摘要
针对常规PID技术不能满足磁轴承(MB)的鲁棒性控制要求,提出一种基于模糊小波神经网络(FWNN)的磁轴承控制方法。将差分驱动模式的MB系统构建成一个非线性动态模型,在模糊神经网络(FNN)中嵌入小波神经网络(WNN)形成F-WNN,并通过监督学习机制对F-WNN的参数进行在线学习,同时利用遗传算法优化学习率因子。最后,以转子位置误差作为F-WNN的输入,以控制电压作为输出,实现转子的自适应控制。仿真结果表明,该方法在存在干扰情况下能够精确控制MB转子位置,具有较高的实际应用价值。
For the issue that the conventional PID technology can＇t meet the requirements of the magnetic bearing （MB） robustness control, a magnetic bearing control method based on fuzzy-wavelet neural network （F-WNN） is pro- posed. Firstly, a nonlinear dynamic model was built for the MB system with differential drive mode. Then, the wavelet neural network （WNN） was embedded in the fuzzy neural network （FNN） to form F-WNN, and the supervised learning mechanism was used for online learning the F-WNN parameters. At the same time, the genetic algorithm （GA） was adopt- ed to optimize the learning rate factor. Finally, the rotor position error is regard as the input of the F-WNN, and the con- trol voltage as the output, realized adaptive control of the rotor. Experimental results show that the proposed method can ac- curately control the position of MB rotor in the presence of interference, and has high practical application value.
出处
《微特电机》
北大核心
2017年第4期77-81,92,共6页
Small & Special Electrical Machines
基金
广西教育厅课题项目(KY2015YB440)
广西工业职业技术学院课题项目(桂工业院[2014]56号)
关键词
磁轴承
转子位置控制
模糊小波神经网络
遗传算法
magnetic bearing （MB）
rotor position control
fuzzy-wavelet neural network （F-WNN）
genetic algorithm （GA）
