Unsteady Model extreme jump or drop in velocity of flow

[Anonymous]

Ive run a dam breach model that is processing stable but its giving me some questionable results. A relatively smooth profile with mostly the same parameters (mannings “n”, etc.), cross sections are nominally similar, yet all of the sudden I get a sudden spike in velocity and serious drop in WSE for a section or two. then it will iron it. Upstream of this the velocities drop to 1 or less and the WSE flattens to unreasonably high number. Error and warning messages give the standard “may need more cross sections”, but ive interpolated down all I can without it going unstable. Have not encountered this before, need some suggestions.

[Anonymous]

I have seen this before for dam break models. It seems to happen more easily for steep streams. My understanding is that the leading edge of the dam break has such a rapid change that the numerical solution goes crazy.

It sounds counter intuitive, but try going to the other extreme. Spread your cross sections farther apart and increase the time step. This spreads the leading edge out in time and space and may keep it from jumping. The longer time steps lose a bit of accuracy. If this approach works, try to get the cross sections back as close together and the time step as short as possible without the extreme jump.

As another possibility, I have found that the 2D solution is sometimes more stable than the 1D for dambreaks.

[Anonymous]

Thanks much, I will give it a whirl

[Anonymous]

just wanted to let you know that worked!! Thanks so much

[Author]

Posts