ML15299A149

Enclosure B, Bechtel Report No. 25593-000-G83-GEG-00016-000, Effect of Laminar Cracks on Splice Capacity of No. 11 Bars Based on Testing Conducted at Purdue University and University of Kansas for Davis-Besse Shield Building. Part 5 of 7

ML15299A149
Person / Time
Site:	Davis Besse
Issue date:	07/30/2012
From:	Liano R, Munshi J, Reilly R Bechtel Power Corp
To:	Office of Nuclear Reactor Regulation
Shared Package
ML15299A142	List: L-15-328, Enclosure B, Bechtel Report No. 25593-000-G83-GEG-00016-000, Effect of Laminar Cracks on Splice Capacity of No. 11 Bars Based on Testing Conducted at Purdue University and University of Kansas for Davis-Besse Shield Building. Part 7 of 7 ML15299A143 ML15299A144 ML15299A146 ML15299A147 ML15299A148 ML15299A149 ML15299A150
References
L-15-328 25593-000-G83-GEG-00016-000
Download: ML15299A149 (90)

Text

{{#Wiki_filter:~bYJ-UUU-UtS3-ULU-UUU1b-UUU Page bbd 0? 1114Effect of Laminar Cracks on Splice Capacity of No. 11 Bars based on Testing Conducted at PurdueUniversity and University of Kansas for Davis-Besse Shield BuildingAppendix BUniversity of Kansas Test ReportUOflJi32flU~I ~J 2012 Becfltel Uorporatcon. LontBeebteland its affihatsd comoankz which shall~in: confidential and/or proprietar,' information tonotbz ussd, disclosed, or reproduced in anyformatby anvnnn Rc:htcl nnt..'withoutEechtelf: oricrwrttten oerm[zian. All ridts rcscr..'rd. --iT ............. -lr..... 0 ....B-iREDACTED VERSION

I-'age OT 1114EFFECT OF SIMULATED CRACKS ON LAP SPLICESTRENGTH OF REINFORCING BARSByJiqiu YuanMatthew O'ReillyAdolfo Matamoros David DarwinResearch supported byFIRSTENERGY NUCLEAR OPERATING COMPANY(FENOC)Structural Engineering and Engineering Materials SL Report 12-2THE UNIVERSITY OF KANSAS CENTER FOR RESEARCH, INC.LAWRENCE, KANSASJune 2012REDACTED VERSION

Mage ti~u atl 1114AbstractThe effect of preexisting cracks, oriented in the plane of and parallel to the reinforcing steel, on the strength of No. 11-bar lap splices was investigated by testing six beams -three witha splice length of 79 in. and three with a splice length of 120 in. One of the beams with a 79-in.splice was cast monolithically and loaded monotonically to failure. To simulate the cracks, theother five beams were cast with a cold joint at the mid-height of the reinforcing steel. Two beams(one with a 79-in, splice and one with a 120-in, splice) with a cold joint were loadedmonotonically to failure. The other three beams were preloaded to develop horizontal cracks inthe face of the cold joint, unloaded and then loaded to failure; those beams developed horizontal cracks with widths of 20, 30 and 35 mils (0.02, 0.03, 0.035 in.) during the first cycle of loadingand just prior to unloading. The nominal concrete compressive strength was 5000 psi.The methods described in this report provide a viable means of simulating a crack in theplane of flexural reinforcement. In the presence of a simulated crack in the plane of thereinforcing bars, the two specimens with lap-spliced No. 11 bars with a 79-in, splice lengthachieved bar stresses of 62 and 57 ksi. In the presence of a simulated crack in the plane of thereinforcing bars, the three specimens with lap-spliced No. 11 bars with a 120-in, splice lengthachieved bar stresses of 72, 67, and 69 ksi.REDACTED VERSION bbIZ-UUU-Lidi-USI-U-UUUIh-UUU I-age btJE1 01" 1114Personnel The research described in this report was performed under the direction of David Darwin,Ph.D., P.E., Deane E. Ackers Distinguished Professor of Civil, Environmental, and Architectural Engineering and Director of the Structural Engineering Materials Laboratory and AdolfoMatamoros, Ph.D., Associate Professor of Civil, Environmental, and Architectural Engineering at the University of Kansas. Darwin and Matamoros are joined by post-doctoral researchers, Matthew O'Reilly and Jiqiu Yuan.David Darwin has extensive experimental and analytical experience in the field of bondand development of reinforcement and has been involved in bond research for over 30 years. Heis a member and past chair ACI Committee 408 on Bond and Development of Reinforcement and a member of ACI Subcommittee 318-B Reinforcement and Development (Structural Concrete Building Code). Darwin developed the ACI Committee 408 expression fordevelopment and splice lengths, which accurately covers straight reinforcing bars with yieldstrengths between 40 and 120 ksi and for concrete with compressive strengths ranging from2,000 to 16,000 psi. Darwin also developed the ASTM A944 Beam-End Test, which is used toevaluate relative bond strength. In addition to Committee 408, Darwin is a member and pastchair of ACI on Committee 224 on Cracking, as well as four other ACI technical committees. Adolfo Matamoros has many years of experience in testing and analysis of reinforced concrete with special expertise in test instrumentation. He is the immediate past chair of ACICommittee 408 and a member of four other ACI technical committees. Matthew O'Reilly and Jiqiu Yuan completed their Ph.D. degrees at the University ofKansas in 2011 and have been serving as senior researchers on a number of experimental projects since receiving their degrees.REDACTED VERSION

P-age bSU2 0 1114ContentsAbstract.................................................................................................... iiiPersonnel .................................................................................................. iv1 Overview / Background............................................................................. 12 Research Program and Test Specimens............................................................ 12.1 Design of test specimens ....................................................................... 12.2 Concrete ......................................................................................... 32.3 Cold joint construction and crack simulation................................................. 42.4 Test methodology............................................................................... 92.4.1 Fabrication ................................................................................. 92.4.2 Casting .................................................................................... 102.4.3 Curing..................................................................................... 112.4.4 Test apparatus ............................................................................ 112.4.5 Loading protocol ......................................................................... 142.4.6 Calibration ................................................................................ 182.5 Test Facilities.................................................................................. 182.6 Section Analysis............................................................................... 193 Test Results ....................................... ................................................. 233.1 General......................................................................................... 23REDACTED VERSION

P~age bu;d oT 11143.2 Beams 1, 2, and 3 with 79-in. splice length................................................. 243.2.1 Concrete strength ........................................................................ 243.2.2 Beam 1 (monolithic concrete)........................................................... 263.2.3 Beam 2 (cold joint, monotonically-loaded)............................................ 323.2.4 Beam 3 (cold joint, cycled)..................... ......................................... 373.3 Beams 4, 5, and 6 with 120-in, splice length ............................................... 413.3.1l Concrete strength ........................................................................ 413.3.2 Beam 4 (cold joint, monotonically-loaded)............................................ 423.3.3 Beam 5 (cold joint, cycled) ............................................................. 4"73.3.4 Beam 6 (cold joint, cycled).............................................................. 534 Summary and Conclusions........................................................................ 59

References:

............................................................................................... 61Appendix A: Pilot Tests -Preliminary Study of the Effect of Simulated Cracks on Lap SpliceStrength of Reinforcing Bars using Beams with Single Splices ..................................... 62Appendix B: Reinforcing Steel Drawings............................................................. 83Appendix C: Detailed crack maps of Beams 1- 6.................................................... 87Appendix D: Reinforcing steel mill certification and deformation measurements............... 106Appendix E: Concrete Mixture Proportions ......................................................... 110Appendix F: Data recording forms.................................................................... 113REDACTED VERSION bbui;-UUU-Ud3-ULhU-UUU1br-UUU H-age boW40 at1114Appendix G: Load cell and displacement transducer calibration .................................. 119Appendix H: Training forms, Trip tickets, concrete properties, specimen dimensions, and crackrecording during beam tests ........................................................................... 131Appendix I: Certificates of calibration for laboratory apparatus ................................... 198REDACTED VERSION 25593-000-G83-GEG-00016-000 Page 695 of 1114This page is intentionally blankREDACTED VERSION

b-U~u I-'age 15ti Ot 11141 Overview / Background Past research on the strength of lapped bar splices in reinforced concrete has focused oninvestigating the performance of various lap splice configurations in monolithic members. Theresearch program described in this report investigates the effect of preexisting cracks, oriented inthe plane of and parallel to the reinforcing steel, on the strength of lapped bar splices. Theresearch program was conducted in two phases, a pilot study investigating various methods tosimulate the preexisting cracks that is described in Appendix A of this report, and a series ofbeam tests described in the main body of the report.Beams in the main study had cold joints in the splice region, along the plane of thereinforcement, to facilitate the initiation of a crack prior to failure. Two No. 3-bar hoops (one oneach side) crossing the plane of the cold joint, in the center of the specimen and on the exterior ofthe lap splices, were used to simulate the effects of the continuity of concrete in an actualstructure. The beams contained two spliced No. 11 bars with 79 or 120-in, long lap splices. Someof the beams were loaded until horizontal cracks had developed along the plane of the cold jointwith a minimum width of 10 mils (0.01 in.); they were then unloaded and subsequently reloadedto failure. The remainder of the beams were loaded monotonically to failure.2 Research Program and Test Specimens 2.1 Design of test specimens A total of six beam-splice specimens were tested in the main study -three specimens with a splice length of 79 in. and three with a splice length of 120 in. For the three specimens with a 79-in, splice length, one was cast with monolithic concrete and the other two were castwith a cold joint in the plane of reinforcing steel. All three specimens with a 120 in. splice lengthwere cast with a cold joint in the plane of reinforcing steel. All specimens With cold joints hadtwo No. 3-bar hoops crossing the plane of the cold joint, outside the spliced bars, at the center ofthe specimen. The beams were subjected to four-point loading to provide a constant moment (excluding deadload) in the middle portion of the member, where the splice was located, as shown in Figure 2.1.REDACTED VERSION bb-LUUU-U-(.5iLU-,1(-UUU I-age Ot 1114The specimens were configured to have a constant moment in the splice region toeliminate the effect of shear forces on splice strength, and also to eliminate the need for shearreinforcement within the splice region. The spacing of the supports was chosen so that thedistance from either end of the splice to the central pin and roller supports was equal to or greaterthan the effective depth of the beam. The span lengths were selected in increments of 3 ft basedon the spacing of load points in the Structural Testing Laboratory of University of Kansas.ShearMomentFigure 2.1 -Configuration and shear and moment diagrams for the testing fixtureThe reinforcement diagrams for the specimens in the study are shown in Appendix B. Thetop reinforcement layer of the beams consisted of two No. 11 reinforcing bars, which were spliced atthe center of the beam, as shown in Figure 2.1. The No. 11 bars used in the specimens were from asingle heat of reinforcement. The bottom layer of reinforcement, placed to maintain the integrity ofthe beam after failure of the splice and to facilitate placement of shear reinforcement in the constantshear regions, consisted of two Grade 60 No. 3 bars. Beam dimensions and effective depths aresummarized in Table 2.1.REDACTED VERSION DbL-UUU-UdiL-(.5UL-UUUI b-UUU r-age or11114The specimens were proportioned to have two splices, each with a nominal side concretecover of 3 in. to the outermost No. 11 bars and a top concrete cover of 3 in. Grade 60 No. 5closed hoops spaced at 5 in. on center were placed in the constant shear region (Figure 2.1) of allsix beams. Mill certifications for the No. 11, No. 5 and No. 3 bars are reported in Appendix D.B179(monolithic) 1125182420.32.8B2 79 Cold joint 11 25 18 24 20.3 2.8B3 79 Cold joint 11 25 18 24 20.3 2.8B4 120 Cold joint 14 28 18 24 20.3 2.8B5 120 Cold joint 14 28 18 24 20.3 2.8B6 120 Cold joint 14 28 18 24 20.3 2.8The deformation properties of the No. 11 bars are summarized in Table 2.2. The meandeformation height and spacing of the No. 11 bars meet the requirements of ASTM A6 15 and therelative rib area, 0.0685, is within the typical range for conventional reinforcement in U.S.practice (0.060 and 0.085) (ACI 408R-03). INo. II V .UOI I I U.IJIL U 1 .*Per ASTM A 615 **Per ACI 408R-03 and ACI 408.3R-09 for calculation of relative rib area2.2 ConcreteThe concrete used to fabricate the test specimens was supplied by a local ready mix plant.The concrete was non-air-entrained with Type I portland cement, 1 1/2/-in. nominal maximum-size crushed coarse aggregate, and a water-cement ratio of 0.42. A trial batch was prepared at the concretelaboratory of the University of Kansas prior to casting the first three beams. The aggregate gradation, REDACTED VERSION

b-UUU r-age 0? 1114mixture proportions, and concrete properties for the trial batch and each of the placements arepresented in Appendix E. The dosage of high-range water reducer was adjusted on site whenconsidered necessary to obtain adequate slump for placement. 2.3 Cold joint construction and crack simulation The specimens with cold joints were cast using two placements, with a cold joint at themid-height of the top layer of reinforcement, to ensure that a longitudinal crack would develop inthe plane of the reinforcing steel before the beam failed. The cold joints spanned the entire lengthof the spliced region and extended approximately 6.5 ft outside of the spliced region (Figures B.2and B.3 in Appendix B).In the first placement, concrete was cast up to the center of the top layer of reinforcement (Figures 2.2, B.2 and B.3). After the concrete was placed, a roughened surface was created tosimulate the roughness of a natural crack by introducing indentations in the concrete while itremained plastic (Figure 2.3). The exposed reinforcing steel was cleaned using sponges tofacilitate adequate bond between the exposed bars and the concrete cast during the secondcasting stage. The specimens were moist cured for a day, and the remainder of the concrete wasplaced no later than 26 hours after the original placement. The concrete for the second placement had the same mixture proportions and was supplied by the same ready-mix plant as the first.Before the second placement, the concrete surface was cleaned using compressed air to removedebris and loose concrete, and maintained in a wet condition until the second placement started(Figures 2.4 and 2.5). After casting, the specimens were moist-cured until the compressive strength of the concrete from the first placement exceeded 3500 psi.Some beams were loaded in two stages to ensure that the preexisting crack of minimumwidth had formed in the plane of the reinforcing steel. To do this, beams were loadedmonotonically until the width of the horizontal cracks at the cold joint exceeded 10 mils (0.01in.). After initial loading, the specimens were unloaded and subsequently reloaded monotonically to failure.REDACTED VERSION

l-'age fUU oft 1114Concrete cast monolithically at the ends of the beamspecimen. Concrete cast to the level ofreinforcing steel in the middleportion of the beam specimens. Figure 2.2 -rfirst stage of casting was completed. REDACTED VERSION

I-'age fU1 Ot 1 114(a) (b)Figure 2.3 -Roughening of the concrete surface at the cold joint. (a) roughening of the concretesurface while the concrete remains plastic. (b) roughened surface after concrete had set.Figur 2.4RemoAlCfTooED cocEteRSIONgopesdar REDACTED VERSION bbL-UUU-UbJ-GW:I-UUUIIh-UUU I-'age /UZ 01" 1114Figure 2.5 -Wetting of concrete surface prior to concrete placement. The flexural strength of the concrete (a measure of its tensile strength) was measured inaccordance with ASTM C78. For each set of beams, two specimens were cast monolithically with concrete from below the cold joint and two were cast with a cold joint at midspan in theflexure specimen using concrete from both below and above the cold joint in the beam. ForBeams 4, 5, and 6, two additional flexure beams were cast monolithically using the concretefrom the second placement (above the cold joint). The specimens with the cold joint were cast sothat half of the total length was filled with concrete from below the cold joint in the splicespecimens; the concrete surface was then roughed (Figure 2.6) following the same procedure used for the beam-splice specimens (Figure 2.3 to Figure 2.5). The second half of the "coldjoint" flexure specimens was cast using concrete from above the cold joint in the splicespecimens. A schematic of the flexure test is shown in Figure 2.6(c). The test results aresummarized in Chapter 3 and indicate that the cold joint had significantly lower tensile strengththan monolithically-cast

concrete, and thus provided a good representation of a preexisting crack.REDACTED VERSION bwb-u-UUU-w.tuJ-uuuiw.-UUUu roUU -ags /us 0?" 1114(a)(b)/ L -' .L ;-4.-- £-Bed Spe Lenth L '-"(c)Figure 2.6 -Flexure beam specimens with cold joint. (a) A flexure beam specimen cast to halfof its length. (b) Roughed concrete surface.

(c) Schematic of flexure test (ASTM C78).REDACTED VERSION bb~-UUU-U~iJ-ULUI--UUU1t)-UUU I-'age fU4 0? 1 1142.4 Test methodology

2.4.1 Fabrication

FormworkThe formwork was fabricated using plywood and dimension lumber with the tolerances specified in Table 2.3. The interior of the forms was coated with a sealant, taped at the seams toprevent leakage, and covered with a thin layer of oil before casting. The dimensions of theformwork were measured and are recorded in Table F. 1 (Appendix F).Table 2.3 -Form tolerances Reinforcement The steel reinforcement was fabricated to meet the dimensions specified in the drawingsshown in Appendix B. In the splice test region, the bar spacing, concrete cover, location ofsimulated cracks, and splice length satisfied the tolerances specified in Table 2.4. Outside of thesplice test region, the bar spacing, concrete cover, and longitudinal bar location satisfied theintended tolerance of+/- 1/2 in. Inside the forms, the reinforcing steel was supported by chairs tiedto the bottom of the hoops outside of the test region (splice region) and to the bottom layer oflongitudinal reinforcing steel in the splice region. Spliced bars were supported by small-diameter threaded rods attached to both sides of the forms. The threaded rods were introduced with theobjective of achieving the specified tolerance in the cover dimensions and preventing bowing ofthe forms at the top of the beams. Cover and reinforcement dimensions in the test region weremeasured and are recorded on Table F.2 (Appendix F).REDACTED VERSION

I-'age (Ub2 Of 1 114i oierance +/-1l~/2 +/-J/6 +/-1]/22.4.2 CastingThe properties of the plastic concrete were measured in accordance with the ASTMstandards cited and are presented in Table F.3. The following properties were recorded: -Unit Weight (ASTM C 138)-Slump (ASTM C 143)-Concrete Temperature (ASTM C 1064)The concrete truck operator delivered a ticket with the batched mixture weights. Theticket was examined to verify that the mixture delivered had the specified proportions and thatthe concrete had arrived less than 45 min. after leaving the batching plant. No water was addedto the concrete after the truck left the plant.The beams were cast in two layers, beginning and ending at the ends of the beams. Thebottom and top layers of concrete in the splice regions of all three beams were placed from themiddle portion of the batch. The concrete was sampled at two points in the middle portion of thebatch in accordance with ASTM C 172, the first sample taken immediately after placing the firstlift, and the second sample taken immediately after placing the second lift in the splice regions.After placing the second lift, excess concrete was removed from the formwork using a screed.The upper surfaces of the specimens were finished using hand floats.The samples were consolidated prior to testing the plastic concrete (for slump, unitweight, and concrete temperature) and making the strength test specimens. Ten plastic and sixsteel 6 x 12-in, cylinder molds were filled in accordance with ASTM C3 1, along with fourflexural beam specimens cast in accordance with ASTM C78. Two of the flexural beamspecimens were cast monolithically and two were cast with cold joints. Cylinders cast in plasticREDACTED VERSION bbJ;-UUU-LSt-J'-GLU-UUU1 I-age I Uti 0? 1]14molds were used for monitoring the strength of the concrete prior to testing the beam; thecylinders cast in steel molds were used to obtain the compressive strength on the day of test ofeach beam. All flexural beam specimens were tested on the day of test of the corresponding beam. Test beams and cylinders were labeled with an identifying mark.For specimens with a cold joint, the concrete above the joint plane was placed no laterthan 26 hours after the initial placement. The concrete above the cold joint had the same mixproportions as the concrete below the cold joint and was supplied by the same ready-mix plant.The concrete slump, unit weight, and temperature were recorded. A minimum of five 6 x 12-in.cylinders (two in plastic molds and three in steel molds) were prepared. The two cylinders castusing plastic forms were tested on the day of form removal when the concrete below the coldjoint had achieved a compressive strength of 3,500 psi. The three cylinders cast using steel moldswere used to determine the concrete compressive strength on the day the beams were tested.2.4.3 CuringTest cylinders and flexure beam specimens were stored and cured next to the beam-splice specimens and under similar conditions of temperature and humidity. The beams were coveredwith wet burlap immediately after finishing of the beam surface. The beams, flexure beams, andthe 6 x 12-in, cylinder specimens were moist-cured by keeping them covered with wet burlapand plastic until the measured compressive strength of the concrete exceeded 3500 psi. Theplastic cylinder molds were sealed with plastic caps during the period in which the beams werewet cured.The beam formwork and the molds were removed after the 3500 psi threshold wasexceeded. After demolding and removal of the forms, the specimens were air-cured until themeasured compressive strength reached 5000 +500 psi.2.4.4 Test apparatus The beam-splice specimens were tested using a four-point loading configuration (Figure2.1 and Figure 2.7). To facilitate inspection of the splice region during the test, the loads wereapplied in the downward direction (Figure 2.7) so that the main flexural reinforcement would belocated at the top of the beam. The splice region was located between the two supports (FigureREDACTED VERSION

I-age IuI 0? 11142.7) in the central constant moment region of the beam. The final location of the supports wasmeasured (to the nearest '/8-in.) and is reported in Table F.4 (Appendix F). As-built externaldimensions of each test beam were recorded using the same form. The maximum deviation fromnominal dimensions in the test region was 1/2/ in.Figure 2.7 -Four-point loading configuration Loads were applied at the ends of the specimen using two loading frames, as shown inFigure 2.7. Each loading frame consisted of two load rods attached to a loading beam that wasplaced above the specimen. Loading was imposed through dual-acting center-hole hydraulic rams attached to the lower surface of the reaction floor. At the start of the test, the lower end ofthe load rods passed through the reaction floor without applying load to the specimen other thanthe weight of the loading frame and the rods. A total of four rams were used, two for eachloading frame. High-pressure hydraulic lines connected the rams to separate pressure and returnmanifolds, which were connected to the pressure and return lines of a single hydraulic pump. Allhoses and other hydraulic hardware were inspected visually before testing began.The beams were instrumented to measure displacement and load. As shown in Figure 2.8,the applied load was measured with load cells mounted on the load rods, and displacements wereREDACTED VERSION bbBJ-UuU-U-(S-.ji-ULU-UUU~t)-UUU t-'age (Ud 0? 1 114recorded using displacement transducers and dial gages (for redundancy) at the center of thebeam and at each of the two load points.Within each specimen, 350-ohm 1/4-in, strain gages with attached leads were bonded tothe spliced bars, approximately 2 in. outside the edges of the splice. One deformation in each barwas removed using low-heat grinding to provide a smooth surface to attach the strain gages.Strain gages were attached to the bars using epoxy cement and sealed following therecommended procedures by the manufacturer for submersion in concrete. The strain gages wereplaced so that the coating used to seal the strain gages covered only deformations outside of thesplice region. The strain gages provided little useful data.Figure 2.8 -Loading apparatus and instrumentation at each load pointREDACTED VERSION

5-UUU I-'age aT 1 1142.4.5 Loading protocolThe double acting rams were fully retracted prior to the start of each test. With theloading rams in the fully retracted

position, slack was taken out of the load rods by tightening thenuts until each load rod was nearly engaged with the fully retracted hydraulic jacks, but withoutapplying any load. This procedure was adopted to prevent rotation of the loading beams andconsequently maintain even loading across all four rods.Before load was applied, all displacement transducers, load cells, and strain gages werezeroed and initial readings were recorded for each of the three dial gages. Data were recordedcontinuously by the data acquisition system with a sampling rate of approximately one sampleper second. Recorded data was continuously appended to a data file to prevent any loss of data incase of system failure.Load was applied using a single manually-controlled hydraulic pump. Loading wasstopped at predetermined load levels to visually inspect the beam, mark visible cracks (identified based on the average value of the load applied at one end of the beam, as illustrated in Figure2.9), measure crack widths using crack comparators, and to record strain and dial gage readings.

The specimens were marked before each test to indicate the locations of the ends of the spliceregion, the beam centerline, the pin and roller (pedestal)

supports, and the load apparatus.

Themarkings, shown in Figure 2.10, were 'SR' to indicate the ends of the splice region, 'CL' for thecenterline of the beam, and 'PS' for the center of the pedestal support. All longitudinal measurements were taken using the centerline of the beam as a reference point to eliminate anyinconsistencies caused by small deviations from the nominal length in the specimens. REDACTED VERSION

s-UUU I-'age flU Ot 1 114Figure 2.9 -Crack inspection and marking during test(a)REDACTED VERSION bbYJ-UUU-UdJL-ULUI--UUU'Ib-UUU I-age t Ii O't 1i 14(b)Figure 2.10 -Beam marks: (a) End of splice region and centerline of the beam; (b) pedestalsupport centerline The initial load increment was chosen to be smaller than one half of the calculated flexural cracking load to ensure that all instruments and the hydraulic system were operating properly. From this point forward, loading proceeded in increments of approximately 5 kips ateach end of the beam. The final load step at which cracks were marked was approximately two-thirds of the estimated failure load. In some of the specimens, the loading protocol was such thatthe specimens were unloaded after the formation of a horizontal crack with a width of at least 10mils in the splice region. After the specimen was fully unloaded, it was loaded to failurefollowing the procedure specified above for monotonically-loaded specimens. The loadingprotocol used for each beam is presented in Table 2.5.A log was maintained to record any meaningful observations during the test, such as loadcorresponding to flexural

cracking, crack widths, file names, and gage readings.

The logs arepresented in Appendix H.After failure, cracks were marked on the specimens with each identified using thepreliminary value of the average maximum end load (this value typically deviated by a fewpercent from the recorded value).The following data were recorded and continuously transferred to disk throughout eachtest:REDACTED VERSION bbi-UUU-Ubi-UL::-UUU1 b-UUU P'age /1 o1 Grill4-Force applied to each load rod-Displacement at midspan and each load application point-Strain in the reinforcing steelTable 2.5 -Detailed loading protocol for each beamBeam Loading Protocol1 (1)Monotonically-increasing load up to an average end load of 40 kips in increments of 5 kips.At the end of the each increment, the beam was inspected for cracks and dial-gage displacement measurements were recorded. (2) Loading resumed with increasing displacement until failure.2 (1)Monotonically-increasing load up to an average end load of 25 kips in increments of 5 kips.At the end of the each increment, the beam was inspected for cracks and dial-gage displacement measurements were recorded. (2) Dial-gage measurements were recorded at an average end load of 30 kips.(3) Loading resumed with increasing displacement until failure.3(1)Monotonically-increasing load up to an average end load of 30 kips in increments of 5 kips.At the end of the each increment, the beam was inspected for cracks and dial-gage displacement measurements were recorded. (2) The beam was frilly unloaded and dial-gage displacement measurements were recorded. (3) The beam was loaded a second time up to an average end load of 35 kips in load increments of 5 kips. At the end of the each increment, dial-gage displacement measurements wererecorded. The beam was inspected for cracks at an average end load of 30 kips.(4) Loading resumed with increasing displacement until failure.4(l)Monotonically-increasing load up to an average end load of 35 kips in increments of 5 kips.At the end of the each increment, the beam was inspected for cracks and dial-gage displacement measurements were recorded. (2) Loading resumed with increasing displacement until failure.5(1)Monotonically-increasing load up to an average end load of 40 kips in increments of 5 kips.The beam was inspected for cracks and dial-gage displacement measurements were recorded atthe end of each increment. (2) The beam was fully unloaded and dial-gage displacement measurements were recorded. (3) The beam was loaded a second time up to an average end load of 40 kips in increments of 5kips. Dial-gage displacement measurements were recorded at the end of each increment. Thebeam was inspected for cracks at average end loads of 20, 30, 35 and 40 kips.(4) Loading resumed with increasing displacement until failure6(1)Monotonically-increasing load up to an average end load of 40 kips in increments of 5 kips.The beam was inspected for cracks and dial-gage displacement measurements were recorded atthe end of the each increment. (2) The beam was fully unloaded and dial-gage displacement measurements were recorded. (3) The beam was loaded a second time. The beam was inspected for cracks and dial-gage displacement measurements were recorded at average end loads of 10, 20, 30, 35, and 40 kips._____(4) Loading resumed with increasing displacement until failure.REDACTED VERSION

b-UUU i-'age (15 T01 111 42.4.6 Calibration Instruments used to measure force and displacement were calibrated following theprocedure specified in this section. The applied load was measured using load cells.Displacement transducers (either linear variable differential transformers or string potentiometers depending on availability) were used to record the vertical beam deflections. Load cells anddisplacement transducers were calibrated using a digitally-controlled hydraulic test framecalibrated annually using NIST-traceable standards. Load cell and displacement transducers werecalibrated following the steps listed below:1) The sensor (load cell or displacement transducer) was connected to the data-acquisition system that was used in the test.2) The sensor was securely mounted on the testing machine.3) A series of known force or displacement increments were applied to the sensor.Calibrations were performed exceeding the displacement and load range expectedduring the tests. In the case of load cells, calibrations were performed between zeroand 100 kips. In the case of displacement, calibrations were performed in a rangebetween zero and 4 in.4) Sensor output was recorded with the data-acquisition system at each knowndisplacement or force increment.

5) A least-squares linear regression analysis was performed on force and displacement versus sensor output to determine the calibration constant.

The load cells and displacement transducers were calibrated before and after testing eachthree beams and the calibration results are reported in Appendix G. The calibration constantdeviated with an average value of 0.28% for all sensors, ranging between 0 to 0.84%.2.5 Test Facilities The tests were performed in the Structural Testing Laboratory at the University of Kansas,a facility of the KU Structural Engineering and Materials Laboratory (SEML). The Laboratory has static and servo-hydraulic test equipment for the testing of steel, concrete, and composites. The structural testing bay has 4000 square feet of open laboratory area with a clear height of 30REDACTED VERSION

I-age 1140oT 1 114ft for large-scale structural testing. Loads up to 100,000 lb can be applied on 3-ft centers over a50 x 80 ft area. The laboratory houses a 600,000-lb universal testing machine for testing steeland concrete. A 450,000-lb MTS Structural Test System supported on a four-column test framemay be used for dynamic and cyclic testing of large scale structural components. 110,000-lb and55,000-lb MTS Structural Test Systems are also used for cyclic and dynamic testing of full-scale structural components within the test bay. Actuators within the test bay are powered by twohydraulic pumps (total flow rate of 110 gpm), meeting the requirements for demanding cyclictest applications. High-speed Mars Labs, National Instruments (used in the current study), andHewlett Packard data acquisition systems are available to monitor and record load, strain, anddisplacement. The structural testing laboratory includes an overhead 20-ton crane with access tothe entire lab floor area. Over 500 beam-end tests and over 200 splice tests have been performed in the KU Structural Testing Lab since 1990.Material tests were performed in the Concrete Laboratory, another SEML facility, whichis equipped to run standard tests on cement, aggregates, and concrete. Equipment is available totest concrete aggregate for deleterious

behavior, including alkali silica reactivity, and to measureaggregate properties as they affect mixture proportioning.

Freeze-thaw equipment is available forrunning tests under both Procedures A and B of ASTM C666. A walk-in freezer is used forscaling tests. Concrete is cured under controlled temperature and humidity in the lab's curingroom. Two hydraulic testing machines, with load capacities of 180 tons (400,000 lb),'oare usedfor concrete strength determination. Certificates of calibration for the equipment used in this study, including for the test frameused to calibrate the sensors, are presented in Appendix I.2.6 Section AnalysisSplice strength was evaluated based the calculated moment in the splice region at failure(ACI 408R-03). Loads, moments, and stresses for the beams were calculated using a two-dimensional analysis in which loads and reactions were assumed to act along the longitudinal centerline of the beam. Reactions and moments were based on load cell readings and the weightof the loading assemblies. The self-weight of the beam was included in the calculations based onaverage beam dimensions and an assumed concrete density of 150 pcf.REDACTED VERSION

r-age (lb o? 1114The test specimens were evaluated using cracked section theory assuming a linear straindistribution through the height of the cross-section. The beams were analyzed using anequivalent rectangular stress block and moment-curvature analyses for comparison. Themoment-curvature relationship was derived using the nonlinear stress-strain relationship forconcrete proposed by Hognestad (1951) and follows the procedure described by Nilson, Darwin,and Dolan (2010). Figure 2.11 shows the assumed stress distribution in the compression zone forthe moment-curvature and the equivalent rectangular stress block analyses. Good agreement inthe calculated bar stress at failure was typically noted between results obtained with the twomethods.O.85f; aC.=bO.85f-b and equivalent rectangular stress block analyses [after Nawy (2003)]In calculating splice strength, the tensile stress in the steel f5 (ksi) was calculated asfollowing the procedures used by ACI Committee 408 (2003):f,= Ex s= 29000 x ts forf, measured yield strengthfy (1)For es >fy/E5, f5 =fy for e, .e'h, where Csh = 0.0086 forty = 60 ksi and 0.0035 forty = 7ksi and above. There is no flat portion of the stress-strain curve forty 101.5 ksi. The modulusof strain hardening ELh = 614 ksi forty = 60 ksi, 713 ksi forty = 75 ksi, and 1212 ksi forty 90ksi. The values of 8sh and ELb forty between 60 and 90 ksi are obtained using linear interpolation. The equivalent rectangular stress block used in the calculations was proposed by Whitneywith the values of the parameter /J1 specified in ACI 318-11. The moment-curvature relationship was calculated using the concrete model proposed by Hoguestad (1951).REDACTED VERSION

M-age fibO0? 1114for= t,. o J(2)Io (i °-zc J1]1 for f"7= 0.85/' (3a)-1.7/'(b% = O.0038 (3c)= 1.8x106 +460fc' (3d)where:fc=concrete stress, psif"= concrete compressive

strength, psif"= peak concrete stress, psiSc= concrete strains0 = concrete strain at peak stressec,= ultimate concrete strain at crushingEc= approximate concrete modulus of elasticity, psiTensile stresses carried by the concrete were neglected in both analyses.

The calculations using both equivalent rectangular stress block and moment-curvature analyses proceed as follows:1. Select top face concrete strain sc in the inelastic range.2. Assume the neutral axis depth, at distance c below the top face.3. Assuming a linear variation in strain throughout the depth of the member, determine thetensile strain in the steel ss (equal to the tensile strain in the concrete at the level of thesteel sc).4. Compute the stress in the reinforcing steel in accordance with the defined stress-strain relationships (above). The tensile force in the steel T=fi x As (see Figure 2.11).REDACTED VERSION btYUi-UUU-UbJ-UbU--UUUlb-UUU P-age (11 oT 11 145. Determine the compressive force C, which equals to 0.85 fcba (Figure 2.11ib) for theequivalent rectangular stress block method, or by numerically integrating the concretestresses as defined by Eq. (2) and (3) for the moment curvature method.6. If C -- T, go to step 7. If not, adjust the neutral axis depth c in step 2 and repeats steps 3 -5.7. Using the internal lever arm z from the centroid of the concrete stress distribution to thetensile resultant, the calculated bending moment M= Cz = Tz.8. If the calculated bending moment M equals the applied bending moment (from test), equals the force in the reinforcing steel. If the calculated bending moment does not equalthe applied bending moment, modify ec and c in steps 1 and 2, respectively, and repeatsteps 3 -7 until the calculated bending moment M equals the applied bending moment.REDACTED VERSION

r~age fib5 oT 11143 Test Results3.1 GeneralThe testing program consisted of six beam-splice specimens. Three of the specimens had alap splice length of 79 in., and three had a lap splice length of 120 in. The measured loads andcalculated bar stresses at failure are presented in Table 3.1. In addition to failure loads, Table 3.1includes measured material properties and bar cover dimensions. Bar stresses at failure listed inTable 3.1 include those calculated using the equivalent rectangular stress block and moment-curvature analysis. Measured specimen dimensions and other details of the beam tests arepresented in Appendix H.Moment-curvature analyses consistently produced calculated higher bar stresses than didthe analysis using the equivalent rectangular stress block. This is to be expected because theparameters of the equivalent stress block were calibrated to reflect the characteristics of thecompression zone when the peak strain in the concrete exceeds 0.003 and the concrete in thecompression zone is well into the nonlinear range. Under these conditions, the depth of thecompression zone is reduced, resulting in a slightly larger distance between the tension andcompression resultants. With the exception of Beam 1, the splices failed prior to crushing of theconcrete in the bottom surface of the beam, so it was to be expected that the equivalent rectangular stress block would slightly overestimate the distance between tension andcompression resultants and consequently underestimate the stress in the reinforcing bars. Thedifference, on average, between the bar stresses at failure calculated by the two methods was 1.5ksi for the six beams tested in this study, with moment-curvature analysis producing the greatervalue. All bar stress values discussed subsequently are those calculated using moment-curvature

analysis, which is considered to be more accurate method for the reasons stated above.REDACTED VERSION bbtSi-UUU-6bI3-UbUr_-UUUlb-UUU Page (19 aT"1114Table 3.1 -Bar stresses at failure for beam-sp lice specimens Total ,CacltdICalculated bar stressBaID-Sl Ce oConcretemomentaat at failure, ksiBemD-Spie

strength, Cocrt splice omnatFailure modelegt pi cover; in.a failure, splice Equiv. Moment-legt ' failure, kip-ft stress "kips block curvature 1 -79 in. Flexural(monolithic) 3// 103 344 70 70 Failure *2 -79 in.(cold joint, loaded 53/ 3/3/3 85 292 59 62 Spiue*monotonically) 5330/+alu&

3 -79 in.(cold joint, Spliceunloaded and 3.25/3.35/2.9 80 270 53 57failure** reloaded) Splice failure4 -120 in. and(cold joint, loaded 3/2.8/3.4 105 350 71 72 secondary monotonically) flexuralfailure** 5 -120 in. Splice failure(cold j oint, 523 0/ andunloaded and 5490k 3.15/3/15/3.15 96 325 66 67 secondary flexuralreloaded) failure*** 6 -l120in. Splice failure(cold joint, andunloaded and 3.15/3.15/2.9 100 338 69 69 secondary flexuralreloaded) failure~*** " Top cover/north side cover/south side cover+ Compressive strength of concrete below and above the cold joint.Test was stopped after reinforcing steel yielded, when crushing of the concrete in the compression zone wasobserved.

• Splice failed prior to yielding of the flexure reinforcement.

Splice failed after yielding of the flexure reinforcement 3.2 Beams 1, 2, and 3 with 79-in, splice length3.2.1 Concrete strengthThe concrete strengths for Beams 1, 2 and 3 are summarized in Table 3.2. Beam 1 wascast monolithically, while Beams 2 and 3 were cast in two stages to accommodate the presenceof a cold joint at the level of the flexure reinforcement. Beam 1 and the concrete below the coldjoint for Beams 2 and 3 were placed on May 24, 2012 and the concrete above the cold joint wasplaced on May 25, 2012. The forms were removed on May 28, 2012, when the average concreteREDACTED VERSION DbL-UUU-U5di-UI-.i-UUU1 b-UUU l-'age /2U Ot 1 114compressive strength for both placements exceeded 3500 psi. All three beams were tested onMay 31, 2012. On that date the concrete from the first placement had an average compressive strength of 5330 psi, and the concrete from the second placement had an average compressive strength of 4330 psi (Table 3.2). The average split cylinder strength and the average modulusrupture were 435 and 570 psi for the concrete below the cold joint in accordance with ASTMC496 and ASTM C78, respectively. The tensile strength for the concrete above the cold jointwas not recorded for the first three beams. The flexural beam specimens with cold joints werealso tested and had an average modulus of rupture of 140 psi, significantly lower than that ofspecimens cast monolithically. The fact that the tensile strength of the flexural beam specimens with cold joints was significantly lower than the strength of monolithic specimens indicates thatthe presence of a cold joint did in fact introduce a weak plane at the level of reinforcing steel.The proportions of the concrete mixture and the properties of the concrete for each placement arereported in Table E.2 of Appendix E.Tahle 3.2 Concrete strengths for Beams 1, 2, and 3Concrete below cold joint Concrete above cold jointAverage Compressive Strength when 4010oa360 forms were removedAverage Compressive Strength at test 5330c430date, psiSplit Cylinder Strength (ASTM 435C -C496), psiModulus of Rupture (ASTM C78), psi 570C -Modulus of Rupture for specimens 140with cold joint, psiaTested at 4 days; btested at 3 days; ctested at 7 days; dtested at 6 daysA segment of the No. 11 bars used in the splice-beam specimens was tested in tensionand the bar strains were recorded using a linear variable differential transformer (LVDT ) used asthe extensometer (gage length = 8.0 in.). The measured stress-strain curve for the No. 11 bar isshown in Figure 3.1. The yield stress calculated using the 0.2% offset method was 67 ksi and themeasured elastic modulus was 28,990 ksi. The maximum measured steel stress was 105 ksi.REDACTED VERSION

I-page (21 0? 1114120000100000600080000-- --40000III Yield (0.2% offset): 67.1 ksi20000 _____-________ Ultimate:-104.7 k si...FIEasicModulus:28990 ksi~lsi00 0.05 0.1 0.15StrainFigure 3.1 -Measured stress-strain curve for No. 11 bar3.2.2 Beam 1 (monolithic concrete) 3.2.2.1 Beam 1 load-deflection curveBeam 1 was cast monolithically with a splice length of 79 in. It was loadedmonotonically to failure (the load protocol is presented in Table 2.5). The load-deflection curvefor Beam 1 is shown in Figure 3.2. The displacement shown in the figure was calculated byadding the average of the displacement at the two load points to the displacement at the beamcenterline. The load shown in the figure corresponds to the total load applied to the beam (thesum of the two end loads). The load-deflection relationship shows that there was a significant reduction in the stiffness of the beam at a total load of approximately 20 kips, which coincided with the first observation of flexural cracks. Another significant reduction in flexural stiffness was observed at a total load of 94 kips and a total displacement of approximately 2.8 in. In thiscase the reduction in stiffness is attributed to the yielding of the flexural reinforcement. Thecalculated bar stress corresponding to the total load of 94 kips is 68 ksi based on moment-curvature analysis (Table 3.1). The positive slope of the load-deflection relationship after a totalload of 94 kips is attributed to the strain hardening of the reinforcing steel. Loading continued until a flexural failure occurred, which was accompanied by crushing of the concrete in theREDACTED VERSION bDJI-UUU-Uditt-UWI:-UUUI h-UUU -'age 122Z Ot 1 114compression zone, near the supports, at a total load of 103 kips, corresponding to a bar stress of70 ksi, and a total deflection of approximately 5 in. (Figure 3.3).0.0d0110100908070605040302010002 3 45Total Deflection, in.6Figure 3.2 -Total load vs. total deflection for Beam 1 (cast monolithically) (Total loadcalculated as the summation of the two end loads and total deflection calculated defined byadding the average deflection at two ends and the deflection in the beam centerline). Figure 3.3 -Flexural failure in the compression region for Beam 1. Numbers indicate maximumaverage end load when cracks marked.REDACTED VERSION

r-age f"J OT II1143.2.2.2 Crack progression-Beam 1Maximum measured crack width versus average end load for Beam 1 is shown in Figure3.4; the crack map for Beam 1 is presented in Figure 3.5 (see figures in Appendix C for greaterdetail). The first flexural cracks formed near the east support at the end of the east splice region,at an average end load of 10 kips (total load of approximately 20 kips). The flexural cracks grewprogressively wider and more numerous as the load increased. The first horizontal crack formednear the support at an average end load of 25 kips (Figure 3.6). Both longitudinal and flexuralcracks continued to increase in width and number as the load increased. At the last crackmarking prior to failure (average end load of 40 kips), the widest flexural crack had a width of 25mils (0.025 in.) and the widest horizontal (bond) crack had a width of 18 mils (0.018 in.).60) 50.E10,0 10 20 30 40 50Average End Load, Kips-E-FlexuraI Crack -+Horizontal CrackFigure 3.4 -Maximum crack width vs. average end load (one-half of total load) for Beam 1.REDACTED VERSION bbJL-UUU-dJtiSULILI-UUU1 U-UUU r-age (Z4 Ot 1 114EaweweNorth FaceSouth FaceTop FaceI aLoading Pedestal Splice Center Splice Pedestal LoadingPoint Support Region Line Region SupportPon Ea$tFigure 3.5 -Crack map for Beam 1. Numbers indicate maximum average end load when cracksmarked. See Figure C. 1 in Appendix C for greater detail.Figure 3.6 -Beam 1, north side of east support with horizontal crack, 25 kip end load.Failure occurred at an average end load of 51 kips (total load of 103 kips). The failuremode was yielding of the bars followed by crushing of the concrete near the supports (Figure3.7). Both flexural and horizontal cracks were present near the splice region (Figure 3.8). At theREDACTED VERSION bbUJ~-UUUJ-U-LG--UUU"Ib-UUU I'-'age /b OT 01 114support (Figure 3.9), flexural cracks extended most of the depth of the beam; no horizontal cracks were present.A detailed autopsy was not performed on Beam 1. Concrete was removed in selectedregions to verify the concrete cover to the splice was within tolerances. Top cover was 3 in. tothe outer bar in the splice and 3-1/4 in. to the inner bar in the splice for both splices.Figure 3.7 -Beam 1, underside near support, failure.REDACTED VERSION

r-age (~t 0T 1 114Figure 3.8 -Beam 1, north side of west splice region, failure.Figure 3.9 -Beam 1, south side of east support, failure.REDACTED VERSION

r-age f(2 0T1 1114Figure 3.10 -Beam 1, centerline, failure.3.2.3 Beam 2 (cold joint, monotonically-loaded) 3.2.3.1 Beam 2 load-deflection curveBeam 2 was cast with a cold joint in the plane of reinforcing steel. It was monotonically loaded with a load increment of approximately 5 kips (average end load, the load protocol ispresented in Table 2.5). The load-deflection curve for Beam 2 is shown in Figure 3.11. The totaldisplacement and total load shown in the figure were calculated in the same manner as for Beam1. The total load corresponding to cracking was very similar to that of Beam 1, approximately 20kips. The beam was loaded to a maximum total load of 85 kips, with a corresponding totaldisplacement of 2.25 in. At this point the beam failed with a sudden splitting of the concretealong the cold joint. Wide horizontal cracks were observed in the plane of the cold joint withinthe splice region (Figure 3.12). The widest horizontal crack was measured to be 1/22 in. wide afterfailure. It is concluded that the beam failed due to failure of the splice at a total load of 85 kips.The calculated bar stress corresponding to the total load of 86 kips is 62 ksi based on moment-REDACTED VERSION

i-8ge 0? 1I14curvature analysis (Table 3.1), above the minimum specified yield strength of 60 ksi for Grade60 reinforcement but 5 ksi below the actual yield strength of 67 ksi.0d9080706050403020100..... ... ..... .85 kip s012Total Deflection, in.34Figure 3.11 -Total load vs. total deflection for Beam 2 (with a cold joint)Wide horizontal crack at failurewithin the splice regionFigure 3.12 -Beam 2 (with a cold joint) failed with wide horizontal crackREDACTED VERSION

I-UUU I-'age hiB ot 11143.2.3.2 Crack progression-Beam 2Maximum measured crack width versus load for Beam 2 is shown in Figure 3.13; thecrack map for Beam 2 is presented in Figure 3.14. The first flexural cracks formed near thesupports and ends of both splice regions at an average end load of 15 kips (total load of 30 kips).Horizontal cracks first formed at an average end load of 20 kips at both ends of the splice regionalong the cold joint (Figure 3.15). Both longitudinal and flexural cracks continued to increase inwidth and number as the load increased, with horizontal cracks propagating along the cold joint.When the last cracks were marked prior to failure (conducted at an average end load of 30 kips),the widest flexural crack had a width of 20 mils (0.02 in.) and the widest horizontal crack had awidth of 13 mils (0.0 13 in.).60~4oU30I..0 10 20 30 40 50Average End Load, Kips--Flexural Crack Horizontal CrackFigure 3.13 -Maximum crack width vs. average end load for Beam 2.REDACTED VERSION DbiL-UUU-U-LJU-S-t-UUUIIb-UUU I-age fiU t 0 1114EaStweS1North FaceSouth FaceTop Face.5 IL 27 IWesp. ~,7- I<" I tw reSLoadingPointPedestal Splice Center SpliceSupport Region Line RegionPedestalSupportEasLoadingPointFigure 3.14 -Crack map for Beam 2. Numbers indicate maximum average end load whencracks marked. See Figure C.2 in Appendix C for greater detail.Figure 3.15 -Beam 2, northeast support with horizontal crack, 20 kip end load.Failure occurred at an average end load of approximately 43 kips (total load of 85 kips).At failure, the concrete above the cold joint separated from the remainder of the beam (Figure3.16). Near the splice region, a large flexural crack was also present (Figure 3.16). TheREDACTED VERSION

w'age ,1i Or 11 14horizontal crack progressed approximately 12 in. past both ends of the splice region, and with theexception of near the centerline, continued along the cold joint. At the centerline, the crack splitthrough the cover and around the single hoop present at the centerline (Figure 3.17), indicating the hoop was effective in preventing the crack from growing near the centerline. As shown inFigure 3.17, the region affected by the hoop was small.Figure 3.16 -Beam 2, southwest splice region showing separation of concrete, 43 kip end load.REDACTED VERSION bbI-UUU-U56.i-{U:-UUU1bJ-UUU r-age 1;S2 0O" 1 114Figure 3.17 -Beam 2, centerline at failure.3.2.4 Beam 3 (cold joint, cycled)3.2.4.1 Beam 3 load-deflection curveBeam 3 was cast in the same manner and at the same time as Beam 2, with a cold joint inthe plane of reinforcing steel. Instead of loading the beams to failure monotonically, Beam 3 wasfirst loaded to a total load of 60 kips, unloaded to zero, and then re-loaded monotonically tofailure (the load protocol is presented in Table 2.5). When the beam was first loaded to a totalload of 60 kips (average end load of 30 kips), the average end load was increased in increments of approximately 5 kips. The specimen was inspected for cracks, which were marked at eachload step. At a total load of 60 kips, the maximum horizontal crack width was 20 mils (0.02 in.).When the beam was loaded for the second time, it was loaded up to a total load of 60 kipswithout inspecting for cracks. The only visual measurement conducted during the second loadingwas the recording of dial gage readings at approximately 5-kip increments (average end load).The beam was inspected for cracks again when the total load reached 60 kips for the secondtime. At this point some of the horizontal cracks widened to a maximum width of 35 mils (0.035in.)REDACTED VERSION

l-'age /'JJ. OT 1114The load-deflection curve for beam 3 is shown in Figure 3.18. Overall, Beam 3performed very similar to Beam 2, except for the peak load. The beam failed at a total load of 80kips (compared with a total load of 85 kips for Beam 2), in the same manner as observed forBeam 2. A wide horizontal crack in the plane of the cold joint, within the splice region, wasobserved after failure (Figure 3.19), with the widest portion of the crack being 3/8-in. It isconcluded that the beam failed due to a splice failure. The calculated bar stress corresponding tothe total load of 80 kips is 57 ksi based on moment-curvature analysis (Table 3.1).rj~Cu0Cu0908070605040302010... ............. .......... .. .... ... ..._ 8 0O k ip s012 3 4Total Deflection, in.Total load vs. total deflection for Beam 3 (with a cold joint)Figure 3.18Figure 3.19 -Beam 3 failure with wide horizontal cracks along cold jointREDACTED VERSION

h-UUU I-sage f 5440t 11143.2.4.2 Crack progression-Beam 3Maximum measured crack width versus load for Beam 3 is shown in Figure 3.20; thecrack map for Beam 3 is presented in Figure 3.21. As seen in both figures, the first flexuralcracks formed near end of the east splice region at an average end load of 10 kips (total load of20 kips). At an average end load of 15 kips, flexural cracks were present at both ends of thesplice region and both supports. A horizontal crack first formed at an average end load of 15 kipsat the west end of the splice region along the cold joint, with additional horizontal cracksforming and reaching a 9-mail (0.009 in.) width at an average end load of 20 kips (Figure 3.22).At an average end load of 30 kips, a 40-mit (0.04-in.) width flexural crack and 20-mait widthhorizontal crack were recorded. At this point, the beam was unloaded. With zero load, themaximum flexural and horizontal crack widths decreased to 13 and 7 mils (0.013 and 0.007 in.),respectively. The load was reapplied, and at the last crack mapping (average end load of 30kips), the widest flexural crack had a width of 55 mils (0.05 5 in.) and the widest horizontal crackhad a width of 35 mils (0.03 5 in.), much wider than the cracks noted at the first loading to a 30-kip average end load.60O°E5040S30E 2100AI0 10 20 30 40 50Average End Load, Kips-~-Flexural Crack (1st loading) Crack (1st Loading)--Flexural Crack (2nd loading) -*- -Horizontal Crack (2nd Loading)Figure 3.20 -Maximum crack width vs. average end load for Beam 3.REDACTED VERSION

I-'age 1.iD 0?" 1114a,iNorth FaceYWe"We tSouth Face-(IE~.. .... .... .... .. tTop FaceWe r1 ........ f e l .. .. ..Pedestal Sp~lceSupport Region.... '. l .[-Cettef Splice Pedestaltine Reglon Supor... I IaEI !adnFigure 3.21 -Crack map for Beam 3. Numbers indicate maximum average end load whencracks marked. See Figure C.3 in Appendix C for greater detail.Figure 3.18 -Beam 3, northwest splice region with horizontal crack, 20 kip end load.Failure occurred at an average end load of 40 kips (total load of 80 kips), a slightly lowerload than the monotonically loaded Beam 2 (total load of 85 kips). At failure, the concrete abovethe cold joint separated from the remainder of the beam, with the horizontal crack propagating along the cold joint in a region that was somewhat larger than the splice region except for a smallREDACTED VERSION bbiL-UUU-bJ5i-GW-(-UUU1ti-UUU Mage I Jti 01" 1114region near the centerline, which was restrained by the No. 3-bar hoop (Figure 3.23). Largeflexural cracks were also present near both ends of the splice region.Figure 3.23 -Beam 3, splice region and centerline showing separation of concrete, 40 kip endload.3.3 Beams 4, 5, and 6 with 120-in, splice length3.3.1 Concrete strengthThe concrete strengths for Beams 4, 5 and 6 are summarized in Table 3.3. The threebeams were cast in two stages to accommodate the presence of a cold joint at the level of theflexural reinforcement. The concrete below the cold joint was placed on June 13, 2012, and theconcrete above the cold joint was placed on June 14, 2012. The forms were removed on June 17,2012 when the average concrete compressive strength for both placements exceeded 3500 psi.The beams were tested on June 20, 2012. On that date, the concrete from the first placement hadan average compressive strength of 5230 psi, and the concrete from the second placement had anaverage compressive strength of 5490 psi (Table 3.3). The higher strength for the secondREDACTED VERSION

-'age (i/ 0T71114placement was likely due to the slightly lower water-cement ratio of the concrete, as shown onthe batch ticket (Appendix H). The average split cylinder strength and average modulus rupturewere, respectively, 370 and 600 psi for the concrete below the cold joint and 470 and 700 psi forthe concrete above the cold joint. The flexural beam specimens with cold joints were also testedand had an average modulus of rupture of 274 psi, significantly below that of specimens castmonolithically. The proportions of the concrete mixture and the properties of the concrete foreach placement are reported in Table E.2 of Appendix E.Table 3.3 -Concrete strengths for Beams 4, 5, and 6Concrete below cold joint Concrete above cold jointAverage Compressive Strength when 4310oa450 Forms were removedAverage Compressive Strength at test 5230c540date, psiSplit Cylinder Strength (ASTM 370° 70C496), psi ___________ Modulus of Rupture (ASTM C78), psi 600c 700dModulus of Rupture for specimens 274d __with cold joint, psiaTested at 4 days; btested at 3 days; ctested at 7 days; dtested at 6 daysThe same reinforcing steel was used for Beams 4, 5, and 6 as for Beams 1, 2, and 3. Themeasured stress-strain curve for the No. 11 bar is shown in Figure 3.1.3.3.2 Beam 4 (cold joint, monotonically-loaded) 3.3.2.1 Beam 4 load-deflection curveBeam 4 was cast with a cold joint in the plane of reinforcing steel. It was subjected tomonotonically-increasing load in increments of approximately 5 kips (average end load, theloading protocol is presented in Table 2.5). The load-deflection curve for Beam 4 is shown inFigure 3.24. The total load and deflection were determined in the same manner as for Beams 1, 2and 3. The flexural stiffness of the beam decreased once the total load exceeded 20 kips,coinciding with the formation of flexural cracks. A sharp decrease in the slope of the load-deflection curve was observed at a total load of about 94 kips and corresponding deflection ofREDACTED VERSION bbdi-UUU-(.ibi-ti-L~.UUUI b-UUU r'age lid Ot 1114approximately 2.8 in. The stress at the end of the spliced bars for a total load of 94 kips was 68ksi. The decrease in the slope of the load-deflection curve at a total load of 94 kips indicates thatthe reinforcing steel yielded. After yielding of the reinforcing steel, the total load continued toincrease but at a lower rate, which is attributed to the strain hardening of the reinforcing steel.The beam was loaded to a total load of 105 kips (and a displacement of 5.5 in.) and at that pointfailed with the sudden splitting of the concrete along the cold joint. Wide horizontal cracks in theplane of the cold joint were observed within the splice region. Wide flexural cracks were alsoobserved near the support (Figure 3.25). It is concluded that the reinforcing steel yielded at atotal load of approximately 94 kips and beam failed at a total load of 105 kips due to failure ofthe splice, the latter corresponding to a bar stress of 72 ksi (Table 3.1).11010090~70*;40~3001000 2 4 68Total Deflection, in.10Figure 3.194 -Total load vs. total deflection for Beam 4 (with a cold joint)REDACTED VERSION

I-sage fJ31 ot 1114Figure 3.20 -Beam 4 (with a cold joint) at failure3.3.2.2 Crack progression-Beam 4Maximum measured crack width versus load for Beam 4 is shown in Figure 3.26; thecrack map for Beam 4 is presented in Figure 3.27. The first flexural cracks formed near end ofthe west support at an average end load of 10 kips (total load of 20 kips). At an average end loadof 15 kips, flexural cracks were present at both ends of the splice region and both supports. Horizontal cracks first formed at an average end load of 20 kips, at the both ends of the spliceregion along the cold joint. Both longitudinal and flexural cracks continued to increase in widthand number as the load increased, with horizontal cracks propagating along the cold joint. At thelast load prior to failure at which cracks were marked (average end load of 35 kips), the widestflexural crack had a width of 30 mils and the widest horizontal crack had a width of 16 mils. Atthis point, the horizontal cracks extended along most of the length of the splice region (Figure3.28), with some of the horizontal cracks that formed at earlier stages merging together. REDACTED VERSION

I-age / 4U Ot 1 11460E20 03 05Average End Load, Kips--Flexural Crack Horizontal CrackFigure 3.21 -Maximum crack width vs. average end load for Beam 4.EaWwtNorth FaceSouth Face-r~ <:9~~I:ii2F~i2 -- -. -i7SWeT~ tIaPointwe$odnPo~fTop Face-IlefleLineStqo uppOrtFigure 3.27 -Crack map for Beam 4. Numbers indicate maximum average end load whencracks marked. See Figure C.4 in Appendix C for greater detail.REDACTED VERSION

I-age (41 OT 1 114Figure 3.28 -Beam 4, south side of west splice region with horizontal cracks, 35-kip end load.At failure, the concrete above the cold joint separated from the remainder of the beam,with the horizontal crack propagating along the cold joint between the pedestal supports exceptfor a small region near the centerline that was restrained by the No. 3-bar hoop (Figure 3.29).Large flexural cracks were also present near both ends of the splice region (Figure 3.30).Figure 3.29 -Beam 4, centerline showing separation of concrete, 52-kip end load.REDACTED VERSION bbJL-UUU-UI4~(..$tLIL-UUUII-UUU I-'age /4>' 0? 1114Figure 3.30 -Beam 4, end of splice region at 52-kip end load.3.3.3 Beam 5 (cold joint, cycled)3.3.3.1 Beami 5 load-deflection curveBeams 5 and 6 were cast in the same manner and at the same time as Beam 4, with a coldjoint at the plane of reinforcing steel. Instead of monotonically loading the beams to failure,Beam 5 was first loaded to a total load of 80 kips, and subsequently unloaded to zero, and thenre-loaded to failure (the load protocol is presented in Table 2.5). When the beam was first loadedto a total load of 80 kips, the average end load was increased in increments of approximately 5kips. The specimen was inspected for cracks and marked at each load step. Horizontal cracks onthe plane of the cold joint within the splice region were observed when the beam was subjected to a total load of 80 kips. The maximum horizontal crack width at this load was 35 mils (0.03 5in.). It should be noted that the beam was unloaded in a rapid manner and that one of the loadcells had large fluctuations after that point (load cell C in Figure 3.31). Although there were clearproblems with the load readings from load cell C for the remainder of this test, the rams were atall times subjected to uniform pressure, and load readings from the other 5 beam tests show thatthe load was evenly applied to the four different load rods at all times. Furthermore, the loadREDACTED VERSION

t5-UUU t-'age f4Jt 0" 1 114beam remained level and the displacement readings were similar at both ends of the beam, strongindicators that although the load cell readings were not accurate, the load was uniformly appliedto the four load rods. Based on these observations, the total load was calculated based on thereadings from load cells A and B. When the beam was loaded for the second time, it was loadedup to a total load of 80 kips at an increment of 5 kips (average end load). At the end of the eachincrement, dial-gage displacement measurements were recorded. The beam was inspected forcracks at total loads of 40, 60, 70, and 80 kips. When the beam was inspected for crack duringthe second loading, some of the horizontal cracks elongated or widened and some new horizontal cracks were noticed. The maximum horizontal crack width was still 35 mils (0.035 in.)35 __-_ __- _-_ Load Cell A(west)30 Load Cell B(west)___ __ __25-___- Load Cell C (east)20_ ___ Load Cell D (east)0 1523 45-Total Deflection, in.Figure 3.22 -Load cell readings for Beam 5The load-deflection curve for Beam 5 is shown in Figure 3.32. Due to the problemdocumented for load cell C, the total load is calculated as twice the summation of load cells Aand B, located at the West loading point. Overall, Beam 5 performed very similar to Beam 4.The slope of the load-deflection curve first decreased at a total load of 20 kips, which coincides with the first observation of flexural cracks. Another decrease in the slope of the load-deflection curve was observed at a total load 91 kips, with a corresponding total displacement ofapproximately 2.7 in, which is attributed to the yielding of the flexural reinforcement. Thecalculated bar stress corresponding to the total load of 91 kips is 66 ksi based on moment-curvature analysis. The positive slope of the load-deflection relationship after a total load of 91kips is attributed to the strain hardening of the reinforcing steel. The beam was loaded to a totalREDACTED VERSION

I-'age f4401" 1114load of 96 kips, with a corresponding total displacement of 3.6 in., at which point the beam failedsuddenly. Wide flexural cracks near the support and horizontal cracks in the plane of cold jointwere observed within the splice region (Figure 3.33). It is concluded that the reinforcing steelyielded at a total load of 91 kips and beam failed at a total load of 96 kips due to failure of thesplice, the latter corresponding to a bar stress of 67 ksi (Table 3.1).110 -_ _ _ _ _ _100 -- 91kips90i60S50 --_ _S40 -S3020-- ---__ _ _ _10 --- * -- _ -_0 ,r0 2 4 6 8 10Total Deflection, in.Figure 3.32 -Total load vs. total deflection for Beam 5 (with a cold joint)Flexural cracks near the supportFigure 3.33 -Beam 5 (with a cold joint) at failureREDACTED VERSION

I-sage 4tb Ot 111l43.3.3.2 Crack progression-Beam 5Maximum measured crack width versus load for Beam 5 is shown in Figure 3.34; thecrack map for Beam 5 is presented in Figure 3.35. The first flexural and horizontal cracks formedat the supports at an average end load of 10 kips (total load of 20 kips). At an average end load of15 kips, flexural and horizontal cracks were present at both ends of the splice region and bothsupports (Figure 3.36). At an average end load of 40 kips, a 45-mil width flexural crack and 35-miu width horizontal crack were recorded. At this point, the beam was unloaded. The load wasreapplied, and at the last load prior to failure at which cracks were marked (average end load of40 kips), the maximum width of the cracks had not increased from first loading (Figure 3.34).Although the crack width was approximately the same, several cracks had increased in length.60oE 50S40~30E_ 200 10 20 30 40 50Average End Load, KipsSFlexural Crack (1st loading) -s Horizontal Crack (1st Loading)--Flexural Crack (2nd loading) -4.0- Horizontal Crack (2nd Loading)Figure 3.34 -Maximum crack width vs. average end load for Beam 5.REDACTED VERSION

t-'age Mb1 oT 1 114awINotth Face,South F,,ce -~717 I ~2 7~4,w$.% z ',14I~ETop FaceWLodnI$~kce PedestalRe~on SupportEodsPonPedestai 5p1ieCenIerFigure 3.35-Crack map for Beam 5. Numbers indicate maximum average end load whencracks marked. See Figure C.5 in Appendix C for greater detail.Figure 3.36 -Beam 5, northeast splice region with horizontal crack, 15 kip end load.Failure occurred at an average end load of 48 kips (total load of 96 kips), slightly lowerthan the failure load for Beam 4 (average end load of 52 kips, total load of 105 kips), which wasREDACTED VERSION

r'age (41 0l11114subjected to monotonically-increasing load up to failure. At failure of Beam 5, the concreteabove the cold joint separated from the remainder of the beam, with the horizontal crackpropagating along the cold joint throughout a region that was somewhat longer than the spliceregion. A small region near the centerline was restrained by the No. 3-bar hoop (Figure 3.37) andhad a tighter horizontal crack and a failure surface that passed through the top of the beam in thevicinity of the hoop, as shown in Figure 3.35. As with the other beams, large flexural crackswere also present near both ends of the splice region (Figure 3.38).Figure 3.37 -Beam 5, centerline showing separation of concrete, 48-kip end load.REDACTED VERSION bbi-UUU-UdtJt-GLUb-UUUIb-UUU I-'age I 4d ot 1 114Figure 3.38 -Beam 5, splice region, 48-kip end load.3.3.4 Beam 6 (cold joint, cycled)3.3.4.1 Beam 6 load-deflection curveThe configuration and loading protocol of Beam 6 were similar to those of Beam 5. Thebeams were cast using the same procedures and at the same time and were tested in the samemanner, except that unloading was much slower for Beam 6 and the beam was inspected forcracks more often during the second loading. The testing protocol for Beam 6 is presented inTable 2.5.The individual load cell readings are plotted versus total deflection in Figure 3.39. Asshown in Figure 3.39, the readings for the four load cells were identical, which verifies theassumption in Section 3.3.3 that the load was evenly distributed on the four load rods.REDACTED VERSION

I-age 141:: 0?" 1 114( 3 ) -Load Cell A (west)~25 ---Load Cell B (west)~20 -Load Cell C (east)105 -..-Load Cell (east)024 6 8 10Total Deflection, in.Figure 3.39 -Individual load cell readings (Beam 6)The total load versus total deflection for Beam 6 is plotted in Figure 3.40. Overall, Beam6 performed very similar to Beam 5. Yielding of the flexural reinforcement was observed at atotal load of 92 kips and a total displacement of 2.7 in., compared with 91 kips and 2.7 in. forBeam 5. The maximum horizontal crack width at the unloading point was 30 mils (0.03 in.),compared with 35 mils (0.035 in.) for Beam 5. Beam 6 also failed due to splice failure (Figure3.41) at a total load of 100 kips, corresponding to a bar stress of 69 ksi, and a total deflection of4.7 in. (versus 96 kips and 3.6 in. for Beam 5).110100 92 kips... .90~'60~50~40~3020100 ,ii.024 6 8 10Total Deflection, in.Figure 3.40 -Total load vs. total deflection for Beam 6 (with a cold joint)REDACTED VERSION bbgt-UUU-~bi.5t-ULUII-UUUllh-UUU r-age (bU 0t1 1114Flxrlcracksnathsupr Figure 3.41 -Beam 6 (with a cold joint) at failure3.3.4.2 Crack progression-Beam 6Maximum measured crack width versus load for Beam 6 is shown in Figure 3.42; thecrack map for Beam 6 is presented in Figure 3.43. The first flexural cracks formed at the eastsplice region and support at an average end load of 10 kips. At an average end load of 25 kips,flexural and horizontal cracks were present at both ends of the splice region and both supports(Figure 3.44). At an average end load of 40 kips, a 35-mil (0.035 in.) wide flexural crack and 30-mul (0.03 in.) wide horizontal crack were recorded. At this point, the beam was unloaded. Theload was reapplied, and at the last load prior to failure at which cracks were marked (average endload of 40 kips), the crack width had not increased with respect to first loading (Figure 3.42).Although the maximum crack widths remained the same, several cracks had increased in length.REDACTED VERSION

I-'age (/,l 01111460E~ 50"o 40~30U.20._ 10n0 10 20 30 40 50Average End Load, Kips Crack (1st loading) Crack (1st Loading)-*Ftexural Crack (2nd loading) --e--Horizontal Crack (2nd Loading)Figure 3.42 -Maximum crack width vs. average end load for Beam 6.North FaceS >.. --7... .... --. -.......... South Facew .LTop Faceloading Spike Ceiner Splice Pedestal LoadingPontn Suppoet Un. Relf Support PolngFigure 3.43 -CTrack map for Beam 6. Numbers indicate maximum average end load whencracks marked. See Figure C7.6 in Appendix C7 for greater detail.REDACTED VERSION

I-'age (b2 Ot l1114Figure 3.44 -Beam 6, splice region with horizontal crack, 25-kip end load.Failure occurred at an average end load of 50 kips, slightly lower than for Beam 4(average end load of 52 kips, total load of 105 kips), and higher than Beam 5 (average end loadof 48 kips, total load of 96 kips). As observed in Beams 2 through 5, at failure occurred at thecold joint with the upper concrete separating from the remainder of the beam, with the horizontal crack propagating along the cold joint between the pedestal supports. As for Beam 5, a smallregion near the centerline was restrained by the No. 3-bar hoop (Figure 3.45) and had a tighterhorizontal crack and a failure surface that passed through the top of the beam in the vicinity ofthe hoop, as shown in Figure 3.49. As in the case of the other beams, large flexural cracks werealso present near both ends of the splice region (Figure 3.46).REDACTED VERSION bb~tUUU-UdiL-U I:U-UUU1 ti-UUU I-'age fbLJ OT 1 114Figure 3.45 -Beam 6, centerline showing separation of concrete, 50 kip end load.Figure 3.46 -Beam 6, splice region, 50-kip end load.REDACTED VERSION bbi-UUU-3-iSIr-.iSU-UUUIb-UUU r-age fb40ot 1 11,44 Summary and Conclusions The effect of preexisting cracks, oriented in the plane of and parallel to the reinforcing steel, on the strength of No. 11-bar lap splices was investigated by testing six beams -three witha splice length of 79 in. and three with a splice length of 120 in. One of the beams with a 79-in.splice was cast monolithically and loaded monotonically to failure. To simulate the cracks, theother five specimens were cast with a cold joint at the mid-height of the reinforcing steel. Twobeams (one with a 79-in, splice and one with a 120-in, splice) were loaded monotonically tofailure. The other three beams were pre-loaded to develop horizontal cracks in the face of thecold joint, unloaded and then loaded to failure; those beams developed horizontal cracks withwidths of 20, 30 and 35 mils (0.02, 0.03, 0.03 5 in.) just prior to unloading. The test results aresummarized below:1. For the beam with a splice length of 79 in. and cast with monolithic

concrete, thereinforcing steel yielded and the beam failed in flexure.2. For the beam with a splice length of 79 in., cast with a cold joint, and subjected tomonotonically-increasing load to failure, splice failure took place at a bar stress of 62 ksi,about 8% below the bar yield strength of 67 ksi.3. For the beam with a splice length of 79 in., cast with a cold joint and subjected to cyclicloading, horizontal cracks with a maximum width of 20 mils (0.02 in) developed prior tofailure.

Splice failure took place prior at a bar stress of 57 ksi, about 15% below the.baryield strength.

4. For the beam with a splice length of 120 in., cast with a cold joint, and subjected tomonotonically-increasing load, the reinforcing steel yielded prior to a splice failure,which occurred in the strain-hardening region of the stress-strain curve at a bar stress of72 ksi.5. For the two beams with a splice length of 120 in., cast with a cold joint, and subjected tocyclic loading, horizontal cracks with maximum widths of 30 and 35 mils (0.03 and0.03 5 in.) developed prior to splice failure, which occurred at bar stresses of 67 and 69ksi, respectively, values that equaled or exceeded the bar yield strength..

The following conclusions are based on the test results and analyses presented in this report.REDACTED VERSION

b-UUU I-age fo Ot 1 1141. The methods described in this report provide a viable means of simulating a crack in theplane of flexural reinforcement.

2. The cyclically load beams incorporating a cold joint to simulate crack in the plane of thereinforcement exhibited slightly reduced lap splice capacity compared to themonotonically loaded beams.3. In the presence of a simulated crack in the plane of the reinforcing bars, the lap-spliced No. 11 bars with a 79-in, splice length achieved bar stresses of 62 and 57 ksi.4. In the presence of a simulated crack in the plane of the reinforcing bars, the lap-spliced No. 11 bars with a 120-in, splice length achieved bar stresses greater than or equal to theyield strength, 67 ksi.REDACTED VERSION

I-age t/b 0? 1114.

References:

ACI Committee 318. (2011). Building Code Requirements for Structural Concrete andCommentary (ACT 318-11), American Concrete Institute, Farmington Hills, MI, 430 pp.ACT Committee 408. (2003). Bond and Development of Reinforcement, Bond and Development of Straight Reinforcing Bars in Tension (ACT 408R-03), American Concrete Institute, Farmington Hills, MI, 49 pp.ACT Committee 408. (2009). Guide for Lap Splice and Development Length of High Relative RibArea Reinforcing Bars in Tension and Commentary (ACT 408.3R-09), American ConcreteInstitute, Farmington Hills, MI, 12 pp.ASTM C3 1. (2010). Standard Practice for Making and Curing Concrete Test Specimens in theField (ASTM 3 1/C3 1 M- 10), ASTM International, West Conshohocken, PA, 6 pp.ASTM C78. (2010). Standard Test Method for Flexural Strength of Concrete (Using SimpleBeam with Third-Point Loading) (ASTM C78/C78M-10), ASTM International, WestConshohocken, PA, 4 pp.ASTM C138 (2012). Standard Test Method for Density (Unit Weight), Yield, and Air Content(Gravimetric) of Concrete (ASTM C 138/C 138M- 12), ASTM International, West Conshohocken, PA, 4 pp.ASTM C 143. (2010). Standard Test Method for Slump of Hydraulic-Cement Concrete (ASTMC 143/C 1 43M- 10Oa), ASTM International, West Conshohocken, PA, 4 pp.ASTM C 172. (2010). Standard Practice for Sampling Freshly Mixed Concrete (ASTMC 172/C 172M-10), ASTM International, West Conshohocken, PA, 3 pp.ASTM C496. (2011). Standard Test Method for Splitting Tensile Strength of Cylindrical Concrete Specimens (ASTM C496/C496M-1 1), ASTM International, West Conshohocken, PA,5 pp.ASTM C 1064. (2011). Standard Test Method for Temperature of Freshly Mixed Hydraulic-Cement Concrete (ASTM C 1064/C 1 064M- 11), ASTM International, West Conshohocken, PA, 3PP.Nawy, Edward G. (2003). Reinforced Concrete: A Fundamental

Approach, Fifth Edition,Pearson Education Inc., Upper Saddle River, N J, 821 pp.Nilson, A. H., Darwin, D., and Dolan, C. W., Design of Concrete Structures, 14th Ed., McGraw-Hill, New York, 2010, 795 pp.Seliem, H. M., Hosny, A., Rizkalla, S., Zia, P., Briggs, M., Miller, S., Darwin, D., Browning, J.,Glass, G. M., Hoyt, K., Donnelly, K., and Jirsa, J. 0. (2009). "Bond Characteristics of High-Strength ASTM A 1035 Steel Reinforcing Bars," ACI Structural

Journal, Vol. 106, No. 4, July-Aug., pp. 530-539.REDACTED VERSION bYWi-UUU-U5dJ-GW__L-UUUIb-UUU l-'age tb/ 0?" 1114Appendix A: Pilot Tests -Preliminary Study of the Effect of Simulated Cracks on LapSplice Strength of Reinforcing Bars using Beams with Single Splices1. Introduction This appendix presents the findings of a pilot study consisting of two lap splice beamtests that were performed to investigate how a test specimen with a preexisting crack parallel tothe plane of the reinforcement could be developed and tested. The test program described in thebody of this report was developed using lessons obtained in this pilot study.The two beams were cast simultaneously and tested monotonically to failure seven daysafter casting.

Because this project involves a larger number of physical simulations, the testing ofthese two beams is referenced throughout the report as Stage 1 of the project. Both beams hadmain flexural reinforcement consisting of three No. 11 bars, two of them continuous and one ofthem spliced at the center of the beam (Figure A. 1). The splice lengths in the two beams were 33in. and 79 in., respectively. All other dimensions and material properties were identical. Thebeam with a splice length of 33 in. will be referenced throughout this appendix as Beam Al1 andthe beam with a splice length of 79 in. will be referenced as Beam A2.The two beams were instrumented with strain gages placed on all bars at the edge of thesplice region (Figure A.2). Beam displacements and applied loads were monitored during thetests using displacement transducers and load cells.The following sections present brief descriptions of the beams, the test process, andoutline the major findings from the tests.2. Beam CastingCasting was performed in two separate stages. The first stage of the casting processconsisted of placing concrete over the full depth of the beam at the end sections and up to themid-height of the flexural reinforcement in the splice region (Figs. A. 1 and A.3). The firstconcrete was placed on May 3, 2012. The concrete surface at the location of the cold joint wasroughened and the beam was wet-cured for 24 hours (Figure A.4). Two layers of painters tapewere placed adjacent to the bars to simulate the effects of a preexisting crack parallel to the planeof the flexural reinforcement (Figure A.5). Concrete was placed above the cold joint on May 4,and the beams were subsequently moist cured for 7 days.REDACTED VERSION bbig-UUU-Ud.5J-tb..d-UUUlth-UUU I--'age /tbt Oi 11 1476" 3321I 2" 66"I12"18" 252" 18"LoadSupport Support(a)76" 79"18" 252" 18"Load Q Load5" 5"'5" 5" 5"5"5" 5"5" 5"5"5" I" " , " " ' " " _,Z .... ...132".12n12"SupportSupport(b)Figure A.1 -Reinforcing steel drawing. (a) Beam Al -specimen with a 33-in, splice length. (b)Beam A2 -specimen with a 79-in, splice lengthREDACTED VERSION bbYJ-UUU-UdJ3-LiLH-UUU1b-UUU I-'age 0t 1114Figure A.2 -Strain gages placed on bars at the edge of the splice region of Beam A2.REDACTED VERSION

r~age ftiu 0? 1114Figure A.3 -Beam A2 after first concrete placement was completed. REDACTED VERSION

I--'age (bi OT 1 114Figure A.4 -Beam Al1 being moist cured after the first placement was completed. REDACTED VERSION bbJL-UUU-(LbJ-LWFt-UUUltb-UUU I-sage (tb2 Otf 1 114Roughed concrete surfaceto provide betterconnection at the cold jointSide stirrup to simulateconfinement provided bysurrounding concrete in alarger structure Figure A.5 -Painters tape placed to simulate a preexisting crack at the plane of thereinforcement in Beam A2.REDACTED VERSION

t-UUU I-sage 0111143. Test apparatus and loading protocolThe two beams were tested using a four-point loading configuration. To facilitate inspection of the splice region during the test, the loads were applied in the downward direction (Figure A.6) so that the main flexural reinforcement would be located at the top of the beam. Thesplice region was located between the two supports (Figure A.7), in the central constant momentregion of the beam.In addition to strain gages, the beams were instrumented to measure displacement andload. The four load rods used in the test were instrumented to record load, and displacements were recorded using displacement transducers and dial gages for redundancy. Threedisplacement transducers were used to monitor the displacement at the center of the beam and ateach of the two load points (Figure A.6). Dial gages were mounted at a distance of 3 in. from theload points.Loads were applied with four hydraulic rams connected to a manual pump through adistribution system with two separate manifolds. The manifold system allowed adjustments inthe pressure of each ram separately and adjustment of the pressure in each pair of rams allowingfor loading in tandem. The force in each of the four load rods (Figure A.6) was monitored throughout the test and the pressure in the rams was constantly adjusted to maintain the force ineach of the rods approximately equal.REDACTED VERSION

-UUU I-age or 1114Figure A.6 -Test apparatus Figure A.7(a) -Splice region of Beam A2 prior to loadingREDACTED VERSION

I-'age (tb~ Ot 1 114Figure A.7(b) -East support of Beam A2 prior to loadingREDACTED VERSION

I-'age (bb OT1 1114The loading protocol consisted of monotonically-increasing load applied at both the endsof the beams. Loading was paused at increments in the total force of 10 kips (5-kip increments applied at each end of the beam) to monitor crack widths, mark crack locations, and record dialgage readings (Figure A.8). After all these quantities were recorded, loading resumed until thenext increment was completed. Given the potential for brittle failure and the large amount ofenergy stored in the beam, crack location, crack width, and dial gage readings were not recordedafter the total load exceeded 140 kips (forces at beam ends exceeded 70 kips). After this point,the load was increased steadily until the end of the test. Measurements from load anddisplacement sensors were recorded without interruption during the test.Figure A.8 -Marking cracks during test4. Material Properties The beams were tested on May 10, 2012, seven days after initial casting. On the day ofthe test the compressive strength of the concrete was 5090 psi in the body of the beam and 5150psi above the cold joint.A segment of the No. 11 bars used in the beams was tested in tension. The stress-strain curve for the No. 11 bar is shown in Figure A.9. To avoid damage, the extensometer wasREDACTED VERSION

i-'age tfb o071114removed at approximately 3% elongation; force was recorded until failure. As shown in thefigure, the No. 1 1 bar did not have a well-defined yield point. The yield stress calculated usingthe 0.2% offset method was 71 ksi, the proportional limit was approximately 67 ksi, and themeasured elastic modulus was 27,666 ksi. The tensile strength of the steel was 108 ksi.1009002% offset th500)4020 Yield (0.2% offset): 70.6 ksi20 I tUltimate: 108.2 ksi10 Elastic Modulus: 27666 ksi100 0.005 0.01 0.015 0.02 0.025 0.03StrainFigure A.9 -Measured stress-strain curve for the No. 11 bar used in the beams5. Test ResultsThe load-deflection curves for Beams 1 (33-in, splice) and 2 (79-in, splice) are shown inFigures A. 10 and A. 11, respectively. The displacement shown in both figures was calculated byadding the average displacement at the two load points and the displacement at the center of thebeam. The load shown in Figures A. 10 and A. 11 corresponds to the total load applied to thebeam. Based on the shape of the load-deflection curves shown in Figures A. 10 and A. 11, it isconcluded that a splice failure took place in Beam Al and a flexural failure occurred in BeamA2.For Beam Al (33-in, splice length), the peak total load recorded was 140 kips, at acorresponding total displacement of 1.14 in. (Figure A. 10). At a total load of 140 kips, the stressin the bars calculated using elastic cracked section theory was approximately 54 ksi. After thedisplacement exceeded 1.14 in., the total load dropped in a sudden manner to approximately 133REDACTED VERSION bb.i-UUU-dJ-UI:-UUU re-'age 01 1i14kips. If it is assumed that the tension force is carried in its entirety by the two continuous bars, atotal force of 133 kips corresponds to a calculated bar stress equal to the yield stress of 71 ksi(based on linear elastic cracked section theory). These calculations indicate that splice failureoccurred at a displacement of 1.14 in. and that the splice lost all its load carrying capacity in asudden manner. The total load tended to increase again at displacements greater than 1.6 in.,which is attributed to the effects of strain hardening in the two continuous bars.The load-deflection curve for Beam A2, with a splice length of 79-in., is presented inFigure A. 11. Loading was stopped when crushing of the concrete in the compression zone wasobserved in the constant moment region, in the areas adjacent to the two beam supports, at a totaldisplacement of approximately 2.5 in. Unlike the curve for Beam A l, there was no sudden dropin load associated with failure of the splice. In the case of Beam A2, a sharp decrease in the slopeof the load-deflection curve was observed at a total load of approximately 172 kips and totaldisplacement of approximately 1.4 in. The stress in the three bars calculated based on moment-curvature analysis at this load is approximately 67 ksi (Table A. 1), which corresponds to theobserved proportional limit of the measured stress-strain relationship of the steel (Figure A.9).The calculated steel stress indicates that the sharp decrease in the slope of the load-deflection curve at 172 kips was caused by yielding of the reinforcing steel, not by failure of the splice.After yielding began, the total load continued to increase with increasing displacement, as thereinforcing steel strain hardened. The maximum load prior to flexural failure was approximately 186 kips, which corresponds to a bar stress of 72 ksi in all three bars (Table A. 1). At a total loadof 186 kips, horizontal splitting cracks on the beam top surface were observed (described in moredetail below).After the tests we completed, the beams were autopsied to determine the actual cover onthe bars. For Beam Al, the top cover was 4 in., and side covers to the continuous bars were 3.5(North) and 3.75 in. (South). For Beam A2, the top cover was 4 in., and side covers to thecontinuous bars were 3.5 in. (North and South). (These values are reflected in the bar stresses inthe previous paragraph and summarized in Table A. 1)REDACTED VERSION

I-'age 1bSU OT 1 114160010100-J0 0.5 1 1.5 2 2.5 3 3.5 4Displacement, in.Figure A.10 -Total load vs. total deflection for Beam Al (33-in, splice length)20012kp _ 1 8 0.. .... ....140.-100 __06000.5 11.5 22.53Displacement, in.Figure A.11 -Total load vs. added deflection for Beam A2 (79-in, splice length)REDACTED VERSION

I.-sage /flU 0? 111"m4Loads, moments, and bar stresses for the beams were calculated assuming that loads andreactions acted along the longitudinal centerline of the beam. Reactions and moments werecalculated based on load cell readings and the weight of the loading assemblies. The self-weight of the beam was included in the calculations based on average beam dimensions and an assumedconcrete density of 150 pcf.The calculated moment, bar stress at splice failure, and calculated bar stress using thesplice strength equation developed by ACI Committee 408 (2003) are shown in Table A. 1. It isimportant to note that the splice strength expression developed by Committee 408 was calibrated on the basis of beams without preexisting cracks in the plane of the flexural reinforcement, andfor this reason are presented only as a reference. For Beam Al (with a 33-in, long splice), the barstress at splice failure calculated based on a moment-curvature analysis was approximately 54ksi. The calculated splice strength using the expression developed by ACI Committee 408 (ACI408R) was 70 ksi. For Beam A2 (with a 79-in, long splice), the calculated bar stress at flexuralfailure was approximately 72 ksi, while the calculated splice strength using the ACI 408expression was 140 ksi.Table A.1 -Bar stresses at splice failureI 33-in. I splice failure j 140 I 35I54I70I[79 in. flex ural 186 [ 472 72 [140]The strain in the No. 11 bars was measured using strain gages located 2 in. outside thesplice region (Figure A.2). The relationships between measured strain and total load are shown inFigures A. 12 and A. 13 for beams 1 and 2, respectively. As shown in Figure A. 12, the strain in the spliced bars (East-center and West-center gages) of Beam Al increased to a maximum of 1750 and 1700 microstrain, respectively, andthen dropped in a sudden manner. The maximum strain in the spliced bar was recorded at a totalload of approximately 130 kips and corresponds to a bar stress of approximately 50 ksi, which isvery close to the failure value of 54 ksi inferred on the basis of moment-curvature analysis(Table A. 1). Strain readings from the east-center gage on the spliced bar show that the strain1REDACTED VERSION I-'age 111 0? 1114dropped from approximately 1700 to approximately 1300 microstrain at a total load of 130 kips,corresponding to a sudden reduction in capacity of approximately 25%. When the total loadreached 140 kips, the strain in the east-center gage dropped suddenly to almost zero. Strainreadings from the east continuous bar (East-Side

1) show a sudden increase from 2100microstrain to more than 2500 microstrain at the failure total load of 140 kips. The strain gagereadings indicate that failure of the splice led to a rapid decrease in the stress in the spliced bars,and that the tension force that was lost due to failure of the splice was transferred to thecontinuous bars, causing yielding of the continuous bars at a total force of 140 kips.For Beam A2 (79-in, long splice),

the recorded strains show a plateau (Figure A. 13) dueto exceeding the limiting strain allowed by the gain in the data acquisition system.25002000* 1500.N10005000-East-center -East-Side I-East-side 2---- West-center 050 100 150Total load, kipsFigure A.12 -Measured strain in the reinforcing bars vs. total load for Beam Al (33-in, splicelength). (Note: The beam was oriented in an east-west direction; "center" identifies strain gageson the spliced bars and "side" means strain gauges on the continuous bars)REDACTED VERSION 'U~ Page (/j' or 1114L.0L..U100 150Total load, kips 200length). (Note: The beam Was oriented in an east-west drcin cne"i~t~on the spliced bars and "side" meaection; ,,...,, sstrain gages6. Bemcr c ...... , tuges on" the continuous bars)Figures A. 14 through A. 18 are photographs taken after the Conclusion of the two tests.For B earn Al (33-in. splice length), splitting cracks Were observed on the top surface betweenthe vertical edges of the cold joint (Figures A. 14 and A. 15). The cracks Were approximate 1y 1/4in. wide, as shown in Figure A. 16. Splitting cracks above the splice Were also noted in Beam A2(79-jn. splice length) (Figures A. 17 and A. 18), although they were much narrow~er than thoseThe crack patterns for both beams show that the side stirrups Were effective in keepingthe COver in place, even after failure of the splice for Beam Al. In the case of Beam Al, thecracks Were wider, which is COnsistent with the sudden drop in bar foc that ocre tslcfailure. For Beam A2, the cracks Were much narrower, and it is apparent that the splice was ableto sustain the same bar force as the continuous bars at displacement ag nuht asflexural failure of the beam .nt la g e ou h o c usREDACTED VERSION

t-UUU r-age i Ot o 1114Splitting crackabove spliceEnd of splice regionLFigure A.14 -Splitting crack at the top of the splice region for Beam Al (33-in, splice length).REDACTED VERSION bbI,-UUU-L

,,-U51"-UUUll-UUU I-'age (140OTll114 Horizontal cold joint in theplane of the reinforcement Figure A.15 -Crack pattern in the splice region for Beam AlI (33-in, splice length).REDACTED VERSION

1-'age ((b 0O" 1114Figure A.16 -Splitting crack at the top of the splice region of Beam Al (33 in. splice length).REDACTED VERSION

I-sage 0O" 1114End o cFigure A.17 -Splitting crack at the top of the splice region of Beam A2 (79 in. splice length).REDACTED VERSION

F-'age f[f 0?' 1114Splitting crackabove spliceHorizontal cold jointin the plane ofreinforcing steelFigure A.18 -Crack pattern in the splice region of Beam A2 (79-in, splice length)REDACTED VERSION ~bYJ-UUU-UtS3-ULU-UUU1b-UUU Page bbd 0? 1114Effect of Laminar Cracks on Splice Capacity of No. 11 Bars based on Testing Conducted at PurdueUniversity and University of Kansas for Davis-Besse Shield BuildingAppendix BUniversity of Kansas Test ReportUOflJi32flU~I ~J 2012 Becfltel Uorporatcon. LontBeebteland its affihatsd comoankz which shall~in: confidential and/or proprietar,' information tonotbz ussd, disclosed, or reproduced in anyformatby anvnnn Rc:htcl nnt..'withoutEechtelf: oricrwrttten oerm[zian. All ridts rcscr..'rd. --iT ............. -lr..... 0 ....B-iREDACTED VERSION

I-'age OT 1114EFFECT OF SIMULATED CRACKS ON LAP SPLICESTRENGTH OF REINFORCING BARSByJiqiu YuanMatthew O'ReillyAdolfo Matamoros David DarwinResearch supported byFIRSTENERGY NUCLEAR OPERATING COMPANY(FENOC)Structural Engineering and Engineering Materials SL Report 12-2THE UNIVERSITY OF KANSAS CENTER FOR RESEARCH, INC.LAWRENCE, KANSASJune 2012REDACTED VERSION

Mage ti~u atl 1114AbstractThe effect of preexisting cracks, oriented in the plane of and parallel to the reinforcing steel, on the strength of No. 11-bar lap splices was investigated by testing six beams -three witha splice length of 79 in. and three with a splice length of 120 in. One of the beams with a 79-in.splice was cast monolithically and loaded monotonically to failure. To simulate the cracks, theother five beams were cast with a cold joint at the mid-height of the reinforcing steel. Two beams(one with a 79-in, splice and one with a 120-in, splice) with a cold joint were loadedmonotonically to failure. The other three beams were preloaded to develop horizontal cracks inthe face of the cold joint, unloaded and then loaded to failure; those beams developed horizontal cracks with widths of 20, 30 and 35 mils (0.02, 0.03, 0.035 in.) during the first cycle of loadingand just prior to unloading. The nominal concrete compressive strength was 5000 psi.The methods described in this report provide a viable means of simulating a crack in theplane of flexural reinforcement. In the presence of a simulated crack in the plane of thereinforcing bars, the two specimens with lap-spliced No. 11 bars with a 79-in, splice lengthachieved bar stresses of 62 and 57 ksi. In the presence of a simulated crack in the plane of thereinforcing bars, the three specimens with lap-spliced No. 11 bars with a 120-in, splice lengthachieved bar stresses of 72, 67, and 69 ksi.REDACTED VERSION bbIZ-UUU-Lidi-USI-U-UUUIh-UUU I-age btJE1 01" 1114Personnel The research described in this report was performed under the direction of David Darwin,Ph.D., P.E., Deane E. Ackers Distinguished Professor of Civil, Environmental, and Architectural Engineering and Director of the Structural Engineering Materials Laboratory and AdolfoMatamoros, Ph.D., Associate Professor of Civil, Environmental, and Architectural Engineering at the University of Kansas. Darwin and Matamoros are joined by post-doctoral researchers, Matthew O'Reilly and Jiqiu Yuan.David Darwin has extensive experimental and analytical experience in the field of bondand development of reinforcement and has been involved in bond research for over 30 years. Heis a member and past chair ACI Committee 408 on Bond and Development of Reinforcement and a member of ACI Subcommittee 318-B Reinforcement and Development (Structural Concrete Building Code). Darwin developed the ACI Committee 408 expression fordevelopment and splice lengths, which accurately covers straight reinforcing bars with yieldstrengths between 40 and 120 ksi and for concrete with compressive strengths ranging from2,000 to 16,000 psi. Darwin also developed the ASTM A944 Beam-End Test, which is used toevaluate relative bond strength. In addition to Committee 408, Darwin is a member and pastchair of ACI on Committee 224 on Cracking, as well as four other ACI technical committees. Adolfo Matamoros has many years of experience in testing and analysis of reinforced concrete with special expertise in test instrumentation. He is the immediate past chair of ACICommittee 408 and a member of four other ACI technical committees. Matthew O'Reilly and Jiqiu Yuan completed their Ph.D. degrees at the University ofKansas in 2011 and have been serving as senior researchers on a number of experimental projects since receiving their degrees.REDACTED VERSION

P-age bSU2 0 1114ContentsAbstract.................................................................................................... iiiPersonnel .................................................................................................. iv1 Overview / Background............................................................................. 12 Research Program and Test Specimens............................................................ 12.1 Design of test specimens ....................................................................... 12.2 Concrete ......................................................................................... 32.3 Cold joint construction and crack simulation................................................. 42.4 Test methodology............................................................................... 92.4.1 Fabrication ................................................................................. 92.4.2 Casting .................................................................................... 102.4.3 Curing..................................................................................... 112.4.4 Test apparatus ............................................................................ 112.4.5 Loading protocol ......................................................................... 142.4.6 Calibration ................................................................................ 182.5 Test Facilities.................................................................................. 182.6 Section Analysis............................................................................... 193 Test Results ....................................... ................................................. 233.1 General......................................................................................... 23REDACTED VERSION

P~age bu;d oT 11143.2 Beams 1, 2, and 3 with 79-in. splice length................................................. 243.2.1 Concrete strength ........................................................................ 243.2.2 Beam 1 (monolithic concrete)........................................................... 263.2.3 Beam 2 (cold joint, monotonically-loaded)............................................ 323.2.4 Beam 3 (cold joint, cycled)..................... ......................................... 373.3 Beams 4, 5, and 6 with 120-in, splice length ............................................... 413.3.1l Concrete strength ........................................................................ 413.3.2 Beam 4 (cold joint, monotonically-loaded)............................................ 423.3.3 Beam 5 (cold joint, cycled) ............................................................. 4"73.3.4 Beam 6 (cold joint, cycled).............................................................. 534 Summary and Conclusions........................................................................ 59

References:

............................................................................................... 61Appendix A: Pilot Tests -Preliminary Study of the Effect of Simulated Cracks on Lap SpliceStrength of Reinforcing Bars using Beams with Single Splices ..................................... 62Appendix B: Reinforcing Steel Drawings............................................................. 83Appendix C: Detailed crack maps of Beams 1- 6.................................................... 87Appendix D: Reinforcing steel mill certification and deformation measurements............... 106Appendix E: Concrete Mixture Proportions ......................................................... 110Appendix F: Data recording forms.................................................................... 113REDACTED VERSION bbui;-UUU-Ud3-ULhU-UUU1br-UUU H-age boW40 at1114Appendix G: Load cell and displacement transducer calibration .................................. 119Appendix H: Training forms, Trip tickets, concrete properties, specimen dimensions, and crackrecording during beam tests ........................................................................... 131Appendix I: Certificates of calibration for laboratory apparatus ................................... 198REDACTED VERSION 25593-000-G83-GEG-00016-000 Page 695 of 1114This page is intentionally blankREDACTED VERSION

b-U~u I-'age 15ti Ot 11141 Overview / Background Past research on the strength of lapped bar splices in reinforced concrete has focused oninvestigating the performance of various lap splice configurations in monolithic members. Theresearch program described in this report investigates the effect of preexisting cracks, oriented inthe plane of and parallel to the reinforcing steel, on the strength of lapped bar splices. Theresearch program was conducted in two phases, a pilot study investigating various methods tosimulate the preexisting cracks that is described in Appendix A of this report, and a series ofbeam tests described in the main body of the report.Beams in the main study had cold joints in the splice region, along the plane of thereinforcement, to facilitate the initiation of a crack prior to failure. Two No. 3-bar hoops (one oneach side) crossing the plane of the cold joint, in the center of the specimen and on the exterior ofthe lap splices, were used to simulate the effects of the continuity of concrete in an actualstructure. The beams contained two spliced No. 11 bars with 79 or 120-in, long lap splices. Someof the beams were loaded until horizontal cracks had developed along the plane of the cold jointwith a minimum width of 10 mils (0.01 in.); they were then unloaded and subsequently reloadedto failure. The remainder of the beams were loaded monotonically to failure.2 Research Program and Test Specimens 2.1 Design of test specimens A total of six beam-splice specimens were tested in the main study -three specimens with a splice length of 79 in. and three with a splice length of 120 in. For the three specimens with a 79-in, splice length, one was cast with monolithic concrete and the other two were castwith a cold joint in the plane of reinforcing steel. All three specimens with a 120 in. splice lengthwere cast with a cold joint in the plane of reinforcing steel. All specimens With cold joints hadtwo No. 3-bar hoops crossing the plane of the cold joint, outside the spliced bars, at the center ofthe specimen. The beams were subjected to four-point loading to provide a constant moment (excluding deadload) in the middle portion of the member, where the splice was located, as shown in Figure 2.1.REDACTED VERSION bb-LUUU-U-(.5iLU-,1(-UUU I-age Ot 1114The specimens were configured to have a constant moment in the splice region toeliminate the effect of shear forces on splice strength, and also to eliminate the need for shearreinforcement within the splice region. The spacing of the supports was chosen so that thedistance from either end of the splice to the central pin and roller supports was equal to or greaterthan the effective depth of the beam. The span lengths were selected in increments of 3 ft basedon the spacing of load points in the Structural Testing Laboratory of University of Kansas.ShearMomentFigure 2.1 -Configuration and shear and moment diagrams for the testing fixtureThe reinforcement diagrams for the specimens in the study are shown in Appendix B. Thetop reinforcement layer of the beams consisted of two No. 11 reinforcing bars, which were spliced atthe center of the beam, as shown in Figure 2.1. The No. 11 bars used in the specimens were from asingle heat of reinforcement. The bottom layer of reinforcement, placed to maintain the integrity ofthe beam after failure of the splice and to facilitate placement of shear reinforcement in the constantshear regions, consisted of two Grade 60 No. 3 bars. Beam dimensions and effective depths aresummarized in Table 2.1.REDACTED VERSION DbL-UUU-UdiL-(.5UL-UUUI b-UUU r-age or11114The specimens were proportioned to have two splices, each with a nominal side concretecover of 3 in. to the outermost No. 11 bars and a top concrete cover of 3 in. Grade 60 No. 5closed hoops spaced at 5 in. on center were placed in the constant shear region (Figure 2.1) of allsix beams. Mill certifications for the No. 11, No. 5 and No. 3 bars are reported in Appendix D.B179(monolithic) 1125182420.32.8B2 79 Cold joint 11 25 18 24 20.3 2.8B3 79 Cold joint 11 25 18 24 20.3 2.8B4 120 Cold joint 14 28 18 24 20.3 2.8B5 120 Cold joint 14 28 18 24 20.3 2.8B6 120 Cold joint 14 28 18 24 20.3 2.8The deformation properties of the No. 11 bars are summarized in Table 2.2. The meandeformation height and spacing of the No. 11 bars meet the requirements of ASTM A6 15 and therelative rib area, 0.0685, is within the typical range for conventional reinforcement in U.S.practice (0.060 and 0.085) (ACI 408R-03). INo. II V .UOI I I U.IJIL U 1 .*Per ASTM A 615 **Per ACI 408R-03 and ACI 408.3R-09 for calculation of relative rib area2.2 ConcreteThe concrete used to fabricate the test specimens was supplied by a local ready mix plant.The concrete was non-air-entrained with Type I portland cement, 1 1/2/-in. nominal maximum-size crushed coarse aggregate, and a water-cement ratio of 0.42. A trial batch was prepared at the concretelaboratory of the University of Kansas prior to casting the first three beams. The aggregate gradation, REDACTED VERSION

b-UUU r-age 0? 1114mixture proportions, and concrete properties for the trial batch and each of the placements arepresented in Appendix E. The dosage of high-range water reducer was adjusted on site whenconsidered necessary to obtain adequate slump for placement. 2.3 Cold joint construction and crack simulation The specimens with cold joints were cast using two placements, with a cold joint at themid-height of the top layer of reinforcement, to ensure that a longitudinal crack would develop inthe plane of the reinforcing steel before the beam failed. The cold joints spanned the entire lengthof the spliced region and extended approximately 6.5 ft outside of the spliced region (Figures B.2and B.3 in Appendix B).In the first placement, concrete was cast up to the center of the top layer of reinforcement (Figures 2.2, B.2 and B.3). After the concrete was placed, a roughened surface was created tosimulate the roughness of a natural crack by introducing indentations in the concrete while itremained plastic (Figure 2.3). The exposed reinforcing steel was cleaned using sponges tofacilitate adequate bond between the exposed bars and the concrete cast during the secondcasting stage. The specimens were moist cured for a day, and the remainder of the concrete wasplaced no later than 26 hours after the original placement. The concrete for the second placement had the same mixture proportions and was supplied by the same ready-mix plant as the first.Before the second placement, the concrete surface was cleaned using compressed air to removedebris and loose concrete, and maintained in a wet condition until the second placement started(Figures 2.4 and 2.5). After casting, the specimens were moist-cured until the compressive strength of the concrete from the first placement exceeded 3500 psi.Some beams were loaded in two stages to ensure that the preexisting crack of minimumwidth had formed in the plane of the reinforcing steel. To do this, beams were loadedmonotonically until the width of the horizontal cracks at the cold joint exceeded 10 mils (0.01in.). After initial loading, the specimens were unloaded and subsequently reloaded monotonically to failure.REDACTED VERSION

l-'age fUU oft 1114Concrete cast monolithically at the ends of the beamspecimen. Concrete cast to the level ofreinforcing steel in the middleportion of the beam specimens. Figure 2.2 -rfirst stage of casting was completed. REDACTED VERSION

I-'age fU1 Ot 1 114(a) (b)Figure 2.3 -Roughening of the concrete surface at the cold joint. (a) roughening of the concretesurface while the concrete remains plastic. (b) roughened surface after concrete had set.Figur 2.4RemoAlCfTooED cocEteRSIONgopesdar REDACTED VERSION bbL-UUU-UbJ-GW:I-UUUIIh-UUU I-'age /UZ 01" 1114Figure 2.5 -Wetting of concrete surface prior to concrete placement. The flexural strength of the concrete (a measure of its tensile strength) was measured inaccordance with ASTM C78. For each set of beams, two specimens were cast monolithically with concrete from below the cold joint and two were cast with a cold joint at midspan in theflexure specimen using concrete from both below and above the cold joint in the beam. ForBeams 4, 5, and 6, two additional flexure beams were cast monolithically using the concretefrom the second placement (above the cold joint). The specimens with the cold joint were cast sothat half of the total length was filled with concrete from below the cold joint in the splicespecimens; the concrete surface was then roughed (Figure 2.6) following the same procedure used for the beam-splice specimens (Figure 2.3 to Figure 2.5). The second half of the "coldjoint" flexure specimens was cast using concrete from above the cold joint in the splicespecimens. A schematic of the flexure test is shown in Figure 2.6(c). The test results aresummarized in Chapter 3 and indicate that the cold joint had significantly lower tensile strengththan monolithically-cast

concrete, and thus provided a good representation of a preexisting crack.REDACTED VERSION bwb-u-UUU-w.tuJ-uuuiw.-UUUu roUU -ags /us 0?" 1114(a)(b)/ L -' .L ;-4.-- £-Bed Spe Lenth L '-"(c)Figure 2.6 -Flexure beam specimens with cold joint. (a) A flexure beam specimen cast to halfof its length. (b) Roughed concrete surface.

(c) Schematic of flexure test (ASTM C78).REDACTED VERSION bb~-UUU-U~iJ-ULUI--UUU1t)-UUU I-'age fU4 0? 1 1142.4 Test methodology

2.4.1 Fabrication

FormworkThe formwork was fabricated using plywood and dimension lumber with the tolerances specified in Table 2.3. The interior of the forms was coated with a sealant, taped at the seams toprevent leakage, and covered with a thin layer of oil before casting. The dimensions of theformwork were measured and are recorded in Table F. 1 (Appendix F).Table 2.3 -Form tolerances Reinforcement The steel reinforcement was fabricated to meet the dimensions specified in the drawingsshown in Appendix B. In the splice test region, the bar spacing, concrete cover, location ofsimulated cracks, and splice length satisfied the tolerances specified in Table 2.4. Outside of thesplice test region, the bar spacing, concrete cover, and longitudinal bar location satisfied theintended tolerance of+/- 1/2 in. Inside the forms, the reinforcing steel was supported by chairs tiedto the bottom of the hoops outside of the test region (splice region) and to the bottom layer oflongitudinal reinforcing steel in the splice region. Spliced bars were supported by small-diameter threaded rods attached to both sides of the forms. The threaded rods were introduced with theobjective of achieving the specified tolerance in the cover dimensions and preventing bowing ofthe forms at the top of the beams. Cover and reinforcement dimensions in the test region weremeasured and are recorded on Table F.2 (Appendix F).REDACTED VERSION

I-'age (Ub2 Of 1 114i oierance +/-1l~/2 +/-J/6 +/-1]/22.4.2 CastingThe properties of the plastic concrete were measured in accordance with the ASTMstandards cited and are presented in Table F.3. The following properties were recorded: -Unit Weight (ASTM C 138)-Slump (ASTM C 143)-Concrete Temperature (ASTM C 1064)The concrete truck operator delivered a ticket with the batched mixture weights. Theticket was examined to verify that the mixture delivered had the specified proportions and thatthe concrete had arrived less than 45 min. after leaving the batching plant. No water was addedto the concrete after the truck left the plant.The beams were cast in two layers, beginning and ending at the ends of the beams. Thebottom and top layers of concrete in the splice regions of all three beams were placed from themiddle portion of the batch. The concrete was sampled at two points in the middle portion of thebatch in accordance with ASTM C 172, the first sample taken immediately after placing the firstlift, and the second sample taken immediately after placing the second lift in the splice regions.After placing the second lift, excess concrete was removed from the formwork using a screed.The upper surfaces of the specimens were finished using hand floats.The samples were consolidated prior to testing the plastic concrete (for slump, unitweight, and concrete temperature) and making the strength test specimens. Ten plastic and sixsteel 6 x 12-in, cylinder molds were filled in accordance with ASTM C3 1, along with fourflexural beam specimens cast in accordance with ASTM C78. Two of the flexural beamspecimens were cast monolithically and two were cast with cold joints. Cylinders cast in plasticREDACTED VERSION bbJ;-UUU-LSt-J'-GLU-UUU1 I-age I Uti 0? 1]14molds were used for monitoring the strength of the concrete prior to testing the beam; thecylinders cast in steel molds were used to obtain the compressive strength on the day of test ofeach beam. All flexural beam specimens were tested on the day of test of the corresponding beam. Test beams and cylinders were labeled with an identifying mark.For specimens with a cold joint, the concrete above the joint plane was placed no laterthan 26 hours after the initial placement. The concrete above the cold joint had the same mixproportions as the concrete below the cold joint and was supplied by the same ready-mix plant.The concrete slump, unit weight, and temperature were recorded. A minimum of five 6 x 12-in.cylinders (two in plastic molds and three in steel molds) were prepared. The two cylinders castusing plastic forms were tested on the day of form removal when the concrete below the coldjoint had achieved a compressive strength of 3,500 psi. The three cylinders cast using steel moldswere used to determine the concrete compressive strength on the day the beams were tested.2.4.3 CuringTest cylinders and flexure beam specimens were stored and cured next to the beam-splice specimens and under similar conditions of temperature and humidity. The beams were coveredwith wet burlap immediately after finishing of the beam surface. The beams, flexure beams, andthe 6 x 12-in, cylinder specimens were moist-cured by keeping them covered with wet burlapand plastic until the measured compressive strength of the concrete exceeded 3500 psi. Theplastic cylinder molds were sealed with plastic caps during the period in which the beams werewet cured.The beam formwork and the molds were removed after the 3500 psi threshold wasexceeded. After demolding and removal of the forms, the specimens were air-cured until themeasured compressive strength reached 5000 +500 psi.2.4.4 Test apparatus The beam-splice specimens were tested using a four-point loading configuration (Figure2.1 and Figure 2.7). To facilitate inspection of the splice region during the test, the loads wereapplied in the downward direction (Figure 2.7) so that the main flexural reinforcement would belocated at the top of the beam. The splice region was located between the two supports (FigureREDACTED VERSION

I-age IuI 0? 11142.7) in the central constant moment region of the beam. The final location of the supports wasmeasured (to the nearest '/8-in.) and is reported in Table F.4 (Appendix F). As-built externaldimensions of each test beam were recorded using the same form. The maximum deviation fromnominal dimensions in the test region was 1/2/ in.Figure 2.7 -Four-point loading configuration Loads were applied at the ends of the specimen using two loading frames, as shown inFigure 2.7. Each loading frame consisted of two load rods attached to a loading beam that wasplaced above the specimen. Loading was imposed through dual-acting center-hole hydraulic rams attached to the lower surface of the reaction floor. At the start of the test, the lower end ofthe load rods passed through the reaction floor without applying load to the specimen other thanthe weight of the loading frame and the rods. A total of four rams were used, two for eachloading frame. High-pressure hydraulic lines connected the rams to separate pressure and returnmanifolds, which were connected to the pressure and return lines of a single hydraulic pump. Allhoses and other hydraulic hardware were inspected visually before testing began.The beams were instrumented to measure displacement and load. As shown in Figure 2.8,the applied load was measured with load cells mounted on the load rods, and displacements wereREDACTED VERSION bbBJ-UuU-U-(S-.ji-ULU-UUU~t)-UUU t-'age (Ud 0? 1 114recorded using displacement transducers and dial gages (for redundancy) at the center of thebeam and at each of the two load points.Within each specimen, 350-ohm 1/4-in, strain gages with attached leads were bonded tothe spliced bars, approximately 2 in. outside the edges of the splice. One deformation in each barwas removed using low-heat grinding to provide a smooth surface to attach the strain gages.Strain gages were attached to the bars using epoxy cement and sealed following therecommended procedures by the manufacturer for submersion in concrete. The strain gages wereplaced so that the coating used to seal the strain gages covered only deformations outside of thesplice region. The strain gages provided little useful data.Figure 2.8 -Loading apparatus and instrumentation at each load pointREDACTED VERSION

5-UUU I-'age aT 1 1142.4.5 Loading protocolThe double acting rams were fully retracted prior to the start of each test. With theloading rams in the fully retracted

position, slack was taken out of the load rods by tightening thenuts until each load rod was nearly engaged with the fully retracted hydraulic jacks, but withoutapplying any load. This procedure was adopted to prevent rotation of the loading beams andconsequently maintain even loading across all four rods.Before load was applied, all displacement transducers, load cells, and strain gages werezeroed and initial readings were recorded for each of the three dial gages. Data were recordedcontinuously by the data acquisition system with a sampling rate of approximately one sampleper second. Recorded data was continuously appended to a data file to prevent any loss of data incase of system failure.Load was applied using a single manually-controlled hydraulic pump. Loading wasstopped at predetermined load levels to visually inspect the beam, mark visible cracks (identified based on the average value of the load applied at one end of the beam, as illustrated in Figure2.9), measure crack widths using crack comparators, and to record strain and dial gage readings.

The specimens were marked before each test to indicate the locations of the ends of the spliceregion, the beam centerline, the pin and roller (pedestal)

supports, and the load apparatus.

Themarkings, shown in Figure 2.10, were 'SR' to indicate the ends of the splice region, 'CL' for thecenterline of the beam, and 'PS' for the center of the pedestal support. All longitudinal measurements were taken using the centerline of the beam as a reference point to eliminate anyinconsistencies caused by small deviations from the nominal length in the specimens. REDACTED VERSION

s-UUU I-'age flU Ot 1 114Figure 2.9 -Crack inspection and marking during test(a)REDACTED VERSION bbYJ-UUU-UdJL-ULUI--UUU'Ib-UUU I-age t Ii O't 1i 14(b)Figure 2.10 -Beam marks: (a) End of splice region and centerline of the beam; (b) pedestalsupport centerline The initial load increment was chosen to be smaller than one half of the calculated flexural cracking load to ensure that all instruments and the hydraulic system were operating properly. From this point forward, loading proceeded in increments of approximately 5 kips ateach end of the beam. The final load step at which cracks were marked was approximately two-thirds of the estimated failure load. In some of the specimens, the loading protocol was such thatthe specimens were unloaded after the formation of a horizontal crack with a width of at least 10mils in the splice region. After the specimen was fully unloaded, it was loaded to failurefollowing the procedure specified above for monotonically-loaded specimens. The loadingprotocol used for each beam is presented in Table 2.5.A log was maintained to record any meaningful observations during the test, such as loadcorresponding to flexural

cracking, crack widths, file names, and gage readings.

The logs arepresented in Appendix H.After failure, cracks were marked on the specimens with each identified using thepreliminary value of the average maximum end load (this value typically deviated by a fewpercent from the recorded value).The following data were recorded and continuously transferred to disk throughout eachtest:REDACTED VERSION bbi-UUU-Ubi-UL::-UUU1 b-UUU P'age /1 o1 Grill4-Force applied to each load rod-Displacement at midspan and each load application point-Strain in the reinforcing steelTable 2.5 -Detailed loading protocol for each beamBeam Loading Protocol1 (1)Monotonically-increasing load up to an average end load of 40 kips in increments of 5 kips.At the end of the each increment, the beam was inspected for cracks and dial-gage displacement measurements were recorded. (2) Loading resumed with increasing displacement until failure.2 (1)Monotonically-increasing load up to an average end load of 25 kips in increments of 5 kips.At the end of the each increment, the beam was inspected for cracks and dial-gage displacement measurements were recorded. (2) Dial-gage measurements were recorded at an average end load of 30 kips.(3) Loading resumed with increasing displacement until failure.3(1)Monotonically-increasing load up to an average end load of 30 kips in increments of 5 kips.At the end of the each increment, the beam was inspected for cracks and dial-gage displacement measurements were recorded. (2) The beam was frilly unloaded and dial-gage displacement measurements were recorded. (3) The beam was loaded a second time up to an average end load of 35 kips in load increments of 5 kips. At the end of the each increment, dial-gage displacement measurements wererecorded. The beam was inspected for cracks at an average end load of 30 kips.(4) Loading resumed with increasing displacement until failure.4(l)Monotonically-increasing load up to an average end load of 35 kips in increments of 5 kips.At the end of the each increment, the beam was inspected for cracks and dial-gage displacement measurements were recorded. (2) Loading resumed with increasing displacement until failure.5(1)Monotonically-increasing load up to an average end load of 40 kips in increments of 5 kips.The beam was inspected for cracks and dial-gage displacement measurements were recorded atthe end of each increment. (2) The beam was fully unloaded and dial-gage displacement measurements were recorded. (3) The beam was loaded a second time up to an average end load of 40 kips in increments of 5kips. Dial-gage displacement measurements were recorded at the end of each increment. Thebeam was inspected for cracks at average end loads of 20, 30, 35 and 40 kips.(4) Loading resumed with increasing displacement until failure6(1)Monotonically-increasing load up to an average end load of 40 kips in increments of 5 kips.The beam was inspected for cracks and dial-gage displacement measurements were recorded atthe end of the each increment. (2) The beam was fully unloaded and dial-gage displacement measurements were recorded. (3) The beam was loaded a second time. The beam was inspected for cracks and dial-gage displacement measurements were recorded at average end loads of 10, 20, 30, 35, and 40 kips._____(4) Loading resumed with increasing displacement until failure.REDACTED VERSION

b-UUU i-'age (15 T01 111 42.4.6 Calibration Instruments used to measure force and displacement were calibrated following theprocedure specified in this section. The applied load was measured using load cells.Displacement transducers (either linear variable differential transformers or string potentiometers depending on availability) were used to record the vertical beam deflections. Load cells anddisplacement transducers were calibrated using a digitally-controlled hydraulic test framecalibrated annually using NIST-traceable standards. Load cell and displacement transducers werecalibrated following the steps listed below:1) The sensor (load cell or displacement transducer) was connected to the data-acquisition system that was used in the test.2) The sensor was securely mounted on the testing machine.3) A series of known force or displacement increments were applied to the sensor.Calibrations were performed exceeding the displacement and load range expectedduring the tests. In the case of load cells, calibrations were performed between zeroand 100 kips. In the case of displacement, calibrations were performed in a rangebetween zero and 4 in.4) Sensor output was recorded with the data-acquisition system at each knowndisplacement or force increment.

5) A least-squares linear regression analysis was performed on force and displacement versus sensor output to determine the calibration constant.

The load cells and displacement transducers were calibrated before and after testing eachthree beams and the calibration results are reported in Appendix G. The calibration constantdeviated with an average value of 0.28% for all sensors, ranging between 0 to 0.84%.2.5 Test Facilities The tests were performed in the Structural Testing Laboratory at the University of Kansas,a facility of the KU Structural Engineering and Materials Laboratory (SEML). The Laboratory has static and servo-hydraulic test equipment for the testing of steel, concrete, and composites. The structural testing bay has 4000 square feet of open laboratory area with a clear height of 30REDACTED VERSION

I-age 1140oT 1 114ft for large-scale structural testing. Loads up to 100,000 lb can be applied on 3-ft centers over a50 x 80 ft area. The laboratory houses a 600,000-lb universal testing machine for testing steeland concrete. A 450,000-lb MTS Structural Test System supported on a four-column test framemay be used for dynamic and cyclic testing of large scale structural components. 110,000-lb and55,000-lb MTS Structural Test Systems are also used for cyclic and dynamic testing of full-scale structural components within the test bay. Actuators within the test bay are powered by twohydraulic pumps (total flow rate of 110 gpm), meeting the requirements for demanding cyclictest applications. High-speed Mars Labs, National Instruments (used in the current study), andHewlett Packard data acquisition systems are available to monitor and record load, strain, anddisplacement. The structural testing laboratory includes an overhead 20-ton crane with access tothe entire lab floor area. Over 500 beam-end tests and over 200 splice tests have been performed in the KU Structural Testing Lab since 1990.Material tests were performed in the Concrete Laboratory, another SEML facility, whichis equipped to run standard tests on cement, aggregates, and concrete. Equipment is available totest concrete aggregate for deleterious

behavior, including alkali silica reactivity, and to measureaggregate properties as they affect mixture proportioning.

Freeze-thaw equipment is available forrunning tests under both Procedures A and B of ASTM C666. A walk-in freezer is used forscaling tests. Concrete is cured under controlled temperature and humidity in the lab's curingroom. Two hydraulic testing machines, with load capacities of 180 tons (400,000 lb),'oare usedfor concrete strength determination. Certificates of calibration for the equipment used in this study, including for the test frameused to calibrate the sensors, are presented in Appendix I.2.6 Section AnalysisSplice strength was evaluated based the calculated moment in the splice region at failure(ACI 408R-03). Loads, moments, and stresses for the beams were calculated using a two-dimensional analysis in which loads and reactions were assumed to act along the longitudinal centerline of the beam. Reactions and moments were based on load cell readings and the weightof the loading assemblies. The self-weight of the beam was included in the calculations based onaverage beam dimensions and an assumed concrete density of 150 pcf.REDACTED VERSION

r-age (lb o? 1114The test specimens were evaluated using cracked section theory assuming a linear straindistribution through the height of the cross-section. The beams were analyzed using anequivalent rectangular stress block and moment-curvature analyses for comparison. Themoment-curvature relationship was derived using the nonlinear stress-strain relationship forconcrete proposed by Hognestad (1951) and follows the procedure described by Nilson, Darwin,and Dolan (2010). Figure 2.11 shows the assumed stress distribution in the compression zone forthe moment-curvature and the equivalent rectangular stress block analyses. Good agreement inthe calculated bar stress at failure was typically noted between results obtained with the twomethods.O.85f; aC.=bO.85f-b and equivalent rectangular stress block analyses [after Nawy (2003)]In calculating splice strength, the tensile stress in the steel f5 (ksi) was calculated asfollowing the procedures used by ACI Committee 408 (2003):f,= Ex s= 29000 x ts forf, measured yield strengthfy (1)For es >fy/E5, f5 =fy for e, .e'h, where Csh = 0.0086 forty = 60 ksi and 0.0035 forty = 7ksi and above. There is no flat portion of the stress-strain curve forty 101.5 ksi. The modulusof strain hardening ELh = 614 ksi forty = 60 ksi, 713 ksi forty = 75 ksi, and 1212 ksi forty 90ksi. The values of 8sh and ELb forty between 60 and 90 ksi are obtained using linear interpolation. The equivalent rectangular stress block used in the calculations was proposed by Whitneywith the values of the parameter /J1 specified in ACI 318-11. The moment-curvature relationship was calculated using the concrete model proposed by Hoguestad (1951).REDACTED VERSION

M-age fibO0? 1114for= t,. o J(2)Io (i °-zc J1]1 for f"7= 0.85/' (3a)-1.7/'(b% = O.0038 (3c)= 1.8x106 +460fc' (3d)where:fc=concrete stress, psif"= concrete compressive

strength, psif"= peak concrete stress, psiSc= concrete strains0 = concrete strain at peak stressec,= ultimate concrete strain at crushingEc= approximate concrete modulus of elasticity, psiTensile stresses carried by the concrete were neglected in both analyses.

The calculations using both equivalent rectangular stress block and moment-curvature analyses proceed as follows:1. Select top face concrete strain sc in the inelastic range.2. Assume the neutral axis depth, at distance c below the top face.3. Assuming a linear variation in strain throughout the depth of the member, determine thetensile strain in the steel ss (equal to the tensile strain in the concrete at the level of thesteel sc).4. Compute the stress in the reinforcing steel in accordance with the defined stress-strain relationships (above). The tensile force in the steel T=fi x As (see Figure 2.11).REDACTED VERSION btYUi-UUU-UbJ-UbU--UUUlb-UUU P-age (11 oT 11 145. Determine the compressive force C, which equals to 0.85 fcba (Figure 2.11ib) for theequivalent rectangular stress block method, or by numerically integrating the concretestresses as defined by Eq. (2) and (3) for the moment curvature method.6. If C -- T, go to step 7. If not, adjust the neutral axis depth c in step 2 and repeats steps 3 -5.7. Using the internal lever arm z from the centroid of the concrete stress distribution to thetensile resultant, the calculated bending moment M= Cz = Tz.8. If the calculated bending moment M equals the applied bending moment (from test), equals the force in the reinforcing steel. If the calculated bending moment does not equalthe applied bending moment, modify ec and c in steps 1 and 2, respectively, and repeatsteps 3 -7 until the calculated bending moment M equals the applied bending moment.REDACTED VERSION

r~age fib5 oT 11143 Test Results3.1 GeneralThe testing program consisted of six beam-splice specimens. Three of the specimens had alap splice length of 79 in., and three had a lap splice length of 120 in. The measured loads andcalculated bar stresses at failure are presented in Table 3.1. In addition to failure loads, Table 3.1includes measured material properties and bar cover dimensions. Bar stresses at failure listed inTable 3.1 include those calculated using the equivalent rectangular stress block and moment-curvature analysis. Measured specimen dimensions and other details of the beam tests arepresented in Appendix H.Moment-curvature analyses consistently produced calculated higher bar stresses than didthe analysis using the equivalent rectangular stress block. This is to be expected because theparameters of the equivalent stress block were calibrated to reflect the characteristics of thecompression zone when the peak strain in the concrete exceeds 0.003 and the concrete in thecompression zone is well into the nonlinear range. Under these conditions, the depth of thecompression zone is reduced, resulting in a slightly larger distance between the tension andcompression resultants. With the exception of Beam 1, the splices failed prior to crushing of theconcrete in the bottom surface of the beam, so it was to be expected that the equivalent rectangular stress block would slightly overestimate the distance between tension andcompression resultants and consequently underestimate the stress in the reinforcing bars. Thedifference, on average, between the bar stresses at failure calculated by the two methods was 1.5ksi for the six beams tested in this study, with moment-curvature analysis producing the greatervalue. All bar stress values discussed subsequently are those calculated using moment-curvature

analysis, which is considered to be more accurate method for the reasons stated above.REDACTED VERSION bbtSi-UUU-6bI3-UbUr_-UUUlb-UUU Page (19 aT"1114Table 3.1 -Bar stresses at failure for beam-sp lice specimens Total ,CacltdICalculated bar stressBaID-Sl Ce oConcretemomentaat at failure, ksiBemD-Spie

strength, Cocrt splice omnatFailure modelegt pi cover; in.a failure, splice Equiv. Moment-legt ' failure, kip-ft stress "kips block curvature 1 -79 in. Flexural(monolithic) 3// 103 344 70 70 Failure *2 -79 in.(cold joint, loaded 53/ 3/3/3 85 292 59 62 Spiue*monotonically) 5330/+alu&

3 -79 in.(cold joint, Spliceunloaded and 3.25/3.35/2.9 80 270 53 57failure** reloaded) Splice failure4 -120 in. and(cold joint, loaded 3/2.8/3.4 105 350 71 72 secondary monotonically) flexuralfailure** 5 -120 in. Splice failure(cold j oint, 523 0/ andunloaded and 5490k 3.15/3/15/3.15 96 325 66 67 secondary flexuralreloaded) failure*** 6 -l120in. Splice failure(cold joint, andunloaded and 3.15/3.15/2.9 100 338 69 69 secondary flexuralreloaded) failure~*** " Top cover/north side cover/south side cover+ Compressive strength of concrete below and above the cold joint.Test was stopped after reinforcing steel yielded, when crushing of the concrete in the compression zone wasobserved.

• Splice failed prior to yielding of the flexure reinforcement.

Splice failed after yielding of the flexure reinforcement 3.2 Beams 1, 2, and 3 with 79-in, splice length3.2.1 Concrete strengthThe concrete strengths for Beams 1, 2 and 3 are summarized in Table 3.2. Beam 1 wascast monolithically, while Beams 2 and 3 were cast in two stages to accommodate the presenceof a cold joint at the level of the flexure reinforcement. Beam 1 and the concrete below the coldjoint for Beams 2 and 3 were placed on May 24, 2012 and the concrete above the cold joint wasplaced on May 25, 2012. The forms were removed on May 28, 2012, when the average concreteREDACTED VERSION DbL-UUU-U5di-UI-.i-UUU1 b-UUU l-'age /2U Ot 1 114compressive strength for both placements exceeded 3500 psi. All three beams were tested onMay 31, 2012. On that date the concrete from the first placement had an average compressive strength of 5330 psi, and the concrete from the second placement had an average compressive strength of 4330 psi (Table 3.2). The average split cylinder strength and the average modulusrupture were 435 and 570 psi for the concrete below the cold joint in accordance with ASTMC496 and ASTM C78, respectively. The tensile strength for the concrete above the cold jointwas not recorded for the first three beams. The flexural beam specimens with cold joints werealso tested and had an average modulus of rupture of 140 psi, significantly lower than that ofspecimens cast monolithically. The fact that the tensile strength of the flexural beam specimens with cold joints was significantly lower than the strength of monolithic specimens indicates thatthe presence of a cold joint did in fact introduce a weak plane at the level of reinforcing steel.The proportions of the concrete mixture and the properties of the concrete for each placement arereported in Table E.2 of Appendix E.Tahle 3.2 Concrete strengths for Beams 1, 2, and 3Concrete below cold joint Concrete above cold jointAverage Compressive Strength when 4010oa360 forms were removedAverage Compressive Strength at test 5330c430date, psiSplit Cylinder Strength (ASTM 435C -C496), psiModulus of Rupture (ASTM C78), psi 570C -Modulus of Rupture for specimens 140with cold joint, psiaTested at 4 days; btested at 3 days; ctested at 7 days; dtested at 6 daysA segment of the No. 11 bars used in the splice-beam specimens was tested in tensionand the bar strains were recorded using a linear variable differential transformer (LVDT ) used asthe extensometer (gage length = 8.0 in.). The measured stress-strain curve for the No. 11 bar isshown in Figure 3.1. The yield stress calculated using the 0.2% offset method was 67 ksi and themeasured elastic modulus was 28,990 ksi. The maximum measured steel stress was 105 ksi.REDACTED VERSION

I-page (21 0? 1114120000100000600080000-- --40000III Yield (0.2% offset): 67.1 ksi20000 _____-________ Ultimate:-104.7 k si...FIEasicModulus:28990 ksi~lsi00 0.05 0.1 0.15StrainFigure 3.1 -Measured stress-strain curve for No. 11 bar3.2.2 Beam 1 (monolithic concrete) 3.2.2.1 Beam 1 load-deflection curveBeam 1 was cast monolithically with a splice length of 79 in. It was loadedmonotonically to failure (the load protocol is presented in Table 2.5). The load-deflection curvefor Beam 1 is shown in Figure 3.2. The displacement shown in the figure was calculated byadding the average of the displacement at the two load points to the displacement at the beamcenterline. The load shown in the figure corresponds to the total load applied to the beam (thesum of the two end loads). The load-deflection relationship shows that there was a significant reduction in the stiffness of the beam at a total load of approximately 20 kips, which coincided with the first observation of flexural cracks. Another significant reduction in flexural stiffness was observed at a total load of 94 kips and a total displacement of approximately 2.8 in. In thiscase the reduction in stiffness is attributed to the yielding of the flexural reinforcement. Thecalculated bar stress corresponding to the total load of 94 kips is 68 ksi based on moment-curvature analysis (Table 3.1). The positive slope of the load-deflection relationship after a totalload of 94 kips is attributed to the strain hardening of the reinforcing steel. Loading continued until a flexural failure occurred, which was accompanied by crushing of the concrete in theREDACTED VERSION bDJI-UUU-Uditt-UWI:-UUUI h-UUU -'age 122Z Ot 1 114compression zone, near the supports, at a total load of 103 kips, corresponding to a bar stress of70 ksi, and a total deflection of approximately 5 in. (Figure 3.3).0.0d0110100908070605040302010002 3 45Total Deflection, in.6Figure 3.2 -Total load vs. total deflection for Beam 1 (cast monolithically) (Total loadcalculated as the summation of the two end loads and total deflection calculated defined byadding the average deflection at two ends and the deflection in the beam centerline). Figure 3.3 -Flexural failure in the compression region for Beam 1. Numbers indicate maximumaverage end load when cracks marked.REDACTED VERSION

r-age f"J OT II1143.2.2.2 Crack progression-Beam 1Maximum measured crack width versus average end load for Beam 1 is shown in Figure3.4; the crack map for Beam 1 is presented in Figure 3.5 (see figures in Appendix C for greaterdetail). The first flexural cracks formed near the east support at the end of the east splice region,at an average end load of 10 kips (total load of approximately 20 kips). The flexural cracks grewprogressively wider and more numerous as the load increased. The first horizontal crack formednear the support at an average end load of 25 kips (Figure 3.6). Both longitudinal and flexuralcracks continued to increase in width and number as the load increased. At the last crackmarking prior to failure (average end load of 40 kips), the widest flexural crack had a width of 25mils (0.025 in.) and the widest horizontal (bond) crack had a width of 18 mils (0.018 in.).60) 50.E10,0 10 20 30 40 50Average End Load, Kips-E-FlexuraI Crack -+Horizontal CrackFigure 3.4 -Maximum crack width vs. average end load (one-half of total load) for Beam 1.REDACTED VERSION bbJL-UUU-dJtiSULILI-UUU1 U-UUU r-age (Z4 Ot 1 114EaweweNorth FaceSouth FaceTop FaceI aLoading Pedestal Splice Center Splice Pedestal LoadingPoint Support Region Line Region SupportPon Ea$tFigure 3.5 -Crack map for Beam 1. Numbers indicate maximum average end load when cracksmarked. See Figure C. 1 in Appendix C for greater detail.Figure 3.6 -Beam 1, north side of east support with horizontal crack, 25 kip end load.Failure occurred at an average end load of 51 kips (total load of 103 kips). The failuremode was yielding of the bars followed by crushing of the concrete near the supports (Figure3.7). Both flexural and horizontal cracks were present near the splice region (Figure 3.8). At theREDACTED VERSION bbUJ~-UUUJ-U-LG--UUU"Ib-UUU I'-'age /b OT 01 114support (Figure 3.9), flexural cracks extended most of the depth of the beam; no horizontal cracks were present.A detailed autopsy was not performed on Beam 1. Concrete was removed in selectedregions to verify the concrete cover to the splice was within tolerances. Top cover was 3 in. tothe outer bar in the splice and 3-1/4 in. to the inner bar in the splice for both splices.Figure 3.7 -Beam 1, underside near support, failure.REDACTED VERSION

r-age (~t 0T 1 114Figure 3.8 -Beam 1, north side of west splice region, failure.Figure 3.9 -Beam 1, south side of east support, failure.REDACTED VERSION

r-age f(2 0T1 1114Figure 3.10 -Beam 1, centerline, failure.3.2.3 Beam 2 (cold joint, monotonically-loaded) 3.2.3.1 Beam 2 load-deflection curveBeam 2 was cast with a cold joint in the plane of reinforcing steel. It was monotonically loaded with a load increment of approximately 5 kips (average end load, the load protocol ispresented in Table 2.5). The load-deflection curve for Beam 2 is shown in Figure 3.11. The totaldisplacement and total load shown in the figure were calculated in the same manner as for Beam1. The total load corresponding to cracking was very similar to that of Beam 1, approximately 20kips. The beam was loaded to a maximum total load of 85 kips, with a corresponding totaldisplacement of 2.25 in. At this point the beam failed with a sudden splitting of the concretealong the cold joint. Wide horizontal cracks were observed in the plane of the cold joint withinthe splice region (Figure 3.12). The widest horizontal crack was measured to be 1/22 in. wide afterfailure. It is concluded that the beam failed due to failure of the splice at a total load of 85 kips.The calculated bar stress corresponding to the total load of 86 kips is 62 ksi based on moment-REDACTED VERSION

i-8ge 0? 1I14curvature analysis (Table 3.1), above the minimum specified yield strength of 60 ksi for Grade60 reinforcement but 5 ksi below the actual yield strength of 67 ksi.0d9080706050403020100..... ... ..... .85 kip s012Total Deflection, in.34Figure 3.11 -Total load vs. total deflection for Beam 2 (with a cold joint)Wide horizontal crack at failurewithin the splice regionFigure 3.12 -Beam 2 (with a cold joint) failed with wide horizontal crackREDACTED VERSION

I-UUU I-'age hiB ot 11143.2.3.2 Crack progression-Beam 2Maximum measured crack width versus load for Beam 2 is shown in Figure 3.13; thecrack map for Beam 2 is presented in Figure 3.14. The first flexural cracks formed near thesupports and ends of both splice regions at an average end load of 15 kips (total load of 30 kips).Horizontal cracks first formed at an average end load of 20 kips at both ends of the splice regionalong the cold joint (Figure 3.15). Both longitudinal and flexural cracks continued to increase inwidth and number as the load increased, with horizontal cracks propagating along the cold joint.When the last cracks were marked prior to failure (conducted at an average end load of 30 kips),the widest flexural crack had a width of 20 mils (0.02 in.) and the widest horizontal crack had awidth of 13 mils (0.0 13 in.).60~4oU30I..0 10 20 30 40 50Average End Load, Kips--Flexural Crack Horizontal CrackFigure 3.13 -Maximum crack width vs. average end load for Beam 2.REDACTED VERSION DbiL-UUU-U-LJU-S-t-UUUIIb-UUU I-age fiU t 0 1114EaStweS1North FaceSouth FaceTop Face.5 IL 27 IWesp. ~,7- I<" I tw reSLoadingPointPedestal Splice Center SpliceSupport Region Line RegionPedestalSupportEasLoadingPointFigure 3.14 -Crack map for Beam 2. Numbers indicate maximum average end load whencracks marked. See Figure C.2 in Appendix C for greater detail.Figure 3.15 -Beam 2, northeast support with horizontal crack, 20 kip end load.Failure occurred at an average end load of approximately 43 kips (total load of 85 kips).At failure, the concrete above the cold joint separated from the remainder of the beam (Figure3.16). Near the splice region, a large flexural crack was also present (Figure 3.16). TheREDACTED VERSION

w'age ,1i Or 11 14horizontal crack progressed approximately 12 in. past both ends of the splice region, and with theexception of near the centerline, continued along the cold joint. At the centerline, the crack splitthrough the cover and around the single hoop present at the centerline (Figure 3.17), indicating the hoop was effective in preventing the crack from growing near the centerline. As shown inFigure 3.17, the region affected by the hoop was small.Figure 3.16 -Beam 2, southwest splice region showing separation of concrete, 43 kip end load.REDACTED VERSION bbI-UUU-U56.i-{U:-UUU1bJ-UUU r-age 1;S2 0O" 1 114Figure 3.17 -Beam 2, centerline at failure.3.2.4 Beam 3 (cold joint, cycled)3.2.4.1 Beam 3 load-deflection curveBeam 3 was cast in the same manner and at the same time as Beam 2, with a cold joint inthe plane of reinforcing steel. Instead of loading the beams to failure monotonically, Beam 3 wasfirst loaded to a total load of 60 kips, unloaded to zero, and then re-loaded monotonically tofailure (the load protocol is presented in Table 2.5). When the beam was first loaded to a totalload of 60 kips (average end load of 30 kips), the average end load was increased in increments of approximately 5 kips. The specimen was inspected for cracks, which were marked at eachload step. At a total load of 60 kips, the maximum horizontal crack width was 20 mils (0.02 in.).When the beam was loaded for the second time, it was loaded up to a total load of 60 kipswithout inspecting for cracks. The only visual measurement conducted during the second loadingwas the recording of dial gage readings at approximately 5-kip increments (average end load).The beam was inspected for cracks again when the total load reached 60 kips for the secondtime. At this point some of the horizontal cracks widened to a maximum width of 35 mils (0.035in.)REDACTED VERSION

l-'age /'JJ. OT 1114The load-deflection curve for beam 3 is shown in Figure 3.18. Overall, Beam 3performed very similar to Beam 2, except for the peak load. The beam failed at a total load of 80kips (compared with a total load of 85 kips for Beam 2), in the same manner as observed forBeam 2. A wide horizontal crack in the plane of the cold joint, within the splice region, wasobserved after failure (Figure 3.19), with the widest portion of the crack being 3/8-in. It isconcluded that the beam failed due to a splice failure. The calculated bar stress corresponding tothe total load of 80 kips is 57 ksi based on moment-curvature analysis (Table 3.1).rj~Cu0Cu0908070605040302010... ............. .......... .. .... ... ..._ 8 0O k ip s012 3 4Total Deflection, in.Total load vs. total deflection for Beam 3 (with a cold joint)Figure 3.18Figure 3.19 -Beam 3 failure with wide horizontal cracks along cold jointREDACTED VERSION

h-UUU I-sage f 5440t 11143.2.4.2 Crack progression-Beam 3Maximum measured crack width versus load for Beam 3 is shown in Figure 3.20; thecrack map for Beam 3 is presented in Figure 3.21. As seen in both figures, the first flexuralcracks formed near end of the east splice region at an average end load of 10 kips (total load of20 kips). At an average end load of 15 kips, flexural cracks were present at both ends of thesplice region and both supports. A horizontal crack first formed at an average end load of 15 kipsat the west end of the splice region along the cold joint, with additional horizontal cracksforming and reaching a 9-mail (0.009 in.) width at an average end load of 20 kips (Figure 3.22).At an average end load of 30 kips, a 40-mit (0.04-in.) width flexural crack and 20-mait widthhorizontal crack were recorded. At this point, the beam was unloaded. With zero load, themaximum flexural and horizontal crack widths decreased to 13 and 7 mils (0.013 and 0.007 in.),respectively. The load was reapplied, and at the last crack mapping (average end load of 30kips), the widest flexural crack had a width of 55 mils (0.05 5 in.) and the widest horizontal crackhad a width of 35 mils (0.03 5 in.), much wider than the cracks noted at the first loading to a 30-kip average end load.60O°E5040S30E 2100AI0 10 20 30 40 50Average End Load, Kips-~-Flexural Crack (1st loading) Crack (1st Loading)--Flexural Crack (2nd loading) -*- -Horizontal Crack (2nd Loading)Figure 3.20 -Maximum crack width vs. average end load for Beam 3.REDACTED VERSION

I-'age 1.iD 0?" 1114a,iNorth FaceYWe"We tSouth Face-(IE~.. .... .... .... .. tTop FaceWe r1 ........ f e l .. .. ..Pedestal Sp~lceSupport Region.... '. l .[-Cettef Splice Pedestaltine Reglon Supor... I IaEI !adnFigure 3.21 -Crack map for Beam 3. Numbers indicate maximum average end load whencracks marked. See Figure C.3 in Appendix C for greater detail.Figure 3.18 -Beam 3, northwest splice region with horizontal crack, 20 kip end load.Failure occurred at an average end load of 40 kips (total load of 80 kips), a slightly lowerload than the monotonically loaded Beam 2 (total load of 85 kips). At failure, the concrete abovethe cold joint separated from the remainder of the beam, with the horizontal crack propagating along the cold joint in a region that was somewhat larger than the splice region except for a smallREDACTED VERSION bbiL-UUU-bJ5i-GW-(-UUU1ti-UUU Mage I Jti 01" 1114region near the centerline, which was restrained by the No. 3-bar hoop (Figure 3.23). Largeflexural cracks were also present near both ends of the splice region.Figure 3.23 -Beam 3, splice region and centerline showing separation of concrete, 40 kip endload.3.3 Beams 4, 5, and 6 with 120-in, splice length3.3.1 Concrete strengthThe concrete strengths for Beams 4, 5 and 6 are summarized in Table 3.3. The threebeams were cast in two stages to accommodate the presence of a cold joint at the level of theflexural reinforcement. The concrete below the cold joint was placed on June 13, 2012, and theconcrete above the cold joint was placed on June 14, 2012. The forms were removed on June 17,2012 when the average concrete compressive strength for both placements exceeded 3500 psi.The beams were tested on June 20, 2012. On that date, the concrete from the first placement hadan average compressive strength of 5230 psi, and the concrete from the second placement had anaverage compressive strength of 5490 psi (Table 3.3). The higher strength for the secondREDACTED VERSION

-'age (i/ 0T71114placement was likely due to the slightly lower water-cement ratio of the concrete, as shown onthe batch ticket (Appendix H). The average split cylinder strength and average modulus rupturewere, respectively, 370 and 600 psi for the concrete below the cold joint and 470 and 700 psi forthe concrete above the cold joint. The flexural beam specimens with cold joints were also testedand had an average modulus of rupture of 274 psi, significantly below that of specimens castmonolithically. The proportions of the concrete mixture and the properties of the concrete foreach placement are reported in Table E.2 of Appendix E.Table 3.3 -Concrete strengths for Beams 4, 5, and 6Concrete below cold joint Concrete above cold jointAverage Compressive Strength when 4310oa450 Forms were removedAverage Compressive Strength at test 5230c540date, psiSplit Cylinder Strength (ASTM 370° 70C496), psi ___________ Modulus of Rupture (ASTM C78), psi 600c 700dModulus of Rupture for specimens 274d __with cold joint, psiaTested at 4 days; btested at 3 days; ctested at 7 days; dtested at 6 daysThe same reinforcing steel was used for Beams 4, 5, and 6 as for Beams 1, 2, and 3. Themeasured stress-strain curve for the No. 11 bar is shown in Figure 3.1.3.3.2 Beam 4 (cold joint, monotonically-loaded) 3.3.2.1 Beam 4 load-deflection curveBeam 4 was cast with a cold joint in the plane of reinforcing steel. It was subjected tomonotonically-increasing load in increments of approximately 5 kips (average end load, theloading protocol is presented in Table 2.5). The load-deflection curve for Beam 4 is shown inFigure 3.24. The total load and deflection were determined in the same manner as for Beams 1, 2and 3. The flexural stiffness of the beam decreased once the total load exceeded 20 kips,coinciding with the formation of flexural cracks. A sharp decrease in the slope of the load-deflection curve was observed at a total load of about 94 kips and corresponding deflection ofREDACTED VERSION bbdi-UUU-(.ibi-ti-L~.UUUI b-UUU r'age lid Ot 1114approximately 2.8 in. The stress at the end of the spliced bars for a total load of 94 kips was 68ksi. The decrease in the slope of the load-deflection curve at a total load of 94 kips indicates thatthe reinforcing steel yielded. After yielding of the reinforcing steel, the total load continued toincrease but at a lower rate, which is attributed to the strain hardening of the reinforcing steel.The beam was loaded to a total load of 105 kips (and a displacement of 5.5 in.) and at that pointfailed with the sudden splitting of the concrete along the cold joint. Wide horizontal cracks in theplane of the cold joint were observed within the splice region. Wide flexural cracks were alsoobserved near the support (Figure 3.25). It is concluded that the reinforcing steel yielded at atotal load of approximately 94 kips and beam failed at a total load of 105 kips due to failure ofthe splice, the latter corresponding to a bar stress of 72 ksi (Table 3.1).11010090~70*;40~3001000 2 4 68Total Deflection, in.10Figure 3.194 -Total load vs. total deflection for Beam 4 (with a cold joint)REDACTED VERSION

I-sage fJ31 ot 1114Figure 3.20 -Beam 4 (with a cold joint) at failure3.3.2.2 Crack progression-Beam 4Maximum measured crack width versus load for Beam 4 is shown in Figure 3.26; thecrack map for Beam 4 is presented in Figure 3.27. The first flexural cracks formed near end ofthe west support at an average end load of 10 kips (total load of 20 kips). At an average end loadof 15 kips, flexural cracks were present at both ends of the splice region and both supports. Horizontal cracks first formed at an average end load of 20 kips, at the both ends of the spliceregion along the cold joint. Both longitudinal and flexural cracks continued to increase in widthand number as the load increased, with horizontal cracks propagating along the cold joint. At thelast load prior to failure at which cracks were marked (average end load of 35 kips), the widestflexural crack had a width of 30 mils and the widest horizontal crack had a width of 16 mils. Atthis point, the horizontal cracks extended along most of the length of the splice region (Figure3.28), with some of the horizontal cracks that formed at earlier stages merging together. REDACTED VERSION

I-age / 4U Ot 1 11460E20 03 05Average End Load, Kips--Flexural Crack Horizontal CrackFigure 3.21 -Maximum crack width vs. average end load for Beam 4.EaWwtNorth FaceSouth Face-r~ <:9~~I:ii2F~i2 -- -. -i7SWeT~ tIaPointwe$odnPo~fTop Face-IlefleLineStqo uppOrtFigure 3.27 -Crack map for Beam 4. Numbers indicate maximum average end load whencracks marked. See Figure C.4 in Appendix C for greater detail.REDACTED VERSION

I-age (41 OT 1 114Figure 3.28 -Beam 4, south side of west splice region with horizontal cracks, 35-kip end load.At failure, the concrete above the cold joint separated from the remainder of the beam,with the horizontal crack propagating along the cold joint between the pedestal supports exceptfor a small region near the centerline that was restrained by the No. 3-bar hoop (Figure 3.29).Large flexural cracks were also present near both ends of the splice region (Figure 3.30).Figure 3.29 -Beam 4, centerline showing separation of concrete, 52-kip end load.REDACTED VERSION bbJL-UUU-UI4~(..$tLIL-UUUII-UUU I-'age /4>' 0? 1114Figure 3.30 -Beam 4, end of splice region at 52-kip end load.3.3.3 Beam 5 (cold joint, cycled)3.3.3.1 Beami 5 load-deflection curveBeams 5 and 6 were cast in the same manner and at the same time as Beam 4, with a coldjoint at the plane of reinforcing steel. Instead of monotonically loading the beams to failure,Beam 5 was first loaded to a total load of 80 kips, and subsequently unloaded to zero, and thenre-loaded to failure (the load protocol is presented in Table 2.5). When the beam was first loadedto a total load of 80 kips, the average end load was increased in increments of approximately 5kips. The specimen was inspected for cracks and marked at each load step. Horizontal cracks onthe plane of the cold joint within the splice region were observed when the beam was subjected to a total load of 80 kips. The maximum horizontal crack width at this load was 35 mils (0.03 5in.). It should be noted that the beam was unloaded in a rapid manner and that one of the loadcells had large fluctuations after that point (load cell C in Figure 3.31). Although there were clearproblems with the load readings from load cell C for the remainder of this test, the rams were atall times subjected to uniform pressure, and load readings from the other 5 beam tests show thatthe load was evenly applied to the four different load rods at all times. Furthermore, the loadREDACTED VERSION

t5-UUU t-'age f4Jt 0" 1 114beam remained level and the displacement readings were similar at both ends of the beam, strongindicators that although the load cell readings were not accurate, the load was uniformly appliedto the four load rods. Based on these observations, the total load was calculated based on thereadings from load cells A and B. When the beam was loaded for the second time, it was loadedup to a total load of 80 kips at an increment of 5 kips (average end load). At the end of the eachincrement, dial-gage displacement measurements were recorded. The beam was inspected forcracks at total loads of 40, 60, 70, and 80 kips. When the beam was inspected for crack duringthe second loading, some of the horizontal cracks elongated or widened and some new horizontal cracks were noticed. The maximum horizontal crack width was still 35 mils (0.035 in.)35 __-_ __- _-_ Load Cell A(west)30 Load Cell B(west)___ __ __25-___- Load Cell C (east)20_ ___ Load Cell D (east)0 1523 45-Total Deflection, in.Figure 3.22 -Load cell readings for Beam 5The load-deflection curve for Beam 5 is shown in Figure 3.32. Due to the problemdocumented for load cell C, the total load is calculated as twice the summation of load cells Aand B, located at the West loading point. Overall, Beam 5 performed very similar to Beam 4.The slope of the load-deflection curve first decreased at a total load of 20 kips, which coincides with the first observation of flexural cracks. Another decrease in the slope of the load-deflection curve was observed at a total load 91 kips, with a corresponding total displacement ofapproximately 2.7 in, which is attributed to the yielding of the flexural reinforcement. Thecalculated bar stress corresponding to the total load of 91 kips is 66 ksi based on moment-curvature analysis. The positive slope of the load-deflection relationship after a total load of 91kips is attributed to the strain hardening of the reinforcing steel. The beam was loaded to a totalREDACTED VERSION

I-'age f4401" 1114load of 96 kips, with a corresponding total displacement of 3.6 in., at which point the beam failedsuddenly. Wide flexural cracks near the support and horizontal cracks in the plane of cold jointwere observed within the splice region (Figure 3.33). It is concluded that the reinforcing steelyielded at a total load of 91 kips and beam failed at a total load of 96 kips due to failure of thesplice, the latter corresponding to a bar stress of 67 ksi (Table 3.1).110 -_ _ _ _ _ _100 -- 91kips90i60S50 --_ _S40 -S3020-- ---__ _ _ _10 --- * -- _ -_0 ,r0 2 4 6 8 10Total Deflection, in.Figure 3.32 -Total load vs. total deflection for Beam 5 (with a cold joint)Flexural cracks near the supportFigure 3.33 -Beam 5 (with a cold joint) at failureREDACTED VERSION

I-sage 4tb Ot 111l43.3.3.2 Crack progression-Beam 5Maximum measured crack width versus load for Beam 5 is shown in Figure 3.34; thecrack map for Beam 5 is presented in Figure 3.35. The first flexural and horizontal cracks formedat the supports at an average end load of 10 kips (total load of 20 kips). At an average end load of15 kips, flexural and horizontal cracks were present at both ends of the splice region and bothsupports (Figure 3.36). At an average end load of 40 kips, a 45-mil width flexural crack and 35-miu width horizontal crack were recorded. At this point, the beam was unloaded. The load wasreapplied, and at the last load prior to failure at which cracks were marked (average end load of40 kips), the maximum width of the cracks had not increased from first loading (Figure 3.34).Although the crack width was approximately the same, several cracks had increased in length.60oE 50S40~30E_ 200 10 20 30 40 50Average End Load, KipsSFlexural Crack (1st loading) -s Horizontal Crack (1st Loading)--Flexural Crack (2nd loading) -4.0- Horizontal Crack (2nd Loading)Figure 3.34 -Maximum crack width vs. average end load for Beam 5.REDACTED VERSION

t-'age Mb1 oT 1 114awINotth Face,South F,,ce -~717 I ~2 7~4,w$.% z ',14I~ETop FaceWLodnI$~kce PedestalRe~on SupportEodsPonPedestai 5p1ieCenIerFigure 3.35-Crack map for Beam 5. Numbers indicate maximum average end load whencracks marked. See Figure C.5 in Appendix C for greater detail.Figure 3.36 -Beam 5, northeast splice region with horizontal crack, 15 kip end load.Failure occurred at an average end load of 48 kips (total load of 96 kips), slightly lowerthan the failure load for Beam 4 (average end load of 52 kips, total load of 105 kips), which wasREDACTED VERSION

r'age (41 0l11114subjected to monotonically-increasing load up to failure. At failure of Beam 5, the concreteabove the cold joint separated from the remainder of the beam, with the horizontal crackpropagating along the cold joint throughout a region that was somewhat longer than the spliceregion. A small region near the centerline was restrained by the No. 3-bar hoop (Figure 3.37) andhad a tighter horizontal crack and a failure surface that passed through the top of the beam in thevicinity of the hoop, as shown in Figure 3.35. As with the other beams, large flexural crackswere also present near both ends of the splice region (Figure 3.38).Figure 3.37 -Beam 5, centerline showing separation of concrete, 48-kip end load.REDACTED VERSION bbi-UUU-UdtJt-GLUb-UUUIb-UUU I-'age I 4d ot 1 114Figure 3.38 -Beam 5, splice region, 48-kip end load.3.3.4 Beam 6 (cold joint, cycled)3.3.4.1 Beam 6 load-deflection curveThe configuration and loading protocol of Beam 6 were similar to those of Beam 5. Thebeams were cast using the same procedures and at the same time and were tested in the samemanner, except that unloading was much slower for Beam 6 and the beam was inspected forcracks more often during the second loading. The testing protocol for Beam 6 is presented inTable 2.5.The individual load cell readings are plotted versus total deflection in Figure 3.39. Asshown in Figure 3.39, the readings for the four load cells were identical, which verifies theassumption in Section 3.3.3 that the load was evenly distributed on the four load rods.REDACTED VERSION

I-age 141:: 0?" 1 114( 3 ) -Load Cell A (west)~25 ---Load Cell B (west)~20 -Load Cell C (east)105 -..-Load Cell (east)024 6 8 10Total Deflection, in.Figure 3.39 -Individual load cell readings (Beam 6)The total load versus total deflection for Beam 6 is plotted in Figure 3.40. Overall, Beam6 performed very similar to Beam 5. Yielding of the flexural reinforcement was observed at atotal load of 92 kips and a total displacement of 2.7 in., compared with 91 kips and 2.7 in. forBeam 5. The maximum horizontal crack width at the unloading point was 30 mils (0.03 in.),compared with 35 mils (0.035 in.) for Beam 5. Beam 6 also failed due to splice failure (Figure3.41) at a total load of 100 kips, corresponding to a bar stress of 69 ksi, and a total deflection of4.7 in. (versus 96 kips and 3.6 in. for Beam 5).110100 92 kips... .90~'60~50~40~3020100 ,ii.024 6 8 10Total Deflection, in.Figure 3.40 -Total load vs. total deflection for Beam 6 (with a cold joint)REDACTED VERSION bbgt-UUU-~bi.5t-ULUII-UUUllh-UUU r-age (bU 0t1 1114Flxrlcracksnathsupr Figure 3.41 -Beam 6 (with a cold joint) at failure3.3.4.2 Crack progression-Beam 6Maximum measured crack width versus load for Beam 6 is shown in Figure 3.42; thecrack map for Beam 6 is presented in Figure 3.43. The first flexural cracks formed at the eastsplice region and support at an average end load of 10 kips. At an average end load of 25 kips,flexural and horizontal cracks were present at both ends of the splice region and both supports(Figure 3.44). At an average end load of 40 kips, a 35-mil (0.035 in.) wide flexural crack and 30-mul (0.03 in.) wide horizontal crack were recorded. At this point, the beam was unloaded. Theload was reapplied, and at the last load prior to failure at which cracks were marked (average endload of 40 kips), the crack width had not increased with respect to first loading (Figure 3.42).Although the maximum crack widths remained the same, several cracks had increased in length.REDACTED VERSION

I-'age (/,l 01111460E~ 50"o 40~30U.20._ 10n0 10 20 30 40 50Average End Load, Kips Crack (1st loading) Crack (1st Loading)-*Ftexural Crack (2nd loading) --e--Horizontal Crack (2nd Loading)Figure 3.42 -Maximum crack width vs. average end load for Beam 6.North FaceS >.. --7... .... --. -.......... South Facew .LTop Faceloading Spike Ceiner Splice Pedestal LoadingPontn Suppoet Un. Relf Support PolngFigure 3.43 -CTrack map for Beam 6. Numbers indicate maximum average end load whencracks marked. See Figure C7.6 in Appendix C7 for greater detail.REDACTED VERSION

I-'age (b2 Ot l1114Figure 3.44 -Beam 6, splice region with horizontal crack, 25-kip end load.Failure occurred at an average end load of 50 kips, slightly lower than for Beam 4(average end load of 52 kips, total load of 105 kips), and higher than Beam 5 (average end loadof 48 kips, total load of 96 kips). As observed in Beams 2 through 5, at failure occurred at thecold joint with the upper concrete separating from the remainder of the beam, with the horizontal crack propagating along the cold joint between the pedestal supports. As for Beam 5, a smallregion near the centerline was restrained by the No. 3-bar hoop (Figure 3.45) and had a tighterhorizontal crack and a failure surface that passed through the top of the beam in the vicinity ofthe hoop, as shown in Figure 3.49. As in the case of the other beams, large flexural cracks werealso present near both ends of the splice region (Figure 3.46).REDACTED VERSION bb~tUUU-UdiL-U I:U-UUU1 ti-UUU I-'age fbLJ OT 1 114Figure 3.45 -Beam 6, centerline showing separation of concrete, 50 kip end load.Figure 3.46 -Beam 6, splice region, 50-kip end load.REDACTED VERSION bbi-UUU-3-iSIr-.iSU-UUUIb-UUU r-age fb40ot 1 11,44 Summary and Conclusions The effect of preexisting cracks, oriented in the plane of and parallel to the reinforcing steel, on the strength of No. 11-bar lap splices was investigated by testing six beams -three witha splice length of 79 in. and three with a splice length of 120 in. One of the beams with a 79-in.splice was cast monolithically and loaded monotonically to failure. To simulate the cracks, theother five specimens were cast with a cold joint at the mid-height of the reinforcing steel. Twobeams (one with a 79-in, splice and one with a 120-in, splice) were loaded monotonically tofailure. The other three beams were pre-loaded to develop horizontal cracks in the face of thecold joint, unloaded and then loaded to failure; those beams developed horizontal cracks withwidths of 20, 30 and 35 mils (0.02, 0.03, 0.03 5 in.) just prior to unloading. The test results aresummarized below:1. For the beam with a splice length of 79 in. and cast with monolithic

concrete, thereinforcing steel yielded and the beam failed in flexure.2. For the beam with a splice length of 79 in., cast with a cold joint, and subjected tomonotonically-increasing load to failure, splice failure took place at a bar stress of 62 ksi,about 8% below the bar yield strength of 67 ksi.3. For the beam with a splice length of 79 in., cast with a cold joint and subjected to cyclicloading, horizontal cracks with a maximum width of 20 mils (0.02 in) developed prior tofailure.

Splice failure took place prior at a bar stress of 57 ksi, about 15% below the.baryield strength.

4. For the beam with a splice length of 120 in., cast with a cold joint, and subjected tomonotonically-increasing load, the reinforcing steel yielded prior to a splice failure,which occurred in the strain-hardening region of the stress-strain curve at a bar stress of72 ksi.5. For the two beams with a splice length of 120 in., cast with a cold joint, and subjected tocyclic loading, horizontal cracks with maximum widths of 30 and 35 mils (0.03 and0.03 5 in.) developed prior to splice failure, which occurred at bar stresses of 67 and 69ksi, respectively, values that equaled or exceeded the bar yield strength..

The following conclusions are based on the test results and analyses presented in this report.REDACTED VERSION

b-UUU I-age fo Ot 1 1141. The methods described in this report provide a viable means of simulating a crack in theplane of flexural reinforcement.

2. The cyclically load beams incorporating a cold joint to simulate crack in the plane of thereinforcement exhibited slightly reduced lap splice capacity compared to themonotonically loaded beams.3. In the presence of a simulated crack in the plane of the reinforcing bars, the lap-spliced No. 11 bars with a 79-in, splice length achieved bar stresses of 62 and 57 ksi.4. In the presence of a simulated crack in the plane of the reinforcing bars, the lap-spliced No. 11 bars with a 120-in, splice length achieved bar stresses greater than or equal to theyield strength, 67 ksi.REDACTED VERSION

I-age t/b 0? 1114.

References:

ACI Committee 318. (2011). Building Code Requirements for Structural Concrete andCommentary (ACT 318-11), American Concrete Institute, Farmington Hills, MI, 430 pp.ACT Committee 408. (2003). Bond and Development of Reinforcement, Bond and Development of Straight Reinforcing Bars in Tension (ACT 408R-03), American Concrete Institute, Farmington Hills, MI, 49 pp.ACT Committee 408. (2009). Guide for Lap Splice and Development Length of High Relative RibArea Reinforcing Bars in Tension and Commentary (ACT 408.3R-09), American ConcreteInstitute, Farmington Hills, MI, 12 pp.ASTM C3 1. (2010). Standard Practice for Making and Curing Concrete Test Specimens in theField (ASTM 3 1/C3 1 M- 10), ASTM International, West Conshohocken, PA, 6 pp.ASTM C78. (2010). Standard Test Method for Flexural Strength of Concrete (Using SimpleBeam with Third-Point Loading) (ASTM C78/C78M-10), ASTM International, WestConshohocken, PA, 4 pp.ASTM C138 (2012). Standard Test Method for Density (Unit Weight), Yield, and Air Content(Gravimetric) of Concrete (ASTM C 138/C 138M- 12), ASTM International, West Conshohocken, PA, 4 pp.ASTM C 143. (2010). Standard Test Method for Slump of Hydraulic-Cement Concrete (ASTMC 143/C 1 43M- 10Oa), ASTM International, West Conshohocken, PA, 4 pp.ASTM C 172. (2010). Standard Practice for Sampling Freshly Mixed Concrete (ASTMC 172/C 172M-10), ASTM International, West Conshohocken, PA, 3 pp.ASTM C496. (2011). Standard Test Method for Splitting Tensile Strength of Cylindrical Concrete Specimens (ASTM C496/C496M-1 1), ASTM International, West Conshohocken, PA,5 pp.ASTM C 1064. (2011). Standard Test Method for Temperature of Freshly Mixed Hydraulic-Cement Concrete (ASTM C 1064/C 1 064M- 11), ASTM International, West Conshohocken, PA, 3PP.Nawy, Edward G. (2003). Reinforced Concrete: A Fundamental

Approach, Fifth Edition,Pearson Education Inc., Upper Saddle River, N J, 821 pp.Nilson, A. H., Darwin, D., and Dolan, C. W., Design of Concrete Structures, 14th Ed., McGraw-Hill, New York, 2010, 795 pp.Seliem, H. M., Hosny, A., Rizkalla, S., Zia, P., Briggs, M., Miller, S., Darwin, D., Browning, J.,Glass, G. M., Hoyt, K., Donnelly, K., and Jirsa, J. 0. (2009). "Bond Characteristics of High-Strength ASTM A 1035 Steel Reinforcing Bars," ACI Structural

Journal, Vol. 106, No. 4, July-Aug., pp. 530-539.REDACTED VERSION bYWi-UUU-U5dJ-GW__L-UUUIb-UUU l-'age tb/ 0?" 1114Appendix A: Pilot Tests -Preliminary Study of the Effect of Simulated Cracks on LapSplice Strength of Reinforcing Bars using Beams with Single Splices1. Introduction This appendix presents the findings of a pilot study consisting of two lap splice beamtests that were performed to investigate how a test specimen with a preexisting crack parallel tothe plane of the reinforcement could be developed and tested. The test program described in thebody of this report was developed using lessons obtained in this pilot study.The two beams were cast simultaneously and tested monotonically to failure seven daysafter casting.

Because this project involves a larger number of physical simulations, the testing ofthese two beams is referenced throughout the report as Stage 1 of the project. Both beams hadmain flexural reinforcement consisting of three No. 11 bars, two of them continuous and one ofthem spliced at the center of the beam (Figure A. 1). The splice lengths in the two beams were 33in. and 79 in., respectively. All other dimensions and material properties were identical. Thebeam with a splice length of 33 in. will be referenced throughout this appendix as Beam Al1 andthe beam with a splice length of 79 in. will be referenced as Beam A2.The two beams were instrumented with strain gages placed on all bars at the edge of thesplice region (Figure A.2). Beam displacements and applied loads were monitored during thetests using displacement transducers and load cells.The following sections present brief descriptions of the beams, the test process, andoutline the major findings from the tests.2. Beam CastingCasting was performed in two separate stages. The first stage of the casting processconsisted of placing concrete over the full depth of the beam at the end sections and up to themid-height of the flexural reinforcement in the splice region (Figs. A. 1 and A.3). The firstconcrete was placed on May 3, 2012. The concrete surface at the location of the cold joint wasroughened and the beam was wet-cured for 24 hours (Figure A.4). Two layers of painters tapewere placed adjacent to the bars to simulate the effects of a preexisting crack parallel to the planeof the flexural reinforcement (Figure A.5). Concrete was placed above the cold joint on May 4,and the beams were subsequently moist cured for 7 days.REDACTED VERSION bbig-UUU-Ud.5J-tb..d-UUUlth-UUU I--'age /tbt Oi 11 1476" 3321I 2" 66"I12"18" 252" 18"LoadSupport Support(a)76" 79"18" 252" 18"Load Q Load5" 5"'5" 5" 5"5"5" 5"5" 5"5"5" I" " , " " ' " " _,Z .... ...132".12n12"SupportSupport(b)Figure A.1 -Reinforcing steel drawing. (a) Beam Al -specimen with a 33-in, splice length. (b)Beam A2 -specimen with a 79-in, splice lengthREDACTED VERSION bbYJ-UUU-UdJ3-LiLH-UUU1b-UUU I-'age 0t 1114Figure A.2 -Strain gages placed on bars at the edge of the splice region of Beam A2.REDACTED VERSION

r~age ftiu 0? 1114Figure A.3 -Beam A2 after first concrete placement was completed. REDACTED VERSION

I--'age (bi OT 1 114Figure A.4 -Beam Al1 being moist cured after the first placement was completed. REDACTED VERSION bbJL-UUU-(LbJ-LWFt-UUUltb-UUU I-sage (tb2 Otf 1 114Roughed concrete surfaceto provide betterconnection at the cold jointSide stirrup to simulateconfinement provided bysurrounding concrete in alarger structure Figure A.5 -Painters tape placed to simulate a preexisting crack at the plane of thereinforcement in Beam A2.REDACTED VERSION

t-UUU I-sage 0111143. Test apparatus and loading protocolThe two beams were tested using a four-point loading configuration. To facilitate inspection of the splice region during the test, the loads were applied in the downward direction (Figure A.6) so that the main flexural reinforcement would be located at the top of the beam. Thesplice region was located between the two supports (Figure A.7), in the central constant momentregion of the beam.In addition to strain gages, the beams were instrumented to measure displacement andload. The four load rods used in the test were instrumented to record load, and displacements were recorded using displacement transducers and dial gages for redundancy. Threedisplacement transducers were used to monitor the displacement at the center of the beam and ateach of the two load points (Figure A.6). Dial gages were mounted at a distance of 3 in. from theload points.Loads were applied with four hydraulic rams connected to a manual pump through adistribution system with two separate manifolds. The manifold system allowed adjustments inthe pressure of each ram separately and adjustment of the pressure in each pair of rams allowingfor loading in tandem. The force in each of the four load rods (Figure A.6) was monitored throughout the test and the pressure in the rams was constantly adjusted to maintain the force ineach of the rods approximately equal.REDACTED VERSION

-UUU I-age or 1114Figure A.6 -Test apparatus Figure A.7(a) -Splice region of Beam A2 prior to loadingREDACTED VERSION

I-'age (tb~ Ot 1 114Figure A.7(b) -East support of Beam A2 prior to loadingREDACTED VERSION

I-'age (bb OT1 1114The loading protocol consisted of monotonically-increasing load applied at both the endsof the beams. Loading was paused at increments in the total force of 10 kips (5-kip increments applied at each end of the beam) to monitor crack widths, mark crack locations, and record dialgage readings (Figure A.8). After all these quantities were recorded, loading resumed until thenext increment was completed. Given the potential for brittle failure and the large amount ofenergy stored in the beam, crack location, crack width, and dial gage readings were not recordedafter the total load exceeded 140 kips (forces at beam ends exceeded 70 kips). After this point,the load was increased steadily until the end of the test. Measurements from load anddisplacement sensors were recorded without interruption during the test.Figure A.8 -Marking cracks during test4. Material Properties The beams were tested on May 10, 2012, seven days after initial casting. On the day ofthe test the compressive strength of the concrete was 5090 psi in the body of the beam and 5150psi above the cold joint.A segment of the No. 11 bars used in the beams was tested in tension. The stress-strain curve for the No. 11 bar is shown in Figure A.9. To avoid damage, the extensometer wasREDACTED VERSION

i-'age tfb o071114removed at approximately 3% elongation; force was recorded until failure. As shown in thefigure, the No. 1 1 bar did not have a well-defined yield point. The yield stress calculated usingthe 0.2% offset method was 71 ksi, the proportional limit was approximately 67 ksi, and themeasured elastic modulus was 27,666 ksi. The tensile strength of the steel was 108 ksi.1009002% offset th500)4020 Yield (0.2% offset): 70.6 ksi20 I tUltimate: 108.2 ksi10 Elastic Modulus: 27666 ksi100 0.005 0.01 0.015 0.02 0.025 0.03StrainFigure A.9 -Measured stress-strain curve for the No. 11 bar used in the beams5. Test ResultsThe load-deflection curves for Beams 1 (33-in, splice) and 2 (79-in, splice) are shown inFigures A. 10 and A. 11, respectively. The displacement shown in both figures was calculated byadding the average displacement at the two load points and the displacement at the center of thebeam. The load shown in Figures A. 10 and A. 11 corresponds to the total load applied to thebeam. Based on the shape of the load-deflection curves shown in Figures A. 10 and A. 11, it isconcluded that a splice failure took place in Beam Al and a flexural failure occurred in BeamA2.For Beam Al (33-in, splice length), the peak total load recorded was 140 kips, at acorresponding total displacement of 1.14 in. (Figure A. 10). At a total load of 140 kips, the stressin the bars calculated using elastic cracked section theory was approximately 54 ksi. After thedisplacement exceeded 1.14 in., the total load dropped in a sudden manner to approximately 133REDACTED VERSION bb.i-UUU-dJ-UI:-UUU re-'age 01 1i14kips. If it is assumed that the tension force is carried in its entirety by the two continuous bars, atotal force of 133 kips corresponds to a calculated bar stress equal to the yield stress of 71 ksi(based on linear elastic cracked section theory). These calculations indicate that splice failureoccurred at a displacement of 1.14 in. and that the splice lost all its load carrying capacity in asudden manner. The total load tended to increase again at displacements greater than 1.6 in.,which is attributed to the effects of strain hardening in the two continuous bars.The load-deflection curve for Beam A2, with a splice length of 79-in., is presented inFigure A. 11. Loading was stopped when crushing of the concrete in the compression zone wasobserved in the constant moment region, in the areas adjacent to the two beam supports, at a totaldisplacement of approximately 2.5 in. Unlike the curve for Beam A l, there was no sudden dropin load associated with failure of the splice. In the case of Beam A2, a sharp decrease in the slopeof the load-deflection curve was observed at a total load of approximately 172 kips and totaldisplacement of approximately 1.4 in. The stress in the three bars calculated based on moment-curvature analysis at this load is approximately 67 ksi (Table A. 1), which corresponds to theobserved proportional limit of the measured stress-strain relationship of the steel (Figure A.9).The calculated steel stress indicates that the sharp decrease in the slope of the load-deflection curve at 172 kips was caused by yielding of the reinforcing steel, not by failure of the splice.After yielding began, the total load continued to increase with increasing displacement, as thereinforcing steel strain hardened. The maximum load prior to flexural failure was approximately 186 kips, which corresponds to a bar stress of 72 ksi in all three bars (Table A. 1). At a total loadof 186 kips, horizontal splitting cracks on the beam top surface were observed (described in moredetail below).After the tests we completed, the beams were autopsied to determine the actual cover onthe bars. For Beam Al, the top cover was 4 in., and side covers to the continuous bars were 3.5(North) and 3.75 in. (South). For Beam A2, the top cover was 4 in., and side covers to thecontinuous bars were 3.5 in. (North and South). (These values are reflected in the bar stresses inthe previous paragraph and summarized in Table A. 1)REDACTED VERSION

I-'age 1bSU OT 1 114160010100-J0 0.5 1 1.5 2 2.5 3 3.5 4Displacement, in.Figure A.10 -Total load vs. total deflection for Beam Al (33-in, splice length)20012kp _ 1 8 0.. .... ....140.-100 __06000.5 11.5 22.53Displacement, in.Figure A.11 -Total load vs. added deflection for Beam A2 (79-in, splice length)REDACTED VERSION

I.-sage /flU 0? 111"m4Loads, moments, and bar stresses for the beams were calculated assuming that loads andreactions acted along the longitudinal centerline of the beam. Reactions and moments werecalculated based on load cell readings and the weight of the loading assemblies. The self-weight of the beam was included in the calculations based on average beam dimensions and an assumedconcrete density of 150 pcf.The calculated moment, bar stress at splice failure, and calculated bar stress using thesplice strength equation developed by ACI Committee 408 (2003) are shown in Table A. 1. It isimportant to note that the splice strength expression developed by Committee 408 was calibrated on the basis of beams without preexisting cracks in the plane of the flexural reinforcement, andfor this reason are presented only as a reference. For Beam Al (with a 33-in, long splice), the barstress at splice failure calculated based on a moment-curvature analysis was approximately 54ksi. The calculated splice strength using the expression developed by ACI Committee 408 (ACI408R) was 70 ksi. For Beam A2 (with a 79-in, long splice), the calculated bar stress at flexuralfailure was approximately 72 ksi, while the calculated splice strength using the ACI 408expression was 140 ksi.Table A.1 -Bar stresses at splice failureI 33-in. I splice failure j 140 I 35I54I70I[79 in. flex ural 186 [ 472 72 [140]The strain in the No. 11 bars was measured using strain gages located 2 in. outside thesplice region (Figure A.2). The relationships between measured strain and total load are shown inFigures A. 12 and A. 13 for beams 1 and 2, respectively. As shown in Figure A. 12, the strain in the spliced bars (East-center and West-center gages) of Beam Al increased to a maximum of 1750 and 1700 microstrain, respectively, andthen dropped in a sudden manner. The maximum strain in the spliced bar was recorded at a totalload of approximately 130 kips and corresponds to a bar stress of approximately 50 ksi, which isvery close to the failure value of 54 ksi inferred on the basis of moment-curvature analysis(Table A. 1). Strain readings from the east-center gage on the spliced bar show that the strain1REDACTED VERSION I-'age 111 0? 1114dropped from approximately 1700 to approximately 1300 microstrain at a total load of 130 kips,corresponding to a sudden reduction in capacity of approximately 25%. When the total loadreached 140 kips, the strain in the east-center gage dropped suddenly to almost zero. Strainreadings from the east continuous bar (East-Side

1) show a sudden increase from 2100microstrain to more than 2500 microstrain at the failure total load of 140 kips. The strain gagereadings indicate that failure of the splice led to a rapid decrease in the stress in the spliced bars,and that the tension force that was lost due to failure of the splice was transferred to thecontinuous bars, causing yielding of the continuous bars at a total force of 140 kips.For Beam A2 (79-in, long splice),

the recorded strains show a plateau (Figure A. 13) dueto exceeding the limiting strain allowed by the gain in the data acquisition system.25002000* 1500.N10005000-East-center -East-Side I-East-side 2---- West-center 050 100 150Total load, kipsFigure A.12 -Measured strain in the reinforcing bars vs. total load for Beam Al (33-in, splicelength). (Note: The beam was oriented in an east-west direction; "center" identifies strain gageson the spliced bars and "side" means strain gauges on the continuous bars)REDACTED VERSION 'U~ Page (/j' or 1114L.0L..U100 150Total load, kips 200length). (Note: The beam Was oriented in an east-west drcin cne"i~t~on the spliced bars and "side" meaection; ,,...,, sstrain gages6. Bemcr c ...... , tuges on" the continuous bars)Figures A. 14 through A. 18 are photographs taken after the Conclusion of the two tests.For B earn Al (33-in. splice length), splitting cracks Were observed on the top surface betweenthe vertical edges of the cold joint (Figures A. 14 and A. 15). The cracks Were approximate 1y 1/4in. wide, as shown in Figure A. 16. Splitting cracks above the splice Were also noted in Beam A2(79-jn. splice length) (Figures A. 17 and A. 18), although they were much narrow~er than thoseThe crack patterns for both beams show that the side stirrups Were effective in keepingthe COver in place, even after failure of the splice for Beam Al. In the case of Beam Al, thecracks Were wider, which is COnsistent with the sudden drop in bar foc that ocre tslcfailure. For Beam A2, the cracks Were much narrower, and it is apparent that the splice was ableto sustain the same bar force as the continuous bars at displacement ag nuht asflexural failure of the beam .nt la g e ou h o c usREDACTED VERSION

t-UUU r-age i Ot o 1114Splitting crackabove spliceEnd of splice regionLFigure A.14 -Splitting crack at the top of the splice region for Beam Al (33-in, splice length).REDACTED VERSION bbI,-UUU-L

,,-U51"-UUUll-UUU I-'age (140OTll114 Horizontal cold joint in theplane of the reinforcement Figure A.15 -Crack pattern in the splice region for Beam AlI (33-in, splice length).REDACTED VERSION

1-'age ((b 0O" 1114Figure A.16 -Splitting crack at the top of the splice region of Beam Al (33 in. splice length).REDACTED VERSION

I-sage 0O" 1114End o cFigure A.17 -Splitting crack at the top of the splice region of Beam A2 (79 in. splice length).REDACTED VERSION

F-'age f[f 0?' 1114Splitting crackabove spliceHorizontal cold jointin the plane ofreinforcing steelFigure A.18 -Crack pattern in the splice region of Beam A2 (79-in, splice length)REDACTED VERSION}}