Selenium-rich solid beverage production whole-process intelligent quality control system and method

Abstract

The invention discloses a selenium-rich solid beverage production whole-process intelligent quality control system and method, and relates to the technical field of food processing and intelligent manufacturing, the selenium-rich solid beverage production whole-process intelligent quality control system comprises a data acquisition module, a process mark analysis module, a feedback control module and a quality control report generation module, the data acquisition module is in communication connection with the process mark analysis module, and the process mark analysis module is in communication connection with the feedback control module; the data acquisition module is used for synchronously acquiring multi-source real-time process parameter data including temperature, humidity, pressure, mixing torque, spraying rotating speed and online selenium content detection value from a plurality of key working sections in the selenium-rich solid beverage production process; the process mark analysis module is used for adding a process flow position label to the acquired data and establishing a process behavior track of the current batch; the system has the advantages of being high in practicability and high in intelligent degree.

Description

Technical Field

[0001] This invention relates to the field of food processing and intelligent manufacturing technology, specifically to an intelligent quality control system and method for the entire production process of selenium-enriched solid beverages. Background Technology

[0002] Selenium-enriched solid beverages have gained widespread attention in the food industry in recent years due to the health benefits of selenium, a functional component that enhances immunity and has antioxidant properties. However, selenium is highly susceptible to decomposition or loss during production due to factors such as heat, moisture, and oxidative environments, resulting in large fluctuations in the selenium content of the final product, poor production stability, and difficulty in controlling quality consistency. Traditional selenium-enriched beverage production mainly relies on manual experience or static quality control models based on single parameters, which are insufficient to address the dynamic changes in selenium content caused by multiple process disturbances.

[0003] Currently, in the field of selenium-enriched foods and functional beverages, there is a lack of a systematic quality control method that can span the entire process, integrate real-time data, intelligently identify anomalies, and provide feedback and intervention. Especially in key stages involving spray drying, homogenization, and stable addition of selenium sources, traditional process control methods often rely solely on set equipment parameters or post-batch sampling data. This fails to dynamically track and provide early warnings of selenium content fluctuations, and even more so to correct potential process drift in real time, leading to fluctuations in product qualification rates and increased energy consumption. Therefore, it is essential to design a practical and highly intelligent intelligent quality control system and method for the entire production process of selenium-enriched solid beverages. Summary of the Invention

[0004] The purpose of this invention is to provide an intelligent quality control system and method for the entire production process of selenium-enriched solid beverages, so as to solve the problems mentioned in the background art.

[0005] To address the aforementioned technical problems, this invention provides the following technical solution: an intelligent quality control system and method for the entire production process of selenium-enriched solid beverages, comprising a data acquisition module, a process marker analysis module, a feedback control module, and a quality control report generation module. The data acquisition module is communicatively connected to the process marker analysis module, and the process marker analysis module is communicatively connected to the feedback control module. The data acquisition module, process marker analysis module, and feedback control module are each communicatively connected to the quality control report generation module. The data acquisition module is used to simultaneously collect multi-source real-time process parameter data, including temperature, humidity, pressure, mixing torque, spray speed, and online selenium content detection value, from multiple key stages in the production process of selenium-enriched solid beverages. The process tagging and analysis module is used to attach process flow location tags to the above-mentioned collected data and establish the process behavior trajectory of the current batch. The feedback control module is used to perform matching analysis between the process behavior trajectory and the historical batch process database, identify potential process deviations or abnormal trends, automatically generate intervention suggestions or adjustment parameter combinations, and execute process correction operations. The quality control report generation module is used to integrate all collected data, warning records, adjustment parameters and production results after the batch is completed to generate a standard electronic quality control report.

[0006] According to the above technical solution, the data acquisition module includes a multi-channel signal access unit, a signal synchronization verification unit, and an image acquisition module; wherein, The multi-channel signal access unit is used to access signal channels from temperature sensor, humidity sensor, pressure gauge, stirring motor, ammeter, and online selenium content analyzer, and to complete data preprocessing; The signal synchronization verification unit is used to calibrate the data sampling timing between channels and construct a unified timestamp to ensure the parallel validity of multi-dimensional process parameter data. The image acquisition module is used to capture and archive images of key visual process sections, and the images can be used by the anomaly identification module or for manual review.

[0007] According to the above technical solution, the process marking analysis module includes a process space polyline construction unit, a historical data matching unit, a similarity analysis unit, and an anomaly detection module; wherein, The process space polyline construction unit is used to construct a spatial process mapping diagram and output a process fluctuation curve with each process node position as the horizontal axis and the selenium content collected at each corresponding time as the vertical axis. The historical data matching unit is used to call up historical records in the historical batch process database that are similar to the current batch of raw material batches, equipment combinations, and target content conditions, and extract their key process paths. The similarity analysis unit is used to calculate and output a similarity index based on the curve and the broken line sequence of the historical model. The anomaly detection module is used to identify whether there are process anomalies or potential fluctuation trends, issue anomaly warning signals in a timely manner, and provide process segment positioning basis for subsequent feedback and adjustment.

[0008] According to the above technical solution, the feedback control module includes a parameter deviation identification unit, a parameter fine-tuning execution unit, a trend observation unit, and an intervention record generation unit; wherein, The parameter deviation identification unit is used to identify whether the current real-time selenium content has deviated from the set allowable fluctuation range based on the difference and trend between the current real-time selenium content and the target selenium content, and to assess whether to trigger the intervention logic. The parameter fine-tuning execution unit is used to adjust the selected process parameters within a preset range of priority after the intervention condition is triggered. The trend observation unit is used to assess the regression trend over two consecutive sampling periods after adjustment. The intervention record generation unit is used to record all intervention processes after the intervention is completed and write them into the electronic quality control archive and model training dataset.

[0009] According to the above technical solution, the anomaly detection module further includes a distribution segment difference modeling submodule and a process anomaly trend judgment submodule; wherein, The distribution segment difference modeling submodule is used to establish a fluctuation early warning model for process flow nodes based on historical batch data; The process anomaly trend judgment submodule is used to analyze the distribution density and continuity of process nodes that meet the judgment criteria in the process flow, and to determine whether they constitute an abnormal trend segment.

[0010] A method for intelligent quality control of the entire production process of selenium-enriched solid beverages, comprising the following steps: Step S1: In response to receiving the production scheduling instruction, initiate the batch plan and initialize key quality control parameters, including the target selenium content of the product. Initial settings for key processes, batch numbers and historical quality scores of raw materials used, and allowable fluctuation range. And the setting of key monitoring points; Step S2: During the production process, temperature values ​​are collected in real time through the data acquisition module. Humidity value ,pressure Spray speed Mixed torque Online selenium content The current data collection location is recorded as the process node. ; Step S3: Record the timestamp corresponding to each process node. To form a triple Attached process location ; Step S4: Based on historical deviation analysis, identify key process sections that may cause fluctuations in selenium content and mark them as key dynamic monitoring sections. Establish a process fluctuation model and compare it with the historical data. Calculate the similarity index γ and assess whether there are any process anomalies. Step S5: In response to the process abnormality warning signal, the system issues an early warning and activates the feedback adjustment mechanism, driving the equipment to correct itself through parameter fine-tuning suggestions, thereby achieving dynamic process control; Step S6: After the batch is completed, the system uses all process data, control records and result information for model regression learning and optimization, generates a standard electronic quality control report, and realizes self-learning and full-process traceability.

[0011] According to the above technical solution, step S4 includes: Step S41: Establish a historical batch process database and compare the current raw material batch number, equipment combination, and target selenium content. The deviation rate from the historical finished products in the database is greater than The sample was analyzed as follows: Traverse the historical database for batches containing the same raw material batches or equipment combinations, extract their process flow, and mark the concentrated areas of selenium content deviation as... ; For the current production batch and Overlapping process nodes Label the section as a "key dynamic monitoring section"; Step S42: Collect real-time data from each time period in step S3. Projected onto the process flow diagram, and with The vertical axis is The horizontal axis is... and To support information channels, a spatial process mapping diagram is constructed; Step S43: For the marked "key dynamic monitoring sections" nodes, establish a spatial broken line model of selenium content fluctuation. Then, the average content sequence of the same work section in history is used to construct a historical average model. ; Step S44: Calculate the spatial broken line model of selenium content fluctuation. Compared with historical average models Similarity index Its calculation expression is: ; in Represents the similarity index. , The lower the value, the stronger the abnormality. This represents the measured value of the i-th selenium content index in actual production; This represents the measured value of the i-th selenium content indicator in historical production; n represents the total number of selenium content indicators.

[0012] According to the above technical solution, step S4 further includes an assessment of whether there is an abnormal trend, as follows: Step S45: Establish a distribution segment difference early warning model and set a matching threshold. When any critical process segment If so, then mark the segment as a "potential process drift zone"; Step S46: Continue to count the number N of consecutive low-matching regions within this segment, using each unit of process distance as a node. If If so, a process abnormality warning signal will be output.

[0013] According to the above technical solution, step S5 further includes: Step S51: After the system identifies a "potential process drift zone," it immediately locates the process section number and historical anomaly tag of the drift zone and triggers a parameter correction request. The system then adjusts the parameters according to the target selenium content of the current batch. With real-time content The difference is used to determine whether intervention is needed; Step S52: When and The difference exceeds the dynamic threshold range If an intervention is initiated, the highest priority parameter among the controllable parameters of that process segment will be selected. Following the strategy of "single-factor trial and step-by-step response," a small adjustment will be made to the single parameter, with the initial adjustment not exceeding the basic set value. ; Step S53: During the two most recent data acquisition cycles after parameter adjustment, the system continuously monitors the selenium content. Does it appear to be directed towards the target? If the regression trend fails to reach the standard of significant regression change or shows an aggravated deviation after two consecutive cycles, the system will pause the adjustment of this parameter and switch to the next priority parameter to perform a new round of fine-tuning. Step S54: If the parameter switching fails after three consecutive attempts... If the trend is reversed, the key process paths of the successful amendment examples in the past ten batches of this process segment are compared, and the adjustment combination with the highest matching degree is extracted. Then, a "combination recovery adjustment" process is performed according to its parameter path. Step S55: After the adjustment is completed, the system generates an abnormal intervention record for this batch and marks the specific parameters, response time, and stabilization time information that were affected by the intervention. This information is written into the electronic quality control file and used for model updates and subsequent batch process prediction optimization.

[0014] Compared with existing technologies, the beneficial effects achieved by this invention are as follows: This invention can dynamically identify fluctuation trends through a similarity index before quality deviations exceed control limits, achieving precise, feedforward-based control throughout the entire process, significantly improving the consistency and stability of key nutritional indicators such as selenium content. Furthermore, by employing multi-level control strategies such as spatial polyline modeling, trend identification, feedback fine-tuning, and combined correction, it not only enhances the sensitivity of abnormal responses and the accuracy of control adjustments but also effectively avoids system disturbances caused by excessive intervention through a dynamic regulation mechanism, thereby ensuring stable product quality output. Attached Figure Description

[0015] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings: Figure 1 This is a schematic diagram of the system module composition of the present invention. Detailed Implementation

[0016] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0017] Please see Figure 1 This invention provides a technical solution: an intelligent quality control system for the entire production process of selenium-enriched solid beverages, comprising a data acquisition module, a process marker analysis module, a feedback control module, and a quality control report generation module. The data acquisition module is communicatively connected to the process marker analysis module, and the process marker analysis module is communicatively connected to the feedback control module. The data acquisition module, process marker analysis module, and feedback control module are each communicatively connected to the quality control report generation module. The data acquisition module is used to simultaneously collect multi-source real-time process parameter data, including temperature, humidity, pressure, mixing torque, spray speed, and online selenium content detection value, from multiple key stages in the production process of selenium-enriched solid beverages. The process tagging and analysis module is used to attach process flow location tags to the above-mentioned collected data and establish the process behavior trajectory of the current batch. The feedback control module is used to perform matching analysis between the process behavior trajectory and the historical batch process database, identify potential process deviations or abnormal trends, automatically generate intervention suggestions or adjustment parameter combinations, and execute process correction operations. The quality control report generation module is used to integrate all collected data, warning records, adjustment parameters and production results after the batch is completed to generate a standard electronic quality control report.

[0018] The data acquisition module includes a multi-channel signal access unit, a signal synchronization verification unit, and an image acquisition module; among which, The multi-channel signal access unit is used to access signal channels from temperature sensor, humidity sensor, pressure gauge, stirring motor, ammeter, and online selenium content analyzer, and to complete data preprocessing; The signal synchronization verification unit is used to calibrate the data sampling timing between channels and construct a unified timestamp to ensure the parallel validity of multi-dimensional process parameter data. The image acquisition module is used to capture and archive images of key visual process sections, and the images can be used by the anomaly identification module or for manual review.

[0019] The process tagging and analysis module includes a process space polyline construction unit, a historical data matching unit, a similarity analysis unit, and an anomaly detection module; among them, The process space polyline construction unit is used to construct a spatial process mapping diagram and output a process fluctuation curve with each process node position as the horizontal axis and the selenium content collected at each corresponding time as the vertical axis. The historical data matching unit is used to call up historical records in the historical batch process database that are similar to the current batch of raw material batches, equipment combinations, and target content conditions, and extract their key process paths. The similarity analysis unit is used to calculate and output a similarity index based on the curve and the broken line sequence of the historical model. The anomaly detection module is used to identify whether there are process anomalies or potential fluctuation trends, issue anomaly warning signals in a timely manner, and provide process segment positioning basis for subsequent feedback and adjustment.

[0020] The feedback control module includes a parameter deviation identification unit, a parameter fine-tuning execution unit, a trend observation unit, and an intervention record generation unit; among them, The parameter deviation identification unit is used to identify whether the current real-time selenium content has deviated from the set allowable fluctuation range based on the difference and trend between the current real-time selenium content and the target selenium content, and to assess whether to trigger the intervention logic. The parameter fine-tuning execution unit is used to adjust the selected process parameters within a preset range of priority after the intervention condition is triggered. The trend observation unit is used to assess the regression trend over two consecutive sampling periods after adjustment. The intervention record generation unit is used to record all intervention processes after the intervention is completed, including adjustment parameters, effective time, response lag, effect evaluation, etc., and write them into the electronic quality control file and model training dataset.

[0021] The anomaly detection module further includes a distribution segment difference modeling submodule and a process anomaly trend judgment submodule; among which, The distribution segment difference modeling submodule is used to establish a fluctuation early warning model for process flow nodes based on historical batch data; The process anomaly trend judgment submodule is used to analyze the distribution density and continuity of process nodes that meet the judgment criteria in the process flow, and to determine whether they constitute an abnormal trend segment.

[0022] A method for intelligent quality control of the entire production process of selenium-enriched solid beverages, comprising the following steps: Step S1: In response to receiving the production scheduling instruction, initiate the batch plan and initialize key quality control parameters, including the target selenium content of the product. Initial settings for key processes, batch numbers and historical quality scores of raw materials used, and allowable fluctuation range. And the setting of key monitoring points; Step S2: During the production process, temperature values ​​are collected in real time through the data acquisition module. Humidity value ,pressure Spray speed Mixed torque Online selenium content The current data collection location is recorded as the process node. ; Step S3: Record the timestamp corresponding to each process node. To form a triple Attached process location ; Step S4: Based on historical deviation analysis, identify key process sections that may cause fluctuations in selenium content and mark them as key dynamic monitoring sections. Establish a process fluctuation model and compare it with the historical data. Calculate the similarity index γ and assess whether there are any process anomalies. Step S5: In response to the process abnormality warning signal, the system issues an early warning and activates the feedback adjustment mechanism, driving the equipment to correct itself through parameter fine-tuning suggestions, thereby achieving dynamic process control; Step S6: After the batch is completed, the system uses all process data, control records and result information for model regression learning and optimization, generates a standard electronic quality control report, and realizes self-learning and full-process traceability.

[0023] Step S4 includes: Step S41: Establish a historical batch process database and compare the current raw material batch number, equipment combination, and target selenium content. The deviation rate from the historical finished products in the database is greater than The sample was analyzed as follows: Traverse the historical database for batches containing the same raw material batches or equipment combinations, extract their process flow, and mark the concentrated areas of selenium content deviation as... ; For the current production batch and Overlapping process nodes Label the section as a "key dynamic monitoring section"; Step S42: Collect real-time data from each time period in step S3. Projected onto the process flow diagram, and with The vertical axis is The horizontal axis is... and To support information channels, a spatial process mapping diagram is constructed; Step S43: For the marked "key dynamic monitoring sections" nodes, establish a spatial broken line model of selenium content fluctuation. Then, the average content sequence of the same work section in history is used to construct a historical average model. ; Step S44: Calculate the spatial broken line model of selenium content fluctuation. Compared with historical average models Similarity index Its calculation expression is: ; in Represents the similarity index. , The lower the value, the stronger the abnormality. This represents the measured value of the i-th selenium content index in actual production; This represents the measured value of the i-th selenium content indicator in historical production; n represents the total number of selenium content indicators.

[0024] Step S4 further includes an assessment of whether there are any abnormal trends, as follows: Step S45: Establish a distribution segment difference early warning model and set a matching threshold. When any critical process segment If so, then mark the segment as a "potential process drift zone"; Step S46: Continue to count the number N of consecutive low-matching regions within this segment, using each unit of process distance as a node. If If so, a process abnormality warning signal will be output.

[0025] Step S5 further includes: Step S51: After the system identifies a "potential process drift zone," it immediately locates the process section number and historical anomaly tag of the drift zone and triggers a parameter correction request. The system then adjusts the parameters according to the target selenium content of the current batch. With real-time content The difference is used to determine whether intervention is needed; Step S52: When and The difference exceeds the dynamic threshold range If an intervention operation is initiated, the highest priority parameter among the controllable parameters of that process segment (including but not limited to drying temperature, atomization speed, and mixing torque) will be selected. Following the strategy of "single-factor trial and step-by-step response," a small adjustment will be made to the single parameter, with the initial adjustment not exceeding the basic set value. ; Step S53: During the two most recent data acquisition cycles after parameter adjustment, the system continuously monitors the selenium content. Does it appear to be directed towards the target? If the regression trend fails to reach the standard of significant regression change or shows an aggravated deviation after two consecutive cycles, the system will pause the adjustment of this parameter and switch to the next priority parameter to perform a new round of fine-tuning. Step S54: If the parameter switching fails after three consecutive attempts... If the trend is reversed, the key process paths of the successful amendment examples in the past ten batches of this process segment are compared, and the adjustment combination with the highest matching degree is extracted. Then, a "combination recovery adjustment" process is performed according to its parameter path. Step S55: After the adjustment is completed, the system generates an abnormal intervention record for this batch and marks the specific parameters, response time, and stabilization time information that were affected by the intervention. This information is written into the electronic quality control file and used for model updates and subsequent batch process prediction optimization. This application constructs an intelligent quality control mechanism for the entire process of selenium-enriched solid beverage production, enabling dynamic tracking of process behavior, comparison with historical models, and early warning of abnormal trends. Its core lies in uniformly calibrating and labeling process nodes with multi-source process parameters (such as temperature, humidity, mixing torque, spray speed, and online selenium content) collected from each production batch, thereby establishing a complete "process behavior trajectory." Based on this trajectory, the system calls upon a historical batch process database and, combined with key variables such as raw material batch number, target selenium content, and equipment combination, extracts representative historical sample data to establish a "historical fluctuation concentration zone" model. Subsequently, through spatial polyline modeling technology, the real-time data of the current batch is projected onto the process flow, and the similarity of the fluctuation curves formed with the historical model is compared to extract key sections with significant selenium content fluctuations. Furthermore, this application constructs a "distribution segment difference early warning mechanism" by setting a similarity threshold and a standard for the number of consecutive low-matching areas. When a certain process section continuously exhibits fluctuations that significantly deviate from historical trends, it can be identified as a potential "process drift zone." The system automatically triggers a feedback control mechanism, which fine-tunes key process parameters through priority rules to achieve dynamic correction control at the source.

[0026] Through the above steps, it is possible to conduct key monitoring and analysis on sections with high fluctuations in actual production, effectively avoiding systematic deviations in key indicators such as selenium content; on the other hand, for most stable sections, redundant calculations and invalid early warnings are reduced, thereby optimizing computing resources and improving the overall system response efficiency.

[0027] Furthermore, through continuous system operation and learning, this invention supports iterative updates and feedback learning mechanisms for the quality control model, continuously accumulating process optimization experience, improving the accuracy of prediction and early warning for new batches, and ultimately forming a "self-evolving" quality control model library. Especially when dealing with changes in different raw material batches and equipment combinations, it can achieve rapid adaptation and strategy adjustment, effectively improving the consistency and nutritional stability of selenium-enriched beverage products. Ultimately, it not only significantly improves the process stability and product quality consistency of selenium-enriched solid beverage production, but also realizes a closed-loop quality control process from monitoring, identification to regulation, achieving the technical effects of early warning, rapid adjustment, and precise control.

[0028] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0029] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0030] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0031] The embodiments of the present invention have been described above with reference to the accompanying drawings. However, the present invention is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of the present invention without departing from the spirit and scope of the claims. All of these forms are within the protection scope of the present invention.

Claims

1. A full-process intelligent quality control system for the production of selenium-rich solid beverages, characterized by, The full-process intelligent quality control system for the selenium-rich solid beverage production comprises a data acquisition module, a process marker analysis module, a feedback control module and a quality control report generation module, the data acquisition module is in communication connection with the process marker analysis module, the process marker analysis module is in communication connection with the feedback control module, the data acquisition module, the process marker analysis module and the feedback control module are in communication connection with the quality control report generation module respectively, and the full-process intelligent quality control system for the selenium-rich solid beverage production comprises the following steps: The data acquisition module is used for synchronously collecting multi-source real-time process parameter data including temperature, humidity, pressure, mixing torque, spraying speed and online selenium content detection value from multiple key sections in the selenium-rich solid beverage production process. The process marker analysis module is used for attaching a process flow position label to the collected data and establishing a process behavior trajectory of the current batch. The feedback control module is used for matching and analyzing the process behavior trajectory with a historical batch process database, identifying potential process deviations or abnormal trends, automatically generating intervention suggestions or adjusting parameter combinations and performing process correction operations. The quality control report generation module is used for integrating all collected data, warning records, adjusting parameters and production results to generate a standard electronic quality control report after the batch ends.

2. The intelligent quality control system for the whole process of selenium-rich solid beverage production according to claim 1, characterized in that: The data acquisition module comprises a multi-channel signal access unit, a signal synchronization verification unit and an image acquisition module, wherein The multi-channel signal access unit is used for accessing signal channels from temperature sensors, humidity sensors, pressure gauges, stirring motors, ammeters and online selenium content analyzers and completing data preprocessing; The signal synchronization verification unit is used for calibrating the data sampling time sequence between channels and constructing a unified timestamp to ensure the parallel effectiveness of multi-dimensional process parameter data; The image acquisition module is used for image shooting and archiving of key visualized process sections, and the images can be used by an abnormality identification module or artificial review. 3.The intelligent quality control system for the whole process of selenium-rich solid beverage production according to claim 1, characterized in that: The process marker analysis module comprises a process space polyline construction unit, a historical data matching unit, a similarity analysis unit and an abnormality identification module, wherein The process space polyline construction unit is used for constructing a space process mapping diagram and outputting a process fluctuation curve with each process node position as the horizontal axis and the selenium content collected at each corresponding time as the vertical axis; The historical data matching unit is used for calling historical records similar to the current batch raw material batch, equipment combination and target content conditions in the historical batch process database and extracting the key process path thereof; The similarity analysis unit is used for calculating a similarity index based on the curve and the polyline sequence of the historical model; The abnormality identification module is used for identifying whether there is a process abnormality or a potential fluctuation trend, timely issuing an abnormality warning signal and providing a process section positioning basis for subsequent feedback adjustment.

4. The intelligent quality control system for the whole process of selenium-rich solid beverage production according to claim 1, characterized in that: The feedback control module comprises a parameter deviation identification unit, a parameter fine-tuning execution unit, a trend observation unit and an intervention record generation unit, wherein The parameter deviation identification unit is used for identifying whether the current real-time selenium content has deviated from the set allowable fluctuation range according to the difference and trend between the current real-time selenium content and the target selenium content and evaluating whether the intervention logic is triggered. The parameter fine-tuning execution unit is configured to perform priority adjustment of the selected process parameter within a preset amplitude after triggering the intervention condition; The trend observation unit is configured to evaluate the regression trend in the two sampling periods after the adjustment; The intervention record generation unit is configured to record all intervention processes after the intervention is completed and write into the electronic quality control archive and the model training data set.

5. The intelligent quality control system for the whole process of selenium-rich solid beverage production according to claim 3, characterized in that: The abnormality discrimination module further comprises a distribution section difference modeling submodule and a process abnormality trend judgment submodule; wherein, The distribution section difference modeling submodule is configured to establish a fluctuation early warning model of the process flow node based on the historical batch data; The process abnormality trend judgment submodule is configured to analyze the distribution density and continuity of the process nodes that meet the judgment criteria in the process flow, and judge whether they constitute an abnormal trend section.

6. A method for intelligent quality control of the whole process of producing selenium-rich solid beverage, characterized in that: The method comprises the following steps: Step S1 : In response to receiving the production scheduling instruction, start the batch plan and initialize key quality control parameters, including product target selenium content , key process initial setting parameters, raw material batch number used and historical quality score, allowable fluctuation range and key monitoring point settings; Step S2: In the production process, real-time acquisition of temperature value by data acquisition module , humidity value , pressure , spray rotation speed , mixing torque , online selenium content and the current acquisition position are recorded as process nodes ; Step S3: record the time stamp corresponding to each process node , constituting a triple , with the process position ; Step S4: Based on historical deviation analysis, identify the key process section that may cause selenium content fluctuation, and mark it as a key dynamic monitoring section. Establish a process fluctuation model and compare it with history to calculate the similarity index γ and evaluate whether there is a process abnormality; Step S5: In response to the process abnormality warning signal, the system issues a warning and starts a feedback adjustment mechanism, and through parameter fine-tuning suggestion, drives the equipment to correct, realizes dynamic process control; Step S6: After the batch is completed, the system uses all process data, control records and result information for model regression learning and optimization, generates a standard electronic quality control report, and realizes self-learning and full-process tracing.

7. The method according to claim 6, wherein the method is characterized by: The step S4 comprises: Step S41: Establish a historical batch process database, compare the current use of raw material batch number, equipment combination, target selenium content , and the sample in the database with a historical product deviation rate greater than , and perform the following analysis: Traverse the batches with the same raw material batch or equipment combination in the history database, extract the process flow thereof, and mark the section with concentrated selenium content deviation as ; Tagging the current production batch with the process node it overlaps with "Critical Dynamic Monitoring Section" Tagging the current production batch with the process node it overlaps with Step S42: Collecting real-time data of each time period in step S3 projected onto a process flow chart and plotted with the vertical axis, the horizontal axis, and auxiliary information channels, a spatial process map is constructed; Step S43: For the marked "key dynamic monitoring section" node, establish a selenium content fluctuation space polyline model , call the average content sequence of the same section position in history again to construct a historical average model ; Step S44: Calculate the selenium content fluctuation space polyline model Similarity index with historical average model The calculation expression is:​ ； wherein represents the similarity index, , The lower the value, the stronger the anomaly; represents the measured value of the i-th selenium content indicator in actual production; represents the measured value of the i-th selenium content indicator in historical production; n represents the total number of selenium content indicators. 8.The method according to claim 7, characterized in that: The step S4 further comprises an abnormal trend evaluation step, specifically as follows: Step S45: Establishing the distribution section difference early warning model, setting the matching threshold When any key process section is Then mark this section as "potential process drift area"; Step S46: Continue to count the number of continuous low matching regions N with each unit process distance as a node in the section, if then output a process abnormality warning signal. 9.The method according to claim 8, characterized in that: The step S5 further comprises: Step S51: When the system identifies the "potential process drift zone", it immediately locates the process section number and the historical abnormality label where the drift zone is located, and triggers a parameter correction request. The system determines whether intervention operation needs to be performed according to the difference between the target selenium content of the current batch and the real-time content . . Step S52: When and The difference exceeds the dynamic threshold range If an intervention is initiated, the highest priority parameter among the controllable parameters of that process segment will be selected. Following the strategy of "single-factor trial and step-by-step response," a small adjustment will be made to the single parameter, with the initial adjustment not exceeding the basic set value. ; Step S53: The system continuously monitors the selenium content in the last two data collection periods after the parameter adjustment Whether the trend of regression occurs If the significant regression change criterion is not met or the deviation intensifies for two consecutive periods, the system suspends the parameter adjustment and proceeds to the next priority parameter to perform a new round of fine-tuning operation. Step S54: If the parameter switching is not achieved for three times in succession If the regression trend is not eliminated, the key process path of the successful correction cases in the past ten batches of this process section is compared, and the adjustment combination with the highest matching degree is extracted. A "combination recovery adjustment" process is performed according to the parameter path. Step S55: After the adjustment is completed, the system generates the abnormal intervention record of the batch, and marks the specific parameters, response time and recovery time information of the intervention, writes it into the electronic quality control archive, and at the same time, is used for model updating and subsequent batch process prediction optimization.

Images