Hi, I am wondering if anybody tried to make a 2d water (or any other liquid) simulation. By 2d, I mean side view. I tried creating height-based water represenation, as described in "Dynamic Simulation of Splashing Fluids" by James O''Brien, but it has serious problems with overflows and boundary conditions, and also movement of water isn''t as fast as I would like. Otherwise, the method seems great, as it has complexity O(n). Thanks.

Check out Jeff Lander''s stuff at:

www.darwin3d.com/confpage.htm

His "Taming a Wild River" article is helpful, although it doesn''t have as many details as you would like.

Graham Rhodes
Senior Scientist
Applied Research Associates, Inc.

Yes, I''ve done liquid simulation (fully 3D with free-surface). Nearly killed me.

What do you want to know?

Thanks for the reply, but unfortunatly, I have already read that article.

Someone mentioned that there is some article in Game Programing Gems 3, that deals with similar problem (and has source?), but I was not yet able to get the book.

Basically, I am looking for 2d fluid simulator that has complexity 0(N) (memory and speed), and that behaves like water.
There is no need for breaking waves, turbulence, etc.

The code from article is very simple and it works, if adjacent water heights are similar. If they are not, simulation behaves very strangely (odd and even water columns gets separated), or water starts looking more like sampled sound (in, say, sound recorder

To Mr/Ms/Mrs Anonymous Poster:

I''m interested of your 3d fluid program. I guess you use MAC discretization? How do you track the surface, do you use a fully particle based method or a level set, or possibly a combination of these? How do you handle with boundary conditions and "extrapolating" the surface in time? Do you use the methodology of Foster and Metaxas or the newer ones by Fedkiw et al. or possibly some older models? If you could give your e-mail address, we could possibly continue this interesting discussion in private.

- Mikko Kauppila