基于电子鼻技术对羊奶粉中掺假牛奶粉的快速检测

Rapid detection of goat milk powder adulterated milk powder based on electronic nose

利用电子鼻技术结合多种化学计量学方法,对羊奶粉掺假牛奶粉进行了快速定性判别和定量分析。将牛奶粉按不同比例掺入到羊奶粉中,并按掺假羊奶粉与水为1∶7.2的比例配成溶液,进行电子鼻检测。通过电子鼻采集不同比例掺假奶挥发性成分的响应值,然后利用主成分分析(PCA)、Fisher线性判别分析(FLDA)以及线性回归拟合分析进行定性判别和定量分析,建立基于电子鼻技术检测羊奶粉掺假的方法。结果表明,FLDA及PCA都能够区分出不同比例的掺假奶,且FLDA区分效果优于PCA;线性回归拟合分析的相关系数为94.8%,预测值与实际掺假值呈现一定的线性关系,表明该模型具有较好的泛化能力。因此利用电子鼻实现羊奶粉掺假的快速定性判别和定量分析是可行的。

The goat milk powder adulterated milk powder was rapidly qualitatively discriminated and quantitatively analyzed by an electronic nose combined with a variety of chemometric methods. Milk powder was added to the goat milk powder according to the different propor- tions, and then mixed with 7.2 times of water for electronic nose to detect. The volatile components emanating from the adulterated goat milk samples were gathered by the electronic nose. The electronic nose technologies for discriminating the different proportions of adulterated goat milk powder were established through principal component analysis （PCA）, fisher linear discriminant analysis （FLDA） and Linear regression fitting analysis. The results showed that both PCA and FLDA can discriminate the different proportions of adulterated goat milk pow- der, but FLDA was more effective than PCA. The correlation coefficient of linear regression analysis was 94.8%, the predicted value and the actual adulteration value showed a certain linear relationship, indicating that the model has better generalization ability. So using electronic nose to discriminate the adulterated goat milk powder rapid qualitative discrimination and quantitative analysis is feasible.