CoreDNA: Combining the analysis of ultra-high resolution images with continuous direct measurements to identify rock types
Tanguy Lhomme, Christophe Germay
Epslog’s CoreDNA solution combines a selection of transdisciplinary, high resolution, non-destructive measurements on whole cores that enable an early yet objective and comprehensive description of cores and the rapid estimation of properties of formations a handful of days after opening core barrels.
Whole cores, which may still be in their half-open liners, are mounted on one unique tabletop test bench and submitted to a battery of tests, all sharing the same depth reference and compatible resolution ranges.
Technologies including ultra-high resolution pictures, pXRF elemental composition, grain size analysis, but also the direct measurement of geomechanical properties such as strength and acoustic velocities, are all deployed along the entire core on the same 3cm wide mini-slab surface.
Ultra-high resolution panoramic pictures (1.8µm/px) are processed to extract textural and colour features but also continuous grain size distribution from wavelet analysis. The grain size distribution profile calculated from the images are backed-up by analysis of 3D topographical images accurately measured with a laser scan.
Results of these fast tests (3ft per hour) are analysed real-time and turned into high resolution, continuous profiles of properties (petrophysical, geomechanical and geochemistry) fed into (unsupervised) machine learning algorithms for the automated identification of lithofacies, enabling the detailed understanding of reservoir architecture and the design of tailored plug selections and the programing of subsequent steps in core analysis programs, even remotely.
Such a detailed and comprehensive knowledge of the distributions of core properties under one unique format for all discipline eases interdisciplinary core analysis work, from the QCing of standard tests (Routine Core Analysis, Rock Mechanical Test) to the upscaling of core data and the calibration of robust predictive models from well logs.
A number of case studies on reservoirs from different regions all underline the benefit of this disruptive technology in core analysis standards when run prior (i) taking any sample (plugs and preserved) and (ii) slabbing the cores.
