CN112784346B

CN112784346B - Building structure autonomous design method, system, terminal and storage medium

Info

Publication number: CN112784346B
Application number: CN202110178988.0A
Authority: CN
Inventors: 史健勇
Original assignee: Dianhui Space Shanghai Information Technology Co ltd
Current assignee: Dianhui Space Shanghai Information Technology Co ltd
Priority date: 2021-02-07
Filing date: 2021-02-07
Publication date: 2022-12-06
Anticipated expiration: 2041-02-07
Also published as: CN112784346A

Abstract

The invention provides an artificial intelligence-based building structure autonomous design method and system, which comprises the following steps: extracting structural design knowledge of historical design, and establishing a structural design field ontology; establishing a Bayesian network model, taking structural design knowledge contained in a structural design field ontology as input of a Bayesian network, calculating conditional probability distribution, and outputting a design recommendation scheme; generating a building information model corresponding to the scheme according to the output design recommendation scheme; and verifying the structural design scheme of the generated building information model, feeding back a verification result to the building information model, and optimizing the building information model. The invention provides technical support for high efficiency and intellectualization of structural design based on artificial intelligence of the ontology and the Bayesian network, reduces the workload of design employees in the stages of scheme design and preliminary design, improves the design efficiency, reduces the design cost, effectively utilizes the historical design experience, and has innovation and expansion significance.

Description

Building structure autonomous design method, system, terminal and storage medium

Technical Field

The invention relates to an auxiliary technology in the field of building structure design, in particular to an artificial intelligence-based building structure autonomous design method, an artificial intelligence-based building structure autonomous design system, a terminal and a storage medium.

Background

At the present stage of the country, the combination of the civil engineering field and modern information technology is greatly promoted, the labor cost and the time cost of building design, construction and delivery are reduced through modern technical means, and the building intellectualization is promoted. The building structure design is taken as an important link of a civil engineering project, and the following problems exist at present:

1. there are duplicative designs of similar structures in conventional structural designs. For large stadiums, super high-rise buildings or urban landmark buildings and the like, the building modeling is complex and is influenced by geology and the like, and specific analysis and design needs to be carried out on projects, but for traditional multi-story buildings such as houses and teaching buildings, the building modeling is regular, stress transmission is clear, the design requirement of a structural designer is not high, and the waste of manpower and time cost is caused.

2. The design experience is difficult to popularize. Design experience is very important in structural design, and abundant engineering experience is favorable for shortening design time and improving engineering efficiency. The building design process involves a great deal of knowledge and experience and is rather complex and cannot be represented by simple rules or models. General design experience usually needs to be accumulated through projects, and rapid succession, popularization and application through teaching in a short period are difficult.

3. There is a lack of archive management for structural designs. The existing management method of CAD construction drawing is continuously used for filing structural design, knowledge management of Building Information Models (BIM) is lacked, knowledge extraction and filing based on design elements such as structural design principles, structural body types and components are not available, search of design history documents can be performed only through classified keywords, and key design Information is difficult to extract from history design quickly.

4. There is a lack of efficient use of knowledge in historical design computer books. In the past, the research on the BIM technology only aims at expanding the BIM related technology and the Industry basic Class (IFC) standard of the acknowledged standard thereof, neglects the combination with the traditional building experience, and especially lacks the effective extraction and utilization of the knowledge in the historical design calculation book.

5. Design project parameterization, modularization level are lower, and the human cost is effectively reduced. Three-dimensional modeling of a complete project by hand still requires a significant amount of time and labor. The high-modularization products in the industrial production field can promote the development and analysis options of the design, and have important application value when being introduced into the building industry. In structural design, the modular design can make the design have stronger logicality, regularity and learnability.

Meanwhile, under the background of the development of the informatization technology, the ontology and the uncertain artificial intelligence are developed vigorously, and important realization bases are provided for promoting the modernization development of the civil engineering field and realizing the building intelligence. The ontology mainly refers to an abstract representation for describing concepts in the world or a certain field and mutual relations among attributes and concepts in computer science, and has a lot of achievements on solving the problems of field knowledge management, information retrieval, logical reasoning and the like; the uncertain artificial intelligence is an important means of data mining nowadays, wherein Bayesian Network (BN) is widely applied, is one of the most effective theoretical models in the uncertain knowledge expression and reasoning fields at present, and has important application in the fields of document classification, medical diagnosis, information retrieval, statistical decision, learning and prediction and the like.

Through search, the following results are found:

the Chinese patent application with the application number of 201811570300.8 and the invention name of building design method based on machine learning and BIM technology breaks through the limitation of the traditional building design method, combines the optimization of machine learning and the information and visualization characteristics of BIM technology with high efficiency, reduces the problem of unreasonable design of the traditional design method, and improves the efficiency of building design and the product quality; meanwhile, the building user can express own design requirements more directly, the participation degree of the building user in the building design stage is increased, and a design team is better helped to cooperate with the building user to complete project design. The method considers the influence of the physical attributes of the building components on the design parameters, establishes an influence relation and an influence equation, divides the design area into regular square sub-areas by using a finite element analysis algorithm on the basis of determining the building boundary, and then performs transmission type analysis and calculation from the building boundary to the center by using the established influence equation. The method takes building designs of different styles as training samples to serve as training samples of an intelligent design model, and the intelligent design model adopts a decision tree model. Through the discussion of a design team and a building user, the priority of target design parameters is determined, and according to the priority, an intelligent design model realizes the screening of a plurality of feasible design scheme sets generated by design, and finally a building design scheme meeting the expectation of the building user is obtained. However, the method still has the following problems:

1. when the influence of the building elements on the design parameters is determined by human, the correctness of an influence equation cannot be ensured, so that the reasonable schemes cannot be ensured in various building schemes generated by the design parameters.

2. The design parameters related in the method only relate to parameters in the aspect of building construction functions, such as building area, natural daylighting rate and the like, only the design of building function layout is completed, and the design of a structural scheme of a building is not realized.

Disclosure of Invention

Aiming at the defects in the prior art, the invention provides an artificial intelligence-based building structure autonomous design method, an artificial intelligence-based building structure autonomous design system, a terminal and a storage medium.

According to an aspect of the present invention, there is provided a building structure autonomous design method, including:

extracting structural design knowledge of historical design, and establishing a structural design field ontology;

establishing a Bayesian network model, taking the structural design knowledge contained in the structural design field ontology as the input of a Bayesian network, calculating conditional probability distribution, and outputting a design recommendation scheme;

generating a building information model corresponding to the scheme according to the design recommendation scheme;

and verifying the structural design scheme of the building information model, feeding back a verification result to the building information model, and optimizing the building information model.

According to another aspect of the present invention, there is provided an autonomous design system of a building structure, comprising:

the structural design field ontology construction module is used for extracting structural design knowledge of historical design and establishing a structural design field ontology;

a design recommendation generation module for establishing a Bayesian network model, taking the structural design knowledge contained in the structural design domain ontology as the input of the Bayesian network, calculating the conditional probability distribution, and outputting a design recommendation;

the building information model generating module generates a building information model corresponding to the scheme according to the output design recommendation scheme;

and the building information model optimizing module is used for verifying the structural design scheme of the generated building information model, feeding back the verification result to the building information model and optimizing the building information model.

According to a third aspect of the present invention, there is provided a terminal comprising a memory, a processor and a computer program stored on the memory and executable on the processor, the processor when executing the program being operable to perform any of the methods described above.

According to a fourth aspect of the invention, there is provided a computer-readable storage medium, having stored thereon a computer program, which, when executed by a processor, is operable to perform the method of any of the above.

Due to the adoption of the technical scheme, compared with the prior art, the invention has the following beneficial effects:

aiming at the problems of design decision, design information lookup, design information management and the like in actual structural design in the prior art, the artificial intelligence based on the ontology and the Bayesian network provides technical support for high efficiency and intellectualization of structural design by means of advanced computers and artificial intelligence technology, reduces the workload of design employees in the stages of scheme design and preliminary design, improves the design efficiency, reduces the design cost, effectively utilizes the historical design experience, and has innovation and expansion significance.

The building structure autonomous design method, the building structure autonomous design system and the building structure autonomous design terminal based on artificial intelligence provided by the invention take the structure intelligent design as a development target, utilize the ontology as a structure design knowledge framework, utilize the Bayesian network as structure design knowledge logic, and utilize the BIM technology as a structure design implementation platform, so that the knowledge extraction, design scheme recommendation and calculation model modularization establishment of the structure design are realized.

The building structure autonomous design method, the building structure autonomous design system and the building structure autonomous design terminal based on artificial intelligence can effectively utilize historical design experience and provide a new idea for the retention and development of design knowledge.

The building structure autonomous design method, the building structure autonomous design system and the building structure autonomous design terminal based on artificial intelligence can solve the problems of design decision, design information query, design information management and the like in actual structure design.

According to the building structure autonomous design method, system and terminal based on artificial intelligence, a Bayesian network method is adopted to realize primary scheme design, the model is obtained by learning from a plurality of real samples, and the influence of design parameters on a design result has higher reliability.

Drawings

Other features, objects and advantages of the invention will become more apparent upon reading of the detailed description of non-limiting embodiments with reference to the following drawings:

fig. 1 is a flowchart of an artificial intelligence-based building structure autonomous design method according to an embodiment of the present invention.

FIG. 2 is a block diagram of an artificial intelligence based architecture design system according to an embodiment of the present invention.

Fig. 3 is a schematic diagram of the operational relationship between the constituent modules of the artificial intelligence-based building structure autonomous design system in a preferred embodiment of the present invention.

FIG. 4 is a block diagram of a framework for extracting knowledge of structural design according to a preferred embodiment of the present invention.

Fig. 5 is a flow chart of building a ontology in the field of structural design according to a preferred embodiment of the present invention.

Fig. 6 is a schematic diagram of the ontology of the structural design domain established in the preferred embodiment of the present invention.

FIG. 7 is a flow chart of extracting basic design information of a building structure from a design calculation book in a preferred embodiment of the present invention.

Fig. 8 is a schematic diagram of the process of extracting the design basic information of the building structure in combination with the actual content in a preferred embodiment of the present invention.

FIG. 9 is a flow chart of extracting geometric information of a building structure from IFC data in a preferred embodiment of the present invention.

FIG. 10 is a diagram illustrating the instantiation results of the structure calculation sub-ontology in a preferred embodiment of the present invention.

Fig. 11 is a schematic diagram illustrating an instantiation result of a building structure sub-ontology in a preferred embodiment of the present invention.

FIG. 12 is a flow chart of a naive Bayes based structure selection design in a preferred embodiment of the invention.

Fig. 13 is a diagram illustrating a naive bayes model established in a preferred embodiment of the invention.

Fig. 14 is a schematic diagram of the sub-network structure of the building structure design according to a preferred embodiment of the present invention.

Fig. 15 is a diagram of a bayesian network after a single pruning in a preferred embodiment of the present invention.

FIG. 16 is a diagram of a Bayesian network after the second pruning in a preferred embodiment of the present invention.

Fig. 17 is a schematic diagram of three common teaching buildings according to a preferred embodiment of the present invention.

FIG. 18 is a flowchart illustrating the intelligent generation of standard models in accordance with a preferred embodiment of the present invention.

FIG. 19 is a schematic diagram of the generation of the axle net, elevation and column net in a preferred embodiment of the present invention.

FIG. 20 is a schematic diagram of a batch modification process of components in a preferred embodiment of the present invention.

FIG. 21 is a flow chart of the Revit-YJK structure design in a preferred embodiment of the present invention.

FIG. 22 is a plan view of a teaching floor in accordance with an embodiment of the present invention.

FIG. 23 is a flow chart of an intelligent design method for a structure according to an embodiment of the present invention

FIG. 24 is a diagram illustrating conditional probability distribution results in an embodiment of the present invention.

Fig. 25 is a schematic diagram of an input interface of a basic structural design scheme according to an embodiment of the present invention.

Detailed Description

The following examples illustrate the invention in detail: the embodiment is implemented on the premise of the technical scheme of the invention, and gives a detailed implementation mode and a specific operation process. It should be noted that various changes and modifications can be made by those skilled in the art without departing from the spirit of the invention, and these changes and modifications are all within the scope of the invention.

As shown in fig. 1, the method for autonomously designing a building structure based on artificial intelligence provided by this embodiment may include the following steps:

s100, extracting structural design knowledge of historical design, and establishing a structural design field body;

s200, establishing a Bayesian network model, taking structural design knowledge contained in a structural design field ontology as input of a Bayesian network, calculating conditional probability distribution, and outputting a design recommendation scheme;

s300, generating a building information model corresponding to the scheme according to the output design recommendation scheme;

and S400, verifying the structural design scheme of the generated building information model, feeding back a verification result to the building information model, and optimizing the building information model.

In S100 of this embodiment, as a preferred embodiment, extracting the structural design knowledge and establishing the structural design domain ontology may include the following steps:

s101, extracting geometric information of a building structure and design basic information of the building structure from IFC data and a design calculation book of historical design respectively to form structural design knowledge;

s102, respectively establishing a building structure sub-body and a structure calculation sub-body;

s103, instantiating the established building structure sub-ontology and the structure calculation sub-ontology respectively by using the extracted structural design knowledge to obtain instantiation files of the building structure sub-ontology and the structure calculation sub-ontology;

and S104, aligning and integrating the instantiation files of the building structure sub-ontology and the structure calculation sub-ontology to obtain a complete instantiation file of the structure design domain ontology, and obtaining the structure design domain ontology.

In S101 of this embodiment, as a preferred embodiment, extracting basic design information of the building structure may include the following steps:

s1011, generating a text relative rule of the calculation book, and acquiring structure calculation related content through the text relative rule generated by text regular processing;

and S1012, adding identification attributes to the related contents calculated for each type of structure to obtain the design basic information of the building structure.

In S101 of this embodiment, as a preferred embodiment, extracting geometric information of the building structure may include the following steps:

S101I, analyzing the IFC data and extracting the building structure related attribute information from the IFC data;

and S101II, retrieving and classifying example elements in the IFC data, sequentially analyzing and acquiring corresponding attribute pairs according to the attribute information extraction link, and finally obtaining example information, namely the geometric information of the building structure.

In S102 of this embodiment, as a preferred embodiment, a seven-step method may be adopted, and a building structure sub-ontology and a structure calculation sub-ontology are built based on a prot g e platform, where the building structure sub-ontology and the structure calculation sub-ontology respectively correspond to a model building stage and a calculation analysis stage in structure design calculation.

In S103 of this embodiment, as a preferred embodiment, instantiating the built building structure sub-ontology may include the following steps:

and S1031, establishing a mapping rule file according to the established building structure sub-ontology and the building structure information in the IFC data file.

S1032, the mapping rule file and the building structure information table data are obtained, and the building structure sub-ontology instantiation file is generated.

In one embodiment of the present invention, the substrate is,

and S1031, establishing a mapping rule file required by the D2RQ according to the established building structure sub-ontology and the building structure information in the IFC data file converted into the relational database MySQL.

And S1032, calling the D2RQ open source tool to read the mapping rule file and the building structure information table data in the MySQL, and generating a building structure sub-ontology instantiation file.

In S103 of this embodiment, as a preferred embodiment, instantiating the built structural computation sub-ontology may include the following steps:

and S103I, establishing a mapping rule file according to the established structure calculation sub-body and the structure calculation information in the structure calculation book file.

S103II, obtaining the mapping rule file and the structure calculation information table data, and generating a structure calculation sub-ontology instantiation file.

In one embodiment of the present invention, the substrate is,

S103I, establishing a mapping rule file required by D2RQ according to the established structure calculation sub-body and the structure calculation information in the structure calculation book file converted into the relational database MySQL.

S103II, calling a D2RQ open source tool to read the mapping rule file and the structure calculation information table data in MySQL, and generating a structure calculation sub-ontology instantiation file.

In S200 of this embodiment, as a preferred embodiment, establishing a bayesian network model, taking the structural design knowledge included in the ontology of the structural design domain as an input of the bayesian network, calculating a conditional probability distribution, and outputting a design recommendation scheme may include the following steps:

s201, generating a logic relation of a Bayesian network by adopting a domain expert system, and constructing a Bayesian network model;

s202, classifying the structural design knowledge contained in the structural design field ontology, and inputting the classified structural design knowledge as a model input sample to obtain conditional probability distribution;

and S203, taking the structural design knowledge contained in the structural design field ontology as an event node of the Bayesian network, and outputting a design recommendation scheme according to the conditional probability distribution.

In S200 of this embodiment, as a preferred embodiment, the following steps may be further performed:

the logical relationships of the bayesian network model are pruned and/or populated.

In S300 of this embodiment, as a preferred embodiment, generating the building information model corresponding to the design recommendation according to the design recommendation may include the following steps:

s301, determining span segments of standard trusses according to the design scheme output based on the Bayesian model, marking key information of the shaft gateway, and constructing the standard model.

S302, the key information of the axle gateway is determined by taking the standard model as a template, the axle layout is completed for a new project based on the standard model, and a modularized axle network is generated.

S303, based on the modularized axle network, completing the generation of the design component, and finally generating the initially designed building information model.

In one embodiment, the shaft gateway key information includes: bay, span, etc.

In one embodiment, the design component comprises: beams, columns, walls, etc.

In S400 of this embodiment, as a preferred embodiment, the method includes performing structural design scheme verification on the generated building information model, feeding back a verification result to the building information model, and optimizing the building information model, and may further include the following steps:

and (4) aiming at the optimized building information model, deriving the structure design knowledge contained in the optimized building information model to be used as a new IFC data and design calculation book data sample.

The building structure autonomous design method based on artificial intelligence provided by the embodiment is based on the ontology, the Bayesian network model and the building information model, and realizes intelligent building structure autonomous design.

Another embodiment of the present invention provides an artificial intelligence-based building structure autonomous design system, as shown in fig. 2, which may include the following modules: the system comprises a structural design field body construction module, a design recommendation scheme generation module, a building information model generation module and a building information model optimization module.

Wherein:

The following describes in detail the preferred embodiments and design principles of the present invention with reference to the accompanying drawings.

The building structure autonomous design method and system based on artificial intelligence provided by the above embodiments of the present invention is an artificial intelligence building structure autonomous design technology based on ontology and bayesian network, and mainly includes four modules: the building structure autonomous design method based on artificial intelligence provided by the embodiment of the invention can be implemented through the system.

As shown in fig. 3, a schematic diagram of an operational relationship between modules is shown, in fig. 3, the operational relationship between the modules includes: the module structure design field ontology construction module is a structure design knowledge extraction module based on an ontology and provides a data base for the system; a second module is used for designing a recommendation scheme generation module, namely a design scheme recommendation module based on a Bayesian model, and providing algorithm support for system application; a building information model generation module III, namely a BIM-based structural calculation model generation function module, which converts an abstract design scheme into a concrete information model; and a fourth module, namely a structural calculation and knowledge reuse function module, is used as a verification module to play a role in scheme verification and can perform further optimization on the basis.

1. The body construction module in the structural design field is a functional module for extracting structural design knowledge based on the body: and extracting structural design knowledge. And establishing a structural design field ontology by taking a data file and a design calculation book corresponding to the universal data standard IFC in the BIM technology as data sources. The intelligent design method needs to be based on data research of historical design cases, so knowledge extraction of the design cases is the basis of the method. For knowledge extraction of cases, a structural design domain ontology is established as a design knowledge framework, and the integrity of information extraction is ensured by extracting two data sources from IFC data and extracting a data source combination form from computer book text.

2. The design recommendation scheme generation module is a design recommendation function module based on a Bayesian model: and finishing data classification aiming at the historical case data, establishing a Bayesian network model, and recommending a design scheme for a new project. The extracted original data is firstly classified according to two attributes, namely text attribute and digital attribute, which are contained in the design project, namely discrete data and continuity data. And analyzing the design logic, and establishing a structural design Bayesian network model by using a method of combining domain expert knowledge and sample data analysis. And parameter learning is completed by adopting an EM (effective expectation maximization) algorithm aiming at the Bayesian network, so that the influence of a sample missing value on network probability distribution is reduced, the accuracy of a Bayesian model is improved, and design reference is provided for structural design.

3. The building information model generation module is a BIM-based structural calculation model generation functional module: according to the output design recommendation scheme, parameterization and modularized generation of the structural scheme design BIM are carried out by utilizing a secondary development environment of a structural column network standard template and BIM related software (in a specific application example, autodesk Revit software can be used), design parameters are adjusted, and the BIM corresponding to the design alternative scheme can be output in batch.

4. The building information model optimization module is a structure calculation and knowledge reuse functional module: and extracting a structural calculation model based on the BIM design model generated by the recommended scheme, importing the structural calculation model into structural analysis software (in a specific application example, the encyclopedia software can be used), verifying the structural design scheme, and proving the rationality of the technical process provided by the system through the scheme after analysis and verification.

The embodiment provides a building structure autonomous design system based on artificial intelligence, and the specific technical scheme is as follows:

module I and body building module in structural design field

The module derived base technology route may be as shown in the upper left dashed box of fig. 3. The ontology-based structural design knowledge extraction involved in the preferred embodiment mainly includes extracting structural design features (i.e., structural design knowledge) from BIM model files (i.e., IFC data) and design computer books involved in historical design and stored in the form of IFC data standards. The structural design knowledge extraction can manage structural design projects, and meanwhile, effective data samples can be provided for a subsequent data mining module through obtaining structural data.

The general architectural structure design parameter attribute comprises two parts: the method comprises the following steps that firstly, geometric information of a building structure comprises geometric parameters of components, building span, building elevation, axis network information and the like, the information determines the design form of the building structure and is information displayed for a structural calculation model; and secondly, design basic information of the building structure, including building name, building geographical position, building site condition, building space usage and the like, is generally used as background setting information during structure calculation, and plays a vital role in the structure design. Therefore, the data carriers are extracted from the IFC data file and the computer book text data file respectively.

Dividing the knowledge content contained in the structural design ontology according to the design depth, wherein the method comprises three stages: the design method comprises the steps of building overall design, building space design and building element design, wherein the design contents respectively correspond to the design contents of a scheme design stage, a primary design stage and a construction drawing design stage of structural design, the structural design contents such as design overall requirements, spatial arrangement, load distribution, earthquake-resistant design and element design are embodied, and as shown in figure 4, a structural design knowledge extraction framework is provided. The method comprises the following steps of constructing a structural design field body based on a prot ge platform by adopting a seven-step method, wherein the general flow is shown in fig. 5, and the constructed structural design field body is shown in fig. 6. The body is divided into two sub-bodies of a building structure and a structure calculation, and the two sub-bodies respectively correspond to two stages of model establishment and calculation analysis in the structure design calculation. By respectively extracting knowledge from the computer books and the IFC data files, the extracted information can instantiate the structural design field ontology, and retrieval and knowledge reasoning can be carried out on the basis.

1) Computer book information extraction

The procedure of extracting the information of the calculation book is shown in fig. 7, and includes: unifying Chinese and English characters, in particular to quotation marks and colon marks of Chinese and English; extracting half-structure data with the value of'; positioning word segmentation to extract information of the calculation book; obtaining structured data; splitting the table according to the categories; and finally writing into the relational database MySQL.

The computer book information extraction is that the computer book text relative rule is generally automatically generated by computing software, and the required 17-class structure including the general design information, the computing control information and the second-order effect information can be obtained through the basic text regular processing to calculate the related content. Thereafter, five identification attributes of "project name", "project code number", "designer", "proof reader", and "calculation date" are added to each type of content so as to perform association and query according to the project information. The extraction process in combination with the actual content is shown in fig. 8.

2) Knowledge extraction based on IFC data

The process of extracting knowledge based on IFC data, including extracting building, floor, room and member attribute information from IFC data files, is shown in fig. 9, and includes: all IFC entities are traversed firstly, and based on the structural characteristics of the IFC file, a part of attributes can be directly subjected to attribute extraction, and besides, attribute pairs need to be further extracted from the attribute groups of IfcRelDefinesByType, ifcRelDefinesByProperties and IfcElementQuantity.

Firstly, an IFC data file is analyzed by adopting an IfcOpenShell open source tool, and building structure related information is extracted from the IFC data file. And (4) searching and classifying the instance elements in the IFC file through types, sequentially analyzing and acquiring corresponding attribute pairs according to the attribute extraction link, and finally inputting the obtained instance information into a relational database such as MySQL and the like.

3) Structural calculation and building structure sub-ontology instantiation and alignment integration into complete structural design ontology

After the information data of the structured calculation book (namely the design basic information of the building structure) and the geometric information of the building structure in the IFC data file are recorded into the database, the table data in the relational database is used for instantiating the structure calculation sub-ontology through mapping by using a D2RQ open source tool. The D2RQ conversion of the database data to the instantiation ontology is mainly divided into two parts, one part is to read the rules in the compiled mapping rule file, and the other part is to read the corresponding relational database data, and then output the data to the target ontology instantiation file. A Mapping rule of conversion of a Resource Description Framework (RDF) of a database to an ontology is defined through D2RQ Mapping, and a Table (Table), a Field (Field), a Value (Value) in the database are converted into a Class (Class), an attribute (Property) and an Instance (Instance) of the RDF. The final formed computer book data and the result of ontology instantiation of the building structure information in the IFC file are shown in fig. 10 and 11, respectively.

And according to the obtained instantiation results of the two sub-ontologies, class, attribute relation and instance relation are further aligned with the complete structural design ontology, so that the original instantiation data is integrated on the structural design ontology. The corresponding mapping rule comprises the following steps: class mapping, attribute mapping, and inferential mapping. The class mapping mainly maps and instantiates class instances of source sub-ontologies which have the same meaning but adopt different references into corresponding classes of a complete structural design ontology; the attribute mapping is used for reducing the length of an attribute query path and directly relating attribute representation information needing to be searched by a multilayer path in an original sub-ontology to an attribute relation of a new ontology by using an explicit method; the inference type mapping is relatively complex comprehensive mapping, and aims to dig out relationships which are not shown in an original graph through conditional relationships on the basis of an original source sub-ontology.

An example of the procedure for the corresponding mapping rule is as follows:

@prefix De1:<http://www.semanticweb.org/b.y/ontologies/2020/2/untitled-ontology-7#>.

@prefix De2:<http://www.semanticweb.org/b.y/ontologies/2020/3/untitled-ontology-14#14>

@prefix Re:<http://www.semanticweb.org/b.y/ontologies/2020/2/untitled-ontology-6#>.

@prefix owl:<http://www.w3.org/2002/07/owl#>.

@prefix rdf:<http://www.w3.org/1999/02/22-rdf-syntax-ns#>.

@prefix rdfs:<http://www.w3.org/2000/01/rdf-schema#>.

@prefix xml:<http://www.w3.org/XML/1998/namespace>.

@prefix xsd:<http://www.w3.org/2001/XMLSchema#>.

# # # class mapping

# IFC data Source

[ rule _ project: (

[ rule _ site: (

[ rule _ building: ((is

[ rule _ store: (

[ rule _ space: (

[ rule _ wall: (

[ rule _ slab: (

[ rule _ column: (

[ rule _ beam: (

# calculation book data Source

[ rule _01: ((is

[ rule _02: ((is

[ rule _03: ((is x rdf: type De2:03 second-order effect information) > (

[ rule _04: (

[ rule _05: (

[ rule _06: (

[ rule _07: (

[ rule _08: ((is

[ rule _09: (

[ rule _10: (is x rdf: type De2:10 fire prevention validation)) > (

[ rule _11: (

[ rule _12: (

[ rule _13: (

[ rule _14: ((is x rdf: type De2:14 material information) > (

[ rule _15: ((is

[ rule _16: ((is

[ rule _17: (

# # # # attribute mapping

Module II, design recommendation scheme generation module

The main content of the module is to obtain joint probability distribution by using design information extracted from the structural design body as a sample, and further construct a Bayesian network. Through the combing of the design logic, the design information extracted from the structural design field body is used as an event node of the Bayesian network, and the learning of the structural design case is completed according to the probability model, so that the design scheme is recommended.

The Bayesian model is an algorithm model for simulating human inference process based on probability statistics. The model represents the correlation relationship by a directed acyclic graph, the causal relationship between two events is represented by the connection of single arrows between event nodes, the starting point of the arrow is the reasons (entries), the pointing point of the arrow is the effect (children), and the two events are connected by the connecting arrow to be the conditional probability event. The variables of the Bayesian model are usually discrete, and the continuous variables can be converted into discrete data for model analysis. Because the data contained in the body of the structural design field is more, the information types are more abundant, so the body can not be directly used, a proper amount of data types need to be selected, and the data types can be converted into sample data of the Bayesian model through effective classification.

The Bayesian network is a network diagram simulating a human propulsion scheme, and the construction of the structural design Bayesian network is based on the premise that the human structural design idea is analyzed, the variable conditions of the design process are defined in a classified manner, and the condition variables of the structural design mainly come from three aspects. The first is objective conditions, including design year, material market price and other conditions, and buildings in different periods have design characteristics and process technologies in different periods; secondly, the requirements of the owner, including design parameters which play a decisive role in building planning and design such as building purposes, building floors and the like; and thirdly, site conditions, geological conditions of the building hidden in the site, earthquake conditions, wind load environments and the like have important influence on load distribution of the structural design.

Based on analysis of structural design logic and information extracted from a BIM model, 11 items of important parameter index information in structural design, such as year, building application, material, building length, building width, structural type selection, floor, seismic fortification intensity, space application, span and section width, are selected as examples, and classification standards are defined. In 11 design parameter indexes, continuous data such as year, building length, building width and span are converted into discrete data according to distance. For the discretized data, the data is classified according to relevant specifications and guidelines (see the column of classification rules in table 1, which are consistent with industry standards), and the classification definition is shown in table 1:

TABLE 1 structural design data Classification List

1) Structure model selection design based on naive Bayes

The naive Bayes model is one of Bayes models, simplification is carried out on the basis of Bayes algorithm, and the simplification precondition is that each conditional feature is independent and has the same influence on the explained vector, and the conditional probability problem of a plurality of father nodes to a child node is researched. The method has the advantages of simplicity and capability of better processing the classification problem, but has the defect of emphasizing that the characteristics are independent, and if the conditional characteristics which are not independent from each other exist, the method is not suitable for a naive Bayes model.

The general normal flow of the structure-selection design is shown in fig. 12. Aiming at the structure model selection design, the key, independence and statistics of design variables are considered, five variables of year, material, building application, floor and structure model selection are selected, wherein the year, the material, the building application and the floor are condition variables, the structure model selection is used as an explanatory variable, and a naive Bayes model is established as shown in figure 13. One of the conditions for establishing the naive Bayes model is that the condition variables need to satisfy the independence assumption. Verification can be made from two aspects for this assumption. Firstly, theoretical logic analysis is used as a verification basis, secondly, data analysis is used as a verification basis, the Cramer' V coefficient of chi-square independence test is used for completing independence test on condition variables in a sample, and the test formula is as follows:

wherein:

event f _ij Actual value of

Event f _ij Theoretical value of (2)

R-number of rows in the list

C-number of columns of the Linked List

The closer to 1 the independence test coefficient V is, the less independence the two condition variables are, and the closer to 0 the V is, the more independence the two condition variables are.

Therefore, combined with historical case data, joint probability distribution of the structure design selection is established, then conditional probability prediction can be carried out on new design conditions, the conditional probability result of each structure selection is judged, and the scheme with the largest result is selected for subsequent design.

The structure model selection design is a precondition of preliminary structure design, the model selection scheme of the structure is basically determined through the structure model selection, and information such as the section of the component is preliminarily determined through the preliminary structure design on the basis of the structure model selection design.

2) Preliminary structure design based on Bayesian network

The general naive Bayes model requires mutual independence between conditional events, but complete logical independence is difficult to achieve between design elements in the structural design, and moreover, the naive Bayes model intelligently describes the influence of a plurality of events on one event, and the structural design usually has mutual influence between design elements. Therefore, a Bayesian network is introduced to predict design parameters in the preliminary design of the structure.

A bayesian network is a graph structure used to represent probabilistic relationships between variables and make probabilistic inferences about these variables. The network is a probability relation model established on the basis of conditional probability, and if a probability space (omega, S, P) is set, A belongs to S, and P (A) > 0, then for any event B belongs to S, the following are provided:

i.e. the probability of event B being at the conditional probability of occurrence of event a.

Constructing bayesian networks typically uses two or a combination of the following methods: 1) Automatically constructing a Bayesian network according to the existing data, namely completing Bayesian network structure learning through the existing sample data; 2) A Bayesian network constructed based on domain experts is used to express complex knowledge domain logic. In the preferred embodiment, a Bayesian network is established by combining two methods, and a logical relationship of the Bayesian network is combed based on a domain expert (domain expert system); then generating a sub-network based on a field expert Bayesian network according to the building characteristics of the design target building; and finally, according to the existing data correlation analysis, carrying out network relationship detection and supplementation on a Bayesian sub-network established based on domain experts, deleting the conditional relationship with weak data set correlation, and supplementing the conditional relationship with strong correlation. The specific creation steps are as follows:

(I) Bayesian network construction based on domain experts

The method comprises the steps of establishing a Bayesian network based on a domain expert system, wherein the domain expert system is generally formed by adopting modes of expert interview, interview investigation, document reading and the like, combing a professional domain logic framework, analyzing event influence relations, and establishing the Bayesian network. The domain expert system logical combing of the preferred embodiment comes from three aspects: first, structural design related specifications and standards. In China, the structural design must meet the requirements specified by the specification, so the specification has extremely high authority and universality in the structural design; secondly, the mechanics law is summarized. The main purpose of structural design analysis is to calculate the stress condition of the structure, so that the mechanics rule needs to be summarized; thirdly, summarizing the objective influence among the design elements. The architectural structure design needs to meet professional requirements and also needs to meet objective use requirements, and objective influences among events have important influences on the structural design.

Based on three parts of design specifications and standards, mechanical rules and objective influences, 30 design elements such as service life, materials, building model selection, space usage and the like are selected, and a Bayesian network is established according to the logical relationship between condition events and inspection events among the elements.

And (II) creating a structural design Bayesian subnetwork according to different design requirements of the building and the logic of the domain expert Bayesian network.

Then on the basis of the analyzed building design characteristics, merging, omitting and refining each design element, determining specific event nodes and event states, and establishing a sub-network according to the obtained event logic. For events with sequential logical form, if the middle event is omitted or merged, the direct action relationship of the two end events is established, i.e. the logical form between the times is X → Z → Y, and if the event Z is omitted or merged into the event X or the event Y, the logical relationship X → Y is established.

In a specific application example, taking the design of a teaching building as an example, 13 parameters of building purpose, material, building length, building width, structure type selection, floor, seismic fortification intensity, classroom span, corridor span, room width, beam section height and column section width are selected to serve as event nodes of the bayesian network. The sequentially established sub-networks of the teaching building structural design are shown in fig. 14.

(III) pruning sub-network architecture

The complex network is pruned, so that the operation efficiency can be effectively improved, the original Bayesian network can be pruned for the first time or the second time or new correlation relations can be filled according to the statistical data characteristics and the correlation analysis of the sample data, the correlation of the sample data can be utilized on the basis of simplifying the model and preventing overfitting, event nodes and time relations can be increased or deleted, and the correctness of model logic can be ensured.

In the statistical data, if the statistical number of certain time states is less, the research value of the event state in the whole Bayesian network is lower, the prediction reference of probability distribution calculation is also lower, and the statistical data capacity can be increased or the Bayesian network event state can be limited by adjusting the statistical data structure. The bayesian network after one pruning is shown in fig. 15. On the basis, the relevance analysis is further carried out on the sample data, whether the situation that event relations are omitted along with the deletion of event nodes exists in the pruning network or not is judged, some existing event relations with low relevance are removed, the network structure is simplified, and the secondary pruning result is shown in fig. 16.

(IV) obtaining event probability distribution by using Bayesian network structure

And generating Bayesian network joint probability distribution and edge probability distribution based on the structure of the concrete teaching building structure design network subjected to secondary trimming and sample data. A Bayesian network and a sample data office are designed by utilizing the structure, the event distribution probability can be directly obtained, and a complete Bayesian network with event nodes, event relations and probability distribution is established. Its advantage lies in convenient and fast, and the classification result is traceablely circulated, and the logic of causal derivation can be shown through the probability chain. When the sample capacity is large enough, the inference result can reflect the joint probability distribution, but when the sample capacity is limited or the sample characteristics have missing values, the inference result is limited by the sample. For the case of sample feature loss, the prediction estimation of the bayesian network parameters by the EM algorithm can be attempted. Through EM parameter learning, the conditional probability distribution has the characteristic of local optimum compared with the original conditional probability distribution.

According to the formation of the Bayesian network, for the structural design of a new project, corresponding design parameters can be input by using the network, the test data is used as a conditional event, the probability distribution of the design parameters focused in the design scheme is obtained, and the probability distribution is integrated to form a complete structural design scheme.

Module III, building information model generation module

The module mainly takes design parameters obtained based on probability analysis of a Bayesian network as a design scheme for primary selection, secondarily develops Revit through a RevitAPI (Revit Application Programming Interface), and performs parametric modeling on the basis of the input design parameters. Revit reserves a development port for a third-party developer, a development community is developed, and a stable data conversion interface is arranged between Revit and structural design software, so that the generation of a structural design BIM model can be performed on the basis of selecting Revit in the preferred embodiment.

For a building with the characteristic of regular planar arrangement, the structure design of the building can use for reference and can learn. Buildings such as office buildings and teaching buildings are usually designed with standard roof trusses or standard rooms, and the standard roof trusses or standard rooms are arrayed and expanded. Therefore, for building structures with standard trusses or standard rooms, the structural arrangement can be used as a standard model, and the model of a new project is rapidly generated by using the standard model.

In a specific application example, taking a teaching building as an example, the teaching building mainly includes two space functions of classroom and corridor in terms of function use, and is matched with a staircase, a toilet and the like, and has two forms of corridor + classroom and classroom in a planar arrangement, and three forms of multi-frame, double-frame and single-span cantilever type are arranged on a column network arrangement, as shown in fig. 17.

According to the regularity of the structural design of the teaching building, a corresponding intelligent generation function of a standard model is developed, and a specific flow chart is shown in FIG. 18. Firstly, design and carding are finished according to a historical design scheme by referring to the historical design scheme, standard trusses or standard rooms are counted to obtain design structure arrangement, span segments of the standard trusses are determined, shaft gateway key information such as opening and span is marked, and a standard model is constructed. Secondly, the Revit secondary development is utilized, the standard shaft network is used as a template, design information such as bay, span and the like is determined, and the shaft line arrangement is completed for a new project based on the standard shaft network template. Thirdly, based on the modularly generated shaft network, the generation of design components such as beams, columns and walls is completed, and finally a new Revit design model is generated. The intelligent generation of the standard model comprises five steps: standard shafting selection, shafting establishment, component establishment and interface development. The axis net, elevation, cylinder net generation interface and results are shown in figure 19.

And then carrying out batch modification development on the three-dimensional model components. In actual engineering, a general structural design can not complete a final design through one-time scheme design, and a model needs to be modified and debugged. The design efficiency can be further improved by performing corresponding development for repeated and time-consuming modification work, such as member naming modification, batch member section parameter adjustment, and the like, as shown in fig. 20.

Module four, building information model optimization module

When the module acts, in order to verify the feasibility of the recommended design scheme generated by the module, the alternative scheme is input into the structure calculation software for analysis so as to achieve the purpose of verification and optimization, and on the premise that the structure meets the requirements of the design specifications of China, the stress safety of the structure is ensured. At present, more software is available for structural calculation and analysis in China, and the calculation software such as YJK, PKPM, midas and the like developed based on the national design specifications in China can meet the design requirements of structural designers in China. In consideration of the portability of the application and the data circulation with the Revit software, the preferred embodiment uses YJK as the structural calculation software to complete the model mechanical analysis in linkage with the Revit platform. A data intercommunication interface is arranged between Revit and YJK, so that requirements of double-platform model creation, component marking, construction diagram reinforcement, steel bar three-dimensional model creation and the like can be met, data sharing of Revit and FJK is realized, and the process is shown in figure 21.

And extracting submodels for structural calculation from the Revit model, wherein the submodels comprise structural calculation models of shaft nets, beams, columns, rods, walls and opening information. When the sub-model is converted and extracted, the conversion between the family parameters used by the Revit component and the YJK definition parameters needs to be noticed, and the conversion needs to be matched with the associated section shape, geometric parameters and material type.

After structural calculation, analysis and optimization and construction drawing design, model parameters can be fed back to the Revit model in a reverse direction, original design model information is enriched, and then IFC data and design calculation book results are led out and are used as new case samples to be put into a structural design knowledge base.

The following describes, by way of a specific example, an autonomous design flow of a building structure implemented by using the method and system provided by the above embodiments of the present invention.

The specific example takes the structural design of a teaching building as an example, a plan view is shown in fig. 22, the classroom span is 6900mm, the corridor span is 3000mm, and the bay span is 6600mm, the project design is completed through structural design field ontology case query, bayesian network design scheme recommendation, BIM model automatic generation and YJK structural mechanics calculation, IFC and calculation book knowledge extraction are performed on a new project, new project case data storage is completed, new case input is provided for a structural design knowledge base, and a structural intelligent design method flow for the new project is shown in fig. 23.

Step 1: when a new project is faced, a designer firstly clearly designs project information such as functions, floors, locations of buildings and the like, and queries historical project design information in a structural design field body by using the SPARQL language. The engineer can directly use the historical project design information in the inquired structural design field ontology to obtain a design scheme according to the design experience of the engineer.

Step 2: and reflecting the design thinking of the historical design case by using the Bayesian network, and providing reference for structure selection and member section initial selection of a new project. And (4) performing conditional probability calculation on the column-beam uniform section interval of the new project by utilizing the established structural design Bayesian network, and finally recommending a primary selection scheme.

According to the newly designed project, the classroom span is 6900mm, the room span is 6600mm, and let event X = { classroom span (classorom) =6900to7200, room width (width) =6600to7500}, the conditional probability distribution result is shown in fig. 24.

In the case sample, P (z _ Column =500 no smoke x) =99.6%, i.e., in the history case, when the classroom span is 6900mm-7200mm and the room width is 6600mm-7500mm, the probability that the pillar width is 500mm is large, so it is recommended that the pillar section is initially selected as a 500mm square pillar. Similar inferences can be made for beam cross-sections.

And 3, carrying out parameterization rapid generation of a design model in Revit by inputting the obtained structural design basic scheme. As shown in fig. 25, the generated structural model is a main bearing frame of the frame structure, and further, the contents of secondary beam design, floor slab design and the like can be completed and used as basic model input of the next structural calculation.

And 4, step 4: by means of the REVIT-YJKS plug-in unit, a structural model for YJK calculation is generated through component matching and model extraction, structural calculation is completed according to a traditional structural calculation process, a structural calculation book is generated, and a result after calculation can be led into Revit and further exported into an IFC model file.

And 5: after the project structural design of the teaching building is completed, the IFC file and the structural design calculation book generated by the project are used for extracting knowledge of the project design information, and the data source body and the IFC data source body of the calculation book are aligned with the structural design field body, so that the project structural design information is managed, and the SPARQL query language is convenient to use for querying the project design information. Knowledge extraction about the design case can refer to the related content of the chapter v module one.

By finishing the design case aiming at the structural design of a certain teaching building, the design steps of case query of a structural design body, preliminary selection of a member section based on a Bayesian network, establishment of a structural calculation model and body alignment in the structural design field are realized, and the logicality and convenience of completing the intelligent structural design based on the combination of the body, the Bayesian network and the BIM technology are embodied.

A third embodiment of the present invention provides a terminal comprising a memory, a processor and a computer program stored on the memory and executable on the processor, the processor being operable to execute the method according to any one of the above embodiments of the present invention when executing the program.

Optionally, a memory for storing a program; a Memory, which may include a volatile Memory (RAM), such as a Static Random Access Memory (SRAM), a Double Data Rate Synchronous Dynamic Random Access Memory (DDR SDRAM), and the like; the memory may also comprise a non-volatile memory, such as a flash memory. The memories are used to store computer programs (e.g., applications, functional modules, etc. that implement the above-described methods), computer instructions, etc., which may be stored in partition in the memory or memories. And the computer programs, computer instructions, data, etc. described above may be invoked by a processor.

The computer programs, computer instructions, etc. described above may be stored in one or more memories in a partitioned manner. And the computer programs, computer instructions, data, etc. described above may be invoked by a processor.

A processor for executing the computer program stored in the memory to implement the steps of the method according to the above embodiments. Reference may be made in particular to the description relating to the preceding method embodiment.

The processor and the memory may be separate structures or may be an integrated structure integrated together. When the processor and the memory are separate structures, the memory, the processor may be coupled by a bus.

A fourth embodiment of the invention provides a computer-readable storage medium, on which a computer program is stored, which program, when being executed by a processor, is adapted to carry out the method of any of the above-mentioned embodiments of the invention.

The building structure autonomous design method, the building structure autonomous design system and the building structure terminal based on artificial intelligence provided by the embodiment of the invention aim at the problems of design decision, design information lookup, design information management and the like in the actual structure design in the prior art, and the artificial intelligence based on the ontology and the Bayesian network provides technical support for high efficiency and intellectualization of the structure design by means of advanced computers and artificial intelligence technology, reduces the workload of design practitioners in the stages of scheme design and preliminary design, improves the design efficiency, reduces the design cost, effectively utilizes the historical design experience, and has innovation and expansion significance; the method comprises the following steps of taking structural intelligent design as a development target, utilizing a body as a structural design knowledge framework, utilizing a Bayesian network as a structural design knowledge logic, and utilizing a BIM technology as a structural design realization platform, and realizing the knowledge extraction, design scheme recommendation and calculation model modularization establishment of structural design; historical design experience can be effectively utilized, and a new thought is provided for the retention and development of design knowledge; in addition, although the invention aims at the design stage of the building structure, the invention has expansion value for the research of each building life stage of other building forms, and is beneficial to improving the automation level of the building engineering; a Bayesian network method is adopted to realize the primary scheme design, the model is obtained by learning from a plurality of real samples, and the influence of the design parameters on the design result has higher reliability.

It should be noted that, the steps in the method provided by the present invention may be implemented by using corresponding modules, devices, units, and the like in the system, and those skilled in the art may implement the step flow of the method with reference to the technical solution of the system, that is, the embodiment in the system may be understood as a preferred example of the implementation method, and details are not described herein.

Those skilled in the art will appreciate that, in addition to implementing the system and its various devices provided by the present invention in purely computer readable program code means, the method steps can be fully programmed to implement the same functions by implementing the system and its various devices in the form of logic gates, switches, application specific integrated circuits, programmable logic controllers, embedded microcontrollers and the like. Therefore, the system and various devices thereof provided by the present invention can be regarded as a hardware component, and the devices included therein for realizing various functions can also be regarded as structures in the hardware component; means for performing the various functions may also be conceived of as structures within both software modules and hardware components of the illustrated method.

The foregoing description of specific embodiments of the present invention has been presented. It is to be understood that the present invention is not limited to the specific embodiments described above, and that various changes and modifications may be made by one skilled in the art within the scope of the appended claims without departing from the spirit of the invention.

Claims

1. A method of autonomous design of a building structure, comprising:

establishing a Bayesian network model, taking the structural design knowledge contained in the structural design domain ontology as the input of a Bayesian network, calculating conditional probability distribution, and outputting a design recommendation scheme;

verifying the structural design scheme of the building information model, feeding back a verification result to the building information model, and optimizing the building information model;

the establishing of the Bayesian network model takes the structural design knowledge contained in the structural design field ontology as the input of the Bayesian network, calculates the conditional probability distribution, and outputs the design recommendation scheme, and the establishing of the Bayesian network model comprises the following steps:

combing the logical relationship of the Bayesian network based on a domain expert system to generate a domain expert Bayesian network;

generating the domain expert Bayesian network sub-network based on the domain expert Bayesian network according to the design requirement of the building;

taking structural design knowledge contained in a structural design domain ontology as event nodes of the domain expert Bayesian network sub-network;

classifying structural design knowledge contained in a structural design domain ontology, and taking the classified structural design knowledge as an input sample of the domain expert Bayesian network sub-network;

deleting and/or filling the time nodes and the time relation of the domain expert Bayesian network sub-network by utilizing the correlation of the input samples to construct a Bayesian network model; inputting the input sample into the Bayesian network model to obtain a conditional probability distribution;

and outputting a design recommendation scheme according to the conditional probability distribution.

2. The method of claim 1, wherein the extracting historical design knowledge and establishing structural design domain ontology comprises:

extracting geometric information of the building structure and design basic information of the building structure from IFC data and a design calculation book of historical design respectively to form structural design knowledge;

respectively establishing a building structure sub-body and a structure calculation sub-body;

respectively instantiating the established building structure sub-ontology and the structure calculation sub-ontology by using the extracted structural design knowledge to obtain instantiation files of the building structure sub-ontology and the structure calculation sub-ontology;

and aligning and integrating the instantiation files of the building structure sub-ontology and the structure calculation sub-ontology to obtain a complete instantiation file of the structure design field ontology, and obtaining the structure design field ontology.

3. The method of autonomous design of a building structure according to claim 2, further comprising any one or more of the following features:

-said extracting basic design information of the building structure, comprising:

generating a text relative rule of the calculation book, and acquiring structure calculation related content through the text relative rule generated by text regular processing;

calculating related content for each type of structure, and adding identification attributes to obtain design basic information of the building structure;

-said extracting geometrical information of the building structure comprises:

analyzing the IFC data, and extracting the related attribute information of the building structure from the IFC data;

retrieving and classifying example elements in the IFC data, sequentially analyzing and acquiring corresponding attribute pairs according to the attribute information extraction link, and finally obtaining example information, namely geometric information of the building structure;

-said separately establishing a building structure sub-ontology and a structure calculation sub-ontology, comprising:

respectively constructing a building structure sub-body and a structure calculation sub-body on the basis of a protege platform, wherein the building structure sub-body and the structure calculation sub-body respectively correspond to a model establishing stage and a calculation analysis stage in structure design calculation;

-said instantiating of the built building structure sub-ontology, comprising: establishing a mapping rule file according to the established building structure sub-body and building structure information in the IFC data file; acquiring a mapping rule file and building structure information table data, and generating a building structure sub-ontology instantiation file;

-said instantiating of the built structure calculation sub-ontology, comprising: establishing a mapping rule file according to the established structure calculation sub-body and the structure calculation information in the structure calculation book file; and reading the mapping rule file and the structure calculation information table data to generate a structure calculation sub-ontology instantiation file.

4. The method of claim 1, wherein the generating a building information model corresponding to a project according to the design recommendation project comprises:

determining span segments of standard trusses according to the design scheme output based on the Bayesian model, marking key information of the shaft gateway, and constructing a standard model;

determining key information of the axis gateway by taking the standard model as a template;

completing axis arrangement for the new project based on the standard model to generate a modular shaft network;

and finishing the generation of the design component based on the modularized axle network, and finally generating a preliminarily designed building information model.

5. The method of autonomous design of a building structure according to claim 1, characterized in that it further comprises:

and (4) aiming at the optimized building information model, deriving the structure design knowledge contained in the optimized building information model to be used as new IFC data and design calculation book data samples.

6. An autonomous design system for a building structure, comprising:

the design recommendation generation module is used for establishing a Bayesian network model, taking structural design knowledge contained in a structural design domain ontology as input of a Bayesian network, calculating conditional probability distribution and outputting a design recommendation; wherein:

deleting and/or filling the time nodes and the time relation of the field expert Bayesian network sub-network by utilizing the correlation of the input samples, and constructing to obtain a Bayesian network model; inputting the input sample into the Bayesian network model to obtain a conditional probability distribution;

outputting a design recommendation scheme according to the conditional probability distribution;

the building information model generating module is used for generating a building information model corresponding to the scheme according to the output design recommendation scheme;

7. A terminal comprising a memory, a processor and a computer program stored on the memory and executable on the processor, characterized in that the processor, when executing the program, is adapted to perform the method of any of claims 1-5.

8. A computer-readable storage medium, on which a computer program is stored which, when being executed by a processor, is adapted to carry out the method of any one of claims 1 to 5.