Accidental gas explosions are a recognised hazard in process industries but they are also common in residential buildings. Whilst process plants have specifically designed vent reliefs to limit the enclosure damage, in homes a similar effect is achieved due to the presence of doors and windows whose failure often protects the building. There are empirically based correlations for predicting overpressure and for vent sizing, however these are limited in application to simple enclosures. In practice, enclosures have interconnected spaces which would potentially increase the flame acceleration considerably. In this paper we present the results of full scale natural gas layer tests in a twin chamber, which consisted of two 22 m3 enclosures connected by an open doorway. Layered natural gas/air mixtures of 8, 10 and 12% by volume, were ignited at the rear of one of the chambers. Explosion relief was provided by vent openings of 2.48, 1.49 or 0.74 m2 on the far walls of both chambers. With tests with equal large vents on each of the chambers, the dominant influence was the external explosion. The maximum overpressure was produced by tests involving a 12% natural gas concentration. The use of a smaller vent in the adjoining enclosure had a significant effect on the maximum overpressure and the mechanism of the explosion development. However, altering the size from 1.49 m2 to 0.74 m2 had little overall effect. This was largely due to the greater generation of turbulence and the venting process which predominantly occurred via the doorway and through the ignition-chamber vent opening. The use of a smaller vent in the ignition enclosure also altered the manner in which the explosion developed. A venting driven ‘jetting’ expanding flame, propagated into the adjoining enclosure and towards the far vent opening, generating the dominant pressure peak in these type of tests.
CORE (COnnecting REpositories)
CORE (COnnecting REpositories)