Prediction of secondary circuit power generation for nuclear units based on WSA-VMD-BP approach

Le résumé fourni par la source

Nuclear energy has been regarded as an environmentally superior and reliable form of energy due to its inherent stability and lengthy operational lifetime. Small modular reactors (SMR) have been positioned as a prominent research focus in recent years, while artificial intelligence has also opened new avenues for exploration the performance of nuclear units. This paper proposes to predict the power generation of nuclear power plants integrating variational modal decomposition (VMD) and back propagation (BP) neural networks to optimize the hyperparameters through the utilization of the wave search algorithm (WSA). The efficacy of the proposed method is validated through the construction of a secondary circuit mathematical model of a nuclear power facility and an in-depth analysis of the thermal parameter relationships. The proposed method demonstrated an MAE of 0.6964 and a WIA of up to 0.99 in the noiseless case, exhibiting a low prediction error. In the context of environmental noise, the prediction accuracy can be maintained at approximately WIA ∼ 0.97. These results demonstrate the high prediction accuracy and strong robustness of the algorithm.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.