Unsteady Flow with Dams in Series

[Anonymous]

Hi,

I’m having problems with my model – my HEC-RAS geometry model has a reservoir (Reservoir-2) that is connected to two reaches – one downstream of reservoir-2 and one upstream. I have modeled dams in series by adding a time series hydrograph that I copied from HEC-HMS (from the most upstream reservoir (reservoir-1) with a dam breach) and pasted that into a flow hydrograph for the most upstream cross section in my hec-ras model. There are several more cross sections downstream until it reaches Reservoir-2 which I modeled in HEC-RAS – I modeled it to breach within HEC-RAS. When the second reservoir breaches in HEC-RAS at around hour 4, the cross sections upstream of the second reservoir and downstream of the first reservoir dry up. I am guessing that because the second reservoir is emptying quickly from the dam break, the upstream cross sections are drying up – but how do I prevent this from happening to minimize errors. I have a WSE error of about 7′ which is pretty large. The weird thing is, the xs immediately upstream of Reservoir 2 does not dry up, but the xs’s further upstream do. I hope this makes sense – anybody have an idea of how to fix this?

[Vince Moody]

You could try adding some additional flow to keep the sections from going dry. Generally, a thousand cfs is insignificant when compared to the magnitude of flow from a dam break scenario. There are a few other things you could try, but I’ve used that approach numerous times.

[Anonymous]

RAS needs to have a finite depth at all sections to solve the unsteady flow equations. If your channel is going dry, perhaps a pilot channel should be used to insure that there is depth everywhere. This is a modification to the cross sections that provides a lower invert for situations such as the upstream end of reservoir where the cross section goes dry when the water surface at the dam drops below the natural upstream invert. The use of pilot channels is described in Chap. 6 of the HEC-RAS 5.0 User’s Manual.

[cameron]

Another option would be modeling the whole thing as 2D or just 2D upstream of the the second reservoir.

[Author]

Posts