Model Precast Concrete Beam-to-Column Connections Subject to Cyclic Loading

Transcription

1 Model Precast Concrete Beam-to-Column Connections Subject to Cyclic Loadin Geraldine S. Cheok Research Structural Enineer Buildin and Fire Research Laboratory National nstitute of Standards and Technoloy Gaithersbur, Maryland H. S. Lew, Ph.D. Chief Buildin and Fire Research Laboratory National nstitute of Standards and Technoloy Gaithersbur, Maryland Experimental results of eiht -scale model precast concrete beam-to-column connections are presented. The test specimens consisted of interior connections desined in accordance with the 1985 Uniform Buildin Code provisions for Seismic Zones 2 and 4. These tests constitute the second and third phases of a multi-year test proram bein conducted at the National nstitute of Standards and Technoloy. The objective of the test proram is to develop uidelines for an economical precast beam-tocolumn connection for reions of hih seismicity. Variables considered in the research proram include location of the post-tensionin steel, the use of post-tensionin bars vs. prestressin strands and fully bonded vs. partially bonded strands. Specimens were subjected to reversed cyclic loadin accordin to a prescribed displacement history. Comparisons were made between the behavior of precast concrete specimens and the monolithic specimens tested previously in Phase. These comparisons were based on connection strenth, ductility and enery dissipation characteristics. Comparison of results with the monolithic test specimens indicates that the post-tensioned precast concrete specimens had comparable connection strenths, hiher ultimate displacement ductilities and total enery dissipation to failure, but lower enery dissipation per cycle. 8 PC JOURNAL

2 Astudy of the behavior of YJscale model precast concrete beam-to-column connections subject to cyclic inelastic loadin was initiated at the National lnstitute of Standards and Technoloy (NST) in The objective of the experimental proram is to develop recommended uidelines for the desin of an economical precast concrete beam-tocolumn connection, that provides for rapid field assembly without the use of corbels, for reions of hih seismicity. These connections were desined in accordance with the provisions of the 1985 Uniform Buildin Code (UBC)' for Seismic Zones 2 and 4. South UPPER CROSS HEAD D 1321 =::) North All dimcnsiom in mm 2S.4mm= inch TEST PROGRAM The test proram consists of four phases. Phase was an exploratory phase in which four monolithic specimens and two precast concrete specimens were tested. Test results of the monolithic specimens served as references for the precast concrete tests. The results from the precast tests were used to determine the viability of the connection details. Phase of the proram involved testin six precast concrete specimens. As a result of the low enery dissipation observed for the precast concrete specimens in Phase, several methods of increasin the enery dissipation of the precast concrete connection were explored. Because of a lack of stiffness in the precast specimens in the later staes of the tests upon load reversal, the use of partially bonded tendons was studied in Phase. Phase V, currently underway, involves the study of usin conventional mild reinforcin steel in conjunction with a post-tensioned precast connection. The results of Phases and of the NST test proram are summarized in this paper. Results from Phase and Phase were reported by Cheok and Lew in Refs. 2, 3 and 4. Test Procedure Boundary conditions for the test specimens are shown schematically in Fi. 1. The column bottom was pinned and the column top and beam ends were roller supported. The test setup July-Auust 1993 Fi. 1. Schematic of the loadin and boundary conditions for the test specimens. Fi. 2. Test setup and loadin arranement. and loadin arranement are shown in Fi. 2. An axial load equal to.1 f A was applied to the column at the beinnin of the test and maintained throuhout the test. The column was then laterally loaded at the column top for the initial elastic cycle. This initial elastic cycle developed a load equal to 75 percent of the maximum strenth of the monolithic beam calculated usin nominal values for the material properties. The top column displacements in the forward (south) and reverse (north) directions were measured. The "yield" displacement, L\y, was defined as the averae of these two displacements divided by.75. As defined in the previous pararaph, the yield displacement obtained for the precast concrete specimens was actually a nominal yield displacement as yieldin of the post-tensionin steel did not occur in the initial elastic cycle. Henceforth, the term yield displacement should be taken to mean the nominal yield displacement. After the initial elastic cycle, the basic loadin history was two cycles each at ± 2L\y and ± 4L\y, three cycles each at ± 6L\y and ± 8L\y, two cycles at ± 1 OL\y, and three cycles at ± l2l\y. 81

3 Table 1. Specimen description. Post-tensionin Test phase Specimens Seismic zone Type* steelt Groutedt Bar distance from extreme fiber (mm) Lenth of debonded post-tensionin steel (mm) A-M-Z2 & B-M-Z2 2 M A-M-Z4 & B-M-Z4 4 M A-P-Z4 & B-P-Z4 4 p B F 89 - A-P-Z2 & B-P-Z2 2 p s F 63 - C-P-Z4 & D-P-Z4 4 p B F E-P-Z4 & F-P-Z4 4 p s F 12 - G-P-Z4 & H-P-Z4 4 p s p * M = Monolithic; P = Precast B =Post-tensionin bars; S =Prestressin strands F =Fully routed; P =Partially routed Note: 25.4 mm = in. This loadin history is similar to those used by other researchers [e.., Bhatt and Kirk, 5 Bull and Park, 6 French et al. 7 ]. The test was stopped whenever failure occurred. Displacement ductility, )1, is defined as the ratio of the maximum displacement achieved in any cycle to the yield displacement. Ultimate displacement ductility, Jlu, is defined as the ratio of the maximum displacement achieved at failure to the yield displacement. Failure was considered to have occurred when the lateral force durin a cycle dropped below 8 percent of the maximum lateral load that was achieved in the first cycle at 2y Strains in the beam steel and ties were measured usin resistance-type strain aues. The aues were placed at specified locations alon the beam flexural steel. n the case of the precast concrete specimens, only the mild steel reinforcin bars in the beam were instrumented (not the post-tensionin steel). n addition, the applied lateral load at the column top, the displacement of the column top, beam loads and concrete strain at various locations alon the beam were measured throuhout the tests. Phases and ll Specimens The results of the Phase tests indicated that post-tensioned, precast concrete specimens desined usin the 1985 UBC Seismic Zone 4 criteria as 82 desin uidelines were as stron and as ductile as the monolithic Zone 4 specimens. However, compared with the monolithic specimens, the cumulative and averae cyclic enery dissipated by the precast concrete specimens were 2 and 7 percent lower, respectively. n an effort to improve the enery dissipation characteristics of the precast concrete connection, several parameters were studied in Phases and. One such parameter was the location of the post-tensionin steel in the beams. t was felt that the interity of the connection would be improved if the post-tensionin steel were moved closer to the beam centroid. By doin this, the post-tensionin steel would experience less strain and would, therefore, retain its clampin force at hiher drift levels. Also, the difference in connection behavior by the use of prestressin strands instead of post-tensionin bars was examined. This variable was included in the study because prestressin strands are used more frequently in practice. When the Phases and tests were conducted, concern was raised reardin the zero slope of the hysteresis loops upon load reversal that was observed in the precast concrete tests. This lack of stiffness observed in the later staes of the tests was thouht to be caused by the inelastic response of the post-tensionin steel. A possible solution to this problem was to leave the strands un- bonded throuh the joint reion. This concept was developed by Priestley and Tao. 8 Briefly, the strains in the post-tensionin steel would be kept in the elastic rane and the post-tensionin force would, therefore, be maintained at hiher drift levels and thereby preserve the connection stiffness upon load reversal. Special interlockin spirals were provided in the beam hine reion because hih compressive strains were expected in this location. A volumetric ratio of 2 percent with a maximum spiral pitch of D/4, where D is the spiral diameter, was used to desin the confinin spirals. A summary of the specimens tested in Phases, and is iven in Table 1. Phase of the test proram consisted of testin three sets of specimens. The Phases and precast specimens were desined in a manner similar to the Phase specimens. The required amount of post-tensionin steel was computed so that the predicted strenth of the precast connection was the same as that of the monolithic connection. Reinforcement details for the precast columns, for both Zones 2 and 4, were the same as those of the correspondin monolithic specimens. The reinforcement details for the Zone 4 precast beams were identical to those of the Phase precast specimens except for the location of the post-tensionin steel. Fi. 3 shows the reinforcement de- PCJOUANAL

4 2-#3 Top and B«tan 2Smm Pcst-k:Mionin bar & 38mm CO'l'UJated duct Smm smlh. 68mm.C j_ llmm Strands Top&Bottom Specimens C & D All dijdimsi111 in 111m m = 1 inch 2-#3 Top and Bottom Specimens E & F - Fully Bonded Specimens G & H - Partially Bonded Fi. 3. Beam cross sections for Phases and ll precast concrete Zone 4 specimens. tails for the precast Zone 4 specimens and Fi. 4 shows the reinforcement details for the Zone 2 specimens. The reinforcin bars located in the corners of the beams did not cross the joint since they were used mainly to hold the ties toether. The main resistance to the applied loads was provided by the post-tensionin steel. Material Properties The stress-strain curves for the #3 reinforcin bars, smooth wire used in the beams and the hih strenth steel are shown in Fis. 5 throuh 12. The stressstrain curves for the smooth wire used in Specimens A-P-Z2 and B-P-Z2 and E-P-Z4 and F-P-Z4 are similar to those for Specimens G-P-Z4 and H-P-Z4. Application of Post-Tensionin A major concern in the use of strands was the load loss in the strand forces due to seatin of the wedes. This loss was expected to be sinificant due to the short lenths of strands involved. A procedure which involved shimmin of the chucks was followed to minimize the load loss. When the test specimens were posttensioned, only two load cells were July-Auust 1993 Table 2. Load cell readins after post-tensionin. Specimen Load Celli* Load Cell2t E-P-Z4.65 F;,u.65 F;,u F-P-Z4.65 F;,u.67 F;,u G-P-Z4.66 F;,ut.66 F;,ut H-P-Z4.68 F;,u:f:.64 F;,u:f: A-P-Z2.71 F;,u.67 F;,u B-P-Z2.69 F;,u.7 F;,u * Top set of strands. Bottom set of strands. Averae of loads from strain aues and load cells. 13 mm (.5 in.) prestressin strands. used - one top and one bottom for Specimens E-P-Z4 and F-P-Z4 and A P-Z2 and B-P-Z2. Six load cells were used for Specimens G-P-Z4 and H-P Z4. n addition to the load cells, the strands in Specimens G-P-Z4 and H P-Z4 were instrumented with strain aues. These aues were located in the unbonded part of the strands. The loads measured immediately after post-tensionin are iven in Table 2. As indicated in Table 2, the loads in the individual load cells were within one standard deviation of the mean as obtained in the trial runs. The post-tensionin steel was routed one day after the post-tensionin operation and the time between post-tensionin and testin the specimens raned from 2 to 4 weeks. TEST RESULTS Load-Displacement and Ductility The hysteresis loops for Phases and specimens are shown in Fis. 13 throuh 2. The load and displacement in these fiures represents the applied lateral force and the displacement at the column top. The hysteresis plots for the Zones 2 and 4 monolithic 83

5 2-#3 Top and Bottom 2-13 nun strands r _j_ 4 -#3,3 -#4 Top and Bottom 5 nun smth nun smth. 2-llnun strands t 63 All mmensions in mm 25.4 mm = 1 inch BEAM CROSS SECTON COLUMN CROSS SECTON Fi. 4. Reinforcement details for Phase precast concrete Zone 2 specimens. 5 r----;--;----,.----,---r-l---, , ,, ,-1----,---=,..., , f.=i===t==t::=:=t:::::::::::::.. j... i j t l : t! -! F = -+---f : ,l... f - f f -.. i..... l:: 2 1!! 14.5 [ '---"'----'----'---'---'---' O.QJ O.Q Strain Fi. 5. Stress-strain curve for #3 reinforcin bars, Specimens A-P-Z4 to D-P-Z4, A-P-Z2 and B-P-Z2. u/...;....! v '----'----'------''------'' '..5.1 O.oJ Strain Fi. 6. Stress-strain curve for #3 reinforcin bars, all monolithic, Specimens E-P-Z4 and F-P-Z4. specimens (from Phase ) are shown in Fis. 21 and 22, respectively, for comparison purposes. The ductilities of the precast Zone 4 specimens, as shown in Fis. 13 throuh 18, are reater than that of the monolithic Zone 4 specimen, as can be seen from Fi. 22. However, the hysteresis loops for the precast concrete specimens are more severely pinched than those for the monolithic specimen. The pinched hysteresis loops probably result from the combination of the concentration of rotation at the column face due to yieldin of the ten- 84 dons and debondin of the strands durin the test which allowed some slippae to occur. The shape of the hysteresis loops for the fully bonded precast specimens (see Fis. 15 and 16) differ from those for the partially bonded specimens shown in Fis. 17 and 18. As seen in these fiures, the partially bonded specimens did not exhibit zero stiffness upon load reversal. However, the loops for the partially bonded specimens are narrower than those for the fully bonded specimens. The experimental yield displacements and ultimate displacement ductilities for all the specimens (Phases, and ) are listed in Table 3. t should be noted that the yield displacements for the precast concrete specimens in Table 3 are nominal yield displacements and J.lu is based on this nominal yield displacement. Also listed in Table 3 are the initial connection stiffnesses and story drifts at failure. Both precast Zone 2 specimens had an ultimate displacement ductility of 4, which is 5 percent less than the ductility of the correspondin monolithic specimens. The yield displace- PC JOURNAL

6 L _[ 72.5 ' ll :::::::-; !l "' < :::.J 1./ 1/ 29. "'!l "' <1.1 < i.5 O.QJ O.Ql Strain Strain 1/... J 1i 29. ":' 14.5 Fi. 7. Stress-strain curve for #3 reinforcin bars, Specimens G-P-Z4 and H-P-Z4. Fi. 8. Stress-strain curve for smooth wire, Specimens A-P-Z4 to D-P-Z4. ments for the Zone 2 specimens, monolithic and precast, were approximately the same. However, due to the hiher ductilities achieved by the monolithic specimens, the story drifts at failure were 65 percent reater for the monolithic specimens. As listed in Table 3, all the Zone 4 precast specimens achieved hiher ultimate displacement ductilities than those of the monolithic specimens. The displacement ductilities listed in Table 3 for Specimens G-P-Z4 and H P-Z4 are minimum values because the tests for these specimens were stopped prior to failure. At Jlu = 14 (story drift ""4 percent), the specimens did not exhibit any sins of strenth deradation, and it was thouht that the specimens would not fail unless the strands fractured. The expected displacement ductility of 15.6 for these specimens 8 compares well with the experimental minimum value of 14. The initial connection stiffnesses for the Zone 4 precast specimens were hiher than those for the companion monolithic specimens. Approximate increases in stiffness were 85 percent for precast Specimens A, B, G and H, 95 percent for precast Specimens C and D, and 135 percent for precast Specimens E and F. The fully bonded precast specimens post-tensioned with strands (Specimens E and F) were stiffer and achieved slihtly hiher story drifts at failure than did the specimens post-tensioned with fully bonded bars (Specimens C and D). Also, the precast specimens with fully bonded steel had slihtly hiher initial connection stiffnesses when the post- Table 3. Yield displacement and ultimate displacement ductility. Concrete compressive Experimental yield Ultimate Specimen strenth, J; * displacement displacement desination MPa (psi) mm (in.) ductility, 11u A-M-Z (631) 9.1 (.359) 6 B-M-Z (596) 9.4 (.371) 6 A-P-Z2 34. (493) 8.5 (.333) 4 B-P-Z (528) 8.4 (.33) 4 A-M-Z4 3.7 (445) 6.7 (.263) 6 B-M-Z (467) 7.4 (.293) 6 A-P-Z4 4.6 (589) 4.1 (.16) 1 B-P-Z (645) 4.5 (.179) 1 C-P-Z (678) 5.3 (.29) 12 D-P-Z (651) 5.4 (.213) 12 E-P-Z (423) 5.7 (.225) 12 F-P-Z (42) 5.5 (.218) 12 G-P-Z4 3.1 (437) 3.6 (.143) 14:j: H-P-Z (468) 3.4 (.135) 14:j: nitial elastic connection Ultimate story stiffnesst drift kn/m (kips/in.) (percent) 7.7 (44) (38) (71) (63) (121) (13) (24) (216) (18) (258) (241) (286) (26) (22) 3.6 * Strenths were obtained at the time of specimen testin. These values are the initial elastic stiffness as obtained from the slope ofihe load-displacement plot for the initial excursion to +.756,. The specimens did not fail at this ductility level. Therefore, these are minimum values for the specimens. July-Auust

7 6 r-----,, ,----;----;---, soo4 :::::--;r-=f=.. =+==:... T=t==+==ln.s... ;+ --! -' l ! ! 43.5 "' R <ll 2 't -,/ !1. <ll 1 ll!l t + l l ::., t lnl 116 <ll O.Ql O.oi o.oos.& Strain Slr.tin Fi. 9. Stress-strain curve for smooth wire, Specimens G-P-Z4 and H-P-Z4. Fi. 1. Stress-strain curve for 7/16 in. (11 mm) strands, Specimens E-P-Z4 to H-P-Z4. tensionin steel was moved closer to the beam centroid. Flexural Strenth The maximum measured and calculated moments are listed in Table 4. The calculated values were based on an ultimate concrete compressive strain of.3 and actual material properties. The calculated ultimate moments for the monolithic specimens had a factor of 1.25 applied to the steel yield stress to account for possible hiher actual material strenths and strain hardenin. No factor for strain hardenin was used to calculate the moments for the precast concrete specimens. The moments for the partially bonded specimens were calculated based on AC-318 Eq. (18-4) 9 for determinin the stress in the strands at nominal strenth. Aain, actual material properties were used. The experimental values were obtained by multiplyin the peak load recorded in the beam load cell and the moment arm to the column face. The measured maximum moments for the monolithic specimens were achieved at or close to the ultimate displacement ductility of the specimens. This was also true for precast Specimens A-P-Z2, B-P-Z2, A-P-Z4 and B P-Z4. However, the measured maximum moments for the precast concrete specimens with the post-tensionin steel closer to the beam centroid (Specimens C-P-Z4 throuh F-P-Z4) were achieved earlier in the tests (at J..l "' 4 whereas J..lu "' 12). The measured maximum moments for the partially bonded Table 4. Experimental and calculated maximum moments. Concrete compressive strenth Calculated maximum Experimelltal maximum Specimen J;* moment momentt A v. exp. moment desination (2) (3) (4) Calc. max. moment () MPa (psi) kn-m kip-ft kn m kip ft (5) A-M-Z (631) B-M-Z (596) A-P-Z2 34. (493) B-P-Z (528) A-M-Z4 3.7 (445) B-M-Z (467) A-P-Z4 4.6 (589) B-P-Z (645) C-P-Z (678) D-P-Z (651) E-P-Z (423) F-P-Z (42) G-P-Z4 3.1 (437) H-P-Z (468) * Strenths were obtained at the time of specimen testin. t Moments are at the two column faces & 8 51 & & 75 52& &54 38 & &54 37 & & & & & & & & & & & & & & & & & & & & & PC! JOURNAL

8 16 t: S' 6 / "' l 8 "' Strain Fi. 11. Stress-strain curve for 1/2 in. (13 mm) strands, Specimens A-P-Z2 and B-P-Z "' l 1:! / 116 ;; 6 6 1'1 -.../ 87.) "' 4 58 "' 2 f O.oJ O.oJ Sttain Fi. 12. Stress-strain curve for 1 in. (25 mm) Dywida bars, Specimens A-P-Z4 to D-P-Z4. specimens (minimum J.lu = 14) were obtained at}.l"" 1. t would appear that the fully bonded precast concrete specimens with the post-tensionin steel located closer to the beam centroid debonded early in the test and behaved in a manner similar to the partially bonded specimens. Debondin of the steel in the specimens with the post-tensionin steel further away from the beam centroid may also have occurred, and the hiher moments achieved in the later staes of the tests may be a result of the strain hardenin of the steel. As shown in Column 4 of Table 4, the fully bonded precast Zone 4 specimens (A-P-Z4 to F-P-Z4) were as stron as or stroner than their monolithic counterparts. When comparin the results of Specimens A-P-Z4 and B P-Z4 with those of Specimens C-P-Z4 and D-P-Z4, the reduction in strenth caused by movin the post-tensionin bars closer to the beam centroid was approximately 8 percent. The expected reduction in connection strenth (Column 3, Table 4) for these specimens was 6 percent. The ratios iven in Column 5 of Table 4 indicate that the precast concrete specimens post-tensioned with strands had more "reserve" strenth than those post-tensioned with hih strenth bars - an averae of 29 percent reater than the calculated moment for the strands as opposed to 17 percent for the bars. The experimental strenth of the partially bonded specimens (G-P-Z4 and H-P Z4) was approximately 1 percent less than that of the fully bonded specimens (E-P-Z4 and F-P-Z4). Failure Modes The failure mode for the precast Zone 2 specimens was beam crushin. This is different from the failure mode for the monolithic Zone 2 specimens, which failed predominantly in shear in the column joint reion. The precast specimens did not sustain sinificant damae in the joint reion. The beamto-column openin for the precast Zone 2 specimens was approximately 5 mm (.2 in.). This is one-fifth of the openin size for the precast Zone 4 specimens. There are insufficient data to determine if this is a result of the hiher initial beam stress resultin Displacement (ifl.) Fully Bonded Bars l Fully Bonded Bars :.::::::::::::::::J::::::::::::::::::r. 1 l lflb'hn'h -(#.flf. : -1 fjjjlij.flll ;;e: :;.; ;;c -+-"' S 22-22! + -4S i--- -i ! -67 i 1 i - L SO so 1 DisplacemeRt (mm) :. :::::::::::::::::::r::::::::::::::::r!! u! i 4S 22 : i 't"""'""""""t"""""""""' t... t" ;... i l -1 -SO SO 1-9 Fi. 13. Hysteresis plot for Specimen C-P-Z4. Fi. 14. Hysteresis plot for Specimen D-P-Z4. J!My-Auust

9 «9 «r-----,1----f-ull-y_b_o_n-s-tr_m_&-, :.::::::::::::::::::l:::::::::::::::::::r: t j [_.Afr#-7/ i \ i! -3 f... -i j i -4 L «-9-1 -SO Fi. 15. Hysteresis plot for Specimen E-P-Z4. Fi. 16. Hysteresis plot for Specimen F-P-Z4. from the post-tensionin of the Zone 2 specimens, or the result of the lower loads experienced by these specimens. The failure modes for all the precast Zone 4 specimens were similar to those for the precast Zone 2 specimens. One of the precast Zone 4 connections after failure is shown in Fi. 23. The precast concrete beams sustained less crackin than did the beams in the monolithic specimens, and concentration of beam hinin at the column face was observed. However, more beam crushin and a wider openin between the beam and col- umn, approximately 5 percent reater, was observed in the precast specimens which had the post-tensionin steel closer to the beam centroid. The partially bonded precast specimens experienced more crushin of the beams than did the fully bonded specimens. However, the width of the openin between the beam and column was similar in both sets of precast concrete specimens. The column joint reions in the partially bonded precast specimens appeared to sustain the same amount of damae as the fully bonded set of precast specimens. Enery Dissipation Characteristics Fi. 24 shows a comparison of the enery dissipated per cycle for all the precast Zone 4 specimens. As seen in the chart, Specimens E-P-Z4 and F-P Z4 (fully bonded strands) performed the best and Specimens G-P-Z4 and H-P-Z4 (partially bonded strands) performed the worst in terms of enery dissipated per cycle. Aain, the low cyclic enery dissipated by the partially bonded specimens was due to the expected bilinear elastic behavior of these specimens «r Fi. 17. Hysteresis plot for Specimen G-P-Z4. Fi. 18. Hysteresis plot for Specimen H-P-Z4. 88 PC JOURNAL

10 r :! Fully Bonded Strands 67 --[-- :! l : J... l... l S _: ,--- i Fully Bonded Strands t t t ; : l :! ; t t -! 45 i! 1 i =:=:-z::_:=:! i! l -2 - t t i -45! i i i -3 r r r ] -22!! l -4 L Fi. 19. Hysteresis plot for Specimen A-P-Z2. Fi. 2. Hysteresis plot for Specimen B-P-Z2. Comparin the results of Specimens A-P-Z4 and B-P-Z4 with those of Specimens C-P-Z4 and D-P-Z4, the enery dissipated per cycle showed an averae increase of 45 percent ( cr = 24 percent) when the post-tensionin bars were moved closer to the beam centroid. This increase is probably due to the increased crushin of the beam concrete in Specimens C-P-Z4 and D P-Z4 as observed durin the tests. An averae increase of 3 percent ( cr = 22 percent) in the enery dissipated per cycle was noted when prestressin strands (Specimens E-P-Z4 and F-P-Z4) were used in place of post-tensionin bars (Specimens C-P Z4 and D-P-Z4). As seen by the lare standard deviations, the percent increase in enery dissipated per cycle was hihly variable. The enery dissipated per cycle decreased by approximately 55 percent (cr = 7 percent) when partially bonded strands were used instead of fully bonded strands. As shown in Fi. 25, the enery dissipated per cycle by precast Specimens E-P-Z4 and F-P-Z4 was an averae of approximately 6 percent of that dissipated by the monolithic Zone 4 specimens. Also, the cyclic enery dissipated by Specimens G-P-Z4 and H-P-Z4 was less than the cyclic enery dissipated by the monolithic Zones 2 and 4 specimens. This is due to the narrow and pinched hysteresis loops for these specimens. The cumulative enery dissipated to failure by all the specimens is shown in Fi. 26. With respect to cumulative enery dissipated, the precast Zone 4 Specimens C throuh F performed better than the monolithic Zone 4 specimens. This was a result of the hiher displacement ductilities achieved by these precast concrete specimens, which meant that these specimens underwent more reversed ,----,----,.----,.----., :! Monlithic :z:one 2 -:.. -: i! i i : = ! +!..! \ j... :! 1!. : t i!!! l -2 j..t....( !......! r l- -r i 45 : : i : -4 L i i i 1 -' \... :... ' !! l -1 r ! -2 r r- l r r r l Fi. 21. Hysteresis plot for Specimen A-M-Z2. Fi. 22. Hysteresis plot for Specimen A-M-Z4. July-Auust

11 specimens. This is due to the concentration of the rotation at the column face for the precast concrete specimens. The recorded strains in ties for the precast Zone 4 specimens were also much lower than for the monolithic specimens. Amon the precast specimens, the specimens post-tensioned with bars had loner reinforcin bar yield lenths than those posttensioned with strands. t would appear that the strains and rotations were more concentrated at the column joint for the specimens post-tensioned with strands. Fi. 23. Representative failure mode for the precast concrete specimens , 142 ll Av. A & B P-Z4 D Av.C&DP-Z4 ll Av. E & F P-Z4 12 1:-:s Av. G&HP-Z4 \ ' Cycle Number Fi. 24. Cyclic enery dissipated by precast Zone 4 specimens. cyclic loadin prior to failure. The cumulative enery dissipated to failure by Specimens G-P-Z4 and H-P Z4 was approximately equal to that for monolithic Zone 4 specimens. However, these specimens had not failed when the tests were stopped. Therefore, the cumulative enery dissipated by Specimens G-P-Z4 and H-P-Z4 as shown in Fi. 26 are minimum values. Reinforcin Bar Yield The yield lenths of the beam reinforcement ive an indication of the strain concentration in the beam and, 9 therefore, an indication of the concentration of the beam rotation. The yield lenth of the reinforcement was measured from the column face. This lenth was obtained by first plottin the strain in the reinforcin bar aainst the distance of the strain aue from the column face. The points were then connected with straiht lines. The yield lenths were defined as the lenths from the column face to the point at which these lines crossed the nominal yield strain level. As expected, the monolithic specimens exhibited loner reinforcin bar yield lenths than did the precast CONCLUSONS 1. Failure Mode - Failure modes for all precast concrete specimens were similar. Failure was characterized by yieldin of the post-tensionin steel, beam crushin, and an openin at the junction between the beam and the column. The width of the openin at the beam-to-column joint increased as the post-tensionin steel was placed closer to the beam centroid. The width of the openin does not appear to be influenced by the type of post-tensionin steel (post-tensionin bars or prestressin strands) used to connect the precast concrete elements or by the use of partially debonded prestressin. t should be noted, however, that the partially bonded specimens were not tested to failure. Unlike the monolithic Zone 2 specimens which failed predominantly in shear in the column joint reion, the precast Zone 2 specimens did not experience severe joint distress. 2. Displacement Ductility - The ultimate displacement ductility for both precast Zone 2 specimens, f.lu = 4, was lower than that obtained for the monolithic Zone 2 specimens, f-lu = 6, and as a result, the precast concrete specimens had lower story drifts at failure. n eneral, the precast concrete specimens had hiher story drifts at failure and hiher initial stiffnesses than the monolithic specimens. The ultimate displacement ductilities of the precast specimens (J.lu = 12 and 14) were reater than those of the monolithic Zone 4 specimens, (J.lu = 6). The experimental value of 14 for the par- PCJOURNAL

12 2 r r.d Av.A&BM-Z2 Av. A & B M-Z4 Av.A&BP-Z4 Av.C&DP-Z :::: Cycle Number Fi. 25. Comparison of enery dissipated per cycle to 6y (Cycle 3). l4u ;E' &i Specimen Number Fi. 26. Comparison of the cumulative enery dissipated to failure. Note: Specimens G-P-Z4 and H-P-Z4 had not failed when testin was stopped. tially bonded specimens compared well with the predicted value of 15.6 by Priestley and Tao. 3. Connection Strenth - The measured maximum connection strenths for all the precast concrete specimens exceeded the calculated values and tlle precast specimens performed as well as the monolithic specimens in most cases. Based on the results obtained thus far, the experimental maximum moments for the precast specimens post-tensioned with strands were approximately 3 percent reater than the predicted mo- July-Auust 1993 ments. Placement of the post-tensionin bars closer to the beam centroid does not appear to have a sinificant adverse effect on the connection strenth. 4. Enery Dissipation - As in Phase, the enery dissipated per cycle by the precast Zone 4 specimens was less than that of the monolithic specimens. Enery dissipation per cycle improved from the use of prestressin strands instead of post-tensionin bars with the post-tensionin steel located closer to the beam centroid, and fully bonded. The enery 3 15 &i dissipated per cycle by Specimens E P-Z4 and F-P-Z4 (the precast specimens which performed best with respect to enery dissipated per cycle) was, on the averae, approximately 6 percent of tllat for tlle monolithic Zone 4 specimens. n eneral, the cumulative enery dissipated to failure by precast Zone 4 specimens was reater than the cumulative enery dissipated by the monolithic Zone 4 specimens. This is due to the hiher displacement ductilities achieved by the precast concrete specimens. RECOMMENDATONS Test results obtained thus far show that a post-tensioned, precast beam-tocolumn connection appears to be a viable system for reions of hih seismic activity. The shear friction between a precast beam and column has been shown to be sufficient to resist tlle applied shear loads witllout the need for corbels or shear keys. This type of post-tensioned precast concrete connection has also behaved as well as, and in most cases, better than similar monolithic connections in terms of connection strenth and ductility. n reions where enery dissipation is not an issue, the partially bonded specimens would be an attractive alternative as tlley have been shown to perform in what is essentially an elastic manner. Also, the amount of spiral steel used in the beam, with a volumetric ratio of 2 percent and a maximum pitch of D/4, was sufficient to confine the beam in the plastic hine reion. FUTURE RESEARCH NEEDS The precast concrete connections exhibited low enery dissipation per cycle as compared with monolithic connections. Phase V of the test proram will examine the use of low strenth (mild) reinforcin steel in conjunction with post-tensionin as a means of improvin the cyclic enery dissipation characteristics of the precast concrete specimens. The premise for tllis concept is that the mild steel will be used as an enery dissipater while the friction between the beam and the column caused by the post- 91

13 tensionin force will be used to pro- reatly appreciated. Members of the to-column Connections Subject to vide the necessary shear resistance. steerin committee are Robert E. En- Cyclic Loadin," PC JOURNAL, To address the concern that the pre- lekirk, S. K. Ghosh, Daniel P. Jenny, V. 36, No.3, May-June 1991, pp vious NST tests did not have ravity (L. S.) Paul Johal and M. J. Niel 5. Bhatt, Prabhakara, and Kirk, D. W., loads superimposed on the beams, Priestley. The authors would also like "Tests on an mproved Beam Column simulated ravity loads will be applied to thank Suzanne D. Nakaki for pro- Connection for Precast Concrete," AC Journal, Proceedins, V. 82, to the beams of the Phase V speci- vidin the desin of the test specimens. No. 6, November-December 1985, mens. Other parameters that require pp further study are the amount of post- 6. Bull, D. K., and Park, Robert, "Seismic tensionin steel and mild reinforcin REFERENCES Resistance of Frames ncorporatin steel, the use of different concrete. Uniform Buildin Code, nternational Precast Prestressed Concrete Beam strenths and joint materials in the Conference of Buildin Officials, Shells," PC JOURNAL, V. 31, No. 4, construction joint, and the influence Whittier, CA, July-Auust 1986, pp of slabs and transverse beams on con- 2. Cheok, G. S., and Lew, H. S., "Perfor- 7. French, Catherine, et al., "Connections nection behavior. mance of /;-Scale Model Precast Con- Between Precast Elements - Failure crete Beam-Column Connections Sub- Within Connection Reion," ASCE ACKNOWLEDGMENT jected to Cyclic nelastic Loads - Structural Journal, V. 114, No. 2, Report No. 1," NSTR 4433, National February 1989, pp The authors would like to express nstitute of Standards and Technoloy, 8. Priestley, M. J. Niel, and Tao, Jian Gaithersbur, MD, October 199. their ratitude to the individuals who Ren, "Seismic Response of Precast Precontributed to this project. The assis- 3. Cheok, G. S., and Lew, H. S., "Perfor- stressed Concrete Frames With Partance and support of the laboratory mance of /;-Scale Model Precast Con- tially Debonded Tendons," PC JOURcrete Beam-Column Connections Sub- NAL, V. 38, No. 1, January-February staff for the Buildin and Fire Rejected to Cyclic nelastic Loads , pp search Laboratory, Structures Divi- Report No. 2," NSTR 4589, National 9. AC Committee 318, "Buildin Code sion, especially that of Frank Rankin, nstitute of Standards and Technoloy, Requirements for Reinforced Concrete is ratefully acknowleded. Gaithersbur, MD, June and Commentary (AC )," n addition, the technical uidance 4. Cheok, Geraldine S., and Lew, H. S., American Concrete nstitute, Detroit, provided by the steerin committee is "Performance of Precast Concrete Beam- M, PC JOURNAL