Seismic behaviour of gravity quay walls built on liquefiable soils

Abstract

In the last 50 years there have been many documented failures of gravity retaining quay walls due to earthquake events. These failures are often associated with significant deformation of liquefiable soil deposits (e. g. major damage occurred at Kobe Port during the 1995 Hyogoken-Nanbu earthquake). Saudi Arabia has similar types of quay walls located in regions that have the potential to experience significant seismic events. These walls have not been designed for seismic resistance and therefore have the potential to suffer serious damage from seismic activity. For many years the design of seismic gravity quay walls has been studied and design codes for engineering practice established; however, the widespread failures of these structures during recent earthquakes demonstrates that these design methods may be insufficient. Such gravity quay wall failures have stimulated progress in the development of a performance-based seismic design method using non-linear inelastic dynamic analysis for quay wall structures. The aim of this study was to develop a methodology for the seismic design of gravity quay walls using a non-linear elasto-plastic dynamic analysis. The final method adopted in this work is based on the generalised elasto-plasiticity constitutive model developed by Pastor et aL (1990), with some minor modifications, which has been incorporated into a finite element procedure. The proposed P-Z sand model was first validated by simulating published monotonic and cyclic test results. Secondly, an effective stress analysis was established by developing a finite element model for Kobe Port Island quay walls using the P-Z sand model. This model was validated by comparing the predicted deformations with those experienced at Kobe. The computed residual deformations from the analysis were in good agreement with published field observations. To develop mitigation strategies, a parametric study of the seismic perfonnance of gravity quay walls, using the effective stress analysis, was conducted. This study assessed the effect of various structural and geotechnical parameters on the seismic performance of quay walls. Twenty-six cases of effective stress analysis with variation in tidal range, soil permeability, soil relative densities, and wall widths were conducted as well as analyses to test the importance of considering multi-directional seismic excitations as opposed to uni-directional. In order to assess the safety of existing quay walls in Jeddah Port, an experimental programme was conducted. This programme consisted of a site investigation (using a standard penetration test (SPT)) to determine the in situ relative density of the existing backfill, a series of laboratory based, monotonic and cyclic triaxial tests (to define the soil properties of Jeddah Port sand) and a two-dimensional effective stress finite element analysis of a typical Jeddah quay wall. The monotonic and cyclic triaxial tests were conducted for three different relative densities of D, = 35 %, 55% and 75%. These represent loose (equal to in situ conditions as established by the site investigation), medium dense and dense sand. The experimental results are discussed and then used to identify the P-Z sand model parameters. These parameters were used in conjunction with a finite element analysis of Jeddah Port quay walls to Predict the seismic deformations. In this analysis the finite element model was subjected to a number of different ground motions, which represented two different levels of earthquake intensity; namely moderate and strong ground shaking. The effect of improvement strategies such as increasing the relative density of the backfill and foundation materials was then assessed The results of the simulations showed that existing Jeddah Port quay walls are not satisfactory to resist either moderate or strong earthquake excitations. However, if the relative density is increased to 55% then satisfactory performance can be achieved for a moderate intensity earthquake. For the case of strong shaking, the analysis showed that the quay walls did not demonstrate the required performance levels; however, they were only under specification by 10%. Finally, a flowchart illustrating a seismic design procedure for gravity quay walls has been proposed, which is applicable to both existing and new gravity quay walls. Key-words: quay wall, liquefaction, earthquake, port, effective stress, constitutive model.