How to simulate buoyant air correctly?
Posted: Sat Sep 03, 2011 7:40 pm
ello all,
I am about to conduct a coupled simulation between code-saturne and syrthes. In my case, I use the losses from a high-power coaxial RF cable as thermal source (thermal flow in SYRTHES) and the air between the two conductors is simulated with the help of code-saturne.
My questions are: How is it possible to incorporate the Prandl-number (see: http://fr.wikipedia.org/wiki/Nombre_de_Prandtl) into the k-epsilon model of code-saturne?
And is there a standard-scheme to determine the maximum possible step-size in order to attain minimum simulation times?
According to:
http://www.google.de/url?sa=t&source=web&cd=6&ved=0CEQQFjAF&url=http%3A%2F%2Fresearch.edf.com%2Ffichiers%2Ffckeditor%2FCommun%2FInnovation%2Flogiciels%2Fcode_saturne%2FDocuments%2F2.0%2Fuser.pdf&rct=j&q=%2Bprandtl%20code-saturne&ei=j2ViTubtNMvusgbgkbyHCg&usg=AFQjCNGi9sle7tL7Kva4Kb2EBtTKgS6yYg&cad=rja
there are 2 possible variables: sigmas and sigmak to incorporate a Prandtl-number of 0.7 (= Prandtl-number of air)
Since buoyancy is an effect due to differences in densities depending on temperature, in the first step I changed usphyv.f:
Density is changed according to: rho=p/(Rs*T), Rs=287.05 J/(kg*K) (see http://fr.wikipedia.org/wiki/Masse_volumique_de_l%27air)
As I think the correct simulation of air (in this case: buoyancy) is important for many different cases, I would be pleased, if someone could help me to setup an exact fluid-dynamic model of air. So, in this first step, air shalt be simulated with a maximum of accuracy. If this is setup, I would like to generate experience driven data to determine some kind of rule of thumb for a fast and accurate simulation which would be posted here.
Best regards,
Christoph
I am about to conduct a coupled simulation between code-saturne and syrthes. In my case, I use the losses from a high-power coaxial RF cable as thermal source (thermal flow in SYRTHES) and the air between the two conductors is simulated with the help of code-saturne.
My questions are: How is it possible to incorporate the Prandl-number (see: http://fr.wikipedia.org/wiki/Nombre_de_Prandtl) into the k-epsilon model of code-saturne?
And is there a standard-scheme to determine the maximum possible step-size in order to attain minimum simulation times?
According to:
http://www.google.de/url?sa=t&source=web&cd=6&ved=0CEQQFjAF&url=http%3A%2F%2Fresearch.edf.com%2Ffichiers%2Ffckeditor%2FCommun%2FInnovation%2Flogiciels%2Fcode_saturne%2FDocuments%2F2.0%2Fuser.pdf&rct=j&q=%2Bprandtl%20code-saturne&ei=j2ViTubtNMvusgbgkbyHCg&usg=AFQjCNGi9sle7tL7Kva4Kb2EBtTKgS6yYg&cad=rja
there are 2 possible variables: sigmas and sigmak to incorporate a Prandtl-number of 0.7 (= Prandtl-number of air)
Since buoyancy is an effect due to differences in densities depending on temperature, in the first step I changed usphyv.f:
Density is changed according to: rho=p/(Rs*T), Rs=287.05 J/(kg*K) (see http://fr.wikipedia.org/wiki/Masse_volumique_de_l%27air)
As I think the correct simulation of air (in this case: buoyancy) is important for many different cases, I would be pleased, if someone could help me to setup an exact fluid-dynamic model of air. So, in this first step, air shalt be simulated with a maximum of accuracy. If this is setup, I would like to generate experience driven data to determine some kind of rule of thumb for a fast and accurate simulation which would be posted here.
Best regards,
Christoph