HAN Ping1, WANG Tian kun1, MENG Yong yi2

（1.Shenhua Guoneng Energy Group Co., Ltd, Beijing 100033, China; 

2.Shanxi Luneng Hequ Power Generation Co., Ltd, Xinzhou 036504, China）

 

 

Abstract: Aiming at the problem of real time monitoring and fault diagnosis of primary air fan in thermal power plant, a data mining based Least Squares Support Vector Machine (LS SVM) primary air fan vibration estimation and fault early warning method was proposed due to it difficult to achieve fault diagnosis through the precise mechanism modeling because of the complex, changeable operation and the cross coupling process variables of the auxiliary equipment. The historical operation data of the 1# primary air fan of unit 1 in Hequ power plant was tested. The results indicate that the method has high estimation accuracy, and can identify the abnormal vibration of the primary air fan in time. The validity of the method is verified.

 

Key words: primary air fan; online monitoring; Least Square Support Vector Machines (LS SVM); fault prediction