Detection of ethanol content in ethanol diesel based on PLS and multispectral method

Abstract

Ultraviolet spectroscopy (UV) and near-infrared (NIR) spectroscopy combined with Partial Least Square (PLS) was used as a destructive technique to detect the ethanol content in ethanol diesel mixture. The spectral data were obtained by ultraviolet (TU-1901/1900) and near infrared spectroscopy (IRTracer-100). The Savitaky-Golay (S-G) method and Max-min normalization method were used to pretreat and normalize the spectral data, respectively. Then the regression models were established by the pretreated spectral data of UV, NIR and UV-NIR combined with PLS, and the evaluation parameters including rc, rp, RMSECV and RMSEP were used to compare the results of three regression models. The results show that the evaluation parameters of the UV-NIR fusion spectrum are rp=1, RMSECP=0.00156, rc=0.8431 and RMSEV=0.0148, respectively. Compared with the evaluation parameters of the UV and the NIR spectrum, the RMSECP is reduced by 0.00169 and 0.01224, and the rc is increased by 0.0902 and 0.1813, respectively. These results indicated that the regression model of the UV-NIR fusion spectrum has more accuracy than that of the single spectrum, and the regression model established by UV-NIR fusion spectrum combined with PLS can detect the ethanol content.