Detonation modelling of non-ideal high explosives

摘要

High explosives (HE) are used in many fields where the energy liberated by the combustion process is used to perform useful work. High explosives normally burn via a detonation; a supersonic wave consisting of a shock wave coupled to chemical energy release. Detonations in conventional HE (CHE) propagate with a typical velocity of 6 - 8 km/s. Insensitive HE (IHE) and non-ideal HE (NIHE) are of particular interest as they areharder to initiate and thus safer to store and transport. Detonations in IHEs and NIHEs are characterized by longer reaction time and length scales than detonations in CHE. NIHEs are typically characterized by their porous, granular structure. Detonations in NIHEs have lower detonation velocities (4 - 6 km/s) than those in CHEs or IHEs due to their lower initial densities.The short time scales (O(ns ??? ??s)), length scales (O(??m ??? mm)) and the opaque nature of HEs and their products make experimental observations, required to calibrate detonation models for reaction flow modelling, challenging. Currently used reactive burn models assume a two component, mechanically equilibrated mixture of reactants and products. Individual components are modelled with an empirical equation of state (EOS). The set of relations which uniquely determine the mixture-averaged state in terms of the states of the mixture constituents, the mixture closure conditions, are also often of a pressure-temperature equilibrium form. The chemical reaction rate law(s) are mostly based on preconceptions of how a detonating HE burns.Typically, such engineering style models are complex and contain a large number of fitting parameters that are calibrated in some form to a limited set of experimental data. Minimal attention has been devoted to the physical and mathematical implications of the fitting process and reactive burn model structure (such as the choice of closure condition) to issues such as detonation stability and interacting oblique shock structure. For a well-posed reactive burn model, such properties should be understood.A majority of this thesis research is devoted to formulating and studying the shock and detonation properties of reactive burn models based on the use of stiffened-gas (SG) equations of state. A SG model allowsan appropriate initial sound speed of a material to be set, an important improvement over ideal gas models when applied to condensed phase reactive burn models. Due to its relative simplicity, a semi-analytical understanding of reactive burn models based on the use of SG EOS models for its constituent components can be obtained. Furthermore, changes in physical aspects of the reactive burn model, such as detonationstability and interacting oblique shock structure, with changes in calibrated fitting parameters, can be better understood. In this context, we establish the ability of SG EOS models to reasonably formulate a reactive burn model for the IHE PBX 9502. The model is designed to capture the fast and slow reaction stages inherent in PBX 9502 detonation using a two-stage reaction model. Different mixture closure conditions are examined, namely the classical pressure-temperature equilibrium assumption and a constant solid entropy closure condition. The stability characteristics of SG EOS based detonation models are examined in the context of varying EOS properties of the reactants and products, as well as closure conditions. The SG EOS based structure of oblique shock and detonation waves are also examined. Finally, in a separate exercise, the implemetation and results of a series of large cylindrical rate-stick experiments with the NIHE ammonium nitrate-fuel oil (ANFO) is reported. A detonation-shock-dynamics calibration to the detonation front curvature data obtained from experiments is also presented.