The Fundamental Research Of Array Azimuthal Laterolog

In the modern petroleum logging,the complex geological conditions of heterogeneous formation and horizontal wells call for new development in well logging instrument and data processing method.The array azimuthal laterolog inherits the multi-measurement superiority from array laterolog;in addition,it also possesses the good azimuthal directivity and can image the formation with azimuthal resistivity.Based on the principal theory of laterolog,this paper studies the basic key technology of array azimuthal laterolog inspired by the struction of traditional laterolog tool,mainly using the finite element numerical simulation method.First,based on the research of the development and application of laterolog tools at home and abroad,this paper designs and optimizes the structure and size of electrode system,and establishes the working mechanism and digital focusing method of the new electrode system using the method of the orthogonal design and the optimization of objective function by the finite element method,furthermore,researches and discusses the basic detection feature,including the borehole effect,radial investigation depth,vertical resolution,azimuthal resolution,etc.The results show that the measurement accuracy of the electrode system is strongly affected by the location and the length of the insulation parts in the apparatus,and the best size assembly can be designed using the method of objective function optimization.The vertical resolution of the electrode system is up to 0.3m and the maximum investigation depth is 1.23 m.The undesirable effect from borehole and surrounding formation can be declined and the resistivity of the formation surrounding the borehole can be measured quantitatively when the sidewall measurement style is adopted.The structural reasonability of the electrode system is verified by the simulated result in the standard Oklahoma formation,laying the foundation for further data processing and the research of azimuthal resistivity imaging.Aiming at the five common detection models,first,a mass of multiple formation-parameter forward calculation has been implemented using the finite element method,as a result,the multi-dimension apparent resistivity data packages of the five common detection models have been established that are the data foundation for the fast inversion.Then a multi-dimension cubic spline interpolation algorithm is used to calculate the apparent resistivity corresponding the given formation parameters.This method can replace the traditional straight-laced forward calculation based on the numerical simulation and realize the fast calculation.Furthermore,a surface fitting method based on the least square method is adopted to fit the scatter data in the multi-dimension apparent resistivity data packages and to narrow the needed data space,which can reduce the required time for data loading.A simplex search method combined with the multi-dimension cubic spline interpolation calculation is used to carry out the parameter inversion for 1D,2D formation and Oklahoma formation,and the results verify the feasibility of the inversion method adopted here.Finally,the support vector regression optimized by the genetic algorithm is used to operate the parameter inversion that can reduce the dependence on the quantity of the prepared forward,which can guide the formation parameter calculation using the machine learning algorithm.The next task is to study the response of the array azimuthal laterolog in the horizontal well and the azimuthal resistivity is mainly focused.First,the mud invasion process in horizontal well is carried out based on the finite difference method,and the results show that the mud invasion shape is obvious asymmetric especially in the high permeable formation and permeability anisotropic formation.Based on the simulated results,constriction factor and mobile factor are introduced to describe the mud invasion shape in horizontal well.The responses of array azimuthal laterolog to the asymmetric mud invasion are studied.The results show that the array azimuthal laterolog inherits the common response of traditional array laterolog,and offers twelve azimuthal resistivity characterizing the asymmetric resistivity distribution around the horizontal well.A primary 1D inversion is implemented to depict the asymmetric invasion profile that further verifies the sonde's ability to heterogeneous formation.Finally,the boundary responses to of this sonde in two-layer and three-layer stratum are researched and correction charts for layer thickness are manufactured.All this work can provide guidance for the use of this sonde in the horizontal well.Finally,the imaging processing of azimuthal resistivity is studied using the measured responses of this sonde in the deviated formation.The twelve azimuthal resistivity curves offered by this sonde can describe the dip angle and direction.In the azimuthal resistivity imaging processing,an edge-directed interpolation method is introduced here to increase the pixel points and improve the imaging continuity and resolution.The weighted image fusion method is used here to remove the step phenomenon occurring in the dynamic image processing.The grey level grouping method is adopted to realize the enhance enhancement processing of the azimuthal resistivity imaging,compared with the traditional histogram specification,and improves the imaging contrast,which lay the foundation for further application.Furthermore,an imaging example of the sonde's five detection models is exhibited that demonstrates the uniformity of each detection model.