论文部分内容阅读
近年来振动信号处理技术在机械设备故障诊断中的应用已经成为研究热点。在实际工程中,设备运行过程得到的诊断信息往往存在信噪比低及源信号混叠等问题,因而加大了识别难度,降低了故障诊断精度。笔者提出了一种基于独立成分分析(Independent Component Andlycis,ICA)的齿轮箱机械故障识别方法,应用确定性混合信号对算法进行了仿真验证,并用该算法对最小均方自适应(Least Mean Square,LMS)采集到的齿轮箱振动时域信号进行处理分析。结果表明,经该算法处理后故障信息明显增强,故障诊断精度也相对提高。
In recent years, the application of vibration signal processing technology in mechanical equipment fault diagnosis has become a research hotspot. In practical engineering, the diagnostic information obtained during the operation of the equipment often has problems of low signal-to-noise ratio and source signal aliasing, thereby increasing the difficulty of identification and reducing the accuracy of fault diagnosis. This paper presents a mechanical component identification method based on Independent Component Andlycis (ICA). The algorithm is verified by using deterministic mixed signal. The algorithm is applied to the Least Mean Square LMS) collected gear box vibration time domain signal processing analysis. The results show that the fault information is obviously enhanced after the algorithm is processed, and the fault diagnosis accuracy is also improved.