Identification Of Flavor Components In Milk,Milk Powder And Fermented Milk Based On Electronic Nose

Milk and dairy products have rich nutritional value and are deeply loved by consumers.Volatile flavor components help to improve the taste of milk and dairy products,which is an important attribute affecting consumers' purchases.Therefore,the volatile flavor components and aroma of milk and dairy products have always been the focus of research.However,a large number of studies on the flavor components of dairy products have focused on the analysis of the taste components of dairy products and their key aroma supplying substances.A few studies were carried out on their flavor formation mechanism,flavor influencing factors,the types of milk,and analysis of dairy products samples.In this paper,the volatile flavor components,aroma components,and aroma components in different fat content and UHT organic milk,milk powder of different ages,and fermented milk with different starters were analyzed by ultra-fast GC electronic nose(UFGC E-nose).On this basis,the root causes of their classification were analyzed more deeply.This thesis consists of the following three parts:(1)UFGC E-nose technology was used to identify the volatile flavor components of UHT whole fat,low-fat,and skimmed milk,which was quantitatively analyzed by the external standard method.The aroma differences and reasons of UHT milk with different fat content were analyzed by odor activity value and aroma radar map.The aroma characteristics of UHT milk were classified and predicted by combining principal component analysis and multilayer perceptron neural network.The results showed that 49 volatile flavor components were identified from UHT milk with different fat contents,of which 29 were common components.Isobutanol,acetoin,1-pentanol,(E)-3-hexenal,and decanal were only detected in UHT whole milk;3-methylheptane,2,6-dimethylpyrazine and(E)-2-nonene-1-ol were only detected in UHT low-fat milk;?-pinene,5-methylfurfural and capric acid were only detected in UHT skimmed milk.In addition,2-pentanone was not detected in UHT organic whole milk;ethanol,cyclopentanone,and nonanal were not detected in UHT organic low-fat milk;dimethyl disulfide was not detected in UHT low-fat milk;2-undecanone and(E)-2-heptanaldehyde were not detected in UHT skimmed milk.UHT milk has vegetable flavor,cream flavor,grass flavor,fruit flavor,and malt flavor.Milk flavor,vegetable flavor,and malt flavor are most widely distributed in UHT whole milk;fruit flavor is most widely distributed in UHT low-fat milk;grass flavor is most widely distributed in UHT skimmed milk.PCA in multivariate statistical analysis can distinguish UHT milk of the same type and different brands,and can also trace the source of UHT milk in different countries;MLP-NN model can classify UHT milk with an accuracy of 98.6%.The established UFGC E-nose detection and multivariate statistical analysis method provide an important method basis for the rapid and accurate classification of different brands and types of UHT milk.(2)UFGC E-nose technology was used to identify the volatile flavor components of milk powder for infants,students,adults,and middle-aged and elderly people.The volatile flavor components were quantitatively analyzed by the internal standard method and external standard method.The aroma differences and causes of milk powder at different ages were analyzed by using odor activity value and aroma radar map.The aroma characteristics of milk powder were classified and predicted by combining linear discriminant analysis and a linear support vector machine.The results showed that 54 volatile flavor components were detected in milk powder of four different ages,of which 37 were common components.Methyl propionate and 1-octene-3-one only exist in infant milk powder;(E,E)-2,4-decadienal and ?-undecanolide are unique component of student milk powder;heptanaldehyde was only detected in adult milk powder;2-butanone and benzaldehyde are only contained in milk powder of middle-aged and elderly people.In addition,only infant milk powder was not detected with 1-propanol,ethylbenzene,vanillin?-Caprolactone;only student and adult milk powder did not contain m-xylene;(E)-2-pentenoaldehyde,butyl acetate,(Z)-2-hexene-1-ol and n-hexanol were not found in milk powder of middle-aged and elderly people;?-decalactone was not detected in milk powder of infants and middle-aged and elderly people.The milk powder of the four ages has vegetable flavor,milk flavor,fruit flavor,lactic acid flavor,tar flavor,flower flavor,grass flavor,and caramel flavor,while the almond flavor is the characteristic smell of adult milk powder.LDA in multivariate statistical analysis can clearly distinguish different brands of milk powder in the same age group;Combined with LDA and LSVM model,milk powder of different ages can be classified and predicted,and the accuracy is as high as 100.0%.The established UFGC E-nose detection and multivariate statistical analysis method provide an important method basis for the rapid and accurate classification of milk powder at different ages.(3)UFGC E-nose technology is used to identify the effects of fermentation of different starter cultures,including traditional starters(Streptococcus thermophilus,Lactobacillus bulgaricus),Lactobacillus fermentans,Bifidobacterium lactis,and Lactobacillus casei,on the volatile flavor components of fermented milk.The aroma differences and reasons of fermented milk with different starters are analyzed by relative odor activity value,aroma radar,and sensory evaluation.The aroma characteristics of fermented milk were classified and predicted by principal component analysis and cluster heat map.The results showed that 43 volatile flavor components were detected in fermented milk,of which 34 compounds were common components.After adding probiotics,the composition of volatile flavor components in fermented milk changed relatively little,but the content changed greatly.The relative odor activity value and aroma radar map showed that the odor of the four kinds of fermented milk was significantly different from the sensory evaluation results.Combined with principal component analysis and cluster heat map,the grouping and volatile flavor components of different fermented milk were analyzed.The grouping results were consistent,including acetaldehyde,acetone,2,3-butanedione,2-methylbutyraldehyde,2,3-glutaraldehyde,pentanol,butyl acetate,3-nonone,decanal,2-undecanone,2-dodecanone,?-decalactone,?-pentadecane,and 10-methyl eicosane are the main reasons for sample separation.The established UFGC E-nose detection and multivariate statistical analysis method has important guiding significance for better understanding the typical volatile compounds in fermented milk and improving product quality.