Study on fault diagnosis of gearbox based on morphological filter and multi-fractal theory

Based on mathematical morphology, the median combined morphological filter and differential morphological filter were designed. With the procession of simulation and real vibration signal, the different effects of the two filters were researched. The results show that the combined filter can filter the impulse and stochastic noises, but the differential filter can extract the impulses in the signal. After the noise interference was reduced effectively, multi-fractal theory was used to analyze the gearbox vibration signals which under different operating states. By calculating the multi-fractal spectrum parameters and the multi-fractal entropy, the different faults can be distinguished. The analysis results show that multi-fractal parameters give a good presentation for gearbox working condition and provide necessary evidence for gearbox fault feature identification.