Power transmission and transformation engineering cost data analysis method and system based on large language model

Abstract

The application provides a power transmission and transformation engineering cost data analysis method and system based on a large language model, which comprises the following steps: obtaining multi-source data of power transmission and transformation engineering cost; configuring corresponding structured data index items in advance based on the file type of the cost file; dividing the cost document into multiple text segments, combining each text segment with the corresponding type of structured data index item to construct a prompt word template, inputting the prompt word template into the large language model to identify and extract structured extraction data; performing data processing on the structured extraction data and the structured data to obtain fused same project data assets, and storing the same project data assets in a relational database; receiving an analysis request of a user, analyzing the analysis request by using the large language model, and converting the analysis request into a database query instruction; calling a preset data analysis model to analyze the project data assets according to the database query instruction, and obtaining analysis result data.

Description

Technical Field

[0001] This application relates to the field of data processing technology, and in particular to a method and system for analyzing the cost data of power transmission and transformation projects based on a large language model. Background Technology

[0002] Power grid transmission and transformation projects involve large-scale investments and long construction periods, generating massive amounts of structured and unstructured data during their cost management. How to effectively manage and deeply analyze this data to achieve precise cost control and provide scientific support for investment decisions is a core challenge facing power grid companies.

[0003] In existing technologies, data extraction from unstructured cost documents mainly relies on manual review and entry, which is inefficient and inconsistent. At the data analysis level, pre-defined dimensions and indicators are primarily used in business intelligence dashboards, requiring technical personnel to perform extensive data modeling and report development beforehand, preventing business personnel from conducting flexible and spontaneous queries. This traditional approach has the following drawbacks: fixed and inflexible analysis dimensions, failing to respond to temporary and personalized natural language analysis needs from business personnel; analytical capabilities remain at the descriptive statistical level, making it difficult to automatically uncover hidden patterns and anomalies in the data, resulting in delayed risk warnings; and the data management platform is independent of upper-level intelligent analysis applications, preventing real-time updates to the analysis model and causing delays in analysis results, creating a barrier between data management and business applications. Summary of the Invention

[0004] This application provides a method and system for analyzing power transmission and transformation engineering cost data based on a large language model, in order to at least address the shortcomings of the aforementioned related technologies.

[0005] In a first aspect, embodiments of this application provide a method for analyzing power transmission and transformation engineering cost data based on a large language model, comprising the following steps: Step 1: Obtain multi-source data on the cost of power transmission and transformation projects, wherein the multi-source data includes structured data and cost documents containing unstructured text; Step 2: Based on the file type of the cost document, pre-configure the corresponding structured data indicator items, wherein the structured data indicator items define the data fields to be extracted from cost documents of different file types; Step 3: Divide the cost document into multiple text segments, combine each text segment with the corresponding type of structured data indicator to construct a prompt word template, and input the prompt word template into the large language model to identify and extract structured data. Step 4: Process the extracted structured data with the structured data to obtain the merged data assets of the same project, and store them in a relational database; Step 5: Receive the user's analysis request, and use the large language model to parse the analysis request and convert it into a database query instruction; Step Six: Based on the database query command, invoke the preset data analysis model to analyze the project data assets and obtain the analysis results data.

[0006] Furthermore, step five includes: The analysis request is semantically parsed using a large language model to obtain the corresponding analysis dimensions and analysis intent; The analytical dimensions and analytical intent are converted into standard query statements and invocation instructions corresponding to the relational database.

[0007] Furthermore, step six includes: The corresponding integrated data assets are retrieved from the relational database according to the standard query statement. The call instruction invokes and executes at least one embedded data analysis model to analyze and calculate the relevant fused data assets, thereby obtaining structured analysis results data. The data analysis model includes at least one of a control line analysis model, a material cost analysis model, and a historical engineering comparison model.

[0008] Furthermore, the control line analysis model is used to automatically match the cost characteristics of the current project with the preset cost control line standard, calculate the deviation value between the actual cost and the cost control line standard, and generate early warning information when the deviation value exceeds the preset threshold. The material cost analysis model is used to perform time-series trend analysis, inter-supplier price comparison analysis, and quantity-price-cost composition analysis of material prices by using equipment and material parameters extracted from cost documents. The historical project comparison model is used to automatically filter out similar historical projects from the historical project database that have a similarity to the current project that is higher than a threshold through clustering or similarity algorithms, and compares and analyzes the cost indicators of the current project with the corresponding cost indicators of the similar historical projects to generate a difference report.

[0009] Furthermore, the method also includes: The analysis results data are converted into natural language analysis text using the large language model, and the natural language analysis text is pushed to the user interface in the form of a dialog interface. Based on the analysis results, a visual chart is dynamically generated and embedded into the corresponding position of the dialog interface.

[0010] Secondly, this invention also proposes a power transmission and transformation engineering cost data analysis system based on a large language model, comprising: The data acquisition module is used to acquire multi-source data on the cost of power transmission and transformation projects, wherein the multi-source data includes structured data and cost documents containing unstructured text; The data processing module is used to pre-configure corresponding structured data indicator items based on the file type of the cost document, wherein the structured data indicator items define the data fields to be extracted from cost documents of different file types; The text processing module is used to divide the cost document into multiple text fragments, combine each text fragment with a corresponding type of structured data indicator item to construct a prompt word template, and input the prompt word template into a large language model to identify and extract structured data. The data fusion module is used to process the structured extracted data and the structured data to obtain the fused data assets of the same project, and store them in a relational database. The request parsing module is used to receive the user's analysis request, parse the analysis request using the large language model, and convert it into a database query instruction; The data analysis module is used to call a preset data analysis model according to the database query command to analyze the project data assets and obtain the analysis result data.

[0011] Furthermore, the request parsing module is specifically used for: The analysis request is semantically parsed using a large language model to obtain the corresponding analysis dimensions and analysis intent; The analytical dimensions and analytical intent are converted into standard query statements and invocation instructions corresponding to the relational database.

[0012] Furthermore, the data analysis module is specifically used for: The corresponding integrated data assets are retrieved from the relational database according to the standard query statement. The call instruction invokes and executes at least one embedded data analysis model to analyze and calculate the relevant fused data assets, thereby obtaining structured analysis results data. The data analysis model includes at least one of a control line analysis model, a material cost analysis model, and a historical engineering comparison model.

[0013] Furthermore, in the data analysis module, the control line analysis model is used to automatically match the cost characteristics of the current project with the preset cost control line standard, calculate the deviation value between the actual cost and the cost control line standard, and generate early warning information when the deviation value exceeds the preset threshold. The material cost analysis model is used to perform time-series trend analysis, inter-supplier price comparison analysis, and quantity-price-cost composition analysis of material prices by using equipment and material parameters extracted from cost documents. The historical project comparison model is used to automatically filter out similar historical projects from the historical project database that have a similarity to the current project that is higher than a threshold through clustering or similarity algorithms, and compares and analyzes the cost indicators of the current project with the corresponding cost indicators of the similar historical projects to generate a difference report.

[0014] Furthermore, the system also includes: The data conversion module is used to convert the analysis result data into natural language analysis text using the large language model, and push the natural language analysis text to the user interface in the form of a dialog interface. The data visualization module is used to dynamically generate visualization charts based on the analysis results and embed the visualization charts into the corresponding positions of the dialog interface.

[0015] Compared to related technologies, the present application provides a method and system for analyzing power transmission and transformation engineering cost data based on a large language model. By combining a large language model with configurable indicators and prompt word engineering, it achieves automated extraction and governance of unstructured cost data, significantly improving data utilization and processing efficiency. Through a natural language interaction engine, it parses user analysis intent and drives embedded professional analysis models, breaking the limitations of traditional fixed reports and enabling business personnel to conduct flexible and in-depth data insights without technical background. Simultaneously, real-time linkage between data management and intelligent analysis realizes a shift from superficial statistics to in-depth mining, enhancing the early warning capability of cost risks and effectively supporting lean cost control in power grid transmission and transformation engineering.

[0016] Details of one or more embodiments of this application are set forth in the following drawings and description to make other features, objects and advantages of this application more readily apparent. Attached Figure Description

[0017] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings: Figure 1 This is a flowchart of the power transmission and transformation engineering cost data analysis method based on a large language model in the first embodiment of the present invention; Figure 2 This is a structural block diagram of the power transmission and transformation engineering cost data analysis system based on a large language model according to the second embodiment of the present invention.

[0018] The following detailed description, in conjunction with the accompanying drawings, will further illustrate the present invention. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of this application clearer, the application is described and illustrated below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application. All other embodiments obtained by those skilled in the art based on the embodiments provided in this application without inventive effort are within the scope of protection of this application.

[0020] Obviously, the accompanying drawings described below are merely some examples or embodiments of this application. Those skilled in the art can apply this application to other similar scenarios based on these drawings without any inventive effort. Furthermore, it is understood that although the efforts made in this development process may be complex and lengthy, for those skilled in the art related to the content disclosed in this application, any changes to design, manufacturing, or production based on the technical content disclosed in this application are merely conventional technical means and should not be construed as insufficient disclosure of the content of this application.

[0021] In this application, the reference to "embodiment" means that a specific feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places in the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment that is mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described in this application may be combined with other embodiments without conflict.

[0022] Unless otherwise defined, the technical or scientific terms used in this application shall have the ordinary meaning understood by one of ordinary skill in the art to which this application pertains. The terms “a,” “an,” “an,” “the,” and similar words used in this application do not indicate quantity limitation and may represent singular or plural. The terms “comprising,” “including,” “having,” and any variations thereof used in this application are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or device that includes a series of steps or modules (units) is not limited to the listed steps or units, but may also include steps or units not listed, or may include other steps or units inherent to such processes, methods, products, or devices. The terms “connected,” “linked,” “coupled,” and similar words used in this application are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. Example 1

[0023] Please see Figure 1 The figure shows a method for analyzing power transmission and transformation engineering cost data based on a large language model in the first embodiment of the present invention. The method specifically includes steps S101 to S106: S101, Obtain multi-source data on the cost of power transmission and transformation projects, wherein the multi-source data includes structured data and cost documents containing unstructured text; In practical implementation, cost-related data for power transmission and transformation projects throughout their entire lifecycle—from planning, feasibility study, preliminary design, construction to final settlement—is collected. This data is divided into two categories. The first category is structured data, primarily sourced from existing business information systems such as Enterprise Resource Planning (ERP), Project Management System (PMS), and Financial Management System. It typically exists in the form of relational database tables, containing clearly defined fields such as: project code, project name, voltage level, project type, construction unit, contract number, cost item code, budget amount, and settlement amount. Access methods include direct JDBC / ODBC database connection, RESTful API scheduled synchronization, or message queue streaming access.

[0024] The second category consists of cost documents containing unstructured text. These are stored as binary files in a document management system (DMS) or distributed file system, with formats including PDF, Word, Excel, plain text files, and scanned images. Specific file types include, but are not limited to: feasibility study investment estimation reports, preliminary design budgets, construction drawing budgets, bills of quantities and tender control price documents, final settlement audit reports, various engineering change orders, equipment and material procurement contracts and invoices, etc. Key cost information within these documents, such as the total estimated cost, equipment unit prices, quantities, and detailed cost breakdowns, exists in a mixed format of natural language text, paragraphs, tables, or key-value pairs, making them difficult to parse directly using traditional tools.

[0025] S102, based on the file type of the cost document, pre-configure corresponding structured data indicator items, wherein the structured data indicator items define the data fields to be extracted from cost documents of different file types; In practice, a set of configurable structured data indicator templates is pre-established for different types of cost documents. For example, for a "preliminary design budget," the configured indicator items may include: total project budget, construction costs, installation costs, equipment purchase costs, other costs, and unit cost indicators (such as yuan / kVA, yuan / km); for a "list of equipment and materials," the indicator items may include: equipment name, specifications, quantity, unit price, total price, and supplier. Each type of document corresponds to a set of indicator item definitions, and different indicator item definitions require data field names, data types, and unit specifications extracted from the document.

[0026] S103, the cost document is divided into multiple text segments, and each text segment is combined with a corresponding type of structured data indicator to construct a prompt word template. The prompt word template is then input into a large language model to identify and extract structured data. In practice, the original cost estimation document is preprocessed and segmented. For PDF or Word documents, the plain text content is extracted and divided into multiple text segments according to the document's natural chapters, paragraph marks, or a fixed length (in this embodiment, the length is 1000 characters per segment), while retaining the position information of the segments in the original text.

[0027] Secondly, construct prompt word templates. For a specific type of cost document and its corresponding list of indicators, combine the name and unit requirements of each indicator with each text fragment to form precise prompt words. For example, for a fragment of a preliminary estimate, to extract "total project budget," the prompt word could be designed as: Extract the value and unit of "total project budget" from the following text, requiring the value to be accurate to the yuan. If it is not mentioned in the text, return "not found."

[0028] Furthermore, a large language model is invoked for information recognition and extraction. The aforementioned prompt words are input in batches into the deployed large language model (in this embodiment, the model is a general-purpose large model). The model understands the contextual semantics according to the instructions, automatically locates and extracts the numerical values ​​and units corresponding to the indicators.

[0029] The extracted results returned by the model are validated and cleaned, such as checking the numerical format and unit normalization, to form structured key-value pair data, i.e., structured extracted data.

[0030] S104, The structured extracted data and the structured data are processed to obtain the merged data assets of the same project, and stored in a relational database; In practice, structured data extracted from various unstructured documents will be integrated with the original structured data mentioned above. The data processing includes: data cleaning to remove duplicate, erroneous, or incomplete records; data standardization to unify project coding, cost item systems, unit dimensions, etc.; and data association, using project number, project name, or contract number as primary keys to link and merge extracted cost details, equipment and material prices, etc., with the original project structure data. The processed data forms a complete, consistent, and high-quality project data asset, which is ultimately stored in a relational database (such as PostgreSQL, MySQL, or Oracle) to provide a reliable data foundation for upper-level analytical applications.

[0031] S105, Receive the user's analysis request, and use the large language model to parse the analysis request and convert it into a database query instruction; Furthermore, step S105 specifically includes steps S1051 to S1052: S1051, Use a large language model to perform semantic parsing on the analysis request to obtain the corresponding analysis dimensions and analysis intent; S1052, the analysis dimensions and analysis intent are converted into standard query statements and call instructions corresponding to the relational database.

[0032] In practical implementation, the user inputs a natural language request in a dialog interface (in this embodiment, the dialog interface includes a web or mobile terminal). The request is then fed into a large language model for intent recognition and slot filling. After fine-tuning with a small number of samples, the model can output a structured semantic framework, parsing out the structured analytical intent, analytical dimensions, and conditions. Based on the parsed structured parameters, standard SQL query statements or API call instructions are automatically generated through preset conversion templates or the code generation capabilities of the large language model itself. For example, an SQL query is generated to retrieve the actual cost and estimated amount of projects at each voltage level, and the deviation rate is calculated. At the same time, an instruction to call the "control line analysis model" is generated to make an over-budget warning judgment.

[0033] For example, when the input request is "Compare the average cost of substation projects of various voltage levels in the first half of the year and find the projects that exceed the budget", after analysis by the large language model, the analysis intent is: comparative analysis and deviation warning. The analysis dimensions and conditions are "time: first half of this year, project type: substation, classification dimension: voltage level, comparison benchmark: budget standard".

[0034] S106, according to the database query instruction, a preset data analysis model is invoked to analyze the project data assets and obtain the analysis result data.

[0035] Furthermore, step S106 specifically includes steps S1061 to S1062: S1061, Query the corresponding relevant fused data assets from the relational database according to the standard query statement; S1062, according to the calling instruction, call and execute at least one embedded data analysis model to analyze and calculate the relevant fused data assets to obtain structured analysis result data, wherein the data analysis model includes at least one of control line analysis model, material cost analysis model and historical engineering comparison model.

[0036] Furthermore, the control line analysis model is used to automatically match the cost characteristics of the current project with the preset cost control line standard, calculate the deviation value between the actual cost and the cost control line standard, and generate early warning information when the deviation value exceeds the preset threshold. The material cost analysis model is used to perform time-series trend analysis, inter-supplier price comparison analysis, and quantity-price-cost composition analysis of material prices by using equipment and material parameters extracted from cost documents. The historical project comparison model is used to automatically filter out similar historical projects from the historical project database that have a similarity to the current project that is higher than a threshold through clustering or similarity algorithms, and compares and analyzes the cost indicators of the current project with the corresponding cost indicators of the similar historical projects to generate a difference report.

[0037] In practical implementation, the control line analysis model incorporates multi-level cost control line standards categorized by project type, voltage level, and region. These standards are derived from accumulated data or industry quotas. During execution, the model automatically matches the cost characteristics of the current project with the corresponding control lines, calculating the absolute deviation and deviation rate between the actual cost and the standard. When the deviation rate exceeds a preset threshold (e.g., ±10%), the model will mark a warning in the results and generate a list of over-budget items.

[0038] The material cost analysis model uses equipment and material parameters extracted from documents to construct material price time series, enabling price trend analysis of core equipment (such as transformers, GIS, cables, etc.); it can perform horizontal price comparison analysis of quotations from different suppliers to identify price gaps; and it can also perform quantity, price, and cost composition analysis to reveal the degree of influence of various factors such as quantity, unit price, and transportation costs in material costs.

[0039] The historical engineering comparison model extracts key feature vectors (such as voltage level, capacity, line length, terrain conditions, and equipment type) from characteristic engineering projects. Then, based on clustering algorithms (such as K-Means) or cosine similarity algorithms, it automatically selects a group of similar historical engineering projects from the historical engineering database whose feature similarity exceeds a set threshold. Next, it compares the main cost indicators of the current project (such as unit cost and the proportion of sub-items) with the average level and distribution of these similar projects, generating a difference analysis report.

[0040] In some optional embodiments, the method further includes steps S107-S108: S107, The analysis result data is converted into natural language analysis text using the large language model, and the natural language analysis text is pushed to the user interface in the form of a dialog interface; S108, dynamically generate a visualization chart based on the analysis results data, and embed the visualization chart into the corresponding position of the dialog interface.

[0041] In practice, the analysis model outputs structured data tables or JSON objects, which are difficult to read directly. In this embodiment, by re-intervening with a large language model, the structured analysis results are used as input, combined with the context of the analysis request, to generate a coherent, clear, and focused natural language analysis text. For example, the generated text might be: "The average cost of 110kV substations in the first half of the year was XX yuan / kVA, of which the actual cost of Project XX exceeded the budget by 15%, with the main overspending item being construction costs, which requires close attention." This text is pushed segment by segment into the user interface's dialog stream in the form of speech bubbles.

[0042] While generating textual analysis conclusions, the system automatically selects the optimal visualization chart format (such as bar charts comparing costs across voltage levels, line charts displaying price trends, and dashboards showing deviation warnings) based on the structure and type of the analysis results data, and dynamically generates the charts using the visualization engine. These charts are directly embedded as interactive components in the corresponding message cards within the dialog interface. Users can directly zoom in and out of charts, hover over data points to view details, and switch chart types (such as from bar charts to pie charts) within the dialog interface without needing to switch to other reporting platforms, thus completing the entire closed loop of "questioning-analysis-insight-interaction" within a unified dialog interface.

[0043] In summary, the power transmission and transformation engineering cost data analysis method based on a large language model in the above embodiments of the present invention achieves automated extraction and governance of unstructured cost data by combining a large language model with configurable indicators and prompt word engineering, significantly improving data utilization and processing efficiency. By parsing user analysis intent through a natural language interaction engine and driving the embedded professional analysis model, it breaks through the limitations of traditional fixed reports, enabling business personnel to conduct flexible and in-depth data insights without technical background. Simultaneously, real-time linkage between data management and intelligent analysis realizes the transformation from superficial statistics to in-depth mining, enhancing the early warning capability of cost risks and effectively supporting lean cost control in power grid transmission and transformation engineering. Example 2

[0044] In another aspect, this invention proposes a power transmission and transformation engineering cost data analysis system based on a large language model. Please refer to [link / reference]. Figure 2 The figure shows a power transmission and transformation engineering cost data analysis system based on a large language model according to the second embodiment of the present invention, comprising: The data acquisition module 11 is used to acquire multi-source data on the cost of power transmission and transformation projects, wherein the multi-source data includes structured data and cost documents containing unstructured text. Data processing module 12 is used to pre-configure corresponding structured data indicator items based on the file type of the cost document, wherein the structured data indicator items define the data fields to be extracted from cost documents of different file types; The text processing module 13 is used to divide the cost document into multiple text segments, combine each text segment with a corresponding type of structured data indicator item to construct a prompt word template, and input the prompt word template into a large language model to identify and extract structured data. The data fusion module 14 is used to process the structured extracted data and the structured data to obtain the fused data assets of the same project, and store them in a relational database. The request parsing module 15 is used to receive the user's analysis request, parse the analysis request using the large language model, and convert it into a database query instruction; Data analysis module 16 is used to call a preset data analysis model to analyze the project data assets according to the database query command and obtain analysis result data.

[0045] Furthermore, the request parsing module 15 is specifically used for: The analysis request is semantically parsed using a large language model to obtain the corresponding analysis dimensions and analysis intent; The analytical dimensions and analytical intent are converted into standard query statements and invocation instructions corresponding to the relational database.

[0046] Furthermore, the data analysis module 16 is specifically used for: The corresponding integrated data assets are retrieved from the relational database according to the standard query statement. The call instruction invokes and executes at least one embedded data analysis model to analyze and calculate the relevant fused data assets, thereby obtaining structured analysis results data. The data analysis model includes at least one of a control line analysis model, a material cost analysis model, and a historical engineering comparison model.

[0047] Furthermore, in the data analysis module 16, the control line analysis model is used to automatically match the cost characteristics of the current project with the preset cost control line standard, calculate the deviation value between the actual cost and the cost control line standard, and generate early warning information when the deviation value exceeds the preset threshold. The material cost analysis model is used to perform time-series trend analysis, inter-supplier price comparison analysis, and quantity-price-cost composition analysis of material prices by using equipment and material parameters extracted from cost documents. The historical project comparison model is used to automatically filter out similar historical projects from the historical project database that have a similarity to the current project that is higher than a threshold through clustering or similarity algorithms, and compares and analyzes the cost indicators of the current project with the corresponding cost indicators of the similar historical projects to generate a difference report.

[0048] Furthermore, the system also includes: The data conversion module is used to convert the analysis result data into natural language analysis text using the large language model, and push the natural language analysis text to the user interface in the form of a dialog interface. The data visualization module is used to dynamically generate visualization charts based on the analysis results and embed the visualization charts into the corresponding positions of the dialog interface.

[0049] The functions or operation steps implemented by the above modules are largely the same as those in the above method embodiments, and will not be repeated here.

[0050] The power transmission and transformation engineering cost data analysis system based on a large language model provided in this embodiment of the invention has the same implementation principle and technical effects as the aforementioned method embodiment. For the sake of brevity, any parts not mentioned in the system embodiment can be referred to the corresponding content in the aforementioned method embodiment.

[0051] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0052] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this patent application should be determined by the appended claims.

Claims

1. A method for analyzing power transmission and transformation engineering cost data based on a large language model, characterized in that, Includes the following steps: Step 1: Obtain multi-source data on the cost of power transmission and transformation projects, wherein the multi-source data includes structured data and cost documents containing unstructured text; Step 2: Based on the file type of the cost document, pre-configure the corresponding structured data indicator items, wherein the structured data indicator items define the data fields to be extracted from cost documents of different file types; Step 3: Divide the cost document into multiple text segments, combine each text segment with the corresponding type of structured data indicator to construct a prompt word template, and input the prompt word template into the large language model to identify and extract structured data. Step 4: Process the extracted structured data with the structured data to obtain the merged data assets of the same project, and store them in a relational database; Step 5: Receive the user's analysis request, and use the large language model to parse the analysis request and convert it into a database query instruction; Step Six: Based on the database query command, invoke the preset data analysis model to analyze the project data assets and obtain the analysis results data.

2. The method for analyzing power transmission and transformation engineering cost data based on a large language model according to claim 1, characterized in that, Step five includes: The analysis request is semantically parsed using a large language model to obtain the corresponding analysis dimensions and analysis intent; The analytical dimensions and analytical intent are converted into standard query statements and invocation instructions corresponding to the relational database.

3. The method for analyzing power transmission and transformation engineering cost data based on a large language model according to claim 1, characterized in that, Step six includes: The corresponding integrated data assets are retrieved from the relational database according to the standard query statement. The call instruction invokes and executes at least one embedded data analysis model to analyze and calculate the relevant fused data assets, thereby obtaining structured analysis results data. The data analysis model includes at least one of a control line analysis model, a material cost analysis model, and a historical engineering comparison model.

4. The method for analyzing power transmission and transformation engineering cost data based on a large language model according to claim 3, characterized in that, The control line analysis model is used to automatically match the cost characteristics of the current project with the preset cost control line standard, calculate the deviation value between the actual cost and the cost control line standard, and generate an early warning information when the deviation value exceeds the preset threshold. The material cost analysis model is used to perform time-series trend analysis, inter-supplier price comparison analysis, and quantity-price-cost composition analysis of material prices by using equipment and material parameters extracted from cost documents. The historical project comparison model is used to automatically filter out similar historical projects from the historical project database that have a similarity to the current project that is higher than a threshold through clustering or similarity algorithms, and compares and analyzes the cost indicators of the current project with the corresponding cost indicators of the similar historical projects to generate a difference report.

5. The method for analyzing power transmission and transformation engineering cost data based on a large language model according to claim 1, characterized in that, The method further includes: The analysis results data are converted into natural language analysis text using the large language model, and the natural language analysis text is pushed to the user interface in the form of a dialog interface. Based on the analysis results, a visual chart is dynamically generated and embedded into the corresponding position of the dialog interface.

6. A power transmission and transformation engineering cost data analysis system based on a large language model, characterized in that, include: The data acquisition module is used to acquire multi-source data on the cost of power transmission and transformation projects, wherein the multi-source data includes structured data and cost documents containing unstructured text; The data processing module is used to pre-configure corresponding structured data indicator items based on the file type of the cost document, wherein the structured data indicator items define the data fields to be extracted from cost documents of different file types; The text processing module is used to divide the cost document into multiple text fragments, combine each text fragment with a corresponding type of structured data indicator item to construct a prompt word template, and input the prompt word template into a large language model to identify and extract structured data. The data fusion module is used to process the structured extracted data and the structured data to obtain the fused data assets of the same project, and store them in a relational database. The request parsing module is used to receive the user's analysis request, parse the analysis request using the large language model, and convert it into a database query instruction; The data analysis module is used to call a preset data analysis model according to the database query command to analyze the project data assets and obtain the analysis result data.

7. The power transmission and transformation engineering cost data analysis system based on a large language model according to claim 6, characterized in that, The request parsing module is specifically used for: The analysis request is semantically parsed using a large language model to obtain the corresponding analysis dimensions and analysis intent; The analytical dimensions and analytical intent are converted into standard query statements and invocation instructions corresponding to the relational database.

8. The power transmission and transformation engineering cost data analysis system based on a large language model according to claim 6, characterized in that, The data analysis module is specifically used for: The corresponding integrated data assets are retrieved from the relational database according to the standard query statement. The call instruction invokes and executes at least one embedded data analysis model to analyze and calculate the relevant fused data assets, thereby obtaining structured analysis results data. The data analysis model includes at least one of a control line analysis model, a material cost analysis model, and a historical engineering comparison model.

9. The power transmission and transformation engineering cost data analysis system based on a large language model according to claim 8, characterized in that, In the data analysis module, the control line analysis model is used to automatically match the cost characteristics of the current project with the preset cost control line standard, calculate the deviation value between the actual cost and the cost control line standard, and generate early warning information when the deviation value exceeds the preset threshold. The material cost analysis model is used to perform time-series trend analysis, inter-supplier price comparison analysis, and quantity-price-cost composition analysis of material prices by using equipment and material parameters extracted from cost documents. The historical project comparison model is used to automatically filter out similar historical projects from the historical project database that have a similarity to the current project that is higher than a threshold through clustering or similarity algorithms, and compares and analyzes the cost indicators of the current project with the corresponding cost indicators of the similar historical projects to generate a difference report.

10. The power transmission and transformation engineering cost data analysis system based on a large language model according to claim 6, characterized in that, The system also includes: The data conversion module is used to convert the analysis result data into natural language analysis text using the large language model, and push the natural language analysis text to the user interface in the form of a dialog interface. The data visualization module is used to dynamically generate visualization charts based on the analysis results and embed the visualization charts into the corresponding positions of the dialog interface.

Images