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.

248 Chapter 9 Flange/Deck Connection: Selection of Most Promising Connection Detail through Two-Phase Experimental Investigation 9.0 Introduction to Two-Phase Experimental Investigation to Finalize Connection Concept Detail for Further Study The headed bar detail and deformed wire reinforcement (DWR) and stainless steel (SS) U-bar details described in Section 8.1.2 were investigated and compared in the first phase of a two-phase experimental program conducted to finalize the best-performing connection detail for further study. The detail with the best performance as deemed from the Phase I experiments was subjected to additional testing in Phase II to investigate variations in parameters including overlap lengths, rebar spacings, and concrete strengths. The two-phase study was conducted on representative longitudinal (flexure) joint specimens and transverse (tension) joint specimens cast monolithically. The detailed connection concept, which was the outcome of the two-phase study, was subsequently investigated in jointed test specimens which included overnight- cure and 7-day cure closure pour (CP) materials (see Chapter 11) in the joint subjected to static and fatigue loadings as described in Chapters 12 and 13 for the longitudinal and transverse connection concepts, respectively. This chapter presents the results of the initial two-phase experimental investigation. 9.1. Test Phase I 9.1.1. Specimen Design As shown in Figure 8.1.1, in a real bridge, the longitudinal joint specimen is representing the specimen across the longitudinal joint and transverse joint specimen is representing the specimen across the transverse joint. The longitudinal and transverse reinforcement are the directions in the real bridge. Therefore, the main reinforcement in the longitudinal joint specimen is transverse reinforcement. And the main reinforcement in the transverse joint specimen is longitudinal reinforcement. Based on typical deck design practice and the desired deck thickness (e.g., DBT top flange thickness is 6 in.), the maximum bar size was limited to #5 bars. This size of reinforcement represented the maximum possible bar size to achieve the required tight bend diameters and clearance requirements for the head. It was assumed that the transverse reinforcement would always be located as the outer reinforcement layers in the deck to provide the largest moment couple to resist flexure of the deck in the transverse direction. Consequently, the specimens were detailed to give the transverse rebar the largest moment arm without violating cover requirements. To simulate the most demanding case to achieve flexural capacity across the longitudinal joint, the deck was minimized in those specimens to represent a situation where a transverse joint may not be required (e.g., a relatively short span, simply-supported bridge). In such cases, it may be possible to achieve thinner decks because there may not be a need to provide headed or U-bars in the longitudinal direction. In such a situation, the deck thickness was controlled by the outer layer of reinforcement achieving the tightest bend diameter. The resulting deck thickness for the longitudinal joint specimens was 6-¼ in. The most demanding load case for a transverse joint is to resist a high tension force (e.g., a transverse joint over the pier in the negative moment region). In such cases, headed or U-bars in the longitudinal direction can be placed inside of the outer reinforcement layers which can also be headed or U-

249 bars in the transverse direction. The resulting deck thickness for the transverse joint only (or the combined transverse and longitudinal joints) was 7-¼ in. to accommodate the required reinforcement clearances. Due to two joint directions being investigated, two design methods were used to determine a realistic spacing for the reinforcement of the joint details. The spacing of the rebar in the longitudinal joint was designed utilizing the AASHTO strip method of deck design (AASHTO LRFD 2010). This method takes into account the largest positive and negative moments that would be experienced by the longitudinal joint. However, to determine the controlling load case for a transverse joint, the transverse joint should be positioned over an interior support in a continuous span bridge system. In this case, if the deck was compositely connected to the girder, the deck would have to resist large tensile forces that would be produced by the negative moment developed there. So, the rebar spacing for the transverse joint was determined using flexural analysis of the deck-girder composite section, conservatively assuming that all the tension force created by the negative moment would be resisted by the deck. The spacing of the rebar in each joint direction was determined by its corresponding deck design. Two PCI design examples were used to design the reinforcement spacings for the joint details. The longitudinal joint spacing was designed using design example 9.8 in the PCI Bridge Design Manual (PCI 2003). The spacing for the transverse joint reinforcement was designed using design example 9.6 in the PCI (2003). Both examples used the same bridge cross section and longitudinal section. The bridge cross section consisted of four BT-72 girders that were spaced at 12 ft on center. The longitudinal section of the bridge consisted of a center span of 120 ft and two side spans of 110 ft. The PCI design examples were used, so that the steel areas determined would represent the amount of steel in a typical bridge deck. An 8 in. deck was first designed for the interior portion of the deck. This design was carried out to obtain an estimate of the area of steel used in decks with a conventional thickness. Table 9.1.1 shows the results of the design for the U-bar detail assuming Grade 75 steel, and Table 9.1.2 shows the design results for the headed bar detail assuming Grade 60 steel for the 8 in. deck thickness. Two different values of exposure factors, class one and class two, were used in the designs to show where the service limit state controlled the design. Exposure factors are directly proportional to the crack widths expected at service level loading. So, the larger the exposure factor the larger the expected crack widths. In order to minimize the deck thickness, two other thicknesses were also investigated. A 6-¼ in. deck was designed for the longitudinal joint direction utilizing both joint details, and a 7-¼ in. deck was designed for the transverse joint direction utilizing both joint details. The transverse joint required a larger deck thickness in order to keep the transverse deck rebar, required to cross the longitudinal joint, as the outer reinforcement layer, as discussed earlier. The 6-¼ in. and 7-¼ in. deck thicknesses were the thinnest deck sections possible when considering the joint details and cover requirements. The 6-¼ in. thick deck section was then designed. The results from this design were used for the design of the reinforcement for the longitudinal joint direction for both joint details. Both yield strengths and exposure classes were used. Table 9.1.3 shows the results of the design for the U-bar detail, and Table 9.1.4 shows the design results for the headed bar detail.

250 Table 9.1.1: Reinforcement required for the U-bar detail in an 8 in. deck Locations fy = 60 ksi eγ = .75 eγ = 1 Bars Size Spacing (in) Bars Size Spacing (in) M+ (Bottom) Transverse # 5 8.0 # 5 8.0 Longitudinal # 5 11.5 # 5 11.5 M- (Top) Transverse # 5 7.0 # 5 8.5 Longitudinal # 4 18.0 # 4 18.0 Table 9.1.2: Reinforcement required for the headed bar detail in an 8 in. deck Locations fy =75 ksi eγ = .75 eγ = 1 Bars Size Spacing (in) Bars Size Spacing (in) M+ (Bottom) Transverse # 5 8.0 # 5 9.5 Longitudinal # 5 11.5 # 5 11.5 M- (Top) Transverse # 5 7.0 # 5 8.5 Longitudinal # 4 18.0 # 4 18.0 Table 9.1.3: Reinforcement required for the U-bar detail in a 6-¼ in. deck Locations fy =75 ksi eγ = .75 eγ = 1 Bars Size Spacing (in) Bars Size Spacing (in) M+ (Bottom) Transverse # 5 4.5 # 5 4.5 Longitudinal # 5 6.5 # 5 6.5 M- (Top) Transverse # 5 4.0 # 5 4.5 Longitudinal # 4 12.0 # 4 12.0

251 Table 9.1.4: Reinforcement required for the headed bar detail in a 6-¼ in. deck - Locations fy = 60 ksi eγ = .75 eγ = 1 Bars Size Spacing (in) Bars Size Spacing (in) M+ (Bottom) Transverse # 5 4.0 # 5 4.0 Longitudinal # 5 5.5 # 5 5.5 M- (Top) Transverse # 5 4.0 # 5 4.0 Longitudinal # 4 12.0 # 4 12.0 Tables 9.1.3 and 9.1.4 show that the service limit state had little impact on the rebar spacing for the 6-¼ in. deck section. The only case that the service limit state governed was the top layer of transverse reinforcement of the U-bar detail design. Also, when comparing the amount of reinforcement required for the 6-¼ in. and 8 in. thick deck sections (see Tables 9.1.3 and 9.1.4 compared to 9.1.1 and 9.1.2), it was found that the 8 in. thick deck required approximately half the reinforcement that was required for the 6-¼ in. thick deck. Although the area of steel required for the 6-¼ in. thick deck was larger than that required for the 8 in. thick deck, the decrease in weight of the slimmer deck section and the labor costs saved by the possibility of accelerated construction provided the incentive to investigate the slimmer deck section. The required steel area for the transverse joint was then determined. The reinforcement in the transverse joint was designed as if the joint was located over an interior pier of a continuous span bridge system with the deck compositely connected to the girder. The negative moment developed in these locations would create large tensile forces in the deck, which would require more longitudinal steel in these regions. The composite cross section and the negative moment used in the flexural calculations were taken from example 9.6 of PCI (2003). The negative moment value that was used in the flexural calculations was 4837 kip-ft. The amount of longitudinal deck reinforcement was determined by a conventional flexural design using the composite section. The centroid of the reinforcement was assumed to be at mid-height of the deck and the required amount of reinforcement was determined for both the 60 ksi and 75 ksi yield strengths. Table 9.1.5 contains the results of the designs.

252 Table 9.1.5: Negative moment longitudinal reinforcement Longitudinal Reinforcement (MU- region) fy (ksi) rebar size spacing (in) 60.9 #5 4.5 75.4 #5 5.5 To facilitate comparison of the results of the different joint details and joint directions, the deck designs were used to develop a rebar spacing configuration that was used for the construction of all specimens. The reinforcement in both the longitudinal (flexure) joint specimens and transverse (tension) joint specimens was selected to be the same for both specimens (across the joint and perpendicular to the joint). The reinforcement spacing was conservatively selected from the tables above to provide the same detailing across the joint in the specimens that used the Grade 75 reinforcement as was used in those that used the Grade 60 reinforcement. Due to the use of the U-bar detail, the top and bottom layers of primary joint reinforcement had to have the same spacing (because by the very nature of the U-bar, it consists of two layers of reinforcement). In summary, the reinforcement that crossed the joint consisted of a top and bottom layer of #5 rebar spaced at 4.5 in. (see Table 9.1.5). The top layer of reinforcement transverse to the joints consisted of #4 rebar spaced at 12 in. and the bottom layer of reinforcement transverse to the joint was determined to be #5 rebar spaced at 6 in. The joint overlap length, which is defined herein as the distance between the reinforcement bearing surfaces, was determined based on the expected development length of a U -bar. The ACI equation for determining the development length of a standard hook in tension was used to calculate the approximate development length of a U-bar. This equation does not directly apply to the U-bars that were used, because the U-bars do not meet the dimensional requirements for a standard hook, namely the 3db bend diameter used in the U-bar fabrication violated the minimum 6db bend diameter specified in ACI 318-08. Eqn. (9.1.1) shows the ACI development length equation for a standard hook in tension. .02 ' e y dh b c fl d f λ Ψ =      (ACI 12.5.2) (9.1.1) where ldh represents the development length from the tail of the hook, in.; ψe is the epoxy coating factor; λ is the lightweight concrete factor; fy is the reinforcement yield strength; √(f’c), the square root of the concrete compressive stress, is expressed in psi; and db is the diameter of the bar, in. The terms, ψe and λ, were both set equal to one in Eqn. (9.1.1), because the rebar that was used was not epoxy coated and the concrete was not lightweight. It should be noted that a number of states do not use epoxy -coated reinforcement in their decks, 4 of 38 states surveyed by Russell (2004) indicated that they do not specify epoxy-coated deck reinforcement. In addition, 14 of the 38 states surveyed, indicated that they specify a

253 metallic coating for the deck reinforcement, and 7 of the 38 states indicated that they were beginning to use solid stainless steel. In order to minimize the joint width between the precast flanges or panels, it is desirable to avoid epoxy-coated reinforcement because that may require a longer lap length to develop the reinforcement. In addition, tighter bend diameters are feasible with stainless steel reinforcement than with ordinary mild reinforcement. Using Eqn. (9.1.1), the development length was calculated for a No. 5 bar, assuming a concrete compressive strength of 6 ksi and a steel yield strength of 75 ksi, because deformed wire reinforcement and stainless steel were materials used for this joint detail. The ACI 318 -08 development length modification factor of 0.7 was used, because the specimens met the bar cover perimeters of having not less than 2.5 in. of side cover and not less than 2 in. of cover beyond the extension of the bar. The development length of a standard hook bar in tension for this situation was calculated to be 8.5 in. In the testing program, an overlap length of 6 in. was used for the U-bar detail. Where the overlap length is taken from the inside bend of the adjacent hooks. This would be 7-¼ in. to the outside of the hooks, less than the 8.5 in. specified by ACI 318-08, but lacer bars were also used in the connection to enhance the mechanical anchorage. Based on Li et al. (2010), one-layer of headed bars with varying overlap lengths, 2.5 in., 4 in. and 6 in., were tested and compared, and the lap length for the headed bar detail was recommended to be 6 in. Although the head was larger in the Li study (Li et al. 2010) as discussed earlier, for the case of the No. 5 headed bars in this study, with the small head size, the overlap length was also taken to be 6 in. Because the behavior of the specimens with both joint details was to be compared, the overlap lengths for the U-bar and headed bar joint details were made the same, i.e., 6 in. Two transverse lacer bars were added to each joint detail to provide continuity and confinement of the joints. The inclusion of lacer bars in both joint types was due to previous research conducted by Gordon and May (2005) on loop bar joints in tension. In their experimental program a loop bar specimen was tested in tension without transverse lacer bars, which resulted in a sudden brittle failure. The transverse lacer bars were installed in the center of the bend in the U-bar detail and in the middle of the headed bar detail. The specimens and connection details designed for the longitudinal joint direction are shown in Figures 9.1.1 and 9.1.2. The specimens and connection details designed for the transverse joint direction are shown in Figures 9.1.3 and 9.1.4.

254 Figure 9.1.1: U-bar longitudinal joint specimen Figure 9.1.2: Headed bar longitudinal joint specimen #5 bars @ 6 in spacing #4 bars @ 12 in spacing 6.25 in 120.0 in 6.0 in 15.0 in 4.5 in 1" 6.0 in 6.25 in #4 bars @ 12 in spacing #5 bars @ 6 in spacing 4.5 in 15.0 in 120.0 in

255 Figure 9.1.3: U-bar transverse joint specimen Figure 9.1.4: Headed bar transverse joint specimen 9.1.2. Experimental Setup and Instrumentation Experimental Set-up Simple static tests were performed for both the longitudinal connections and transverse connections. The specimens representing the longitudinal joint direction were tested in bending. The specimens representing the transverse joint direction were tested in tension, because of the tensile forces assumed to be generated in the deck created by negative moment regions of the bridge system. Both joint details (i.e., U-bar and headed bar) were tested in both joint directions. 4.5 in 6.0 in 72.0 in 15.0 in #4 Lacer Bars 7.25 in #4 bars @ 12 in spacing #5 bars @ 6 in spacing 72.0 in #4 bars @ 12 in spacing #5 bars @ 6 in spacing 4.5 in 15.0 in 7.25 in #4 Lacer Bars 6.0 in

256 As stated previously, the specimens representing the longitudinal joint direction were tested in bending. A modified version of the four point bending tested was used for the flexural test set-up. The actuators used to apply force were located on the outside of the supports; this set-up produced upward deflection in the specimen producing tension on the top surface of the specimen. Tension on the top surface of the specimen produced safer conditions for observing cracks and crack propagation. This set-up, like the four point bending test, produced a constant maximum moment between the supports where the joint was located. Figure 9.1.5 shows the experimental set up used to test the longitudinal joint specimens (Flexural Test Set-Up). Figure 9.1.5: Flexural test set-up (Longitudinal Joint Test) 120.0 in 6.25 in 40.0 in 28.0 in 28.0 in Actuator JointExt./Comp. Readings (LDVT's) Actuator Deflection Deflection Reading (LVDT) 96.0 in Reading (LVDT) (LVDT) Reading Deflection

257 As stated previously the specimens representing the transverse joint direction were tested in tension. The tension test set-up was slightly more complicated than the flexural test set-up. The longitudinal reinforcement in the transverse joint specimens was welded to ¾ in. threaded rods. These threaded rods were used to bolt the tension specimen to the support and loading beams. The support beam was connected to the specimens and then placed on top of the load frame. The support beam was then braced and clamped into position, so it would remain stationary. The loading beam was then connected to the specimen and the actuators. The actuators pushed the loading beam down, which applied a tension force to the specimens. Figure 9.1.6 shows the tension test set-up. 72.0 in Actuator Joint Reading (LVDT) Displacement Loading Beam Force Direction Load Frame Longitudinal Beam Load Frame Support Beam Column

258 Figure 9.1.6: Tension test set-up (Transverse Joint) Instrumentation The specimens were instrumented to achieve an understanding of the U-bar and headed bar details in bending and tension in the longitudinal and transverse joint tests, respectively. Loads cells were used to measure the loads applied to the specimens. Linear motion transducers (LMT’s) were used to measure the deflection of the specimens. Also, horizontally placed LMT’s were used to determine the curvature of the longitudinal joint specimens tested in flexure. The strain in the joint reinforcement was measured using strain gages. The main purpose of the strain gages was to determine whether the reinforcement could achieve yielding in the joint, which would indicate that the joint details could produce a precast deck system that would emulate monolithic behavior. For the U-bar detail, strain gages were installed on the top and bottom of the three interior U-bars. These gages were installed 2, 6, 8, and 10 in. away from the bend of the bar. Also, a strain gage was installed on the outside apex of the bend on each gaged U-bar. For the headed bar joint detail, strain gages were installed on the top and bottom headed bars on the three interior sets of headed bars. The term headed bar set refers to the top and bottom headed bars at a specific location. These gages were installed 1, 4, 6, and 8 in. away from the bearing surface of the head on headed bar sets 2 and 3. Strain gages were installed 4, 6, 8 and 10 in. away from the bearing surface of the head on headed bar set 4. Figures 9.1.7 and 9.1.8 show the strain gage configurations used for the U-bar and headed bar details, respectively.

259 Figure 9.1.7: U-bar joint detail strain gage configuration Figure 9.1.8: Headed bar joint detail strain gage configuration The strain gage notation used in Figures 9.1.7 and 9.1.8 indicates the U-bar or the headed bar set where the gage was located and the relative position of that gage. For example, strain gage 2-3 indicated that the gage was located on U-bar 2 or headed bar set 2 and that it was the third gage away from the bearing surface of that bar. In the strain gage results section of this paper, the strain gages are labeled additionally with a “T” or “B” indicating that they are located on the top or bottom of a U-bar or located on the top or bottom bar of a headed bar set. The distances to the centerlines of the strain gages are given at the bottom of Figures 9.1.7 and 9.1.8 in inches. The first length given for each bar is the distance from the centerline of the first gage to the bearing surface of the reinforcement. The other distances shown in the figures represent the center to center spacing between consecutive strain gages in inches. Centerline U-Bar 1 U-Bar 3 U-Bar 5 U-Bar 2 U-Bar 4 U-Bar 2 and 4 Lacer Bar 2Lacer Bar 1 U-Bar 3 2''2'' 4'' 2''2'' 4'' 2'' 2'' 2-1 2-2 2-3 2-4 3-4 3-3 3-2 3-1 4-1 4-2 4-3 4-4 LB 2-1 LB 1-2 LB 1-1 LB 2-2 Centerline Headed Bar Set 1 Headed Bar Set 3 2-1 Headed Bar Set 5 Headed Bar Set 2 Headed Bar Set 4 3-4 3-3 3-2 4-1 4-2 4-3 4-4 2-2 2-3 2-4 3-1 2'' 2'' 2''2'' 3'' Headed Bar Set 2Headed Bar Set 3 Lacer Bar 1 LB 1-1 LB 1-2 2''4'' 2''2'' Headed Bar Set 4 1'' 1'' 3''

260 Strain gages were also installed on the transverse lacer bars. A strain gage was installed 1 in. away from the bearing surface of the head, and another strain gage was installed in the center. All lacer bars used for the construction of the specimens were gaged. Figure 9.1.9 shows the strain gage configuration of the lacer bars. Figure 9.1.9: Lacer bar strain gage configuration LMT’s were used to determine specimen deflections at various locations. For the longitudinal joint bending specimens, LMT’s were installed at the center of the specimen and at both ends. For the transverse joint tension specimens, LMT’s were installed on the top and the bottom of the joint to measure joint elongation and at the bottom of the specimen to measure the total deflection of the specimens. The curvature of the longitudinal joint bending specimens was also measured using LMT’s. The LMT’s were installed parallel to the specimen on the top and bottom of the joint zone. Top and bottom surface strains could then be calculated based on the initial gage length of the LMT wire and the change in the readings taking into consideration the distance between LMTs and concrete surface. These surface strains were then used to determine the curvature of the specimens throughout the duration of the test. Figures 9.1.5 and 9.1.6 show the positions of all previously discussed LMT’s. 9.1.3. Specimen Construction, Reinforcement Cost and Fabrication Specimen Construction Figure 9.1.10 shows several specimens during the construction process. As can be seen in the photos, the specimens were monolithically cast. In practice the joint would consist of precast deck panels with staggered, protruding reinforcement that would then be anchored into a cast-in-place joint. In this experimental program the behavior of the reinforcement of the different joint details was of main concern. The monolithic specimens allowed the behavior of the joint details to be observed without the additional variable of a grouted joint. Additional testing with the best performing detail contained the grouted joint between the precast panel sections. These tests are described in Chapters 12 and 13. LB1-2LB1-1 1.0 in 5.75 in

261 (a) Reinforcement in the forms (b) Plate used to ensure proper layer separation (c) Threaded rods extending through the form (d) Casting concrete (e) Vibrator use in the joint zone (f) Concrete finishing Figure 9.1.10: Specimen construction

262 Comparing the constructability of the U-bar and the headed bar joint details, the U-bar detail seemed to be the easiest to construct for two reasons. The U-bar detail produces a joint that is less congested than the headed bar detail and therefore allows for easier placement of precast deck components. The headed bar detail is more congested due to the bearing heads. Instead of considering just the diameter of the bar in considering clearances, which is the case for the U-bar detail, the outside diameter of the bearing head must be accommodated. The head reduces the construction tolerances, which may lead to placement problems in the field. The reinforcement of the U-bar detail was also easier to tie and set in place compared to the headed bar detail. After the top and bottom of the U-bars were set to the correct height by tack welding a thin plate between the free ends, the U-bars acted as a single reinforcement cage, which made the installation of the rebar easier than the two separate layers of reinforcement produced by the headed bar detail. During shipment, storage, and placement of precast deck components, unintentional bending of the protruding joint reinforcement due to handling is a concern. The U-bar detail provides benefits considering this aspect. In the headed bar detail, one headed bar could be accidentally bent at a time, but for the U-bar detail two bars would have to be bent at one time, because the top and bottom layer of reinforcement consists of one bent bar. The bending of two bars at once would require a greater force and would therefore be more unlikely to happen. With the U-bar detail, it may be possible to bend a mesh of reinforcement to prepare the rebar cage for placement. This would reduce construction time in the precast plant, which would reduce overall project cost. Reinforcement Cost and Fabrication In February 2009, reinforcement suppliers were contacted to determine an approximate cost of reinforcing materials and their fabrication. The cost information received from the suppliers was a representative snap shot of the reinforcement prices at that particular time for construction scale orders. The ease of fabrication of the U-bars was also discussed with the suppliers. The lowest price quote was given for conventional A615 rebar with attached Lenton Terminator® bearing heads. The price per ton of this reinforcement was 800 dollars with an additional cost of 25 dollars for the application of each Lenton Terminator® bearing head. Deformed wire reinforcement was very competitive with the headed bar price. Two price ranges were given for fabricated deformed wire reinforcement, the first price of 850-950 dollars a ton was given for single cut and fabricated wires. The second price range of 900-1000 dollars a ton was given for fabricated wire mesh. The highest price was the price of fabricated stainless steel reinforcement. The price quote given was for Enduramet 32 stainless steel, which was 5000 dollars a ton. The stainless steel price also included fabrication in a stainless steel only production line which eliminated contamination from black carbon dust. Even though stainless steel had the largest initial material cost, the life cycle cost of a deck using this material could offset the high initial cost. The ease of the U-bar fabrication process was also discussed with the reinforcement distributors. The U-bar fabrication process was of concern, because of the tight bends that were required in the U-bar detail so the deck thickness could be minimized. The stainless steel supplier stated that the tight bends were not a problem for the stainless steel material, because of its high ductility. The stainless steel supplier also stated that it may be possible to reduce the bend diameter to less than three times the diameter of the bar (3db) without breaking the material. The deformed wire reinforcement supplier stated that the tight bend was

263 also not a problem for the material. After determining the best fabrication method for the bends, no deformed wire reinforcement was broken during fabrication. 9.1.4. Material Testing Concrete Testing Tables 9.1.6 and 9.1.7 show the results of the concrete compressive strength tests. Table 9.1.6: Concrete compressive strengths, U-bar specimens U-Bar Specimens ( Four Specimens, Cast on July 24, 2008) 7 Day Test 28 Day Test Cylinder # Stress (psi) Cylinder # Stress (psi) 1 8710 1 10350 2 9753 2 11651 3 9820 3 11575 Average 9428 11192 Table 9.1.7: Concrete compressive strengths, headed bar specimens Headed Bar Specimens (Two Specimens, Cast on August 28, 2008) 7 Day Test 28 Day Test Cylinder # Stress (psi) Cylinder # Stress (psi) 1 8301 1 9948 2 8068 2 9383 3 8114 3 9416 Average 8161 9582 Reinforcing Materials Testing Tension tests were performed on the deformed wire reinforcement and stainless steel reinforcement to obtain accurate material properties. The conventional Grade 60 reinforcement used for the headed bars was not tested. Four samples of each reinforcing material were tested. An Instron Universal Testing Machine (UTM) was used to test the reinforcement and to obtain the data necessary to construct stress versus strain curves for both reinforcement types.

264 An extensometer with the required 2 in. gage length was used to determine the strain in the deformed wire reinforcement (DWR) and the stainless steel reinforcement (SS) materials. When the yield strength of the reinforcing materials was approached, the extensometer was removed as a precautionary measure. The remainder of the strain readings was determined using an initial gage length of the entire specimen and the position of the moving platen of the UTM. The stress in the specimens was calculated by using the force readings taken from the load cell in the UTM divided by the area of the #5 bars tested. This allowed for the construction of a stress versus strain curve for both the deformed wire reinforcement and the stainless steel reinforcement. The modulus of elasticity for the materials was taken as the slope of the stress-strain trend line using the extensometer data. As shown in Table 9.1.8, the modulus of elasticity of the DWR and SS was determined to be 29918 ksi and 26802 ksi, respectively. A modulus of elasticity of 29000 ksi was used in all theoretical calculation. The ultimate strength data is also given in Table 9.1.8. The stress versus strain curves for the DWR and the SS reinforcement is plotted in Figure 9.1.11, which shows that stainless steel reinforcement was extremely ductile, compared to the deformed wire reinforcement. Table 9.1.8 Modulus of elasticity and ultimate strength Modulus (ksi) Ultimate strength (ksi) DWR SS DWR SS Sample1 31536.0 26913.0 117.0 117.0 Sample 2 28621.0 28537.0 115.0 116.9 Sample 3 28781.0 24902.0 116.4 117.0 Sample 4 30734.0 26856.0 113.7 117.0 Average 29918.0 26802.0 115.5 117.0

265 Figure 9.1.11: Stress versus strain curves for Deformed Wire Reinforcement (DWR) and Stainless Steel (SS) Weld Testing A preliminary concern of the tension test set-up was the strength of the welds between the longitudinal rebar in the specimens and the threaded rods that were used to connect the specimens to the loading and support beams. If the welds were to fail before the specimen, then the test would terminate before valuable information about the connection behavior could be ascertained. Tension tests were performed on the rebar to threaded rod welds to ensure that the weld strengths were greater than the rebar yield strengths. The top connection detail used in the tension test set-up is shown in Figures 9.1.12 and 9.1.13. Figure 9.1.14 shows the weld test set-up.

266 Figure 9.1.12: Connection detail, conceptual drawing Figure 9.1.13: Photo of the top connection detail

267 Figure 9.1.14: Weld test set-up The welds tested were made using the Metal Inert Gas, Tungsten Inert Gas and Shielded Metal Arc Welding (MIG, TIG and SMAW) welding methods on all rebar materials. Again, conventional reinforcement had a yield strength of 60 ksi, and deformed wire reinforcement and stainless steel reinforcement had a yield strength of 75 ksi. The threaded rods that were used in the experimental program had a diameter of ¾ in. and a minimum yield strength of 110 ksi. The first set of weld tests placed were one pass welds made by both MIG and TIG welding methods with a 70 ksi carbon welding wire. This welding method produced a small ring of weld material around the outside of the threaded rod and the rebar, and did not fuse the threaded rod and reinforcement together in the center. This welding method produced small cross-sectional areas in the weld region, which led to small capacities. The desired capacity for each of the welds was the capacity of the #5 bar (i.e., 0.31in.2x60ksi (or 75ksi)=18.6(or 23.3)kip). Table 9.1.9 shows the testing results of the one pass weld tests and Figure 9.1.15 shows the typical failure of the one pass weld specimens.

268 Table 9.1.9: Weld test results, one pass welds Material Welding Method Capacity (kips) Failure Mode Stainless Steel, fy = 75 ksi MIG 15.5 Weld Broke Deformed wire Reinforcement, fy = 75 ksi MIG 13 Rebar Pulled Out of Weld Conventional Rebar, fy = 60 ksi MIG 13.7 Rebar Pulled Out of Weld Stainless Steel fy = 75 ksi TIG 11.8 Weld Broke Deformed wire Reinforcement fy = 75 ksi TIG 15 Weld Broke Conventional Rebar fy = 60 ksi TIG 15.8 Weld Broke Figure 9.1.15: One pass weld failure The second set of weld tests consisted of the same welding methods and materials, but the threaded rods and rebar were beveled, so larger weld areas and capacities could be produced. This set of weld tests all produced capacities that were larger than the 23.3 kip force required to yield a #5 bar with a yield strength

269 of 75 ksi. These results indicated that the welds could yield the reinforcing material, but to account for any unforeseen forces in the experiment, larger weld capacities were preferred. Table 9.1.10 shows the results of the second set of weld tests using the beveled weld geometry. The third set of weld tests used a 110 ksi welding stick and the beveled weld geometry. The welding method was also changed to shielded metal arc welding (SMAW). These welds were tested on all reinforcing materials. The results of the third set of welding tests are given in Table 9.1.11. Table 9.1.10: Weld test results, Beveled Welds Material Welding Method Capacity (kips) Failure Mode Stainless Steel, fy = 75 ksi MIG 23.8 Weld Broke Deformed wire Reinforcement, fy = 75 ksi MIG 26.7 Weld Broke Conventional Rebar, fy = 60 ksi MIG 26.5 Weld Broke Stainless Steel fy = 75 ksi TIG 25.9 Weld Broke Deformed wire Reinforcement fy = 75 ksi TIG 24.5 Weld Broke Conventional Rebar fy = 60 ksi TIG 23.8 Weld Broke

270 Table 9.1.11: Weld test results, beveled welds and 110 ksi welding stick Material Welding Method Capacity (kips) Failure Mode Stainless Steel, fy = 75 ksi SMAW 21.4 Weld Broke Deformed wire Reinforcement, fy = 75 ksi SMAW 33.5 Weld Broke Conventional Rebar, fy = 60 ksi SMAW 29.5 Weld Broke Table 9.1.11 shows that the 110 ksi welding sticks and the SMAW welding method produced very strong welds with the deformed wire reinforcement and the conventional rebar, but the strength of the weld decreased for the stainless steel reinforcement. The decrease in weld strength for the stainless steel reinforcement may have been because of a weld incompatibility between the 110 ksi welding stick and the stainless steel reinforcement. So the threaded rod to stainless steel weld was tested again using a stainless steel welding stick. The SMAW welding method using a 308 stainless steel welding rod was used in conjunction with the beveled weld geometry to increase the weld strength between the threaded rod and the stainless steel reinforcement. Only one weld test was conducted using these parameters, due to time constraints and an adequate result. The capacity of the weld using the previously described parameters produced a strength of 28.4 kips, which was adequate strength for the transverse joint (tension) tests. The results of the weld tests indicated that the previously described tension test set-up was feasible and that the weld strength between the threaded rod and the longitudinal reinforcement was not a limiting factor. 9.1.5. Results and Discussion Flexural Capacity (Longitudinal Joint Behavior) The flexural test specimens were designated as HB-1, SB-1 and WB-1, where the first letter corresponds to H=headed bar, S=stainless steel U-bar or W=deformed wire reinforcement U-bar; and the second letter, “B,” stands for “bending.” The numerical designation, “1,” represents the first specimen. The AASHTO strength I and service limit states were used in the determination of both the positive and negative moments. A 2 in. future wearing surface, 6-¼ in. deck thickness, and the live load determined from AASTHO Table A4-1, were used to calculate the following moment demands. Table 9.1.12 and Table 9.1.13 show the service and strength I moments for the specimens. It should be noted that the original design was based on the Strip Method using Grade 60 reinforcement; however, the size and spacing of the

271 reinforcement was kept the same when the Grade 75 reinforcement was used. Considering the specimens with the Grade 75 reinforcement could resist a higher moment, a reinforcement ratio was used to adjust the loading demands to reflect this difference. Table 9.1.12: Moment demands for specimens containing the U-bar detail U-Bar Detail Specimens (SB-1 and WB-1) M+ (kip-ft) M- (kip-ft) Service 10.1 8.3 Strength I 16.9 13.6 Table 9.1.13: Moment demands for specimens containing the headed bar detail Headed Bar Detail Specimen (HB-1) M+ (kip-ft) M- (kip-ft) Service 9.5 7.8 Strength I 16.0 13.1 All specimens produced similar flexural capacities. Specimen HB-1, which contained the headed bar joint detail, produced the lowest flexural capacity, which was 29.1 kip-ft. Specimen WB-1, which contained the U-bar detail made of deformed wire reinforcement, produced a flexural capacity of 31.0 kip-ft. The largest flexural capacity was produced by specimen SB-1, which contained the U-bar joint detail made of the stainless steel material; the flexural capacity of this specimen was 31.9 kip-ft. Comparing the flexural capacities of the specimens to the required capacities (Strength I) shown in Tables 9.1.12 and 9.1.13, it can be seen that all specimens produced capacities that were much greater than the required Strength I limit state moments. These results show that both the U-bar and headed bar joint details are capable of creating precast deck systems that can emulate the behavior of cast-in-place deck systems in terms of strength requirements. Moment versus deflection and moment-versus-curvature curves were constructed for each specimen. The total applied moment was determined from the actuator load cell. The deflection data were taken from the LMT located in the center of the joint. The curvature data were taken from the horizontally placed LMT’s located above and below the joint. Figure 9.1.16 shows the moment versus deflection curves and Figure 9.1.17 shows the moment-versus-curvature curves for all of the flexural specimens. Also plotted in the figures is the Strength I limit state moment requirements for the U-bar (UB) and headed bar (HB) cases.

272 Figure 9.1.16: Moment versus deflection curves

273 Figure 9.1.17: Moment-versus-curvature curves From Figures 9.1.16 and 9.1.17, the ductility of the specimens can be seen. Figure 9.1.16 shows that the U- bar specimens, SB-1 and WB-1, produced larger deflections than the headed bar detail contained in specimen HB-1. The same trend is also shown in Figure 9.1.17. Specimens SB-1 produced larger curvatures than HB-1 and WB-1. Even though specimens SB-1 and WB-1 were more ductile then HB-1, they did not sacrifice capacity. Although the concrete age at structural testing was older than 28 days for all six specimens, the concrete strength test at 28 days, shown in Tables 9.1.6 and 9.1.7, was used to determine the calculated section capacities discussed below. The respective nominal yield strength of either 60ksi or 75ksi was used to determine the calculated section capacities. The higher capacities of SB-1 and WB-1 can be attributed to the joint detail, the higher reinforcement yield strength, and the higher compressive strength of the concrete. The higher ductility produced by the U-bar specimens compared with the headed bar specimen may have been due to the U-bar joint detail: U-bars can deform (bent around lacer bars) while the head of headed bars bears against concrete. When comparing the ductilities of SB-1 and WB-1, it can be seen that specimen SB-1 had a significantly larger deflection and curvature. Specimens SB-1 and WB-1 had the same measured concrete strength, rebar yield strength, and joint detail; the only difference was the reinforcing materials used in them. SB-1 consisted of the stainless steel reinforcing material and WB-1 consisted of deformed wire reinforcement. Stainless steel is extremely ductile compared to deformed wire reinforcement as shown in Figure 9.1.11. Theoretical moment-versus-curvature curves were constructed, so that the behavior of the longitudinal joint (flexural) specimens could be compared to theoretical values. The theoretical moment-versus-

274 curvature curves were constructed for both the U-bar detail and the headed bar detail. Two separate moment-versus-curvature curves had to be determined because of the different concrete strengths, assumed rebar yield strengths and reinforcement configurations used in the specimens. The cracking moment was calculated considering the modulus of concrete rupture stated in ACI 318-08, the additional stiffness added to the cross sections by the rebar and the uncracked moment of inertia. The moment at yield was calculated considering the cracked section moment of inertia and a steel strain equal to the yield strain. The steel yield strain was determined by dividing the yield stress by a modulus of elasticity of 29,000ksi. The nominal moment capacity was calculated based on ACI 318-08 procedures, which considered an equivalent rectangular concrete stress block and no strain hardening of the reinforcement. Both layers of reinforcement were considering during the calculation of all moments. The specimen cross sections used for the theoretical calculations consisted of the six bar side shown in Figure 9.1.18 (U-bar specimens) and Figure 9.1.19 (headed bar specimens). The calculated theoretical moments and corresponding curvatures are given in Table 9.1.14 for the six bar (three loop) configuration. The actual behavior of the specimen was controlled by the joint zone where the steel changed from six bars (three loop bars) to four bars (two loop bars) in the loop splice region. Figure 9.1.18: Cross Section used for theoretical calculations of the U-bar specimens Figure 9.1.19: Cross section used for the theoretical calculations of the headed bar specimen 15.0 in 4.5 in 3.0 in 4.5 in 4.9 in 2.3 in 3.0 in 6.25 in 15.0 in 4.5 in 3.0 in 4.5 in 4.4 in 2.8 in 3.0 in 6.25 in

275 Table 9.1.14: Calculated moments and curvature (six-bar side) Headed Bar Detail (fy=60ksi) U-bar Detail (fy=75ksi) Moment (kip-ft ) Curvature (10-5/in) Moment (kip-ft) Curvature (10-5/in) Mcr 6.19 Φcr 4.27 Mcr 6.82 Φcr 4.27 My 20.7 Φy 73.15 My 25.82 Φy 76.96 Mn 20.7 Φn 213.87 Mn 25.82 Φn 214.44 When comparing the tested specimen moment capacities and the calculated moment capacities, it can be seen that the specimens behaved better than expected. The theoretical calculations were based on 60ksi and 75 ksi nominal yield strengths and tested concrete 28-day compressive strengths. For HB-1, the calculated moment, when the nominal design yield strength was used, was 20.7 kip-ft, but the measured capacity was 29.1 kip-ft. The measured capacity of HB-1 was 41 percent greater than the calculated value. SB-1 and WB-1 both had the same theoretical nominal moment capacity because they had the same nominal design yield strength of 75 ksi and reinforcement configurations; the calculated moment capacity was 25.8 kip-ft. The measured moment capacities of SB-1 and WB-1 were 31.9 kip-ft and 31.0 kip-ft respectively. The measured moment capacities of SB-1 and WB-1 were 23.5 and 20.4 percent greater than the calculated value. The main point shown here was that although reinforcement was not continuous across the joint zone (i.e., the U-bars were spliced), the specimen resisted a moment capacity as if continuous reinforcement was provided across the joint with bars that had the nominal design yield strength. The theoretical moment-versus-curvature curves are plotted with the corresponding measured moment- versus-curvature curves obtained from experimental data in Figures 9.1.20 and 9.1.21 for the U-bar and headed bar details, respectively.

276 Figure 9.1.20: Measured and theoretical moment-versus-curvature curves for the U-bar details Figure 9.1.21: Measured and theoretical moment-versus-curvature curves for headed bar details

277 Figures 9.1.20 and 9.1.21 show that all specimens produced a larger capacity than the theoretical capacity. Also, the specimens were more ductile than predicted by the theoretical curvature calculations based on nominal properties; the detailed explanation for this difference is provided later in reference to Figure 9.2.22. The behavior of the specimens shows that the U-bar and headed bar joint details produced joints deemed to emulate the behavior of a cast-in-place bridge deck. Flexural Specimen Behavior (Longitudinal Joint Behavior) The first cracks to appear in all of the flexural specimens were transverse cracks uniformly spaced along the length of the specimens. In the joint zone the transverse cracks formed at the ends of the longitudinal reinforcement. The transverse cracks formed in the top tension face of the specimens first (see Figure 9.1.5 for test set-up) and then propagated deeper into the specimens as the loading progressed. At higher moments, lateral cracks formed in the direction of the longitudinal reinforcement, which corresponded to the direction of transverse reinforcement in an actual bridge deck. These lateral cracks formed over the longitudinal reinforcement that comprised the lightly reinforced half of the specimens. Diagonal cracks then formed from the longitudinal rebar comprising the lightly reinforced half of the specimens and extended to the outside edge of the specimen. Figure 9.1.22 shows the crack patterns of specimens SB-1, WB-1 and HB- 1 at failure.

278 (a) Specimen SB-1 (b) Specimen WB-1 (c) Specimen HB-1 Figure 9.1.22: Flexural crack patterns at failure

279 The numbers written next to the cracks shown in Figure 9.1.22 represent the total force, in kips, applied to the specimen when the cracks formed. All flexural specimen failures were ductile, producing yielding in the reinforcement and crushing of the concrete on the compression face of the specimens, under the joint zone. Flexural Crack Widths at Service Level Loading (Longitudinal Joint Behavior) Flexural crack widths were visually measured systematically throughout the testing of all bending specimens using a crack width gage. The crack width gage contained numerous lines of labeled widths. The cracks in the specimens were compared to the lines on the crack width gage; the width of the line that most accurately represented the crack was recorded as the crack width. The positive service level moments given in Tables 9.1.12 and 9.1.13 were used to compare the crack widths of the specimens because the flexural specimens represented a longitudinal joint that would primarily resist positive moment. The positive service moments for the U-bar detail and the headed bar detail were 10.1 kip-ft and 9.5 kip-ft, respectively. For specimens SB-1 and WB-1, crack widths were measured at 9.29 kip-ft and 11.65 kip-ft. For specimen HB-1, the crack widths were measured at 8.11 kip-ft and 10.47 kip-ft. Because the crack widths were not measured at the corresponding service moments, the crack widths at the service-level loading were found by interpolating between the measured values. SB-1 was found to have an average crack width of 0.0075 in. at service level loading. The average crack width at service level loading for specimen HB-1 was found to be 0.0083 in. WB-1 was found to have an average crack width of 0.01 in. at service level loading. So the U-bar detail produced both the largest and the smallest crack widths at service level loading, which were 0.01 in and 0.0075 in. The headed bar detail produced the mid-range crack widths. Strain Gage Data (Longitudinal Joint Behavior) Strain gages were installed on the specimen reinforcement in the configurations shown in Figures 9.1.7 and 9.1.8. The purpose of the strain gage data was to verify the required overlap length needed to develop the yield strength of the reinforcement. The strain gage data were expected to show increasing strain as the distance away from the bearing surface of the reinforcement increased (i.e., as the spliced bars transferred the tension across the joint region). The strain gage data did not show this trend because of the effect of the cracks. The strain gage locations that were crossed by a crack produced high strain readings, while the other gage locations had smaller strain readings. The joint cracking patterns produced scattered results that could not be used to verify the required overlap length; however, the strain gage data do show that both joint details allowed the reinforcement to develop large strains, and the required forces were transferred across the joints. Figures 9.1.23, 9.1.24 and 9.1.25 show the moment versus reinforcement strain curves for specimens SB-1, WB-1 and HB-1, respectively.

280 (a) Bottom of U-bar 2 (Specimen SB-1) (b) Top of U-bar 2 (Specimen SB-1) (c) Bottom of U-bar 3 (Specimen SB-1) (d) Top of U-bar 3 (Specimen SB-1) (e) Bottom of U-bar 4 (Specimen SB-1) (f) Top of U-bar 4 (Specimen SB-1)

281 (g) Lacer Bars 1 and 2 (Specimen SB-1) Figure 9.1.23: Moment versus rebar strain curves for SB-1 (a) Bottom of U-bar 2 (Specimen WB-1) (b) Top of U-bar 2 (Specimen WB-1) (c) Bottom of U-bar 3 (Specimen WB-1) (d) Top of U-bar 3 (Specimen WB-1)

282 (e) Bottom of U-bar 4 (Specimen WB-1) (f) Top of U-bar 4 (Specimen WB-1) (g) Lacer Bars 1 and 2 (Specimen WB-1) Figure 9.1.24: Moment versus rebar strain curves for WB-1

283 (a) Bottom of U-bar 2 (Specimen HB-1) (b) Top of U-bar 2 (Specimen HB-1) (c) Bottom of U-bar 3 (Specimen HB-1) (d) Top of U-bar 3 (Specimen HB-1) (e) Bottom of U-bar 4 (Specimen HB-1) (f) Top of U-bar 4 (Specimen HB-1)

284 (g) Lacer Bars 1 and 2 (Specimen HB-1) Figure 9.1.25: Moment versus rebar strain curves for HB-1 Tensile Capacity (Transverse Joint Behavior) Specimens ST-1, WT-1 and HT-1 were tested in tension to simulate the conditions in the deck at a transverse joint where the compression forces were assumed to be resisted by the girder. As mentioned in the section on Flexural Capacity (Longitudinal Joint Behavior), the specimen naming convention indicated the type of reinforcement detail investigated (i.e., S=stainless steel; W=deformed wire; H=headed bar); the “T” stands for tension test, and the “1,” indicates the specimen number of that type. Specimen HT-1 utilized a headed bar joint detail made of conventional reinforcement. Specimens ST-1 and WT-1 both utilized the U-bar joint detail. The reinforcement used in ST-1 was stainless steel, and the reinforcement used in specimen WT-1 was deformed wire reinforcement (DWR). The largest tension capacity that was expected from the tension specimens was the force determined by multiplying the area of steel of the lightly reinforced side of the specimens by the appropriate rebar yield strength. HT-1 was reinforced with conventional rebar with a nominal yield strength of 60 ksi and was expected to have a maximum tensile capacity of 74.4 kips. The maximum capacity of HT-1 was calculated considering the four #5 bars comprising the lightly reinforced side of the specimen and the rebar yield strength. The U-bar specimens both used reinforcement that had a nominal yield strength of 75 ksi, considering the yield strength of the reinforcement and the area of the four #5 bars in the lightly reinforced side of the specimen, the maximum expected tensile capacity of the U-bar specimens was 93.0 kips. All specimens produced similar tensile capacities. HT-1 produced the lowest tensile capacity, which was 89.81 kip. This was to be expected because HT-1 contained conventional rebar with the lowest nominal rebar yield strength. The second highest tensile capacity was produced by the U-bar detail using stainless steel reinforcement (ST-1). ST-1 produced a tensile capacity of 91.8 kip. The largest tensile capacity was produced by WT-1, which was 93.3 kips. Both WT-1 and HT-1 exceeded the expected tensile capacity. The additional capacity of specimens WT-1 and HT-1 might be attributed to the actual yield strength being higher than the nominal design yield strength or the steel may have developed strain hardening. However, specimen ST-1 did not meet the expected capacity of 93 kips, it only had a capacity of 91.78 kips. ST-1 had a capacity that was 1.3 percent

285 less than the expected tensile capacity. The low capacity could have been due to the fact that the welds broke during testing at a load of approximately 65 kips. The specimen was rewelded and tested to failure, but the specimen may have been damaged during the first unsuccessful test. ST-1 may have experienced damage that could have affected its behavior and tensile capacity during the second successful test. Load deflection curves were constructed for the test specimens. During the second test of ST-1, no strain gage or LMT data was collected. The only data collected from the second test were the applied load and actuator displacement. The actuator displacement allowed for the determination of the total specimen deflection. The data collected from both tests were spliced together to form a complete load versus deflection curve. Load versus deflection curves were also constructed for specimens HT-1 and WT-1. The applied load was determined from the load data from the MTS actuator system. The total specimen deflection was used in the construction of the curves; these data were taken from the LMT that was attached to the bottom of the specimens. The load versus deflection curves for specimens HT-1, WT-1 and ST-1 were plotted together and can be seen in Figure 9.1.26. The end of the bold line for specimen ST-1 signifies where the data from the first test ends and where the data from the second test begins. Figure 9.1.26: Total applied force versus deflection curves Figure 9.1.26 shows the load deflection curves of all the specimens leveling off at approximately the force required to yield the reinforcement as determined from the nominal capacities. The increase in deflection while holding constant load signifies that the reinforcement in the specimens was yielding. Both the shape of the load versus deflection curves and the capacities of the specimens indicate that it can be assumed that the reinforcement in all the specimens yielded. The behavior of the specimens shows that the U-bar and the headed bar joint details can successfully yield the joint reinforcement without brittle failure. This

286 result shows that both joint details could effectively be used as a transverse joint in a negative moment region, which would mainly produce global tension in the deck. Tension Specimen Behavior (Transverse Joint Behavior) All specimens produced similar crack patterns up to and beyond the service loading. The tensile service load was calculated by using the service level negative moment found in design example 9.6 of the PCI Bridge Design Manual (PCI 2003). The neutral axis was then found for the cracked composite cross section, which was then used in conjunction with the service level negative moment to determine the total tensile service load. The area of steel in the specimens was compared to the total required area of steel to determine the tensile service loads for the specimens. The tensile service load for HT-1 was determined to be 44.1 kips (i.e., 35.6 ksi stress times the reinforcement area of 1.24 in.2), and the tensile service load for ST-1 and WT-1 was determined to be 44.1 x (75/60) = 55.1 kips. Different tensile service loads were determined for the U- bar and headed bar specimens because of the different nominal yield strengths of their reinforcing materials. The first cracks to appear were transverse cracks evenly spaced along the length of the specimens. The joint zone usually experienced transverse cracking after several other transverse cracks had already formed in other locations. Delayed transverse cracking in the joint zone may have been due to the larger area of reinforcement in the joint region compared to that of the body of the specimens. The transverse cracks initially were found only on the surface of the concrete, and as the loading progressed the cracks propagated through the entire thickness of the specimens. Additional loading produced longitudinal cracks that appeared above the main longitudinal reinforcement in the specimens. These longitudinal cracks appeared above the longitudinal reinforcement that comprised the lightly reinforced half of the specimen, or the top half of the specimens in this particular set-up. When approaching the capacities of the specimens, diagonal cracks appeared close to the sides of the specimens. These diagonal cracks would usually propagate toward a transverse crack in the joint zone and cause the failure surface for the specimens. Figure 9.1.24 shows the crack patterns at failure for specimens ST-1, WT-1 and HT-1. The numbers written by the cracks in Figure 9.1.27 represent the total force applied in kips when the crack was formed.

287 (a) Specimen ST-1 (b) Specimen WT-1 (c) Specimen HT-1 Figure 9.1.27: Tension crack patterns at failure Tensile Crack Widths at Service Level Loading (Transverse Joint Behavior) Crack widths in the tension specimens were measured in the same way as in the flexural specimens. A crack width gage was used to visually determine the widths of the specimens. Crack widths were measured

288 systematically throughout the testing of HT-1 and ST-1, but only two crack widths were measured at two different loads for specimen WT-1. The average crack width of all cracks within the joint zone at 55 kips (estimated service load level for ST-1 and WT-1) was found to be 0.010 in for ST-1 (a comparable crack width measurement was taken at 45 kips for specimen ST-1, and the average crack width was found to be 0.006 in.). The average crack width at 44 kips (estimated service load level for HT-1) for specimen HT-1 was determined to be 0.012 in. The last crack width measurement for specimen WT-1 was taken at 40 kips and the average crack width was found to be 0.0079 in. Comparing the crack widths of the two joint details, one can see that the headed bar detail of specimen HT-1 generated the largest crack widths at approximately 40 kips and at its service level loading, compared to the crack widths created by the U-bar details of specimens ST-1 at its service level loading and WT-1 at approximately 40 kips. Strain Gage Data (Transverse Joint Behavior) Strain gages were installed on the specimen reinforcement in the configurations shown in Figures 9.1.7 and 9.1.8. Figures 9.1.28 and 9.1.29 show the total force versus reinforcement strain curves for specimens WT-1 and HT-1, respectively. The strain gage data for ST-1 were unfortunately lost for this specimen which suffered a weld failure. The strain gage data show that both joint details allow the reinforcement to develop large strains, and the required forces were transferred across the joints.

289 (a) Bottom of U-bar 2 (b) Top of U-bar 2 (c) Bottom of U-bar 3 (d) Top of U-bar 3 (e) Bottom of U-bar 4 (f) Top of U-bar 4 (g) Transverse Lacer Bars 1 and 2 Figure 9.1.28: Total force versus rebar strain for WT-1

290 (a) Bottom Bar of Headed Bar Set 2 (Specimen HT-1) (b) Top Bar of Headed Bar Set 2 (Specimen HT-1) (c) Bottom Bar of Headed Bar Set 3 (Specimen HT-1) (d) Top Bar of Headed Bar Set 3 (Specimen HT-1) (e) Bottom Bar of Headed Bar Set 4 (Specimen HT-1) (f) Top Bar of Headed Bar Set 4 (Specimen HT-1)

291 (g) Transverse Lacer Bars 1 and 2 Figure 9.1.29: Total force versus rebar strain for HT-1 9.1.6. Conclusions for Phase I U-Bar (SS, DWR) and Headed Bar (HB) Tests As stated previously, the main objective of this research was to test two joint details (i.e., U-bar and headed bar details) and select the best performing joint detail for further testing. The additional testing of the best performing detail described in Section 9.2 to investigate variations in parameters was the first step toward the development of design guidelines and standard details for longitudinal and transverse precast deck joints. Specimens containing both headed bar and U-bar joint details were tested in tension and in flexure to simulate the loading conditions in transverse and longitudinal joints, respectively, to ensure that the proposed joints could produce a precast deck system that would act monolithically. The behavior and capacities of the joint details were compared to select the best performing joint detail. All joint details produced adequate capacities and ductility in both the tension and flexure tests. Specimens containing the U-bar joint detail produced the largest capacities in both the bending and tension tests, as expected because they were fabricated with a higher grade of steel. Specimen WT-1 produced the largest tensile capacity which was 93.24 kips, and specimen SB-1 produced the largest flexural capacity which was 31.88 kip-ft. The U-bar detail produced the largest capacities without compromising ductility. Smaller crack widths at service level loading were also produced by the U-bar detail compared to the headed bar detail. The development of small crack widths increase durability by decreasing reinforcement corrosion, thus leading to longer deck life. The constructability and reinforcement costs of the joint details were also compared. The U-bar detail created a less congested joint, which made it the easiest to construct. The bearing heads of the headed bar detail require more space to accommodate the larger diameter of the rebar heads. This extra space would reduce construction tolerances and could therefore cause problems in placement of precast deck components. The U-bars can also be easily tied together to form a rebar cage, which would allow for easy construction in the precast yard when compared to the two single layers of reinforcement in the headed bar detail. The lowest material cost was the conventional rebar used for the headed bars. The material costs were competitive between the conventional rebar used in the headed bars and the deformed wire reinforcement. The cost for conventional reinforcement was approximately 800 dollars a ton with an additional cost of 25 dollars for the installation of each Lenton Terminator bearing head. The deformed wire reinforcement cost was 850-900 dollars a ton for single fabricated cut wires or 900-1000 dollars a ton for

292 fabricated wire mesh. The stainless steel reinforcement had the highest cost of 5000 dollars a ton including fabrication. Even though the initial cost of stainless steel is high, it should still be considered due to the potential to increase the life span of the structure. After consideration of capacity, service level crack widths, constructability, and cost, the U-bar detail constructed of deformed wire reinforcement was chosen for further testing. This detail produced adequate capacities in both the tension and flexural testing, while still producing adequate ductility. The constructability of the U-bar detail as well as the cost of the deformed wire reinforcement made this detail an economical choice. It should be noted however, that none of the reinforcement studied in the NCHRP 10-71 project was coated reinforcement. If the reinforcement were protected with epoxy coating, it may have an impact on the required length to transfer forces across the joint. Because the objectives of the project were to minimize the joint thickness, in regions where reinforcement protection is required, it is recommended that deformed wire reinforcement be used. The Phase I experiments showed that the U-bar detail could develop adequate capacity with an overlap length of 6 in., a rebar spacing of 4.5 in. and two transverse lacer bars, but the Phase I tests did not provide a sense of how variations in these parameters would affect the behavior of the joint. The Phase II tests, discussed in the next section, provided insight regarding the effects of these parameters. 9.2. Test Phase II Testing Parameters Based on the test results from Phase I, U-bar details with deformed wire reinforcement were selected for further testing in Phase II. The testing parameters of the U-bar joint detail are listed in Table 9.2.1. As discussed earlier, each specimen is labeled where the “W” represents deformed wire reinforcement, the “B” represents bending test, the “T” represents tension test, and the number represents the specimen number tested in chronological order. WB-1 was the first specimen to be tested in flexure, which was completed in Test Phase I. The testing parameters for WB-2, WB-3 and WB-4 were then specified based on testing results of WB-1. Also, WT-1 was the first specimen to be tested in pure tension, which was conducted as part of the Phase I tests. The testing parameters for WT-2, WT-3, and WT-4 were specified based on the testing results of WT-1.

293 Table 9.2.1: Testing parameters Specimen ID Concrete Strength Bar Spacing Joint Overlap Length (ksi) (in) (in) WB-1 WT-1 10.0 4.5 6.0 WB-2 WT-2 7.0 4.5 6.0 WB-3 WT-3 10.0 4.5 4.0 WB-4 WT-4 10.0 6.0 6.0 As shown in Table 9.2.1, two different concrete strengths (10 and 7 ksi), two different rebar spacings (4.5 and 6 in.), and two different overlap lengths (6 and 4 in.) were considered in this second phase tests. The details of the longitudinal joint test specimens are provided in Figures 9.2.1 through 9.2.4. And details of the transverse joint test specimens are provided in Figures 9.2.5 through 9.2.8. Figure 9.2.1: WB-1 longitudinal joint specimen (Tested in Phase I) Figure 9.2.2: WB-2 longitudinal joint specimen #5 bars @ 6 in spacing #4 bars @ 12 in spacing 6.25 in 120.0 in 6.0 in 15.0 in 4.5 in #5 bars @ 6 in spacing #4 bars @ 12 in spacing 6.25 in 120.0 in 6.0 in 15.0 in 4.5 in

294 Figure 9.2.3: WB-3 longitudinal joint specimen Figure 9.2.4: WB-4 longitudinal joint specimen Figure 9.2.5: WT-1 transverse joint specimen (Tested in Phase I) 6.25 in #4 bars @ 12 in spacing #5 bars @ 6 in spacing #4 Lacer Bars 4.0 in 15.0 in 120.0 in 4.5 in 20.0 in 6.0 in 120.0 in 6.0 in #5 bars @ 6 in spacing #4 bars @ 12 in spacing 6.25 in 4.5 in 6.0 in 72.0 in 15.0 in #4 Lacer Bars 7.25 in #4 bars @ 12 in spacing #5 bars @ 6 in spacing

295 Figure 9.2.6: WT-2 transverse joint specimen Figure 9.2.7: WT-3 transverse joint specimen 4.5 in 6.0 in 72.0 in 15.0 in #4 Lacer Bars 7.25 in #4 bars @ 12 in spacing #5 bars @ 6 in spacing 4.5 in 72.0 in 15.0 in 4.0 in #4 Lacer Bars #5 bars @ 6 in spacing #4 bars @ 12 in spacing 7.25 in

296 Figure 9.2.8: WT-4 transverse joint specimen 9.2.1. Experimental Setup and Instrumentation The Phase I test setups and instrumentation described in Section 9.1.2 were also used for the Phase II test series. The only change was that the strain gage configuration was modified based on the test results obtained for WB-1 and WT-1during the first phase of testing. In Phase II, gages were placed on either side of the expected development length location so that the region of yielding in the bar could be identified. In Phase I, the gages were placed on both the top layer and the bottom layer of the U-bar in WB-1 and WT-1. In Phase II, the gages were placed only on the bottom layer of the U-bar. For WB-3 and WT-3, the specimens with a 4 in. joint overlap length, gages were placed at 4, 6, and 8 in. away from the bend of the U-bar. For WB-2, WT-2, WB-4 and WT-4, the specimens with a 6 in. joint overlap length, gages were placed at 6, 8, and 10 in. away from the bend of the U-bar. Figures 9.2.9 and 9.2.10 show the strain gage configurations for each joint overlap length of the specimens tested during Phase II. The strain gage diagrams have notations indicating the U-bar identifier and the location of the gage on the bar. The U-bars are represented by “UB” and the lacer bars are indicated by “LB.” The distance from the bend of the U-bar to each gage is shown at the bottom of the diagram. All distances indicated on the diagrams are in inches and measured from center-to-center. A gage was placed on one lacer bar at 1 in. from the bearing surface of the head and a second gage was placed at the midpoint of the lacer bar. The strain gage configuration of the lacer bar is shown previously in Figure 9.1.9. 7.25 in #4 bars @ 12 in spacing #5 bars @ 6 in spacing 6.0 in 72.0 in 6.0 in 20.0 in #4 Lacer Bars

297 Figure 9.2.9: Strain gage configuration for WB-3 and WT-3 Figure 9.2.10: Strain gage configuration for WB-2, WT-2, WB-4, and WT-4 UB-3 UB-4 UB-2 4-1 4-2 4-3 3-13-23-3 2-1 2-2 2-3 LB1-2 LB1-1 LB-2LB-1 UB-5 UB-4 UB-3 UB-2 UB-1 4"2"2" 2"2"4" 4" 2"4" 2" 4-34-24-1 3-13-23-3 2-1 2-2 2-3 LB1-2 LB1-1 LB-2LB-1 UB-5 UB-4 UB-3 UB-2 UB-1 6" 2" 2" 6" 6" 2" 2" 2"2"6" UB-2 UB-4 UB-3

298 9.2.2. Material Testing Concrete Testing The longitudinal joint specimens were cast on September 16, 2009. When these three specimens were poured, fifteen companion cylinders were also cast. To get accurate concrete compressive strengths, three cylinders were cast for each specimen to be tested on the day of the actual flexural test conducted in the laboratory. Three cylinders were also cast for 7-day and 28-day compressive strength measurements. Fifteen companion cylinders were also cast with the transverse joint specimens which were poured on November 3, 2009. All cylinder tests complied with the ASTM C 39 standards when tested to determine the concrete compressive strength (ASTM C 39, 2005). The cylinders were loaded as specified in the standards and within the limit of 35±7 psi/s. The cylinder’s compressive forces were recorded at failure. Due to a machine malfunction, some cylinders were not able to be tested at the 7-day benchmark, as denoted by “n/a” in Table 9.2.2. The compressive strength test results are recorded in Tables 9.2.2 and 9.2.3. Table 9.2.2: Concrete compressive strengths (longitudinal joint specimens) Cylinder 7-Day Test Day of Test 28-Day Test ID (psi) (psi) (psi) W B- 1 1 8710 Ages exceeded 28 days. Results from “28-Day Test” were used. 10350 2 9753 11651 3 9820 11575 Average 9428 11192 W B- 2 1 8323 9359 10743 2 8780 9308 10504 3 8989 9103 11220 Average 8697 9257 10822 W B- 3 1 n/a 10544 10743 2 n/a 10385 10464 3 n/a 10180 10265 Average n/a 10370 10491 W B- 4 1 n/a 9759 10743 2 n/a 11711 10464 3 n/a 10097 10265 Average n/a 10522 10491

299 Table 9.2.3: Concrete compressive strengths (transverse joint specimens) Cylinder 7-Day Test Day of Test 28-Day Test ID (psi) (psi) (psi) W T- 1 1 8301 Ages exceeded 28 days. Results from “28-Day Test” were used. 9948 2 8068 9383 3 8114 9416 Average 8161 9582 W T- 2 1 7600 7600 9231 2 7958 7958 9111 3 7600 7600 8992 Average 7719 7719 9111 W T- 3 1 9072 9629 10743 2 9390 9231 10265 3 9231 9629 10743 Average 9231 9496 10584 W T- 4 1 9072 9589 10743 2 9390 9152 10265 3 9231 9987 10743 Average 9231 9576 10584 9.2.3. Results and Discussion Flexural Capacity (Longitudinal Joint Behavior) The four longitudinal joint specimens, WB-1, WB-2, WB-3, and WB-4, were tested in flexure. As stated before, the “W” represented deformed wire reinforcement, the “B” represented bending test, and the number identified the specimen tested in chronological order. The service level moments were calculated for WB-1 based on the AASHTO service limit states. WB-1, as well as WB-2 and WB-3, had a width of 15 in., therefore the service moments were be the same for those specimens. However, the width of WB-4 was 20 in., so the service moments increased by the width ratio of 1.33. The results of these moment calculations are presented in Table 9.2.4.

300 Table 9.2.4: Service moments M+ M- WB-1, WB-2, WB-3 WB-4 WB-1, WB-2, WB-3 WB-4 (kip-ft) (kip-ft) (kip-ft) (kip-ft) 10.1 13.4 8.3 11.0 To compare the behavior among the specimens, moment versus deflection and moment-versus-curvature plots were created, as shown in Figures 9.2.11 and 9.2.12, respectively. The deflection measurements were obtained from the LMT’s placed at the center location and end locations on the tension side of the specimen. Because padding was secured between the specimen and the supports, the LMT’s could have given an inaccurate value for the total deflection. To improve the accuracy of the data, the deflection readings of the LMT’s at both ends of the specimen were averaged together. The averaged end deflections and the middle deflection were combined to obtain the total deflection of each specimen. The curvature values were derived based on the data provided by the horizontally placed LMT’s across the joint zone on the tension and compression side of the specimen. The moment capacities were derived from the forces applied by the MTS actuators. As the loading increased toward ultimate capacity, the deflection LMT’s were removed in order to prevent damage. By removing these LMT’s, the data was not completely representative of the behavior of the specimen. The moment versus deflection curves should reach a peak and then taper off to ultimate failure. This tapering effect was evident in WB-1 and WB-4. However, the deflection readings were not available at ultimate for WB-2 and WB-3. These trends can be viewed in Figure 9.2.11. The curvature LDVT’s were also removed as the loading increased to ultimate capacity. As the cracks propagated and the crack widths increased, the LMT on the top of the specimen was removed because the measuring device went out of range. The LMT located on the bottom of the specimen was removed when the concrete began to spall due to high compressive stresses. The moment-versus-curvature curves should also taper off toward ultimate. The data for WB-1 and WB-2 displayed this tapering effect, but the top and bottom LMT’s were removed before adequate data could be recorded for WB-3 and WB-4. These trends are displayed in Figure 9.2.12.

301 Figure 9.2.11: Moment versus deflection Figure 9.2.12: Moment-versus-curvature From Figures 9.2.11 and 9.2.12, the behavior of each specimen can be compared to one another. Because ACI 318-08 does not provide a specific method for calculating the moment capacity for a staggered U-bar detail at the joint, two continuously reinforced beam sections were analyzed using two different steel reinforcement patterns: (1) cross-section with As=1.24 in 2 representing the side of the longitudinal joint specimen with two U-bars and (2) cross-section with As=1.86 in 2 representing the other side of the

302 longitudinal joint specimen with three U bars. The cross-sections used in theoretical calculations are displayed in Figures 9.2.13 and 9.2.14. In order to observe the behavior of each test specimen in comparison to the continuously reinforced cross sections, the moment versus deflection and moment- versus-curvature curves of each specimen were plotted along with the continuously reinforced calculated curves in Figures 9.2.15 and 9.2.16, respectively. The curves relating to the continuously reinforced cross sections were plotted using analysis software called Response 2000. This program was used to predict the moment curvature behavior of a continuously reinforced specimen with either 15 in wide or 20 in wide cross section. The nominal yield stress of 75ksi, steel elastic modulus of 29000ksi and concrete strength of 10ksi or 7ksi were used in the Response 2000 analysis. The material properties used in Response 2000 are shown in Figure 9.2.17. (a) WB-1, WB-2, and WB-3 (b) WB-4 Figure 9.2.13: Cross sections, As of 1.24 in 2 4.5 in 4.9 in 2.3 in 15.0 in 6.25 in 20.0 in 2.3 in 4.9 in 6.0 in 6.25 in

303 (a) WB-1, WB-2, and WB-3 (b) WB-4 Figure 9.2.14: Cross sections, As of 1.86 in 2 15.0 in 2.3 in 4.9 in 4.5 in 4.5 in 6.25 in 6.0 in 6.0 in 4.9 in 2.3 in 20.0 in 6.25 in

304 (a) WB-1 (b) WB-2

305 (c) WB-3 (d) WB-4 Figure 9.2.15: Measured and calculated moment versus deflection

306 (a) WB-1 (b) WB-2

307 (c) WB-3 (d) WB-4 Figure 9.2.16: Measured and calculated moment-versus-curvature

308 Figure 9.2.17: Material properties used in Response 2000 Because the U-bars were staggered in the test specimens, the moment capacities, curvatures, and deflections differed from a continuously reinforced section. The joint was a combination of two U-bars, As=1.24 in 2, and three U-bars, As=1.86 in 2. Ideally the measured moment versus deflection curve would fall in between the two continuously reinforced calculated curves. The results of the flexural tests are shown in Tables 9.2.5 and 9.2.6.

309 Table 9.2.5: Flexural test results, nominal moments (Mn Specimen ID ) Mn (kip-ft) Test Calculation 2 Bars 3 Bars WB-1 31.05 27.48 36.00 WB-2 29.31 24.93 32.20 WB-3 25.50 27.48 36.00 WB-4 29.09 29.22 40.05 Table 9.2.6: Flexural test results, curvatures (Φn Specimen ID ) Φn (x10-5 1/in) Test Calculation 2 Bars 3 Bars WB-1 736.9 595.9 541.8 WB-2 550.4 532.6 492.3 WB-3 1459.2 595.9 541.8 WB-4 670.8 595.9 586.0 From Tables 9.2.5 and 9.2.6, WB-1, WB-2 and WB-4 produced adequate capacities in comparison to a continuously reinforced beam with As=1.24 in 2. WB-3 did not produce adequate moment capacity. WB-3 had the reduced joint overlap length of 4 in. All four U-bar detailed specimens were extremely flexible in comparison to what would be expected from a continuously reinforced section.

310 Flexural Specimen Behavior (Longitudinal Joint Behavior) According to the data, each specimen cracked at the following moments: WB-1 cracked at approximately 5.9 kip-ft; WB-2 cracked at 2.4 kip-ft; WB-3 cracked at 4.7 kip-ft; and WB-4 cracked at 6.8 kip-ft. Early in the loading, transverse cracks appeared during the flexural tests, as expected, on the tension side of the specimens. As the loading slowly increased, longitudinal cracks began to form inside the joint zone. These longitudinal cracks corresponded to the transverse joint reinforcement within the bridge deck. The longitudinal cracks in the joint did not appear until the following: 7.6 kip-ft for WB-2; 19.9 kip-ft for WB-3; and 18.0 kip-ft for WB-4. After the longitudinal cracks formed, the cracks continued to propagate deeper in the longitudinal and transverse directions until diagonal cracks appeared over the joint zone. Each specimen displayed the same diagonal crack patterns on the heavier-reinforced side of the joint. These diagonal cracks appeared near the failure point of each specimen. WB-3 produced the widest crack width in the center of the joint at failure. Also, as the specimen neared ultimate failure, the concrete on the compression side began to crush and fall to the ground. The cracks at failure can be seen in Figure 9.2.18. The numbers written on the specimens represent the forces that were applied when the cracks occurred.

311 (a) WB-1 (b) WB-2 (c) WB-3 (d) WB-4 Figure 9.2.18: Flexural cracks at failure Flexural Crack Widths at Service Level Loading (Longitudinal Joint Behavior) Cracks are a concern for several reasons. The appearance of cracks in a structure causes public concern, and deterioration of concrete and corrosion of reinforcement is a major concern among engineers. As discussed earlier in this chapter, the service level positive moment was 10.1 kip-ft for WB-2 and WB-3 and 13.4 kip-ft for WB-4. Because the longitudinal joint in a bridge deck would primarily resist positive bending, the service level positive moment of 10.1 kip-ft was used to compare crack widths. For WB-1, which was tested as part of Phase I, a crack width ruler was used to measure the crack widths. At service level loading, the crack width for WB-1 was 0.01 in. The crack widths were measured incrementally while the test was conducted; therefore the crack widths at the service level loading were interpolated values. At service level loading, the crack width for WB-2 was 0.009’’. The other crack widths were as follows: 0.006 in. for WB-3; and 0.003 in. for WB-4.

312 Strain Gage Data (Longitudinal Joint Behavior) As discussed in Section 9.2.1, strain gages were applied to the reinforcement to determine whether the reinforcement yielded within the overlap length. Several of the strain gages were destroyed during either transport or casting, so they were not able to be used for testing. For the longitudinal joint tests, moment-versus-strain curves for each gage location are plotted in Figure 9.2.19. Note that the strain gages for the WB-1 specimen shown in Figure 9.2.19 are actually gages 2-2, 2-3, 2-4, 3-2, 3-3, 3-4, 4-2, 4-3, 4-4 for comparison due to different gage configurations in Phases I and II. Theoretically, as the loading increases, the gages should indicate increasing strains in the reinforcement due to the increased strains in the reinforcement. For the longitudinal joint tests, most of the data proved this to be true and all the U-bars were observed to yield on the tension side of the specimen. Due to cracking of the concrete at some of the gage locations, a few of the strain gage readings produced noisy data and were discarded. The strain gages were placed near the approximate development length of the U-bars. Because of cracks occurring over some of the gages, no consistent data was present throughout all four specimens to verify the development length. However, the data was consistent in showing that the reinforcement developed large strains where diagonal cracks occurred.

313 (a) Gage 2-1 (b) Gage 2-2 (c) Gage 2-3 (d) Gage 3-1 (e) Gage 3-2 (f) Gage 3-3

314 (g) Gage 4-1 (h) Gage 4-2 (i) Gage 4-3 (j) Gage LB 1-2 Figure 9.2.19: Moment versus rebar strain for longitudinal joint tests Tensile Capacity (Transverse Joint Behavior) The four transverse joint specimens, WT-1, WT-2, WT-3, and WT-4, were tested in tension. As stated before, the “W” represented deformed wire reinforcement, the “T” represented tension test, and the number identified the specimen tested in chronological order. The tensile capacity of the U-bar specimen was calculated as the product of the lightly reinforced area of steel, As=1.24 in 2, and the nominal U-bar yield strength of 75 ksi. The expected nominal tensile capacity of the specimens was 93 kips. The service level was determined earlier. The results of the tests are tabulated below in Table 9.2.7.

315 Table 9.2.7: Tensile test results Specimen ID Tensile Capacity Deflection (kip) (in) Test Calculation Test Nominal WT-1 93.2 93.0 0.446 WT-2 88.7 93.0 0.411 WT-3 75.6 93.0 0.295 WT-4 106.6 93.0 0.350 Typically, the tensile capacity of a specimen under pure tension is a function of the amount and strength of steel if the steel is continuous. All four specimens exceeded the service level load. However, only two of the specimens, WT-1 and WT-4, exceeded the theoretical tensile capacity. Because the amount of steel was not varied among the test specimens, the tensile capacity must be attributed to the interaction between the concrete and steel as well as the steel arrangement. The tensile capacity of WT-2, which had a decrease in f’c from 10 to 7 ksi, was 4.6% less than the expected capacity based on nominal properties, and the tensile capacity of WT-3, which had a decrease in joint overlap length from 6 to 4 in., was 18.7% less than the expected capacity based on nominal properties. In the joint zone, the staggered U-bars tied with two lacer bars created a truss-like model. This truss model can also be considered a strut-and-tie model where the compression in the concrete represents the strut and the tension in the reinforcement represents the tie. As a test specimen reaches failure, the load will begin to decrease while the displacement continues to increase. During the WT-4 test, the actuators did not display this behavior. The actuators would reach about 92 kips, and then the load and displacement would level off due to a problem with the system hydraulic pressure. After the system was back in full operating mode, the tension test was completed, but the only data collected beyond 92 kips were the forces and displacements of the actuators, because the other instrumentation had been removed anticipating specimen failure. Once all the transverse joint tests were completed, load versus deflection curves were plotted. The LMT, placed at the bottom of the specimen, recorded the total deflection of the specimen under the tensile loading. Specimens had similar slopes after cracking of the concrete. As the loading increased beyond theoretical tensile capacity, the concrete began to form large cracks and crumble at the joint zone. To avoid damaging the deflection measuring device below the joint zone, the LMT was removed as the loading caused this crumbling effect. Figure 9.2.20 displays the load versus deflection curves of all four tension specimens prior to the removal of the LMT’s. In the case of WT-3 where the joint overlap length was reduced from 6 to 4 in., the specimen was observed to fail at a load that was almost 20% less than the expected failure load of 93 kips.

316 Figure 9.2.20: Load versus deflection Tensile Specimen Behavior (Transverse Joint Behavior) For all specimens, the first visible surface cracks developed in the transverse direction and were located outside the joint zone. As the tension loading increased, transverse cracks continued to appear in various locations outside the joint zone. For WT-1, the first cracks to appear were transverse cracks evenly spaced along the length of the specimens. The transverse cracks initially were found only on the surface of the concrete, and as the loading progressed the cracks propagated through the entire thickness of the specimens. For WT-2, the first transverse crack to develop at the joint occurred at approximately 28 kips, which was at about 30% of the nominal tensile capacity. For WT-3, the first transverse crack occurred at nearly 34 kips, at about 45% of the nominal tensile capacity. For WT-4, the first transverse crack to develop near the joint occurred at 54 kips, at about half of the nominal tensile capacity. All initial transverse cracks occurred on the side of the specimen that had a 2 in. cover measured from the reinforcement surface to the surface of concrete (the cover on the other face was 1 in. as shown in Fig. 9.2.14). As testing progressed, the cracks began to form throughout the thickness of the specimen. For WT-1, additional loading produced longitudinal cracks that appeared above the main longitudinal reinforcement in the specimen. These longitudinal cracks appeared above the longitudinal reinforcement that comprised the lightly reinforced half of the specimen, or the top half of the specimens in this particular set-up. Longitudinal cracks began forming inside the joint zone at the following loads: 60 kips for WT-2, 44 kips for WT-3, and 70 kips for WT-4. The longitudinal cracks formed above the longitudinal reinforcement in the specimen, which relates to the longitudinal reinforcement that crosses the transverse joint in the deck of precast bridge deck system.

317 Diagonal cracks appeared in the joint as the specimens approached capacity. These diagonal cracks propagated toward the first transverse cracks that developed in the joint. The concrete could be easily removed from the specimen where the diagonal and transverse cracks met. The crack patterns at tensile failure for WT-1, WT-2, WT-3, and WT-4 can be seen in Figure 9.2.21. The lacer bars provided confinement of concrete within the joint and served as restraints for the U-bars. The lacer bars allowed ductile failure in all four specimens (compared to the behavior of the headed bar specimens discussed earlier). An example of the deformation of the lacer bars can be seen in Figure 9.2.22. Please note that the location of the lacer bar in reference to the U-bars; the lacer bar also serves as bearing to U-bars. These “bearing” forces caused the lacer bar to bend. This interaction between the lacer bar and U-bars helps to explain the more ductile failure mode observed in the U-bar tests.

318 (a) WT-1 (b) WT-2 (c) WT-3 (d) WT-4 Figure 9.2.21: Tensile cracks at failure

319 Figure 9.2.22: Deformation of lacer bar Tensile Crack Widths at Service Level Loading (Transverse Joint Behavior) As calculated in Section 9.1, the tensile service load for the U-bar detail was determined to be 55.1 kips if the yield strength of reinforcement was 75.0 ksi. For the tension tests, the crack widths were measured using a crack comparator, as shown in Figure 9.2.23. However, the crack widths inside the joint zone were measured incrementally as for the flexural tests. To determine the crack widths at service-level loading, the values were interpolated as described earlier. For each specimen, the first transverse cracks that developed in the joint zone, as discussed above, were the ones measured throughout the tension tests for crack widths. The crack widths for WT-2 and WT-4 at service level were 0.020 and 0.008 in., respectively. For WT- 3, the crack width at service- level loading had exceeded the limits of the comparator of 0.050 in. WT-3, with the joint overlap length of 4 in., created the largest crack width at service-level loading.

320 Figure 9.2.23: Crack comparator used to measure tension crack widths Strain Gage Data (Transverse Joint Behavior) In Figure 9.2.24, the total applied force versus strain curves for each gage location were plotted for the transverse joint tests. For the transverse joint tests, the data provided by the strain gages was not as consistent in proving increasing strain in the reinforcement. Also, the strains were smaller at each gage located further from the bearing surface in the joint. According to the strain data, only a couple of U-bars in WT-2 and WT-4 yielded at failure which was observed at the strain gage closest to the bend. As in the longitudinal joint tests, the data from the transverse joint tests were consistent in showing that the reinforcement developed large strains where diagonal cracks occurred.

321 (a) Gage 2-1 (b) Gage 2-2 (c) Gage 2-3 (d) Gage 3-1

322 (e) Gage 3-2 (f) Gage 3-3 (g) Gage 4-1 (h) Gage 4-2

323 (i) Gage 4-3 (j) Gage LB 1-2 Figure 9.2.24: Force versus rebar strain for transverse joint tests 9.2.4. Conclusions for Phase II Tests The purpose of the research in Phase II was to test variations in U-bar joint details to identify the most influential variables. Once the significance of the variables was identified, the best performing U-bar detail could be fabricated for large-scale jointed tests used to develop the design guidelines and details for both longitudinal and transverse joints in precast bridge deck systems. The longitudinal and transverse joint specimens with a U-bar joint detail were tested in flexure and tension to represent the behavior of a precast bridge deck system. Based on the testing results of WB-1 and WT-1 of the Phase I test series, the variables considered in Phase II included concrete strength, joint overlap length, and spacing between the U-bar reinforcement. A summary of the tested variables is given in Table 9.2.8. Table 9.2.8: Summary of tested parameters Specimen ID Testing Parameter Concrete Strength Bar Spacing Joint Overlap Length (ksi) (in) (in) WB-1 WT-1 Original Test Specimen 10 4.5 6 WB-2 WT-2 Decreased f'c 7 4.5 6 WB-3 WT-3 Decreased Joint Overlap Length 10 4.5 4 WB-4 WT-4 Increased Longitudinal Reinforcement Spacing 10 6.0 6 By reducing the concrete strength, both the flexural and tensile capacities were reduced. When decreasing the joint overlap length from 6 to 4 in., the crack widths were observed to be significantly larger, the

324 flexural capacity was decreased by 17.8%, and the tensile capacity was decreased by 18.9%. Increasing the spacing of the U-bar reinforcement from 4.5 to 6 in. did not change the behaviors of longitudinal and transverse joints very much in terms of crack width, flexural capacity, or tensile capacity. In summary, to provide adequate ductility without significant loss of strength at ultimate, the joint overlap length should not be less than 6 in. where using #5 joint reinforcement. The #4 lacer bars were observed to provide restraint to help facilitate anchorage of the U-bar details for the joint zones in both flexure and in tension, and should be included in the joint detail. It is important that the lacer bars be located at the bearing face of the U-bar. 9.3. Conclusions Chapter 9 presents the results of a study that assessed potential alternate details for longitudinal and transverse joints based on constructability, performance and cost. In the Phase I test series, six reinforced concrete specimens were tested. Three of the specimens represented longitudinal joint connections (flexural specimens) and three represented transverse joint connections (tension specimens). The three joint details investigated included (1) lapped headed reinforcement, (2) lapped U-bar reinforcement fabricated with deformed wire, and (3) lapped U-bar reinforcement fabricated with stainless steel. The three specimens tested in flexure were subjected to forces that would be experienced in a longitudinal deck joint, and three specimens tested in tension were subjected to forces that would be experienced in a transverse joint over an interior pier. The capacities of the joint details were used for comparison and the selection of the best performing joint detail. All joint details produced adequate capacities and ductility in both the tension and flexural tests. Specimens with U-bar details and headed bar details both produced a capacity corresponding to their respective nominal design yield strengths. Because the U-bars had a higher nominal design yield strength (i.e., 75 ksi) than the headed bars (i.e., Grade 60), specimens containing the U-bar joint detail produced the largest capacities in both the bending and tension tests without compromising ductility. Smaller crack widths at service-level loading were also produced by the U-bar detail when compared to the headed bar detail. The constructability and reinforcement costs of the joint details were also compared. The U-bar detail created a less congested joint, which made it the easiest to construct. The bearing heads of the headed bar detail require more space due to the larger diameter of the rebar heads. This extra space reduces construction tolerances and could therefore cause problems in placement of precast deck components. The U-bars can also be easily tied together to form a rebar cage, which would allow for easy construction in the precast yard when compared to the two single layers of reinforcement in the headed bar detail. The lowest material cost was the conventional rebar used for the headed bars. The material costs were competitive between the conventional rebar used in the headed bars and the deformed wire reinforcement. The stainless steel reinforcement had the highest cost. After consideration of capacity, service level crack widths, constructability, and cost, the U-bar detail, with an overlap length of 6 in., a rebar spacing of 4.5 in. and two transverse lacer bars, constructed of deformed wire reinforcement was recommended for the Phase II tests.

325 In the Phase II tests, another six specimens with the 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. Based on results of the Phase II tests, the following conclusions can be made. By reducing the concrete strength, both the flexural and tensile capacities are reduced. When decreasing the joint overlap length from 6 to 4 in., the crack widths were significantly enlarged, the flexural capacity was decreased by 17.8%, and the tensile capacity was decreased by 18.9%. Increasing the spacing of the U- bar reinforcement from 4.5 to 6.0 in. was not observed to significantly change the behaviors of longitudinal and transverse joints in terms of their crack widths, flexural capacities, and tensile capacities. In order to provide adequate ductility without significant loss of strength at ultimate, the joint overlap length should not be less than 6 in. and #4 lacer bars should be provided to enhance the mechanical anchorage of the U- bars.