Study, The Mechanical Fault Feature Extraction Method Based On Full Information

The vibration signal of mechanical equipment contains plenty of information.,which is very useful to determine the running status of equipment. The great use of feature extraction is to transform this information into diagnosis symptoms. Effective extraction of eigenvector is key tache on fault diagnosis.The development of technology makes the structure more and more complicated, it causes more and more fault category, and the fault conditions and diagnosis symptoms also increases accordingly. The importance of these characteristic parameters to fault diagnosis are different, even some characteristics is redundant. Redundant features not only waste diagnosis resources but directly influence the formation of concise and effective decision rules, and then reduce the accuracy and the rapidity of fault diagnosis. Therefore the extraction of characteristics which make the greatest contribution to the diagnosis and combination of the most effective characteristics were needed.Due to the insufficiency of traditional rotary machinery feature extraction with single channel signal and the deficiency of rapidity. A new rotary machinery fault feature extraction method combining full vector spectrum theory and rough set theory with ant colony algorithm is proposed.The Vector-spectrum based on the same source data fusion of rotary machinery is a joint name of vector spectrum and its patulous analysis methods. It's the data fusion technology based on data hierarchy, it fuse double or ternate channels' information according to some rules, and then finally come to conclusion that be in accord with actual condition. The introduction of information fusion technology makes up the defects of single channel analysis from the essence, and it brings out more reliable and more accurate monitoring and diagnosis.The rough set theory was proposed in the early 1980s, it is a kind of mathematical tool processing the fuzziness and the uncertainties. It did good execution in dealing with large amount of data and eliminating redundant information. It was usually used as the tools of data reduction. The ant colony algorithm simulating natural ant colony foraging is a kind of newest bionic optimization algorithm. It has obtained good results in solving all kinds of combinatorial optimization problems. Its strong point is it was not affected by the scale of the problems and in solving large-scale problems can still play a superior performance. The usefulness of rough set theory and ant colony algorithm in fault diagnosis of rotary machinery can eliminate the redundant information of the characteristic parameters, find out optimal combination, and extract effective rules of the fault diagnosis.The text combined Vector-spectrum with rough set theory and ant colony algorithm, discussed the basic principle of Vector-spectrum, the algorithm of rough set theory, the basic algorithm improved ant colony algorithm, and performed these functions with Matlab.