Characterization of shale oil and water micro-occurrence based on a novel method for fluid identification by NMR T2 spectrum

Nuclear magnetic resonance (NMR) has become a crucial technique for determining complex fluid occurrence characteristics. However, the overlap of fluid signals limits the application of the one-dimensional (1D) T-2 (transverse relaxation time). Conventional methods are unable to accurately detect fluid signals that overlap entirely and cannot directly determine the petrophysical significance of each component spectrum. Leveraging the additional T-1 (longitudinal relaxation time) information of two-dimensional (2D) NMR T-1-T-2, this study introduced a novel T-2 spectra fluid identification method using the mixed Gaussian function. Six states of T-2 and T-1-T-2 experiments were conducted on shale samples, including as-received (AR), water restoration of AR (WR-AR), water-oil restoration of AR (WOR-AR), oil-washed and dry (dry), oil-saturated (SO), and water-saturated (SW). By introducing the T-2 projection spectrum, linear relationships between the 1D and 2D NMR signal amplitudes of various fluids were established to help separate the overlapping signals in the T-2 spectrum. The results have demonstrated an inconsistent linear correlation between the 1D and 2D NMR signal amplitudes of various fluids, influenced by fluid composition and occurrence state. The conversion coefficient of adsorbed oil is 1.745 times higher than that of bound oil in the SO state. The adsorbed-to-bound ratio conversion coefficient was proposed as an additional input in the mixed Gaussian function to identify overlapping fluids in the oil-water coexistence state. The conversion coefficient of bound water determined by peak area analysis of 1D and 2D NMR before and after water restoration validated the identification results. The pore water altered the occurrence characteristics of oil, reducing the conversion coefficient. Furthermore, the pore-size characteristics of fluid occurrence were described. Compared with the conventional method, the proposed method bridges the gap between 1D and 2D NMR, significantly enhancing the accuracy and reliability of fluid identification by T-2 spectra and providing new insights for NMR fluid evaluation in shale reservoirs.