Study On Response And Data Processing Of Horizontal Well Array Laterolog

With the increasing difficulty of oil exploration and production,the horizontal well logging technology has been greatly improved,and has been widely used in various kinds of oil and gas exploration and development.In the complex borehole and formation environment,new requirements for logging instruments are put forward.Based on the basic theory of array laterolog,this paper studies the response of array laterolog and data processing in horizontal Wells.The main research achievements are divided into the following three parts:In the first part,the basic principle and response model of array laterolog are studied.Based on the development of array laterolog technology and the literature investigation of horizontal well logging data processing methods,the structure,working principle,focusing conditions and current focusing methods of array laterolog instruments are determined.The formation model and mathematical model were established.Based on the basic idea of finite element,COMSOL multiphysical field simulation software was used to establish the formation model of horizontal well array laterolog in different formation environments,and the establishment of geometric model,parameter setting,grid division and data processing methods were studied.Finally,the rationality of the model was verified to prove the accuracy of the numerical simulation results.In the second part,based on the working principle of array laterolog electrode system,the forward response characteristics of horizontal well array laterolog are studied by using threedimensional finite element numerical simulation method.The forward response database was established to study the influence of formation environmental factors such as borehole,surrounding rock,mud intrusion,resistivity anisotropy and borehole tilt angles on logging responses under five operating modes of electrode system in horizontal wells.Plot the influence curve of formation parameters and analyze the sensitivity of parameters.The results show that under the five operating modes,tool eccentricity,layer thickness,wall rock resistivity,invasion radius,intrusion zone resistivity,borehole tilt angles and resistivity anisotropy coefficient all affect the logging response results,which lays a foundation for the subsequent parameter inversion research.The third part studies the processing method of array laterolog data in horizontal well.The inversion database was established under the influence of multiple parameters such as borehole,surrounding rock,mud intrusion,resistivity anisotropy and dip Angle.Combining with the basic theory of inversion,the inversion method of BP neural network and support vector regression(SVR)is studied,and all kinds of models are trained.The BP neural network and SVR regression model were used to quickly invert parameters such as intrusion depth,invaded zone resistivity,formation resistivity and anisotropy coefficient in horizontal Wells,and the inversion results were evaluated.The results show that the BP neural network has strong applicability in the processing of horizontal well array laterolog data,and the determination coefficient of the inversion value can reach more than 97%,which is of high efficiency and can meet the requirements of engineering application.In the inversion of small sample size,the accuracy of the SVR model is better,but the effect of the inversion of the invasion depth is poor,and the running time is longer than that of BP neural network.When the sample size increases,the accuracy of the electrical and length parameters of the SVR model is poor.Therefore,the effect of BP neural network is obviously superior to SVR model in the study of array laterolog inversion method in horizontal Wells.