基于EMD与多重分形去趋势法的轴承智能诊断方法

来源期刊:中南大学学报(自然科学版)2015年第2期

论文作者:贾峰 武兵 熊晓燕 熊诗波

文章页码:491 - 498

关键词:多重分形;去趋势波分析;经验模态分解;遗传算法;支持向量机

Key words:multifractal; detrended fluctuation analysis; empirical mode decomposition; genetic algorithm; support vector machine

摘    要:引入经验模态分解(EMD)方法去除故障信号的趋势项,提出EMD-MFDFA(multifractal detrended fluctuation analysis)的多重分形分析方法,并通过仿真分析EMD方法去趋势效果的有效性。然后将采用EMD-MFDFA方法提取的电机滚动轴承振动信号的多重分形特征向量作为训练集,利用混合遗传算法搜索全局最优的能力优化支持向量机(SVM)模型参数,建立电机滚动轴承的智能诊断模型。最后,通过对电机滚动轴承不同状态的振动信号进行分析。研究结果表明:EMD-MFDFA方法能很好地揭示滚动轴承的振动信号多重分形特性,对滚动轴承正常状态、单一故障与多故障耦合等状态具有很强的辨识能力;所建立的智能诊断模型可以有效地诊断滚动轴承不同的故障状态,能够作为滚动轴承故障在线监测的有效工具。

Abstract: The empirical mode decomposition (EMD) method was introduced to remove the fault signal local trends. A modified multifractal analysis method, i.e EMD-MFDFA (multifractal detrended fluctuation analysis), was proposed and the performance of the EMD-MFDFA method was proved by simulation. Then taking the multifractal feature vectors of motor bearing vibration signals extracted by the EMD-MFDFA method as the training sets, an intelligent diagnosis model was established to diagnose motor bearing early faults according to the support vector machine (SVM) theory, in which the parameters of SVM were optimized by using the hybrid genetic algorithm. Finally, the different states of the motor bearing vibration signals were analyzed. The results show that the EMD-MFDFA method can reveal the multifractal characteristics of the motor bearing signals and identify the bearing normal state, a single fault state and multiple faults state accurately. And the established model is able to diagnose bearing different fault effectively and suitable for on-line monitoring for motor bearing faults.

相关论文

  • 暂无!

相关知识点

  • 暂无!

有色金属在线官网  |   会议  |   在线投稿  |   购买纸书  |   科技图书馆

中南大学出版社 技术支持 版权声明   电话:0731-88830515 88830516   传真:0731-88710482   Email:administrator@cnnmol.com

互联网出版许可证:(署)网出证(京)字第342号   京ICP备17050991号-6      京公网安备11010802042557号