Abstract
Accurate state of charge (SOC) plays a dominant role in safety control and energy management of battery system. In this paper, an improved adaptive extended Kalman filter (IAEKF) is proposed to realize co-estimation for battery capacity and SOC. Firstly, the online OCV identified by forgetting factor recursive least squares (FFRLS) is innovatively regarded as the observation state, so battery capacity and SOC can be integrated into a second-order filter to realize co-estimation. Secondly, the polynomial relationship between OCV, SOC and T is established to enhance the temperature adaptability of IAEKF. Notably, compared with adaptive extended Kalman filter (AEKF), IAEKF can improve estimation process by producing a fast convergence speed under big OCV errors and assure a slower slope when the OCV errors turn small. Besides, the forgetting factor can simplify the moving window of the adaptive update for the process and measurement noise. With sophisticated Federal Urban Driving Schedule test (FUDS), the estimation accuracy of the proposed algorithm is verified with SOC error band controlled between ±1.2% within the first 50s and relative capacity error limited within 2% after convergence. Furthermore, the verification results also prove that IAEKF based co-estimation algorithm has strong anti-interference ability when facing disturbance, including erroneous initial SOC settings, unknown battery capacity and various ambient temperatures.
| Original language | English |
|---|---|
| Article number | 102559 |
| Journal | Journal of Energy Storage |
| Volume | 40 |
| DOIs | |
| State | Published - Aug 2021 |
UN SDGs
This output contributes to the following UN Sustainable Development Goals (SDGs)
-
SDG 7 Affordable and Clean Energy
Keywords
- Capacity
- Improved adaptive extended Kalman filter
- Lithium-ion battery
- State of charge
- Temperature
Fingerprint
Dive into the research topics of 'Co-estimation for capacity and state of charge for lithium-ion batteries using improved adaptive extended Kalman filter'. Together they form a unique fingerprint.Cite this
- APA
- Author
- BIBTEX
- Harvard
- Standard
- RIS
- Vancouver