de Villeroché , M Armand et Mouradi, Mme Rem-Sophia et Le Guen, M. Vincent et Massin, M. Patrick (2023) Neural surrogates for atmospheric dispersion in built areas PFE - Projet de fin d'études, ENSTA.

Il s'agit de la dernière version de ce document.

Fichier(s) associé(s) à ce document :

[img]PDF
Restricted to Accès restreint
795Kb

Résumé

Studies of atmospheric dispersion of pollutants on a local scale are usually done with computational fluid dynamics (CFD) simulations. However, CFD computations are expensive, so can only be performed on a limited number of cases. Machine Learning (ML) approaches offer a promising alternative for faster results, including results of cases not present in the initial CFD-generated database. Here, a multi-layer perceptron (MLP) is trained on CFD simulation results with varying wind directions. The impact of the preprocessing steps is studied. Then, a sub-sampling method of spatial points is presented. The method allows to greatly reduce training time with little to no impact on the accuracy of the training and validation by selecting only a small number of spatial points from what is available in the CFD mesh. Finally, the trained MLP is compared to a simple interpolation-based baseline. It is shown that the model is more accurate and has better generalization capabilities

Type de document:Rapport ou mémoire (PFE - Projet de fin d'études)
Mots-clés libres:Machine learning
Sujets:Mathématiques et leurs applications
Mécanique des fluides et énergétique
Code ID :9880
Déposé par :Armand JODON DE VILLEROCHE
Déposé le :15 nov. 2023 15:57
Dernière modification:15 nov. 2023 15:57

Versions disponibles de ce document

Modifier les métadonnées de ce document.