NEUTRON FLUCTUATIONS IN TRADITIONAL AND ACCELERATOR DRIVEN REACTORS

Abstract

Classical particle transport, such as the diffusion and slowing down of neutrons in a multiplying system, is a random process. As a result, the number of neutrons in the system, or the number of detector counts, is a random number. Because of the branching property of the cascade, i.e. producing two or more neutrons in the fission process, correlations will exist between the neutrons in the system. This will lead to non-trivial statistics in the cascade, which, among others, carries information on the dynamical properties of the system. Neutron fluctuations have thus been used for a long time to measure either nuclear physics parameters such as the effective delayed neutron fraction, or the so-called multiplication constant, which describes the criticality properties of the system. In addition to the above described inherent random property of the particle transport in steady-state systems, another possibility of neutron fluctuations arises in systems whose properties fluctuate in time. This type of noise is observed in systems with high power. The reason for neutron fluctuations, that is neutron noise, is the fluctuations of the reactor material. There is boiling of the coolant, vibration of various components such as control rods, the core barrel etc., which all will lead to parametric excitation of the neutron noise. In this paper we shall however mostly deal with the neutron noise in low power systems, i.e. with inherent fluctuations of the neutron transport process, and the case of power reactor noise will only be touched upon. The neutron noise in low power systems has received a renewed strong interest recently, due to the appearance of a new concept, the so-called accelerator driven systems (ADS). An ADS is a subcritical reactor, driven by a strong external source, which utilizes spallation. The advantages of such a system is that it has excellent operational safety properties (a criticality accident is practically impossible), can utilize fertile nuclides as fuel such as thorium or U-238 easier than traditional reactors (hereby having access to more abundant fuel reserves as the current reactors), and finally, on the long run and with proper design, it would produce less high-level nuclear waste. As a matter of fact, an ADS can incinerate more waste than it produces, so it can also be used primarily for transmutation of nuclear waste with energy production as a by-product. In an ADS, however, the statistical properties of the neutron source are different from those of the classical radioactive sources that were used in the development of the theory in the past. A radioactive source emits one neutron at a time, and thus all source neutrons are independent (they obey Poisson statistics). In a spallation source, on the other hand, a large number of neutrons are produced in each spallation event, and the number of neutrons generated is a random number. Thus the source neutrons are correlated. In addition to time correlations, energy correlations may exist between spallation neutrons even if the neutron energies in a single spallation even are independent. This fact will be illustrated i the paper. All of the above mentioned (i.e. time and energy) correlations will influence the statistics of the neutrons in an ADS and thus they need to be accounted for.