中红外光谱在润滑油分类识别中的应用

Application of Mid-infrared Spectrum in Classification of Lubricant

由中红外光谱能够获得不同种类润滑油的组成与结构信息。分别用聚类分析法和主成分分析法提取润滑油的中红外光谱信息,对齿轮油、内燃机油和液压油这3大类共113个润滑油样品进行了分类识别。采用系统聚类法对其进行分类,比较了不同类间距离、样品间距离和光谱区间对聚类分析结果的影响。采用主成分分析法对齿轮油、内燃机油、液压油及不同黏度等级的通用车用油（内燃机油）与齿轮油进行了分类,比较了数据处理方法与光谱区间对分类结果的影响。结果表明：合理选取中红外光谱区间和数据处理方法,则润滑油的分类可以取得令人满意的效果;中红外光谱技术可以作为润滑油种类识别的新方法。

The composition and structure of different types of lubricants is gained by mid-infrared spectrum（MIR）.113 oil samples containing gear lubricant,engine lubricant and hydraulic lubricant are identified by extracting the MIR information of lubricant with system cluster analysis and principle component analysis（PCA）.All these samples are classified by system cluster method,and the influence of the distance between clusters,samples and spectral ranges on the classification result is compared.Gear oil,engine oil,hydraulic oil,and different viscosity grades of general purpose vehicles oil（also a kind of engine oil） and gear oil of different viscosity grades are classified by PCA,respectively.The influence of both data processing methods and different spectral range on the result is compared,too.The results indicate,with information of infrared spectrum and data preprocessing method being selected,that the classification is satisfactory and MIR can be used as a new technique to identity the type of lubricant.