friction and stiction

[GrahamMac]

What is the best parameter or combination of parameters to control friction or stiction of liquids resting on inclined surfaces? I am finding that my liquid test sims are creating droplets which like to cling to edges and sides of objects when they should be running off to the ground.
Thanks.

[Ivaylo Katev]

there is no friction control for the liquid-object interaction. it should be zero by design, however the objects are voxelized in the grid and this produces artificial roughness that holds the droplets to the wall. definitively we have improve this and some other related issues at some stage.

regards,
ivo

[GrahamMac]

Thank you Ivo for clarifying this!
I assume then the best current solution is to up the density of the grid to reduce the roughness?

[Ivaylo Katev]

this will decrease the droplet's size, but the behavior will remain the same. if is possible avoid the inclined surfaces, instead this you can incline the gravity.

[GrahamMac]

So increasing grid density will not increase resolution of object voxelization?
Also, by inclining gravity do you mean to set gravity direction to an angle different than vertical? Perhaps one could instead rotate the simulation grid to an angle that matches that of an inclined object surface inside the grid, so that voxelization of that surface is effectively smoother for the fluid.

[Ivaylo Katev]

So increasing grid density will not increase resolution of object voxelization?

no, it will, but the roughness remains, i.e. there will be droplets on the walls, but smaller by size.
rotating the simulator will not affect the embedded gravity, it is generated toward th z direction in local coordinates.

[GrahamMac]

Thanks again for the tips.
So to clarify: by inclining gravity you mean one should turn the Phoenix simulators own internal gravity to zero, add a gravity space warp to the scene at say 45 degrees to the vertical and rotate the inclined surface in the grid by the same angle. This should then alter gravitation direction with respect to the grid?
Not that I mind too much but I imagine this method would push up simulation times as we are using a space warp which I believe is one of the more costly processes to run in Phoenix right?

[Ivaylo Katev]

yes correct. and for the slowing down too