Research On Susceptibility Inversion Technology Of Near-field Magnetic Sources

Susceptibility inversion of near-field magnetic sources is of great application significance in the field of magnetic target detections,such as wood nail detection,underground pipeline detection,mine or abandoned bomb clearances,antiquities exploration,shipwreck positioning,etc.The traditional method which is currently used mainly applied for long-distance magnetic body of kilometer level,and defaults the magnetization direction of the magnetic body to the same direction as the geomagnetic field,and the traditional inversion method discretizes the magnetic body into a rectangular parallelepiped unit and uses the analytical formula of the rectangular parallelepiped to calculate the nuclear matrix.But in reality,most magnetic bodies are not only magnetized by the background magnetic field,but also have strong remanence,and their shape is not limited to the combination of cuboids.Aiming at the shortcomings of the existing methods,this paper transforms the total field norm anomaly on the observation surface into the total field vector anomaly to suppress the effects of remanence,calculates the magnetic matrix with the principle of magnetic dipole equivalent,and proposes a method for the three-dimensional inversion and imaging of magnetic susceptibility of magnetic sources with unknown shape,position and magnetization state under near-field conditions.In this paper,the unknown space to be tested is divided into a grid-like magnetic susceptibility model.The objective function is constructed with the magnetic field vector anomaly modulus and optimized to get the solution.The Gauss-Newton algorithm is used to linearize the nonlinear inversion problem in the objective function,and then conjugate gradient algorithms is used to solve the high-dimensional linear equations.Finally,the Tikhonov regularization curve is utilized to determine the optimal regularization parameters,and the nonlinear transformation method is used to constrain the range of the magnetic susceptibility in the whole inversion process.Simulations and experiments demonstrate that,with magnetic data from a finite number of points in the plane,the proposed method is capable to realize 3D inversions of curved pipe,spatial discrete cuboid magnetic source,laminated cuboid magnetic source,"V" shaped pipe,cross pipe and "L" shaped steel pipe under near-field conditions within two meters.Through the magnetic field measurement experiments in the actual environment and the implementation of the proposed optimization inversion algorithm,accurate inversions of the magnetic susceptibility distribution and magnetic imaging of two strong magnets in the same plane and “L” shaped steel tube are realized,testifying that the proposed method is feasible to be applied to the detection of spatial discrete magnetic body and buried pipelines.And it is confirmed that the accuracy of this method in the depth direction is limited so that the method does not apply to the model inversion where spatial depth differences exist.