Discriminant Analysis Of Orange Juice Beverage Quality Based On Synchronous Fluorescence Method

In recent years, along with the social life level, Orange Juice drinks with its rich nutrition, taste the public has become the first choice of drinks in the daily life of people. The attendant is about its products adulterated, fake phenomenon more and more serious. Synchronous fluorescence spectroscopy and compared with other fluorescent technology, has many unique advantages, such as simplification can reduce spectrum, Raman optical scattering light effect, select the appropriate wavelength difference reduced spectral overlap. The advantages of fluorescence spectroscopy in order to complex mixtures in similar to the "orange juice" and other components in the show. Therefore, this paper uses a combination of this method with principal component analysis and partial least squares two chemometric method, discriminant analysis for a variety of quality of different sources of different brands of orange juice and orange juice beverage. The main contents and research results are as follows:1. pairs of fresh Orange Juice and a commercial brand Orange Juice synchronous fluorescence scanning and then use the Origin7.0 software to obtain the three-dimensional fluorescence spectra, and extract the characteristic peaks of three groups from the chart. Fluorescent substance and then look for the Orange Juice in, and analyzes its structure characteristics, that the three groups of the characteristic peak of material respectively, vitamin B2, vitamin B6 and flavonoids. At the same time, using the principal component analysis method with excitation wavelength 240 nm and emission wavelength of 640 nm, with the excitation wavelength difference of 30 nm conditions for the spectral scanning and data synchronous fluorescence spectra obtained by clustering on all Orange Juice samples, then the recognition by clustering information, the effect is good.2. using synchronous fluorescence method is used to scan the spectrum of Orange Juice drinks Kangshifu, unity of three kinds of brand Minute Maid, and then use Origin7.0 software to obtain the three-dimensional fluorescence spectra. Analysis of spectrum shape and characteristic parameters from the summarized, results showed that:three brand Orange Juice three-dimensional fluorescence spectrum contour map has many similarities, but in the fluorescence peak is not the same performance as the fluorescence peak wavelength range, intensity of fluorescence and excitation wavelength interval and so different, combined with principal component analysis cluster information scores plot in the rapid identification of different Orange Juice beverage brands; several different batches of Minute Maid and Kangshifu two brand samples at the excitation wavelength 240-640 nm, wavelength difference spectrum scanning at 30nm, and then also use principal component analysis to the income spectrum data inversion, the results showed that:the two brand beverage properties in the shelf period with batches of change, but the degree of change is different. For the same operation to open the cover of room temperature storage of Minute Maid beverage, the best time to drink to get within 24h, the storage limit is 48h.3. The volume fraction of the Orange Juice raw juice content of quantitative analysis of 3 using synchronous fluorescence spectroscopy combined with partial least squares. Data synchronous fluorescence spectra of samples were pretreated and then select different spectral bands to establish prediction model of raw juice content combined with partial least squares method, and use this model to partition validation set samples were predicted, to verify the robustness of the model and the feasibility of this method with the results, results showed:the establishment of smoothing of spectral data and then add a derivative treated more than the original data suitable for the conditions of the model; the use of the spectral band model established by the better performance than the interval of the established model. Predict income prediction mean squared error set with optimal model, is finally obtained for verificationis 0.035832the coefficient of determination 0.972570. Quantitative analysis of this shows the volume fraction content method used in this experiment can achieve the original juice.