Development of Module-Integrated Cell Monitoring Units and Model-Based State-of-Charge Estimation for an Innovative Battery Management System Architecture

Abstract

The growing complexity of electric vehicles asks for new battery management system (BMS) designs that are simpler, more reliable, and cost-efficient. This thesis presents the development of compact cell monitoring units (CMUs) that are mounted directly onto battery modules, eliminating bulky harnesses, reducing wiring complexity, and maintaining measurement accuracy. The CMUs combine redundant voltage and temperature sensing, passive balancing, and galvanic isolation in a custom four-layer printed circuit board optimized for thermal management and validated through simulation and experiments. Since these CMUs will be integrated into a future electrical/electronic architecture where the BMS master functions reside in the vehicle control unit, this design reduces component count and wiring complexity while maintaining functional safety. In parallel, a model-based state-of-charge (SoC) estimator was implemented as a redundant method to conventional Coulomb counting (CC), which suffers from drift, initial offset, and dependence on usable capacity. Through estimating the open-circuit voltage (OCV) during dynamic driving conditions with Kalman and H-infinity filters, a redundant SoC estimation method is demonstrated that compensates for the known drawbacks of CC. To prove the concept under realistic conditions, the model was validated not only on laboratory drive cycles but also on a real-life drive cycle containing measurement noise. Together, these innovations provide a cost-effective and scalable pathway towards robust battery monitoring and reliable SoC estimation in future electric vehicles.