Validation of the Results of 2D Image Analysis Using Laboratory Measurements of Porosity, Permeability, and NMR Measurements

Thin section imaging and image analysis will provide robust estimates of petrophysical properties when measurements cannot be made (e.g. percussion sidewall cores), when funds are limited, or to augment existing rock properties measurements in a well. Although high resolution micro CT imaging is preferred for modeling flow properties, there are limitations when modeling geomechanical or acoustic properties. This is due to resolution constraints as well as the inability to quantify mineralogy, particularly the volume of overgrowth cements, which dramatically impact both static and dynamic properties. This paper illustrates the use of 2D image analysis in both thin section and SEM for estimating pore throat and pore body size distributions, total porosity, and absolute permeability. A plug in for ImageJ was developed that allows for simple pore system analysis (total porosity, pore body and pore throat size distribution, and fractal dimension corrected estimates of specific surface area). In addition to characterizing the pore system, quantitative mineralogic and textural analysis can also be performed including grain size and sorting, grain contact length and orientation analysis, and the volume and distribution of detrital and authigenic phases. The influence of applied pore system segmentation parameters on the results is quantified and the influence of each step in the process on reported estimates of the characteristics of the pore system is illustrated. Image analysis results are compared with laboratory measurements including Boyle’s Law and Archimedes porosity, and NMR estimates of total porosity, pore body size distribution, and measured brine permeability for a range of sandstones. An example of MICP measurement of pore throat size distribution in a carbonate is also illustrated and compared with pore throat sizes obtained in thin section. Thin section orientation with respect to bedding is an important parameter. End trim thin sections are typically cut from both horizontal (porosity/permeability measurements) and vertical plugs (geomechanical measurements) for quantifying mineralogy and porosity. Pore body size distributions have been shown to vary as a function of orientation with respect to bedding, resulting in different predicted rock properties. A vertical thin section from the carcass material adjacent to vertical plugs for geomechanical measurements should be utilized to capture the variability of the sample as well as to correspond with the orientation of end trims from horizontal plugs for pore system analysis. The entire thin section obtained from each sample is scanned at high resolution using a transmitted light microscope in plane and cross polarized light. Representative regions of interest are selected for analysis. Each ROI is segmented for porosity, a smoothing algorithm is applied, and then the watershed segmentation is performed. Finally, small, artifact pores are removed from the segmented pore system. Applying this process, very good agreement is obtained between mean pore body size in 2D and values obtained from NMR measurements. The results of permeability estimation using a Carmen-Kozeny model with inputs obtained from image analysis of the pore system in thin section fall nearly on a 1:1 line with laboratory measurements of brine permeability over a range of permeabilities from 10 mD to 1D (R2=0.94). In contrast, permeability estimates made using laboratory-based NMR measurements with an application of the SDR model showed substantially more scatter in the predicted permeability with an R2=0.66.