单神经元-PID算法在磁轴承系统的研究

Research on Single-nerve-cell PID Arithmetic Application in Magnetic Bearing System

常规PID控制器参数往往整定不良,性能欠佳,对磁轴承系统运行工况的适应性很差。而由具有自学习和自适应能力的单神经元构成的单神经元自适应PID控制器,不但结构简单,而且能适应环境变化,具有较强的鲁棒性。将神经网络控制思想引入磁轴承系统中,提出了适合于磁轴承单自由度控制的单神经元-PID算法,通过实验探讨了其控制效果,并与传统PID算法控制效果作了比较。

The routine PID controller had a bad adaptability on the operation of magnetic bearing system due to its inefficient parameter coordination. The single-nerve-cell-PID self-adaptive controller, which consisted of single-nerve-cells with self- study and self-adaptive abilities, was excellent for its simple configuration, flexible adaption in changeful environment and strong robustness. The single-nerve-cell-PID arithmetic, which was appropriate for single-degree-of-freedom control of magnetic bearing, was provided and its control effect was estimated based on results of experimentation. In the end, such control effect was reasonably compared with that of traditional PID arithmetic.