[Bf-committers] Unified physic simulation (Lagrangian)

Xavier Thomas xavier.thomas.1980 at gmail.com
Tue Mar 16 18:17:17 CET 2010


Hello,

I found an interesting page on Computer Graphic recently
(http://www2.cs.kuleuven.be/~graphics/index.php/publications) and may
attention was caught by this work:
http://www2.cs.kuleuven.be/~graphics/publications/LenaertsPhD/

It present a unified method based on SPH for simulating everything:
- fluids: water/fire/smoke (multiphases)
- cloth
- rigid body (with elastic and plastic deformation, melting, breaking, ...)
- soft body
- porous body
- thin shell
...
and their interaction mixed all together.

I think it is pretty similar to the new particles simulation of max
and maya. Anyhow it is a really interesting read and once I started I
couldn't stop reading it till the end. It sums up pretty much every
part of the simulation process, including the base mathematics and the
surface reconstruction. It has a lot of nice graphs explaining all the
operation and is really didactic.

Further he explains that it is straight forward to modify an SPH
implementation to obtain this unified architecture. A downside I guess
is that the number of total particles will increase because collision
object are simulated as particles too so the simulation time increase.
Real-time simulation being still possible with GPU implementation.

Also Blender is in the bibliography :)


I hope it will interest somebody here and if it is not ok to post this
kind of stuff here tell me and I will never do it again.


Xavier


More information about the Bf-committers mailing list