--> A Multiscale Method for Characterizing Shales by Downscaling and Upscaling Statistically Sampled Datasets From Multiple Imaging Modalities

AAPG Annual Convention and Exhibition

Datapages, Inc.Print this page

A Multiscale Method for Characterizing Shales by Downscaling and Upscaling Statistically Sampled Datasets From Multiple Imaging Modalities

Abstract

Shale is well-known for its high degree of heterogeneity, requiring imaging across many orders of length scale in order to capture information spanning from lithological structure down to the grain and pore scale. Recent advances in 2D microscopy allow imaging of organic-hosted porosity in shales down to sub-nanometer resolution, but constructing large field-of-view images (by stitching many images into mosaics) at these resolutions, while possible, is not practical, as it would require weeks of imaging time for standard size 1-inch plug or petrographic thin section. A method for statistical sampling is therefore desirable to characterize the mudrock, utilizing instead an approach based on “downscaling” and “upscaling”. In this abstract, we demonstrate a practical method for this approach - by using lower resolution plane and cross-polarized light micrograph images, and SEM backscatter and secondary electron images, we sub-divide the whole rock sample area into regions classified as different rock types or lithologies. Then, we statistically sample (“downscale”) each lithotype at much higher resolutions sufficient to discriminate desired grain-scale mineralogy, organic content, and organic-hosted porosity. Properties of the grain and pore-scale resolved images such as pore-size distribution, pore connectivity in 2D, organic content, and mineralogy can then be used to extrapolate average properties within each lithotype (“upscaling”), which can then be linked to local depositional environment and micro-facies. Furthermore, within this process, it is beneficial to use multiple imaging modalities (polarized-light, electron, and energy dispersive X-ray microscopy) for determination of the above properties. We demonstrate the implementation of this technique on standard petrographic thin sections and end-trim of 1-inch core plugs, extracted from a standard 4-inch core. Upscaled data extracted via this method has the potential to be linked to core-level data and could be used to calculate or support calculations of such properties as net pay or original hydrocarbons-in-place.