摘要
导热油在设备热传递领域应用广泛,但目前市场上导热油的品质良莠不齐,优质导热油掺杂劣质导热油会导致一系列安全问题。常用的导热油鉴定手段耗时长,检测复杂。本实验借助近红外光谱技术结合偏最小二乘回归法(Partial Least Squares Regression,PLSR)用于鉴定导热油品质。结果表明:鉴定导热油品质的模型,经标准正态变量(standard normal variate,SNV)预处理后的定量分析模型的拟合程度最佳,预测相关系数(regression coefficient of prediction,Rp)达到0.9999,预测均方根误差(the root mean square error of prediction,RMSEP)小于0.02。因此,近红外光谱技术结合PLSR可为导热油的品质鉴定提供一种快速、便捷的方法。
Heat conducting oil （HCO） is widely used in different areas as the device heat transfer, but HCO on the market are mixed, adulteration of HCO can cause a range of security issues. The common methods of HCO identifica- tion are time-consuming and complicated. Near infrared spectroscopy combined with partial least squares regression （PLSR） were used in quantitatively analysing the adulteration of HCO. The results showed that the best model of quantitatively analysis the adulteration of HCO was achieved by the processing of standard normal variate （SNV） , the coefficient rate of prediction （Rp） reached 0.9999 and the root mean square error of prediction （RMSEP） was less than 0.02. Therefore, near infrared spectroscopy combined PLSR provided a fast and convenient way for the char- acterization of HCO.
出处
《福建分析测试》
CAS
2016年第5期13-17,共5页
Fujian Analysis & Testing
关键词
导热油
偏最小二乘回归法
近红外光谱
heat conducting oil
Partial Least Squares Regression
Near Infrared Spectrum
