Improvement of the hydrostatic reconstruction scheme to get fully discrete entropy inequalities

Christophe Berthon (1) , Arnaud Duran (2) , Françoise Foucher (1, 3) , Khaled Saleh (2, 4, 5) , Jean de Dieu Zabsonré

1 LMJL - Laboratoire de Mathématiques Jean Leray
2 MMCS - Modélisation mathématique, calcul scientifique
3 ECN - École Centrale de Nantes
4 ICJ - Institut Camille Jordan
5 UCBL - Université Claude Bernard Lyon 1

This work is devoted to the derivation of an energy estimate to be satisfied by numerical schemes when approximating the weak solutions of the shallow water model. More precisely, here we adopt the well-known hydrostatic reconstruction technique to enforce the adopted scheme to be well-balanced; namely to exactly preserve the lake at rest stationary solution. Such a numerical approach is known to get a semi-discrete (continuous in time) entropy inequality. However, a semi-discrete energy estimation turns, in general, to be insufficient to claim the required stability. In the present work, we adopt the artificial numerical viscosity technique to increase the desired stability and then to recover a fully discrete energy estimate. Several numerical experiments illustrate the relevance of the designed viscous hydrostatic reconstruction scheme.

Domaines: Analyse numérique [math.NA]; Equations aux dérivées partielles [math.AP]

Christophe Berthon, Arnaud Duran, Françoise Foucher, Khaled Saleh, Jean de Dieu Zabsonré. Improvement of the hydrostatic reconstruction scheme to get fully discrete entropy inequalities. Journal of Scientific Computing, 2019, 80 (2), pp.924-956. ⟨10.1007/s10915-019-00961-y⟩. ⟨hal-01952341⟩