[Chris G.]

Yes. Use the Depth Blocked entry. You can also have a different n value for the bed sediment in the culvert.

[Chris G.]

Unfortunately those runtime messages usually don’t help user out too much. However, given that it mentions Coordinates adn “Division by error”, I’d be inclined to think maybe one of your cross section coordinates is mis-typed, or just incorrect. Wild guess though. I’d send your data set and a very brief explanation to HEC. Runtime messages are always “bugs” and they like to fix those.

[Chris G.]

Why a flow hydrograph for the downstream boundary? I’ve had very little success in the past using flow hydrographs for the downstream boundary, because if you don’t exactly match what RAS is calculating for flow, then you get errors at the d/s boundary that can lead to instabilities. And, as you noticed, sometimes the How about normal depth for your downstream boundary? This should work very nicely for a canal, as long as your downstream boundary is not in the middle of a backwater curve from a downstream control (if so, you could still use normal depth, just move the downstream boundary out of the backwater influence).

[Chris G.]

Yes! I forgot about the Minor Loss Table. If you go with lids, definately use that.

[Chris G.]

Great question! I would say if you expect flows to rise in the main channel, over top the structure and flow up the tributary, then model it as a lateral structure off the main channel. Otherwise, use an inline structure. Keep in mind, that if you use an inline structure, you’ll have to have at least 2 cross sections downstream of the inline structure before the junction.

Even better, try both and let us all know which works better!

Good luck-
Chris
@RASModel

[Chris G.]

Hi Phillip-

It is indeed possible, and a great application fo rHEC-RAS. First this needs to be an unsteady flow model (I think it is, but you don’t specifically state this so I want to make sure).

To model the attenuation through a reservoir you can use a storage area to simulate the reservoir (easier, less accurate) or you can model it dynamically with cross sections (a little harder, more accurate). To see attenuation, you can view the stage and flow hydrographs and look at the flow entering the reservoir compared to the flow leaving the reservoir over the spillway. While in the stage and flow hydrograph viewer, make sure to go to the “Type” menu item to switch between different node types. If you do model the reservoir dynamically, you can tell RAS to produce stage and flow hydrographs for all cross sections by going to Options…Stage and Flow Hydrograph Locations, in the Unsteady Flow Analysis window. By default, RAS only produces S&F hydrographs for nodes the bound internal and external boundaries.

Good Luck

[Chris G.]

It’s hard to say what the problem might be over this forum. First make sure you model is stable. When I find a somewhat complex situation with unexplainable results, I’ll usually rebuild my model in stages, starting simple, and adding in one compenent at a time. Checking results in between each change. This usually helps me to pinpoint what the problem is.

In your case, try the model without pumps. Try stabilizing the pool by opening the gate until inflow to the header pond equals outflow through the gates. Then try adding in a pump (one at a time) and adjust the gate settings accordingly.

Hope this helps
Chris G.
@RASModel

[Chris G.]

Not with a csv file. Unfortunately that only brings in stations and elevations.

[Chris G.]

some of that is bound to happen around structures, especially when one profile may be low flow, and the next pressure flow. Or pressure flow to pressure and weir flow. Switching between equations (which is what RAS does when switching between these regimes at bridges) can cause some overlapping of profiles just upstream of the structure.

This can also happen when ineffective flow areas are just “overtopped” and turn off. Again, this should be a local phenomenon.

Sometimes, it’s numerical instability. I suggest plot critical depth and see if you are defaulting to critical depth as some of the “jumpy” areas. You may need more cross sections, or better cross sections in these areas.

Good Luck-
Chris
@RASModel

[Chris G.]

This can happen if cross sections cross-cross each other. Also, it happens when you draw your flow lines the wrong direction. However, since only some of the reach lengths are negative, I’d assume it’s a criss-crossing issue.

[Chris G.]

Are you trying to capture the head loss from the bend? If the culvert is long enough, the loss from the 45 degree bend is probably fairly minor compared to the friction loss and the entrance and exit losses. Plus, RAS does not explicitely calculate minor losses in pipes (other than the entrance and exit losses). You would have to put in some forced energy loss. Unless there is a particular reason you have so show the 45 degree bend, I’d ignore the bend altogether. If you really want to include some loss, bump up the mannings n value slightly in the culvert to simulate the minor loss from the 45 degree bend.

If you have to show the 45 degree bend, lidded cross sections would work, but again RAS will not caluculate losses due to this 45 degree bend-it will purely be a visual thing.

Keep in mind that RAS is not designed to be a pipe network model. The culvert routines were designed for typical highway culverts. That being said, if you understand the limitations, you can simulate long tunnels and pipes to some degree. Lids do provide flexibility to do that.

[Chris G.]

Hi Caroline-

Excuse me if you already understand this, this next paragraph is for others who may not understand how RAS deals with bends.

Please keep in mind that RAS is a 1-D model and it does not explicitely calculate any effects from a bend (i.e. added turbulence, super elevation, etc.). That being said, there is added friction loss on the outside of a bend, because RAS uses the longer reach length (i.e. for a left-hand turn, the right overbank will have a longer reach length). For the same reason there is less friction loss on the inside of the bend, because it will have a shorter reach length.

If the flow remains between the bank stations through the bend, RAS only uses the main channel reach length to calculate friction loss, so the added friction length on the outside of the bend and the subtracted friction length on the inside of the bend is not accounted for. So the answer to your question is: No. But, if you wanted to put bank stations inside the channel to somehow simulate the variation in friction length loss due to a bend, you can.

[Chris G.]

Joan-

The problem you bring up always frustrates me. Make a cross section’s flow area wider and you get a slight rise. It’s counter intuitive and many people don’t get it-unfortunately that includes reviewers. If you indeed increased the flow area, you should see a slight decrease in stage upstream of your cut. Can you plead your case to the reviewer given that?

You could insert some ineffective flow areas in the new cut area, to reduce the flow area back to pre-project conditions, thereby bringing the stage at that cross section back down. That may also increase the stage upstream slightly-just make sure it is still at or less than pre-project. You’ll need to justify the ineffective flow areas, but if you cut 40% over pre-project, I’m guessing most of that will be ineffective, unless you smooth out that expansion/contraction over a number of cross sections.

Good Luck-
@RASModel

[Chris G.]

Hi Nils-

Ideally, you’d get a few more from GIS (enough so that interpolation doesn’t give you this problem). But if you don’t want to do that, when interpolating, try using the feature “Generate for display as perpendicular segments to reach invert” in the “Cut Line GIS Coordinates” pane of the Interpolation window. That will keep the cross sections’ inverts on the stream centerline. The geopositioning might not be perfect, but neither is the other method and they are, after all, interpolated cross sections.

Second Question: It could cause instabilities, but I’d also be worried about inaccuracies (even if you get it stable). If you have to go that route though, better not to use wide flat-bottomed cross sections-they tend to cause stability problems. Parabolic would work, or even easier, just give the rectangle an invert point in the middle that is slightly lower than the end points. You are correct that the higher the stages, the less of an issue this becomes.
Good Luck-
Chris
@RASModel

[Chris G.]

There is a new presentation up on RASModel.com that discusses in detail the use of Rules for inline structures in HEC-RAS.

http://hecrasmodel.blogspot.com/2013/09/advanced-rules-for-unsteady-flow.html

Chris
@RASModel

[Author]

Posts