Identification And Prediction Of Lithology And Fracture Of Volcanic Rocks In Jinlong 2 Well Area,Junggar Basin

At present,volcanic reservoirs are attracting widespread attention,which are more complex than other types of reservoirs.The volcanic rocks of Jiamuhe Formation in Jinlong 2 well area of Junggar Basin are complex and changeable lithology,which is difficult to accurately identify by conventional methods.The development of fractures is closely related to volcanic reservoirs,and it is of great significance to accurately identify fractures for the study of volcanic reservoirs.This paper mainly studies the lithology and fracture identification of volcanic reservoirs.Based on the analysis of the geological characteristics of volcanic reservoirs in this area,the log response characteristics of different lithologies are analyzed according to core description,thin section and imaging data.Two sensitive parameters M and N of volcanic lithology are constructed.The characteristic parameters that ultimately identify lithology are GR,DT,RHOB,CNL,RT,RI,M and N.These features are normalized to eliminate the influence of logging instrument.According to logging characteristic parameters and lithology labels,four lithology recognition and prediction models are established by using four different methods in machine learning.By comparing and evaluating different models,the Random Forest lithology recognition model was selected,and the lithology recognition accuracy rate was over 0.9.In the aspect of fracture identification,the response characteristics of conventional logging to fractures are analyzed,and the characteristics of fractures in conventional logging and imaging logging are determined.In addition to conventional neutron,acoustic and density logging,wavelet transform and rock physical parameters are introduced to identify fractures.According to the analysis of the influencing factors of fracture development,the different lithology has a great influence on the degree of fracture development.Therefore,this paper takes the identified lithologic results as characteristic parameters to participate in the establishment of fracture identification model.Finally,the regression model of fracture development degree is determined according to the 14 characteristic parameters screened,and the correlation coefficient is about 0.8.