With the upcoming release of distributed rain-on-grid, there is of course the welcomed potential for surface water flood modelling.
I have a question related to urban areas. I have seen comments about the model not being suited to storm water networks. Is this true. Will the release of rain-on-grid mean that modelling an urban area is still either impossible or impractical? By urban area I'm thinking about something around 10km2 and 5km of pipeline with manholes and culverts.
RAS doesn't have built in capabilities for manholes and drop inlets. It is not impossible to do an urban storm network, but it is awkward. You end up having to model the underground network as cross sections with lids. You can get good results as long as you don't have stability issues. Version 5.1 is supposed to have a finite volume solution for 1D, the same approach that 2D uses. Hopefully this helps a lot with the stability issues you can get modeling pipes as cross sections with lids.
It would be nice if RAS added some bells and whistles for urban drainage such as drop inlets. Making it easier to block flow from going through buildings in the 2D mesh would also be nice.
If I need to get some subsurface flow accounting I'll typically use a storage area connector with culverts for the pipes. Requires be to "burn" in the manhole into the terrain to meet the minimum cell elevation being equal to or lower than the culvert flowline. If you play with the n value and entrance/exit loss coefficients you can adjust some of the losses to match a typical storm sewer calc. Gets cumbersome though if you need to model a lot of storm sewer.
I wouldnt even consider all the tricks. The reality is most SS networks are designed for ARI's that lower than that used for inundation mapping. Ive done a complete urban model for testing (fun) at home in XPSWMM utilizing rain on grid with a interconnected underground SS network that would take inflow from the surface. However the results were terrible to say the least. I would never recommend analyzing a urban area this way. MIKE seemed to do it better than SWMM. I spoke with XPSWMM (innovyze) and one of their engineers/tech people, did not recommend modeling as such. He recommended utilizing traditional methodology (rational method) convert that flow to a hydrograph via whatever method you are comfortable with and enter it directly into the manhole. Then rain at the same time, this of course requires certain losses to be exaggerated because you are double dipping. But if you try and only utilize rain the amount that gets into your inlet will be DRASTICALLY less than that which what is calculated via traditional methodology.
The question becomes which method is correct. In a sense if you are raining you are playing god and it should represent the most realistic floodplain mapping; but then ask yourself am i underestimating flow? Is it worth the work to match/calibrate each and every node vs traditional methodology via precipitation? Right now i dont think so.
HCFCD mappped most of houston for the Harvey storm and matched realistic conditions very very closely without considering a SS system at all. All done in RAS. Some people who have never modeled really had issues with that; but if any of them have that experience they would understand the difference is negligible.