Time-frequency image identification using non-negative matrix factorization and principal component analysis

Here,a method for machine condition identification was put forward based on non-negative matrix factorization(NMF) and principal component analysis(PCA).A vibration signal was used to construct a Hilbert two-dimensional time frequency image after pre-processing.Then,NMF was used to determine the feature vector for the time frequency image.Principal component analysis(PCA) was used to reduce the dimension number of the feature vector,it was useful for three-dimension condition identification.Different condition identifications of rolling bearing were as examples to testify the effectiveness of this method.It was concluded that this method can improve accuracy of machine condition identification;it is helpful for machine fault diagnosis.