摘要
针对柴油机工作环境恶劣,故障信号特征难以提取且诊断效率较低,引入最大相关峭度解卷积(MCKD)降噪、小波包和希尔伯特变换相结合的能量熵特征提取,最后用灰狼优化的支持向量机(GWO-SVM)分类的方法。首先利用排列熵、包络谱稀疏度选取最优MCKD的滤波器长度L与冲击周期T;然后通过小波包分解和希尔伯特变换分别提取各节点能量熵和IMF分量的Hilbert边际谱能量熵构成高维特征向量,最后用GWO-SVM进行诊断。通过水泵不同状态下的试验数据验证了该方法的有效性。
In view of the bad working environment of diesel engine,it is difficult to extract fault signal features and low diagnosis efficiency,The MCKD de-noising,wavelet packet and Hilbert transform combined energy entropy feature extraction are introduced.Finally,gray wolf optimized support vector machine classification method is used.Firstly,the filter length L and the impact period T of the optimal MCKD are selected by using the permutation entropy and the sparsity of the envelope spectrum.Then,the energy entropy of each node and the Hilbert marginal spectral energy entropy of IMF component are extracted by wavelet packet decomposition and Hilbert transform to form a high-dimensional feature vector.Finally,GWO-SVM is used for diagnosis.The effectiveness of the proposed method is verified by the experimental data of different pump states.
作者
许昕
李磊磊
潘宏侠
景雪瑞
Xu Xin;Li Leilei;Pan Hongxia;Jing Xuerui(School of Mechanical Engineering,North University of China,Taiyuan 030051,China)
出处
《国外电子测量技术》
北大核心
2022年第2期132-137,共6页
Foreign Electronic Measurement Technology
基金
内燃机可靠性国家重点实验室基金(skler-201911)项目资助
关键词
MCKD
小波包分解
HHT变换
能量熵
故障诊断
MCKD
wavelet packet decomposition
HHT transform
energy entropy
fault diagnosis
作者简介
许昕,博士,讲师,主要研究方向为过程装备运行状态监测与故障诊断。E-mail:1097518571@qq.com
