Experimental Study On Adsorption Characteristics Of Methane In Water-bearing Shale

The abundant shale gas resources have gradually become the focus of oil and gas exploration and development.Different from conventional oil and gas reservoirs,there are a large amount of adsorbed gas in shale gas reservoirs.The production and development depend on hydraulic fracturing and other stimulation measures,resulting in the existence of stagnant water in the reservoir.In addition,the specific shale reservoirs naturally contain water.Therefore,the study of adsorption characteristics of water-bearing shale has important guiding significance for shale gas reservoir reserves assessment and production development.In this paper,for the Wufeng Formation and Longmaxi shale rock samples,the equilibrium moisture absorption experiment in the environment of 25 °C and relative humidity 10%?100% was carried out,and the applicability of 10 kinds of moisture absorption models and the main factors of moisture absorption were investigated.The results showed that the equilibrium water content shows a "three-stage" and "S"-type ascending trend with the increase of relative humidity.The Peleg model has the best characterization effect on the isothermal hygroscopic curve.The water content shows an upward trend of "first fast,then slow" and "staircase" with the increase of equilibrium time.The ultimate equilibrium moisture content is mainly affected by specific surface area,average pore diameter,carbonate and pyrite,and respectively meets significant positive correlation,negative correlation and positive correlation.Based on the optimized design of shale saturated water device,adsorption device and adsorption amount calculation method,the methane adsorption experiment in 30??50°C and 0MPa?20MPa environment was carried out,and the isotherm adsorption curve was analyzed,and the adaptability of Langmuir model was explored.The results showed that water has an inhibitory effect on shale adsorption.With the increase of water saturation,the adsorption amount shows a downward trend of "single slide type"(powder state)or "double slide"(particle state),and the "double slide type" curve has critical water saturation,which is mainly affected by organic matter carbon content,organic matter maturity and pore structure,and satisfies significant negative correlation,negative correlation and positive correlation,respectively.In addition,the adsorption amount curve corresponding to high pressure is higher than the low pressure condition;the Langmuir model is not suitable for adsorption characterization under high water content.Based on BP neural network model improved by trial and error method,genetic algorithm and nuclear density estimation method,the prediction model of water-bearing shale adsorption amount is constructed(the topological structure is “6-10-1”).The model prediction accuracy is higher than the stepwise regression model and the traditional BPNN model,which is suitable for the prediction of water-bearing shale adsorption capacity.Through R-type cluster analysis and orthogonal experiment,it is found that shale adsorption is mainly affected by organic carbon content,clay mineral content,BET specific surface area,water content,temperature and pressure,and they are positively correlated,positively correlated,and positively correlated,Negative correlation,negative correlation and positive correlation,the order of significance is: water content > BET specific surface area > pressure > temperature > clay mineral content > organic carbon content.Finally,an intelligent prediction software for water-bearing shale adsorption is compiled,which provides software support for shale gas reservoir reserve prediction and evaluation.