Data-Driven Lithium-Ion Battery Degradation Evaluation Under Overcharge Cycling Conditions

Abstract

Accurately assessing degradation and detecting abnormalities of overcharged lithium-ion batteries is critical to ensure the health and safe adoption of electric vehicles. This paper proposed a data-driven lithium-ion battery degradation evaluation framework. First, a multi-level overcharge cycling experiment was conducted. Second, the battery degradation behaviours and features were analyzed and extracted using incremental capacity analysis and pearson correlation coefficient. Above all, a data-driven lithium-ion battery degradation evaluation method based on machine learning and model integration method was developed. The proposed integrated model was compared with other state-of-the-art methods and reached a mean squared error of 1.26×10 <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">-4</sup> . Finally, based on prediction results, rate of degradation was calculated and classified to different degrees, and overcharged cells can be effectively identified. Moreover, to verify the feasibility of the proposed overall framework, the paper carried out an experiment by connecting overcharge-induced degraded cell and fresh cells in series to simulate the real-world battery assembly and function of battery management systems. Based on the proposed scheme, the overcharged batteries in the battery series can be detected efficiently likewise.