摘要
针对往复压缩机振动信号的非平稳、非线性和特征耦合特性,提出了基于ITD与排列熵的往复压缩机轴承故障特征提取方法。利用ITD方法将各状态振动信号分解为一系列PR分量,依据相关性系数选择代表故障状态主要信息的PR分量,计算其排列熵形成有效的特征向量。以平均样本距离为特征向量可分性标准,对比了ITD与近似熵方法所提取特征向量,结果表明此法具有更好的可分性。
Based on non-stationary,nonlinearity and feature coupling characteristics of reciprocating compressor vibration signal,a feature extraction method of reciprocating compressor fault based on ITD and permutation entropy is proposed.Vibration signals in each state are decomposed into a series of PR components with ITD method, and the PR components which contain the main information of fault state are chosen according to relative coefficient.Permutation entropy of the selected PR components was calculated as eigenvectors.Average sample distance is used as severability standard of eigenvector,and superiority of this method is proved by comparing the average sample distance of eigenvectors extracted by ITD and approximate entropy method.
出处
《压缩机技术》
2014年第5期17-20,共4页
Compressor Technology
基金
国家科技支撑计划项目(2012BAH28F00)
黑龙江省教育厅科学技术研究重点项目(12521Z007)
黑龙江省自然基金项目(E201335)
东北石油大学青年科学基金项目(ky120221)
关键词
ITD
排列熵
往复压缩机
特征提取
故障诊断
ITD
permutation entropy
reciprocating compressor
feature extraction
fault diagnosis
