Journal of Earthquake Engineering, Vol. 9, No.

6 (2005) 777804 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 dierent 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 rst 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 dierent 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 Modied 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 specic 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 13 storeys tall, mid rise frames that are 47 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 specic reinforced concrete frames as part of a loss assessment study of Memphis buildings. Similarly, Erberik and Elnashai [2004] computed fragility curves for 5-storey at-slab structures. In the probabilistic framework being developed at the Pacic Earthquake Engineering Research (PEER) Center, damage assessment is performed in two steps. In the rst 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 oor accelerations, etc.) computed in the rst 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. Denition of Damage States In this study, four discrete damage states of slab-column connections have been introduced. These damage states have been dened based on specic 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).

Fragility Assessment of Slab-Column Connections

779

DS1

DS2

DS3

DS4

Fig. 1. Schematic view of the four damage states considered for slab-column connections: (a) Light cracking; (b) severe cracking; (c) punching shear failure; and loss of vertical carrying capacity.

Figure 1(a) shows a schematic view of a typical cracking pattern on the top face of the slab at this damage state. These crack widths can be tolerated in reinforced concrete in humid or moist air conditions [ACI 224R, 2001]. Hence, 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, 1998]. The purpose of this repair action is to improve the aesthetic appearance of the slab or to provide an additional barrier against water inltration into the slab. This repair action is not intended to provide any increase in strength and stiness to the cracked slab. DS 2 Severe Cracking: This damage state involves extensive cracking with crack widths between 0.3 mm (0.013 in) and 2 mm (0.08 in). Figure 1(b) shows a typical cracking pattern on the top face of the slab at this damage state. For this level of cracking most concrete repair guidelines suggest epoxy injection [ACI 224.1, 1984; ACI 546, 1996; ACI 548, 1997]. The repair action associated with this damage state typically consists of crack epoxy injection which provides a structural binding agent to ll the crack and adhere to the substrate material [FEMA 308,

780

H. Aslani & E. Miranda

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

Table 1. Specimen properties Slab Column

Properties of slab-column specimens used in this study.

Specimen number fy (Mpa) (6) 354.4 410.9 373.7 315.8 315.8 460.0 462.6 455.0 460.0 460.0 462.6 462.6 475.8 475.8 475.8 322.7 330.3 335.1 317.2 339.9 290.0 0.02 0.02 0.02 0.05 0.11 0.04 0.31 0.33 0.30 0.86 0.81 Yes Yes Yes 0.33 0.45 0.45 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 0.18 0.23 0.22 Yes Yes Yes 22.5 22.5 22.5 22.6 22.5 22.7 0.04 0.05 30 15 30 15 33.0 32.8 31.9 29.7 29.7 34.8 23.4 22.0 32.3 23.2 26.5 26.5 36.7 33.3 39.9 45.8 35.1 33.9 34.9 35.2 36.3 (7) (8) (9) (10) Section (cm cm) fc (Mpa) (4) 7.50 7.50 8.75 8.75 8.75 15.00 15.00 15.00 15.00 15.00 15.00 15.00 15.00 15.00 15.00 7.50 7.50 7.50 7.50 7.50 7.50 45.8 35.1 33.9 34.9 35.2 36.3 36.7 33.3 39.9 23.2 26.5 26.5 34.8 23.4 22.0 32.3 31.9 29.7 29.7 33.0 32.8 (5)

References

Label

Thickness (cm)

fc (Mpa)

Vg V0

Bottom reinforcement through the column fy (Mpa) (11) 354.4 410.9 373.7 315.8 315.8 460.0 462.6 455.0 460.0 460.0 462.6 462.6 475.8 475.8 475.8 322.7 330.3 335.1 317.2 339.9 290.0

(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. [1974]

S1 S2 S3 S4

10 11 12

Seymonds et al. [1976]

S6 S7 S8

13 14 15

Ghalli et al. [1976]

SM 0.5 SM 1.0 SM 1.5

Fragility Assessment of Slab-Column Connections

16 17 18 19 20 21

Morrison and Sozen [1981]

S1 S2 S3 S4 S5 D1

781

782

Table 1. Specimen properties Slab Column

(Continued)

Specimen number fy (Mpa) (6) 327.0 355.0 435.1 472.0 472.0 472.0 472.0 0.37 0.36 0.18 0.19 Yes Yes Yes Yes 27 27 27 27 27 27 27 27 0.21 Yes 13.5 13.5 37.2 33.3 33.3 31.4 31.4 0.04 0.04 25 25 25 25 33.9 36.5 (7) (8) (9) (10) (11) 327.0 355.0 435.1 472.0 472.0 472.0 472.0 Section (cm cm) fc (Mpa) fy (Mpa) (4) 7.50 7.50 6.00 12.00 12.00 12.00 12.00 33.3 33.3 31.4 31.4 37.2 33.9 36.5 (5)

References

Label

Thickness (cm)

fc (Mpa)

Vg V0

Bottom reinforcement through the column

H. Aslani & E. 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. [1989]

29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45

Hawkins et al. [1989]

6AH 9.6AH 14AH 6AL 9.6AL 14AL 7.3BH 9.5BH 14.2BH 7.3BL 9.5BL 14.2BL 6CH 9.6CH 14CH 6CL 14CL

15.00 15.00 15.00 15.00 15.00 15.00 11.25 11.25 11.25 11.25 11.25 11.25 15.00 15.00 15.00 15.00 15.00

31.3 30.7 30.3 22.7 28.9 27.0 22.2 19.8 29.5 18.1 20.0 20.5 52.4 57.2 54.7 49.5 47.7

415.0 415.0 415.0 415.0 415.0 415.0 415.0 415.0 415.0 415.0 415.0 415.0 415.0 415.0 415.0 415.0 415.0

0.43 0.48 0.53 0.72 0.68 1.38 0.38 0.48 0.42 0.69 0.71 0.80 0.37 0.41 0.48 0.55 0.74

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

31.3 30.7 30.3 22.7 28.9 27.0 22.2 19.8 29.5 18.1 20.0 20.5 52.4 57.2 54.7 49.5 47.7

415.0 415.0 415.0 415.0 415.0 415.0 415.0 415.0 415.0 415.0 415.0 415.0 415.0 415.0 415.0 415.0 415.0

Table 1. Specimen properties Slab Column

(Continued)

Specimen number fy (Mpa) (6) 415.0 415.0 415.0 415.0 415.0 415.0 415.0 415.0 415.0 415.0 45.0 415.0 415.0 500.2 524.9 524.9 524.9 0.18 0.51 0.37 0.18 0.85 0.65 0.52 Yes Yes Yes Yes 25 25 25 25 25 25 25 25 0.35 0.49 0.61 0.81 0.43 0.45 0.52 0.63 0.80 0.80 0.47 0.50 0.45 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) 30.0 31.7 28.7 24.3 25.9 33.8 31.2 25.9 18.1 26.5 24.7 26.3 27.0 33.0 32.2 30.8 39.3 Section (cm cm) fc (Mpa) (4) 15.00 15.00 15.00 15.00 15.00 15.00 15.00 15.00 15.00 15.00 15.00 15.00 15.00 11.25 11.25 11.25 11.25 33.0 32.2 30.8 39.3 30.0 31.7 28.7 24.3 25.9 33.8 31.2 25.9 18.1 26.5 24.7 26.3 27.0 (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 10.2FHI 10.2FHO 14FH 6FLI 10.2FLI 10.2FLO 9GH2 9.6GH0.5 9.6GH3

415.0 415.0 415.0 415.0 415.0 415.0 415.0 415.0 415.0 415.0 415.0 415.0 415.0 500.2 524.9 524.9 524.9

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

784

Table 1. Specimen properties Slab Column

(Continued)

H. Aslani & E. Miranda

Specimen number fy (Mpa) 372.3 372.3 372.3 372.3 524.9 457.6 457.6 457.6 457.6 379.0 413.7 413.7 413.7 413.7 413.7 413.7 420.0 0.25 0.23 0.28 0.47 0.24 0.28 0.20 0.08 No No No No No No No Yes 0.32 0.32 0.49 0.56 Yes Yes Yes Yes 0.25 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 0.37 0.23 0.27 No No No No 25 25 25 25 25 25 25 25 35.3 25.7 24.6 19.1 39.3 35.1 35.1 15.0 15.0 20.7 29.6 32.3 24.1 26.3 18.9 39.2 35.4 Section (cm cm) fc (Mpa) 1 2 3 4 11.25 8.00 8.00 8.00 8.00 11.25 11.40 11.40 11.40 11.40 11.40 11.40 11.5 35.4 29.6 32.3 24.1 26.3 18.9 39.2 20.7 35.1 35.1 15.0 15.0 39.3 11.25 11.25 11.25 11.25 35.3 25.7 24.6 19.1

References

Label

Thickness (cm)

fc (Mpa)

Vg V0

Bottom reinforcement through the column

fy (Mpa) 372.3 372.3 372.3 372.3 524.9 457.6 457.6 457.6 457.6 379.0 413.7 413.7 413.7 413.7 413.7 413.7 420.0

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. [1993]

75

Luo et al. [1994]

76 77 78 79 80 81 1C

Johnson and Robertson [2001], from Enomoto and Robertson [2001]

ND1C ND4LL ND5XL ND6HR ND7LR ND8B

82

Robertson et al. [2002]

Information was not available

Fragility Assessment of Slab-Column Connections

785

to properly establish the level of interstorey drift at which the damage state was observed. The later situation was particularly common for the rst two damage states (light cracking and severe cracking), primarily because until very recently, earthquake provisions have mainly been concerned with life safety and not damage control. Only ve investigations included detailed information about the level of interstorey drift at which light cracking was observed. For these specimens, the reported interstorey drift at which clearly visible cracking occurred, was accompanied with a signicant reduction in lateral stiness. Therefore, in order to expand the number of data points associated with this damage state, the interstorey drift at which a 30% or more sudden reduction in lateral stiness was observed in the hysteresis loop was considered. Figure 2 presents a graphic representation of a point in loading history where this sudden reduction occurred. Table 2, summarises the interstorey drifts associated with this rst damage state. Interstorey drifts presented in Table 2 have been calculated using the slab centreline to slab centreline interstorey height. The indirect way of determining the interstorey drift associated with rst damage state was only used in specimens where a good quality hysteresis loop at early loading cycles was included in the report. Consequently, Table 2 reports interstorey drifts associated with light cracking for only 43 specimens. It can be seen that the rst damage state in slab-column connections occurs at very low deformations raging from 0.19% to 0.8% drift ratios. A few studies systematically reported cracking patterns and crack widths at various levels of lateral deformation to be used for the second damage state. Careful study of such investigations [Pan and Moehle, 1988; Robertson and Durrani, 1990] suggests that damage state two, typically occurs when top steel reinforcement is at yield. Therefore, in order to gather more data points associated with the second damage state, it was assumed that this damage state occurs at peak interstorey

Lateral Load

Significant change in stiffness

IDRDS1

IDR

Fig. 2. Typical drop in lateral stiness observed in the hysteresis loop of slab-column specimens used to identify the rst damage state, light cracking, in some of the specimens.

786

H. Aslani & E. Miranda Table 2. Interstorey drift ratios used to develop the fragility functions.
IDRDS3 (%) (4) 3.80 5.80 4.72 4.22 4.81 3.75 1.82 2.18 2.46 1.39 1.04 *** 6.00 2.70 2.70 4.47 2.89 4.25 4.30 4.70 3.11 2.89 4.32 3.30 1.60 1.50 3.70 3.50 2.27 1.70 1.36 1.48 1.08 0.97 1.62 1.97 1.27 1.62 1.73 1.48 2.59 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) ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** 0.64 0.54 0.40 0.80 ** ** ** 0.50 0.50 0.50 0.50 ** 0.19 0.19 0.28 0.28 0.40 0.74 0.54 0.43 0.71 0.43 0.50 0.70 IDRDS2 (%) (3) ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** 1.50 ** 1.00 1.00 ** ** ** ** ** ** ** ** 1.63 1.27 1.20 0.82 1.00 1.48 0.76 1.00 1.07 1.00 1.14 1.40 IDRDS3 (%) (4) 1.25 1.25 1.93 0.91 2.39 1.65 1.48 0.77 0.80 1.25 0.91 1.08 0.79 1.09 0.80 1.48 0.57 5.00 1.00 1.50 4.00 0.90 1.20 1.40 4.50 2.00 4.50 4.70 5.0 3.48 2.53 2.40 3.16 5.00 8.00 5.00 2.00 4.00 5.00 5.00 3.50 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) ** ** 0.43 0.30 0.28 0.65 0.22 0.32 0.43 0.42 0.28 0.35 ** ** ** 0.31 0.31 0.28 0.22 0.22 0.31 0.40 0.36 0.28 0.20 0.50 0.25 0.50 ** ** ** ** ** ** ** ** ** ** ** ** **

IDRDS2 (%) (3) ** ** 1.11 1.11 1.11 0.76 0.93 0.76 0.90 0.69 0.49 0.69 ** ** ** ** ** ** ** ** 0.89 0.89 0.89 0.98 0.77 0.75 0.80 0.79 ** ** ** ** ** ** ** ** ** ** ** ** **

4.90 4.90

2.00 3.00 5.00

2.60 1.60 5.00 3.50

5.00 5.00 6.00 4.70 3.70 3.30 3.80

5.00

4.00

*Centreline dimensions were used to compute IDR **Information was not available ***Punching shear failure did not occur ****IDR corresponds to a exural 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. Figure 3 schematically illustrates the latter situation. The third column in Table 2 lists the interstorey drifts that were identied or inferred corresponding to this

Fragility Assessment of Slab-Column Connections

787

Lateral Load

IDR DS 2

IDR

Permanent residual displacement

Fig. 3. The rst occurrence of a residual lateral drift in the hysteresis loop of a slab-column connection specimen used to identify the second damage state, severe cracking, in some of the specimens.

damage state. Interstorey drifts associated with severe cracking were obtained for 33 specimens which range between 0.49% and 1.63% drift ratios. With the exception of ve specimens in which shear failure was not observed, 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. The fourth column in Table 2 lists interstorey drift ratios at which punching shear failures occurred in various specimens. It can be seen that drifts at which these failures have occurred exhibit a very large dispersion ranging from 0.57% to as large as 8% drift ratios. Most experimental studies have investigated parameters that inuence the moment capacity and deformation at which punching shear failure occurred. Hence, in most cases testing was stopped after punching shear failure was observed. There are no investigations in which testing continued until LVCC occurred. However, 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. The main goal of these subsequent cycles was the investigation of post-punching behaviour of the slab-column connections. In this study, 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. Maximum drift ratios at which testing were stopped in dierent specimens range between 1.6% and 6.0%.

4. Fragility Functions As shown in Table 2, the interstorey drift at which each damage state in slab-column connections was observed shows relatively large variations from one specimen to

788

H. Aslani & E. Miranda

another. 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. In particular, a drift-based fragility function provides the probability of experiencing or exceeding a particular damage state conditioned on the peak interstorey drift. At each damage state, a cumulative frequency distribution function was obtained by plotting ascending-sorted interstorey drifts at which the damage state was experimentally observed against (i0.5)/n, where i is the position of the peak interstorey drift and n is the number of specimens. 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. Figures 4(a) and 4(b) show cumulative distribution functions for the rst and second damage states, respectively. As can be seen in Fig. 4(a), the rst damage state (light cracking) was observed in half of the specimens when an interstorey drift ratio of about 0.4% was reached. The second damage state (severe cracking) occurred in half of the specimens when the interstorey drift ratio reached 0.9%, Figure 4(b). Also plotted in these gures are tted lognormal cumulative distribution functions of the interstorey drift ratios given by P (DS dsi |IDR = idr) = Ln(idr) Ln (IDR) , Ln IDR (1)

where P (DS dsi |IDR = idr) is the probability of experiencing or exceeding damage state i, IDR is the median of the interstorey drift ratios, IDRs, at which

(a)

(b)

Fig. 4. Fragility functions tted to interstorey drift ratios corresponding to the rst and second damage states in slab-column connections.

Fragility Assessment of Slab-Column Connections

789

damage state i was observed, Ln IDR is the standard deviation of the natural logarithm of the IDRs, and is the cumulative standard normal distribution. It can be seen that the lognormal distribution ts the data relatively well. Parameters for the tted lognormal probability distribution are given in Table 3. In order to verify if the cumulative distribution function could be assumed as lognormally distributed, a Kolmogorov-Smirnov goodness-of-t test [Benjamin and Cornell, 1970] was conducted. Also shown in Figs. 4(a) and 4(b) are graphical representations of this test for 10% signicance levels. The hypothesis that the assumed cumulative probability distribution adequately ts the empirical data is accepted if all data points lie between the two gray lines. It can be seen that for both damage states this assumption is adequate. Empirical and tted cumulative distribution functions for the third damage state are shown in Fig. 5. Also shown in the gure are the 10% signicance levels of the Kolmogorov-Smirnov goodness-of-t test which indicate that the lognormal assumption is also valid for the third damage state. The corresponding parameters

Table 3. Statistical parameters estimated for interstorey drift ratios corresponding to the four damage states in slab-column connections. Damage state (1) DS1: DS2: DS3: DS4: Light cracking Severe cracking Punching shear failure Loss of vertical carrying capacity IDR(%) (2) 0.40 0.95 2.00 4.28 Ln IDR (3) 0.39 0.25 0.62 0.36 Number of specimens (n) (4) 43 33 77 18

P (DSds3|IDR) 1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0 0.00 0.02 0.04 IDR
Fig. 5. Fragility function tted to interstorey drift ratios corresponding to the third damage state, punching shear failure, in a slab-column connection.
Lognormal fit K-S test,10% significance

DS 3

ln IDR

= 0.62

Data

0.06

0.08

790

H. Aslani & E. Miranda

for the cumulative distribution of this damage sate are given in Table 3. The dispersion parameter (Ln IDR ) which is approximately comparable to the coecient of variation of the data is shown in Column 3 of Table 3 for each damage state. It can be seen that for the third damage state Ln IDR is 0.62 which is 1.6 times larger than the level of dispersion corresponding to the rst damage state and 2.5 times larger than the dispersion corresponding to the second damage state. 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. Experimental research has shown that the deformation capacity of slab-column connections is a function of the level of gravity load. In particular, Pan and Moehle [1988], Moehle et al. [1988], 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 , where Vg is the vertical shear that acts on the slab critical section dened at a distance d/2 from the column face in which d is the average slab eective depth, and the normalising shear force V0 is the theoretical punching shear strength in the absence of moment transfer in the connection, which according to ACI-352 [1988] is given by V0 = 1 1 + 6 3c fc b0 d 1 3 fc b0 d, (2)

where c is the ratio of long-to-short cross-sectional dimensions of the supporting column, fc is the compressive strength of concrete in MPa and not to exceeding 40 MPa, b0 is the perimeter length of the slab critical section in mm, and d is the slab eective width in mm. For most columns encountered in practice the maximum limit shown in Eq. (2) governs. More recently, Hueste and Wight [1999] proposed a trilinear equation to estimate the drift at which punching shear failure occurs. On the basis of the above observations, a fragility surface was developed for the third damage state, in which the probability of experiencing a punching shear failure was computed as a function of the peak IDR and the vertical shear ratio. Column seven in Table 1 lists the gravity shear ratio in all specimens where punching shear failure was observed. Figure 6(a) shows the drift at which punching shear failure occurred as a function of the vertical shear ratio, Vg /V0 . Also shown in the gure is the trilinear variation of the drift capacity at punching shear proposed by Hueste and Wight [1999]. This gure shows results from 77 specimens, which represent 44 data points in addition to those originally gathered by Hueste and Wight. In the proposed fragility surface, the probability of experiencing punching shear failure is computed with Eq. (1), 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 = 0.014 + 0.03 exp 4440 Vg V0
8

(3)

Fragility Assessment of Slab-Column Connections

791

IDRDS3 0.08
Data

0.07 0.06 0.05 0.04 0.03 0.02 0.01 0.00 0.0 0.2 0.4

Hueste et al. (1999) This study, Eq. (3)

0.6 Vg

0.8

1.0

V0

(a)

(b)

Fig. 6. (a) Variations of interstorey drift ratio at punching shear failure, IDR DS3 , with changes in the level of gravity shear, Vg /V0 , in slab-column connections; (b) variations of the dispersion of the IDR D S3 with changes in Vg /V0 .

Figure 6(b) shows the variations of the dispersion of IDRs at which punching shear failure occurred around IDRDS3 , as a function of the gravity load ratio. It can be seen that the level of dispersion signicantly increases as the level of gravity shear ratio increases. Points shown in this graph were obtained using a moving window analysis with a gravity load ratio width of 0.2 moving at increments of 0.05. This variation of dispersion with changes with the level of gravity load ratio was approximated as follows: Ln IDR DS3 = 0.62 0.4 exp 11.6 Vg V0
2.9

(4)

Figure 6(b) also shows the variation of dispersion computed with Eq. (4), 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. Figure 7(a) shows the fragility surface resulting from the use of Eqs. (1), (3) and (4). Figure 7(b) shows fragility functions for the third damage state for slabcolumn connections with gravity shear ratios equal to 0.1, 0.3 and 0.5. Comparison of Figs. 5 and 7(b) shows that computing the two parameters of the lognormal distribution as a function of the gravity shear ratio using Eqs. (3) and (4) leads to much better estimations of the probability of experiencing or exceeding a damage corresponding to a punching shear failure. For example, in a slab-column connection with a vertical shear ratio of 0.1 subjected to an interstorey drift ratio of 0.02, the probability of experiencing or exceeding a punching shear failure is essentially

792

H. Aslani & E. Miranda

(a)

(b)

Fig. 7. Proposed fragility surfce to account for high levels of specimen-to-specimen variability for the third damage state: (a) 3-dimensional presentation; (b) 2-dimensional presentation.

zero, while if the eect of the gravity load ratio is neglected (as done in Fig. 5), this probability would be estimated as 46%, hence, signicantly overestimating the fragility of the connection. Similarly, if a slab-column connection with a vertical shear ratio of 0.5 is subjected to the same level of drift, there is a very high probability of experiencing or exceeding a punching shear failure, 75%, whereas if the eect of the gravity load ratio is neglected, as done in Fig. 5, this probability would be estimated as 46%, underestimated by 29%. In order to obtain an estimate of the probability of experiencing LVCC, for each specimen that was subjected to further cyclic loading after the punching shear failure was observed, the following ratio was computed = IDRDS4 , 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. This ratio of drifts can be interpreted as an amplication factor acting on the drift corresponding to punching shear failure in order to provide a conservative estimate of the drift at which LVCC occurs. Figure 8 shows a cumulative frequency distribution of , P ( < ), corresponding to the 18 specimens in which post-punching shear failure was studied. Also shown in this gure is a lognormal t of this parameter with median equal to 1.23 and logarithmic standard deviation of 0.2, and 10% signicance levels of the Kolmogorov-Smirnov goodness-of-t test. As shown in this gure all data points lie within the condence bands, indicating that it is adequate to assume that this ratio is also lognormally distributed.

Fragility Assessment of Slab-Column Connections

793

Fig. 8. Lognormal cumulative distribution function tted to the cumulative frequency distribution of .

The interstorey drift ratio at which LVCC occurs is assumed to be lognormally distributed with median and logarithmic standard deviations given by IDRDS4 = 1.23 IDRDS3 , Ln IDRDS4 =
2 Ln IDRDS3 0.3Ln IDRDS3 + 0.04.

(6) (7)

The value of 0.3 in Eq. (7) corresponds to the correlation between the drift at which punching shear failure was reported and the amplication factor . Figure 9(a) presents the fragility surface corresponding to the LVCC damage state computed with Eqs. (1), (6), and (7). Figure 9(b) shows fragility functions corresponding to LVCC for slab-column connections with vertical shear ratios of 0.1, 0.3 and 0.5. It can be seen that if the vertical shear ratio has a low value such as 0.1, LVCC is not likely to occur provided that IDRs are smaller than 3.8%. However, if one wants to avoid the possibility of having LVCC at high levels of gravity shear ratio, such as 0.5, the IDR needs to be limited to 0.6%. As can be seen in Fig. 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. Furthermore neglecting the eect of gravity load can lead to results that in some cases are too conservative while in other are unconservative. For example, in a slab-column connection with a vertical shear ratio of 0.1 subjected to an IDR of 0.04, the probability of experiencing a loss of vertical carrying capacity is approximately 2%. However, if the eect of the gravity load ratio is neglected this probability would be estimated as 55% which is signicantly larger. Similarly, if a slab-column connection with a vertical shear ratio of 0.5 is subjected to the same level of IDR, there is a 97% probability

794

H. Aslani & E. Miranda

(a)

(b)

Fig. 9. Fragility surface developed for the loss of vertical carrying capacity damage state in slab-column connections; (a) 3-dimensional presentation, (b) 2-dimensional presentation.

of experiencing LVCC, whereas if the eect of gravity load ratio is neglected this probability would be 55%, underestimated by 42%. The drift level at which LVCC damage state occurs in a slab-column connection is not only inuenced by the level of gravity load, but also by whether or not the slab has longitudinal steel reinforcement that is continuous through the column reinforcement cage. In particular, slab-column connections with continuous bottom reinforcement may resist signicant vertical loads following initial punching failure of the slab which leads to a signicant increase in the drift capacity associated with experiencing LVCC. Column 8 in Table 1 summarises the information on whether the slab bottom reinforcement in the various specimens passed through the column reinforcement cage. It can be seen that of 11 specimens that did not contain any continuous bottom reinforcement, post-punching behaviour was only studied in 4 specimens. Results from these 4 specimens indicate that, even when there is no continuous bottom reinforcement, the connection does not loose its vertical carrying capacity immediately after experiencing punching shear failure, since top slab reinforcement, though less reliable and ecient than bottom slab reinforcement, is capable of resisting vertical load following initial punching [Pan and Moehle, 1992]. For example, even though specimen 69 had no bottom slab reinforcement passing through the column, 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. 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, results of the four specimens with no continuous bottom reinforcement do not show a clear trend of decreased drift capacity compared to

Fragility Assessment of Slab-Column Connections

795

that of the specimens with continuous bottom reinforcement. However, this should not be interpreted as the bottom reinforcement not increasing the drift at which LVCC occurs, and simply that the test setup, or loading protocols did not permit this to be observed in the laboratory. As mentioned before, 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. 5. 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 specic damage state when it is subjected to a certain level of IDR. This probability is a primary input when performing loss estimation in buildings [Miranda and Aslani, 2003] and can be estimated as the dierence 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 , i=m P (DS < dsi |IDR = idr)

(8)

where i = 0 corresponds to the state of no damage in the component, P (DS dsi |IDR = idr) is the fragility function for the ith damage state of a component which is computed form Eq. (1), and m is the number of damage states dened for the component, and for the case of slab-column connections is 4. In order to apply Eq. (8) to a specic slab-column connection, it is rst required to estimate fragility curves corresponding to all the damage states dened in the connection. Figures 10(a) and 10(c) present fragility functions corresponding to the four damage states dened for slab-column connections at two levels of gravity shear ratio, e.g. at two dierent locations in a building. Fragility functions presented in these gures can then be used in Eq. (8) to estimate the probability of being at each damage state as a function of the level of IDR in the slab-column connection. Figures 10(b) and 10(d) show the probability of being at each damage state for slabcolumn connections with gravity shear ratios equal to 0.2 and 0.5, respectively. As can be seen in Fig. 10(b), for a connection with a gravity load ratio of 0.2 and at 1% drift, the probability that the connection is in the rst damage state is 40% while there is a 58% probability for the connection to be in the second damage state. The probability that the connection has not experienced any physical damage is, therefore, 2%. At 5% drift, however, the connection for sure has experienced physical damage and there is a 30% probability that the connection is in the second damage state, 35% to be in the third damage state and 35% to be in the last damage state. As the level of gravity shear ratio increases, punching shear failure and LVCC damage states occur at signicantly lower levels of interstorey drift ratio, as can be seen in Fig. 10(d).

796

H. Aslani & E. Miranda

P(DSds P(DS ds i | IDR) 1.0 0.8 0.6 0.4 0.2 0.0 0.000

P(DS=ds i | IDR) 1.0 0.8 0.6 0.4

DS1 DS2 DS3 DS4

DS1 DS2 DS3 DS4

Vg V0

= 0.2
0.2 0.0 0.000

Vg V0

= 0.2

0.025

0.050 IDR

0.075

0.100 0.

0.025

0.050 IDR

0.075

0.100

(a)
P(DSds P(DS ds i | IDR) 1.0 0.8 0.6 0.4 0.2 0.0 0.000
P( =ds P(DS=ds i | IDR) R) ) 1.0 0.8 0.6 0.4

(b)

DS1 DS2 DS3 DS4

DS1 DS2 DS3 DS4

Vg V0

= 0.5
0.2 0.0 0.000

Vg V0

= 0.5

0.025

0.050 IDR

0.075

0.100

0.025

0.050 IDR

0.075

0.100

(c)

(d)

Fig. 10. (a), (c) Component-specic fragility curves for slab-column connections with gravity shear ratios equal to 0.2 and 0.5, respectively; (b), (d) Probability of being at each damage state for the two connections computed with Eq. (8).

6. Epistemic Uncertainty on Fragility Functions In addition to specimen-to-specimen variability, another source of uncertainty that need to be considered in estimating the level of damage in slab-column connections is epistemic uncertainty. Two sources of epistemic uncertainty were incorporated in this study. The rst source is the uncertainty caused by using fragility functions whose parameters have been obtained from a limited number of specimens. This source of epistemic uncertainty is referred here as nite-sample uncertainty. The second source is the uncertainty produced by the fact that damage observations in

Fragility Assessment of Slab-Column Connections

797

the specimen are only collected at peak values of the loading protocol. Therefore, 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. This source of uncertainty is associated with the drift increment in the loading protocol used to test each specimen. A quantitative measure for nite-sample uncertainty can be obtained by computing condence intervals of the statistical parameters of the deformation levels corresponding to each damage state. Dierent methods can be used to estimate the condence intervals of a statistical parameter. If the underlying probability distribution is normal or lognormal, conventional statistical methods can be used [Crow et al., 1960; Benjamin and Cornell, 1970]. For cases in which the underlying probability distribution is not normal or lognormal, a generic method of nding the condence interval is to use re-sampling techniques, such as bootstrap statistics [Efron and Tibshirani, 1993]. Since it was veried that the fragility curves of a slab-column connection can be assumed lognormal, a conventional approach was used to estimate condence intervals of the median and logarithmic standard deviation of the drifts corresponding to dierent damage states. The condence intervals of the median of a lognormally distributed sample can be approximated as follows [Crow et al., 1960]: IDR exp z/2 Ln IDR , n (9)

where z/2 is the value in the standard normal distribution such that the probability of a random deviation numerically greater than z/2 is , and n is the number of data points. The number of slab-column specimens used to develop fragility functions are given in the last column of Table 3. Condence intervals for the logarithmic standard deviation of the data are nonsymmetric and can be computed as [Crow et al., 1960]:
2 (n 1)Ln IDR 2 /2,n1 1/2

and

2 (n 1)Ln IDR 2 1/2,n1

1/2

(10)

2 where 2 /2,n1 is the inverse of the distribution with n 1 degrees of freedom 2 and a probability of occurrence of /2, similarly, 2 /2,n1 is the inverse of the distribution with n1 degrees of freedom and a probability of occurrence of 1/2. For each damage state, lower and upper condence intervals of IDR and Ln IDR were estimated using Eqs. (9) and (10). These condence intervals were used to build a condence band on the original fragility functions. 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. Once these drift increments were collected, a bias and dispersion on the empirical median drift capacity was estimated by assuming that the median capacity is a normally distributed random

798

H. Aslani & E. Miranda

variable. Finally, the two sources of epistemic uncertainty were combined assuming that they are uncorrelated. Figure 11 presents the eects of epistemic uncertainty on the fragility curves for each damage state. The black line in each graph corresponds to the fragility curve in the absence of epistemic uncertainty, while the grey bands correspond to a 90% condence band on the fragility curves when both sources of epistemic uncertainty have been considered. The corresponding parameters are listed in Table 4. Please note that the second source of uncertainty in addition to increasing the

P(DSds 1|IDR) 1.0 0.9 0.8 0.7 0.6 Original 0.5 fragility 0.4 0.3 Epistemic 0.2 uncertainty 0.1 envelope 0.0 0.000 0.003 0.006 0.009 0.012 0.015 IDR
(a)

P(DSds2|IDR) 1.0 0.9 0.8 0.7 0.6 Original 0.5 fragility 0.4 0.3 Epistemic 0.2 uncertainty 0.1 envelope 0.0 0.000 0.005 0.010 0.015 0.020 0.025 IDR
(b)

P(DSds3|IDR) 1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0 0.00

P(DSds3|IDR) 1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0 0.00

Vg V0

= 0.2

Vg V0

= 0.5

Original fragility Epistemic uncertainty envelope 0.02 0.0 4 0.06 IDR

(c)

Original fragility Epistemic uncertainty envelope 0.02 0.04 0.06 IDR

(d)

0.0 8

0.10

0.08

0.10

Fig. 11. Incorporating epistemic uncertainty to fragility functions of slab-column connections; (a) Damage state 1, (b) Damage state 2, (c), (d) Damage state 3 With Vg /V0 = 0.2 and 0.5, (e), (f) Damage state 4 with Vg /V0 = 0.2 and 0.5.

Fragility Assessment of Slab-Column Connections

799

P (DS4ds4| IDR) 1.0 0.9 0.8 Vg = 0.2 0.7 V0 0.6 Original 0.5 fragility 0.4 0.3 Epistemic 0.2 uncertainty 0.1 envelope 0.0 0.00 0.02 0.04 0.06 0.08 0.10 0.12 IDR
(e) Fig. 11.

P (DS4ds4| IDR) 1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0 0.00

Vg V0

= 0.5

Original fragility Epistemic uncertainty envelope 0.02 0.04 0.06 IDR

(f)

0.08

0.10

(Continued)

condence band on the median drift also decreases the drift capacity (i.e. shifts the fragility curve to the left) As shown in these graphs, the inuence of epistemic uncertainty is signicant and needs to be accounted for when performing sensitivity studies in loss estimation [Aslani and Miranda, 2004]. For example, a 50% probability of experiencing or exceeding the second damage state can occur anywhere from 0.61% to 0.95% drift ratio, as can be seen in Fig. 11(b). Furthermore, epistemic uncertainty can cause important changes in the probability of experiencing or exceeding a damage state. For example, for a slab-column connection with a gravity shear ratio equal to 0.5, it can be seen in Fig. 11(f) that at 2% interstorey drift ratio, probability of experiencing or exceeding LVCC varies anywhere from 64% to 95%.

7. 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. In these fragility functions the damage is estimated as a function of the peak interstorey drift imposed on the connection. These new fragility information can be used in estimating damage at the component level and with damage states associated to specic repair actions. Furthermore, by using results from experimental studies, the study has identied and quantied the uncertainties associated with estimating damage in slab-column connections. Experimental data from 16 investigations conducted in the last 36 years, with a total of 82 slab-column specimens was used to develop fragility functions for slab-column connections.

800

H. Aslani & E. Miranda

Table 4. Statistical parameters estimated to incorporate epistemic uncertainty for interstorey drift ratios corresponding to the damage states in slab-column connections at 90% condence interval. Parameters considering both sources of epistemic uncertainty IDR (4) 0.0027 1.35 0.0027 0.74 0.0076 1.25 0.0076 0.80 (IDR DS3 0.0024 ) 1.21 (IDR DS3 0.0024 ) 0.82 (IDR DS4 0.0032 ) 1.18 (IDR DS3 0.0024 ) 0.85 Ln IDR (5) 0.39 + 0.09 0.39 0.06 0.25 + 0.07 0.25 0.04 Ln IDRDS3 + 0.10 Ln IDRDS3 0.07 Ln IDRDS4 + 0.14 Ln IDRDS4 0.08

Parameters considering only nite sample epistemic uncertainty IDR (2) 0.39 + 0.09 0.39 0.06 0.25 + 0.07 0.25 0.04 Ln IDRDS3 + 0.10 Ln IDRDS3 0.07 Ln IDRDS4 + 0.14 Ln IDRDS4 0.08 (3) Ln IDR

Damage state

(1)

DS1: Light cracking

0.0033 1.10 0.0033 0.90

DS2: Severe cracking

0.009 1.07 0.009 0.93

DS3: Punching shear failure

IDRDS3 1.12 IDR DS3 0.89

DS4: Loss of vertical carrying capacity

IDR DS4 1.15 IDR DS4 0.87

*IDR DS3 and IDR DS4 are computed from Eq. (3) and Eq. (6), respectively. **Values 0.0024 and 0.0032 correspond to epistemic uncertainty caused by drift increments in the loading protocol used to test each specimen for damage states 3 and 4, respectively.

Fragility Assessment of Slab-Column Connections

801

Two sources of uncertainty were incorporated in the fragility functions: Specimen-to-specimen variability and epistemic uncertainty. Specimen-to-specimen variability corresponds to the fact that dierent specimens experience the various damage states at levels of deformation that in general are dierent for each specimen. The specimen-to-specimen variability was incorporated by developing drift-based fragility functions at each damage state. It was observed that, for damage states involving punching shear failure and loss of vertical carrying capacity, 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. Two kind of epistemic uncertainty were considered. The rst one is associated to the fact that the parameters of the fragility functions have been obtained from a limited number of specimens, 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 inuenced by the drift increment used in the loading protocol that was employed to test of each specimen. Quantitative measures for each of these two kinds of epistemic uncertainty were developed using statistical procedures. Results indicate that in some cases the eects of epistemic uncertainty on the probability of experiencing each damage state are signicant and therefore should be incorporated in probabilistic damage assessment.

Acknowledgements The authors would like to acknowledge the support by the Pacic Earthquake Engineering Research (PEER) Center with support from the Earthquake Engineering Research Centers Program of the National Science Foundation under Award Number EEC-9701568. The authors also would like to express their gratitude for comments and suggestions from the director of the PEER Center, Professor Jack P. Moehle at University of California at Berkeley, and Professor Ian N. Robertson at University of Hawaii at Manoa, during the course of this investigation.

References
ACI 224.1. [1984] Causes, evaluations, and repair of cracks, ACI Journal 81(3), 211230 Detroit, Michigan. ACI 224R. [2001] Control of cracking in concrete structures, ACI Technical Document, Detroit, Michigan, 46 pages. ACI 546. [1996] Concrete repair guide. (ACI 546R-96), ACI Technical Document, Detroit, Michigan, 126 pages. ACI 548. [1997] Guide for the use of polymers in concrete (ACI 548.1R-97), ACI Technical Document, Detroit, Michigan, 29 pages.

802

H. Aslani & E. Miranda

ACI-ASCE 352. [1988] Recommendations for design of slab-column connections in monolithic reinforced concrete structures (ACI 352.1R-89), ACI Structural Journal 85(6), 675696. ASCE-ACI 426. [1974] The shear strength of reinforced concrete members-slabs, ASCE Journal of the Structural Division 100(ST8), 15431591. Aslani, H. and Miranda, E. [2004] Component-level and system-level sensitivity study for earthquake loss estimation, Proc. 13th World Conference on Earthquake Engineering, Canadian Association for Earthquake Engineering under auspices of International Association of Earthquake Engineering, Vancouver, BC, Canada. ATC-13. [1985] Earthquake damage evaluation data for California, Applied Technology Council, Redwood City, California. Benjamin, J. and Cornell, C. A. [1970] Probability, Statistics and Decision for Civil Engineers, McGraw-Hill, New York, 684 pages. Crow, E. L., Davis, F. A. and Maxeld, M. W. [1960] Statistics Manual, Dover Publication, New York, 288 pages. Dilger, W. and Cao, H. [1991] Behavior of slab-column connections under reversed cyclic loading, Proc. 2nd Int. Conf. on High-Rise Buildings, China. Dovich, L. and Wight, J. K. [1996] Lateral response of older at slab frames and the economic eect on retrot, Earthquake Spectra 12(4), 667691. Durrani, A. J. and Du, Y. [1992] Seismic resistance of slab-column connections in existing non-ductile at-plate buildings, Technical Report NCEER-92-0010, State University of New York at Bualo, New York, 75 pages. Durrani, A. J., Du, Y. and Luo, Y. H. [1995] Seismic resistance of non-ductile slabcolumn connections in existing at-slab buildings, ACI Structural Journal 92(4), 479487. Erberik, M. A. and Elnashai, A. S. [2004] Fragility analysis of at-slab structures, Engineering Structures 26, 937948. Efron, B. and Tibshirani, R. J. [1993] An Introduction to the Boot-Strap, Chapman and Hall, New York, 436 pages. Enomoto, B. and Robertson, I. N. [2001] Hysteretic Modeling of Flat Slab-Column Connections, Research report, UHM/CE/01/03, Department of Civil Engineering, University of Hawaii, Honolulu, Hawaii, 235 pages. Farhey, D. N., Moshe, A. A. and Yankelevsky, D. Z. [1993] RC at slab-column subassemblages under lateral loading, Journal of Structural Engineering, ASCE 119(6), 19031916. FEMA 308, [1998] Repair of earthquake damaged concrete and masonry wall buildings, Prepared by the Applied Technology Council (ATC-43 Project) for the Federal Emergency Management Agency, Washington, DC (FEMA Publication No. 308) 65 pages. Ghali, A., Elmasri, M. Z. and Dilger, W. [1976] Punching of at plates under static and dynamic horizontal forces, ACI Journal, Title No. 73-47 73(10), 566572. Hamburger, R. O. [1996] Department store, Northridge fashion center, 1994 Northridge Earthquake Buildings Case Studies Project, Seismic safety commission, State of California, Proposition 122: Product 3.2, Sacramento, California, pp. 83100. Hanson, N. W. and Hanson, J. M. [1968] Shear and moment transfer between concrete slabs and columns, Journal of the Portland Cement Association, Research and development laboratories 10(1), 216. Hawkins, N. M., Aibin, B. and Yamazaki, J. [1989] Moment transfer from concrete slabs to columns, ACI Structural Journal 86(6), 705717. Hawkins, N. M. and Mitchell, D. [1979] Progressive collapse of at plate structures, ACI Structural Journal 76(7), 775808.

Fragility Assessment of Slab-Column Connections

803

Hawkins, N. M., Mitchell, D. and Sheu, M. S. [1974] Cyclic Behavior of Six Reinforced Concrete Slab-Column Specimens Transferring Moment and Shear, Progress report 19731974, Division of Structures and Mechanics, Department. of Civil Engineering, University of Washington, Seattle, Washington, Various paging. Hueste, M. B. D. and Wight, J. K. [1999] Nonlinear punching shear failure model for interior slab-column connections, ASCE Journal of Structural Engineering, 125(9), 9971008. Hwang, H. and Huo, J. R. [1997] Chapter 7.b: Development of fragility curves for concrete frame and shear wall buildings. Loss assessment of Memphis buildings, Technical Report NCEER 97-0018, pp. 113137. Hwang, S. J. and Moehle, J. P. [1993] An experimental study of at-plate structures under vertical and lateral loads, Report UCB/EERC-93/03, Department of Civil Engineering, University of California, Berkeley, California, 278 pages. Islam, S. and Park, R. [1973] Limit design of reinforced concrete slab-openings and slab-column connections, PhD Thesis, University of Canterbury, Christchurch, New Zealand, 214 pages. Islam, S. and Park, R. [1976] Tests on slab-column connections with shear and unbalanced exure, American Society of Civil Engineers 102(ST3), 549568. Johnson, G., Enomoto, B. and Robertson, I. N. [2001] Slab-column connections with discontinuous slab reinforcement subjected to gravity and cyclic lateral loading, Report Draft, Department of Civil Engineering, University of Hawaii, Honolulu, Hawaii. Krauss, P. D., Scanlon, J. M. and Hanson, M. A. [1995] Evaluation of injection materials for the repair of deep cracks in concrete structures, Technical Report REMR-CS-48, US Army Corps of Engineers, Washington, DC. Luo, Y. H. and Durrani, A. J. [1995] Equivalent beam model for at-slab buildings-Part I: Interior connections, ACI Structural Journal 92(1), 115124. Luo, Y. H., Durrani, A. J. and Conte, J. P. [1994] Equivalent frame analysis of at plate buildings for seismic loading, Journal of Structural Engineering, ASCE 120(7), 21372155. Megally, S. and Ghali, A. [1994] Design considerations for slab-column connections in seismic zones, ACI Structural Journal 91(3), 303314. Megally, S. and Ghali, A. [2000] Punching shear design of earthquake-resistant slabcolumn connections, ACI Structural Journal 97(5), 720730. Miranda, E. and Aslani, H. [2003] Building specic loss estimation for performance-based design, Proc. Pacic Conference on Earthquake Engineering, New Zealand Society for Earthquake Engineering, Christchurch, New Zealand. Mitchell, D. and Cook, W. D. [1984] Preventing progressive collapse of slab structures, Journal of Structural Engineering, ASCE 110(7), 15131532. Moehle, J. P., Kreger, M. E. and Leon, R. [1988] Background to recommendations for reinforced concrete slab-column connections, ACI Structural Journal 85(6), 636644. Morrison, D. G. and Sozen, M. A. [1981] Response of reinforced concrete plate-column connections to dynamic and static horizontal loads, Report UILU-ENG-81-2004, Civil Engineering Studies, Structural Research Series No. 490, University of Illinois at Urbana-Champaign, Urbana, Illinois. 560 pages. Morrison, D. G., Hirasawa, I. and Sozen, M. A. [1983] Lateral-load tests of R/C slabcolumn connections, Journal of Structural Engineering, ASCE 109(11), 26982713. National Institute of Building Sciences (NIBS), [1999] Standardized earthquake loss estimation methodology (HAZUS 99 technical manual), Federal Emergency Management Agency, Washington, DC.

804

H. Aslani & E. Miranda

Pan, A. D. and Moehle, J. P. [1988] Reinforced concrete at plates under lateral loading: An experimental study including biaxial eects, Report UCB/EERC-88/16, University of California, Berkeley, California, 262 pages. Pan, A. D. and Moehle, J. P. [1989] Lateral displacement ductility of reinforced concrete at plates, ACI Structural Journal 86(3), 250258. Pan, A. D. and Moehle, J. P. [1992] An experimental study of slab-column connections, ACI Structural Journal 89(6), 626638. Robertson, I. N. [1997] Analysis of at slab structures subjected to combined lateral and gravity loads, ACI Structural Journal 94(6), 723729. Robertson, I. N. and Durrani, A. J. [1990] Seismic response of connections in indeterminate at-slab subassemblies, Structural Research at Rice Report 41, Department of Civil Engineering, Rice University, Houston, Texas, 266 pages. Robertson, I. N. and Durrani, A. J. [1993] Gravity load eect on seismic behavior of interior slab-column connections, ACI Structural Journal 89(1), 3745. Robertson, I. N., Kawai, T., Lee, J. and Enomoto, B. [2002] Cyclic testing of slab-column connections with shear reinforcement, ACI Structural Journal 99(5), 605613. Rosenblueth, E. and Meli, R. [1986] The earthquake of 19 September 1985: Causes and eects in Mexico City, Concrete International 8(5), 2334. Seymonds, D. W., Mitchell, D. and Hawkins, N. M. [1976] Slab-column connections subjected to high intensity shears and transferring reversed moments, Report SM 76-2, Division of Structures and Mechanics, Department of Civil Engineering, University of Washington, Seattle, Washington. Various paging. Singhal, A. and Kiremidjian, A. S. [1996] A method for earthquake motion-damage relationships with application to reinforced concrete frames, Report No. 119, The John A. Blume Earthquake Engineering Center, Department of Civil and Environmental Engineering, Stanford University, Stanford, CA. Wald, D. J., Quitoriano, V., Heaton, T. H., Kanamori, H., Scrivner, C. W. and Worden, C. B. [1999] TriNet ShakeMaps: Rapid generation of instrumental ground motion and intensity maps for earthquakes in southern California, Earthquake Spectra 15, 537556. Wey, E. H. and Durrani, A. J. [1992] Seismic response of interior slab-column connections with shear capitals, ACI Structural Journal 89(6), 682691. Zee, H. L. and Moehle, J. P. [1984] Behavior of interior and exterior at plate connections subjected to inelastic load reversals, UCB/EERC -84/07, University of California, Berkeley, California, 130 pages.