Although I love POPs I'd like to take this approach: Simulating sand with the FLIP solver, now I don't know where to start so please help, I'm not afraid of anything complicated but please provide clarity in explanations, this is for a project that requires it to be emitted from cracks in the ground, any one who helps out will get credit when I put the video out there! thanks.

Does anyone know anything on this?

It's probably not suited for dry sand, but you could def use it for wet / runny sand. You can use viscosity to get something closer to sand, but it will be really hard getting a convincing granular property that sand has – the inter-particle collisions just don't exist with flip. Really though, It really depends on the shot and how close you'll be to it. Anyways, definetly play with variable viscosity. You can set up the source points with a viscosity attribute (can vary per point) and use that in the flip solver. Eetu has an example that looks like it could be used for sand. http://eetu.iki.fi/hou/mud/mudsolver_03.mov [eetu.iki.fi] Also, you'll probably want to turn off reseeding since that will likely be adding and removing particles from the sim, causing flickering if you intend to instance particles to the points, or even just render as points (again it depends on the shot you're after)

For something close up, bullet might be a better choice. just tons of small boxes or spheres…

FLIP's built of the old sand solver, so i had the idea that this would suit, what should I use instead? And can you link to some resources or give me some other info? Bullet's really inefficient from where I'm standing, I've done tests and have got 4-5 million flip particles fine, but that many with RBD would kill my computer.

Well again, it depends on the shot. It can definetly work, and I'm thinking viscosity is probably the key here. You can still use the old sand solver (though I don't think it's been worked on recently), but I think it may even give you similar results to flip with viscosity. I don't have any experience with it, but from what I remember of the paper the stuff is based on, it also didn't have much granularity. You said 4-5 million particles, so I'm guess it's not close up. For a far away shot, flip is probably great (but again, it won't be as granular and dry as a rbd sim or something custom), using the viscosity settings, and by turning reseeding off.

Here's some interesting videos and stuffs:

http://www.cs.ubc.ca/~rbridson/docs/zhu-siggraph05-sandfluid.pdf [cs.ubc.ca] the paper for the sand solver / flip solver – the sand solver still doesn't do the granularity, but to me it looks the same as flip with viscosity would.

http://www.youtube.com/watch?v=2Dx8hwwWMe4 [youtube.com] some interesting approaches in here

http://www.geonak.com/blog/?p=656 [geonak.com]
https://vimeo.com/album/2035007 [vimeo.com] custom plugin for Houdini (not available as of yet) but it does ‘grains’ very well.

Thanks, well the shot seems to be a pop one, as It's only secondary CG that adds to a ground crushing, If i finnish it i'll be sure to include you in the who helped me credits.

I also have the same problem,but i found a solution:
Take a look at CUDA SDK example “particle”,it is real sand(see its reference paper),it computes particle self-collision and friction.

@ Last poster, can you link me kind sir?

Just in case, there is a hip file for that muddy sand thing in here. [forums.odforce.net]

eetu
Just in case, there is a hip file for that muddy sand thing in here. [forums.odforce.net]

Wow thanks, Wouldn't something like Lagoa in Houdini be amazing. So powerful, more powerful than within ICE I'd reckon.

Yeah. hopefully SimSOM [sidefx.com] will be it.