SEISMIC RESPONSE OF MULTIPLE-ANCHOR CONNECTIONS TO CONCRETE Zhang, Yong-gang*, Richard E. Klingner**, and Herman L. Graves, III*** * Han-Padron Associates, Houston, Texas, USA. ** Dept. of Civil Engineering, The University of Texas at Austin, Austin, Texas, USA. *** U.S. Nuclear Regulatory Commission, Washington, D.C., USA. Abstract Under the sponsorship of the US Nuclear Regulatory Commission, a research program was carried out on the dynamic behavior of anchors (fasteners) in concrete. As part of that program, full-scale seismic tests were conducted at dynamic loading rates on 16 multiple-anchor connections to concrete. Test variables included anchor type, loading history, and the presence of cracks. Multiple-anchor connections designed for ductile behavior in uncracked concrete under static loading, in general behaved in a ductile manner in cracked concrete under dynamic loading. 1. Background A few studies have investigated the behavior of connections under impact loading, seismic loading and reversed loading [1, 2, 3, 4]. Loading patterns primarily involved dynamic loading far below the anchor's ultimate capacity, followed by monotonic loading to failure [3, 4]. Only a few investigations [5, 6, 7] have studied the influence of loading rate on the overall load-displacement behavior of anchors. Some tests have been conducted in cracked concrete or in high-moment regions [1, 3, 5, 8]. To the best of the authors knowledge, the testing program described here is the first published testing on multiple-anchor connections under seismic loading. 2. Anchors, Test Setups and Procedures Based on their use in nuclear applications, this research program involved one wedgetype expansion anchor ( Expansion Anchor II ), with some tests on one normally opening undercut anchor with large bearing area ( UC Anchor 1 ). Anchors were 5/8 in. (16 mm) diameter, installed with an effective embedment of 7 in. (178 mm) to develop ductile response. Expansion Anchor II (EAII) is described in Figure 1. Undercut Anchor 1 () is described in Figure 2. 596
wedge dimple wedge mandrel (cone) D D 1 D 2 l c Figure 1 Expansion Anchor II threaded shank extension sleeve expansion sleeve cone D D 1 D 2 l ef l c Figure 2 Undercut Anchor 1 The target concrete compressive strength for this testing program was 47 lb/in. 2 (32.4 MPa), with a permissible tolerance of ±5 lb/in. 2 (±3.45 MPa) at the time of testing. Aggregate was river gravel. The overall test setup for Task 4 is shown in Figure 3. Reversed cyclic loads were applied to the connection through a loading attachment, shown in Figure 4. The stiff baseplate was 2 in. (51 mm) thick; the flexible one, 1 in. (25 mm). Both baseplates had stiffeners. External load on the connections was measured with a load cell. Tension in each anchor was measured with a force washer placed between the normal washer and the baseplate. Slip of the baseplate was measured with a potentiometer placed against the back of the baseplate, and displacement of the vertical beam was measured at 12 in. (35 mm) from the surface. Loading Attachment Load Cell DCDT Hydraulic Actuator Clamping Beams Concrete Specimen Reaction Frame --- Lab Floor --- Tie-Down Rods On Floor Figure 3 Test setup for multiple-anchor tests of Task 4 597
Stiffeners 14 in. Elevation 6 in. 1 in. 12 in. 9 in. 12 in. Plan View Figure 4 Loading attachment used for Task 4 The prescribed displacement history that simulated earthquake loading was developed by idealizing the attachment as a bilinear singledegree-of-freedom (SDOF) system with a concentrated mass at 12 in. (35 mm) above the concrete surface, sufficient to give a realistic period of vibration, and to produce yielding of the attachment under the selected ground motion. The response of the SDOF system was calculated using as input the earthquake history of El Centro 194 (NS component). The most significant portion, consisting of the first 6. sec. of that record (Figure 5), was used as the prescribed displacement history..8 Displacement Input.6 Displacement (in.).4.2 -.2 -.4 -.6 -.8 2 4 6 8 1 12 14 Time (sec.) Figure 5 History of estimated attachment displacements for seismic tests Some specimens had cracks with an initial width of.12 in. (.3 mm), introduced using wedge-type splitting tubes of high-strength steel. Anchors were installed to the torque specified by the manufacturer. To simulate the reduction of prestressing force in anchors in service due to concrete relaxation, anchors were first fully torqued, then released after about 5 minutes to permit relaxation, and finally torqued again, but only to 5% of the specified torque. For multiple-anchor shear tests, two separate cracks were initiated parallel to the loading direction. Crack widths were monitored during tests, but not controlled. 598
For static tests, the load was applied slowly and monotonically under displacement control. For dynamic tests, the loading pattern was dynamic reversed cyclic loading (Figure 5), applied under displacement control. During tests, the loading sequence was first applied with a maximum displacement of.6 in. (15.2 mm). If the connection did not fail, loading sequences were applied with maximum displacements of 1. in. (25 mm) and then 1.5 in. (38 mm). Further cycles, to a maximum displacement of 15 in. (38 mm), were applied until failure. Before each loading sequence, the anchors were finger-tightened to eliminate any initial lack-of-fit, which would have increased the displacement required to reach any particular load level. Tests are listed in Table 1. For all tests, the edge distance was 5 in. (127 mm), and the embedment was 7 in. (178 mm). Table 1 Test matrix for eccentric shear tests on multiple-anchor connections Test Description Anchor 411 static, 4-anchor group, rigid baseplate, uncracked concrete, e = 12 in. (35 mm) 412 static, 4-anchor group, rigid baseplate, uncracked concrete, e = 18 in. (457 mm) 423 dynamic, 4-anchor group, flexible baseplate, uncracked concrete, e = 12 in. (35 mm) 424 dynamic, 4-anchor group, rigid baseplate, uncracked concrete, e = 12 in. (35 mm) EAII 425 dynamic, 4-anchor group, rigid baseplate, uncracked concrete, e = 12 in. (35 mm) 426 dynamic, 4-anchor group, rigid baseplate, uncracked concrete, e = 18 in. (457 mm) 437 dynamic, 4-anchor group, rigid baseplate, cracked concrete, e = 12 in. (35 mm) EAII 438 dynamic, 4-anchor group, rigid baseplate, cracked concrete, e = 12 in. (35 mm) 439 dynamic, 4-anchor group, rigid baseplate, cracked concrete, e = 18 in. (457 mm) 431 dynamic, 4-anchor group, rigid baseplate, cracked concrete, e = 18 in. (457 mm) EAII 4411 static, near-edge, 4-anchor group, rigid baseplate, uncracked concrete, no hairpins, e = 12 in. (35 mm) 4412 static, near-edge, 4-anchor group, rigid baseplate, uncracked concrete, no hairpins, e = 18 in. (457 mm) 4513 dynamic, near-edge, 4-anchor group, rigid baseplate, uncracked concrete, no hairpins, e = 12 in. (35 mm) 4514 dynamic, near-edge, 4-anchor group, rigid baseplate, uncracked concrete, no hairpins, e = 18 in. (457 mm) 4615 static, near-edge, 4-anchor group, rigid baseplate, uncracked concrete, close hairpins, e = 12 in. (35 mm) 4616 dynamic, near-edge, 4-anchor group, rigid baseplate, uncracked concrete, close hairpins, e = 12 in. (35 mm) 4617 dynamic, near-edge, 4-anchor group, rigid baseplate, uncracked concrete, close hairpins, e = 18 in. (457 mm) 599
3. Test Results In Figures 6 and 7, the load-displacement envelopes of dynamic tests under eccentric shear at 12 in. (35 mm) and 18 in. (457 mm) respectively, are compared with the static tests on the same configuration (Test 411 versus Test 423, and Test 412 versus Test 426). The following observations can be made: 1) The dynamic load-displacement curves follow the static load-displacement curves over most of the displacement range, differing only near the ultimate load. 2) The maximum dynamic capacity was close to the maximum static capacity. It was 7% higher at a 12 in. (35 mm) eccentricity, and 7% smaller at an 18 in. (457 mm) eccentricity. Due to the small number of tests, however, this observation is not definitive. 3) The most significant effect of dynamic reversed cyclic loading is the increase in total displacement, due to spalling of the concrete in front of the anchors, to the gaps between the baseplate and the anchors and between the anchors and the concrete, and to the larger tensile displacement of the anchors under dynamic cyclic loading. Seismic tests on a connection with EAII under dynamic reversed cyclic loading (Test 437) showed large displacements of about 1 in. (25.4 mm) measured at 12 in. (35 mm) above the concrete, although there is no corresponding static test with which this can be compared. In Figures 8 and 9, load-displacement curves for tests with dynamic loading in cracked concrete (Tests 438 and 439) are compared with the corresponding curves for tests with static loading in uncracked concrete (Tests 411 and 412). The dynamic load-displacement envelopes for these tests also follow the static load-displacement curves well, except near the ultimate load. Horizontal Displacement at 12 in. (35 mm) (mm) 6-38.1-25.4-12.7 12.7 25.4 38.1 266.88 Applied Shear (kips) 4 2-2 -4 Static 177.92 88.96-88.96-177.92 Applied Shear (kn) -6-1.5-1 -.5.5 1 1.5-266.88 Horizontal Displacement at 12 in. (35 mm) (in.) Figure 6 Static versus seismic load-displacement curves, multiple-anchor connections, Anchors, at 12 in. (35 mm) eccentricity (Tests 411 and 423) 6
Horizontal Displacement at 12 in. (35 mm) (mm) -63.5-5.8-38.1-25.4-12.7 12.7 25.4 38.1 5.8 63.5 6 266.88 Applied Shear (kips) 4 2-2 -4 Static 177.92 88.96-88.96-177.92 Applied Shear (kn) Figure 7-6 6-2.5-2 -1.5-1 -.5.5 1 1.5 2 2.5 Horizontal Displacement at 12 in. (35 mm) (in.) -266.88 Comparison of static and seismic load-displacement behaviors of multiple-anchor connections with Anchors under shear at 18 in. (457 mm) eccentricity (Tests 412 and 426 respectively) Horizontal Displacement at 12 in. (35 mm) (mm) -63.5-5.8-38.1-25.4-12.7 12.7 25.4 38.1 5.8 63.5 266.88 Applied Shear (kips) 4 2-2 -4 Dynamic Static 177.92 88.96-88.96-177.92 Applied Shear (kn) -6-266.88-2.5-2 -1.5-1 -.5.5 1 1.5 2 2.5 Horizontal Displacement at 12 in. (35 mm) (in.) Figure 8 Comparison of seismic load-displacement behavior of multipleanchor connections with Anchors at 12 in. (35 mm) eccentricity in cracked concrete (Test 438) with static behavior in uncracked concrete (Test 411) 61
Comparison of Test Results and Analytical Predictions Test results were compared with predictions of BDA5, a macro-model program developed at the University of Stuttgart for the static analysis of multiple-anchor connections loaded by eccentric shear [9]. It requires as input data a complete set of load-displacement curves of the anchor under oblique loading at angles from to 9 degrees, and assumes a rigid baseplate. Comparisons were reasonable, and are discussed in more detail in Zhang [1]. 6 Horizontal Displacement at 12 in. (35 mm) (mm) -38.1-25.4-12.7 12.7 25.4 38.1 266.88 Applied Shear (kips) 4 2-2 -4 Dynamic Static 177.92 88.96-88.96-177.92 Applied Shear (kn) -6-266.88-2.5-2 -1.5-1 -.5.5 1 1.5 2 2.5 Horizontal Displacement at 12 in. (35 mm) (in.) Figure 9 Comparison of seismic load-displacement behavior of multipleanchor connections with Anchors at 18 in. (457 mm) eccentricity in cracked concrete (Test 439) with static behavior in uncracked concrete (Tests 412) Comparison of Test Results with Plastic Analysis Methods The Plastic Method [11] and the Modified Plastic Method [7] predict the capacity of multiple-anchor connections with large edge distances, loaded in shear and failing by steel fracture. For the connection with a loading eccentricity of 18 in. (457 mm), the capacities calculated by the Plastic Method and the BDA5 program are very close to the test results. The Modified Plastic Method [7], however, underestimated the static capacity by as much as 1%. For the connection with a loading eccentricity of 12 in. (35 mm), both the Plastic Method and the BDA5 program overestimated the static capacity. The Modified Plastic Method [7] was very close to the test results. 62
4. Conclusions and Recommendations 1) Multiple-anchor connections in uncracked or cracked concrete, with or without edge effects, and with or without hairpins, loaded dynamically under reversed cyclic loading histories representative of seismic response, behaved consistently with the results of previous single- and double-anchor tests of this study. Previous observations regarding the load-displacement behavior, and failure mechanisms of single and double anchors, were applicable in predicting the behavior of complex, multiple-anchor connections under simulated seismic loading. The implications of this are clear. Multiple-anchor connections designed for ductile behavior in uncracked concrete under static loading, will probably still behave in a ductile manner in cracked concrete under dynamic loading. 2) Anchors that show relatively good performance when tested individually in cracked concrete (CIP headed anchors,, and 2 mm diameter Sleeve) will probably also show relatively good performance in multiple-anchor connections subjected to seismic loading. Anchors that show relatively poor performance when tested individually in cracked concrete (Grouted Anchor, EAII, and 1 mm diameter Sleeve) will probably also show relatively poor performance in multiple-anchor connections subjected to seismic loading. 3) Cyclic load-displacement behavior of multiple-anchor connections is accurately bounded by the corresponding static load-displacement envelope, and also by the static load-displacement envelope predicted by the BDA5 program. Dynamic cycling does not significantly influence the fundamental load-displacement behavior of multiple-anchor connections. 4) Under dynamic reversed cyclic loading in both uncracked and cracked concrete, the load-displacement envelopes of multiple-anchor connections with the Anchor basically follow the static curves in uncracked concrete over most displacements, differing only near the ultimate load. Dynamic reversed loading did not significantly affect the maximum dynamic capacity. In uncracked concrete, the connection had larger displacements under reversed dynamic than under static loading. Under dynamic reversed loading, connections in cracked concrete had slightly larger displacements than those in uncracked concrete. 5) Under dynamic reversed cyclic loading, multiple-anchor connections with Expansion Anchor II had very large displacements. In both uncracked and cracked concrete, the connections loaded at 12 in. (35 mm) eccentricity failed by steel fracture. The test in cracked concrete had a larger displacement and smaller capacity than that in uncracked concrete. The connection loaded at an 18 in. (457 mm) eccentricity experienced gross pull-out failure of the anchors. 63
6) The capacity of multiple-anchor connections at large edge distances was predicted with reasonable accuracy by the plastic design procedures [7, 1, 11]. 7) Capacities were reasonably predicted by the BDA5 program [9]. 5. Acknowledgement and Disclaimer This paper presents partial results of a research program supported by the U.S. Nuclear Regulatory Commission (NRC) (NUREG/CR-5434, Anchor Bolt Behavior and Strength during Earthquakes ). The technical contact is Herman L Graves, III, whose support is gratefully acknowledged. The conclusions in this paper are those of the authors only, and are not NRC policy or recommendations. 6. References 1. Cannon, R. W., Expansion Anchor Performance in Cracked Concrete, ACI Journal, Proceedings, Vol. 78, No. 6, November-December 1981, pp. 471-479. 2. Malik, J. B., Mendonca, J. A., and Klingner, R. E., Effect of Reinforcing Details on the Shear Resistance of Short Anchor Bolts under Reversed Cyclic Loading, Journal of the American Concrete Institute, Proceedings Vol. 79, No. 1, January-February 1982, pp. 3-11. 3. Copley, J. D. and E. G. Burdette, Behavior of Steel-to-Concrete Anchorage in High Moment Regions, ACI Journal, Proceedings, Vol. 82, No. 2, March-April 1985, pp. 18-187. 4. Collins, D., R. E. Klingner and D. Polyzois, Load-Deflection Behavior of Cast-in Place and Retrofit Concrete Anchors Subjected to Static, Fatigue, and Impact Tensile Loads, Research Report CTR 1126-1, Center for Transportation Research, The University of Texas at Austin, February 1989. 5. Eibl, J. and E. Keintzel (1989): Zur Beanspruchung von Befestigungsmitteln bei dynamischen Lasten, Forschungsbericht T2169, Institut für Massivbau und Baustofftechnologie, Universität Karlsruhe, 1989. 6. Rodriguez, M., Behavior of Anchors in Uncracked Concrete under Static and Dynamic Loading, M.S. Thesis, The University of Texas at Austin, August 1995. 7. Lotze, D. and Klingner, R. E., Behavior of Multiple-Anchor Connections to Concrete From the Perspective of Plastic Theory, PMFSEL Report No. 96-4, The University of Texas at Austin, March 1997. 8. Eligehausen, R. and Balogh, T., Behavior of Fasteners Loaded in Tension in Cracked Reinforced Concrete, ACI Structural Journal, Vol. 92, No. 3, May-June, 1995, pp. 365-379. 64
9. Li, L., BDA: Programm zur Berechnung des Trag- und Verformungsverhaltens von Gruppenbefestigungen unter kombinierter Schragzug- und Momentenbeanspruchung (Programmbeschreibung), The University of Stuttgart, June 1994. 1. Zhang, Yong-gang, Dynamic Behavior of Multiple-Anchor Connections in Cracked Concrete, Ph.D. Dissertation, Dept. of Civil Engineering, The University of Texas at Austin, August 1997. 11. Cook, R. A. and Klingner, R. E., Ductile Multiple-Anchor Steel-to-Concrete Connections, Journal of Structural Engineering, ASCE, Vol. 118, No. 6, June 1992, pp. 1645-1665. 65