Fire coal purchasing decision-making method, system and equipment based on large language model

Abstract

The invention relates to the technical field of fire coal purchase decision support, and particularly discloses a fire coal purchase decision method, system and equipment based on a big language model.The fire coal purchase decision method based on the big language model.The fire coal purchase decision method based on the big language model comprises the steps that target data are acquired; inputting the target data into the large language model, and outputting a fire coal purchase basic data analysis result; according to a fire coal purchasing basic data analysis result, further analysis is carried out through a large language model, and a fuel purchasing strategy direction suggestion is generated; according to a strategy direction selected by a user, constructing a multi-target optimization model, and solving the multi-target optimization model to obtain an optimal combination of purchased coal types, quantity and purchasing routes; according to the optimal combination of the types, the quantity and the routes of the purchased coal, a customized fire coal purchasing scheme is generated through the natural language generation capability of the large language model, the fire coal purchasing decision is intelligent, scientific and efficient, and powerful support is provided for reducing cost, improving efficiency and guaranteeing safety for thermal power enterprises.

Description

Technical Field

[0001] The present invention relates to the technical field of coal procurement decision support, and in particular to a coal procurement decision method, system and device based on a large language model. Background Art

[0002] The main challenge facing current coal procurement decisions is over-reliance on manual experience and judgment, making it difficult to achieve multi-objective collaborative optimization. Traditional decision-making models have three major limitations: First, it is highly subjective, making it impossible to quantitatively assess the interplay of key factors such as coal type structure, fuel costs, transportation cycles, environmental protection indicators, and inventory structure. Currently, coal procurement by thermal power companies relies primarily on manual decision-making based on empirical evidence. This involves manual calculations based on indicators such as future power generation, environmental protection, inventory levels, and available coal types, combined with Excel. Each proposed coal type is considered eligible if it meets average calorific value and sulfur requirements. This approach, when faced with a large number of coal types to be purchased, makes it difficult to guarantee the optimal number of coal types to be purchased. Even when the number of coal types is determined, manual calculations cannot guarantee the optimal purchase quantity for each individual coal type.

[0003] Second, response is delayed, making it difficult to promptly address dynamic changes in coal prices, transportation capacity, and weather conditions. Coal procurement and coal blending are mutually reinforcing. Scientific coal procurement is the key to ensuring the effectiveness of future coal blending. An irrational distribution of coal types and qualities will inevitably lead to problems such as inability to support high loads, exceeding environmental standards, and using high-calorific-value coal for low loads. However, manual decision-making cannot track market changes in real time, leading to delayed adjustments to procurement strategies.

[0004] Third, the strategy is narrowly focused, making it difficult to balance economic and environmental considerations while ensuring supply. Manual procurement can only provide a single procurement plan, not a comprehensive one. This lack of flexibility and speed in procurement is unreliable. Traditional approaches typically prioritize minimizing procurement costs while neglecting the comprehensive optimization of multiple factors, including combustion efficiency, environmental compliance, and transportation efficiency.

[0005] At present, the technologies related to coal procurement optimization mainly include the following categories: Procurement guidance models based on operations research optimization models: For example, Chinese patent CN108846538A discloses a fuel procurement guidance model for large CFB boilers. Although this model achieves a certain degree of optimization, it lacks the ability to flexibly handle complex constraints and does not integrate the latest artificial intelligence technologies.

[0006] Procurement optimization system based on virtual blending: For example, Chinese patent CN202410797510 discloses a cascade coal procurement optimization system based on virtual blending. Although it takes into account the factors of coal blending, it does not fully utilize the natural language processing and knowledge reasoning capabilities of the large language model.

[0007] Large-model-based procurement applications, such as ChatGPT, are used in coal enterprise procurement management, primarily for procurement demand forecasting, price analysis and evaluation, and compliance risk management. While these applications incorporate large language models, they lack deep integration with specialized operational optimization models, making it difficult to achieve multi-objective optimization for complex procurement decisions.

[0008] Existing technologies lack multi-objective collaborative optimization capabilities and typically only optimize for a single objective, such as minimizing procurement costs. This makes it difficult to simultaneously address multiple objectives, including economy, safety, efficiency, environmental protection, transportation cycles, and inventory ratios. The level of intelligence is limited, and the system lacks natural language interaction and knowledge reasoning capabilities, making it impossible to conduct efficient human-machine collaborative decision-making with users. Data integration and processing capabilities are insufficient, making it difficult to effectively integrate multi-source heterogeneous data, including complex data such as equipment parameters, inventory information, supplier information, coal prices, and transportation routes. Dynamic adaptability is poor, making it difficult to respond in real time to dynamic factors such as coal price fluctuations, changes in transportation capacity, and weather conditions, resulting in a lack of the ability to respond quickly to market changes. Summary of the Invention

[0009] Based on this, it is necessary to provide a coal procurement decision-making method, system and equipment based on a large language model to address the above technical problems. By integrating the large language model with the operations optimization model, the coal procurement decision-making is made intelligent, scientific and efficient, providing strong support for thermal power companies to reduce costs, improve efficiency and ensure safety.

[0010] In a first aspect, a coal purchasing decision-making method based on a large language model is provided, comprising the following steps: Obtain target data, including boiler equipment design parameters, minimum coal inventory indicators, basic information for fuel procurement plan formulation, coal supplier and price information, historical coal procurement information, and coal transportation routes and cycles; Input the target data into a pre-trained large language model, and analyze it through the large language model according to preset prompt words, and output the analysis results of the basic data of coal procurement; Based on the coal procurement basic data analysis results, policies, regulations and industry standards, further analysis is performed through the large language model to generate fuel procurement strategy direction recommendations; Based on the strategy direction selected by the user from the fuel procurement strategy direction recommendations, combined with the planned total procurement volume and the estimated procurement ratios of high and low ash melting points output by the basic calculation, a multi-objective optimization model is constructed, and the multi-objective optimization model is solved to obtain the optimal combination of procurement coal type, procurement quantity, and procurement route; Based on the optimal combination of the purchased coal type, purchase quantity and purchase route, a customized coal purchasing plan is generated through the natural language generation capability of the large language model.

[0011] In some examples, the analyzing by the large language model and outputting the analysis results of the basic coal procurement data include: Verifying and cleaning the target data and generating a data verification report; Extract key information from the verified and cleaned target data and convert it into structured data; Analyzing the correlation between different data items in the structured data to obtain a preliminary purchasing decision; Based on the abnormal data items in the data verification report, the impact on the procurement decision is evaluated, and the basic data analysis results of coal procurement are obtained based on the evaluation results.

[0012] In some examples, the large language model is further analyzed based on the coal procurement basic data analysis results, policies, regulations, and industry standards to generate fuel procurement strategy direction recommendations, including: Obtain information on environmental protection policies, energy policies, and coal industry standards to analyze policy constraints in the procurement environment; Obtain industry standards based on power plant type and unit capacity parameters to develop a procurement strategy that complies with industry standards; Analyze historical procurement data and current market information to predict market trends in coal prices and supply tightness; Generate multiple fuel procurement strategy direction recommendations based on the policy constraints of the procurement environment, procurement strategies that comply with industry standards, and market trends in coal prices and supply tightness.

[0013] In some examples, based on the strategy direction selected by the user from the fuel procurement strategy direction recommendations, combined with the planned total procurement volume and the estimated procurement ratios of high and low ash melting points output by the basic calculation, a multi-objective optimization model is constructed, and the multi-objective optimization model is solved to obtain the optimal combination of procurement coal type, procurement quantity, and procurement route, including: Calculate the total planned purchase volume and the estimated purchase ratios of high and low ash melting point coals based on the power generation plan and power supply coal consumption index information; Defining decision variables of the large language model based on multiple fuel procurement strategy direction suggestions; Construct an objective function based on the fuel procurement strategy direction recommended by the user; Setting constraints for the large language model based on boiler equipment parameters, inventory requirements, and environmental standards; The large language model is solved to generate an optimal combination of purchased coal type, purchase quantity and purchase route.

[0014] In some examples, generating a customized coal procurement plan based on the optimal combination of the purchased coal type, purchase quantity, and purchase route through the natural language generation capability of the large language model includes: Generate a detailed data table based on the solution of the multi-objective optimization model, wherein the detailed data table includes the purchase quantity, supplier, price, transportation method, and arrival time of each type of coal; Generate a structured procurement report based on preset prompt words, including current coal procurement and inventory status, next month's demand status, and next month's fuel procurement plan; Present key data from the coal procurement plan in chart form; Output the generated detailed data tables and structured procurement reports to the designated path and upload them to the company's document management system.

[0015] In some examples, this also includes: Through the natural language interactive interface, real-time interaction with users is achieved to receive users' evaluation, adjustment and confirmation of the generated coal procurement plan, and the parameters and strategy recommendations of the large language model are optimized based on user feedback.

[0016] In some examples, the natural language interactive interface is used to implement real-time interaction with the user to receive user evaluation, adjustment, and confirmation of the generated coal procurement plan, and optimize the parameters and strategy recommendations of the large language model based on user feedback, including: Providing a text-based or voice-based interactive interface to communicate with the user in natural language; Analyze the natural language text input by users to identify their real needs and intentions; Adjust and optimize the coal procurement plan based on user feedback and needs.

[0017] In some examples, this also includes: Query expertise related to coal procurement through an interactive interface and receive accurate answers based on an external knowledge base; Record the user's interaction history and decision preferences, and continuously optimize the parameters and strategy recommendations of the large language model through continuous learning.

[0018] Secondly, a coal procurement decision-making system based on a large language model is provided, including: An acquisition module is used to acquire target data, including boiler equipment design parameters, minimum coal inventory indicators, basic information for fuel procurement plan formulation, coal supplier and price information, historical coal procurement information, and coal transportation routes and cycles; An analysis module is used to input the target data into a pre-trained large language model, and analyze it through the large language model according to preset prompt words, and output the analysis results of the basic coal procurement data; A generation module, configured to generate fuel procurement strategy direction recommendations based on the coal procurement basic data analysis results, policies, regulations and industry standards, and further analyze them through the large language model; A solution module is used to construct a multi-objective optimization model based on the strategy direction selected by the user from the fuel procurement strategy direction recommendations, combined with the planned total procurement volume and the estimated procurement ratio of high and low ash melting points output by the basic calculation, and solve the multi-objective optimization model to obtain the optimal combination of procurement coal type, procurement quantity, and procurement route; The decision-making module is used to generate a customized coal procurement plan based on the optimal combination of the purchased coal type, purchase quantity and purchase route through the natural language generation capability of the large language model.

[0019] In a third aspect, a computer device is provided, comprising a processor and a computer program. When the processor executes the computer program, the coal purchasing decision-making method based on the large language model according to the first aspect is implemented.

[0020] By adopting the embodiments of the present invention, by integrating the large language model with the operations optimization model, intelligent, scientific and efficient coal procurement decisions are achieved, providing strong support for thermal power companies to reduce costs, improve efficiency and ensure safety. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Other features, objects and advantages of the present invention will become more apparent upon reading the detailed description of non-limiting embodiments made with reference to the following drawings: Figure 1 A flowchart of a coal purchasing decision-making method based on a large language model provided by one embodiment of the present invention; Figure 2 A flowchart of a coal purchasing decision-making method based on a large language model provided in another embodiment of the present invention; Figure 3 A detailed flowchart of the multi-objective optimization model construction and solution of the present invention; Figure 4A structural block diagram of a coal purchasing decision-making system based on a large language model provided by an embodiment of the present invention; Figure 5 This is a structural block diagram of a computer device provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0022] The present invention will be further described in detail below with reference to the embodiments and accompanying drawings. It should be understood that the specific embodiments described herein are intended only to illustrate the relevant invention and are not intended to limit the invention. It should also be noted that, for ease of description, only portions relevant to the invention are shown in the accompanying drawings.

[0023] It should be noted that, in the absence of conflict, the embodiments of the present invention, that is, the features of the embodiments, can be combined with each other. The present invention will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0024] The following describes in detail the coal procurement decision-making method, system and device based on a large language model according to an embodiment of the present invention with reference to the accompanying drawings.

[0025] Figure 1 FIG. 1 is a flow chart of a coal purchasing decision-making method based on a large language model according to an embodiment of the present invention. Figure 1 As shown, combined with Figure 2 According to an embodiment of the present invention, a coal purchasing decision-making method based on a large language model includes the following steps: S101: Acquire target data, which includes boiler equipment design parameters, minimum coal inventory indicators, basic information for fuel procurement plan formulation, coal supplier and price information, historical coal procurement information, and coal transportation routes and cycle information.

[0026] S102: Input the target data into a pre-trained large language model, and analyze the target data through the large language model according to preset prompt words, and output the analysis results of the basic coal procurement data.

[0027] In one embodiment of the present invention, analysis is performed using the large language model to output analysis results of basic coal procurement data, including: verifying and cleaning the target data to generate a data verification report; extracting key information from the verified and cleaned target data and converting it into structured data; analyzing the correlation between different data items in the structured data to obtain a preliminary procurement decision; evaluating the impact on the procurement decision based on abnormal data items in the data verification report, and obtaining analysis results of basic coal procurement data based on the evaluation results.

[0028] S103: Based on the analysis results of the coal procurement basic data, policies, regulations and industry standards, further analysis is performed through the large language model to generate fuel procurement strategy direction recommendations.

[0029] In one embodiment of the present invention, based on the analysis results of the basic coal procurement data, policies, regulations and industry standards, the large language model is further analyzed to generate fuel procurement strategy direction recommendations, including: obtaining environmental protection policies, energy policies, and coal industry standard information to analyze the policy constraints of the procurement environment; obtaining industry standards based on power plant type and unit capacity parameters to obtain procurement strategies that comply with industry norms; analyzing historical procurement data and current market information to predict market trends in coal prices and supply tightness; generating multiple fuel procurement strategy direction recommendations based on the policy constraints of the procurement environment, procurement strategies that comply with industry norms, and market trends in coal prices and supply tightness.

[0030] S104: Based on the strategy direction selected by the user from the fuel procurement strategy direction suggestions, combined with the planned total procurement volume and the estimated procurement ratios of high and low ash melting points output by the basic calculation, a multi-objective optimization model is constructed, and the multi-objective optimization model is solved to obtain the optimal combination of procurement coal types, procurement quantities, and procurement routes.

[0031] In one embodiment of the present invention, based on the strategy direction selected by the user from the fuel procurement strategy direction suggestions, combined with the planned procurement total amount and the expected procurement ratio of high and low ash melting points output by the basic operation, a multi-objective optimization model is constructed, and the multi-objective optimization model is solved to obtain the optimal combination of procurement coal types, procurement quantities and procurement routes, including: calculating the planned procurement total amount and the expected procurement ratio basic data of high and low ash melting point coal types based on the power generation plan and power supply coal consumption index information; defining the decision variables of the large language model based on multiple fuel procurement strategy direction suggestions; constructing the objective function based on the fuel procurement strategy direction suggestions selected by the user; setting the constraints of the large language model based on boiler equipment parameters, inventory requirements, and environmental protection standards; solving the large language model to generate the optimal combination of procurement coal types, procurement quantities and procurement routes.

[0032] S105: Generate a customized coal purchasing plan based on the optimal combination of the purchased coal type, purchase quantity and purchase route through the natural language generation capability of the large language model.

[0033] In one embodiment of the present invention, a customized coal procurement plan is generated based on the optimal combination of the purchased coal types, purchase quantities, and purchase routes through the natural language generation capability of the large language model, including: generating a detailed data table based on the solution results of the multi-objective optimization model, wherein the detailed data table includes the purchase quantity, supplier, price, transportation method, and arrival time of each coal type; generating a structured procurement report based on preset prompt words, wherein the structured procurement report includes the current coal procurement and inventory status, next month's demand status, and next month's fuel procurement plan; displaying key data in the coal procurement plan in the form of charts; outputting the generated detailed data table and structured procurement report to a specified path, and uploading them to the enterprise's document management system.

[0034] In one embodiment of the present invention, the coal purchasing decision-making method based on the large language model also includes: realizing real-time interaction with the user through a natural language interaction interface to receive the user's evaluation, adjustment and confirmation of the generated coal purchasing plan, and optimizing the parameters and strategy recommendations of the large language model based on user feedback.

[0035] In this example, the system implements real-time interaction with the user through a natural language interactive interface to receive user evaluation, adjustment, and confirmation of the generated coal procurement plan, and optimizes the parameters and strategy recommendations of the large language model based on user feedback. This includes: providing a text-based or voice-based interactive interface to communicate with the user through natural language; analyzing the natural language text input by the user to identify the user's real needs and intentions; and adjusting and optimizing the coal procurement plan based on user feedback and needs. The coal procurement decision-making method based on a large language model is characterized by further including: querying professional knowledge related to coal procurement through an interactive interface and providing accurate answers based on an external knowledge base; recording the user's interaction history and decision preferences, and continuously optimizing the parameters and strategy recommendations of the large language model through continuous learning.

[0036] For example, combining Figure 3 As shown, in this embodiment, data acquisition and integration obtains various types of data related to coal procurement through multiple channels, including: 1. Equipment design parameters: Key parameters such as boiler model, rated evaporation capacity, design coal type, and burner type are obtained from the power plant's equipment management system. These parameters will be used to evaluate the adaptability of different coal types to unit operation.

[0037] 2. Minimum inventory index: Obtain the minimum inventory requirements for each type of coal from the fuel management system to ensure that the procurement plan can meet the safety stock demand.

[0038] 3. Basic information for formulating fuel procurement plans: Obtain information such as monthly power generation plans, power supply coal consumption indicators, and blending plans from the production planning department to provide a basis for formulating procurement plans.

[0039] 4. Supplier information: Obtain basic information, supply capacity, quality assurance system, historical performance of each coal supplier from the supplier management system for supplier evaluation and selection.

[0040] 5. Coal price information: Through data interfaces with coal trading platforms and price index publishing agencies, real-time market price information for various types of coal is obtained, including different pricing methods such as pithead price, ex-factory price, and price including tax.

[0041] 6. Transportation route and cycle information: Obtain information such as transportation routes, transportation methods (rail, road, water), transportation cycles, and transportation costs from each coal production area to the power plant from the logistics management system to optimize transportation plans.

[0042] 7. Historical coal purchase information: Obtain purchase records for the past 12 months from the procurement management system, including purchased coal types, quantities, prices, suppliers, transportation methods, etc., for analyzing procurement trends and predicting market changes.

[0043] This data interacts with external systems through the data docking module to ensure real-time and accuracy. The data integration module converts these multi-source heterogeneous data into a unified structured format and stores it in the system's database for subsequent analysis.

[0044] In this embodiment, a pre-trained large language model, such as the DeepSeek-R1 model, is used to analyze the integrated data, as follows: Data Verification and Cleansing: The large model validates input data based on a pre-set prompt: "You are a power plant fuel coal procurement expert. Please summarize and analyze the verification results of the above data from the perspective of fuel coal procurement, and analyze the impact of data items with abnormal verification results." It identifies missing values, outliers, and inconsistent data and generates a data verification report. For example, the large model might discover that the calorific value data of coal provided by a supplier deviates by more than 10% from historical records and mark it as an abnormal data item.

[0045] Data feature extraction: The large model uses natural language processing technology to extract key information from unstructured data (such as supplier qualification documents and transportation contract texts), such as supplier credit rating and transportation contract terms, and converts it into structured data for subsequent analysis.

[0046] Data association analysis: The large model analyzes the correlation between different data items. For example, it analyzes the correlation between coal prices and quality indicators such as calorific value and sulfur content to identify coal types with abnormal prices; it analyzes the relationship between transportation costs, transportation distance, and transportation methods to find the most economical transportation plan.

[0047] Abnormal Data Impact Assessment: For abnormal data items discovered during verification, the large-scale model evaluates their potential impact on procurement decisions. For example, if the sulfur content of a batch of coal is abnormally high, the large-scale model will analyze the environmental compliance risks and increased desulfurization costs that may result, and will indicate this in the data analysis results.

[0048] The output of the large-scale model data analysis is a basic data analysis report on coal procurement, including data quality assessment, key data characteristics, abnormal data and its impact, etc., providing a basis for the subsequent generation of procurement strategy recommendations.

[0049] In this embodiment, the procurement strategy recommendation generation module generates multiple procurement strategy direction recommendations based on the data analysis results of the large model and information such as policies and industry standards in the external knowledge base. The specific steps are as follows: Policy and Regulatory Compliance Analysis: The large-scale model draws on the latest environmental protection policies, energy policies, coal industry standards, and other information from external knowledge bases to analyze the policy constraints of the current procurement environment. For example, if the latest environmental protection policy requires that the sulfur content of coal burned by power plants does not exceed 0.8%, the large-scale model will incorporate this requirement into the constraints of the procurement strategy.

[0050] Industry standard matching: The large model obtains corresponding industry standards from external knowledge bases based on parameters such as power plant type and unit capacity, such as the "Guidelines for Coal Quality Used in Coal-fired Power Plants", to ensure that procurement strategies comply with industry regulations.

[0051] Market Trend Analysis: The large-scale model analyzes historical procurement data and current market information to predict market trends such as coal price trends and supply constraints, providing reference for procurement strategies. For example, if the large-scale model predicts that the price of a certain type of coal will increase by 10% within the next month, it will recommend increasing the purchase volume of that type of coal in advance.

[0052] Strategic direction generation: Based on the above analysis, the large model generates a variety of procurement strategy direction recommendations, including: The lowest unit price of standard coal entering the factory (focusing on cost optimization): With the main goal of reducing procurement costs, priority is given to low-priced and cost-effective coal types, while considering factors such as transportation costs and quality indicators.

[0053] Optimal unit energy consumption (maximum thermal efficiency): With the main goal of improving unit thermal efficiency and reducing coal consumption, priority is given to coal with high calorific value and good combustion performance, even if the price is relatively high.

[0054] Optimal environmental indicators (ensuring emission compliance): With the main goal of meeting environmental protection requirements and reducing pollutant emissions, priority is given to clean coal with low sulfur and low ash content. This may increase procurement costs but reduce environmental protection treatment costs.

[0055] Shortest delivery cycle (select the best transportation route and shortest transportation cycle): With the main goal of ensuring timely supply of coal and reducing inventory pressure, priority is given to suppliers with short transportation distances and reliable transportation methods, which may increase transportation costs.

[0056] Optimal inventory structure (optimal ratio of high and low ash melting point coal): With the main goals of optimizing inventory structure and improving inventory turnover efficiency, the optimal ratio of high and low ash melting point coal types is determined based on the unit operating requirements and inventory status.

[0057] The large model generates a strategy suggestion based on the preset prompt "Fuel procurement strategy direction recommendation: You are a power plant fuel coal procurement expert. Based on the above data, recommend a procurement direction among the lowest unit price of standard coal entering the plant (focusing on cost optimization), optimal unit energy consumption (maximizing thermal efficiency), optimal environmental indicators (ensuring emission compliance), shortest delivery cycle (improving supply chain responsiveness), and optimal inventory structure (optimizing inventory structure). The model also provides an explanation (only the reasoning is required, no suggestions are required)." and analyzes the advantages and disadvantages of each strategy direction for the user to choose.

[0058] In this embodiment, the multi-objective optimization model construction module constructs a multi-objective optimization model based on the user-selected strategy direction and information such as the planned total purchase volume and the expected purchase ratio of high and low ash melting points output by the basic calculation. The specific steps are as follows: Basic Calculations: Based on information such as the power generation plan and coal consumption indicators, basic data such as the planned total purchase volume and the estimated purchase ratio of high- and low-ash melting point coals are calculated. For example, based on the monthly power generation plan and coal consumption indicators, the total required standard coal volume can be calculated. The approximate ratio of high- and low-ash melting point coals can then be determined based on historical blending data and unit characteristics.

[0059] Decision variable definition: Define the model's decision variables based on the procurement strategy. For example, in a cost optimization strategy, the decision variables might include the purchase quantity of each type of coal, supplier selection, and transportation method selection. In an environmental performance optimization strategy, the decision variables might include quality indicators such as the sulfur content and ash content of each type of coal.

[0060] Objective function construction: Based on the strategy direction selected by the user, the corresponding objective function is constructed. For example: Cost optimization objective function: Minimize the unit price of standard coal entering the factory, while considering factors such as procurement cost, transportation cost, and inventory cost.

[0061] Thermal efficiency maximization objective function: maximize the thermal efficiency of the unit, mainly considering the combustion characteristics of coal such as calorific value and volatile matter.

[0062] The optimal objective function of environmental protection indicators: minimize pollutant emissions, mainly considering indicators such as sulfur content, ash content, and nitrogen content of coal.

[0063] Constraint setting: Set the model's constraints based on factors such as equipment parameters, inventory requirements, and environmental standards, including: Supply capacity constraint: the maximum supply capacity limit of each supplier.

[0064] Quality index constraints: The quality indicators of coal such as calorific value, sulfur content, ash content, etc. must meet the unit operation and environmental protection requirements.

[0065] Transport capacity constraint: the maximum transport capacity limit of a transport route.

[0066] Inventory constraints: Purchase quantities must meet minimum inventory requirements while not exceeding maximum inventory capacity.

[0067] Policy and regulatory constraints: must comply with the latest environmental protection policies, energy policies and other requirements.

[0068] Multi-objective optimization: Use an appropriate multi-objective optimization algorithm to solve the model and generate a set of optimal solutions that represent the trade-offs between different objectives. Users can select the most satisfactory solution from these optimal solutions based on their actual needs.

[0069] The coal procurement decision-making method based on a large language model in this embodiment of the present invention can simultaneously balance multiple objectives, including economic efficiency, safety, energy efficiency, environmental protection, transportation cycle, and inventory ratio, to generate a procurement plan with the best overall performance, avoiding the limitations of traditional methods that rely on single-objective optimization. By combining the large language model with an operations research optimization model, an end-to-end solution is achieved, from natural language requirements to mathematical model solutions. Leveraging the large language model's natural language understanding and knowledge reasoning capabilities, it provides intelligent decision support, understanding complex user requirements and generating procurement strategy recommendations that meet those requirements. The natural language interactive interface significantly improves the system's usability and interaction efficiency. It can effectively integrate multi-source heterogeneous data, including complex data such as equipment parameters, inventory information, supplier information, coal prices, and transportation routes. Through the analytical capabilities of the large model, it extracts key information, providing comprehensive data support for decision-making. It can also perceive dynamic factors such as market changes, capacity fluctuations, and weather conditions in real time, allowing for rapid adjustments to procurement strategies to ensure the optimal procurement plan. By connecting to external data systems, real-time data updates and dynamic model adjustments are possible.

[0070] In addition, it integrates a rich external knowledge base, including policies and regulations, industry standards, coal type information, unit information, etc., to provide comprehensive knowledge support for decision-making and improve the scientific nature and compliance of decision-making. The knowledge reasoning ability of the large model can apply this knowledge to specific procurement scenarios and provide more targeted suggestions. Through automated data processing and model solving, the workload of manual analysis and calculation is greatly reduced, and decision-making efficiency is improved. At the same time, the precise solution of the multi-objective optimization model and the intelligent analysis of the large model ensure the quality and accuracy of decision-making. By optimizing the procurement portfolio, transportation plan and inventory structure, the procurement cost and the whole life cycle cost are effectively reduced. Through the comprehensive analysis of factors such as supplier risk, market risk, and environmental risk, potential risks are identified and assessed in advance, risk response strategies are provided, and the company's risk management and control capabilities are enhanced.

[0071] In summary, the coal procurement decision-making method based on the large language model according to the embodiment of the present invention realizes the intelligent, scientific and efficient coal procurement decision-making by integrating the large language model with the operations optimization model, providing strong support for thermal power companies to reduce costs, improve efficiency and ensure safety.

[0072] Figure 4 FIG is a structural block diagram of a coal purchasing decision system based on a large language model according to an embodiment of the present invention. Figure 4 As shown, a coal purchasing decision system based on a large language model according to an embodiment of the present invention includes: an acquisition module 410, an analysis module 420, a generation module 430, a solution module 440 and a decision module 450, wherein: Acquisition module 410, for acquiring target data, including boiler equipment design parameters, minimum coal inventory indicators, basic information for fuel procurement plan formulation, coal supplier and price information, historical coal procurement information, and coal transportation routes and cycles; The analysis module 420 is configured to input the target data into a pre-trained large language model, analyze the target data using the large language model according to preset prompt words, and output an analysis result of the basic coal procurement data; A generation module 430 is configured to generate fuel procurement strategy direction recommendations based on the coal procurement basic data analysis results, policies, regulations and industry standards, and further analyze the large language model; A solution module 440 is configured to construct a multi-objective optimization model based on the strategy direction selected by the user from the fuel procurement strategy direction recommendations, combined with the planned total procurement volume and the estimated procurement ratios of high and low ash melting points output by the basic calculations, and solve the multi-objective optimization model to obtain the optimal combination of procurement coal type, procurement quantity, and procurement route; The decision module 450 is used to generate a customized coal purchasing plan based on the optimal combination of the purchased coal type, purchase quantity and purchase route through the natural language generation capability of the large language model.

[0073] Specifically, data acquisition is responsible for obtaining coal procurement-related data from multiple data sources, including: Equipment parameter interface: connects to the power plant's equipment management system to obtain equipment design parameters such as boiler model, rated evaporation capacity, designed coal type, and burner type.

[0074] Inventory management interface: connects to the fuel management system to obtain the minimum inventory indicators and current inventory data of various types of coal.

[0075] Production planning interface: Connects to the production planning department's system to obtain basic information for fuel procurement planning, such as monthly power generation plan, power supply coal consumption indicators, and blending plan.

[0076] Supplier management interface: Connect to the supplier management system to obtain basic information, supply capacity, quality assurance system, historical performance of each coal supplier, etc.

[0077] Price information interface: Data interface with coal trading platforms and price index publishing agencies to obtain real-time market price information of various types of coal.

[0078] Logistics management interface: Connect with the logistics management system to obtain information such as transportation routes, transportation methods, transportation cycles, and transportation costs from each coal production area to the power plant.

[0079] Procurement history interface: connects to the procurement management system to obtain procurement records for the past 12 months, including purchased coal type, quantity, price, supplier, transportation method, etc.

[0080] These interfaces can be web service interfaces based on standard protocols such as HTTP / REST and SOAP, or direct data access interfaces based on database connections, ensuring efficient data acquisition and real-time updates.

[0081] Data integration converts heterogeneous data obtained from multiple data sources into a unified structured format for subsequent processing, including: Data cleaning component: cleans the acquired data, removes noise, processes missing values ​​and outliers, and improves data quality.

[0082] Data conversion component: Converts data in different formats and encodings into a unified structured format, such as JSON or XML.

[0083] Data integration component: Integrates data from different data sources into a unified data model to resolve data conflicts and inconsistencies.

[0084] Data storage component: This component stores the consolidated data in the system's database for use by subsequent modules. This database can be a relational database (such as MySQL or Oracle) or a NoSQL database (such as MongoDB). The appropriate storage method should be selected based on the data characteristics and processing requirements.

[0085] Large model analysis is responsible for analyzing and processing the integrated data, including: Large language model: Use pre-trained large language models, such as the DeepSeek-R1 model, which has powerful natural language understanding, knowledge reasoning, and generation capabilities.

[0086] Prompt word management component: manages and maintains prompt words used for different tasks, such as data verification prompt words, strategy recommendation prompt words, report generation prompt words, etc., to ensure that the large model can accurately understand user needs and generate output that meets the requirements.

[0087] Model calling interface: Provides a unified interface to facilitate other modules to call the functions of the large model, such as data analysis, strategy recommendation generation, report generation, etc.

[0088] Model output processing component: processes and parses the output results of the large model, extracts key information and converts it into structured data for use by subsequent modules.

[0089] The large model analysis module can be deployed on a high-performance computing server, using GPU acceleration to improve processing efficiency and ensure that the system can respond to user requests in real time.

[0090] The external knowledge base stores various types of knowledge related to coal procurement, providing background knowledge support for large-scale model analysis, including: Policy and regulation knowledge base: stores the latest information on environmental protection policies, energy policies, coal industry standards, etc. to ensure the compliance of procurement decisions.

[0091] Industry Standard Knowledge Base: stores industry standards such as the "Guidelines for Coal Quality Used in Coal-fired Power Plants" to provide a basis for coal type selection and quality assessment.

[0092] Coal type characteristics knowledge base: stores information such as the physical and chemical properties, combustion characteristics, and applicable scenarios of various types of coal to help evaluate the adaptability of different coal types to unit operation.

[0093] Supplier evaluation knowledge base: stores knowledge such as supplier evaluation indicators, evaluation methods, risk factors, etc., providing a reference for supplier selection.

[0094] Transportation knowledge graph: Build a knowledge graph of the coal transportation network, including information such as transportation routes, transportation methods, transportation costs, and transportation time, to support transportation plan optimization.

[0095] The external knowledge base module can store knowledge in various forms such as knowledge graphs, relational databases or document databases to ensure efficient retrieval and application of knowledge.

[0096] Multi-objective optimization is responsible for building and solving multi-objective optimization models to generate the optimal procurement plan, including: Model building component: Based on the strategy direction and basic data selected by the user, build the corresponding multi-objective optimization model and define decision variables, objective functions and constraints.

[0097] Algorithm library: Integrates a variety of multi-objective optimization algorithms and selects the appropriate algorithm to solve the problem according to the characteristics of the problem.

[0098] Model solving component: Use the selected algorithm to solve the model and generate a set of optimal solutions for users to choose.

[0099] Result analysis component: Analyzes and evaluates the solution results, provides performance indicators, risk analysis, sensitivity analysis and other information of the solution, and helps users make decisions.

[0100] The multi-objective optimization module can be implemented using mathematical modeling tools such as MATLAB and Python, or a dedicated optimization solution engine can be developed to ensure the accuracy of the model and the efficiency of the solution.

[0101] Solution Generation: Generate customized procurement solutions based on the solution results of the multi-objective optimization model, including: Excel generation component: converts the solution results into a detailed data table in Excel format, including detailed information such as the purchase quantity, supplier, price, transportation method, arrival time, etc. of each type of coal.

[0102] Word generation component: Utilizes the natural language generation capabilities of the large model to generate structured procurement reports, including current coal procurement and inventory status, next month's demand, next month's fuel procurement plan, and other content.

[0103] Chart generation displays key data in the procurement plan in the form of charts, such as bar charts, line charts, scatter charts, etc., to enhance the visualization of the plan.

[0104] The solution output will output the generated Excel tables and Word reports to the specified path and automatically upload them to the company's document management system for easy reference and use by relevant departments.

[0105] The solution generation module can be integrated with Microsoft Office or wps components or implemented using open source office document generation libraries (such as Python's openpyxl and python-docx libraries) to ensure that the generated documents meet corporate standards.

[0106] The interactive interface provides an interface for users to interact with the system, including: Natural language interaction component: provides a text-based or voice-based interactive interface, allowing users to communicate with the system through natural language.

[0107] User intent understanding component: Analyzes the natural language text input by the user, identifies the user's real needs and intentions, and converts them into instructions that the system can understand.

[0108] Solution display component: displays the generated procurement solutions in an intuitive way, including tables, charts, reports, etc., to facilitate users to understand and compare the advantages and disadvantages of different solutions.

[0109] User feedback collection component: collects users' evaluation, suggestions and adjustment requirements for the generated solutions to provide a basis for system optimization.

[0110] Knowledge query component: supports users to query professional knowledge related to coal procurement, such as coal type characteristics, environmental protection policies, market trends, etc., and provides accurate answers.

[0111] The interactive interface module can be implemented in various forms such as web applications, desktop applications or mobile applications, ensuring that users can use the system conveniently on different devices.

[0112] Data connection is responsible for data exchange and synchronization with external data systems, including: Data lake connection component: connects with the enterprise's data lake (ERP, CRM, etc.) to obtain and synchronize relevant data.

[0113] Meteorological docking component: connects to the meteorological system, obtains meteorological data, and analyzes the impact of meteorological conditions on coal production, transportation, and inventory.

[0114] Plant-side intelligent fuel system docking component: docks with the plant-side intelligent fuel system to obtain real-time information such as coal quality data and inventory data, and supports dynamic optimization of procurement plans.

[0115] Data synchronization strategy component: defines the strategy and frequency of data synchronization to ensure that the data in the system is consistent with the external system and improve the real-time and accuracy of the data.

[0116] The data docking module can be implemented using ETL tools, data integration platforms or custom-developed data interfaces to ensure efficient docking with external systems.

[0117] The practical application process of the present invention is described below through a specific application scenario example.

[0118] A thermal power plant has four 300MW coal-fired generating units, designed to use bituminous coal from northern Shanxi. Its current inventory is 150,000 tons, with a minimum inventory requirement of 100,000 tons. According to the monthly production plan, next month's planned power generation is 360 million kWh, with a target coal consumption of 300 grams per kWh. Currently, coal market prices are fluctuating significantly, and environmental protection authorities are requiring the implementation of new emission standards starting next month, limiting the sulfur content of coal to no more than 0.8%. The power plant needs to develop a coal procurement plan for the following month to ensure it meets production needs, environmental requirements, and control procurement costs.

[0119] 1. Data acquisition and integration The equipment parameter interface obtains information such as boiler model and designed coal type.

[0120] The inventory management interface obtains the current inventory coal quantity and minimum inventory requirements.

[0121] The production planning interface obtains the monthly power generation plan and power supply coal consumption target.

[0122] The price information interface obtains the market prices of various types of coal, including northern Shanxi bituminous coal (Qnet,ar=5500kcal / kg, S=0.6%, ex-factory price 650 yuan / ton), Shaanxi coal (Qnet,ar=5000kcal / kg, S=0.7%, ex-factory price 580 yuan / ton), Inner Mongolia coal (Qnet,ar=5800kcal / kg, S=0.9%, ex-factory price 680 yuan / ton), etc.

[0123] The logistics management interface obtains the transportation cycle and transportation cost of each type of coal.

[0124] The procurement history interface obtains procurement records for the past three months and analyzes procurement trends.

[0125] 2. Large Model Data Analysis The large model verified the acquired data and found that the sulfur content of Inner Mongolia coal (0.9%) exceeded the new environmental protection standard (0.8%). It marked it as an abnormal data item and evaluated its impact on procurement decisions.

[0126] An analysis of the cost-effectiveness of various types of coal revealed that Shaanxi coal has a lower price but lower calorific value, Inner Mongolia coal has a higher calorific value but exceeds the sulfur content standard, and northern Shanxi bituminous coal meets environmental protection requirements but has a higher price.

[0127] By predicting the coal price trend for the next month, it was found that due to production restrictions in the production areas, the price of bituminous coal in northern Shanxi may rise by 5%.

[0128] 3. Procurement strategy recommendation generation: Based on the data analysis results, the big model generates three procurement strategy recommendations: Cost optimization strategy: Prioritize the lower-priced Shaanxi coal, but the desulfurization cost will increase when mixing high-sulfur coal types.

[0129] The optimal strategy for environmental indicators: Choose northern Shanxi bituminous coal with a sulfur content that meets the requirements, but there may be a risk of price increases.

[0130] Comprehensive balance strategy: Mixed procurement of northern Shanxi bituminous coal and Shaanxi coal to meet environmental protection requirements while controlling procurement costs.

[0131] The big model recommends a comprehensive balanced strategy on the grounds that it can achieve a better balance between environmental compliance and cost control, while taking into account the risk of future price fluctuations.

[0132] 4. Construction of multi-objective optimization model: Basic calculation shows that the amount of standard coal required next month is 108,000 tons (360 million kWh × 300 g / kWh ÷ 1000).

[0133] The decision variables are defined as the purchase volume of bituminous coal in northern Shanxi (x1), the purchase volume of coal in Shaanxi (x2), and the purchase volume of coal in Inner Mongolia (x3).

[0134] Build a multi-objective optimization model: Minimize procurement costs: 650x1+580x2+680x3 Minimum sulfur content: (0.6x1+0.7x2+0.9x3) / (x1+x2+x3)≤0.8% Maximized average calorific value: (5500x1+5000x2+5800x3) / (x1+x2+x3)≥5300 kcal / kg Set up constraints: x1+x2+x3≥108,000 tons (to meet production needs) x1+x2+x3≤150,000 tons (not exceeding inventory capacity) x3=0 (Inner Mongolia coal sulfur content exceeds the standard and purchase is prohibited) x1≥20,000 tons (ensure a certain proportion of designed coal types to ensure combustion stability) 5. Procurement plan generation and output: The optimal solution obtained by the multi-objective optimization model is x1=40,000 tons, x2=68,000 tons, with a procurement cost of 65.44 million yuan, an average sulfur content of 0.72%, and an average calorific value of 5,200 kcal / kg.

[0135] Generate an Excel spreadsheet to list in detail the purchase quantity, supplier, price, shipping method and other information.

[0136] Generate a Word report including current inventory status, next month's demand analysis, recommended procurement plans and their advantages and risks.

[0137] 6. System interaction and feedback: The user viewed the generated procurement plan through the interactive interface and asked, "Can we increase the proportion of Shaanxi coal to reduce costs?" The system re-ran the multi-objective optimization model, adjusted the constraints, and generated a new plan: x1 = 30,000 tons, x2 = 78,000 tons, a procurement cost of 64.44 million yuan, an average sulfur content of 0.75%, and an average calorific value of 5,100 kcal / kg.

[0138] After the user compares the two options and chooses the second one, the system records the user's preference and optimizes the model parameters.

[0139] Through this invention, the power plant successfully developed a procurement plan that balanced environmental protection requirements with cost control, reducing procurement costs by approximately 1.5% while ensuring that the sulfur content of the coal met new environmental standards. The system's intelligent analysis and multi-objective optimization capabilities help users quickly assess the impact of different strategies, improving decision-making efficiency and quality.

[0140] The coal procurement decision-making system based on a large language model, as described in this embodiment of the present invention, can simultaneously balance multiple objectives, including economic efficiency, safety, energy efficiency, environmental protection, transportation cycle, and inventory ratio, to generate a procurement plan with optimal comprehensive indicators, thus avoiding the limitations of traditional single-objective optimization methods. By combining the large language model with an operations research optimization model, an end-to-end solution is achieved, from natural language requirements to mathematical model solutions. Leveraging the large language model's natural language understanding and knowledge reasoning capabilities, it provides intelligent decision support, understanding complex user requirements and generating procurement strategy recommendations that meet those requirements. The natural language interactive interface significantly improves the system's usability and interaction efficiency. It can effectively integrate multi-source heterogeneous data, including complex data such as equipment parameters, inventory information, supplier information, coal prices, and transportation routes. Through the analytical capabilities of the large model, it extracts key information, providing comprehensive data support for decision-making. It can also perceive dynamic factors such as market changes, transportation capacity fluctuations, and weather conditions in real time, allowing for rapid adjustments to procurement strategies to ensure the procurement plan remains optimal. By integrating with external data systems, real-time data updates and dynamic model adjustments are achieved.

[0141] In summary, the coal procurement decision-making system based on the large language model according to the embodiment of the present invention realizes the intelligent, scientific and efficient coal procurement decision-making by integrating the large language model with the operations optimization model, providing strong support for thermal power companies to reduce costs, improve efficiency and ensure safety.

[0142] Reference below Figure 5 , Figure 5A schematic diagram of the structure of a computer device suitable for implementing an embodiment of the present invention is shown.

[0143] like Figure 5 As shown, computer system 1000 includes a central processing unit (CPU) 1001, which can perform various appropriate actions and processes according to programs stored in read-only memory (ROM) 1002 or programs loaded from storage 1008 into random access memory (RAM) 1003. RAM 1003 also stores various programs and data required for the system's operating instructions. CPU 1001, ROM 1002, and RAM 1003 are connected to each other via bus 1004. An input / output (I / O) interface 1005 is also connected to bus 1004.

[0144] The following components are connected to the I / O interface 1005: an input section 1006 including a keyboard, mouse, and the like; an output section 1007 including devices such as a cathode ray tube (CRT), a liquid crystal display (LCD), and speakers; a storage section 1008 including devices such as a hard disk; and a communication section 1009 including a network interface card such as a LAN card or a modem. The communication section 1009 performs communication processing via a network such as the Internet. A drive 1010 is also connected to the I / O interface 1005 as needed. Removable media 1011, such as a magnetic disk, an optical disk, a magneto-optical disk, or a semiconductor memory, is installed in the drive 1010 as needed, so that computer programs read therefrom can be installed in the storage section 1008 as needed.

[0145] In particular, according to an embodiment of the present invention, the above reference flow chart Figure 1 The described process may be implemented as a computer-readable storage medium. For example, an embodiment of the present invention includes a computer-readable storage medium that includes a computer program that includes a computer program for executing the process. Figure 1 The program code of the method shown.

[0146] In particular, according to an embodiment of the present invention, the above reference flow chart Figure 1 The described process may be implemented as a computer software program. For example, an embodiment of the present invention includes a computer program product comprising a computer program carried on a computer readable medium, the computer program including a computer program for executing the process. Figure 1 The program code of the method shown.

[0147] In such an embodiment, the computer program includes program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network via communication section 1009 and / or installed from removable media 1011. When the computer program is executed by central processing unit (CPU) 1001, the above-described functions defined in the system of the present invention are performed.

[0148] The above embodiments merely illustrate several implementations of the present invention, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the patent. It should be noted that a person skilled in the art would be able to make various modifications and improvements without departing from the spirit of the present invention, all of which fall within the scope of protection of the present invention. Therefore, the scope of protection of the patent for this invention shall be determined by the appended claims.

Claims

1. A coal purchasing decision-making method based on a large language model, characterized in that: The following steps are involved: Obtain target data, including boiler equipment design parameters, minimum coal inventory indicators, basic information for fuel procurement plan formulation, coal supplier and price information, historical coal procurement information, and coal transportation routes and cycles; Input the target data into a pre-trained large language model, and analyze it through the large language model according to preset prompt words, and output the analysis results of the basic data of coal procurement; Based on the coal procurement basic data analysis results, policies, regulations and industry standards, further analysis is performed through the large language model to generate fuel procurement strategy direction recommendations; Based on the strategy direction selected by the user from the fuel procurement strategy direction recommendations, combined with the planned total procurement volume and the estimated procurement ratios of high and low ash melting points output by the basic calculation, a multi-objective optimization model is constructed, and the multi-objective optimization model is solved to obtain the optimal combination of procurement coal type, procurement quantity, and procurement route; Based on the optimal combination of the purchased coal type, purchase quantity and purchase route, a customized coal purchasing plan is generated through the natural language generation capability of the large language model.

2. The coal purchasing decision-making method based on a large language model according to claim 1 is characterized in that: The analysis is performed using the large language model to output the analysis results of the basic coal procurement data, including: Verifying and cleaning the target data and generating a data verification report; Extract key information from the verified and cleaned target data and convert it into structured data; Analyzing the correlation between different data items in the structured data to obtain a preliminary purchasing decision; Based on the abnormal data items in the data verification report, the impact on the procurement decision is evaluated, and the basic data analysis results of coal procurement are obtained based on the evaluation results.

3. The coal purchasing decision-making method based on a large language model according to claim 1 is characterized in that: The fuel procurement strategy direction recommendations are generated based on the coal procurement basic data analysis results, policies, regulations and industry standards, and further analyzed by the large language model, including: Obtain information on environmental protection policies, energy policies, and coal industry standards to analyze policy constraints in the procurement environment; Obtain industry standards based on power plant type and unit capacity parameters to develop a procurement strategy that complies with industry standards; Analyze historical procurement data and current market information to predict market trends in coal prices and supply tightness; Generate multiple fuel procurement strategy direction recommendations based on the policy constraints of the procurement environment, procurement strategies that comply with industry standards, and market trends in coal prices and supply tightness.

4. The coal purchasing decision-making method based on a large language model according to claim 1 is characterized in that: The multi-objective optimization model is constructed based on the strategy direction selected by the user from the fuel procurement strategy direction suggestions, combined with the planned total procurement volume and the estimated procurement ratio of high and low ash melting points output by the basic calculation, and the multi-objective optimization model is solved to obtain the optimal combination of procurement coal type, procurement quantity and procurement route, including: Calculate the total planned purchase volume and the estimated purchase ratios of high and low ash melting point coals based on the power generation plan and power supply coal consumption index information; Defining decision variables of the large language model based on multiple fuel procurement strategy direction suggestions; Construct an objective function based on the fuel procurement strategy direction recommended by the user; Setting constraints for the large language model based on boiler equipment parameters, inventory requirements, and environmental standards; The large language model is solved to generate an optimal combination of purchased coal type, purchase quantity and purchase route.

5. The coal purchasing decision-making method based on a large language model according to claim 1 is characterized in that: The method generates a customized coal procurement plan based on the optimal combination of the purchased coal type, purchase quantity, and purchase route through the natural language generation capability of the large language model, including: Generate a detailed data table based on the solution of the multi-objective optimization model, wherein the detailed data table includes the purchase quantity, supplier, price, transportation method, and arrival time of each type of coal; Generate a structured procurement report based on preset prompt words, including current coal procurement and inventory status, next month's demand status, and next month's fuel procurement plan; Present key data from the coal procurement plan in chart form; Output the generated detailed data tables and structured procurement reports to the designated path and upload them to the company's document management system.

6. The coal purchasing decision-making method based on a large language model according to claim 1, characterized in that: Also includes: Through the natural language interactive interface, real-time interaction with users is achieved to receive users' evaluation, adjustment and confirmation of the generated coal procurement plan, and the parameters and strategy recommendations of the large language model are optimized based on user feedback.

7. The coal purchasing decision-making method based on a large language model according to claim 6 is characterized in that: The natural language interactive interface is used to achieve real-time interaction with users to receive user evaluation, adjustment, and confirmation of the generated coal procurement plan, and optimize the parameters and strategy recommendations of the large language model based on user feedback, including: Providing a text-based or voice-based interactive interface to communicate with the user in natural language; Analyze the natural language text input by users to identify their real needs and intentions; Adjust and optimize the coal procurement plan based on user feedback and needs.

8. The coal purchasing decision-making method based on a large language model according to claim 7 is characterized in that: Also includes: Query expertise related to coal procurement through an interactive interface and receive accurate answers based on an external knowledge base; Record the user's interaction history and decision preferences, and continuously optimize the parameters and strategy recommendations of the large language model through continuous learning.

9. A coal purchasing decision system based on a large language model, characterized in that: include: An acquisition module is used to acquire target data, including boiler equipment design parameters, minimum coal inventory indicators, basic information for fuel procurement plan formulation, coal supplier and price information, historical coal procurement information, and coal transportation routes and cycles; An analysis module is used to input the target data into a pre-trained large language model, and analyze it through the large language model according to preset prompt words, and output the analysis results of the basic coal procurement data; A generation module, configured to generate fuel procurement strategy direction recommendations based on the coal procurement basic data analysis results, policies, regulations and industry standards, and further analyze them through the large language model; A solution module is used to construct a multi-objective optimization model based on the strategy direction selected by the user from the fuel procurement strategy direction recommendations, combined with the planned total procurement volume and the estimated procurement ratio of high and low ash melting points output by the basic calculation, and solve the multi-objective optimization model to obtain the optimal combination of procurement coal type, procurement quantity, and procurement route; The decision-making module is used to generate a customized coal procurement plan based on the optimal combination of the purchased coal type, purchase quantity and purchase route through the natural language generation capability of the large language model.

10. A computer device comprising a processor and a computer program, characterized in that When the processor executes the computer program, the coal purchasing decision-making method based on the large language model is implemented according to any one of claims 1 to 8.

Images