Analysis and validation of integral pool spreading model of LNG spills on concrete
Conference Paper
Overview
Research
Identity
Additional Document Info
View All
Overview
abstract
2018 IChemE. When a loss of primary containment of liquefied natural gas (LNG) occurs on the ground, a pool, that simultaneously spreads and vaporizes, is formed posing cryogenic, asphyxiating, and flammable hazards to its surrounding. Determining the pool size and vapour generation upon release play key roles in the accuracy of dispersion and consequence models. This work focuses on LNG source term model validation through the evaluation of an existing model with a newly generated data. A field-scale experimental setup was designed to study the pool temperature, pool spreading and heat flux from the concrete following a release of cryogenic liquid. In this work, liquid nitrogen (LIN) was used as a safer analogue to LNG as it is a non-toxic non-flammable cryogen. The experiments were carried out inside a 651.2 m (width length depth) pit. The Gas Accumulation over a Spreading Pool (GASP) model was implemented and used to predict the vaporization and pool spreading rates of the experiment and a spill from selected literature. The model assumes that the pool boils until it completely vaporizes, meaning that there is no switch to the evaporation regime, which is consistent with obtained experimental results. The pool radius estimated by GASP was found to be relatively close with the experimental data, but it did not simulate the pool growth as consistently as the simulation with Moorhouse and Carpenter (1986) data. Vaporization inside the discharge hose was speculated to have decreased the discharge rate of LIN onto the ground, causing the pool to spread slower at the beginning of the spill and thus not match the input data to the model. It is also possible that the model does not simulate slow spills well, as reproduced by the GASP simulations of Nguyen et al. (2015)'s LIN spill. The accuracy of Fourier's one-dimensional conduction equation for modelling conduction through the concrete was also tested by comparing the predicted temperature with experimental temperature. The study showed that the model is sufficient to predict heat transfer inside the concrete as long as the substrate thermal properties change as a function of temperature.
