Hot air cold riveting production execution synergy integration method

Abstract

The present application relates to the technical field of industrial data integration service, in particular to a hot air cold riveting production line manufacturing execution collaborative integration method. The method receives heterogeneous manufacturing data of hot air process, cold riveting process, batch flow, work order state, process parameters and quality determination through industrial data integration service, converts the manufacturing data into process event objects containing manufacturing object identification, process state, time identification and source identification according to a process event semantic model, determines process meaning by using field semantic mapping relationship, checks continuous process sequence position by using process anchor point sequence, merges repeated reporting and delayed reporting events by using event fingerprints, and then encapsulates the events into a manufacturing execution collaborative data package and writes the data package into a manufacturing execution data chain. The method can reduce the error association caused by the mispositioning of process semantics of isomer data fields, and make the hot air process result, cold riveting execution state and quality determination result form a continuous and checkable data chain under the same manufacturing object.

Description

Technical Field

[0001] This invention relates to the field of industrial data integration service technology, specifically a collaborative integration method for manufacturing execution in hot air cold riveting production lines. Background Technology

[0002] The manufacturing execution process on a hot-air cold riveting production line typically requires integrating data such as hot-air processes, cold riveting processes, work order flow, batch status, process parameters, and quality judgments into the manufacturing execution system (MES). Current practices often employ industrial data integration services as an intermediate layer. This layer retrieves data from various source systems according to pre-defined interfaces and then writes it into the MES database based on field mappings. Conventional integration solutions generally use data tables, interface fields, message topics, or file records as processing objects, mapping source fields to target fields. For example, hot-air processing completion status is mapped to a preceding completion field, cold riveting execution status to a subsequent processing field, and quality judgment results to inspection result fields. The key to this type of solution is ensuring data transmission, field conversion, and record storage. Issues such as inconsistent field names, formats, or different source systems are typically resolved by configuring field dictionaries, conversion rules, and interface retransmission mechanisms.

[0003] In similar existing technologies, the Manufacturing Execution System (MES) establishes production line execution records around work order numbers or batch numbers. After the hot air process is completed, the data integration service receives status data and updates the work order progress. The cold riveting process reads the updated work order progress and generates subsequent execution statuses. The quality system returns quality results after inspection. To handle data duplication and interface delays, existing solutions generally use timestamp sorting, primary key deduplication, interface log verification, and exception retransmission for data maintenance. This type of processing belongs to the integration control oriented towards the data transmission link. Its judgment is mainly based on whether the interface is called successfully, whether the fields are complete, whether the time meets the chronological order, and whether the primary key is duplicated. For the differences in the status expression of the same manufactured object in different systems in the continuous hot air and cold riveting processes, conventional solutions usually rely on manual maintenance of field meanings or adding fixed verification rules to the MES, without forming a unified process semantic constraint before the data enters the MES chain.

[0004] The main technical problem with the aforementioned existing technologies is that, even after interface transmission and field mapping are completed, the heterogeneous manufacturing data in hot air and cold riveting production lines may still suffer from misalignment of process semantics. Because the hot air and cold riveting processes have different data sources, status naming, acquisition times, and batch flow granularities, the same field may correspond to different process meanings in different source systems. Furthermore, batches or workpieces under the same work order may experience misalignment due to delayed or duplicate reporting. Existing integration methods, primarily based on field mapping and time sorting, can only determine data format and transmission status, but cannot establish a linkage verification relationship between manufacturing objects, process sequences, and event uniqueness. This can lead to the manufacturing execution system incorrectly associating the previous batch of hot air completion data, the current batch of cold riveting status, and subsequent quality results, resulting in a manufacturing execution data chain inconsistent with the actual process flow. Summary of the Invention

[0005] The purpose of this invention is to provide a collaborative integration method for manufacturing execution in a hot air cold riveting production line, which can solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0007] The collaborative integration method for manufacturing execution in hot air cold riveting production lines includes: receiving heterogeneous manufacturing data such as hot air processes, cold riveting processes, batch flow, work order status, process parameters, and quality judgments through industrial data integration services;

[0008] Based on a preset process event semantic model, the heterogeneous manufacturing data is converted into process event objects containing manufacturing object identifiers, process status, time identifiers, and source identifiers;

[0009] The process meaning of the process event object is determined according to the field semantic mapping relationship; the sequential position of the process event object in the hot air cold riveting continuous process is verified according to the process anchor sequence; and the process event objects that are repeatedly reported and delayed are merged according to the event fingerprint.

[0010] The validated and merged process event objects are encapsulated into a manufacturing execution collaboration data package and written into the manufacturing execution data chain.

[0011] Preferably, the step of converting the heterogeneous manufacturing data into process event objects based on a preset process event semantic model includes: extracting the work order number, batch number, workpiece flow identifier, process name, status value, generation time, parameter version, and quality result identifier from the heterogeneous manufacturing data;

[0012] Establish binding indexes between work orders, batches, and workpieces according to the manufacturing object hierarchy;

[0013] The binding index is combined with the process name, status value, and generation time to form the event primary key;

[0014] Map the source system, field name, and field value to the event attribute set under the event primary key.

[0015] Preferably, determining the process meaning of the process event object based on the field semantic mapping relationship includes: setting field semantic categories for hot air treatment status, cold riveting execution status, process parameter version, and quality judgment result respectively;

[0016] Map the original field names from the source system to the semantic categories of the fields;

[0017] When the same original field name corresponds to multiple field semantic categories, the target field semantic category is determined based on the manufacturing object identifier, process status, and source identifier of the process event object.

[0018] Process event objects for which the semantic category of the target field could not be determined are marked as events to be verified.

[0019] Preferably, the step of verifying the sequential position of the process event object in the hot air cold riveting continuous process according to the process anchor point sequence includes: establishing a process anchor point sequence including hot air treatment start, hot air treatment end, cold riveting preparation, cold riveting execution completion and quality confirmation;

[0020] Match the process state corresponding to the process event object to the target anchor point in the process anchor point sequence;

[0021] Generate sequence verification results based on the order of recorded anchor points and target anchor points under the same manufacturing object identifier;

[0022] Process event objects whose sequence relationship does not satisfy the process anchor point sequence are written into the conflict event set.

[0023] Preferably, the step of establishing a binding index between work orders, batches, and workpieces according to the manufacturing object hierarchy includes: using the work order number as a first-level index, the batch number as a second-level index, and the workpiece transfer identifier as a third-level index;

[0024] When the heterogeneous manufacturing data lacks any level identifier, a candidate binding index is searched based on the status of adjacent processes, generation time, and source system in the received process event objects.

[0025] After performing hierarchical consistency verification on the candidate binding indexes, the completed binding indexes are generated and the basis for completion is recorded.

[0026] Preferably, the semantic category of the target field is determined based on the manufacturing object identifier, process status, and source identifier of the process event object, including: when there is a conflict in the semantic category of the field, reading the process status of adjacent process event objects under the same manufacturing object identifier;

[0027] The process status of the adjacent process event objects is compared with the semantic category of the field corresponding to the source identifier.

[0028] When the combined comparison result is consistent with the state transition relationship in the semantic model of the hot air cold riveting process event, the semantic category of the corresponding field is determined as the semantic category of the target field.

[0029] Preferably, generating the sequence verification result based on the sequential relationship between the recorded anchor points and the target anchor points under the same manufacturing object identifier includes: reading the recorded anchor point queue according to the manufacturing object identifier;

[0030] The target anchor point is compared with the last anchor point in the recorded anchor point queue.

[0031] When the target anchor point is earlier than the last anchor point, it is determined whether the event is late based on the generation time and event fingerprint of the process event object.

[0032] When the target anchor point is later than the last anchor point and intermediate anchor points are missing, the missing anchor point and the target anchor point are written together into the pending completion verification queue.

[0033] Preferably, the step of merging duplicate and delayed reported process event objects based on event fingerprints includes: generating event fingerprints based on manufacturing object identifier, process status, parameter version, source identifier, and generation time;

[0034] Match the newly generated event fingerprint with the already stored event fingerprint;

[0035] When event fingerprints are the same but event attribute sets differ, merge event attribute sets according to source priority and attribute completeness.

[0036] When event fingerprints are different but process anchors and binding indexes are the same and the generation time is within the same process time period, associated fingerprints are generated and the delayed event merging process is initiated.

[0037] Preferably, after writing the missing anchor point and the target anchor point together into the queue to be filled in the verification, the method further includes: reading the process event objects located before and after the missing anchor point under the same manufacturing object identifier from the manufacturing execution data chain;

[0038] Generate candidate events for missing anchor points based on the field semantic categories, parameter versions, and source identifiers of the preceding and following process event objects;

[0039] Perform a sequence consistency check between the missing anchor point candidate events and the target anchor points in the queue to be filled in;

[0040] Missing anchor candidate events that pass the verification are written into the manufacturing execution data chain.

[0041] Preferably, the step of generating associated fingerprints and entering the delayed event merging process includes: reading stored process event objects and newly received process event objects according to the associated fingerprints;

[0042] Compare the consistency between the stored process event object and the newly received process event object in terms of field semantic category, process anchor, binding index, and quality result identifier;

[0043] When the semantic category of the field and the binding index are consistent and the quality result identifiers are complementary, the newly received process event object will be merged into the already stored process event object;

[0044] When process anchor points are inconsistent, the process event object corresponding to the associated fingerprint will be transferred to the conflict event set.

[0045] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0046] 1. Through industrial data integration services, heterogeneous manufacturing data from hot air processes, cold riveting processes, batch flow, work order status, process parameters, and quality judgments are converted into process event objects. Field semantic mapping relationships, process anchor sequence, and event fingerprint merging mechanisms are introduced into the same processing chain, transforming the data integration process from field-level transmission to process event-level verification. Field semantic mapping relationships determine the actual category of the source field in the hot air and cold riveting processes. Process anchor sequence defines the sequential position of process events between the start of hot air processing, the end of hot air processing, cold riveting preparation, cold riveting execution completion, and quality confirmation. Event fingerprints identify and merge duplicate and delayed reported process events. With these technical features working together, each event in the manufacturing execution data chain includes a manufacturing object identifier, process status, time identifier, and source identifier. This enables joint verification of object ownership, process location, and event uniqueness before writing, resolving the issue of successful heterogeneous manufacturing data transmission but misaligned process semantics.

[0047] 2. By using hierarchical binding indexes for work orders, batches, and workpieces, comparing adjacent process statuses, establishing a pending verification queue, implementing a delayed event merging process, and creating conflict event sets, data gaps, semantic conflicts, and sequence anomalies from different source systems are handled hierarchically. For manufacturing data lacking hierarchical identifiers, candidate binding indexes can be searched based on adjacent process statuses, generation times, and source systems, and the basis for completion can be recorded. For cases where the same field corresponds to multiple semantic categories, the semantic category of the target field can be determined by combining the manufacturing object identifier, process status, and source identifier. For cases where the target anchor point is earlier than the recorded anchor point or an intermediate anchor point is missing, late events, missing events, and conflict events can be distinguished. For cases where event fingerprints are different but process anchor points and binding indexes are the same, merging or transferring to conflict handling can be performed based on the field semantic category, binding index, and quality result identifier. Thus, hot air process results, cold riveting execution status, and quality judgment results form a continuous and verifiable data chain under the same manufacturing object. Attached Figure Description

[0048] Figure 1 A flowchart illustrating the overall process of implementing a collaborative integration method for hot air cold riveting production lines;

[0049] Figure 2 Flowchart for creating object hierarchy binding and field semantic mapping;

[0050] Figure 3 Flowchart for process anchor point sequence verification and missing data completion;

[0051] Figure 4 Write a flowchart for event fingerprint merging, associated fingerprint processing, and manufacturing execution data chain. Detailed Implementation

[0052] In one embodiment, a collaborative integration method for manufacturing execution in a hot-air cold riveting production line includes: receiving heterogeneous manufacturing data such as hot-air processes, cold riveting processes, batch flow, work order status, process parameters, and quality judgments through an industrial data integration service; converting the heterogeneous manufacturing data into process event objects containing manufacturing object identifiers, process status, time identifiers, and source identifiers based on a preset process event semantic model; determining the process meaning of the process event objects based on field semantic mapping relationships; verifying the sequential position of the process event objects in the continuous hot-air cold riveting processes based on process anchor point sequences; merging duplicate and delayed reported process event objects based on event fingerprints; and encapsulating the verified and merged process event objects into a manufacturing execution collaborative data package and writing it into the manufacturing execution data chain.

[0053] In this embodiment, reference Figure 1The industrial data integration service serves as the entry point for manufacturing execution data into the unified processing chain. The receiving objects are limited to heterogeneous manufacturing data from hot air processes, cold riveting processes, batch flows, work order status, process parameters, and quality judgments. This heterogeneous manufacturing data can originate from data tables, message records, interface messages, or file records in the manufacturing execution system, process parameter recording system, quality judgment recording system, or production line execution recording system. Upon receipt, the data is not directly written into the manufacturing execution data chain. Instead, it undergoes source registration, field decomposition, and record merging. Source registration stores the data source system, receiving time, and original record identifier. Field decomposition breaks down the original field names and values ​​into identifiable field fragments. Record merging temporarily merges original records that appear repeatedly within the same receiving batch from the same source system. After these processes, a set of original data awaiting semanticization is formed.

[0054] In this embodiment, a preset process event semantic model is used to convert the original data set from field records into process event objects. The process event objects include manufacturing object identifiers, process status, time identifiers, and source identifiers. The manufacturing object identifier is used to characterize the attribution relationship between work orders, batches, and workpieces. The process status is used to characterize the process position in hot air treatment, cold riveting execution, or quality confirmation. The time identifier is used to characterize the relationship between the event generation time and the data reception time. The source identifier is used to characterize the business system from which the data comes and the original data record. The conversion process is executed in the order of manufacturing object priority, process status second, time identifier supplementation, and source identifier verification, so that data of the same manufacturing object in different source systems can be classified into the same event semantic space.

[0055] In this embodiment, the field semantic mapping relationship is used to solve the problem of inconsistency between the original field name and the process meaning, the process anchor sequence is used to solve the problem of verifying the order of the hot air process and the cold riveting process, and the event fingerprint is used to solve the problem of event uniqueness caused by duplicate reporting and delayed reporting. The three types of technical means are sequentially associated in the same processing chain. The process meaning output by the field semantic mapping relationship is used as the input for process anchor matching, the process anchor matching result is used as the verification condition for event fingerprint merging, and the event fingerprint merging result is used as the basis for encapsulating the manufacturing execution collaborative data packet. This avoids the situation where only field transmission is completed without process semantic alignment.

[0056] In this embodiment, field semantic matching can be calculated using a semantic matching score, which represents the degree of correspondence between process event objects and candidate field semantic categories. The calculation formula is as follows:

[0057] ;

[0058] Where e represents the process event object to be processed, and c represents the semantic category of the candidate field. This indicates the matching value between the original field name and the semantic category of the candidate field. This indicates the matching values ​​between the process status and the semantic category of the candidate field. This indicates the matching value between the manufacturing object identifier and the semantic category of the candidate field. This indicates the matching value between the source identifier and the semantic category of the candidate field. , , and These represent the weights of the corresponding items. When a certain hot air completion field name is the same as the hot air processing status category, the value is 1; when the process status is at the end of hot air processing, the value is 1; when the manufacturing object identifier is missing, the value is 0; and when the source identifier comes from the hot air process record system, the value is 1, the semantic matching score is determined by the three types of consistent information and the one type of missing information. The calculation result is used to select the semantic category of the candidate field or enter the event to be verified.

[0059] Furthermore, the manufacturing execution collaboration data package is encapsulated according to process event objects, verification basis, merging basis, and write index. The process event object stores the transformed event content, the verification basis stores the field semantic mapping results and process anchor sequence matching results, the merging basis stores the event fingerprint, association fingerprint, and attribute merging record, and the write index stores the work order index, batch index, and workpiece index in the manufacturing execution data chain. When the encapsulated data package enters the manufacturing execution data chain, it carries the reference relationship of the original source record, so that subsequent quality judgment and anomaly investigation can trace back to the original data source along the same manufacturing object.

[0060] In a preferred embodiment, reference Figure 2 Based on a preset process event semantic model, the heterogeneous manufacturing data is converted into process event objects, including: extracting the work order number, batch number, workpiece flow identifier, process name, status value, generation time, parameter version, and quality result identifier from the heterogeneous manufacturing data; establishing a binding index between work orders, batches, and workpieces according to the manufacturing object hierarchy; combining the binding index with the process name, status value, and generation time to form an event primary key; and mapping the source system, field name, and field value to the event attribute set under the event primary key.

[0061] Specifically, the work order number is used to determine the production task boundary, the batch number is used to determine the manufacturing batch boundary under the same work order, the workpiece transfer identifier is used to determine the transfer boundary of a single piece or group of objects within a batch, the process name is used to determine the hot air process, cold riveting process, or quality confirmation stage, the status value is used to determine the start, end, preparation, completion, or judgment status, the generation time is used to determine the sequence of events, the parameter version is used to determine the effective relationship of process parameter records, and the quality result identifier is used to determine the correspondence between quality judgment data and process events. During the extraction process, the field names are standardized, and spaces, separators, differences in capitalization, and custom prefixes from the source system are unified into internal field identifiers, and then written into the process event semantic model according to the field semantic mapping relationship.

[0062] In this embodiment, the binding index forms a manufacturing object hierarchy based on work orders, batches, and workpieces. The work order index is used to limit the production task to which the data belongs, the batch index is used to limit the production batch to which the data belongs, and the workpiece index is used to limit the specific flow object to which the data belongs. The event primary key is composed of the binding index, process name, status value, and generation time. This ensures that hot air completion data from different batches under the same work order will not be mistakenly assigned to the same cold riveting status record due to identical field names, and that the quality results of different workpieces under the same batch will not be mistakenly bound to the wrong workpiece due to similar times.

[0063] In this embodiment, the event attribute set stores the source system, field name, and field value. It also stores the original record identifier, reception time, field conversion record, and exception flag. When the event attribute set does not participate in the event primary key uniqueness judgment, it serves as supplementary explanatory information. When participating in conflict verification, it serves as the basis for semantic category judgment. When participating in event merging, it serves as the basis for attribute completeness judgment. When there are field attributes from multiple source systems under an event primary key, the attribute set is grouped according to the source system, field semantic category, and reception order, so that subsequent merging can distinguish between complementary attributes and conflicting attributes.

[0064] In a preferred embodiment, a binding index is established between work orders, batches, and workpieces according to the hierarchical relationship of manufacturing objects, including: using the work order number as a first-level index, the batch number as a second-level index, and the workpiece transfer identifier as a third-level index; when the heterogeneous manufacturing data lacks any level identifier, a candidate binding index is searched based on the status of adjacent processes, generation time, and source system in the received process event objects; after performing hierarchical consistency verification on the candidate binding indexes, a completed binding index is formed and the basis for completion is recorded.

[0065] Specifically, the hierarchical consistency check is performed in the manner of constraining the lower-level index by the upper-level index. After the first-level index is determined, the second-level index can only be searched within the corresponding work order. After the second-level index is determined, the third-level index can only be searched within the corresponding batch. When the second-level index is missing, the candidate batch is searched based on the generation time and source system of the adjacent process status under the same work order. When the third-level index is missing, the candidate workpiece is searched based on the time proximity between the adjacent hot air end event, cold riveting preparation event and quality confirmation event under the same batch. After the completed binding index is formed, the original missing fields, candidate sources and check results are written into the completion basis.

[0066] In this embodiment, the candidate binding index can be calculated using binding confidence, whereby the binding confidence represents the strength of the attribution relationship between the original record with the missing level identifier and the candidate manufacturing object index. The calculation formula is as follows:

[0067] ;

[0068] Where x represents heterogeneous manufacturing data lacking a hierarchical identifier, and b represents a candidate binding index. The value indicates whether the states of adjacent processes can be connected. This value indicates whether the generation time falls within the process time period of the candidate object. The value indicates whether the source system and the candidate object record have the same source. This indicates whether the candidate binding index has a traceable value in the manufacturing execution data chain, showing the preceding and following links. , , and The value represents the weight of the corresponding item. When the cold riveting preparation data with missing batch number can be connected with the hot air end status of a certain candidate batch, the value is 1; when the generation time is within the corresponding process time period, the value is 1; when the source system is consistent, the value is 1; and when the link exists, the value is 1. The binding credibility reaches the maximum value and can form a complete index. When the link does not exist, the value is 0. The calculation result is reduced and enters the manual confirmation or pending verification event processing path.

[0069] In a preferred embodiment, determining the process meaning of the process event object based on the field semantic mapping relationship includes: setting field semantic categories for hot air treatment status, cold riveting execution status, process parameter version, and quality judgment result respectively; mapping the original field names in the source system to the field semantic categories; when the same original field name corresponds to multiple field semantic categories, determining the target field semantic category based on the manufacturing object identifier, process status, and source identifier of the process event object; and marking process event objects for which the target field semantic category cannot be determined as events to be verified.

[0070] In this embodiment, the semantic categories of fields are established according to the process flow, not directly according to the field names of the source system. The hot air processing status category is used to include the status meanings of hot air start, hot air completion, hot air stop, and hot air verification. The cold riveting execution status category is used to include the status meanings of cold riveting preparation, cold riveting execution, cold riveting completion, and cold riveting abnormality. The process parameter version category is used to include the version meanings of hot air parameter version, cold riveting parameter version, and effective time. The quality judgment result category is used to include the judgment meanings of qualified, unqualified, re-judged, and pending confirmation. When mapping the original field names, the field paths and value interpretations of the source system are saved to avoid fields with the same name being directly regarded as having the same meaning in different source systems.

[0071] In this embodiment, when the same original field name corresponds to multiple field semantic categories, the manufacturing object identifier, process status, and source identifier are combined for judgment. The manufacturing object identifier determines the object to which the field belongs, the process status determines the process position of the field, and the source identifier determines the field generating system. The three together define the semantic category of the target field. For example, if there is an original field named "status" in the source system, it corresponds to the hot air processing status in the hot air process record, the cold riveting execution status in the cold riveting execution record, and the quality judgment result in the quality judgment record. If it is not possible to determine based on the field name alone, the process status and source identifier are read for combination matching.

[0072] In a preferred embodiment, determining the target field semantic category based on the manufacturing object identifier, process status, and source identifier of the process event object includes: when a field semantic category conflict occurs, reading the process status of adjacent process event objects under the same manufacturing object identifier; combining and comparing the process status of adjacent process event objects with the field semantic category corresponding to the source identifier; when the combination comparison result is consistent with the state transition relationship in the hot air cold riveting process event semantic model, determining the corresponding field semantic category as the target field semantic category.

[0073] Specifically, field semantic category conflict refers to the situation where the same original field name, the same field value, or the same source path can be mapped to multiple field semantic categories. When reading adjacent process event objects, the same manufacturing object identifier is used as the boundary. The confirmed preceding events are read forward, and the received but not written data events are read backward. The process status of the preceding event, the candidate semantic category of the current event, and the process status of the subsequent event are combined into a state transition fragment. This fragment is then compared with the legal state transition relationship in the process event semantic model. Candidate categories that conform to the transition relationship are written into the target field semantic category, while candidate categories that do not conform to the transition relationship are moved into the event to be verified.

[0074] In this embodiment, the state transition relationship can be represented as a combination of anchor point position constraints and field semantic constraints. The anchor point position constraints define the sequential relationship between the start of hot air processing, the end of hot air processing, cold riveting preparation, the completion of cold riveting execution, and quality confirmation. The field semantic constraints define the allowed field semantic categories under a certain anchor point. For example, the anchor point for the end of hot air processing allows the hot air processing status and hot air parameter version, the anchor point for the preparation of cold riveting allows the cold riveting execution status and cold riveting parameter version, and the anchor point for quality confirmation allows the quality judgment result and quality result identifier. When comparing the combinations, the correspondence between the anchor point position and the field semantic category must be satisfied simultaneously.

[0075] In a preferred embodiment, reference Figure 3 The process of verifying the sequential position of the process event object in the continuous hot air and cold riveting process according to the process anchor point sequence includes: establishing a process anchor point sequence including hot air treatment start, hot air treatment end, cold riveting preparation, cold riveting execution completion, and quality confirmation; matching the process state corresponding to the process event object to the target anchor point in the process anchor point sequence; generating a sequence verification result based on the sequential relationship between the recorded anchor points and the target anchor point under the same manufacturing object identifier; and writing process event objects whose sequential relationship does not satisfy the process anchor point sequence into the conflict event set.

[0076] In this embodiment, the process anchor sequence is the minimum sequence constraint set of the hot air and cold riveting continuous processes. Hot air processing start indicates the start of the preceding processing process, hot air processing end indicates the completion of the required preceding state before cold riveting, cold riveting preparation indicates that the subsequent process has entered the pending execution state, cold riveting execution completion indicates that the cold riveting action state is completed, and quality confirmation indicates that the manufacturing execution data chain can form a closed loop with the quality result. When matching the target anchor, the semantic category of the process event object is read first, and then the status value and parameter version are read. If the semantic category of the field belongs to the hot air processing state and the status value indicates the end, then the target anchor is matched as hot air processing end. If the semantic category of the field belongs to the quality judgment result and there is a quality result identifier, then the target anchor is matched as quality confirmation.

[0077] In this embodiment, the sequential verification result is generated according to the positional relationship between the recorded anchor queue and the target anchor under the same manufacturing object identifier. The recorded anchor queue stores anchors that have passed verification and been written into the manufacturing execution data chain. Before the target anchor enters the queue, it is compared with the last anchor in the queue. If the target anchor is after the last anchor and no intermediate anchors are missing, a pass result is generated. If the target anchor is before the last anchor, it enters the late event judgment. If the target anchor is after the last anchor but intermediate anchors are missing, it enters the pending completion verification queue. If the target anchor and the last anchor are the same but the event fingerprints are different, it enters the duplicate or conflict judgment.

[0078] In this embodiment, the anchor point order difference can be represented by a position difference value. The position difference value is used to determine the order relationship between the target anchor point and the recorded anchor point queue. The calculation formula is as follows:

[0079] ;

[0080] in, This indicates the position code of the target anchor point in the process anchor point sequence. This indicates the position code of the last anchor point in the recorded anchor point queue under the same manufacturing object identifier. When the hot air treatment start position code is 1, the hot air treatment end position code is 2, the cold riveting preparation position code is 3, the cold riveting execution completion position code is 4, and the quality confirmation position code is 5, if the recorded anchor point is hot air treatment completion and the target anchor point is cold riveting preparation, then... If the target anchor point is 1 and the path is entered sequentially, and the anchor point is already recorded as "cold riveting completed" and the target anchor point is "cold riveting preparation", then... If the value is negative, proceed to the late event judgment. If the anchor point has been recorded as having completed hot air processing and the target anchor point has been recorded as having completed cold riveting, then... If the result is greater than 1, it will be added to the pending verification queue.

[0081] In a preferred embodiment, generating a sequence verification result based on the sequential relationship between recorded anchor points and target anchor points under the same manufacturing object identifier includes: reading the recorded anchor point queue according to the manufacturing object identifier; comparing the position of the target anchor point with the last anchor point in the recorded anchor point queue; when the target anchor point is earlier than the last anchor point, determining whether it is a late event based on the generation time and event fingerprint of the process event object; when the target anchor point is later than the last anchor point and intermediate anchor points are missing, writing the missing anchor point and the target anchor point together into the pending completion verification queue.

[0082] Specifically, the determination of late events is not based solely on the reception time. Instead, it compares the generation time, the event fingerprint, and the corresponding anchors in the recorded anchor queue. If the target anchor is earlier than the last anchor but the generation time is within the time period that the corresponding anchor should fall within, and the event fingerprint can form a correlation fingerprint with the recorded process event object, then it is identified as a late event and enters the delayed event merging process. If the target anchor is earlier than the last anchor and the generation time conflicts with the time period of the corresponding anchor, or if the event fingerprint cannot form a correlation fingerprint with the recorded process event object, then it is written into the conflict event set to avoid reverse-order data directly overwriting the existing manufacturing execution data chain.

[0083] In this embodiment, the verification queue to be completed stores missing anchors, target anchors, manufacturing object identifiers, preceding anchors, generation time, and source identifiers. Missing anchors represent anchors that have not appeared, derived from the anchor position difference. Target anchors represent the positions of events that have been received but cannot be directly written. Preceding anchors represent the last anchor in the recorded anchor queue. The records in the verification queue to be completed do not change the confirmed state of the manufacturing execution data chain; they only serve as temporary data structures for the subsequent generation of candidate events for missing anchors and for sequential consistency verification.

[0084] In a preferred embodiment, after writing the missing anchor point and the target anchor point together into the completion verification queue, the method further includes: reading process event objects located before and after the missing anchor point under the same manufacturing object identifier from the manufacturing execution data chain; generating candidate events for missing anchor points based on the field semantic category, parameter version, and source identifier of the preceding and following process event objects; performing a sequence consistency check between the candidate events for missing anchor points and the target anchor point in the completion verification queue; and writing the candidate events for missing anchor points that pass the check into the manufacturing execution data chain.

[0085] Specifically, missing anchor point candidate events are not generated directly from a blank state, but are constrained by the process event objects before and after the missing anchor point. The preceding process event object limits the starting boundary of the missing anchor point, the subsequent target anchor point limits the ending boundary of the missing anchor point, the field semantic category limits the types of events that can be generated, the parameter version limits the range of process parameters that the candidate event can inherit, and the source identifier limits the data source path that the candidate event can use. When the missing anchor point is cold riveting preparation, the preceding hot air processing completion event provides the manufacturing object identifier and parameter version boundary, and the subsequent cold riveting execution completion event provides the state constraint that the cold riveting process has already occurred. A candidate event can only enter the sequence consistency check if it meets the constraints at both ends at the same time.

[0086] In this embodiment, the consistency of missing anchor point candidate events can be represented by candidate consistency degree. The candidate consistency degree is used to measure the semantic and sequential matching degree between the candidate event and the preceding and following process event objects. The calculation formula is as follows:

[0087] ;

[0088] Where c represents the missing anchor candidate event, This indicates the anchor point connection value between the candidate event and the preceding process event object. This indicates the anchor point connection value between the candidate event and the subsequent target anchor point. This indicates whether the version of the candidate event parameter falls within the allowed version range of the preceding and following events. The value indicates whether the candidate event source identifier is connectable with the preceding and following event source paths. , , and The value represents the weight of the corresponding item. When the candidate cold riveting preparation event can be connected with the end of the preceding hot air processing, the value is 1; when it can be connected with the completion of the subsequent cold riveting execution, the value is 1; when the parameter version is consistent, the value is 1; and when the source path is missing, the value is 0, the candidate consistency is determined by 3 types of consistent information and 1 type of missing information. The calculation result is used to determine whether the candidate event is written into the manufacturing execution data chain or continues to be retained in the pending verification queue.

[0089] In a preferred embodiment, reference Figure 4 The process merges duplicate and delayed reported process event objects based on event fingerprints, including: generating event fingerprints based on manufacturing object identifier, process status, parameter version, source identifier, and generation time; matching the newly generated event fingerprints with the stored event fingerprints; merging event attribute sets according to source priority and attribute completeness when event fingerprints are the same but event attribute sets differ; and generating associated fingerprints and entering the delayed event merging process when event fingerprints are different but process anchor points and binding indexes are the same and generation times are within the same process time period.

[0090] In this embodiment, the event fingerprint is generated based on the manufacturing object identifier, process status, parameter version, source identifier, and generation time. The manufacturing object identifier ensures that the event belongs to the same object. The process status ensures that the event is in the same process position. The parameter version ensures that the event references the same process parameter context. The source identifier ensures that the source of the event is traceable. The generation time ensures that the event is in the same occurrence time context. Before generating the fingerprint, the generation time is normalized to the process time period. The normalization process does not change the original generation time, but only maps multiple reports within the same process execution range to the same time period number, which is used to distinguish duplicate reports, supplementary reports, and cross-process false reports.

[0091] Specifically, event fingerprints can be expressed using event digests, calculated using the following formula:

[0092] ;

[0093] in, This represents the event fingerprint of process event object e. Indicates the identifier of the manufactured object. Indicates the process status. Indicates the parameter version. Indicates the source identifier. This indicates the process time period after the generation time is normalized. When two hot air processing completion events have the same manufacturing object identifier, the same process status, the same parameter version, the same source identifier, and the same process time period, the summary operation outputs the same event fingerprint and enters the attribute merging path. When the process anchor point and binding index are the same but the source identifier or generation time period is different, the event fingerprints are different and enter the association fingerprint judgment.

[0094] In this embodiment, when there are differences in the event attribute set, the existing event attributes are not directly overwritten. Instead, they are merged according to source priority and attribute completeness. The source priority is predetermined based on the correspondence between the source system and the semantic category of the field. The attribute completeness is calculated based on the existence of required and supplementary fields in the event attribute set. During merging, field values ​​with higher source priority and higher attribute completeness are retained first. If two field values ​​belong to complementary fields, they are retained and written into the merged record. If two field values ​​belong to the same semantic category and have different values, the original record is retained and written into the conflict event set.

[0095] In this embodiment, attribute completeness can be calculated using a weighted average of field occupancy ratios. Attribute completeness represents the sufficiency of the event attribute set in supporting the process event object, and the calculation formula is as follows:

[0096] ;

[0097] in, This represents the completeness of the attributes of the process event object e, where n represents the number of fields involved in the completeness calculation. This represents the weight of the j-th field item. This indicates whether the j-th field item exists in the event attribute set and passes semantic validation. When the process event object contains manufacturing object identifier, process status, generation time, and source identifier, and all of them pass semantic validation, the corresponding field item is 1. When the quality result identifier has not yet been returned, the corresponding field item is 0. The completeness calculation result is used to select the source of the field to retain and to determine whether it is necessary to continue waiting for supplementary reporting when merging attributes.

[0098] In a preferred embodiment, generating an associated fingerprint and entering a delayed event merging process includes: reading stored process event objects and newly received process event objects according to the associated fingerprint; comparing the consistency between stored process event objects and newly received process event objects in terms of field semantic category, process anchor point, binding index, and quality result identifier; when the field semantic category and binding index are consistent and the quality result identifiers are complementary, merging the newly received process event object into the stored process event object; when the process anchor points are inconsistent, transferring the process event object corresponding to the associated fingerprint into the conflict event set.

[0099] Specifically, association fingerprints differ from event fingerprints. Event fingerprints are used to identify the same event, while association fingerprints are used to identify delayed supplementary events that may belong to the same process time period. Association fingerprints consist of binding indexes, process anchors, and process time periods. They do not directly use source identifiers and parameter versions as exclusion conditions, allowing delayed quality results from the quality judgment record system to be associated with stored cold riveting completion events. When comparing consistency relationships, consistent field semantic categories indicate that the two events belong to a mergeable semantic range, consistent process anchors indicate that the two events are at the same process position, consistent binding indexes indicate that the two events belong to the same manufacturing object, and complementary quality result identifiers indicate that the newly received event supplements the missing quality judgment content in the stored events.

[0100] In this embodiment, when the field semantic category and binding index are consistent and the quality result identifiers are complementary, the newly received process event object is merged into the stored process event object. The merging process retains the original source records of the two events and adds the merging source, merging time, associated fingerprint, and supplementary field name to the event attribute set. When the process anchor points are inconsistent, the merging process is not performed. Instead, the process event object corresponding to the associated fingerprint is transferred to the conflict event set. The conflict event set stores the conflict anchor point, conflict field, binding index, and source record to prevent events from different process locations from being mistakenly merged due to close time or identical objects.

[0101] Furthermore, the manufacturing execution data chain adopts a chain-like record structure organized by manufacturing object identifier. Each chain node corresponds to a process event object after semantic verification, anchor point verification, and fingerprint merging. Nodes establish sequential reference relationships according to the process anchor point sequence. The node stores the event primary key, event attribute set, field semantic category, process anchor point, event fingerprint, association fingerprint, merged record, and conflict flag. When a new manufacturing execution collaboration data packet is written, the writing logic first locates the data chain corresponding to the manufacturing object identifier, then locates the target anchor point position, and then determines the addition, merging, completion, or conflict handling path based on the event fingerprint.

[0102] In this embodiment, the conflict event set is not part of the final manufacturing execution data chain, but rather serves as an independent exception temporary storage structure. Events entering the conflict event set include events whose field semantic categories cannot be determined, events whose anchor point sequence does not satisfy the process anchor point sequence, duplicate events with different event fingerprints but conflicting semantic categories, and erroneous association events with the same association fingerprints but inconsistent process anchor points. The conflict event set stores the event entry reason, original source record, candidate semantic categories, candidate anchor points, and manufacturing object identifier. After subsequent manual confirmation or rule updates, the conflict events can be re-sent into the semantic mapping and anchor point verification chain.

[0103] Preferably, the industrial data integration service maintains field semantic mapping relationships, process anchor sequence, event fingerprint index, association fingerprint index, binding index table and conflict event set during operation. The field semantic mapping relationship is saved in version when the source system field changes. The process anchor sequence is saved in version when the hot air cold riveting process is adjusted. The event fingerprint index is used to query whether there are duplicate reports of stored events. The association fingerprint index is used to query whether there are delayed supplementary events. The binding index table is used to query the manufacturing object hierarchy relationship. The conflict event set is used to store events that cannot be directly written into the manufacturing execution data chain.

[0104] In this embodiment, consistency confirmation is performed before the manufacturing execution collaborative data packet is written into the manufacturing execution data chain. Consistency confirmation includes confirmation of the existence of binding index, confirmation of field semantic category, confirmation of target anchor order, confirmation of event fingerprint merging, and confirmation of conflict mark. If the binding index does not exist, it enters the hierarchical completion path. If the field semantic category is uncertain, it enters the event to be verified path. If the target anchor order is not satisfied, it enters the late event or the verification queue to be completed. If the event fingerprint already exists, it enters the attribute merging path. If the conflict mark exists, it is prohibited to write directly into the manufacturing execution data chain.

[0105] In a preferred embodiment, the manufacturing execution collaboration data package generates a write index after being written. The write index includes a manufacturing object index, an anchor point location index, an event fingerprint index, and an original source index. The manufacturing object index is used to find events along the work order, batch, and workpiece levels. The anchor point location index is used to find events along the sequence of hot air treatment start, hot air treatment end, cold riveting preparation, cold riveting execution completion, and quality confirmation. The event fingerprint index is used to identify duplicate reports and delayed reports. The original source index is used to trace back the original field names and field values ​​in the source system. The four types of indexes jointly support subsequent quality judgment and process traceability.

[0106] Furthermore, the maintenance of field semantic mapping relationships adopts a versioned mapping table. The versioned mapping table stores the mapping version, source system, original field name, field value range, field semantic category, applicable process anchor point, and effective time. When newly received heterogeneous manufacturing data hits multiple mapping versions, the applicable mapping version is determined based on the event generation time and parameter version. If the event generation time is within the overlapping range of multiple mapping versions, a second judgment is made by combining the source identifier and the status of adjacent processes. If the judgment still cannot determine the version, the event to be verified is written.

[0107] In this embodiment, the process parameter version serves as an important constraint for semantic judgment and event merging, but is not directly used as the basis for judging whether the process is qualified or not. The parameter versions of the hot air process and the cold riveting process are written into the event attribute set respectively. The parameter version is used as a contextual condition in field semantic category conflict, binding index completion, and delayed event merging. When the parameter version of the same manufacturing object changes between the end of hot air processing and the preparation of cold riveting, the system records the version change node and requires the subsequent cold riveting execution completion event to form a consistent record with the changed parameter version.

[0108] In this embodiment, the quality result identifier does not determine the sequential position of the manufacturing execution data chain independently, but is used as a closed-loop field in the quality confirmation anchor point and event attribute set. When the quality judgment record is delayed in being returned, the quality result identifier establishes a relationship with the stored cold riveting execution completion event through the association fingerprint. If the field semantic category, binding index and process anchor point are consistent and the quality result identifier is complementary, the quality result identifier is incorporated into the stored event attribute set. If the quality result identifier points to different manufacturing objects or different anchor points, it enters the conflict event set.

[0109] Preferably, when reprocessing records in the set of events to be verified and conflict events, the process is still executed in the order of field semantic mapping, process anchor verification, event fingerprint merging, and manufacturing execution data chain writing. It is not allowed to bypass any verification step and directly write to the manufacturing execution data chain. When reprocessing, the original entry time, reprocessing time, processing basis version, and processing result are saved so that the same event can maintain traceability of source and semantic comparability in multiple processing processes.

[0110] In this embodiment, the industrial data integration service can replay and verify the data chain of the same manufacturing object. The replay verification takes the manufacturing object identifier as input, and sequentially reads the chain nodes corresponding to the start of hot air processing, the end of hot air processing, cold riveting preparation, cold riveting execution completion, and quality confirmation. It recalculates the field semantic matching score, anchor point order difference, event fingerprint, and attribute completeness, and compares the recalculated results with the verification basis saved during writing. If the comparison results are consistent, the data chain state remains unchanged. If the comparison results are inconsistent, the difference nodes are written into the conflict event set.

[0111] Furthermore, the writing of the manufacturing execution data chain adopts idempotent processing. Idempotent processing uses event fingerprints as the core judgment basis. When the industrial data integration service receives process event objects with the same event fingerprint, it does not generate new chain nodes, but enters the attribute merging path or directly saves duplicate reporting records. When it receives process event objects with the same associated fingerprint but different event fingerprints, it does not directly write new nodes, but first performs a consistency comparison of field semantic categories, process anchors, binding indexes and quality result identifiers. After the comparison is completed, the merging or conflict handling path is determined.

[0112] In this embodiment, the operational logic of the hot air cold riveting production line manufacturing execution collaborative integration method can be deployed as a continuous software process of data access processing, semantic conversion processing, anchor point verification processing, fingerprint merging processing, and data chain writing processing. Data access processing is responsible for receiving heterogeneous manufacturing data and saving source records. Semantic conversion processing is responsible for generating process event objects and event attribute sets. Anchor point verification processing is responsible for determining the order and position of events. Fingerprint merging processing is responsible for handling duplicate reporting and delayed reporting. Data chain writing processing is responsible for encapsulating manufacturing execution collaborative data packets and writing them into the manufacturing execution data chain. Data is transferred between each process through process event objects, without relying on new production line hardware structures.

[0113] In this embodiment, the hot air process result, cold riveting execution status, and quality judgment result are entered into the same manufacturing execution data chain through the manufacturing object identifier. The field semantic mapping relationship avoids the same field being misjudged as having the same process meaning in different source systems. The process anchor point sequence avoids the cold riveting execution event being directly written into the confirmation chain when there is a lack of evidence for the end of hot air processing. The event fingerprint avoids the repeated reporting and generation of multiple co-occurring events. The association fingerprint avoids the delay in quality results being unable to be merged into the stored events. The binding index completion avoids data lacking hierarchical identifiers being directly discarded or incorrectly bound.

[0114] Preferably, when the manufacturing execution system needs to read the collaborative status of the current workpiece or batch, the industrial data integration service reads the last anchor point, last event fingerprint, event attribute set, and conflict flag under the corresponding manufacturing object identifier from the manufacturing execution data chain. If the last anchor point is the end of hot air processing and there are no records to be filled in the verification queue, a data packet reference that can enter the cold riveting preparation is output. If the last anchor point is the completion of cold riveting and there are delayed events with complementary quality result identifiers, a merged data packet reference is output. If there is a conflict flag, a conflict event set reference is output instead of an executable data packet reference.

[0115] In this embodiment, the storage format of process event objects can adopt structured key-value records. The key part includes manufacturing object identifier, target anchor point and event fingerprint, and the value part includes event attribute set, field semantic category, binding index, generation time, source identifier, parameter version, quality result identifier, merged record and conflict flag. Structured key-value records facilitate retrieval according to manufacturing object dimension, anchor point dimension and event fingerprint dimension, avoiding reliance on a single work order number or a single timestamp for locating manufacturing execution data.

[0116] Furthermore, the source priority is not fixedly sorted according to the source system name, but determined according to the reliable correspondence between the field semantic category and the source system. For example, the hot air processing status takes priority to the hot air process record source, the cold riveting execution status takes priority to the cold riveting execution record source, the quality judgment result takes priority to the quality judgment record source, and the process parameter version takes priority to the process parameter record source. When the source priority is the same but the attribute completeness is different, the main attribute record is selected according to the attribute completeness. When neither the source priority nor the attribute completeness can be distinguished, the corresponding field is written into the conflict event set.

[0117] In this embodiment, after the missing anchor candidate events in the verification queue to be completed are written into the manufacturing execution data chain, the target anchor does not immediately pass by default. Instead, the order verification is performed again based on the completed and recorded anchor queue. If the completion event makes the anchor order difference between the target anchor and the last anchor satisfy the continuity relationship, the event corresponding to the target anchor can continue to enter the event fingerprint merging path. If the parameter version or source path of the completion event is inconsistent with that of the target anchor, the target anchor and the completion event are jointly written into the conflict event set.

[0118] In this embodiment, the delayed event merging process retains time difference information. The generation time of the stored process event object, the generation time and the reception time of the newly received process event object are all stored in the merging record. The merging record does not change the original generation time of the stored events, but only adds supplementary attributes and merging basis to the event attribute set, so as to avoid the delayed reporting of quality results or parameter version records from changing the order position of the confirmed process events in the manufacturing execution data chain.

[0119] Preferably, when processing abnormal data, the industrial data integration service distinguishes between events to be verified, late events, missing events, and conflict events. Events to be verified indicate that the semantic category of a field has not yet been determined. Late events indicate that the event occurred within an existing process time period but the reception time is delayed. Missing events indicate that there are unreceived anchors in the process anchor sequence. Conflict events indicate that there are contradictions in the semantic category, anchor position, binding index, or quality result identifier that cannot be directly merged. Each event type enters a different temporary storage structure and retains a processing path, so that the data chain writing process is not disturbed by the processing logic of a single abnormal type.

[0120] In this embodiment, after the collaborative data packet is written, a data chain node verification record is generated. The data chain node verification record includes at least one of the following: field semantic matching score, binding credibility, anchor order difference, event fingerprint, association fingerprint, attribute completeness, and candidate consistency. The verification record is bound and saved with the chain node and is not used as external display data, but as an internal basis for subsequent replay verification, conflict investigation, and mapping relationship version adjustment.

[0121] Furthermore, if the source system field name changes, the industrial data integration service will not directly map the new field name to the target manufacturing execution field. Instead, it will first create a new field semantic mapping version and use historically confirmed process event objects as comparison samples. The status value, source identifier, and process anchor point corresponding to the new field name will be compared with the historical field semantic category. If the comparison result can stably fall into the same field semantic category, the new mapping version will be put into use. If the comparison result spans multiple field semantic categories, the new field name will be put into the event path to be verified.

[0122] In this embodiment, the manufacturing execution data chain can be maintained in parallel for different batches under the same work order. The data chains of each batch share the field semantic mapping relationship and the process anchor sequence, but the binding index, event fingerprint, association fingerprint and conflict event set are stored in isolation by batch. When a batch has a delayed reporting or conflict event, it does not affect the manufacturing execution data chains of other batches that have passed the verification, so that the hot air cold riveting manufacturing data of multiple batches under the same work order can be processed separately according to the manufacturing object boundary.

[0123] In this embodiment, when the workpiece transfer identifier is missing but the work order number and batch number exist, the industrial data integration service does not directly summarize and write by batch. Instead, it reads the status and quality result identifiers of adjacent processes under the same batch, searches for candidate workpiece transfer identifiers that are in the same process time period as the original record generation time. If the candidate workpiece transfer identifier is unique and passes the hierarchical consistency check, a completed three-level binding index is generated. If there are multiple candidate workpiece transfer identifiers, the original record is retained and written to the event to be checked.

[0124] In this embodiment, the data merging of quality confirmation anchor points uses the quality result identifier as the main supplementary field. If the quality judgment record arrives before the cold riveting completion record, the industrial data integration service will generate a process event object with the quality judgment record as the target anchor point for quality confirmation. Due to the lack of intermediate anchor points, it will be written into the pending verification queue. After the cold riveting completion event arrives and passes the anchor point verification, the quality confirmation event in the pending verification queue will be read for sequential consistency verification and associated fingerprint merging.

[0125] Preferably, the processing of process parameter versions follows the principle of prioritizing the event generation time. If there are multiple parameter version records for the same manufacturing object, the industrial data integration service establishes a correlation between the effective time of the process parameter version and the generation time of the process event object. Events whose generation time falls within the effective range of a certain parameter version inherit the corresponding parameter version. If the generation time overlaps with the boundary of two parameter versions, the source identifier and the status of adjacent processes are read for a second judgment. If it is still uncertain, it enters the conflict event set.

[0126] In this embodiment, the manufacturing execution collaboration data package output by the industrial data integration service does not directly expose all internal verification processes. Instead, it saves the data package identifier, manufacturing object identifier, current anchor point, previous anchor point reference, event attribute set summary, merging basis summary, and conflict status. The manufacturing execution system can read the current process status and traceability index based on the data package identifier. The quality judgment system can read the corresponding process event object based on the manufacturing object identifier and quality result identifier. The process parameter recording system can read the associated events based on the parameter version index.

[0127] In this embodiment, if the hot air processing completion event, cold riveting preparation event, and cold riveting execution completion event have all passed verification, but the quality confirmation event has not yet arrived, the manufacturing execution data chain is saved to the cold riveting execution completion anchor point. After the quality confirmation event arrives, a relationship is established with the cold riveting execution completion event through the associated fingerprint. If the quality result identifier is complementary and the binding index is consistent, the quality confirmation event is written to the subsequent anchor point and the chain node reference is updated. If the quality result identifier points to other manufacturing objects, it is written to the conflict event set.

[0128] Furthermore, the hot air cold riveting production line manufacturing execution collaborative integration method can form a closed processing path at the software level, from data reception, semantic conversion, hierarchical binding, anchor point verification, fingerprint merging, missing data completion, conflict temporary storage to data chain writing. Each processing path is bounded by the manufacturing object identifier, with the process event object as the basic processing unit, and with the field semantic category, process anchor point, and event fingerprint as the core constraints. This allows the hot air process results, cold riveting execution status, process parameter version, and quality judgment results to be saved in the manufacturing execution data chain according to the same object and the same process sequence.

Claims

1. A collaborative integration method for manufacturing execution in a hot air cold riveting production line, characterized in that, include: The industrial data integration service receives heterogeneous manufacturing data from hot air processes, cold riveting processes, batch flow, work order status, process parameters, and quality judgment. Based on a preset process event semantic model, the heterogeneous manufacturing data is converted into process event objects containing manufacturing object identifiers, process status, time identifiers, and source identifiers; The process meaning of the process event object is determined according to the field semantic mapping relationship; the sequential position of the process event object in the hot air cold riveting continuous process is verified according to the process anchor sequence; and the process event objects that are repeatedly reported and delayed are merged according to the event fingerprint. The validated and merged process event objects are encapsulated into a manufacturing execution collaboration data package and written into the manufacturing execution data chain.

2. The collaborative integration method for manufacturing execution in a hot air cold riveting production line according to claim 1, characterized in that, The step of converting the heterogeneous manufacturing data into process event objects based on a preset process event semantic model includes: extracting the work order number, batch number, workpiece flow identifier, process name, status value, generation time, parameter version, and quality result identifier from the heterogeneous manufacturing data; Establish binding indexes between work orders, batches, and workpieces according to the manufacturing object hierarchy; The binding index is combined with the process name, status value, and generation time to form the event primary key; Map the source system, field name, and field value to the event attribute set under the event primary key.

3. The collaborative integration method for manufacturing execution in a hot air cold riveting production line according to claim 1, characterized in that, The step of determining the process meaning of the process event object based on the field semantic mapping relationship includes: setting field semantic categories for hot air treatment status, cold riveting execution status, process parameter version and quality judgment result respectively; Map the original field names from the source system to the semantic categories of the fields; When the same original field name corresponds to multiple field semantic categories, the target field semantic category is determined based on the manufacturing object identifier, process status, and source identifier of the process event object. Process event objects for which the semantic category of the target field could not be determined are marked as events to be verified.

4. The collaborative integration method for manufacturing execution in a hot air cold riveting production line according to claim 1, characterized in that, The step of verifying the sequential position of the process event object in the hot air cold riveting continuous process according to the process anchor point sequence includes: establishing a process anchor point sequence including hot air treatment start, hot air treatment end, cold riveting preparation, cold riveting execution completion and quality confirmation; Match the process state corresponding to the process event object to the target anchor point in the process anchor point sequence; Generate sequence verification results based on the order of recorded anchor points and target anchor points under the same manufacturing object identifier; Process event objects whose sequence of operations does not satisfy the process anchor point sequence are written into the conflict event set.

5. The collaborative integration method for manufacturing execution in a hot air cold riveting production line according to claim 2, characterized in that, The method of establishing a binding index between work orders, batches, and workpieces according to the hierarchical relationship of manufacturing objects includes: using the work order number as a first-level index, the batch number as a second-level index, and the workpiece transfer identifier as a third-level index. When the heterogeneous manufacturing data lacks any level identifier, a candidate binding index is searched based on the status of adjacent processes, generation time, and source system in the received process event objects. After performing hierarchical consistency verification on the candidate binding indexes, the completed binding indexes are generated and the basis for completion is recorded.

6. The collaborative integration method for manufacturing execution in a hot air cold riveting production line according to claim 3, characterized in that, Determine the semantic category of the target field based on the manufacturing object identifier, process status, and source identifier of the process event object, including: when there is a conflict in the semantic category of the field, read the process status of adjacent process event objects under the same manufacturing object identifier; The process status of the adjacent process event objects is compared with the semantic category of the field corresponding to the source identifier. When the combined comparison result is consistent with the state transition relationship in the semantic model of the hot air cold riveting process event, the semantic category of the corresponding field is determined as the semantic category of the target field.

7. The collaborative integration method for manufacturing execution in a hot air cold riveting production line according to claim 4, characterized in that, The step of generating a sequence verification result based on the sequential relationship between the recorded anchor points and the target anchor points under the same manufacturing object identifier includes: reading the recorded anchor point queue according to the manufacturing object identifier; The target anchor point is compared with the last anchor point in the recorded anchor point queue. When the target anchor point is earlier than the last anchor point, it is determined whether the event is late based on the generation time and event fingerprint of the process event object. When the target anchor point is later than the last anchor point and intermediate anchor points are missing, the missing anchor point and the target anchor point are written together into the pending completion verification queue.

8. The collaborative integration method for manufacturing execution in a hot air cold riveting production line according to claim 1, characterized in that, The process of merging duplicate and delayed reported process event objects based on event fingerprints includes: generating event fingerprints based on manufacturing object identifier, process status, parameter version, source identifier, and generation time; Match the newly generated event fingerprint with the already stored event fingerprint; When event fingerprints are the same but event attribute sets differ, merge event attribute sets according to source priority and attribute completeness. When event fingerprints are different but process anchors and binding indexes are the same and the generation time is within the same process time period, associated fingerprints are generated and the delayed event merging process is initiated.

9. The collaborative integration method for manufacturing execution in a hot air cold riveting production line according to claim 7, characterized in that, After writing the missing anchor point and the target anchor point together into the queue to be filled in the verification, it also includes: reading the process event objects located before and after the missing anchor point under the same manufacturing object identifier from the manufacturing execution data chain; Generate candidate events for missing anchor points based on the field semantic categories, parameter versions, and source identifiers of the preceding and following process event objects; Perform a sequence consistency check between the missing anchor point candidate events and the target anchor points in the queue to be filled in; Missing anchor candidate events that pass the verification are written into the manufacturing execution data chain.

10. The collaborative integration method for manufacturing execution in a hot air cold riveting production line according to claim 8, characterized in that, The process of generating associated fingerprints and entering the delayed event merging process includes: reading stored process event objects and newly received process event objects according to the associated fingerprints; Compare the consistency between the stored process event object and the newly received process event object in terms of field semantic category, process anchor, binding index, and quality result identifier; When the semantic category of the field and the binding index are consistent and the quality result identifiers are complementary, the newly received process event object will be merged into the already stored process event object; When process anchor points are inconsistent, the process event object corresponding to the associated fingerprint will be transferred to the conflict event set.

Images