关键词: 游梁式抽油机, 智能控制器, 模糊神经网络, 模糊Takagi-Sugeno模型, 模糊专家数据库

Abstract: There are the problems of "pump emptiness" and "evacuation" in rocker pumping unit during petroleum exploitation of the low-production well.So,a new RL fuzzy neyral network model for controlling the interim operation of pumping unit was presented.This scheme can decrease the backlash of pumping unit and save much electrical energy.The control system possesses the favorable self-adaptability.With the change of environment,it can regulate the parameters and structure of itself and determine the reasonable idle hour according to the sampled data and fuzzy-expert database.The controller of pumping unit based on the model has been successfully used in the oil recovery plants of Daqing,Shengli,Liaohe,Zhongyuan and Dagang Oil Fields. Experiments showed that the works of controllers and the start-stop time of pumping unit were reasonable.The energy-saving rate ran up to 30 percent with the precondition of oil output equality.The automatic and intelligent control of well pumping unit may be realized.

Key words: rocker pumping unit, intelligent controller, fuzzy neural network, fuzzy Takagi-Sugeno model, fuzzy-expert database