Application of Least Square Support Vector Machine in Core Power Distribution Reconstruction

The application of the least square support vector machine (LS-SVM) to core axial power distribution reconstruction was researched, and 18-node powers were reconstructed from six-level in-core detector signals. Axial power distributions of 7 740 cases of ACP-100 modular reactor were used to verify the accuracy of the LS-SVM reconstruction method. The results show that the LS-SVM method performs much better than the alternating conditional expectation (ACE) method and the LS-SVM method has good robustness.