High-temperature Daqu liquor identification method based on ultraviolet spectrum
Through the random forest model based on ultraviolet spectroscopy combined with stoichiometric methods, the problem of high-temperature Daqu liquor identification was solved, and the rapid and accurate identification and Maillard reaction product prediction were achieved, and the cumbersome and time-consuming identification methods in the existing technology were solved. Problem, providing a highly sensitive and simple detection solution.
Patent Information
- Application Number
- CN202411058159.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-02
- Publication Date
- 2024-11-12
AI Technical Summary
The existing technology is difficult to quickly and economically identify high-temperature Daqu liquor. The traditional method has problems such as cumbersome sample preprocessing, long time-consuming, low resolution ability and strong subjectivity, which makes it difficult to reflect the quality of liquor. Market demand urgently requires a simplicity. Fast detection method.
A random forest model was constructed through spectral signals generated by Maillard reaction products such as pyrazine and furfural to achieve the identification of high-temperature Daqu liquor, and combined with stoichiometric methods to perform rapid identification and Maillard reaction product prediction.
It has achieved rapid identification of high-temperature Daqu liquor and prediction of Maillard reaction products. It has the advantages of high sensitivity, fast response speed, simple operation and wide application range. The accuracy rate reaches 100%, which significantly improves classification accuracy.
Smart Images

Figure CN118937261A_ABST
Abstract
Description
Technical Field
[0001] The invention belongs to the field of liquor identification, and in particular relates to a high-temperature Daqu liquor identification method based on ultraviolet spectroscopy, which can be used for liquor type identification. Background Art
[0002] There are twelve major flavors of Chinese liquor, among which sauce-flavored, mixed-flavored, sesame-flavored and some strong-flavored liquors all partially or completely use high-temperature Daqu technology to increase their elegant and delicate flavor. This is because liquor brewed by high-temperature Daqu technology has the characteristics of long brewing cycle, high cost, elegant and delicate aroma, rich fragrance and full body, so more than 70% of my country's precious liquors use high-temperature Daqu technology. In the market, due to the high quality and high value of such liquors, they are often used by illegal vendors to fill low-grade liquors in high-end bottles to pass off inferior ones as good ones, which not only infringes on the rights and interests of consumers but also seriously disrupts the order of the liquor market. Therefore, it is of great significance to develop a method for rapid identification of high-temperature Daqu liquors.
[0003] At present, the classification and identification of liquor mainly focuses on quality grading, identification of storage years, origin and aroma type. The method used is to rely on large-scale analytical instruments. Although it has the advantages of accurate separation performance, high precision and high sensitivity, it still has the disadvantages of sample pretreatment, time-consuming, laborious and cumbersome measurement steps. The second is to rely on human senses, and professional wine tasters observe, analyze, describe and rate the color, aroma and taste of liquor, and finally make a comprehensive evaluation. However, there are many problems in the traditional liquor evaluation method, such as low resolution, strong subjectivity and large errors, which make it difficult to reflect the quality of liquor and difficult to define its characteristics. Therefore, it is necessary to establish an economical, easy-to-operate and rapid detection method for the identification of high-temperature Daqu liquor to meet market demand. Summary of the invention
[0004] In order to solve the problems raised in the background technology, the present invention provides a method for identifying high-temperature Daqu liquor based on ultraviolet spectroscopy. A random forest model is constructed based on the spectral signals generated by Maillard reaction products such as pyrazine and furfural, and finally the identification of high-temperature Daqu liquor is achieved. The identification method has the characteristics of high sensitivity, fast response speed and wide application range.
[0005] The technical solution adopted by the present invention to solve the above-mentioned problems is: A method for identifying high-temperature Daqu liquor based on ultraviolet spectroscopy comprises the following steps: Step 1: Prepare the test solution Dilute the liquor sample 10 times with an ethanol solution of the same alcohol concentration to obtain a test solution; Step 2: Detect the absorbance change of the solution Place the test solution prepared in step 1 into a two-way quartz cuvette with a light path of 10 mm, place the cuvette into a UV spectrophotometer to scan the UV spectrum of the test solution, set the instrument parameters to wavelength scanning range: 200 400 nm, scanning interval: 1 nm, prepare 10 parallels for each wine sample, scan 10 times, and save the absorbance value in an Excel document; Step 3: Establish RF model to identify high-temperature Daqu liquor The absorbance data of the single sample obtained in step 2 were stacked into a complete two-dimensional data matrix in Excel, converted into csv format and imported into Matlab2018b software. First, the samples were classified according to their divisions. Then, two parameters were optimized, namely, the classification parameter (kkk) and the latent variable (nTree). Finally, the optimized parameters were used to construct a random forest for classification and prediction. The random forest regression algorithm was run in Matlab2018b software to perform regression analysis on the absorbance values measured in step 2 and the concentrations of liquor flavor compounds identified by headspace solid phase microextraction / gas chromatography-mass spectrometry, so as to predict the content of Maillard reaction products. High-temperature Daqu liquor is identified based on the accuracy of the training set and the prediction set. The higher the accuracy, the better the model classifies the samples.
[0006] Preferably, the ethanol solutions with the same alcohol concentration in step 1 are 53% ethanol solution, 52% ethanol solution, 50% ethanol solution, 43% ethanol solution, and 42% ethanol solution, respectively.
[0007] Preferably, the prediction method in step three is specifically as follows: the data matrix obtained in step three and the concentration data of the compound are merged and converted into csv format and then imported into Matlab2018b software, first determining the number of columns N, the number of categories and the number of samples in each category of the variable, and then optimizing three parameters, namely, the classification parameter kkk, the latent variable nTree and the regression parameter M, and finally using the optimized parameters to construct a regression model for prediction, and obtaining the correlation coefficient and root mean square error of the training set and the prediction set. The closer the correlation coefficient is to 1, the smaller the root mean square error is, indicating that the compound has a stronger correlation with the classification of liquor. This identification method can realize the prediction of the components of high-temperature Daqu liquor.
[0008] Preferably, the Maillard reaction products are mainly pyrazine compounds and furfural compounds.
[0009] Preferably, the ultraviolet band related to the Maillard reaction product and having a larger classification contribution weight is 243-321 nm.
[0010] Compared with the prior art, the present invention has the following beneficial effects: The method provided by the present invention realizes rapid identification of high-temperature Daqu liquor and prediction of Maillard reaction products based on ultraviolet spectrum combined with chemometric method, which can be realized by only detecting the ultraviolet spectrum of liquor and combining it with chemometric method. The method does not require sample pretreatment, long-term instrument detection and complex data preprocessing, and has the advantages of high sensitivity, fast detection speed, simple operation and wide application range. BRIEF DESCRIPTION OF THE DRAWINGS
[0011] Figure 1 The following are the UV spectra of different flavor liquors. A) Sauce-flavor liquor, B) Mixed-flavor liquor, and C) Luzhou-flavor liquor.
[0012] Figure 2 The linear correlation diagram of the RF regression model based on UV spectrum and the Maillard reaction products, including A) 2,6-dimethylpyrazine, B) 2,6-dimethyl-5-ethylpyrazine, C) tetramethylpyrazine, D) 2,5-dimethyl-3-pentyl-pyrazine, E) 1,2-dimethoxybenzene, F) 2-acetyl-5-methylfuran, G) 3-methylphenol, H) phenylacetaldehyde, I) α-ethylphenylacetaldehyde, J) 5-methyl-2-furancarboxaldehyde, K) furfural, L) 1-(2-furanyl)-ethanone. DETAILED DESCRIPTION
[0013] The following examples are provided to facilitate a better understanding of the present invention, but are not intended to limit the present invention. The identification of high-temperature Daqu liquor is used as an example, and the technical solution of the present invention is further described in detail in conjunction with the accompanying drawings.
[0014] In the following examples, the liquor sample does not need to be pre-treated, and only needs to be diluted 10 times with an ethanol solution of the same alcohol concentration to obtain a liquor test solution, and then placed in a quartz cuvette with a light path of 10 mm, and a UV spectrophotometer is used to detect and record the spectrum, with a wavelength scanning range of 200 400 nm and a scanning interval of 1 nm. The absorbance data is used to construct a random forest (RF) model to achieve rapid identification of high-temperature Daqu liquor, and the content of compounds in liquor is predicted by the random forest regression (RFR) algorithm.
[0015] In this example, 31 kinds of high-temperature Daqu liquors were selected as test samples (Table 1).
[0016] In this embodiment, the liquor sample does not need to be pre-treated. It only needs to be diluted 10 times with the corresponding ethanol solution of the same alcohol concentration to obtain the liquor test solution, and then placed in a quartz cuvette with an optical path of 10 mm, and the spectrum is detected and recorded using an ultraviolet spectrophotometer. The wavelength scanning range is 200 400 nm, and the scanning interval is 1 nm. Ten parallel samples are prepared for each liquor sample, scanned 10 times, and an Excel document of the absorbance value is saved.
[0017] Figure 1 The UV absorption spectra of different flavor liquors (A: Maotai-flavor liquor; B: Mixed-flavor liquor; C: Luzhou-flavor liquor) are shown. According to the results, the UV absorption spectra of the three flavor liquors only have one absorption band (250-300nm), which is the absorption peak of furfural and other substances. This type of substance is also an important aroma component in liquor. Its structure contains an aldehyde group and a furan ring, two double bonds and a cyclic ether bond, and its chemical properties are active. When the liquor sample is irradiated with continuous wavelength ultraviolet light, the electrons in the unsaturated bonds in furfural absorb energy and transition from low energy levels to high energy levels, forming a strong characteristic absorption at 276nm. Furfural in liquor is mainly produced by Maillard reaction and microbial metabolism. Different flavor liquors are affected by factors such as brewing raw materials, geographical origin, fermentation temperature, fermentation time, and microorganisms, and their furfural content must be different. From the results, we can see that the furfural content is the highest in the sauce-aroma type, followed by the mixed-aroma type, and the lowest in the strong-aroma type. This is related to the brewing process of the sauce-aroma type liquor and is one of the important bases for distinguishing the sauce-aroma type from other aroma types.
[0018] The chemometric random forest algorithm was used to establish a high-temperature Daqu liquor discrimination model. The model performance was optimized by optimizing the latent variables (nTree). Table 2 shows that the training set and prediction set of 31 high-temperature Daqu liquors were completely correctly discriminated, with an accuracy rate of 100%. The classification accuracy of the model training set and prediction set was significantly improved compared with the accuracy of the original UV spectrum. According to the RF results, the main contribution band related to classification is 243-321nm, which overlaps with the absorption band (250-300nm) of Maillard reaction products such as furfural on the UV spectrum, which further illustrates that the Maillard reaction products are markers that affect the classification of liquor.
[0019] In order to further verify the correlation between absorbance and liquor flavor compounds, the absorbance data of high-temperature Daqu liquor and the concentration of liquor flavor compounds identified by HS-SPME / GC-MS were regressed by random forest regression algorithm. Table 3 shows that UV spectrum can accurately predict the content of Maillard reaction products such as pyrazine and furfural, and the regression correlation coefficient is greater than 0.98. Figure 2 It also shows that the true and predicted values of the contents of these 12 Maillard reaction products (Table 3) have a good linear correlation.
[0020] The embodiments of the present invention are described in detail above, but the present invention is not limited to the above embodiments. In addition, all equivalent changes and modifications made without departing from the method and scope of the present invention should be included in the scope of the present invention.
Claims
1. A method for identifying high-temperature Daqu liquor based on ultraviolet spectroscopy, characterized in that: The following steps are involved: Step 1: Prepare the test solution Dilute the liquor sample 10 times with an ethanol solution of the same alcohol concentration to obtain a test solution; Step 2: Detect the absorbance change of the solution Place the test solution prepared in step 1 into a cuvette, place the cuvette into a UV spectrophotometer to scan the UV spectrum of the test solution, set the instrument parameters to wavelength scanning range: 200-400 nm, scanning interval: 1 nm, prepare 10 parallel scans for each wine sample 10 times, and save the absorbance value in an Excel document; Step 3: Establish RF model to identify high-temperature Daqu liquor The absorbance data of the single sample obtained in step 2 is stacked into a complete two-dimensional data matrix in Excel, converted into csv format and imported into Matlab2018b software. First, classification is performed according to the division of the samples. Then, two parameters are optimized, namely, the classification parameter kkk and the latent variable nTree. Finally, the optimized parameters are used to construct a random forest for classification and prediction. The random forest regression algorithm is run in Matlab2018b software to perform regression analysis on the absorbance values measured in step 2 and the concentrations of liquor flavor compounds identified by headspace solid phase microextraction / gas chromatography-mass spectrometry, so as to predict the content of Maillard reaction products. High-temperature Daqu liquor is identified based on the accuracy of the training set and the prediction set. The higher the accuracy, the better the model classifies the samples.
2. The method for identifying high-temperature Daqu liquor based on ultraviolet spectroscopy according to claim 1, characterized in that: The ethanol solutions with the same alcohol concentration in step 1 are 53% ethanol solution, 52% ethanol solution, 50% ethanol solution, 43% ethanol solution, and 42% ethanol solution respectively.
3. The method for identifying high-temperature Daqu liquor based on ultraviolet spectroscopy according to claim 1, characterized in that: The cuvette in step 2 is a two-light quartz cuvette with an optical diameter of 10 mm.
4. The method for identifying high-temperature Daqu liquor based on ultraviolet spectroscopy according to claim 1, characterized in that: The prediction method in step three is specifically as follows: the data matrix obtained in step three and the concentration data of the compound are merged and converted into csv format and then imported into Matlab2018b software. The number of columns N, the number of categories and the number of samples in each category of the variable are first determined. Then three parameters are optimized, namely the classification parameter kkk, the latent variable nTree and the regression parameter M. Finally, the regression model is constructed using the optimized parameters for prediction, and the correlation coefficient and root mean square error of the training set and the prediction set are obtained. The closer the correlation coefficient is to 1, the smaller the root mean square error is, indicating that the compound has a stronger correlation with the classification of liquor. This identification method can realize the prediction of the components of high-temperature Daqu liquor.
5. A method for identifying high-temperature Daqu liquor based on ultraviolet spectroscopy as claimed in claim 1 or 4, characterized in that: The Maillard reaction products are mainly pyrazine compounds and furfural compounds.
6. The method for identifying high-temperature Daqu liquor based on ultraviolet spectroscopy according to claim 1, characterized in that: The ultraviolet band related to the Maillard reaction product and having a larger classification contribution weight is 243-321 nm.
Citation Information
Patent Citations
Method for rapidly determining content of hydroxymethylfurfural in high fructose corn syrup
CN110658146A
White spirit identification method based on ultraviolet absorption spectrum
CN112858201A
Baijiu brand identification method
CN114279991A
High-temperature Daqu liquor identification and compound prediction method based on Maillard reaction product fluorescence signal analysis
CN116403661A
Non-contact measurement system for alcohol concentration
CN117451652A
Cited By
Traditional Chinese medicinal material rapid detection system and method based on ultraviolet spectrum and machine learning
CN121558660A
