[FrankW]

Hi,
how can i simulate (transient) natural convection of a liquid (like coffee in a cup). Is there a posibility to avoid vof. And when how define i the boundary conditions of the top (free surface). My first ideas was to use a symmetry condition at top, but it does'nt work (i get no convergence). The reason could be the increase (decrease) of volume. Has anyone ideas.

thx Frank

[A_Lyaskin]

You can try with slip wall. But you'll also have to specify some boundary condition for temperature/heat flux in this case.

[wiebe]

Hi Frank,

One possibility to avoid the problem that a change of volume in the liquid poses is to include a pressure relief. This is implemented in different ways in different programs, but what it does is that it fixes the pressure on one point.

Another problem with the coffee cup is the different heat losses that occur. There is transmission to all sides, where the heat loss largely depends on air speed around the cup. The heat loss on the surface is even more prominent, as there is evaporation involved. (That's why you blow the surface of the coffee, and not the side of the cup to cool it down...).

How you simulate the coffeecup depends on what you want to know. If you want to simulate how the water-vapour escapes the mouth of the cup, for example because you want to optimize the shape of the top of the cup, then you need to use VOF, possibly even without simulating the coffee itself. If you want to know when it is safe to sip, but not safe to gulp it down, you really want to know the temperature distribution in the fluid, and a first order investigation might want to simulate the coffee only.

Kind regards,
Wiebe.

[FrankW]

Hi Wiebe,

thanks, the solution with the pressure relief solved my problem. By the way the cup of coffee was only an example of my problem. Because i have a long period to simulate and so i would avoid a vof simulation with short timesteps. The used termal boundary condition are predefind for a horizontal and vertial plane.

regards
Frank