基于改进MSET的一次风机故障预警及诊断方法  

Early warning and diagnosis of primary fans based on improved MSET

作　　者：余兴刚 宾谊沅 陈文 魏鑫 刘明[3] 邱斌斌[3] YU Xing-gang;BIN Yi-yuan;CHEN Wen;WEI Xin;LIU Ming;QIU Bin-bin(Hunan Province Key Laboratory of Efficient&Clean Power Generation Technologies,State Grid Hunan Electric Power Corporation Limited Research Institute,Changsha 410007,China;Hunan Xiangdian Test&Research Institute Co.,Ltd.,Changsha 410004,China;State Key Laboratory of Multiphase Flow in Power Engineering,Xi'an Jiaotong University,Xi'an 710049,China)

机构地区：[1]国网湖南省电力有限公司电力科学研究院高效清洁发电技术湖南省重点实验室,湖南长沙410007 [2]湖南省湘电试验研究院有限公司,湖南长沙410004 [3]西安交通大学动力工程多相流国家重点实验室,陕西西安710049

出　　处：《机电工程》2023年第4期535-541,共7页Journal of Mechanical & Electrical Engineering

基　　金：国家自然科学基金资助项目(52022079)。

摘　　要：针对电站风机故障预警和故障点追溯等问题,基于改进多元状态估计技术(MSET)模型和误差分量,提出了一次风机故障预警和诊断方法。首先,介绍了多元状态估计技术的概念;选取了建模变量,并进行了数据预处理,通过构建改进动态记忆矩阵D,进行了MSET模型有效性的验证;然后,使用相似度函数作为故障预警依据,并利用滑动窗口法降低了噪音干扰,采用两种动态记忆矩阵构建方法,分别建立了模型,并进行了模型的有效性验证;最后,采用人为增加扰动的方式,模拟了温度、振动故障数据,进行了故障预警模拟,并通过计算各参数的误差分量进行了故障点追溯。研究结果表明:改进的动态记忆矩阵建模方法具有更高的准确度和更强的抗干扰能力;改进MSET和误差分量模型可成功实现故障的提前预警和故障点追溯功能。该模型能够为电站设备故障预警和检修提供借鉴。Aiming at the problems of wind turbine fault early warning and fault point tracing in power station,a wind turbine fault early warning method based on improved multivariate state estimation technique(MSET)and error component was proposed.Firstly,the multivariate state estimation technology(MSET)was introduced.The modeling variables were selected,data preprocessing was performed,and the validity of model validity was verified by constructing an improved dynamic memory matrix D.Then,the similarity function was used as the basis for fault early warning,the sliding window method was used to reduce noise interference,and two kinds of dynamic memory matrix construction methods were used to build models and verify their effectiveness.Finally,the temperature and vibration fault data were simulated by artificially increasing the disturbance to carry out fault warning simulation,and the fault point was traced by calculating the error components of each parameter.The results show that the improved dynamic memory matrix modeling method has higher accuracy and stronger anti-interference ability.The improved MSET and error component model can successfully realize the early warning of faults and the traceability of fault points.This model can provide effective guidance for early warning and maintenance of power station equipment failures.

关 键 词：离心式鼓风机 多元状态估计技术 误差分量 故障点追溯 动态记忆矩阵建模方法 人为增加扰动 相似度函数 

分 类 号：TH442[机械工程—机械制造及自动化] TM62[电气工程—电力系统及自动化]