Seminar: Stochastic Modeling for System Remaining Life Prognosis

Advancement in sensing and system estimation have established the foundation for predictive maintenance as a promising health management strategy. The goal of predictive maintenance is to predict the remaining useful life (RUL) of a machine system based on information extracted from sensor measurements, and make cost- effective decisions on maintenance scheduling before the equipment breaks down. Stochastic modeling, rooted in probabilistic theory and integrating physical knowledge, provides a promising tool, for RUL prediction.

This seminar presents research on Particle Filtering (PF) for RUL prediction. The fundamentals of PF first reviewed. Next, research on enhanced resampling strategy and multi-mode filtering for improved prediction accuracy and computational efficiency are highlighted, using aircraft engine performance degradation and RUL prediction as the application context. The presentation demonstrates the potential of enhancing physical science with data science, to achieve smarter machine health management and prognostics.

About the Speaker

Dr. Gao’s research focuses on physics-based sensing methods, design, modeling, and characterization of measurements systems, multi- resolution signal analysis, and energy efficient sensor networks for improving the observability of dynamical system and product quality control.