yade-dev team mailing list archive

[Christian Jakob <jakob@xxxxxxxxxxxxxxxxxxx>]

Hi Jerome,

I introduced isBroken some years ago to deactivate liquid bridges. I  
used it to simulate rising water level:

O.engines=O.engines+[PyRunner(iterPeriod=100,command='eraseLiquidBridges()')]

def eraseLiquidBridges():
  for i in O.interactions:
    if i.geom.contactPoint[2] < waterLevel and i.isReal:
      i.phys.isBroken=True	#delete liquid bridges below water level

meniscus is set to true by [1] or [2] automatically in the next time  
step, so in the current version it is not redundant with isBroken and  
can not be absorbed by it.

> This seems to me contradictory (zero volume meniscii might "form"  
> between contacting spheres depending on the succion) and, associated  
> to the test l.139 [3], it leads, in some cases, to interactions that  
> are never deleted, even if there is no more geometrical overlap  
> neither a meniscus with a non-zero amount of water (because erase on  
> l.160 might be switched).

This is more or less a bug, but not harmful since the contact is  
inactive (does not influence forces on bodies). I use it to remember  
contacts, that were set to isBroken manually.
The code is indeed quite confusing. I think that second condition in  
[3] can be deleted, since it will never be fulfilled.

[1]  
https://github.com/yade/trunk/blob/master/pkg/dem/Law2_ScGeom_CapillaryPhys_Capillarity.cpp#L123
[2]  
https://github.com/yade/trunk/blob/master/pkg/dem/Law2_ScGeom_CapillaryPhys_Capillarity.cpp#L126
[3]  
https://github.com/yade/trunk/blob/master/pkg/dem/Law2_ScGeom_CapillaryPhys_Capillarity.cpp#L161

Zitat von Jerome Duriez <Jerome.Duriez@xxxxxxxxxxx>:

> Hi,
>
> * CapillaryPhys/MindlinPhys.isBroken is a bool attribute that is  
> false by default, never updated, and never usefully used neither in  
> c++ nor py files of source. Furthermore, its meaning according to  
> the doc seems quite redundant with "meniscus". Do you agree if I  
> remove it ? (it = isBroken)
>
> * Furthermore, the design of "meniscus" itself might maybe be  
> improved. Around l. 119  [1], this bool is set to true as soon as  
> there is contact, or if distantMeniscii should be created
> And, around l.149 [2], it is set to false when volume is zero.
>
> This seems to me contradictory (zero volume meniscii might "form"  
> between contacting spheres depending on the succion) and, associated  
> to the test l.139 [3], it leads, in some cases, to interactions that  
> are never deleted, even if there is no more geometrical overlap  
> neither a meniscus with a non-zero amount of water (because erase on  
> l.160 might be switched).
>
> I propose to suppress the (meniscus = 0 if volume = 0), and the test  
> on "meniscus" l.139.
> I could attach an illustrative script if necessary
>
> Jerome
>
>
>
> [1]  
> https://github.com/yade/trunk/blob/master/pkg/dem/Law2_ScGeom_CapillaryPhys_Capillarity.cpp#L119
> [2]  
> https://github.com/yade/trunk/blob/master/pkg/dem/Law2_ScGeom_CapillaryPhys_Capillarity.cpp#L149
> [3]  
> https://github.com/yade/trunk/blob/master/pkg/dem/Law2_ScGeom_CapillaryPhys_Capillarity.cpp#L139