Investigations of the influence of feedback and coupling effects on neutron noise in a nuclear reactor

Abstract

The neutron power spectral density of power reactors is influenced by many parameters. These can be summarized into two groups. The first being reactivity perturbations whose statistical characteristics are unknown in most cases, and the second being various feedback and spatially dependent transfer functions. In the case of all effects being simultaneously present, which is the case for power reactors, it is extremely difficult to understand the spectra in detail. However some of the reactivity perturbations and their influence on the neutron noise in large power reactors can be simulated in a zero power reactor. In this way the effects can be studied more flexibly without the interference of other effects.

This paper describes investigations of the influence of a known feedback—namely a control loop with known transfer function—on the spectra of neutron chamber signals. Theoretical formulas for the spectra are derived using the point reactor model. These formulas were verified by noise measurements in a zero power reactor. In addition, special attention is given to the noise generated by the control loop. The influence of this feedback noise on the spectra is verified experimentally. In large reactors space dependent transfer functions must be taken into account. As a first approximation to handle the spatial dependence, the afore-mentioned investigations were extended to the two-point reactor model. Corresponding experimental work was done for the Argonaut Reactor Karlsruhe (ARK) with a symmetrical two-slab core loading.

As an application to a more realistic situation coolant boiling in a BWR has been investigated. The boiling must be considered as a feedback mechanism as well as an external reactivity perturbation. In order to simulate the steam bubble content, nitrogen gas was injected into the water moderator of the ARK. By modulating the total gas flow according to the instantaneous reactor power the feedback effect was simulated. The gas flow produced a bandlimited, white reactivity noise. The upper break frequency could be used to determine the travelling time of the bubbles through the core.