Method Research On Fault Diagnosis Of Industrial Equipment Based On Fuzzy Multiple Attribute Decision Making

With the rapid development of industrial intelligent manufacturing industry,due to the long-term vibration and impact,the parts of industrial equipment are wear and aging.As a result,industrial equipment faults occur frequently.On the one hand,there is usually no one-to-one correspondence between the fault phenomenon and the cause of the fault,there is ambiguity in the diagnostic process.On the other hand,the cause of the fault is the result of the joint action of multiple fault feature,and different fault features have different ability to identify the same fault,which requires a comprehensive decision analysis.Therefore,the application of fuzzy multiple attribute decision making concepts in fault diagnosis methods has a strong exploration.This paper takes steam turbines and rolling bearings as research objects,and mainly studies the method of industrial equipment fault diagnosis based on fuzzy multiple attribute decision making.It mainly includes three aspects: firstly,how to deal with fuzzy fault feature data;secondly,how to select the optimal fault feature;thirdly,how to cluster the fault modes.Aiming at the above three problems,the main research contents of this paper are as follows:(1)Aiming at the problem that the fault feature cannot be expressed with accurate values,this paper proposes a method of fault diagnosis based on ordered triangular fuzzy numbers.First of all,this method is based on the theory of ordered triangular fuzzy numbers and improves the formula of Hamming nearness degree,Maximum and Minimum nearness degree and Euclidean nearness degree;then,determines the cause of the fault by comparing the test sample with the values of the nearness degree of each fault mode.This new method is applied to the fault diagnosis of steam turbines.The experimental results show that new method reduces the difficulty of fault diagnosis.(2)Aiming at the problem that the single evaluation criterion feature selection method cannot comprehensively evaluate the recognition degree of each fault feature to the fault mode,this paper proposes a fault feature selection method based on grey correlation vlsekriterijuska optimizacija i komoromisno resenje(GCVIKOR)multiple attribute decision making.First of all,this method introduces grey correlation degree to improves the vlsekriterijuska optimizacija i komoromisno resenje(VIKOR)method and proposes the multiple attribute decision making method;then,integrate the three evaluation criteria of Relief F algorithm,dispersion ratio method and distance method to establish a decision system of multiple features and multiple criteria(attributes);finally,according to the magnitude of the integrated value,the optimal fault feature are determined.This new method is applied to the fault feature selection of rolling bearings.The experimental results show that new method effectively improves the accuracy of fault feature selection which compared with other fault feature selection methods.(3)Aiming at the problem that the traditional fuzzy c-means clustering algorithm(FCM)is particularly sensitive to initial clustering prototypes and easy to fall into local optimization,and,the distance method is used to construct the fuzzy membership function without considering the influence of population variation on the distance,which reduces the accuracy of fault mode clustering,this paper proposes a nearness degree fuzzy c-means clustering fault diagnosis method based on particle swarm optimization algorithm(PSO)(PSO-NFCM).First of all,this method uses PSO to determine the clustering center of the training sample;then,introduces the nearness degree to improve the fuzzy membership degree formula and compare the fuzzy membership degree values of the test sample and the clustering center;finally,cluster fault modes.On the basis of feature selection,apply this new method to the fault pattern recognition of rolling bearings.The experimental results show that new method effectively improves the accuracy of fault diagnosis compared with the traditional FCM method.