Propagation of a vapor cloud detonation from a congested area into an uncongested area: Demonstration test and impact on blast load prediction

Abstract

A test was conducted which demonstrates that a detonation wave, once formed due to a deflagration to detonation transition (DDT) within a congested region, will propagate as a detonation from the congested region into an uncongested region. This is the expected behavior based on the general behavior of detonation waves as well as other tests reported in literature. The impact of a detonation wave propagating beyond the congested volume in which it is initiated on the resulting blast load was evaluated parametrically. As would be expected, the impact on the blast load is large for flammable clouds which extend well beyond the congested volume. The test rig was 16.5 m (54 ft) long with the first 9.1 m (30 ft) of the rig length comprised of a congested section 3.7 m (12 ft) in width and 1.8 m (6 ft) high. The congestion was made up of a regular array of vertical circular tubes [6 cm (2.375 in.) diameter, pitch-to-diameter ratio of 4.1, area and volume blockage ratios of 23% and 4.2%, respectively]. The last 7.3 m (24 ft) of the test rig length was completely uncongested. The test rig was configured without any confinement (i.e., no wall or roof sections). A near-stoichiometric ethylene-air mixture completely filled both the congested and uncongested portions of the test rig. Prior testing with a similar rig configuration had shown that this flammable mixture would undergo a DDT within the congested portion of the rig.

Improved far-field blast predictions from fast deflagrations, DDTs and detonations of vapour clouds using FLACS CFD

“…While this method may be appropriate for single free-field blast predictions it will not be applicable for pressures reflected at buildings or for general correction of 3D pressure field, and the approach is thus of limited value. Hydrogen Fraunhofer ICT lane McGill detonation tubes DNV GL Merge A, B, D ethylene experiments (Mercx, 1994) BakerRisk ethylene and hydrogen experiments (Thomas et al, 2013) DNV GL Buncefield propane tree experiments (New Scientist, 2012;Johnson et al, 2014) NIOSH 73 m D ¼ 1.05 m pipe natural gas (97.5% methane) tests (Zipf, 2012;Gamezo et al, 2013), see Fig. 3.…”

Are unconfined hydrogen vapor cloud explosions credible?

“…A detonation (i.e., a DDT) would be predicted for hydrogen releases at medium levels of congestion [16], as would be present within many processing areas. A detonation could propagate into unburnt fuel outside of the congested region [22,23], as was discussed previously.…”

“…Furthermore, a large portion of the flammable cloud is outside the process module. This is relevant since, if the flammable cloud were ignited inside the module and accelerated such that a DDT occurred, the detonation could propagate into the portion of the flammable cloud outside the module such that it could also contribute to the explosion energy .…”

“…In the case of a deflagration, the explosion energy will be essentially limited to the portion of the flammable cloud within the congested/confined volume, with the portion of the gas mixture outside of this volume contributing little to the explosion energy. In the case of a detonation, essentially the entire flammable cloud can contribute to the explosion energy, since the detonation wave can propagate out of the congested region into the unburnt portion of the cloud in the open . The standoff distance is simply the distance between the target of interest (e.g., an occupied building) and the explosion.…”

Effect of inert species on the laminar burning velocity of hydrogen and ethylene

“…In the BST VCE blast load prediction method, a fuel with an LBV equal to or greater than 75 cm/s is considered a high reactivity fuel. BakerRisk has observed deflagration-to-detonation transitions (DDTs) with high reactivity fuels (e.g., ethylene and hydrogen) at moderate levels of congestion in the absence of confinement (Thomas et al, 2003(Thomas et al, , 2010(Thomas et al, , 2013.…”