High‐Precision Tracking of Sandstone Deformation From Micro‐CT Images. Issue 9 (14th September 2021)

• Record Type: Journal Article
• Authors: Liang, Jiabin; Lebedev, Maxim; Gurevich, Boris; Arns, Christoph Hermann; Vialle, Stephanie; Glubokovskikh, Stanislav
• Is Part Of: Journal of geophysical research. Volume 126:Issue 9(2021)
• Page Start: n/a
• Page End: n/a
• Publication Date: 2021-09-14
• Subjects: Geomagnetism -- Periodicals; Geochemistry -- Periodicals; Geophysics -- Periodicals; Earth sciences -- Periodicals; 551.1
• DOI: 10.1029/2021JB022283
• Languages: English
• ISSNs: 2169-9313

Abstract

Abstract: Elastic properties of sandstones can be highly pressure dependent. However, the pressure‐induced deformation of intact sandstones in X‐ray micro‐Computed Tomography (micro‐CT) images are below the resolving power of conventional image analysis methods. Also, most digital rock physics studies of the elastic properties operate with micro‐CT images acquired at ambient pressure. Here, we perform a comprehensive analysis of an intact sandstone sample with pronounced stress‐sensitivity. A purpose‐built X‐ray‐transparent pressure cell enables scanning micro‐CT images at confining pressures of up to 36 MPa. To detect the small pressure‐induced changes from these images, we design an overall deformation detection workflow with slice‐searching technique. We apply this workflow to a set of micro‐CT images of a 2 × 2 × 2 mm sample scanned at several pressures with a voxel size length of ∼1 μm. The results give the static modulus of 2.78 GPa, which is ∼0.5 times of the dynamic modulus derived from ultrasonic measurements. This ratio is consistent with literature data, showing the accuracy of the deformation detection workflow. Images scanned at different pressures are then segmented into two‐phase labels. The porosity change detected from segmented labels is consistent with the value derived from static modulus using poroelastic theory. The two‐phase segmented images are utilized to simulate the elastic properties with a finite element method. Most of the pressure dependency of elastic moduli shown in ultrasonic measurements is caused by closure of compliant pores at grain contacts, which cannot be resolved by the micro‐CT. The accuracy improvement of the estimated elastic properties from images scanned at higher pressures is negligible. Plain Language Summary: X‐ray micro‐Computed Tomography (micro‐CT) technique is the main tool for imaging three‐dimensional rock microstructure. Often, researchers attempt to substitute the real laboratory measurements on rock samples ‐ time‐consuming and expensive ‐ by numerical simulations of diverse physical processes in these images. However, such an approach ignores the effect of stress on the rock properties, because micro‐CT images are usually scanned at ambient pressures. This study estimates the stress‐induced changes on the images of a porous sandstone. To this end, we acquired a set of images of a sandstone sample at different confining pressures with a purpose‐built X‐ray‐transparent pressure cell and accompanied the images with a comprehensive laboratory data set. Standard algorithms for time‐lapse analysis of rock images failed at our data set, because the deformations are at the limit of the algorithms' sensitivity, hence, we designed a novel deformation detection workflow. The subsequent numerical elastic simulations using these images, ultrasonic measurements and petrophysical experiments suggest that the main stress effects are concentrated in thin intergranular contact that may not be resolved directly by the modern CT‐machines. Hence, the effect of stress should be estimated using indirectly from the analysis of multiphysics measurements and the images of sandstones may be taken at ambient pressure. Key Points: Micro‐Computed Tomography (micro‐CT) images of a sandstone are scanned in a X‐ray transparent pressure cell at confining pressures of up to 36 MPa An overall deformation detection workflow is designed to track the sandstone deformation from micro‐CT images The differences of the computed moduli based on scans at different pressures are demonstrated