Date of Award

9-7-2007

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Department

Department of Engineering Physics

First Advisor

Glen P. Perram, PhD

Abstract

Time-resolved infrared spectra were recently collected via a Fourier-transform spectrometer (FTS) from the detonation fireballs of two types of conventional military munitions (CMM) as well as uncased TNT and four types of enhanced novel explosives (ENEs). The CMM spectra are dominated by continuum emission, and a single-temperature Planckian distribution, modified for atmospheric attenuation, captures most of the variation in the data. Some evidence of selective emission is identified by systematic patterns in the fit residuals. The behavior of these systematic residuals affords a distinction between the two types of CMMs studied. The uncased TNT and ENE spectra appear strongly influenced by both continuum and selective emission. A physics-based spectral model is developed consisting of seven parameters: size, temperature, particulate absorption coefficient, and gas concentrations for H2O, CO2, CO, and HCl. Fitting affords a high-fidelity representation with features that correlate with HE characteristics. The hydrogen-to-carbon ratio (R) separates the TNT and ENE events and is consistent with stoichiometric expectations. Average values of R are compared with stoichiometry (in parenthesis): TNT 1.13 (0.79); ENE0B 9.2 (21.3); ENE1 4.9 (6.7); ENE2A 4.6 (5.8); ENE2B 6.5 (6.7). Bayesian discrimination boundary between TNT and ENE is R = 1.67 and the mean probability of error is less than 0.3% for this two-class problem. The Fisher ratio is 17.4 and ENE can be distinguished from TNT with 99% detection rate with corresponding false-alarm rate of less than 10-4%.

AFIT Designator

AFIT-DS-ENP-07-03

DTIC Accession Number

ADA472261

Included in

Nuclear Commons

Share

COinS