Rotor Fault Identification of Induction Motor Based on Discrete Fractional Fourier Transform

Abstract

While the rotor of a variable frequency induction motor malfunctions, the signal change of stator current will reflect this fault. This paper proposes an identification method of rotor fault for induction motor based on discrete fractional Fourier transform of stator current. The induction motor's three-phase stator current is firstly measured and then discrete fractional Fourier transform is utilized to identify the rotor's fault. Discrete fractional Fourier transform can transform time domain signal at any angle. This study divides angles 0~2π into 20 equal parts. Each part of angle can then be sequentially transformed to build a characteristic matrix. In order to make the characteristic information more identifiable, fractal theory is used to extract two features of fractal dimension and lacunarity from characteristic matrix. Finally, extension theory is applied to identify the defective type. In the identification process, ±5%~±15% random noise is added to evaluate the anti-interference ability of the proposed method. To verify the effectiveness three defective models are specially built according to common induction motor faults, including three broken rotor bar, one hollow rotor bar, and bearing ball damage. Normal motor is also added for identification analysis. The test results show that the identification method can effectively identify rotor fault type and has the advantage of strong anti-interference.