Hello, very new Houdini user here!

I have a very simple concept where I want to create a stone wall (glued voronoi fractured RBD pieces) that is attached to a ground plane (static object). If the wall is hit by an object with enough force then it will obviously break up, which is what I want. The problem I am having is that I can't get the wall to break up from the impact while the bottom of the wall that is touching the ground stays in place, as it would in real life if a wall was properly built into the ground. In my scene currently, because I can't find a way to constrain/glue the base of the wall to the ground, the base just tips over unrealistically (while the crumbled pieces above it fall down on top).

How do I constrain the base of this fractured geometry to the static ground? I have already tried to use the “Glue Adjacent” shelf tool to glue them together but I can't figure that out for some reason. I have also tried to use some combination of the “Constraint Network” node but I also don't know how to work that either. I finally tried the “RBD Pin Constraint” node by connecting the wall to the ground but that stops working after I make the stone wall into a voronoi fractured object.

This seems like such an easy thing to do but I can't figure it out. I've been working at this for 2 weeks and can't find a solution! Please help!

Thanks!

Hello! Not sure what solver engine you're using for this (RBD vs Bullet). From what you've described, it would be easy to set up a simulation like this using the bullet solver.

As far as creating a base for the wall, here's a simple setup that accomplishes this without gluing the wall itself to a ground plane. It just tells the bullet solver not to simulate pieces below a certain distance threshold. This is more of a work around, because I'm essentially de-activating the simulation for certain pieces that are close to the ground plane. They still act as collision objects, and the glue constraints work as they normally would, the bullet solver is just not applying any forces to any non-active pieces.

In this setup i've just initialized an integer point attribute called “active” in SOPs before feeding that into the DOP simulation. Once in DOPs the Bullet solver will read that integer attribute and won't simulate any pieces when @active==0.

hope this is helpful! I think if you go through this setup node-by-node it will make more sense.