HEC RAS Dam Break

[jdelliott]

I am in the process of running my first nonsteady state HEC RAS (4.1.0) model for a small DAM BREACH. This is an in stream detention pond with low baseflow (roughly 2 cfs). I am trying to develop a sunny day breach at full pool. The model functions well under steady state and I have developed a number of boundary and initial conditions with the steady state model at 5 cfs. However, the output of the nonsteady state model demonstrates large fluctuations in the energy grade line while under boundary conditions, and once the DAM BREACH is initiated and boundary conditions are removed the water surface elevations are highly suspect. I have tried a variety of scenarios, but nothing thus far has resulted in reliable results. I am in the process of trouble shooting. Depending on the computation interval, the model also commonly goes unstable after the breach. Any advice would be greatly appreciated.

 

[Drew08]

The best advice is to forget everything you know about steady-state modeling. Unsteady-state breach modeling is a whole different animal.

Unsteady-state models hate sudden changes between sections and over time, so some generalizations need to be made. Here are a few suggestions to start with:

1. Channel profile is too variable, average out the longitudinal slopes.
2. Manning’s numbers are too high (0.8?) and too variable, find an average characteristic Manning’s reach wide.
3. Time step is too long (1 hour). Time step should be <1/20 of the rising limb of the hydrograph (<1 minute, <10 seconds is more typical).
4. Initial flow (5cfs) and corresponding initial depth is too low. Try higher flows (50 cfs is a good starting point). As long as initial flow is <10% of peak, the model is still valid, though some states have a maximum initial flow rule.
5. Section spacing might be too large for a few sections (<200 feet is typical). Use interpolates were necessary.
6. An upstream constant inflow hydrograph is needed at the upstream boundary equal to the initial condition.

 

[jdelliott]

Thank you for those suggestions. I am working through them and will let you know.

I have tried to model this by both placing a lateral structure in front of cross sections and with a storage area. Under both scenarios, the water surface tends to elevate at extremes well above the dam height. This occurs right at the dam and continues upstream. This then appears to lead to a surplus of energy and causes numerous problems. Does anyone have an explanation for this? Or more importantly a solution?

Thank you for your help,

 

[Drew08]

Extreme results are from a computational instability, which can be cause by virtually anything. There is usually no one silver bullet for cause and effect. One common instance is a zero depth situation. One extreme causes another. See item 6 above. With downstream discharge occurring without upstream inflow, the system is running dry. I’m sorry to say that trouble shooting an unstable breach model is well outside the scope of what can be addressed in a few typed paragraphs.

Good luck.

 

[jdelliott]

We got the model to run by generating the breach flows in HMS, applying them to the basin, following your guidance, and a great deal of trouble shooting. The results seem realistic. For an unknown reason, the versions of the HEC-RAS model that include the Pond and Dam never wanted to work properly. Thanks again for your help. We have what we need for the project, but I'm still interested to learn why the model has additional problems when the pond is included. It does not seem that HEC RAS should be concerned with all of the potential energy of the water behind the dam, until it breaks and turns kinetic. However, it has my curiosity because these versions of the model reports unrealistic flow depths and energies even under high base flows.