Comparison of nodal and edge basis functions for the forward modelling of three‐dimensional frequency‐domain wire source electromagnetic data using a potentials formulation

Abstract

ABSTRACTAn investigation of forward modelling 3‐D wire source electromagnetic data in the frequency domain using nodal and edge finite‐element basis functions is presented. The equations solved are those for the coupled secondary vector–scalar potentials (i.e. magnetic vector potential A and electric scalar potential φ), which give rise to a less ill‐conditioned system of equations to be solved. The effectiveness of these two methods is validated with two 3‐D examples, i.e. multi‐blocks models and topography models, which considered the computation of the A–φ potentials and the electromagnetic fields due to a grounded electric line source by inter‐comparing the numerical solutions computed from each method and by comparing them against the equivalent solutions for the same model computed using other methods. Based on the multiple‐blocks models, we compare the relative contributions of the inductive and galvanic parts to the E‐field for different frequencies. Results show that for the type of source considered in this paper, the galvanic parts dominate for a relatively low frequency (10 Hz), but that for a relatively high frequency (10 kHz), the inductive parts also contribute. The extent to which the Coulomb gauge condition is satisfied when using nodal basis functions for A is also investigated, with results showing that the divergence of the numerically calculated A is almost zero as desired. Based on the topography models, we analyse the electromagnetic response characteristics at different survey planes, with results of this investigation indicating that the electromagnetic responses measured over a plane close to the target will carry more accurate geometry information (such as position, shape and size) of the target, which is expected, and less distortion due to topography compared to those measured at the earth's surface. This suggests that when conducting electromagnetic prospecting in an area with complex topography, it is possible to obtain better anomaly detection affected less by topography if the electromagnetic measurement is carried out underground (such as in a borehole).