特征性峰位法结合化学计量学构建奶粉脂肪含量定量模型

Establishment of Quantitative Model of Fat Content of Milk Powder by FTIR Spectroscopy Based on Characteristic Peak Position and PLS Methods

目的:在对FTIR和UV两种定量方法效果比较的基础上,将二者结合起来,以特征性峰位为指导,选择合适的光谱范围以期获得更为稳定高效的奶粉脂肝含量定量分析模型。方法:以市售奶粉为研究对象,利用化学测定的脂肪含量作为参考值,以脂肪的特征性峰位为备选范围,优化获得奶粉脂肪含量的定量分析模型。结果:所获得的FTIR定量分析模型定量模型的R^2=0.931,RMSECV=1.69,RPD=4.01,维数=6。结论:单独依靠特征峰位法或偏最小二乘法(PLS)的自动优化,都很难获得稳定而高效的定量分析模型,但二者的结合使用,可以有效地改善FTIR定量模型的稳定性,并降低了维数,提高了在实际检测中的应用潜力。

Objective: The two methods (FTIR and UV) are firstly compared and finally combined, to obtain a more stable and efficient quantitative analysis model. Methods: The fat contents of milk powder were regarded as the reference values obtained by using the chemical method, and the characteristic peaks of fat were used as the alternative range for choosing the quantitolive analysis model was finally obtained. Results: The quantitative model of milk powder was obtained by using characteristic peak position and PLS methods. The R^2=0.931, RMSECV=1.69, RPD=4.01, Rank=6. Conclusion: It is difficult to obtain a stable and efficient quantitative model depending on one single method, but the combination of the two methods can effectively improve the stability and efficiency of FTIR quantitative model. The rank is reduced and the possibility of application is improved for future use.