Forward modeling of radio imaging (RIM) data with the Comsol RF module
Publication Type:
Journal ArticleSource:
Computers & GeosciencesComputers & Geosciences, Volume 85, p.60-67 (2015)ISBN:
0098-3004Accession Number:
WOS:000365051100006Keywords:
cross-hole imaging, finite element, geophysics, numerical modeling, radio imaging method, sirt algorithm, tomographyAbstract:
Radio imaging method (RIM) is an electromagnetic (EM) tomographic method, which can be applied to image the electrical properties (principally the conductivity) in the plane defined by two boreholes. RIM employs the EM waves at radio frequencies between 100 kHz and 10 MHz, and the manner that these waves propagate around subsurface ore bodies has not been studied thoroughly. We studied the wave propagation using the finite element modeling (FEM) algorithm as implemented in the Comsol RF module. An appropriate element size is quantified by comparing the Comsol modeling results of 6 types of element sizes at 4 frequencies with the analytical solution in the homogeneous whole space. The FEM model data with 5 elements per wavelength have errors less than 5%; 7-8 elements per wavelength provide the errors around 1%; when there are 10 elements per wavelength, the errors are less than 1%. Comparison studies for more complicated models with anomalous conductivity structures show that the Comsol modeling results are consistent with results derived from analytical solutions, finite-difference time-domain methods and integral equations. To illustrate the flexibility of the Comsol method for RIM modeling, we provide an example with two moderately conductive bodies between boreholes. Receiver profiles and a relative variation map show that when the conductive bodies are two wavelengths away from the source, the EM wave attenuation and reflection by the conductive bodies can be observed. The amplitude tomography of the model data reveals that with the SIRT algorithm, the location of the conductive anomalies can be reconstructed successfully, although, some limitations exist such as low resolution, incorrect conductivity estimation, and some artifacts. From our work, we conclude that Comsol modeling is helpful to study radio wave propagation and imaging methods. (C) 2015 Elsevier Ltd. All rights reserved.
Notes:
A<br/>Cw5qi<br/>Times Cited:1<br/>Cited References Count:22