基于神经网络和分数阶滑模的无轴承永磁同步电机悬浮控制系统  

BPMSM Suspension Control System Based on Neural Network and Fractional Order Sliding Mode

作　　者：宝金[1] 张呼和[2] 

机构地区：[1]包头师范学院,包头014030 [2]包头轻工职业技术学院,包头014030

出　　处：《微特电机》2015年第10期64-67,87,共5页Small & Special Electrical Machines

摘　　要：无轴承永磁同步电机悬浮力控制中,其强耦合特性会造成系统的抖动。针对该问题建立了无轴承永磁同步电机径向悬浮力模型;基于微分几何控制理论对强耦合悬浮系统进行精确线性化处理,并得到两个独立子系统;采用滑模控制对子系统进行解耦控制,并结合神经网络以解决滑动模切换过程中的抖振问题。同时设计了基于分数阶的神经网络滑模控制器,实现了系统不确定性的鲁棒控制,进一步避免了抖振的发生。通过仿真验证了基于分数阶的神经网络滑模控制系统响应速度快、抗干扰性能强、鲁棒性能好。The strong coupling characteristic of bearingless permanent magnet synchronous motor（ BPMSM） levitation force control can cause system chattering. Aimed at this problem the radial levitation force model of BPMSM was established in this paper. The precise linearization of strong coupling suspension system was done based on differential geometry control theory and two independent were achieved. The subsystems Decoupling was conducted based on sliding mode control,and combined with neural network the chattering problem in the process of sliding mode switch was solved. At the same time the fractional order sliding mode controller was designed based on neural network. The system uncertainty robust control was also realized. The simulation results verify that the neural network sliding mode control system based on fractional order has a lot of advantages such as high response speed,strong antijamming capability,and good robust performance.

关 键 词：无轴承永磁同步电机 悬浮力控制 神经网络 滑模控制 分数阶 

分 类 号：TM341[电气工程—电机] TM351