yade-users team mailing list archive
Mailing list archive
Re: About the moment calculation in cohesion-friction model
I am so sorry. You are right.This Law should have no problem. I just tried it, it works. In the previous modelling, I changed the triaxial test files, so got errors when the moment function was opened. But if the moment function was closed, the modified files can run normally.
I should check it carefully. I am appreciate for your help.
At 2010-10-12 10:19:54，"焦利青" <jiaoliqing@xxxxxxx> wrote:
Thank you for your reply.
You are right. If the moment function is closed, this law is no problem. But if the user need to consider the moment, it will occur the errors using this law. In the file, the code showed "if (momentRotationLaw && (!currentContactPhysics->cohesionBroken || always_use_moment_law))" to switch the moment function. The "momentRotationLaw" can be set as True or False in Law2_ScGeom_CohFrictPhys_ElasticPlastic. But the "currentContactPhysics->cohesionBroken" just transfers from other files. Although the user can give True to always_use_moment_law, the program still cannot run normally and output errors. On the other hand, if the "currentContactPhysics->cohesionBroken" is forced as True, the program still output errors. In the triaxial test, this variable has already been set as False before the Law function run. So maybe that's the reason that you cannot find the problem immediately.
In the past several weeks, I engaged on the post-process part. Sorry I haven't done work about it. But in the coming weeks, I must consider the moment in the program.
Thank you again.
At 2010-10-11 22:56:45，"Bruno Chareyre" <bruno.chareyre@xxxxxxxxxxx> wrote:
I finally gave Law2_ScGeom_CohFrictPhys_ElasticPlastic a try in the cohesive triaxial test (with moment law), but I didn't see any problem. How is the situation on your side?
On 20/09/10 13:52, 焦利青 wrote:Hello Vaclav,
You are right. I see the force is nan indeed. Thank you so much. Thank you for so warm help. I will check the contact law carefully. Thank you again.
At 2010-09-20 17:25:06，"Vaclav Smilauer" <eudoxos@xxxxxxxx> wrote:
>> I am sorry to tell you maybe there are some problems if the rolling
>> resistance is considered in Law2_ScGeom_CohFrictPhys_ElasticPlastic.
>> Yade give the error and stop the calculation. I thought maybe it is
>> derived from the velositybins. But, I still cannot get the force value
>> even if I deleted the all parameters in InsertionSortCollider.
>> The mentioned error is described as follow:
>> 55217 FATAL
>> yade.VelocityBins /home/jiaoliqing/yade/pkg/common/DataClass/VelocityBins.cpp:77 setBins: Body #12179, velSq=-nan not put in any bin!?
>> Could you help me to check it? Thank you so much.
>That error is coming almost certainly from the contact law (unhandled
>division by zero or such); you can try saving the simulation just before
>the problem occurs (it should be always at the same place, if you run
>the same simulation) and the sub-stepping through the iteration using
>the inspector, you will see that the force will be nan after
>Mailing list: https://launchpad.net/~yade-users
>Post to : yade-users@xxxxxxxxxxxxxxxxxxx
>Unsubscribe : https://launchpad.net/~yade-users
>More help : https://help.launchpad.net/ListHelp
Mailing list: https://launchpad.net/~yade-users
Post to : yade-users@xxxxxxxxxxxxxxxxxxx
Unsubscribe : https://launchpad.net/~yade-users
More help : https://help.launchpad.net/ListHelp
ENSE³ - Grenoble INP
BP 53 - 38041, Grenoble cedex 9 - France
Tél : +33 4 56 52 86 21
Fax : +33 4 76 82 70 43