
[Sponsors] 
December 6, 2011, 12:05 
ERROR #001100279  CFX 13  transient run

#1 
Senior Member
Join Date: May 2011
Location: Germany
Posts: 130
Rep Power: 7 
Hey,
I just started my transient calculation and it run quite good. I calculated 9 setps successfully (with normal residuals) but in my 10th timestep I get the following error after 12 of 15 iterations: Hope you can help me. Could this be, because somebody else started a calculation on the same machine and it run out of memory?  edit: only chance from a previous (running without error) one is Turbulence Numerics : Option : First Order > High Resolution  ++  ERROR #001100279 has occurred in subroutine ErrAction.   Message:   c_fpx_handler: Floating point exception: Invalid Operand            ++ ++  ERROR #001100279 has occurred in subroutine ErrAction.   Message:   Stopped in routine FPX: c_fpx_handler            ++ ++  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. Last edited by Zigainer; December 6, 2011 at 12:24. 

December 6, 2011, 19:47 

#2 
Super Moderator
Glenn Horrocks
Join Date: Mar 2009
Location: Sydney, Australia
Posts: 12,638
Rep Power: 98 
Been asked a million times, I really should write an FAQ on this.
Floating point error means your simulation has diverged. You need to improve numerical stability. To fix it consider smaller time steps, better initial conditions, tighter convergence and lots of other options. 

December 7, 2011, 03:48 

#3 
Senior Member
Join Date: May 2011
Location: Germany
Posts: 130
Rep Power: 7 
Hey,
I can't believe that my calculation diverged. Here is the part of the outfile: ================================================== ====================  Timestepping Information    Timestep  RMS Courant Number  Max Courant Number  ++++  5.0000E02  999.99  999.99   ================================================== ==================== TIME STEP = 10 SIMULATION TIME = 5.0000E01 CPU SECONDS = 5.386E+04  COEFFICIENT LOOP ITERATION = 1 CPU SECONDS = 5.386E+04   Equation  Rate  RMS Res  Max Res  Linear Solution  ++++++  UMom  1.14  1.6E05  5.9E03  1.0E01 ok  VMom  1.17  5.5E05  1.9E02  1.2E01 ok  WMom  1.13  2.6E05  7.6E03  1.4E01 ok  PMass  0.91  1.3E07  1.9E05  10.2 6.0E02 OK ++++++  HEnergy  0.85  3.3E03  1.5E01  3.0E02 OK  TEnergy  1.19  8.2E04  2.0E02  6.2 3.0E02 OK ++++++  KTurbKE  1.19  3.0E05  8.8E03  6.2 6.4E02 OK  OTurbFreq  0.90  5.2E06  2.1E03  7.6 4.6E05 OK ++++++  COEFFICIENT LOOP ITERATION = 2 CPU SECONDS = 5.425E+04   Equation  Rate  RMS Res  Max Res  Linear Solution  ++++++  UMom  0.83  1.4E05  2.9E03  1.1E01 ok  VMom  1.29  7.1E05  1.7E02  1.1E02 OK  WMom  0.88  2.3E05  5.7E03  1.2E01 ok  PMass  2.08  2.6E07  2.7E05  23.9 8.3E02 OK ++++++  HEnergy  0.07  2.4E04  1.3E02  7.9E02 OK  TEnergy  0.30  2.4E04  4.2E03  16.5 7.9E02 OK ++++++  KTurbKE  1.46  4.4E05  3.6E03  11.4 2.9E02 OK  OTurbFreq  9.17  4.7E05  1.9E03  7.6 5.6E05 OK ++++++  COEFFICIENT LOOP ITERATION = 3 CPU SECONDS = 5.472E+04   Equation  Rate  RMS Res  Max Res  Linear Solution  ++++++  UMom  1.32  1.8E05  4.2E03  1.0E01 ok  VMom  0.69  4.9E05  1.1E02  9.4E02 OK  WMom  0.90  2.1E05  3.4E03  1.1E01 ok  PMass  1.23  3.2E07  3.3E05  14.8 6.8E02 OK ++++++  HEnergy  0.88  2.1E04  1.8E02  9.9E02 OK  TEnergy  0.63  1.5E04  3.3E03  6.2 9.9E02 OK ++++++  KTurbKE  0.91  4.0E05  8.2E03  6.2 9.6E02 OK  OTurbFreq  1.05  5.0E05  3.6E03  7.6 9.1E05 OK ++++++  COEFFICIENT LOOP ITERATION = 4 CPU SECONDS = 5.510E+04   Equation  Rate  RMS Res  Max Res  Linear Solution  ++++++  UMom  1.03  1.8E05  4.3E03  1.0E01 ok  VMom  1.11  5.4E05  1.6E02  1.1E01 ok  WMom  1.18  2.4E05  4.8E03  1.2E01 ok  PMass  0.84  2.7E07  3.3E05  10.2 6.9E02 OK ++++++  HEnergy  0.72  1.5E04  2.4E02  8.5E02 OK  TEnergy  0.23  3.5E05  8.9E04  6.2 8.5E02 OK ++++++  KTurbKE  0.73  2.9E05  7.4E03  6.2 8.3E02 OK  OTurbFreq  0.28  1.4E05  1.8E03  7.6 1.1E04 OK ++++++  COEFFICIENT LOOP ITERATION = 5 CPU SECONDS = 5.548E+04   Equation  Rate  RMS Res  Max Res  Linear Solution  ++++++  UMom  0.78  1.4E05  3.0E03  1.0E01 OK  VMom  0.92  5.0E05  1.1E02  1.1E01 ok  WMom  0.93  2.3E05  4.6E03  1.3E01 ok  PMass  0.71  1.9E07  2.5E05  10.2 4.8E02 OK ++++++  HEnergy  0.73  1.1E04  1.8E02  7.6E02 OK  TEnergy  0.57  2.0E05  3.1E04  6.2 7.6E02 OK ++++++  KTurbKE  0.83  2.4E05  6.2E03  6.2 7.5E02 OK  OTurbFreq  0.53  7.6E06  1.2E03  7.6 1.6E04 OK ++++++  COEFFICIENT LOOP ITERATION = 6 CPU SECONDS = 5.585E+04   Equation  Rate  RMS Res  Max Res  Linear Solution  ++++++  UMom  0.91  1.3E05  2.5E03  1.1E01 ok  VMom  0.91  4.6E05  9.7E03  1.3E01 ok  WMom  0.83  1.9E05  3.3E03  1.5E01 ok  PMass  0.83  1.6E07  2.1E05  10.2 5.0E02 OK ++++++  HEnergy  0.83  9.2E05  1.9E02  7.5E02 OK  TEnergy  0.45  8.8E06  2.1E04  6.2 7.5E02 OK ++++++  KTurbKE  0.95  2.3E05  5.5E03  6.2 7.7E02 OK  OTurbFreq  0.75  5.7E06  1.8E03  7.6 2.3E04 OK ++++++  COEFFICIENT LOOP ITERATION = 7 CPU SECONDS = 5.623E+04   Equation  Rate  RMS Res  Max Res  Linear Solution  ++++++  UMom  1.15  1.5E05  4.7E03  1.1E01 ok  VMom  1.02  4.7E05  1.3E02  1.3E01 ok  WMom  1.05  2.0E05  5.1E03  1.4E01 ok  PMass  0.93  1.5E07  1.6E05  10.2 5.8E02 OK ++++++  HEnergy  1.03  9.4E05  2.3E02  7.2E02 OK  TEnergy  0.68  6.0E06  1.7E04  6.2 7.2E02 OK ++++++  KTurbKE  1.25  2.9E05  8.2E03  6.2 7.3E02 OK  OTurbFreq  0.98  5.5E06  2.1E03  7.6 2.7E04 OK ++++++  COEFFICIENT LOOP ITERATION = 8 CPU SECONDS = 5.660E+04   Equation  Rate  RMS Res  Max Res  Linear Solution  ++++++  UMom  1.15  1.7E05  7.5E03  9.9E02 OK  VMom  1.15  5.4E05  2.1E02  1.1E01 ok  WMom  1.21  2.4E05  8.3E03  1.2E01 ok  PMass  0.95  1.4E07  2.0E05  10.2 5.9E02 OK ++++++  HEnergy  1.07  1.0E04  2.4E02  6.9E02 OK  TEnergy  0.86  5.1E06  2.6E04  6.2 6.9E02 OK ++++++  KTurbKE  1.12  3.3E05  8.7E03  6.2 6.5E02 OK  OTurbFreq  0.88  4.9E06  1.7E03  7.6 2.5E04 OK ++++++  COEFFICIENT LOOP ITERATION = 9 CPU SECONDS = 5.698E+04   Equation  Rate  RMS Res  Max Res  Linear Solution  ++++++  UMom  0.95  1.6E05  5.5E03  9.5E02 OK  VMom  0.99  5.3E05  1.5E02  1.0E01 ok  WMom  1.13  2.7E05  7.8E03  1.2E01 ok  PMass  1.02  1.5E07  1.7E05  10.2 4.8E02 OK ++++++  HEnergy  0.90  9.1E05  1.8E02  7.4E02 OK  TEnergy  0.95  4.9E06  3.5E04  6.2 7.4E02 OK ++++++  KTurbKE  0.86  2.8E05  6.4E03  6.2 6.3E02 OK  OTurbFreq  0.79  3.9E06  1.1E03  7.6 2.4E04 OK ++++++  COEFFICIENT LOOP ITERATION = 10 CPU SECONDS = 5.735E+04   Equation  Rate  RMS Res  Max Res  Linear Solution  ++++++  UMom  0.85  1.4E05  2.9E03  1.0E01 ok  VMom  0.90  4.8E05  1.3E02  1.3E01 ok  WMom  0.85  2.3E05  5.6E03  1.4E01 ok  PMass  1.07  1.6E07  1.7E05  10.2 4.4E02 OK ++++++  HEnergy  0.92  8.4E05  1.4E02  9.9E02 OK  TEnergy  0.97  4.7E06  2.5E04  6.2 9.9E02 OK ++++++  KTurbKE  0.84  2.4E05  3.1E03  6.2 6.7E02 OK  OTurbFreq  1.19  4.6E06  1.8E03  7.6 2.4E04 OK ++++++  COEFFICIENT LOOP ITERATION = 11 CPU SECONDS = 5.773E+04   Equation  Rate  RMS Res  Max Res  Linear Solution  ++++++  UMom  1.10  1.5E05  4.1E03  1.0E01 ok  VMom  1.05  5.0E05  1.1E02  1.4E01 ok  WMom  0.79  1.8E05  3.6E03  1.3E01 ok  PMass  0.98  1.5E07  2.1E05  10.2 5.3E02 OK ++++++  HEnergy  1.12  9.4E05  1.6E02  2.9E02 OK  TEnergy  0.96  4.5E06  2.4E04  11.3 2.9E02 OK ++++++  KTurbKE  1.06  2.5E05  5.7E03  6.2 6.8E02 OK  OTurbFreq  1.06  4.9E06  2.0E03  7.6 1.8E04 OK ++++++  COEFFICIENT LOOP ITERATION = 12 CPU SECONDS = 5.810E+04   Equation  Rate  RMS Res  Max Res  Linear Solution  ++++++  UMom  1.01  1.5E05  3.6E03  9.4E02 OK  VMom  1.12  5.6E05  1.2E02  1.2E01 ok  WMom  1.18  2.1E05  4.5E03  1.2E01 ok  PMass  0.90  1.4E07  2.3E05  10.2 5.3E02 OK ++++++  HEnergy  1.08  1.0E04  2.4E02  9.1E02 OK  TEnergy  1.06  4.8E06  2.9E04  6.2 9.1E02 OK ++++++ ++  ERROR #001100279 has occurred in subroutine ErrAction.   Message:   c_fpx_handler: Floating point exception: Invalid Operand            ++ ++  ERROR #001100279 has occurred in subroutine ErrAction.   Message:   Stopped in routine FPX: c_fpx_handler            ++ ++  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. 

December 7, 2011, 06:13 

#4 
New Member
Join Date: Oct 2011
Posts: 6
Rep Power: 6 
Hi Zigainer!
I'm not an expert of transient simulations, not even ANSYS. You should listen to ghorrocks. But I had the same error recently, and there are many good possible solutions on this forum. I advice you to use the search option. Also, according to your last reply, I would decrease the timestep, instead of increasing the coefficient loops. Double precision may can also help, to avoid rounding errors. Another possible solution of this problem is to start your transient simulation from a converged steady state solution. 

December 7, 2011, 06:37 

#5 
Senior Member
Join Date: May 2011
Location: Germany
Posts: 130
Rep Power: 7 
Hey,
my calculation starts from a converged stead state calculation with all data from the first timestep. And again, I had a calculation (same setup) which was running. The only change was in "solver control" > Turbulence Numerics > from First Order to High Resolution. And I can't believe that this small change destroyed my convergence. 

December 7, 2011, 18:23 

#6  
Super Moderator
Glenn Horrocks
Join Date: Mar 2009
Location: Sydney, Australia
Posts: 12,638
Rep Power: 98 
Quote:
And in this case you seem to be using a massive time step so the first thing I would try is running a time step far smaller than you currently. 

December 7, 2011, 18:55 

#7 
Senior Member
Join Date: May 2011
Location: Germany
Posts: 130
Rep Power: 7 
Hey,
I tried a timestep of 0.01 (one fifth of original one) and it also interrupted. and again the only change to a running one is turbulence numerics. And if I watch the residual I can't see a divergence sorry..... residual are getting smaller and smaller and then the error occurce 

December 7, 2011, 21:12 

#8 
Super Moderator
Glenn Horrocks
Join Date: Mar 2009
Location: Sydney, Australia
Posts: 12,638
Rep Power: 98 
You need a much smaller timestep than that.
Floating point error divergences do not always show up in residuals. 

December 8, 2011, 06:58 

#9 
Senior Member
Join Date: May 2011
Location: Germany
Posts: 130
Rep Power: 7 
Alright thanks.
I actually do not have enough time for a smaller timestep. The calculation for the 0.01s timestep would run almost 1 month. So a smaller timestep is not really possible. I stick to the First Order Turbulence Numerics. 

December 8, 2011, 07:11 

#10 
Super Moderator
Glenn Horrocks
Join Date: Mar 2009
Location: Sydney, Australia
Posts: 12,638
Rep Power: 98 
Do you want it to run or not? Up to you.
You will probably be able to use some small timesteps to start it off and increase them from there. Adaptive time steps (adapting to 35 coeff loops per iteration) is very useful here as it does it automatically. 

December 8, 2011, 07:18 

#11 
Senior Member
Join Date: May 2011
Location: Germany
Posts: 130
Rep Power: 7 
Meanwhile I've got some changes (inlet channel before the ribbed channel) and with these simulation is no problem. Even with the "large" timestep of 0.05s
But thanks for your help!!! 

December 8, 2011, 07:31 

#12 
Super Moderator
Glenn Horrocks
Join Date: Mar 2009
Location: Sydney, Australia
Posts: 12,638
Rep Power: 98 
Why didn't you say that before? So it is more likely then you have a setup problem with this simulation but not the other.
Can you post an image of both geometries? 

December 8, 2011, 08:38 

#13 
Senior Member
Join Date: May 2011
Location: Germany
Posts: 130
Rep Power: 7 
The ribbed channel part is the same for both. The new calculation just got an extra unribbedinletchannel.
The calculation which is not running doesn't use a velocity and temperatureprofile at the inlet. edit: btw its a conjugated calculation 

December 8, 2011, 17:21 

#14 
Super Moderator
Glenn Horrocks
Join Date: Mar 2009
Location: Sydney, Australia
Posts: 12,638
Rep Power: 98 
If the longer simulation runs OK then it is likely to be something wrong with your setup of the shorter one. If you also changed the inlet BC type between runs then that is likely to be the problem.


December 9, 2011, 14:22 

#15 
Senior Member
Join Date: May 2011
Location: Germany
Posts: 130
Rep Power: 7 
The INLET condition is the same for both simualtions.
The new longer calcualtion is running, so tahts fine for me. I don't need the shorter one, anyway. There've been an other problem with velocityprofil anyway. Thanks for your help. 

Thread Tools  
Display Modes  


Similar Threads  
Thread  Thread Starter  Forum  Replies  Last Post 
Time average results in Transient CFX Simulation  BalanceChen  CFX  28  September 28, 2016 02:11 
courant number for transient flow on CFX ?  amine  CFX  13  August 20, 2013 19:02 
Can workbench and cfx run correctly on vista?  Sandam  CFX  1  December 2, 2007 17:32 
mesh deformation in transient run  Ichi  CFX  1  July 27, 2005 18:26 
run Build Topology in IcemCFD 4 CFX  Ahmed  CFX  3  March 19, 2004 17:59 