Refining State-of-Charge Estimation for Battery Energy Storage System using Historical Operating Data

Abstract

In a battery energy storage system (BESS), an accurate estimation of state-of-charge (SOC) is of great significance to prevent batteries from over-charging or over-discharging. However, existing SOC estimator implemented in battery management system (BMS) may suffer from significant error, accumulating along with time. This paper discusses an online approach to refine SOC estimation from BMS, taking advantage of historical operating data. After locating SOC reference point from historical time-series data, the maximum available capacity of charge or discharge is tracked online using a weighted least squares (WLS) formulation. Then, a refined SOC value can be determined by coulomb counting. Based on the operation data from a practical BESS, the proposed SOC refining approach is proved to be effective in providing a more accurate estimation.