Patchy nuclear chain reactions

Stochastic fluctuations of the neutron population within a nuclear reactor are typically prevented by operating the core at a sufficient power, since a deterministic (i.e., exactly predictable) behavior of the neutron population is required by automatic safety systems to detect unwanted power excursions. In order to characterize the reactor operating conditions at which the fluctuations vanish, an experiment was designed and took place in 2017 at the Rensselaer Polytechnic Institute Reactor Critical Facility. This experiment however revealed persisting fluctuations and striking patchy spatial patterns in neutron spatial distributions. Here we report these experimental findings, interpret them by a stochastic modeling based on branching random walks, and extend them using a “numerical twin” of the reactor core. Consequences on nuclear safety will be discussed. Understanding and predicting fluctuations of the neutron population within a nuclear reactor is of fundamental importance for nuclear safety, especially in connection with reactor control at startup and shutdown. The study presents experiments and Monte Carlo simulations of persistent neutron fluctuations and correlations (stochastic noise and neutron clustering) in nuclear reactors which are interpreted by stochastic modelling based on branching random walks.