Assessing Tail-Probability Stimulus Levels for Explosive Devices and Explosives

Understanding the response of explosive devices and charges as a function of the magnitude of an applied stimulus is essential to understanding the mechanisms which determine the performance and safety performance characteristics and to inform performance and safety assessments.

Some traditional testing methods for quantifying the relationship between stimulus and outcome are not practical or cost-effective.  In this paper sequential statistical methods that combine an initial exploratory phase, typically targeting stimuli around P = 0.5, before switching to target stimuli for tail probabilities, like P = 0.001 or P = 0.999 are considered.  Obviously, such extrapolations with small numbers of tests are dependent on the statistical model, and the strengths and the weaknesses of current approaches are discussed and illustrated with test data for a high-voltage detonator. The approach has a greater applicability than the analysis of detonator data.