The Applicability of Sliding Block Analyses for the Prediction of Lateral Spreading Displacements
Author
Makdisi, Andrew James
Metadata
Show full item recordAbstract
Liquefaction-induced lateral spreading deformations can pose significant risks to bridge pile foundations, underground utilities, and shallow foundation systems, particularly at sites in seismically active areas near embankments and levees, in close proximity to natural or man-made free faces. Current approaches for lateral spreading predictions, consisting primarily of empirical models and strain potential methods, largely neglect complex and critical factors that influence such deformations, such as nonlinear dynamic site response, dilation pulses, void redistribution, and water film formation, as well as complex two- or three-dimensional topographic effects. It has also been proposed by several researchers that lateral spreading predictions can be potentially approximated using Newmark-type rigid block analyses, with an appropriate reduction in the yield acceleration to account for soil liquefaction. The appropriateness of the Newmark-type framework in lateral spread characterization was assessed in this study. Particular focus was placed on characterizing any potential structural biases inherent to the sliding block framework, including those related to the assumption of strain concentration along a discrete failure plane and rigid dynamic behavior, as well as practical issues related to the uncertainty in sliding block displacement predictions, and how the sliding block method compares to other approaches currently in use for predicting lateral spreading displacements. When evaluating solely the effects of the assumption of a discrete failure surface, sliding block analyses predicted displacements on the order of 30 to 70% lower than displacements predicted using simple dynamic analyses of a soil continuum. Further analyses showed that, when compared to both finite-element methods used for liquefaction modeling and established lateral spreading empirical models, sliding block methods largely failed to predict substantial deformations over a broad range of lateral spreading conditions, and have the potential to be highly un-conservative in characterizing lateral spreading hazards in general.
Collections
- Civil engineering [413]