Digital twinning-based fruit juice HPP sterilization production line operation monitoring method

Abstract

The invention belongs to the technical field of state monitoring, and discloses a fruit juice HPP sterilization production line operation monitoring method based on digital twinning. Comprising the steps that bottled conveying state data and production line operation data are collected in real time, and data cleaning is executed to obtain a state monitoring original data set; state analysis is conducted on the conveying state, rhythm recognition is executed, and a bottle flow state data set is obtained; performing change trend analysis on the liquid filling process, and outputting a water injection state data set; pressurizing stage state parameters in the water injection state data set are extracted, fluctuation recognition is carried out, and bottle body matching state information is obtained; carrying out bottle body stress evaluation, and generating pressure relief adjustment reference data; equipment offset adjustment is executed, the synchronization error of multiple parallel cavities is monitored, and a synchronous regulation and control data set is obtained; the production line operation monitoring results are counted to obtain production line operation monitoring information; and the whole-process digital sensing capability and the multi-stage coordinated regulation flexibility of the operation monitoring of the sterilization production line are improved.

Description

A Digital Twin-Based Operation Monitoring Method for a Fruit Juice HPP Sterilization Production Line Technical Field

[0001] This invention relates to the field of status monitoring technology, and more specifically, to a method for monitoring the operation of a fruit juice HPP sterilization production line based on digital twins. Background Technology

[0002] Digital twin technology is a technique that uses virtual mapping models of physical entities to achieve real-time state perception and dynamic simulation prediction. It has been widely used in fields such as intelligent manufacturing and industrial production line monitoring. In particular, it has the potential to achieve full-process operation monitoring and multi-stage collaborative control in the HPP sterilization production line of juice. However, traditional operation monitoring methods still face many technical bottlenecks when dealing with complex production line operation scenarios.

[0003] Before the juice bottles enter the HPP sterilization equipment, there is a buffer section where the bottle flow state often changes dynamically. For example, excessively fast or slow bottle discharge can cause the equipment to idle or jam. However, traditional operation monitoring methods rely solely on fixed time parameters or conveyor belt length for cycle time estimation, neglecting the real-time nature of the bottle flow rhythm. This can easily lead to misalignment of the entry cycle time, resulting in material accumulation. Furthermore, the water injection rate in the few seconds before the equipment starts during the pre-filling stage affects the stability of the subsequent pressurization curve. For example, even slight differences in the pre-filling rate can easily cause residual air bubbles in the chamber. However, traditional operation monitoring methods often use a constant-rate water injection strategy and do not dynamically model the liquid level change trend, resulting in the inability to identify the risk of residual gas retention in a timely manner, affecting the stability of the subsequent pressurization process. Simultaneously, empty bottles or bottles with insufficient liquid levels often... This process generates pressure waves that absorb disturbances, resulting in localized micro-pressure wave oscillations. Traditional operation monitoring methods tend to overlook these minute changes in pressure waves during pressurization, limiting the ability to accurately identify abnormal bottles. On the other hand, the pressure release rate during the depressurization phase directly affects the stability of the bottle structure. However, traditional operation monitoring methods only set depressurization as a fixed time parameter without dynamically adjusting the depressurization gradient for different bottle states, leading to abnormal noises or disordered bubbles in some juice bottles after processing. Furthermore, during the operation of multiple parallel chambers, there are often very small synchronization deviations between different chambers. Traditional operation monitoring methods use a uniform start-up time configuration for multiple chambers without compensating for the actual response differences of each chamber, which may cause mismatches during downstream bottle unloading, affecting the overall operating efficiency of the production line.

[0004] In view of this, the present invention proposes a digital twin-based method for monitoring the operation of a fruit juice HPP sterilization production line to solve the above problems. Summary of the Invention

[0005] To overcome the aforementioned deficiencies of the prior art and achieve the above objectives, the present invention provides the following technical solution: a digital twin-based method for monitoring the operation of a fruit juice HPP sterilization production line, comprising: S1. Real-time acquisition of bottle conveying status data and production line operation data during the fruit juice production line process, and performing data cleaning to obtain a raw status monitoring dataset; S2. Based on the raw status monitoring dataset, performing status analysis on the conveying status to obtain bottle status information, and performing rhythm recognition based on the bottle status information to obtain a bottle flow status dataset; S3. Based on the bottle flow status dataset, performing trend analysis on the liquid filling process to identify abnormal filling results that interfere with the stable operation of the production line. S4. Output the water injection status dataset; S5. Extract the pressurization stage status parameters from the water injection status dataset, and perform fluctuation identification based on these pressurization stage status parameters to obtain the bottle matching status information; S6. Combine the bottle matching status information and the depressurization stage status parameters to perform bottle stress assessment and generate depressurization adjustment reference data; S7. Obtain the real-time operating status information of multiple parallel cavities, perform equipment offset adjustment in combination with the depressurization adjustment reference data, monitor the synchronization error of multiple parallel cavities, and obtain the synchronization control dataset; S8. Statistically analyze the production line operation monitoring results and obtain the production line operation monitoring information through archiving and integration, and send the production line operation monitoring information to the preset production line monitoring terminal.

[0006] Furthermore, the method for performing state analysis includes: dividing the original state monitoring dataset into periodic groups to obtain operating cycle parameter groups; extracting bottle count parameters from the operating cycle parameter groups and obtaining the corresponding production line conveyor belt speed; calculating the bottle quantity change rate for the operating cycle based on the bottle count parameters and the conveyor belt speed; constructing a standard virtual production line operating scenario and extracting simulated operating data from the virtual production line operating scenario; calculating the difference in bottle quantity change rate and the conveyor belt speed change rate between adjacent consecutive operating cycles, and comparing the conveyor belt speed change rate with the simulated speed change rate in the simulated operating data to obtain the speed change deviation; matching the speed change deviation and the bottle quantity change rate difference with the corresponding preset threshold interval to output a conveying state trend label; binding the conveying state trend label with the corresponding operating cycle number and the operating cycle parameter group to output bottle state information.

[0007] Furthermore, the execution rhythm recognition method includes: sorting the delivery status trend tags in the bottle status information according to time order, and extracting the trend type field corresponding to the running cycle; counting the number of consecutive running cycles with the same trend type field, and filtering the time segment consisting of running cycles with a number of consecutive running cycles higher than a preset continuous trend threshold as the continuous trend change period; calculating the number of consecutively pushed bottles based on the length of the continuous trend change period and the rate of change of the number of bottles corresponding to each running cycle; calculating the bottle delivery time range based on the number of consecutively pushed bottles, and taking the time segment corresponding to the bottle delivery time range as the bottle flow coverage time period; comparing the bottle flow coverage time period with the expected bottle entry time in the simulated running data, and outputting the actual bottle push deviation time; comparing the actual bottle push deviation time with the beat offset threshold, judging the bottle flow rhythm state of the corresponding continuous trend change period, and outputting the bottle flow status tag; using the bottle flow status tag to mark the corresponding running cycle, and outputting the bottle flow status dataset.

[0008] Furthermore, the method for performing the trend analysis includes: identifying the pre-filling water start command field and mapping it to the timestamp corresponding to the bottle flow state dataset; extracting the initial liquid level value at that timestamp and the corresponding operating cycle number, which are recorded as the pre-filling initial state parameters; obtaining the liquid filling end timestamp and using the time interval between the timestamp corresponding to the pre-filling water start command field and the liquid filling end timestamp as the liquid filling process segment; identifying the liquid level height of each operating cycle in the liquid filling process segment and constructing a liquid level height change time series; calculating the unit time liquid level change amount of the liquid level change time series and calculating the liquid level change rate based on the unit time liquid level change amount; and plotting the liquid level change curve based on the liquid level change rate. Based on simulated operation data, a theoretical liquid level change curve is plotted, and the theoretical liquid level change curve is compared with the actual liquid level change curve to output a liquid level deviation value sequence. Time intervals in the liquid level deviation value sequence where consecutive liquid level deviations are higher than a preset deviation threshold are identified and recorded as abnormal liquid level intervals. The duration of abnormal liquid level intervals is calculated, and the abnormality type is determined based on this duration, outputting a liquid level abnormality type field. The abnormality percentage of the abnormal liquid level interval in the liquid filling process is calculated; if this percentage is higher than a preset safety percentage threshold, the corresponding liquid filling process section is determined to have a risk of residual gas retention. The corresponding time intervals are labeled using the liquid level abnormality type field, the abnormality percentage, and the label of residual gas retention risk, outputting a water injection status dataset.

[0009] Furthermore, the method for determining the anomaly type includes: setting an anomaly duration range, comparing the duration length with the anomaly duration range, and classifying it into short-term anomalies, medium-term anomalies, and long-term anomalies; extracting the liquid level deviation value sequence of the anomaly segment, calculating the average deviation amplitude, and simultaneously counting the number of positive and negative alternations of the deviation values; if it is a short-term anomaly and the average deviation amplitude exceeds a preset short-term deviation threshold, it is determined to be either a fast instantaneous rate or a slow instantaneous rate based on the positive or negative direction of the deviation value; if it is a medium-term anomaly and the number of positive and negative alternations exceeds a preset frequency threshold, it is determined to be a fluctuation anomaly; otherwise, it is determined to be either a sustained rate that is too fast or a sustained rate that is too slow based on the positive or negative direction of the deviation value; if it is a long-term anomaly, it is determined to be a long-term anomaly; binding all anomaly type tags with the corresponding time segment and average deviation amplitude, and outputting the liquid level anomaly type field.

[0010] Furthermore, the state parameters of the pressurization stage include: the instantaneous pressure sequence corresponding to the pressurization process and the theoretical pressure value data generated based on the simulation operation data.

[0011] Further, the method for fluctuation identification includes: dividing the instantaneous pressure sequence corresponding to the pressurization process into segments and outputting instantaneous pressure segments; calculating the pressure standard deviation of the instantaneous pressure values ​​within each instantaneous pressure segment, and identifying instantaneous pressure segments with pressure standard deviations higher than a stability threshold as pressure unstable segments; identifying local extreme points in the instantaneous pressure sequence and statistically analyzing the distribution density of these local extreme points; if the distribution density of these local extreme points is higher than a preset density threshold, then the corresponding pressurization process is determined to have pressure oscillations; calculating the ratio of the pressure rise amplitude to the corresponding time length of the pressure unstable segment to obtain the average pressure rise rate of the pressure unstable segment; and calculating the pressure rise rate in the theoretical pressure value data relative to the pressure unstable segment. The theoretical pressure rise rate for the same time period is calculated, and the average pressure rise rate is compared with the theoretical pressure rise rate to calculate the pressure wave velocity attenuation ratio. The actual timestamp of reaching the preset target pressure value in the instantaneous pressure sequence is identified, and the theoretical timestamp of reaching the preset target pressure value in the theoretical pressure value data is identified simultaneously. The time difference between the actual timestamp and the theoretical timestamp is calculated as the pressure propagation deviation time. If pressure oscillation is satisfied, the pressure wave velocity attenuation ratio exceeds the preset attenuation threshold, and the pressure propagation deviation time is higher than the propagation deviation threshold, then the corresponding pressurization process is marked as a liquid level deficiency anomaly. The liquid level deficiency anomaly label is bound to the corresponding pressurization stage state parameters, and the bottle matching state information is output.

[0012] Further, the method for assessing bottle stress includes: extracting the instantaneous pressure drop sequence from the state parameters during the depressurization phase, calculating the pressure drop rate of adjacent drop segments, and constructing a pressure drop rate sequence; identifying labels in the bottle matching state information, setting a high-sensitivity bottle stress coefficient if the label is insufficient liquid level anomaly, and setting a sensitive bottle stress coefficient if the label is residual gas retention risk; calculating the stress change amplitude during the depressurization process based on the pressure drop rate sequence and the corresponding bottle stress coefficient; plotting the theoretical depressurization rate curve for the corresponding depressurization process based on simulated operation data, comparing the difference between the theoretical depressurization rate curve and the pressure drop rate sequence, and identifying drop segments where the difference is higher than the depressurization safety threshold as risk segments; if a drop segment is a risk segment and the stress change amplitude during the depressurization process is higher than a preset stress threshold, marking the drop segment as a depressurization gradient anomaly; calculating the predicted residual pressure inside the bottle after the depressurization process ends, combining it with the label of the depressurization gradient anomaly, the stress change amplitude, and the corresponding time segment, and updating the water injection state dataset to obtain depressurization adjustment reference data.

[0013] Furthermore, the method for adjusting the offset of the execution device includes: extracting the start and end timestamps of the multiple parallel cavities during the pressurization, pressure holding, and pressure release phases; selecting one cavity from the multiple parallel cavities as the time reference cavity, calculating the end time difference between the remaining cavities and the end timestamp of the time reference cavity in each phase, and constructing a multi-cavity synchronization time deviation sequence; obtaining the theoretical multi-cavity synchronization time range based on simulated operation data, comparing each value in the multi-cavity synchronization time deviation sequence with the theoretical multi-cavity synchronization time range, identifying the cavity number and corresponding phase label that exceed the theoretical multi-cavity synchronization time range; calculating the cumulative offset duration of each cavity in any operating cycle, and if the cumulative offset duration exceeds a preset cumulative threshold, performing time compensation on the start timestamp of the next operating cycle; and integrating the adjusted operating status information of all multiple parallel cavities with the pressure release adjustment reference data to obtain a synchronization control dataset.

[0014] Furthermore, the method of obtaining production line operation monitoring information includes: integrating the synchronous control dataset, bottle matching status information, and all tags into production line operation monitoring results, storing the production line operation monitoring results into a data structure, and obtaining production line operation monitoring information.

[0015] The technical effects and advantages of this invention's digital twin-based method for monitoring the operation of a fruit juice HPP sterilization production line are as follows: By real-time acquisition of bottle conveying status data and production line operation data during the fruit juice production line process, and performing data cleaning, a raw dataset for status monitoring is obtained. Based on this dataset, status analysis, liquid filling change trend analysis, pressurization stage fluctuation identification, bottle stress assessment, and multi-parallel cavity synchronous error monitoring are subsequently performed, realizing a digital twin-based method for monitoring the operation of a fruit juice HPP sterilization production line. Compared with existing technologies, by constructing a standard virtual production line operation scenario and extracting simulated operation data for comparison, the efficiency of operation monitoring is improved. Furthermore, by combining the bottle flow coverage time period with the expected bottle loading time... By comparing time intervals within the chamber, the flow rhythm of bottles in the buffer section of the chamber was visualized and monitored. By identifying abnormal liquid level sections, the accuracy of identifying the risk of residual gas retention during the pre-filling stage was improved. Fine modeling and empty bottle anomaly identification were performed on minute pressure wave changes during pressurization. By identifying risk sub-segments, the adaptive control capability for different bottle states during the depressurization stage was enhanced. By identifying deviations in multiple parallel cavities and performing time compensation, the probability of synchronization deviations in multiple parallel cavities was reduced. The digital perception capability and multi-stage collaborative control flexibility of the entire process of the fruit juice HPP sterilization production line operation monitoring were improved, while ensuring the accuracy of status identification and the timeliness of anomaly response in the production line operation monitoring. Attached Figure Description

[0016] Figure 1 is a schematic diagram of the operation monitoring method for the fruit juice HPP sterilization production line based on digital twins according to the present invention; Figure 2 is a schematic diagram of the operation monitoring system for the fruit juice HPP sterilization production line based on digital twins according to the present invention. Detailed Implementation

[0017] 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. Embodiment 1

[0018] Please refer to Figure 1. The method for monitoring the operation of a juice HPP sterilization production line based on digital twins described in this embodiment includes: S1. Real-time acquisition of bottle conveying status data and production line operation data during the juice production line process, and data cleaning to obtain a raw status monitoring dataset; S2. Status analysis of the conveying status based on the raw status monitoring dataset to obtain bottle status information, and rhythm recognition based on the bottle status information to obtain a bottle flow status dataset; S3. Trend analysis of the liquid filling process based on the bottle flow status dataset to identify abnormal filling results that interfere with the stable operation of the production line, and output a water injection status dataset; S4. Extract the pressurization stage state parameters from the water injection state dataset, and perform fluctuation identification based on these pressurization stage state parameters to obtain bottle matching state information; S5. Combine the bottle matching state information and the depressurization stage state parameters to perform bottle stress assessment and generate depressurization adjustment reference data; S6. Obtain the real-time operating status information of multiple parallel cavities, perform equipment offset adjustment in combination with the depressurization adjustment reference data, monitor the synchronization error of multiple parallel cavities, and obtain a synchronization control dataset; S7. Statistically analyze the production line operation monitoring results and obtain production line operation monitoring information through archiving and integration, and send the production line operation monitoring information to the preset production line monitoring terminal.

[0019] In this embodiment, the bottle conveying status data includes bottle attribute information such as the conveying time period range, bottle number, liquid level in the bottle, and bottle spacing, which reflects the bottle attributes of each bottle of juice during its entry into the HPP sterilization production line; the production line operation data includes attribute information such as conveyor belt speed, motor speed, and control cycle time, which reflects the operating status of relevant equipment in the production line; data cleaning is achieved by filling missing values ​​and deduplicating duplicate records, and the bottle conveying status data and production line operation data after data cleaning are integrated to obtain a more accurate and complete original dataset for status monitoring.

[0020] The methods for performing status analysis include: dividing the original status monitoring dataset into cycles to obtain a set of operating cycle parameters. The original status monitoring dataset is obtained by cleaning real-time collected data, and therefore includes data from multiple existing operating cycles. The duration of each operating cycle is set based on the operating specifications of the production line equipment. The original status monitoring dataset is then divided into cycles according to these operating cycles, and the operating cycle parameter set for each operating cycle is output.

[0021] Extract the bottle count parameter from the operating cycle parameter group and obtain the corresponding production line conveyor belt speed. Calculate the bottle quantity change rate for the operating cycle based on the bottle count parameter and the conveyor belt speed. The bottle count parameter refers to the number of bottles entering the designated detection section in each operating cycle. The bottle quantity change rate within a unit operating cycle is obtained by calculating the difference in bottle count parameters between adjacent operating cycles and the time length of the operating cycle. It should be noted that the calculation process is dimensionless.

[0022] A standard virtual production line operation scenario is constructed, and the simulated operation data of the virtual production line operation scenario is extracted. In this embodiment, the standard virtual production line operation scenario is constructed using a digital twin simulation model. The relevant plan content of the sterilization process of juice bottles entering the HPP sterilization production line is used as training data to construct a simulated operation scenario that reflects the original sterilization plan process. The extracted simulated operation data all conform to the ideal situation in the plan content and are aligned with the actual real-time data in the time dimension.

[0023] Calculate the difference in the rate of change of the number of bottles and the rate of change of the conveyor belt speed in adjacent consecutive operating cycles, and compare the rate of change of the conveyor belt speed with the simulated rate of change of speed in the simulated operating data to obtain the speed change deviation.

[0024] The difference in the rate of change of the number of bottles in adjacent consecutive operating cycles is used to reflect the acceleration or deceleration behavior of bottle conveying. The difference between the rate of change of the conveyor belt speed and the simulated rate of change of the speed in the ideal situation is calculated to obtain the difference between the actual change of the conveyor belt speed and the ideal speed change in each operating cycle, which is the speed change deviation.

[0025] The speed change deviation and the difference in the bottle quantity change rate are matched with the corresponding preset threshold range to output the conveying status trend label.

[0026] Based on the production line operation specifications, a preset threshold range related to the difference in the rate of change of bottle quantity and the deviation of speed change is set to meet the normal operation of the production line. The speed change deviation and the difference in the rate of change of bottle quantity are used as a set of associated judgment conditions. By matching the judgment conditions with the corresponding preset threshold range, the conveying status trend label that the speed change deviation and the difference in the rate of change of bottle quantity can match for the corresponding operating cycle is output, such as "speeding up", "slowing down" and "stable" labels that reflect the conveying status.

[0027] The delivery status trend label is bound to the corresponding running cycle number and the running cycle parameter group to output the bottle status information. Each running cycle parameter group is bound to the corresponding status trend label to obtain a data unit. All data units are integrated to obtain the bottle status information.

[0028] The execution rhythm recognition method includes: sorting the delivery status trend labels in the bottle status information according to time order, and extracting the trend type field corresponding to the running cycle. The trend type field refers to the field type information such as acceleration or deceleration in the delivery trend label.

[0029] The number of consecutive running cycles with the same statistical trend type field is used to filter the time intervals formed by running cycles whose number of consecutive running cycles is higher than the preset continuous trend threshold, and these are used as the continuous period of trend change.

[0030] The continuous trend threshold is set based on historical rhythm recognition experience to meet the continuity requirements. If there are multiple consecutive trend type fields with the same running cycle, the number of these consecutive running cycles is compared with the continuous trend threshold. If it is higher than the continuous trend threshold, the time segment formed by the corresponding consecutive running cycles is taken as the continuous period of trend change.

[0031] The number of bottles continuously pushed is calculated based on the length of the continuous period of trend change and the rate of change of the number of bottles in each corresponding operating cycle.

[0032] The formula for calculating the number of bottles continuously pushed is as follows: ;in This indicates the number of bottles continuously pushed out within a continuous period of trend change. Indicating the first continuous period of trend change The rate of change in the number of bottles per operating cycle; Indicating the first continuous period of trend change The duration of each running cycle; It indicates the number of cycles in a continuous period of trend change.

[0033] The bottle delivery time range is calculated based on the number of bottles continuously pushed, and the time segment corresponding to this bottle delivery time range is taken as the bottle flow coverage time period.

[0034] The formula for calculating the time range of bottle delivery is as follows: ;in Indicates the time range for bottle delivery; This represents the average operating speed of the conveyor belt over a continuous period corresponding to the trend change. Indicates the number of bottles continuously pushed; Indicates the width of each bottle; This indicates the spacing between consecutive bottles.

[0035] The bottle flow coverage time period is compared with the estimated bottle entry time in the simulated operation data to output the actual bottle push deviation time. The estimated bottle entry time refers to the time period when the bottle queue arrives at the sterilization chamber under the ideal rhythm in the simulated operation data. The actual bottle push deviation time is obtained by calculating the difference between the bottle flow coverage time period and the estimated bottle entry time, which is used to determine whether the bottle queue docks with the chamber on time in the time dimension.

[0036] The actual bottle pushing deviation time is compared with the beat offset threshold to determine the bottle flow rhythm state during the corresponding continuous time period of trend change, and the bottle flow state label is output.

[0037] The beat offset threshold is a set of thresholds set based on the sterilization production line operation specifications to reflect rhythm differences. It is divided into multiple threshold ranges. The actual bottle flow push deviation time is compared with the beat offset threshold to determine the bottle flow rhythm status of the continuous time period corresponding to the actual bottle flow push deviation time, such as lagging or leading. At the same time, the field corresponding to the bottle flow rhythm status is output as the bottle flow status label.

[0038] Use bottle flow status labels to mark the corresponding running cycles and output the bottle flow status dataset.

[0039] The methods for performing trend analysis include: identifying the pre-filling water start command field, mapping the pre-filling water start command field to the timestamp corresponding to the bottle flow status dataset, extracting the initial liquid level value at that timestamp and the corresponding operating cycle number, and recording them as the pre-filling initial state parameters.

[0040] The pre-fill water start command field refers to the automatic water injection start command sent by the preset production line control terminal to the production line. It is the start of the liquid filling process, maps the timestamp to the bottle flow status data set, and locates the initial liquid level value at which the liquid filling process begins at this timestamp. The initial liquid level value and the corresponding running cycle number are combined as the pre-fill initial state parameters.

[0041] Obtain the liquid filling end timestamp, and take the time interval between the timestamp corresponding to the pre-fill water start command field and the liquid filling end timestamp as the liquid filling process segment.

[0042] The liquid filling end timestamp refers to the timestamp at which the liquid filling process ends. It is the timestamp at which the preset production line control terminal executes the closing command on the water injection valve. The time difference between the timestamp corresponding to the pre-filling start command field and the liquid filling end timestamp is the actual water injection duration, which is used as the liquid filling process segment.

[0043] Identify the liquid level height in each operating cycle of the liquid filling process segment, and construct a time series of liquid level height changes, wherein the liquid level height of each operating cycle is sorted in chronological order to obtain the time series of liquid level height changes.

[0044] The liquid level change per unit time is calculated in the liquid level height change time series, and the liquid level change rate is calculated based on the liquid level change per unit time. The liquid level change curve is plotted based on the liquid level change rate. The liquid level change per unit time is obtained by calculating the liquid level height difference between adjacent operating cycles in the liquid level height change time series. At the same time, the liquid level change rate of adjacent operating cycles is calculated based on this value. The liquid level change rate is used as the slope to plot the liquid level change curve.

[0045] The theoretical liquid level change curve is plotted based on the simulated operation data, and then compared with the liquid level change curve to output a liquid level deviation value sequence. Specifically, the theoretical liquid level change curve is plotted based on the simulated operation data for the same time period as the liquid level change curve. The difference between the values ​​of each point in the theoretical liquid level change curve and the liquid level change curve is compared, and the differences are sorted to obtain the liquid level deviation value sequence.

[0046] The time interval in the liquid level deviation value sequence where the continuous liquid level deviation is higher than the preset deviation threshold is identified as the liquid level abnormal segment. The preset deviation threshold is set based on historical trend analysis experience. This value is used as a threshold to determine whether the liquid level deviation exceeds the acceptable range of normal conditions. If there are consecutive cases in the liquid level deviation value sequence that are higher than the preset deviation threshold, it indicates that the liquid level change in the corresponding time interval deviates from the theoretical situation, and therefore it is determined to be a liquid level abnormal segment.

[0047] The system calculates the duration of abnormal liquid level sections and determines the abnormality type based on this duration, outputting the abnormality type field. Specifically, the abnormality type determination is performed by using the duration of the abnormal liquid level section as one of the criteria, and the corresponding abnormality type field is output.

[0048] Calculate the percentage of abnormal liquid level sections in the liquid filling process. If the percentage of abnormal liquid level sections is higher than the preset safety ratio threshold, it is determined that there is a risk of residual gas retention in the corresponding liquid filling process section.

[0049] The preset safety ratio threshold is a threshold set based on the production line operation specifications. If the duration of the abnormal liquid level section accounts for a higher proportion of the total time of the liquid filling process section than the preset safety ratio threshold, it is considered that there is residual gas retention in the corresponding liquid filling process section, which affects the normal water injection operation.

[0050] The corresponding time periods are labeled using the liquid level anomaly type field, the anomaly percentage value, and the label of residual gas retention risk, and the water injection status dataset is output.

[0051] The methods for determining the type of anomaly include: setting an anomaly duration range, comparing the duration length with the anomaly duration range, and classifying them into short-term anomalies, medium-term anomalies, and long-term anomalies. The anomaly duration range is a threshold interval set based on experience summarized from a large number of historical anomaly judgment records. The duration length is compared with this threshold interval to determine which type of time anomaly the corresponding liquid level anomaly segment belongs to, which is divided into short-term anomalies, medium-term anomalies, and long-term anomalies.

[0052] Extract the liquid level deviation value sequence of the abnormal liquid level section and calculate the average deviation amplitude. At the same time, count the number of positive and negative alternations of the deviation values. The average deviation amplitude is used to represent the intensity of the abnormality. The larger the value, the more serious the liquid level loss of control. The number of positive and negative alternations refers to the number of times the deviation direction changes between consecutive deviation values ​​in the liquid level deviation value sequence. The larger the value, the more frequent the liquid level control switching occurs, reflecting the unstable state.

[0053] If it is a short-term anomaly and the average deviation exceeds the preset short-term deviation threshold, it is determined to be either a fast instantaneous rate or a slow instantaneous rate based on the positive or negative direction of the deviation value. If the duration of the anomaly meets the criteria for a short-term anomaly and the average deviation is greater than the short-term deviation threshold set based on historical experience, it is determined to be a fast instantaneous rate if the change in the corresponding deviation value is in the positive direction and a slow instantaneous rate if the change is in the negative direction.

[0054] If the time interval is abnormal and the number of positive and negative alternations exceeds the preset frequency threshold, it is judged as a fluctuation abnormality. If the duration of the abnormality is within the range of the time interval abnormality and the number of positive and negative alternations is higher than the preset frequency threshold set based on historical experience, it means that water injection has basically started but there are fluctuations and the overall situation is not stable. This state is judged as a fluctuation abnormality.

[0055] Conversely, the direction of the deviation value determines whether the continuous rate is too fast or too slow. If the number of alternating positive and negative values ​​does not exceed the preset frequency threshold, it indicates that the water injection rate is abnormal. Therefore, the direction of the liquid level change determines whether the continuous rate is too fast or too slow.

[0056] If it is a long-term anomaly, it is classified as a long-term anomaly. Since long-term anomalies cannot be ignored, they are directly classified as long-term anomalies, which are classified as severe anomalies.

[0057] All anomaly type labels are bound to the corresponding time segment and average deviation range, and the liquid level anomaly type field is output. Each anomaly type label is encoded and bound to the corresponding time segment and average deviation range. The liquid level anomaly type field is obtained through field encoding.

[0058] The state parameters during the pressurization phase include: the instantaneous pressure sequence corresponding to the pressurization process and the theoretical pressure value data generated based on the simulated operation data. The instantaneous pressure sequence is the pressure reading sequence collected in real time by the pressure sensor during the pressurization process, while the theoretical pressure value data is the pressure value related data representing the theoretical situation in the scenario simulated by the digital twin model.

[0059] The method for fluctuation identification includes: dividing the instantaneous pressure sequence corresponding to the pressurization process into sub-segments and outputting instantaneous pressure sub-segments. The entire time interval of the pressurization process is divided into several sub-segments. Those skilled in the art can adjust the degree of division based on historical experience and specific working conditions to ensure that the time interval corresponding to the entire pressurization process can be completely divided, and the local instantaneous pressure value corresponding to each sub-segment is taken as the instantaneous pressure sub-segment.

[0060] Calculate the pressure standard deviation of the instantaneous pressure values ​​within the instantaneous pressure sub-segment, and define the instantaneous pressure sub-segments with pressure standard deviations higher than the stability threshold as pressure unstable sub-segments.

[0061] The standard deviation of all instantaneous pressure values ​​within each instantaneous pressure segment is calculated to obtain the pressure standard deviation. A stability threshold is set based on the production line operation specifications. If the pressure standard deviation is higher than the stability threshold, it indicates that the dispersion of pressure changes in the corresponding instantaneous pressure segment is too high. Therefore, the segment is judged as an unstable pressure segment.

[0062] Identify local extreme points in the instantaneous pressure sequence and count the distribution density of these local extreme points. If the distribution density of these local extreme points is higher than a preset density threshold, the corresponding pressurization process is determined to have pressure oscillation.

[0063] Local extreme points refer to the points in the instantaneous pressure sequence where the pressure value of the middle point among three consecutive points is higher than that of the two adjacent points. The distribution density of local extreme points is obtained by counting the number of local extreme points in a unit instantaneous pressure segment in the entire instantaneous pressure sequence. A preset density threshold is set based on historical fluctuation identification experience. If the distribution density of local extreme points is higher than the preset density threshold, it indicates that a high-frequency oscillation phenomenon occurs during the pressurization process.

[0064] The average pressure rise rate of the unstable pressure sub-segment is obtained by calculating the ratio of the pressure rise amplitude to the corresponding time length. The pressure rise amplitude is the difference between the minimum and maximum values ​​of the unstable pressure sub-segment. The average pressure rise rate of the sub-segment is calculated by comparing this value with the time length of the unstable pressure sub-segment.

[0065] Calculate the theoretical pressure rise rate in the same time period as the pressure instability segment in the theoretical pressure value data, and compare the average pressure rise rate with the theoretical pressure rise rate to calculate the pressure wave velocity attenuation ratio. The theoretical pressure rise rate in the theoretical pressure value data is calculated simultaneously, and the ratio of the average pressure rise rate to the theoretical pressure rise rate in the pressure instability segment is calculated to obtain the pressure wave velocity attenuation ratio. The larger this value is, the lower the actual pressure value is than the theoretical expectation.

[0066] Identify the actual timestamp of reaching the preset target pressure value in the instantaneous pressure sequence, and simultaneously identify the theoretical timestamp of reaching the preset target pressure value in the theoretical pressure value data. Calculate the time difference between the actual timestamp and the theoretical timestamp as the pressure propagation deviation time.

[0067] The preset target pressure value is a value set based on the original pressurization plan. The pressure propagation deviation time is obtained by comparing the time difference between the actual timestamp and the theoretical timestamp of reaching the preset target pressure value in the instantaneous pressure sequence. This value is the time delay in which the actual response is slower than the theoretical rhythm.

[0068] If pressure oscillation is simultaneously met, the pressure wave velocity attenuation ratio exceeds the preset attenuation threshold, and the pressure propagation deviation time is higher than the propagation deviation threshold, then the corresponding pressurization process is marked as an abnormality of insufficient liquid level. The preset attenuation threshold and propagation deviation threshold are set based on the production line equipment operation specifications. If pressure oscillation occurs during the pressurization process, the difference between the pressure wave velocity attenuation ratio and the attenuation threshold is large, and the pressure propagation deviation time is significantly delayed, it indicates that the liquid level in the bottle is insufficient or the bottle is completely empty, causing the pressure wave to be unable to propagate in a steady state and instead being absorbed and delayed.

[0069] The label indicating insufficient liquid level is bound to the corresponding pressurization stage status parameter, and the bottle matching status information is output.

[0070] The methods for assessing bottle stress include: extracting the instantaneous pressure drop sequence from the state parameters during the depressurization phase, calculating the pressure drop rate of adjacent drop segments, and constructing a pressure drop rate sequence.

[0071] The pressure relief phase state parameters refer to the relevant parameters corresponding to the pressure relief phase. Similar to the pressure increase phase state parameters, the corresponding pressure value sequence is divided into segments. The segmentation scale is based on experience from processing a large amount of historical pressure relief phase data. The instantaneous pressure drop sequence is divided into several drop segments. The pressure drop rate sequence is obtained by calculating and sorting the pressure drop rates between adjacent drop segments.

[0072] Identify the labels in the bottle matching status information. If the label indicates insufficient liquid level, set a high-sensitivity bottle stress coefficient. If the label indicates residual gas retention risk, set a medium-sensitivity bottle stress coefficient.

[0073] The system differentiates and sets the stress coefficient for sensitive bottles by identifying labels. A high-sensitivity stress coefficient is set for labels indicating insufficient liquid level. This sensitive bottle stress coefficient is used to adjust the weighting of information related to the corresponding bottle during stress assessment. When the label indicates insufficient liquid level, it means the liquid level in the bottle is severely insufficient. Due to the weakened energy absorption of the liquid, the bottle is prone to structural deformation during rapid depressurization; therefore, a high-sensitivity stress coefficient is set. When the label indicates residual gas risk, it means there is some air entrainment in the bottle, not extreme, but it may lead to phenomena such as air sealing; therefore, a medium-sensitivity stress coefficient is set. The specific values ​​of both coefficients are based on historical experience.

[0074] The stress change amplitude during the pressure relief process is calculated based on the pressure drop rate sequence and the corresponding bottle stress coefficient. This involves multiplying the drop rate value at any point in the pressure drop rate sequence with the corresponding bottle stress coefficient, and then multiplying this product with the time length of the corresponding drop segment to obtain the stress change amplitude during the pressure relief process of that drop segment.

[0075] Based on simulated operation data, a theoretical pressure relief rate curve for the corresponding pressure relief process is plotted. The difference between the theoretical pressure relief rate curve and the pressure drop rate sequence is compared, and the drop segment where the difference is higher than the pressure relief safety threshold is identified as a risk segment. Specifically, a point-to-point difference comparison is performed between the theoretical pressure relief rate curve and the pressure drop rate sequence. If there are drop segments where the difference is higher than the pressure relief safety threshold set based on the production line pressure relief operation specifications, then the segment is identified as a risk segment, indicating that the pressure drop rate in that segment may pose a risk.

[0076] If a descending segment is identified as a risky segment and the stress change during the pressure relief process exceeds a preset stress threshold, then the descending segment is marked as having an abnormal pressure relief gradient. The preset stress threshold is set based on the production line operation specifications. If a descending segment has been identified as a risky segment and the stress change during the pressure relief process exceeds the preset stress threshold, it indicates that it is not only a rate problem, but may be caused by excessive pressure accumulation leading to an abnormal pressure relief. Therefore, this segment is marked as having an abnormal pressure relief gradient.

[0077] The predicted residual pressure inside the bottle after the depressurization process is calculated, and combined with the labels of the depressurization gradient anomalies, the stress change amplitude, and the corresponding time segments, the water injection state dataset is updated to obtain depressurization adjustment reference data.

[0078] The process involves calculating the pressure value corresponding to the end timestamp of the depressurization process, then predicting the residual pressure based on relevant pressure prediction theories and the gas-liquid state inside the bottle, thus obtaining the predicted residual pressure inside the bottle. The predicted residual pressure inside the bottle, along with the depressurization gradient anomaly label and stress change amplitude, are combined to form a data unit, which is then bound to the corresponding time segment and added to the depressurization stage state parameters of the corresponding time segment in the water injection state dataset to obtain depressurization adjustment reference data.

[0079] The methods for performing equipment offset adjustment include: extracting the start and end timestamps of multiple parallel chambers during the pressurization, pressure holding, and pressure release phases. The multiple parallel chambers refer to multiple sterilization chambers operating in parallel in the HPP sterilization production line. The start and end timestamps of the pressurization, pressure holding, and pressure release phases are recorded as the data basis for subsequent operations.

[0080] Select one cavity from the multiple parallel cavities as the time reference cavity, calculate the end time difference between the other cavities and the end timestamp of the time reference cavity in each stage, and construct a multi-cavity synchronization time deviation sequence.

[0081] Based on the production line operation specifications, a stable cavity is selected from multiple parallel cavities as the time base cavity. At the same time, the time difference between the end timestamp of each stage of all other cavities and the time base cavity is calculated, and the time deviation sequence between each cavity and the time base cavity at different stages is constructed, which is the multi-cavity synchronization time deviation sequence.

[0082] Based on simulated operation data, the theoretical multi-cavity synchronization time range is obtained. Each value in the multi-cavity synchronization time deviation sequence is compared with the theoretical multi-cavity synchronization time range to identify the cavity number and corresponding stage label that exceed the theoretical multi-cavity synchronization time range.

[0083] The theoretical multi-cavity synchronization time range refers to the synchronization deviation threshold between multiple cavities allowed in each stage under the theoretical state set by the production line equipment design specifications. The deviation value of each timestamp in the multi-cavity synchronization time deviation sequence of each cavity is compared with the theoretical multi-cavity synchronization time range. If the deviation value is higher than the theoretical multi-cavity synchronization time range, the corresponding cavity number and the corresponding stage label that exceeds the range are recorded.

[0084] The cumulative offset duration of each cavity in any operating cycle is calculated. If the cumulative offset duration exceeds the preset cumulative threshold, time compensation is performed on the start timestamp of the next operating cycle.

[0085] The system sets a preset cumulative threshold based on the production line operation specifications. It calculates the duration of the cumulative offset generated by each cavity in any operating cycle and compares it with the corresponding threshold. If the duration is higher than the preset cumulative threshold, it means that the corresponding cavity has a timing misalignment in the operating cycle. At this time, the time is shifted by adjusting the start timestamp of each stage of the next operating cycle, for example, delaying the entry into the stage by several seconds, to achieve time compensation.

[0086] The adjusted operating status information of all parallel cavities is integrated with the pressure relief adjustment reference data to obtain a synchronous control dataset.

[0087] The methods for obtaining production line operation monitoring information include: integrating the synchronous control dataset, bottle matching status information, and all tags into the production line operation monitoring result, storing the production line operation monitoring result into a data structure, and obtaining the production line operation monitoring information. In this process, the synchronous control dataset obtained through multiple steps, the bottle matching status information obtained by identifying fluctuations during the pressurization stage, and various tags obtained in all steps are archived and integrated, and a unified format data structure is output. This data structure is the production line operation monitoring information.

[0088] This embodiment obtains a raw dataset for status monitoring by real-time acquisition of bottle conveying status data and production line operation data during the juice production line process and performing data cleaning. Based on this dataset, it performs status analysis, liquid filling change trend analysis, pressurization stage fluctuation identification, bottle stress assessment, and multi-parallel chamber synchronization error monitoring, realizing a digital twin-based method for monitoring the operation of a juice HPP sterilization production line. Compared with existing technologies, it improves operation monitoring efficiency by constructing a standard virtual production line operation scenario and extracting simulated operation data for comparison. Furthermore, by comparing the bottle flow coverage time period with the expected bottle entry time, it achieves... Visual monitoring of the bottle flow rhythm in the buffer section of the chamber; improved identification accuracy of residual gas retention risk during the pre-filling stage by identifying abnormal liquid level sections; refined modeling and empty bottle anomaly identification of minute pressure wave changes during pressurization; enhanced adaptive control capability for different bottle states during the depressurization stage by identifying risk sub-segments; reduced the probability of synchronization deviations in multiple parallel cavities by identifying deviations in multiple parallel cavities and performing time compensation; improved the full-process digital perception capability and multi-stage collaborative control flexibility of the fruit juice HPP sterilization production line operation monitoring, while ensuring the accuracy of status identification and the timeliness of anomaly response in the production line operation monitoring. Example 2

[0089] Please refer to Figure 2. For details not described in this embodiment, please refer to the description in Embodiment 1. A digital twin-based monitoring system for the operation of a fruit juice HPP sterilization production line is provided, including: a data acquisition module, which collects real-time bottle conveying status data and production line operation data during the fruit juice production line process, and performs data cleaning to obtain a raw status monitoring dataset; a bottle analysis module, which performs status analysis on the conveying status based on the raw status monitoring dataset to obtain bottle status information, and performs rhythm recognition based on this bottle status information to obtain a bottle flow status dataset; a trend recognition module, which performs trend analysis on the liquid filling process based on the bottle flow status dataset, identifies abnormal filling results that interfere with the stable operation of the production line, and outputs a water injection status dataset; and a pressurization analysis module. The system comprises several modules: a pressure relief adjustment module, a multi-cavity adjustment module, and a multi-cavity control module. The pressure relief adjustment module extracts the pressurization stage state parameters from the water injection state dataset, identifies fluctuations based on these parameters, and obtains the bottle matching state information. The pressure relief adjustment module combines the bottle matching state information and the pressure relief stage state parameters to assess bottle stress and generate pressure relief adjustment reference data. The multi-cavity adjustment module acquires real-time operating status information for multiple parallel cavities, performs equipment offset adjustments based on the pressure relief adjustment reference data, monitors the synchronization error of the multiple parallel cavities, and obtains a synchronization control dataset. The archiving and integration module statistically analyzes the production line operation monitoring results and obtains production line operation monitoring information through archiving and integration, then sends this information to a preset production line monitoring terminal. All modules are connected via wired and / or wireless means.

[0090] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

[0091] All formulas in this manual are dimensionless and calculated numerically. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters and thresholds in the formulas are set by those skilled in the art according to the actual situation.

[0092] Although embodiments of the invention have been shown and described, those skilled in the art will understand that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the claims and their equivalents.

Claims

1. A method for monitoring the operation of a fruit juice HPP sterilization production line based on digital twins, characterized in that, include: S1. Real-time acquisition of bottle conveying status data and production line operation data during the juice production line process, and data cleaning to obtain the raw dataset for status monitoring; S2. Based on the original data set of status monitoring, perform status analysis on the conveying status to obtain bottle status information, and perform rhythm recognition based on the bottle status information to obtain bottle flow status dataset; S3. Based on the bottle flow status dataset, perform trend analysis on the liquid filling process, identify abnormal filling results that interfere with the stable operation of the production line, and output water injection status dataset; S4. Extract the pressurization stage state parameters from the water injection state dataset, and perform fluctuation identification based on the pressurization stage state parameters to obtain the bottle matching state information; S5. Combine bottle body matching status information and pressure relief stage status parameters to evaluate bottle body stress and generate pressure relief adjustment reference data; S6. Real-time acquisition of the operating status information of multiple parallel cavities, combined with pressure relief adjustment reference data to perform equipment offset adjustment, monitoring the synchronization error of multiple parallel cavities, and obtaining a synchronization control dataset; S7. Statistical analysis of production line operation monitoring results and archiving and integration to obtain production line operation monitoring information, and sending the production line operation monitoring information to the preset production line monitoring terminal.

2. The method for monitoring the operation of a fruit juice HPP sterilization production line based on digital twins according to claim 1, characterized in that, The method for performing state analysis includes: dividing the original state monitoring dataset into cycles to obtain a set of operating cycle parameters; extracting the bottle count parameters from the set of operating cycle parameters and obtaining the corresponding conveyor belt speed of the production line; calculating the rate of change of the number of bottles in the operating cycle based on the bottle count parameters and the conveyor belt speed; constructing a standard virtual production line operating scenario and extracting the simulated operating data of the virtual production line operating scenario; calculating the difference in the rate of change of the number of bottles and the rate of change of the conveyor belt speed in adjacent consecutive operating cycles, and comparing the rate of change of the conveyor belt speed with the simulated rate of change of speed in the simulated operating data to obtain the speed change deviation; matching the speed change deviation and the difference in the rate of change of the number of bottles with the corresponding preset threshold interval and outputting a conveying state trend label; binding the conveying state trend label with the corresponding operating cycle number and the set of operating cycle parameters and outputting the bottle state information.

3. The method for monitoring the operation of a fruit juice HPP sterilization production line based on digital twins according to claim 2, characterized in that, The execution rhythm recognition method includes: sorting the delivery status trend tags in the bottle status information according to time order, and extracting the trend type field corresponding to the running cycle; counting the number of consecutive running cycles with the same trend type field, and filtering the time segment consisting of running cycles with a number of consecutive running cycles higher than a preset continuous trend threshold as the continuous trend change period; calculating the number of consecutively pushed bottles based on the length of the continuous trend change period and the rate of change of the number of bottles corresponding to each running cycle; calculating the bottle delivery time range based on the number of consecutively pushed bottles, and taking the time segment corresponding to the bottle delivery time range as the bottle flow coverage time period; comparing the bottle flow coverage time period with the expected bottle entry time in the simulated running data, and outputting the actual bottle push deviation time; comparing the actual bottle push deviation time with the beat offset threshold, determining the bottle flow rhythm state of the corresponding continuous trend change period, and outputting the bottle flow status tag; using the bottle flow status tag to mark the corresponding running cycle, and outputting the bottle flow status dataset.

4. The method for monitoring the operation of a fruit juice HPP sterilization production line based on digital twins according to claim 3, characterized in that, The method for performing trend analysis includes: identifying the pre-filling water start command field and mapping it to the timestamp corresponding to the bottle flow state dataset; extracting the initial liquid level value at that timestamp and the corresponding operating cycle number, which are recorded as the pre-filling initial state parameters; obtaining the liquid filling end timestamp and using the time segment between the timestamp corresponding to the pre-filling water start command field and the liquid filling end timestamp as the liquid filling process segment; identifying the liquid level height for each operating cycle in the liquid filling process segment and constructing a liquid level height change time series; calculating the unit time liquid level change in the liquid level height change time series and calculating the liquid level change rate based on the unit time liquid level change, and plotting the liquid level change curve based on the liquid level change rate; based on... Simulated operation data is used to plot theoretical liquid level change curves, and these curves are compared with the actual liquid level change curves to output a liquid level deviation value sequence. Time intervals within the liquid level deviation value sequence where consecutive liquid level deviations exceed a preset deviation threshold are identified and recorded as abnormal liquid level intervals. The duration of abnormal liquid level intervals is calculated, and anomaly type is determined based on this duration, outputting an abnormal liquid level type field. The abnormal proportion of abnormal liquid level intervals within the liquid filling process is calculated; if this proportion exceeds a preset safety proportion threshold, the corresponding liquid filling process segment is determined to have a risk of residual gas retention. The corresponding time intervals are labeled using the abnormal liquid level type field, the abnormal proportion value, and the label of residual gas retention risk, outputting a water injection status dataset.

5. The method for monitoring the operation of a fruit juice HPP sterilization production line based on digital twins according to claim 4, characterized in that, The method for determining the anomaly type includes: setting an anomaly duration range, comparing the duration length with the anomaly duration range, and classifying it into short-term anomalies, medium-term anomalies, and long-term anomalies; extracting the liquid level deviation value sequence of the anomaly segment and calculating the average deviation amplitude, while counting the number of positive and negative alternations of the deviation values; if it is a short-term anomaly and the average deviation amplitude exceeds a preset short-term deviation threshold, it is determined to be either a fast instantaneous rate or a slow instantaneous rate based on the positive or negative direction of the deviation value; if it is a medium-term anomaly and the number of positive and negative alternations exceeds a preset frequency threshold, it is determined to be a fluctuation anomaly; otherwise, it is determined to be either a sustained rate that is too fast or a sustained rate that is too slow based on the positive or negative direction of the deviation value; if it is a long-term anomaly, it is determined to be a long-term anomaly; binding all anomaly type tags with the corresponding time segment and average deviation amplitude, and outputting the liquid level anomaly type field.

6. The method for monitoring the operation of a fruit juice HPP sterilization production line based on digital twins according to claim 5, characterized in that, The state parameters of the pressurization stage include: the instantaneous pressure sequence corresponding to the pressurization process and the theoretical pressure value data generated based on the simulation operation data.

7. The method for monitoring the operation of a fruit juice HPP sterilization production line based on digital twins according to claim 6, characterized in that, The method for fluctuation identification includes: dividing the instantaneous pressure sequence corresponding to the pressurization process into segments and outputting instantaneous pressure segments; calculating the pressure standard deviation of the instantaneous pressure values ​​within each instantaneous pressure segment, and identifying instantaneous pressure segments with pressure standard deviations higher than a stability threshold as unstable pressure segments; identifying local extreme points in the instantaneous pressure sequence and calculating the distribution density of these local extreme points; if the distribution density of these local extreme points is higher than a preset density threshold, then the corresponding pressurization process is determined to have pressure oscillations; calculating the ratio of the pressure rise amplitude to the corresponding time length of the unstable pressure segment to obtain the average pressure rise rate of the unstable pressure segment; and calculating the theoretical pressure value data corresponding to the unstable pressure segment. Simultaneously, the theoretical pressure rise rate is calculated, and the average pressure rise rate is compared with the theoretical pressure rise rate to calculate the pressure wave velocity attenuation ratio. The actual timestamp of reaching the preset target pressure value in the instantaneous pressure sequence is identified, and the theoretical timestamp of reaching the preset target pressure value in the theoretical pressure value data is identified simultaneously. The time difference between the actual timestamp and the theoretical timestamp is calculated as the pressure propagation deviation time. If pressure oscillation is satisfied, the pressure wave velocity attenuation ratio exceeds the preset attenuation threshold, and the pressure propagation deviation time is higher than the propagation deviation threshold, then the corresponding pressurization process is marked as a liquid level deficiency anomaly. The liquid level deficiency anomaly label is bound to the corresponding pressurization stage state parameters, and the bottle matching state information is output.

8. The method for monitoring the operation of a fruit juice HPP sterilization production line based on digital twins according to claim 7, characterized in that, The method for assessing bottle stress includes: extracting the instantaneous pressure drop sequence from the state parameters during the depressurization phase, calculating the pressure drop rate of adjacent drop segments, and constructing a pressure drop rate sequence; identifying labels in the bottle matching state information; if the label indicates insufficient liquid level anomaly, setting a high-sensitivity bottle stress coefficient; if the label indicates residual gas retention risk, setting a sensitive bottle stress coefficient; calculating the stress change amplitude during the depressurization process based on the pressure drop rate sequence and the corresponding bottle stress coefficient; plotting the theoretical depressurization rate curve for the corresponding depressurization process based on simulated operation data, comparing the difference between the theoretical depressurization rate curve and the pressure drop rate sequence, and identifying drop segments where the difference exceeds the depressurization safety threshold as risk segments; if a drop segment is a risk segment and the stress change amplitude during the depressurization process exceeds a preset stress threshold, marking the drop segment as a depressurization gradient anomaly; calculating the predicted residual pressure inside the bottle after the depressurization process ends, combining it with the label of the depressurization gradient anomaly, the stress change amplitude, and the corresponding time segment, and updating the water injection state dataset to obtain depressurization adjustment reference data.

9. The method for monitoring the operation of a fruit juice HPP sterilization production line based on digital twins according to claim 8, characterized in that, The method for adjusting the offset of the execution device includes: extracting the start and end timestamps of the pressurization, pressure holding, and pressure relief phases of the multiple parallel cavities; selecting one cavity from the multiple parallel cavities as the time reference cavity, calculating the end time difference between the other cavities and the end timestamp of the time reference cavity in each phase, and constructing a multi-cavity synchronization time deviation sequence; obtaining the theoretical multi-cavity synchronization time range based on simulated operation data, comparing each value in the multi-cavity synchronization time deviation sequence with the theoretical multi-cavity synchronization time range, identifying the cavity number and corresponding phase label that exceed the theoretical multi-cavity synchronization time range; calculating the cumulative offset duration of each cavity in any operating cycle, and if the cumulative offset duration exceeds a preset cumulative threshold, performing time compensation on the start timestamp of the next operating cycle; and integrating the adjusted operating status information of all multiple parallel cavities with the pressure relief adjustment reference data to obtain a synchronization control dataset.

10. The method for monitoring the operation of a fruit juice HPP sterilization production line based on digital twins according to claim 9, characterized in that, The method of obtaining production line operation monitoring information includes: integrating the synchronous control dataset, bottle matching status information, and all tags into production line operation monitoring results, storing the production line operation monitoring results into a data structure, and obtaining production line operation monitoring information.

Images