摘要
对中红外光谱测定发动机油运动黏度的方法进行了研究。收集了81个发动机油样,并随机抽取其中的61个油样作为建模集,剩余的20个油样作为预测集。筛选出发动机油运动黏度的特征波段1758.8 cm^(-1)~1305.6 cm^(-1)。用建模集建立了特征波段的主成分分析-偏最小二乘法数学模型。该模型预测集的相关系数和预测误差均方根分别为0.97279和0.43433,满足中红外光谱测定发动机油运动黏度的要求。特征波段的主成分分析-偏最小二乘法数学模型可用于测定发动机油的运动黏度。
The determination method of kinematic viscosity of engine oils by mid-infrared spectroscopy was studied.81 engine oils samples were collected and in it 61 engine oils samples were randomly selected as the modeling set.The remaining 20 engine oils samples were used as the prediction set.The selected characteristic bands of the kinematic viscosity of engine oils were 1758.8 cm-1~1305.6 cm-1.The principal component analysis- partial least square mathematical model was established by modeling set.Correlation coefficient and prediction error mean square root of the model prediction set were 0.97279 和 0.43433 respectively.This could satisfy the requirement of kinematic viscosity of engine oils determined by mid-infrared spectroscopy.The principal component analysis-partial least square mathematical model could be used for the determination of kinematic viscosity of engine oils.
出处
《合成润滑材料》
CAS
2017年第2期12-15,共4页
Synthetic Lubricants
基金
国家质检总局科技计划项目。编号2015IK192
关键词
中红外光谱
发动机油
运动黏度
特征波段
主成分分析
偏最小二乘法
数学模型
mid-infrared spectroscopy
engine oils
kinematic viscosity
characteristic wave band
principalcomponent analysis
partial least square
mathematical model