A data-driven strategy for xenon dynamical forecasting using dynamic mode decomposition

Abstract

Detection and prediction of xenon induced oscillations are an important part in the operation of pressurized water reactors. Several models have been proposed for the prediction or estimation of xenon oscillations with drawbacks e.g. strongly depend on the initial xenon and iodine distributions, hard-to-implement or computationally costly. In this article, we proposed a fast, model-free and easy-to-implement data-driven strategy based on dynamic mode decomposition (DMD) to forecast the power distribution during the process of xenon oscillations. Various comparative experiments based on HPR1000 reactor show that the proposed data driven strategy is able to capture the complex relational characteristics of the temporal and spatial data of xenon oscillations. It can efficiently discover the hidden dynamicities and thus offers an accurate prediction of the system behavior.