Finding a better fit for lithium ion batteries: A simple, novel, load dependent, modified equivalent circuit model and parameterization method

Equivalent circuit models (ECM) of lithium ion batteries are used in many applications because of their ease of implementation and low complexity. The accuracy of an ECM is critical to the functionality and usefulness of the battery management system (BMS). The ECM accuracy depends on the parametrization method, and therefore different experimental techniques and model parameter identification methods (PIM) have been widely studied. Yet, how to account for significant changes in time constants between operation under load and during relaxation has not been resolved. In this work a novel PIM and modified ECM is presented that increases accuracy by 77.4% during drive cycle validation and 87.6% during constant current load validation for a large format lithium iron phosphate prismatic cell. The modified ECM uses switching RC network values for each phase, which is significant for this cell and particularly at low state-of-charge for all lithium ion batteries. Different characterisation tests and the corresponding experimental data have been trained together across a complete State-of-Charge (SoC) and temperature range, which enables a smooth transition between identified parameters. Ultimately, the model created using parameters captured by the proposed PIM shows an improved model accuracy in comparison with conventional PIM techniques.

Automated summary

This study introduces a novel parameter identification method and modified equivalent circuit model for lithium ion batteries, showing significant improvements in accuracy during drive cycle and constant current load validations for a large format lithium iron phosphate prismatic cell. The modified ECM considers switching RC network values for each phase, resulting in better accuracy especially at low state-of-charge for all lithium ion batteries. Training different characterization tests and experimental data across a complete State-of-Charge (SoC) and temperature range enables a smooth transition between identified parameters, leading to an improved model accuracy compared to conventional techniques.