Augmented real-time lumped-parameter model for enhanced reliability in battery management systems

Accurate and computationally efficient battery management systems (BMSs) rely heavily on lumped-parameter models for real-time monitoring and control. However, these models often fail to maintain precision under rapidly fluctuating operating conditions due to limitations in conventional parameter identification methods. This paper presents a new framework that augments a modified equivalent circuit model (ECM) with a lumped-parameter thermal representation of a cell (LPTM) and a reliable real-time parameter estimation algorithm. The core novelty of the framework is the new representation of the cell and the Modified Recursive Least Squares (ModRLS) algorithm, which addresses challenges of data saturation, parameter sensitivity, and initial condition uncertainty. Simulations demonstrate a significant improvement in parameter tracking accuracy, with root mean square errors as low as 3% not only for key electrical parameters but also thermal ones. The proposed framework minimises the reliance on extensive sensor networks, offering a cost-effective and scalable solution for dynamic applications such as electric vehicles. This work lays the foundation for more reliable and longer-lasting energy storage systems through advanced monitoring.