Analysis of a radiation attribute from uranium in storage

Storage of radioactive nuclear material necessitates the measurement of gamma and X-ray count rate to discourage tampering. However, in long-term, unattended storage scenarios, it is vital that the analysis of this count rate account for perturbations in the count rate caused by drifting equipment response, and changes in natural (radon levels, for example) or man-made background. In this paper, we report the long-term behavior of 420 silicon P–I–N diodes used as radiation detectors. Data were obtained over the course of 16–20 months at both 1-h and 1-min intervals, and, in addition to perfectly stable count rates, effects attributable to equipment failure, long-term gain drift, short-term gain drift, and software failures were observed. The results of statistical analyses of the time series will be given as well as a description of an event classification algorithm.