Water Levels at a Dam

[chicopee]

Applying HEC-RAS 4 steady flow conditions on a 3.83 miles of a sinous river above a "run of the river" dam(low head) with little useable storage capacity. The dam is fitted with two sluice gates.
The HEC-RAS tells me that with a flow of 6000cfs and the sluice gate fully opened, the water level at the dam is 3' higher than when the sluice gates are opened. Is that possible?
Note, that my river stations start at the dam and that the property of interest is ~2 1/2 miles upstream. The water level results there were what you would expect with the gates fully closed and opened.

 

[chicopee]

Oops! first paragraph statement "when the sluice gates are opened" should have been "when the sluice gates are closed"

 

[cvg]

is the dam diverting any water? How deep is the water going over the dam? what is the capacity of the sluices? how wide and high and deep is the dam and river? What is the slope of the river? Is sediment filling the river upstream of the dam? How have you set this up in HECRAS? Depending on the hydraulic conditions, opening the sluices may have a large or a small effect on the water level at the dam and potentially little to no effect 2.5 miles upstream. But how could the sluices possibly cause the water level to rise, especially when you have not modelled anything downstream of the dam?

 

[Drew08]

Double and triple check all parameters and assumptions. One bad assumption or misunderstood entry could cause crazy things to happen.

If all entries are right, it could be the result of each type of flow, overtopping and sluice gate flow, interacting with the boundary condition differently. Because of the uncertainty associated with any boundary condition assumption, it is good practice to extend the model at least 0.5 miles downstream.

Another option, to ensure the stage at the reservoir is as intended, develop a stage-discharge rating curve at the dam using basic dam hydraulics, apply that as your downstream boundary condition and delete the dam element. With your location of interest at least 0.5 miles upstream, any inherent error in the computation would be mostly "washed out".

A sensitivity analysis would be easy to perform to test impacts of any uncertainties (i.e. discharge coefficients) on the result at your property of interest. I'd bet you'd see little influence beyond the inherent error in the model methodology.

 

[chicopee]

I am in agreement with starting the model at least 1/2mile below the dam and I'll do that.
The dam was constructed to power two hydraulic turbines during the 1800 and the sluice gates are designed to maintain a minimum flow below the dam and the turbine outfall for aquatic life.
At the moment I can not go into more details about this dam and I'll probably have more questions as time goes.