基于MF-DFA和SVM的齿轮箱故障诊断  

Fault Diagnosis of Gearbox based on MF-DFA and SVM

作　　者：刘春林[1] 潘宏侠[1] 史斐娜 蒋红军[1] 

机构地区：[1]中北大学机械与动力工程学院,山西太原030051

出　　处：《机械传动》2014年第12期119-123,共5页Journal of Mechanical Transmission

基　　金：国家自然科学基金(50875247)

摘　　要：针对齿轮箱的故障信号是具有非平稳性、非线性等复杂特征的信号,利用多重分形去趋势波动分析(MF-DFA)和支持向量机(SVM)相结合的方法对齿轮箱进行故障诊断。利用多重分形去趋势波动分析,提取齿轮箱故障的多重分形最大奇异指数,最小奇异指数,奇异谱的宽度,极值点所对应的奇异指数等4个分形参数,将其作为特征参数。然后建立齿轮箱的不同故障状态的样本,通过利用支持向量机的方法实现齿轮箱故障的诊断与识别。研究表明,这两种方法的结合为机械故障诊断提供了一种新的方法,对齿轮箱的故障诊断识别具有重要意义。The fault signal of the gearbox is the complex signal of the non--stationary and nonlin- ear characteristics, by using the multifractal detrended fluctuation analysis （MF--DFA） and the sup- port vector machine （SVM） to diagnose the fault of the gearbox. The multifractal detrended fluctuation analysis is employed to the extract multifractal singularity index of the gearbox fault, largest singular index, minimum singular index, the width of the singular spectrum, the singular index of extreme value point and so on, and this four fractal parameters are as characteristic parameter. And then the sample of the gearbox different fault state is built, by using the new method of support vector ma- chine, the gearbox fault diagnosis and identification is achieved. Through the study, a new method for mechanical fault diagnosis is provided through the combination of this two methods. There is of great significance for the fault diagnosis and identification of gearbox.

关 键 词：多重分形 去趋势波动分析 支持向量机 故障诊断 

分 类 号：TH132.41[机械工程—机械制造及自动化] TH165.3