Simple Dam Breach using Hec-Ras 1

[red9]

I am relatively new to Hec-Ras, and have a dam breach to perform. Using NRCS SITES program, I have determined the maximum water surface level of the dam. I can input cross section data; steady flow analysis works using a trial flow, but unsteady flow analysis does not. Looks like I should be able to input the maximum stage as a stage hydrograph at the upstream section of the reservoir, normal depth slope at the downstream of the breach area, and it should run. Simple earthen weir. But it will not run - error messages at several cross sections. Cross section data is high enough. Any suggestions?

 

[Drew08]

One: there is no such thing as a simple unsteady state model, and two: a dam breach model is one of the most difficult unsteady state situations. Even the most experienced dam breach modelers spend significant time trouble shooting.

If you really want to get experience with dam breach studies, there are few organizations that offer courses. There are far too many potential problems with corresponding tips and tricks to cover here.

Your state dam safety authority may provide some guidence. Many states have pre-set rules for performing dam breach studies including recomendations for some of the more uncertain breach parameters.

Otherwise, this thread might be helpful:

http://www.eng-tips.com/viewthread.cfm?qid=243136&page=1

The biggest misconception I've seen is making the assumption that the model should work because the model exactly matches reality. Remember, reality changes during a dam breach and many parameters have significant uncertainty. Many of your typical assumptions about roughness, channel geometry, etc made when doing flood studies is irrelevant during the extremes of a dam breach. You need to consider the effects of highly turbid and turbulent water laden with significant debris (not leaves and branches, but trees). More importantly, you may have to take some artistic discretion with these uncertain parameters to get the working model.

Start by simplifying: smooth out longitudinal slopes and spatial changes in roughness factors. Avoid sudden expansions or contractions. Any kind of sudden change is not your friend. The extreme flood wave of a breach will likely “wash out” these sudden changes anyhow in reality so don’t sweat trying to match exactly pre-existing conditions. The model accuracy is at best perhaps 1-2 feet, but changing some parameters (channel inverts, etc) by 0.1-0.5 feet to smooth out the model can make a big difference.

Good Luck.

 

[red9]

Thanks for the reply and thread reference. As a further note, if I use a stage hydrograph at the upstream end, and a uniform lateral flow beginning at the first full station in the lake, downstream to the embankment, the program will run, results look reasonable, but I am unsure whether they are correct. May try the Maryland version of SMPDBK for comparision I do need to add additional sections downstream of the embankment. Sections are close enough in the reservoir.