研究了一种永磁同步电动机单神经元直接模型参考自适应矢量控制系统,控制器采用了新的误差函数,通过梯度下降法对单神经元参数进行在线调整使误差函数值趋于零,电机输出更好地跟踪参考模型输出,分析了系统的稳定性。控制器结构简单,算法易于数字化实现。仿真结果表明控制器能保证电机转速快速、准确地跟随给定速度曲线,并对负载扰动有较强的抑制能力。
A novel single neuron speed controller of permanent magnet synchronous motor drive was presented,which was on-line trained by a gradient descent algorithm based on direct model reference adaptive control(MRAC) and field-oriented control.A new error function was used to ensure motor speed tracking better.The permanent magnet synchronous motor drive was proved to be stable.The controller was very simple and easy to be achieved on the digital computer.The simulation results revealed that proposed method can ensure motor speed tracking quickly and precisely and had good performance of the load disturbance attenuation.