State of Charge Estimation of an Electric Battery using the Theory of Evidence

Accurate monitoring of the state of charge (SOC) of a battery is crucial in battery management systems. Coulomb counting is one of the easiest and most widely used methods to estimate the SOC of the battery. While it is computationally efficient, its accuracy is limited by the accuracy of the initial estimate of the capacity of the battery, which depends on several factors such as battery's state of health (SOH) and internal conditions, such as leakage. SOC can also be estimated using the open circuit voltage (OCV) of the battery. However, this method is sensitive to the estimate of the OCV and to noise in the measurement system. There are numerous other approaches available in the literature that utilize machine learning techniques. While these are much more accurate, the computational complexity makes them unsuitable in battery management systems operating in real time. This article proposes a method that provides a SOC estimate by combining the SOC estimate given by Coulomb counting and the SOC estimate obtained from the OCV. The combination is achieved using the rules of combination defined in the Theory of Evidence, resulting in a more accurate estimation.