关键词: 马氏距离, 阴性选择, 风电机组, 故障诊断

Abstract:

In order to correctly judge wind turbine vibration fault types,reduce  its impacts  on power generation efficiency and personal property safety,this paper proposed  an improved algorithm.In the traditional negative selection algorithm  Mahalanobis distance  was introduced for preliminary screening of vibration data,and the prediction   algorithm was applied to the wind turbine vibration faults.At the same time with the  two algorithms the diagnostic accuracy of fault diagnosis  was  improved.Studies show that the improved negative selection algorithm can be more quickly and accurately determine the fault type units and the diagnostic accuracy of 97.5% are achieved.And it can improve the wind turbine operation reliability and power efficiency.

Key words: Mahalanobis distance;negative selection;wind , turbine;fault diagnosis