Experimental and inverse-analysis data for matrix-to-fiber load transfer in hybrid concrete reinforced with BFRP minibars
Description
This dataset accompanies the manuscript "Matrix-to-fiber load transfer in hybrid concrete reinforced with BFRP minibars: Experiments and inverse analysis". It contains experimental measurements, processed DIC/AE data and inverse-analysis results for four mixtures: S1/Run15 (BFRP only), B4/Run10 (BFRP-BF), B9/Run16 (BFRP-PP), and T3/Run11 (BFRP-BF-PP). BF and PP dosages in the hybrid groups are 0.15 vol.% each. Each main mixture has three independent beams from the same casting batch, including one synchronously monitored by DIC and AE and two mechanical repeats. Four BFRP-free control beams are also included. The current release is JBE_Data_20261117_Part01.zip through Part13.zip. Extract all 13 standard ZIP archives into one common directory to combine their data folders; see the deposited README.md and archive_file_map.csv. Its adopted calibration is the 6 November 2026 result before embedment-length compensation: S1-L15, B4-L125, B9-L10 and T3-L71. It contains 475 history frames, 455 equally weighted scored frames and 300 independent fiber records. The package includes measured load-CMOD curves, original mechanical exports and AE acquisition files, compressive-strength tables, COD and neutral-axis inputs, numerical fiber geometry, eight-parameter calibration results, optimization records, constitutive curves and section contributions. Further contents comprise control-beam inversion, parameter profiles, 40 joint geometry-perturbation scenarios (independent root/tip localization within 3-pixel disks combined with independent exposed-length errors in [-0.05, 0.05] mm), eight inclination-resistance scenarios (f=0.2 and 0.5), and design scenarios with r from 0 to 0.50. Current figure numbers and source tables are mapped in figure_source_map.csv. The design analysis originally numbered Figure 23 is now Figure 26. Scenario bands are finite-scenario minima and maxima; solid curves and filled symbols denote the adopted calibration. Matrix and BFRP contributions are model estimates. No fracture-surface photographs, annotated photographs, manuscript files or reviewer-response documents are included. The complete continuous raw DIC acquisition (approximately 2 TB) is not included; supplied DIC arrays are a subset. README.md describes the scope, units, verified reproduction entry points and remaining source-coverage notes. File inventories and SHA-256 checksums are included. Historical published versions retain earlier calibrations and should not be mixed with this release.
Files
Steps to reproduce
1. Download JBE_Data_20261117_Part01.zip through Part13.zip and extract ALL 13 standard ZIP archives into the SAME directory, preserving and combining the data folders. No binary concatenation is required. Verify the downloads using archive_checksums.sha256 and consult archive_file_map.csv. Read the deposited README.md, then data/README.md, group_code_mapping.csv and figure_source_map.csv. 2. Use 01_Adopted_Inversion for the adopted calibration. The current analysis uses 475 history frames, of which 455 have equal weights in the scored objective, and 300 fiber records. 3. Experimental mechanical/AE tables are in 02_Mechanical_and_AE; original exports are in 05_Raw_Mechanical_and_AE_Data. Numerical fiber geometry is in 03_Fiber_Geometry; current COD and neutral-axis inputs are in 04_DIC_COD_Neutral_Axis. 4. Run python verify_release.py from data to check counts, calibration identities, scenario coverage and file hashes. 5. Run the numerical replot scripts within 07_Design_Fig26, 08_Inclination_Resistance and 12_Geometry_Perturbations. The Figure 26 script requires an --output directory. These reproduce saved numerical curves rather than rerunning inverse optimization. 6. To regenerate the 40 joint geometry inputs, run python 14_Model_Source/perturbation/joint_pixel_scenarios.py --output <new_directory> from data. Endpoint positions are sampled independently within 3-pixel disks and exposed lengths receive independent uniform errors in [-0.05, 0.05] mm (seed 20261110). The 40 geometry tables and 3000 recorded draws were regenerated byte-for-byte against the executed inputs. The eight f=0.2/0.5 inclination scenarios remain unchanged. 7. Consult 06_Optimization_Records, 09_Control_Beam_Inversion, 10_Section_Contribution_Source and 11_Parameter_Profiles for associated results. Full inverse-model source snapshots are in 14_Model_Source; some historical environment paths require adaptation. A fresh complete inversion and raw-image DIC processing were not rerun during packaging. 8. Preserve units and missing-value flags. Scenario bounds represent minima/maxima across computed scenarios, not confidence intervals. See 13_Validation/SOURCE_COVERAGE.md for supplementary source tables not yet located.
Institutions
- Tongji UniversityShanghai, Shanghai
Categories
Funders
- National Natural Science Foundation of ChinaGrant ID: 52378408
- Key Science and Technology Program of the Department of Transport of Yunnan ProvinceGrant ID: Yun-Jiao-Ke-Jiao-Bian [2024] 13
- National Key Research and Development Program of ChinaGrant ID: 2023YFC3806705