Physics-Informed Fault Prediction in BMS Using EKF

Created using ChatSlide

This research presentation delves into advanced Battery Management System (BMS) techniques for predicting unobservable states and detecting faults prior to failures, emphasizing the use of model versus data-driven detectors. We explore the 3-RC Thevenin Battery Model, focusing on its physical attributes and graphite-anode chemistry. The implementation of the Extended Kalman Filter (EKF) is highlighted, detailing process and measurement models alongside essential numerical safeguards....