Abstract: Taking the advantage of the fuzzy-neural-network's superiority in the non-linear system control,a new fuzzy CMAC (cerebellar model articulation controller) neural network model was used to control the loading system in pumping unit.The control structure and algorithm were presented.This model has been applied successfully to a passive loading system of pumping unit with the nonlinear, determinable interference and random disturbance.The simulation results demonstrate that the proposed control method is insensitive to the time-varying external disturbance and uncertainty of parameters for the system and has good generalizing property, adaptability and robustness. It is also suitable to other electricity-hydraulic servo system.

Key words: well pumping unit, fuzzy logic method, cerebellar model articulation controller, neural network, loading control

摘要： 利用模糊神经网络能够在不能获得精确数学模型的非线性系统中达到最优控制的特性,提出了采用模糊小脑模型神经网络(FCMAC)控制器来提高抽油机电液伺服被动加载系统性能的方法,并给出了具体的控制结构和算法。实验结果表明,该控制方法用于被动变载荷加载的跟踪控制具有良好的泛化能力,且对于位移等扰动具有较强的自适应性和鲁棒性

关键词: 抽油机, 模糊逻辑, 小脑模型, 神经网络, 加载控制