Research On Electromagnetic Properties Of Seabed Sulfide Minerals Based On RBFNN Algorithm

The integral equation method is one of the commonly used methods to analyze the electromagnetic characteristics of marine mineral resources,and its core is the calculation of dyadic Green's function.In the calculation process of dyadic Green's function,Sommerfeld integral operation is required,but its high oscillation and slow attenuation will cause the calculation process to be time-consuming.The introduction of neural network method can improve the calculation cost of dyadic Green's function.It has important research significance.This thesis proposes a method for calculating the dyadic Green's function in a plane layered medium by using the Radial Basis Function Neural Network(RBFNN),and applies it to the study of electromagnetic properties of submarine sulfide mines.In this paper,based on the transmission line theory,a method for solving the dyadic Green's function in a plane layered medium is firstly given.Then,by analyzing the solution process of the dyadic Green's function in a plane layered medium,it is determined that RBFNN can solve the dyadic Green's function in a plane layered medium.And the orthogonal least square method(Orthogonal Least Square,OLS)is used to train the RBFNN.Through a large number of training,the nonlinear relationship between the parameters of the model and the dyadic Green function in the plane layered medium is expressed by the linear relationship.In the subsequent calculation of the dyadic Green's function in the plane layered medium,it can be quickly calculated by simply calling the corresponding linear relationship.In the numerical simulation,the Swidinsky double half-space model was used to verify the effectiveness of the RBFNN algorithm to solve the dyadic Green's function.The results show that when the relative errors of each component are controlled within 2%,the RBFNN solution method is faster than the direct numerical integration method 14 times.Finally,the RBFNN algorithm is applied to the study of electromagnetic characteristics of submarine sulfide mines,and the electromagnetic characteristics of the receiving array of layered submarine sulfide mines in linear,surface and body shapes are numerically calculated based on the RBFNN algorithm and the direct numerical integration method.Numerical simulation results show that when the relative error of each component is controlled within 2%,the RBFNN solution method is faster than the direct numerical integration method by 21 times,23 times and 31.8 times.And the RBFNN algorithm is faster than the direct numerical integration method.It is more efficient and provides a certain theoretical reference for the electromagnetic exploration and research of submarine sulfide mines in the actual natural environment.