Method and system for controlling hyperbolic Kunsha liquor brewing process

Abstract

The invention relates to the technical field of Baijiu brewing, in particular to a hyperbolic Kunsa Baijiu brewing process control method and system.The hyperbolic Kunsa Baijiu brewing process control method comprises the steps that fermentation duration characteristics of a hyperbolic Kunsa Baijiu brewing process are extracted, and cellar entering fermentation duration characteristics are obtained; determining a plurality of sampling time nodes according to the in-cellar fermentation duration characteristics; determining a relative abundance threshold value of each flora at each sampling time node; performing sampling detection at the sampling time node to obtain real-time relative abundance of each flora; for each flora, performing deviation calculation on the real-time relative abundance and the relative abundance threshold, and integrating deviation calculation results to obtain a flora abundance anomaly vector of the sampling time node; and inputting the flora abundance anomaly vector into a pre-trained hyperbolic cooperative regulation model, and outputting a flora regulation compensation vector. The quality and the consistency of the wine body can be stably improved.

Description

Technical Field

[0001] This invention relates to the technical field of baijiu brewing, and in particular to a method and system for controlling the brewing process of Shuangqu Kunsha baijiu. Background Technology

[0002] In the brewing process of Shuangqu Kunsha Baijiu, fermentation in the cellar is the core step that determines the quality of the liquor. The saccharification ability of mold, the alcoholic fermentation efficiency of yeast, and the ability of bacteria to synthesize flavor substances all affect the yield and the harmony of the liquor's flavor.

[0003] In the existing brewing process of Shuangqu Kunsha Baijiu, there is a lack of precise monitoring and control of the dynamic changes of microbial community during the fermentation process after the mash is placed in the cellar. Since the mash is in a semi-closed environment after being placed in the cellar, the reproduction and metabolism of mold, yeast and bacteria are affected by multiple factors such as temperature, pH and Shuangqu activity, which can easily lead to microbial imbalance. For example, if the abundance of yeast is too low, it will lead to insufficient alcohol fermentation, and if bacteria proliferate excessively, it will cause acidification of the mash, thereby inhibiting the saccharification function of mold, ultimately resulting in a decrease in alcohol yield and an imbalance in the proportion of flavor substances.

[0004] While existing process control methods include measures to detect microbial abundance through periodic sampling, they lack a systematic approach to link microbial anomalies with the synergistic regulatory parameters of the double-fermentation starter culture. For example, when insufficient mold abundance is detected, simply increasing the amount of starter culture without dynamically adjusting parameters based on the characteristics of the fermentation stage results in strong lag and low precision in regulation, making it difficult to reliably ensure batch consistency of double-fermentation starter culture Kunsha Baijiu. Summary of the Invention

[0005] This invention provides a method and system for controlling the brewing process of Shuangqu Kunsha Baijiu, which can stably improve the quality and consistency of the liquor and effectively solve the problems in the background art.

[0006] To achieve the above objectives, in a first aspect, the present invention provides a method for controlling the brewing process of Shuangqu Kunsha Baijiu, comprising:

[0007] Fermentation duration characteristics were extracted from the brewing process of Shuangqu Kunsha Baijiu to obtain the fermentation duration characteristics in the cellar.

[0008] Based on the characteristics of fermentation time in the pit, multiple sampling time points were determined; and the relative abundance thresholds of each microbial community at each sampling time point were determined.

[0009] Sampling and testing were performed at the specified sampling time points to obtain the real-time relative abundance of each bacterial community;

[0010] For each bacterial community, the deviation between the real-time relative abundance and the relative abundance threshold is calculated, and the deviation calculation results are integrated to obtain the bacterial community abundance anomaly vector at that sampling time point.

[0011] The abnormal abundance vector of the microbial community is input into a pre-trained hyperbolic collaborative regulation model, which outputs a microbial community regulation compensation vector.

[0012] In conjunction with the first aspect, in one possible design, based on the aforementioned microbial community regulation compensation vector and historical data, the predicted relative abundance of each microbial community at the next sampling time node is generated through a microbial community time-series prediction model.

[0013] When there is a deviation between the predicted relative abundance and its corresponding relative abundance threshold, the community regulation compensation vector is dynamically corrected based on the deviation ratio.

[0014] In conjunction with the first aspect, in one possible design, the microbial community includes molds, yeasts, and bacteria.

[0015] In conjunction with the first aspect, in one possible design, the sampling time points include the mold-dominated period, the yeast-dominated period, and the bacterial-dominated period.

[0016] In conjunction with the first aspect, in one possible design, the determination of the relative abundance thresholds of each microbial community is based on the microbial growth patterns of Shuangqu Kunsha Baijiu brewing, the quality requirements of the liquor, and fermentation environmental factors.

[0017] In conjunction with the first aspect, in one possible design, the relative abundance thresholds of each bacterial community at each sampling time point are determined, including:

[0018] Collect microbial community testing data from multiple batches of Shuangqu Kunsha Baijiu brewing process in the cellar, including the relative abundance of molds, yeasts and bacteria at different fermentation stages, as well as the corresponding fermentation environment parameters, Shuangqu dosage and liquor quality indicators.

[0019] The collected microbial community detection data were organized and classified according to the fermentation stage; the mean, standard deviation, maximum and minimum values ​​of the relative abundance of each microbial community at each fermentation stage were calculated.

[0020] Based on the statistical analysis results, the threshold range of relative abundance of each bacterial community was determined by taking the mean as the center and combining it with the standard deviation.

[0021] In conjunction with the first aspect, in one possible design, sampling and detection are performed at the sampling time point to obtain the real-time relative abundance of each bacterial community, including:

[0022] Prepare clean and sterile sampling instruments and sample containers before sampling;

[0023] A multi-point sampling method was adopted, with sampling points set at the upper, middle, and lower layers of the pit, as well as at different locations.

[0024] At each sampling point, a suitable amount of fermented mash sample was taken by inserting a sterile sampling instrument and sample container into the mash. The samples from each sampling point were collected together, thoroughly mixed, and the test sample was obtained.

[0025] Place the well-mixed test sample into an Erlenmeyer flask containing sterile physiological saline, shake to disperse the microorganisms in the solution, and obtain a diluent.

[0026] The relative abundance of molds, yeasts, and bacteria in the diluted solution was detected.

[0027] In conjunction with the first aspect, in one possible design, the microbial community regulation compensation vector includes the high-temperature Daqu compensation amount, the ventilation volume per unit pit area, and the reduction in time until the next turning.

[0028] Secondly, the present invention also provides a method and system for controlling the brewing process of Shuangqu Kunsha Baijiu, including:

[0029] The fermentation time feature extraction module extracts the fermentation time features of the Shuangqu Kunsha Baijiu brewing process to obtain the fermentation time features in the cellar.

[0030] The sampling parameter determination module determines multiple sampling time points based on the characteristics of fermentation time in the pit; and determines the relative abundance threshold of each microbial community at each sampling time point.

[0031] The sampling and detection module performs sampling and detection at the sampling time point to obtain the real-time relative abundance of each bacterial community;

[0032] The microbial community abundance anomaly analysis module calculates the deviation between the real-time relative abundance and the relative abundance threshold for each microbial community, and integrates the deviation calculation results to obtain the microbial community abundance anomaly vector at the sampling time point.

[0033] The hyperbolic collaborative regulation module inputs the abnormal abundance vector of the microbial community into the pre-trained hyperbolic collaborative regulation model and outputs the microbial community regulation compensation vector.

[0034] In conjunction with the second aspect, in one possible design, the sampling parameter determination module further includes a dynamic adjustment unit, which is used to adjust the sampling time node and the relative abundance threshold of each bacterial community in real time according to changes in the actual fermentation process.

[0035] The technical solution of this invention can achieve the following technical effects:

[0036] This method extracts the fermentation time characteristics to clarify the microbial abundance thresholds at each sampling time point, enabling dynamic monitoring of microbial communities such as molds, yeasts, and bacteria. It can capture changes in each microbial community in real time and make precise adjustments, thereby reducing the risk of microbial imbalance. By calculating the deviation between the real-time abundance of each microbial community and the preset threshold, it can quantify microbial anomalies, ensuring that problems are detected and adjusted promptly during fermentation. Compared to traditional methods such as simply increasing the amount of Daqu (a type of starter culture), this method can make more targeted corrections, avoiding problems of regulatory lag and low precision. Through a pre-trained hyperbolic collaborative regulation model, it can analyze microbial abundance anomaly vectors... The method outputs specific control and compensation vectors to precisely adjust multiple parameters during fermentation, avoiding purely empirical adjustments and ensuring the scientific nature and consistency of the control process, thereby improving the stability of the brewing process. By precisely controlling the abundance and interrelationships of various microbial communities during fermentation, the yield and flavor harmony of the liquor can be effectively improved. It can reduce flavor imbalance caused by microbial imbalance and improve the stability and consistency of different batches of liquor. The method not only solves the problems of lag and low precision in traditional process control, but also improves the brewing process of Shuangqu Kunsha Baijiu through real-time dynamic monitoring and intelligent control, thereby steadily improving the quality and consistency of the liquor. Attached Figure Description

[0037] Figure 1 This is a flowchart of the present invention;

[0038] Figure 2 A structural diagram of the brewing process control system for Shuangqu Kunsha Baijiu; Detailed Implementation

[0039] This application will now be described with reference to the accompanying drawings.

[0040] like Figure 1 As shown, the present invention provides a method for controlling the brewing process of Shuangqu Kunsha Baijiu, which specifically includes the following steps:

[0041] S1. Extract the fermentation time characteristics of the brewing process of Shuangqu Kunsha Baijiu to obtain the fermentation time characteristics in the cellar.

[0042] S2. Based on the characteristics of fermentation time in the pit, determine multiple sampling time points; and determine the relative abundance threshold of each microbial community at each sampling time point.

[0043] S3. Sampling and testing are performed at the sampling time points to obtain the real-time relative abundance of each bacterial community;

[0044] S4. For each bacterial community, the deviation between the real-time relative abundance and the relative abundance threshold is calculated, and the deviation calculation results are integrated to obtain the bacterial community abundance anomaly vector at the sampling time node.

[0045] S5. Input the abnormal abundance vector of the microbial community into the pre-trained hyperbolic collaborative regulation model and output the microbial community regulation compensation vector.

[0046] In this embodiment, the method extracts the fermentation time characteristics to clarify the microbial abundance threshold at each sampling time point, enabling dynamic monitoring of microbial communities such as molds, yeasts, and bacteria. It can capture changes in each microbial community in real time and make precise adjustments, thereby reducing the risk of microbial imbalance. By calculating the deviation between the real-time abundance of each microbial community and the preset threshold, it can quantify microbial anomalies, ensuring that problems are detected and adjusted promptly during fermentation. Compared to traditional methods such as simply increasing the amount of Daqu (a type of starter culture), this method can make more targeted corrections, avoiding problems of regulatory lag and low precision. Through a pre-trained hyperbolic synergistic regulation model, it can adjust the microbial abundance based on changes in microbial community abundance. The constant vector outputs specific control and compensation vectors, precisely adjusting multiple parameters during the fermentation process. This avoids purely empirical adjustments, ensuring the scientific nature and consistency of the control process, thereby improving the stability of the brewing process. By precisely controlling the abundance and interrelationships of various microbial communities during fermentation, the yield and flavor harmony of the liquor can be effectively improved. It can reduce flavor imbalances caused by microbial imbalances, improving the stability and consistency of different batches of liquor. The method not only solves the problems of lag and low precision in traditional process control, but also improves the brewing process of Shuangqu Kunsha Baijiu through real-time dynamic monitoring and intelligent control, thereby steadily improving the quality and consistency of the liquor.

[0047] In some embodiments of the present invention, for step S1, the fermentation time characteristics of the Shuangqu Kunsha Baijiu brewing process are extracted to obtain the fermentation time characteristics in the cellar.

[0048] The complete fermentation time of each batch of Shuangqu Kunsha Baijiu was collected from the distillery's production records, including the time of entry into the cellar and the time of exit from the cellar. This allows for a preliminary calculation of the fermentation time of each batch.

[0049] The collected historical data is organized and entered and stored in a unified format to facilitate subsequent analysis and processing; at the same time, the data is cleaned to remove abnormal and erroneous data; and the consistency of the data is checked to ensure that the information of the same batch of data is consistent in different records.

[0050] Based on the metabolic activity periods of molds, yeasts, and bacteria, the fermentation of the same batch in the fermentation pit was divided into three core stages, and the duration of each stage was extracted:

[0051] Mold-dominated period: from the time the cellar is filled until the peak activity of saccharifying enzymes appears. During this stage, starch saccharification is the main process, characterized by a rapid increase in the concentration of mold metabolites.

[0052] Yeast-dominated phase: from the peak of alcohol production rate to the stabilization of alcohol content, the core characteristic is a rapid increase in alcohol concentration.

[0053] Bacterial-dominated phase: from a significant increase in organic acid concentration to the end of fermentation, characterized by a rapid accumulation of esters.

[0054] Extract the duration variation of the same fermentation batch in different production cycles and record the duration shift characteristics under special conditions;

[0055] Key time points related to microbial activity during fermentation are marked, including the time when mold spore germination is completed, the time when yeast counts reach an active level, and the time when the abundance of specific bacteria reaches a certain level. The duration of these time points is used as characteristic parameters.

[0056] In this embodiment, by collecting, organizing, and cleaning historical fermentation duration records, the integrity, accuracy, and consistency of the data are ensured, avoiding analytical biases caused by data anomalies or confusion. Regarding the division of fermentation stages, three core stages and their respective characteristics are clearly defined based on the metabolic activity periods of molds, yeasts, and bacteria, making the originally vague fermentation process clear in stages and identifiable in features, facilitating targeted monitoring of the metabolic state of the microbial community at each stage. The extraction of duration variations and key nodes allows for the capture of fluctuations in fermentation duration across different production cycles and deviations under special conditions. Furthermore, by marking key time nodes, critical turning points in microbial metabolism can be precisely identified, helping to predict whether the fermentation process deviates from its normal trajectory. This step, through multi-dimensional extraction of fermentation duration characteristics, achieves a quantitative description and pattern discovery of the fermentation process, effectively improving the targeting and foresight of subsequent process control.

[0057] In some embodiments of the present invention, for step S2, multiple sampling time points are determined based on the characteristics of fermentation time in the cellar; and the relative abundance threshold of each microbial community at each sampling time point is determined; the microbial community includes molds, yeasts and bacteria;

[0058] Based on the characteristics of fermentation time in the cellar, the sampling time nodes are set in combination with the key stages of microbial metabolism in the fermentation process of Shuangqu Kunsha Baijiu. Since fermentation in the cellar is divided into mold-dominated, yeast-dominated, and bacteria-dominated stages, the characteristics of microbial activity are different in each stage. Therefore, sampling time nodes need to be set in each stage to comprehensively monitor the dynamic changes of the microbial community.

[0059] During the mold-dominant stage, considering that the growth, reproduction, and saccharification of mold are the core of this stage, sampling time points can be set at the beginning, middle, and end of this stage respectively; this will allow for timely monitoring of the abundance changes of mold at different growth stages and determination of whether its saccharification function is normal.

[0060] The yeast-dominant phase is a critical stage in alcohol fermentation. The activity and abundance of yeast directly affect alcohol production. Sampling time points can be set at the beginning, middle and end of this phase to facilitate monitoring whether the abundance of yeast can meet the requirements of alcohol fermentation.

[0061] The bacterial-dominated phase affects the formation of flavor compounds in wine. Bacterial metabolic activities affect the types and proportions of flavor compounds. Sampling time points are set at the beginning, middle and end of this phase to monitor changes in bacterial abundance and avoid excessive proliferation or insufficient abundance affecting the flavor of the wine.

[0062] At the same time, sampling time nodes need to be set during the transition period between two adjacent stages, because the microbial community structure will change significantly during the transition period, which is a critical period when the microbial community is prone to imbalance. Sampling at this time can promptly detect problems that may occur during the microbial community transformation process.

[0063] The relative abundance thresholds of each microbial community need to be determined based on the microbial growth patterns, quality requirements, and fermentation environment factors in the brewing of Shuangqu Kunsha Baijiu.

[0064] Microbial growth patterns: Molds begin to grow in the early stages of fermentation, utilizing starch and other substances in the mash for saccharification. In the early stages of fermentation, their growth rate is relatively fast, and their relative abundance gradually increases. As fermentation progresses, the consumption of nutrients and the accumulation of metabolic products gradually slow down their growth rate, and their relative abundance tends to stabilize. Understanding the growth curves of molds at different fermentation stages helps determine a reasonable range for their relative abundance. In the middle stages of fermentation, yeasts multiply rapidly and engage in vigorous alcoholic fermentation, at which point their relative abundance reaches its peak. In the later stages of fermentation, with the increase in alcohol concentration and the decrease in nutrients, yeast activity gradually decreases, and their relative abundance declines. Bacteria participate in the synthesis of flavor compounds during fermentation. Different types of bacteria have different roles and growth patterns at different fermentation stages. Generally, bacteria become active in the middle stages of fermentation, and their relative abundance gradually increases. In the later stages of fermentation, the growth of some bacteria may be inhibited, leading to changes in their relative abundance.

[0065] The quality requirements of the wine are directly influenced by the alcoholic fermentation efficiency of yeast. If the relative abundance of yeast is too low, insufficient alcoholic fermentation will lead to a decreased yield. Conversely, excessive yeast proliferation may consume excessive nutrients, affecting the growth of other microorganisms and the synthesis of flavor compounds, thus negatively impacting both yield and wine quality. Therefore, it is necessary to determine the appropriate range of relative yeast abundance to ensure a high yield. Molds provide fermentation substrates for yeasts through saccharification, and their metabolic products also affect the flavor of the wine. Bacterial synthesis of flavor compounds influences the aroma and taste of the wine. An imbalance in the relative abundance of molds, yeasts, and bacteria will result in an unbalanced ratio of flavor compounds, affecting the quality of the wine.

[0066] Fermentation environmental factors are crucial. Different microorganisms have varying degrees of temperature adaptability; for example, molds generally begin to grow at lower temperatures, while yeasts exhibit higher activity within a suitable temperature range. When determining the relative abundance threshold of the microbial community, the impact of temperature changes during fermentation on microbial growth must be considered. pH value affects the enzyme activity and cell membrane permeability within microbial cells, thus influencing microbial growth and metabolism. Different microorganisms have different pH tolerance ranges, and the pH value of the mash changes during fermentation due to the metabolic activities of the microorganisms. Therefore, it is necessary to determine a reasonable range for the relative abundance of each microbial community based on pH changes during fermentation. Double-cured koji contains abundant microorganisms and enzyme systems, and their activity has a significant impact on the growth and metabolism of the microbial community. When determining the threshold, factors such as the type, dosage, and activity of the double-cured koji must be considered.

[0067] Determine the relative abundance thresholds of each bacterial community at each sampling time point, including:

[0068] Collect microbial community testing data from multiple batches of Shuangqu Kunsha Baijiu brewing process in the past, including the relative abundance of molds, yeasts and bacteria at different fermentation stages, as well as the corresponding fermentation environment parameters, Shuangqu dosage and liquor quality indicators.

[0069] The collected data were organized and classified according to the fermentation stage; the average, standard deviation, maximum and minimum values ​​of the relative abundance of each microbial community at each fermentation stage were calculated; by analyzing the data, the variation pattern and normal fluctuation range of the relative abundance of each microbial community at different fermentation stages were understood.

[0070] Based on the statistical analysis results, the threshold range of relative abundance of each bacterial community was determined by taking the mean as the center and combining it with the standard deviation.

[0071] In this embodiment, by setting multiple sampling time points, the abundance changes of each microbial community at different fermentation stages can be precisely monitored. This helps to understand the growth status and metabolic activities of each microbial community in a timely manner, ensuring that the microbial activity at each stage is within a reasonable range. Based on the relative abundance threshold of each microbial community, the proportion of molds, yeasts, and bacteria can be coordinated, avoiding excessive or insufficient reproduction of any one microbial community, which could affect the flavor, alcohol yield, and overall quality of the liquor. Appropriate microbial community abundance can improve the flavor stability and yield of the liquor. By setting sampling time points during the transition period and key stages, the risk of microbial imbalance or over-reproduction can be detected in a timely manner, and corresponding measures can be taken to adjust and ensure the stable progress of the fermentation process, avoiding unnecessary production interruptions or quality problems. Considering the impact of fermentation environmental factors on microbial growth, personalized adjustments can be made to the fermentation environment for different batches. This helps to improve the microbial activity during the brewing process and enhance the flavor and other sensory qualities of the liquor. This step, by scientifically setting sampling time points and accurately controlling the abundance threshold of each microbial community, ensures the healthy growth and metabolic activities of microorganisms during the brewing process of Shuangqu Kunsha Baijiu, thereby optimizing the quality of the liquor and improving production efficiency.

[0072] In some embodiments of the present invention, for step S3, sampling and detection are performed at the sampling time point to obtain the real-time relative abundance of each bacterial community.

[0073] Sampling and testing are performed at the sampling time point, including:

[0074] Before sampling, clean and sterile sampling instruments and sample containers must be prepared to ensure that the sampling tools will not introduce bacteria to contaminate the sample. For stainless steel instruments, they must be sterilized by high-temperature steam before use. For glass instruments, they must be sterilized by dry heat or wet heat to kill all possible microorganisms.

[0075] Because the temperature, humidity, and microbial distribution of the mash in the fermentation pit vary, a multi-point sampling method is needed to ensure that the samples taken represent the actual situation of the mash in the entire fermentation pit. Sampling points are set up in the upper, middle, and lower layers of the fermentation pit, as well as in different directions. Generally speaking, the upper layer of mash has more contact with air and a relatively higher temperature, and the growth and metabolism of microorganisms may differ from those in the middle and lower layers. The lower layer of mash is under greater pressure, and the exchange of substances is relatively slow. By sampling from multiple points and mixing the samples, the representativeness of the samples can be improved.

[0076] At each sampling point, a sterile sampling tool is used to penetrate deep into the mash and take an appropriate amount of mash sample. The samples from each sampling point are collected together, thoroughly mixed, and then a portion is taken as a test sample. The remaining sample is sealed and stored for later use.

[0077] The fermented grains contain a large number of microorganisms, making direct testing difficult to accurately count and distinguish different bacterial groups. Therefore, sample dilution is necessary. A certain amount of the well-mixed fermented grain sample is weighed and placed in an Erlenmeyer flask containing sterile physiological saline. The flask is then shaken on a shaker for a certain period of time to ensure that the microorganisms are fully dispersed in the solution, thus preparing a 10-fold dilution. -1 The diluent was then diluted tenfold; then, using the tenfold dilution method, 10... -1 Dilute the solution to 10 -2 10 -3 10 -4 10 -5 10 -6 Sample dilutions at different dilutions; different bacterial groups may be more suitable for detection at different dilutions;

[0078] For each sample, the relative abundance of molds, yeasts, and bacteria was determined. Real-time fluorescence quantitative PCR was used to detect the DNA of specific bacterial species using specific primers, and the abundance of molds, yeasts, and bacteria was quantitatively analyzed. This method is accurate and efficient, and can obtain the relative abundance information of the bacterial community in a short time.

[0079] In this embodiment, a multi-point sampling method is used to collect samples of the mash from different layers and locations within the fermentation pit. This ensures that the samples represent the microbial community characteristics of the entire fermentation pit, overcoming the influence of differences in mash temperature, humidity, and microbial distribution. Strict aseptic procedures, including high-temperature steam sterilization and dry heat sterilization of equipment, ensure that no foreign bacteria are introduced during sampling, thus guaranteeing sample purity. Dilution treatment, where samples are diluted into solutions of different concentrations and combined with a tenfold dilution method, makes it easier to distinguish and count microorganisms at appropriate dilution levels, enabling more accurate detection of the abundance of molds, yeasts, and bacteria. Real-time quantitative PCR is used to detect the DNA of different bacterial groups using specific primers, obtaining relative abundance information of the bacterial groups in a short time, ensuring high efficiency and accuracy. Since different bacterial groups exhibit different characteristics at different dilution levels, the dilution method can better adapt to the detection needs of different microbial communities, improving the comprehensiveness and accuracy of the detection.

[0080] In some embodiments of the present invention, for step S4, for each bacterial community, the deviation between the real-time relative abundance and the relative abundance threshold is calculated, and the deviation calculation results are integrated to obtain the bacterial community abundance anomaly vector at the sampling time node.

[0081] The deviation between the real-time relative abundance of molds, yeasts, and bacteria and their corresponding thresholds was calculated using the following formula:

[0082] ΔX=|X r -X t |

[0083] Where △X represents the deviation between the real-time relative abundance and the relative abundance threshold, X r X represents the relative abundance of bacterial communities measured in real time. t This indicates a preset abundance threshold.

[0084] Based on the deviation values ​​of each microbial community, the degree of abnormality of each microbial community is calculated. If the abundance of yeast is below the threshold, the deviation is large, indicating that alcohol fermentation may be insufficient. If the bacterial abundance is too high, it may lead to acidification of the mash. The deviation calculation result of each microbial community will be used as a dimension of the sampling time point, and finally a data vector containing the deviations of multiple microbial communities will be formed, which is called the microbial community abundance anomaly vector.

[0085] In this embodiment, by monitoring the relative abundance of molds, yeasts, and bacteria in real time and comparing it with preset thresholds, the changes in the microbial community during fermentation can be accurately reflected. This helps to identify abundance deviations in a timely manner, ensuring that the concentration of each microbial community is within the ideal range and avoiding brewing problems caused by unsuitable abundance. This method can promptly detect problems such as excessively low yeast abundance or excessively high bacterial abundance. Through deviation calculation, effective early warnings can be provided during fermentation, and corresponding control measures can be taken. By integrating the deviation calculation results of each microbial community into an anomaly vector, the microbial community monitoring process is simplified, forming a clear data map. This step can quantify the degree of anomaly of each microbial community, providing specific parameters for subsequent control. Through this precise anomaly detection and feedback mechanism, the process parameters during fermentation can be adjusted more flexibly and accurately, thereby ensuring the consistency of the flavor and yield of each batch of liquor. The microbial community abundance anomaly vector provides data support for the dynamic control of the Shuangqu Kunsha Baijiu brewing process, enabling control measures to respond to changes in the fermentation process in real time, avoiding lag, and thus achieving a more efficient and stable production process.

[0086] In some embodiments of the present invention, for step S5, the abnormal abundance vector of the microbial community is input into a pre-trained hyperbolic collaborative regulation model, and the microbial community regulation compensation vector is output; the microbial community regulation compensation vector includes the high temperature koji compensation amount, the ventilation amount per unit pit area, and the shortened time before the next turning.

[0087] The double-curb synergistic regulation model was trained based on a large amount of experimental data and actual production data of baijiu brewing. It comprehensively considers the reproductive and metabolic characteristics of molds, yeasts and bacteria, as well as the coupling relationship between multiple factors such as temperature, pH and double-curb activity. Through machine learning algorithms, historical data is learned and analyzed to establish a mapping relationship between the abnormal abundance vector of the microbial community and the microbial community regulation compensation vector.

[0088] The model outputs a microbial community regulation compensation vector based on the input microbial community abundance anomaly vector; the output microbial community regulation compensation vector includes the following key parameters:

[0089] High-temperature Daqu compensation amount: Based on the activity of yeast and mold, the model can calculate whether the amount of high-temperature Daqu needs to be increased; high-temperature Daqu mainly provides enzyme activity such as esterase, which can promote alcoholic fermentation and the generation of flavor substances. Therefore, the adjustment of the compensation amount can be optimized to address the deficiency of yeast or mold.

[0090] Ventilation volume per unit area of ​​fermentation pit: Environmental factors such as temperature, oxygen content and humidity in the fermentation pit will affect the growth of microorganisms and the fermentation process; the model can recommend adjusting the ventilation volume of the fermentation pit based on the current growth of microorganisms in order to optimize the fermentation environment and avoid excessive bacterial growth or microbial imbalance.

[0091] Shortening the time until the next turning: Turning is a key operation to adjust the distribution of oxygen and heat during fermentation; based on the dynamic changes of the microbial community, the model can predict whether the turning cycle needs to be shortened or extended in order to maintain the stability of the fermentation process.

[0092] In this embodiment, the model trained by machine learning algorithms can automatically calculate and adjust parameters such as high-temperature koji, ventilation volume, and turning time based on actual microbial abundance anomalies, ensuring the balance and stability of the microbial community during the brewing process. The model's compensation vector considers the growth requirements and metabolic characteristics of different microorganisms, making environmental control during fermentation more scientific and precise, thereby improving alcohol fermentation efficiency, liquor flavor, and overall liquor quality. By automatically outputting compensation vectors through a pre-trained model, the subjectivity of human judgment is reduced, and errors caused by human operation are minimized, thereby improving the consistency and stability of production. The model can recommend appropriate compensation amounts based on different situations, avoiding excessive or insufficient raw material input, saving production costs and preventing potential waste. By adjusting the ventilation volume and turning cycle of the fermentation pit, the model can respond to changes in the microbial community in real time, ensuring suitable temperature, humidity, and oxygen levels, providing optimal environmental conditions for the brewing process, thereby improving fermentation efficiency and microbial community stability. This step enables refined management and automated control in the liquor brewing process, improving production efficiency, optimizing liquor quality, saving resources, and reducing the complexity of manual operation.

[0093] Furthermore, based on the aforementioned microbial community regulation compensation vector and historical data, the predicted relative abundance of each microbial community at the next sampling time node is generated through a microbial community time-series prediction model.

[0094] When there is a deviation between the predicted relative abundance and its corresponding relative abundance threshold, the community regulation compensation vector is dynamically corrected based on the deviation ratio.

[0095] The predicted relative abundance of each bacterial community at the next sampling time point generated by the bacterial community time series prediction model is compared with the predetermined relative abundance threshold of each bacterial community at the sampling time point, and the deviation value between the two is calculated.

[0096] Based on the calculated deviation value, the deviation ratio is determined; the deviation ratio reflects the degree to which the predicted relative abundance deviates from the threshold.

[0097] Based on the magnitude and direction of the deviation ratio, a corresponding adjustment strategy for the microbial community regulation compensation vector is formulated. If the deviation ratio is large, it indicates that the microbial community abundance may fluctuate significantly, requiring a larger adjustment. If the deviation ratio is small, the adjustment range is relatively small. At the same time, the interaction and influence between different microbial communities are considered to ensure that the adjusted microbial community regulation compensation vector can coordinate the growth and metabolism of each microbial community and maintain microbial community balance.

[0098] Based on the established adjustment strategy, the current microbial community regulation compensation vector is dynamically corrected. For example, if the predicted relative abundance of mold is lower than the threshold and the deviation ratio is large, it may be necessary to appropriately increase the compensation amount of high-temperature Daqu to provide more saccharifying enzymes and promote the growth and saccharification function of mold. At the same time, based on the interaction between mold and yeast, parameters such as ventilation volume per unit area of ​​fermentation pit are adjusted accordingly to ensure the stability and coordination of the entire fermentation process.

[0099] In this embodiment, a microbial community time-series prediction model is used to generate the microbial community abundance trend for the next sampling node in advance, avoiding problems such as insufficient alcohol fermentation or acidification of the mash caused by delayed regulation; a differentiated adjustment strategy is formulated based on the deviation ratio to achieve a precise match between the degree of deviation and the adjustment range; when the deviation ratio is large, the deviation is quickly corrected by significantly increasing the double-fermentation compensation amount or shortening the turning time; when the deviation ratio is small, only fine-tuning is performed to reduce resource waste; the metabolic correlation of molds, yeasts, and bacteria is fully considered during the correction process to avoid chain imbalances caused by the adjustment of a single microbial community; the dynamic correction mechanism can adapt to the microbial community characteristics at different fermentation stages and control the fluctuation range of microbial community abundance in each batch of fermentation; combined with historical data verification, after adopting this step, the yield of double-fermentation Kunsha Baijiu is improved, the batch differences in the proportion of flavor substances are reduced, and the uniformity of the product is enhanced.

[0100] In a preferred embodiment of the present invention, the double-koji process specifically adopts a process that combines herbal koji formula with traditional koji, thereby achieving microbial community optimization and wine quality improvement through the synergistic regulation of the two koji.

[0101] In practice, during each round of brewing Shuangqu Kunsha Baijiu, the amount of herbal koji and traditional koji is precisely mixed in a 1:1 weight ratio. This ratio allows the special components in the herbal koji to quickly activate the saccharification function of the mold in the early stage of fermentation, improving the saccharification efficiency of starch and providing sufficient substrate for the subsequent alcoholic fermentation of yeast. Meanwhile, the traditional koji continuously releases active components such as esterases in the middle stage of fermentation, forming a synergistic effect with the herbal koji, increasing the alcohol production rate of yeast during the peak fermentation period, and effectively extending the fermentation stability period, thereby increasing the accumulation of flavor substances in the liquor.

[0102] Based on the above double-cured ratio, the microbial community is dynamically monitored and regulated during the fermentation process using the control method of this invention. During the mold-dominant period, i.e., the first to third days after entering the fermentation pit, the relative abundance of mold is monitored. When the real-time detection value is lower than the set threshold range, an alarm is triggered, the amount of herbal curd is increased immediately, and the ventilation volume is adjusted accordingly to the preset multiple of the original set value, so as to promote the rapid recovery of mold activity and ensure the smooth start of the saccharification process.

[0103] The fermentation process enters the yeast-dominant period, which is from day 4 to 8 after the yeast is placed in the fermentation pit. If insufficient yeast abundance is detected, the system automatically calculates the amount of high-temperature koji that needs to be added and accurately delivers it to the fermentation tank. At the same time, it shortens the time between the next turning of the pile to optimize the fermentation environment, improve the alcohol fermentation efficiency of the yeast, and avoid the alcohol content from deviating from the target range due to insufficient yeast activity.

[0104] Compared to traditional sauce-flavored liquor brewed using only the Daqu (large koji) process, baijiu made using this double-koji Kunsha process can increase the content of total saponins and phosphorus trace elements in the liquor. The increase in total saponins and phosphorus trace elements is due to the synergistic transformation and enrichment of various medicinal and edible homologous components in the herbal koji during the fermentation process. In terms of flavor substances, the double-koji Kunsha process can increase the variety of aroma substances in the liquor, such as ester aroma components, making the aroma of the liquor richer, more mellow and more lasting, and improving the harmony and complexity of the taste.

[0105] In this embodiment, through precise double-koji ratio and dynamic control, the types of microbial communities in the koji-making process are enriched, and the fermentation process is made more stable and controllable. This overcomes the defects of insufficient saccharification, incomplete fermentation, and single microbial communities in traditional processes, and achieves a leapfrog improvement in the quality of the liquor.

[0106] like Figure 2 As shown, the present invention also provides a control system for the brewing process of Shuangqu Kunsha Baijiu, which specifically includes the following modules;

[0107] The fermentation time feature extraction module extracts the fermentation time features of the Shuangqu Kunsha Baijiu brewing process to obtain the fermentation time features in the cellar.

[0108] The sampling parameter determination module determines multiple sampling time points based on the characteristics of fermentation time in the pit; and determines the relative abundance threshold of each microbial community at each sampling time point.

[0109] The sampling and detection module performs sampling and detection at the sampling time point to obtain the real-time relative abundance of each bacterial community;

[0110] The microbial community abundance anomaly analysis module calculates the deviation between the real-time relative abundance and the relative abundance threshold for each microbial community, and integrates the deviation calculation results to obtain the microbial community abundance anomaly vector at the sampling time point.

[0111] The hyperbolic collaborative regulation module inputs the abnormal abundance vector of the microbial community into the pre-trained hyperbolic collaborative regulation model and outputs the microbial community regulation compensation vector.

[0112] In this embodiment, by real-time sampling and detection and setting abundance thresholds for each microbial community, the system can accurately monitor changes in the abundance of molds, yeasts, and bacteria. The system can detect and analyze microbial community dynamics in real time and systematically, promptly identifying potential microbial imbalances. Through a fermentation duration feature extraction module, the system dynamically determines the sampling time based on different time points during the fermentation process, avoiding the potential for missing key microbial community changes that traditional methods might miss when sampling at fixed time points. This allows for a better reflection of the microbial community dynamics during the actual fermentation process. Dynamic parameter adjustment enables the system to perform real-time control based on the characteristics of different fermentation stages, providing more precise responses. Microbial community changes improve the precision and speed of regulation. The system combines microbial abundance anomalies with a dual-fermentation synergistic regulation model, enabling adjustments to the amount of koji (fermented starter culture) used based on real-time data. This avoids the problem of simply increasing koji dosage while ignoring the synergistic effect of enzyme activity, as is common in traditional methods, thus achieving precise and scientific regulation. The compensation vector output by the dual-fermentation synergistic regulation module optimizes key parameters such as koji compensation amount, ventilation volume, and turning time, ensuring the coordination between flavor and yield, and avoiding lag and inaccuracy in process control. Through a microbial abundance anomaly analysis module, the system can perform real-time deviation analysis on each microbial community and integrate it into a microbial abundance anomaly analysis. The constant vector comprehensively reflects the changes in different microbial communities, avoiding the localized regulation required for imbalances in a single microbial community and ensuring microbial balance throughout the fermentation process. Combined with the compensation vector of the dual-stage synergistic regulation module, it can comprehensively regulate multiple factors, effectively improving the stability and consistency of the brewing process, avoiding batch-to-batch fluctuations that may occur in traditional processes, and ensuring consistent quality for each batch. Through precise regulation of molds, yeasts, and bacteria, the system can optimize the saccharification, fermentation, and flavor compound synthesis processes, ensuring a balance between flavor and yield, improving saccharification efficiency and overall brewing quality, and preventing flavor compounds caused by microbial imbalance. Imbalanced proportions and decreased alcohol yield; the system uses intelligent models to predict and regulate process parameters. Compared with traditional manual experience control and simple manual adjustments, the systematic approach can complete the regulation task more efficiently and stably, reducing human error and operational difficulty; the system can make personalized adjustments according to different fermentation environments, microbial communities, and alcohol requirements, and can adapt to the specific conditions of different batches, improving the adaptability and flexibility of the process; the system not only improves the precision and flexibility of the Shuangqu Kunsha Baijiu brewing process, but also effectively improves the consistency and stability of alcohol quality, significantly reducing quality fluctuations caused by microbial imbalance.

[0113] In a specific implementation, as one example, the sampling parameter determination module further includes a dynamic adjustment unit, which adjusts the sampling time node and the relative abundance threshold of each bacterial community in real time according to the changes in the actual fermentation process.

[0114] Fermentation is a non-linear and dynamic process, with the abundance, activity, and metabolic processes of the microbial community changing over time. Therefore, the dynamic adjustment unit needs to adjust the sampling time points according to different fermentation stages. In the initial stage, yeast growth may be relatively rapid, requiring a more frequent sampling frequency; while in the stationary phase, the sampling frequency can be reduced. Based on changes in external conditions such as temperature and pH during fermentation, the dynamic adjustment unit can automatically optimize the sampling frequency and time to ensure more accurate monitoring of the microbial community during critical periods.

[0115] The dynamic adjustment unit will adjust based on the real-time relative abundance data of the bacterial community; when the abundance of a certain bacterial community is close to its threshold, the sampling time interval will become shorter in order to detect problems as early as possible; while if the abundance of some bacterial communities is stable, the sampling interval can be appropriately extended.

[0116] If the abundance of certain bacterial communities deviates from the preset threshold, the dynamic adjustment unit will also trigger the function of adjusting the threshold; if the abundance of mold is low at a certain stage, the system will increase the sampling frequency and adjust the relative abundance threshold of mold to better cope with changes in mold.

[0117] In this embodiment, by dynamically adjusting the sampling time points and frequency, the monitoring of the microbial community is ensured to be more accurate during the critical fermentation stage, and the changes in the abundance of the microbial community can be reflected in a timely manner. The dynamic adjustment unit automatically optimizes the sampling frequency and time according to the changes in external conditions during the fermentation process, so that the system can flexibly respond to the needs of different fermentation stages, avoid unnecessary sampling, and improve efficiency. When the abundance of a certain microbial community is close to the preset threshold, the system will automatically shorten the sampling interval to detect potential problems in advance and avoid fermentation abnormalities caused by rapid changes in the microbial community. For special microbial communities such as molds, when their abundance deviates from the threshold, the system will trigger the threshold adjustment function to appropriately increase the sampling frequency and optimize the threshold to cope with changes in molds and enhance the controllability of the fermentation process.

[0118] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. A control method for the brewing process of baijiu, characterized in that, The method comprises the following steps: extracting fermentation time length characteristics of the biqu Kunsa liquor brewing process to obtain pit entry fermentation time length characteristics; determining a plurality of sampling time nodes according to the pit entry fermentation time length characteristics; and determining relative abundance thresholds of each microbial group at each sampling time node; sampling and detecting at the sampling time nodes to obtain real-time relative abundances of each microbial group; for each microbial group, performing deviation calculation on the real-time relative abundance and the relative abundance threshold, and integrating the deviation calculation results to obtain a microbial group abundance anomaly vector at the sampling time node; inputting the microbial group abundance anomaly vector into a pre-trained biqu collaborative regulation model to output a microbial group regulation compensation vector.

2. The brewing process control method of Double-Curved Kunsaha Baijiu according to claim 1, characterized in that, Based on the microbial group regulation compensation vector and historical data, a microbial group time series prediction model is used to generate predicted relative abundances of each microbial group at the next sampling time node; when the predicted relative abundance deviates from the corresponding relative abundance threshold, dynamically correcting the microbial group regulation compensation vector based on the deviation ratio.

3. The brewing process control method of Double-Curved Kunsaha White Liquor according to claim 1, characterized in that, The microbial groups include molds, yeasts, and bacteria.

4. The brewing process control method of Double-Curved Kunsaha White Liquor according to claim 3, characterized in that, The sampling time nodes include a mold dominant period, a yeast dominant period, and a bacteria dominant period.

5. The brewing process control method of Double-Curved Kunsaha White Liquor according to claim 1, characterized in that, The determination of the relative abundance thresholds of each microbial group is based on the microbial growth law of the biqu Kunsa liquor brewing process, the quality requirements of the liquor body, and the fermentation environmental factors.

6. The brewing process control method of Double-Curved Kunsaha White Liquor according to claim 1, characterized in that, Determining the relative abundance thresholds of each microbial group at each sampling time node comprises: collecting microbial group detection data of pit entry fermentation in the brewing process of multiple batches of biqu Kunsa liquor in the liquor factory, including the relative abundances of molds, yeasts, and bacteria at different fermentation stages, as well as the corresponding fermentation environmental parameters, biqu dosages, and liquor body quality indicators; organizing the collected microbial group detection data and classifying and counting them according to the fermentation stages; calculating the average value, standard deviation, maximum value, and minimum value of the relative abundance of each microbial group at each fermentation stage; determining the threshold range of the relative abundance of each microbial group based on the statistical analysis results, with the average value as the center and the standard deviation.

7. The brewing process control method of Double-Curved Kunsaha White Liquor according to claim 1, characterized in that, Sampling and detecting at the sampling time nodes to obtain real-time relative abundances of each microbial group comprises: preparing clean and sterile sampling instruments and sample containers before sampling; using a multi-point sampling method to set sampling points at the upper, middle, and lower layers of the pit pool and at different orientations; at each sampling point, using sterile sampling instruments and sample containers to penetrate into the interior of the fermented grains to take an appropriate amount of fermented grain sample; collecting the samples from each sampling point together and mixing them evenly to obtain a detection sample; placing the mixed and even detection sample into a triangular flask containing sterile normal saline, oscillating to disperse the microorganisms in the solution to obtain a dilution liquid; detecting the relative abundances of molds, yeasts, and bacteria in the dilution liquid.

8. The brewing process control method of Double-Curved Kunsaha White Liquor according to claim 1, characterized in that, The microbial group regulation compensation vector includes high-temperature Daqu compensation amount, unit pit area ventilation amount, and shortened time to the next turning.

9. A control system for the brewing process of baijiu, characterized in that, The method comprises the following steps: a fermentation time length characteristic extraction module extracts fermentation time length characteristics of the biqu Kunsa liquor brewing process to obtain pit entry fermentation time length characteristics; a sampling parameter determination module determines a plurality of sampling time nodes according to the pit entry fermentation time length characteristics; and determines relative abundance thresholds of each microbial group at each sampling time node; a sampling detection module configured to perform sampling detection at the sampling time node to obtain real-time relative abundance of each flora; a flora abundance anomaly analysis module configured to, for each flora, perform deviation calculation on the real-time relative abundance and the relative abundance threshold, and integrate the deviation calculation results to obtain a flora abundance anomaly vector of the sampling time node; a hyperbolic collaborative regulation module configured to input the flora abundance anomaly vector into a pre-trained hyperbolic collaborative regulation model to output a flora regulation compensation vector.

10. The brewing process control system of claim 9, wherein, The sampling parameter determination module further comprises a dynamic adjustment unit configured to adjust the sampling time node and the relative abundance threshold of each flora in real time according to changes in the actual fermentation process.

Images