Equivalent Circuit Model Generation for Batteries Using Non-ideal Test Data

Abstract

Modeling is a key component in the development of battery products. While there are multiple levels of complexity which may be achieved in model development, equivalent circuit modeling is able to quickly produce reliable and accurate predictions for battery behavior. While the use of equivalent circuit models has been described in great detail for lithium ion batteries, it is also desirable to use this methodology regardless of chemistry, specifically with respect to lead-acid technology. When developing battery models for predicting battery behavior in a vehicle, the testing methods meant to mimic vehicle applications often cause non-ideal data for model generation. Specifically, periods of constant voltage charging can limit the model’s capabilities and accuracy. This is due to the imposed voltage limit required for constant voltage charging which is not an inherent battery behavior. By thoroughly examining equivalent circuit models of increasing complexity, it is shown that lead-acid and lithium ion batteries behave similarly so that minimal impact is had on model development. Additionally, three methods are considered for modifying the fitting process so that test data which contains voltage limits may still be considered useful for model development.