An investigation into non-linear soil structure interaction

Abstract

The influence of non-linear soil behaviour on the dynamic response of a structure under earthquake loading was investigated by a number of numerical analyses, modelling the soil as a non-linear solid. Properties of the structure and soil layer were chosen to emphasise the influence of soil-structure interaction (SSI). For the linear elastic analysis, used as a reference of the subsequent non-linear analyses, the fundamental frequency of the structure decreased from 5.0 Hz for the fixed base structure to 3.49 Hz. There was a corresponding increase in the damping ratio from 0.02 to 0.056. The response of the structure was found to be dominated by the first mode, and little evidence of higher modes were found in the frequency range of interest (up to 15 Hz). The principal influences of non-linear soil behaviour on the response of the SSI structure were the decrease in the fundamental frequency, increase in the damping, and an increase in the distortion of the amplification characteristics (transfer function) relative to the SSI structure on the linear soil layer. These were considered consistent with the reduction of stiffness and increase in material damping associated with non-liner soil behaviour in the region of the foundation. The motion recorded at the soil-foundation interface was compared to the corresponding free-field motion. The structure was found to influence the surface motion in two ways; a local distortion of the amplification at the fundamental frequency of the SSI structure and a suppression of high frequency motion. This suppression of high frequency motion became increasingly influential in the non-linear analyses. A discrete mass system of equations was developed to model the structure and foundation of the numerical model. Two additional parameters were introduced to account for the principal effects of non-linear soil behaviour observed in the numerical analyses; the shear modulus ratio (Gequiv∕Gmax) and the material damping ratio (ξmateriai)∙ The use of these parameters provided an excellent approximation of the non-linear behaviour observed in the numerical models. The shear modulus ratio ranged from 1.0 in Iow amplitude analyses linear to 0.31 in the largest loading analyses. The material damping ratio had a corresponding variation of between 0.01 up to 0.23. When plotted as functions of the foundation PGA, both were of the same form as those used to represent the secant shear modulus and damping in cyclic soil testing. This suggested that the development of these parameters would provide a useful means to incorporate the influence of non-linear soil behaviour into the analysis and design of a structure. The influence of the structure on the surface motion into the structure was largely negated by the amplification characteristics of the structure, which strongly de-amplified higher frequency motion where the difference was most significant. In all comparison analyses the response of the structure was greater when the free-field motion was used. This suggested the use of free-field motions (recorded or generated for a specific site) would produce conservative analysis results. The alteration of the backbone relationship of the soil directly had a corresponding effect on the equivalent linear parameters, producing distinctly different curves. The earthquake time history applied at the base of the structure was found to be relatively Uninfluential on the equivalent linear parameters.