Fault Diagnosis Of Aero-engine Rotor System Based On Grid Refinement Wave Cluster

With the enhancement of our national economic strength and the rising international status, the development step of the modern aviation industry speeds up gradually, at the same time,the aircraft engine performance improves constantly, the structure design of aircraft engine becomes complex increasingly, the probability of various fault will be rising. Thus, the requirements of its stability and reliability are also becoming higher and higher. The rotor system vibration faults become one of the main causes of the aero-engine faults, and often lead to serious vibration problems, or even lead to the machine crash accident. Therefore, the research on faults diagnosis of the rotor system, not only has important academic value, but also has a very important practical significance to economic and the national security.In this paper, the wavelet clustering algorithm is regarded as the vibration fault diagnosis method of the aero-engine rotor system, after analyzing the accuracy of fault diagnosis which is expressed as clustering quality in the MATLAB environment, the improved wavelet clustering algorithm was proposed in order to determine fault types more accurately and improve the clustering quality. The grid separation technology is applied to the improved wavelet clustering algorithm. In the implementation process, refining the grids in the feature space can extract the boundary points. Finally, the theoretical research and experimental verification of the clustering quality of the before and after improved algorithm has been carried on.This paper, firstly, introduces the purpose and significance of the research, and also describes the present situation and the development trend of aero-engine fault diagnosis. Secondly, the aero-engines are classified, and introduces several kinds of aero-engine. Then, fault reasons and vibration characteristic of several common aero-engine rotor failures are introduced in detail. Thirdly, the specific improvement methods are given on the basis of comprehending the fundamental principle of wavelet clustering algorithm. Fourthly, the grid separation technology and the two-dimensional characteristic are introduced in detail. The complexity of the algorithm is also analyzed, and what’s more, the concept of cellular arrays is cited to classify and label the grid cell. Last but not least, the rotor system fault simulation experiment device and the experimental methods of several common vibration faults are introduced; The rotor fault data which were collected by the rotor system simulation experience can be used to verify the feasibility and effectiveness of the improved wavelet cluster algorithm. The final experimental results show that the grid refinement wavelet cluster algorithm can get the better clustering quality than before,and the fault diagnosis result of rotor system vibration will be more accurate.