电磁离合器电流的神经网络整定PID控制  

PID control on the current of electromagnetic clutch tuned by neural network

作　　者：吴晓刚[1] 王旭东[1] 余腾伟[1] 谢先平[1] 

机构地区：[1]哈尔滨理工大学电气与电子工程学院,黑龙江哈尔滨150040

出　　处：《电机与控制学报》2007年第4期335-339,共5页Electric Machines and Control

基　　金：黑龙江省科技攻关项目(GC04A519)

摘　　要：针对电磁离合器驱动电路存在的非线性,及汽车运行时复杂环境使其应用传统的PID控制难以在控制参数整定上达到最优的问题,依据神经网络收敛速度快,全局逼近能力强的优点,提出了基于径向基函数(Radial Basis Function)神经网络整定PID控制电磁离合器电流的方法,在保留传统PID控制优点的同时,利用RBF神经网络对PID控制参数进行在线整定。仿真与试验结果证明,基于该方法驱动控制的电磁离合器电流动态效果与跟踪效果较好,抗干扰能力好于传统的PID控制,系统具有较好的自适应性。The driver circuit of the electromagnetism clutch has some nonlinearity, and the running environment of vehicles is complex, so the traditional PID （proportioual-integral-derivative） control strategy couldn＇ t work at its best on parameter tuning. The Neural Network （NN） is global optimum and has best approximation performance, as well as fast convergence speed. Accordingly a PID tuning strategy to control the current of the clutch based on the radial basis function Neural Networks （ RBF NN） is proposed in this paper, which keeps the advantages of traditional PID controllers. In addition, parameters of PID controller can be tuned on-line. Simulation and experiments prove that the dynamic and tracing performance of electromagnetic clutch driver is good, the anti-jamming ability is better than that of traditional PID control strategy, and the system has nice adaptability.

关 键 词：RBF神经网络 辨识 PID控制 电磁离合器 电流跟踪 

分 类 号：TM46[电气工程—电器]