Sudden crash

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

October 3, 2013, 12:20
Sudden crash
#1
Member

Join Date: Jul 2010
Posts: 40
Rep Power: 9
Hi everyone;

I am having some problems with my simulations. They crashed all of a sudden. Usually the iterations of k increase (sometime is the velocity), then deltaT becomes really small and a couple time steps after the sim crashes. This is an example of the output:

Code:
```Time = 56.3014205783

Courant Number mean: 0.118327767573 max: 0.19896182766
deltaT = 0.00153197241597
DILUPBiCG:  Solving for Ux, Initial residual = 0.00174185441499, Final residual = 4.59171443112e-09, No Iterations 3
DILUPBiCG:  Solving for Uy, Initial residual = 0.0180925560235, Final residual = 5.51828116172e-08, No Iterations 3
DILUPBiCG:  Solving for Uz, Initial residual = 0.0208771839438, Final residual = 5.52651987859e-08, No Iterations 3
GAMG:  Solving for p, Initial residual = 0.0786783298856, Final residual = 5.28569419341e-06, No Iterations 9
time step continuity errors : sum local = 2.29595919923e-10, global = 1.15432066317e-19, cumulative = -8.84673137504e-16
GAMG:  Solving for p, Initial residual = 0.00333261520514, Final residual = 9.92201073873e-07, No Iterations 7
time step continuity errors : sum local = 4.30705321016e-11, global = -1.31772997273e-18, cumulative = -8.85990867476e-16
GAMG:  Solving for p, Initial residual = 0.000138442060363, Final residual = 7.07661072868e-07, No Iterations 6
time step continuity errors : sum local = 3.53898112737e-11, global = -7.05569157603e-19, cumulative = -8.86696436634e-16
DILUPBiCG:  Solving for omega, Initial residual = 7.00026119366e-07, Final residual = 3.58270229366e-09, No Iterations 1
DILUPBiCG:  Solving for k, Initial residual = 0.00115615362982, Final residual = 2.81404703888e-09, No Iterations 3
bounding k, min: 0 max: 0.586278690722 average: 0.0996495165825
ExecutionTime = 75598.44 s  ClockTime = 75928 s

Calculating averages

Time = 56.3029525507

Courant Number mean: 0.118847270725 max: 0.198956109975
deltaT = 0.0015387360911
DILUPBiCG:  Solving for Ux, Initial residual = 0.00174945388057, Final residual = 4.68860801279e-09, No Iterations 3
DILUPBiCG:  Solving for Uy, Initial residual = 0.018172002406, Final residual = 5.60114718505e-08, No Iterations 3
DILUPBiCG:  Solving for Uz, Initial residual = 0.0209706723416, Final residual = 5.64116133548e-08, No Iterations 3
GAMG:  Solving for p, Initial residual = 0.0789601498487, Final residual = 6.87653606352e-06, No Iterations 9
time step continuity errors : sum local = 3.17562391533e-10, global = 3.71130246033e-19, cumulative = -8.86325306388e-16
GAMG:  Solving for p, Initial residual = 0.00338734796122, Final residual = 9.66446829089e-07, No Iterations 31
time step continuity errors : sum local = 5.06496323068e-11, global = -5.9961362602e-19, cumulative = -8.86924920014e-16
GAMG:  Solving for p, Initial residual = 0.000140735396748, Final residual = 6.27831105034e-07, No Iterations 5
time step continuity errors : sum local = 2.65259351068e-11, global = -2.82758134623e-19, cumulative = -8.87207678149e-16
DILUPBiCG:  Solving for omega, Initial residual = 7.02406726747e-07, Final residual = 3.61836938681e-09, No Iterations 1
DILUPBiCG:  Solving for k, Initial residual = 0.00116106584398, Final residual = 166169.798741, No Iterations 1001
bounding k, min: -312205129.035 max: 269273680.127 average: -10606.0281394
ExecutionTime = 75608.43 s  ClockTime = 75938 s

Calculating averages

Time = 56.3044912868

Courant Number mean: 0.119372303823 max: 0.198738491473
deltaT = 0.00154557491817
DILUPBiCG:  Solving for Ux, Initial residual = 0.994347884301, Final residual = 9.44225468604e-08, No Iterations 189
DILUPBiCG:  Solving for Uy, Initial residual = 0.908777437471, Final residual = 9.09721357325e-08, No Iterations 201
DILUPBiCG:  Solving for Uz, Initial residual = 0.918085844451, Final residual = 8.23188861914e-08, No Iterations 266
GAMG:  Solving for p, Initial residual = 0.999894333283, Final residual = 9.98918535975e-05, No Iterations 534
time step continuity errors : sum local = 5.88051660082e-06, global = 1.07090381147e-16, cumulative = -7.80117297001e-16
GAMG:  Solving for p, Initial residual = 0.600725392574, Final residual = 5.96105999062e-05, No Iterations 429
time step continuity errors : sum local = 7.48767985205e-06, global = 7.86420631943e-17, cumulative = -7.01475233807e-16
GAMG:  Solving for p, Initial residual = 0.191758582659, Final residual = 9.96774342026e-07, No Iterations 817
time step continuity errors : sum local = 1.22807359407e-07, global = -4.48319516578e-16, cumulative = -1.14979475038e-15
DILUPBiCG:  Solving for omega, Initial residual = 0.908220760948, Final residual = 3.49400568529, No Iterations 1001
bounding omega, min: -8784737312 max: 13428850393.2 average: 5500413.94033
DILUPBiCG:  Solving for k, Initial residual = 0.999995075281, Final residual = 0.00454874595929, No Iterations 1001
bounding k, min: -1.41020644379e+12 max: 2.66411047567e+12 average: 111889394.608
ExecutionTime = 75668.23 s  ClockTime = 75998 s```
The most puzzling thing is that I cannot reproduce the "crash", to save it and visualize it for example. If I rerun the simulation without any change (starting at the last saved time), it runs pass the previous crashing point.

I have tried changing pretty much every option in the fvSolution and fvSchemes. But I have not been able to identify the problem. Any suggestions or ideas?

- Simulations of a simplified atmospheric boundary layer so it is basically a channel flow.
- I am using a hybrid model (k-w SST-IDDES).
- There is nothing fancy about the mesh. Rectangle with rectangular cells, that are really elongated close to the wall

Thanks a lot,

Mary
Attached Files
 system.zip (3.7 KB, 4 views)

 October 4, 2013, 02:32 #2 Senior Member   Bernhard Join Date: Sep 2009 Location: Delft Posts: 790 Rep Power: 14 Two remarks: - Try to reduce the maximum Courant number. I've commonly seen these kind of errors for to large time steps. - Your "writeFormat" is probably set to "ascii". This causes the the number behind the decimal point are truncated on re-read, does giving a (slightly) different initial condition. Changing this to binary will solve this issue.

 October 4, 2013, 06:56 #3 Member   Join Date: Jul 2010 Posts: 40 Rep Power: 9 Bernhard, Thanks for your reply. I am using a max CFL of 0.2 and writing in binary in all my simulations. And I still have that issue. Any other ideas? Thanks again...

October 6, 2013, 14:13
#4
Super Moderator

Bruno Santos
Join Date: Mar 2009
Location: Lisbon, Portugal
Posts: 9,978
Blog Entries: 39
Rep Power: 108
Greetings to all!

@Mary:
Quote:
 Originally Posted by MaryBau The most puzzling thing is that I cannot reproduce the "crash", to save it and visualize it for example.
This could mean that your machine is having a serious hardware problem, possibly CPU or RAM related. Check the CPU temperature while the simulation is running, just in case it's related to overheating.

Quote:
 Originally Posted by MaryBau If I rerun the simulation without any change (starting at the last saved time), it runs pass the previous crashing point.
That's actually normal to occur in CFD, when there is some very small numerical instability. What you are observing is actually one of the common workarounds used, for when similar situations occur. One can only wish were so lucky that this was possible every time such crashes occurred.

Quote:
 Originally Posted by MaryBau Rectangle with rectangular cells, that are really elongated close to the wall
This could be the source of the numerical instability.

Best regards,
Bruno
__________________

 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 vainilreb OpenFOAM Running, Solving & CFD 23 August 20, 2013 15:09 cbritan OpenFOAM Bugs 3 January 6, 2011 04:58 ruzfactor Main CFD Forum 0 November 26, 2010 19:21 HervĂ© CFX 2 June 16, 2008 06:40 adarsh Main CFD Forum 0 June 11, 2002 12:08