Research On Lithology Identification Method Of Drilling Debris Based On XRF Analysis

With the continuous increase in energy demand for social and economic development,the output of petrochemical energy is also rising,and the requirements for oilfield geological exploration technology are also increasing.The division of lithology is one of the important research fields related to formation evaluation,reservoir description,and real-time drilling monitoring,and is also an important basic work in the process of petroleum exploration and development.At the drilling site,geological workers generally distinguish lithology based on experience,which is affected by factors such as personal knowledge and experience.Therefore,it is very important to establish a model for identifying lithology quickly,easily and effectively.In this paper,with the support of the national key research and development program "research and software development of high-resolution airborne gamma spectrum detection method technology"(Project No.: 2017yfc0602105),some cuttings of three wells in the exploration area of Huaqing oilfield are analyzed.XRF cuttings analysis equipment is used to measure the content of main elements in cuttings samples,and the measured cuttings elements are analyzed by mathematical statistics analysis method Through the comparison and analysis of element content data and lithology,the characteristic elements corresponding to each lithology are determined,and then the elements closely related to each lithology are selected as the lithology discrimination parameters by the multiple stepwise regression method.Element contains The quantity is used as the discrimination parameter of mudstone(subdivision).Random sampling is used to extract 50% of each lithology data from 1118 debris samples by segment to establish a discriminant model,and the data of the remaining samples are used to verify the established lithologic discriminant model.By running SPSS25.0 software,the data was analyzed by Fischer discriminant analysis and Bayes discriminant analysis,and three discriminant models of sandstone,mudstone and coal were established.Off-white medium sandstone,gray argillaceous siltstone,light gray fine sandstone,5 discriminative models of gray-brown oil spot fine sandstone and light gray oil trace fine sandstone,4 discriminative models of gray-green mudstone,dark gray mudstone,black carbonaceous mudstone and gray-black mudstone,cross the accuracy of each model verification.The discrimination accuracy rate of sandstone,mudstone and coal is 88.8%;the overall discrimination accuracy rate of sandstone subdivision classification is 82.3%,and the overall discrimination accuracy rate of mudstone subdivision classification is 89.2%.The discrimination effect is good,indicating that the established The discriminant model is accurate and reliable,and this method is feasible.