摘要
通过对润滑油红外光谱区间选择、平滑、求导等预处理和识别方法的选择,建立润滑油红外光谱专家识别系统,并介绍该系统的功能和使用方法。红外光谱预处理可以有效消除仪器等因素对光谱质量的影响;选择指纹区为润滑油的识别区间,避开了润滑油基础油的红外谱峰的影响,提高了识别准确度;普鲁克识别方法可以有效消除测样方式等对光谱质量的影响,进行润滑油指纹区的轨迹识别比通常的特征峰检索方法能更有效地识别润滑油红外谱图。利用所建的润滑油红外光谱专家识别系统对不同油品进行识别验证,结果表明所建的专家系统使用方便,能有效识别润滑油。
The expert recognized system of lubricating oil infrared spectrum was built through the choosing of recognized methods, and the pre-processing of original infrared spectrum data, including the choosing of wave number bound, the process of smoothing and deriving. The function and using methods of the system were introduced. The effect of the instru- ment on infrared spectrum quality can be effectively eliminated by pretreatment. The effect of the peak of base oil can be e- liminated by choosing the fingerprint region as the recognized bound of lubricating oil, and the recognized accuracy is im- proved. Procrustes analysis method has a better recognition for lubricating oil infrared spectrum than the characteristic peak index method, which can reduce the effect on infrared spectrum quality resulting from experimental conditions. Various oil samples were used to validate the recognized system. The results show that the syslem is convenient and has high recogniz- ing effect on lubricating oil.
出处
《润滑与密封》
CAS
CSCD
北大核心
2012年第10期95-98,共4页
Lubrication Engineering
关键词
润滑油
红外光谱
专家识别系统
lubricating oil
infrared spectrum
expert recognized system
