Evaluation of present accelerated temperature testing and modeling of batteries

Abstract

Battery manufacturers and device companies often test batteries at high temperature to accelerate the degradation process. The data collected from these accelerated tests are then used to determine battery performance and reliability over specified nominal operating temperatures. In many cases, companies assume an Arrhenius model, or prescribe a decade rule to conduct the data analysis. This paper presents the flaws in accelerated temperature testing of batteries using the Arrhenius model and the decade rule, with the emphasis on lithium-ion batteries. Experimental case studies demonstrate the inaccuracy of the Arrhenius model. Alternative methods based on reliability science are then provided.