Simulation of nuclear magnetic resonance response based on 3D CT images of sandstone core

Abstract

In recent years, nuclear magnetic resonance (NMR) logging has become increasingly prevalent in the characterization of rock properties such as porosity, permeability, saturation, and pore size distribution. However, interpreting such properties accurately from NMR logging data is challenging for some reservoirs. In particular, the impact of salinity, viscosity, and saturation on NMR measurements is not always clear, which can lead to inaccuracies in the resulting data interpretation. Properly accounting for these factors is essential in order to obtain accurate and reliable measurements for effective characterization of subsurface formations. This study utilized a random walk technique to simulate the NMR response of homogeneous sandstones using 3D CT images and conducted a sensitivity analysis under various salinity, crude oil viscosity, and water saturation conditions. The results indicate that the T2 relaxation time slightly shifts toward the short relaxation direction as the salinity of the formation water increases. In addition, longer echo intervals result in a more significant forward shift in the T2 value than shorter intervals. Whereas, for crude oil, the T2 relaxation time becomes shorter as its viscosity increases. Furthermore, the effect of echo interval on the forward T2 shift is less pronounced for crude oil than it is for formation water. Under water wet conditions, the T2 spectrum of crude oil exhibits a peak at the volume relaxation position. As the water saturation decreases, the left two peaks in the T2 spectrum shift toward shorter relaxation times. Under oil wet conditions, the T2 spectrum exhibits a complex three-peak structure. The method provides a physical basis for interpreting NMR macroscopic responses, and the simulated NMR responses can help identify fluids in reservoirs.