BLASTNetv3: Multiphysics Simulation Dataset

arXiv (Cornell University)

Abstract

Analysis of compressible turbulent flows is essential for applications related to propulsion, energy generation, and the environment. Here, we present BLASTNet 2.0, a 2.2 TB network-of-datasets containing 744 full-domain samples from 34 high-fidelity direct numerical simulations, which addresses the current limited availability of 3D high-fidelity reacting and non-reacting compressible turbulent flow simulation data. With this data, we benchmark a total of 49 variations of five deep learning approaches for 3D super-resolution - which can be applied for improving scientific imaging, simulations, turbulence models, as well as in computer vision applications. We perform neural scaling analysis on these models to examine the performance of different machine learning (ML) approaches, including two scientific ML techniques. We demonstrate that (i) predictive performance can scale with model size and cost, (ii) architecture matters significantly, especially for smaller models, and (iii) the benefits of physics-based losses can persist with increasing model size. The outcomes of this benchmark study are anticipated to offer insights that can aid the design of 3D super-resolution models, especially for turbulence models, while this data is expected to foster ML methods for a broad range of flow physics applications. This data is publicly available with download links and browsing tools consolidated at https://blastnet.github.io.

Authors 37

1. Wai Tong Chung
2. Matthias Ihme

   Stanford University

   Affiliation as printed

   Stanford University

3. Ki Sung Jung

   Sandia National Laboratories California

   Affiliation as printed

   Sandia National Laboratories California

4. Jacqueline H Chen

   Sandia National Laboratories California

   Affiliation as printed

   Sandia National Laboratories California

5. Jack Guo

   Lawrence Livermore National Laboratory

   Affiliation as printed

   Lawrence Livermore National Laboratory

6. Davy J. Brouzet

   Stanford University

   Affiliation as printed

   Stanford University

7. Mohsen Talei

   The University of Melbourne

   Affiliation as printed

   The University of Melbourne

8. Bin Jiang

   The University of Melbourne

   Affiliation as printed

   The University of Melbourne

9. Bruno Savard

   Polytechnique Montréal

   Affiliation as printed

   Polytechnique Montréal

10. Alexei Poludnenko

    University of Connecticut

    Affiliation as printed

    University of Connecticut

11. Bassem M. Akoush

    Stanford University

    Affiliation as printed

    Stanford University

12. Pushan Sharma
13. Alex Tamkin

    Stanford University

    Affiliation as printed

    Stanford University

14. Qing Wang

    Google (United States)

    Affiliation as printed

    Google (United States)

15. Shantanu Shahane

    Google (United States)

    Affiliation as printed

    Google (United States)

16. Yifan Chen

    Google (United States)

    Affiliation as printed

    Google (United States)

17. Victor Coulon

    Centre Européen de Recherche et de Formation Avancée en Calcul Scientifique

    Affiliation as printed

    CERFACS

18. Corentin Lapeyre

    Nvidia (United Kingdom) · Nvidia (United States)

    Affiliation as printed

    NVIDIA

19. Michael Gauding Aachen

    RWTH Aachen University

    Affiliation as printed

    RWTH Aachen University

20. Roshan J. Samuel

    Technische Universität Ilmenau

    Affiliation as printed

    Technische Universität Ilmenau

21. Mathis Bode

    Forschungszentrum Jülich

    Affiliation as printed

    Forschungszentrum Jülich

22. Jörg Schumacher

    Technische Universität Ilmenau

    Affiliation as printed

    Technische Universität Ilmenau

23. Jen Zen Ho

    Stanford University

    Affiliation as printed

    Stanford University

24. Quentin Malé

    Harvard University

    Affiliation as printed

    Harvard University

25. Cenk Gazen

    Google (United States)

    Affiliation as printed

    Google (United States)

26. Daniel L. Dotson

    Georgia Institute of Technology

    Affiliation as printed

    Georgia Institute of Technology

27. P. K. Yeung

    Georgia Institute of Technology

    Affiliation as printed

    Georgia Institute of Technology

28. Myoungkyu Lee

    University of Houston

    Affiliation as printed

    University of Houston

29. Zhang Yanqi

    Kyoto University

    Affiliation as printed

    Kyoto University

30. Jiangkuan Xing

    Kyoto University

    Affiliation as printed

    Kyoto University

31. Ryoichi Kurose

    Kyoto University

    Affiliation as printed

    Kyoto University

32. Gino Biondini

    University at Buffalo, State University of New York

    Affiliation as printed

    University at Buffalo, State University of New York

33. Alexander Bivolcic

    Embry-Riddle Aeronautical University Worldwide & Online

    Affiliation as printed

    Embry-Riddle Aeronautical University Worldwide & Online

34. Lluís Jofre

    Universitat Politècnica de Catalunya

    Affiliation as printed

    Universitat Politècnica de Catalunya

35. Guillem Barea

    Universitat Politècnica de Catalunya

    Affiliation as printed

    Universitat Politècnica de Catalunya

36. Carlos Monteiro

    Universitat Politècnica de Catalunya

    Affiliation as printed

    Universitat Politècnica de Catalunya

37. Cheng Chi

    Otto-von-Guericke-Universität Magdeburg

    Affiliation as printed

    Otto-von-Guericke-Universität Magdeburg