Multiparameter Inversion Of Electromagnetic And Acoustic Wave In Frequency Domain

Under the background of growing demand for detailed description of underground target,we need a more robust,detailed and high resolution geophysical method to map the underground mineral.But traditionaly,each kind of geophysical method only provides partial information of the underground targets,individual methods have certain limitation and ambiguity inherent which leads to non-uniqueness when solving inverse problem.For obtaining a more robust and detailed inversion results,two inversion strategies were discussed: joint inversion and muti-parameter inversion.Firstly,for getting a more detailed and robust inversion results,we proposed a joint inversion method to integrate seismic information and Magnetotelluric(MT)data to ameliorate the non-uniqueness in the inversion procedure.We discussed seismic and MT inversion methods seperatley and finally get the joint inversion algorithm.Seismic full wave inversion(FWI)is a powerful data-fitting procedure to reconstruct velocity model,which can reveal detailed information of the structure and lithology under complex geological background.And frequency domain FWI with Limited-Memory-Broyden–Fletcher–Goldfarb–Shanno(L-BFGS)algorithm need to store only a few gradients of the previous nonlinear iterations and significantly improves calculation efficiency.In order to verify the applicability of the algorithm,we applied the algorithm to Overthrust model.Since seismic method has advantages of high depth penetration and resolution,it can be used to map structures or contacts between lithologies where metallic ore deposits are known to accumulate and compared to other geophysical methods.It can make up for the deficiency of other geophysical method in the field of detecting hidden mines.Therefore,in this paper,the FWI method based on L-BFGS algorithm is used to calculate the metallic ore models and we get a good result.MT method is a type of passive electromagnetic(EM)geophysical technique and has advantage in mapping low resistance structures.In this paper,the finite element method is used to calculate the forward modeling,and the calculation results are compared with the analytical solution to verify the correctness of the algorithm.Gauss-newton inversion method is used for MT inversion.ACB method is applied in the inversion procedure to automatically obtain the regularization factor value,and the feasibility of the algorithm is verified by simple model.In our work,we proposed a joint inversion method with a cross-gradient constraint for L-BFGS seismic FWI and ACB based MT Gauss–Newton inversion.Using the joint inversion algorithm,we simultaneously inverted the resistivity model and p-wave velocity model.In order to verify the applicability of the algorithm,we applied the algorithm to a simple synthetic model;the numerical simulation results showed that joint inversion significantly improved the resolution of MT inversion result,as compared to the result obtained with separate Gauss-Newton inversion.The stability of the FWI procedure has also been improved.Then,we applied the method to Jinchuan Ni-Cu-(PGE)sulfide deposit model and achieved results that were better than those obtained for separate inversions.In the traditional ground-penetrating radar(GPR)application,people always considered the influence of permittivity and neglected conductivity.For inverting permittivity and conductivity models simultaneously,the influence of different parameters on the misfit function is analyzed,and nonlinear transformation of parameter and normalized gradient is used to balance the sensitivities of the objective function to different parameters.Then,we apply memoryless quasi-Newton(MLQN)method to solve inverse problem.MLQN can attain acceptable results with low computational cost and small memory storage requirements.Numerical tests are examined from on-ground multioffset GPR data with and without white Gaussian noise.Numerical tests show that our inversion strategies are feasible and reliable in simultaneous inversion of permittivity and conductivity from on-ground GPR data.