基于神经网络的永磁同步电机广义预测控制

Generalized Predictive Control for Permanent Magnet Synchronous Motor Based on Neural Network

介绍了一种基于神经网络的永磁同步电机矢量控制系统的广义预测控制方法。通过分析永磁同步电机数学模型,采用带有延时结构的多层前向神经网络作为预测模型,进行非线性广义预测控制。控制算法是基于非线性激励函数的局部线性思想,将预测模型处理成线性和非线性两部分,并用线性预测控制方法求得控制律,简化了计算。仿真结果表明,利用该法建立的永磁同步电机调速系统,具有良好的控制效果。

Based on neural network model a generalized predictive control(GPC)approach for the permanent magnet synchronous motor(PMSM)servo system is proposed.By analyzing the math model of the PMSM,the multiplayer feedforward neural network with a time delay structure is adopted as predictive model,and the nonlinear GPC is done.The predictive model is separated into a linear part and nonlinear part by the control algorithm based on the idea of the local linearization of nonlinear activation functions,and the control law is gotten by using simple linear predictive control methods.The counting is simplified.Simulation results show that this system is effective.