On-Line State Estimation for Li-Ion Batteries in Traction Applications

Climate change is a fast-growing global threat, mainly caused by increasing of CO2 emissions due to multiple sources, including transport sector. Globally, electric vehicles (EVs) have been identifying as a promising solution to achieve decarbonization in the transport sector. EVs use Lithium-ion (Li-ion) batteries as energy source that require a battery management system (BMS) to ensure their properly operating condition. To work correctly BMS needs to estimate: State of Charge (SoC), State of Health (SoH) and State of Power (SoP); they can be computed or off-line or on-line. The second one is the preferred approach, as it applies to battery pack installed in moving vehicles by using a single model, one for each parameter. In this paper, a hybrid approach based on both model-based methods and artificial neural networks is proposed for real-time estimation of the SoC, SoH and SoP. These parameters are estimated considering effective EV driving conditions. The effectiveness of the proposed model is proved by analyzing its performance in terms of root mean square error (RMSE) and percentage error (MAPE) and comparing it with existing methods in technical literature.