RBF网络在横向磁场永磁电机调速系统中的应用

An Application of RBF Neural Network ( NN ) on Speed Control System of Transverse Magnetic Field Magnet Motor

针对横向磁场永磁电机的调速系统特点及常规PID控制参数难以精确整定而引起的控制效果不佳,提出一种RBF神经网络整定PID参数的控制方法;该方法通过比较网络输出值与系统输入值,利用RBF神经网络的自学习、自适应能力,实时调整神经元输入权值、中心节点及结点宽参数值,结合网络的整定指标修正PID系统控制参数,改善系统的运行状态;实验中,在额定工作电压110V设定转速80rpm的条件下,实际转速输出为78.8rpm;根据仿真和实验结果,验证该方法具有良好的控制效果。

Due to the difficulty of accurate adjustment for control parameter, frequently, control performance of conventional PID algo- rithm is dissatisfactory. Consider the characteristics and requirements of speed control system for transverse magnetic field magnet motor, a novel control algorithm that utilizing RBF Neural Network （NN） to adjust PID parameter is proposed in this paper. After comparing input and output value of neural network, this newly--elaborated method uses RBF Neural Network which has the capacity of self--learning and self--adapting to precisely adjust neuron input weights and parameters of central point. Meanwhile, networking setting is also applied to cor- rect control parameter so as to improve system performance. In the experimenting we set the rated working voltage ll0V and the rotate speed 80rpm, the actual rotate speed is 78. 8rmp. The result of simulation and experiment shows that this new developmental PID algorithm per- forms favorable control effect.