Production control method and system for aloe beverage

Abstract

The invention relates to the technical field of food production control, and discloses a production control method and system for an aloe beverage. The method comprises the following steps: acquiring key data of production equipment, and sending the key data and a unique identifier of the production equipment to a first server; the first server analyzes the production state based on the key data and sends an analysis result to a second server, the second server judges whether the production equipment is abnormal or not based on the analysis result, and if yes, abnormal information is sent to the user terminal; if no abnormity occurs, the second server sends an analysis result to a third server, and the third server performs abnormity prediction on the production equipment based on the target model; an abnormal mode of each target model is established, deep analysis is performed on the production equipment based on the abnormal modes, the types and reasons of abnormity are obtained and sent to a user terminal, and the user terminal optimizes and adjusts the production process based on an analysis result; the production efficiency of the aloe beverage is improved.

Description

Technical Field

[0001] This application relates to the technical field of food production control, and particularly to a production control method and system for aloe vera beverages. Background Art

[0002] As a functional beverage, aloe vera beverage is favored by more and more consumers due to its health care effects and unique flavor. However, the production process of aloe vera beverage is relatively complex and involves multiple key links. In the above production process, the judgment of abnormal situations and the parameter control of each link are crucial for the quality of the final product. Traditional production control methods for aloe vera beverages mainly rely on manual experience and preset parameter setting values.

[0003] To solve the above problems, the food production industry has begun to explore intelligent production control methods. For example, the Chinese patent application with the publication number CN118466426A discloses an optimization control method for the production process of fermented rice buns based on artificial intelligence, which specifically relates to the technical field of artificial intelligence. The specific steps include installing sensors during the production process of fermented rice buns to collect production characteristics and environmental optimization information, and transmitting them to the database for storage and management. Training a machine learning model using historical data, analyzing the influence of data on the fermentation effect, establishing a mathematical model, predicting the fermentation effect, and prompting production personnel. According to the monitoring data and prediction results, dynamically adjusting the raw material ratio, introducing a closed-loop control system, and continuously optimizing the production state. Deploying a computer vision system to detect the production effect in real time and feedback the results to the production system, and further optimizing production control in combination with model analysis, which can effectively avoid waste of raw materials caused by abnormal production failures, reduce production costs, and effectively evaluate production efficiency in the case of gas transfer or leakage in the fermentation container. However, the above-mentioned prior art mainly focuses on the optimization of the production process and does not clearly explain how to handle abnormal situations.

[0004] Another example is the Chinese patent application with the publication number CN110398907A, which discloses an automatic control method applicable to the production of meatballs. The method includes obtaining parameters associated with meatballs in real time and sending the collected meatball images to a computer through an industrial camera component; receiving the meatball parameter comparison results completed by the computer through a preset algorithm; when the meatball parameter comparison results meet the preset meatball state parameter values, completing the automatic control of meatball production through the meatball production numerical control system acting on shaping structures, steaming structures, etc. This method controls the automation of meatball production through the monitoring of the water environment state of meatballs by multiple sensors and the calculation of the industrial camera component and the computer through a preset algorithm. However, although the above-mentioned prior art uses an industrial camera for real-time detection, it does not elaborate on how to quickly respond to abnormal situations.

[0005] Therefore, a production control method and system for aloe vera beverages are needed to quickly respond to abnormal situations and optimize according to the causes of the abnormalities. Summary of the Invention

[0006] This application provides a production control method and system for aloe vera beverages to improve the production efficiency of aloe vera beverages.

[0007] In a first aspect, this application provides a production control method for aloe vera beverages, and the method includes: Step S1: The sensor module acquires key data of the production equipment and sends the key data and the unique identifier of the production equipment itself to the first server; Step S2: The first server analyzes the production status of the corresponding production equipment based on the unique identifier to obtain a first analysis result, and sends the first analysis result to the second server. The second server determines whether the production equipment has an abnormality based on the first analysis result. If an abnormality occurs, it directly sends the first analysis result to the user terminal and executes step S5; Step S3: If no abnormality occurs, the second server sends the key data, the unique identifier, and the first analysis result to the third server. The third server includes multiple target models, and analyzes the key data and the first analysis result based on the target models to obtain a second analysis result; Step S4: Establish an abnormal mode for each target model. The third server deeply analyzes the production equipment based on the abnormal mode and the second analysis result to obtain a third analysis result, and sends the third analysis result to the user terminal; Step S5: The user terminal optimizes and adjusts the production process based on the obtained first analysis result or third analysis result.

[0008] In combination with the first aspect, in the first implementation manner of the first aspect of this application, the first server analyzes the production status of the corresponding production equipment based on the unique identifier, including: Dividing the key data into multiple data categories based on different data performances, including alarm data, parameter data, general data, and performance index data. Based on the data categories, the analysis of the production status is divided into two implementation manners. For the key data of the same category, the first server compares each data item in the key data with a preset threshold to determine whether it meets the normal range of the corresponding data category. If it does not meet, it determines that the status of the key data is an abnormal status; For the key data of different categories, the first server combines the key data of different data categories based on the detection target to obtain multiple sample data, and performs multi-dimensional analysis on the sample data to obtain the analysis result of whether the production equipment is in an abnormal state.

[0009] Combined with the first aspect, in the second implementation manner of the first aspect of the present application, directly sending the first analysis result to the user terminal further includes: Dividing the abnormal state from simple to complex into multiple abnormal types, including basic type, associated type, type to be predicted, and type to be optimized, setting an urgency level for each abnormal type, where the urgency level decreases successively based on the simplicity to complexity of the abnormal type, and selecting different notification methods based on the urgency level to send the first analysis result to the user terminal.

[0010] Combined with the first aspect, in the third implementation manner of the first aspect of the present application, analyzing the key data and the first analysis result based on the target model includes: Obtain the key data during the normal production process of the aloe vera beverage, defined as historical normal data, extract multiple variables from the historical normal data, and define explanatory variables and target variables therefrom. Based on the explanatory variables, establish a prediction model for the target variables, where the prediction model is a mathematical model or a machine learning model. Each production stage includes one or more target models. The target model outputs the predicted value of the target variable. Based on the unique identifier, identify the production stage where each key data is currently located, select the corresponding target model based on the production stage, and input the key data into the corresponding target model to obtain the actual value of each target variable for each production stage, and define the actual value as the second analysis result.

[0011] Combined with the first aspect, in the fourth implementation manner of the first aspect of the present application, establishing the abnormal mode of each target model includes: Determine the perturbation method based on each explanatory variable in the historical normal data. The perturbation method is to add abnormal values to each explanatory variable within a preset range while keeping other explanatory variables unchanged. The changed explanatory variables and other explanatory variables are combined to generate abnormal data. Input each abnormal data into the corresponding target model. The target model outputs the predicted value of the target variable. Compare the predicted value of the target variable with the predicted value of the historical normal data to obtain the deviation value for each target variable. Combine all the deviation values to obtain the abnormal mode of each target model. Each target model has one or more abnormal modes.

[0012] In combination with the first aspect, in the fifth implementation manner of the first aspect of the present application, the third server performs in-depth analysis on the production equipment based on the abnormal mode and the second analysis result, including: Calculate the deviation value between the actual value of each target variable in the second analysis result and the predicted value obtained from the historical normal data, generate a deviation mode of the current production equipment based on the deviation value combination, calculate the similarity between the deviation mode and each abnormal mode, select the abnormal mode with the similarity greater than the first threshold as the current abnormal mode, establish a fault prediction model based on a deep learning model, obtain a variety of fault data related to the abnormal modes, the fault data including the fault type, the fault occurrence time, and the fault cause, input the abnormal mode and the corresponding fault data into the fault prediction model for training, and input the current abnormal mode into the trained fault prediction model to obtain the possible fault type and the cause of the fault in the current abnormal mode, and define it as the third analysis result.

[0013] In combination with the first aspect, in the sixth implementation manner of the first aspect of the present application, optimize and adjust the production process based on the obtained first analysis result or third analysis result, including: Record the optimization operations performed on the production process based on the first analysis result and the third analysis result, obtain the production status of the corresponding production stage after the optimization operation, and label the optimization operation. Establish an optimization model based on reinforcement learning, input the optimization operation and the production status after the optimization operation into the optimization model, use the optimization operation as the action and the production status as the reward, set a reward function based on the production target, evaluate each action based on the reward function to obtain the optimal action, and perform the optimization operation based on the optimal action.

[0014] In the second aspect, the present application provides a production control system for aloe vera beverages, and the system includes: A collection module, the sensor module obtains the key data of the production equipment and sends the key data and the unique identifier of the production equipment itself to the first server; An abnormality judgment module, the first server analyzes the production status of the corresponding production equipment based on the unique identifier to obtain a first analysis result, and sends the first analysis result to the second server. The second server judges whether the production equipment has an abnormality based on the first analysis result. If an abnormality occurs, directly send the first analysis result to the user terminal and execute step S5; Anomaly prediction module. If no anomaly occurs, the second server sends the critical data, the unique identifier, and the first analysis result to the third server. The third server includes multiple target models, and analyzes the critical data and the first analysis result based on the target models to obtain a second analysis result; Anomaly analysis module. An anomaly pattern for each target model is established. The third server deeply analyzes the production equipment based on the anomaly pattern and the second analysis result to obtain a third analysis result, and sends the third analysis result to the user terminal; Optimization module. The user terminal optimizes and adjusts the production process based on the obtained first analysis result or third analysis result.

[0015] Compared with the prior art, the beneficial effects of the present invention are at least as follows: By assigning a unique identifier to each device, the present invention avoids data confusion, ensures traceability of data sources. The sensor module collects critical data in real time and sends it to the first server. The first server quickly completes basic status analysis. If an anomaly is detected, it directly notifies the user to make adjustments to reduce losses; when no anomaly is detected, the critical data is sent to the second server. The second server defines dedicated target models for each production stage, and automatically matches the corresponding target model for each production stage based on the unique identifier. The target model can predict changes in target variables according to the critical data in the stage; an anomaly pattern library corresponding to the target model is established, and in combination with the current data and the pattern library, the probability and cause of a fault occurring in the current device are generated, potential faults are pre-warned, and planned maintenance is guided; the user terminal automatically obtains optimization actions using a reinforcement learning model according to the analysis results, adapts to changes and uncertainties in the production process, reduces manual operations, and improves production efficiency. Description of the Drawings

[0016] To more clearly illustrate the technical solutions of the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the following drawings are some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0017] Figure 1 It is a schematic diagram of an embodiment of the production control method for aloe vera beverage in the embodiments of the present application; Figure 2 It is a flowchart of production status analysis in the embodiments of the present application; Figure 3 It is a schematic diagram of the types of abnormal states and corresponding emergency levels in the embodiments of the present application; Figure 4This is a schematic diagram of an embodiment of the production control system for aloe vera beverages in the embodiments of the present application. Detailed implementation manners

[0018] The embodiments of the present application provide a production control method and system for aloe vera beverages. Terms such as "first", "second", "third", "fourth", etc. (if any) in the specification, claims, and above-mentioned drawings of the present application are used to distinguish similar objects and do not necessarily describe a specific order or sequence. It should be understood that such data can be interchanged under appropriate circumstances so that the embodiments described here can be implemented in an order different from that illustrated or described here. In addition, the term "including" or "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or units does not necessarily limit to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products, or devices.

[0019] For ease of understanding, the specific process of the embodiments of the present application is described below. Please refer to Figure 1 , an embodiment of a production control method for aloe vera beverages in the embodiments of the present application includes: Step S1: The sensor module acquires the key data of the production equipment and sends the key data and the unique identifier of the production equipment itself to the first server.

[0020] Specifically, according to the production process of aloe vera beverages, the production equipment is divided into extraction equipment, mixing equipment, sterilization equipment, canning equipment, and other auxiliary equipment. A unique identifier is assigned to each production equipment, and this identifier is associated with the relevant information of the equipment, such as equipment name, model, location, etc. The key data on the production equipment is collected in real time through sensors, such as temperature data, pressure data, etc. during the extraction stage. The collected data is also preliminarily processed, and the key data and the unique identifier are packaged into a data packet and sent to the first server through a selected communication method (such as Wi-Fi, Bluetooth, LPWAN, etc.). The first server is an edge server, which can process the key data collected by the sensor module in real time and perform a preliminary analysis locally, enabling the control system to quickly respond to abnormal situations during the production process.

[0021] Step S2: The first server analyzes the production status of the corresponding production equipment based on the unique identifier, obtains a first analysis result, and sends the first analysis result to the second server. The second server determines whether the production equipment has an abnormality based on the first analysis result. If an abnormality occurs, the first analysis result is directly sent to the user terminal, and step S5 is executed.

[0022] Specifically, asFigure 2 As shown, it is a flowchart for production status analysis. The first server receives data packets from the sensor module, identifies the corresponding production equipment according to the unique identifier in the data packet, conducts a preliminary analysis of the production status of the production equipment, generates a first analysis result. The first analysis result is, for example, the operating efficiency of the equipment, whether the operating parameters of the equipment are within the normal range. If it exceeds the normal range, the production equipment is determined to be in an abnormal state. The second server receives the first analysis result from the first server. The second server is a cloud server, mainly responsible for comprehensive anomaly judgment and user notification.

[0023] Based on the first analysis result, the cloud server determines whether the production equipment has an anomaly. If there is an anomaly, it directly sends the first analysis result to the user terminal, and the specific information of the anomaly, such as the equipment name and the parameter information of the anomaly, is displayed on the user terminal, and step S5 is executed, which will be further described later.

[0024] Step S3: If there is no anomaly, the second server sends the key data, unique identifier, and the first analysis result to the third server. The third server includes multiple target models, and analyzes the key data and the first analysis result based on the target models to obtain a second analysis result.

[0025] Specifically, the third server is an analysis server, responsible for further data analysis and optimization. Multiple target models are pre-configured in the third server. Each production stage corresponds to a different target model. According to the unique identifier, the production equipment and its production stage are identified, and the corresponding target model is selected. For example, if the unique identifier corresponds to an extraction device, the target model of the extraction stage is selected, such as an extraction rate prediction model. Assume the key data: extraction temperature = 85°C, extraction pressure = 5.5 bar, extraction time = 2.5 hours, etc., and the first analysis result: "The extraction equipment is operating normally". The key data and the first analysis result are input into each target model, and each target model conducts in-depth analysis on the input data to predict the extraction rate of aloe in the raw materials in the extraction stage. The predicted values of other target variables such as the extraction rate are used as the second analysis result.

[0026] Step S4: Establish an abnormal mode for each target model. The third server conducts in-depth analysis on the production equipment based on the abnormal mode and the second analysis result to obtain a third analysis result, and sends the third analysis result to the user terminal.

[0027] Specifically, the third server identifies possible anomalies in the current production status based on the second analysis result, matches the possible anomaly patterns of the current device from the anomaly patterns, predicts the likelihood and time of occurrence of a failure in the current production device based on the anomaly patterns, and analyzes the cause of the failure to obtain a third analysis result, such as "the predicted probability of a failure in the extraction device is 25%, the expected failure time is within 24 hours, and the cause of the failure may be a sensor failure or equipment wear".

[0028] Step S5: The user terminal optimizes and adjusts the production process based on the obtained first analysis result or third analysis result.

[0029] Specifically, based on the received analysis result, which may be the first analysis result (information indicating device anomalies) or the third analysis result (predicting which device may have a failure and the cause of the failure), the user terminal adjusts the production equipment. During the adjustment process, a reinforcement learning model is set up. Since the device status and environmental conditions in the production process may change over time, the reinforcement learning model can dynamically adjust the parameters of the production equipment according to real-time data, thereby ensuring that the production process is always in an optimal state. By evaluating the effect of the adjustment actions in real time, the reinforcement learning model can quickly find the optimal adjustment strategy and reduce waste and downtime in the production process.

[0030] In the present invention, by assigning a unique identifier to each device, data confusion is avoided, and the data source is ensured to be traceable. The sensor module collects key data in real time and sends it to the first server. The first server quickly completes the basic status analysis. If an anomaly is detected, the user is directly notified to make adjustments to reduce losses; when no anomaly is detected, the key data is sent to the second server. The second server defines a dedicated target model for each production stage, automatically matches the corresponding target model for each production stage based on the unique identifier. The target model can predict the change of the target variable according to the key data in the stage; an anomaly pattern library corresponding to the target model is established, and by combining the current data with the pattern library, the probability and cause of a failure in the current device are generated, potential failures are warned in advance, and planned maintenance is guided; the user terminal automatically obtains optimization actions according to the analysis result, adapts to the changes and uncertainties in the production process, reduces manual operations, and improves production efficiency.

[0031] In a specific embodiment, the steps for the first server to analyze the production status of the corresponding production device based on the unique identifier are as follows: Critical data is divided into multiple data categories based on different data performances, including alarm data, parameter data, general data, and performance indicator data. The analysis of the production status is divided into two implementation methods based on the data categories. For critical data of the same category, the first server compares each data item in the critical data with a preset threshold to determine whether it meets the normal range of the corresponding data category. If it does not meet the requirement, the status of the critical data is determined as an abnormal status; For critical data of different categories, the first server combines the critical data of different data categories based on the detection target to obtain multiple sample data, and conducts multi-dimensional analysis on the sample data to obtain the analysis result of whether the production equipment is in an abnormal state.

[0032] Specifically, alarm data represents data on equipment failures or abnormalities, such as excessively high temperature sensor data; parameter data represents real-time data of key parameters in the production process, such as extraction temperature, sterilization temperature, extraction pressure, extraction liquid flow rate, stirring speed, etc.; general data represents data related to production plans, product quality, equipment status, etc., such as planned output, planned running time, etc.; performance indicator data is data used to evaluate production efficiency and equipment performance, such as hourly output, cumulative output, etc.

[0033] The received critical data is divided into different data categories. For data of the same type, assuming the data category is parameter data and the data item includes extraction temperature, then the actually obtained extraction temperature is compared with the preset normal range (for example, 40°C to 50°C) to determine whether it is within the normal range. If the extraction temperature is 60°C and does not meet the normal range, the first analysis result including the abnormal temperature in the extraction stage is directly sent to the user terminal, and the user terminal can quickly respond to the abnormality in a short time.

[0034] However, a single data category may not comprehensively reflect the complex situations in the production process. The present invention also sets a second method to judge the abnormal state of the production process. Key data of different categories are combined according to the detection target. Suppose the detection target is whether the production equipment frequently stops briefly. General data 1: the running time of production equipment A in the past 10 minutes, performance index data 1: the number of brief stops of production equipment A, such as 2 times. The combined sample data 1: the number of brief stops of production equipment A within 10 minutes is 2. Compare this number with the number in the past 10 minutes. If the number of brief stops increases by more than 50%, it is determined that the production equipment is abnormal. Another example: the target is to analyze whether the number of unqualified products increases abnormally. General data 2: the number of unqualified products is 5, performance index data 1: the inspection time (for example, within the past 1 hour). The combined sample data 2: within the past 1 hour, the number of unqualified products is 5. Compare sample data 2 with the number at the previous inspection. If the data of unqualified products increases by more than the preset threshold, it is determined to be in an abnormal state.

[0035] There may be an association relationship between different categories of data. For example, the equipment running time (general data) may affect the production efficiency (performance index data), and the changes in temperature and pressure (parameter data) may affect the product quality. By comprehensively analyzing different categories of data, the present invention can more comprehensively evaluate the running state of the equipment.

[0036] In a specific embodiment, directly sending the first analysis result to the user terminal specifically further includes the following steps: The abnormal states are divided into multiple abnormal types from simple to complex, including basic type, associated type, to-be-predicted type, and to-be-optimized type. An urgency level is set for each abnormal type, and the urgency level decreases successively based on the simplicity to complexity of the abnormal type. Different notification methods are selected based on the urgency level to send the first analysis result to the user terminal.

[0037] Specifically, as Figure 3As shown in the figure, it is a schematic diagram of the types of abnormal states and the corresponding emergency levels. The abnormal states include the following four types: Basic type: related to the basic operation status of equipment, such as equipment stop time, simple alarms, etc.; Association type: related to the association analysis of multiple data categories, such as the relationship between temperature and production efficiency; To-be-predicted type: related to algorithm-predicted abnormalities, such as the prediction of production end time, which will be elaborated later; To-be-optimized type: related to the causes of abnormalities and optimization suggestions, such as production plan optimization. The emergency levels of the above four types decrease in order from high to low. Basic type: The highest emergency level (e.g., equipment stop, immediate handling required); Association type: Medium emergency level (e.g., production efficiency decline, timely adjustment required); To-be-predicted type: Lower emergency level (e.g., production end time prediction, advance preparation required); To-be-optimized type: The lowest emergency level (e.g., production plan optimization, can be arranged for subsequent processing).

[0038] According to the emergency levels of the abnormal types, different notification methods are selected to send the first analysis result to the user terminal. For example, the abnormal state of the basic type is notified through the monitor at the production site to prompt the operator to immediately check the equipment; the abnormal state of the association type is sent to the user terminal to prompt the operator to adjust the equipment parameters; the abnormal state of the to-be-predicted type is sent to the user terminal and relevant managers are notified by email; the notification information of the to-be-optimized type is sent to the production planner by email for optimization in subsequent work.

[0039] In a specific embodiment, the key data and the first analysis result are analyzed based on the target model: Obtain the key data during the normal production process of aloe vera beverage, defined as historical normal data. Extract multiple variables from the historical normal data, and define the explanatory variable and the target variable therefrom. Based on the explanatory variable, establish a prediction model for the target variable. The prediction model is a mathematical model or a machine learning model. Each production stage includes one or more target models. The target model outputs the predicted value of the target variable. Identify the production stage where each key data is currently located based on the unique identifier, select the corresponding target model based on the production stage, and input the key data into the corresponding target model to obtain the actual value of each target variable for each production stage. Define the actual value as the second analysis result.

[0040] Specifically, historical normal data of aloe vera beverages during normal production processes are obtained from the historical database. The data content includes key data for each production stage. Multiple variables are extracted from the historical normal data. For example, in the extraction stage: extraction temperature (x1), extraction pressure (x2), extraction time (x3), extraction rate (y1); in the mixing stage: stirring speed (x4), stirring time (x5), raw material ratio (x6), mixing uniformity (y2); in the sterilization stage: sterilization temperature (x7), sterilization time (x8), product pH value (x9), microbial inactivation rate (y3). Among them, the variables to be predicted are defined as target variables, such as extraction rate, mixing uniformity, and microbial inactivation rate. The variables that affect the target variables are used as explanatory variables, such as extraction temperature, extraction pressure, etc. Based on the explanatory variables, a prediction model for the target variables is constructed. The prediction model can be a mathematical model (such as multiple linear regression) or a machine learning model (such as neural network, support vector machine, etc.). Each production stage can include one or more target models. For example, in the extraction stage: one target model is used to predict the extraction rate; in the mixing stage: two target models are respectively used to predict the mixing uniformity and stirring power; in the sterilization stage: one target model is used to predict the microbial inactivation rate.

[0041] Assume that the target model is a mathematical model. Then, the extraction rate prediction model: extraction rate = f(extraction temperature, extraction pressure, extraction time). The second server identifies the production stage according to the unique identifier, selects the corresponding target model, and inputs the current key data of the production equipment into the target model for predictive analysis, and outputs the predicted value of the target variable, which is defined as the second analysis result. For example, the predicted extraction rate is 95%.

[0042] In a specific embodiment, the steps for establishing the abnormal pattern of each target model are specifically as follows: Based on each explanatory variable in the historical normal data, the perturbation method is determined. The perturbation method is to add abnormal values to each explanatory variable within a preset range while keeping other explanatory variables unchanged. The changed explanatory variables and other explanatory variables are combined to generate abnormal data. Each abnormal data is input into the corresponding target model. The target model outputs the predicted value of the target variable. The predicted value of the target variable is compared with the predicted value of the historical normal data to obtain the deviation value for each target variable. All the deviation values are combined to obtain the abnormal pattern of each target model. Each target model has one or more abnormal patterns.

[0043] Specifically, for each explanatory variable, determine the disturbance method. Assuming that the explanatory variable is the extraction temperature, the corresponding disturbance method is to increase the extraction temperature by 5°C and keep other explanatory variables unchanged. The disturbance range can be determined based on the actual situation. If the explanatory variable has a small impact on the target variable, a larger disturbance range can be set to see the abnormal changes in the target variable. If the explanatory variable has a large impact on the target variable, the target variable can be changed within a smaller disturbance range. Each disturbed explanatory variable is combined with other unperturbed explanatory variables to generate abnormal data, such as abnormal data 1: [extraction temperature +5°C, extraction pressure, extraction time, stirring speed, stirring time].

[0044] Each abnormal data is input into the corresponding target model, that is, the extraction rate prediction model. The target model outputs the predicted target variable value, compares the predicted value with the predicted value of the historical normal data, calculates the deviation value, and combines all the deviation values to form the abnormal mode of each target model. For example, abnormal mode 1: [-5%, -3%, +2%], which corresponds to a 3% decrease in extraction rate, a 3% decrease in mixing uniformity, and a 2% increase in microbial inactivation rate. Each target model can generate one or more abnormal modes to fully cover various abnormal situations. By determining the abnormal mode, potential problems can be warned in advance, the cause of the abnormality can be accurately identified, production parameters can be optimized, manual intervention can be reduced, and the stability of the production process can be improved.

[0045] In a specific embodiment, the third server performs in-depth analysis on the production equipment based on the abnormal pattern and the second analysis result, specifically including the following steps: Calculate the deviation between the actual value of each target variable in the second analysis result and the predicted value obtained from the historical normal data, generate the deviation pattern of the current production equipment based on the deviation value combination, calculate the similarity between the deviation pattern and each abnormal pattern, select the abnormal pattern with a similarity greater than the first threshold as the current abnormal pattern, establish a fault prediction model based on the deep learning model, obtain fault data related to multiple abnormal patterns, the fault data includes fault type, fault occurrence time and fault cause, input the abnormal pattern and the corresponding fault data into the fault prediction model for training, input the current abnormal pattern into the trained fault prediction model, obtain the type of fault that may occur in the current abnormal pattern, and the cause of the fault, and define it as the third analysis result.

[0046] Specifically, a deviation pattern of the current production equipment is generated based on the deviation value, and is matched with the abnormal patterns in the predefined abnormal pattern library. The average value of the differences of the same category of deviation values is calculated. The smaller the average value, the greater the similarity. The abnormal pattern with a similarity greater than the first threshold is defined as the abnormal pattern of the current production equipment. The second server selects a corresponding fault prediction model according to the identified abnormal pattern. For example, abnormal pattern 1: [-5%, -3%, +2%], and obtains the fault data related to this abnormal pattern, including the fault type such as equipment failure, the fault occurrence time: 3 times occurred within the past 6 months, and the fault causes, such as temperature sensor failure and pressure sensor failure. The fault data is input into the fault prediction model for training. The current abnormal pattern is input into the trained fault prediction model to obtain the possible fault types and fault causes. The user terminal sends the fault information to the user terminal to display a fault warning: for example, "Extraction equipment fault warning: Temperature sensor failure may occur", and the fault cause analysis: display a fault cause analysis report, for example, "The fault cause may be the aging or damage of the temperature sensor." In a specific embodiment, optimizing and adjusting the production process based on the obtained first analysis result or third analysis result specifically includes the following steps: Record the optimization operations performed on the production process based on the first analysis result and the third analysis result, obtain the production status of the corresponding production stage after the optimization operations are performed, and label the optimization operations. Establish an optimization model based on reinforcement learning, input the optimization operations and the production status after the optimization operations into the optimization model, use the optimization operations as actions and the production status as rewards, set a reward function based on the production target, evaluate each action based on the reward function to obtain the optimal action, and perform the optimization operation based on the optimal action.

[0047] Specifically, the user terminal receives the first analysis result or the third analysis result from the first server or the third server. The user makes adjustments according to the equipment with abnormal states in the analysis result. However, when making adjustments, the range of parameters may not be particularly determined. The present invention establishes a reinforcement learning model based on the deep Q-network (DQN). The reinforcement learning model defines the current production status as the state, and the state vector: [extraction temperature, extraction pressure, stirring speed, product quality index, equipment status]. The adjusted parameters or the replaced equipment are defined as actions. The user terminal designs a reward function according to the production target to evaluate the quality of each action.

[0048] For example, the goal is to increase the extraction rate while reducing energy consumption. Then the corresponding reward = w1×(actual extraction rate - target extraction rate) / target extraction rate×100% - w2×(actual energy consumption - target energy consumption) / target energy consumption×100%, where w1 and w2 are weight coefficients. The reinforcement learning model evaluates the expected cumulative reward of each action based on the current state and the reward function, and selects the optimal action. For example, if the expected cumulative reward for increasing the extraction temperature by 1°C is the highest, then increasing the extraction temperature by 1°C is selected as the optimal action. The optimization and adjustment method based on reinforcement learning can significantly improve the automation adjustment efficiency of the aloe vera beverage production process, adapt to the changes and uncertainties in the production process, reduce manual operations, and improve production efficiency.

[0049] The above describes a production control method for aloe vera beverages in the embodiments of the present application. Next, a production control system for aloe vera beverages in the embodiments of the present application will be described. Please refer to Figure 4 , an embodiment of a production control system for aloe vera beverages in the embodiments of the present application includes: A collection module, the sensor module obtains the key data of the production equipment and sends the key data and the unique identifier of the production equipment itself to the first server; An abnormality judgment module, the first server analyzes the production status of the corresponding production equipment based on the unique identifier to obtain a first analysis result, and sends the first analysis result to the second server. The second server judges whether the production equipment has an abnormality based on the first analysis result. If an abnormality occurs, the first analysis result is directly sent to the user terminal, and step S5 is executed; An abnormality prediction module, if no abnormality occurs, the second server sends the key data, the unique identifier, and the first analysis result to the third server. The third server includes multiple target models, and analyzes the key data and the first analysis result based on the target models to obtain a second analysis result; An abnormality analysis module, an abnormality pattern of each target model is established. The third server deeply analyzes the production equipment based on the abnormality pattern and the second analysis result to obtain a third analysis result, and sends the third analysis result to the user terminal; An optimization module, the user terminal optimizes and adjusts the production process based on the obtained first analysis result or third analysis result.

[0050] Those skilled in the art can clearly understand that for the convenience and brevity of description, the specific working processes of the above-described system, system, and unit can refer to the corresponding processes in the foregoing method embodiments and will not be repeated here.

[0051] When the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of this application. The aforementioned storage medium includes: various media that can store program codes, such as USB flash drives, mobile hard disks, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical discs.

[0052] As described above, the above embodiments are only used to illustrate the technical solutions of this application, rather than to limit them; although this application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of various embodiments of this application.

Claims

1. A production control method for aloe vera beverage, characterized in that, The method includes: Step S1: The sensor module acquires the key data of the production equipment and sends the key data and the unique identifier of the production equipment itself to the first server; Step S2: The first server analyzes the production status of the corresponding production equipment based on the unique identifier, obtains a first analysis result, and sends the first analysis result to the second server. The second server determines whether the production equipment is abnormal based on the first analysis result. If it is abnormal, the first analysis result is directly sent to the user terminal, and step S5 is executed; Step S3: If no abnormality occurs, the second server sends the key data, the unique identifier, and the first analysis result to the third server. The third server includes multiple target models, and analyzes the key data and the first analysis result based on the target models to obtain a second analysis result; Step S4: An abnormal mode for each target model is established. The third server deeply analyzes the production equipment based on the abnormal mode and the second analysis result to obtain a third analysis result, and sends the third analysis result to the user terminal; Step S5: The user terminal optimizes and adjusts the production process based on the obtained first analysis result or third analysis result.

2. The method according to claim 1, wherein The first server analyzing the production status of the corresponding production equipment based on the unique identifier includes the following steps: The key data is divided into multiple data categories based on different data performances, including alarm data, parameter data, general data, and performance index data. Based on the data categories, the analysis of the production status is divided into two implementation methods. For the key data of the same category, the first server compares each data item in the key data with a preset threshold to determine whether it meets the normal range of the corresponding data category. If it does not meet, the status of the key data is determined to be an abnormal state; For the key data of different categories, the first server combines the key data of different data categories based on the detection target to obtain multiple sample data, and performs multi-dimensional analysis on the sample data to obtain the analysis result of whether the production equipment is in an abnormal state.

3. The method according to claim 2, wherein Directly sending the first analysis result to the user terminal further includes the following steps: The abnormal state is divided into multiple abnormal types from simple to complex, including basic type, associated type, type to be predicted, and type to be optimized. An urgency level is set for each abnormal type, and the urgency level decreases successively from simple to complex of the abnormal type. Different notification methods are selected based on the urgency level to send the first analysis result to the user terminal.

4. The method according to claim 1, wherein Analyzing the key data and the first analysis result based on the target model includes the following steps: Obtain the key data of the aloe vera beverage during the normal production process, which is defined as historical normal data. Extract multiple variables from the historical normal data, and define the explanatory variable and the target variable from them. Based on the explanatory variable, establish a prediction model for the target variable. The prediction model is a mathematical model or a machine learning model. Each production stage includes one or more target models. The target model outputs the predicted value of the target variable. Identify the production stage where each key data is currently located based on the unique identifier, select the corresponding target model based on the production stage, and input the key data into the corresponding target model to obtain the actual value of each target variable for each production stage. Define the actual value as the second analysis result.

5. The method according to claim 4, wherein The steps for establishing the abnormal pattern of each target model are as follows: Determine the perturbation method based on each explanatory variable in the historical normal data. The perturbation method is to add abnormal values to each explanatory variable within a preset range while keeping other explanatory variables unchanged. The changed explanatory variable and other explanatory variables are combined to generate abnormal data. Input each abnormal data into the corresponding target model. The target model outputs the predicted value of the target variable. Compare the predicted value of the target variable with the predicted value of the historical normal data to obtain the deviation value for each target variable. Combine all the deviation values to obtain the abnormal pattern of each target model. There is one or more abnormal patterns for each target model.

6. The method according to claim 5, characterized in that, The in-depth analysis of the production equipment by the third server based on the abnormal pattern and the second analysis result includes the following steps: Calculate the deviation value of the actual value of each target variable in the second analysis result from the predicted value obtained from the historical normal data. Generate the deviation pattern of the current production equipment based on the deviation value combination, and calculate the similarity between the deviation pattern and each abnormal pattern. Select the abnormal pattern with a similarity greater than the first threshold and define it as the current abnormal pattern. Establish a fault prediction model based on a deep learning model. Obtain various fault data related to the abnormal patterns. The fault data includes the fault type, the fault occurrence time, and the fault cause. Input the abnormal pattern and the corresponding fault data into the fault prediction model for training, and input the current abnormal pattern into the trained fault prediction model to obtain the possible fault type and the cause of the fault in the current abnormal pattern, and define it as the third analysis result.

7. The method according to claim 1, wherein The steps for optimizing and adjusting the production process based on the obtained first analysis result or third analysis result are as follows: Record the optimization operations performed on the production process based on the first analysis result and the third analysis result, obtain the production status of the corresponding production stage after the optimization operations are performed, and label the optimization operations. Establish an optimization model based on reinforcement learning, input the optimization operations and the production status after the optimization operations into the optimization model, use the optimization operations as actions, and the production status as rewards. Set a reward function based on the production target, evaluate each action based on the reward function to obtain the optimal action, and perform the optimization operations based on the optimal action.

8. A production control system for aloe vera beverages, which is used to implement a production control method for aloe vera beverages as described in any one of claims 1-7, characterized in that, The system includes: A collection module, where the sensor module obtains the key data of the production equipment and sends the key data and the unique identifier of the production equipment itself to the first server; An abnormality judgment module, where the first server analyzes the production status of the corresponding production equipment based on the unique identifier to obtain a first analysis result, and sends the first analysis result to the second server. The second server judges whether the production equipment has an abnormality based on the first analysis result. If an abnormality occurs, it directly sends the first analysis result to the user terminal and executes step S5; An abnormality prediction module, if no abnormality occurs, the second server sends the key data, the unique identifier, and the first analysis result to the third server. The third server includes multiple target models, and analyzes the key data and the first analysis result based on the target models to obtain a second analysis result; An abnormality analysis module, establish the abnormality patterns of each target model, and the third server deeply analyzes the production equipment based on the abnormality patterns and the second analysis result to obtain a third analysis result, and sends the third analysis result to the user terminal; An optimization module, where the user terminal performs optimization adjustments on the production process based on the obtained first analysis result or third analysis result.

Images