EARTHQUAKE ENGINEERING AND STRUCTURAL DYNAMICS Earthquake Engng Struct. Dyn.

2007; 36:19151934 Published online 13 June 2007 in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/eqe.711

Inuence of structural capacity uncertainty on seismic reliability of buildings under narrow-band motions
M. A. Montiel1 and S. E. Ruiz,
1 Instituto

de Ingeniera, Universidad Nacional Aut noma de M xico, Coyoac n, 04510 M xico, D.F., M xico o e a e e

SUMMARY The inuence of parameter uncertainties on the seismic reliability of several structural reinforced concrete buildings is analysed. Drift hazard curves of buildings considering mean values of their material properties and of instantaneous live loads are compared with those considering uncertain values. Inuence of uncertainty in the storey drift capacity and, alternatively, in intensity capacity on the structural annual failure rates is evaluated. The structural capacity is estimated by incremental dynamic analysis (IDA). It is shown that for the analysis of structures that soften into a peak of their corresponding spectra associated with narrow-band motions it could be more adequate to use the peak storey drift as a measure of structural capacity, rather than the intensity that causes the failure. The study makes emphasis on the importance of the acceleration spectral shape on the reliability of structures. Copyright q 2007 John Wiley & Sons, Ltd.
Received 4 April 2004; Revised 26 August 2004; Accepted 22 March 2007 KEY WORDS:

structural reliability; structural softening; narrow-band motions; uncertainty in structural capacity; failure intensity

1. INTRODUCTION The seismic reliability implicit in a structure is normally inuenced by uncertain parameters that are related with both, the seismic demand and the structural capacity [1, 2]. The uncertain parameters are associated with the following variables, among others: (a) the material properties of which the structure is made and the mechanical properties of the structural elements, (b) the analytical algorithms used for the response and the probabilistic analyses, (c) the stochastic loads acting on

Correspondence

to: S. E. Ruiz, Instituto de Ingeniera, Universidad Nacional Aut noma de M xico, Coyoac n, o e a 04510 M xico, D.F., M xico. e e E-mail: sruizg@iingen.unam.mx

Contract/grant sponsor: DGAPA-UNAM; contract/grant number: IN-106205

Copyright q

2007 John Wiley & Sons, Ltd.

1916

M. A. MONTIEL AND S. E. RUIZ

the structure (i.e. live, seismic, wind loads, etc.), and (d) other uncertainties associated with the randomness of the processes. The inuence of the uncertainty implicit in some of the parameters listed above on the structural reliability has been analysed by several authors. Some of them were involved in the 2000 SAC/FEMA project [3]. The present paper is focused on evaluating the inuence of the uncertainty of some parameters mentioned above on the reliability of reinforced concrete moment-frame buildings subjected to seismic loads. Special attention is given to the inuence of the structural capacity on the structural reliability. For this purpose, three structures (5-, 10- and 15-storey, three-bay buildings) located on soft soil in Mexico City are analysed under the action of narrow-band motions.

2. OBJECTIVES The particular objectives of the study are: (1) To analyse the inuence of the uncertainties related to material properties and to instantaneous live loads on the structural demand hazard curves. (2) To evaluate the inuence related to structural capacity uncertainty on the annual failure rates of several building structures. The inuence of both displacement-based and intensity-based capacities is analysed. The uncertainty in the structural capacity is calculated by means of incremental dynamic analysis (IDA). The results obtained by means of this approach are compared with those obtained from static nonlinear (push-over) analysis. (3) To underline the importance of the acceleration spectral shape on the reliability of structures evaluation.

3. CHARACTERISTICS OF THE STRUCTURES AND OF THE GROUND MOTIONS 3.1. Structures analysed In order to reach the objectives mentioned before three reinforced concrete ofce buildings are analysed. The buildings were designed in accordance with the last version of Mexico City Design Regulations (RCDF-2004) [4] by three well recognized Mexican engineering rms. The geometric properties of the structures are shown in Figure 1(a). The mean values of the fundamental periods of vibration (T0 ) of the 5-, 10- and 15-storey structural frames are equal to 0.67, 1.17 and 1.65 s, respectively; and their yield strength coefcients (Cs ) are 0.34, 0.4 and 0.17, respectively [5]. Each building is composed of exterior and interior structural frames. The dynamic interaction between these was taken into account by means of two-dimensional structural models in which the exterior and the interior frames are connected by hinged links (see Figure 1(b)). The frames were constituted by exural beams and columns. The momentrotation ratios for each element were calculated assuming the model for conned concrete originally proposed by Kent and Park [6] and modied by Park et al. [7]. The axial stressstrain ratios corresponding to the steel bars were represented by means of Mander model [8]. For example, the dimensions, bending yield moments (My ) and yield ( y ) and ultimate ( u ) rotations of the beams corresponding to the interior and
Copyright q 2007 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. 2007; 36:19151934 DOI: 10.1002/eqe

INFLUENCE OF STRUCTURAL CAPACITY UNCERTAINTY

1917

Interior Frame Exterior Frame 3@10m 3@10m 3@10m 15 @ 4m 10 @ 4m 5 @ 4m (a) To=0.67s Interior frame To=1.17s 1E+01 1E+00 Variable number of stories (1/year) 1E-01 1E-02 1E-03 1E-02 (b) (c) To=0.67 s To=1.17 s To=1.65 s To=1.65s

3@10m

Exterior frame LINKS

1E-01

1E+00 S (m/s2)

1E+01

1E+02

Figure 1. (a) Plan and elevation of the buildings; (b) two-dimensional structural model; and (c) seismic hazard curves for the SCT site.

exterior frames of the 15-storey building are shown in Table I. The hysteretic structural behaviour was assumed bilinear with the ratio of the post-yielding to the initial stiffness equal to 3.0%. 3.2. Ground motions and seismic hazard curves The structures were subjected to ground motions recorded at the Ministry of Communications and Transportation station (SCT), located on soft soil in Mexico City. The corresponding seismic hazard curves (associated with T0 = 0.67, 1.17 and 1.65 s) at the SCT site are shown in Figure 1(c) [9]. The ground motions have dominant periods between 1.5 and 2.2 s and correspond to subduction events with magnitude M 5.6. Table II shows the characteristics of the seismic events. The table indicates (with small numbers in parentheses) the Bin II motions that are described later in this paper. The response spectra corresponding to 5% of critical damping are shown in Figure 2(a), which presents logarithmic vertical axis. The gure gives an idea about the scaling factors used in the analysis.
Copyright q 2007 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. 2007; 36:19151934 DOI: 10.1002/eqe

1918

M. A. MONTIEL AND S. E. RUIZ

Table I. Properties of the beams of the exterior and interior frames of the 15-storey building.
(+) Stories Dimensions (m) My (Ton m) 119.05 119.04 71.08 71.11 514.59 296.68 196.05 237.84
y (rad) u (rad)

() My (Ton m) 169.86 147.17 131.17 93.92 561.99 329.55 321.28 321.99

y (rad) u (rad)

Interior frame 16 0.30 1.0 69 0.30 1.0 912 0.30 1.0 1215 0.30 1.0 Exterior frame 16 0.40 1.60 69 0.40 1.60 912 0.40 1.60 1215 0.40 1.60

0.005 0.005 0.005 0.005 0.002 0.002 0.002 0.004

0.0554 0.0554 0.0554 0.0554 0.0505 0.0505 0.0505 0.0505

0.00667 0.00667 0.00667 0.00667 0.00267 0.00267 0.00200 0.00267

0.0605 0.0605 0.0605 0.0594 0.05371 0.0529 0.0505 0.0505

Table II. Characteristics of the seismic events.

Epicentral coordinates Date and component 19-Sep-85 EW 19-Sep-85 NS 25-Apr-89 EW 25-Apr-89 NS 24-Oct-93 EW 24-Oct-93 NS 10-Dec-94 EW 10-Dec-94 NS 22-May-97 EW 22-May-97 NS 31-May-90 EW 31-May-90 NS 23-May-94 EW 23-May-94 NS Magnitude, M 8.1 8.1 6.9 6.9 6.6 6.6 6.3 6.3 6.0 6.0 5.9 5.9 5.6 5.6 Epicentral distance (km) 441.69 441.69 308.16 308.16 314.37 314.37 305.39 305.39 312.25 312.25 316.60 316.60 216.75 216.75 Dominant period Ts (s) 2.03 2.05 2.07 2.19 1.83 1.73 1.87 2.02 1.90 1.50 1.98 1.50 1.60 1.65

Event 1 2 3 4(1,2,3) 5 6(3) 7 8(2) 9(2) 10(1,2,3) 11(3) 12(2,3) 13(3) 14(1,2,3)
(1) (2) (3)

Latitude N 18.081 18.081 16.603 16.603 16.540 16.540 18.020 18.020 18.410 18.410 17.106 17.106 18.030 18.030

Long W 102.942 102.942 99.400 99.400 98.980 98.980 101.560 101.560 101.810 101.810 100.893 100.893 100.570 100.570

Bin II, 5-storey building. Bin II, 10-storey building. Bin II, 15-storey building.

Figures 2(b)(d) show the response spectra of the motions scaled to the same spectral acceleration level [10] that correspond to a return period TR = 50 years (from Figure 1(c)) associated with the corresponding seismic hazard curve. Figures 2(b)(d) also show with a thick black line the arithmetic mean values of the 14-scaled spectra. In these gures, the Bin II motions are indicated inside an oval. The inuence of the
Copyright q 2007 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. 2007; 36:19151934 DOI: 10.1002/eqe

INFLUENCE OF STRUCTURAL CAPACITY UNCERTAINTY

1919

1E+01

1E+00

S19se85.ew S25ab89.ew S24oc93.ew S10di94.ew S22my97.ew S31my90.ew S23my94.ew

S19se85.ns S25ab89.ns S24oc93.ns S10di94.ns S22my97.ns S31my90.ns S23my94.ns Sa(g)

1.2 1.1 1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0 0.0 (b) 1.2 1.1 1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0 0.0 (d)

Mean

Sa(g)

1E-01

1E-02

Bin II 0.5 1.0 1.5 T(s) 2.0 2.5 3.0

1E-03 0.0 (a) 1.3 1.2 1.1 1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0 0.0

0.5

1.0

1.5 T(s)

2.0

2.5

3.0

Mean

Mean

Sa(g)

Bin II

Sa(g)

Bin II 0.5 1.0 1.5 T (s) 2.0 2.5 3.0

0.5

1.0

1.5 T (s)

2.0

2.5

3.0

(c)

Figure 2. Pseudo-acceleration spectra of the ground motions. = 5%, TR = 50 years: (a) without scaling; (b) T0 = 0.67 s; (c) T0 = 1.17 s; and (d) T0 = 1.65 s.

motions contained in Bin II on the conclusions about structural capacity is studied at the end section of this paper.

4. EVALUATION OF THE STRUCTURAL RELIABILITY 4.1. Description of the method used for the analyses Here, the structural reliability is represented by means of demand hazard curves D (d) and annual failure rates F of the multi-storey frames. The former represents the annual rates of exceeding a certain storey drift value (d), and the latter represents the mean annual failure rate of the structures ( F ), for different limit states.
Copyright q 2007 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. 2007; 36:19151934 DOI: 10.1002/eqe

1920

M. A. MONTIEL AND S. E. RUIZ

Here, we used the numerical integration approach. It establishes that the seismic demand hazard curve ( D (d)) for a given structure can be obtained by means of the integral [11, 12]
D (d) = 0

d (y) P(D d|y)dy dy

(1)

where d (y)/dy represents the derivative of the seismic hazard curve, y is the seismic intensity, and P(D d|y) is the conditional probability that the peak drift D over the multiple stories in the building exceeds a given value d, given an intensity level y. In this study, the intensity y is dened as the ordinate of the scaled spectrum (Sa /g) associated with the fundamental period of vibration of the structure (T0 ), and the demand is the peak storey drift over the multiple stories in the building. The structural failure rate value F is calculated by means of the following expression, which takes into account the uncertainty in the structural capacity C [13]:
F= 0

D (d)

dd

P(C d) dd

(2)

Here, P(C d) represents the probability that the storey drift capacity C is smaller than or equal to a given value, d. Equations (1) and (2) were solved by means of the trapezoidal integration technique. The probability distribution function (pdf) of the peak drift for a given intensity was assumed as lognormal. 4.2. Mean values and standard deviations of the logarithm of the peak storey drift as functions of the seismic intensity The three buildings shown in Figure 1(a) were excited with the 14 ground motions mentioned in Section 3.2. These were scaled so that the pseudo-acceleration spectral ordinate (Sa /g, where g = gravity) associated with the structural fundamental period of vibration (T0 ) corresponds to a given return period, TR [10]. The results mentioned in this section correspond to structures material properties and live loads taken equal to their mean values. The median values ( D) and standard deviations of the logarithm of the peak storey drifts ( ln D ) as functions of the seismic intensity Sa /g are presented in Figures 3(a) and (b), respectively. Other intermediate results can be found in [5]. Figures 3(a) and (b) show that D and ln D values grow with the seismic intensity (Sa /g); however, the increment rates of D and ln D corresponding to the 5- and to the 10-storey buildings are much higher than that corresponding to the 15-storey structure. (It is noticed that P effects were not included in the present study. In such a case, some of the D versus Sa /g curves shown in Figure 3 would increase more rapidly, and the standard deviations of ln D would be larger). In what follows a brief explanation is presented about the structural behaviour of each frame as a function of Sa /g. (a) 5-storey building: This building has a vibration period T0 = 0.67 s and lateral resistant seismic coefcient Cs = 0.34. It was excited with ground motions with response spectra similar to those shown in Figure 2(b). As the seismic intensity increases, the structure presents inelastic nonlinear behaviour and its effective vibration period becomes longer, which implies that it becomes closer to the peak of the spectrum. As a consequence, the structural response grows very rapidly. In other words, the structure is softening into the peak of the spectrum, as shown in Figure 3.
Copyright q 2007 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. 2007; 36:19151934 DOI: 10.1002/eqe

INFLUENCE OF STRUCTURAL CAPACITY UNCERTAINTY

1921
5-story frame 10-story frame 15-story frame

0.027 0.024 0.021 Median value, D 0.018 0.015 0.012 0.009 0.006 0.003 0.000 0.0 (a) 0.2 0.4 0.6 0.8

5-story frame 10-story frame 15-story frame Standard deviation, ln D

0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1

1.0

1.2

1.4 (b)

0.0 0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

S /g

S /g

Figure 3. (a) Median values, D and (b) standard deviations ln D of the logarithm of peak drift over the multiple stories in the buildings.

(b) 10-storey building: This building has a vibration period T0 = 1.17 s and a lateral resistant seismic coefcient Cs = 0.40. The structure was excited with ground motions whose linear response spectra have characteristics similar to those shown in Figure 2(c). The nonlinear behaviour of this structure as a function of the seismic intensity is similar to that of the 5-storey building. Figure 3(a) shows that the median drift D of the structure subjected to very small Sa /g values (<0.4) results smaller for the 5-storey than for the 10-storey frame. This is because the spectral displacement (in the near linear response range) is smaller for T0 = 0.67 s than for T0 = 1.17 s. (c) 15-storey building: This building has a vibration period T0 = 1.65 s and lateral resistant seismic coefcient Cs = 0.17. It was excited with ground motions with response spectra sim ilar to those shown in Figure 2(d). Figure 3(a) shows that the median values of D associated with Sa /g up to 0.4 are larger for the 15-storey building than for the other two structures. This is because the set of ground motions used gives place to spectral displacements (in the near linear response range) with higher ordinates for a system with a fundamental period of vibration equal to T0 = 1.65 s than for the other systems, which have shorter vibration periods; however, for higher intensities (for example: Sa /g 0.8) the response of the 15-storey building is smaller than those of the others (because this building does not soften into the peak of the spectrum). In order to verify the inuence of the seismic intensity on the structural behaviour of the three buildings mentioned above, Figure 4 presents the hysteretic behaviour (seismic coefcient Cs versus storey drift) of the corresponding critical stories of the buildings under the SCT-1985 ground motion scaled to an intensity Sa = 0.4g (hysteretic curves at the left side in Figure 4) and to Sa = 0.6g (curves at the right side). The ratios between the secant stiffness and the initial stiffness of the systems shown at the right side in Figure 4 are equal to 10.7, 4.3 and 3 for the 5-, 10- and 15-storey buildings, respectively. This indicates that the 5-storey building present a softening behaviour more pronounced than the other two structures.
Copyright q 2007 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. 2007; 36:19151934 DOI: 10.1002/eqe

1922
0.4 0.3 0.2 0.1 Cs -0.04 -0.02 0 0.00 -0.1 -0.2 -0.3 -0.4 (a)

M. A. MONTIEL AND S. E. RUIZ

0.4 0.3 0.2 0.1 Cs 0.02 0.04 -0.04 -0.02 0 0.00 -0.1 -0.2 -0.3 -0.4 Second story drift 0.5 0.4 0.3 0.2 0.1 0 -0.02 -0.1 0 0.02 -0.2 -0.3 -0.4 -0.5 Third story drift 0.02 0.04

Second story drift 0.5 0.4 0.3 0.2 0.1 0 -0.02 -0.1 0 -0.2 -0.3 -0.4 -0.5 Third story drift

Cs

Cs -0.04

-0.04

0.02

0.04

0.04

(b)

0.25 0.20 0.15 0.10 0.05 0.00 -0.05 -0.0151 -0.01 -0.005 0 -0.10 -0.15 -0.20 -0.25

0.005 0.01 0.015

0.25 0.20 0.15 0.10 0.05 0.00 -0.015 -0.01 -0.005 0 -0.05 -0.20 -0.10 -0.15 -0.20 -0.25

Cs

Cs

0.005 0.01 0.015

(c)

Fifth story drift

Fifth story drift

Figure 4. Hysteretic behaviour curves. Left side: SCT ground motion scaled to 0.4g. Right side: SCT ground motion scaled to 0.6g: (a) 5-storey frame; (b) 10-storey frame; and (c) 15-storey frame.

The analysis performed above indicates that the narrow bandness of the motions has a signicant inuence on the D versus Sa /g curves (as shown in Figures 3 and 4) and as a consequence, on the structural reliability. On the other hand, Figure 3(b) indicates that the standard deviations of the logarithm of D are very small (0.03< ln D <0.06) when the structures are subjected to low seismic intensities (Sa /g<0.3), and the structural behaviour is linear. Those standard deviations are larger (0.06< ln D <0.3) when the buildings are subjected to seismic intensities Sa /g between 0.3 and
Copyright q 2007 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. 2007; 36:19151934 DOI: 10.1002/eqe

INFLUENCE OF STRUCTURAL CAPACITY UNCERTAINTY

1923

0.3) when the structures soften into the peak 0.5 approximately; and are much larger ( ln D of the spectrum, as happens with the 5- and the 10-storey buildings. 5. INFLUENCE OF THE UNCERTAINTY IN MECHANICAL PROPERTIES AND LIVE LOADS ON THE DEMAND HAZARD CURVES In order to better understand the inuence on the demand hazard curves of the uncertainty in the mechanical properties and in the instantaneous live loads acting on the reinforced concrete structures shown in Figure 1(a), the following cases were analysed: Case a: structural model using nominal properties. Each building model was dened using the nominal values of concrete strength ( f c ), of steel bars yield stress ( f y ) and of instantaneous live loads acting on the building structures (L). The nominal values were those used for the design of the buildings. Case b: structural model using mean properties. Each building model was dened using the mean values of the parameters: fc , fy and L [1416]. Case c: nine structural models using simulated properties. The frame models were constituted using simulated values of f c , f y and L. For the rst two variables, the Gaussian pdfs were assumed [14, 17], and for the instantaneous live load magnitude, a Gamma distribution was supposed [16, 18]. The nominal, mean values and coefcients of variation used are as follows: f c = 2.5 106 kg/m2 , fc = 2.68 106 kg/m2 , CV f c = 0.17; f y = 4.2 107 kg/m2 , fy = 4.68 107 kg/m2 , CV f y = 0.096; L = 180 kg/m2 , L = 75 kg/m2 , CV L = 0.3. Except that for the 15-storey building a larger concrete strength value was used ( f c = 3.0 106 kg/m2 , fc = 3.21 106 kg/m2 , CV f c = 0.13). Each building was subjected to the action of the 14 time acceleration histories mentioned in Section 3.2. 5.1. Demand hazard curves of the buildings analysed The demand hazard curves ( D (d)) for the three buildings, and for Cases a, b and c, are presented in Figures 5(a)(c). In these, it can be seen that the curves associated with Cases b (mean values) and c (simulated values) are very similar for annual exceedance rates corresponding to small storey drifts (which correspond to small seismic intensities); however, as the storey drifts increase the curves present larger differences among them. It can also be seen from Figure 5 that the annual exceedance rate values corresponding to Case a (nominal parameters) are larger than those corresponding to Cases b and c. From Figure 5, it is concluded that the reliability analysis of the buildings studied here gives place to reasonable approximation when the mean values of the mechanical properties and of the instantaneous live load magnitudes are used; and uncertainties about both parameters can be taken into account by means of correction factors (as suggested by FEMA 351). 6. INFLUENCE OF THE UNCERTAINTY IN THE STRUCTURAL CAPACITY ON THE ANNUAL FAILURE RATES In order to calculate the annual rate of failure of the buildings ( F ) using Equation (2), it is rst necessary to evaluate the uncertainty on the structural capacity (C).
Copyright q 2007 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. 2007; 36:19151934 DOI: 10.1002/eqe

1924

M. A. MONTIEL AND S. E. RUIZ

1.E+00 1.E-01

1.E+00 Case a Case b Case c 1.E-01 Case a Case b Case c

D( d )

1.E-03 1.E-04 1.E-05 0.001

D( d )
1.E-02 1.E-03 0.001 (b)

1.E-02

(a)

0.010 d 1.E+00

0.100

0.010 d

0.100

1.E-01

Case a Case b Case c

D( d )

1.E-02

1.E-03

1.E-04 0.001 (c)

0.010 d

0.100

Figure 5. Demand hazard curves for the three buildings: (a) 5-storey frames; (b) 10-storey frames; and (c) 15-storey frames.

6.1. Incremental dynamic analyses (IDAs) 6.1.1. Description of Bin I and Bin II motions. In order to evaluate the structural capacity of the buildings several IDAs were performed. The algorithm followed was similar to that proposed by Vamvatsikos and Cornell [19]. Each building model was subjected to 14 ground motions. In order to better understand the behaviour of the IDAs results (curves at the right side in Figures 6 and 7) the motions were grouped into two bins. The rst, called Bin I, contains ground motions with a positive epsilon ( ) value [20] ( >0). The second group, called Bin II, contains records with a negative epsilon ( ) value [20] ( <0). Bin I is composed mainly of narrow-band motions, and Bin II of broad-band motions (see Figures 2(b)(d)). IDAs results are shown in Figures 6 and 7 with thin lines, and those corresponding to static nonlinear (push-over, PO) analyses with thick lines. The IDAs curves at the left side are interrupted when the drifts present an extremely large value under a small increment in the motion amplitude [21]. The drifts where the curves are interrupted are the same drifts indicated with circles (capacity points) in the curves at the right side in Figures 6 and 7. The black full circles correspond to Bin I motions, and white circles to Bin II motions.
Copyright q 2007 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. 2007; 36:19151934 DOI: 10.1002/eqe

INFLUENCE OF STRUCTURAL CAPACITY UNCERTAINTY

1925

19sep85EW 19sep85NS 25apr89EW 25apr89NS 24oct93EW 24oct93NS 10dec94EW 10dec94NS 22may97EW 22may97NS 23may94EW 23may94NS 31may90EW 31may90NS Push-over Bin II

0.6 0.5 Cs2 = V2 / W2 0.4 0.3 0.2 0.1 0.0 0.00 S (T1,5%) /(g)

1.6 1.4 1.2 1.0 0.8 0.6 0.4 0.2 0.02 0.04 d2 0.06 0.08

Capacity point, Bin I Capacity point, Bin II Push-over IDAs

0.0 0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 d2

(a)
19sep85.EW 19sep85.NS 25apr89.EW 25apr89.NS 24oct93.EW 24oct93.NS 10dec94.EW 10dec94.NS 22may97.EW 22may97.NS 23may94.EW 23may94.NS 31may90.EW 31may90.NS Push-over Bin II

0.55 0.50 0.45 0.40 0.35 0.30 0.25 0.20 0.15 0.10 0.05 0.00 0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07 d3

2.25 2.00 1.75 S (T1, 5%)/(g) 1.50 1.25 1.00 0.75 0.50 0.25

Capacity point, Bin I Capacity point, Bin II Push-over IDAs

Cs3 =V3 / W3

0.00 0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07 d3

(b)
19sep85.EW 19sep85.NS 25apr89.EW 25apr89.NS 24oct93.EW 24oct93.NS 10dec94.EW 10dec94.NS 22may97.EW 22may97.NS 23may94.EW 23may94.NS 31may90.EW 31may90.NS Push-over Bin II

0.40 0.35 0.30 Cs5 =V5 / W5 0.25 0.20 0.15 0.10 0.05 0.00 0.00 0.02 0.04 d5 0.06 0.08

11 10 9 8 7 6 5 4 3 2 1 0 0.00

Capacity point, Bin I Capacity point, Bin II Push-over IDAs

S (T1, 5%) / (g)

0.02

0.04 d5

0.06

0.08

(c)

Figure 6. Comparison between local results of IDAs and PO analyses: (a) 5-storey frame; (b) 10-storey frame; and (c) 15-storey frame.

Copyright q

2007 John Wiley & Sons, Ltd.

Earthquake Engng Struct. Dyn. 2007; 36:19151934 DOI: 10.1002/eqe

1926
19sep85EW 19sep85NS 25apr89EW 25apr89NS 24oct93EW 24oct93NS 10dec94EW 10dec94NS 22may97EW 22may97NS 23may94EW 23may94NS 31may90EW 31may90NS Push-over Bin II

M. A. MONTIEL AND S. E. RUIZ

0.50 0.45 0.40 0.30 0.25 0.20 0.15 0.10 0.05 0.00 0.00 0.01 0.02 0.03 0.04 0.05 0.06 d roof S (T1,5%) / (g) Cs = Vb / WT 0.35

2.0 1.8 1.6 1.4 1.2 1.0 0.8 0.6 0.4 0.2 0.0 0.00

Capacity point, Bin I Capacity point, Bin II Push-over IDA's

0.01

0.02

0.03 d roof

0.04

0.05

0.06

(a)
19sep85.EW 19sep85.NS 25apr89.EW 25apr89.NS 24oct93.EW 24oct93.NS 10dec94.EW 10dec94.NS 22may97.EW 22may97.NS 23may94.EW 23may94.NS 31may90.EW 31may90.NS Push-over Bin II

0.50 0.45 0.40 0.30 0.25 0.20 0.15 0.10 0.05 0.00 0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07 d roof S (T1,5%) / (g) 0.35 Cs =Vb / WT

Capacity point,Bin I 2.75 Capacity point,Bin II 2.50 Push-over 2.25 IDAs 2.00 1.75 1.50 1.25 1.00 0.75 0.50 0.25 0.00 0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07 d roof

(b)
19sep85.EW 19sep85.NS 25apr89.EW 25apr89.NS 24oct93.EW 24oct93.NS 10dec94.EW 10dec94.NS 22may97.EW 22may97.NS 23may94.EW 23may94.NS 31may90.EW 31may90.NS Push-over BIN II

0.40 0.35 0.30 Cs =Vb/ WT 0.25 0.20 0.15 0.10 0.05 0.00 0.00 0.02 0.04 0.06 d roof 0.08 S (T1,5%) / (g)

12 11 10 9 8 7 6 5 4 3 2 1 0 0.00

Capacitypoint, Bin I Capacitypoint, Bin II Push-over IDAs

0.02

0.04 d roof

0.06

0.08

(c)

Figure 7. Comparison between global results of IDAs and PO analyses: (a) 5-storey frame; (b) 10-storey frame; and (c) 15-storey frame.

Copyright q

2007 John Wiley & Sons, Ltd.

Earthquake Engng Struct. Dyn. 2007; 36:19151934 DOI: 10.1002/eqe

INFLUENCE OF STRUCTURAL CAPACITY UNCERTAINTY

1927

6.1.2. Analysis of IDAs results presented at the right side in Figures 6 and 7. The IDAs results are shown in Figures 6(a)(c) (corresponding to storey response), and 7(a)(c) (corresponding to global response), for the 5-, 10- and 15-storey buildings, respectively. The graphs at the right side in Figure 6 are the 5% damped spectral acceleration at the structures rst-mode period (Sa (T1 , 5%)) divided by the acceleration of gravity (g) versus the peak storey drift over the multiple stories. The maximum value of the peak storey drift occurs at the second and at the third storey for the 5- and 10-storey buildings, respectively. For the 15-storey building, the maximum values of the peak storey drift occur at different stories. Figure 6(c) shows the response corresponding to the fth storey (where the maximum drift occurs more frequently). The IDAs curves associated with the 5- and 10-storey buildings subjected to Bin I motions (curves with smaller ordinates, in general, at the right side in Figures 6 and 7) reect the fact that the structures soften into the peak of the 14 spectra (as explained in Section 4.2), then recover their stiffness, and nally become at. It can also be seen in Figures 6 and 7 (at the right side) that for Bin II records the structures fail, in general, for larger Sa (T1 , 5%) values. These are larger because they are associated with higher motion scaling factors (see Figure 2(a)). Bin II motions correspond to accelerograms (with < 0) that do not have a large spectral peak ordinate located at the right side of the structural vibration period T0 , and consequently, the structure is not affected by a structural softening behaviour. From this, the strong inuence of the spectral shape on the structural behaviour of the buildings becomes obvious. On the other hand, the IDAs curves associated with the 15-storey building (Figures 6(c) and 7(c) at the right side) subjected to Bin I motions differ from those corresponding to the 5and the 10-storey buildings (Figures 6(a) and (b), and 7(a) and (b)). The curves are different because the initial vibration period of the 15-storey structure is close to the dominant period of the ground motion (see Figure 2(d)) and contrary to what happens with the 5- and the 10-storey structures, the 15-storey building does not soften into a spectral peak. In this case, the Sa (T1 , 5%) values grow almost linearly as the storey drift increases (see Figures 6(c) and 7(c)). This happens with the curves corresponding to Bin I as well as to Bin II motions, but in the latter case (Bin II motions) larger values of scaling motion factors are needed. The points where the IDAs curves become at (at the right side in Figures 6 and 7) correspond to the spectral acceleration level at the elastic-rst-mode frequency of the building required to induce near-collapse to the structure. That intensity level is called here failure intensity, and is denoted as (Sa near-collapse /g)storey or (Sa near-collapse /g)global . The rst corresponds to the storey response, and the latter to the global one. 6.1.3. Analysis of IDAs results shown at the left side in Figures 6 and 7. The graphs at the left side in Figure 6 represent the storey seismic shear ratio (Csi ) versus the peak storey drift over the multiple stories in the building. The seismic coefcient (Csi ) is the ratio of the structural shear force (Vi ) to the weight of the structure (Wi ) above the ith storey (where the maximum drift occurred). The graphs at left side in Figure 7 represent the base shear ratio (Cs ) versus the peak roof drift. Cs is the ratio of the structural shear base force (Vb ) to the total weight of the structure (WT ). The white circles correspond to the capacity points (Bin II) indicated also with white circles (Sa near-collapse /g values) at the right side in Figures 6 and 7. It can be noticed that the results at the left side in Figures 6 and 7 are less sensitive to the frequency content of the motions than those presented at the right side in the same gures.
Copyright q 2007 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. 2007; 36:19151934 DOI: 10.1002/eqe

1928

M. A. MONTIEL AND S. E. RUIZ

The curves at the left side in Figure 6(c) are very irregular because those curves correspond to the fth storey peak drifts; however, the maximum value of the peak drift over the stories does not occur always at the fth storey, but at different stories.

6.2. Comparison of IDAs curves with those obtained from static nonlinear (PO) analysis In order to see the inuence of the dynamic effects on the structural capacity evaluation, the three buildings were also analysed by means of a static nonlinear (push-over) analyses (PO) [22]. For this purpose, a modied version of the DRAIN-2D program [23] was used. The two-dimensional analysis is performed by subjecting the building to a linearly increasing base acceleration, up to the structural failure. The response is obtained by means of a step-by-step analysis in time. The results of the analyses (represented with thick black lines in the graphs at the left side in Figures 6 and 7) indicate that when the ultimate capacity occurs, there is a sudden drop on the curves. After this drop, a recovery of the structure can take place. In this study, the P- effects were not taken into account. Graphs at the left side in Figure 6 show that the seismic storey resistant coefcients (Csi ) and the storey drift capacities obtained by means of PO analysis are in general equal to or smaller than those obtained by means of IDAs, from which it is concluded that the PO analysis results are on the safe side. Curves in Figures 7(a) and (b) (at the left) show that PO analyses give place to slightly larger lateral load resistance values. It is noticed that the denition of the structural capacity (global or local) is very debatable because it depends on the criteria used to select the capacity point in each curve. This introduces an uncertainty that should be taken into account in the structural analysis, especially for the displacement-based capacity format.

6.3. Coefcients of variation of some parameters The coefcients of variation (COVs) of several parameters corresponding to storey and to global properties were calculated from the information presented in Figures 6 and 7. The coefcients are shown in Table III. The parameters are: structural stiffness (K storey and K global ), yield displacement (dystorey and dyglobal ), peak storey drift and roof drift (dmax storey and droof ), ductility capacity ( storey and global ) and the failure intensity ((Sa near-collapse /g)storey and (Sa near-collapse /g)global ). The coefcients of variation shown in Table III correspond to structures excited with Bin motions IIII. The latter is composed of Bin I plus Bin II motions. It was found that the COVs of the parameters associated with the storey response are in general smaller than those associated with the global one. For example, considering the COV corresponding to Bins I and II, the storey stiffness ranges between 1.0 and 5.5%; however, for the global stiffness it is larger (5.37.3%). The COV of the storey yield drift lies between 1.9 and 7.5%, and that of the global yield drift is also larger (3.99.11%). The COV of the storey ductility capacity is between 13.6% and 34.6%, and that of the global ductility is larger as well (10.151.0%). The COV of the storey drift capacity lies between 15.3 and 33.4%; however, that associated with the peak roof drift is between 9.1 and 47.9%. The COV of the failure intensity (Sa near-collapse /g) associated with the storey response is between 21.7 and 37.0% and that corresponding to the global response lies between 22.7 and 37.1%.
Copyright q 2007 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. 2007; 36:19151934 DOI: 10.1002/eqe

INFLUENCE OF STRUCTURAL CAPACITY UNCERTAINTY

1929

Table III. Coefcients of variation of some parameters of interest: Storey and global response.
Bin Storey response VK storey Vdy storey Vd max storey V storey V(Sa
near-collapse /g)

5-storey frame (second storey) 0.022 0.010 0.026 0.042 0.019 0.050 0.153 0.334 0.200 0.136 0.346 0.214 0.217 0.308 0.278 5-storey frame 0.071 0.0711 0.069 0.091 0.0911 0.102 0.159 0.1599 0.184 0.168 0.1680 0.212 0.298 0.2982 0.327

10-storey frame (third storey) 0.036 0.019 0.028 0.041 0.025 0.034 0.184 0.322 0.278 0.203 0.312 0.279 0.329 0.313 0.375 10-storey frame 0.053 0.056 0.058 0.041 0.039 0.039 0.403 0.479 0.426 0.418 0.510 0.446 0.346 0.295 0.366

15-storey frame (fth storey) 0.055 0.045 0.048 0.075 0.060 0.066 0.261 0.168 0.272 0.258 0.196 0.281 0.275 0.370 0.508 15-storey frame 0.073 0.059 0.074 0.069 0.051 0.063 0.091 0.321 0.310 0.101 0.347 0.323 0.227 0.371 0.531

storey

I II III I II III I II III I II III I II III Bin

Global response VK global Vdy global Vdroof V global V(Sa

near-collapse /g)

global

I II III I II III I II III I II III I II III

Some authors [24] report coefcients of variation of the global displacement capacity (Vdroof ) of 18 and 30% approximately, which are of the order of those found in the present study. It is noticed that the COVs of the storey drift capacity (Vd max storey ) of the 5- and of the 10-storey buildings are smaller than those corresponding to the failure intensity (V(Sa near-collapse /g)storey ) when the structures are subjected to Bin I motions (which are in general narrow-banded); however, when the frames are subjected to Bin II motions the contrary occurs (Vd max storey >V(Sa near-collapse /g)storey ). The latter inequality is usually found in the literature [10]; however, those studies deal mainly with structures subjected to broad-band motions (recorded on hard rock) and not with structures that soften into the peak of response spectra corresponding to narrow-band motions (as those contained in Bin I).
Copyright q 2007 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. 2007; 36:19151934 DOI: 10.1002/eqe

1930

M. A. MONTIEL AND S. E. RUIZ

Table IV. Median values and standard deviations of the logarithm of the structural capacities associated with two limit states (Bin III motions).
Frame (1) Storey (2) Cy (3) ln Cy C near-collapse (4) (5) 0.045 0.040 0.052 ln Cnear-collapse Say /g (6) 0.18 0.26 0.29 (7) 0.28 0.40 0.18 ln Say /g Sa near-collapse /g (8) 0.027 0.008 0.017 (9) 0.86 1.13 4.02
ln Sa
near-collapse /g

(10) 0.28 0.38 0.48

5-storey 2nd 0.0037 0.051 10-storey 3rd 0.0072 0.034 15-storey 5th 0.0045 0.065

6.4. Structural performance levels represented on IDAs curves Some authors have suggested different points in the IDAs curves such that each represents a limit to a certain structural performance level [25]. Here, two specic values of structural capacities were selected as performance levels. These are the following: (a) The yielding drift value for the storey of interest (dy = y /h): This value is dened at the intersection of the initial stiffness and the post-yielding stiffness of the structure. The median values (Cy ) and the standard deviation of the logarithm ( ln Cy ) of the storey yielding drift (based on the curves shown at the left side in Figures 6(a)(c)) are shown in the third and fourth columns in Table IV, respectively (the values correspond to structures under Bin III motions). Alternatively, the structural capacity can be dened as the spectral acceleration level at the elastic-rst-mode frequency of the structure required to induce yielding in the structure. The median values ( Say /g) and the standard deviations of the logarithm ( ln Say /g ) of that spectral acceleration are shown in the seventh and eighth columns in Table IV, respectively. (b) The peak drift associated with the near-collapse limit state (dnear-collapse = near-collapse / h): This corresponds to the capacity points indicated with circles on the thin lines of the curves at the right side in Figures 6(a)(c). This is a transition point (capacity point) beyond which there is a large increase in storey drift. This is the threshold of the near-collapse state [24] (see fth and sixth columns in Table IV). In a similar way, the structural capacity associated with the near-collapse limit state can be expressed by the failure intensity (Sa near-collapse /g), de ned in Section 6.1.2. The median values ( Sa near-collapse /g) and the standard deviations of the logarithm ( ln Sa near-collapse /g ) of that spectral acceleration are shown in the last two columns in Table IV. Table IV shows that for the three buildings, the standard deviations of the logarithms of the structural capacities are smaller for the yield drift performance level ( ln Cy ) than for the nearcollapse state ( ln Cnear-collapse ). Table IV also shows that the standard deviations of the logarithms of the values associated with the yield capacity level are smaller for the intensity-based approach ( ln Say /g ) than for the displacement-based approach ( ln Cy ) (see columns 8 and 4); however, the opposite occurs for the near-collapse performance level ( ln Sa near-collapse /g > ln Cnear-collpase ) (see columns 10 and 6). 6.5. Inuence of the uncertainty in the storey drift capacity on the structural failure rates In order to quantify the inuence of the uncertainty in the structural capacity on the structural annual failure rates, the F values associated with the two structural performance levels were calculated for each building. The numerical integration method (Equations (1) and (2)) were used
Copyright q 2007 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. 2007; 36:19151934 DOI: 10.1002/eqe

INFLUENCE OF STRUCTURAL CAPACITY UNCERTAINTY

1931

1E+0 D(d) 1E-1 F v D (d), v F 1E-2 1E-3 1E-4 1E-5 0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07 (a) d,C 1E+0 1E-1 v D (d), v F 1E-2 1E-3 1E-4 (b) v D (d), v F

1E+0 D(d) 1E-1 F

1E-2

1E-3

1E-4 0.00

0.01

0.02 d,C

0.03

0.04

D(d) F

1E-5 0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07 (c) d,C

Figure 8. Mean annual failure rates and demand hazard curves for the three frames: (a) 5-storey frame; (b) 10-storey frame; and (c) 15-storey frame.

for the analysis. From Table IV it can be anticipated that the inuence of the uncertainty in the peak storey drift capacity is equally important for the three buildings, and it is more important for the near-collapse performance level than for the yield drift level. Figures 8(a)(c) show a comparison of the demand hazard curves ( D (d), indicated with white circles) and the annual failure rates F (indicated with triangles) for the three buildings (subjected to Bin III motions). The two full triangles appearing in each gure from left to right are associated with the drift capacities Cy and Cnear-collapse , respectively. From Figure 8 it is concluded that the inuence of the uncertainty in the storey drift capacity on the structural reliability is not signicant for the yield structural performance; however, it becomes more important for the near-collapse limit state (see Figure 8(c)). It is noticed that for the latter case, the standard deviations ln Cnear-collapse are equal to 0.18, 0.26 and 0.29 for the 5-, 10- and 15-storey buildings (see Table IV, column 6), respectively; therefore, the inuence of that uncertainty is more signicant for the 15-storey building because the standard deviation is the largest ( ln Cnear-collapse = 0.29).
Copyright q 2007 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. 2007; 36:19151934 DOI: 10.1002/eqe

1932

M. A. MONTIEL AND S. E. RUIZ

Table V. Mean annual failure rates. Yield limit state and near-collapse limit state. Displacement- and Intensity-based capacities. Bin III.
Yield limit state Equation
F F

Near-collapse limit state

5-storey frame 10-storey frame 15-storey frame 5-storey frame 10-storey frame 15-storey frame 1.95 102 1.47 102 1.11 101 1.16 101 1.12 104 1.93 104 9.32 104 9.88 104 0.81 104 0.84 104

(Equation (2)) 1.42 102 (Equation (3)) 1.58 102

6.6. Inuence of the uncertainty in the failure intensity on the structural reliability The mean annual failure rate F can be calculated assuming that the structural capacity is given by the peak storey drift (Equation (2)), or assuming that the capacity is represented by the failure intensity. In the latter case, F can be calculated as follows [10]:
F= 0

d (y) P(y Sa dy

LState ) dy

(3)

where (y) is the seismic hazard curve as a function of the intensity (y) and Sa LState represents the spectral acceleration level at the elastic-rst-mode frequency of a structure required to induce certain damage level in that structure (which is generally associated with a given limit sate (LState); for example, yielding or near-collapse of the structure). In this study, Sa LState is assumed lognormal. One advantage of the latter formulation (Equation (3)) is that the design and/or assessments are performed in the spectral acceleration ordinates and do not explicitly involve the displacementbased response. Another advantage is that the method is more direct and the required computational effort is lower. Table V show the F values calculated numerically with Equation (2) and, alternatively, with Equation (3) for the yielding and the near-collapse limit states, respectively. The rst part of Table V indicates that for the yield performance condition the failure rates obtained with both equations are similar; however, the second part of the Table V (which corresponds to the near-collapse state) shows that the F values calculated with the displacement-based method (Equation (2)) are slightly smaller than those calculated with the intensity-based method (Equation (3)). This may be due to the fact that the standard deviation of the logarithm of the peak storey drift capacity ( ln Cnear-collapse ) is smaller than that corresponding to the intensity capacity ( ln Sa near-collapse ) (see Table IV, columns 6 and 10). 7. CONCLUSIONS The reliability analyses of the three buildings lead to the following conclusions: 1. For small values of peak storey drifts, the demand hazard curves corresponding to building models with mean values of material properties and live loads are similar to those corresponding to models where the uncertainties about these parameters are included in the analysis (see Figure 5); however, as the peak drift values increase, the differences become more signicant. Those differences can be taken into account by means of correction factors, as in FEMA 351.
Copyright q 2007 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. 2007; 36:19151934 DOI: 10.1002/eqe

INFLUENCE OF STRUCTURAL CAPACITY UNCERTAINTY

1933

2. The buildings were analysed by means of incremental dynamic analysis (Figures 6 and 7). The results were used to calculate coefcients of variation of several parameters associated with storey and with global structural properties and responses. The coefcients of variation of the storey drift capacity of the 5- and 10-storey buildings subjected to Bin I motions ( <0, in general narrow-band motions) are smaller than those corresponding to the failure intensity; however, for Bin II motions, the contrary occurs (see Table III). Due to this, the authors came to the conclusion that for structures that soften into the peak of narrow-band motions it may be more convenient to use as a measure of structural capacity the storey drift rather than the intensity that causes failure. The results show that the standard deviations of the logarithms of the structural capacities ( ln C ) are smaller for the yield drift performance level ( ln Cy ) than for the near-collapse state ( ln Cnear-collapse ) (Table IV). It was also found that the ln C values associated with the yield capacity level are smaller for the intensity-based approach ( ln Say /g ) than for the displacementbased approach ( ln Cy ); however, the opposite occurs for the near-collapse performance level ( ln Sa near-collapse/g > ln Cnear-collapse ). 3. The annual structural failure rates associated with the near-collapse limit state, calculated with the displacement-based capacity approach (Equation (2)), resulted slightly smaller than those obtained with the intensity-based approach (Equation (3)). This is due to the fact that, for the Bin I motions (mainly narrow-band motions) used in this study, the dispersion in the displacement-based capacity resulted smaller than that based on the intensity (see Table III).

ACKNOWLEDGEMENTS

The authors thank L. Esteva for reading the nal manuscript, and R. Granados, F. Garca Jarque and J. Alonso for providing the designs of the three buildings analysed in this study. The support by DGAPAUNAM under Project IN-106205 is appreciated. Thanks are given to Gerardo Rangel for his enthusiastic collaboration.
REFERENCES 1. Cornell CA. Reliability-based earthquake-resistant design: the future. Eleventh World Conference on Earthquake Engineering, Acapulco, Mexico, June 1996; Paper no. 2166. 2. Wen YK. Building reliability and code calibration. Earthquake Spectra 1995; 11:269296. 3. SAC Joint Venture for the Federal Emergency Management Agency. Recommended seismic evaluation and upgrade criteria for existing welded steel moment-frame buildings. FEMA-351. SAC Joint Venture for the Federal Emergency Management Agency, Washington, DC, 2000. 4. Gaceta Ocial. Mexico City Design Regulations. January 2004. 5. Montiel MA. Conabilidad Implcita en Estructuras Convencionales como Base para Establecer Criterios para el Dise o Ssmico de Estructuras Reforzadas con Disipadores de Energa. Ph.D. Thesis, Graduate School of n Engineering, National University of Mexico, 2006 (in Spanish). 6. Kent DC, Park R. Flexural members with conned concrete. Journal of Structural Division (ASCE) 1971; 97(7):19691990. 7. Park R, Priestley MJN, Gill WD. Ductility of square conned concrete columns. Journal of Structural Division (ASCE) 1982; 108(4):929950. 8. Mander J. Seismic design of bridge piers. Report 84-2, Department of Civil Engineering, University of Cantenbury, New Zealand, 1984. 9. Alamilla JL. Personal communication, 2004. 10. Shome N, Cornell CA. Probabilistic seismic demand analysis of nonlinear structures. Report No. RMS-35, Department of Civil Engineering, Stanford University, 1999. Copyright q 2007 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. 2007; 36:19151934 DOI: 10.1002/eqe

1934

M. A. MONTIEL AND S. E. RUIZ

11. Esteva L, Bases para la formulaci n de decisiones de dise o ssmico. Ph.D. Thesis, Graduate School of o n Engineering, National University of Mexico, 1968 (in Spanish). 12. Cornell CA. Engineering seismic risk analysis. Bulletin of Seismological Society of America 1968; 58(5):1583 1606. 13. Cornell CA, Jalayer F, Hamburger RO, Foutch DA. The probabilistic basis for the 2000 SAC/FEMA steel moment frame guidelines. Journal of Structural Engineering (ASCE) 2002; 128(4). 14. Mendoza CJ. Propiedades Mec nicas de los Concretos Fabricados en el Distrito Federal. Research Report, a Institute of Engineering, UNAM, 1983 (in Spanish). 15. Rodrguez M, Botero JC. Aspectos del Comportamiento Ssmico de Estructuras Considerando las Propiedades Mec nicas del Acero Producido en M xico. Research Report, Institute of Engineering, UNAM, Mexico, 1994 a e (in Spanish). 16. Ruiz SE, Soriano A. Design live loads for ofce buildings in Mexico and the United States. Journal of Structural Engineering 1997; 123(6):816822. 17. Mirza A, McGregor J. Variability of mechanical properties of reinforcing bars. Journal of the Structural Division (ASCE), 1979; 105(ST5):921937. 18. Peir J, Cornell CA. Spatial, temporal variability of live loads. Journal of the Structural Division (ASCE) 1973 5:903922. 19. Vamvatsikos D, Cornell CA. Incremental dynamic analysis. Earthquake Engineering and Structural Dynamics 2002; 31:491514. 20. Baker JW, Cornell CA. A vector valued ground motion intensity measure consisting of spectral acceleration and epsilon. Earthquake Engineering and Structural Dynamics 2005; 34:11931217. 21. Wen YK, Foutch DA. Proposed statistical and reliability framework for comparing and evaluating predictive models for evaluation and design, and critical issues in developing such framework. Report No. SAC/BD-97/03, 1997. 22. Krawinkler H, Seneviratna GDPK. Pros and cons of a pushover analysis of seismic performance evaluation. Engineering Structures 1998 20(46):452464. 23. Campos D, Esteva L. Modelo de Comportamiento Hister tico y de Da o para Vigas de Concreto Reforzado. XI e n CongresoNacional de Ingeniera Ssmica, Veracruz, Mexico, 1997 (in Spanish). 24. Wen YK, Ellingwood BR, Bracci J. Vulnerability function framework for consequence-based engineering. MAE Center Project DS-4 Report, April 2004. 25. Vamvatsikos D, Cornell CA. The incremental dynamic analysis and its application to performance-based earthquake engineering. Proceedings of the 12th European Conference on Earthquake Engineering, London, U.K., 2002; Paper no. 479.

Copyright q

2007 John Wiley & Sons, Ltd.

Earthquake Engng Struct. Dyn. 2007; 36:19151934 DOI: 10.1002/eqe