Computational Fluid Dynamics Analysis of Liquefied Natural Gas Dispersion for Risk Assessment Strategies

Abstract

Liquefied natural gas (LNG) spill events can create hazardous vapour clouds due to the dense gas behaviour of evaporated LNG. Accurate prediction of LNG vapour dispersion is critical for risk assessment and site planning at LNG facilities. This study employs computational fluid dynamics (CFD) simulations to model dense gas dispersion of LNG, accounting for gravity-driven spreading, time-dependent downwind and crosswind dispersion, and terrain effects.

The CFD model was validated using experimental data from the Burro series of large-scale LNG field tests and compared with the integral model DEGADIS. CFD predictions compared more favourably with experimental data than DEGADIS, with an average relative error in maximum downwind gas concentration of 19.62%. The validated model was applied to risk assessment for the most-likely-spill scenarios at LNG stations in accordance with NFPA 59A standards, examining gas dispersion behaviour around impounding dikes. Simulations revealed that while dikes had minimal effect on affected areas for elevated releases (e.g., from tank tops), they significantly influenced wind velocity fields and generated swirl that confined the dispersion cloud inside the dike, retaining approximately 75% of dispersed vapour in most scenarios. These findings demonstrate the utility of CFD as a tool for designing LNG plant layout and site selection to minimise hazard exposure.

@article{sun2013jolpitpi,
  title         = {Computational Fluid Dynamics Analysis of Liquefied Natural Gas Dispersion for
                  Risk Assessment Strategies},
  author        = {Sun, Biao and Utikar, Ranjeet P. and Pareek, Vishnu K. and Guo, Kaihua},
  year          = 2013,
  journal       = {Journal of Loss Prevention in the Process Industries},
  volume        = 26,
  number        = 1,
  pages         = {117--128},
  doi           = {10.1016/j.jlp.2012.10.002},
  issn          = {09504230},
  url           = {https://linkinghub.elsevier.com/retrieve/pii/S095042301200157X}
}