CN118939769A - A BIM visualization safety technology briefing method based on large language model - Google Patents

A BIM visualization safety technology briefing method based on large language model Download PDF

Info

Publication number
CN118939769A
CN118939769A CN202410973108.2A CN202410973108A CN118939769A CN 118939769 A CN118939769 A CN 118939769A CN 202410973108 A CN202410973108 A CN 202410973108A CN 118939769 A CN118939769 A CN 118939769A
Authority
CN
China
Prior art keywords
language model
risk
technology
risks
large language
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202410973108.2A
Other languages
Chinese (zh)
Inventor
钟宇君
李健
王齐炫
王旭东
孔镜棋
陈丹
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Guangdong Jianke Innovation Technology Research Institute Co ltd
Original Assignee
Guangdong Jianke Innovation Technology Research Institute Co ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Guangdong Jianke Innovation Technology Research Institute Co ltd filed Critical Guangdong Jianke Innovation Technology Research Institute Co ltd
Priority to CN202410973108.2A priority Critical patent/CN118939769A/en
Publication of CN118939769A publication Critical patent/CN118939769A/en
Pending legal-status Critical Current

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
    • G06F16/33—Querying
    • G06F16/332—Query formulation
    • G06F16/3329—Natural language query formulation
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G06Q10/063—Operations research, analysis or management
    • G06Q10/0635—Risk analysis of enterprise or organisation activities
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/10—Office automation; Time management
    • G06Q10/103—Workflow collaboration or project management
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
    • G06Q50/08—Construction
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
    • Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
    • Y02P90/30—Computing systems specially adapted for manufacturing

Landscapes

  • Engineering & Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • Human Resources & Organizations (AREA)
  • Strategic Management (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Economics (AREA)
  • General Physics & Mathematics (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Tourism & Hospitality (AREA)
  • Marketing (AREA)
  • General Business, Economics & Management (AREA)
  • Quality & Reliability (AREA)
  • Data Mining & Analysis (AREA)
  • Mathematical Physics (AREA)
  • Operations Research (AREA)
  • Development Economics (AREA)
  • Game Theory and Decision Science (AREA)
  • Educational Administration (AREA)
  • Primary Health Care (AREA)
  • General Health & Medical Sciences (AREA)
  • Artificial Intelligence (AREA)
  • Human Computer Interaction (AREA)
  • Health & Medical Sciences (AREA)
  • Computational Linguistics (AREA)
  • Databases & Information Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

A BIM visual security technology engagement method based on a large language model comprises the following steps: step 1), constructing a security risk database, wherein the security risk database is formed by cutting and vectorizing a general file related to security risks; step 2) utilizing a customized prompting word template to question a large language model based on the RAG technology, so that the large language model refines engineering characteristics of the project from special files related to the actual project; the prompt word template is a combination of summarized problems of engineering characteristics of a series of auxiliary extraction projects; step 3) inputting the engineering characteristics into the large language model, and requesting the large language model to search risks and corresponding processing measures possibly caused by the engineering characteristics by referring to the safety risk database; step 4) building a building model by using BIM visualization technology, and corresponding the extracted risks to corresponding positions in the building model. The invention can improve the safety technology bottoming efficiency and effect in the building construction process.

Description

BIM visual security technology engagement method based on large language model
Technical Field
The invention belongs to the technical field of building engineering safety production management and artificial intelligence, and particularly relates to a BIM (building information modeling) visual safety technology mating method based on a large language model.
Background
In the conventional construction safety management, safety bottoming is an important link, and aims to ensure that constructors know and master necessary safety knowledge and skills so as to prevent and reduce the occurrence of safety accidents. However, the conventional form of secure engagement is too single, and usually requires manual investigation of risk sources, review of specifications to make treatment measures, and then relies on oral explanation or distribution of some paper information to make secure technical engagement with production personnel. On the one hand, the identification of the safe risk sources in this way is completely dependent on the experience and cognitive scope of the individual, which is time-consuming and labor-consuming, and can be subject to omission and subjective deviations. On the other hand, such a bottoming method is not intuitive enough, and thus is likely to cause insufficient information transfer and variations in understanding.
Disclosure of Invention
The invention aims to provide a method capable of improving the safety technology bottoming efficiency and effect in the building construction process.
The technical problems of the invention are solved by the following technical scheme: a BIM visual security technology engagement method based on a large language model comprises the following steps:
step 1), constructing a safety risk database, wherein the safety risk database is formed by cutting and vectorizing safety production management regulations, safety regulations, construction technical specifications, laws and regulations and other safety risk-related general files;
step 2) utilizing a customized prompting word template to question a large language model based on the RAG technology, so that the large language model refines engineering characteristics of a specific project from specific files related to the actual project, such as investigation design files, construction schemes, major risk source lists and the like of the project; the prompt word template is a combination of summarized problems of engineering characteristics of a series of auxiliary extraction projects;
Step 3) inputting the engineering characteristics into the large language model, and requesting the large language model to search risks and corresponding processing measures possibly caused by the engineering characteristics by referring to the safety risk database;
Step 4) building a building model by using BIM visualization technology, and corresponding the extracted risks to corresponding positions in the building model.
The large language model (Large Language Models, LLMs) is an important development in the field of artificial intelligence in recent years. These models are capable of understanding and generating natural language text through deep learning and natural language processing techniques. With the increase of computing power and the enrichment of data sets, large language models exhibit strong application potential including text generation, semantic understanding, machine translation, and the like. The invention optimizes the large language model based on RAG technology. The mechanism of RAG is particularly applicable to scenarios where information needs to be updated continuously, by which large language models can directly access up-to-date information without retraining in order to generate more reliable output.
In the field of building construction, management of security risks is of paramount importance. Conventional risk identification relies on artificial experience and cognitive scope, and this approach has limitations such as inefficient information induction and is likely to be incomplete. According to the invention, through the application of the large language model based on the RAG technology, a large amount of texts can be rapidly analyzed, potential risks are automatically identified and extracted, and the risk identification efficiency and comprehensiveness are improved. The physical and functional characteristics of the building can be digitally expressed by using BIM visualization technology, so that participants of the building project can intuitively see the content related to the project. According to the invention, the risk information identified by the large language model is combined with the BIM model, so that the risk source is intuitively displayed, and the understanding and understanding of people on the risk source are facilitated, thereby improving the efficiency, the comprehensiveness and the accuracy of the safety technology communication on the whole.
In addition, the invention firstly refines engineering characteristics, and then enables the large language model to search risks and processing measures according to the engineering characteristics, thereby avoiding the problems of weaker comprehensiveness and pertinence of the output result of the large language model caused by weak questioning directivity and the like.
The prompting word template suggests project characteristics of excavation projects from five dimensions of personnel, management, equipment materials, construction process and environment, and the organization mode is more beneficial to ensuring the comprehensiveness of the collected project characteristics.
The specific steps of cutting and vectorizing in the step 1) are as follows:
Document cutting
Cutting a material document to be analyzed to form a first generation document block, cutting the first generation document block to form a second generation document block, forming a hierarchical structure with a multi-layer (more than two layers) father-son relationship, and forming the father document block after all the son document blocks under the same father document block are summarized;
Vectorization
And vectorizing the file blocks of each level of the hierarchical structure, storing the file blocks into a vector database, and constructing a security risk database.
Step 3) belongs to the retrieval generation stage of the RAG technology, and the stage adopts an automatic merging retrieval optimization technology, and specifically comprises the following steps:
After engineering characteristics of the project are extracted by using a large language model based on the RAG technology, vectorizing the project, then carrying out correlation degree calculation on the project as an input source and the smallest sub-document block in the vector database, selecting n smallest sub-Wen Dangkuai with the highest correlation degree, judging whether the quantity or proportion of high-correlation sub-document blocks contained in a father document block reaches a threshold value, if so, providing the father document block for the large language model to identify a potential risk source in the project, otherwise, directly providing the high-correlation sub-document block.
If the hierarchical structure has more levels, the parent document block reaching the threshold value can be used as a new high-correlation document block to further judge whether the large language model should be provided with a parent document block of a higher layer or not.
For large language models based on RAG technology, document cutting is a common technical route that can increase the speed of vector computation, but also creates the possibility of missing context information for the document block. On the basis of the traditional technical route, the invention automatically judges whether to return the child document block or the merged father Wen Dangkuai, and provides a preferable coordination scheme in the aspects of operation speed and information richness.
The method classifies and prioritizes the potential risks after they are identified.
This step can help construction teams to know more clearly which risks are most urgent, requiring major attention and handling. Meanwhile, by classifying risks, basis can be provided for subsequent risk management strategies.
In the step 4), when the risk information extracted by the large language model is associated with the corresponding position in the BIM model, the risk source is displayed in the BIM model by different modes including marking, color coding and three-dimensional graphics, and the applicability of the different display modes is as follows:
labeling: adapted to point out a specific risk point;
The implementation method comprises the following steps: adding text labels at corresponding positions of the BIM model, and directly describing the risk types existing at the positions;
Color coding: the risk classification method is suitable for distinguishing risks of different types or grades and classifying risk areas;
The implementation method comprises the following steps: assigning different colors to different types of risks;
three-dimensional graphics: the method is suitable for displaying complex risk structures or dynamically-changing risk conditions;
The implementation method comprises the following steps: three-dimensional graphics are used to represent the source of risk and possible propagation paths.
The beneficial effects are that:
1) According to the invention, a complete set of flow is designed, so that intelligent analysis of project files is realized, risks are automatically extracted and intuitively displayed in the BIM model, so that constructors can understand safety requirements more intuitively, comprehensively and accurately, and the efficiency and effect of safety technology mating are improved;
2) According to the invention, engineering characteristics are extracted firstly, and then the large language model searches risks and processing measures according to the engineering characteristics, so that the problems of low comprehensiveness and weak pertinence of an output result of the large language model caused by weak questioning directivity and the like are avoided;
3) The invention adopts the technology of automatic merging, searching and optimizing on the traditional RAG technology route, and provides a preferable coordination scheme in the aspects of operation speed and information richness;
4) In addition, the project engineering characteristics are mined from five dimensions of personnel, management, equipment materials, construction process and environment, and the organization mode can better ensure the comprehensiveness of the collected project characteristics.
Drawings
FIG. 1 is a flow chart of the main steps of the BIM visual security technology engagement method based on the large language model.
Detailed Description
The invention aims to improve the efficiency and effect of safety technology bottoming in the building construction process, so as to provide a BIM visual safety technology bottoming method or system based on a large language model, as shown in fig. 1, wherein the method comprises the following steps or the system is configured to execute the following steps:
Step 1) constructing a safety risk database, wherein the safety risk database is formed by cutting and vectorizing safety production management regulations, safety regulations, construction technical specifications, laws and regulations and other safety risk-related general files. This part belongs to the data preparation phase of the RAG technology.
1.1 Document cutting
Cutting Guan Wendang the safety production management regulations, safety regulations, construction technical specifications, laws and regulations and the like according to a set hierarchical structure to form a plurality of file blocks with father-son relations. If the document is cut according to 1000 token blocks to form a first generation of sub-document blocks, then the first generation of sub-document blocks are cut, if the second generation of sub-document blocks are cut according to 500 token blocks, the lower-level document blocks are the sub-Wen Dangkuai of the upper-level document blocks, the sub-document blocks at the same level are mutually independent, and all sub-document blocks under the same father document block are assembled to form the father document block.
1.2 Vectorization
And vectorizing each level of cut file blocks (for example, step 1.1, forming a second generation sub-file block to illustrate three levels of the second generation sub-file block), for example, vectorizing by adopting a embedding model, and storing the vector data into a vector database to construct a security risk database.
And 2) asking questions of the large language model based on the RAG technology by utilizing a customized prompt word template, so that the large language model extracts engineering characteristics of a specific project from specific files related to the actual project, such as investigation design files, construction schemes, major risk source lists and the like of the project.
The cue word templates are a combination of questions that summarize a series of engineering features that assist in extracting the project. The prompt word template suggests engineering features to mine projects from five dimensions of personnel, management, equipment materials, construction process and environment.
Wherein personnel related includes, but is not limited to, special operations, overhead operations, fire operations, limited space operations, temporary electricity use, and the like;
management related includes, but is not limited to, organizational structure, liability system, emergency plan, etc.;
Equipment material related includes, but is not limited to, forms (slipforms, creeping forms, flying forms, tunnel forms, concrete forms, high formwork), lifting (tower cranes, construction lifts, material lifts, etc.), scaffolds (scaffold types, baskets, discharge platforms, handling platforms), construction machinery (shield machines, pipe-jacking machines, drills, cranes, excavators, crushers, rollers, concrete delivery pump trucks, bulldozers, concrete tank trucks, concrete spreader, cutters, electric welding machines, rammers, etc.), flammable and explosive material items, and the like;
Construction process related includes, but is not limited to, foundation pit engineering (deep foundation pit engineering, basement), undermining engineering, steel structure, net rack and cable membrane structure installation engineering, manual hole digging pile engineering, concrete prefabricated member installation engineering, high slope cutting engineering, high slope filling engineering, demolition engineering, and the like;
environmental concerns include, but are not limited to, geological conditions (soft foundation), groundwater conditions, ambient conditions (road, underground pipeline, rail traffic, river), working environments (tunnel, civil air defense, high temperature, conductive dust, moisture), extreme weather, winter construction, flood season rain season, etc.
The prompting word template can have various organization modes, the problem can not be set completely according to the five classifications provided above, a proper classification system can be established by combining project reality, for example, only new materials and new technologies used in engineering are concerned, the problem can be set only from two angles of equipment materials and construction processes, but in general, engineering characteristics of projects can be mined according to the five angles suggested in the invention, and the engineering characteristics can be collected more comprehensively.
Step 3) inputting the engineering characteristics into the large language model, and requesting the engineering characteristics to search risks and corresponding treatment measures possibly caused by the engineering characteristics against a security risk database. This part belongs to the search generation stage of the RAG technology, and in particular, the stage adopts an automatic merging search optimization technology.
Automatic merging search optimization
After engineering characteristics of the project are extracted by using a large language model based on the RAG technology, vectorizing the project, then carrying out correlation degree calculation on the project as an input source and the smallest sub-document block in the vector database, selecting n smallest sub-Wen Dangkuai with the highest correlation degree, judging whether the quantity or proportion of high-correlation sub-document blocks contained in a father document block reaches a threshold value, if so, providing the father document block for the large language model to identify potential risks and corresponding processing measures in the project, otherwise, directly providing the high-correlation sub-document block. If more layers are cut, whether the large language model should be provided with a document block of a higher layer or not can be further judged upwards, and logic similarity is judged.
For large language models based on RAG technology, document cutting is a common technical route that can increase the speed of vector computation, but also creates the possibility of missing context information for the document block. On the basis of the traditional technical route, the invention automatically judges whether to return the child document block or the merged father Wen Dangkuai, and provides a preferable coordination scheme in the aspects of operation speed and information richness.
Preferably, the present invention classifies and prioritizes potential risks after they are identified. This step can help construction teams to know more clearly which risks are most urgent, requiring major attention and handling. Meanwhile, by classifying risks, basis can be provided for subsequent risk management strategies.
The traditional risk source identification method is mainly characterized in that the construction scheme is manually read, the history data is empirically analyzed and consulted, and the research and judgment efficiency is low. The invention can analyze a large amount of text information by utilizing the rapid processing capability of the large language model, automatically identify and extract potential risks and corresponding processing measures, not only can improve the efficiency of risk identification, but also is beneficial to ensuring comprehensiveness.
In addition, the invention firstly refines engineering characteristics, and then enables the large language model to search risks and processing measures according to the engineering characteristics, thereby avoiding the problems of weaker comprehensiveness and pertinence of the output result of the large language model caused by weak questioning directivity and the like.
Examples of the above steps 1) to 3):
as shown in part by the steps of fig. 1, a security risk database records a number of descriptions of risks and corresponding treatments that may be caused by soft soil floors
As shown in step two of fig. 1, a question is set to a large language model (abbreviated LLM) according to the questions related to the environment in the prompt word template: what is the geological features of the foundation of the project? LLM answer: the project base is mainly composed of deep flowing plastic mucky soil, belongs to a soft soil foundation, is one of engineering characteristics of the project, and can enable a large language model to extract possible risks and corresponding treatment measures in a safety risk database according to the engineering characteristics, as shown in a step three part of fig. 1.
Step 4) building a building model by using BIM visualization technology and corresponding the extracted risk to a corresponding position in the building model, wherein the step comprises the following steps:
building a BIM model: according to the concrete content of the construction scheme, building information, such as building structures, construction equipment, operation environments and the like, is built by using BIM software.
Risk source data integration: and (3) corresponding the risk information obtained by analyzing the large language model to a specific position in the BIM model to form a complete risk source list.
The risk source is combined with the building, so that the risk source is intuitively displayed, and the accuracy of people in understanding and understanding the risk source is improved. More importantly, constructors can quickly, clearly and comprehensively know the possible risks at different positions of the building model through checking the positions, so that a better safe bottoming effect is achieved.
In addition, when the risk information extracted by the large language model is associated with the corresponding position in the BIM model, the risk source can be displayed in the BIM model in various different modes such as labeling, color coding or three-dimensional graphics. The applicability of different display modes is specifically as follows:
labeling: is suitable for indicating a specific risk point, such as a specific component, equipment or construction area.
The implementation method comprises the following steps: text labels are added to corresponding positions of the BIM model, and risk types existing in the positions are directly described, for example, a 'high-falling risk area', 'electric risk', and the like. The method is simple and visual and is easy to understand.
Color coding: the method is suitable for distinguishing risks of different types or grades and dividing risk areas.
The implementation method comprises the following steps: different colors are assigned to different types of risks, e.g. red for high risk areas, yellow for medium risk and blue for low risk. By means of the color change, constructors can quickly identify risk areas and risk levels.
Three-dimensional graphics: the method is suitable for displaying complex risk structures or dynamically-changing risk situations.
The implementation method comprises the following steps: three-dimensional graphics, such as arrows, icons, or other symbols, are used to represent the sources of risk and possible propagation paths. For example, red arrows may be used to demonstrate drainage from low risk pit to high risk pit surges.
The BIM model with the risk source labeling completed is utilized to carry out the safe bottoming of constructors, and the key of the step is as follows:
Interactive security education: through the interactive function of BIM model, let constructor can know the position, the nature and the possible influence of risk source from a plurality of angles and layers.
Detailed safety regulations explain: and combining the detailed safety regulations and operation guidelines provided by the large language model, carrying out comprehensive safety education on constructors, and ensuring that the constructors understand and can obey all safety requirements.
Simulating emergency drilling: and simulating various possible safety accident situations in the BIM model, so that constructors can perform emergency response exercise in the virtual environment, and the emergency processing capacity of the constructors is improved.
Through verification, the efficiency and the effect can be greatly improved compared with the traditional specification mode by adopting the method or the system for carrying out the safety technical engagement.

Claims (9)

1.一种基于大语言模型的BIM可视化安全技术交底方法，其特征在于，包括如下步骤：1. A BIM visualization safety technology disclosure method based on a large language model, characterized by comprising the following steps: 步骤1)构建安全风险数据库，所述安全风险数据库通过将与安全风险相关的通用文件做切割和向量化后形成；Step 1) constructing a security risk database, wherein the security risk database is formed by cutting and vectorizing common files related to security risks; 步骤2)利用定制的提示词模板对基于RAG技术的大语言模型进行提问，让大语言模型从与实际项目相关的专用文件中提炼出该项目的工程特点；所述提示词模板是总结的一系列辅助提取项目的工程特点的问题的组合；Step 2) using a customized prompt word template to ask questions to the large language model based on RAG technology, so that the large language model can extract the engineering characteristics of the project from the dedicated files related to the actual project; the prompt word template is a combination of a series of questions that are summarized to assist in extracting the engineering characteristics of the project; 步骤3)将所述工程特点输入所述大语言模型，要求其对照所述安全风险数据库检索该工程特点可能造成的风险和对应的处理措施；Step 3) inputting the engineering features into the large language model, requiring it to retrieve the risks that may be caused by the engineering features and the corresponding treatment measures by comparing with the security risk database; 步骤4)利用BIM可视化技术建立建筑模型，并将提取的风险对应到所述建筑模型中的相应位置。Step 4) Use BIM visualization technology to establish a building model, and correspond the extracted risks to corresponding positions in the building model. 2.根据权利要求1所述的BIM可视化安全技术交底方法，其特征在于，所述通用文件包括安全生产管理条例、安全规程、施工技术规范、法律法规。2. The BIM visualization safety technology briefing method according to claim 1 is characterized in that the general documents include production safety management regulations, safety regulations, construction technical specifications, and laws and regulations. 3.根据权利要求1所述的BIM可视化安全技术交底方法，其特征在于，所述专用文件包括具体项目的勘察设计文件、施工方案、重大风险源清单。3. The BIM visualization safety technology briefing method according to claim 1 is characterized in that the special files include survey and design documents, construction plans, and a list of major risk sources for specific projects. 4.根据权利要求1所述的BIM可视化安全技术交底方法，其特征在于，所述提示词模板从人员、管理、设备材料、施工工艺和环境五个维度去挖掘项目的工程特点。4. The BIM visualization safety technology briefing method according to claim 1 is characterized in that the prompt word template explores the engineering characteristics of the project from five dimensions: personnel, management, equipment and materials, construction technology, and environment. 5.根据权利要求1所述的BIM可视化安全技术交底方法，其特征在于，步骤1)中做切割和向量化的具体步骤如下：5. The BIM visualization safety technology disclosure method according to claim 1 is characterized in that the specific steps of cutting and vectorization in step 1) are as follows: 文档切割Document cutting 将待分析的材料文档进行切割，形成第一代文档块，再对第一代文档块进行切割，形成第二代文档块，如此形成具有多层父子关系的层次结构，同一个父文档块下的全部子文档块汇总后，构成其该父文档块；The material document to be analyzed is cut into first-generation document blocks, and then the first-generation document blocks are cut into second-generation document blocks, so as to form a hierarchical structure with multiple layers of parent-child relationships. All child document blocks under the same parent document block are aggregated to form the parent document block; 向量化Vectorization 将所述层次结构的各级文档块进行向量化，存入向量数据库，组建出一个安全风险数据库。The document blocks at each level of the hierarchical structure are vectorized and stored in a vector database to construct a security risk database. 6.根据权利要求5所述的BIM可视化安全技术交底方法，其特征在于，步骤3)属于RAG技术的检索生成阶段，该阶段采用了自动合并检索优化技术，具体如下：6. The BIM visualization safety technology disclosure method according to claim 5 is characterized in that step 3) belongs to the retrieval generation stage of the RAG technology, which adopts the automatic merging retrieval optimization technology, which is as follows: 利用基于RAG技术的大语言模型提炼出项目的工程特点后，对其进行向量化，然后作为输入源与所述向量数据库中的最小子文档块进行相关程度计算，选择相关程度最高的n个最小子文档块，判断其父文档块包含的高相关子文档块的数量或比例是否达到阈值，达到则将该父文档块提供给所述大语言模型以识别项目中的潜在风险源，否则直接提供所述高相关子文档块。After the engineering characteristics of the project are extracted using a large language model based on RAG technology, they are vectorized and then used as an input source to calculate the correlation with the minimum sub-document block in the vector database. The n minimum sub-document blocks with the highest correlation are selected to determine whether the number or proportion of highly correlated sub-document blocks contained in their parent document blocks reaches a threshold. If so, the parent document block is provided to the large language model to identify potential risk sources in the project. Otherwise, the highly correlated sub-document blocks are directly provided. 7.根据权利要求6所述的BIM可视化安全技术交底方法，其特征在于，若所述层次结构的级数较多，将达到所述阈值的父文档块作为新的高相关文档块进一步向上判断是否应该给所述大语言模型提供更上层的父文档块。7. The BIM visualization safety technology briefing method according to claim 6 is characterized in that if the number of levels of the hierarchical structure is large, the parent document block that reaches the threshold is used as a new high-relevant document block to further determine whether a higher-level parent document block should be provided to the large language model. 8.根据权利要求1所述的BIM可视化安全技术交底方法，其特征在于，所述方法在识别出潜在风险后，对这些风险进行分类和优先级排序。8. The BIM visualization safety technology briefing method according to claim 1 is characterized in that after identifying potential risks, the method classifies and prioritizes these risks. 9.根据权利要求1所述的BIM可视化安全技术交底方法，其特征在于，步骤4)中将大语言模型提取的风险信息与BIM模型中的相应位置进行关联时，所述方法通过包括标注、颜色编码和三维图形这些不同的方式在BIM模型中展示风险源，不同展示方式的适用性具体如下：9. The BIM visualization safety technology disclosure method according to claim 1 is characterized in that when the risk information extracted by the large language model is associated with the corresponding position in the BIM model in step 4), the method displays the risk source in the BIM model in different ways including annotation, color coding and three-dimensional graphics, and the applicability of different display methods is as follows: 标注：适用于指出具体的风险点；Marking: suitable for pointing out specific risk points; 实施方法：在BIM模型的相应位置添加文本标注，直接说明该处存在的风险类型；Implementation method: Add text annotations at the corresponding locations of the BIM model to directly indicate the type of risk that exists there; 颜色编码：适用于区分不同类型或等级的风险，以及对风险区域进行划分；Color coding: suitable for distinguishing different types or levels of risks and dividing risk areas; 实施方法：为不同类型的风险分配不同的颜色；Implementation method: Assign different colors to different types of risks; 三维图形：适用于展示复杂的风险结构或动态变化的风险情况；Three-dimensional graphics: suitable for displaying complex risk structures or dynamically changing risk situations; 实施方法：利用三维图形来表示风险的来源和可能的传播路径。Implementation method: Use three-dimensional graphics to represent the sources of risks and possible transmission paths.
CN202410973108.2A 2024-07-19 2024-07-19 A BIM visualization safety technology briefing method based on large language model Pending CN118939769A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202410973108.2A CN118939769A (en) 2024-07-19 2024-07-19 A BIM visualization safety technology briefing method based on large language model

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202410973108.2A CN118939769A (en) 2024-07-19 2024-07-19 A BIM visualization safety technology briefing method based on large language model

Publications (1)

Publication Number Publication Date
CN118939769A true CN118939769A (en) 2024-11-12

Family

ID=93353717

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202410973108.2A Pending CN118939769A (en) 2024-07-19 2024-07-19 A BIM visualization safety technology briefing method based on large language model

Country Status (1)

Country Link
CN (1) CN118939769A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN119516287A (en) * 2025-01-21 2025-02-25 四川观筑数智科技有限公司 A material attribute acquisition method and system based on large language model and BIM

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN119516287A (en) * 2025-01-21 2025-02-25 四川观筑数智科技有限公司 A material attribute acquisition method and system based on large language model and BIM

Similar Documents

Publication Publication Date Title
Ding et al. Safety risk identification system for metro construction on the basis of construction drawings
Yuan et al. Accident prevention through design (PtD): Integration of building information modeling and PtD knowledge base
Peng et al. A GIS-based evaluation method of underground space resources for urban spatial planning: Part 1 methodology
Zhang et al. BIM-based risk identification system in tunnel construction
Bang et al. Context-based information generation for managing UAV-acquired data using image captioning
Zhou et al. Building information modelling review with potential applications in tunnel engineering of China
Karimi The spatial logic of organic cities in Iran and the United Kingdom
Luo et al. A BIM-based code compliance checking process of deep foundation construction plans
CN114358507B (en) Risk assessment and prevention method for water transmission buildings based on knowledge graph
CN102609774B (en) System and method for identifying and reasoning security risks of metro construction
Chen et al. Integrating knowledge management and BIM for safety risk identification of deep foundation pit construction
Li et al. Managing safety hazards in metro subway projects under complex environmental conditions
CN106897445A (en) It is a kind of that the novel method that operation delimited on border is improved based on ArcGIS
CN111859779A (en) Early warning method and device for preventing third-party construction damage risk of gas pipe network
CN115186052A (en) Earthquake disaster situation rescue information determination method and device
Cai et al. Improved hybrid reasoning approach to safety risk perception under uncertainty for mountain tunnel construction
CN118939769A (en) A BIM visualization safety technology briefing method based on large language model
CN117274010A (en) Scene coping method and system for mine disaster emergency rescue
Liu et al. Risk assessment of deep excavation construction based on combined weighting and nonlinear FAHP
Chen Hazards identification and characterisation of the tailings storage facility dam failure and engineering applications
Chen et al. Research on the impact of underground excavation metro on surface traffic safety and assessment method
CN117473989A (en) Geological document key information automatic extraction and report generation method
Liu et al. A pre-control method in safety risk management of metro construction
Nguyen et al. Discovering workers’ actions leading to severe construction accidents using accident report data and sequence mining techniques
Ma et al. Quality risk evaluation of urban rail transit construction based on AHP–FCE method

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination