| After our oil exploration in China has undergone a first oil recovery and secondary oil recovery,it has reached the phase of tertiary oil recovery;during this phase of the oil field,the water cut tends to increase and the oil production will decrease accordingly.Under the new situation,the new technology of reducing water cut and increasing oil recovery is very urgent.The ASP flooding is a new technology that can effectively increase oil recovery.It has already begun trials in many large oil fields in China and has also achieved very good practical results.However,as a chemical flooding technology,it has the disadvantages of high cost of the flooding agent,long injection cycle time,complicated model,and slow effectiveness.In order to improve the overall economic benefits of the oil field,we started to study the ASP flooding dynamics optimization in this topic.Taking into account the volatility and uncertainty of crude oil prices,this paper establishes a fuzzy multi-objective optimization model which contains the net present value and crude oil production as objective functions;proposes an improved probabilistic planning algorithm,introduces and solves the switch control variables to achieve reduced cost.For the purpose of cost,the genetic algorithm is used to find the optimal control mining scheme for the ASP.In view of the uncertainty in the actual mining process of ASP flooding and the low accuracy of the solution process,we fully consider the fluctuation of crude oil prices and use fuzzy numbers to represent the price of crude oil;at the same time,we consider that the decision makers have different requirements for crude oil production.A fuzzy multi-objective optimization model with net present value and crude oil production as objective function is established,which contains both the fuzzy coefficient equation and the fuzzy inequality equation.This model is closer to the real crude oil extraction process than the previous traditional ASP flooding model,and is more in line with our exploration needs.For the established ASP fuzzy multi-objective optimization model,we use the probability distribution algorithm based on Gaussian probability distribution to transform the fuzzy multiobjective model into a deterministic single-objective model.The algorithm is improved on the original probability programming algorithm.Because the Gaussian probability distribution has more complex membership functions and can reflect the authenticity of the fuzzy data better,the Gaussian probability distribution is introduced to describe the fuzzy number and performed on some algorithm steps.We tested it through examples of oil and gas exploration and distribution problems and compared it with the most available mining plan to verify its accuracy and effectiveness.Then the genetic algorithm is used to solve the deterministic single-objective model after the simplification processing,and the optimal mining plan for the ASP flooding is obtained.This paper then proposes to add a switch control variable in the fuzzy multi-objective optimization model of the ASP flooding to divide the mining cycle into several periods,and consider adding a switch control in the injection well in order to save electricity and human resources,etc.Then a fully implicit finite difference algorithm is introduced here to approximate treatment the complex state equations,and the improved probability programming algorithm based on Gaussian probability distribution is used to process and solve the improved dynamic programming model.By solving the optimal switch control variables and comparing the previous test results,an optimal exploitation scheme can be obtained and the algorithm verifies the feasibility of the ASP flooding technology based on the switch control.For the problem of dynamic optimization of injection strategy for ASP flooding,this paper is divided into the following chapters to discuss the solution: First,in the third chapter we put forward an improved probability programming algorithm and introduce an oil field mining distribution instance to verify the accuracy and effectiveness of the algorithm;in Chapter 4,a fuzzy multi-objective optimization model for ASP flooding is established and an improved probability programming algorithm is used to simplify the model into a deterministic singleobjective model,and then a genetic algorithm is introduced to solve the problem.In the fifth chapter,we introduce the switch control variables related to the ASP flooding injection concentration,and establish a new dynamic optimization model,use the probabilistic programming algorithm and the genetic algorithm to tackle the problem.Through the comparison of the simulation results,the final ASP flooding optimal scheme can be obtained. |
PDF Full Text Request