无刷直流电机神经网络自校正控制研究

Research on Neural Network Self-Tuning Control for Brushless DC Motor

针对无刷直流电机的非线性、强耦合、变参数等特点,利用多层神经网络对非线性函数的任意精度拟合性,提出了基于BP神经网络的无刷直流电机神经网络自校正控制。神经网络自校正控制不需要被控对象的准确数学模型,并可以通过在线学习来适应系统工作环境和系统本身参数的变化,给出了控制系统软硬件的具体实现方案。仿真和实验结果表明,神经网络自校正控制的速度跟随性、动态响应过程较之PI控制更好,具有很好的控制效果。

For nonlinearity, strong-coupling and variable parameters of brushless DC motor （ BLDCM）, neural network self-tuning control （NSTC） is employed to control brushless DC motor applying the neural network＇ s capability of approximating any nonlinear function to arbitrary accuracy. NSTC does not need the plant＇ s accurate mathematic model, and can adapt the change of operational environment and self parameters by self learning on line. The hardware and software scheme of the system is introduced in detail. The results of simulation and experiment indicate that dynamic response process and speed following of NSTC are better then ones of PI control.