OpenFOAM for Combustion Simulations
The hands-on block course OpenFOAM for Combustion Simulations focuses on the fundamentals of numerical modeling of reactive flows. Theoretical concepts are immediately applied in practical computer exercises, and newly acquired knowledge is reinforced through the analysis of simulation results.
The course covers the theoretical foundations of perfectly premixed flames (reactive Navier–Stokes equations, chemical kinetics, flame front kinematics, and thermoacoustics), as well as methods for modeling these phenomena. The practical exercises build on one another and make use of the open-source tools Cantera, OpenFOAM, and ParaView. You will learn how to independently set up, run, and analyze simulations, including 0D reactors, ignition processes, 1D flames, and 2D flames.
In the second half of the course, you will apply the skills and knowledge you have acquired by working on a group project focused on analyzing the physical differences between conventional methane flames and sustainable hydrogen flames.
Prerequisites: The course can be completed without extensive prior knowledge; however, basic knowledge in the following areas is helpful and will make it easier to get started: fluid mechanics, CFD, thermodynamics, combustion, and Python.
The official course language is English, although the teaching staff also speaks German.
Machine Learning for Dynamical Systems
This course offers you a hands-on introduction to machine learning—specifically designed for engineers.
The course is structured as a block course: in the mornings, you’ll learn the theoretical fundamentals, and in the afternoons, you’ll put everything into practice yourself using tools like TensorFlow and Keras. In the first half of the course, you’ll receive an introduction to neural networks, deep learning, recurrent networks, convolutional neural networks, autoencoders, and physics-informed neural networks—always with direct practical application in the computer lab.
In the second half, you will work in groups on a project of your choice:
- Modeling flame dynamics using NNs/RNNs
- Autoencoding of a Kolmogorov flow
- Autoencoding of flame images
- Physics-Informed Neural Networks for incompressible flows
Prerequisite: Basic programming experience with Python
The course is taught in English.
