IFC model data and fire-fighting specification knowledge semantic mapping method and system

Abstract

The invention relates to an IFC model data and fire-fighting specification knowledge semantic mapping method and system. According to the method, on one hand, a vector database is constructed and called, labels are added to IFC model data to be analyzed, and an ontology knowledge base is obtained; on the other hand, performing standard semantic analysis on the fire-fighting standard text to construct a structured rule base; based on the ontology knowledge base and the structured rule base, semantic mapping and semantic reasoning are carried out by utilizing a large language model, and a mapping and reasoning result is obtained; mapping and reasoning results are fed back to the ontology knowledge base to complete knowledge updating; and finally, performing compliance verification based on a reasoning result, and generating and outputting a fire-fighting compliance check report of the current BIM model. Compared with the prior art, the method has the advantages that a computer can accurately understand the meaning of the IFC model in the context of the fire-fighting specification, and the accuracy and reliability of intelligent examination are improved.

Description

Technical Field

[0001] This invention relates to the field of fire protection engineering technology, and in particular to a semantic mapping method and system for IFC model data and fire protection code knowledge. Background Technology

[0002] Significant technical bottlenecks exist in the current field of Building Information Modeling (BIM) IFC (Industry Foundation Classes) data parsing. Traditional IFC parsing methods exhibit two major limitations: entity-based methods require professionals to spend considerable time manually mapping data against standard documents, resulting in repetitive and tedious work; while dedicated parsing software suffers from insufficient scalability due to its rigid architecture. More critically, in practical applications, the IFC standard commonly prioritizes geometric data while weakening semantic information, leading to severe semantic integrity loss during cross-platform interaction. This semantic gap directly restricts the in-depth application value of BIM (Building Information Modeling) models throughout their entire lifecycle.

[0003] In the professional field of fire safety code review, existing technologies face three systemic challenges. First, the review process relies heavily on human experience and interpretation, resulting in low efficiency and significant subjective variability. Studies show that the time cost of reviewing a single project is 3-5 times higher than traditional methods. Second, code provisions contain numerous complex spatial constraints, such as the dynamic matching relationship between the clear width of evacuation staircases and personnel loads specified in GB50016, while existing BIM systems lack the ability to accurately parse such high-level semantics. Third, the lack of an effective ontology mapping mechanism between building component attributes and the fire safety code knowledge system makes it difficult for intelligent review accuracy to break through the 70% technical bottleneck.

[0004] Based on current research progress, there is an urgent need for a cross-modal dynamic mapping method to bridge the semantic gap between IFC data and fire protection codes, and to solve the problems of inaccurate semantic mapping and insufficient accuracy and reliability of intelligent review caused by the disconnect between IFC model data and fire protection codes, low efficiency of manual comparison, and the inability of computers to understand the meaning of IFC models in the context of fire protection codes. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings of the existing technology and provide a semantic mapping method and system for IFC model data and fire protection code knowledge.

[0006] The objective of this invention can be achieved through the following technical solutions: According to one aspect of the present invention, a method for semantic mapping between IFC model data and fire protection code knowledge is provided, the method comprising the following steps: S1. Collect IFC standard documents of various versions, preprocess them and convert them into semantic vectors through vector model, and build a vector database of IFC standard documents of multiple versions; S2. Convert the BIM model to IFC format to obtain the IFC model data to be parsed; S3. Call the vector database to add labels to the IFC model data to be parsed and build the ontology knowledge base; S4. Perform semantic parsing on the fire protection code text to construct a structured rule base; S5. Based on the ontology knowledge base and structured rule base, semantic mapping and semantic reasoning are performed using a large language model to obtain the mapping and reasoning results; S6. Feed the mapping and reasoning results back to the ontology knowledge base to complete the knowledge update; S7. Based on the reasoning results, perform compliance verification and generate and output a fire safety compliance inspection report for the current BIM model.

[0007] As a preferred technical solution, the specific process of S3 includes: Parse IFC data and decompose it into component semantics, association semantics, and coordinate semantics; Create dynamic IfcEntity tags for component semantic nodes and static relation attribute tags for associated semantic nodes to obtain parsed semantic data. The vector database is called to label the parsed semantic data, thus constructing the ontology knowledge base.

[0008] As a preferred technical solution, when S5 uses a large language model for semantic mapping and semantic reasoning, it includes the IFC semantic enhancement process and the heterogeneous data fusion process. The IFC semantic enhancement process is as follows: Based on the ontology knowledge base and structured rule base, attribute tags are enhanced. Each attribute tag is divided into automatic attributes, built-in attributes, added attributes, and derived attributes according to the attribute source. Among them, automatic attributes are geometric parameter attributes related to entities that are automatically extracted; built-in attributes are information filled in by the modeler and defined by the modeling software; added attributes are manually added attributes that meet the requirements of the project review; and derived attributes are new attributes derived from existing information in the model through calculation and reasoning. The heterogeneous data fusion process is as follows: Based on the fire protection domain ontology in the structured rule base and the IFC ontology in the ontology knowledge base, an alignment matrix is ​​constructed according to semantic equivalence, an IFC-ontology alignment matrix is ​​constructed, and an equivalent association is established between the IFC ontology and the fire protection domain ontology categories.

[0009] By classifying attributes into four categories, the source and characteristics of IFC attributes are clarified, which solves the problem of "judgment bias caused by confusion of attribute types" in fire protection review. Based on semantic equivalence, an IFC-fire protection ontology alignment matrix is ​​constructed to achieve accurate docking between IFC technical language and fire protection standard language, break down the barriers of cross-domain data heterogeneity, and upgrade semantic reasoning from single parameter matching to multi-attribute collaborative judgment, which significantly improves the comprehensiveness of reasoning results.

[0010] As a preferred technical solution, the specific process of S4 includes: Key textual elements of each clause in fire protection codes are extracted using named entity recognition technology. Based on key text elements, triple rules are generated to construct a structured rule base; Key textual elements include any one or more combinations of spatial entities, spatial relationships, object entities, attribute constraints, logical operators, and behavioral restrictions.

[0011] As a preferred technical solution, the triplet rule Specifically represented as , where L is a finite set of natural language specification clauses; To pass through the function A finite set of rule-based logical expressions obtained through mapping; functions To convert from L The generation rules.

[0012] As a preferred technical solution, the S5 large language model enhances its reasoning ability by combining knowledge graphs and designs specific prompting engineering strategies for the field of building fire protection.

[0013] By enhancing the reasoning ability of large language models through knowledge graphs, the shortcomings of general large models in the field of building fire protection are compensated for. The targeted design of domain-specific prompting engineering strategies enables the model to quickly focus on the core correlation logic of IFC-fire protection, avoids reasoning from deviating from the review objectives, solves the problem of divergent reasoning and unreliable results in professional domains of large language models, and significantly improves the efficiency and accuracy of semantic mapping and reasoning.

[0014] As a preferred technical solution, the specific process of semantic mapping and semantic reasoning using a large language model in S5 includes: according to the preset prompt word template, the ontology knowledge base, the structured rule base and the query requirements are concatenated into a complete large language model prompt word; the prompt word is input into the fine-tuned large language model for the building fire protection field, and combined with the ontology knowledge base and the structured rule base, semantic mapping and semantic reasoning are completed.

[0015] As a preferred technical solution, the specific process of compliance verification based on inference results in S7 includes: Identify components and properties in the IFC model that do not comply with fire protection codes; The specific fire safety regulations and clauses violated should be clearly indicated. Provide modification suggestions and compliance reference values.

[0016] As a preferred technical solution, the fire safety compliance inspection report output in S7 includes visualization operations, including: Non-compliant building components are highlighted using 3D models; The mapping relationship between IFC data and fire protection code provisions is presented in the form of a semantic network. It provides an interactive exploration interface that supports detailed mapping relationship queries.

[0017] According to another aspect of the present invention, an IFC model data and fire protection code knowledge semantic mapping system is provided, the system including an IFC parsing module, a fire protection code parsing module, a semantic mapping reasoning module and a compliance verification module; The IFC parsing module collects standard documents from various versions of IFC, preprocesses them, and converts them into semantic vectors using vector models to build a vector database of standard documents from multiple versions of IFC. It then converts BIM models to IFC format to obtain IFC model data to be parsed. Finally, it calls the vector database to add tags to the IFC model data to be parsed, thus building an ontology knowledge base. The fire protection code parsing module performs semantic parsing on fire protection code texts to construct a structured rule base; The semantic mapping and reasoning module is based on the ontology knowledge base and the structured rule base. It uses a fine-tuned large language model to perform semantic mapping and semantic reasoning to obtain the mapping and reasoning results. The mapping and reasoning results are then fed back to the ontology knowledge base to complete the knowledge update. The compliance verification module performs compliance verification based on the reasoning results, and generates and outputs a fire safety compliance inspection report for the current BIM model.

[0018] Compared with the prior art, the present invention has the following beneficial effects: 1. This invention, through the technical chain of IFC data standardization, fire protection rule structuring, large-scale model semantic mapping, and compliance result output, completely solves the pain points of traditional fire protection review, such as the disconnect between IFC model data and fire protection regulations, low efficiency of manual comparison, and easy errors in judgment results. Furthermore, this invention forms a data closed loop through a knowledge update mechanism, enabling the method to be continuously optimized with project accumulation, significantly improving the universality and accuracy of fire protection compliance review of different building BIM models, and reducing over-reliance on the professional experience of reviewers. This allows computers to accurately understand the meaning of IFC models in the context of fire protection regulations, improving the accuracy and reliability of intelligent review.

[0019] 2. In this invention, scattered IFC data is decomposed into three semantic categories: components, associations, and coordinates, and exclusive labels are established. This solves the core problem of ambiguity and chaotic relationships in IFC data. The IfcEntity dynamic label ensures that semantics are not lost when component attributes are updated, and the static relationship attribute label clarifies the entity association logic, providing structured basic data for subsequent semantic mapping. This transforms IFC data from raw numerical values ​​into a semantic carrier that can be associated with fire protection codes, improving data reusability.

[0020] 3. This invention employs named entity recognition technology to extract key elements of fire protection codes, transforming natural language provisions into structured information of "entity-relationship-constraint." This addresses the pain points of traditional fire protection codes being unreadable by machines and relying on manual rule extraction. It clearly defines the coverage of key elements, ensuring comprehensive code parsing and providing accurate and unambiguous rules for subsequent S5 semantic mapping, transforming fire protection codes from textual data into calculable judgment standards. Through the formal expression of triples, a rigid association is formed between natural language provisions, logical expressions, and transformation rules, resolving the problems of ambiguous rule expressions and uncontrollable transformation processes in fire protection code conversion. The function clarifies the transformation logic from natural language to machine rules, making rule generation reproducible and verifiable. This improves the efficiency of building a structured rule base and provides clear operational standards for subsequent rule modification and updates, enhancing the rigor of the technical solution.

[0021] 4. This invention uses preset templates to form complete prompt words, making the reasoning task of the large language model clearer and solving the problem of model comprehension deviation caused by non-standard prompt words; it adopts a fine-tuned large language model in the field of building fire protection, combined with dual knowledge bases for reasoning, to ensure that the model output results not only comply with IFC standards but also meet fire protection code requirements, avoiding the problem of poor domain adaptability of general models and improving the engineering practicality of reasoning results.

[0022] 5. This invention identifies components and attributes in the IFC model that do not comply with fire safety regulations; marks the specific fire safety regulations clauses and provisions that are violated; and provides modification suggestions and compliance reference values. It solves the pain points of traditional compliance verification results being vague and lacking a basis for rectification; by directly linking verification results to specific components and attributes, it enables reviewers to quickly locate the root cause of problems, and the compliance reference values ​​provide clear targets for design optimization, significantly reducing rectification costs and timelines.

[0023] 6. In this invention, the problems of abstract results and unintuitive relationships in traditional written reports are solved by using visualization methods such as 3D model highlighting, semantic network display, and interactive query. 3D highlighting makes non-compliant components clear at a glance, semantic network clearly presents the mapping logic of IFC data and normative clauses, and interactive interface allows reviewers to trace details as needed, improving the readability and efficiency of the report, and significantly enhancing practicality and user experience. Attached Figure Description

[0024] Figure 1 This is a schematic diagram illustrating the steps of a semantic mapping method between IFC model data and fire protection code knowledge in this invention. Figure 2 This is a schematic diagram illustrating the overall implementation approach in the embodiments; Figure 3 This is a diagram illustrating the basic structure of the general building fire protection standard in the embodiment. Figure 4 This is a schematic diagram illustrating the text processing in the embodiment; Figure 5 This is a diagram illustrating the compliance check between the building structure and the code provisions in the embodiment. Figure 6 This is a schematic diagram of the implementation process in the example; Figure 7 This is a schematic diagram illustrating the ontology construction of building data for a certain building in the embodiment; Figure 8 This is an example diagram of a single attribute in the BIM model in the embodiment. Detailed Implementation

[0025] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0026] It is noteworthy that large language model technology has made groundbreaking progress in recent years, and its superior performance in complex semantic parsing provides a new paradigm for solving the aforementioned problems. Especially after integration with architectural knowledge graphs, this technological combination has been successfully applied to scenarios such as structural design optimization and intelligent code parsing. Its multimodal data processing capabilities can effectively capture the implicit spatial topological relationships and attribute constraints in IFC models. This technological approach provides a reliable theoretical foundation and technical feasibility support for building a semantic understanding-based intelligent IFC parsing and automatic fire protection code review system.

[0027] From a technological evolution perspective, the latest extension mechanism of the IFC standard provides a standardized approach to semantic enhancement, while the continuous optimization of the large language model solves the challenges of knowledge representation and reasoning in specialized domains. The synergistic innovation of these two approaches is expected to overcome the current imbalance between geometry and semantics in BIM applications, propelling building information technology (BIM) towards a cognitive intelligence stage. Specifically, by combining the IFC entity extension mechanism with the contextual understanding capabilities of the large language model, automated parsing and verification of complex constraints such as the "relationship between evacuation distance and safety exit configuration" in fire protection codes can be achieved, thereby significantly improving the accuracy and reliability of intelligent review systems.

[0028] This method can efficiently parse and transform IFC data, accurately map it to fire protection code requirements, and has adaptive capabilities to cope with code updates. Furthermore, by combining professional knowledge, it performs in-depth analysis of IFC data, identifies key fire safety information, and accurately matches it with code clauses.

[0029] Example 1 In this embodiment, a semantic mapping method between IFC model data and fire protection code knowledge is adopted. The method steps are as follows: Figure 1 As shown, it specifically includes: S1. Collect IFC standard documents of various versions, preprocess them and convert them into semantic vectors through vector model, and build a vector database of IFC standard documents of multiple versions; S2. Convert the BIM model to IFC format to obtain the IFC model data to be parsed; S3. Call the vector database to add labels to the IFC model data to be parsed and build the ontology knowledge base; S4. Perform semantic parsing on the fire protection code text to construct a structured rule base; S5. Based on the ontology knowledge base and structured rule base, semantic mapping and semantic reasoning are performed using a large language model to obtain the mapping and reasoning results; S6. Feed the mapping and reasoning results back to the ontology knowledge base to complete the knowledge update; S7. Based on the reasoning results, perform compliance verification and generate and output a fire safety compliance inspection report for the current BIM model.

[0030] In this method, the optimization strategies for the large language model include: We combine knowledge graphs to enhance the reasoning ability of large language models; and we design domain-specific prompting engineering strategies to improve the accuracy of semantic mapping.

[0031] IFC data parsing includes: decomposing IFC data into three categories: component semantics, association semantics, and coordinate semantics; Create the IfcEntity dynamic label for the component node and the static relationship attribute label for the associated node.

[0032] The compliance inspection report generation includes: identifying components and attributes in the IFC model that do not comply with fire protection codes; marking the specific code provisions and clauses violated; and providing modification suggestions and compliance reference values.

[0033] The visualization includes: highlighting non-compliant building components using 3D models; displaying the mapping relationship between IFC data and standard clauses in the form of a semantic network; and providing an interactive exploration interface to support detailed mapping relationship queries.

[0034] This method also includes: supporting adaptive parsing and mapping across versions of IFC standards; and realizing collaborative analysis and mapping of geometric and semantic information.

[0035] The implementation process is as follows: The overall implementation plan is as follows: Figure 2 As shown, the fire protection code text is parsed by LLM (Large Language Model) to obtain the code logical expression, and the IFC parser is used to parse the IFC building model to obtain the entity semantic description. Finally, the code logical expression and the entity semantic description are input into the semantic mapping to complete the semantic mapping between the IFC model data and the fire protection code knowledge.

[0036] Specifically, it includes: Construct a vector database of multiple versions of IFC standard documents, and convert the preprocessed IFC standard documents into vector representations; Convert the BIM model to IFC format to obtain the IFC data to be parsed; A fire protection knowledge graph is constructed based on fire protection regulations, and a domain ontology is automatically constructed using an entity relationship joint extraction model. Convert natural language specification clauses into structured logical expressions that can be recognized by computers; The IFC data to be parsed and the database query results are input into the finely tuned large language model for semantic reasoning; The system combines IFC data and query results according to a preset prompt word template to generate prompt words for a large language model. By using a large language model to analyze the relationships between IFC entities, semantic mapping between IFC data and fire protection code knowledge is achieved. A compliance inspection report is generated and visualized based on the semantic mapping results.

[0037] Step 1: Standardize text structuring and construct knowledge ontology Standardized text processing, such as Figure 4As shown, the fire protection code text is first subjected to LLM semantic parsing, then triple rules are generated based on the parsing results, and finally the rule base is dynamically updated by the triple rules to form a structured rule base.

[0038] 1. Standardize deep semantic parsing Key textual elements from GB55037-2022 "General Code for Fire Protection of Buildings," such as spatial entities like "fire compartment," attribute constraints like "fire resistance rating ≥ 2h," and behavioral restrictions like "no crossing," are extracted using named entity recognition technology. These elements are then transformed into triple structures to construct a semantic parsing engine for the code, which is processed through the following layered workflow: Entity identification layer: Identifies key objects (such as firewalls and evacuation doors), attribute constraints (fire resistance rating, net width), spatial relationships (distance, containment), and logical operators (should / should not, and / or) in the specification clauses.

[0039] Rule deconstruction layer: Decomposes complex clauses into atomic rules. For example, "firewall fire resistance limit ≥ 3h" is deconstructed into IfcWall ∧ isFireWall=true → fireResistance≥3.

[0040] Ontology Mapping Layer: Maps the parsed results to the fire protection code ontology (built based on OWL), inheriting the class hierarchy (such as building component class, area class) and attribute relationships (such as hasFireResistance) defined in the paper.

[0041] A semantic bridge is constructed between the IFC model and fire protection codes. This is achieved by parsing the natural language code text using a large language model, dynamically generating logical expressions, and binding them to IFC entity attributes. This enables the computer to understand the "contextual meaning of the code." In this embodiment, the constructed general building fire protection code ontology is as follows: Figure 3 As shown.

[0042] Step 2: IFC semantic enhancement and heterogeneous data fusion 1. IFC heterogeneous data processing and semantic missing completion; By analyzing the logic rules of fire safety review, a single compliance review clause needs to retrieve attributes related to one or more objects. Each attribute that needs to be checked can be divided into four types according to its source: automatic attributes, built-in attributes, added attributes, and derived attributes.

[0043] Extract entity-related automatic attributes and directly obtain geometric parameters such as IfcDoor.Width and IfcStair.TreadDepth; Built-in attributes are information that has been defined by the modeling software and needs to be filled in by the modeler. Manually add attributes that meet the project requirements review, such as the "firewall" attribute; Derived attributes are new attributes derived from existing information in the model through computational reasoning, and can be automatically added to the model.

[0044] Building entity and code clause compliance checks, such as Figure 5 As shown.

[0045] 2. Multi-source data fusion; construct an IFC-ontology alignment matrix to establish equivalent associations (owl:equivalentClass) between IFC entities (such as IfcDoor) and ontology classes (such as evacuation doors) to achieve semantic unification.

[0046] Step 3, rule generation: Transform clauses such as "net width of evacuation route ≥ 1.1m" into triplet parameters retrieved from the IFC file using SPARQL. The pseudocode is shown below: The function is to verify whether the clear width of the evacuation staircase meets the specifications. Write a function to get the net width of the staircase (SPARQL client, component ID instance_id): Step 1: Construct a SPARQL query sparql_query = """ PREFIX definition: standard prefixes such as rdf / ifc / express, etc. Objective: Query the actual width of the stair flight associated with a given component. Query logic: 1. Locate the aggregate component using inst:{instance_id} 2. Locate the stair flight (IfcStairFlight type) within the aggregate component. 3. Attempt to retrieve the "actual stair width" attribute of this stair flight. 4. Note: If an attribute does not exist, the data needs to be manually added. """ Step 2: Execute the query Query result = client.execute query(sparql_query) Step 3: Result Processing If the query result is empty: The message "Manual intervention required: Please add the 'IfcStairFlight_24474224' attribute to the component {instance_id} or assign a correct value" appears. otherwise: Width value = Convert to floating-point number(query result.first record.width value) Return width value End function The BIM evacuation staircase clear width check process involves checking the returned width value. start Input: Unique identifier for stair components Set instance_id = "IfcStair_24474010" to the IFC standard component code. Process: Call the net width calculation service result ← Calculate the clear width of the evacuation staircase ( Model client: client, BIM model connection handle Staircase identifier: instance_id Target component location identifier ) Output: Formatting check results Output: "[{instance_id}]Result of clear width detection of evacuation staircase: {result}" Finish Subprocess definition The function calculates the clear width of the evacuation staircase (parameter client, parameter instance_id): Implementation logic: 1. Query component properties via the IFC SPARQL endpoint 2. Extract or calculate ActualStairWidth parameters 3. Returns a width value with units (e.g., "1.2m") or an error message. Example return value: "1.5m" End function In this embodiment, this solution is used to process the clauses of the GB55037-2022 specification using large language model technology, and the data is classified by attribute. The text data of the GB55037-2022 specification is shown in Table 1. The specific process is as follows: Figure 6 As shown.

[0047] Table 1 GB55037-2022 Standard Text Data Based on an analysis of the "General Code for Fire Protection of Buildings," this paper decomposes the hierarchical structure of building types in a specific building design example, according to the basic relationships between classes. Referring to the IFC standard, the constructed building data ontology is as follows: Figure 7 As shown. Through analysis of the fire safety review logic rules, a single compliance review clause requires retrieving attributes related to one or more objects, extracting and adding various review-required attributes. For example, for the IFC type "ifcDoor" in a building, attributes and corresponding values ​​are added, specifically as shown below. Figure 8 As shown.

[0048] The compliance assessment is as follows: When reviewing the compliance of the net width of a certain evacuation route in the building, the input is displayed in pseudocode form. The function is to verify whether the net width of the evacuation staircase meets the specifications. A function is written to first obtain the net width of the staircase (SPARQL client, component ID instance_id): Step 1: Construct the SPARQL query, sparql_query = """ PREFIX is defined using standard prefixes such as rdf / ifc / express; the goal is to query the actual width of the stair section associated with a given component. The query logic is as follows: 1. Locate the aggregate component using inst:{instance_id}; 2. Locate the stair flight (IfcStairFlight type) within the aggregate component; 3. Try to obtain the "actual stair width" attribute of the stair section; If the attribute does not exist, the data needs to be manually added.

[0049] Step 2: Execute the query, query result = client.execute query(sparql_query); Step 3: Result processing. If the query result is empty, return the following: "Manual intervention required: Please add the 'IfcStairFlight_24474224' attribute to the component {instance_id} or assign it a correct value." Otherwise, convert the width value to a floating-point number (query result.first record.width value) and return the converted width value.

[0050] The BIM process for checking the net width of evacuation staircases, based on the returned width value, is as follows: Enter the unique identifier for the stair component; set instance_id = "IfcStair_24474010" / / IFC standard component code; during processing, call the net width calculation service; Calculate the net width of the evacuation staircase and output the result: The model client uses the client parameter as the BIM model connection handle; the stair identifier uses the instance_id parameter as the target component location identifier.

[0051] Finally, output the formatted inspection result: "[{instance_id}] Inspection result of the net width of the evacuation staircase: {result}" The sub - process of this process is defined as follows: When calculating the net width of the evacuation staircase, use the sub - function (parameter client, parameter instance_id).

[0052] Implementation logic: 1. Query component attributes through the IFC SPARQL endpoint 2. Extract or calculate the ActualStairWidth parameter 3. Return the width value with unit (such as "1.2m") or error message: For example, return "1.5m".

[0053] When checking whether handrails are set for the evacuation staircase in this building, the BIM evacuation staircase handrail inspection process is as follows: The input is to set stair_instance = "IfcStair_24474010" for the IFC stair component instance, which is the IFC standard component type + unique ID; the specific processing is to check the handrail setting, check the stair handrail: The SPARQL client uses the client parameter as the BIM model query interface; the target component uses the stair_instance parameter as the identifier of the stair to be detected.

[0054] Finally, output the following formatted inspection result: "{stair_instance}: Whether handrail is set for the evacuation staircase: {has_railing}"

[0055] The sub - process of this process uses the function to check the stair handrail, and this function contains the parameters client and target_component; the implementation logic of this function is as follows: 1. Construct the SPARQL query statement sparql_query = """ PREFIX is defined as: rdf: <Standard RDF namespace> ifc: <IFC ontology namespace> inst: <Project instance namespace> The query target is to find the IfcRailing type sub-components associated with the target staircase component; The query criteria include: a. Locate the aggregation relationship using ifc:isDecomposedBy; b. Filter ifc:IfcRailing type instances.

[0056] 2. Execute the query and analyze the results: The query results are for the client. Execute a SPARQL query (sparql_query). 3. Return the judgment result: If the query result is not empty, it returns True, indicating that a handrail exists; otherwise, it returns False, indicating that no handrail was found.

[0057] In summary, this solution constructs a bridge to semantically link the implicit or explicit information about building components, spaces, and relationships within the IFC model with fire safety code provisions, typically expressed in natural language as logical rules, numerical requirements, and relational constraints. This allows computers to "understand" the meaning of the IFC model within the context of fire safety codes and to perform inference, verification, or report generation accordingly. This method efficiently parses and transforms IFC data, accurately mapping it to fire safety code requirements, and possesses adaptive capabilities to cope with code updates. Furthermore, by incorporating professional knowledge, it performs in-depth analysis of IFC data, identifies key fire safety information, and precisely matches it with code clauses.

[0058] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in the present invention, and these modifications or substitutions should all be covered within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A semantic mapping method between IFC model data and fire protection code knowledge, characterized in that, The method steps include: S1. Collect IFC standard documents of various versions, preprocess them and convert them into semantic vectors through vector model, and build a vector database of IFC standard documents of multiple versions; S2. Convert the BIM model to IFC format to obtain the IFC model data to be parsed; S3. Call the vector database to add labels to the IFC model data to be parsed and build the ontology knowledge base; S4. Perform semantic parsing on the fire protection code text to construct a structured rule base; S5. Based on the ontology knowledge base and structured rule base, semantic mapping and semantic reasoning are performed using a large language model to obtain the mapping and reasoning results; S6. Feed the mapping and reasoning results back to the ontology knowledge base to complete the knowledge update; S7. Based on the reasoning results, perform compliance verification and generate and output a fire safety compliance inspection report for the current BIM model.

2. The semantic mapping method between IFC model data and fire protection code knowledge according to claim 1, characterized in that, The specific process of S3 includes: Parse IFC data and decompose it into component semantics, association semantics, and coordinate semantics; Create dynamic IfcEntity tags for component semantic nodes and static relation attribute tags for associated semantic nodes to obtain parsed semantic data. The vector database is called to label the parsed semantic data, thus constructing the ontology knowledge base.

3. The semantic mapping method between IFC model data and fire protection code knowledge according to claim 2, characterized in that, When S5 uses a large language model for semantic mapping and semantic reasoning, it includes the IFC semantic enhancement process and the heterogeneous data fusion process. The IFC semantic enhancement process is as follows: Based on the ontology knowledge base and structured rule base, attribute tags are enhanced. Each attribute tag is divided into automatic attributes, built-in attributes, added attributes, and derived attributes according to the attribute source. Among them, automatic attributes are geometric parameter attributes related to entities that are automatically extracted; built-in attributes are information filled in by the modeling personnel and predefined by the modeling software; added attributes are manually added attributes that meet the requirements of project review; and derived attributes are new attributes derived from existing information in the model through calculation and reasoning. The heterogeneous data fusion process is as follows: based on the fire protection domain ontology in the structured rule base and the IFC ontology in the ontology knowledge base, an alignment matrix is ​​constructed according to semantic equivalence, an IFC-ontology alignment matrix is ​​constructed, and an equivalent association is established between the IFC ontology and the fire protection domain ontology category.

4. The semantic mapping method between IFC model data and fire protection code knowledge according to claim 1, characterized in that, The specific process of S4 includes: Key textual elements of each clause in fire protection codes are extracted using named entity recognition technology. Based on key text elements, triple rules are generated to construct a structured rule base; Key textual elements include any one or more combinations of spatial entities, spatial relationships, object entities, attribute constraints, logical operators, and behavioral restrictions.

5. The semantic mapping method between IFC model data and fire protection code knowledge according to claim 4, characterized in that, The triple rule Specifically represented as , where L is a finite set of natural language specification clauses; To pass through the function A finite set of rule-based logical expressions obtained through mapping; functions To convert from L The generation rules.

6. The semantic mapping method between IFC model data and fire protection code knowledge according to claim 1, characterized in that, The S5 large language model enhances its reasoning ability by combining knowledge graphs and designs targeted prompt engineering strategies in the field of building fire protection.

7. The semantic mapping method between IFC model data and fire protection code knowledge according to claim 1, characterized in that, The specific process of semantic mapping and semantic reasoning using a large language model in S5 includes: according to the preset prompt word template, the ontology knowledge base, the structured rule base and the query requirements are concatenated into a complete large language model prompt word; the prompt word is input into the fine-tuned large language model for the building fire protection field, and combined with the ontology knowledge base and the structured rule base, semantic mapping and semantic reasoning are completed.

8. The semantic mapping method between IFC model data and fire protection code knowledge according to claim 1, characterized in that, The specific process of compliance verification based on reasoning results in S7 includes: Identify components and properties in the IFC model that do not comply with fire protection codes; The specific fire safety regulations and clauses violated should be clearly indicated. Provide modification suggestions and compliance reference values.

9. A semantic mapping method for IFC model data and fire protection code knowledge according to claim 8, characterized in that, The output of the fire safety compliance inspection report in S7 includes visualization operations, including: Non-compliant building components are highlighted using 3D models; The mapping relationship between IFC data and fire protection code provisions is presented in the form of a semantic network. It provides an interactive exploration interface that supports detailed mapping relationship queries.

10. A semantic mapping system for IFC model data and fire protection code knowledge, characterized in that, The system operates using a semantic mapping method for IFC model data and fire protection code knowledge as described in any one of claims 1-9. The system includes an IFC parsing module, a fire protection code parsing module, a semantic mapping reasoning module, and a compliance verification module. The IFC parsing module collects standard documents of various versions of IFC, preprocesses them, and converts them into semantic vectors through a vector model to build a vector database of standard documents of multiple versions of IFC; it converts the BIM model into IFC format to obtain IFC model data to be parsed; it calls the vector database to add tags to the IFC model data to be parsed and builds an ontology knowledge base. The fire protection code parsing module performs semantic parsing on the fire protection code text to construct a structured rule base; The semantic mapping and reasoning module is based on the ontology knowledge base and the structured rule base. It uses a fine-tuned large language model to perform semantic mapping and semantic reasoning to obtain the mapping and reasoning results. The mapping and reasoning results are then fed back to the ontology knowledge base to complete the knowledge update. The compliance verification module performs compliance verification based on the reasoning results, and generates and outputs a fire safety compliance inspection report for the current BIM model.

Images