Fault feature extraction of planetary gear using an optimized matching pursuit decomposition

This paper presents a novel method for fault feature extraction of planetary gear using optimized matching pursuit algorithm, which is applied to vibration sensor signal feature extraction and fault diagnosis of planetary gearbox. This method combined optimization algorithm, planetary gearbox vibration model, and matching pursuit to implement planet gear fault characteristic extraction. Genetic algorithm and matching pursuit were integrated to decompose the vibration sensor signals into parameterized planetary gearbox model bases. Then the parameters of model bases were composited to extract fault feature of planetary gear. Different planet gear faults were tested for verifying the presented methods in different scenarios. All results of experiment analysis demonstrated its effectiveness and reliability.