基于BP神经网络的堆芯三维功率重构方法研究

Research on 3D Power Reconstruction of Reactor Core Based on BP Neural Network

摘要: 堆芯功率分布包含了堆芯内的大量信息，由于在反应堆运行过程中无法直接测量堆芯内所有位置的功率，因此需通过其他方法得到堆芯三维功率分布的情况。本文以秦山一期工程为对象，利用堆外中子探测器在不同棒位和不同功率下的计数及BP神经网络对堆芯三维功率分布进行重构计算，并利用REMARK程序对该计算结果进行验证。结果表明，该功率重构方法能在反应堆运行的50%～100%功率范围内，较好地呈现堆芯三维功率分布。

Abstract: The in-core power distribution contains plenty of informations in the core. Because the power distribution of all the parts of the core can not be directly measured when the reactor is operating, it is necessary to obtain the in-core power distribution by other ways. A method was proposed to calculate 3D power distribution of Qinshan project by using of ex-core neutron detecting system and BP neural network. The results were compared with those obtained by REMARK core physics real-time simulation program. It is shown that this BP reconstruction method can reconstruct the 3D distribution of reactor core power well in 50%-100% power range.