Drone-based electromagnetic survey system for geophysical applications

Geophysical electromagnetic (EM) methods are used in geological mapping, mineral exploration, groundwater studies and geotechnical investigations. Airborne EM methods have the benefit of avoiding terrain obstacles such as lakes, rivers, swamps, and ravines. Compared to manned aircrafts, drones or unmanned aerial vehicles (UAVs) have benefits of their own. Drone-based surveys are versatile, fast to deploy, economical and ecologically more friendly. Presently, magnetic surveying is the only geophysical method that is routinely conducted with drones. The modest maximum payload limit of drones imposes severe restrictions on the applicability of other methods including EM and radiometric methods, for example. Finnish company, Radai Ltd has been developing Louhi, a novel drone-based frequency-domain EM survey system, in an EU funded Horizon 2020 project NEXT – New Exploration Technologies. The EM system has two operation options – the first uses a large loop on the ground as an EM source and the other uses a small portable EM transmitter loop. Both systems utilize a stand-alone and light-weight three-component EM receiver that can be towed by a drone. This article presents the theoretical background of the EM methods, the solution developed by Radai Ltd, the current version of the EM device, and results from field and flight tests that demonstrate the applicability of the drone-based EM system under development.

Evaluation of a Semi-Airborne Electromagnetic Survey Based on a Multicopter Aircraft System

The semi-airborne electromagnetic (EM) method has the potential to reach deeper exploration depths than purely airborne EM approaches. The concept of the method is to deploy high-power transmitters on the ground, which excite subsurface currents and induce strong magnetic fields, and to measure the corresponding EM fields with a passive airborne receiver instrument. Following recent conceptual developments of the semi-airborne EM technique deployed on helicopters, we performed a 10 km2 semi-airborne EM survey near Münster (Germany) based on a multicopter aircraft system. For this purpose, horizontal electric dipole (HED) transmitters were installed in the survey area and were surveyed individually. Magnetic transfer functions were determined and a model of the conductivity of the study area was derived. Despite restrictions such as low payload capacity and multicopter-related EM noise, we were able to estimate spatially and spectrally consistent transfer functions of high quality up to a distance of 2 km from the respective transmitter. Our results could be validated with independent results from a magnetotelluric and a direct current sounding. The study demonstrates that an unmanned aircraft system (UAS) is suitable for semi-airborne EM application and that such a system can be beneficial where ground-based methods and manned techniques become impractical.

Forward Modeling and 3D Inversion of EM Data Collected Over the McArthur River Uranium Deposit in the Athabasca Basin, Canada

Detection and assessment of the deeply buried high-grade uranium deposits in the Athabasca Basin rely on geophysical methods to map conductive rocks. Variable Quaternary surface cover can mask the anomalous signals from depth and affect interpretation of inverted conductivity models. We present the analysis of a number of EM modeling studies and two field data sets, to demonstrate the effects of varying Quaternary cover resistivity and thickness, on the ability to resolve the parameters of underlying sandstone, alteration, and basement conductors. Synthetic data, assuming a typical shallow EM sounding system and realistic resistivities found in the Athabasca Basin, show that resistivity and thickness parameters of the Quaternary cover can be separately recovered in cases where this cover is more conductive than the underlying sandstone, but not when the cover is significantly more resistive. A 3D modeling study shows that using airborne EM data, it is possible to detect a basement conductor of 20 S at a depth of at least 600 m below surface, even in the presence of Quaternary cover thickness variations of the up to 20% (40 m to 60 m). Furthermore, while Quaternary cover variations and deeper sandstone alteration can produce comparable anomalous signal amplitudes in a time-domain EM response, their effects are most visible in distinctly separate time windows. Analysis of a GPR field data set to map the thickness of Quaternary cover in the McArthur River area, indicates that this cover consists mostly of sandy tills and ranges in thickness from 0 to 117 m. Constrained 3D inversion of an airborne EM data set from the same area shows basement conductors consistent with the depth and location of a known fault. Elevated conductivity in the sandstone by up to a factor of two over the background values could indicate possible alteration.

3D Wavelet Finite-Element Modeling of Frequency-Domain Airborne EM Data Based on B-Spline Wavelet on the Interval Using Potentials

We present a wavelet finite-element method (WFEM) based on B-spline wavelets on the interval (BSWI) for three-dimensional (3D) frequency-domain airborne EM modeling using a secondary coupled-potential formulation. The BSWI, which is constructed on the interval (0, 1) by joining piecewise B-spline polynomials between nodes together, has proved to have excellent numerical properties of multiresolution and sparsity and thus is utilized as the basis function in our WFEM. Compared to conventional basis functions, the BSWI is able to provide higher interpolating accuracy and boundary stability. Furthermore, due to the sparsity of the wavelet, the coefficient matrix obtained by BSWI-based WFEM is sparser than that formed by general FEM, which can lead to shorter solution time for the linear equations system. To verify the accuracy and efficiency of our proposed method, we ran numerical experiments on a half-space model and a layered model and compared the results with one-dimensional (1D) semi-analytic solutions and those obtained from conventional FEM. We then studied a synthetic 3D model using different meshes and BSWI basis at different scales. The results show that our method depends less on the mesh, and the accuracy can be improved by both mesh refinement and scale enhancement.