Li-ion Batteries Parameter Estimation Using Converter Excitation and Fusion Methods

Abstract

This paper proposes a fusion method to estimate the low frequency impedance of Lithium-Ion batteries by using excitation signals. The proposed excitation is delivered by the converter used for the interface of the battery. The fusion method is conducted in two steps. Firstly, the impedance is estimated by using a frequency method known as vector fitting. Despite its high accuracy, it has a moderate computational burden that makes it difficult to use in real-time application running on the embedded systems often used for battery and converter control. For that reason, in a second stage, recursive least squares algorithm will be used for the parameter estimation in real-time. The discussion includes not only the methods description and testing, but also the effects of the data used for the estimation (terminal current and voltage) as well as the discretization method used for the digital implementation. The proposal is validated by simulation and with preliminary experimental results.

An experimental validation of the adoption of dc-dc converters for the impedance measurement in Li-Ion batteries

“…In [12] a DC current pulse is injected into the battery and the voltage response is collected. The obtained impedance is also sensitive to the amplitude, the pulse width, and its frequency resolution [13]. The unidirectional power converter used in [14] does not allow to inject zero average current and consequently is not used to perform the offline EIS analysis.…”