The increasing demand for electric mobility underscores the need for precise and efficient battery diagnostics, particularly within Battery Management Systems (BMS) that monitor important parameters such as State of Charge (SoC) and State of Health (SoH). Conventional methods like Coulomb Counting and Open-Circuit Voltage measurements provide limited precision and fail to capture the underlying chemical processes within cells. Electrochemical Impedance Spectroscopy (EIS), a proven laboratory method, offers a more detailed analysis by measuring impedance across a wide frequency range. However, challenges related to complexity, cost, and system integration have restricted its use in mobile applications.
This study explores the integration of an onboard EIS module into a BMS for real-time SoC and SoH diagnostics in lithium-ion batteries. A prototype system for an 8-cell lithium-ion module was validated under diverse conditions, including varying SoC levels, temperatures, and load currents. Key achievements include the validation of the onboard EIS prototype against laboratory systems.
The Development of a machine learning-based SoH classification algorithm (Random Forest), comparing various frequency ranges has been achieved. The 1 kHz to 1 Hz range demonstrated high accuracy alongside a reduced measurement time of 33 seconds.
The results confirm the feasibility of onboard EIS for real-time, precise state diagnostics, addressing challenges such as robustness under real load conditions and external disturbances.
Future work will focus on module-level integration, enhanced algorithms, and expanded testing across different battery chemistries and operating conditions.