HEC-RAS Cross Section Requirements 3

[TomHydro]

Great forum, folks! Brand new here, and a HEC newbie, also, so please bear with me ...

I'm beginning a HEC-RAS analysis of a creek for a bridge design. What are the required limits (upstream and downstream) for the creek cross sections?

(I used the search function and noted that this question had been answered, but I didn't see the actual answer anywhere. So, a link to an old thread would be great, also.)

Thanks a lot.

 

[gbam]

I take a stab. The upstream & downstream limits are set by the influence resulting from your proposed structure. For instance, if your uppermost XSEC is still showing an increase it may warrant additional XSECs or a larger structure. The idea is to establish a distance outside the influence. There are ways to evaluate the impact from a hydraulic structure. However, with the software today it does not take alot of time to cut and add more XSECs. Hope that helps.

 

[TomHydro]

Thanks a lot, gbam. Thing is, I have to send my surveyors out to topo the creek, so I need to know the limits before I do any analysis. I've seen a few things referenced, such as, "at least to a point above the 500-year flood elevation", etc., and, "500 feet upstream and downstream", but these seem like project-specific criteria, and not general requirements.

Is there a concise, quantifiable requirement for the minimum limits of the cross sections?

 

[gbam]

Tom have you looked or requested mapping from your local agency for the area? Is there any Lidar data available to get you started and then send your surveyors out for additional or fine tuned XSECs. If you have GIS capabilities in your office maybe you can get some of the information cheap if not free.

 

[gbam]

I forgot. Is this a FEMA regulated watercourse? If it is you can typically obtain the effctive model from FEMA.

 

[SMIAH]

The downstream boundary is usually selected at a location that is independent of flow conditions below the boundary.

You have to look if there is any control downstream of the bridge. There can be structures such as dams or waterfalls or even other bridges. Empirical water-surface elevation-discharge relations such as that for weir flow can be utilized for simulating rapidly-varying flow.

So if the creek is on subcritical flow, I suggest you take more sections downstream of the bridge than upstream.

I'm not talking about storm surges, tides or large tributary inflows here...

 

[cvg]

if it's subcritical than you may need plenty of sections upstream also because your bridge will influence the water surface elev for some distance upstream. Distance up and downstream depend greatly on the slope of the river. How many feet per mile does the river slope? 1, 10 100?

 

[francesca]

I just completed modeling over a dozen proposed bridge sites for a new interstate. We started with the guideline of "4-cross-section models, 500-ft upstream and downstream." It was hideously inadequate in almost all cases.

Your four bridge cross-sections need to be situated at (#1) the end of the expansion reach, (#2) base of downstream abutment, (#3) base of upstream abutment, and (#4) beginning of the contraction reach. XSs #s 1 and 4 will not be the same distance from the intermediate XSs; in fact, 4 will be half as far from 3 as 1 is from 2. (Contraction reach has a 1:1 slope, expansion reach has a 2:1 slope.) You need at least one cross-section upstream of #4 to show that the point of maximum backwater has been reached.

You need additional XSs (1-2) downstream to let the calculation settle from your downstream boundary condition and similarly upstream if you have mixed/supercritical flow.

Finally, what are your design criteria? On our job, we had two options: 0.14 ft of backwater at the ROW line or up to 1 ft of backwater at the ROW line with the purchase of flood easement upstream until the point that backwater had dissipated to 0.14 ft. (Those models required XSs a long way upstream.) If you have to dissipate your backwater by your ROW line, you'll need to model up to there and only beyond if the ROW line is inside the contraction reach/the backwater hasn't started to reduce. Obviously you need a cross-section at your ROW line, and if it's within the contraction reach, then you need to note that and make sure you're using the proper XS #4 for your approach section for your scour calculations.

For your survey, you need a good handle on top of bank, bank width, and an average cross-section. You also need a good handle on the average (linear regression) bottom slope. If you're using normal depth as your downstream boundary condition, that should really take precedence. Have the surveyor take 100s of photographs, especially if the site isn't on Google Street View.) A good aerial/GIS topo is essential because the majority of your flow will be on the overbanks. Outside of our topo (we had a 1,000 ft wide corridor and then 25-yr old Quad maps) we extrapolated cross-sections. (Same channel shape, vertical offset for average slope, widened/contracted the floodplain as measured off the quad map.)

Make sure your cross-sections #s 2 & 3 are the same under the 100-yr elevation as what's happening a the bridge site. The structural engineer will cut a cross-section under the bridge centerline and will need to design for the flow area your modes specifies. If the XS at the base of the abutments isn't representative of under the bridge, use the internal bridge XS option.

 

[SMIAH]

your bridge will influence the water surface elev for some distance upstream

Question: Can you perform a simulation without the bridge and then add it to see what's the difference of WSE at the bridge (or a couple of feet upstream). If there is, for example, 1' between the two conditions, then you can work a scenario with a different bridge depending (widened, for example). Is such a way to do acceptable in a design study? When you don't necessarily want to access the values of the increase in WSE upstream of the bridge.

You need additional XSs (1-2) downstream to let the calculation settle from your downstream boundary condition and similarly upstream if you have mixed/supercritical flow.

That's want I wanted to say.

 

[cvg]

SMIAH
if you have plenty of right of way, no environmental impacts and lots of money you can span the entire floodplain with a very long bridge and create no rise at all. However, I have seen very few bridges that actually do that. Most bridges are designed with very little skew and as short of spans as possible and create a reasonable rise in the WSE in order to save money on the bridge. so the answer to your question is yes, but probably not feasible.

 

[SMIAH]

The bridge might cause a constriction of flow (choke) as long as you keep the required freeboard under the soffit. Well, waterway opening width no greater than necessary to meet basic hydraulics requirement (including permissible velocity, scour potential, etc.). There might already be flooding with the "without bridge" condition.

In a perfect world, I'd like to have 5 miles of survey upstream of a bridge to see what's the increase on the WSE introduced by the bridge. But that is not economical and not necessary in most of my studies (except studies to access environmental impacts of a bridge, for example).

 

[gbam]

SMIAH - Regardless of existing flooding it is necessary to not worsen the condition with new structure unless there are other circumstances. There may be purchase options and such to counter the increased water surface.

I typically find free or inexpensive sources of mapping (LIDAR) to perform preliminary analysis and then update with current topo-survey. This way you have an approximate need for survey. Of course, if the free or cheap information is not available then I use 500-1000 feet Upstream and the downstream control as my initial limits for survey. I also include a caveat for additional survey, if needed (T&M).

 

[cvg]

the main problem with aerial, lidar etc is that in general, you can't get the elevations of bottom of the river channel which you need for your modeling. You have water and trees which get in the way. Without that data, it may be totally useless for hydraulic modeling. Field survey is usually the best you can do. USGS quads are of limited value, but may give you an idea of what the terrain looks like.

 

[SMIAH]

And I still think that if you're running on subcritical state and approximating the downstream boundary with the bottom slope, you will need more cross sections downstream than upstream of the bridge.