Abstract
This report, Development of Preliminary Strut-Tie Models for Precast Bent Cap Connections (Cast-in-place and Grouted Duct), develops strut-tie models (STM) for precast bent cap-column connections using the NCHRP 12-74 Grouted Duct (GD) and Cast-in-Place (CIP) specimens. In development of these STM’s, the bent cap bar strains from the specimens were compared against three theoretical models: beam theory, 2D strut-tie models, and 3D strut-tie models. The beam theory analysis used statics, moment-curvature analysis, and actual material properties of the specimens to determine the theoretical bent cap bar strains. The 2D STM, based on the modified external strut force transfer model (EFTM) proposed in the literature, was established using the computer aided strut-and-tie (CAST) program. Through an iterative process, a refined 2D STM was developed for both the push and pull test directions by comparing specimen strain data to CAST output. An important modification to the EFTM-based STM was the addition of a tension tie at the bottom cap face (as tested) for the pull direction. This corresponded to tension strain present in test data and made the CAST model stable. Based on the 2D STM, the 3D STM was created in the SAP 2000 structural analysis program. The 3D STM incorporates out-of-plane effects related to actual column bar positions and allows a more accurate representation of the two primary mechanisms assumed in anchoring column tension forces: clamping mechanism and splice transfer mechanism. The beam theory results included limited comparisons of actual-to-theoretical flexural strains for two locations adjacent to the joint, top vs. bottom bars, and CIP vs. GD specimens. The average percent differences of the actual to the theoretical strains for the CIP was 49 and 146 for the compression bar, 12 inches away from the cap face and at the cap face, respectively. Over the entire range of loading stages, differences in actual-to-theoretical strains were generally larger for locations closer to the joint, indicating a more pronounced local disturbance compared to locations further away from the joint. Bars that were in compression for most of the loading sequences exhibited a much closer match to theoretical strains than bars that were primarily subjected to tension. Local cracking and other effects are believed to have influenced gage readings. CIP and GD strains for the same locations generally displayed similar trends and values, especially for bars in compression. Compared to beam theory, results of the 2D STM analysis indicated a closer correlation between actual and theoretical strain. The difference between actual and theoretical strains for the CIP specimen were limited to 28 percent and averaged approximately 16 percent for both push and pull directions. For the GD specimen, the differences were as large as 44 percent except for one location, which reached a 98 percent difference. On average, the differences averaged 27 percent in the push direction and the 45 percent in the pull direction. This increased accuracy reflects the more realistic representation of the flow of forces within a joint and their effects. The 3D STM showed the closest correlation between the test data and theoretical analysis. Actual to theoretical strains for the 3D STM’s differed by no more than 42 percent and only 14 percent on average. These values were smaller than for any other analytical method. The reason is because the 3D mechanisms associated with anchoring the column tension force were more accurately detailed and accounted for in the 3D model. Conclusions from these analyses include: 1) beam theory does not accurately represent strains that develop in longitudinal reinforcement at the face of CIP and precast bent cap joints; 2) the limit of the disturbed (D) region appears to extend a distance of approximately half of the bent cap depth (hb/2) from the face of the joint; 3) the developed 2D STM, including the additional tension tie, provides a reasonably simple and accurate model for the flow of forces through a bent cap joint using CIP or GD connections; 4) the modified EFTM requires an additional tension tie in the pull direction to accurately represent tension that develops in the exterior face of bent caps; 5) the developed 3D STM is the most complicated yet accurate model for analyzing joint forces and associated reinforcement strain in a CIP or precast bent cap joint; 6) the presence of ducts in the GD specimen did not noticeably affect specimen strain values compared to the CIP specimen nor affect the development or results of the 2D or 3D STM’s; and 7) analytical results from this study do not indicate the need for any changes to existing NCHRP 12-74 recommendations for non-integral precast bent caps using CIP or GD connections. Based on results of this analysis, the following are recommended for future study: 1) perform finite element analysis (FEA) of the CIP and GD specimens in the pull direction to confirm the need for the additional tension tie; 2) further develop the 2D STM, with special focus on determination of strains in the joint hoops and joint stirrups (interior and exterior) and comparison to test data; 3) perform FEA for CIP and GD precast bent caps to validate the three splice transfer mechanisms and their impact on joint behavior; and 4) incorporate data from the CSUS Preliminary Grouted Duct specimen to supplement these analytical results.