A method and system for generating a 110kV power transmission and transformation project feasibility study report based on multi-source data fusion

Abstract

The present application relates to a kind of 110kV transmission and transformation engineering feasibility study report generation method and system based on multi-source data fusion, obtain the multi-source input data related to target engineering, based on the semantic analysis and key information extraction of pre-set large language model, the intermediate data generated is filled to the corresponding position according to pre-set chapter template, optimization processing, generate the first draft of 110kV transmission and transformation engineering feasibility study report;System includes data acquisition module, data processing and analysis module configured with large language model, report automatic generation module and optimization output module.The present application realizes the whole process automation processing from multi-source unstructured document to structured data extraction, finally integrates standard report, reduces manual intervention link, data consistency is good, response speed is fast, avoids the information conflict caused by human understanding deviation, ensures that report meets local regulatory requirements, provides scientific and reasonable cost reference basis, has strong transplanting ability.

Description

Technical Field

[0001] This invention relates to the technical field of electrical digital data processing, and in particular to a method and system for generating a feasibility study report for a 110kV power transmission and transformation project based on multi-source data fusion. Background Technology

[0002] Against the backdrop of rapid development in power grid construction, 110kV transmission and transformation projects, as an important component of urban distribution networks, directly impact construction progress and investment returns due to the efficiency and quality of their preliminary work. The feasibility study report (hereinafter referred to as the "feasibility report") is the core document for project approval in this type of project. It covers multiple professional areas, including geological surveys, system access, primary / secondary electrical configuration, environmental protection, and technical and economic analysis. It requires the integration of a large amount of cross-disciplinary documents, including but not limited to geological survey reports, communication project proposals, primary system cost estimates, protection system cost estimates, and regional noise control standards.

[0003] Currently, feasibility study reports are mainly completed manually, which has the following prominent problems: (1) The professional documents are not uniform in format and have different structures. Technical personnel need to read them one by one and manually extract key parameters, which is time-consuming and prone to errors, resulting in low information extraction efficiency. (2) Information transmission between different professions relies on manual communication, which can easily lead to inconsistent or missing data versions and poor data consistency. (3) There are generally differences in the regulations on noise levels and environmental protection requirements of substations in different regions. It is necessary to consult local policy documents, which can easily lead to oversights and lack an automatic identification mechanism. (4) Technical and economic personnel usually rely on experience to select reference projects for investment comparison, lacking objective quantitative support; (5) Although the report template has been solidified, the content filling still depends on the individual understanding of the engineers, resulting in inconsistent output results and insufficient standardization.

[0004] In recent years, natural language processing and large language model technology have developed rapidly, showing great potential in text understanding, information extraction and content generation. In existing technologies, although some research has attempted to use AI for power document parsing, it often focuses on processing a single type of document and has not yet formed a technical system for complex multi-source heterogeneous data fusion and end-to-end intelligent generation of feasibility study reports. Summary of the Invention

[0005] This invention solves the problems existing in the prior art and provides a method and system for generating feasibility study reports for 110kV power transmission and transformation projects based on multi-source data fusion. It can integrate multi-source professional data, realize semantic-level information understanding and structured output, and automatically generate high-quality feasibility study reports to improve the intelligence level of early-stage work in power engineering. It is suitable for power engineering design units to achieve automated report preparation, efficient information extraction and scheme auxiliary decision-making in the early stage of projects.

[0006] The technical solution adopted in this invention is a method for generating a feasibility study report for a 110kV power transmission and transformation project based on multi-source data fusion, which obtains multi-source input data related to the target project; Based on a pre-defined large language model, semantic parsing and key information extraction are performed on the multi-source input data to generate intermediate data; the intermediate data includes structured geological exploration information, standardized investment data, investment comparison analysis results, preliminary design schemes, and regional standard identification results; According to the preset chapter template, the intermediate data is filled into the corresponding positions of the chapter template and optimized. A draft of the feasibility study report for the 110kV power transmission and transformation project was generated.

[0007] Preferably, the multi-source input data includes geological survey report documents; A preset set of query conditions is used to perform semantic understanding on the geological survey report document using a preset large language model. The relevant text paragraphs containing the target fields in the set of query conditions are located. Entity recognition and relation extraction are performed on the located text paragraphs to extract key parameter values ​​corresponding to the target fields. A corresponding confidence score is generated for the key parameter values ​​to quantify the reliability of the matching between the extracted key parameter values ​​and the target fields. The output includes a geological information table containing structured geological exploration information.

[0008] Preferably, the query condition set includes at least one target field among topographic features, groundwater level, soil resistivity, seismic intensity, and adverse geological phenomena.

[0009] Preferably, the multi-source input data also includes data submission documents, and corresponding parsing templates and information extraction rule sets are configured for different types of data submission documents; Based on a pre-defined large language model, the document submission is semantically parsed according to type, key data items in the document are identified, and the key data items are mapped to a unified standardized document submission data model according to field semantics, data type and engineering meaning, and a document submission database is built on this basis. Based on the aforementioned capital contribution database, the standardized capital contribution data is processed to generate a comprehensive capital contribution order that includes the standardized capital contribution data.

[0010] Preferably, key data items include main transformer capacity, outgoing line scale, communication channel requirements, and protection configuration principles.

[0011] Preferably, the multi-source input data further includes basic project parameters; Obtain a historical project database containing historical project metadata, preprocess the historical project metadata, combine the structured or semi-structured historical project metadata according to field semantics to form a project feature description, and input the project feature description into a preset large language model for semantic encoding to obtain the corresponding historical project semantic vector; input the basic parameters of the target project into the preset large language model and generate the target project semantic vector using the same semantic encoding method. Calculate the semantic similarity between the semantic vector of the target project and the semantic vectors of each historical project, and select historical projects with semantic similarity higher than a preset threshold as the target project reference set; A comparative analysis is performed on the investment component indicators in the target project reference set to generate an investment comparison analysis table including the results of the investment comparison analysis and investment estimation suggestions for decision support.

[0012] Preferably, the multi-source input data also includes basic project information, which aggregates structured geological exploration information, standardized financial data, investment comparison analysis results and basic project information according to a preset data association relationship to form a unified engineering semantic dataset; Based on the engineering semantic dataset, a preliminary design scheme is generated by using a pre-set large language model combined with a pre-set knowledge graph for reasoning and control constraints.

[0013] Preferably, the multi-source input data also includes project geographical location information. Based on the project geographical location information, corresponding regional regulatory documents are obtained, and the regional regulatory documents are semantically parsed using a preset large language model to identify the rules and standards applicable to the target project. Based on the identification results, a descriptive document is generated as the identification result of the regional standard.

[0014] Preferably, based on the preset chapter template, the intermediate data is subjected to chapter semantic matching and structural mapping, and different types of intermediate data are automatically allocated to the corresponding positions according to the chapter functional requirements. During the filling process, the content structure, information granularity and logical order are adaptively adjusted and optimized based on chapter semantic constraints, professional expression norms and context consistency rules. The source and confidence information of the generated content are associated in the generated draft.

[0015] A system for generating feasibility study reports for 110kV power transmission and transformation projects based on multi-source data fusion, comprising: The data acquisition module is used to acquire multi-source input data related to the target project; The data processing and analysis module is equipped with a preset large language model for processing and analyzing multi-source input data to obtain structured intermediate data. The automatic report generation module is used to populate the corresponding chapters with intermediate data to obtain the filled report text; The optimized output module is used to call a preset large language model to optimize the filled report text and output the first draft of the feasibility study report for the 110kV power transmission and transformation project.

[0016] This invention relates to a method and system for generating a feasibility study report for a 110kV power transmission and transformation project based on multi-source data fusion. The method involves acquiring multi-source input data related to the target project, performing semantic parsing and key information extraction based on a preset large language model to generate intermediate data, filling the corresponding positions in the preset chapter templates with the intermediate data, and performing optimization processing to generate a draft of the 110kV power transmission and transformation project feasibility study report. The system includes a data acquisition module, a data processing and analysis module configured with a preset large language model, an automatic report generation module, and an optimization output module.

[0017] The beneficial effects of this invention are as follows: (1) It realizes the full-process automated processing of extracting structured data from multiple sources of unstructured documents and finally integrating it into standard reports, reducing manual intervention by more than 70%; (2) All data originates from a unified semantic parsing engine, avoiding information conflicts caused by human misunderstandings, and ensuring good data consistency; (3) The time for generating a draft report for a single project has been reduced from an average of 8 hours to within 1 hour, resulting in a faster response time; (4) By utilizing the contextual understanding capabilities of the large language model, key information can be accurately identified and extracted from multi-source unstructured documents, thereby generating structured data, avoiding information conflicts caused by human misunderstanding, and improving data consistency; the accuracy rate of key information extraction can reach over 92%; this capability runs through the entire process from original documents to the generation of the first draft report, and provides scientific basis in the historical project matching and cost reference stages; (5) It can automatically identify local technical regulations to ensure that the report complies with local regulatory requirements; (6) Provide scientific and reasonable cost references by matching historical projects through semantic similarity; (7) The methodology can be extended to 220kV and above power transmission and transformation projects or other infrastructure fields, and has strong adaptability. Attached Figure Description

[0018] Figure 1This is a flowchart of the method of the present invention; Figure 2 This is a schematic diagram of the system structure of the present invention. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] This invention relates to a method for generating feasibility study reports for 110kV power transmission and transformation projects based on multi-source data fusion. The method aims to achieve intelligent understanding, extraction, and fusion of multi-source heterogeneous engineering data through large language model technology, and automatically generate high-quality, standardized draft feasibility study reports, thereby significantly improving the efficiency, accuracy, and consistency of report preparation.

[0021] The method includes the following steps: (1) Obtain multi-source input data related to the target project; (2) Based on the preset large language model, semantic parsing and key information extraction are performed on the multi-source input data to generate intermediate data; the intermediate data includes structured geological exploration information, standardized investment data, investment comparison analysis results, preliminary design schemes and regional standard identification results; (3) According to the preset chapter template, fill the intermediate data into the corresponding position of the chapter template and perform optimization processing; (4) Generate the first draft of the feasibility study report for the 110kV power transmission and transformation project.

[0022] (1) Obtain multi-source input data related to the target project; In this invention, multi-source input data includes, but is not limited to: Geological exploration report documents are generally in electronic formats, such as PDF and Word. Multiple professional documents, including communication project proposal, system initial proposal, relay protection professional proposal, preliminary civil engineering opinions, etc. The basic parameters of the current project include, but are not limited to, the number of main transformers and total capacity; Project geographical location information, such as geographical coordinates or administrative division code; Basic project information, such as substation site selection area, power load forecast, and upstream power source.

[0023] Generally, the above data can be uploaded in batches via an interface or automatically extracted from the project management system.

[0024] (2) Based on the preset large language model, semantic parsing and key information extraction are performed on the multi-source input data to generate intermediate data; the intermediate data includes structured geological exploration information, standardized investment data, investment comparison analysis results, preliminary design schemes and regional standard identification results; In this invention, a large language model (LLM) drives multiple parallel processing sub-processes. The preset large language model here is a special model that has been fine-tuned and optimized on a large amount of professional language data such as power engineering design specifications, technical reports, and project documents, so that it has a deep understanding of professional terms, common expressions and logical structures in the power industry.

[0025] The preset large language model is not a direct call to a general model, but rather a targeted adjustment and optimization at the underlying inference and output control levels of the model, specifically for the business scenario of 110kV power transmission and transformation project feasibility studies. This includes: A professional semantic constraint layer for the field of power transmission and transformation engineering is constructed. In the model decoding stage, a special vocabulary of engineering fields, priority weight of professional fields and technical expression constraint rules are introduced to dynamically guide the model generation process, so that the model prioritizes the power engineering professional terminology system and industry expression standards in the semantic generation process. A multi-source data consistency verification submodule is embedded in the inference chain within the model. Semantic consistency comparison is performed on fields or related parameters with the same name from different document sources. When a conflict or inconsistency is detected, a conflict resolution strategy is triggered to rearrange or recalculate the inference path, thereby suppressing the propagation of unreliable inference results. In response to the chapter structure characteristics of the feasibility study report, the intermediate representation layer of the large language model is subject to chapter semantic conditional control. The inference state of the model is constrained by the chapter identifier vector, so that the model adopts differentiated information selection strategy and generation granularity control strategy in different chapters. By introducing an interpretable reasoning tagging mechanism, key generated conclusions are bound to their corresponding data source nodes and reasoning confidence weights within the model, enabling a traceable mapping between the model output and the underlying semantic reasoning process.

[0026] (2-1) Structured Geological Information The multi-source input data includes geological survey report documents; A preset set of query conditions is used to perform semantic understanding on the geological survey report document using a preset large language model. The relevant text paragraphs containing the target fields in the set of query conditions are located. Entity recognition and relation extraction are performed on the located text paragraphs to extract key parameter values ​​corresponding to the target fields. A corresponding confidence score is generated for the key parameter values ​​to quantify the reliability of the matching between the extracted key parameter values ​​and the target fields. The output includes a geological information table containing structured geological exploration information.

[0027] The query criteria set includes at least one target field among topographic features, groundwater level, soil resistivity, seismic intensity, and adverse geological phenomena.

[0028] In this invention, a set of query conditions is preset, including target fields such as topographic features, groundwater level, soil resistivity, seismic intensity, and adverse geological phenomena. Adverse geological phenomena mainly include collapses, landslides, debris flows, karst, soil caves, river erosion, and seepage deformation. The full text of the geological survey report is input into a preset large language model. Based on the understanding of the document's semantics, the model locates and highlights relevant text paragraphs containing the above target fields. Entity recognition and relation extraction operations are performed on the located text paragraphs to accurately extract the key parameter values ​​corresponding to each target field, such as "groundwater level: -2.5m". In the specific implementation process, the confidence score can be determined comprehensively based on factors such as the semantic explicitness of the target field in the text paragraph, contextual consistency, consistency verification of extraction results at different text positions, and the degree of matching between the model output results and the rule verification results. The confidence score is generally a number between 0 and 1. Finally, a structured geological information table is output, which includes field names, extracted values, source locations, and confidence levels (confidence scores).

[0029] (2-2) Standardized data submission The multi-source input data also includes data submission documents, and corresponding parsing templates and information extraction rule sets are configured for different types of data submission documents; Based on a pre-defined large language model, the document submission is semantically parsed according to type, key data items in the document are identified, and the key data items are mapped to a unified standardized document submission data model according to field semantics, data type and engineering meaning, and a document submission database is built on this basis. Based on the aforementioned capital contribution database, the standardized capital contribution data is processed to generate a comprehensive capital contribution order that includes the standardized capital contribution data.

[0030] Key data items include main transformer capacity, outgoing line scale, communication channel requirements, and protection configuration principles.

[0031] In this invention, corresponding parsing templates and information extraction rule sets are pre-configured for different types of information submission documents. The information extraction rule sets are used to constrain and verify the semantic parsing results, including at least one of field triggering rules, position constraint rules, numerical rationality verification rules, and conflict resolution rules, so as to improve the stability and consistency of the information extraction results.

[0032] The system invokes a pre-defined large language model to perform semantic parsing on various types of documents, identify and extract key data items, such as main transformer capacity "2×50MVA", outgoing line scale such as "4 outgoing lines of 110kV", communication channel requirements, protection configuration principles, etc. The extracted results are mapped to a unified, well-defined, standardized data model according to field semantics, data type, and engineering meaning, thereby constructing a structured professional data database. The data database is used to realize the unified expression, consistency verification, and cross-professional reuse of multi-source data. Based on this database, the system can automatically generate a "Comprehensive Data Submission Form for 110kV Transmission and Transformation Project" that conforms to the standard format by summarizing, filtering and combining the standardized data submissions, and can export it to Excel or PDF.

[0033] (2-3) Investment Comparison Analysis Results The multi-source input data also includes basic project parameters; Obtain a historical project database containing historical project metadata, preprocess the historical project metadata, combine the structured or semi-structured historical project metadata according to field semantics to form a project feature description, and input the project feature description into a preset large language model for semantic encoding to obtain the corresponding historical project semantic vector; input the basic parameters of the target project into the preset large language model and generate the target project semantic vector using the same semantic encoding method. Calculate the semantic similarity between the semantic vector of the target project and the semantic vectors of each historical project, and select historical projects with semantic similarity higher than a preset threshold as the target project reference set; A comparative analysis is performed on the investment component indicators in the target project reference set to generate an investment comparison analysis table including the results of the investment comparison analysis and investment estimation suggestions for decision support.

[0034] In this invention, the system maintains a historical project database, which stores metadata of at least one completed 110kV transmission and transformation project, including structured or semi-structured data such as project scale, voltage level, site conditions, main equipment configuration, investment composition indicators and technical and economic parameters. The basic parameters of the current project are input into the large language model and encoded into a high-dimensional semantic vector, namely the target project semantic vector. At the same time, the semantic vectors of historical projects are obtained. The weighted cosine similarity algorithm is used to calculate the semantic similarity between the current project vector and the vectors of each historical project. The calculation comprehensively considers the project scale parameters, engineering type characteristics, main equipment configuration characteristics and investment composition characteristics, and is used to characterize the similarity between the overall engineering attributes of the project. The weights can be dynamically configured according to key factors such as main transformer capacity and location. Historical projects with similarity higher than a preset threshold, such as 0.8, are selected as target projects. Subsequently, at least one of the following components of these target projects—equipment purchase cost, construction and installation cost, other expenses, and contingency fund—is compared and analyzed as an investment composition indicator. Based on the historical investment data distribution of the target project reference set, the investment composition of the target project is compared and analyzed across intervals to generate an investment comparison analysis table including the investment comparison analysis results. Based on the investment level range, unit capacity investment index, and key cost influencing factors of the target project set, investment estimation suggestions for decision support are generated.

[0035] (2-4) Preliminary Design Scheme The multi-source input data also includes basic project information. Structured geological exploration information, standardized financial data, investment comparison analysis results and basic project information are aggregated according to preset data association relationships to form a unified engineering semantic dataset. Based on the engineering semantic dataset, a preliminary design scheme is generated by using a pre-set large language model combined with a pre-set knowledge graph for reasoning and control constraints.

[0036] In this invention, the basic information of the project includes the project scale, construction location, voltage level, and construction conditions, etc. Maintain a knowledge graph that includes at least one of the following: electrical primary system topology, electrical secondary configuration rules, civil engineering and structural constraints, equipment selection rules, and regional specification requirements; The structured geological exploration information, standardized data, technical and economic analysis results extracted in the aforementioned steps are aggregated and integrated with the basic project information input by the user. The aggregation includes semantic-based association mapping of fields, cross-professional parameter references and consistency verification, which is used to realize cross-utilization and collaborative constraints between different data sources. By utilizing a pre-defined large language model and a pre-defined knowledge graph for reasoning, and based on all input information and the constraints of the knowledge graph, a preliminary technically compliant solution text can be generated. The content covers core design solutions such as electrical main wiring suggestions, main transformer selection recommendations, power distribution device layout, grounding system design points, and lightning protection measures. At the same time, through rule constraints, parameter dependencies, and typical design solution templates in the knowledge graph, the generated content of the model is constrained and controlled to ensure that the preliminary design solution meets the pre-defined technical compliance requirements in terms of electrical configuration, equipment selection, and structural conditions.

[0037] In practice, the process supports interaction, allowing users to provide feedback on the generated solution, which in turn enables iterative optimization of the model.

[0038] (2-5) Regional Standard Identification Results The multi-source input data also includes project geographical location information. Based on the project geographical location information, corresponding regional regulatory documents are obtained. The regional regulatory documents are semantically parsed using a preset large language model to identify the rules and standards applicable to the target project. Based on the identification results, a descriptive document is generated as the identification result of the regional standard.

[0039] In this invention, based on the project's geographical location information, regional regulatory documents such as acoustic environment functional zoning and electromagnetic radiation control standards published in the project location are automatically associated and retrieved through geocoding. These regulatory documents are semantically parsed using a preset large language model to accurately identify the noise limit standards applicable to the project, such as daytime and nighttime A-weighted sound level requirements, protection distance requirements, and environmental impact assessment scope. Based on the identification results, normative chapters such as "Noise Control Requirements Description for Project Location" are automatically generated.

[0040] In the specific implementation process, if there are multiple applicable standards, the system calculates and provides priority ranking and selection suggestions for the optimal standard based on the matching degree of the applicable scope and historical project experience.

[0041] (3) According to the preset chapter template, fill the intermediate data into the corresponding position of the chapter template and perform optimization processing; Specifically, based on the preset chapter template, the intermediate data is semantically matched and structurally mapped. Different types of intermediate data are automatically assigned to the corresponding positions according to the chapter's functional requirements. During the filling process, the content structure, information granularity, and logical order are adaptively adjusted and optimized based on chapter semantic constraints, professional expression standards, and contextual consistency rules. The source and confidence information of the generated content are associated in the generated draft.

[0042] In this invention, a standard feasibility study report chapter template is configured, including chapters such as overview, necessity of construction, system planning, site selection, electrical components, civil engineering components, environmental protection and energy conservation, investment estimation, and conclusions and recommendations. All generated structured intermediate data, including geological exploration information tables, comprehensive investment proposals, investment comparison analysis tables, preliminary scheme texts, noise control instructions, etc., are automatically filled into the corresponding chapter positions of the template according to preset mapping rules.

[0043] After the filling is completed, the preset large language model is invoked to polish and optimize the coherence of the entire report text, including eliminating mechanical splicing traces, unifying terminology, optimizing paragraph connections, and correcting grammar, so that the report text is fluent and natural.

[0044] Outputs an optimized and fully formatted draft of the "Feasibility Study Report for 110kV Transmission and Transformation Project", supporting export in multiple formats such as Word and PDF.

[0045] Furthermore, the output text provides revision mode tags, allowing all automatically generated content to be tagged and its source data and confidence information to be displayed in conjunction with the data, facilitating manual review and verification.

[0046] (4) Generate the first draft of the feasibility study report for the 110kV power transmission and transformation project.

[0047] Based on the semantic parsing, structured processing, and chapter-level content mapping of multi-source professional data, and through consistency control and professional standard constraints, the initial draft of the feasibility study report for the 110kV power transmission and transformation project is automatically generated, which has complete engineering logic, reusable data foundation, and decision support capabilities.

[0048] This invention also relates to a system for generating feasibility study reports for 110kV power transmission and transformation projects based on multi-source data fusion, comprising: The data acquisition module is used to acquire multi-source input data related to the target project. The multi-source input data here is data related to the target project from various sources, such as user uploads and database interface collection. It includes at least electronic documents of geological survey reports, multi-disciplinary data provision documents, current project basic parameters, project geographical location information and project basic information. The data processing and analysis module is equipped with a preset large language model for processing and analyzing multi-source input data to obtain structured intermediate data. In this embodiment, the module includes multiple sub-processing units, which respectively perform geological survey report parsing, data provision document standardization, technical and economic comparison analysis, intelligent scheme generation, and regional standard identification, and finally output structured intermediate data; Specifically, the module performs semantic parsing on geological exploration report documents to extract structured geological exploration information; The submitted documents are parsed and standardized to generate standardized submitted data; Based on the project's basic parameters and historical project database, an investment comparison analysis is conducted to generate investment comparison analysis results. Based on the project's geographical location information, identify and extract applicable regional regulatory requirements; By aggregating structured geological exploration information, standardized data, investment comparison analysis results, regional standard identification results, and basic project information, a preliminary design scheme is generated. In practice, the module is also equipped with a historical project database and knowledge graph to support analysis and decision-making.

[0049] The automatic report generation module is used to fill intermediate data into the corresponding chapters (templates) to obtain the filled report text; specifically, it automatically fills the structured geological exploration information, standardized financial data, investment comparison analysis results, preliminary design schemes, and regional standard identification results into the corresponding chapters of the template. The output optimization module is used to call a preset large language model to optimize the filled report text, including text polishing, coherence optimization and formatting, and output the first draft of the feasibility study report for the 110kV power transmission and transformation project. The first draft can be exported as a Word or PDF document, and the revision mode has the function of tracing the source.

[0050] In practical applications, the entire system is deployed on the enterprise's private server or private cloud environment, and all data processing is completed on the internal network to ensure that sensitive engineering data is not leaked and to meet the enterprise's data security and compliance requirements.

[0051] The following is a specific embodiment of the technical implementation process of the present invention.

[0052] The goal is to prepare a feasibility study report for the "110kV substation project in Zhangjiang Science City, Pudong New Area, Shanghai". (a) Data preparation stage Upload one PDF geological survey report to the data acquisition module, import five Word documents including the communication project proposal, system initial data submission form, and protection professional data submission, enter the target center coordinates (longitude 121.78°E, latitude 31.20°N), and fill in the basic parameters: 2×50MVA main transformers in this phase, 3 units in the future, and double busbar connection.

[0053] (II) System Processing Flow The data processing and analysis module extracted the following: (1) Key parameters such as “Site Category III”, “Peak ground acceleration 0.10g”, and “Soil resistivity approximately 85Ω·m”; (2) "Two fiber optic channels need to be opened", "Dual bus differential protection needs to be configured", and "No less than 30 meters of space needs to be reserved for expansion"; (3) It was found that the similarity with "Minhang Huacao 110kV Substation" was 0.86, and the recommended unit capacity construction cost range was RMB11.5 million to RMB12.8 million / MVA; (4) It was found that Zhangjiang area belongs to Class 3 sound environment zone, with nighttime noise limit ≤55dB(A), and noise reduction measures suggestions were automatically generated; (5) Based on the above information, it is recommended to adopt a fully indoor prefabricated structure, a three-phase double-winding self-cooled main transformer, and GIS combined electrical appliances arranged on the second-floor platform.

[0054] (iii) Report pre-assembly The automatic report generation module fills in all the content into the template, resulting in the filled-in report text.

[0055] (iv) Deliverables The output optimization module was used to optimize the filled report text, generating a preliminary feasibility study report of approximately 40 pages. Output PDF report, structured investment proposal form, and investment comparison analysis attachments.

[0056] Once completed, engineers only need to review the key technical points and economic indicators before submitting it for review, which greatly saves labor costs.

[0057] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0058] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0059] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0060] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0061] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including both the preferred embodiments and all changes and modifications falling within the scope of the invention.

[0062] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A method for generating a feasibility study report for a 110kV power transmission and transformation project based on multi-source data fusion, characterized in that: Acquire multi-source input data related to the target project; Based on a pre-defined large language model, semantic parsing and key information extraction are performed on the multi-source input data to generate intermediate data; the intermediate data includes structured geological exploration information, standardized investment data, investment comparison analysis results, preliminary design schemes, and regional standard identification results; According to the preset chapter template, the intermediate data is filled into the corresponding positions of the chapter template and optimized. A draft of the feasibility study report for the 110kV power transmission and transformation project was generated.

2. The method for generating a feasibility study report for a 110kV power transmission and transformation project based on multi-source data fusion as described in claim 1, characterized in that: The multi-source input data includes geological survey report documents; A preset set of query conditions is used to perform semantic understanding on the geological survey report document using a preset large language model. The relevant text paragraphs containing the target fields in the set of query conditions are located. Entity recognition and relation extraction are performed on the located text paragraphs to extract key parameter values ​​corresponding to the target fields. A corresponding confidence score is generated for the key parameter values ​​to quantify the reliability of the matching between the extracted key parameter values ​​and the target fields. The output includes a geological information table containing structured geological exploration information.

3. The method for generating a feasibility study report for a 110kV power transmission and transformation project based on multi-source data fusion as described in claim 2, characterized in that: The query criteria set includes at least one target field among topographic features, groundwater level, soil resistivity, seismic intensity, and adverse geological phenomena.

4. The method for generating a feasibility study report for a 110kV power transmission and transformation project based on multi-source data fusion as described in claim 1, characterized in that: The multi-source input data also includes data submission documents, and corresponding parsing templates and information extraction rule sets are configured for different types of data submission documents; Based on a pre-defined large language model, the document submission is semantically parsed according to type, key data items in the document are identified, and the key data items are mapped to a unified standardized document submission data model according to field semantics, data type and engineering meaning, and a document submission database is built on this basis. Based on the aforementioned capital contribution database, the standardized capital contribution data is processed to generate a comprehensive capital contribution order that includes the standardized capital contribution data.

5. The method for generating a feasibility study report for a 110kV power transmission and transformation project based on multi-source data fusion as described in claim 4, characterized in that: Key data items include main transformer capacity, outgoing line scale, communication channel requirements, and protection configuration principles.

6. The method for generating a feasibility study report for a 110kV power transmission and transformation project based on multi-source data fusion as described in claim 1, characterized in that: The multi-source input data also includes basic project parameters; Obtain a historical project database containing historical project metadata, preprocess the historical project metadata, combine the structured or semi-structured historical project metadata according to field semantics to form a project feature description, and input the project feature description into a preset large language model for semantic encoding to obtain the corresponding historical project semantic vector; input the basic parameters of the target project into the preset large language model and generate the target project semantic vector using the same semantic encoding method. Calculate the semantic similarity between the semantic vector of the target project and the semantic vectors of each historical project, and select historical projects with semantic similarity higher than a preset threshold as the target project reference set; A comparative analysis is performed on the investment component indicators in the target project reference set to generate an investment comparison analysis table including the results of the investment comparison analysis and investment estimation suggestions for decision support.

7. The method for generating a feasibility study report for a 110kV power transmission and transformation project based on multi-source data fusion as described in claim 1, characterized in that: The multi-source input data also includes basic project information. Structured geological exploration information, standardized financial data, investment comparison analysis results and basic project information are aggregated according to preset data association relationships to form a unified engineering semantic dataset. Based on the engineering semantic dataset, a preliminary design scheme is generated by using a pre-set large language model combined with a pre-set knowledge graph for reasoning and control constraints.

8. The method for generating a feasibility study report for a 110kV power transmission and transformation project based on multi-source data fusion as described in claim 1, characterized in that: The multi-source input data also includes project geographical location information. Based on the project geographical location information, corresponding regional regulatory documents are obtained. The regional regulatory documents are semantically parsed using a preset large language model to identify the rules and standards applicable to the target project. Based on the identification results, a descriptive document is generated as the identification result of the regional standard.

9. The method for generating a feasibility study report for a 110kV power transmission and transformation project based on multi-source data fusion as described in claim 1, characterized in that: Based on the preset chapter template, the intermediate data is semantically matched and structurally mapped. Different types of intermediate data are automatically assigned to the corresponding positions according to the functional requirements of the chapter. During the filling process, the content structure, information granularity and logical order are adaptively adjusted and optimized based on chapter semantic constraints, professional expression standards and contextual consistency rules. The source and confidence information of the generated content are associated in the generated draft.

10. A system for generating feasibility study reports for 110kV power transmission and transformation projects based on multi-source data fusion, characterized in that: include: The data acquisition module is used to acquire multi-source input data related to the target project; The data processing and analysis module is equipped with a preset large language model for processing and analyzing multi-source input data to obtain structured intermediate data. The automatic report generation module is used to populate the corresponding chapters with intermediate data to obtain the filled report text; The optimized output module is used to call a preset large language model to optimize the filled report text and output the first draft of the feasibility study report for the 110kV power transmission and transformation project.

Images