Research On Modeling And Optimization Method For Sucker-rod Pumping Process

Sucker rod pumping is one of the traditional oil production methods in the domestic and foreign oil industry,and it plays an important role in the oil exploitation in our country,which has the advantages of simple structure,high reliability,convenient maintenance,high adaptability for site conditions and so on.Therefore,in the oil fields of China,about 80%of the oil wells which sucker rod pumping system are used in.However,it also has some shortcomings,such as low efficiency,high production cost,and high energy consumption and so on.Therefore,under the premise of ensuring the safe and stable production of the oil production process,pursuing the improvement of liquid production,the reduction of energy consumption and production cost,and the improvement of economic efficiency has become an urgent problem to be solved.In another words,the above requirements can be realized by the optimal control of oil production process.This dissertation is on the background of actual production process of sucker rod pump,which is in an oil production working area of Liaohe oil field.Aiming at the difficulties in online measuring and optimization of comprehensive production indexes,including the overall liquid production and energy consumption for per ton liquid production,basing on depth analysis of the characteristics of the sucker-rod pumping process,this dissertation carried out the research on the modeling and optimization of the relevant production indexes,which can realize the optimal operation direction of the sucker rod pump.In this dissertation,the main researches are summarized as follows:1.Through the analysis of the main factors affecting the overall oil production process,combined with the characteristics of sucker-rod pumping process,and on the basis of the research in previous literatures,the mechanism model of related production index of sucker-rod pumping process is established.Besides,the unknown parameters are identified and determined.The model lays the foundation for the establishment of hybrid model.2.Aiming at the large error when the mechanism model is applied in the industrial field,a hybrid modeling method for sucker-rod pumping process is proposed based on parallel structure.The method is composed of mechanism model and least square support vector machine(LS-SVM)error compensation model.The mechanism model is used to describe the overall characteristics of oil production process,and LS-SVM error compensation model is used to compensate the error which can't be established by model.The advantages of the two models are adequately adopted by hybrid model,and the advantages of the mechanism model and the data model are complementary,which improve the accuracy and the generalization ability of model effectively.The validity of the hybrid model is verified by simulation experiments.3.In order to adapt to the requirements of changes in the actual process,an adaptive correction strategy of hybrid model was proposed.Firstly,the error distribution characteristics of hybrid model output of oil production process are described by the Gauss Mixture Model(GMM),and the statistic of error distribution is constructed and used to evaluate the performance of hybrid model.Then,the model is corrected based on the evaluation results of model,which mainly includes the short-term correction and the long-term correction.Finally,the actual production data is used to verify the proposed hybrid model and the model correction strategy,the obtained results are satisfactory.It can effectively improve the model accuracy and the ability of adapting to different production conditions,which lays a model foundation for the optimization and control of sucker-rod pumping process.4.Due to the optimization for oil production process is a multi-objective optimization problem,multi-objective optimization algorithm is studied in depth.Basing on non-dominated sorting genetic algorithm with elitist strategy,that is,non-dominated sorting genetic algorithm II(NSGA-II),a chaotic non-dominated sorting genetic algorithm with hybrid gradient operator(IG-NSGA-II)is proposed combined with the ideas of local search and chaos theory,which introduces some strategies,including the hybrid chaotic mapping model for initializing population,the hybrid operator based on gradient and the substitution operation of chaotic population candidate.Simulation experiments on some standard test functions show that the proposed algorithm has better convergence,diversity and distribution uniformity,and is suitable for solving complex multi-objective optimization problems.5.At present,there are some common problems in sucker-rod pumping process,such as low oil production,high energy consumption and so on.Therefore,the optimization objectives for oil production process are maximizing the overall liquid production and minimizing energy consumption for per ton liquid production in the block.Besides,the decision variables and constraints are analyzed.Then the proposed hybrid model with adaptive correction strategy is used to describe oil production process.Finally,the proposed IG-NSGA-II is used to solve the multi-objective optimization model of oil production process.The results verify the effectiveness of the proposed optimization strategy.