Performance assessment of self-interrogation neutron resonance densitometry for spent nuclear fuel assay

摘要

Self-Interrogation Neutron Resonance Densitometry (SINRD) is one of several nondestructive assay (NDA) techniques being integrated into systems to measure spent fuel as part of the Next Generation Safeguards Initiative (NGSI) Spent Fuel Project. The NGSI Spent Fuel Project is sponsored by the US Department of Energy's National Nuclear Security Administration to measure plutonium in, and detect diversion of fuel pins from, spent nuclear fuel assemblies. SINRD shows promising capability in determining the ~(239)Pu and ~(235)U content in spent fuel. SINRD is a relatively low-cost and lightweight instrument, and it is easy to implement in the field. The technique makes use of the passive neutron source existing in a spent fuel assembly, and it uses ratios between the count rates collected in fission chambers that are covered with different absorbing materials. These ratios are correlated to key attributes of the spent fuel assembly, such as the total mass of ~(239)Pu and ~(235)U. Using count rate ratios instead of absolute count rates makes SINRD less vulnerable to systematic uncertainties. Building upon the previous research, this work focuses on the underlying physics of the SINRD technique: quantifying the individual impacts on the count rate ratios of a few important nuclides using the perturbation method; examining new correlations between count rate ratio and mass quantities based on the results of the perturbation study; quantifying the impacts on the energy windows of the filtering materials that cover the fission chambers by tallying the neutron spectra before and after the neutrons go through the filters; and identifying the most important nuclides that cause cooling-time variations in the count rate ratios. The results of these studies show that ~(235)U content has a major impact on the SINRD signal in addition to the ~(239)Pu content. Plutonium-241 and ~(241)Am are the two main nuclides responsible for the variation in the count rate ratio with cooling time. In short, this work provides insights into some of the main factors that affect the performance of SINRD, and it should help improve the hardware design and the algorithm used to interpret the signal for the SINRD technique. In addition, the modeling and simulation techniques used in this work can be easily adopted for analysis of other NDA systems, especially when complex systems like spent nuclear fuel are involved. These studies were conducted at Los Alamos National Laboratory.