A K-Value Dynamic Detection Method Based on Machine Learning for Lithium-Ion Battery Manufacturing

During the manufacturing process of lithium-ion battery, metal foreign matter is likely to be mixed into which seriously influences safety performance battery. In order reduce outflow such defect cells, production line universally adopted K-value test process. traditional test, detection threshold determined empirically, has poor dynamic characteristics and probably leads missing or false detection. Based on comparing screening effect different machine learning algorithms for data this paper proposes a algorithm cell based local outlier factor algorithm. The analysis results indicate that proposed method can adaptively adjust threshold. Furthermore, we validated its effectiveness through implantation experiment conducted in pilot line. Experiment show method’s rate improved significantly. increase defects beneficial improving battery quality safety.