new

Get trending papers in your email inbox!

Subscribe

Daily Papers

byAK and the research community

Nov 20

AeroZoo: A Benchmark of Well-Posed Flow Problems for Foundation Models of Aerodynamics

Neural surrogates for computational fluid dynamics (CFD) are trained and ranked on datasets that each vary one design family in one flow domain. Such datasets cannot test the least that an aerodynamic foundation model must do: predict CFD-quality incompressible flow in seconds for new shapes in new enclosures. On DrivAerML, larger models and longer training barely lower Transolver's error; doubling the training cases cuts it by about 12% up to 400 cases, exhausting the dataset. We introduce AeroZoo, a benchmark and training corpus of 3,815 time-mean solutions to the unsteady Reynolds-averaged Navier--Stokes (RANS) equations with a transition model, across seven geometry families, eight enclosure configurations and Reynolds numbers from 10^3 to 10^8. Each sample poses a well-posed PDE problem in the form a solver receives it: the boundary meshes grouped by condition type, the prescribed values and the fluid properties. The 2,816 samples with a stable time mean form the canonical benchmark: two tasks, engineering metrics and 22 frozen splits, disjoint in geometry, that hold out geometry families, enclosures, boundary-condition layouts and Reynolds ranges. On the random split, boundary-condition tokens cut GeoTransolver's median velocity error from 76.9% to 8.4% of the inlet speed in single-seed runs (uniform inlet flow: 11.0%). ISLA, a transformer for boundary-driven PDEs, reaches 5.3% over four seeds, about 4 times the change from a finer mesh. On the same test samples, a held-out geometry family barely changes ISLA's error relative to uniform flow, and a held-out enclosure or boundary-condition layout raises it 2.4 to 9.2 times. The data and the scoring code will be released.

  • 5 authors
·
Nov 11

ISLA: A Coordinate-Free, Measure-Weighted Transformer for Boundary-Driven PDEs

Symmetry is a foundational principle of physics: the behavior of a PDE-governed system does not depend on the coordinate frame, unit system, or mesh used to write it down, and neither should a surrogate model. Yet today's leading neural surrogates lack these invariances and fail under simple transformations of the same problem: for models trained on PDEs of cars in one pose (i.e., translation and rotation), rotating the full PDE raises the surface-pressure error from 4% to more than 100%, and re-sampling its boundary raises it to about 30%. We introduce ISLA (Invariant Slice Attention), a slice-attention transformer for steady, boundary-driven PDEs. Because the boundary determines the interior solution, ISLA encodes the posed boundary once into learned slices and evaluates arbitrary interior points independently through a tree-accelerated learned boundary kernel. Inputs are nondimensionalized in scales computed from the boundary, sums over boundary cells are area-weighted quadratures that converge under mesh refinement, and geometry enters only as distances and relative angles to equivariant slice anchors, making the model equivariant to rigid motions and reflections up to optional token subsampling and tree clustering. On AeroZoo, 2,816 3D flows spanning diverse geometries, poses, physical scales, and boundary conditions, ISLA's mean reference-scaled velocity and pressure errors are at least 1.9 and 2.2 times lower than any baseline's. On full AirfRANS, ISLA is second only to an explicit boundary-integral network (GLOBE), with errors at least 1.6 times lower than Transolver's and GeoTransolver's in every field.

  • 1 authors
·
Nov 14

PhysicsNeMo-Mesh: GPU-Native Meshes and Simulation Cases for Physics AI

PhysicsNeMo-Mesh is the mesh module of the open-source PhysicsNeMo library. It provides mesh processing similar to PyVista, but on the GPU, in batches, and inside PyTorch's automatic differentiation. Like xarray for unstructured meshes, it attaches semantics to data: fields on points, cells, and whole meshes have names and shapes, and a simulation case keeps the complete PDE problem specification, including boundary conditions and their prescribed values. It addresses three problems that plague machine learning for physical simulation. First, GPU utilization is bottlenecked by I/O, because samples are very large relative to the compute per forward pass. Second, because traditional preprocessing is expensive, it is necessarily done once, during curation, rather than differently for each model and workflow at the point of use. Finally, advanced equivariant and mesh-native architectures require information that curation ordinarily discards. Fast GPU mesh processing makes it practical to preprocess at the point of use: on one GB200 GPU, mesh operations run 6.2--1{,}068times faster than VTK with 72 threads, and loading a whole case is 8.0--88times faster than reading its VTK files (8.0times when both read the same fields). Overlapped with training, this preprocessing costs 2.1% of throughput for a 127,ms Transolver step. So one dataset serves many models and workflows: in PhysicsNeMo's public recipe, one stored DrivAerML dataset feeds Transolver, GeoTransolver, and GLOBE. PhysicsNeMo-Mesh combines the speed of an intermediate array store with the flexibility of an archival format such as VTK. When shapes and wall conditions both vary, per-boundary conditions cut a heat-conduction error from 87% to 6--9%. And by keeping field ranks and named boundaries, it retains the semantic structure that next-generation architectures will need.

  • 4 authors
·
Oct 28