Nuclear magnetic resonance (NMR) relaxation measurements provide powerful, non-invasive probes of porous media microstructure, yet their interpretation remains challenging due to the complex interplay of surface relaxation, diffusion coupling, internal magnetic field gradients, and heterogeneous relaxation regimes. Traditional interpretation frameworks rely on simplifying assumptions—constant surface relaxivity, weak coupling, and fast diffusion—that often break down in realistic reservoir rocks. This presentation demonstrates how integrating high-resolution imaging with physics-based forward modeling enables robust extraction of intrinsic physical properties and advanced interpretation of NMR responses.
A workflow is presented combining micro-CT imaging, random walk simulations, and Bayesian optimization to bridge the gap between digital rock physics and NMR petrophysics. The forward modeling approach utilizes random walk algorithms on high-resolution segmented images to naturally incorporate structural heterogeneity and diffusive motion without limiting assumptions. This enables quantitative assessment of effects like diffusion coupling between different relaxation regimes and the influence of internal gradients at varying field strengths.
Applications are illustrated for Bentheimer sandstone using a Bayesian optimization framework. First, NMR T2 relaxation measurements are matched by simulation and surface relaxivity of quartz and effective relaxation and effective diffusion properties of clay regions are simultaneously extracted. Second, the framework is applied to dual T1 and T2 relaxation measurements extracting 5 parameters. T1 and T2 surface relaxivities of quartz, the effective relaxation times and effective diffusion coefficient of clay regions are determined simultaneously. The approach is then extended to T2 measurements at two different temperatures, extracting surface relaxivities, effective clay relaxation times and effective diffusion coefficients simultaneously.
Finally, I will discuss an analysis of internal gradient effects and microstructure distributions on an artificial “soil” sample at fuall and partial saturation.
References
[1] R. Li, I. Shikhov, C.H. Arns, Phys. Rev. Appl. 15 (5), 054003 (2021)
[2] R. Li, I. Shikhov, C.H. Arns, Water Resources Research 58 (9), e2021WR031590 (2022)
[3] R. Li, I. Shikhov, C.H. Arns, E3S Web of Conferences 367, 01002:1-9 (2022)
[4] C.H. Arns & I. Shikhov, New Developments in NMR: NMR in Plants and Soils 37, 58-84 in “New Developments in NMR: NMR in Plants and Soils”, Royal Society of Chemistry, 2025.