Simulating Water Splash Using a Particle System

Zi-Wei Chiou

If you're interesting in my research: http://www.csie.nctu.edu.tw/~jwchiu/research.htm

***SUMMARY***

This is a computer graphics question. I want to simulation the splash of water using a particle system.

***PROBLEM***

I use a Navier-Stokes equation solver (FDM), and the free surface is tracked by Hirt's VOF(Volume-of-Fraction) algorithm. VOF works great, except that in order to increase the fidelity of the animation, the computation mesh must be very fine to be able to caputure very small droplets.

Such heavy computation is not feasible in computer graphics. A common approach is to add a "particle system" to represent the droplets. Many shperical particles are added into the system to represent small droplets, but they're not involved in the computation of Navier-Stokes equations, similar like the "marker particles" in MAC (Marker-And-Cell) algorithm.

Here is a short review of how most existing papers do:

(1) Compute the water motion by solving shallow water equation and model the water surface as a height field.

(2) When the velocity of a part of surface exceeds a predefined threshold (T1), some particles will be generated.

(4) The volume of new particles will be deducted from the water body.

(5) Particles follow simple physics law of motion, and the only governing force is the gravity, so computing their trajectories are very simple (assume we don't deal with collisions).

(6) A particle may separate into two particles if its energy exceeds another predefined threshould (T2) and sometime randomly.

(7) The volume of particles will be added into water body if these particles fall back to water surface, and be eliminated if they fall to the ground or other objects.

***QUESTION***

(A) In (2) particles are generated when the velocity of a portion of water exceeds a predefined threshold value. This criterion is too simple, and it should take more physics laws into consideration, e.g. surface tension.

(B) Most papers model the droplet as shperes, but it's not the case in the real life. Except the gravity and surface tension, what laws dominate the shape of falling droplets?

I've search my university and Internet (water+splash+particle) exhaustedly, but perhaps I don't have much experience in CFD, I can not find any useful reference.

Thanks for anyone who nicely points me to another reference (book or paper).