摘要
采用超谱M型原子发射光谱仪对配制的CD40、CF40柴油机油、液压油油样进行分析。以该数据为样本,使用系统聚类法(类平均法)和主成分分析法分别进行元素聚类分析和样本聚类分析。结果表明,系统聚类法更能有效地进行元素聚类,区分出主要磨损元素、主要添加剂元素以及微量元素和干扰元素;而在分析主要添加剂元素时,主成分分析法能更合理地对油样的类别进行聚类。应用系统聚类法和主成分分析法对型船主机的油样数据进行分析,结果表明,该方法可准确推断出该主机换油时间,判断所使用润滑油油品类型。
The oil samples made by lubricating oil of CIM0, CF40 diesel engine oil and hydraulic oil were analyzed by atomic emission Spectrometric M Instrument. The element clustering spectral data as samples were carried out respectively by system cl analysis （PCA）. The results show that the system clustering analysis the main wear elements, main additional elements, microelements analysis and sample clustering analysis of the obtained ustering analysis （average） and principle component can more effectively cluster elements, and distinguish and interferential elements. When analyzing the main additional elements, the PCA can carry out the reasonable oil sample clustering. The spectral data of main engine oil sam- ples of some ships were analyzed by system clustering analysis and PCA method. It is shown that this method can accurately deduce oil-renewal time and estimate lubricating oil types in use.
出处
《润滑与密封》
CAS
CSCD
北大核心
2013年第5期99-103,共5页
Lubrication Engineering
基金
湖北省自然科学基金项目(2010CDB01505)
关键词
油液光谱分析
系统聚类
主成分分析
润滑油
spectrometric oil analysis
system clustering
principle component analysis
lubricating oil
