Lithium plating detection of lithium-ion batteries based on the improved variance entropy algorithm

Abstract

Lithium plating is an important causation leading to capacity loss and thermal runaway of lithium-ion batteries. A detection method and alarm strategy of abnormal lithium plating can mitigate the risk form lithium plating. This paper presents a comprehensive and stable detection method for abnormal lithium plating based on variance entropy. An overvoltage-induced lithium plating experiment has been conduction on a 19-series lithium-ion battery pack, and the abnormal lithium plating was detected by the electrochemical characteristics and variance entropy. The variance entropy can analyze the subtle voltage inconsistencies of cells and first detect abnormal cells in Cycle 13. By comparing the electrochemical characteristics and variance entropy values, the effective application conditions of variance entropy were summarized. Moreover, the variance entropy and voltage platform were combined to extend the application condition and improve the stability. This research can provide the basis for the stable and robust detection method of abnormal lithium plating.