The construction of carbonate digital rock with hybrid superposition method

Carbonate reservoirs are inherently heterogeneous and the pore sizes can vary over orders of magnitudes, the pores at different scales have a great impact on interconnectivity and flow properties. It is necessary to describe the microscopic pore characteristics of carbonate rocks at different scales and its influence to flow mechanism. In this paper, 2D thin section carbonate images with different resolutions are collected with scanning-electron microscopy (SEM). The lower resolution image shows macropore properties while the higher resolution image shows micropore properties. With two different scale resolution images, a hybrid superposition method is proposed to construct superposition digital rock with two steps, the first step is to reconstruct macropore digital rock with simulated annealing method and micropore digital rock with Markov Chain Monte Carlo (MCMC) method, and the second step is to construct the carbonate digital rock with superposition method. Finally, pore space microscopic structure analysis method and lattice Boltzmann method are used to analyze the pore structure and flow properties. Results show that, the hybrid superposition method could combine the advantages of simulated annealing Method and Markov Chain Monte Carlo method, which could reconstruct macropore digital rock with a better morphology description and micropore digital rock with a less computation time. The carbonate superposition digital rock has a bimodal pore size distribution which could describe the macropore and micropore characteristics simultaneously, it also has a higher percolating volume fraction and absolute permeability than both of macropore and micropore digital rock, which shows that micropores have an important influence on the total connectivity in carbonate rocks. This method has a research platform for the study of multiscale pore characteristics and microscopic flow mechanism in carbonate rocks. (C) 2013 Elsevier B.V. All rights reserved.