Well Log Interpretation And Productivity Prediction Of Tight Sandstone Reservoirs Based On Seepage And Conductivity

With the development of economy and science,the demand of oil and gas resources for the daily life and industrial production are increasing,and unconventional oil and gas resources gave attracted more and more attention.As a piece of cake which has the most exploration and development significance in unconventional oil and gas resources,tight sandstone and the effective identification of reservoir,reservoir parameter evaluation and capacity prediction have become a hot problem.At the same time,due to the low reservoir parametersand the complex seepage mechanism,it is difficult to interpret porosity,permeability and water saturation accurately;the existence of low resistivity gas layer also makes the identification of gas and water layer difficult;in the actual production,due to the complex seepage mechanism,it makes the application of the empirical formula from other areas have bad results,and reservoir productivity prediction is faced with many problems and challenges.In view of these hot and difficult problems,this paper has carried on the research and the discussion,the study area is selected in tight sandstone reservoir of Sulige area.From the original data collection and collation to the experimental design and data analysis,and finally based on the relationship between the permeability and conductivity of the tight sandstone we set up the porosity,permeability and water saturation interpretation model.Through the study of gas and water layer identification and productivity prediction,the gas and water distribution in Sulige West area is obtained.Detailed regional data and experimental data are essential for the interpretation and evaluation of tight sandstone reservoirs.From the macroscopic regional geological structure,reservoir genesis and distribution,to reservoir lithology,rock mineral content,and finally to microscopic pore scale,the microstructures of rock,such as pore structure,seepage and conductivity,are experimentally tested.Through the presentation of this multi-scale feature,it lays the foundation for further research and provides effective technical support.The heterogeneity of the tight sandstone reservoir in the Sulige region is very strong,which makes the conventional interpretation method often have problems in the evaluation,and the reservoir parameter results can not meet the logging accuracy requirements.From the mercury intrusion curve,the relative permeability curve and the NMR T₂ distribution data is not difficult to see the complex pore structure and the larger range of physical properties are typical features of the Sulige region.In this paper,the experimental cores are classified into three types according to the mercury intrusion curve.From the first to the third types,the pore structure and physical properties are deteriorated in turn.There are also large differences in the physical parameters between different types.If a unified interpretation parameter is used,it is necessary to produce a large error,which proves the necessity of classification.At the same time,the experimental data show that the micropores are well developed and further research needs to be done at the pore scale.According to the classification of capillary pressure curves,the relative permeability curves and T₂ distribution of the corresponding cores also have a good classification effect,which indicates that there are some relationships between the petrophysical parameters of the cores,which need further study.In addition,if the traditional method is limited,new methods and new models are needed to apply to the Sulige sandstone reservoir.In response to the questions raised,based on the experimental data,this paper builds an effective model to complete the comprehensive logging interpretation and production forecast of the tight sandstone reservoir in the Suligexi area.In order to solve this problem,based on the seepage and conductivity characteristics,the relationship between the existing capillary pressure?PC?and the transverse relaxation time?T₂?and the resistivity index?I?,the fractal theory is used to derive the new T₂-I model.The relationship between the relative permeability?K_r?and the transverse relaxation time?T₂?which named T₂-K_r model is also deduced by the petrophysical data.The validity of the model is verified by the experimental data,and the significance of the relevant parameters and the influence on the model are analyzed and discussed.In order to study the influence of different physical parameters on the conversion model,this paper studies the relative permeability and resistivity increase rate conversion model as the starting point.The results show that the experimental conditions have an effect on the results of the transformation model,butthe result shows that the relative permeabilities are similar.If cores are fragile and not resistant to high temperature or high pressure,it is possible to use resistivity index measured in conventional condition to get the relative permeability which provides a convenient way for experimental measurement and reservoir evaluation.In practical applications,the results show that the pore structure make a great influence on the relationship between wetting phase tortuosity ratio and resistivity index.Therefore,in this study we took the difference of pore structure into account when deciding model parameters by core analysis and the verification results are basically consistent with the laboratory measurement results.At present,the method of interpretation of porosity and permeability in tight sandstone is more mature.After summarized the work of predecessors,the core fitting method and mathematical method are more commonly used,but there are still many problems to be solved.The results show that the GA-SVM?genetic algorithm-support vector machine?is better than the core data fitting method,and the prediction errorsof the porosity permeability and the porosityare small.The calculation of the water saturation of the tight sandstone reservoir has always been difficult.Conventional methods such as the Archie formula,the mathematical method inevitably have some problems in the accuracy.The traditional Archie formula method uses the petrophysical experimental data to obtain a set of static parameters?a,b,m,n?applied to the whole area,and the calculated results of this method got large errors in Sulige region.Similarly,the mathematical method has a similar problem in the prediction of water saturation,and the unified prediction sample is used to establish the prediction model.The average absolute error of the predicted results is more than 5%.The T₂-I model is applied to calculate water saturation,and the resistivity index curve is reconstructed by using the NMR T₂ distribution in the log evaluation to obtain the dynamic change Saturation parameters b and n.Finally,the water saturation is obtained.This paper also improves the way of obtaining the new Three-water model parameters,and obtains the dynamic Archie parameters with the depth change to replace the past empirical values.The average absolute error of these new methods is less than that of the traditional method,which provides a new way for the calculation of the water saturation in tight sandstone reservoirs.The methods of identifying the gas and water layer in tight sandstone reservoirs are plenty.The conventional qualitative methods such as using logging overlap curves to identify the gas layer.The disadvantage of this method is the high negative rate,and it is clearly affected by the shale content.In this paper,the method of PSO-SVM and the random forest method are used to identify the gas-water layer.The random forest method is used in the identification reservoir for the first time,and the methodis fast and accurate.By using T₂-K_r model to identify the gas-water layer.In the practical application,the continuous relative permeability curves are obtained by using NMR logging data.This provides a basis for the identification of gas and water layers in tight sandstone reservoirs.In the end,According to the production data and K_r-I model,the KHK capacity splitting method is improved,and the single layer capacity prediction model is established by PSO-SVM and RBF neural network algorithm.RBF neural network is superior to SPSO-SVM in time and precision.RBF model is used to analyze the production and to obtain the distribution of gas and water in tight sandstone of Sulige region Reservoirs,which is of great significance to the optimization of production wells and production areas.