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

6 (2005) 777–804 c Imperial College Press

FRAGILITY ASSESSMENT OF SLAB-COLUMN CONNECTIONS IN EXISTING NON-DUCTILE REINFORCED CONCRETE BUILDINGS

HESAMEDDIN ASLANI and EDUARDO MIRANDA Dept. of Civil and Environmental Engineering Stanford University, Stanford, CA, USA Received 12 November 2004 Reviewed 11 April 2005 Accepted 14 April 2005 Fragility functions that estimate the probability of exceeding diﬀerent levels of damage in slab-column connections of existing non-ductile reinforced concrete buildings subjected to earthquakes are presented. The proposed fragility functions are based on experimental data from 16 investigations conducted in the last 36 years that include a total of 82 specimens. Fragility functions corresponding to four damage states are presented as functions of the level of peak interstory drift imposed on the connection. For damage states involving punching shear failure and loss of vertical carrying capacity, the fragility functions are also a function of the vertical shear in the connection produced by gravity loads normalised by the nominal vertical shear strength in the absence of unbalanced moments. Two sources of uncertainty in the estimation of damage as a function of lateral deformation are studied and discussed. The ﬁrst is the specimen-to-specimen variability of the drifts associated with a damage state, and the second the epistemic uncertainty arising from using small samples of experimental data and from interpreting the experimental results. For a given peak interstorey drift ratio, the proposed fragility curves permit the estimation of the probability of experiencing diﬀerent levels of damage in slab-column connections. Keywords: Performance evaluation; damage assessment; experimental data; uncertainty analysis; slab-column connections.

1. Introduction The goal of performance-based seismic design (PBSD) is to design facilities with predictable levels of seismic performance. An adequate implementation of PBSD requires a relationship between the ground motion intensity and the damage in the structure. Previous studies in damage assessment have typically provided global estimates of damage in the facility in a single step through the use of damage probability matrices relating a damage ratio with Modiﬁed Mercalli Intensity [ATC13, 1985], empirical functions relating peak ground motion parameters to damage ratios [Wald, 1999] and with functions relating damage ratios to spectral ordinates [NIBS, 1999]. More recent work has been aimed at improved estimates of damage in speciﬁc types of reinforced concrete buildings as a function of spectral ordinates.

777

778

H. Aslani & E. Miranda

For example, Singhal and Kiremidjian [1996] computed fragility of three classes of reinforced concrete frame buildings, namely low rise concrete frames that are 1–3 storeys tall, mid rise frames that are 4–7 storeys tall, and high rise frames that are 8 storeys or taller, to estimate the probability of reaching or exceeding a certain damage level in the building conditioned on the ground motion with a certain level of linear spectral acceleration. In another study, Hwang and Huo [1997] developed fragility functions for speciﬁc reinforced concrete frames as part of a loss assessment study of Memphis buildings. Similarly, Erberik and Elnashai [2004] computed fragility curves for 5-storey ﬂat-slab structures. In the probabilistic framework being developed at the Paciﬁc Earthquake Engineering Research (PEER) Center, damage assessment is performed in two steps. In the ﬁrst step, a probabilistic description of the structural response at increasing levels of ground motion intensity is obtained through a series of response history analyses. Then, in a second step, damage to individual structural and nonstructural components is estimated as a function of structural response parameters (e.g. peak interstory drift demands, peak ﬂoor accelerations, etc.) computed in the ﬁrst step. This approach requires fragility functions for various damage states for each component in the facility as a function of structural response parameters. The objective of this work is to summarise research aimed at the development of fragility functions that estimate damage in slab-column connections of non-ductile reinforced concrete buildings as a function of the peak interstorey drift imposed on the connection. In particular, cast-in-place slab-column connections built prior to 1976 which have no shear reinforcement have been studied. An important amount of research has been conducted on the seismic design and behaviour of slab-column connections. Moehle et al. [1988], ACI-352 [1988], Hueste and Wight [1999] and Enomoto and Roberston [2001] provide excellent summaries of previous research. These previous research has been aimed at improving design recommendations for slab-column connections in new buildings located in seismic regions, and in particular design recommendations to avoid shear failures. However, results from previous experimental studies also provide valuable information to estimate the level of damage that can occur in slab-column connections in existing reinforced concrete buildings as they are subjected to increasing levels of lateral deformation.

2. Deﬁnition of Damage States In this study, four discrete damage states of slab-column connections have been introduced. These damage states have been deﬁned based on speciﬁc actions that would have to be taken as a result of the observed damage. This approach facilitates the estimation of economic losses and other types of consequences (e.g. repair time, possible casualties, etc.) resulting from the occurrence of the damage. The following damage states have been considered: • DS 1 Light Cracking: Light cracking corresponds to crack widths smaller than 0.3 mm (0.013 in) which become visible at distances of about 2.0 m (6.6 ft).

2001]. ACI 548. Figure 1(a) shows a schematic view of a typical cracking pattern on the top face of the slab at this damage state. Figure 1(b) shows a typical cracking pattern on the top face of the slab at this damage state. 1998]. (b) severe cracking. The purpose of this repair action is to improve the aesthetic appearance of the slab or to provide an additional barrier against water inﬁltration into the slab. ACI 546. 1996. 1984.1. and loss of vertical carrying capacity.013 in) and 2 mm (0. actions associated with this damage state typically consist of either “no repair” or a “light repair” by applying a coating on the concrete surface to conceal the projection of cracks [FEMA 308. This repair action is not intended to provide any increase in strength and stiﬀness to the cracked slab. 1. . For this level of cracking most concrete repair guidelines suggest epoxy injection [ACI 224.08 in). These crack widths can be tolerated in reinforced concrete in humid or moist air conditions [ACI 224R. (c) punching shear failure. Schematic view of the four damage states considered for slab-column connections: (a) Light cracking.Fragility Assessment of Slab-Column Connections 779 DS1 DS2 DS3 DS4 Fig. Hence. 1997]. • DS 2 Severe Cracking: This damage state involves extensive cracking with crack widths between 0.3 mm (0. The repair action associated with this damage state typically consists of crack epoxy injection which provides a structural binding agent to ﬁll the crack and adhere to the substrate material [FEMA 308.

• DS 3 Punching Shear Failure: This damage state corresponds to severe cracking characterised by a roughly circular tangential cracking around the column area of slab. which includes 17 experimental research investigations for a total of 82 specimens. Data considered in this study was limited to interior slabcolumn connections without shear reinforcement. this damage state has disastrous consequences. radial cracks extending from that area. 1986.g. In most cases there were not enough information to establish the interstorey drifts at which all four damage states took place. sometimes it is necessary to cut out the damaged length of reinforcing bars in order to replace it with new bars. Therefore. 1984. This was either because the damage state did not occur (e. 1992]. since it can lead to a local collapse. 1998]. 3. If there is no possibility to redistribute the vertical load to other members. Results of experimental research conducted over the last 36 years in 10 diﬀerent research universities were considered. When repairing this type of damage. 1979. in most cases the level of lateral deformation that produced damage in various ﬂoor levels is unknown. If reinforcing bars are exposed. Experimental Results Used in this Study Estimation of the probability of experiencing various damage states in slab-column connections requires gathering information about the level of lateral deformation at which various damage states have been observed. local collapses produced by the LVCC have triggered a global progressive collapse in buildings [Rosenblueth and Meli.. Information about the material properties and characteristics of all specimens considered in this study is summarised in Table 1. Pan and Moehle. • DS 4 Loss of Vertical Carrying Capacity (LVCC ): At this damage state component loses its vertical carrying capacity. Figure 1(d) shows a schematic view of a slab-column connection at this damage state. Mitchell and Cook. one would use information from actual buildings that have undergone various levels of earthquake damage.780 H. Aslani & E. results from experimental studies have been used as the basis for establishing levels of lateral deformations associated with diﬀerent damage states. 1996]. 1995]. and collapses under its gravity load. Figure 1(c) shows typical cracking pattern on the top face of the slab at this damage state. a punching shear failure or LVCC was not observed during the test) or because the report did not document well enough information . However. The main purpose of this repair action is to partially restore the original strength and stiﬀness of the connection [Krauss et al. Loose concrete is removed with chipping hammer. and consist of concrete spall repair and rebar replacement [FEMA 308. In many cases. Hamburger. Miranda 1998]. concrete is removed to provide suﬃcient clearance around the bar for the patch to bond to the full diameter. Repair actions involve signiﬁcant labour and cost. Ideally. 1974]. and considerable spalling of the concrete cover [ASCE-ACI 426. Previous studies have shown that LVCC occurs at larger levels of deformation than those associated with punching shear failure [Hawkins and Mitchell.

9 34.8 475.1 317.8 460.Table 1.8 31.50 7.0 462.33 0.5 × 22.7 34.50 7.9 35. [1976] S6 S7 S8 13 14 15 Ghalli et al.45 0.00 15.22 Yes Yes Yes 22.11 0.40 Yes Yes Yes Yes 25 × 25 25 × 25 25 × 25 25 × 25 25 × 25 25 × 25 25 × 25 25 × 25 25 × 25 25 × 25 25 × 25 25 × 25 25 × 25 25 × 25 25 × 25 25 × 25 0.04 0.0 462. Specimen number fy (Mpa) (6) 354.8 315.0 32.9 29.8 322.00 15.8 475.7 315.9 373.3 335.50 7.8 475.6 475.8 35.0 32.75 15.8 (5) References Label Thickness (cm) fc (Mpa) Vg V0 Bottom reinforcement through the column fy (Mpa) (11) 354. Specimen properties Slab Column Properties of slab-column specimens used in this study.0 0.9 35.8 460.00 15.8 23.50 7.6 455.2 339.4 22.5 26.81 — Yes Yes Yes 0.6 475.3 335.5 34.3 23.23 0.6 455.00 7.5 26.86 0.2 26.3 39.50 7.04 0.7 33.50 8. [1976] SM 0.7 33.18 0.1 33.30 — — — — — — — — — 0.9 290.0 460.0 32.9 45.8 35.5 22.0 462.5 SM 1.0 460.1 33.6 462.3 31.75 8.00 15.2 339.9 29.2 36.1 317.6 462.3 (7) (8) (9) (10) Section (cm × cm) fc (Mpa) (4) 7.00 15.2 36.75 8.9 23.00 15.9 373.00 15.50 7.7 33.00 15.4 410.02 0.05 — — 30 × 15 30 × 15 33.7 330.5 36.3 39.02 0.8 315.9 34.4 22.33 0.7 29.7 0.8 23.8 322.50 45.5 Fragility Assessment of Slab-Column Connections 16 17 18 19 20 21 Morrison and Sozen [1981] S1 S2 S3 S4 S5 D1 781 .7 315.00 15.6 22.7 330.5 × 22.3 36.0 SM 1.0 460.7 29.31 0.0 (1) (2) (3) 1 2 Hanson and Hanson [1968] P7 C8 3 4 5 Islam and Park [1973] IP2 IP3C IP1 6 7 8 9 Hawkins et al.0 460.9 290.45 0.2 26.0 32.5 × 22.02 0. [1974] S1 S2 S3 S4 10 11 12 Seymonds et al.8 475.05 0.4 410.0 462.

2 54.0 415.9 27.5BH 14.3 22.68 1.7 49.00 15.0 415.4 31.48 0.69 0.0 415.41 0.3 30.7 28.00 15.4 31.25 11. [1989] 6AH 9.0 415. Miranda (1) (2) (3) 22 23 D2 D3 24 1 2 3 4 25 26 27 28 Zee and Moehle [1984] Pan and Moehle [1988] INT Test Test Test Test Hawkins et al.25 11.0 22.0 435.00 15.0 472.5 × 13.00 15.74 Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes 30 × 30 30 × 30 30 × 30 30 × 30 30 × 30 30 × 30 30 × 30 30 × 30 30 × 30 30 × 30 30 × 30 30 × 30 30 × 30 30 × 30 30 × 30 30 × 30 30 × 30 31. Specimen properties Slab Column (Continued) Specimen number fy (Mpa) (6) 327.2BH 7.5 47.3 31.0 472.00 15.4 57.0 415.5 37.5 52.9 27.6CH 14CH 6CL 14CL 15.19 Yes Yes Yes Yes 27 × 27 27 × 27 27 × 27 27 × 27 0.0 415.5 18.2 54.7 30.1 20.0 355.72 0.0 415.0 415.43 0.0 415.38 0.0 472.0 415.53 0.1 472.48 0.00 33.2 19.0 415.4 57.25 11.00 15.0 415.0 415.0 415.0 20.7 28.0 472.00 12.00 12.00 15. [1989] 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 Hawkins et al.0 415.42 0.25 15.2 33.18 0.3BH 9.04 0.6AL 14AL 7. Aslani & E.00 11.0 415.0 Section (cm × cm) fc (Mpa) fy (Mpa) (4) 7.71 0.8 29.0 415.5 47.7 415.0 415.3 33.0 22.0 355.50 7.38 0.3 31.0 0.37 0.5 18.25 11.3 22.0 435.80 0.48 0.00 12.782 Table 1.6AH 14AH 6AL 9.55 0.0 415.5 52.00 15.04 — — 25 × 25 25 × 25 33.1 20.0 415.0 415.0 415.00 31.0 472.0 415.1 472.0 415.0 415.0 415.3BL 9.2 19.5 (7) (8) (9) (10) (11) 327.9 36.3 30.5 (5) References Label Thickness (cm) fc (Mpa) Vg V0 Bottom reinforcement through the column H.3 33.21 Yes 13.0 415.0 415.7 30.4 37.0 415.2BL 6CH 9.25 11.5BL 14.2 33.7 415.36 0.7 49.0 415.9 36.0 472.0 20.4 0.0 0.8 29.37 0.50 6.00 12.0 415.0 415.0 .00 15.

00 15.81 0.52 Yes Yes Yes Yes — — — 25 × 25 25 × 25 25 × 25 25 × 25 — — — 0.0 415.00 15.5 24.25 11.0 31.7 26.0 415.0 415.9 33.0 415.00 15.0 32.0 415.3 — — — Section (cm × cm) fc (Mpa) (4) 15.50 0.0 415.00 15.35 0.9 524.6GH3 415.0 415.49 0.85 0.0 415.0 415.00 15.52 0.0 415.9 — — — 0.3 25.0 415.9 524.0 33.0 415.61 0.0 415.45 0.0 415.0 415.0 415.45 Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes 30 × 30 30 × 30 30 × 30 30 × 30 30 × 30 30 × 30 30 × 30 30 × 30 30 × 30 30 × 30 20 × 40 40 × 20 45 × 15 (7) (8) (9) (10) 30.00 15.00 15.0 415.9 524.3 25.7 28.9 18.47 0.2FHI 10.2FHO 14FH 6FLI 10.2FLO 9GH2 9.0 500.00 11.18 0.2 524.6GH0.8 31.3 30.7 26.Table 1.2 25.2 30.9 — — — 59 60 61 62 Robertson and Durrani [1990] 2C 6LL 7L 8I Fragility Assessment of Slab-Column Connections 63 64 65 Dilger and Cao [1991] CD1 CD2 CD8 783 .0 (5) References Label Thickness (cm) fc (Mpa) Vg V0 Bottom reinforcement through the column (11) fy (Mpa) (1) (2) (3) 46 47 48 49 50 51 52 53 54 55 56 57 58 6DH 14DH 6DL 14DL 10.0 415.0 415.00 15.2FLI 10.2 25.8 31.9 18.7 28.0 415.80 0.5 24.25 11.3 27.00 15.37 0.8 39. Specimen properties Slab Column (Continued) Specimen number fy (Mpa) (6) 415.9 33.0 31.0 415.00 15.9 524.25 11.1 26.5 9.63 0.80 0.43 0.0 415.1 26.3 27.2 30.0 32.8 39.2 524.7 24.7 24.00 15.0 45.0 500.0 415.18 0.00 15.51 0.65 0.25 — — — — — — 33.

00 11.1 References Label Thickness (cm) fc (Mpa) Vg V0 Bottom reinforcement through the column fy (Mpa) 372.7 413.7 24.6 379.23 0.3 372.0 0.3 35.7 420.32 0.40 11.6 379. [1993] 75 Luo et al.3 24.7 413.28 0.49 0.3 372.3 372.2 20.25 0.0 66 67 68 69 SCO 1 2 3 4 II Durrani and Du [1992] DNY DNY DNY DNY 70 Wey and Durrani [1992] 71 72 73 74 Farhey et al.6 32.24 0.0 15.1 26.3 11.40 11. [2002] — Information was not available .40 11.23 0.6 19.56 Yes Yes Yes Yes 0.37 0.9 39.6 457.3 18.25 Yes 25 × 25 30 × 20 30 × 20 30 × 20 30 × 12 25 × 25 25 × 25 25 × 25 25 × 25 25 × 25 25 × 25 25 × 25 25 × 25 0.26 0.08 No No No No No No No Yes 0.9 457.00 8.2 35.0 39.5 35.3 524.7 29.9 39.3 372.3 524.1 35.32 0.40 11.3 25.25 11.27 No No No No 25 × 25 25 × 25 25 × 25 25 × 25 35.7 413. from Enomoto and Robertson [2001] ND1C ND4LL ND5XL ND6HR ND7LR ND8B 82 Robertson et al.6 457.1 35.3 372.40 11.6 19.6 32.0 20.25 11.28 0.40 11.6 457.20 0.7 24.1 15.3 18.9 457.00 8.1 15.4 Section (cm × cm) fc (Mpa) 1 2 3 4 11.25 11.47 0.3 372.6 457.25 8.7 413.7 35.1 39.0 413.0 15. [1994] 76 77 78 79 80 81 1C Johnson and Robertson [2001].25 11. Specimen properties Slab Column (Continued) H.7 413.3 24. Aslani & E.00 8.4 29. Miranda Specimen number fy (Mpa) 372.6 457.1 26.0 413.7 413.25 35.6 457.7 413.3 25.7 413.784 Table 1.7 413.7 413.7 420.

light cracking. primarily because until very recently. Only ﬁve investigations included detailed information about the level of interstorey drift at which light cracking was observed. Table 2. A few studies systematically reported cracking patterns and crack widths at various levels of lateral deformation to be used for the second damage state. in some of the specimens. was accompanied with a signiﬁcant reduction in lateral stiﬀness. Interstorey drifts presented in Table 2 have been calculated using the slab centreline to slab centreline interstorey height. It can be seen that the ﬁrst damage state in slab-column connections occurs at very low deformations raging from 0. Careful study of such investigations [Pan and Moehle. For these specimens. it was assumed that this damage state occurs at peak interstorey Lateral Load Significant change in stiffness IDRDS1 IDR Fig. . Consequently. 1988. 1990] suggests that damage state two. typically occurs when top steel reinforcement is at yield. Typical drop in lateral stiﬀness observed in the hysteresis loop of slab-column specimens used to identify the ﬁrst damage state. the interstorey drift at which a 30% or more sudden reduction in lateral stiﬀness was observed in the hysteresis loop was considered. The later situation was particularly common for the ﬁrst two damage states (light cracking and severe cracking). Therefore. Therefore. in order to gather more data points associated with the second damage state. Table 2 reports interstorey drifts associated with light cracking for only 43 specimens. 2. Figure 2 presents a graphic representation of a point in loading history where this sudden reduction occurred.19% to 0.8% drift ratios. the reported interstorey drift at which clearly visible cracking occurred. Robertson and Durrani. in order to expand the number of data points associated with this damage state. The indirect way of determining the interstorey drift associated with ﬁrst damage state was only used in specimens where a good quality hysteresis loop at early loading cycles was included in the report.Fragility Assessment of Slab-Column Connections 785 to properly establish the level of interstorey drift at which the damage state was observed. summarises the interstorey drifts associated with this ﬁrst damage state. earthquake provisions have mainly been concerned with life safety and not damage control.

35 ** ** ** 0.80 5.00 ** ** ** ** ** ** ** ** 1.40 0.54 0.79 ** ** ** ** ** ** ** ** ** ** ** ** ** 4.00 6.00∗∗∗∗ 1.73 1.22 0.69 ** ** ** ** ** ** ** ** 0.42 0.50 4.72 4.76 0.0 3.19 0.43 0.48 1.786 H.28 0.82 1.20 0.31 0.39 1.00 1.91 1.74 0.25 0.20 1.40 0.90 4.00 *Centreline dimensions were used to compute IDR **Information was not available ***Punching shear failure did not occur ****IDR corresponds to a ﬂexural failure drifts at which top steel reinforcement was reported to yield or at peak interstorey drifts at which a residual drift in the hysteresis loop was observed after unloading.50 5.93 0.36 0.50 ** ** ** ** ** ** ** ** ** ** ** ** ** IDRDS2 (%) (3) ** ** 1.48 0.80 4.80 1.98 0.07 1.50 0.43 0.00 1.11 0.70 IDRDS2 (%) (3) ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** 1.00∗∗∗∗ 8.22 0.47 2.00 2.09 0.00 1.25 4.00 3.97 1.00 4. Interstorey drift ratios∗ used to develop the fragility functions.64 0.70 3.32 3.31 0. Miranda Table 2. Aslani & E.40 3.69 0.00 5.90 0.28 0.50 ** 1.50 2.77 0.11 2.71 0.00 1. IDRDS3 (%) (4) 3.77 0.89 4.25 0.65 1.70 5.25 1.79 1.59 IDRDS4 (%) (5) Specimen IDRDS1 number (%) (1) 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 (2) ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** 0.65 0.50 IDRDS4 (%) (5) Specimen IDRDS1 number (%) (1) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 (2) ** ** 0.89 4.48 0.19 0.22 4.04 *** 6.00 4.89 0.82 2.50 0.53 2.50∗∗∗∗ 4.00 3.50 3.00 2.00 5.62 1.00 2.00 5.28 0.40 IDRDS3 (%) (4) 1.70 2.70 3.57 5.60 1.00 0.27 1.70 3.18 2.50 0.80 ** ** ** 0.75 1.11 1.00 4.81 3.97 1.28 0.75 0.39 1.70 4.40 4.30 3.89 0.80 0.16 5.89 0.91 2.43 0.50∗∗∗∗ 2.27 1.50 0.50 0.76 0.70 1. The third column in Table 2 lists the interstorey drifts that were identiﬁed or inferred corresponding to this .48 2.90 2.28 0.11 1.50 ** 0.54 0. Figure 3 schematically illustrates the latter situation.00 5.30 1.31 0.00 5.80 1.93 0.14 1.20 0.28 0.00 3.00 1.00 4.43 0.60 1.48 0.49 0.36 1.80 5.40 0.08 0.32 0.90 1.60 5.70 3.76 1.22 0.08 0.30 0.63 1.46 1.48 2.27 1.62 1.30 4.25 1.

57% to as large as 8% drift ratios. The fourth column in Table 2 lists interstorey drift ratios at which punching shear failures occurred in various specimens. However.6% and 6. 3. The main goal of these subsequent cycles was the investigation of post-punching behaviour of the slab-column connections.Fragility Assessment of Slab-Column Connections 787 Lateral Load IDR DS 2 IDR Permanent residual displacement Fig. The ﬁrst occurrence of a residual lateral drift in the hysteresis loop of a slab-column connection specimen used to identify the second damage state. It can be seen that drifts at which these failures have occurred exhibit a very large dispersion ranging from 0. Fragility Functions As shown in Table 2. With the exception of ﬁve specimens in which shear failure was not observed. these specimens have been used to obtain information of the post-punching deformation capacity and of a conservative estimate of the probability of experiencing LVCC in the connection. in some of the specimens. Most experimental studies have investigated parameters that inﬂuence the moment capacity and deformation at which punching shear failure occurred. there are six studies representing a total of 18 cases in which the specimens were subjected to one or more cycles or larger deformations after punching shear failure was observed.49% and 1. Hence. Interstorey drifts associated with severe cracking were obtained for 33 specimens which range between 0. In this study. 4. Maximum drift ratios at which testing were stopped in diﬀerent specimens range between 1.0%. There are no investigations in which testing continued until LVCC occurred. the interstorey drift at which each damage state in slab-column connections was observed shows relatively large variations from one specimen to . in most cases testing was stopped after punching shear failure was observed.63% drift ratios. all other specimens experienced a punching shear failure and the research reports contained information about the interstorey drift at which this damage state was observed. severe cracking. damage state.

788 H. 4. respectively. Fragility functions ﬁtted to interstorey drift ratios corresponding to the ﬁrst and second damage states in slab-column connections. This cumulative frequency distribution function provides information about the portion of the data set that does not exceed a particular value of drift and represents an empirically derived cumulative distribution function. The second damage state (severe cracking) occurred in half of the specimens when the interstorey drift ratio reached 0. Aslani & E. IDR is the median of the interstorey drift ratios. a drift-based fragility function provides the probability of experiencing or exceeding a particular damage state conditioned on the peak interstorey drift. Figure 4(b).9%.4% was reached. σLn IDR (1) where P (DS ≥ dsi |IDR = idr) is the probability of experiencing or exceeding damage state i. 4(a). Figures 4(a) and 4(b) show cumulative distribution functions for the ﬁrst and second damage states. at which (a) (b) Fig. Miranda another. the ﬁrst damage state (light cracking) was observed in half of the specimens when an interstorey drift ratio of about 0. This specimen-to-specimen variability has taken into account by developing drift-based fragility functions that estimate the likelihood of experiencing a particular damage state.5)/n. a cumulative frequency distribution function was obtained by plotting ascending-sorted interstorey drifts at which the damage state was experimentally observed against (i−0. In particular. At each damage state. . As can be seen in Fig. IDR’s. Also plotted in these ﬁgures are ﬁtted lognormal cumulative distribution functions of the interstorey drift ratios given by P (DS ≥ dsi |IDR = idr) = Φ Ln(idr) − Ln (IDR) . where i is the position of the peak interstorey drift and n is the number of specimens.

In order to verify if the cumulative distribution function could be assumed as lognormally distributed. Lognormal fit K-S test. Statistical parameters estimated for interstorey drift ratios corresponding to the four damage states in slab-column connections.00 0.9 0. and Φ is the cumulative standard normal distribution. It can be seen that for both damage states this assumption is adequate.4 0.06 0. σLn IDR is the standard deviation of the natural logarithm of the IDR’s.04 IDR Fig.39 0.3 0.40 0.0 0.95 2.1 0. Damage state (1) DS1: DS2: DS3: DS4: Light cracking Severe cracking Punching shear failure Loss of vertical carrying capacity IDR(%) (2) 0.25 0. punching shear failure.28 σLn IDR (3) 0.5 0.Fragility Assessment of Slab-Column Connections 789 damage state i was observed.62 0.6 0. It can be seen that the lognormal distribution ﬁts the data relatively well. 5. 1970] was conducted. 4(a) and 4(b) are graphical representations of this test for 10% signiﬁcance levels. a Kolmogorov-Smirnov goodness-of-ﬁt test [Benjamin and Cornell.0 0. Empirical and ﬁtted cumulative distribution functions for the third damage state are shown in Fig. Parameters for the ﬁtted lognormal probability distribution are given in Table 3. The hypothesis that the assumed cumulative probability distribution adequately ﬁts the empirical data is accepted if all data points lie between the two gray lines. 5.02 0.8 0.00 4.7 0.10% significance DS 3 σ ln IDR = 0. Also shown in Figs.62 Data 0.36 Number of specimens (n) (4) 43 33 77 18 P (DS≥ds3|IDR) 1. in a slab-column connection. Fragility function ﬁtted to interstorey drift ratios corresponding to the third damage state.08 .2 0. The corresponding parameters Table 3. Also shown in the ﬁgure are the 10% signiﬁcance levels of the Kolmogorov-Smirnov goodness-of-ﬁt test which indicate that the lognormal assumption is also valid for the third damage state.

where Vg is the vertical shear that acts on the slab critical section deﬁned at a distance d/2 from the column face in which d is the average slab eﬀective depth. In the proposed fragility surface.5 times larger than the dispersion corresponding to the second damage state. but the median interstorey drift capacity is computed as a function of the level of gravity load present in the slab at the time of earthquake as follows: IDRDS3 = 0. Pan and Moehle [1988]. (1). Miranda for the cumulative distribution of this damage sate are given in Table 3. a fragility surface was developed for the third damage state. b0 is the perimeter length of the slab critical section in mm. In particular. which represent 44 data points in addition to those originally gathered by Hueste and Wight. Also shown in the ﬁgure is the trilinear variation of the drift capacity at punching shear proposed by Hueste and Wight [1999]. the probability of experiencing punching shear failure is computed with Eq. It can be seen that for the third damage state σLn IDR is 0. This means that the estimation of whether a punching shear failure is likely to occur in a slab-column connection is very uncertain if only the peak interstorey drift ratio is used. and d is the slab eﬀective width in mm. On the basis of the above observations. (3) . Hueste and Wight [1999] proposed a trilinear equation to estimate the drift at which punching shear failure occurs. This ﬁgure shows results from 77 specimens. fc is the compressive strength of concrete in MPa and not to exceeding 40 MPa. [1988].6 times larger than the level of dispersion corresponding to the ﬁrst damage state and 2. Column seven in Table 1 lists the gravity shear ratio in all specimens where punching shear failure was observed. which according to ACI-352 [1988] is given by V0 = 1 1 + 6 3βc fc b0 d ≤ 1 3 fc b0 d. More recently. Aslani & E.03 exp −4440 Vg V0 8 . in which the probability of experiencing a punching shear failure was computed as a function of the peak IDR and the vertical shear ratio. (2) governs. (2) where β c is the ratio of long-to-short cross-sectional dimensions of the supporting column. Robertson and Durrani [1990] indicated that the displacement ductility ratio at which punching shear failure occurs is a function of the gravity load ratio Vg /V0 . Moehle et al. Figure 6(a) shows the drift at which punching shear failure occurred as a function of the vertical shear ratio. The dispersion parameter (σLn IDR ) which is approximately comparable to the coeﬃcient of variation of the data is shown in Column 3 of Table 3 for each damage state. Vg /V0 .790 H. Experimental research has shown that the deformation capacity of slab-column connections is a function of the level of gravity load.014 + 0. For most columns encountered in practice the maximum limit shown in Eq. and the normalising shear force V0 is the theoretical punching shear strength in the absence of moment transfer in the connection.62 which is 1.

4 Hueste et al. (4) Figure 6(b) also shows the variation of dispersion computed with Eq.Fragility Assessment of Slab-Column Connections 791 IDRDS3 0.0 0.01 0. Figure 7(b) shows fragility functions for the third damage state for slabcolumn connections with gravity shear ratios equal to 0. Figure 7(a) shows the fragility surface resulting from the use of Eqs. (4).1 subjected to an interstorey drift ratio of 0.02 0. 6. For example. in slab-column connections.8 1. IDR DS3 . Comparison of Figs.6 Vg V0 2.62 − 0. (a) Variations of interstorey drift ratio at punching shear failure. (3) 0. where it can be seen that this equation captures relatively well the variation of the dispersion parameter as a function of the gravity shear ratio.4 exp −11. Eq.02. with changes in the level of gravity shear.06 0. 5 and 7(b) shows that computing the two parameters of the lognormal distribution as a function of the gravity shear ratio using Eqs. in a slab-column connection with a vertical shear ratio of 0.05 0.9 .05.6 Vg 0. as a function of the gravity load ratio.03 0.2 moving at increments of 0.2 0. (3) and (4).1. Vg /V0 . the probability of experiencing or exceeding a punching shear failure is essentially .08 Data 0. (1). (1999) This study. It can be seen that the level of dispersion signiﬁcantly increases as the level of gravity shear ratio increases.0 V0 (a) (b) Fig. Points shown in this graph were obtained using a moving window analysis with a gravity load ratio width of 0. Figure 6(b) shows the variations of the dispersion of IDR’s at which punching shear failure occurred around IDRDS3 .00 0.04 0.3 and 0. This variation of dispersion with changes with the level of gravity load ratio was approximated as follows: σLn IDR DS3 = 0.07 0. (3) and (4) leads to much better estimations of the probability of experiencing or exceeding a damage corresponding to a punching shear failure. 0.5. (b) variations of the dispersion of the IDR D S3 with changes in Vg /V0 .

the following ratio was computed Γ= IDRDS4 . signiﬁcantly overestimating the fragility of the connection. Aslani & E. . zero. IDRDS3 (5) where IDRDS4 is the interstorey drift ratio at which the test was stopped in specimens where the post-punching shear failure behaviour was studied.2. underestimated by 29%. corresponding to the 18 specimens in which post-punching shear failure was studied. Similarly. 5). Proposed fragility surfce to account for high levels of specimen-to-specimen variability for the third damage state: (a) 3-dimensional presentation. if a slab-column connection with a vertical shear ratio of 0. This ratio of drifts can be interpreted as an ampliﬁcation factor acting on the drift corresponding to punching shear failure in order to provide a conservative estimate of the drift at which LVCC occurs. 5. and 10% signiﬁcance levels of the Kolmogorov-Smirnov goodness-of-ﬁt test. Miranda (a) (b) Fig. while if the eﬀect of the gravity load ratio is neglected (as done in Fig. indicating that it is adequate to assume that this ratio is also lognormally distributed. hence. P (Γ < γ). (b) 2-dimensional presentation. As shown in this ﬁgure all data points lie within the conﬁdence bands.23 and logarithmic standard deviation of 0. for each specimen that was subjected to further cyclic loading after the punching shear failure was observed. Also shown in this ﬁgure is a lognormal ﬁt of this parameter with median equal to 1. 7.792 H. this probability would be estimated as 46%. whereas if the eﬀect of the gravity load ratio is neglected. Figure 8 shows a cumulative frequency distribution of Γ. 75%. there is a very high probability of experiencing or exceeding a punching shear failure. as done in Fig.5 is subjected to the same level of drift. this probability would be estimated as 46%. In order to obtain an estimate of the probability of experiencing LVCC.

(6). there is a 97% probability .1 subjected to an IDR of 0. and (7). 0.5 is subjected to the same level of IDR. such as 0.1.3σLn IDRDS3 + 0. σLn IDRDS4 = 2 σLn IDRDS3 − 0. For example. As can be seen in Fig. if one wants to avoid the possibility of having LVCC at high levels of gravity shear ratio.04. Lognormal cumulative distribution function ﬁtted to the cumulative frequency distribution of γ. the IDR needs to be limited to 0. However.3 and 0. (7) corresponds to the correlation between the drift at which punching shear failure was reported and the ampliﬁcation factor Γ. the probability of experiencing a loss of vertical carrying capacity is approximately 2%. Figure 9(a) presents the fragility surface corresponding to the LVCC damage state computed with Eqs. 9 estimating the two parameters of the lognormal distribution for DS 4 as a function of the gravity shear ratio leads to much better estimations of the probability of experiencing LVCC. (6) (7) The value of −0.04.6%. if a slab-column connection with a vertical shear ratio of 0. Figure 9(b) shows fragility functions corresponding to LVCC for slab-column connections with vertical shear ratios of 0.3 in Eq.23 · IDRDS3 . 8.1. Furthermore neglecting the eﬀect of gravity load can lead to results that in some cases are too conservative while in other are unconservative. in a slab-column connection with a vertical shear ratio of 0. However. (1).5. It can be seen that if the vertical shear ratio has a low value such as 0. Similarly. LVCC is not likely to occur provided that IDR’s are smaller than 3.5. if the eﬀect of the gravity load ratio is neglected this probability would be estimated as 55% which is signiﬁcantly larger. The interstorey drift ratio at which LVCC occurs is assumed to be lognormally distributed with median and logarithmic standard deviations given by IDRDS4 = 1.8%.Fragility Assessment of Slab-Column Connections 793 Fig.

Column 8 in Table 1 summarises the information on whether the slab bottom reinforcement in the various specimens passed through the column reinforcement cage. For example. (b) 2-dimensional presentation. underestimated by 42%. even though specimen 69 had no bottom slab reinforcement passing through the column. It can be seen that of 11 specimens that did not contain any continuous bottom reinforcement. Aslani & E. though less reliable and eﬃcient than bottom slab reinforcement. post-punching behaviour was only studied in 4 specimens. the connection does not loose its vertical carrying capacity immediately after experiencing punching shear failure. The drift level at which LVCC damage state occurs in a slab-column connection is not only inﬂuenced by the level of gravity load. Miranda (a) (b) Fig. is capable of resisting vertical load following initial punching [Pan and Moehle. even when there is no continuous bottom reinforcement. (a) 3-dimensional presentation. since top slab reinforcement. whereas if the eﬀect of gravity load ratio is neglected this probability would be 55%. slab-column connections with continuous bottom reinforcement may resist signiﬁcant vertical loads following initial punching failure of the slab which leads to a signiﬁcant increase in the drift capacity associated with experiencing LVCC. results of the four specimens with no continuous bottom reinforcement do not show a clear trend of decreased drift capacity compared to . Since the maximum drift shown in Table 2 corresponds to the drift at which testing was stopped and not to the drift at which LVCC occurred. but also by whether or not the slab has longitudinal steel reinforcement that is continuous through the column reinforcement cage. Fragility surface developed for the loss of vertical carrying capacity damage state in slab-column connections. of experiencing LVCC. Results from these 4 specimens indicate that. 1992]. In particular.794 H. the test was stopped at 6% drift ratio without experiencing LVCC. which corresponds to a drift 28% larger than the one at which the punching shear failure occurred. 9.

As can be seen in Fig. it is ﬁrst required to estimate fragility curves corresponding to all the damage states deﬁned in the connection. (1). i=m P (DS < dsi |IDR = idr) (8) where i = 0 corresponds to the state of no damage in the component. 10(b). and m is the number of damage states deﬁned for the component. The probability that the connection has not experienced any physical damage is. Figures 10(b) and 10(d) show the probability of being at each damage state for slabcolumn connections with gravity shear ratios equal to 0. or loading protocols did not permit this to be observed in the laboratory. 35% to be in the third damage state and 35% to be in the last damage state. the probability that the connection is in the ﬁrst damage state is 40% while there is a 58% probability for the connection to be in the second damage state. At 5% drift. e. and for the case of slab-column connections is 4. this should not be interpreted as the bottom reinforcement not increasing the drift at which LVCC occurs.2 and at 1% drift. Fragility functions presented in these ﬁgures can then be used in Eq. As the level of gravity shear ratio increases. P (DS ≥ dsi |IDR = idr) is the fragility function for the ith damage state of a component which is computed form Eq. for a connection with a gravity load ratio of 0. drifts associated to DS 4 provide a conservative estimate of the probability of experiencing a LVCC because this damage state was not observed in any of the specimens.g. Figures 10(a) and 10(c) present fragility functions corresponding to the four damage states deﬁned for slab-column connections at two levels of gravity shear ratio. Probability of Being at Each Damage State Fragility functions developed for a slab-column connection can be used to estimate the probability that the connection is at a speciﬁc damage state when it is subjected to a certain level of IDR.Fragility Assessment of Slab-Column Connections 795 that of the specimens with continuous bottom reinforcement.2 and 0.5. 2%. However. As mentioned before. respectively. (8) to a speciﬁc slab-column connection. In order to apply Eq. 5. however. at two diﬀerent locations in a building. punching shear failure and LVCC damage states occur at signiﬁcantly lower levels of interstorey drift ratio. (8) to estimate the probability of being at each damage state as a function of the level of IDR in the slab-column connection. 2003] and can be estimated as the diﬀerence between fragility functions corresponding to two consequent damage states as follows P (DS = ds i |IDR = idr ) i=0 1 − P (DS ≥ dsi+1 |IDR = idr) = P (DS ≥ dsi |IDR = idr) − P (DS ≥ ds i+1 |IDR = idr ) 1 ≤ i ≤ m . therefore. This probability is a primary input when performing loss estimation in buildings [Miranda and Aslani. as can be seen in Fig. . the connection for sure has experienced physical damage and there is a 30% probability that the connection is in the second damage state. and simply that the test setup. 10(d).

0 0.025 0.4 0.2 0.4 DS1 DS2 DS3 DS4 DS1 DS2 DS3 DS4 Vg V0 = 0.100 (a) P(DS≥ds P(DS ds i | IDR) 1. (c) Component-speciﬁc fragility curves for slab-column connections with gravity shear ratios equal to 0. The second source is the uncertainty produced by the fact that damage observations in .6 0. (d) Probability of being at each damage state for the two connections computed with Eq.050 IDR 0. Two sources of epistemic uncertainty were incorporated in this study.075 0.0 0.4 (b) DS1 DS2 DS3 DS4 DS1 DS2 DS3 DS4 Vg V0 = 0. (b). Miranda P(DS≥ds P(DS ds i | IDR) 1.5. (8).025 0.8 0.796 H.100 0.000 Vg V0 = 0.025 0. Epistemic Uncertainty on Fragility Functions In addition to specimen-to-specimen variability.075 0.5 0.0 0.100 (c) (d) Fig. another source of uncertainty that need to be considered in estimating the level of damage in slab-column connections is epistemic uncertainty.000 P(DS=ds i | IDR) 1. (a).2 0.6 0.0 0.075 0.2 0.000 Vg V0 = 0.8 0.050 IDR 0.0 0.2 0. 0.2 0. respectively.8 0.2 0.025 0. 10.0 0.4 0.0 0.6 0.6 0.0 0.000 P( =ds P(DS=ds i | IDR) R) ) 1.5 0. Aslani & E.2 and 0. The ﬁrst source is the uncertainty caused by using fragility functions whose parameters have been obtained from a limited number of specimens. 6.050 IDR 0. This source of epistemic uncertainty is referred here as ﬁnite-sample uncertainty.100 0.075 0.8 0.050 IDR 0.

n−1 1/2 . 1960. a generic method of ﬁnding the conﬁdence interval is to use re-sampling techniques. For each damage state. such as bootstrap statistics [Efron and Tibshirani.Fragility Assessment of Slab-Column Connections 797 the specimen are only collected at peak values of the loading protocol. Once these drift increments were collected. This source of uncertainty is associated with the drift increment in the loading protocol used to test each specimen. Since it was veriﬁed that the fragility curves of a slab-column connection can be assumed lognormal. These conﬁdence intervals were used to build a conﬁdence band on the original fragility functions. a bias and dispersion on the empirical median drift capacity was estimated by assuming that the median capacity is a normally distributed random .n−1 1/2 and 2 (n − 1)σLn IDR 2 χ1−α/2.. 1993]. similarly. The conﬁdence intervals of the median of a lognormally distributed sample can be approximated as follows [Crow et al. (9) and (10).n−1 is the inverse of the χ distribution with n − 1 degrees of freedom 2 and a probability of occurrence of α/2. and n is the number of data points. conventional statistical methods can be used [Crow et al. a conventional approach was used to estimate conﬁdence intervals of the median and logarithmic standard deviation of the drifts corresponding to diﬀerent damage states. A quantitative measure of the second source of epistemic uncertainty was obtained by identifying the drift increment used in the loading protocol in the cycle at which each damage state occurred in each specimen. A quantitative measure for ﬁnite-sample uncertainty can be obtained by computing conﬁdence intervals of the statistical parameters of the deformation levels corresponding to each damage state. the drifts reported at a certain damage state in a specimen will typically correspond to the peak drift that was imposed to the loading cycle in which the damage state was observed. 1970]. Diﬀerent methods can be used to estimate the conﬁdence intervals of a statistical parameter.. Benjamin and Cornell. Conﬁdence intervals for the logarithmic standard deviation of the data are nonsymmetric and can be computed as [Crow et al. If the underlying probability distribution is normal or lognormal. lower and upper conﬁdence intervals of IDR and σLn IDR were estimated using Eqs. χ2 α/2.n−1 is the inverse of the χ distribution with n−1 degrees of freedom and a probability of occurrence of 1−α/2. Therefore. The number of slab-column specimens used to develop fragility functions are given in the last column of Table 3. (10) 2 where χ2 α/2. 1960]: 2 (n − 1)σLn IDR 2 χα/2. n (9) where zα/2 is the value in the standard normal distribution such that the probability of a random deviation numerically greater than zα/2 is α. For cases in which the underlying probability distribution is not normal or lognormal.. 1960]: IDR · exp ± zα/2 σLn IDR √ .

4 0.3 Epistemic 0.1 0.0 0.9 0.6 0. The corresponding parameters are listed in Table 4.5.9 0.5 fragility 0.08 0.8 0. 11.9 0.02 0. Aslani & E.003 0.2 Vg V0 = 0.0 0.3 0.010 0. while the grey bands correspond to a 90% conﬁdence band on the fragility curves when both sources of epistemic uncertainty have been considered.7 0.5 0. Incorporating epistemic uncertainty to fragility functions of slab-column connections.000 0.6 0.8 0.0 8 0. (c).2 and 0.02 0. (a) Damage state 1.2 0.3 Epistemic 0. (d) Damage state 3 With Vg /V0 = 0.015 0. the two sources of epistemic uncertainty were combined assuming that they are uncorrelated.1 envelope 0.000 0.0 0.006 0.0 4 0.06 IDR (c) Original fragility Epistemic uncertainty envelope 0.4 0.2 0. Please note that the second source of uncertainty in addition to increasing the P(DS≥ds 1|IDR) 1.2 uncertainty 0.9 0.1 envelope 0.005 0.10 Fig.6 Original 0.015 IDR (a) P(DS≥ds2|IDR) 1.8 0.6 Original 0.7 0. (b) Damage state 2.0 0.009 0.00 Vg V0 = 0.5.00 P(DS≥ds3|IDR) 1.04 0.798 H. Figure 11 presents the eﬀects of epistemic uncertainty on the fragility curves for each damage state. Finally. .4 0.1 0.012 0.5 Original fragility Epistemic uncertainty envelope 0. (f) Damage state 4 with Vg /V0 = 0.7 0.020 0.0 0.06 IDR (d) 0. Miranda variable.5 0.4 0. (e).2 and 0.10 0.0 0.0 0.5 fragility 0.8 0.2 uncertainty 0.025 IDR (b) P(DS≥ds3|IDR) 1.7 0.3 0. The black line in each graph corresponds to the fragility curve in the absence of epistemic uncertainty.0 0.

2 0. 11(f) that at 2% interstorey drift ratio. by using results from experimental studies. probability of experiencing or exceeding LVCC varies anywhere from 64% to 95%.1 0.08 0.Fragility Assessment of Slab-Column Connections 799 P (DS4≥ds4| IDR) 1.08 0.2 0.8 Vg = 0.3 Epistemic 0.9 0.00 0. with a total of 82 slab-column specimens was used to develop fragility functions for slab-column connections. the study has identiﬁed and quantiﬁed the uncertainties associated with estimating damage in slab-column connections. For example.5 Original fragility Epistemic uncertainty envelope 0. 11.1 envelope 0.00 Vg V0 = 0.06 IDR (f) 0.04 0.0 0. For example.9 0. P (DS4≥ds4| IDR) 1. 11(b).95% drift ratio. for a slab-column connection with a gravity shear ratio equal to 0. Conclusions Fragility functions that provide a probabilistic estimation of the level of damage experienced in slab-column connections of non-ductile reinforced concrete buildings have been presented.61% to 0. These new fragility information can be used in estimating damage at the component level and with damage states associated to speciﬁc repair actions. Furthermore.8 0.12 IDR (e) Fig.0 0.06 0. In these fragility functions the damage is estimated as a function of the peak interstorey drift imposed on the connection. .e.5 0.2 uncertainty 0.10 0. 7.4 0.7 0.7 V0 0. as can be seen in Fig.3 0.0 0.5 fragility 0. epistemic uncertainty can cause important changes in the probability of experiencing or exceeding a damage state.5.04 0. it can be seen in Fig. a 50% probability of experiencing or exceeding the second damage state can occur anywhere from 0. the inﬂuence of epistemic uncertainty is signiﬁcant and needs to be accounted for when performing sensitivity studies in loss estimation [Aslani and Miranda. Experimental data from 16 investigations conducted in the last 36 years.4 0. shifts the fragility curve to the left) As shown in these graphs. 2004]. Furthermore.02 0.02 0.6 Original 0.10 (Continued) conﬁdence band on the median drift also decreases the drift capacity (i.0 0.6 0.

21 (IDR DS3 − 0.0033 × 0.93 DS3: Punching shear failure IDRDS3 × 1.0076 × 0. Parameters considering both sources of epistemic uncertainty IDR (4) 0.07 0.0024∗∗ ) × 0.25 − 0.25 − 0.15 ∗ IDR DS4 × 0.07 0.07 σLn IDRDS4 + 0.0033 × 1.0027 × 1.74 0. Aslani & E.10 σLn IDRDS3 − 0.07 σLn IDRDS4 + 0.14 σLn IDRDS4 − 0.25 + 0.09 0.39 − 0.0024∗∗ ) × 0. Miranda Table 4. .0032 correspond to epistemic uncertainty caused by drift increments in the loading protocol used to test each specimen for damage states 3 and 4.25 0.06 0.25 + 0.09 0.10 σLn IDRDS3 − 0.18 (IDR DS3 − 0.87 ∗ *IDR DS3 and IDR DS4 are computed from Eq.85 σLn IDR (5) 0.12 ∗ IDR DS3 × 0.08 (3) σLn IDR Damage state (1) DS1: Light cracking 0.0076 × 1.0024∗∗ ) × 1.39 + 0.14 σLn IDRDS4 − 0.82 (IDR DS4 − 0.04 σLn IDRDS3 + 0.009 × 0.08 Parameters considering only ﬁnite sample epistemic uncertainty IDR (2) 0. (6).0024 and 0.80 (IDR DS3 − 0.39 + 0. respectively.800 H.009 × 1.0032∗∗ ) × 1.39 − 0. respectively. **Values 0.07 0.35 0.90 DS2: Severe cracking 0. Statistical parameters estimated to incorporate epistemic uncertainty for interstorey drift ratios corresponding to the damage states in slab-column connections at 90% conﬁdence interval. (3) and Eq.89 ∗ DS4: Loss of vertical carrying capacity IDR DS4 × 1.06 0.10 0.04 σLn IDRDS3 + 0.0027 × 0.

1.” ACI Technical Document. Acknowledgements The authors would like to acknowledge the support by the Paciﬁc Earthquake Engineering Research (PEER) Center with support from the Earthquake Engineering Research Centers Program of the National Science Foundation under Award Number EEC-9701568. Quantitative measures for each of these two kinds of epistemic uncertainty were developed using statistical procedures. developing fragility functions both as a function of interstorey drift ratio and as a function of the gravity shear ratio leads to improved estimates of the probability of experiencing these damage states compared to damage estimates obtained only as a function of interstorey drift ratio.” ACI Journal 81(3). The specimen-to-specimen variability was incorporated by developing drift-based fragility functions at each damage state. 46 pages. Detroit. Professor Jack P. Michigan. [1996] “Concrete repair guide. [2001] “Control of cracking in concrete structures.” ACI Technical Document. [1997] “Guide for the use of polymers in concrete (ACI 548. during the course of this investigation. 29 pages. It was observed that. ACI 548. Results indicate that in some cases the eﬀects of epistemic uncertainty on the probability of experiencing each damage state are signiﬁcant and therefore should be incorporated in probabilistic damage assessment. References ACI 224. and the second is associated to the fact that damage observations in the various specimen were typically collected only at the end of each loading cycle and therefore the resulting fragility functions are inﬂuenced by the drift increment used in the loading protocol that was employed to test of each specimen. Two kind of epistemic uncertainty were considered. Robertson at University of Hawaii at Manoa. . for damage states involving punching shear failure and loss of vertical carrying capacity. Detroit.” ACI Technical Document. Michigan. (ACI 546R-96). The ﬁrst one is associated to the fact that the parameters of the fragility functions have been obtained from a limited number of specimens.1R-97). Specimen-to-specimen variability corresponds to the fact that diﬀerent specimens experience the various damage states at levels of deformation that in general are diﬀerent for each specimen. and Professor Ian N. Michigan. Detroit. evaluations. ACI 546. 126 pages. Moehle at University of California at Berkeley.Fragility Assessment of Slab-Column Connections 801 Two sources of uncertainty were incorporated in the fragility functions: Specimen-to-specimen variability and epistemic uncertainty. 211–230 Detroit. ACI 224R. [1984] “Causes. The authors also would like to express their gratitude for comments and suggestions from the director of the PEER Center. Michigan. and repair of cracks.

A. O. 73-47 73(10).” Engineering Structures 26. Y.” Proc. New York.” ASCE Journal of the Structural Division 100(ST8). [1979] “Progressive collapse of ﬂat plate structures.” ACI Structural Journal 85(6). 705–717. 436 pages. 13th World Conference on Earthquake Engineering. A. and Cao. H. Erberik. Sacramento. Hawkins. [1993] An Introduction to the Boot-Strap. and Du. N.” ACI Structural Journal 86(6).. N. S. Hanson. New York. Crow. W. 937–948. D. Hamburger. [1992] “Seismic resistance of slab-column connections in existing non-ductile ﬂat-plate buildings. Ghali. M..” 1994 Northridge Earthquake Buildings Case Studies Project. Aslani. [1996] “Lateral response of older ﬂat slab frames and the economic eﬀect on retroﬁt. . 1903–1916. and Luo.” Technical Report NCEER-92-0010. State of California. and Miranda. ASCE-ACI 426. Hawaii.” ACI Journal. Aslani & E. M. 83–100. Title No. A. Aibin. and Cornell. ATC-13. 684 pages. K. R. A. [1998] “Repair of earthquake damaged concrete and masonry wall buildings.” Journal of the Portland Cement Association. Efron.. B. Canada. J. [1985] “Earthquake damage evaluation data for California. M. State University of New York at Buﬀalo. N. B. W. Dover Publication. J. F. M.2. 1543–1591. E. [2001] Hysteretic Modeling of Flat Slab-Column Connections. Dovich. New York.” Proc.” ACI Structural Journal 92(4). Proposition 122: Product 3. ASCE 119(6). J. [1974] “The shear strength of reinforced concrete members-slabs. BC. Davis. Redwood City. W. 75 pages. New York. Seismic safety commission. E. N. [1988] “Recommendations for design of slab-column connections in monolithic reinforced concrete structures (ACI 352.. [1996] “Department store. Dilger. 288 pages. China. Benjamin. Enomoto. 235 pages. [1960] Statistics Manual. Washington. Conf. 2–16. A. and Maxﬁeld. California. Z. Y. M. [2004] “Component-level and system-level sensitivity study for earthquake loss estimation. J. and Elnashai. DC (FEMA Publication No. M.802 H. Honolulu. and Robertson. B. H.” ACI Structural Journal 76(7). pp. 2nd Int. and Dilger. J. 308) 65 pages. California. 675–696. W. FEMA 308. 566–572. Z. I. 775–808. L. Chapman and Hall. D. [2004] “Fragility analysis of ﬂat-slab structures. C. Moshe. Department of Civil Engineering. N. Research report. J.” Applied Technology Council. Elmasri.1R-89). and Wight. Farhey. [1995] “Seismic resistance of non-ductile slabcolumn connections in existing ﬂat-slab buildings.” Prepared by the Applied Technology Council (ATC-43 Project) for the Federal Emergency Management Agency. Research and development laboratories 10(1). 479–487.” Journal of Structural Engineering. Vancouver. Northridge fashion center. Canadian Association for Earthquake Engineering under auspices of International Association of Earthquake Engineering. A. A. A. and Yamazaki. on High-Rise Buildings. and Yankelevsky.” UHM/CE/01/03. Hawkins. University of Hawaii. Du. D. [1991] “Behavior of slab-column connections under reversed cyclic loading. [1989] “Moment transfer from concrete slabs to columns.. [1970] Probability. Statistics and Decision for Civil Engineers. and Hanson. A. J.” Earthquake Spectra 12(4). L. R. H. Miranda ACI-ASCE 352. [1968] “Shear and moment transfer between concrete slabs and columns. [1993] “RC ﬂat slab-column subassemblages under lateral loading. and Tibshirani. [1976] “Punching of ﬂat plates under static and dynamic horizontal forces. Y. Durrani. 667–691. Durrani. and Mitchell.” McGraw-Hill.

549–568. H. 113–137. R. and Durrani. Urbana. D.” American Society of Civil Engineers 102(ST3). Moehle. Islam. G. Department of Civil Engineering. [1973] “Limit design of reinforced concrete slab-openings and slab-column connections. Durrani. S.” ASCE Journal of Structural Engineering.” Federal Emergency Management Agency. A. 214 pages. J. Division of Structures and Mechanics. D. M. Structural Research Series No. A. G. J. New Zealand Society for Earthquake Engineering. P. and Sozen.” Journal of Structural Engineering. 125(9). J.” Report UCB/EERC-93/03. 720–730. [1999] “Nonlinear punching shear failure model for interior slab-column connections. Krauss. W. N.Fragility Assessment of Slab-Column Connections 803 Hawkins. Megally.” Report Draft. Islam.. and Wight. [1995] “Evaluation of injection materials for the repair of deep cracks in concrete structures. DC. B. Y. Christchurch.b: Development of fragility curves for concrete frame and shear wall buildings. Morrison. D.” ACI Structural Journal 91(3). P. and Huo. and Hanson. University of Hawaii. and Ghali. Illinois. 1513–1532. Progress report 1973–1974. 490. Hwang. and Moehle. B. S. and Leon. Washington. P. Various paging. H. [1994] “Equivalent frame analysis of ﬂat plate buildings for seismic loading. [1999] “Standardized earthquake loss estimation methodology (HAZUS 99 technical manual). 560 pages. M.” ACI Structural Journal 97(5). [1984] “Preventing progressive collapse of slab structures. 636–644. Johnson. Civil Engineering Studies. J.” ACI Structural Journal 85(6). Luo. Morrison. K. Mitchell. J.” ACI Structural Journal 92(1). Luo. Paciﬁc Conference on Earthquake Engineering. N.” Journal of Structural Engineering.. 278 pages. Christchurch. New Zealand. H.. New Zealand. California.” Journal of Structural Engineering. J. Megally. [1983] “Lateral-load tests of R/C slabcolumn connections. Department of Civil Engineering. R. S. D. S. [1993] “An experimental study of ﬂat-plate structures under vertical and lateral loads.” Technical Report REMR-CS-48. S. [1997] “Chapter 7.” Report UILU-ENG-81-2004. M. D. and Sozen. H.. and Conte. [1995] “Equivalent beam model for ﬂat-slab buildings-Part I: Interior connections. and Park.” Technical Report NCEER 97-0018. E.. University of Illinois at Urbana-Champaign. M. A. I. R. M. J. Hueste. M. 2698–2713. Seattle. Hawaii. ASCE 120(7).” Proc. Hwang. . and Park. M. Berkeley. S. [2003] “Building speciﬁc loss estimation for performance-based design.” PhD Thesis. A. 115–124. J. [2000] “Punching shear design of earthquake-resistant slabcolumn connections. Scanlon. M. Mitchell. [1981] “Response of reinforced concrete plate-column connections to dynamic and static horizontal loads. E. Kreger. DC. [1994] “Design considerations for slab-column connections in seismic zones. A. A. R. University of Washington. Department. Washington. and Sheu. US Army Corps of Engineers. Hirasawa. Enomoto. 303–314. and Ghali. D. and Cook. and Robertson. J. [1976] “Tests on slab-column connections with shear and unbalanced ﬂexure. and Aslani. University of Canterbury. I. Washington. Loss assessment of Memphis buildings. A. Y. ASCE 109(11). University of California.. [1988] “Background to recommendations for reinforced concrete slab-column connections. pp. ASCE 110(7). 2137–2155. National Institute of Building Sciences (NIBS). G. D. Honolulu. 997–1008. [1974] Cyclic Behavior of Six Reinforced Concrete Slab-Column Specimens Transferring Moment and Shear. P. Miranda. of Civil Engineering. [2001] “Slab-column connections with discontinuous slab reinforcement subjected to gravity and cyclic lateral loading.

H.” Earthquake Spectra 15. I. N. [1976] “Slab-column connections subjected to high intensity shears and transferring reversed moments.” UCB/EERC -84/07. 37–45. Department of Civil Engineering.. 605–613. Mitchell. [1992] “Seismic response of interior slab-column connections with shear capitals. 626–638. and Meli. P. 537–556. D. B. W. [1989] “Lateral displacement ductility of reinforced concrete ﬂat plates. W. [1984] “Behavior of interior and exterior ﬂat plate connections subjected to inelastic load reversals. Heaton. N. J. [1988] “Reinforced concrete ﬂat plates under lateral loading: An experimental study including biaxial eﬀects. California. Texas. C. Department of Civil and Environmental Engineering. A. Stanford University. J. CA. Stanford. I. 119. J. Scrivner.” ACI Structural Journal 86(3). . Kanamori. Berkeley. J. L. A. Berkeley. Wey. A. and Hawkins. Kawai. T. Seymonds. California. [1996] “A method for earthquake motion-damage relationships with application to reinforced concrete frames. P.. T. Washington. J. [1990] “Seismic response of connections in indeterminate ﬂat-slab subassemblies. I. [1993] “Gravity load eﬀect on seismic behavior of interior slab-column connections. 262 pages. and Moehle. and Durrani. Singhal. Department of Civil Engineering. Seattle. and Durrani. [1992] “An experimental study of slab-column connections. A. A.” Report SM 76-2. [1999] “TriNet “ShakeMaps”: Rapid generation of instrumental ground motion and intensity maps for earthquakes in southern California. 682–691. N. E. Rice University. and Durrani. J. C. Rosenblueth. P. Pan. A. University of California. D. Robertson. Aslani & E. and Moehle. and Moehle. 266 pages. Miranda Pan. The John A.. [2002] “Cyclic testing of slab-column connections with shear reinforcement. D.” ACI Structural Journal 99(5). University of Washington. D. Robertson. J. Quitoriano. [1986] “The earthquake of 19 September 1985: Causes and eﬀects in Mexico City. Lee.” ACI Structural Journal 94(6). Pan. H.” Concrete International 8(5). 250–258. I. 723–729. Zee. B. Various paging.” ACI Structural Journal 89(6). P. N.. [1997] “Analysis of ﬂat slab structures subjected to combined lateral and gravity loads.. and Kiremidjian. J. D. 23–34. M. A. V.” ACI Structural Journal 89(6). Division of Structures and Mechanics.804 H. 130 pages. University of California. Wald. D.. R.” Report No.” Structural Research at Rice Report 41. Blume Earthquake Engineering Center. H. H. and Enomoto. Robertson.. A. E.” ACI Structural Journal 89(1). and Moehle. Robertson. N. Houston. and Worden. S.” Report UCB/EERC-88/16. J.

Sign up to vote on this title

UsefulNot useful- Group 2
- Ambiguity Aversion
- Philosophical Basis - Rev1 (Submitted)
- Uncertainty AI
- DerKiureghian Paper
- Determin_Probalistic Resources- P90, P50
- Reasoning About The Reliability Of Diverse Two-Channel Systems In Which One Channel Is “Possibly Perfect”
- zhangEPES04
- Interpretation With Expanded Uncertainty 2007 v1
- Interpretation With Expanded Uncertainty 2007 v1
- Practice Final Solutions Sp06
- logic
- Aumann's Agreement Theorem
- Heisenberg Proff
- 10.1.1.195.8221
- Probabilistic Design - The Future of Rock Engineering
- Robust Counterpart
- 82
- Hw1Sol
- math
- Problems With Economics
- Conditional 4
- Section33 Probability Some Basic Terms
- 9781107673311 w 10
- Fundamentals Probability 08072009
- Standard Normal Table (PDF)
- unchance_2013
- skilyrt líkindi frá MIT
- OIML
- Topic Modelling
- Aslani and Miranda - Fagility Slab Column Connections