https://surrogatemodels.com/ News, papers, datasets, and funding and deals in Physics AI for engineers: Structures & crash. en https://surrogatemodels.com/papers/r-bilogohttps://surrogatemodels.com/papers/r-bilogoThu, 12 Nov 2026 09:00:00 GMTR-BiLoGo predicts the change in displacement and thinning caused by ribs on the side walls of hot-stamped corner components, from the baseline response and the geometric perturbation of the ribs; the local-feature effect is learned on one global geometry and transferred to others.Structures & crash https://surrogatemodels.com/papers/physicsbenchhttps://surrogatemodels.com/papers/physicsbenchMon, 09 Nov 2026 09:00:00 GMTA benchmark and leaderboard that ranks 66 models on nine datasets, expanded into 28 configurations over seven generation and prediction tasks in 1D, 2D and 3D, at data scales from S to XL, with a metric suite and a PageRank-based ranking (BenchRank) that debiases correlated metrics.Aero & CFDStructures & crash https://surrogatemodels.com/papers/d-jepahttps://surrogatemodels.com/papers/d-jepaMon, 09 Nov 2026 09:00:00 GMTA JEPA-style encoder computes a 128-number latent from the geometry alone; the operating conditions enter only through small FiLM-modulated neural-field decoders, one per physics. A training loss forces the design parameters to be linearly readable from the latent, and case-level variance terms stop all geometries collapsing to the same code.Aero & CFDStructures & crash https://surrogatemodels.com/news/luminary-shift-nvhhttps://surrogatemodels.com/news/luminary-shift-nvhTue, 27 Oct 2026 09:00:00 GMTLuminary Cloud presents SHIFT-NVH, a large physics model trained on finite-element solves of the 2010 Toyota Yaris body-in-white from the public NCAC crash model. It takes the gauges of 15 parts, the material and the vehicle speed and returns RMS acceleration in x, y and z at all 130,713 surface nodes under ISO 8608 class B road excitation. Trained on 840 of 1,212 cases (100 Latin-hypercube designs plus baseline, in steel, aluminium and magnesium, at four speeds), it reaches a median relative L2 field error of 1.1% (worst 2.6%) on 192 held-out cases. The mesh is fixed, so the model learns how thickness and material redistribute the response, not geometry changes.Structures & crash https://surrogatemodels.com/news/luminary-coupled-lpm-aerostructuralhttps://surrogatemodels.com/news/luminary-coupled-lpm-aerostructuralMon, 26 Oct 2026 09:00:00 GMTLuminary Cloud replaces the solvers in a static aeroelastic loop on a transonic swept wing. With SHIFT-Wing (pressure and shear on the surface) coupled to the TACS structural solver, a converged loop took 8 to 11 minutes on one NVIDIA T4 against about 4 hours for the CFD loop on three H100 GPUs; on three wings both converged in four iterations and the predicted lift reduction differed by 0.2 to 0.8 percentage points. A second model, SHIFT-Wingbox, trained on about 1,400 TACS solutions, then replaces the structure, and the two surrogates agree with the SHIFT-Wing and TACS loop within 1 percentage point of lift reduction and 0.04° of tip twist.Aero & CFDStructures & crash https://surrogatemodels.com/news/physicsx-plaid-tensile2d-agenthttps://surrogatemodels.com/news/physicsx-plaid-tensile2d-agentTue, 11 Aug 2026 09:00:00 GMTDouglas Boubert of PhysicsX describes how a coding agent reverse-engineered Tensile2D, a benchmark in Safran's PLAID collection of two-dimensional quasistatic structural simulations generated with the Z-set finite-element solver. The benchmark gives the mesh, the applied pressure and five material parameters and withholds the constitutive law. Working from the 500 labelled training simulations, the agent wrote its own simulator and refined the hardening law until its error on held-out samples was around 10⁻⁸ on all eight metrics, at a median of 0.437 seconds per simulation. PhysicsX reports that the simulator ranks first on the public leaderboard with zero error on every metric, and that the agent also inferred, after 85 hypotheses, that only the even-indexed test samples are scored publicly. The post concludes that public benchmarks which withhold simulation details are vulnerable to capable agents.Structures & crash https://surrogatemodels.com/datasets/shift-crashhttps://surrogatemodels.com/datasets/shift-crashWed, 01 Apr 2026 09:00:00 GMTAbout 5,200 OpenRadioss simulations of a 56 km/h frontal impact into a rigid wall by a 2010 Toyota Yaris model morphed over four body dimensions and front-rail thickness and yield strength, with displacement, stress, plastic strain and energy at 13 instants over 120 ms. Licence: CC BY-NC 4.0.Structures & crash https://surrogatemodels.com/news/physicsnemo-2-0https://surrogatemodels.com/news/physicsnemo-2-0Tue, 10 Mar 2026 09:00:00 GMTNVIDIA releases PhysicsNeMo v2.0.0, a refactor of the open-source framework that the 26.03 release notes describe as a major architectural change toward a modular toolkit. Reusable layers move to physicsnemo.nn, the base Module class to physicsnemo.core, and physicsnemo.launch is removed; a compatibility layer (physicsnemo.compat) keeps old import paths working, and a migration guide lists the changes. New packages are physicsnemo.diffusion (noise schedules, samplers, losses, multi-diffusion), physicsnemo.mesh (GPU mesh processing for simplicial meshes of any dimension) and refactored GPU datapipes. The 26.03 notes also add training recipes for crash dynamics with Transolver and MeshGraphNet, for car aerodynamics volume fields with GeoTransolver, and for StormScope with a diffusion transformer.Aero & CFDStructures & crashWeather & climate