RUL prediction Using Bayesian polynomial Regression

Maintaining the reliability of complex systems is crucial in today's technological landscape. Maintenance strategies have evolved from corrective and time-based maintenance to condition-based maintenance and prognostics and health management. Typical remaining useful lifetime (RUL) prediction methods require substantial historical data, posing challenges in data-limited scenarios. To address this, we propose an efficient Bayesian polynomial regression approach with informative priors that predicts RUL even with sparse data. Regression parameters are continuously updated as new data are collected, ensuring accuracy and responsiveness. We validate our algorithm on simulated power module run-to-failure degradation data.

Keywords: Bayesian regression, Prognostics and health management, Remaining useful lifetime.