近红外光纤光谱法快速检测葡萄酒中酒精度

Rapid Detection of Alcohol in Wine by Near Infrared Optical Fiber Spectroscopy

目的构建近红外光纤传感检测系统,结合近红外光谱分析技术和化学计量学,对葡萄酒中酒精度进行快速检测。方法以葡萄酒为研究对象,构建近红外光纤传感检测系统,分别使用偏最小二乘法和主成分回归对葡萄酒中酒精度进行近红外光谱分析,进行模型参数的比较。采用国标GB/T 15038-2006《葡萄酒、果酒通用分析方法》中气相色谱法对近红外预测结果进行验证。结果采用偏最小二乘法建模的预测性能优于主成分回归分析。在偏最小二乘法建模中,其决定系数(R^(2))为0.9534,交叉验证均方根误差(RMSECV)为0.0283,预测均方根误差(RMSEP)为0.0179,相对分析误差(RPD)为3.0607。统计学分析表明近红外分析的预测值与气相色谱法测定值之间的差异无统计学意义。结论研究表明,近红外光纤光谱法用于葡萄酒中酒精度的检测,操作简便、快速。近红外技术在酒类品质监测中具有良好的应用前景。

Objective To establish a method for rapid determination of alcohol in wine using near-infrared optical fiber sensing detection system based on near infrared spectroscopy and chemometrics.Methods Taking wine as the research object,the near infrared optical fiber sensor detection system was constructed,using partial least squares(PLS)and principal component regression(PCR)to analyze alcohol content in wine by near-inrared spectroscopy,and compare the model parameters.The near infrared prediction results are verified by gas chromatography in GB/T 15038-2006"General Analysis Methods for Wine and Fruit Wine".Results The results showed that the prediction ability of PLS was better than PCR.The coefficient of determination(R^(2))was 0.9534,the root mean square error of cross validation(RMSECV)was 0.0283,the root mean square error of prediction(RMSEP)was 0.0179,and the relative percent deviation(RPD)was 3.0607 in the PLS analysis model.Statistical analysis showed that there was no significant difference between the predicted value of near infrared analysis and the measured value of gas chromatography.Conclusion The studies have shown that near infrared optical fiber spectroscopy is simple,fast,and can be used for alcohol detection in wine.This technology has good application prospects in wine quality monitoring.