Research on Diagnosing the Gearbox Faults Based on Near Field Acoustic Holography

Abstract

The gearbox fault diagnosis was developed for some decades. The current diagnosis techniques were mostly based on analyzing the shell vibration signals especially close to the bearing seat of gearbox. In order to utilize the spatial distribution information of fault signal, the near field acoustic holography (NAH) is employed for the condition monitoring and fault diagnosis of the gearbox in this presentation. The distribution images of sound pressure on the surface of gearbox are reconstructed by NAH, and the feature extraction and pattern recognition can be made by image processing techniques. A gearbox is studied in a semi-anechoic chamber to verify the fault diagnosis technique based on NAH. The pitting and partial broken tooth faults of gears are artificially made on one gear as the fault statuses, and the differences of acoustic images among normal and fault working states under the idling condition are analyzed. It can be found that the acoustic images of gearbox in the three different situations change regularly, and the main sound sources can be recognized from the acoustic images which also contain rich diagnosis information. After feature extraction of the acoustic images, the pattern reorganization technique is employed for diagnosis. The results indicate that this diagnosis procedure based on acoustic images is available and feasible for the gearbox fault diagnosis.