turbulent parameters and convergence

joaopffg · March 10, 2015, 05:44

Hi there,
im simulating a kitchen air exhaust (Teka) using solidworks flow simulation. I created all the parts (box, rotor, rotating domains, etc...),
added the conditions (rotation, real walls on the box and rotor, inlet and outlet atm pressure) and refined the mesh (i used one part to refine
the volume of the rotor and outlet). I chose the inlet and outlet mass rate to control convergence (since it is a transient analisys and there
isnt much to converge inside). I ran the program and it ended up converging just fine and the values were as expected.
Then some questions appeared.

1st) i let the intensity and lenght of the turbulence with the default values. I know i cant know for sure the values, but i read (from the wiki on this site) that for the intensity its between 5% and 20% for turbomachinery (in general) and less for more "calm" flows.
Wich value should i choose in this case (inlet and outlet)? And will it affect the results that much? Im not interested in obtaining the exact
velocity values, just an estimate of the flow directions (stream lines). I will then perform PIV on the outlet to confirm the model.

2nd) i then had to change the rotor's velocity (from 190 rad/s to 140 rad/s). I just started a new project, in all equal to the one done before but with the new velocity. But now i've been having problems with the convergence... I dont know why since the only thing that changed is the velocity.

if you guys need more info just ask,
thanks

Boris_M · March 16, 2015, 07:32

Leave it with the default. I've only changed it once or twice when I had to compare it with wind tunnel measurements where the wind tunnel had given paramters for turbulence but other than that you don' have to change it 99.9% ever.

The convergence behaviour can change due to changes in the RPM or pressure difference for fan and pump applications. It might take longer to converge.
What inlet and outlet boundary conditions did you use?
Also what rotation method did you use (Averaging plane or sliding mesh)?

Boris

joaopffg · March 16, 2015, 14:11

First here's a picture of the system.
Picture1.jpg
I'm interested in the air direction in the rotor's exit (beginning of the exit tube).
The inlet and outlet BC are, since this is a kitchen air exhaust system, atm pressure.
And i used the sliding mesh. I dont know if SW flow sim provides us with other rotating regions methods.

Boris_M · March 17, 2015, 03:07

Yes, SWFS did have the rotating region or as you will find it in SWFS "Averaging plane" method. This method can have slight differences depending on how it is realized. We just renamed it in the menu to averaging plane because the rotating region as a body is necessary for both the sliding and the averaging plane method and we wanted to avoid confusion.
But you were correct to use the sliding mesh because if I see it right you have a tangential fan and not a radial fan and for that you need the sliding mesh. Reason for that it that the averaging plane method basically does what the name says. It cuts the rotating region into slices and averages the parameters on the circumference of the slices. So with a tangential fan you will have inflow on the one side and outflow on the other and simply said If you average +5m/s and -5m/s you will get 0m/s.

March 10, 2015, 05:44	turbulent parameters and convergence	#1
joaopffg New Member Join Date: Feb 2015 Posts: 21 Rep Power: 11	Hi there, im simulating a kitchen air exhaust (Teka) using solidworks flow simulation. I created all the parts (box, rotor, rotating domains, etc...), added the conditions (rotation, real walls on the box and rotor, inlet and outlet atm pressure) and refined the mesh (i used one part to refine the volume of the rotor and outlet). I chose the inlet and outlet mass rate to control convergence (since it is a transient analisys and there isnt much to converge inside). I ran the program and it ended up converging just fine and the values were as expected. Then some questions appeared. 1st) i let the intensity and lenght of the turbulence with the default values. I know i cant know for sure the values, but i read (from the wiki on this site) that for the intensity its between 5% and 20% for turbomachinery (in general) and less for more "calm" flows. Wich value should i choose in this case (inlet and outlet)? And will it affect the results that much? Im not interested in obtaining the exact velocity values, just an estimate of the flow directions (stream lines). I will then perform PIV on the outlet to confirm the model. 2nd) i then had to change the rotor's velocity (from 190 rad/s to 140 rad/s). I just started a new project, in all equal to the one done before but with the new velocity. But now i've been having problems with the convergence... I dont know why since the only thing that changed is the velocity. if you guys need more info just ask, thanks

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
convergence issue for transonic turbulent case	aeroiitkgp	SU2	5	May 12, 2015 16:44
Turbulence parameters and convergence	saisanthoshm88	CFX	3	October 15, 2013 06:00
Turbulence parameters and convergence	saisanthoshm88	FLUENT	2	October 14, 2013 14:41
convergence parameters	amv	Phoenics	4	September 21, 2005 06:49
Problem of Turbulent Viscosity Ratio Limited	David Yang	FLUENT	3	June 3, 2002 06:13

March 16, 2015, 07:32		#2
Boris_M Disabled Join Date: Jul 2009 Posts: 616 Rep Power: 23	Leave it with the default. I've only changed it once or twice when I had to compare it with wind tunnel measurements where the wind tunnel had given paramters for turbulence but other than that you don' have to change it 99.9% ever. The convergence behaviour can change due to changes in the RPM or pressure difference for fan and pump applications. It might take longer to converge. What inlet and outlet boundary conditions did you use? Also what rotation method did you use (Averaging plane or sliding mesh)? Boris

March 16, 2015, 14:11		#3
joaopffg New Member Join Date: Feb 2015 Posts: 21 Rep Power: 11	First here's a picture of the system. Picture1.jpg I'm interested in the air direction in the rotor's exit (beginning of the exit tube). The inlet and outlet BC are, since this is a kitchen air exhaust system, atm pressure. And i used the sliding mesh. I dont know if SW flow sim provides us with other rotating regions methods.

March 17, 2015, 03:07		#4
Boris_M Disabled Join Date: Jul 2009 Posts: 616 Rep Power: 23	Yes, SWFS did have the rotating region or as you will find it in SWFS "Averaging plane" method. This method can have slight differences depending on how it is realized. We just renamed it in the menu to averaging plane because the rotating region as a body is necessary for both the sliding and the averaging plane method and we wanted to avoid confusion. But you were correct to use the sliding mesh because if I see it right you have a tangential fan and not a radial fan and for that you need the sliding mesh. Reason for that it that the averaging plane method basically does what the name says. It cuts the rotating region into slices and averages the parameters on the circumference of the slices. So with a tangential fan you will have inflow on the one side and outflow on the other and simply said If you average +5m/s and -5m/s you will get 0m/s.