1D Flow Optimization Over Lateral Weir Results in Water Going Wrong Direction?

classic Classic list List threaded Threaded
2 messages Options
Reply | Threaded
Open this post in threaded view
|

1D Flow Optimization Over Lateral Weir Results in Water Going Wrong Direction?

Jared D.
I'm working with a complicated steady-flow 1D model (not built by myself) that has three flow splits from the main channel with lateral weirs allowing flow exchange between the splits and the main channel.  The model builder determined the steady-state flows in the various channels at various locations by running the automated flow optimization routine in HEC-RAS.

I'm looking at one particular lateral weir between a flow split and the main channel.  As received, the steady flow model results show a gradient from the main channel to the split (i.e. higher WSELs in the channel relative to the split), and the split does indeed gain flow from the channel at that location.  It makes sense to me that the split would be the one gaining flow here given the gradient from the channel to the split.

However, I decided to re-optimize this particular lateral weir in order to reflect some minor changes I made to the geometry in the vicinity.  My solution converged, but the results were unexpected.  What I ended up with is a reversed gradient (i.e. now lower WSELS in the channel relative to the split), however the model continues to show the split gaining flow from the channel at that location.  This doesn't make a whole lot of sense to me: if the WSEL in the split is actually higher than the WSEL in the main channel, why then is the split gaining flow from the channel?  Shouldn't the reverse be occurring (i.e. the split losing flow to the main channel)?

If anyone has thoughts on this, I'd appreciate it.
Reply | Threaded
Open this post in threaded view
|

Re: 1D Flow Optimization Over Lateral Weir Results in Water Going Wrong Direction?

Luis Partida
If this is a steady state model your best bet in finding the solution for debugging issues is to split up the complicated model into multiple steady state models. If you can isolate the reach and find the solution then try those parameters back into the complicated model