Method for optimizing fragrance quality of thick broad-bean sauce based on electronic nose technology and application of method

Abstract

The invention provides a thick broad-bean sauce aroma quality optimization method based on an electronic nose technology and application thereof, and relates to the technical field of food processing. According to the method, five representative pepper varieties are selected and are respectively combined with the shaded bean halves and the sun-dried bean halves to prepare a thick broad-bean sauce sample, an electronic nose technology is combined with multivariate statistical analysis to evaluate the overall aroma profile of the thick broad-bean sauce sample, and the correlation between sensor response and sample physicochemical indexes is discussed. Load analysis shows that W1S and W5S sensors are main contributors causing sample aroma difference. In addition, the response value of the electronic nose sensor has significant correlation with part of physical and chemical quality indexes. The key effect of raw material selection and combination on thick broad-bean sauce aroma formation is disclosed, the application potential of an electronic nose technology in fast evaluation and quality control of the flavor of the thick broad-bean sauce is proved, and a scientific basis is provided for product optimization and characteristic development.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of food processing, and specifically relates to a fermented bean paste aroma quality optimization method based on electronic nose technology and application thereof. BACKGROUND

[0002] Fermented bean paste is a traditional compound condiment made of chili peppers and broad beans as main raw materials through microbial fermentation. It is praised as the "soul of Sichuan cuisine" due to its bright red color, rich sauce aroma and unique flavor. It plays a key role in cooking by adding aroma, color and flavor. The sensory quality of fermented bean paste, especially its complex aroma characteristics, is the core indicator for evaluating its quality and determining consumer acceptance. These aromas are mainly derived from various volatile organic compounds formed during fermentation based on the characteristics of different raw materials. There are many factors that affect the final flavor quality of fermented bean paste, among which the selection of raw materials is crucial. First, the diversity of chili pepper varieties directly determines the basic flavor skeleton of fermented bean paste. Different chili pepper varieties have natural differences in chemical components such as capsaicinoids (which impart spiciness), pigments (which impart color), sugars, amino acids, and aromatic substance precursors. These differences not only directly affect the sensory characteristics of fermented bean paste, but also provide different substrates for microbial growth and metabolism during fermentation, thereby indirectly regulating the generation and accumulation of characteristic aroma substances. Therefore, exploring the influence of different chili pepper varieties on the aroma characteristics of fermented bean paste has important theoretical and practical value for optimizing raw material ratios, improving product flavor quality, and developing fermented bean paste products with regional characteristics. Second, as another main raw material of fermented bean paste, the preparation process of broad beans, i.e. "bean paste", also has a profound impact on the flavor of the final product. In the traditional production of fermented bean paste, there are different processing methods such as "shaded bean paste" and "sun-dried bean paste". The physical and chemical changes and microbial flora succession experienced during the preparation of these two types of bean paste are different, which may lead to differences in protein degradation, sugar transformation, and the types and contents of flavor precursors, thereby significantly affecting the formation pathways and final profiles of aroma substances during subsequent fermentation with chili peppers. However, there is still a lack of systematic comparative studies on the contribution of different bean paste raw materials to the aroma of fermented bean paste.

[0003] Electronic nose technology has become an important analytical tool for evaluating flavor, monitoring quality, and identifying origin in the field of food science due to its ability to simulate the biological olfactory system for overall recognition and differentiation of complex odors. It can quickly, non-destructively and objectively reflect the aroma differences between samples by capturing the volatile compound fingerprint of the sample headspace using a sensor array, providing effective technical support for in-depth analysis of the influence of different factors on the flavor of fermented bean paste. Currently, there have been reports on the influence of chili pepper varieties on the flavor of fermented bean paste, but there are few studies on the synergistic effect of different combinations of chili pepper varieties and different bean paste on the aroma components of fermented bean paste. SUMMARY

[0004] Therefore, the present application aims to provide a fermented bean paste aroma quality optimization method based on electronic nose technology, which can quickly and objectively evaluate the synergistic effect of different pepper varieties and fermented bean paste raw materials on aroma characteristics, and realize precise screening and ratio optimization of raw materials.

[0005] Another object of the present application is to provide an application of the fermented bean paste aroma quality optimization method in screening and / or preparing target aroma characteristic fermented bean paste products.

[0006] In order to achieve the above-mentioned application purposes, the present application provides the following technical solutions:

[0007] The present application provides a fermented bean paste aroma quality optimization method based on electronic nose technology, which comprises the following steps:

[0008] Preparation of fermented bean paste sample set: select fermented bean paste raw materials of different pepper varieties and different treatment methods, and prepare fermented bean paste sample set by combining them in proportion;

[0009] Electronic nose detection: use an electronic nose system to detect the aroma components of the fermented bean paste sample set, and obtain sensor array response data;

[0010] Multivariate statistical analysis: perform principal component analysis, cluster analysis and loading analysis on the electronic nose data to identify key aroma contribution sensors;

[0011] Correlation analysis: perform correlation analysis on the sensor response values and the physicochemical indexes of the samples to establish an aroma-quality correlation model;

[0012] Optimization screening: according to the aroma characteristics and physicochemical indexes of fermented bean paste, the optimal combination of pepper varieties and fermented bean paste raw materials is screened out;

[0013] Product preparation: based on the screening results, fermented bean paste products with target aroma characteristics are prepared.

[0014] Preferably, the pepper varieties include one or more of Hongguan No. 3, Hongguan No. 4, Hongguan No. 5, Chuanteng No. 6 and Chuanteng No. 10.

[0015] Preferably, the fermented bean paste raw materials of different treatment methods include shaded fermented bean paste and / or sun-dried fermented bean paste.

[0016] More preferably, the shaded fermented bean paste refers to fermented bean paste prepared in a cool and dark environment; and the sun-dried fermented bean paste refers to fermented bean paste prepared by sunlight.

[0017] Preferably, the electronic nose system comprises a metal sensor, response characteristics of the metal sensor to specific gases, an aroma characteristic detection system, response characteristics of the metal sensor to physicochemical indexes, a physicochemical index detection system and a data processing and analysis system.

[0018] More preferably, the model of the metal sensor comprises one or more of W1C, W5C, W3C, W6S, W5S, W1S, W1W, W2S, W2W and W3S.

[0019] Preferably, the aroma feature detection method comprises the following steps: cleaning the electronic nose for 100-150 s, detecting for 60-70 s, and taking data for 56-58 s for analysis.

[0020] Preferably, the physicochemical index comprises one or more of alcohol dehydrogenase content, carotenoid content, total amino acid content, lactic acid content, total pectin content, flavonoid content, pH value, reducing sugar content, total phenol content, gamma-aminobutyric acid content, moisture content, capsaicin content, color value, soluble solid content and nitrite content.

[0021] Preferably, the preparation method of the set of soybean paste samples comprises the following steps: mixing different processing methods of soybean paste raw materials with different pepper varieties, and fermenting to obtain a set of underground soybean paste samples and / or a set of sun-dried soybean paste samples.

[0022] The application also provides an application of the soybean paste aroma quality optimization method in screening and / or preparing a target aroma characteristic soybean paste product.

[0023] Compared with the prior art, the application has the following beneficial effects:

[0024] (1) The application provides a soybean paste aroma quality optimization method based on electronic nose technology. A plurality of representative pepper varieties are selected, and underground soybean paste and sun-dried soybean paste are combined as main fermentation substrates to prepare different combinations of soybean paste samples. By using electronic nose technology, the influence of different combinations of pepper varieties and soybean paste raw materials on the overall aroma profile of soybean paste is systematically analyzed, and the internal law is revealed by combining multivariate statistical analysis methods. The electronic nose can effectively distinguish soybean paste samples of different processing combinations. Load analysis shows that W1S (sensitive to methane) and W5S (sensitive to alkanes and aromatic components) sensors are the main contributors to the aroma difference of the samples. Principal component analysis and cluster analysis both show that both pepper varieties and soybean paste raw material types have a significant influence on the aroma profile of soybean paste. Among them, the aroma characteristics of samples prepared from the HG3 variety are the most unique. It is worth noting that the processing method of the soybean paste raw material has a greater influence on the aroma profile than the basic difference between pepper varieties, resulting in higher aroma similarity between samples of different pepper varieties prepared by the same processing method (especially sun-dried soybean paste).

[0025] (2)The present application first discloses that there is a significant correlation between the electronic nose aroma profile of soybean paste and a plurality of key physicochemical indexes, which deeply reflect the fermentation depth and nutritional value of the product. Although the traditional physicochemical index detection method can accurately represent the internal quality of the soybean paste, it has limitations such as complicated process and long time-consuming; while the electronic nose technology has the advantages of rapid, non-destructive and objective acquisition of sensor response value. The present application establishes a correlation model between the sensor response signal and the physicochemical index, and for the first time realizes the correlation between the fast and non-destructive electronic nose measurement results and the traditional time-consuming but quality-essential physicochemical indexes. This breakthrough makes it possible to indirectly and efficiently predict or evaluate the internal chemical quality of soybean paste by using electronic nose. Based on the correlation analysis of electronic nose sensor response and physicochemical quality index, it is found that W1C, W3C and W5C sensor groups show extremely significant (p<0.001) or significant (p<0.01, p<0.05) positive correlation with ADH, total amino acid content and total flavonoids. This indicates that the increase in the response of these sensors may be related to the increase in the content of these nutritional or flavor precursor substances; and show extremely significant (p<0.001) or significant (p<0.01, p<0.05) negative correlation with reducing sugar, pH, total phenol and GABA; and also show significant negative correlation with carotenoids. W1W, W2W, W6S, W1S, W2S and W5S sensor groups: the response mode of this group of sensors shows opposite trend to that of W1C group. They show extremely significant (p<0.001) or significant (p<0.01, p<0.05) negative correlation with ADH, total amino acid content and total flavonoids; show extremely significant (p<0.001) or significant (p<0.01, p<0.05) positive correlation with reducing sugar, pH, total phenol and GABA; also show significant positive correlation with carotenoids; and sensor W5S also shows significant positive correlation with sugar acid ratio and soluble solids. The relationship between other physicochemical indexes and sensors: moisture content shows certain positive correlation with W1C, W3C, W5C and ADH, total amino acid content and total flavonoids, while shows negative correlation with W1W, W2W, W6S, W1S, W2S and reducing sugar, total phenol and GABA; sugar acid ratio and soluble solids show significant positive correlation with W5S, W1S and W2S, while show negative correlation with W1C and W3C. The core value of the present application lies in the conversion of the invisible internal quality attributes into visible and quantifiable electronic nose signals, which provides a new technical approach for real-time quality evaluation and control of soybean paste production.

[0026] (3)The present application aims to clarify the synergistic mechanism of pepper varieties and soybean paste raw materials on the aroma characteristics of soybean paste, and to provide scientific basis and technical reference for the optimization of raw materials, improvement of process and development of characteristic flavor products in soybean paste production, which has positive significance for improving the quality and added value of traditional fermented foods. BRIEF DESCRIPTION OF DRAWINGS

[0027] Figure 1 Load analysis chart for electronic nose 10 sensor response values;

[0028] Figure 2 Radar chart for different bean paste electronic nose odor response;

[0029] Figure 3 Principal component analysis chart for different bean paste aroma components;

[0030] Figure 4 Cluster analysis chart for different bean paste aroma components;

[0031] Figure 5 Electronic nose data and physicochemical quality correlation analysis results. DETAILED DESCRIPTION

[0032] The application provides a bean paste aroma quality optimization method based on electronic nose technology, and the method preferably comprises the following steps: bean paste sample set preparation: selecting bean paste raw materials of different pepper varieties and different treatment methods, and preparing a bean paste sample set by proportioning combination; electronic nose detection: using an electronic nose system to detect the aroma components of the bean paste sample set, and obtaining sensor array response data; multivariate statistical analysis: performing principal component analysis, cluster analysis and load analysis on the electronic nose data, and identifying key aroma contribution sensors; correlation analysis: performing correlation analysis on the sensor response values and the physicochemical indexes of the samples, and establishing an aroma-quality correlation model; optimization screening: screening the optimal pepper variety and bean paste raw material combination according to the aroma characteristics and physicochemical indexes of the bean paste; product preparation: preparing a bean paste product with target aroma characteristics based on the screening results.

[0033] In the application, the pepper varieties are preferably but not limited to Hongguan No. 3, Hongguan No. 4, Hongguan No. 5, Chuanteng No. 6 and Chuanteng No. 10. The pepper materials used in the application are derived from the varieties independently bred by the Sichuan Academy of Agricultural Sciences Institute of Horticulture.

[0034] In the present application, the different processing methods of the bean paste raw material preferably include shadow bean paste and / or sun-dried bean paste. The shadow bean paste and / or sun-dried bean paste are bean paste raw materials obtained by different processing methods known in the art. In the present application, the shadow bean paste refers to bean paste prepared by fermentation in a cool environment; and the sun-dried bean paste refers to bean paste obtained by sun-drying. In the present application, as an optional embodiment, the preparation method of the shadow bean paste and / or sun-dried bean paste is as follows: vetch bean paste is soaked at room temperature for 6-8 h, and then drained; the vetch bean paste is mixed with flour at a mass ratio of 8.5:1.5, inoculated with 0.3% Aspergillus oryzae koji, and then prepared in a koji room at 30°C for 48 h to obtain vetch koji; and Aspergillus oryzae koji, salt and water are uniformly mixed at a mass ratio of 3:1:1 and then loaded into a fermentation tank. The shadow bean paste is covered with a food-grade plastic film, pressed with salt, and sealed for natural fermentation at room temperature in a cool environment for 8 months; and the sun-dried bean paste is transferred into a sun-drying jar for artificial sun-drying for 4 months after natural fermentation at room temperature in the fermentation tank for 8 months.

[0035] In the present application, the electronic nose system preferably includes a metal sensor, response characteristics of the metal sensor to specific gases, an aroma feature detection system, response characteristics of the metal sensor to physicochemical indicators, a physicochemical indicator detection system, and a data processing and analysis system. In the present application, the model of the metal sensor preferably includes one or more of W1C, W5C, W3C, W6S, W5S, W1S, W1W, W2S, W2W and W3S. Each metal sensor has response characteristics to specific gases.

[0036] In the present application, the aroma feature detection method preferably includes the following steps: cleaning the electronic nose for 100-150 s, detecting for 60-70 s, and taking data for 56-58 s for analysis; and more preferably, cleaning the electronic nose for 120 s, detecting for 70 s, and taking data for 56-58 s for analysis.

[0037] In the present application, the physicochemical indicators preferably include one or more of ethanol dehydrogenase content, carotenoid content, total amino acid content, lactic acid content, total pectin content, flavonoid content, pH value, reducing sugar content, total phenol content, γ-aminobutyric acid content, moisture content, capsaicin content, color value, soluble solid content, and nitrite content.

[0038] In the present application, the preparation method of the sample set of the bean paste includes the following steps: fully mixing the chili pepper variety and the bean paste raw material with different treatment methods, and fermenting to obtain the sample set of the sun-dried bean paste and / or the sample set of the dried bean paste. In the present application, as a preferred embodiment, the preparation method of the sample set of the bean paste includes the following steps: fully mixing the chili pepper, the sun-dried mature sun-dried bean paste / dried bean paste and salt according to the mass ratio of 5-8:2:1, and then fermenting at room temperature of 27-30 DEG C. The clog is manually turned over every day, and the sun-drying and night exposure are performed. The sun-drying is performed for 2-3 months to ferment into the traditional sun-dried / parched Pi County bean paste. The mass ratio of the chili pepper, the sun-dried mature sun-dried bean paste and salt is more preferably 7:2:1, and the mass ratio of the chili pepper, the sun-dried mature dried bean paste and salt is more preferably 7:2:1.

[0039] The present application also provides an application of the aroma quality optimization method of the bean paste in screening and / or preparing a target aroma characteristic bean paste product.

[0040] The technical solutions provided by the present application will be described in detail below in combination with the embodiments, but they should not be understood as limitations to the protection scope of the present application.

[0041] In the following examples, the experimental methods are conventional methods unless otherwise specified. The test materials used in the following examples are commercially available products unless otherwise specified.

[0042] Example 1

[0043] An aroma quality optimization method of bean paste based on electronic nose technology includes the following steps:

[0044] Preparation of the sample set of the bean paste: different chili pepper varieties are de-tied, washed, broken, and then put into a jar. The sun-dried mature sun-dried bean paste / dried bean paste and salt are added according to the mass ratio of 7:2:1, and then fully mixed and fermented at room temperature of 27-30 DEG C. The clog is manually turned over every day, and the sun-drying and night exposure are performed. The sun-drying is performed for 2 months to ferment into the traditional sun-dried / parched Pi County bean paste.

[0045] Electronic nose detection: the aroma components of the sample set of the bean paste are detected by using an electronic nose system to obtain sensor array response data. The metal sensor array equipped in the system and its response characteristics to specific gases are shown in Table 1. The aroma component determination steps are as follows: the aroma flavor is detected by using the electronic nose, 5g of the bean paste is divided into 50mL centrifuge tubes, and the tube opening is sealed by using a sealing film. The detection is performed according to the instructions of the electronic nose, the cleaning time is 120s, the detection time is 70s, 55s is basically stable, and the data of 56-58s is taken for analysis. Each sample is determined in parallel for 3 times.

[0046] Multivariate statistical analysis: Load analysis and principal component analysis were performed using the Winmuster system of the electronic nose system; response radar chart was drawn using Origin 2024 software; variance correlation analysis was performed using IBM SPSS Statistics 23 software, and principal component PCA and cluster analysis were performed using cloud tools in the Meivian cloud platform.

[0047] Correlation analysis: The sensor response value was correlated with the physicochemical index of the sample to establish an aroma-quality correlation model; wherein the physicochemical index included the content of alcohol dehydrogenase, carotenoids, total amino acids, lactic acid, total pectin, flavonoids, pH, reducing sugar, total phenol, gamma-aminobutyric acid, water content, capsaicin, color value, soluble solids content, and nitrite content.

[0048] Optimization screening: According to the aroma characteristics and physicochemical indexes of the bean paste, the optimal combination of pepper varieties and bean paste raw materials was screened.

[0049] Product preparation: Based on the screening results, bean paste products with target aroma characteristics were prepared.

[0050] Example 2

[0051] According to the method of Example 1, five representative pepper varieties were combined with Yin bean paste and sun-dried bean paste to prepare a set of bean paste samples, and the overall aroma profile was evaluated using electronic nose technology combined with multivariate statistical analysis, and the correlation between sensor response and sample physicochemical index was explored. The steps are as follows:

[0052] 1. Test materials

[0053] Aspergillus oryzae 3.042 was purchased from Chengdu Qufu Biotechnology Co., Ltd., and mung bean paste, flour, and salt were purchased from the market. The pepper materials used in the test were sourced from the Red Crown No. 3 (HG3), Red Crown No. 4 (HG4), Red Crown No. 5 (HG5), Chuanteng No. 6 (CT6), and Chuanteng No. 10 (CT10) independently developed by the Sichuan Academy of Agricultural Sciences Institute of Horticulture.

[0054] 2. Test method

[0055] Sample preparation: The test pepper varieties were sown on November 15, 2021. The seedlings were planted in a plastic greenhouse on March 25, 2022. The test plot was 80 cm wide and 4 m long. The planting density was 70 cm apart and 40 cm apart, with 20 plants per plot. During the entire growth period, uniform field management measures were used in each test plot to ensure consistency of test conditions. When the peppers reached the red ripe stage (mid-July 2022), healthy fruits with uniform shape and size and no obvious diseases and pests were selected for collection.

[0056] Bean paste making process: Soaked broad bean paste at room temperature for 6-8 hours, drained the water, mixed broad bean paste with flour at a ratio of 8.5:1.5 (mass ratio), inoculated with 0.3% Aspergillus oryzae koji, and fermented at 30°C for 48 hours to obtain broad bean koji; evenly mix the koji, salt and water at a ratio of 3:1:1 and load into the fermentation tank. The Y bean paste (Y) is covered with food-grade plastic film, salt-pressed and sealed, and naturally fermented at room temperature in a cool place for 8 months; the sun-dried bean paste (S) is naturally fermented in the fermentation tank at room temperature for 8 months, then transferred to a sun-drying jar for artificial sun-drying for 4 months.

[0057] Using the above pepper varieties, after removing the stems, washing, crushing, and adding into the jar, according to the ratio of 7:2:1, respectively adding sun-dried mature Y / S bean paste and salt, and fully mixing, then fermenting at room temperature at 27-30°C. Artificially turn the pestle every day, and expose to sunlight during the day and to the moonlight at night. After 2 months of sun-drying, the traditional Y / S bean paste is fermented. The prepared bean paste samples are HG3+S / Y (Hongguan No. 3 + Y / S bean paste), HG4+S / Y (Y / S bean paste + Y / S bean paste), HG5+S / Y (Hongguan No. 5 + Y / S bean paste), CT6+S / Y (Chuanteng No. 6 + Y / S bean paste), and CT10+S / Y (Chuanteng No. 10 + Y / S bean paste), which are used for subsequent electronic nose flavor detection. The PEN3 portable electronic nose system produced by AIRSENSE Company of Germany is used for flavor analysis of the samples. The metal sensor array provided by the system and its response characteristics to specific gases are shown in Table 1.

[0058] Table 1: Electronic nose sensor model and response substance

[0059] Sensor model Function description Reference substance Reference substance detection limit W1C Sensitive to aromatic components Toluene 10 W5C High sensitivity, sensitive to nitrogen oxides NO2 1 W3C Sensitive to ammonia, aromatic components Benzene 10 W6S Mainly selective to hydrogen [H2] 100 W5S Alkanes, aromatic components Propane 1 W1S Sensitive to methane CH3 100 W1W Sensitive to sulfides [H2S] 1 W2S Sensitive to ethanol CO 100 W2W Aromatic components, sensitive to organic sulfides [H2S] 1 W3S Sensitive to alkanes CH4 10

[0060] Aroma component determination: The aroma flavor was detected by electronic nose, 5g of bean paste was divided into 50mL centrifuge tubes, the tube opening was sealed with sealing film, each treatment was repeated 3 times. The detection was carried out according to the electronic nose instruction, the cleaning time was 120s, the detection time was 70s, the basic stability was 55s, and the data of 56-58s was analyzed, each sample was determined in triplicate.

[0061] Physicochemical quality determination: alcohol dehydrogenase (ADH) content determination used CheKine TM The alcohol dehydrogenase activity detection kit (Wuhan Yakeyin Biotechnology Co., Ltd.) was used for carotenoid content determination using a carotenoid content biochemical kit (Shanghai Jianke Biological Technology Co., Ltd.); total amino acid content determination used total amino acid colorimetric test box (Wuhan Elreter Biological Technology Co., Ltd.); lactic acid content determination used lactic acid content kit (Shanghai Xiaoya Biological Technology Co., Ltd.); total pectin content used total pectin content kit (Shanghai Walan Biological Technology Co., Ltd.); flavonoid content determination used plant flavonoid content detection kit (Shanghai Hengfei Biological Technology Co., Ltd.); pH value determination used a desktop digital pH meter (Shanghai Yisheng Biological Technology Co., Ltd.); reducing sugar content determination used reducing sugar content detection kit (Wuhan Yakeyin Biotechnology Co., Ltd.); total phenol content determination used plant total phenol detection kit (Beijing Yita Biological Technology Co., Ltd.); γ-aminobutyric acid (GABA) content determination used CheKine TM γ-aminobutyric acid content detection kit (Wuhan Yakeyin Biotechnology Co., Ltd.); moisture content was determined by drying method; capsaicin content was determined by high performance liquid chromatography; color value was determined by spectrophotometer; soluble solids content was determined by handheld refractometer; nitrite content determination used food nitrite content test box (Shanghai Youlikelife Science Co., Ltd.).

[0062] Data processing: electronic nose data were analyzed by loading and principal component analysis (PCA) using the Winmuster system; response value radar chart was drawn by Origin2024 software; variance correlation analysis was performed using IBM SPSS Statistics 23 software, and principal component PCA and cluster analysis were performed using cloud tools in the Maiwei cloud platform (https: / / cloud.metware.cn).

[0063] 3. Results and analysis

[0064] (1) Loading analysis of the contribution of electronic nose sensors to principal components

[0065] In order to clarify the contribution of each sensor of the electronic nose to the aroma characteristics of different samples, Figure 1The loadings of the 10 sensors of the electronic nose on the first principal component (PC1) and the second principal component (PC2) are shown. PC1 explained 97.20% of the total variance of the original data, while PC2 explained 2.23%, and the cumulative contribution of the two was as high as 99.43%, indicating that the two principal components had highly summarized the main variation information of the original sensor data.

[0066] Sensor W1S had a very high positive loading value on PC1 (close to 0.9), much higher than all other sensors. This indicated that PC1 was mainly driven by the response variation of the W1S sensor. Therefore, the distribution difference of the samples on the PC1 axis mainly reflected the difference in the volatile substance spectrum detected by the W1S sensor. Other sensors such as W5S (about 0.35-0.4) and W2S (about 0.2) also had certain positive loadings on PC1, but their contribution was much smaller than that of W1S. Sensor W5S had a very high positive loading value on PC2 (about 0.9), which was the main contributor to PC2. This indicated that the W5S sensor captured the minor but unique odor difference information between samples. The remaining sensors such as W1W, W2W, W1C, W3C and W5C had either small loading values on PC1 and PC2 or were closely clustered around the origin, indicating that these sensors had less contribution to distinguishing the main odor difference between samples in this study. Therefore, WIS (methane) and W5S (alkanes, aromatic components) were the main flavor substances of different samples.

[0067] (2) Analysis of the response of the electronic nose to different samples

[0068] To visually compare the differences in volatile substances between different samples, the response intensity radar chart of different samples on each sensor was drawn, and the analysis results are shown in Figure 2 Sensor W1S (methane) showed strong response to most samples, especially in samples such as CT6+S, HG4+S, CT10+Y, HG5+S, etc., indicating that W1S was sensitive to the key volatile components released by these samples. Secondly, sensor W5S (alkanes, aromatic components) also had a relatively obvious response to these samples, but the intensity was generally lower than that of W1S. This result is consistent with the analysis of the loadings chart. In contrast, sensors W2W, W2S, W5C, W6S, etc. had low response values in all samples, indicating that these sensors were not sensitive to the volatile substances in the measured samples, or the concentration of these substances was low.

[0069] (3) Principal component analysis of the aroma components of different samples

[0070] To further explore the overall difference in aroma profile between different varieties and visualize the clustering, the present application performed principal component analysis (PCA) on the response data of the electronic nose sensor array. The PCA results are shown in Figure 3As shown, the contribution rate of the first principal component (PC1) was 72.33%, and the contribution rate of the second principal component (PC2) was 20.21%, with a cumulative contribution rate of 92.54%, indicating that the two principal components could fully explain most of the variation information in the original data and effectively reflect the aroma differences between samples. The sample points of most treatment groups formed relatively independent clusters in the two-dimensional PCA space, indicating that the electronic nose could effectively identify and distinguish the aroma characteristics of these varieties. In the first quadrant region (PC1 and PC2 are both positive values), HG3+Y was significantly separated from all other sample groups, showing its unique aroma characteristics; in the second quadrant region (PC1 is negative and PC2 is positive), CT6+S and CT10+S each formed an independent cluster and had good discrimination from other groups; in the third quadrant region (PC1 and PC2 are both negative values), CT10+Y, CT6+Y, and HG4+S, although relatively close to each other, showed a certain separation trend, indicating that they had both similarities and differences in aroma composition. In particular, the CT10+Y sample occupied a relatively independent space; in the fourth quadrant region (PC1 is positive and PC2 is negative), HG5+Y and HG4+Y were highly aggregated and even partially overlapped, indicating that the aroma characteristics of these two varieties were very similar and difficult to distinguish using the current electronic nose sensor array and PCA analysis. HG3+S near them was also in this quadrant, but slightly separated from the other two.

[0071] (4) Hierarchical cluster analysis of aroma characteristics of different samples

[0072] To further explore the similarities and differences in aroma characteristics between different samples, based on the response data of the electronic nose sensor array, all samples were classified by clustering, and the results were presented in the form of a tree diagram. Figure 4

[0073] 1) Main cluster branches:

[0074] At a higher level of dissimilarity, all samples were first clearly divided into two main clusters. The first main cluster included HG3+S and HG3+Y, which showed significant differences from all other samples. The second main cluster included all other samples.

[0075] 2) Uniqueness of HG3 group:

[0076] ​Samples HG3+S and HG3+Y, though belonging to the same HG3 group, first separated from each other and then aggregated at a distance of about 4.3, indicating that there is a significant aroma difference between them. At the same time, they have a larger aggregation distance with all other samples, highlighting the unique aroma characteristics of the HG3 group as a whole, which is consistent with the observation of the radar chart analysis, indicating that they each form a unique aroma profile that is distinct from other samples.

[0077] 3) Internal structure analysis of the second largest cluster:

[0078] Within the second largest cluster, subcluster one (right side), CT10+Y first separated from the others, showing that its aroma characteristics are somewhat different from those of the other samples. Subsequently, CT6+S and CT10+S aggregated at a lower distance level (about 1.6) to form a tight subcluster, indicating that the aroma characteristics of these two samples are very similar. Subcluster two (middle), HG5+S and HG4+Y aggregated at a distance of about 2.2, indicating that their aromas are relatively close. Subcluster three (left side), HG4+S first separated from the others. Subsequently, CT6+Y and HG5+Y tightly aggregated at a very low distance level (about 1.2), indicating that the aroma characteristics of these two samples are highly similar.

[0079] 4) Effect of processing technology for Y and S

[0080] The clustering results show that different processing technologies have a significant impact on clustering. In the HG3 group, HG3+S and HG3+Y, although homologous, have a long clustering distance; in the HG4 group, HG4+S and HG4+Y are classified in different subclusters. Similarly, HG5+S and HG5+Y of the HG5 group also belong to different subclusters; in the CT6 group, CT6+S clusters with CT10+S, while CT6+Y clusters with HG5+Y; in the CT10 group, CT10+S clusters with CT6+S, while CT10+Y is relatively independent. The above results show that different processing technologies have a greater impact on the aroma profile of the samples than the base varieties, making different base varieties with the same processing technology exhibit higher aroma similarity.

[0081] (5) Correlation analysis between electronic nose sensor response and physicochemical quality indicators

[0082] To further explore the internal relationship between the electronic nose sensor response and the physicochemical quality indicators of the samples, the inventors conducted a correlation analysis of the response data of all electronic nose sensors and the measured physicochemical quality indicators (including ADH, total amino acids, total flavonoids, reducing sugars, pH, total phenols, GABA, carotenoids, moisture content, sugar acid ratio, and soluble solids), and visualized the results as a heatmap using two-way hierarchical clustering. Figure 5 )。

[0083] 1) Cluster pattern of variables:

[0084] Sensor cluster: The electronic nose sensors apparently formed several clusters. For example, sensors W1C, W3C, W5C clustered together, indicating that they had a high similarity in the response pattern to the sample odor. Another large cluster contained W1W, W2W, W6S, W1S, W2S, and W5S, which also showed a strong positive correlation with each other, suggesting that they might be sensitive to similar types of volatile compounds;

[0085] Physico-chemical index cluster: The physico-chemical indexes also showed certain clustering characteristics. For example, total amino acids, total flavonoids, and ADH clustered together, indicating that they might have consistent trends in the sample. Reducing sugar, total phenol, GABA, and pH also clustered together. Moisture content, sugar-acid ratio, and soluble solids formed another cluster;

[0086] Mixed sensor and physico-chemical index cluster: The overall cluster showed that some sensors were closer to certain physico-chemical indexes in terms of correlation patterns. For example, W1C, W3C, W5C were closer to ADH, total amino acids, and total flavonoids in the larger branch of the cluster tree. W1W, W2W, W6S, W1S, W2S, W5S, on the other hand, clustered with reducing sugar, pH, total phenol, GABA, and carotenoids in another main branch.

[0087] 2) Significant correlation between electronic nose sensor response and physico-chemical quality indicators

[0088] W1C, W3C, W5C sensor group: This group of sensors showed extremely significant (p<0.001) or significant (p<0.01, p<0.05) positive correlation with ADH, total amino acids, and total flavonoids. This indicates that the increase in response of these sensors may be related to the increase in the content of these nutrients or flavor precursors. They showed extremely significant (p<0.001) or significant (p<0.01, p<0.05) negative correlation with reducing sugar, pH, total phenol, and GABA. They also showed significant negative correlation with carotenoids.

[0089] W1W, W2W, W6S, W1S, W2S, W5S sensor group: The response pattern of this group of sensors showed the opposite trend to the W1C group. They showed extremely significant (p<0.001) or significant (p<0.01, p<0.05) negative correlation with ADH, total amino acids, and total flavonoids. They showed extremely significant (p<0.001) or significant (p<0.01, p<0.05) positive correlation with reducing sugar, pH, total phenol, and GABA. They also showed significant positive correlation with carotenoids. Sensor W5S also showed significant positive correlation with sugar-acid ratio and soluble solids.

[0090] Other physicochemical indicators and sensor relationships: moisture content was positively correlated with W1C, W3C, W5C, ADH, total amino acid content, total flavonoids, etc., and negatively correlated with W1W, W2W, W6S, W1S, W2S, reducing sugar, total phenol, GABA, etc.; sugar acid ratio and soluble solids were significantly positively correlated with W5S, W1S, W2S, etc., and negatively correlated with W1C, W3C, etc.

[0091] 3) Correlation between physicochemical quality indicators:

[0092] Total amino acid content was extremely significantly positively correlated with total flavonoids; reducing sugar was extremely significantly positively correlated with total phenol and GABA; soluble solids were extremely significantly positively correlated with sugar acid ratio; moisture content was significantly negatively correlated with soluble solids and sugar acid ratio.

[0093] The present application system evaluates the synergistic effect of different pepper varieties and different bean paste raw materials (Yin bean paste / sun-dried bean paste) on the aroma characteristics of bean paste. As a rapid and effective aroma fingerprint analysis tool, the electronic nose successfully distinguished the overall aroma profiles of different processed bean paste samples in this study. Load analysis results showed that W1S (sensitive to methane) and W5S (sensitive to alkanes and aromatic components) sensors were the main contributors to distinguishing the aroma differences of samples, which was consistent with the results of radar chart that these two sensors had higher response values for most samples, indicating that these two types of volatile substances were the key components that constituted and distinguished the aroma characteristics of different bean pastes. This may be related to the degradation of amino acids to produce amines and alcohols during the fermentation of bean paste, as well as the oxidation of lipids to produce aldehydes, ketones, and alkanes.

[0094] PCA analysis and cluster analysis both clearly distinguished the different processed bean paste samples. It is worth noting that the HG3 group (Hongguan No. 3) showed significant aroma differences from all other sample groups, suggesting that this variety may have a very unique volatile component spectrum or content, and further indicating that the flavor components of bean paste may differ due to differences in varieties. In the cluster analysis, samples with the suffix "+Y" (Yin bean paste) (such as CT6+Y, HG5+Y, HG4+Y) and samples with the suffix "+S" (sun-dried bean paste) (such as CT6+S, CT10+S, HG5+S, HG4+S) tended to form their own clustering trends, indicating that the pretreatment method of bean paste had a decisive shaping effect on the volatile aroma components of the final product, and its influence even exceeded the basic differences of the samples. This may be related to the differences in microbial community structure, physicochemical indicators, and flavor substances produced during the fermentation of Yin bean paste and sun-dried bean paste. Sun-dried bean paste, due to sun-drying, has a lower water activity, which may be more conducive to the growth of aroma-producing yeasts and molds, while Yin bean paste may be more conducive to the activity of bacteria such as lactic acid bacteria in a relatively humid environment, leading to differences in aroma components.

[0095] In summary, the responses of different sensors of the electronic nose to the volatile components of the samples were significantly and complexly correlated with the physicochemical quality indicators. In particular, the response patterns of sensors W1C, W3C and W5C were positively correlated with ADH, total amino acids and total flavonoids, but negatively correlated with reducing sugars and total phenols; sensors W1W, W2W, W6S, W1S, W2S and W5S showed opposite trends. These results indicated that the electronic nose could not only distinguish the overall aroma profiles of the samples, but also had the potential to indirectly predict or indicate the levels of certain key physicochemical quality indicators in the samples. This provided important theoretical basis and data support for the rapid and non-destructive quality evaluation using electronic nose technology.

[0096] The above only describes the preferred embodiments of the present application, and it should be noted that those skilled in the art can make several improvements and refinements without departing from the principles of the present application, and these improvements and refinements should also be considered as the protection scope of the present application.

Claims

1. A method for optimizing the aroma quality of fermented soybean paste based on electronic nose technology, characterized in that, Includes the following steps: Preparation of Doubanjiang (fermented broad bean paste) sample set: Doubanjiang raw materials of different chili varieties and different processing methods were selected and combined in proportion to prepare a Doubanjiang sample set; Electronic nose detection: The electronic nose system is used to detect the aroma components of a soybean paste sample set and obtain sensor array response data; Multivariate statistical analysis: Principal component analysis, cluster analysis, and loading analysis were performed on the electronic nose data to identify key aroma-contributing sensors; Correlation analysis: Correlation analysis was performed between sensor response values ​​and physicochemical indicators of samples to establish an aroma-quality correlation model; Optimization screening: Based on the aroma characteristics and physicochemical indicators of fermented soybean paste, the optimal combination of chili varieties and fermented soybean paste raw materials was selected. Product preparation: Based on the screening results, a fermented soybean paste product with the target aroma characteristics was prepared.

2. The method for optimizing the aroma and quality of fermented soybean paste according to claim 1, characterized in that, The chili pepper varieties include one or more of the following: Hongguan No. 3, Hongguan No. 4, Hongguan No. 5, Chuanteng No. 6, and Chuanteng No.

10.

3. The method for optimizing the aroma and quality of fermented soybean paste according to claim 1, characterized in that, The different processing methods for fermented soybean raw materials include sun-dried soybeans and / or fermented soybeans.

4. The method for optimizing the aroma and quality of fermented soybean paste according to claim 3, characterized in that, The term "yin douban" refers to douban produced by sealing and fermenting in a cool environment; the term "shai douban" refers to douban produced by sun-drying.

5. The method for optimizing the aroma and quality of fermented soybean paste according to claim 1, characterized in that, The electronic nose system includes a metal sensor, the response characteristics of the metal sensor to specific gases, an aroma feature detection system, the response characteristics of the metal sensor to physicochemical indicators, a physicochemical indicator detection system, and a data processing and analysis system.

6. The method for optimizing the aroma and quality of fermented soybean paste according to claim 5, characterized in that, The metal sensor model includes one or more of W1C, W5C, W3C, W6S, W5S, W1S, W1W, W2S, W2W, and W3S.

7. The method for optimizing the aroma and quality of fermented soybean paste according to claim 1, characterized in that, The aroma feature detection method includes the following steps: cleaning the electronic nose for 100-150 seconds, detecting for 60-70 seconds, and taking data from 56-58 seconds for analysis.

8. The method for optimizing the aroma and quality of fermented soybean paste according to claim 1 or 5, characterized in that, The physicochemical indicators include one or more of the following: alcohol dehydrogenase content, carotenoid content, total amino acid content, lactic acid content, total pectin content, flavonoid content, pH value, reducing sugar content, total phenol content, γ-aminobutyric acid content, moisture content, capsaicin content, color value, soluble solids content, and nitrite content.

9. The method for optimizing the aroma and quality of fermented soybean paste according to claim 1, characterized in that, The preparation method of the fermented soybean paste sample set includes the following steps: thoroughly mixing chili varieties with fermented soybean raw materials of different treatment methods, and fermenting to obtain a fermented soybean paste sample set and / or a sun-dried fermented soybean paste sample set.

10. The application of the method for optimizing the aroma quality of fermented soybean paste according to any one of claims 1 to 9 in screening and / or preparing fermented soybean paste products with target aroma characteristics.

Images