An intelligent production fluctuation monitoring system for giant oilfield development

To guarantee production stability of oilfields, an intelligent production fluctuation monitoring and early warning system is developed based on methodologies of fuzzy comprehensive evaluation (FCE) and support vector machine (SVM). A novel early warning indicator system established through deep analysis of historical production data from Shengli oilfield is adopted in the FCE model and SVM model. Through performing field test, both the two models are proved to be capable of accurately predicting abnormal production decline of oilfields and the SVM model is proved to be of high prediction accuracy of 94%. Uncertainty analysis of early warning results can be realised in the system thanks to the mutual examination of the two models. This system also integrates modules of data reading, data processing, and result displaying, which facilitates application of it. The production monitoring and early warning system developed in this paper has been successfully applied in Shengli oilfield which is the second largest oilfield in China. It can play important role in ensuring production stability in oilfields, especially for giant oilfields and oilfields in high water-cut development stage.