基于KNN方法的熔盐堆系统瞬态识别模型开发及分析  

Development and analysis of a K-nearest-neighbor-based transient identification model for molten salt reactor systems

作　　者：周天泽 虞凯程 程懋松[2,3] 戴志敏 ZHOU Tianze;YU Kaicheng;CHENG Maosong;DAI Zhimin(ShanghaiTech University,Shanghai 201210,China;Shanghai Institute of Applied Physics,Chinese Academy of Sciences,Shanghai 201800,China;University of Chinese Academy of Sciences,Beijing 100049,China)

机构地区：[1]上海科技大学,上海201210 [2]中国科学院上海应用物理研究所,上海201800 [3]中国科学院大学,北京100049

出　　处：《核技术》2023年第11期122-132,共11页Nuclear Techniques

基　　金：中国科学院科技战略先导项目资助。

摘　　要：熔盐堆作为第四代先进核能系统之一,在安全性、经济性、防核扩散和可持续性等方面具有独特的优势。为了保障熔盐堆运行安全,需要快速、准确地识别瞬态工况,目前的瞬态识别方法主要依赖于操作员人工识别,这会引入较大的人为因素,严重影响核电安全。为了减少熔盐堆系统瞬态识别过程中引入的人为因素,提高熔盐堆运行安全,使用RELAP5-TMSR程序对美国橡树岭国家实验室建造运行的熔盐实验堆(Molten Salt Reactor Experiment,MSRE)的瞬态工况进行建模与仿真,产生数据集,基于K近邻(K-Nearest Neighbor,KNN)机器学习方法,建立了熔盐堆系统瞬态识别模型,并对识别模型在噪声下的鲁棒性进行了分析和优化。结果显示:基于KNN方法建立的熔盐堆系统瞬态识别模型在测试集上的F1分数达到99.99%;在噪声下的识别F1分数达到94.32%,具有较高的鲁棒性;进一步优化后的熔盐堆系统瞬态识别模型在噪声下的F1分数达到99.73%,能较为准确地识别MSRE的瞬态工况,满足熔盐堆系统瞬态识别需求。基于KNN方法的熔盐堆系统瞬态识别模型能够有效识别系统瞬态工况,可应用于熔盐堆智能运维,确保熔盐堆运行安全。[Background]Molten salt reactors(MSRs)are fourth-generation advanced nuclear energy systems that exhibit characteristics such as high safety,high economy,nonproliferation,and sustainability.To ensure the safe operation of MSRs,identifying transient conditions promptly and accurately is crucial.However,current system transient identification methods rely on manual identification by operators,introducing significant human factors seriously affecting nuclear power safety.[Purpose]This study aims to establish a transient identification model for an MSR system based on the K-nearest neighbor(KNN)method,so as to reduce human factors introduced during the traditional system transient identification process,and improve the operational safety of the MSR.[Methods]Datasets for the system transient identification model were generated by using the RELAP5-TMSR code to simulate 11 operating conditions of the molten salt reactor experiment(MSRE)built and operated at Oak Ridge National Laboratory in the United States.Subsequently,a system transient identification model based on the KNN method was developed by training,optimizing,and validating these datasets.Four metrics,i.e.,accuracy,precision,recall,and F1-score were applied to evaluating the system transient identification model.Finally,the robustness of the model was tested and optimized under noisy conditions.[Results]The results demonstrate that the KNN-based transient identification model for the MSR system achieves a 99.99%F1-score on the test datasets.The system transient identification model also exhibits high robustness,with an F1-score of 94.32%under noisy conditions.The optimized system transient identification model achieves a 99.73%F1-score when identifying transient conditions under noise,accurately identifying the transient conditions of the MSRE.[Conclusions]The KNN-based transient identification model for the MSR system can satisfy the requirements of transient identification of the MSR system,hence be applied to intelligent MSR operations and maintenance,ensur

关 键 词：熔盐堆 K近邻 系统瞬态识别 鲁棒性 

分 类 号：TL426[核科学技术—核技术及应用]