Prediction of state transitions in 3D core dynamics and xenon transients based on dynamic mode decomposition

This study proposes a method based on Dynamic Mode Decomposition (DMD) for predicting the state transitions of core dynamics and xenon transients in nuclear reactors. Both types of transients, occurring at different time or spatial scales, have significant impacts on the stability and safety of nuclear reactors. The DMD method is employed to model and predict the transient behaviors in nuclear reactors. DMD is a data-driven model simplification technique that extracts the dynamic modes, i.e., characteristic patterns, from time-series data. By applying DMD to the data of core dynamics and xenon transients, we obtain the characteristic modes of the system under different states, which are then used for predicting state transitions. Simulation experiments are conducted using the benchmark problems of the 3D LWR transient cores provided by NEACRP. The results demonstrate that the DMD-based method effectively predicts the state transitions of core dynamics and xenon transients in nuclear reactors. The accuracy and reliability of the method are validated through the analysis and comparison of predicted results with actual data. Additionally, the impact of model parameters on prediction performance is investigated, and optimization strategies are proposed to further improve the prediction results. The findings of this study offer a novel approach and insights for predicting transient behaviors in the field of nuclear energy.