摘要
采用InfraXact型多功能近红外分析仪,利用正交试验设计研究了不同预处理和校正方法对建模效果的影响,建立了牛奶中青霉素定标模型,利用目标函数法对模型进行评定,并对评定的最优模型适用性进行验证。试验结果表明:在最优组合处理下,模型的验证决定系数为0.975,相对标准差为12.87%,相对分析误差为6.1758。结果表明:近红外光谱定量分析牛奶中青霉素的含量是可行的,模型的稳定性和预测能力较好。
The calibration model for quantification of penicillin in milk was developed by using the InfraXact NIR analyzers, which was evaluated by objective function method and applicability of the optimal models was validated. Effects of different pretreatment and calibration methods on model- ing were studied through orthogonal test design. The results showed that under optimal conditions, the coefficient of determination was 0.975, RSD was 12.87%, RPD was 6.1758. It is indicated that quantification of the content of penicillin in milk with NIRS was feasible, and the model is stable and predictable.
出处
《中国酿造》
CAS
2012年第12期124-127,共4页
China Brewing
基金
陕西师范大学大学生创新性实验计划项目(CX10035)
关键词
近红外光谱技术
青霉素
牛奶
near infrared spectroscopy
penicillin
milk
