Real-time Learning Embedding Next Gen Battery Management Systems for EV Safety

This 41-minute lecture by Dr. Avimanyu Sahoo from the AI Institute at UofSC explores the evolution of Battery Management Systems (BMS) for electric vehicles with a focus on safety enhancement. Delve into the challenges of current BMS technology, particularly regarding battery pack overheating and fire incidents that reveal limitations in detecting cell-level anomalies and internal faults. Discover an innovative approach that combines physics-driven models with real-time learning capabilities to create advanced BMS solutions. Learn about the State of Health (SOH)-coupled Electro-Thermal-Aging (ETA) model for lithium-ion cells that integrates equivalent circuit, thermal parameters, and capacity fade to provide comprehensive insights into cell behavior under various operating conditions. Examine physics-driven fault detection mechanisms that can differentiate between actual faults and normal cell degradation. Gain perspective on the future roadmap for BMS technology with embedded learning capabilities that promises to usher in a new era of electric vehicle safety and reliability.