I have a series of micro-tomography images of a porous material (received from experimental partners) with a resolution 10 $\mu m$ with total size 1cm $\times$ 1cm. Porosity of the porous material lies in the range 7-10$\%$ and is anisotropic. I am supposed to run fluid dynamical calculations through such porous media.
The problem I have is that, most probably due to poor resolution of the images, segmented or binary images obtained from the original micro-tomography images differ from each other. The permeability values of the reconstructed porous material is 10-50 times larger than the experimental values. Given that fluid solver is well benchmarked, the only possibility is that reconstructed structure does not corresponds to the original structure. The experimental partners maintain that this is the ideal resolution to include heterogeneity on mm-cm range.
I want to know if there are algorithms (like multiple points statistics for isotropic media), that can stochastically generate structure and also account for anisotropy. A quick look through Google scholar shows that there are abundant algorithms for isotropic media with porosity higher than 20$\%$. given that I do not have much experience in this field, any suggestions for algorithms or its impossibility will be highly appreciated. I am not sure of tags for this question (porous media, anisotropy, reconstruction none of these are available)