BASEMENTBasic Simulation Environment for computation of environmental flow and natural hazard simulationLaboratory of Hydraulics, Hydrology and Glaciology (VAW)ETH Zurich |
You are not logged in.
Pages: 1
Hi everyone
I would like to add a further chapter to the mass balance problems saga and ask if any of you were able to solve it or implement some tricks that made the problem go away.
I am modeling a flood with only hydraulics (no sediment) over a pretty big area. The first steps generate mass balance problems (h < 0m) over a handful of cells. This seems to be some sort of noise, that goes away pretty soon and does not impact either the rest of the calculation nor the end result.
The problem with this is that mass balance problems are not admissible when using GPU. The simulation is aborted right away. With CPU there seems to be no problems.
It has been already said here that mass balance problems can be ignored if the results are plausible, but I find it really unfavorable that one is compelled to use the much slower CPU for an error that does not really exist.
Did any of you already experience this and were able to solve it somehow?
Offline
Hi Liliana,
although this is not encouraged, you can deactivate the safe_mode in the hydraulics parameters and also run your simulations with single precision. This should be enough.
Anyway, try and understand which cells are causing the mass balance problems and try to correct their level/slope to avoid the mass balance problem.
Cheers
Matteo
Offline
Pages: 1