Source:Harquail School of Earth Sciences, Laurentian University, Volume PhD, p.139 (2017)
Keywords:2D inversion, Back scattering inversion, Comsol, Contrast source inversion (CSI), Cross-hole, electromagnetic, Finite-element method (FEM), Forward modelling, Radio-imaging method (RIM), SIRT, TE
The radio imaging method (RIM) is a cross-hole electromagnetic (EM) method which employs radio frequency EM waves to estimate the electric properties between boreholes. RIM is applied in hard rock mining to find and delineate sulfide mineral deposits. A basic and relatively simple method for imaging and interpreting RIM data is the straight-ray method. However, the strengths and weaknesses of the straight-ray method and other more sophisticated methods have not been studied thoroughly.<br/>In the first part this research, I modelled RIM data using a finite element package, Comsol Multiphysics. To validate the Comsol approach, I compared the Comsol model data with the analytical solution of an electric dipole in a homogeneous whole-space model, and some published analytical solutions and numerical solutions of models with conductive objects. The Comsol generated data are consistent with the analytical and the published results.<br/>Secondly, I used Comsol synthetic data to assess the effectiveness of the straight-ray method for interpreting RIM data and to study the characteristics of the radio-frequency EM fields. I studied four sets of models with conductive objects embedded in resistive environments, which resemble ore deposits in mining settings. The experiments show that the characteristics of the EM fields mainly depend on the wavelength. Longer wavelengths are associated with lower frequencies. In this condition, EM induction is strong. Shorter wavelengths are associated with higher frequencies. In this condition, the scattering effects of EM waves dominate. In the radio-frequency range, I concluded that the straight-ray method cannot always provide high quality imaging results for RIM data.<br/>To account for the scattering effects, I adopted the contrast source inversion (CSI) method, which was originally developed for microwave tomography in medical imaging, to invert the RIM data. The CSI method was tested with Comsol synthetic data and field data. The synthetic studies show that the CSI method provides images with more accurate locations and shapes of the conductive objects when compared with the straight-ray method. The case studies show that CSI imaging results are more consistent for data collected at different frequencies and are easier to interpret geologically.