# fluent convergence problem

 Register Blogs Members List Search Today's Posts Mark Forums Read

March 12, 2014, 13:59
fluent convergence problem
#1
New Member

Join Date: Feb 2014
Posts: 8
Rep Power: 5
hi guys,

I am having difficulty with convergence,especially continuity residuals.
My model and boundary conditions are shown below.Air flows in a swirling pattern and creates negative pressure in the central region.
the nozzle is 1.5 mm and the gap between skirt and work-piece is 0.3 mm.
I want to find the pressure distribution on the work-piece..
The peak velocity is around 100 m/s only and so pressure based solver is sufficient...right?

The mesh created has a max skewness of .7 with average skewness of .2
Is the mesh too coarse.do i have to increase the no. of elements where thickness is very less.( I am modelling the fluid domain only and considered an outer envelope to apply pressure boundary condition)

How to attain faster convergence for this case.
should I use simplec or coupled ??
do i need to modify the under-relaxation factors ?
also,how to reduce back-flow ?

Any help is appreciated...
Attached Images
 111.PNG (78.2 KB, 35 views) 222.PNG (22.4 KB, 30 views) Capture.jpg (100.5 KB, 31 views)

March 13, 2014, 03:24
#2
Member

Simas
Join Date: Mar 2012
Posts: 37
Rep Power: 7
Quote:
 Originally Posted by itzme do i need to modify the under-relaxation factors ?
Yes you can do this higher under-relaxation factors will accelerate. If iterations grahps go up it means you have too high relaxation factors and need to lower it.

 March 13, 2014, 08:31 #3 Senior Member     Philipp Join Date: Jun 2011 Location: Germany Posts: 1,297 Rep Power: 20 itzme, 1) What turbulence model do you use? 2) Can you show the residuals? 3) Numerics ("Solution Methods") are? __________________ The skeleton ran out of shampoo in the shower.

March 13, 2014, 08:46
#4
New Member

Join Date: Feb 2014
Posts: 8
Rep Power: 5
Quote:
 Originally Posted by RodriguezFatz itzme, 1) What turbulence model do you use? 2) Can you show the residuals? 3) Numerics ("Solution Methods") are?

K-e RNG model
I dont currently have the residuals picture...everything except continuity converges...
Simplec... Initially with first order and then to second order...

 March 13, 2014, 10:34 #5 Senior Member     Philipp Join Date: Jun 2011 Location: Germany Posts: 1,297 Rep Power: 20 Di you try to run the whole case with much lower massflow? Try e-5 inlet massflow instead of e-4, and tell us what happens. __________________ The skeleton ran out of shampoo in the shower.

March 13, 2014, 10:51
#6
New Member

Join Date: Feb 2014
Posts: 8
Rep Power: 5
Quote:
 Originally Posted by RodriguezFatz Di you try to run the whole case with much lower massflow? Try e-5 inlet massflow instead of e-4, and tell us what happens.
Well...I managed to achieve convergence... but the pressure distribution on the plate is not accurate...
Will improving mesh or increasing no. of elements yield correct results??

 March 13, 2014, 10:52 #7 Senior Member     Philipp Join Date: Jun 2011 Location: Germany Posts: 1,297 Rep Power: 20 You mean convergence for the e-5 case? Can you post a picture of the y+ value on your surfaces? __________________ The skeleton ran out of shampoo in the shower.

March 13, 2014, 11:15
#8
New Member

Join Date: Feb 2014
Posts: 8
Rep Power: 5
Quote:
 Originally Posted by RodriguezFatz You mean convergence for the e-5 case? Can you post a picture of the y+ value on your surfaces?
No...for the e-4 case itself...and I am not using any type of inflation in the mesh....is y+ value significant for accuracy and convergence of solution....
(sorry, I am only a beginner in cfd analysis)
Thanks

 March 13, 2014, 11:18 #9 Senior Member     Philipp Join Date: Jun 2011 Location: Germany Posts: 1,297 Rep Power: 20 y+ is the normalized size of your cells, perpendicular to the surface, so its always plotted on the surface of your domain. If it is too large, the boundary conditions of the turbulence model become garbage. What did you change to achieve convergence? To your question: Of course results can get better if you make the grid even smaller. But you already have 1.3 mio cells for such a simple geometry. Do you have ICEM for meshing or just the ansys meshing? __________________ The skeleton ran out of shampoo in the shower.

March 13, 2014, 11:32
#10
New Member

Join Date: Feb 2014
Posts: 8
Rep Power: 5
Quote:
 Originally Posted by RodriguezFatz y+ is the normalized size of your cells, perpendicular to the surface, so its always plotted on the surface of your domain. If it is too large, the boundary conditions of the turbulence model become garbage. What did you change to achieve convergence? To your question: Of course results can get better if you make the grid even smaller. But you already have 1.3 mio cells for such a simple geometry. Do you have ICEM for meshing or just the ansys meshing?
Thanks for the info
I refined mesh a bit and changed the relaxation parameters.
I am not familiar with icem meshing.

 March 14, 2014, 07:11 #11 Senior Member     Philipp Join Date: Jun 2011 Location: Germany Posts: 1,297 Rep Power: 20 You should have a look at the y+ anyway. Maybe post the pictures! __________________ The skeleton ran out of shampoo in the shower.

 March 16, 2014, 00:23 #12 New Member   Join Date: Feb 2014 Posts: 8 Rep Power: 5 Is it possible to do a 2-D analysis for this case ???

March 16, 2014, 06:49
#13
Senior Member

Daniele
Join Date: Oct 2010
Location: Italy
Posts: 998
Rep Power: 17
Quote:
 Originally Posted by itzme Is it possible to do a 2-D analysis for this case ???
you mean axissymmetric; anyway no, because there is a tangential nozzle.

March 17, 2014, 10:11
weird CP distribution near trailing edge for NACA0012
#14
New Member

shubham jain
Join Date: Dec 2013
Posts: 25
Rep Power: 5
i am simulating naca0012 with Gurney flap
mesh using ICEM CFD
simulation using FLUENT, k-e RNG, Re=3e06
I have a problem in CLEAN Airfoil (without any gurney flap) in its Static pressure distribution (CP).

1. CP values of pressure and suction surface intersect each other near the trailing edge for every angle of attack.

2. Also there are many riggles (wavy nature of CP) in the CP distribution

Though the pathlines seem to be fine
can anybody explain this weird CP distribution and how can i rectify it???

Attached Images
 cp distribution cfd forum pic 1.PNG (11.3 KB, 7 views) cp distribution cfd forum pic 2.PNG (17.8 KB, 7 views) cp distribution cfd forum pic 3.jpg (72.4 KB, 7 views) cp distribution cfd forum pic 4.jpg (76.6 KB, 6 views)

 Thread Tools Display Modes Linear Mode

 Posting Rules You may not post new threads You may not post replies You may not post attachments You may not edit your posts BB code is On Smilies are On [IMG] code is On HTML code is OffTrackbacks are On Pingbacks are On Refbacks are On Forum Rules

 Similar Threads Thread Thread Starter Forum Replies Last Post Centurion2011 FLUENT 44 June 20, 2017 09:57 behest FLUENT 11 July 22, 2014 03:49 Tleja FLUENT 3 September 13, 2013 10:54 Marcin FLUENT 1 March 4, 2005 12:41 limseokmin FLUENT 3 November 14, 2004 13:43