method to generate packings with different void ratio

Hi Ning,

When using LSMGenGeo or the RandomBoxPacker in ESyS-Particle, the only major controlling factor on the final porosity (i.e. amount of void space) is the range of particle sizes. The algorithm works by looking for void space and trying to fill that space with a spherial particle. If the void is smaller than the specified minimum particle radius, the void cannot be filled (an insertion failure). When you increase maxInsertFails, the algorithm will search harder for voids to fill so you should see the porosity decrease slightly. Eventually there will be no voids left that are larger than the minimum particle radius and the algorithm will terminate. The only recourse to fill those voids is to decrease the minimum particle radius. This result in packings with a lot more particles though.

Also, the practical range of particle sizes is limited to about 20:1 (i.e. largest particle is 20 times larger than the smallest). If you use this maximal range, the minimum porosity you can achieve is about 20%. This is a "feature" of the packing algorithm we use and there's not a whole lot that can be done about it. I think for the most part, the advantages of this algorithm out-weight its disadvantages though.

If you want to achieve a lower porosity, you might consider the algorithm described in Potyondy and Cundall (2004) "A bonded particle model for rock". This is a "dynamical" packing algorithm because you need to run simulations to generate the packing. I once tried to implement this in ESyS-Particle and almost got it working but gave up at the "zap floaters" phase. I'm sure it could be done if you were more persistent than I was!

Cheers,

Dion.