You are on page 1of 10

ACI STRUCTURAL JOURNAL

Title no. 99-S49

TECHNICAL PAPER

Impulsive Seismic Response of Bridge Column-Cap Beam Joints


by Natalie Gibson, Andre Filiatrault, and Scott A. Ashford
The results of dynamic tests on six half-scale reinforced concrete bridge column-cap beam joints are discussed. The loading protocol used included the effects of impulsive near-field ground motions. Cyclic loading with increasing amplitudes followed this impulsive loading. Three of the specimens incorporated nominally ductile reinforcement details typical of pre-1971 construction practice in California. The three other specimens incorporated ductile reinforcement details typical of current design practice in California. The impulsive loading was particularly detrimental to the pre-1971 specimens. Shear failure in the joint was initiated during the impulsive loading, causing early-strength degradation. The joints incorporating ductile reinforcement details suffered minimal damage during the impulsive loading and continued to perform well during the subsequent loading cycles. The increased capacity exhibited by both types of joints during the impulsive loading can be accounted for by modifying the steel and concrete properties for the high strain rate induced during the impulsive loading.
Keywords: dynamic loading; joint; strain; test.

INTRODUCTION Strong ground-motion records obtained during the 1994 Northridge earthquake in California and the 1995 Kobe earthquake in Japan have provided clear evidence on the short duration and impulsive characteristics of near-field ground motions. These characteristics are the result of a directivity pulse phenomenon occurring when the propagation of the fault proceeds at nearly the same rate as the shear wave velocity. The directivity pulse phenomenon can be explained by the simple model illustrated in Fig. 1 (Abrahamson 1998). The fault rupture is divided into a number of discrete subfaults. Beginning at the epicenter, the rupture propagates along the fault in the direction of the arrow toward Site A. If the velocity of the shear waves propagating from the fault is close to the rupture velocity, most of the energy in the forward direction will reach Site A within a short time differential and cause an impulsive ground motion. This phenomenon is similar to the sonic boom in aerodynamics. Forward-directivity effects only occur when the rupture propagates toward the site and the slip direction on the fault is aligned with the site. Not all near-fault locations will experience forward rupturedirectivity effects in a given seismic event. It can be seen in Fig. 1 that Site B, for example, which is located opposite the faulting direction, will experience an increased time differential between the arrivals of shear waves. Thus, the ground motion at Site B will not be impulsive. There is a growing concern that near-field impulsive ground motions can be particularly detrimental to reinforced concrete bridge structures. Impulsive ground motions can cause large ductility demands in a very short time on the structural elements. Also, the effects of the high strain rate induced by impulsive dynamic loading can change the hysteretic 470

Fig. 1Directivity pulse phenomenon. response of the structural elements. It is unclear how bridge structures behave under such impulsive seismic excitations. The objective of this study is to contribute to a better understanding of the seismic response of reinforced concrete bridge structures under impulsive near-field ground motions. For this purpose, the results of dynamic tests on six halfscale reinforced concrete bridge column-cap beam joints are presented. The loading protocol used for these tests included the effects of impulsive near-field ground motions. Cyclic loading with increasing amplitudes followed this impulsive loading. Three of the test specimens incorporated nominally ductile reinforcement details typical of pre-1971 construction practice in California. The three other specimens incorporated modern ductile reinforcement details. RESEARCH SIGNIFICANCE Recent earthquakes in California and Japan have demonstrated the impulsive characteristics of near-field ground motions. It is unclear how bridge structures behave under such seismic excitations. The experimental results obtained herein characterize the impulsive seismic response of bridge column-cap beam joints incorporating nominally or fully ductile reinforcement details. The observation that impulsive loading is particularly detrimental to nominally ductile joints is significant and can
ACI Structural Journal, V. 99, No. 4, July-August 2002. MS No. 01-312 received September 30, 2001, and reviewed under Institute publication policies. Copyright 2002, American Concrete Institute. All rights reserved, including the making of copies unless permission is obtained from the copyright proprietors. Pertinent discussion will be published in the May-June 2003 ACI Structural Journal if received by January 1, 2003.

ACI Structural Journal/July-August 2002

Natalie Gibson is a project engineer with KPFF Consulting Engineers in San Francisco, Calif. She received her MS in structural engineering in 2001 from the University of California, San Diego, Calif. ACI member Andre Filiatrault is a professor of structural engineering at the University of California, San Diego. His research interests include the seismic behavior of structures. Scott A. Ashford is an assistant professor of geotechnical engineering at the University of California, San Diego. His research interests include various aspects of geotechnical earthquake engineering, including the behavior of deep foundations in liquefied sand; the effects of large-velocity pulses on bridge structures; and the earthquake-induced loss of lifeline facilities.

have several impacts on the state of practice, including the reassessment of retrofit priorities in regions susceptible to near-field ground motions and the consideration in design of high strain-rate effects induced by impulsive loading. DESCRIPTION OF TEST SPECIMENS Figure 2 and 3 show the geometry and reinforcement details for the nominally ductile and ductile half-scale reinforced concrete bridge column-cap beam joints considered in this investigation. All specimens were constructed and tested in an inverted position. Each test specimen incorporated half the column height, along with a cap beam representing portions of the surrounding deck and the soffit slab. Identical longitudinal and transverse reinforcement details were incorporated in the column for all test specimens. The circular column had a diameter of 610 mm (24 in.). The height of the column was 1829 mm (6 ft) from the soffit side of the cap beam to the loading axis. This column height corresponds to the distance between the joint interface and a point of inflection assumed to be located halfway up the column. The column was reinforced with

14 No. 7 longitudinal bars (diameter = 22 mm). The transverse column reinforcement was a No. 3 spiral (diameter = 10 mm) spaced at 97 mm (3.8 in.) over the full height of the column. The cap beam was 610 mm (24 in.) deep and 686 mm (27 in.) wide. The length of the beam was 1219 mm (48 in.) from the center of the column to each simple support. All reinforcement was of Grade 60 steel. The nominally ductile specimens (Fig. 2) incorporated reinforcement details in the joint region that were typical of the construction practice used in California prior to the 1971 San Fernando earthquake. The longitudinal bars of the column were anchored into the joint region at a distance of 406 mm (16 in.), corresponding to 16 times the bar diameter. The soffit side of the cap beam was reinforced with four No. 7 and two No. 3 bars. The deck side of the cap beam was reinforced with four No. 7, two No. 3, and three No. 2 (diameter = 6.4 mm) bars. The bars on the deck side ran continuously for the length of the cap beam. On the soffit side, all of the No. 2 longitudinal bars and one of the No. 7 bars of the cap beam were spliced within the joint region (Gibson, Filiatrault, and Ashford 2001). Transverse No. 2 open stirrups were provided at a spacing of 114 mm (4.5 in.) in the cap beam. No joint reinforcement was provided. The column spiral ended at the soffit face of the cap beam, and the transverse stirrups in the cap beam ended at the column face. The ductile specimens (Fig. 3) incorporated reinforcement details in the joint region that were typical of the current (early 1990s) design practice used in California. The longitudinal bars in the column extended fully into the joint region at a distance of 530 mm (21 in.) to meet the longitudinal steel on the deck side of the cap beam. The longitudinal bars of the

Fig. 2Reinforcement details for nominally ductile test specimens. ACI Structural Journal/July-August 2002 471

Fig. 3Reinforcement details for ductile test specimens. Table 1Description of test specimens
Specimen no. 1 2 3 4 5* 6
*Tested

Joint reinforcement details Nominally ductile Ductile Nominally ductile Ductile

Loading protocol Quasistatic Dynamic Quasistatic Dynamic Quasistatic, no pulse

Table 2Compressive strength of concrete at time of testing


Test Joint reinforcement specimen details Structural member 1 Nominally ductile 2 3 Ductile 4 Column Cap beam/joint Column Cap beam/joint Column Cap beam/joint Column Cap beam/joint Compressive strength MPa 40.8 38.1 43.0 39.7 41.4 42.9 43.9 41.2 ksi 5.9 5.5 6.2 5.8 6.0 6.1 6.4 6.0

by McRae, Priestley, and Seible (1994). Tested by Sritharan, Priestley, and Seible (1996).

column were confined transversely in the joint region by a No. 3 spiral spaced at 57 mm (2.25 in.). The cap beam was reinforced with seven No. 7 bars at the top and bottom. The bars ran continuously over the length of the cap beam. Transverse reinforcement of the cap beam was provided by No. 3 closed double stirrups spaced at 172 mm (6.75 in.). Within the joint region, the cap beam was reinforced transversely by two sets of double No. 3 hairpins. Table 1 lists the characteristics of the six specimens considered in this study. Three of the test specimens incorporated nominally ductile reinforcement details, while the other three test specimens incorporated ductile reinforcement details. Specimens 1 through 4 were tested in this study. Specimens 5 and 6 were tested previously under nonimpulsive quasistatic loading by MacRae, Priestley, and Seible (1994) and by Sritharan, Priestley, and Seible (1996), respectively, using 472

the same setup as that used for Specimens 1 through 4. These test results were used herein for comparative purposes. MATERIAL PROPERTIES Compressive tests on concrete cylinders and tensile tests on steel coupons were performed in order to determine the actual material properties of the concrete and reinforcing steel. Table 2 presents the compressive strength of concrete just prior to testing of the first four specimens. Table 3 lists the yield strength and the ultimate strength of steel for each reinforcing bar type used in the test specimens. EXPERIMENTAL SETUP Each test specimen was tested in an inverted position as shown in Fig. 4. Both ends of the cap beam were simply ACI Structural Journal/July-August 2002

Fig. 4Experimental setup. Table 3Yield and ultimate strength of reinforcing steel
Yield strength Bar location Column and cap beam longitudinal Column spirals Joint spirals Cap beam stirrups Bar size No. 7 No. 3 No. 3 No. 2 MPa 437 434 362 489 ksi 63 62 53 71 Ultimate strength MPa 709 679 672 590 ksi 103 98 97 86

connected to the strong floor of the laboratory. Two vertical actuators were used to apply a constant vertical loading of 40 kN (90 kips) simulating the gravity loads on the column. This vertical loading corresponds to approximately 3.5% of the nominal axial-load capacity of the column. The lateral cyclic loading on the column was applied by a displacementcontrolled horizontal actuator mounted between the strong wall of the laboratory and the loading stub at the top of the column. LOADING PROTOCOL Figure 5(a) illustrates the horizontal displacement timehistory applied at the loading stub of the column during the dynamic tests. This loading protocol starts with one high amplitude, short duration, impulsive cycle followed by a cyclic loading sequence with increasing amplitudes. The latter part of the loading protocol was used by MacRae, Priestley, and Seible (1994) and by Sritharan, Priestley, and Seible (1996) for the quasistatic testing of Specimens 5 and 6, respectively (Table 1). The initial impulsive portion of the loading protocol was developed by Cox and Ashford (2001) based on a statistical analysis of 34 near-field historical ground motions containing impulsive characteristics. The records considered originated ACI Structural Journal/July-August 2002

(a)

(b) Fig. 5Loading protocol: (a) complete protocol; and (b) impulsive portion of protocol. 473

Fig. 6Final cracking patterns of test specimens. from earthquakes having moment magnitudes between 6.1 and 7.4, and were measured at stations located within 18 km of the causative fault. The average peak velocity obtained from the impulsive portions of these records was 940 mm/s (37 in./s) and the average duration of these pulses was 2.2 s. It was also found that, typically, only a single pulse occurred with a peak velocity and displacement occurring during the first half-cycle. Following these observations, a nonlinear time-history dynamic analysis of a full-scale prototype model of the test specimens were performed to determine the resulting displacements at the top of the column caused by this impulsive base excitation. The single pulse used as base excitation in these analyses had a peak velocity of 1 m/s (40 in./s) and a duration of 2 s. From these analyses, and using appropriate scale factors, the resulting impulsive portion of the loading protocol was obtained as shown in Fig. 5(b). The positive amplitude of the pulse is 69 mm (2.71 in.), while its negative amplitude is 14 mm (0.55 in.). The duration of the pulse is 474 0.85 s. The maximum velocity induced by the pulse was 405 mm/s (16 in./s). The subsequent cyclic loading was then applied at a constant velocity of 254 mm/s (10 in./s). Note also that a short pause was introduced at the end of each sequence containing three cycles at the same amplitude. The displacement ductility ratio corresponding to the amplitude of each cyclic sequence is also indicated in Fig. 5. This displacement ductility ratio is defined as

= ---y

(1)

where is the amplitude of a given cyclic sequence, and y is the first yield displacement assumed to be equal to 12.4 mm (0.49 in.) for all test specimens. This value is the same as the one taken by MacRae, Priestley, and Seible (1994) and Sritharan, Priestley, and Seible (1996). Note that the ACI Structural Journal/July-August 2002

amplitude of the first impulsive half-cycle of the loading protocol shown in Fig. 5 corresponds to a displacement ductility ratio of approximately 4.0. As shown in Table 1, two specimens were tested under dynamic conditions, while two others were tested under quasistatic conditions. For the quasistatic tests, the same loading protocol shown in Fig. 5 was applied, but at a very slow rate. The last two specimens, Specimens 5 and 6, previously tested by MacRae, Priestley, and Seible (1994) and by Sritharan, Priestley, and Seible (1996), contained only the cyclic loading portion of the loading protocol without the initial impulsive cycle. EXPERIMENTAL RESULTS General behavior of test specimens Figure 6 shows the final cracking patterns of the four specimens tested in this study (Specimens 1 through 4, refer to Table 1). Clearly, the failure modes for both types of specimens are very different. The two specimens with nominally ductile reinforcement details (Specimens 1 and 2) suffered shear failures in the joint region. No longitudinal column bars fractured in either of these two specimens. For Specimen 1, loaded under quasistatic conditions, three main diagonal cracks originated from the soffit side of the cap beam on the north side of the column and propagated to the deck side of the cap beam on the south side of the column. These diagonal cracks appeared during the first impulsive half-cycle of the loading protocol and enlarged as the amplitude of the subsequent cyclic loading increased. In the other impulsive half-cycle, only one main diagonal crack occurred. The cap-beam longitudinal bars buckled out of plane in the joint region near the soffit side. Specimen 2, loaded dynamically, exhibited similar cracking patterns with more concrete spalling in the joint region than the statically loaded Specimen 1. The failure modes were similar for both ductile test specimens (Specimens 3 and 4). Because of the improved reinforcement details in the joint region, the full flexural capacity of the column was developed without premature shear failure of the joint. Failure of Specimen 3, loaded under quasistatic condition, occurred when several longitudinal bars in the columns buckled in the plastic hinge region and, subsequently, fractured in tension. Failure of the dynamically loaded Specimen 4 was much more dramatic. During the second positive cycle at a ductility level of 8, the spiral bar fractured in the plastic-hinge region on the north side face of the column. After fracture of the spiral, the column concrete core began to degrade quickly, followed by buckling in compression and fracture in tension of several longitudinal bars in the plastichinge region. The only major difference between the two ductile specimens was the behavior of the joint region. The quasistatically loaded Specimen 3 suffered much more cracking in the joint region than did the dynamically loaded Specimen 4. Two longitudinal bars on the deck side of the cap beam of Specimen 3 yielded slightly at a ductility level of 6, whereas no yielding was observed in Specimen 4. Global hysteretic response Figure 7 and 8 present the global top of column loaddisplacement hysteresis loops obtained for the nominally ductile and ductile specimens, respectively. It can be seen that the initial impulsive loading had a much greater influence for the nominally ductile specimens (Fig. 7) than for the ductile specimens (Fig. 8). ACI Structural Journal/July-August 2002

Fig. 7Global hysteretic response of nominally ductile specimens. During the impulsive portion of the loading protocol, the nominally ductile Specimens 1 and 2 developed higher capacities than Specimen 5 tested quasistatically by MacRae, Priestley, and Seible (1994) without the impulsive portion of the loading protocol. After the impulsive portion of the loading protocol, however, the capacities of both Specimens 1 and 2 drop by approximately 50%. Also, both specimens exhibit very pinched hysteresis loops with substantially lower stiffness and strength than Specimen 5. This result indicates that the joint shear failure occurred during the impulsive portion of the loading protocol for the nominally ductile specimens, leaving limited stiffness and strength for the remaining cyclic loading. For the ductile specimens, however, the impulsive portion of the loading protocol has little effect on their behaviors under the remaining cyclic loading. Similar stiffness, 475

Fig. 9Moment-curvature hysteretic response of nominally ductile specimens. near the soffit face of the cap beam. To investigate this flexural behavior, various sections along the column were instrumented by pairs of linear potentiometers that measured the average elongation and contraction of the north and south faces of the column along a given gage length. At a given section, the displacement measured by the north potentiometer N and the south potentiometer S can be converted to an average curvature by

N S = -----------------hclg
Fig. 8Global hysteretic response of nominally ductile specimens. strength, and good energy-dissipation capacities can be observed for Specimens 3, 4, and 6. The capacity developed during the first impulsive half-cycle of loading is significantly higher for the specimens tested under dynamic conditions than for the specimens tested under quasistatic conditions. This phenomenon is observed for both nominally ductile and ductile test specimens. This increased capacity is a result of the high strain rate induced during the impulsive loading. A high rate of loading causes an increase in the apparent strength of the concrete in compression and of the steel in tension. Finally, the capacities developed by the dynamically loaded Specimens 2 and 4 are similar. Moment-curvature hysteretic response The ductile behavior of a bridge column-cap joint assembly involves the formation of a flexural plastic hinge in the column 476

(2)

where hc is the horizontal distance between the potentiometers, and lg is the gage length. The curvature measured along the column section nearest to the soffit side of the cap beam was plotted against the maximum moment in the column to obtain the momentcurvature hysteresis loops in the potential plastic region of the column. The maximum moment at the base of the column was calculated from the horizontal force in the actuator multiplied by the distance from the actuator loading point to the soffit face of the cap beam (1829 mm [72 in.]). Figure 9 and 10 present the resulting moment-curvature hysteresis loops for the nominally ductile and ductile specimens tested in this study, respectively. The maximum curvature reached in the columns of the nominally ductile specimens was about 65% of the maximum curvature measured in the ductile specimens. The dynamically tested and nominally ductile Specimen 2 exhibited very little curvature in the column. The impulsive portion of the loading protocol caused similar responses in all specimens. After the impulsive ACI Structural Journal/July-August 2002

Fig. 11Cumulative energy absorbed.

Fig. 10Moment-curvature hysteretic response of ductile specimens. portion of the loading protocol, the ductile specimens dissipated much more energy in flexure than did the nominally ductile specimens. The curvature values are greater, and the loops are wider. The nominally ductile specimens, conversely, have very thin moment-curvature hysteresis loops. Energy dissipation The energy-dissipation capacity of each specimen can be assessed by computing the absorbed energy Ea during each test Ea =

Fig. 12Strain rate time-history during impulsive portion of loading protocol, Specimen 4, longitudinal bar on south face of column. loading protocol, however, caused a similar rapid initial increase in the energy absorbed for the first four specimens. For the ductile joints, Specimen 4, when tested under dynamic conditions, absorbed more energy than the specimens tested under quasistatic conditions (Specimens 3 and 6). This can be explained by the apparent increased capacity arising from the high strain rate induced during dynamic loading. The behavior of the ductile Specimen 6, tested under quasistatic conditions and without the impulsive portion of the loading protocol, was similar to the Specimen 3 until a ductility level of = 8. For higher ductility levels, the rate of energy absorption of Specimen 6 was reduced compared with that of Specimen 3. At = 8 for Specimen 6, it was observed by Sritharan, Priestley, and Seible (1996) that significant shear deformations occurred in the joint region accompanied by a sudden drop in horizontal capacity. Much more concrete degradation and spalling was seen in the joint region of Specimen 6 than in that of Specimens 3 and 4. From these results, it can be inferred that the impulsive portion of the loading protocol is not overly detrimental to the overall behavior of the ductile test specimens. For the nominally ductile specimens, the situation is different. Specimens 1 and 2, which were tested with the impulsive portion of the loading protocol, absorbed less energy than Specimen 5, which was tested without the impulsive portion of the loading protocol. This result indicates that the impulsive loading has a significant effect on reducing the energydissipation capacity of nominally ductile specimens due to the early brittle shear failure in the joint region induced during the impulsive loading. 477

Vcol dcol

(3)

where Vcol is the applied horizontal shear force in the column, and col is the horizontal displacement at the top of the column. Figure 11 presents the cumulative energy absorbed by the six test specimens considered in this study. The crest of each ductility increment represents the most deformed configuration of a specimen under maximum shear force. The corresponding trough occurs when a specimen is unloaded and the elastic portion of the strain energy is recovered. The numerical value at that point corresponds to the cumulative energy dissipated. Note that the absorbed energy is plotted on a pseudo-time scale. This pseudo-time was determined so that the peaks of the cycles occurred simultaneously for the quasistatic and dynamic tests. As was expected, the ductile specimens (Specimens 3, 4, and 6) absorbed much more energy than the nominally ductile specimens (Specimens 1, 2, and 5). The impulsive portion of the ACI Structural Journal/July-August 2002

Fig. 13Experimental and predicted envelope of momentcurvature hysteresis loops for Specimen 4; strain rate effects neglected. Table 4Amplification factors for strain rate of 130,000 /s (0.13/s) on concrete and steel properties
Material Property Reference Wright and Hall (1964) Soroushian and Choi (1987) Yield strength Steel Kaneko (1997) Ammann and Nussbaumer (1995) Wakabayashi et al. (1984) Soroushian and Choi (1987) Ultimate strength Kaneko (1997) Ammann and Nussbaumer (1995) Youngs modulus Concrete Ammann and Nussbaumer (1995) Mean Amplification amplification factor factor 1.47 1.38 1.15 1.11 1.02 1.11 1.05 1.02 1.24 1.26 1.22 1.24 1.06 1.22

Fig. 14Experimental and predicted envelope of momentcurvature hysteresis loops for Specimen 4; strain rates included. such analysis include the geometry of the cross section and the material properties defining the stress-strain curves of the concrete in compression and the steel in tension and compression. The concrete model must also include the effect of confinement provided by the transverse reinforcement. Figure 13 compares the experimental moment-curvature hysteresis loops of Specimen 4 with the prediction of a moment-curvature analysis (Gibson, Filiatrault, and Ashford 2001) using the computer program SEQMC (SEQAD Inc. 2000). This program determines the moment and curvature characteristics of planar reinforced concrete sections and their effect on the idealized moment-rotation behavior of a flexural hinge at the end of a prismatic concrete column. The curvature values measured in the column were modified in Fig. 13 to account for the additional rotation due to strain penetration into the joint (Priestley, Seible, and Calvi 1996, Sritharan, Priestley, and Seible 1996). The analysis was conducted with the static material properties given in Table 2 and 3 and for the confined concrete model proposed by Mander, Priestley, and Park (1988). The model predicts reasonably well the initial stiffness and strength of the specimen for the cyclic loading following the impulsive portion of the loading protocol. The model, however, underpredicts the strength developed by the specimen during the first impulsive half-cycle of the loading protocol. It is believed that this is the result of the high strain rate that changes the material properties of the concrete and steel. Within the last few decades, a significant research effort has been devoted to the behavior of steel and concrete under high strain rates. This research effort has been summarized recently (Filiatrault and Holleran 2001). Most of these investigations reported that the yield strength and tensile (ultimate) strength of steel increase linearly with logarithmically increasing strain rate. It was also recognized that the strain rate had negligible effects on the elastic modulus and the stiffness of steel in the strain-hardening range. On the concrete side, there is still a lack of information concerning the strain-rate effect on the seismic behavior of concrete members. Many past studies have been carried out for impact conditions, with strain rate several magnitudes above earthquake-type strain rate. Most of the research results available indicate increases in elastic modulus, tensile strength, flexural strength, and compressive strength with increasing strain rate. Table 4 presents amplification factors for the steel and concrete properties as predicted by various empirical expressions available in the literature for the mean peak strain rate ACI Structural Journal/July-August 2002

Ammann and Compres- Nussbaumer (1995) sive strength Wakabayashi et al. (1984)

Induced strain rate The column longitudinal bars were instrumented with electrical strain gages at several locations in the plastichinge region. The strain rate in the longitudinal bars can be computed as the change in strain over each time step. Figure 12 shows, for example, the strain-rate time-history measured in a longitudinal bar of Specimen 4 located on the south face of the column at a distance of 152 mm (6 in.) from the soffit face of the cap beam. This bar experienced a maximum strain rate of over 150,000 /s (0.15/s) during the first impulsive half-cycle. From an analysis of several instrumented sections in the plastic-hinge region of the column, the mean maximum strain rate induced in Specimen 4 was estimated at 130,000 /s (0.13/s). ANALYSIS The flexural behavior of the ductile specimens can be approximated by a moment-curvature analysis of the column cross section in the plastic-hinge region. The input parameters of 478

of 130,000 /s (0.13/s) induced in Specimen 4. These amplification factors represent the ratio between dynamic and static material properties. The mean values across the various expressions are also given in Table 4. The static properties given in Table 2 and 3 were modified by the mean amplification factors given in Table 4 and then used in the same moment-curvature analysis described above (Gibson, Filiatrault, and Ashford 2001). Figure 14 compares the experimental moment-curvature hysteresis loops of Specimen 4 with the prediction of this new momentcurvature analysis that incorporates material properties modified for strain-rate effects. The model predicts very well the initial stiffness of the specimen, as well as the flexural capacity achieved during the first impulsive half cycle of loading. SUMMARY AND CONCLUSIONS This study has characterized the impulsive seismic response of bridge column-cap beam joints incorporating nominally ductile or fully ductile reinforcement details. Based on the results obtained, the following conclusions can be drawn: The impulsive seismic loading was particularly detrimental to the specimens incorporating nominally ductile reinforcement details in the joint region. Shear failure in the joint was initiated during the impulsive loading, causing early strength and stiffness degradations; The joints incorporating ductile reinforcement details suffered minimal damage during impulsive loading and continued to perform in a ductile manner during the subsequent loading cycles; and The increased capacity exhibited by both types of specimens during impulsive loading can be accounted for in a moment-curvature analysis of the column in the plastic-hinge region by modifying the steel and concrete properties for the high strain rate induced during impulsive loading. ACKNOWLEDGMENTS
This research project was supported primarily by the Pacific Earthquake Engineering Research (PEER) Center through the Earthquake Engineering Research Centers Program of the National Science Foundation under Award EEC-9701568. The opinions, findings, conclusions, and recommendations expressed herein are those of the authors and do not necessarily reflect the views of PEER.

S y

= = = = = = =

displacement measured by south potentiometer displacement amplitude of given cyclic sequence first yield displacement strain average curvature over gage length displacement ductility factor microstrain (106)

REFERENCES
Abrahamson, N., 1998, Seismological Aspects of Near-Fault Ground Motions, Proceedings of the Fifth Caltrans Seismic Research Workshop, California Department of Transportation Engineering Service Center, Sacramento, Calif., 5 pp. Ammann, H., and Nussbaumer, H., 1995, Behavior of Concrete and Steel Under Dynamic Actions, Vibrations Problems in Structures, Practical Guide, Chapter F, Birkhuser Verlag, Boston, pp. 177-183. Cox, K., and Ashford. S., 2001, Characterization of a Large Velocity Directivity Pulse for Laboratory Testing, Structural Systems Research Report No. SSRP-2001/11, Department of Structural Engineering, University of California, San Diego, La Jolla, Calif., 62 pp. Filiatrault, A., and Holleran, M., 2001, Stress-Strain Behavior of Reinforcing Steel and Concrete under Seismic Strain Rates and Low Temperatures, Materials and Structures, V. 34, pp. 235-239. Gibson, N.; Filiatrault, A.; and Ashford, S. A., 2001, Performance of Bridge Joints Subjected to a Large Velocity Pulse, Structural Systems Research Project Report No. SSRP-2001/10, Department of Structural Engineering, University of California, San Diego, La Jolla, Calif, 100 pp. Kaneko, H., 1997, Influence of Strain-Rate on Yield Ratio, Kobe Earthquake Damage to Steel Moment Connections and Suggested Improvement, JSSC Technical Report No. 39/96, 125 pp. MacRae, G. A.; Priestley, M. J. N.; and Seible, F., 1994, Santa Monica Viaduct RetrofitLarge-Scale Column-Cap Beam Joint Transverse Test Preliminary Report, Report No. TR-94/02, Structural Systems Research Project, University of California at San Diego, La Jolla, Calif., 117 pp. Mander, J. B.; Priestley, M. J. N.; and Park, R., 1988, Theoretical Stress-Strain Model for Confined Concrete, Journal of Structural Engineering, ASCE, V. 114, pp. 1804-1826. Priestley, M. J. N.; Seible, F.; and Calvi, G. M., 1996, Seismic Design and Retrofit of Bridges, Wiley-Interscience Publications, New York, 686 pp. SEQAD, Inc., 2000, SEQad Moment-Curvature Analysis Program, SC Solutions, http://www.scsolutions.com/seqmc.html. Soroushian, P., and Choi, K., 1987, Steel Mechanical Properties at Different Strain Rates, Journal of Structural Engineering, ASCE, V. 113, No. 4, pp. 863-872. Sritharan, S.; Priestley, M. J. N.; and Seible, F., 1996, Seismic Response of Column/Cap Beam Tee Connections with Cap Beam Prestressing, Report No. SSRP-96/09, Structural Systems Research Report, University of California at San Diego, Calif, 311 pp. Udagavva, K.; Takanashi, K.; and Kato, B., 1984, Effects of Displacement Rates on the Behavior of Steel Beams and Composite Beams, Eighth World Conference on Earthquake Engineering, San Francisco, Calif., V. IV, pp. 177-184. Wakabayashi, M.; Nakamura, T.; Iwai, S.; and Hayashi, Y., 1984, Effects of Strain Rate on the Behavior of Structural Members, Eighth World Conference on Earthquake Engineering, San Francisco, Calif., V. IV, pp. 491-498. Wright, R. N., and Hall, W. J., 1964, Loading Rate Effects in Structural Steel Design, Journal of the Structural Division, ASCE, V. 90, No. 55, pp. 11-37.

NOTATION
Ea hc lg Vcol col N = = = = = = absorbed energy horizontal distance between two potentiometers gage length horizontal shear force in column horizontal displacement at top of column displacement measured by north potentiometer

ACI Structural Journal/July-August 2002

479

You might also like