Abstract

Soil supporting foundations for infrastructural assets such as bridges are repeatedly subjected to dynamic loading from passing traffic. The load-displacement characteristics are nonlinear at even small to medium strains, however this effect is mostly modelled using secant stiffness models with effective linear stiffness. This simplification enables a pseudo linear model to encapsulate the effects of a nonlinear system with some accuracy. Linear models have a frequency related to the amplitude of the load used to specify the secant stiffness. Nonlinear models, however, have a range of frequencies due to the changing stiffness of the system resulting from the amplitude of the input load. There is growing uncertainty surrounding the response of foundation systems to damage effects. In particular scour erosion, which is the term used to describe removal of soil from around foundation elements by hydraulic action, presents a significant hazard to infrastructure and is becoming more prevalent with increased flooding risk due to climate change effects. This paper describes a numerical framework to incorporate a basis for nonlinearities in the load-displacement response modelling of pile foundations. The Newmark-Beta nonlinear integration solver is implemented in the study. A case study is presented of a pile foundation subjected to a range of load types and progressive scour damage, and the calculated response from the various conditions is evaluated. Periodic loads of varying amplitude are implemented. Effective linear models are developed based on the secant stiffness approach and compared to the nonlinear systems. The paper presents the numerical basis for the analysis, results of the case study and advice on modelling issues relevant to analyses of this nature.

Abstract

We present an application on springback of an anisotropic elastoplastic finite strain model. The computational procedure allows the sole use of internal elastic variables as it is fully hyperelastic. Based on the Lee-type multiplicative decomposition, which permits an additive solution in a logarithmic strain approach, it resolves the 'rate issue' by the employment of a corrector rate of elastic strains and its algorithmic implementation [1]. The associative nature and Clausius-Duhem inequality recover the formulation of Simo's isotropic strain-hardening particular case. Its application to metal and soft materials has been corroborated as it is also valid for anisotropic yield functions and for any anisotropic stored energy (linear and nonlinear in logarithmic strains) [2]. The computation of the stress-point algorithm is done using a simple backward-Euler scheme. Its validation has been performed by the implementation of the procedure into a subroutine which allows the user to compute benchmark models in the commercial program ADINA [3]. This permits to explore its usefulness in application for the numerical prediction of springback in thin sheet metal forming processes. The results are compared with both experimental and other authors' results. Moreover, many simulations were performed such as the draw-bending test, the unconstrained bending problem, the square cup or the 3DS warping benchmark.