Data Fusion Of Raman And Near-Infrared Spectroscopies For The Rapid Quantitative Analysis Of Methanol Content In Methanol?Gasoline

Methanol-gasoline is a new type of fuel to replace traditional gasoline,for which methanol content has a significant impact.Therefore,rapid analysis of methanol content in methanol-gasoline is of great significance to monitor the quality of methanol-gasoline.In this work,two different data-fusion strategies based on Raman and NIR spectroscopies coupled with chemometrics method were constructed and applied for the rapid and accurate analysis of methanol content in methanol-gasoline.The paper is divided into four chapters,the main research contents are as follows:?1?A rapid quantitative analysis method of methanol content in methanol gasoline based on Raman spectroscopy and PLS was established.The influence of five spectral preprocessing methods?Normalization,D1st,MSC,SNV and WT?on the predictive performance of PLS calibration model based on Raman spectra was investigated.5-flod CV was used to optimize the input variables,latent variables and variable importance threshold of the model.The results show that the PLS model has good predictive performance,with the R²_CV and R_P² of 0.9737 and0.9604,respectively,and the RMSECV and RMSEP of 0.0311 and 0.0341%,respectively.This method has many advantages,such as fast detection speed,accurate analysis results and so on,from which a new method can be obtained to detect the methanol content in methanol-gasoline quickly and accurately.?2?A rapid quantitative analysis method of methanol content in methanol gasoline based on NIR spectroscopy and PLS was established.The influence of five spectral preprocessing methods?Normalization,D1st,MSC,SNV and WT?on the predictive performance of PLS calibration model based on NIR spectra was investigated.5-flod CV was used to optimize the input variables,latent variables and variable importance threshold of the model.The results show that the PLS model has good predictive performance,with the R²_CV and R_P² of 0.9847 and0.9846,respectively,and the RMSECV and RMSEP of 0.0237 and 0.0174%,respectively.With the advantages of fast detection speed,accurate analysis results and so on,this method can provide new ideas and methods for the rapid and accurate quantitative analysis of methanol content in methanol-gasoline.?3?A rapid quantitative analysis method of methanol content in methanol gasoline based on Raman-NIR?low-level and mid-level fusion data?fusion spectral data coupled with PLS was established.The influence of four spectral preprocessing methods?D1st,MSC,SNV and WT?on the predictive performance of PLS calibration model based on Raman-NIR fusion spectra was investigated.5-flod CV was used to optimize the input variables,latent variables and variable importance threshold of the model.Normalization was used for low-level data fusion,and VIP was used for mid-level data fusion.And then,the predictive performance of PLS calibration model based on different input variables was compared,and the results show that the PLS calibration model based on mid-level fusion data has the best predictive performance,with the R²_CV and R_P² of 0.9988 and 0.9905,respectively,and the RMSECV and RMSEP of 0.0068 and0.0288%,respectively.Compared with the single spectral analysis method based on Raman/NIR,this method can provide a new idea and method with more accurate analysis results for the rapid and accurate quantitative analysis of methanol content in methanol-gasoline.