基于RBF神经网络的永磁同步直线电机全局滑模控制

摘要

Considering on unknown nonlinear, parameter pertubations, external disturbances and end effects of permanent magnet linear synchronous motor (PMLSM) system, a global sliding mode control based on recurrent neural network is proposed. The proposed method has global robustness, which makes the control system states on the sliding mode from the beginning. Then, RBF neural network is used to online estimate system uncertainty boundaries, and the saturation function is adapted to attenuates the chatting level. The simulation results show that the proposed control strategy has better effectiveness than traditional ones,such as PID control and traditional SMC.%针对永磁直线同步电机直接驱动伺服系统存在负载扰动、参数时变、非线性摩擦及端部效应等不确定性因素,提出了一种神经网络增益切换全程滑模控制策略.该策略使控制系统一开始就处于滑动模态上,使系统具有全局鲁棒性.其次,通过递归神经网络的在线学习来对不确定界限实时估计,并引入饱和函数,以进一步降低控制中的抖振,改善系统的性能.通过仿真对所提方案与PID控制和传统滑模控制进行对比,验证了所提出的方案有更好的鲁棒性和跟踪能力.