A sorting method for retired lithium batteries based on charge-discharge curve features and improved K-means clustering
In order to improve the consistency of the sorting and reorganization of retired lithium batteries, a sorting method that fusion voltage curve and energy curve, consider the numerical characteristics and morphological characteristics of the curve, and usethe euclidean distance and morphological distance to improve the K-means clustering sorting method is proposed. The charge-discharge curve of retired lithium battery is obtained by experiment, and the voltage curve and energy curve are fused as the sorting basis. The euclidean distance is used to measure the numerical difference of the fusion curve, the morphological distance is used to measure the morphological difference of fusion curve. When obtaining the morphological distance, the fusion curve is first transformed into a feature sequence describing the morphological change of the curve by using the quantile method, and then the morphological distance of the feature sequence is extracted by using the longest common subsequence algorithm. Based on the Euclidean distance of the fusion curve and the morphological distance of the feature sequence, the K-means clustering algorithm is used to cluster the retired lithium batteries. The results show that compared with voltage curve or capacity curve sorting, using fusion curve sorting, the consistency of capacity, charging voltage and discharge voltage is increased by about 23％, 93％ and 16％. Compared with the Euclidean distance method, using the improved K-means algorithm, the capacity, charging voltage, and discharge voltage consistency are increased by about 67％, 40％, and 51％, respectively.
retired power batteryinconsistencyseparation methodimproved K-means
国家科技期刊平台
NSTL
万方数据
