HEC-RAS: semi-perched bridge, road parallel to stream

[nadler64]

We're modeling a bridge that has some funky geometry...original model was set up by folks not terribly proficient in HEC-RAS, but we're at a loss as to how to model this correctly. I've attached a sketch of the old hydraulic site plan.

Project is in a FEMA detailed study area, designated floodway. The FEMA data, as usual, is old and of no help whatsoever.

Looking downstream: on the left, the overbank area goes up from the stream and there a narrow floodplain. On the right, the floodplain is low and very broad.

Roadway on the right overbank/floodplain (moving station ahead) parallels the stream. Just before the bridge, roadway turns sharply left and sharply up (10%) to cross the bridge, then proceeds uphill (9%) from the bridge. Low chord is about 6' above the roadway elevation on the floodplain.

So basically we have a perched bridge on the right side. The deck/roadway section, if we were to extend it the length of the project, would cross our upstream cross sections, which is a no-no. If we truncate the the bounding cross sections, we get an artificially small flow area through the bridge which bumps up our upstream WSEs and the V's through the structure skyrocket. We know we don't have high V's through the structure; the stream is stable, there is very little scour.

Has anyone dealt with a situation like this before? How did you model it?

 

[jhathorn]

Don't try to model the roadway parallel to the stream as a bridge since this portion of the roadway will be at grade. If you want you can change the manning's n value for your cross sections crossing this portion of the roadway. For the bridge, just tie your bridge deck to natural ground at the cross section just upstream and downstream of the bridge. Also and this is just a suggestion, you don't have to zig-zag your cross sections that much. I understand you are trying keep the cross sections perpendicular to the contour intervals, but you can generalize the cross sections a little. Hope this helps.

 

[nadler64]

You just confirmed that we were on the right track - that's pretty much what we did. Thanks!

 

[Crystalct1]

Good point about not zigzagging your cross-sections so much. It distorts the sectional area in the calculations. My advice is a straight shot bank-to-bank and one bend at the banks and one straight segment across the floodplain in most cases. Of course, if it obviously deserves one more bend within the floodplain, that would be okay. Just never bend the section within the stream banks - and it looks like you did not.