摘要
通过气相色谱法分析花生油和棕榈油的混合油的脂肪酸组成,建立了人工神经网络分析二组分食用油混合模型的方法。分别基于混合油样品中棕榈酸和亚油酸含量变化的一元线性回归模型和基于全部脂肪酸组成的BP神经网络建立定量模型对花生油中棕榈油的掺杂量进行预报。结果表明,BP神经网络的预报准确率为96.7%,当棕榈油掺杂量≥0.050(V/V)时,相对偏差≤6%,其准确度高,能够实现二组分混合油掺混量的准确预报,为调和油的组成分析提供了新思路。
The fatty acid composition of the mixed oil of peanut oil and palm oil was analyzed by gas chromatography,and an artificial neural network method was established to analyze the two-component edible oil mixed model.Based on the unary linear regression model of palmitic acid and linoleic acid content in mixed oil samples and the BP neural network based on the composition of all fatty acids,a quantitative model was established to predict the doping amount of palm oil in peanut oil.The result showed that the forecast accuracy rate of BP neural network was 96.7%,when palm oil amount greater than or equal than 0.050(V/V),the relative deviation was below 6%.BP neural network can accurately predict the mixing of two components of mixed oil,provides new ideas for composition analysis of blending oil.
作者
王李平
林晨
张方圆
杨熙
麦小漫
范华均
WANG Li-ping;LIN Chen;ZHANG Fang-yuan(Guangdong Provincial Engineering Research Center for Efficacy Component Testing and Risk Substance Rapid Screening of Health Food,Guangdong Provincial Key Laboratory of Emergency Test for Dangerous Chemicals,Guangdong Institute of Analysis(China National Analytical Center Guangzhou),Guangdong Academy of Sciences,Guangzhou,Guangdong 510070)
出处
《安徽农业科学》
CAS
2020年第21期202-204,209,共4页
Journal of Anhui Agricultural Sciences
基金
广东省科学院青年科技工作者引导专项(2019GDASYL-0105013)
广东省科技计划项目(2015B090906023)。
关键词
人工神经网络
食用油
脂肪酸组成
混合模型
气相色谱法
二元组分分析
Artificial neural network
Edible oil
Fatty acid composition
Mixed model
Gas chromatography
Binary component analysis
