How could I refine this interpenetration??

[rootboot00]

Hi I'm just studying about RBD solver and its fracturing with first picture's object.

and When I setup those with solvers (rigidbody solver and sop solver will be added),

I could find some interpenetration in it.

 

 

Those boxes were having interpenetration, so I've looking for some tutorials about it.

and I found 'Glue constrain' but that tutorial used it for fractured(and packed) object.

Does 'Glue constrain' be able to used in my situation ??

Thank you for read and always sorry for my poor English. Have a good day

 

[f1480187]

Constraints used to hold simulation pieces together. For Bullet sim you can model simplified simulation geometry out of watertight convex shapes and use glue to make it solid concave object again.

convex_glue.hipnc

[Noobini]

is strength -1 infinite ? (ok so it's a multiplier...so is negative strength infinite ?)

Edited January 11, 2018 by Noobini

[Jesper Rahlff]

1 minute ago, Noobini said:

is strength -1 infinite ? (ok so it's a multiplier...so is negative strength infinite ?)

yep. -1 means unbreakable.

[Noobini]

but...this is me foolin' around again....so to test my theory, I made strength = 1 but in the sim, glueconrel, I set strength to -10000...so

-1 * 10000

or

1 * -10000

same same, they should not break...but the 2nd case does break...what gives ?

[Jesper Rahlff]

4 minutes ago, Noobini said:

but...this is me foolin' around again....so to test my theory, I made strength = 1 but in the sim, glueconrel, I set strength to -10000...so

-1 * 10000

or

1 * -10000

same same, they should not break...but the 2nd case does break...what gives ?

I will have to get back to you on that, as I have not actually opened the file.

I assumed the strength is set in sops to -1?

the real test would be if you removed that multiplication factor, and just set the strength to be -1 in the glueconrel. if it still breaks then it could because the glueconrel only works with positive numbers, but as everything gets solved the mult of -1 is used. hence not triggering an error in the glueconrel

 

this is only a theory of a noob. Don't rely on it to heavily yet

Edited January 11, 2018 by Jesper Rahlff

[Noobini]

yeah if strength is undefined, then DOP works as advertised....even the helpcard says that about -1....but if sop strength is introduced...things get a bit more 'not what I expect'...as in the 2nd case..

[Jesper Rahlff]

wait a minute...!

3 minutes ago, Jesper Rahlff said:

.so to test my theory, I made strength = 1

if you set strength = 1 in sops its not  a mult anymore? 

[Jesper Rahlff]

you want to set strength *= 1;

this way you can do 10000 in gluecon and mult by -1

or mult by 1 and set gluecon to -10000

Edited January 11, 2018 by Jesper Rahlff

[Noobini]

but that's my 2nd test above, set strength = 1 (so really if this is just a multiplier, then it's superfluous, DOP should determine the effective strength....) and then in DOP, strength = -10000...so I expect effective strength = -10000...which is infinite...but it breaks...

[Noobini]

my deduction is this:

01) if sop strength is undefined, DOP strength works as advertised perfectly...(its sign is effective)

02) if sop strength IS defined (it is a multiplier), then DOP strength is used as MAGNITUDE only...DOP sign is ineffectual.

[Jesper Rahlff]

if your sop strength = 1 then that will overwrite whatever you set in your glueconrel. thats why you should mult it

edit: I am gonna open the file now and have a look. hard to argue without any proof

Edited January 11, 2018 by Jesper Rahlff

[Noobini]

huh ? isn't 'effective' strength = SOP strength (if defined) x DOP strength ?

(so I don't mult anything myself...Houdini calcs the strength for me as above ?)

[Jesper Rahlff]

Just now, Noobini said:

huh ? isn't 'effective' strength = SOP strength (if defined) x DOP strength ?

(so I don't mult anything myself...Houdini calcs the strength for me as above ?)

I do not believe that is how it works. I believe that sops will overwrite dop

[Noobini]

ok, the SOP strength is NOT a multiplier now ? I assume it was...

[Jesper Rahlff]

4 minutes ago, Noobini said:

ok, the SOP strength is NOT a multiplier now ? I assume it was...

actually I am not sure, because it does not behave is I expected.

[Noobini]

i remember from some various videos that it's a multiplier...but if you say it overrides...that makes more sense to what i am observing..

[Jesper Rahlff]

but if it overrides then you should be able to set strength = 1000000 in sops and then 1 in dops and it should not break?

edit: and it does not break.. so I guess it is an override. 

what puzzles me though is, why if I do f@strength *= value , it still overrides and does not multiply as it should..

Edited January 11, 2018 by Jesper Rahlff

[Noobini]

I'm testing if 'effective' strength is negative....is it infinite....that was my orig question

so fine if effective strength is -1 x 10000 = -10000...works

then i tested 1 x -10000 = -10000 doesn't work...

of course, the above assumes the effective strength is a product of SOP strength and DOP strength....but seems the assumption is wrong..

(no I have no interest in testing an insanely large positive number to see if it breaks, i'm testing a negative number that should NOT break)

[Noobini]

I can see in the spreadsheet that in SOP land, it is set to say -1.0....but in DOP land, the GlueConRel is its own value, say 10000...so I cannot find anywhere that it is a 'product' of -1x10000....but....it is also not being overridden by the -1 !!!