Detection and Localisation of Multiple In-core Perturbations with Neutron Noise-based Self-Supervised Domain Adaptation

The use of non-intrusive techniques for monitoring nuclear reactors is becoming more
vital as western fleets age. As a consequence, the necessity to detect more frequently occurring operational anomalies is of upmost interest. Here, noise diagnostics—the analysis of small stationary deviations of local neutron flux around its time-averaged value — is employed aiming to unfold from detector readings the nature and location of driving perturbations. Given that in-core instrumentation of western-type light-water reactors are scarce in number of detectors, rendering formal inversion of the reactor transfer function impossible, we propose to utilise advancements in Machine Learning and Deep Learning for the task of unfolding. This work presents an approach to such a task doing so in the presence of multiple and simultaneously occurring perturbations or anomalies. A voxel-wise semantic segmentation network is proposed to determine the nature and source location of multiple and simultaneously occurring perturbations in the frequency domain. A diffusion-based core simulation tool has been employed to provide simulated training data for two reactors. Additionally, we work towards the application of the aforementioned approach to real measurements, introducing a self-supervised domain adaptation procedure to align the representation distributions of simulated and real plant measurements.