Behaviour of piles in liquefiable deposits during strong earthquakes

Soil liquefaction has caused major damage to pile foundations in many previous earthquakes. Pile foundations are relatively vulnerable to lateral loads such as those from earthquake shaking; during liquefaction this vulnerability is particularly pronounced due to a loss of strength and stiffness in the liquefied soil. In this research seismic assessment methods for piles in liquefied soil are studied; a simplified approach and a detailed dynamic analysis are applied to a case study of a bridge founded on pile foundations in liquefiable soils. The likely effects of liquefaction, lateral spreading and soil-structure interaction on the bridge during a predicted future earthquake are examined. In the simplified approach, a pseudo-static beam-spring method is used; this analysis can be performed using common site investigation data such as SPT blow count, yet it captures the basic mechanism of pile behaviour. However, the phenomenon of soil liquefaction is complex and predictions of the seismic response are subject to a high level of aleatoric uncertainty. Therefore in the simplified analysis the key input parameters are varied parametrically to identify key features of the response. The effects of varying key parameters are evaluated and summarised to provide guidance to designers on the choice of these parameters. The advanced analysis was based on the effective stress principle and used an advanced constitutive model for soil based on a state concept interpretation of sand behaviour. The analysis results give detailed information on the free field ground response, soil-structure interaction and pile performance. The modelling technique is described in detail to provide guidance on the practical application of the effective stress methodology and to illustrate its advantages and disadvantages when compared to simplified analysis. Finally, a two-layer finite element modelling technique was developed to overcome the limitations conventional two-dimensional (2-D) models have when modelling three-dimensional (3-D) effects. The technique, where two 2-D finite element meshes are overlapped and linked by appropriate boundary conditions, was successful in modelling 3-D characteristics of both deep-soil-mixing walls for liquefaction remediation and pile groups in laterally spreading soil. In both cases the new two-layer model was able to model features of the response that conventional one-layer models cannot; for cases where such aspects are important to the overall response of the foundation, this method is an alternative to the exhaustive demands of full 3-D analysis