A PLM persistent storage and semantic reasoning method for CAD / CAE / CAM collaboration

Abstract

The application discloses a PLM persistent storage and semantic reasoning method for CAD / CAE / CAM collaboration, and the method comprises the following steps: obtaining a submission request, writing the content of the submission request into a content storage according to the content hash value after the content is divided and hashed, generating an expression record and recording the expression record into a cross-domain expression set, and pointing to an old expression record through a derived pointer; generating a semantic annotation record and an explicit semantic relationship record based on a target object identifier contained in the submission request; performing consistency verification, and performing incremental ontology reasoning and rule reasoning based on a semantic subgraph range limited by a change set after the verification is passed, and generating a derived semantic relationship and an explanation chain; generating a submission record, appending the submission record to a metadata storage, updating an index, and performing affected analysis according to an expression dependency graph, generating a to-be-recomputed or to-be-repaired mark, publishing a change event, and triggering a collaborative response action. The application has the advantages that the cost of cross-domain collaboration interruption chain rework, storage redundancy and manual intervention is significantly reduced.

Description

Technical Field

[0001] This application relates to the field of collaborative engineering software, specifically to a PLM persistent storage and semantic reasoning method for CAD / CAE / CAM collaboration. Background Technology

[0002] In the integrated process of product design, simulation, and manufacturing, Computer-Aided Design (CAD) is used for geometric modeling and assembly design, Computer-Aided Engineering Analysis (CAE) is used for mesh generation, material and boundary condition definition, and solution analysis, and Computer-Aided Manufacturing (CAM) is used for process planning, tool / toolpath calculation, and CNC output. With the advancement of intelligent manufacturing and digital engineering, efficient collaboration among CAD / CAE / CAM has become a key link in shortening product development cycles and improving manufacturing quality. However, existing collaborative solutions still have many shortcomings in terms of data structure, association stability, traceability and auditing, storage efficiency, and semantic consistency, specifically manifested in the following ways:

[0003] Existing PLM systems typically employ heterogeneous storage, with CAD assembly structures, CAE calculation structures, and CAM process structures stored in different data models or files, lacking a unified representation framework. When design changes occur, data synchronization between CAD and CAE / CAM often relies on manual export / import or script conversion, easily leading to data inconsistencies and deviations. Furthermore, the lack of explicit records of reference relationships between different domains makes cross-domain traceability difficult.

[0004] Loads and constraints in CAE analysis, as well as machining areas in CAM, are typically directly bound to the instantaneous IDs (such as face IDs, edge IDs, or feature names) of topological entities in the CAD model. When CAD engineers iterate on the model (e.g., reorder features, modify parameters, or reconstruct geometry), the IDs of these topological entities often change, leading to the loss of CAE constraints and the failure of CAM machining areas. Engineers must manually rebind and recalculate, resulting in repeated rework and severely impacting project timelines and data quality. This problem is particularly prominent in scenarios involving complex assemblies and frequent design changes.

[0005] Existing collaborative solutions mostly use overlay updates, making it difficult to form a reliable version diagram and derivation chain. The derivation relationship from CAD version to CAE mesh version and CAM toolpath version lacks explicit records. When quality issues or analysis anomalies occur, it is difficult to trace the source of the problem, resulting in high costs for problem reproduction, responsibility determination, and compliance audits.

[0006] In cross-domain collaboration, CAD geometry, CAE meshes, CAM toolpaths, and other representations are often repeatedly stored in different systems or directories, lacking content addressing and deduplication mechanisms. The same geometric model is stored multiple times in different projects or versions, causing storage space bloat. In network transmission scenarios, the lack of incremental synchronization and breakpoint resume capabilities based on content chunking leads to bandwidth waste and low synchronization efficiency.

[0007] The same engineering intent (such as "sealing surface," "supporting surface," and "high roughness area") is expressed using different terms and structures in CAD features, CAE boundary conditions, and CAM processes, making cross-domain semantic alignment difficult. Existing systems lack a unified engineering semantic ontology, making semantic-level verification and reasoning impossible. This hinders stable support for automated configuration, intelligent recommendation, and interpretable analysis, limiting the deep application of artificial intelligence technology in product development processes.

[0008] In summary, existing CAD / CAE / CAM collaborative solutions have significant shortcomings in terms of data structure, association stability, traceability and auditing, storage efficiency, and semantic uniformity. Therefore, there is an urgent need for a PLM persistence solution that can uniformly represent CAD / CAE / CAM data, stably maintain cross-domain associations, support version traceability and auditing, achieve storage deduplication and efficient transmission, and provide semantic uniformity and interpretable reasoning capabilities. Summary of the Invention

[0009] This application provides a PLM persistent storage and semantic reasoning method for CAD / CAE / CAM collaboration to solve the technical problem that the association between CAE constraints and CAM processing areas is broken due to CAD version iteration, requiring a lot of rework and rebinding and repeated recalculation.

[0010] This application provides a PLM persistent storage and semantic reasoning method for CAD / CAE / CAM collaboration, specifically including steps for obtaining submission requests, writing content, generating expression records, generating semantic annotations, verifying consistency, inferential reasoning, generating submission records, and analyzing the impact.

[0011] The step of obtaining a submission request is used to obtain a submission request from the client. The submission request includes at least a target object identifier and the expression content to be written. The content writing step performs block partitioning and hash calculation on the expression content to be written, generates a content hash value, and writes the expression content to be written into the content storage according to the content hash value. The expression record generation step is used to generate an expression record, which is recorded in a cross-domain expression set and points to the derived old expression record through a derived pointer. The semantic annotation generation step is used to obtain the object identifier, version identifier, instance identifier, and expression identifier associated with the target object identifier, and maps the object identifier, version identifier, instance identifier, expression identifier, or attribute item to the concept identifier in the project semantic ontology to generate semantic annotations. The process involves several steps: First, defining the record and explicit semantic relationship record. Second, the consistency verification step verifies the records and their relationships involved in the submission request according to a consistency verification rule set. If the verification is successful, the next step is executed. If the verification fails, the generation of a publishable submission record is prevented, or the corresponding object is marked as pending repair. Third, the incremental reasoning step performs incremental ontology reasoning and rule reasoning based on a limited semantic subgraph scope of the change set, generating deduced semantic relationship records and explanation chain records, and storing them permanently. Fourth, the submission record generation step generates submission records and appends them to metadata storage, updates index storage and semantic index storage. Fifth, the affected analysis step performs affected analysis based on the submission records and the expression dependency graph, generates recalculation markers or repair markers, publishes change events, and triggers collaborative response actions.

[0012] Furthermore, after the expression record generation step, an anchor point mapping record acquisition step is also included. The anchor point mapping record acquisition step is to acquire the anchor point set record when the submission request contains anchor point definitions, parse the candidate entity set in the expression record, perform selection expression parsing on each anchor point, and filter the candidate entity set using geometric predicates and attribute constraints to obtain entity identifiers, and generate an anchor point mapping record corresponding to the expression record.

[0013] Furthermore, the expression record generation step includes: incorporating metadata updates, anchor mapping updates, semantic relationship updates, and semantic index updates into the same atomic commit unit; after the atomic commit unit is completed, ensuring that the data visible to the external query end simultaneously satisfies structural consistency and semantic consistency.

[0014] Further, the semantic annotation generation step includes: mapping the object identifier to the object class concept identifier in the project semantic ontology; mapping the version identifier to the version class concept identifier in the project semantic ontology; mapping the instance identifier to the occurrence class concept identifier in the project semantic ontology; mapping the expression identifier to the expression class concept identifier in the project semantic ontology; mapping the anchor identifier to the anchor class concept identifier in the project semantic ontology; mapping the attribute item to the attribute class concept identifier in the project semantic ontology; generating the semantic annotation record, recording the mapping relationship between each identifier and the concept identifier; generating the explicit semantic relationship record, recording the semantic association between each identifier, and marking the explicit semantic relationship record as an assertion type.

[0015] Furthermore, the consistency verification steps include: performing reference integrity verification to verify the reachability of version reference records pointing to cross-domain expression set records, the reachability of cross-domain expression set records pointing to expression records, the reachability of anchor set records pointing to anchors, and the reachability of anchors pointing to anchor mapping records; performing mode version compatibility verification to verify whether the mode version written by the client matches the registry; performing unit consistency verification to verify whether the cross-domain physical quantity units are convertible and conform to the preset strategy; performing coordinate system consistency verification to verify whether the coordinate reference is convertible and conforms to the preset strategy; performing anchor point resolvability verification to verify whether key anchors are successfully resolved or are in an allowed pending repair state; and performing semantic consistency verification based on ontology constraints to verify whether the boundary condition type matches the type of the acting entity, whether the process type matches the material requirements, and whether the anchor point semantic type matches the semantic type of the resolved entity.

[0016] Furthermore, the incremental reasoning step includes: determining the set of affected semantic nodes based on the change set; performing reasoning closure calculation on the set of affected semantic nodes to generate inferred semantic relations; generating an explanation chain summary corresponding to the inferred semantic relations and writing it into the reasoning result record.

[0017] Furthermore, the commit record generation step includes: incorporating metadata updates, anchor mapping updates, semantic relationship updates, and semantic index updates into the same atomic commit unit; after the atomic commit unit is completed, ensuring that the data visible to the external query end simultaneously satisfies structural consistency and semantic consistency.

[0018] Furthermore, the affected analysis steps include: performing graph traversal along the derived pointer chain and the expression dependency graph to determine the minimum recalculation closure set of CAE expression records and CAM expression records affected by the CAD expression record update; generating a task queue and marking the expression records in the minimum recalculation closure set as pending recalculation; and publishing a change event to the corresponding domain client to trigger the automatic recalculation process.

[0019] Furthermore, the PLM persistent storage and semantic reasoning method for CAD / CAE / CAM collaboration also includes a copy-on-write step, which includes: generating new version records for version reference records and instance records on the affected path based on the copy-on-write strategy; reusing old version records for unaffected structures through predecessor pointers or references; and pointing the newly generated commit records to the old commit records through parent pointers to form a version graph.

[0020] Furthermore, the PLM persistent storage and semantic reasoning method for CAD / CAE / CAM collaboration also includes: encoding the change set into incremental transmission packets; and, during network transmission, performing missing block filling and breakpoint resumption on the incremental transmission packets based on the block hash value.

[0021] This application provides a PLM persistent storage and semantic reasoning method for CAD / CAE / CAM collaboration. It replaces instantaneous topology numbering with semantic anchors, automatically rebinding CAE constraints and CAM processing areas when the CAD model changes, reducing rework due to broken links from hours to seconds of automated processing. Through content hashing for deduplication and incremental transmission, the amount of cross-domain synchronized data is reduced to less than 10% of the original data volume, significantly reducing storage and bandwidth consumption. Immutable records and version graphs enable full traceability of the collaborative process, meeting the auditing requirements of highly compliant industries. Automatic recalculation of minimum closures driven by expression dependency graphs reduces the frequency of human-computer interaction in cross-domain collaboration by more than 80%. Furthermore, by combining engineering semantic ontology and incremental reasoning, it provides an auditable reasoning foundation for intelligent verification, automatic configuration, and process recommendation, achieving quantifiable technical effects in areas such as broken link repair, storage and transmission, traceability auditing, automatic recalculation, and intelligent applications. Attached Figure Description

[0022] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0023] Figure 1 This is the flow of the PLM persistent storage and semantic reasoning method for CAD / CAE / CAM collaboration described in the embodiments of this application. Figure 1 ;

[0024] Figure 2 This is a schematic diagram of the overall architecture of the PLM persistent storage and semantic reasoning method for CAD / CAE / CAM collaboration described in the embodiments of this application;

[0025] Figure 3This is the flow of the PLM persistent storage and semantic reasoning method for CAD / CAE / CAM collaboration described in the embodiments of this application. Figure 2 ;

[0026] Figure 4 This is the flow of the PLM persistent storage and semantic reasoning method for CAD / CAE / CAM collaboration described in the embodiments of this application. Figure 3 . Detailed Implementation

[0027] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0028] Existing technologies in cross-domain collaboration for product lifecycle management suffer from problems such as fragmented data structures, poor cross-domain consistency, frequent rework due to topology changes leading to broken links between CAE constraints and CAM processing areas, difficulties in version tracing and auditing, redundant storage and inefficient transmission, and inconsistent semantics lacking interpretable reasoning capabilities. This application aims to fundamentally solve practical problems such as broken links and rework, storage redundancy, difficulty in tracing, and semantic isolation in cross-domain collaboration by using a unified three-layer PLM object persistence structure, an automatic rebinding mechanism for semantic anchors, content addressing deduplication and incremental transmission, immutable version graphs and expression dependency graphs, and engineering semantic ontology and incremental reasoning. This will achieve intelligent collaboration with unified structure, stable associations, traceability and auditability, efficient storage and transmission, and interpretable reasoning capabilities.

[0029] To achieve the above-mentioned technical effects, this application includes the following technical content.

[0030] like Figure 1 as well as Figure 2As shown, in the submission chain of this application, after the CAD client, CAE client, or CAM client is authenticated by the access layer and authentication module, the submission request enters the submission processing module. The submission processing module calls the cross-domain adaptation module to generate expression records, calls the anchor point parsing module to generate anchor point mappings, calls the unified pattern registration and consistency management module to perform verification, and calls the ontology management and semantic reasoning module to generate semantic annotations and derive semantic relationships. Subsequently, metadata is appended to the metadata storage, expression content is written to the content storage, the index is updated to the index storage, and semantic relationships are written to the semantic graph storage and semantic index storage. Finally, the affected analysis module is triggered to generate tasks and write them to the task queue, and the event distribution module publishes change events to the client. In the query chain, after the client is authenticated by the access layer and authentication module, the query and traversal module first accesses the cache module. If no match is found, the record is located through the index storage and semantic index storage, and then the data is read from the metadata storage and content storage and returned.

[0031] Specifically, in the above submission chain, this application provides a PLM persistent storage and semantic reasoning method for CAD / CAE / CAM collaboration, which specifically includes step S1) obtaining the submission request step, step S2) content writing step, step S3) expression record generation step, step S4) semantic annotation generation step, step S5) consistency verification step, step S6) incremental reasoning step, step S7) submission record generation step, and step S8) affected analysis step.

[0032] Step S1) Obtaining the submission request step: Obtain the submission request from the client. The submission request includes at least the target object identifier and the content to be written.

[0033] In this application, the client includes a CAD client, a CAE client, and a CAM client; the target object identifier includes at least one of an object identifier, a version identifier, or an instance identifier, used to point to the identity, version revision, or instance occurrence of the product main object involved in this submission; the expression content to be written includes at least one of CAD geometric expression, CAE mesh expression, CAE case expression, CAM process expression, or CAM toolpath expression, and the expression content to be written is encapsulated in binary stream or structured data form; when the submission request is used to establish or update cross-domain associations, the submission request also includes an anchor point definition, the anchor point definition includes an anchor point identifier and a selection expression, and the selection expression is used to describe the geometric entity, mesh entity, or process feature to be bound in an engineering intent manner.

[0034] Step S2) Content writing step: Perform block division and hash calculation on the content to be written to generate a content hash value, and write the content to be written to the content storage according to the content hash value.

[0035] In one feasible embodiment, if the same content hash value already exists in the content storage, the stored content is reused to achieve deduplication.

[0036] In this application, performing block partitioning and hash calculation on the expression content to be written further includes: dividing the expression content to be written into multiple data blocks of a fixed size, calculating a block hash value for each data block, and organizing the block hash values ​​of the multiple data blocks into a hash list, wherein the content hash value is the Merkle root or combined digest of the hash list; the content storage is an object storage or block storage addressed by the content hash value, and each expression content is written and read using the content hash value as a unique storage key; when the same content hash value already exists in the content storage, the content writing operation is skipped, and the reference address of the stored content is returned to achieve content-level deduplication across versions and expression records; when the same content hash value does not exist in the content storage, each data block of the expression content to be written is written to the storage according to the block hash value, and the mapping relationship between the content hash value and each block hash value is recorded to support incremental transmission and breakpoint resume.

[0037] Its technical advantages lie in generating content hash values ​​by performing block-based hash calculations on the content to be written, and then writing the content to the content storage according to the content hash values. When the same content hash value already exists in the content storage, the stored content is directly reused, thereby achieving content-level deduplication of cross-version, cross-project, and cross-domain expressions at the expression record level, avoiding the repeated storage of the same geometry, mesh, or toolpath data. At the same time, the storage organization method based on block hash values ​​supports missing block completion and breakpoint resumption in the form of incremental transmission packets during network transmission, compressing the amount of cross-domain synchronized data to less than 10% of the original data volume, significantly reducing storage space occupation and network bandwidth consumption, and achieving efficient storage and transmission collaboration.

[0038] Step S3) Expression record generation step: Generate expression records, which are recorded to cross-domain expression sets and pointed to by derived pointers.

[0039] Specifically, step S3) expresses the record generation step, which includes: incorporating metadata update, anchor mapping update, semantic relationship update and semantic index update into the same atomic commit unit; after the atomic commit unit is completed, the data visible to the external query end simultaneously satisfies structural consistency and semantic consistency.

[0040] In this application, an expression identifier is assigned to the expression record, and fields such as domain type, format, mode version, content hash value, unit configuration, and coordinate reference are set. The domain type is selected from one of CAD domain, CAE domain, or CAM domain, and the content hash value points to the expression content that has been stored or reused in step S2.

[0041] Incorporating the expression record into the cross-domain expression set record includes: obtaining the version reference record pointed to by the submission request, locating the current cross-domain expression set record through the version reference record, generating a new cross-domain expression set record based on the copy-on-write strategy, adding the newly generated expression record reference to the expression reference set of the new cross-domain expression set record, and reusing the expression references in the old cross-domain expression set record for expression records that have not changed.

[0042] Pointing to the derived old expression record via a derived pointer includes: when the expression record is a modification or derivation of an existing expression record, setting the derived pointer to the expression identifier of the derived old expression record to form a derivation chain between expression records, supporting the construction of cross-domain expression dependency graphs and impact analysis.

[0043] The expression record generation step, metadata update, anchor mapping update, semantic relationship update, and semantic index update are incorporated into the same atomic commit unit. After the atomic commit unit is completed, the data visible to the external query end simultaneously satisfies structural consistency and semantic consistency.

[0044] Its technical effectiveness lies in integrating expression record generation, metadata updates, anchor mapping updates, semantic relationship updates, and semantic index updates into the same atomic commit unit. This ensures that the data visible to the external query end at any given time simultaneously satisfies both structural and semantic consistency, fundamentally avoiding intermediate inconsistencies such as "CAD expression updated but anchor mapping not updated" or "semantic relationship stored but semantic index not updated." Simultaneously, it generates new cross-domain expression set records based on a copy-on-write strategy and shares the structure of unchanged expression records, significantly reducing write amplification and storage overhead. Furthermore, it explicitly records the derivation relationships between expression records through derived pointers, forming a traversable derivation chain. This provides a structured dependency tracking foundation for subsequent expression dependency graph construction, impact analysis, and automatic recalculation, elevating data consistency assurance, storage efficiency, and traceability in cross-domain collaboration to a transaction-level reliability level.

[0045] like Figure 3 As shown, in one feasible embodiment, after the expression record generation step S3), the method further includes step S31) anchor point mapping record acquisition step. When the submission request contains anchor point definitions, the anchor point set record is acquired and the candidate entity set is obtained by parsing the expression record. For each anchor point, the selection expression is parsed and the candidate entity set is filtered using geometric predicates and attribute constraints to obtain entity identifiers and generate an anchor point mapping record corresponding to the expression record.

[0046] In this application, when the submission request includes anchor point definitions, a new anchor point set record is generated or an existing anchor point set record is updated. The anchor point set record includes at least one anchor point definition, and each anchor point definition includes at least an anchor point identifier and a selection expression. The selection expression is used to describe the geometric entity, mesh entity, or process feature to be bound in an engineering intent manner.

[0047] Parsing the expression record to obtain a candidate entity set includes: applying the selection expression in the anchor point definition to the expression record of a specified domain type, including CAD expression record, CAE expression record, or CAM expression record; locating candidate entities that conform to the engineering intent description in the expression record by executing the selection expression; and forming a candidate entity set.

[0048] The process of parsing the selection expression for each anchor point and filtering the candidate entity set using geometric predicates and attribute constraints includes: when the candidate entity set contains multiple candidate entities, filtering the candidate entity set using geometric predicates and attribute constraints, wherein the geometric predicates include at least one of normal constraint, curvature constraint, and adjacency constraint, and the attribute constraints include at least one of feature path matching, attribute label matching, and parameter constraint matching, thereby obtaining a unique or preferred entity identifier through filtering.

[0049] Obtaining entity identifiers includes: using the filtered entity identifiers as anchors in the parsing results of the expression record, calculating the parsing confidence and recording the parsing status, wherein the parsing status includes at least one of parsing success, parsing failure requiring manual repair, or parsing ambiguity requiring confirmation.

[0050] Generating an anchor mapping record corresponding to the expression record includes: generating an anchor mapping record for each anchor point, wherein the anchor mapping record includes at least an anchor point identifier, an expression identifier, an entity identifier, a resolution confidence level, and a resolution status. When the resolution status is "resolution failed", the reason for the failure is recorded to trigger a manual or rule-based repair process.

[0051] Its technical advantages lie in the fact that by selecting an expression to describe the entity to be bound in an engineering intent manner, and parsing the candidate entity set in the specified expression record, multi-dimensional filtering is performed by combining geometric predicates and attribute constraints to ensure that the target entity can still be stably located after topology reconstruction or feature renumbering. At the same time, by recording the parsing confidence and parsing status, the reason for the failure is identified when parsing fails and a manual or rule-based repair process is triggered. This reduces the rework caused by topology changes due to broken links from several hours to second-level automated rebinding, significantly reducing the cost of manual intervention and collaboration waiting time in cross-domain collaboration.

[0052] Step S4) Semantic annotation generation step: Obtain the object identifier, version identifier, instance identifier and expression identifier associated with the target object identifier, map the object identifier, version identifier, instance identifier, expression identifier or attribute item to the concept identifier in the engineering semantic ontology, and generate semantic annotation records and explicit semantic relationship records.

[0053] In this application, when the target object identifier is an object identifier, it is directly used as the object identifier, and the currently effective version identifier is queried through the object identifier. The associated instance identifier and expression identifier are obtained through the version identifier. When the target object identifier is a version identifier, it is directly used as the version identifier, and the object identifier is retrieved through the version identifier. The associated instance identifier and expression identifier are obtained through the version identifier. When the target object identifier is an instance identifier, it is directly used as the instance identifier, and the version identifier is retrieved through the instance identifier. The object identifier is retrieved through the version identifier. The associated expression identifier is obtained through the version identifier.

[0054] Specifically, step S4) the semantic annotation generation step includes: mapping the object identifier to the object class concept identifier in the project semantic ontology; mapping the version identifier to the version class concept identifier in the project semantic ontology; mapping the instance identifier to the occurrence class concept identifier in the project semantic ontology; mapping the expression identifier to the expression class concept identifier in the project semantic ontology; mapping the anchor identifier to the anchor class concept identifier in the project semantic ontology; mapping the attribute item to the attribute class concept identifier in the project semantic ontology; generating the semantic annotation record, recording the mapping relationship between each identifier and the concept identifier; generating the explicit semantic relationship record, recording the semantic association between each identifier, and marking the explicit semantic relationship record as an assertion type.

[0055] In this application, mapping the object identifier to the object class concept identifier in the engineering semantic ontology includes: locating the corresponding object class concept in the engineering semantic ontology based on the type classification and lifecycle metadata of the object identifier, wherein the object class concept includes at least one of the following: part class, assembly class, standard part class, or material class.

[0056] Mapping the version identifier to the version class concept identifier in the project semantic ontology includes: locating the corresponding version class concept in the project semantic ontology based on the status and validity information of the version identifier, wherein the version class concept includes at least one of working version, release version, frozen version, or obsolete version.

[0057] Mapping the instance identifier to the occurrence class concept identifier in the engineering semantic ontology includes: locating the corresponding occurrence class concept in the engineering semantic ontology based on the context type and configuration expression of the instance identifier, wherein the occurrence class concept includes at least one of assembly occurrence, process occurrence, or simulation occurrence.

[0058] Mapping the expression identifier to the expression class concept identifier in the engineering semantic ontology includes: locating the corresponding expression class concept in the engineering semantic ontology according to the domain type and format of the expression identifier, wherein the expression class concept includes at least one of CAD geometric expression, CAE mesh expression, CAE example expression, CAM process expression, or CAM toolpath expression.

[0059] Mapping anchor point identifiers to anchor point class concept identifiers in the engineering semantic ontology includes: when the submission request contains an anchor point definition or has an associated anchor point identifier, locating the corresponding anchor point class concept in the engineering semantic ontology based on the selection expression type of the anchor point identifier and the engineering intent semantics, wherein the anchor point class concept includes at least one of the following: load application surface class, constraint application surface class, processing area class, or measurement reference class.

[0060] Mapping attribute items to attribute class concept identifiers in the engineering semantic ontology includes: locating the corresponding attribute class concept in the engineering semantic ontology based on the name, data type, unit, and semantic tag of the attribute item. The attribute class concept includes at least one of physical attribute class, geometric attribute class, material attribute class, or process parameter class.

[0061] Generating the semantic annotation record includes: generating at least one semantic annotation record, each semantic annotation record containing at least an annotation identifier, an annotated identifier, a concept identifier, annotation type and scope information, wherein the annotated identifier includes one of object identifier, version identifier, instance identifier, expression identifier or anchor identifier, and the annotation type includes at least one of type annotation, instance annotation or attribute annotation.

[0062] Generating the explicit semantic relation record includes: generating at least one explicit semantic relation record based on the association between the object identifier, the version identifier, the instance identifier, the expression identifier, or the anchor identifier. Each explicit semantic relation record includes at least a relation identifier, a subject identifier, a predicate identifier, an object identifier, relation type, and scope information. The explicit semantic relation record is marked as an assertion type to distinguish it from the inferred semantic relation record generated by subsequent incremental reasoning.

[0063] For step S4, the technical effect is that by uniformly mapping object identifiers, version identifiers, instance identifiers, expression identifiers, anchor identifiers, and attribute items to concept identifiers in the engineering semantic ontology, a unified semantic association is established between the object layer, version layer, occurrence layer, expression layer, anchor layer, and attribute layer, forming structured semantic annotation records and explicit semantic relationship records. This fundamentally solves the problem of terminology fragmentation and semantic inconsistency between CAD features, CAE boundary conditions, and CAM processes. By marking explicit semantic relationships as assertion types, a clear distinction is formed between them and the inferential semantic relationships generated by subsequent incremental reasoning. This provides a structured semantic foundation for auditable and reproducible intelligent reasoning, supporting upper-level intelligent applications such as intelligent verification, automatic configuration, process recommendation, and engineering decision support.

[0064] Step S5) Consistency verification step: The records involved in the submission request and their associated relationships are verified according to the consistency verification rule set. If the verification is successful, the next step is executed; if the verification fails, the generation of publishable submission records is prevented or the corresponding objects are marked as pending repair.

[0065] Step S5) The consistency verification step includes: performing reference integrity verification, verifying the reachability of the version reference record to the cross-domain expression set record, the reachability of the cross-domain expression set record to the expression record, the reachability of the anchor set record to the anchor, and the reachability of the anchor to the anchor mapping record.

[0066] Specifically, the system verifies the reachability of version reference records pointing to cross-domain expression set records, ensuring that the cross-domain expression set identifier in the version reference record has a corresponding cross-domain expression set record in storage; it also verifies the reachability of cross-domain expression set records pointing to expression records, ensuring that each expression identifier in the expression reference set of the cross-domain expression set record has a corresponding expression record in storage; it verifies the reachability of anchor set records pointing to anchors, ensuring that the anchor identifier in the anchor set record has a corresponding anchor definition in storage; and it verifies the reachability of anchors pointing to anchor mapping records, ensuring that the anchor identifier has a corresponding anchor mapping record in the specified expression record. When any reference is unreachable, the system determines that the reference integrity check has failed, records the missing reference path and blocks the commit, or marks the affected object as pending repair.

[0067] Step S5) The consistency verification step includes: performing a mode version compatibility verification to verify whether the mode version written by the client matches the registry.

[0068] Specifically, the pattern version field in the expression record is obtained and compared with the pattern version list registered in the unified pattern registration and consistency management module to verify that the pattern version written by the client matches the registry. When the pattern version does not match the registry, it is further verified whether there is a version compatibility policy. If there is a compatibility policy, it is allowed to pass; otherwise, the pattern version compatibility verification is determined to have failed, the version incompatibility information is recorded and the submission is blocked or the corresponding object is marked as pending repair.

[0069] Step S5) Consistency verification includes: performing unit consistency verification to verify whether cross-domain physical quantity units are convertible and conform to the preset strategy.

[0070] Specifically, the unit configuration field in the expression record and the unit configuration between cross-domain expression records involved in the submission request are obtained. The unit system adopted by each expression record is verified to be convertible and conforms to the preset strategy. When there is an unconvertible unit or the unit conversion causes the numerical precision to exceed the allowed range, the unit consistency verification is determined to fail, the unit conflict information is recorded and the submission is blocked or the corresponding object is marked as pending repair.

[0071] Step S5) Consistency verification includes: performing coordinate system consistency verification to verify whether the coordinate datum is convertible and conforms to the preset strategy;

[0072] Specifically, the coordinate reference field in the expression record and the coordinate reference between cross-domain expression records involved in the submission request are obtained. The coordinate reference used by each expression record is verified to be convertible and in accordance with the preset strategy. When there is an unconvertible coordinate reference or the spatial position deviation caused by the coordinate transformation exceeds the allowable range, the coordinate system consistency verification is determined to be unsuccessful. The coordinate system conflict information is recorded and the submission is blocked or the corresponding object is marked as pending repair.

[0073] Step S5) The consistency verification step includes: performing anchor point resolvability verification to verify whether the key anchor points are successfully resolved or in an allowed pending repair state.

[0074] Specifically, the parsing status field in the anchor mapping record involved in the submission request is obtained, and the parsing status of the key anchor is verified to be successful. When the parsing status of the key anchor is failed or ambiguous, if the preset strategy allows the state to be repaired, it is marked as the state to be repaired. Otherwise, the anchor parsing validity check is determined to have failed, the reason for failure is recorded and the submission is blocked or the corresponding object is marked as the state to be repaired.

[0075] Step S5) Consistency verification includes: performing semantic consistency verification based on ontology constraints to verify whether the boundary condition type matches the type of the active entity, whether the process type matches the material requirements, and whether the anchor point semantic type matches the parsed entity semantic type.

[0076] Specifically, the semantic annotation records and explicit semantic relationship records are obtained, and the constraint rules in the engineering semantic ontology are combined to verify whether the boundary condition type matches the type of the active entity, whether the process type matches the material requirements or surface requirements, and whether the anchor point semantic type matches the parsed entity semantic type. When any semantic constraint is violated, the semantic consistency verification is determined to have failed, a violation summary is recorded and submission is blocked, or the corresponding object is marked as pending repair.

[0077] When all the above checks pass, the consistency check is considered successful, and the next step is executed; when any check fails, the generation of a publishable commit record is prevented, or the corresponding object is marked as pending repair and the reason for failure is recorded to trigger a manual or rule-based repair process.

[0078] The technical effect of step S5 is that by incorporating reference integrity, schema version compatibility, unit consistency, coordinate system consistency, anchor point resolvability, and ontology-based semantic consistency into the same set of verification rules, a comprehensive verification of the records involved in this submission and their relationships is performed before the submission record is generated. This ensures that only data that meets structural integrity, numerical semantic consistency, and engineering semantic constraints can enter the version graph and index storage, fundamentally avoiding collaborative errors caused by broken references, unit conflicts, coordinate misalignment, or semantic mismatch. When verification fails, by marking the status to be repaired and recording the reason for failure, precise repair guidance is provided to engineers, bringing the discovery and location of problems forward to the submission stage, significantly reducing cross-domain collaborative failures and rework costs caused by data inconsistency.

[0079] Step S6) Incremental reasoning step: Based on the limited semantic subgraph range of the change set, perform incremental ontology reasoning and rule reasoning to generate inferred semantic relationship records and interpretation chain records, and solidify and store the inferred semantic relationship records and interpretation chain records in the semantic graph storage.

[0080] The semantic graph storage is used to store semantic annotation records, explicit semantic relationship records, and inferred semantic relationship records. It is organized in a directed labeled graph structure, where nodes are concept identifiers or annotated identifiers in semantic annotation records, and edges are semantic associations in explicit semantic relationship records or inferred semantic relationship records. Each edge contains subject identifier, predicate identifier, object identifier, relationship type, and scope information. The semantic graph storage supports traversal and querying by node identifier, edge type, subgraph range, or scope identifier, providing a semantic subgraph extraction interface for incremental reasoning.

[0081] Step S6) The incremental reasoning step includes: determining the set of affected semantic nodes based on the change set; performing reasoning closure calculation on the set of affected semantic nodes to generate inferred semantic relations; generating an explanation chain summary corresponding to the inferred semantic relations and writing it into the reasoning result record.

[0082] In this application, determining the set of affected semantic nodes based on the change set includes: obtaining the change set corresponding to the current submission, wherein the change set includes at least a list of newly added records, a list of deleted records, a list of updated records, and a set of affected roots; locating the associated semantic nodes in the semantic graph storage according to the record identifiers in the change set, wherein the semantic nodes include concept identifiers in semantic annotation records and subject identifiers and object identifiers in explicit semantic relationship records; and constructing the affected semantic node set by combining the located semantic nodes and their adjacent nodes within a one- or two-hop range in the semantic graph, thereby limiting the scope of the semantic subgraph for incremental reasoning.

[0083] Performing inference closure computation on the affected semantic node set to generate inferred semantic relations includes: loading the rule set record corresponding to the engineering semantic ontology, wherein the rule set record contains at least ontology inference rules and custom rule inference rules; taking the explicit semantic relations in the affected semantic node set as the starting point of inference, applying the inference rules in the rule set record to perform forward chain inference or backward chain inference, and performing inference closure computation until no new semantic relations are derived; taking the newly derived semantic relations as inferred semantic relations, wherein each inferred semantic relation contains at least a subject identifier, a predicate identifier, an object identifier, a relation type, and an inference basis.

[0084] Generating an interpretation chain summary corresponding to the derived semantic relations and writing it into the reasoning result record includes: for each derived semantic relation, recording the inference rule identifier applied in its derivation process and the explicit semantic relation identifier or intermediate derived semantic relation identifier it depends on, forming an interpretation chain; performing hash calculation on the interpretation chain to generate an interpretation chain summary, or serializing the interpretation chain into a structured summary; generating a reasoning result record, the reasoning result record at least includes a reasoning result identifier, a list of derived semantic relations, an interpretation chain summary, a derivation timestamp, and a scope identifier, and storing the reasoning result record in a semantic graph storage and a semantic index storage, the scope identifier at least including a commit record identifier or a version identifier, so as to achieve scope isolation of reasoning results by version or by commit.

[0085] When the reasoning process detects a derivation result that conflicts with a preset constraint rule, a violation summary is generated and written into the reasoning result record. The violation summary includes at least the violation type, the conflict node identifier, and the identifier of the violated constraint rule to support semantic consistency review and automated repair.

[0086] Its technical advantages lie in the fact that by performing incremental ontology reasoning and rule reasoning based on the scope of the semantic subgraph limited by the change set, it only performs reasoning closure calculation on the set of semantic nodes affected by the current submission, avoiding repeated reasoning on the entire semantic graph, reducing the reasoning computation resource consumption to less than 10% of that of full reasoning, and significantly improving the reasoning efficiency in large-scale semantic graph scenarios; by generating and storing deducing semantic relationship records and interpretation chain summaries, it binds the reasoning results with the submission record, making the reasoning results auditable, reproducible, and rollbackable; by recording interpretation chains and violation summaries, it provides traceable reasoning basis for intelligent verification, automatic configuration, and process recommendation, supporting the reliable invocation and interpretable blocks of engineering semantics by upper-layer artificial intelligence models and automated processes.

[0087] Furthermore, this application also includes: encoding the change set into incremental transmission packets; and, during network transmission, performing missing block filling and breakpoint resumption on the incremental transmission packets based on the block hash value.

[0088] In this application, the change set corresponding to the current submission is obtained. The change set includes at least a list of newly added records, a list of deleted records, a list of updated records, and a set of affected roots. The record identifiers, record content references, and change operation types in the change set are serialized into an incremental transport packet format. The incremental transport packet includes at least a packet identifier, a source version identifier, a target version identifier, a change operation sequence, and a content hash value reference list. When the change set involves changes to the expressed content, the incremental transport packet carries the content hash value and block hash value list corresponding to the expressed content, rather than carrying the complete binary data of the expressed content, in order to minimize the transmitted content.

[0089] After receiving the incremental transmission packet, the receiving end parses it to obtain a content hash value reference list and a block hash value list. For each content hash value, the receiving end checks whether the corresponding content already exists in the local content storage. When a block is missing in the local content storage, the receiving end sends a block replacement request to the sending end based on the missing block hash value. The sending end reads the corresponding block data from the content storage based on the block hash value and returns it to the receiving end. After obtaining the missing block, the receiving end reassembles the complete content and writes it to the local content storage according to the content hash value.

[0090] When network transmission is interrupted, the receiving end records a list of successfully received block hash values ​​and a list of unreceived block hash values. When the network is restored, the receiving end sends a resume request to the sending end. The resume request includes at least a transmission session identifier and a list of received block hash values. The sending end determines the untransmitted blocks based on the list of received block hash values ​​and sends only the untransmitted block data to the receiving end, thus achieving resume transmission after transmission interruption and avoiding duplicate transmission of completed blocks.

[0091] When the incremental transmission packet is used for cross-site synchronization, the sending end and the receiving end maintain their own content storage and metadata storage respectively. Incremental synchronization is achieved by comparing the content hash value and the block hash value. Only the missing expression content blocks and change set metadata are transmitted, compressing the amount of cross-domain synchronization data to less than 10% of the total transmission.

[0092] Its technical advantages lie in the fact that by encoding change sets into incremental transport packets, and carrying only content hash values ​​and a list of block hash value references instead of the complete binary data representing the content during transmission, the amount of network transmission data is compressed to less than 10% of the total transmission volume; through a block completion mechanism based on block hash values, the receiving end only requests the missing data blocks locally from the sending end, avoiding the repeated transmission of existing content; and through a breakpoint resumption mechanism, after network interruption is restored, only the incomplete blocks are resumed, without retransmitting the entire incremental transport packet, significantly improving the synchronization efficiency and network transmission reliability in cross-regional and cross-site collaborative scenarios, and reducing bandwidth consumption and synchronization waiting time.

[0093] Step S7) The commit record generation step is used to generate commit records and append them to the metadata storage, update the index storage and semantic index storage.

[0094] Step S7) The commit record generation step includes: incorporating metadata update, anchor mapping update, semantic relationship update and semantic index update into the same atomic commit unit; after the atomic commit unit is completed, the data visible to the external query end simultaneously satisfies structural consistency and semantic consistency.

[0095] In this application, generating a commit record includes: assigning a commit identifier to the current commit; setting a parent pointer to point to one or more parent commit records, wherein the parent pointer is used to represent the previous version on which the current commit is based, and when there are multiple parent pointers, it indicates a branch merge; setting a root reference set, wherein the root reference set contains one or more version reference record identifiers involved in the current commit, used to locate the core change scope of the current commit; setting a changeset reference, pointing to the changeset record corresponding to the current commit, wherein the changeset record at least includes a list of added records, a list of deleted records, a list of updated records, and a set of affected roots; and generating the commit record, wherein the commit record at least includes a commit identifier, a list of parent pointers, a root reference set, a changeset reference, a commit timestamp, and a committer identifier.

[0096] The process of appending commit records to metadata storage includes: using a log structure merge tree or append log structure as the underlying storage engine for the metadata storage; writing the commit records to the storage medium in a sequential append manner; not modifying existing records during the writing process, only adding new commit records through appending; generating a write position pointer after the append writing is completed; when multiple commit records are appended sequentially, a traversable version graph is formed through the parent pointer and the predecessor pointer, supporting branching, merging, and tracing; during the merge and compaction operation in the background, multiple sequentially appended commit records and their associated records are merged and organized to optimize query performance and storage space utilization.

[0097] Incorporating metadata updates, anchor mapping updates, semantic relationship updates, and semantic index updates into a single atomic commit unit includes: packaging the metadata record updates, anchor mapping record updates, semantic relationship record updates, and semantic index updates involved in this commit into a single atomic commit unit. All update operations within this atomic commit unit either succeed entirely and are committed to disk, or fail entirely and are rolled back; there is no partial success or partial failure. The metadata updates at least include adding or modifying object records, version reference records, instance records, cross-domain expression set records, expression records, and anchor set records. The anchor mapping updates at least include adding or modifying anchor mapping records. The semantic relationship updates at least include adding or modifying semantic annotation records, explicit semantic relationship records, and inferred semantic relationship records. The semantic index updates at least include adding or modifying index entries in the semantic index storage based on concept identifiers, predicate identifiers, or anchor identifiers.

[0098] After the atomic commit unit is completed, the data visible to the external query client must simultaneously satisfy structural consistency and semantic consistency. This includes: before the atomic commit unit is successfully committed, the external query client cannot query any intermediate state data of this commit; after the atomic commit unit is successfully committed, the data queried by the external query client through the index storage and semantic index storage are all complete states after this commit is completed. The structural consistency is at least reflected in the complete reachability of the reference relationship between records, and the semantic consistency is at least reflected in the synchronous update of the semantic annotation record, explicit semantic relationship record, and inferred semantic relationship record with the annotated record, without intermediate inconsistent states such as "CAD expression has been updated but anchor mapping has not been updated" or "semantic relationship has been stored but semantic index has not been updated".

[0099] Updating the index storage and semantic index storage includes: during or after the execution of an atomic submission unit, updating the main index, anchor index, content index, and expression index in the index storage. The main index is used for quick location by object identifier, version identifier, or instance identifier. The anchor index is used to look up which loads, constraints, or processing regions reference it by anchor identifier. The content index is used to locate expression records by content hash value. The expression index is used for retrieval by domain type, format, or mode version. Updating the semantic index in the semantic index storage involves using the semantic index to look up semantic annotation records and semantic relationship records by concept identifier, predicate identifier, or anchor identifier, providing a searchable semantic foundation for artificial intelligence reasoning and automation processes.

[0100] Its technical effectiveness lies in ensuring that the data visible to external queries at any given time satisfies both structural and semantic consistency by incorporating metadata updates, anchor mapping updates, semantic relationship updates, and semantic index updates into the same atomic commit unit. This fundamentally avoids intermediate inconsistencies such as "CAD expressions have been updated but anchor mappings have not" or "semantic relationships have been stored but semantic indexes have not been updated." By using an LSM-tree or append log structure for appending commit records, random writes are converted to sequential writes, significantly improving concurrent commit throughput. This is further enhanced by background merging and compaction to optimize storage space utilization. By generating commit records containing a set of parent pointers and root references, a traversable version graph structure is formed, supporting branching, merging, and full version tracing. This provides a structured version foundation for version management, change auditing, and issue backtracking in cross-domain collaboration.

[0101] Step S8) The affected analysis step is to perform affected analysis based on the submitted records and the expression dependency graph, generate recalculation markers or repair markers, publish change events, and trigger collaborative response actions.

[0102] Step S8) The affected analysis steps include: performing graph traversal along the derived pointer chain and the expression dependency graph to determine the minimum recalculation closure set of CAE expression records and CAM expression records affected by the CAD expression record update; generating a task queue and marking the expression records in the minimum recalculation closure set as pending recalculation; and publishing change events to the corresponding domain client to trigger the automatic recalculation process.

[0103] In this application, the process of traversing the graph along the derived pointer chain and the expression dependency graph to determine the minimum recalculation closure set of CAE and CAM expression records affected by the CAD expression record update includes: obtaining the commit record generated in this commit; locating the changed version reference record from the root reference set of the commit record; obtaining the cross-domain expression set record associated with the version reference record; extracting the set of changed expression record identifiers from the cross-domain expression set record; starting from the changed expression record identifiers, performing a graph traversal upstream or downstream along the derived pointer chain in the expression record to obtain all other expression records that directly or indirectly depend on the changed expression record; and simultaneously loading the cross-domain table. The expression dependency graph in the set of records is a directed acyclic graph, where nodes are expression identifiers and edges represent derivation or dependency relationships. In the expression dependency graph, with the changed expression record identifier as the source node, a breadth-first traversal or depth-first traversal is performed to collect all reachable target nodes. The expression records belonging to the CAE or CAM domain in the set of expression records obtained from the graph traversal are identified as the CAE and CAM expression records affected by this change, and the minimum recalculation closure set is selected from them. The minimum recalculation closure set refers to the minimum subset of expression records that satisfies the integrity of dependency relationships, such that all expression records in this subset need to be recalculated, and it does not contain redundant expression records that do not need to be recalculated.

[0104] Generating a task queue and marking the representation records in the minimum recalculation closure set as pending recalculation includes: generating corresponding recalculation tasks based on the domain type and computation type of each representation record in the minimum recalculation closure set; each recalculation task includes at least a task identifier, representation identifier, task type, priority, and a list of dependent tasks; writing the generated recalculation tasks into the task queue, which supports priority scheduling and dependency management to ensure the execution order of CAE mesh recalculation being completed before CAE solution recalculation, and CAE solution recalculation being completed before CAM toolpath recalculation; and marking each representation record in the minimum recalculation closure set as pending recalculation and writing the pending recalculation state into the metadata of the representation record to support subsequent queries and audits.

[0105] The process of publishing change events to the corresponding domain clients and triggering the automatic recalculation process includes: publishing change events to CAE clients or CAM clients respectively according to the domain type of each representation record in the minimum recalculation closure set; the change events include at least an event identifier, a change type, a list of affected representation identifiers, and a recalculation task reference; the change events are sent to subscribed clients via push through the event distribution module, or are retrieved by clients asynchronously through a message queue; when a client receives a change event, it retrieves a recalculation task from the task queue according to the recalculation task reference and automatically executes the corresponding recalculation process, the recalculation process including at least one of CAE mesh repartitioning, CAE solution recalculation, and CAM toolpath regeneration.

[0106] When the anchor mapping record is found to be in a state of pending repair or a violation summary is generated during the affected analysis process, generating a pending repair marker and triggering manual processing includes: during graph traversal along the expression dependency graph, detecting the parsing status of anchor mapping records associated with the affected expression record; when there is an anchor mapping record with a parsing status of pending repair, generating a pending repair marker; writing the pending repair marker into the metadata of the corresponding expression record or the metadata of the anchor mapping record, and generating a repair task and writing it into the task queue; publishing a repair event to the corresponding domain client to trigger the manual processing flow, which includes at least one of the following: displaying a list of anchors to be repaired on the client interface, providing an entry point for re-editing the selected expression, or providing a candidate entity manual selection interface.

[0107] When the affected analysis step is linked with the incremental reasoning step, determining the set of affected objects or the set of constraint conflicts based on the derived semantic relations includes: obtaining the derived semantic relation records generated in step S6, and extracting the derived semantic relations related to this change; locating the affected expression records, anchor mapping records, or semantic annotation records based on the subject identifier, predicate identifier, and object identifier in the derived semantic relations; including the located records in the affected analysis scope, and generating a comprehensive affected set by combining the expression dependency graph traversal results, so as to support more comprehensive automatic recalculation triggering and constraint conflict detection.

[0108] Its technical advantages lie in the fact that by traversing the graph along the derived pointer chain and the expression dependency graph, it accurately determines the minimum set of recalculation closures of CAE and CAM expression records affected by the update of CAD expression records. This compresses the full recalculation that might otherwise be triggered to the minimum necessary range, avoiding unnecessary mesh repartitioning, solution recalculation, and toolpath regeneration, reducing the computational resource consumption of cross-domain recalculation by more than 80%. By generating a task queue and marking the status to be recalculated, combined with priority scheduling and dependency management, it ensures that the recalculation process is executed automatically in the correct order, without the need for engineers to manually identify and trigger it. By publishing change events to the corresponding domain clients, it triggers automatic recalculation or manual processing, compressing the rework time caused by design changes in cross-domain collaboration from several hours to minutes or seconds of automated response. When linked with incremental inference steps, it uses inference semantic relationships to locate the set of affected objects and the set of constraint conflicts, further expanding the coverage of the affected analysis, improving the completeness and accuracy of automated processing, and providing a reliable execution foundation for intelligent collaboration and automated processes.

[0109] like Figure 4 As shown, in a feasible embodiment, this application further includes step S32) a copy-on-write step: based on the copy-on-write strategy, a new version record is generated for the version reference record and instance record on the affected path; the unaffected structure reuses the old version record through the predecessor pointer or reference; the newly generated commit record is pointed to the old commit record through the parent pointer to form a version graph.

[0110] In this application, generating new version records for version reference records and instance records on the affected path based on the copy-on-write strategy includes: obtaining the commit record corresponding to the current commit; locating the set of version reference record identifiers involved in the current commit from the root reference set of the commit record; for each version reference record in the set of version reference record identifiers, determining whether it has changed due to the current commit, wherein the change includes at least one of expression set reference change, anchor set reference change, state change, validity change, or attribute change; when a version reference record changes, generating a new version reference record and assigning a new version identifier to the new version reference record. Copy the unchanged field values ​​to the new version reference record, and set the predecessor pointer of the new version reference record to point to the old version reference record before the change; when the version reference record has not changed, reuse the old version reference record and do not generate a new record; for instance records associated with the version reference record, when the instance record is affected by pose change, quantity change, configuration expression change, or context change, generate a new instance record, assign a new instance identifier to the new instance record, and set the predecessor pointer of the new instance record to point to the old instance record before the change; when the instance record is not affected, reuse the old instance record and do not generate a new record.

[0111] Unaffected structures reuse old version records via predecessor pointers or references, including: for unchanged version reference records, the original version identifier is directly referenced in the root reference set of this commit, without generating a new record; for unaffected instance records, the old instance record is directly pointed to in the parent instance record or version reference record via the reference identifier, without generating a new record; for unchanged cross-domain expression set records, the old cross-domain expression set record is directly referenced in the version reference record via the cross-domain expression set identifier, without generating a new record; for unchanged expression records, the old expression record is directly referenced in the expression reference set of the cross-domain expression set record via the expression identifier, without generating a new record. Through the above reuse mechanism, structure sharing is achieved, and only records on the affected paths are generated with new versions, while all unaffected structures are reused, thereby reducing write amplification and storage space occupation.

[0112] The process of creating a version graph by pointing newly generated commit records to older commit records via parent pointers includes: when generating the commit record for the current commit, setting a parent pointer list for the commit record, which contains one or more pointers pointing to parent commit records; when the current commit is based on a single preceding version, the parent pointer list contains a single parent pointer pointing to the commit record of that preceding version, forming a linear version chain; when the current commit merges multiple branches, the parent pointer list contains multiple parent pointers, each pointing to the latest commit record of each branch, forming a branch merge node; through the parent pointer list and the preceding pointer chain in each commit record, a traversable directed acyclic graph structure is formed. The version graph supports traversing all ancestor versions backward from any commit record and traversing all derived versions backward, realizing full history traversal of versions and branch management.

[0113] When the copy-on-write step is linked with the affected analysis step, the version reference record and instance record on the affected path are determined according to the minimum recalculation closure set determined in step S8. New version records are generated only for the version reference record and instance record paths on which the minimum recalculation closure set depends, further narrowing the range of new version record generation and reducing write amplification.

[0114] Its technical advantages lie in the fact that, based on the copy-on-write strategy, new version records are generated only for version reference records and instance records on the affected path. Unaffected structures directly reuse old version records through predecessor pointers or references, reducing the write amplification of each commit to only involve records on the changed path, avoiding full copying of the version tree and instance tree, and significantly reducing storage space occupation and write overhead. By pointing the newly generated commit records to the old commit records through parent pointers, a traversable directed acyclic graph version graph is formed, supporting complex version management scenarios such as linear version chains, branches, and merges, providing a structured version foundation for version tracing, change auditing, branch development, and merge integration in cross-domain collaboration. When linked with the affected analysis step, the affected path is further accurately determined based on the minimum recalculated closure set, compressing the generation range of new version records to the minimum necessary set, thereby optimizing storage efficiency and write performance.

[0115] In complex product development scenarios, this application forms an automated closed loop from design change triggering, cross-domain association maintenance, dependency impact analysis to collaborative response execution. It realizes an integrated system of "unified structure, stable association, traceability, efficient storage, and intelligent reasoning" for cross-domain collaborative processes, which significantly reduces CAE / CAM chain breakage and rework caused by CAD topology changes, reduces cross-domain data duplication storage and transmission bandwidth occupation, improves concurrent collaboration efficiency and traceability, and provides a structured and reasonable semantic foundation for intelligent verification, automatic configuration, and AI-assisted engineering.

[0116] The advantages of this application are as follows: It provides a PLM persistent storage and semantic reasoning method for CAD / CAE / CAM collaboration. By replacing the instantaneous topology number with semantic anchors, it automatically rebinds CAE constraints and CAM processing areas when the CAD model changes, reducing rework due to chain breakage from hours to second-level automated processing. Through content hashing for deduplication and incremental transmission, it reduces the amount of cross-domain synchronized data to less than 10% of the original data volume, significantly reducing storage and bandwidth consumption. It achieves full traceability of the collaborative process through immutable records and version graphs, meeting the audit requirements of highly compliant industries. By using expression dependency graphs to drive automatic recalculation of minimum closures, it reduces the frequency of human-computer interaction in cross-domain collaboration by more than 80%. Furthermore, by combining engineering semantic ontology and incremental reasoning, it provides an auditable reasoning foundation for intelligent verification, automatic configuration, and process recommendation, forming quantifiable technical effects in chain breakage repair, storage and transmission, traceability auditing, automatic recalculation, and intelligent applications.

[0117] The foregoing has provided a detailed description of a PLM persistent storage and semantic reasoning method for CAD / CAE / CAM collaboration. Specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of this application. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this application. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A PLM persistent storage and semantic reasoning method for CAD / CAE / CAM collaboration, characterized in that, Specifically, the steps include the following: The step of obtaining a submission request involves obtaining a submission request from the client, wherein the submission request includes at least the target object identifier and the content to be written. The content writing step involves performing block division and hash calculation on the content to be written, generating a content hash value, and writing the content to be written into the content storage according to the content hash value. The expression record generation step generates an expression record, which is recorded to a cross-domain expression set and points to the derived old expression record through a derivation pointer; The semantic annotation generation step involves obtaining the object identifier, version identifier, instance identifier, and expression identifier associated with the target object identifier, mapping the object identifier, version identifier, instance identifier, expression identifier, or attribute item to the concept identifier in the engineering semantic ontology, and generating semantic annotation records and explicit semantic relationship records. The consistency verification step involves verifying the records and their relationships involved in the submission request according to the consistency verification rule set. If the verification is successful, the next step is executed. If the verification fails, prevent the generation of publishable submission records or mark the corresponding object as pending repair. The incremental reasoning step involves performing incremental ontology reasoning and rule reasoning based on the limited semantic subgraph scope of the change set, generating deduced semantic relationship records and interpretation chain records, and storing them in a fixed manner. The commit record generation step involves generating a commit record and appending it to the metadata storage, as well as updating the index storage and semantic index storage. The affected analysis step involves performing affected analysis based on the submitted records and the expression dependency graph, generating recalculation markers or repair markers, publishing change events, and triggering collaborative response actions.

2. The PLM persistent storage and semantic reasoning method for CAD / CAE / CAM collaboration as described in claim 1, characterized in that, The process further includes the following steps after the expression record generation step: The anchor point mapping record acquisition step involves obtaining the anchor point set record when the submission request contains anchor point definitions, parsing the expression record to obtain a candidate entity set, performing selection expression parsing on each anchor point, filtering the candidate entity set using geometric predicates and attribute constraints to obtain entity identifiers, and generating an anchor point mapping record corresponding to the expression record.

3. The PLM persistent storage and semantic reasoning method for CAD / CAE / CAM collaboration as described in claim 1, characterized in that, The expression record generation step includes: Incorporate metadata updates, anchor mapping updates, semantic relationship updates, and semantic index updates into the same atomic commit unit; After the atomic submission unit is completed, the data visible to the external query end simultaneously satisfies both structural consistency and semantic consistency.

4. The PLM persistent storage and semantic reasoning method for CAD / CAE / CAM collaboration as described in claim 2, characterized in that, The semantic annotation generation step includes: Map the object identifier to the object class concept identifier in the engineering semantic ontology; Map the version identifier to the version class concept identifier in the project semantic ontology; Map the instance identifier to the occurrence class concept identifier in the project semantic ontology; Map the expression identifier to the expression class concept identifier in the engineering semantic ontology; Map the anchor point identifier to the anchor point class concept identifier in the engineering semantic ontology; Map the attribute items to the attribute class concept identifiers in the engineering semantic ontology; Generate the semantic annotation record, which records the mapping relationship between each identifier and the concept identifier; Generate the explicit semantic relationship record, record the semantic association between each identifier, and mark the explicit semantic relationship record as an assertion type.

5. The PLM persistent storage and semantic reasoning method for CAD / CAE / CAM collaboration as described in claim 1, characterized in that, The consistency verification step includes: Perform reference integrity checks to verify the reachability of version reference records to cross-domain expression set records, the reachability of cross-domain expression set records to expression records, the reachability of anchor set records to anchors, and the reachability of anchors to anchor mapping records. Perform a mode version compatibility check to verify whether the mode version written by the client matches the registry; Perform unit consistency verification to check whether cross-domain physical quantity units are convertible and conform to the preset strategy; Perform a coordinate system consistency check to verify whether the coordinate datum is convertible and conforms to the preset strategy; Perform anchor point resolvability verification to check whether critical anchor points are successfully resolved or in an allowed pending repair state; Perform semantic consistency checks based on ontology constraints to verify whether the boundary condition type matches the type of the acting entity, whether the process type matches the material requirements, and whether the anchor point semantic type matches the parsed entity semantic type.

6. The PLM persistent storage and semantic reasoning method for CAD / CAE / CAM collaboration as described in claim 1, characterized in that, The incremental reasoning steps include: The set of affected semantic nodes is determined based on the change set; Perform inference closure computation on the affected set of semantic nodes to generate inferred semantic relations; Generate an explanation chain summary corresponding to the derived semantic relationship and write it into the reasoning result record.

7. The PLM persistent storage and semantic reasoning method for CAD / CAE / CAM collaboration as described in claim 1, characterized in that, The submission record generation step includes: Incorporate metadata updates, anchor mapping updates, semantic relationship updates, and semantic index updates into the same atomic commit unit; After the atomic submission unit is completed, the data visible to the external query end simultaneously satisfies both structural consistency and semantic consistency.

8. The PLM persistent storage and semantic reasoning method for CAD / CAE / CAM collaboration as described in claim 1, characterized in that, The affected analysis steps include: A graph traversal is performed along the derived pointer chain in the expression record and the expression dependency graph to determine the minimum set of recalculation closures of the CAE expression record and CAM expression record affected by the update of the CAD expression record; Generate a task queue and mark the expression records in the minimum recalculation closure set as pending recalculation. Publish the change event to the corresponding domain client to trigger the automatic recalculation process.

9. The PLM persistent storage and semantic reasoning method for CAD / CAE / CAM collaboration as described in claim 1, characterized in that, It also includes a copy-on-write step, which includes: Based on the copy-on-write strategy, new version records are generated for version reference records and instance records on the affected paths; Unaffected structures reuse older versions of records via predecessor pointers or references; The newly generated commit record is linked to the old commit record through a parent pointer, forming a version graph.

10. The PLM persistent storage and semantic reasoning method for CAD / CAE / CAM collaboration as described in claim 1, characterized in that, Also includes: The change set is encoded into an incremental transport packet; During network transmission, the incremental transmission packets are padded with missing blocks and resumed from breakpoints based on the block hash value.

Images