严重事故条件下安全壳混凝土气体渗透性预测  

Prediction of gas permeability of containment concrete under severe accident conditions

作　　者：郭俊营 白冰 高健琳 范宸宇 许强 毛继泽[2] 肖梦霞 GUO Junying;BAI Bing;GAO Jianlin;FAN Chenyu;XU Qiang;MAO Jize;XIAO Mengxia(China Nuclear Power Engineering Design Co.,Ltd.,Shenzhen 518031,China;College of Aerospace and Civil Engineering,Harbin Engineering University,Harbin 150001,China)

机构地区：[1]中广核工程设计有限公司,广东深圳518031 [2]哈尔滨工程大学航天与建筑工程学院,黑龙江哈尔滨150001

出　　处：《应用科技》2023年第1期100-105,共6页Applied Science and Technology

基　　金：中广核安全壳本体(混凝土结构)密封性试验研究项目(007-EC-B);哈尔滨工程大学高水平科研引导专项(3072022QBZ0201)。

摘　　要：核电站发生严重事故时安全壳内部形成的高温、高湿、高压条件对安全壳混凝土的抗气体渗透能力有着极大的考验。本文基于理论分析模型建立数据样本,环境温度、水蒸气分压、升温速率和混凝土的水扩散系数作为为输入变量,安全壳混凝土气体渗透深度作为输出变量,利用BP算法设计了包含4个输入层节点、12个隐含层节点、1个输出层节点的人工神经网络结构,对严重事故工况下安全壳混凝土的抗气体渗透性能进行分析预测。模型测试结果的相对误差在5%以内,相关系数达到了0.999 7,研究结果表明反向传播神经网络(BPNN)有着有效可靠的预测能力,能对安全壳混凝土的气体渗透深度进行高精度预测。The high temperature,high humidity and high pressure conditions formed inside the containment of a nuclear power plant during a serious accident have a great impact on the gas permeation resistance of the containment concrete.In this paper,data samples are established based on the theoretical analysis model,with ambient temperature,partial pressure of water vapor,heating rate and water diffusion coefficient of concrete materials as input variables and gas permeation depth of containment concrete as output variables.Using back propagation(BP)algorithm to design an artificial neural network,containing 4 input layer nodes,12 implicit layer nodes and 1 output layer node structure,so as to predict the gas permeation resistance of containment concrete under severe accident conditions.The relative error of the model test result is within 5%,and the correlation coefficient reaches 0.9997.The research results show that the BP artificial neural network has effective and reliable prediction ability,and it can predict the gas penetration depth of containment concrete with high accuracy.

关 键 词：严重事故 安全壳 混凝土 环境条件 耦合作用 气体渗透深度 BP算法 预测模型 

分 类 号：TG178[金属学及工艺—金属表面处理] TM623[金属学及工艺—金属学]