[ANSYS Meshing] how to deal with unavoidable high skew cells

vipulramk · September 29, 2017, 04:49

Hello,

I have 3D conjugate heat transfer problem with around 5 crore cells, I am having a condition where few cells are having a skewness greater than 0.98 (~4000 cells), I am in a situation where I cant improve the mesh since I have "RAM" restrictions. I am presently using RNG K-epsilon, enhanced wall function turbulence model. I have observed that in these cells my turbulence model is unable to converge and producing misleading velocity results.
but my energy equation is has reached stability with an error 1e-6 very early during the simulation (around 50 iterations) can someone shed some light on how to tackle this problem?
Also, I am using ANSYS Fluent to solve and the geometry is an Electric motor and is quite complex.

Thank you.

plucas · October 2, 2017, 13:22

I would look into maybe a cloud solving solution where you can get a better mesh like Nimbix

vasava · October 6, 2017, 06:40

I recommend the following:

Convert mesh to polyhedral. This should significantly reduce number of cells.
Use fluent commands to remove mesh anomalies.
Make a simpler and coarse mesh. See if everything works ok. If so then make a finer and complicated mesh.

September 29, 2017, 04:49	how to deal with unavoidable high skew cells	#1
vipulramk New Member vipul ram k p Join Date: Apr 2014 Posts: 6 Rep Power: 12	Hello, I have 3D conjugate heat transfer problem with around 5 crore cells, I am having a condition where few cells are having a skewness greater than 0.98 (~4000 cells), I am in a situation where I cant improve the mesh since I have "RAM" restrictions. I am presently using RNG K-epsilon, enhanced wall function turbulence model. I have observed that in these cells my turbulence model is unable to converge and producing misleading velocity results. but my energy equation is has reached stability with an error 1e-6 very early during the simulation (around 50 iterations) can someone shed some light on how to tackle this problem? Also, I am using ANSYS Fluent to solve and the geometry is an Electric motor and is quite complex. Thank you.

October 6, 2017, 06:40		#3
vasava Senior Member Paritosh Vasava Join Date: Oct 2012 Location: Lappeenranta, Finland Posts: 732 Rep Power: 23	I recommend the following: Convert mesh to polyhedral. This should significantly reduce number of cells. Use fluent commands to remove mesh anomalies. Make a simpler and coarse mesh. See if everything works ok. If so then make a finer and complicated mesh.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
[ICEM] Problem with prism cells	sidharath	ANSYS Meshing & Geometry	0	September 1, 2015 08:09
use of high aspect ratio cells near walls	Usman	FLUENT	0	January 23, 2008 08:16
wedge cells pridicting high T-KE why?	shan	Siemens	0	September 21, 2006 15:12
physical boundary error!!	kris	Siemens	2	August 3, 2005 01:32
Multicomponent fluid	Andrea	CFX	2	October 11, 2004 06:12

October 2, 2017, 13:22		#2
plucas Senior Member Anonymous Join Date: Dec 2011 Location: USA Posts: 108 Rep Power: 15	I would look into maybe a cloud solving solution where you can get a better mesh like Nimbix