基于分子光谱技术的浓香型白酒基酒品质检测研究

发布时间：2018-04-08 23:00

本文选题：浓香型白酒基酒　切入点：分子光谱技术　出处：《江苏大学》2017年硕士论文

【摘要】：白酒作为六大蒸馏酒之一,是我国的特色酒类,已逐渐成为人们日常生活和交际过程中重要的食品。现有的白酒质量等级评定方法主要包括人工感官评定和常规理化分析,这些方法因主观性强、方法繁琐、时效性差和成本昂贵等不足而无法实现实时快速检测。近年来,光谱检测技术因其操作简便、检测快速、灵敏度高和重现性好等优点,在食品品质检测领域得到广泛应用。本研究主要采用了四种分子光谱技术(近红外光谱、中红外光谱、紫外-可见光谱、拉曼光谱)对不同等级的浓香型白酒基酒及其主要酯类化合物(己酸乙酯、乳酸乙酯、丁酸乙酯、乙酸乙酯)含量进行检测,旨在建立一套较为完整的浓香型白酒基酒质量等级的无损快速检测体系,具体研究内容和相关结论如下:(1)浓香型白酒基酒样本的收集及感官等级评定。收集了不同车间、不同班组的白酒基酒样本共75个,经车间工人初步等级划分后获得一级酒、二级酒和三级酒。经初步划分的白酒基酒样本再由专业的感官评定人员进行严格的等级评定,最终获得一级好酒样本10个,一级差酒样本22个,二级好酒样本10个,二级差酒样本14个,三级酒样本19个。(2)不同等级浓香型白酒基酒样本的快速判别。研究采集了不同等级白酒基酒样本的四种光谱信息,并分别结合主成分分析(PCA)、线性判别分析(LDA)、反向传播人工神经网络分析(BPANN)三种化学计量学方法建立不同等级白酒基酒的等级判别模型,并比较各模型效果。结果表明,四种光谱技术均能对不同等级样本进行区分,模型效果较好。其中,中红外光谱的LDA和BPANN模型的训练集和测试集识别率均达到100%,模型效果最好,近红外、紫外和拉曼光谱虽未达到100%,但识别率均较高,也能达到判别目的,能够对不同等级浓香型白酒基酒进行判别区分。(3)浓香型白酒基酒中主要酯类化合物含量的快速检测。研究采用气相色谱法测定白酒基酒样本中各酯类化合物的含量,所测得的酯类含量作为参考值分别与四种光谱建立联合区间偏最小二乘(SiPLS)定量模型。结果表明,四种光谱技术均能实现白酒基酒中主要酯类化合物含量的快速测定,其中中红外光谱技术对白酒基酒中己酸乙酯、乳酸乙酯、乙酸乙酯三种酯类的定量模型效果最好,其训练集和测试集的相关系数分别为0.9866和0.9847、0.9942和0.9937、0.9908和0.9852;近红外光谱技术对丁酸乙酯的定量模型效果最好,其训练集和测试集的相关系数为0.9262和0.9707,各模型均能满足白酒生产过程中各酯类化合物含量的检测要求。本研究实现了不同等级浓香型白酒基酒的准确判别,同时所建的酯类化合物快速定量模型的效果较好,能满足白酒生产中的分析检测要求,为浓香型白酒基酒等级的快速判别提供一种客观而准确的分析方法,有效提升了白酒的智能化生产水平。
[Abstract]:As one of the six distilled wines, liquor is the characteristic liquor of our country, and it has gradually become an important food in people's daily life and communication.The existing methods of liquor quality grade evaluation mainly include artificial sensory evaluation and routine physical and chemical analysis. These methods can not realize real-time and rapid detection because of their strong subjectivity, cumbersome methods, poor timeliness and high cost.In recent years, the spectral detection technology has been widely used in the field of food quality detection because of its advantages of simple operation, rapid detection, high sensitivity and good reproducibility.In this study, four kinds of molecular spectroscopic techniques (near infrared spectrum, middle infrared spectrum, UV-Vis spectrum, Raman spectrum) were applied to Luzhou-flavor liquor and its main ester compounds (ethyl caproate, ethyl lactate).The content of ethyl butyrate and ethyl acetate was determined in order to establish a complete nondestructive and rapid detection system for the quality grade of Luzhou-flavor liquor.The specific research contents and related conclusions are as follows: 1) the collection and sensory grade evaluation of Luzhou-flavor liquor base liquor.A total of 75 samples of liquor base liquor were collected from different workshops and groups. The first grade, the second grade and the third grade liquor were obtained after the preliminary grading of the workers in the workshop.The samples of liquor base liquor were evaluated strictly by professional sensory assessors. Finally, 10 samples of first grade good wine, 22 samples of first grade good wine, 10 samples of second grade good liquor and 14 samples of second grade good liquor were obtained.The rapid discrimination of base liquor samples with different grades of Luzhou-flavor liquor was obtained from 19 samples of third grade liquor.Four kinds of spectral information of liquor samples of different grades were collected.Combined with three chemometrics methods, principal component analysis (PCA), linear discriminant analysis (LDAA) and backpropagation artificial neural network (Ann), three chemometrics methods were used to establish the grade discriminant models of liquor base liquor of different grades, and the effects of each model were compared.The results show that the four spectral techniques can distinguish the samples of different grades and the model is effective.Among them, the training set and test set of the LDA and BPANN models of the mid-infrared spectrum have the highest recognition rate, and the model has the best effect. Although the near infrared, ultraviolet and Raman spectra are not up to 100, the recognition rate is high, and the recognition rate can also reach the purpose of discrimination.The content of main ester compounds in Luzhou-flavor liquor could be detected quickly by discriminating and distinguishing the base liquor of Luzhou-flavor liquor.The determination of ester compounds in liquor samples by gas chromatography was studied. The measured esters were used as reference values to establish a combined interval partial least square siprs quantitative model with four kinds of spectra.The results showed that the four spectral techniques could be used to determine the contents of main ester compounds in liquor, and the content of ethyl caproate and ethyl lactate in liquor was determined by mid-infrared spectroscopy.The correlation coefficients between the training set and the test set were 0.9262 and 0.9707 respectively. Each model could meet the requirements of the determination of the contents of ester compounds in liquor production.This study realized the accurate discrimination of Luzhou-flavor liquor base liquor with different grades, and established the fast quantitative model of ester compounds, which can meet the requirements of analysis and detection in liquor production.It provides an objective and accurate analysis method for the fast discrimination of liquor grade of Luzhou-flavor liquor, and effectively improves the level of intelligent production of liquor.
【学位授予单位】：江苏大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：O657.3;TS262.3

【参考文献】