Sarah K. Bowlin
4 min
This dissertation investigates the shear behavior of post-tensioned bridge girders, specifically focusing on the impact of internal bonded and unbonded tendons located within the web. As bridge design evolves to include flexible fillers for easier inspection and strand replacement, understanding the shear performance of these unbonded systems has become critical. The research aims to fill the gap in current design guidance by evaluating how different bond types, duct geometries, and reinforcement ratios influence shear strength and failure mechanisms.
To address these questions, the study conducted two comprehensive test series at Purdue University. Test Series 1 involved 25 moderate-scale girder specimens (I-shaped and box-shaped) to isolate variables such as tendon bond type, transverse reinforcement ratios, and duct diameter-to-web width ratios. Test Series 2 validated these findings with four full-scale girder tests representative of field conditions. The specimens were subjected to monotonic loading until failure, with extensive instrumentation—including strain gauges on stirrups and prestressing strands—to monitor internal force distribution and crack development.
The experimental results reveal that the bond type of web tendons has a negligible impact on overall shear capacity. While bonded tendons provide a slight increase in strength, the failure mechanism for both bonded and unbonded cases is consistently characterized by localized concrete crushing near the duct level. The study found that increasing the duct diameter-to-web width ratio decreases shear capacity, while increasing transverse reinforcement effectively enhances it. Based on these observations, the author proposes refinements to the AASHTO LRFD Bridge Design Specifications. These include using the full web width for concrete contribution calculations and removing the reduction factor for transverse reinforcement efficiency, providing a more consistent and accurate design approach for both grouted and ungrouted systems.
Sam: And the fix Bowlin proposes is to apply a consistent reduction to both—subtracting the duct diameter directly from the effective web width, regardless of bond condition.
Alex: That's the unified model. The duct diameter is treated as a direct deduction from the web width when calculating nominal shear resistance. The mechanism is the same for both bond conditions, so the design equation should reflect that. And when you validate this against the broader literature, the correlation holds well across configurations. [[RP_SECTION:upper-limit-on-strength|Upper limit on strength]]
Sam: You mentioned an upper limit on shear strength. What's the role of that?
Alex: Without an upper bound, the model can overestimate capacity in heavily reinforced sections—cases where you have dense stirrups but the concrete compressive strength is the actual bottleneck. Bowlin evaluated several candidate limits, and the one tied to concrete compressive strength is the most practical. It's consistent with existing AASHTO logic, and it catches the unconservative outliers without disrupting the rest of the design space.
Sam: So the full proposal is a unified web width reduction based on duct geometry, plus a compressive-strength-based upper limit on nominal shear resistance. That's a fairly targeted set of changes to push into a code revision cycle.
Alex: And that's the point. These aren't sweeping changes to the code framework—they're specific, data-driven corrections to provisions that were built on an incomplete experimental base. The segmental box girder market has been growing, and engineers choosing between grouted and unbonded systems deserve design equations that actually reflect the physics.
Sam: It's a good example of how a single well-designed experimental program can resolve a question that's been sitting in the literature as an assumption. The mechanism was always plausible—now there's direct evidence behind it.
Alex: And the practical consequence is real. If you're designing a segmental bridge and your shear capacity estimate assumes full web width for bonded tendons, you may be operating with less margin than you think. Bowlin's framework gives engineers a more defensible basis—and gives code committees a dataset to act on.
Sam: Thanks for walking through the mechanics, Alex. And thanks for listening to ResearchPod.