ON/OFF State Classification of a Reactor Facility Using Gas Effluence Measurements

Inferring the ON/OFF operational state of a reactor facility using measurements from an independent monitoring system is critical to the assessment of its compliance to agreements. We consider the problem of inferring the ON/OFF state of a reactor facility using the measurements of Ar-41, Cs-138, and Xe-138 gas effluence types collected at the facility's off-gas stack. We present classifiers based on thresholding measurements of individual effluence types, and then present fusers that combine their outputs or measurements. We present five fusers based on the simple majority rule, Chow's pattern recognition function, Fisher's combined p-value statistic, the physics-based Poisson radiation counts model, and the correlation coefficient (CC) method. In addition, we also test five machine learning methods based on nonlinear classifiers, which are available as R packages. Our results show that: 1) these gas effluence measurements are effective in inferring the ON/OFF state of a reactor facility, for example, best fusers achieve ~97% detection at ~1% false alarm rate and 2) fusers that combine all effluence types based on physics-based models, CC, and Fisher's method outperform the simple majority rule, Chow's fusers, and the machine learning methods, as well as when they are applied to individual and pairs of effluence types.