Discussion Forum

Posted: 1 decade ago 10 avr. 2013, 06:03 UTC−4

I forgot of mention this: The heat source is should be variable in time, but I emulated it with a constant one that creates an amount of heat even to the medium value of the real one in one period... is correct? Or non linear properties of the materials could make this setting a not realistic one?

Posted: 1 decade ago 10 avr. 2013, 12:50 UTC−4

Hi your "air" inside the bioxes, how is it simulated ? I would say you should rather use conjugated heat flow NITF to get convective bulk, but also internal transport, conduction and possibly radiation exchange between solid and fluids Now I'm not sure any longer how that was implemented in 3.5, sorry its getting to long since I last run that version ;) -- Good luck Ivar

Posted: 1 decade ago 11 avr. 2013, 08:58 UTC−4

Thanks for your help. Actually, I have considered just the conductive heat flow through the air inside the box. I considered the assuption reasonable due to the following: 1)The volume is not very big 2)The volume is totally surrounded by metal, so the thermal gradients are negligible. 3)The heat source is in direct contact with the metal walls of the box, so there is a 'thermic short circuit' from the heat source to the external air, and I expect that the heat flow through air is quite negligible, while air should be considered justa as thermal capacity. Do you have experience that in similar condition I should consider better the internal convection?

Posted: 1 decade ago 11 avr. 2013, 14:35 UTC−4

Hi sounds reasonable (I do not know you model but still, you have common arguments) One way to tke a look is to integrate and see what part of the heat flux is going through the air and how much through the solid walls, then double the heat flux through air (or double the air conductivity) just to get a sensitivity feeling (or perform a full sensitivity analysis,but that takes more time) -- Good luck Ivar

Posted: 1 decade ago 11 avr. 2013, 14:45 UTC−4

Hi Francesco, It’s not surprising that the temperature boundary condition dominates since the conduction through the metal is far more significant than the convection to the outside air. I would avoid that condition if possible and try to do something more true to the physics along that boundary. Based on what you described, I agree that internal convection should not be significant. I believe if you are only interested in the steady state solution then it’s fine to not explicitly model the internal convection. If you want the transient behavior then you may find that the internal convection makes a bigger difference. It may still be insignificant though … it all depends on the geometry and properties. Nagi Elabbasi Veryst Engineering

Posted: 1 decade ago 12 avr. 2013, 04:28 UTC−4

[QUOTE] Hi sounds reasonable (I do not know you model but still, you have common arguments) One way to tke a look is to integrate and see what part of the heat flux is going through the air and how much through the solid walls, then double the heat flux through air (or double the air conductivity) just to get a sensitivity feeling (or perform a full sensitivity analysis,but that takes more time) -- Good luck Ivar [/QUOTE] I will definitively follow your suggestion for having a more reliable confirm of my assumption... also if I expect that in a steady state simulation the impact of the heat flow through air will not be very important. Thank you again. [QUOTE] Hi Francesco, It’s not surprising that the temperature boundary condition dominates since the conduction through the metal is far more significant than the convection to the outside air. I would avoid that condition if possible and try to do something more true to the physics along that boundary. Nagi Elabbasi Veryst Engineering [/QUOTE] Hi, thanks for the help to you, too. I was thinking the same, but I have seen that setting the temperature of the inlet surface is the most common setting present in similar situations on other works... and above all I did non succeded in finding any other possible options... do you have any suggestions? I tried also to set an external convection to the inlet boundary...but, depending on the chosen parameter h, I have obtained similar results, or problems of convergence.

Posted: 1 decade ago 12 avr. 2013, 15:13 UTC−4

You're welcome. Depends on what is connected to that surface. If it is a solid such a pipe, wall or ground I would model part of that structure. If it is a large good-conducting connection then it’s likely that the prescribed temperature condition you already use is accurate enough. Nagi Elabbasi Veryst Engineering

Posted: 1 decade ago 22 avr. 2013, 05:24 UTC−4

[QUOTE] You're welcome. Depends on what is connected to that surface. If it is a solid such a pipe, wall or ground I would model part of that structure. If it is a large good-conducting connection then it’s likely that the prescribed temperature condition you already use is accurate enough. Nagi Elabbasi Veryst Engineering [/QUOTE] I managed to avoid the prevail of the temperature condition making a longer tube and setting this condition at a bigger distance from the metal box, but now I have a lot of problems of meshing and convergence, and I still have problems to understand the right coundition for the Navier-Stokes physics.