一种用于故障诊断的多信息熵监测方法研究  

Study on a Multi-entropy Monitoring Method Used in Fault Diagnosis

作　　者：谢平[1] 林洪彬[1] 王霄[1] 刘彬[1] 

机构地区：[1]燕山大学电院,秦皇岛066004

出　　处：《仪器仪表学报》2004年第z3期541-543,共3页Chinese Journal of Scientific Instrument

基　　金：河北省自然科学基金资助项目(603334)。

摘　　要：采用信息熵计算模型,在对信号进行多分辨率分析和多重分形分析的基础上,研究了对信号能量进行定量评价的新方法—多分辨率奇异谱熵和多重分形熵,该指标能够分别表征不同尺度上信号能量分布的奇异状况和几何特征分布几率,可用于非平稳信号内在特征信息的精细刻画和准确提取,并作为设备故障诊断和状态监测的有效途径。通过对不同信号的模拟分析,提出的信息熵模型可作为设备状态评价的综合特征指标。The information entropy model is used and two new entropies—multi resolution singular spectrum entropy and multi fractal dimension entropy which can evaluate the energy distribution of the signal are proposed based on the multi resolution analysis and multi fractal analysis. Because the feature entropy values can describe the singular condition and geometry distribution odds of the signal energy in different scales, so it can be used to extract the internal feature information of non stational signals and act as an efficient method in fault diagnosis and state monitoring of the machines. Finally, the new method is simulated in analysis of several engineering signals and it is proved that this information entropy model could be as effective synthesis indexes for the recognition of system running condition.

关 键 词：信息熵 多分辨率分析 多重分形 故障诊断 

分 类 号：TH7-55[机械工程—仪器科学与技术]