| Time domain induced polarization method(TDIP)is an electrical exploration method based on induced polarization effect of rocks and ores to solve related geological problems.It has played an important role in the geological exploration work in the fields of metal mining,energy,hydrology,engineering and environment.However,the cause of IP effect observed in actual work is complex.The results of a large number of mineral IP exploration show that the high polarization anomaly is reflected by both metal deposits and graphitization or pyrite formations.In addition,engineering,hydrological and environmental exploration work contains a large number of weak polarization anomalies that are difficult to identify and distinguish.These IP interference factors make TDIP have limitations in practical application.A large number of relevant studies show that the polarization anomaly can be well identified and resolved by the time spectrum characteristic parameters of different states of the polarization body.In order to comprehensively and accurately identify and evaluate abnormal bodies,researchers further studied and applied the IP spectrum characteristics of polarization bodies,that is,different polarization bodies have different time spectrum characteristics of different states.Time domain spectral induced polarization method(referred to as time spectral induced polarization method)is the main development direction of TDIP,which is characterized by observing the time-varying resistivity spectrum of the earth(or the attenuation curve of the induced field)and analyzing the difference of time spectral characteristics to carry out exploration work.The observed response characteristics are highly related to the rock structure,stratigraphic structure and rock composition.By identifying different types of time spectrum parameters,we can obtain rich geological information,and then conduct structural analysis and lithologic identification.At present,the time spectrum analysis methods used for polarization anomaly identification mainly include half-life,deviation degree,characteristic parameter inversion based on Cole Cole model,etc.However,the application of half life and deviation degree is based on the assumption that the polarizability or secondary field changes linearly,which is difficult to accurately reflect the characteristics of time spectrum.The parameter inversion of Cole Cole model is usually based on the optimization method.Because of the multiple solutions of the optimization method itself and the similarity of time spectrum attenuation characteristics,its solution is not stable.The time spectrum of induced electric field is composed of the polarizability of different observation times.Although the data of multiple time traces have universal similarity,they still have individual characteristics,so they have certain statistical laws.Therefore,the related multivariate statistical analysis method has the potential to statistically classify such problems.K-L transform is an orthogonal transform,which is applicable to any probability density function,can be used to eliminate the correlation between components,and effectively compress information through the eigenvalues and eigenvectors of the covariance matrix.Based on K-L transform,multiple observation data in the same observation sequence of time spectrum can be described,which can avoid the problem of strong multi solution of inversion.This paper aims to identify and distinguish the time-varying states of different IP effects,and regards the time spectrum of IP field as the pattern to be identified in statistics.The K-L transform compression principle is used to establish the analysis method of time spectrum of IP field.This analysis method aims at the IP effect samples at different times and changes the state mode in time.By reducing the number of features of the IP time spectrum,that is,reducing the feature space dimension of the IP time spectrum,the statistical analysis results of the IP time spectrum are obtained to identify and distinguish different types of polarization anomalies.According to the above methods and theories,a practical program was compiled based on Matlab language,and the theoretical data were calculated and analyzed to verify the feasibility,applicability and stability of the method.The measured data obtained from metal mine and hydrological IP exploration were further processed,interpreted and analyzed.The results show that the time-spectrum analysis method based on K-L transform is feasible,which can eliminate the correlation between time-spectrum pattern features,highlight their differences,and realize the time-spectrum identification of polarization anomalies.The results obtained in this paper can provide reference for finding polarization anomaly bodies,provide technical support for exploration,and provide a more convenient method for analyzing IP time spectrum. |
