Abhijit Dasgupta, Ravi Doraiswami, Michael Azarian, Michael Osterman, Sony Mathew, and Michael Pecht
Center for Advanced Life Cycle Engineering (CALCE), University of Maryland,
College Park, MD 20742, United States
Reliability concerns in state-of-the-art electronic systems have led researchers and engineers to develop innovative real-time prognostics and adaptive health management methods to assure desired availability. Prognostics techniques described here use a novel concept of canaries, along with data analysis, failure mechanism models and integrated fusion techniques to determine the remaining useful life of a system. A canary is a device that provides data to generate early warning of functional degradation and impending functional failure. Three types of canaries are discussed. Expendable canaries experience accelerated degradation (compared to functional degradation) by design, so that early warning of impending failure can be generated. Sensory canaries provide early warning by observing non-functional manifestation of functional degradation. Conjugate-stress canaries provide measurement of life-cycle stress history so that failure models can be used to estimate consumed life and remaining life. This paper focuses on expendable canaries, in particular, and provides three examples to illustrate the underlying design concepts.
Key words: canary; electronic systems; failure mechanisms; precursors; prognostics; reliability; remaining useful life.
Complete article available to CALCE consortium members