Numerical model for simulating shallow water hydrodynamics on the GPU using an Adaptive Mesh Refinment type grid. The model was designed with the goal of simulating inundation (River, Storm surge or tsunami). The model uses a Block Uniform Quadtree approach that runs on the GPU but the adaptive/multi-resolution/AMR is being implemented and not yet operational. The core SWE engine and adaptivity has been inspired and taken from St Venant solver from Basilisk and the CUDA GPU memory model has been inspired by the work from Vacondio _et al._2017)
GNU General Public License v3.0
34
stars
15
forks
source link
New Engine (Buttinger et al.) and fix adaptive simulation #48
The previous iteration of the refine linear scheme was wrong. this new one is now, more in line with Basilisk.
New Fill-Halo for fluxes
This new fill halo for fluxes is bug fix from the previous iteration. This scheme now conserves mass.
New Engine
Implement the Buttinger et al. 2019 Scheme as a replacement for the kurganov scheme. The new scheme is identical to the basilisk implementation in BFlood but fixes a Lake at rest issue.
More stable==Faster
This new scheme handles the flow over steep topography very well, removing some spurious velocity that plagued the model time step. Early real world test in Waikanae suggest a 40% speed gain!
Two new tests
I also added 2 test to verify some of the validity f the new scheme:
Parabolic Lake-at-rest
12 scenario of river flow in in a valley (left right top botom uniform coasrse to fine et fine to coarse)
New stability check
An additional limiter applied at advection step prevents negative depth. This seem to help on some instability at partially wet cell or cells drying out rapidely
CyprienBosserelle
commented
3 years ago
Fix flow bug in adaptive case
New refine linear scheme
The previous iteration of the refine linear scheme was wrong. this new one is now, more in line with Basilisk.
New Fill-Halo for fluxes
This new fill halo for fluxes is bug fix from the previous iteration. This scheme now conserves mass.
New Engine
Implement the Buttinger et al. 2019 Scheme as a replacement for the kurganov scheme. The new scheme is identical to the basilisk implementation in BFlood but fixes a Lake at rest issue.
More stable==Faster
This new scheme handles the flow over steep topography very well, removing some spurious velocity that plagued the model time step. Early real world test in Waikanae suggest a 40% speed gain!
Two new tests
I also added 2 test to verify some of the validity f the new scheme:
New stability check
An additional limiter applied at advection step prevents negative depth. This seem to help on some instability at partially wet cell or cells drying out rapidely