Data-Driven Prognosis of Multiscale and Multiphysics Complex System Anomalies: Its Application to Lithium-ion Batteries Failure Detection

Abstract

Advancements in experimental and modeling techniques allow for in-depth analysis of multiphysics phenomena in complex systems with unprecedented sophistication and details at discrete spatial and temporal scales. Energy systems are crucial for reliability, making health monitoring vital to prevent failures. Balancing experimental complexity and computational cost is challenging, leading to the need for predictive capabilities in prognostics and health monitoring (PHM). Using lithium-ion batteries as an example, we summarize PHM predictive modeling for remaining useful life, anomalies, and failure detection. Additionally, we introduce data-driven prognosis (DDP) as a new approach for detecting failures in such systems.