Dataset for experimental design of ply-blocked and helicoidal carbon-fibre laminates for ballistic protection

Dataset for experimental design of ply-blocked and helicoidal carbon-fibre laminates for ballistic protection Description This dataset contains the raw data and processing code behind the paper Experimental design of ply-blocked and helicoidal carbon-fibre laminates for ballistic protection by M. Lubout, L. Amaral and S. G. P. Castro (submitted). Ten symmetric 24-ply carbon/epoxy laminates were tested. They were made of M30SC/DT120 unidirectional prepreg, about 5.1 mm thick, with an areal density of about 7.4 kg/m². The set comprises a cross-ply baseline (L01) and nine ply-blocked, helicoidal, zoned and mixed architectures selected with a Taguchi L9 orthogonal array. The laminates were impacted by a 5.0 mm, 0.51 g steel sphere at the TNO Laboratory for Ballistics Research in June 2026, in 95 shots: 15 shots on L01 to determine its ballistic limit; eight shots per laminate at nominal velocities of 450–520 m/s, to compare residual velocity and dissipated energy. Residual velocities were measured from high-speed video at 40 000 fps, with a side view and a mirror view. The dataset is organised in four folders: 00_Materials: material card, cure cycle, and the mass, size and thickness of every specimen. 01_Test_matrix: Python code for the Taguchi design and the stacking sequences, stacking plots, planned test matrices and the projectile hardness measurements. 02_Experiments: impact and residual velocities of every shot (CSV) and the TNO shot log; 92 high-speed videos (AVI) with camera metadata, the video-analysis code and its calibration; photographs of the test set-up and of the impacted specimens, and a technical drawing of the sabot; pre-impact ultrasonic C-scans (raw matrices and screenshots); cross-sectional micrographs of one specimen per laminate, as printed (JPG) and at full resolution (TIFF, 1.589 µm/px). 03_Paper_reproduction: the shot-by-shot data tables of Appendix B of the paper (Tables B.1 and B.2), and the Python scripts that check them against the raw records and reproduce the paper's statistics, Table 4 and data figures. The statistics are the probit ballistic limit, the Lambert–Jonas fits, and ANCOVA with Dunnett and HC3 intervals. Each folder has a README.md describing its files. The root README.md gives the folder map, the specimen identifiers and instructions for running the code. Table C.1 of the paper links each table, figure and derived quantity to its files in this dataset. Software: Python 3.10 or newer with numpy, scipy, pandas, statsmodels, matplotlib and openpyxl, plus opencv-python for the video analysis. Acknowledgements This research was partially funded by the Ministry of Defence of The Netherlands via Doelfinanciering. This research was conducted within the research and innovation programme Luchtvaart in Transitie, which is co-funded by the Dutch National Growth Fund. It is part of the ‘HAPSS Phase 2’ project, and part of The CrashProof Knowledge Centre (CrashProofLab, doi:10.5281/zenodo.11500722). Metadata Field Value Resource type Dataset Creators Lubout, Mitchell (Delft University of Technology; TNO); Amaral, Lucas (TNO); Castro, Saullo G. P. (Delft University of Technology, ORCID 0000-0001-9711-0991) Keywords ballistic impact; CFRP; ply blocking; helicoidal laminate; Taguchi method; ballistic limit; high-speed video Licence to be confirmed Related identifier is supplement to the DOI of the paper, once published Language English Upload The package has 319 files (4.1 GB: 2.0 GB of AVI videos and 2.0 GB of TIFF micrographs). The upload is the root README.md, unzipped, and one zip archive per top-level folder: 00_Materials.zip 01_Test_matrix.zip 02_Experiments.zip 03_Paper_reproduction.zip Each archive contains its folder at the top level (for example, 01_Test_matrix/README.md), so extracting all four next to the root README.md rebuilds the package. The paths inside the archives are exactly those given in the READMEs and in Table C.1 of the paper.