HEC RAS Internal Bridge Sections 2

[Crystalct1]

Does anyone know for sure how the program creates the two internal bridge sections? I know it copies them from upstream and downstream, but where are the actual "cut lines" located with respect to the stream centerline? And how are they oriented? Notice how if the two upstream and downstream sections are different lengths, the bridge, as shown on the geometric plot looks as if as trapezoid has been formed using the 0 and last station points of the upstream and downstream sections--and that the bridge was formed inside those boundaries. Better yet!! It asks for a distance from upstream section and deck width, but is the deck actually the same width all the way across when the two upstream and downstream sections are not parallel? Where is that distance measured along? the centerline? or is it just a straight distance? Any insight?

 

[francesca]

This image pretty much sums it up:

See this page for explanation of the image.

Between your downstream reach lengths and the bridge station/width/geometry, I believe HEC-RAS can calculate it sufficiently for its purposes. Remember that HEC-RAS is a one-dimentional model and all your inputs are reduced to a head loss calculation, effectively.

Also try the Hydraulic Reference Manual as it's better than the user manual.

 

[Crystalct1]

I've seen that image many times. Actually, it does NOT answer my question. I'm refering to the actual cut lines in plan view. What are the lengths of the sections and where (again, in plan view) are they oriented with repect to the stream centerline and those two cross sections. That picture depicts a theoretical situation whose geometry is unrealistic. In reality, streams meander and sometimes all sections are not perfect straight lines whose orientations appear as a single straight up-and-down line like that picture. In plan view, your two adjacent sections may not be parallel. That is my question: How does the program decide where to sit those internals when the two defining cross sections are two different lenghths and not parallel?

 

[francesca]

Between your downstream reach lengths and the bridge station/width/geometry, I believe HEC-RAS can calculate it sufficiently for its purposes. Remember that HEC-RAS is a one-dimentional model and all your inputs are reduced to a head loss calculation, effectively.

Also contributing are the up- and downstream expantion lengths.

 

[blueoak]

I would recommend reading the Hydraulic Reference manual chapter 5 on modeling bridges. I am a little confused as to your question. The bridge model allows for skewing elements or the whole bridge. It sounds a little like you are assuming that the plan view means it is a 2D or psuedo 2D model or you are assuming an accuracy in HEC-RAS that just isn't there. For bridges, the upstream cross section is from your upsream cross section and the same for the downstream.

It is important to note that while there are items such as Width in Hec-RAS, they may not actually do anything. Width for structures is typically just so it shows up on the plan view and is not used in calculations.

As for meanders, you really have to have the survey data for those. All the squiggles on your plan view don't mean anything between cross sections.

 

[Crystalct1]

I have read all of that already. Actually, I answered my own question with trial and error and trying different schematics, etc. Hec Ras INTERPOLATES between the two cross sections bounding the bridge. If your bounding sections (the GIS cut lines) are of diferent lengths, then your BU and BD sections will be too. It connects the endpoints of the bounding sections and forms BU and BD exactly as it would if you interpolated (within a reach, not between 2 sections) between them. Only thing it does NOT do is change the internal bridge ground points accordingly. It leaves that up to you. But you need to know the length and location of actual cut line so that you can define the internal cross sections correctly. Remember, if your schematic cut line and your cross section length (from left-most station to right-most station) are two different lengths, HEC RAS will squeeze and/or pull the data points to match the cut line--resulting in inaccuracies in the model. Always look at your 3-D plot to make sure your internals are not shifted or squeezed. You should see a opening flowing smoothly from upstream of bridge to downstream. My collegues and I have seen many models where the internal sections are off. The ONLY way to be sure is to look at the 3-D plot. You'll be surprised at what this program does when your bounding sections are not parallel and equal like the (unrealistic) picture given in the literature.

Thanks for your posts everyone. Unfortunately, it seems that no one has delved into HEC RAS the way my collegues and I have.

 

[blueoak]

I think the interpolation point is well made. I find it most often with dikes and levees since I haven't really interpolated a cross section for a bridge. My upstream distance is normally about 1 ft. I always have survey data there.

Something I am curious of, are people running sensitivity analysis on the roughness values? It seems that the uncertainty of roughness is often larger than geometry changes. I am curious how this is done for frequent events? Especially if you have a lot of debris dropped in your channels.

I get to estimate bulking now, which can add up to 50% for some areas, so my models have a pretty wide range for expected flow depths, more fun but less precision.

 

[francesca]

The HEC-RAS models I've developed have been exclusively to determine a 100-year flood elevation for subdivisions. Because of this, I haven't needed to perform sensitivity analysis on my manning's values as I've needed to err on the conservative side. Particularly with stream buffer regulations which took effect in my area last year, I've been able to assume woods for the first 30' perpendicular to the channel, which has meant high manning's values and correspondingly high water surface elevations.

 

[blueoak]

I can see how for 100yr regulations they would want just 1 map of the conservative case. Thanks.