采用符号序列Shannon熵的机器信息提取方法  

Extracting Machine Information by Shannon Entropy of Symbolic Series

作　　者：赵建华[1] 张雨[1] 

机构地区：[1]南京工程学院车辆工程系,南京211167

出　　处：《武汉理工大学学报（交通科学与工程版）》2009年第2期321-324,共4页Journal of Wuhan University of Technology（Transportation Science & Engineering）

基　　金：国家自然科学基金项目资助(批准号:50176006)

摘　　要：讨论了将时间序列转化为符号序列,即信号粗粒化问题,给出二进制符号化和角区间符号化2种规则,得到了二进制符号化规则下的符号序列Shannon熵计算方法.对于汽油机瞬态排放信号、柴油机机身振动信号、转子轴振动信号,采用Shannon熵提取了信号特征,结果表明描述符号序列总体特征的Shannon熵能够便捷地反映机器的运行状况.The problem of how to translate a time series into symbolic series, which means signal coarsegrained, is discussed. The symbolizing regulations of binary and angle section are given. The method of calculating symbolic series Shannon entropy under binary symbolizing regulation is also introduced. The signal characteristics, which are in the gasoline-engine transient emission, diesel-engine block vibration and rotor shaft vibration, are extracted base on the Shannon entropy. The results indicate that the Shannon entropy, which describes the collectivity characteristics of symbolic series, could reflect the machine＇s operation condition conveniently.

关 键 词：符号序列分析 Shannon熵 机器 特征提取 

分 类 号：U464.1[机械工程—车辆工程] O212[交通运输工程—载运工具运用工程]