| Camellia oil is derived from the seeds of the Theaceae, and it’s performance is very well and it’s use is widespread.Camellia oil is clear, delicious, free of cholesterol, erucic acid, and has good hygienical function.Therefore, camellia oil had developed for multi-purpose and it caused the extreme imbalance of supply and demand. It contributes a sharp rise in camellia oil price which is 3-6 times market price of common edible vegetable oil. In order to seek exorbitant profits, some illegal businessmen adulterate pure camellia oil with cheap edible vegetable oil. The doped camellia oil is sold as pure camellia oil, so the consumers are deceived and violated their benefit. However, due to some edible vegetable oil has the similar fatty acid composition with the camellia oil, the conventional chemical methods are difficult to identify the low-cost edible vegetable oil and its content added into the camellia oil.Some manufactures even mix with mineral oil and poisonous vegetable oil,it will harm public health. These cases seriously affect the brand reputation of the camellia oil and restrict the development of the camellia oil industry. In addition, the market prices of the camellia oils are quite different because their different preparation methods, so some illegal production operators don’t label the actual manufacturing technique to make a killing.At present, the detection of the doped camellia oil and the preparation methods of camellia oil are saponification, freezing test, gas chromatograph, fluorescent spectrometry, etc. These methods are rough, labor-intensive, time-consuming and complex. But for quality supervision, the determination that the camellia oil is doped or not must be given,it is sure that what the kind and amount of the mixed oils is, and the process methods of camellia oil is. To solve these problems and meet the needs of the modern quality supervision, a simple, rapid, accurate and low cost method is urgently needed in order.The camellia oil samples were gotted by compression and extraction process. The doping identification,extraction process determination and temperature on the performance of camellia samples were built by near-infrared spectroscopy, fluorescence spectroscopy,raman spectros- copy and chemometrics method.It studied temperature on the performance of dopping identification with two- dimensional correlation method. The main contents of this thesis are as follows:1. The production process of camellia oil was distinguished by near infrared spectros-copy technology and the model were setup by UVE-PLS-LDA and MIA-SVM methods to remove useless band. The correct rate of UVE-PLS-LDA classification method for identify-cation of calibration set and prediction set samples was 100%; the sensitivity, specificity and accuracy of MIA-SVM classification methods of calibration set and prediction set were 100%,87.50%,93.75 and 100%,87.50%,93.75%.2. The camellia oil with soybean oil or rape seed oil were qualitatively and quantitatively studied by near infrared spectroscopy. The qualitative identification and classification model were established by the methods of UVE-GA-LDA.this model was applying to recognize doped samples with more than 2% rapeseed oil, and the correct rate of classification was 100%, but the correct classification rate is only 50% when rape seed oil was below 1%;Near infrared spectroscopy combined with CARS-PLS-LDA method can effectively identify 1%～50% soybean oil in camellia oil, the recognition correct rate, sensitivity and specificity were 100%; near infrared spectroscopy can be used for quantitative detection of rapeseed oil add to camellia oil, the correlation coefficient and RMSEP of the UVE-SPA-MLR model were 0.980 and 0.670%.3. The Raman spectra can be used to qualitative and quantitative detect rape seed oil or soybean oil in camellia oil samples with different volume fraction. The PLS was used to quantitative detecting doped oil and PLSDA methods are used to recognize the type of doped oil. The results show that the airPLS method can effectively eliminate the background fluorescence of Raman spectrum, SNV pretreatment method is effective to improve the spectral quality, airPLS-SNV-PLS was applied to quantitative analysis the volume fraction of camellia oil mixed with rapeseed oil and soybean oil and the correlation coefficient of prediction model is greater than 0.950.The correct rate of model was as high as 97.67% when using airPLS-SNV-PLSDA discriminant model to distinguish pure camellia oil, camellia oil mixing rapeseed oil and camellia oil mixing soybean oil.4. In order to study the influence of temperature on the quality of camellia oil, the model of quantitative analysis of dopping camellia oil was built by fluorescence spectroscopy method with the effect of temperature. It found that, in the range of 100-240℃, the QRFST model can be build by partial least squares regression method.the model can identify the camellia oil, rapeseed oil and palm oil.According to the different temperature of mixed oil (camellia oil and palm oil) with different mixing ratio (range 0.1-0.5), the quantitative scores of working curve was setup. The results show that, the correlation coefficient of QRFST model with cross validation was higher than 0.99 for prediction temperature and the actual temperature measurement, and the mixed oil temperature can be predicted by using the QSTR model and the fluorescence spectrum; The PLS model with cross validation was build by heating mixed oil with he volume ratio of palm oil in the range of 0.1-0.5.The mixed volume ratio of camelliaoil and palm oil wereachieved by means of the intercept and volumeratio after calculating the intercept of the predicted temperature and the actual temperature regression line.5. The pure camellia oil and doped oil samples (mixed rapeseed oil or palm oil) were identified and analyzed based on the methods of fluorescence spectroscopy and two-dimensional correlation spectroscopy with external disturbance of the temperature change. For the pure camellia oil samples and other doped oil, automatically peak of two-dimensional map of pure camellia oil samples in the range of 650-700 nm wavelength is not exist, but the automatically peak of two-dimensional map can be observed in the range of 650-700 nm; NPLS-DA method for detecting the doped oil have good rate of correct recognition, the rate of predicting sample and test sample sets were 93.94% and 97.73% respectively.We provides a rapid analysis method to determine the doping camellia oil and the different production process of camellia oil for quality supervision departments, and provides a fast, nondestructive identification method for the relevant departments to crack down on illegal business. |
PDF Full Text Request