摘要
针对机械故障信号的非线性、故障征兆的多样性和复杂性等诊断问题,提出了一种基于小波包样本熵和流形学习的故障特征提取模型。该模型首先利用小波包的分解和重构,计算重构细节信号的样本熵,初步提取滚动轴承故障特征,然后利用流形学习法对初步的样本熵故障特征进行进一步的提取,在保留故障特征的整体几何结构信息的同时降低了特征数据的复杂度,增强了故障模式识别的分类性能。最后通过支持向量机对该模型提取的特征进行分类,通过比较初提取特征和再提取特征分类效果来验证该模型的优越性。
Aiming at nonlinearities of mechanical fault signals,and diversities and complexities of fault symptoms,a model of fault feature extraction was proposed based on wavelet packet sample entropy and manifold learning.Firstly,to extract initial rolling bearing fault features,the sample entropy of the reconstructed signal was calculated with the model by using the wavelet packet decomposition and reconstruction method.Then the local tangent space alignment （LTSA） method for the further extraction of the fault features was applied.The complexity of the feature data was reduced,meanwhile the structural information of the total geometry of the fault features was reserved.Moreover,with the proposed model,the classification performance of the entire fault mode identification was enhanced.Finally,a support vector machine （SVM） was used to classify the fault features extracted with the proposed model.The initial feature extraction and further feature extraction of the classification results were compared to validate the effectiveness of the proposed model.
出处
《振动与冲击》
EI
CSCD
北大核心
2014年第11期1-5,共5页
Journal of Vibration and Shock
基金
广东省教育厅育苗工程(自然科学)(2013LYM-0052)
广东省高等学校优秀青年教师培养计划(Yq2013110)
关键词
小波包
样本熵
流形学习
特征提取
支持向量机
wavelet packet
sample entropy
manifold learning
feature extraction
SVM
