- 作者: Ta-Peng Tsao ; Rong-Ching Wu
- 中文摘要: In the past, the research on fault recognition for induction motors only concentrated on spectrum amplitudes which are based on a constant load. However, the frequency and amplitude of the spectrum analyzed under different fault conditions are also affected significantly by load variation. Hence, simply using spectrum amplitudes to recognize motor faults is not sufficient in a practical system. Both various load conditions and different types of faults will influence the spectrum structure. In order to recognize faults under various load conditions, we have to consider band shift and amplitude as two major factors. In this paper, we use band shift and amplitude techniques to solve the spectrum problem under various load conditions and different types of faults. We also use the methods of frequency axis adjustment and feature exaction to solve the band shift and amplitude variation problems respectively. After the above-mentioned procedures, efficient features are obtained. Then, we use the back propagation artificial neural network (ANN) to train and recognize fault conditions. In addition, we compare the recognition ability between the artificial neural network and traditional method. All the theories and methods used in the paper are validated by means of different experimental results on motors.
- 英文摘要: --
- 中文關鍵字: fault recognition, artificial neural network, feature exacting, induction motor
- 英文關鍵字: --