A data mining approach is proposed for evaluating the effects of battery production factors in cathode coating stage on both battery capacity and internal resistance for the first time. Specifically, an effective neural network model is built based on real data form designed experiments for obtaining reference cathode coating for coin cells. The purpose is to analyze and predict how the battery quality in both charge and discharge scenarios changes with respect to the key factors of coating including its weight and thickness. The results highlight the correlation between mentioned factors and battery quality indices, which could guide manufacturer to identify efficient ways for producing high-quality batteries.
Keywords Battery manufacturing, Data mining, Capacity, Modelling