Study On Logging Evaluation Method Of Tight Reservoir Effectiveness In Block X Of Jiyang Depression

With the further deepening of exploration,the exploration and development of conventional oil and gas reservoirs has come to an end.The exploration and exploitation of new oil and gas reservoirs such as tight oil and shale gas has become a key research area of major oilfields in recent years.The reservoirs of the fourth member of the Shahejie Formation in the Bohai Bay Basin are typical deep tight sandstone reservoirs with complex four-sex relationships,versatile lithology,diverse composition,low porosity and low permeability,complex pore structure and strong heterogeneity,which bring great challenges to reservoir effectiveness evaluation.Tight reservoir rock mass identification,parameter interpretation,and effectiveness evaluation are all difficult to explain in current reservoir evaluation.It is necessary to establish new theories and methods to evaluate according to actual conditions.In this paper,the tight sandstone reservoirs of the fourth member of the Shahejie Formation in the X block of the Jiyang Depression are the main research objects,and the characteristics of the reservoir lithology,physical properties,oil and electrical properties and their interrelationships are comprehensively analyzed;Based on the analysis of lithology sensitivity curves,combined with core physical properties,logging and thin section analysis,BP neural network and convolutional neural network are used to identify the lithology of tight reservoirs.Based on the idea of core calibration conventional logging,fine reservoir parameters are modeled by classified modeling method.The pore and pore canals types in the study area are described by scanning electron microscopy and casting thin section.The pore structure is classified by mercury injection data.The lithology,physical property and oil-bearing lower limit of reservoir are discussed.Finally,the evaluation criteria of tight reservoir effectiveness are established by synthesizing lithology,physical property and pore structure.The results show that the reservoir component maturity and textural maturity of the study area are low,and the particle size distribution is wide.It belongs to low porosity and low permeability reservoirs.The reservoir lithology in the study area is fine sandstone and siltstone,and the oil quality is good.Electrical characteristics are complex.The BP neural network method was used to establish the lithology discriminant model.The training accuracy was84.3%,the test accuracy was 81.1%,and the convolutional neural network model training accuracy was 98.42%,and the regression accuracy was 97.89%.The convolutional neural network model has high precision,fast convergence and real-time performance.The logging phase analysis method is used to divide the stratum into four types of petrophysical facies,and the reservoir parameter model is established by classification.The calculation effect is better.According to the mercury intrusion curve,the pore structure is divided into four categories.The lower limit of effective reservoir lithology is gray sandstone.The lower limit of oiliness is oil trace.The lower limit of reservoir porosity is 10.8%,and the lower limit of permeability is 0.5x10^-3mm².Finally,based on the classification and evaluation of petrophysical facies and pore structure,combined with the test oil test data to establish a comprehensive evaluation standard of reservoir effectiveness,the application effect is better.