Application of digital rock physics without segmentation to thermal conductivity estimation of Berea and Shirahama sandstones

• A new Digital Rock Physics method for estimating the thermal conductivity of sandstones was established. • The effectiveness of the segmentation-less method by low-resolution XCT was confirmed. • An appropriate mixing law was identified after comparison with the measurements. Thermal conductivity is one of the most important physical properties of rocks and sediments for the industrial development of energy resources from geological formations, including geothermal energy extraction, enhanced oil recovery, and gas hydrate exploitation, as well as for understanding heat-related phenomena such as crustal heat flow and the subsurface temperature structure of active tectonic regions. Although conventional measurement techniques using probes and rock or sediment samples are well established, it is often impractical to measure all samples of interest because of limited sample availability or the large number of target specimens. In this study, to estimate the thermal conductivity of sandstone samples in a nondestructive, contactless, and efficient manner, this study developed a novel Digital Rock Physics (DRP) method that does not require segmentation of X-ray Computed Tomography (XCT) images. This method enables the use of XCT data with lower spatial resolution than existing DRP approaches by calculating intermediate physical properties for voxels representing multiple mineral or fluid phases. Raw XCT images were processed via continuous functions to convert CT values to density and subsequently to porosity; mixing laws were then applied to derive thermal conductivity, and arithmetic averaging was used to estimate specific heat capacity. Following numerical simulations of heat conduction, the DRP approach without segmentation was demonstrated to accurately estimate the effective thermal conductivity of two types of sandstones with different mineralogy and structures under dry and water-saturated conditions, in contrast to existing studies of thermal conductivity estimation. The proposed method enables the use of low-resolution XCT images for thermal conductivity estimation and significantly expands the availability of reliable thermal conductivity data efficiently and cost-effectively.