Parametric study of the explosivity of graphite-metals mixtures

The explosivity of dust clouds is greatly influenced by several parameters which depend on the operating conditions, such as the initial turbulence, temperature or ignition energy, but obviously also on the materials composition. In the peculiar case of a mixture of two combustible powders, the physical and chemical properties of both dusts have an impact on the cloud flammability and on its explosivity. Nevertheless, no satisfactory 'mixing laws' predicting the mixture behavior are currently available and the composition variable to be considered for such models greatly depend on the safety parameters which have to be determined: from volume ratios for some thermal exchanges and ignition phenomena, to surface proportions for some heterogeneous reactions and molar contents for chemical reactions. This study is mainly focused on graphite/magnesium mixtures as they are encountered during the decommissioning activities of UNGG reactors (Natural Uranium Graphite Gas). Due to the different nature and reactivity of both powders, these mixtures offer a wide range of interests. Firstly, the rate-limiting steps for the combustion of graphite are distinct from those of metals (oxygen diffusion or metal vaporization). Secondly, the flame can be thickened by the presence of radiation during metal combustion, whereas this phenomenon is negligible for pure graphite. Finally, the turbulence of the initial dust cloud is modified by the addition of a second powder. In order to assess the explosivity of graphite/magnesium clouds, a parametric study of the effects of storage humidity, particle size distribution, ignition energy, and initial turbulence has been carried out. In particular, it was clearly demonstrated that the turbulence significantly influences the explosion severity by speeding up the rate of heat release on the one hand and the oxygen diffusion through the boundary layer surrounding particles on the other hand. Moreover, it modifies the mean particle size and the spatial dust distribution in the test vessel, impacting the uniformity of the dust cloud. Thus, the present work demonstrates that the procedures developed for standard tests are not sufficient to assess the dust explosivity in industrial conditions and that an extensive parametric study is relevant to figure out the explosive behavior of solid/solid mixtures subjected to variations of operating conditions.