Taste Quality Detection Of Black Tea Based On Electronic Tongue Techniques

Black tea is one of the world’s most popular beverage.With consumption growing up, the quality control becomes important increasingly. The traditional detection methods such as sensory evaluation lacks sufficient reliability, the study of technology which can be used to identify the tea quality quickly and accurately is very important. Electronic tongue,a new instrument simulates human taste system,is fast, accurate, convenient, objective, etc., in the sample identification.It has a very broad application prospects. In this work, black tea were used as the main material object and the aim was to study the application of the electronic tongue on qualifing tea samples by means of the pattern recognition methods,and analyze the relationship between the signals of the electronic tongue and the chemical components or sensory scores.The results and conclusions were listed as followed:1. The study of classificating black tea samples according to electronic tongue signals. By studying the characteristics of the electronic tongue response signals in the detection process,determined the first120s sensor response signals had relatively small deviation, so we regarded120s as data acquisition time. To contrast different pattern recognition classification methods,we found that principal component analysis reflected the general classification of samples’types; stepwise discriminant analysis classified tea samples correctly; stepwise discriminant analysis and neural network both had the ability to predict unknown samples accurately.2. The study of the correlation between flavor substances and taste scores of black tea samples:It was found that there were differences in flavor substances and sensory scores of samples among different regions and grades.The sensory results explained and demonstrated the classification of the samples by the e-tongue.3. The study of quantitative analysis of black tea samples’quality by e-tongue:With multiple stepwise linear regression method,we established linear regression models between the electronic response value and flavor substances, taste score of black tea samples in different areas and grades. Statistical models were significant related and had good fitting degrees, the models of Qimen black tea were the best, Dianhong tea followed,tea samples in different religions were the worst. The correction coefficient of polyphenols, catechins and simple catechins were the largest among the chemical substances in the study.It was the samples in different producing areas that had the largest calibration coefficient of astringent,concentration,thickness,while the samples in different grades had the largest calibration coefficient of chen taste, freshness and smoothness.In addition, the sensors’response degree were different among the models, both of sensor GA and sensor JB played important roles in affecting the quality of flavor and taste components in three trials.