采集六一散混合过程中样品近红外光谱, 建立甘草酸含量近红外(NIR)偏最小二乘(PLS)定量模型。 结果校正集相关系数rcal=0.998 5, RMSEC=0.044 0 mg·g-1, 预测集rval=0.947 4, RMSEP=0.124 mg·g-1, 表明近红外光谱法可作为六一散混合过程中甘草酸含量的快速测定方法。 在定量模型建立的基础上, 设计验证试验, 采用由Liao等提出的基于蒙特卡罗仿真的方法, 估计β-容度-γ-置信容许区间, 并计算NIR定量分析不确定度, 绘制不确定度轮廓。 结果表明甘草酸含量高于1.56 mg·g-1时, 测量不确定度在可接受范围(λ=±20%)内, 表明所建不确定度评估方法可有效评价不同浓度水平下的甘草酸含量NIR定量模型的准确性和可靠性, 可为其他中药NIR定量分析方法的不确定度评估提供借鉴。

The near infrared (NIR) spectra of Liuyi San samples were collected during the mixing process and the quantitative models by PLS (partial least squares) method were generated for the quantification of the concentration of glycyrrhizin. The PLS quantitative model had good calibration and prediction performances (rcal=0.998 5, RMSEC=0.044 mg·g-1; rval=0.947 4, RMSEP=0.124 mg·g-1), indicating that NIR spectroscopy can be used as a rapid determination method of the concentration of glycyrrhizin in Liuyi San powder. After the validation tests were designed, the Liao-Lin-Iyer approach based on Monte Carlo simulation was used to estimate β-content-γ-confidence tolerance intervals. Then the uncertainty was calculated, and the uncertainty profile was drawn. The NIR analytical method was considered valid when the concentration of glycyrrhizin is above 1.56 mg·g-1 since the uncertainty fell within the acceptable limits (λ=±20%). The results showed that uncertainty assessment can be used in NIR quantitative models of glycyrrhizin for different concentrations and provided references for other traditional Chinese medicine to finish the uncertainty assessment using NIR quantitative analysis.