An encoder signal-based approach for low-speed planetary gearbox fault diagnosis

Low-speed rotating machines are extensively used in heavy industry. Among those, the planetary gearbox is a pivotal component with a high power-weight ratio and large loadbearing capacity, which inevitably fail due to the tough working conditions. The fault signature in such conditions is rather weak due to the complex planetary structure and the low rotating speed. Hence, the diagnosis of planetary gearbox problems in low-speed working conditions is considered as a bottleneck issue. In view of this, a rotary encoder signal, instead of conventional vibration, is initially applied to capture the fault-related information from the low-speed planetary gearbox. Then, a periodic group sparse-robust principal component analysis (PGS-RPCA) model with adaptive parameter programming, called adaptive PGS-RPCA (APGS-RPCA) is presented to extract the weak fault transient immersed in harmonic interferences and heavy noise. Finally, the effectiveness of the presented APGS-RPCA approach is verified via an experimental encoder signal at a very low input frequency. The diagnostic results show that the presented approach is superior to the conventional approach, and it may provide a promising solution for health monitoring of low-speed rotating machinery.