← Back to team overview

yade-users team mailing list archive

Re: [Question #678799]: O.load but still in the first code


Question #678799 on Yade changed:

Jan Stránský proposed the following answer:
>> Do you have an idea what is the evolution of unbalanced force before? like how much it is oscillating etc.?
> Yes!I know about servo systems.

Sorry for not being specific enough.. I meant this specific case. Maybe the unbalanced force oscillates like between 0.005 to 0.2 and then the behavior what you described is ok..
If it monotoncally converges to 3e-3 and then is 0.16, then it is another case

Sorry, I did not get the second part at all..

> The reason i need to separate the PeriTriaxController() into 2 parts
is the stiffness of particle is large. First isotropic compression, then
one-dimensional compression.

"large" is too vague term. The value "young=64e9" is quite normal I would say.
I see no reason how stiffness implies the need of "separate the PeriTriaxController() into 2 parts"..
Maybe the change of loading is the problem, see the last paragraph.

> If O.dt is big,the second part will distort.

yes, but then also the first part should distort..

> If O.dt is small,it will waste time in the first part. So,is there any
method to save simulation time,such as change O.dt to a the specified
value in the second part.

simulation time (O.time) is of course saved.
you can set O.dt to whatever value..

Do you know how PeriTriaxController behaves when changing stressMask and goal?
I don't use it and have no idea..
If this would be a problem, you can replace the controller by
triax = PeriTriaxController(...) # completely new instance with no history
O.engines = O.engines[:3] + [triax] + O.engines[4:]
It is a good idea to test both approaches (creating a new controller and modifying the existing one) to see if there is a significant difference.


You received this question notification because your team yade-users is
an answer contact for Yade.