A Data-Driven Fleet Service: State of Health Forecasting of Lithium-Ion Batteries, Friedrich von Bülow, 2023
The dissertation of Dr.-Ing. Friedrich von Bülow deals with the prediction of the state of health (SOH) of lithium-ion batteries using machine learning models. The focus is on the application in the practical context of battery systems in battery electric vehicles (BEVs) using fleet operating data.
We asked Friedrich about his dissertation:
What was the context of your dissertation? What projects or other factors particularly influenced your dissertation?
My dissertation was written as an industrial doctorate as part of the Volkswagen doctoral program at Group Innovation of the Volkswagen Group. This industrial-automotive environment of the Volkswagen Group directly influenced the motivation for the dissertation and the research questions. The discussions with colleagues from the various Group brands were important. Therefore, a central factor of my dissertation from the beginning was the applicability of the methodology in practice. The collaboration and exchange with the other doctoral students and alumni of the VW doctoral program and the other Group brands also helped my work progress positively. Finally, the TMDT completed this diverse ecosystem of different disciplines and perspectives.
.
How does your work contribute to the field of research?
My goal was to make the operating data of battery electric vehicle fleets usable in order to make statements about the aging behavior of the batteries under real vehicle use.
My work focuses on using battery system data from real fleet operation. These differ characteristically from laboratory data of individual battery cells in terms of data volume and variability, which have mostly been used in research to date.
In addition, the model should also be usable in practice by e.g. fleet managers of a battery electric vehicle fleet who want to predict the ageing condition of the batteries of their vehicles and optimize the fleet operation with regard to battery ageing.
What does the future hold for you and the topic?
The results of my dissertation were taken up by the relevant specialist departments in various Group brands and have influenced further development there. Overall, more and more work in this field of research is dedicated to the use of real fleet data, so that some progress can still be expected here. Many questions, such as the continuous training of models with new data (continuous learning), are still open. New developments in battery technology such as new battery chemistries and pack variants will also have to be taken into account.
I look forward to continuing to work with vehicle data and the data-driven vehicle functions and services based on it at Volkswagen.
