Application of Regularized Radial Basis Function Neural Network in Core Axial Power Distribution Reconstruction

The core axial power shape reconstruction method based on the regularized radial basis function neural network（r RBFNN） is proposed, and the axial power distribution can be reconstructed from 6-segment ex-core detector signals. The 7740 axial power distributions of ACP-100 modular reactor and the corresponding simulated ex-core detector readings are used to verify the r RBFNN reconstruction method. The results show that r RBFNN reconstruction method can reconstruct the core axial power distribution accurately, and this method is robust enough to deal with the inherent ill-posedness of the power distribution reconstruction problem.