Thesis Work - Machine Learning Modeling of Thermal System Components for BEVs

About this opportunity - Background

This master's thesis work will be carried out within the Thermal & Energy Management Section, which is part of the Energy Department within the Propulsion & Energy Unit, and has the responsibility for the Thermal Management of the complete vehicle. You will work with different teams within Thermal Management, understanding both system and component needs and limitations. Battery Electric Vehicles have essential components like batteries and motors that need controlled heating or cooling to function optimally. The Thermal Management System is made up of heat exchangers, pumps, valves, and fans, and will regulate these temperatures using coolant or refrigerant. The Control System is dependent on the full functionality of its components to maintain required temperatures. This thesis focuses on creating machine learning models to know in advance about the behavior and the life expectancy of those components.

Scope of the thesis work

The student(s) should complete the following tasks:

•    Data Acquisition & Preprocessing – Collect operational and environmental data (temperatures, flow rates, vibration, duty cycles, ambient conditions) from pumps, fans, and valves; handle missing/noisy data.
•    Feature Engineering for Degradation – Extract health indicators such as run-time hours, thermal cycling, and switching frequency to characterize wear and tear.
•    Machine Learning Modeling – Develop predictive models (e.g., survival analysis, regression, or deep learning) to estimate the remaining useful life (RUL) of cooling components.
•    Validation & Evaluation – Compare ML predictions with actual field/lab test data; assess accuracy, robustness, and generalizability across vehicle platforms.
•    Integration & Practical Impact – Propose how such models could be integrated into BEV diagnostics/maintenance systems for predictive maintenance and reduced downtime.

This thesis aims to provide tools to help achieve optimal control by avoiding foreseeable component failures. The goal is to know the RUL of its components to minimize the impact on customers.
Students will have access to plant models to use for the development. They can choose to use MATLAB, Python, C++ or C for the work.

What you'll bring

Your strong theoretical foundation gives you the confidence to be innovative. You approach problems methodically, are self-motivated, and always aim for success. You also put in extra effort to achieve your goals. While the project is designed for two students, one can complete it too. Please share any specific requests you might have in the cover letter. We are looking with people with the following qualifications:

•    Master's students within System Controls and Mechatronics, Mobility/Automotive Engineering, Applied Mechanics, or similar education
•    Software experience and skills in Matlab, Simulink is a must
•    Experience with Software in Loop and Model in Loop techniques is a plus
•    Fluent in English, written and spoken

Duration

•    The work will start January 15 (Dates can be flexible within 1 week).
•    The duration for this thesis work is 6 months.
•    30 ECTS in agreement with your Thesis Advisor at the University
•    Number of Students: Preferably 2 students, but 1 works as well