Investigation of the frequency-dependence of the MTC noise estimator

The Moderator Temperature Coefficient of reactivity (MTC) was earlier proven to be properly determined via noise analysis when the neutron noise is measured locally and when the radial average of the moderator temperature noise is used. Nevertheless, measurements revealed that the MTC was correctly estimated on a much narrower frequency band (0.6 - 0.7 Hz) than expected (0.1 - 1.0 Hz). The present paper investigates the effect of the feedback on the MTC determination using this new MTC noise estimator (which could explain the deviation at low frequencies, i.e. between 0.1 - 0.5 Hz) and the effect of the detector time constants (which could explain the deviation at high frequencies, i.e. between 0.7 - 1.0 Hz). For that purpose, a coupled neutronic/thermalhydraulic model of the Ringhals-3 PWR is developed in the frequency-domain, based on a radial 2-dimensional heterogeneous representation of the core. The transfer function of the detectors is also modelled in this representation. The underestimation of the MTC in the lower end of the frequency band is qualitatively well reproduced by the present model and is explained by the feedback mechanism already important at these frequencies. The effect of the detector time constants is to slightly overestimate the MTC in the upper end of the frequency band. This effect is nevertheless rather small and the underestimation of the MTC at these frequencies is still unexplained. (author)