What are particular features of CpmMat model?

Hi Alexander,

> I'm so sorry about my lack of knowledge but i'm trying to understand what
> type of interactions are used in CpmMat:
>

you can read [1], chapter 3 "Concrete particle model"

>
> 1) Is CpmMat a bonded particle model (BPM) which is described for example
> in "A bonded-particle model for rock"
> D.O. Potyondya, P.A. Cundall. (
> http://mech.spbstu.ru/images/b/bd/Potyondy_Cundall_2004_A_bonded-particle_model_for_rock.pdf).
> Generally, are all models in yade that support cohesion can be viewed like
> BMP?
>

After a short look to the article, it is similar approach, although CpmMat
has no bending nor torsion stiffness. I don't know other materials very
well, but in *my opinion* "bonded particle model" equals particle model,
that supports cohesion. Maybe you can ask another question concerning
cohesive models and also other poeple (not confused by CpmMat in question
title) would give their opinions :-)

>
> 2) If first is true, does CpmMat use contact bond model or beam bond mode,
> like here (http://i10.pixs.ru/storage/0/9/8/picPNG_2192511_18182098.png).
>

Cpm has only normal and shear behavior, no bending nor torsion

>
> 3) Do i understand correctly for CpmMat that:
> a) cohesive contact is created only at the beginning before simulation has
> been started?
> b) during simulation new cohesion contacts aren't created only old can be
> broken?
>

by default yes (if the script is written correctly :-)

>
> 3) Is it possible to use CpmMat for modeling damages (e.g. break on
> tension) for metallic and composite materials (not concrete and rock)? (I
> mean that concrete and metal have different behavior under tensile load
> because metal is streched fist and concrete breaks immediately) .
>

This is really a tricky question :-) The model is inspired by continuous
mechanics damage and plasticity and in tension it is designed to behave
elastically, then short plastic behavior and then exponential-like
softening. Furthermore damage has the effect of stiffness reduction, which
is not the case of metals (at least for the most part of stress-strain
diagram).

it also strongly depends on definition of "damages (e.g. break on tension)
for metallic and composite materials" and loading scenarios. For pure
tension only maybe it could work, but if it should describe tension and
compression failure at the same time, then definitely not, as the
compressive/tensile strength ratio is usually >8 in Cp,, wich is definitely
not the case of metals :-)

So a general answer is no, it is not possible, but there is some very
little chance it could work in very special circumstances.

cheers
Jan

[1] http://beta.arcig.cz/lide/eudoxos/smilauer2010-phd-thesis.pdf

Hi Alexander,
Your are maybe looking for Law2_ScGeom6D_CohFrictPhys_CohesionMoment (second one in [1]).
It contains, but it is much more general that the BPM you mention. It allows different combinations of friction/cohesion on every contact degrees of freedom (including torsion and bending). And possibly creep.
Bruno

[1] https://yade-dem.org/wiki/ConstitutiveLaws

Hello Bruno and Jan, thank's a lot for answers.

So questions specially for you Bruno.

> well, but in *my opinion* "bonded particle model" equals particle model,
> that supports cohesion. Maybe you can ask another question concerning
> cohesive models and also other poeple (not confused by CpmMat in question
> title) would give their opinions :-)

Do you agree with Jan that in general "bonded particle model" equals particle model, that supports cohesion?

Is it possible to use Law2_ScGeom6D_CohFrictPhys_CohesionMoment for modeling damages (e.g. break on tension) for metallic and composite materials?

with best regards Alexander

And what is the difference between the application area of CmpMat and Law2_ScGeom6D_CohFrictPhys_CohesionMoment for example Why there is a lot of constituve laws and they are not united together?

with regards Alexander

>Do you agree with Jan that in general "bonded particle model" equals particle model, that supports cohesion?

It depends. First, how do you define "bonded particle model"? ;)
I can tell you what Law2_ScGeom6D_CohFrictPhys_CohesionMoment does.

>Is it possible to use Law2_ScGeom6D_CohFrictPhys_CohesionMoment for modeling damages (e.g. break on tension) for metallic and composite materials?

The implementation lets you decide if the interactions are fragile or ductile.

>And what is the difference between the application area of CmpMat and Law2_ScGeom6D_CohFrictPhys_CohesionMoment for example Why there is a lot of constituve laws and they are not united together?

If there was one single force-torque-nonlinear-elasto-visco-plasto-frictional-cohesive law including damage and creep its documentation would be a nightmare for you to read and maintainance a nightmare for us dev. And it would still not cover the full range of things (bubble interactions are missing...).

This being said, they are not all justified. Some of them are just "CohesionMoment" or "CundallStrack" (the simpler torqueless version) with a very little change which could have been merged in the original one.

First things first:
1/ which are the equations you want in your interaction model? then
2/ is there a functor available to do that?

If you skip 1/ the discussion becomes confused usually...

>
> Why there is a lot of constituve laws and they are not united together?
>

because each law has its own purpose :-) there are probably overlaps in
some cases. Usually the simpler law is a special case of more general one
(I would guess that setting proper parameters you could get
Law2_ScGeom_FrictPhys_CundallStrack law from most of the laws). However,
the more general laws differ and this is simply the reason why there are
many of them :-)

Theoretically it would be possible to create one "united" material/law with
1000 parameters (996 default), but it would be just mess and from user's
point of view useless..

cheers
Jan

Bruno and Jan thank's for answers.

> It depends. First, how do you define "bonded particle model"? ;)

 Here i mean that bonds described in (http://mech.spbstu.ru/images/b/bd/Potyondy_Cundall_2004_A_bonded-particle_model_for_rock.pdf) are something like cohesive interactions in yade.

So i will think about your comments

Thanks a lot again

with best regards Aexander

Hello,

Regarding the comparison with Potyondy's BPM:
in my opinion, yes all Yade contact laws supporting cohesion can be compared to such contact model, especially if you consider the simplified "contact bond" version of the BPM.

And, just to add more confusion about Yade contact laws, let me mention https://yade-dem.org/doc/yade.wrapper.html#yade.wrapper.Law2_ScGeom_JCFpmPhys_JointedCohesiveFrictionalPM

This last one (supporting cohesion too: it is designed for rock) having as particular feature a controlled interaction range so that distant particles may be bonded (as CpmMat maybe ?), and the possibility to modify contact normal direction for contacts corresponding to planar discontinuity surfaces.

>Here i mean that bonds described in...

So #2 applies: "It contains, but it is much more general that the BPM you mention"

@Jerome
>This last one (supporting cohesion too: it is designed for rock) having as particular feature a controlled interaction range so that distant particles may be bonded

It is not specific really. All cohesive laws do that, else how could they work in traction if "distant" means "no interaction"?
I once modelized a truss with CohesionMoment...

Then let me precise that, in JCFpm, the initial strictly positiv distance at the bond creation is directly associated with zero contact force, which is a difference with Law2_ScGeom6D_CohFrictPhys_CohesionMoment.

You mean that there is no need to add this in the script:
"for i in O.interactions: i.phys.unp=i.phys.un" #set current=equilibrium distance in CM

You were right, it adds confusion, and it illustrates very well the "they are not all justified" part of #5. ;)

Yes Bruno, I know this "trick" too, which may require one or two more lines than the one you quote (see https://answers.launchpad.net/yade/+question/266828, #4 especially).

Yade suffers probably from having too many laws, but as for Law2_ScGeom_JCFpmPhys_JointedCohesiveFrictionalPM, I do not think this not so interesting discussion may conclude it is not justified, since I mentioned earlier a second specific feature about the discontinuity contacts.....