基于动态轮廓误差估计的三轴运动平台NNLARC控制  

Three-axis motion platform NNLARC control based on dynamic contour error estimation

作　　者：王丽梅[1] 赵晨 孙宜标[1] WANG Limei;ZHAO Chen;SUN Yibiao(School of Electrical Engineering,Shenyang University of Technology,Shenyang 110870,CHN)

机构地区：[1]沈阳工业大学电气工程学院,辽宁沈阳110870

出　　处：《制造技术与机床》2019年第5期116-121,共6页Manufacturing Technology & Machine Tool

基　　金：国家自然科学基金资助项目(51875366;51175349)

摘　　要：在直线电动机驱动三轴运动平台中,为提高轮廓加工精度,需要解决两个主要问题,即轮廓误差估计和轮廓误差控制。采用基于牛顿极值搜索算法的动态轮廓误差估计的方法,建立更为精确的轮廓误差模型。由于永磁同步直线电动机(PMLSM)易受外部扰动和摩擦力等因素影响,为使系统具有良好的跟踪性能和抗干扰性能,采用神经网络学习的自适应鲁棒控制(NNLARC)进行单轴位置控制器的设计,削弱各种因素对于系统性能影响。另外,采用位置环交叉耦合控制器,解决三轴间的动态特性不匹配问题,进一步减小轮廓误差。仿真结果表明,所设计的方法能有效提高系统的加工精度和鲁棒性。In the linear motor driven three-axis motion platform, in order to improve the contour machining accura-cy, two main problems need to be solved, namely contour error estimation and contour error control. In this paper, a dynamic contour error estimation method based on Newton^s extreme value search algorithm is used to establish a more accurate contour error model. Since the permanent magnet synchronous linear motor ( PMLSM) is susceptible to external disturbances and frictional forces, the neural network learning adaptive robust control (NNLARC) is used for single-axis position in order to make the system have good tracking performance and anti - interference performance. The design of the controller attenuates the effects of various factors on system performance. In addition, the position loop cross-cou-pling controller is used to solve the problem of dynamic characteristic mismatch between the three axes, further reducing the contour error. The simulation results show that the designed method can effectively improve the processing accuracy and robustness of the system.

关 键 词：多轴运动控制 牛顿极值搜索算法 轮廓误差 自适应鲁棒控制 神经网络 

分 类 号：TM359[电气工程—电机] TP273[自动化与计算机技术—检测技术与自动化装置]