[bigfoot47]

Hello everybody,


I want to condensate water vapor to water (liquid) in Ansys fluent. At the moment I am working with a 2D geometry, a rectangle with a height of 1 cm and a length of 100 cm. I tried a lot of different lengths from 1 cm to 100 cm. In my mesh there are 60000 knots with a sizing of 0.5 mm and with 10 inflation layers at the heat transfer side. I tried a lot of different meshs too but when i use a smaller number of knots it seems like nothing would happen.

I am using: pressure-based, transient and planar with gravity. As model the volume of Fluid model with two eulerian phases (water as primary and vapor as secondary) implicit formulation and the implicit body force. The energy equation is turned on. The standard k-epsilon turbulence model with the standard k-epsilon Model and enhanced wall treatment. As material of the condenser I kept aluminium by default.
The condenser is a stationary wall with a constant temperature of 273.15 K. At the velocity inlet steam comes in with a water vapor volume fraction of 1 and a temperature of 373.15 K and a velocity of 5 cm/s.

The Phase interaction is turned on from water to vapor (evaporation- condensation model)
should this be the other way round or is it not important? The Model is the Lee-model with data kept by default (saturation temp. 373.15 K).
By solution methods i tried every single one of them but it seems that the PISO scheme with Third-order Muscl at tubulence, dissipation, momentum and energy works best.


I am initializing with standrad initialization and a temperature of 373.15 K and the box tatally filley with vapor and patching the volume fraction of vapor with 0.999999 because when there is only vapor and no water nothing condensates in de box. I don´t now why there has to be a little water in it, can somebody maybe explain?


As time stepsize i used almost everything from 0.5 s to 0.00001 sn and 20- 50 max iterations per step.

I have a lot of trouble with reversed flow when i am using only a velocity of 5 cm/s. When i use 10 cm/s there is no problem with backflow. But i think 10 cm/s is too fast that the vapor can totally condensate.

When i simulate a shorter period of time the "water layer" looks bigger but the water volume fraction is about 1 x 10^-5. When i simulate a longer period of time ( maybe about 200s) the "water layer gets smaller but the water volume fraction increases to about 1 x 10^-3.
A further problem is that i can not really imagine the volume fraction. I wanted to show the results of the phases as mass fraction but i think it is not possible in fluent? Is it maybe? Water steam should take a much larger amount of space as liquid water or am i wrong? But even when i simulate longer like 5-10 minutes the water volume fraction is not increasing. Shouldn´t there be a formation of water droplets at the condenser sometime?


I would be so glad if somebody could help me. I am working on a solution for almost half a year and there is not really a progress.

Thanks a lot,
Raphael.