Multi-stage emergency disposal scheduling method for heat supply system containing large model

Abstract

The invention discloses a heat supply system multi-stage emergency disposal scheduling method containing a large model. The method comprises the steps of forming a heat supply system emergency disposal scheduling large model; in the emergency prevention preparation stage, a large model is used for building a heat supply system multi-type emergency event question and answer database, and emergency knowledge intelligent question and answer, emergency event reason analysis, risk hidden danger query and emergency disposal scheduling measure recommendation are conducted; in the emergency response stage, the large model is guided to output an optimal feature combination representing various types of emergency events, and machine learning is adopted to identify the various types of emergency events; heat supply loss caused by an emergency event is evaluated by using the large model, a multi-level emergency disposal scheduling strategy is set, and an optimal emergency disposal scheduling scheme is obtained; and in an emergency disposal post-stage, analyzing the emergency disposal scheduling process by using the large model, establishing an emergency disposal scheduling plan library, and performing plan simulation, identifying potential vulnerabilities and strengthening the emergency disposal scheduling process by using the intelligent agent.

Description

Technical Field

[0001] The present invention belongs to the technical field of heating systems, and particularly relates to a multi-stage emergency response and dispatching method for a heating system containing a large model. Background Art

[0002] With the advancement of urbanization, the urban heating industry has been continuously developing, and the operation and management difficulty has increased accordingly. The frequent occurrence of emergency events in the heating system has also brought serious losses to heating enterprises. Therefore, it is particularly important to respond, dispose, and dispatch in a timely manner after an emergency event occurs. The emergency response and dispatching of the heating system is to, when an emergency event such as a system failure or the impact of extreme weather occurs, formulate an optimal emergency response and dispatching plan in a timely and accurate manner according to the type of emergency event, control and prevent the expansion of the situation, and minimize the losses and impacts.

[0003] In recent years, with the emergence of artificial intelligence technologies such as natural language large models and machine learning, industries such as power and industry have gradually gained the capabilities of knowledge learning, comprehensive perception, and autonomous decision-making. Studying how to efficiently empower the emergency response and dispatching scenarios of the heating system based on technologies such as natural language large models and machine learning, making the emergency prevention and preparation stage, emergency response stage, and emergency response and aftermath stage of the heating system more intelligent and efficient, and improving the emergency response and dispatching management level of the heating system is an urgent problem to be solved currently.

[0004] Based on the above technical problems, a new multi-stage emergency response and dispatching method for a heating system containing a large model needs to be designed. Summary of the Invention

[0005] The technical problem to be solved by the present invention is to overcome the deficiencies of the prior art and provide a multi-stage emergency response and dispatching method for a heating system containing a large model, which integrates large models, machine learning, and agent technologies, gives full play to the advantages of different technologies, understands and generates emergency knowledge through the large language model of the heating system formed by training, identifies emergency events through large models and machine learning, evaluates heating losses through large models, and outputs the optimal emergency response and dispatching plan through multi-agent emergency response and dispatching decisions, etc., making each stage of the heating system emergency more intelligent and scientific, and comprehensively improving the emergency response and dispatching management level of the heating system.

[0006] To solve the above technical problems, the technical solution of the present invention is: The present invention provides a multi-stage emergency response and dispatching method for a heating system containing a large model, including: an emergency prevention and preparation stage, an emergency response stage, and an emergency response and aftermath stage; The emergency prevention and preparation stage includes: Construct a multi-type emergency event Q&A database for the heating system based on a large model. The multi-type emergency event Q&A database for the heating system extracts key event information from the emergency response dispatching data of the heating system, generates targeted questions and answers. The targeted questions and answers include intelligent Q&A for emergency knowledge. The multi-type emergency event Q&A database for the heating system provides recommendations for emergency event cause analysis, risk and hidden danger query, and emergency response dispatching measures; The emergency response stage includes: Use prompt engineering to guide the large model to output the optimal feature combination representing each type of emergency event for the identification of each type of emergency event; Evaluate the heating loss caused by the identified emergency event. Based on the goal of quickly reconstructing the hydraulic balance condition of the heat network and reducing heating loss, set the emergency response dispatching decision-making task; Multiple emergency response dispatching intelligent agents respectively output emergency response dispatching plans based on the emergency response dispatching decision-making task; Each intelligent agent judges the emergency response dispatching plans output by other intelligent agents; Based on the emergency response collaborative intelligent agent and the judgment and analysis of each intelligent agent, obtain the optimal emergency response dispatching plan; The post-emergency response stage includes: Analyze and summarize the emergency response dispatching process using a large model, and use intelligent agents to simulate the emergency response dispatching plan to strengthen the emergency response dispatching process.

[0007] Furthermore, in the emergency response stage, set up a workflow for outputting the feature combination of each type of emergency event. The feature combination workflow includes uploading the emergency response dispatching data set, emergency event description, task requirement description for feature extraction using machine learning, selecting the feature combination mechanism, and outputting the optimal feature combination; Use prompt engineering to guide the large model to sequentially implement the workflow for outputting the feature combination, and output the optimal feature combination representing each type of emergency event; Input the optimal feature combination of each type of emergency event into a preset machine learning model for training and learning to establish a heating emergency event identification model to identify the types of emergency events occurring in the heating system.

[0008] Furthermore, constructing a heating emergency event identification model includes: setting up a two-layer machine learning model to extract features from the emergency event data set. The input end of the first-layer machine learning model is the original emergency response dispatching data set, and the original emergency response dispatching data set is feature-abstracted through unsupervised learning; The second-layer machine learning model uses local feature extraction capabilities to perform deep extraction of each data feature; Evaluate different feature combinations using the cross - validation method, optimize and iterate the feature combinations according to the evaluation results, and output the optimal feature combination for identifying emergency events in the heating system.

[0009] Furthermore, the steps for evaluating heating losses include: constructing an economic loss calculation model and calculating the economic losses caused by emergency events based on the data identified in the emergency events.

[0010] Furthermore, the data in the emergency events include the number of affected heat users, heating area, missing values of heat load, and repair time.

[0011] Furthermore, the method for constructing the economic loss calculation model includes using the economic loss data of historical emergency events and currently determined loss factors. Taking the number of affected heat users, heating interruption duration, unit price of heating fees, etc. as independent variables and the economic loss value as the dependent variable, training the model to learn the relationship between loss factors and economic losses, and constructing the economic loss calculation model.

[0012] Furthermore, quickly reconstructing the hydraulic balance condition of the heat network based on the heat network hydraulic model includes constructing the heat network hydraulic model based on the heat network topology structure data, fluid mechanics principles, and historical operation data, and calculating the impact of emergency events on the hydraulic balance of the heat network.

[0013] Furthermore, set a multi - level emergency response dispatching strategy: In the first layer, multiple emergency response dispatching agents are set up to analyze and reason about the information of emergency events occurring in the heating system respectively, and output their respective emergency response dispatching plans; In the second layer, information interaction is carried out among the emergency response dispatching agents to judge, discuss, and give opinions on the emergency response dispatching plans output by other emergency response dispatching agents respectively; In the third layer, the set emergency response coordination agent integrates the current emergency event problems and the inferences of each emergency response dispatching agent, makes a judgment and analysis, and obtains the optimal emergency response dispatching plan.

[0014] Furthermore, the emergency response dispatching decision - making task is expressed as: ; Where: D is the generated emergency response dispatching plan; is the large - model for emergency response dispatching of the heating system; is the goal of quickly reconstructing the hydraulic balance condition of the heat network and reducing heating losses; is the current operating information of the heating system; is the knowledge of emergency response dispatching of the heating system.

[0015] Furthermore, it also includes extracting key event information from the emergency response dispatching data of the heating system, generating targeted questions and answers, and automatically constructing a multi-type emergency event Q&A database for the heating system using a large model. The beneficial effects of the present invention are as follows: (1) By forming a large model for emergency response dispatching of the heating system, the present invention has the capabilities of semantic understanding, knowledge reasoning, and content generation of emergency response dispatching knowledge for the heating system, enabling it to deeply understand the business knowledge and emergency scenarios of the heating system, establishing a learning and reasoning foundation for emergency events and emergency response dispatching knowledge in the subsequent emergency prevention and preparation stage, emergency response stage, and emergency response aftermath stage, improving the efficiency of emergency knowledge learning and application for heating business personnel, making each stage of the heating system emergency more intelligent and scientific, and comprehensively improving the emergency response dispatching management level of the heating system; in the emergency prevention and preparation stage of the present invention, by using a large model to construct a multi-type emergency event Q&A database for the heating system, heating business personnel can quickly obtain relevant emergency knowledge in advance, conduct risk and hidden danger queries, and analyze the causes of emergency events, etc., deepening the understanding of emergency events by heating business personnel, thoroughly investigating potential risk points, and preventing the occurrence of emergency events in advance; (2) In the emergency response stage of the present invention, using prompt engineering to guide the large model to output the optimal feature combination and combining with a machine learning model for emergency event recognition can significantly improve the accuracy and efficiency of emergency event recognition, quickly and accurately identify various emergency events, and win precious time for subsequent emergency response dispatching; the assessment of heating losses by the large model is more comprehensive and accurate, and it can formulate more targeted strategies to reduce losses, minimizing the impact of emergency events on the heating system and users; moreover, through the emergency response dispatching of multi-level intelligent agents, it can better adapt to the complex emergency event response environment of the heating system, understand and analyze the current emergency event environment and related emergency problems, and propose an interpretable emergency response dispatching plan for the heating system, facilitating heating business personnel to quickly respond to and handle emergency events, and better achieving the goals of quickly reconstructing the hydraulic balance condition of the heat network and reducing heating losses; (3) In the emergency response aftermath stage of the present invention, using the large model to analyze and summarize the emergency response dispatching process of emergency events in the heating system helps to accumulate valuable experience, integrate this experience into the safety emergency response dispatching plan library of the heating system, continuously improve the emergency plan, make it more in line with the actual situation, and improve the practicality and operability of the plan; moreover, using intelligent agents to simulate the emergency response dispatching plan can discover potential loopholes in advance, and through the analysis of the simulation results, strengthen the emergency response dispatching process targeted, improve the response speed and handling ability of the heating system in dealing with similar events in the future, realize the continuous optimization of the emergency response process, and ensure the safe and stable operation of the heating system. Description of the Drawings

[0016] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following will briefly introduce the accompanying drawings required for the description of the specific embodiments or the prior art. Obviously, the accompanying drawings in the following description are some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other accompanying drawings can also be obtained based on these drawings.

[0017] Figure 1 is a flowchart of an embodiment of the present invention; Figure 2 is a schematic diagram of the implementation principle of a multi-level emergency response dispatching strategy in an embodiment of the present invention. Specific embodiments

[0018] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the following will clearly and completely describe the technical solutions of the present invention with reference to the accompanying drawings. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts fall within the scope of protection of the present invention.

[0019] As Figure 1 shown, an embodiment of the present invention provides a multi-stage emergency response dispatching method for a heating system containing a large model, which includes: Based on the heating system text data, a heating system large language model is pre-trained through a base large model, and the heating system large language model is fine-tuned through the heating system emergency response dispatching data to form a heating system emergency response dispatching large model; In the emergency prevention and preparation stage, by extracting the key event information of the heating system emergency response dispatching data, generating targeted questions and answers, using the large model to automatically construct a multi-type emergency event Q&A database for the heating system, and supporting heating business personnel to conduct intelligent Q&A of emergency knowledge, analysis of the causes of emergency events, query of risk hazards, and recommendation of emergency response dispatching measures through the Q&A method; In the emergency response stage, prompt engineering is used to guide the large model to output the optimal feature combination representing various types of emergency events, and a machine learning model is used to identify various types of emergency events; After using the large model to evaluate the heating losses caused by the identified emergency events, with the goal of quickly reconstructing the hydraulic balance condition of the heat network and reducing heating losses, an emergency response dispatching decision-making task is set, and after analysis and reasoning by different intelligent agents, their respective emergency response dispatching plans are output. Then, each intelligent agent judges the emergency response dispatching plans of other intelligent agents, and finally, the optimal emergency response dispatching plan is obtained through the judgment and analysis of the emergency response coordination intelligent agent; In the aftermath stage of emergency response, use large models to analyze and summarize the emergency response dispatching process of heating system emergency events, establish a safety emergency response dispatching plan library for the heating system, and use intelligent agents to simulate emergency response dispatching plans to identify potential vulnerabilities and strengthen the emergency response dispatching process.

[0020] It should be noted that in the aftermath stage of emergency response, use the large model for emergency response dispatching of the heating system to analyze and summarize the emergency response dispatching process of heating system emergency events, and generate a detailed analysis, including the detailed process of the event, the effectiveness evaluation of various dispatching measures, whether the resource utilization is reasonable, and whether the personnel dispatching is efficient, etc.; according to the analysis report of the large model, combined with past successful emergency response cases and industry standards and specifications, use the large model to generate a general emergency response dispatching plan template. For example, for a pipeline rupture event, the plan template generated by the large model may include: Emergency response initiation: After receiving the pipeline rupture alarm, confirm the location and general situation of the event; Emergency measures: Immediately close the valves upstream and downstream of the ruptured pipeline to prevent further leakage of hot water; and perform heat network isolation and network switching operations to make up for heat losses; Repair preparation: Dispatch repair personnel and required materials (such as pipelines of corresponding diameters, welding equipment, etc.) to the site; On-site repair: Repair the pipeline according to the standard process; Restoration of heating: After the repair is completed, conduct pressure tests and trial runs. After ensuring no leakage, gradually restore heating.

[0021] Use intelligent agents to simulate emergency response dispatching plans to identify potential vulnerabilities and strengthen the emergency response dispatching process, including: Set up an agent to be responsible for simulating the operating environment of the heating system, generating an emergency event scenario and implementing an emergency response dispatching plan. For example, if a heating pipeline breaks in a certain area, implement the disposal operation according to the emergency response dispatching plan, and at the same time track various data during the simulation process, such as whether the valve relationship is successful and whether the heating system parameters gradually return to normal, etc.; during the simulation process, identify potential loopholes by comparing the actual simulation results with the expected results of the plan. For example, in the plan, when the valve is closed, the pressure in the pipeline should drop to the safe range within the preset time, but the simulation results show that the pressure drops slowly and exceeds the expected time, indicating that there may be loopholes in the valve operation or pressure control in the plan. At the same time, conduct in-depth analysis of the data during the simulation process to dig out potential systematic problems. For example, through the analysis of the resource allocation situation in multiple simulations, it is found that the emergency repair resource reserves in some areas are insufficient, resulting in an extended emergency response time, which is also a loophole that needs to be improved. According to the potential loopholes identified by the agent simulation, use the large model to generate improvement suggestions, and feedback the improvement suggestions to the emergency response dispatching plan library to update and improve the corresponding plan. At the same time, regularly re-run the agent simulation to verify whether the improved plan is effective, continuously strengthen the emergency response dispatching process, and improve the ability of the heating system to respond to emergencies.

[0022] In this embodiment, based on the heating system text data, a heating system large language model is pre-trained through a base large model, and the heating system large language model is fine-tuned through the heating system emergency response dispatching data to form a heating system emergency response dispatching large model, including: Obtain heating system text data, including: design documents of the heating system, operation manuals, historical operation data, and industry standard specifications; Input the heating system text data into the pre-selected base large model for pre-training, learn the language patterns, terms, and knowledge related to the heating system, and continuously adjust the large model parameters to form a heating system large language model with the ability to understand heating system expertise, knowledge reasoning, and content generation; Obtain the emergency response dispatching data of the heating system when equipment failures, extreme weather impacts, and heat network failures occur, and fine-tune the heating system large language model. During the fine-tuning process, on the basis of maintaining the original heating system knowledge, learn specific knowledge related to emergency response dispatching to form a heating system emergency response dispatching large model.

[0023] In this embodiment, in the emergency prevention and preparation stage, by extracting the key event information of the heating system emergency response dispatching data, generating targeted questions and answers, and automatically constructing a heating system multi-type emergency event Q&A database by using the large model, including: In the emergency prevention and preparation stage, utilize the parsing ability of the large model for heat supply system emergency response dispatching to automatically extract key event information from the heat supply system emergency response dispatching data, including emergency event named entity recognition and event description extraction; According to the key event information of different emergency events, guided by the large model for heat supply system emergency response dispatching, generate questions and answers related to the emergency events; Organize and structurally store the generated questions and answers to construct a question and answer database for various types of emergency events in the heat supply system.

[0024] Utilize the text parsing ability of the large model to identify named entities in emergency events, such as event types, names of involved equipment, geographical locations, time, etc.; also use the large model to extract key event description information from the emergency event text, such as the process of the event occurrence, the affected range, and the measures already taken. Based on the extracted key event information, use the large model again to generate relevant questions and answers. For example, for a pipeline rupture event, questions such as "Where does a pipeline rupture usually occur?" and "How to quickly locate the rupture point after a pipeline rupture?" can be generated; for the question "How to quickly locate the rupture point after a pipeline rupture?", the large model generates the answer "It is possible to use the pressure monitoring system to compare the normal pressure distribution with the current pressure data. The area with a significant sudden drop in pressure may be the location of the rupture point. It is also possible to use thermal imaging technology to detect abnormal temperature areas on the pipeline surface to locate the rupture point."

[0025] It should be noted that in the heat supply system, common emergency events include: Pipeline rupture: Usually accompanied by a sudden drop in pressure and an abnormal increase in flow rate, and the temperature in the area near the rupture point drops. For example, at 2 am on a winter day, a main heat supply pipeline in the city center area ruptured, affecting the heat supply of multiple surrounding communities; Equipment failure: Such as a failure of the circulation pump, which will cause a significant reduction in the circulation flow rate of the heat supply system, resulting in a drop in the heat supply temperature; a failure of the heat exchanger may cause a reduction in the heat exchange efficiency, resulting in a situation where the supply water temperature is high but the return water temperature is also high. For example, a circulation pump in a heat exchange station suddenly failed during operation, causing the heat supply temperature in the area served by the heat exchange station to drop by 5°C within a short period of time; Heat source interruption: May be caused by reasons such as a failure of the thermal power plant or fuel supply problems. At this time, the supply water temperature of the entire heat supply system will drop rapidly, and the pressure will also decrease accordingly. For example, due to a mistake in the equipment maintenance of the thermal power plant, the heat supply source was interrupted for 2 hours, and the temperature of the city's heat supply system generally dropped. In this embodiment, the emergency event entity recognition includes: the time, location, equipment name, emergency event type, severity, and corresponding emergency response dispatching knowledge of the emergency event; Emergency event types include: heat source failures, heat network failures, pump and valve failures, and the inability of the heat source to be processed normally according to the dispatching plan caused by extreme weather impacts; Emergency response dispatching knowledge includes: analyzing the causes of events, heat losses, economic losses, and emergency response dispatching plans for different types of emergency events; the emergency response dispatching plan includes: adjusting the heat source to increase the load, starting standby equipment, switching the heat supply network line, valve disconnection, and regulating the pump and valve of the heat substation; Event description extraction includes: by analyzing the sentence structure and semantic relationships, locating key descriptive statements, and extracting descriptions of the occurrence process, impacts, and handling measures of emergency events.

[0026] In this embodiment, in the emergency response phase, prompt engineering is used to guide the large model to output the optimal feature combinations representing various types of emergency events, and a machine learning model is used to identify various types of emergency events, including: In the emergency response phase, set up a workflow for outputting the feature combinations of various types of emergency events, including uploading the emergency response dispatching data set, emergency event description, task requirement description for feature extraction using machine learning, selecting a feature combination mechanism, and outputting the optimal feature combination; Use prompt engineering to guide the large model to sequentially implement the workflow of outputting feature combinations, and output the optimal feature combinations representing various types of emergency events; Input the optimal feature combinations of various types of emergency events into a preset machine learning model for training and learning, establish a heat supply emergency event recognition model, and identify the types of emergency events occurring in the heat supply system.

[0027] The task requirement description for feature extraction using machine learning refers to being able to extract key features for accurately identifying different types of emergency events from the huge and complex operation data of the heat supply system. The extracted features should have the following characteristics: High discrimination: It can significantly distinguish different types of emergency events. For example, the sudden pressure drop feature during pipeline rupture and the abnormal flow feature during equipment failure should have obvious differences, so that the model can accurately identify the event type; Stability: The features are not affected by the normal daily fluctuations of the heat supply system, ensuring that they can reliably indicate the occurrence of emergency events under different operating conditions. For example, the normal flow changes during the morning and evening peak periods of the heat supply system should not interfere with the judgment of abnormal flow caused by equipment failure; Computational efficiency: Considering the real-time requirements of emergency response, the calculation process of the features should be simple and efficient, and can be completed in a short time to provide support for rapid decision-making.

[0028] In this embodiment, the task requirements for feature extraction using machine learning include: setting up a two-layer machine learning model to perform feature extraction on the emergency response dispatch dataset. The input end of the first-layer machine learning model is the original emergency response dispatch dataset, which is used to perform feature abstraction through unsupervised learning. The second-layer machine learning model uses local feature extraction capabilities to perform deep extraction of each data feature; The feature combination selection mechanisms include: the exhaustive search mechanism, the greedy search mechanism, and the heuristic algorithm mechanism; Outputting the optimal feature combination includes: using the cross-validation method to evaluate different feature combinations, and optimizing and iterating the feature combinations according to the evaluation results to output the optimal feature combination applicable to the identification of emergency events in the heating system.

[0029] It should be noted that for the first-layer machine learning model: an autoencoder can be selected as the unsupervised learning model for feature abstraction. The autoencoder consists of an encoder and a decoder. The encoder compresses the original high-dimensional emergency response dispatch dataset into a low-dimensional feature representation, and the decoder then reconstructs these features back into the original data space. By training the autoencoder to be able to reconstruct the input data as accurately as possible, the potential patterns and features in the data can be learned. For the second-layer machine learning model: a convolutional neural network (CNN) is used for deep feature extraction. Since the CNN has strong local feature extraction capabilities, it is suitable for processing the feature data abstracted by the first-layer autoencoder.

[0030] The genetic algorithm in the heuristic algorithm mechanism can be selected to perform feature combination. The genetic algorithm simulates the biological evolution process and gradually optimizes the feature combination through operations such as selection, crossover, and mutation. The specific steps are as follows: Coding: Encode each feature combination as a chromosome. For example, binary coding is used, where 1 indicates that the feature is selected and 0 indicates that it is not selected; Initializing the population: Randomly generate a certain number of chromosomes (feature combinations) to form the initial population; Fitness evaluation: Use a fitness function based on a machine learning model (such as a decision tree classifier) to evaluate the quality of each chromosome. On the training data, the classification accuracy of the model under this feature combination is used as the fitness value; Selection: The roulette wheel selection method is adopted to select chromosomes according to the fitness. The higher the fitness, the greater the probability of being selected, so as to retain excellent feature combinations; Crossover: With a certain crossover probability, perform single-point crossover operations on the selected chromosomes to generate new chromosomes and introduce new feature combination possibilities; Mutation: With a low mutation probability, flip some gene positions in the chromosome to increase the diversity of the population; Iterative optimization: Repeat the above fitness evaluation, selection, crossover, and mutation operations until the termination condition is met (such as reaching the maximum number of iterations or the fitness no longer improves).

[0031] In this embodiment, a large model is used to evaluate the heat supply losses caused by the identified emergency events, including: Using the large model for heat supply system emergency disposal and scheduling, according to the identified emergency events, obtain the heat supply operation data before and after the emergency events occur, and calculate the number of affected heat users, heat supply area, heat load missing value, and economic loss value.

[0032] It should be noted that the evaluation of heat supply losses specifically includes: 1) The large model can process and analyze a large amount of heat network topology data, historical operation data, etc., help identify patterns, trends, and correlations in the data, and provide deeper insights for constructing a heat network hydraulic model. For example, through the analysis of historical operation data, the large model can discover the variation laws of the heat network hydraulic characteristics in different seasons and time periods; Secondly, using the heat network topology data (pipe connection relationship, pipe diameter, length, etc.), fluid mechanics principles, and historical operation data, construct a heat network hydraulic model (this model can simulate the hydraulic characteristics of the heat network under different working conditions, such as flow distribution, pressure change, etc.), and then use the optimization algorithm of the large model to optimize the parameters of the heat network hydraulic model to improve the accuracy and reliability of the model. The large model can simulate different working conditions, adjust parameters such as pipe diameter and length in the model, so that the model can better simulate the hydraulic characteristics of the actual heat network; 2) Simulate the impact of emergency events on the heat network hydraulic balance in the heat network hydraulic model. Determine the affected area range, flow change situation, and pressure fluctuation situation. For example, when a pipe bursts, the model can quickly calculate the sudden decrease in flow and pressure drop in the area near and downstream of the burst point; 3) Loss factor analysis: Use the large model to comprehensively analyze the emergency event type, the number of affected heat users, heat supply area, heat load missing value, and economic data of the heat supply system (such as heat fee unit price, equipment repair cost, user compensation standard, etc.). Through in-depth mining of these data, the large model can more comprehensively determine the constituent factors of heat supply losses. For example, for a pipe burst event, the economic losses include emergency repair costs, heat fee losses caused by heat supply interruption, and possible compensation costs for users; In addition, based on the analysis of loss factors, the large model can evaluate and warn of the economic loss risks that may be brought by different types of emergency events. For example, according to the current heat network operation status and the development trend of emergency events, the large model can predict in advance possible heat fee losses, equipment repair costs, etc., and provide decision-making support for emergency disposal; 4) Construction of loss calculation model: The large model utilizes its powerful learning ability. Based on the economic loss data of historical emergency events and the currently determined loss factors, taking the number of affected heat users, the duration of heat supply interruption, the unit price of heat fees, etc. as independent variables and the economic loss value as the dependent variable, the model is trained to learn the relationship between these factors and the economic loss, and an economic loss calculation model is constructed. 5) Loss assessment: When an emergency event occurs, the large model can obtain relevant data in real time and use the constructed economic loss calculation model to input the relevant data of the current emergency event (such as the number of affected heat users, heating area, heat load missing value, repair time, etc.), and quickly calculate the economic loss value caused by this emergency event; at the same time, the large model can also timely feedback the calculation result to relevant personnel to provide real-time support for the emergency dispatch plan.

[0033] In this embodiment, with the goal of quickly reconstructing the hydraulic balance condition of the heat network and reducing heat supply losses, an emergency disposal dispatch decision-making task is set. After analysis and reasoning by different agents, their respective emergency disposal dispatch plans are output. Then, each agent judges the emergency disposal dispatch plans of other agents. Finally, the optimal emergency disposal dispatch plan is obtained through the judgment and analysis of the emergency disposal collaborative agent, including: Setting the emergency disposal dispatch decision-making task: The large model for emergency disposal dispatch of the heating system aims to quickly reconstruct the hydraulic balance condition of the heat network and reduce heat supply losses. Combining the current operation information of the heating system and based on the knowledge of emergency disposal dispatch of the heating system, it makes understanding judgments and generates the optimal emergency disposal dispatch plan. Setting multi-level emergency disposal dispatch strategies: In the first layer, multiple emergency disposal dispatch agents are set to analyze and reason about the emergency event information occurring in the heating system respectively and output their respective emergency disposal dispatch plans; in the second layer, information interaction is carried out among the emergency disposal dispatch agents to judge and discuss and put forward opinions on the emergency disposal dispatch plans output by other emergency disposal dispatch agents respectively; in the third layer, the set emergency disposal collaborative agent integrates the current emergency event problems and the inferences of each emergency disposal dispatch agent for judgment and analysis to obtain the optimal emergency disposal dispatch plan.

[0034] As shown in Figure 2, in an embodiment of the present invention, the specific steps for implementing the multi-level emergency disposal dispatch strategy include: 1) In the first layer, ChatGPT, ChatGLM, and Gemini large models can be selected and integrated into the emergency response dispatching agent. Driven by the large models, multiple emergency response dispatching agents (Agent I, Agent II, and Agent III) are generated. According to the emergency events occurring in the current heating system, the operation parameters of the heating system are sensed, valuable information is collected, analyzed, and inferred, and then their respective emergency response dispatching plans are output; 2) In the second layer, based on the respective emergency response dispatching plans of Agent I, Agent II, and Agent III, each agent conducts information interaction, examines the emergency response dispatching plans of the other two agents, makes judgment and discussion, and puts forward its own opinions; 3) In the third layer, the emergency response collaborative agent integrates the current emergency event problems and the inferences of Agent I, Agent II, and Agent III, makes judgment and analysis, and obtains the optimal emergency response dispatching plan that conforms to the current heating system emergency event, providing strong decision-making support for heating operation personnel.

[0035] Suppose a serious pipeline rupture emergency event occurs during the winter peak period in the heating system. The following is a specific case of generating an emergency response dispatching plan based on the multi-level emergency response dispatching strategy: 1) Initialization of the plan generation agent driven by the large model: Integrate ChatGPT, ChatGLM, and Gemini large models into the emergency response dispatching agent to generate Agent I, Agent II, and Agent III. After the system identifies the pipeline rupture event, the emergency response dispatching agent starts to work.

[0036] Data collection and analysis: Agent I, through connecting to the sensor network of the heating system, senses the sudden drop in temperature, rapid decrease in pressure, and abnormal increase in flow rate near the pipeline rupture point. At the same time, it collects that the number of heat users in this area is 500 households, the heating area reaches 300,000 square meters, the supply water temperature in the primary network of the pipe network is 80°C, the return water temperature is 50°C, and the diameter of the ruptured pipeline is 500mm, which belongs to the key main pipeline. Based on this information, using the large model for analysis and inference, an emergency response dispatching plan I is generated: First, immediately close the valves upstream and downstream of the ruptured pipeline to prevent further leakage of hot water, and it is expected that the valve closing operation can be completed within 1 hour; mobilize the nearby repair team to carry a 500mm diameter spare pipeline and related welding equipment to the scene, and it is expected to arrive within 2 hours; at the same time, start the nearby standby heat source and increase its output to compensate for part of the heating demand in the affected area. It is expected that the standby heat source can be increased to full load operation within 30 minutes and can meet about 30% of the heating demand in the affected area.

[0037] Agent II obtains data from the equipment monitoring system and the user information management system, and learns that the rupture of this pipeline may affect 3 surrounding heat exchange stations, involving 100 commercial users and 20 industrial users. After analyzing with the large model, it outputs the emergency response dispatch plan II: first reduce the operating load of the 3 surrounding heat exchange stations, reduce the overall heat network flow, and reduce the pipeline pressure to reduce the hot water leakage. It is expected to complete the load adjustment within 15 minutes; notify the surrounding commercial and industrial users of the possible interruption and approximate restoration time of the heating supply, and make relevant preparations; organize a professional inspection team to use pipeline inspection robots to inspect the area around the ruptured pipeline to evaluate whether there are other potential hidden dangers. The inspection work is expected to be completed within 4 hours.

[0038] Agent III combines the emergency event monitoring system and historical data, and finds that the rupture location of this pipeline is at a key node of the heat network, with a large influence range. With the help of the large model, it generates the emergency response dispatch plan III: urgently allocate some heating resources from other areas, and through adjusting the regulating valves in the heat network, deliver part of the heat to the affected area. It is expected to achieve the redistribution of heating resources within 2 hours; at the same time, arrange personnel to conduct a comprehensive inspection of the heating pipeline network in the affected area, focusing on checking the branch pipelines connected to the ruptured pipeline. The inspection work is expected to be completed within 6 hours; for the affected residential users, provide temporary heating equipment such as electric heaters, and it is expected to start distributing within 3 hours.

[0039] 2) Information interaction and judgment among agents Agent I's judgment on the plans of other agents: When Agent I examines Plan II, it believes that although reducing the operating load of the heat exchange station can reduce the leakage volume, it may cause the heating temperature in the surrounding non-directly affected areas to drop, expanding the influence range. It is recommended to closely monitor the heating temperature in the non-affected areas while reducing the load and make timely adjustments. For Plan III, Agent I believes that the idea of allocating heating resources from other areas is feasible, but the 2-hour time is too long. It is recommended to optimize the dispatching algorithm to shorten the allocation time.

[0040] Agent II's judgment on the plans of other agents: Agent II analyzes that the operations of closing the valve and mobilizing the repair team in Plan I are reasonable, but the start-up time of the standby heat source is relatively long. It is recommended to make preparations for preheating the standby heat source in advance to shorten the start-up time. For Plan III, Agent II points out that the work arrangement of inspecting the branch pipelines is relatively reasonable, but the plan of distributing temporary heating equipment may not be able to meet the needs of all residents due to limited quantity. It is recommended to coordinate the opening of public heating places in the surrounding communities at the same time.

[0041] Agent III's judgment on the plans of other agents: When Agent III evaluated Plan I, it found that the time for the emergency repair team to arrive at the scene was relatively long and suggested optimizing the dispatching route to improve traffic efficiency. For Plan II, Agent III believed that the way to notify users could be diversified. In addition to direct notification, information could also be released through channels such as social media and the heating company's APP to expand the scope of notification.

[0042] 3) The emergency response collaborative agent determines the optimal plan Comprehensive evaluation: The emergency response collaborative agent collects the plans of Agent I, Agent II, and Agent III and the mutual judgment information between them. From the perspective of the recovery effect of the heat network hydraulic balance, Plan I closing the valve and starting the standby heat source, and Plan III allocating heating resources have a positive effect on restoring the hydraulic balance; from the aspect of reducing the degree of heat supply loss, Plan II reducing the load of the heat exchange station to reduce the leakage volume, and Plan I and III taking measures to ensure partial heat supply can reduce the heat fee loss; in terms of implementation cost, Plan I mobilizing the emergency repair team and equipment, and Plan III allocating resources, etc. involve human and material costs; for the impact on users' heat supply experience, Plan III providing temporary heating equipment, and Plan II notifying users, etc. all have a certain role.

[0043] Weight setting: Considering the current serious pipeline rupture and large heat supply loss, the weight of the degree of heat supply loss reduction is set to 0.4; the weight of the heat network hydraulic balance recovery effect is 0.3; the weight of the implementation cost is 0.2; the weight of the impact on users' heat supply experience is 0.1.

[0044] Plan selection: The emergency response collaborative agent quantitatively scores each plan. The comprehensive score of Plan I is 80 points, the comprehensive score of Plan II is 75 points, and the comprehensive score of Plan III is 82 points. Finally, it is determined that integrating Plan I and Plan III is the optimal plan, closing the corresponding valves, starting the standby heat source, allocating heating resources, and mobilizing the emergency repair team and equipment to repair as soon as possible, and outputting it to the heating business personnel. The heating business personnel quickly organize the implementation according to the optimal plan, allocate heating resources, check the branch pipelines, and if economically permitted, provide temporary heating equipment for residents or provide necessary economic compensation, effectively responding to this pipeline rupture emergency event, minimizing the heat supply loss to the greatest extent, and ensuring the stable operation of the heating system and the basic heat supply needs of users.

[0045] In this embodiment, the emergency response dispatching decision-making task is expressed as: ; D is the generated emergency response dispatching plan; is the large model for emergency response dispatching of the heating system; is the goal of quickly reconstructing the heat network hydraulic balance working condition and reducing heat supply loss; is the operation information of the current heating system; is the emergency response dispatching knowledge of the heating system.

[0046] In this embodiment, each emergency response dispatching agent is driven by a pre-selected different large model for emergency response dispatching of the heating system.

[0047] In several embodiments provided in the present application, it should be understood that the disclosed systems and methods can also be implemented in other ways. The system embodiments described above are merely illustrative. For example, the flowcharts and block diagrams in the accompanying drawings show the possible architectures, functions, and operations of systems, methods, and computer program products according to multiple embodiments of the present invention. In this regard, each block in the flowchart or block diagram may represent a module, a program segment, or a part of code, and the module, program segment, or part of code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order than marked in the accompanying drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in the reverse order, depending on the functions involved. It should also be noted that each block in the block diagram and / or flowchart, as well as the combination of blocks in the block diagram and / or flowchart, can be implemented by a dedicated hardware-based system for performing the specified functions or actions, or can be implemented by a combination of dedicated hardware and computer instructions.

[0048] In addition, the functional modules in each embodiment of the present invention can be integrated together to form an independent part, or each module can exist alone, or two or more modules can be integrated to form an independent part. If the function is implemented in the form of a software functional module and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods in each embodiment of the present invention. The aforementioned storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memories, random access memories, magnetic disks, or optical discs that can store program codes.

[0049] Enlightened by the above-described ideal embodiments of the present invention, through the above description, relevant staff can completely make various changes and modifications without departing from the technical idea of this invention. The technical scope of this invention is not limited to the content in the specification, and its technical scope must be determined according to the scope of the claims.

Claims

1. A multi-stage emergency response scheduling method for a heating system containing a large model, characterized in that: include: Emergency prevention and preparation stage, emergency response stage, emergency disposal and aftermath stage; The emergency preparedness phase includes: Based on the big model, a Q&A database of multiple types of emergency events in the heating system is constructed. The database generates targeted questions and answers by extracting key event information of emergency disposal and dispatching data of the heating system. The targeted questions and answers include intelligent Q&A of emergency knowledge. The database provides cause analysis of emergency events, risk hidden danger query and recommendation of emergency disposal and dispatching measures. The emergency response phases include: Use prompt word engineering to guide the large model to output the optimal feature combination that characterizes various types of emergency events, and identify various types of emergency events; Assess the heat loss caused by the identified emergency event, and set emergency response scheduling decision tasks based on the goal of quickly reconstructing the hydraulic balance of the heat network and reducing heat loss; Multiple emergency response dispatch agents respectively output emergency response dispatch plans based on the emergency response dispatch decision-making tasks; Each agent judges the emergency response dispatch plan output by other agents; Obtain the optimal emergency response scheduling plan based on the emergency response collaborative agent and the judgment and analysis of each agent; The emergency response and aftermath phase includes: The emergency response and dispatch process is analyzed and summarized using a large model, and emergency response and dispatch plan simulation is carried out using intelligent agents to strengthen the emergency response and dispatch process.

2. The multi-stage emergency response scheduling method for a heating system with a large model according to claim 1 is characterized in that: In the emergency response stage, a feature combination workflow is set up to output various types of emergency events, and the feature combination workflow includes uploading emergency response scheduling data sets, emergency event description, task requirement description for feature extraction using machine learning, selecting feature combination mechanism, and outputting optimal feature combination; Use prompt word engineering to guide the large model to implement the workflow of output feature combination in sequence, and output the optimal feature combination to characterize various types of emergency events; The optimal feature combination of each type of emergency event is input into the preset machine learning model for training and learning, and a heating emergency event recognition model is established to identify the types of emergency events occurring in the heating system.

3. The multi-stage emergency response scheduling method for a heating system with a large model according to claim 2 is characterized in that: The construction of the heating emergency event recognition model includes: setting a two-layer machine learning model to extract features from the emergency event data set, the input end of the first layer of the machine learning model is the original emergency disposal scheduling data set, and the features of the original emergency disposal scheduling data set are abstracted through unsupervised learning; The second-layer machine learning model uses local feature extraction capabilities to perform deep extraction of each data feature; The cross-validation method is used to evaluate different feature combinations, and the feature combinations are optimized and iterated according to the evaluation results to output the optimal feature combination for emergency event identification in the heating system.

4. A multi-stage emergency response dispatching method for a heating system with a large model according to any one of claims 1 to 3, characterized in that: The heating loss assessment step includes: constructing an economic loss calculation model, and calculating the economic losses caused by the emergency event based on the data identified in the emergency event.

5. The multi-stage emergency response scheduling method for a heating system with a large model according to claim 4 is characterized in that: The data in the emergency event include the number of affected heat users, heating area, missing value of heat load, and emergency repair time.

6. The multi-stage emergency response scheduling method for a heating system with a large model according to claim 5 is characterized in that: The method for constructing an economic loss calculation model includes taking the number of affected heat users, the duration of heating interruption, the unit price of heating fee, etc. as independent variables and the economic loss value as the dependent variable based on the economic loss data of historical emergency events and the currently determined loss factors, training the model to learn the relationship between the loss factors and the economic losses, and constructing an economic loss calculation model.

7. A multi-stage emergency response scheduling method for a heating system with a large model according to claim 1, characterized in that: The method of quickly reconstructing the hydraulic balance condition of the heating network based on the heating network hydraulic model includes constructing the heating network hydraulic model based on the heating network topological structure data, fluid mechanics principles and historical operation data, and calculating the impact of emergency events on the hydraulic balance of the heating network.

8. A multi-stage emergency response scheduling method for a heating system with a large model according to any one of claims 1-3 or 5-7, characterized in that: A multi-level emergency response and dispatch strategy is set up: in the first layer, multiple emergency response and dispatch intelligent agents are set up to analyze and reason about the emergency event information of the heating system respectively, and output their own emergency response and dispatch plans; in the second layer, each emergency response and dispatch intelligent agent exchanges information, and judges, discusses and puts forward opinions on the emergency response and dispatch plans output by other emergency response and dispatch intelligent agents respectively; in the third layer, the set emergency response collaborative intelligent agent integrates the current emergency event problem and the inferences of each emergency response and dispatch intelligent agent for judgment and analysis, and obtains the optimal emergency response and dispatch plan.

9. The multi-stage emergency response and dispatching method for a heating system according to claim 8, characterized in that: Said The emergency response dispatch decision task is expressed as: ; Where: D To generate emergency response dispatch plan; A large-scale model for emergency response and dispatching of heating systems; To quickly reconstruct the hydraulic balance of the heating network and reduce the heat loss; Provides current heating system operation information; Provide emergency response and dispatch knowledge for heating systems.

10. The multi-stage emergency response scheduling method for a heating system with a large model according to claim 1 is characterized in that: It also includes extracting key event information from the heating system emergency response and dispatching data, generating targeted questions and answers, and using a large model to automatically build a Q&A database for multiple types of emergency events in the heating system.

Images