Li-ion Battery Pack State-of-Charge Estimation Disturbed by Colored Noises

In electric vehicles, onboard lithium-ion batteries need an accurate state of charge (SOC) to enhance safety, improve efficiency, and extend lifetime. However, there are noises to disturb electric quantities of batteries, especially the battery current for SOC calculation. Based on an equivalent circuit model, the battery SOC is well estimated by a residua-sequence-based adaptive extended Kalman filter (AEKF) when the battery current is polluted by colored noises. This adaptive filtering technique was implemented on the experiment data of a real lithium-ion battery pack, the current values of which were contaminated by the non-zero mean Gaussian noise. Results showed that the SOC estimation produced by the proposed AEKF was much more accurate and reliable than that caused by the conventional extended Kalman filter (EKF) in the colored noise environment. (C) 2017 The Authors. Published by Elsevier Ltd.