Lake, Dam, Gates... Water level 1

[SMIAH]

I have to prepare the hydraulic study of a dam. This dam is located 1 mile downstream from the outlet of a lake (where it represents the only possible outlet).

On the lake are recorded maximum daily water levels for nearly 100 years now. The maximum instantaneous water levels are also available for over 40 years, but they do not vary dramatically from the daily maximum levels (the max deviation is around 1 inch).

The dam is equiped with multiple gates and it controls the water level of the lake. Thus, closing the gates in order to reduce the peak flows evacuated downstream (e.g. to protect people and infrastructure downstream) has a significant impact on the lake’s water level.

However, the opening height of these gates is rarely noted, and it is difficult to access the rules of exploitations used back in the 50s, for example.

I think that the basic method to estimate the hydrologic conditions in the lake would be to simulate the passage of a flood hydrograph, taking into account the presence of the dam. However, the watershed of this lake has an area of nearly 775 square miles...

I already have estimated the probabilities associated with the recorded water levels using the Adjustment law of Log-Pearson III (WRC). Then I think that I could simulate the water level reached at the dam using the water levels of the lake as the upstream boundary condition in a model like HEC RAS (I have cross section data).

But I doubt that these water levels are representative of the hydrological conditions and this based on the fact that it is possible that the dam operators have used different strategies like, for example, letting the water level of the lake increase by closing the gates in order to reduce peak flows and protect downstream populations as there were less issues related to flooding around the lake back then. And this is not shown in the maximum recorded water levels...

I'm stuck right now.

Tips? Similar situations solved?

 

[cvg]

what is the purpose of the study? Is it routing the IDF through the dam for the purpose of determining the factor of safety / freeboard? Is it for a risk assessment? Is it a forensic investigation? Is it for a hydroelectric or irrigation study? Your strategy would be different for each type of analysis.

 

[SMIAH]

It's for risk assessment.

 

[cvg]

You haven't said how much variation the levels have with respect to the normal operating level. Do they cause the emergency spillway to operate? Are there no gages downstream of the dam that recorded releases? Even if you go quite a ways downstream, you should be able to find a river stage gage that you could use to help estimate what the release from the dam might have been. However, for a risk assessment, if this risk assessment is related to dam safety and not downstream flooding, than you would probably be dealing with 100-year flood or greater. I would assume that for modeling purposes, the lake is full and gates wide open at the beginning of a 100-year flood analysis or for any greater flood.

 

[SMIAH]

Water level of 10,000-year is 7 feet over the normal operating level. We have some records downstream but they do not cover the period of maximum flooding.

"The lake is full and gates wide open at the beginning of a 100-year flood analysis or for any greater flood."

Well, I'm looking for the elevation associated with the full lake level.

 

[cvg]

majority of the dams I have worked on the emergency spillway operates during the 100-year flood. So assuming the lake is full when the water gets to the crest of the emergency spillway may be a good starting point.

 

[beej67]

Agree with cvg's last comment. I'd pick a lake level equal to that which activates the emergency spillway, route the 100 year (or whatever) flow through the emergency spillway alone ignoring the other outlet works (assume clogging) and start with that. Not only is it conservative, it's reasonably easy to model as well as defend.

Full disclosure: my experience with lakes of this size is very limited.



Hydrology, Drainage Analysis, Flood Studies, and Complex Stormwater Litigation for Atlanta and the South East -

http://www.campbellcivil.com

 

[SMIAH]

Thanks for the answers.
I don't have any hydrograph at the moment. I know that the inflow hydrograph is usually derived from a design-storm model (did I say the watershed is 775 square miles…).

From this study we would like to access the different outflow and WSE at the dam associated to floods ranging from 2-year to 10,000-year.

I thought that using frequency analysis from the recorded 100 years of peak instant WSE and relating them to flow with respect to the rating curve of the dam (e.g all gates opened), I would be able to save myself from using a design-storm model and deriving a inflow hydrograph.

Probably not

 

[beej67]

But your data isn't for a fixed rating curve, is it? Didn't you say the dam operators monkey with the sluice gates all the time?

I'm a little out of my comfort zone with this, but here's a thought:

Suppose there's an emergency spillway, and it's got fixed characteristics to it that can't be monkey'd with. Suppose it gets activated at the (X) Storm Event. That would mean that the stage-discharge relationship for storms less intense than (X) is entirely at the whim of the dam operators, but storms greater than (X) have a stage-discharge relationship that is partly due to the emergency spillway and partly due to the other variable outlet works. Storms *much* greater than (X) are routed more through the emergency spillway than the gates, so the gates mean less by ratio to the stage-discharge relationship. At some very large event, the gates mean very little compared to the discharge of the emergency spillway, right?

So plot yourself a stage-discharge relationship with the gates open, and another one with the gates closed, and at some stage those two relationships should start to converge. (plot it on a log scale) Then pick a point near the convergence, call that point the point where "dam operations can be neglected for the purpose of analysis." Note that lake stage, and fish through your 100 years worth of data to see which data points exceed that stage. If any of them do, then you could presumably use that subset of data to do your analysis without having to worry about what the dam operators were doing for those storms. It won't work for smaller storms, but could work for the big ones, which is what you're most concerned with in risk analysis.

Does that seem right? Am I missing something?



Hydrology, Drainage Analysis, Flood Studies, and Complex Stormwater Litigation for Atlanta and the South East -

http://www.campbellcivil.com

 

[SMIAH]

Yes that does seems right (excellent idea I'd add). And yes, I do have a rating curve for different gate opening and height.

Thing is, this dam is equiped with 6x wooden stop logs (logs now fixed togueter) and... there is no emergency spillway (shocking truth revealed). If the dam operator isn't there to move the gate, the water will probably overflow by the crest of the dam (let's hope that this won't happend).

Well there is an opening height between both gate and dam's crest. When left in place, there will be water conveyed through this space. I doubt to find a convergence point though. This might worth a try though.