论文部分内容阅读
提出一种新型的基于模型参考神经网络的异步电机驱动系统鲁棒速度控制方法。由带负载转矩观测器的两层神经网络对象辨识器(NNPI)对未知的电机动态参数进行实时的自适应辨识与估计。由双层神经网络PI控制器(NNC)对异步电机转子速度进行鲁棒控制。神经网络使用学习算法以自动调节NNPIC的参数并有效地降低系统对参数变化以及负载扰动的敏感度。仿真结果表明该方法对于参数变化和负载转矩扰动具有很强的自适应能力,能够提高异步电机的性能,并减小其对参数变化、非线性影响以及负载扰动的敏感度。
A new robust speed control method for induction motor drive system based on model reference neural network is proposed. The unknown motor dynamic parameters are adaptively identified and estimated in real time by a two-layer Neural Network Object Recognizer (NNPI) with a load torque observer. The rotor speed of asynchronous motor is robustly controlled by a two-layer neural network PI controller (NNC). Neural networks use learning algorithms to automatically adjust the parameters of NNPIC and effectively reduce the sensitivity of the system to parameter changes and load disturbances. The simulation results show that this method has strong adaptability to parameter variation and load torque disturbance, which can improve the performance of induction motor and reduce its sensitivity to parameter variation, nonlinear influence and load disturbance.