Neutron monitoring based on the higher order statistics of fission chamber signals

Abstract: The work in this thesis corresponds to the safety aspect of Generation IV nuclear systems. One of the safety aspects concerns the enhancement of the performance of the in-vessel on-line core monitoring with neutron flux measurements. It was concluded earlier that fission chambers are the best candidate to provide in-vessel measurements in sodium cooled fast reactors.

This thesis focuses on the performance of signal processing methods in order to unfold the count rate of fission chambers. The main goal is to investigate the possible application of processing methods based on the higher order statistics of the signal in order to provide accurate count rate estimation over a wide range both for stationary and transient signals. The work also consists of the study of self-monitoring capabilities in order to detect fission chamber malfunctions at an early stage. The investigation is based on analytic assessments, on simulation of fission chamber responses and signals, and on experimental application of processing methods.

The thesis covers five main studies. The first part presents the theoretical description of fission chamber signals. The second part investigates the performance of the traditionally applied methods (pulse and Campbelling mode) through simulations. It is shown that these methods are not capable to cover the whole count rate range of the chamber. Therefore, the third part studies the possible application of methods based on higher order statistics of the signal through simulations and experiments. It is shown that these methods can provide accurate estimations over a wide count rate range. The fourth part covers the theoretical background of self-monitoring capabilities based on the spectral properties of the signal. Finally, the fifth part presents the implementation of the methods in a real-time neutron monitoring system based on a System on a Chip, which embeds a field-programmable gate array.

By the methods elaborated in this thesis, a faster, more effective and more accurate monitoring of the reactor power is possible than with the methods used so far, even when the normal operating state is changing.