LNG - Quantifying Major Accident Risks

Abstract

This reference is for an abstract only. A full paper was not submitted for this conference. Abstract Throughout much of the world, the demand for natural gas has been steadily increasing and is projected to double in the next 15 years. A large portion of the demand can only be met through long-distance marine transport, and for economic reasons, ships carry the product as liquefied natural gas (LNG). While the LNG industry lays claim to an exceptional safety record, safety concerns have threatened to delay or even prevent the expansion as regulators and general public remain unconvinced. In order to identify the major hazards involved it is useful to review the physical processes which take place in the spillage of LNG on water. In this work we have concentrated on one of the crucial factors that determines the formation and the future behaviour of the hazardous vapour cloud, namely the rate of vaporization of the LNG. A model has been developed for estimating the rate of vaporisation of LNG mixture spreading on confined and unconfined water surfaces. In the context of this presentation the model is used to illustrate the importance of correctly describing the thermodynamics of LNG vaporisation and to examine the influence of chemical composition of LNG on its vaporisation rate. The detailed results indicate that the vaporisation rate of LNG mixture is markedly different to that of pure methane. The difference can be attributed primarily to the contributions of the direct and indirect component of the total, differential, isobaric latent heat to the boiling process. For LNG, as the liquid mixture gets rich in ethane, the total, differential, isobaric latent heat increases rapidly, leading to a large decrease in the vaporisation of LNG compared to pure methane. The overall results suggest that treating an LNG spill as a pure methane spill results in underestimation of the total spillage time and in qualitatively wrong dynamics of the rate of vapour formation. Furthermore, we report on the conditions under which the water surface will freeze. It was concluded that when LNG spills on a confined, shallow-water surface the surface temperature of water will decrease rapidly leading to ice formation. The formation of an ice layer, that will continue to grow for the duration of the spill, will have a profound effect upon the vaporization rate. The decreasing surface temperature of ice will decrease the temperature differential between LNG and ice that drives the heat transfer and will lead to a change of the boiling regime. The overall effect would be that the vaporization flux would first decrease during the film boiling; followed by an increase during the transition boiling and a steady decrease during the nucleate boiling. The developed model reduces the uncertainty associated with predicting the crucial vaporization rate of LNG and allows for a more secure quantification, for the purposes of risk assessment, of all hazards associated with accidental or terrorist-related LNG release.