CAJ | 학술논문

对红外光谱技术应用于航空合成润滑油的快速判别进行了研究。采用红外光谱衰减全反射附件采集5种不同牌号的航空合成润滑油的红外光谱。基于偏最小二乘判别分析方法，建立了5种润滑油牌号快速判别模型。讨论了建模主成分数与不同建模光谱区间对模型预测能力的影响。通过优化建模参数，提高了模型的预测精度。使用建立的分析模型对预测集样品进行预测，判别正确率为100％。
대홍외광보기술응용우항공합성윤활유적쾌속판별진행료연구。채용홍외광보쇠감전반사부건채집5충불동패호적항공합성윤활유적홍외광보。기우편최소이승판별분석방법，건립료5충윤활유패호쾌속판별모형。토론료건모주성분수여불동건모광보구간대모형예측능력적영향。통과우화건모삼수，제고료모형적예측정도。사용건립적분석모형대예측집양품진행예측，판별정학솔위100％。
The identification of synthesized lubricant for the aero-engine was studied by the means of Fourier transform infrared spectrum technique. Five different grades of lubricants were analyzed by the attenuated total reflection sampler of Fourier transform infrared spectrum. The quick grade identification model was built with the method of partial least squares discriminant analysis. The influence of the principle components number and the different spectra range to the prediction was discussed. The prediction precision of the model was improved by optimizing the parameters. The model was used to predict the sample. The results show the model has 100% accuracy to predict the samples.