毋文峰,陈小虎,苏勋家,江克侠
WU Wen-feng1,CHEN Xiao-hu1,SU Xun-jia1,JIANG Ke-xia2(1.The Second Artillery Engineering College

• 1:第二炮兵工程学院
• 2:武警工程学院

摘要(Abstract)：

齿轮泵动态压力信号蕴含了丰富的状态信息,是齿轮泵的关键性能参数之一;负熵是随机变量独立性的自然测度,反映了机械信号信息的动态变化特征。为了进行基于齿轮泵动态压力信号的故障诊断,在引入经验模态分解技术的基础上,提出了齿轮泵压力信号的负熵基ICA特征提取方法,并进而联合最小二乘支持向量机,探讨了基于ICA和LS-SVM的齿轮泵特征提取和故障诊断方法;理论和试验研究表明,基于负熵的ICA特征提取和故障诊断方法是有效的。
Gear pumps' dynamic pressure signals contain abundant condition information.They are key parameters of gear pumps.Negentropy is a natural measurement for independent property of random variables.It denotes dynamic information features of mechanical signals.To study gear pumps' fault diagnosis based on their dynamic pressure signals,ICA feature abstraction algorithm based on negentropy and EMD was proposed.It combined excellent characteristics of negentropy and empirical mode decomposition.After negentropy ICA features abstracted,least squares support vector machines was applied to recognize different fault patterns.Their applications in gear pumps indicate that ICA feature abstraction and fault diagnosis algorithm based on negentropy is effective.

关键词(KeyWords)： 压力信号;负熵;经验模态分解;最小二乘支持向量机;特征提取;故障诊断
pressure signal;negentropy;empirical mode decomposition;least squares support vector machine;feature abstraction;fault diagnosis

Abstract:

Keywords:

基金项目(Foundation): 总装备部预研重点基金项目

作者(Author): 毋文峰,陈小虎,苏勋家,江克侠
WU Wen-feng1,CHEN Xiao-hu1,SU Xun-jia1,JIANG Ke-xia2(1.The Second Artillery Engineering College

参考文献(References)：

• [1]吉林工业大学.工程机械液压与液力传动[M].机械工业出版社,1988.
• [2]胥永刚,李强,王正英,等.基于独立分量分析的机械故障信息提取[J].天津大学学报,2006,39(9):1066-1071.
• [3]郝志华,张一杨,刘岩.独立成分分析在转子故障诊断中的应用[J].汽轮机技术,2007,49(2):137-139.
• [4]陈长征,程锦生,韩丽娅,等.基于盲源分离的齿轮箱状态检测与故障诊断[J].沈阳工业大学学报,2008,30(4):444-448.
• [5]焦卫东.基于独立分量分析的旋转机械故障诊断方法研究[D].浙江:浙江大学,2003.
• [6]Comon P Independent component analysis,Anewcon-cept[J].Signal Processing,1994,36(3):287-314.
• [7]Aapo Hyvarinen,Juha Karhunen,Erkki Oja.Inde-pendent component analysis[M].John Wiley&SonsInc.,2001.
• [8]杨福生,洪波.独立分量分析的原理与应用[M].北京:清华大学出版社,2006.
• [9]史习智.盲信号处理—理论和实践[M].上海:上海交通大学出版社,2008.
• [10]邓乃杨,田英杰.支持向量机—理论、算法与拓展[M].北京:科学出版社,2009.
• [11]Suykens J AK,Vandewalle J.Least Squares SupportVector Machine Classifiers[J].Neural ProcessingLetters(S1370-4621),1999,9(3):293-300.
• [12]于德介,程军圣,杨宇.机械故障诊断的Hilbert-Huang变换方法[M].北京:科学出版社,2006.
• [13]李志农,吕亚平,范涛,等.基于经验模态分解的机械故障欠定盲源分离方法[J].航空动力学报,2009,24(8):1886-1892.
• [14]陈建国,张志新,郭正刚,等.独立分量分析方法在经验模式分解中的应用[J].振动与冲击,2009,28(1):109-111.
• [15]Jang Gil Jin,Te Won Lee.Single-channel signal sepa-ration using time-domain basis function[J].IEEE sig-nal processing letters,2003,10(6):168-171.