Modelling of dynamic pile behavior during an earthquake using PLAXIS 2D

Abstract

In the last few years the research activities on the effect of human induced earthquakes on (piled) foundations has grown enormously. Apart from the complex numerical models for masonry and the effect of earthquake loading on these structures, a lot of research is currently performed on the soil properties and its effect on the wave propagation. However, the influence of these human induced earthquakes on pile foundations and an easy and time efficient FEM calculation method is not yet available or well described in Dutch literature. There are several ways of approximating the soil-pile behaviour during an earthquake. The main one, which is also applied in this master thesis, is by performing 2D FEM analyses on the soil and subsequently the pile behaviour. In this aspect a study on the possibilities and limitations of the embedded beam (row) within PLAXIS 2D will be performed. Static pushover analysis, free field site response analysis of the Groningen situation and the kinematic loading of the embedded pile will be addressed. The embedded beam (row) within PLAXIS 2D is in fact a 2.5D situation, where the pile (or beam) is connected to the soil elements of the FEM by special interfaces. These interfaces (springs) are defined by the interface stiffness factor, which are determined and validated in previous master thesis projects for mostly axial loading. They depend mainly on the pile to pile distance specified for the embedded beam. In the latest version of PLAXIS (2015) a limiting lateral soil resistance can be defined. With this option, plastic behaviour is now incorporated into the 2D pile behaviour. This should give a better approximation of the pile response when compared to 3D calculations or measurements. Based on soil investigation from Groningen, a silty clay layer was defined and used as a one layered soil profile of 20 m depth. During an earthquake, the soil behaviour is assumed to be undrained. In order to define the right effective strength parameters for this model, based on the undrained shear strength, a small investigation was performed. The results from the static pushover analyses were compared with a 3D volume pile in PLAXIS and D-Pile Group calculations. From this comparison the best way of specifying the limiting lateral soil resistance for this thesis was determined. It also became evident that the default lateral interface stiffness factor should be improved, either by making it stress or strain dependent. Based on KNMI data, which is the Royal Netherlands Meteorological institute, a deconvolution and scaling of the Huizinge earthquake signal was performed based on reports from amongst other Deltares. This signal was then used for the free field site response analysis in both EERA and PLAXIS. After the implementation of the earthquake loading in the PLAXIS model, the behaviour of the embedded beam was evaluated and compared to analytical design methods for kinematic bending moments and pseudo-static calculations with both D-Sheet Piling and PLAXIS 2D. Also the cyclic loading behaviour of the embedded beam in comparison with the 3D volume pile was elaborated. The main conclusion of this thesis is that the embedded beam (row) in 2D does show capabilities for modelling (dynamic) lateral loaded pile behaviour. However, the option of the limiting lateral soil resistance should be optimized in combination with an alternative way of defining the interface stiffness factors. The plastic behaviour of the pile-soil system should be improved of the embedded beam(row) in order to show similar damping behaviour as was obtained in 3D.