Consistency and bicharacteristic analysis of integral porosity shallow water models. Explaining model oversensitivity to mesh design - Archive ouverte HAL
Article Dans Une Revue Advances in Water Resources Année : 2017

Consistency and bicharacteristic analysis of integral porosity shallow water models. Explaining model oversensitivity to mesh design

Résumé

The Integral Porosity and Dual Integral Porosity two-dimensional shallow water models have been proposed recently as ecient upscaled models for urban oods. Very little is known so far about their consistency and wave propagation properties. Simple numerical experiments show that both models are unusually sensitive to the computational grid. In the present paper, a two-dimensional consistency and characteristic analysis is carried out for these two models. The following results are obtained: (i) the models are almost insensitive to grid design when the porosity is isotropic, (ii) anisotropic porosity elds induce an articial polarization of the mass/momentum uxes along preferential directions when triangular meshes are used and (iii) extra rst-order derivatives appear in the governing equations when regular, quadran-gular cells are used. The hyperbolic system is thus mesh-dependent, and with it the wave propagation properties of the model solutions. Criteria are derived to make the solution less mesh-dependent, but it is not certain that these criteria can be satised at all computational points when real-world situations are dealt with.
Fichier principal
Vignette du fichier
PorosConsistency_ForHAL_170617.pdf (1.54 Mo) Télécharger le fichier
Origine Fichiers produits par l'(les) auteur(s)
Loading...

Dates et versions

hal-01541070 , version 1 (17-06-2017)

Identifiants

Citer

Vincent Guinot. Consistency and bicharacteristic analysis of integral porosity shallow water models. Explaining model oversensitivity to mesh design. Advances in Water Resources, 2017, 107, pp.43-55. ⟨10.1016/j.advwatres.2017.06.008⟩. ⟨hal-01541070⟩
369 Consultations
131 Téléchargements

Altmetric

Partager

More