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.

83 Chapter 5 PCSSS: Large-Scale Laboratory Bridge Investigation of System Behavior 5.0 Introduction The performance of the precast composite slab span system in a global sense depends on the interaction between neighboring precast panels as well as panels in the adjacent spans. This interaction between components of the PCSSS warranted the development of two large-scale laboratory test specimens constructed and tested in the University of Minnesota Structures Laboratory. The first bridge, Concept 1, was developed as a continuous two-span system with two adjacent precast panels per span. A number of variables were investigated with this bridge, which are discussed in detail in Section 5.1.1. A portion of the Concept 1 specimen emulated one of the first implementations of PCSSS bridges in the State of Minnesota, Mn/DOT Bridge 13004 built in Center City. The second bridge, designated as the Concept 2 specimen, was a simply-supported bridge with two adjacent precast panels. The Concept 2 laboratory bridge was developed to augment the information obtained from testing of the Concept 1 laboratory bridge specimen, and is discussed in detail in Section 5.1.2. This chapter discusses the development, testing, and results of the large-scale laboratory testing. 5.1. Selection and Design of Laboratory Bridge Specimens The Concept 1 specimen which was based on the Center City Bridge with modifications as required to investigate variations in some of the parameters, was constructed as part of a Mn/DOT research project which consisted of a companion field and laboratory investigation described by Smith et al. 2008. The results of the field study are summarized in Section 3.2. At the conclusion of the Mn/DOT study, the Concept 1 laboratory bridge was made available to the NCHRP 10-71 study, which included cyclic tests to simulate fatigue, investigation of initiation of reflective cracking simulated to emulate observed crack initiation in the Center City field bridge, as well as tests to investigate the ultimate strength of the system and composite action. The Concept 2 laboratory bridge, developed at the conclusion of laboratory testing on the Concept 1 laboratory bridge, was designed to incorporate information accrued from the Concept 1 tests, as well as to investigate variations in additional parameters. Relevant design details for both specimens are discussed in the following sections. Many of the design details associated with the Concept 1 and Concept 2 large-scale laboratory bridge specimens, as well as the subassemblage specimens discussed in Chapter 6, include variations in the size and spacing of the transverse reinforcement traversing the longitudinal precast joint in the trough region created by adjacently abutted precast sections. To provide a means of qualifying the reinforcement in this region, a transverse reinforcement ratio was defined. The presence of transverse reinforcement for both transverse load transfer and reflective crack control warranted separate definitions for the reinforcement ratio in each case. In the case of transverse load transfer, the area of reinforcement considered was only that which was embedded or mechanically anchored through the precast web in the lowest region of the CIP trough such that all load developed in the reinforcement could be adequately transferred through the adjacent panels. For a given longitudinal joint, the reinforcement protruding from a single precast panel was considered (i.e., for each pair of transverse embedded bars, only one was considered in the reinforcement ratio). It was assumed that the load was transferred from the bars protruding from one panel to the bars protruding from the adjacent panel through lap splices. Furthermore, the area of

84 concrete considered in the calculation was taken as that between the top of the precast flanges and the top of the section. An illustration of the area of steel and concrete considered for the calculation of the reinforcement ratio for transverse load transfer is highlighted in yellow in Figure 5.1.1. Figure 5.1.1: Reinforcement and depth of concrete considered in the calculation of the reinforcement ratio for transverse load transfer (highlighted in yellow) The reinforcement ratio for crack control, on the other hand, considered the area of all reinforcement crossing the precast joint near the bottom of the CIP trough. Therefore, the bottom leg of the cage stirrups and all of the embedded transverse reinforcement was included in the calculation (i.e., for each pair of transverse embedded bars, both were included in the calculation because both were assumed to be effective above the longitudinal joint between the adjacent flanges). Furthermore, the area of concrete used in the calculation was only that which was located between the top of the precast flanges and the top of the precast webs. An illustration of the area of steel and concrete considered in the calculation of the reinforcement ratio for crack control is highlighted in yellow in Figure 5.1.2. It should be noted that this crack control reinforcement would only be effective in the region above the longitudinal joint between adjacent precast panels. For potential cracks that may develop at the precast web-CIP interface, only the reinforcement protruding from the precast webs would be effective. CIP PRECAST Unit Length

85 Figure 5.1.2: Reinforcement and depth of concrete considered in the calculation of the reinforcement ratio for reflective crack control (highlighted in yellow) 5.1.1. Concept 1 Laboratory Bridge The continuous two-span Concept 1 laboratory bridge specimen was selected to provide insight into the effects of restraint moment developed at the pier, load distribution, and the effect of a number of variations in parameters. Restraint moments will develop at continuous piers due to both time- dependent effects (i.e., creep and shrinkage) and thermal gradients. Because the laboratory bridge was constructed and tested within the temperature-controlled environment of the University of Minnesota Structures Laboratory, restraint moments due to thermal effects were not experimentally studied in the laboratory specimen, however an effort to induce a longitudinal curvature due to a uniformly applied thermal heat source was implemented on Span 2 of the Concept 1 laboratory bridge, which is discussed later in this chapter. Variables investigated in the two-span Concept 1 laboratory bridge are summarized in Table 5.1.1. The variables included the precast flange thickness, flange surface roughness, and reinforcement detailing, in addition to the amount of CIP deck reinforcement. Span 2 of the bridge was constructed identically to the exterior spans of the Center City Bridge, allowing for a relative comparison between observed results in the field and laboratory bridges. Span 1 of the bridge specimen was constructed with the application of the variables of interest, such as a reduced flange depth; the reduction in flange depth from 5.25 to 3 in. was introduced in an effort to reduce the distance between the transverse hooked bars and the precast flange, which would help to intercept a potential reflective crack at the precast joint region at a lower depth as well as increase the transverse moment capacity. Additionally, a smooth CIP PRECAST Unit Length

86 flange surface was used in Span 1. The smooth flange was designed to reduce the stress concentration above the flanges directly at the joint, as well as aid in the removal of the formwork during fabrication. The stirrup spacing for horizontal shear reinforcement was increased to 24 in. on center in Span 1, primarily because the 12 in. stirrup spacing utilized in Span 2 was expected to be overly conservative. The horizontal shear reinforcement was also detailed with increased clear spacing below the hooked bars in Span 1, as the 1/4 in. nominal clear spacing provided in Span 2 was expected to inhibit bond development between the CIP concrete and the stirrup hook. The longitudinal and transverse deck reinforcement was also considered during the study, with the north half of the bridge consisting of a reduced longitudinal reinforcement scheme. The transverse deck reinforcement was reduced to No. 4 bars at 12 in. in Span 1, from No. 5 bars at the same spacing in Span 2. The four precast panels used in the Concept 1 laboratory bridge provided eight regions (i.e., one region at each end of each precast panel) that were used in the end zone stress study, discussed in Section 4.5. The bursting reinforcement locations in the Concept 1 laboratory bridge are shown in Figure 5.1.3, and were identical for Spans 1 and 2. The behaviors of interest to this study are discussed herein; additional results pertinent to the Concept 1 laboratory bridge have been documented by Smith et al. (2008) in association with the previous Mn/DOT study. Table 5.1.1: Original and modified design criteria in Spans 1 and 2 of the Concept 1 laboratory bridge Span 1 (Modified Section) Span 2 (Original Section) Decreased flange thickness (3 in.) Original flange thickness (5-¼ in.) Smooth flange surface Original roughened flange surface Increased stirrup spacing for horizontal shear reinforcement (No. 5 Stirrups at 24 in.) Original stirrup spacing for horizontal shear reinforcement (No. 5 Stirrups at 12 in.) Increased clear spacing under hooks (1-3/8 in. nominal clear spacing between horizontal shear reinforcement stirrups and precast section) Original clear spacing under hooks (1/4 in. nominal clear spacing between horizontal shear reinforcement stirrups and precast section) Decreased transverse deck reinforcement (No. 4 bars at 12 in.) Original transverse deck reinforcement (No. 5 bars at 12 in.) The longitudinal deck steel in the south half of the bridge was two No. 7 and one No. 8 bars per 12 in. at the continuous pier (Original design) The longitudinal deck steel in the north half of the bridge was reduced to No. 6 bars at 6 in. spacing at the continuous pier (Modified design)

87 Figure 5.1.3: Bursting reinforcement details used in the Concept 1 laboratory specimen. Configuration numbers correspond to the discussion on bursting in Section 4.5

88 A simplified plan view of the Concept 1 laboratory bridge, including support locations and relevant dimensions is included in Figure 5.1.4; the transverse reinforcement near the longitudinal precast joint region is not shown for clarity. The reinforcement present in each unique cross section (i.e., east, west, or midspan) of each precast panel (i.e., 1N, 1S, 2N, 2S), as well as an elevation and layout views illustrating the location of reinforcement along the length of each beam are shown in Figures 5.1.5 through 5.1.20; four figures are included for each precast panel, in the following order: east cross section, west cross section, midspan cross section, and elevation/plan. A photograph of the Concept 1 laboratory bridge shortly after the completion of the continuity deck pour is shown in Figure 5.1.21. Figure 5.1.4: Plan view of the Concept 1 laboratory bridge, including support locations and relevant dimensions. Transverse reinforcement near longitudinal precast joint is not included for clarity

89 Figure 5.1.5: Cross section and individual reinforcement details for the east end of precast beam 1N

90 Figure 5.1.6: Cross section and individual reinforcement details for west end of precast beam 1N

91 Figure 5.1.7: Cross section and individual reinforcement details for midspan of precast beam 1N

92 Figure 5.1.8: Elevation and plan views of reinforcement layout for precast beam 1N

93 Figure 5.1.9: Cross section and individual reinforcement details for east end of precast beam 1S

94 Figure 5.1.10: Cross section and individual reinforcement details for west end of precast beam 1S

95 Figure 5.1.11: Cross section and individual reinforcement details for midspan of precast beam 1S

96 Figure 5.1.12: Elevation and plan views of reinforcement layout for precast beam 1S

97 Figure 5.1.13: Cross section and individual reinforcement details for east end of precast beam 2N

98 Figure 5.1.14: Cross section and individual reinforcement details for west end of precast beam 2N

99 Figure 5.1.15: Cross section and individual reinforcement details for midspan of precast beam 2N

100 Figure 5.1.16: Elevation and plan views of reinforcement layout for precast beam 2N

101 Figure 5.1.17: Cross section and individual reinforcement details for east end of precast beam 2S

102 Figure 5.1.18: Cross section and individual reinforcement details for west end of precast beam 2S

103 Figure 5.1.19: Cross section and individual reinforcement details at midspan of precast beam 2S

104 Figure 5.1.20: Elevation and plan views of reinforcement layout for precast beam 2S

105 Figure 5.1.21: Photograph of the Concept 1 laboratory bridge shortly after completion of the continuity pour The reinforcement ratios for each span were not identical because of the reduction in the flange thickness in Span 1. The reinforcement ratio for transverse load transfer and crack control in Span 2 (which was similar to the Center City Bridge) was 0.0029 and 0.0147, respectively; whereas the reinforcement ratio for transverse load transfer and crack control in Span 1 was 0.0024 and 0.0110, respectively. The center pier supporting the Concept 1 laboratory bridge was designed to replicate the piers present in the Center City Bridge (see Figure 4.6.1 for the bearing and pier detail in the Center City Bridge). The pier cap extended 42 in. in the direction corresponding to the longitudinal direction of the bridge, identified hereafter as the width of the pier. A 6 in. wide by 1/2 in. thick elastomeric bearing pad provided the primary bearing at the pier, with the bearing pad occupying the area between 4 and 10 in. from the ends of the precast beams. The remaining area between the pier cap and precast beams was filled with 1/2 in. polystyrene foam, which was selected to prevent the egress of CIP concrete during the closure pour, while also remaining relatively crushable to allow the elastomeric bearing pad to act as the primary means of bearing. The supports at the free ends of the bridge were constructed with a 12 in. wide HSS tube section resting on a wide flange section. A 1/2 in. thick by 12 in. wide elastomeric bearing pad was provided between the HSS tube and precast inverted-T. The end support detail was selected to provide little rotational Span 2 Span 1

106 restraint, and therefore simulate a roller connection. Lateral stability at the end supports was provided via bracing between the wide flange sections and the strong floor, as shown in Figure 5.1.22. Figure 5.1.22: Support and bearing detail of the end supports of the laboratory bridge specimens 5.1.2. Concept 2 Laboratory Bridge The Concept 2 laboratory bridge was developed to address issues that remained unanswered at the conclusion of the Concept 1 tests. A statically determinate simply-supported bridge specimen was selected to provide a direct means of determining the applied moment during testing. The symmetrical nature of the simply-supported structure enabled investigation of two unique design parameters in each half span of the bridge. Table 5.1.2 highlights the variations in parameters in the Concept 2 laboratory bridge relative to those of Span 1 of the Concept 1 laboratory bridge. Both half spans of the Concept 2 laboratory bridge were constructed with No. 4 transverse bars spaced at 18 in. protruding from the precast sections. The west half span of the bridge was constructed using traditional hooked bars as defined by AASHTO (2010) Article 5.10.2.1. The east half span of the Concept 2 laboratory bridge was constructed with threaded transverse bars without a hook return or headed feature, as shown in Figure 5.1.23, extending approximately 22.5 in. from the face of the web, which left approximately 1.5 in. clear to the face of the adjacent precast web. The use of hooked threaded bars were deemed impractical due to the difficulty in ensuring the hooked portion of the bar would be oriented vertically upward when the threaded portion of the bar was fully tightened into the anchor; positional threaded inserts that could resolve this issue would be costly. Headed bars were considered as an alternative to hooked bars, but the clearance required around the head would require the transverse bars to be raised higher in the section to accommodate the required clearance of the head to the horizontal precast flange face. As a consequence, it was decided to use straight transverse bars

107 threaded into the anchors embedded in the precast sections. The development length of the straight bars was calculated to be 28.5 in. using ACI 318-08 Section 12.2.2, with an assumed concrete compressive strength of 4,000 psi and ψt=1.2, suggesting that approximately 80 percent of the yield strength of the straight bars would be developed over the distance between the free end of the bar and the vertical precast web. The transverse reinforcement (both the hooked and straight bars) in the trough region provided the primary mechanism for transverse load transfer between adjacent precast panels. For this reason, it is imperative that the transverse reinforcement be continuous through the precast member. In the case of the mechanically anchored reinforcement, continuity was provided using straight bars embedded in the web region of each precast member, which were threaded half way into the mechanical anchor at the interior longitudinal precast joint, as shown in Figure 5.1.24; a vertical hook was maintained over the outside precast flange to assist with the location of the internal reinforcement during the study. When mechanical anchors are used in a real world application, the embedded continuous reinforcement could be connected to mechanical anchors at the vertical web faces on both sides of the precast member. Note that in Figure 5.1.24 the transverse reinforcement to be mechanically anchored across the joint region is not shown for clarity. Table 5.1.2: Comparison of design parameters between Span 1 of Concept 1 and Concept 2 laboratory bridge specimen Span 1 of Concept 1 Laboratory Bridge Concept 2 Laboratory Bridge No 6 transverse hooks spaced at 12 in. No. 4 transverse bars spaced at 18 in. No. 5 cage spaced at 12 in., in line with hooks No. 3 cage spaced at 18 in., offset 9 in. from bars 2 1/2 in. clear from face of flange to bottom of transverse hook 1 in. clear from face of flange to bottom of transverse bar All transverse bars embedded into web and terminated with a standard hook East half span included threaded straight bars West half span included embedded hooked bars No. 5 stirrups at 24 in. for horizontal shear reinforcement No reinforcement for horizontal shear transfer

108 Figure 5.1.23: Threaded connection and adjacent termination detail of straight bars in east half span of the Concept 2 laboratory bridge specimen Figure 5.1.24: Conceptual section illustrating continuous nature of embedded reinforcement utilized in conjunction with mechanical anchors when threaded transverse reinforcement is present; figure shown represents configuration in east half span of the Concept 2 laboratory bridge specimen (transverse reinforcement in joint region to be mechanically anchored to reinforcement in precast web is not shown)

109 As for the Concept 1 specimen, a supplementary cage was used in the joint region to control potential reflective cracks that could originate over the longitudinal joint between the adjacent precast flanges. The reinforcement cage in the Concept 2 laboratory bridge was reduced in size to No. 3 stirrups and increased in spacing to 18 in. in comparison with that of the Concept 1 laboratory bridge. The cage reinforcement was offset from the protruding transverse bars by 9 in. to ensure that the maximum spacing of the transverse reinforcement across the joint did not exceed 9 in. The reinforcement ratio of the transverse section for crack control was 0.0031. The specimen was intended to provide a lower bound design based on the ACI (318-08) and AASHTO (2010) specifications for potential crack control reinforcement, specifically satisfying a spacing limit of 18 in. and a minimum reinforcement ratio of 0.0018. The reinforcement ratio of the section for transverse load transfer was 0.0007. The reinforcement ratios for both the Concept 1 and Concept 2 laboratory bridges are summarized in Table 5.1.3. Table 5.1.3: Transverse load transfer and crack control reinforcement ratios for the Concept 1 and Concept 2 large-scale laboratory bridge specimens Specimen description Reinforcement ratio for crack control Reinforcement ratio for transverse load transfer Concept 1, Span 1 0.0110 0.0024 Concept 1, Span 2 0.0147 0.0029 Concept 2 0.0031 0.0007 Another modification implemented in the Concept 2 laboratory bridge included the removal of all horizontal shear reinforcement, leaving only the roughened surface of the precast web to provide all necessary horizontal shear transfer. This modification was selected to address the required minimum horizontal shear reinforcement in the AASHTO (2010) LRFD specification, as the literature suggested that requirement was overly conservative (see review of work by Naito et al. (2006) in Section 2.3). Furthermore, the large horizontal interface between the top of the precast webs and the CIP deck provided a horizontal shear interface area that was considerably larger than is generally provided by traditional girder bridges, where the width of the top flange is relatively narrow. The sectional capacity of the Concept 2 laboratory bridge was calculated using the measured concrete compressive strength of 5,800 psi, which resulted in the determination of the largest expected internal compression and tension forces for the specimen, which were equal to approximately 2450 kips. Using the global force equilibrium method for calculating the horizontal shear demand, the total change in compression force was divided by the area over which that change occurs. Because the Concept 2 laboratory bridge was simply supported, the compression force was assumed to be a maximum of 2450 kips at midspan, and 0 kips at the center of bearing. The shear area was therefore calculated as the product of half of the center to center of bearing (256 in. / 2 = 128 in.) multiplied by the full width of the web in compression, or 10 ft.; the area over the longitudinal trough region was included in the width of the shear area, as this area was expected to provide horizontal shear resistance as effectively as the horizontal precast-CIP interface [Note that CIP was not placed over the exterior flanges, only in the trough region between the two adjacent precast sections.]. Using this method, the horizontal shear demand at the predicted ultimate capacity of the Concept 2 laboratory bridge was calculated to be approximately 160 psi. The cohesion factor for a horizontal shear interface that is clean and intentionally roughened provided in Article 5.8.4.3 of the AASHTO LRFD specification (2010) was 240 psi, suggesting that the largest

110 horizontal shear stress that was expected to be developed in the Concept 2 laboratory bridge was unlikely to result in loss of composite action. A simplified plan view of the Concept 2 laboratory bridge specimen, including support locations and relevant dimensions is included in Figure 5.1.25; the transverse reinforcement near the longitudinal precast joint region is not shown for clarity. The Concept 2 laboratory bridge was designed with the same overall dimensions as a single span in the Concept 1 laboratory bridge. Therefore, the Concept 2 laboratory bridge was placed to the east of the center pier, with the west bearing being located on the concrete pier, and the east bearing located on the 12 in. HSS tube. Based on the work outlined in Section 4.5, no bursting reinforcement was necessary in the precast elements utilized in the Concept 2 specimen. The reinforcement present in each cross section (i.e., east, west, and midspan), as well as elevation and layout views documenting the location of reinforcement along the length of each beam is shown in Figures 5.1.26 through 5.1.32. A photograph of the Concept 2 laboratory bridge specimen is shown in Figure 5.1.33. Figure 5.1.25: Plan view of the Concept 2 laboratory bridge specimen, including support locations and relevant dimensions. Transverse reinforcement near longitudinal precast joint is not included for clarity

111 Figure 5.1.26: Cross section and individual reinforcement details for the east end of the precast beam 1N in the Concept 2 laboratory bridge specimen

112 Figure 5.1.27: Cross section and individual reinforcement details for the west end of the precast beams 1N and 1S in the Concept 2 laboratory bridge specimen

113 Figure 5.1.28: Cross section and individual reinforcement details at midspan of precast beam 1N in the Concept 2 laboratory bridge specimen

114 Figure 5.1.29: Elevation and plan views of the reinforcement layout for precast beam 1N in the Concept 2 laboratory bridge specimen

115 Figure 5.1.30: Cross section and individual reinforcement details for the east end of the precast beam 1S in the Concept 2 laboratory bridge specimen

116 Figure 5.1.31: Cross section and individual reinforcement details at midspan of precast beam 1S in the Concept 2 laboratory bridge specimen

117 Figure 5.1.32: Elevation and plan views of the reinforcement layout for precast beam 1S in the Concept 2 laboratory bridge specimen

118 Figure 5.1.33: Photograph of the Concept 2 laboratory bridge specimen prior to placement of CIP concrete The Concept 2 laboratory bridge specimen occupied the same space in the University of Minnesota Structures Laboratory as Span 2 of the Concept 1 laboratory bridge specimen. The west end of the Concept 2 laboratory bridge specimen was located a distance of 2 in. from the center of the concrete pier, with the east end of the specimen bearing completely on the 12 in. HSS tube/wide flange support. See Figure 4.6.1 and 5igure 5.1.22 for details on the concrete and steel supports, respectively. 5.1.3. Instrumentation of Concept 1 and Concept 2 Laboratory Bridge Specimens The instrumentation schemes used in the Concept 1 and Concept 2 laboratory bridge specimens facilitated the monitoring of the transverse behavior and health of the joint region between precast panels. In addition, vertical sets of longitudinally oriented instrumentation were provided as a means of calculating the longitudinal curvature during load transfer and ultimate tests. The primary instrumentation type was concrete embedment resistive strain gages due to their relatively inexpensive cost, which allowed for an adequately dense instrumentation field. A small number of longitudinally oriented concrete embedment VW strain gages were also used to allow for the measurement of prestress losses, as well as to provide an estimate of the absolute levels of strain in the section without the effects of gage drift. Transversely oriented concrete embedment VW and spot-weldable VW strain gages were also included to provide an estimate of the strain due to transverse shrinkage near the precast joint, as well as to provide an estimate of the absolute levels of transverse strain throughout the laboratory tests. West End East End

119 The instrumentation layout for the Concept 1 laboratory bridge specimen is shown in Figure 5.1.34 (Smith et al. 2008). Each span had a set of transversely oriented instruments at each end, each quarter point, and at midspan. At each section, five transversely oriented concrete embedment resistive gages were located near the precast flange equally dispersed between adjacent precast webs. Some cross sections had an additional ten transversely oriented concrete embedment resistive gages located above the precast web corners. The presence of the gages above the precast web is indicated by the longer transverse line extending over the webs in Figure 5.1.34. A representative cross-sectional view of the fully instrumented section is shown in Figure 5.1.35; the aspect ratio of the instrumentation has been modified for clarity. In addition, adjacent transverse gages overlapped (not shown in figure) to provide redundancy and a means to better refine the determination of crack locations by investigating whether a crack may have intersected a single gage or two adjacent gages. The nominal and measured gage locations in the Concept 1 laboratory bridge are given in Appendix E. Figure 5.1.34: Instrumentation layout for Concept 1 laboratory bridge specimen (Smith et al. 2008)

120 Figure 5.1.35: Typical instrumentation layout near precast joint in the Concept 1 laboratory bridge specimen The instrumentation layout developed for the Concept 2 laboratory bridge closely followed the original instrumentation plan for the Concept 1 laboratory bridge. Additional transversely oriented instruments were added at the eighth points to allow for increased refinement in detection of the extent of longitudinal reflective cracking. Similarly, an additional layer of transversely oriented concrete embedment strain gages was added through the depth of the cross section to improve the accuracy of the measurement of the depth of the crack in the section. The number of gages across the width of the joint region was reduced in response to the localized behavior of the crack observed in the Concept 1 laboratory bridge near the precast joint. Furthermore, fewer vertical sets of instruments were selected for longitudinal instrumentation, while five gages were included in each vertical set to improve the accuracy of the curvature calculation, especially in the case of faulty or broken instrumentation due to construction. The instrumentation layout for the Concept 2 laboratory bridge and the standard cross sections are shown in Figure 5.1.36 and Figure 5.1.37, respectively. The nominal and measured gage locations for the Concept 2 laboratory bridge are given in Appendix F. The origin and positive x and y ordinate directions for the placement of the instrumentation for the Concept 1 and Concept 2 laboratory bridges is shown in Figure 5.1.38. The origin was located at the center of the concrete pier for both bridges. The positive x ordinate was directed towards the east wall of the lab, with the positive y ordinate towards the north. The vertical ordinate, z, was measured from the bottom of the precast sections.

121 Figure 5.1.36: Instrumentation layout for the Concept 2 laboratory bridge specimen

122 Figure 5.1.37: Typical 9 and 6 gage transverse instrumentation layout in Concept 2 laboratory bridge specimen Figure 5.1.38: Location of origin and definition of positive x and y ordinates for instrumentation layout in the Concept 1 and Concept 2 laboratory bridge specimen

123 5.2. Construction of Laboratory Bridge Specimens and Material Properties Fabrication of the precast panels for both bridges was completed at County Materials in Roberts, Wisconsin. The CIP concrete selected for the project was a standard Mn/DOT bridge design mix, designated as 3Y33H68, with a nominal compressive strength of 4,000 psi at 28 days. Due to the interest in the effects of restraint moment on the system, the CIP concrete was placed on the Concept 1 laboratory bridge when the precast panels were at a young age of 7 days. Material testing was completed on the Concept 1 laboratory bridge to determine measured concrete compressive strengths, tensile strengths, and elastic modulus, and was tabulated by Smith et al. (2008), which is summarized in Table 5.2.1 and Table 5.2.2 for the CIP and precast concrete, respectively. A more relaxed construction schedule was sufficient for the Concept 2 laboratory bridge because of the simply-supported design. The CIP concrete for the Concept 2 laboratory bridge was poured when the precast panels had reached an age of 83 days. The concrete 28-day CIP concrete compressive strength was measured to be 5750 psi, measured according to ASTM C36 Test Method for Compressive Strength of Cylindrical Concrete Specimens (ASTM C36, 2009). The CIP concrete tensile strength was measured to be 640 psi at an age of 28 days, measured according to ASTM C78 Standard Test Method for Flexural Strength of Concrete (Using Simple Beam with Third-Point Loading) (ASTM C78, 2009). The elastic modulus of the CIP concrete at an age of 28 days was measured to be 4,020 ksi, which was measured in accordance with ASTM C469 Standard Test Method for Static Modulus of Elasticity and Poisson’s ratio of Concrete in Compression (ASTM C469, 2002). The measured material properties are summarized in Table 5.2.1 and Table 5.2.2 for the CIP and precast concrete, respectively. Table 5.2.1: Measured CIP material properties at an age of 28 days Specimen Description Compressive Strength, fc’ Modulus of Rupture, fr Elastic Modulus, E Concept 1, Span 1 4160 psi 411 psi 3540 ksi Concept 2, Span 2 4590 psi 457 psi 3780 ksi Concept 2 5750 psi 640 psi 4020 ksi Table 5.2.2: Measured precast concrete material properties for the Concept 1 and Concept 2 bridges Specimen Description Compressive Strength, fc’ Modulus of Rupture, fr Elastic Modulus, E Concept 1, both spans 12,900 psi 927 psi 6160 ksi Concept 2 12,150 psi 783 psi 6029 ksi The surfaces of the precast panels were not pre-wetted prior to the placement of the CIP concrete for either of the bridge specimens. No deleterious behavior appeared to result in casting the CIP on non- wetted precast surfaces; however, it is recommended in future applications of constructing PCSSS bridges in the field, that the precast surfaces be pre-wetted prior to casting the CIP. Both specimens were moist-cured for 8 days based on the ACI 308.1-98 Standard Specification for Curing Concrete.

124 The yield strength of the No. 4 hooked bars that provided transverse reinforcement continuity and crack control reinforcement above the longitudinal joint between the precast panels in the Concept 2 laboratory bridge specimen was found to be approximately 70 ksi through the testing of three representative samples. 5.3. Laboratory Testing Program and Results An extensive laboratory testing program was developed for each specimen. The requirements for the testing plan included simulated fatigue loading, reproducing levels of strain in the precast joint region as observed in the Center City Bridge, as well as the introduction of reflective cracking, monitoring load transfer between precast panels in common spans, and loading to the ultimate capacity of the bridge specimens or the maximum capacity of the load frame. Because the tests used on the Concept 1 and Concept 2 laboratory bridges were similar, the testing plans and results of the laboratory tests are discussed below based on the controlling behavior, rather than by laboratory specimen. The results of the tests presented herein were all associated with the NCHRP 10-71 study. The previous tests conducted on the Concept 1 laboratory bridge specimen as part of a Mn/DOT study are summarized in Smith et al. (2008). 5.3.1. Simulated Traffic Loading Good fatigue performance of any highway structure is essential to ensure stable behavior over extended periods of time. Traffic loading was simulated on the Concept 1 and 2 bridge specimens using a representative patch load of dimensions 10 by 20 in. oriented with the long direction perpendicular to the joint. Two million cycles of fatigue loading were completed on each span of the laboratory bridge specimens. The loading was placed at midspan of both spans of the Concept 1 laboratory bridge. The complete two million cycles of fatigue loading were completed on Span 2 prior to the implementation of fatigue loading on Span 1, while equal patch loads were cycled at the quarter points of the Concept 2 laboratory bridge using a spreader beam. The nominal patch load used for all fatigue loading was 35 kips (i.e., Concept 1 laboratory bridge: 35 kips was applied to midspan; Concept 2 laboratory bridge: 35 kips was applied to each quarter point simultaneously); the development of the magnitude of the laboratory patch loading was discussed in Section 4.7. The magnitude of the patch load was developed to replicate the magnitude of transverse stress at the precast flange-CIP interface in the laboratory specimen as was expected in a 30-30-30 ft. three-span continuous bridge with twice the AASHTO (2010) tandem design load applied. A magnitude of twice the AAHSTO tandem load was selected to account for the possibility of two truck wheel loads being located directly above a longitudinal joint and situated immediately adjacent to one another. As discussed in Section 4.7, the stress developed near the precast flange at midspan of the Concept 1 laboratory bridge and at the quarter points of the Concept 2 laboratory bridge were nearly identical, which accounts for the equal patch loads used in both specimens. Also note that the maximum moment developed due to a single patch load at midspan was equal to the maximum moment developed due to two equal patch loads placed at the quarter points. During the fatigue and all associated patch loading, the load was cycled between 2 and 35 kips. A minimum load of 2 kips was maintained during all tests to prevent the actuator from “walking” during cycling. Steel angles were also adhered to the surface of the deck to help constrain the movement of the actuators during testing. Load distributed to the quarter points of the Concept 2 laboratory bridge using the spreader beam was cycled between 4 and 70 kips.

125 As discussed in Section 3.2 reflective cracking was observed in two of the three instrumented joints in the Center City Bridge via transversely oriented concrete embedment VW strain gages located at midspan of the center span. Furthermore, cracking in the field bridge was observed to occur during the first spring after construction was completed, revealing the potential for reflective cracking to be present during the majority of the service life for the precast composite slab span system (PCSSS) bridge. In an effort to investigate the fatigue performance of the PCSSS both with and without the presence of reflective cracking, the fatigue study for each specimen was conducted in two parts. As mentioned above, the fatigue study for each span of the Concept 1 laboratory Bridge (Span 2 first, followed by Span 1) was conducted in series; followed by the tests of the Concept 2 laboratory bridge after completion of the Concept 1 laboratory bridge tests. The subsequent discussion of the fatigue testing is relevant to each span separately. The first portion of fatigue loading consisted of the completion of one million cycles of fatigue loading on the virgin laboratory specimens, with no reflective cracking observed prior to the initiation of fatigue loading. At the conclusion of each 100,000 cycles, fatigue loading was suspended to allow for a quasi- static 35 kip patch load to be applied to the section at the same location as the fatigue patch load (i.e., Concept 1 laboratory bridge: 35 kips was applied to midspan; Concept 2 laboratory bridge: 35 kips was applied to each quarter point simultaneously). During the quasi-static loading, data from the complete set of transversely oriented strain gages was collected. A minimum of three quasi-static cycles were completed during this process to provide redundancy. This process provided a sample of the condition of the longitudinal joint in 100,000 cycle increments. At the completion of the first million cycles of loading, a longitudinal reflective crack was introduced near the precast joint by increasing the applied loads until target strain values were reached to replicate the levels of transverse strain measured via concrete embedment VW strain gages located near the precast flange in the Center City Bridge. The target transverse strain value of 160 µε was selected to represent the maximum daily change in strain observed in Joint 1 during the first summer after construction. The target strain value of 160 µε is hereafter referred to as part of the environmental effect simulation, which is discussed in further detail in Section 5.3.2. During the initiation of the reflective crack in the laboratory study, mechanical loading was applied to the same 10 by 20 in. patch dimension utilized during the fatigue study while the transverse strains measured via the concrete embedment resistive strain gages located nearest the precast flanges were monitored in real time, until a change in strain of approximately 160 µε was observed, which was also generally associated with observed nonlinearity in the strain measurements, suggesting that a crack was generated. For Span 2 of the Concept 1 laboratory bridge, the crack was initiated at midspan (i.e., same location as during fatigue tests). In Span 1 of the Concept 1 laboratory bridge and the Concept 2 laboratory bridge, the crack was first initiated at the same longitudinal location as the fatigue tests (i.e., midspan of Span 1 and at the quarter points of the Concept 2 laboratory bridge, though cracking at the west quarter of the Concept 2 laboratory bridge was not clearly observed, which is discussed in Section 5.3.2. For these two spans, the crack was extended longitudinally using a 10 ft. long spreader beam, as it was expected that the uniform nature of a thermal gradient induced in a field bridge would tend to extend the crack along the length of the span. Bearing plates were placed a distance of 2.5 and 5 ft. to either side of midspan (corresponding with an effective 5 and 10 ft. spreader length). The locations of the applied patch loads for fatigue loading and locations of the patch loads used to extend the length of the reflective crack, where applicable, are shown in Figures 5.3.1 and 5.3.2 for the Concept 1 and Concept 2 laboratory bridge specimens, respectively.

126 Figure 5.3.1: Placement of patch loads during fatigue loading and extension of longitudinal reflective cracking (applicable in Span 1 only) for the Concept 1 laboratory bridge specimen Figure 5.3.2: Placement of patch loads during fatigue loading and extension of longitudinal reflective cracking for the Concept 2 laboratory bridge specimen

127 The vertical location of the transversely oriented concrete embedment resistive strain gages varied between each span of the Concept 1 laboratory bridge and the Concept 2 laboratory bridge due to the variations in the thickness of the precast flange and the vertical depth of the transverse reinforcement traversing the longitudinal precast joint. The measured vertical locations of the bottommost layers of transversely oriented strain gages that were monitored during initiation of cracking in the laboratory bridge specimens, as well as the relative instrumentation in the Center City Bridge, are given in Table 5.3.1. Because of the variations in the depths of the instrumentation, the transverse strain observed in the Center City Bridge was not simulated exactly in the laboratory specimens. The target strain was not modified for each span based on the relative depth of the instrumentation because a relatively consistent magnitude of strain was desired to be induced in the transverse reinforcement (which was vertically collocated with the instrumentation). The transverse strains measured during the introduction of reflective cracking in each span are shown in Table 5.3.2, along with the total applied load required to achieve each level of strain. The transverse strains reported here were the largest strains measured at the cross sections associated with fatigue loading (i.e., midspan for Concept 1 laboratory bridge, and quarter points for the Concept 2 laboratory bridge). When cracking was completed using the spreader load, the applied load shown in Table 5.3.2 represents the total load applied to the spreader. The distribution of load to symmetrically placed patch loads was measured via load cells and was found to be approximately equal. Table 5.3.1: Measured vertical locations of transversely oriented strain gages that were utilized in the observation of reflective cracking in the Center City Bridge and laboratory bridge specimens Specimen Description Type of Instrumentation Vertical depth of instrumentation, measured from the bottom of precast section Center City Field Bridge Concrete embedment VW strain gage 8.25 in. Concept 1, Span 1 laboratory bridge Concrete embedment resistive strain gages 6 in. Concept 1, Span 2 laboratory bridge Concrete embedment resistive strain gages 7.25 in. Concept 2 laboratory bridge Concrete embedment resistive strain gages 4 in. Table 5.3.2: Measured transverse strains during introduction of reflective cracking after the completion of one million fatigue cycles in each specimen Specimen Description Load applied to patch or spreader Measured transverse strain Associated applied load Concept 1, Span 1 Patch load 161 µε 89 kip Concept 1, Span 2 Patch load 200 µε 95 kip Concept 2, east quarter point (straight bars) Spreader load 150 µε 210 kip Concept 2, west quarter point (hooked bars) Spreader load 85µε 210 kip

128 The largest strain that was developed in the west quarter point of the Concept 2 laboratory bridge was 85 µε, when a total load of 210 kips was applied to a spreader beam and a transverse strain of 150 µε was measured at the east quarter point. A second million cycles of fatigue loading were subsequently completed on each specimen after reflective cracking was introduced in each section. As before, cyclic loading was suspended every 100,000 cycles to investigate the condition of the longitudinal precast joint via the measurement of the transversely oriented strain gage data during application of a quasi-static 35 kip patch load. The behavior of the large-scale bridge specimens subjected to the fatigue loading is summarized in Table 5.3.3, which illustrates the change in strain measured under the quasi-static 35 kip patch load at relevant points during the two million cycles of loading. The first column of the table identifies the specimen designation. For the Concept 2 specimen, fatigue loading was completed at the quarter points using a spreader beam; in addition, a 35 kip quasi-static load was applied at midspan to investigate the performance of the joint at that location, and is included in the table. The initial strain was the strain measured prior to the fatigue tests (at zero cycles). Because little variation in the transverse strain was observed during the first and second million cycles, only the end points of these loading periods are shown. The transverse instrumentation used to ascertain the results in Table 5.3.3 was determined after the introduction of cracking at one million cycles. During and after cracking, only one or two transversely oriented concrete embedment resistive strain gages located at the loaded cross section recorded a crack with a significant permanent strain increase, suggesting that, at those cross sections, the crack traversed those gages. For all cases, this instrumentation was located in the lowest level of gages, near the precast flanges, and was coincident with the location of loading (i.e., at midspan if loading was applied at midspan or at the quarter points when loading was applied there). Figure 5.3.3 illustrates the gages that were used to create Table 5.3.3 for each specimen. Note that for the Concept 1 specimen, gage number 3 corresponded to the gage centered over the precast joint; whereas, in the Concept 2 specimen, gage number 2 corresponded to the gage centered over the precast joint. Table 5.3.31 Specimen : Change in transverse strain measured at 35 kips throughout the course of 2M cycles of fatigue loading in each span of Concept 1 and Concept 2 laboratory specimens Initial Strain (µε) Change from initial strain (µε)2 Before Cracking After Cracking After 2M Cycles Concept 1, Span 1 18 0 24 25 Concept 1, Span 2 23 1 26 24 Concept 2, Midspan 33 -2 9 9 Concept 2, East Quarter span (straight bars) 33 -1 32 30 Concept 2, West Quarter span (hooked bars) 30 -3 5 6 1The instrumentation used to measure the values in this table is illustrated in Figure 5.3.3 2The initial strain was measured at zero cycles. The data taken to represent the before cracking state was recorded immediately after the completion of one million cycles. The data representing the after cracking state was recorded immediately after cracking with no cycling completed between.

129 (a) Concept 1 Span 1; patch load at midspan (b) Concept 1 Span 2; patch load at midspan (c) Concept 2; patch load at midspan (d) Concept 2; patch load at west quarter point (e) Concept 2; patch load at east quarter point Figure 5.3.3: Transversely oriented concrete embedment resistive gages located nearest the precast flange1 1The instrumentation is shown in two layers to illustrate the overlap between gages, however the gages were nominally located in the same vertical layer and at the same depth as the horizontal legs of the embedded transverse reinforcement. Extent of overlap between gages is shown to scale . Instrumentation locations which detected reflective cracking and were used for measurement of transverse strain values during fatigue loading are highlighted in black and annotated

130 Good fatigue behavior was observed for both the Concept 1 and Concept 2 laboratory bridge specimens. No significant degradation of the joint was observed. The transverse strains remained stable throughout the two million cycles of load regardless of the presence of the longitudinal reflective crack that was imposed near the precast joint. The performance of each specimen during the second million cycles of fatigue loading can be inferred from Table 5.3.3 by subtracting the measured strain values in the two rightmost columns. The largest increase in the strain measured during the second million cycles of loading was observed to be 2 µε in both Span 2 of the Concept 1 laboratory bridge and the east quarter of the Concept 2 laboratory bridge. Because the resistive gages were considered accurate to approximately +/- 6µε, this small increase in strain was considered negligible. Furthermore, the increase in the transverse strain due to the introduction of the longitudinal reflective crack (calculated by subtracting the after cracking strain value from the before cracking strain value in Table 5.3.3) varied slightly depending on the specimen. The increase in strain in Span 1 and 2 of the Concept 1 laboratory bridge was 24µε and 25µε, respectively, while the increase under the spreader load in the Concept 2 laboratory bridge was approximately 33µε and 8µε for the east and west quarters, respectively. The increase in strain due to cracking under the midspan patch load in the Concept 2 laboratory bridge was approximately 11µε. The similar performance level attributed to the Concept 1 laboratory bridge and Concept 2 laboratory bridge in regards to reflective crack control under traffic loading suggested that the reinforcement details from either span were adequate for the design of PCSSS bridges. 5.3.2. Environmental Effect Simulation As mentioned in Section 5.3.1, transverse strain measurements in the Center City Bridge suggested that a reflective crack initiated in the joint region during the first spring after construction. Of the three adjacent instrumented joints at midspan of the center span of the field bridge, the two outermost joints showed evidence of cracking. The transverse strains measured in the bridge immediately after construction in October 2005 until July 2009 are shown in Figure 3.2.7. The instruments directly over the joint and in the position immediately next to the joint (blue and red series, respectively) illustrated the presence of a reflective crack, initiating on April 25th, 2006. The ranges of strains observed during the summer of 2008 and summer of 2009 were approximately 150µε and 220µε respectively. The seasonal fluctuation of the strain measurements, in addition to the fact that the 35 kip patch load used for fatigue testing had induced a transverse strain of only approximately 40µε in the laboratory bridge, further substantiated that the large strains observed in the transverse joints of the field bridge were likely due to environmental effects rather than due to traffic load. The increase in transverse strains from the summer of 2008 to the summer of 2009 warranted the investigation of the degradation of the joint region under large strains. It was expected that the increase in strains was caused by one of three sources: (1) fatigue traffic loading on the cracked system caused the degradation of the joint, (2) the cyclic influence of the thermal gradient change between day and night caused the increase in transverse strain, or (3) a larger thermal gradient may have been experienced that caused the crack to extend. Therefore, as mentioned in Section 5.3.1, a crack was induced in the laboratory bridge at the conclusion of one million traffic fatigue cycles to allow for the investigation of traffic fatigue loading on the cracked system. At the conclusion of the second million cycles of traffic fatigue loading, additional cyclic loading was applied to the laboratory bridge specimens at larger strain levels to simulate cyclic strains due to environmental effects observed in the Center City Bridge. The strain cycles simulated the large daily strain fluctuations observed in the field bridge that occurred during the seasons that experienced the largest effects of solar radiation. For each level of strain considered, a total of 15,000 cycles were applied. The 15,000 cycles were selected to simulate approximately 100 years of significant temperature

131 fluctuations, assuming that a total of 150 days a year were expected to induce considerable thermal gradients. Each of the three test spans were subjected to similar levels of transverse strains throughout the testing procedure. Two qualitative levels of transverse strain were considered during the environmental study, specified from here forward as environmental simulation strain level A, which corresponded to a target strain of approximately 180 µε, and environmental simulation strain level B, which corresponded to a target strain of approximately 300 µε. Target level A was selected based on the observed daily fluctuations in the transverse strain in Joints 1 and 3 of the Center City Bridge soon after reflective cracking was observed. Target level B was defined during the environmental simulation tests on Span 1 of the Concept 1 specimen; a strain of approximately 300 µε was measured in the transverse instrumentation near the precast joint when a total patch load of 210 kips, which was the maximum practically available load to the 210 kip actuators, was applied at midspan of Span 1. This magnitude of transverse strain was comparable to the daily transverse strain fluctuations observed in Joint 3 of the Center City Bridge during the summer of 2008. At both levels of strain to simulate environmental effects, quasi-static and cyclic loading was applied. As with the initiation of reflective cracking in the specimens after the first million fatigue cycles, quasi-static loading was applied to each specimen, while the instrumentation designated in Figure 5.3.3 (same instrumentation that was utilized during the fatigue tests) was monitored in real time, until a change in strain approximately equal to the target levels was reached. The full range of transversely oriented instrumentation was also monitored during this process, however only the gages that had previously indicated cracking exhibited large increases in strain with load, suggesting that no new cracks developed during this process. Quasi-static loading was applied using a load controlled program, with load applied at a rate of approximately 6 kips per minute until a strain of 80 percent of the target strain was observed, at which point load was applied at a rate of 1 kip per minute. All quasi-static and cyclic loading associated with environmental simulation strain level A was completed prior to the initiation of quasi- static and cyclic loading to strain level B. A spreader beam was also employed in each of the spans to help increase the longitudinal extent of the reflective crack. Because the methodology utilized in the environmental effect simulation evolved during the tests, the specific spreader orientation for each span is discussed later in this section. Cyclic loading during the environmental simulation was completed using a displacement-controlled test program, which effectively ensured that the level of strain induced near the precast joint during each cycle was approximately equal to the target strain level. Displacement limits were determined during the preceding quasi-static load-control tests at both target levels. The displacement limits were selected such that each cycle ranged from a maximum displacement which approximately induced the level of strain selected, while the minimum displacement was selected to ensure that a compressive force of no less than 2 kips was present at each patch load, which was primarily to keep the loading actuator from “walking” during testing. The magnitude of strain in the selected instrumentation (i.e., Figure 5.3.3) and the minimum load was monitored throughout each test. The displacement limits were modified in real time as necessary to maintain the desired values, though this process was seldom required. The maximum levels of strain achieved in the joint region in each of the specimens throughout the environmental effect simulation are tabulated in Table 5.3.4. For both spans of the Concept 1 specimen, the cyclic loading to induce the target strain levels was applied at midspan through a single patch load, though both quasi-static patch and spreader loads were first utilized to induce and extend the longitudinal crack. For the Concept 2 laboratory bridge, the cyclic environmental loading to induce the target-level strains was applied using a spreader beam, with the patch loads placed at the quarter points of the span, though both quasi-static patch and spreader loads were utilized to induce and extend the crack prior to cyclic loading. The column designated by ‘P/S’ in Table 5.3.4 represents the use of a patch

132 load, spreader load, or both (in which case the order shown is that applied to the specimen) during the quasi-static cracking process. Generally, the use of the spreader test did not affect the strains measured at midspan, and vice-versa, which indicated that the effects of loading were relatively localized and did not affect the measurements 1/4 span away. The number of cycles completed at each load, if applicable, is represented in parentheses immediately following the respective strain level to which the sections were cycled. The column headers in Table 5.3.4 represent the target environmental simulation strain levels. Deviations from the measured strain and target strain at each level were generally due to limitations in the capacities of the loading actuators used during the study, as was observed for the Concept 2 specimen at the 300µε target strain. Table 5.3.4: Maximum transverse strains and number of cycles completed at given strain level during laboratory environmental effect simulation Max. strain observed in joint regions and (number of cycles completed at that strain) Prior to cracking at 1M cycles1 P/S2 Cracking at 1M Cycles P/S1 Environmental simulation strain level A P/S1 Environmental simulation strain level B P/S1 Approximate target strain value 160 µε 180 µε 300 µε Concept 1, Span 1 18µε P 161µε S,P 213µε (15,000) S,P 303µε (15,000) S,P Concept 1, Span 2 23µε P 200µε P 183µε (15,000) S,P 233µε (15,000) S,P Concept 2, east quarter 33µε S 150µε S,P 177µε (15,000) S,P 253µε (15,000) S,P Concept 2, west quarter 30µε S 85µε S,P 81µε (15,000) S,P 104µε (15,000) S,P Concept 2, midspan 33µε P 174µε S,P 176µε (15,000) S,P 256µε (15,000) S,P 1Measured strain was induced due to applied 35 kip patch load at midspan of the Concept 1 laboratory bridge spans, and at each quarter span, simultaneously, in the Concept 2 laboratory bridge 2Patch (P) or spreader (S) load orientation utilized to apply initial quasi-static cracking load, prior to cyclic loading Span 2 of the Concept 1 laboratory bridge (which emulated the Center City Bridge section) was the first specimen to undergo loading to simulate environmental effects and as a result was subjected to the most unique loading procedure of all of the bridge specimens. Specifically, the use of a 110 kip capacity actuator initially limited the loading capabilities on the specimen. Also unique to the environmental simulation history on Span 2 of the Concept 1 laboratory bridge was that a preliminary load test was completed which induced a transverse strain of approximately 160µε near the joint at which 200 cycles were completed. This process was completed to investigate the performance of the joint to the 160µε level after the completion of the second million cycles of fatigue load. The specimen was subsequently cycled for an additional 15,000 cycles to the environmental simulation levels A and B each. The reflective crack introduced in the section after the completion of the first million cycles of fatigue load was developed using a quasi-static patch load at midspan, and was not extended with use of a spreader beam. Cracking induced in the section to correlate with the environmental simulation strain levels A and B was generated with the use of both spreader and patch loading. The reflective crack was extended to

133 the outer (i.e., east) quarter point of Span 2, however the crack was not observed to have extended to the instrumentation at the interior quarter point, which may have been attributed to transverse restraint provided by the proximity to the nearby interior support. Figure 5.3.4 illustrates the performance of Span 2 of the Concept 1 laboratory bridge during the environmental loading simulation. The diamond, square, and triangle data series represent strain measurements at midspan, the outer quarter point and the inner quarter point, respectively, with the data measured from the gages centered over the precast joint at each cross section. The vertical axis represents the magnitude of the transverse strain measured under a 35 kip patch load applied at midspan. The number of cycles completed, the strain that was induced in the section and the load required to induce that strain has been tabulated for Span 2 of the Concept 1 laboratory bridge in Table 5.3.5. Figure 5.3.41 1Figure 5.3.4 shows the transverse strain measured at midspan, the outer quarter point and the inner quarter point under a 35 kip patch load applied at midspan of Span 2. The horizontal axis contains a description of the behavior and qualitative target strain values. For example, the data points labeled “After 15,000 cycles to enviro. strain level A” represent the transverse strain measured with a 35k patch load applied at midspan after a load to induce a strain of 183µε (from Table 5.3.4) was applied quasi- statically, and a spreader beam was utilized to extend the crack. : Transverse strains measured with a 35 kip patch load applied at midspan during environmental effect simulation in Span 2 of Concept 1 laboratory bridge 0 20 40 60 80 100 120 Tr an sv er se S tr ai n [u e] Mid-Span Outer Quarter Pt Inner Quarter Pt. Prior to cracking at 1M cycles of fatigue load After cracking at 1M cycles of fatigue load After cracking to induce strain of approx. 160µε After 200 cycles to induce strain of approx. 160µε After 15,000 cycles to enviro. strain level A After 15,000 cycles to enviro. strain level B

134 Table 5.3.5: Measured load and transverse strain, and number of cycles completed during environmental effect simulation on Span 2 of the Concept 1 laboratory bridge specimen Concept 1, Span 2 Prior to cracking at 1M cycles of fatigue load After cracking at 1M cycles of fatigue load After cracking to target strain of 160µε After cycling to target strain of 160µε After cycling to enviro. strain level A After cycling to enviro. strain level B Magnitude of target strain NA1 160µε 160µε 160µε 180µε 300µε Number of cycles completed at each stage 3 3 3 200 15000 15000 Max. transverse strain achieved 23µε 200µε 160µε 160µε 183µε 233µε Load required to achieve strain 35k2 95k2 95k2 95k2 108k2 155k3 1The first column of Table 5.3.5 represents the strain measured during the fatigue loading. Fatigue loading was not based on strain levels, but rather was based on an applied patch load of 35 kip, therefore no target strain was applicable to fatigue loading 2The load recorded represents the patch load applied at midspan. 3The load recorded represents the total load applied to the 10 ft. longitudinal spreader beam. The load applied at each quarter point was half of this value. Span 2 of the Concept 1 laboratory bridge indicated degradation of the joint due to the static and cyclic application of loads after reflective cracking was first introduced into the span (which corresponds to the data point labeled as “After cracking at 1M cycles of fatigue loading”). Degradation of the joint is evident in Figure 5.3.4 in that the transverse strain measured at midspan with a 35 kip patch load applied at midspan increased as the environmental effect simulation progressed from initially cracking the joint after the first million cycles of fatigue load to the last environmental simulation test of applying cyclic loading to induce strains on the order of the environmental simulation strain level B. The increase in strain under the 35 kip patch load applied at midspan, however was relatively small after the crack was first introduced in the section. The increase in transverse strain at midspan under the 35 kip patch load due to the initiation of the crack after the first million cycles was approximately 25µε (i.e., 50µε - 25µε). Increases in the transverse strain due to cracking and cyclic loading to induce strains on the order of the environmental target levels A and B were observed to be between 5-8µε (i.e., difference in measured strains between the second and third data points, the third and fourth data points, etc.). Furthermore, a total increase in strain at the 35 kip patch load of about 15µε at midspan due to the application of 30,000 cycles to simulate environmental loading (i.e., 77µε after cycling to strain level B - 62µε after 200 cycles to strain level A), suggests a relatively stable joint region when loaded to strains on the order of 200µε. Recall that the application of 15,000 cycles was assumed to represent approximately 100 years of thermal effect loading, where there was assumed to be 150 days in a year with a significant thermal gradient. A similar environmental simulation process was completed on Span 1 of the Concept 1 laboratory bridge. The primary difference in testing Span 1 involved the early use of an actuator with a 220 kip capacity. The reflective crack introduced in the section after the completion of the first million cycles of fatigue load was developed using a quasi-static patch load at midspan, and was also extended with use

135 of a spreader beam during two separate load cycles, the first with bearing points symmetrically located 2.5 ft. from midspan, and the second with bearing points symmetrically located 5 ft. from midspan (referred to as 5 and 10 ft. spreader tests, respectively). Cracking induced in the section to correlate with the environmental simulation strain levels A and B was generated with the use of both the 5 and 10 ft. spreader layout as well as patch loading; however, cyclic loading was always applied through a patch load at midspan. The reflective crack was extended to the outer (i.e., west) quarter point of Span 2, however the crack was not observed to have extended to the instrumentation at the interior quarter point, which was attributed to the transverse restraint provided by proximity to the nearby interior support, and was similar to the observed results from Span 1 of the Concept 1 laboratory bridge. The results of the environmental effect laboratory test on Span 1 are shown in Figure 5.3.5. In addition, Table 5.3.6 outlines the number of cycles completed, the strain that was induced in the section and the load required to induce that strain for Span 1 of the Concept 1 laboratory bridge. Figure 5.3.5: Transverse strains measured with 35 kip patch load applied at midspan during environmental effect simulation in Span 1 of Concept 1 laboratory bridge 0 20 40 60 80 100 120 Tr an sv er se S tr ai n [u e] Mid-Span Outer Quarter Pt. Inner Quarter Pt. Prior to cracking at 1M cycles of fatigue load After cracking at 1M cycles of fatigue load After cracking to enviro. strain level A After 15000 cycles to enviro. strain level A After cracking to enviro strain level B After 15,000 cycles to enviro. strain level B

136 Table 5.3.6: Measured load and transverse strain, and number of cycles completed during environmental effect simulation on Span 1 of the Concept 1 laboratory bridge specimen Span 1, Concept 1 Prior to cracking at 1M cycles of fatigue load After cracking at 1M cycles of fatigue load After cracking to enviro. strain level A After cycling to enviro. strain level A After cracking to enviro. strain level B After cycling to enviro. strain level B Magnitude of target strain NA1 160µε 180µε 180µε 300µε 300µε Number of cycles completed 3 3 3 15000 3 15000 Max. transverse strain achieved 18µε 161µε 213µε 213µε 280µε 303µε Load required to achieve strain 35k2 89k2 160k3 105k3 206k3 210k3 1The first column of Table 5.3.6 represents the strain measured during the fatigue loading. Fatigue loading was not based on strain levels, and but rather was based on an applied patch load of 35 kip, therefore no target strain was applicable to fatigue loading 2The load recorded represents the patch load applied at midspan. 3The load recorded represents the total load applied to the 10 ft. longitudinal spreader beam setup. Somewhat significant degradation of the joint with the application of the 35 kip patch load was observed due to both cracking and cycling at approximately 210 kips, or a measured strain of approximately 300µε. Furthermore, degradation of the outer quarter point at the environmental strain level A and above suggested that the use of the 10 and 5 ft. spreader beam for the application of the quasi-static patch loads reduced the stiffness toward the exterior support along the joint. It should be noted that although degradation was observed under the application of loading, a forensic examination of the specimen discussed in Section 5.4 indicated negligible residual damage in the unloaded specimen. The Concept 2 laboratory bridge consisted of two half spans of interest. For this reason, simulation of environmental effects was completed at the quarter points and at midspan separately, with the tests for environmental simulation strain level A completed first at the quarter points and then at midspan, with the same order of testing used at environmental simulation strain level B. The spreader setup consisted of load applied quasi-statically and dynamically through a 10 ft. long spreader beam centered at midspan. The reflective crack introduced in the section after the completion of the first million cycles of fatigue load was developed using a quasi-static patch load at the quarter points through the use of the 10 ft. spreader beam, and was also extended to midspan via a single patch load at that location, though cracking was observed only between midspan and just beyond the east quarter point, which corresponded to the half span constructed with straight transverse reinforcement in the trough region. Cracking induced in the section to correlate with the environmental simulation strain levels A and B was generated with the use of both a 5 and 10 ft. spreader layout as well as patch loading at midspan, though for all tests cracking was not observed in the instrumentation in the west half span of the specimen. An additional load test was designed with a spreader beam bearing a distance of 2.5 ft. towards the west half span, and 5 ft. towards the east half span, in an effort to increase the reaction force in the west half span and subsequently induce cracking near that location. Loading was applied in

137 a quasi-static load controlled load test, however the transverse strains measured at the midspan cross section reached the target values prior to the indication of the cracking at the west quarter point, and therefore no additional load was applied in this modified setup. The lack of observed cracking was attributed to three possible explanations. The west half span corresponded to the half span of the bridge bearing on the concrete pier (i.e., corresponding to the interior half spans in the Concept 1 laboratory bridge), in which cracking was also not observed attributed to transverse restraint provided by the support. The natural roughness of the concrete pier may have provided additional restraint to the transverse expansion of the joint at the pier, which may have reduced the transverse stress demands in the CIP concrete near the joint. The west half span was also constructed with hooked reinforcement rather than the straight bars in the adjacent half span, though cracking was observed in the previous specimens where larger diameter hooked bars at closer spacing were used. A third possibility was that a potential reflective crack in the west half span was not observed by the instrumentation. This may have been the case especially if the crack propagated along the precast flange-CIP interface to the vertical web interface of the precast section, where instrumentation was not present (A forensic examination of the specimen following testing did not reveal a crack at the vertical web interface; however, the forensic exam was conducted under no load such that cracks that had opened under loading may have closed upon unloading.). Figure 5.3.6 illustrates the behavior of the Concept 2 laboratory bridge under a 35 kip quasi-static patch load applied at the quarter points (i.e., a load of 70 kips applied to a 10 ft. spreader beam from midspan) at various times during the environmental simulation. Even though the Concept 2 laboratory bridge specimen had a single simply-supported span, the three data series in Figure 5.3.6 are named in the same way as for the Concept 1 laboratory bridge: the inside quarter was located nearest the concrete pier (i.e., west end of the Concept 2 laboratory bridge), while the outside quarter was located nearest the steel pier (i.e., east end of the Concept 2 laboratory bridge). Table 5.3.7 outlines the number of cycles completed, the strain that was induced in the section at both the east and west quarter points, and the associated load required to induce that strain for the Concept 2 laboratory bridge specimen.

138 Figure 5.3.6: Transverse strains measured with 35 kip patch load applied at quarter points during environmental effect simulation in the Concept 2 laboratory bridge 0 20 40 60 80 100 120 Tr an sv er se S tr ai n [u e] Mid-Span Outer (east) Quarter Pt. Inner (west) Quarter Pt. Prior to cracking at 1M cycles of fatigue load After cracking at 1M cycles of fatigue load After cracking to enviro. strain level A After 15000 cycles to enviro. strain level A After cracking to enviro strain level B After 5,000 cycles to enviro. strain level B

139 Table 5.3.7: Measured load and transverse strain, and number of cycles completed during environmental effect simulation at the quarter points of the Concept 2 laboratory bridge specimen Concept 2, load applied at quarter points Prior to cracking at 1M cycles of fatigue load After cracking at 1M cycles of fatigue load After cracking to enviro. strain level A After cycling to enviro. strain level A After cracking to enviro. strain level B After cycling to enviro. strain level B Magnitude of target strain NA1 160µε 180µε 180µε 300µε 300µε Number of cycles Completed 3 3 3 15000 3 5000 Max. transverse strain measured at the east quarter point 33µε 150µε 177µε 177µε 253µε 253µε Max. transverse strain measured at the west quarter point 30µε 85µε 81µε 81µε 104µε 104µε Load required to achieve strain2 70k 210k 195k 153k 210k 207k 1The first column of Table 5.3.7 represents the strain measured during the fatigue loading. Fatigue loading was not based on strain levels, but rather was based on an applied patch load of 35 kip, therefore no target strain was applicable to fatigue loading 2The load recorded represents the total load applied to the 10 ft. longitudinal spreader beam setup Relatively similar levels of degradation were observed in the Concept 2 laboratory bridge compared to Span 1 of the Concept 1 laboratory bridge. At the conclusion of cracking and cycling to a load that induced a strain representative of the environmental target strain level A, the strain measured when a 35 kip patch load was applied to each quarter point had increased by 22µε (i.e., 87µε - 65µε), compared to 9µε (i.e., 52µε - 43µε) and 17µε (i.e., 67µε - 50µε) for Spans 1 and 2 of the Concept 1 laboratory bridge, respectively. Furthermore, when loaded to the maximum capacity of the 220 kip actuator, which was limited to 210 kips due to the available system hydraulic pressure at the time of testing, and cycling to induce a strain of 253µε at the east quarter point, the strain measured with an applied 35 kip quasi- static patch load at each quarter point increased by 32µε (i.e., 119µε - 87µε), compared to a 51µε (i.e., 103µε - 52µε) increase in Span 1 of the Concept 1 laboratory bridge. Note however that the Concept 2 bridge only underwent 5000 cycles at that strain level as opposed to the 15000 cycles subjected to the Concept 1 bridge spans and the magnitude of strain induced in each span during the environmental simulation varied. The absence of cracking near the west quarter span was further supported by the minimal increase in the strain measured at that location over the duration of the environmental simulation. Cyclic loading was applied at midspan of the Concept 2 laboratory bridge after the completion of loading at the quarter points for both environmental simulation strain levels A and B (i.e., cyclic loading to level A was applied at quarter points, then midspan, followed by cyclic loading to level B at the quarter points, then midspan). The target strain level at the environmental simulation level B was modified for the

140 midspan tests, with the revised target strain to be the maximum transverse strain measured at the quarter points (i.e., 253µε at the east quarter point). The behavior of the specimen under a 35 kip quasi- static patch load at midspan through the duration of the environmental simulation is shown in Figure 5.3.7. The three data series in Figure 5.3.7 are named in the same way as for the Concept 1 laboratory bridge (i.e., the inside quarter is located nearest the concrete pier (west end of the Concept 2 laboratory bridge), while the outside quarter was located nearest the steel pier (east end of the Concept 2 laboratory bridge)). Table 5.3.8 outlines the number of cycles completed, the strain that was induced in the section at midspan, and the associated load required to induce that strain for the Concept 2 laboratory bridge. Figure 5.3.7: Transverse strains measured with 35 kip patch load applied at midspan during environmental effect simulation in the Concept 2 laboratory bridge 0 20 40 60 80 100 120 Midspan Outer (east) Quarter Pt. Inner (west) Quarter Pt. Prior to cracking at 1M cycles of fatigue load After cracking at 1M cycles of fatigue load After cracking to enviro. strain level A After 15000 cycles to enviro. strain level A After cracking to enviro strain level B After 15,000 cycles to enviro. strain level B

141 Table 5.3.8: Measured load and transverse strain, and number of cycles completed during environmental effect simulation at midspan of the Concept 2 laboratory bridge specimen Concept 2, load applied at midspan Prior to cracking at 1M cycles of fatigue load After cracking at 1M cycles of fatigue load After cracking to enviro. strain level A After cycling to enviro. strain level A After cracking to enviro. strain level B After cycling to enviro. strain level B Magnitude of target strain NA1 160µε 180µε 180µε 300µε 300µε Number of cycles Completed 3 3 3 15000 3 15000 Max. transverse strain measured at midspan 33µε 174µε 176µε 176µε 256µε 256µε Load required to achieve strain 35k 136k 116k 116k 124k 134k 1The first column of Table 5.3.8 represents the strain measured during the fatigue loading. Fatigue loading was not based on strain levels, but rather was based on an applied patch load of 35 kip; therefore no target strain was applicable to fatigue loading. Little degradation of the longitudinal joint was observed at midspan of the Concept 2 laboratory bridge during the environmental effect simulation. Note that despite the fact that loading was applied at midspan in Figure 5.3.7, a larger increase in the transverse strain was measured at the outer (east) quarter point than at midspan after cracking and cycling to simulate the environmental strain target level B, which suggested that the level of degradation achieved near the east quarter point was more severe than that at midspan. The resiliency of the joint region at midspan was likely due to the limited success in extending the reflective crack into the west half span of the specimen. The rate of degradation of the joint during cycling was also monitored during the environmental simulation on the Concept 2 laboratory bridge. Figure 5.3.8 illustrates the strain measured at midspan periodically with the 35 kip patch load located at midspan throughout the 15000 cycles that were completed to the maximum capacity of the actuator. The cycling portion of the environmental simulation to induce a strain to the target level B was ended at after the completion of 5,000 cycles because of a fracture failure of the steel spreader beam used for the test. The rate of degradation of the joint at the outside quarter point was observed to be larger during the early cycles and tended to taper off between 2,000 and 5,000 cycles. The increase in strain at the outer quarter point measured with the application of the 35 kip patch load over the first 2,000 cycles was 22µε, while an additional 3,000 cycles increased the strain by only 10µε. This suggested that the 5,000 cycles applied for the environmental simulation of the Concept 2 specimen provided an adequate estimate of the degradation that might be expected under the cyclic loads to a given strain condition.

142 Figure 5.3.8: Transverse strains measured at midspan under the 35 kip patch load applied at midspan during the 15,000 cycles completed during the environmental effect simulation In summary, Span 2 of the Concept 1 laboratory bridge showed excellent crack control during the cycling procedure at the environmental simulation strain level B, which corresponded to a measured value of approximately 233µε. Span 1 of the same bridge, which would be expected to have superior crack control capabilities because of the reduced flange depth, showed a degradation of the joint under the same target strain level, which corresponded to an approximate transverse strain of 303µε. Finally, the Concept 2 laboratory bridge, which was cycled to a maximum strain level of approximately 256µε, measured at the east quarter point, also showed degradation of that portion of the joint region during cycling. The Concept 2 span was constructed with less total reinforcement, however with a tighter spacing. In addition, the Concept 2 laboratory bridge was constructed with a 3 in. flange, which would be expected to outperform Span 2 of the Concept 1 laboratory bridge. Because of the similarity in the performance of the two bridge concepts, the details from either were expected to be acceptable for PCSSS bridge construction. Furthermore, the adequate performance of the Concept 1 laboratory bridge (with 12 in. maximum transverse reinforcement spacing) indicated that a maximum spacing of 12 in. was suitable for the design of PCSSS bridges based on this research. Recall that the variations in the depths of the instrumentation utilized to measure the transverse strains (see Table 5.3.1) in each specimen prevented direct comparison of the relative magnitudes of measured strain among the specimen. Assuming that the transverse strain varied linearly between the compression and tension fibers of the section, instrumentation that was located lower in the section was associated with a smaller transverse curvature than instrumentation that was located higher in the section when both gages measured identical strains. A forensic exam was completed on each of the laboratory bridge specimens in an effort to visually identify reflective cracking by means of an investigation of core samples. After saw cutting the specimens into several sections, as discussed in detail in Section 5.4, negligible evidence of reflective cracking was observed in the forensic exam, despite the fact that many locations selected for core samples coincided with the location of concrete embedment instrumentation which reported large, 0 20 40 60 80 100 120 0 2000 4000 6000 8000 10000 12000 14000 16000 Tr an sv er se S tr ai n [u e] Cycles of Load Inducing 256ue

143 inelastic strains during the load tests. Because the forensic exam was conducted following the tests, the cracks may have closed due to the load removal. In addition to the environmental effect simulation using the procedures described above with the application of mechanical loading, a setup using commercial grade heated thermal blankets was applied to Span 2 of the Concept 1 laboratory bridge in an attempt to induce the same thermal gradient in the laboratory bridge as measured in the Center City Bridge. This method was found to be incapable of inducing the required thermal gradient in the specimen, and was not considered feasible in the future. 5.3.3. Load Transfer Between Precast Panels The ability of the section to effectively transfer load between adjacent precast panels was expected to depend on the condition of the longitudinal joint and therefore was expected to change throughout the extent of laboratory testing as cracking was induced along the longitudinal joint. To investigate the load transfer between adjacent panels, patch loading was applied at the middle of the south precast panel, approximately 36 in. from the precast joint, directly at midspan of Span 2 of the Concept 1 laboratory bridge as well as separate tests at midspan and the quarter points of the Concept 2 laboratory bridge. The locations of the applied load for the load transfer tests are shown in Figure 5.3.9. Load transfer was not investigated during testing of Span 1 of the Concept 1 laboratory bridge. Load transfer tests were completed on Span 2 prior to traffic fatigue or introduction of cracking, as well as at the conclusion of all fatigue loading and environmental simulations. Several load transfer tests were completed on the Concept 2 laboratory bridge throughout the series of traffic fatigue cycles as well as intermittently during the environmental simulation. Monitoring of the longitudinal instrumentation during the load transfer tests allowed for the calculation and comparison of the longitudinal curvature in the loaded and unloaded panels. Span 2 of the Concept 1 laboratory bridge was first loaded on the south panel prior to any fatigue or cracking loading in November 2007. The second load test on the south panel was completed at the conclusion of all fatigue, cracking, and environmental simulation on both spans in October 2008 at the completion of roughly 2,030,200 cycles of fatigue and environmental loading on Span 2. Figure 5.3.10 illustrates the measured longitudinal curvature in the north and south panels at midspan in Span 2 of the Concept 1 laboratory bridge. The gages utilized through the depth in both panels were nominally located at midspan approximately 21 in. from the precast joint.

144 (a) Load placement during transverse load distribution tests for Concept 1 laboratory bridge (b) Load placement during transverse load distribution tests for Concept 2 laboratory bridge Figure 5.3.9: Load placement during transverse load distribution tests for Concept 1 and Concept 2 laboratory specimens Span 2 35 kip Span 1 35 kip 35 kip 35 kip Together as spreader load

145 Figure 5.3.10: Longitudinal curvature in north and south panels in Span 2 of Concept 1 laboratory bridge under 35 kip patch load applied at midspan centered over south panel The curvature at each data point was calculated using three longitudinally oriented strain gages in a vertical line. The coefficient of variation (R2) value for the curvature ranged from 0.961 to 0.988 for the four data points shown in Figure 5.3.10. The reduction in measured curvature between the tests at zero and two million cycles might be attributed to variations in the loading and instrumentation setup between the two measurements, as the zero and two million cycles were completed by different personnel. Load transfer was more thoroughly investigated in the Concept 2 laboratory bridge, with the longitudinal curvature measured several times throughout the laboratory tests conducted by the same personnel. As for the tests on Span 2 of the Concept 1 laboratory bridge, the south panel of the Concept 2 laboratory bridge was loaded both at midspan as well as at the quarter points with a 35 kip patch load at each location. Figure 5.3.11 illustrates the curvature calculated throughout the fatigue and environmental cycles with a 35 kip patch load applied to each quarter point through a longitudinal spreader. The last data points in the figure were taken after completion of environmental cycling to the 160µε strain level. The longitudinal curvature was not obtained after the environmental simulation cycles to the 250µε strain level due to failure of the spreader beam during those tests. -16 -14 -12 -10 -8 -6 -4 -2 0 0 2 Lo ng it ud in al C ur va tu re [u e/ in ] Number of Cycles [Millions] North South

146 Figure 5.3.11: Longitudinal curvature in north and south panels in Concept 2 laboratory bridge under 35 kip patch load applied at the quarter points centered over south panel. The longitudinal curvatures were observed to increase in the Concept 2 laboratory bridge. The increase in curvature observed in the Concept 2 laboratory bridge was not suspected to be a result of longitudinal flexural cracking, as the predicted cracking loads for Span 2 of the Concept 1 laboratory bridge and the Concept 2 laboratory bridge were 271 and 224 kips, respectively, while the largest loads applied to Span 2 of the Concept 1 laboratory bridge and the Concept 2 laboratory bridge were 155 and 210 kips, respectively. It was expected that the longitudinal curvatures measured in the north precast panels would decrease as reflective cracking was introduced and extended near the precast joint because of reduced ability for the section to transfer load across the joint region, though this was not observed. Rather, good transverse load transfer was observed across the longitudinal joint regardless of the application of fatigue loading to simulate traffic loads, the introduction of reflective cracking, or even the application of cyclic loading to simulate the large daily strain fluctuations observed in the Center City field bridge due to environmental effects. 5.3.4. Composite Action Each of the three bridge spans (i.e., Concept 1 Span 1 and 2, and Concept 2) provided an opportunity to investigate the effects of variations in horizontal shear reinforcement on composite action. Span 2 of the Concept 1 laboratory bridge was constructed with horizontal shear ties spaced at 12 in., though the hook returns provided a nominal 1/4 in. of clearance to the top of the precast web, which was unlikely to provide adequate clearance for proper bond. The horizontal shear reinforcement in this span reflected that used in the Center City Bridge, where, according to Article 5.8.4.1 of the AASHTO LRFD -16 -14 -12 -10 -8 -6 -4 -2 0 0 0.5 1 1.5 2 2.5 Lo ng it ud in al C ur va tu re [u e/ in ] Number of Cycles Completed [Millions] North South

147 specification (2004), the maximum stirrup spacing at the ends of the precast beam was 15 in., which could be reduced to 24 in. at locations away from the ends (where the shears were smaller). The horizontal shear reinforcement spacing was doubled to 24 in. in Span 1 of the Concept 1 laboratory bridge and the clearance between hook returns and the top of the precast web was increased to 1-3/8 in., while no horizontal shear reinforcement was provided in the Concept 2 laboratory bridge. The 2010 AASHTO LRFD specification required the presence of a minimum area of horizontal shear reinforcement, though Naito et al. (2006) suggested that this requirement was overly conservative (see Section 2.3), which prompted the investigation of the performance of the PCSSS in the absence of horizontal shear reinforcement. In all three cases, a roughened surface was provided by means of raking, which created a surface roughness amplitude of approximately 1/4 in, and was provided on the horizontal top web surface as well as both vertical web faces in each specimen. The top surface of the precast flanges was also roughened in Span 2 of the Concept 1 laboratory bridge. A photograph of the surface condition of one of the precast members utilized in the Concept 2 laboratory bridge is shown in Figure 5.3.12. Figure 5.3.12: Intentionally roughened surface, by means of raking, of top web of precast beam used for the construction of Concept 2 laboratory bridge specimen The ability of the precast slab span system to maintain composite action throughout all loading scenarios warranted investigation of composite action at the largest possible load levels available in the

148 laboratory. For this reason, load tests were conducted on each span to determine the maximum capacity of each specimen, and subsequently to determine whether composite action was maintained throughout each test. A total of between 3 and 6 longitudinally oriented concrete embedment resistive gages aligned vertically in the CIP and precast concrete at several cross sections in each specimen were utilized to calculate the longitudinal curvature during the ultimate tests. An analysis of the curvatures provided an assessment of the state of continuity between the precast and CIP concrete. The application of load during the ultimate tests was expected to create an artificially compressed region near the location of applied load, which was likely to increase the horizontal shear friction (due to the localized increase in the normal force) at that location. Therefore, the location of the applied loading during the ultimate tests was modified in an effort to reduce the artificial effects of loading on the longitudinally oriented instrumentation. In both spans of the Concept 1 laboratory bridge, the instrumentation utilized for the investigation of composite action was located within a longitudinal distance of 3 in. from the application of load at midspan. The load frame was therefore modified, and relocated 20 in. closer to the center pier, which provided a center to center distance of 17 in. from the load beam to the instrumentation. The longitudinally oriented instrumentation in the Concept 2 laboratory bridge was designed to allow for the load frame to remain in the same position for the ultimate loading as was utilized for the fatigue and environmental studies, with a center to center distance of 21 in. between the instrumentation and load beam. Each test was preceded by the construction of an even grouted surface between the top of the bridge and the loading beam, which extended across the entire width of the bridge and had a flange width of 12 in. Three actuators, a 220 kip in the center with a 110 kip actuator to each side, were used to apply load to the transverse load beam. Load was applied to each specimen using a displacement-controlled program, primarily to prevent the rapid collapse of the specimens, if applicable. The displacement rate was selected such that the load was increased by approximately 2 kips per minute. A photograph of the loading system utilized for the ultimate tests is provided in Figure 5.3.13.

149 Figure 5.3.13: Photograph of tri-actuator load setup and transverse loading beam utilized during ultimate load tests; shown for the Concept 2 laboratory bridge test The calculation of the sectional ultimate moment capacity for each span was completed using Response- 2000. The assumed tensile strength of the prestressing strand was taken to be 270 ksi, the yield strength of the mild reinforcement as 60 ksi, and the concrete compressive strain at crushing was assumed to be 0.003. The ultimate sectional moment capacity of each section varied slightly, due to the depths of the longitudinal cage reinforcement based on the flange thickness, number of strands in each span, as well as variations in the CIP concrete compressive strength and the finished depth of each span due to construction tolerances. The maximum measured depth of the Concept 1 laboratory bridge was 19.25 in., with concrete compressive strengths of 4.47 ksi and 4.55 ksi at an age of 432 days in Spans 1 and 2, respectively. The measurement of the compressive strength at 432 days suggests that the value should reasonably represent the maximum strength and reasonably predict the strength during ultimate loading, which occurred when the CIP concrete was 855 and 784 days old in Spans 1 and 2, respectively. The maximum measured section depth in the Concept 2 laboratory bridge was 18.5 in. with a CIP concrete compressive strength of 6.9 ksi measured 10 days prior to ultimate loading on the Concept 2 laboratory bridge. The ultimate sectional moment capacity of Span 2 of the Concept 1 laboratory bridge was calculated to be 2544 ft.-kip, and the calculated capacity of Concept 1 Span 1 was 2256 ft.-kip. The sectional capacity of the Concept 2 laboratory bridge was calculated to be 2313 ft.-kip.

150 A numerical analysis of the continuous two-span Concept 1 laboratory bridge suggested that the ultimate capacity of the specimen would not be reached, as the maximum load available during the Concept 1 ultimate tests was approximately 420 kips, which corresponded to an applied maximum moment of 1272 ft.-kips, which was considerably less than the ultimate sectional capacities of either Span 1 or 2. The continuous bridge was modeled as a three-span system, with the center span modeled with an 18 in. span length (i.e., the distance between the center of bearing of Spans 1 and 2 at the center pier). The center span was modeled with a moment of inertia that was an order of magnitude larger than that of the exterior spans, to simulate the stiffness provided by the pier. Ultimate loading was initiated on Span 2 of the Concept 1 laboratory bridge, with a maximum applied load of approximately 420 kips. The ultimate capacity of the section was not observed to be imminent, which corresponded with the previously stated predictions. In an effort to reach the ultimate capacity of Spans 1 and 2, saw cutting was completed at the center pier in an attempt to reduce the continuity in the bridge (i.e., create approximately simple-support conditions) to reach the maximum flexural capacity of the laboratory bridge specimens with the available actuators. The cutting blade had a usable cutting radius of approximately 5 in, and therefore only the longitudinal deck reinforcement was severed, while the longitudinal reinforcement for continuity at the bottom of the reinforcing cage through the joint remained intact. Spans 1 and 2 of the Concept 1 laboratory bridge after continuity was modified at the center pier are referred to as semi-simple spans hereafter. The ultimate capacities of the two semi-simple spans in the Concept 1 laboratory bridge and the simple span in the Concept 2 laboratory bridge were not achieved during testing due to limitations of the actuators. The maximum load available during the Concept 1 tests was approximately 420 kips, while an applied load of 458 kips was achieved during the Concept 2 test, due to an increase in available hydraulic pressure. The applied loads and respective moments, as well as the calculated sectional capacities of each specimen are given in Table 5.3.9. Table 5.3.9: Maximum loads applied to laboratory bridge specimens during ultimate loading, calculated applied moments, and predicted moment capacities Specimen Max. applied load Calculated max. applied moment Predicted ultimate sectional capacity Concept 1, continuous span 420 kip 1272 ft.-kip 2256 ft.-kip for Span 1 2544 ft.-kip for Span 2 Concept 1, Span 1, semi-simple span 420 kip 2158 ft.-kip 2256 ft.-kip Concept 1, Span 2, semi-simple span 420 kip 2158 ft.-kip 2544 ft.-kip Concept 2 458 kip 2450 ft.-kip 2313 ft.-kip In all cases, ultimate failure of the specimens was not observed to be imminent. Loading was conducted at a displacement controlled rate; no reduction in applied load was observed during the tests, suggesting that composite action was maintained throughout the applied load range. In addition, visual inspections of the CIP-precast interface did not reveal any indication of slippage between the CIP and precast section on the external faces. A composite section should experience linearly varying strains through the full depth of the section, while loss of composite action would be evinced if distinct curvatures were observed in each section

151 (i.e., the CIP top section and the precast bottom section). Therefore, an analysis of the longitudinal curvatures in each section was conducted. The Concept 2 laboratory bridge specimen was well instrumented for the calculation of longitudinal curvature. At least three longitudinally oriented instruments were included in both the precast and CIP concrete. Good linear correlation through the section was observed among the instruments used to calculate the curvature at most locations. The change in curvature, measured between an applied load of 0 kips and 458 kips, and correlation coefficients, among other quantities determined during the ultimate test on the Concept 2 laboratory bridge are summarized in Table 5.3.10. Table 5.3.10: Longitudinal curvature and relevant values during ultimate loading on Concept 2 laboratory bridge specimen Location1 Number of Gages Load [kip] Curvature [1/in.] NA Depth [in.] R2 N-Mid-40” 7 458 -454.5 14.0 0.991 N-Mid-20” 6 458 -341.3 12.8 0.977 N-Mid-3” 6 458 -476.2 14.3 0.930 S-Mid-20” 6 458 -416.7 12.9 0.975 N-Qtr-20” 6 458 -201.0 13.0 0.995 N-Qtr-40” 7 458 -196.1 13.0 0.981 N-Qtr-3” 5 458 -213.2 12.4 0.954 1The instrumentation location was designated as follows: North/South precast panel – Mid- span/Quarter point – Distance to gage set from precast joint As an example, the longitudinal curvature measured at midspan of the north panel, 20 in. from the precast joint (i.e., N-Mid-20”) is shown in Figure 5.3.14. The curvature measured at this location reasonably represents the results of the ultimate load test at all locations. The linear trend line appears to provide a good fit to the instrumentation both above and below a vertical depth of 12 in. measured from the bottom of the precast, the location of the top of the precast web-CIP interface, suggesting that composite action was maintained at the applied load of 458 kips. The measured curvature was calculated based on the linear trend line, where the curvature was the reciprocal of the slope of the line.

152 Figure 5.3.14: Longitudinal strains measured through the section depth at midspan of north panel 20 in. from precast joint with 458 kip applied Load in Concept 2 laboratory bridge specimen The compression force developed in the section was calculated by integrating the stress profile using the measured strain profile and assuming the modified Kent and Park (Park et al., 1982) concrete stress- strain model with a nominal concrete compressive strength, f’c, of 5865 psi, which was equal to 85 percent of the measured concrete compressive strength of 6.9 ksi, and ε0 = 0.002. Eighty-five percent of the measured concrete compressive cylinder strength was utilized to provide an estimate of the unconfined concrete compressive strength in a structural member. Integrating the compressive stress above the neutral axis and multiplying by the deck width, the total compressive stress developed in the Concept 2 laboratory bridge was 2134 kips. The reduction in compressive force transferred through the composite surface due to the concrete in tension must also be considered. Using the measured modulus of rupture, the uncracked concrete in tension was assumed to provide a tensile force of 46 kips in the CIP. Therefore, the compressive force transferred between the CIP and precast concrete was 2088 kips. Dividing the total force by the width of the deck, including the area above the longitudinal trough, and half of the center to center of bearing span length, the measured stress transferred through the unreinforced composite surface was determined to be 135 psi. Because the unreinforced Concept 2 laboratory bridge span provided good horizontal shear capacity throughout loading to the maximum load of the actuators, which was nearly the ultimate capacity of the section, it was recommended that AASHTO Article 5.8.4 be modified to allow sections unreinforced for horizontal shear to develop a nominal horizontal shear strength of up to 135 psi. y = -0.00293x + 12.83136 R² = 0.97775 0 2 4 6 8 10 12 14 16 18 20 -3000 -2000 -1000 0 1000 2000 3000 4000 V er ti ca l D ep th in S ec ti on [i n] Longitudinal Strain [ue] precast inverted-T CIP concrete

153 5.4. Destructive Testing of Large-Scale Laboratory Specimens At the conclusion of the tests, cores were taken at locations of interest through the laboratory bridge specimens to investigate the location and extent of cracks that were believed to have developed during the tests based on the strain measurements. The information obtained from the cores is summarized in Section 5.4.1. The specimens were also sliced into segments small enough such that they could be removed from the laboratory with the overhead crane. The information obtained from a visual inspection of the sliced surfaces to determine the extent of any residual cracking is summarized in Section 5.4.2. 5.4.1. Inspection of Cores taken from Laboratory Bridge Specimens At the conclusion of laboratory testing a series of cores were removed from both the Concept 1 and Concept 2 laboratory bridges to provide a means of physical investigation of the cracking behavior near the precast joint. The cores were utilized to document a maximum crack width and length in each core, the magnitude of the crack width was dependent on the maximum loading and specific boundary conditions in each span, and therefore would not be directly comparable between spans. The diameter of cores ranged between 2, 3, and 4 in. Each of the core specimens was examined both with the naked eye and the aid of an Olympus SZX12 stereo-microscope to find cracks on the core surface. The level of magnification used to examine the cores ranged between 2.1X to 27X, which was the full capacity of the microscope. All cracking that was identified in each core was tabulated, regardless of the size or anticipated origin. The observed crack widths were documented in classification categories, defined in Table 5.4.1. The vertical depth of the cracking identified in the cores was referenced from the line created by the horizontal precast flange-CIP interface, as shown in Figure 5.4.1. A complete description of the characteristics and measured values obtained from each core sample is tabulated in Appendix G. Table 5.4.1: Crack width classification categories for analysis of core specimens Crack Classification Crack Width (W) 0.002 W < 0.002 in. A 0.002 in. ≤ W < 0.008 in. B 0.008 in. ≤ W < 0.023 in. C 0.023 in. ≤ W < 0.200 in. D W ≥ 0.200 in.

154 Figure 5.4.1: Location of reference line for measurement of vertical location of cracking in core specimens Core samples were extracted from both spans of the Concept 1 large-scale laboratory bridge. A total of five cores were removed from Span 1, while four cores were extracted from Span 2. All the cores from the Concept 1 laboratory bridge specimen were nominally located directly over the longitudinal precast joint (corresponding to the red core outline in Figure 5.4.1), at various locations along the length of each span. Table 5.4.2 shows the core description and location, as well as observed crack width and depth. The specimen identification consists of the bridge concept number-span number-core number. The x- coordinate of the core location corresponds to the coordinate system with the origin at the center of the pier, in accordance with the locations of the instrumentation; a negative x-coordinate corresponds to Span 1; while a positive x-coordinate corresponds with Span 2. After the first core was removed from Span 1 of the Concept 1 laboratory bridge, a saw cut was completed along the entire length of the bridge near the precast joint, after which specimens C1-S1-2 through C1-S1-5 were removed from the span. These cores therefore consisted of two half cylinders. The thickness of the cylinder removed due to the saw cut was approximately 1/8 in. wide through the length of the core near where the precast joint was located. Because of this, evidence of cracking in these specimens may have been removed from the saw cut, however the only core from the Concept 1 laboratory bridge where a crack was clearly measured was in a core specimen that had been cut. The other potential cracks were too small to be seen even with the magnification of the microscope.

155 Table 5.4.2: Summary of maximum height and width of cracks measured in core specimens from the Concept 1 laboratory specimen Specimen x-Coordinate of Core (in.) Saw Cut? Diameter of Core (in.) Maximum height of crack1 (in.) Maximum width of crack2 (in.) C1-S1-1 -194.5 No 1.75 Undefined4 Undefined C1-S1-2 -26 Yes 2.75 NO3 NO C1-S1-3 -134 Yes 2.75 NO NO C1-S1-4 -192 Yes 2.75 NO NO C1-S1-5 -197.5 Yes 2.75 2.75 (up) A C1-S2-1 74 No 1.75 NO NO C1-S2-2 146 No 1.75 NO NO C1-S2-3 195.5 No 1.75 NO NO C1-S2-4 198 No 1.75 NO NO 1The height of crack was measured from the reference line, defined in Figure 5.4.1; only “upward” values are recorded in this table 2The width of crack was documented by crack classification, as defined in Table 5.4.1 3”NO” represents “No reflective cracks observed” 4 “Undefined” represents that the reflective crack was continuous with a shrinkage crack from the deck surface, therefore beginning and end points were undefined The only evidence of internal cracking observed via the investigation of the core samples was in specimen C1-S1-5, which was in the outer quarter of Span 1, despite clear evidence from the instrumentation that cracking occurred at several other locations. Strain increases as large as 303 µε and 233 µε were measured in Spans 1 and 2, respectively, via the concrete embedment resistive gages under loading. Considering the 4.7 in. gage length, a maximum associated crack width of 0.0011 and 0.0014 inches might be expected in all the core samples except C1-S1-2 because of the proximity to the continuous pier. While the expected crack widths of approximately 0.001 inches were smaller than the 0.002 inch minimum reading on the crack gage, the use of the microscope allowed for significantly smaller cracks to be identified, especially those on the order of 0.001 inches. From the forensic examination, it appears that the cracks may have closed upon unloading. The Concept 2 laboratory bridge was similarly cored at the conclusion of testing. In this case, eight cores were removed from the specimen, four from each quarter point. At each quarter point, two cores were removed from above the joint, 4 in. apart, while the remaining two were centered over the vertical precast webs. Cores were taken from the web areas to document any cracks or separation at the CIP- precast web interface at those locations, which should be avoided because the cage reinforcement provides no benefit near the vertical web interfaces. Table 5.4.3 includes a summary of the measured crack widths and depths in the cores from the Concept 2 specimen. No saw cutting was present in the Concept 2 laboratory bridge specimen prior to the removal of the core samples. The measured core locations in the Concept 2 laboratory bridge corresponded with the origin and sign convention utilized for the placement of the instrumentation.

156 Table 5.4.3: Summary of maximum height and width of cracks measured in core specimens from the Concept 2 laboratory specimen Specimen x-Coordinate of Core (in.) y- Coordinate of Core (in.) Diameter of Core (in.) Maximum height of crack1 (in.) Maximum width of crack2 (in.) C2-S1-1 193 0 3.75 NO3 NO C2-S1-2 189 0 3.75 2.25 0.003 C2-S1-3 193 +12 3.75 3.5 A C2-S1-4 193 -12 3.75 NO NO C2-S1-5 76 0 3.75 NO NO C2-S1-6 80 0 3.75 NO NO C2-S1-7 76 +12 3.75 NO NO C2-S1-8 76 -12 3.75 NO NO 1The height of crack was measured from the reference line, defined in Figure 5.4.1; only “upward” values are recorded in this table 2The width of crack was documented by crack classification, as defined in Table 5.4.1 3”NO” represents “No reflective cracks observed” Reflective cracking was observed in one of the four cores located over the joint in the Concept 2 laboratory bridge specimen, and was located in the east half span (C2-S1-2). In addition, a vertical crack was also observed at the vertical CIP-precast web interface (C2-S1-3), also located in the east half span. It is unclear if the discontinuity observed at this location was truly a crack, or was separation of the precast web and CIP concrete due to poor bond, or was simply an illusion due to the interface between the two types of concrete, however it was expected that cracking or separation at the vertical web interface would relieve transverse stresses at the longitudinal joint, which would likely hinder crack growth over the joint, however cracking was measured extensively near the east quarter point of the span, suggesting that the vertical web face cracking was unlikely. Furthermore, the fact that no cracking or separation of vertical CIP-precast web interface at the west quarter point suggested that it was unlikely that the lack of observed cracking indicated by the instrumentation over the joint in the west half span of the Concept 2 specimen was due to cracking at the vertical web interface and subsequent relief of transverse stresses near the longitudinal precast joint. The largest transverse strain measured during load testing on the Concept 2 laboratory bridge was approximately 256 µε, measured near midspan, which over the 4.7 in. gage length of the concrete embedment resistive gage would correspond to a maximum crack width of approximately 0.0014. This was a strain measured under loading to simulate the environmental effects. A likely reason that the cracks were not identifiable with the microscope was because the cores were removed from the bridge upon unloading and examined under no load conditions.

157 5.4.2. Visual Inspection of Internal Surfaces after Sectioning the Laboratory Bridge Specimens by Means of Saw Cutting Capacity constraints of the overhead crane system present in the University of Minnesota Structures Laboratory necessitated that both the Concept 1 and Concept 2 laboratory specimens be split into segments of no more than 30,000 lbs in total weight for removal. This created an opportunity to visually inspect the internal cut faces of the specimens for signs of cracking. For this reason, the laboratory bridge specimens were segmented such that the locations of the cuts were near locations of expected cracking, as determined via the data analyses. The two span Concept 1 laboratory bridge specimen was cut into a total of eight sections, as shown in Figure 5.4.2. Figure 5.4.2: Concept 1 laboratory bridge specimen partitioning for saw cutting procedure The longitudinal saw cut, which ran the length of the Concept 1 laboratory bridge specimen was nominally located directly through the precast joint, though drift of the saw during the cutting process caused the cut to vary between +/- 1 in. to either side of the precast joint. Unfortunately, the location of the cut line near the precast joint allowed for the possibility of the saw kerf to remove the evidence of reflective cracking in the section. No signs of reflective cracking were observed on any of the cut faces in the Concept 1 laboratory bridge specimen, despite the use of a wetting and drying process in the region where reflective cracking was expected (i.e., the precast trough region) in which water was sprayed on the surface and allowed to air dry. This process was expected to highlight small cracks, where the ingress of water into the cracks would emphasize any cracking on the cut surfaces. The vertical and horizontal interfaces between the precast concrete and CIP concrete however tended to feature unique characteristics. In the case of the horizontal precast – CIP interface, delamination of the PC flanges from the CIP concrete was observed, as shown in Figure 5.4.3. This documented evidence

158 of delamination at the horizontal interface was observed in all cross-sectional cuts, regardless of the presence of the smooth or roughened precast flange. No evidence of delamination or cracking was observed at the vertical PC web interface on any of the cut surfaces of the Concept 1 laboratory bridge. Figure 5.4.3: Delamination at the PC – CIP interface in the East transverse face of section number 7 of the Concept 1 Laboratory bridge specimen The Concept 2 laboratory bridge specimen was similarly sectioned to accommodate removal from the laboratory. The cutting plan for this specimen included a total of three transverse cuts, at each quarter point of the span. Recall that the quarter points of the span coincided with the locations of the loading during the tests. The cutting plan and numbering convention is shown in Figure 5.4.4. Flange thickness = 5.25 in. Location of precast joint in Concept 1 i Evidence of delamination at interface

159 Figure 5.4.4: Concept 2 laboratory bridge specimen partitioning for saw cutting procedure Evidence of reflective cracking was not observed in any of the cross sections in the Concept 2 laboratory bridge specimen. As observed in the Concept 1 laboratory bridge specimen, some delamination of the horizontal PC – CIP interface was observed at midspan on both the west face of panel 3 and the east face of panel 2, as illustrated in Figure 5.4.5. In addition, a diagonal crack was observed on the east face of panel 2 originating near the precast flange interface and propagating towards the corner of the precast section, as shown in Figures 5.4.6 and 5.4.7. The crack was measured to be approximately 0.006 in. As discussed in Section 5.4, cracking was observed in the east quarter of the Concept 2 laboratory bridge north of the joint via the core analysis, though was not detected via the analysis of the cut faces at the corresponding location. Figure 5.4.5: Delamination at the PC – CIP interface observed on the West face of section number 3 of the Concept 2 laboratory bridge specimen Flange thickness = 3in. Evidence of delamination at interface

160 Figure 5.4.6: Evidence of diagonal cracking in the trough region on the east face of panel 2 in the Concept 2 laboratory bridge PC flange interface Diagonal crack Corner of PC web Cage reinforcement

161 Figure 5.4.7: Close up view of diagonal crack identified on east face of panel 2 in the Concept 2 laboratory bridge Precast – CIP interface Diagonal cracking