Abstract: In drinks production, quick identification of quality stability is very important for insuring drinks quality and improving production technique. Gas sensor array technology and neural network pattern classification technique are used to study quick identification of drinks quality stability. The results of identification experiments showed that characteristic information of drinks could be tested by appropriate sensor array. The proper identification model of drinks quality could be set up by RBF neural network based on genetic algorithms, and the quick identification of quality stability in one drink production could also be achieved.

Key words: drink, quality, stability, sensor array, neural network