An Automated Method for Dual T 2 Cutoff Determination and Permeability Prediction in Fine-Grained Hydrate-Bearing Sediments

The dynamic evolution of permeability during depressurization-driven production plays a critical role in determining the deliverability of natural gas hydrate (NGH) reservoirs. The reliability of nuclear magnetic resonance (NMR)-based permeability models fundamentally hinges on the accurate determination of the transverse relaxation time cutoff ( T 2 cutoff, T 2C ). However, conventional empirical approaches for T 2C determination perform poorly in fine-grained sediments and hydrate-bearing formations, posing a major bottleneck to the broader application of NMR techniques. To overcome this limitation, this study introduces a fully automated and physically interpretable framework for Automated Dual T 2 Cutoff Determination via Wavelet-Gaussian Decomposition, termed ADWD. This method combines multiscale wavelet peak detection with multilog-Gaussian spectral fitting to objectively identify clay-bound and free-fluid T 2C values. Building on these objectively derived fluid partitions, we develop a physically interpretable permeability model ( K ADWD ) that integrates effective flow porosity and bound porosity to reflect pore-structure controls on fluid transport, and field validation shows that our model significantly outperforms classical Coates and SDR models in both hydrate-bearing and hydrate-free intervals. In addition, an ADWD flow-through critical transition criterion is proposed to facilitate the recognition of critical low-permeability transitions. Overall, this study overcomes the applicability limitations of traditional T 2C selection and permeability modeling, providing a more reliable petrophysical evaluation tool for assessing the development potential of NGH and other unconventional reservoirs, with strong implications for both theoretical advancement and engineering practice.