磁悬浮柔性转子系统变学习速率单神经元PID控制

磁悬浮轴承柔性转子系统的转速范围较大，固定的控制参数很难同时保证系统所需的低频和高频特性。本文在一般单神经元PID控制的基础上，将转速信号引入磁悬浮轴承数字控制器，在不同转速段选择不同的学习速率以调节控制参数得到合适的磁悬浮轴承的刚度和阻尼。试验结果表明，变学习速率单神经元自适应PID控制能够明显减小转子在各阶临界转速时的振幅，有助于磁悬浮轴承柔性转子系统的安全稳定运行。

Adaptive PID control for flexible rotor supported by active magnetic bearings based on single neuron with varying learning rate

The range of the rotation speed is so large for flexible rotor supported by active magnetic bearings that invariable control parameters can not satisfy the requirements of the system that operates on both low frequency and high frequency. In the paper, based on normal single neuron PID control, the signal of the rotation speed is introduced into the control, and learning rate varies with the rotation speed to change the control parameters to bring the active magnetic bearings with suitable stiffness and damping. The experimental results show that adaptive PID control based on single neuron with varying learning rate can reduce vibration of the rotor obviously at critical speeds, and help the system operate safely.