姬嘉益,马洁

• 1:北京信息科技大学机电工程学院

摘要(Abstract)：

在核电站运行过程中，及时检测故障的发生并进行故障诊断，对核电站的安全稳定运行至关重要。选取核动力一回路系统作为主要研究对象，采用分布式思想，结合深度学习和信息融合的诊断方法，建立了核动力一回路分布式故障诊断系统。首先，通过皮尔逊相关系数以及排序融合的卡方得分和轻量级梯度提升机（light gradient boosting machine,LightGBM）特征重要性进行特征选择。随后，利用深度可分离密集残差网络（deep separable dense residual network,DSDRN）进行故障诊断，通过加深网络深度提取更深层次的特征信息，采用深度可分离卷积减少网络参数量，并通过残差连接降低梯度消失的风险，保证网络训练稳定性。所提模型的各类故障识别准确率均大于98.5%，可为核电站故障诊断工程提供参考。

关键词(KeyWords)： 核电站;分布式故障诊断;特征选择;残差网络;深度可分离卷积网络;密集块

Abstract:

Keywords:

基金项目(Foundation): 国家自然科学基金项目(61973041)

作者(Author): 姬嘉益,马洁

DOI: 10.16508/j.cnki.11-5866/n.2025.04.004

参考文献(References)：

• [1]何鑫，段晨，刘闯.基于改进探路者算法优化支持向量机的断路器故障诊断[J].红水河，2023,42(6):90-95.HE X,DUAN C,LIU C. Circuit breaker fault diagnosis based on support vector machine optimized by improved pathfinder algorithm[J]. Hongshui River,2023,42(6):90-95.(in Chinese)
• [2]孔维林，曾冀，刘雪涛，等.基于多特征融合和深度信念网络的断路器元件故障识别研究[J].电子设计工程，2025,33(3):63-67.KONG W L,ZENG J,LIU X T,et al. Research on fault identification of circuit breaker components based on multi feature fusion and deep belief network[J]. Electronic Design Engineering,2025,33(3):63-67.(in Chinese)
• [3]吴与伦，肖谭南，陈颖.基于故障枚举树生成模糊Petri网的变电站故障诊断方法[J].工矿自动化，2025,51(1):85-94.WU Y L,XIAO T N,CHEN Y. Fault diagnosis method for substations based on fault enumeration tree to generate fuzzy Petri net[J]. Industry and Mine Automation,2025,51(1):85-94.(in Chinese)
• [4]PARK J H,JO H S,LEE S H,et al. A reliable intelligent diagnostic assistant for nuclear power plants using explainable artificial intelligence of GRU-AE,LightGBM and SHAP[J].Nuclear Engineering and Technology,2022,54(4):1271-1287.
• [5]戴滔，隋阳，郑梦琰.核电厂反应堆冷却剂系统故障诊断模型[J].科学技术与工程，2024,24(35):15042-15047.DAI T,SUI Y,ZHENG M Y. Model for fault diagnosis of reactor coolant system for nuclear power plants[J]. Science Technology and Engineering,2024,24(35):15042-15047.(in Chinese)
• [6]陈波，曹中才，姚祥英，等.核电厂化学控制异常检测与故障诊断系统研究与实现[J].核动力工程，2024,45(S1):13-18.CHEN B,CAO Z C,YAO X Y,et al. Research and implementation of anomaly detection and fault diagnosis system for chemical control in nuclear power plants[J]. Nuclear Power Engineering,2024,45(S1):13-18.(in Chinese)
• [7]TONG T,LI G,LIU X J,et al. Image super-resolution using dense skip connections[C]//2017 IEEE International Conference on Computer Vision(ICCV). New York, NY,USA:IEEE,2017:4809-4817.
• [8]LAN R S,SUN L,LIU Z B,et al. MADNet:a fast and lightweight network for single-image super resolution[J]. IEEE Transactions on Cybernetics,2021,51(3):1443-1453.
• [9]MIKI D,DEMACHI K. Bearing fault diagnosis using weakly supervised long short-term memory[J]. Journal of Nuclear Science and Technology,2020,57(9):1091-1100.
• [10]NGUYEN T N,PONCIROLI R,BRUCK P,et al. A digital twin approach to system-level fault detection and diagnosis for improved equipment health monitoring[J]. Annals of Nuclear Energy,2022,170:109002.
• [11]AYO-IMORU R M,ALI A A,BOKORO P N. An enhanced fault diagnosis in nuclear power plants for a digital twin framework[C]//2021 International Conference on Electrical,Computer and Energy Technologies(ICECET). New York, NY,USA:IEEE,2021:2072-2077.
• [12]AYODEJI A,LIU Y K,XIA H. Knowledge base operator support system for nuclear power plant fault diagnosis[J].Progress in Nuclear Energy,2018,105:42-50.
• [13]AMIN M T,YAO Y T,YU J,et al. Probabilistic monitoring of nuclear plants using R-vine copula[J]. Annals of Nuclear Energy,2023,190:109867.
• [14]薛时雨.基于深度学习的核电站冷却剂流失事故诊断及趋势预测[D].长沙：湖南大学，2020.XUE S Y. Diagnosis and prediction for loss of coolant accidents in nuclear power plants using deep learning methods[D].Changsha:Hunan University,2020.(in Chinese)
• [15]王梦蝶，王俊玲，于成波.基于CNN的核电站主冷却剂系统压力预测[J].哈尔滨商业大学学报（自然科学版），2021,37(5):576-581.WANG M D,WANG J L,YU C B. Main coolant system pressure prediction in nuclear power plants based on CNN[J]. Journal of Harbin University of Commerce(Natural Sciences Edition),2021,37(5):576-581.(in Chinese)
• [16]ZHANG Y,ZHAN H X,ZHANG H,et al. Fault location method of distributed generation sources in distribution networks based on optimal configuration of voltage sag monitoring points[J].Journal of Physics. Conference Series,2022,2310(1):012010.
• [17]张鑫.面向多模态数据记录的相似度分类方法研究与实现[D].西安：西安理工大学，2024.ZHANG X. Study and implementation of similarity classification method for multimodal data recording[D]. Xi′an:Xi′an University of Technology,2024.(in Chinese)
• [18]HOORN Q A,ZAYAS G A,RODRIGUEZ E E,et al.Identification of quantitative trait loci and associated candidate genes for pregnancy success in Angus-Brahman crossbred heifers[J]. Journal of Animal Science and Biotechnology,2024,15(1):162-170.
• [19]曹希彧，陈鑫，魏同权.基于RISC-V的深度可分离卷积神经网络加速器[J].计算机学报，2024,47(11):2536-2551.CAO X Y,CHEN X,WEI T Q. Efficient accelerator for depthwise separable convolutional neural networks based on RISC-V[J]. Chinese Journal of Computers,2024,47(11):2536-2551.(in Chinese)
• [20]刘志锋，唐俊贤，林芝宁，等.基于深度可分离卷积混合网络模型的地浸采铀注液量预测研究[J].铀矿冶，2025,44(1):9-17.LIU Z F,TANG J X,LIN Z N,et al. Injection volume prediction for in-situ leaching of uranium based on depthwise separable convolution mixer network model[J]. Uranium Mining and Metallurgy,2025,44(1):9-17.(in Chinese)
• [21]金钰森，丁飞，陈竺，等.基于深度可分离卷积神经网络的轴承故障诊断模型[J].无线电通信技术，2024,50(1):193-202.JIN Y S,DING F,CHEN Z,et al. Bearing fault diagnosis based on deep separable convolutional neural network[J]. Radio Communications Technology,2024,50(1):193-202.(in Chinese)
• [22]包锋，庄泽堃.基于CNN-BiGRU-ResNet的网络入侵检测研究[J].计算机与数字工程，2024,52(2):468-472.BAO F,ZHUANG Z K. Research on network intrusion detection based on CNN-BiGRU-ResNet[J]. Computer and Digital Engineering,2024,52(2):468-472.(in Chinese)
• [23]陈红，闫建国，杨华，等.基于结构重参数化的深度可分离卷积神经网络[J/OL].北京航空航天大学学报，2024:1-15[2025-01-18]. https：//doi. org/10. 13700/j. bh. 1001-5965.2024. 0287.CHEN H,YAN J G,YANG H,et al. Deep separable convolutional neural network based on structural reparameterization[J/OL]. Journal of Beijing University of Aeronautics and Astronautics,2024:1-15[2025-01-18].https：//doi. org/10. 13700/j. bh. 1001-5965. 2024. 0287.(in Chinese)