近红外光谱仪在浓酱兼香型白酒酒醅检测中的应用研究  

Application of Near-Infrared Spectrometer in the Detection of Fermented Grains of Nongxiang-Jiangxiang Baijiu

作　　者：邓丽娟 邹小月 熊笠君 唐宁 DENG Lijuan;ZOU Xiaoyue;XIONG Lijun;TANG Ning(Baiyunbian Distillery Co.Ltd.,Songzi,Hubei 434200,China)

机构地区：[1]湖北白云边酒业股份有限公司,湖北松滋434200

出　　处：《酿酒科技》2022年第5期128-132,共5页Liquor-Making Science & Technology

摘　　要：选用浓酱兼香型白云边酒酿造生产过程中的酒醅作为样品集,采用近红外光谱方法与对应的手工检测值匹配后形成数据集,通过主成分分析,基于最小二乘法建立酒醅数据分析模型。经定标验证证明模型测得的酒醅水分、酸度、还原糖和淀粉值与实际值拟合良好,各相关系数RSQ和经系统偏差修正后的预测偏差SEP(C)分别为0.935、0.916、0.940、0.981和0.306、0.070、0.092、0.212。经盲样检测,表明构建的模型有良好的预测能力且稳定性较高,能够应用于酿造中大批量酒醅的快速检测。Using the fermented grains of Nongxiang-Jiangxiang Baijiu as the sample set,the data set was formed by matching the de-tection results of near-infrared spectroscopy with that of manual detection.Through principal component analysis and least square method,the fermented grains analysis model was established.Through verification by calibration equation,the values of moisture,acidity,reducing sugar and starch measured by the model fitted well with the actual values.The correlation coefficients RSQ were 0.935,0.916,0.940,and 0.981,and the SEP(C)were 0.306,0.070,0.092,and 0.212,respectively.Blind sample detection proved that the model had good prediction performance and high stability,and could be used for the rapid detection of large batch of fermented grains in production practice.

关 键 词：近红外光谱仪 酒醅 模型 快速检测 

分 类 号：TS262.3[轻工技术与工程—发酵工程] TS261.7[轻工技术与工程—食品科学与工程]