The Data Mining Modeling Method And Application Technology Research For The Oilfield Development Evaluation

With the development of China’s main oil fields gradually entered the middle or later stage, the distribution of oil and gas water in the reservoir and the production of injectionproduction are very complicated. This topic researched data mining methods and application technology which based on the intelligent models. Using the geological development data of oil blocks, this topic adopted data mining technology and pointed at the typical problems in the oilfield development evaluation such as diagnostic mining and predictive mining.Based on cutting-edge technology of data mining and the typical problems of the reservoir development evaluation, the paper set up the oilfield geological development data source for typical data mining problems. For oilfield development index prediction problem, this paper built RBFPNN based on discrete Walsh transform mining model. Then used the super linear convergence algorithm for the optimization of model training, and improved the learning properties of mining model to increase the convergence speed and forecast accuracy. Aiming at the problems of low efficient wells’ diagnosis in oilfield development, this topic put forward a classification mining model based on PSVM. The model can directly take oilfield geological parameters and development timely data as the input. Meanwhile, using genetic algorithm in the parameters optimization of the PSVM mining model improved the performance of the mining model and had a good applicability in practical application. In the study of applied technology, this paper combined the mining model and the typical problems in oilfield development and evaluation to establish a system framework weakly coupled by data organization, evaluation index system, mining model and other functional modules. It realized the typical problems such as diagnostic mining and predictive mining in the oilfield development evaluation according to the actual geological development data.The subject developed the software prototype system on the basis of mining model and algorithm for oilfield development and evaluation. And it used two typical problems, oilfield development index prediction and low efficient wells, to conduct the actual data processing, achieving better application results. The results of the research have some theoretical and practical value for the oilfield development of diagnostic problems and the predictive problems and the data mining modeling.