hi, I have a model wich combine natural convection and radiation. It is enclosed volume ( actually one of the surface is an outlet for radiation but there are no real wall but because there are no flow of air inside the volume , i assume it being wall). So one face is at more than 1000 degC (from where is coming the radiation), and i want to see how this face is going to radiate the entire volume. But for my bondary conditions i assume the temperature of the others face: the side wall at 110 deg C (because of cooling) and the outlet wall (actually not real wall) at 300 degC (just a guess). I am using DO model, but i cant get convergence. I would like to have some advice from you to set up my model. I tried to follow the tutorial 5 of fluent, but first their gradient of temperature is less important and secondly i dont get why they adjust the value of g. Thanks for your help cos i am really screwed .

Furthermore, is the boussinesq density model right to use in this case? I think my gradient of temperature between wall is to big to use constant density, right? which density model should i use? I am a beginner so i will really appreciate your help. thanks

For my density model i am using incompressible ideal gas, is that right?

radiation seems to work, cos my model is converging when i remove gravity. But when i want to introduce gravity then my residual doesnt go below 0.1. I tried to play with under relaxation factor without sucess. I am using presto for pressure and simple pressure velocity solver. Still waiting for your help please

If you'd have read the Fluent's Doc more carefully, you could see that for the Boussinesq approx. it is clearly stated that it cannot be used if the temperature gradient in the flow domain is too big (consequently the density variation is important). So you must use at least the (compressible) ideal gas law for density.

If I remember correctly, the reason for modifying the g constant in that tutorial was to obtain certain values for the dimensionless criteria, but you don't need to do that usually.

Razvan

thank you razvan, actually i just wanted to be sure that boussinesq was not right in my case. I am using incompressible ideal gas. But i still cant get convergence. Actually i am going to try to set up a turbulent flow instead of laminar, cos my gradiant being important and the rayleigh number as well , it can be the cause of my non convergence.

Even with turbulence model, my solution doesnt converge. I am wondering if anybody has experience to set up BC for this kind of problem. I actually use only wall evenif 2 of my opposite vertical wall are not real wall. I know for sure the temperature of the face which radiate my geometry and I assume the other. But would it be better to use pressure inlet and outlet with the same pressure (cos there are no flow only natural convectiopn and radiation)?

what is real wall