Stormdrain Design Procedure Hypothetical Question

[Thom93]

I will start off with a hypothetical example, and then ask my questions rearding stormdrain design procedures.

The hypothetical example is this: there is a single storm drain system with a single on-grade inlet at it terminus. The inlet bypasses offsite. The drainage area to the inlet is large; the runoff is 100 cfs by the Rational Method only. However, the inlet is small; its capacity is only 2 cfs. The remainder, 98 cfs, bypasses the inlet and goes off-site. This is the only information that is given, so the Tc, gutter cap., intensity, runoff coeff., design storm, etc...is not important to the example.

Ordinarily, you would go through an inlet sizing and spacing exercise, but in this example, we cannot. The inlet will pick up 2 cfs under all conditions.

Now, according to HEC-22 and AASHTO Drainage Design Criteria, the assumption is that the storm drain system picks up all of the drainage basin. There is no inlet bypass in the storm drain design. All of the basin contributes and I have to size my system for 100 cfs. StormCAD and Hydroflow Storm Sewers will do the same thing; they will size a storm pipe for 100 cfs. HEC-22 stresses the importance of Tc, so as you move through the system, your Tc increases, Intensity decreases, Q decreases, etc...

You can set the programs to use only inlet captured flows, however. If you use inlet captured flow, the Q will be 2 cfs, but the Tc will not change. Try it yourself. space 3-4 inlets on grade and use inlet captured flows only and you'll find that the system will just "add up" the Qs and Tc will remain the same throughout the system.

So, here's the question: what would be the correct method?

Using HEC-22 does not seem intuitively correct because the inlet can only pick up 2 cfs. What happens to the bypass? It doesn't flow down and enter the system downstream; it all gets in the inlet using HEC-22 and so the pipe would have to be very large.

Using inlet captured flows seems correct, because the pipe can be very small and the bypass can just get into the system at the next inlet, but this is not the correct "procedure". It is not technically correct per HEC-22.

I don't need to know how to design storm systems, I just want to know how others have handled situations where inlet bypass kept accumulating downstream.

Thanks.

 

[gbam]

I will try to address this. First by definition the Tc cannot stay the same of which you mentioned. The Tc will increase by the travel time from inlet to inlet either within the stormdrain system or overland, if you are using the Rational Method. As a result your intensity will decrease and so will your r/o discharge.

As to HEC22, I do not recall anywhere in the manual it stating that each inlet will capture 100% of the discharge unless you size it that way. Please point out where in the manual you found this requirement. You may want to step back and reexamine the process and maybe crunch it out by hand first then employ the use of computer modeling.

Two things are going on if your system is connected with downstream inlets and pipes. You have to keep track of the pipe travel time and the overland travel time and evaluate each inlet/node accordingly. Typically I ignore the overland traveltime and add the bypass directly to the computed runoff for the next downstream structure. This is a conservative approach. Typically onsite R/O will be in the flowrate range of a few cfs not 100. With a discharge rate of 100-cfs I consider this an offsite condition and potentially capture is upstream and try to keep it seperated from an onsite facility.

Hopefully i understood your situation and answered it appropriately. Good Luck!

 

[Thom93]

Thank you for your comments, but what happens to the bypass? At the inlet, Q=CiA, not Q=CiA-Bypass (the Q=CiA-Bypass is only true for inlet spacing).

And at the next junction, Q=CiA (where Tc increases and i decreases) but not Q=CiA-bypass.

So, HEC-22 states implicitly 100% pickup at the inlet, because there is no bypass. It all gets into the system. Otherwise if time of concentration increased, how would you calculate Q=CiA-bypass at some junction far downstream from the inlet? You can't.

The result: I have to size my pipe for 100 cfs even though the inlet can only pick up 2 cfs.

Again, we're talking hypothetically. How do you factor bypass into your system? It doesn't make sense to size a pipe for all the runoff when 98% of the runoff bypasses the inlet.

 

[gbam]

Let's think about this. You can adjust the area by the factor to properly account for the major bypass. Realisticly you should not encounter this condition when evaluating a system as in HEC22.

 

[Thom93]

The reason for my questions is that one of the SD requirements at local municipality requires the inlet bypass to enter the system at the next downstream inlet. but the problem is you end up having to mix and match different Tc. The inlet bypass is based on a Tc= 15 min. but the SD system might end up having a Tc of 22 minutes. and I can't find any data to support that approach.

But, you may be on to something. It might be possible to size the upstream catchment so that the runoff is only what the inlet can hadle. the rest of the catchment would be added to the catchment of the next downstream inlet. That way, the whole system can use the same Tc. Good idea. Thanks. But would it be an acceptable practice?

 

[beej67]

I think you need to recheck your bypass targets in your stormwater model.

I know for a FACT if you enter your inlet properties properly in StormCAD, whatever bypasses your inlet won't be applied to your pipe system, so you'd only be sizing your pipes for 2 cfs. You'll want to enter all the appropriate inlet geometry for an on-grade inlet and set something downstream (outlet I think) as your bypass target. I do not think the software allows flow bypasses for inlets in sag, because they're in sag, they're supposed to catch everything given enough ponding depth. The software is set up so that the default bypass target is the next downstream inlet in your pipe system, but you can point it manually anywhere in your network, such as if the surface bypass flowed overland to an inlet on a different branch of your network.

That's for StormCAD 8. My experience with the new fangly dangly GIS version is limited, because it takes twice as long to input anything and is four times as complicated to interpret and get running, so I just use the old(er) one.

Try that and report back.

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

http://www.campbellcivil.com

 

[Thom93]

Beej67: Thank you for your comments.

You are correct. You can specify the bypass to go into another downstream inlet. But what happens to the Tc throughout the system? It will stay the same to the outfall, because the software cannot know what the overland flow time is from one inlet to the other. So, if you specify inlet captured flows, the software simply sums the flows and ignores Tc. But doing this is not the HEC-22 procedure.

So, the question still stands. You can specify that bypass goes to the next inlet, but is this an appropriate design procedure? I cannot find any info that supports this.

 

[TerryScan]

Thom93,

You repeatedly state that the software "sums the flows". It does NOT just sum the flows. It recomputes Q at each structure based on system CA and system Tc. It is often less than the simple sum of flows (due to system Tc vs sum).

In an attempt to address your observation that the Tc can stay the same at the outfall:
If one has Drainage areas (DA) A,B, & C in series with respective inlet Tc(s) of 10m, 5m, 15m and in series, and the pipe flow time from A to C is 4m, the system Tc will be 15m (because it is the longest in the system). This which would not change if the Tc for A was reduced, but would if the Tc for A or the pipe flow time was increasd more than 1m.

It just occurred to me that we may be missing your main concern. Is your concern that the flow time of bypass flow can affect the system Tc? A conventional Rational model is not routing hydrographs. It is more of an attempt at snapshots of peak flows. One is normally designing inlets to capture the majority of flow going to them. The bypasses are accounted for, but typically not the greater influence on the system.

In the confines of your example, 2cfs go in the inlet, through the pipe, and 98cfs is bypassed...follwoing HEC22 guidelines..end of story. Your example is an extreme.

If one needs to deal with both the bypass and pipe flow downsteam..that's another story. Normal modeling may not apply in some other extreme example, but this other example is unlikely to be a sensible real-world design.

 

[Thom93]

I think everybody is getting lost in the reeds and that may be my fault for not articulating the problem well enough. The extreme example is just to illustrate a point. I could have said that 10 cfs enters the inlet and 0.10 cfs bypasses. It doesn't matter what the actualy values are, because the point is :what happens to the bypass?

(TerryScan. You are right and you are wrong about the Tc. The Tc of the system will not be 15 min. Inlet time is 15 min. System time will be the Tc to the inlet(15 min) + pipe travel time)

What I am trying to say is that bypass is normally not accounted for in pipe design.

In HEC-22 the flow rate in a pipe segment is Q=CiA. It is not Q=CiA-bypass (this is pipe sizing chapter 7 not inlet spacing chapter 4). And for the next inlet, it is not Q=CiA+bypass. It's not. I'm sorry, but it's not in there in chapter 7. Most people think it states this because this seems intuitively correct. The pipe should only carry what the inlet can capture, right? Wrong. HEC-22 states Q=CiA not Q=CiA-bypass.

Test it yourself with your software. Make a system with 3 or 4 inlets on-grade so that each inlet bypasses to the one downstream. Make the first drainage area really large. make all of the inlets really small. Really exaggerate the values. Check the results and you will see that the first pipe segment will have a very large flow. It will be much higher than the inlet can handle. Because the software is modelling HEC-22 where Q=CiA. and at a downstream junction, Q still equals CiA and not Q=CiA-bypass. This is how most designers do this. Most designers assume that all of the flow from a drainage area will enter the pipe at the first segment.

(You can set a parameter in the software so that the pipe system will model inlet captured flows only, so that Q=CiA-bypass, but is not normally done. This procedure is not supported in HEC-22 Ch.7)

Check AASHTO, check your State's DOT design guidelines. They are all the same for pipe design. The assumption is all the flow gets into inlet at each drainage area.

 

[beej67]

Ahh, I get it now.

Here's what I THINK StormCAD does: (from memory here)

I do not think it sums the flows. I think it figures an efficiency ratio (percentage) and applies that percentage to the "CA" of the inlet, then sums those "CAs" down the network, then uses the Tc at any point in the network to determine flow via Q=I*sum(CA).

So in your first-post example of 100 cfs to your inlet, you WOULD actually need to know the watershed properties generating that 100 cfs to understand what the software is doing. Lets say A=20ac, C=0.5, I=10in/hr, for round numbers. The software would calculate the CA separately (=10) and the inlet efficiency (2%) and would assign 0.02*CA to that node in the network, and 0.98*CA to the downstream node in the network provided that inlet could carry it, otherwise pass it down.

Etc.

I'm not 100% positive, but the StormCAD help menus should be able to explore this in more detail, and I'm reasonably sure that's also how HEC22 handles it, though I'm not sure. In any case, this approach of apportioning the CA by inlet efficiency should be the correct approach to make sure your Tcs are right.

Or right enough anyway. The whole thing is sorta hoodoo when you realize that you're modeling each successive downstream pipe for a different storm event.

I do not know anything about Hydraflow Storm Sewers, but my opinion of that software suite is very poor, and it wouldn't surprise me at all to discover that it was simply dumping bypass flows off into space.

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

http://www.campbellcivil.com

 

[TerryScan]

Thom93, I may be mistaken, but in my example the Tc for the drainage are to 'C' exceeds the sum of other Tc's + pipe travel time...so it would prevail. From HEC22:

The time of concentration for pipe sizing is defined as the time required for water to travel from
the most hydraulically distant point in the total contributing watershed to the design point.
Typically, this time consists of two components: (1) the time for overland and gutter Flow to
reach the first inlet, and (2) the time to Flow through the storm drainage system to the point of interest"

I am interpreting the point of interest to not include the downstream pipe time (and it seems Stormcad makes the same assumption)

And..yes...Q=CIA. A being the contributing watershed. As beej67 points out, Stormcad uses assumptions to get at what's contributing by adjusting CA by the capture percentage.

I have an older version of Stormcad (4.1) I opened an old project:

At one inlet: DA=0.31ac, C=0.44 (CA=0.14ac), and Tc=15min generating 0.68cfs. The inlet intercepts 0.53cfs with a capture efficiency of 78%. Stormcad calculates an intercepted CA of 0.11 (0.14 x 0.78).

The upstream piped flow is 7.89cfs with a CA of 2.12ac and Tc of 28.2min.

The system CA is 2.22 (there must be rounding error from 0.11 + 2.12 =2.23...but this is definately not 2.22+0.14), system flow time=28.2min with a resulting flow of 8.28cfs. This represents only the area capture...the bypassed area will be accounted for at the appropriate downstream point. This is absolutely supported as HEC22 chapter 7 specifically calls for contributing area. If it bypasses, it is not contributing.

Stormcad's method may not be explicitly described in HEC22, but I do not see where it is a "violation" of any directives in the manual either.