Abnormal large spread on curb inlet in sag

Question

While doing some testing with SUDA on a previous project that was done in GEOPAK Drainage, I noticed that I was getting some abnormal large spreads for curb inlets in a sag. After I imported the .gdf file into a new file for SUDA testing, the spread was 10.12' while the spread in GEOPAK Drainage was 3.1138'. They should be somewhat close. Why is this??

One thing I did change in SUDA was in the inlet catalog, I changed the default curb opening length from 5 to 60 (60"=5'). Then I ran the computation again and it was somewhat close on the spread. 
 
 Maybe the program is figuring the curb opening length incorrectly?? Please look into this. Thanks.

Jesse Dringoli · Accepted Answer

Hello Chad, 
 Thanks for your patience on this. I have discussed this at length with our developers. In short, the results will make more sense in an upgraded version, and GEOPAK drainage may be making a different assumption with regard to the spread and depth calculations in certain conditions. 
 Update: this is now documented in this wiki article: Unexpected spread and depth results for a curb inlet in sag with local depression 
 First, to directly answer your two questions: 
 Chad Green said: #1 - Why isn't the gutter depth changing on the different local depression depth I gave the program, if it is measuring from the low point on the depression for the inlet? 
 The version you're using does not measure gutter depth from the bottom of the local depression. This was an improvement we made for the CONNECT Edition of the Haestad products, and that code should be part of the latest OpenRoads Designer version (which replaces Power GEOPAK). 
 Chad Green said: #2 - Shouldn't the "0" local depression inlet have the smaller gutter depth than the inlet that has a local depression value? 
 See above - in the latest version of the Haestad products, it does have a smaller gutter depth when compared to the cases with the gutter depression (and a slightly higher spread, due to the impact of not having the additional capacity from the local depression) 
 Here are the results of the three cases you mentioned, in the latest version of StormCAD: 
 Case 1: 4" gutter depression 
 Depth: 0.6525 ft Spread: 10.12 ft Explanation: with the assumption that the depth is measured from the bottom of the local depression, the depression depth is added to the depth above the top of the depression (the normal datum point for the weir calculations - see red line in post at the bottom of this thread ) and the spread is computed accordingly. 
 Case 2: zero/no gutter depression 
 Depth: 0.377 ft Spread: 13.0 ft Explanation: Without a local depression, there is no local depression depth to add, so the overall gutter depth is lower when compared to the cases with the local depression. Equation 4-30 is used in place of 4-28 though, when there is no local depression, and hence the spread is a bit larger (since the inlet is not quite as efficient without the local depression). 
 Case 3: 6" gutter depression 
 Depth: 0.819 ft Spread: 10.12 ft Explanation: increasing the depression lowers the bottom/low point but the datum point above which the weir head is measured is still the top of the local depression (see red line in post at the bottom of this thread ) Hence, the spread stays the same (WSE stays the same) but the depth increases (bottom point is lower). 
 Here is the general flow of how the program (latest version) is computing depth and spread for these inlets, using the 4" depression as an example: 
 1) Two versions of the weir equation are compared and the one with the largest depth is selected - Checks HEC-22 equation 4-28, gets depth of 0.202 ft - Checks HEC-22 equation 4-30 (applicable to curb inlets without depression) and subtracts the local depression, resulting in a depth of -0.0728 - The first is selected; 0.202 ft 
 2) Gutter depression height is added. The weir equations 4-28 and 4-30 solve for a HEC-22 "d", which does yet include gutter depression. - Gutter depression for a conventional gutter = W * (effective gutter slope - road slope) = 0.1168 ft - 0.202 + gutter depression of 0.1168 = 0.319 ft 
 3) Orifice flow is considered, with local depression subtracted - This yields a negative number in this case 
 4) The larger of the weir (step 2) and orifice (step 3) depths is selected - 0.319 ft is selected 
 5) Local depression height is added - 0.319 + 0.333 = 0.6525 ft (the depth reported in the latest version) 
 6) Compute spread - See help topic "Composite Gutter Section" - With a composite gutter section with multiple different slopes, the spread is computed in steps, and depending on whether it extends past the local and/or gutter depression, it determines which slope(s) need to be considered to compute the final spread. - In this case it extends past the depression and gutter width of 2.5 ft, and a portion of it extends into the road slope. - Based on this, the overall spread is 10.12 ft.

Jesse Dringoli · Answer

While we wait for a copy of the model file to confirm, here are a few possible factors involved: 
 1) Road and gutter cross slopes - can you confirm that they are the same between the two applications? 
 2) Differences in interpretation of HEC-22 between GEOPAK and the Haestad solvers. There are many gray areas in the HEC-22 inlet calculation standards, which can often account for differences in results. You can look at the StormCAD Help documentation for details on the equations used for curb inlets. Here are some examples which may or may not be involved in your particular case: 
 
 Gutter depth measurement 
 Unexpected results with long curb opening length 
 In-depth forum discussion on calculations for curb inlets in sag 
 
 3) Version-specific problems. Based on the version number provided, I believe you are using a "SUDA" version that uses the V8i SELECTseries 5 version of the Haestad numerical solver. There is always the chance that some fixes and enhancements have been made in more recent versions. Do you have StormCAD to test this, or only SUDA? We can certainly test in the latest version (or in OpenRoads) once we have a copy of the model (.DGN file, for SUDA).

Yashodhan Joshi · Answer

Hello Chad, 
 Thank you for sharing your model DGN file. I was able to open it in Power GeoPak V8i (08.11.09.878). To test I ran the scenario "NW Base Analysis 2 year storm" wherein I got the spread for inlet DI 71 as 10.12 ft. 
 Since GeoPak uses storm-sewer models up to SS4 only, I had to convert your .stsw file linked to the DGN to the latest version of CivilStorm (10.01.00.72). 
 I then attached this .stsw to a blank DGN in Open Roads Designer. I analyzed the same scenario in SUDA but the spread value was same (10.12 ft). I am using Open Roads Designer (10.03.00.43). 
 So it seems that the HEC-22 equations are working similarly for both GeoPak and Open Roads Designer. 
 However I require the .gdf file for this project to open in GeoPak Drainage which you have not shared. To understand the difference in spread calculations please share the .gdf file. 
 Also please share the workflow you are adopting to open the file. 
 Additionally also share details about the scenario you tested, and the .stsw model.

Jesse Dringoli · Answer

Hello Chad, 
 We're having some trouble opening the GDF file, but I'm not sure if we need it. See my previous post - it is very likely that the Haestad solver (used in SUDA/OpenRoads) has a different interpretation of HEC-22 in this specific case, yielding different results. 
 First, there appears to be a difference in input in the screenshot of the GEOPAK file, the SUDA model, and the model you sent. Here are some observations: 
 - In the GEOPAK screenshot, the Flow is 1.81 (CFS?) but in the SUDA screenshot and the model, the flow is 1.99 CFS 
 - In the GEOPAK screenshot, there appears to be a local depression of 3 inches with a depression width of 3 ft, in the SUDA screenshot it is 4 inches and 2 ft, and in the model file, there is no local depression. 
 - The "computed head" in the GEOPAK screenshot appears to be from HEC-22 equation 4-28 (I matched the result with a hand calculation) 
 - The depth (and thus spread) in the Haestad solver (and in SUDA) is a bit more complex, as explained in the forum discussion previously linked to.

Abnormal large spread on curb inlet in sag

Top Replies