Experimental Validation of a Modular Low-Voltage Battery Management System with Model-Based Architecture

The design and experimental validation of a modular low-voltage battery management system (LV-BMS) for low-voltage electric vehicle subsystems and second-life battery applications are presented in this study. A key contribution of the study is the embedded integration and experimental validation of functions that are typically addressed separately in the literature, including OCV-based initialization, Coulomb counting, a 3RC equivalent circuit model with an Extended Kalman Filter (EKF), coulomb-throughput-based State of Health (SoH) estimation, and passive single-cell balancing, within a unified architecture using repurposed cells. While the model-based voltage feedback layer reduces long-term drift under dynamic loading conditions, the current-integration layer provides fast short-term response. A four-cell battery pack was used for experimental validation under controlled operating conditions. At the end of the discharge test, the Coulomb Counting method produced a 39.5 % relative SoC estimation error with respect to the reference value, whereas the EKF-based method reduced this error to approximately 3.8 %. The results demonstrate that the proposed architecture provides an embedded-feasible and experimentally validated LVBMS solution for safe and reliable battery operation.