电工技术学报  2018, Vol. 33 Issue (17): 4044-4051    DOI: 10.19595/j.cnki.1000-6753.tces.180019
电机与电器 |
基于函数链径向基神经网络的PMLSM自适应反推控制
吴勇慷, 赵希梅
沈阳工业大学电气工程学院 沈阳 110870
Adaptive Backstepping Control Based on Functional Link Radial Basis Function Neural Network for PMLSM
Wu Yongkang, Zhao Ximei
School of Electrical Engineering Shenyang University of Technology Shenyang 110870 China
全文: PDF (17829 KB)  
输出: BibTeX | EndNote (RIS)      
摘要 为提高永磁直线同步电动机(PMLSM)伺服系统的控制性能,解决参数变化、外部扰动和摩擦力等不确定性因素对系统影响的问题,提出一种基于函数链径向基神经网络(FLRBFNN)的自适应反推控制(ABC)方法。首先建立含有不确定性因素的PMLSM动态模型;其次,利用ABC中的自适应律对系统总不确定性进行估计,但在设计ABC时存在大量求导运算,以至于产生“微分爆炸”现象。因此,为解决这一问题并进一步提高系统性能,采用FLRBFNN在线学习并调整控制器参数,FLRBFNN将径向基神经网络(RBFNN)和函数链神经网络(FLNN)相结合,利用FLNN增大神经网络搜索空间,提高网络收敛速度和收敛精度,从而提高RBFNN估计系统不确定性的能力,有效降低不确定性因素对系统的影响。实验结果表明,该方法切实可行,与ABC相比,能够使系统具有较强的鲁棒性能和跟踪性能。
服务
把本文推荐给朋友
加入我的书架
加入引用管理器
E-mail Alert
RSS
作者相关文章
吴勇慷
赵希梅
关键词 永磁直线同步电动机不确定性因素函数链径向基神经网络自适应反推控制    
Abstract:An adaptive backstepping control (ABC) based on functional link radial basis function neural network (FLRBFNN) was proposed to improve the control performance of permanent magnet linear synchronous motor (PMLSM) servo system and solve the influence on the system of the uncertain factors such as the changes in system parameters, external disturbances, frictions and so on. Firstly, the dynamic model of PMLSM with the uncertain factors was established. Then, the adaptive law in ABC is used to estimate the lumped uncertainties of the system, but the “differential explosion” phenomenon is produced because of a large number of derivation operations during the design of ABC. Thus, in order to solve the problem and further improve the system performance, FLRBFNN is used to learn and adjust the controller parameters online. FLRBFNN is combined with radial basis function neural network (RBFNN) and functional link neural network (FLNN), FLNN is used to increase the searching space to improve the convergence speed and convergence accuracy of neural network so that it can improve the ability to estimate the uncertainties of RBFNN and reduce the influence of uncertain factors to the system. The experimental results show that the proposed method is effective. Compared with ABC, the system has stronger robust performance and tracking performance.
Key wordsPermanent magnet linear synchronous motor    uncertain factors    functional link radial basis function neural network    adaptive backstepping control   
收稿日期: 2018-01-08      出版日期: 2018-09-14
PACS: TP273  
基金资助:辽宁省自然科学基金计划重点项目(20170540677)和辽宁省教育厅科学技术研究项目(LQGD2017025)资助
通讯作者: 赵希梅 女,1979年生,副教授,博士生导师,研究方向为电机控制、机器人控制、智能控制等。E-mail:zhaoxm_sut@163.com   
作者简介: 吴勇慷 男,1992年生,硕士研究生,研究方向为电机控制、智能控制。E-mail:wuyongkangch@gmail.com
引用本文:   
吴勇慷, 赵希梅. 基于函数链径向基神经网络的PMLSM自适应反推控制[J]. 电工技术学报, 2018, 33(17): 4044-4051. Wu Yongkang, Zhao Ximei. Adaptive Backstepping Control Based on Functional Link Radial Basis Function Neural Network for PMLSM. Transactions of China Electrotechnical Society, 2018, 33(17): 4044-4051.
链接本文:  
https://dgjsxb.ces-transaction.com/CN/10.19595/j.cnki.1000-6753.tces.180019          https://dgjsxb.ces-transaction.com/CN/Y2018/V33/I17/4044