摘要
为实现锅炉辅机设备潜在故障预警功能,保障锅炉系统正常运行,提出基于偏互信息和改进多元状态估计技术的故障预警方法。偏互信息法被用于提取故障特征信息及筛选故障相关变量;改进多元状态估计技术被用于构建辅机设备正常状态模型,采用概率分布采样法代替传统采样法构造过程记忆矩阵。采用观测状态与模型估计正常状态之间的滑动平均相似度作为故障预警标准,当相似度低于阈值时触发警报。以循环流化床锅炉引风机为例,结果表明该方法能够有效预警辅机设备故障。
This paper proposed a fault early warning method based on partial mutual information and improved multivariate state estimation technology in order to realize the potential fault early warning function of boiler auxiliary equipment and ensure the normal operation of boiler system.The partial mutual information method helped extract fault feature information and screen fault-related variables.The improved multivariate state estimation technique was for constructing the normal state model of auxiliary equipment.Besides,this paper substituted probability distribution sampling method for the traditional sampling method to construct the process memory matrix.This paper defined the fault early warning standard as the sliding average similarity between the observed state and the normal state estimated by improved multivariate state estimation technique.When the similarity is lower than the threshold,the alarm will be triggered.This paper verified the effectiveness of the method in early warning the fault of auxiliary equipment by an induced draft fan of a circulating fluidized bed boiler.
作者
张维
高明明
伯运鹤
翟海涛
ZHANG Wei;GAO Mingming;BO Yunhe;ZHAI Haitao(State Key Lab of Alternate Electrical Power System with Renewable Energy Sources,North China Electric Power University,Beijing 102206,China;Linhuan Joinlion Power Co.,Ltd.,Huaibei 235039,China)
出处
《华北电力大学学报(自然科学版)》
CAS
北大核心
2019年第6期73-80,共8页
Journal of North China Electric Power University:Natural Science Edition
基金
国家重点研发计划资助(2016YFB0600200)
中央高校基本科研业务费专项资金项目资助(2017XS072,2018ZD05,2019MS019)
关键词
偏互信息
改进多元状态估计技术
辅机设备
故障预警
partial mutual information
improved multivariate state estimation technology
auxiliary equipment
fault early warning