A study of detonation propagation and diffraction with compliant confinement

Abstract

Previous computational studies of diffracting detonations with the ignition-and-growth (IG) model demonstrated that, contrary to experimental observations, the computed solution did not exhibit dead zones. For a rigidly confined explosive it was found that while diffraction past a sharp corner did lead to a temporary separation of the lead shock from the reaction zone, the detonation re-established itself in due course and no pockets of unreacted material remained. The present investigation continues to focus on the potential for detonation failure within the IG model, but now for a compliant confinement of the explosive. The aim of the present paper is two-fold. First, in order to compute solutions of the governing equations for multi-material reactive flow, a numerical method is developed and discussed. The method is a Godunov-type, fractional-step scheme which incorporates an energy correction to suppress numerical oscillations that occur near material interfaces for standard conservative schemes. The accuracy of the solution method is then tested using a two-dimensional rate-stick problem for both strong and weak confinements. The second aim of the paper is to extend the previous computational study of the IG model by considering two related problems. In the first problem, the corner-turning configuration is re-examined, and it is shown that in the matter of detonation failure, the absence of rigid confinement does not affect the outcome in a material way; sustained dead zones continue to elude the model. In the second problem, detonations propagating down a compliantly confined pencil-shaped configuration are computed for a variety of cone angles of the tapered section. It is found, in accord with experimental observation, that if the cone angle is small enough, the detonation fails prior to reaching the cone tip. For both the corner-turning and the pencil-shaped configurations, mechanisms underlying the behaviour of the computed solutions are identified.