Master Thesis - Fault Detection in Heat Pumps Using Transfer Learning on Time-Series Encodings

The Fraunhofer-Gesellschaft (www.fraunhofer.com) is one of the world's leading organizations for application-oriented research. 75 institutes develop pioneering technologies for our economy and society – more precisely: 32 000 people from technology, science, administration and IT. They know: Anyone who comes to Fraunhofer wants to and can make a difference. For themselves, for us and for the markets of today and tomorrow.

The Fraunhofer Research Institution for Energy Infrastructures and Geotechnologies IEG carries out research in the fields of integrated energy infrastructures, geothermal energy and sector coupling for a successful energy transition at seven locations. We develop ideas, technologies and strategies for the next phase of the transformation of the energy systems and see ourselves as independent pioneers in politics, economics, regulation and society. By founding Fraunhofer IEG, the Fraunhofer Society is making a significant contribution to exploiting the markets in a more targeted way for the use of geothermal energy systems, the storage of energy sources and technologies to couple the energy sectors of heat, electricity and transportation.

In light of the increasing importance of energy efficiency and sustainability, the heating and cooling sector is under pressure to enhance reliability and minimize operational costs. Heat pumps, as a critical technology for energy-efficient heating and cooling, can experience faults that affect their performance and energy consumption. Effective and timely fault detection is essential to maintain optimal operation and reduce downtime.
In this context, the application of transfer learning on time-series holds significant potential. By converting heat pump time-series data into image formats, we can leverage pre-trained vision models to improve fault detection accuracy. This approach not only enhances the detection process but also reduces the need for extensive labeled datasets.
The main goal of this project is to develop and evaluate a transfer-learning approach that converts heat pump time-series data into images and fine-tunes pre-trained vision models for accurate and timely fault detection.

What you will do

Literature review on methods for time-series data encoding into image formats suitable for fault detection.
Exploration of existing transfer learning techniques in the context of image classification and fault detection.
Implementation of various time-series-to-image encoding methods (e.g., GAF, MTF, Recurrence Plots) and evaluation of their effectiveness for fault detection.
Training and fine-tuning of pre-trained vision models (e.g., ResNet, Vision Transformer) on the generated image datasets.
Conducting experiments to assess the performance of different models and encoding techniques.

What you bring to the table

Enrolled Student in Engineering, Mathematics, Computer Science, or other related STEM programs.
Strong understanding of mathematical modeling and simulation of dynamical systems, with a focus on time-series analysis.
Experience with machine learning and deep learning methods.
Proficient programming skills in Python, with experience in libraries such as TensorFlow or PyTorch for implementing deep learning models.

What you can expect

Practice-oriented work environment that complements your studies with an attractive remuneration.
Supervisors who will strengthen and support you to become successful.
Targeted and individual guidance and mentoring.
Well-equipped technical infrastructure, flexible working hours and possibility to work remotely.

We value and promote the diversity of our employees' skills and therefore welcome all applications - regardless of age, gender, nationality, ethnic and social origin, religion, ideology, disability, sexual orientation and identity. Severely disabled persons are given preference in the event of equal suitability. Remuneration according to the general works agreement for employing assistant staff.

With its focus on developing key technologies that are vital for the future and enabling the commercial utilization of this work by business and industry, Fraunhofer plays a central role in the innovation process. As a pioneer and catalyst for groundbreaking developments and scientific excellence, Fraunhofer helps shape society now and in the future.

Interested? Apply online now. We look forward to getting to know you!

If you have any questions about this position, please contact

Mehran Ahmadpour

Contact via Mail

If you have any questions about the application process, please contact

Philipp Steinborn

Contact via Mail

Phone: +49 355 35540 172

Fraunhofer Research Institution for Energy Infrastructures and Geotechnologies IEG

www.ieg.fraunhofer.de

Requisition Number: 81756 Application Deadline: 11/30/2025

Job Segment: Sustainability, Computer Science, HVAC, Training, Engineer, Energy, Technology, Operations, Education, Engineering

Apply now »