Hybrid simulation digital twin production line virtual-real interaction verification system

Abstract

The application provides a hybrid simulation digital twin production line virtual-real interaction verification system, and relates to the technical field of data processing.The system is used for: encoding the change trend between the running parameter points, converting the numerical change into a symbol sequence, combining the symbol sequence into a pattern unit, obtaining pattern unit data, combining and operating the change amplitude and the duration of each pattern unit, obtaining a dynamic consistency coefficient, constructing the pattern unit into a time sequence matrix according to the process sequence, introducing a position index into the time sequence matrix, recording the arrangement relationship of each pattern unit for item-by-item accumulation, obtaining a time sequence accumulation factor, comparing the virtual simulation result with the real equipment data, extracting a difference unit in the comparison process, structurally updating the virtual model, and performing real-time interaction between the virtual model and the real equipment data to obtain a verification result.The application ensures that the virtual model and the actual production line are always consistent through virtual-real interaction verification.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data processing, in particular to a hybrid simulation digital twin production line virtual-real interactive verification system. BACKGROUND

[0002] The verification of existing production lines is usually completed through digital twin technology, which is based on actual equipment and process flow to construct a virtual model, and then uses simulation means to verify the production process. In this way, the running state of the production line can be predicted in a virtual environment, the process layout can be optimized, and large-scale shutdown and rework in production can be avoided. Especially in electronic product assembly production lines, virtual simulation has been widely used to evaluate process connection and capacity allocation.

[0003] However, the existing simulation method generally lacks dynamic docking of virtual results and real device data. The virtual model often relies on preset parameters during operation. Once these parameters deviate from the real production environment, the simulation results may differ from the actual situation. For example, when the virtual model sets the workpiece conveying speed that does not match the real-time speed of the physical conveying belt, the simulation result may appear to be normally coordinated in pace, but in actual production, it is easy to cause process congestion and material accumulation. SUMMARY

[0004] The purpose of the present application is to provide a hybrid simulation digital twin production line virtual-real interactive verification system to solve the problems mentioned in the background.

[0005] To solve the above technical problems, the technical solution of the present application is as follows:

[0006] The hybrid simulation digital twin production line virtual-real interactive verification system comprises:

[0007] A parameter module for obtaining the running parameters of the production line process and encoding the change trend between the running parameter points to convert the numerical change into a symbol sequence to obtain symbolized parameter data;

[0008] A combination module for combining the symbol sequence into a pattern unit according to the symbolized parameter data and introducing a process identifier in the combination process to obtain pattern unit data;

[0009] A coefficient module for combining the change amplitude and duration of each pattern unit according to the pattern unit data to obtain a dynamic consistency coefficient;

[0010] A construction module for constructing a time sequence matrix according to the pattern unit data in the order of the process and introducing a position index in the time sequence matrix to record the arrangement relationship of each pattern unit to obtain time sequence matrix data;

[0011] a factor module configured to accumulate each item according to the timing matrix data and correct the accumulation by the position index to obtain timing accumulation factors;

[0012] a comparison module configured to compare the virtual simulation result with the real device data according to the dynamic consistency coefficient and the timing accumulation factors, extract difference units in the comparison process, and obtain correction parameter data;

[0013] a verification module configured to update the virtual model according to the correction parameter data, and perform real-time interaction with the real device data to obtain a verification result.

[0014] Further, the parameter module comprises:

[0015] a parameter acquisition unit configured to acquire the running process of the production line process, record the real-time running parameters of each process point by point, and obtain running parameter data;

[0016] a trend calculation unit configured to calculate the numerical difference between adjacent parameter points according to the running parameter set, and obtain change trend data;

[0017] a symbol generation unit configured to divide the change trend into rising, falling and stable according to the change trend data, and obtain a direction symbol sequence;

[0018] an encoding conversion unit configured to assign a unique code to each symbol according to the direction symbol sequence and form serialized data, and obtain symbolized parameter data.

[0019] Further, the combination module comprises:

[0020] a sequence segmentation unit configured to segment the symbolized parameter data, split the symbol sequence into a plurality of symbol segments according to the change of adjacent symbols, and obtain a segment sequence;

[0021] a feature aggregation unit configured to aggregate the segment sequence, merge the symbol segments according to the symbol change direction and the duration, and obtain an aggregated segment set;

[0022] a process identification unit configured to bind each aggregated segment to a corresponding process identification according to the aggregated segment set, and obtain a labeled segment set;

[0023] a unit combination unit configured to uniformly form a structured record of the symbol features of the segments and the process identification according to the labeled segment set, and obtain mode unit data.

[0024] Further, the coefficient module comprises:

[0025] a dynamic consistency coefficient unit configured to calculate an enhancement amount of the amplitude channel according to the mode cell data to obtain an amplitude enhancement term, calculate an enhancement amount of the length channel to obtain a length enhancement term, calculate an inhibition amount of the main channel according to a difference between lengths of adjacent mode cells to obtain an inhibition term, and fuse the amplitude enhancement term and the length enhancement term to form a ratio relationship with the inhibition term to obtain a main channel term;

[0026] a coupling deviation is calculated according to a difference between the amplitude and the length of the adjacent mode cells to obtain a coupling residual term, and a transition amplification term is calculated according to a direction mutation amount of the adjacent mode cells to obtain a transition amplification term, and the coupling residual term and the transition amplification term are fused to calculate a coupling channel contribution value to obtain a coupling channel term;

[0027] the main channel term and the coupling channel term are summed term by term to calculate a total metric to obtain a dynamic consistency coefficient.

[0028] Further, the construction module comprises:

[0029] a sequential arrangement unit configured to arrange the mode cells according to a sequence of the processes according to the mode cell data, and attach a serial number identifier to each mode cell to obtain a sequential sequence;

[0030] a matrix framework unit configured to divide rows according to the process categories and divide columns according to the serial number identifiers according to the sequential sequence to establish a matrix framework for storing the mode cells to obtain a framework matrix;

[0031] a position indexing unit configured to generate a unique index number for each mode cell position of the framework matrix, and record a correspondence between the index and the process category and the serial number identifier to obtain a matrix index table;

[0032] a matrix mapping unit configured to place the mode cells in the sequential sequence one by one to corresponding positions of the framework matrix according to the matrix index table, and record the arrangement relationship in combination with the index table to obtain time sequence matrix data.

[0033] Further, the factor module comprises:

[0034] a time sequence accumulation factor unit configured to calculate a position attenuation weight term according to a position index in the time sequence matrix data, calculate a prefix accumulation amount in a row according to values of the cells from a first column to a previous column of a current column in the same row to obtain a prefix accumulation term in the row, and calculate an accumulation attention amount according to the position attenuation weight term, the prefix accumulation term in the row, and a current cell value to obtain an accumulation attention term;

[0035] According to the accumulated attention items and the corresponding position attenuation weight items of each column, a total coverage of the row is calculated to obtain a logarithmic sum item; according to the unit values of all columns of the row, a logarithmic product channel quantity is calculated to obtain a logarithmic product item; according to the logarithmic sum item and the logarithmic product item, a multiplicative coupling strength of the two is calculated to obtain a row-level comprehensive item;

[0036] According to the accumulated attention items of all columns of the row, an overlap suppression item is calculated; according to the row-level comprehensive items of each row, a weighted convergence quantity is calculated to obtain a weighted convergence item; and the overlap suppression item and the weighted convergence item are fused to obtain a timing accumulation factor.

[0037] Further, the comparison module comprises:

[0038] A result alignment unit is configured to construct an index relationship of the dynamic consistent coefficient and the timing accumulation factor, correspondingly arrange the virtual simulation data and the real data on the same reference dimension to obtain an aligned data set.

[0039] A difference detection unit is configured to calculate deviation values of the virtual simulation result and the real device data on each corresponding dimension according to the aligned data set to obtain a difference distribution diagram.

[0040] A unit extraction unit is configured to identify a region with a deviation value deviation exceeding a preset deviation threshold according to the difference distribution diagram, mark the region as a difference unit to obtain a difference unit set.

[0041] A parameter correction unit is configured to adjust the simulation parameters according to the amplitude feature, the timing feature and the position index of the difference unit set to obtain corrected parameter data.

[0042] Further, the difference detection unit comprises:

[0043] A deviation extraction unit is configured to calculate difference values between the virtual simulation result and the real device data on each dimension according to the aligned data set to obtain an original deviation sequence.

[0044] A deviation normalization unit is configured to proportionally scale the deviation values of each dimension according to the original deviation sequence to eliminate dimensional differences to obtain a normalized deviation sequence.

[0045] A feature weighting unit is configured to amplify the deviation values of the amplitude dimension by the dynamic consistent coefficient and correct the deviation values of the time dimension by the timing accumulation factor to obtain a weighted deviation sequence.

[0046] A timing fusion unit is configured to calculate the continuity and accumulation degree of the deviation values in the time dimension according to the weighted deviation sequence to obtain a fused deviation sequence.

[0047] A distribution generation unit is configured to map the fused deviation values of each dimension and the position index to a spatial coordinate system according to the fused deviation sequence to obtain the difference distribution diagram.

[0048] Further, the verification module comprises:

[0049] The model updating unit is configured to write the correction parameter data into corresponding structural elements of the virtual model, and form a new model structure in a parameterized manner to obtain an updated model.

[0050] The structural mapping unit is configured to establish a corresponding relationship between the structural elements of the virtual model and the functional modules of the real device according to the updated model, and generate a structural mapping table.

[0051] The interactive execution unit is configured to perform interaction according to the structural mapping table, establish a data channel between the virtual model and the real device, realize real-time exchange of input and output data, and obtain an interactive data set.

[0052] The verification result unit is configured to analyze the matching degree of the virtual model output and the real device data according to the interactive data set, and record it in real time to obtain a verification result.

[0053] Further, the interactive execution unit comprises:

[0054] The data channel establishing unit is configured to connect the input and output modules of the virtual model and the data interfaces of the real device according to the structural mapping table, form a bidirectional data channel, and obtain a data channel.

[0055] The data transmission unit is configured to send the input data of the virtual model to the real device through the data channel, and transmit the feedback data of the real device back to the virtual model to obtain preliminary interactive data.

[0056] The data synchronization unit is configured to perform time alignment and sequence coordination on the input and output between the virtual model and the real device according to the preliminary interactive data to obtain synchronized interactive data.

[0057] The quality inspection unit is configured to verify the validity of the synchronized interactive data, judge whether the data has packet loss, repetition or error, and obtain a qualified interactive data set.

[0058] The interactive data summarizing unit is configured to classify, integrate and archive all data generated in the interactive process according to the qualified interactive data set to obtain an interactive data set.

[0059] The above-mentioned scheme of the present application at least has the following beneficial effects:

[0060] The application realizes the structured expression of the process mode by splitting and aggregating the symbol sequence, converts the continuous production parameters into highly simplified and meaningful mode units by splitting the symbol sequence and aggregating similar features, associates different production modes with the process identifier by introducing the process identifier, ensures that the features of each process can be accurately recorded and correspond to their positions in the production line, and the system can efficiently model each process, reducing the errors of manual labeling and calculation in traditional methods, and can identify potential correlations and dependencies between multiple processes, providing accurate and structured data support for subsequent processes.

[0061] The application can accurately quantify the dynamic changes between processes by calculating the dynamic consistency coefficient, reflect the change trend and mutual influence of each mode unit, improve the adaptability and precision of the simulation model, enable the virtual model to adjust in time when facing changes in the actual production environment, ensure that the simulation results are consistent with the real production data, and more comprehensively reflect the complex relationships between processes through comprehensive amplitude enhancement, length enhancement, suppression terms and coupling bias, provide important reference for subsequent simulation verification, and enable the system to accurately distinguish the influence degree of different mode units and realize accurate production line optimization.

[0062] The application can integrate the influencing factors of each process by accumulating and correcting the time sequence matrix data item by item, derive the time sequence cumulative factor, dynamically adjust the time sequence characteristics of the data through the correction of the position index, enable the model to more accurately reflect the fluctuations and changes in the actual production process, accurately capture the interaction between each process in the production process, especially in long-time span production processes, effectively handle the deviation caused by long-time data accumulation, and avoid the situation of ignoring the overall process influence due to separate processing of each process.

[0063] The application can effectively identify the differences between virtual simulation data and real production data by comparing the dynamic consistency coefficient with the time sequence cumulative factor, accurately locate the potential problem areas in the production line by extracting the difference units, provide an automatic difference detection method, can reflect the deviation between simulation and reality in real time, provide a reliable basis for subsequent parameter correction, and deeply explore the potential optimization space between processes.

[0064] The application can establish a real-time data channel between the virtual model and the actual production equipment through real-time interaction and updating of the correction parameter data, dynamically adjust the structure of the virtual model, enable the virtual model to respond in a timely manner and be corrected when facing changes in the actual production environment, enable the data exchange between the virtual model and the real equipment to be real-time synchronized, ensure that the simulation result and the actual equipment running state always remain consistent, make the verification process of the production line more intelligent, reduce the error of manual intervention, at the same time, provide an efficient and reliable production optimization scheme, ensure the operability and reliability of the virtual simulation result, and provide more accurate decision support for the optimization of the production line. BRIEF DESCRIPTION OF DRAWINGS

[0065] Figure 1 is a flow chart of a hybrid simulation digital twin production line virtual-real interaction verification system provided by the embodiment of the application. DETAILED DESCRIPTION

[0066] Exemplary embodiments of the present disclosure will be described in greater detail below with reference to the accompanying drawings. Although exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. Rather, these embodiments are provided so that the present disclosure can be more thoroughly understood and the scope of the present disclosure can be accurately conveyed to those skilled in the art.

[0067] As Figure 1 shown, the embodiment of the application proposes a hybrid simulation digital twin production line virtual-real interaction verification system, which comprises:

[0068] A parameter module is configured to obtain running parameters of a production line process, encode the change trend between running parameter points, convert numerical changes into symbol sequences, and obtain symbolized parameter data.

[0069] A combination module is configured to combine symbol sequences into pattern units according to the symbolized parameter data, and introduce process identifiers in the combination process to obtain pattern unit data.

[0070] A coefficient module is configured to combine the change amplitude and duration of each pattern unit according to the pattern unit data to obtain dynamic consistency coefficients.

[0071] A construction module is configured to construct the pattern unit data into a time sequence matrix according to the process order, and introduce a position index in the time sequence matrix to record the arrangement relationship of each pattern unit, and obtain time sequence matrix data.

[0072] A factor module is configured to accumulate each item according to the time sequence matrix data, and correct the accumulation process through the position index to obtain time sequence accumulation factors.

[0073] A comparison module is configured to compare the virtual simulation result with the real equipment data according to the dynamic consistency coefficient and the time sequence accumulation factor, and extract a difference unit in the comparison process to obtain the modified parameter data.

[0074] A verification module is configured to perform a structured update on the virtual model according to the modified parameter data, and perform real-time interaction between the virtual model and the real equipment data to obtain a verification result.

[0075] In the embodiment of the present application, the parameter module is configured to obtain the running parameters of the production line process, encode the change trend between the running parameter points, convert the numerical change into a symbol sequence, and obtain symbolized parameter data, so that the numerical data is converted into a symbol sequence, the subtle changes in the production process can be accurately tracked and recorded, and potential process problems in the production process can be captured; the combination module is configured to combine the symbol sequence into a mode unit according to the symbolized parameter data, and introduce a process identifier in the combination process to obtain mode unit data, so that the features of each production process are accurately represented in the data through segmentation, aggregation and structured combination of the symbol sequence, and structured data is provided for subsequent processes; the coefficient module is configured to combine the change amplitude and the duration of each mode unit according to the mode unit data to obtain a dynamic consistency coefficient, so that the overall quantification of the process change is realized, and the simulation result can accurately reflect the mutual influence of each process in the production process.

[0076] The construction module is configured to construct the mode unit data into a time sequence matrix according to the process order, introduce a position index in the time sequence matrix, and record the arrangement relationship of each mode unit to obtain time sequence matrix data, so that the order relationship between the processes is clearly expressed, the order and timeliness in the production process are captured, and a basis is provided for subsequent data analysis and modification; the factor module is configured to accumulate each item according to the time sequence matrix data, and correct the time sequence matrix data through the position index in the accumulation process to obtain a time sequence accumulation factor, so that the time sequence features between the processes are accurately captured, the superposition effect of the processes in the production process is fully reflected, and a data basis is provided for subsequent processes; the comparison module is configured to compare the virtual simulation result with the real equipment data according to the dynamic consistency coefficient and the time sequence accumulation factor, and extract a difference unit in the comparison process to obtain the modified parameter data, so that the errors caused by time or process differences are effectively eliminated, the virtual simulation is always consistent with the actual production line state, and accurate parameter adjustment basis is provided for subsequent optimization; the verification module is configured to perform a structured update on the virtual model according to the modified parameter data, and perform real-time interaction between the virtual model and the real equipment data to obtain a verification result, so that dynamic synchronization between the virtual model and the real equipment is realized, real-time feedback and correction in the simulation process are ensured, and data support is provided for subsequent production line optimization.

[0077] In a preferred embodiment of the present application, the parameter module comprises:

[0078] a parameter collection unit configured to collect running processes of the production line processes and record real-time running parameters of each process point by point to obtain running parameter data;

[0079] a trend calculation unit configured to calculate numerical differences between adjacent parameter points according to the running parameter set to obtain change trend data;

[0080] a symbol generation unit configured to divide the change trend into rising, falling and stable according to the change trend data to obtain a direction symbol sequence;

[0081] an encoding conversion unit configured to assign a unique code to each symbol according to the direction symbol sequence and form serialized data to obtain symbolized parameter data.

[0082] In the embodiment of the present application, the parameter collection unit is configured to collect running processes of the production line processes and record real-time running parameters of each process point by point to obtain running parameter data, which provides a basis for trend analysis and avoids information distortion caused by data loss or too large sampling interval; the trend calculation unit is configured to calculate numerical differences between adjacent parameter points according to the running parameter set to obtain change trend data, which eliminates part of redundant information and creates conditions for subsequent symbolization operation; the symbol generation unit is configured to divide the change trend into rising, falling and stable according to the change trend data to obtain a direction symbol sequence, which avoids misjudgment caused by numerical fluctuations and quickly locates the change direction of production parameters; and the encoding conversion unit is configured to assign a unique code to each symbol according to the direction symbol sequence and form serialized data to obtain symbolized parameter data, which avoids operation incompatibility caused by symbol processing and provides input for subsequent data mining and simulation verification.

[0083] The parameter collection unit is configured to collect running processes of the production line processes and record real-time running parameters of each process point by point to obtain running parameter data, which includes:

[0084] First, data interface connection is established with various sensors on the production line, such as speed sensors, temperature sensors, pressure sensors, displacement detectors and signal output ports of controllers. When the production line enters a normal running state, the running state of each process is sampled point by point according to a preset time sampling interval, and the sampling result is provided with a corresponding time stamp to ensure the time continuity of the parameter data. The collected data is stored in the buffer area according to the process order as a classification basis, and is transmitted to the data processing module through a bus or a network when necessary to form a continuous and complete running parameter sequence.

[0085] The symbol generation unit is configured to divide the change trend into rising, falling and stable according to the change trend data, and obtain a sequence of direction symbols, including:

[0086] First, the difference value is compared with the preset threshold value. When the difference value is greater than the positive threshold value, it is determined that the running parameter at the point presents an upward trend, and an upward symbol is generated. When the difference value is less than the negative threshold value, it is determined that the running parameter at the point presents a downward trend, and a downward symbol is generated. When the difference value is between the positive threshold value and the negative threshold value, it is determined that the change amplitude of the running parameter at the point is insufficient to constitute a significant trend, and a stable symbol is generated. The threshold value can be adjusted according to the process characteristics of the specific process. For example, in a high-speed production line, the threshold value is set to be large to avoid excessive sensitivity, and in a precision process, the threshold value is set to be small to ensure that small fluctuations can be identified. The generated direction symbol sequence is arranged in time sequence, and maintains a one-to-one correspondence relationship with the original parameter point and the trend data, forming a symbolic representation representing the change direction of the production line running state.

[0087] In a preferred embodiment of the present application, the combination module comprises:

[0088] The sequence segmentation unit is configured to segment the symbolic parameter data, and split the symbol sequence according to the change of adjacent symbols to obtain a segment sequence.

[0089] The feature aggregation unit is configured to aggregate the segment sequence, and merge the symbol segments according to the symbol change direction and the duration to obtain an aggregated segment set.

[0090] The process identification unit is configured to bind each aggregated segment and the corresponding process identification according to the aggregated segment set, and obtain a labeled segment set.

[0091] The unit combination unit is configured to unify the symbol features of the segments and the process identification to form a structured record according to the labeled segment set, and obtain pattern unit data.

[0092] In the embodiment of the present application, the sequence splitting unit is used for splitting according to the symbolization parameter data, splitting the symbol sequence according to the change of adjacent symbols into a plurality of symbol segments to obtain a segment sequence, avoiding feature ambiguity caused by information mixing in a long sequence, and providing more accurate pattern analysis for subsequent modules; the feature aggregation unit is used for aggregating according to the segment sequence, merging the symbol segments according to the symbol change direction and the duration length to obtain an aggregated segment set, effectively removing random fluctuations and short-term disturbances in the data, and ensuring that the data in the pattern analysis is more accurate; the process identification unit is used for binding each aggregated segment and the corresponding process identification according to the aggregated segment set to obtain an identified segment set, mapping the abstract trend segment to the actual process on the production line, and ensuring that the pattern data has the process attribute; and the unit combination unit is used for combining the symbol features of the segments and the process identification to form a structured record, obtaining pattern unit data, ensuring that the data can maintain logical consistency and integrity when entering the next step of processing, avoiding misplacement or omission caused by scattered data sources, and providing a stable input basis for subsequent processing.

[0093] The sequence splitting unit is used for splitting according to the symbolization parameter data, splitting the symbol sequence according to the change of adjacent symbols into a plurality of symbol segments to obtain a segment sequence, and specifically includes:

[0094] First, the symbol sequence is received and the relationship between the symbols is scanned in time sequence. In the scanning process, the system focuses on the change trend between adjacent symbols. When it is detected that the symbol is switched from rising to falling, or from falling to stable, or from stable to rising, the system takes the change point as a splitting boundary, thereby disconnecting the sequence at this position. In this way, a longer symbol sequence is decomposed into a plurality of symbol segments with internal consistency, and each segment corresponds to a continuous change trend. At the same time, the system detects the switching amplitude between adjacent symbols. If the switching amplitude is lower than the preset minimum difference threshold, the splitting is not triggered, but the data is continued to be included in the current segment to avoid excessive splitting caused by data noise. After splitting, the system generates a number for all segments and arranges them in time sequence to finally form a segment sequence.

[0095] The feature aggregation unit is used for aggregating according to the segment sequence, merging the symbol segments according to the symbol change direction and the duration length to obtain an aggregated segment set, and specifically includes:

[0096] First, the duration of each segment, i.e. the number of symbols, is calculated, and its direction feature is extracted. Then the system compares the direction and length features between adjacent segments. If the adjacent segments are consistent in direction and the duration is within a preset similarity threshold, the segments are merged into a new segment, thus extending the trend. If there are some segments with opposite direction but very short duration, the system identifies them as interference segments and automatically absorbs them into the adjacent main segments to eliminate invalid information caused by acquisition fluctuations. In the merging process, the system not only saves the direction and length information of the aggregated segment, but also updates the time coverage of the segment, and finally obtains the aggregated segment.

[0097] In a preferred embodiment of the present application, the coefficient module comprises:

[0098] The dynamic consistent coefficient unit is configured to calculate an enhancement amount of the amplitude channel according to the mode unit data to obtain an amplitude enhancement term; calculate an enhancement amount of the length channel to obtain a length enhancement term; calculate an inhibition amount of the main channel according to a difference value of lengths of adjacent mode units to obtain an inhibition term; fuse the amplitude enhancement term and the length enhancement term, and form a ratio relationship with the inhibition term to obtain a main channel term;

[0099] The coupling deviation is calculated according to a difference between the amplitude and the length of adjacent mode units to obtain a coupling residual term; the influence degree of the coupling deviation is calculated according to a direction mutation amount of adjacent mode units to obtain a transition amplification term; the coupling residual term and the transition amplification term are fused to calculate a coupling channel contribution value to obtain a coupling channel term;

[0100] The main channel term and the coupling channel term are summed item by item to calculate a total metric to obtain a dynamic consistent coefficient.

[0101] In the embodiment of the present application, the dynamic consistency coefficient unit is used for calculating the enhancement amount of the amplitude channel according to the mode unit data, obtaining an amplitude enhancement term, highlighting the intensity of the process parameter change; calculating the enhancement amount of the length channel, obtaining a length enhancement term, emphasizing the persistence of the process change; calculating the suppression amount of the main channel according to the difference of the lengths of adjacent mode units, obtaining a suppression term, avoiding the excessive influence of some abnormal length changes on the overall result; fusing the amplitude enhancement term and the length enhancement term, and forming a ratio relationship with the suppression term, obtaining a main channel term, quantifying the importance of the unit in the sequence, reflecting the consistency contribution of the mode unit to the overall process from multiple dimensions; calculating the coupling deviation according to the difference of the adjacent mode units in amplitude and length, obtaining a coupling residual term, capturing the subtle mismatch between adjacent processes, and identifying the potential inconsistency between processes; calculating the influence degree of the coupling deviation according to the direction mutation amount of the adjacent mode units, obtaining a transition amplification term, highlighting the influence of the sudden change on the system consistency; fusing the coupling residual term and the transition amplification term, calculating the coupling channel contribution value, obtaining a coupling channel term, accurately reflecting the interaction intensity between different mode units; summing up the main channel term and the coupling channel term item by item, calculating the overall measurement, obtaining the dynamic consistency coefficient, realizing the unified quantification of the mode unit in the monomer dimension and the coupling dimension, enabling the system to compare the simulation data with the actual data in a refined manner, effectively making up for the shortcomings of the traditional method which only relies on static parameters, and ensuring that the simulation verification process is more accurate and reliable.

[0102] In a preferred embodiment of the present application, the configuration module comprises:

[0103] The sequential arrangement unit is used for arranging each mode unit according to the sequence of the process, and attaching a serial number to each mode unit, to obtain a sequential sequence.

[0104] The matrix framework unit is used for dividing the rows according to the process categories and dividing the columns according to the serial numbers according to the sequential sequence, to establish a matrix framework for storing the mode units, and to obtain a framework matrix.

[0105] The position index unit is used for generating a unique index number for each mode unit position of the framework matrix, and recording the correspondence between the index and the process category and the serial number, to obtain a matrix index table.

[0106] The matrix mapping unit is used for placing each mode unit in the sequential sequence into the corresponding position of the framework matrix according to the matrix index table, and recording the arrangement relationship in combination with the index table, to obtain time sequence matrix data.

[0107] In the embodiment of the present application, the sequential arrangement unit is used to arrange each mode unit according to the mode unit data in the order of the process and to attach a serial number to each mode unit to obtain a sequential sequence, so as to eliminate the confusion caused by the parallel execution of different processes or the inconsistent data collection sequence and to make the subsequent data matrix construction have a unified time reference and sequence standard; the matrix framework unit is used to divide the rows according to the process category and the columns according to the serial number based on the sequential sequence, to establish a matrix framework for storing the mode units, and to obtain a framework matrix, which directly reflects the hierarchical and sequential relationship between the processes and provides a highly ordered carrier for the subsequent process; the position index unit is used to generate a unique index number for each mode unit position of the framework matrix based on the framework matrix, to record the corresponding relationship between the index and the process category and the serial number, to obtain a matrix index table, and to eliminate the positioning uncertainty caused by the repetition of the process category or the confusion of the serial number and to provide a stable foundation for the subsequent process; and the matrix mapping unit is used to place each mode unit in the sequential sequence into the corresponding position of the framework matrix based on the matrix index table, to record the arrangement relationship in combination with the index table, to obtain time sequence matrix data, and to clearly reflect the arrangement relationship of each process and to avoid the arrangement errors caused by the manual method.

[0108] The matrix framework unit is used to divide the rows according to the process category and the columns according to the serial number based on the sequential sequence, to establish a matrix framework for storing the mode units, and to obtain a framework matrix, which directly reflects the hierarchical and sequential relationship between the processes and provides a highly ordered carrier for the subsequent process; the position index unit is used to generate a unique index number for each mode unit position of the framework matrix based on the framework matrix, to record the corresponding relationship between the index and the process category and the serial number, to obtain a matrix index table, and to eliminate the positioning uncertainty caused by the repetition of the process category or the confusion of the serial number and to provide a stable foundation for the subsequent process; and the matrix mapping unit is used to place each mode unit in the sequential sequence into the corresponding position of the framework matrix based on the matrix index table, to record the arrangement relationship in combination with the index table, to obtain time sequence matrix data, and to clearly reflect the arrangement relationship of each process and to avoid the arrangement errors caused by the manual method.

[0109] First, the process identifier and sequence number identifier carried by each pattern unit in the sequence sequence are extracted, and the process identifier is de-duplicated and sorted according to the established sequence of the production process, to obtain a row set arranged by process category; At the same time, the integrity of the sequence number identifier is checked, and the minimum sequence number and the maximum sequence number form a continuous or quasi-continuous column set, and if a missing sequence number identifier is detected, an empty placeholder item is generated for the sequence number position to maintain the continuity of the column index. According to the row set and the column set, the skeleton of the two-dimensional data structure is allocated, wherein the row dimension is bound to the process category, and the column dimension is bound to the sequence number identifier. When initialized, an empty cell record is preloaded for each unit, and the empty cell record includes an occupation flag bit, a position index placeholder, a timestamp placeholder, and an empty reference pointer pointing to the pattern unit. At the same time of establishing the two-dimensional skeleton, the row header and column header metadata of the matrix are generated; The row header records the process category, the row number, the row check code and the expansion pointer, and the column header records the sequence number identifier, the column number, the column time window and the expansion pointer; The check code is used to quickly verify the consistency of subsequent mapping stage, and the expansion pointer is used for online expansion when new process or new sequence number appears without destroying the existing address space. Subsequently, according to the Cartesian product rule of process category and sequence number identifier, a position index encoding function is defined, which maps any pair of process category and sequence number identifier to a globally unique position index string. The position index adopts a reversible encoding format of row-column combination, and the encoding format is written into the empty cell record at the initialization of the unit placeholder, which is used as the primary key for subsequent matrix mapping unit table lookup. After the skeleton and metadata are prepared, consistency constraint loading is performed: the constraint of binding a single row unique process to each row, the constraint of binding a single column unique sequence number to each column, and the constraint strategy of registering the same coordinate single main value at the matrix level to ensure that only one pattern unit can occupy a main slot of a row-column coordinate in the subsequent process. The matrix framework is divided into rows according to process category and columns according to sequence number identifier, and the initialization is completed to obtain the framework matrix containing row header, column header, empty cell record, position index rule and constraint strategy.

[0110] The matrix mapping unit is configured to place the pattern units in the sequence sequence into the corresponding positions of the framework matrix one by one according to the matrix index table, and record the arrangement relationship in combination with the index table to obtain the time sequence matrix data, and specifically includes:

[0111] Firstly, the mode units in the sequential sequence are read one by one, the process identifier and the sequence identifier are extracted from the mode unit, and then the process identifier and the sequence identifier are used as the query key to search the corresponding row and column coordinates and position index in the matrix index table; when the query hits, the corresponding empty unit record in the frame matrix is located, the occupation flag of the unit is set as occupied, the reference pointer of the mode unit is written, the timestamp and the source sequence position of the mode unit in the original sequential sequence are written, and the position index is backfilled to the position information field of the mode unit to complete the bidirectional association; when the query does not hit, the controlled expansion process is triggered: according to the missing type, new entries are expanded in the row dimension or the column dimension, the row header or the column header is updated, and the new position index is generated, and then the new coordinate empty unit record is materialized in the frame matrix, and then the writing process is returned to complete the placement of the mode unit, so that any valid process and sequence combination can obtain a legal main slot in the matrix. At the same time of completing the unit placement, the structured record of the arrangement relationship is constructed: for the adjacent column coordinates in the same row, the predecessor column index and the successor column index are established to form the row-in-link list which is monotonously increasing according to the sequence identifier; for the adjacent row coordinates in the same column, the upper neighbor row index and the lower neighbor row index are established to reflect the vertical corresponding relationship of different processes under the same sequence number; and a global arrangement relationship table is maintained to record the four-way adjacent pointers of each occupied position index, the checksum snapshot of the row header and the column header, and the transaction number written at one time. If the target coordinate already exists a main value, the occupation detection will enter the conflict processing branch: according to the same coordinate single main value constraint strategy, the earliest written or the mode unit meeting the preset priority is retained as the main value, and the newly arrived mode unit is registered as the slave value of the coordinate and stored in the subsidiary list and the conflict reason and comparison summary are recorded. As the sequential sequence is mapped one by one, the matrix mapping unit updates the effective column span in the column header and the effective row span in the row header after each write, which is used to quickly define the boundary of the filled area; when a row or a column changes from all empty to non-empty for the first time, the activation mark is triggered, and the row or the column is included in the active dimension set; after this batch mapping is completed, the visible state of the current frame matrix is frozen, and the frame matrix, the matrix index table, the arrangement relationship table and the transaction log are packaged together as the time sequence matrix data output.

[0112] In a preferred embodiment of the present application, the factor module comprises;

[0113] The time sequence accumulation factor unit is configured to calculate a position attenuation weight term according to the position index in the time sequence matrix data; calculate a row-in prefix accumulation amount according to the values of the units from the first column to the previous column in the same row, to obtain a row-in accumulation term; and calculate an accumulation attention amount according to the position attenuation weight term, the row-in accumulation term and the value of the current unit, to obtain an accumulation attention term.

[0114] According to the accumulated attention items of each column and the corresponding position attenuation weight items, the overall coverage of the row is calculated to obtain a logarithmic sum item; according to the cell values of all columns of the row, a logarithmic product channel quantity is calculated to obtain a logarithmic product item; and according to the logarithmic sum item and the logarithmic product item, the multiplicative coupling strength of the two is calculated to obtain a row-level comprehensive item;

[0115] According to the accumulated attention items of all columns of the row, an overlap suppression item is calculated; according to the row-level comprehensive items of each row, a weighted convergence quantity is calculated to obtain a weighted convergence item; and the overlap suppression item and the weighted convergence item are fused to obtain a timing accumulation factor.

[0116] In the embodiment of the present application, the timing accumulation factor unit is used to calculate the position attenuation weight item according to the position index in the timing matrix data, identify the influence difference of different processes in the production process, and balance the weight of the influence of different processes on the entire production line; calculate the intra-row prefix accumulation quantity according to the cell values from the first column to the current column before the current column in the same row to obtain an intra-row accumulation item, capture the dependency relationship between different processes, accurately reflect the cumulative effect of the previous process, and avoid ignoring the potential influence of historical data on the current process; calculate the accumulated attention quantity according to the position attenuation weight item, the intra-row accumulation item and the current cell value to obtain the accumulated attention item, comprehensively evaluate the contribution of each cell to the current production process, and efficiently identify the key nodes in the timing data; calculate the overall coverage of the row according to the accumulated attention items of each column and the corresponding position attenuation weight items to obtain a logarithmic sum item, quantify the contribution of each process of the production line to the overall production, and identify the process or time period that has a greater impact on the production process; calculate the logarithmic product channel quantity according to the cell values of all columns of the row to obtain a logarithmic product item, which is helpful for the system to process data with extreme values and greater volatility, so that the model can run more stably; calculate the multiplicative coupling strength of the logarithmic sum item and the logarithmic product item to obtain a row-level comprehensive item, accurately capture the nonlinear interaction and coupling relationship in the production process, quantify the mutual influence between different processes, and reveal the potential complex process interaction effect; calculate the overlap suppression item according to the accumulated attention items of all columns of the row, eliminate the redundant information that may exist in the timing data, and effectively avoid the calculation deviation caused by data redundancy; calculate the weighted convergence quantity according to the row-level comprehensive items of each row to obtain a weighted convergence item, enhance the system's evaluation of the influence of each row, i.e., each process, and ensure that each process has a high priority in the comprehensive evaluation; and fuse the overlap suppression item and the weighted convergence item to obtain the timing accumulation factor, which comprehensively reflects the relative importance of the process and the suppression of the redundant influence in the production process.

[0117] In a preferred embodiment of the present application, the comparison module comprises:

[0118] The result alignment unit is configured to construct an index relationship between the dynamic consistency coefficient and the timing accumulation factor, correspondingly arrange the virtual simulation data and the real data in the same reference dimension to obtain an aligned data set.

[0119] a difference detection unit configured to calculate deviation values of the virtual simulation result and the real device data in each corresponding dimension according to the aligned data set, and obtain a difference distribution map;

[0120] a unit extraction unit configured to identify a region with a deviation value deviation exceeding a preset deviation threshold according to the difference distribution map, mark the region as a difference unit, and obtain a difference unit set;

[0121] a parameter correction unit configured to adjust the simulation parameters according to the amplitude feature, the time sequence feature and the position index of the difference unit set, and obtain corrected parameter data.

[0122] In the embodiment of the present application, the result alignment unit is configured to construct an index relationship of the dynamic consistency coefficient and the time sequence cumulative factor, arrange the virtual simulation data and the real data in a corresponding manner in the same reference dimension, and obtain an aligned data set, so as to ensure effective comparison of the two groups of data in the same reference framework and eliminate potential errors caused by inconsistent data sources; the difference detection unit is configured to calculate deviation values of the virtual simulation result and the real device data in each corresponding dimension according to the aligned data set, and obtain a difference distribution map, so as to automatically identify the deviation between the virtual simulation result and the real device data and avoid errors caused by manual intervention; the unit extraction unit is configured to identify a region with a deviation value deviation exceeding a preset deviation threshold according to the difference distribution map, mark the region as a difference unit, and obtain a difference unit set, so as to accurately find out the main deviation region affecting the production line operation and reduce redundant irrelevant data; and the parameter correction unit is configured to adjust the simulation parameters according to the amplitude feature, the time sequence feature and the position index of the difference unit set, and obtain corrected parameter data, so as to automatically optimize the virtual simulation model and ensure that the production line state reflected by the model is more consistent with the actual situation.

[0123] The parameter correction unit is configured to adjust the simulation parameters according to the amplitude feature, the time sequence feature and the position index of the difference unit set, and obtain corrected parameter data, and specifically includes the following steps:

[0124] First, the difference unit set is received, and each difference unit contains deviation information between the virtual simulation data and the real device data, which includes not only the numerical value of the deviation, but also the specific process, time point and position index where the deviation occurs.

[0125] In terms of amplitude characteristics, the correction is made according to the size of the deviation of the difference unit. For example, for a certain process, there is a large difference between the conveying speed or processing speed set in the virtual simulation and the speed of the real device. At this time, the adjustment is made according to the amplitude characteristics of the difference. If the deviation is large, the related parameters such as speed and pressure in the simulation model are corrected by a large amplitude. In this way, the system can better adjust the parameters of the virtual model to make them more consistent with the data of the real device, thereby avoiding large errors in the virtual simulation results.

[0126] In terms of timing characteristics, the time dimension characteristics in the difference unit are considered. For example, assuming that the actual execution time of a certain process deviates from the preset time in the simulation, the timing characteristics are analyzed to evaluate the difference between the time span of the process in the production line and the actual operation. If it is found that the process rhythm in the simulation is out of sync with the actual production rhythm, the time-related parameters are adjusted. This adjustment is not limited to the correction of parameters at a certain moment, but may affect the time sequence of multiple related processes in the entire production process, ensuring that the simulation model can be synchronized with the data flow of the real device.

[0127] According to the position index, the adjustment is made to ensure that the parameter correction of each process matches its physical position in the production line. Different positions of processes may be affected by different external factors, such as the transmission speed of materials, the working state of devices, etc., so the correction of each process needs to be handled separately. In the actual production environment, the position of a process may affect its coordination with other processes. According to the actual position of each process, the adjustment is made to ensure that the parameter correction between different processes can maintain coordination in the operation of the overall production line.

[0128] In a preferred embodiment of the present application, the difference detection unit comprises:

[0129] The deviation extraction unit is configured to calculate the difference between the virtual simulation result and the real device data in each dimension according to the aligned data set, to obtain an original deviation sequence;

[0130] The deviation normalization unit is configured to scale the deviation values of each dimension according to the original deviation sequence, to eliminate the dimensional difference, and to obtain a normalized deviation sequence;

[0131] The feature weighting unit is configured to amplify the deviation values of the amplitude dimension by a dynamic consistency coefficient, and to correct the deviation values of the time dimension by a timing accumulation factor, to obtain a weighted deviation sequence;

[0132] The timing fusion unit is configured to calculate the continuity and accumulation degree of the deviation values in the time dimension according to the weighted deviation sequence, to obtain a fused deviation sequence;

[0133] The distribution generation unit is configured to map the fusion deviation values of each dimension and the position index to a spatial coordinate system according to the fusion deviation sequence to obtain a difference distribution map.

[0134] In the embodiment of the present application, the deviation extraction unit is configured to calculate the difference between the virtual simulation result and the real device data in each dimension one by one according to the aligned data set to obtain an original deviation sequence, accurately identify the deviation between the virtual simulation result and the actual production process, and provide data support for the subsequent; the deviation normalization unit is configured to scale the deviation values of each dimension according to the original deviation sequence, eliminate the dimensional difference, obtain a normalized deviation sequence, so that the deviation values of different dimensions can be compared in the same dimension, and avoid analysis deviation caused by dimensional difference; the feature weighting unit is configured to amplify the deviation values of the amplitude dimension by a dynamic consistency coefficient, correct the deviation values of the time dimension by a time sequence accumulation factor, obtain a weighted deviation sequence, and give different importance to different types of deviation, effectively focusing on the deviation that has greater impact on the production process; the time sequence fusion unit is configured to calculate the continuity and accumulation degree of the deviation values in the time dimension according to the weighted deviation sequence to obtain a fusion deviation sequence, effectively capture the deviation fluctuation characteristics in the time dimension, and identify the deviation accumulation effect generated in a short time; and the distribution generation unit is configured to map the fusion deviation values of each dimension and the position index to a spatial coordinate system according to the fusion deviation sequence to obtain a difference distribution map, clearly present the error distribution between different processes and different parameters in the production process, and quickly locate the problem area.

[0135] The feature weighting unit is configured to amplify the deviation values of the amplitude dimension by a dynamic consistency coefficient and correct the deviation values of the time dimension by a time sequence accumulation factor to obtain a weighted deviation sequence, and specifically includes:

[0136] First, when processing the deviation of the amplitude dimension, the dynamic consistency coefficient is used to amplify the amplitude deviation value. The dynamic consistency coefficient is a coefficient dynamically calculated according to the actual situation of each process in the production line, which reflects the relative importance between different processes. For example, if the deviation of a process has a significant impact on the quality of the final product in a production link, the dynamic consistency coefficient will give greater weight to the deviation of this process. Specifically, the system will multiply the amplitude deviation of the process by the dynamic consistency coefficient, thereby amplifying the amplitude deviation.

[0137] Secondly, when dealing with the deviation in the time dimension, the deviation value is corrected by using a time accumulation factor. The time accumulation factor is a factor calculated based on historical data, aiming to capture the cumulative effect of deviation in the time dimension. Specifically, the system calculates the change rate of deviation in the time dimension according to the deviation difference between the previous time point and the current time point. When a certain process continuously produces deviation in the time dimension, the time accumulation factor will increase the correction strength of the deviation. In this way, the system can effectively identify and adjust the deviation that continues or accumulates in time, avoiding the continuous expansion of these deviations in the production line, leading to greater problems.

[0138] The feature weighting unit, the time fusion unit, are configured to calculate the continuity and accumulation degree of the deviation value in the time dimension according to the weighted deviation sequence, obtain a fusion deviation sequence, and specifically include:

[0139] The weighted deviation sequence contains deviation data at each time point, which has been corrected according to the weighting coefficients in the amplitude dimension and the time dimension. The system first checks the relationship between the deviation value at each time point in the weighted deviation sequence and the deviation value at the previous time point. By comparing the deviations at consecutive time points, the system can calculate the change trend and rate of the deviation value in the time dimension. If the deviation of a certain process continuously increases or decreases over a period of time, the system will identify this trend and adjust the correction strength of the deviation according to the time characteristics. For processes with relatively stable deviation values, the system will give less correction to avoid over-adjustment. Next, the cumulative effect of the deviation is calculated, which refers to how the deviation gradually accumulates over multiple consecutive time points to form a larger deviation value. For example, a process may have a small deviation for a short period of time, but due to the continuous deviation for a long time, it may accumulate into a larger problem, ultimately affecting the normal operation of the production line. By calculating the cumulative value of the deviation and combining it with the time factor, a fusion deviation sequence is generated, which better reflects the overall impact of the deviation in the time dimension.

[0140] In a preferred embodiment of the present application, the verification module comprises:

[0141] The model updating unit is configured to write the correction parameter data into the corresponding structural element of the virtual model and form a new model structure in a parameterized manner to obtain an updated model.

[0142] The structure mapping unit is configured to establish a corresponding relationship between the structural elements of the virtual model and the functional modules of the real device according to the updated model, and generate a structure mapping table.

[0143] An interaction execution unit is configured to interact according to the structure mapping table, establish a data channel between the virtual model and the real device, realize real-time exchange of input and output data, and obtain an interaction data set.

[0144] A verification result unit is configured to analyze the matching degree of the virtual model output and the real device data according to the interaction data set, and record the matching degree in real time to obtain a verification result.

[0145] In the embodiment of the present application, the model updating unit is configured to write the correction parameter data into corresponding structural elements of the virtual model, form a new model structure in a parameterized manner, obtain an updated model, and ensure that the virtual simulation model and the actual production environment are dynamically consistent; the structure mapping unit is configured to establish a corresponding relationship between the structural elements of the virtual model and the functional modules of the real device according to the updated model, generate a structure mapping table, and ensure accurate docking between the virtual simulation and the actual production device; the interaction execution unit is configured to interact according to the structure mapping table, establish a data channel between the virtual model and the real device, realize real-time exchange of input and output data, and obtain an interaction data set, so as to ensure real-time and accurate data exchange between the two; and the verification result unit is configured to analyze the matching degree of the virtual model output and the real device data according to the interaction data set, record the matching degree in real time, and obtain a verification result, so as to timely find the deviation between the two and provide an important basis for production line optimization.

[0146] The model updating unit is configured to write the correction parameter data into corresponding structural elements of the virtual model, form a new model structure in a parameterized manner, obtain an updated model, and specifically includes:

[0147] First, the correction parameter data is received, and the correction data represents the deviation between the virtual simulation model and the actual device. The correction parameters can be adjustment values related to process time, speed, material flow, device performance, etc. The data is cleaned and preprocessed to ensure that all correction parameters meet the requirements of actual application and are consistent with the parameters of each structural element in the virtual model.

[0148] Next, according to these modified parameters, the relevant structural elements in the virtual model are adjusted one by one. These structural elements can include each workstation in the production line, material transfer system, equipment configuration, etc. The adjustment method can be to directly modify the parameter values of the corresponding modules in the model, such as updating the running rate of the equipment or adjusting the process time of a certain link of the production line. The adjustment process is parameterized, that is, the modified parameters are mapped to the corresponding model components through a specific parameterization interface or model updating rule, ensuring that all changes to the model have a unified mathematical expression and logical relationship. Once these modified parameters are successfully written into the virtual model, the structure of the virtual model will change, and at this time the system will generate an updated model according to the new structure.

[0149] The verification result unit is configured to analyze the matching degree of the virtual model output and the real device data according to the interaction data set, and record the matching degree in real time to obtain a verification result, specifically including:

[0150] First, the interaction data set is received, which contains the bidirectional data exchange content between the virtual model and the real device, including the output data of the virtual model such as operation instructions, production results, etc., and the feedback data of the actual device such as device state, output, etc.

[0151] After receiving the interaction data set, data comparison will begin. First, the output data of the virtual model and the feedback data of the real device are matched to check their consistency in the time dimension. The system analyzes the output difference between the virtual model and the actual device at each time point, compares the production efficiency, process progress, device state and other key indicators of the two, and calculates a difference value to reflect the error degree between the virtual simulation and the actual production. Next, according to the preset deviation threshold, it is judged whether the difference exceeds the acceptable range. If the difference exceeds the predetermined allowable error range, the system will automatically mark the area as abnormal and record the specific time, component or process where the difference occurs. The analysis process of the verification result unit is real-time, which can continuously update the data and calculation results during production. Whenever a new interaction data set is received and processed, the system will immediately update the verification result and generate a real-time verification report. The report will show the matching degree of the virtual simulation model and the actual device data, the difference area and possible improvement suggestions.

[0152] In a preferred embodiment of the present application, the interaction execution unit comprises:

[0153] The data channel establishment unit is configured to connect the input and output modules of the virtual model with the data interfaces of the real device according to the structure mapping table to form a bidirectional data channel and obtain a data channel.

[0154] a data transmission unit configured to transmit input data of the virtual model to the real device through the data channel and transmit feedback data of the real device back to the virtual model to obtain preliminary interaction data;

[0155] a data synchronization unit configured to perform time alignment and sequence coordination on input and output between the virtual model and the real device according to the preliminary interaction data to obtain synchronized interaction data;

[0156] a quality inspection unit configured to verify validity of the synchronized interaction data to determine whether the data has packet loss, repetition or error, and obtain qualified interaction data set;

[0157] an interaction data summarization unit configured to classify, integrate and archive all data generated in the interaction process according to the qualified interaction data set to obtain an interaction data set.

[0158] In the embodiment of the present application, the data channel establishment unit is configured to connect input and output modules of the virtual model with data interfaces of the real device according to the structure mapping table to form a bidirectional data channel, thereby obtaining the data channel, ensuring deep integration of virtual simulation and actual production system, eliminating the gap between the virtual model and the actual device, and providing necessary conditions for subsequent data synchronization, transmission and verification; the data transmission unit is configured to transmit input data of the virtual model to the real device through the data channel and transmit feedback data of the real device back to the virtual model to obtain preliminary interaction data, thereby ensuring bidirectional flow of data between the virtual model and the real device and eliminating time difference between virtual simulation and actual device; the data synchronization unit is configured to perform time alignment and sequence coordination on input and output between the virtual model and the real device according to the preliminary interaction data to obtain synchronized interaction data, thereby ensuring high consistency of data between the virtual model and the real device in the time dimension and avoiding data inconsistency caused by time difference or data sequence error; the quality inspection unit is configured to verify validity of the synchronized interaction data to determine whether the data has packet loss, repetition or error, and obtain qualified interaction data set, thereby ensuring integrity and consistency of data transmitted between the virtual model and the real device, effectively eliminating various potential problems in the communication process, and avoiding mismatch or misoperation between the virtual model and the real device caused by data error; and the interaction data summarization unit is configured to classify, integrate and archive all data generated in the interaction process according to the qualified interaction data set to obtain an interaction data set, thereby providing a comprehensive production line data view through integration of all data, effectively supporting subsequent analysis, optimization and decision-making in the production process, ensuring consistency and traceability of data, and providing reliable basis for production line optimization.

[0159] The data channel establishing unit is configured to connect the input and output modules of the virtual model with the data interfaces of the real device according to the structure mapping table, form a bidirectional data channel, and obtain a data channel, specifically including:

[0160] Firstly, the system identifies the data interface mapping relationship between the virtual model and the real device through the structure mapping table. The structure mapping table contains the correspondence between the input and output modules of the virtual model and the sensors and actuators of the device. Through the analysis of the mapping table, the data channel establishing unit can find the specific data interfaces of each input and output module in the virtual model and the device. Then, the system configures the communication according to the communication protocol of the device, such as TCP / IP, Modbus, Ethernet / IP, etc. The input and output modules of the virtual model establish physical or logical connection with the control system or sensor data interface of the real device through the communication protocol. At this time, the output data of the virtual model can be transmitted to the device control system in real time, and the data collected by the device sensor can also be fed back to the virtual model for simulation update. The direction and mode of data transmission need to be adjusted and optimized according to the real-time production needs. A bidirectional data channel is formed between the virtual model and the real device, ensuring that the virtual model can obtain the device state data and feed back the adjustment scheme to the device.

[0161] The data synchronization unit is configured to time-align and sequence-coordinate the input and output between the virtual model and the real device according to the preliminary interaction data, and obtain synchronized interaction data, specifically including:

[0162] Firstly, the preliminary interaction data between the virtual model and the device is time-stamped. Each input and output data is attached with an accurate time stamp when collected, so that the system can know the time position corresponding to each data point. Then, the system analyzes the data time axis of the virtual model and the real device to identify the time difference between them. Next, the data is adjusted through time sliding window or time difference compensation mechanism. If the system finds that the data timestamp of the virtual model lags behind the collected data of the real device, or vice versa, the system will automatically align the data timestamp, so that the data of the virtual model and the device can correctly correspond in the same time dimension. In addition, in the actual production environment, the data collection frequency of the device may be higher or lower than that of the virtual model, which needs to be adjusted through interpolation or data downsampling method to ensure the consistency of data sequence and the time continuity of data exchange. Through the above operations, the data between the virtual model and the real device can be accurately aligned in the time dimension, avoiding the problem of inaccurate interaction caused by data time disorder or time delay.

[0163] The quality checking unit is configured to verify the validity of the synchronization interaction data, determine whether the data has packet loss, repetition or error, and obtain qualified interaction data sets, and specifically includes:

[0164] Firstly, it is detected whether there is packet loss phenomenon, in the process of bidirectional data transmission between virtual and reality, packet loss may occur due to network instability or transmission error and other problems, therefore, the received data is compared with the data sent by the sender to check whether there is data loss, if it is found that the data is lost, the system will request the device to resend the lost data through the retransmission mechanism, to ensure the integrity of the data. Secondly, it is checked whether there is repetition phenomenon, in the process of high-frequency data interaction, due to transmission delay or retransmission mechanism, etc., repeated data may occur, by comparing with the previous data packet, the repeated part is identified and removed, to avoid the interference of redundant data. Then, the format verification and error checking are performed on each piece of received data, the data format between the virtual model and the device needs to be unified, therefore, it is ensured that all data is transmitted according to the predetermined data structure, any data not conforming to the format is marked as error data. In addition, the legality of the content of the data is checked, to ensure that the value range and the calculation result meet the expected requirements. Through the checking and correction of the packet loss, repetition and error data, the qualified interaction data is effectively screened, to ensure the accuracy of subsequent data processing and decision-making.

[0165] The above is the preferred embodiment of the present application, it should be noted that for those skilled in the art, without departing from the principles of the present application, a number of improvements and refinements can be made, these improvements and refinements should also be considered as the protection scope of the present application.

Claims

1. A hybrid simulation digital twin production line virtual-real interactive verification system, characterized in that, The system includes: The parameter module is used to acquire the operating parameters of the production line process, encode the changing trends between the operating parameter points, and convert the numerical changes into symbolic sequences to obtain symbolic parameter data. The combination module is used to combine symbol sequences into pattern units based on symbolized parameter data, and introduce process identifiers during the combination process to obtain pattern unit data; The coefficient module is used to perform combined calculations on the variation amplitude and duration of each mode unit based on the mode unit data to obtain the dynamic consistency coefficient. The construction module is used to construct a time sequence matrix based on the pattern unit data according to the process sequence, and introduce position indexes into it to record the arrangement relationship of each pattern unit, thereby obtaining the time sequence matrix data; The factor module is used to accumulate time series matrix data item by item and correct it through position index during the accumulation process to obtain the time series accumulation factor; The comparison module is used to compare the virtual simulation results with the real equipment data based on the dynamic consistency coefficient and the time-series accumulation factor, and extract the difference units during the comparison process to obtain the correction parameter data; The verification module is used to perform structured updates on the virtual model based on the corrected parameter data and to interact with the real device data in real time to obtain the verification results. The coefficient module includes: The dynamic consistency coefficient unit is used to calculate the enhancement amount of the amplitude channel based on the mode cell data to obtain the amplitude enhancement term; calculate the enhancement amount of the length channel to obtain the length enhancement term; calculate the suppression amount of the main channel based on the difference in length between adjacent mode cells to obtain the suppression term; fuse the amplitude enhancement term and the length enhancement term, and form a ratio relationship between them and the suppression term to obtain the main channel term; Based on the differences in amplitude and length between adjacent mode units, the coupling deviation is calculated to obtain the coupling residual term; based on the directional abrupt changes of adjacent mode units, the degree of influence on the coupling deviation is calculated to obtain the transition amplification term; the coupling residual term and the transition amplification term are fused to calculate the coupling channel contribution value to obtain the coupling channel term; The main channel term and the coupled channel term are summed one by one to calculate the overall metric and obtain the dynamic consistency coefficient. The factor module includes: The time-series accumulation factor unit is used to calculate the position attenuation weight based on the position index in the time-series matrix data; calculate the intra-row prefix accumulation based on the cell values ​​from the first column to the current column in the same row to obtain the intra-row accumulation term; and calculate the cumulative attention based on the position attenuation weight, the intra-row accumulation term, and the current cell value to obtain the cumulative attention term. Based on the cumulative attention terms and corresponding positional attenuation weights of each column, calculate the overall coverage of the row to obtain the logarithmic sum term; based on the cell values ​​of all columns in the row, calculate the logarithmic product channel quantity to obtain the logarithmic product term; based on the logarithmic sum term and the logarithmic product term, calculate the multiplicative coupling strength between the two to obtain the row-level synthesis term. Calculate the overlap suppression term based on the cumulative attention terms of all columns in the row; calculate the weighted convergence term based on the row-level comprehensive terms of each row; and merge the overlap suppression term and the weighted convergence term to obtain the time-series cumulative factor.

2. The hybrid simulation digital twin production line virtual-real interactive verification system according to claim 1, characterized in that, The parameter module includes: The parameter acquisition unit is used to collect data according to the operation process of the production line, and to record the real-time operating parameters of each process point by point to obtain the operating parameter data. The trend calculation unit is used to calculate the numerical differences between adjacent parameter points based on the set of operating parameters, and to obtain the trend data. The symbol generation unit is used to divide the trend of change into rising, falling, and stable based on the trend data, and obtain a sequence of directional symbols. The encoding conversion unit is used to assign a unique code to each symbol according to the direction symbol sequence and form serialized data to obtain symbolized parameter data.

3. The hybrid simulation digital twin production line virtual-real interactive verification system according to claim 2, characterized in that, The combined module includes: The sequence segmentation unit is used to segment the symbol sequence according to the symbolization parameter data, dividing the symbol sequence into several symbol segments according to the changes of adjacent symbols, and obtaining the segment sequence; The feature aggregation unit is used to aggregate fragment sequences, merging symbol fragments according to the direction of symbol change and duration to obtain an aggregated fragment set; The process identification unit is used to bind each aggregated fragment to the corresponding process identification according to the aggregated fragment set, so as to obtain the identified fragment set; The unit combination unit is used to combine the symbolic features of the segments with the process identifiers to form a structured record based on the set of identified segments, thus obtaining pattern unit data.

4. The hybrid simulation digital twin production line virtual-real interactive verification system according to claim 3, characterized in that, The construction module includes: The sequence arrangement unit is used to arrange each pattern unit according to the order of the processes based on the pattern unit data, and to add a serial number identifier to each pattern unit to obtain the sequence sequence. The matrix frame unit is used to establish a matrix frame for storing pattern units by dividing rows according to process category and columns according to serial number identifier according to the sequential sequence, thus obtaining the frame matrix; The location index unit is used to generate a unique index number for each pattern unit position based on the frame matrix, and to record the correspondence between the index and the process category and sequence number identifier to obtain a matrix index table; The matrix mapping unit is used to place the pattern units in the sequential sequence into the corresponding positions in the frame matrix one by one according to the matrix index table, and record their arrangement relationship in combination with the index table to obtain the time series matrix data.

5. The hybrid simulation digital twin production line virtual-real interactive verification system according to claim 4, characterized in that, The comparison module includes: The result alignment unit is used to construct the index relationship between the dynamic consistency coefficient and the time-series cumulative factor, and to arrange the virtual simulation data and the real data in the same reference dimension to obtain the aligned dataset. The difference detection unit is used to calculate the deviation values ​​between the virtual simulation results and the real device data in each corresponding dimension based on the aligned dataset, and to obtain the difference distribution map; The unit extraction unit is used to identify areas where the deviation value exceeds a preset deviation threshold based on the difference distribution map, and mark them as difference units to obtain a set of difference units; The parameter correction unit is used to adjust the simulation parameters based on the amplitude characteristics, timing characteristics, and location index of the difference unit set to obtain corrected parameter data.

6. The hybrid simulation digital twin production line virtual-real interactive verification system according to claim 5, characterized in that, The difference detection unit includes: The deviation extraction unit is used to calculate the difference between the virtual simulation result and the real device data in each dimension based on the aligned dataset, and obtain the original deviation sequence. The deviation normalization unit is used to scale the deviation values ​​of each dimension according to the original deviation sequence, eliminate the difference in dimensions, and obtain a normalized deviation sequence. The feature weighting unit is used to amplify the deviation value in the amplitude dimension through a dynamic consistency coefficient and to correct the deviation value in the time dimension through a time-series accumulation factor, thereby obtaining a weighted deviation sequence. The temporal fusion unit is used to calculate the continuity and accumulation of the deviation values ​​in the time dimension based on the weighted deviation sequence, and obtain the fused deviation sequence. The distribution generation unit is used to map the fusion deviation values ​​and position indices of each dimension to a spatial coordinate system based on the fusion deviation sequence, thereby obtaining a difference distribution map.

7. The hybrid simulation digital twin production line virtual-real interactive verification system according to claim 6, characterized in that, The verification module includes: The model update unit is used to write the corrected parameter data into the corresponding structural elements of the virtual model and form a new model structure in a parameterized manner to obtain the updated model. The structure mapping unit is used to establish a correspondence between the structural elements of the virtual model and the functional modules of the real device based on the updated model, and generate a structure mapping table. The interactive execution unit is used to interact according to the structure mapping table, establish a data channel between the virtual model and the real device, realize the real-time exchange of input and output data, and obtain the interactive dataset. The verification result unit is used to analyze the degree of matching between the virtual model output and the real device data based on the interactive dataset, and to record the results in real time to obtain the verification results.

8. The hybrid simulation digital twin production line virtual-real interactive verification system according to claim 7, characterized in that, The interactive execution unit includes: The data channel establishment unit is used to connect the input and output modules of the virtual model with the data interface of the real device according to the structure mapping table to form a bidirectional data channel. The data transmission unit is used to send the input data of the virtual model to the real device through the data channel, and to transmit the feedback data of the real device back to the virtual model to obtain preliminary interactive data. The data synchronization unit is used to perform time alignment and sequence coordination of the input and output between the virtual model and the real device based on the preliminary interaction data to obtain synchronized interaction data; The quality inspection unit is used to verify the validity of the synchronous interactive data, determine whether there is packet loss, duplication or error in the data, and obtain a qualified interactive dataset. The interactive data aggregation unit is used to classify, integrate, and archive all data generated during the interaction process based on the qualified interactive dataset, thus obtaining the interactive dataset.

Images