Extracting and identifying essential characteristics of gear pitting and cracks with consideration of fault mechanisms

Pitting and cracking, common faults in gear transmissions, manifest distinct characteristics in vibration signals due to their differing generation and characteristics. Extracting and identifying these essential characteristics that reflect the unique traits of pitting and cracks presents a challenging task in gear fault diagnosis. This article initially employs simple Fourier transform method considering vibrational mechanisms to extract the impact waveforms caused by pitting and cracks from dynamic responses, and further proposes that three faults indicate assessing the characteristics of pitting and crack. Subsequently, a novel weak dictionary-based sparse enhancement method has been presented inspired by signal modulation, which can adaptively enhance the frequencies of pitting and crack in the frequency domain while preserving the lower components related to gear meshing frequencies. Ultimately, the effectiveness of the proposed method and indicators was validated through the experiments involving different severities of gear pitting and cracks.