Fault diagnosis method of centrifugal pump based on improved hierarchical sample entropy and extreme learning machine
To improve the accuracy of early fault diagnosis models for centrifugal pumps,a fault diag-nosis method for centrifugal pump based on the improved hierarchical sample entropy(IHSE)and ex-treme learning machine(ELM)was proposed.Firstly,aiming at the problem of weak algorithm stability of traditional hierarchical sample entropy at high level,a new time series signal complexity evaluation tool named IHSE was proposed by using moving average and moving difference process in-stead of the traditional hierarchical model.Secondly,IHSE was used to extract fault features of centri-fugal pump vibration signal.Finally,the fault features were input into ELM model to realize the effec-tive identification of different operating states of centrifugal pumps.The results show that the proposed method achieves diagnostic rates of 99.58％and 99.68％on two different types of centrifugal pump fault data sets,respectively,and performs the best among all diagnostic models,indicating that the proposed model has good diagnostic performance.This study provides a new method for fault diagnosis of centrifugal pump,and has good reference value and application prospects.
万方数据
维普
