Evaluation of Accuracy for Automated Method of Predicting Capacity Loss in Li-Ion Batteries

This paper provides an overview, methodology, and accuracy assessment detailing the collaborative efforts between Ridgetop Group Inc. (Ridgetop) and Idaho National Laboratory (INL) in advancing the system architecture of the CellSage battery health modeling, simulation, and analysis software platform to allow end users to adapt the CellSage battery modeling tool to make predictions for batteries for which an end user has necessary baseline cycle life and calendar life aging data. These efforts resulted in the design and development of automated regression analysis methods for baseline battery aging data, aiming to extract essential model parameters for Sigmoidal Rate Expressions (SREs). Such SRE parameters are utilized in CellSage to predict battery aging under reasonable use, such as arbitrary conditions of temperature, state of charge, and other operating conditions related to a specific battery duty cycle. The underlying technology core is based on a patented approach where the sum of SREs has been demonstrated to accurately model the loss of lithium inventory (LLI) and loss of active material (LAM) as the dominant aging mechanisms within lithium-ion batteries. Throughout this paper there is a discussion on various types of battery State-of-Health (SoH) Prediction models, an overview of how capacity loss is calculated with SREs, and a comparative analysis that evaluates the accuracy of the new automated regression analysis methods for eight baseline aging datasets from various battery chemistries and cell types.