Showing 1–5 of 5 results for author: Deleersnyder, W

A multidimensional AI-trained correction to the 1D approximate model for Airborne TDEM sensing

Authors: Wouter Deleersnyder, David Dudal, Thomas Hermans

Abstract: The computational resources required to solve the full 3D inversion of time-domain electromagnetic data are immense. To overcome the time-consuming 3D simulations, we construct a surrogate model, more precisely, a data-driven statistical model to replace the 3D simulations. It is trained on 3D data and predicts the approximate output much faster. We construct a surrogate model that predicts the di… ▽ More The computational resources required to solve the full 3D inversion of time-domain electromagnetic data are immense. To overcome the time-consuming 3D simulations, we construct a surrogate model, more precisely, a data-driven statistical model to replace the 3D simulations. It is trained on 3D data and predicts the approximate output much faster. We construct a surrogate model that predicts the discrepancy between a 1D subsurface model and a deviation of the 1D assumption. The latter response is fastly computable with a semi-analytical 1D forward model. We exemplify the approach on a two-layered case. The results are encouraging even with few training samples. Given the computational cost related to the 3D simulations, there are limitations in the number of training samples that can be generated. In addition, certain applications require a wide range of parameters to be sampled, such as the electrical conductivity parameters in a saltwater intrusion case. The challenge of this work is achieving the best possible accuracy with only a few thousand samples. We propose to view the performance in terms of learning gain, representing the gain from the surrogate model whilst still acknowledging a residual discrepancy. Our works open new avenues for effectively simulating 3D TDEM data. △ Less

Submitted 9 July, 2024; v1 submitted 23 November, 2023; originally announced November 2023.

Journal ref: Computers & Geosciences, 188, 105602 (2024)

Determining the optimal focusing parameter in sparse promoting inversions of EMI surveys

Authors: Wouter Deleersnyder, David Dudal, Benjamin Maveau, Marieke Paepen

Abstract: If the magnetic field caused by a magnetic dipole is measured, the electrical conductivity of the subsurface can be determined by solving the inverse problem. For this problem a form of regularisation is required as the forward model is badly conditioned. Commonly, Tikhonov regularisation is used which adds the $\ell_2$-norm of the model parameters to the objective function. As a result, a smooth… ▽ More If the magnetic field caused by a magnetic dipole is measured, the electrical conductivity of the subsurface can be determined by solving the inverse problem. For this problem a form of regularisation is required as the forward model is badly conditioned. Commonly, Tikhonov regularisation is used which adds the $\ell_2$-norm of the model parameters to the objective function. As a result, a smooth conductivity profile is preferred and these types of inversions are very stable. However, it can cause problems when the true profile has discontinuities causing oscillations in the obtained model parameters. To circumvent this problem, $\ell_0$-approximating norms can be used to allow discontinuous model parameters. Two of these norms are considered in this paper, the Minimum Gradient Support and the Cauchy norm. However, both norms contain a parameter which transforms the function from the $\ell_2$- to the $\ell_0$-norm. To find the optimal value of this parameter, a new method is suggested. It is based on the $L$-curve method and finds a good balance between a continuous and discontinuous profile. The method is tested on synthetic data and is able to produce a conductivity profile similar to the true profile. Furthermore, the strategy is applied to newly acquired real-life measurements and the obtained profiles are in agreement with the results of other surveys at the same location. Finally, despite the fact that the Cauchy norm is only occasionally used to the best of our knowledge, we find that it performs at least as good as the Minimum Gradient Support norm. △ Less

Submitted 22 November, 2022; originally announced November 2022.

Novel Airborne EM Image Appraisal Tool for Imperfect Forward Modelling

Authors: Wouter Deleersnyder, David Dudal, Thomas Hermans

Abstract: Full 3D inversion of time-domain Airborne ElectroMagnetic (AEM) data requires specialists' expertise and a tremendous amount of computational resources, not readily available to everyone. Consequently, quasi-2D/3D inversion methods are prevailing, using a much faster but approximate (1D) forward model. We propose an appraisal tool that indicates zones in the inversion model that are not in agreeme… ▽ More Full 3D inversion of time-domain Airborne ElectroMagnetic (AEM) data requires specialists' expertise and a tremendous amount of computational resources, not readily available to everyone. Consequently, quasi-2D/3D inversion methods are prevailing, using a much faster but approximate (1D) forward model. We propose an appraisal tool that indicates zones in the inversion model that are not in agreement with the multidimensional data and therefore, should not be interpreted quantitatively. The image appraisal relies on multidimensional forward modelling to compute a so-called normalized gradient. Large values in that gradient indicate model parameters that do not fit the true multidimensionality of the observed data well and should not be interpreted quantitatively. An alternative approach is proposed to account for imperfect forward modelling, such that the appraisal tool is computationally inexpensive. The method is demonstrated on an AEM survey in a salinization context, revealing possible problematic zones in the estimated fresh-saltwater interface. △ Less

Submitted 12 October, 2022; originally announced October 2022.

Flexible quasi-2D inversion of time-domain AEM data, using a wavelet-based complexity measure

Authors: Wouter Deleersnyder, Benjamin Maveau, David Dudal, Thomas Hermans

Abstract: Regularization methods improve the stability of ill-posed inverse problems by introducing some a priori characteristics for the solution such as smoothness or sharpness. In this contribution, we propose a multidimensional, scale-dependent wavelet-based L1-regularization term to cure the ill-posedness of the airborne (time-domain) electromagnetic induction inverse problem. The regularization term i… ▽ More Regularization methods improve the stability of ill-posed inverse problems by introducing some a priori characteristics for the solution such as smoothness or sharpness. In this contribution, we propose a multidimensional, scale-dependent wavelet-based L1-regularization term to cure the ill-posedness of the airborne (time-domain) electromagnetic induction inverse problem. The regularization term is flexible, as it can recover blocky, smooth and tunable in-between inversion models, based on a suitable wavelet basis function. For each orientation, a different wavelet basis function can be used, introducing an additional relative regularization parameter. We propose a calibration method to determine (an educated initial guess for) this relative regularization parameter, which reduces the need to optimize for this parameter, and, consequently, the overall computation time is under control. We apply our novel scheme to a time-domain airborne electromagnetic data set in Belgian saltwater intrusion context, but the scheme could equally apply to any other 2D or 3D geophysical inverse problem. △ Less

Submitted 25 January, 2023; v1 submitted 13 May, 2022; originally announced May 2022.

Inversion of electromagnetic induction data using a novel wavelet-based and scale-dependent regularization term

Authors: Wouter Deleersnyder, Benjamin Maveau, Thomas Hermans, David Dudal

Abstract: The inversion of electromagnetic induction data to a conductivity profile is an ill-posed problem. Regularization improves the stability of the inversion and, based on Occam's razor principle, a smoothing constraint is typically used. However, the conductivity profiles are not always expected to be smooth. Here, we develop a new inversion scheme in which we transform the model to the wavelet space… ▽ More The inversion of electromagnetic induction data to a conductivity profile is an ill-posed problem. Regularization improves the stability of the inversion and, based on Occam's razor principle, a smoothing constraint is typically used. However, the conductivity profiles are not always expected to be smooth. Here, we develop a new inversion scheme in which we transform the model to the wavelet space and impose a sparsity constraint. This sparsity constrained inversion scheme will minimize an objective function with a least-squares data misfit and a sparsity measure of the model in the wavelet domain. A model in the wavelet domain has both temporal as spatial resolution, and penalizing small-scale coefficients effectively reduces the complexity of the model. Depending on the expected conductivity profile, an optimal wavelet basis function can be chosen. The scheme is thus adaptive. Finally, we apply this new scheme on a frequency domain electromagnetic sounding (FDEM) dataset, but the scheme could equally apply to any other 1D geophysical method. △ Less

Submitted 7 September, 2020; originally announced September 2020.