采用PEN3电子鼻检测中式香肠发酵过程中(0、2、6、11、15 d)气味指纹变化数值,通过主成分分析(PCA)、线性判断分析(LDA)和偏最小二乘法(PLS)预测模型构建,研究电子鼻气味指纹参数与挥发性盐基氮(TVB-N)、过氧化值(POV)指标的相关性。结果表明:在进样温度35℃,顶空体积100 mL,富集时间30min,采集时间55~60 s的条件下,PEN3电子鼻的传感器阵列对中式香肠的风味变化区分度最大,PEN3电子鼻的传感器阵列可以较好地识别后熟发酵时间不同的香肠样品的挥发性气味差异。运用PLS预测模型分析,发现香肠气味指纹的预测值变化与其TVB-N、POV值的变化呈显著的相关关系(R2:0.994 2和0.991 1),从而可以建立一种基于气味指纹识别,快速预测香肠理化品质指标(TVB-N、POV值)的新型检测方法。在对25个样本的模拟测试中,该方法对香肠样品的TVB-N和POV值预测的标准误差分别为0.155 7和0.003 0。

In our study,electronic nose was used to measure the flavor finger-printing changing of Chinese fermented sausage in the process of fermentation( Day 0,2,6,11,15). The relevance between chemometry quality indicator( Total volatile Basic Nitrogen( TVB-N),Peroxide Value( POV)) and flavor finger-printing was studied by the way of principal component analysis( PCA),linear discriminant analysis( LDA) and model analysis which based on partial least squares method( PLS). The sensor array of PEN3 electronic nose showed the most outstanding capability distinguishing the flavor changing during the sausage fermentation under the conditions as follows: temperature35℃,headspace volume 100ml,accumulation time 30min,data-collection interval 55-60s. In addition,the relationship between flavor finger-printing changing and chemometry modification was also identified. Our study provided a novel PLS prediction model for sensitive and precise detection of chemometry indexes of Chinese fermented sausage by flavor finger-printing identification. In a mock test for 25 samples,the standard errors of TVB-N and POV measured by this new method were 0. 1557 and 0. 0030 respectively.