CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   FLUENT (http://www.cfd-online.com/Forums/fluent/)
-   -   Simulation of filling a pool by Vof (http://www.cfd-online.com/Forums/fluent/72676-simulation-filling-pool-vof.html)

Asghari_M February 14, 2010 05:53

Simulation of filling a pool by Vof
 
Hi all;

I gonna simulate filling a pool(pond) by VOF method. My Dimensions are: 10m leng and 10 m wide and 1 m depth. I did a initial and rough 2d simulation but it didn't work well. This pond is connected to free air. Pool should be filled by a feeding pipe between 1 to 10 days. Certainly, Should I consider this problem as 2d? How many can pipe size be ? How should I mesh my geometry by gambit?

Thx a lot for any suggestion.

jack1980 February 14, 2010 06:13

Starting out 2D is an excellent idea. Start out with the most simple approach and if that works, make it your improvements. Some suggestions:

- VOF requires setting a lot of options. Start out by copying the settings from a 2D vof tutorial.

- Start out with simple boundary conditions. For example, use a velocity_inlet and an outflow. This will speed up your simulation and avoid backflow, that you are likely to get with a (to large?) pressure_outlet.

- Use a rectangular grid. Have your velocity_inflow in the side of the pool, at the corner near the floor, such that the water will flow in horizontally over the floor. Have your outflow somewhere in the ceiling, such that the excessive air can flow out (again you could also opt for a pressure_outlet, although in my experience this will give more trouble). Make your grid a bit finer in the inflow corner.

Good luck!

Asghari_M February 14, 2010 07:50

I have to pour water by a globe or floating valve in 0.5 m distance over floor. So, I can't locate inflow condition near the floor corner. Everyone is willing to have my geometry percisely, I can give him my configuration, if anyone send his/her email for me.

Thx a lot.

jack1980 February 14, 2010 11:20

having your inflow 0.5 m high is more complicated. it is possible of course but you'll have to be carefull how you mesh it.

i can only suggest: start out with the most simple geometry. if that works, move to the actual problem step by step.

Philipov February 14, 2010 12:51

Try to define you information needs... what do you want to find with this simulation?! Filling time or what?! for this problem it is possible to solve main parameters by hand...
And it is true that out flow is working for BC BUT.... see UG ... there it is written... outflow is not suitable for VOF!!!!

jack1980 February 14, 2010 13:56

Ah I see what you mean. I have been using outflow bc for open channel flows. But indeed, for this case a pressure_outlet seems best after all.

So I have quickly set up a simple case. The velocity_inlet is at bottom left. The top half right is a pressure_outlet. It serves as an overflow, so the pool should fill up half way.

Here you can see the water flowing in:
http://www.freeimagehosting.net/uploads/7856435461.jpg

This is the pool partly filled:
http://www.freeimagehosting.net/uploads/ac7a1f38b2.jpg

And this is the final steady state solution:
http://www.freeimagehosting.net/uploads/74a623aaf7.jpg

However doesn't run very smoothly since there is some backflow here already. If you want I can send you the case. You might get rid of the backflow, move the inflow up and then go 3D.

Anyway, good luck; it seems like a fun problem.

Asghari_M February 15, 2010 01:22

Ok, I will apply your suggestions and comback to discuss you about the consequences soon. I am studying on thermo-hydrolic transient behaviour of a pool. In next step, I will study thermal modeling of this pool such as thermosyphon forces in this pool. For example this pool might be heated by cheap source of energies such as solar or heat exchanger by ground heat source or other sources. I'll return soon.
Thx for your suggestions. See u soon.

Asghari_M February 16, 2010 02:47

1 Attachment(s)
I present my configuration in following picture:
Attachment 2275http://www.cfd-online.com/Forums/C:\...s\untitled.JPG
It seems to be a stable solution in time step=0.02 & inlet velocity= 0.0001, but when I selected inlet velocity =0.0.0001m/s and time step=0.2, I saw some divergences. Then, I alter gravity acceleration to 0.981 instead of 9.81 m/s2, I observed some convergences. This was very interesting for me. I think that I should modify mesh for better solution & a stable result in g=9.81 m/s2. What’s your idea?
L=2m. H=2m . Velocity inlet is located in h0= 0.69~0.7 m.


Philipov February 16, 2010 03:30

Do not even try to solve the problem with this mesh...it is almost impossible to get useful solution with this mesh quality.....
Increase the number of nodes...


All times are GMT -4. The time now is 23:21.