Study On Magnetotelluric Data Inversion Using Step-by-Step Regularization Strategy Based On Adaptive Unstructured Mesh

Regularization technology can include a priori information in the form of data and model covariance matrices,which effectively improve the stability of inverse and reduce the ill-posedness of the inversion.So it is widely used in geophysical inversion.This is an extremely useful approach when sufficient a priori information about the subsurface exists,but it may produce misleading results when the assumptions and a priori information are insufficient and possibly flawed.Another option for reducing the underdetermined aspect of the inversion problem is improving the quality of datasets and increasing the quantity of datasets.However,this method cannot change the ill-posedness of the inverse problem caused by the equivalence of geophysical field.In addition,reducing the number of inversion unknowns is another important method to effectively reduce the ill-posedness of the inversion,however this method will impact the resolution of the inversion.On the other hand,if the discrete inversion mesh can not reconstruct the actual anomaly body,it may lead to multiple solutions of inversion and affect the resolution of inversion.A step-by-step regularization inversion scheme based on adaptive mesh is investigated in this paper.In the initial stage of the inversion,coarse mesh is adopted for the inversion,and the ill-posedness of the inversion is decreased by reducing the number of inversion elements.During the iterative inversion process,mesh is adaptively refined according to mesh refinement strategies to get better imaging of abnormal bodies.The inversion results of the previous mesh are used as the reference model and the initial model in the inversion of the next mesh,so as to ensure the model improvements along the correct direction of the inversion,and then improve the inversion stability and inversion results.Four mesh refinement strategies were proposed,including model sensitivity,model variation,model gradient and edge-angle detection.The characteristics of the four mesh optimization schemes are analyzed by Hessian matrix eigenvalue distribution,and the adaptive inversion results of four mesh refinement schemes are compared.The selection of mesh refinement ratio is discussed and adaptive inversion is compared with traditional fixed mesh inversion.Finally,the step-by-step regularization inversion scheme based on adaptive mesh is used for two-dimensional and three-dimensional magnetotelluric(MT)measured data inversion,and the results show that the adaptive inversion algorithm is stable and reliable.