基于小波包分解与K-L变换的齿轮泵振动信号故障特征提取方法  

Fault Feature Extraction Method of Gear Pumps Vibration Signal Based on Wavelet Packet Decomposition and K-L Transform

作　　者：蔡伟[1] 黄坤阳 戴民强[1] 杨志勇[1] 

机构地区：[1]第二炮兵工程大学兵器发射理论与技术国家重点学科实验室,陕西西安710025

出　　处：《机床与液压》2016年第21期159-163,共5页Machine Tool & Hydraulics

基　　金：国家自然科学基金资助项目(61102170)

摘　　要：针对齿轮泵故障成因复杂、模糊性强的特点,结合小波包分解与K-L变换,提出一种适用于支持向量机故障诊断的特征提取方法。通过小波包对样本故障振动信号进行分解得到特征向量,而后利用K-L变换处理得到新的特征向量集,达到降维去噪的目的。将处理后的特征向量集用于支持向量机的模型训练,分析结果表明:该方法能够有效提高故障模式识别准确率和识别效率。According to the characteristics of the gear pumps fault of complicated formation, strong fuzziness, combined with wavelet packet decomposition and K-L transform, presented a suitable support vector machine fault diagnosis feature extraction method The feature vector was decomposed by wavelet packet, and the new feature vector set was processed by K-L transform. The feature vector set was used to support vector machine model training. The result shows that the method can effectively improve the accuracy and efficiency of fault pattern recognition.

关 键 词：故障诊断 小波包分解 K-L变换 特征提取 

分 类 号：TH137[机械工程—机械制造及自动化]