作者: Hongkun Chen , Yujia Zhang , Hanbing Qi , Dong Li
DOI: 10.1016/J.IJLEO.2019.05.067
关键词: Materials science 、 Near-infrared spectroscopy 、 Analytical chemistry 、 Ultraviolet 、 Regression analysis 、 Normalization (statistics) 、 Spectroscopy 、 Diesel fuel 、 Ethanol 、 Ultraviolet visible spectroscopy
摘要: Abstract Ultraviolet spectroscopy (UV) and near-infrared (NIR) combined with Partial Least Square (PLS) was used as a destructive technique to detect the ethanol content in diesel mixture. The spectral data were obtained by ultraviolet (TU-1901/1900) near infrared (IRTracer-100). Savitaky-Golay (S-G) method Max-min normalization pretreat normalize data, respectively. Then regression models established pretreated of UV, NIR UV-NIR PLS, evaluation parameters including rc, rp, RMSECV RMSEP compare results three models. show that fusion spectrum are rp=1, RMSECP=0.00156, rc=0.8431 RMSEV=0.0148, Compared UV spectrum, RMSECP is reduced 0.00169 0.01224, rc increased 0.0902 0.1813, These indicated model has more accuracy than single PLS can content.