Study On Log Evaluation Of Rock Mechanics Parameters And Brittleness For Shale Gas Reservoir

In the brittleness index evaluation for the shale gas reservoir,due to the complexity of the mineral composition of the shale and being expensive and time-consuming of the rock mechanics experiment,the previous methods based on the brittleness evaluation from the other research areas are often very different from that of the study area.Therefore,it is urgent to develop a method of the brittleness shale gas reservoirs.The thesis takes the Longmaxi Formation and the Wufeng Formation shale of the Jiaoshiba Shale Gas Field in Fuling,Chongqing,as the research object,combines the regional geological background and basic characteristics of the reservoir,and applies mechanical experimental data of rock samples,XRD diffraction data of rocks,and Jiaoye 1 logs,multiple linear regression,and data mining methods to evaluate the relationship between shale gas reservoir brittleness index and logs response.Comparing the brittleness evaluation methods based on the elastic parameter and the mineral component,the deficiencies and defects of the traditional brittleness evaluation were analyzed.Then the quantitative relationship between brittleness index and mineral content,porosity and density were established with data mining techniques and multiple linear regression methods.Python data mining library was used for data exploration and analysis.Based on univariate and combinatorial variables' analysis,the correlations between various factors and brittleness index were found to select suitable features for regression.The results shows that brittleness is the comprehensive reflection of the composition of various minerals and pore structure in shale,and only considering the mineral components could not evaluate the brittleness.The method in the thesis is only based on a limited number of data points for fitting,applies a statistical point of view to carry out brittleness evaluation,and the fitting result only uses the data of Jiaoye 1.It is necessary to further broaden the ideas and combine logging theory.