The Technology Research Of Fault Detection Of Gear Based On SVD And ELM

Gear box is a kind of key components of rotating machinery.Whether it could operate normally or not maked a significant influence on the normal work of the whole set of mechanical equipment and even the complete set of machinery.The gear box decide other components whether operate normally or not because it is use in special environment.Diagnosing of the different fault feature during the operation process of gear box accurately,identifying the characteristics of them in time,making the fast diagnosis for fault effectively and replacing the defective parts in time have the very vital importance on the improvement of reliability of the overall rotating machinery operation.This article describe the applications of the ensemble empirical mode decomposition,singular value decomposition and Extreme Learning Machine in the field of gear box of fault diagnosis and recognition.Based on studying of gear box of fault diagnosis and recognition model,the new methods of EEMD-SVD and EEMD-ELM are proposed.In this method,the data is decomposed by EEMD and a group of intrinsic mode function are obtained,which is selected and reconstructed effectively.Reconstructed signals are used to create a Hankel matrix,then SVD operation of each matrix is made to obtain its orthogonal decomposition results.furthermore,the singular values are selected and then by dint of difference spectrum of singular value for the SVD reconstruction.The feature information can be extracted through these procedures.It extracts the energy of the IMF that has larger correlation with the original signal.Then,fault classification model of extreme learning machine of gear is established.Afterwards the feature vector of energy index being inputed to the model.Finally,the method of EEMD-SVD can extract the feature information of gear box effectively.The experimental results show gear fault diagnosis method of extreme learning machine has faster speed and higher classification accuracy by the contrast of SVM.