CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   CFX (https://www.cfd-online.com/Forums/cfx/)
-   -   How to model a lake? (https://www.cfd-online.com/Forums/cfx/19179-how-model-lake.html)

 Bowel September 21, 2002 11:43

How to model a lake?

Hi everyone,

I want to model a river entering into a lake. Sand and water will accumulate into the lake. Apparently this is a two-phase proble. I encount two difficulties:

First, how to set the domain which initially contain water and sand. There is a sharp interface between upper layer of clear water and lower layer of sand.

Second, how can I set the outlet condition. In reality, there is no outlet and the water surface will increase with time as more and more sand and water accumulte into the lake.

Anybody can tell me whether this simulation is possible in CFX 5.5.1.

Many thanks!

Bowel

 Holidays September 24, 2002 02:59

Re: How to model a lake?

Initial conditions are set up using cel (or fortran) esp as density difference (include buyoancy) will favour any stratification there is.

Regarding the free surface change, if I were you I would consider a mobile mesh (ie only model water and sand). Since you will know the inflow, you can deduce the level increase per timestep and deform a surface patch accordingly. Such an assumption would be more than adequate for a lake in my opinion. Additionally such deforming grid can be implemented either in CFX-4 (fortran) or 5 (ccl). Try this on a simplified case for a start.

 Astrid September 24, 2002 15:47

Re: How to model a lake?

Although I have never performed such a simulation, I think it is possible to use an imaginary box (intially containing air) above the lake with openings as BC, p=0. Then, when water comes in, air can flow out.

Astrid

 Holidays September 26, 2002 02:32

Re: How to model a lake?

I agree Astrid, but this might be expensive, esp in view of the fact Bowel wants to include a sediment phase as well. Maybe it is better to limit the number of phase to 2 here,ie exclude the air, esp if the lake increase in level is low (its surface being flat anyway), which is often the case (large lake, sea)and implies that lots of such simulations might even include a fixed level or a sin function if some sort of tide has to be accounted for.

 Bowel September 30, 2002 10:30

Re: How to model a lake?

Thank you Holiday. But how can I set the mobile mesh using CCL. Can you give me some hints?

Thanks,

Bowel

 Holidays October 1, 2002 11:12

Re: How to model a lake?

Hi,

You need to speak to the CFX techies/vendors; I know it works because I have seen the files for a piston test case. What you do is that you set up two BCs and set a motion between them, which will displace the nodes contained within that zone proportionally. It is quit neat. This was presented in various CFX communications and CFX conferences I attended.

I believe this is the most adequate method for you as (1) the surface motion is simple and 'flat', therefore this tool is physically correct, and (2) you will have other constraint with sediments!

 Jon October 3, 2002 11:23

Re: How to model a lake?

You can do this, but as dz/dt will be so slow, why dont you have the whol of the surface as an outlet and make the assumption that the chnage in water height is negligable?

 Holidays October 3, 2002 13:02

Re: How to model a lake?

Hi Jon,

Suggested that as well before. I actually think it would be the most viable IF the discharge is low in comparison with the lake area...

Hope all is well around Guilers ;-)

 Jon October 4, 2002 06:44

Re: How to model a lake?

So you did.

J'espere tous va bien pour toi et les famaille de trois!

a+ (peut-etre a Paris) Jon

 Bowel October 10, 2002 21:18

Thanks