基于数据挖掘的锅炉在线运行状态监测
|
摘要
为了解锅炉的运行状态,提出了一种基于数据挖掘的锅炉在线运行状态的监测方法:利用K-means算法对锅炉正常运行时的工况进行划分,然后依据距离度量的状态监测思想,在各工况内计算每个采样点的马氏距离,利用高斯阈值构建决策函数,得到锅炉的实时健康指标。利用重庆某新能源技术有限公司提供的2台燃气锅炉半年的运行数据,对该状态监测方法进行验证,表明该方法能够准确地跟踪锅炉的运行状态,在锅炉发生异常时及时报警。
This paper proposes an online operational state monitoring method of boiler based on data mining in order to monitor the operational status of boiler. Firstly,the K-means clustering algorithm is used to divide the operating condition of boiler using historical data under normal operation. Then,according to the distance measurement of condition monitoring,the Mahalanobis distance at each sampling point is calculated in each subspace. The decision function is constructed based on Gaussian threshold in each subspace. Finally,the real-time health index( HI) of boiler is obtained according to the corresponding decision function. We use the operational report data provided by a new energy technology company in Chongqing to verify the method proposed in this paper. The results show that the method can track the operational status of the boiler accurately and send out alarm timely when the boiler is abnormal.
This paper proposes an online operational state monitoring method of boiler based on data mining in order to monitor the operational status of boiler. Firstly,the K-means clustering algorithm is used to divide the operating condition of boiler using historical data under normal operation. Then,according to the distance measurement of condition monitoring,the Mahalanobis distance at each sampling point is calculated in each subspace. The decision function is constructed based on Gaussian threshold in each subspace. Finally,the real-time health index( HI) of boiler is obtained according to the corresponding decision function. We use the operational report data provided by a new energy technology company in Chongqing to verify the method proposed in this paper. The results show that the method can track the operational status of the boiler accurately and send out alarm timely when the boiler is abnormal.
引文
[1]胡昌华,施权,司小胜,等.数据驱动的寿命预测和健康管理技术研究进展[J].信息与控制,2017,46(1):72-82.HU Chang-hua,SHI Quan,SI Xiao-sheng,et al. Data-driven lifeprediction and health management:state of the art[J]. Informa-tion and Control,2017,46(1):72-82.
[2] HE W,MIAO Q,AZARIAN M,et al. Health monitoring of coolingfan bearings based on wavelet filter[J]. Mechanical Systems andSignal Processing,2015,64:149-161.
[3] HU Y,LI H,LIAO X,et al. A probability evaluation method ofearly deterioration condition for the critical components of wind tur-bine generator systems[J]. Mechanical Systems and Signal Pro-cessing,2016,76:729-741.
[4] BANGALORE P,TJERNBERG L B. An artificial neural networkapproach for early fault detection of gearbox bearings[J]. IEEETransactions on Smart Grid,2015,6(2):980-987.
[5]严英杰,盛戈皞,陈玉峰,等.基于大数据分析的输变电设备状态数据异常检测方法[J].中国电机工程学报,2015,35(1):52-59.YAN Ying-jie,SHENG Ge-hao,CHEN Yu-feng,et al. An methodfor anomaly detection of state information of power equipment basedon big data analysis[J]. Proceedings of the CSEE. 2015,35(1):52-59.
[6]郑心伟,孙瑜,王晓军.增压锅炉汽包低周疲劳寿命计算方法研究[J].热能动力工程,2010,25(2):181-189.ZHENG Xin-wei,SUN Yu,WANG Xiao-jun. Study of the methodsfor calculating the low-cycle fatigue life of a supercharged boilerdrum[J]. Journal of Engineering for Thermal Energy and Power,2010,25(2):181-189.
[7] NIKULA R P,RUUSUNEN M,LEIVISKK. Data-driven frame-work for boiler performance monitoring[J]. Applied Energy,2016,183:1374-1388.
[8]黄景涛,任志伟,罗威.电站锅炉监测数据的异常值检测算法研究[J].计算机与应用化学,2013(10):15.HUANG Jing-tao,REN Zhi-wei,LUO Wei. Research on abnormalvalue detection algorithm for monitoring data of boiler[J]. Com-puters and Applied Chemistry,2013(10):15.
[9] WU G,WU C,XU S. Application of wireless sensor network in themonitoring system of boiler[C]//The 26th Chinese Control andDecision Conference(2014 CCDC),Changsha,IEEE,2014:1751-1754.
[10] ROMEO L M,GARETA R. Neural network for evaluating boilerbehaviour[J]. Applied Thermal Engineering,2006,26(14):1530-1536.
[11]付冬梅,张雪松,刘同锋,等.基于层次分析和模糊原理的燃油锅炉状态评估系统的研究[J].锅炉技术,2006,37(B4):20-23.FU Dong-mei,ZHANG Xue-song,LIU Tong-feng,et al. Researchon oil boiler state evaluation system based on analytic hierarchyprocess and fuzzy principle[J]. Boiler and Technology,2006,37(B4):20-23.
[12]高飞.基于数传电台与PLC的锅炉状态预警系统设计[J].山西科技,2015(3):127-129.GAO Fei. Design of boiler state early warning system based ondigital radio and PLC[J]. Shanxi Science and Technology,2015(3):127-129.
[13] MACKIEWICZ A,RATAJCZAK W. Principal components analy-sis(PCA)[J]. Computers and Geosciences,1993,19(3):303-342.
[14]杨宏宇.锅炉工况变动对锅炉机组的影响[J].电力学报,2009,24(3):256-258.YANG Hong-yu. Influence of boiler operating conditions on boilerunit[J]. Journal of Electric Power,2009,24(3):256-258.
[15] Rokach L. A survey of clustering algorithms[M]. Boston:Spring-er,2009:269-298.
[16]冯超. K-means聚类算法的研究[D].大连:大连理工大学,2007.FENG Chao. Research on K-means clustering algorithm[D].Dalian:Dalian University of Technology,2007.
[17]朱连江,马炳先,赵学泉.基于轮廓系数的聚类有效性分析[J].计算机应用,2010(12):139-141.ZHU Lian-jiang,MA Bing-xian,ZHAO Xue-quan. Clustering va-lidity analysis based on silhouette coefficient[J]. Journal of Com-puter Applications,2010(12):139-141.
[18]林彬,宋东,和麟.基于马氏距离与组距估计的复杂系统健康评估[J].仪器仪表学报,2016,37(9):2022-2028.LIN Bin,SONG Dong,HE Lin. Complex system health assess-ment based on Mahalanobis distance and bin-width estimationtechnique[J]. Journal of Scientific Instrument,2016,37(9):2022-2028.
[19] XIANG S,NIE F,ZHANG C. Learning a Mahalanobis DistanceMetric for Data Clustering and Classification[J]. Pattern Recog-nition,2008,41(12):3600-3612.
[20] WANG Y,MIAO Q,MA E W M,et al. Online anomaly detectionfor hard disk drives based on Mahalanobis distance[J]. IEEETransactions on Reliability,2013,62(1):136-145.
[21] JIN X,MA E W M,CHENG L L,et al. Health monitoring of cool-ing fans based on Mahalanobis distance with m RMR feature selec-tion[J]. IEEE Transactions on Instrumentation and Measure-ment,2012,61(8):2222-2229.
[22]胡雷,范彬,胡茑庆,等.变工况下轴承健康监测的相关向量机与自适应阈值模型方法[J].国防科技大学学报,2016,35(1):168-174.HU Lei,FAN Bin,HU Niao-qing,et al. Bearings health monito-ring under varying operation conditions using relevance vector ma-chine and adaptive threshold model[J]. Journal of National Uni-versity of Defense Technology,2016,35(1):168-174.
[2] HE W,MIAO Q,AZARIAN M,et al. Health monitoring of coolingfan bearings based on wavelet filter[J]. Mechanical Systems andSignal Processing,2015,64:149-161.
[3] HU Y,LI H,LIAO X,et al. A probability evaluation method ofearly deterioration condition for the critical components of wind tur-bine generator systems[J]. Mechanical Systems and Signal Pro-cessing,2016,76:729-741.
[4] BANGALORE P,TJERNBERG L B. An artificial neural networkapproach for early fault detection of gearbox bearings[J]. IEEETransactions on Smart Grid,2015,6(2):980-987.
[5]严英杰,盛戈皞,陈玉峰,等.基于大数据分析的输变电设备状态数据异常检测方法[J].中国电机工程学报,2015,35(1):52-59.YAN Ying-jie,SHENG Ge-hao,CHEN Yu-feng,et al. An methodfor anomaly detection of state information of power equipment basedon big data analysis[J]. Proceedings of the CSEE. 2015,35(1):52-59.
[6]郑心伟,孙瑜,王晓军.增压锅炉汽包低周疲劳寿命计算方法研究[J].热能动力工程,2010,25(2):181-189.ZHENG Xin-wei,SUN Yu,WANG Xiao-jun. Study of the methodsfor calculating the low-cycle fatigue life of a supercharged boilerdrum[J]. Journal of Engineering for Thermal Energy and Power,2010,25(2):181-189.
[7] NIKULA R P,RUUSUNEN M,LEIVISKK. Data-driven frame-work for boiler performance monitoring[J]. Applied Energy,2016,183:1374-1388.
[8]黄景涛,任志伟,罗威.电站锅炉监测数据的异常值检测算法研究[J].计算机与应用化学,2013(10):15.HUANG Jing-tao,REN Zhi-wei,LUO Wei. Research on abnormalvalue detection algorithm for monitoring data of boiler[J]. Com-puters and Applied Chemistry,2013(10):15.
[9] WU G,WU C,XU S. Application of wireless sensor network in themonitoring system of boiler[C]//The 26th Chinese Control andDecision Conference(2014 CCDC),Changsha,IEEE,2014:1751-1754.
[10] ROMEO L M,GARETA R. Neural network for evaluating boilerbehaviour[J]. Applied Thermal Engineering,2006,26(14):1530-1536.
[11]付冬梅,张雪松,刘同锋,等.基于层次分析和模糊原理的燃油锅炉状态评估系统的研究[J].锅炉技术,2006,37(B4):20-23.FU Dong-mei,ZHANG Xue-song,LIU Tong-feng,et al. Researchon oil boiler state evaluation system based on analytic hierarchyprocess and fuzzy principle[J]. Boiler and Technology,2006,37(B4):20-23.
[12]高飞.基于数传电台与PLC的锅炉状态预警系统设计[J].山西科技,2015(3):127-129.GAO Fei. Design of boiler state early warning system based ondigital radio and PLC[J]. Shanxi Science and Technology,2015(3):127-129.
[13] MACKIEWICZ A,RATAJCZAK W. Principal components analy-sis(PCA)[J]. Computers and Geosciences,1993,19(3):303-342.
[14]杨宏宇.锅炉工况变动对锅炉机组的影响[J].电力学报,2009,24(3):256-258.YANG Hong-yu. Influence of boiler operating conditions on boilerunit[J]. Journal of Electric Power,2009,24(3):256-258.
[15] Rokach L. A survey of clustering algorithms[M]. Boston:Spring-er,2009:269-298.
[16]冯超. K-means聚类算法的研究[D].大连:大连理工大学,2007.FENG Chao. Research on K-means clustering algorithm[D].Dalian:Dalian University of Technology,2007.
[17]朱连江,马炳先,赵学泉.基于轮廓系数的聚类有效性分析[J].计算机应用,2010(12):139-141.ZHU Lian-jiang,MA Bing-xian,ZHAO Xue-quan. Clustering va-lidity analysis based on silhouette coefficient[J]. Journal of Com-puter Applications,2010(12):139-141.
[18]林彬,宋东,和麟.基于马氏距离与组距估计的复杂系统健康评估[J].仪器仪表学报,2016,37(9):2022-2028.LIN Bin,SONG Dong,HE Lin. Complex system health assess-ment based on Mahalanobis distance and bin-width estimationtechnique[J]. Journal of Scientific Instrument,2016,37(9):2022-2028.
[19] XIANG S,NIE F,ZHANG C. Learning a Mahalanobis DistanceMetric for Data Clustering and Classification[J]. Pattern Recog-nition,2008,41(12):3600-3612.
[20] WANG Y,MIAO Q,MA E W M,et al. Online anomaly detectionfor hard disk drives based on Mahalanobis distance[J]. IEEETransactions on Reliability,2013,62(1):136-145.
[21] JIN X,MA E W M,CHENG L L,et al. Health monitoring of cool-ing fans based on Mahalanobis distance with m RMR feature selec-tion[J]. IEEE Transactions on Instrumentation and Measure-ment,2012,61(8):2222-2229.
[22]胡雷,范彬,胡茑庆,等.变工况下轴承健康监测的相关向量机与自适应阈值模型方法[J].国防科技大学学报,2016,35(1):168-174.HU Lei,FAN Bin,HU Niao-qing,et al. Bearings health monito-ring under varying operation conditions using relevance vector ma-chine and adaptive threshold model[J]. Journal of National Uni-versity of Defense Technology,2016,35(1):168-174.