Building fire simulation model construction method based on knowledge graph and IFC

Abstract

The invention relates to the technical field of knowledge maps, in particular to a building fire simulation model construction method based on a knowledge map and IFC, and the method comprises the steps: obtaining an associated entity set and an attribute set of each entity, and an associated command class set and an attribute set of each command class; according to the distance between each entity and each command class, the similarity degree of the belonging fields, the similarity degree of the attribute classes between the attribute sets, the hierarchical difference in the knowledge graph and the difference between the number of the associated words, determining the attribute classes in the knowledge graph; obtaining a map mapping confidence coefficient between each entity and each command class according to the association entity set of each entity and an average level of all cross-domain semantic association degrees between each entity in the association entity set of each entity and each command class in the association command class set of each command class so as to obtain all mapping command classes of each entity; and obtaining a building fire simulation model. The node mapping precision of the cross-domain knowledge graph is improved by analyzing semantics of entities and command classes from multiple dimensions.

Description

Technical Field

[0001] The present application relates to the field of knowledge graph technology, and specifically to a method for constructing a building fire simulation model based on knowledge graph and IFC. Background Art

[0002] In the construction industry, the IFC (Industry Foundation Classes) standard, as a universal specification for building information model (BIM) data exchange, can achieve seamless integration between different types of software and promote efficient information sharing at all stages of the project. As the scale of buildings continues to expand, the issue of building fire safety has received increasing attention. As a technical means to effectively predict the occurrence, development process and impact of fires, building fire simulation technology has become a key tool to ensure the safe design of buildings. Through fire model simulation, scientific and reasonable personnel evacuation strategies can be formulated and the layout of fire protection systems can be optimized. Therefore, constructing an automated building fire simulation model based on knowledge graphs can provide intelligent decision-making support for fire safety management throughout the entire life cycle of building design, construction, operation and maintenance, and create a safer building environment.

[0003] Existing techniques for matching node mappings in knowledge graphs typically rely solely on node names or simple attributes, but fail to fully consider the fact that nodes in different domains may have inconsistent names but similar semantics. Therefore, relying solely on name or simple attribute similarity thresholds for matching makes it difficult to identify instances where names differ but semantics are similar. This results in poor node mapping matching between knowledge graphs in IFC format and those in FDS (Fire Dynamics Simulator) format, which in turn affects the accuracy of the final building fire model. Summary of the Invention

[0004] In order to solve the above technical problems, this application provides a method for constructing a building fire simulation model based on knowledge graph and IFC to solve the existing problems.

[0005] The method for constructing a building fire simulation model based on knowledge graph and IFC in this application adopts the following technical solutions:

[0006] One embodiment of the present application provides a method for constructing a building fire simulation model based on a knowledge graph and IFC, the method comprising the following steps:

[0007] Get the associated entity set and attribute set of each entity in the IFC knowledge graph; get the associated command class set and attribute set of each command class in the FDS knowledge graph;

[0008] The root domain set of each entity and each command class is obtained based on the top-level root node corresponding to all synonyms of each entity and each command class; the cross-domain semantic association between each entity and each command class is obtained based on the distance between each entity and each command class and the similarity of the root domain set, and the graph node matching degree between each entity and each command class is obtained by combining the similarity between the categories of each attribute in the attribute set of each entity and the categories of each attribute in the attribute set of each command class, as well as the difference between the distance between each entity and the top-level node in the knowledge graph in which it is located and the distance between each command class and the top-level node in the knowledge graph in which it is located;

[0009] According to the average level of all cross-domain semantic associations between each entity in the associated entity set of each entity and each command class in the associated command class set of each command class, the difference between the total number of entities in the associated entity set of each entity and the total number of command classes in the associated command class set of each command class, and the graph node matching degree between each entity and each command class, the graph mapping confidence between each entity and each command class is obtained and compared with the preset threshold to obtain all mapped command classes of each entity, and then obtain the building fire simulation model.

[0010] Preferably, the process of obtaining the root domain set of each entity and each command class is as follows: respectively obtain the synonym sets of each entity and each command class; record the set consisting of the top-level root nodes corresponding to all words in the synonym set of each entity in the WordNet thesaurus as the root domain set of each entity; record the set consisting of the top-level root nodes corresponding to all words in the synonym set of each command class in the WordNet thesaurus as the root domain set of each command class.

[0011] Preferably, the calculation formula for the cross-domain semantic association between each entity and each command class is: Where A u,v is the cross-domain semantic association between entity u and command class v, B u,v is the Jaccard similarity between the root domain set of entity u and the root domain set of command class v, C u,v is the normalized Google distance between entity u and command class v, and τ is a preset constant.

[0012] Preferably, the process of obtaining the graph node matching degree between each entity and each command class is as follows:

[0013] Obtaining a first similarity between each entity and each command class based on a similarity between a category of each attribute in the attribute set of each entity and a category of each attribute in the attribute set of each command class;

[0014] Obtain a first difference between each entity and each command class based on the difference between the distance between each entity and the top-level node in the knowledge graph in which it resides and the distance between each command class and the top-level node in the knowledge graph in which it resides;

[0015] Calculate the graph node matching degree between each entity and each command class: Where, F u,v A is the graph node matching degree between entity u and command class v, u,v is the cross-domain semantic association between entity u and command class v, G u,v is the first similarity between entity u and command class v, D u,v is the first difference between entity u and command class v, and τ is a preset constant.

[0016] Preferably, the process of obtaining the first similarity between each entity and each command class is as follows: respectively obtaining the category of each attribute in the attribute set of each entity and each command class through the bSDD dictionary library, and recording the set consisting of the categories of all attributes in the attribute set of each entity as the attribute category set of each entity; recording the set consisting of the categories of all attributes in the attribute set of each command class as the attribute category set of each command class; and recording the Jaccard similarity between the attribute category set of each entity and the attribute category set of each command class as the first similarity between each entity and each command class.

[0017] Preferably, the process of obtaining the first difference between each entity and each command class is as follows: the shortest number of edges in the path from the top-level node to the entity u in the IFC knowledge graph is recorded as the depth value of the entity u; the shortest number of edges in the path from the top-level node to the command class v in the FDS knowledge graph is recorded as the depth value of the command class v; and the absolute difference between the depth values ​​of each entity and each command class is recorded as the first difference between each entity and each command class.

[0018] Preferably, the process of obtaining the confidence level of the graph mapping between each entity and each command class is as follows:

[0019] Obtain the mean value of the associated semantics between each entity and each command class according to the average level of all cross-domain semantic associations between each entity in the associated entity set of each entity and each command class in the associated command class set of each command class;

[0020] Calculate the confidence of the graph mapping between each entity and each command class: Where Z u,v is the confidence of the graph mapping between entity u and command class v, F u,v is the graph node matching degree between entity u and command class v, L u,v is the mean value of the association semantics between entity u and command class v, H u,vis the absolute difference between the total number of elements in the entity set associated with entity u and the total number of elements in the command class set associated with command class v, and τ is a preset constant.

[0021] Preferably, the mean value of the association semantics between each entity and each command class refers to the mean value of all cross-domain semantic associations between each entity in the associated entity set of each entity and each command class in the associated command class set of each command class.

[0022] Preferably, the specific process of comparing with the preset threshold to obtain all mapping command classes of each entity is: the graph mapping confidence between each entity and all command classes in the FDS knowledge graph, and the command class corresponding to the graph mapping confidence greater than the preset threshold is recorded as the mapping command class of each entity.

[0023] Preferably, the process of obtaining a building fire simulation model is: constructing a semantic mapping rule base through all mapping command classes of all entities in the IFC knowledge graph, and constructing a building fire simulation model through the IFC knowledge graph, the FDS knowledge graph and the semantic mapping rule base.

[0024] This application has at least the following beneficial effects:

[0025] This application addresses the problem in the prior art that the node matching is one-sided and the matching effect is poor because the influence of the differences in the environmental semantics and topological structures of nodes in different knowledge graphs is not fully considered. By constructing cross-domain semantic correlation, the core semantic correlation between entities and command classes in different fields can be reflected, and even if there are differences in the names, the semantic consistency can be accurately captured; by constructing graph node matching, the alignment effect of entities and command classes in dimensions such as attribute categories and hierarchical structures can be reflected, thereby improving the accuracy and rationality of node mapping; by constructing graph mapping confidence, it is possible to combine the topological structure similarity and the semantic consistency of associated nodes, introduce a multi-dimensional semantic, attribute and environmental information comprehensive evaluation mechanism, which can significantly improve the node mapping accuracy of cross-domain knowledge graphs, thereby ensuring the construction accuracy of building fire simulation models, and providing reliable support for intelligent fire safety decision-making throughout the life cycle of buildings. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] In order to more clearly illustrate the technical solutions and advantages of the embodiments of the present application or the prior art, the following is a brief introduction to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0027] Figure 1A flowchart of the steps of the method for constructing a building fire simulation model based on knowledge graph and IFC provided in this application;

[0028] Figure 2 This is a flowchart for obtaining the confidence level of the graph mapping between each entity and each command class provided in this application. DETAILED DESCRIPTION

[0029] To further illustrate the technical means and effectiveness of this application to achieve the intended invention objectives, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effectiveness of the method for constructing a building fire simulation model based on a knowledge graph and IFC proposed in this application. In the following description, different references to "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics of one or more embodiments may be combined in any suitable manner.

[0030] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs.

[0031] The specific scheme of the method for constructing a building fire simulation model based on knowledge graph and IFC provided by this application is described in detail below with reference to the accompanying drawings.

[0032] An embodiment of the present application provides a method for constructing a building fire simulation model based on a knowledge graph and IFC. Specifically, the following method for constructing a building fire simulation model based on a knowledge graph and IFC is provided. Figure 1 , the method comprises the following steps:

[0033] Step 1: Obtain the associated entity set and attribute set of each entity in the IFC knowledge graph; obtain the associated command class set and attribute set of each command class in the FDS knowledge graph.

[0034] First, build a BIM building model using building modeling software (such as Revit). Then, export the building model in the IFC standard format. Then, synchronize the IFC-formatted building files exported from the building model using the following two methods:

[0035] Method 1: Use a parsing tool like ifcopenshell to parse IFC-formatted building files to obtain building geometry information, such as component coordinates, dimensions, and spatial relationships; building material properties, such as thermal conductivity and combustion characteristics; and building ventilation system information, such as the layout of ventilation duct networks and firefighting equipment. After parsing the building files, use the open-source knowledge extraction tool DeepDive to extract entities, relationships, and attributes from the parsed data. An IFC knowledge graph is constructed based on the extracted entity objects, relationships between entities, and attribute information.

[0036] Method 2: Import the IFC-formatted building file into the fire simulation software Pyrosim to generate an initial fire simulation model and output the initial FDS file in the FDS data format. This initial FDS file is then processed using the FDS editable program to generate an editable FDS file. Once the editable FDS file is generated, the DeepDive open-source knowledge extraction tool is used to extract entities, relationships, and attributes from the information in the editable FDS file. The FDS knowledge graph is constructed based on the extracted entity objects, relationships between entities, and attribute information.

[0037] It should be noted that the construction of IFC knowledge graph and FDS knowledge graph are both well-known technologies and will not be elaborated here.

[0038] Because the IFC knowledge graph and the FDS knowledge graph are different types, to better distinguish the entities in the two knowledge graphs and avoid confusion, this application refers to various building model entities in the IFC knowledge graph, such as "ifcWall, ifcDoor, ifcMaterial (wall, door, material)", as entities, and various command-class entities in the FDS knowledge graph, such as "&OBST, &VENT, &MATL (obstacle, vent, material)", as command classes. Later, during semantic label matching and mapping, the IFC entities will be linked to the FDS command classes to achieve mapping, thereby completing the construction of the building fire simulation model.

[0039] Since the knowledge graph has already identified entities, the relationships between entities, and the attributes of each entity during its creation, this application uses all entity names in the IFC knowledge graph and all command class names in the FDS knowledge graph to construct entity sets and command class sets, respectively. Based on the entities directly associated with each entity and the attributes of each entity, the associated entity set and attribute set of each entity are constructed respectively; similarly, the associated command class set and attribute set of each command class are constructed.

[0040] Convert all elements in the entity set to lowercase and remove the first three "ifc" characters. If the first three characters of an element are not "ifc", the element is not processed. Convert all elements in the command class set to lowercase and remove all non-alphabetic characters such as "&" and "_". Process all entity names and command class names in the same way (for example, the processed entity set is shown as {"wall", "door", ..., "material"}; the processed command class set is shown as {"obst", "vent", ..., "matl"}).

[0041] Step 2: Obtain the root domain set of each entity and each command class based on the top-level root node corresponding to all synonyms of each entity and each command class; obtain the cross-domain semantic association between each entity and each command class based on the distance between each entity and each command class and the similarity of the root domain set, and combine the similarity between the categories of each attribute in the attribute set of each entity and the categories of each attribute in the attribute set of each command class, as well as the difference between the distance between each entity and the top-level node in the knowledge graph and the distance between each command class and the top-level node in the knowledge graph, to obtain the graph node matching degree between each entity and each command class.

[0042] Because IFC and FDS are built based on different specifications, standards, and application scenarios, even if they originate from the same initial BIM model, the names of the graph nodes in the IFC and FDS knowledge graphs may differ. For example, a "wall" in the architectural field may be represented as an "obstacle" in the fire field. Therefore, it is necessary to combine semantic and attribute information to more accurately map entities and command classes to improve the construction of building fire simulation models.

[0043] Take the u-th entity in the IFC knowledge graph and the v-th command class in the FDS knowledge graph as examples for analysis.

[0044] First, we can analyze the association relationship by analyzing the overlap between the conceptual domains of entity u and command class v. We retrieve synonym sets for entity u and command class v from the WordNet lexicon, and then retrieve the hypernyms of each word in the two synonym sets until the top-level root node. Based on each top-level root node, we construct the root domain set for entity u and command class v.

[0045] As a preferred implementation, the cross-domain semantic association between each entity and each command class is obtained based on the distance between each entity and each command class and the similarity of the root domain set, which is used to characterize the degree of association between each entity and each command class in semantics and conceptual domain.

[0046] In this embodiment, the cross-domain semantic association between entity u and command class v is denoted as A u,v , its specific expression is: Where A u,v is the cross-domain semantic association between entity u and command class v, B u,v is the Jaccard similarity between the root domain set of entity u and the root domain set of command class v, C u,v is the normalized Google distance between entity u and command class v, τ is a preset constant. To avoid the denominator being 0, its value range is (0.001, 0.01). The value has little effect on the calculation and can be ignored. In this embodiment, τ is set to 0.005, and the implementer can choose the value at his own discretion.

[0047] B u,v It can reflect whether entity u and command class v have the same semantic root. The larger the value, the more common conceptual domains entity u and command class v have, and the stronger the degree of association. u,v It can reflect the co-occurrence frequency between entity u and command class v. The smaller the value, the higher the frequency of entity u and command class v appearing together, thus reflecting the closer the association between entity u and command class v. Cross-domain semantic association degree A u,v It can reflect whether entity u and command class v have semantic and conceptual domain associations. The larger the value, the greater the possibility that entity u and command class v belong to different domains but have similar core semantics.

[0048] It should be noted that the use of the WordNet lexicon, Jaccard similarity, and normalized Google distance are all well-known technologies in the art and will not be described in detail here.

[0049] Furthermore, considering that cross-domain semantic relevance only calculates the semantic similarity between entity u and command class v, it does not fully consider the correspondence between the attributes carried in the specific application scenario, which may lead to errors. Therefore, it is necessary to further analyze the relationship between entity u and command class v in combination with attribute information.

[0050] Considering that attributes may have inconsistent names but belong to the same category, we use the bSDD (building SMART Data Dictionary) dictionary to obtain the categories of each attribute in the attribute sets of entity u and command class v. The set consisting of the categories of all attributes in the attribute set of entity u is recorded as the attribute category set of entity u; the set consisting of the categories of all attributes in the attribute set of command class v is recorded as the attribute category set of command class v; and the Jaccard similarity between the attribute category set of entity u and the attribute category set of command class v is recorded as the first similarity between entity u and command class v. The first similarity reflects whether the attributes of entity u and command class v belong to the same category. A larger value indicates that the attribute categories of entity u and command class v are also highly similar.

[0051] Furthermore, we can analyze the importance of entity u and command class v in their respective knowledge graphs to perform hierarchical deep matching to further analyze the possibility of association mapping.

[0052] Although knowledge graphs can be constructed from top to bottom or bottom to top, both approaches can lead to the top-level node in the knowledge graph. Therefore, the depth of entity u is calculated by denoting the shortest number of edges between the top-level node and entity u in the IFC knowledge graph, and the depth of command class v is calculated by denoting the shortest number of edges between the top-level node and command class v in the FDS knowledge graph.

[0053] The absolute difference between the depth values ​​of entity u and command class v is recorded as the first difference between entity u and command class v. The first difference can reflect the difference in importance between entity u and command class v in their respective knowledge graphs; the larger the value, the greater the hierarchical difference between entity u and command class v in the knowledge graph, and the greater the possibility that they are not suitable for association mapping.

[0054] As a preferred implementation, based on the first similarity, first difference and cross-domain semantic correlation between each entity and each command class, the graph node matching degree between each entity and each command class is obtained to characterize the degree of matching between each entity and each command class in semantics, attributes and hierarchy.

[0055] In this embodiment, the graph node matching degree between entity u and command class v is denoted as F u,v , its specific expression is: Where, F u,v A is the graph node matching degree between entity u and command class v, u,v is the cross-domain semantic association between entity u and command class v, G u,v is the first similarity between entity u and command class v, D u,vis the first difference between entity u and command class v, and τ is a preset constant.

[0056] The graph node matching degree can reflect the semantic consistency between entity u and command class v across domains and the alignment effect of the hierarchical attribute structure. The larger the value, the better the multi-dimensional matching effect between entity u and command class v, and the greater the possibility that entity u and command class v are corresponding mapping nodes in the IFC and FDS knowledge graphs.

[0057] Step 3: Based on the average level of all cross-domain semantic associations between each entity in the associated entity set of each entity and each command class in the associated command class set of each command class, the difference between the total number of entities in the associated entity set of each entity and the total number of command classes in the associated command class set of each command class, and the graph node matching degree between each entity and each command class, the graph mapping confidence between each entity and each command class is obtained, and compared with the preset threshold to obtain all mapped command classes of each entity, and then obtain the building fire simulation model.

[0058] Furthermore, since there is a connection relationship between entities in the knowledge graph, the matching degree between entity u and command class v can be finally evaluated by combining the associated nodes of entity u and command class v.

[0059] Calculate the mean of the cross-domain semantic associations between each entity in entity u's associated entity set and each command class in command class v's associated command class set, denoting this as the mean association semantics between entity u and command class v. The mean association semantics reflects the semantic consistency between entity u and command class v at the associated nodes. A larger value indicates that entity u and command class v have stronger semantic commonalities in their respective domains, enabling effective cross-domain matching and mapping.

[0060] As a preferred implementation, the graph mapping confidence between each entity and each command class is obtained based on the mean value of the associated semantics between each entity and each command class, the difference between the total number of entities in the associated entity set of each entity and the total number of command classes in the associated command class set of each command class, and the graph node matching degree between each entity and each command class. This is used to characterize the matching degree between each entity and each command class when they are mapped in the knowledge graphs of the two fields of IFC and FDS. The flowchart for obtaining the graph mapping confidence between each entity and each command class is shown in the following figure: Figure 2 shown.

[0061] In this embodiment, the confidence of the graph mapping between entity u and command class v is recorded as Z u,v , its specific expression is: Where Z u,v is the confidence of the graph mapping between entity u and command class v, F u,vis the graph node matching degree between entity u and command class v, L u,v is the mean value of the association semantics between entity u and command class v, H u,v is the absolute difference between the total number of elements in the entity set associated with entity u and the total number of elements in the command class set associated with command class v, and τ is a preset constant.

[0062] The graph mapping confidence can comprehensively reflect the degree of mapping matching between entity u and command class v when mapping in the knowledge graphs of the two fields of IFC and FDS. The larger the value, the stronger the semantic consistency and the more consistent the attribute structure between entity u and command class v, and the more consistent the semantics of the local environment in the knowledge graph, which reflects the greater the possibility that entity u and command class v are corresponding mapping nodes.

[0063] Using the above-mentioned graph mapping confidence calculation method, the graph mapping confidence between entity u and all command classes in the FDS knowledge graph is calculated. Then, all graph mapping confidences are normalized, and the command class corresponding to the graph mapping confidence that is greater than the preset threshold after normalization is recorded as the mapping command class of entity u. In this embodiment, the preset threshold is 0.5, and the implementer can choose the value according to the actual situation. Normalization methods are not limited to Z-score, maximum and minimum normalization, sigmoid normalization, etc. This embodiment adopts maximum and minimum normalization.

[0064] In the same way, all mapping command classes of all entities in the IFC knowledge graph are obtained, and a semantic mapping rule base is constructed through all mapping command classes of all entities in the IFC knowledge graph. Finally, a building fire simulation model is constructed through the IFC knowledge graph, the FDS knowledge graph and the semantic mapping rule base. The automatic conversion from the IFC building information model to the FDS fire simulation model is realized through the building fire simulation model. Among them, the construction of the building fire simulation model is a well-known technology and will not be repeated here.

[0065] If the data of the BIM model changes, the newly added or modified IFC entity is identified and its information in the IFC knowledge graph is updated; then the mapping confidence between the added or modified entity and the command class in all FDS knowledge graphs is calculated to update the semantic mapping rule base; finally, the FDS data is updated according to the updated semantic mapping rule base to realize the data update of the building fire simulation model.

[0066] It should be noted that the order in which the embodiments of the present application are presented is for illustrative purposes only and does not necessarily represent the superiority or inferiority of the embodiments. Furthermore, the foregoing descriptions of specific embodiments of this specification are provided. Furthermore, the processes depicted in the accompanying drawings do not necessarily require the specific order or sequential sequence shown to achieve the desired results. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0067] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on the differences from other embodiments.

[0068] The above-described embodiments are only used to illustrate the technical solutions of the present application, and not to limit them. Modifications to the technical solutions described in the aforementioned embodiments, or equivalent replacements of some of the technical features therein, do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present application, and should all be included in the scope of protection of the present application.

Claims

1. A method for constructing a building fire simulation model based on knowledge graph and IFC, characterized in that: The method comprises the following steps: Get the associated entity set and attribute set of each entity in the IFC knowledge graph; get the associated command class set and attribute set of each command class in the FDS knowledge graph; The root domain set of each entity and each command class is obtained based on the top-level root node corresponding to all synonyms of each entity and each command class; the cross-domain semantic association between each entity and each command class is obtained based on the distance between each entity and each command class and the similarity of the root domain set, and the graph node matching degree between each entity and each command class is obtained by combining the similarity between the categories of each attribute in the attribute set of each entity and the categories of each attribute in the attribute set of each command class, as well as the difference between the distance between each entity and the top-level node in the knowledge graph in which it is located and the distance between each command class and the top-level node in the knowledge graph in which it is located; According to the average level of all cross-domain semantic associations between each entity in the associated entity set of each entity and each command class in the associated command class set of each command class, the difference between the total number of entities in the associated entity set of each entity and the total number of command classes in the associated command class set of each command class, and the graph node matching degree between each entity and each command class, the graph mapping confidence between each entity and each command class is obtained and compared with the preset threshold to obtain all mapped command classes of each entity, and then obtain the building fire simulation model.

2. The method for constructing a building fire simulation model based on knowledge graph and IFC according to claim 1, characterized in that: The process of obtaining the root domain set of each entity and each command class is as follows: respectively obtain the synonym set of each entity and each command class; record the set consisting of the top-level root nodes corresponding to all words in the synonym set of each entity in the WordNet thesaurus as the root domain set of each entity; record the set consisting of the top-level root nodes corresponding to all words in the synonym set of each command class in the WordNet thesaurus as the root domain set of each command class.

3. The method for constructing a building fire simulation model based on knowledge graph and IFC according to claim 1, characterized in that: The calculation formula for the cross-domain semantic association between each entity and each command class is: Where A u,v is the cross-domain semantic association between entity u and command class v, B u,v is the Jaccard similarity between the root domain set of entity u and the root domain set of command class v, C u,v is the normalized Google distance between entity u and command class v, and τ is a preset constant.

4. The method for constructing a building fire simulation model based on knowledge graph and IFC according to claim 1, characterized in that: The process of obtaining the graph node matching degree between each entity and each command class is as follows: Obtaining a first similarity between each entity and each command class based on a similarity between a category of each attribute in the attribute set of each entity and a category of each attribute in the attribute set of each command class; Obtain a first difference between each entity and each command class based on the difference between the distance between each entity and the top-level node in the knowledge graph in which it resides and the distance between each command class and the top-level node in the knowledge graph in which it resides; Calculate the graph node matching degree between each entity and each command class: Where, F u,v A is the graph node matching degree between entity u and command class v, u,v is the cross-domain semantic association between entity u and command class v, G u,v is the first similarity between entity u and command class v, D u,v is the first difference between entity u and command class v, and τ is a preset constant.

5. The method for constructing a building fire simulation model based on knowledge graph and IFC according to claim 4, characterized in that: The process of obtaining the first similarity between each entity and each command class is as follows: respectively obtain the category of each attribute in the attribute set of each entity and each command class through the bSDD dictionary library, record the set composed of the categories of all attributes in the attribute set of each entity as the attribute category set of each entity; record the set composed of the categories of all attributes in the attribute set of each command class as the attribute category set of each command class; record the Jaccard similarity between the attribute category set of each entity and the attribute category set of each command class as the first similarity between each entity and each command class.

6. The method for constructing a building fire simulation model based on knowledge graph and IFC according to claim 4, characterized in that: The process of obtaining the first difference between each entity and each command class is as follows: the shortest number of edges in the path from the top-level node to entity u in the IFC knowledge graph is recorded as the depth value of entity u; the shortest number of edges in the path from the top-level node to command class v in the FDS knowledge graph is recorded as the depth value of command class v; the absolute difference between the depth values ​​of each entity and each command class is recorded as the first difference between each entity and each command class.

7. The method for constructing a building fire simulation model based on knowledge graph and IFC according to claim 1, characterized in that: The process of obtaining the confidence of the graph mapping between each entity and each command class is as follows: Obtain the mean value of the associated semantics between each entity and each command class according to the average level of all cross-domain semantic associations between each entity in the associated entity set of each entity and each command class in the associated command class set of each command class; Calculate the confidence of the graph mapping between each entity and each command class: Where Z u,v is the confidence of the graph mapping between entity u and command class v, F u,v is the graph node matching degree between entity u and command class v, L u,v is the mean value of the association semantics between entity u and command class v, H u,v is the absolute difference between the total number of elements in the entity set associated with entity u and the total number of elements in the command class set associated with command class v, and τ is a preset constant.

8. The method for constructing a building fire simulation model based on knowledge graph and IFC according to claim 7, characterized in that: The mean value of the association semantics between each entity and each command class refers to the mean value of all cross-domain semantic associations between each entity in the associated entity set of each entity and each command class in the associated command class set of each command class.

9. The method for constructing a building fire simulation model based on knowledge graph and IFC according to claim 1, characterized in that: The specific process of comparing with the preset threshold to obtain all mapping command classes of each entity is: the graph mapping confidence between each entity and all command classes in the FDS knowledge graph is recorded as the mapping command class of each entity.

10. The method for constructing a building fire simulation model based on knowledge graph and IFC according to claim 1, characterized in that: The process of obtaining the building fire simulation model is as follows: constructing a semantic mapping rule base through all mapping command classes of all entities in the IFC knowledge graph, and constructing the building fire simulation model through the IFC knowledge graph, the FDS knowledge graph and the semantic mapping rule base.

Images