RealFlow Official Forum

The workflow for making this behave is narrow compared to working with fluids, but I got it working by 1) not setting the rigid bodies too close together at first, 2) keeping the rigid bodies below the granular particles instead of resting on top and 3) not using Initial State for the rigid bodies but simulating it all at once from setup to end. I had tons of strange behaviors before and set this test aside twice before getting this far. Multiphysics is neat when it works. Hopefully it's undergoing significant improvement.

Some errors from setting the rigid bodies too close and on top at the start of the simulation.

Interestingly, the rigid bodies can be set close together when the substeps are set low 2-4, but there is some interpenetration. The problem is when working with granular particle interactions, because the granular particle domain's substeps must be set high at 16-32 substeps. For a single project, each domain can have its own settings, but when all domains are being simulated at once, the domain with the highest settings is used for all domains. Would a low substep domain interact well with a high substep domain? Maybe that's why it uses the highest substep domain for all domains when simulated together. In which case it just needs to function better in general.

It might be neat if dust volumes could be generated for Dyverso granular particles the way mist volumes can be generated for Hybrido particles. Mist volumes are pretty basic though, and don't look realistic in animations. So, the whole mist volume simulation would need to be improved before it would be worthwhile to do a version for dust volumes. Otherwise, people will have to do it afterward in the main DCC app, using plug-ins like Turbulence FD as a secondary simulation based on the RealFlow particles.

Better render with white HDR background lighting in Maxwell Render:
https://youtu.be/jbueuC6AZbg

This was an earlier test, and one that better matched what I wanted to do with this type of multiphysics simulation. Some of the geometry has an Iso Surface that is too simple for the hard angles of the rigid body geometry. Is there a way to tune that tighter when using Dyverso rigids, because some of the granular particles get stuck there in the corners.

Yes, it does. The Dyverso rigid body will follow the accuracy of the geometry's volume cell size. So, just increase the cell size to fit the accuracy needed for angular geometry. In the image below, the left uses the default cell size of 0.1 for the cross' geometry and the right uses a setting of 0.01, which fits the 90 degree angles of the geometry just fine.