Research On Near Infrared Technology For Rapid Detection Of Key Parameters Of Apple Vinegar During Brewing Process

The development of apple vinegar industry is very rapid in our country, and it has a huge consumer popularity and market capacity. However, there are many quality problems about the apple vinegar in market which seriously disturb the market order. Therefore, an fasty, efficient and accurate analysis technique is needed to improve the quality control level of apple vinegar industry. In this article, near infrared spectroscopy was used in combination with the stoichiometry method, and this article studied the fast and quantitative detection of the key components of apple vinegar during brewing process, such as the acceptance check of raw materials, the quality of cider and the quality of apple vinegar, aiming to provide a new technical reference for the real time nondestructive testing of the quality of apple vinegar during brewing process.The main research contents are as follows:(1)The research of fast and quantitative determination of the content of titratable acids and soluble solids of apple juice was done. Extracted the optimal numbers of main factors to optimize the data by using FD, SD, SNV and MSC, establish the model by combining partial least squares (PLS) method. As a result,R2 which is the verification decision coefficient, RMSEP which is the titration standard deviation and RPD which is the relative analysis error of the validation set of the model of the titratable acids and the soluble solids are 0.981,0.129,5.97 and 0.992,1.474,11.26.It was shown that the near infrared spectroscopy was feasible and effective for the determination of the main contents of apple juice.(2)The research on a rapid and quantitative determination method of reducing sugar and soluble solid content of apple vinegar by high fructose syrup was done. Two band filtering methods, CARS and none information variable elimination method were used to reduce the number of variables, and four kinds of pretreatment methods, FD, SD, SNV and MSC were combined to optimize the model, the partial least squares model of CARS filter was more ideal, the R2, RMSEP and RPD of reducing sugar and soluble solid model are 0.964,0.432,3.80,0.901,0.048,3.75, and by the test of paired t, there was no significant difference between the predicted value and the measured value. It indicated that the selected optimization method significantly reduced the modeling complexity, and established the foundation for the accuracy and stability of the model.(3)In order to control the brewing process of apple vinegar, the quantitative analysis of the alcohol content in apple wine was done. The quantitative analysis model of alcohol content of apple vinegar was established by PLS method, at the same time the interval partial least squares (iPLS), backward interval partial least squares (BiPLS), synergy interval partial least squares (SiPLS) and Genetic partial least squares (GA-PLS) were used toscreen spectral characteristic band of the whole spectral region and determined the best modeling method. As a result, the filtering of Feature band could be used to optimized the model, and the speed of model operation could be improved. Among that, the optimization effects of the GA-BiPLS and GA-SiPLS were most obvious, the modeling variables of alcohol content in apple wine were reduced greatly, at the same time, the R2 of the model reached 0.987 and 0.988, the RMSEP were 0.459 and 0.468, the stability and accuracy of the model was improved effectively, a new way to study the quality of apple wine was provided. At the same time, the application of near infrared spectroscopy technology in the production process of apple vinegar had was provided.(4)A rapid and quantitative detection method for the quality of original apple vinegar was studied. The characteristic wavelengths extracted by genetic algorithm partial least squares(GA-PLS) was used as the input variables of least squares support vector machines(LS-SVM), the article built NIR quantitative models of total acids and soluble solids in original apple vinegar. And it was compared with the result of the partial least squares (PLS) model established to determine the best modeling method. As a result, compared with PLS model, the R2, RMSEP and RPD values of the LS-SVM quantitative model of total acids and soluble solids content were better, which reached 0.964,0.039,5.01,0.979, 0.021,6.01.When verifying the independent test, the prediction accuracy of LS-SVM model was also significantly better than that of PLS model. The model could be applied to the rapid detection of total acids and soluble solids content in original apple vinegar. A reference basis had been provided for real-time and accurate monitoring of the content changes of main indexes of original apple vinegar and scientific optimization of near infrared spectral model.