Aiming at the problem of weak feature extraction and recognition in the bearing acoustic emission signal processing

multi-scale principal component coordinate transformation is proposed to enhance fault feature.Firstly

the signal is decomposed by wavelet package

and the sub-band reconstruction component is converted to a new linear space by principal component analysis

and the signal fault feature is enhanced.Finally

the method performance is verified by simulation signal and testing signal

the results show that multi-scale principal component coordinate transformation has obvious enhancement effect in bearing fault detection.