A GNN-BASED FRAMEWORK TO IDENTIFY FLOW PHENOMENA ON UNSTRUCTURED MESHES -PARCFD2023 - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année : 2023

A GNN-BASED FRAMEWORK TO IDENTIFY FLOW PHENOMENA ON UNSTRUCTURED MESHES -PARCFD2023

Résumé

Driven by the abundant data generated from Computational Fluid Dynamics (CFD) simulations, machine learning (ML) methods surpass the deterministic criteria on flow phenomena identification in the way that is independent of case-by-case thresholds by combining the flow field properties and the topological distribution of the phenomena. The current most popular and successful ML models based on convolutional neural networks are limited to structured meshes and unable to directly digest the data generated from unstructured meshes which are more widely used in the real industrial CFD simulations. We propose a framework based on graph neural networks with the proposed Fast Gaussian Mixture Model as the convolution kernel and U-Net architecture to detect flow phenomena resided on a graph hierarchy generated by the algebraic multigrid method embedded in the open-source CFD solver, code saturne. We demonstrate the superiority of the proposed kernel and U-Net architecture, along with the generality of the framework to unstructured mesh and unseen case on detecting the vortexes behind the backward-facing step. Our proposed framework can be trivially extended to detect other flow phenomena in 3D cases which is ongoing work.
Fichier principal
Vignette du fichier
A_GNN_based_framework_to_identify_flow_phenomena_on_unstructured_meshes____ParCFD2023 (1).pdf (4.56 Mo) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-04510114 , version 1 (18-03-2024)

Licence

Paternité

Identifiants

  • HAL Id : hal-04510114 , version 1

Citer

L. Wang, Y. Fournier, Jf. Wald, Youssef Mesri. A GNN-BASED FRAMEWORK TO IDENTIFY FLOW PHENOMENA ON UNSTRUCTURED MESHES -PARCFD2023. 33rd Parallel CFD International Conference, May 2023, Cuenca, Ecuador. ⟨hal-04510114⟩
0 Consultations
1 Téléchargements

Partager

Gmail Facebook X LinkedIn More