Modeling Lithium-Ion Batteries Using Machine Learning Algorithms for Mild-Hybrid Vehicle Applications

The prediction of voltage levels in an automotive 48V mild hybrid power supply system is safety-relevant while also enabling greater efficiency. The high power-to-energy ratio in these power supply systems makes exact voltage prediction challenging, so that a method is established to model the behavior of the lithium-ion batteries by means of a recurrent neural network. The raw data are consequently pre-processed with over- and undersampling, normalization and sequentialization algorithms. The resulting database is used to train the constructed recurrent neural network models, while hyperparameter tuning is carried out with the optimization framework optuna. This training methodology is performed with two battery types. Validation shows a maximum error of 2.34 V for the LTO battery and a maximum error of 3.39 V for the LFP battery. The results demonstrate performance of the proposed methodology in an appropriate error range for utilization as a tool to generate a battery model based on available data.     «

The prediction of voltage levels in an automotive 48V mild hybrid power supply system is safety-relevant while also enabling greater efficiency. The high power-to-energy ratio in these power supply systems makes exact voltage prediction challenging, so that a method is established to model the behavior of the lithium-ion batteries by means of a recurrent neural network. The raw data are consequently pre-processed with over- and undersampling, normalization and sequentialization algorithms. The r...     »