Sediment boundaries

[Anonymous]

Hello forum,

I am currently working on a sediment transport model in HEC Ras 5.0 and i have some issues with the boundaries.

I have a river and set at the inflow cross section the boundary as “grading curve” as there is a function for a river’s sediment concentration depending on its discharge.
( C= 0.000209 * Q^1.5544, i change this to tonnes/day obviously)

Now when i run a model for a month with the maximum daily discharges observed (last 10 years) with Yang’s sediment transport function, I get a very big accumulation of sediment in the first cross section (looks like a triangle). the problem is that this triangle grows and grows the longer i run the simulation.

I have already been able to find out that it has something to do with the total load for high flows, obviously the incoming sediment cannot be transported and just settles right at the beginning. So I tried different loads for the bigger discharges but that does not represent our measured curve…

did I miss something?

[Anonymous]

Well I made some progress.

1. I was able to deal with the problem of the first XS. I just didn’t have enough XS’s to get a stable numerical solution.

2. Now I get outputs which are pretty oscillating… meaning alternating unrealistic highs and lows, but at least in the chainages where one would expect sedimentation phenomena. So obviously my solution is numerically not stable. Now there are a few options I have,

-lowering the computational increment (im having 1hr for 24hr flows at the moment and spatially XS every 5m)

-changing the number of updating the sediment rate in each timestep (is by default 10)

– changing other sediment output parameters

Final question, somebody knows a paper or the formula regarding the CFL-criterion when i use sediment transport mechanisms?

Thx

[cameron]

It appears you are running the model in unsteady with sediment transport instead of quasi-steady.

5 meters is pretty close for cross-sections and I would assume you would need a smaller time step. I would recommend getting the model working with good results without sediment transport. Once you are happy with the results, turn sediment transport back on.

Note that the program is keeping track of sediment transport, it will only adjust the cross-sections based on what you specify and what you pick is dependent on your time step.

For the CFL criteria, I would use the same one you would without sediment transport.

Once the final version of HEC 5.0 is released there will be a brand new manual just for sediment transport which will have a lot of good information on all of the questions you have.

[Anonymous]

Well, yes I use Quasi steady flow with time series of flow for 24hrs.

Back to the CFL criterion, so I have now XS’s in 25m and an average velocity of 4 m/s.
So my delta_t should be delta_t<=25m / 4 m/s ... which is around 6.25 s Thats quite small for a time series of a year haha... I saw HEC RAS 5 is using two cores, is there a way to use more than two?

[cameron]

Quasi steady is different than full unsteady with sediment transport and your time step will be different. The CFL criterion is generally for unsteady, unsteady sediment, and 2D. I would think hourly would work with Quasi steady.

As for cores, the only place I know of specifying the number is in the Unsteady calculation options window.

[Author]

Posts