Solutions to speed-up "simple" 3D simulations

[Henri Colaux]

Dear all,

Thanks you all for your previous replies. That removed the last bits of hassle I had and it now gives terrific results.

While I get the paper done for the current project, there is a question that I would like to ask as I write a proposal for a future project. I am soon to try to simulate diffusion inside porous materials, which requires using 3D simulation models. Yet, I realised from the current 2D simulations that I need fairly low space and time increment to get a precision comparable to that of experiments. This makes the resolution of the PDE really time-consuming, and that will become even worst when we switch to 3D models.
Therefore, I am investigation for solutions to make 3D simulation faster. The 3d "unit cell" that we will be using are fairly simple structures, so I was wondering if there are algorithm that offer, for instance, decomposition of the 3D structure onto bases of functions such as spherical harmonics, for which the resolution would be faster and incremental. All this is far beyond my knowledge as a mere physical-chemist.
In short, if anyone knows any research paper - maybe even your own work - that could be of valuable contribution to this end, I would be really grateful.

Cheers !

Henri Colaux