基于机器学习的超临界水传热恶化判定研究  

Research on Judgment of Supercritical Water Heat Transfer Deterioration Based on Machine Learning

作　　者：马栋梁 周涛[2,3] 黄彦平 Ma Dongliang;Zhou Tao;Huang Yanping(School of Information Engineering and Computer Science,Hebei Finance University,Baoding,Hebei,071051,China;School of Nuclear Science and Engineering,North China Electric Power University,Beijing,102206,China;School of Energy and Environment,Southeast University,Nanjing,211189,China;CNNC Key Laboratory on Nuclear Reactor Thermal Hydraulics Technology,Nuclear Power Institute of China,Chengdu,610213,China)

机构地区：[1]河北金融学院信息工程与计算机学院,河北保定071051 [2]华北电力大学核科学与工程学院,北京102206 [3]东南大学能源与环境学院,南京211189 [4]中国核动力研究设计院中核核反应堆热工水力重点实验室,成都610213

出　　处：《核动力工程》2021年第4期91-95,共5页Nuclear Power Engineering

基　　金：“河北省科技型企业发展评价监测平台”专项课题(KJQY202003);东南大学学科振兴计划和教师启动基金(1103007005-2020);北京市自然科学基金(3172032)。

摘　　要：为了进一步提高超临界水堆的安全稳定性,避免超临界水传热恶化的发生,在已有的超临界水传热实验数据基础之上,利用几种主要的机器学习算法,对超临界水的实验参数状态点是否发生了传热恶化进行分类判断和预测精度分析。研究表明:随机森林算法对于测试数据的平均预测精度最高,达到了97.8%左右;K近邻(KNN)分类算法的平均预测精度最低,但是也达到了90%以上。同时对各种不同的影响参数对传热恶化的选取重要度的分析可知,与传热恶化判定关系最重要的参数是比焓,其次为传热系数;与传热恶化重要度选择关系最小的是管径。In order to further improve the safety and stability of supercritical water reactors,avoid the occurrence of the heat transfer deterioration in supercritical water,based on the existing experimental data of supercritical water heat transfer,using several main machine learning algorithms,the classification and judgment and prediction accuracy analysis of the experimental parameter state points of supercritical water were made to determine the occurrence of the heat transfer deterioration.The research results show that the random forest algorithm has the highest average prediction accuracy for the test data,reaching about 97.8%.The average prediction accuracy of the K-nearest neighbor algorithm is the lowest,but it also reaches about 91%.At the same time,the importance of various influence parameters on the selection of heat transfer deterioration was analyzed.The most important parameter related to the heat transfer deterioration judgment is the specific enthalpy,and the second important parameter is the heat transfer coefficient.The third important parameter with contribution to the heat transfer deterioration is the pipe diameter.

关 键 词：超临界水 传热恶化 机器学习 随机森林 K近邻 

分 类 号：TP391[自动化与计算机技术—计算机应用技术] TL331[自动化与计算机技术—计算机科学与技术]