In nuclear reactor power control, the temporal and spatial distribution characteristics of core power are the core research focus. To realize the observation of the temporal variation and spatial distribution of axial power, an axial multi-point reactor model is established based on the distributed parameter model. Due to the existence of the diffusion term in the model, the power at each point is correlated with its upper and lower neighboring points. Solving the axial spatial distribution of power requires the complete operation of the entire model and the calculation of all axial spatial domains. However, the most critical objective of reactor power control is to limit excessively high power peaks and prevent excessive downward shift of the peak location. Accordingly, attention is mainly focused on the local power spatial distribution near the peak position. Solving for the power distribution over the entire axial space introduces redundant computations. This paper proposes a prediction method for the local spatial distribution of nuclear reactor power. Based on the dynamic characteristics of the distributed parameter model, the neutron concentration vector is reconstructed, and the predicted model coefficient matrix can be obtained without operating the complete axial multi-point reactor model. The proposed method can be directly applied to observe the spatial distribution of axial local power. Simulation results verify that the predicted local spatial power distribution is highly consistent with that obtained by the full-scale model. Meanwhile, the proposed method significantly reduces the computational load, providing a reference for the safe and stable operation of nuclear reactors.

Nuclear reactor axial power distribution; Distributed parameter model; Local power prediction; State vector reconstruction