The objective of this project is to optimize performance, improve safety, minimize over-design, and reduce cost of battery systems in electrified vehicles. If successful, higher adoption of electrified vehicles is foreseeable in the near future to revolutionize the future transportation systems so that environmental challenges can be addressed for conserving energy and cutting back on carbon emissions and pollution. This research will also have significant impact on other applications such as aircraft electric systems, military portable devices, and aerospace battery applications, where lithium ion batteries are used for energy storage. These impacts will result in economic and social benefit and help the U.S. stay competitive globally. The theory and methodology will be integrated into education through modifications in related courses in order to engage graduate and undergraduate students.
This project studies the uncertainty management and proactive maintenance of lithium ion batteries in electrified vehicles. The scope of the proposed work is to: i) increase accuracy of a baseline battery model; ii) develop an intelligent uncertainty management system; and iii) facilitate proactive maintenance decisions for lithium ion batteries. The investigators will address four issues including: i) characterization of battery model uncertainty under various battery operating conditions; ii) quantification of various uncertainties and their coupling effects in the battery system; iii) prediction of battery remaining useful life (RUL) under various battery operating conditions; and iv) validation of the proposed approaches on an electrified vehicle. Traditionally, researchers employ electrochemical models to improve battery model accuracy. However, a full order electrochemical model is not feasible for the battery management system (BMS) in electrified vehicles, and a significant level of model simplification must be established to meet the computational efficiency requirement which will cause unwanted model error. The first contribution of the research project will be to study an effective model uncertainty characterization approach to improve model prediction accuracy of a low-fidelity model (i.e., an equivalent circuit model) so that its accuracy is comparable to that of a high fidelity model (i.e., an electrochemical model) but with much higher computational efficiency. The second contribution of the project will be to develop an intelligent uncertainty management system for more effective battery performance estimation so that battery safety can be improved and lifetime risk minimized. Finally, the challenge of battery proactive maintenance stems from the complexity of battery devices. Even the best models cannot predict the complex degradation and failure mechanisms. As a result, accurate prediction of the battery RUL based on the underlying degradation theory is almost impossible. The third contribution of the research will be to study proactive maintenance decisions for batteries through accurate RUL prediction using innovative data-driven prognostics and health management (PHM) technologies.
