IPE 14-2022 Bachelor / Master Thesis
Smart energy: Federated Learning for secure load forecasting

Organisationseinheit

Institut für Prozessdatenverarbeitung und Elektronik (IPE)

Ihre Aufgaben

The energy transition and the resulting expansion of renewable energy resources increasingly pose a challenge to the energy system due to their volatile and intermittent nature. In this context, energy management systems are central as they monitor and forecast energy flows. In order to predict future energy flows, suitable machine learning models need to be selected and trained with smart meter data. However, using smart meter data for load forecasting is challenging due to data privacy requirements. One solution to improve data privacy is federated learning, as data is kept private, and only the trained machine learning models are merged and updated on a global server. In your thesis, you analyze how federated learning can improve data privacy and accuracy for load forecasting.

Your Thesis:

• Review: You will review the state-of-the-art research on artificial neural networks, including deep neural networks, recurrent neural networks, and transformers. Further, you consider publications in the areas of load forecasting and federated learning.
• Architecture concept: Based on your research, you propose a federated learning architecture for load forecasting.
• Implement: You will implement your concept with a selected algorithm on real datasets.
• Evaluate: You will thoroughly evaluate the experiments.

Eintrittstermin

Immediately

Ihre Qualifikation

• You study Computer Science, engineering, industrial engineering, or a related course of study
• You are deeply interested in topics such as artificial intelligence, deep learning, energy systems, or load forecasting
• You are able to read and write scientific texts in English or German
• You already have some experience in Python
• You show an above-average degree of initiative and commitment, as well as a thorough way of working