Research on FNN Energy Saving Control for Light Load Oil Well with Intermittent Oil Extraction

Based on the situation that average load for 80% of all the pumping units is under 30% of the entire load, the intermittent operation scheme for pumping units is required to meet with the need of energy saving. Considering that oil extraction is a complicated process, fuzzy neural network (FNN) is used to ascertain the intermittent time. In this paper: first, an improved Takagi-Sugeno(T-S) fuzzy system is used to compose the antecedent and consequence of FNN. Second, combining characteristics of oil extraction, the rules and structure of FNN are simplified to meet the engineering need. Third, combining the sample data from oil fields and expert knowledge, the samples are achieved. FNNs are trained with the samples. Finally, BP engineering algorithm for FNN is presented. The algorithm is implemented in "Intelligent Pumping Unit Energy Saving Controller", and the energy saving effect is prominent