For seismically active areas, designers must consider how bridge foundations will respond to ground liquefaction, particularly lateral spreading of the soil. Current design procedures generally rely on simplified analytical methods based on a two-dimensional description of the site geometry. However, numerous bridges affected by lateral spreading during past earthquakes have displayed three-dimensional soil deformation that cannot be captured in a two-dimensional analysis, and an analysis that assumes 2-D conditions may be overly conservative, leading to unnecessarily expensive design solutions. Therefore, a simplified analysis procedure for estimating lateral spreading forces that provides guidance on when 2-D or 3-D assumptions are most applicable would be a valuable tool for bridge design.
The objectives of this research were to identify and quantify the mechanisms that may result in potential reductions in the bridge foundation demands that develop during lateral spreading by considering the 3-D geometry of the bridge site. The work involved a review of relevant case histories and the development of numerical models that capture the kinematic loading conditions and 3-D effects of the problem.
Several Chilean bridge sites affected by lateral spreading during the 2010 offshore Maule earthquake were examined for potential use as case studies, and the Puente Mataquito and Llacolén bridges were selected. At the Puente Mataquito bridge, the earthquake caused widespread lateral spreading in the surrounding soil; however, only insignificant structural damage was observed in the bridge itself. The discrepancy between the amount of soil deformation and structural damage suggested that design procedures for this load case had not adequately considered the 3-D soil deformation mechanisms and led to overly conservative solutions. In contrast, observed lateral spreading and damage near the Llacolén bridge were more consistent with the amount of soil deformation. The Llacolén bridge approaches showed fewer 3-D effects on both sides of the bridge and therefore larger loads on the structural components, resulting in the collapse of one of the approach sections.
Three types of numerical analyses with varying levels of complexity were applied to the two chosen case study bridges. These models included beam on nonlinear Winkler foundation models, dynamic effective stress models of the bridge-foundation-soil system in plane strain, and 3-D models of the bridge abutments, approach embankments, and surrounding soils.
The results of the numerical modeling for the Mataquito and Llacolén bridges, along with a preliminary parameter study conducted with an independent set of 3-D finite element models, indicated that by considering the 3-D geometry of the bridge site and structure, the resulting estimates for foundation bending demands and abutment displacements may be substantially smaller than the demands and displacements resulting from a plane strain description of the problem or a simplified analysis based on 1-D models. This difference increases with decreases in the depth of the liquefiable layer and the effective width of the approach embankment. The researchers modified an existing simplified analysis procedure to better reflect the findings of this work, and they recommend it for the design of bridge foundations subjected to lateral spreading.
The researchers also developed a preliminary procedure for assessing the expected amount of lateral pinning resistance for a given combination of bridge foundation, soil profile, and approach embankment. This procedure is based on results from a preliminary parametric study and will require further development and validation before it can be used in practice.
WA-RD 874.1(Caution: 20MB file)
Authors:
Pedro Arduino
Christopher R. McGann
Alborz Ghofrani
UW Department of Civil and Environmental Engineering
Sponsor: WSDOT
WSDOT Technical Monitor: Tony Allen
WSDOT Project Manager: Lu Saechao