基于激光诱导荧光光谱原理, 提出一种常见机油的快速识别方法。利用激光器发射波长为355 nm的紫外激光, 诱导九种常见机油样品发射荧光, 共采集450组荧光光谱数据, 其中360组数据用于分类训练, 90组数据用于识别。分析发现不同机油的荧光光谱特征有明显差异, 利用主成分分析结合聚类分析法实现了对90组待识别光谱数据的快速识别, 识别率可达97.8%。实验证明, 激光诱导荧光光谱结合多元分析可以实现不同机油的快速识别与检测。

A rapid recognition method of common engine oils, based on the principle of laser induced fluorescence (LIF), is proposed in the present paper. A 355 nm ultraviolet laser is used to induce fluorescence emission of 9 kinds of common machine oil samples. In total 450 groups of fluorescence spectral data are collected, of which 360 groups of data are used for classification training and 90 sets of data for identification. It was found that the fluorescence spectra of engine oils are distinct from each other visibly. The rapid identification of 90 groups of data is realized by using clustering analysis combined with principal component analysis. The recognition rate could reach up to 97.8%. Experiment demonstrated that the fast identification of diverse engine oils could be realized by using LIF combined with multivariate analysis method.