When the transmission gear has cracks, broken teeth and other faults, it will produce impact characteristics. Because of the impact component, the transmission gear vibration signal shows certain morphological characteristics in time domain. The morphological Haar wavelet is used to denoise the gear vibration signal, and the fuzzy entropy of the gear vibration signal after denoising is further calculated. The gear fault is identified and classified by comparing the fuzzy entropy differences of different types of fault signals of transmission gears.
The calculation process of gear fault classification based on morphological wavelet and fuzzy entropy is given
1) Using morphological Haar wavelet to decompose gear fault signal, the first layer, the second layer Approximate signal and detail signal of layer J;
2) Put layer 1, layer 2 The detail signal of layer J is processed by soft threshold denoising;
3) The fault signal is reconstructed by using the wavelet detail coefficients of each layer after de-noising;
4) The fuzzy entropy of the reconstructed fault signal is calculated, and the gear fault state is effectively classified according to the fuzzy entropy of the gear fault signal.