Fault prediction method of centrifugal pump in water injection station based on identification rules
In view of the inadequacy of the traditional manual diagnostic methods such as"looking,hearing,inquiring and touching"used in oilfield water injection stations for analyzing the operational status of centrifugal pump units,an intelligent fault diagnosis method based on identification rules is proposed in this paper,which realize real-time fault diagnosis of centrifugal pumps.Firstly,parameters from the normal operation and fault states of the centrifugal pumps at the water injection station are collected.Then,by analyzing the trend of these parameter changes under both conditions and comparing them,the characteristic differences in parameter variations between the normal and abnormal states of the centrifugal pumps are identified.Subsequently,by examining the change rate and threshold value of these parameters,the boundary conditions for detecting anomalies in the centrifugal pump units were established.Finally,a set of rules for identifying centrifugal pump faults is developed,enabling intelligent prediction of centrifugal pump faults.Field validation of this method at oilfield centrifugal pump units shows that the accuracy rate of predicting faults reaches 99.5％,ensuring the normal production and operation of the centrifugal pump units.
万方数据
维普
