Clustering of Self Powered Neutron Detectors: Combining Prompt and Slow Dynamics

The focus of this work is on clustering Self Powered Neutron Detectors (SPNDs) with different dynamic characteristics into smaller groups, with each group containing highly correlated SPNDs. In order to cluster the SPNDs correctly, we propose novel ways to compensate for the effect of different dynamic response characteristics. In particular, two types of SPNDs: (i) cobalt, which give a prompt response, and (ii) vanadium, which give a delayed response, are considered. We propose and compare three compensation methods to cluster both types of SPNDs together using their measurement data: (i) pure delay applied to cobalt SPND data, (ii) cobalt SPND data `slowed down´ to match the vanadium SPND dynamics by passing it through vanadium SPND transfer function, and (iii) vanadium SPND data `speeded up´ to match the cobalt SPND dynamics by passing it through the inverse of vanadium SPND transfer function. Based on extensive simulations, it is found that slowing down cobalt SPND measurements to match the vanadium SPND dynamics yields the best results. This method is then used to obtain clusters from data obtained from a nuclear reactor in India for both vanadium and cobalt SPNDs and the resulting clusters appear reasonable.