35639: Master student Information technologies, Mathematics, Aerospace Engineering or similar - Artificial Neural Networks for Data-Driven Aerodynamic Applications

German Aerospace Center (DLR)

Braunschweig, Germany

Work group:

Institute of Aerodynamics and Flow Technology



Area of research:

Diploma & Master Thesis



Job description:


In the past few years the application of machine learning (ML) methods has increased rapidly throughout various domains including several engineering disciplines. In particular artificial neural networks (ANN) have been heavily investigated for reconstruction, clustering and classification problems. In aerodynamics, the main interest is in applying ML methods to predict different quantities of interest such as the lift coefficient or the surface pressure distributions at changing flight conditions based on just a few, well-chosen highly accurate and expensive simulations. The current approach to achieve such predictions is to combine dimensionality reduction methods and interpolation or regression techniques offering reasonable results for a range of aerodynamic problems. Nevertheless, a smart combination of established approaches with carefully selected ANN might yield another boost in accuracy and efficiency.


The emerging potential offered by ANN should be investigated during this master thesis and compared to existing and more established methods. Therefore the following steps could be seen as a rough guideline:


identification of potential artificial neural network architectures with respect to aerodynamic problems selection of one or a few ANN approaches which seem promising for specific applications demonstration of the performance of these architectures using simplified test cases application to one or several aerodynamic test cases depending on previously drawn conclusions

Please apply via recruiter’s website.

Quote Reference: 35639

Favorite