基于NARX的蒸汽发生器液位异常检测方法研究
DOI:
作者:
作者单位:

1.三峡大学理学院;2.中核武汉核电运行技术股份有限公司;3.三峡大学电气与新能源学院

作者简介:

通讯作者:

中图分类号:

TL48

基金项目:

中核集团核工业仿真技术重点实验室对外开放基金项目(B220631)


Research on Anomaly Detection of Steam Generator Water Level Based on NARX
Author:
Affiliation:

College of Science, China Three Gorges University

Fund Project:

  • 摘要
  • |
  • 图/表
  • |
  • 访问统计
  • |
  • 参考文献
  • |
  • 相似文献
  • |
  • 引证文献
  • |
  • 资源附件
  • |
  • 文章评论
    摘要:

    蒸汽发生器液位是评价核电机组运行状态的重要参数指标之一,由于传统预设固定液位报警阈值的监测方法无法在触发报警信号前及早发现异常,通过建立异常检测模型对蒸汽发生器液位进行状态监测和预警很有必要。基于蒸汽发生器复杂非线性系统的特点,将含外源输入的非线性自回归(NARX)方法引入,通过历史正常运行数据,建立正常工作模式下液位及相关参数间的耦合关系模型。模型以历史液位值和相关参数作为输入回归得到下一时刻的液位预测值,并通过预测值与实际观测值残差的大小,来判断蒸汽发生器多传感器系统当前工作状态是否异常。与触发预设液位阈值后再报警的传统状态监测方法相比,该方法能够检测到液位与相关参数间的耦合关系偏移,并在微小变化发生时就检测到异常,从而实现蒸汽发生器液位的状态监测和预警。经真实核电厂数据验证,该模型能够对液位实现准确的回归预测,并在依照真实故障类型构建的异常数据集验证实验中,取得了较好的异常检测效果。

    Abstract:

    The steam generator water level is one of the most important parameters for evaluating the operating status of a nuclear power unit. Since traditional condition monitoring methods with preset alarm thresholds do not allow for early detection of anomalies, it’s necessary to enable early warning and condition monitoring of SG water level through anomaly detection models. Due to the fact that the SG is a complex nonlinear system, historical normal operating data was used to build the coupling relationship model within water level values and relevant parameters by the methods based on nonlinear polynomial autoregressive models with exogenous inputs (NARX). The model makes use of historical water level values and the relevant parameters as input to obtain the water level regression prediction, by which the residuals from the observed values can be used to evaluate the operating status of SG water level. The method can detect anomalies as soon as small changes occur, enabling condition monitoring and early warning of SG water level, which differs from traditional methods with preset threshold alarms. The model has been validated with real nuclear plant data to achieve accurate regression predictions for SG water level and has achieved good detection results in validation experiments on the constructed anomaly data sets.

    参考文献
    相似文献
    引证文献
引用本文

周光荣,杨森权,郑胜,等. 基于NARX的蒸汽发生器液位异常检测方法研究[J]. 科学技术与工程, , ():

复制
分享
文章指标
  • 点击次数:
  • 下载次数:
  • HTML阅读次数:
  • 引用次数:
历史
  • 收稿日期:2023-10-15
  • 最后修改日期:2024-05-27
  • 录用日期:2024-05-29
  • 在线发布日期:
  • 出版日期:
×
亟待确认版面费归属稿件,敬请作者关注