An Ergodic Method With Harmonic Signal-To-Noise Ratio In The Squared Envelope Spectrum Matrix For Planetary Gearbox Fault Diagnosis

Vibration signal of planetary gearbox contains plenty of amplitude modulated and frequency modulated (AM-FM) components, which carry much information about the operation status of planetary gearbox. In the process of fault diagnosis, it is necessary to decompose its vibration signal into mono-components and then use demodulation algorithm to extract fault features. In this paper, variational mode decomposition (VMD), harmonic L2/L1 norm (HL21N) based method, and harmonic signal-to-noise ratio (HSNR) are utilized to form a new method that can take their advantages. VMD can decompose AM-FMs signal into mono-components and it is robust against noise. HL21N, proposed recently, can reduce influence of outliers. In addition, HSNR, proposed in this paper, can screen out the best diagnosis result among all the squared envelope spectrums given by VMD-HL21N method. For the purpose of validating the proposed method, vibration signal collected from a planetary gearbox with ring gear tooth crack and planet gear tooth broken are investigated, respectively. In comparison study, fast kurtogram and HL21N-based methods are also used to analyze the same data. Finally, in the abovementioned two experiments, the proposed method is doing better than fast kurtogram and HL21N-based method.