基于小波去噪的冷水机组传感器故障检测  

Chiller sensor fault detection using wavelet de-noising

作　　者：胡云鹏[1] 陈焕新[2] 周诚[2] 徐荣吉[3] 

机构地区：[1]华中科技大学环境科学与工程学院,湖北武汉430074 [2]华中科技大学能源与动力工程学院,湖北武汉430074 [3]北京建筑工程学院供热供燃气通风及空调工程北京市重点实验室,北京100044

出　　处：《华中科技大学学报（自然科学版）》2013年第3期16-19,24,共5页Journal of Huazhong University of Science and Technology(Natural Science Edition)

基　　金：国家科技支撑计划资助项目(2008BAJ12B03);供热供燃气通风及空调工程北京市重点实验室研究基金资助课题(NR2012K06);2010年度湖北省建设科技项目;2011年度湖北省建设科技项目

摘　　要：对基于主元分析方法的冷水机组传感器故障检测效率取决于训练数据和被测数据的质量的问题进行了研究.采用小波变换剔除测量数据中的噪声,提高数据质量,从而提高故障检测效率.结果表明:在-1.0℃故障下,基于小波去噪的主元分析方法的故障检测效率达到了91%.在同等数值的正负偏差故障下,基于小波去噪的主元分析方法的故障检测效率对称性更好.故障检测效率与小波基函数的分解层次关系密切.分解层次越多,故障检测效率越高.所有的db族小波基函数在5层分解的情况下,-0.5℃故障下的检测效率均能达到90%以上.Chiller sensor fault detection based on principal component analysis is a data-based analysis method. The fault detection efficiency relies on the quality of the training data and the measured data. The measurement noise was removed by wavelet transfer. The fault detection efficiencies were promoted because of the promotion of the data quality. Results show that the fault detection efficiency is 91G on the --1.0 ℃ introduced fault level by the PCA-based method combined with wavelet de-noising. On the same values of the positive and negative fault levels, the symmetry of the presented method is well than the normal PCA-based method. The fault detection efficiencies rely on the decomposed layer of the wavelet transfer. The more decomposed layers are, the well the fault detection efficiencies are. On the --0.5 ℃ fault level, the fault detection efficiencies of all the db＇s wavelet function on the 5 layers decomposition are greater than 90%.

关 键 词：主元分析 小波变换 传感器 故障检测 噪声 冷水机组 

分 类 号：TB65[一般工业技术—制冷工程]