Abstract:

Feature extraction and health status are urgently needed to be solved in health status assessment for complex systems. A new method of health status assessment is proposed to deal with these problems. This method has three steps.Firstly, feature variable of vibration signal is extracted by empirical model decomposition (EMD) and singular value decomposition (SVD).Then, Mahalanobis space is constructed and optimized by using Mahalanobis-Taguchi system (MTS), which can reduce dimension of feature variable.Finally, a new concept of health index (HI) is proposed for health status assessment of complex systems. The method is successfully applied to the health status assessment.