基于层次分析法和深度学习的凝汽器健康状况趋势  

Condenser health condition trend prediction using analytic hierarchy process and deep learning

作　　者：钱虹 严庆云 王海心 王广基 QIAN Hong;YAN Qingyun;WANG Haixin;WANG Guangji(College of Automation Engineering,Shanghai University of Electric Power,Shanghai 200090,China;China Nuclear Power Engineering Co.,Ltd.,Shenzhen 518052,China;Empower Information Technology(Shanghai)Co.,Ltd.,Shanghai 201508,China)

机构地区：[1]上海电力大学自动化工程学院,上海200090 [2]中广核工程有限公司,广东深圳518034 [3]因博信息科技(上海)有限公司,上海201508

出　　处：《哈尔滨工程大学学报》2024年第12期2391-2399,共9页Journal of Harbin Engineering University

基　　金：国家自然科学基金青年科学基金项目(51906133)。

摘　　要：核电站凝汽器的海洋运行环境导致其工作效率下降,进而产生对二回路冷源损失的影响,本文提出一种基于层次分析法的特征参数贡献度融合方法对其健康度状态进行科学评估。在考虑特征参数对健康度趋势影响的情况下,进一步建立卷积神经网络和长短期记忆神经网络的组合预测模型CNN-LSTM,对凝汽器的劣化演变情况做出趋势分析和预判。研究表明:构建的组合预测模型的预测误差最小,低至0.076%。本文可以有效地预见凝汽器设备劣化演变的趋势,对清洗胶球投放时间的确定提供一定的参考。A feature parameter contribution degree fusion method based on the analytic hierarchy process is proposed to scientifically evaluate the health condition of nuclear power plant condensers.This method addresses the effect of condenser efficiency reduction due to the ocean operating environment on the loss of the secondary loop cold source.Furthermore,a combined prediction model,CNN-LSTM,which integrates convolutional neural networks and long short-term memory neural networks,is established to analyze and predict the future deterioration evolution of the condenser,considering the effect of feature parameters on the health degree trend.Experimental results demonstrate that the prediction error of the combined prediction model constructed in this paper is the smallest,as low as 0.076%.This method effectively predicts the deterioration trend of the condenser and provides valuable insights for determining the optimal timing for deploying cleaning rubber balls.

关 键 词：健康度 清洗胶球 层次分析法 空间距离 卷积神经网络 长短期记忆神经网络 凝汽器 趋势分析 

分 类 号：TL35[核科学技术—核技术及应用]