Multi-level processing method and system for automobile manufacturing industry data based on digital twin

Abstract

The present application relates to the field of data processing technology, and discloses a multi-level processing method and system for automobile manufacturing industry data based on digital twins. The method includes: hierarchically processing automobile manufacturing data through a five-level hierarchical identification mechanism to obtain a hierarchical data set; establishing a process-equipment-material ternary coupling relationship matrix based on the hierarchical data set; dynamically assigning weights to the coupling relationship matrix to obtain multi-level processing weights; dynamically reconstructing the digital twin structure according to changes in the coupling relationship matrix; and collaboratively deciding the reconstructed structure and the processing weights to generate multi-level control instructions. The present application solves the problem in the prior art of the lack of dynamic perception and adaptive reconstruction capabilities of changes in the ternary coupling relationship of process parameters, equipment status, and material properties, and improves the accuracy and real-time performance of the digital twin system in tracking and responding to complex multi-factor collaborative changes in the automobile manufacturing process.

Description

Technical Field

[0001] The present application relates to the field of data processing technology, and in particular to a multi-level processing method and system for automobile manufacturing industry data based on digital twins. Background Art

[0002] As the automotive manufacturing industry continues its transformation toward intelligent and digitalization, digital twin technology, a key technology connecting physical manufacturing systems with the digital world, has been widely adopted across all aspects of automotive manufacturing. Existing digital twin technology primarily builds a digital mirror of the physical manufacturing system, enabling real-time monitoring, simulation analysis, and predictive optimization of the production process. In the automotive manufacturing sector, traditional digital twin systems typically employ a hierarchical data processing architecture, categorizing and processing manufacturing data at different levels, such as equipment, workshop, and factory. Data collection, storage, and analysis are then implemented through information systems such as MES, ERP, and PLM. These systems can effectively process large amounts of manufacturing data, supporting essential functions such as production planning, equipment status monitoring, and quality control, providing crucial technical support for the digital transformation of automotive manufacturers. Furthermore, existing technologies have developed process-based data processing methods that optimize the control of the manufacturing process by establishing correlation models between process parameters, equipment status, and product quality.

[0003] However, the existing digital twin-based automotive manufacturing data processing methods have significant technical defects in practical applications. First, the traditional hierarchical data processing architecture adopts a relatively rigid hierarchical structure, which is unable to adaptively adjust the data processing strategy according to the dynamic changes of the manufacturing process, resulting in low data processing efficiency and delayed response when facing complex and changeable automotive manufacturing scenarios. Secondly, the existing technology lacks in-depth modeling of the complex coupling relationship between process parameters, equipment status, and material properties. These manufacturing elements are often treated as independent modules, ignoring the dynamic correlation between them, resulting in a low degree of match between the data processing results and the actual manufacturing process. In addition, the existing digital twin system usually adopts a static weight strategy for data weight allocation, which cannot dynamically adjust the processing priority of data at different levels according to the real-time changes in production status, resulting in untimely processing of key data, affecting the real-time control and optimization decision-making of the manufacturing process. Summary of the Invention

[0004] This application provides a multi-level processing method and system for automobile manufacturing industrial data based on digital twins, which is used to solve the problem in the existing technology of lacking dynamic perception and adaptive reconstruction capabilities of changes in the ternary coupling relationship of process parameters, equipment status, and material properties, and improves the accuracy and real-time performance of the digital twin system in tracking and responding to complex coordinated changes of multiple factors in the automobile manufacturing process.

[0005] In a first aspect, the present application provides a multi-level processing method for automobile manufacturing industry data based on digital twins, the multi-level processing method for automobile manufacturing industry data based on digital twins comprising:

[0006] The automobile manufacturing industry data is processed through a five-level classification identification mechanism to obtain a hierarchical data set;

[0007] Establishing a ternary coupling relationship among process parameters, equipment status, and material attributes based on the hierarchical data set to obtain a coupling relationship matrix;

[0008] Performing dynamic weight distribution processing on the coupling relationship matrix to obtain multi-level data processing weights;

[0009] Performing difference calculation and threshold comparison between the current coefficient value of the coupling relationship matrix and the historical benchmark coefficient value to obtain a reconstruction trigger condition, identifying and classifying coupling paths of the coupling relationship matrix according to the reconstruction trigger condition to obtain a change path list, and dynamically reconstructing the digital twin structure according to the change path list to obtain a reconstructed data structure;

[0010] The reconstructed data structure and the multi-level data processing weights are subjected to collaborative decision-making processing to obtain multi-level control instructions.

[0011] In a second aspect, the present application provides a multi-stage processing system for automobile manufacturing industry data based on digital twins, the multi-stage processing system for automobile manufacturing industry data based on digital twins comprising:

[0012] The classification module is used to classify and identify the automotive manufacturing industry data through a five-level classification identification mechanism to obtain a classified data set;

[0013] A coupling module is used to establish a ternary coupling relationship among process parameters, equipment status, and material attributes based on the hierarchical data set to obtain a coupling relationship matrix;

[0014] An allocation module, configured to perform dynamic weight allocation processing on the coupling relationship matrix to obtain multi-level data processing weights;

[0015] A reconstruction module is used to perform difference calculation and threshold comparison between the current coefficient value of the coupling relationship matrix and the historical benchmark coefficient value to obtain a reconstruction trigger condition, identify and classify the coupling paths of the coupling relationship matrix according to the reconstruction trigger condition to obtain a change path list, and dynamically reconstruct the digital twin structure according to the change path list to obtain a reconstructed data structure;

[0016] The collaborative module is used to perform collaborative decision-making processing on the reconstructed data structure and the multi-level data processing weights to obtain multi-level control instructions.

[0017] In a third aspect, a multi-level processing device for automobile manufacturing industry data based on digital twin is provided, comprising: a memory and at least one processor, wherein the memory stores instructions; the at least one processor calls the instructions in the memory to enable the multi-level processing device for automobile manufacturing industry data based on digital twin to execute the above-mentioned multi-level processing method for automobile manufacturing industry data based on digital twin.

[0018] In a fourth aspect, a computer-readable storage medium is provided, in which instructions are stored. When the computer-readable storage medium is run on a computer, the computer executes the above-mentioned multi-level processing method of automobile manufacturing industry data based on digital twins.

[0019] In the technical solution provided by the present application, accurate hierarchical processing from millisecond-level equipment control data to daily-level factory management data is achieved through a five-level hierarchical identification mechanism, effectively solving the problem of inefficiency caused by mixed processing of data of different time granularities. By establishing a ternary coupling relationship among process parameters, equipment status, and material properties and constructing a coupling relationship matrix, the present invention has for the first time realized the mathematical expression and dynamic modeling of the complex correlation relationship between multiple factors in the automobile manufacturing process, overcoming the limitation of the existing technology of independently processing manufacturing factors, and significantly improving the mapping accuracy of the digital twin model to the actual manufacturing process. The dynamic weight allocation processing mechanism adaptively adjusts the processing priority of data at all levels according to the real-time changes of the abnormal urgency index, the hierarchical importance index, and the state deviation index, which completely changes the technical defect of the traditional static weight allocation strategy that cannot respond to changes in production status, and ensures that key data is given priority processing and timely response.

[0020] The dynamic reconstruction processing function of the digital twin structure enables the digital twin system to autonomously adapt to changes in the manufacturing process by monitoring changes in the coupling relationship matrix and triggering structural adjustments, solving the problem that the traditional digital twin structure is rigid and cannot track process adjustments. The collaborative decision-making processing mechanism intelligently integrates the reconstructed data structure with the multi-level data processing weights to generate a full range of control instructions from PLC control parameters to workshop logistics scheduling, realizing integrated decision-making optimization from bottom-level equipment control to high-level production scheduling. Especially in the complex process scenarios of automobile manufacturing, the Pearson correlation calculation algorithm accurately quantifies the linear correlation strength between process parameter changes and equipment status changes, providing a mathematical basis for establishing a reliable coupling relationship matrix, and the incremental update mechanism ensures the stability and continuity of the digital twin structure reconstruction process, avoiding the risk of system interruption caused by full reconstruction. The synergistic effect of these technical features enables the present invention to achieve precise control and intelligent optimization of the entire automobile manufacturing process while ensuring stable system operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0022] Figure 1 This is a schematic diagram of an embodiment of a multi-stage processing method for automobile manufacturing industry data based on digital twins in an embodiment of the present application;

[0023] Figure 2 This is a schematic diagram of an embodiment of a multi-stage processing system for automobile manufacturing industry data based on digital twins in an embodiment of the present application;

[0024] Figure 3 It is a schematic block diagram of the structure of a multi-stage processing device for automobile manufacturing industry data based on digital twins in an embodiment of the present invention. DETAILED DESCRIPTION

[0025] The embodiments of the present application provide a multi-level processing method and system for automotive manufacturing industry data based on digital twins. The terms "first", "second", "third", "fourth", etc. (if any) in the specification and claims of this application and the above-mentioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate so that the embodiments described herein can be implemented in an order other than that illustrated or described herein. In addition, the terms "including" or "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or that are inherent to these processes, methods, products or devices.

[0026] For ease of understanding, the specific process of the embodiment of the present application is described below. Figure 1 In the embodiments of the present application, an embodiment of a multi-level processing method for automobile manufacturing industry data based on digital twins includes:

[0027] Step S101: Perform hierarchical identification processing on automobile manufacturing industry data through a five-level hierarchical identification mechanism to obtain a hierarchical data set;

[0028] Step S102: establishing a ternary coupling relationship among process parameters, equipment status, and material attributes based on the hierarchical data set to obtain a coupling relationship matrix;

[0029] Step S103: Dynamically assign weights to the coupling relationship matrix to obtain multi-level data processing weights;

[0030] Step S104: Calculate the difference between the current coefficient value of the coupling relationship matrix and the historical benchmark coefficient value and compare the threshold value to obtain the reconstruction trigger condition. According to the reconstruction trigger condition, identify and classify the coupling paths of the coupling relationship matrix to obtain a list of change paths. Dynamically reconstruct the digital twin structure according to the list of change paths to obtain a reconstructed data structure.

[0031] Step S105: Perform collaborative decision-making processing on the reconstructed data structure and the multi-level data processing weights to obtain multi-level control instructions.

[0032] It is understandable that the execution subject of this application can be a multi-level processing system for automotive manufacturing industrial data based on digital twins, or a terminal or server, which is not limited here. The embodiment of this application is explained by taking the server as the execution subject as an example.

[0033] Specifically, by establishing a process-equipment-material ternary coupling relationship and a dynamic reconstruction mechanism, this approach addresses the technical issues of data processing hierarchies and insufficient cross-level collaboration in traditional automotive manufacturing. A five-level hierarchical identification mechanism classifies and processes data based on its temporal granularity and importance. The L1 equipment control data subset uses millisecond-level time windows to identify PLC output signals and servo motor feedback signals, performing real-time classification by detecting changes in signal frequency and amplitude. The L2 process monitoring data subset uses time series annotation based on welding current waveform characteristics and assembly torque variation curves, identifying process states by analyzing waveform peaks and variation trends. The L3 production status data subset uses workstation cycle deviation values ​​and equipment utilization statistics input into a minute-level identifier for classification, calculating the difference between the actual cycle and the standard cycle to determine the production status level. The L4 workshop scheduling data subset uses aggregate identification based on production plan execution progress and material consumption rate, and classifies data by comparing the degree of deviation between planned and actual values. The L5 factory management data subset classifies and identifies data based on capacity analysis and cost accounting data. A unique identifier is generated by combining data source code and timestamp to complete the hierarchical dataset construction.

[0034] The welding temperature time series and assembly pressure change series are extracted from the process monitoring data subset, and the process parameter vector is obtained through numerical vectorization. Then, the equipment vibration frequency amplitude and current load peak in the equipment control data subset are standardized and converted into a device state vector. Then, the material property vector is obtained by quantitative coding mapping based on the material strength test value and geometric dimension deviation value in the production status data subset. Subsequently, the process equipment correlation coefficient is obtained by Pearson correlation calculation based on the temperature change rate in the process parameter vector and the vibration intensity in the equipment state vector. The Pearson correlation calculation determines the degree of linear correlation by analyzing the ratio of the covariance of the two variables to their respective standard deviations. Finally, the process equipment correlation coefficient and the intensity coefficient in the material property vector are constructed into a three-dimensional matrix and the coupling strength is assigned to complete the coupling relationship matrix construction.

[0035] The dynamic weight allocation processing extracts the coupling coefficient change rate from the coupling relationship matrix to perform abnormal state discrimination processing to obtain the abnormal urgency index. The abnormal state discrimination determines the urgency by setting a change rate threshold and comparing the relationship between the current change rate and the threshold. Then, the data importance is hierarchically divided according to the three-dimensional correlation strength of process-equipment-material in the coupling relationship matrix to obtain the hierarchical importance index. Then, the current production status and the standard coupling mode in the coupling relationship matrix are compared and analyzed for deviation to obtain the state deviation index. Based on the abnormal urgency index, hierarchical importance index and state deviation index, dynamic weight calculation is performed to obtain the real-time weight coefficient. The real-time weight coefficient is assigned weight values ​​according to the five levels of equipment level, process level, production level, workshop level and factory level to complete the multi-level data processing weight generation.

[0036] The dynamic reconstruction processing calculates the difference between the current coefficient value of the coupling relationship matrix and the historical benchmark coefficient value and compares the threshold to obtain the reconstruction trigger condition. When the difference exceeds the preset threshold, the reconstruction mechanism is triggered. Then, according to the reconstruction trigger condition, the changed coupling paths in the coupling relationship matrix are identified and classified to obtain a list of changed paths. The path is identified by detecting the change amplitude and change direction of the correlation coefficients between the process parameter vector and the equipment state vector, the equipment state vector and the material attribute vector, and the material attribute vector and the process parameter vector. Then, the coupling relationship type in the change path list is mapped and matched with the corresponding data flow node in the digital twin structure to obtain a set of nodes to be adjusted. Based on the set of nodes to be adjusted, the data processing connection relationship within the digital twin structure is reconfigured and the path is optimized to obtain a new connection architecture. The new connection architecture and the original digital twin structure are incrementally updated and structurally verified to complete the generation of the reconstructed data structure.

[0037] Collaborative decision processing fuses the new connection architecture in the reconstructed data structure with the device-level weight value in the multi-level data processing weight to obtain the PLC control parameter adjustment instruction. The fusion calculation multiplies the reconstructed connection weight and the data processing weight by weighted average, and then adjusts the process-level weight value in the multi-level data processing weight based on the data flow node configuration in the reconstructed data structure to obtain the welding current and assembly torque optimization instructions. Then, based on the PLC control parameter adjustment instructions and the welding current and assembly torque optimization instructions, the automobile production line rhythm is collaboratively optimized to obtain the production line scheduling instructions. The collaborative optimization calculation determines the optimal production rhythm by comprehensively considering the equipment response time and process execution time. The production line scheduling instructions are then input into the decision tree node of the reconstructed data structure for workshop material distribution and AGV path planning to obtain the workshop logistics scheduling instructions. Based on the workshop logistics scheduling instructions and the full-process optimization goals of automobile manufacturing of the reconstructed data structure, multi-level instruction encapsulation and execution timing arrangement are performed to complete the generation of multi-level control instructions.

[0038] In a specific embodiment, the process of executing step S101 may specifically include the following steps:

[0039] Based on the millisecond time window, the PLC output signal and the servo motor feedback signal are classified into L1 level identification to obtain the device control data subset;

[0040] According to the welding current waveform characteristics and assembly torque change curve, the process parameters are marked at level L2 time series to obtain the process monitoring data subset;

[0041] The workstation beat deviation value and equipment utilization statistics are input into the minute-level identifier for L3 classification processing to obtain the production status data subset;

[0042] Perform L4 aggregation identification on the shop floor scheduling information based on the production plan execution progress and material consumption rate to obtain a shop floor scheduling data subset;

[0043] Perform L5 classification and identification on factory management information based on capacity analysis data and cost accounting data to obtain a subset of factory management data;

[0044] The equipment control data subset, process monitoring data subset, production status data subset, workshop scheduling data subset and factory management data subset are uniquely identified and coded by combining data source code and timestamp to obtain a hierarchical data set.

[0045] Specifically, when the five-level hierarchical identification mechanism uses the millisecond time window to perform L1 identification classification on the PLC output signal and the servo motor feedback signal, the millisecond time window refers to the data sampling time interval of 1 millisecond. The PLC output signal includes the switching signal and the analog signal, and the servo motor feedback signal includes the position feedback, speed feedback and torque feedback. During the data processing process, the collected original signal is first digitized and then classified and identified according to the frequency characteristics and amplitude range of the signal. The high-frequency control signal is classified as the real-time control category, and the low-frequency status signal is classified as the status monitoring category. The L1 identification classification is completed by matching the signal feature vector and a subset of the equipment control data is generated. In the L2 level timing annotation processing of process parameters based on the welding current waveform characteristics and the assembly torque change curve, the welding current waveform characteristics include timing characteristics such as current rise time, peak current, current holding time and current fall time. The assembly torque change curve reflects the change law of torque over time during the bolt tightening process. During data processing, the waveform data is first filtered and denoised, and then the key feature points including the starting point, inflection point, peak point and end point are extracted. The normal process mode, abnormal process mode and fault process mode are classified and labeled through timing feature pattern matching to form a process monitoring data subset.

[0046] When the workstation beat deviation value and equipment utilization statistical results are input into the minute-level identifier for L3 classification processing, the workstation beat deviation value is the difference between the actual production beat and the standard beat, and the equipment utilization statistical result is the ratio of the equipment's effective operating time to the total time. The minute-level identifier is a module that aggregates data with minutes as the cycle. During the data processing process, the cumulative beat deviation value of each workstation within one minute is first calculated, and then the average utilization rate of the equipment within this time period is calculated. The beat deviation value is then divided into three levels according to the degree of deviation: slight deviation, medium deviation and severe deviation. The equipment utilization rate is divided into three levels according to the degree of utilization: efficient utilization, normal utilization and inefficient utilization. Nine production status combinations are formed through cross-classification and the L3 classification processing is completed to generate a production status data subset. When performing L4 aggregation identification on workshop scheduling information based on the production plan execution progress and material consumption rate, the production plan execution progress is the ratio of the current completed output to the planned output, and the material consumption rate is the amount of material consumed per unit time. During the data processing process, the difference between the actual output and the planned output of each production line is first counted, and then the difference between the actual consumption and the theoretical consumption of various materials is calculated. Then, a comprehensive evaluation is performed based on the execution progress deviation and the material consumption deviation, and the workshop scheduling status is divided into four levels: ahead of schedule completion, normal execution of the plan, slight delay of the plan, and serious delay of the plan, forming a subset of workshop scheduling data.

[0047] In the L5 level classification and identification processing of factory management information based on capacity analysis data and cost accounting data, the capacity analysis data includes indicators such as actual capacity, designed capacity and capacity utilization rate, and the cost accounting data includes cost elements such as direct material cost, direct labor cost, manufacturing cost and management cost. When processing data, the capacity data of each workshop is first summarized and counted, and the overall capacity utilization rate and capacity gap are calculated. Then, the various cost data are classified and summarized, and the unit product cost and cost change trend are calculated. Then, a comprehensive evaluation is conducted based on the capacity utilization level and cost control situation, and the factory management status is divided into four levels: efficient operation, normal operation, efficiency to be improved and operation difficulties. After completing the L5 level classification and identification, a subset of factory management data is generated. When uniquely identifying and encoding data subsets at all levels through a combination of data source code and timestamp, the data source code includes hierarchical information such as equipment number, workstation number, and workshop number. The timestamp records the exact time when the data is generated. The coding rules adopt a hierarchical structure format, including a first-level code to represent the data level, a second-level code to represent the data source, a third-level code to represent the data type, and a fourth-level code to represent the time information. A complete unique identifier is generated by string splicing to ensure the uniqueness and traceability of each data in the entire system, forming a hierarchical data set containing data at all levels.

[0048] In a specific embodiment, the process of executing step S102 may specifically include the following steps:

[0049] Extracting welding temperature time series and assembly pressure change series from the process monitoring data subset of the hierarchical data set, and constructing numerical vectors of the welding temperature time series and assembly pressure change series to obtain process parameter vectors;

[0050] The equipment vibration frequency amplitude and current load peak value in the equipment control data subset are converted into standardized values ​​to obtain the equipment state vector;

[0051] Based on the material strength test values ​​and geometric dimension deviation values ​​in the production status data subset, quantitative coding mapping is performed to obtain the material attribute vector;

[0052] The Pearson correlation calculation is performed based on the temperature change rate in the process parameter vector and the vibration intensity in the equipment state vector to obtain the process equipment correlation coefficient;

[0053] The three-dimensional matrix of process equipment correlation coefficient and the intensity coefficient in the material attribute vector is constructed and the coupling strength is assigned to obtain the coupling relationship matrix.

[0054] Specifically, welding temperature time series and assembly pressure change series are extracted from the process monitoring data subset of the hierarchical data set. The welding temperature time series refers to the continuous numerical record of temperature changes over time during the welding process, and the assembly pressure change series refers to the numerical sequence of pressure changes over time during the assembly of parts. Numerical vectorization construction is the process of converting time series data into mathematical vectors. In the specific processing, the welding temperature time series is first segmented into time windows, and the continuous temperature data is divided into several segments according to fixed time intervals. Then, the statistical features of each segment of data are extracted, including mean, variance, maximum, minimum and change rate. Then, the same segmentation and feature extraction operations are performed on the assembly pressure change series, and the extracted feature values ​​are arranged in chronological order to form a numerical vector. The vector contains temperature characteristic components and pressure characteristic components, which constitute the basic structure of the process parameter vector. When performing standardized numerical conversion processing on the equipment vibration frequency amplitude and current load peak in the equipment control data subset, the equipment vibration frequency amplitude refers to the amplitude value of each frequency component of the vibration signal in the frequency domain analysis when the equipment is running, and the current load peak refers to the maximum instantaneous value of the current consumption during the operation of the equipment. The standardized numerical conversion processing is a mathematical transformation process that converts data of different dimensions and numerical ranges into a unified standard range. During the processing, the mean and standard deviation of the vibration frequency amplitude are first calculated, and then the mean is subtracted from each amplitude and divided by the standard deviation to obtain the standardized vibration intensity value. At the same time, the same standardized processing is performed on the current load peak to obtain the standardized current intensity value. The standardized vibration intensity value and current intensity value are combined to form an equipment state vector, which reflects the operating status characteristics of the equipment and is comparable.

[0055] When performing quantitative coding mapping based on the material strength test values ​​and geometric dimension deviation values ​​in the production status data subset, the material strength test value refers to the material mechanical performance index obtained through tensile test, compression test or bending test, and the geometric dimension deviation value refers to the difference between the actual processing size and the design size. Quantitative coding mapping is a data transformation process that converts continuous numerical values ​​into discrete codes. During the processing, a material strength grade classification standard is first established, and the strength test values ​​are divided into three grades of high strength, medium strength and low strength according to the preset threshold, and the coding values ​​are assigned to them respectively. Then, a geometric dimension deviation grade classification standard is established, and the deviation values ​​are divided into three grades of qualified, critical and out-of-tolerance according to the tolerance range, and the coding values ​​are assigned to them respectively. Then, the strength grade code and the deviation grade code are combined and operated, and the comprehensive quality coefficient is calculated by weighted summation to form a material attribute vector, which contains material performance information and processing quality information. When performing Pearson correlation calculation based on the temperature change rate in the process parameter vector and the vibration intensity in the equipment state vector, the temperature change rate refers to the amplitude of change of the welding temperature per unit time, and the vibration intensity refers to the intensity level of the equipment vibration signal. Pearson correlation calculation is a statistical method to measure the degree of linear correlation between two continuous variables. During the calculation process, all temperature change rate data in the process parameter vector are first extracted to form a data sequence, and all vibration intensity data in the equipment state vector are extracted to form a corresponding data sequence. Then, the covariance and respective standard deviations of the two data sequences are calculated, and then the covariance is divided by the product of the two standard deviations to obtain the Pearson correlation coefficient. The coefficient range is from negative one to positive one. A positive value indicates positive correlation, a negative value indicates negative correlation, and the absolute value indicates the strength of the correlation. The obtained process equipment correlation coefficient quantifies the strength of the correlation between process parameter changes and equipment state changes.

[0056] When constructing a three-dimensional matrix and assigning coupling strength values ​​to the process equipment correlation coefficients and the strength coefficients in the material property vector, the strength coefficient refers to the numerical component in the material property vector that reflects the mechanical properties of the material. Three-dimensional matrix construction refers to establishing a data structure with three-dimensional indexes to store multivariate relational data. Coupling strength assignment refers to the process of assigning values ​​to matrix elements based on the degree of correlation between different elements. The process first establishes a three-dimensional matrix structure, with the first dimension representing the process parameter category, the second dimension representing the equipment status category, and the third dimension representing the material property category. The process equipment correlation coefficient is then used as the matrix element value at the intersection of the process dimension and the equipment dimension. The correlation coefficient between the equipment status and the material properties is then calculated. The equipment-material correlation coefficient is obtained by analyzing the correlation between the equipment vibration intensity and the material strength coefficient. The correlation coefficient between the material properties and the process parameters is also calculated. The material-process correlation coefficient is obtained by analyzing the correlation between the material strength coefficient and the temperature change rate. The three correlation coefficients are weighted and fused according to preset weights. The comprehensive coupling strength value of each matrix element is calculated to complete the construction of the coupling relationship matrix, which comprehensively reflects the complex correlations between process parameters, equipment status, and material properties.

[0057] In a specific embodiment, the process of executing step S103 may specifically include the following steps:

[0058] The coupling coefficient change rate is extracted from the coupling relationship matrix to perform abnormal state discrimination processing and obtain the abnormal emergency index;

[0059] The importance of data is divided into different levels according to the three-dimensional correlation strength of process-equipment-material in the coupling relationship matrix to obtain the hierarchical importance index;

[0060] Compare and analyze the deviation between the current production status and the standard coupling mode in the coupling relationship matrix to obtain the status deviation index;

[0061] Based on the abnormal urgency index, level importance index and state deviation index, dynamic weight calculation is performed to obtain the real-time weight coefficient;

[0062] The real-time weight coefficient is assigned weight values ​​according to the five levels of equipment level, process level, production level, workshop level, and factory level to obtain multi-level data processing weights.

[0063] Specifically, abnormal state discrimination processing is performed by extracting the coupling coefficient change rate from the coupling relationship matrix. The coupling coefficient change rate refers to the speed of change of the values ​​of each element in the coupling relationship matrix within a continuous time period. The abnormal state discrimination processing is a data analysis process that identifies abnormal situations by setting a change rate threshold. During data processing, each coupling coefficient value is first extracted from the coupling relationship matrix at the current moment, and then the difference with the coupling coefficient value at the previous moment is calculated to obtain the change amount, and then the change amount is divided by the time interval to obtain the coupling coefficient change rate. Subsequently, a change rate grading standard is established, and the change rate is divided into four levels according to the value size: normal change, slight abnormality, moderate abnormality and severe abnormality. When the change rate exceeds the preset normal range threshold, it is determined to be an abnormal state. The larger the change rate value, the higher the degree of abnormality. According to the abnormality level, the corresponding urgency value is assigned to obtain an abnormal urgency index, which reflects the urgency of the current manufacturing process deviating from the normal state. When data importance is divided into levels according to the three-dimensional correlation strength of process-equipment-material in the coupling relationship matrix, the three-dimensional correlation strength refers to a quantitative indicator of the degree of mutual influence among process parameters, equipment status and material properties. The hierarchical division of data importance is the process of classifying data according to the importance based on the size of the correlation strength. During the processing, the absolute value of the correlation strength of each element in the coupling relationship matrix is ​​first calculated, and then all the correlation strength values ​​are sorted and counted. Then, based on the statistical distribution, the importance level division standard is established, and the correlation strength values ​​are divided into four levels according to percentiles: core importance, general importance, secondary importance and low importance. The larger the correlation strength value, the higher the data importance level. Each importance level is assigned a corresponding weight coefficient to obtain a hierarchical importance index, which reflects the difference in the importance of different data in the entire manufacturing process.

[0064] When performing a deviation comparison analysis between the current production status and the standard coupling mode in the coupling relationship matrix, the standard coupling mode refers to the ideal state mode of the process-equipment-material ternary coupling relationship under normal production conditions. Deviation comparison analysis is a data analysis method that evaluates the degree of deviation by calculating the difference between the current state and the standard state. During the processing process, a reference matrix of the standard coupling mode is first established. This matrix contains the standard values ​​of the coupling coefficients under various normal production conditions. Then, the difference between the current coupling relationship matrix and the standard coupling mode matrix is ​​calculated at the element level. The root mean square of the difference is then calculated to obtain the overall deviation measurement value. At the same time, the sub-item deviation values ​​of process deviation, equipment deviation, and material deviation are calculated respectively. Subsequently, a deviation grade classification standard is established based on the size of the deviation value, and the deviation degree is divided into four levels: slight deviation, small deviation, obvious deviation, and severe deviation. The corresponding deviation degree value is assigned according to the deviation level to obtain the state deviation degree index, which quantifies the degree of deviation of the current production status from the standard state. When dynamic weight calculation is performed based on the abnormal urgency index, the hierarchical importance index and the state deviation index, the dynamic weight calculation is a mathematical calculation process that comprehensively considers multiple influencing factors to determine the data processing priority. During the calculation process, the three indicators are first normalized to ensure that the numerical range of each indicator is consistent. Then, a weight calculation formula is established to weight the abnormal urgency index, the hierarchical importance index and the state deviation index according to the preset weight coefficients. Among them, the abnormal urgency index weight coefficient is set to reflect the impact of the urgency on the weight distribution, the hierarchical importance index weight coefficient is set to reflect the role of data importance on the weight distribution, and the state deviation index weight coefficient is set to represent the contribution of the deviation degree to the weight distribution. The sum of the three weight coefficients is equal to one to ensure the rationality of the calculation results. The comprehensive real-time weight coefficient is obtained through weighted summation. This coefficient dynamically reflects the processing priority of various types of data at the current moment.

[0065] The real-time weight coefficient is assigned according to the five levels of equipment, process, production, workshop, and factory. The five-level hierarchy is a hierarchical system established based on the management hierarchy and data processing granularity of the automotive manufacturing process. Weight allocation is the process of breaking down the overall weight coefficient according to the importance and impact of each level. The allocation process first determines the basic weight ratio based on the time sensitivity of each level. The equipment level processes millisecond-level real-time data with the highest time sensitivity and is assigned a higher weight. The process level processes second-level control data with high time sensitivity and is assigned the second highest weight. The production level processes minute-level status data with medium time sensitivity and is assigned a medium weight. The workshop level processes hour-level scheduling data with low time sensitivity and is assigned a lower weight. The factory level processes daily management data with the lowest time sensitivity and is assigned the lowest weight. The basic weight ratio is then dynamically adjusted based on the current real-time weight coefficient. When the real-time weight coefficient is large, the weight ratio of the lower level is increased to enhance real-time response capabilities. When the real-time weight coefficient is small, the weight ratio of the higher level is increased to enhance overall planning capabilities. The specific weight values ​​of each level are calculated through proportional allocation to form a multi-level data processing weight. This weight system guides the processing priority and resource allocation strategy of data at different levels.

[0066] In a specific embodiment, the process of executing step S104 may specifically include the following steps:

[0067] Perform difference calculation and threshold comparison between the current coefficient value of the coupling relationship matrix and the historical benchmark coefficient value to obtain the reconstruction trigger condition;

[0068] According to the reconstruction triggering conditions, the changed coupling paths in the coupling relationship matrix are identified and classified to obtain a list of changed paths;

[0069] Map and match the coupling relationship types in the change path list with the corresponding data flow nodes in the digital twin structure to obtain the set of nodes to be adjusted;

[0070] Based on the set of nodes to be adjusted, the data processing connection relationship within the digital twin structure is reconfigured and the path is optimized to obtain a new connection architecture;

[0071] The new connection architecture and the original digital twin structure are incrementally updated and structurally verified to obtain the reconstructed data structure.

[0072] Specifically, the reconstruction trigger condition is determined by performing difference calculation and threshold comparison between the current coefficient value of the coupling relationship matrix and the historical benchmark coefficient value. The historical benchmark coefficient value refers to the standard value of each element in the coupling relationship matrix recorded during the past stable production period, and the current coefficient value refers to the latest value in the coupling relationship matrix calculated in real time. The difference calculation is a mathematical operation process of subtracting the historical benchmark coefficient value from the current coefficient value to obtain the change amount. The threshold comparison is a judgment process of comparing the calculated difference with the preset change threshold. During data processing, the average value of the coupling relationship matrix in the past month is first extracted from the historical database as the historical benchmark coefficient value. Then, the difference between each element in the coupling relationship matrix at the current moment and the corresponding historical benchmark coefficient value is calculated one by one. The absolute value of the difference is then compared with the preset change threshold. When the absolute value of the difference exceeds the threshold, it indicates that the coupling relationship has changed significantly and reconstruction needs to be triggered. The number of coupling relationships that exceed the threshold and the degree of change are counted to form a reconstruction trigger condition, which includes information on the location and change amplitude of the changed coupling relationship.

[0073] When identifying and classifying the changed coupling paths in the coupling relationship matrix based on the reconstruction trigger conditions, the coupling path refers to the interaction transmission path between process parameters, equipment status, and material properties. Change path identification is the process of locating the specific coupling relationship that has changed based on the position information in the reconstruction trigger conditions. Classification processing is a data processing method that classifies the identified change paths according to the nature of the change. During the processing process, the changed coupling relationship is first located according to the coordinates of the matrix elements recorded in the reconstruction trigger conditions. Then, the numerical difference before and after the change is analyzed to determine the direction of change. When the coupling coefficient increases, it indicates that the correlation between the relevant elements has increased and is classified as an enhanced coupling path. When the coupling coefficient decreases, it indicates that the correlation between the relevant elements has weakened and is classified as a weakened coupling path. When the sign of the coupling coefficient changes, it indicates that the correlation nature between the relevant elements has fundamentally changed and is classified as a reverse coupling path. Then, according to the type of elements involved in the change, it is further subdivided into process-equipment coupling paths, equipment-material coupling paths, and material-process coupling paths, forming a list of change paths that includes path type, change nature, and impact range.

[0074] When mapping and matching the coupling relationship types in the change path list with the corresponding data flow nodes in the digital twin structure, the data flow node refers to the logical unit in the digital twin structure responsible for specific data processing functions. Mapping and matching is the association process of establishing the corresponding relationship between the coupling relationship type and the data flow node. During the processing process, a mapping table of the coupling relationship type and the data flow node is first established. The process equipment coupling relationship corresponds to the process control node and the equipment monitoring node, the equipment material coupling relationship corresponds to the equipment monitoring node and the quality inspection node, and the material process coupling relationship corresponds to the quality inspection node and the process control node. Then, according to the coupling relationship type in the change path list, the mapping table is searched to determine the corresponding data flow node. Then, the connection status and data processing capability of these nodes in the digital twin structure are checked. When the connection strength between nodes does not match the coupling relationship change, the node is marked as a node that needs to be adjusted. All nodes that need to be adjusted are collected to form a set of nodes to be adjusted. The set contains information such as node identification, current status, and adjustment requirements.

[0075] When the data processing connection relationship within the digital twin structure is reconfigured and the path is optimized based on the set of nodes to be adjusted, the data processing connection relationship refers to the data transmission and processing logical relationship between different nodes in the digital twin structure. Reconfiguration is the process of adjusting the connection parameters between nodes according to new requirements. Path optimization is the algorithmic processing process of finding the optimal data transmission path. During the configuration process, the adjustment requirements of each node in the set of nodes to be adjusted are first analyzed. The connection weights between the corresponding nodes need to be increased according to the enhancement of the coupling relationship. The connection weights between the corresponding nodes need to be reduced according to the weakening of the coupling relationship. The connection directions between the corresponding nodes need to be changed according to the reverse need of the coupling relationship. Then, the optimal connection parameters between the nodes are recalculated, including connection weights, transmission delays, and processing priorities. Then, the shortest path algorithm is used to find the optimal transmission path for data in the reconfigured structure. The algorithm selects the path with the lowest cost as the optimal path by calculating the total cost of all possible paths, forming a new connection architecture containing new connection parameters and optimized paths.

[0076] When the new connection architecture and the original digital twin structure are incrementally updated and structurally verified, incremental update refers to an update method that only modifies the changed parts while keeping the other parts unchanged. Structural verification is a verification process to check the correctness and stability of the updated structure. During the update process, the configuration parameters and connection relationships of the original digital twin structure are first backed up, and then the connection parameters in the new connection architecture are gradually applied to the original structure. For connection relationships that need to be added, new connections are directly established between the corresponding nodes. For connection relationships that need to be modified, the corresponding connection parameters are updated. For connection relationships that need to be deleted, the corresponding node connections are disconnected. Then, structural verification is performed to check whether the updated structure has loops, deadlocks, or data flow interruptions. The correctness and efficiency of data processing are verified by simulating the transmission process of data flow in the new structure. When problems are found in the verification, roll back to the backed-up original structure and reconfigure it. When the verification passes, the update is confirmed to be successful and a reconstructed data structure is formed.

[0077] In a specific embodiment, the process of performing the step of identifying and classifying the changed coupling paths in the coupling relationship matrix according to the reconstruction triggering condition may specifically include the following steps:

[0078] Based on the reconstruction trigger condition, the correlation coefficient between the process parameter vector and the equipment state vector in the coupling relationship matrix is ​​detected to obtain the process equipment change mark;

[0079] Input the reconstruction trigger condition into the coupling relationship detector of the equipment state vector and the material attribute vector to perform change pattern recognition and obtain the equipment material change mark;

[0080] According to the reconstruction trigger conditions, the correlation strength between the material attribute vector and the process parameter vector is trend analyzed and the deviation is calculated to obtain the material process change mark;

[0081] Based on the process equipment change mark, equipment material change mark and material process change mark, the change type is summarized and the path is traced to obtain the coupling path change type;

[0082] The coupling path change types are classified into enhanced coupling paths, weakened coupling paths and reverse coupling paths to obtain a list of changed paths.

[0083] Specifically, based on the reconstruction trigger condition, the change amplitude of the correlation coefficient between the process parameter vector and the equipment state vector in the coupling relationship matrix is ​​detected. The change amplitude detection is a data analysis process that quantifies the degree of change by comparing the current value of the correlation coefficient with the historical benchmark value. During the processing, the correlation coefficient corresponding to the process parameter vector and the equipment state vector is first extracted from the coupling relationship matrix. The coefficient reflects the intensity of the impact of the process parameter change on the equipment state. Then, the specific changed correlation coefficient is located according to the change position identified in the reconstruction trigger condition. Then, the difference between the current correlation coefficient and the historical benchmark correlation coefficient is calculated to obtain the change amplitude. When the change amplitude is positive, it indicates that the impact of the process on the equipment has increased. When the change amplitude is negative, it indicates that the impact of the process on the equipment has weakened. Subsequently, a change level classification standard is established according to the size of the change amplitude. The change amplitude is divided into four levels according to the numerical range: small change, slight change, obvious change and significant change. A process equipment change mark is generated according to the change level and change direction. The mark contains key information such as the change position, change level and change direction. When the reconstruction trigger condition is input into the coupling relationship detector between the equipment state vector and the material attribute vector for change pattern recognition, the coupling relationship detector is a data processing module specifically used to analyze the correlation between the equipment state and the material attribute. Change pattern recognition is a pattern recognition process that identifies the change type by analyzing the time series characteristics of data changes. During the processing, the change information in the reconstruction trigger condition is first input into the input interface of the detector. The detector locates the correlation between the equipment state vector and the material attribute vector based on the input information, and then extracts the change sequence data of the correlation in the time dimension. Then, the characteristic patterns of the change sequence are analyzed, including monotonically increasing patterns, monotonically decreasing patterns, periodic fluctuation patterns and random change patterns. The actual change sequence is calculated with the predefined standard pattern through the pattern matching algorithm. The standard pattern with the highest similarity is selected as the recognition result to generate an equipment material change mark, which describes the pattern type of the impact of the equipment state change on the material attribute.

[0084] When performing trend analysis and deviation calculation on the correlation strength between the material attribute vector and the process parameter vector according to the reconstruction trigger condition, trend analysis is a statistical analysis method that analyzes the direction and speed of change by fitting the trend line of data change, and deviation calculation is a mathematical calculation process that quantifies the degree of deviation by calculating the difference between the actual value and the expected value. During the processing, the correlation strength between the material attribute vector and the process parameter vector to be analyzed is first determined according to the reconstruction trigger condition, and then the historical data of the correlation strength in the past period of time is collected to form a time series. Then, linear regression fitting is performed on the time series data to obtain a trend line. The slope of the trend line reflects the changing speed of the correlation strength. A positive slope indicates that the correlation strength is on an upward trend, and a negative slope indicates that the correlation strength is on a downward trend. Then, the deviation between the current correlation strength value and the trend line predicted value is calculated. A positive deviation indicates that the actual value is higher than expected, and a negative deviation indicates that the actual value is lower than expected. A material process change mark is generated according to the trend direction and the degree of deviation. The mark reflects the changing trend and deviation characteristics of the correlation relationship between the material attribute and the process parameter. When performing change type induction and path tracing based on process equipment change tags, equipment material change tags, and material process change tags, change type induction is a data organization process that classifies and integrates different types of change tags. Path tracing is a logical analysis process that tracks the change propagation path by analyzing the correlation between change tags. During the processing, the change information contained in the three change tags is first collected. Then, based on the impact range of the change, the change type is divided into two categories: local change and global change. Local changes only affect a single coupling relationship, while global changes affect multiple coupling relationships. Then, a change path diagram is established based on the propagation direction of the change. The process equipment change tag represents the impact path from process parameters to equipment status, the equipment material change tag represents the impact path from equipment status to material properties, and the material process change tag represents the feedback path from material properties to process parameters. By connecting these paths, a complete change propagation network is formed. The coupling path change type is determined based on the structural characteristics of the path network, including different types such as one-way propagation, two-way interaction, and loop feedback.

[0085] When coupling path change types are categorized into enhancing, weakening, and reverse coupling paths, enhancing coupling paths refer to paths where coupling strength increases, weakening coupling paths refer to paths where coupling strength decreases, and reverse coupling paths refer to paths where coupling direction changes. Classification and organization is a data management process that categorizes and organizes path change types according to predefined classification criteria. The process first extracts the magnitude and direction of change information contained in each coupling path change type. Then, a classification rule is established: when the magnitude of change is positive and exceeds the enhancing threshold, it is classified as an enhancing coupling path; when the magnitude of change is negative and exceeds the weakening threshold, it is classified as a weakening coupling path; and when the sign of the coupling relationship changes, it is classified as a reverse coupling path. Then, based on the classification results, each coupling path change type is assigned to a corresponding classification group. The specific change characteristics and impact range of each path are recorded, forming a structured list of change paths. This list contains three main categories and their corresponding specific path information. Each path entry in the list includes detailed attributes such as path identification, change type, influencing factors, and degree of change.

[0086] In a specific embodiment, the process of executing step S105 may specifically include the following steps:

[0087] The new connection structure in the reconstructed data structure is combined with the device-level weight value in the multi-level data processing weight to obtain the PLC control parameter adjustment instruction;

[0088] Based on the data flow node configuration in the reconstructed data structure, the process level weight value in the multi-level data processing weight is adjusted to obtain the welding current and assembly torque optimization instructions;

[0089] According to the PLC control parameter adjustment instructions and the welding current and assembly torque optimization instructions, the automobile production line beat is collaboratively optimized and calculated to obtain the production line scheduling instructions;

[0090] Input the production line scheduling instructions into the decision tree nodes of the reconstructed data structure to perform workshop material distribution and AGV path planning to obtain workshop logistics scheduling instructions;

[0091] Based on the workshop logistics scheduling instructions and the automobile manufacturing full process optimization goal of the reconstructed data structure, multi-level instruction packaging and execution timing arrangement are performed to obtain multi-level control instructions.

[0092] Specifically, a PLC control parameter adjustment instruction is generated by fusing the new connection architecture in the reconstructed data structure with the device-level weight value in the multi-level data processing weight. The new connection architecture refers to the data node connection relationship and weight configuration formed after dynamic reconstruction. The device-level weight value refers to the weight coefficient specifically allocated to the device-level data in the multi-level data processing weight. Fusion calculation is a data processing process that combines weight information from two different sources through mathematical operations. During the processing, the connection weight matrix of the new connection architecture is first extracted from the reconstructed data structure. The matrix contains the connection strength and transmission priority between each data node. Then, the device-level weight value is extracted from the multi-level data processing weight. The weight value reflects the importance of the device-level data in the overall data processing. Then, the device-related weight value in the connection weight matrix and the device-level weight value are weighted averaged. The fusion weight value is obtained by weighted summing the two weight values ​​according to a preset ratio. Based on the fusion weight value, a parameter adjustment instruction for the PLC control system is generated. The instruction contains control information such as the adjustment direction, adjustment amplitude and execution priority of the control parameter. When adjusting the process-level weight value in the multi-level data processing weight based on the data flow node configuration in the reconstructed data structure, the data flow node configuration refers to the connection method and data transmission path configuration of each data processing node in the reconstructed data structure, and the process-level weight value refers to the weight coefficient specifically allocated to the process-level data in the multi-level data processing weight. The adjustment process is a data processing process in which the original weight value is corrected according to the node configuration change. During the processing, the configuration changes of the data flow nodes in the reconstructed data structure are first analyzed to identify which nodes' connection relationships have changed and the degree of change. Then, the influence coefficient on the process-level weight value is calculated based on the node configuration change. When the node connection is enhanced, the influence coefficient is positive and the process-level weight needs to be increased. When the node connection is weakened, the influence coefficient is negative and the process-level weight needs to be reduced. Then, the influence coefficient is calculated with the original process-level weight value to obtain the adjusted weight value. According to the adjusted weight value, welding current and assembly torque optimization instructions are generated. The instructions accurately adjust the current control parameters in the welding process and the torque control parameters in the assembly process.

[0093] When collaborative optimization calculation is performed on the automobile production line rhythm based on PLC control parameter adjustment instructions and welding current and assembly torque optimization instructions, collaborative optimization calculation is a mathematical optimization process that comprehensively considers the impact of multiple control instructions on the production rhythm. During the processing, the impact of PLC control parameter adjustment instructions on the equipment operation rhythm is first analyzed, and the change in equipment response time caused by parameter adjustment is calculated. Then, the impact of welding current and assembly torque optimization instructions on the process execution rhythm is analyzed, and the change in process execution time caused by process parameter adjustment is calculated. Then, a production line rhythm optimization model is established. This model takes equipment response time and process execution time as constraints, and the overall rhythm of the production line as the optimization target. The optimal production line rhythm configuration is solved by linear programming method. During the solution process, the synchronization requirements between each workstation and the capacity limit of the buffer area are considered to obtain the production line scheduling instructions for coordinating the rhythm of each workstation. This instruction contains the target rhythm time and rhythm synchronization control strategy of each workstation. When the production line scheduling instructions are input into the decision tree nodes of the reconstructed data structure for workshop material distribution and AGV path planning, the decision tree nodes are the logical processing units in the reconstructed data structure specifically responsible for decision analysis. AGV path planning refers to the calculation of the optimal driving path of the automatic guided vehicle in the workshop. During the processing, the beat information in the production line scheduling instructions is first input into the decision tree nodes. The decision tree calculates the material demand time and quantity of each workstation according to the beat requirements, and then formulates the material distribution plan based on the material demand information, including distribution batches, distribution time and distribution routes. Then, the AGV path planning calculation is performed. The optimal driving path of the AGV is determined by analyzing the workshop layout and traffic flow. The path planning takes into account multiple factors such as path length, traffic congestion and safety. The shortest path algorithm is used to calculate the optimal path from the material storage area to each workstation, and generate a workshop logistics scheduling instruction, which includes a material distribution schedule and AGV driving path planning.

[0094] When multi-level instruction packaging and execution timing are performed based on the workshop logistics scheduling instructions and the full-process optimization goal of automobile manufacturing based on the reconstructed data structure, the full-process optimization goal of automobile manufacturing refers to the comprehensive optimization goal covering the entire automobile manufacturing process. Multi-level instruction packaging is the data processing process of uniformly formatting and packaging control instructions from different levels. Execution timing is the logical processing process of sorting instructions based on their temporal dependencies. During this process, an instruction priority evaluation standard is first established based on the full-process optimization goal of automobile manufacturing. This standard comprehensively considers multiple dimensions such as production efficiency, product quality, resource utilization, and cost control. The workshop logistics scheduling instructions are then uniformly packaged with previously generated instructions at all levels. During the packaging process, each instruction is assigned a unique identifier, execution priority, and dependency tag. Then, the instructions are time-sequenced based on their logical dependencies. Equipment-level instructions have the highest execution priority and must be executed first. Process-level instructions that depend on the execution results of equipment-level instructions must be executed after equipment adjustment is completed. Production-level instructions and workshop-level instructions are executed sequentially according to the production process sequence, forming a multi-level control instruction sequence arranged in chronological order and priority. This sequence ensures the coordinated execution of instructions at all levels and the overall optimization effect.

[0095] The above describes the multi-level processing method of automobile manufacturing industry data based on digital twin in the embodiment of the present application. The following describes the multi-level processing system of automobile manufacturing industry data based on digital twin in the embodiment of the present application. Figure 2 In the embodiments of the present application, an embodiment of a multi-level processing system for automobile manufacturing industry data based on digital twins includes:

[0096] The classification module is used to classify and identify the automotive manufacturing industry data through a five-level classification identification mechanism to obtain a classified data set;

[0097] A coupling module is used to establish a ternary coupling relationship among process parameters, equipment status, and material attributes based on the hierarchical data set to obtain a coupling relationship matrix;

[0098] an allocation module, configured to perform dynamic weight allocation processing on the coupling relationship matrix to obtain multi-level data processing weights;

[0099] A reconstruction module is used to perform difference calculation and threshold comparison between the current coefficient value of the coupling relationship matrix and the historical benchmark coefficient value to obtain a reconstruction trigger condition, identify and classify the coupling paths of the coupling relationship matrix according to the reconstruction trigger condition to obtain a change path list, and dynamically reconstruct the digital twin structure according to the change path list to obtain a reconstructed data structure;

[0100] The collaborative module is used to perform collaborative decision-making processing on the reconstructed data structure and the multi-level data processing weights to obtain multi-level control instructions.

[0101] above Figure 2 The multi-level processing system for automobile manufacturing industry data based on digital twins in the embodiment of the present invention is described in detail from the perspective of modular functional entities. The multi-level processing device for automobile manufacturing industry data based on digital twins in the embodiment of the present invention is described in detail from the perspective of hardware processing.

[0102] Reference Figure 3 In the embodiment of the present invention, a multi-stage processing device for automobile manufacturing industry data based on digital twin is also provided. The multi-stage processing device for automobile manufacturing industry data based on digital twin can be a server, and its internal structure can be as follows: Figure 3 As shown. The multi-stage processing device for automobile manufacturing industry data based on digital twin includes a processor, memory, display screen, input device, network interface and database connected through a system bus. Among them, the computer-designed processor is used to provide computing and control capabilities. The memory of the multi-stage processing device for automobile manufacturing industry data based on digital twin includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and computer program in the non-volatile storage medium. The database of the multi-stage processing device for automobile manufacturing industry data based on digital twin is used to store the corresponding data in this embodiment. The network interface of the multi-stage processing device for automobile manufacturing industry data based on digital twin is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, the above method is implemented.

[0103] Those skilled in the art will understand that Figure 3 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present invention, and does not constitute a limitation on the multi-stage processing equipment for automobile manufacturing industry data based on digital twins to which the solution of the present invention is applied.

[0104] The present invention also provides a computer-readable storage medium, which may be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium. The computer-readable storage medium stores instructions. When the instructions are executed on a computer, the computer executes the steps of the multi-level processing method of automobile manufacturing industry data based on digital twins.

[0105] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described systems, systems and units can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0106] If the integrated unit is implemented as a software functional unit 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 the present invention, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a multi-stage processing device for automotive manufacturing industry data based on digital twins (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk, and other media that can store program code.

[0107] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A multi-level processing method for automobile manufacturing industry data based on digital twins, characterized by: The method comprises: The automobile manufacturing industry data is processed through a five-level classification identification mechanism to obtain a hierarchical data set; Establishing a ternary coupling relationship among process parameters, equipment status, and material attributes based on the hierarchical data set to obtain a coupling relationship matrix; The coupling relationship matrix is ​​subjected to dynamic weight distribution processing to obtain multi-level data processing weights, including: extracting the coupling coefficient change rate from the coupling relationship matrix to perform abnormal state discrimination processing to obtain an abnormal urgency index; hierarchically dividing the data importance according to the three-dimensional correlation strength of process-equipment-material in the coupling relationship matrix to obtain a hierarchical importance index; performing deviation comparison analysis between the current production status and the standard coupling mode in the coupling relationship matrix to obtain a state deviation index; performing dynamic weight calculation based on the abnormal urgency index, the hierarchical importance index and the state deviation index to obtain a real-time weight coefficient; and distributing weight values ​​of the real-time weight coefficient according to the five levels of equipment level, process level, production level, workshop level and factory level to obtain the multi-level data processing weights; Performing difference calculation and threshold comparison between the current coefficient value of the coupling relationship matrix and the historical benchmark coefficient value to obtain a reconstruction trigger condition, identifying and classifying coupling paths of the coupling relationship matrix according to the reconstruction trigger condition to obtain a change path list, and dynamically reconstructing the digital twin structure according to the change path list to obtain a reconstructed data structure; The reconstructed data structure and the multi-level data processing weights are subjected to collaborative decision-making processing to obtain multi-level control instructions.

2. The multi-level processing method for automobile manufacturing industry data based on digital twin according to claim 1 is characterized in that: The automobile manufacturing industry data is processed by hierarchical identification using a five-level hierarchical identification mechanism to obtain a hierarchical data set, including: Based on the millisecond time window, the PLC output signal and the servo motor feedback signal are classified into L1 level identification to obtain the device control data subset; According to the welding current waveform characteristics and assembly torque change curve, the process parameters are marked at level L2 time series to obtain the process monitoring data subset; The workstation beat deviation value and equipment utilization statistics are input into the minute-level identifier for L3 classification processing to obtain the production status data subset; Perform L4 aggregation identification on the shop floor scheduling information based on the production plan execution progress and material consumption rate to obtain a shop floor scheduling data subset; Perform L5 classification and identification on factory management information based on capacity analysis data and cost accounting data to obtain a subset of factory management data; The equipment control data subset, the process monitoring data subset, the production status data subset, the workshop scheduling data subset and the factory management data subset are uniquely identified and coded by combining data source code and timestamp to obtain the hierarchical data set.

3. The multi-level processing method for automobile manufacturing industry data based on digital twin according to claim 2 is characterized in that: The ternary coupling relationship of process parameters, equipment status, and material attributes is established based on the hierarchical data set to obtain a coupling relationship matrix, including: Extracting a welding temperature time series and an assembly pressure change series from a process monitoring data subset of the hierarchical data set, and numerically vectorizing the welding temperature time series and the assembly pressure change series to obtain a process parameter vector; Performing standardized numerical conversion processing on the device vibration frequency amplitude and current load peak value in the device control data subset to obtain a device state vector; Performing quantitative coding mapping based on the material strength test value and the geometric dimension deviation value in the production status data subset to obtain a material attribute vector; Performing a Pearson correlation calculation based on the temperature change rate in the process parameter vector and the vibration intensity in the equipment state vector to obtain a process equipment correlation coefficient; A three-dimensional matrix is ​​constructed and coupling strength is assigned by performing coupling on the process equipment correlation coefficient and the intensity coefficient in the material attribute vector to obtain the coupling relationship matrix.

4. The multi-level processing method for automobile manufacturing industry data based on digital twin according to claim 1 is characterized in that: The method comprises performing difference calculation and threshold comparison on the current coefficient value of the coupling relationship matrix and the historical benchmark coefficient value to obtain a reconstruction trigger condition, performing coupling path identification and classification on the coupling relationship matrix according to the reconstruction trigger condition to obtain a change path list, and dynamically reconstructing the digital twin structure according to the change path list to obtain a reconstructed data structure, including: Performing difference calculation and threshold comparison between the current coefficient value of the coupling relationship matrix and the historical benchmark coefficient value to obtain a reconstruction trigger condition; Identifying and classifying changed coupling paths in the coupling relationship matrix according to the reconstruction triggering condition to obtain a changed path list; Mapping and matching the coupling relationship types in the change path list with the corresponding data flow nodes in the digital twin structure to obtain a set of nodes to be adjusted; Reconfiguring and optimizing the data processing connection relationship within the digital twin structure based on the set of nodes to be adjusted to obtain a new connection architecture; The new connection architecture and the original digital twin structure are incrementally updated and structurally verified to obtain the reconstructed data structure.

5. The multi-level processing method for automobile manufacturing industry data based on digital twin according to claim 4 is characterized in that: The step of identifying and classifying the changed coupling paths in the coupling relationship matrix according to the reconstruction triggering condition to obtain a changed path list includes: Based on the reconstruction trigger condition, a change amplitude detection is performed on the correlation coefficient between the process parameter vector and the equipment state vector in the coupling relationship matrix to obtain a process equipment change mark; Inputting the reconstruction trigger condition into a coupling relationship detector of the device state vector and the material attribute vector to perform change pattern recognition to obtain a device material change mark; Performing trend analysis and deviation calculation on the correlation strength between the material attribute vector and the process parameter vector according to the reconstruction trigger condition to obtain a material process change mark; Based on the process equipment change mark, the equipment material change mark and the material process change mark, change type induction and path tracing are performed to obtain a coupling path change type; The coupling path change types are classified and sorted according to enhanced coupling paths, weakened coupling paths, and reverse coupling paths to obtain the changed path list.

6. The multi-level processing method for automobile manufacturing industry data based on digital twin according to claim 1 is characterized in that: The step of performing collaborative decision-making processing on the reconstructed data structure and the multi-level data processing weights to obtain multi-level control instructions includes: The new connection structure in the reconstructed data structure is combined with the device-level weight value in the multi-level data processing weight to obtain a PLC control parameter adjustment instruction; Adjusting the process-level weight values ​​in the multi-level data processing weights based on the data flow node configuration in the reconstructed data structure to obtain welding current and assembly torque optimization instructions; Performing collaborative optimization calculation on the automobile production line beat according to the PLC control parameter adjustment instruction and the welding current and assembly torque optimization instruction to obtain a production line scheduling instruction; Inputting the production line scheduling instructions into the decision tree nodes of the reconstructed data structure to perform workshop material distribution and AGV path planning to obtain workshop logistics scheduling instructions; Based on the workshop logistics scheduling instructions and the automobile manufacturing full process optimization target of the reconstructed data structure, multi-level instruction packaging and execution timing arrangement are performed to obtain the multi-level control instructions.

7. A multi-level processing system for automobile manufacturing industry data based on digital twins, characterized by: For implementing the multi-stage processing method for automobile manufacturing industry data based on digital twin according to any one of claims 1 to 6, the multi-stage processing system for automobile manufacturing industry data based on digital twin comprises: The classification module is used to classify and identify the automotive manufacturing industry data through a five-level classification identification mechanism to obtain a classified data set; A coupling module is used to establish a ternary coupling relationship among process parameters, equipment status, and material attributes based on the hierarchical data set to obtain a coupling relationship matrix; An allocation module is used to perform dynamic weight allocation processing on the coupling relationship matrix to obtain multi-level data processing weights, including: extracting the coupling coefficient change rate from the coupling relationship matrix to perform abnormal state discrimination processing to obtain an abnormal urgency index; hierarchically dividing the data importance according to the three-dimensional correlation strength of process-equipment-material in the coupling relationship matrix to obtain a hierarchical importance index; performing deviation comparative analysis between the current production status and the standard coupling mode in the coupling relationship matrix to obtain a state deviation index; performing dynamic weight calculation based on the abnormal urgency index, the hierarchical importance index and the state deviation index to obtain a real-time weight coefficient; and allocating weight values ​​of the real-time weight coefficient according to the five levels of equipment level, process level, production level, workshop level and factory level to obtain the multi-level data processing weight. A reconstruction module is used to perform difference calculation and threshold comparison between the current coefficient value of the coupling relationship matrix and the historical benchmark coefficient value to obtain a reconstruction trigger condition, identify and classify the coupling paths of the coupling relationship matrix according to the reconstruction trigger condition to obtain a change path list, and dynamically reconstruct the digital twin structure according to the change path list to obtain a reconstructed data structure; The collaborative module is used to perform collaborative decision-making processing on the reconstructed data structure and the multi-level data processing weights to obtain multi-level control instructions.

8. A multi-stage processing device for automobile manufacturing industry data based on digital twin, characterized by: It includes a memory and a processor, the memory stores a computer program that can be run on the processor, and the processor implements the multi-level processing method of automobile manufacturing industry data based on digital twins as described in any one of claims 1 to 6 when executing the computer program.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the processor executes the multi-level processing method for automobile manufacturing industry data based on digital twins as claimed in any one of claims 1 to 6.

Images