食品科学 ›› 2016, Vol. 37 ›› Issue (24): 142-148.doi: 10.7506/spkx1002-6630-201624022

• 安全检测 • 上一篇    下一篇

基于电子鼻和神经网络对广式香肠脂肪氧化的评价

顾欣哲,吴振川,刘芮瑜,尹 涛,何淑文,屠 康,潘磊庆   

  1. 南京农业大学食品科技学院,江苏省肉类生产与加工质量安全控制协同创新中心,江苏 南京 210095
  • 出版日期:2016-12-25 发布日期:2016-12-21
  • 基金资助:
    国家自然科学基金面上项目(31571769);江苏高校优势学科建设工程资助项目;安徽省科技攻关计划项目(1604a0702031)

Evaluation of Fat Oxidation in Cantonese Sausage during Processing and Storage Using an Electronic Nose and Artificial Neural Network

GU Xinzhe, WU Zhenchuan, LIU Ruiyu, YIN Tao, HE Shuwen, TU Kang, PAN Leiqing   

  1. Jiangsu Collaborative Innovation Center of Meat Production and Processing Quality Safety Control, College of Food Science and Technology, Nanjing Agricultural University, Nanjing 210095, China
  • Online:2016-12-25 Published:2016-12-21

摘要: 采用电子鼻技术对广式香肠加工和贮藏过程中的脂肪氧化程度进行检测。在加工和贮藏中,分别提取香肠烘干0~120 h和贮藏0~20 周的电子鼻响应值,同时检测香肠的酸价(acid value,AV)和过氧化值(peroxidevalue,POV)来评价香肠的脂肪氧化程度,建立两者之间相关性。通过Loading分析、方差分析和Pearson相关性分析评价10 个传感器对香肠气味的贡献率,选出最佳传感器阵列,通过人工神经网络方法建立香肠AV和POV的预测模型。结果表明:S4、S6、S7、S8和S9为香肠加工过程中对脂肪氧化表征的最佳传感器阵列,S6、S7、S8和S9为香肠贮藏过程中的最佳传感器阵列。在加工和贮藏过程中模型预测效果较好。其中,对于加工过程,基于全部传感器阵列模型对AV和POV预测的R2分别为0.959和0.930,而基于最佳传感器阵列对AV和POV预测的R2分别为0.930和0.914;对于贮藏过程,基于最佳传感器对POV预测模型R2为0.805外,其余皆在0.9以上。因此,电子鼻在广式香肠加工和贮藏过程中对其脂肪氧化程度的检测有着比较好的效果,可以进一步应用到广式香肠的商业生产和贮藏。

关键词: 广式香肠, 电子鼻, 脂肪氧化, 神经网络, 预测

Abstract: In this study, an electronic nose was used to detect the flavor of Cantonese sausage during processing (0, 12, 24, 36, 42, 48, 54, 60, 72, 96 and 120 h) and storage (0, 2, 4, 6, 8, 12 and 20 weeks). The acid value (AV) and peroxide value (POV) were simultaneously measured to evaluate fat oxidation in Cantonese sausage. The contribution rates of 10 sensors to the flavor of sausages were evaluated through loading analysis, variance analysis and Pearson analysis. The results indicated that as the optimal sensor arrays for monitoring fat oxidation, S4, S6, S7, S8 and S9 for processing and S6, S7, S8 and S9 for storage were selected. Artificial neural network (ANN) models were developed to predict the degree of fat oxidation in Cantonese sausage using electronic nose data. The R2 values of the models based on all sensors for AV and POV prediction during processing were 0.959 and 0.930, respectively, while those were 0.930 and 0.914 based on the optimal sensors, respectively. During storage, all the R2 values were greater than 0.9, except for the POV prediction model based on the optimal sensors with R2 value of 0.805. Therefore, the electronic nose was suitable for evaluating fat oxidation in Cantonese sausages during processing and storage, which could be further applied to guide the commercial production and storage of Cantonese sausage.

Key words: Cantonese sausage, electronic nose, fat oxidation, artificial neural network, prediction model

中图分类号: