Hello, I've recently started learning about Houdini's fracture and RBD tools and have some questions on some of my early test work. I'm trying to figure out how to have a RBD object pass through a dynamic or pre-fractured column.

The main problem I'm having is that the RBD collision object seems to always bounce or get stuck instead of passing through the column and pushing the pieces out of the way. I've tried various combos of settings: (velocity transfer in the fractureparms, different densities, object collision speeds, stiffness/bounce/friction, glue strengths on my pre-fractured column, etc.) but am having no luck with anything I try.

It would seem to me that larger densities and higher speeds on the RBD object should get the result I want but no luck so far. I've tried lower densities on the column object as well, but it just seems to make the pieces in the back of the pillar opposite the impact point fly out faster, while the areas in the front of the column seem to collide together and fall downward. What settings am I missing here?

* files are made in Houdini master 11 non-commercial

just by having a quick look at your scene, it seems that you played around with the settings of the sphere, but not with the physical settings of the column. Try setting the mass of the “box_object2” to something like 10 and you will immediately see an impact.

regards,

Manu

Thanks Manu! I had actually tried adjusting the density to 10 in a previous version, but it tended to leave the front of the column intact while blowing out the back of it. Mass = 10 is I'm guessing significantly less mass than when it's calculated auto when Density = 10.

In any case, when running simulations like this is the best practice not to worry about getting the density of the physical object correct and instead just worry about how it looks? I.E. if it was a concrete pillar it's not important that it's density be ~2000. Or is it best to use physically correct values and compensate in other areas? I had thought my solution would be to leave the pillar's density at the correct value for concrete and instead increase the values of the projectile object, its speed, density, mass, etc. but that did not seem to give me correct results either.

Seems like what would work best is a way for the projectile object not to be affected as much by the impacts on geometry and for it to mostly push other geometry out of the way, since it always seems to deflect off and not push through the object like I'd like it to.

you know, CG is all about cheating. Think about it, when a little sphere like that would hit a big concrete pillar like in your example it would never make the whole thing collapse, maybe make the impact area crumble a little. Thus you got to set up your simulation either in a realistic way or deal with the fact that it's not and trick your way around it