基于神经网络优化的永磁同步直线电机迭代学习位置控制方案

Iterative Learning Position Control Scheme of PMLSM Based on Neural Network

针对开环迭代学习控制的局限性和跟踪轨迹收敛速度慢的问题,提出利用径向基函数网络优化控制律的开闭环迭代学习控制算法。在每次迭代过程中,用神经网络拟合合适的学习增益,并分别采用开环迭代学习、开闭环迭代学习和优化的开闭环迭代学习算法在永磁同步直线电机中进行了位置控制仿真,结果表明:优化后的算法比传统方法具有更快的收敛速度和更小的位置跟踪误差。

Considering the limitation of open-loop iterative learning control and the slow convergence speed of tracking trajectory,an open-closed loop iterative learning control algorithm based on RBF network optimization control law was proposed. In each iteration process,the neural network was adopted to fit a suitable learning gain,and the open loop iterative learning,open-closed loop iterative learning and optimized open-closed loop iterative learning algorithm were employed respectively to simulate the position control in the permanent magnet linear synchronous motor( PMLSM). The results show that the optimized algorithm has faster convergence and smaller position location tracking error than traditional methods.