Résumé
Imaging reservoir rocks in 3D using X-ray microtomography with spatial resolution ranging from about 1 to 10 mm provides us a unique opportunity not only to characterize pore space geometry but also for simulating hydrodynamical processes. Yet, pores and throats displaying sizes smaller than the resolution cannot be distinguished on the images and must be assigned to a so called microporous phase during the process of image segmentation. Accordingly one simulated mass transfers caused by advection and diffusion in the connected pores (mobile domain) and diffusion in the microporous clusters (immobile domain) using Time Domain Random Walk (TDRW) and developed a set of metrics that can be used to monitor the different mechanisms of transport in the sample, the final objective being of proposing a simple but accurate upscaled 1D model in which the particle travel times in the mobile and immobile domain and the number of mobile-immobile transfer events (called trapping events) are independently distributed random variables characterized by PDFs. For TDRW the solute concentration is represented by the density distribution of non-interacting point-like solute particles which move due to advection and dispersion. The set of metrics derives from different spatial and temporal statistical analyses of the particle motion, and is used for characterizing the particles transport (i) in the mobile domain in relation with the velocity field properties, (ii) in the immobile domain in relation with the structure and the properties of microporous phase and at the mobile-immobile interface. We specifically focused on how to model the trapping frequency and rate into the immobile domain in relation with the structure and the spatial distribution of the mobile-immobile domain interface. This thorough analysis of the particle motion for both simple artificial structures and real rock images allowed us to derive the parametrization of the upscaled 1D model.