Residual signal–based condition monitoring of planetary gearbox using electrical signature analysis
Electrical signature–based technique, due to its non-intrusive nature, is very useful for condition monitoring of rotary machines. Major challenge is the dominance of line frequency in the current signature. The present study focused on decreasing this dominance of the line frequency by obtaining the residual signals through autoregressive modeling. The residual signals are used further to extract the health-related features. The recently developed weighted multi-scale fluctuation-based dispersion entropy features are extracted as health indicators. The extracted health indicators are used for classification of different types of planetary gearbox faults. The results reflect that the proposed methodology has the potential for diagnosing different types of planetary gearbox faults with acceptable accuracy values.