Methods On Intelligent Logging Identification For Lithology Of Terrestrial Shale Oil Reservoir In Jiyang Depression

The exploration potential of shale oil resources in Jiyang depression is great,but its reservoir lithology is complex,resulting in inconsistent lithology division scheme in the study area and difficult logging lithology identification.It is necessary to study the logging identification method of shale oil reservoir lithology,so as to provide support for logging interpretation and subsequent fracturing transformation.Taking the terrestrial shale in Jiyang depression as the research object,comprehensively using the data of core description,thin sections,X-ray diffraction,whole rock mineral analysis,physical properties,electron microscope and electrical imaging,this paper deeply studies the shale lithology classification scheme,lithology logging response characteristics and lithology logging identification methods.(1)Through thin section identification and core description data,the shale lithology is divided into 8 categories in detail from the perspective of structure,organic matter content and rock type.Finally,the shale lithology division scheme combining the triple information of "organic matter + structure + lithology" is determined.(2)Based on the conventional logging curve data,imaging data and element logging data,the response characteristics of logging,imaging and element logging are analyzed and compared respectively,and the characteristic differences are analyzed.Finally,the characteristic differences of eight types of lithology are defined,which lays a foundation for the accurate identification method of lithology and the division of favorable intervals.(3)Build the petrophysical volume model of mineral components,calculate the volume content of shale mineral components in the study area by using the optimal inversion method,and then combine the inverted mineral volume content model with the pattern recognition and classification method of cluster analysis to establish the lithology recognition model according to the inversion results of mineral content,with an accuracy of 78% and good effect.(4)According to the logging curve data,the sensitive logging curves of different lithology are intersected in pairs to identify the lithology,which has a good effect on the identification of gray matter,argillaceous and sandy lithology;According to the fractal self similarity of logging curve,the box dimension of fractal theory is used to analyze the logging curve of terrestrial shale oil reservoir,and the fractal dimension of well section logging curve is calculated one by one.The results show that the fractal dimension optimizes the identification method of lithologic structure;geochemical parameter method can be used to judge the type and content of lithologic organic matter.(5)The identification research is carried out for the triple information of lithology.Using the sensitive logging curve and the intelligent identification method of cyclic neural network,the identification accuracy is more than 80%,and the lithology can be identified quickly and accurately;Finally,the convolution neural network and clustering algorithm are used to automatically classify the electrical imaging images in the study area.The overall recognition effect is about 90%,which provides a reliable method for identifying lithology.Through the establishment of a series of lithology identification methods,the accuracy of terrestrial shale lithology logging identification is effectively improved.The lithology research is closely combined with logging methods and computer algorithms,and its advantages of continuous data information,high resolution and computer automation are brought into full play,which is very beneficial to the division of favorable intervals and "dessert" prediction of shale reservoir by using logging information.