基于相空间重构和聚类分析的空气预热器热点检测  

Hot Spot Detection Based on Phase Space Reconstruction and Cluster Analysis of the Air Preheater

作　　者：苏芳[1] 董红生[1] 

机构地区：[1]兰州工业学院电气工程学院,甘肃兰州730050

出　　处：《兰州工业学院学报》2015年第6期77-83,共7页Journal of Lanzhou Institute of Technology

摘　　要：结合相空间重构理论和聚类分析法提出了一种新的空气预热器热点检测方法.该方法首先利用相空间重构理论对空气预热器的温度时间序列进行相空间重构,然后通过提出的一种改进聚类算法对重构后的温度信号进行聚类,并结合空气预热器实际运行工况对聚类结果进行分析,获得热点对应的类别,以此来实现一种新的空气预热器热点检测方法.仿真结果表明,聚类异常检测思想应用于空气预热器的热点检测的行性且有效,所提出的空气预热器热点检测法适用于在线热点检测.A new method for air preheater hot spot detection was proposed by combining phase space reconstruc- tion and fuzzy clustering algorithm. This algorithm firstly reconstructed the temperature sequence of air preheater by using the theory of phase space reconstruction, then through an improved clustering algorithm reconstructed temperature signal clustering, combined with the actual air preheater operating conditions implemented a new air preheater hot spot detection methods. Simulation results show that clustering anomaly detection used in air pre- heater hot spot detection is feasible and effective, the proposed air preheater hot spot detection method in this pa- per is suitable for online hot spot detection.

关 键 词：空气预热器 热点 混沌 相空间重构 聚类 

分 类 号：TP277[自动化与计算机技术—检测技术与自动化装置]