Master’s Thesis: Feature Engineering for Wind Energy Forecasting

29.01.2026, Diplomarbeiten, Bachelor- und Masterarbeiten

You are passionate about applying cutting-edge information technology to solve the energy and climate crisis and would like to work in a vibrant research environment? Then let’s design the energy systems of the future together!

Our research focus:

The Professorship of Energy Management Technologies focuses on the design and evaluation of innovative information technology to integrate more renewable energy into our energy systems and make energy use more efficient. We develop new algorithms and prototypical systems controlling complex energy systems like the electric grid for a sustainable future. These systems coordinate distributed renewable generation like solar and wind, flexible loads like heat pumps and electric vehicles, and distributed energy storage like stationary batteries and hydrogen storage to maximize energy efficiency while keeping the grid reliable and secure. Our research method is engineering-oriented, prototype-driven, and highly interdisciplinary.

For wind energy, one of the most volatile renewable sources, accurate forecasts remain a ma-jor scientific challenge. In the project WindCast we are building holistic forecasting models for both day-ahead and intraday horizons, powered by advanced physics-informed machine learning. By providing open-source tools, WindCast will equip the wind industry to compete confidently in mod-ern markets and accelerate the shift toward a renewable energy future.

Feature engineering addresses a critical bottleneck in wind power forecasting, where raw SCADA and meteorological data suffer from noise, nonlinearity, and redundancy that degrade model performance. This research quantifies all influencing factors, reviews state-of-the-art prepro-cessing techniques, and develops a reproducible modular pipeline to streamline data preparation for training and evaluation.

Research Questions:

• How do different factors influence wind power generation?
• How can feature engineering improve forecast performance, reliability, complexity and computational costs?

Possible Approach:

• Investigate the impact of key factors on forecasting performance
• Research and apply advanced feature engineering methods
• Employ state-of-the-art machine learning models to assess downstream impacts using metrics for predictive accuracy (e.g., MAE, MSE), reliability (e.g., calibration, sharpness), model complexity (e.g., parameter count), and computational costs (e.g., training time relative to baselines)

Our offer:

• Scientific work in a highly motivated research team
• Pleasant working environment and intensive supervision
• Learning to apply modern data science and machine learning methods on real-world challeng-es
• Opportunity to contribute to a scientific publication
• Immediate start possible

Your application:

Please submit your application as one single PDF file via email, containing the following docu-ments:

• Cover letter
• Curriculum vitae
• Academic transcripts

If you are interested, please contact Jonas Betscher at jonas.betscher@tum.de.

Kontakt: jonas.betscher@tum.de

More Information