Research On Gas Productivity Prediction Based On Logs For Post-frac Tight Sandstone Reservoirs In Sulige Area

Tight sandstone gas reservoirs are extremely rich in our country. As far as the provedreserve and technical strength, tight sandstone gas is the most realistic resource which isworth exploration and development in unconventional natural gas field. Tight sandstone gashas become one of the most important supplies of natural gas, and it will continue to be thehighlight of the growth of reserve and production. As to the tight sandstone reservoirs arealways with the characteristics such as low porosity, low permeability, low gas saturation,complex log responses, various factors of fracturing productivity, and so on. Thesecharacteristics bring a series of challenges to the evaluation of tight sandstone gas reservoirs.Correct evaluation of productivity contributes to the implementation of explorationresults and scientificly guides to the rational development of oil and gas fields. So, researchon gas productivity prediction for post-frac tight sandstone reservoirs based on logs isnecessary and urgent. It is the inevitable requirement for logging in the future exploration.Regarding tight sandstone gas reservoirs of P2h and P1s of Permian in easten SuligeArea as the research objectives, this paper studied on the gas productivity prediction forpost-frac tight sandstone reservoirs based on logs.Starting from macro-geological perspective and micro-reservoir characteristics, thispaper researched on the characteristics of production and well logging features of tightsandstone gas reservoirs in Sulige Area, used6methods to determine the lower limit ofporosity and permeability, and determined the lower limit of logging parameters according tothe plate methods, eventually established the lower limit standards of the target reservoirs inthe study area.Discussed the factors of gas productivity in post-frac tight sandstone reservoirs fromtwo aspects, one is the reservoir itself, and the other is fracturing. Studied on therelationships between the parameters and fracturing productivity, used gray relationalanalysis method to calculate the associated degrees between the parameters and the fracturing productivity. discussed the the main factors affecting the post-frac productivity inthe study area by sorting the associated degrees.Through using the method of "core calibrate logging", established the porosity,permeability and water saturation calculation models with hierarchical groups. Based on coreporosity, permeability, saturation parameters, logs and fracturing parameters, used multipleregression analysis to establish fracturing productivity rapid calculation model withhierarchical group in the study area.On the basis of preferred factors of fracturing productivity in the study area, used threemethods established the gas production forecasting models for post-frac reservoirs in EasternSulige Area, they are BP neural network, wavelet neural network and Elman neural network.The models use the reservoir lower limit standards as cutoff values, and calculate theproductivity from point by point, and get the cumulative productivity through accumulatingthe test interval. By discussing the advantages and disadvantages of the three models formthree aspects of principle, training convergence speed and precision, ultimately determinedthe Elman neural network model is the optimal fracturing productivity prediction model inthe study area, achieved the quantitative calculation of fracturing productivity.In the application of Eastern Sulige Area, Elman neural network model successfullypredicted the gas produvtivity for post-frac tight sandstone reservoirs.