基于RBF神经网络的压水堆堆芯三维功率分布方法研究

Research on 3D Power Distribution of PWR Reactor Core Based on RBF Neural Network

堆芯三维功率分布的实时监测对核电厂的安全高效运行和控制系统优化均有重大意义。本文利用堆外核测量系统及RBF神经网络构建了一个实时堆芯功率三维分布监测系统,以提高监测的实时性及减小三维功率分布的拟合误差。在300MW压水堆核电厂全范围仿真机上进行了一系列仿真实验,结果表明,该监测系统能在燃料循环周期的一定燃耗范围内,实时呈现堆芯三维功率分布,并通过几种方法对模型的精度进行了有效改进。

Real-time monitor for 3Dpower distribution is critical to nuclear safety and high efficiency of NPP’s operation as well as the control system optimization.A method was proposed to set up a real-time monitor system for 3Dpower distribution by using of ex-core neutron detecting system and RBF neural network for improving the instantaneity of the monitoring results and reducing the fitting error of the 3Dpower distribution.A series of experiments were operated on a 300 MW PWR simulation system.The results demonstrate that the new monitor system works very well under condition of certain burnup range during the fuel cycle and reconstructs the real-time 3D distribution of reactor core power.The accuracy of the model is improved effectively with the help of several methods.