Coating Feature Analysis and Capacity Prediction for Digitalization of Battery Manufacturing: An Interpretable AI Solution

Battery production line is crucial for determining the performance of batteries, further significantly affecting the industrial applications of relevant energy systems. As a complex and multidisciplinary system involving electrical, mechanical, and chemical processes, efficient prediction of manufactured battery properties and explainable analysis of strongly coupled battery production variables becomes an important but challenging issue for the wider application of batteries. In this article, an interpretable AI solution based on generalized additive model with interactive features and interpretability (GAM-IFI) is proposed to effectively predict battery capacities in the early phase of battery manufacturing and explain the effects of involved coating features. The designed solution is evaluated by using reliable production data from a real battery manufacturing line. Illustrative results show that the proposed solution is able to accurately predict three different types of battery capacities with an <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$R^{2}$ </tex-math></inline-formula> over 0.98. Moreover, information regarding the importance ratio of both main effect and pairwise interaction terms derived from three coating features is identified, while global and local interpretations of the effects of these terms can be well explained. The developed interpretable solution opens a promising avenue to identify the importance of battery production features and explain how the variation of these features influences the properties of battery products. This can help engineers to better understand the underlying complex behaviors in battery production, which in turn will benefit the digitalization of battery manufacturing.