Research On Transient Electromagnetic Nonlinear Inversion Of Central Loop Devices Based On Improved PSO

Transient electromagnetic method(TEM)is a time-domain electromagnetic detection method.Due to its high resolution and detection efficiency,and its sensitivity to low-resistance objects,TEM is widely used in hydrological surveys,resource detection,engineering investigation and other fields.At present,the forward theory of TEM has developed rapidly,and good results have been achieved in the construction of models and numerical calculations of two-dimension and three-dimension.However,in actual production,TEM inversion problem is a nonlinear parameter estimation problem with multiple solutions and has complicate solution process.Therefore,the inversion work is still mainly one-dimensional.Further research about how to further improve the interpretation accuracy of TEM one-dimensional inversion is urgently needed.The nonlinear inversion method is widely used in TEM inversion due to its better optimization ability,and the search for a more efficient nonlinear inversion algorithm has become a current research hotspot.Based on the forward theory of the center loop device of TEM,the objective function of the TEM nonlinear inversion issue is set up,and the particle swarm optimization(PSO),which is the heuristic intelligent optimization algorithm,is put forward.The PSO algorithm has the superiority of simple operation and rapid convergence and great potential in solving complex multivariable optimization problems,which has been extensively used in TEM inversion.For all that,it has the shortcomings of "premature" phenomenon and insufficient population diversity.To solve the limitation of PSO,this paper proposes the improved hybrid particle swarm algorithm(IH-PSO).On the basis of IH-PSO,the improvement of convergence accuracy is further studied,and an improved hybrid particle swarm joint shuffled frog leaping algorithm(IHPSO-SFLA)is proposed.The main contents of this paper are as follows:(1)On the basis of the PSO algorithm,the IH-PSO algorithm is obtained as follow:use reverse learning strategy to improve population initialization and enrich population diversity;optimize learning factors through sine and cosine acceleration coefficients;utilize sine mapping to adjust inertial weight and expand the search space of the particles;introduce the mutation operator to help the particles jump out of the local optimum.The performance test of the improved algorithm shows that the IH-PSO algorithm can better get rid of local stagnation and obtain higher convergence accuracy than PSO algorithm and HPSO.The IH-PSO algorithm is applied to the inversion of TEM.The simulation of one-dimensional layered geoelectric model and low-resistance anomalous body model show that it can effectively deal with TEM data,with high inversion accuracy,strong stability and good robustness.(2)Based on the cooperative search strategy of the shuffled frog leaping algorithm(SFLA)and the swarm intelligence idea of the IH-PSO algorithm,this paper proposes a joint algorithm: IHPSO-SFLA.The improved algorithm enhances the local search ability,fully integrates the individual evolution information of each subgroup,and avoids the excessive concentration of particles.The IHPSO-SFLA is utilized to the inversion optimization problem of TEM one-dimensional geoelectric model to explore its operation performance.The outcome demonstrate that the IHPSO-SFLA can have preferable search capability and convergence precision,and is more suitable for dealing with TEM inversion problems.(3)TEM inversion study in the landslide area of Xishan Village,Sichuan Province is carried out.First,wavelet threshold denoising is used to preprocess the measured data,and then the data is qualitatively processed through smoke ring inversion.On this basis,the PSO and IHPSO-SFLA are used to carry out the apparent resistivity profile of the survey line in the survey area.With the best effect,IHPSO-SFLA is selected to perform the stratigraphic configuration in the survey area to obtain a pseudo three-dimensional electrical structure map,and to determine the location of the aquifer,which is consistent with the actual geology and proves the practicability of the algorithm.