arxiv: 2605.07098 · v1 · submitted 2026-05-08 · 💻 cs.LG · physics.comp-ph

Recognition: no theorem link

CarCrashNet: A Large-Scale Dataset and Hierarchical Neural Solver for Data-Driven Structural Crash Simulation

Mohamed Elrefaie , Dule Shu , Matthew Klenk , Faez Ahmed

Authors on Pith no claims yet

Pith reviewed 2026-05-11 00:50 UTC · model grok-4.3

classification 💻 cs.LG physics.comp-ph

keywords crash simulationfinite element analysismachine learningstructural mechanicsvehicle safetydataset benchmarkneural solverdata-driven modeling

0 comments

The pith

CarCrashNet provides a validated open benchmark of 15,000 crash simulations and a neural model to predict full-vehicle responses.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper introduces CarCrashNet as a public dataset combining more than 14,000 component-level bumper simulations with 825 full-vehicle crashes drawn from three standard models of increasing complexity. It validates the underlying open-source simulation workflow against both physical experiments and the commercial LS-DYNA solver. The authors also present CrashSolver, a machine-learning model trained to forecast full-vehicle crash outcomes directly from high-resolution finite-element data, and benchmark it against existing geometric deep-learning and transformer approaches.

Core claim

CarCrashNet is a multi-modal, high-fidelity benchmark that pairs extensive component-scale pole-impact data with full-vehicle crash simulations generated in OpenRadioss. The dataset is shown to be reliable through direct comparison with experimental results and the commercial Ansys LS-DYNA solver. CrashSolver is introduced as a hierarchical neural model that learns to predict the time-evolving structural response of complete vehicles from this data, with performance evaluated across the released component and vehicle subsets.

What carries the argument

CarCrashNet dataset together with CrashSolver, a machine-learning model that takes high-resolution finite-element crash data as input and outputs predicted full-vehicle crash behavior.

If this is right

The released dataset allows direct comparison of multiple neural architectures on the same high-resolution crash problems.
CrashSolver demonstrates that learned models can capture nonlinear contact, plasticity, and failure on full-vehicle meshes.
Validated open-source data lowers the barrier to exploring AI methods for crashworthiness without commercial licenses.
The multi-scale structure (component plus full vehicle) supports hierarchical modeling strategies that respect different length scales.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

Similar dataset-plus-solver pipelines could be constructed for other nonlinear structural problems such as aerospace impact or building collapse.
If the hierarchical design of CrashSolver generalizes, it may reduce the data requirements for training on even larger meshes by first learning component behavior.
Public availability of both the raw simulation files and the trained model enables community extension to new vehicle classes or material models.

Load-bearing premise

The open-source finite-element simulations match both real crash-test measurements and commercial LS-DYNA results closely enough to serve as trustworthy training data.

What would settle it

If CrashSolver predictions on a new vehicle geometry outside the training set show large discrepancies in intrusion depth, acceleration traces, or energy absorption compared with independent LS-DYNA runs or physical tests, the central claim would be falsified.

Figures

Figures reproduced from arXiv: 2605.07098 by Dule Shu, Faez Ahmed, Matthew Klenk, Mohamed Elrefaie.

**Figure 1.** Figure 1: Overview of our CARCRASHNET framework. Left: the released datasets include a large-scale bumper-beam pole-impact dataset with more than 14k simulations and three full-vehicle crash datasets based on the Toyota Yaris sedan, Dodge Neon, and Chevrolet Silverado. Middle: the machine learning tasks considered in this work include full-field crash prediction using FIGConvUNet, Transolver, GeoTransolver, and our … view at source ↗

**Figure 2.** Figure 2: Representative low- and high-velocity crash cases for the three vehicle-scale models. [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗

**Figure 3.** Figure 3: Structural regions edited by the vehicle-scale DoE. Highlighted front support and lower [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗

**Figure 4.** Figure 4: Overview of CrashSolver for full-vehicle crash prediction. Starting from the undeformed [PITH_FULL_IMAGE:figures/full_fig_p008_4.png] view at source ↗

**Figure 5.** Figure 5: Scalar wall-force validation summary. NHTSA 5677 and 6221 values are digitized from [PITH_FULL_IMAGE:figures/full_fig_p019_5.png] view at source ↗

**Figure 6.** Figure 6: Qualitative comparison of post-impact deformation fields between OpenRadioss and LS [PITH_FULL_IMAGE:figures/full_fig_p019_6.png] view at source ↗

**Figure 7.** Figure 7: Validation comparison of wall-force and energy responses for the same Yaris full-frontal [PITH_FULL_IMAGE:figures/full_fig_p020_7.png] view at source ↗

**Figure 8.** Figure 8: Baseline vehicle models used for the dataset generation in [PITH_FULL_IMAGE:figures/full_fig_p022_8.png] view at source ↗

**Figure 9.** Figure 9: Front-support components whose thicknesses are varied as design parameters for vehicle [PITH_FULL_IMAGE:figures/full_fig_p025_9.png] view at source ↗

**Figure 10.** Figure 10: Lower-rail and subframe components whose thicknesses are varied as design parameters [PITH_FULL_IMAGE:figures/full_fig_p025_10.png] view at source ↗

**Figure 11.** Figure 11: Representative deformed configurations from the Dodge Neon campaign shown in [PITH_FULL_IMAGE:figures/full_fig_p027_11.png] view at source ↗

**Figure 12.** Figure 12: Representative deformed configurations from the Dodge Neon campaign shown in side [PITH_FULL_IMAGE:figures/full_fig_p028_12.png] view at source ↗

**Figure 13.** Figure 13: Representative deformed configurations from the Toyota Yaris campaign shown in [PITH_FULL_IMAGE:figures/full_fig_p029_13.png] view at source ↗

**Figure 14.** Figure 14: Representative deformed configurations from the Toyota Yaris campaign shown in side [PITH_FULL_IMAGE:figures/full_fig_p030_14.png] view at source ↗

**Figure 15.** Figure 15: Representative deformed configurations from the Chevrolet Silverado campaign shown in [PITH_FULL_IMAGE:figures/full_fig_p031_15.png] view at source ↗

**Figure 16.** Figure 16: Representative deformed configurations from the Chevrolet Silverado campaign shown [PITH_FULL_IMAGE:figures/full_fig_p032_16.png] view at source ↗

**Figure 17.** Figure 17: Geometry and loading setup of the simplified bumper beam assembly used for dataset [PITH_FULL_IMAGE:figures/full_fig_p039_17.png] view at source ↗

**Figure 18.** Figure 18: Representative bumper-beam pole-impact configurations sampled from the design space. [PITH_FULL_IMAGE:figures/full_fig_p044_18.png] view at source ↗

**Figure 19.** Figure 19: Representative deformed configurations from the bumper-beam dataset used in [PITH_FULL_IMAGE:figures/full_fig_p045_19.png] view at source ↗

**Figure 20.** Figure 20: Column-normalized feature importance averaged over XGBoost, LightGBM, and CatBoost. The plotted values are model-internal importance scores averaged over the three boosters, then normalized independently within each target so that every target column sums to one. AutoML versus hand-picked boosters. AutoGluon’s best_quality preset reaches a mean R2 of 0.760, close to the bagged-tree baselines but below … view at source ↗

read the original abstract

Crash simulation is a cornerstone of modern vehicle development because it reduces the need for costly physical prototypes, accelerates safety-driven design iteration, and increasingly supports virtual testing workflows. At the same time, modeling structural crash mechanics remains exceptionally challenging: the response is governed by nonlinear contact, large deformation, material plasticity, failure, and complex multi-body interactions evolving over space and time on high-resolution finite-element meshes. In this work, we introduce \textsc{CarCrashNet}, a public high-fidelity open-source benchmark for data-driven structural crash simulation. \textsc{CarCrashNet} combines component-scale and full-vehicle simulations in a multi-modal format, including more than 14{,}000 bumper-beam pole-impact simulations with varying geometry, materials, and boundary conditions, together with 825 full-vehicle crash simulations built from three industry-standard vehicle models of increasing structural complexity: Dodge Neon, Toyota Yaris, and Chevrolet Silverado. To establish the reliability of the benchmark, we validate our open-source finite-element workflow based on OpenRadioss against both experimental crash data and the commercial solver Ansys LS-DYNA. We also introduce \textsc{CrashSolver}, a machine-learning model designed for full-vehicle crash prediction from high-resolution finite-element crash data. We further perform extensive benchmarking across the released datasets and evaluate \textsc{CrashSolver} against state-of-the-art geometric deep learning and transformer-based neural solvers. Our results position \textsc{CarCrashNet} as a foundation for reproducible research in structural simulation, crashworthiness modeling, and AI-driven virtual crash testing. The dataset is available at https://github.com/Mohamedelrefaie/CarCrashNet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

CarCrashNet offers a large public crash dataset and solver but validation for full vehicles needs more scrutiny.

read the letter

The main point to know is that this paper releases a sizable public dataset of structural crash simulations along with a hierarchical neural solver, but the validation for the full-vehicle cases may not be as robust as claimed. They have put together over 14,000 component simulations of bumper beams under pole impacts with different geometries and materials. On top of that, there are 825 full-vehicle simulations using three real models: Dodge Neon, Toyota Yaris, and Chevrolet Silverado. The data comes in a multi-modal format and is made available on GitHub. They also built CrashSolver to predict full-vehicle crashes from the finite element data and benchmarked it against other neural approaches. This scale of public data is new and useful. Releasing it openly allows others to train and test models without the high cost of running their own simulations. The validation against both experiments and LS-DYNA is a good step toward making the data trustworthy. The weak area is the lack of detailed quantitative results in the abstract. It mentions validation and benchmarking but does not show error metrics or convergence studies. If the full paper only provides strong checks for the component cases and lighter ones for the full vehicles, then the central claim about reliability for training the solver could be overstated. The stress-test note points to a real possibility here that needs addressing. Readers in AI for mechanical engineering or automotive design would get the most from this. It serves as a benchmark for data-driven crash simulation. The paper deserves peer review because the dataset itself adds value to the field, provided the validation holds up under scrutiny. I recommend sending it for review with specific attention to the methods and results sections on validation and model performance.

Referee Report

2 major / 2 minor

Summary. The paper introduces CarCrashNet, a public benchmark dataset combining over 14,000 component-scale bumper-beam pole-impact simulations (varying geometry, materials, and BCs) with 825 full-vehicle crash simulations from three industry models (Dodge Neon, Toyota Yaris, Chevrolet Silverado). It validates an open-source OpenRadioss finite-element workflow against experimental crash data and the commercial Ansys LS-DYNA solver, presents CrashSolver as a hierarchical neural model for full-vehicle crash prediction from high-resolution FE data, and benchmarks the model against geometric deep learning and transformer baselines, positioning the dataset as a foundation for reproducible AI-driven crash simulation research.

Significance. If the validation establishes sufficient fidelity, the work would be significant for supplying the first large-scale, open, multi-modal benchmark spanning component to full-vehicle scales in structural crash mechanics. This directly addresses the scarcity of high-fidelity training data for nonlinear contact, plasticity, and failure problems, enabling reproducible data-driven surrogate modeling and virtual testing workflows that could reduce reliance on physical prototypes and proprietary solvers.

major comments (2)

[§4] §4 (Validation of OpenRadioss workflow): The central claim that CarCrashNet provides reliable high-fidelity data for training CrashSolver requires quantitative agreement metrics (e.g., force-displacement curve RMSE, peak force error, or energy absorption discrepancy) specifically for the 825 full-vehicle simulations. The abstract asserts validation against experiments and LS-DYNA, but if these metrics are reported only for the 14k bumper-beam cases (as the skeptic concern indicates), systematic differences in contact, plasticity, or failure modes for full-vehicle runs would undermine the benchmark's suitability as training data.
[§5] §5 (CrashSolver architecture and training): The hierarchical neural solver is positioned as effective for full-vehicle prediction, yet the manuscript must demonstrate that any validation discrepancies in the underlying FE data do not propagate into learned dynamics (e.g., via ablation on data fidelity or error propagation analysis). Without this, the benchmarking results against geometric DL and transformer baselines cannot fully support the claim of reliable data-driven crash prediction.

minor comments (2)

[§3] The multi-modal dataset description would benefit from an explicit table summarizing mesh sizes, simulation durations, and output variables for both component and full-vehicle subsets to improve reproducibility.
[Figures in §3 and §6] Figure captions for the vehicle models and crash setups should include quantitative scale bars or mesh density indicators for clarity.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the detailed and constructive review. The comments highlight important aspects of validation and model robustness that we address below. We have revised the manuscript to incorporate additional quantitative metrics and analyses as suggested.

read point-by-point responses

Referee: [§4] §4 (Validation of OpenRadioss workflow): The central claim that CarCrashNet provides reliable high-fidelity data for training CrashSolver requires quantitative agreement metrics (e.g., force-displacement curve RMSE, peak force error, or energy absorption discrepancy) specifically for the 825 full-vehicle simulations. The abstract asserts validation against experiments and LS-DYNA, but if these metrics are reported only for the 14k bumper-beam cases (as the skeptic concern indicates), systematic differences in contact, plasticity, or failure modes for full-vehicle runs would undermine the benchmark's suitability as training data.

Authors: We agree that explicit quantitative metrics for the full-vehicle simulations are essential to support the benchmark's reliability. The original manuscript provides detailed validation (including force-displacement curves, RMSE, peak force, and energy metrics) primarily for the 14,000+ component-scale cases against both experiments and LS-DYNA. For the 825 full-vehicle simulations, comparisons to LS-DYNA were performed but not reported with the same granularity. In the revised manuscript, we will add a dedicated subsection with quantitative agreement metrics (RMSE on force-displacement curves, peak force error, energy absorption discrepancy, and deformation mode similarity) for representative full-vehicle cases across the three vehicle models. We note that direct experimental data for these exact full-vehicle configurations is limited in the public domain, so the primary validation remains against the commercial solver; this limitation will be explicitly stated. revision: yes
Referee: [§5] §5 (CrashSolver architecture and training): The hierarchical neural solver is positioned as effective for full-vehicle prediction, yet the manuscript must demonstrate that any validation discrepancies in the underlying FE data do not propagate into learned dynamics (e.g., via ablation on data fidelity or error propagation analysis). Without this, the benchmarking results against geometric DL and transformer baselines cannot fully support the claim of reliable data-driven crash prediction.

Authors: We concur that robustness to potential data discrepancies must be demonstrated to substantiate the claims. The current manuscript includes benchmarking of CrashSolver against geometric deep learning and transformer baselines on the released datasets but does not include explicit ablation on data fidelity or error propagation. In the revised version, we will add an analysis subsection that (1) trains CrashSolver on subsets of the data with controlled fidelity variations (e.g., lower-resolution meshes or perturbed material parameters) and reports the resulting prediction degradation, and (2) performs a limited error propagation study by comparing model outputs when trained on OpenRadioss-generated data versus available LS-DYNA equivalents. These additions will directly address whether discrepancies propagate into the learned dynamics and will strengthen the interpretation of the baseline comparisons. revision: yes

Circularity Check

0 steps flagged

No significant circularity; empirical dataset and model introduction

full rationale

The paper introduces CarCrashNet as a dataset of OpenRadioss simulations (component and full-vehicle) and CrashSolver as a neural model trained on it, with validation against external experimental crash data and the independent commercial solver LS-DYNA. No load-bearing derivation chain exists that reduces predictions or results to fitted inputs, self-definitions, or self-citation chains by construction. The central claims rest on data generation, external benchmarking, and empirical performance evaluation rather than any self-referential mathematical step.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The paper relies on standard finite-element assumptions for nonlinear mechanics and introduces no new physical constants, particles, or ad-hoc entities; the central contribution is empirical data and an applied ML model.

axioms (1)

domain assumption Standard assumptions of the finite-element method for modeling large-deformation contact, plasticity, and failure on high-resolution meshes.
Invoked when describing the simulation workflow and its validation against experiments.

pith-pipeline@v0.9.0 · 5615 in / 1342 out tokens · 52412 ms · 2026-05-11T00:50:28.612933+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

113 extracted references · 15 canonical work pages · 2 internal anchors

[3]

International Journal of Impact Engineering , volume=

Axial crushing of thin-walled high-strength steel sections , author=. International Journal of Impact Engineering , volume=. 2006 , doi=

2006
[4]

2026 , url =

Crash Simulation Vehicle Models , howpublished =. 2026 , url =

2026
[5]

2026 , url =

Finite Element Models , howpublished =. 2026 , url =

2026
[6]

2016 , doi =

2010 Toyota Yaris Finite Element Model Validation: Coarse Mesh , howpublished =. 2016 , doi =

2010
[7]

2016 , doi =

2010 Toyota Yaris Finite Element Model Validation: Detail Mesh , howpublished =. 2016 , doi =

2010
[8]

2011 , number =

Development and Validation of a Finite Element Model for the 2010 Toyota Yaris Passenger Sedan , institution =. 2011 , number =

2010
[9]

2012 , number =

Extended Validation of the Finite Element Model for the 2010 Toyota Yaris Passenger Sedan , institution =. 2012 , number =

2010
[10]

Computer Methods in Applied Mechanics and Engineering , volume=

A machine learning framework for accelerating the design process using CAE simulations: An application to finite element analysis in structural crashworthiness , author=. Computer Methods in Applied Mechanics and Engineering , volume=. 2021 , publisher=

2021
[11]

Computer Methods in Applied Mechanics and Engineering , volume=

Introducing finite element method integrated networks (FEMIN) , author=. Computer Methods in Applied Mechanics and Engineering , volume=. 2024 , publisher=

2024
[12]

Computer Methods in Applied Mechanics and Engineering , volume=

Accelerating crash simulations with Finite Element Method Integrated Networks (FEMIN): Comparing two approaches to replace large portions of a FEM simulation , author=. Computer Methods in Applied Mechanics and Engineering , volume=. 2025 , publisher=

2025
[13]

The Thirty-ninth Annual Conference on Neural Information Processing Systems Datasets and Benchmarks Track , year=

Bubbleformer: Forecasting Boiling with Transformers , author=. The Thirty-ninth Annual Conference on Neural Information Processing Systems Datasets and Benchmarks Track , year=
[14]

SIA Simulation num

Comparing traditional surrogate modelling and neural fields for vehicle crash simulation data , author=. SIA Simulation num
[16]

International Journal of Crashworthiness , volume=

Crashworthiness performance of thin-walled structures towards design configuration in vehicle crash boxes application: a review , author=. International Journal of Crashworthiness , volume=. 2025 , publisher=

2025
[17]

International Journal of Impact Engineering , volume=

Neural network modelling of mechanical joints for the application in large-scale crash analyses , author=. International Journal of Impact Engineering , volume=. 2023 , publisher=

2023
[18]

International Journal of Crashworthiness , volume=

A review on finite element modelling and experimental analysis of crashworthiness design of automotive body , author=. International Journal of Crashworthiness , volume=. 2025 , publisher=

2025
[19]

International Journal for Numerical Methods in Engineering , volume=

Nonintrusive Local/Global Coupling With Local Deep Learning-Based Models for the Effective Simulation of Spotwelded Structures Under Impact , author=. International Journal for Numerical Methods in Engineering , volume=. 2025 , publisher=

2025
[20]

Thin-walled structures , volume=

Design and analysis of an automotive bumper beam in low-speed frontal crashes , author=. Thin-walled structures , volume=. 2009 , publisher=

2009
[21]

The Journal of Engineering , volume=

The Automotive Bumper Beam in the Era of the 4th Industrial Revolution , author=. The Journal of Engineering , volume=. 2025 , publisher=

2025
[22]

Inventions , volume=

Intelligent Damage Prediction During Vehicle Collisions Based on Simulation Datasets , author=. Inventions , volume=. 2025 , publisher=

2025
[23]

PAMM , volume=

Machine learning enhanced optimisation of crash box design for crashworthiness analysis , author=. PAMM , volume=. 2023 , publisher=

2023
[24]

Computers in Industry , volume=

Explainable artificial intelligence for enhancing system understanding and interpretability of numerical crash simulations , author=. Computers in Industry , volume=. 2026 , publisher=

2026
[26]

Journal of complexity , volume=

Scrambling Sobol'and Niederreiter--Xing Points , author=. Journal of complexity , volume=. 1998 , publisher=

1998
[27]

Proceedings of the 7th International Symposium on Ballistics, Am

A constitutive model and data for metals subjected to large strains, high strain rates and high temperatures , author=. Proceedings of the 7th International Symposium on Ballistics, Am. Def. Prep. Org.(ADPA), Netherlands, 1983 , pages=

1983
[28]

Computer Methods in Applied Mechanics and Engineering , volume =

Explicit algorithms for the nonlinear dynamics of shells , author =. Computer Methods in Applied Mechanics and Engineering , volume =. 1984 , doi =

1984
[29]

Technometrics , volume=

A comparison of three methods for selecting values of input variables in the analysis of output from a computer code , author=. Technometrics , volume=. 2000 , publisher=

2000
[30]

Journal of statistical planning and inference , volume=

Minimax and maximin distance designs , author=. Journal of statistical planning and inference , volume=. 1990 , publisher=

1990
[34]

Journal of Machine Learning Research , volume=

Neural operator: Learning maps between function spaces with applications to pdes , author=. Journal of Machine Learning Research , volume=
[35]

International Conference on Learning Representations , year =

Pfaff, Tobias and Fortunato, Meire and Sanchez-Gonzalez, Alvaro and Battaglia, Peter , title =. International Conference on Learning Representations , year =
[36]

Advances in neural information processing systems , volume=

Attention is all you need , author=. Advances in neural information processing systems , volume=
[37]

International Conference on Learning Representations , year =

Fourier Neural Operator for Parametric Partial Differential Equations , author =. International Conference on Learning Representations , year =
[38]

Nature machine intelligence , volume=

Learning nonlinear operators via DeepONet based on the universal approximation theorem of operators , author=. Nature machine intelligence , volume=. 2021 , publisher=

2021
[39]

International conference on machine learning , pages=

Learning to simulate complex physics with graph networks , author=. International conference on machine learning , pages=. 2020 , organization=

2020
[40]

Advances in Neural Information Processing Systems , volume=

Towards foundation models for scientific machine learning: Characterizing scaling and transfer behavior , author=. Advances in Neural Information Processing Systems , volume=
[41]

Advances in Neural Information Processing Systems , volume=

Drivaernet++: A large-scale multimodal car dataset with computational fluid dynamics simulations and deep learning benchmarks , author=. Advances in Neural Information Processing Systems , volume=
[42]

Advances in neural information processing systems , volume=

Pdebench: An extensive benchmark for scientific machine learning , author=. Advances in neural information processing systems , volume=
[43]

Advances in Neural Information Processing Systems , volume=

Airfrans: High fidelity computational fluid dynamics dataset for approximating reynolds-averaged navier--stokes solutions , author=. Advances in Neural Information Processing Systems , volume=
[45]

Advances in Neural Information Processing Systems , volume=

Lagrangebench: A lagrangian fluid mechanics benchmarking suite , author=. Advances in Neural Information Processing Systems , volume=
[46]

Advances in Neural Information Processing Systems , volume=

The well: a large-scale collection of diverse physics simulations for machine learning , author=. Advances in Neural Information Processing Systems , volume=
[47]

Advances in Neural Information Processing Systems , volume=

Apebench: A benchmark for autoregressive neural emulators of pdes , author=. Advances in Neural Information Processing Systems , volume=
[48]

Advances in Neural Information Processing Systems , volume=

Pinnacle: A comprehensive benchmark of physics-informed neural networks for solving pdes , author=. Advances in Neural Information Processing Systems , volume=
[49]

Advances in Neural Information Processing Systems , volume=

Poseidon: Efficient foundation models for pdes , author=. Advances in Neural Information Processing Systems , volume=
[50]

Journal of Advances in Modeling Earth Systems , volume=

WeatherBench: a benchmark data set for data-driven weather forecasting , author=. Journal of Advances in Modeling Earth Systems , volume=. 2020 , publisher=

2020
[51]

0: A benchmark for data-driven climate projections , author=

ClimateBench v1. 0: A benchmark for data-driven climate projections , author=. Journal of Advances in Modeling Earth Systems , volume=. 2022 , publisher=

2022
[52]

Scientific Data , year=

WxC-Bench: A novel dataset for weather and climate downstream tasks , author=. Scientific Data , year=
[54]

Acs Catalysis , volume=

Open catalyst 2020 (OC20) dataset and community challenges , author=. Acs Catalysis , volume=. 2021 , publisher=

2020
[56]

International Design Engineering Technical Conferences and Computers and Information in Engineering Conference , volume=

Blendednet: A blended wing body aircraft dataset and surrogate model for aerodynamic predictions , author=. International Design Engineering Technical Conferences and Computers and Information in Engineering Conference , volume=. 2025 , organization=

2025
[58]

Journal of Mechanical Design , volume=

Drivaernet: A parametric car dataset for data-driven aerodynamic design and prediction , author=. Journal of Mechanical Design , volume=. 2025 , publisher=

2025
[59]

International Design Engineering Technical Conferences and Computers and Information in Engineering Conference , volume=

Drivaernet: A parametric car dataset for data-driven aerodynamic design and graph-based drag prediction , author=. International Design Engineering Technical Conferences and Computers and Information in Engineering Conference , volume=. 2024 , organization=

2024
[62]

Adams, R

Corey Adams, Rishikesh Ranade, Ram Cherukuri, and Sanjay Choudhry. Geotransolver: Learning physics on irregular domains using multi-scale geometry aware physics attention transformer. arXiv preprint arXiv:2512.20399, 2025

work page arXiv 2025
[63]

Altair Radioss : Crash and safety dynamic analysis software

Altair . Altair Radioss : Crash and safety dynamic analysis software. Official product page, 2026. URL https://altair.com/radioss/. Accessed: 2026-05-04

2026
[64]

Neural network modelling of mechanical joints for the application in large-scale crash analyses

Victor Andr \'e , Miguel Costas, Magnus Langseth, and David Morin. Neural network modelling of mechanical joints for the application in large-scale crash analyses. International Journal of Impact Engineering, 177: 0 104490, 2023

2023
[65]

Ansys LS-DYNA : Crash simulation software

Ansys . Ansys LS-DYNA : Crash simulation software. Official product page, 2026. URL https://www.ansys.com/products/structures/ansys-ls-dyna. Accessed: 2026-05-04

2026
[66]

Lin, and C

Ted Belytschko, Jerry I. Lin, and C. S. Tsay. Explicit algorithms for the nonlinear dynamics of shells. Computer Methods in Applied Mechanics and Engineering, 42 0 (2): 0 225--251, 1984. doi:10.1016/0045-7825(84)90026-4

work page doi:10.1016/0045-7825(84)90026-4 1984
[67]

Airfrans: High fidelity computational fluid dynamics dataset for approximating reynolds-averaged navier--stokes solutions

Florent Bonnet, Jocelyn Mazari, Paola Cinnella, and Patrick Gallinari. Airfrans: High fidelity computational fluid dynamics dataset for approximating reynolds-averaged navier--stokes solutions. Advances in Neural Information Processing Systems, 35: 0 23463--23478, 2022

2022
[68]

Machine learning enhanced optimisation of crash box design for crashworthiness analysis

Aditya Borse, Rutwik Gulakala, and Marcus Stoffel. Machine learning enhanced optimisation of crash box design for crashworthiness analysis. PAMM, 23 0 (4): 0 e202300145, 2023

2023
[69]

2010 toyota yaris finite element model validation: Coarse mesh

Center for Collision Safety and Analysis . 2010 toyota yaris finite element model validation: Coarse mesh. Validation presentation, 2016 a . URL https://www.ccsa.gmu.edu/wp-content/uploads/2016/11/2010-toyota-yaris-coarse-validation-v1.pdf

2010
[70]

2010 toyota yaris finite element model validation: Detail mesh

Center for Collision Safety and Analysis . 2010 toyota yaris finite element model validation: Detail mesh. Validation presentation, 2016 b . URL https://www.ccsa.gmu.edu/wp-content/uploads/2016/10/2010-toyota-yaris-detailed-validation-v2.pdf

2010
[71]

Finite element models

Center for Collision Safety and Analysis . Finite element models. Official model repository, 2026. URL https://www.ccsa.gmu.edu/models/. Accessed: 2026-05-04

2026
[72]

Open catalyst 2020 (oc20) dataset and community challenges

Lowik Chanussot, Abhishek Das, Siddharth Goyal, Thibaut Lavril, Muhammed Shuaibi, Morgane Riviere, Kevin Tran, Javier Heras-Domingo, Caleb Ho, Weihua Hu, et al. Open catalyst 2020 (oc20) dataset and community challenges. Acs Catalysis, 11 0 (10): 0 6059--6072, 2021

2020
[73]

arXiv preprint arXiv:2503.17400 , year=

Qian Chen, Mohamed Elrefaie, Angela Dai, and Faez Ahmed. Tripnet: Learning large-scale high-fidelity 3d car aerodynamics with triplane networks. arXiv preprint arXiv:2503.17400, 2025

work page arXiv 2025
[74]

Factorized implicit global convolution for automotive computational fluid dynamics prediction,

Chris Choy, Alexey Kamenev, Jean Kossaifi, Max Rietmann, Jan Kautz, and Kamyar Azizzadenesheli. Factorized implicit global convolution for automotive computational fluid dynamics prediction. arXiv preprint arXiv:2502.04317, 2025

work page arXiv 2025
[75]

Drivaernet: A parametric car dataset for data-driven aerodynamic design and graph-based drag prediction

Mohamed Elrefaie, Angela Dai, and Faez Ahmed. Drivaernet: A parametric car dataset for data-driven aerodynamic design and graph-based drag prediction. In International Design Engineering Technical Conferences and Computers and Information in Engineering Conference, volume 88360, page V03AT03A019. American Society of Mechanical Engineers, 2024 a

2024
[76]

Drivaernet++: A large-scale multimodal car dataset with computational fluid dynamics simulations and deep learning benchmarks

Mohamed Elrefaie, Florin Morar, Angela Dai, and Faez Ahmed. Drivaernet++: A large-scale multimodal car dataset with computational fluid dynamics simulations and deep learning benchmarks. Advances in Neural Information Processing Systems, 37: 0 499--536, 2024 b

2024
[77]

Drivaernet: A parametric car dataset for data-driven aerodynamic design and prediction

Mohamed Elrefaie, Angela Dai, and Faez Ahmed. Drivaernet: A parametric car dataset for data-driven aerodynamic design and prediction. Journal of Mechanical Design, 147 0 (4): 0 041712, 2025 a

2025
[78]

Car- Bench: A Comprehensive Benchmark for Neural Sur- rogatesonHighFidelity3DCarAerodynamics,

Mohamed Elrefaie, Dule Shu, Matt Klenk, and Faez Ahmed. Carbench: A comprehensive benchmark for neural surrogates on high-fidelity 3d car aerodynamics. arXiv preprint arXiv:2512.07847, 2025 b

work page arXiv 2025
[79]

AutoGluon-Tabular: Robust and Accurate AutoML for Structured Data

Nick Erickson, Jonas Mueller, Alexander Shirkov, Hang Zhang, Pedro Larroy, Mu Li, and Alexander Smola. Autogluon-tabular: Robust and accurate automl for structured data. arXiv preprint arXiv:2003.06505, 2020

work page internal anchor Pith review arXiv 2003
[80]

Pinnacle: A comprehensive benchmark of physics-informed neural networks for solving pdes

Zhongkai Hao, Jiachen Yao, Chang Su, Hang Su, Ziao Wang, Fanzhi Lu, Zeyu Xia, Yichi Zhang, Songming Liu, Lu Lu, et al. Pinnacle: A comprehensive benchmark of physics-informed neural networks for solving pdes. Advances in Neural Information Processing Systems, 37: 0 76721--76774, 2024

2024
[81]

Bubbleformer: Forecasting boiling with transformers

Sheikh Md Shakeel Hassan, Xianwei Zou, Akash Dhruv, and Aparna Chandramowlishwaran. Bubbleformer: Forecasting boiling with transformers. In The Thirty-ninth Annual Conference on Neural Information Processing Systems Datasets and Benchmarks Track, 2025

2025
[82]

Poseidon: Efficient foundation models for pdes

Maximilian Herde, Bogdan Raoni \'c , Tobias Rohner, Roger K \"a ppeli, Roberto Molinaro, Emmanuel De Bezenac, and Siddhartha Mishra. Poseidon: Efficient foundation models for pdes. Advances in Neural Information Processing Systems, 37: 0 72525--72624, 2024

2024
[83]

TabPFN: A Transformer That Solves Small Tabular Classification Problems in a Second

Noah Hollmann, Samuel M \"u ller, Katharina Eggensperger, and Frank Hutter. Tabpfn: A transformer that solves small tabular classification problems in a second. arXiv preprint arXiv:2207.01848, 2022

work page internal anchor Pith review arXiv 2022
[84]

Realpdebench: A benchmark for complex physical systems with real-world data.arXiv preprint arXiv:2601.01829, 2026

Peiyan Hu, Haodong Feng, Hongyuan Liu, Tongtong Yan, Wenhao Deng, Tianrun Gao, et al. Realpdebench: A benchmark for complex physical systems with real-world data. arXiv preprint arXiv:2601.01829, 2026

work page arXiv 2026
[85]

A constitutive model and data for metals subjected to large strains, high strain rates and high temperatures

Gordon R Johnson. A constitutive model and data for metals subjected to large strains, high strain rates and high temperatures. In Proceedings of the 7th International Symposium on Ballistics, Am. Def. Prep. Org.(ADPA), Netherlands, 1983, pages 541--547, 1983

1983
[86]

Minimax and maximin distance designs

Mark E Johnson, Leslie M Moore, and Donald Ylvisaker. Minimax and maximin distance designs. Journal of statistical planning and inference, 26 0 (2): 0 131--148, 1990

1990
[87]

Apebench: A benchmark for autoregressive neural emulators of pdes

Felix Koehler, Simon Niedermayr, R \"u diger Westermann, and Nils Thuerey. Apebench: A benchmark for autoregressive neural emulators of pdes. Advances in Neural Information Processing Systems, 37: 0 120252--120310, 2024

2024
[88]

A machine learning framework for accelerating the design process using cae simulations: An application to finite element analysis in structural crashworthiness

Christopher P Kohar, Lars Greve, Tom K Eller, Daniel S Connolly, and Kaan Inal. A machine learning framework for accelerating the design process using cae simulations: An application to finite element analysis in structural crashworthiness. Computer Methods in Applied Mechanics and Engineering, 385: 0 114008, 2021

2021
[89]

Neural operator: Learning maps between function spaces with applications to pdes

Nikola Kovachki, Zongyi Li, Burigede Liu, Kamyar Azizzadenesheli, Kaushik Bhattacharya, Andrew Stuart, and Anima Anandkumar. Neural operator: Learning maps between function spaces with applications to pdes. Journal of Machine Learning Research, 24 0 (89): 0 1--97, 2023

2023
[90]

Comparing traditional surrogate modelling and neural fields for vehicle crash simulation data

Yves Le Guennec, Thibaut Defoort, Jose Vicente Aguado, and Domenico Borzacchiello. Comparing traditional surrogate modelling and neural fields for vehicle crash simulation data. In SIA Simulation num \'e rique 2025 , 2025

2025
[91]

A new graph-based surrogate model for rapid prediction of crashworthiness performance of vehicle panel components

Haoran Li, Yingxue Zhao, Haosu Zhou, Tobias Pfaff, and Nan Li. A new graph-based surrogate model for rapid prediction of crashworthiness performance of vehicle panel components. arXiv preprint arXiv:2503.17386, 2025

work page arXiv 2025
[92]

Fourier neural operator for parametric partial differential equations

Zongyi Li, Nikola Kovachki, Kamyar Azizzadenesheli, Burigede Liu, Kaushik Bhattacharya, Andrew Stuart, and Anima Anandkumar. Fourier neural operator for parametric partial differential equations. In International Conference on Learning Representations, 2021

2021
[93]

Intelligent damage prediction during vehicle collisions based on simulation datasets

Sheng Liu, Conghao Liu, Xunan An, Xin Liu, and Liang Hao. Intelligent damage prediction during vehicle collisions based on simulation datasets. Inventions, 10 0 (3): 0 40, 2025

2025

Showing first 80 references.