Temperature sensor-free parameter and state joint estimation for battery pack in electric vehicles

Abstract

Model based state estimation methods rely heavily on a relatively stable open-circuit-voltage (OCV)-state-of-charge (SOC) curve which is described as temperature dependent in traditional definition. However, the temperature of not every cell can be measured accurately in EVs, which restricts their application. To decouple the influences of temperature on battery characteristics, this paper proposes a new method to describe the relation between OCV and SOC, which simplifies a traditional four-dimensional mapping between temperature, aging, OCV and SOC into a temperature-independent three-dimensional mapping. These newly-defined capacity and OCV-SOC curve are independent of battery temperature, which are verified based on a large number of test data under different temperatures and aging conditions. A cooperative estimation of model parameters and state based on dual extended Kalman filter (DEKF) algorithm for a battery pack is developed by building a parameter database of noise characteristics into DEKF algorithm. Battery experimental and hardware-in-the-loop platform is built to verify the performances of the proposed method for battery management systems (BMSs) in all climate EVs. The results indicate that the proposed method can accurately estimate battery SOC with the error within 3 % after fast convergence regardless of battery initial SOC and capacity degradation.