Discussion Groups

Simulation Moldflow Insight

Reply
Contributor
petr_AL
Posts: 17
Registered: ‎02-23-2011
Accepted Solution

Fiber orientation analysis computing problem

286 Views, 7 Replies
05-30-2014 02:51 AM

Hi,

 

do you have somebody similar problem?

 

We computed two Moldflow studies using fiber filled material. The difference between studies were only number and positions of injection points - see image.
First study was computed completely (Fill+pack+warp).
Second study computing "freezed" on 44% of compute - during filling phase (about 88% of filled volume) - no error in log file, but computing stay.
Time to compute first study was about 5hrs, for second study we wait about 24hrs, still the same status. I stoped then computing using Job manager.
Both log files see as attachments.

 

I had the same problem some time ago with other part using also fiber material (PBT GF30). One study was computed succesfully, second using
different gate locations "freezed". Here I changed Fiber Solver parameters and also second study was computed.

Now I tried also several Fiber Solver settings, but study with two injection points was not computed completely any time. It stoped on the same filling step.

 

Thank you for response to this issue

 

Petr

Please use plain text.
Distinguished Contributor
mayur_dhumal
Posts: 165
Registered: ‎05-05-2011

Re: Fiber orientation analysis computing problem

05-30-2014 04:40 AM in reply to: petr_AL
Hi Petr,

Right now i don't have any solution on this, but i have also came across this issue previously.
I was having defected mesh, it was not repaired, so once i got the solve complete, but it failed for every run after that. I am trying to find out the exact reason behind it.

Thanks,
Mayur
Please use plain text.
Employee
Alex.Bakharev
Posts: 47
Registered: ‎04-10-2013

Re: Fiber orientation analysis computing problem

05-30-2014 05:35 AM in reply to: petr_AL

There was a 3D fiber orientation bug ithat in some rare circumstances led to hanging like this. It was fixed in 2015 (might be in some 2014 service pack as well). 

 

Try to upgrade to 2015, if you can. If not I will check what service pack of 2014 fixes the problem

Please use plain text.
Employee
Alex.Bakharev
Posts: 47
Registered: ‎04-10-2013

Re: Fiber orientation analysis computing problem

05-30-2014 05:39 AM in reply to: Alex.Bakharev

If you could send me the model, I would test whether it works in 2015

Please use plain text.
Contributor
petr_AL
Posts: 17
Registered: ‎02-23-2011

Re: Fiber orientation analysis computing problem

05-30-2014 05:50 AM in reply to: Alex.Bakharev
Hi Alex,

thank you for you response. We will upgrade to 2015 in 2 weeks latest. I wil inform you about computed results in 2015.
Now I try to make new mesh and compute new studis.
Do you think, if runner system as beam elements would be used, the computing would be better?

Thanks, Petr
Please use plain text.
Employee
Alex.Bakharev
Posts: 47
Registered: ‎04-10-2013

Re: Fiber orientation analysis computing problem

06-01-2014 06:58 PM in reply to: petr_AL

Hi, Peter

 

If this is the 2014 bug I am thinking of then it has the tendency to disappear with slight changes in the simulation setting - changing the time step, remeshing, etc. might make the simulation to run normally. I guess modelling runners as beams might make the bug to hide.

 

Usually it is a good idea to model long runners as beams because it provides better resolution across the runners.

Please use plain text.
Employee
Alex.Bakharev
Posts: 47
Registered: ‎04-10-2013

Re: Fiber orientation analysis computing problem

06-01-2014 07:02 PM in reply to: petr_AL

BTW, the bug and the hot fix are described by Stefan at http://autodesk.typepad.com/beyondmoldflowinsight/2013/09/new-moldflow-hotfix-for-autodesk-simulatio... . If this is the result of this bug applying the hot fix would solve the problem (upgrading to 2015 would solve it too).

Please use plain text.
Contributor
petr_AL
Posts: 17
Registered: ‎02-23-2011

Re: Fiber orientation analysis computing problem

06-03-2014 06:24 AM in reply to: Alex.Bakharev

Hi Alex,

 

thank you for your answers.

 

On last Friday I prepared new mesh using slightly lower Global edge length, both gate variations were computed OK incl. Fiber orintation analysis.

Yesterday I prepared two studies with original mesh, but using beam elements. Two injection points variation was computed OK now, but study with 3 gates freezed during computing....

 

The problem is in 2014 solver, as you wrote, I think.

 

We have not installed the hot fix for 2014. Next week we will install 2015, I will compute then the freezed study in 2015 to compare it.

 

Thanks

 

Petr

Please use plain text.