液压泵故障诊断稀疏编码方法研究

发布时间：2018-06-12 18:28

本文选题：液压泵 + 奇异值分解　；参考：《北京理工大学学报》2017年05期

【摘要】：针对液压泵故障特征难以提取、诊断过程复杂、自动化程度低等特点,将稀疏编码方法应用于液压泵故障诊断.通过对液压泵泵壳处振动信号进行时频域变换,将变换后的信号作为样本,采用K-SVD算法对训练样本进行字典学习以获取字典,利用正交匹配追踪算法对测试信号进行分解与重构,通过不同类别字典对测试信号的重构率大小进行故障种类识别,实现液压泵故障分类.通过试验验证并与BP神经网络、支持向量机对比,结果表明稀疏编码方法具有对故障识别速度快、准确率高、稳定性好等优点,可以有效地实现对液压泵故障的诊断.
[Abstract]:In view of the characteristics of hydraulic pump fault features such as difficult to extract, complex diagnosis process and low automation, the sparse coding method is applied to hydraulic pump fault diagnosis. Through the time-frequency domain transformation of the vibration signal at the pump shell of the hydraulic pump, the transformed signal is taken as the sample, and the K-SVD algorithm is used to study the dictionary of the training sample to obtain the dictionary. The orthogonal matching tracking algorithm is used to decompose and reconstruct the test signal, and the fault classification of hydraulic pump is realized by identifying the fault types of the reconstruction rate of the test signal by different kinds of dictionaries. Compared with BP neural network and support vector machine, the results show that the sparse coding method has the advantages of fast fault identification, high accuracy and good stability, and can effectively realize the fault diagnosis of hydraulic pump.
【作者单位】：第四军医大学;解放军理工大学;
【基金】：国家自然科学基金项目资助(51175511)
【分类号】：TH137.51

【相似文献】