A tri-fold hybrid classification approach for diagnostics with unexampled faulty states. (January 2015)

• Record Type: Journal Article
• Authors: Tamilselvan, Prasanna; Wang, Pingfeng
• Is Part Of: Mechanical systems and signal processing. Volume 50/51(2015)
• Page Start: 437
• Page End: 455
• Publication Date: 2015-01
• Subjects: Fault diagnostics -- Health monitoring -- Unexampled faulty; Structural dynamics -- Periodicals; Vibration -- Periodicals; Constructions -- Dynamique -- Périodiques; Vibration -- Périodiques; Structural dynamics; Vibration; Periodicals; 621
• DOI: 10.1016/j.ymssp.2014.05.016
• Languages: English
• ISSNs: 0888-3270

Abstract

Abstract: System health diagnostics provides diversified benefits such as improved safety, improved reliability and reduced costs for the operation and maintenance of engineered systems. Successful health diagnostics requires the knowledge of system failures. However, with an increasing system complexity, it is extraordinarily difficult to have a well-tested system so that all potential faulty states can be realized and studied at product testing stage. Thus, real time health diagnostics requires automatic detection of unexampled system faulty states based upon sensory data to avoid sudden catastrophic system failures. This paper presents a trifold hybrid classification (THC) approach for structural health diagnosis with unexampled health states (UHS), which comprises of preliminary UHS identification using a new thresholded Mahalanobis distance (TMD) classifier, UHS diagnostics using a two-class support vector machine (SVM) classifier, and exampled health states diagnostics using a multi-class SVM classifier. The proposed THC approach, which takes the advantages of both TMD and SVM-based classification techniques, is able to identify and isolate the unexampled faulty states through interactively detecting the deviation of sensory data from the exampled health states and forming new ones autonomously. The proposed THC approach is further extended to a generic framework for health diagnostics problems with unexampled faulty states and demonstrated with health diagnostics case studies for power transformers and rolling bearings. Highlights: A tri-fold hybrid classification approach is presented for structural health diagnostics. A grand challenge of unexampled health states (UHS) in health diagnostics is addressed. A new thresholded Mahalanobis distance classifier is developed for the THC. Power transformers and rolling bearing health diagnostics applications are employed. Case studies results indicate that the THC approach can successfully diagnose UHS.