Gaussian Process Regression based State of Health Estimation of Lithium-Ion Batteries using Indirect Battery Health Indicators

In a Battery Management System (BMS), the State of Health (SoH) of a battery is a key parameter to be estimated. This parameter shows whether the capacity of the battery has degraded substantially from its original value. However, in online applications, it is difficult to directly measure the capacity of a battery. In this paper, Indirect Health Indicators (IHIs) are extracted from the curves of terminal voltage, current, and temperature during the process of charging and discharging the batteries, which reflect the battery capacity degradation. Out of a number of such indirect indicators, some significant ones are selected as the inputs of an SoH estimation algorithm by applying the Principal Component Analysis (PCA) technique. Based on these, the Gaussian Process Regression (GPR) method is used for the final SoH estimation. The results show that the proposed method has quite high estimation accuracy.