Hello!
I am having problems with setting up boundary conditions properly. The subject is natural convection.
I would like to have an open air volume with 2 solid walls (side and bottom). I set up the project and it works fine (image 1)
I also need to specify the temperature of 1 side wall to be 16 deg C. Ambient temperature must be 20 deg C. I have tried with many set up options but results are weird. The air flow is very different from results that I had with only ambient temperature set up to 20 deg C (image 2)
I even have tried with an extra wall with temperature/film coefficient boundary condition. (image 3)
Flag "use_property_ref_temp value" is set to 1 according to this soulution strategy guide:
I would be grateful if someone could give me tips/solution how to set up scene/boundary conditions properly.
Thank you.
Pablo, Student.
Solved! Go to Solution.
Solved by Jon.Wilde. Go to Solution.
Hi Pablo,
Your model looks over cosntrained to me, we should not need anything on the side walls (other than your fixed 20C - might a film coefficient be more realistic though?).
Please check out the guide here for how to set these models up - use the chimney approach.
Kind regards,
Jon
Hi wildej, thank you for your answer.
I just wanted to simplify model in order to save time needed for solving . The situation I want to simulate is a radiator situated near one wall (which have temperature 16 deg C) when the air temperature is 20 deg C.
I followed THIS guides to set up film coefficient coundary condition on the top surface of the air. I also set those conditions to 3 other sides of the air to simulate situation when air comes from the side. Image 1 in the first post shows that it worked.
Do I have to model the room or there is other approach I could do to simulate this?
Pablo
Hi Pablo,
It is probably better to model the room and use a 'bucket' approach with just the conditions on the top surface of the air. P=0 and a Film Coefficient.
If you set all the walls open you will run into issues as we ideally want air only entering or exiting the domain over a P=0. If it recirculates we can end up with poor results.
Kind regards,
Jon