Trying to reproduce another engineer's results 1

[beej67]

See the attached file. First page is their hydrograph, second page is my hydrograph. I'm looking for a combination of HydroCAD settings than can best emulate the other engineer's (Hydraflow Hydrographs) results, and can't seem to get it any closer than 5%, which seems like too big a bust to me. I'd be happy with 1% or 2%, but 5% makes me feel like I'm missing something. If they made some weird assumptions to get their number, I need to figure out what those weird assumptions were. I find it especially fishy that our peak discharges are off by 5% even though our total runoff volumes are within 0.008% of each other.

So far I've tried varying my calculation interval, which didn't help much if at all, and using the Georgia 484 unit hydro, which helped a little but not enough. I cant find a way to get my model to match his number without monkeying with my Ia/S ratio, (not shown on his printout) but I don't think he would have done that for his report.

(note, I realize our Tcs are off by 0.03 mins, I checked that and it doesn't fundamentally change the results)



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

http://www.campbellcivil.com

 

[psmart]

The matching volumes suggest that the CN and Ia/S ratio are probably the same. Since the volumes match, but the peaks are different, there is probably a difference in one of the time factors, such as the Tc, Unit Hydrograph, or rainfall distribution.

The Tc seems to match, so I would suspect the use of a (slightly) different Unit Hydrograph or rainfall distribution. I would get the actual rainfall table that was used in the Hydraflow calculation and compare it to the HydroCAD data. There have been slight changes to the Type II storm over the years, and there is always the possibility that someone has altered the standard table.

For short Tc values, the peak runoff is especially sensitive to the time step that is used to define the rainfall distribution. To address this, HydroCAD uses a polynomial form of the Type II distribution which provides good accuracy (slightly higher peaks) for small Tc values. But this is the opposite of the difference you're seeing. For comparison, I tried the tabular form of the Type II distribution with HydroCAD, and this gives an (even) lower peak, as expected, so it doesn't shed any light on this difference.


Peter Smart
HydroCAD Software

http://www.hydrocad.net

 

[psmart]

HydroCAD runoff calculations have been extensively cross-checked against TR-20, and generally match the TR-20 runoff volumes and peaks to within 1%. (Of course, this assumes that all the input parameters are the same)

Although I've already summarized the potential issues in this particular case, the following web page provides a more general guide to comparing runoff results:

http://www.hydrocad.net/scsresults.htm

Peter Smart
HydroCAD Software

http://www.hydrocad.net