利用者:Sebbas/GSoC 2017/Ideas

Ideas (GSoC 2017 - Blender Mantaflow Integration)

Ideas for my GSoC 2017 proposal. They are based on previous integration steps of Mantaflow into Blender.

Fluids

• Node based fluid solver
  • "Everything nodes" is one of Blenders goals for 2.8. Since we communicate with Mantaflow via a flexible Python interface a node-based fluid solver seems feasible.
• Accelerate Blender 2.8 / master branch integration
• Multi-fluid domain
  • Currently, a fluid domain is either of type 'Gas' or type 'Liquid' and thus can only hold flow objects of such type. A multi-fluid domain would allocate resources for both smoke and liquids (e.g. all smoke and liquid grids + particle system) and let liquid and smoke interact (e.g. they share the velocity grid).
• Primal-Dual Optimization for Fluids
  • The CG Group at TUM published a new paper and their results on fluid flow guiding are already in Mantaflow 0.10.
  • More: http://www.ntoken.com/pubs.html#Inglis_2017_pdfluids

Liquids

• Secondary particle system
  • Elbeem had this functionality and with the internal Mantaflow particle system, similar effects should be possible. The Mantaflow particle system would need to be hooked up to the Blender cache and an appropriate UI should let the user control the particles.
• Viscous liquids
  • Not implemented in Mantaflow yet. A viscosity parameter to control FLIP particle adhesion would need to be introduced.