ASB-Elasto2D: An Aircraft Seat Bracket–Inspired Dataset of Elastodynamics with Geometric and Topological Variations, e.g. for GNNs applications

Overview This database, ASB-Elasto2D, contains results from linear elastodynamic simulations performed on 2D geometries inspired by aircraft seat brackets. The dataset includes 640 examples, generated from three distinct topology types with varying geometric and loading configurations. All simulations were carried out using linear T3 finite elements, and time integration was performed via the Newmark scheme. This dataset is particularly suited for benchmarking reduced-order modeling techniques and graph-based machine learning approaches (e.g., GNNs). Matlab Files: asb_i.mat Each asb_i.mat file contains the following data: M: Mass matrix K: Stiffness matrix F: Time-dependent loading term ddlu: Boolean indices indicating free degrees of freedom lt: List of 400 time steps Uref: Primal solution field (displacements) over time The following equation is solved in all simulations: M d2Uref/dt2 + K Uref = F. GMSH Mesh Files: asb_i.msh Each sample also includes a GMSH mesh file (asb_i.msh) describing the geometry and triangulation used for the simulation. These files follow the standard GMSH format and are directly usable for visualization or further meshing operations using open-source tools. Python Reader: transfer_mat2py.py A Python script, transfer_mat2py.py, is provided to facilitate the reading of .mat files into a Python environment. It allows users to load the simulation data (mass matrix, stiffness matrix, loading terms, etc.) into Python for analysis or model training workflows. Python Files: asb_i.pt Each asb_i.pt file is provided in the torch_geometric.data format, suitable for graph neural network applications. Each file contains: Node features: X: Node positions and local stiffness contributions N: Node type (Dirichlet, non-zero Neumann, or zero Neumann) F: Nodal loading term Edge features: edge_attr: Stiffness contributions associated with each edge edge_index: Graph connectivity Output fields: s1: First spatial mode of the displacement field s2: Second spatial mode s3: Third spatial mode This dataset is intended for researchers and practitioners in computational mechanics, model reduction, and geometric deep learning, particularly for problems involving varying topologies and complex boundary conditions.