yade-users team mailing list archive

[Bruno Chareyre <bruno.chareyre@xxxxxxxxxxx>]

kan a écrit :
> Thanks, Bruno,
>
> On Wed, Sep 23, 2009 at 11:04 AM, Bruno Chareyre 
> <bruno.chareyre@xxxxxxxxxxx <mailto:bruno.chareyre@xxxxxxxxxxx>> wrote:
>
>     > The algorithm is : displacement = (stress offset) / (total
>     box-spheres stiffness)
>
> hmmmm..... for the unit...
> stress offset unit is  N/m^2
> stiffness unit is   N/m  ?am I right?
> then (N/m^2) / (N/m) =(1/m), this is not the unit of displacement of 
> m, I think if we use force may be better.
> then Force/stiffness=(N)/(N/m)=m.
>  
Yes. Or consider that the stifness is a stress/meter (increment of 
stress on the boundary associated to a unit displacement).


>     On one hand you compute the total stiffness on boundary N = the
>     sum of all sphere-box(N) contact stiffnesses.
>
> ok, here the stiffness is due to the sphere-box are parallel connected 
> along the boundary.
>
>
>     On the other hand you have the current stress (or, say, the
>     current force if you prefer) applied on boundary N. From this you
>     can define the "offset", which is (target stress - current stress).
>
> so from here, we can get the stress offset, if the applied force 
> (stress) is not the target force (stress)
>  
>
>
>     Then, the displacement computed in the equation above is the one
>     "so that the new force will be exactly the one we want after the
>     displacement". (assuming spheres are not moving in the meantime,
>     which is not true, but it will be corrected again on the next time
>     step, so it finally converge relatively well)
>
> well, if the sphere move fast (such as compressing loose sand to a 
> solid condition), then this may not sufficient to be corrected in the 
> next timestep (I am not very sure , but I feel it may not sufficient 
> to...) . but for static condition, I believe it is correct.
>  
> One more question:
> Assume the boundary spheres will not change (always those spheres 
> nomatter they are moving or not), how about I just always apply a 
> constant force on each sphere (if the stress on this boundary is S, 
> then assume the area of the sphere is A, then we can get the force 
> F=S*A, so that since the stress S is a constant, so the force is also 
> a constant ) and then nomatter it is static or it is dynamic, can we 
> say it is also under a constant stress boundary? what do you think?
>  
It sounds complex. How will you define which spheres are on the 
boundary? What will happen if one sphere is going out of the packing 
even with this constant force on it? You need a stiffness here, not a force.

Bruno

> Thanks .
> Yongfeng
>  
>  
>
>     This algorithm has no name that I know.
>
>     Bruno
>
>     kan a écrit :
>
>         Thanks, Bruno,
>
>          I checked the code, but I still did not understand the
>         physical meaning of the code.
>         May I ask : what is the physical meaning behind the code? or
>         what is the mathematic of the algorithm?
>         how is it called ?
>          Thanks a lot.
>         yongfeng
>
>
>          On Fri, Sep 11, 2009 at 10:22 AM, Bruno Chareyre
>         <bruno.chareyre@xxxxxxxxxxx
>         <mailto:bruno.chareyre@xxxxxxxxxxx>
>         <mailto:bruno.chareyre@xxxxxxxxxxx
>         <mailto:bruno.chareyre@xxxxxxxxxxx>>> wrote:
>
>            More precisely, it is in TriaxialStressController. This engine
>            needs to be assigned 6 boxes (one for each boundary). The boxes
>            are used to define the surfaces (since you need a surface to
>            define a stress).
>            If you don't have 6 boxes, you need to find a workaround.
>
>            The algorithm is : displacement = (stress offset) / (total
>            box-spheres stiffness)
>
>            Bruno
>
>
>
>
>     -- 
>
>     _______________
>     Chareyre Bruno
>     Maître de Conférences
>
>
>     Grenoble INP
>     Laboratoire 3SR - bureau E145
>     BP 53 - 38041, Grenoble cedex 9 - France
>     Tél : 33 4 56 52 86 21
>     Fax : 33 4 76 82 70 43
>     ________________
>
>
>     _______________________________________________
>     Mailing list: https://launchpad.net/~yade-users
>     <https://launchpad.net/%7Eyade-users>
>     Post to     : yade-users@xxxxxxxxxxxxxxxxxxx
>     <mailto:yade-users@xxxxxxxxxxxxxxxxxxx>
>     Unsubscribe : https://launchpad.net/~yade-users
>     <https://launchpad.net/%7Eyade-users>
>     More help   : https://help.launchpad.net/ListHelp


--

_______________
Chareyre Bruno
Maître de Conférences

Grenoble INP
Laboratoire 3SR - bureau E145
BP 53 - 38041, Grenoble cedex 9 - France
Tél : 33 4 56 52 86 21
Fax : 33 4 76 82 70 43
________________