Model-Based Software Development for Predictive Battery Temperature Estimation

In modern vehicles, developing reliable and efficient software systems is crucial for optimal performance, safety, and cost-effectiveness. This paper suggests a model-driven method for developing automotive software, particularly emphasizing the creation of a predictive system for estimating battery temperature. The proposed approach employs Simulink models and the KF algorithm to develop a software system for the temperature control of the battery in electric cars. The accuracy and performance of the KF algorithm are assessed by comparing it with the EKF method for non-linear battery models. Furthermore, the reliability and performance of the proposed approach are assessed through MATLAB Simulink MIL, SIL, and PIL testing. The study findings suggest that the EKF is recommended for estimating battery temperature in rapidly fluctuating systems with non-linear behavior, where accurate estimation is critical. The developed system showed good scalability, efficiency, and outperformed other methods by providing more precise and timely temperature estimates for the battery.