Skip to Main Content
As for non-destructive condition monitoring method, vibration signals are measured and analyzed to diagnose mechanical faults. A method of mechanical fault levels identification is proposed in this paper using vibration signals. Firstly, Hilbert-Huang transform is used to analyze the signal. Fault information is obtained. And energy ratio of intrinsic mode functions is taken as fault characteristics. Secondly, different levels of fault are identified by applying Mahalanobis-Taguchi System. Standard Mahalanobis space is constructed. And slight and severe fault are diagnosed based on its Mahalanobis Distance to standard space. The method is validated by real rotating machinery seeded-fault-test vibration signals, results showed that this method performed well.