[Scott Miller]

Hello Saksit,

Yes, for unsteady 1D. Is that what you are asking about? It is in chapter 16 of the User’s Manual.

[Scott Miller]

Yes. It’s about time I get your book. Thank you, Chris. It’s on its way.

And thank you Jarvus, Con, and Paolo for your suggestions.

[Scott Miller]

Thank you, Jarvus. I haven’t harvested information from hdf files. I know that is where results are written, but it sounds like 1D geometry data may be stored there too. Is that correct?

I’m interested in the text geometry file. There is very little change that needs to be made to translate a HEC RAS cross section to SWMM transects. Both use an awkward format, which I believe was the HEC 2 format – written for limited storage and computational capacity. Here is a cross section:

Type RM Length L Ch R = 1 ,750 ,25,25,25
XS GIS Cut Line=3
1422669.09648924177400.4128280281422668.86708273177410.410196314
1422668.63767623177420.407564601
Node Last Edited Time=Aug/07/2019 09:54:45
#Sta/Elev= 16
0 767.2 .46 767.11 4.95 765.86 6.73 765.39 7.6 765.24
8.42 765.29 10 765.23 10.85 765.2 11.51 765.23 12.85 765.63
15.67 766.06 16.9 766.25 17.45 766.1 19.29 765.99 19.68 765.99
20 765.97
#Mann= 0 , 0 , 0
Bank Sta=0,20
XS Rating Curve= 0 ,0
XS HTab Horizontal Distribution= 5 , 5 , 5
Exp/Cntr=0.3,0.1

It’s largely a bunch of lists – northing/easting pairs, station/elevation pairs, some of them mixed – like bank stations then downstream lengths. It’s all formatted. But I don’t know which one of those “25”s is the downstream length of the main channel, not without tweaking a model and seeing what changes in the text file.

Is there a document that describes the formatting?

[Scott Miller]

Look at the 2D Modeling User’s Manual page 3-37.

Also, it looks like the 1D cross-sections could be reconfigured so they are perpendicular to flow.

[Scott Miller]

Flow across a 1D/2D boundary at initial conditions can throw an otherwise stable 1D model. Is the initial 1D flow high enough to pass into the 2D area? Is there an initial wse in the 2D flow area that drops into the 1D reach? It may be necessary to balance the initial conditions, by trial and error, to get inundation in the 2D flow area into a realistic distribution.

With lateral structures passing 1D/2D flow, you can allow the calculations to iterate. Unsteady Flow – Options >> Calculation Options and Tolerances >> 1D/2D Options tab. (The default is no iterations, so try this.) Choose a minimum flow tolerance that works for the scale of your model. The 1D/2D flow errors during a run give you an idea where to set the tolerance.

[Scott Miller]

Yes. In the Geometry Data window, click on the 2D area connector that has the culvert in it. Choose ‘plot stage and flow hydrograph’. Headwater and tailwater time series are plotted, like in the graphic above.

[Scott Miller]

I’ve had this happen when importing cross-sections to geometries with interpolated cross sections, or added cross sections for bridges. Check your reach lengths and make sure they make sense. Also check the distance upstream from the bridge to cross section.

[Scott Miller]

The Courant number has to be calculated before you view it. The easiest way to view it for all cells is to map them. It has to be a single time step. After they’re calculated you can hover the pointer over individual cells for the digits. Otherwise the map color scheme is useful for finding problem areas.

In Mapper, right click in the legend on “Results”. Choose “Manage Result Maps”. Beside the plan you’re interested in, click “Add New Map”. Choose “Courant” and the time step you’re interested in. Make it a stored raster. Click “Add Map”, then highlight the row of the new map in the Manage Results Maps window. Now you can click “Compute/update stored map” and hold your breath for the results.

[Scott Miller]

There must be something more to it. Is there a reason flash boards would not work?

[Scott Miller]

Hi Giovanni. Take a look at this blog post to improve your unsteady model. I found the Brunner paper to be especially helpful.

General tips to stabilize the model would be to decrease the time step and distance between cross sections. Check to see if there is a reasonable Courant number for the distance and time step in the model. Try to step through output intervals and look at what happens in profile view to locate where problems are.

[Scott Miller]

It looks like attempting a lateral structure to an off line storage area was a good approach. Unless there is a significant lake in line with the system, it makes sense to go with one structure rather than two. You tried the lateral structure with a single reach, right?

In the unsteady flow analysis window, take a look into Options>Calculation Options and Tolerances. There are inline and lateral structure flow stability factors you can turn up. They dampen the calculation of flow over the structure. That can help stability. There is also a number of iterations that can be increased. The default is a single calculation, which may be way off. Try iterating the 1D/2D flow calculation (within each unsteady iteration). There is also a minimum flow tolerance, which allows those iterations to cut off at a reasonably low value. It is different for each model.

Consider how much storage is available in the storage area, and whether the water surface elevation would shoot up rapidly in a single time step. Was the extrapolation warning because the water surface was unreasonable?

[Scott Miller]

Hi Sumedha. Please draw a simple sketch of what you’re trying to do. 1D alignments and a storage area, right? Pencil and paper, if you can scan it.

[Scott Miller]

It may be necessary to further calibrate the hydrology so the rising limb, peak, recession, baseflow, and period volumes follow the same pattern as the observed stage. Consider modeling a longer time period, so the suitability of the hydrology is more visible.

Beyond that it looks like storage may be over-utilized – too much attenuation. What Manning n values have you tried? What flow range, channel width, and bed material are we looking at?

[Scott Miller]

That sounds like a good approach.

You will be able to import the final mesh from an existing geometry to new geometries. That keeps the mesh consistent, in case nodes have to be moved manually.

[Scott Miller]

In the plan window, pull down the options menu to ‘view computation log file’, but if there is not any .bco## file, this may not work.

The flow trials 3 and 4 look like they agree well. What you might do from there is to test sensitivity of grid cell size, while allowing the Courant condition to adjust the time step.

[Author]

Posts