Sequential analysis of ageing on lithium-ion batteries

• Applicant: Dr.-Ing. Philipp Dechent
• Subject Area: Electrical Energy Systems, Power Management, Power Electronics, Electrical Machines and Drives
• Term: since 2022
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 511349305

Current lithium-ion cells do not yet meet some applications' required power and energy densities. Therefore, research on new materials is dynamic, and ageing behaviour is an essential part of the evaluation. Furthermore, new applications for batteries with particular requirements - such as the electrification of ships, trucks, aircraft, tractors and construction machinery - often bring new load profiles, which will most likely induce a different ageing behaviour. Therefore, adapted test and evaluation methods are necessary for a reliable lifetime prediction. Batteries are highly complex systems with physical, chemical and electrical effects taking place simultaneously requiring a lot more effort to accurately model these effects themselves. In the short and medium term therefore data-driven and empirical models will be used. These methods include diagnostic methods such as deep learning or neural networks based on systematically generated data from accelerated ageing tests in the laboratory. The ageing of lithium-ion batteries depends on the complex interaction of numerous stress factors such as current rate and temperature, which necessitates an extensive test matrix. In addition, the transfer of test results to new batteries with varying materials and dimensions to create models for new cells is very limited. Currently, complex testing is carried out on a small scale on random samples due to the lack of testing resources. This limits the scope of a test regarding the number of different stress factors, the resolution of the influence, and the statistical aspects of cell-to-cell variation. All battery degradation prediction is limited by the amount of data available for either creating empirical models or parameterising physics based or data-based models. Furthermore, due to the vast parameter space of stress factors influencing battery degradation, tests only can provide meaningful data when those stress factors are consistently considered. When a test is finished, and the end-of-life of the cell is reached, the equipment gets free again, and a question arises: Should you do more of the same or do different stress factors, in order to get the maximum information in a given time? This project will cover the research gap identified through the following main research objectives:1. Establish the design space of battery testing stress factors and benchmark the impact on the battery performance and lifetime.2. Determine rules for the design of experiment of large-scale ageing tests with more than 1000 cells, including a quantifier of the incremental usefulness of each measurement.3. Develop a tool that automatically creates an optimised test matrix and updates it continuously while measuring with a feedback loop.

• DFG Programme: WBP Fellowship
• International Connection: United Kingdom
• Host: Professor David Howey, Ph.D.