Impedance technique combined with supervised algorithms-based internal degradation state classification and its economic analysis for safety in retired battery pack

Abstract

As the growth of lithium-ion batteries (LIBs) applications accelerates, future problems for waste LIB disposal are also appeared. When LIB reaches 80 % of the initial capacity, it is defined as the End-of-life (EOL). Since power and capacity of the LIB abruptly decreases at this point, it is determined that the LIB can't be used. However, in the case of the battery pack, even if it reaches 80 % of its initial capacity, it can be reused with secondary applications such as energy storage devices and uninterruptible power supplies (UPS). Therefore, in order to reduce environmental problems, such a battery pack reuse technique is required. However, as the internal degradation state of the LIB differs depending on the operating environment, research is needed to select the LIB in the same state through classification algorithm and configure the waste battery pack. Since the internal degradation of LIB cannot be evaluated qualitatively with existing technology, like capacity and internal resistance, so a new qualitative evaluation factor is required. This research classified an internal degradation state according to LIB environment using an impedance technique and supervised algorithms (Decision tree, Support vector machine, K-nearest neighbor) that could qualitatively represent LIB degradation mechanism. And it is validated that new parameters are good at classifying operating environment compared with existing parameters. Also, Economic analysis was performed for comparison with existing methods, and as a result, it was proved that it showed superior performance compared to existing techniques in both classification accuracy and economic analysis.