Source Volume Collision vs Static Object plus Static Solver

[art3mis]

Hi

Have been working on a simulation now for a couple weeks but still unclear about the best method to obtain collision between fast moving geometry and smoke. Think fast moving fan or turbine.

For the moment disregard how fast this geometry is moving to displace the smoke(rotation) which I know requires lots of substeps.

How should I bring my SDF collision geometry(rotating fan) into my DOP?

1- Source Volume set to Collision

or

2- Static Object + Static Solver

In the attached you can see I am using the latter method. Not sure if my settings are correct.
For the former method, Source Volume Collision, most people use the last input on your Pyrosolver although in one Peter Quint video I see he uses the 3rd velocity input on the solver.

 

Edited November 12, 2019 by art3mis

[vtrvtr]

Here are some comparisons:

Volume source collision:

Incorrect Static object collision:

Correct Static object collision:

As you can see the static object and the volume source are the same, as they should be since they do the same thing. Note that the I didn't have a collisionvel field, which can make the difference depending on the situation.

The "incorrect" static is "incorrect" because you're putting a vdb in the sop path parameter, but the solver is recreating this vdb every frame, which is likely not what you want. To fix is you need to put the path the vdb in the Proxy Volume parameter and change the mode to "volume sample".

Here is the hip file

pyro_col_ODF.hip

Edited November 12, 2019 by vtrvtr
too many gifs

[art3mis]

Fantastic! Thanks for this.

So in the visualization would it be correct to say that the collision vdb is advecting the velocity field of the smoke object?

One thing I am unclear about in Volume Source is how to visualize the 'collisionvel' Target Field? Simply omitting this field and leaving only collision has no effect.

Edited November 12, 2019 by art3mis