Open Circuit Voltage Estimation for Lithium-Ion Batteries using Partial Relaxation Voltage Data

Accurate open circuit voltage (OCV) estimation is critical for estimating the state and ensuring the safe operation of lithium-ion battery systems. As an important indicator of battery's electrochemical behavior, the terminal voltage measurement during the relaxation period, or called the relaxation voltage (RV), offers a practical approach to capturing the relationship between the battery's terminal voltage and OCV. This study introduces a robust and efficient estimation method for battery OCVs across a wide SoC range based on RV data and Gaussian Process Regression (GPR). The proposed method utilizes the strong correlation between RV and OCV, extracting key features from the RV data to train the GPR model for accurate OCV prediction. Experimental test results show that the proposed OCV estimation method only using the RV data during the first 30 seconds can achieve high accuracy with a mean absolute error of 0.9 mV. Additionally, the proposed method demonstrates high accuracy under various SoC levels and operating conditions. Thus, the proposed method featuring high accuracy and low data requirement has the potential to advance online battery management and support sustainable energy applications.