POROUS MEDIA MICROSTRUCTURE RECONSTRUCTION USING PIXEL-BASED AND OBJECT-BASED SIMULATED ANNEALING – COMPARISON WITH OTHER RECONSTRUCTION METHODS

Abstract

In this contribution the issue of the stochastic reconstruction of particulatemedia from 2D photomicrographic images is addressed with particular reference to pore space connectivity. The reconstruction of porous bodies in 2D or 3D space was achieved by using simulated annealing techniques. Two methods were proposed to reconstruct a well connected pore space. The first, named PSA (Pixel-based Simulated Annealing), a pixel-movement based, three constraints were found to be necessary for the successful reconstruction of well connected pore space: the two-pointcorrelation function, the d3-4 distance transform distribution and the linealpath function for the pore phase. The second, named OSA (Object-based Simulated Annealing), only constrains the two-point correlation function. Following several researches which tried to reconstruct porous media using pixel-movement based simulated techniques, we propose a new parameter to add a microstructure descriptor, but we also propose a new technique, based in moving the microstructure grains (spheres) instead of the pixels. Both methods were applied to reconstruct reservoir rocks microstructures, and the 2D and 3D results were compared with microstructures reconstructed by truncated Gaussian methods. The PSA resulted in microstructures characterized by poor pore space connectivity, and by artificial patterns, while the OSA reconstructed microstructures with good pore space connectivity. These results indicate that the OSA method can reconstruct better microstructures than the present methods.