基于EMMD和AR奇异值熵的故障特征提取方法研究

提出了一种基于EMMD(extremum field mean mode decomposition)和AR(auto-regressive)奇异值熵的故障特征提取方法。该方法在对故障信号的EMMD分解基础上,选取有限个固有模态函数(IMF,intrinsic mode function)的AR模型参数向量作为故障的初始特征向量矩阵,对初始特征向量矩阵求取奇异值熵,通过奇异值熵的大小表征故障类型。对转子故障数据的分析结果表明该方法能够有效地应用于非线性和非平稳故障信号的特征提取。

Research on Fault Feature Extraction Approach Based on EMMD and Singular Value Entropy of AR Model

A fault feature extraction approach based on extremum field mean mode decomposition(EMMD) and singular value entropy of auto-regressive(AR) model is proposed.The fault signal is decomposed with EMMD into a number of intrinsic mode functions(IMF),and the AR model of each IMF component is established.Then the AR parameters and the variance of remnant are regarded as the initial feature vectors,from which the initial feature vector matrix is formed.By applying the singular value decomposition to the initial feature vector matrix,the singular value entropy is obtained,which can discriminate the fault type.The experimental results of rotor system fault show that the proposed method can be applied to fault feature extraction for nonlinear and nonstationary fault signals efficiently.