Code, quick-start examples and full documentation: <a href="https://github.com/BaumSebastian/DDACS">https://github.com/BaumSebastian/DDACS</a>.<br><br>The benchmark dataset was generated through a comprehensive simulation study of the deep drawing process for DP600 sheet metal, incorporating variations in geometry, material properties, and process parameters. The simulations were based on modified quadratic cups with a length of 210 mm and a drawing depth of 30 mm. Three distinct base geometries (Concave, Convex, and Rectangular) were derived from a rectangular reference shape, with the geometric parameters (curvature radius, bottom radius, wall angle) varied at their minimum and maximum corner values.<br><br>For each geometry corner, the material and process parameters (material scaling factor, friction coefficient, sheet metal thickness, blank holder force) were systematically varied across their full Cartesian product, yielding approximately 32,000 unique simulations. Each simulation captures the forming operation (OP10) and a subsequent cutting operation (OP20), with stress, strain, thickness, and nodal displacement fields at multiple timesteps including the configuration after springback.<br><br>Each HDF5 file carries named scalar root attributes (geometry, curvature_radius, material_scaling_factor, friction_coefficient, sheet_metal_thickness, blank holder force, etc.) matching the columns of process_parameters.csv. This CSV is the central index mapping every simulation to its parameters and a recommended ML train, test and val split. A single simulation file (258864.zip) is provided for fast preview without downloading the full corner packages. The dataset also includes a complementary RDDAC study covering intermediate geometric and process parameter combinations.<br><br>The HDF5 files use gzip and shuffle compression for compact downloads while remaining readable by standard tooling (h5py, MATLAB HDF5, Julia HDF5). The structured layout supports surrogate modeling, springback prediction, and machine learning benchmarks for sheet metal forming.<br><br>The simulation provenance fields (program, tool deformability, model symmetries, material model, meshing, contact, etc.) follow the SIM-KAx schema (Schumann et al., Production Engineering, 2026) for cross-institution interoperability of forming-simulation datasets.

- Added RDDAC Simulations - Replaces metadata with enriched process_parameters file - Replaces Simulations with bigger geometric range - Added Croissant manifest

LS-Dyna, R14