Tunnel BIM intelligent collaborative modeling method and system

Abstract

The invention provides a tunnel BIM intelligent collaborative modeling method and system, and the method comprises the steps: receiving user natural language input information and tunnel-related engineering data files through an interactive interface, performing integration processing on information in the engineering data file and natural language input information of the user through the preposed information and the structured information template to obtain integrated information; sending the integrated information to a large language model for analysis to obtain structured data; obtaining a modeling function call sequence based on modeling scene information in the structured data, and generating structured call information based on the modeling function call sequence and the structured data; and on the basis of the calling information, executing the modeling instruction to obtain target modeling, so that full-life-cycle tunnel collaborative modeling can be realized, and intelligent IFC delivery decision is realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of tunnel modeling, in particular to a tunnel BIM intelligent collaborative modeling method and system, and in particular to BIM intelligent collaborative modeling and delivery and operation management of a tunnel project in the whole life cycle of design, construction, delivery and operation. BACKGROUND

[0002] In related technologies, for the modeling of building infrastructure, some modeling methods that integrate AI technology are adopted. However, there are still deficiencies in the aspects of tunnel BIM deepening intelligent collaborative modeling and delivery in the field of underground engineering, such as 1) field limitations of AI-based modeling; 2) lack of interactive intelligence; 3) lack of data verification; 4) weak delivery support; 5) rigid function calling; 6) separation of engineering data and model components, leading to difficulty in quality tracing; and lack of data embedding capability in traditional IFC delivery. SUMMARY

[0003] The present application provides a tunnel BIM intelligent collaborative modeling method and system to at least partially solve one of the technical problems in related technologies. The technical solution of the present disclosure is as follows: In a first aspect, the present application provides a tunnel BIM intelligent collaborative modeling method, applied to a tunnel BIM intelligent collaborative modeling system, comprising the following steps: Receiving user natural language input information and tunnel-related engineering data files through an interactive interface, and integrating information in the engineering data files and the user natural language input information through pre-prepared front-end information and structured information templates to obtain integrated information, wherein the front-end information includes modeling-related information, the structured information template is used to instruct a large language model to generate structured data, and the user natural language input information includes modeling-related information; Sending the integrated information to the large language model for analysis to obtain structured data; Based on the modeling scenario information in the structured data, a modeling function calling sequence is obtained, wherein the modeling function calling sequence includes a sequence of interface functions to be executed, the sequence of interface functions stores the required calling software and the function sequence corresponding to the calling software that needs to be executed in order, and the modeling scenario information includes at least one modeling behavior information; Based on the modeling function calling sequence and the structured data, structured calling information is generated; Based on the calling information, a modeling instruction is executed to obtain a target modeling. In a second aspect, the present application provides a tunnel BIM intelligent collaborative modeling device, configured in a tunnel BIM intelligent collaborative modeling system, comprising: The data integration module is configured to receive user natural language input information and tunnel-related engineering data files through an interactive interface, integrate and process information in the engineering data files and the user natural language input information through pre-prepared front information and a structured information template, and obtain integrated information, wherein the front information includes modeling-related information, the structured information template is used to instruct a large language model to generate structured data, and the user natural language input information includes modeling-related information. The data generation module is configured to send the integrated information to the large language model for analysis, and obtain structured data. The function acquisition module is configured to obtain a modeling function call sequence based on modeling scenario information in the structured data, wherein the modeling function call sequence includes a sequence of interface functions to be executed, the sequence of interface functions stores required call software and a function sequence to be executed corresponding to the call software in sequence, and the modeling scenario information includes at least one modeling behavior information. The call acquisition module is configured to generate structured call information based on the modeling function call sequence and the structured data. The modeling implementation module is configured to execute modeling instructions based on the call information, and obtain target modeling. In a third aspect, an electronic device is provided, including a processor and a memory connected with the processor in communication; the memory stores computer execution instructions; the processor executes the computer execution instructions stored in the memory to implement the method of the first aspect.

[0004] In a fourth aspect, a computer readable storage medium is provided, and the computer readable storage medium stores computer execution instructions; the computer execution instructions are executed by a processor to implement the method of the first aspect.

[0005] In a fifth aspect, a computer program product is provided, including a computer program; the computer program is executed by a processor to implement the method of the first aspect.

[0006] The tunnel BIM intelligent collaborative modeling method and system provided by the application can support collaborative modeling of multiple components and realize whole life cycle tunnel collaborative modeling through a structured information template and a dynamic modeling function call sequence; a natural language driven dynamic interaction mechanism is realized based on a large language model, error correction and data supplement efficiency are improved; structured data is automatically checked to ensure input / output consistency; an AI dynamically generated export decision scheme is realized to achieve intelligent IFC delivery decision and open up a data transmission channel between a construction and operation and maintenance management platform; a modeling function sequence generation technology driven by prompt template information is proposed to adapt to flexible requirements of BIM deep modeling; relevant engineering data during the construction period is automatically analyzed by AI, engineering data and models are intelligently bound, and construction quality traceability efficiency is improved.

[0007] Additional aspects and advantages of the application will be set forth in part in the description that follows, and in part will become apparent to those skilled in the art upon examination of the following description and drawings or can be learned by practice of the application. BRIEF DESCRIPTION OF DRAWINGS

[0008] The above and / or additional aspects and advantages of the application will become apparent and be readily appreciated from the following description, including the accompanying drawings, in which: Figure 1 A flowchart of a tunnel BIM intelligent collaborative modeling method provided by an embodiment of the application; Figure 2 An example diagram of a data structure of prompt template information provided by an embodiment of the application; Figure 3 An example diagram of prompt template information corresponding to a tunnel main structure and a reinforcement model provided by an embodiment of the application; Figure 4 An example diagram of a structure of an API dependency graph provided by an embodiment of the application; Figure 5 An example diagram of a data structure of call information provided by an embodiment of the application; Figure 6 A flowchart of a tunnel BIM intelligent collaborative modeling system provided by an embodiment of the application; Figure 7 A block diagram of a tunnel BIM intelligent collaborative modeling device provided by an embodiment of the application; Figure 8 A block diagram of an electronic device provided by an embodiment of the application. DETAILED DESCRIPTION

[0009] Embodiments of the present application are described below in detail, examples of which are shown in the accompanying drawings, wherein the same or similar notations represent the same or similar elements or elements having the same or similar functions throughout. The embodiments described below by reference to the accompanying drawings are exemplary and are intended to explain the present application, and cannot be understood as a limitation of the present application.

[0010] Terminology: Platform: refers to system architecture.

[0011] The application scenario of the tunnel BIM intelligent collaborative modeling method and system of the embodiments of the present application is, under the premise of having design results (design drawings, survey reports), construction period engineering data (construction log, detection records, etc.) and operation and maintenance data (asset information, operation and maintenance records, etc.), to perform automatic modeling based on AI and data exchange standard IFC (Industry Foundation Classes) data export, complete the deepening design and delivery of the BIM model.

[0012] The tunnel BIM intelligent collaborative modeling method, device and equipment of the embodiments of the present application are described below with reference to the accompanying drawings.

[0013] Figure 1 A flowchart of a tunnel BIM intelligent collaborative modeling method provided by the embodiments of the present application.

[0014] It should be noted that the execution subject of the tunnel BIM intelligent collaborative modeling method of the embodiments of the present application is the tunnel BIM intelligent collaborative modeling device of the embodiments of the present application, which can be configured in an electronic device so that the electronic device can perform tunnel BIM intelligent collaborative modeling functions.

[0015] As shown in Figure 1 The tunnel BIM intelligent collaborative modeling method includes the following steps: Step S101, receiving user natural language input information and tunnel-related engineering data files through an interactive interface, and integrating the information in the engineering data files and the user natural language input information through pre-prepared front-end information and structured information templates to obtain integrated information, wherein the front-end information includes modeling-related information, the structured information template is used to instruct a large language model to generate structured data, and the user natural language input information includes modeling-related information.

[0016] The tunnel BIM intelligent collaborative modeling method of the embodiments of the present application is applied to a tunnel BIM intelligent collaborative modeling system.

[0017] In the embodiment, natural language interaction is performed with the system through the interaction interface, the system processes the received information, sends the information to a large language model for analysis, analyzes the returned information, and displays dialogue information or performs the next step according to the analysis result.

[0018] In some embodiments, the modeling-related information can include a modeling type, operation file information, a name of a route, a mileage range of a model, a cross section used, a cutting interval, and the like. The operation file information includes path information of an operation file, that is, a file to be processed or generated for a three-dimensional model. The operation file and the calling software are both present on the current local computer, and the path of the operation file to be processed and the information of the calling software required are automatically obtained through large model interaction.

[0019] For example, the modeling-related information includes: generating a section of a route according to default parameters, and the following information: route name: left line; annotated short line interval: 10; horizontal curve file path: G:\Program Files\Tunnel Construction Deepening Software (18BIM-Tunnel) Professional Edition\Excel\Horizontal Curve Element Table.xlsx; horizontal curve file sheet name: NewSheet; vertical curve file path: G:\Program Files\Tunnel Construction Deepening Software (18BIM-Tunnel) Professional Edition\Excel\Vertical Curve Element Table.xlsx; vertical curve file sheet name: NewSheet; do not use the chain break file, and do not consider the chain break file sheet, both are empty. Through this natural language input, the AI can automatically find the software and file, and automatically open the file to generate a route model. According to the file "G:\1-1.dgn", a section of a channel model is generated according to the following channel modeling information: "tunnel name: test tunnel", "place information file path: G:\Channel Statistics Table Template.xlsx". Through this sentence, the channel model can be placed in the existing model file. According to the file "G:\1-1.dgn", a section of a steel bar model is generated according to the following steel bar modeling information: "tunnel name: test tunnel", "start mileage: K0+030", "end mileage: K0+080", "ring steel bar spacing (m): 0.2", "longitudinal steel bar spacing (m): 0.2", "ring steel bar diameter (m): 0.02", "longitudinal steel bar diameter (m): 0.02", "stirrup diameter (m): 0.01", "protection layer thickness (m): 0.02", and a lining steel bar can be generated.

[0020] For example, the engineering data file can be a file generated during the entire construction process, such as a budget estimate, a material list, a material list, and the like.

[0021] In some embodiments, the preface information can include, but is not limited to, modeling software information (name, category, version number, function list, etc.), supported modeling capability range, product summary information, other modeling-related information. The preface information is used to process the information sent to the large language model for analysis. In some embodiments, the method for constructing the structured information template comprises: obtaining all modeling behaviors involved in the automated modeling; determining the modeling information involved in each modeling behavior and the field type corresponding to each modeling information; and integrating the modeling behaviors and their corresponding modeling information based on the field type corresponding to each modeling information to obtain a structured information template. For example, all modeling behaviors involved in automated modeling are determined, such as tunnel body modeling, tunnel advance support modeling, tunnel body reinforcement modeling, three-dimensional geological modeling, drilling modeling, underground pipeline modeling, etc. The modeling information involved in each modeling behavior is determined, such as tunnel body modeling involving "route name, starting mileage, ending mileage, cross section". The field type of the field corresponding to each modeling information contained in each modeling behavior is determined, such as route name being a string type and starting mileage being a string type. The above information is integrated using the Json format to obtain a structured information template, such as {tunnel main structure modeling: {route name: A1, starting mileage: 0+000, ending mileage: 0+900, cross section: V-cr}, tunnel reinforcement modeling: {tunnel name: test tunnel, starting mileage: 0+000, ending mileage: 0+900, database path: ***.db}, project data: {data type: construction log, associated component ID: 668899, file content: ******}}. That is, the template defines the corresponding format for each modeling scenario (such as: {tunnel main structure modeling: {route name: A1, starting mileage: 0+000, ending mileage: 0+900, cross section: V-cr}, tunnel reinforcement modeling: {tunnel name: test tunnel, starting mileage: 0+000, ending mileage: 0+900, database path: ***.db}, project data: {data type: construction log, associated component ID: 668899, file content: ******}}). The template is used to process the information sent to the large language model for analysis, and structured data can be generated in a suitable dialogue process to prepare recognizable and accurate language for the next modeling step.

[0022] Natural language input and engineering data file upload can be performed in the interactive interface. The system reads the information in the uploaded engineering data file and integrates the input information in combination with the pre-prepared preface information and structured information template, so as to send the integrated information to the large language model for analysis in the subsequent step.

[0023] In step S102, the integrated information is sent to the large language model for analysis to obtain structured data.

[0024] In the embodiment, the integrated information is sent to the large language model for analysis, and structured data returned by the large language model is obtained.

[0025] In some embodiments, after obtaining the structured data, the following is performed: based on the structured information template, it is verified whether there is an error or a missing field or data in the structured data; if there is, an error prompt information is generated and displayed on the interactive interface to remind the user to re-input the user natural language input information and the engineering data file through the interactive interface according to the error prompt information to generate the structured data.

[0026] That is, the returned information after analysis by the large language model is parsed to determine whether it matches the pre-prepared structured information template. If it does not match, the returned information is displayed to a dialog box; if it matches, subsequent processes are performed.

[0027] That is, the returned structured data is verified, and if there is an error or a missing field or data, an error prompt is generated and displayed on the interactive interface; dialog information is input in response to the prompt, steps S101-S102 are executed again to obtain new returned structured data, and verification is performed again until there is no error or missing field or data, and the next step is performed.

[0028] In one example, the method for verifying the returned structured data includes: comparing the data structure, field name, and field value type (such as integer, floating point, etc.) against the structured information template corresponding to the current modeling scenario, recording the error information and error type obtained by the comparison; organizing the recorded error information and error type into list information; based on the list information, generating error prompt information and displaying it to the interactive interface.

[0029] The application realizes a natural language driven dynamic interaction mechanism through an interactive interface, improving the efficiency of error correction and data supplementation; that is, by introducing a natural language real-time correction mechanism, when the structured data verification finds a missing field or an error (such as a missing “lining type” parameter), the system automatically generates a precise error prompt (such as “please supplement the lining type code for the 0+000-0+900 section”), and after the user supplements the parameter through a dialog, the system regenerates the structured data. This mechanism will improve the parameter correction efficiency and significantly reduce the modeling failure rate caused by input errors.

[0030] The present application guarantees the input / output consistency through automatic checking of structured data. Specifically, the data returned by the AI is automatically checked in step S102 through a pre-prepared JSON structured template (such as a tunnel modeling template that needs to include mandatory fields such as {route name, starting mileage, cross section}). The system compares the field name and data type (such as "starting mileage" must be in string format "0+000"), and records illegal data and generates error report information in real time. This mechanism ensures the accuracy and attribute integrity of the tunnel BIM model information, and improves the first-pass rate of model automatic generation.

[0031] In step S103, a modeling function call sequence is obtained based on the modeling scene information in the structured data, wherein the modeling function call sequence includes a sequence of interface functions to be executed, and the sequence of interface functions stores the required calling software and the function sequence to be executed corresponding to the calling software in sequence, and the modeling scene information includes at least one modeling behavior information.

[0032] In some embodiments, the method of obtaining the modeling function call sequence includes: searching for prompt template information matching the modeling scene information from a database based on the modeling scene information in the structured data; obtaining the modeling function call sequence based on the prompt template information; wherein the prompt template information is used to describe the modeling function call sequence under different modeling scenes. A plurality of prompt template information is stored in the database, and the plurality of prompt template information corresponds to different modeling scenes. That is, in the database, according to various modeling scenes, some prompt template information is pre-prepared, and each prompt template information describes a sequence of interface functions to be executed under a modeling scene, which is an array that stores the required calling software and the function sequence to be executed corresponding thereto in sequence. For example, the data structure of the prompt template information is shown in Figure 2 For example, to generate a tunnel main structure and a steel model, the prompt template information corresponding to the tunnel main structure and the steel model is shown in Figure 3

[0033] In one example, the modeling scene information is extracted from the structured data, such as: generating a tunnel main structure model + generating a tunnel steel model information; according to the obtained modeling scene information, searching for matching prompt template information from the database; and generating a modeling function call sequence according to the prompt template information, which includes: calling software, and a calling function interface sequence corresponding to the calling software.

[0034] ​In some embodiments, the method for obtaining the modeling function call sequence comprises: if prompt template information matching the modeling scene information is searched from the database, constructing an API dependency graph based on BIM modeling software related information; generating the modeling function call sequence based on the modeling scene information and the API dependency graph, combining an algorithm model based on knowledge graph reasoning rules and reinforcement learning dynamic decision-making; performing conflict checking on the modeling function call sequence; if there is a conflict, adjusting the modeling function call sequence to obtain an adjusted modeling function call sequence.

[0035] In one example, using a large language model, the structured data is semantically parsed to extract modeling scene information, such as: generating a V-class surrounding rock tunnel main structure model + generating a tunnel reinforcement model; based on the API documents, historical call logs, and interface dependency rule library of each BIM modeling software, an API dependency graph is constructed, and the structure of the API dependency graph is as shown in Figure 4 Based on the modeling scene information and the API dependency graph, a rule reasoning based on a knowledge graph + a reinforcement learning dynamic decision-making algorithm model is used to implement shortest path search based on the modeling scene information and the API dependency graph, and a preliminary modeling function call sequence is generated; it is checked whether the generated modeling function call sequence has conflicts (such as: software startup sequence conflict (such as: calling an interface function of a tunnel modeling software without starting the software), parameter passing conflict (such as: tunnel modeling requires section information, but the data has not been generated), resource occupation conflict (such as: parallel tasks competing for CPU)); if there is a conflict, the discovered conflict is processed (such as: inserting an interface function at a certain position in the function call sequence, changing a parallel task to a serial task), and a final modeling function call sequence is obtained.

[0036] In this step, after obtaining the structured data, the matching prompt template information is searched from the database to obtain the modeling function call sequence, and if there is no matching prompt template information in the database, the modeling function call sequence is automatically generated.

[0037] Step S104, based on the modeling function call sequence and the structured data, generating structured call information.

[0038] In some embodiments, the method for generating structured call information comprises: obtaining operation file information based on structured data; obtaining call software and its corresponding function sequence and parameter information from the modeling function call sequence for each operation file in the operation file information, and assigning each parameter in the parameter information from the structured data; generating structured call information based on the operation file information and the call software and its corresponding function sequence and parameter information corresponding to each operation file according to a pre-prepared call information data structure template; prompting whether the call information needs to be modified through an interactive interface; receiving user input adjustment requirement information through the interactive interface, and sending the preface information, call information and adjustment requirement information to a large language model to regenerate the call information.

[0039] It can be understood that the trained large language model can automatically obtain the path of the operation file to be processed and the information of the call software to be processed.

[0040] The pre-prepared call information data structure template is a data structure (such as a Json type) of pre-prepared call information; the call information is generated based on the generated structured data and the modeling function call sequence. The data structure of the call information is pre-prepared, combined with the modeling software information and the function interface list, and the call information is generated based on the structured data and the function call sequence. The call information can include operation files, call software, function interfaces (also called call interfaces) and corresponding parameters (i.e. interface input parameters).

[0041] In one example, the method for generating call information comprises: preparing a general data structure of call information (in Json format) that can meet the modeling call requirements in various situations. As shown in Figure 5 The data object of one call information can include multiple operation files to be operated, each operation file to be operated corresponds to multiple call software to be called, and each call software includes all call interfaces to be called and input parameters; when executing one call information, multiple operation files can be operated simultaneously; when operating one file, the call software to be called and the call interface are executed in sequence; the file information (such as file path, file format, etc.) of the operation file to be operated is obtained from the modeling function call sequence; for each operation file, the call software and its corresponding call interface and input parameter information are obtained from the modeling function call sequence, and the input parameters of each call interface are assigned from the structured data to obtain the call information.

[0042] In some embodiments, it is confirmed in the interactive interface whether the generated calling information needs to be modified. If modification is needed, the adjustment requirements are input in the interactive interface. The system collates the existing "preliminary information, calling information, adjustment requirements", sends them to the large language model for the regeneration of calling information, and confirms again whether to modify until it is confirmed that no modification is needed.

[0043] In step S105, based on the calling information, the modeling instruction is executed to obtain the target modeling.

[0044] In some embodiments, based on the calling information, the method for executing the modeling instruction comprises: obtaining all operation files according to the calling information; for each operation file, obtaining a calling software sequence from the calling information, opening the calling software in the calling software sequence in sequence according to the order, opening the corresponding operation file through the currently opened calling software, and obtaining the function sequence and parameter sequence information corresponding to the currently opened calling software from the calling information, sending the function sequence and parameter sequence information corresponding to the currently opened calling software to the software process of the currently opened calling software, completing the corresponding modeling operation; after completing the component modeling, binding the geometry and non-geometry information corresponding to the obtained component model to the current model in the form of attributes; and searching for the corresponding engineering data information according to the number of the current model, and binding the engineering data information to the component model.

[0045] In one example, according to the call information generated in step S104, modification confirmation is performed, the automation modeling and attribute binding command is executed, and the modeling process is completed. It is selected whether to continue to confirm, if necessary, the generated call information is displayed to the interactive interface, modification is performed, and then new call information is generated, and it is selected again whether to continue to confirm; if not, according to the call information, for each operation file, the corresponding call software is opened in turn, then the operation file is opened through the software, after the file is opened, the operation file opening completion event is monitored through the listening function, and the interface function and parameter instruction sequence sent by the system are executed, and the modeling and attribute binding process is completed. The specific implementation process includes: reading all operation files that need to be operated from the call information, and storing the file path and file format of the operation file in a temporary variable; for each operation file, read the corresponding call software sequence from the call information, open the call software in order, and then open the corresponding operation file with the call software; for the currently opened operation file and the call software, read the call interface function and parameter sequence information from the call information, receive the data request sent in the next “file opening end” listening event, and send the data obtained based on the data request to the software process; in the “file opening end” listening function in the call software, an automation modeling processing program is written in advance, when the operation file is opened, the “call interface function and parameter sequence” request is sent to the system through the processing program, and when the return information is received, the call interface function and parameter sequence information transmitted by the system is executed in order, and the modeling process is completed; when each tunnel component modeling is completed, the geometry and non-geometry information corresponding to the current component is bound to the model in the form of attributes; according to the current model number ID, the corresponding engineering data information is searched, and the corresponding “data type, file content / file path” information is bound to the current component model.

[0046] In some embodiments, after step S105, step S106 of deriving an IFC file based on target modeling is further included, which includes: receiving user inputted export setting information through the interactive interface, the export setting information including export range, export precision and project information; sending the export IFC software and its interface information, the export setting information and the pre-prepared export IFC structured information template to the large language model to obtain a structured export decision scheme; generating an export function call sequence based on the structured export decision scheme; and extracting relevant data from the export decision scheme based on the export function call sequence to complete IFC file export.

[0047] In one example, the user is prompted to select whether to export IFC, and the delivery is completed. If not, it is ended; if yes, export settings are inputted in the interactive window, such as: export range (file path, model type, mileage range, component type), export precision (such as: only display the lightened three-dimensional model, only display the three-dimensional model, including three-dimensional model and attribute information, including three-dimensional model, attribute information and engineering information), project information. The export setting information, export IFC software and interface information, export IFC structured information template are sent to the large language model after processing, and the structured data of the export decision scheme is returned after analysis, and the export function call sequence is generated according to the structured data, and the corresponding software is opened in turn, the interface function is called, the data extracted from the export decision scheme is transmitted, and the IFC file export is completed. That is, the export IFC corresponding export IFC structured information template is defined in advance, designed in Json format, including: calling software path, calling function interface, file path, model type, mileage range, component type, exported attribute and other fields; export setting information, export IFC software and interface information, export IFC structured information template are merged, and then sent to the large language model for analysis; obtain the structured data of the export decision scheme returned by the large language model, parse it into a data object, and generate a function call sequence according to the data object; based on the function call sequence, the corresponding software is opened in turn, the interface function is called, the data extracted from the export decision scheme is transmitted, and the specific steps of IFC file export include: based on the constructed API dependency graph and historical data, using the "knowledge graph based rule reasoning + multi-objective constraint optimization" algorithm to train the model, get the recommended "function call sequence", that is, the export function call sequence; based on "the structured data of the export decision scheme and the export function call sequence, get the recommended preliminary scheme; the recommended scheme of the export function call sequence is detected for conflict, and the conflict is solved by inserting the interface function, changing the parallel or serial task, to get the final export function call sequence; according to the export function call sequence, start the software, and open the file to be operated; in the end of opening the file listening event of the software, the export IFC processing program is written in advance, when the event is listened, the request of calling the export IFC interface function and parameter information is sent to the system through the program; after receiving the return information based on the request, the corresponding interface function and parameter are called in turn, and the IFC export of the current file is completed; if the engineering information is exported, that is, if the engineering information content is exported, the engineering information content is converted to Base64 encoding, and then to string, and then the engineering information content is exported through the IFC custom attribute set "IfcPropertySet". For each file to be operated, the above export function call sequence steps and the steps after them are executed in turn, and the whole IFC export is completed.

[0048] For example, the data of the export decision scheme is as follows: { Tunnel model: { Initial support: {Geometry model: export, geometric properties: export, design properties: export, construction properties: export, operation and maintenance properties: export, engineering data: export}, Secondary lining: {Geometry model: export, geometric properties: export, design properties: export, construction properties: export, operation and maintenance properties: export, engineering data: export}, Inverted arch backfill: {Geometry model: export, geometric properties: export, design properties: export, construction properties: export, operation and maintenance properties: export, engineering data: export}, Pavement paving: {Geometry model: export, geometric properties: export, design properties: export, construction properties: export, operation and maintenance properties: export, engineering data: export} }, Steel model: { Circumferential reinforcement: {Geometry model: export, geometric properties: export, design properties: export, construction properties: export, operation and maintenance properties: export, engineering data: export}, Longitudinal reinforcement: {Geometry model: export, geometric properties: export, design properties: export, construction properties: export, operation and maintenance properties: export, engineering data: export}, Tension reinforcement: {Geometry model: export, geometric properties: export, design properties: export, construction properties: export, operation and maintenance properties: export, engineering data: export} } The application supports the collaborative execution of multi-specialty tasks such as tunnel body, advance support, and steel reinforcement modeling, combined with structured information templates (such as tunnel modeling information templates in JSON format) and dynamic modeling function call sequences. For example, the user triggers the tunnel main structure modeling and tunnel steel reinforcement modeling processes simultaneously through natural language instructions, and the system automatically integrates parameters such as route name, mileage range, and cross-section type to generate a unified model across specialties. This mechanism significantly expands the application scenarios of AI modeling, providing an integrated solution for the entire life cycle of projects in the fields of railways, bridges, and underground engineering, and realizing collaborative modeling of the entire life cycle of tunnels.

[0049] The application realizes intelligent IFC delivery decision-making, opens up data transmission channels between construction and operation and maintenance management platforms, and the IFC export decision engine dynamically generates export decision schemes through AI: the user inputs export settings such as export range (such as "export only lining model, mileage 0+500-0+800") and precision requirements (such as "include attribute information"), and the system automatically generates the optimal export decision scheme (such as calling the export IFC interface function to export the IFC4.0 file of the specified component) based on the API dependency graph. Through testing, this scheme supports precision grading export from LOD300 to LOD500, adapts to the compliance requirements of model delivery, and greatly improves the delivery precision and rationality.

[0050] The application realizes prompt template information driven multi-process function dynamic call, adapts to the flexible demand of BIM deepening modeling, and proposes a prompt template information driven function sequence generation technology. The system matches preset prompt template information (including tunnel software main structure modeling function + tunnel software reinforcement modeling function call sequence) from the database according to the modeling scene (such as “generate V-level surrounding rock tunnel main structure + tunnel reinforcement model”), or dynamically generates a function sequence (such as “first call route, tunnel main structure modeling interface function → then call tunnel reinforcement modeling interface function”) through a trained AI model and API dependency map. This mechanism supports the simultaneous start of multiple tunnel design software processes, seamless collaboration between independent functions, and effective improvement of the efficiency of tunnel BIM deepening modeling.

[0051] The application realizes intelligent binding of engineering data and models, can improve construction quality traceability efficiency, automatically analyzes non-structured data such as construction period engineering data (construction log, detection report), operation and maintenance data (asset information, operation and maintenance record) through AI, and binds to the corresponding tunnel component (such as binding monitoring data to the lining model). In IFC delivery, data links or direct embedding are supported, making BIM models become comprehensive carriers of construction and operation and maintenance stage data, and greatly improving the quality evaluation and problem traceability efficiency.

[0052] The tunnel BIM intelligent collaborative modeling method of the embodiment supports multi-component collaborative modeling through structured information templates and dynamic modeling function call sequences, can realize whole life cycle tunnel collaborative modeling, realizes a natural language driven dynamic interaction mechanism based on a large language model, improves error correction and data supplement efficiency, realizes automatic checking of structured data, guarantees input / output consistency, dynamically generates export decision schemes through AI, realizes intelligent IFC delivery decision, and opens up data transmission channels between construction and operation and maintenance management platforms, proposes a prompt template information driven modeling function sequence generation technology, which can adapt to the flexible demand of BIM deepening modeling, and automatically analyzes relevant engineering data during the construction period through AI, realizes intelligent binding of engineering data and models, and improves construction quality traceability efficiency. The method of the application can also be extended to whole life cycle intelligent modeling and delivery and operation and maintenance management of bridges, subways, underground comprehensive pipe galleries and other infrastructures.

[0053] In order to clearly illustrate the above-mentioned embodiments, specific examples are described. Figure 6 A flowchart of a tunnel BIM intelligent collaborative modeling system provided by the embodiment of the application is shown in FIG. 1. Figure 6As shown, the tunnel BIM intelligent collaborative modeling method of the present application is applied to a tunnel BIM intelligent collaborative modeling system, which includes the following implementation process: entering an interactive interface, receiving natural language information input by a user and performing arrangement, sending the arranged information and engineering data files to a large language model, and receiving return information from the large language model; checking whether the return information is structured data, if not, prompting the user to re-input natural language information; if yes, verifying whether the structured data has data errors or omissions, if there are data errors or omissions, generating an error prompt and returning to the user to prompt the user to re-input correct modeling information according to the error prompt; if there are no data errors or omissions, searching for matched prompt template information, if matched, calling the modeling function interface sequence built-in the prompt template information; if not matched, generating a modeling function interface sequence. According to the structured data and the modeling function interface sequence, generate calling information, determine whether the user needs to confirm the calling information, if yes, prompt the user to input adjustment requirements, and re-generate the calling information according to the user's input adjustment requirements; if the user does not need to confirm, execute the automatic modeling and attribute binding command, and complete the modeling. After completing the modeling, determine whether the file needs to be exported, if not, end the current processing process; if yes, obtain the export requirements input by the user, generate an IFC export decision scheme according to the export requirements, call the IFC export interface, and output the IFC file.

[0054] To implement the above-mentioned embodiments, the present application further provides a tunnel BIM intelligent collaborative modeling device configured in a tunnel BIM intelligent collaborative modeling system. Figure 7 A structural schematic diagram of a tunnel BIM intelligent collaborative modeling device provided by an embodiment of the present application is shown in FIG. 7. Figure 7 As shown, the tunnel BIM intelligent collaborative modeling device can include a data integration module 70701, a data generation module 702, a function acquisition module 703, a calling acquisition module 704, and a modeling implementation module 705.

[0055] The data integration module 70701 is configured to receive user natural language input information and tunnel-related engineering data files through an interactive interface, and perform integration processing on the information in the engineering data files and the user natural language input information through pre-prepared front-end information and structured information templates, to obtain integrated information, wherein the front-end information includes modeling-related information, and the structured information template is used to instruct the large language model to generate structured data, and the user natural language input information includes modeling-related information. The data generation module 702 is configured to send the integrated information to the large language model for analysis, to obtain structured data. The function obtaining module 703 is configured to obtain a modeling function calling sequence based on the modeling scene information in the structured data, wherein the modeling function calling sequence comprises a sequence of interface functions to be executed, the sequence of interface functions stores required calling software and a function sequence to be executed corresponding to the calling software in sequence, and the modeling scene information comprises at least one modeling behavior information. The calling obtaining module 704 is configured to generate structured calling information based on the modeling function calling sequence and the structured data. The modeling implementing module 705 is configured to execute a modeling instruction based on the calling information to obtain target modeling.

[0056] Further, in a possible implementation of the embodiment of the application, the data integrating module 70701 is further configured to: obtain all modeling behaviors involved in the automatic modeling; determine modeling information involved in each modeling behavior and a field type corresponding to each modeling information; integrate the modeling behavior and the corresponding modeling information in a Json format based on the field type corresponding to each modeling information to obtain a structured information template.

[0057] Further, in a possible implementation of the embodiment of the application, after obtaining the structured data, the data generating module 702 is further configured to: verify whether there is an error or a missing field or data in the structured data based on the structured information template; if there is, generate error prompt information and display the error prompt information through an interactive interface to remind a user to input user natural language input information and engineering data files through the interactive interface again to generate the structured data according to the error prompt information.

[0058] Further, in a possible implementation of the embodiment of the application, the function obtaining module 703 is specifically configured to: search prompt template information matched with the modeling scene information from a database based on the modeling scene information in the structured data; obtain the modeling function calling sequence based on the prompt template information; wherein the prompt template information is used to describe the modeling function calling sequence in different modeling scenes; if the prompt template information matched with the modeling scene information is not searched from the database, construct an API dependency graph based on BIM modeling software related information; generate the modeling function calling sequence based on the modeling scene information and the API dependency graph, combined with an algorithm model based on knowledge graph reasoning rules and reinforcement learning dynamic decision-making; perform conflict verification on the modeling function calling sequence; If there is a conflict, the modeling function call sequence is adjusted to obtain an adjusted modeling function call sequence.

[0059] Further, in a possible implementation manner of the embodiment of the application, the modeling related information includes operation file information, the operation file information being used to indicate a file to be generated into a model, the calling obtaining module 704 is specifically used to: obtain the operation file information based on the structured data; for each operation file in the operation file information, obtain calling software and corresponding function sequences and parameter information of the calling software from the modeling function call sequence, and assign each parameter in the parameter information from the structured data; generate structured calling information based on the operation file information and the calling software and corresponding function sequences and parameter information of the calling software corresponding to each operation file according to a pre-prepared calling information data structure template; prompt whether the calling information needs to be modified through an interactive interface; receive adjustment requirement information input by a user through the interactive interface, and send the pre-position information, the calling information and the adjustment requirement information to a large language model to regenerate the calling information.

[0060] Further, in a possible implementation manner of the embodiment of the application, the modeling implementation module 705 is specifically used to: obtain all operation files according to the calling information; for each operation file, obtain a calling software sequence from the calling information, open the calling software in the calling software sequence in sequence, open a corresponding operation file through the currently opened calling software, obtain function sequences and parameter sequences information corresponding to the currently opened calling software from the calling information, send the function sequences and the parameter sequences information corresponding to the currently opened calling software to a software process of the currently opened calling software, and complete a corresponding modeling operation; after completing component modeling, bind geometry and non-geometry information corresponding to a component model to a current model in an attribute form, search for corresponding engineering data information according to a number of the current model, and bind the engineering data information to the component model.

[0061] Further, in a possible implementation manner of the embodiment of the application, the apparatus further includes a file exporting module 706, configured to: receive export setting information input by a user through an interactive interface, the export setting information including an export range, an export precision and project information; send the export IFC software and interface information, the export setting information and a pre-prepared export IFC structured information template to a large language model to obtain structured export decision schemes; Based on the structured export decision scheme, an export function call sequence is generated; Based on the export function call sequence, relevant data is extracted from the export decision scheme to complete the IFC file export.

[0062] It should be noted that the foregoing explanation and description of the tunnel BIM intelligent collaborative modeling method embodiment also applies to the tunnel BIM intelligent collaborative modeling device of this embodiment, which will not be described here.

[0063] In order to realize the above-mentioned embodiments, the present application further provides an electronic device. Please see Figure 8 , Figure 8 is a structural schematic diagram of an electronic device provided by an embodiment of the present application. As shown in Figure 8 , the electronic device 800 includes a processor 801 and a memory 802 connected with the processor 801; the memory 802 stores computer execution instructions; the processor 801 executes the computer execution instructions stored in the memory to realize the method provided by the foregoing embodiments.

[0064] In order to realize the above-mentioned embodiments, the present application further provides a computer readable storage medium, and the computer readable storage medium stores computer execution instructions, and the computer execution instructions are executed by a processor to realize the method provided by the foregoing embodiments.

[0065] In order to realize the above-mentioned embodiments, the present application further provides a computer program product, which includes a computer program, and the computer program is executed by a processor to realize the method provided by the foregoing embodiments.

[0066] In the foregoing embodiment description, the description of the terms “one embodiment”, “some embodiments”, “example”, “specific example”, or “some examples” means that the specific features, structures, materials or characteristics described in combination with the embodiment or example are included in at least one embodiment or example of the present application. In the description of the present application, the illustrative description of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any embodiment or example in a suitable manner. In addition, the person skilled in the art can combine and combine the different embodiments or examples described in the description and the features of the different embodiments or examples without contradiction.

[0067] In addition, the terms “first”, “second” are only for descriptive purposes, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of indicated technical features. Therefore, the features limited by “first”, “second” can explicitly or implicitly include at least one of the features. In the description of the present application, the meaning of “multiple” is at least two, for example, two, three, etc., unless otherwise specifically limited.

[0068] Any processes or methods described in the flowcharts or otherwise described herein can be understood as representing code modules, segments, or portions of code which include one or more executable instructions for implementing specific logic functions (or steps) of the application, and alternate implementations are possible. The various steps or functions described in the flowcharts or otherwise described herein can be implemented as program instructions (i.e., as one or more modules of computer program code) in any of a variety of programming languages. The various steps or functions described in the flowcharts or otherwise described herein can be implemented as machine or computer readable code on a computer readable medium. Thus, the various aspects of the application can be embodied in hardware and / or in software (including firmware, resident software, micro-code, etc.) that runs on a processor such as a digital signal processor, which can collectively be referred to as "an appropriate circuit" or "a module." A software module can comprise one or more instructions that, when executed by a processor, carry out the designated computation or step. As will be understood by those skilled in the art, the software modules can be stored on a computer readable medium, including the memory 110, before, during or after execution by a processor. Accordingly, the various aspects of the application can be embodied in a number of different forms, all of which have been contemplated to be within the scope of the applicable patent princi¬ ples described herein. For example, the various embodiments can take the form of a computer program product on a computer readable storage medium having computer system readable data and / or computer program code embodied in the fabric of the medium that can be used to program computers and / or other

[0069] The logic and / or steps represented in the flowcharts or otherwise described herein, for example, can be embodied in computer executable code, such as in the form of software modules or portions of code that are written in any of a variety of computer programming languages or scripting languages. The logic and / or steps represented in the flowcharts or otherwise described herein can be embodied in any of a variety of ways, including as a stand-alone software package, as a software module, component or section of a program, as a utility for use with

[0070] It should be understood that aspects of the application can be implemented in hardware, software, firmware, or combinations thereof. In the above embodiments, various steps or methods can be implemented in software or firmware that is stored in memory and executed by a suitable instruction execution system. As well, if desired, the various steps or methods can be implemented using any of a variety of technologies that are well known in the art, including but not limited to: discrete logic circuitry having logic gates for implementing logic functions upon data signals, application specific integrated circuits having appropriate combinational logic gates, programmable gate arrays (PGA), field programmable gate arrays (FPGA), and / or the like.

[0071] Those skilled in the art of the present technology can understand that all or part of the steps carried out by the above-mentioned embodiment methods can be completed by programs instructing related hardware, and the programs can be stored in a computer readable storage medium. When the program is executed, it includes one of the steps of the method embodiment or a combination thereof.

[0072] In addition, each functional unit in each embodiment of the present application can be integrated into one processing module, or each unit can exist physically alone, or two or more units can be integrated into one module. The integrated module can be realized in the form of hardware or in the form of a software functional module. When the integrated module is realized in the form of a software functional module and sold or used as an independent product, it can also be stored in a computer readable storage medium.

[0073] The storage medium mentioned above can be a read-only memory, a magnetic disk or an optical disk, etc. Although the embodiments of the present application have been shown and described above, it should be understood that the above-mentioned embodiments are exemplary and cannot be understood as limiting the present application, and those skilled in the art can make changes, modifications, replacements and variations to the above-mentioned embodiments within the scope of the present application.

Claims

1. A tunnel BIM intelligent collaborative modeling method, characterized in that, The method is applied to a tunnel BIM intelligent collaborative modeling system and comprises the following steps: Receiving user natural language input information and tunnel-related engineering data files through an interactive interface, integrating information in the engineering data files and the user natural language input information through prefabricated front information and a structured information template to obtain integrated information, wherein the front information comprises modeling-related information, the structured information template is used to instruct a large language model to generate structured data, and the user natural language input information comprises modeling-related information; Sending the integrated information to the large language model for analysis to obtain structured data; Obtaining a modeling function call sequence based on modeling scenario information in the structured data, wherein the modeling function call sequence comprises a sequence of interface functions to be executed, the sequence of interface functions stores required calling software and a function sequence to be executed corresponding to the calling software in sequence, and the modeling scenario information comprises at least one modeling behavior information; Generating structured calling information based on the modeling function call sequence and the structured data; Executing modeling instructions based on the calling information to obtain target modeling.

2. The method of claim 1, wherein, The method for constructing the structured information template comprises: Obtaining all modeling behaviors involved in automatic modeling; Determining modeling information involved in each modeling behavior and field types corresponding to each modeling information; Integrating the modeling behaviors and their corresponding modeling information through a Json format based on the field types corresponding to the modeling information to obtain a structured information template.

3. The method of claim 1, wherein, After obtaining the structured data, the method comprises: Verifying whether there are errors or omissions in fields or data in the structured data based on the structured information template; If there are, generating error prompt information and displaying the error prompt information through the interactive interface to remind the user to input user natural language input information and engineering data files through the interactive interface again to generate structured data.

4. The method of claim 1, wherein, Obtaining a modeling function call sequence based on modeling scenario information in the structured data, the method comprises: Searching prompt template information matching the modeling scenario information from a database based on the modeling scenario information in the structured data; Obtaining a modeling function call sequence based on the prompt template information; wherein the prompt template information is used to describe modeling function call sequences in different modeling scenarios; If the prompt template information matching the modeling scenario information is not searched from the database, constructing an API dependency graph based on BIM modeling software-related information; Generating the modeling function call sequence based on the modeling scenario information and the API dependency graph, combining an algorithm model based on knowledge graph reasoning rules and reinforcement learning dynamic decision-making; Performing conflict checking on the modeling function call sequence; If there is a conflict, adjusting the modeling function call sequence to obtain an adjusted modeling function call sequence.

5. The method of claim 1, wherein, The modeling-related information includes operation file information, which is used to indicate a file to be generated into a model. The structured call information is generated based on the modeling function call sequence and the structured data, including: Operation file information is obtained based on the structured data; For each operation file in the operation file information, call software and its corresponding function sequence and parameter information are obtained from the modeling function call sequence, and each parameter in the parameter information is assigned from the structured data; According to the pre-prepared call information data structure template, the operation file information and the call software and its corresponding function sequence and parameter information corresponding to each operation file are used to generate structured call information; Whether the call information needs to be modified is prompted through the interactive interface; The pre-prepared information, the call information, and the adjustment requirement information input by the user are sent to a large language model to re-generate the call information.

6. The method of claim 1, wherein, The modeling instruction is executed based on the call information to obtain a target modeling, including: All operation files are obtained according to the call information; For each operation file, the call software sequence is obtained from the call information, the call software in the call software sequence is opened in sequence, the corresponding operation file is opened through the currently opened call software, the function sequence and the parameter sequence information corresponding to the currently opened call software are obtained from the call information, the function sequence and the parameter sequence information corresponding to the currently opened call software are sent to the software process of the currently opened call software, and the corresponding modeling operation is completed. After the component modeling is completed, the geometry and non-geometry information corresponding to the obtained component model is bound to the current model in the form of attributes, and the corresponding engineering information is searched according to the number of the current model, and the engineering information is bound to the component model.

7. The method of claim 1, wherein, The method further includes: Export setting information including export range, export precision, and project information is received through the interactive interface; The export IFC software and its interface information, the export setting information, and the pre-prepared export IFC structured information template are sent to a large language model to obtain structured export decision scheme; Based on the structured export decision scheme, an export function call sequence is generated; Based on the export function call sequence, related data is extracted from the export decision scheme to complete IFC file export.

8. A tunnel BIM intelligent collaborative modeling device, characterized in that, The device is configured in a tunnel BIM intelligent collaborative modeling system, and the device includes: A data integration module is configured to receive user natural language input information and tunnel-related engineering information files through an interactive interface, integrate information in the engineering information files and the user natural language input information through pre-prepared pre-prepared information and structured information templates, and obtain integrated information, wherein the pre-prepared information includes modeling-related information, the structured information template is used to indicate a large language model to generate structured data, and the user natural language input information includes modeling-related information. The data generation module is configured to send the integrated information to a large language model for analysis to obtain structured data. The function acquisition module is configured to obtain a modeling function call sequence based on modeling scenario information in the structured data, wherein the modeling function call sequence includes a sequence of interface functions to be executed, the sequence of interface functions stores required call software and a function sequence corresponding to the call software that needs to be executed in sequence, and the modeling scenario information includes at least one modeling behavior information. The call acquisition module is configured to generate structured call information based on the modeling function call sequence and the structured data. The modeling implementation module is configured to execute modeling instructions based on the call information to obtain a target modeling.

9. An electronic device, comprising: It comprises: a processor and a memory connected in communication with the processor; the memory stores computer execution instructions; the processor executes the computer execution instructions stored in the memory to implement the method of any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, The computer readable storage medium stores computer execution instructions, and the computer execution instructions are executed by the processor to implement the method of any one of claims 1-7.

Images