Research On The Intelligent Water Injection Optimization Model And Algorithm For The Classified Oil Layer By Water-drive

Water injection of classified oil layer By water-drive refers to the method of developing oilfield in order to maintain the pressure of the oil layer,realize the high and stable production of the oilfield and improve the effect of oilfield development.It is the main research direction to continue to improve the oil recovery of water driving for the later stage of high water cut and the post productive period.With the oilfield of water driving entering the later stage of high water cut,conditions of oil pool are more complicated,reservoir characteristic and the rule of oil-water movement increasingly complex,water-drive efficiency continues to decline,and the development problems and contradictions are constantly exposed,such as the high water-cut of oil well.Therefore,it has become the urgent requirement in the separate layer injection of water-drive to resolve the conflict between the different layers of oilfield development,achieve the efficient water injection of different layers,and improve the oilfield economics.In view of the above problem,firstly,this paper researches the computing method of different layers to maintain the reasonable injection-production ratio,and it establishes the calculation model of injection-production ratio and oil recovery rate modeled by introducing invalid water injection,with the derivation of equation according to the principle of material balance.Then,it is the work of preparing the data well to filter and preprocess the data by comparing the production data of oil-water well with the numerical simulation data.After that,based on the analysis results of the K-Means clustering algorithm,the calculation model of the oil producing rate based on dynamic data is established.At the same time,the important influence factors of oil producing rate are selected as the input feature with the grey relational analysis.And the improved CART decision tree regression algorithm based on ELM is proposed to establish the prediction model of oil producing rate,and the effectiveness of the algorithm in predicting the oil producing rate is proved by the experimental comparison.Finally,the study establishes the economic benefit evaluation model,which is solved by particle swarm optimization algorithm to get optimal injection-production ratio and oil producing rate,and combining the computing method of the injection water quantity of the interval of the injection well to get the injection water quantity of each interval,so the optimal control scheme of the water-drive injection-production is obtained.Based on the research on the intelligent water injection optimization model and algorithm for the classified oil layer by water-drive,the system on the intelligent water injection optimization for the classified oil layer by water-drive is developed,which is proved to be feasible with taking the actual data as the sample example and has higher application value.