Statistical investigation of temperature-dependent cycle lifetime and cell-to-cell variance in lithium-ion batteries: A model-based approach
Nikolay I. Nikolov, Ahmed Chahbaz, Felix Hildenbrand, Maria Kateri, Dirk Uwe Sauer
Abstract
It is widely recognized that temperature has a significant influence on the cycle lifetime of lithium-ion batteries (LIBs). Although there are several studies in the literature exploring the effect of elevated ambient temperature on the cyclic aging behavior of LIBs, statistically robust conclusions regarding the capacity-temperature relation remain challenging due to the limited sample sizes used in the available experiments. In this work, we perform cyclic aging tests on 48 NCA/Gr-SiOx cells at six temperature levels, ranging from 25 °C to 55 °C. First, we classify the tested cells into two groups with the help of a normal mixture model based on their initially extracted capacity. Then, a temperature dependent regression model is presented and fitted to the capacity and resistance results after 600 and 1200 partial cycles. Our investigation shows, that cycling within an ambient temperature range of 35 °C to 40 °C strikes a balance between achieving the highest mean capacity and minimizing cell-to-cell variance. Furthermore, the presented classification and regression models can be applied to enhance and manage the overall reliability of LIB packs.