Study On Optimization Of Ground Process For Development Of Upper Palaeozoic Gas Pool In Jingbian Gas Field

The Jingbian gas field is located at the junction of Inner Mongolia in Shaanxi and is one of the main producing areas of natural gas.After long-term development,with the gas reservoir resources and stratum energy are rapidly depleted,most gas wells have intermittent production problems.How to improve the production process to ensure stable production in Jingbian gas field and provide reliable energy guarantee for economic development.It has become the first problem to be solved.This paper took the ground technology equipment of an upper palaeozoic gas field in northern Shaanxi as the research object to optimize the ground technology.Based on the site investigation,a neural network prediction model for methanol injection rate was established,Pipe Flow Expert software was used to study the distribution of methanol flow in the pipeline and a feasible improvement plan was proposed.Based on the actual situation of the gas gathering station,the necessity of natural gas throttling heating was analyzed,the operation of the separator was evaluated,and corresponding solutions were proposed.Through laboratory experiments,the effect of condensate oil single composition on TEG?Triethylene Glycol Anhydrous?solution was studied,and a neural network was used to build a prediction model.Finally,based on the research results,the software for analysis and prediction of condensate oil in natural gas was developed.Specific research results included:?1?When establishing a model for predicting the amount of methanol injected,the decision coefficients between the output value and the expected value of the model in the validation stage and test stage were 0.9997 and 0.9959,respectively.The results showed that the output value had a high degree of correlation with the expected value.Thus,the neural network model could accurately predict the methanol injection rate.?2?Jing 99-67H2,Jing 98-66,G7-4,G4-2,and Jing 79-01 wells had an over-injection phenomenon.The total over-injection of methanol was about 57.58 L/h.Jing 99-65,Jing99-66H2,Jing 98-65,Jing 88-7H1,Jing 88-7,and Jing 79-03 wells had an under-injection phenomenon.And their total under-injection of methanol was 24.57 L/h.When the method of changing the displacement and changing the connection mode was adopted,the purpose of injecting alcohol on demand could not be achieved.Finally,a flow distribution device was designed and developed and tested on site with good results.?3?For the Xinbei 5 station,when the natural gas throttling pressure drops to 3.0 MPa,2.0 MPa,and 1.0 MPa,the temperature of the natural gas was 4.908?,7.255?,and 6.863?lower than the critical temperatures for forming the water-hydrocarbon mixture.Therefore,heating was required to prevent the formation of natural gas hydrates.For the Wu 2 station,when the natural gas throttling pressure drops to 5.2 MPa and 4.2 MPa,the temperature of the natural gas was 3.586?and 0.164?higher than the critical temperature of the water-hydrocarbon mixture,and heat treatment was not required.When the pressure was reduced to3.2 MPa and 2.2 MPa,the temperature of natural gas was 3.167?and 5.836?lower than the critical temperature for forming a water-hydrocarbon mixture,and heat treatment was required.?4?For Xinbei 5 and Wu 2 stations,when the gas-liquid separator met the separation requirements?solid particle size was 4?m?,its maximum natural gas processing capacity was lower than the specified processing capacity.Therefore,it was recommended to replace the horizontal high-efficiency coalescence separator for field reconstruction.?5?Through laboratory experiments,the effect of different condensate oil components on the TEG solution was studied,and the effect size of different condensate oil components was analyzed by analysis of variance.For the absorption ability of the TEG solution,the order of influence was:n-C₅>C₉>C₈>C₁₀>i-C₅>C₆>C₇;for the desorption ability of the TEG solution,the effect of C₉ was not significant,and the order of influence wass:i-C₅>n-C₅>C₆>C₈>C₁₀>C₇;For the foaming ability of the TEG solution,the order of its impact size wass:C₉>C₇>C₁₀>n-C₅>i-C₅>C₈;For the defoaming ability of the TEG solution,the order of influence was:C₁₀>C₇>n-C₅>C₆>C₉>i-C₅>C₈.?6?When establishing the prediction model of the influence ability of the condensate oil,the fuzzy neural network was trained by the sample date,and the models could meet the experimental accuracy requirements with a mean square error less than 0.001.In the validation stage,the decision coefficients between the output value and the expected value of the model were 0.9961,0.9938,0.9962,and 0.9951,respectively.In the test stage,the decision coefficients between the output value and the expected value of the model were 0.9983,0.9943,0.9920,and 0.9923,respectively.The results showed that the output value had a high degree of correlation with the expected value.Thus,the neural network model could accurate predict influence ability of the condensate oil.?7?According to the experimental research method,the natural gas condensate oil analysis and prediction software had been developed.This software had the function of predicting influence ability of the condensate oil,variance analysis,and data management for technical personnel's to reference.