6th 7th 8th 9th Uniform 0.7 26.9 15.436 Error (%) NL Uniform RHA 0.11E-03 9.372 1.344 0.3 Peak values of hinge plastic rotations (radians) from FEMA force distributions and MPA Hinge Plastic Rotation (rad) MPA FEMA ELF SRSS 4.0 Table 5.4 16.399 1.015 0.927 1.318 2.13E-03 5.0 -100.998 1.2 16.560 1.6 4.53E-02 1.351 0.128 1.992 1.00E+00 Error (%) FEMA Uniform ELF SRSS MPA 24.858 1.88E-03 1.490 0.367 1.55E-03 3.1 -3.2 -100.877 1.7 -60.6 -11.1 -26.154 1.323 1.3 -14.830 2.875 0.524 0.2 16.5 22.1 0.355 0.0 -71.8 17.9 16.221 1.611 0.35E-10 3.708 0.6 21.00E+00 8.417 1.2 9.9 -70.00E+00 4.399 1.0 -30.860 1.5 -29.2 -4.09E-03 4.5 -33.26E-03 3.998 1.168 1.262 1.23E-03 3.5 -33.7 19.9 16.0 37.0 -63.530 1.9 -20.8 -2.5 -29.736 0.5 -3.195 0.58E-04 6.19E-03 1.214 1.00E+00 2.93E-03 1.2 10.17E-03 9.9 11.623 1.6 14.00E-10 Floor 1st 2nd 3rd 4th 5th 6th 7th 8th 9th Uniform 1.854 0.414 1.8 16.0 -59.5 -32.9 28.234 1.353 1.00E+00 0. Peak values of floor displacements (as % of building height = 37.975 1.6 17.75E-03 0.673 Error (%) NL RHA 1.5 61 .399 -27.9 SRSS -22.9 0.6 -73.0 -100.672 1.7 MPA -2.3 29.3 31.6 -4.16E-03 0.

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

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

7.5 2 GM Multiplier 2.5 2 GM Multiplier 2.5 1 1.(a) FEMA: Uniform 100 60 50 9 8 (b) FEMA: ELF Error Envelope Error for Story No. 5.5 1 1.5 3 0 0 0. Error in story drifts from three force distributions in FEMA-273 and from MPA including three “modes”.5 3 (c) FEMA: SRSS 60 50 40 60 50 40 3 2 1 (d) MPA (3 "Modes") Error (%) 30 20 Error (%) 30 5 4 8 9 20 9 6 7 7 10 0 0 4 10 2 3 1 8 5 6 0.5 2 GM Multiplier 2.5 2 GM Multiplier 2.5 1 1. Noted 80 40 9 1 Error (%) 7 Error (%) 60 30 20 10 2 5 6 4 8 40 1 6 20 2 5 3 4 3 7 0 0 0.5 1 1.5 3 0 0 0. gravity loads included 64 .5 3 Fig.

where the nature and magnitude of errors arising from approximate modal combination rules are well understood. pushover analysis is inherently limited in the sense that it cannot provide any cumulative measure of response. SRSS) leads to the modal pushover analysis (MPA) procedure. 65 . the energy dissipated in yielding or the cumulative rotation of a plastic hinge.g. and the structure is pushed to the roof displacement determined from the peak deformation Dn of the nth-mode elastic SDF system. etc.g. joint rotations. where m is the mass matrix n of the building and φ n its nth mode. Combining these peak modal responses by an appropriate modal combination rule (e. Pushover analysis of a one-story inelastic system predicts perfectly peak seismic demands: deformation. 2. 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 .. ζ n —of the nth-mode of the MDF system. However. 3. e. Dn is available from the elastic response (or design) spectrum.. which retains the conceptual simplicity and computational attractiveness of the procedure with invariant force distribution. This system has vibration properties—natural frequency w n and damping ratio. For this system. It has led to the following conclusions: 1.6 Conclusions This investigation was aimed toward developing an improved pushover analysis procedure based on structural dynamics theory. This MPA procedure for elastic buildings is shown to be equivalent to the standard response spectrum analysis (RSA) procedure. now common in structural engineering practice. hinge plastic rotation.

2. p eff (t ) = . To enable systematic extension of the MPA procedure to inelastic systems. p eff . 66 . 5.4.n t is determined by pushover analysis..n t = . These underlying assumptions and approximations are evaluated and the errors in the UMRHA procedure relative to the “exact” nonlinear response history analysis are documented.g. an uncoupled modal response history analysis (UMRHA) is developed by (1) neglecting the coupling among modal coordinates arising from yielding of the system.sn ug t —n denotes mode number—in the modal expansion of the effective earthquake forces. …) according to an appropriate combination rule (e. (ii) Idealize the pushover curve as a bilinear curve and convert it to the bilinear force-deformation relation for the nth-“mode” inelastic SDF system. n where φ n is now the nth natural mode for small-amplitude linear vibration. and (2) superposing responses of the inelastic MDF system to individual terms. 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 . the SRSS rule). and (ii) the total response ro is determined by combining the rno (n = 1. with vibration properties in the linear range same as those of the nth-mode elastic SDF system. 6.urn bg bg bg bg b g curve from a pushover analysis of the structure for the force distribution s* = mφ n . (iii) Compute the peak deformation Dn of this system with unit mass by nonlinear response history analysis or from the inelastic response (or design) spectrum.mι ug t . The response value rno is determined by implementing the following steps: (i) Develop the base-shearroof-displacement Vbn .

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.(iv) At the roof displacement determined from Dn .” 10. and lead to unacceptably large errors in the hinge plastic rotations. 67 . Thus the MPA procedure is accurate enough for practical application in building evaluation and design. and identifying locations of most plastic hinges. However. MPA. and hinge plastic rotations. all pushover analysis procedures do not seem to compute with acceptable accuracy local response quantities. 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. 9. the pushover analysis provides the peak value rno of any response quantity: floor displacements.0. 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. Based on results presented for El Centro ground motion scaled by factors varying from 0. story drifts. Comparing the earthquake-induced demands for the selected 9-story building determined by pushover analysis using three force distributions in FEMA-273. demonstrates that the FEMA force distributions greatly underestimate the story drift demands. etc. the errors in the MPA procedure are shown to be only weakly dependent on ground motion intensity. and nonlinear RHA. plastic hinge rotations. The MPA procedure is more accurate than all the FEMA force distributions in estimating floor displacements. such as hinge plastic rotations. 8. it fails to compute with acceptable accuracy plastic rotations of the hinges. Pushover analyses seem to be inherently limited in computing accurately hinge plastic rotations. 7.25 to 3. joint rotations. story drifts.

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

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

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

develop the base-shear – roof-displacement (Vbn . 3. Apn . ω n .2. (3. respectively. 2. Idealize the pushover curve as a bilinear curve (Fig.g. 1977). For the nth-“mode”.. Since the target roof displacement may not be known at the start of the procedure. 3.2. This step can be conveniently implemented in any commercially available software.g. B. Apply force distribution of Step 2. 3. 1.. Let the roof displacement and base shear at the anchor point be urno and Vbno . Calculate the area under the actual pushover curve. Define the force distribution s* from Eq. The structure should be pushed just beyond the target (or expected) roof displacement in the selected mode. iterations may be necessary. using any numerical integration method. and modes.1. 71 .1 incrementally and record the base shears and associated roof displacements.1.Appendix A Uncoupled Modal Response History Analysis A. A. Compute natural frequencies. Define the anchor point.urn ) pushover curve for the force distribution s* : n 2. e. of the bilinear curve at the target roof displacement.1) using the FEMA-273 procedure (Building Seismic Safety Council. 1988).20): s* = mφn n n 2. trapezoidal rule. for linear-elastic vibration of the building.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. DRAIN2DX (Allahabadi and Powell. e. φn .

Calculate the yield displacement. Calculate area under the bilinear curve OAB.4 to 3.2. 3. 3. kn = 0. i i i 3. 3. 72 .9. This step gives the secant stiffness at a base shear equal to 60% of the yield base shear.9. Vbny . i 3. Replace i+1 with i and repeat Steps 3. If desired. i i 3. From the pushover data. at base shear i equal to 0. If the error exceeds some pre-specified tolerance. will be refined by an iterative procedure that seeks to equate areas under the actual and the idealized pushover curves. Estimate the yield base shear.0.8. and a point on the actual pushover curve with base shear equal to i 0. Let the point with base shear = Vbny and roof displacement = urny be denoted as A.6 × Vbny . i 3.i 3.5. Calculate the post-yielding strain-hardening ratio. urny = Vbny kn . O. other appropriate methods can be ( used. Calculate the slope.1. and B with straight-line segments to obtain the idealized bilinear curve. Vbny .1. Draw the curve OAB by connecting the three points O. corresponding to the estimated yield i i i base shear.3.4.6 × Vbny ( ) i urn . kn . Calculate Vbny = Vbny × Apn Abn . Calculate initial slope of the idealized bilinear curve. i 3. obtained by judgment. by connecting a straight line between origin. determine the roof displacement. iterations are necessary.8. i +1 i i 3. urn.7.6 .4.6 . i i i αn = Vbno Vbny − 1 urno urny − 1 i 3. This value.6 × Vbny .2.9. A.0.6. Abn . Calculate the error = 100 × Abn − Apn ( ) ( ) ( ) ) Apn .4.

(3. In general first two or three modes will suffice.2. 7. Compute deformation history.2. Vbn Idealization of Pushover Curve Vbno Vbny 0. of the nth“mode” inelastic SDF system (Fig. and pseudo-acceleration history. A. of the combined responses obtained in Step 8.3. 6.3b) with unit-mass and force-deformation relation of Fig. 9. Scale the horizontal axis by Γnφrn to obtain Dno = urno Γ nφrn and Dny = urny Γ n φrn (Eqs. A. 5.10a and 4.4) and effective modal mass from M n = Ln Γn . A.11b).12) and (3. (3. Develop the Fsn Ln − Dn relation (Fig. Calculate peak values.2). Scale the vertical axis with * Mn to obtain * Fsno Ln = Vbno M n and * Fsny Ln = Vbny M n (Eqs. (3. 4.13).16).11a). 4. r o .1 Idealization of nth-“mode” pushover curve 73 . Repeat Steps 3 to 6 for as many modes as required for sufficient accuracy. Compute the Ln and Γ n from Eq.6 ur n y ur n o ur n Fig.6 × Vbny A • Idealized αnkn B • 1 Actual k 1 • O n figA_1. Combine the “modal” responses using Eqs.10b and 4. Dn (t ) . * 4.eps ur n.0. 4.1.4.15) and (3. 4. 4. An (t ) . Calculate histories of various responses using Eqs. 8.

respectively. generated using DRAIN-2DX.). 3.urn ) pushover curve for the force distribution s* : n 2. Following is step-by-step implementation of the procedure described in Section A.). A. A.4 cm (10 in. Target displacements used to generate these pushover curves are 63. 3. 74 . The force distributions. 3.3. First three mode shapes and frequencies of the selected building were computed and are shown in Fig.6 kN. second.2. The anchor point.4. and 12. The base-shear – roof-displacement (Vbn .5 cm (25 in.Fsn / Ln Fsn / Ln − Dn Relationship Vbno / M* n Vbny / Mn * 1 α ω2 n n ω2 1 n figA_2.3. ur1o = 63. A. is defined at the target roof displacement. for the first. 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. 25. Area under the actual pushover curve. Ap1 = 360777 kN-cm. B.2.2 EXAMPLE The UMRHA procedure is implemented to calculate the response of the 9-story building described in Section 3.1.). s* . to the north-south component of the El Centro (1940) ground motion scaled up by a factor of 1.5. The pushover curves for the first three modes. 3. and third mode. n 2. 2.4.1. At this point. computed for the first three modes are shown in Fig. 3.5 cm and Vb1o = 8729.1. are shown in Fig.7 cm (5 in.2 Properties of the nth-“mode” inelastic SDF system A. 1. Idealized bilinear curves for each of the three modes are included in Fig.3.

8 22. The next estimate of the yield shear is Vbny = 8006. The results of the iterative procedure are summarized in Table A. is calculated as follows. The results for other modes are summarized in Table A.7.4 kN.18 = 38. The procedure converged after nineteen cycles to give ur1 y = 36.198%.5 38.4 ) − 1 (63. k1 = 0.4 × (360777 365100 ) = 7911.1.6 8006.09 cm. i 3.3666 = 3740189 kg. i i i 3.6 = 22.9.9 kN. i 3.5. ur1 y = Vb1 y k1 = 8006. L1 = 2736789 kg. 75 .13).0. and B with straight-line segments gives the idealized bilinear curve. Γ1 = 1. α1 = Vb1o Vb1 y − 1 ur1o ur1 y − 1 = ( ) ( ) (8729.9.4 210. k1 .09 ) − 1 = 0. and α1 = 0. 3.6 kN. The first estimate of the yield base shear Vb1 y = 8006. The post-yielding strain-hardening ratio.6.6 × Vb1 y = 4803.86 = 210.i 3. 3. The initial slope of the idealized bilinear curve. Ab1 = 365100 kN-cm.1. and M1 = 2736789 × 1. Determined from the pushover database.135.3.9. (4.4.0. 4.23 cm.6 × Vb1 y ( ) i ur1. i i 3.8 kN. * 4.8.194.86 cm at 0. This value exceeds the prespecified tolerance of 0. Also included are the modal damping ratios and the periods calculated from Eq.4.09 cm and Vb1 y = 8006. 3.4.18 kN/cm. iterations are necessary.2. i i i 3. Therefore. i +1 3. A. The yield displacement.2.4 kN. The Fs1 L1 − D1 relation for the first “mode” is developed as follows. The point A on the i i bilinear curve is defined by ur1 y = 38.01%. i i 3.3666. ur1. Error = 100 × (365100 − 360777 ) 360777 = 1. Vb1 y = 7615. The curve OAB obtained by connecting the three points O.6 = 4803.2. Area under the bilinear curve OAB.1.1.

The combined modal responses are presented in Fig.7. Scaling the vertical axis by Fs1 y L1 = 203. A.4. The results were generated for first three “modes’ and are included in Fig. 8.4.2.8. 4. 5.3. 4. 4. 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.7.51 cm. Histories of roof displacement and top story drifts for the first “mode” are computed and presented in Fig. Scaling the horizontal axis by Γ1φr1 gives D1o = 46. The peak values are also plotted in Fig.40 (cm/sec2) and 76 . 4. 9.2. * M1 gives Fs1o L1 = 233.1 and 4.7.46 cm and D1y = 26. The peak values are computed and are summarized in Tables 4.62 (cm/sec2). 4. 6. 7.

V = 7616 kN.eps 10 20 30 40 50 Roof Displacement (cm) (b) "Mode" 2 Pushover Curve 60 8000 u = 9.9 cm.19 y by Base Shear (kN) 8000 6000 4000 2000 0 0 Actual Idealized figA_3a. α = 0. α = 0. V = 5210 kN.(a) "Mode" 1 Pushover Curve 12000 10000 u = 36. “Modal” pushover curves for the example building 77 .13 6000 Base Shear (kN) 4000 2000 Actual Idealized figA_3b.3. V y by = 4952 kN.eps 0 0 5 10 15 20 Roof Displacement (cm) (c) "Mode" 3 Pushover Curve 25 8000 u = 4.14 y by 6000 Base Shear (kN) 4000 2000 Actual Idealized figA_3c. A.2 cm. α = 0.6 cm.eps 0 0 2 4 6 8 10 Roof Displacement (cm) 12 Fig.

2769 D1 (cm) 0 • 35.223 −80 80 n=3 −3 3 n=3 D3 (cm) 0 A3 (g) • 10.5 × El Centro ground motion for the first “mode.” second “mode.” and third “mode” inelastic SDF systems.m −80 0 −3 0 25 30 Fig.52 5 10 15 20 Time (sec) 25 30 0 • 1.06 0 • 1. A.747 5 10 15 20 Time (sec) figa_4. Histories of deformation and pseudo-acceleration due to 1.(a) Deformation 80 n=1 3 (b) Pseuso−acceleration n=1 • 0. 78 .4.33 A1 (g) n=2 0 −80 80 −3 3 n=2 D2 (cm) 0 A1 (g) • 22.

35 36.18 210.44 36.86 21.910 0.25 36.25 36.191 0.2.32 36.8525 1.2 4571.4 7911.193 0.4 7672.78 21.139 0.193 0.1 1013.010 (kN) 8006.9 7615.24 36.18 210.18 210.09 18.037 0.18 210.082 0.193 0.85 0.62 26.529 0.18 210.18 210.176 0.18 210.18 210.18 210.6 × Vb1 y i ur1.3 4603.063 0.103 “Mode” 3 696400 0.2671 1.3 7658.4927 1.5309 488839.1 4574.2 7690.188 0.39 36.5 (cm) 22.38 22.182 0.162 0.23 0.0 4577.3 7786.1 7616.9 4573.194 0.135 0.0 7619.18 210.65 1226.8 4747.51 2 “Mode” 2 -920860 -0.192 0.18 210.9 4570. No.5 7633.70 36.194 0.64 37.85 36.26 36.180 0.59 22.18 (cm) 38.05 52.05 36.5 4614.107 0.309 0.81 21.3666 3740189 203.404 0.9 4570.40 46.5 3109.237 0.0 4588.198 0.76 21.18 210.186 0.18 210.18 210.86 22.029 0.18 210.46 2.77 21.4 7714.0 4704.8 7618.136 Fsny Ln (cm/sec2) Dny (cm) Fsno Ln (cm/sec ) Dno (cm) Tn (sec) 233.18 210.184 0.2 4671.56 47.09 37.194 0.Table A.95 21.194 (kN-cm) 365100 364060 363276 362684 362235 361892 361631 361432 361279 361162 361073 361004 360951 360911 360880 360856 360838 360824 360813 Table A. Properties of “modal” inelastic SDF systems Properties Ln (kg) Γn * M n (kg) “Mode” 1 2736789 1.28 36.18 210.56 19.79 0.693 0.83 21.8 7622.170 0.11 22.151 0. i 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 i Vb1 y i 0.013 0.29 36.75 21.7 7639.4 4595.9 (kN) 4803.6 i k1 i ur1 y i α1 i Ab1 Error (%) 1.90 21.7 4580.30 37.190 0.74 (kN/cm) 210.18 210.2406 167531.1.02 21.74 21.017 0.948 ζ n (%) 79 .23 22.022 0. Results of iterative procedure to develop the idealized bilinear curve for the first “mode” inelastic SDF system Itr.2 4628.3 7628.5 7624.6 4583.1 4569.048 0.0.4 7647.6 7840.3 7745.79 21.75 21.8 4647.12 3876.50 36.59 36.

80 .

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. 10. Determine the total response by combining the peak “modal” responses using the SRSS combination rule of Eq. From the total hinge rotation. extract from the pushover database values of other desired responses.21).18). Typically. rno .5. subtract the yield hinge rotation to determine the plastic hinge rotation. 12. 11.4. B.1 to the north-south component of the El Centro (1940) ground motion scaled up by a factor of 1. At urno . or from the inelastic response (or design) spectrum. Following is step-by-step implementation of the procedure described in Section B. 10. (3.6b by solving Eq.8).2 EXAMPLE The MPA procedure is implemented to calculate the response of the 9-story building described in Section 3. Steps 1 to 4 of the MPA are the same as those for UMRHA presented in Appendix A. Solving Eq. 4. 4.1. Calculate the peak roof displacement urno associated with the nth-“mode” inelastic SDF system from Eq.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. 13. Dn . the first two or three “modes” will suffice. Repeat Steps 3 to 7 for as many “modes” as required for sufficient accuracy.3b) with unit mass and force-deformation relation of Fig. (4. of the nth-“mode” inelastic SDF system (Fig. (4.33 cm.Appendix B Modal Pushover Analysis B. Compute the peak deformation. (3. This is the same example as solved in Appendix A. 6. 81 .

3 for the floor displacements normalized by the building height and in Table 4.18) are also included in Tables 4. (3. Results of Steps 5 and 6 are summarized in Table B. 8.3 and 4.1.28 cm.3 and 4. Peak roof displacement ur1o = Γ1 × φr1 × D1 = 1. 82 .5 are the plastic hinge rotations computed by subtracting the yield hinge rotation from the total hinge rotations. The total response computed by combining the peak “modal” responses using the SRSS combination rule of Eq.4. The values are summarized in Table 4. 12.” and the results are included in Tables 4. At ur1o = 48. where results for other ground motion intensities are also included. Also included in Table 4.11.28 cm.4.4 for the story drifts normalized by the story height. values of floor displacements and story drifts are extracted. Steps 3 to 7 are repeated for first three “modes. 9.366 × 1 × 35.33 = 48.

748 63.436 7.154 78.25 20.35 1.13 2.71 1.37 57.007 36.504 18.35 0.467 14.969 0.332 48.023 0.1.139 µ = Dn Dny urno = Γnφrn Dn (cm) Dn (cm) 0.36 1.755 0.312 1.59 0.513 0.229 8.18 27.38 22.0 µ = Dn Dny urno = Γnφrn Dn (cm) Dn (cm) 3.50 35.8451 5.856 31.0 µ = Dn Dny urno = Γnφrn Dn (cm) 83 .0 µ = Dn Dny urno = Γnφrn Dn (cm) Dn (cm) 1.03 0.267 5.577 16.755 27.691 0. Calculation of roof displacements urno from peak deformation of inelastic SDF systems Ground Motion Multiplier 0.457 12.117 5.05 0.676 6.660 14.06 1.185 11.126 13.07 “Mode” 2 4.85 µ = Dn Dny urno = Γnφrn Dn (cm) Dn (cm) 1.225 2.52 0.03 26.332 13.37 1.75 µ = Dn Dny urno = Γnφrn Dn (cm) Dn (cm) 0.Table B.690 10.33 1.52 “Mode” 3 1.901 8.367 1.5 µ = Dn Dny urno = Γnφrn Dn (cm) Dn (cm) 0.4222 3.73 24.450 4.28 46.735 3.79 0.535 14.27 0.25 Quantity Dn (cm) “Mode” 1 6.913 22.379 21.5 µ = Dn Dny urno = Γnφrn Dn (cm) Dn (cm) 2.395 0.82 1.275 1.200 0.184 0.70 0.551 2.268 0.678 0.252 9.766 7.

84

**Appendix C FEMA Force Distribution Calculations
**

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

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

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

Floor, j

mj (10-3×kg)

s* = j

∑i mi

0.112 0.110 0.110 0.110 0.110 0.110 0.110 0.110 0.119

mj

1 2 3 4 5 6 7 8 9

503.5 494.7 494.7 494.7 494.7 494.7 494.7 494.7 534.1

C.2

EQUIVALENT LATERAL FORCE (ELF) DISTRIBUTION 85

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

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

Floor j 1 2 3 4 5 6 7 8 9

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

-3 k

s* j

=

m jhk j

∑i mihik

0.007 0.020 0.038 0.062 0.091 0.126 0.165 0.210 0.281

C.3

SRSS DISTRIBUTION

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

N

∑ n (V jn )

2

to get column 8 of

8 430.4 1842.5 320.2 105.9 1683.070 0.7 (10) 0.1 -967.9 1446.3.4 400.042 0.2 285.9 366.6 366.090 0.2 95.7 374.7 Vj (kN) (8) 2268.7 2065.0 381.0 176.045 0.6 286.367 87 . the lateral forces are normalized by the base shear to obtain column 10 in Table C.6 -352.0 136.4 1759.3 -646.9 446.0 354.8 1381.0 1476.098 0.2 148.6 319.177 0.8 374.5 215.7 -46.1 Story Shears V j2 (kN) (6) 1114.9 -153.7 1622.3.3 -6.7 95.5 -973.2 fj = ∑ fi i fj (2) (3) (4) 59.0 980.9 832.0 -5.6 V j3 (kN) (7) 478.7 355.2 277.9 -166.5 159.7 101. Table C.7 694.2 Lateral Force fj (kN) (9) 203.1 1857.5 -350.1 -525.8 -326.2 97.0 1231.065 0. 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) 1917.7 234.1 832.6 -732.7 525.1 87.047 0.6 -359.3 240.5 -320.9 880.3 222.For convenience.6 1233.1 -438.2 200.4 250.7 1578.1 -646.