Depth Peak at the beginning of a simple channel – 2D Modeling

[Anonymous]

Hi !

I created a simple channel with a trapezoidal section : 150 m long ; 20 m wide ; a slope of 2 %.
The boundary conditions are :
– at the beginning of the channel : a flow hydrograph with a flow of 50 m3/s for 3 hours of simulation time

– at the end of the channel : normal depth with a friction slope of 0.02

My mesh is made of 1m-squares and I created 2 breaklines : one for each bank of my trapezoidal channel.

After the unsteady flow computation, an element surprises me : as regards the depth for a longitudinal profile in the middle of the channel, I notice a depth peak of 0.46 m at the beginning of the channel. Then, the depth returns to a steady value of 0.28 m.

I do not understand why I observe this peak which is not physically explicable.
Does anyone have ideas ?

Thank you in advance for your answers,

[cameron]

What is your timestep and what slope did you enter for the upstream boundary condition? Are you using the Full Momentum Solver?

[gustbadell]

I am having a similar issue.

I’m running a 2d model, diffusion wave, 2m x 2m cell, and 1m x 1 m cell on brakelines around my channel.

The model behaves ok until peak flow comes, it seems flow spreads first towards banks and then goes downstream.

I established normal depth for boundary condition upstream with a 0.005 m/m slope. I have tested several time steps and cell dimension configuration.

I don´t know if th result is correct because I´m modeling a low slope to flat area with a big peak flow.

I´d really appreciate if someone tells me what I´m doing wrong.

[cameron]

what is your timestep? If you switch to Full Momentum and the model crashes, your timestep is too high. Do you need to use a cell of 2m or can you use a larger value?

[Anonymous]

I am modeling a kind of small stream, when I set a bigger cell size the model tends to make the flow advance very quickly and it spreads even wider, and I have tested several timesteps, since less than a second to a half and hour, When I set the 30 minutes timestep, the flow dissapears for few timesteps, like if the flow dries due to all volume goes off towards my model outlet

My hydrograph has a high peak flow, very big for the stream, so, It is logic to have overbank flow

[cameron]

You need to follow the Courant criteria when selecting a time step. A 30 min timestep is way to high. It should be on the order of seconds genearally and with a 2m grid it would be 1 second or smaller most likely.

[Anonymous]

I revised the results and I notice the depth at upstream boundary is ok. I analyzed flow spreading and realized the most of the flow is in the main channel, overbank flow is only few centimeters depth so I think the result is logic taking into account I’m modeling a mild to low slope area, draining a peak flow of 100 years return period.

I will open a new thread to ask some questions about weir coefficients to model flow over road embankments.

I really appreciate your help

[Author]

Posts