Reactor Noise Analysis Applications in Ontario Hydro: A Statistical Technique Used for Systems Surveillance and Conditions Assessment

Abstract

Reactor noise analysis is a non-intrusive statistical technique regularly used in surveillance and diagnostics tasks. Valuable information on reactor system dynamics can be extracted from the fluctuations of instrumentation signals measured during steady-state operation. The small and measurable fluctuations of process signals are the results of stochastic effects inherent in physical processes, such as heat transfer, boiling, coolant flow turbulence, fission process, structural vibrations and pressure oscillations. The goal of reactor noise analysis is to monitor and assess the conditions of technological processes and their instrumentation in the nuclear reactor in a non-intrusive passive way. The noise measurements are usually performed at steady-state operation, while the availability of the signals in their respected systems (i.e. shutdown systems, regulating system) is not interrupted. This paper concentrates on recent applications of reactor noise analysis in Ontario Hydro's CANDU stations, related to the dynamics of in-core flux detectors (ICFDs) and ion chambers. These applications include (1) detecting anomalies in the dynamics of ICFDs and ion chambers, (2) estimating the effective prompt fractions of ICFDs in pourer rundown tests and in noise measurements, (3) detecting the mechanical vibration of ICFD instrument tubes induced by moderator flow, (4) detecting the mechanical vibration of fuel channels induced by coolant flow, (5) identifying the cause of excessive signal fluctuations in certain flux detectors, (6) validating the dynamic coupling between liquid zone control signals.