Detection Method And Research For The Quality Indexs Of Edible Vegetable Oil Based On Raman Spectroscopy

Edible vegetable oil is one of necessities in people’s daily life. Its quality was closely related to people’s health. There were many quality indexes to measure the quality of edible vegetable oil. Traditional methods exist the disadvantage of determination process cumbersome, low detecting and accuracy rate. Because of many different kinds and complex compositions of edible vegetable oil, we need to continue to explore edible vegetable oil in the experiment and production process in order to meet the needs of easily with lower consumption, rapidly nondestructive testing with accurate real-time quality detection of edible vegetable oil. With its simple operation, fast nondestructive detection, high sensitivity and accuracy rate, using Raman spectroscopy detection had been widely applied to explore edible vegetable oil quality indexes. Selecting several edible vegetable oil as the research object, The main contents and conclusions are as follows.(1) This paper explored acid value for edible vegetable oil based on rapid quantitative detection method using the Raman spectrum. The spectra processing methods as five points smoothing(S.G.), Normalize, three points first-order differential(1st D), three points second order differential(2nd D), three points 3 third-order differential(3rd D) and fourth-order differential(4th D), baseline correction(Baseline), standard normal variable transformation(SNV), and so on, totally 11 kinds of methods were using after outlier sample were carried out. The division of all samples sets were discussed to achieve the best selection of several main principle component for Raman spectroscopy. Then the linear PCR and PLS quantitative analysis model were been established. The result showed that after using dependence on the three order derivative on original spectra PLS was the best calibration model. The Rc and RMSEC of the calibration set were 0.942 and 0.212.(2) The paper explored edible vegetable oils peroxide values based on rapid quantitative detection method using the Raman spectrum. First, the 4801 spectral points was compressed into the contribution of wavelength data by SPA spectral wavelength bands selection method. The selection of the optimal value for models by support vector kernel parameters and kernel function was discussed. The nonlinear SVM and LS-SVM model worked more effective than the linear PCR and PLS model methods. For the view of parameters selection, SVM kernel function was better than RBF. The model of points filtered by SPA is better than others, the Rc and RMSEC of the calibration set were 0.949 and 1.869.(3) The use of Raman spectroscopy for the three components of the gradient blend oil sample was applied to conduct quantitative determination. Respectively, after Standard Normal Variable(SNV) Transformation and de-trending algorithm and Orthogonal Signal Correction(OSC) preprocessing algorithm the quantitative analysis model and partial least squares support vector machine algorithm were established by external cross-validation of the best model for verification. The results showed that three-component food reconcile oil containing canola oil, peanut oil and sesame oil to build PLS model to predict the most effect Rp and RMSEP after the OSC. The Rp and RMSEP were 0.990, 0.966, 0.977 and 0.019, 0.038, 0.026 respectively.(4)The rapid quantitative models of acid value and peroxide value for edible vegetable oil and the components in ternary blended edible vegetable oil based on Raman spectroscopy were feasible and high-accuracy. It can provide theory and reference for detecting the quality of edible vegetable oil. The paper can supply a new method of rapid non-destructive testing technology for the detection of oil freshness, quality level and discriminating of each component.