A water conservancy and hydropower construction scheme generation method based on construction object feature modeling

Abstract

The application provides a water conservancy and hydropower construction scheme generation method based on construction object feature modeling, comprising the following steps: step 1) construction object feature identification and modeling; step 2) construction scheme generation constraint construction: combining engineering specifications, historical schemes, expert experience and other data, a set of multiple source heterogeneous constraint conditions is constructed, and the set of constraint conditions is in a structured form to participate in the combination and generation of subsequent construction scheme content; step 3) combination and generation of construction scheme content: the construction scheme is disassembled into multiple controllable content units, and combination and generation are performed according to different constraint conditions in the generation process. The method adopts a content generation mechanism of "constraint driving + content combination", realizes unified support of a set of technical schemes for multiple construction scenes, and significantly improves the application range of the technical scheme.

Description

Technical Field

[0001] This invention relates to the field of building construction technology, and more specifically, to a method for generating water conservancy and hydropower construction schemes based on modeling the characteristics of construction objects. Background Technology

[0002] Water conservancy and hydropower projects are important national infrastructure construction projects, and the scientific nature and rationality of their construction plans directly affect the quality, safety, and progress of the projects. With the development of information technology, the application of technologies such as artificial intelligence and big data in various industries is becoming increasingly widespread, providing possibilities for the digital and intelligent transformation of the water conservancy and hydropower industry. Traditional water conservancy and hydropower construction plan preparation mainly relies on manual review of materials and writing based on experience. This not only requires the writers to have solid professional knowledge but also consumes a lot of time and energy. With the continuous expansion of the scale and increasing technical complexity of water conservancy and hydropower projects, higher demands are placed on the efficiency and quality of construction plan preparation, and traditional preparation methods can no longer meet the actual needs.

[0003] With the development of computer technology, the industry has developed some tools to assist in the preparation of construction plans, mainly including the following two categories: Document template tools: These tools provide fixed construction plan templates, requiring users to manually fill in the specific content. These tools can only achieve format standardization; they cannot automatically generate or optimize content, still relying on manual completion of the core compilation work. Furthermore, template matching requires users to manually select based on experience, resulting in low efficiency.

[0004] A simple information retrieval system can perform keyword searches on standards, specifications, and historical project data to help users find relevant information. However, this system cannot understand the deeper meaning and logical relationships of information, cannot automatically integrate the retrieved information into construction plans, and does not have the ability to generate targeted content based on project characteristics.

[0005] For example, the construction plan module integrated into some engineering management software mainly provides template libraries and data retrieval portals, while the core content writing and integration are still done manually.

[0006] Current large-scale modeling technologies used in engineering for construction plan development go a step further than those in the conventional approach. However, their implementation is typically based on fixed processes or generic text generation logic, failing to systematically model and represent the inherent differences between construction objects. Because they do not clearly distinguish and quantify the structural characteristics, engineering environment characteristics, and construction business characteristics of the construction objects, existing technologies struggle to accurately reflect the essential differences between various construction objects when generating construction plans. Specific problems are analyzed below: (1) Differences in construction objects are difficult to express effectively. Due to the lack of a unified modeling method for the characteristics of construction objects, different types of caverns, buildings or construction units often adopt the same or highly similar content structure during the scheme generation process, resulting in insufficient targeting of the generated construction schemes.

[0007] (2) The construction plan has poor reusability and insufficient consistency. In existing technologies, construction plans are mainly generated through manual copying, modification, or based on general rules. There is a lack of clear feature mapping relationships between different construction objects, which can easily lead to problems such as redundancy, omissions, or mismatch between the plan content and the actual construction objects.

[0008] (3) Difficult to adapt to combined scenarios of multiple business types and complex engineering conditions When different construction business types such as blasting, excavation and support, installation and demolition coexist, existing technologies are unable to dynamically adjust the content of the plan according to the business attributes of the construction object and the engineering environment conditions, which limits the adaptability of the construction plan and the feasibility of the project.

[0009] (4) The process of generating construction plans is highly dependent on human experience. Due to the lack of constraint mechanisms based on the characteristics of the construction object, existing technologies rely heavily on manual judgment in the process of generating and adjusting solutions. The efficiency and quality of the solutions are limited by the experience level of the compilers, making it difficult to form a stable and reusable technical path.

[0010] In summary, the lack of consideration for different construction objects in existing technologies results in construction plans that are more akin to templates, lacking differentiated expression, lacking practical significance, and severely limiting the application scenarios.

[0011] In order to solve the above problems, people have been seeking an ideal technological solution. Summary of the Invention

[0012] The purpose of this invention is to address the shortcomings of existing technologies by providing a water conservancy and hydropower construction scheme generation method based on construction object feature modeling. This method involves systematically modeling the characteristics of different construction objects in water conservancy and hydropower projects, transforming the structural features, engineering environment features, and construction business features of the construction objects into constraints for generating construction schemes, thereby driving the selection, combination, and generation of construction scheme content.

[0013] To achieve the above objectives, the technical solution adopted by this invention is: a method for generating water conservancy and hydropower construction schemes based on construction object feature modeling, comprising the following steps: Step 1) Construction object feature identification and modeling: By decomposing the multi-dimensional features of the construction object, extracting features at the structural, environmental and business levels and modeling them, a construction object feature model is constructed to characterize the differences of the construction object at the structural, environmental and business levels. Step 2) Construction of constraints for construction scheme generation: Based on the construction object feature model obtained in Step 1), and combined with data such as engineering specifications, historical schemes, and expert experience, a set of multi-source heterogeneous constraints is constructed, and this set of constraints is used in a structured form in the subsequent combination and generation of construction scheme content. Step 3) Combination and generation of construction plan content: The construction plan is broken down into multiple controllable content units. Based on the set of constraints generated in Step 2), the units are combined and generated according to different constraints during the generation process.

[0014] Preferably, in step 1), structural features refer to the features used to describe the geometric and structural attributes of the construction object, including: object type, cross-sectional form, dimensional parameters, burial depth conditions, and spatial positional relationships; Engineering environmental characteristics refer to the features used to describe the engineering environmental conditions in which the construction object is located, including: surrounding rock grade, geological category, hydrological conditions, groundwater conditions, special geological risks, and construction risk factors; Construction business characteristics refer to the characteristics used to describe the construction process and business type corresponding to the construction object, including: construction method type, whether it belongs to special construction conditions, and special requirements for construction organization and safety control.

[0015] Preferably, in step 1), the construction of the feature model is carried out by the following method: collecting multiple historical construction plans of implemented or approved water conservancy and hydropower projects as sample data. The historical construction plans include construction plan texts formed under different construction object types, different construction business types, and different engineering environment conditions. Because historical construction plans come from diverse sources and have inconsistent formats, they are standardized before feature identification. This includes: analyzing the chapter structure of the construction plan text; standardizing terminology and description formats; and establishing the relationship between the construction object and the corresponding construction plan. After obtaining candidate features, the extracted features are processed by structured modeling to transform unstructured text features into computable and matchable construction object feature models. The feature models include structural features, engineering environment features, and construction business features. Engineering experts are brought in to review and revise the feature model, confirm or correct the candidate features extracted from the large model, and supplement the features of construction objects that are not explicitly described in the text but have a key impact on the construction plan; the feature classification or feature value is adjusted to conform to the actual engineering situation, and the modeling of the construction object is completed.

[0016] Preferably, in step 2), the process of constructing the set of constraints is as follows: Constraints are modeled using a rule-based approach and parameterized description. The constraints include: Element triggering conditions: One or more feature items in the feature model of the corresponding construction object; Target objects: Chapters, procedures, or parameters in the construction plan; Constraints: Generation conditions that are allowed, prohibited, or required to be met; Constraint priority: Used for decision-making when multiple constraints conflict.

[0017] Preferably, in step 3), the process of combining and generating the construction plan content is as follows: First, the content generation is supported by chapter-based logic. Based on the general structure of water conservancy and hydropower construction schemes, it is structurally decomposed and automatically divided into multi-dimensional chapter units to form a standardized generation framework. Then, based on the basic information of the input project and the core parameters of the solution, solutions with high similarity are matched as candidate context content for the chapters of the large model generation section; Based on the generated chapter framework and matching content, a three-tiered progressive logic is adopted to construct the content: For chapters strongly related to the project, the system calls the project library data interface and directly extracts the original project information through set constraints, ensuring that the content is highly consistent with the actual project; For general chapters, the system selects suitable general standards and specifications from the standard and specification library and the general scheme library according to the current scheme category, realizing the rapid reuse of standardized content; For core professional chapters, the system first extracts the chapter framework and key technical points of high-quality schemes of the same category through text parsing technology, and then inputs these structured elements with the constraints and construction object characteristics of the current project into a large model fine-tuned by water conservancy and hydropower industry data. Through contextual semantic reconstruction and professional terminology calibration, the system matches and generates content that conforms to the actual project and meets industry standards.

[0018] Preferably, in step 3), the combination and generation of the construction plan content also involves a format conversion process, as detailed below: The structured data extracted from general basic information such as project, plan, and construction process includes: title, full text, tables, images, and flowcharts. This data is converted into an intermediate data structure for OnlyOffice. All intermediate data is provided to the OnlyOffice platform in a streaming manner. OnlyOffice parses the intermediate data structure and uses Canvas rendering technology to present all the current content in the collaborative document in a streaming manner.

[0019] This invention has significant substantive features and remarkable progress compared to existing technologies. Specifically, this invention constructs a feature model of the construction object, including structural features, engineering environment features, and construction business features. This feature model enables precise matching between the construction plan and the specific construction object. Before generation, a set of constraints for generating the construction plan based on the features of the construction object is constructed, explicitly transforming engineering specifications, process logic, and safety requirements into computable constraints. Based on a content generation mechanism of "constraint-driven + content combination," the construction plan is decomposed into multiple controllable content units, which are combined and generated according to the constraints during the generation process, rather than being generated as a whole text at once. This makes the plan flexible and widely adaptable.

[0020] This invention effectively solves the problem of difficult unified modeling of various types of water conservancy and hydropower construction objects. It takes the feature modeling of construction objects as a unified entry point, and adapts to different construction objects and construction business types through feature combination. It realizes unified support for multiple construction scenarios with a single technical solution, significantly improving the applicability of the technical solution. Furthermore, through feature-driven constraint construction, it improves the engineering rationality of the generated construction plan, is not affected by the personal experience and level of the writer, and ensures the professionalism and reliability of the plan. The comprehensive intelligent auxiliary functions provided simplify the compilation process and improve the user experience and convenience of plan compilation. Attached Figure Description

[0021] Figure 1 This is a schematic diagram of the overall framework for generating the construction scheme in this invention.

[0022] Figure 2 This is the identification diagram of the feature model of the construction object in this invention.

[0023] Figure 3 This is a flowchart of constraint rule matching in this invention.

[0024] Figure 4 This is a schematic diagram of the operation interface for modeling the feature parameters of the construction object in this invention.

[0025] Figure 5 This is a schematic diagram of the parameters for generating constraint matching rules in this invention.

[0026] Figure 6 This is one of the visualization images generated based on the feature recognition of construction objects in this invention.

[0027] Figure 7 This is the second visualization diagram generated based on the feature recognition of construction objects in this invention.

[0028] Figure 8 This is the third visualization diagram generated based on the feature recognition of construction objects in this invention.

[0029] Figure 9 This is a flowchart illustrating the parsing of external files referenced in this invention.

[0030] Figure 10 This is a diagram illustrating the generation of construction plan content, knowledge recommendation, and citation in this invention. Detailed Implementation

[0031] The technical solution of the present invention will be further described in detail below through specific embodiments.

[0032] like Figures 1-10 As shown, a method for generating water conservancy and hydropower construction schemes based on construction object feature modeling includes the following steps: Step 1) Identification and modeling of construction object features: Modeling the features of construction objects is a key technical step in this invention. Its purpose is to transform the engineering differences of construction objects into structured feature information that can be used to generate construction plans.

[0033] Basic ideas of feature modeling This invention extracts and models the multidimensional features of construction objects to construct a feature model of construction objects, which is used to characterize the differences of construction objects in terms of structure, environment and business.

[0034] The characteristics of a construction object, in specific implementation, include at least the following three categories: (1) Structural features Used to describe the geometric and structural properties of the construction object, including but not limited to: Object type (tunnel, factory, building, etc.); Cross-sectional shape and dimensional parameters; Burial depth conditions or spatial location relationships.

[0035] (2) Engineering environmental characteristics Used to describe the engineering environmental conditions in which the construction object is located, including but not limited to: Rock grade or geological category; Hydrological conditions and groundwater situation; Special geological risks or construction risk factors.

[0036] (3) Characteristics of construction business Used to describe the construction process and business type corresponding to the construction object, including but not limited to: Construction method type (blasting, excavation, support, etc.); Does it fall under special construction conditions? Special requirements for construction organization and safety control.

[0037] The algorithm for constructing the feature model is as follows: We collected historical construction plans from multiple implemented or approved water conservancy and hydropower projects as sample data. These historical construction plans included texts of construction plans formed under different construction object types, different construction business types, and different engineering environment conditions.

[0038] Because historical construction plans come from diverse sources and have inconsistent formats, they are standardized before feature recognition. This includes: analyzing the chapter structure of the construction plan text; standardizing terminology and description formats; and establishing the relationship between the construction object and the corresponding construction plan.

[0039] After obtaining candidate features, the extracted features undergo structured modeling processing, transforming unstructured text features into a computable and matchable feature model of the construction object. The feature model includes structural features, engineering environment features, and construction business features. Through the above feature modeling, the multidimensional differences of the construction object are expressed in a unified feature model form.

[0040] Engineering experts are brought in to review and revise the feature model, confirm or correct the candidate features extracted from the large model, and supplement the features of construction objects that are not explicitly described in the text but have a key impact on the construction plan; feature classification or feature values ​​are adjusted to conform to the actual engineering situation.

[0041] Step 2) Constructing constraints for generating construction plans: By constructing a structured and computable set of constraints for construction scheme generation before the construction scheme is generated, the conditions that need to be met in the engineering field are explicitly introduced into the generation process, thereby achieving "constraint-driven large model generation" to ensure accurate matching between the generated content and the business environment.

[0042] 1. Overall approach to constraint construction Based on the construction object feature model obtained in step 1), and combined with engineering specifications, historical schemes, and expert experience, a set of multi-source heterogeneous constraint conditions is constructed, and this constraint set is used in a structured form to participate in the combination and generation of subsequent construction scheme content.

[0043] 2. Constraint Construction Process Constraints are modeled using a combination of "rule-based expression + parameterized description" to make them computable. Constraints include element triggering conditions (corresponding to one or more feature items in the construction object feature model), the object of action (a chapter, process, or parameter in the construction plan), constraint content (permitted, prohibited, or mandatory generation conditions), and constraint priority (used for adjudicating conflicts between multiple constraints).

[0044] Step 3) Combining and generating construction plan content: Overall, this step uses a content generation mechanism of "constraint-driven + content combination" to break down the construction plan into multiple controllable content units, which are then combined and generated according to constraints during the generation process, rather than generating the entire text all at once.

[0045] The specific implementation process adopts the logic of "structured decomposition - feature matching - layered generation - format adaptation", and the specific process and technical applications are as follows: First, the content generation is supported by chapter-based logic. Based on the general structure of water conservancy and hydropower construction plans (including core modules such as project overview, compilation basis, construction technology, and quality control), the content is structured and automatically divided into multi-dimensional chapter units to form a standardized generation framework. Then, when the user inputs basic project information (such as project code, dam type, and construction period requirements) and core parameters of the scheme through the system interface, the features of the scheme classification are compared with the features of the same category of schemes in the knowledge base. Based on the number of successful matches, the schemes with higher similarity are obtained and used as the context for some chapters of the large model generation section.

[0046] Based on the generated chapter framework and matched content, a three-tiered progressive logic is used to construct the content: For content in strongly related chapters of a project, the system calls the project library data interface and extracts the project's original information directly through set constraints, including engineering geology, construction scale, and schedule milestones, ensuring that the content closely matches the actual project.

[0047] For the content of general sections: This section refers to the sections with strong cross-project reusability, such as the basis for compilation and safety specifications. The system adopts the "scheme category - library table adaptation" retrieval logic, and selects the appropriate industry standards, technical specifications and other content from the standard specification library and general scheme library according to the current scheme category (such as concrete pouring and tunnel excavation), so as to realize the rapid reuse of standardized content.

[0048] For the core professional chapters: The system first extracts the chapter framework and key technical points (such as template installation process and seepage prevention construction parameters) of high-quality solutions of the same category through text parsing technology. Then, it inputs these structured elements with the constraints of the current project and the characteristics of the construction object (such as geological parameters and construction period requirements) into a large model that has been fine-tuned with water conservancy and hydropower industry data. Through contextual semantic reconstruction and professional terminology calibration, the system matches and generates content to produce personalized content that conforms to the actual project and meets industry standards.

[0049] Furthermore, since the entire construction plan requires teamwork to complete and advance, the presentation method utilizes a collaborative technology foundation to build an online collaborative editing environment that natively supports real-time collaborative editing by multiple users, thus ensuring the team collaboration efficiency of water conservancy and hydropower construction plan preparation from the tool level.

[0050] Its content performs a format conversion from "document XML - intermediate data - canvas", generating content including: Headings: Use different HTML headings as tags, depending on the heading level; Full text: Provided directly as a string to the collaboration platform; Tables: Provided to the collaboration platform using table tags; Images: Combine image tags and provide the corresponding image IDs to the collaboration platform; Flowchart: When parsing the document, the flowchart will be generated as a separate Word document, using a group of labeled images, and the generated file ID will be recorded and provided to the collaboration platform.

[0051] For content requiring similar structure, the data generated using LLM is in Markdown format. Since the large model only generates headings, full text, and tables, only the Markdown content needs to be parsed.

[0052] All intermediate layer data is provided to the collaboration platform in a streaming manner. The collaboration platform parses the intermediate format data structure and uses Canvas rendering technology to present all current content in the collaborative document in a streaming manner, allowing users to perceive the writing effect.

[0053] After receiving the globally compiled document data, users can also redraw titles, full text, tables, images, and flowcharts based on the collaborative platform.

[0054] The following content demonstrates some of the functionalities of this solution during its implementation: like Figure 4 As shown, based on the name of the input project and the uploaded content, the relevant parameters of the structured feature information are identified.

[0055] like Figure 5 As shown, based on the generated constraint rules, the most similar content is matched.

[0056] After the user selects fully automatic writing, the system will enter the construction plan writing interface, generate a table of contents outline using a matched plan template, and generate content based on the writing logic of each chapter confirmed by the user's research. There are three main types of generation logic: like Figure 6 As shown, the first step is to extract content from background materials. This part of the content will be generated by the system by extracting relevant content from the background information of this project.

[0057] like Figure 7As shown, secondly, it is generated from similar solutions. The system will match the relevant content of this chapter in similar solution cases based on the characteristics of the previous solutions and integrate them to generate the solution.

[0058] like Figure 8 As shown, thirdly, it is generated from the dependent business knowledge base. The system will match business knowledge data with similar characteristics in the established business knowledge base according to the characteristics of the solution and integrate them to generate the solution.

[0059] like Figure 9 As shown in the figure, if an external file is called during the generation process, the flowchart for obtaining and parsing the content of the external file is as follows.

[0060] like Figure 10 The image shown is a screenshot of the interface displaying the generated content, as well as a screenshot of the related referenced content.

[0061] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them; although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications can still be made to the specific implementation of the present invention or equivalent substitutions can be made to some technical features without departing from the spirit of the technical solutions of the present invention, and all such modifications and substitutions should be covered within the scope of the technical solutions claimed in the present invention.

Claims

1. A method for generating water conservancy and hydropower construction schemes based on construction object feature modeling, characterized in that: Includes the following steps: Step 1) Construction object feature identification and modeling: By decomposing the multi-dimensional features of the construction object, extracting features at the structural, environmental and business levels and modeling them, a construction object feature model is constructed to characterize the differences of the construction object at the structural, environmental and business levels. Step 2) Construction of constraints for construction scheme generation: Based on the construction object feature model obtained in Step 1), and combined with data such as engineering specifications, historical schemes, and expert experience, a set of multi-source heterogeneous constraints is constructed, and this set of constraints is used in a structured form in the subsequent combination and generation of construction scheme content. Step 3) Combination and generation of construction plan content: The construction plan is broken down into multiple controllable content units. Based on the set of constraints generated in Step 2), the units are combined and generated according to different constraints during the generation process.

2. The method for generating water conservancy and hydropower construction schemes based on construction object feature modeling according to claim 1, characterized in that: In step 1), structural features refer to the features used to describe the geometric and structural attributes of the construction object, including: object type, cross-sectional form, dimensional parameters, burial depth conditions, and spatial positional relationships; Engineering environmental characteristics refer to the features used to describe the engineering environmental conditions in which the construction object is located, including: surrounding rock grade, geological category, hydrological conditions, groundwater conditions, special geological risks, and construction risk factors; Construction business characteristics refer to the characteristics used to describe the construction process and business type corresponding to the construction object, including: construction method type, whether it belongs to special construction conditions, and special requirements for construction organization and safety control.

3. The method for generating water conservancy and hydropower construction schemes based on construction object feature modeling according to claim 2, characterized in that: In step 1), the construction of the feature model is carried out by the following method: collecting multiple historical construction plans of implemented or approved water conservancy and hydropower projects as sample data. The historical construction plans include construction plan texts formed under different construction object types, different construction business types, and different engineering environment conditions. Because historical construction plans come from diverse sources and have inconsistent formats, they are standardized before feature identification. This includes: analyzing the chapter structure of the construction plan text; standardizing terminology and description formats; and establishing the relationship between the construction object and the corresponding construction plan. After obtaining candidate features, the extracted features are processed by structured modeling to transform unstructured text features into computable and matchable construction object feature models. The feature models include structural features, engineering environment features, and construction business features. Engineering experts are brought in to review and revise the feature model, confirm or correct the candidate features extracted from the large model, and supplement the features of construction objects that are not explicitly described in the text but have a key impact on the construction plan; the feature classification or feature value is adjusted to conform to the actual engineering situation, and the modeling of the construction object is completed.

4. The method for generating water conservancy and hydropower construction schemes based on construction object feature modeling according to claim 3, characterized in that: In step 2), the process of constructing the set of constraints is as follows: Constraints are modeled using a rule-based approach and parameterized description. The constraints include: Element triggering conditions: One or more feature items in the feature model of the corresponding construction object; Target objects: Chapters, procedures, or parameters in the construction plan; Constraints: Generation conditions that are allowed, prohibited, or required to be met; Constraint priority: Used for decision-making when multiple constraints conflict.

5. The method for generating water conservancy and hydropower construction schemes based on construction object feature modeling according to claim 4, characterized in that: In step 3), the process of combining and generating the construction plan content is as follows: First, the content generation is supported by chapter-based logic. Based on the general structure of water conservancy and hydropower construction schemes, it is structurally decomposed and automatically divided into multi-dimensional chapter units to form a standardized generation framework. Then, based on the basic information of the input project and the core parameters of the solution, solutions with high similarity are matched as candidate context content for the chapters of the large model generation section; Based on the generated chapter framework and matching content, a three-tiered progressive logic is adopted to construct the content: For chapters strongly related to the project, the system calls the project library data interface and directly extracts the original project information through set constraints, ensuring that the content is highly consistent with the actual project; For general chapters, the system selects suitable general standards and specifications from the standard and specification library and the general scheme library according to the current scheme category, realizing the rapid reuse of standardized content; For core professional chapters, the system first extracts the chapter framework and key technical points of high-quality schemes of the same category through text parsing technology, and then inputs these structured elements with the constraints and construction object characteristics of the current project into a large model fine-tuned by water conservancy and hydropower industry data. Through contextual semantic reconstruction and professional terminology calibration, the system matches and generates content that conforms to the actual project and meets industry standards.

6. The method for generating water conservancy and hydropower construction schemes based on construction object feature modeling according to claim 1, characterized in that: In step 3), the combination and generation of the construction plan content also involves a format conversion process, as detailed below: The structured data extracted from general basic information such as project, plan, and construction process includes: title, full text, tables, images, and flowcharts. This data is converted into an intermediate data structure for OnlyOffice. All intermediate data is provided to the OnlyOffice platform in a streaming manner. OnlyOffice parses the intermediate data structure and uses Canvas rendering technology to present all the current content in the collaborative document in a streaming manner.

Images