I have a pretty complex deforming geometry and I am looking for an efficient way to collide/advect/influence particles with it.

I converted it into triangles and used it with an collision pop. It works, but it is slow and heavy.

I noticed that velocity fields from dops are pretty fast to advect in a particle network. So, I am wondering if I can create a custom velocity field from my deforming geometry.

However, velocities are vectors, so how does the dop vel field store those? When I import them from a dopnet they look like a standard volume so I'd have thought its just a scalar field, but that wouldn't make sense.

Could I convert the $VX $VY $VZ of my deforming geometry into a velocity field, ready to advect?

This is the closest solution I have come up with. It merges 3 volumes and uses a dopnet to calculate the velocity field. But…hmmmm…something isn't working right.

I don't think I can use $X $Y $Z in the volumes to represent velocity because surely it's just representing the position of the voxels?

On the other hand using $VX $VY $VZ gives an interesting but wrong (?) result (does it recognize velocity variables at all?)

This seems to work. I used a gas microsolver for it.