ICA算法及其在齿轮箱故障诊断中的应用  

ICA algorithm and its application in gearbox fault diagnosis

作　　者：张睿[1,2] 黄晋英[2] 张永梅[3] 

机构地区：[1]中北大学信息与通信工程学院,山西太原030051 [2]中北大学机械工程与自动化学院,山西太原030051 [3]北方工业大学信息工程学院,北京100144

出　　处：《矿山机械》2013年第5期121-125,共5页Mining & Processing Equipment

基　　金：山西省自然科学基金项目(2009011026-1)

摘　　要：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。The application of vibration signal processing technology in the fault diagnosis of mechanical equipment has become the research focus in the field of engineering. In the practical engineering, the diagnosis information from operative equipment always has low ratio of signal to noise, aliasing of source signal and so on, which led to increase in the recognition difficulty and reduce in the fault diagnosis accuracy. The paper presented a method of diagnosing the gearbox fault based on ICA （independent component analysis）, and the deterministic mixed signal was applied to simulate and verify the algorithm. In addition, the algorithm was applied to dispose and analyze the vibration signal of the gearbox acquired by LMS. The results showed that the fault signal obviously boosted, which brought about higher diagnosis accuracy.

关 键 词：独立分量分析 齿轮箱 故障诊断 诊断精度 

分 类 号：TH165.3[机械工程—机械制造及自动化] TN911.7[电子电信—通信与信息系统]