An automobile engine body production line digital twin monitoring system

Abstract

The application relates to the field of intelligent production and industrial digital monitoring of automobile manufacturing, in particular to a digital twin monitoring system for an automobile engine body production line, which comprises the following: a data receiving module, which is used for receiving data of numerical control equipment, a programmable logic controller, a conveying controller, a workpiece code reading device and an online measuring device, and generating unified event records; a time determining module, which is used for determining event absolute time, extracting key nodes and generating event effective time intervals; a data correlation module, which is used for correlating workpieces, pallets and stations; a state updating module, which is used for performing current updating, pending correction or historical supplement according to state confirmation time points; and a result output module, which is used for driving digital twin model synchronous updating; the application can avoid disorder and wrong judgment, and solve the problem of persistent state drift.

Description

Technical Field

[0001] This invention relates to the field of intelligent production and industrial digital monitoring in automobile manufacturing, specifically a digital twin monitoring system for an automobile engine production line. Background Technology

[0002] With the development of intelligent manufacturing and digital twin technology, online monitoring and synchronous control of automobile engine production lines has become an important research direction in the field of discrete manufacturing. In order to achieve accurate perception of the production line's operating status and real-time mapping of digital models, unified access, time-series analysis, and state reconstruction of multi-source production data have become particularly important. In the digital twin monitoring process of an automotive engine production line, there are usually multiple sources of equipment on site, such as CNC equipment, programmable logic controllers, conveyor controllers, workpiece coding and reading devices, and online measuring devices. Different devices have heterogeneous parameters and inconsistent cycle times in terms of control cycle, communication method, and data reporting time. This can easily lead to problems such as out-of-order event arrival, inconsistent time bases, and confused state associations. Existing solutions often rely more on data reception time or a single timestamp for state updates. It is difficult to accurately determine the correspondence between workpieces, pallets, and workstations under conditions such as network jitter, breakpoint retransmission, rework backflow, and delayed arrival of measurement results, and to maintain the consistency between the digital twin model and the physical production line status. This affects the reliability of production process monitoring, quality traceability, and anomaly detection. Therefore, it is crucial to ensure the safe, efficient and precise operation of the automotive engine production line by uniformly aligning and effectively constraining production events from different devices, and by combining workpiece flow relationships, process rules and historical state boundaries to complete accurate correlation and state correction, so as to obtain monitoring results that can truly reflect the physical state of the production line and facilitate the stable and synchronous updating of the digital twin model. Summary of the Invention

[0003] The purpose of this invention is to provide a digital twin monitoring system for an automotive engine production line, addressing the following technical problems: Existing digital twin monitoring technologies suffer from time reference deviations and state association failures when facing out-of-order arrival of multi-source data and in complex operating conditions. There is an urgent need for a digital twin monitoring system for an automotive engine production line that can more accurately handle multi-source event time constraints and dynamically correct the physical state of the production line. To achieve the above objective, the technical solution adopted by this invention is: A digital twin monitoring system for an automotive engine production line is applied to a production line containing multiple workstations and conveyor lines. The system includes a data receiving module, a time determination module, a data association module, a status update module, and a result output module. The data receiving module is used to communicate with and receive data from at least one source device, which includes at least a CNC machine, a programmable logic controller, a conveyor controller, a workpiece code reading device, and an online measuring device. For source equipment with periodic control parameters, its reference time, control cycle, cycle number, and time offset ratio within the cycle are parsed; for source equipment without periodic control parameters, its event time identifier is received, and a unified event record is generated; the time determination module is used to determine the absolute time of the event based on the unified event record, extract key nodes representing changes in the status of the workpiece, pallet, or workstation based on the event type and event parameters in the unified event record, and generate the effective time interval of the event by combining the preset equipment communication delay, the node allowable time range set relative to the planned cycle time, and the process allowable time range. The data association module is used to maintain the correspondence between workpieces, pallets, and workstations, using workpieces as monitoring objects. It filters workpieces to be matched based on the effective time interval of the event and the corresponding workstation, and calculates a comprehensive matching value based on route consistency, workstation occupancy consistency, process consistency, coding consistency, and time difference, thus associating the event with the corresponding monitoring object. The status update module is used to maintain the status confirmation time point of each monitoring object, and performs current digital twin status updates, corrects pending status data in the processing queue, or rewrites historical status records based on the chronological relationship between the effective time interval of the event and the status confirmation time point. The result output module is used to output the monitoring information of the monitoring objects and drive the digital twin model of the production line to update synchronously based on the updated status data.

[0004] In one alternative: the unified event record includes source device identifier, event type, event parameters, receiving time, cycle number, time offset ratio within the cycle, pallet identifier, workstation identifier, and process identifier; the time determination module determines the absolute time of the event based on the reference time, control cycle, cycle number, and time offset ratio within the cycle.

[0005] In one alternative approach: the effective time interval of the event is determined based on the absolute time of the event, the tolerance time of the source device, the allowable time range of the node, and the allowable time range of the process; when the lower limit of the effective time interval of the event is greater than the upper limit, the corresponding event is placed in the processing queue, waiting for subsequent key nodes to recalculate, and is not used to change the current digital twin state.

[0006] In one alternative: the key nodes include at least one of pallet arrival at the workstation, completion of workpiece code reading, completion of fixture clamping, workpiece departure and release, and completion of online measurement result archiving; each key node is defined by a logical production line segment, a physical workstation, and a preset coordinate range; the time determination module is also used to configure a preset process flow rule table for each workstation, and only when the current process to the target process corresponding to the event belongs to the flow relationship allowed by the process flow rule table is the event allowed to update the process status of the corresponding monitored object.

[0007] In one alternative: the data association module calculates a comprehensive matching value for the selected workpieces to be matched, and the comprehensive matching value is determined based on each consistency, time difference and their corresponding weights; wherein, the time difference is determined based on the difference between the absolute time of the event and the expected arrival time of the workpiece to be matched, and the expected arrival time of the workpiece to be matched is determined based on the workpiece's historical flow information and the production line's operating status; When the number of workpieces to be matched is one and the overall matching value is not lower than the preset minimum matching threshold, or when the number of workpieces to be matched is greater than one, the highest overall matching value is not lower than the preset minimum matching threshold, and the difference between the highest overall matching value and the second highest overall matching value is not lower than the preset matching difference threshold, the event will be associated with the workpiece corresponding to the highest overall matching value; otherwise, the event will be placed in the pending queue, waiting for subsequent key nodes to recalculate, and will not be used to change the current digital twin state.

[0008] In one alternative: the status update module determines the status confirmation time point according to the upper limit of the valid time interval of the confirmed key node; when the upper limit of the valid time interval of the event is earlier than the status confirmation time point, the event is treated as a historical event, and only the historical status record is updated. When the lower limit of the effective time interval of an event is earlier than and the upper limit is later than the status confirmation time, the event is treated as a cross-period event, the location status and workstation occupancy status in the current digital twin status are retained, and the process data in the historical status record is corrected; when the lower limit of the effective time interval of an event is later than or equal to the status confirmation time, the event is treated as the current event and used to perform the current digital twin status update.

[0009] In one alternative: the data association module is further configured to determine whether the return belongs to an allowed flow path according to a preset rework flow rule table when the workpiece returns from the current workstation to the already processed rework target workstation; when it belongs to an allowed flow path, generate a processing round identifier containing the workpiece code and rework number identifier, and re-determine the expected arrival time of the workpiece entering the rework workstation according to the most recent rework release time, rework path length and rework workstation queuing status.

[0010] In one alternative: for online measurement events generated by the online measurement device, the data association module determines the measurement matching value based on the workpiece code, processing path, measurement time, and the degree of correspondence between the measurement result and the processing parameters; the degree of correspondence between the measurement result and the processing parameters is determined based on the correspondence between the processing parameter sequence of the upstream station of the workpiece to be matched and the historical measurement sequence of the same workpiece. Measurement time consistency is determined by the time interval between the absolute time of the online measurement event and the expected arrival time of the workpiece to be matched at the online measurement device. The online measurement event with the highest measurement matching value that meets the preset measurement matching threshold is associated with the corresponding workpiece and the corresponding processing round identifier.

[0011] In one alternative solution: it also includes a parameter adjustment module, which is used to adjust the advance tolerance time and lag tolerance time according to the statistical results of the event time deviation of the source equipment; set the historical statistical period as a preset time window; when the misjudgment rate confirmed by manual review within the historical statistical period is higher than the first threshold, the preset minimum matching threshold is increased by a preset step size; and when the proportion of reworked workpieces within the historical statistical period is higher than the second threshold, the preset matching difference threshold is increased by a preset step size.

[0012] Compared with the prior art, the present invention has the following advantages: 1. This system determines the absolute time of an event by parsing parameters such as the control cycle of the source equipment or the event time identifier, and generates an effective time interval for the event by combining preset communication delay and the allowable time range of the process. This solution effectively solves the problem of out-of-order arrival of data and inconsistent time base caused by the difference in control cycles of multiple source equipment and communication delay, ensuring unified time constraints for production line events and improving the reliability of the digital twin model in perceiving the operating status. 2. This system calculates a comprehensive matching value based on route and process consistency and time difference, and performs current status updates, pending queue corrections or historical record rewriting according to the relationship between the effective time interval and the status confirmation time point. This mechanism overcomes the problem of chaotic status association caused by network jitter, rework backflow and measurement lag, and maintains the consistency between the digital twin model and the physical production line while avoiding the error of historical late data overwriting the current real status. Attached Figure Description

[0013] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0014] Figure 1 This is a block diagram of a digital twin monitoring system for an automobile engine production line according to this application. Detailed Implementation

[0015] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings.

[0016] Reference Figure 1 One embodiment of this application provides a digital twin monitoring system for an automotive engine production line, applied to a production line containing multiple workstations and conveyor lines, including: The data receiving module communicates with and receives data from source devices that include at least CNC equipment, programmable logic controllers, conveyor controllers, workpiece code reading and online measurement devices; it parses the reference time, control cycle, cycle number and time offset ratio within the cycle of source devices with cycle control parameters; it receives event time identifiers from source devices without cycle control parameters; and it generates a unified event record. The time determination module determines the absolute time of events based on the unified event record, extracts key nodes representing changes in the status of workpieces, pallets, or workstations based on the event type and event parameters of the unified event record, and generates the effective time interval of events by combining the node allowable time range and process allowable time range set by the preset communication delay and relative planned cycle time. The data association module takes the workpiece as the monitoring object, maintains the correspondence between the workpiece and the pallet and workstation, filters the workpiece to be matched according to the effective time interval of the event and the corresponding workstation, calculates the comprehensive matching value according to the consistency of route, workstation occupancy, process, code and time difference, and associates the event with the corresponding monitoring object. The status update module maintains the status confirmation time point of each monitored object. Based on the relationship between the effective time interval of the event and the status confirmation time point, it performs the current digital twin status update, the correction of pending status data in the pending queue, or the rewriting of historical status records. The results output module outputs the monitoring information of the monitored objects and drives the production line digital twin model to update synchronously based on the updated status data.

[0017] The unified event log includes source device identifier, event type, event parameters, reception time, cycle number, time offset ratio within the cycle, pallet identifier, workstation identifier, and process identifier; the time determination module determines the absolute time of the event based on the reference time, control cycle, cycle number, and time offset ratio within the cycle.

[0018] The effective time interval of an event is determined based on the absolute time of the event, the tolerance time of the source device, the allowable time range of the node, and the allowable time range of the process. When the lower limit of the effective time interval of an event is greater than the upper limit, the corresponding event is placed in the pending queue, waiting for subsequent key nodes to recalculate, and is not used to change the current digital twin state.

[0019] Key nodes include at least one of the following: pallet arrival at the workstation, completion of workpiece code reading, completion of fixture clamping, workpiece departure and release, and completion of online measurement result archiving. Each key node is jointly defined by the logical production line segment representing the process flow stage, the physical workstation of the actual equipment, and the preset coordinate interval used to limit the spatial range. The time determination module is also used to configure a preset process flow rule table for each workstation. Only when the current process corresponding to the event and the target process belong to the flow relationship allowed by the process flow rule table, the event is allowed to update the process status of the corresponding monitored object.

[0020] This embodiment provides a digital twin monitoring system for an automotive engine production line, which solves the problem of out-of-order data arrival caused by multi-source heterogeneous equipment under different control cycles and different communication delays, and achieves accurate synchronization of the physical state of the production line. Acquire production data output from the source device, parse cycle parameters or received event time identifiers according to the device type, encapsulate the received data in a unified manner, and generate a unified event record; the unified event record is a standard data structure, including a dataset containing source device identifier, event type, event parameters, received time, cycle number, time offset ratio within the cycle, pallet identifier, workstation identifier, and process identifier; For source equipment with periodic control parameters, such as CNC equipment, the system extracts the internal operating reference time, control cycle, cycle number, and time offset ratio within the cycle. For asynchronous equipment without periodic control parameters, such as online measuring devices, the system directly extracts the event time identifier generated when it is triggered. For this type of asynchronous equipment, the event time identifier adopts a time value with the same dimensions as the monitoring system's unified time axis. If the equipment reports a relative timing value, it is first converted to an absolute time on the unified time axis based on the corresponding time reference of the equipment before being written into the unified event record. The system establishes a bidirectional alignment mechanism between the local clock and the unified time axis; the system periodically sends out timestamp probe messages and receives acknowledgments, calculates the round-trip delay (i.e., the time of receiving acknowledgment minus the time of sending the probe message), calculates the one-way transmission delay for each communication, and obtains the initial time deviation between the system unified time and the device local time at that moment, thereby establishing a deviation sample set. The system uses the least squares method to perform linear regression on multiple sets of time deviations. Specifically, it uses the system's transmission time of the probe message as the independent variable and the corresponding initial time deviation as the dependent variable to fit a straight line of deviation change. The slope of this line is the local clock drift rate of the asynchronous device, and the intercept is the initial intercept deviation. Based on this time synchronization reference, the system performs dynamic slope compensation within a preset statistical window according to the drift rate and calculates the absolute time. The corresponding formula for calculating the absolute time is: When the statistical window is exceeded or a communication interruption and reconnection occur, the intercept is updated as a whole according to the most recent time handshake result, so as to ensure that the time values ​​of asynchronous events within the same window can be strictly aligned and directly compared. Based on the above embodiments, a unified event record is obtained, the absolute time of the event is determined according to the period parameters of the source device, and the effective time interval of the event is generated in combination with multi-dimensional physical constraints. Specifically, for source devices with period control parameters, the absolute time of the event is determined by the reference time of the device, the current period number, the device control period, and the time offset ratio within the period. This is used to map the independent internal control beats of each device onto the unified time axis of the monitoring system, eliminating the interference of reported beat differences on timing judgment. For source devices without period control parameters, the absolute time of the event is directly adopted using the event time identifier, or the absolute time converted by the aforementioned time reference. The system determines the earliest possible time of an event (lower limit) and the latest possible time of an event (upper limit) to construct the effective time interval of the event; the source equipment tolerance time is used to characterize the time uncertainty between the physical occurrence of the event and its perceptibility by the system. It is determined by the preset equipment communication delay and the statistical results of the equipment's historical time deviation, and is divided into advance tolerance time and lag tolerance time; the node allowable time range and the process allowable time range are first converted into allowable intervals on a unified time axis, and then participate in the subsequent upper and lower bound trimming; Among them, the node allowable time range is used to limit the possible time period of the same critical node near the planned cycle time, and the process allowable time range is used to limit the possible time period of the same workpiece between adjacent processes to satisfy the shortest and longest flow constraints. Both are represented as closed intervals on the time axis; the system finds the intersection through a preset boundary clipping strategy. The system reads the lower and upper limits of device communication latency, and dynamically sets the advance tolerance time and lag tolerance time based on the statistical root mean square error of time deviation. The specific setting rules are as follows: the sum of the upper limit of device communication latency and a preset multiple, such as 3 times the root mean square error of time deviation, is used as the advance tolerance time to cover the latest arrival boundary caused by network fluctuations; the absolute value obtained by subtracting the preset multiple of the root mean square error of time deviation from the lower limit of device communication latency is used as the lag tolerance time. The earliest possible time is determined by subtracting the advance tolerance time from the absolute time of the event to obtain the initial lower limit. When this initial lower limit is less than the lower limit of the allowed time range of the critical node or the lower limit of the allowed time range of the process, the maximum value among the above lower limits is taken for truncation and replacement. The latest possible time is determined by adding the lag tolerance time to the absolute time of the event to obtain the initial upper limit. When this initial upper limit is greater than the upper limit of the allowed range of the node or the upper limit of the allowed range of the process, the minimum value among the upper limits is taken for truncation and replacement. When the truncation replacement causes the lower limit of the effective time interval to be greater than the upper limit, the system puts it into the pending queue to wait for subsequent key nodes to be recalculated, without directly changing the current digital twin state; recalculation means that after the arrival of subsequent new key nodes, the system uses the newly added workstation location, process progress or pallet mapping information to correct the node allowable time range or process allowable time range corresponding to the event, and then determines whether it can be used for state update. Specifically, when a new critical node arrives later, which is the arrival event of the workpiece at the downstream station, the system extracts the lower limit of the time interval of the arrival event and tightens the latest allowed occurrence boundary of the event to be processed to no later than the time after deducting the minimum theoretical flow time between the two stations from the lower limit. If a valid interval cannot be formed after tightening, it will continue to be kept in the queue of events to be processed and will not participate in the current state update. In practical implementation, key nodes representing changes in the status of workpieces, pallets, or workstations are extracted from highly reliable data sources, such as pallet arrival at workstation, workpiece code reading completion, and fixture clamping completion. A preset process flow rule table is configured for each workstation, and the system determines whether the process transfer corresponding to the event is legal. Only when the current process to the target process corresponding to the event belongs to the flow relationship allowed by the rule table, such as a one-way flow from the workstation to the clamping completion and processing, is the event allowed to enter the status update process. If it does not belong to the allowed flow relationship, the event only retains the original record and enters the pending queue, without directly updating the process status of the corresponding monitored object. This embodiment transforms discrete events into effective time intervals with spatiotemporal boundaries through a combination of source period normalization and physical migration constraints; it drives the synchronous update of the digital twin model, ensuring that the model state and the physical production line state remain logically consistent under a unified time axis and state confirmation boundary, without using the message reception time as the sole basis for update. This embodiment generates an effective time interval for events based on a time determination module to synchronize the physical state of the production line; In one embodiment, the data association module calculates a comprehensive matching value for the selected workpieces to be matched. The comprehensive matching value is determined based on route consistency, workstation occupancy consistency, process consistency, coding consistency, time difference, and their corresponding weights. The time difference is determined based on the difference between the absolute time of the event and the expected arrival time of the workpiece to be matched. The expected arrival time of the workpiece to be matched is determined based on the workpiece's historical flow information and the production line's operating status. When the number of workpieces to be matched is 1 and the overall matching value is not lower than the preset minimum matching threshold, or when the number of workpieces to be matched is greater than 1, the highest overall matching value is not lower than the preset minimum matching threshold, and the difference between the highest overall matching value and the second highest overall matching value is not lower than the preset matching difference threshold, the event will be associated with the workpiece corresponding to the highest overall matching value; otherwise, the event will be placed in the pending queue, waiting for subsequent key nodes to recalculate, and will not be used to change the current digital twin state.

[0021] The status update module determines the status confirmation time point according to the upper limit of the valid time interval of the confirmed key nodes; when the upper limit of the valid time interval of the event is earlier than the status confirmation time point, the event is treated as a historical event, and only the historical status record is updated. When the lower limit of the valid time interval of an event is earlier than and the upper limit is later than or equal to the status confirmation time, the event is treated as a cross-period event, the location status and workstation occupancy status in the current digital twin status are retained, and the process data in the historical status record is corrected; when the lower limit of the valid time interval of an event is later than or equal to the status confirmation time, the event is treated as the current event and used to perform the current digital twin status update.

[0022] The data association module is also used to determine whether the return belongs to an allowed flow path according to the preset rework flow rule table when the workpiece returns from the current station to the already processed rework target station. When it belongs to an allowed flow path, it generates a processing round identifier containing the workpiece code and rework number identifier, and re-determines the estimated arrival time of the workpiece entering the rework station according to the most recent rework release time, rework path length and rework station queuing status.

[0023] The system acquires a set of workpieces to be matched, calculates a comprehensive matching value for the selected workpieces based on multidimensional consistency features, and determines the unique association between the event and the workpiece. The comprehensive matching value is a quantitative indicator that measures the degree of fit between the physical trajectory of the current event and the candidate workpiece. The system extracts the route consistency, workstation occupancy consistency, process consistency, code consistency, and time difference of the event. For discrete state features such as route, workstation, process, and code, the system uses binary logic to calculate the consistency score, that is, the score is recorded as 1 when the feature attributes match and 0 when they do not match. The time difference is used to characterize the deviation between the time of the event occurrence and the theoretical time when the workpiece flows at a normal pace. It is determined based on the difference between the absolute time of the event and the expected arrival time of the workpiece to be matched. The system uses the absolute value of the difference as the deviation. The system then linearly normalizes this deviation according to the preset maximum allowable time difference to calculate the time matching score. The calculation formula is as follows: in, Match scores to time. This refers to the time deviation. This is the maximum permissible time difference; the maximum permissible time difference is determined based on the sum of the theoretical cycle time of this section of the production line and the maximum queuing buffer time. If a workstation is designated as a no-queue buffer workstation, the maximum allowable time difference is based on the theoretical cycle time of the corresponding process at that workstation. The consistency scores of the route, workstation, and process are weighted and summed with the time matching score according to preset weights, and the absolute physical constraints are directly multiplied as Boolean veto multipliers to calculate the final comprehensive matching value. The preset weights are pre-configured based on the confidence level of each feature information source. The system assigns the highest weight to the consistency of the hardware-read code, which is in turn greater than the weight of the route, workstation, and process consistency calculated by the location and the weight of the time matching score. To ensure the stability of the weighted results, the sum of all weights is preferably set to 1. If some consistency features are not enabled due to missing data in a certain matching scenario, only the weights corresponding to the enabled features are normalized before calculating the comprehensive matching value. When the number of workpieces to be matched is 1 and the comprehensive matching value is not lower than the preset minimum matching threshold (example value is 0.75), direct association is performed. When there are multiple workpieces to be matched, if the highest comprehensive matching value is not lower than the preset minimum matching threshold, and the difference between the highest score and the second highest score is not lower than the preset matching difference threshold (example value is 0.15), then the event is associated with the workpiece corresponding to the highest score; if the above threshold conditions are not met, the event is placed in the pending queue; waiting for subsequent key nodes to recalculate, so that after a new encoding reading event, arrival event or departure event occurs, the set of workpieces to be matched is reduced again and the comprehensive matching value is recalculated. The associated events are acquired, and the monitoring objects are updated in a hierarchical manner according to the relationship between the effective time interval of the events and the status confirmation time point. The status confirmation time point refers to the irreversible time boundary determined according to the upper limit of the effective time interval of the confirmed key nodes. When there are multiple confirmed key nodes, the upper limit of the effective time interval of the latest confirmed key node can be selected as the current status confirmation time point. The system classifies and judges events: when the upper limit of the effective time interval of an event is earlier than the status confirmation time, it is judged as a historical event. The system only performs historical status record supplementation, such as supplementing historical tool life consumption, and prohibits modification of the current position. When the lower limit of the effective time interval of an event is earlier than and the upper limit is later than the status confirmation time, it is determined to be a cross-period event. The system performs a cross-period separation operation at the database level: retains the location status and workstation occupancy status label in the current in-memory digital twin state to ensure the uninterrupted flow of real-time control, and performs incremental updates on the historical record copy of the corresponding time node in the persistent storage layer. For example, it directly overwrites or appends and corrects non-location business attributes such as processing energy consumption, tool wear, or temperature curve at this historical moment through timestamp index. After the update is successful, the historical confidence label of this time period is refreshed, thereby restoring historical detailed data without destroying the confirmed subsequent physical topology state machine evolution. When the lower limit of the effective time interval of an event is later than or equal to the state confirmation time point, it is determined to be the current event and is directly used to perform the current digital twin state update; for the pending state data in the aforementioned queue, if subsequent key nodes clarify the relationship between the effective time interval of the event and the state confirmation time point, the corresponding correction is performed according to the above classification results of historical events, cross-period events or current events. In practical implementation, for the return of workpieces in the production line, the system obtains the flow information of the rework workpieces, verifies the legality of the path according to the preset rework flow rule table, and generates a processing round identifier to isolate the event. The processing round identifier includes the workpiece code and the rework number identifier, for example, the identifier is workpiece 205 - 1st rework. When the workpiece enters the rework station, the system does not reuse the time base of the original processing stage, but re-deduces the estimated arrival time based on the most recent rework release time, the rework path length, and the queuing status of the rework station. The specific calculation rule is as follows: the estimated arrival time is determined by the most recent rework release time, the theoretical transmission time corresponding to the rework path, and the queuing waiting time in front of the rework target station. That is, the system adds the values ​​of the above three on a unified time axis to obtain the absolute time scale of the current workpiece entering the rework target station and uses it as the estimated arrival time; among them, the queuing waiting time is obtained by accumulating the number of workpieces to be processed in the order of the rework target station and their corresponding estimated release time, which are represented by the queuing status of the rework station. When the difference in cycle time of the preceding workpiece is less than the preset cycle time fluctuation threshold, the number of preceding workpieces can be multiplied by the average expected release cycle time to approximate the calculation, thereby avoiding the incorrect inclusion of rework workpieces into the normal processing cycle; the system prioritizes comparing candidate sets with consistent processing cycle identifiers under the same workpiece code, and when the cycle identifier is missing, the workpiece code, path and time consistency are used for joint determination. This embodiment uses the maintenance status confirmation time point and processing round identifier to isolate and correct late and return events; This embodiment can achieve dynamic correction of the state based on a multi-dimensional consistency comprehensive matching and confirmation boundary control mechanism, avoiding the rollback of historical late events or incorrect rework data to the current true workpiece position and station occupancy status, and ultimately solving the problem of monitoring and control judgment failure caused by the lack of state boundary isolation in the existing technology. In one embodiment, for online measurement events generated by the online measurement device, the data association module determines the measurement matching value based on the workpiece code, processing path, measurement time, and the degree of correspondence between the measurement result and the processing parameters; the degree of correspondence between the measurement result and the processing parameters is determined based on the correspondence between the processing parameter sequence of the upstream station of the workpiece to be matched and the historical measurement sequence of the same workpiece; the measurement time consistency is determined according to the time interval between the absolute time of the online measurement event and the expected arrival time of the workpiece to be matched at the online measurement device, and the online measurement event corresponding to the highest measurement matching value and satisfying the preset measurement matching threshold is associated with the corresponding workpiece and the corresponding processing round identifier.

[0024] The system also includes a parameter adjustment module, which is used to adjust the advance tolerance time and lag tolerance time according to the statistical results of the event time deviation of the source equipment; set the historical statistical period as a preset time window; when the misjudgment rate confirmed by manual review within the historical statistical period is higher than the first threshold calibrated based on the upper limit of the benchmark misjudgment rate, the preset minimum matching threshold is increased by a preset step size; and when the proportion of reworked workpieces within the historical statistical period is higher than the second threshold calibrated based on the maximum allowable rework rate, the preset matching difference threshold is increased by a preset step size.

[0025] The system acquires online measurement events generated by the online measurement device, especially delayed measurement events that have occurred but arrive at the system later than some processing events. Based on the correspondence between multi-dimensional processing and measurement features, it calculates measurement matching values ​​to accurately assign quality results to the corresponding workpieces. The measurement matching value is an indicator that measures the confidence level of the matching between the quality inspection data output by the online measurement device and the upstream processing entity. The system extracts the consistency of workpiece codes, processing paths, and measurement times, and focuses on evaluating the degree of correspondence between measurement results and processing parameters. The degree of correspondence characterizes the relationship between process variables and final quality performance, which is determined by comparing the processing parameter sequence of the upstream station of the workpiece to be matched with the historical measurement sequence of the same workpiece; the historical measurement sequence of the same workpiece is the historical online measurement result sequence that has been archived under the same processing round identifier for that workpiece. When no archived historical online measurement results exist for the workpiece, the corresponding score is downgraded to a calculation based on the processing parameter sequence and the parameter-measurement baseline in the process template to which the workpiece belongs. This ensures that the judgment can still be completed in the case of the first piece or the first round of rework. The parameter-measurement baseline is composed of the average processing parameter sequence of historical qualified workpieces in the same batch at the corresponding process. During calculation, the system extracts the mean square error between each sampling point of the current processing parameter sequence and the corresponding point of the baseline, and uses an exponential decay function to make it dimensionless. The specific mapping formula is as follows: in, To assign a score based on the degree of difficulty, Mean square error, To preset the maximum permissible mean square error, This is the sensitivity attenuation coefficient, whose value is obtained through calibration using historical data. For example, it is set that when the MSE equals a preset maximum permissible mean square error, the corresponding score attenuates to a specific confidence level. For example, when the score drops to a preset lower limit, the attenuation coefficient takes a preset fixed parameter, thereby ensuring that the parameter sequence whose deviation from the baseline exceeds a preset deviation threshold obtains the corresponding score according to a nonlinear function, while strictly converging the dimensions to... interval; Before calculation, the processing parameter sequence and historical measurement sequence are aligned according to the same processing round identifier, the same process sequence, and a unified sampling window. Because of the temporal and dimensional mismatch between the high-frequency temporal characteristics of the processing process and the spatial discrete characteristics of the online measurement results, direct temporal downsampling would lead to the loss of dynamic features. Therefore, before spatial mapping, a preset workpiece-machine tool homogeneous transformation matrix is ​​required. Transform the three-dimensional motion data of the CNC equipment to the local coordinate system of the workpiece to ensure that the data tensors are aligned in the same topological space; The specific alignment rule is as follows: A speed-time model of the machining trajectory is introduced, that is, the composite feedback velocity vector of the three-dimensional feed axis of the CNC equipment at each time stamp ti is extracted. Combined with initial spatial coordinates Instead of the original feed axis drive speed, the kinematic integral formula is used: Calculate, where, For the initial timestamp of processing, This provides the timestamp corresponding to the current high-frequency machining parameters; thus, the timestamp of the high-frequency machining parameters is accurately mapped to the three-dimensional spatial coordinates of the corresponding part along the machining trajectory. Based on the spatial sampling points of low-frequency online measurement, feature aggregation is performed on the high-frequency processing parameters located in the spatial interval of adjacent measurement points, such as extracting the mean or extreme value in the interval, thereby converting the continuous high-frequency time-series processing data into a one-dimensional ordered numerical vector that corresponds one-to-one with the discrete spatial measurement points and has a strictly equal number of samples. Based on the above alignment results, the system only calculates the correlation between the processing parameter vector and the historical measurement vector when the number of samples is not less than 2 and neither of them are constant vectors, and maps the correlation result to a corresponding score in the range of 0 to 1; if the number of samples after effective alignment is less than 2, or the variance of any vector is 0, causing the correlation to lose its discriminative significance, the correlation coefficient is not used directly, but the parameter-measurement corresponding baseline is used instead, or a preset neutral corresponding score is used to participate in subsequent matching. Meanwhile, the absolute value of the time deviation is obtained based on the absolute time of the event and the expected arrival time. After normalization, a measurement time consistency score is generated. The measurement time consistency score decreases monotonically as the time deviation increases, and is recorded as 0 when the time deviation reaches the upper limit of the measurement tolerance window. The measurement tolerance window is determined according to the standard quality inspection cycle and upper limit of communication delay of the measurement equipment. The system uses a weighted summation method to combine the corresponding degree score with the consistency of the workpiece code and processing path to obtain the measurement matching value. The consistency of the workpiece code and processing path is constructed using binary logic, where matching attributes are marked as 1 and non-matching attributes are marked as 0. In order to ensure the uniqueness of the route for quality traceability, the system presets that the difference between the sum of the weights of the consistency of the workpiece code and the consistency of the processing path and the weight of the corresponding degree score reaches the preset weight difference lower limit and assigns it to the weight of the corresponding degree score. This ensures that hard spatial topology matching is used as the primary association basis, and the online measurement event with the highest measurement matching value that meets the preset measurement matching threshold is associated with the specific processing round identifier of the correct workpiece. To prevent the same online measurement event from being repeatedly assigned, when multiple candidate workpieces have the same measurement matching value and all reach the threshold, the system prioritizes the candidate with the same processing round identifier and the smaller deviation of the expected arrival time; if it still cannot be distinguished, the online measurement event is temporarily stored in the pending queue, and will be associated again after subsequent processing or departure events provide supplementary evidence. The system acquires statistical indicators during system operation and dynamically adjusts judgment parameters based on event time deviation and business conditions. The historical statistical period is set as a preset time window, with the example value being the most recent 500 event periods. The system continuously monitors the jitter of the communication network, specifically by extracting the difference between the reception time and the absolute time of each event within the historical statistical period to generate a time deviation distribution sequence. To avoid boundary divergence caused by occasional extreme outliers, the system performs outlier removal on the time deviation distribution sequence, for example, by utilizing outliers containing... The isolated forest algorithm with 10 decision trees is used to construct the evaluation model, with the number of subsamples set to 1. Calculate the deviation samples Average path length in a forest And through the abnormal score mapping formula: Calculate the outlier score, where, For the sample Abnormal scores, For the sample size The average path length normalization constant is used to remove outlier scores. Outlier samples exceeding a set threshold, such as 0.6; Based on the confidence sample intervals retained after stripping, the absolute values ​​of the preset lower quantiles, such as the 5th percentile, and the absolute values ​​of the preset upper quantiles, such as the 95th percentile, are calculated respectively. These two values ​​are used as the updated advance tolerance time and lag tolerance time, and the above tolerance time boundaries are smoothly shrunk or expanded in a sliding window mode. Meanwhile, when the false positive rate confirmed by manual review within this time window exceeds the first threshold, where the first threshold is determined based on the upper limit of the baseline false positive rate during the historical trial operation of the system under normal network conditions, for example, set to 3%, the system will proceed in fixed preset steps. Example of a value: Increase the preset minimum matching threshold Its update mechanism uses an upper limit truncation function. in, The updated preset minimum matching threshold. The preset minimum matching threshold before the update. To prevent the system from rejecting the match, the absolute upper limit of the preset matching threshold is set to 0.95 in the example, so as to raise the threshold for association and reduce false associations; When the proportion of reworked parts exceeds the second threshold within the statistical period, where the second threshold is calibrated according to the maximum allowable rework rate of the production line process design, for example, set to 10%, it indicates that the mixing density of normal and reworked parts in the production line has increased. The system increases the preset matching difference threshold by a preset step size, requiring that the difference between the highest comprehensive matching value and the second highest comprehensive matching value reach the preset matching difference threshold before association can be performed. To keep the parameter adjustment direction consistent, the statistical results of the event time deviation of the source equipment are preferably statistically analyzed using the historical distribution of the difference between the event reception time and the event absolute time, and the quantile boundaries corresponding to the advance tolerance time and the lag tolerance time are updated respectively. Unlike traditional systems that rely solely on timestamps for data association and have fixed parameters, this embodiment analyzes the correspondence between processing and measurement sequences and introduces a threshold adaptive tuning mechanism based on working condition feedback, transforming the static judgment method into a dynamic equilibrium mechanism. This embodiment can achieve reliable traceability of lagging quality data and improve system robustness based on measurement matching value and parameter adjustment module, avoid the failure of association rules caused by network delay variance exceeding the tolerance limit or return rate exceeding the preset threshold, and finally solve the problem of system monitoring distortion under extreme working conditions caused by the lack of adaptive adjustment mechanism in the existing technology. In a specific application embodiment, the aforementioned digital twin monitoring system for the automotive engine production line is deployed on a manufacturing execution system server; the system continuously receives asynchronous data streams generated by CNC machine tools, programmable logic controllers, etc., in the workshop; After new data flows in, the system adaptively completes cycle normalization and spatiotemporal interval construction. Without suspending the production line, it determines the validity and scope of the event through the correlation between workpieces, pallets, and workstations, as well as the control of status confirmation time points; triggering synchronous updates or historical compensation of workpiece positions, processing stages, and quality traceability chains in the digital twin model.

[0026] The above are merely preferred embodiments of the present invention. It should be noted that those skilled in the art can make several improvements and modifications without departing from the principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A digital twin monitoring system for an automotive engine production line, applied to a production line containing multiple workstations and conveyor lines, characterized in that, include: The data receiving module communicates with and receives data from the source device, which includes at least a CNC machine, a programmable logic controller, a conveyor controller, a workpiece code reader, and an online measurement device; it also parses the reference time, control cycle, cycle number, and time offset ratio within the cycle from the source device, which has cycle control parameters. Receive event time identifiers from source devices that do not have periodic control parameters; Generate a unified event log; The time determination module determines the absolute time of the event based on the unified event record, extracts key nodes representing changes in the status of the workpiece, pallet, or workstation based on the event type and event parameters of the unified event record, and generates the effective time interval of the event by combining the node allowable time range and the process allowable time range set by the preset communication delay and relative planned cycle time. The data association module takes the workpiece as the monitoring object, maintains the correspondence between the workpiece and the pallet and workstation, filters the workpiece to be matched according to the effective time interval of the event and the corresponding workstation, calculates the comprehensive matching value according to the consistency of route, workstation occupancy, process, code and time difference, and associates the event with the corresponding monitoring object. The status update module maintains the status confirmation time point of each monitored object, and performs current digital twin status update, correction of pending status data in the pending queue, or rewriting of historical status records according to the order of the effective time interval of the event and the status confirmation time point. The results output module outputs the monitoring information of the monitored objects and drives the production line digital twin model to update synchronously based on the updated status data.

2. The digital twin monitoring system for an automobile engine production line according to claim 1, characterized in that, The unified event log includes source device identifier, event type, event parameters, receiving time, cycle number, time offset ratio within the cycle, pallet identifier, workstation identifier, and process identifier; the time determination module determines the absolute time of the event based on the reference time, control cycle, cycle number, and time offset ratio within the cycle.

3. The digital twin monitoring system for an automobile engine production line according to claim 1, characterized in that, The effective time interval of the event is determined based on the absolute time of the event, the tolerance time of the source device, the allowable time range of the node, and the allowable time range of the process. When the lower limit of the effective time interval of the event is greater than the upper limit, the corresponding event is placed in the queue to be processed, waiting for subsequent key nodes to be recalculated, and is not used to change the current digital twin state.

4. The digital twin monitoring system for an automobile engine production line according to claim 1, characterized in that, The key nodes include at least one of the following: pallet arrival at the workstation, completion of workpiece code reading, completion of fixture clamping, workpiece departure and release, and completion of online measurement result archiving. Each key node is jointly defined by a logical production line segment representing a stage of the process flow, a physical workstation of the actual equipment, and a preset coordinate interval used to limit the spatial range. The time determination module is also used to configure a preset process flow rule table for each workstation. Only when the current process to the target process corresponding to the event belongs to the flow relationship allowed by the process flow rule table is the event allowed to update the process status of the corresponding monitored object.

5. The digital twin monitoring system for an automobile engine production line according to claim 1, characterized in that, The data association module calculates the comprehensive matching value for the selected workpieces to be matched. The comprehensive matching value is determined based on route consistency, workstation occupancy consistency, process consistency, coding consistency, time difference, and their corresponding weights. The time difference is determined based on the difference between the absolute time of the event and the expected arrival time of the workpiece to be matched. The expected arrival time of the workpiece to be matched is determined based on the workpiece's historical flow information and the production line's operating status. When the number of workpieces to be matched is 1 and the comprehensive matching value is not lower than a preset minimum matching threshold, or when the number of workpieces to be matched is greater than 1, the highest comprehensive matching value is not lower than a preset minimum matching threshold, and the difference between the highest and second-highest comprehensive matching values ​​is not lower than a preset matching difference threshold, the event is associated with the workpiece corresponding to the highest comprehensive matching value. Otherwise, the event is placed in the pending processing queue, awaiting recalculation at subsequent key nodes, and is not used to change the current digital twin state.

6. The digital twin monitoring system for an automobile engine production line according to claim 1, characterized in that, The status update module determines the status confirmation time point according to the upper limit of the valid time interval of the confirmed key node; when the upper limit of the valid time interval of the event is earlier than the status confirmation time point, the event is treated as a historical event, and only the historical status record is updated. When the lower limit of the effective time interval of the event is earlier than and the upper limit is later than or equal to the state confirmation time point, the event is treated as a cross-period event, the location status and workstation occupancy status in the current digital twin status are retained, and the process data in the historical status record is corrected. When the lower limit of the effective time interval of the event is later than or equal to the state confirmation time point, the event is taken as the current event and used to perform the current digital twin state update.

7. The digital twin monitoring system for an automobile engine production line according to claim 1, characterized in that, The data association module is also used to determine whether the return belongs to an allowed flow path according to a preset rework flow rule table when the workpiece returns from the current workstation to the already processed rework target workstation; when it belongs to an allowed flow path, it generates a processing round identifier containing the workpiece code and rework number identifier, and re-determines the expected arrival time of the workpiece entering the rework workstation according to the most recent rework release time, rework path length and rework workstation queuing status.

8. The digital twin monitoring system for an automobile engine production line according to claim 1, characterized in that, For online measurement events generated by the online measurement device, the data association module determines the measurement matching value based on the workpiece code, processing path, measurement time, and the degree of correspondence between the measurement result and the processing parameters. The degree of correspondence between the measurement result and the processing parameters is determined based on the correspondence between the processing parameter sequence of the upstream station of the workpiece to be matched and the historical measurement sequence of the same workpiece. The measurement time consistency is determined according to the time interval between the absolute time of the online measurement event and the expected arrival time of the workpiece to be matched at the online measurement device. The online measurement event with the highest measurement matching value that meets the preset measurement matching threshold is associated with the corresponding workpiece and the corresponding processing round identifier.

9. The digital twin monitoring system for an automobile engine production line according to claim 1, characterized in that, It also includes a parameter adjustment module, used to adjust the advance tolerance time and lag tolerance time according to the statistical results of the event time deviation of the source equipment; set the historical statistical period as a preset time window, and when the misjudgment rate confirmed by manual review within the historical statistical period is higher than the first threshold calibrated based on the upper limit of the benchmark misjudgment rate, increase the preset minimum matching threshold by a preset step size; and when the proportion of reworked workpieces within the historical statistical period is higher than the second threshold calibrated based on the maximum allowable rework rate, increase the preset matching difference threshold by a preset step size.

Images