Numerical modeling of a bridge system & its application for performance-based earthquake engineering
Many bridges have been damaged during earthquakes. The severity of the recorded damage varies from column failures and bridge deck collapse due to joint unseating to pile and abutment foundation failure due to lateral spreading. In most of these cases, soil pile-structure interaction (SPSI) has played an important role. SPSI is affected by soil conditions (e.g., competent soil or liquefiable soil) and loading conditions (e.g., earthquake intensity, motion characteristics, and shaking orientation). In this context, two bridges with different soil and loading conditions were considered focusing on different aspects of their response. One was a two-span portion of a bridge supported by drilled-shaft foundations embedded in competent soil (dry dense sand) and subjected to shaking in the transverse direction. The other represented a typical five-span highway bridge system supported by pile group foundations embedded in liquefiable soils and subjected to shaking in the longitudinal direction.In the first part of this research, centrifuge experimental tests and OpenSees numerical simulations were performed to understand and validate numerical modeling strategies to capture soil-pile-structure interaction of bridge structures. The study included shaking in several directions, pile embedment lengths, and various levels of nonlinearity. Sensitivity studies were conducted for a bent and a single pile to study the effect of soil motion and p-y spring parameter variation on the superstructure and pile response. For a prototype bridge model, the soil and p-y spring models verified in the centrifuge simulation study were coupled with a reinforced concrete bridge model (Ranf 2007, Johnson et al. 2006 and Ramirez et al. 2007). Using the prototype reinforce concrete bridge model an equivalent-cantilever approach was investigated to understand appropriate fixity depths and to provide a reasonable dynamic SPSI approximation for motions with different intensity.In the second part of this research, the PEER PBEE methodology was applied to a typical highway bridge system founded in liquefaction-susceptible soils. A comprehensive numerical model was developed for the bridge system by coupling a well-defined OpenSees bridge model (Mackie and Stojadinovic 2003) with various types of SPSI modeling components, including: pile, pile cap, and abutment structures. Using the comprehensive bridge model, the global bridge behavior and the effect of lateral spreading on the bridge were investigated. In this study, various sources of uncertainties were considered, including: earthquake motion uncertainty, model parameter uncertainty, and soil spatial variability uncertainty. Motions corresponding to four different hazards are used to evaluate record-to-record uncertainties. From this study, structural and geotechnical Engineering Demand Parameters (EDPs) related to bridge damage were identified and relative efficiencies for several Intensity Measures (IMs) were investigated. Parameters that influence the bridge performance were investigated through sensitivity analysis and Tornado diagrams. Then, First-Order Second-Moment (FOSM) analyses were performed to estimate parametric uncertainties. Gaussian stochastic random fields for clay and liquefiable soils were generated to investigate spatial variability uncertainties. Using the combined EDP uncertainties with IM hazard curves and the PEER PBEE framework, EDP hazard curves were developed for several EDPs.
- Civil engineering