International audienceThe results of a large test series on vented gas explosions are herein reported (with vessel sizes ranging from 1 to 100 m(3) mixture of methane, propane, and hydrogen with air) The pressure was not only measured internally but also externally The flame propagation was also recorded with a speed camera It is shown that a very significant explosion could result from the formation and ignition of the combustible cloud outside the vent and that, in a number of circumstances, this "external" explosion would overcome the internal one. Surprising results were also obtained showing, for Instance, that near the vent the maximum overpressure may be lower than further downstream This is not because the external explosion is developing Jar downstream but because of the local influence of the chamber pumping the outside atmosphere after the flame went out (thus creating a lower than atmosphere pressure) produced at the vent. The mode of combustion of the external cloud is still outstanding and cannot be described by the traditional outward expanding, flame kernel

Some years ago, one of the authors (Proust, 2015) published the conclusions of a rather large experimental work devoted to the gas flame acceleration down a long pipe. It was concluded that the flame propagation could be represented by a constantly accelerating piston. The acceleration parameter seems to be primarily linked to the expansion velocity of the burnt product. Other parameters seemed of secondary importance questioning in particular the respective roles of the turbulence of the flame and of the instabilities. Further experiments were performed using perfectly smooth and rough tubes (figure 1), varying the diameter of the pipe (150 and 250 mm) and the reactivity of the mixtures (methane-air and hydrogen air at various equivalence ratios). The smooth pipe is transparent enabling a direct visualization of the flame during the flame propagation and a refined resolution of the flame trajectory (in the steel pipes standard flame sensors were used). The pressure was measured at various locations but also the flow velocities in the boundary to try and detect any turbulence development. Only homogeneous and quiescent mixtures were studied and the flame was propagated from a closed ignition end toward the other open end. The results of the parametric study are presented in this paper.

The research activities in the project Assessing the Influence of Real Releases on Explosions (AIRRE) included a unique series of large-scale explosion experiments with ignited high-momentum jet releases directed into congested geometries. The primary objective for the AIRRE project was to gain improved understanding of the effect realistic releases and turbulent flow conditions have on the consequences of accidental gas explosions in the petroleum industry, to develop methodology that can facilitate safe and optimal design of process facilities. This paper presents selected results from experiments involving ignition of a highly turbulent gas cloud, generated by a large-scale, pressurised release of natural gas. The paper gives an overview of the effect on maximum explosion overpressures of varying the ignition position relative to the release point of the jet and a congested region placed inside the flammable cloud, with either a high or medium level of congestion. For two of the tests involving a jet release and the medium congestion rig, the maximum overpressures significantly exceeded those obtained in a quiescent reference test. The paper presents detailed results for selected tests in the campaign. The effects on the relevant explosion phenomena from introducing initial turbulence, initial flow momentum and concentration gradients, as generated by the realistic release, are discussed.

No abstract

No abstract