单神经元自适应PID控制器在主动磁悬浮系统中的应用研究  

The Application Research of Single Neuron Adaptive PID Controller to Active Magnetic Levitation System

作　　者：郑仲桥[1,2] 

机构地区：[1]常州工学院电子信息与电气工程学院,江苏常州213002 [2]上海大学机电工程与自动化学院,上海200072

出　　处：《机械设计与制造》2014年第4期268-271,共4页Machinery Design & Manufacture

摘　　要：针对主动磁悬浮控制系统的非线性、开环不稳定性、参数不确定等特点,设计了单神经元自适应PID控制器。以主动磁悬浮系统中的转子实际位置和平衡位置之间的偏差作为神经网络控制器的一个输入信号,引入有监督的Hebb学习规则,推导出Kp、Ki和Kd与加权系数之间的关系。通过对加权系数的调整来实现自适应,控制器的输出信号调节电磁铁线圈中电流的大小,从而调整了转子的位置。仿真结果表明,单神经元自适应PID控制器能极大地改善系统的稳态性能和动态性能,具有较强的抗干扰能力和鲁棒性。A kind of single neuron adaptive PID controller is designed in response to nonlinearity, open-loop instability and uncertainty of active magnetic levitation control system. Introducing supervised Hebb learning rule, the relations between Kp, Ki, Kd and weighting coefficients is deduced, using the deviation between actual position and equilibrium of the rotor in active magnetic levitation system as a input signal of neural network controller. The output signal of the controller tunes the coil current of electromagnet to modulate the position of the rotor by adjusting weighting coefficients to realize adaptive control The simulation results show single neuron adaptive PID controller can effectively improve static and dynamic performanee of the system, and has noise immunity and robustness.

关 键 词：主动磁悬浮 非线性 单神经元 自适应PID控制器 鲁棒性 

分 类 号：TH16[机械工程—机械制造及自动化] TP273[自动化与计算机技术—检测技术与自动化装置]