Physics-Informed Two-Stage Framework for Small-Data Prediction of Steel Fiber Pull-Out Load-Slip Response of Ultra-High-Performance Concrete
Published: 14 October 2026| Version 1 | DOI: 10.17632/vkzcn9bpnb.1
Contributors:
qiao zhang, Zemei Wu, Xinyan Zheng, Kunpeng Li, Yupu Wang, Kamal H. KhayatDescription
A dataset was compiled with 12 input variables, including fiber shape, embedment length (Le), inclination angle (θ), fiber tensile strength (ft), fiber length (L), diameter (D), aspect ratio (L/D), specimen shape, loading rate (V), water-to-binder ratio (W/B), concrete compressive strength (fc), and curing age (A). The outputs included the peak pull-out load (Pmax), corresponding slip, and complete load–slip curves. A total of 212 experimental datasets were collected from 21 peer-reviewed academic articles, with two additional datasets included for external validation. The original load–slip curves were interpolated into 251 points over a slip range of 0–5 mm. For more information, please see the paper linked to this dataset.
Files
Institutions
- Hunan UniversityHunan, Changsha
- Missouri University of Science and TechnologyMissouri, Rolla
Categories
Data Science, Machine Learning, Concrete Technology, Fiber