The Information Fusion Of Rotating Machinery Fault Diagnosis Based On Multi-vibration Signal

Rotating Machineries are used in industrial machinery widely.Once the equipments arc failure of running,it will cause enormous economic losses.With rotating machinery developing towards high-speed,high-power,high reliability and large,the information need to be considered is more and more and the difficulty of extracting information effectively is increased because of the mutual interference between Signals and signals,signals and noises, so the traditional fault diagnosis technology have failed adapting to complicated equipment fault diagnosis.Based on this consideration,the signal integration theory was used to fuse the useful information to enhance the accuracy and the reliability of fault diagnosis.In this paper,research on integration and fault diagnosis id made from the three levels of vibration signals based on information theory and technology integration.On this basis,a detailed Description of the methods to every signal integration layers' achieving are made.1.The fault diagnosis method of Multi - sensors vibration signal's data fusion based on correlation.A signal's data fusion based on correlation function is proposed for data-level integration to improve the signal accuracy,fault-tolerant.In conjunction with the examples were analyzed.Through to carry on the practical application in revolves diagnosis,indicated that this method is feasible and effective in the practical application.2.The fault diagnosis method of characteristics information fusion based on neural network technology.The process of network design is described in detail.A method combining the wavelet packet decomposition and Empirical Mode Decomposition(EMD)is used to extract the feature of the signals,First signal was decomposed into a series of narrow-band signals using wavelet packet decomposition and feature signals is fall on the different frequency band generally.Then various band signals gottcn from wavelet packet decomposition are made further decomposed using EMD decomposition and signal characteristics are extracted from the Intrinsic Mode Function(IMF)component which the feature signals are contained.The extraction of the signal features are put into the neural networks separately,and then the feature-signal fusion results are gained.The detailed process description is made through a combination of spccific cxalnples.3.The fault diagnosis method of decision-making information fusion based on D-S evidence theory.The detailed description on the basic concepts theory of D-S evidence theory and the integration and reasoning methods of D-S evidence is made in this paper.A decision-making level signal fusion is set up to make a further integration base on the Diagnosis of neural network and draw a final conclusion.Finally this method is proved to be high reliability and low uncertainty by multi-characteristie fusing.