Numerical simulation and parameter analysis of NMR T2–D distributions of tight sandstone saturated with a gas–water two-phase fluid

Abstract

Two-dimensional nuclear magnetic resonance (NMR) transverse relaxation time–diffusion coefficient (T2–D) distributions can simultaneously provide with pore-fluid parameters of the relaxation and the diffusion, making it possible to distinguish fluid types based on the differences in the two parameters. Hence, T2–D distributions are advantageous over one-dimensional T2 distributions for the identification of fluids because the T2 distributions of different fluids sometimes overlap. In this paper, we simulated NMR measurements with diffusion-editing pulse sequences in tight sandstone saturated with a gas–water two-phase fluid using the random-walk method, and the simulated echo trains were inverted to obtain the T2–D distributions. The effect of fluid saturation on the T2–D distributions of tight sandstone was analyzed, and we studied the effects of the observation parameters–echo spacing TE, magnetic field gradient G, and the group of echo spacing TE1–in diffusion-editing pulse sequences and data signal-to-noise ratios (SNRs) on T2–D distributions of tight sandstone with different gas saturations. The simulated results showed that the water signal in T2–D distributions of tight sandstone gradually moved to a lower diffusion coefficient and a shorter transverse relaxation time with decreasing water saturation, but the gas signal remained fixed in position. With increasing echo spacing TE or magnetic field gradient G, the gas signal in the T2–D distributions moved to shorter transverse relaxation times. The effects of echo spacing TE or magnetic field gradient G on the position of the water signal in the T2–D distributions can be ignored. T2–D distributions for the hybrid echo spacing TE1 group is much better at reflecting pore-fluid information than ones with the short echo spacing TE1 group or the long echo spacing TE1 group. When data SNRs are low, identifying pore fluids based on T2–D distributions is not reliable due to the partial overlap of the water and gas signals.