Research Into Warning Mechanism Of Dynamic Development In Oil Field Based On Data Mining

Oilfield warning is a problem with many uncertainty factors. There are many factors which have no historical data, so scientific calculations are unable to be realized. The traditional warning methods have many limitations, and therefore, analysis and prediction results are not accurate. The research focus in the current artificial intelligence and database field is data mining technology. It can analyze large quantities of data in database, dig abundant and objective useful knowledge. The data mining technology will be applied in the warning area, which can not only enrich oilfield warning forecasting theory and method, but also provide scientific basis and technical support for oil resources reasonable planning and sustainable development. In order to keep normal production of oilfield, the paper tries to find the change rule of historical production data, find the alarm production mode, and these patterns will be used in early warning in order to guide the oil production effectively.According to the early warning theory, the data mining technology of artificial neural network theory, oil & gas development theory, and oilfield production actual situation, the research and application framework is established in this paper. Secondly, summing up many research, according to the expert advice, several rounds of constant consideration of oilfield development indexes and the corresponding oilfield history, 10 main factors related to oil output are chosen as production early-warning indexes. Finally, according to the actual conditions, oilfield production early-warning index system is established. After several traditional warning models based on data mining technology is analyzed, early warning model based on BP artificial neural network is established. Through training and learning about historical production data and warning degrees interval in the BP artificial neural network early warning model, the potential change rule is obtained, so we can understand the future production trend. According to calculation results of the oil well production warning model, warning pattern affecting oil field production is obtained based on index variation analysis. Simultaneously, these patterns will be applied in developing dynamic warning system, so warning results are more precise, timely warning is achieved. All of these will provide guidance for oil production.