Solution To The Two-tier Planning Model Of Oilfield Development And Evaluation Of The Planning Scheme

In recent years,China's crude oil demand has steadily increased,while China's oil resources are insufficient.The proven oil reserves per capita are only 1/10 of the global per capita value,which is difficult to meet the needs of economic development.In order to improve the ability to cope with risks,China should not only vigorously develop new technologies and technologies for oil and gas exploration and development,but also use scientific development planning methods to improve the efficiency of oil and gas field development.At present,most of the oilfields in our country have entered the middle and late stages of exploitation.Large input cost,small output and low profit have become the prominent characteristics of these oilfields.Especially with the increase of water cut,the cost of oilfield development rises sharply.How to control water and stabilize oil,increase production and reduce cost is a serious problem facing oilfields at present.On the basis of previous work,this paper puts forward a Group Method of Data Handling method(GMDH)to predict the oil increment(stock)of old wells in old areas,and further establishes a two-tier model of oilfield development planning considering both increment and stock optimization.Then,based on the improved differential evolution algorithm,the solving algorithm of the model is designed.Finally,we improved cloud model and grey clustering method are used to calculate the oil increment(stock).Optimum selection and sequencing of development planning schemes for various oilfields.This paper mainly carries out the following research:1.In view of the difficulty in determining the factors affecting the oil increment index of old wells in old areas and the large random interference,many prediction methods do not have specific analytical formulas.In this paper,GMDH method is used to predict the oil increment of old wells in old areas.Taking the effect prediction of carbon dioxide huff and puff measures in an oil field as an example,this paper establishes a prediction model of oil increment of carbon dioxide huff and puff measures in old wells in old areas.Comparing the results with the support vector machine(SVM)prediction method,it is found that the GMDH method is more effective in predicting the oil increment of wells in old areas,and the analytical expression between the oil increment and the influencing factors of the measures can be obtained.This method lays a foundation for the establishment of a two-level programming model for non-linear oilfield development considering incremental measures.2.Combining with the field practice and optimization method of an oil field in China,a bilevel programming model for oilfield development considering increment and stock optimization is established by adding the increment of old wells into the stock on the basis of the original work.The optimization model of water flooding and heavy oil in the upper and lower layers of the model is a non-linear programming model with workload as decision variables.The optimization model of offshore and chemical flooding in the lower layers is a 0-1 programming model aiming at project optimization.The whole model is a non-linear bilevel programming problem.3.The optimization model established above is a NP-hard problem.In this paper,a heuristic algorithm for solving the model is designed based on the improved differential evolution algorithm.Aiming at the shortcomings of traditional differential evolution algorithm,such as easily falling into local optimum,this paper adjusts the fixed scaling factor and probability crossover factor CR into adaptive parameters.Then,the results of the improved algorithm are compared and analyzed to prove the feasibility of the improved algorithm.Finally,based on the improved differential evolution algorithm,an algorithm for solving the nonlinear bilevel programming model of oilfield development is designed and implemented.Taking the data of an oil field in China as an example,the numerical experiments of the model are carried out.The test results conform to the law of oilfield development and verify the reliability of the model and algorithm.4.Aiming at the disadvantage of traditional cloud model that can not distinguish the advantages and disadvantages of Clouds with very close distances,this paper puts forward a method of optimizing development planning schemes based on improved cloud model and improved grey clustering method,and optimizes and ranks several oilfield development schemes.This method has good application prospects for the comprehensive evaluation of complex systems with stochastic and fuzzy uncertainties.