Experimental application of FRF-based model updating approach to estimate soil mass and stiffness mobilised under pile impact tests

Abstract

The dynamic response of structures in contact with soil is receiving increasing interest and there is a growing need for more accurate models capable of simulating the behaviour of these systems. This is particularly important in the field of offshore wind turbines, where accurate estimates of system frequency are needed to avoid resonance, and in the structural health monitoring fields, where accurate reference damage models are used. Previous work has shown that there is significant uncertainty in how to specify mobilised soil stiffness for dynamic soil-pile interaction modelling. Moreover, the contribution of soil mass in dynamic motion is often ignored. This paper applies a finite-element iterative model updating approach previously developed by the authors to two experimental piles to ascertain the mobilised soil stiffness and mass profiles from impact test data. The method works by obtaining a frequency response function (FRF) from an impact test performed on a test pile, developing a numerical model of this system, applying initial estimates of soil mass and stiffness, and updating these properties to match the experimental FRF with that generated in the numerical model. A range of elements are investigated including multiple runs of the approach to test repeatability, the influence of different starting estimates for stiffness, the effect of variability in experimental test data, and the influence of the pile length over which masses are distributed. Moreover, potential sources of error are discussed. The method provides reasonably consistent estimates of the soil stiffness and mass acting in the lateral dynamic motion of a given pile tested in this paper. The approach may be useful in the continued improvement of Soil-Structure Interaction (SSI) modelling for dynamic applications.