Designing Complex Behaviour: Novel Pathways for Assisting Design Based on Dynamics-Informed Machine Learning in Structural Mechanics

Prof. Dr.-Ing. Merten Stender (Berlin)

Goal of the project

Novel machine learning approaches as design assistants for the nonlinear dynamic behavior of structures under complex transient loads

Preliminary work and contents

Mapping and prediction of nonlinear structural dynamics using deep learning

  • Huge static black box models
  • Data hunger
  • Training duration, resource consumption

Reservoir computer = nonlinear dynamic (computing) systems

  • Inherent dynamics
  • Small amount of data
  • Elimination of the training process
  • Speed, efficiency

Dynamics-integrated reservoir computer

Collaborations & Modules

How can we support other projects?

  • Prediction modules / architectures
  • Computing Power (GPU Server)

How could other projects support our work?

  • Exchange on benchmark cases and challenges
  • Measurement/Data for complex dynamic systems


Prof. Dr.-Ing. Merten Stender
Technische Universität Berlin
Straße des 17. Juni 135 
10623 Berlin
Room H 2024 (Hauptgebäude)


Dr. Manish Yadav


To the top of the page