一种基于FNN的感应电机定子电阻辨识建模方法

直接转矩控制技术的关键问题是如何在定子电阻不确定的情况下，准确辨识定子磁链。通过转子磁链定向的方法，推导出转差的表达式，此表达式中包含定子电阻的信息，当定子电阻发生变化时，计算出来的转差与实际的转差不相等，基于这种思想构建了模糊神经网络，采用梯度下降法训练网络的参数，实现对定子电阻的在线辨识。针对所提出的方法，在matlab/simulink中搭建模型进行仿真，结果表明在定子电阻受干扰的情况下，此方法能准确、迅速地跟随定子电阻的变化，实现实时调整，感应电机的低速性能得到改善。

Modeling method of induction motor stator resistance identification based on Fuzzy Neural Network

The key problem of direct torque control technology is how to acquire the stator flux linkage accurately when the stator resistance is uncertain. In this paper, taking the method of rotor flux orientation, and deducing the expression of slip, this expression contains the information of the stator resistance. As the stator resistance is changing, the slip of the calculation and the actual slip is not equal. The fuzzy neural network is constructed based on this idea, and the gradient descent method is used to train network parameters. It implements the online identification of the stator resistance. For the mentioned methods in this paper, a simulation is done in matlab/simulink. The results are that this method can accurately and quickly follow the change of the stator resistance under the condition of the stator resistance interference. Therefore, it realizes stator resistance real-time adjustment and improves motor’s low speed performance.