I am modelling a steep channel of constant slope operating under supercritical flow conditions between two small weirs. I have set up two different models to represent this:
1. A single link between the upstream and downstream weirs.
2. I have also modelled the channel as a number of smaller links with the same slope separated by nodes spaced arbitrarily along the channel length
The two models give very different results. The former results in an unstable channel (oscillates between sub- and supercritical flow) and a maximum velocity in the channel of around 1.5m/s. It seems to me that the velocity in the upper section of the channel should be higher. I assume this has something to do with the manner in which SSA averages the velocities along the entire channel link.
The second model gives more realistic velocities in the upper channel links (around 3.6m/s) and a hydraulic jump and transition to subcritical flow upstream of the weir in the lower channel. This seems more realistic. I think the trick here is to make sure that the downstream link is long enough to enable realistic formation of the hydraulic jump in the model.
Can anyone give any feedback on the above? I would be interested to here any suggestions for improvement, similar experiences, advice etc.
Thanks in advance.
Not sure what SSA is doing, but SWMM isn't the best tool for this reach due to the intertia term handling choices in the St Venant equations, etc.,.... if you are really focused on this reach.
If you are really focused on this reach, then use HEC RAS and interpolate sections, etc. for stability as the subsequent feedback from Warnings may suggest.
I can see though too if you want to WAG a bit it in SSA/SWMM due to this being part of larger network routing...
Hope this helps...
Thanks Fred. I will model the link in HECRAS as well and do a comparison. As you suggested, the link is part of a larger network, so I would like to continue using SSA, but I will see what I can do to make the link behave as HECRAS says it should.
Much appreciate the advice.
Thanks for the offer to have a look at this Matt. I will firstly do as Fred suggests and model the link in HECRAS. If I still can't work out what is going wrong, I will upload my model for you to take a look at.