The Research Of Fault Diagnosis Method Based On Fuzzy Neural Network For Rotating Machinery

In these years, machinery fault diagnosis technology has grown rapidly and research approaches and means are increasingly updated, all of which have been used in nearly every part of industrial field. However, for rotating machinery, it is very difficult to carry out fault diagnosis for its complicated structures and the ubiquitous fuzziness and complexity in character and causation of the fault. Although much research work has been done and some research fruits have been obtained, its diagnostic level is still low and does not match its wide application status. Therefore, the research on rotating machinery fault diagnosis is of great significance. The research work in this dissertation is started just in this technical background.Research on the feature extraction method of rotating machinery vibration signals. According to the time variation and feature extraction difficulty of rotating machinery vibration signals, the rule of choosing wavelet basis function and wavelet denoising soft-thresholding value is proposed after making further research on wavelet transform technique, using for denoising rotating machinery vibration signals The conception of "energy" is proposed, based on the theory that signals energy in all frequency can be affected by faults deeply, to construct feature vectors of rotating machinery vibration signals which can give a convenient disposal way to fault feature extraction and fault intellectual diagnosis. The vibration signals analytical result of rotating machinery mass imbalance and oil film turbulence fault verified its feasibility and validity.Researched on the fault diagnosis method of neural network and fuzzy system. Fuzzy system lacks self-study ability and its membership functions and fuzzy rule are chosen by experts subjectivity, and input/output relation obtained by neural network can not be expressed in acceptable way and exists the quality of absoluteness, all of which make diagnosis result not live up to the fact. So a rotating mechanical failure diagnosis method base on fuzzy neural network (ANFIS) is put forward and be applied to the fault diagnosis of rotating machinery. The experimental result indicates that this method, compared with the common one, can make up the shortcoming of the single-handed application of fuzzy classification or neural network. Moreover, it owns the better validity and popularity .It has a good application prospects in rotating machinery fault diagnosis.Based on the deep analysis of the diagnosis process for rotating machinery fault diagnosis system, the research has accomplished the design and developed system software of rotating machinery fault diagnosis prototype by the MATLAB language, which has strong function and toolboxes. The validity of the fault diagnosis system was well tested by eigenvectors, which are common appeared in rotating machinery fault status.