Learning Forum

[akban]

Hi,

I want to simulate open channel flows where flow velocity and waves on interface co-exist.

My 2D geometry is shown here:

[Rob]

The surface Free surface needs to be away from the water free surface and set as a pressure boundary but not part of the open channel. Open channel is for where the waves enter and leave the domain.n

[akban]

Dear sir,nI did what you suggest. The free surface in my image is at z=0.5 m and water free surface is at z=0.2 m. Please check pressure outlet boundary condition.n

[Rob]

What is pressure outlet set as? You need more spare height and if the domain is only a few metres long you may want to consider how big the waves should be and what the model was designed for. n

[akban]

The pressure outlet is se as pressure outlet. The domain dimension: length, L=5 m, water depth, d=0.2 m, domain height, H=0.5 m. Velocity of water, v=0.2 m/s.nnPlease suggest.n

[Rob]

I'd make the height enough that the waves aren't anywhere near the top of the domain. With a shallow depth you'll need to choose the wave conditions carefully, you also need more mesh in the domain to track the free surface. n

[akban]

Thanks.nFew points:nI provided a wave of zero amplitude to check if I get water flow velocity contour properly. I thought upon getting that, I would introduce waves of non-zero amplitude for future simulations. But the velocity contour came to be weird -- my main issue is with simulating the water flow.nIf my way of producing open channel flow (without wave) is wrong, please suggest the better way.nnRegards,nAyann

[Rob]

Did you patch in the water? We do use the model without waves regularly for various applications. n

[akban]

Hi,nnI did try patching the water. Even with this my velocity came to be unreal. nFor example, I used only open channel option (leaving open channel wave BC) :nI used water density as compressible:nI did patch the domain of water. nBoundary conditions:nInlet- velocity inlet, noutlet & top of the domain- pressure outletnbottom- wallnnThe solution I got after 24 seconds is:nPlease suggest.nn

[YasserSelima]

You can ignore the compresability of water ... I am not sure is your inlet 0.5 m height or 0.2 m height (Whole domain height, or just water height)nSecond with the dimension you mention and the mesh on the first post, you will not get the free surface. You need fine mesh around the interface ...nFinally, if you simulated 24 seconds between your comments ... you are using a super super fast HPC or you are using very large time step. And most probably you don't do enough iterations to convergen

[akban]

Hi,nMy domain height is 0.5 m, water height is 0.2 m. My entire inlet is 0.5 m.nIs my method is correct?.Regards,nAKBn

[YasserSelima]

I think you should have the inlet velocity inlet of a height of 0.2 m, not 0.5 m. If you keep it 0.5 m, you will have flow rate more than you want. Also. it might be source of instability of the free surface.n

[akban]

Hi,nI tried to do so. In that case, the problem is what condition should I use for top 0.3 m? If I use pressure inlet for top 0.3 m, then it takes secondary phase by default. That's the problem.n

[YasserSelima]

You are right, It should take the secondary phase. But you should check the open channel wave BC to be able to put the free surface at the inletn

[akban]

But if it takes secondary phase in the top 0.3 m, where secondary phase (water) does not exist, won't it give wrong results? n

[YasserSelima]

No, it will take airn

[akban]

Thanks. I will try and inform again.nnRegards,nAyann

[akban]

One point:nI am using velocity inlet in the water inlet. I have clicked open channel wave BC & Segregated velocity inputs, where I have provided velocity for the incoming water. Primary phase has no velocity. Please refer the image in the first message. Is it okay ?.

[akban]

Hi,nIf I use height of 0.2 m to be the inlet, then the problem is what condition should I use for top 0.3 m ? If I take pressure inlet, it takes secondary phase by default.nThat's the problem.nn

[Rob]

If you pick waves then the velocity boundary gives nnSo you specify inlet velocity for both phases. Note, you need sufficient resolution to do this. If you don't have waves we'd tend to use a velocity inlet for the water & pressure for air that's above it. n

[akban]

Hi,nI have followed all the suggestions.nMy domain height is 0.5 m, water height is 0.2 m. I have two inlets- one water inlet , and another is air inlet. Water flows in the channel and air remains stationary. I am using velocity inlet for the water & pressure inlet for air that's above it. I am using open channel wave BC option along with segregated velocity input (water velocity -0.15 m/s, air vel- 0 m/s).nMy resolution is high, time step is 0.001 (initially) and 0.01 later on.nBut velocity contour is not acceptable.nKindly suggest.nnRegards,nAya

[Rob]

Is it converged, and is the mesh resolution sufficient? n

[akban]

Yes, it is converged and the resolution is pretty highn

[Rob]

Pictures pleasen

[akban]

n

[Rob]

And convergence? n

[akban]

Simulation did converge at every step.nI ran in transient. n