Prediction and uncertainty analysis of power peaking factor by cascaded fuzzy neural networks

Nuclear reactor cores should be maintained within various safety limits such as the local power density (LPD). Therefore, a detailed three-dimensional core power distribution monitoring is required during reactor operation. In addition, LPD must be predicted to prevent nuclear fuel melting. In this study, the most important parameter related to LPD-the power peaking factor-was predicted. A cascaded fuzzy neural network (CFNN) methodology was utilized to predict the power peaking factor in the reactor core. A CFNN model was developed using the numerical simulation data of the optimized power reactor 1000 and its performance was analyzed. Additionally, its uncertainty analysis was conducted to determine the prediction accuracy of the CFNN model. The prediction intervals were found to be pretty narrow, which confirms that the predicted value is reliable. The accuracy of the proposed CFNN model proves to be able to assist nuclear reactor operators in monitoring the power peaking factor.