Three-dimensional forward modelling and inversion of controlled-source electromagnetic data using the edge-based finite-element method
Abstract: Electromagnetic geophysical methods are applied to investigate anomalous subsurface structures exhibiting contrasts in electrical resistivity, as for example in mineral exploration or geothermal areas. When evaluating electromagnetic data, one big challenge is to account for full three-dimensional measurement setups and subsurface scenarios. To address this challenge, this work seeks to report on the development of a three-dimensional forward modelling code for controlled-source electromagnetic data and its integration into an inversion framework to enable the computation of three-dimensional resistivity models of the subsurface.This thesis outlines the technical details of the developed forward and inverse modelling routines and describes synthetic tests to verify their effectiveness. The implemented forward modelling routines, based on the linear edge-based finite-element formulation on tetrahedral elements, consist of a standalone code containing a goal-oriented mesh refinement option. The forward modelling code was included with several modules into an existing inversion framework. Electromagnetic field components or impedance tensors and vertical magnetic transfer functions serve as input data for the inversion. Data gradients, required for the used non-linear conjugate-gradient inversion algorithm, were implemented for all input data types. Furthermore, regularisation options were employed and tested. The aforementioned forward and inverse modelling codes were applied for different field settings: A forward modelling study of an ore deposit in central Sweden aimed at finding an adequate receiver distribution for the field measurements. In another modelling project, the mesh design for subsurface models containing steel-cased wells was investigated. Finally, an inversion of controlled-source electromagnetic data collected on a frozen lake in Stockholm for a tunnel-project complemented previously obtained two-dimensional resistivity models with a three-dimensional model. The benefit of the presented research is twofold: First, the new modules added to the inversion framework can be used by the academic community to obtain improved resistivity models. Second, the reported models and field settings provide detailed information for measurement design, meshing of three-dimensional domains and local resistivity distributions of the subsurface allowing to answer geological, hydrological and geotechnical questions of selected field sites.
CLICK HERE TO DOWNLOAD THE WHOLE DISSERTATION. (in PDF format)