基于神经网络的高性能直线电机伺服系统分数阶滑模控制  

Neural-network Based Fractional Order Sliding Mode Control for Linear Motors Servo System

作　　者：张涛 唐传胜[2,3] 李冠甲 

机构地区：[1]平高集团有限公司,河南平顶山467001 [2]南阳理工学院,河南南阳473004 [3]电子科技大学,成都611731

出　　处：《微电机》2016年第8期50-53,共4页Micromotors

基　　金：国家火炬计划产业化示范项目(2014GH041090)

摘　　要：永磁同步直线电机伺服系统是一个强耦合的复杂非线性系统,该系统具有时变性和对外部负载变化敏感等特点,因此给系统的控制带来了难度。针对永上述问题,结合分数阶系统理论和滑模控制理论,提出了一种新型的永磁同步直线电机分数阶全局滑模智能控制方法。该方法在控制作用时系统便处于分数阶滑模面上,以保证系统的全局鲁棒性。通过RBF网络实现系统不确定界限的在线估计与补偿。同时,引入S函数来降低滑模控制带来的抖振,改善系统的性能。最后,将本文所提出的方法与常规控制方法进行对比,验证了所提出的方案优越性。Permanent magnet synchronous linear motor（PMSLMS） servo system is a complex and strong coupling nonlinear system. It is difficult to control PMSLMs servo system because the system is sensitive to the changes in motor parameter external disturbances. Based on fractional order systems theory and sliding mode control theory, a fractional order global sliding mode intelligent control approach was proposed. The proposed method has global robustness, which makes the control system states on the fractional order sliding mode sur- face when the controller acts. Then, the uncertainty boundaries of the system was estimated and compensated via RBF neural network. Next, the S function was introduced to attenuate the chatting level and improve system performance. Finally, the proposed method was compared with the conventional control to verify the superiority of the proposed scheme.

关 键 词：永磁直线同步电机(PMLSM) 分数阶 全局滑模控制 RBF网络 

分 类 号：TM351[电气工程—电机] TM341