白鲢鱼糜低气味本底模型的构建研究

鱼糜制品(如火锅鱼丸)的风味是消费者关心的质量属性之一,而关键气味活性物质的吸附释放规律并不明确。现有气味研究主要在配置溶液中进行,与真实的气味活性物质—固态鱼糜之间的相互作用存在一定差异,因此,基于固态鱼糜进行气味研究是十分必要的,其关键在于一个无气味或低气味的鱼糜本底模型,从而可进一步研究各气味成分与鱼糜本底模型的互作关系。本研究考察了8种不同漂洗介质对鱼糜本底模型气味残留的影响。结果表明,白鲢(Hypophthalmichthys molitrix)鱼糜经SPME-GC-MS共检出65种挥发性物质,气味活性物质(OAV>1)有18种;经8种漂洗介质处理后,鱼糜样品中分别含有6、8、7、9、6、12、9和9种气味活性物质,挥发性气味物质的残留率依次为(0.380±0.120)%、(0.610±0.086)%、(0.280±0.033)%、(0.480±0.037)%、(0.150± 0.018)%、(4.330±0.160)%、(18.680±0.081)%和(0.490±0.003)%。综合SPME-GC-MS、电子鼻和感官评价结果比较,1% NaCl (W/W) + 1% Na2CO3 (W/W) + 4.0% C2H5OH (V/W)漂洗介质处理后,白鲢鱼糜的挥发性气味物质残留少,总含量降低为(6.57±0.77) μg/kg,17种气味活性物质的OAV<1,仅壬醛的OAV为1.34±0.05,可构建出低气味的鱼糜本底模型。     

基于电子鼻和高斯过程的秸秆固态发酵过程监测技术

基于电子鼻监控数据,建立基于高斯过程的状态监控分类器,实现对秸秆饲料固态发酵过程的有效监测。秸秆饲料固态发酵过程实验周期为7 d,每隔24 h利用电子鼻系统对发酵气体监测数据进行采集。该发酵实验共分20批次,其中10批次实验数据用来训练高斯过程分类器,其余10批次实验数据用来测试所训练分类器的性能。实验结果表明所采用电子鼻系统可以对秸秆饲料固态发酵过程状态进行有效监控。将所训练高斯过程分类器与支持向量机、神经网络分类器进行比较表明,基于高斯过程分类器的正确率为100%,高于基于支持向量机、神经网络分类器的正确率85.71%、94.29%,能够更好地实现对秸秆饲料固态发酵过程的监测。     

Scent profiling of Cymbidium ensifolium by electronic nose

The performance of electronic nose (E-nose) for Chinese Cymbidium scent profiling has been evaluated. Changes in scent profiles of two Cymbidium ensifolium cultivars have been monitored at different flowering stages (initial flowering, full flowering, and terminal flowering) and different times combined with two gas collecting devices. Samples were collected by static headspace (SHS) method. How E-nose can be used for pattern recognition and for studying the releasing of flower scent were proposed. Data obtained were subjected to principal component analysis (PCA) and discriminant function analysis (DFA). PCA was performed on the initially instrumental data to explore the structure of each data set and such result showed that the sensory data contained information related to the cultivar and to time spots. DFA was performed to improve the results, leading to clear separations between the sample groups. Gas collecting device did not seriously affect the result of PCA and DFA. Relative aroma intensity (RAI) was proposed as an alternative concept to compare scent intensity between samples on different time points. These results demonstrate the potential application of the E-nose to evaluate the scent profile of flower.     

基于空载数据的鉴别食醋电子鼻信号漂移校正方法

由于传感器老化,环境温度等因素,电子鼻信号的漂移是不可避免的,且严重降低电子鼻的长期稳健检测能力。为了实现电子鼻对6种食醋样品的长期稳健检测,该文提出了一种基于空载数据的小波包分解系数的漂移递归校正方法。通过小波包对电子鼻空载数据的分解,给出空载阈值函数(no-load threshold function,NLTF),然后将NLTF转换为适合样本数据的样本阈值函数(samplethresholdfunction,STF)。在获得的STF基础上,构建样本检测数据小波包分解系数的校正函数。借助于所构建的样本测试数据的校正函数,对6种食醋样品的电子鼻数据进行漂移校正。同时,运用"样本测量时间窗口(sample measurement time window,SMTW)"的概念,实现电子鼻数据的递归校正,进而建立了可实现长期稳健检测的递归鉴别模型。针对6种食醋样品,进行了为期16个月的间歇式测试。当SMTW选为4个月的测试样本及每次递推前移1个月样本数据时,建立的基于递归校正的Fisher判别分析(Fisher discriminant analysis,FDA)模型可完全实现6种食醋样品的长期稳健鉴别,正确鉴别率达到100%,使紧随SMTW后1个月内的测试样本能得到准确鉴别。该校正方法能够有效的去除漂移并且实现了电子鼻的长期稳健检测。     

Analysis of the Volatile Chemicals of Longjing Tea from Different Production Locations Using Electronic Nose

Electronic nose has been applied in analysis of food products by monitoring their flavours.But it gives an overall response to a mixture of volatile components.In this work,multianalysis and chemometrics technique were employed to analyse the E-nose and GC-MS data of Longjing tea produced in different locations in Zhejiang province,China.A good discrimination of Longjing tea samples was obtained by E-nose according to the producing area,and a total of 38 compounds were commonly identified from the GC-MS data.Each individual volatile component was related to E-nose response using PLSR and the significance of each component was evaluated by the coefficient R2.Results indicated that isoamyl isovalerate,cis-3-hexenyl hexanoate,cadinene,phenylethylalcohol and linalool played an important role in response on E-nose sensors.It was possible to give a complementary information concerning the individual chemicals interacts with the sensors of E-nose instead of the total volatile components.