Sorting method for retired power batteries – Principal component analysis and clustering based on dynamic parameters

Abstract:
In view of the large scale of retired power batteries and the lack of a fast sorting means, a sorting method is presented based on principal component analysis and clustering. Perform charging and discharging experiments on battery cells, and after obtaining a complete charging and discharging curve, extract characteristic variables from the curve. Using the multivariate statistical method of principal component analysis, feature variables are extracted for principal component extraction. Determine the best K value, and complete the battery sorting and reorganization through the K-means clustering algorithm.