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.

233 Chapter 7 PCSSS: Conclusions and Recommendations 7.0 Introduction Several numerical and experimental investigations were completed and reviewed during the NCHRP 10-71 project related to issues of importance to the design and performance of precast composite slab span system (PCSSS) bridges. The results from each of these parts of the study were synthesized to develop a relevant design guide for precast composite slab span bridges. The following sections provide a brief summary of the results that were obtained during these studies. Numerical studies included an investigation of bursting and spalling stresses in the end zones of precast inverted-T sections, effects of spacing of transverse reinforcement in the joint region, and an investigation of the applicability of the AASHTO (2010) specifications for slab-type bridges to the design of PCSSS bridges for live load distribution factors and for consideration of effects of skewed supports. Experimental studies were completed on two large-scale laboratory bridge specimens, including a Concept 1 two-span continuous bridge and a Concept 2 simply-supported bridge specimen. The Concept 1 laboratory bridge included variations in a number of parameters including precast flange depth and end zone reinforcement details. It was also instrumented in a study by the Minnesota Department of Transportation (Mn/DOT) to investigate the effects of restraint moment and potential development of reflective cracking (Smith et al. 2008). The Concept 1 laboratory bridge was made available to the NCHRP 10-71 project for further study. In the NCHRP 10-71 study, the performance of both bridge specimens was investigated under various types of loading, including cyclic loading to simulate traffic, loading to simulate environmental effects, as well as to investigate load transfer between adjacent precast panels (both longitudinally and transversely). One of the keys to the effectiveness of the PCSSS bridge, which also makes it unique relative to slab-type bridges, is the need to control potential reflective cracking in the cast-in-place (CIP) concrete above the longitudinal joint between the adjacent precast flanges. A significant effort in the project was devoted to the investigation of crack control reinforcement and the simulation of potential crack development which was noted in a field instrumented PCSSS in Center City Minnesota. The Concept 1 specimen included No. 6 transverse hooked reinforcement embedded into the precast webs to provide load transfer and crack control in the joint region, as well as No. 5 cage stirrups which contributed to the crack control reinforcement. The nominal maximum spacing between transverse reinforcement was 12 in. (similar to the detail of the Center City Bridge). The Concept 2 specimen included No. 4 embedded hooked reinforcement in the west half of the simple span, while No. 4 straight embedded bars mechanically connected to reinforcement in the precast webs were provided in the east half span. No. 3 cage stirrups were staggered in the Concept 2 laboratory bridge relative to the transverse reinforcement spaced at 18 in. to provide a maximum spacing of 9 in. between transverse reinforcement. In addition to the two large-scale laboratory bridge specimens, six subassemblage specimens were tested to investigate the relative performance of various reflective crack control reinforcement details. The subassemblage specimens were loaded to flexurally induce cracking above the longitudinal joint between the precast flanges. The size, quantity, and location of cracking were documented through a range of quasi- static and cyclic load tests.

234 The culmination of the NCHRP 10-71 was the development of a design guide and examples which is included in Appendices A and B for application to PCSSS bridges. The following sections provide a summary of the investigation and conclusions from the components of the study. The order of the topics focuses on issues related to the precast inverted-T element of the bridge and progresses to system issues. 7.1. Bursting, Splitting and Spalling Stresses The AASHTO (2007) specification included design procedures to address bursting stresses at the ends of precast prestressed concrete beams due to the effects of the development of prestress, which were likely developed specifically for I-girders. These guidelines required large amounts of vertical reinforcement to be placed within a distance of h/4 from the end of the member. The shallow section depths of inverted-T precast beams resulted in limited area for the placement of vertical reinforcement, which led to significant congestion. Significant changes to the specification in regards to end zone stresses have been incorporated since 2007, specifically in the terminology related to end zone stresses. The 2008 interim AASHTO LRFD specification indicate that “bursting” has been replaced with “splitting” in terms of the resistance of pretensioned anchorage zones, and also represented a wider range of precast member shapes (i.e., not necessarily developed specifically for I-girders), as opposed to the specifications up to and including the 2007 LRFD specifications. The 2008 Interim specifications relaxed the placement requirements for wide-shallow sections, by allowing the designer to spread the end zone reinforcement, termed “splitting” reinforcement over a larger distance. In the case of pretensioned solid or voided slabs, the designer can substitute the section width for “h,” rather than using the section depth for “h.” According to the NCHRP 10-71 study, this may not be appropriate when trying to control spalling stresses. In addition, the terminology for the reinforcement described in this section of the AASHTO specifications should be termed “spalling” reinforcement rather than “splitting” or “bursting” reinforcement. Experimental and numerical studies were completed in the NCHRP 10-71 study to investigate the effects of end zone stresses on the precast prestressed inverted-T sections used in the PCSSS. The experimental results from the Concept 1 and 2 laboratory bridge investigations indicated that the 12 in. deep concrete sections had sufficient strength to resist tensile stresses induced in the transfer zone of the precast inverted-T sections at the time of release. Four unique end regions of the Concept 1 laboratory bridge specimen precast members, did not exhibit any evidence of cracking in those regions, even where vertical reinforcement was not provided in the end zones of those specimens. These findings were corroborated with the results of numerical studies that showed certain inverted-T members did not require spalling reinforcement, specifically those members with depths less than 22 in. for which the expected concrete strength was larger than the expected vertical tensile stresses due to the development of prestress. It was also found that for deep inverted-T sections, in particular, the existing requirements specified by the LRFD specification (AASHTO 2010) may be unconservative. Larger amounts of spalling reinforcement than specified by AASHTO 2010 were found to be required through a numerical study. It was also found that the reinforcement should be placed as close to the end of the member as possible (i.e., within h/4 of the end of the member, where “h” represents the depth of the member). The end region was the most critical region for the reinforcement to be located to address spalling stresses, even for the case of wide sections.

235 7.2. Restraint Moment It is important to consider the effects of restraint moment in the design of PCSSS or design the systems as a series of simple spans. The current specification allows that, when the age of the girder is at least 90 days at the time of continuity, the computation of restraint moments is not required (AASHTO 2010 Article 5.14.1.4.4). The reasoning lies in the fact that, when the girder has aged beyond 90 days, the positive restraint moments caused by the precast beams due to time-dependent effects are minimal, and the negative restraint moments that may be generated can be accommodated by the negative moment reinforcement over the piers. In the case of PCSSS bridges made continuous by casting the CIP concrete on relatively young girders (e.g., 7 to 14 days old) to complete the continuous composite bridge system, the effects of positive restraint moments should be considered. In these cases the positive restraint moments due to time-dependent effects are typically dominated by the creep of the precast sections. It is recommended that the resulting positive time-dependent restraint moments developed at the piers be computed using the P-method. Research completed during the Mn/DOT study by Smith et al. (2008) and the current study has shown that restraint moments that develop due to thermal gradients applied to the section are significant, and should be considered in either case (i.e., whether or not time-dependent effects generate positive or negative restraint moments). The positive restraint moment effects attributed to the design thermal gradients can be an order of magnitude larger in some climates than the positive restraint moments due to time- dependent effects. The thermal gradients provided by the AASHTO 2010 specification should be taken into consideration by calculating the resulting expected curvatures of each span treated as simply supported. Subsequent derivation of the resulting restraint moments due to the thermal gradient effects is summarized in Section 3.3. The development of positive restraint moments due to thermal gradient effects would also be expected to be significant in continuous monolithic slab span systems, and should be considered during design. There may be little or no economic gain in continuity because of the large thermal restraint moments that develop and in some cases, continuity may require additional reinforcement in the precast sections (i.e., larger than would be required for a simply-supported design). 7.3. Live Load Distribution Factors Numerical modeling was combined with observations from a live load truck test on the Center City Bridge along with load distribution tests on the laboratory bridge specimens (i.e., Concept 1 and Concept 2) to determine the applicability of current live load distribution factors in the AASHTO LRFD (2010) specification for slab-type bridges to the PCSSS. The numerical models illustrated that the longitudinal curvatures measured in the precast slab span system with a reflective crack extending to within 3 in. of the extreme compression fiber and a tandem load greater than that which could be physically applied in the field resulted in longitudinal curvatures which were only 84 percent of the longitudinal curvatures predicted using the AASHTO LRFD (2010) load distribution factors for monolithic concrete slab span bridges, suggesting that PCSSS-type superstructures could reasonably and conservatively be designed using the current live load distribution factors for monolithic slab-type bridges. Furthermore, the live load truck tests on the Center City Bridge suggested that the measured longitudinal curvatures were approximately three times less than those calculated using monolithic slab span equations. In addition, the measured longitudinal curvatures were consistently conservative when compared to

236 monolithic slab span FEM models. The conservatism in the factors for monolithic slab span bridges was sufficient to cover the cases of the PCSSS bridges even considering the potential effects of reflective cracking as discussed above. Load distribution tests on Span 2 of the Concept 1 laboratory bridge and the Concept 2 laboratory bridge included an investigation of the transverse load distribution between adjacent precast panels. For each specimen, loading was applied to the south precast panel, and the longitudinal curvatures in the north and south precast panels were calculated based on measured longitudinal strains located throughout the depth of each section. The longitudinal curvatures were measured at various times throughout the range of testing that occurred on each specimen, and therefore provided an estimate of the curvatures of the sections before cracking was induced near the precast joint, as well as after traffic fatigue loading and loading to simulate cracking along the longitudinal precast joint due to environmental effects. The test was designed such that a reduction in the ability of the PCSSS to transfer load to adjacent precast panels would be indicated by a relative reduction in the measured longitudinal curvature of the unloaded panel. Both Span 2 of the Concept 1 laboratory bridge and the Concept 2, laboratory bridge showed good load transfer capabilities across the longitudinal joint during intermittent tests conducted throughout the investigation of the laboratory bridge specimens to extend the reflective crack. In both cases, little variation in the measured longitudinal curvatures was observed in the unloaded panels, which suggested that load was effectively transferred across the longitudinal joint from the loaded panel despite the presence and increase in the size of reflective cracking induced in/near the joint. In summary, the numerical and experimental studies in regards to live load distribution factors indicated that the PCSSS was well represented by monolithic FEM models, suggesting that the discontinuity at the precast joint did not significantly affect the load distribution characteristics of the system. Also, the performance of the large-scale laboratory bridge specimens reinforced the notion that the system provided sufficient transverse load distribution, with and without the presence of reflective cracking near the joint region. 7.4. Skew Numerical modeling was applied to simply-supported monolithic and jointed (to simulate PCSSS discontinuity at the adjacent precast flange interface) bridge models with skewed supports ranging from 0 to 45 degrees. Three independent load cases were investigated, which included a 35 kip load individually applied over a 12 by 12 in. patch at both quarter points and at midspan for each model. For each load case, the largest horizontal shear stress in the plane above the precast joint nearest the loading was determined. The maximum shear stress from these three load cases defined envelopes for the monolithic and jointed models that varied with skew angle. This maximum horizontal shear stress envelope remained relatively constant through the range of skew angles considered for both jointed and monolithic models. With increasing skew angle, the shear stress envelope increased by approximately 15% for the monolithic models and by less than 10% for the jointed models. The small variation and consistency between the models considering a joint between precast sections with a 3 in. flange and a monolithic structure suggested that the effect of the joint in precast composite slab span construction was not expected to significantly affect the performance of the system in skewed applications, and the design of skewed PCSSS bridges could be completed assuming a monolithic slab span system.

237 7.5. Composite Action and Horizontal Shear Strength To conclude the laboratory tests, the large-scale bridge specimens were loaded to near ultimate levels of load to investigate the ability for the precast slab span sections to remain composite with the CIP concrete topping. Placement of reinforcement for horizontal shear was observed to be difficult and time consuming for the fabricator, especially when finishing the top web surfaces. Furthermore, the reinforcement extending from the precast webs for horizontal shear extended out of the precast section with minimal clearance between the hook and the precast web surface to avoid interference with placement of the deck reinforcement in the field. In initial field applications of the PCSSS, the low clearance of this horizontal shear reinforcement may have limited its effectiveness because aggregate was unable to flow below the returned stirrups. In addition, past research by Naito et al. (2006) suggested that concrete girders loaded to induce positive moments in the section without horizontal shear ties were observed to achieve sufficiently large levels of horizontal shear stress and remain composite. Span 2 of the Concept 2 laboratory bridge was designed with the same horizontal shear layout utilized in the Center City Bridge, which satisfied the design requirements of AAHSTO (2005). Span 1 of the Concept 1 laboratory bridge was designed with fewer horizontal shear ties than were used in Span 2 and in the Center City Bridge, and which did not satisfy the minimum horizontal shear reinforcement requirements of the LRFD specification (AASHTO 2005). The minimum horizontal shear reinforcement requirements have not been modified between the 2005 and 2010 versions of the AASHTO specification. The Concept 2 laboratory bridge was designed and constructed with no horizontal shear ties. In both bridges, the surface condition of the precast member was roughened to a surface consistent with a 1/4 in. rake. In the tests on both spans of the Concept 1 laboratory bridge and on the Concept 2 laboratory bridge, the sections were observed to remain composite well beyond service load levels, through the full range of loading to the maximum capacity of the loading system, which was in excess of the predicted nominal capacity of the Concept 1 and 2 bridges. The longitudinal strains measured during the tests on the Concept 2 laboratory bridge (which had no horizontal shear ties) indicated linear distributions through the cross sections. The Kent and Park model (1971) was used to determine the corresponding compressive stress distribution in the CIP section assuming unconfined concrete models. Integrating the nonlinear stress distribution, resulted in an estimate of the maximum compression force achieved in the slab during loading to the ultimate capacity. The horizontal shear stress estimated in the system at the precast-CIP interface was subsequently calculated by dividing the total compression force by half of the center to center of bearing span length and the total width of the bridge structure, and was determined to be 135 psi. The results of the laboratory tests suggest that the AASHTO LRFD specification should allow for the design of precast slab span structures without horizontal shear ties, and allow for the development of a maximum factored horizontal shear stress of 135 psi in sections with intentionally roughened surfaces (i.e., 1/4 in. rake) unreinforced for horizontal shear. 7.6. Control of Reflective Cracking across Longitudinal Joint between Precast Flanges Control of reflective cracking was investigated through the development and testing of two large-scale laboratory bridge specimens (Concept 1 and Concept 2 laboratory bridges), as well as through a series of subassemblage tests.

238 Reflective cracking was intentionally induced in the Concept 1 and Concept 2 large-scale laboratory specimens to investigate the performance of the PCSSS through a range of loading that was designed to simulate both fatigue performance due to vehicular loading, as well as the influence of environmental effects. Two million cycles of fatigue loading were applied near the longitudinal precast joint with a patch load to simulate tire traffic on both spans of the Concept 1 laboratory bridge, as well as on the Concept 2 simply-supported laboratory bridge. The magnitude of measured transverse strain under loading to simulate traffic was on the order of 30µε, which was much less than the strains measured in the Center City field bridge after cracking was observed. Therefore, in each case, a reflective crack was introduced near the precast joint region by inducing transverse strains above the horizontal precast flange-CIP interface to replicate strain measurements (i.e., approximately 160 µε) observed in a field application of a PCSSS bridge in Center City, Minnesota, that were attributed to reflective cracking resulting from thermal gradient effects. Reflective cracking was induced in each specimen after the completion of one million fatigue cycles and then the laboratory bridge specimens were subjected to an additional one million cycles of simulated traffic loading. This enabled investigation of the fatigue performance of the virgin and cracked systems. In the Concept 2 laboratory bridge, reflective cracking was induced only in the east half span (i.e., associated with straight embedded transverse bars), despite equal loading simultaneously applied at the quarter points of both half spans. During both the first (i.e., uncracked) and second (i.e., after cracking to approximately 160 µε) million cycles of traffic loading, little degradation of the joint was observed in either of the large-scale bridge specimens. The resiliency of the system under simulated traffic load suggested that the PCSSS would provide a durable bridge solution even with the presence of reflective cracking. Because traffic loading was observed to promote little to no degradation of the system, additional tests were conducted to simulate the large increases in transverse strains that had been measured in the Center City Bridge suspected to be the result of daily fluctuations due to thermal gradient effects. To simulate thermal gradient effects in the laboratory, mechanical loading was applied over the precast joint to induce nominal strains on the order of 180µε (target level A) and 300 µε (target level B), which corresponded to approximate magnitudes measured in the Center City Bridge. To simulate the repeated yearly effects of daily fluctuations of the thermal gradients, a total of 15,000 cycles of load were applied to induce each of the target level strains (except in the Concept 2 laboratory bridge, where the loading beam fractured after 5000 cycles of load to target level B). The 15,000 cycles at the target strain levels (A and B) each corresponded to approximately 100 years of environmental effects, conservatively assuming thermal gradients large enough to induce the target strain levels would be expected to occur 150 days a year. Transverse strains measured under a 35 kip patch load, to simulate the adjacent wheel loads of two truck tires, were documented at various times during the environmental effect simulation for both spans of the Concept 1 laboratory bridge, as well as the Concept 2 laboratory bridge to investigate potential degradation of the joint during these tests. For each specimen, measured increases in the transverse strains under the 35 kip patch load were observed as an increasing number of cycles to induce the target-level strains were completed. After the cycles to the target-level strain B (i.e., 300 µε) were completed (i.e., 15,000 cycles for the Concept 1 bridge and 5000 cycles for the Concept 2 bridge), the increases in strains due to the application of the 35 kip patch load were approximately 34, 6, and 23 percent for Spans 1 and 2 of the Concept 1 laboratory bridge, and the east (i.e., straight embedded transverse bars) quarter point of the Concept 2 laboratory bridge, respectively. Degradation of the joint in Span 2 of the Concept 1 laboratory bridge was expected to be most severe because the precast flange was thicker than the remaining specimens causing the transverse reinforcement to be located higher in the section and therefore intercept

239 the reflective cracking at a higher depth, and also because the instrumentation was located higher in the section, and therefore strains measured in Span 2 would have been larger if they had been measured lower in the section, at the same approximate depths as in Span 1 of the Concept 1 laboratory bridge and the Concept 2 laboratory bridge. Also documented during the environmental effect simulation was the rate of degradation of the joint under the loads required to induce strains of target-level strain B (approximately 300 µε). It was observed that the increase in the strain occurred quickly during the first few thousand cycles, and then tended to stabilize. The performance of both spans of the Concept 1 laboratory bridge and the Concept 2 laboratory bridge was observed to adequately control cracking in the precast joint region throughout loading to simulate traffic and environmental effects related to the thermal gradient. Reflective cracking was also monitored throughout the range of testing for seven subassemblage specimens to quantify the relative performance of the respective design details for reflective crack control in each specimen. A primary metric recorded during these tests was the width and stability of the observed cracks throughout the range of applied loads. The ability for each specimen to control the width of cracking was desirable, as large cracks were expected to cause degradation of the longitudinal joint region including providing a potential avenue for the ingress of moisture and chlorides. Each of the subassemblage specimens performed adequately throughout the range of loading, though variations in the extent of cracking indicated some relative differences. The two specimens with the largest reinforcement ratios, SSMBLG5-No.6Bars (ρcr=0.0061, as defined in Figure 5.1.2) and SSMBLG6-Frosch (ρcr=0.0052), performed well relative to the remaining specimens. In these two specimens, measured crack widths were consistently smaller than the remaining specimens. SSMBLG7-Control2 also indicated better than average performance through visual observations; however the analysis of the embedment instrumentation suggested that the behavior of this specimen was similar to the specimens in the group not including SSMBLG5-No.6Bars and SSMBLG6-Frosch. The behavior of SSMBLG7-Control2 was attributed to a relatively smooth precast flange surface achieved prior to the placement of the CIP concrete (which was done in anticipation of studying a debonded flange surface, which was abandoned to allow for a second control specimen to be tested). The relatively smooth flange surface was expected to better distribute transverse stresses across the precast flanges in the joint region, thereby reducing the potential stress concentration at the interface between the adjacent precast flanges which created a longitudinal joint, however it was observed via an analysis of the horizontal crack propagation using the concrete embedment resistive strain gages that a single crack was present internally in the specimen, suggesting that the smooth flange surface did not distribute the transverse stress adequately well so as to promote the development of multiple cracks. A completely debonded surface, however, was not expected to be desirable, as it would likely promote delamination of the horizontal precast flange-CIP interface, which was expected to promote cracking at the vertical precast web, where cage reinforcement was not present to aid in the control of cracking. In the subassemblage tests, the best performing specimens, SSMBLG5-No.6Bars and SSMBLG6-Frosch, had the largest amounts of steel as noted above (ρcr of 0.0061 and 0.0052, respectively). Although SSMBLG6- Frosch was named after the Frosch et al. (2006) recommendations, the reinforcement ratio of the SSMBLG5-No.6Bars specimen more closely correlated with that of the Frosch recommendations for crack control. SSMBLG5-No.6Bars had No. 6 transverse hooks spaced at 18 in., which was supplemented with No. 3 bars spaced at 18 in. in a cage that was staggered relative to the transverse hooks such that the maximum spacing was 9 in. to provide crack control over the longitudinal joint region between the precast flanges.

240 SSMBLG6-Frosch was fabricated with No. 4 transverse hooks spaced at 18 in., supplemented with No. 3 bars spaced at 4.5 in. in a cage that was offset around the transverse bars. The maximum spacing in the SSMBLG6-Frosch specimen was 4.5 in., and narrowed in the region where the cage reinforcement bounded the transverse hooks. The tight spacing of the SSMBLG6-Frosch specimen required careful construction tolerances. In the subassemblage study, the maximum transverse 9 in. spacing for crack control appeared to be sufficient as long as enough reinforcement was provided to ensure that the reinforcement did not yield upon cracking. This was evident through the good performance of both of these specimens. The maximum transverse reinforcement spacing was further investigated by evaluating the performance of the Concept 1 and 2 laboratory bridges which provided more realistic boundary conditions in the longitudinal joint region above the precast flanges. In this study, it was found that the 9 in. maximum transverse reinforcement spacing provided in the Concept 2 laboratory bridge did not correlate with an improvement in the control of cracking near the longitudinal trough area relative to the 12 in. maximum spacing provided in the Concept 1 spans, and therefore an economical design may favor 12 in. transverse reinforcement spacing to 9 in. spacing with no expected reduction in performance. An increase in the maximum transverse reinforcement spacing to 18 in. is not recommended, primarily because cracking in SSMBLG2-NoCage (which was reinforced with only transverse No. 4 bars spaced at 18 in.) was generally largest and crack widths increased with the least increase in the applied load relative to the other subassemblage specimens which had transverse reinforcement spacings on the order of 9 in. (except for the SSMBLG6-Frosch which had the 4.5 in. spacing), which were observed to provide acceptable crack control. Furthermore, little difference was observed between the performance of the sections of the Concept 1 laboratory bridge where reflective cracking was observed, with No. 6 transverse hooked bars, and the performance of the Concept 2 laboratory bridge where reflective cracking was observed, with No. 4 transverse hooks. There was, however, a noticeable increase in the relative performance of SSMBLG5- No.6Bars and SSMBLG1-Control1, in which the only nominal difference was the larger bars in the former specimen. Because the increased performance observed in SSMBLG5-No.6Bars, which was on the order of SSMBLG6-Frosch, was achieved with larger bars and a maximum transverse reinforcement spacing of 9 in., it was suggested that a design with No. 6 bars and less cage reinforcement was likely to be more economical and easier to implement in the field than the closely spaced reinforcement cage provided in SSMBLG6-Frosch. 7.7. PCSSS Design Recommendations The research completed during the NCHRP 10-71 study resulted in the development of a comprehensive design guide for the design and construction of precast composite slab span system bridges. The design guide in reference to PCSSS was developed based on previous work by the researchers (Smith et al. 2008, Eriksson 2008) summarized in Chapter 3; numerical studies to investigate the performance of the PCSSS details, as well as considering the system as a whole as described in Chapter 4; and an extensive large-scale laboratory research program summarized in Chapters 5-6. The design guide includes proposed modifications to the AASHTO LRFD 2010 Bridge Design Specification, as well as the AASHTO LRFD Bridge Construction Specifications. The design guide developed during the NCHRP 10-71 study is included in Appendix A, and corresponding design examples of PCSSS bridges are included in Appendix B.