Seismic Response of Steel Tube Concrete Arch Bridge Considering Pile-Soil Dynamic Interaction

Overview of the Study

This paper by Zhang Bo and Zai Jinmin from Nanjing Tech University addresses a critical gap in bridge seismic engineering: the dynamic interaction between pile foundations and surrounding soil in steel tube concrete (CFT) arch bridges. The study focuses on an 85-meter main-span CFT arch bridge supported by pile foundations with a 60-meter overburden soil layer. A three-dimensional finite element model was established, employing the Drucker-Prager yield criterion for the soil's elastoplastic behavior and a contact pair method to simulate the interaction between the pile lateral surface and the surrounding soil. The results reveal that dynamic soil-structure interaction (SSI) has a profound influence on the transverse seismic response of the bridge, a finding that carries significant implications for the design of long-span arch bridges in seismic zones.

Core Technical Content and Modeling Approach

The modeling approach deserves careful attention from practicing engineers. The Drucker-Prager yield criterion was selected for the soil because it provides a computationally efficient approximation of the Mohr-Coulomb criterion while maintaining smoothness in the yield surface, which facilitates numerical convergence in nonlinear finite element analysis. The contact pair method used to model pile-soil lateral interaction is essentially a penalty-based or Lagrange multiplier approach that captures the non-linear frictional and normal contact behavior at the soil-pile interface. This is particularly important because the soil-pile interface undergoes significant slip and gap formation during seismic loading, and ignoring this behavior leads to non-conservative estimates of lateral displacement and shear demand.

The comparison between the SSI model and the fixed-base model highlights a critical engineering insight: the transverse (cross-bridge) seismic response is significantly affected by SSI effects. This is counter-intuitive for many engineers who might expect the longitudinal direction to be more sensitive due to the arch's primary load path. However, in the transverse direction, the bridge deck and arch ribs rely more heavily on lateral restraint provided by the pile foundations, and the soil-structure interaction directly modifies this restraint mechanism. The longitudinal response is more governed by the arch's inherent stiffness and the longitudinal restraint at the springings, which are less sensitive to soil flexibility.

Engineering Implications and Practice Recommendations

From a design practice perspective, this study reinforces the necessity of incorporating SSI effects in the seismic design of arch bridges, particularly those with deep pile foundations in soft or medium-dense soils. Traditional approaches that model the pile top as a fixed support or use simplified subgrade reaction methods may underestimate the transverse displacement demand and the associated secondary stresses in the arch ribs and deck. The following practical recommendations emerge from this study:

1. For CFT arch bridges with spans exceeding 80 meters in seismic intensity 7 or higher zones, SSI should be explicitly modeled in the seismic analysis rather than being accounted for through simplified reduction factors.
2. The transverse direction warrants particular attention in SSI analysis, as the interaction effects are more pronounced in this direction than in the longitudinal direction.
3. The pile-soil contact interface should be modeled with appropriate friction coefficients derived from site-specific geotechnical investigations, as the interface behavior is highly sensitive to soil type, density, and moisture content.
4. When the overburden soil depth exceeds 50 meters, as in this case, the amplification and filtering effects of the soil column on seismic waves become significant, and the input motion at the pile top may differ substantially from the free-field surface motion.

Key Questions and Critical Reflections

Several questions arise from this study that deserve further investigation. First, the Drucker-Prager model, while computationally efficient, does not capture the pressure-dependent dilatancy behavior of soils accurately. For soils with a friction angle exceeding approximately 30 degrees, the Drucker-Prager approximation deviates notably from the Mohr-Coulomb criterion. Would the use of a more sophisticated soil model, such as the Mohr-Coulomb with associated flow or a cap model, yield different conclusions regarding the severity of SSI effects? Second, the study does not appear to address the potential for pile-soil separation during strong seismic shaking, which is a well-documented phenomenon in liquefiable soils. If the 60-meter soil column contains any liquefiable layers, the contact pair method must be calibrated to allow for complete separation, not just frictional sliding. Third, the study focuses on a single bridge configuration; sensitivity analyses varying the span, soil profile, and pile geometry would strengthen the generalizability of the conclusions.

Study Insights and Implications for Practice

The most valuable contribution of this paper is the demonstration that SSI is not merely a second-order effect that can be handled through empirical adjustment factors. In the transverse direction of a CFT arch bridge with deep pile foundations, the dynamic interaction fundamentally alters the response spectrum characteristics, the displacement distribution along the arch, and the stress state in the CFT members. This finding has direct implications for the seismic design provisions in codes such as GB 50011 and AASHTO LRFD, which currently rely on simplified SSI models for most bridge types. For future projects involving long-span CFT arch bridges in seismic regions, the practice should include a dedicated SSI analysis phase in the design workflow, with particular emphasis on the transverse response, and the results should be compared against the fixed-base analysis to quantify the interaction effect magnitude before finalizing the design.