
[Sponsors] 
June 3, 2012, 08:04 
MUT/MU CLIPPEDFineTurbo

#1 
New Member
Join Date: May 2011
Posts: 3
Rep Power: 7 
I am performing a turbomachinery steady simulation for transonic compressor stage using FineTurbo. The inputs are
Inlet : Total pressure extrpolated(The profile values have been defined except for constant turbulent viscosity 0.0001 m2/s) Outet : Static pressure extrpolated Numerical Model : CFL 3 Midway during the simulation I recieve a message "MUT/MU CLIPPED FOR X CELLS" and the the convergence curve goes wild. I increased the MUCLIP value as suggested in the manual. But it shows the same message and behaviour as mentioned. My concern is that the the constant turbulent viscosity may be the reason for this behaviour. I am not aware of a method to calculate the turbulent viscosity. Any suggestions on the possible causes for this error message/ methods to calculate turbulent viscosity are welcome. Thanks 

June 7, 2012, 10:54 

#2  
New Member
Alberto
Join Date: Jan 2012
Posts: 1
Rep Power: 0 
Quote:
Hi, granted that I use NUMECA just from a couple of months, and I've found the same problem in some simulations, but the residul didn't have a wild behaviour. Personally I think that the reason is due to the profiles that you use as boundary conditions; they could be too "hard" for your machine. In particular if you have a transonic field, it doesn't help the convergence. Let me know. 

June 10, 2012, 01:49 

#3 
Member

Hi,
Please do check your mesh quality and make sure the qualities are following the criteria. Good luck.
__________________
Regards CJ NUMIT Enterprise, Malaysia +6 017 656 3436 http://maccomnumit.blogspot.com http://www.numit.com.my Email : seiko.q@numit.com.my 

June 17, 2012, 06:01 

#4 
New Member
Join Date: May 2011
Posts: 3
Rep Power: 7 
@ Alberto
The stage is transonic field. I tried using constant profile conditions but the result is the same. @ Old shoes My mesh conforms to the numeca recommended values wrt aspect ratio, orthogonality etc. The current mesh size is about 700000 for the stage. i will refine the mesh and try checking the results. Thanks Vikram 

July 30, 2012, 10:32 

#5 
Member

Hi,
Is there any improve computed results after your refinement? What is the Reynolds Number of your case?
__________________
Regards CJ NUMIT Enterprise, Malaysia +6 017 656 3436 http://maccomnumit.blogspot.com http://www.numit.com.my Email : seiko.q@numit.com.my 

July 31, 2012, 14:08 
Help need to simulate flow around Contra rotating Fan

#6 
Member
venkatesh
Join Date: May 2012
Posts: 93
Rep Power: 6 
I am trying to simulate flow around Contra rotating Fan using Full Non Matching
Mixing Plane. I meshed my geometry in Autogrid5 and I am trying to simulate in FineTurbo. The flow is steady and incompressible with a blade speed of 1029 rpm. I use Spalart–Allmaras turbulence model with velocity inlet (68 m/s), pressure outlet (Radial equilibrium) as the boundary condition and I imposed Farfield velocity as 68 m/s. The Periodic angle for the front rotor is 32.7272 degree and 40 degree. There are nearly 6.5 million cells. In the initialization of simulation, I gave 98500 pascal as inlet pressure and 90000 pascal as pressure in mixing plane. I ran simulation, after two iteration it say solution blow out, Process killed, Crashed. I tried with different boundary condition but nothing helped me. I am not able to figure out the problem. Please Can anyone give me some suggestions to run my simulation. 

August 1, 2012, 07:39 

#7 
New Member
Mark
Join Date: Mar 2012
Posts: 20
Rep Power: 7 
Hello,
I think it is a problem with your boundary conditions. Which value of CFL are you using? try to reduce it Sorry but i am not an expert in that problems... 

August 1, 2012, 07:53 

#8 
Member
venkatesh
Join Date: May 2012
Posts: 93
Rep Power: 6 
Hi daxterss,
Thanks for your reply. The value of CFL that I use is 3. I tried to reduce it to 1 but again the calculation blows up after two iteration. Currently I am using velocity inlet and pressure outlet. Could you please tell me what mistake am I doing in boundary condition. 

August 1, 2012, 07:58 

#9 
New Member
Mark
Join Date: Mar 2012
Posts: 20
Rep Power: 7 
Sorry but I am not an expert in this topic.
Are you running your simulation in the finest grid? or a coarse level? Try to get run in the coarsest level. And when you get the convergence, move to the next level using as inital solution, the solution you got before. 

August 1, 2012, 08:07 

#10 
Member
venkatesh
Join Date: May 2012
Posts: 93
Rep Power: 6 
Again thanks for your quick reply. I have started my calculation at coarsest grid level. I did as you said but still I have the blow up problem. I don't know what to do.


August 1, 2012, 08:11 

#11 
New Member
Mark
Join Date: Mar 2012
Posts: 20
Rep Power: 7 
Ok, if you try to converge in the coarsest grid and you can't (the coarsest is the grid, the easiest is to converge) I am pretty sure that the problem is with your b.c, but I can't help with this... Sorry
Hope you can solve your problem 

August 1, 2012, 08:15 

#12 
Member
venkatesh
Join Date: May 2012
Posts: 93
Rep Power: 6 
Thanks for your suggestion. one quick question. can i try with pressure inlet instead of velocity inlet. The problem is I don't know to set up pressure inlet. At inlet it show me only parameters associated to velocity.


August 1, 2012, 08:20 

#13 
New Member
Mark
Join Date: Mar 2012
Posts: 20
Rep Power: 7 
Sorry but I don't know which b.c are the best for your problem


August 1, 2012, 08:21 

#14 
Member
venkatesh
Join Date: May 2012
Posts: 93
Rep Power: 6 
Thanks for your suggestion. one quick question. can i try with pressure inlet instead of velocity inlet. The problem is I don't know to set up pressure inlet. At inlet it show me only parameters associated to velocity.


August 1, 2012, 08:22 

#15 
Member
venkatesh
Join Date: May 2012
Posts: 93
Rep Power: 6 
ok I will work on it and post my issues with my simulation.


August 1, 2012, 09:54 

#16 
Member

Hi,
What is the number of mesh you have within your fine grid and what is the number of mesh you have within your coarse grid?
__________________
Regards CJ NUMIT Enterprise, Malaysia +6 017 656 3436 http://maccomnumit.blogspot.com http://www.numit.com.my Email : seiko.q@numit.com.my 

August 1, 2012, 16:23 

#17 
Member
venkatesh
Join Date: May 2012
Posts: 93
Rep Power: 6 
Hi,
I dont know how to check number of mesh in fine grid and coarse grid. My Mesh has around 5.8 million points. The Number of grid level in quality report shows that 4 3 2. I stated the calculation with coars grid initialization with a CFL number 3. I dont know much about preconditioning parmater so I kept the default value. I will post some images so that you can check my boundary conditions and figure out my mistake. The flow is incompressible and Steady. 

August 1, 2012, 16:27 

#18 
Member
venkatesh
Join Date: May 2012
Posts: 93
Rep Power: 6 
In Preconditoning parameter I also tried Merkel method but it does not work,solution blow up after two iteration.


August 1, 2012, 17:44 

#19 
Member
venkatesh
Join Date: May 2012
Posts: 93
Rep Power: 6 
Hi,
My Calculation is running now. I reduced the CFL number to 0.0000001. In the user manual it was suggested to have a minimum CFL number between 0.5 to 1. Problem is that MAX_RES (Maximum residual) increases with iteration. MIN_RES (Minimum Residual) decreases with iteration which is a normal thing. There is almost no variation in the Force quantities with increase in iteration. Dont know whether the calculation is running in right path. How can we check whether the calculation is running in right path. 

August 1, 2012, 21:51 

#20 
Member
venkatesh
Join Date: May 2012
Posts: 93
Rep Power: 6 
I ran the calculation with CFL number = 0.000001. I started the calculation in coarse grid for 200 iteration. the residuals were decreasing slowly. After 200 iteration, I started to run in fine grid but the solution blew up after 213 iteration.
Can you tell me the reason why the solution blew up in finest grid. thanks in advance for suggestions 

Thread Tools  
Display Modes  


Similar Threads  
Thread  Thread Starter  Forum  Replies  Last Post 
Error importing mesh from Hexpress to Fineturbo  cheche  NUMECA  4  January 27, 2011 04:35 
Running FineTurbo on a Linux Cluster  Philipp Höfliger  NUMECA  3  May 6, 2004 08:47 