Student (f/m/x) - ML-Based Composite Panel Prediction for Wing Preliminary Design

What to expect:

Composite structures are key to lightweight aircraft design but require complex optimization to balance stiffness and mass. Conventional analysis of composite panels under varying loads is computationally intensive. By generating a dataset of optimized CLT plates with different dimensions and load cases, a machine learning model can be trained to predict stiffness and mass directly. This enables rapid structural assessments of full wing assemblies without detailed sizing.
The objective of this thesis is to develop and train a machine learning–based surrogate model to predict the stiffness and mass of composite panels as a function of geometry and loading conditions.

You will be responsible for the development of your solution from conception to final implementation. As part of a student job or internship, you will work 10-20 hours per week. In general, it would be desirable to combine this activity with a student research project or thesis.

Your tasks:

• Literature review on composite plate theory and ML applications in structural analysis
• Parametric variation of plate geometry, loading conditions, stringer geometry and panel optimization
• Dataset generation and preprocessing
• Development and training of the ML model
• Validation against reference data set
• Application to wing structural weight prediction
• Documentation and analysis of findings

Your profile:

• Ongoing or completed basic studies in aerospace engineering, computational engineering, mathematics or comparable
• Basic knowledge in machine learning
• Basic knowledge of fibre composite structures and of aircraft design
• Structured way of working
• Willingness to familiarise yourself quickly and largely independently with a new subject area, motivation and initiative