Internal Airflow Analysis of a Car Cabin

mjkey · September 7, 2014, 20:01

Hey guys,

I'm working on a project for the internal airflow of a car cabin. This includes alot of swirling flows and recirculation, so are there any specific turbulence models recommended? I'm assuming RNG-Keps or Realizable Keps.

I've tried modelling with Realizable Keps but have been having convergence issues. It is an unstructured grid using tetrahedrons all around. I'm using the "Coupled" setting and the solution blew up "Divergence Detected in AMG Solver" with the default settings of Courant: 200 and URF: default.

Do you guys have any recommendations of how to approach this problem to get convergence? The geometry is pretty complicated so it's very difficult to have a structured grid, so I'm gona stick to an unstructured tetrahedrons grid.

Should I start off with default values and slowly reduce the Courant number and URF's?

Cheers!

adruma · September 8, 2014, 17:32

Did you try K-omega SST? try that and also try to use a fine mesh. Sometimes it diverges due to the mesh quality.

Adriana Druma
Aavid Thermalloy

mjkey · September 8, 2014, 19:06

Quote:

Originally Posted by adruma

Did you try K-omega SST? try that and also try to use a fine mesh. Sometimes it diverges due to the mesh quality.

Adriana Druma
Aavid Thermalloy

Hey adruma, I'm trying K-w SST now. Is there any specific reasons to use this model instead of k-epsilon? I'm also trying a fine grid with the CutCell Method. The tetrahedral mesh definitely had bad quality due to the complex geometry.

ghost82 · September 9, 2014, 04:14

I would try the rng k-eps; usually it's better for convergence than the realizable.
K-omega SST is good too, for transitional flows, but you need a finer grid.
Check your mesh and spend time to improve the quality; it is a must for convergence.

Daniele

September 7, 2014, 20:01	Internal Airflow Analysis of a Car Cabin	#1
mjkey New Member Michael Join Date: May 2014 Posts: 28 Rep Power: 11	Hey guys, I'm working on a project for the internal airflow of a car cabin. This includes alot of swirling flows and recirculation, so are there any specific turbulence models recommended? I'm assuming RNG-Keps or Realizable Keps. I've tried modelling with Realizable Keps but have been having convergence issues. It is an unstructured grid using tetrahedrons all around. I'm using the "Coupled" setting and the solution blew up "Divergence Detected in AMG Solver" with the default settings of Courant: 200 and URF: default. Do you guys have any recommendations of how to approach this problem to get convergence? The geometry is pretty complicated so it's very difficult to have a structured grid, so I'm gona stick to an unstructured tetrahedrons grid. Should I start off with default values and slowly reduce the Courant number and URF's? Cheers!

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Boundary Conditions	Thomas P. Abraham	Main CFD Forum	20	July 7, 2013 05:05
Car park ventilation and impulse fans	guillaume	Main CFD Forum	12	June 27, 2011 11:16
cfd analysis of a solar car	Sandile Peta	Main CFD Forum	0	January 29, 2006 06:05
Aerodynamic analysis of car	Vikas Kumar	Main CFD Forum	3	April 10, 2003 15:15
Simulation of Car Cabin Air Flow	Erwin Susanto	Main CFD Forum	0	May 19, 2000 11:33

September 8, 2014, 17:32		#2
adruma New Member Adriana Join Date: Sep 2014 Location: san jose, CA Posts: 9 Rep Power: 11	Did you try K-omega SST? try that and also try to use a fine mesh. Sometimes it diverges due to the mesh quality. Adriana Druma Aavid Thermalloy

September 9, 2014, 04:14		#4
ghost82 Senior Member Rick Join Date: Oct 2010 Posts: 1,016 Rep Power: 26	I would try the rng k-eps; usually it's better for convergence than the realizable. K-omega SST is good too, for transitional flows, but you need a finer grid. Check your mesh and spend time to improve the quality; it is a must for convergence. Daniele