| About this Abstract |
| Meeting |
2024 ASC Technical Conference, US-Japan Joint Symposium, D30 Meeting
|
| Symposium
|
2024 ASC Technical Conference, US-Japan Joint Symposium, D30 Meeting
|
| Presentation Title |
Predicting Macroscopic Mechanical Properties of Fiber Reinforced Composites Based on Microscale Fiber Morphologies |
| Author(s) |
Jamal F. Husseini, Scott E. Stapleton, Farhad Pourkamali-Anaraki |
| On-Site Speaker (Planned) |
Jamal F. Husseini |
| Abstract Scope |
Fiber reinforced composites are ideal for structural components where high strength and stiffness to weight ratios are desired. While the benefits are known, composites are not always used in practice because it is difficult to predict how microscale variability may impact macroscale mechanical performance under loading. Variability, such as randomness in the microscale fiber arrangements, may introduce localized regions where failure can initiate and propagate sooner than expected. Capturing this variability and simulating mechanical failure using computer models is typically time and computationally expensive, resulting in models that do not capture large features formed by fiber arrangements such as fiber clusters or matrix pockets. The scope of this work is to examine how features formed by microscale fiber arrangements impact mechanical performance under varied loading conditions. Artificial microstructures with tailored fiber morphologies were generated using a machine learning model. These microstructures were then simulated under varied loading conditions using a novel reduced order, finite element-based model which can more efficiently simulate large microstructures for stiffness, strength, and fracture toughness. Through the efficiency of these tools, large quantities of microstructures can be generated and simulated, and an understanding of how variations in fiber arrangements impact mechanical response can be examined. Results of this study aim to show these complex relationships and a way to predict mechanical properties based on variations in microscale fiber morphologies using a novel machine learning method. |
| Proceedings Inclusion? |
Definite: Post-meeting proceedings |