Wednesday, February 13, 2008

Fluid Readings for V-Day

Jos Stam. (1999). Stable Fluids. Proc. ACM SIGGRAPH, 121-128. The corresponding talk.

And Chapters 3-4 of Bridson and Müller-Fischer from last class.

1 comments:

Unknown said...

Stable Fluids:
One thing the authors mention is that toroidal fluids make for good for making evolving texture maps. I understand that this is good for textures in general because there are no boundary discontinuities, but what do they mean by evolving textures and are there more reasons than just the above for why they are ideal for these cases?

He claims this is a "stable solver" but I wonder if or how these ideas could be extended to take turbulence and such into account.

SIGGRAPH Course:
This paper also talks about boundary conditions but in a slightly more confusing manner (to me). It is stated that numeric error could cause a particle to fly outside of where the fluid actually is. That's all well and good, but their lengthy discussion on the solution seems to just boil down to "throw it back in at the nearest place". Also, they seem to oversimplify some things. In the case of a grate allowing water to flow in I'm pretty sure not every molecule of water has the same velocity once it passes the threshold of the grate. I'm thinking of eddies, side currents, areas outside the influence of the grate, etc etc. I assume when they say boundary they mean an enclosing boundary, but what if it is not as in a half-pipe or something. Water could easily flow out, how would they tell what are valid outliers and what are not? I'll also tip my hat to them in that their equations looks pretty impressive.