Gear Fault Feature Extraction Based on OMP and K-SVD

In mechanical transmission systems, gears are the most widely used transmission component. Once the gear fails, it will cause huge economic losses. Therefore, it is extremely important to perform gear fault diagnosis. In the process of gear fault diagnosis, the key is to extract the fault impact characteristics. In order to separate and extract the fault impact characteristics of the gear, a method for separation and extraction of gear fault impact characteristics based on Orthogonal Matching Pursuit (OMP) algorithm and K-SVD is proposed. First, the OMP algorithm and the Fourier dictionary are used to separate the modulated signal from the impact signal in the gear signal. Then, the separated impact signal is divided into Hankel matrices. Finally, the Hankel matrix is denoised and restored using the K-SVD algorithm. The results show that this method can well extract the fault impact characteristics of gears.