针对轴承微弱故障声发射信号处理中,故障特征微弱难以有效提取识别的问题,引入多尺度主元坐标变换对故障特征进行增强。首先对信号进行小波包分解并重构各子频带分量,然后通过主成分分析将各分量变换到新的线性空间,实现信号故障特征的增强。通过仿真信号及实验信号对该方法进行验证,结果表明该方法在轴承故障增强检测中效果明显。 Aiming at the acoustic emission signal processing of weak fault bearing, it is difficult to effectively extract and identify the weak fault features. The multi-scale principal component coordinate transformation is introduced to enhance the fault features. Firstly, the signal is decomposed by wavelet packet and the sub-band components are reconstructed. Then the components are transformed into a new linear space by principal component analysis to realize the enhancement of signal fault features. The method is validated by simulation signal and experimental signal. The results show that the method is effective in bearing fault detection.