hec-ras inline weir

[troywoodyard]

I am modeling a stream that goes under a bridge at a skew and then takes a hard left turn and runs parallel to the road. The bridge is higher than the roadway, so the water will back up behind the bridge opening and then spill over the road, but not overtop the bridge. When it does this it will not enter the next downstream section, but will actually go three to four sections downstream and bypass the channel for several hundred feet. Ideally i could set up the road as an inline weir that when overtopped empties into the left overbank of the stream at a downstream section several hundred feet downstream, but I don't think this is possible. Any ideas how to model this behavior?

Just and EIT
Troy

 

[gbam]

Troy;
If I understand correctly the roadway parallels the watercourse and is not across the channel. This would be a split flow scenario rather than inline. you can model this split using the geometry of the roadway and an appropriate weir coefficient. once split from the main channel you can either return the flow to a downstream xsec or you can model a secondary reach down to the xsec. This is not an uncommon condition.

 

[Drew08]

I understand that the channel parallels and borders the road on one side, does a hair pin (180 deg) turn through the bridge, where then the channel parallels and borders the road on the opposite side.

How about modeling the road as a lateral weir, separate from its bridge (modeled separately as a bridge or inline weir), where the lateral weir connects the cross-sections upstream of the bridge to those downstream of the bridge.

 

[troywoodyard]

I am thinking about these ideas... The problem I have with your idea (gbam), and I am not very experienced in HEC RAS so I'm not sure this matters, is that there is no discernable second channel for the water to take. I am familiar with split flow optimization, but it isn't like a second channel goes over the road. I tried drawing a stream channel that ran from upstream, parallel to the road (even though in real life there isn't one) and placing a lateral weir to spill over into the DS sections, but I was losing 1/3 of the original flow at the end of this imaginary channel. I thought about using an inline weir as a dam at the end of this secondary channel, forcing the water to queue in the floodplain and then spill over the road but didn't know if this was reasonable? Any other ideas?

Drew, you understand the geometry. My problem, unless you know a way around this, is that the reach cross sections are roughly parallel to the road upstream, and perpindicular DS. There isn't really a lateral wier upstream because the angle is more inline. Are you suggesting that I make a lateral weir Downstream, and input all the flow that overtops at the overbank of the upstream sections? Then essentially allow the program to compute this as a negative flow? Would that work? Now that you mention it I think I have heard that this works? Have I understood you correctly?

Thanks
Troy

 

[gbam]

Troy,
A sketch would help. Have you considered that maybe the flow that weirs over the road does not completely return to the main channel? If it does not return then just split the flow without returning it. The lateral weir concept as Drew mentioned is the same concept I stated with the split, I just use HEC2terms. If your flowline makes a 90^ bend at the road under the bridge, I would model at a split or lateral weir; but use momentum rather than watersurface.

 

[troywoodyard]

So originally I did the model with water flowing the wrong way (rookie mistake... feel sheepish ) This is my geometry, but the arrows should go the other way. Water is flowing south to north, not as indicated on the picture. The left branch is a dummy branch I used to pick up the weir flow over the road when water was flowing the other direction. Let me know if you have other questions/suggestions...

If I were to use the same geometry and make a lat weir DS, emptying upstream, would the negative flow be accounted for correctly by HEC RAS, since in real life water would spill over weir from Upstream to downstream? Does my question even make sense?



Just an EIT,
Troy

 

[troywoodyard]

Also, low spot in the road is on the left side of the bridge looking downstream (ie left side of page)

Just an EIT,
Troy

 

[gbam]

Do you have a PDF of a map? I can't tell where things are from your geo file; maybe post an aerial from google earth or something.

 

[troywoodyard]

Here is an aerial shot. Does this help? Also Am I correct that a lateral weir dowwnstream could have a negative flow, thus allowing backe dup water from upstream to move downstream (TW to HW of weir)? Or would HEC RAS lose this flow since it computes from Downstream to Upstream?

Just an EIT,
Troy

 

[Drew08]

Your question doesn't really make sense to me. You connect a lateral were between two reaches, what you call upstream and downstream of the weir is arbitrary.

With out a scale or purpose in mind, it’s hard to judge. But, do you really need that dense of cross-sections? I'd suggest keeping the section immediately downstream of the bridge, but delete the next 4 or 5, so that the second cross-section downstream of the bridge is at the low point of the road.

Then, I'd consider two options:

1.) Extend the downstream cross-sections left across the road, so that the road is a levee within the cross-section and not an adjacent parallel feature. Include the road as XS Levee element in the downstream cross-sections. Essentially keep the cross-sections as shown in your figure except connect the left branch and north branch cross-sections, and eliminate the left branch. It looks like the road (i.e., levee) actually extends upstream through your upstream cross-sections as well, where you have ineffective flow designated? Or,


2.) Set the left over-bank downstream lengths between the sections downstream of the bridge to something very low. Channel and right overbank downstream lengths would remain as shown. This would allow you to model the bridge as a skewed weir with a low overtopping point on the left side, where overtopping flow, which is overbank flow, would not travel the several hundred feet of channel, but rather the insignificant overbank length.

 

[gbam]

Assuming that the flow is from south to north bottom of page to top of page or Right to left in your photo. Two scenarios come to mind.

1. Flow continues through the bridge without overtopping road/bridge and backwater occurs left and right on the upstreamside of the bridge.

2. Flow can escape the left along the left side of the bridge and not re-enter the original watercolurse with out over topping the roadway.

I agree with Drew. re-examine your xsecs and apply flow diversions as needed. It may be an iterative process. If I think there will be a flow breakout, I model one. HECRAS will compute losses and gains if properly coded.

 

[troywoodyard]

Thanks for the help... What I am looking at is a lateral structure downstream with a TW at the first upstream bridge section. This seems to be working (whether or not it is correct I do not know.) I am starting with 1400 cfs, 925 goes under the bridge, and the difference is being added back into the model by the end of the reach (by way of overtopping the road.) The assumption is that no water is lost, which is more conservative for my project. I am working on a bridge replacement and I am just trying to show that the bridge is not overtopped so that I can add a barrier wall on the bridge. I want to leave the opening the same so I just need to prove that adding a barrier wall will not cause additional flooding.

Thanks for all of the help. Let me know if something I just said is wrong...



Just an EIT,
Troy

 

[troywoodyard]

Here is a pic of new cross sections... I am assuming that the upstream channel will T into the bridge, creating backwater, which will just queue until it can go over road or under bridge. Is this what you guys were talking about. I was confused because drew seemed to indicate removing the secondary channel and gbam seemed to indicate that he agreed, but that I should split the flow...

I have a lateral structure DS that feeds back into the reach just US of the bridge.

Thoughts?

Just an EIT,
Troy