基于RBF神经网络的直线磁悬浮同步电动机控制优化

Control Optimization of Linear Maglev Synchronous Motor Based on RBF Neural Network

选择数控机床进给控制系统的励磁直线电机作为研究对象,并根据该系统运行机制构建数学模型,设置误差函数以及RBF神经网络来实现逼近控制,通过自适应律验证了系统运行的稳定性,并开展仿真分析。空载启动下,RBF控制达到了最快的响应速率,经过0.17 s就进入制定悬浮高度处,与PID和SMC控制相比调节效率依次提升42.2%和24.1%。突加负载下,RBF控制悬浮气隙高度下降5.0×10^(-5) m,经过0.065 s恢复到原先状态,相对之前PID和SMC控制,动态降落显著减小,恢复时间也明显缩短。端部扰动下,RBF控制形成基本稳定响应,有助于获得更加稳定的气隙高度,使控制系统对端部效应起到明显抵抗作用。经测试可知:采用本控制策略能够实现系统抗干扰性能的显著提升。

The exciting linear motor of CNC machine tool feed control system is selected as the research object to construct the mathematical model according to the operation mechanism of the system.Error function and RBF neural network are set to realize the approximation control,the operating stability of the system is verified by the adaptive law,and the simulation analysis is carried out.Under no-load starting,RBF control reaches the fastest response rate and enters the designated suspension height after 0.17 s.Compared with PID control and SMC control,the adjustment efficiency increases by 42.2%and 24.1%respectively.Under sudden loading,the suspension air gap height under RBF control decreases by 5.0×10^(-5) m,and recovers to the original state after 0.065 s.Compared with the previous PID and SMC control,the dynamic landing is significantly reduced,and the recovery time is evidently shortened as well.Under end disturbance,RBF control forms a basically stable response,which helps to obtain a more stable air gap height,so that the control system can significantly resist the end effect.The test shows that the anti-interference performance of the system can be remarkablely improved by using the control strategy.