Major/Minor System Modeling

Andrea;

We also recognized that inlet capacity can be an important parameter. We found that old curb inlets, without grates in the gutter, would not pass 5 year return level peak flows and round inlets were similar. All curb and gutter combination grates had ample capacity.

Even where inlet capacity was ample it is still useful to include in the models if you include street storage. It is necessary to build up some amount of depth on the street gutter before inlets are flowing at significant rates and this street storage has a dampening effect on peak flows - not much for high capacity inlets because not much storage depth is required, more so for lower capacity inlets where deeper depths are required to pass peak flows.

Since you don't state it, I wonder if you are referring to coarse modeling or fine modeling. In either case you could accumulate inlet rating curves and street storage area/depth curves for coarse catchments.

In our fine modeling we are accumulating a rating curve to cover form one to 6 inlets at a model inlet location and applying a single area/depth curve. This smoothes out the inlet hydrographs and makes it a simple matter to model inlet restriction during relief design.

If your drainage basin has mostly large capacity inlets then you will see minimal effects. It is a large amount of work to add inlets and street storage to models so you should investigate the probability of seeing worthwhile benefits. I would do a detailed examination of at least two catchments to see the benefits of such detailed modeling and then develop a method of doing the modeling.

We have added what we call an Intermediate Extran model to our work process. The surface catchment Runoff interface file is directed to this intermediate model which contains only street storage curves and inlet rating curves. Each 'inlet' is actually an outfall from the intermediate model for this purpose as it's computed flows are routed to the pipe Extran model via an interface file. Thus it is necessary to have larger array sizes than standard as many outfalls are required. We are working with the regular EPA SWMM engine and using PCSWMM.

If you need more help, let me know.

Reinhard Sprenger
rsprenger@UMAGROUP.COM

Reinhard -

Your "Intermediate Extran Model" is very interesting. As I understand, this is generated by new Fortran Code that reads the RUNOFF.INT file and produces its own .INT file for input to EXTRAN. So then, it routes the runoff to storage based on the storage elevation curve, and routes discharge to EXTRAN based on head-discharge curves. How does it handle nodal flooding beyond capacity of the Storage curve?

I suppose it would be possible to implement the same idea within EXTRAN, using a stoage elevation curve and a pipe connected to the system with a Manning's n equivalent to the loss through the inlet. I am sure you considered this; could you comment on the advantages of the "Intermediate Extran Model"

This sounds like a great tool and I am sure many folks would like to try it. Any possibility of releasing it, for sale or otherwise?

Regards

Ron Kilmartin
ronkilmartin@attbi.com

Ron;

Actually this isn't a new tool - just a way of using the regular SWMM.

Our intermediate Extran model is just an extra model that we insert into the process.

For each modeled inlet location it contains; - a variable storage node - a Type 4 pump

The variable storage node is based on actual field surveys - shoot enough points to plot 0.05 m contours around a set of inlets. Almost all of our inlets are sag locations. In order to save a lot of time and expense we use a bit of artistic license, I call it executive decisions. I get about a dozen locations surveyed to cover the range of differing pavement widths, asphalt overlays, slopes, etc. Then I assign a category from A to E for 5 of the curves - even though locations appear quite different many of the depth/area curves are close together when plotted. This simplifies applying this procedure.

Then I take photos of the 5 typicals and drive the whole study area and assign a curve number to each inlet location. There are a few locations that can't be generalized like this so I get them surveyed separately for a total of about 15 curves but most are an A to E.

We start our studies by having team members drive the entire area with air photo maps onto which we note pavement highs, inlets (including which type of inlet), parking lot drainage directions, connected/disconnected roofs, gravel lots, lot drainage splits, etc.

From the drive around we have enough info that, combined with other drawings, we can draw our catchments. Since we have the number and type of inlet for each catchment it is a simple thing to import the catchment input parameters into a spreadsheet. The spreadsheet is then modified to include the number and type of inlets in each catchment.

For each type of inlet we compute a rating curve and thus the spreadsheet can accumulate a total rating curve for each catchment.

The rating curves are output as a PRN file in the format that the Extran Type 4 pump wants.

I have a SWMM engine with array sizes increased to 1,000 pumps and outfalls for the intermediate Extran model use. Each inlet location is thus a simple routing of the Runoff flows through a variable storage node controlled by a Type 4 pump to an outfall. This then becomes the inflow interface file to the pipe model. That's why I call it the intermediate Extran model - it is inserted between regular Runoff and Extran models.

There is no technical reason why this work requires a separate model - we could just insert the variable storage node beside each inlet node and connect then to the pipe node by a Type 4 pump. That way you wouldn't need a new engine with larger pump and outfall arrays.

We have broken this routing out into a separate model for purely logistical reasons. We are doing basement flood relief studies using fine models with about one acre catchments and including every pipe. This gives use hundreds of catchments and pipes - adding the storage nodes and pumps would almost double the number of pipes and nodes and that is what I am avoiding with a separate model. My screen is way to crowded as it is and working with greater density of pipes and nodes is just too much trouble.

Also, our work uses only a few Runoff simulations followed by dozens or hundreds of Extran pipe model runs. The intermediate Extran is required only for the Runoff runs - don't need a new intermediate run unless inlet modifications are wanted. Thus adding the street and inlet features would add to computational time, output review and screen clutter for the many Extran runs that have no mods at all to the intermediate data. Thus efficiency of work results from a separate model.

Having a separate model also helps for reviewing it's output - it is not mired in the hundreds of pipe nodes. The street modeling does help quite a bit for spotting locations with surface flooding problems. We have used it to design extra inlet capacity in locations which need it and to store runoff in locations with ample surface storage capacity.

So there you have it Ron, for free. Nothing to sell actually.

Reinhard

Reinhard - thanks much for your detailed reply. That is a very cool use of SWMM capabilities.

Regards,

Ron K.
ronkilmartin@ATTBI.COM