Building engineering automatic compliance examination method based on AI and BIM

Abstract

The invention discloses a building engineering automatic compliance examination method based on AI and BIM, and the method comprises the steps: obtaining a to-be-examined standard text, carrying out the natural language processing, and generating a semantic information group; by adopting a semantic mapping rule, mapping the semantic information group subjected to the information conversion algorithm and entity attributes in the IFC ontology structure, and by utilizing a logic rule generator, converting a mapped target NLP deep learning rule knowledge model into a logic rule; obtaining a to-be-examined BIM model, constructing a plurality of specific IDM views, extracting entity attributes corresponding to a specific field from the to-be-examined BIM model, mapping the entity attributes in each specific IDM view and the entity attributes in the IFC ontology structure, and converting the mapped target IFC engineering feature semantic model into a logic fact by using a preset conversion tool; and performing automatic compliance review on the logic rules and the logic facts by utilizing a rule engine through a rule reasoning algorithm and an ontology semantic reasoning algorithm to obtain a rule check report.

Description

Technical Field

[0001] The present invention belongs to the technical field of construction engineering, and specifically relates to an automatic compliance review method for construction engineering based on AI and BIM. Background Art

[0002] Traditional construction engineering specification review adopts manual review based on two-dimensional drawings, which has many disadvantages such as time-consuming and labor-intensive, low efficiency and high error rate. Existing computer specification inspection is often based on hard coding and expression rule technology, which requires time-consuming manual rule extraction and coding storage, and has the problem of low efficiency. Summary of the invention

[0003] In view of this, the purpose of the present invention is to provide an automatic compliance review method for construction projects based on AI and BIM to solve the above-mentioned technical problems.

[0004] In order to achieve the above object, the present invention provides the following technical solutions:

[0005] An automatic compliance review method for construction projects based on AI and BIM, comprising:

[0006] Obtain the standard text to be reviewed for natural language processing to generate a semantic information group;

[0007] The semantic mapping rule is used to map the semantic information group after the information conversion algorithm with the entities and attributes in the IFC ontology structure, and the logical rule generator is used to convert the mapped target NLP deep learning rule knowledge model into logical rules;

[0008] Obtain the BIM model to be reviewed, build multiple specific IDM views, extract entities and attributes corresponding to specific fields from the BIM model to be reviewed, and map the entities and attributes in each specific IDM view with the entities and attributes in the IFC ontology structure. Use the preset conversion tool to convert the mapped target IFC engineering feature semantic model into logical facts.

[0009] By using the rule engine, rule reasoning algorithms and ontology semantic reasoning algorithms, automatic compliance review of logical rules and logical facts is performed to obtain a rule inspection report.

[0010] Furthermore, the standard text to be reviewed is obtained for natural language processing to generate a semantic information group, including:

[0011] Based on the three-level model of building specification text, combined with the seven-step method in the ontology construction method and the text characteristics of the specification text, the relevant concepts and terms are extracted from the specification text to be reviewed. After preprocessing, a concept term set is formed to establish the meta-ontology, domain ontology and instance ontology of the specification text, and generate semantic information groups.

[0012] The three-level model of building specification text is based on the characteristics of building standard specification text and the knowledge model of building standard specification text, and is constructed with reference to the structural hierarchy of the IFC standard. The three-level model of building specification text includes meta-ontology, domain layer and instance layer.

[0013] Furthermore, the meta-ontology is a knowledge meta-model defined based on the knowledge model of the building standard specification text, which contains abstract concepts and their corresponding relationships to define common concepts and general relationships; the meta-ontology can be reused and inherited;

[0014] Domain ontology is used to describe the concepts in a specific domain and the relationships between them;

[0015] The domain ontology inherits the concepts and relationships of the meta-ontology and has the expression framework of the meta-ontology;

[0016] An instance ontology is a specific specification object.

[0017] Furthermore, the construction process of the NLP deep learning rule knowledge model includes:

[0018] Complete the domain ontology development by inheriting the concepts and relationships in the meta-ontology;

[0019] Based on the completed domain ontology, it participates in the word segmentation NLP process, node prediction process, and relationship prediction process in the knowledge graph structure prediction model to construct a Seq2Graph model;

[0020] According to the completed domain ontology and Seq2Graph model, the construction process of the standardized knowledge graph is realized to obtain the NLP deep learning rule knowledge model.

[0021] Furthermore, the construction process of the IFC engineering feature semantic model includes: obtaining all business requirements in the building standard specification text knowledge model to form the IFC ontology structure;

[0022] The entities and attributes in all specific IDM views are mapped with the entities and attributes in the IFC ontology structure, and the IDM view ontology is edited using the protégé software to construct ontology subsets of different specific IDM views, which are used to convert the BIM models in all projects into several independent engineering feature semantic models.

[0023] An automatic compliance review method for construction projects based on AI and BIM also includes: when there is text in a rule check report that does not comply with building code standards, extracting the non-compliant text in the rule check report and the target rules associated with the non-compliant text in the NLP deep learning rule knowledge model to automatically mark the non-compliant text and form a review feedback report with the rule check report.

[0024] An automatic compliance review method for construction projects based on AI and BIM also includes: when the review feedback report lacks text that meets the business requirements of the BIM model to be reviewed, obtaining the missing requirement entity in the review feedback report and the target requirement associated with the missing requirement entity in the IFC project feature semantic model, so as to automatically annotate the review feedback report and form a new review feedback report to feed back to the user.

[0025] The beneficial effects of the present invention are:

[0026] The present invention proposes an automatic compliance review method for construction projects based on AI and BIM. By establishing a semantic description framework of IFC and building code standards, a semantic model that can fully express the characteristics of BIM models and specification clauses based on IFC is constructed, laying a foundation for computer semantic understanding and reasoning of IFC and specification knowledge; by combining with AI technologies such as rule reasoning and NLP deep learning, the ontology of specification clauses and the automatic construction mechanism of specification libraries are realized, and a BIM model intelligent specification review method based on workflow technology and artificial intelligence technology integrated is generated, which is used to solve the problem that existing computer specification checks are often based on hard coding and expression rule technology, and require time-consuming manual rule extraction and coding warehousing work, which is inefficient.

[0027] Other advantages, objectives and features of the present invention will be described in the following description and will be apparent to those skilled in the art to some extent, or they may be taught from the practice of the present invention. The purposes and other advantages of the present invention may be realized and obtained by the structures particularly pointed out in the written description and the accompanying drawings.

[0028] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0030] Figure 1 A method flow chart of an automatic compliance review method for construction projects based on AI and BIM in an embodiment of the present invention;

[0031] Figure 2 This is a schematic diagram of a specific execution process of an automatic compliance review method for construction projects based on AI and BIM in an embodiment of the present invention;

[0032] Figure 3A schematic diagram of a brief execution process of an automatic compliance review method for construction projects based on AI and BIM in an embodiment of the present invention;

[0033] Figure 4 A schematic diagram of a three-layer ontology model in an automatic compliance review method for construction projects based on AI and BIM in an embodiment of the present invention;

[0034] Figure 5 A schematic diagram of a deep learning rule knowledge model of a specification text in an automatic compliance review method for construction projects based on AI and BIM in an embodiment of the present invention;

[0035] Figure 6 This is a schematic diagram of the difference between IFC and IDM in an automatic compliance review method for construction projects based on AI and BIM in an embodiment of the present invention. DETAILED DESCRIPTION

[0036] The preferred embodiments of the present invention are described below in conjunction with the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0037] like Figure 1 , Figure 2 and Figure 3 As shown, the present invention proposes an automatic compliance review method for construction projects based on AI and BIM, comprising:

[0038] S101, obtaining the standard text to be reviewed, performing natural language processing, and generating a semantic information group;

[0039] S102. Using a semantic mapping rule, the semantic information group after the information conversion algorithm is mapped with the entities and attributes in the IFC ontology structure, and using a logical rule generator, the mapped target NLP deep learning rule knowledge model is converted into a logical rule; the target NLP deep learning rule knowledge model is corresponding to the specification text to be reviewed, and is obtained by mapping the semantic information group after the information conversion algorithm with the entities and attributes in the IFC ontology structure to the total NLP deep learning rule knowledge model;

[0040] S103, obtaining the BIM model to be reviewed, constructing multiple specific IDM views, extracting entities and attributes corresponding to specific fields from the BIM model to be reviewed, and mapping the entities and attributes in each specific IDM view with the entities and attributes in the IFC ontology structure, and using a preset conversion tool to convert the mapped target IFC engineering feature semantic model into a logical fact; the preset conversion tool includes but is not limited to converting the target IFC engineering feature semantic model into a logical fact using basic conversion rules and linear transformation rules; the target IFC engineering feature semantic model is corresponding to the BIM model to be reviewed, and is obtained by mapping the entities and attributes in each specific IDM view with the entities and attributes in the IFC ontology structure to the overall IFC engineering feature semantic model;

[0041] S104. Using the rule engine, through the rule reasoning algorithm and the ontology semantic reasoning algorithm, automatically conduct compliance review on the logic rules and logic facts to obtain a rule check report;

[0042] The beneficial effects of the above technical solution are as follows: based on the introduction of ontology and semantic web technology to model IFC text and building specification standard knowledge, this application establishes a rule reasoning mechanism by integrating artificial intelligence technologies such as natural language processing and deep learning, and constructs a technical framework for intelligent specification review of building projects based on IFC (Industry Foundation Classes) ontology to complete automatic compliance review of building projects; by establishing a semantic description framework of IFC and building specification standards, a semantic model that can fully express the model characteristics and specification provisions of BIM (Building Information Modeling) based on IFC is constructed, laying the foundation for computer semantic understanding and reasoning of IFC and specification knowledge; by combining with AI technologies such as rule reasoning and NLP (Natural Language Processing) deep learning, the ontology of specification provisions and the automatic construction mechanism of the specification library are realized, and a BIM model intelligent specification review method based on workflow technology and artificial intelligence technology is generated to solve the problem that existing computer specification checks are often based on hard coding and expression rule technology, and require time-consuming manual rule extraction and coding and warehousing work, which is inefficient.

[0043] In one embodiment, the standard text to be reviewed is obtained and natural language processing is performed to generate a semantic information group, including:

[0044] Based on the three-level model of building specification text, combined with the seven-step method in the ontology construction method and the text characteristics of the specification text, the relevant concepts and terms are extracted from the specification text to be reviewed. After preprocessing, a concept term set is formed to establish the meta-ontology, domain ontology and instance ontology of the specification text, and generate semantic information groups.

[0045] The three-level model of building specification text is based on the characteristics of building standard specification text and the knowledge model of building standard specification text, and is constructed with reference to the structural hierarchy of the IFC standard. The three-level model of building specification text includes meta-ontology, domain layer and instance layer.

[0046] Meta-ontology is a knowledge meta-model defined based on the knowledge model of the building standard specification text, which contains abstract concepts and their corresponding relationships, and is used to define common concepts and general relationships; meta-ontology can be reused and inherited;

[0047] Domain ontology is used to describe the concepts in a specific domain and the relationships between them;

[0048] The domain ontology inherits the concepts and relationships of the meta-ontology and has the expression framework of the meta-ontology;

[0049] The instance ontology is a specific normative object;

[0050] The working principle and beneficial effects of the above technical solution are as follows: Based on the characteristics of the building code itself and the knowledge model of the building standard specification text, refer to the structural hierarchy of the IFC standard, establish Figure 4 The three-layer ontology model shown in the figure, among which, the meta-ontology is a knowledge meta-model defined based on the above knowledge model, which contains the most abstract concepts and their relationships, and mainly defines common concepts and general relationships; it is the public semantic framework of the normative knowledge model, which can be reused and inherited; the domain layer ontology is used to describe the concepts in a specific field and the relationship between concepts; each standard specification has its specific applicable domain scope, corresponding to the specific concepts and terms in the field, and constitutes the specific domain knowledge; the domain ontology inherits the concepts and relationships of the meta-ontology and has the expression framework of the meta-ontology; the instance ontology is a specific normative object, such as the concrete engineering-frame structure-frame beam instance;

[0051] Based on the above three-layer structure, combined with the seven-step method in the ontology construction method and the characteristics of the normative text document, relevant concepts and terms are extracted from the normative text. After continuous streamlining, refinement, supplementation and modification, a basic set of concepts and terms is formed, thereby establishing the normative text meta-ontology, domain ontology and instances; normative knowledge modeling and retrieval are closely related. The normative knowledge model will directly affect the information available for retrieval and the design of the retrieval scheme, and effective retrieval is the basis for the intelligent application of normative knowledge.

[0052] In one embodiment, the process of building the NLP deep learning rule knowledge model includes:

[0053] Complete the domain ontology development by inheriting the concepts and relationships in the meta-ontology;

[0054] Based on the completed domain ontology, it participates in the word segmentation NLP process, node prediction process, and relationship prediction process in the knowledge graph structure prediction model to construct a Seq2Graph model;

[0055] According to the completed domain ontology and Seq2Graph model, the construction process of the standardized knowledge graph is realized to obtain the NLP deep learning rule knowledge model;

[0056] The NIP deep learning rule knowledge model is a general NLP deep learning rule knowledge model. The concepts and relationships in the meta-ontology include the concepts and relationships in all building code standard texts.

[0057] The working principle and beneficial effects of the above technical solution are as follows: As described in the background technology, it is difficult to implement a large number of building specification ontology and rule reasoning and transformation by manual means. This application preferably adopts a deep learning method combining GCN (Graph Convolutional Network: graph convolutional neural network) and LSTM (Long short-term memory, long short-term memory recursive neural network), establishes a suitable learning model to automatically annotate the ontology concepts corresponding to the specification text, establishes a mapping relationship between the vocabulary set and the concept model, and forms a rule knowledge graph;

[0058] The deep learning rule knowledge model of the standard text is formed as follows Figure 5 As shown in the figure, for the standard text, the text needs to be preprocessed first; after sentence segmentation and regular expression, the required standards are extracted to form a rule set; for each rule in the rule set, the NLPIR Chinese word segmentation system is used to segment it, that is, the sentence is divided into several complete words; the commonly used word segmentation method is based on the existing common word library, and the high-frequency word graph scanning is realized through the parsing tree structure, and the word segmentation problem is converted into a dynamic programming problem to solve the maximum segmentation combination based on word frequency; for the unregistered, the HMM hidden Markov model and the Viterbi algorithm are used for word segmentation; the semantic analysis of the rules after word segmentation is carried out to obtain its ontology features, and these features are semantically annotated, that is, the so-called ontology annotation;

[0059] By extracting and analyzing the knowledge points contained in the standard clauses, the annotation ontology is constructed; the concepts, attributes, and relationships between concepts contained in the annotation ontology are used to semantically annotate the standard clauses to form a semantic annotation library of the standard clauses; the annotation information of the standard clauses is stored in the annotation library in the form of information and attachments; the annotation library associates the concepts and relationships in the annotation ontology with the corresponding standard clauses in the form of relationship instances or concept instances;

[0060] After defining the semantic annotation method of normative clauses, it is necessary to solve the annotation problem of a large amount of text; through the semantic annotation of the contents of these normative documents, the knowledge points of the clauses distributed in various parts of the document or in different documents are reorganized in the form of an ontology knowledge model, realizing the ontology-based organization and expression of normative clause knowledge with mutual relevance and logical relationships; the annotated information has semantic associations based on the semantics implied in the annotated ontology, which constitutes the basis for semantic association retrieval of normative clauses.

[0061] In one embodiment, the construction process of the IFC engineering feature semantic model includes: obtaining all business requirements in the building standard specification text knowledge model to form an IFC ontology structure;

[0062] Map the entities and attributes in all specific IDM views to the entities and attributes in the IFC ontology structure, and use the protégé software to edit the IDM view ontology to construct ontology subsets of different specific IDM views, which are used to convert the BIM models in all projects into several independent engineering feature semantic models;

[0063] The IFC engineering feature semantic model is the overall IFC engineering feature semantic model. The entities and attributes in all specific IDM views refer to the project requirements of all building project types. The entities and attributes in the above IFC ontology structure refer to all individual entities and attributes after ontologizing the project requirements in all building code standard texts.

[0064] The working principle and beneficial effects of the above technical solution are as follows: The purpose of IFC is to support all business requirements of all project stages, that is, the sum of all information; but usually the information to be exchanged is about a specific subject (such as structural design, specification review, etc.), and its purpose is to support a business requirement in one or several stages. Its main task is to determine which basic elements of IFC meet this business requirement, that is, IDM, such as Figure 6 As shown;

[0065] Compliance reviews based on different building standards and specifications also require different engineering information and data. For example, compliance reviews based on fire protection standards require the extraction of fire protection-related building information. This application establishes independent building model IDM views, and extracts a certain field of data subsets and related information from the building information model through specific IDM views.

[0066] After extracting relevant building information, it is necessary to map the entities and attributes in each IDM view with the entity attributes in the IFC ontology relationship, and construct ontology subsets of different IDM views, so as to convert the BIM model in the project into engineering feature semantic models; by constructing various types of building model IDM views, extracting the building information data required for specification review and constructing engineering feature semantic models, the efficiency of compliance review based on specifications and standards can be greatly improved; this application uses the protégé developed by Stanford University to edit the IDM (Internet Download Manager) view ontology, and realizes the construction of the IFC engineering feature semantic model through ontology mapping between the IDM view ontology and ifcOWL (IFC ontology, OWL-Web Ontology Language).

[0067] In one embodiment, when there is text in the rule check report that does not comply with building code standards, the non-compliant text in the rule check report and the target rules associated with the non-compliant text in the NLP deep learning rule knowledge model are extracted to automatically annotate the non-compliant text and form a review feedback report with the rule check report.

[0068] In one embodiment, when the review feedback report lacks text that meets the business requirements of the BIM model to be reviewed, the missing requirement entity in the review feedback report and the target requirement associated with the missing requirement entity in the IFC engineering feature semantic model are obtained to automatically annotate the review feedback report and form a new review feedback report to be fed back to the user.

[0069] Finally, it should be noted that the above preferred embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail through the above preferred embodiments, those skilled in the art should understand that various changes can be made in form and details without departing from the scope defined by the claims of the present invention.

Claims

1. An automatic compliance review method for construction projects based on AI and BIM, characterized in that: include: Obtain the standard text to be reviewed for natural language processing to generate a semantic information group; The semantic mapping rule is used to map the semantic information group after the information conversion algorithm with the entities and attributes in the IFC ontology structure, and the logical rule generator is used to convert the mapped target NLP deep learning rule knowledge model into logical rules; Obtain the BIM model to be reviewed, build multiple specific IDM views, extract entities and attributes corresponding to specific fields from the BIM model to be reviewed, and map the entities and attributes in each specific IDM view with the entities and attributes in the IFC ontology structure. Use the preset conversion tool to convert the mapped target IFC engineering feature semantic model into logical facts. By using the rule engine, the logic rules and logic facts are automatically checked for compliance through the rule reasoning algorithm and the ontology semantic reasoning algorithm to obtain a rule checking report.

2. According to claim 1, the method for automatic compliance review of construction projects based on AI and BIM is characterized in that: Obtain the standard text to be reviewed for natural language processing and generate a semantic information group, including: Based on the three-level model of building specification text, combined with the seven-step method in the ontology construction method and the text characteristics of the specification text, the relevant concepts and terms are extracted from the specification text to be reviewed. After preprocessing, a concept term set is formed to establish the meta-ontology, domain ontology and instance ontology of the specification text, and generate semantic information groups. The three-level model of building specification text is based on building standard specification text characteristics and building standard specification text knowledge model, and is constructed with reference to the structural hierarchy of the IFC standard. The three-level model of building specification text includes a meta-ontology, a domain layer and an instance layer.

3. The method for automatic compliance review of construction projects based on AI and BIM according to claim 2 is characterized in that: The meta-ontology is a knowledge meta-model defined based on the knowledge model of the building standard specification text, including abstract concepts and their corresponding relationships, which are used to define common concepts and general relationships; the meta-ontology can be reused and inherited; The domain ontology is used to describe the concepts in a specific domain and the relationships between the concepts; The domain ontology has the expression framework of the meta-ontology by inheriting the concepts and relationships of the meta-ontology; The instance ontology is a specific specification object.

4. The method for automatic compliance review of construction projects based on AI and BIM according to claim 3 is characterized in that: The construction process of the NLP deep learning rule knowledge model includes: Complete the domain ontology development by inheriting the concepts and relationships in the meta-ontology; Based on the completed domain ontology, it participates in the word segmentation NLP process, node prediction process, and relationship prediction process in the knowledge graph structure prediction model to construct a Seq2Graph model; According to the completed domain ontology and Seq2Graph model, the construction process of the standardized knowledge graph is realized to obtain the NLP deep learning rule knowledge model.

5. The method for automatic compliance review of construction projects based on AI and BIM according to claim 1 is characterized in that: The construction process of the IFC engineering feature semantic model includes: Acquire all business requirements in the building standard specification text knowledge model to form the IFC ontology structure; The entities and attributes in all specific IDM views are mapped with the entities and attributes in the IFC ontology structure, and the IDM view ontology is edited using the protégé software to construct ontology subsets of different specific IDM views, which are used to convert the BIM models in all projects into several independent engineering feature semantic models.

6. The method for automatic compliance review of construction projects based on AI and BIM according to claim 4 is characterized in that: Also includes: When there is text in the rule check report that does not comply with building code standards, the non-compliant text in the rule check report and the target rules associated with the non-compliant text in the NLP deep learning rule knowledge model are extracted to automatically annotate the non-compliant text and form a review feedback report with the rule check report.

7. An automatic compliance review method for construction projects based on AI and BIM according to any one of claims 5 or 6, characterized in that: Also includes: When the review feedback report lacks text that meets the business requirements of the BIM model to be reviewed, the missing requirement entity in the review feedback report and the target requirement associated with the missing requirement entity in the IFC engineering feature semantic model are obtained to automatically annotate the review feedback report and form a new review feedback report to be fed back to the user.

Images