yade-dev team mailing list archive

[Bruno Chareyre <bruno.chareyre@xxxxxxxxxxxxxxx>]

Christian,
We better move this discussion back to yade-dev.


>
>> The fluid force Ff[i] = p*A[i] is already included if you read
>> O.forces.f[wallId]. As soon as PFV comes before TriaxController, you
>> don't have to compute it again.
>
> You mean PFV must be defined before TSC in O.engines to get wall-fluid
> interaction automatically? Or do you mean the force is already
> calculated, but it does not act on the wall?

If PFV is before TSC in engines it is automatic, yes.

>
>> The problem is in the definition of kn, not in the definition of dF.
>> Currently kn is the stiffness of the solid phase. It should be the
>> cumulated stiffness of both phases.
>
> What is the cumulated stiffness of both phases? kn + k_fluid?

The stiffness should reflect the stress-strain relation in uniaxial
undrained deformation.
Should be Young+k_fluid/porosity I think, but it is not required to be
absolutely exact (e.g. assuming porosity=0.3 is good enough).

>
>> If the fluid is incompressible the cumulated stiffness becomes infinite,
>> which is consistent since an undrained test with incompressible fluid is
>> a ill-posed problem.
>> You could test this idea by using a user defined stiffness in order to
>> see if you can stabilise the engine by defining a higher stiffness.
>> If so, TSC could automatically search a PFV engine in O.engines, and if
>> found it would use pfv.bulkModulus to update the stiffness.
>
> Fine, but I still have a problem in understanding what a fluid
> stiffness is. How can I update the stiffness from bulk modulus? Is it
> kn += K/A? Then what is A?
Fluid stiffness is the inverse of fluid compressibility. ;)
PFV.bulkModulus is the stiffness. Maybe not very well named, since bulk
is more commonly used for solids or mixtures.

Cheers

Bruno