Predicting the effective gas permeability of gas shale, at scales from a few centimeters to hundreds of meters or even larger, is a requirement for assessing the productivity of shale gas plays, for developing appropriate production and stimulation plans, and, perhaps more importantly, for quantifying associated uncertainties. Like other fine-grained sediments, gas shales are heterogeneous at every scale, from nanometers to kilometers, but these rocks are also rich in organic matter. At microscopic scales, gas shales contain mostly sub-nano to sub-micron sized, intra-/interinorganic pores and fissures, as well as intra-kerogen pores that are probably due to gas generation. Gas transport takes place through these small void spaces in a pressure and temperature or simply state dependent non-Darcy manner, with gas sorption being a key feature in intra-kerogen pores. A multi-scale upscaling is necessary for predicting effective transport properties at useful scales. The upscaling must account for these two important aspects of gas shale flow arising from the sub-micron scales.
We consider here the extension of a multi-scale upscaling framework that was developed for predicting a wide range of properties of organic-lean shales/mudstones (Aplin et al., 2012; Ma et al., 2013), applying it for predicting the effective gas permeability of organic-rich gas shales. Unlike for organic-lean shales, no empirical relationship has been established between effective gas permeability and the lithological and/or physico-chemical properties for organic-rich shales, although a relationship of that form is a critical aspect of the framework. Directly modeling gas shale at micron scales and calculating the effective gas permeability is a possibility for overcoming the lack of robust measured data. However, despite advances in image-based gas shale characterization down to even submicron, this approach can only be applied to samples of tens of microns in dimension, and it is known not to provide sufficient information alone for constructing representative grain/pore models for predicting effective gas permeability. Noting the similarities between organic-lean and organic-rich shales, we propose to employ the approach for constructing pore-space models for organic-lean shales, and apply these methods, with extension, to gas shales. Specifically, the sample-scale properties of the organic-lean shales are adopted as constraints for constructing representative pore-space models using image-based, local gas shale characterizations at micron scales. On such models, effective gas permeability is estimated and is used to populate models of layered shale systems via the correlation of total organic contents (TOC) and the degree of pore formation in kerogens. This paper illustrates the extensions we propose for the existing framework, with justifications, and outlines the necessary techniques.