Structural Analysis Of Porosity And Reservoir Interpretation Evaluation Of Tight Sandstone In Hangjinqi,Ordos

In this thesis,taking the dense sandstone reservoir of the Lower Stone Box Formation in the Jin58 well area of the Hangjinqi Area of ordos Basin as an example,based on the data of core analysis and testing,logging curve and gas test,and on the basis of briefly introducing the geological characteristics of the study area,the four-sex relationship analysis of the reservoir and the identification of reservoir fluid were carried out;the petrophysical experiments such as casting sheet,X-ray diffraction analysis and core nuclear magnetic resonance were integrated to classify and divide and quantitatively evaluate the pore structure of the lower stone box group in the study area;and the calculation model of pore permeability of five reservoirs was compared and analyzed.Including the traditional multiple regression model and four machine learning models,the high-precision pore infiltration calculation model is optimized,and then the longitudinal continuous division of the pore structure of a single well and the comprehensive evaluation of logging are carried out on the reservoir in the study area.The results show that the pore types of the stone box group under the study area are:intrapartronic lysate pores,primary pores and heterozygous lysate pores,and the throat types mainly include: flake,dot,tube bundle and fine tubular;according to the core nuclear magnetism and casting sheet and other data,the reservoir is qualitatively divided into three types of pore structures: class I large pore laryngeal pore structure,class II medium pore fine throat pore structure and class III medium pore micro-throat pore structure,and the reservoir quality index RQI quantitatively evaluates the three types of pore structure,the better the pore structure,The corresponding RQI value is larger;the three-porosity logging is more sensitive to the gas-bearing reservoir,and the curve overlap method can well identify the gas layer;compared with the five porosity calculation methods,the porosity calculation model based on random forest and the permeability calculation model based on extreme gradient enhancement are preferred,and the root mean square error RMSE of the two models in the blind well test is the smallest,namely: 1.038 and 0.184,and the decision coefficient R2 is the largest.0.929 and 0.877 respectively show that the established porosity calculation model has good computational accuracy and applicability.