Determination of T2 Cutoff Values for Complex Reservoirs in NMR Logging Based on Spectrum Morphology Analysis

The T2 cutoff value is a critical parameter in nuclear magnetic resonance (NMR) logging for evaluating reservoir saturation. In conventional sandstone reservoirs, a fixed T2 value (33 ms) is typically adopted as the cutoff threshold. However, this empirical value proves inadequate for the complex Shahejie Formation reservoirs in the Jinzhou Oilfield, Bohai Bay Basin, which exhibit heterogeneous lithology, strong heterogeneity, and diverse pore structures. Core experiments reveal that the T2 cutoff values in this area are dynamic (generally <33 ms) and do not converge to a fixed value, rendering the conventional 33 ms threshold inconsistent with actual conditions. To address this issue, this study proposes an automated T2 cutoff determination method based on NMR T2 spectrum morphology by integrating core experimental data and logging information. First, the fully water-saturated core NMR T2 spectrum from 24 samples are classified into three distinct morphological types: single-peak macropore type (dominant peak at larger pores), single-peak micropore type (dominant peak at smaller pores), and bimodal type. Subsequently, normal distribution functions are employed to fit the movable water spectrum for single-peak macropore types, bound water spectrum for single-peak micropore types, and bound water spectrum for bimodal types, respectively. Based on the reconstructed T2 spectrum, bound water saturation is calculated, and the T2 cutoff value is iteratively optimized using a progressive approximation algorithm. The primary advantage of this method lies in its capability to directly derive T2 cutoff values from the intrinsic features of NMR T2 spectrum. Validation results demonstrate that the proposed method achieves high accuracy: the average absolute error between calculated T2 cutoff values and core-derived experimental values is 1.1 ms, while the absolute error in bound water saturation estimation remains below 5.0%. This approach significantly enhances the reliability of NMR logging in evaluating complex reservoirs, providing a robust technical solution for saturation characterization in heterogeneous formations with diverse pore systems.