| With the continuous development of society, the human demand for energy isgrowing today. Oil is one of the primary energy demanded by human society. Due to thelarge number of uses of petroleum material, petroleum material waste and derivatives onthe environment pollution is more and more serious. Especially the petroleum substance inthe aquatic pollution has started to cause the attention of people. Therefore, to detect waterquality has become the primary problem of water pollution regulation and processing.In the background of the detection of trace mineral oil in water，the research contentof this paper is to test the components in mixture of diesel and gasoline, the purpose ofwhich is to provide experimental and theoretical basis for the water oil testing experience.Based on diesel and gasoline mixture as samples, this paper uses Fourier transformnear infrared spectroscopy technology on experimental samples collection. By using thePLS model regression algorithm and iterative Bagging PLS model regression algorithmwe can analyze the near infrared spectra of the samples. By steps of preparing the testsamples, processing spectrum of the samples, selecting the calibration set and test set,establishing regression model, and analyzing prediction results, we can carry on thequantitative analysis for diesel and gasoline components. Established regression model isevaluated by the parameters such as the Root Mean Square Error of Correction (RMSEC),the Root Mean Square Error Of Cross Validation (RMSECV), the Root Mean SquaredError of Prediction (RMESEP) and correlation coefficient (R). The model is evaluatedfrom multiple spectral absorption bands. And through the analysis system development ofthe content of diesel fuel and gasoline we can analyze the content of diesel fuel andgasoline under interface.In this paper, the research shows that iteration Bagging PLS can well analyze andforecast the near infrared spectrum data and get a better regression model and forecastingability. |
PDF Full Text Request