永磁直线电机运输系统中的模糊增益神经控制

A Neuron Controller with Fuzzy Self-Tuning Gains for PMLSM Transportation

永磁直线同步电动机的理论研究还不十分完善,在初级分段的形式下该运输系统整体数学模型变化很大且边缘效应的影响更加剧烈.因此不能利用其数学模型或常规的控制方法对垂直运动的PMLSM实现良好控制.本文设计了直线伺服系统的神经元控制器.首先应用神经元控制器,它可以自动调节权值获得多个控制参数,然后考虑到被控对象的多动态特性,加入输出比例因子的在线模糊修改部分,从而改善了控制器的动态和静态特性.实验证明,该伺服系统能够消除PMLSM固有的负载扰动、边缘效应、参数时变对系统伺服性能的影响;且具有自学习能力强、快速跟踪、定位精确和鲁棒性强等特点.

The theory of permanent magnet linear synchronous motor （PM LSM） doesn＇ t consummate, the whole math model of the transportation system changes a lot under the primary partition model, and the fringe effects is very fierce, so the servo-system of the PMLSM vertical movement can＇ t be designed according to its math model or by using simple controller. A neuron model-free controller with self-tuning fuzzy rules controllers was proposed for the servo-system in this paper. First, neuron model-free controller was designed and it can get different control parameters automatically. Then, multidynamic characteristics of PM LSM were considered, and a self-adjusting fuzzy layer was made to modify the proportional factor and improve its static and dynamic characteristics. The experimental results shown that the designed system can eliminate the influence of the load disturbances, the fringe effects, and the changes of system parameters. In fact, it has a strong robustness to uncertainties of control object, a strong ability of self-organization, a good tracking and positioning performance.