电子舌在黄酒检测中的应用

Application of Electronic Tongue in the Detection of Yellow Rice Wine

运用Astree型电子舌对不同的黄酒样品进行了检测,并结合PCA、DFA及BP神经网络,对3个厂家不同酒龄的9个样品进行分析。结果表明,电子舌不仅可以区分不同厂家的产品,同时也可以很好地区分同一厂家不同陈酿时间的黄酒样品。对古越龙山品牌3个酒龄各8个批次的样品进行了理化指标的分析,并将数据与电子舌传感器数据信息建立BP神经网络模型,结果显示除氨态氮含量的拟合与预测较为准确外,总酸、p H值等其他理化指标的拟合与预测效果都不理想。

In the experiment, Astree-type electronic tongue combined with principal component analysis （PCA）, discriminant function analysis （DFA） and back propagation neural network （BPNN） as pattern recognition technology was employed for the detection of nine yellow rice wine samples of different wine age from three different manufacturers. The results showed that, the use of electronic tongue could differentiate the samples from different manufacturers, meanwhile, it would discriminate wine samples of different wine age from the same manufacturer. Through the analysis of the physiochemical indexes of Guyuelongshan wine samples of three different wine age and eight different production batches, BPNN models were established based on the analytic data and electronic tongue data. The model was good in the fitting and prediction of ammoniacal nitrogen but not ideal in the fitting and prediction of other physiochemical indexes such as total acids and pH value etc.