基于电子舌的料酒味觉特征辨识与定量分析

Taste Identification and Quantitative Analysis of Cooking Wines Based on Electronic Tongue

为探明电子舌对调味料酒生产工艺的判别能力和理化指标的预测能力,本研究采用电子舌和理化检测手段,结合不同统计方法,对54份料酒样品分别建立定性和定量分析模型。结果表明,应用主成分分析(PCA)可以区分不同生产工艺的料酒样品,第一主成分为鲜味,贡献率62.4%,第二主成分为酸味,贡献率33.2%;应用簇类独立软模式法(SIMCA)可以准确判别酿造料酒和配制料酒,各传感器区分能力(DP)>5,识别率达到100%;应用偏最小二乘法(PLS)将传感器信号与行标方法检测结果进行拟合,总酸、氨基酸态氮和食盐的验证集标准偏差与预测标准偏差的比值(RPD)分别为12.1、6.5和14.1,建立的模型效果良好,可进行准确的定标和预测;酒精度RPD值为2.7,也可进行定量分析,但模型稳定性较弱。本研究结果为应用电子舌对调味料酒进行品质区分和检测提供了理论和实践基础。

In order to investigate the capability of electronic tongue to distinguish manufacturing progress and predict physicochemical properties of cooking wines, the qualitative and quantitative analysis models of 54 cooking wines were established using electronic tongue and physical and chemical detection methods, combined with different statistical methods. The results showed that principal component analysis(PCA) could be used to differentiate cooking wines from different types of production processes. The first principal component was the signal of fresh taste flavour sensor with the contribution of 62.4%, and the second one was the sour taste with a contribution of 33.2%. The soft independent modeling of class analogy(SIMCA) could be used to accurately discriminate brewed and prepared cooking wines. The discrimination power(DP) of each sensor was more than 5, and the recognition rate was 100%.The partial least squares(PLS) method was used to fit the sensor signal with the results of the trade standard method. The RPD values of total acid, amino acid nitrogen and salt were 12.1, 6.5 and 14.1, respectively. The established models exhibited good effect, and could be used for accurate calibration and prediction. That of alcohol content was 2.7. The model could be predicted but was not stable enough. The results of this study provided theoretical and practical basis for the application of electronic tongue in the quality discrimination and detection of cooking wines.