Master’s thesis
© IKVTopic of the work:
Today, FE simulations provide detailed information about the time-dependent behavior of short-fiber-reinforced thermoplastics, but are often too computationally intensive for fast development cycles.
Therefore, this thesis aims to develop a data-driven surrogate model that efficiently predicts macroscopic stress-strain responses from micromechanical RVE data while considering physical boundary conditions. The focus is on viscoelastic effects and the question of how load history, material condition and model fidelity can be combined in a learning-based approach.
The work is related to this research project:
In the associated research project, an FE-based, physics-informed substitute model for the non-linear behavior of fiber-reinforced plastics is being developed. The starting point is micromechanical simulations at RVE level, from which macroscopic stress-strain curves and internal stress fields are derived. The aim is to make this information usable for fast and robust predictions in industrial design loops.
Objective:
The aim of your work is the development of a physically-informed surrogate model for the prediction of the time-dependent behavior of fiber-reinforced plastics.
Your task:
| For a Bachelor's thesis, you will work on the following tasks | For a Master's thesis, you will work on the following tasks |
| Familiarization with the basics of time-dependent material models and current approaches to physics-guided machine learning | Familiarization with the basics of time-dependent material models and current approaches to physics-guided machine learning |
| Preparation and structuring of existing FE data | Development of a method for determining physically supported points for the generation of virtual data |
| Implementation of a surrogate model taking into account physical constraints | Generation, preparation and structuring of FE data |
| Validation of the model in terms of accuracy and computing time using defined training, validation and extrapolation cases | Development of a surrogate model based on recurrent networks, taking into account physical constraints |
| Validation of the model in terms of accuracy and calculation time using defined training, validation and extrapolation cases |
Your profile:
- Very good analytical skills and enjoyment of technical issues
- Basic knowledge of numerical simulation or finite element methods (FEM)
- Initial experience in data analysis or Python-based evaluation is advantageous
- Ideally initial experience in the field of plastics processing or fiber-reinforced plastics
- Technical degree course (e.g. mechanical engineering, SiSc, CES, etc.)
These are your benefits:
- Collaboration on a cutting-edge research topic at the interface of simulation and AI
- Independent work with intensive support
- Co-design of current research projects
- Varied mix of material modeling, data analysis and simulation
- Direct insight into the construction of modern surrogate models for industrial applications
- Immediate start possible
Has this vacancy piqued your interest? Then get in touch with me by e-mail or phone and send me your application documents.