Study On Logging Evaluation And Productivity Prediction Methods Of Tight Sandstone Gas Reservoirs

With the continuous deepening of exploration and development,tight gas resources have become one of the key areas of unconventional oil and gas exploration.Tight sandstone gas reservoirs are characterized by complex pore structures,strong heterogeneity,and unclear seepage mechanism,which pose challenges to reservoir logging evaluation and productivity prediction.This article takes the tight sandstone gas reservoirs of the Shaximiao Formation in the central Sichuan Basin as the research object,and focuses on key scientific issues in the logging evaluation process.By integrating data from petrophysical experiments,logging,and production testing,in-depth research is conducted on the petrophysical parameters evaluation,pore structure evaluation,reservoir classification,and productivity prediction of tight gas reservoirs.The reservoir characteristics of the Shaximiao Formation are further researched through petrophysical experimental data.Aiming at the difficulty of evaluating reservoir parameter through logging,the porosity and permeability prediction method has been developed based on principal component analysis（PCA）and Gaussian process regression（GPR）.The logging data after dimensionality reduction using PCA is used as input,and a porosity and permeability prediction model is established using GPR.The relative error with core porosity is less than 8%,and the relative error with core permeability is less than 30%.The results indicate that the porosity and permeability model based on PCA-GPR has high prediction accuracy and strong generalization ability.The key reservoir parameters such as mud content and saturation are obtained using the classical interpretation model by combining the test and analysis data.The accurate interpretation of reservoir parameters lay a good foundation for subsequent well logging evaluation.Combining conventional logging and new logging techniques（nuclear magnetic resonance（NMR）logging and array acoustic logging）,the fluid properties in the Shaximiao Formation are qualitatively identified by using differential spectrum method,density-NMR porosity method,P-and S-wave velocity ratio method,respectively,with an interpretation accuracy rate of over 85%.Through the analysis of NMR experimental data,the influencing factors of reservoir movable fluid are clarified.The results show that the movability of reservoir fluids is comprehensively influenced by the macroscopic petrophysical properties and microscopic pore structure.By conducting gas-water phase permeability experiments,the gas-water seepage mechanism in tight sandstone reservoirs is further researched.Based on the transverse relaxation time（T₂）distribution,a prediction method for gas-water relative permeability is proposed for tight sandstone reservoirs with complex pore structures.The effectiveness and applicability of the method are verified through relative permeability experiments and oil testing data,which achieve quantitative evaluation of seepage parameters in tight reservoirs.Petrophysical experiments are carried out to clarify the pore structure characterization of the Shaximiao Formation reservoir.Two methods for evaluating and classifying the pore structure of tight sandstone reservoirs are proposed.On the basis of capillary pressure curve and bimodal Weibull distribution function,the pore structure characterization model of tight sandstone is established.The correlation between the bimodal Weibull distribution parameters（weights,means and standard deviations）and petrophysical properties and pore structure parameters is analyzed,and then the classification criteria are established of tight sandstone reservoirs in the Shaximiao Formation.This method realizes the effective classification of tight sandstone reservoirs.NMR technology can non-destructive and quantitatively characterize the pore characteristics of unconventional reservoirs,and can continuously provide information on the pore structure of underground formations.The three-peak Gaussian function is used to fit NMR T₂ distribution,and the parameters that can characterize pore structure and have petrophysical significance are obtained.Taking into account porosity and permeability,a model for classification of tight sandstone reservoirs is established using the naive Bayesian classification method based on hierarchical clustering.The accuracy of model is verified by k-fold cross validation,with an average accuracy of 96.67%.Finally,the effectiveness of the proposed method is verified using processed actual NMR logging data and oil testing data.The method does not require conversion of the T₂ distribution into a capillary pressure curve and is capable of fine-grained evaluation of tight sandstone reservoirs.Based on the characteristics of tight sandstone gas reservoirs in the Shaximiao Formation,a productivity prediction model is established based on the principle of plane radial flow.Through actual logging data and oil testing conclusions,it is verified that the model has good application effectiveness,but it requires more parameters and has strong regional empiricism.Therefore,this article combines petrophysical,pore structure,and logging parameters to analyze the production influencing factors through crossplots.At the same time,the importance of production influencing factors is analyzed based on the feature selection methods,and the main controlling influences of capacity in tight sandstone reservoirs are preferred by integrating multiple methods.A method for predicting production is proposed.Combined with the production demand,the reservoir production classification standard is established based on the unimpeded flow rate.The preferred production main-controlling factors are taken as inputs,and the production category prediction model is established using the naive Bayesian method to qualitatively evaluate the reservoir production category.The productivity prediction model is established by e Xtreme Gradient Boosting（XGBoost）method to quantitatively predict the production of tight gas reservoirs.The error between the productivity prediction results and the field gas testing results is 12.17%,which is relatively small.The results indicate that the productivity prediction model based on XGBoost has achieved good application results for tight sandstone reservoirs in the Shaximiao Formation,and it provides a methodological basis for the productivity prediction of tight sandstone reservoirs.