【摘要】 为实现准确、高效预测铅铋快堆关键热工参数，提高铅铋快堆热工安全评价能力，提出了一种基于自适应径向基函数(RBF)神经网络的铅铋快堆燃料元件表面温度预测方法。利用子通道分析程序SUBCHANFLOW建立小型铅铋快堆SPALLER-100堆芯子通道模型，以计算得到的2 000组堆芯功率分布和各冷却剂流道质量流量分布数据作为训练样本，对自适应RBF神经网络模型进行训练，实现对铅铋快堆燃料元件表面温度的预测。通过对比，证明了自适应RBF神经网络方法的有效性、优越性和泛化能力。研究表明：自适应RBF神经网络方法预测燃料包壳最高温度相对误差不超过0.5%,可用于铅铋快堆热工水力参数的快速预测。更多还原

【Abstract】 In order to accurately and efficiently predict the key thermal parameters of lead-bismuth fast reactor and improve the ability of thermal safety evaluation of lead-bismuth fast reactor, a surface temperature prediction method of for the fuel elements of lead-bismuth fast reactor fuel element based on adaptive radial basis function(RBF) neural network was proposed in this paper. The core subchannel model of small lead-bismuth fast reactor SPALLER-100 was established by using subchannel analysis program SUBCHANFLOW. 2 000 groups of core power distribution and coolant mass flow distribution data were used as training samples to train the adaptive RBF neural network model to predict the surface temperature of lead-bismuth fast reactor fuel elements. By comparison, the effectiveness, superiority and generalization ability of the adaptive RBF neural network method were proved. The research shows that the relative error of the adaptive RBF neural network method for predicting the maximum temperature of fuel cladding is less than 0.5%, which can be used for the rapid prediction of thermal and hydraulic parameters of lead-bismuth fast reactor.更多还原