Machine-Learning-basierte Modellierung in der Strukturdynamik

What it is about

You will learn how machine learning is used for modeling in structural dynamics and vibroacoustics. In the end you will be able to apply mathematical and numerical methods for data-driven modeling, perform model order reduction, and use ML models (including neural networks / deep learning) to predict system behavior as well as assess their possibilities and limitations.

What you will be able to do

• Apply mathematical and numerical methods for ML models
• Carry out data-driven modeling
• Use model order reduction and spectral decomposition
• Understand and use dynamic mode decomposition
• Apply artificial neural networks and deep learning
• Assess the possibilities and limitations of ML-based modeling

What the module consists of

• VorlesungConveying the theoretical foundations of machine-learning-based modeling
• ÜbungComputation exercise for applying the methods; problems are solved in advance and discussed in the exercise

Teaching method

• Vortrag / Tafelanschrieb / Tablet-PC mit BeamerExplanation of the theory and demonstration of content
• RechenübungPractical application, clarification of open questions and numerics with MATLAB and ANSYS
• Script via TUM-MoodleProvision of lecture and exercise materials