Validate fiber length distribution by comparison with experiment
Validation: FiberFind
This validation study was performed using a CT-scan of a polypropylene sample reinforced with 40% long glass fibers. Fiber identification was carried out using the artificial intelligence–based workflow implemented in GeoDict (FiberFind), enabling reliable segmentation and reconstruction of individual fibers from the voxelized microstructure.
Based on the identified fibers, key statistical fiber characteristics were computed, including the fiber diameter distribution, fiber orientation distribution, fiber curvature and curliness, as well as the fiber length distribution.
For validation, the computed fiber length distribution was compared with independent experimental reference data obtained from the FASEP system. The resulting distributions showed very good agreement, demonstrating that the AI-based fiber identification provides accurate fiber reconstructions and enables quantitative characterization of fiber-reinforced materials.
