Ganoderma lucidum spore oil exosome production and preparation control method and system

Abstract

The invention discloses a Ganoderma lucidum spore oil exosome production and preparation control method and system, and belongs to the technical field of biological medicines.The control method specifically comprises the following steps that I, multi-type variable data of Ganoderma lucidum spore oil exosomes are collected and preprocessed, features of the processed various types of variable data are extracted, and a multi-dimensional data set is formed; according to the method, the negative influence of single production link adjustment on the whole production can be avoided, the production fluctuation is effectively reduced, a sufficient time window is provided for timely intervention, and the quality stability and batch consistency of the ganoderma lucidum spore oil exosome product are guaranteed; meanwhile, production risks caused by unreasonable parameter adjustment can be effectively avoided, the feasibility and safety of the scheme are improved, it is guaranteed that the quality of the ganoderma lucidum spore oil exosome product is stable, and parameter adjustment requirements in different production scenes can be flexibly met.

Description

Technical Field

[0001] This invention relates to the field of biomedicine, and in particular to a method and system for controlling the production and preparation of Ganoderma lucidum spore oil exosomes. Background Technology

[0002] Ganoderma lucidum spore oil is the core active product of Ganoderma lucidum, rich in key active ingredients such as triterpenoids and polysaccharides. These ingredients are the core basis for its medicinal value and play an irreplaceable role in biomedical fields such as immune regulation, anti-tumor, and anti-inflammation. Exosomes, as natural active ingredient delivery carriers, can effectively overcome biological barriers due to their excellent biocompatibility, low immunogenicity, and targeted delivery capabilities, significantly improving the bioavailability and in vivo targeted enrichment efficiency of Ganoderma lucidum spore oil active ingredients, and greatly enhancing its pharmacological efficacy. Therefore, the efficient preparation technology of Ganoderma lucidum spore oil exosomes has become a key research hotspot in the biomedical field and has attracted much attention from the industry.

[0003] Existing methods and systems for controlling the production and preparation of Ganoderma lucidum spore oil exosomes often cause chain fluctuations in the overall production process when adjusting parameters at any stage of production, making it difficult to guarantee the stability and consistency of product quality across batches. Furthermore, existing methods and systems for controlling the production and preparation of Ganoderma lucidum spore oil exosomes lack flexibility in parameter adjustment when dealing with different production scenarios, often relying on manual experience, which carries production risks due to unreasonable adjustments. Therefore, we propose a method and system for controlling the production and preparation of Ganoderma lucidum spore oil exosomes. Summary of the Invention

[0004] The purpose of this invention is to address the deficiencies in the existing technology by proposing a method and system for controlling the production and preparation of Ganoderma lucidum spore oil exosomes.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: A method for controlling the production and preparation of Ganoderma lucidum spore oil exosomes, the specific steps of which are as follows: Ⅰ: Collect and preprocess multi-type variable data of Ganoderma lucidum spore oil exosomes, and extract the features of each type of variable data after processing to form a multi-dimensional dataset; II: Based on a multi-dimensional dataset, predict the current production trend of Ganoderma lucidum spore oil exosomes, and analyze the direction and degree of the corresponding production deviation; III: Based on the prediction results, coordinate the control directions and adjustment needs of different production stages to generate corresponding globally optimal control commands; IV: Collect the adjusted Ganoderma lucidum spore oil exosome production data, determine the matching degree between the current control parameters and the preset production target, and adjust the control parameters according to the judgment results; V: Collect all control data, forecast data, and feedback data during the production process of the corresponding batch to form a complete production data archive.

[0006] A control system for the production and preparation of Ganoderma lucidum spore oil exosomes includes a data acquisition and processing module, a variable prediction module, a collaborative control module, a control execution module, a real-time monitoring module, an adjustment and optimization module, a storage management module, and a standard configuration module. The data acquisition and processing module is used to collect various variable data during the production process of Ganoderma lucidum spore oil exosomes and to preprocess the collected variable data. The variable prediction module fits the current production data change trajectory based on the processed variable data and generates production trend prediction curves for each indicator. The collaborative control module is used to receive prediction results, extract deviation information and risk points of each production link, and generate corresponding global optimal control instructions. The control execution module is used to receive and parse the corresponding global optimal control instructions, and adjust the execution parameters of the corresponding production equipment according to the parsed information; The real-time monitoring module is used to collect the indicators after the adjustment of each control link in real time and record the corresponding dynamic change curve of the parameters. The adjustment and optimization module is used to analyze the matching degree between the current control parameters and the preset production target, and generate an optimized parameter scheme based on the analysis results; The storage management module is used to store the entire production process data of Ganoderma lucidum spore oil exosomes and to classify and archive the data according to production batch, data type and time node. The standard configuration module is used to store the standard numerical range of each indicator, the standard production trend curve, and the process control threshold.

[0007] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. This invention collects multi-type variable data on the production of Ganoderma lucidum spore oil exosomes. After removing outlier data and supplementing missing data, the data is uniformly mapped to the [0,1] interval. Features of each type of variable data are extracted and integrated to form a multi-dimensional dataset. Then, the raw material characteristics, process parameters, and historical production data of the current production batch are obtained, and similarity is calculated. Historical production data with high similarity exceeding the preset standard are screened out. Then, each type of variable data is matched with the screened historical production data according to time nodes and variable types to form a fused dataset. At the same time, a variable prediction model is constructed and trained. The fused dataset is input into the model, and the inherent mapping relationship between different variables is explored through multi-dimensional correlation analysis. The current production trajectory is fitted to generate prediction curves for each indicator, and the prediction curves are compared with the standard curve to predict subsequent changes, forming a production trend prediction report. Then, the numerical difference between all production trend prediction curves and the preset standard production trend curve at each time node is quantitatively calculated to determine the direction of deviation. At the same time, the predicted values ​​of each indicator and the corresponding standard values ​​are calculated. The absolute value of the difference and its percentage ratio to the standard value are used to determine the deviation level. Then, the impact of different deviation levels on product quality and production efficiency is analyzed, and a deviation assessment report is compiled. Subsequently, the distributed collaborative control center receives and parses the prediction results and distributes them to the corresponding sub-controllers. Each sub-controller provides feedback on the operating conditions and preliminary adjustment requirements, forming an "operating condition-prediction" comparison dataset. Simultaneously, based on the "operating condition-prediction" comparison dataset and the distributed consensus algorithm, the distributed collaborative center coordinates the global control direction and calls the global optimization algorithm to construct a multi-objective function. Iteratively solving the problem, it obtains the adjustment parameters of each production link and the execution tasks of each sub-controller, generates and issues the corresponding global optimal control instructions. After each sub-controller executes the parameter adjustment, it continuously collects and records the parameter change curves after the adjustment and then synchronously transmits them back to the collaborative control module. This can avoid the negative impact of a single production link adjustment on the overall production, effectively reduce production fluctuations, provide sufficient time window for timely intervention, and ensure the stability and batch consistency of Ganoderma lucidum spore oil exosome product quality.

[0008] 2. This invention receives real-time monitoring data collected by each sub-controller and decision information from the collaborative control module. It extracts actual control parameters for each production stage, real-time characteristic indicators of Ganoderma lucidum spore oil exosomes, and parameter change trend curves from the monitoring data. Then, it extracts the global production target, control thresholds for each stage, deviation levels, and baseline values ​​of control parameters from the decision information. Subsequently, it establishes a "control parameter-actual indicator-production target" correspondence table after classifying and integrating the data according to the production stage. Simultaneously, it combines the global production target with the standard parameters for Ganoderma lucidum spore oil exosome production to determine the standard target range and optimal target value for each control parameter. It also calculates the quantitative deviation index between the actual parameters and the optimal target value, analyzes the fit between the real-time characteristic indicators of Ganoderma lucidum spore oil exosomes and the target indicators, and examines the impact of the parameters on product quality. Finally, it classifies the control parameter matching level according to preset rules and locks the matching state below the preset level. Standard deviation control parameters are used to analyze the influencing factors of each deviation control parameter and their chain effects on other control parameters, production efficiency, and product quality. Then, based on the parameter type and optimization requirements, an adaptive learning algorithm is selected and constraints are set. The locked deviation control parameters, actual working condition data, production targets, and constraints are input into the algorithm. The algorithm iterates and optimizes with the goal of minimizing deviation and maximizing matching degree until a preliminary solution is output. Then, the feasibility of the solution is verified by production process constraints, equipment limit parameter verification, and variable prediction model simulation. If the target is not met, the iteration is returned and restarted. Finally, the verified optimization solution is integrated and transmitted to the collaborative control module. This can effectively avoid production risks caused by unreasonable parameter adjustments, improve the feasibility and safety of the solution, ensure the stable quality of Ganoderma lucidum spore oil exosomes, and flexibly respond to parameter adjustment needs under different production scenarios. Attached Figure Description

[0009] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used together with the embodiments of the invention to explain the invention and do not constitute a limitation thereof.

[0010] Figure 1 This is a flowchart illustrating a method for controlling the production and preparation of Ganoderma lucidum spore oil exosomes proposed in this invention. Figure 2 This is a system block diagram of a control system for the production and preparation of Ganoderma lucidum spore oil exosomes proposed in this invention. Detailed Implementation

[0011] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0012] Example 1, referring to Figure 1This embodiment discloses a method for controlling the production and preparation of Ganoderma lucidum spore oil exosomes. The specific steps of this control method are as follows: We collected and preprocessed multi-variable data of Ganoderma lucidum spore oil exosomes, and extracted features of each type of variable data after processing to form a multi-dimensional dataset.

[0013] Specifically, data on multiple types of variables were collected during the production process of Ganoderma lucidum spore oil exosomes. Data exceeding the preset range of values ​​in each type of variable data were removed. Then, interpolation was used to supplement the missing data in each type of variable data. Subsequently, each type of variable data was uniformly mapped to [0,1]. At the same time, the features of each type of variable data were extracted and integrated to form a multi-dimensional dataset.

[0014] It should be further noted that the data includes multiple types of variables such as raw material extraction purity, workshop temperature and pressure, stirring rate of the reactor, heating power, feed rate, and discharge pressure.

[0015] Based on a multi-dimensional dataset, we predict the current production trend of Ganoderma lucidum spore oil exosomes and analyze the direction and extent of the corresponding production deviations.

[0016] Specifically, historical production data of Ganoderma lucidum spore oil exosomes, raw material characteristics of the current production batch, and production process parameters are obtained. The similarity between the current production batch data and historical production data is calculated, and historical production data with similarity exceeding a preset standard is selected. Then, multi-type variable data is matched with the selected historical production data according to production time nodes and variable types to form a fused dataset. Simultaneously, a variable prediction model is constructed and trained. The fused dataset is input into the trained variable prediction model, which, based on preset algorithm logic, performs multi-dimensional correlation modeling and temporal feature extraction on various types of variable data in the fused dataset, while also mining different variables. The inherent mapping relationship between quantities, combined with the trend evolution pattern in historical production data, is used to reason and analyze the change trajectory of current production data within the prediction time window, outputting the predicted numerical results of each type of variable at multiple time nodes. Based on the predicted numerical results, time series reconstruction and trend fitting processing are performed on the corresponding indicators of each type of variable to generate production trend prediction curves corresponding to each type of variable data. The various production trend prediction curves are compared and analyzed with the preset standard production trend curves to determine the degree of fit between the current production trend and the standard trend, and to predict the change direction and rate of change of each type of variable in the subsequent production process, forming a production trend prediction report.

[0017] It should be further explained that the specific training steps for the variable prediction model are as follows: Historical production data of Ganoderma lucidum spore oil exosomes were collected and preprocessed to form a standardized feature dataset. This dataset was then divided into training, validation, and test sets according to a preset ratio to obtain the temporal and multivariate correlation characteristics of Ganoderma lucidum spore oil exosome production. Based on these characteristics, a variable prediction model architecture was built, and core parameters and hyperparameters for model training were set. A smoothed L1 loss function was selected, and the Adam optimizer and initial learning rate were configured. Data from the training set was input into the built variable prediction model in preset batches and forward propagation was performed. Each batch of data sequentially passed through the input layer, normalization layer, convolutional layer, recurrent layer, fully connected layer, and output layer for feature fusion and nonlinear transformation, outputting the prediction results for each indicator. The deviation between the predicted and true values ​​was calculated using the smoothed L1 loss function. Subsequently, the gradient of each layer's parameters was solved in a chain using the backpropagation algorithm, and the model weights and biases were updated using the Adam optimizer and the set learning rate. The model performance is evaluated using the validation set data of the current batch. The smoothed L1 loss function value and various error metrics are recorded. After each iteration, the smoothed L1 loss function value, validation set prediction error, and parameter status are recorded. The current performance metrics of the model are compared with the preset termination conditions. If the smoothed L1 loss function value does not converge, the validation set error is higher than the preset termination condition, or shows an upward trend, the reasons for not meeting the standards are analyzed. The model hyperparameters are adjusted using the Adam optimizer, and the iterative training process is repeated until the smoothed L1 loss function converges and the validation set prediction error converges to the preset range. The test set data is then input into the optimized variable prediction model, and the predicted values ​​of each metric at the corresponding prediction time step are output. The error metrics between the predicted values ​​and the actual values ​​of the test set are calculated, and each error metric is compared with the preset threshold. If all metrics are within the preset threshold, the model performance is considered to have met the standards; otherwise, the model architecture adjustment or hyperparameter optimization steps are returned, and training and validation are restarted.

[0018] Specifically, the numerical differences between all production trend prediction curves and preset standard production trend curves at each time point are calculated to identify corresponding production deviations. The standard numerical ranges, standard trend curves, and corresponding process control thresholds for various indicators in Ganoderma lucidum spore oil exosome production are obtained. The predicted values ​​of each indicator output by the variable prediction model are compared one by one with the corresponding standard values. If the predicted value is higher than the standard value, it is determined as a positive deviation; if the predicted value is lower than the standard value, it is determined as a negative deviation; if the predicted value is within the standard value range, it is determined as no deviation. The absolute value of the difference between the predicted value and the corresponding standard value of each indicator is calculated, and then the ratio of the absolute value of the deviation to the corresponding standard value is calculated and converted into a percentage. Subsequently, two quantitative parameters of each indicator are calculated and recorded to form a deviation quantitative parameter table. Simultaneously, based on the production process requirements of Ganoderma lucidum spore oil exosomes, deviation level classification rules are set. If the deviation is relatively... If the deviation ratio is between 0% and 5%, it is considered a slight deviation; if the relative deviation ratio is between 5% and 15%, it is considered a moderate deviation; if the relative deviation ratio exceeds 15%, it is considered a severe deviation. The deviation level is then linked to corresponding indicators and quantitative parameters. Based on the deviation level determination results, corresponding data of the same deviation level in the preset production database are obtained. The impact of different deviation levels on the final Ganoderma lucidum spore oil exosome product quality and production efficiency is analyzed. If it is a slight deviation, it has no significant impact on product quality and production efficiency, and no emergency intervention is required. If it is a moderate deviation, it will lead to fluctuations in product quality or a decrease in production efficiency, requiring timely adjustment of control parameters. If it is a severe deviation, it will cause substandard product quality or production process interruption, requiring immediate intervention measures. Finally, a deviation assessment report is compiled according to a standardized format, including the deviation direction, deviation quantitative parameter table, and the expected impact assessment data of the deviation level determination results.

[0019] It should be further explained that the specific calculation of the numerical differences at each time point is as follows: ; In the formula, Indicates any core indicator in The numerical difference at any given time; This represents discrete time points in the production process. Indicates any core indicator in Predicted values ​​for each moment Indicates any core indicator in Standard value for time; The specific formula for calculating the absolute value of the difference is as follows: ; In the formula, This represents the absolute value of the deviation from the core indicator; This indicates the predicted value of the core indicator; This represents the standard value of the core indicator; The specific formula for calculating the ratio is as follows: ; In the formula, This indicates the relative proportion of deviations in the core indicators; This indicates the predicted value of the core indicator; This indicates the standard value of the core indicator.

[0020] Based on the prediction results, the control directions and adjustment needs of different production stages are coordinated to generate corresponding globally optimal control commands.

[0021] Specifically, the system receives the prediction results from the variable prediction model in real time and performs hierarchical analysis of these results through the collaborative control module. Based on the prediction results, it comprehensively evaluates the prediction deviation magnitude, rate of change trend, and risk score of key indicators in each production link, identifies production links with deviation risks exceeding preset thresholds, and determines the scope of production links requiring adjustment. Then, according to the preset division of labor in the distributed control architecture, the analyzed prediction information is distributed to the corresponding responsible sub-controllers. Each sub-controller receives the distributed prediction information, obtains the indicator data for the current link, and compares and analyzes it with the corresponding indicators in the prediction results. It assesses the fit between the current operating condition and the predicted trend, determines the initial adjustment requirements for its own control link, and then synchronously feeds back the current operating condition and initial adjustment requirements to the distributed collaborative control center via the communication network, forming an "operating condition-prediction" comparison dataset. Based on this dataset and combined with a distributed consensus algorithm, each sub-controller exchanges data in real time via the communication network, reporting the current operating condition, initial adjustment requirements, and adjustments for its own control link. The potential impact of other processes is then considered. Subsequently, the distributed collaborative control center coordinates the adjustment needs of each sub-controller based on the global optimization goal. For production processes with deviations, a unified control direction is negotiated and determined. Based on the negotiated control direction, the global optimization algorithm is invoked to conduct collaborative calculations. With the goal of minimizing the impact of deviations, ensuring stable product quality, and improving production efficiency, a corresponding objective function is constructed. The current operating conditions fed back by each sub-controller, the deviation parameters in the prediction results, and the production process constraints are input. The objective function is then solved through iterative calculations to obtain the specific adjustment parameters for each production process. At the same time, specific execution tasks are assigned to each sub-controller. Then, based on the collaborative calculation results, globally optimal control instructions for each production process are generated. The globally optimal control instructions are divided according to the sub-controllers, and the encapsulated control instructions are classified and organized to form an instruction list for each sub-controller and issued. After receiving the corresponding globally optimal control instructions, each sub-controller adjusts the corresponding production equipment parameters. Afterward, it continuously collects the adjusted parameters of its own control process and records the dynamic change curve of the parameter adjustment, and then transmits it back to the collaborative control module.

[0022] It should be further explained that the specific formula for calculating the objective function is as follows: ; In the formula, This represents the number of production processes involved in collaborative optimization. A vector representing the control variables for each production stage; Representative production process In control parameters Actual operating indicators; Representative production process Standard target indicators; The production process is represented by the output of the variable prediction model. The prediction results; Representative production process The control cost function; Represents the weighting coefficient for quality stability; This represents the weighting coefficient for predicting trend following. The weighting coefficient represents production efficiency. The objective function constraints specifically include process parameter constraints, rate of change constraints, process coupling constraints, and product quality constraints, and their specific constraint forms are as follows: Process parameter constraints: ; Rate of change constraint: ; Process coupling constraints: ; Product quality constraints: ; Among the above constraints, Representing the One control variable; Representing the The minimum allowed value for each control parameter; Representing the The maximum allowed value for each control parameter; Representing the Each control parameter in the current control cycle or current time step The value to be taken below; Representing the The control parameter was in the previous control cycle or the previous time step. The value of ; Representing the The maximum allowable variation of a control parameter within an adjacent control cycle; Representing the A process coupling constraint function; This represents a vector of control variables composed of control parameters from each production stage. This represents the number of pre-defined process coupling constraints. Represents the vector of control variables Under the action, the first The actual operating values ​​of each production link or corresponding product quality indicator; Representative No. The lower limit threshold allowed for each product quality indicator; Representing the The upper limit threshold allowed for each product quality indicator.

[0023] Collect adjusted Ganoderma lucidum spore oil exosome production data, determine the matching degree between the current control parameters and the preset production target, and adjust the control parameters according to the determination results.

[0024] Collect all control data, forecast data, and feedback data during the production process of the corresponding batch to form a complete production data archive.

[0025] Example 2, refer to Figure 2 This embodiment discloses a control system for the production and preparation of Ganoderma lucidum spore oil exosomes, including a data acquisition and processing module, a variable prediction module, a collaborative control module, a control execution module, a real-time monitoring module, an adjustment and optimization module, a storage management module, and a standard configuration module; The data acquisition and processing module is used to collect various variable data during the production process of Ganoderma lucidum spore oil exosomes and to preprocess the collected variable data. The variable prediction module fits the current production data change trajectory based on the processed variable data and generates production trend prediction curves for each indicator.

[0026] The collaborative control module receives the prediction results, extracts deviation information and risk points from each production link, and generates corresponding global optimal control instructions.

[0027] The control execution module is used to receive and parse the corresponding global optimal control commands, and adjust the execution parameters of the corresponding production equipment according to the parsed information; the real-time monitoring module is used to collect the indicators after the adjustment of each control link in real time, and record the corresponding parameter dynamic change curves.

[0028] The adjustment and optimization module is used to analyze the matching degree between the current control parameters and the preset production target, and generate an optimized parameter scheme based on the analysis results.

[0029] Specifically, the system receives real-time monitoring data from each sub-controller and decision information from the collaborative control module. It extracts the current actual control parameters for each production stage, real-time characteristic indicators of Ganoderma lucidum spore oil exosomes, and parameter trend curves from the monitoring data. Then, it extracts the global production target, control thresholds for each stage, deviation levels, and baseline values ​​of the currently executed control parameters from the decision information. Finally, it categorizes and integrates the parsed information according to production stages, establishing a correspondence table of "control parameters - actual indicators - production targets" to obtain the corresponding standard parameters for Ganoderma lucidum spore oil exosome production. Finally, it determines the standard targets for each control parameter based on the global production target. The system first defines the target range and optimal target value, then calculates the quantitative index of the deviation between the current actual control parameters and the optimal target value. Simultaneously, it analyzes the fit between the real-time characteristic indicators of Ganoderma lucidum spore oil exosomes and the target indicators, assessing the impact of the current control parameters on product quality. Then, according to a preset matching degree classification rule, the matching state of each control parameter is divided into corresponding levels. Based on the matching degree analysis results, the deviation control parameters that are below the preset matching state are identified, and the influencing factors of each deviation control parameter are analyzed. Furthermore, the chain effect of each deviation control parameter on other control parameters, production efficiency, and product quality is outlined. Finally, based on the deviation... Based on the parameter type, coupling characteristics, and optimization requirements, a corresponding adaptive learning algorithm is selected, and appropriate constraints are set. The locked deviation control parameters, current actual operating data, production target parameters, and constraints are input into the configured adaptive learning algorithm. The algorithm uses minimizing the control parameter deviation and maximizing the production target matching degree as its objective function. It adjusts the deviation parameter values ​​through multiple iterations. After each iteration, the matching degree between the optimized parameters and the target value is calculated, and the deviation changes before and after the iteration are compared. If the matching degree does not meet the preset requirements, the algorithm parameters are adjusted, and iteration continues; otherwise, the iteration is stopped, and a preliminary result is output. The parameter optimization scheme is compared with the production process constraints and equipment operating limits. Parameter values ​​that exceed the normal constraint range are then eliminated. Simultaneously, the preliminary optimized parameter scheme is input into a variable prediction model for simulation calculation. This model predicts the parameter change trends in each production stage, the quality indicators of Ganoderma lucidum spore oil exosomes, and any new deviations that may arise after the scheme is implemented. The simulation results are then compared with the global production target. If the simulation results meet the requirements, the optimization scheme is deemed feasible; otherwise, iterative optimization is returned, the algorithm parameters are adjusted, and the calculation is repeated. Finally, the validated optimization schemes are integrated and synchronously transmitted to the collaborative control module.

[0030] It should be further explained that the specific formula for calculating the matching degree is as follows: ; In the formula, Indicates the first The production target matching degree evaluation value after round of iteration; This represents the total number of production target indicators; Indicates the first After the first iteration The achievement rate of each production target indicator; Indicates the first The weighting coefficients of each production target indicator.

[0031] The storage management module is used to store the entire production process data of Ganoderma lucidum spore oil exosomes and to classify and archive the data according to production batch, data type and time node; the standard configuration module is used to store the standard value range of each indicator, standard production trend curve and process control threshold.

Claims

1. A method for controlling the production and preparation of Ganoderma lucidum spore oil exosomes, characterized in that, The specific steps of this control method are as follows: Ⅰ: Collect and preprocess multi-type variable data of Ganoderma lucidum spore oil exosomes, and extract the features of each type of variable data after processing to form a multi-dimensional dataset; II: Based on a multi-dimensional dataset, predict the current production trend of Ganoderma lucidum spore oil exosomes, and analyze the direction and degree of the corresponding production deviation; III: Based on the prediction results, coordinate the control directions and adjustment needs of different production stages to generate corresponding globally optimal control commands; IV: Collect the adjusted Ganoderma lucidum spore oil exosome production data, determine the matching degree between the current control parameters and the preset production target, and adjust the control parameters according to the judgment results; V: Collect all control data, forecast data, and feedback data during the production process of the corresponding batch to form a complete production data archive.

2. The method for controlling the production and preparation of Ganoderma lucidum spore oil exosomes according to claim 1, characterized in that, The specific steps for forming the multidimensional dataset described in step I are as follows: S1.1: Collect multi-type variable data during the production process of Ganoderma lucidum spore oil exosomes, and remove data that exceed the preset value range in each type of variable data. Then, use interpolation to supplement the missing data in each type of variable data. Subsequently, map each type of variable data to [0,1] and extract and integrate the features of each type of variable data to form a multi-dimensional dataset. Among them, the multi-type variable data includes the purity of raw material extraction, workshop temperature and pressure, stirring rate of reaction vessel, heating power, feeding rate and discharge pressure.

3. The method for controlling the production and preparation of Ganoderma lucidum spore oil exosomes according to claim 1, characterized in that, The specific steps for predicting the current production trend of Ganoderma lucidum spore oil exosomes, as described in step II, are as follows: S2.1: Obtain historical production data of Ganoderma lucidum spore oil exosomes, raw material characteristics and production process parameters of the current production batch, calculate the similarity between the current production batch data and historical production data, then filter out historical production data with similarity exceeding the preset standard, and then match multi-type variable data with the filtered historical production data according to production time nodes and variable types to form a fusion dataset, while building and training a variable prediction model. S2.2: Input the fused dataset into the trained variable prediction model. Based on the preset algorithm logic, the variable prediction model performs multi-dimensional correlation analysis on the various types of variable data in the fused dataset, explores the inherent mapping relationship between different variables, and combines the trend patterns in historical production data to fit the change trajectory of the current production data, generating production trend prediction curves for the corresponding indicators in each type of variable data. S2.3: Compare and analyze various production trend forecast curves with preset standard production trend curves to determine the degree of fit between the current production trend and the standard trend, and predict the direction and rate of change of various types of variables in the subsequent production process to form a production trend forecast report.

4. The method for controlling the production and preparation of Ganoderma lucidum spore oil exosomes according to claim 3, characterized in that, The specific steps for analyzing the direction and extent of the corresponding production deviations described in Step II are as follows: S3.1: Calculate the numerical differences between all production trend prediction curves and the preset standard production trend curves at each time node, identify the corresponding production deviations, and obtain the standard numerical ranges, standard trend curves, and corresponding process control thresholds for various indicators of Ganoderma lucidum spore oil exosome production. Then, compare the predicted values ​​of each indicator output by the variable prediction model with the corresponding standard values ​​one by one. If the predicted value is higher than the standard value, it is determined to be a positive deviation; if the predicted value is lower than the standard value, it is determined to be a negative deviation; if the predicted value is within the standard value range, it is determined to be no deviation. S3.2: Calculate the absolute value of the difference between the predicted value and the corresponding standard value of each indicator, then calculate the ratio of the absolute value of the deviation to the corresponding standard value and convert it into a percentage. Subsequently, calculate and record the two quantitative parameters of each indicator to form a deviation quantitative parameter table. At the same time, in combination with the production process requirements of Ganoderma lucidum spore oil exosomes, set deviation level classification rules. If the relative deviation ratio is in the range of 0-5%, it is judged as a slight deviation; if the relative deviation ratio is in the range of 5-15%, it is judged as a moderate deviation; if the relative deviation ratio exceeds 15%, it is judged as a serious deviation. Then, the deviation level is associated and bound with the corresponding indicator and quantitative parameter. S3.3: Based on the deviation level determination results, obtain the corresponding data of the same deviation level in the preset production database, analyze the impact of different deviation levels on the final Ganoderma lucidum spore oil exosome product quality and production efficiency. If it is a slight deviation, it will not have a significant impact on product quality and production efficiency, and no emergency intervention is required. If it is a moderate deviation, it will lead to fluctuations in product quality or a decrease in production efficiency, and the control parameters need to be adjusted in time. If it is a serious deviation, it will cause the product quality to fail to meet the standards or the production process to be interrupted, and intervention measures need to be initiated immediately. Finally, the deviation direction, deviation quantification parameter table, and deviation level determination results are used to prepare a deviation assessment report in a standardized format.

5. The method for controlling the production and preparation of Ganoderma lucidum spore oil exosomes according to claim 3, characterized in that, The specific training steps for the variable prediction model described in S2.2 are as follows: P1.1: Collect historical production data of Ganoderma lucidum spore oil exosomes, preprocess the collected historical production data of each group to form a standardized feature dataset, and then divide it into training set, validation set and test set according to a preset ratio; P1.2: Obtain the time-series characteristics and multivariate association characteristics of Ganoderma lucidum spore oil exosome production, build a variable prediction model architecture based on the time-series characteristics and multivariate association characteristics, set the core parameters and hyperparameters for model training, select the smooth L1 loss function, and configure the Adam optimizer and initial learning rate. P1.3: Input the data in the training set into the completed variable prediction model according to the preset batches, and perform forward propagation. Each batch of data passes through the input layer, normalization layer, convolutional layer, recurrent layer, fully connected layer and output layer in sequence for feature fusion and nonlinear transformation, and outputs the prediction results of each index. Then, the deviation between the predicted value and the true value is calculated by the smoothed L1 loss function. Subsequently, the gradient of the parameters of each layer is solved in a chain by the backpropagation algorithm, and the model weights and biases are updated by combining the Adam optimizer and the set learning rate. At the same time, the model performance is evaluated by using the validation set data of the current batch, and the smoothed L1 loss function value and various error indicators are recorded. P1.4: After each iteration, record the model's smoothed L1 loss function value, validation set prediction error, and parameter status. Compare the model's current performance metrics with the preset termination conditions. If the smoothed L1 loss function value does not converge, the validation set error is higher than the preset termination conditions, or shows an upward trend, analyze the reasons for not meeting the criteria, adjust the model's hyperparameters using the Adam optimizer, and repeat the iterative training process until the smoothed L1 loss function converges and the validation set prediction error converges to the preset range. P1.5: Input the test set data into the optimized variable prediction model, calculate the error index between the model's predicted value and the actual value of the test set, and then compare each error index with the preset threshold. If all indexes are within the preset threshold, the model performance is deemed satisfactory; otherwise, return to the model architecture adjustment or hyperparameter optimization step and retrain and validate.

6. The method for controlling the production and preparation of Ganoderma lucidum spore oil exosomes according to claim 3, characterized in that, The specific steps for generating the corresponding globally optimal control command in step III are as follows: S4.1: Receive the prediction results output by the variable prediction model in real time, and perform hierarchical analysis of the prediction results through the built-in distributed collaborative control center. Extract the indicator deviation information, trend change patterns and potential deviation risks corresponding to each production link, determine the scope of production links that need to be adjusted, and then distribute the analyzed prediction information to the corresponding responsible sub-controllers according to the preset division of labor of the distributed control architecture. S4.2: After receiving the distributed prediction information, each sub-controller obtains the data of each indicator in the current stage and compares and analyzes them with the corresponding indicators in the prediction results. It then analyzes the fit between the current operating condition and the prediction trend, determines the preliminary needs for adjustment of its own control stage, and then each sub-controller synchronously feeds back the current operating condition and preliminary adjustment needs to the distributed collaborative control center through the communication network to form an "operating condition-prediction" comparison dataset. S4.3: Based on the "condition-prediction" comparison dataset and combined with the distributed consensus algorithm, each sub-controller exchanges data in real time through the communication network, reports the current condition of its own control link, preliminary adjustment needs, and potential impact on other links. Then, the distributed collaborative control center coordinates the adjustment needs of each sub-controller according to the global optimization goal, and negotiates and determines a unified control direction for production links with deviations. S4.4: Based on the control direction determined through negotiation, the global optimization algorithm is invoked to carry out collaborative calculations. With the goals of minimizing the impact of deviations, ensuring stable product quality, and improving production efficiency, a corresponding objective function is constructed. The current operating conditions fed back by each sub-controller, the deviation parameters in the prediction results, and the production process constraints are input. The objective function is then solved through iterative calculations to obtain the specific adjustment parameters for each production link. At the same time, the specific execution tasks of each sub-controller are assigned. Then, based on the results of the collaborative calculations, the globally optimal control instructions for each production link are generated. S4.5: All global optimal control instructions are assigned to sub-controllers according to their functions. The encapsulated control instructions are classified and organized to form an instruction list for each sub-controller and then issued. After receiving the corresponding global optimal control instructions, each sub-controller adjusts the corresponding production equipment parameters. Then, it continuously collects the adjusted parameters of its own control loop and records the dynamic change curve of the parameter adjustment, and then transmits it back to the collaborative control module.

7. A control system for the production and preparation of Ganoderma lucidum spore oil exosomes, used to implement the production and preparation control method for Ganoderma lucidum spore oil exosomes according to any one of claims 1-6, characterized in that, It includes a data acquisition and processing module, a variable prediction module, a collaborative control module, a control execution module, a real-time monitoring module, an adjustment and optimization module, a storage management module, and a standard configuration module; The data acquisition and processing module is used to collect various variable data during the production process of Ganoderma lucidum spore oil exosomes and to preprocess the collected variable data. The variable prediction module fits the current production data change trajectory based on the processed variable data and generates production trend prediction curves for each indicator. The collaborative control module is used to receive prediction results, extract deviation information and risk points of each production link, and generate corresponding global optimal control instructions. The control execution module is used to receive and parse the corresponding global optimal control instructions, and adjust the execution parameters of the corresponding production equipment according to the parsed information; The real-time monitoring module is used to collect the indicators after the adjustment of each control link in real time and record the corresponding dynamic change curve of the parameters. The adjustment and optimization module is used to analyze the matching degree between the current control parameters and the preset production target, and generate an optimized parameter scheme based on the analysis results; The storage management module is used to store the entire production process data of Ganoderma lucidum spore oil exosomes and to classify and archive the data according to production batch, data type and time node. The standard configuration module is used to store the standard numerical range of each indicator, the standard production trend curve, and the process control threshold.

8. The control system for the production and preparation of Ganoderma lucidum spore oil exosomes according to claim 7, characterized in that, The specific steps by which the adjustment and optimization module generates the optimized parameter scheme based on the analysis results are as follows: S5.1: Receives real-time monitoring data collected by each sub-controller and decision information from the collaborative control module, extracts the current actual control parameters of each production link, real-time characteristic indicators of Ganoderma lucidum spore oil exosomes and parameter change trend curves from the monitoring data, and then extracts the global production target, control thresholds of each link, deviation level and current control parameter benchmark value from the decision information. After that, the parsed information is classified and integrated according to the production link to establish a correspondence table of "control parameters-actual indicators-production target". S5.2: Obtain the relevant standard parameters for the production of Ganoderma lucidum spore oil exosomes, and in conjunction with the overall production target, determine the standard target range and optimal target value of each control parameter. Then, calculate the quantitative index of the deviation between the current actual control parameter and the optimal target value. At the same time, analyze the fit between the real-time characteristic index of Ganoderma lucidum spore oil exosomes and the target index, evaluate the impact of the current control parameters on product quality, and then classify the matching status of each control parameter into the corresponding level according to the preset matching degree classification rules. S5.3: Based on the matching degree analysis results, identify the deviation control parameters that are in a matching state lower than the preset state, analyze the influencing factors of each deviation control parameter, and sort out the chain effect of each deviation control parameter on other control parameters, production efficiency and product quality. Then, according to the type, coupling characteristics and optimization requirements of the deviation parameters, select the corresponding adaptive learning algorithm and set the corresponding constraints. S5.4: Input the locked deviation control parameters, current actual operating data, production target parameters, and constraints into the configured adaptive learning algorithm. The algorithm uses minimizing the control parameter deviation and maximizing the production target matching degree as its objective function. It adjusts the deviation parameter values ​​through multiple iterations. After each iteration, it calculates the matching degree between the optimized parameters and the target value, compares the deviation changes before and after the iteration, and if the matching degree does not meet the preset requirements, it adjusts the algorithm parameters and continues iterating; otherwise, it stops the iteration calculation and outputs the preliminary optimized parameter scheme. The specific formula for calculating the matching degree is as follows: ； In the formula, Indicates the first The production target matching degree evaluation value after round of iteration; This represents the total number of production target indicators; Indicates the first After the first iteration The achievement rate of each production target indicator; Indicates the first Weighting coefficients for each production target indicator; S5.5: Compare the preliminary optimized parameter scheme with the production process constraints and equipment operating limit parameters, and then eliminate parameter values ​​that exceed the normal constraint range. At the same time, input the preliminary optimized parameter scheme into the variable prediction model for simulation calculation, predict the parameter change trend of each production link, the quality index of Ganoderma lucidum spore oil exosomes, and the new deviations generated after the scheme is implemented. Then, compare the simulation results with the global production target. If the simulation results meet the requirements, the optimization scheme is determined to be feasible; otherwise, return to iterative optimization, adjust the algorithm parameters and recalculate. Then, integrate the verified optimization scheme and transmit it synchronously to the collaborative control module.

Images