Fine Interpretation Of Coalbed Methane Logging In DJ Block,southern Section Of Jinxi Flexural Fold Belt

At present,coalbed methane(CBM)development object in DJ block is mainly no.5 coal in Shanxi Formation.In order to strengthen the comprehensive geological study of coal reservoir no.5 in DJ block and meet the needs of the policy formulation of CBM stable production technology,it is urgent to carry out fine CBM logging interpretation research based on newly obtained test and analysis data.This paper based on extensive literature investigation and data collected in the study area,combed the CBM detailed interpretation technology present situation and the key technology,carry out the logging data pretreatment,coalbed methane reservoir logging response characteristics,coal structure recognition method,based on the analysis of test data optimization of coal seam is established industrial composition,porosity,permeability and gas content prediction model,Combined with the actual data,the prediction model of rock mechanics parameters of coal roof and floor is established,and finally the fine logging interpretation of coalbed methane reservoir is realized.The results show that the natural gamma-compensated neutron cross plot has the best identification effect for five lithologies.The stratigraphic division and correlation revealed that the thickness of the main target strata in the study area was relatively stable,and the microstructural characteristics of the target strata were high in the southeast and gradually decreased to the northwest.The XGBoost coal structure identification model is superior to the traditional graph method.The results show that the no.5 coal seam in the study area is dominated by class II and III coal,and there is less class I coal.Random forest method,regression analysis method and volume model method were used to predict the industrial components of no.5 coal seam.It was verified that the effect of random forest model was better than the other two methods.The total porosity and fracture porosity of coal and rock were predicted by using variable skeleton density method and double lateral iteration method,and the permeability of coal and rock was calculated based on F-S method and Darcy law method.The accuracy of predicting gas content by neural network method is 96%,which is better than regression analysis method,KIM method and LAN coal rank equation method.The elastic parameters of roof and floor of coal seam were predicted by acoustic time difference method,and the prediction model of roof and floor strength parameters was built based on support vector machine method.