Predictive Limits and Potential Improvements using Detonation Shock Dynamics for High Explosives

The Detonation Shock Dynamics (DSD) methodology is commonly used for predicting detonation arrival times of a reacting high explosive in a given geometry. The method is built upon the notion that the curvature is small relative to the inverse of the reaction zone length. The detonation wave is then transported as a level set through the geometry, and in standard DSD, the local normal wave velocity is directly correlated to the curvature of the detonation shock front. It is critical to accurately define this velocity-curvature relation to predict detonation propagation properties in varied geometries. In this work, we will show how different DSD model forms, each calibrated to the same data set, compare with one another. First, we will define a baseline calibration, following our standard best practices for a wide range of explosive materials. Then, through uncertainty analysis, we will show that adding more parameters to this model does not fundamentally improve the predictive capabilities, suggesting that the foundational assumptions of the model itself are insufficient. Using higher-order methods, we will show that the inclusion of more physics substantially improves the model's predictive capabilities.