SIMULATING DUST EXPLOSIONS WITH THE FIRST VERSION OF DESC T. SKJOLD 1,2 , B. J. ARNTZEN 2 , O. R. HANSEN 1 , O. J. TARALDSET 2 , I. E. STORVIK 1 and R. K. ECKHOFF 2 1 GexCon AS, Bergen, Norway 2 University of Bergen, Department of Physics and Technology, Bergen, Norway D ESC is a new CFD-code that is being developed for simulating dust explosions in complex geometries. In the methodology followed, dust – air mixtures are ignited to deflagration in laboratory test vessels to provide information on fundamental flame propagation parameters. These quantities are subsequently used as input for combustion models in the CFD-code. Experiments were performed with maize starch in a 20-litre explosion vessel, the laminar burning velocity was extracted from the experimental pressure–time curve, and subsequently used to predict what might happen if the same mix- tures would explode in larger geometries, namely, a vented silo, and, a system involving two interconnected vessels. The results presented in this paper are preliminary and serve as a proof of principle. Uncertainties and gaps in knowledge are identified in the light of this approach, and future challenges discussed. Keywords: DESC; dust explosion; modelling. INTRODUCTION Dust explosions represent a hazard to both personnel and equipment in industries that handles combustible powders. Primarily, one seeks to reduce the risk posed by dust explosions by preventing them from taking place, either by eliminating all possible ignition sources, or by avoiding the formation of combustible dust clouds altogether. How- ever, if the possibility of an explosion cannot be ruled out, measures for minimizing damage have to be considered. In some cases, the enclosure containing the combustible dust – air mixture can be made strong enough to withstand an internal explosion. In this case, only the maximum explosion pressure is needed as design parameter. More often, however, the enclosure will not be able to withstand the total explosion load, and mitigatory measures, such as venting, isolation and automatic suppression, must be implemented in the design. Safe dimensioning of mitigating measures usually requires adequate knowledge about the burning rate of dust clouds in actual process situations. Traditionally, the reactivity of explosive dust clouds is characterized by the K St value, defined as the maximum rate of pressure rise determined in constant volume explosion vessels, multi- plied by the cube root of the vessel volume. Bartknecht (1971) presented experimental results that indicated that the so-called cube-root-law could be used to scale turbulent dust explosions between vessels with volumes larger than 40 litres. Results presented by Siwek (1977, 1988) suggested that a 20-litre spherical vessel could produce K St values that agree with data from the standardized 1-m 3 ISO-vessel (ISO, 1985). However, the cube-root-law can only be regarded as a valid scaling relationship under hypothetical circumstances (Eckhoff, 1984; Bradley et al., 1988; Dahoe et al., 1996, 2001a), such as: near spherical vessels, central point ignition, spherical propagation of a thin flame, the same mass burning rate in both vessels, and so on. Several so-called integral balance models have been introduced in order to overcome some of the limit- ations with the cube-root-law (Dahoe et al., 2001a). Although such models are limited to relatively simple geo- metries, they may prove useful for estimating fundamental flame propagation parameters of combustible mixtures. Although acceptable levels of risk usually can be achieved with design according to experience, empirical formulas, or existing guidelines; better prediction of flow, flame propagation and pressure build-up in complex geo- metries can be accomplished by computational fluid dynamics (CFD). Solutions based on CFD have much higher potential for being optimised with respect to risk/ cost, especially for complex geometries, compared to sim- pler methods. It appears that the new ATEX directives have created a demand for a more differentiated approach to design of explosion mitigation systems in Europe; a prop- erly validated CFD-code for dust explosions will be a most useful tool to meet this need. The code could be useful both with respect to risk assessments required by the user directive (ATEX 1999/92/EC, 1999), and for  Correspondence to: Mr T. Skjold, GexCon AS, Fantoftveien 38, 5892 Bergen, Norway. E-mail: trygve@gexcon.com 151 0957–5820/05/$30.00+0.00 # 2005 Institution of Chemical Engineers www.icheme.org/journals Trans IChemE, Part B, March 2005 doi:10.1205/psep.04237 Process Safety and Environmental Protection, 83(B2): 151–160