Predicting Graphite/Metals Explosion Severity: from Experimentation to Modelling

Abstract

During decommissioning of UNGG (Uranium Natural Graphite Gas) nuclear reactors and wastes reconditioning operations, mixtures of graphite and metal dusts can be encountered. In this perspective, an extensive experimentation has been realized on graphite/magnesium and graphite/iron mixtures to determine their ignition sensitivity and explosivity. The influence of parameters such as the metal content, the ignition energy, the storage humidity, the particle size distribution and the initial turbulence has been considered. It appears that the introduction of such metals can strongly changes the behavior of graphite powders: the maximum rate of pressure rise, for example, increases of roughly 355 % comparing to that of simple graphite when adding 30 % wt. of magnesium dust. By means of a vertical explosion tube, an estimation of the laminar burning velocity has been deduced from the apparent flame velocity through the image processing of explosion videos. This experimental study has provided a solid background to better understand these explosion phenomena, to characterize the mixtures behavior by semi-empirical correlations and also to develop and validate a model predicting the severity of a dust cloud in a confined space. Initially conceived for graphite dusts, this model aspires to provide a description of the flame front propagation, leading to the prediction of the explosivity parameters. The dust cloud combustion has been regarded as the combustion of isolated graphite particles, taking into consideration the oxidation kinetic mechanisms and its rate-limiting steps. A finite volume method has been used for the resolution of the transient conservation equations applied to a one–dimension scheme.