Power grid strategy multi-level consistency evaluation method and device

Abstract

The embodiment of the invention provides a power grid strategy multi-level consistency evaluation method and device, and the method innovatively proposes a cross-stage strategy element embedding expression and consistency evaluation method, and carries out the element extraction, semantic embedding and cross-stage mapping of strategy, planning, project and plan multi-level texts. A strategic semantic conduction path between a strategic target and a power grid project execution link is established and optimized, so that consistency analysis and optimization in a strategic landing process are realized, the scientificity and intelligence level of strategic landing are improved, and finally a power grid project and strategic consistency evaluation method is constructed. Effects and defects of the project in response strategy aspects are obtained, and thoughts are provided for project optimization.

Description

Method and apparatus for multi-level consistency assessment of power grid strategy Technical Field

[0001] This application relates to the field of data processing, specifically to a method and apparatus for multi-level consistency assessment of power grid strategy. Background Technology

[0002] In the fields of strategic planning and project management, embedding strategic elements into project processes is crucial. Its core objective is to ensure that project goals are highly aligned with high-level national and industry strategies. To achieve this, a method capable of intelligently analyzing the relationship between projects and strategies is needed. The quality of strategic element embedding directly impacts the complete communication of strategic intent, the precise matching of indicators and specifications, and the effective assessment of strategic consistency across project phases.

[0003] Currently, the main challenges facing this process are: 1) Inconsistent embedding and expression of strategic elements across stages: The embedding methods of strategic elements are not uniform in different stages such as project reserves, planning, and comprehensive planning, and the semantic expression is difficult to maintain coherence, resulting in distortion or attenuation of strategic intent in the transmission process, making it difficult to conduct effective cross-stage alignment and consistency assessment.

[0004] 2) Lack of strategic element retrospective analysis and dynamic verification capabilities: There is a lack of methods for dynamic tracking, mapping and consistency verification of strategic elements throughout their entire life cycle, from project reserves to comprehensive planning, making it difficult to achieve strategic-oriented closed-loop management and decision optimization.

[0005] This invention aims to solve the above problems by establishing a cross-stage embedding specification framework and realizing dynamic tracking and verification of elements, thereby improving the quality and consistency of strategic element embedding and providing reliable support for strategic-oriented projects in complex fields such as power grids. Summary of the Invention

[0006] To address the problems in the prior art, this application provides a method and apparatus for multi-level consistency assessment of power grid strategy, which can improve the quality and consistency of strategic element embedding and provide reliable support for strategic-oriented projects in complex fields such as power grids.

[0007] To address at least one of the aforementioned problems, this application provides the following technical solution: Firstly, this application provides a multi-level consistency assessment method for power grid strategies, comprising: collecting energy-related policy documents and extracting strategic elements related to the strategy from these documents to obtain a strategic element library; the strategic elements include standardized terminology, quantitative indicators, responsible entities, and constraints; collecting power grid-related management documents and dividing them into three stages—reserve, feasibility study, and comprehensive planning—according to the entire lifecycle; inputting the power grid project-related management documents and the strategic element library into an LLM (Limited Language Model); using the LLM model, based on the macro-strategies and preset prompts in the energy-related policy documents, analyzing the strategic elements contained in the project texts of each stage and the rules governing their manifestation, and outputting a rule table for strategic elements at each stage; extracting stage elements from the project texts of each stage according to the rule table; performing consistency checks on the extracted stage elements at each stage and retaining semantically consistent items to obtain a set of stage elements for each stage; evaluating the consistency between the strategic elements and the stage elements of the three stages, and outputting the consistency assessment result.

[0008] Furthermore, the step of extracting strategic elements related to the strategy from the documents to obtain a strategic element library includes: using an LLM (Large Language Model) to perform semantic understanding on energy-related policy documents, identifying and extracting strategic-related data from the documents; using a named entity recognition model to identify core entity information in the extracted strategic-related data to obtain a set of strategic elements; the core entity information includes normative terminology, quantitative indicators, responsible entities, and constraints.

[0009] Furthermore, the steps of using the LLM (Large Language Model) to analyze the strategic elements and their manifestation rules in project texts at each stage, based on the macro-strategy and preset prompts in relevant power grid project policy documents, and outputting rule tables for strategic elements at each stage, include: Project Reserve Stage: Using the LLM and preset prompts, analyze how strategic elements are manifested in the project reserve stage, clarify the strategic elements that the macro-strategy needs to respond to in the project reserve stage, and generate strategic element rules for the project reserve stage; Project Feasibility Study Stage: Using the LLM and preset prompts, analyze the specific elements into which the macro-strategy is decomposed, and analyze the specific elements contained in the strategic elements, outputting strategic element rules for the project feasibility study stage; the specific elements include technical indicators, economic benefit assessment, and environmental constraints; Comprehensive Planning Stage: Using the LLM, analyze the integration and optimization of strategic elements of each project in the overall comprehensive plan at this stage, forming strategic element rules for the comprehensive planning stage.

[0010] Furthermore, the step of extracting stage elements from the project text of each stage according to the rule table includes: inputting the project text of the reserve stage and the rule table into a rule matching model, using the rule matching model to scan the project text of the reserve stage, and extracting the stage elements in the project text of the reserve stage according to the rule table; inputting the project text of the feasibility study stage and the rule base into a named entity recognition model, and combining the named entity recognition model with regular expressions to extract the stage elements in the project text of the feasibility study stage; inputting the project text of the comprehensive planning stage and the rule table into a rule matching model, using the rule matching model to scan the project text of the comprehensive planning stage, and extracting the stage elements in the project text of the comprehensive planning stage according to the rule table.

[0011] Furthermore, the step of performing consistency checks on the stage elements extracted from each stage and retaining semantically consistent items to obtain the stage element set for each stage includes: aligning the stage element sets for each stage with data, and using set intersection calculation to identify semantically consistent items and inconsistent elements in the three element sets; converting inconsistent elements into high-dimensional semantic vectors, and calculating the cosine similarity between the vectors: if the similarity is greater than a preset threshold, it is determined to be a semantically consistent item; if the similarity is less than the preset threshold, it is determined to be a conflicting item and discarded; merging the semantically consistent items of each stage to form the stage element set for each stage.

[0012] Furthermore, the step of evaluating the consistency between strategic elements and the stage elements of the three stages, and outputting the consistency evaluation results, includes: Semantic level: converting strategic elements and stage elements into high-dimensional semantic vectors, and calculating the cosine similarity between the semantic vectors of strategic elements and the stage elements of each stage; Data level: using a rule matching model to extract quantifiable indicators of strategic elements and stage elements of each stage, calculating whether they are within a threshold range, and calculating the matching score linearly according to the actual degree of deviation; the quantifiable indicators include capacity and investment amount; Structural level: intersecting the set of strategic elements with the set of stage elements of each stage, and calculating the coverage rate of stage elements of each stage of the project; the semantic similarity, matching score, and coverage rate obtained from the three levels are weighted and calculated to obtain the final consistency score of the project.

[0013] Furthermore, the aforementioned multi-level consistency assessment method for power grid strategy also includes: establishing a graph knowledge base with strategic elements and stage elements of each stage as nodes, and establishing an evolution chain of strategic elements in each stage starting from the initial strategic elements; using the LLM large language model, taking the evolution chain and stage elements of each stage as input, and back-generating the cross-stage required embedded expression form of each strategic element; and using the cross-stage required embedded expression form to optimize the strategic consistency of the three stages.

[0014] Secondly, this application provides a multi-level consistency assessment device for power grid strategy, comprising: a strategic element extraction module, used to collect energy-related policy documents and extract strategic elements related to the strategy from the documents to obtain a strategic element library; the strategic elements include standardized terminology, quantitative indicators, responsible entities, and constraints; a rule analysis module, used to collect power grid-related management documents and divide the power grid project-related management documents into three stages according to the entire life cycle: reserve, feasibility study, and comprehensive planning; inputting the power grid project-related management documents and the strategic element library into an LLM (Large Language Model), and using the LLM, based on the macro-strategy in the energy-related policy documents and preset prompts, analyzing the strategic elements contained in the project text of each stage and the rules for the manifestation of the strategic elements, and outputting a rule table of strategic elements in each stage; a stage element extraction module, used to extract stage elements from the project text of each stage according to the rule table; performing consistency checks on the stage elements extracted from each stage and retaining semantically consistent items to obtain a stage element set for each stage; and a result evaluation module, used to evaluate the consistency between the strategic elements and the stage elements of the three stages, and output the consistency evaluation result.

[0015] Thirdly, this application provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the steps of the aforementioned multi-level consistency assessment method for power grid strategy.

[0016] Fourthly, this application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the aforementioned multi-level consensus assessment method for power grid strategy.

[0017] Fifthly, this application provides a computer program product, including a computer program / instructions that, when executed by a processor, implement the steps of the aforementioned multi-level consistency assessment method for power grid strategy.

[0018] As can be seen from the above technical solution, this application provides a method and device for multi-level consistency assessment of power grid strategy. This method innovatively proposes a cross-stage strategic element embedding expression and consistency evaluation method. By extracting elements, embedding semantics, and mapping cross-stages of multi-level texts of strategy, planning, project, and program, it establishes and optimizes the strategic semantic transmission path between strategic objectives and power grid project execution links, thereby realizing consistency analysis and optimization in the process of strategy implementation, improving the scientific and intelligent level of strategy implementation, and finally constructing a consistency evaluation method for power grid projects and strategies. It obtains the effects and shortcomings of projects in responding to strategies, and provides ideas for project optimization. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 is a flowchart of one of the multi-level consistency assessment methods for power grid strategy in this application embodiment; Figure 2 is a flowchart of another of the multi-level consistency assessment methods for power grid strategy in this application embodiment; Figure 3 is a structural diagram of the multi-level consistency assessment device for power grid strategy in this application embodiment; Figure 4 is a structural diagram of the electronic device in this application embodiment.

[0021] Reference numerals: Electronic device 9600, Central processing unit 9100, Memory 9140, Communication module 9110, Input unit 9120, Audio processor 9130, Display 9160, Power supply 9170, Buffer memory 9141, Application / function storage unit 9142, Data storage unit 9143, Driver storage unit 9144, Antenna 9111, Speaker 9131, Microphone 9132. Detailed Implementation

[0022] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0023] The acquisition, storage, use, and processing of data in this application all comply with the relevant provisions of national laws and regulations.

[0024] In view of the problems existing in the prior art, this application provides a method and device for multi-level consistency assessment of power grid strategy. This method innovatively proposes a cross-stage strategic element embedding expression and consistency evaluation method. By extracting elements, embedding semantics, and mapping cross-stages of multi-level texts of strategy, planning, project, and program, it establishes and optimizes the strategic semantic transmission path between strategic objectives and power grid project execution links, thereby realizing consistency analysis and optimization in the process of strategy implementation, improving the scientific and intelligent level of strategy implementation, and finally constructing a consistency evaluation method for power grid projects and strategies. It obtains the effectiveness and shortcomings of projects in responding to strategies, and provides ideas for project optimization.

[0025] In order to improve the quality and consistency of strategic element embedding and provide reliable support for strategic-oriented projects in complex fields such as power grids, this application provides an embodiment of a multi-level consistency assessment method for power grid strategies. Referring to Figures 1 and 2, the multi-level consistency assessment method for power grid strategies specifically includes the following: Step S101: Collect energy-related policy documents and extract strategic elements related to strategy from the documents to obtain a strategic element library; the strategic elements include normative terminology, quantitative indicators, responsible entities, and constraints.

[0026] Optionally, in this embodiment, heterogeneous documents from multiple sources, such as national and governmental policies and regulations related to energy, as well as industry-related strategies and management regulations, can be collected. A unified file format and encoding method can be used to convert the content of documents from different sources into plain text format encoded in UTF-8. Regular expression matching methods can be used to identify the chapter hierarchy in the documents, automatically extracting titles such as "Chapter ×" and "Section ×.×" and their corresponding content for subsequent semantic segmentation. All documents are indexed according to their source, time, and type, and stored uniformly in a database (such as MongoDB or Neo4j) to form an initial strategic text library with a consistent structure and standardized fields, providing a unified data foundation for subsequent model processing.

[0027] Optionally, in this embodiment, this step provides a method for extracting strategic features and forming a top-level strategic element library based on policy document management regulations. This method includes: firstly, from a horizontal perspective, inputting the collected relevant strategic and management regulations documents for the two types of power grid projects targeted by this invention into an LLM (Large Language Model, such as ChatGLM, GPT, LLaMA, etc.) large language model; structurally extracting the title and corresponding text content of each part of the document for semantic understanding. For example, the temperature parameter can be set to 0.3, and the maximum generated length to 4096. Structured prompts such as "Please parse the document title and corresponding content paragraph by paragraph to identify strategic intent and complex terminology" are used to extract the chapter titles and their corresponding text content of the document, thus structurally extracting them. The system outputs a JSON-formatted document containing necessary data such as title, abstract, keywords, and key terms. From a vertical perspective, structured documents derived from the LLM framework's existing understanding of strategy are used as input to train the model. Specialized small models (such as BERT+CRF named entity recognition, GlobalPointer, and rule-matching models) are employed. Taking BERT+CRF as an example, the model performs fine-grained element recognition on the text. For instance, the model can be set with a hidden layer dimension of 768, a learning rate of 2e-5, a batch size of 32, and 20 training epochs. Specific parameters can be adjusted according to actual needs. The labeling system uses BIO encoding. By combining model recognition with regular rule matching, core entity information such as standardized terms, quantitative indicators (such as investment amount, capacity, and technical standards), responsible parties, and constraints are extracted from the text. All extracted results are written into a graph database in the form of structured fields, forming a top-level strategic element library containing semantic relationships and attributes.

[0028] Currently, single models have limitations in the identification and extraction of strategic elements: while large models possess broad semantic understanding advantages, they may lack accuracy in specific normative matching; small models excel at rule-based precise matching, but their generalization understanding ability is limited. Existing technologies lack a collaborative mechanism that combines the semantic understanding capabilities of large models with the normative processing capabilities of small models, resulting in deficiencies in the completeness and confidence of the final constructed set of strategic elements.

[0029] To address the aforementioned issues, this invention proposes a strategic element embedding and consistent expression method based on a synergistic approach of large and small models. This scheme utilizes a large model for macro-level semantic analysis and initial element screening to ensure broad coverage and semantic integrity of the strategic element set. Simultaneously, a small model is used for refined specification matching and consistency verification to ensure the accuracy and high confidence of the elements. Through this synergistic mechanism, this invention can construct a high-confidence strategic element set at different stages, such as project preparation, planning, and comprehensive planning, and achieve consistency and full lifecycle traceability of strategic embedding, thereby providing reliable support for strategically oriented projects in complex fields such as power grids.

[0030] Step S102: Collect relevant management documents for the power grid and divide them into three stages according to the entire life cycle: reserve, feasibility study, and comprehensive planning. Input the relevant management documents and strategic element library of the power grid project into the LLM language model. Using the LLM language model, based on the macro strategies and preset prompts in energy-related policy documents, analyze the strategic elements contained in the project texts of each stage and the rules for the manifestation of strategic elements, and output the rule table of strategic elements in each stage.

[0031] Optionally, in this embodiment, this step provides a method for analyzing the multi-stage strategic action mechanism. This method includes: utilizing the reasoning and analysis capabilities of LLM to deeply analyze the action mechanism and transformation logic of strategy at different stages of the project's entire lifecycle. First, three stages are defined: project reserve, feasibility study, and comprehensive planning. The strategic constraint logic of each stage is independently reasoned. The strategic element library is input into the model via prompts, requiring the output of the manifestation, constraint indicators, and target objects of strategic elements at different stages. The model output is then converted into a rule table after regularization. The specific focus of the rules for each stage is as follows: Project Reserve Stage: Combining the LLM's understanding of the strategic mechanism at this stage, the manifestation of macro-strategy in the project reserve stage is analyzed. The system prompt is set to "Please analyze the manifestation of strategy in the project reserve stage and clarify the strategic elements that need to be responded to," generating rules for the strategic elements that should be included in the project reserve stage, such as "New energy project reserves must include 'renewable energy quota indicators' and 'grid access capacity reservation' elements." Feasibility Study Phase: This phase studies how the macro strategy is broken down into specific elements such as technical indicators, economic benefit assessments, and environmental constraints in the feasibility study. Using the same LLM prompt, "Please break down the macro strategy into technical indicators, economic benefit assessments, and environmental constraints in the feasibility study," semi-structured rules are output, such as "Technical indicators must include 'investment per kilowatt ≤ 8500 yuan', and environmental constraints must include 'carbon emission intensity ≤ 0.5 kg CO2 / kWh'." Integrated Planning Phase: This phase analyzes the integration and optimization of strategic elements for each project within the annual or overall integrated plan, forming strategic element mapping rules for the integrated planning phase. This rule table includes fields such as phase name, strategic elements, constraints, and quantitative indicators, forming a multi-phase strategic element evolution rule base to provide a rule basis for the generation of elements in subsequent phases.

[0032] Optionally, in this embodiment, the strategic mechanism analysis method for the project reserve stage includes: 1. Constructing staged analysis prompts and element input. First, the system retrieves policy text fragments that are highly related to the strategic elements to be analyzed (such as "rural revitalization" and "new energy access") from the "initial strategic text library" (i.e., the original set of macro strategies). For example, to analyze the "rural revitalization" element, the system will extract the specific paragraph about "improving the power supply guarantee capacity of rural power grids" from the "Opinions on Implementing the Rural Revitalization Strategy".

[0033] Next, design a highly structured system prompt to guide the LLM into the role of a "power grid project reserve expert." This prompt should clearly define the task background, output format requirements, and include the characteristics of the reserve phase. For example: "You are a power grid planning expert. Based on the provided macro-strategic text, analyze how a certain strategic element should be reflected in the initial stage of 'project reserve.' This stage focuses on project selection and initial assessment of strategic compliance. Please output specific and verifiable element rules." Subsequently, the macro-strategic text fragment and a list of specific strategic elements are submitted to the LLM as input. The strategic elements are input as a structured list (e.g., JSON format) as the user prompt, for example, input element list: ["Improve power supply reliability", "Rural power grid consolidation and improvement"].

[0034] Set the LLM generation parameters, such as temperature=0.1 to ensure output stability and max_tokens=2048 to ensure content integrity.

[0035] 2. LLM's Phase-Adaptive Reasoning and Rule Generation: Based on the macro-strategic text and its built-in strategic knowledge and logical reasoning capabilities, LLM performs "reserve phase-adaptive transformation" on each strategic element. Its internal reasoning process can be analyzed as follows: Interpreting Strategic Intent: Deeply interpreting the macro-strategic text to understand its policy objectives, principles, and underlying intentions. For example, understanding the strategic intent of "improving the power supply guarantee capacity of rural power grids" is to address the insufficient power supply reliability in rural areas and support the development of rural industries.

[0036] Define the characteristics of the stage: Clarify that the core task of the "reserve stage" is project selection and initial assessment of strategic compliance, focusing on the necessity of project initiation and the correctness of the strategic direction.

[0037] Element concretization: Transforming abstract strategic intentions into key attributes, eligibility criteria, or core elements that must be addressed in project proposals or reserve application materials at this stage. For example, LLM would deduce that: a power grid project responding to this strategy must clearly define its "project type" as "rural power grid renovation and upgrading" during the reserve stage; its "necessity for construction" must explain "solving power supply problems"; and its "preliminary construction content" should include "adding / renovating distribution transformers," etc.

[0038] 3. Rule Extraction, Structuring, and Storage: First, the LLM organizes the above reasoning results into preliminary natural language rules using binding language such as "must," "should include," and "must be explicitly stated."

[0039] Subsequently, using regular expressions or parsing scripts, key fields such as "stage name", "strategic elements", and "constraints" are precisely extracted from the output text and then standardized and formatted.

[0040] Ultimately, the formatted rules (e.g., stage: reserve; strategic element: rural revitalization; constraints: project type must be "rural power grid transformation and upgrading", construction necessity must clarify "solving low voltage / power supply bottleneck", preliminary construction content must include "distribution transformer capacity increase / line extension") will be written into the multi-stage strategic element evolution rule base.

[0041] Optionally, in this embodiment, the specific operational steps for strategic mechanism analysis during the project feasibility study stage are as follows: 1. Constructing feasibility study-specific prompts and element inputs. First, the system constructs the context and retrieves macro-strategic text fragments directly related to the strategic element to be analyzed (such as "clean energy development" and "improving system efficiency") from the "initial strategic text library". For example, to analyze "clean energy development", the specific statements in the policy regarding "accelerating the development of new energy sources such as wind power and photovoltaics, and promoting the transformation of the energy structure" will be extracted.

[0042] Next, design dedicated system prompts for the feasibility study phase. These prompts must clearly define the task characteristics of this phase, for example: "You are an economic and technical review expert for a power grid project. Currently in the 'feasibility study' phase, please decompose the following strategic elements based on the provided macro-strategic background into specific technical indicators, economic benefit assessment indicators, and environmental constraints that must be calculated and demonstrated in detail at this stage. The output should be in the form of quantifiable and verifiable indicators." Finally, the original macro-strategic text and the strategic elements to be analyzed (such as "clean energy share," "power grid investment benefits," and "carbon emission control") are used as user input and passed to the Large Language Model (LLM).

[0043] Configure LLM generation parameters, such as setting temperature=0.1 to keep the output stable, and enable JSON output format to facilitate subsequent automated processing.

[0044] 2. LLM performs multi-dimensional indicator decomposition and generation. LLM, based on its powerful domain knowledge, performs indicator decomposition on each strategic element. Its internal reasoning logic follows the standard structure of feasibility study reports: 1) Tracing back to strategic intent: Combining the original macro-strategic text, understand the deep policy objectives behind the elements. For example, understand the macro-intent of "clean energy development" as "optimizing the energy structure, ensuring energy security, and promoting low-carbon transformation".

[0045] 2) Define the tasks of the feasibility study: The core of the "feasibility study stage" is to conduct a quantitative demonstration of the project's technical feasibility, economic rationality, and environmental impact.

[0046] 3) Strategic Indicator Translation: This involves accurately translating macro-level strategic intentions into quantifiable and assessable engineering and economic language that requires detailed calculation and verification during the feasibility study phase. Technical Indicator Conversion: This transforms strategic intentions into specific technical parameters. For example, regarding the element of "improving system flexibility," LLM will break it down into technical indicators such as "peak-shaving capacity demand (MW)" and "energy storage deployment duration (hours)."

[0047] Linking with economic indicators: Quantifying strategic requirements into economic benefit assessment standards. For example, for the element of "power grid investment benefits", LLM will link it with key financial indicators such as "internal rate of return on total project investment (%)" and "investment payback period (years)".

[0048] Specific constraints: Strategic constraints are clearly defined as hard thresholds in areas such as environmental protection and safety. For example, for the element of "ecological protection", LLM will specify it as binding indicators such as "the distance (in meters) between the project site and the ecological red line" and "the noise control value (decibels) during the construction period".

[0049] LLM organizes the above analysis results into semi-structured content that includes "indicator categories", "specific indicator items", and "thresholds or requirements".

[0050] 3. Rule extraction, quantitative verification, and rule base update: Using regular expressions or a dedicated parser, key fields such as "stage name: feasibility study stage", "strategic elements", "quantitative indicators" (such as "investment per kilowatt ≤ 8500 yuan") and "constraints" (such as "carbon emission intensity ≤ 0.5 kg CO2 / kWh") are accurately extracted from the text generated by LLM.

[0051] The extracted quantitative indicators are logically and unit-wise verified to ensure that their numerical range and units conform to common sense in engineering (for example, the investment amount is in "ten thousand yuan" or "hundred million yuan", rather than an unreasonable huge number).

[0052] Finally, the validated rules are standardized and formatted, and written into the "Multi-Stage Strategic Element Evolution Rule Base" as executable review standards. These rules will be directly used for subsequent automated element extraction and compliance assessment of the "Project Feasibility Study Report".

[0053] Optionally, in this embodiment, the specific operational steps for the strategic mechanism analysis in the comprehensive planning phase are as follows: 1. Constructing a global overall analysis context. First, the system retrieves macro-strategic text fragments emphasizing global balance, resource optimization, and annual goals from the "initial strategic text library." For example, it extracts policy statements with a global orientation, such as "strengthen the overall planning of power grid investment, optimize the layout, and ensure key areas" or "ensure the smooth completion of the annual new energy consumption target."

[0054] Next, a dedicated system prompt is designed, instructing the LLM to act as an "expert in the power grid company's planning department." Its core task is: based on the provided macro-strategic background, analyze how the following strategic elements should be integrated across projects, prioritized, and resources optimized during the 'integrated planning' phase to form a coordinated and unified annual execution plan.

[0055] Finally, the selected macro-strategy excerpts and the list of strategic elements to be analyzed (which have been decomposed in the previous stage) are submitted to LLM as input.

[0056] 2. LLM conducts global integration and optimization reasoning based on strategic intent. LLM conducts global reasoning based on the received information. The process is as follows: 1) Understanding the overall intent: interpreting the overall requirements of the macro strategy such as "overall planning", "optimization" and "ensuring key points", and understanding that its core is to maximize the overall strategic benefits under limited resources.

[0057] 2) Define the characteristics of the planning phase: Clarify that the core task of the "comprehensive planning phase" is to package, prioritize, allocate resources to, and set annual milestones for all reserve and feasibility study projects.

[0058] 3) Element Integration and Rule Generation: Elevate discrete, project-level strategic element requirements to planning-level integrated rules, priority criteria, and resource allocation logic. For example, for the "investment benefit" element: deduce the need to set an "annual total investment limit for the company" and establish a "project priority evaluation model, with weights tilted towards projects with high internal rates of return."

[0059] Regarding the "carbon emission reduction" element: it is deduced that a "company's annual total carbon emission reduction target" needs to be set, and this target should be broken down into "carbon emission reduction quotas to be undertaken by each region / project group".

[0060] Regarding the "security guarantee" element: the rule is deduced that "projects involving supply guarantee in core cities or repair of critical lines will have their planned priority automatically increased".

[0061] 3. Generate and structure planning-level coordination rules. The LLM outputs the above global reasoning results as planning-level coordination rules.

[0062] Use regular expressions or parsing scripts to accurately extract key fields from the output text, such as "Phase Name: Comprehensive Planning Phase", "Strategic Elements", "Integration / Optimization Rules" (e.g., "Setting an annual total investment limit") and "Priority Factors" (e.g., "Priority Increase for Supply Guarantee Projects").

[0063] Ultimately, the formatted rules (e.g., Stage: Comprehensive Plan; Strategic Elements: Investment Benefits and Carbon Emission Reduction; Integration Rules: Establish a Project Priority Evaluation Model, comprehensively considering the internal rate of return on investment and carbon emission reduction per unit of investment; Constraints: Total annual investment shall not exceed XX billion yuan, and total carbon emission reduction shall not be less than YY tons) will be written into the "Multi-Stage Strategic Element Evolution Rule Base".

[0064] Step S103: Extract stage elements from the project text of each stage according to the rule table; perform consistency checks on the stage elements extracted from each stage and retain semantically consistent items to obtain the stage element set of each stage.

[0065] Optionally, in this embodiment, the method for generating and structuring multi-stage strategic element sets includes: using the multi-stage evolution rule base of strategic elements obtained from the LLM analysis in the third step as input, and using a small model to generate and structure the strategic element sets for each stage. In the project reserve stage, a rule matching model is invoked to scan the reserve project text line by line, extracting basic project elements based on the keywords and constraints of the corresponding stage in the rule base, and standardizing the units and formats of indicators such as investment amount, capacity, and construction period to generate a strategic element set for the reserve stage. In the feasibility study stage, combining the BERT model and regular expression recognition module, data such as economic indicators, technical parameters, and feasibility coefficients in the feasibility study report are extracted and indexed to generate a structured feasibility study stage element set. In the comprehensive planning stage, consistency checks are performed on the element sets of the first two stages. Conflicting items are eliminated using set intersection calculation and a semantic similarity threshold of 0.85, while semantically consistent items are retained, ultimately forming the strategic element set for the comprehensive planning stage, achieving the unification and comparability of multi-stage data. By comprehensively leveraging the semantic understanding capabilities of large language models and the rule-matching advantages of small models, we can achieve accurate analysis of national and industry strategic intentions and standardized mapping of multi-stage strategic elements.

[0066] Optionally, this embodiment also includes reverse derivation and optimization of the strategic embedding method: the output results of the first three stages, i.e., the set of elements that the project should have in the three stages obtained in step S4, are input into a graph knowledge base (Neo4j, etc.). Using "strategic elements" and "stage elements" as node types, an "evolution" relationship is established, and the Node2Vec algorithm is used to generate 256... The dimensional node embedding vectors are calculated using the BERT model to determine the cosine similarity between semantic vectors of the same strategic element at different stages, with a threshold of 0.8. When the similarity exceeds the threshold, a semantic alignment relationship is established. This identifies different expressions of the same element in the three stages, resulting in the evolutionary chain of the strategic element and the corresponding set of stage elements for each stage. Using a large model, the evolutionary chain and the multi-stage strategic element library are taken as input. Semi-structured embedding expression paths are generated through prompt word engineering, such as "Please deduce the cross-stage embedding expression path of the strategic element based on the evolutionary chain." The element set is used as keywords, and the evolutionary chain is used as the generation idea to deduce the cross-stage semi-structured embedding expression path of the strategic element. At the same time, the semantic weight of each element in the strategic document is calculated based on the TF-IDF algorithm, and the priority level is adjusted (e.g., the weight of "carbon emission reduction indicator" is set to 0.8, and the weight of "investment amount" is set to 0.6). This yields the cross-stage embedding expression forms of each element in the top-level strategic element set obtained in step S2, thereby optimizing the strategic consistency expression methods in the project reserve, planning, and comprehensive planning stages.

[0067] This embodiment addresses the problem of inconsistent embedding methods and semantic expressions of strategic elements in various stages of current power grid projects, such as reserves, planning, and comprehensive planning. It studies and establishes a cross-stage strategic element embedding standard framework to achieve accurate transmission and semantic coherence of strategic intent throughout the entire project lifecycle, providing basic support for multi-stage strategic alignment and consistency assessment.

[0068] Step S104: Assess the consistency between the strategic elements and the phase elements of the three phases, and output the consistency assessment results.

[0069] Optionally, in this embodiment, the strategic consistency assessment method includes: using this method to compare the matching degree between the strategic elements of the three stages of power grid project reserves, planning, and integrated planning and the top-level strategy, outputting quantitative consistency assessment results to intuitively reflect the response of each stage to the strategic objectives. The specific consistency assessment is divided into semantic, data, and structural levels: At the semantic level, the process documents of the three stages are used as input. Based on the semantic understanding capability of LLM, it identifies whether each stage of the project contains the strategic elements that should be present at that stage. That is, LLM is used to identify and extract the elements of each stage of the project to be assessed, and combined with a pre-trained language model (BERT, etc.), the elements are converted into corresponding high-dimensional semantic vectors. The cosine similarity of the semantic vectors of the stage elements and the top-level strategic elements is calculated, with a threshold set to 0.85, to determine whether semantic consistency is achieved. At the data level, a small model is used to extract quantifiable indicators (such as capacity and investment amount) using a rule-matching model, and it is calculated whether they are within the threshold range (e.g., "installed capacity must be within ±10% of the planned value"). The matching score is calculated linearly according to the actual deviation. At the structural level, the intersection of two element sets is taken, and the coverage rate p of the elements of each stage of the project is calculated. The calculation formula is:

[0070] After obtaining the consistency scores at the three levels, a weighted average is calculated to obtain the final consistency score for the project. The score calculation formula is as follows:

[0071] Where the weighting coefficient α+β+γ=1, Sim semantics For semantic vector cosine similarity at the semantic level, Sim semantics For data-level matching scores, Sim semantics The coverage is represented at the structural level, and the score is the strategic consistency evaluation score for the power grid project. The weighting coefficients are determined through 5-fold cross-validation on the power grid project dataset.

[0072] Optionally, in the above technical solution, the similarity between semantic text vectors is calculated using cosine similarity, and the calculation formula is as follows:

[0073] Where cos(θ) is the final similarity value, which is the angle between the two text vectors, and x and y represent the semantic text vectors of the two elements, respectively. i and y i These represent the feature values ​​of the i-th dimension of the corresponding semantic text vector, which are mainly used in this invention to calculate the similarity between the elements of each stage of an existing project and the top-level strategic elements.

[0074] Optionally, this embodiment also includes a dynamic update mechanism: establishing a long-term operational element database update process, and using this method to re-update and re-evaluate the strategic element database when the state or industry issues new policies and management documents, ensuring the real-time effectiveness of the consistency assessment results.

[0075] This embodiment constructs a multi-dimensional evaluation framework consisting of a semantic layer, a data layer, and a structural layer. At the semantic level, a large language model is used to extract elements from the project to be evaluated, and semantic similarity calculation is performed by combining the top-level strategic element set with BERT embedding vectors. At the indicator level, a small model is used to compare and analyze quantitative indicators. At the structural level, the element coverage rate is calculated. Finally, a weighted fusion is used to obtain a quantitative score for strategic consistency, thereby achieving a comprehensive measurement of the project's strategic response and providing data support and directional guidance for project optimization.

[0076] As described above, the multi-level consistency assessment method for power grid strategy provided in this application innovatively proposes a cross-stage strategic element embedding expression and consistency evaluation method. By extracting elements, embedding semantics, and mapping cross-stages of multi-level texts such as strategy, planning, project, and program, it establishes and optimizes the strategic semantic transmission path between strategic objectives and power grid project execution links, thereby realizing consistency analysis and optimization in the process of strategy implementation, improving the scientific and intelligent level of strategy implementation, and finally constructing a consistency evaluation method for power grid projects and strategies. This method obtains the effectiveness and shortcomings of projects in responding to strategies, providing ideas for project optimization.

[0077] To improve the quality and consistency of strategic element embedding and provide reliable support for strategically oriented projects in complex fields such as power grids, this application provides an embodiment of a multi-level consistency assessment device for power grid strategies, which implements all or part of the aforementioned multi-level consistency assessment method. Referring to Figure 3, the multi-level consistency assessment device specifically includes the following components: a strategic element extraction module 10, used to collect energy-related policy documents and extract strategic elements related to the strategy from these documents to obtain a strategic element library; the strategic elements include standardized terminology, quantitative indicators, responsible entities, and constraints; and a rule analysis module 20, used to collect power grid-related management documents and analyze them according to the entire lifecycle of power grid projects. The project is divided into three phases: reserve, feasibility study, and comprehensive planning. Relevant management documents and strategic element libraries for the power grid project are input into the LLM (Language Model). Using the LLM, based on macro-strategies and pre-defined prompts in energy-related policy documents, the strategic elements and their manifestation rules in the project texts of each phase are analyzed, and a rule table for strategic elements in each phase is output. A phase element extraction module 30 extracts phase elements from the project texts of each phase according to the rule table. Consistency checks are performed on the extracted phase elements, and semantically consistent items are retained to obtain the phase element set for each phase. A result evaluation module 40 evaluates the consistency between the strategic elements and the phase elements of the three phases, and outputs the consistency evaluation result.

[0078] As described above, the power grid strategy multi-level consistency assessment device provided in this application innovatively proposes a cross-stage strategic element embedding expression and consistency evaluation method. By extracting elements, embedding semantics, and mapping cross-stages of multi-level texts such as strategy, planning, project, and schedule, it establishes and optimizes the strategic semantic transmission path between strategic objectives and power grid project execution links. This enables consistency analysis and optimization during the strategy implementation process, improves the scientific and intelligent level of strategy implementation, and ultimately constructs a consistency evaluation method for power grid projects and strategies. It obtains the effectiveness and shortcomings of projects in responding to strategies, providing ideas for project optimization.

[0079] From a hardware perspective, in order to improve the quality and consistency of strategic element embedding and provide reliable support for strategically oriented projects in complex fields such as power grids, this application provides an embodiment of an electronic device for implementing all or part of the multi-level consistency assessment method for power grid strategy. The electronic device specifically includes the following components: a processor, memory, a communication interface, and a bus. The processor, memory, and communication interface communicate with each other via the bus. The communication interface is used to realize information transmission between the multi-level consistency assessment device for power grid strategy and core business systems, user terminals, and related databases. The logic controller can be a desktop computer, tablet computer, or mobile terminal, etc., and this embodiment is not limited to these. In this embodiment, the logic controller can be implemented with reference to the embodiments of the multi-level consistency assessment method for power grid strategy and the embodiments of the multi-level consistency assessment device for power grid strategy, the content of which is incorporated herein, and repeated details are not repeated.

[0080] It is understood that the user terminal may include smartphones, tablet computers, network set-top boxes, portable computers, desktop computers, personal digital assistants (PDAs), in-vehicle devices, smart wearable devices, etc. Among these, the smart wearable devices may include smart glasses, smartwatches, smart bracelets, etc.

[0081] In practical applications, parts of the multi-level consistency assessment method for power grid strategy can be executed on the electronic device side as described above, or all operations can be completed in the client device. The choice can be made based on the processing power of the client device and the limitations of the user's usage scenario. This application does not impose any limitations on this. If all operations are completed in the client device, the client device may further include a processor.

[0082] The aforementioned client device may have a communication module (i.e., a communication unit) that can communicate with a remote server to achieve data transmission with the server. The server may include a server on the task scheduling center side; in other implementation scenarios, it may also include a server on an intermediate platform, such as a server on a third-party server platform that has a communication link with the task scheduling center server. The server may include a single computer device, a server cluster consisting of multiple servers, or a distributed server structure.

[0083] Figure 4 is a schematic block diagram of the system configuration of an electronic device 9600 according to an embodiment of this application. As shown in Figure 4, the electronic device 9600 may include a central processing unit 9100 and a memory 9140; the memory 9140 is coupled to the central processing unit 9100. It is worth noting that Figure 4 is exemplary; other types of structures may also be used to supplement or replace this structure to achieve telecommunications functions or other functions.

[0084] In one embodiment, the functionality of the multi-level consistency assessment method for power grid strategy can be integrated into the central processing unit 9100. The central processing unit 9100 can be configured to perform the following control: Step S101: Collect energy-related policy documents and extract strategic elements related to strategy from the documents to obtain a strategic element library; the strategic elements include standardized terminology, quantitative indicators, responsible entities, and constraints; Step S102: Collect power grid-related management documents and divide the power grid project-related management documents into three stages according to the entire life cycle: reserve, feasibility study, and comprehensive planning; input the power grid project-related management documents and the strategic element library into the LLM large language model, and use the LLM large language model to analyze the strategic elements contained in the project text of each stage and the rules for the manifestation of strategic elements based on the macro strategy and preset prompt words in the energy-related policy documents, and output the rule table of strategic elements in each stage; Step S103: Extract stage elements from the project text of each stage according to the rule table; perform consistency checks on the stage elements extracted from each stage and retain semantically consistent items to obtain the stage element set of each stage; Step S104: Evaluate the consistency between the strategic elements and the stage elements of the three stages, and output the consistency evaluation result.

[0085] As described above, the electronic device provided in this application embodiment innovatively proposes a cross-stage strategic element embedding expression and consistency evaluation method. By extracting elements, embedding semantics, and mapping cross-stages of multi-level texts such as strategy, planning, project, and program, it establishes and optimizes the strategic semantic transmission path between strategic objectives and power grid project execution links. This enables consistency analysis and optimization during the strategy implementation process, improves the scientific and intelligent level of strategy implementation, and ultimately constructs a consistency evaluation method for power grid projects and strategies. It obtains the project's effectiveness and shortcomings in responding to the strategy, providing ideas for project optimization.

[0086] In another embodiment, the power grid strategy multi-level consistency assessment device can be configured separately from the central processing unit 9100. For example, the power grid strategy multi-level consistency assessment device can be configured as a chip connected to the central processing unit 9100, and the power grid strategy multi-level consistency assessment method function can be realized through the control of the central processing unit.

[0087] As shown in Figure 4, the electronic device 9600 may further include: a communication module 9110, an input unit 9120, an audio processor 9130, a display 9160, and a power supply 9170. It is worth noting that the electronic device 9600 does not necessarily include all the components shown in Figure 4; furthermore, the electronic device 9600 may also include components not shown in Figure 4, as can be found in existing technologies.

[0088] As shown in Figure 4, the central processing unit 9100, sometimes also referred to as a controller or operating control, may include a microprocessor or other processor device and / or logic device. The central processing unit 9100 receives input and controls the operation of various components of the electronic device 9600.

[0089] The memory 9140 may be, for example, one or more of a cache, flash memory, hard drive, removable media, volatile memory, non-volatile memory, or other suitable devices. It may store the aforementioned failure-related information, and also store a program for executing that information. The central processing unit 9100 may execute the program stored in the memory 9140 to perform information storage or processing, etc.

[0090] Input unit 9120 provides input to central processing unit 9100. Input unit 9120 may be, for example, a keypad or touch input device. Power supply 9170 provides power to electronic device 9600. Display 9160 displays images and text. Display may be, for example, an LCD display, but is not limited thereto.

[0091] The memory 9140 can be a solid-state memory, such as a read-only memory (ROM), random access memory (RAM), a SIM card, etc. It can also be a memory that retains information even when power is off, can be selectively erased, and contains more data; examples of this type of memory are sometimes referred to as EPROMs. The memory 9140 can also be some other type of device. The memory 9140 includes a buffer memory 9141 (sometimes referred to as a buffer). The memory 9140 may include an application / function storage unit 9142 for storing application programs and function programs or processes for executing the operation of the electronic device 9600 via the central processing unit 9100.

[0092] The memory 9140 may also include a data storage unit 9143 for storing data, such as contacts, digital data, pictures, sounds, and / or any other data used by the electronic device. The driver storage unit 9144 of the memory 9140 may include various drivers for the electronic device for communication functions and / or for performing other functions of the electronic device (such as messaging applications, address book applications, etc.).

[0093] The communication module 9110 is a transmitter / receiver that sends and receives signals via the antenna 9111. The communication module 9110 (transmitter / receiver) is coupled to the central processing unit 9100 to provide input signals and receive output signals, which is the same as in a conventional mobile communication terminal.

[0094] Based on different communication technologies, multiple communication modules 9110 can be configured in the same electronic device, such as cellular network modules, Bluetooth modules, and / or wireless LAN modules. The communication module 9110 (transmitter / receiver) is also coupled to a speaker 9131 and a microphone 9132 via an audio processor 9130 to provide audio output via the speaker 9131 and receive audio input from the microphone 9132, thereby realizing typical telecommunications functions. The audio processor 9130 may include any suitable buffer, decoder, amplifier, etc. Additionally, the audio processor 9130 is coupled to a central processing unit 9100, enabling on-device recording via the microphone 9132 and on-device playback of stored audio via the speaker 9131.

[0095] Embodiments of this application also provide a computer-readable storage medium capable of implementing all steps of the multi-level consistency assessment method for power grid strategy with a server or client as the execution subject in the above embodiments. The computer-readable storage medium stores a computer program, which, when executed by a processor, implements all steps of the multi-level consistency assessment method for power grid strategy with a server or client as the execution subject in the above embodiments. For example, when the processor executes the computer program, it implements the following steps: Step S101: Collect energy-related policy documents and extract strategic elements related to strategy from the documents to obtain a strategic element library; the strategic elements include normative terminology, quantitative indicators, responsible entities, and constraints; Step S102: Collect power grid-related management documents. The relevant management documents for power grid projects are divided into three stages according to the entire life cycle: reserve, feasibility study, and comprehensive planning. The relevant management documents and strategic element library for power grid projects are input into an LLM (Language Model). Using the LLM model, based on macro-strategies and preset prompts in energy-related policy documents, the strategic elements contained in the project texts of each stage and the rules governing their manifestation are analyzed, outputting a rule table for strategic elements at each stage. Step S103: Extract stage elements from the project texts of each stage according to the rule table; perform consistency checks on the extracted stage elements for each stage and retain semantically consistent items to obtain the stage element set for each stage; Step S104: Evaluate the consistency between the strategic elements and the stage elements of the three stages, and output the consistency evaluation results.

[0096] As described above, the computer-readable storage medium provided in this application embodiment innovatively proposes a cross-stage strategic element embedding expression and consistency evaluation method. By extracting elements, embedding semantics, and mapping cross-stages of multi-level texts such as strategy, planning, project, and program, it establishes and optimizes the strategic semantic transmission path between strategic objectives and power grid project execution links, thereby realizing consistency analysis and optimization in the strategic implementation process, improving the scientific and intelligent level of strategic implementation, and finally constructing a consistency evaluation method for power grid projects and strategies. This method obtains the effectiveness and shortcomings of projects in responding to strategies, providing ideas for project optimization.

[0097] The embodiments of this application also provide a computer program product capable of implementing all steps in the multi-level consistency assessment method for power grid strategy, where the execution subject is a server or client, as described in the above embodiments. When executed by a processor, this computer program / instruction implements the steps of the multi-level consistency assessment method for power grid strategy. For example, the computer program / instruction implements the following steps: Step S101: Collect energy-related policy documents and extract strategic elements related to strategy from the documents to obtain a strategic element library; the strategic elements include standardized terminology, quantitative indicators, responsible entities, and constraints; Step S102: Collect power grid-related management documents and divide the power grid project-related management documents according to the entire life cycle. For the three stages of reserve, feasibility study, and comprehensive planning, relevant management documents and strategic element library of the power grid project are input into the LLM large language model. Using the LLM large language model, based on the macro strategy and preset prompt words in energy-related policy documents, the strategic elements contained in the project text of each stage and the rules for the manifestation of strategic elements are analyzed, and the rule table of strategic elements in each stage is output. Step S103: Extract stage elements from the project text of each stage according to the rule table; perform consistency checks on the stage elements extracted from each stage and retain semantically consistent items to obtain the stage element set of each stage; Step S104: Evaluate the consistency between the strategic elements and the stage elements of the three stages, and output the consistency evaluation results.

[0098] As described above, the computer program product provided in this application innovatively proposes a cross-stage strategic element embedding expression and consistency evaluation method. By extracting elements, embedding semantics, and mapping cross-stages of multi-level texts such as strategy, planning, project, and schedule, it establishes and optimizes the strategic semantic transmission path between strategic objectives and power grid project execution links. This enables consistency analysis and optimization during the strategy implementation process, improves the scientific and intelligent level of strategy implementation, and ultimately constructs a consistency evaluation method for power grid projects and strategies. It obtains the project's effectiveness and shortcomings in responding to the strategy, providing ideas for project optimization.

[0099] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, apparatus, 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.

[0100] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (devices), 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, create means for implementing the functions specified in one or more blocks of the flowchart illustrations and / or one or more blocks of the block diagrams.

[0101] 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 that implement the functions specified in one or more flowcharts and / or one or more block diagrams.

[0102] These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process, such that the instructions, which execute on the computer or other programmable apparatus, provide steps for implementing the functions specified in one or more flowcharts and / or one or more block diagrams.

[0103] Specific embodiments have been used to illustrate the principles and implementation methods of this invention. The descriptions of the embodiments above are only for the purpose of helping to understand the method and core ideas of this invention. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this invention. Therefore, the content of this specification should not be construed as a limitation of this invention.

Claims

1. A multi-level consistency assessment method for power grid strategy, characterized in that, The method includes: collecting energy-related policy documents and extracting strategic elements related to strategy from these documents to obtain a strategic element library; the strategic elements include standardized terminology, quantitative indicators, responsible entities, and constraints; collecting power grid-related management documents and dividing them into three stages—reserve, feasibility study, and comprehensive planning—according to the entire life cycle; inputting the power grid project-related management documents and the strategic element library into an LLM (Large Language Model); using the LLM model, based on the macro-strategies and preset prompts in the energy-related policy documents, analyzing the strategic elements contained in the project texts of each stage and the rules governing their manifestation, and outputting a rule table for strategic elements at each stage; extracting stage elements from the project texts of each stage according to the rule table; performing consistency checks on the stage elements extracted from each stage and retaining semantically consistent items to obtain a set of stage elements for each stage; evaluating the consistency between the strategic elements and the stage elements of the three stages, and outputting the consistency evaluation results.

2. The multi-level consistency assessment method for power grid strategy according to claim 1, characterized in that, The steps of extracting strategic elements related to the strategy from the documents and obtaining a strategic element library include: using an LLM (Large Language Model) to perform semantic understanding on energy-related policy documents, identifying and extracting strategic-related data from the documents; using a named entity recognition model to identify core entity information in the extracted strategic-related data, and obtaining a set of strategic elements; the core entity information includes normative terminology, quantitative indicators, responsible entities, and constraints.

3. The multi-level consistency assessment method for power grid strategy according to claim 1, characterized in that, The steps of using the LLM (Large Language Model) to analyze the strategic elements and their manifestation rules in project texts at each stage, based on macro-strategies and pre-set prompts in energy-related policy documents, and outputting rule tables for strategic elements at each stage, include: Project Reserve Stage: Using the LLM and pre-set prompts, analyze how strategic elements are manifested in the project reserve stage, clarify the strategic elements that the macro strategy needs to respond to in the project reserve stage, and generate strategic element rules for the project reserve stage; Project Feasibility Study Stage: Using the LLM and pre-set prompts, analyze the specific elements into which the macro strategy is decomposed, and analyze the specific elements contained within the strategic elements, outputting strategic element rules for the project feasibility study stage; the specific elements include technical indicators, economic benefit assessment, and environmental constraints; Comprehensive Planning Stage: Using the LLM, analyze the integration and optimization of strategic elements of each project in the overall comprehensive plan at this stage, forming strategic element rules for the comprehensive planning stage.

4. The multi-level consistency assessment method for power grid strategy according to claim 1, characterized in that, The step of extracting stage elements from project texts of each stage according to the rule table includes: inputting project texts and the rule table of the reserve stage into a rule matching model, scanning the project texts of the reserve stage using the rule matching model, and extracting stage elements from the project texts of the reserve stage according to the rule table; inputting project texts and the rule base of the feasibility study stage into a named entity recognition model, and extracting stage elements from the project texts of the feasibility study stage by combining the named entity recognition model with regular expressions; inputting project texts and the rule table of the comprehensive planning stage into a rule matching model, scanning the project texts of the comprehensive planning stage using the rule matching model, and extracting stage elements from the project texts of the comprehensive planning stage according to the rule table.

5. The multi-level consistency assessment method for power grid strategy according to claim 1, characterized in that, The step of performing consistency checks on the stage elements extracted from each stage and retaining semantically consistent items to obtain the stage element set for each stage includes: aligning the stage element sets for each stage with data, and using set intersection calculation to identify semantically consistent items and inconsistent elements in the three element sets; converting inconsistent elements into high-dimensional semantic vectors and calculating the cosine similarity between the vectors: if the similarity is greater than a preset threshold, it is determined to be a semantically consistent item; if the similarity is less than the preset threshold, it is determined to be a conflicting item and discarded; merging the semantically consistent items of each stage to form the stage element set for each stage.

6. The multi-level consistency assessment method for power grid strategy according to claim 1, characterized in that, The steps for evaluating the consistency between strategic elements and the stage elements of the three stages, and outputting the consistency evaluation results, include: Semantic level: converting strategic elements and stage elements into high-dimensional semantic vectors, and calculating the cosine similarity between the semantic vectors of strategic elements and the stage elements of each stage; Data level: using a rule matching model to extract quantifiable indicators of strategic elements and stage elements of each stage, calculating whether they are within a threshold range, and calculating the matching score linearly according to the actual degree of deviation; the quantifiable indicators include capacity and investment amount; Structural level: intersecting the set of strategic elements with the set of stage elements of each stage, and calculating the coverage rate of stage elements of each stage of the project; the semantic similarity, matching score, and coverage rate obtained from the three levels are weighted to calculate the final consistency score of the project.

7. The multi-level consistency assessment method for power grid strategy according to claim 1, characterized in that, Also includes: A graph knowledge base is established with strategic elements and stage elements of each stage as nodes, and the evolution chain of strategic elements in each stage is established starting from the initial strategic elements. Using the LLM large language model, the evolution chain and stage elements of each stage are used as input to generate the cross-stage embedded expression of each strategic element. The cross-stage embedded expression is used to optimize the strategic consistency of the three stages.

8. A multi-level consistency assessment device for power grid strategy, characterized in that, The device includes: a strategic element extraction module, used to collect energy-related policy documents and extract strategic elements related to strategy from the documents to obtain a strategic element library; the strategic elements include standardized terminology, quantitative indicators, responsible entities, and constraints; a rule analysis module, used to collect power grid-related management documents and divide the power grid project-related management documents into three stages according to the entire life cycle: reserve, feasibility study, and comprehensive planning; inputting the power grid project-related management documents and the strategic element library into an LLM (Large Language Model), and using the LLM, based on the macro-strategy and preset prompts in the energy-related policy documents, analyzing the strategic elements contained in the project text of each stage and the rules for the manifestation of strategic elements, and outputting a rule table of strategic elements in each stage; a stage element extraction module, used to extract stage elements from the project text of each stage according to the rule table; performing consistency checks on the stage elements extracted from each stage and retaining semantically consistent items to obtain a stage element set for each stage; and a result evaluation module, used to evaluate the consistency between the strategic elements and the stage elements of the three stages, and output the consistency evaluation result.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the steps of the multi-level consistency assessment method for power grid strategy as described in any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the computer program implements the steps of the multi-level consistency assessment method for power grid strategy as described in any one of claims 1 to 7.

Images