近红外光谱法测定菜籽油中水分含量

Determination of moisture content in rapeseed oil by near-infrared spectroscopy

以菜籽油为试验材料,用干燥乙腈萃取菜籽油中水分,用傅里叶变换近红外光谱技术(FT-NIRS)结合偏最小二乘法(PLS)建立乙腈水分定量分析模型,并用乙腈水分模型间接预测菜籽油水分含量。结果表明:校正样品集的标准偏差(SEC)为0.014 6,预测样品集的标准偏差(SEP)为0.039,预测值与实测值的决定系数(R2)为0.921 6。菜籽油水分含量用FT-NIRS分析和卡尔费休法测定结果之间有良好的一致性,模型具有较高的预测能力,结果准确可靠,可代替常规分析方法。

Rapeseed oil was used as a raw material,and the moisture in the rapeseed oil was extracted by dry acetonitrile,then Fourier transform near-infrared spectroscopy（FT-NIRS） combining partial least squares method（PLS）was used to develop the moisture quantitative analysis model.The results showed that the SEC,SEP and coefficient of determination（R2） were 0.014 6,0.039,and 0.921 6,respectively.The model had high predictive ability to forecast the moisture content of rapeseed oil,and FT-NIRS could replace the conventional analysis methods.