CFD Online Logo CFD Online URL
www.cfd-online.com
[Sponsors]
Home > Forums > CFX

2 Way FSI Problem

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

Reply
 
LinkBack Thread Tools Display Modes
Old   September 16, 2010, 10:23
Default 2 Way FSI Problem
  #1
New Member
 
Ali Bas
Join Date: Sep 2010
Location: İstanbul, Turkey
Posts: 15
Rep Power: 6
alialibas is on a distinguished road
Hi all,

I have a problem with building a 2-Way FSI case. First of all, i would like to use both heat transfer and mesh motion in the same run. I am using CFX for fluid solver and ANSYS for structural side. I am preparing a mechanical ".inp" file for Multifield Analysis. Mechanical input file is built by multiphysics, i mean it can solve both thermal and structural. The element type is Solid226. I prepared the inp file in workbench and embeded a command to change the element type. Because in workbench, Transient Structural analysis system give aoutomatically Solid186 or similar..

Now, the case is running, i dont have a problem with it. But the problem is in this case, only mesh deformation is transfering between CFX and ANSYS. But i need thermal stresses.

CFX needs an "interface" boundary domain between solid and fluid domains to be able to run a heat transfer problem. But on the other hand, for FSI problem, i need a "wall" boundary domain to be able to transfer the mesh motion(especially) and heat transfer. For heat transfer i am trying to send to Ansys "Wall Heat Flow", and receiving "Temperature" from Ansys. and for mesh motion;from CFX, i am trying to send "Total Force" to Ansys and receiving from Ansys "Total Mesh Displacement". But as it seems that only mesh motion transfer works..!



Also i am receiving a note from Ansys output:


************************************************** ***************************
*** NOTE *** CP = 11106.921 TIME= 17:49:12
New Solution Control Option is only valid for a single field structural
or thermal analysis. The SOLCONTROL,ON command has been de-activated.
************************************************** ***************************


But at the same run, i can see this also in Ansys run:
At the bottom line, it says about HT Flow Convergence :s by the way thermal dofs are ON! what does it mean??? but why i cannot see any temp in the solid body!

************************************************** ************************
L O A D S T E P O P T I O N S

LOAD STEP NUMBER. . . . . . . . . . . . . . . . 29
TIME AT END OF THE LOAD STEP. . . . . . . . . . 0.29000E-02
NUMBER OF SUBSTEPS. . . . . . . . . . . . . . . 1
MAXIMUM NUMBER OF EQUILIBRIUM ITERATIONS. . . . 25
STEP CHANGE BOUNDARY CONDITIONS . . . . . . . . YES
TRANSIENT (INERTIA) EFFECTS
STRUCTURAL DOFS. . . . . . . . . . . . . . . ON
THERMAL DOFS . . . . . . . . . . . . . . . . ON
TRANSIENT INTEGRATION PARAMETERS
GAMMA. . . . . . . . . . . . . . . . . . . . 0.10000
ALPHA. . . . . . . . . . . . . . . . . . . . 0.30250
DELTA. . . . . . . . . . . . . . . . . . . . 0.60000
USING HHT TIME INTEGRATION
ALPHAF . . . . . . . . . . . . . . . . . . . 0.10000
ALPHAM . . . . . . . . . . . . . . . . . . . 0.0000
THETA. . . . . . . . . . . . . . . . . . . . 0.50000
OSCILLATION LIMIT CRITERION. . . . . . . . . 0.50000
TOLERANCE. . . . . . . . . . . . . . . . . . 0.20000
TERMINATE ANALYSIS IF NOT CONVERGED . . . . . .YES (EXIT)
CONVERGENCE CONTROLS. . . . . . . . . . . . . .USE DEFAULTS
PRINT OUTPUT CONTROLS . . . . . . . . . . . . .NO PRINTOUT
DATABASE OUTPUT CONTROLS
ITEM FREQUENCY COMPONENT
ALL NONE
NSOL ALL
RSOL ALL
NLOA ALL
STRS ALL
EPEL ALL
EPPL ALL
V ALL
A ALL

FORCE CONVERGENCE VALUE = 0.2163E-12 CRITERION= 0.1752
HT FLOW CONVERGENCE VALUE= 0.8834E-14 CRITERION= 0.1000E-08
EQUIL ITER 1 COMPLETED. NEW TRIANG MATRIX. MAX DOF INC= -0.3626E-17
FORCE CONVERGENCE VALUE = 0.9861E-13 CRITERION= 0.1752 <<< CONVERGED
HT FLOW CONVERGENCE VALUE= 0.8506E-14 CRITERION= 0.1000E-08 <<< CONVERGED
************************************************** *******************************


What should i do for both mesh motion transfer and heat transfer in a FSI problem?
alialibas is offline   Reply With Quote

Old   September 16, 2010, 13:34
Default
  #2
Senior Member
 
Join Date: Apr 2009
Posts: 511
Rep Power: 12
stumpy is on a distinguished road
In CFX you should have a wall boundary condition, with the mesh motion settings set to send forces and receive diplacements, and the heat transfer settings set to send heat flows and receive temperature. The CFX setup should not include any solid domains that are included in ANSYS (i.e. you cannot solve a CHT case in CFX and send the interface thermal data to ANSYS when using 2-way FSI), so you shouldn't have any interface boundaries in CFX that need to send data to ANSYS.
stumpy is offline   Reply With Quote

Old   September 16, 2010, 13:52
Default
  #3
New Member
 
Ali Bas
Join Date: Sep 2010
Location: İstanbul, Turkey
Posts: 15
Rep Power: 6
alialibas is on a distinguished road
Actually, i am also trying to solve heat transfer in Ansys side. i did exactly what you said. but problem is ansys didnt solve the heat transfer...
alialibas is offline   Reply With Quote

Old   December 30, 2010, 22:54
Default
  #4
New Member
 
abdul khader
Join Date: Dec 2010
Posts: 19
Rep Power: 6
smak.quadri is on a distinguished road
Salam Ali bas,

I have a doubt in FSI related to ANSYS MFX FSI b/w CFX and structural.

I tried to couple wit ANSYS Structure and CFX but I have some problems in hand. I am working of FSI of flow through flexible tube.

Can u tel me, how did u transfer ansys structural file to workbench...?

I have no problem wit Fluid, but for structure, I tried to mesh in gambit and bring it to ansys classic, applied boundary conditions. but stuck up wit writing input file. (.inp). I tried using saveas option and created it and then linked wit CFX, but during solution its not recognising.

Can u tel me what is the procedure in transfering mesh from clasic to workbench and method of creating inp file in classic.
smak.quadri is offline   Reply With Quote

Old   December 31, 2010, 07:56
Default
  #5
New Member
 
Ali Bas
Join Date: Sep 2010
Location: İstanbul, Turkey
Posts: 15
Rep Power: 6
alialibas is on a distinguished road
Hi Smak,

First of all, i use only workbench modules for fsi cases. it is a little bit different in classic mode. to export a input file in classic, in file tab there may be write input file tab. and in in workbench, ,n the transient or static structural analysis, in the tools tab there is "write input file" you can prepare all your case in workbench and get input file then use it inside the cfx module...
alialibas is offline   Reply With Quote

Old   February 2, 2011, 09:28
Default Problem in FSI using CFX-ANSYS
  #6
New Member
 
abdul khader
Join Date: Dec 2010
Posts: 19
Rep Power: 6
smak.quadri is on a distinguished road
the problem of FSI using CFX and ANSYS classic is solved.

But there is another hurdle popping up.

The CFX solves in cylindrical coordinate system, when the results are viewed in CFX post for structure, its showing in cartesian system, May i know is there any way of solving this problem.

Also I tried to see the structural results in ANSYS classic, but there also its in cartesian only, when I tried to view in cylindrical also,, the results are shown in cartesian only.
smak.quadri is offline   Reply With Quote

Old   April 5, 2011, 08:08
Default
  #7
New Member
 
Felipe Langellotti
Join Date: Jan 2011
Location: São Paulo, Brazil
Posts: 3
Rep Power: 6
felipelang is on a distinguished road
Quote:
Originally Posted by smak.quadri View Post
the problem of FSI using CFX and ANSYS classic is solved.
How did you generate the ansys .inp file for CFX from Ansys Classic?

My input file generated from Workbench is very different from that one of Classical generates. When I try to run the coupled analysis from CFX, it says that:

"No boundary condition that is compatible with the following data transfer definition could be identified: "

However, in Ansys Classic, before I generate the .inp file, I set the FSIN flag....

Thanks in advance
felipelang is offline   Reply With Quote

Old   December 8, 2011, 04:44
Default Problem in transfering temperature data between two domains
  #8
Member
 
Mina
Join Date: Apr 2011
Posts: 88
Rep Power: 6
Mina_Shahi is on a distinguished road
Hello every body

I am also modeling 2-way FSI with Workbench version 13.
i want to transfer temperature data from CFX to Ansys, so i need to define non adibatic boundary condition in CFX for walls to active heat transfer between inner flow and the CFD walls, i did that and i selected heta transfer coefficient, so it asks two things: heat transfer coefficient and outside temperature , now my problem is setting the outside temperature because outside temperature is an unknow and actually the temperature of inner flow which we have to define. Untitled.jpg
Mina_Shahi is offline   Reply With Quote

Old   June 18, 2012, 13:01
Default
  #9
Senior Member
 
Saima
Join Date: Apr 2009
Location: Canada
Posts: 140
Rep Power: 8
Saima is on a distinguished road
Hello Guys,

I am doing stress analysis for a compressor blade.

I have done flow analysis in CFX and have result file. Now wants to do stress analysis in Ansys classic. How can i import pressure loading through a CGX result file?

I know it is easy to do i workbench but dur to some conditions i have to import in ansys classic. Pls let me know how can it possible?

Regards,



Quote:
Originally Posted by Mina_Shahi View Post
Hello every body

I am also modeling 2-way FSI with Workbench version 13.
i want to transfer temperature data from CFX to Ansys, so i need to define non adibatic boundary condition in CFX for walls to active heat transfer between inner flow and the CFD walls, i did that and i selected heta transfer coefficient, so it asks two things: heat transfer coefficient and outside temperature , now my problem is setting the outside temperature because outside temperature is an unknow and actually the temperature of inner flow which we have to define. Attachment 10366
__________________
Best Redards,
Saima
Saima is offline   Reply With Quote

Old   December 11, 2012, 10:48
Default
  #10
Senior Member
 
Govindaraju
Join Date: Apr 2010
Posts: 206
Rep Power: 8
kmgraju is on a distinguished road
Quote:
Originally Posted by stumpy View Post
In CFX you should have a wall boundary condition, with the mesh motion settings set to send forces and receive diplacements, and the heat transfer settings set to send heat flows and receive temperature. The CFX setup should not include any solid domains that are included in ANSYS (i.e. you cannot solve a CHT case in CFX and send the interface thermal data to ANSYS when using 2-way FSI), so you shouldn't have any interface boundaries in CFX that need to send data to ANSYS.
dear friend
I am doing two way fsi. My structural model has a fixed temperature. I have to send heat to the fluid domain and receive heat from fluid to solid.
similarly, to send forces and displacements.
please clarify the way of setting in the ansys workbench which has been attached for your reference.

Thank you so much for your timely advice

Regards

Govind
Attached Images
File Type: png Untitled1.png (35.1 KB, 38 views)
kmgraju is offline   Reply With Quote

Old   December 12, 2013, 07:30
Default how to rectify this
  #11
New Member
 
prashant gour
Join Date: Dec 2013
Posts: 1
Rep Power: 0
PRASHANT GOUR is on a distinguished road
Run mode: serial run

Host computer: HTML2-PC (PID:4180)
Job started: Thu Dec 12 16:57:45 2013

Connecting to the following master process:
Host Name : HTML2-PC
Port Number : 60288

+--------------------------------------------------------------------+
| ERROR #002100049 has occurred in subroutine CPLG_SetupBcp. |
| Message: |
| No boundary condition that is compatible with the following data |
| transfer definition could be identified: |
| |
| Sending process : ANSYS |
| Sending region : 1 |
| Sending variable : DISP |
| Receiving process : CFX |
| Receiving region : interface1 |
| Receiving variable: Total Mesh Displacement |
| |
| Please review the ANSYS CFX boundary condition and external data |
| transfer definitions. |
+--------------------------------------------------------------------+

+--------------------------------------------------------------------+
| ERROR #001100279 has occurred in subroutine ErrAction. |
| Message: |
| Failed in solver setup synchronization point. |
| |
| |
| |
| |
| |
+--------------------------------------------------------------------+

+--------------------------------------------------------------------+
| ERROR #001100279 has occurred in subroutine ErrAction. |
| Message: |
| CFX encountered the error: |
| -11000. |
| |
| |
| |
| |
+--------------------------------------------------------------------+

+--------------------------------------------------------------------+
| ERROR #001100279 has occurred in subroutine ErrAction. |
| Message: |
| Stopped in routine SYNP_SOLVERSETUP |
| |
| |
| |
| |
| |
+--------------------------------------------------------------------+

+--------------------------------------------------------------------+
| An error has occurred in cfx5solve: |
| |
| The ANSYS CFX solver exited with return code 1. No results file |
| has been created. |
+--------------------------------------------------------------------+

End of solution stage.

+--------------------------------------------------------------------+
| An error has occurred in cfx5solve: |
| |
| ANSYS Solver terminated with return code 3840 |
+--------------------------------------------------------------------+


+--------------------------------------------------------------------+
| The results from this run of the ANSYS solver have been written to |
| C:\Users\html2\Desktop\abcd_pending_tasks\dp0_CFX_ Solution\Fluid |
| Flow CFX_008.ansys |
+--------------------------------------------------------------------+


This run of the ANSYS CFX Solver has finished.
PRASHANT GOUR is offline   Reply With Quote

Old   April 29, 2014, 07:50
Default 2 Way FSI runs only with maximum 2 cores in ANSYS v14.5
  #12
New Member
 
David
Join Date: Aug 2013
Posts: 27
Rep Power: 3
mrkmrk is on a distinguished road
Hi all,

I have a problem with my total solution time,

I am modelling a flexible artery in Ansys v14.5 by an eight core system (3.4 GHz) with 8 GB ram, with 2300 elements for solid part and 94000 elements for fluid part and laminar flow, my total solution time is about 2 days, my first question is that is it normal? And my second question is when I hold mouse cursor on Ansys mechanical APDL Batch running window it shows that the solver is running with 2 cores, I changed every solver settings to 8 cores, but it does not help, what should I do?

Hope to see a response.

Many thanks in advance
mrkmrk is offline   Reply With Quote

Old   April 29, 2014, 09:50
Default
  #13
New Member
 
Ali Bas
Join Date: Sep 2010
Location: İstanbul, Turkey
Posts: 15
Rep Power: 6
alialibas is on a distinguished road
Hi mrkmrk,

I think, run time is not proper. But are you sure the problem is about ANSYS FEA? or the problem is on the CFD side? First of all, you need to check where the problem is.

I think, your CFD time increment is very small since ANSYS tries to get every increment for stagger loops.. IF Your CFD analysis is laminar flow, you should check time increment first, then decrease the stagger loops.

Hope this helps

Ali
alialibas is offline   Reply With Quote

Old   April 30, 2014, 02:59
Exclamation Hi,
  #14
New Member
 
Naveen B
Join Date: Feb 2014
Location: Bangalore, Shimoga
Posts: 2
Rep Power: 0
naveenmyself is on a distinguished road
i am doing CFD analysis of Turbocharger Turbine for Producer gas Engines, can anybody please give me the hints regarding about Turbine Performance characters like shaft power, then power coefficient please.
please mail me on naveenmyself@ymail.com
naveenmyself is offline   Reply With Quote

Old   April 30, 2014, 03:20
Default
  #15
New Member
 
Ali Bas
Join Date: Sep 2010
Location: İstanbul, Turkey
Posts: 15
Rep Power: 6
alialibas is on a distinguished road
Hi prashant gour,

Did you define FSI surfaces for both ansys and cfx side?

I think, cfx cannot transfer displacement data to ansys in regard to below error;

""No boundary condition that is compatible with the following data |
| transfer definition could be identified:""

Thanks,
Ali
alialibas is offline   Reply With Quote

Old   May 1, 2014, 14:58
Default
  #16
New Member
 
David
Join Date: Aug 2013
Posts: 27
Rep Power: 3
mrkmrk is on a distinguished road
Hi Ali,
Thanks for your attention; I really appreciate your help.


The time step in my simulation is 0.01 and the total time (end time) is 8 sec. The convergence in any part of solid or fluid is very quick, whereas (I think) the problem is the interface convergence. In ‘solver control external coupling’ section,’ the relaxation factor’ is set to 0.5 (more values lead to divergence) while ‘the convergence target’ is set to 0.001. Is there any way to run the Mechanical APDL with more than 2 cores?"
Thanks in advance.
mrkmrk is offline   Reply With Quote

Old   May 5, 2014, 08:34
Default
  #17
Senior Member
 
Matthias Voß
Join Date: Mar 2009
Location: Berlin, Germany
Posts: 446
Rep Power: 11
mvoss is on a distinguished road
I am not sure if you really gain speed up when running in parallel with 2300 elements for solid part and 94000 elements for fluid part because the information transfer for the parallel run is more than the actual element number per core. The transfer is eating up the speedup. Depending on total time/timestep and number of stagger iterations per timestep 2 days sounds not that wrong. Anyways, you should definitely run not more cores then the system has in total (means 8 cores = 6 cfd + 2 csm instead of 8+8 or 8+2)
mvoss is offline   Reply With Quote

Old   October 31, 2014, 03:04
Default
  #18
New Member
 
Ali Bas
Join Date: Sep 2010
Location: İstanbul, Turkey
Posts: 15
Rep Power: 6
alialibas is on a distinguished road
Hi mrkmrk,

You can increase Ansys core number by pop-up window on cfx side. The window i say is all the settings you are doing. Look more close to tabs. I could not remember exact place now.

Thanks,
Ali
alialibas is offline   Reply With Quote

Old   October 31, 2014, 03:07
Default
  #19
New Member
 
Ali Bas
Join Date: Sep 2010
Location: İstanbul, Turkey
Posts: 15
Rep Power: 6
alialibas is on a distinguished road
mrkmrk,

For data transfer by FSI interfaces propoerly, you may need to have similar mesh size on both size. I recommend this also.

Thanks
alialibas is offline   Reply With Quote

Old   October 31, 2014, 05:22
Default
  #20
New Member
 
David
Join Date: Aug 2013
Posts: 27
Rep Power: 3
mrkmrk is on a distinguished road
Hi Ali, Matthias

Thank you for your useful responses,

The problem was solved,

Good luck.
mrkmrk is offline   Reply With Quote

Reply

Thread Tools
Display Modes

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 Off
Trackbacks are On
Pingbacks are On
Refbacks are On


Similar Threads
Thread Thread Starter Forum Replies Last Post
FSI nonlinear problem mead4454 CFX 0 March 1, 2010 06:24
what's rhe problem when CFX solve the FSI justdo CFX 2 February 26, 2010 03:58
FSI MODELING PROBLEM (pinball)!!! smn CFX 1 September 16, 2009 00:33
Negative volume problem in Two-way FSI coupling fred CFX 3 August 16, 2006 10:03
Is this problem well posed? Thomas P. Abraham Main CFD Forum 5 September 8, 1999 14:52


All times are GMT -4. The time now is 04:00.