Simulation of In-Core Neutron Noise Measurements for Axial Void Profile Reconstruction in Boiling Water Reactors

This paper reports on the development and application of a method of emulating bubbly flow by generating bubbles with random sampling methods. The purpose of the modeling is that by using the simulated random two phase flow as input, one can generate "synthetic" neutron noise signals by convoluting the input with a simplified neuronic transfer function, on which the possibility of reconstructing the axial void profile from in-core neutron noise measurements can be studied by standard spectral noise analysis methods. The long term goal of this work is to elaborate methods of neutron noise analysis, by which the local void fraction in a boiling water reactor can be determined by measurements. In this preliminary stage, two methods for the reconstruction of the axial void and the velocity profiles are discussed. The first method is based on the break frequency of the neutron auto-power spectrum, whereas the second method only utilizes the information in the transit time of the void fluctuations between axial pairs of neutron detectors. A clear and monotonic relationship between the chosen observables and the two-phase flow properties was found, but an accurate determination of the void fraction requires further development and testing of the various unfolding alternatives.