Cast-in-Place Concrete Connections for Precast Deck Systems (2011)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

392 Chapter 12 Longitudinal Joint Details for Accelerated Bridge Construction: Fatigue Evaluation 12.0 Introduction Chapter 9 presented results of a preliminary study that assessed potential longitudinal and transverse joint connection details for CIP joints between DBTs or full-depth precast panel to panel connections. The investigation was conducted in two phases. During Phase I, lapped headed reinforcement and two kinds of lapped U-bar reinforcement (i.e., deformed wire (DWR) and stainless steel (S)) were tested to determine the most promising connector. Monolithic reinforced concrete flexural specimens were constructed for each of the three types of connections to simulate the forces that would be experienced in a longitudinal deck joint, and monolithic reinforced concrete tension specimens were fabricated for each of the three types of connections to simulate the forces that would be experienced in a transverse joint over an interior pier. Based on that study, the DWR U-bar detail was deemed the most promising in terms of overall performance, constructability and cost. In the second phase of experiments, six specimens with the DWR U- bar detail were tested, three in flexure and three in tension, to investigate effects of variables including overlap lengths, rebar spacings, and concrete strengths. A DWR U-bar detail with a 6 in. overlap length and 4.5 in spacing was selected for additional testing to further investigate replacing the current welded steel connector detail. This chapter describes the test program and presents results of this additional testing to investigate its feasibility for the longitudinal joint, and Chapter 13 investigates the U-bar detail for the transverse joint. Four pairs of large-scale slabs were fabricated to investigate the flexure and flexure-shear behavior of the longitudinal joints fabricated using the DWR U-bar detail with a 6 in. overlap length. Two closure pour (CP) materials were investigated based on the most promising overnight and 7-day cure CP materials summarized in Chapter 10: SET® 45 HW without extension for the overnight cure and HPC Mix1 for the 7- day cure material. In the longitudinal joint study, two of the specimens incorporating CP overnight-cure materials were fabricated with grout SET® 45 HW with 60% extension. A total of eight tests were conducted, which required the reuse of the four pairs of large-scale slabs. Each slab had U-bars and shear keys along two opposite edges, so after completing tests on the first joint, the panels were separated and another joint was assembled by using the other two edges of the slabs to create a second specimen. The eight tests consisted of static flexure (SF), static shear (SS), fatigue flexure (FF) and fatigue shear (FS) tests conducted on longitudinal joints with two different types of closure pour (CP) materials (overnight and 7-day cure materials). The loads applied to the joints were based on the analytical parametric study summarized in Chapter 10, which was conducted to provide a database of maximum forces used to determine the service live and fatigue loading demand for the slab tests described in this and the subsequent chapter. The results of the static and fatigue tests under four-point pure-flexural loading, as well as three-point flexural-shear loading, are described in this chapter. The test results were evaluated based on flexural capacity, curvature

393 behavior, cracking, deflection and steel strain. Based on these test results, the U-bar detail was deemed to be a viable connection system for the longitudinal joint. 12.1. Experimental Program 12.1.1. Slab Dimension A total of eight slabs with the same dimensions were fabricated for the static and fatigue testing, with two different closure pour materials used in the longitudinal joint. Each specimen consisted of two panels connected by a longitudinal joint as shown in Figure 12.1.1. Each panel was 72 in. wide, 64 in. long and 6-¼ in. deep. The female-to-female shear key was provided at the vertical edge of both ends in the specimen length direction. This allowed each slab to be used for two tests. Figure 12.1.1: Dimensions of longitudinal joint specimen 12.1.2. Reinforcement Layout and Strain Gage Instrumentation Figure 12.1.2 displays the reinforcement layout used in the longitudinal joint specimen. There were four layers of reinforcement in each panel along the specimen depth direction with 2 in. cover at the top and 1 4'' 5.5'' 0.625'' 2.5'' 4'' 5.5'' 0.625'' 2.5'' Shear Key Detail 64'' 64''6 .2 5 '' 72 '' Panel 1 Panel 2 See "Shear Key Detail" Centerline of Joint

394 in. cover at the bottom. The straight bars simulated longitudinal reinforcement while the U-bars simulated the transverse reinforcement in the bridge deck. The reinforcement details in the specimen were as follows: #5 straight bar spaced at 6 in. at the bottom along the slab length direction; #4 straight bar spaced at 12 in. at the top along the slab length direction. Note that the longitudinal reinforcement was located within the U-bars to enable the largest diameter bend as possible for the U-bar while still meeting concrete cover requirements. The #5 U- bars projected out of the panel to splice with the U-bars in the adjacent panel in the longitudinal joint. The spacing of the U-bars was 4.5 in. and the overlap length (the distance between bearing surfaces of adjacent U-bars) was 6 in. The interior diameter of bend of the U-bar was 3db. See "Joint Reinforcement Detail" Centerline of Joint 4. 5' ' ( T yp .) #5 U bar spacing 4.5'' (Typ.) #4 bar spacing 12'' (Typ.) #5 bar spacing 6'' (Typ.) #4 lacer bar (Typ.) 2'' 1'' 6'' 3 1 8'' (5d) Figure 12.1.2: Reinforcement layout in longitudinal joint specimen The U-bars around the joint zone were instrumented with strain gages to gain an understanding of the behavior of the slab connected by the longitudinal joint. Figure 12.1.3 depicts the strain gage layout in the slab for the four-point pure-flexure test and the three-point flexure-shear test.

395 L1 L5 L8 R1 R4 R7 a (a) Bars instrumented with strain gages (marked with red) 2"7" L1-7 (R1-7) L1-9 (R1-9) g 6" 4" L5-6 (R4-6) L5-10 (R4-10) (b) Strain gage locations for Bar L1 and R1 (c) Strain gage locations for Bar L5 and R4 4"2" L8-0 (R7-0) L8-2 (R7-2) L8-6 (R7-6) strain gauge in the center (a-2) strain gauge at the end (a-1) (d) Strain gage locations for Bar L8 and R7 (e) Strain gage locations for Lacer Bar a Figure 12.1.3: Strain gage layout There were six U-bars with installed strain gages. The U-bars were numbered from the edge of the slab (number 1) to the middle (number 8) along the slab width direction. Strain gages were labeled to identify whether it was attached to a U-bar in the left (L) or right (R) slab, the U-bar designator (1-8), and the distance from the outside bend of the bar. For example, strain gage “L1-7” meant the strain gage on U-bar #1 of the left slab, located 7 in. away from the outside bend of the bar. Two strain gages were also placed at the end and middle of the lacer bar, labeled as a-1 and a-2 respectively. 12.1.3. Panel Fabrication The concrete panels were fabricated locally at Ross Prestressed Concrete Inc. in Knoxville, TN. Figure 12.1.4- a and c show the panel reinforcement before placement of the concrete. The two ends of the wood form, in

396 the length direction, were slotted at a spacing of 6 in. to fix the U-bars in place. Foam wedges were used to form the configuration of the shear key at the vertical edge of the panel. The design concrete compressive strength at 28 days was 7000 psi. Companion concrete cylinders were cast with the panels (Figure 12.1.4-b and d). (a) Before pouring (FS & SS Specimens) (b) After pouring (FS & SS Specimens) (c) Before pouring (FF & SF Specimens) (d) After pouring (FF & SF Specimens) Figure 12.1.4: Panel fabrication 12.1.4. Joint Surface Preparation The surfaces of the shear key were sandblasted to prepare the joint for the closure pour. The purpose of the surface preparation was to remove all contaminants that could interfere with adhesion and to develop a surface roughness to promote mechanical bond between the CP material and panel concrete. After the

397 removal of the deteriorated concrete, proper preparation should provide a dry, clean and sound surface offering a sufficient profile to achieve adequate adhesion. There were many methods of surface preparation considered such as chemical cleaning, mechanical cleaning and blast cleaning. Sandblasting was chosen, which uses compressed air to eject a high speed stream of sand onto the surface which needed to be prepared. This method is very effective to process the surface of precast members under industrial conditions. Black Beauty 2050 sand was chosen for sandblasting to prepare the surface in this study. The profiles of the surface before and after sandblasting are shown in Figure 12.1.5. (a) Before sandblasting (b) After sandblasting Figure 12.1.5: Profile of joint surface 12.1.5. Closure Pour (CP) Materials The longitudinal joint was filled with closure pour (CP) material to complete the connection which simulated the longitudinal joint connection at the interface of the top flange of adjacent DBT girders or full- depth precast panel to panel connections, considered to be the structural element of the bridge deck. To facilitate accelerated bridge construction, it is important for the selected CP material to reach its design compressive strength in a relatively short period of time. In this study, it was decided to use two primary CP materials, SET® 45 HW for overnight cure and HPC Mix1 for 7 day cure, selected in Chapter 11. The grout SET® 45 HW used in the longitudinal joint study was investigated both without extension in two joints and with 60% extension in two joints for comparison. The uniform-sized sound 0.25-0.5 in. round pea gravel used to extend the grouts was tested with 10% HCL to confirm that it was not calcareous. 12.1.6. Testing Plan and Setup

398 A total of eight slab specimens were made. Each slab specimen consisted of two concrete panels connected with an overlapping U-bar detail and one of the selected closure pour (CP) materials. During the test setup, each panel was placed on a steel I-beam, which was leveled to ensure that the two panels were on the same plane. At the joint zone, the two panels were positioned to satisfy the 6 in. overlap length and the 4.5 in. spacing of the U-bar (Figure 12.1.6-a). A wood form was use at the bottom and at both ends of the joint to prevent leakage when casting the CP joint. After the CP material was placed, the longitudinal joint slab specimen consisting of 2 panels connected by the joint was ready for testing (Figure 12.1.6-b). Because each panel had U-bars and shear keys along two edges, each set of two panels was used to fabricate two test specimens. After completion of testing the first joint, the panels were separated, and then another joint was reassembled by the other two edges to create the second test specimen. (a) Before grouting (b) After grouting Figure 12.1.6: Longitudinal joint specimen before and after grouting Eight slab specimens (four for each of two selected CP materials) were tested under different parameters: 1) static flexure (SF) test; 2) static shear (SS) test; 3) fatigue flexure (FF) test, and 4) fatigue shear (FS) test. Table 12.1.1 presents the loading matrix for the eight specimens. Figure 12.1.7 shows the testing setup and the linear motion transducer (LMT) instrumentation layout for each test. All slab specimens were simply supported with a 72 in. span and the joint zone located in the center of the span. A neoprene pad, with two layers of plastic sheets placed between the pad and slab bottom, was used at one end; only the neoprene pad was used at the other end, as shown in Figure 12.1.8. A 10 in. by 20 in. neoprene pad and steel plate were used to simulate the truck tire contact area and the pressure loading. The LMT’s were used to measure the specimen deflection, settlement and curvature. Four LMT’s (Nos. 4-7 in Figure 12.1.7) were

399 used to measure the vertical deflection along the joint. LMT’s 4, 6 and 7 were placed along the centerline of the joint while LMT 5 was placed at the panel edge off the interface of the joint. In this way, the relative deflection between the side of the joint interface and joint center could be measured. LMT’s 1-3 and 8-10 were used to measure potential support settlement. Two LMT’s, placed horizontally across the top and bottom of the joint, were used to measure the average curvature of the joint zone. DEMEC points were used to measure the width of crack opening at the joint interface with a DEMEC mechanical strain gage. The DEMEC points were glued onto the slab, as shown in Figure 12.1.9. Table 12.1.1: Slab specimen loading matrix Overnight Cure 7 Day Cure Flexure Flexure-Shear Flexure Flexure-Shear Static Fatigue Static Fatigue Static Fatigue Static Fatigue SET® 45 HW extended SET® 45 HW SET® 45 HW SET® 45 HW extended HPC Mix 1

400 AA A - A P P 1 2 3 9 10 4 7 5 6 8 LMT 1, 2 and 3LMT 8, 9 and 10 10 in. 20 i n. LMT 4, 6 and 7 LMT 5 12 in.12 in. 36 in. 36 in. 4 in . 32 i n. 32 i n. 4 in . LMTs for Curvature AA A - A P 20 in . LMT 1, 2 and 3LMT 8, 9 and 10 LMT 4, 6 and 7 LMT 5 1 2 3 8 9 10 4 7 6 5 LMTs for Curvature 12 in. 4 in . 32 in . 32 in . 4 in . 10 in. 36 in. 36 in. (a) Static Flexure (SF) test (b) Static Shear (SS) test AA A - A 2P 1 2 3 8 9 10 4 7 5 6 LMT 1, 2 and 3LMT 8, 9 and 10 LMT 5 LMT 4, 6 and 7LMT for Curvature 12 in.12 in. 36 in. 36 in. 4 in . 32 in . 32 in . 4 in . 10 in. 20 in . 26 in. 42 in. 42 in. 26 in. AA A - A 36 in. 36 in. P1 P2 12 in.12 in. LMT 1, 2 and 3LMT 8, 9 and 10 LMT 4, 6 and 7 LMT 5 1 2 3 8 9 10 4 7 5 6 LMTs for Curvature 10 in. 20 in . 4 in . 32 in . 32 in . 4 in . (c) Fatigue Flexure (FF) test (d) Fatigue Shear (FS) test Figure 12.1.7: Longitudinal joint specimen test setup

401 Figure 12.1.8: Simple support boundary condition at edge of longitudinal joint specimen AA A - A 36 36 DEMEC 1-1 DEMEC 2-2 DEMEC 1-1 DEMEC 2-2 (a) DEMEC points for SF, SS, FS tests

402 AA A - A 36 36 2P DEMEC 1-1 DEMEC 2-2 DEMEC 1'-1' DEMEC 2'-2' DEMEC 1'-1' DEMEC 2'-2' BB B- B DEMEC 1-1 DEMEC 2-2 (b) DEMEC points for FF Tests Figure 12.1.9: DEMEC points

403 The static flexural (SF) test specimens were loaded with two equal loads spaced at 12 in. about the center of the span using Material Test System (MTS) actuators until the specimen failed. The joint zone experienced the maximum constant moment without shear. The static shear (SS) test specimens were loaded with one load located at 12 in. off the center of the span until the specimen failed. The joint zone experienced a combination of moment and shear. Similar to the SF test specimens, the fatigue flexural (FF) test specimens were loaded with two equal loads spaced at 12 in. about the center of the span. Figure 12.1.10 shows the apparatus used to apply the fatigue forces to the joint zone of the specimen. Figure 12.1.10: Test setup for applying fatigue forces One side of the swivel rod end was screwed to the actuator tightly while the other side was bolted to the load spreader (i.e., spread tube) at midspan by four steel rods. The spread tube was soldered to two steel hinges, which were located 12 in. away from the middle of the spread tube. The other end of each steel hinge was soldered to the 10 in. by 20 in. steel plate. The use of steel hinges between the spread tube and the steel plates was intended to eliminate any extra moment from being applied to the slab specimen produced by the bending of the spread tube. The steel plate and neoprene pad at the bottom of the slab were bolted to the correspondent top steel plate and neoprene pad through the slab by 4 steel rods, which applied the fatigue forces to the slab. The boundary condition was provided by the steel girder below the slab and by the steel girder above the slab. The two steel girders at each support-end (one below the slab and the other above the slab) were connected by bolts. The steel girder below the slab was fixed to the strong floor as shown in Figure 12.1.10. A neoprene pad, with two layers of plastic sheets placed between the pad and slab surface, was used at one end above and below the slab; only the neoprene pad was used

404 at the other end above and below the slab. The two loads spaced at 12 in. about the center of the span in the FS specimens, designated P1 and P2, were applied out-of-phase on each side of the joint during the fatigue test. For example, when “P1” reached the maximum force, “P2” was zero. The joint zone experienced fatigue shear by applying load reversals at the single point load (i.e., reversing the loading directions). The compressive strength of concrete panel (f’c) and the compressive strength of grouted joint (f’cj) at the time of testing for each specimen are given in Table 12.1.2. The specimens were designated based on the type of test (e.g., static flexure (SF), fatigue flexure (FF), static shear (SS) and fatigue shear (FS)) followed by the type of cure material used (e.g., overnight cure (O) and 7-day cure (7)). For example, “SS-O” denoted the static shear longitudinal joint connection test with the overnight cure closure pour (CP) material. The joint concrete compressive strength used in the FE models in Chapter 9 to develop a database of maximum forces was 4000 psi.

405 Table 12.1.2: Compressive strength of concrete panel and grouted joint Panel (psi) Joint (psi) Specimen Start of Test End of Test Start of Test End of Test SS-O 11512* 7586 SS-7 11512* 8740 FS-O 10687 11512* 6321 6572 FS-7 11512* 11512* 7861 9417** SF-O 12441 5939 SF-7 12441 6966 FF-O 11711 11632 4592 5345 FF-7 11035 11711 10796*** 12361 * The panels for FS-O, FS-7, SS-O, and SS-7 were fabricated with the same batch of concrete and were tested or the test was finished more than 120 days after the panels were fabricated. The 148-day concrete compressive strength of 11,512 psi is the reported strength. ** The test for FS-7 started 5 days after the joint was cast, and finished 13 days after the joint was cast. The strength reported here is the 21-day joint strength. *** The FF-7 test started at the age of 22 days, and the joint strength reported in the table is the 8-day strength. 12.1.7. Fatigue Loading Determination FE models of the test specimens (Figure 12.1.11) were developed to determine the loadings to be used in the fatigue tests in order to produce the maximum moment or the maximum shear in the joint zone corresponding to the results from the previous parametric studies summarized in Chapter 10. Figure 12.1.11-b shows the moment distribution for the model loaded at one pad to produce an average moment along the joint of 10.0 kip-ft/ft. Figure 12.1.11-c shows the shear distribution and average shear for the same loading. These plots indicate that the moment was not expected to vary significantly from the average or nominal moment as compared to the shear which varied significantly along the joint length.

406 (a) Static shear test (FE Model) (b) Moment distribution along Joint (c) Distribution and average vertical shear along joint Figure 12.1.11: FE model for load determination

407 For the FF specimen, static loading was applied with the two patch loads on either side of the joint in several increments up to 44.6 kips (22.3 kips on each pad) in order to produce the maximum positive moment of 7.92 kip-ft per unit length (Section 10.1.5.8) in the joint to crack the joint (see Section 10.3, bullet point 9, for a summary of the sources of these applied moments). After unloading to zero, a negative static load of -12.0 kips (-6.0 kips on each pad), corresponding to a negative moment of -2.15 kip-ft per unit length (as determined to crack the joint in negative bending) was applied and unloaded to zero. During the fatigue test, the applied load was cycled between 11.4 kips (5.7 kips on each pad) corresponding to a positive moment of 1.99 kip-ft per unit length and -2.0 kips (-1.0 kips on each pad) corresponding to a negative moment of -0.35 kip-ft per unit length for a total of 2 million cycles at a frequency of 4Hz. At the end of 0.5, 1.0, 1.5, and 2.0 million cycles, an interim static loading test was conducted. During each of these static tests, the static loading was applied in several increments up to 26.2 kips (13.1 kips on each pad) corresponding to a positive moment of 4.55 kip-ft per unit length after cracking. After unloading to zero, a negative static load of -8.0 kips (-4.0 kips on each pad) corresponding to a negative moment of -1.40 kip-ft per unit length after cracking was applied and unloaded to zero. Finally, the slab specimen was loaded to failure. For the FS test, fatigue loads “P1” and “P2” were applied by the two MTS rams having the same frequency but out-of-phase, as discussed in Section 12.1.6. Before applying cyclic loading, static loading was applied through P1 in several increments up to 49.9 kips in order to produce the maximum shear of 6.091 kips per unit length in the joint to crack the joint. After unloading to zero, a static load of -49.9 kips was applied through P2, corresponding to a negative shear of -6.091 kips per unit length, then the specimen was unloaded to zero. Figure 12.1.12 shows the first few cycles of the fatigue loading history for the FS specimen. As discussed earlier, fatigue loads “P1” and “P2” were applied by the two MTS actuators at the same frequency but out- of-phase. The slab was subjected to fatigue loading with the resultant magnitude of “P1+P2” as shown in Figure 12.1.12. The peak P1 was 25.0 kips, corresponding to a positive shear of 2.84 kips per unit length, which was the combination of the fatigue shear of 2.34 kips/ft plus the camber leveling shear of 0.5 kips/ft. The “Average” value of “P1+P2” of 4.4 kips provided the camber leveling shear of 0.5 kips/ft at the middle of the joint zone all the time. Consequently, the maximum value of P2 was equal to 4.4 kips × 2 – P1= -16.2 kips. An interim static loading test (applying “P1” and “P2” separately) was conducted at the end of 0.5, 1.0, 1.5, and 2.0 million cycles. During each of these static tests, the static loading P1 was applied in several increments up to 46.9 kips corresponding to a positive shear of 5.34 kips per unit length after cracking. After unloading to zero, a static load P2 of -46.9 kips corresponding to a negative shear of -5.34 kips per unit length after cracking was applied and unloaded to zero. The specimen was loaded to failure after the fatigue cycles.

408 Figure 12.1.12: First two cycles of fatigue shear (FS) loading applied 12.1.8. Moment Capacity and Curvature Figure 12.1.13 shows the curvature versus the number of fatigue cycle curves (C-N) for the fatigue tests. The curvature represents the average curvature of the joint zone under a specific loading and is plotted relative to the number of fatigue cycles completed. For example, the curve labeled with “M=4.0 k-ft/ft” in Figure 12.1.13-a represents the change in the curvature of the joint zone for the FF-O specimen with respect to the numbers of fatigue cycles completed, where the curvature was measured at the loading level corresponding to a moment of 4.0 kip-ft/ft during each of the interim static load tests. As shown in Figure 12.1.13, the curvature increased with increasing joint moment for all specimens. There were problems with the initial curvature measurements for the FS-7 (under P1 loading) specimen before fatigue cycling, so the initial results for this specimen are not included in Figure 12.1.13-e. Comparing the results among the different joint moment levels, the impact of fatigue on the curvature was observed to be similar for all loading levels in each of the specimens. It appeared that the fatigue loading had little effect on the curvature for the FF-O and FS-O (under P1 or P2 loading) specimens while initial increases in curvature were observed for the FF-7 and FS-7 (under P2 loading) specimens during the first set of 0.5- million cycles, after which the damage accumulations due to fatigue ceased and the curvatures stabilized for the last 1.5-million cycles. Because the initial data was missing for the FS-7 (under P1 loading) specimen, no conclusions can be made with respect to its fatigue behavior relative to its initial performance. In general, there was no significant influence of fatigue cycles on the curvature for joints fabricated with the overnight CP material, and no significant influence after the first 0.5-million cycles for the 7-day cure CP material. -20 -10 0 10 20 30 0 1/16 1/8 3/16 1/4 5/16 3/8 7/16 1/2 Time (sec) Lo ad (k ip P1 P2 P1+P2 Average

409 (a) FF-O Specimen (b) FF-7 Specimen 0 100 200 300 400 0.0 0.5 1.0 1.5 2.0 Number of Fatigue Cycles (million) C ur va tu re ( 10 ^- 6 ra d/ in ) M=4.0 kips-ft/ft M=3.5 kips-ft/ft M=3.0 kips-ft/ft M=2.5 kips-ft/ft M=2.0 kips-ft/ft 0 100 200 300 400 0.0 0.5 1.0 1.5 2.0 Number of Fatigue Cycles (million) C ur va tu re ( 10 ^- 6 ra d/ in ) M=4.0 kips-ft/ft M=3.5 kips-ft/ft M=3.0 kips-ft/ft M=2.5 kips-ft/ft M=2.0 kips-ft/ft

410 (c) FS-O Specimen under P1 (d) FS-O Specimen under P2 0 100 200 300 400 0.0 0.5 1.0 1.5 2.0 Number of Fatigue Cycles (million) C ur va tu re ( 10 ^- 6 ra d/ in ) M=7.5 kips-ft/ft M=6.5 kips-ft/ft M=5.5 kips-ft/ft M=4.5 kips-ft/ft M=3.5 kips-ft/ft M=2.5 kips-ft/ft 0 100 200 300 400 0.0 0.5 1.0 1.5 2.0 Number of Fatigue Cycles (million) C ur va tu re ( 10 ^- 6 ra d/ in ) M=7.5 kips-ft/ft M=6.5 kips-ft/ft M=5.5 kips-ft/ft M=4.5 kips-ft/ft M=3.5 kips-ft/ft M=2.5 kips-ft/ft

411 (e) FS-7 Specimen under P1 (f) FS-7 Specimen under P2 Figure 12.1.13: C-N curves for flexure and flexure-shear specimens 0 100 200 300 400 0.0 0.5 1.0 1.5 2.0 Number of Fatigue Cycles (million) C ur va tu re ( 10 ^- 6 ra d/ in ) M=7.5 kips-ft/ft M=6.5 kips-ft/ft M=5.5 kips-ft/ft M=4.5 kips-ft/ft M=3.5 kips-ft/ft M=2.5 kips-ft/ft 0 100 200 300 400 0.0 0.5 1.0 1.5 2.0 Number of Fatigue Cycles (million) C ur va tu re ( 10 ^- 6 ra d/ in ) M=7.5 kips-ft/ft M=6.5 kips-ft/ft M=5.5 kips-ft/ft M=4.5 kips-ft/ft M=3.5 kips-ft/ft M=2.5 kips-ft/ft

412 The test results for the moment capacity are compared with the calculated values in Table 12.1.3. The strengths shown in the table for the fatigue tests, FS and FF, were the cylinder strengths measured at the end of the tests, and Table 12.1.2 presents detailed information about the compressive strengths. The service live load was the maximum positive calculated moment after cracking of 4.546 kip-ft/ft reported in Section 10.1.5.9 for the flexure tests; for the flexure-shear tests the service live load was the corresponding moment (CM) that occurred with the maximum shear, which was 3.372 kip-ft/ft after cracking reported in Section 10.1.5.9. Table 12.1.3: Measured and calculated loading capacity Specimen Measured Calculated Panel Compressive Strength (f’c) (psi) Joint Compressive Strength (f’cj) (psi) Failure Load (kip- ft/ft) Service Live Load (kip- ft/ft) Ultimate Load (kip- ft/ft) SS-O 11512 7586 19.34 3.37 19.31 FS-O 11512 6572 22.18 3.37 19.31 SS-7 11512 8740 25.44 3.37 19.31 FS-7 11512 9417 23.30 3.37 19.31 SF-O 12441 5939 24.21 4.55 19.31 FF-O 11632 5345 19.39 4.55 19.31 SF-7 12441 6966 31.5* 4.55 19.31 FF-7 11711 12361 31.5* 4.55 19.31 * SF-7 and FF-7 specimens were beyond the MTS capacity (31.5 kip-ft/ft) and couldn’t be failed by MTS. The measured ultimate capacities of all of the specimens obtained following the service fatigue loading cycles exceeded their calculated capacities. The joints with the overnight cure materials had lower capacities than those with the 7-day cure, due to the lower strength of the joint material. Figure 12.1.14-a compares the moment-curvature curves obtained for the specimens tested in flexure under static loading (FS) with those tested in flexure following 2.0 million cycles of fatigue loading (FF).

413 Figure 12.1.14-b compares the moment-curvature curves obtained for the specimens tested in flexure- shear under static loading (SS) with those tested in flexure-shear following 2.0 million cycles of fatigue loading (SF). The vertical axis labeled “Moment/Joint Length” represents the applied moment assumed to be uniformly distributed along the joint (i.e., applied moment divided by the length of the joint). Both the SF and SS specimens were loaded with applied moments assuming uncracked sections while the FF and FS specimens were loaded with applied moments assuming cracked sections after 2.0 million fatigue cycles. As a result, the slope of the curves (i.e., stiffnesses of the specimens) for SF and SS were steeper (larger) than those of the FF and FS specimens at the beginning of the plots. After the applied moment exceeded the cracking moment level, the slopes of the two curves were similar, indicating that the fatigue cycles had no significant effect on the curvature development, as discussed earlier. In Figure 12.1.14-a, the strengths of SF-7 and FF-7 specimens were beyond the MTS capacity (31.5 kip-ft/ft) and as a consequence could not be tested to failure. The maximum curvature for FF-7 cannot be reported because the LMT’s were removed during the test. At service loading, the SF-O and FF-O specimens had about the same curvature, while FF-7 had more curvature than SF-7. FF-O had a similar failure load and curvature as calculated, while SF-O, SF-7 and FF-7 had higher failure loads and curvatures than calculated. When comparing overnight cure and 7-day cure CP materials, SF-7 had higher failure load and curvature than SF-O, and FF-7 performed better than FF-O. This was because 7-day cure material developed higher strength than the overnight cure material in the tests, as shown in Table 12.1.2. In Figure 12.1.14-b, SS-7 had a higher failure load and more curvature than FS-7, so the fatigue cycles had some influence on the failure load of the joint with the 7-day cure CP material. The failure load for FS-O was higher than that of SS-O. The maximum curvature for SS-O and FS-O cannot be reported, because the LMT’s were removed during the tests. As shown in Figure 12.1.14-b, for the range of loading for which curvature was measured, SS-O and FS-O developed almost the same curves after the service load. The influence of fatigue cycles on the joint with the overnight cure CP material was not significant. At service loading, the SS- O and FS-O specimens had essentially the same curvature, and it was the same for the SS-7 and FS-7 specimens. When comparing with the calculation, SS-7 and FS-7 had higher failure loads and curvatures than the calculated, while SS-O and FS-O had similar failure loads as the calculated. Comparing between different CP materials, the joints with the 7-day cure CP material developed higher strength capacity and more curvature, due to the higher strength developed by the 7-day cure material.

414 (a) Flexure tests (b) Flexure-Shear tests Figure 12.1.14: Moment-curvature curves for flexure and flexure-shear specimens 0 5 10 15 20 25 30 35 0 1000 2000 3000 4000 Curvature (10^-6 rad/in) M om en t/J oi nt L en gt h (k ip -f t/f t SF-7 FF-O FF-7Calculation MTS Capacity Measured Failure Load-SF-O Measured Failure Load-FF-O Service Live Load SF-O 0 5 10 15 20 25 30 35 0 1000 2000 3000 4000 Curvature (10^-6 rad/in) M om en t/J oi nt L en gt h (k ip -f t/f t SS-7 FS-7 Service Live Load Calculation Measured Failure Load-SS-7 Measured Failure Load-FS-7 Measured Failure Load-FS-O Measured Failure Load-SS-O FS-O SS-O

415 12.1.9. Deflection Development Figure 12.1.15-a and b show a comparison of the load-deflection curves of the fatigue specimens after 2 million cycles and the specimens subjected to static loading without fatigue cycles, for the flexure and flexure-shear tests, respectively. The vertical axis labeled “Load/Joint Length” represents the nominal distributed load along the joint, which was the applied load, P, divided by the width of the specimen (i.e., length of the joint). Note that the load P was the load applied to one loading pad as shown in Figure 12.1.7. The LMT’s were removed from the specimen for their protection before the specimens reached their maximum capacities. a. Flexure tests 0 3 6 9 12 15 18 0.0 0.2 0.4 0.6 0.8 1.0 Deflection (in) L oa d/ Jo in t L en gt h (k ip /f t Calculation SF-O SF-7 FF-O FF-7 Service Live Load

416 b. Flexure-Shear tests Figure 12.1.15: Load-deflection curves of flexure and flexure-shear specimens The calculated response in the plots represents the expected load-deflection curve before cracking, after cracking to reinforcement yield, and the stage of plastic hinge development at midspan after reinforcement yielding. Similar to Figure 12.1.14, the slopes of the SF and SS curves were steeper than the slopes of the curves for the FF and FS specimens from the initial loading until the cracking load was reached. After cracking, the load-deflection behavior for the statically-loaded specimen was similar to that of the fatigue- loaded specimens, except for SS-O and FS-O. This indicates that the fatigue cycles had no significant effect on the deflection at that stage. The service live load shown in Figure 12.1.15-a was the Load/Joint Length of 2.18 kips/ft which corresponded to the maximum positive calculated moment of 4.55 kip-ft/ft after cracking reported in Section 10.1.5.9. The service live load shown in Figure 12.1.15-b was the Load/Joint Length of 7.82 kips/ft, which corresponded to the maximum calculated shear near the pad load of 5.34 kips/ft based on analyses using the finite element model shown in Figure 12.1.11. During the shear tests, differential vertical deflection was observed at the interface of the CP and precast panel. Figure 12.1.16 shows the relative displacement (RD) measured between the side of the joint 0 2 4 6 8 10 12 14 16 18 0.0 0.2 0.4 0.6 0.8 1.0 Deflection (in) L oa d/ Jo in t L en gt h (k ip /f t FS-7 FS-O SS-O Service Live Load Calculation SS-7

417 interface and joint center versus fatigue-cycle (N) curves for FS under specific loading levels during interim static load tests. For the FF specimens, the relative displacement of the joint interface was zero under service live load and thus was not studied here. From Figure 12.1.16, it can be seen that the relative displacement was dependent upon the applied load. As expected, the relative displacement increased with increasing applied load. The change in relative displacement observed over the course of the fatigue tests was not significant, so it was concluded that the fatigue cycles did not have an impact on the performance of the joints subjected to service live load. (a) FS-O Specimen under P1 0.000 0.005 0.010 0.015 0.020 0.0 0.5 1.0 1.5 2.0 Number of Fatigue Cycles (million) D ef le ct io n (i n) P1=7.5 kip/ft P1=6.5 kip/ft P1=5.5 kip/ft P1=4.5 kip/ft P1=3.5 kip/ft P1=2.5 kip/ft 0.000 0.005 0.010 0.015 0.020 0.0 0.5 1.0 1.5 2.0 Number of Fatigue Cycles (million) D ef le ct io n (i n) P2=7.5 kip/ft P2=6.5 kip/ft P2=5.5 kip/ft P2=4.5 kip/ft P2=2.5 kip/ft P1=2.5 kip/ft

418 (b) FS-O Specimen under P2 c. FS-7 Specimen under P1 d. FS-7 Specimen under P2 Figure 12.1.16: RD-N curves for flexure and flexure-shear specimens 0.000 0.005 0.010 0.015 0.020 0.0 0.5 1.0 1.5 2.0 Number of Fatigue Cycles (million) D ef le ct io n (i n) P1=7.5 kip/ft P1=6.5 kip/ft P1=5.5 kip/ft P1=4.5 kip/ft P1=3.5 kip/ft P1=2.5 kip/ft 0.000 0.005 0.010 0.015 0.020 0.0 0.5 1.0 1.5 2.0 Number of Fatigue Cycles (million) D ef le ct io n (i n) P2=7.5 kip/ft P2=6.5 kip/ft P2=5.5 kip/ft P2=4.5 kip/ft P2=3.5 kip/ft P2=2.5 kip/ft

419 12.1.10. Crack Width Development During the tests, the cracks at the interface between the grouted joint and the concrete panel were measured. The two cracks marked as “①” and “②” shown in Figure 12.1.17 show the locations of “Crack 1” and “Crack 2,” respectively. Figure 12.1.17: Cracks at joint-panel interface

420 Figure 12.1.18 shows the moment-crack width relationship for the SF and SS specimens. The service live load shown in Figure 12.1.18-a was the maximum positive calculated moment of 4.55 kip-ft/ft, after cracking, reported in Section 10.1.5.9. The service live load shown in Figure 12.1.18-b was 3.37 kip-ft/ft, which was the corresponding moment (CM) occurring with the maximum shear of 5.34 kips/ft after cracking based on analyses using the finite element model discussed in Section 10.1.5.9. The crack width was measured using a Crack Comparator, and represents the maximum width along the crack. From Figure 12.1.18-a, it can be seen that the width of the two cracks monitored during the tests of SF-O and SF-7 developed at different rates. The width of “Crack 1” grew faster than that of “Crack 2” due to the reinforcement yielding that developed at the joint interface of the “Crack 1” location. The crack widths at the service live load level were relatively small compared to the cracks that developed after that. At the same load level, the crack widths of SF-O were greater than those of SF-7, but the difference was not significant. The crack widths of SF-7 at failure were larger than those of SF-O, because SF-7 had a higher failure load. In Figure 12.1.18-b, the two cracks of the same test, SS-O or SS-7, widened at approximately the same rate with the increasing level of the loading. The crack widths at the service live load level were relatively small compared to the cracks that developed after that. At the same load level, the crack widths of SS-O and SS-7 were similar. The crack widths of SS-7 at failure were larger than those of SS-O, because SS-7 had a higher failure load.

421 (a) SF Specimens 0 5 10 15 20 25 30 35 0.00 0.04 0.08 0.12 0.16 Crack Width (in) M om en t/J oi nt L en gt h (k ip -f t/f t Measured Failure Load for SF-O Service Live Load Crack 1 for SF-O Crack 2 for SF-O Crack 1 for SF-7Crack 2 for SF-7 MTS Capacity

422 (b) SS Specimens Figure 12.1.18: Moment-crack width curves for static flexure and static flexure-shear specimens Figure 12.1.19 shows the crack width-fatigue cycle curves (CW-N) for the fatigue tests. The crack widths represent the maximum crack widths (Crack 1 or Crack 2) measured within the joint at specified load levels after various numbers of fatigue cycles were completed. For the FF tests, 2.18 kips /ft corresponded to the maximum static loading of 26.2 kips (13.1 kips on each pad) during the interim static loading tests, as discussed in Section 12.1.7. And 1.25 kips /ft corresponded to the static loading of 15 kips (7.5 kips on each pad). For the FS tests, 5.83 kips /ft and 4.17 kips /ft corresponded to the static loading of 35 kips and 25 kips, respectively, applied during the interim static loading tests. From Figure 12.1.19, it can be seen that the crack widths measured within the joint were dependent upon the applied load, as would be expected. At similar levels of load, changes in crack widths were not significant, except for specimen FF-7. So the influence of fatigue cycles on the crack width within the joint was concluded to be negligible under service live load for the joints with overnight cure materials, while fatigue cycles had an influence on the joint with the 7-day cure material. For FF-7, a transverse crack was observed to develop at the middle of the joint length, shown circled in Figure 12.1.20; this was the largest 0 5 10 15 20 25 30 35 0 0.04 0.08 0.12 0.16 Crack Width (in) M om en t/J oi nt L en gt h (k ip -f t/f t Crack 1 for SS-O Measured Failure Load for SS-O Service Live Load Crack 2 for SS-O Crack 1 for SS-7 Crack 2 for SS-7 Measured Failure Load for SS-7

423 crack measured under service live load, 0.012 in. compared with 0.008 in. for Crack 1 and 2, and as the load was approaching the failure, this crack was getting smaller and Crack 1 and 2 kept increasing and had the largest crack widths. For FF-O, a transverse crack was observed to develop at a similar location to that of FF- 7; however, in the case of FF-O, the transverse crack width was similar in size to the cracks observed along the joint interfaces under service live load. For FS-O and FS-7, the crack along the joint interface, Crack 1 or 2 as shown in Figure 12.1.17, had the maximum width under service live load. Under P1 loading, Crack 1 was the widest crack, while Crack 2 had the maximum crack width under P2 loading. This resulted because P1 and Crack 1 were on the same side of the joint and P2 and Crack 2 were on the other side. In the fatigue flexure tests (FF), the joint with 7-day cure material developed larger crack widths than those of the overnight cure material, while in the fatigue shear tests (FS) the crack width of the joint with the 7-day cure material was similar to that of the overnight cure material.

424 (a) FF-O (b) FF-7 (c) FS-O (d) FS-7 Figure 12.1.19: CW-N curves for fatigue flexure and fatigue flexure-shear specimens 0.000 0.005 0.010 0.015 0.020 0.0 0.5 1.0 1.5 2.0 Number of Fatigue Cycles (million) C ra ck W id th ( in ) P=2.18 kip/ft P=1.25 kip/ft 0.000 0.005 0.010 0.015 0.020 0.0 0.5 1.0 1.5 2.0 Number of Fatigue Cycles (million) C ra ck W id th ( in ) P=2.18 kip/ft P=1.25 kip/ft 0.000 0.005 0.010 0.015 0.020 0.0 0.5 1.0 1.5 2.0 Number of Fatigue Cycles (million) C ra ck W id th ( in ) P2=5.83 kip/ft P1=5.83 kip/ft P2=4.17 kip/ft P1=4.17 kip/ft 0.000 0.005 0.010 0.015 0.020 0.0 0.5 1.0 1.5 2.0 Number of Fatigue Cycles (million) C ra ck W id th ( in ) P2=5.83 kip/ft P1=5.83 kip/ft P2=4.17 kip/ft P1=4.17 kip/ft

425 Figure 12.1.20: Transverse crack in FF-7 test (circled in red) 12.1.11. Strain Development Figure 12.1.21 shows the strain-fatigue cycle curves (MS-N) for the fatigue tests representing the reinforcement strain in the joint measured during intermittent static tests conducted over the course of the fatigue cycles under service live load. The strain gage number and the loading are shown in the figure. For example, “R1-9” denotes the strain gage on U-bar #1 of the right slab that was located 7 in. away from the outside bend of the bar, as shown in Figure 12.1.3. The loading denoted “M = 4 k-ft/ft” signifies the moment applied to the joint per unit length. From Figure 12.1.21, it can be seen that negligible variations in reinforcement strain were observed at the peak static loads applied intermittently during the course of the fatigue tests. Strain increases were observed with increase in distance from the bend of the U-bars, as shown in Figure 12.1.21. The R4-10 and L5-10 strains were exceptions, because they were located outside of the joint. Figure 12.1.22 shows the moment-strain curves representing the maximum strain values in the joint zone for each specimen, which also show the variation of the reinforcement strain after fatigue cycles is not significant.

426 (a) FF-O (b) FF-7 (c) FS-O (d) FS-7 Figure 12.1.21: MS-N curves for fatigue flexure and fatigue flexure-shear specimens M=4.0 kip-ft/ft 0 300 600 900 1200 0.0 0.5 1.0 1.5 2.0 Number of Fatigue Cycles (million) M ic ro st ra in R1-9 R1-7 R4-10 R4-6 R7-6 M=4.0 kip-ft/ft 0 300 600 900 1200 0.0 0.5 1.0 1.5 2.0 Number of Fatigue Cycles (million) M ic ro st ra in L5-6 R1-9 R1-7 L5-10 a-2 M=7.5 kip-ft/ft 0 300 600 900 1200 0.0 0.5 1.0 1.5 2.0 Number of Fatigue Cycles (million) M ic ro st ra in R1-9 R4-10 R1-7 L5-6 a-2 M=7.5 kip-ft/ft 0 300 600 900 1200 0.0 0.5 1.0 1.5 2.0 Number of Fatigue Cycles (million) M ic ro st ra in L5-6 R1-9 R4-6 L5-10 a-2

427 (a) Flexure tests (b) Flexure-shear tests Figure 12.1.22: Moment-strain curves for flexure and flexure-shear specimens 0 5 10 15 20 25 30 0 1000 2000 3000 4000 5000 Microstrain M om en t/J oi nt L en gt h (k ip -f t/f t SF-O FF-O FF-7 SF-7 Service Live Load 0 5 10 15 20 25 30 0 1000 2000 3000 4000 5000 Microstrain M om en t/J oi nt L en gt h (k ip s- ft /f t SS-O FS-O Service Live Load FS-7 SS-7

428 12.1.12. Failure of Specimen As shown in Figure 12.1.23, the failure modes exhibited by the SF-O and FF-O specimens were typical flexural failures. After the U-bar reinforcement yielded, the CP material in the joint zone eventually crushed at the top. Two large cracks developed along the two interfaces of the joint during the tests and two developed within the joint at the bottom of the slab, as shown in Figure 12.1.24-a and b. The widths of the two large cracks within the joint increased significantly when the load was large at approximately 18 kip- ft/ft. The two cracks along the interfaces of the joint had the largest crack widths through the flexural failure tests. The flexure specimens experienced a ductile failure (i.e. The U-bars yielded and then concrete or grout at the top of the joint crushed). After reaching peak loads, ductile specimens can develop large displacements. It was not possible to fail specimens SF-7 and FF-7 due to the limiting capacity of the MTS testing equipment. In specimens SF-7 and FF-7, a horizontal crack was observed to develop at the bottom of the slab starting from the edges of the slab, as shown in Figure 12.1.24-c and d, which increased significantly at the end of the test, when the load was close to the capacity of the MTS equipment. Flexure-shear failure modes were observed in specimens SS-O, FS-O and SS-7, FS-7. A shear crack developed across the joint zone when the specimen failed, as shown in Figure 12.1.25. For SS-O and FS-O specimens, a crack along the joint interface at the loading side developed though the specimen thickness, and the CP material crushed at the top along the joint interface at failure. For SS-7 and FS-7 specimens, a crack developed from the lower part of the joint interface at the loading side and then propagated vertically, not following the interface. The grout also crushed at the top along the joint interface at failure. For SS-O, FS-O and SS-7, FS-7 specimens, one horizontal crack within the joint, as shown in Figure 12.1.26, increased tremendously at the bottom when the slabs were close to failure. (a) SF-O (b) FF-O Figure 12.1.23: Specimen failures for flexure specimens

429 (a) SF-O (b) FF-O (c) SF-7 (d) FF-7 Figure 12.1.24: Cracks at the bottom of the flexure specimens (the joint interfaces are marked with dashed lines)

430 (a) SS-O (b) FS-O (c) SS-7 (d) FS-7 Figure 12.1.25: Specimen failures for flexure-shear specimens

431 (a) SS-O (b) FS-O (c) SS-7 (d) FS-7 Figure 12.1.26: Cracks at the bottom of the flexure-shear specimens (the joint interfaces are marked with dashed lines) 12.2. Conclusions Based on the parametric study and the experimental program, the following conclusions can be made: 1. Fatigue loading had little influence on the behavior of the longitudinal joints (flexure and flexure- shear test specimens) in terms of average curvature of the joint, deflection at midspan, relative displacement of the joint interface and joint center as well as reinforcement strain under service live load. 2. Fatigue loading was observed to have an effect on the loading capacity of the flexure specimens using the overnight cure material. After 2 million cycles, the specimens fabricated with the overnight cure material had less load capacity than the corresponding specimens subjected to the

432 static load tests. For the specimens with 7-day cure material in the joint, fatigue loading had negligible effect on the results for the flexure-shear tests. In the case of the flexure tests, the failure load was not reached due to limitations of the MTS test equipment. 3. Joints with the 7-day cure material performed better than those with the overnight cure material in some cases. Examples included the flexure-shear tests, SS and FS, where the joints with the 7-day cure material had larger failure loads and curvatures than those of the specimen with the overnight cure material. This was because the 7-day cure material used developed higher strengths than could be achieved with the overnight cure material in the tests. Based on these tests, the U-bar detail was deemed to be a viable connection system for longitudinal joints between full-depth precast deck panels and decked bulb-Ts.