see video I posted in p.b.a

I'm beginning the process of simulating bubbles in water using my 
innaccurate ideas about how physics works. I've started with particles that 
attract eachother  when too far, and repel when too close. 

(eventually I'll add a threshold beyond which particles will no longer 
attract eachother.)

The result is pretty good, but that animation took 5 hours to parse total. 
so I'll need to find a better way of organizing my program. Maybe I shall 
take the lead of Tim Nikias, and simply write all the data at the beginning 
and simply re-read the file in, instead of going through the whole 
simulation again and again for each frame. 

At any rate, I need a method for generating a surface or surfaces from 
these clumps of points, and a way of identifying upper most particles to 
apply force. I think to identify upper most particles, I can simply add a 
point above, and check which particles are closest to that point, However 
that has some problems with accuracy.

I could attempt to simulate particles bombarding the top with "water 
molecules" but then I'm a bit shady on how I should go about detecting 
collisions. (though I have some rough ideas)

as for the surface, I could just use blobs, but they're slow and lumpy. Or 
I could try some kind of triangulation algorithem. I could use the same 
approach of particle bombardment, and simply note where the colissions 
occured. But then how should I determine where one bubble ends and the next 
begins?

I thought about identifying groups of air particles by proximity, but the 
detailed logistics of that kind of bog me down.

Oh well, I apologize for thinking out loud, but I was hoping to spark some 
discussion on the subject.

Thanks, and I hope you enjoy the brief animation.

Post a reply to this message