基于核主元分析法的核电厂设备状态监测技术研究
Research on Condition Monitoring Technology for Nuclear Power Plant Equipment Based on Kernel Principal Component Analysis
-
摘要: 为解决核电厂传统监测手段的局限性,提出将核主元分析法(KPCA)引入核电厂设备在线监测领域中,并设计了监测模型建设方法以及在线监测策略。为验证算法的有效性,将其应用在国内某核电机组电动主给水泵的真实监测案例中。仿真结果表明,KPCA算法可适应核电厂设备监测的要求,能比现有阈值监测手段提供更为早期的故障预警。同时,相比于常规的主元分析法(PCA),KPCA算法能够提取各变量之间的非线性关系,识别出设备不同的运行模式,有效减少误报警。Abstract: In order to solve the limitations of the traditional monitoring methods for nuclear power plants, this paper proposes to introduce Kernel Principal Component Analysis (KPCA) into the online monitoring field of nuclear power plant equipment, and design the monitoring method and online monitoring strategy. In order to verify the effectiveness of the algorithm, it has been applied in the real monitoring case of the motor driven main feed water pump a nuclear power plant in China. The simulation results show that the KPCA algorithm can adapt to the requirements of nuclear power plant equipment monitoring, and can provide earlier warnings of failure than the existing threshold monitoring methods. At the same time, compared with the conventional PCA algorithm, the KPCA algorithm can extract the nonlinear relationship between variables, identify different operating modes of the device, and effectively reduce false alarms.
点击查看大图
计量
- 文章访问数: 328
- HTML全文浏览量: 56
- PDF下载量: 8
- 被引次数: 0