# 3D melting problem with natural convection

 Register Blogs Members List Search Today's Posts Mark Forums Read

March 11, 2013, 10:58
3D melting problem with natural convection
#1
New Member

manuel
Join Date: Mar 2013
Posts: 5
Rep Power: 13
Hello fluent users,
my3D geometry consists of a steel pipe which is surrounded by a pcm (phase change material). The boundary condition of the inner pipe wall is convection with heat transfer coefficient and free steam temperature. The free steam temperature is above the pcm melting temperature. So the pcm should melt. There is also gravity. Thus I expect natural convection.
When the melting process starts on the interface between the steel pipe and the pcm, it happens in longitudinal oriented lines (see the picture). It does not melt uniform over the whole surface. After a few time steps the calculation dos not converge anymore.
My settings are:
All surfaces of the pcm are symmetrie. All properties of the pcm are a function of temperature (written in a UDF). I use Simple scheme, standard pressure discretization and the under relaxations factors are set by default. When i lower the relaxations factors, i have the same problem. Timesteps are 2 second. I also tried a much smaller time step (0,02s), but I did not work either.
What do I have to do to make the melting process going??
Attached Images
 pcm_melting.JPG (28.6 KB, 51 views) pcm_melting_residuals.JPG (35.2 KB, 46 views)

 March 12, 2013, 02:21 #2 New Member   manuel Join Date: Mar 2013 Posts: 5 Rep Power: 13 Please can anybody help me!

 January 16, 2014, 11:50 #3 New Member   Yo Join Date: Mar 2010 Posts: 2 Rep Power: 0 you are considering the flow equation in your model, the pcm has high viscosity compared to water, the dominant heat transfer mode is conduction and not convection, how did you model the density?

 Tags convergence problem, pcm, solidification/melting