一种改进的永磁交流伺服系统滑模控制器设计  

Design of modified sliding-mode controller for permanent magnet AC servo system

作　　者：王兴[1] 郭小定[1] 柏达 范婷[1] 

机构地区：[1]湖南科技大学信息与电气工程学院,湖南湘潭411201

出　　处：《电气传动自动化》2016年第2期13-17,共5页Electric Drive Automation

基　　金：国家自然科学基金资助项目(51577057)

摘　　要：针对传统整数阶PID伺服控制精度低、鲁棒性差等问题,结合滑模控制、分数阶理论、和神经网络参数整定技术各自的优点,提出了一种分数阶神经滑模控制的控制策略。首先对PMSM伺服系统的结构模型和数学模型进行了详细的阐述。然后利用滑模控制和等效控制理论设计了一种新的分数阶滑模控制器,提高了其综合性能。最后利用神经网络的强学习能力对分数阶滑模控制率参数进行了有效整定。实验结果表明,此分数阶神经滑模控制策略鲁棒性强、响应速度快、定位精度高,能有效消除抖振,具有良好的综合性能。In view of the traditional integer order PID servo control with low precision and poor robustness,combining with the respective advantages of the sliding mode control,the fractional theory and the neural network parameter setting,a neural sliding mode fractional order control strategy is proposed. Firstly,the full-digital fuzzy servo system structure model and mathematical model have been detailed,then the design of a new kind of fractional order sliding mode controller by using of the sliding mode control and equivalent control theory for improving the comprehensive performance is presented,and finally the parameters setting of the fractional order sliding mode control using the strong learning ability of the neural network is given. The experimental results show that this fractional order neural sliding mode control strat egy enables the system of strong robustness,fast response and high positioning accuracy,and can effectively eliminate the chattering and makes the system good comprehensive performance.

关 键 词：PMSM 伺服系统 滑模控制 分数阶 神经网络 鲁棒性 

分 类 号：TM351[电气工程—电机]