[Scott Miller]

Three percent is a moderate slope on a natural stream, however, you could say a channel is steep based on dominantly supercritical flow. There needs to be an upstream boundary condition, as well as downstream, in order to model a mixed flow regime. Normal depth is usually a good choice. Just make sure your boundary conditions are far enough away from the reach you’re studying. (Calculate backwater distance.) An inlet controlled culvert will not flow full. If you have water surface elevation for a known flow rate, you can calibrate the inlet coefficient, but don’t expect a full culvert unless something on the downstream end make it outlet controlled.

[Scott Miller]

I mean HEC RAS was not developed to model such a long pipe. A separate model, like EPA SWMM, may need to be used to model pipe hydraulics. I am not fully familiar with all new HEC RAS features, but to my knowledge HEC RAS still uses FHWA calculations for culverts, inlet controlled or outlet controlled. I would want to look further into the HEC RAS Hydraulic Reference Manual, or FHWA documentation, to make sure what I was doing with RAS made sense.

If the box culvert does not flow under pressure, i.e. there is always airspace at the top, I would model it as an open channel.

• This reply was modified 3 years ago by Scott Miller.

[Scott Miller]

Make sure the entire cross-section of the culvert (inlet) is at least as high as the terrain. If it is not, you can modify either the terrain or the culvert characteristics. The wording of the error message is different than I’ve see before, so this could be a different issue. An image of the geometry could be helpful.

I would be concerned with modeling such a long pipe as a culvert. It is essentially a network.

[Scott Miller]

I would guess you have set up a 2D model with a breach in an embankment. Correct? How well does flow pass from upstream to downstream boundary conditions without the breach? Check elevations of the breach with regard to the terrain. If the breach goes lower than the terrain, the problem may be there.

[Scott Miller]

This looks like a configuration I will encounter also. For clarification, we are looking at modeling this in 1D. Correct?

[Scott Miller]

Without taking a look at the model, my first guess would be that the 1D reach is getting dewatered where it drains into the downstream 2D domain. Since the discharge from 1D to 2D would be at the high end of the 2D domain, the last cross-sections of the 1D reach may be draining very quickly into an empty channel, or lower water surface elevation. Set the output interval match the computational timestep (or to its minimum), and see if it runs even briefly. The downstream end of the 1D reach can dewater in the first time step, causing instability. Adjust the timestep, and consider adding cross-sections.

[Scott Miller]

It would be helpful to see those screen shots. Are you modeling 1D/2D unsteady, or 1D with a storage area? Do the lateral structures connect to the tailwater only through the culverts, or do you expect them to be overtopped? Are you defining the lateral structure entirely in the geometry editor, or in Mapper and the geometry editor. What progress have you made since your post, and are you seeing convergence issues or volume error?

[Scott Miller]

Rocko. RAS 2D is unsteady, so you’re doing quasi-steady right by using a timeseries. The courant number tells you how well flow is calculated in each cell. Particular cells with high courant numbers control the time step, if you’re using courant condition to control the timestep. Whether it is accuracy for a steady time step, or time step control, you’ll want to modify the size of cells with high courant numbers. Courant numbers can be mapped for a timestep, just like other flow properties, in Mapper.

[Scott Miller]

We need to know more about your model. Is this 1D unsteady? Are you intending to model sediment? Please describe.

[Scott Miller]

Good point. Thank you, Cameron!

[Scott Miller]

Thank you, Jarvus. I did a bit of testing with smaller time steps. Had lock them in rather than allowing reasonable Courant conditions from lengthening the time step. I used a ten day period with storms, and got the following results:

So, it took 56 hours at 0.2 seconds to model 1 year. The same model without the box culvert takes 5 1/2 hours at 2 seconds, Courant number entirely below 0.5. The smaller time step was needed entirely because of the box.

[Scott Miller]

Thank you, Jarvus. These geometries are all 2D, with internal connections.

I’ve run the 96, 72, and 48 foot meshes at both 4 seconds and 2 seconds, and each mesh gets the same results regardless of the time step. The volume error is clearly smaller with the smaller time step – same results.

As it turns out I put Courant criteria in a plan without checking the box. At a glance the Courant numbers range from 0.01 to 0.35, well below the criteria.

I think the model is solid enough I can go forward with it, and that minute differences in results are artifacts of the mesh alignments on terrain. Calibration would have more bearing on the results than mesh artifacts.

[Scott Miller]

I take it the entire channel is offset by the same distance and direction. Is that right? How far is it off?

What projections are the terrain and image data in? Check to see if a datum transformation is needed.

[Scott Miller]

Make sure your initial and boundary conditions keep all cross sections wetted. A steeper reach between cross sections can project a water surface below the downstream cross section. Is there anywhere in your model this might happen?

[Scott Miller]

1D, 2D, or a mix?

[Author]

Posts