2 replies to this topic

[Scratch]

Hey odforce!

I'm experimenting with voronoi dynamic fracturing at the moment and I'm getting some strange behaviour in my simulation which I'm sure you can help me fix.

My testsetup is simple. A Utha Teapot falling to the ground, shattering on impact (allowing 2 shatter steps). You can download and watch my test-video to get a better impression on what I'm talking about next:

Video Download:
www.scratch-arts.net/Downloads/Teapot_fracture.zip

My Hip-file is attached to this post aswell.


Problem 1:
Afte the majority of the (bigger) shattered pieces come to a rest, some small pieces still keep moving like there is no friction or dampening or anything (but of course there is). They just keep spinning on the ground, or sliding on the ground with no counterforce to bring them to a halt. What is causing this strange behaviour? Can I somehow add a global dampening to the simulation to bring everything to a rest?

Problem 2:
When I allow multiple shatter steps (eg. that a big piece can shatter again once more it hits the ground), it looks like a explosion rather than shattering. Meaning it seams as if the parts are getting new energy after shattering, flying up in the air when hitting the ground again, which looks as said like little explosions rather than just shattering.
I'm not sure which setting causes this behaviour. I guess its related to the velocity transfere settings, but I dont understand how to set them correctly. Any hints here?

Problem 3:
I know RBD simulations have to be modeld in real life size in order to be simulated "physically correct". But when setting the grid unit to 1 unit = 1m, and 1unit = 1kg mass, and modeling e.g a teapot with a radius of 0.15 which would be a teapot of 30cm in with, it won't shatter when hitting the ground (falling from 1m above the ground). This happens independently from the  min/max impact settings. Even setting min impact to 0 still gives me no shattering. The simulation just does not pic up the impact. When I increas its size of the teapot to radius 1m  in size (=1unit), it shatters correctly. I solved the problem by setting the scene to 1 unit = 1 decimeter, and  1 unit = 100g mass, but I would rather like to work with meters and kg since it is easier to set up all the physical parameters like density and stuff because they are always given in kg/m^3 and so on. Can you explain this strange behaviour?

Furthermore, if you find anything else I might do wrong or if you have any suggestions for better (sim-)settings, please do not hesitate to tell me. I'm very happy/eager to learn and improve!


Thx in advance for your time, help and tips!
Looking forward to your answers!

Attached Files

•  Teapot_Fracture.hipnc   1.63MB   63 downloads

[JuriBryan]

ok i am thinking that you are using houdini 12...
problem1:
H12 defaults to the bullet system for its RBD solver... that means that Houdini will create lots of little cubes or spheres to resample your mesh... this makes the simulations a lot faster the problem though is that as you sad strange movement can happen.
so what you have to do is up the substeps on the RBD solver... you can go very high and its still fast... but dont go to crazy on it:P

problem2:
there is a setting on the fracture node... i think it is right under the multi fracture parameter... its called minVolume or something close to that(sorry dont have houdini at school)
turn that setting up... what it will do is check the volume size of each piece and see if it is smaller or bigger then your value... bigger means yes break again smaller means no.
the second problem of this problem lays with in the velocity tap of the fracture node... there are settings like velocity transfer and radial velocity... just animate then down when the second fracture appears...
if you know how to script in parameter windows you could even say if a second fracture appears have this value if not use that... so you can change speed angle and all with out having to rekey those values.

problem3:
i actually have no idea:)
sorry...
try to change the collision resolution on the tea pro... it could be that its to small for the standard collision volume...

[Scratch]

Thx alot! I will try your tips as soon as I'm back from FMX which will bee next week! Cheers!