Strain Signal-Based Fault Diagnosis Method for the Planet Gear in Planetary Gearboxes

Planetary gearboxes are widely used in mechanical equipment. However, their failures occur with high frequency because of the harsh working conditions. Conventional fault diagnosis methods of planetary gearboxes rely heavily on the acceleration signal. Although acceleration signals are easy to measure, they are susceptible to noise interference, making it challenging to identify fault types. Considering the close relationship between the tooth root strain of the ring gear and the mesh force, this paper aims to identify various tooth faults of the planet gear in the planetary gearbox based on the ring gear tooth root strain signal. Firstly, the ring gear strain is dynamically detected using the fiber Bragg grating (FBG) sensor. Secondly, an improved time synchronous averaging algorithm is proposed to enhance the fault information in the strain signal. Compared with the conventional time synchronous averaging method, the improved method replaces the arithmetic mean value with the root mean square value, sets the averaging time to an integer multiple of the number of ring teeth, and filters out the rotation frequency of the carrier and its harmonics from the strain signal. The proposed method is validated on a planetary gearbox test rig, and the experimental results show that the fault types of the planet gear can be effectively identified.