State of Health Estimation of Lithium-ion Batteries based on Indirect Health Indicators and Gaussian Process Regression Model

Abstract

Achieving accurate and reliable battery state of health (SOH) estimation is significantly important to ensure the reliability and safety of the electrical system operation. The capacity and internal resistance are often used as direct health indicators (HIs) for degradation modeling and SOH estimation of lithium-ion batteries. However, it is difficult to directly measure the battery capacity in online applications due to the complex operating environment. In addition, the measurement of battery resistance is very expensive on-line. In this paper, a new method combining indirect HIs and Gaussian process regression (GPR) model is presented for battery health conditions prediction. First, considering the whole charge-discharge process of lithium-ion battery and the influence of temperature, the appropriate and easy to measure indirect HIs are extracted from the curves of current, voltage and temperature. Then, the important indirect HIs are selected to describe comprehensively the aging of battery performance, and the correlation analysis methods are used to analyze the correlation between indirect HIs and SOH. Next, a GPR model is built based on the extracted indirect HIs for battery SOH estimation. Experimental results show that the proposed approach can provide accurate and effective online SOH estimation information of lithium-ion battery.