Monitoring of Industrial Machine Using a Novel Blind Feature Extraction Approach

Machinery with several rotating and stationary components tends to produce non-stationary random vibration signatures due the fluctuations in input loads process defects long hours of operation. Traditional heuristics methods are suitable for detection fault signatures, however, they become more complicated when level uncertainty or randomness exceeds beyond control. A novel methodology identify these using optimal filtering data is proposed eliminate any false alarms expected provide a higher probability correct diagnosis. In this paper, detailed pipeline algorithms presented along results investigation that was carried out. These investigations performed open-source published by NASA prognostics centre. The performance evaluated based on ground truth researchers. Based parameters fine-tuned ensure generalisation reliable performance.