strange response in unsteady flow, at high max stage

[Anonymous]

Hi,
Im trying to run an unsteady flow model on a stream but keep running into a problem. the stream has two tributaries each abou 7km long with cross-sections spaced 25-100 m’ apart from each other. most cross-section are concrete and trapeze-shaped. the slope of the streams are about 0.0015. they are connected by a junction to one low slope stream (leading to the ocean). Im using HEC-RAS and not GEO-RAS (I dont know if it matters…)
When I run steady state flows, it all works really well, even for very high flows (600 m^2/s), it also works well when I run unsteady state (hydrograph) with maximal low low flows (upto about 250 m^2/s), however when I try to run unsteady state of higher flows (with a max stage of 400-600 m^2/s), I get strange responses. At certain points I will get at a certain cross-section very high flow that does not make sense.
I read around but did not find any explanation or fixes to this strange phenomenon. I noticed that when I decreased the density of cross-sections, it enabled me to get to a higher max stage without the model “blowing up”, but couldnt find any rule by which I this works or happens
I can take a screenshot of how this looks if needed.
Do anyone have an insight, explanation or possible fixes for this?
Thank you,
Ronel

[Anonymous]

btw my initial conditions are a hydrograph from the upstream and normal depth (of 0.00015) at the downstream

[jadams]

Your model is likely going into critical flow and the transition from subcritical to critical can often unbalance the model. You can attempt to add cross sections, modify mannings “n”, modify your time step and several other things to mitigate this effect. For time step use the rule of thumb of 20 times less than floodwave or the courant Condition. For spacing you may check the Samuels equation to give you an idea if you need to interpolate a few more.

[Author]

Posts