Fermentation treatment equipment and process for bamboo wine data optimization learning

Abstract

The invention provides fermentation processing equipment and process for bamboo wine data optimization learning, and relates to the technical field of bamboo wine preparation, the fermentation processing equipment comprises main control equipment and a fermentation tank, the main control equipment is mounted on one side of the fermentation tank, the main control equipment comprises a fermentation processing system, a tank cover is mounted at the top of the fermentation tank, and a stirring motor is mounted in the middle of the top end of the tank cover; the bottom end of the stirring motor is in transmission connection with an output shaft, a stirrer is installed at the bottom end of the output shaft, and winding assemblies are installed on the two sides, the front face and the back face in the stirrer. The fermentation processing system comprises a fermentation intelligent sensing and monitoring system, an automatic fermentation condition adjusting system, a fermentation microbial community accurate regulation and control system, a bamboo wine data optimization learning system, a fermentation stirring and mixing system and a fermentation intelligent control system. The ultra-miniature sensor can comprehensively monitor multi-dimensional data in real time, data noise and errors are effectively reduced, and a powerful basis is provided for accurate regulation and control.

Description

Technical Field

[0001] The invention relates to the technical field of bamboo wine preparation, in particular to fermentation processing equipment and a process for optimizing bamboo wine data learning. Background Art

[0002] Bamboo wine, a unique alcoholic beverage, has been gaining increasing attention in recent years. Combining bamboo with alcohol, the wine absorbs natural flavors and nutrients during its growth, giving the wine its unique taste and aroma. Traditional bamboo wine production relies on experience and simple tools. After the bamboo reaches a certain growth stage, the wine is injected into the bamboo cavity through manual drilling and other methods. Fermentation then occurs within the bamboo's natural growth environment and limited natural conditions.

[0003] With the rapid development of modern technology and rising consumer expectations for product quality, the limitations of traditional bamboo wine fermentation methods are becoming increasingly apparent. As industrialized production becomes increasingly mainstream, precise control of various parameters during the fermentation process plays a crucial role in improving bamboo wine quality and production efficiency. Currently, intelligent fermentation equipment and processes are widely used in fermentation industries such as wine and beer. For example, high-precision sensors are used to monitor fermentation environmental parameters in real time, and automated control systems use preset optimization algorithms to precisely control key parameters such as temperature, humidity, and oxygen content, significantly improving product quality stability and production efficiency. These successful experiences provide valuable insights and technical support for the innovation of bamboo wine fermentation processes.

[0004] The existing bamboo wine fermentation processing equipment and fermentation process will have the following problems in the actual winemaking process: 1. First, the data collection methods of traditional bamboo wine fermentation are relatively simple, and are usually only equipped with basic temperature and humidity measurement tools, such as simple thermometers and hygrometers. It is difficult to conduct comprehensive and real-time monitoring of multi-dimensional data such as gas composition and pH that are crucial to the fermentation process. The limitations of this data collection make it impossible to accurately grasp the complex chemical and biological changes in the fermentation process. Data management is scattered and isolated, and lacks systematic integration and in-depth mining mechanisms; 2. Secondly, the temperature control method mainly relies on manual experience and simple heating or cooling equipment for adjustment. It is impossible to achieve precise dynamic adjustment based on the fine temperature requirements of different stages of bamboo wine fermentation. Humidity and ventilation control lack scientific and reasonable planning and coordination mechanisms. In the traditional fermentation process, ventilation or moisturizing operations are often simply performed, and it is impossible to adjust the fermentation process according to the oxygen demand of microorganisms and the specific humidity requirements at different stages; 3. Traditional technologies lack in-depth understanding and effective prediction methods for the evolution of microbial communities during bamboo wine fermentation, and are also unable to optimize learning based on data; 4. Traditional bamboo wine fermentation stirring equipment is relatively simple and lacks intelligent design concepts. The shape and angle of the stirring blades are fixed and cannot be automatically adjusted according to changes in the concentration and viscosity of the materials in the fermentation tank at different fermentation stages; these problems will affect the mixed fermentation of bamboo wine.

[0005] Therefore, a fermentation processing equipment and process for bamboo wine data optimization learning are needed to solve the above problems. Summary of the Invention

[0006] Technical problems solved

[0007] In view of the shortcomings of the existing technology, the present invention provides a fermentation processing equipment and process for bamboo wine data optimization learning, which solves the problems mentioned in the above background technology.

[0008] Technical Solution

[0009] To achieve the above objectives, the present invention is implemented through the following technical solutions: a fermentation processing equipment and process for bamboo wine data optimization learning, comprising a main control device and a fermentation tank, the main control device being installed on one side of the fermentation tank, the main control device comprising a fermentation processing system, a tank cover being installed on the top of the fermentation tank, a stirring motor being installed in the middle of the top of the tank cover, the bottom end of the stirring motor being transmission-connected with an output shaft, the bottom end of the output shaft being installed with an agitator, winding assemblies being installed on both sides, the front and back sides of the agitator, the winding assemblies being driven by a motor, a pull rope being wound around the outside of the winding assemblies, embedded columns being welded on both sides, the front and back sides of the outside of the agitator, a sliding opening being opened on the surface of the embedded column, the sliding opening being in the shape of an arc, a sliding plate being slidably connected in the sliding opening, the sliding plate and the sliding opening being adapted to each other, a connecting column being embedded on one side of the sliding plate, a leaf being clamped on one side of the connecting column, and one end of the pull rope being connected to the top of the leaf.

[0010] Preferably, the outside of the leaves are slidably connected to the leaf sleeve, a pushing cylinder is installed between the leaf sleeve and the leaves, the top and bottom of one side of the leaves are provided with blocking blocks, an ultra-miniaturized sensor is installed inside the fermentation tank, a feeding port is provided on one side of the top of the tank cover, an automatic feeding device is provided outside the feeding port, a heating device is installed on one side of the top of the tank cover, a refrigeration device is installed on the other side of the top of the tank cover, a vent is provided on one side of the top of the fermentation tank, and a breathing valve is externally connected to the vent.

[0011] The main control device is installed on one side of the fermentation tank and controls the entire fermentation tank through the internal fermentation processing system. Stirring blades are installed in four directions outside the agitator. The complete stirring blade is composed of a blade and a blade sleeve. The pushing cylinder on the blade side will push the external blade sleeve to slide outward to adjust the length of the entire stirring blade. The pull rope can be reeled in through the winding assembly in the agitator. The winding assembly is driven by a motor. When the winding assembly reels the pull rope, the bottom end of the pull rope will pull the blade upward. The blade side slides in the sliding mouth through the slide, which can indirectly adjust the stirring angle of the entire stirring blade.

[0012] Preferably, the fermentation processing system includes a fermentation intelligent sensing and monitoring system, an automated fermentation condition adjustment system, a fermentation microbial community precision control system, a bamboo wine data optimization learning system, a fermentation stirring and mixing system, and a fermentation intelligent control system.

[0013] Preferably, the fermentation intelligent sensing and monitoring system includes a multi-sensor data fusion and feature extraction unit and a blockchain-based data encryption and sharing unit. The multi-sensor data fusion and feature extraction unit combines the Kalman filter algorithm and the principal component analysis algorithm, and performs real-time filtering and state estimation on the time series data collected by the Kalman filter algorithm and the ultra-miniaturized sensor during the growth stage and fermentation process. The ultra-miniaturized sensor 6 includes a temperature sensor, a humidity sensor, a gas sensor, and a pH sensor. The principal component analysis algorithm is used to extract features from the high-dimensional data of the sensor, and the generated variables are converted into mutually uncorrelated principal components for subsequent extraction of characteristic spectral information related to the content of bamboo wine flavor substances by a spectrum analyzer. The blockchain-based data encryption and sharing unit in the fermentation intelligent sensing and monitoring system combines the elliptic curve encryption algorithm and the distributed hash table algorithm, encrypts the sensor data by the elliptic curve encryption algorithm, and constructs the data storage and index structure of the blockchain network by the distributed hash table algorithm, thereby realizing efficient storage, query, sharing and collaborative work of data among various departments within the enterprise and upstream and downstream industry chain partners.

[0014] The principal component analysis algorithm formula is as follows:

[0015] Covariance matrix calculation:

[0016]

[0017] Where C is the covariance matrix, X is the standardized data matrix, and n is the number of samples;

[0018] Eigenvalues and eigenvectors:

[0019] Cv=λv

[0020] Among them, v is the eigenvector and λ is the corresponding eigenvalue;

[0021] Data after dimensionality reduction:

[0022] Z=XV

[0023] Among them, Z is the data after dimensionality reduction, and V is the matrix composed of all eigenvectors;

[0024] The elliptic curve encryption algorithm formula is as follows:

[0025] Elliptic curve equation:

[0026] y 2 =x 3 +ax+b

[0027] Where a and b are the parameters of the curve, and x and y are the coordinates of the point on the curve;

[0028] Public and private key generation:

[0029] P=dG

[0030] Among them, P is the public key, d is the private key, and G is the base point of the elliptic curve;

[0031] The distributed hash table formula is as follows:

[0032] Node hash function:

[0033] h(k)=(k mod N)

[0034] Where k is the key value, N is the total number of nodes, and h(k) is the position of key value k in the hash table.

[0035] Preferably, the automated fermentation condition regulation system includes a temperature adaptive control unit and a humidity and ventilation coordinated control unit. The automated fermentation condition regulation system integrates a fuzzy logic control algorithm and a model predictive control algorithm through the temperature adaptive control unit. The fuzzy logic control algorithm formulates a preliminary control strategy based on the fuzzy requirements for temperature in the fermentation stage and real-time temperature sensor data. The model predictive control algorithm predicts future temperature changes based on the bamboo wine fermentation temperature dynamic model, and optimizes and calculates the control signals of the refrigeration and heating equipment in combination with the fuzzy logic control strategy to achieve precise dynamic regulation of the temperature in each fermentation stage. In the yeast activation stage, the temperature is precisely controlled at 25-28°C, the main fermentation period is stabilized at 30-32°C, and the later aging stage is slowly reduced to 18-20°C, and the temperature regulation accuracy needs to reach ±0.5°C. The humidity and ventilation coordinated control unit combines the proportional-integral-differential algorithm and the particle swarm optimization algorithm. The proportional-integral-differential algorithm performs conventional control on the humidity regulation system according to the real-time feedback data of the humidity and ventilation volume sensors. The particle swarm optimization algorithm is used to optimize the proportional-integral-differential algorithm parameters to achieve precise regulation of the ventilation volume at 0.1-1m according to the microbial oxygen demand and the fermentation stage. 3 / (h·m 3 ) range and maintain the synergistic control effect of humidity in the ideal range of 60%-80%, optimizing gas exchange and water balance;

[0036] The formula of the fuzzy logic control is as follows:

[0037] Fuzzification:

[0038]

[0039] Among them, μ A (x) is the membership function, c is the center position, d is the width parameter, and x is the input variable;

[0040] Reasoning process:

[0041] z=∑wiμi(x)

[0042] Among them, w i is the weight, μ i (x) is the membership of the fuzzy set and z is the output.

[0043] Preferably, the fermentation microbial community precision control system has a microbial community dynamic prediction unit and a microbial niche construction and optimization unit. The microbial community dynamic prediction unit integrates the long short-term memory network algorithm and the Bayesian network algorithm. The long short-term memory network algorithm is used to analyze the learning and evolution laws of the microbial community time series data during the bamboo wine fermentation process. The Bayesian network algorithm constructs a microbial interaction network model based on microbial gene sequencing and metagenomic analysis data. The two are combined to predict the future dynamic changes of the microbial community and provide a decision-making basis for the microbial intelligent delivery system. The microbial niche construction and optimization unit adopts a genetic algorithm and an NSGA-I multi-objective optimization algorithm. The genetic algorithm searches for the optimal addition combination scheme of microbial growth promoters and inhibitors. The microbial growth promoters include vitamins, amino acids and nutrients, and the microbial inhibitors include natural antibacterial substances and biological preservatives. The NSGA-II multi-objective optimization algorithm processes the target optimization problem, weighs the growth and reproduction of beneficial microorganisms and the inhibition of harmful microorganisms, determines the optimal strategy for regulating the ecological balance of the microbial community, and constructs a microbial niche that is conducive to bamboo wine fermentation.

[0044] Preferably, the fermentation stirring and mixing system includes a stirring blade shape and angle adaptive adjustment unit and a stirring speed and time intelligent optimization unit. The stirring blade shape and angle adaptive adjustment unit combines a neural network algorithm and a finite element analysis algorithm. The neural network algorithm learns the corresponding relationship data of the material concentration, viscosity and fermentation stage in the fermentation tank and the optimal shape and angle of the stirring blade to establish a prediction model. The finite element analysis algorithm simulates and analyzes the mechanical properties of the blade according to the neural network prediction parameters to ensure that the blade automatically adjusts the shape and angle according to the material state. The stirring speed and time intelligent optimization unit integrates a reinforcement learning algorithm and a dynamic programming algorithm. The reinforcement learning algorithm adjusts the speed of the stirring motor 4 The integration and stirring time control are used as the action space, and the bamboo wine fermentation efficiency and flavor quality indicators are used as reward functions to explore the optimal strategy in the fermentation stage. The dynamic programming algorithm optimizes the strategy in the reinforcement learning process to determine the global optimal stirring speed and time control strategy. In the early stage of fermentation, the stirring speed is controlled at 50-80 rpm, and stirring is carried out for 15-20 minutes every 2-3 hours; in the middle stage of fermentation, the stirring speed is increased to 80-120 rpm, and the stirring time interval is adjusted to 1-2 hours; in the late stage of fermentation, the stirring speed is reduced to 30-50 rpm, and stirring is carried out 1-2 times a day, each time for 5-10 minutes, to avoid excessive stirring affecting the flavor quality of bamboo wine, and to achieve efficient, stable and high-quality fermentation.

[0045] Preferably, the fermentation intelligent control system includes an intelligent decision-making engine based on multimodal data fusion and a real-time feedback adjustment unit. The intelligent decision-making engine combines deep neural networks and convolutional neural networks, processes time series data from sensors through deep neural networks, learns the changing patterns of bamboo wine flavor substances, temperature and humidity, and gas composition during the fermentation process, and automatically adjusts various control parameters in the fermentation process according to the flavor target. The convolutional neural network is used to extract spatial features from spectral sensor data during the fermentation process to optimize the fermentation environment adjustment decision. The real-time feedback adjustment unit combines incremental learning algorithms and adaptive control algorithms to dynamically adjust fermentation conditions according to real-time feedback information during the bamboo wine fermentation process to ensure that the generation and concentration of bamboo wine flavor substances are maintained within a preset range.

[0046] Preferably, the bamboo wine data optimization learning system includes an intelligent prediction and simulation module, which integrates an adaptive genetic algorithm and a multi-task learning algorithm. During the bamboo wine fermentation process, a mathematical model of the fermentation process is constructed in real time based on data such as temperature, humidity, gas composition, and microbial community, and the flavor evolution trend during the bamboo wine fermentation stage is predicted. The adaptive genetic algorithm is used to adaptively optimize model parameters in each fermentation stage to improve prediction accuracy. The multi-task learning algorithm shares learning knowledge of different tasks, so that the model achieves global optimization effects in temperature and humidity control, flavor optimization, and microbial community regulation, thereby providing accurate prediction and optimization solutions for the bamboo wine fermentation process.

[0047] Preferably, the fermentation intelligent sensing and monitoring system, the automated fermentation condition adjustment system, the fermentation microbial community precise control system, the fermentation stirring and mixing system, and the fermentation intelligent control system work together to perform comprehensive intelligent control and optimization of the bamboo wine fermentation process, thereby improving the quality stability, flavor uniqueness, and production efficiency of the bamboo wine. The specific process is as follows:

[0048] Sp1. First, the main control device 1 is installed on one side of the fermentation tank 2, and the tank cover 3 is ensured to be well sealed with the fermentation tank 2. The processed raw materials and the required brewing raw materials are fed into the fermentation tank 2 through the feeding port 7 using the automatic feeding device, and the fermentation process is ready to start. The ultra-miniaturized sensor 6 installed inside the fermentation tank 2 starts to collect environmental data after the raw materials are added, and transmits it to the fermentation intelligent sensing and monitoring system in the main control device 1. The fermentation intelligent sensing and monitoring system uses the Kalman filter algorithm and the principal component analysis algorithm to process the data, and ensures data security and sharing through the data encryption and sharing unit based on the blockchain:

[0049] Sp2, heating device 8 and refrigeration device 9 are controlled by the temperature adaptive control unit of the automated fermentation condition regulation system. The fuzzy logic control algorithm formulates a preliminary strategy based on the fermentation stage and real-time temperature data. The model predictive control algorithm establishes a temperature dynamic model and predicts future temperature changes. The combination of the two determines the final control signal of the refrigeration or heating equipment, accurately controlling the yeast activation stage at 25-28°C, the main fermentation stage at 30-32°C, and the late aging stage at 18-20°C.

[0050] Sp3, the microbial community dynamics prediction unit in the fermentation microbial community precision control system uses the LSTM algorithm to collect microbial community data to construct a time series dataset and predict change trends. The Bayesian network algorithm constructs an interaction network model based on data such as microbial gene sequencing. The combination of the two can more accurately predict the future dynamics of the microbial community and provide a basis for intelligent microbial delivery.

[0051] Sp4 and the stirring motor 4 drive the stirrer 5 to work. The winding component 15 in the stirrer 5 is driven by the motor to reel in or release the pull rope 16 according to the prediction results of the neural network algorithm, pulling the blade 10 to adjust the angle of the stirring blade, and at the same time pushing the cylinder 21 to push the blade cover 11 to adjust the blade length, adapting to the precise control of the stirring requirements of the three fermentation stages of yeast activation, main fermentation period and late aging. The reinforcement learning algorithm uses the speed adjustment of the stirring motor 4 and the stirring time control as the action space, and the fermentation efficiency and flavor quality of bamboo wine as the reward function for experimentation and learning.

[0052] Beneficial effects

[0053] The present invention provides a fermentation processing equipment and process for optimizing bamboo wine data learning.

[0054] Beneficial effects:

[0055] 1. In terms of data acquisition and monitoring, the present invention's ultra-miniaturized sensors enable comprehensive, real-time monitoring of multi-dimensional data. Processing with Kalman filtering and principal component analysis algorithms effectively reduces data noise and error, allowing precise extraction of characteristic spectral information related to the content of bamboo wine flavor compounds, providing a powerful basis for precise control. A blockchain-based data encryption and sharing unit ensures data security and confidentiality while promoting data collaboration within the enterprise and across the entire supply chain, facilitating optimized planting for raw material suppliers and targeted promotion for distributors. Regarding fermentation condition regulation, temperature control utilizes a fuzzy logic control algorithm to develop a preliminary strategy. Combined with a model predictive control algorithm, this algorithm integrates multiple factors to establish a dynamic model and optimize control signals, enabling precise dynamic temperature regulation at each fermentation stage with an accuracy of ±0.5°C. This creates an ideal temperature environment for bamboo wine fermentation, promoting smooth yeast activation, primary fermentation, and aging. Humidity and ventilation control utilizes a PID algorithm combined with a particle swarm optimization algorithm to optimize PID parameters. This allows for precise, coordinated control of ventilation volume and humidity based on microbial oxygen demand and fermentation stage, optimizing gas exchange and water balance, and promoting microbial growth and proliferation, as well as the development of bamboo wine flavor.

[0056] 2. At the microbial community regulation level, the present invention uses the LSTM algorithm and the Bayesian network algorithm to accurately predict the future dynamic changes of the microbial community, providing a decision-making basis for the intelligent placement of microorganisms, and realizing the targeted regulation and enhancement of the bamboo wine flavor; the microbial niche construction and optimization unit adopts the genetic algorithm and the NSGA-II multi-objective optimization algorithm to determine the optimal addition combination of microbial growth promoters and inhibitors, construct a microbial niche that is conducive to the fermentation of bamboo wine, maintain the stable and healthy development of the microbial community, and ensure the stability and quality of the bamboo wine fermentation process.

[0057] 3. During the stirring and mixing stage, the present invention uses a neural network algorithm to automatically adjust the shape and angle of the stirring blade according to the material concentration, viscosity and fermentation stage to meet the stirring requirements of different stages and improve the fermentation efficiency; the stirring speed and time are intelligently optimized through reinforcement learning algorithm and dynamic programming algorithm to explore the optimal stirring speed and time strategy for different fermentation stages, avoid excessive stirring affecting the flavor quality, and achieve efficient, stable and high-quality bamboo wine fermentation.

[0058] 4. The present invention demonstrates remarkable advantages in data optimization and learning. The Kalman filter algorithm performs real-time filtering and state estimation on data collected by ultra-miniaturized sensors, effectively reducing data noise and measurement errors, laying a solid foundation for subsequent precise analysis. The principal component analysis algorithm converts high-dimensional data from multiple sensors into uncorrelated principal component data, accurately extracting characteristic spectral information closely related to the content of bamboo wine flavor compounds. This makes the data more targeted and usable, facilitating a deeper understanding of the inherent laws of the fermentation process and enabling optimized regulation accordingly. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] Figure 1 It is the overall structural diagram of the present invention;

[0060] Figure 2 It is the internal structure diagram of the present invention;

[0061] Figure 3 This is a diagram showing the overall structure of the agitator of the present invention;

[0062] Figure 4 This is a diagram showing the installation structure of the agitator of the present invention;

[0063] Figure 5 This is a schematic diagram of the leaf connection of the present invention;

[0064] Figure 6 It is a component structure diagram of the present invention;

[0065] Figure 7 This is a process flow chart of the fermentation treatment system of the present invention;

[0066] Figure 8 This is a simulation change diagram of the present invention.

[0067] Legend:

[0068] 1. Main control equipment; 2. Fermentation tank; 3. Tank cover; 4. Stirring motor; 5. Agitator; 6. Ultra-miniaturized sensor; 7. Inlet; 8. Heating device; 9. Refrigeration device; 10. Leaf; 11. Leaf cover; 12. Ventilation port; 13. Breathing valve; 14. Output shaft; 15. Winding assembly; 16. Pull rope; 17. Embedded column; 18. Connecting column; 19. Slide; 20. Slide; 21. Push cylinder; 22. Stop block. DETAILED DESCRIPTION

[0069] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention. Specific embodiment one:

[0071] like Figure 1-8 As shown, a fermentation processing equipment and process for bamboo wine data optimization learning, including a main control device 1 and a fermentation tank 2, the main control device 1 is installed on one side of the fermentation tank 2, the main control device 1 includes a fermentation processing system, a tank cover 3 is installed on the top of the fermentation tank 2, a stirring motor 4 is installed in the middle of the top of the tank cover 3, the bottom end of the stirring motor 4 is connected to the output shaft 14, the bottom end of the output shaft 14 is installed with a stirrer 5, and winding components 15 are installed on both sides, front and back of the stirrer 5. The wire assembly 15 is driven by a motor, and a pull rope 16 is wrapped around the outside of the winding assembly 15. Embedded columns 17 are welded on both sides of the outside of the agitator 5 and on the front and back. A sliding opening 20 is opened on the surface of the embedded column 17. The shape of the sliding opening 20 is an arc. A slide 19 is slidably connected in the sliding opening 20. The slide 19 and the sliding opening 20 are adapted to each other. A connecting column 18 is embedded on one side of the slide 19. A leaf 10 is clamped on one side of the connecting column 18. One end of the pull rope 16 is connected to the top of the leaf 10.

[0072] The outside of the leaf 10 is slidably connected to the leaf sleeve 11, and a pushing cylinder 21 is installed between the leaf sleeve 11 and the leaf 10. Block blocks 22 are provided on the top and bottom of one side of the leaf 10. An ultra-miniaturized sensor 6 is installed inside the fermentation tank 2. A feeding port 7 is provided on one side of the top of the tank cover 3. An automatic feeding device will be provided outside the feeding port 7. The automatic feeding device here can select the existing 304 powder closed automatic feeding device produced by Xinxiang Jianyi Intelligent Equipment Co., Ltd. A heating device 8 is installed on one side of the top of the tank cover 3, and a refrigeration device 9 is installed on the other side of the top of the tank cover 3. A vent 12 is provided on one side of the top of the fermentation tank 2, and a breathing valve 13 is connected to the vent 12.

[0073] The main control device 1 is installed on one side of the fermentation tank 2 and controls the entire fermentation tank 2 through the internal fermentation processing system. Figure 3-5As shown, stirring blades are installed at four positions outside the agitator 5. The complete stirring blade is composed of a blade 10 and a blade sleeve 11. The pushing cylinder 21 on one side of the blade 10 will push the external blade sleeve 11 to slide outward to adjust the length of the entire stirring blade. At the same time, the blocking block 22 on the blade 10 is used to prevent the blade sleeve 11 from falling off. A mouth that is adapted to the blocking block 22 will be provided inside the blade sleeve 11 to limit the position. The pull rope 16 can be reeled in by the winding assembly 15 in the agitator 5. The winding assembly 15 is driven by a motor. When the winding assembly 15 reels in the pull rope 16, the bottom end of the pull rope 16 will pull the blade 10 upward. One side of the blade 10 slides in the sliding mouth 20 through the slide 19, which can indirectly adjust the stirring angle of the entire stirring blade. Specific embodiment two:

[0075] like Figure 1-8 As shown, the following is the bamboo wine fermentation processing equipment and process working principle combined with the system and equipment:

[0076] 1. Data collection and monitoring stage

[0077] The ultra-miniaturized sensors 6 (including temperature, humidity, gas, and pH sensors) installed inside the fermentation tank 2 begin collecting internal environmental data from the growth stage. Once used for winemaking and placed in the fermentation tank 2, they continuously monitor the fermentation environment. The collected data is transmitted to the fermentation intelligent sensing and monitoring system in the main control device 1. Within this system, the Kalman filter algorithm performs real-time filtering and state estimation on the data based on the state transition matrix constructed from the physical characteristics of the growth and fermentation process. It combines the measurement matrix with the actual measurement values to accurately calculate the Kalman gain, thereby reducing data noise and measurement errors and ensuring high data reliability. Subsequently, the principal component analysis algorithm calculates the covariance matrix of the sensor data, and by solving the eigenvalues and eigenvectors to form a projection matrix, the high-dimensional data from multiple sensors is converted into uncorrelated principal component data, so that characteristic spectral information closely related to the content of bamboo wine flavor substances can be accurately extracted from the spectral analysis sensor data. At the same time, the blockchain-based data encryption and sharing unit uses the elliptic curve encryption algorithm to encrypt sensor data to ensure the security and confidentiality of data during transmission and storage, and then uses the distributed hash table algorithm to build the data storage and index structure of the blockchain network, so that various departments within the enterprise and upstream and downstream industry chain partners can efficiently store, query, share and collaborate on data.

[0078] 2. Fermentation Condition Adjustment Stage

[0079] Temperature control

[0080] The heating device 8 and cooling device 9 installed on the top cover 3 of the fermentation tank 2 are controlled by the temperature adaptive control unit in the automated fermentation condition control system. The fuzzy logic control algorithm first formulates a preliminary control strategy based on the specific temperature requirements of different fermentation stages and real-time temperature sensor data. Specifically, specific fuzzy sets are defined for the temperature error (the difference between the set temperature and the actual temperature) and the temperature error change rate (the speed of temperature change), such as {negative large (NB), negative medium (NM), zero (ZO), positive medium (PM), positive large (PB)}, and corresponding membership functions are determined. This function can accurately describe the degree to which a specific temperature error or error change rate belongs to a fuzzy set. Based on these, a fuzzy rule table is constructed. For example, in the early stages of fermentation, when the temperature error is negative and the error change rate is also negative, the corresponding control output (cooling or heating intensity) is positive, indicating that a higher heating intensity is required to increase the temperature. When the actual temperature data is input, its membership in each fuzzy set is determined according to the membership function, and then accurate fuzzy reasoning is performed according to the fuzzy rule table to obtain the fuzzy control output. The center of gravity method and other methods are then used to convert the fuzzy control output into an actual control signal, thereby preliminarily controlling the fermentation temperature.

[0081] The model predictive control algorithm establishes a dynamic model of bamboo wine fermentation temperature that comprehensively considers multiple factors, including heat production from microbial metabolism, heat conduction in the fermentation tank, and heat exchange in the cooling and heating equipment. Based on the state estimates provided by the Kalman filter algorithm, it predicts temperature changes over a period of time, for example, predicting temperature trends within the next 1-2 hours. A performance indicator function is then established that comprehensively considers temperature tracking error (the deviation between the actual and set temperature) and the cost of control inputs (such as the energy consumption of the cooling and heating equipment). Constraints such as the power limit and temperature change rate limit of the cooling and heating equipment are also considered. A specific optimization algorithm is used to solve the control sequence that minimizes the performance indicator function under these constraints. This is the optimal control variable sequence for the cooling or heating equipment over the next period of time. Finally, the optimized control signal obtained by the model predictive control algorithm is combined with the preliminary control signal obtained by the fuzzy logic control algorithm to comprehensively determine the final control signal for the cooling or heating equipment. This achieves precise dynamic temperature regulation at each fermentation stage. The temperature is precisely controlled at 25-28°C during the yeast activation stage, stabilized at 30-32°C during the main fermentation stage, and slowly dropped to 18-20°C during the aging stage, with a temperature regulation accuracy of ±0.5°C.

[0082] Humidity and ventilation control

[0083] The proportional-integral-derivative (PID) algorithm calculates the humidity error (the difference between the set humidity and the actual humidity) based on real-time feedback data from the humidity sensor and the ventilation volume sensor, allowing the spray humidifier or ventilation equipment to respond quickly to change the humidity. If the humidity suddenly rises, the differential term will output a large negative signal, prompting the ventilation equipment to increase the ventilation volume and quickly reduce the humidity.

[0084] The particle swarm optimization algorithm uses the proportional coefficient, integral coefficient and differential coefficient of the PID algorithm as the position vector of the particle, and connects an external breathing valve 13 to the ventilation port 12 on the top side of the fermentation tank 2. Its ventilation volume is related to humidity regulation. The particle swarm optimization algorithm initializes a group of particles, each particle represents a set of PID parameter combinations, and assigns each particle a random velocity vector. The fitness value of each particle is calculated, and the fitness function is set according to the effect of the coordinated control of humidity and ventilation. For example, factors such as humidity control accuracy, the degree of matching between ventilation volume and microbial oxygen demand can be considered. Each particle updates its own speed and position according to its own historical optimal position and the historical optimal position of the group, as well as its own speed, through specific speed update rules and position update rules. Repeat the above steps until the preset number of iterations is reached or the convergence condition is met. The PID parameters corresponding to the historical optimal position of the group are the optimized parameters. Using the optimized PID parameters, the ventilation volume can be accurately adjusted within 0.1-1m according to the microbial oxygen demand and the fermentation stage. 3 / (h·m 3 ) range and maintain the synergistic control effect of humidity in the ideal range of 60%-80%, optimizing gas exchange and water balance.

[0085] 3. Microbial Community Regulation Stage

[0086] Prediction of microbial community dynamics

[0087] The microbial community dynamics prediction unit within the precise fermentation microbial community control system comes into play. A long-short-term memory network (LSTM) algorithm collects microbial community data at different time points during the bamboo wine fermentation process, including information on the types and abundance of various microorganisms, to construct a time-series dataset. This dataset is fed into the LSTM network, and the network's memory units learn how the microbial community structure and abundance evolve at different fermentation stages, capturing the long-term dependencies of microbial community dynamics. For example, the memory units can remember how the growth trend of a particular microorganism in the early stages affects the growth of other microorganisms in the later stages, thereby predicting the changing trends of the microbial community over time.

[0088] The Bayesian network algorithm constructs a network model of microbial interactions based on microbial gene sequencing and metagenomic analysis data. This model describes the causal relationships and conditional probability distributions between microorganisms. For example, the presence of a certain aroma-producing microorganism may promote the growth of another microorganism while inhibiting the reproduction of still others. These relationships are represented by nodes (representing microorganisms), edges (representing interactions), and conditional probability tables in the Bayesian network. The temporal changes in the microbial community predicted by the LSTM algorithm are input into the Bayesian network model, and combined with the interaction relationships in the network, the future dynamic changes of the microbial community can be more accurately predicted. For example, the prediction of the population trend of a key microbial strain under specific fermentation conditions and its impact on other microorganisms provides a decision-making basis for the microbial intelligent delivery system, allowing the precise delivery of specific microbial strains or bacterial community combinations at key nodes to achieve targeted regulation and enhancement of the bamboo wine flavor.

[0089] Microbial niche construction and optimization

[0090] The microbial niche construction and optimization unit uses a genetic algorithm to encode combinations of microbial growth promoters (such as vitamins, amino acids, and other nutrients) and inhibitors (such as natural antimicrobial substances and biopreservatives) as chromosomes. For example, the gene positions of a chromosome can represent the type and dosage of different nutrients or inhibitors. A population is initialized, with each individual representing a specific addition combination. A fitness function is defined, with the growth and reproduction rate of beneficial microorganisms and the inhibitory effect on harmful microorganisms as optimization objectives. For example, the fitness value can be calculated by measuring the growth rate of beneficial microorganisms and the decrease rate of harmful microorganisms during fermentation. The population is iteratively evolved through selection operations (such as roulette wheel selection, where the probability of selection is determined based on the individual's fitness value), crossover operations (such as single-point or multi-point crossover, where chromosomes of two individuals are exchanged to generate new individuals), and mutation operations (such as randomly changing a gene position on a chromosome to introduce new genetic information). After multiple generations of evolution, the individual with the highest fitness value is obtained, and its corresponding addition combination is the optimal addition combination of microbial growth promoters and inhibitors.

[0091] The NSGA-II multi-objective optimization algorithm is similar to the genetic algorithm. First, the addition combination of microbial growth promoters and inhibitors is encoded and the population is initialized. Two objective function values are calculated for each individual in the population, namely the growth and reproduction rate of beneficial microorganisms and the inhibitory effect of harmful microorganisms. According to the objective function values, the population is non-dominated and sorted, and the individuals are divided into different non-dominated levels (i.e., Pareto frontiers). The non-dominated level means that an individual cannot be surpassed by other individuals on a certain goal without compromising other goals. Within each non-dominated level, the crowding distance of the individual is calculated. This distance reflects the density of individuals around the individual and is used to maintain the diversity of the population. The population is evolved through selection, crossover, and mutation operations. The selection operation is based on the non-dominated level and crowding distance, and individuals with high non-dominated levels and large crowding distances are given priority. Repeat the above steps until the preset number of iterations is reached or the convergence condition is met, and finally a set of Pareto optimal solutions are obtained. These solutions represent the best strategy for balancing the two goals of beneficial microbial growth and reproduction and the inhibition of harmful microorganisms. The best strategy for regulating the ecological balance of the microbial community is determined, a microbial ecological niche conducive to bamboo wine fermentation is constructed, and the stability and healthy development of the microbial community is maintained.

[0092] 4. Mixing stage

[0093] Adaptive adjustment of stirring blade shape and angle

[0094] The stirring motor 4 is mounted at the top middle of the tank cover 3 at the top of the fermentation tank 2. The bottom end of its output shaft 14 is connected to the stirrer 5. Winding assemblies 15 are installed on both sides, front and back of the stirrer 5. The winding assembly 15 is driven by the motor. A pull rope 16 is wrapped around the outside of the winding assembly 15, and one end of the pull rope 16 is connected to the top of the blade 10. In the stirring blade shape and angle adaptive adjustment unit of the fermentation stirring and mixing system, a neural network algorithm collects a large amount of data on the correspondence between the material concentration, viscosity and fermentation stage in the fermentation tank and the optimal shape and angle of the stirring blade, including the optimal blade shape and angle data under different fermentation stages (such as early, middle and late fermentation) and different material states (such as thin and thick). This data is divided into a training set, a validation set and a test set to construct a neural network model. The input layer of the model includes material concentration, viscosity and fermentation stage information, and the output layer is the shape and angle parameters of the stirring blade. The training set is used to train the neural network. By adjusting the weights and bias of the neural network, the model can learn the accurate mapping relationship between input and output. During the training process, the validation set is used to monitor the performance of the model to prevent overfitting. After the training is completed, the real-time material concentration, viscosity and fermentation stage data are input into the trained neural network model. The model predicts the optimal shape and angle parameters of the stirring blade under the current working conditions. At this time, the winding assembly 15 drives the motor to reel in or release the pull rope 16 according to the prediction results. The pull rope 16 pulls the top of the leaf 10, so that one side of the leaf 10 slides in the sliding port 20 through the slide 19, thereby indirectly adjusting the stirring angle of the entire stirring blade. At the same time, the outer part of the leaf 10 is slidably connected to the blade sleeve 11. A pushing cylinder 21 is installed between the blade sleeve 11 and the leaf 10. The pushing cylinder 21 will push the outer blade sleeve 11 to slide outward or inward according to the prediction results, thereby adjusting the length of the entire stirring blade, realizing adaptive adjustment of the shape and angle of the stirring blade, and meeting the stirring requirements of different fermentation stages. For example, in the early stage of fermentation, the blade is flattened and expanded to achieve rapid and uniform mixing, and in the middle stage, the blade is bent into a spiral shape to increase the stirring force and shear force, thereby improving the fermentation efficiency.

[0095] Intelligent optimization of stirring speed and time

[0096] The reinforcement learning algorithm uses the speed adjustment of the stirring motor 4 and the stirring time control as the action space. For example, the stirring speed can be adjusted within a certain range (e.g., 30-120 rpm), and the stirring interval can be set to different time periods (e.g., 10 minutes to 3 hours). The reward function uses bamboo wine fermentation efficiency (e.g., fermentation completion time, raw material conversion rate, etc.) and flavor quality (e.g., flavor compound content, taste score, etc.) as indicators. For example, if fermentation efficiency improves and flavor quality is good, a higher reward is given; if fermentation efficiency is low or flavor quality deteriorates, a lower reward or even a penalty is given. The agent (representing the stirring control strategy) continuously experiments during the fermentation process, selecting different stirring speed and time combinations as actions, observing environmental feedback (i.e., changes in various indicators during the fermentation process), and receiving rewards based on the reward function. Through the specific learning method in the reinforcement learning algorithm, the agent continuously learns and optimizes its strategy, exploring the optimal stirring speed and time strategy for different fermentation stages.

[0097] The dynamic programming algorithm divides the entire fermentation process into multiple stages (e.g., early, mid, and late fermentation), treating each stage as a subproblem. Based on the Bellman optimality principle, the algorithm recursively solves the problem from the last stage (late fermentation) forward. At each stage, the state of the current stage (such as the current material state, microbial community state), the actions that can be taken (stirring speed and time adjustment), and the cost of the actions (such as energy consumption, potential impact on flavor quality) and benefits (such as improved fermentation efficiency and increased flavor substance production) are considered. The optimal strategy for each stage (i.e., the optimal stirring speed and time control strategy) is calculated through a dynamic programming algorithm, and the optimal solutions of these sub-problems are combined to obtain the global optimal stirring speed and time control strategy. For example, in the early stage of fermentation, the stirring speed is controlled at 50-80 rpm, and stirring is carried out for 15-20 minutes every 2-3 hours; in the middle stage of fermentation, the stirring speed is increased to 80-120 rpm, and the stirring time interval is adjusted to 1-2 hours; in the late stage of fermentation, the stirring speed is reduced to 30-50 rpm, and stirring is carried out 1-2 times a day, each time for 5-10 minutes, to avoid excessive stirring affecting the flavor quality of bamboo wine, and to achieve high efficiency, stability and high quality of the bamboo wine fermentation process.

[0098] 5. Intelligent Control and Collaboration Stage

[0099] The intelligent decision-making engine of the fermentation intelligent control system combines deep neural networks and convolutional neural networks. The deep neural network processes sensor time series data, learning the changing patterns of flavor compounds, temperature, humidity, and gas composition during bamboo wine fermentation, and automatically adjusting fermentation control parameters based on flavor targets. The convolutional neural network extracts spatial features from spectral sensor data to optimize fermentation environment control decisions. The real-time feedback adjustment unit combines incremental learning algorithms with adaptive control algorithms to dynamically adjust fermentation conditions based on real-time fermentation feedback, ensuring that the production and concentration of bamboo wine flavor compounds remain within preset ranges. The intelligent prediction and simulation module of the bamboo wine data optimization learning system integrates adaptive genetic algorithms and multi-task learning algorithms. Based on temperature, humidity, gas composition, and microbial community data, it constructs a fermentation mathematical model in real time to predict flavor evolution trends. The adaptive genetic algorithm optimizes model parameters at each fermentation stage to improve accuracy, while the multi-task learning algorithm shares knowledge learned from different tasks, enabling the model to achieve global optimization across temperature and humidity control, flavor optimization, and microbial community regulation, providing accurate prediction and optimization solutions for fermentation. Finally, the fermentation intelligent sensing and monitoring system, the automated fermentation condition adjustment system, the fermentation microbial community precision control system, the fermentation stirring and mixing system, and the fermentation intelligent control system work together to comprehensively and intelligently control and optimize the bamboo wine fermentation process, thereby improving the quality stability, flavor uniqueness, and production efficiency of the bamboo wine.

[0100] It should be noted that the present invention mainly improves the system control aspects and the overall process of the fermentation equipment. Some common equipment such as the heating device 8, the refrigeration device 9, the stirring motor 4 and the breathing valve 13 are not described in detail here. The heating device 8 here uses an electric heating rod or a steam heating coil, and the refrigeration device 9 uses an air-cooled refrigerator or a water-cooled refrigerator. The installation position shown in the figure is not fixed, and the position setting can be selected according to actual needs. Specific embodiment three:

[0102] like Figure 1-8 As shown, the key algorithms mentioned in the above embodiment are analyzed in detail below:

[0103] 1. Kalman filter algorithm

[0104] Kalman filtering is used to remove noise from sensor data and perform state estimation. Its core goal is to predict and correct the state of the system through a linear model.

[0105] The basic equation of Kalman filter is:

[0106] Prediction steps:

[0107]

[0108] in, is the predicted state at time k, A is the state transfer matrix, is the state estimate of the previous moment, B is the control matrix, u k is the control input, is the forecast error covariance, and Q is the process noise covariance.

[0109] Update steps:

[0110]

[0111] Among them, K k is the Kalman gain, H is the observation matrix, R is the observation noise covariance, z k is the actual observed value, is the updated state estimate, P k is the updated error covariance.

[0112] 2. Principal Component Analysis (PCA)

[0113] PCA is used to reduce the dimensionality of multidimensional data, extract the main features in the data, reduce redundancy and improve computational efficiency.

[0114] The core formula of PCA is:

[0115] Covariance matrix calculation:

[0116]

[0117] Where C is the covariance matrix, X is the standardized data matrix, and n is the number of samples.

[0118] Eigenvalues and eigenvectors:

[0119] Cv=λv

[0120] Here, v is the eigenvector and λ is the corresponding eigenvalue.

[0121] Data after dimensionality reduction:

[0122] Z=XV

[0123] Among them, Z is the data after dimensionality reduction, and V is the matrix composed of all eigenvectors.

[0124] PCA effectively reduces the dimension of the data and enhances the interpretability of the data by extracting the most significant directions (principal components) in the data.

[0125] 3. Elliptic Curve Cryptography (ECC)

[0126] ECC is used to ensure data encryption and secure transmission, and mainly achieves security through the elliptic curve discrete logarithm problem.

[0127] The core formula of ECC:

[0128] Elliptic curve equation:

[0129] y 2 =x 3 +ax+b

[0130] Where a and b are the parameters of the curve, and x and y are the coordinates of the point on the curve.

[0131] Public and private key generation:

[0132] P=dG

[0133] Among them, P is the public key, d is the private key, and G is the base point of the elliptic curve.

[0134] ECC provides a highly secure encryption method with a relatively small key size by performing addition and multiplication operations on elliptic curves.

[0135] 4. Distributed Hash Table (DHT)

[0136] DHT is used to store and share blockchain data, ensuring decentralized and efficient data storage and query.

[0137] The core formula of DHT:

[0138] Node hash function:

[0139] h(k)=(k mod N)

[0140] Where k is the key value, N is the total number of nodes, and h(k) is the position of key value k in the hash table.

[0141] DHT allows for efficient storage and retrieval of data in a decentralized environment, ensuring data consistency through hashing algorithms.

[0142] 5. Fuzzy Logic Control (FLC)

[0143] Fuzzy logic control is used to process fuzzy inputs of a system and generate precise control outputs. Decisions are made through fuzzy sets and fuzzy reasoning:

[0144] FLC's core formula:

[0145] Fuzzification:

[0146]

[0147] Among them, μ A (x) is the membership function, c is the center location, d is the width parameter, and x is the input variable.

[0148] Reasoning process:

[0149] z=∑wiμi(x)

[0150] Among them, w i is the weight, μ i (x) is the membership of the fuzzy set, z is the output, and fuzzy logic control is used to deal with uncertainty and fuzziness and can generate flexible control strategies.

[0151] 6. Particle Swarm Optimization (PSO)

[0152] PSO is used to optimize the coordinated control of humidity and ventilation, and finds the optimal control parameters by simulating the search behavior of particle swarms:

[0153] The core formula of PSO:

[0154] Particle update formula:

[0155] v i =wv i-1 +c1r1(p i -x i )+c2r2(gx i )

[0156] x i =x i-1 +v i

[0157] Among them, v i is the velocity of the particle, x i is the position of the particle, p i is the individual optimal position of the particle, g is the global optimal position, r1, r2 are random numbers, c1, c2 are acceleration constants, and w is the inertia weight. PSO simulates the flight of particles in the search space to find the optimal solution and optimize humidity and ventilation control through group collaboration and individual exploration. Specific embodiment four:

[0159] like Figure 1-8 As shown, the following is a description of the specific application logic steps of each module and algorithm in the fermentation processing equipment and process of bamboo wine data optimization learning:

[0160] 1. Fermentation intelligent sensing and monitoring system

[0161] Function: The main task of the system is to collect multi-dimensional data of the fermentation process in real time, and perform data processing, feature extraction and encrypted sharing.

[0162] Application steps:

[0163] Ultra-miniaturized sensors (such as temperature, humidity, gas sensors, pH sensors, etc.) are used to collect environmental data in the fermentation tank (2) in real time.

[0164] The Kalman filter algorithm is used to filter the high-dimensional time series data collected by the sensor, remove noise and perform state estimation to ensure data accuracy and continuity.

[0165] The principal component analysis (PCA) algorithm was used to reduce the dimensionality of the collected multidimensional data and extract the main features related to the flavor substances of bamboo wine (such as alcohols, esters, etc.).

[0166] Sensor data is encrypted using elliptic curve cryptography to ensure data security. At the same time, the encrypted data is stored and shared on the blockchain network using a distributed hash table algorithm, enabling efficient collaboration within the enterprise and across the industry chain.

[0167] 2. Automated fermentation condition adjustment system

[0168] Function: The system automatically adjusts the temperature, humidity and ventilation during the fermentation process based on real-time data to ensure that the fermentation environment meets the growth requirements of microorganisms.

[0169] Application steps:

[0170] Temperature regulation: Based on the fuzzy logic control algorithm, a preliminary control strategy is designed according to the temperature requirements of different fermentation stages and real-time sensor data (such as temperature sensor feedback).

[0171] Model Predictive Control: Using the Model Predictive Control (MPC) algorithm, the bamboo wine fermentation temperature dynamic model is used to predict future temperature changes, and the fuzzy logic control strategy is combined to optimize temperature regulation to ensure precise temperature control during the fermentation process.

[0172] Coordinated control of humidity and ventilation: PID algorithm is used to perform conventional humidity adjustment control based on real-time data from humidity sensors and ventilation volume sensors.

[0173] Particle Swarm Optimization (PSO): The parameters in the PID algorithm are optimized using the particle swarm optimization algorithm to accurately adjust the ventilation rate according to the oxygen demand of the microorganisms and the fermentation stage, ensure that the humidity is in the range of 60%-80%, and optimize gas exchange and water balance.

[0174] 3. Precision control system of fermentation microbial communities

[0175] Function: The system precisely regulates the microbial community to ensure the microbial ecological balance during the fermentation process, thereby promoting the formation of bamboo wine flavor.

[0176] Application steps:

[0177] Prediction of microbial community dynamics:

[0178] Long Short-Term Memory (LSTM) Network: Use LSTM networks to analyze the time series data of microbial communities during bamboo wine fermentation and learn the evolution patterns of microbial communities.

[0179] Bayesian network: Based on microbial gene sequencing and metagenomic data, the Bayesian network algorithm is used to construct an interaction network model between microorganisms to predict the dynamic changes of microbial communities.

[0180] Microbial niche construction and optimization:

[0181] Genetic algorithm: Genetic algorithm is used to optimize the addition combination of microbial growth promoters (such as vitamins, amino acids) and inhibitors (such as natural antibacterial substances).

[0182] NSGA-II multi-objective optimization algorithm: The NSGA-II multi-objective optimization algorithm is used to balance the growth of beneficial microorganisms and the inhibition of harmful microorganisms, ensure the ecological balance of the microbial community, and construct a microbial niche suitable for bamboo wine fermentation.

[0183] 4. Fermentation stirring and mixing system

[0184] Function: The system optimizes the mixing effect during the fermentation process by intelligently adjusting the shape, angle and stirring speed of the stirring blades to ensure the flavor and quality of the bamboo wine.

[0185] Application steps:

[0186] Blade shape and angle adjustment:

[0187] Neural Network Algorithm: Use a neural network algorithm to learn and predict the optimal blade shape and angle based on the material concentration, viscosity and fermentation stage data in the fermenter.

[0188] Finite Element Analysis: Finite Element Analysis (FEA) is used to simulate the mechanical properties of the blades at different fermentation stages to ensure that the blades can automatically adjust their shape and angle according to the material state to improve stirring efficiency.

[0189] Stirring speed and time control:

[0190] Reinforcement learning algorithm: Through the reinforcement learning algorithm, the stirring motor speed and stirring time are optimized, and the reward function is designed based on the bamboo wine fermentation efficiency and flavor quality indicators to explore the optimal stirring strategy.

[0191] Dynamic programming algorithm: Use the dynamic programming algorithm to optimize the strategy in the reinforcement learning process and determine the globally optimal stirring speed and time control strategy to avoid the impact of excessive stirring on the flavor of bamboo wine.

[0192] 5. Fermentation intelligent control system

[0193] Function: The system intelligently adjusts and optimizes all parameters in the fermentation process in real time to ensure the stability of bamboo wine flavor and product quality.

[0194] Application steps:

[0195] Intelligent decision-making engine:

[0196] Deep Neural Network (DNN): A deep neural network is used to process time series data from sensors, learn the changing patterns of bamboo wine flavor substances, temperature, humidity, and gas composition during the fermentation process, and automatically adjust various fermentation control parameters according to the target flavor.

[0197] Convolutional Neural Network (CNN): A convolutional neural network is used to extract spatial features from spectral sensor data collected during the fermentation process to optimize the regulation of the fermentation environment and ensure the accurate generation of bamboo wine flavor substances.

[0198] Real-time feedback adjustment:

[0199] Incremental learning algorithm: The incremental learning algorithm receives feedback information in real time and dynamically adjusts fermentation conditions (such as temperature, humidity, etc.) to ensure that the concentration of flavor substances in bamboo wine remains within the preset range.

[0200] Adaptive control algorithm: Through the adaptive control algorithm, the fermentation environment is automatically adjusted according to the real-time changing fermentation data to further ensure the flavor and quality of bamboo wine.

[0201] 6. Bamboo Wine Data Optimization Learning System

[0202] Function: The system is responsible for predicting and optimizing the fermentation process of bamboo wine based on data from different stages of the fermentation process, thereby improving the stability and uniqueness of the flavor.

[0203] Application steps:

[0204] Intelligent prediction and simulation:

[0205] Adaptive genetic algorithm (AGA): The adaptive genetic algorithm is used to optimize the model parameters based on the temperature, humidity, gas composition and other data during the bamboo wine fermentation process to improve the prediction accuracy.

[0206] Multi-task learning algorithm: The multi-task learning algorithm achieves global optimization effects during the bamboo wine fermentation process by sharing learning knowledge from different tasks, and provides optimization solutions for flavor optimization, temperature and humidity control, and microbial community regulation at different fermentation stages. Specific embodiment five:

[0208] like Figure 1-8 As shown, based on the above content, the detailed hardware composition and hardware in each module are described:

[0209] Precision control system for fermentation microbial communities

[0210] Hardware composition:

[0211] Microbiological analysis instruments

[0212] Flow cytometer: used to detect the types and quantities of microorganisms in real time during the fermentation process and analyze the dynamic changes of microbial communities.

[0213] Microbial gene sequencing equipment: used to sequence the genes of microorganisms in the bamboo wine fermentation process and analyze their contribution to flavor substances.

[0214] Environmental control equipment

[0215] Temperature control equipment: adjust the temperature according to the needs of the microbial community.

[0216] Oxygen supply system: Provides appropriate oxygen concentration for specific microbial communities to ensure optimal growth of microorganisms.

[0217] Hardware features:

[0218] Through the microbial monitoring module, combined with flow cytometers, gene sequencers and other equipment, the dynamic changes of microbial communities can be monitored in real time, and fermentation conditions can be adjusted based on feedback.

[0219] Microbial regulator: By controlling the gas concentration and nutrient content in the fermentation tank, the microbial community is precisely adjusted to ensure the flavor characteristics of bamboo wine fermentation. Specific embodiment six:

[0221] like Figure 1-8 As shown in the attached Figure 1 The figure shows the state of the overall structure being placed vertically. Figure 2 This is a horizontal placement state diagram. The fermentation tank 2 can be placed vertically and horizontally at the same time. The usage steps are the same. In actual application, the placement angle can be selected according to application requirements.

[0222] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variants thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or device comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such process, method, article, or device. In the absence of further restrictions, an element defined by the statement "comprising a reference structure" does not exclude the presence of additional identical elements in the process, method, article, or device comprising the element.

[0223] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. A fermentation processing device and process for bamboo wine data optimization learning, characterized by: The invention comprises a main control device (1) and a fermentation tank (2), wherein the main control device (1) is installed on one side of the fermentation tank (2), and the main control device (1) contains a fermentation treatment system. A tank cover (3) is installed on the top of the fermentation tank (2), and a stirring motor (4) is installed in the middle of the top of the tank cover (3). The bottom end of the stirring motor (4) is connected to an output shaft (14) in a transmission manner, and an agitator (5) is installed at the bottom end of the output shaft (14). Winding components (15) are installed on both sides, the front and the back of the agitator (5), and the winding components (15) are driven by a motor. The outside of the winding assembly (15) is wound with a pull rope (16), and both sides of the outside of the agitator (5) and the front and back sides are welded with embedded columns (17). The surface of the embedded column (17) is provided with a sliding opening (20), and the shape of the sliding opening (20) is an arc. A slide (19) is slidably connected in the sliding opening (20), and the slide (19) and the sliding opening (20) are adapted to each other. A connecting column (18) is embedded on one side of the slide (19), and a leaf (10) is clamped on one side of the connecting column (18). One end of the pull rope (16) is connected to the top of the leaf (10).

2. The fermentation processing equipment and process for bamboo wine data optimization learning according to claim 1 is characterized in that: The outer portion of the leaf (10) is slidably connected to the leaf sleeve (11), a pushing cylinder (21) is installed between the leaf sleeve (11) and the leaf (10), a blocking block (22) is provided at the top and bottom of one side of the leaf (10), an ultra-miniaturized sensor (6) is installed inside the fermentation tank (2), a feeding port (7) is provided on one side of the top of the tank cover (3), a heating device (8) is installed on one side of the top of the tank cover (3), a refrigeration device (9) is installed on the other side of the top of the tank cover (3), a vent (12) is provided on one side of the top of the fermentation tank (2), and a breathing valve (13) is externally connected to the vent (12).

3. The fermentation processing equipment and process for bamboo wine data optimization learning according to claim 2, characterized in that: The fermentation processing system includes a fermentation intelligent sensing and monitoring system, an automated fermentation condition adjustment system, a fermentation microbial community precision control system, a bamboo wine data optimization learning system, a fermentation stirring and mixing system, and a fermentation intelligent control system.

4. The fermentation processing equipment and process for bamboo wine data optimization learning according to claim 3 is characterized in that: The fermentation intelligent sensing and monitoring system includes a multi-sensor data fusion and feature extraction unit and a blockchain-based data encryption and sharing unit. The multi-sensor data fusion and feature extraction unit combines a Kalman filter algorithm and a principal component analysis algorithm. The Kalman filter algorithm is used to collect time series data in the fermentation process and is equipped with an ultra-miniaturized sensor (6) for real-time filtering and state estimation. The ultra-miniaturized sensor (6) includes a temperature sensor, a humidity sensor, a gas sensor, and a pH sensor. The principal component analysis algorithm is used to extract features from the high-dimensional data of the sensor, and the generated variables are converted into mutually unrelated principal components for subsequent extraction of characteristic spectrum information related to the content of bamboo wine flavor substances by a spectrum analyzer. The blockchain-based data encryption and sharing unit in the fermentation intelligent sensing and monitoring system combines an elliptic curve encryption algorithm and a distributed hash table algorithm. The sensor data is encrypted by the elliptic curve encryption algorithm. The distributed hash table algorithm is used to construct a data storage and index structure of the blockchain network, thereby realizing efficient storage, query, sharing, and collaborative work of data among various departments within the enterprise and upstream and downstream industry chain partners. The principal component analysis algorithm formula is as follows: Covariance matrix calculation: Where C is the covariance matrix, X is the standardized data matrix, and n is the number of samples; Eigenvalues and eigenvectors: Cv=λv Among them, v is the eigenvector and λ is the corresponding eigenvalue; Data after dimensionality reduction: Z=XV Among them, Z is the data after dimensionality reduction, and V is the matrix composed of all eigenvectors; The elliptic curve encryption algorithm formula is as follows: Elliptic curve equation: y 2 =x 3 +ax+b Where a and b are the parameters of the curve, and x and y are the coordinates of the point on the curve; Public and private key generation: P=dG Among them, P is the public key, d is the private key, and G is the base point of the elliptic curve; The distributed hash table formula is as follows: Node hash function: h(k)=(kmodN) Where k is the key value, N is the total number of nodes, and h(k) is the position of key value k in the hash table.

5. The fermentation processing equipment and process for bamboo wine data optimization learning according to claim 3 is characterized in that: The automated fermentation condition regulation system includes a temperature adaptive control unit and a humidity and ventilation coordinated control unit. The automated fermentation condition regulation system integrates a fuzzy logic control algorithm and a model predictive control algorithm through the temperature adaptive control unit. The fuzzy logic control algorithm formulates a preliminary control strategy based on the fuzzy requirements for temperature in the fermentation stage and real-time temperature sensor data. The model predictive control algorithm predicts future temperature changes based on the bamboo wine fermentation temperature dynamic model, and optimizes and calculates the control signals of the refrigeration and heating equipment in combination with the fuzzy logic control strategy to achieve precise dynamic regulation of the temperature in each fermentation stage. In the yeast activation stage, the temperature is precisely controlled at 25-28°C, stabilized at 30-32°C during the main fermentation period, and slowly dropped to 18-20°C in the later aging stage, and the temperature regulation accuracy needs to reach ±0.5°C. The humidity and ventilation coordinated control unit combines the proportional-integral-differential algorithm and the particle swarm optimization algorithm. The proportional-integral-differential algorithm performs conventional control on the humidity regulation system according to the real-time feedback data of the humidity and ventilation volume sensors. The particle swarm optimization algorithm is used to optimize the proportional-integral-differential algorithm parameters to achieve precise regulation of the ventilation volume at 0.1-1m according to the microbial oxygen demand and the fermentation stage. 3 / (h·m 3 ) range and maintain the synergistic control effect of humidity in the ideal range of 60%-80%, optimizing gas exchange and water balance; The formula of the fuzzy logic control is as follows: Fuzzification: Among them, μ A (x) is the membership function, c is the center position, d is the width parameter, and x is the input variable; Reasoning process: z=∑wiμi(x) Among them, w i is the weight, μ i (x) is the membership of the fuzzy set and z is the output.

6. The fermentation processing equipment and process for bamboo wine data optimization learning according to claim 3 is characterized in that: The fermentation microbial community precision control system is equipped with a microbial community dynamic prediction unit and a microbial niche construction and optimization unit. The microbial community dynamic prediction unit integrates the long-short-term memory network algorithm and the Bayesian network algorithm. The long-short-term memory network algorithm is used to analyze the learning and evolution laws of the microbial community time series data during the bamboo wine fermentation process. The Bayesian network algorithm constructs a microbial interaction network model based on microbial gene sequencing and metagenomic analysis data. The two are combined to predict the future dynamic changes of the microbial community and provide a decision-making basis for the microbial intelligent delivery system. The microbial niche construction and optimization unit adopts a genetic algorithm and an NSGA-I multi-objective optimization algorithm. The genetic algorithm searches for the optimal addition combination scheme of microbial growth promoters and inhibitors. The microbial growth promoters include vitamins, amino acids and nutrients, and the microbial inhibitors include natural antibacterial substances and biological preservatives. The NSGA-II multi-objective optimization algorithm processes the target optimization problem, weighs the growth and reproduction of beneficial microorganisms and the inhibition of harmful microorganisms, determines the optimal strategy for regulating the ecological balance of the microbial community, and constructs a microbial niche that is conducive to bamboo wine fermentation.

7. The fermentation processing equipment and process for bamboo wine data optimization learning according to claim 3 is characterized in that: The fermentation stirring and mixing system includes a stirring blade shape and angle adaptive adjustment unit and a stirring speed and time intelligent optimization unit, the stirring blade shape and angle adaptive adjustment unit combines a neural network algorithm and a finite element analysis algorithm, and uses the neural network algorithm to learn the corresponding relationship data of the material concentration, viscosity and fermentation stage in the fermentation tank and the optimal shape and angle of the stirring blade to establish a prediction model, the finite element analysis algorithm simulates and analyzes the mechanical properties of the blade according to the neural network prediction parameters, and ensures that the blade automatically adjusts the shape and angle according to the material state, and the stirring speed and time intelligent optimization unit integrates a reinforcement learning algorithm and a dynamic programming algorithm, and the reinforcement learning algorithm adjusts the speed of the stirring motor (4) and stirring time control as the action space, and bamboo wine fermentation efficiency and flavor quality indicators as reward functions to explore the optimal strategy in the fermentation stage. The dynamic programming algorithm optimizes the strategy in the reinforcement learning process to determine the global optimal stirring speed and time control strategy. In the early stage of fermentation, the stirring speed is controlled at 50-80 rpm, and stirring is carried out for 15-20 minutes every 2-3 hours; in the middle stage of fermentation, the stirring speed is increased to 80-120 rpm, and the stirring time interval is adjusted to 1-2 hours; in the late stage of fermentation, the stirring speed is reduced to 30-50 rpm, and stirring is carried out 1-2 times a day, each time for 5-10 minutes, to avoid excessive stirring affecting the flavor quality of bamboo wine, and to achieve efficient, stable and high-quality fermentation.

8. The fermentation processing equipment and process for bamboo wine data optimization learning according to claim 3 is characterized in that: The fermentation intelligent control system includes an intelligent decision-making engine based on multimodal data fusion and a real-time feedback adjustment unit. The intelligent decision-making engine combines deep neural networks and convolutional neural networks. The deep neural network processes time series data from sensors to learn the changing patterns of bamboo wine flavor substances, temperature and humidity, and gas composition during the fermentation process, and automatically adjusts various control parameters in the fermentation process according to the flavor target. The convolutional neural network is used to extract spatial features from spectral sensor data during the fermentation process to optimize the fermentation environment adjustment decision. The real-time feedback adjustment unit combines incremental learning algorithms and adaptive control algorithms to dynamically adjust fermentation conditions according to real-time feedback information during the bamboo wine fermentation process to ensure that the generation and concentration of bamboo wine flavor substances are maintained within a preset range.

9. The fermentation processing equipment and process for bamboo wine data optimization learning according to claim 3, characterized in that: The bamboo wine data optimization learning system includes an intelligent prediction and simulation module, which integrates an adaptive genetic algorithm and a multi-task learning algorithm. During the bamboo wine fermentation process, a mathematical model of the fermentation process is constructed in real time based on data such as temperature, humidity, gas composition, and microbial community, and the flavor evolution trend during the bamboo wine fermentation stage is predicted. The adaptive genetic algorithm is used to adaptively optimize model parameters in each fermentation stage to improve prediction accuracy. The multi-task learning algorithm shares learning knowledge from different tasks, enabling the model to achieve global optimization effects in temperature and humidity control, flavor optimization, and microbial community regulation, thereby providing accurate prediction and optimization solutions for the bamboo wine fermentation process.

10. The fermentation processing equipment and process for bamboo wine data optimization learning according to claim 3, characterized in that: The fermentation intelligent sensing and monitoring system, the automated fermentation condition adjustment system, the fermentation microbial community precision control system, the fermentation stirring and mixing system, and the fermentation intelligent control system work together to perform comprehensive intelligent control and optimization of the bamboo wine fermentation process, thereby improving the quality stability, flavor uniqueness, and production efficiency of the bamboo wine. The specific process is as follows: Sp1. First, the main control device 1 is installed on one side of the fermentation tank 2, and the tank cover 3 is ensured to be well sealed with the fermentation tank 2. The processed raw materials and the required brewing raw materials are fed into the fermentation tank 2 through the feeding port 7 using the automatic feeding device, and the fermentation process is ready to start. The ultra-miniaturized sensor 6 installed inside the fermentation tank 2 starts to collect environmental data after the raw materials are added, and transmits it to the fermentation intelligent sensing and monitoring system in the main control device 1. The fermentation intelligent sensing and monitoring system uses the Kalman filter algorithm and the principal component analysis algorithm to process the data, and ensures data security and sharing through the data encryption and sharing unit based on the blockchain: Sp2, heating device 8 and refrigeration device 9 are controlled by the temperature adaptive control unit of the automated fermentation condition regulation system. The fuzzy logic control algorithm formulates a preliminary strategy based on the fermentation stage and real-time temperature data. The model predictive control algorithm establishes a temperature dynamic model and predicts future temperature changes. The combination of the two determines the final control signal of the refrigeration or heating equipment, accurately controlling the yeast activation stage at 25-28°C, the main fermentation stage at 30-32°C, and the late aging stage at 18-20°C. Sp3, the microbial community dynamics prediction unit in the fermentation microbial community precision control system uses the LSTM algorithm to collect microbial community data to construct a time series dataset and predict change trends. The Bayesian network algorithm constructs an interaction network model based on data such as microbial gene sequencing. The combination of the two can more accurately predict the future dynamics of the microbial community and provide a basis for intelligent microbial delivery. Sp4 and the stirring motor 4 drive the stirrer 5 to work. The winding component 15 in the stirrer 5 is driven by the motor to reel in or release the pull rope 16 according to the prediction results of the neural network algorithm, pulling the blade 10 to adjust the angle of the stirring blade, and at the same time pushing the cylinder 21 to push the blade cover 11 to adjust the blade length, adapting to the precise control of the stirring requirements of the three fermentation stages of yeast activation, main fermentation period and late aging. The reinforcement learning algorithm uses the speed adjustment of the stirring motor 4 and the stirring time control as the action space, and the fermentation efficiency and flavor quality of bamboo wine as the reward function for experimentation and learning.

Images