Research Of Iterative Finite Element Forward And Inverse Method Of Magnetotelluric With Induced Polarization Effect

The numerical methods of magnetotelluric forward modeling mainly include the finite difference method,the finite element method and the integral equation method.The finite element method is widely used because it can be well applied to complex terrain.However,because the accuracy of the method is proportional to the number of subdivision units in the solution area,there is also the problem that the calculation time increases with the increase of subdivision units,which ensures the accuracy of the solution To improve the operation speed of the finite element method under the premise is a major problem faced by the finite element method forward.The traditional methods of magnetotelluric inversion,such as gradient method and conjugate gradient method,linearize the non-linear problem,first give the initial model,then carry out forward modeling,get the data and compare with the measured data,if it does not meet the target fitting difference,modify the model,then carry out forward modeling,then compare,such cycle until the model parameter value reaches the required fitting difference.The model curve obtained in this process generally has more peaks and troughs intuitively.Relatively speaking,it is easy to have some wrong eigenvalues.How to reduce these wrong eigenvalues as much as possible is a difficult problem for magnetotelluric inversion.At the same time,when the electromagnetic wave propagates in the underground medium,it will produce the induced polarization effect.The traditional magnetotelluric forward and inversion generally do not consider this point,which will cause the error between the calculation results and the actual situation.In view of the above three problems,this paper studies and analyzes the existing data at home and abroad,considers the IP effect in the forward and inverse calculation,selects the DIAS model as the equivalent IP model,and selects the iterative finite element method for forward modeling and Occam method for inversion to try to improve the above three problems.Dsigning The iterative forward algorithm and Occam inversion algorithm for DIAS model,and write programs for the two algorithm,then test them.The forward and inversion trial of G-type,O-type,H-type and prism model of layered media are carried out respectively,which verifies the correctness of the algorithm and the program,and verifies that the iterative finite element method can ensure the accuracy of the solution and improve the speed of forward operation.Discussed the changes of apparent resistivity and phase curve when the polarizability of O-type and H-type model changes,and concludes that H-type model is more suitable to use this forward modeling method to judge whether there is a polarizable layer underground.Occam inversion test is carried out for five parameters of DIAS model,and it is verified that considering IP effect can explore underground structure from multi parameter perspective.