Digital Twin-Driven Collaborative Control Method and System for LCD Backlight Production

Abstract

This invention discloses a digital twin-driven collaborative management and control method and system for LCD backlight panel production, specifically relating to the field of production management technology. It addresses the resource misallocation, quality risks, and efficiency losses caused by existing static planning methods that neglect the dynamic coupling relationship between the capacity differences of production lines at specific processes and the process paths of work-in-process. The method involves constructing a digital twin model corresponding to the physical production line, integrating process capability vectors and process path vectors, identifying candidate production line sets based on vector data, assessing parameter resonance risk levels, locating key bottleneck processes in the process network topology, generating dynamic routing schemes to avoid high-risk paths and bottlenecks, and triggering real-time adjustments to the work-in-process production path through the manufacturing execution system to achieve dynamic collaborative optimization of production resources and order requirements.

Description

Technical Field

[0001] This invention relates to the field of production management technology, and in particular to a method and system for collaborative control of LCD backlight production driven by digital twins. Background Technology

[0002] In the LCD backlight panel manufacturing industry, to achieve large-scale production and order delivery goals, companies generally deploy multiple production lines and use manufacturing execution systems for production management. In existing technologies, production scheduling and task allocation mainly rely on static planning methods based on order delivery dates and overall equipment utilization. This method decomposes production orders into process sequences and allocates suitable production lines to each process based on preset, relatively fixed production line capacity attributes. This management model can maintain basic production order in production environments with stable orders and minimal resource conflicts.

[0003] However, due to the failure to fully consider the differences in capabilities of different production lines in each specific process and the dynamic coupling relationship between the complete process path of work-in-process, the initial static task allocation may lock the work-in-process on the production route that has a capability gap for its subsequent key processes. This structural contradiction between production path and resource capability leads to potential quality risks and efficiency losses. Moreover, this defect is preset at the beginning of task allocation and is difficult to correct through subsequent local adjustments. Summary of the Invention

[0004] This invention addresses the technical problems existing in the prior art by providing a digital twin-driven collaborative control method and system for LCD backlight production.

[0005] The technical solution of the present invention to solve the above-mentioned technical problems is as follows:

[0006] A digital twin-driven collaborative management method for LCD backlight production includes:

[0007] S1. Construct a digital twin model corresponding to the physical production line. The digital twin model includes the process capability vectors of each production line at different processes and the process path vectors of each work-in-process.

[0008] S2. Based on the process capability vector and process path vector, identify the set of candidate production lines corresponding to each process in the current complete process path of work-in-process;

[0009] S3. Identify the sensitive process parameters in the process path vector of the current work-in-process and compare them with the inherent fluctuation characteristics corresponding to the process capability vector of each production line in the candidate production line set to assess the risk level of parameter resonance.

[0010] S4. Based on the parameter resonance risk level, identify the intersection nodes of production paths with high parameter resonance risk levels in the process network topology, and determine the intersection nodes as key bottleneck processes that trigger systemic chain reactions.

[0011] S5. Generate a dynamic routing scheme based on the parameter resonance risk level and key bottleneck process. The dynamic routing scheme avoids production paths with high parameter resonance risk level and prioritizes bypassing key bottleneck processes.

[0012] S6. Send the dynamic routing scheme to the manufacturing execution system to trigger the adjustment of the work-in-process production path.

[0013] Furthermore, a digital twin model corresponding to the physical production line is constructed. This digital twin model includes process capability vectors for each production line at different stages and process path vectors for each work-in-process, including:

[0014] The process capability vector is generated by collecting historical production data from different processes on each production line and extracting key performance indicators, including process accuracy and process cycle time.

[0015] The process path vector is generated by parsing the production process documents of each work-in-process and mapping the process sequence. The process sequence includes the process order and process parameter requirements.

[0016] The digital twin model integrates process capability vectors and process path vectors to form a virtual mapping that is synchronized with the physical production line in real time.

[0017] Furthermore, based on the process capability vector and the process path vector, the candidate production line set corresponding to each process in the current work-in-process complete process path is identified, including:

[0018] Based on the process capability vector and the process path vector, for each process in the process path vector, the capability parameters of each production line in the corresponding process are extracted from the process capability vector.

[0019] Then, it is matched with the parameter requirements of the corresponding process in the process path vector, and the successfully matched production line is included in the candidate production line set.

[0020] Furthermore, sensitive process parameters in the current work-in-process process path vector are identified and compared with the inherent fluctuation characteristics corresponding to the process capability vectors of each production line in the candidate production line set to assess the risk level of parameter resonance, including:

[0021] Identify the process parameters in the current work-in-process process path vector that have a decisive impact on the quality of the final product as sensitive process parameters;

[0022] The long-term statistical distribution characteristics of the corresponding process parameters are extracted from the process capability vectors of each production line in the candidate production line set as the inherent fluctuation characteristics.

[0023] Analyze the degree of overlap between the allowable deviation range of sensitive process parameters and the actual distribution range characterized by inherent fluctuation characteristics;

[0024] The risk level of parameter resonance is determined based on the comparison between the degree of overlap and the preset threshold.

[0025] When the degree of overlap is lower than the first threshold, the parameter resonance risk level is high risk; when the degree of overlap is between the first threshold and the second threshold, it is medium risk; and when the degree of overlap is higher than the second threshold, it is low risk.

[0026] Furthermore, extracting the long-term statistical distribution characteristics of the corresponding process parameters from the process capability vectors of each production line in the candidate production line set as inherent fluctuation characteristics includes: calculating the mean, standard deviation, and distribution range of each production line on the corresponding process parameters based on the historical production data recorded in the process capability vectors.

[0027] Furthermore, analyzing the degree of overlap between the allowable deviation range of sensitive process parameters and the actual distribution range characterized by inherent fluctuation characteristics includes: determining the boundary between the allowable deviation range and the actual distribution range, and calculating the ratio of the overlapping area of ​​the allowable deviation range and the actual distribution range to the total range area as the degree of overlap.

[0028] Furthermore, based on the parameter resonance risk level, the intersection nodes of production paths with high parameter resonance risk levels in the process network topology are identified, and these intersection nodes are determined to be key bottleneck processes that trigger systemic chain reactions, including:

[0029] Based on the assessment results of the parameter resonance risk level, all production paths whose parameter resonance risk level is assessed as high risk are screened out. Among them, a parameter resonance risk level assessed as high risk indicates a high parameter resonance risk level.

[0030] In the process network topology, process nodes that are simultaneously traversed by high-risk production paths with a preset number of risk items or more are identified as high-potential-energy convergence nodes.

[0031] The number of upstream and downstream process nodes connected to the high-potential-energy convergence node in the process network topology is compared with the average number of connections of all process nodes in the network topology.

[0032] When the number of connections at a high-potential-energy convergence node exceeds a preset multiple of the average number of connections, the corresponding high-potential-energy convergence node is determined to be a key bottleneck process that triggers a systemic chain reaction.

[0033] Furthermore, a dynamic routing scheme is generated based on the parameter resonance risk level and key bottleneck processes. This dynamic routing scheme avoids production paths with high parameter resonance risk levels and prioritizes bypassing key bottleneck processes, including:

[0034] Production paths with all parameters resonating at a high risk level are excluded from the candidate production line set.

[0035] Among the unresolved paths, select those that do not pass through any critical bottleneck processes to form a preliminary routing scheme.

[0036] If the primary routing scheme is an empty set, then the secondary routing scheme is formed by selecting the path with the fewest critical bottleneck processes from the paths that pass through the critical bottleneck processes.

[0037] Based on the decision rule that the primary routing scheme is preferred when it is not empty and the secondary routing scheme is used otherwise, the final dynamic routing scheme is generated.

[0038] Furthermore, the dynamic routing scheme is sent to the manufacturing execution system, triggering adjustments to the work-in-process production path, including:

[0039] Convert dynamic routing schemes into production instruction formats that can be recognized by the manufacturing execution system;

[0040] Production instructions are transmitted through the data interface between the manufacturing execution system and the production line controller.

[0041] The real-time process path identifier of work-in-process is updated according to the production order, and the material distribution system is triggered to adjust the distribution path synchronously with the real-time process path identifier.

[0042] On the other hand, the present invention provides a digital twin-driven collaborative control system for LCD backlight panel production, comprising:

[0043] The model building module is used to build a digital twin model corresponding to the physical production line. The digital twin model includes the process capability vectors of each production line at different processes and the process path vectors of each work-in-process.

[0044] The set identification module is used to identify the set of candidate production lines corresponding to each process in the current complete process path of work-in-process based on the process capability vector and the process path vector.

[0045] The rating assessment module is used to identify sensitive process parameters in the process path vector of the current work-in-process and compare them with the inherent fluctuation characteristics corresponding to the process capability vector of each production line in the candidate production line set to assess the risk level of parameter resonance.

[0046] The process judgment module is used to identify the intersection nodes of production paths with high parameter resonance risk levels in the process network topology based on the parameter resonance risk level, and to determine the intersection nodes as key bottleneck processes that trigger systemic chain reactions.

[0047] The solution generation module is used to generate dynamic routing solutions based on the parameter resonance risk level and key bottleneck processes. The dynamic routing solutions avoid production paths with high parameter resonance risk levels and prioritize bypassing key bottleneck processes.

[0048] The path adjustment module is used to send dynamic routing schemes to the manufacturing execution system, triggering adjustments to the work-in-process production path.

[0049] The beneficial effects of this invention are:

[0050] 1. By constructing a digital twin model that is synchronized with the physical production line in real time, the system integrates the process capability vectors of each production line at different processes and the process path vectors of each work-in-process, forming a virtual mapping and dynamic representation of the production process. Based on the collaborative analysis of process capability vectors and process path vectors, the system can accurately identify the candidate production line set corresponding to each process in the complete process path of the current work-in-process, thereby achieving a high degree of matching between production line capacity and process requirements. This effectively solves the resource mismatch problem caused by neglecting the capability differences of production lines at specific processes in traditional static planning, significantly improving the scientificity and adaptability of production path selection. By continuously monitoring the fluctuation characteristics of sensitive process parameters and assessing the parameter resonance risk level, the system can provide early warnings and avoid potential quality hazards, ensuring that work-in-process is always on the optimal production path, thereby enhancing the stability of the production process and the consistency of product quality.

[0051] 2. By identifying the intersection nodes of production paths with high parameter resonance risk levels in the process network topology, the system accurately locates key bottleneck processes that trigger systemic chain reactions and generates dynamic routing schemes accordingly. This proactively avoids high-risk paths and prioritizes bypassing key bottlenecks, enabling real-time optimization of production resources and order requirements. This avoids efficiency losses and quality fluctuations caused by structural contradictions. The dynamic routing scheme triggers immediate adjustments to work-in-process production paths through the manufacturing execution system and synchronizes material delivery paths, ensuring the synergy and continuity of the production process. This not only improves the utilization efficiency of production line resources and the reliability of order delivery but also enhances the production system's responsiveness to abnormal situations and the agility of overall control, providing a comprehensive and sophisticated collaborative management solution for LCD backlight panel manufacturing. Attached Figure Description

[0052] Figure 1 This is a flowchart of the digital twin-driven collaborative control method for LCD backlight production according to the present invention.

[0053] Figure 2 This is a schematic diagram of the structure of the digital twin-driven collaborative control system for LCD backlight production of the present invention. Detailed Implementation

[0054] 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.

[0055] Example 1: Figure 1 This invention presents a digital twin-driven collaborative control method for LCD backlight panel production, comprising:

[0056] S1. Construct a digital twin model corresponding to the physical production line. The digital twin model includes the process capability vectors of each production line at different processes and the process path vectors of each work-in-process.

[0057] S2. Based on the process capability vector and process path vector, identify the set of candidate production lines corresponding to each process in the current complete process path of work-in-process;

[0058] S3. Identify the sensitive process parameters in the process path vector of the current work-in-process and compare them with the inherent fluctuation characteristics corresponding to the process capability vector of each production line in the candidate production line set to assess the risk level of parameter resonance.

[0059] S4. Based on the parameter resonance risk level, identify the intersection nodes of production paths with high parameter resonance risk levels in the process network topology, and determine the intersection nodes as key bottleneck processes that trigger systemic chain reactions.

[0060] S5. Generate a dynamic routing scheme based on the parameter resonance risk level and key bottleneck process. The dynamic routing scheme avoids production paths with high parameter resonance risk level and prioritizes bypassing key bottleneck processes.

[0061] S6. Send the dynamic routing scheme to the manufacturing execution system to trigger the adjustment of the work-in-process production path.

[0062] S1. Construct a digital twin model corresponding to the physical production line. The digital twin model includes the process capability vectors of each production line at different processes and the process path vectors of each work-in-process. The specific implementation is as follows:

[0063] A digital twin model corresponding to the physical production line is constructed. This model includes process capability vectors for each production line at different processes and process path vectors for each work-in-process. The process capability vectors are generated by collecting historical production data from each production line at different processes and extracting key performance indicators (KPIs). These KPIs include process accuracy and process cycle time. The specific implementation process is as follows: First, historical production data is retrieved from the production line's manufacturing execution system database, for example, data from the past six months. The data is stored in structured tables, containing fields such as process identifier, timestamp, and measurement value. For example, for the film deposition process of LCD backlight panels, the collected data includes film thickness measurements in nanometers and deposition time records in seconds, with a data collection frequency of once per minute. Then, the data is cleaned to remove outliers. For example, by calculating the mean and standard deviation of each process's measurement values, data points exceeding three times the standard deviation are considered outliers and excluded. The three-times-standard-deviation threshold is set based on the principle of statistical control charts and is used to identify points where the process is out of control. The calculation uses the sample standard deviation formula based on historical data sequences. Next, key performance indicators (KPIs) are extracted. Process accuracy is quantified by calculating the deviation of product dimensions or parameters from standard values. For example, for a cutting process, the accuracy indicator is the average value of the cutting edge deviation, measured in micrometers. The calculation is based on the average of at least 100 samples, with the sample size set to ensure a 95% confidence level. The deviation is calculated using the absolute deviation formula, which averages the absolute values ​​of the differences between each sample value and the standard value. Process cycle time is obtained by calculating the average time interval from the start to the end of the process, measured in seconds. For example, the time difference between the start signal and the finish signal is recorded. Abnormally long or short time intervals are excluded during cycle time calculation; for example, data with time intervals less than 1 second or greater than 1 hour are considered invalid. The determination of invalid data is based on process knowledge, such as time ranges set by engineers. Finally, the extracted indicators are organized into a process capability vector. The vector data structure includes process identifier, accuracy value, and cycle time value, stored in the memory of the digital twin platform for subsequent analysis, such as in risk level assessment.

[0064] The process path vector is generated by parsing the manufacturing process documents of each work-in-process and mapping the process sequence. The process sequence includes the process order and process parameter requirements. The specific implementation process is as follows: First, the manufacturing process documents of the work-in-process are obtained from the Enterprise Resource Planning (ERP) system or Product Data Management (PDM) system. The document format includes XML or JSON. For example, the document contains a process list, parameter table, and dependencies. The document is accessed via HTTP protocol or file system. Then, a parser is used to extract the process sequence. The parser identifies the process name and order based on regular expressions or document structure. For example, for LCD backlight panel assembly, the sequence is substrate cleaning, film deposition, heat treatment, and inspection. The parser uses a string matching algorithm to identify the sequence. The process keywords are then mapped to process parameter requirements, including temperature, pressure, and time. For example, the parameter requirements for the film deposition process are a temperature range of 150°C to 200°C and a pressure range of 0.1 MPa to 0.5 MPa. These parameters are read from specified fields in the document and converted into numerical data. The parameter range is set based on process specifications, such as through engineer input or historical best values. The range values ​​are stored as minimum and maximum values. Finally, a process path vector is generated. The vector data structure includes in-process identifiers, a process sequence list, and parameter requirements for each process. This data is stored in the database of the digital twin platform to ensure consistency with the process specifications of the physical production line, for example, for production path planning.

[0065] The digital twin model integrates process capability vectors and process path vectors to form a virtual mapping that is synchronized with the physical production line in real time. The specific implementation process is as follows: First, a data model is established in the digital twin platform, associating the process capability vectors and process path vectors, for example, by matching process identifiers as key values ​​to form a unified data structure. The data structure uses relational database tables or in-memory objects. Then, a real-time synchronization mechanism is implemented, acquiring real-time data from sensors and controllers on the physical production line through an application programming interface (API). For example, process status and parameter values ​​are updated every 5 seconds. The data interface uses a REST API or MQTT protocol, and the transmitted data is in JSON format. The synchronization process includes data verification, such as checking whether the real-time data is within the parameter requirements defined by the process path vector. If it exceeds the range, an alarm is triggered. The alarm threshold is set based on the allowable deviation of the process, for example, when the temperature deviation exceeds ±5 degrees Celsius. The deviation is calculated using the difference between the real-time value and the value within the required range. Finally, the virtual mapping displays the production status through a graphical interface or data stream, such as simulating the flow of work-in-process and process execution, ensuring the dynamic consistency between the digital twin model and the physical production line. This is used to monitor and optimize the production process, for example, displaying real-time progress in a web interface.

[0066] S2. Based on the process capability vector and process path vector, identify the candidate production line set corresponding to each process in the current complete process path of work-in-process, specifically implemented as follows:

[0067] Based on process capability vectors and process path vectors, this method identifies the candidate production line set corresponding to each process in the complete process path of the current work-in-process. This includes extracting the capability parameters of each production line at the corresponding process from the process capability vector for each process in the process path vector. The specific implementation process is as follows: First, based on the process sequence defined in the process path vector of the work-in-process, each process node is traversed one by one. For example, for the film deposition process in LCD backlight panel production, the identifier of this process is read from the digital twin model. The process identifier is stored in string format, for example, process code DEP001. Then, the process capability vector database is accessed to retrieve the capability parameters recorded by all production lines at this specific process. The capability parameters include the average process accuracy and the average process cycle time. For example, the unit of the average process accuracy is micrometers. The unit for the time average of the sequence period is seconds. These parameters are derived from the statistical results of historical production data. The historical data storage time range is the most recent 6 months, and the data collection frequency is once a day. During the extraction process, data format conversion is required to convert the stored numerical data into a comparable floating-point format. For example, string-type capability parameters are converted into double-precision floating-point numbers. The conversion rules are based on a predefined data type mapping table, which defines the conversion rules from strings to numbers. For example, the string "2.5" is converted into the floating-point number "2.5". At the same time, the integrity of the data is verified. If the capability parameter of a certain production line is missing, the production line is temporarily excluded until the data is replenished. The determination of missing data is based on null value checks. For example, if the parameter value is empty or null, it is considered missing, and a missing log is recorded for subsequent data replenishment processes.

[0068] The extracted capability parameters are matched with the parameter requirements of the corresponding processes in the process path vector. The specific implementation process is as follows: First, the parameter requirement ranges defined in the process path vector are read. For example, the temperature requirement range for the film deposition process is 150 degrees Celsius to 200 degrees Celsius, and the pressure requirement range is 0.1 MPa to 0.5 MPa. The parameter requirement ranges are stored in the form of minimum and maximum values ​​in JSON object format. Then, matching rules are established. The matching rules include parameter inclusiveness judgment, that is, the capability parameter values ​​of the production line must completely fall within the parameter requirement range defined in the process path vector to be considered a successful match. For example, if the average process accuracy of a certain production line is 2.5 micrometers, while the accuracy range required by the process path vector is 1.5 micrometers to... If the deviation is 3.5 micrometers, it is considered a match. During the matching process, parameter tolerance thresholds are set. For example, the tolerance threshold for temperature parameters is ±2 degrees Celsius. This threshold is set based on the experience of process experts. The allowable deviation range is collected through expert interviews. Interview questions include allowable temperature deviation values, which are used to handle measurement errors. For multi-parameter matching, all parameters need to meet the matching conditions simultaneously. For example, process accuracy and process cycle time must meet the requirements. The matching calculation adopts the parameter-by-parameter comparison method, comparing the value of each parameter with the minimum and maximum values ​​of the required range in turn. The comparison operation uses the greater than or equal to and less than or equal to operators. For example, for temperature parameters, it checks whether the production line temperature value is greater than or equal to 150 degrees Celsius and less than or equal to 200 degrees Celsius.

[0069] The successful matching production lines are included in the candidate production line set. The specific implementation process is as follows: First, an empty candidate production line set data structure is created. This data structure is stored in list form, and each element contains a production line identifier and the corresponding process capability parameters. Then, the matching results of all production lines are traversed, and the identifiers and capability parameters of the successfully matched production lines are added to the set. At the same time, the matching timestamp and matching score are recorded. The matching score is calculated based on the degree of deviation of the parameters from the center value of the required range. For example, the deviation degree is calculated using linear interpolation. The smaller the deviation degree, the higher the score. The score is calculated as 1 minus the deviation degree divided by the parameter range width. Finally, the candidate production line set is sorted according to the matching score, and production lines with higher scores are selected as preferred candidates. The sorting algorithm adopts the bubble sort method, which is implemented by exchanging positions through pairwise comparisons. The number of sorting iterations is the list length minus 1. The candidate production line set is updated in real time to the digital twin platform for subsequent parameter resonance risk level assessment, such as re-executing the matching process when the process path vector changes.

[0070] S3. Identify sensitive process parameters in the current work-in-process process path vector and compare them with the inherent fluctuation characteristics corresponding to the process capability vectors of each production line in the candidate production line set to assess the risk level of parameter resonance. Specifically, this is implemented as follows:

[0071] The process identifies sensitive process parameters in the current work-in-process's process path vector and compares them with the inherent fluctuation characteristics corresponding to the process capability vectors of each production line in the candidate production line set. It then assesses the risk level of parameter resonance. This includes identifying process parameters in the current work-in-process's process path vector that have a decisive impact on the final product quality as sensitive process parameters. The specific implementation process is as follows: First, based on the list of process parameters recorded in the process path vector, key parameters are determined through quality characteristic correlation analysis. For example, for the film thickness parameter in the LCD backlight panel production process, its sensitivity is determined by analyzing the correlation between this parameter and product yield in historical quality data. The correlation degree is calculated using the Pearson correlation coefficient method, which requires at least 100 sets of historical data points. Each set of data points includes parameter measurements and corresponding product yield values. A strong correlation is defined as a correlation coefficient with an absolute value greater than 0.7. The 0.7 threshold is set based on statistical significance, and the significance of the correlation is determined through hypothesis testing. Then, the data is verified by the experience of process experts. For example, more than three process engineers are organized to score the importance of the parameters using a scale of 1 to 5, where 1 indicates no importance and 5 indicates very important. Parameters with an average score of 4 or higher are considered sensitive process parameters. Finally, a list of sensitive process parameters is formed, which includes the parameter name, unit, and allowable deviation range. For example, the allowable deviation range for the film thickness parameter is ±5 nanometers. This range is derived from the product design specifications, which are provided by the product design department and updated regularly.

[0072] The long-term statistical distribution characteristics of corresponding process parameters are extracted from the process capability vectors of each production line in the candidate production line set as inherent fluctuation characteristics. The specific implementation process is as follows: First, access the historical production data stored in the process capability vector database. The data time span is no less than 6 months. For example, extract the thickness measurement value sequence of a certain production line in the film deposition process. The number of measurements is no less than 1000 to ensure statistical significance. Then, calculate the statistical characteristics of each process parameter, including mean, standard deviation, and distribution range. The mean is calculated using the arithmetic mean method, that is, the sum of all measurements divided by the number of measurements. Data precision issues are handled during the calculation, such as retaining 3 decimal places. The standard deviation is calculated using the sample standard deviation formula, that is, the square root of the sum of the squares of the differences between each measurement and the mean divided by the number of measurements minus 1. The distribution range is determined by calculating the minimum and maximum values ​​of the measurements. Finally, the statistical characteristics are verified. For example, check whether the standard deviation is within a reasonable range. If the standard deviation exceeds twice the historical average, the data is re-checked. The historical average is obtained by calculating the average standard deviation of all production lines in the past year. The multiple of 2 is set based on empirical data to identify abnormal fluctuations.

[0073] The degree of overlap between the allowable deviation range of sensitive process parameters and the actual distribution range characterized by inherent fluctuation characteristics is analyzed. The specific implementation process is as follows: First, the upper and lower boundaries of the allowable deviation range are determined. For example, the allowable deviation range for film thickness parameters is 95 nm to 105 nm, and the boundary values ​​are read from the process path vector. Then, the upper and lower boundaries of the actual distribution range are determined. The actual distribution range is calculated based on the mean plus or minus 3 times the standard deviation. For example, if the mean is 100 nm and the standard deviation is 1.5 nm, then the actual distribution range is 95.5 nm to 104.5 nm. Next, the ratio of the overlapping area of ​​the two ranges to the total area of ​​the range is calculated as the degree of overlap. The degree of overlap is calculated by the ratio of the intersection length to the union length of the ranges. For example, if the length of the allowable deviation range is 10 nm, the length of the actual distribution range is 9 nm, and the length of the overlapping range is 8 nm, then the degree of overlap is 0.8. During the calculation, different cases of range boundaries are considered, including complete overlap, partial overlap, and no overlap. The degree of overlap is calculated separately for each case. When the ranges do not overlap, the degree of overlap is 0, and when the ranges completely overlap, the degree of overlap is 1.

[0074] The risk level of parameter resonance is classified based on the comparison between the degree of overlap and a preset threshold. A degree of overlap below the first threshold indicates high risk; an overlap between the first and second thresholds indicates medium risk; and an overlap above the second threshold indicates low risk. The specific implementation process is as follows: First, set the values ​​of the first and second thresholds, for example, the first threshold is set to 0.6 and the second threshold to 0.8. These thresholds are determined based on historical production accident data analysis. By statistically analyzing the distribution of overlap in parameter resonance cases that occurred in the past year, the 20th percentile is taken as the first threshold and the 80th percentile as the second threshold. The percentile calculation uses linear interpolation. Then, establish risk level classification rules using an interval judgment method: an overlap less than 0.6 is considered high risk, an overlap between 0.6 and 0.8 is considered medium risk, and an overlap greater than 0.8 is considered low risk. Finally, generate risk level labels for each sensitive process parameter, for example, high risk is marked in red, medium risk in yellow, and low risk in green, and store the risk level results in the risk assessment database of the digital twin platform.

[0075] The long-term statistical distribution characteristics of corresponding process parameters are extracted from the process capability vectors of each production line in the candidate production line set as inherent fluctuation characteristics. This includes calculating the mean, standard deviation, and distribution range of each production line for the corresponding process parameters based on historical production data recorded in the process capability vectors. The specific implementation process is as follows: First, the original measurement data sequence is read from the process capability vector. The data sequence must meet the integrity requirements, for example, the number of measurement values ​​for each process parameter is not less than 500. If the data volume is insufficient, the data collection time range is extended. Then, data preprocessing is performed, including removing obvious outliers. Outlier judgment adopts the 3 sigma principle, that is, measurement values ​​that deviate from the mean by more than 3 times the standard deviation are considered outliers. After removing outliers, the mean and standard deviation are recalculated. Next, the mean is calculated, which is the arithmetic mean of all valid measurement values. The calculation process adopts an iterative method to successively accumulate and then divide by the number of valid data. The standard deviation is calculated, that is, first calculate the square of the difference between each measurement value and the mean, then calculate the average of these squares, and finally take the square root to obtain the standard deviation. The distribution range is determined, that is, find the minimum and maximum values ​​among the valid measurement values ​​to form the minimum and maximum value interval.

[0076] Analyzing the overlap between the allowable deviation range of sensitive process parameters and the actual distribution range characterized by inherent fluctuation characteristics involves determining the boundaries of the allowable deviation range and the actual distribution range, and calculating the ratio of the overlapping area of ​​the allowable deviation range and the actual distribution range to the total area as the degree of overlap. The specific implementation process is as follows: First, read the upper and lower limits of the allowable deviation range defined in the process specification document. For example, the lower limit of the allowable deviation range for film thickness parameters is 95 nm, and the upper limit is 105 nm. Then, obtain the upper and lower limits of the actual distribution range from the inherent fluctuation characteristics. The actual distribution range is calculated using the mean plus or minus three times the standard deviation. For example, if the mean is... With a standard deviation of 1.5 nm and a range of 100 nm, the lower limit of the actual distribution range is 95.5 nm, and the upper limit is 104.5 nm. Next, the intersection of the two ranges is calculated, which is to take the smaller of the upper limits of the two ranges and the larger of the lower limits of the two ranges to form an overlapping range. For example, the lower limit of the overlapping range is 95.5 nm, and the upper limit is 104.5 nm. Finally, the degree of overlap is calculated, which is the ratio of the length of the overlapping range to the length of the union of the two ranges. For example, if the length of the overlapping range is 9 nm and the length of the union range is 10 nm, then the degree of overlap is 0.9. The special case of non-overlapping ranges is considered in the calculation process. When the two ranges have no intersection, the degree of overlap is 0.

[0077] S4. Based on the parameter resonance risk level, identify the intersection nodes of production paths with high parameter resonance risk levels in the process network topology, and determine the intersection nodes as key bottleneck processes that trigger systemic chain reactions. The specific implementation is as follows:

[0078] Based on the parameter resonance risk level, the intersection nodes of production paths with high parameter resonance risk levels in the process network topology are identified. These intersection nodes are then determined as critical bottleneck processes that trigger systemic chain reactions. This includes screening all production paths assessed as high-risk based on parameter resonance risk level evaluation results. A high-risk parameter resonance risk level indicates a high parameter resonance risk level. The specific implementation process is as follows: First, parameter resonance risk level data is read from the risk assessment database of the digital twin platform. This data originates from the evaluation results of step S3 and is stored in a structured table format, containing production path identifiers, risk level values, and evaluation times. The indirect stamp field, for example, includes three types of risk level values: high risk, medium risk, and low risk. Then, all production path records are traversed, and conditional statements are used to filter out production paths with a risk level field value of high risk. For example, in the programming implementation, a loop structure is used to check the risk level value one by one. When the risk level value is equal to high risk, the production path identifier is added to the high-risk production path list. Finally, a high-risk production path list is formed. The list data structure is stored in the form of an array. Each element contains a production path identifier and the corresponding process sequence information. For example, the production path identifier is in string format, and the process sequence information is in list structure, storing the process node identifiers that the path passes through.

[0079] In the process network topology, process nodes that are simultaneously traversed by high-risk production paths with a preset number of risk items or more are identified as high-potential intersection nodes. The specific implementation process is as follows: First, the process network topology data is loaded and stored in the form of a directed graph. Nodes represent processes, and edges represent the sequential relationships between processes. For example, an adjacency list data structure is used to store the topology relationships, and each node records its direct predecessor and successor nodes. Then, the number of times each process node is traversed by high-risk production paths is analyzed. Traversal is defined as the process sequence containing the node in the production path. For example, by traversing the list of high-risk production paths, the process sequence of each path is scanned, and the total number of times each node appears in all paths is counted. Next, a preset risk item threshold is set, for example, the preset risk item threshold is set to 3. This threshold is determined based on the analysis of historical production anomaly data. By statistical analysis, intersection nodes that cause systemic problems in the past year are usually simultaneously traversed by at least 3 high-risk paths. Finally, process nodes that appear more than or equal to the preset risk item threshold are selected and marked as high-potential intersection nodes. The number of paths traversed by each node is recorded, for example, using a hash table to store the node identifier and the corresponding count.

[0080] The number of upstream and downstream process nodes connected to a high-potential-energy intersection node in the process network topology is compared with the average number of connections of all process nodes in the network topology. The specific implementation process is as follows: First, calculate the number of connections for each high-potential-energy intersection node. The number of connections refers to the total number of nodes directly adjacent to that node in the topology graph, including all predecessor and successor nodes. For example, by querying the adjacency list of the topology graph, all adjacent nodes of the node are obtained and counted. Then, calculate the average number of connections for all process nodes in the network topology, which is the sum of the number of connections of all nodes divided by the total number of nodes. For example, by traversing all nodes in the topology graph, accumulating the number of connections of each node, and then dividing by the number of nodes, the number of nodes is the total number of process nodes in the topology graph. Finally, compare the number of connections for each high-potential-energy intersection node with the calculated average number of connections. The comparison operation uses the numerical value to determine the value. For example, if the number of connections for a certain high-potential-energy intersection node is 8 and the average number of connections is 5, the comparison result is that the number of connections is greater than the average number of connections.

[0081] When the number of connections at a high-potential-energy convergence node exceeds a preset multiple of the average number of connections, the corresponding high-potential-energy convergence node is determined to be a critical bottleneck process that triggers a systemic chain reaction. The specific implementation process is as follows: First, a preset multiple threshold is set, for example, a multiple of 1.5. This threshold is determined based on network connectivity analysis, and the multiple range is obtained through experiments simulating the propagation range of node failure under a production interruption scenario. Then, for each high-potential-energy convergence node, the ratio of the number of connections to the average number of connections is calculated. For example, if the number of connections is 8 and the average number of connections is 5, the ratio is 1.6. Next, this ratio is compared with the preset multiple threshold. When the ratio is greater than the preset multiple threshold, the node is determined to be a critical bottleneck process. For example, if the ratio is 1.6, which is greater than the preset multiple of 1.5, it is determined to be a critical bottleneck. Finally, a list of critical bottleneck processes is generated, which includes process node identifiers, number of connections, average number of connections, and ratio data. The list is stored in the bottleneck analysis database of the digital twin platform for subsequent dynamic routing scheme generation.

[0082] S5. Generate a dynamic routing scheme based on the parameter resonance risk level and key bottleneck processes. The dynamic routing scheme avoids production paths with high parameter resonance risk levels and prioritizes bypassing key bottleneck processes. The specific implementation is as follows:

[0083] A dynamic routing scheme is generated based on the parameter resonance risk level and key bottleneck processes. This scheme avoids production paths with high parameter resonance risk levels and prioritizes bypassing key bottleneck processes. This includes excluding all production paths assessed as high-risk from the candidate production line set. The specific implementation process is as follows: First, the candidate production line set data is obtained. This data originates from the output of step S2 and is stored in a list structure. Each element contains a production line identifier and the corresponding process path information. For example, the production line identifier is in string format, and the process path information is stored as an ordered list of process node identifiers. The list length dynamically changes according to the actual number of production paths. Then, the parameter resonance risk level assessment result is read. This result originates from the output of step S3 and is stored in a key-value pair structure, where the key is the production path identifier. The value represents the risk level, which includes three types: high risk, medium risk, and low risk. The high-risk identifier is predefined as a specific string constant. Next, each production path in the candidate production line set is traversed, and its corresponding parameter resonance risk level is checked to see if it is high risk. For example, by comparing whether the risk level field value is equal to the high-risk identifier. The comparison operation uses a string comparison function. If they are equal, the production path is removed from the candidate set. The exclusion operation is implemented using a loop structure. The number of iterations is equal to the initial length of the candidate production line set to ensure that all high-risk paths are removed. Finally, an updated candidate production line set is generated, which only contains production paths with parameter resonance risk levels of low risk and medium risk. The updated set is stored in the path planning database of the digital twin platform for subsequent routing scheme generation.

[0084] Among the unexcluded paths, paths that completely bypass critical bottleneck processes are selected to form a preliminary routing scheme. The specific implementation process is as follows: First, the list of critical bottleneck processes is read. This list comes from the output of step S4 and is stored in a list structure, containing a set of identifiers for critical bottleneck processes. For example, the identifiers for critical bottleneck processes are in string format, and the list length dynamically changes according to the actual number of critical bottleneck processes. Then, each production path in the updated candidate production line set is traversed, and each process sequence of the path is checked to see if it contains any critical bottleneck process identifiers. For example, this is done by performing an intersection operation between the process sequence of the production path and the set of critical bottleneck process identifiers. The intersection operation is performed by traversing each node of the process sequence and checking if the node is a critical bottleneck process identifier. The system checks whether a production path exists in the list of critical bottleneck processes. If it does, the intersection of these processes is non-empty; if the intersection is empty, the path is determined not to pass through a critical bottleneck process. The checking process involves iterating through the process sequence of each production path and using the inclusion relationship to determine whether each process node is in the list of critical bottleneck processes. The inclusion relationship is determined using a linear search algorithm. Next, all production paths that do not pass through critical bottleneck processes are added to the primary routing scheme set. The primary routing scheme set is stored in a list structure, with each element containing a production path identifier and a complete process sequence. Finally, the system verifies whether the primary routing scheme set is empty by checking whether the set length is greater than 0. If the set length is greater than 0, the system proceeds to the next processing step; otherwise, it triggers the generation of a secondary routing scheme.

[0085] If the primary routing scheme is empty, the secondary routing scheme is formed by selecting the path with the fewest critical bottleneck processes from the paths passing through the critical bottleneck processes. The specific implementation process is as follows: First, the secondary routing scheme generation process is triggered when the length of the primary routing scheme set is 0. Then, iterate through all production paths passing through critical bottleneck processes in the updated candidate production line set, and calculate the number of critical bottleneck processes for each production path. The calculation method is to count the number of processes in the process sequence of the path that appear in the critical bottleneck process list. For example, by iterating through the process sequence, checking whether each process is in the critical bottleneck list, and incrementing the counter if it is. Next, compare the number of critical bottleneck processes of all production paths passing through critical bottleneck processes, and find the production path with the smallest number. For example, by iterating and comparing, the minimum number of processes is initialized to a large value, such as 1000, and the number of critical bottleneck processes of each production path is compared and the minimum number of processes is updated, while recording the corresponding path. Finally, add the production path with the smallest number of critical bottleneck processes to the secondary routing scheme set. If multiple paths have the same minimum number of processes, all of them are included in the secondary routing scheme set to ensure that all optimal paths are considered.

[0086] Based on the decision rule that the primary routing scheme is prioritized when it is not empty, and the secondary routing scheme is used otherwise, the final dynamic routing scheme is generated. The specific implementation process is as follows: First, the length of the primary routing scheme set is checked. If the length is greater than 0, the primary routing scheme is selected as the final dynamic routing scheme; if the length of the primary routing scheme set is 0, the secondary routing scheme is selected as the final dynamic routing scheme. The decision process is implemented using a conditional judgment structure, such as using if-else conditional statements for branch selection in the programming implementation, with the condition judgment based on the set length value. The final dynamic routing scheme contains the selected production path set and its complete process sequence information. The scheme data is stored in a structured format, including the routing scheme identifier, the route type used, and detailed path data. Finally, the final dynamic routing scheme is used to trigger subsequent production path adjustment operations, such as sending the routing scheme to the manufacturing execution system through a message queue or API interface.

[0087] S6. Send the dynamic routing scheme to the manufacturing execution system to trigger the adjustment of the work-in-process production path. The specific implementation is as follows:

[0088] The dynamic routing scheme is sent to the Manufacturing Execution System (MES) to trigger adjustments to the work-in-process (WIP) production paths. This includes converting the dynamic routing scheme into a production instruction format recognizable by the MES. The specific implementation process is as follows: First, dynamic routing scheme data is read from the routing scheme database of the digital twin platform. This data originates from the output of step S5 and is stored in a structured format, containing routing scheme identifiers, production path sets, and process sequence information. For example, the routing scheme identifier uses a globally unique identifier format, and the production path set uses a list structure to store detailed information for multiple production paths. The list length dynamically changes according to the number of production paths included in the actual routing scheme. Then, the data format of the production instruction is defined according to the MES interface specification, such as using JSON or XML format. Data fields include the work-in-process identifier, target production line identifier, process sequence, and priority parameter, where the priority parameter's value range is specified. The priority level ranges from 1 to 10, with 1 representing the lowest priority and 10 representing the highest priority. This range is determined based on the task scheduling rules of the Manufacturing Execution System (MES). Next, data format conversion is performed, mapping the production path information in the dynamic routing scheme to the corresponding fields of the production instructions. For example, each path in the production path set is converted into an independent production instruction, each containing a complete sequence of process executions. This conversion process uses a field mapping table, which defines the correspondence between the fields of the dynamic routing scheme and the fields of the production instructions. Data validation is performed during the conversion process, such as checking the validity of the work-in-process identifier and whether the target production line is in an available state. The validation rules are based on the equipment status database of the MES, which updates the production line's working status in real time. Finally, a standardized set of production instructions is generated. This set is stored in a list structure, with each element containing complete production instruction data for subsequent transmission to the MES.

[0089] The transmission of production instructions through the data interface between the Manufacturing Execution System (MES) and the production line controller is implemented as follows: First, a communication connection is established with the MES, using either HTTP or Industrial Ethernet protocols. For example, a RESTful API interface can be used for data transmission, with the interface address pre-configured in the system configuration file. Next, data transmission parameters are configured, including a transmission timeout of 30 seconds and 3 retries. These parameters are determined based on network environment testing, and their values ​​are optimized by simulating transmission success rates under different network conditions. Then, the set of production instructions is encapsulated into data packets. The data packet structure includes a header and a body. The header contains the number of instructions and a timestamp, while the body contains the specific production instruction content. The data packet size is limited to 1MB, a limit set based on the processing capacity of the MES. Data encryption is implemented during transmission, for example, using the TLS protocol to encrypt the transmitted data, ensuring data transmission security. Finally, the production instructions are sent to the MES through the data interface. The sending process uses asynchronous transmission, while simultaneously monitoring the transmission status, such as checking the HTTP response status code. A status code of 200 indicates successful transmission; otherwise, a retry mechanism is triggered. After transmission is complete, a transmission log is recorded, containing the sending time, the number of instructions, and the transmission status.

[0090] The real-time process path identifier of work-in-process (WIP) is updated according to the production instruction, and the material distribution system is triggered to adjust the delivery path synchronously with the real-time process path identifier. The specific implementation process is as follows: First, the Manufacturing Execution System (MES) receives the production instruction and parses the instruction content to obtain the WIP identifier and target process path information. Then, the real-time process path identifier of the corresponding WIP is updated in the WIP management database of the MES. The update operation uses database transactions to ensure data consistency. For example, an SQL update statement is executed to modify the current path field of the WIP, and the update timestamp is recorded. At the same time, a path adjustment instruction is sent to the material distribution system. The instruction format is consistent with the production instruction and includes the WIP identifier and the new process path information. The material distribution system calculates the optimal delivery path based on the new process path. The calculation is based on a path planning algorithm, such as using Dijkstra's algorithm to find the shortest path, considering path length and equipment load factors. The path length is in meters and the equipment load is in kilograms. Finally, the delivery equipment is triggered to perform path adjustment, such as controlling the AGV to change its driving route. During the adjustment process, the equipment status is monitored in real time to ensure that the path switch is completed smoothly. The entire synchronization process ensures that the WIP production path and the material distribution path remain consistent in real time. For example, the path consistency is verified by periodic status checks, with the check period set to 5 seconds.

[0091] Example 2: Figure 2A schematic diagram of the digital twin-driven collaborative management and control system for LCD backlight panel production is provided. The digital twin-driven collaborative management and control system for LCD backlight panel production includes:

[0092] The model building module is used to build a digital twin model corresponding to the physical production line. The digital twin model includes the process capability vectors of each production line at different processes and the process path vectors of each work-in-process.

[0093] The set identification module is used to identify the set of candidate production lines corresponding to each process in the current complete process path of work-in-process based on the process capability vector and the process path vector.

[0094] The rating assessment module is used to identify sensitive process parameters in the process path vector of the current work-in-process and compare them with the inherent fluctuation characteristics corresponding to the process capability vector of each production line in the candidate production line set to assess the risk level of parameter resonance.

[0095] The process judgment module is used to identify the intersection nodes of production paths with high parameter resonance risk levels in the process network topology based on the parameter resonance risk level, and to determine the intersection nodes as key bottleneck processes that trigger systemic chain reactions.

[0096] The solution generation module is used to generate dynamic routing solutions based on the parameter resonance risk level and key bottleneck processes. The dynamic routing solutions avoid production paths with high parameter resonance risk levels and prioritize bypassing key bottleneck processes.

[0097] The path adjustment module is used to send dynamic routing schemes to the manufacturing execution system, triggering adjustments to the work-in-process production path.

[0098] All calculations involved in the embodiments are dimensionless numerical calculations, and the preset parameters and thresholds in the calculations are set by those skilled in the art according to the actual situation.

[0099] It should be noted that this invention can be deployed on the device itself to realize embedded applications, or it can run on a PC or other terminal with a user interface, thereby meeting various hardware environments and usage requirements.

[0100] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. A computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions according to the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. Computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wireless or wired transmission; wired transmission methods include optical fiber, twisted pair, coaxial cable, etc.; wireless transmission includes infrared, microwave, etc. Computer-readable storage media can be any available medium that a computer can access or a data storage device such as a server or data center that contains one or more sets of available media. Available media can be magnetic media (e.g., floppy disks, hard disks, magnetic tapes), optical media (e.g., DVDs), or semiconductor media. Semiconductor media can be solid-state drives.

[0101] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and modules described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0102] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or modules may be electrical, mechanical, or other forms.

[0103] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical modules; they may be located in one place or distributed across multiple network modules. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0104] In addition, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0105] If a function is implemented as a software module and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0106] The above are merely specific embodiments of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0107] In conclusion, the above are merely preferred embodiments of the present invention and are not intended to limit the present invention. 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.

Claims

1. A digital twin-driven collaborative control method for LCD backlight panel production, characterized in that, include: S1. Construct a digital twin model corresponding to the physical production line. The digital twin model includes the process capability vectors of each production line at different processes and the process path vectors of each work-in-process. S2. Based on the process capability vector and process path vector, identify the set of candidate production lines corresponding to each process in the current complete process path of work-in-process; S3. Identify sensitive process parameters in the current work-in-process process path vector and compare them with the inherent fluctuation characteristics corresponding to the process capability vectors of each production line in the candidate production line set, and assess the risk level of parameter resonance, including: Identify the process parameters in the current work-in-process process path vector that have a decisive impact on the quality of the final product as sensitive process parameters; The long-term statistical distribution characteristics of the corresponding process parameters are extracted from the process capability vectors of each production line in the candidate production line set as the inherent fluctuation characteristics. Analyze the degree of overlap between the allowable deviation range of sensitive process parameters and the actual distribution range characterized by inherent fluctuation characteristics; The risk level of parameter resonance is determined based on the comparison between the degree of overlap and the preset threshold. When the degree of overlap is lower than the first threshold, the risk level of parameter resonance is high; when the degree of overlap is between the first threshold and the second threshold, it is medium risk; and when the degree of overlap is higher than the second threshold, it is low risk. S4. Based on the parameter resonance risk level, identify the intersection nodes of production paths with high parameter resonance risk levels in the process network topology, and determine the intersection nodes as key bottleneck processes that trigger systemic chain reactions. S5. Generate a dynamic routing scheme based on the parameter resonance risk level and key bottleneck process. The dynamic routing scheme avoids production paths with high parameter resonance risk level and prioritizes bypassing key bottleneck processes. S6. Send the dynamic routing scheme to the manufacturing execution system to trigger the adjustment of the work-in-process production path.

2. The digital twin-driven collaborative control method for LCD backlight panel production according to claim 1, characterized in that, Construct a digital twin model corresponding to the physical production line. The digital twin model includes the process capability vectors of each production line at different processes and the process path vectors of each work-in-process, including: The process capability vector is generated by collecting historical production data from different processes on each production line and extracting key performance indicators, including process accuracy and process cycle time. The process path vector is generated by parsing the production process documents of each work-in-process and mapping the process sequence. The process sequence includes the process order and process parameter requirements. The digital twin model integrates process capability vectors and process path vectors to form a virtual mapping that is synchronized with the physical production line in real time.

3. The digital twin-driven collaborative control method for LCD backlight panel production according to claim 2, characterized in that, Based on process capability vectors and process path vectors, the candidate production line set corresponding to each process in the current work-in-process complete process path is identified, including: Based on the process capability vector and the process path vector, for each process in the process path vector, the capability parameters of each production line in the corresponding process are extracted from the process capability vector. Then, it is matched with the parameter requirements of the corresponding process in the process path vector, and the successfully matched production line is included in the candidate production line set.

4. The digital twin-driven collaborative control method for LCD backlight panel production according to claim 1, characterized in that, Extracting the long-term statistical distribution characteristics of the corresponding process parameters from the process capability vectors of each production line in the candidate production line set as inherent fluctuation characteristics includes: calculating the mean, standard deviation, and distribution range of each production line on the corresponding process parameters based on the historical production data recorded in the process capability vectors.

5. The digital twin-driven collaborative control method for LCD backlight panel production according to claim 1, characterized in that, The analysis of the degree of overlap between the allowable deviation range of sensitive process parameters and the actual distribution range characterized by inherent fluctuation characteristics includes: determining the boundary between the allowable deviation range and the actual distribution range, and calculating the ratio of the overlapping area of ​​the allowable deviation range and the actual distribution range to the total area as the degree of overlap.

6. The digital twin-driven collaborative control method for LCD backlight panel production according to claim 3, characterized in that, Based on the parameter resonance risk level, the intersection nodes of production paths with high parameter resonance risk levels in the process network topology are identified, and these intersection nodes are determined to be key bottleneck processes that trigger systemic chain reactions, including: Based on the assessment results of the parameter resonance risk level, all production paths whose parameter resonance risk level is assessed as high risk are screened out. Among them, a parameter resonance risk level assessed as high risk indicates a high parameter resonance risk level. In the process network topology, process nodes that are simultaneously traversed by high-risk production paths with a preset number of risk items or more are identified as high-potential-energy convergence nodes. The number of upstream and downstream process nodes connected to the high-potential-energy convergence node in the process network topology is compared with the average number of connections of all process nodes in the network topology. When the number of connections at a high-potential-energy convergence node exceeds a preset multiple of the average number of connections, the corresponding high-potential-energy convergence node is determined to be a key bottleneck process that triggers a systemic chain reaction.

7. The digital twin-driven collaborative control method for LCD backlight panel production according to claim 6, characterized in that, Dynamic routing schemes are generated based on the parameter resonance risk level and key bottleneck processes. These schemes avoid production paths with high parameter resonance risk levels and prioritize bypassing key bottleneck processes, including: Production paths with all parameters resonating at a high risk level are excluded from the candidate production line set. Among the unresolved paths, select those that do not pass through any critical bottleneck processes to form a preliminary routing scheme. If the primary routing scheme is an empty set, then the secondary routing scheme is formed by selecting the path with the fewest critical bottleneck processes from the paths that pass through the critical bottleneck processes. Based on the decision rule that the primary routing scheme is preferred when it is not empty and the secondary routing scheme is used otherwise, the final dynamic routing scheme is generated.

8. The digital twin-driven collaborative control method for LCD backlight panel production according to claim 7, characterized in that, The dynamic routing scheme is sent to the manufacturing execution system, triggering adjustments to the work-in-process production path, including: Convert dynamic routing schemes into production instruction formats that can be recognized by the manufacturing execution system; Production instructions are transmitted through the data interface between the manufacturing execution system and the production line controller. The real-time process path identifier of work-in-process is updated according to the production order, and the material distribution system is triggered to adjust the distribution path synchronously with the real-time process path identifier.

9. A digital twin-driven collaborative control system for LCD backlight production, used to implement the digital twin-driven collaborative control method for LCD backlight production as described in any one of claims 1-8, characterized in that, include: The model building module is used to build a digital twin model corresponding to the physical production line. The digital twin model includes the process capability vectors of each production line at different processes and the process path vectors of each work-in-process. The set identification module is used to identify the set of candidate production lines corresponding to each process in the current complete process path of work-in-process based on the process capability vector and the process path vector. The rating assessment module is used to identify sensitive process parameters in the process path vector of the current work-in-process and compare them with the inherent fluctuation characteristics corresponding to the process capability vector of each production line in the candidate production line set to assess the risk level of parameter resonance. The process judgment module is used to identify the intersection nodes of production paths with high parameter resonance risk levels in the process network topology based on the parameter resonance risk level, and to determine the intersection nodes as key bottleneck processes that trigger systemic chain reactions. The solution generation module is used to generate dynamic routing solutions based on the parameter resonance risk level and key bottleneck processes. The dynamic routing solutions avoid production paths with high parameter resonance risk levels and prioritize bypassing key bottleneck processes. The path adjustment module is used to send dynamic routing schemes to the manufacturing execution system, triggering adjustments to the work-in-process production path.

Images