Design and Construction of Deck Bulb Tee Girder Bridges with UHPC Connections (2022)

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.

56 Experimental Testing 3.1 Introduction Experimental testing was performed on the UHPC longitudinal joint and the continuity con- nection. This was to verify analytical results and determine differences that cannot always be modeled. 3.2 Testing of Longitudinal Joint The testing of the longitudinal joint involved two series of tests. Each series contained a system of three PCEF 39-in.-deep girders with a 30° skew. The first series of tests involved applying a temperature gradient to the system, followed by fatigue loading in combination with a thermal gradient. The second series of tests involved differential camber between the girders prior to UHPC placement into the joint. The differential camber was forced out of the system while the UHPC was placed. Fatigue loading was then applied to the system without a thermal gradient. 3.2.1 Series 1 Tests Figure 3.1 shows a top view of the first three girders in place for thermal and fatigue testing. The 30° skew can be seen on the left side of the figure. An insulated box was fabricated above the girders so heating could be applied to generate a thermal gradient. Figure 3.2 shows the end view of the initial framing for the insulating box and Figure 3.3 shows a top view. 3.2.1.1 Thermal Testing with Open Joints Several heating test runs were performed with the high-capacity heaters in November and December 2020. Figure 3.4 shows the surface temperature of the top flanges of the girders at various locations over a heating cycle on November 20. The locations of the thermistor sensors on the top surface and the definition of girders 1, 2, and 3 are provided in Figure 3.5. As shown in Figure 3.4, it took approximately 5 hours to raise the surface temperature of the flanges by 29–48°F. The differences in temperatures at the various locations was due to placement of thermistor sensors relative to heat sources. Figure 3.4 also shows the cooling of the flange surfaces after the heaters were shut down. The sensors embedded at the midspan of each girder were also monitored during the heat cycle test runs. Sensor locations are shown in Figure 3.6. The temperature profile for each girder at the beginning of the heat test run is shown in Figure 3.7. As expected, the temperature was approximately uniform through the depth of the girders from being in the laboratory environ- ment. Note that the average temperature from sensors in the top flanges was used for the 2-in. C H A P T E R 3

Experimental Testing 57   Figure 3.1. Top view of girder layout. Figure 3.2. End view of insulated box framing. Figure 3.3. Top view of insulated box.

58 Design and Construction of Deck Bulb Tee Girder Bridges with UHPC Connections Figure 3.4. Surface temperature. Figure 3.5. Surface thermistor locations. depth in all profiles. Figure 3.8 shows the temperature profiles for the three girders at 2:21 p.m., 5½ hours after heating the top surface initiated. The top surface of the girders (assumed to be the air temperature) showed an increase in temperature of approximately 35°F, the internal temperatures of the flanges 2 in. from the surface showed approximately a 15°F increase, and there was a minor increase in temperature at the top flange/web interface (10¾ in. from the top surface). The remaining depth of the girders did not increase in temperature. The temperature gradient obtained in this heating cycle was similar to that specified in AASHTO 2020 Article 3.12.3; the temperature gradients specified for zones 1–4 are also shown

Experimental Testing 59   Figure 3.6. Internal sensor locations. Figure 3.7. Temperature profiles, 8:49 a.m. in Figure 3.8. The top temperature is not quite at the values specified in AASHTO, which range from 38°F to 54°F depending on the zone. However, the drop in temperature with depth is simi- lar. According to AASHTO, the drop in temperature from the top to 4 in. into the girder ranges from 40°F to 29°F. From Figure 3.8, the temperature drop in the first 4 in. is approximately 25°F. In addition, AASHTO specifies the temperature gradient at 16 in. from the top should be 0 for girders 16 in. in depth or greater. The girders’ temperature gradient at this depth is approximately 4°F. Figure 3.9 shows the temperature profiles of the girders at 4:51 p.m., 2½ hours after heating stopped. The top surfaces of the girders cooled faster than the girders at a 2-in. depth. This cooling of the top surface will likely cause transverse contraction of the top flange and could create critical transverse tensile stress in the joint. The cambers for each of the girders were also monitored during this thermal cycle and are shown in Figure 3.10. The cambers of girders 1 and 2 were identical; girder 3 showed a larger

60 Design and Construction of Deck Bulb Tee Girder Bridges with UHPC Connections Figure 3.8. Temperature gradient, 2:21 p.m. Figure 3.9. Temperature profiles, 4:51 p.m. camber. Intuitively, it was thought that differences in camber may occur between the outside girders (1 and 3) compared to the middle girder (2) due to the middle girder being more con- tained within the insulating box. The minor difference in camber may be due to a slightly uneven heating of the girders or slight differences in girder properties. Although insulation was placed in the open joints, there may have been some leakage leading to uneven heating. It was useful to have information on the movement of the unrestrained joint during the thermal cycle in order to compare to data with the joint restrained from placement of the UHPC. Therefore, instruments were installed across the joints to measure displacement

Experimental Testing 61   of the unrestrained joints due to the thermal cycle. Figure 3.11 shows the location of the vibrating wire gauges used to measure the joint displacements. Gauges JT and JB were placed across the joint between girders 1 and 2 on the top and bottom of the flange, respectively. JT and JB were perpendicular to the joint. The ST and SB gauges were placed along the skew of the bridge across the joint between girders 2 and 3 on the top and bottom of the flange, respectively. Gauge U was used for temperature corrections of the other gauges and was not attached to the girders. Figure 3.10. Girder camber during thermal cycle. Figure 3.11. Joint displacement instrumentation.

62 Design and Construction of Deck Bulb Tee Girder Bridges with UHPC Connections Figure 3.12 shows the results of the joint gauges during a thermal cycle run on November 20. As shown by Figure 3.12, the top of the joint contracted more than the bottom. This was expected since the top of the girders was warmer, resulting in more closing of the joint. The transverse gauges also showed higher movements compared to the gauges placed along the skew. This was likely due to the skew but could also have been affected by a slight difference in girder cambers. The gauges reached their measurement limitations toward the end of the test. However, this was not expected to be an issue once the joints were restrained by placement of the UHPC. 3.2.1.2 UHPC Placement in the Longitudinal Joints On December 20, 2020, the girders were heated from approximately 11 a.m. until about 9:30 p.m. Due to safety concerns about monitoring, the heating system was turned off for the night and the girders began to cool. The girders were heated again at approximately 4:30 a.m. on December 21 and continued until 8:30 a.m. This process was performed to have the girders as warm as possible during the placement of the UHPC. For the safety of personnel, the heating system was turned off while a fine water mist was applied to each joint and top forms screwed back into the girder flanges. Mixing of the UHPC began as the misting of the joints and place- ment of the top forms was completed. High shear mixers were used in the mixing of the UHPC. The mixers had a capacity equal to that required for one full longitudinal joint. Ductal JS1000 manufactured by LafargeHolcim was the UHPC material used for this project. This product is commercially available and has been used in numerous projects throughout Figure 3.12. Joint strains.

Experimental Testing 63   the United States. The research team (RT) was also most familiar with this UHPC product to ensure proper placement and to reduce any issues that might arise from mixing and placement. Table 3.1 provides some of the properties of Ductal JS1000, from the LafargeHolcim website. In other projects, the RT has found that compressive strength reaches 14 ksi in 2–3 days instead of the 4 days listed in Table 3.1. It should also be noted that the UHPC sat in the laboratory for approximately 1 year prior to use due to delays caused by the COVID-19 shutdown. The UHPC seemed to have a slightly more doughy consistency during placement than the RT was accustomed to. The UHPC mix proportions used for this project are shown in Table 3.2. Fresh UHPC proper- ties are shown in Table 3.3. Slightly higher amounts of water were used due to the premix being approximately 1 year old. A LafargeHolcim representative was present and supervised the mixing; he indicated that the age of the material and the additional water did not take the UHPC out of specification. Density 150–160 lb./ft.3 (2,400–2,565 kg/m3) Flow 7–10 in. (175–250 mm) diameter without visible sign of fiber segregation Working Time/Set Time approx. 120 min./15–18 hr. Compressive Strength(2) >14 ksi (100 MPa)(3) at 4 days (4,5) >21 ksi (150 MPa) at 28 days Tensile Strength(6) >725 psi (5 MPa) at 28 days Modulus of Elasticity >6,500 ksi (45 GPa) at 28 days Long-Term Shrinkage <800 microstrain at 28 days Chloride Ion Penetrability <250 coulombs (very low) at 56 days Freeze-Thaw Resistance >96% RDM at 300 cycles Source: https://www.ductal.com/sites/ductal/files/atoms/files/2_ductal_js1000.pdf. (1) Field results may differ depending on mixing/test methods, equipment used, temperature, and site/curing conditions. (2) Compression tests are performed on 3 in. × 6 in. (75 mm × 150 mm) cylinders with ends ground flush prior to testing. (3) 14 ksi (100 MPa) is the typical minimum compressive strength required before application of design live load for most closure pour applications; consult the Engineer or project specifications to verify. (4) 4 days or less is typical when the ambient curing temperature is greater than 60°F (16°C). For colder temperatures, an accelerating admixture may be required to obtain 14 ksi (100 MPa) in 4 days. (5) For 14 ksi (100 MPa) compressive strength in 12–36 hours, consider using rapid-set Ductal JS1212. (6) This test measures the sustainable post-cracking direct tension strength of a mix with 2% by volume steel fibers. Table 3.1. Ductal JS1000 published properties.(1) Constituent Quantity (lb./yd. 3) Batch 1 (Joint 2-3) Batch 2 (Joint 1-2) UHPC Premix Water Fibers Superplasticizer 3,690 216 264 50.4 3,690 219 264 51.2 Table 3.2. UHPC mix. Batch Time of Mixing Lab Temp (°F) UHPC Temp (°F) Static Slump (in.) Dynamic Slump (in.) 1 10:33 a.m. 51 70 8.75 9.35 2 12:49 p.m. 51 71 8.75 9.25 Table 3.3. UHPC fresh properties.

64 Design and Construction of Deck Bulb Tee Girder Bridges with UHPC Connections e UHPC was transferred from the mixer in buckets into chimneys (orange and white buckets in Figure 3.13) at approximately the third points along the joint. is allowed the UHPC to ow in each direction. A hydraulic head was maintained in the chimneys to ensure the entire joint was lled. Aer UHPC completely lled both joints, insulating panels were reinstalled to the framing. e insulating box was closed and reheating of the girders was initiated. On December 24, the forms were removed from the joints. Both joints turned out well. Figure 3.14 shows the top of the joint between girders 2 and 3 at the west end. e sealer material on each side of the joint Figure 3.13. Top of anges before UHPC placement. Figure 3.14. UHPC joint between girders 2 and 3.

Experimental Testing 65   was used to ensure the UHPC did not leak out of the joint and gave an approximately ⅛-in. height above the joint. UHPC is often placed approximately ⅜–⅛ in. higher than the adjoining members in the field and ground flush. Cylinders 3 × 6 in. were cast during the UHPC placement. Results of compression testing the cylinders are shown in Table 3.4. The 28 (thermal) results were cylinders that were stored in the insulated box built on the girders. Several thermal cycles were performed while these cylinders were inside the box. The purpose of using these cylinders was to match the conditions of the UHPC in the joint. Though the results of the compression testing are reasonable, it is believed the compressive strength was likely slightly higher. Great care must be taken with the preparation of the cylinder ends prior to testing. The high compressive strengths UHPC do not allow the use of capping compounds or neoprene caps for testing. The ends need to be flat and square with the cylinder. Typically, the RT has used a milling machine to achieve the desired cylinder-end preparation. Unfortunately, access to the milling machine was not available for the 7- and 14-day tests because Ohio University closed during the holidays. It was later determined the milling machine is no longer available. This was determined too late to make other accommodations. Therefore, testing for the 28-day samples used the same preparation technique. Great care was taken during saw cutting to achieve the best possible end preparation given the available equipment. Samples were tested at 120 days by Lafarge to alleviate the preparation issue. The 120-day test results were performed to coincide with fatigue and static load testing. 3.2.1.3 Thermal Behavior during UHPC Placement Figure 3.15 shows the temperature gradient from the average temperatures of the gauges near the top surface of the girder (gauges 5 and 6 in Figure 3.6) compared to the internal gauge near the bottom surface of the girder (gauge 1 in Figure 3.6). Figure 3.15 shows the heating process that occurred on December 20 and the cooling through the night as the girders began to lose some of the temperature gradient. It also shows the reheating of the girders in the early morning before UHPC mixing and placement, the heating loss while the UHPC was placed, and the reheating which began again at approximately 4:30 p.m. The heating system was shut down at 8 p.m. Multiple heating cycles were also run at later dates. As shown in Figure 3.16, girders 2 and 3 had similar camber and girder 1 had less, even though girder 2 had the largest thermal gradient (see Figure 3.15). Rebar strains increased positively with an increase in temperature. The strains across joints 12 and 23 were positive prior to UHPC place- ment, as shown in Figures 3.17 and 3.18, respectively. The labeling for the sensors across the joints is provided in Figure 3.19. The positive strain results may have been due to the top formwork being in place and differences in the girder cambers. The sensors were removed during top form removal, joint misting, top form installation, and UHPC placement to avoid damage. 3.2.1.4 Thermal Behavior the Day after UHPC Placement A complete heating cycle was run on December 22, 2020, the day after UHPC placement, with the formwork still in place. The temperature gradient for December 22 is shown in Figure 3.20. Age (days) Average Compressive Strength (psi) 7 15,465 14 17,348 28 21,749 28 (thermal) 22,384 120 26,039 Table 3.4. Measured UHPC compressive strength.

66 Design and Construction of Deck Bulb Tee Girder Bridges with UHPC Connections Figure 3.15. Temperature gradient during UHPC placement, 12/21/2020. Figure 3.16. Girder camber, 12/21/2020.

Experimental Testing 67   Figure 3.17. Strains across joint 12, 12/21/2020. Figure 3.18. Strains across joint 23, 12/21/2020.

68 Design and Construction of Deck Bulb Tee Girder Bridges with UHPC Connections Figure 3.19. Vibrating wire gauges across the joints. Figure 3.20. Temperature gradient the day after UHPC placement, 12/22/2020.

Experimental Testing 69   The temperature gradient was determined from the average temperatures of the gauges near the top surface of the girder (gauges 5 and 6 in Figure 3.6) compared to the internal gauge near the bottom surface of the girder (gauge 1 in Figure 3.6). The camber of all girders was similar, as shown in Figure 3.21, with girder 3 having the largest camber. This occurred even though girder 2 had the largest thermal gradient. However, the difference in cambers between girders was reduced compared to the UHPC placement day. Overall, the similar behavior to the day of the UHPC placement implies the UHPC had not gained sufficient strength for the girders to behave as a system. All reactions at the ends of the girders were positive, as shown in Figure 3.22. Rebar strains within the joints increased positively with an increase in temperature. The strains across the tops of the joints were now negative instead of positive, as shown in Figure 3.23. The strain across the bottom of joint 23 (gauge J23-B) was positive. 3.2.1.5 Thermal Behavior 2 Days after UHPC Placement On December 23, 2020, girder 2 had the largest thermal gradient and girders 1 and 3 had similar lower gradients, as shown in Figure 3.24. Note that the data for thermal gradients was only taken during heating and not cooling. The camber of girder 2 now became the largest, while girders 1 and 3 were approximately the same, as shown in Figure 3.25. Girder 2 reactions were reduced under heating while the other girders showed an increase in their reactions (see Figure  3.26); this implies the girders were beginning to behave as a system. Rebar strains increased positively with the increase in temperature. The strains across the joints remained negative during the increase in temperature. Figure 3.21. Girder camber, 12/22/2020.

70 Design and Construction of Deck Bulb Tee Girder Bridges with UHPC Connections Figure 3.22. Girder reactions, 12/22/2020. Figure 3.23. Joint strains, 12/22/2020.

Experimental Testing 71   Figure 3.24. Temperature gradient 2 days after UHPC placement, 12/23/2020. Figure 3.25. Camber, 12/23/2020.

72 Design and Construction of Deck Bulb Tee Girder Bridges with UHPC Connections 3.2.1.6 Thermal Behavior 5 Days after UHPC Placement and Formwork Removal On December 26, 2020, girder 2 had the largest thermal gradient and girders 1 and 3 had similar lower gradients, as shown in Figure 3.27 for this thermal cycle. The camber of girder 2 remained the largest, while girder 1’s camber was slightly more than that of girder 3, as shown in Figure 3.28. The reaction for girder 2’s west end showed an increase at the start, shown in Figure 3.29. Both reactions for girder 2 then decreased under heating while the other girders showed an increase in their reactions. Upon cooling, the girder reactions began to coincide. This implied the girders were beginning to behave as a system. Rebar strains increased positively with an increase in temperature. The strains across the joints remained negative during the increase in temperature but became positive upon the drop in temperature, as shown in Figure 3.30. 3.2.1.7 Thermal Behavior over a Month after UHPC Placement and Formwork Removal Unlike on December 26, 2020, on January 25, 2021, girder 3 had the largest thermal gradient and girders 1 and 2 had similar lower gradients, as shown in Figure 3.31 for this thermal cycle. The camber was typical compared to other tests but girder 3’s was slightly higher than girder 1’s, shown in Figure 3.32. The reaction for girder 2’s west end showed an increase at the start. Both reactions for girder 2 then reduced under heating while the other girders showed an increase in their reactions (see Figure 3.33). This was very similar to the behavior on December 26, 2020. Upon cooling, the girder reactions began to coincide. Rebar strains increased positively with an increase in temperature. However, the strain at joint 12-1 near the west end showed an unusual large increase (see Figure 3.34), which may have signified cracking. The strains across the joints became negative during the increase in temperature but then became positive upon the removal of the temperature, as shown in Figure 3.35. In addition, the gauge at location J12-2 increased positively with the temperature increase and then dropped off during cooling. This was not typi- cal behavior compared to previous test data and may have signified cracking. Figure 3.26. Girder reactions, 12/23/2020.

Experimental Testing 73   Figure 3.27. Temperature gradient, 12/26/2020. Figure 3.28. Camber, 12/26/2020.

74 Design and Construction of Deck Bulb Tee Girder Bridges with UHPC Connections Figure 3.29. Girder reactions, 12/26/2020. Figure 3.30. Strains across joints, 12/26/2020.

Experimental Testing 75   Figure 3.31. Temperature gradient, 1/25/2021. Figure 3.32. Camber, 1/25/2021.

76 Design and Construction of Deck Bulb Tee Girder Bridges with UHPC Connections Figure 3.33. Girder reactions, 1/25/2021. Figure 3.34. Joint 12 rebar strains, 1/25/2021.

Experimental Testing 77   3.2.1.8 Thermal Behavior over a Month after UHPC Placement and Formwork Removal The thermal cycle on January 27, 2021, was very similar to that on January 25. However, unlike December 26, 2020, and January 25, 2021, girders 1 and 3 had larger thermal gradients than girder 2, as shown in Figure 3.36. The camber data was very similar to that of January 25. The reactions for girder 2 were what was expected. Both reactions for girder 2 reduced under heating while the other girders showed an increase in their reactions, as depicted in Figure 3.37. Upon cooling, the girder reactions began to coincide. Rebar strains increased positively with an increase in temperature. However, similar to January 25, the strain at location joint 12-1 near the west end showed an unusual large increase. The strains across the joints became negative during the increase in temperature but became positive upon the removal of the temperature. In addition, the gauge at location J12-2 again increased positively with the increase in temperature and then dropped off during cooling. 3.2.1.9 Thermal Behavior over a Month after UHPC Placement and Formwork Removal The final thermal cycle on January 29, 2021, was the largest thermal gradient generated, as shown in Figure 3.38. For comparison purposes, the AASHTO zones 1–4 thermal gradients are shown in Figure 3.39. As can be seen, the thermal gradient generated during this cycle exceeded the AASHTO thermal gradients for all zones. The generated gradients were close to AASHTO at the 2-in. and 10-in. depths from the top surfaces but exceeded AASHTO at the 4-in. and 16-in. depths. The camber and girder reaction data were very similar to that of the previous tests. Rebar strains were also similar to January 25 and 27, with the strain at joint 12-1 near the west end show- ing an unusually large strain increase. Strains across the joints were also similar to January 27, Figure 3.35. Strains across joints, 1/25/2021.

78 Design and Construction of Deck Bulb Tee Girder Bridges with UHPC Connections Figure 3.36. Temperature gradient, 1/27/2021. Figure 3.37. Girder reactions, 1/27/2021.

Experimental Testing 79   with the strain at location J12-2 again increasing positively with the increase in temperature and then dropped off during cooling. 3.2.1.10 Cyclic Live Loading After the thermal cycles were completed, cyclic live loading was applied to the system from a load of 0 kips to 70 kips for 100,000 cycles. Figure 3.40 provides a drawing of the load setup without the insulating box in place. Figure 3.38. Temperature gradient, 1/29/2021. Figure 3.39. AASHTO design and measured midspan peak temperature gradients, 1/29/2021.

80 Design and Construction of Deck Bulb Tee Girder Bridges with UHPC Connections (a) End view with midspan loading also shown: test 1 with girders 1–3 (b) Elevation view Figure 3.40. Longitudinal flange connection loading configuration. After the initial 100,000 cycles were completed, thermal loading was also applied during the cyclic live loading. Figures 3.41 and 3.42 show the load frame above the insulating box and the hydraulic cylinder and load cell, respectively. A total of 85,000 to 100,000 cycles of live load were applied during one thermal cycle. At the end of each thermal cycle, cyclic loading was stopped, and a 70-kip load was applied statically to monitor potential degradation of connection integrity in terms of load distribution, differential joint movement, and/or deflection of individual girders. This was necessary in order to read instrumentation, which could not be read during the approx- imately 2 Hz cyclic loading. By separating the thermal and cyclic load at the beginning, the effect of each, if any, was hoped to be seen. The process continued until 1,000,000 live load cycles had been applied. The results of the static load tests were consistent over the period of cyclic loading. Figure 3.43 shows the camber of the girders at various locations under static loading on April 26, 2021. As expected, girder 2 had the largest value at midspan, as it was directly loaded. The reactions are shown in Figure 3.44. Girder 2 had the largest reactions and were similar at both ends; the obtuse corners (girder 3 west end and girder 1 east end) were the next largest, and the acute corners had the lowest reactions. The total of the reactions matched the applied load of 70 kips. Figure 3.45 shows the strains on the instrumented rebar in joint 12; location J12-1 shows a large

Experimental Testing 81   Figure 3.41. Load frame with insulated box. Figure 3.42. Hydraulic cylinder and load cell for loading. tensile strain that differs from the other rebar strain measurements. This may have indicated cracking within the joint. Figure 3.46 shows the strains across the joints. As expected, the strains near midspan were the highest and the strains measured at the bottom of the joint at midspan showed tensile strains. 3.2.1.11 Joint Flooding After the thermal and live loading cycles were completed, no cracks were visible. In addi- tion, the nondestructive testing pulse velocity measurements taken throughout the testing did

82 Design and Construction of Deck Bulb Tee Girder Bridges with UHPC Connections Figure 3.43. Camber, 4/26/2021. Figure 3.44. Girder reactions, 4/26/2021.

Experimental Testing 83   Figure 3.45. Joint 12 rebar strains, 4/26/2021. Figure 3.46. Strains across joints, 4/26/2021.

84 Design and Construction of Deck Bulb Tee Girder Bridges with UHPC Connections not indicate any cracking. However, the RT had concerns about cracking based on some of the instrument strain reading jumps. Therefore, dams were constructed and sealed around the top surface of the joints. The dams were then flooded with water (Figure 3.47). The flooding revealed severe leakage and cracking of the north joint between girders 1 and 2 (joint 12). The amount of water on the lab floor beneath joint 12 is shown in Figure 3.48. The leakage was along the inter- face (see Figure 3.49) and from longitudinal cracks within the UHPC (see Figure 3.50). It should Figure 3.47. Flooding of joint. Figure 3.48. Joint 12 leakage.

Experimental Testing 85   Interface crack Figure 3.49. Interface leakage. be noted that foam tape was used between the forms and the girders to ensure a tight seal. This resulted in UHPC extending past the joint interface along some of the edges, as shown in the figures. The South joint between girders 2 and 3 (joint 2-3) had a few small cracks (Figure 3.51). A crack map is shown in Figure 3.52. The pulse velocity system data did not indicate any cracking. It is believed this was due to the fibers allowing the pulse wave to travel through the UHPC with or without cracking. The cracking could have occurred for several reasons or combinations of them. The interface could have dried out during delays in placing the UHPC. However, cracking existed within the UHPC in addition to along the interface. Shrinkage of the UHPC during curing could have resulted UHPC crack Figure 3.50. UHPC cracking.

86 Design and Construction of Deck Bulb Tee Girder Bridges with UHPC Connections Overlapping UHPC UHPC crack Figure 3.51. Joint 2-3 UHPC cracking. in cracking along the interface and within the UHPC. Increases in strain measurements during thermal loading may indicate cracking during the thermal cycle. The thermal cycle loading was severe and may not often be experienced during UHPC placement. Lastly, the UHPC was in excess of 1 year old due to COVID-19 project delays. Flow of the UHPC did appear to slow for the joint experiencing more cracking. 3.2.2 Series 2 Tests The purpose of the second series of tests was to investigate the effects of differential camber. Girder 2 was removed from the series 1 test assembly. Girders 1 and 3 swapped positions so their outside flanges not used in the series-1 testing could be used, and girder 4 was then placed in between girders 1 and 3 to create the new system. Girder 4 had six additional prestressing strands to ensure differential camber in comparison to girders 1 and 3. The measured girder cambers at midspan are provided in Table 3.5. Figures 3.53 and 3.54 show the differential camber between the girders. The differential camber between girders was removed by loading the middle girder (girder 4) to approximately 90 kips, as shown in Figure 3.55. This loading resulted in a downward deflec- tion of 0.5 in. The hydraulic cylinder was locked off to hold the girder in position. The loading actually slightly decreased from seepage of the hydraulic pressure, creep of the girder, and settle- ment of the system. However, the downward deflection remained at approximately 0.5 in. to match flange heights, as shown in Figure 3.56. Formwork was then attached to the girders and the UHPC was placed, as shown in Figure 3.57. The UHPC mix proportions and fresh properties are provided in Tables 3.6 and 3.7, respectively. Table 3.8 provides the compressive strength of the UHPC. Three days after the UHPC placement, the top forms were removed. Dams were erected around the joints and waterproofed. On the fourth day after UHPC placement, the dams were filled with water and allowed sit for over 30 minutes. No cracking was observed from leakage of the joint.

Experimental Testing 87   (a) West end (b) East end Figure 3.52. Crack map. Girder Camber (in.) Differential Camber (in.) 3 1 ⅜ 4 1⅜ ½ 1 ⅞ Table 3.5. Girder cambers.

88 Design and Construction of Deck Bulb Tee Girder Bridges with UHPC Connections Figure 3.53. Differential camber. Figure 3.54. Differential camber between girders 1 and 4. Figure 3.55. Loading to remove differential camber.

Experimental Testing 89   Figure 3.56. Removal of differential camber. Figure 3.57. UHPC placement.

90 Design and Construction of Deck Bulb Tee Girder Bridges with UHPC Connections The load used to remove the differential camber was then released gradually in steps as shown by the girder reactions in Figure 3.58. The girder cambers are provided in Figure 3.59: girder 4 had the largest midspan camber of slightly more than 0.20 in. The midspan cambers of girders 1 and 3 were approximately 0.12 in. Since the cambers were measured at the bottom, this implies a slight bending of the flange and/or slight rotation of the outside girders. Upon removal of the first load increment, leakage and cracking of the north joint (joint 34) occurred. The cracking occurred within the UHPC as well as at the interface between the UHPC and the girder flanges. The south joint (joint 41) also had minor cracking. Figure 3.60 shows the cracking map. The majority of the cracking occurred on the north joint and mostly toward the east end. The UHPC was placed in the west ends of both joints first followed by placement in the east ends. The UHPC premix material was in excess of 1 year old due to delays caused by the COVID-19 shutdown. After the load was fully released from the system, cyclic testing of 72 kips was performed. A total of 1,000,000 cycles was applied to the system. At various periods throughout, the cyclic loading was stopped and a static 72-kip load was applied. Several of the static loadings also included flooding the joints. Minimal crack growth was observed (approximately no more than 1 in. growth in some of the existing cracks). Figure 3.61 provides the camber from static loading on June 16, 2021, after 100,000 cycles of loading. Figure 3.62 is the same plot on June 28, 2021, after 1,000,000 cycles of loading. The cambers are nearly identical, indicating the system did not degrade over the loading cycles. Figures 3.63 and 3.64 provide the reactions after 100,000 and 1,000,000 cycles of loading, respec- tively. The reactions are also nearly identical, again indicating that the system did not degrade. After loading was completed, cores were taken from the east end of joint 34. This location showed the larger frequency of cracking. Cores were then cut to observe the cracking within the depth of the joint. Figure 3.65 shows a core with a crack 2 in. from the bottom of the joint. The distribution of fibers appears to be good. Figure 3.66 is a magnified view of the crack. The crack width was approximately 0.014 in. (0.35 mm). Constituent Quantity ( lb./yd. 3) Batch 1 (west ends) Batch 2 (east ends) UHPC Premix 3,840 3,840 Water 215 215 Fibers 264 264 Superplasticizer 50.4 52.6 Table 3.6. UHPC mix. Batch Time of Mixing Lab Temp (°F) UHPC Temp (°F) Static Slump (in.) Dynamic Slump (in.) 1 10:33 a.m. 75 91 8.75 9.75 2 12:49 p.m. 79 91 8.50 9.50 Table 3.7. UHPC fresh properties. Age (days) Average Compressive Strength (psi) 3 10,410 7 15,507 14 17,536 28 20,024 Table 3.8. UHPC strength.

Experimental Testing 91   Figure 3.58. Girder reactions from load release. Figure 3.59. Girder cambers from load release.

92 Design and Construction of Deck Bulb Tee Girder Bridges with UHPC Connections (a) West end (b) East end Figure 3.60. Cracking map.

Figure 3.61. Camber from static loading, 6/16/2021. Figure 3.62. Camber from static loading, 6/28/2021. 

94 Design and Construction of Deck Bulb Tee Girder Bridges with UHPC Connections Figure 3.63. Reactions from static loading, 6/16/2021. Figure 3.64. Reactions from static loading, 6/28/2021.

Experimental Testing 95   Figure 3.65. Core slice with crack. Figure 3.66. Magnified crack in core slice.

96 Design and Construction of Deck Bulb Tee Girder Bridges with UHPC Connections 3.3 Continuity Joint Testing The continuity joint test consisted of two 20-ft. PCEF 39-in.-deep girders placed end to end. The width between the ends of the girders was 7.5 in. The girders were embedded approximately 2.25 in. into the diaphragm. Conventional concrete formed the bottom 33 in. of the diaphragm, and the upper portion, equivalent to the flange depth of approximately 6 in., was UHPC. The ends of the girder flanges and the top of the conventional concrete had roughened surfaces to promote bond. Figure 3.67 shows the details of the joint. The reinforcement extending in from the top flange consisted of two rows of No. 6 bars spaced at 8 in. in each flange. Staggered spacing resulted in 4-in. spacing within the joint. Two 0.6-in.-diameter strands protruded into the diaphragm from each girder. One No. 5 bar was placed in each bend of the hooked strands along the diaphragm transverse to the girders. Welded wire reinforcement (12 × 12–D3.0 × D3.0) was also placed in the diaphragm but no reinforcement crossed the interface between the UHPC and conventional concrete. Figure 3.68 shows the joint before closing of the formwork and UHPC placement. Instrumentation was installed in the joint and on the top reinforcement. Figures 3.69 and 3.70 provide the instrumentation layout. The mix design for the conventional concrete is provided in Table 3.9. Eleven days after con- ventional concrete placement, the UHPC was placed in the upper portion of the joint. Since the quantity of UHPC was relatively small, it was mixed in a mortar mixer. Unfortunately, the mixer had limited energy and tripped the circuit breaker. This required the addition of water to ensure the UHPC could be continually mixed. Tables 3.10 and 3.11 provide the UHPC mix proportions and compressive strengths, respectively. 3.3.1 Positive Moment Testing Creep and shrinkage of the girders after continuity is formed creates a positive moment effect in the joint. The magnitude of the positive moment depends on when continuity is formed Figure 3.67. Continuity detail.

Experimental Testing 97   Figure 3.68. Continuity joint reinforcement. Figure 3.69. Instrumentation in continuity joint. relative to the properties and age of the girders. It was assumed that continuity was formed at approximately 28 days and a positive moment of approximately 78 kip-ft. would be created in the joint, as noted in Table 2.22 for case 1. Hydraulic cylinders were placed 12 ft. from the face of the joint under each girder (Figure 3.71) and a frame was placed over the joint to keep it from moving upward (Figure 3.72). By applying an upward force on both sides of the joint under the girders, a positive moment was created in the joint. Figure 3.73 provides the load applied by both hydraulic cylinders. A maximum load of 7,300 lb. was applied, resulting in a positive moment of 87.6 kip-ft. The strains within the joint,

98 Design and Construction of Deck Bulb Tee Girder Bridges with UHPC Connections Embedded Strain Gauge Figure 3.70. Continuity joint top reinforcement instrumentation. Constituent Quantity (/yd.3) Cement 1,100 lb. Fine Aggregate 1,450 lb. Course Aggregate 1,596 lb. Water 352 lb. High Range Water Reducer 1,100 ml Table 3.9. Conventional concrete mix.

Experimental Testing 99   Age (days) Average Compressive Strength (psi) 3 12,105 7 14,755 Table 3.11. UHPC strength. Constituent Quantity per joint (lb./yd.3) Batch 1 (bottom) Batch 2 (midheight) Batch 3 (top) UHPC Premix Water Fibers Superplasticizer 3,690 250 250 46.7 3,690 247 246 46.2 3,690 216 264 50.5 Table 3.10. UHPC mix. Figure 3.71. Hydraulic cylinder for +M. Figure 3.72. Frame over continuity joint.

100 Design and Construction of Deck Bulb Tee Girder Bridges with UHPC Connections shown in Figure 3.74, were very small. As expected, gauge 1 near the bottom of the connection experienced tension, while the gauges in the flanges (gauges 3–5) experienced compression. No cracking was observed. 3.3.2 Negative Moment Testing Live loading results in negative moments for the continuity joint. The magnitude of the positive moment depends on when continuity is formed relative to the properties and age of the girders. From the analytical work performed, the ultimate negative moment the joint would experience was 1,152 kip-ft., as noted in Table 2.23 for case 1. A frame was placed over the east girder 12 ft. from the continuity joint. A load cell was placed between the top flange of the girder and the frame to measure the reaction. Another frame was placed 12 ft. from the continuity joint over the west girder. A hydraulic cylinder was mounted to the frame to load the west girder. The end support for the west girder was removed to create a cantilever and a negative moment in the joint. Figure 3.75 shows the setup in the lab. The self-weight of the west girder created a negative moment of approximately 230 kip-ft. Figure 3.76 shows the incremental loading of the west girder to produce the negative moment in the joint. The maximum load was 114 kips, which was equivalent to 1,311 kip-ft. The total moment applied with the self-weight was 1,541 kip-ft., well above the estimated capacity. Strains in the top reinforcement from the west and east girders embedded in the continuity joint are provided in Figures 3.77 and 3.78, respectively. At a load of 98 kips (third to last load increment), the strains in both bars at the middle of the joint (W3 and E3) showed an abrupt Figure 3.73. Continuity joint loading for positive moment.

Experimental Testing 101   Gauge 1 Gauge 2 Gauge 3 Gauge 4 Gauge 5 Figure 3.74. Continuity joint strains from positive moment. Figure 3.75. Negative moment continuity joint setup.

Figure 3.77. Strains in reinforcement from west girder. Figure 3.76. West girder loading to produce negative moment.

Experimental Testing 103   increase and indicated yielding. The load of 98 kips corresponds to a total moment of 1,357 kip-ft., including the self-weight. This moment still exceeded the predicted moment capacity of the joint. The strain in the reinforcement W2 yielded at the next load increment, but the remaining bars did not indicate yielding even on the final load increment. 3.4 Summary of Experimental Testing Based on the experimental results, the following observations and general conclusions were made: 1. Though the cracking from the series 1 system testing was not fully discovered until the end of cyclic testing, it is believed the cracking occurred early in the thermal cycles. The thermal gradient applied to the system was slightly higher than specified in AASHTO 2020 and started immediately after UHPC placement. Significant cracking occurred in the second joint pour and minimal cracking occurred in the first joint placement. This may have been the result of the UHPC gaining sufficient strength in the first joint prior to thermally generated strains and not gaining enough strength in the second joint prior to the thermally generated strains. The second joint UHPC placement also showed slightly lower flow, which could have affected performance. 2. Although cracking may have existed in the joints of the series 1 system test, load transfer continued to occur between girders under combined static and thermal load. 3. Cracking from the series 2 system testing was discovered right after the first unloading incre- ment was removed from the differential camber leveling. The cracking occurred primarily Figure 3.78. Strains in reinforcement from east girder.

104 Design and Construction of Deck Bulb Tee Girder Bridges with UHPC Connections in the east end of the joints. The UHPC was placed in the east ends last and seemed to have slower flow. 4. Though cracking of the series 2 system test occurred during removal of the differential camber leveling load, crack growth from 1,000,000 load cycles was 1 in. or less. In addition, new cracks were not formed. 5. The 7.5-in. continuity joint consisting of UHPC in the top 6 in. and conventional concrete in the remaining height of the diaphragm did not show any distress under a positive moment expected from continuity being formed at an age of 28 days. 6. The top reinforcement in the continuity joint began to yield at a negative moment in excess of the calculated negative moment capacity. Additional negative moment caused additional bars to yield and result in a capacity far exceeding the calculated capacity.