Adjacent Box Beam Connections: Performance and Optimization

CHAPTER 6. SUMMARY AND CONCLUSIONS

Full-scale testing on adjacent box beam connection methods was conducted as part of this study. Four shear key connection designs were evaluated. The first two designs utilized both partial- and full-depth conventional grouted connections with various levels of transverse PT. The other two designs utilized partial- and full-depth UHPC connections with non-contact lap-spliced reinforcement. Transverse PT was not used in the UHPC connections. The beams were subjected to 10 cycles of thermal loading and millions of cycles of structural loading.

SUMMARY

The main findings are summarized as follows:

• The thermal loading generated in the study produced a temperature gradient between the top and bottom flanges of approximately 50 °F (28 °C) and resulted in an upward deflection of approximately 0.47 inch (11 mm). The applied thermal loading cycles did not initiate any significant cracks in the connections.
  
  
• The cyclic structural loading applied in this study was severe. The most extreme case in this study utilized a maximum loading range of 90 kip (400 kN) with a 5-kip (22-kN) minimum load. Within the most restrained test setup, this created an M_equivalent of 498 kip-ft (673 kN-m) transferred through the connection
  
  
• When a connection was uncracked, cyclic structural loading was not seen to initiate cracking. This was true regardless of level of PT in conventionally grouted connections.
  
  
• The calculated shear forces transferred through the connection were small. Maximum shear stress in the partial-depth beams was calculated to be 23 psi (161 kPa).
  
  
• When there were preexisting cracks in conventionally grouted connections, cyclic structural loading was observed to propagate the cracks independent of the level of transverse PT force applied. Cracks propagated more quickly under lower levels of PT.
  
  
• With higher levels of transverse PT force, the cracked connection could still effectively transfer the load, though Δ_δ increased slightly. When the transverse PT force was removed, the cracked connection could quickly lose its ability to limit Δ_δ. This could reduce its capability to effectively transfer the applied loads between adjacent beams.
  
  
• If the transverse PT force was applied before casting the grout, the loss of the PT force after casting may cause transverse tensile forces to develop in the connection. This could lead to cracking if the beams exhibited a large enough amount of relative sweep in their as-fabricated shape.
  
  
• When the connection was uncracked, beams with conventional grout connections had similar load distribution performance as beams with UHPC connections. However, the interface between the conventional grout and box beam concrete was the weak link of the system and could crack if a sufficient load or deformation occurs.
  
  
• The behavior of the adjacent box beam bridges with UHPC connections could be expected to be comparable with an equivalent structural system with no field-cast connections. The mechanical capacity of the UHPC connection was observed to enhance connection capacity so that, under the application of large transverse tensile stresses, tensile rupture occurred in the precast concrete box beams.
  
  
• Full-depth connections showed slight improvements in load distribution between beams. This is likely due to the increased depth of the connection, which significantly increased the transverse flexural and shear stiffnesses of the connection. However, increasing the depth of the connection increases construction costs and possibly construction complexity.
  
  
• A partial-depth UHPC connection appears to be sufficient to achieve the performance requirements.

CONCLUSIONS AND RECOMMENDATIONS

Conclusions and recommendations related to adjacent box beam connection design and performance evaluation include the following:

• The performance and efficiency of the shear key can be evaluated for load transfer by determining the proportion of moment carried by the loaded beam.
  
  
• Δ_δ between adjacent beams can be a good indicator of the serviceability performance of a connection. Based on the tests in this study, Δ_δ for in-tact connections were below 0.005 inch (0.127 mm), while the Precast Prestressed Concrete Bridge Design Manual seeks to limit Δ_δ between adjacent box beams to 0.02 inch (0.51 mm) for spans up to 100 ft (30.5 m).⁽³⁾
  
  
• M_equivalent, which calculates the moment transferred through the connection from a loaded beam to adjacent beams, can be used to compare the test results from this study with other bridge designs that have different geometries and loading conditions.
  
  
• Transverse PT can limit differential movement between beams, compensate for some transverse tensile strains across the connections, and assist with load transfer between beams after connection cracking. Increased transverse PT force distribution along the length of the connections could enhance system performance as the keyway shear strength increases with more confinement force. However, as commonly deployed today, transverse PT only effectively confines a small area near the PT locations. This transverse PT would likely be most valuable after connection degradation has already begun, thus serving to limit large Δ_δ between adjacent beams.
  
  
• Based on the concurrent research by De la Varga et al., a minimum interface bond strength of 150 psi (1.0 MPa) is recommended when selecting a grout material.⁽¹⁷⁾ This helps to avoid interface cracking due to eccentrically placed external loads and may assist with the mitigation of transverse tensile forces and deformations due to thermal loads and material shrinkage.