Xray scans of polypod particles piles

These data are associated with the results presented in the article “Origin of geometric cohesion in non-convex granular materials: interplay between interdigitation and rotational constraints enhancing frictional stability” by J. Barés, A. Regazzi, D. Aponte, S. Buonomo, M. Renouf, N. Estrada, and É. Azéma (https://doi.org/10.48550/arXiv.2601.00262).Additional details on the experimental protocol can be found in this publication. The post-processed data correspond to analyses of the raw datasets performed using the numerical tools available in the following repository:https://git-xen.lmgc.univ-montp2.fr/Image/starxrscan/-/tree/main/SingleShot DATA ORGANIZATIONThe raw and post-processed data from 20 experiments are provided. Each experiment is stored in a compressed archive named:{A}_{B}_{Nx}_{Dx}_{C}_{E}.zipwhere:{A} is the chronological order in which the experiment was performed.{B} (PA12, EPDM, PEHD) specifies the particle material, which determines frictional properties.{Nx} is the number of branches of the particles.{Dx} is the branch diameter.{C} (t, nt) indicates whether the confining tube was kept (t) or removed (nt).{E} (v, nv) indicates whether the system was vibrated (v) or not (nv). FOLDER STRUCTURE FOR EACH EXPERIMENTEach experiment archive contains the following directories and files: meta_data/This folder contains the scanning and reconstruction parameter files generated by the X-Act software suite provided by RX-Solutions. slice_y/This folder contains 8-bit PNG images corresponding to vertical slices of the reconstructed density matrix. result/This folder contains the post-processed numerical data: position.npyCartesian coordinates (x, y, z) of the center of each detected particle.Units: pixel sizeDimensions: [3, number of particles] orientation.npyOrientation of each particle branch expressed as azimuthal and polar angles.Units: radiansDimensions: [number of branches, 2, number of particles] fitting_goodness.npyQuality of particle and branch orientation detection obtained from a minimization procedure.Values closer to 1 indicate better detection.Dimensions: [number of particles] clustering_goodness.npyQuality of branch clustering based on the inertia metric from sklearn.cluster.KMeans.Values closer to 1 indicate better clustering.Dimensions: [number of particles] overlap_goodness.npyOverlap between real and virtual particles used to assess detection quality.Values closer to 1 indicate better agreement.Dimensions: [number of particles] limit_xyz.npySpatial limits of the Region of Interest (ROI).Units: pixel sizeFormat: [x_min, x_max, y_min, y_max, z_min, z_max] volume.npyVolume of each detected particle.Units: cubic pixels local_packing_fraction.npyLocal packing fraction of each particle computed from the Voronoi tessellation.Dimensions: [number of particles] voronoi_matrix.npyApproximated Voronoi tessellation matrix. Each element gives the index of the nearest detected particle.The matrix resolution is 15 times coarser than that of the original density matrix. neighbor.npyNumber of neighboring particles located within a distance of three particle radii.Dimensions: [number of particles] process_data.ymlDictionary containing the parameters used for post-processing. coordination.npyNumber of detected contacts per particle using the optimal contact-detection threshold.Dimensions: [number of particles] contact_distance.npyDistance from each detected contact point to the particle center.Units: pixel sizeDimensions: [number of contacts] contact_array.npyPercolation values for each potential contact as a function of the density threshold.A contact is considered real when the percolation value exceeds 1.Dimensions: [number of potential contacts, length of contact_density_threshold] contact_density_threshold.txtVector of density thresholds used for contact detection. figure/This folder contains visualization outputs: movie_contact.aviReconstructed movie of the packing with detected contacts. view_1.pngSnapshot of the reconstructed packing with contacts highlighted.