摘要
针对当前火电机组热力系统间泄漏状况主要依靠人工监测,隐蔽泄漏不易被发现,易导致机组能耗无谓升高,影响机组运行的安全性及经济性的情况,提出基于模糊C均值聚类算法对系统运行数据进行挖掘,并将相关数据划分为正常状态或泄漏状态两类区域;再通过加动量误差反向传播(BP,back propagation)神经网络对系统实时表征参数进行分类。设计了火电机组多能级系统泄漏智能在线监测系统,在实际应用中该监测系统所监测火电机组热力系统的范围包含锅炉疏放水系统、高低压旁路系统等,涉及隔离界面150个以上。统计结果显示:监测成功率达81.8%,监测结果更加可靠。
At present,the leakage condition in thermal systems of thermal power units mainly depends on manual monitoring,but much hidden leakage is usually not easy to be detected,which leads to unnecessary increase of unit energy consumption and affects the safety and economy of unit operation.Based on the fuzzy C-means clustering algorithm,the system operation data is mined,and relevant data is divided into two types of areas concerning normal state or leakage state.Then the real-time characterization parameters of the system are classified by the momentum error back propagation(BP)neural network.An intelligent online leakage monitoring system for multi-level system of thermal power unit is designed,the thermal systems monitored by the monitoring system include boiler drainage system,high-low pressure bypass system,etc.in the practical application,involving more than 150 isolation interfaces.The statistical results show that the monitoring success rate is 81.8%and the monitoring result is more reliable.
作者
刘建东
丁启磊
张勇
LIU Jian-dong;DING Qi-lei;ZHANG Yong(Jianbi Power Plant of CHN Energy,Zhenjiang,China,Post Code:212006)
出处
《热能动力工程》
CAS
CSCD
北大核心
2022年第7期27-33,共7页
Journal of Engineering for Thermal Energy and Power
关键词
泄漏监测
模糊C均值聚类
BP神经网络
火电机组
leakage monitoring
fuzzy C-means clustering
BP neural network
thermal power unit
