Development of Approximate Entropy as a Prognostic Parameter for Rotating Equipment

Bearings are an essential component throughout most rotating equipment and power-generating systems. Current maintenance practices allow for premature replacement of bearings in a scheduled fashion. By moving to a condition-based maintenance strategy, the health of the component is assessed and used to determine the remaining useful life. This can improve plant safety and efficiency. My research focuses on vibration analysis and developing approximate entropy - a statistic that quantifies a signal's regularity and complexity - as a health indicator for bearings. Preliminary results show that this statistic has earlier fault detection and better trendability than other commonly-used features.