Simulation of nuclear magnetic resonance response based on the high-resolution three-dimensional digital core

Abstract

Numerical simulation of nuclear magnetic resonance (NMR) can simulate experimental scenarios and quantify the impact of each factor on the physical characteristics. However, general simulation methods lack authentic pore structure information and fail to accurately model the complex geometry of rocks. High-resolution digital rock cores can effectively reflect pore structure. In this paper, a high-resolution digital core of Berea sandstone is taken as the research object, the pore parameters of the core (e.g., pore volume and surface-to-volume ratio) are quantified, and the 12 529 pores extracted from the three-dimensional digital core are statistically analyzed. Subsequently, the pores are classified based on their surface-to-volume ratio and volume. After the simulation parameters are calibrated by the experiments, NMR response of different water-saturated pores is simulated. Finally, the NMR response of the core with different oil saturation is simulated. We find that the distribution of pore quantity in terms of volume and surface area both follows a power function. There is a strong correlation between pore volume and surface area, especially for smaller pores. The T2 (transverse relaxation time) spectrum can generally reflect the volume distribution, but it may not accurately reflect the volume distribution when the pores are large. We also observe that the spectrum peak reflecting oil bulk relaxation is positioned to the left of the peak of the oil bulk relaxation due to the combined effects of surface relaxation of residual water and diffusion relaxation. These simulation results provide a physical basis for interpreting NMR measurements and can help identify fluids in reservoirs.