An adaptive remaining energy prediction approach for lithium-ion batteries in electric vehicles

• An adaptive SoE estimation approach is established.
• A data-driven model is established for SoE estimation.
• The forgetting factor RLS method is employed for parameter identification.
• Dynamic current and temperature profiles are performed on the LiFePO4 cells.

With the growing number of electric vehicle (EV) applications, the function of the battery management system (BMS) becomes more sophisticated. The accuracy of remaining energy estimation is critical for energy optimization and management in EVs. Therefore the state-of-energy (SoE) is defined to indicate the remaining available energy of the batteries. Considering that there are inevitable accumulated errors caused by current and voltage integral method, an adaptive SoE estimator is first established in this paper. In order to establish a reasonable battery equivalent model, based on the experimental data of the LiFePO4 battery, a data-driven model is established to describe the relationship between the open-circuit voltage (OCV) and the SoE. What is more, the forgetting factor recursive least-square (RLS) method is used for parameter identification to get accurate model parameters. Finally, in order to analyze the robustness and the accuracy of the proposed approach, different types of dynamic current profiles are conducted on the lithium-ion batteries and the performances are calculated and compared. The results indicate that the proposed approach has robust and accurate SoE estimation results under dynamic working conditions.