基于2 维非负矩阵分解的时频图像压缩在柴油机故障诊断中的应用

针对1 维非负矩阵分解技术对2 维矩阵特征降维时，会产生数据量巨大、计算效率低下和丢失原始数据结构信息的问题，引入2 维非负矩阵分解技术。通过S 变换得到振动信号的时频图像，用1DNMF 和2DNMF 分别压缩时频图像，对压缩后的图像信息进行分类，对柴油机在8 种状态下的振动信号进行采集，并采用最近邻分类器、朴素贝叶斯分类器和支持向量机分类器进行实验对比。结果表明，2 维非负矩阵分解技术比原始的1 维技术计算效率更高，故障诊断更精准。

Application of Time-frequency Image Compression Based on2D Non-negative Matrix Factorization in Engine Fault Diagnosis

For the 1D non-negative matrix factorization (1DNMF) technique, when the dimension of the 2D matrix isreduced, the problem of huge data volume, low computational efficiency and loss of original data structure information isgenerated. 2D non-negative matrix factorization (2DNMF) technique is introduced. The time-frequency image of thevibration signal is obtained by S-transformation, and the time-frequency image is compressed by 1DNMF and 2DNMFrespectively, and the compressed image information is classified, and the vibration signals of the diesel engine in 8 statesare collected, and the nearest neighbor classifier is adopted. The naive Bayes classifier and the support vector machineclassifier are used for experimental comparison. The results show that the 2DNMF is more efficient and accurate in faultdiagnosis than the original 1DNMF.