Research On Noise Reduction Of Transient Electromagnetic Measurement Data Based On Variational Mode Decomposition

Transient electromagnetic method(TEM)is widely used in many fields such as geological exploration,mineral resource exploration and water resources development due to its strong ability to penetrate high-resistance layers,large exploration depth,high resolution efficiency and high work efficiency.In the actual detection process,transient electromagnetic data contains a variety of noise,because the influence of noise will bring false abnormal information in the subsequent data interpretation stage,affect the accuracy of data interpretation,and cause errors in cognition of underground geological structures.Therefore,it is necessary to denoise the measured transient electromagnetic signal before data interpretation work to improve the signal-to-noise ratio of the transient electromagnetic signal.Therefore,this paper proposes to use VMD method to denoise transient electromagnetic data and improve the accuracy of data interpretation.According to the actual parameters of the transient electromagnetic detection system of Jilin University,the theoretical data of the transient electromagnetic signal was simulated,and after analyzing the noise types and characteristics that affect the transient electromagnetic signal,the Gaussian white noise and power frequency noise were added to the theoretical signal.Use VMD to denoise the noisy signal,the time-frequency image,signal-to-noise ratio and mean square error of the signal before and after denoising prove that VMD can effectively eliminate the noise in the TEM signal and achieve the purpose of denoising.The inversion model verifies that the de-noising results can retain effective information,and the inversion results can clearly reflect the underground medium and its structure.Aiming at the problem that the parameters of the VMD method need to be selected based on empirical values,a genetic algorithm(GA)is proposed to find the optimal parameter combination.The basic principle of genetic algorithm is introduced,the steps of GA to optimize VMD are given,the sample entropy function is used as the fitness function of GA,and the principle of sample entropy and its calculation formula are given.The comparison between the optimal parameters found by the GA optimized VMD and the optimal results obtained through experiments shows that the parameter combination found by the GA algorithm can be considered as the optimal VMD parameters.Using GA optimized VMD to process noisy TEM signals can achieve the best denoising effect.Use VMD and GA-VMD to process the measured data respectively,and invert the data before and after the processing respectively to obtain the apparent resistivity depth profile.The practicability of this method is proved by comparing with the geological data given by the survey.