Model-based Prognostics with Fixed-lag Particle Filters
Matthew J. Daigle, Kai Goebel · 2009
Model-based prognostics exploits domain knowl-edge of the system, its components, and how they fail by casting the underlying physical phenom-ena in a physics-based model that is derived from first principles. In most applications, uncertain-ties from a number of sources cause the predic-tions to be inaccurate and imprecise even with accurate models. Therefore, algorithms are em-ployed that help in managing these uncertainties. Particle filters have become a popular choice to solve this problem due to their wide applicability and ease of implementation. We present a gen-eral model-based prognostics methodology using particle filters. In order to provide more accu-rate and precise estimates, and, therefore, more accurate and precise predictions, we investigate the use of fixed-lag filters. We develop a detailed physics-based model of a pneumatic valve, and perform comprehensive simulation experiments to illustrate our prognostics approach. The exper-iments demonstrate the advantages that fixed-lag filters may provide in the context of prognostics, as measured by prognostics performance metrics. 1