Keep my head straight on this: connecting to a culvert under inlet control

[beej67]

Situation:

Existing culvert is fairly over capacity. Watershed is 600 acres of mostly undetained urbanized and suburbanized watershed. FIS says there's 480 cfs in it at the 100 year storm. This causes a pretty big head differential across the culvert - upstream head is about five feet above the downstream head, which lends me to believe the thing is inlet control for the 100 year event.

My development is just upstream of this culvert, and my land plan really lends itself to putting by stormwater management system, which will be an underground concrete vault with an offline cistern, at an elevation well below the 100 year elevation upstream of the culvert. But we're close enough to the culvert that I have the opportunity to not discharge into the creek at all, and carry my discharge into the street, dog-housing the culvert and dropping my water in downstream of the head wall that's pretty clearly forcing the thing into inlet control.

What tailwater depth should I assume in this design, for my stormwater management system? Seems to me, unless I'm thinking about this wrongly, that I should be able to calculate the 100 year HGL in the pipe and use that as a tailwater. Right? Am I missing anything?

The flood profile on the FIS doesn't really show it that way. I'm not sure how they generated the drawing, but it shows the HGL for the culvert being high through the whole culvert, and only dropping off after the downstream headwall. Graphical error in the FIS maybe? (profile attached - Hermance Drive is the culvert in question)

Unfortunately, the FIS was done in XP-SWMM instead of HEC-RAS so I can't easily get at the guts of their model.

Disclaimer: I haven't had my coffee today.



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

http://www.campbellcivil.com

 

[gbam]

Beej - sounds like you need to compute the backwater from the outlet of the culvert to your junction or from the inlet of the culvert to your junction depending on the hydraulics within the culvert. Double check the losses resulting from your proposed junction to the inlet; the additional junction losses may pop the culvert to outlet control which may increase the head water beyond the existing condition. If you don't have a spreadsheet you could use HY-8 to determine the profile through the culvert.

 

[LincolnPE]

480 cfs over 600 acres for the 100 year for an undetained semi-urbanized watershed sounds really low. Less than 1 cfs/acre.

 

[agua84]

beej, are the lines on the FIS depicting HGL or water surface elevation? I see the 10% annual chance profile appears to go through the culvert while the 0.2% and 1% are above, then drop sharply beyond the culvert. Without any more embankment info, I would see the profiles as (highest*) water surface and thus overtopping the roadway, but I could be wrong.

*if the culvert is as you suspect not flowing full, there is another water surface profile inside the culvert (not shown) that you'll need to find to determine the allowable flow you can inject mid-pipe without making things worse.

Then again, looks like the slope is adverse to grade, very possible to be outlet control.

 

[LincolnPE]

Since you are connecting to and discharging into the culvert you need to start your hydraulic analysis below the culvert, using the appropriate tailwater elevation in the channel.

 

[TerryScan]

The FIS profiles show that the culvert is submerged for the 100 & 500 yr event and overtopping the roadway by a foot+.

 

[LincolnPE]

The FIS profiles show that the culvert is submerged for the 100 & 500 yr event and overtopping the roadway by a foot+.

Yep, that's how we design many typical culverts...pass the 10yr and no more than 1' overtopping for the 100yr.

 

[LincolnPE]

Those are water surface profiles. FEMA policy and procedure states the profiles are to be used to determine BFE's. Do not use plan view to try and interpolate and determine BFE.

When you add the connection you are creating a storm sewer analysis scenario because you are introducing junction losses at the connection. Many storm sewer packages now can test for inlet control also using the HDS 5 equations.

 

[beej67]

Thanks for all the input.

I agree 480 cfs seems low, but a lot of the watershed is in undeveloped B soil stream buffers, and a lot more consists of university athletic fields, so that probably helps some. I'm inclined to stick with the FEMA flow rates, especially given that their watershed area and the one I did as a check are nearly identical.

A HY-8 analysis utilizing the FIS flows and the tailwater shown in the FIS profile shows inlet control (as I suspected), overtopping, and a lower HGL in the culvert. I'm using a 'n' of .015 to account for junction boxes in the line. I'll utilize the pipe HGL shown in the HY-8 run as my system tailwater.



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

http://www.campbellcivil.com

 

[LincolnPE]

I should be able to calculate the 100 year HGL in the pipe and use that as a tailwater.

No, because your new incoming junction flow you are inserting into the culvert will change that existing culvert HGL.

 

[beej67]

.015 Mannings should give me enough SWAG without counting it, at least according to Chow.

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

http://www.campbellcivil.com

 

[LincolnPE]

Not SWAG...that's a WAG!

There are two problems with that:

1) If you try and add your flow at the inlet of the overtopping existing culvert, most of your flow will get eaten up by the weir flow that is running to the 3/2 power vs the inlet running to the 1/2 power.
2) If this culvert is not running supercritical, your connection may create tailwater control below your connection, then possibly all the way back up to the inlet of the culvert.

This all affects the tailwater hydraulics of your incoming system.

I suggest analyzing the system together as one. Calibrate your model first to the existing condition, then add your condition.

 

[beej67]

I probably wasn't clear earlier.

I don't intend to add any new junction boxes to the culvert run, just to drop my flows into one of the existing ones along the reach - there are several downstream of the headwall. I'd be diverting existing flow from coming into the culvert, managing it, and then discharging it into the culvert, for a watershed much much smaller than the culvert's watershed. And because the flow is detained, there should be insignificant change in velocities, and therefore head losses, along the culvert run. Plus, I'm not required by code to check the culvert, because of the ratio of my watershed size (very small) to the culvert's (very large). I'm merely trying to figure out the tailwater on my discharge pipe. For that, the Chow (1959) recommendation of n=0.015 for "Concrete sewer with manholes, inlets, etc, straight" should produce an accurate HGL for the calculation. Better than SWAG, actually.

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

http://www.campbellcivil.com

 

[LincolnPE]

Still the same principles.

I think that 1959 Chow recommendation was for the singular culvert itself, not to be used like you are proposing for modeling incoming sewer junction hydraulic systems.

 

[beej67]

I'm not so sure. If you're actually modeling a junction loss at each node, then there would be no need to adjust the Manning's to account for the junction losses at each node.

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

http://www.campbellcivil.com

 

[LincolnPE]

Of course..

Still don't think we are communicating.

 

[LincolnPE]

You propose to just slap down n=0.015 to cover all lateral branch flow conditions...

No one can argue the change in momentum and disturbance caused by a 1cfs lateral flow is different than that caused by a 25cfs lateral flow.

The standard engineering method is to calibrate your model to the FIS (or if the use of new and better data can be justified), for the reach of interest, then add your proposed condition. This simple culvert scenario would not take that long to do in this manner.

Is it a software issue? Take a look at the FEMA approved software for some ideas on which software can help you.