Data Mining-based Vibration Fault Monitoring of Boiler Feed Pump in Thermal Plants
Current methods of monitoring vibration faults of boiler feed pumps are inadequate in monitoring accuracy,and unable to promptly identify potential hidden dangers in operating feed pumps.Based on the traditional vibration fault mo-nitoring methods,this paper puts forward a new feed pump vibration fault monitoring method by introducing data mining technology.The method realizes vibration fault monitoring by extracting frequency band characteristics of vibration fault through utilizing the principle of frequency division control technology,determining judging threshold for fault monitoring according to NSET fault monitoring model and similarity theory,and designing logic flow of fault monitoring using the da-ta mining technology.Experimental analysis shows that the proposed method can achieve monitoring results with accuracy higher than 95.47%and significantly improved monitoring speed,and has high feasibility.
data miningthermal power plantboiler feed pumpfault monitoringvibration fault