A study on battery model parametrisation problem: application-oriented trade-offs between accuracy and simplicity

This study is focused on fast low-fidelity battery modelling for online applications. Because the battery parameters change due to variations of battery’s states, the model may need to be updated during operation. This can be achieved through the use of an online parameter identification technique, making use of online current-voltage measurements. The parametrisation algorithm’s speed is a crucial issue in such applications. This paper describes a study exploring the trade-offs between speed and accuracy, considering equivalent circuit models with different levels of complexity and different parameter-fitting algorithms. A visual investigation of the battery parametrisation problem is also proposed by obtaining battery model identification surfaces which help us to avoid unnecessary complexities. Three standard fitting algorithms are used to parametrise battery models using current-voltage measurements. For each level of complexity, the algorithms performances are evaluated using experimental data from a small NiMH battery pack. An application-oriented view on this trade-offs is discussed which demonstrates that the final target of the battery parametrisation problem can significantly affect the choice of the fitting algorithm and battery model structure