Divided Flow Warning

[Anonymous]

I am currently trying to troubleshoot a dam break model where the dam outlets into the main river but in the floodplain is a fish hatchery that includes three storage areas in series just outside the main channel.

My current setup has these areas modeled as permanent ineffective flow areas considering the fact that those areas will already have water in them for spawning fish, etc. I use a slightly higher Manning’s value for this overbank as well.

My model is blowing up at the first sign of a dam break, this may or may not be the cause of why it is blowing up but I still question it none the less. Why is RAS still computing divided flow here even with the ineffective flow areas? Should I just forget the ineffective flow completely and manually put a solid line across the top of the storage area?

Thanks!

[Chris G.]

Steve-
One of the most common reasons for a dam breach model blowing up right at the start of the breach is sudden increase in both the local and convective derivatives in the St. Venant Equation of conservation of momentum. In other words, stage and flow are changing quite suddenly over a short space and a short period of time and RAS doesn’t do well with this under default conditions. Addint more depth to the base conditions in front of the dam breach flood wave can help to stabilize this condition. You can add more depth by:
1. Providing more base flow.
2. Increase Manning’s n values.

You can also help to stabilize the model under these conditions by dampening out the local acceleration terms. You do this by checking the “Mixed Flow” box in the unsteady flow analysis window. Please read up on this before using it so you understand what’s happening. By the way, this sometimes makes things worse, but it’s easy to try, so see if it works.

Read these posts for more info…
http://hecrasmodel.blogspot.com/2011/04/mixed-flow-regime-options-lpi-method.html
http://hecrasmodel.blogspot.com/2013/10/stabilizing-dynamic-unsteady-hec-ras.html

My guess is the permanent ineffective flow areas, if set up properly, are not causing any trouble.

Good Luck
Chris
@RASModel

[Anonymous]

Good to hear Chris, I ended up changing the LPI factors after interpolating some additional cross sections near the critical flow areas and the model runs stable now. The problem, I believe, was that I was breaching the dam down to the toe even though my actual cross section did not extend down to the toe, so I think this was confusing HEC-RAS and causing the model to blow up instantly when the breach occurred. I ended up changing the bounding cross sections to include this breached area and the model was able to run with some instability which I corrected with the additional XS/LPI factors as stated earlier.

Thanks for the help as always!

[Author]

Posts