Industrial Fire Journal - Fire & Rescue - Hemming Group Ltd
Home
Menu
Vapour-cloud explosions
Published:  25 September, 2017

A review of major vapour-cloud incidents has been carried out by the US Pipeline & Hazardous Materials Safety Administration and the UK Health and Safety Executive.

The report outlines a new type of explosion and reveals the most significant contributory factors to major vapour-cloud explosions.

Although the report primarily aims to improve understanding of vapour-cloud development and explosion in LNG export plants, its findings are also relevant to gasoline storage depots, tanker terminals, refineries and chemical processing sites.

While the report acknowledges that there are no historical records of LNG vapour cloud explosions in open areas with severity sufficient to cause secondary damage to tanks and pipes, it points out that some LNG sites hold substantial amounts of refrigerant gases and blends containing ethane, propane, ethylene and isobutane.

In addition to comprehensive background and discussion sections on vapour-cloud explosions, the 320-page report reviews 24 major vapour-cloud explosion incidents including Jaipur, Buncefield, Amuay Refinery, Flixborough, San Juan, Brehnam, Newark, Baton Rouge, Big Spring, Pasadena, and Skikda.

The report finds that a high proportion of the major vapour cloud incidents studied occurred in nil/low wind conditions, reflecting the fact that relatively small but sustained leaks can accumulate near the source and build a substantial cloud. In these conditions the cloud spreads through the action of gravity, with a flammable zone reaching >500 m from the source on several occasions. Well away from the source an increase of ground elevation of more than two metres is often sufficient to arrest cloud spread.

The study points out that regular occurrence of severe explosions extending to the whole cloud has been recognised in the years since Buncefield, with a general presumption that this means that such incidents are detonations. However, it states that results of detonation tests cast doubt on this assumption. “There appear to be serious discrepancies between the effect of experimental detonation on a variety of objects and what has been observed at most VCE incidents. For example, normal impact by a detonation typically leaves slender column-like objects with continuous curvature ie plastic deformation distributed along their length rather than being concentrated in a hinge. This unusual type of deformation is caused by the extreme shortness and severity of the impulse associated with a detonation. No objects with this type of deformation have been observed at the sites presumed to have been detonations: Buncefield, Jaipur, Amuay and San Juan.”

The report says that the assumption that has underpinned VCE assessment for the last 30 years, namely that high overpressures are confined to congested areas, should be critically examined: “The data however suggests that severe explosions can progress by a different mechanism: one that has not yet been observed in experimental tests on congestion arrays in gas tents. There is a large gap between the scale of clouds in real incidents and available test data.”

The data, says the report, also suggests that this new type of explosion is episodic in nature, where rapid phases of burning are punctuated by pauses.

To download the report click here.