Identification of dynamic soil properties relevant for offshore wind turbines through full waveform inversion of in-situ measured seismic data

Abstract

Wind energy needs to become cheaper in order to take on the competition with current non-durable energy sources. Innovative design procedures and better understanding of structural behavior are aimed for to improve the current wind turbine design. A better understanding of turbine behavior consists of several fields and one of these fields is soil-structure interaction. Currently input parameters for soil-structure interaction models are based on empirical methods. The objective is to estimate the dynamic soil parameters in the first 50m of soil in an offshore environment. In this thesis, a non-invasive seismic measurement is performed in order to identify dynamic soil parameters relevant for offshore wind turbines. The seabed surface response of an active source is measured through a streamer of hydrophones and results in a so called shot record. The data obtained from this measurement is used in a newly developed routine that can can identify an estimate of the shear wave velocity profile. Furthermore the routine is adapted to incorporate a second unknown parameter, the compressional wave velocity structure. These two parameters are the main contributors to the shear modulus, bulk modulus and Poissons ratio of the soil. An analytical linear-elastic full-waveform calculation is performed for a homogeneous, isotropic representation of the soil where the soil is assumed to be perfectly horizontally stratified and is overlain by a water layer. A genetic inversion algorithm is then used to update an initial guess of the soil model. The results show that full waveform inversion has high potential to successfully estimate the complex shear wave profiles. In addition, material damping has a large influence on the shape of the spectrum and should be included in the inversion.