An Automated Petrographic Image Analysis System: Capillary Pressure Curves Using Confocal Microscopy

Abstract

Abstract Laboratory mercury injection capillary pressure (MICP) analyses are key inputs into reservoir simulators. Although laboratory MICP measurements are accurate, they are slow, expensive, not friendly to the environment, destroy samples, and are not useful for pore-throats larger than 100 microns. This paper presents a fast and accurate automated petrographic image analysis system, used to compute pore-body and pore-throat size distributions and simulated capillary pressure curves. Input includes high-resolution petrographic images, such as those from confocal microscopy (0.25-microns resolution), and laboratory-derived porosity measurements. Our system can be used in conjunction with laboratory Special Core Analysis (SCAL) to obtain more accurate representations of reservoir properties. To enable us to compute numerical SCAL measurements, we introduce the Expanding Flow Model (EFM) to show fluid-flow behavior inside pores. This model identifies the largest attached pore-throat as the throat that controls fluid flow in, or out, of a particular pore-body. Pore-throats are extracted using a novel image-analysis technique, based on watershedding algorithms and differences between processed images. We then compute: (a) pore-body and pore-throat size distributions, (b) numerical SCAL in the form of simulated capillary pressure curves for pore bodies and pore-throats, (c) representative element areas (REAs), which are the smallest areas that statistically capture heterogeneity, and (d) values that represent image quality to measure confidence in the results. We have analyzed a number of samples with different porosity and permeability characteristics. Simulated capillary pressure results show a close match to laboratory data if good-quality representative images and accurate laboratory-derived porosity values are used. Rocks with lower porosity tend to produce better results. Simulated results match raw laboratory data, and eliminate closure-correction factors used in Thomeer capillary-pressure analyses.