基于多重分形去趋势波动的机械故障诊断新方法  

A New Method of Machinery Fault Diagnosis Based on Multifractal Detrended Fluctuation Analysis

作　　者：王书涛[1] 李梅梅[1] 张淑清[1] 张金敏[1] 赵玉春[1] 

机构地区：[1]燕山大学电气工程学院河北省测试计量技术及仪器重点实验室,河北秦皇岛066004

出　　处：《计量学报》2012年第3期232-235,共4页Acta Metrologica Sinica

基　　金：基金项目：国家自然科学基金（51075349,61077071）;河北省自然科学基金（F2011203207）

摘　　要：针对旋转机械系统故障信号的非平稳性、非线性等复杂特征，给出一种基于多重分形去趋势波动分析的机械故障诊断方法。该方法首先对原始信号进行去趋势处理，再结合多重分形理论提取多重分形谱面积和多重分形熵两个分形参数，并将其作为故障诊断的新判据，最后通过实验结果证明了方法的有效性，从而为机械故障诊断提供了一种新的方法。Aiming at the complex features of the fault signal of rotating machinery such as nonstationary and nonlinearity, a new method of machinery fault diagnosis based on multifractal detrended fluctuation analysis （ MF-DFA ） was introduced. The original signal is proceed by the MF-DFA , then the fractal parameters （the area of multifractal spectrum and muhifractal entropy） were extracted as new criterions to diagnose machinery faults. Results of experimental analysis indicate that the method is effective and it provides a new way in fault diagnosis machinery of rotating machinery.

关 键 词：计量学 故障诊断 多重分形去趋势波动分析 多重分形谱面积 多重分形熵 

分 类 号：TB936[一般工业技术—计量学]