Citation:
D. Frank, M. Dzur, S. Chirayath, F. Naqvi, R. Weinmann-Smith, P. Simon, P. O’Neal, Q. Burke, “Localization and Quantification of Nuclear Material in Hot Laboratories: A Geometric Solution for NMC&A”, Institute of Nuclear Materials Management Annual Meeting, 21-25 July 2024, Portland, Oregon.
Abstract:
The advancement of experimental and computational techniques is crucial in efficiently controlling and accounting for nuclear materials. This plays a pivotal role in preventing the diversion of special nuclear materials (SNMs) from peaceful applications. Experimental investigations into the metallurgic properties of SNM occasionally result in the dispersion of minute amounts of proliferable material. Conducted in laboratories with heightened background radiation levels, these experiments, despite meticulous cleaning efforts, may leave trace amounts of SNM within the experimental setup, thereby introducing uncertainties into safeguards measures.rnrnThis paper introduces an algorithm designed to address this challenge by initially approximating the 3D position of radioactive material. Subsequently, the algorithm utilizes this position approximation to quantify the mass of the SNM. The algorithm’s efficacy was evaluated through Monte Carlo N-Particle® (MCNP®) simulations, employing a glovebox surrounded by helium-3 neutron detectors. Calibration involved simulating a point source at 400 different positions within the experimental chamber, with the resulting calibration data used to correlate count rates with radial distance from each detector.rnrnUtilizing the correlations between count rate and distance with testing data, the algorithm projects a thick spherical-like shell from each detector. The shape of the shell is influenced by the detector’s efficiency characteristics, and the shell’s thickness is determined by statistical errors from the calibration data. Along the shell’s thickness, weighted voxels are assigned based on a Gaussian distribution. The algorithm determines that by summing these voxelated shells, the highest degree of intersection, represented by the voxel with the highest weight, indicates the approximate location of a radioactive source.rnrnBy applying the algorithm to testing data, the uncertainty regarding the material’s location is narrowed down to about 18cm. Simultaneously, the average uncertainty in the distance between each detector and the material is reduced to approximately 4cm. Since the radioactive material is positioned between 1.5m and 3.5m from each detector, error propagation following an inverse square law allows for mass quantification uncertainties of less than 5%.