Embedded System-Based Extended Kalman Filter for Real-Time Soc Estimation of Lithium-Ion Cells

• Main Authors: Surya Praakash, H N, Krishna Reddy, Aala Kalananda Vamsi, Meenakshi, G, Sharan, S, Ananya, R, Rashmi, N
• Format: Conference Proceeding
• Language: English
• Subjects: Accuracy; Battery Management System (BMS); Computational modeling; Coulomb Counting (CC); Drive Cycle analysis; Electric Vehicle (EV); Estimation; Extended Kalman Filter (EKF); Kalman filters; Lithium-ion batteries; NMC cell; Offline parameters; One-RC model; Parameter estimation; Real-time systems; Robustness; SoC estimation; State estimation; Testing
• ISSN: 2469-5556
• DOI: 10.1109/ICACCS60874.2024.10716819

The State-of-charge (SoC) estimation in Lithium-ion batteries is a crucial and integral component in battery management systems (BMS) for electric vehicles (EVs), The adoption of Kalman filters (KF) for SoC estimation including extended KF (EKF) has been at the forefront of SoC estimation on account of their accuracy and reliability. This article comprehensively analyzes the efficacy of EKF-based SoC estimation techniques including EKF followed by the conventional Coulomb counting (CC) for Lithium-ion battery pack on a low-budget BMS with limited computational complexity. A custom-designed simple BMS setup with STM32 L5-based microcontroller and TI's current sensor is deployed on a cell-cycler. Various testing scenarios have been analyzed for both charging and discharging cycles with real-world e-bike driving data for the Indian urban traffic conditions in Noida. The test results demonstrate the dominance of EKF over CC with SoC estimation error being confined under 4% and RMSE error under 0.5%.