Study On Fault Diagnosis Of Rotating Machinery Based On Self-organizing Mapping Network

As one of the most widely used machine, the fault diagnosis for rotating machinery is of great practical significance and paramount importance. it has attracted a lot of researchers' interest on this problem at home and abroad.We do a lot of research on the fault mechanism of rotating machinery, and make our research focus on the subsequent two facets: feature extraction method and fault recognition method. This paper proposes a new feature extraction method for the fault vibration signal, and develops an improved self-organizing mapping network model for the fault pattern classification.In the aspect of feature extraction method study, with investigating the feature of impact factor in vibration signals and considering the non-placidity and non-linear of vibration diagnosis signals, we import wavelet analysis and fractal theory as the tools of faulty signal feature description. Experimental results proved the validity of this method. To some extent, this method provides a good approach of resolving the wholesome problem of fault feature symptom description.In the aspect of fault pattern classification study, considering such problems as: the complexity of rotating machinery, the samples of diagnosis signal not easy to obtain, more than one faults come up at the same time, we introduce self-organizing mapping(SOM) network as the pattern classification technique. Two dimension SOM network bears many merits such as: unsupervised, can work without standard model samples and reflect the topology fabric of the input samples, etc. This paper develops an improved model of SOM network. Experimental results proved it has higher right classification ratio compared with the classic SOM network. It overcomes many drawbacks of the classic arithmetic, besides, the output result is in the form of feature map which is easy to understand and make convenience for the diagnostic decision making.