Envelope Extraction Method For Surface Nuclear Magnetic Resonance Signal Based On Local Mean Decomposition

As global water consumption continues to rise,regional water scarcity is increasing,leading to serious over-extraction of groundwater,which in turn causes ground subsidence,foundation deformation,and other geological hazards that seriously damage people's lives and property,so it is important to use advanced technology to detect groundwater.Surface Nuclear Magnetic Resonance(SNMR)is an efficient and non-destructive geophysical method to detect groundwater by detecting the resonance jump of hydrogen protons in groundwater,which can quantitatively characterize the groundwater content and the pore size of the water storage medium.However,the SNMR signal obtained with the geomagnetic field as the background field is very weak,only on the order of nanovolts,and in the field practical measurements cannot be shielded from external environmental interference.As a result,the SNMR signal is often drowned in complex environmental noise,leading to poor quality of measurement data and inaccurate interpretation of results,which limits the practical application of the SNMR method.To address this problem,how to achieve effective separation of SNMR signals from noise and realize high-precision extraction of signal envelopes under multiple operating conditions is the technical bottleneck that restricts the development of SNMR methods.Local Mean Decomposition(LMD)is an adaptive processing method for nonlinear non-stationary signals.This method can decompose the signal into a series of single-component AM-FM signals with instantaneous physical meaning,so as to realize the time-frequency analysis of complex signals.Therefore,this paper develops a method based on local mean decomposition to extract the SNMR signal envelope,which has important theoretical significance and practical application value.Firstly,on the basis of fully investigating the SNMR signal and noise characteristics,the basic principle of the LMD algorithm is expounded,and the advantages of the LMD algorithm compared with the Empirical Mode Decomposition(EMD)algorithm are more suitable for SNMR signal processing.The problems such as end effect,small disturbance,modal aliasing and pseudo-product function(PF)components that exist in the data processing of the algorithm are analyzed,and corresponding solutions are given.Then,the simulation study of SNMR envelope extraction based on LMD algorithm is carried out.The LMD algorithm is improved to solve the problem of endpoint effect and difficult to determine the PF component corresponding to the SNMR signal.The simulation results show that for SNMR signals with different noise types,noise intensities and average transverse relaxation times,the initial amplitude fitting error of the envelope extracted by the algorithm is within ±5 %,and the relaxation time fitting error is within ±6 %.The signal-to-noise ratio can be improved by 20?35 d B.In addition,the comparative analysis of LMD algorithm,EMD algorithm and harmonic modeling algorithm is carried out,which further shows the advantages of the algorithm in signal envelope extraction.Finally,the data collection of the field experiment is completed,and the algorithm of the measured signal is carried out.The results show that the improved LMD algorithm can quickly and accurately extract the SNMR signal envelope,and the signalto-noise ratio can be improved by 25?30 d B after extraction,which proves the effectiveness and practicability of the algorithm.The innovative work of this paper is as follows:1.The LMD algorithm is applied to the signal preprocessing in the SNMR field,which can adapt to the complex noise environment,effectively extract the signal envelope,and the initial amplitude fitting error and relaxation time fitting error of the extracted envelope are within ±6 %.The signal-to-noise ratio is improved by 20?35d B,which lays a foundation for the inversion and interpretation of hydrogeological information.2.The LMD algorithm is improved according to the characteristics of the SNMR signal,and the optimal range of the small disturbance of the loop is determined.The occurrence of the overlapping problem is effectively suppressed,and the ability of the LMD algorithm to process SNMR signals is deepened.The improved algorithm is more suitable for application in the field of ground nuclear magnetic resonance.