yade-dev team mailing list archive

["Jan Stránský" <_honzik@xxxxxxxxxx>]

Hi,

in PeriIsoCompessor, the Vector3r stiff contains normal ("pseudo") stiffnesses for each axis (stiff[i] = stress[i] / strain[i], where stress[i] and strain[i] are normal components, too, as Bruno answered in previous email). It is not "real" stiffness due to the absence of Poisson's ratio, but it is completely sufficient for purpose of PeriTriaxController (you only need to predict stress/strain for next step from known value of the dual quantity using this normal "pseudo" stiffness).

Very similar approach is used in Peri3dController (with addition of shear components as well), evaluating something like Vector6r stiff instead of full 6x6 stiffness matrix.

>From my point of view, Vector3r stiff is ok, maybe it could be renamed to normalStiff to prevent confusion, but it is only a detail..

regards
Jan

______________________________________________________________
> Od: "Bruno Chareyre" <bruno.chareyre@xxxxxxxxxxx>
> Komu: yade-dev@xxxxxxxxxxxxxxxxxxx
> Datum: 04.04.2011 18:25
> Předmět: Re: [Yade-dev] Stiffness tensor
>
>This is more a question for Vaclav or Jan, but here are my two cents:
>There is no simple definition of stiffness in the general case. The
>definition you find in periEngines is based on the implicit assumptions
>that there is no strong heterogeneity of the displacement field, and
>that rotations of particles are negligible, which can happen if you have
>more than ~10 contacts per particle (it was the case in Vaclav's PhD).
>For e.g. Cundall law or HM, it doesn't apply.
>
>I don't know precise references but I think the main points of the
>derivation are, with the assumptions mentioned above:
>1- relative displacement at contact is du=epsilon*dl, with dl the
>distance between centers
>2- forces increments are df = dun*kn+dus*ks
>3- average df's somehow to get a dsigma (they could be averaged using
>Love Weber stress, but I think there is a different averaging currently)
>4- derive dsigma/depsilon = stiffness
>
>Bruno
>
>
>On 04/04/11 16:54, Chiara Modenese wrote:
>> Hi there,
>>
>> Do you have any reference which shows the formula for the stiffness
>> tensor? I see that in the PeriIsoCompressor.*pp there is a vector
>> called stiff, is this correct?
>>
>> Cheers.
>> Chiara
>>
>>
>> _______________________________________________
>> Mailing list: https://launchpad.net/~yade-dev
>> Post to     : yade-dev@xxxxxxxxxxxxxxxxxxx
>> Unsubscribe : https://launchpad.net/~yade-dev
>> More help   : https://help.launchpad.net/ListHelp
>
>
>
>_______________________________________________
>Mailing list: https://launchpad.net/~yade-dev
>Post to     : yade-dev@xxxxxxxxxxxxxxxxxxx
>Unsubscribe : https://launchpad.net/~yade-dev
>More help   : https://help.launchpad.net/ListHelp
>
>