This report summarizes the findings and recommendations on the impact of foundation settlements on the reliability of bridge superstructures. As a collaborative effort of an overall initiative for the development of LRFD foundation design specifications, this study is focused on the investigatio of pros and cons for including foundation settlements in bridge designs under gravity loads. Settlement was modeled both probabilistically and deterministically. In the case of a random settlement variable, a lognormal distribution was used in reliability analysis with a fixed coefficient of variation of 0.25. Dead and live loads were modeled as random variables with normal and Gumbel Type I distributions, respectively. Considering the regional traffic condition on Missouri roadways, the effect of a live load reduction factor on bridge reliability was also investigated. Therefore, a total of eight cases were discussed with a complete combination of settlement modeling (mean and extreme values), design consideration (settlements included and excluded), and live load reduction (unreduced and reduced live loads). Based on extensive simulations on multi-span bridges, bridges designed without due consideration on settlements can tolerate an extreme settlement of L/3500 - L/450 under unreduced live loads and up to L/3500 under reduced live loads without resulting in a reliability index below 3.5 (L=span length). Depending upon span lengths and their ratio, the reliability of existing steel-girder bridges is consistently higher than prestressed concrete and solid slab bridges. The shorter and stiffer the spans, the more significant the settlement's effect on the reliability of bridge superstructures. As the span length ratio becomes less than 0.75, the girder and solid slab bridges' reliability drops significantly at small settlements. A concrete diaphragm is very susceptible to the differential settlement of bridges, particularly for moment effects. Two recommended were made to address settlement effects in bridge design: (1) settlement is considered in structural design and no special requirement is needed for foundation designs unless settlement exceeds the AASHTO recommended settlement limit of L/250, and (2) settlement is not considered in structural design as in the current MoDOT practice but ensured below the tolerable settlement (e.g. L/450 for steel girders, L/2500 for slabs, and L/3500 for prestressed concrete girders). The first method provides a direct approach to deal with settlements and has potential to reduce overall costs in bridge design. The second method may result in oversized foundations.
Z. Wang et al., "Calibration of Load and Resistance Factors in LRFD Foundation Design Specifications," Center for Transportation Infrastructure and Safety/NUTC program, Missouri University of Science and Technology, Feb 2011.
Civil, Architectural and Environmental Engineering
United States. Department of Transportation
Research and Innovative Technology Administration (RITA)
Keywords and Phrases
Reliability index; random variable; probabilistic distribution; tolerable settlement; girder bridge; prestressed concrete bridge; solidslab bridge; concrete diaphragm
© 2011 Center for Transportation Infrastructure and Safety/NUTC program, Missouri University of Science and Technology, All rights reserved.
01 Feb 2011