Yunhe Liu

3D forward modeling and response analysis for marine CSEMs towed by two ships

A dual-ship-towed marine electromagnetic (EM) system is a new marine exploration technology recently being developed in China. Compared with traditional marine EM systems, the new system tows the transmitters and receivers using two ships, rendering it unnecessary to position EM receivers at the seafloor in advance. This makes the system more flexible, allowing for different configurations (e.g., in-line, broadside, and azimuthal and concentric scanning) that can produce more detailed underwater structural information. We develop a three-dimensional goal-oriented adaptive forward modeling method for the new marine EM system and analyze the responses for four survey configurations. Oceanbottom topography has a strong effect on the marine EM responses; thus, we develop a forward modeling algorithm based on the finite-element method and unstructured grids. To satisfy the requirements for modeling the moving transmitters of a dual-ship-towed EM system, we use a single mesh for each of the transmitter locations. This mitigates the mesh complexity by refining the grids near the transmitters and minimizes the computational cost. To generate a rational mesh while maintaining the accuracy for single transmitter, we develop a goal-oriented adaptive method with separate mesh refinements for areas around the transmitting source and those far away. To test the modeling algorithm and accuracy, we compare the EM responses calculated by the proposed algorithm and semi-analytical results and from published sources. Furthermore, by analyzing the EM responses for four survey configurations, we are confirm that compared with traditional marine EM systems with only in-line array, a dual-ship-towed marine system can collect more data.

Efficient Modeling of Time-domain AEM using Finite-volume Method

ABSTRACT We present an efficient three-dimensional (3D) time-domain airborne electromagnetic (EM) modeling based on finite-volume method, in combination with the merits of solving the secondary field, local mesh, and the direct solver. Taking the strategy of separating primary from secondary field for the calculation of time-domain EM field, we greatly reduce the number of grids in comparison to direct solution of the total field. The primary field is obtained by Hankel transform from the frequency-domain for a half-space or layered earth model. The techniques of local mesh and direct solver are adopted to further improve the modeling efficiency. We compare and discuss the characteristics of the method developed in this paper from time consumption and grid sizes. The EM responses for different transmitting waveforms are calculated via a convolution between step-wave responses and the transmitting current or its derivatives. All numerical experiments on models with multiple geological bodies and fracture z...

3-D Modeling for Airborne EM using the Spectral-element MethodYin et al.: EM Modeling with the Spectral Element Method

ABSTRACT The 3-D spectral-element method (SE) based on Gauss–Lobatto–Legendre (GLL) polynomials is a very efficient and accurate solver for high-frequency computational electromagnetic (EM) modeling. Although the SE method is based on a weighted residual technique similar to the finite-element method (FE), it utilizes polynomial interpolation functions rather than linear interpolation functions. The GLL polynomials have the characteristic of exponential convergence with the order of polynomial that helps improve the modeling accuracy. We apply the SE method in the frequency range (900 Hz to 56 kHz) to airborne EM modeling. Starting from the vector Maxwells equations, we use the SE method to establish spatial-discrete forms of 3-D vector Helmholtz equations and adopt GLL integration to calculate the matrix elements. The magnetic field is calculated using Faradays law. To check the accuracy of the SE algorithm, we compare our results with semi-analytical solutions for a homogeneous half-space and a layere...

Gravity compression forward modeling and multiscale inversion based on wavelet transform

The main problems in three-dimensional gravity inversion are the non-uniqueness of the solutions and the high computational cost of large data sets. To minimize the high computational cost, we propose a new sorting method to reduce fluctuations and the high frequency of the sensitivity matrix prior to applying the wavelet transform. Consequently, the sparsity and compression ratio of the sensitivity matrix are improved as well as the accuracy of the forward modeling. Furthermore, memory storage requirements are reduced and the forward modeling is accelerated compared with uncompressed forward modeling. The forward modeling results suggest that the compression ratio of the sensitivity matrix can be more than 300. Furthermore, multiscale inversion based on the wavelet transform is applied to gravity inversion. By decomposing the gravity inversion into subproblems of different scales, the non-uniqueness and stability of the gravity inversion are improved as multiscale data are considered. Finally, we applied conventional focusing inversion and multiscale inversion on simulated and measured data to demonstrate the effectiveness of the proposed gravity inversion method.

3D anisotropic modeling and identification for airborne EM systems based on the spectral-element method

The airborne electromagnetic (AEM) method has a high sampling rate and survey flexibility. However, traditional numerical modeling approaches must use high-resolution physical grids to guarantee modeling accuracy, especially for complex geological structures such as anisotropic earth. This can lead to huge computational costs. To solve this problem, we propose a spectral-element (SE) method for 3D AEM anisotropic modeling, which combines the advantages of spectral and finite-element methods. Thus, the SE method has accuracy as high as that of the spectral method and the ability to model complex geology inherited from the finite-element method. The SE method can improve the modeling accuracy within discrete grids and reduce the dependence of modeling results on the grids. This helps achieve high-accuracy anisotropic AEM modeling. We first introduced a rotating tensor of anisotropic conductivity to Maxwell’s equations and described the electrical field via SE basis functions based on GLL interpolation polynomials. We used the Galerkin weighted residual method to establish the linear equation system for the SE method, and we took a vertical magnetic dipole as the transmission source for our AEM modeling. We then applied fourth-order SE calculations with coarse physical grids to check the accuracy of our modeling results against a 1D semi-analytical solution for an anisotropic half-space model and verified the high accuracy of the SE. Moreover, we conducted AEM modeling for different anisotropic 3D abnormal bodies using two physical grid scales and three orders of SE to obtain the convergence conditions for different anisotropic abnormal bodies. Finally, we studied the identification of anisotropy for single anisotropic abnormal bodies, anisotropic surrounding rock, and single anisotropic abnormal body embedded in an anisotropic surrounding rock. This approach will play a key role in the inversion and interpretation of AEM data collected in regions with anisotropic geology.

3D regularized focusing inversion of gravity data with a new stabilizing functional

Regularized focusing inversion has been proposed to generate clearer and more focused images of geological structures with sharp boundaries. It is based on special stabilizing functionals called minimum support (MS) functional and minimum gradient support (MGS) functional. In this paper, we have developed a new stabilizing functional that minimizes the area where strong model parameter variations and discontinuity occur. Compared with the MS stabilizer, the new stabilizer presented in this paper greatly reduces computing cost by reducing the number of iterations. Finally, we apply the new stabilizer to synthetic 3D gravity data to prove the effectiveness and efficiency of the method.