Simulating Fractured Pieces Dop problems

[JuriBryan]

Hi everyone,

I have a nother problem:)

I am doing some R&D at the moment on fractured wood effects and I am running in to two big problems that I just can't seem to find a reason for.

First:

I am simulating a few wooden planks hold in place by pin constraints and a Glue network. I am using a field force driven by a particle simulation to trigger the explosion.

That works all great, but the pieces fly straight with out any rotation even though the sampling mode on the field force is set to sphere (that should take the shape of the object into consideration, right?).

And as soon as the pieces hit the ground they start spinning like crazy, intersect with the ground and each other, just a really really bad simulation.

And I have no clue why that is happening.

The second problem is, that I am thinking about using foreach loops to further fracture pieces to get more random looking pieces in the end.

So I would fracture a plank, then check connectivity and through them all into a foreach loop where every piece gets cut up again.

The problem now is though, that the dop network does not recognize that they are different pieces, it simply puts all the pieces with the same group as one object which results in a very strange simulation.

I tried deleting all the groups and creating new groups based on connectivity and to feed those back into simulation but that did not help.

Any help for either of those problems would be highly appreciated!

SplinterSimulation_v01.hipnc

Cheers,

Juri

[nFrame]

If your pieces are super splintery. You'll need to adjust the collision padding in the Bullet Data. Turning on show guide geometry, will most likely not show any collision geo for those super small pieces, so lower the padding until it does. You also may try simming them at a bigger scene scale. That should take care of it, but you may still get a few bad pieces depending on your fracture, just cluster them to a bigger piece. A proper mass is super important too, If one of those wood splinters is the size of a finger nail, but weighs 1000lbs - well you get the idea

Not sure about the secondary fracture issue your having, I vaguely remember the topic being discussed in the most recent Digital Tutors "controlling your fractures in Houdini"

Goodluck!

[JuriBryan]

Sooo I think I solved all the problems.

nFrame you are right, the collision geometry was none existing.

I changed the padding and now it works, some of the really small pieces explode but a simple delete that kills all points below 0 in y fixes that.

Now I only have to tweak some settings in the glue network so that pieces no longer connected to the "barrier" fall down. Any tips for that?

Tried checking how many primitives each point of the glue network is connected to and delete the ones with less then two... but that is just a gigantic mess:)

And then I need to get the auto freeze right... since pieces move to much for my liking but that is not all to bad:)

As for the second problem, no fix in sight just yet but I will take a look at the digital tutor video and see if they have a solution.

I also have one more question.

now that the rbd simulation is pretty much working I want to add some smoke simulations to it.

I delete all pieces that are not moving and I am only keeping pieces on the first (I can easily control for how many frames I keep it) frame that it moves.

Then I use those pieces as source... that works great but I am wondering if that is the most efficient way to solve this.

Thanks for the help!

SplinterSimulation_v02.hipnc

Cheers,

Juri

PS: the smokeSource node imports the rbd simulation... just use the existing rob to cache it out and see the full effect.