Adaptive finite element for 3D time-domain airborne electromagnetic modeling based on hybrid posterior error estimation

Bo Zhang,Changchun Yin,Xiuyan Ren,Yunhe Liu,Yan-Fu Qi

DOI: https://doi.org/10.1190/GEO2017-0544.1

IF: 3.264

2018-01-01

Geophysics

Abstract:Airborne electromagnetic (AEM) forward modeling has been extensively developed in past years. However, not much attention has been paid to the adaptive numerical algorithms for time-domain electromagnetic modeling. We have created an adaptive method that can generate an effective mesh for time-domain 3D AEM full-wave modeling using an unstructured finite-element method and a backward Euler scheme. For the estimation of the posterior error in the adaptive process, we use a hybrid technique based on the continuity of the normal current density for modeling the off-time channels, and on the continuity of the tangential magnetic field for the on-time channels. To improve the stability of the forward modeling and control the number of grids in the adaptive process, a random grid-selection technique is applied. We check the modeling accuracy of the algorithm by comparing our adaptive results with the semianalytical solution for a time-domain AEM system over a homogeneous half-space. Furthermore, we test the effectiveness of our algorithm for multiple-source time-domain AEM systems by analyzing the meshes generated by the adaptive method and the model results. Finally, we study the topographic effect by calculating time-domain AEM responses over a hill model with an abnormal body embedded.