基于饱和约束LMS算法的磁悬浮隔振器控制研究

摘要

The anti-saturated control for vibration reduction with the electromagnetic suspension vibration isolator was studied. The conventional magnitude-limited anti-saturated LMS algorithm may easily lead to the resonance of the isolator system and result in deterioration of control effet. Moreover, it is difficult to determine the threshlod value in Boaz Rafaely algorithm; if the threshold is too high, the resonance of isolator system may be induced, if the threshold is too low, the control effect may become worse. In consideration of those facts, a novel approach, in which a continuous punishment function was introduced, was proposed in the framework of adaptive feedforward control on the basis of LMS algorithm. The improvement in LMS algorithm was presented and the anti-saturation adaptive feedforward control rule was designed accordingly. In order to verify the method is more superior than the magnitude-limited anti-saturated LMS algorithm and Boaz Rafaely algorithm, simulations and active vibration control experiments on the electromagnetic suspension vibration isolation system were carried out.%研究磁悬浮隔振器的抗饱和减振控制,针对限幅抗饱和法容易激起磁悬浮隔振器共振和Boaz Rafaely抗饱和算法的阈值很难确定,如果阈值选择过大容易激起磁悬浮隔振器共振,如果选择过小则损失减振效果这些问题,提出一种饱和约束处理方法,即加入连续惩罚函数方法.推导相应的改进LMS算法并设计出抗饱和自适应前馈控制律.通过仿真和在磁悬浮隔振器系统上进行实验对比,验证这种方法相对于限幅抗饱和算法和Boaz Rafaely抗饱和算法具有一定的优越性.