PNNL

Abstract

The International Atomic Energy Agency (IAEA) will require the development of advanced technologies to effectively safeguard nuclear material at increasingly large-scale nuclear recycling facilities. Ideally, the envisioned technologies would be capable of nondestructive, near-real-time, autonomous process monitoring. This paper describes recent results from model simulations designed to test the Multi-Isotope Process (MIP) monitor, a novel approach to safeguarding reprocessing plants. The MIP monitor combines the detection of intrinsic gamma ray signatures emitted from process solutions with multivariate analysis to detect off-normal conditions in process streams nondestructively and in near-real-time. Three computer models including ORIGEN-ARP, AMUSE, and SYNTH were used in series to predict spent nuclear fuel composition, estimate element partitioning during separation, and simulate spectra from product and raffinate streams using a variety of gamma detectors, respectively. Simulations were generated for fuel with various irradiation histories and under a variety of plant operating conditions. Principal component analysis (PCA) was applied to the simulated gamma spectra to investigate pattern variations as a function of acid concentration, burnup, and cooling time. Hierarchical cluster analysis (HCA) and partial least squares (PLS) were also used in the analysis. The MIP monitor was found to be sensitive to induced variations of several operating parameters including distinguishing ±2.5% variation from normal process acid concentrations. The ability of PLS to predict burnup levels from simulated spectra was also demonstrated to be within 3.5% of measured values.