Researches On Technology Of Reducing Nitrosamine In Tobacco And Correlation Between Aroma Components And Tobacco Quality

In recent years, the relationship of smoking and health is more concerned and anti-smoking campaign is springing up around the world. So how to reduce the content of harmful substance in tobacco has become one of key factors that concern rise and fall of tobacco trade. And on the other hand, researching the cigarette of Chinese style with character of "high flavor and low harmfulness", improving the scientific characters of formulation of cigarette and enhancing the competitive capability of Chinese tobacco trade are taking on great significance with our country joining WTO. To achieve those, the following research works were carried out on the basis of extensive literature investigation. 1. A method for reducing nitrosamine in cigarette smoke using antiscorbutic vitamin(VC) ,fertility vitamin(VE) and sepiolite was discussed because nitrosamine was strong carcinogens. Tobacco samples were dealt with 2.5% VC, 4.0% VE respectively, And sepiolite powder placed on cigarette filter bottom acted directly on the cigarette smoke. The results showed that the content of N-nitrosamines in cigarette smoke was separately decreased 33.8% and 34.3% after tobacco leaves dealt with 2.5% vitamin C,4.0% vitamin E solution were deposited 12 hours, and decreased 24.1% by sepiolite powder. This means that a technique method was provided for enhancing the security of smoking cigarette. 2. Different tobacco samples were clustered using fuzzy c-means method on the basis of composition and content of aroma component in tobacco firstly. Collating the results of smoking quality, this method was showed great capability of clustering tobacco samples. And a good result of clustering for system of tobacco which partly had transcendent knowledge of smoking quality was provided using fuzzy clustering method, which had some instructing significance for the choose of substitute tobacco and exploitation or rectification of formulation of cigarette in cigarette production. And the effect of aroma substance in tobacco on flavour of cigarette was distinguished by adjust the eigenvalue weight of aroma component which could instruct the work of match flavour of cigarette. 3. A pattern recongnition model for tobacco quality was firstly built by using backpropagation(BP) neural network combined with genetic algorithm based on the relationship between tobacco quality and chemical component in tobacco and expert experience, thus providing a new recognition method for identifying aromatic characteristics of tobacco. The results showed that neural network structure of tobacco aroma components was global optimized, and the neural network trained by genetic algorithm was more efficient and correct than by gradient descent method . 4. The database of aroma components in tobacco was built and tobacco assistant system was improved by introducing aroma components into tobacco expert system. Meanwhile, the fuzzy clustering method and genetic neural network algorithm were applied to tobacco assisitant system, thus enhancing the ability of tobacco system clustering and identifying unknown tobacco.