【作者】 秦辉；

【导师】 霍丹群； 张良；

【作者基本信息】 重庆大学 ， 生物医学工程， 2011， 博士

【摘要】 白酒是中国的传统酒精饮料,有着悠久的酿造历史;在漫长的发展过程中形成了独特的酿造工艺和风格特征,被誉为世界六大蒸馏酒之一。近年来,随着白酒工业的迅速发展,白酒假冒产品日益严重,特别是年份酒和原产地标志方面表现的尤为突出。现今对白酒的检测主要是通过色谱、光谱等分析方法,这些方法虽能对白酒进行定量定性分析,但检测其过程需要大型仪器、费时费力,限制了其在流通领域的实地实时检测。针对现今白酒检测技术存在的不足,自主开发设计了用于白酒鉴别可视化仿生嗅觉系统;本文以浓香白酒作为主要研究对象,利用构建的可视化仿生嗅觉系统结合模式识别分析方法,开展了白酒年份和原产地鉴别的技术研究。本文所完成的主要工作如下:(1)设计开发了用于白酒鉴别的可视化传感系统。该系统由气体发生装置、气体检测装置、系统加热装置、系统动力系统和系统控制和数据处理系统组成。气体发生装置主要用于白酒液体的挥发,以便于可视化传感阵列的检测;气体检测装置主要由可视化传感阵列、图像采集系统和光源组成。系统控制和数据处理系统主要由电脑完成,其功能对整个系统的控制和对所得数据的处理。通过集成化的系统实现了采样、图像扫描、数据处理及样品鉴别的快速化。(2)利用GC-MS技术对白酒中12种微量成分含量进行了检测,并利用方差分析(ANOVA)、主成分分析(PCA)、线性判别分析(LDA)等分析方法对色谱数据进行分析,用以筛选能够区分不同品牌和不同年份白酒的特性香味成分。单因素方差分析表明:丁酸乙酯、乙酸乙酯、乳酸乙酯、己酸乙酯、正丙醇、仲丁醇、异丁醇、丁醇、己醇、甲醇等11种香味成分对不同年份有显著性影响。利用模式识别方法对不同年份白酒进行区分。结果表明:对不同年份白酒的识别能力和预测能力均为96.4%。结果表明这些物质能够表达年份酒的特征信息。(3)利用本文构建的可视化仿生嗅觉系统对白酒中醇、酸、酯、醛等香味物质进行了定量检测。结果分析表明,不同浓度的物质具有特异性的响应差谱图,通过响应差谱图可以直观的对不同浓度的物质进行区分。聚类分析结果显示本文所构建的可视化仿生嗅觉系统能对不同浓度的醇、酸、醛、酯等物质进行正确的区分;也显示出该系统具有好的重复性和重现性。此外,还对固态法白酒和用醇、酸、酯等物质和食用酒精勾兑的勾兑白酒进行了区分。通过对比响应后的可视化差谱图可以直观的对这两种不同白酒进行区分。这为可视化系统用于白酒质量控制、白酒年份和原产地鉴别奠定了基础。(4)利用分子对接方法研究了酒体中醇、酸、醛和酯与TPPS4和ZnTpp之间的相互作用机理。并通过得分函数、Polar值和最高得分结果下的构象分析可以得出,TPPS4主要通过分子间的氢键作用与醇、醛、酸和酯类分子发生相互作用。ZnTpp主要通过配位键与醇、酸和酯发生相互作用,而与糠醛间的相互作用主要是通过π-π作用。本文为研究卟啉分子与白酒微量物质间的作用机理,提供了新的参考。(5)利用本文构建的可视化传感系统对不同年份白酒进行了鉴别。通过对比不同年份酒响应差谱图可以发现,不同年份白酒具有特异性响应差谱图,此差谱图代表了该年份白酒的特征性指纹信息。此外,还利用模式识别分析方法对响应差谱图数据进行分析。主成分分析和聚类分析结果表明:可视化系统能对不同年份成品酒进行较好的区分,判别分析也能对不同年份成品酒进行100%的识别,对未知不同年份成品酒的预测能力为93.8%。在不同年份基酒鉴别中,不同年份基酒具有特性的响应差谱图,该图谱代表该年份基酒的特征性指纹图谱。不同年份基酒通过主成分分析和聚类分析被正确区分。聚类分析还表现出一定的规律性,这一规律性有助于对未知年份基酒进行正确的分类。判别分析能对不同年份基酒进行100%的区分和预测。结果表明,本文所构建的可视化系统在白酒年份鉴别检测中具有很强的实用价值;能做作为一种快速、简便的白酒检测方法。(6)利用本文构建的可视化传感系统对不同原产地浓香白酒和清香型白酒进行了鉴别。通过响应差谱图的研究发现,不同原产地白酒具有独特的差谱图,代表了该原产地白酒的特征性指纹信息。主成分分析和聚类分析能对不同原产地白酒进行正确的区分,判别分析能对不同原产地白酒进行100%的区分,对未知样本的预测能力为100%。结果表明本文所构建的可视化传感系统在对不同原产地白酒检测过程中能够对酒体中所含有的特性信息进行有效的捕捉和识别,并且以形象直观的方式表达出来,这为建立不同原产地白酒指纹图谱数据库提供了一个重要的参考。更多还原

【Abstract】 As one Chinese traditional alcohol beverage, Chinese liquor has a long brewing history in China. In the long development of Chinese liquor, its unique brewing process and style characteristics has been formed. There are six top distilled liquors in the world, and Chinese liquor is among of them. In recent years, with the quick development of Chinese liquor industry, Chinese liquor adulteration in regard to aged liquor and geographic origin exists extensively. Chromatography, spectroscopy and other analytical methods can be used to qualitative and quantitative analysis of compounds of Chinese liquor. However, these methods need to use expensive and large-scale instruments, and detection procedure is time-consuming. So they cannot be applied in real-time detection in the field of circulation. Due to deficiency of liquor detection technology, visualization bionic olfactory system was designed to identify liquor. In this research, visualization bionic olfactory system combined with pattern recognition was applied to discriminate aged liquor and to classify Chinese liquor from different geographic origins. The main works were listed as follows:(1) Visualization bionic olfactory system was designed and manufactured to identify Chinese liquor. This system was composed of gas generator equipment, gad detection equipment, heating equipment, dynamical system, and control and date acquisition system. Gas generator equipment was used to generate volatile gas. Gas detection equipment included colorimetric sensor array, image acquisition system and light source. Control and date acquisition system was used to control system and processing data. This integration system realized the sampling, scanning images, processing data and samples of identification quickly.(2) Gas chromatography-mass spectrometry (GC-MS) was used to determine the 12 volatile compounds. One way analysis of variance (ANOVA), principal component analysis (PCA) and linear discriminant analysis (SLDA) were applied to process data obtained by GC-MS and to develop discriminant models. ANOVA showed that ethyl acetate, ethyl butyrate, ethyl lactate, ethyl caproate, 1-propanol, 2-butanol, 1-butanol, isopentanol, isobutanol, 1-hexanol, methanol were most correlated with aged liquor. The data was analyzed by pattern recognition (PCA, LDA). LDA showed that a 96.4% of recognition capability for aged liquor, and the prediction ability of aged liquor was 96.4%. The results suggested that volatile compounds express the characteristic information of aged liquor.(3) Visualization bionic olfactory system was used for quantitative detection of aroma compositions in Chinese liquor, such as alcohols, acids, esters and aldehyde. The results demonstrated that different concentrations of aroma composition provided special color change profiles. Difference between different concentrations could be observed by visual inspection of color change profiles. The digital date library generated was analyzed with Hierarchical cluster analysis (HCA). Using a HCA dendrogram, alcohols, acids, esters and aldehyde were correctly distinguished and this system showed a good reproducibility and repeatability. In addition, Visualization bionic olfactory system was succeeded to identify the adulterated liquor by observing color change profiles. The results showed the potential application of visualization bionic olfactory system for monitoring quality, discriminating aged liquor and geographic origins.(4) The interaction between porphyrin (TPPS4, ZnTpp) and alcohols, acids, esters, furfural was analyzed by molecular docking. Total-score, Crash, Polar and Cscore were used to describe results. Complex conformation showed that the interaction between TPPS4 and alcohols, acids, esters, furfural was via hydrogen bond. The interaction between ZnTpp and alcohols, acids, esters was via coordinate bond, while ZnTpp interacted with furfural was viaπ-πeffect. This paper provided a new method to study the interaction between porphyrin and aroma composition of Chinese liquor.(5) Visualization bionic olfactory system was applied to discriminate aged liquor. Comparing with color change profiles of aged liquor, it was found that different aged liquor provided special color change profiles as unique fingerprints for each one. The digital data library generated was analyzed with pattern recognition, including PCA, HCA and LDA. PCA and HCA analysis showed that visualization bionic olfactory system can distinguish different aged liquor. LDA model showed a 100% of recognition capability and 93.8% prediction ability for aged liquor. This system was also used in discriminating based liquor. The developed visualization bionic olfactory system showed a unique pattern of color change to different aged liquor. The RGB values of the colorimetric sensor array were analyzed by pattern recognition method, including HCA, PCA and LDA. HCA and PCA suggested different based liquor can be easily discriminated by visualization bionic olfactory system. Certain regularity showed in HCA, it was useful to distinguish based liquor of unknown year. LDA analysis showed a 100% recognition capability and prediction ability for based liquor. These results suggested that visualization bionic olfactory system might provide a rapid and practicable method for detection Chinese liquor.(6) Visualization bionic olfactory system was used to characterize and identify strong aroma style Chinese liquor and light aroma style Chinese liquor from different geographic origins. Colorimetric sensor array before and after exposure to Chinese liquor provided unique color change profiles as fingerprint characteristic information for Chinese liquor from different geographic origins. Data analysis was performed by pattern recognition: Hierarchical cluster analysis (HCA), principal component analysis (PCA) and linear discriminant analysis (LDA). LDA model showed a 100% of recognition capability and prediction ability for Chinese liquor. The results showed that visualization bionic olfactory system was able to classify Chinese liquors from different geographic origins. Differentiation between samples was observed by visual inspection of color change profile. This method provides an important reference for constructing of Chinese liquor fingerprint database.更多还原