Woven Composite

Woven composites are being intensely studied to understand and predict their complex mechanical behavior and optimize their design. Current research employs machine learning, particularly recurrent neural networks and physics-constrained neural networks, often incorporating transfer learning to address data limitations and improve prediction accuracy of material properties under various loading conditions. Advanced image processing techniques, such as panoptic segmentation of X-ray computed tomography scans, are enabling automated 3D modeling of woven structures, facilitating more efficient analysis and design optimization. This work has significant implications for diverse applications, ranging from aerospace and automotive industries to textiles and water filtration, by enabling the creation of lighter, stronger, and more precisely tailored composite materials.