| Title | Implications for the Accuracy of Porosity and Permeability Estimates Derived from X-ray Computed Tomography in Reservoir Characterization |
|---|---|
| Authors | Mathias NEHLER, Ferdinand STOECKHERT, Erik SAENGER, Jörg RENNER, Rolf BRACKE |
| Year | 2020 |
| Conference | World Geothermal Congress |
| Keywords | X-ray computed micro-tomography, flow-through experiments, porosity, permeability |
| Abstract | Geophysical rock properties such as porosity and permeability are fundamental for the characterization of geothermal reservoirs. The geometry of individual pores and fractures, but also the tortuosity associated with their spatial arrangement control the transport properties of rocks. The flow properties may evolve with time in response to the state of stress or due to the dissolution or precipitation of minerals. Computed tomography using X-ray radiation is a nondestructive imaging method providing insights into the geometrical characteristics of pore networks on the laboratory scale. Monitoring the evolution of rock properties quasi time-resolved during flow-through experiments offers the opportunity to enhance our understanding of flow processes in porous media or fractured media on the microscale. We investigated the porosity of different sandstone varieties representing typical reservoir rocks in standardized laboratory experiments. We compared the porosity estimates gained through the evaluation of X-ray tomography images with the laboratory measurements. The total porosity was determined from the ratio between bulk and matrix density. Effective porosity was independently deduced from water imbibition and gas expansion using an argon pycnometer. The sensitivity of segmentation of the reconstructed images was investigated for (a) resolution of the intensity images, (b) denoising filters (median and non-local means), and (c) different global thresholding methods. Porosity characterization was complemented by flow-through experiments to derive the permeability of low-porosity samples with a single fracture induced artificially or natural. All fractures were oriented parallel to the flow direction. Experiments were conducted at confining pressures between 2.5 MPa and 30 MPa. At each pressure level, effective permeability was determined from steady-state flow based on Darcys law. In addition, we placed the flow-through cell including the samples in a micro-CT scanner to determine the fracture aperture. Employing the cubic law enabled us to calculate a geometrical permeability then. Porosity of the examined sandstones ranged from about 3 % to 25 %. Distinctly bimodal intensity distributions enabled fairly reliable porosity estimates. These intensity distributions are typical for high porous, quartz dominated sandstones. In contrast, low-porosity sandstones exhibited overlapping intensity distributions, which yield highly variable porosity estimates. A resolution exceeding the characteristic length scale of pores is mandatory for accurate porosity estimates. Application of denoising filters does not improve segmentation results. The thresholding algorithms are biased by the amount of unresolved pore space. Resolution is equally significant for the geometrical characterization of fractures, characterized by two adjacent surfaces of large lateral extension at a small distance. The limited resolution of X-ray micro-tomography leads to non-resolved apertures that result in erroneous permeability estimates when using the cubic law. |