Study on the hydrogen leakage and diffusion behavior of long-distance high-pressure buried pure-hydrogen pipelines

As a cornerstone of large-scale hydrogen energy systems, high-pressure long-distance buried hydrogen pipelines are critical for mitigating regional resource imbalances and integrating production, storage, and utilization. Despite their importance, leakage risks induced by corrosion, external disturbances, and construction flaws can trigger hydrogen diffusion and soil accumulation, posing severe fire and explosion threats to public safety. While understanding leakage-induced diffusion dynamics is essential for hazard zone delineation and emergency response optimization, current models often oversimplify multi-parameter interactions. This study employs numerical simulations grounded in multi-phase flow theory to characterize hydrogen diffusion in soil environments, prioritizing parameters with industrial failure relevance: soil properties (porosity, viscous/inertial resistance coefficients), pipeline pressure, burial depth, and leakage aperture. Quantitative analysis demonstrates that these parameters collectively govern leakage mass flow rates. A time-dependent model quantifies the duration for ground-level hydrogen accumulation to reach the lower flammability limit (LFL, 4 % vol). Building on this temporal framework, an adaptive multi-source data fusion model is proposed, leveraging nonlinear regression with dynamic parameter weighting to decode variable inter-dependencies. Validated against simulation data, the model achieves a mean absolute error of 6.48 % and a determination coefficient (R2) of 0.964, outperforming conventional static-weight approaches. These advancements establish a methodology for realtime risk mapping and adaptive emergency strategy formulation in buried hydrogen infrastructure management.