Research On Stator Resistor Identification Of Induction Motor Based On Neural Networks

The direct torque control(DTC) theory as a kind of new method in the field of AC drive system. It calculates and controls the torque of AC motor in stator coordinate using the space vector analysis and stator-flux-oriented method. It is dynamic response fastly,simple and easy to realize the advantages. However, There are still some imperfections and areas for improvement.This paper first introduce development and situation of study for induction motor in direct torque control(DTC),then introduce fundamental theory and mathematical model for DTC system. Analyse in detail the problem of low speed in DTC system.In low speed ,the stator resistance has effected much by temperature,stator frequency and stator current .fastly effected system performance and analyse in quantifying.To solve the effect of stator resistance change of the low-speed performance in direct torque control (DTC),this paper use the intelligent control to identify stator resistance change,in order to improving the low-speed performance in DTC.This paper proposes a scheme of Model Reference Adaptive Control Using BP neural network and RBF neural network identifying stator resistance used in DTC. The later scheme Using gradation algorithm to training parameters of RBFNN. Applying this scheme identifies stator resistor, which have high precision and adaptability. Simulations are carried out to verify the proposed strategy based on the simulink , and compared to BP neural network identifying stator resistance. The simulation results show that the scheme is superior than the BP neural network and capable of effectively improving the low-speed performance in DTC. Therefore,this paper which proposed sdutying on stator resistor identification of induction motor based on neural networks is valid.