针对传统盲源分离(BSS)技术在处理非线性变化的复合故障信号时的不足,将核独立分量分析(KICA)技术应用于齿轮箱的状态检测与故障诊断。介绍了KICA的基本原理,进行了仿真说明。最后应用该方法并结合包络阶次谱分析对设置了2种故障的齿轮箱进行故障诊断,最终找到了故障特征,成功判别出了这2种故障,并验证了该方法的有效性。
Aiming at the shortage of the blind source separation (BSS) in processing non-linear compound fault signals, the kernel independent component analysis (KICA) was used in the performance test and fault diagnosis for the gearbox. The principle of KICA was introduced and then illustrated with simulations. Finally, combined with envelope order spectrum analysis, this method was used to diagnose the gearbox with two types of faults. The fault characteristics were found out, and the faults were distinguished successfully.