A new method for stator resistance identification based on wavelet network is presented to improve the operating performance of induction motor. The ripple of the torque is kept constant, resulting in a variable switching frequency, which depends on the tolerance band of the control. Due to wavelet transform behaving good localization property in both time and frequency space and multi-scale property, the wavelet function is adopted as the basic function of neural network. The current error and the change in the current error are the inputs of the wavelet network and the stator resistance error is the output of the wavelet network. The network parameters are initialized by the improved least squares algorithm. The parameters of wavelet network are tuned online to approximate the unknown nonlinear model with an appropriately chosen adaptive mechanism. The proposed method is proved to be efficient to reduce the torque ripple and current ripple by detailed comparison simulation results.

Date Published: 13 October 2008
PDF: 4 pages
Proc. SPIE 7129, Seventh International Symposium on Instrumentation and Control Technology: Optoelectronic Technology and Instruments, Control Theory and Automation, and Space Exploration, 712914 (13 October 2008); doi: 10.1117/12.807385

Published in SPIE Proceedings Vol. 7129:
Seventh International Symposium on Instrumentation and Control Technology: Optoelectronic Technology and Instruments, Control Theory and Automation, and Space Exploration
Jiancheng Fang; Zhongyu Wang, Editor(s)