Study on PID tuning strategy based on dynamic stiffness for radial active magnetic bearing.
Study on PID tuning strategy based on dynamic stiffness for radial active magnetic bearing.
复制标题
DOI：
10.1016/j.isatra.2018.07.036
复制
发表时间：
2018-09
期刊：
影响因子：
7.3
通讯作者：
Jinji Sun;Han Zhou;Xin Ma;Ziyan Ju
中科院分区：
文献类型：
--
作者：
Jinji Sun;Han Zhou;Xin Ma;Ziyan Ju
工业技术的发展要求磁轴承工作在高速条件下。然而，由于涡流效应的影响，磁轴承的电流刚度和位移刚度会显著降低，这将使系统不稳定，甚至导致转子脱落和仪器损坏。因此，传统的基于常刚度的比例积分微分（PID）控制方法不适用于高速工况。提出了一种基于动刚度的径向磁轴承PID参数整定方法。通过分析参数随频率变化的等效磁路模型，建立了涡流效应下的动刚度模型。根据特征方程和Routh-Hurwitz准则，提出了包含动刚度模型的RAMB控制系统的PID参数整定方法。在仿真中设置了不同的PID参数，并进行了相应的实验。实验结果表明，所提出的PID参数整定方法具有良好的控制效果，与理论分析相一致。本文的仿真和实验为控制器参数的设计提供了理论指导，对RAMB的结构优化具有研究意义。
The development of industry technology requires magnetic bearings to work in high speed conditions. However, the current stiffness and displacement stiffness of the magnetic bearing will decrease significantly due to the consequent eddy current effect, and that decrease will make the system unstable and even result in the rotor drop and instrument damage. Therefore, the traditional Proportional-Integral-Derivative (PID) method based on constant stiffness is not adaptable for high speed conditions. This paper proposes a PID parameters tuning strategy based on dynamic stiffness for the radial active magnetic bearing (RAMB). The dynamic stiffness model under eddy current effect is established by analyzing the equivalent magnetic circuit model in which parameters are frequency-dependent. The PID parameters tuning method for RAMB control system including dynamic stiffness model is put forward according to the characteristic equation and Routh-Hurwitz criterion. Different PID parameters are set in simulations and several corresponding experiments are conducted. Satisfactory control effects consistent with the theoretical analysis are obtained and thus the proposed PID tuning strategy is verified to be good. Simulations and experiments in this paper provide theoretical guidance for the design of controller parameters and have research significance for structural optimization of RAMB.