This page showcases the proof-of-concept for using NASGRO models and machine learning techniques to simulate crack propagation in aircraft structures. The results demonstrate the feasibility of building surrogate models that can optimize airframe design and analysis.
Crack propagation models help predict the behavior of cracks in aircraft structures under various stress conditions. NasgroPy leverages models like:
These models have been integrated with a machine learning-based approach, using methods like Gaussian Process and Neural Networks, to create surrogate models that can predict crack growth efficiently.
The proof-of-concept demonstrates that surrogate models, created using a Design of Experiments (DoE) methodology, can be effectively used to predict crack propagation in real-time. The surrogate models are based on NASGRO simulations and are optimized using machine learning techniques.
You can explore the details of the proof of concept by following the links to individual crack propagation cases:
This proof of concept provides a foundation for further development. In future iterations, the models will be expanded to include more complex crack configurations, additional NASGRO models, and refined machine learning algorithms.