Paper Title
Robust Features For Load Independent Machine Fault Diagnosis Using Feature Selection
Abstract
Machine Fault Diagnosis (MFD) is of vital importance in aerospace applications that consist of complex systems.
In this paper, load independent MFD on a three phase 3kVA synchronous generator has been described. Statistical features
were extracted in both time and frequency domains. Chi squared algorithm was used for feature selection and classification
was done using C4.5 algorithm. Voltage and current analyses were performed in both domains, and a set of robust features
were identified.