An intelligent heating decision-making system based on large artificial intelligence models

Abstract

The present invention discloses a heating intelligent decision-making system based on an artificial intelligence large model, including: a data layer for storing various types of data of the source-network-load-storage of the heating system; a knowledge base establishment and large model training layer for establishing a diversified task knowledge base for heating intelligent decision-making and establishing large models for each task; a human-machine enhanced optimization large model layer for putting forward improvement suggestions to the large model, optimizing the generated answers, evaluating multiple answers generated by the large model, and performing parameter optimization training on the large model; an intelligent agent decision-making layer for integrating the large models of each task into each intelligent agent for environmental perception, collaborative interaction, and autonomous adaptive decision-making; a simulation layer for providing a simulation operation environment, a user interaction interface, and dynamic parameter simulation of scenarios; an intelligent agent evaluation and optimization layer for constructing performance indicators for evaluating each intelligent agent, and evaluating and optimizing the behavior and decision-making of each intelligent agent.

Description

Technical Field

[0001] The present invention belongs to the technical field of intelligent heating systems, and particularly relates to a heating intelligent decision-making system based on an artificial intelligence large model. Background Art

[0002] Intelligent decision-making for a heating system refers to the process of using advanced information technology and intelligent algorithms to make autonomous decisions on aspects such as the operation monitoring, fault diagnosis, and optimal scheduling of the heating system, so as to achieve the safe, efficient, energy-saving, and reliable operation of the heating system.

[0003] A large language model is a large-scale natural language processing model based on deep learning technology, capable of understanding and generating natural language text, and possessing powerful language expression and reasoning capabilities. Applying the large language model in an intelligent agent, with the large language model serving as the brain of the intelligent agent, providing language understanding, reasoning capabilities, and a knowledge base. Through continuous training and optimization, the large language model can generate more accurate and fluent text, providing richer information and support for the intelligent agent. The intelligent agent commands the large model to complete specific tasks through prompt words, and continuously optimizes its behavior through learning to adapt to different environments and tasks, with higher autonomy and intelligence. However, due to the complexity of the heating system, the current intelligent decision-making for the heating system results in difficulties in integrating heating system data, limited data analysis capabilities, poor model adaptability, and a lack of self-learning capabilities. Therefore, how to apply large model and intelligent agent technologies to heating intelligent decision-making, improve the intelligent level of decision-making, and achieve environmental perception, collaborative interaction, and autonomous adaptive decision-making for the heating system is an urgent problem to be solved currently.

[0004] Based on the above technical problems, it is necessary to design a new heating intelligent decision-making system based on an artificial intelligence large model. 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 heating intelligent decision-making system based on an artificial intelligence large model. By using a knowledge base, large model, and multi-intelligent agent technologies, a multi-level heating intelligent decision-making system is constructed to improve the intelligent level of decision-making for the heating system, enabling the intelligent agent integrating the large model and the knowledge base to have the capabilities of heating system environmental perception, collaborative interaction, and autonomous adaptive decision-making.

[0006] To solve the above technical problems, the technical solution of the present invention is:

[0007] The present invention provides a heating intelligent decision-making system based on an artificial intelligence large model, which includes:

[0008] Data layer: used to collect, store, and manage various types of data of the source-network-load-storage of the heating system, and serve as a large-scale data set for the establishment of the knowledge base and the training layer of the large model;

[0009] Knowledge base establishment and large model training layer: It is used to preprocess, annotate, extract and represent the large-scale data set of the heating system, and then establish a knowledge base for diversified tasks of heating intelligent decision-making; it is also used to train a large model according to the knowledge base for diversified tasks of heating intelligent decision-making and the learning strategy of heating intelligent decision-making prompts. The large model captures the knowledge and behavior characteristics of each task of heating intelligent decision-making, establishes the internal connection between each prompt word and high-quality answer at the same time, and fine-tunes the large model according to the specific scenario requirements, task objectives and constraints of each heating task to establish a large model for each task, adapting to specific business scenarios, decision-making tasks and each intelligent agent;

[0010] Human-machine enhanced optimization large model layer: It is used for the heating dispatching control personnel to put forward improvement suggestions to the large model during the dialogue process of each task of heating intelligent decision-making, and the large model optimizes the generated answers according to the suggestions; it is also used for the heating dispatching control personnel to evaluate multiple answers generated by the large model according to the prompt words, and train and optimize the large model parameters according to the evaluation results;

[0011] Intelligent agent decision-making layer: It is used to integrate the large models for each task of heating intelligent decision-making after training and optimization into each set intelligent agent, use the large model to drive the intelligent agent, and guide each intelligent agent to perform environment perception, collaborative interaction and autonomous adaptive decision-making in the simulation environment;

[0012] Simulation layer: It is used to provide a simulation operation environment, user interaction interface and simulation of scene dynamic parameters during the interaction of heating intelligent decision-making. After the heating dispatching control personnel initiate a question, each intelligent agent generates decision simulation data, supporting the display of answers to users in the form of text, pictures and videos;

[0013] Intelligent agent evaluation and optimization layer: It is used to construct performance indicators for evaluating each intelligent agent, and evaluate and optimize the behavior and decision-making effects of each intelligent agent according to the performance indicators and decision simulation results.

[0014] Furthermore, the multi-type data of the heat supply system source-network-load-storage include: historical operation data of the heat supply system source-network-load-storage, heat supply operation and maintenance and dispatching management manuals, dynamically changing parameters, and physical attribute data of each system device;

[0015] Among them, the historical operation data includes: operation parameters of the heat source equipment on the source side, heat metering data of each pipeline in the heat network, indoor temperature and heat consumption data, operation status and heat storage / discharge of the heat storage equipment, historical heat load data, on-demand scheduling decision-making strategies, historical energy consumption analysis data, and abnormal diagnosis data; the heat supply operation and maintenance and scheduling management manual includes heat supply inspection content, equipment maintenance strategies, fault handling procedures, scheduling procedures, and emergency scheduling response handling rules; the dynamically changing parameters include outdoor environmental parameters and sudden abnormal fault parameters; the physical attribute data of each device in the system includes fixed parameters, operation characteristics, and service life of each device in the source-network-load-storage and the pipeline network.

[0016] Furthermore, after performing data preprocessing, data annotation, knowledge extraction, and representation on the large-scale dataset of the heat supply system, a diversified task knowledge base for heat supply intelligent decision-making is established, including:

[0017] Clarify the diversified tasks of heat supply intelligent decision-making, including operation perception, parameter prediction, fault diagnosis, scheduling planning, and energy consumption assessment, and determine the knowledge scope and knowledge types covered by the knowledge base according to the diversified tasks; the knowledge types include factual knowledge, rule-based knowledge, and process knowledge;

[0018] According to the diversified tasks, knowledge scope, and knowledge types of heat supply intelligent decision-making, perform data preprocessing, annotation, and classification on the large-scale dataset of the heat supply system;

[0019] Knowledge extraction: For clear rules and experiences, directly formulate rules for knowledge extraction; use machine learning algorithms to discover potential knowledge from data; use natural language processing technology to extract key information and knowledge for unstructured text data;

[0020] Knowledge representation: Represent the extracted knowledge according to the production rule representation method, frame representation method, and semantic network representation method;

[0021] Store the extracted and represented knowledge into the corresponding database tables or graph structures according to the format and requirements of the knowledge base management system, establish indexes and association relationships between knowledge, and provide a knowledge retrieval interface to support keyword retrieval, semantic retrieval, and association retrieval methods, and finally establish a diversified task knowledge base for heat supply intelligent decision-making.

[0022] Furthermore, perform large model training according to the diversified task knowledge base of heat supply intelligent decision-making and the heat supply intelligent decision-making prompt learning strategy, capture the knowledge and behavior characteristics of each task of heat supply intelligent decision-making through the large model, and at the same time establish the internal connection between each prompt word and high-quality answers, including:

[0023] During the large model training stage, the diversified knowledge base for intelligent heating decision-making is input into the base large model as part of the training corpus to learn the heating professional terms, semantic relationships, and logical relationships between various factors in the heating system in the knowledge base, analyze the operation modes, behavioral characteristics, and intelligent decision-making processes of the heating system under different working conditions, and according to the characteristics of the intelligent heating decision-making tasks, preset the thinking processes and rules in the way of chain of thought, decompose the decision-making task problems into multiple sub-problems, and provide logical and clear steps in the form of prompts, design different types of prompt templates to guide the large model to analyze and process the input information according to the steps and processes;

[0024] The designed prompt templates are combined with the content of the knowledge base and input into the large model to enable the model to learn the ability to extract relevant knowledge from the knowledge base according to the prompt information and generate answers, and collect the feedback information of users on the answers to establish the internal connection between the prompt words and the answers.

[0025] Furthermore, fine-tuning the large model according to the specific scenario requirements, task objectives, and constraint conditions of each heating task to establish each task large model includes:

[0026] For the operation perception task of the heating system, its specific scenario requirement is to perceive the operation states of each device in the heat source, network, load, and storage of the heating system, the task objective is to obtain the operation data of each device comprehensively and accurately, and the constraint condition is that the data acquisition is compatible with the data interfaces of different types and different manufacturers' devices;

[0027] For the parameter prediction task of the heating system, its specific scenario requirement is to predict the heat load and heat network parameters, the task objective is to predict in advance and accurately predict the changes in the operation parameters of the heating system, and the constraint condition is to consider the influencing factors of weather changes, user heat consumption behavior, and historical data laws;

[0028] For the fault diagnosis task of the heating system, its specific scenario requirement is to diagnose the types of faults occurring in the heating system and judge the causes of the faults, the task objective is to accurately identify the fault types and locate the causes of the faults, and the constraint condition is the complexity of the fault scenarios;

[0029] For the dispatching and planning task of the heating system, its specific scenario requirement is to generate dispatching schemes with different time scales for each device in the heat source, network, load, and storage according to the supply-demand relationship of the heat load, the task objective is to meet the heating demand, minimize the economic cost, and have low energy consumption, and the constraint conditions are the operation parameter constraints of each device and the heat power balance constraint;

[0030] For the energy consumption assessment task of the heating system, its specific scenario requirement is to evaluate the energy consumption situation of the heating system, analyze the influencing factors of energy consumption, and put forward energy-saving suggestions, the task objective is to reduce the energy consumption of the heating system, and the constraint condition is to consider the energy consumption differences under different operation conditions;

[0031] Fine-tune the large model according to the specific scenario requirements, objective functions, and constraints of each task to establish large models for each task, including operation perception large model, parameter prediction large model, fault diagnosis large model, scheduling planning large model, and energy consumption assessment large model.

[0032] Furthermore, during the dialogue process of each task in the intelligent heating decision-making, the heating scheduling control personnel put forward improvement suggestions to the large model, and the large model optimizes the generated answers according to the suggestions, including:

[0033] During the dialogue process of each task in the intelligent heating decision-making, the heating scheduling control personnel can input improvement suggestions through a text box or voice, and use the natural language understanding function to convert the input improvement suggestions into instructions that the large model can understand;

[0034] The large model adjusts the internal knowledge base and reasoning logic according to the improvement suggestions to optimize the generated answers.

[0035] Furthermore, the heating scheduling control personnel evaluate multiple answers generated by the large model according to the prompt words, and train and optimize the large model parameters according to the evaluation results, including:

[0036] The heating scheduling control personnel use the trained large model to generate multiple different answers according to the prompt words, construct evaluation indicators according to language comprehensibility indicators, semantic rationality indicators, context coherence indicators, heating knowledge accuracy indicators, and heating reasoning ability indicators, and evaluate and assign values to the multiple different answers;

[0037] Take different questions, answers, and evaluation assignment results as inputs, and use the reinforcement learning algorithm to update and optimize the large model parameters.

[0038] Furthermore, integrate the large models of each task of the trained and optimized intelligent heating decision-making into the set intelligent agents, and use the large model to drive the intelligent agents to guide the intelligent agents to perform environment perception, collaborative interaction, and autonomous adaptive decision-making in the simulation environment, including:

[0039] Set corresponding task intelligent agents according to the diverse tasks of intelligent heating decision-making, including operation perception intelligent agent, parameter prediction intelligent agent, fault diagnosis intelligent agent, scheduling planning intelligent agent, and energy consumption assessment intelligent agent;

[0040] Set up bottom-layer device intelligent agents, including heat source intelligent agent, heat network intelligent agent, heat storage intelligent agent, and heat user intelligent agent, which respectively collect data of heat source devices, heat networks, heat storage devices, and heat users and transmit them to each task intelligent agent, as well as issue and execute decision instructions transmitted by each task intelligent agent;

[0041] Optimize the input and output interfaces of the large models for each task to interface with the intelligent agents of each task in heat supply intelligent decision-making, and allocate large model call channels for each task intelligent agent, integrating the large models for each task into the corresponding task intelligent agents;

[0042] Each task intelligent agent obtains the collected data through its perception unit for environmental perception, extracts the key information of the data, and transmits it to the large model of the decision-making unit. At the same time, a collaborative interaction mechanism is established between the task intelligent agent and the underlying device intelligent agent for collaborative interaction of data, optimization of task execution, and task resource allocation. Finally, through in-depth analysis and inference learning of the large model of the decision-making unit, each task intelligent agent makes autonomous decisions to generate decision instructions to adapt to the dynamically changing operating conditions of the heat supply system.

[0043] Furthermore, the simulation operation environment provided during heat supply intelligent decision-making interaction includes: managing the simulation time to analyze the operating state of the heat supply system at different times, updating the state of the simulation environment, and providing a flexible time control interface to allow immediate operation of the simulation time during system operation;

[0044] The user interaction interface provided during heat supply intelligent decision-making interaction includes: providing a graphical parameter configuration interface, setting simulation parameters through slider, drop-down menu, text box interaction elements, supporting the loading of parameter configuration files, and visualizing and analyzing the simulation results in the form of text, pictures, and videos;

[0045] The simulation of dynamic parameters of the scenario provided during heat supply intelligent decision-making interaction includes: using the large model to construct a simulation model of the actual heat supply system and simulating the system in a dynamic environment.

[0046] Furthermore, the construction of performance indicators for evaluating each intelligent agent, and based on the performance indicators and decision simulation results, evaluating and optimizing the behavior and decision-making effects of each intelligent agent, including:

[0047] Construct performance indicators for evaluating each task intelligent agent in intelligent heat supply intelligent decision-making, including task completion rate, task completion efficiency, decision adaptability, and cooperation between intelligent agents;

[0048] According to the behavior trajectories, decision-making processes, and output decision simulation results of each task intelligent agent, calculate the performance indicator values of each task intelligent agent, and compare them with the preset standards to evaluate the performance of each task intelligent agent;

[0049] According to the evaluation results, optimize and adjust the behavior and decision-making strategies of each task intelligent agent, including: improving the algorithms and model structures within the intelligent agent, enhancing the environmental perception ability and adaptability of the intelligent agent, and optimizing the cooperation mechanism and communication protocol between intelligent agents.

[0050] The beneficial effects of the present invention are:

[0051] (1) By collecting various types of data from the source-network-load-storage of the heating system and preprocessing, annotating, extracting knowledge and representing a large-scale data set, the present invention establishes a knowledge base for diverse tasks of heating intelligent decision-making, integrates professional knowledge and experience in the field of heating systems, provides a knowledge source for the training of large models, and helps the large models better understand and process heating-related decision-making tasks;

[0052] (2) The present invention trains large models according to the knowledge base for diverse tasks of heating intelligent decision-making and prompting learning strategies, enabling the large models to capture the knowledge and behavioral characteristics of each task of heating intelligent decision-making, and at the same time establishing an internal connection between the prompt words and high-quality answers. This enables the large models to generate more accurate and practical answers and decision-making suggestions for different tasks of the heating system;

[0053] (3) The present invention fine-tunes the large models according to the specific scenario requirements, task objectives and constraints of each heating task, and establishes large models for each task that are adapted to specific business scenarios, decision-making tasks and agents. This fine-tuning mechanism improves the pertinence and adaptability of the large models and can better meet the complex and changeable actual needs of the heating system;

[0054] (4) The present invention establishes a human-machine enhanced optimization large model layer, enabling the large models to continuously learn and improve, gradually improving the accuracy and reliability of decision-making, and realizing the dynamic optimization of the models;

[0055] (5) The present invention integrates the large models for each task after training and optimization into the agents, and uses the large models to drive the agents, enabling the agents to perform efficient environment perception, collaborative interaction and autonomous adaptive decision-making in the simulation environment. The agents can quickly make reasonable decisions according to the actual operation conditions of the heating system, improving the intelligent level and response speed of the heating system;

[0056] (6) The simulation layer of the present invention provides a simulation operation environment, a user interaction interface and simulation of dynamic scenario parameters during heating intelligent decision-making interaction, and supports presenting answers to users in various ways. This enables heating dispatch control personnel to more intuitively understand the decision-making process and results of the system, facilitating operation and adjustment, and enhancing the user experience and decision-making efficiency;

[0057] (7) The agent evaluation and optimization layer of the present invention constructs performance indicators for evaluating each agent, and can comprehensively and objectively evaluate the behavior and decision-making effects of the agents according to the performance indicators and decision-making simulation results. This helps to discover the problems and deficiencies of the agents and provides a basis for optimization and adjustment.

[0058] Other features and advantages will be set forth in the following description, and in part will be obvious from the description, or may be learned by practice of the invention. The objects and other advantages of the invention are realized and attained by the structure particularly pointed out in the specification and the drawings.

[0059] To make the above objects, features and advantages of the present invention more comprehensible, preferred embodiments accompanied by the drawings are described in detail as follows. Brief Description of the Drawings

[0060] In order 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 drawings required for the description of the specific embodiments or the prior art. Obviously, the drawings in the following description are some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0061] Figure 1 It is a block diagram of a heating intelligent decision-making system based on an artificial intelligence large model of the present invention;

[0062] Figure 2 It is a schematic diagram of the heating intelligent decision-making principle based on an artificial intelligence large model of the present invention;

[0063] Figure 3 It is a schematic block diagram of the heating fault diagnosis principle based on an artificial intelligence large model of the present invention. Detailed Description of the Embodiments

[0064] To make the objectives, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below with reference to the drawings. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art without creative efforts based on the embodiments of the present invention belong to the scope of protection of the present invention.

[0065] As Figure 1 shown, this embodiment provides a heating intelligent decision-making system based on an artificial intelligence large model, which includes:

[0066] Data layer: used to collect, store and manage various types of data of the heat supply system's source network load storage, and serve as a large-scale data set for knowledge base establishment and large model training layer;

[0067] Knowledge base establishment and large model training layer: It is used to preprocess, annotate, extract and represent knowledge from a large-scale dataset of the heating system, and then establish a knowledge base for diverse tasks of heating intelligent decision-making. It is also used to train a large model according to the knowledge base for diverse tasks of heating intelligent decision-making and the learning strategy of heating intelligent decision-making prompts. The large model captures the knowledge and behavior characteristics of each task of heating intelligent decision-making, establishes the internal connection between each prompt word and high-quality answers at the same time, and fine-tunes the large model according to the specific scenario requirements, task objectives and constraint conditions of each heating task to establish large models for each task, adapting to specific business scenarios, decision-making tasks and each intelligent agent;

[0068] Human-machine enhanced and optimized large model layer: It is used for the heating dispatch control personnel to put forward improvement suggestions to the large model during the dialogue process of each task of heating intelligent decision-making, and the large model optimizes the generated answers according to the suggestions. It is also used for the heating dispatch control personnel to evaluate multiple answers generated by the large model according to the prompt words, and train and optimize the large model parameters according to the evaluation results;

[0069] Intelligent agent decision-making layer: It is used to integrate the large models for each task of heating intelligent decision-making after training and optimization into each set intelligent agent, use the large model to drive the intelligent agent, and guide each intelligent agent to perform environment perception, collaborative interaction and autonomous adaptive decision-making in the simulation environment;

[0070] Simulation layer: It is used to provide a simulation operation environment, user interaction interface and simulation of scene dynamic parameters during the interaction of heating intelligent decision-making. After the heating dispatch control personnel initiate a question, each intelligent agent generates decision simulation data, supporting the display of answers to users in the form of text, pictures and videos;

[0071] Intelligent agent evaluation and optimization layer: It is used to construct performance indicators for evaluating each intelligent agent, and evaluate and optimize the behavior and decision-making effects of each intelligent agent according to the performance indicators and decision simulation results.

[0072] Such as Figure 2The overall technical principle of the intelligent heating decision-making is as follows. The heating system consists of a heat source plant, a primary network, heat substations, heat storage devices, and heat users. These components generate a large amount of data sources, providing the original data for subsequent intelligent decision-making. Foundation large model: It is the basic model that provides general artificial intelligence capabilities and learning frameworks, offering underlying technical support for the large model of intelligent heating decision-making. Large model of intelligent heating decision-making: Based on the foundation large model, combined with the large amount of data sources generated by the heating system and various task knowledge sources in the knowledge base for diverse tasks of intelligent heating decision-making (factual, rule-based, and procedural knowledge for tasks such as operation perception, parameter prediction, fault diagnosis, scheduling planning, and energy consumption assessment), it is trained and optimized to possess the intelligent decision-making ability in the heating field. Integrate the large model of intelligent heating decision-making into the heating intelligent agent, and use the large model to drive the intelligent agent, enabling the intelligent agent to have the capabilities of environmental perception, collaborative interaction, and autonomous adaptive decision-making, realizing the intelligent control and management of the heating system, and improving the operation efficiency, reliability, and energy-saving effect of the heating system.

[0073] As Figure 3 shown in the principle of heating fault diagnosis based on the large model of artificial intelligence. Taking the heating fault diagnosis task as an example, the technical principles of large model training and large model application mainly include:

[0074] 1) Foundation large model and data input: The foundation large model is the cornerstone of the entire technical system, providing general artificial intelligence capabilities. Multiple data sources such as fault diagnosis teaching materials data, regulation data, scientific research data, and historical data of the heating system are input. These data contain various information during the operation of the heating system, such as equipment status, fault records, etc. The heating natural language large model is constructed based on the foundation large model. Through learning the input data, it has the capabilities of semantic understanding, knowledge reasoning, content generation, and code generation, and can process and analyze natural language information related to the heating system.

[0075] 2) Fault diagnosis sample data and model fine-tuning: Fault diagnosis reports on the source-network-load side, fault diagnosis regulation documents, operation manuals, etc. constitute the fault diagnosis sample data. Using methods such as model fine-tuning, prompt engineering, and generative retrieval, based on the heating natural language large model and the fault diagnosis sample data, a large model for heating system fault diagnosis is trained. This process enables the model to be optimized for the heating system fault diagnosis task and better understand and handle specific problems in this field.

[0076] 3) The role of the knowledge base: The knowledge base covers the equipment operation instruction knowledge base, equipment fault problem knowledge base, equipment fault handling knowledge base, etc. These knowledge bases provide rich professional knowledge for the large model of heating system fault diagnosis. When the model conducts fault diagnosis, it can obtain relevant information from the knowledge base to assist in making accurate judgments.

[0077] 4) Heating Fault Diagnosis Agent and Application Interface: The heating fault diagnosis agent is constructed based on the large model for heating system fault diagnosis and has the capabilities of fault diagnosis decision-making, perception and insight, learning and evolution, and control execution. Components such as interface service APIs, service encapsulation SDKs, and service plugins provide the agent with a way to interact with external systems, enabling it to play a role in the actual heating system and achieve the fault diagnosis function.

[0078] The heating fault diagnosis task has the following advantages through large model training and large model application:

[0079] 1) Precise Fault Diagnosis: Through the training of a large amount of historical data and professional knowledge, the large model for heating system fault diagnosis can accurately identify various faults in the heating system. Combining the information in the knowledge base, it improves the accuracy and reliability of fault diagnosis, reducing the situations of misjudgment and missed judgment.

[0080] 2) Intelligent Decision Support: The heating fault diagnosis agent has the decision-making ability and can provide reasonable handling suggestions and decision support according to the diagnosis results, helping operation and maintenance personnel quickly and effectively solve fault problems and improving the stability and reliability of the heating system.

[0081] 3) Self-Evolution and Learning: The agent has the ability of learning and evolution. It can continuously optimize and improve the fault diagnosis model with the continuous input of new data and the operation of the system, adapt to the changes and developments of the heating system, and continuously enhance the ability and level of fault diagnosis.

[0082] 4) Convenient System Integration: The rich interface components enable the heating fault diagnosis agent to be easily integrated with other heating systems or management platforms, realizing data sharing and collaborative work, and improving the intelligent management level and operation efficiency of the entire heating system.

[0083] 5) Knowledge Accumulation and Inheritance: The existence of the knowledge base not only provides support for model training and fault diagnosis, but also plays a role in knowledge accumulation and inheritance, facilitating new operation and maintenance personnel to learn and understand the relevant knowledge and experience of heating system fault diagnosis.

[0084] In this embodiment, the multi-type data of the heating system source-network-load-storage includes: historical operation data of the heating system source-network-load-storage, heating operation and scheduling management manuals, dynamically changing parameters, and physical attribute data of each device in the system.

[0085] Among them, the historical operation data includes: the operation parameters of the heat source equipment on the source side, the heat metering data of each pipeline in the heat network, the indoor temperature and heat consumption data, the operation status and heat storage / discharge of the heat storage equipment, the historical heat load data, the on-demand scheduling decision strategy, the historical energy consumption analysis data, and the abnormal diagnosis data; the heat supply operation and maintenance and scheduling management manual includes the heat supply inspection content, equipment maintenance strategy, fault handling process, scheduling process, and emergency scheduling response handling rules; the dynamic change parameters include outdoor environmental parameters and sudden abnormal fault parameters; the physical attribute data of each device in the system includes the fixed parameters, operation characteristics, and service life of each device of the source-network-load-storage and the pipe network.

[0086] In this embodiment, after performing data preprocessing, data annotation, knowledge extraction, and representation on the large-scale dataset of the heat supply system, a diversified task knowledge base for heat supply intelligent decision-making is established, including:

[0087] Define diversified tasks for heat supply intelligent decision-making, including operation perception, parameter prediction, fault diagnosis, scheduling planning, and energy consumption assessment, and determine the knowledge scope and knowledge types covered by the knowledge base based on the diversified tasks; the knowledge types include factual knowledge, rule-based knowledge, and process knowledge;

[0088] According to the diversified tasks, knowledge scope, and knowledge types of heat supply intelligent decision-making, perform data preprocessing, annotation, and classification on the large-scale dataset of the heat supply system;

[0089] Knowledge extraction: For explicit rules and experiences, directly formulate rules for knowledge extraction; use machine learning algorithms to discover potential knowledge from data; use natural language processing techniques to extract key information and knowledge for unstructured text data;

[0090] Knowledge representation: Represent the extracted knowledge according to the production rule representation method, frame representation method, and semantic network representation method;

[0091] Store the extracted and represented knowledge into the corresponding database tables or graph structures according to the format and requirements of the knowledge base management system, establish indexes and association relationships between knowledge, provide a knowledge retrieval interface, support keyword retrieval, semantic retrieval, and association retrieval methods, and finally establish a diversified task knowledge base for heat supply intelligent decision-making.

[0092] In actual applications, the specific process of establishing a task knowledge base includes:

[0093] 1) Define diversified tasks, knowledge scope, and types of heat supply intelligent decision-making

[0094] For example, for the parameter prediction task: predicting heating load, heat network parameters (such as temperature, flow rate, pressure change trends); knowledge scope: historical heating data, meteorological data, building characteristic data, user heat consumption behavior data, etc.; factual knowledge: the maximum heating load in a certain heat network area reaches 100 MW; rule-based knowledge: when the outdoor temperature drops by 1 °C, the heating load in this area is expected to increase by 5 MW; procedural knowledge: the parameter prediction process, including steps of data preprocessing, model selection, training, and evaluation.

[0095] For the fault diagnosis task: quickly and accurately identifying heating system equipment failures (such as pipeline leaks, equipment damage, etc.) and the causes of failures; knowledge scope: equipment failure phenomena, causes of failures, maintenance methods, etc.; factual knowledge: a certain heat network pipeline shows the phenomena of sudden pressure drop and abnormal increase in flow rate; rule-based knowledge: if the heat network pressure drops sharply and the flow rate increases abnormally, it may be a pipeline leak; procedural knowledge: the fault diagnosis process, a series of steps from fault detection to location, cause analysis, and then to repair.

[0096] For the scheduling and planning task: optimizing the scheduling strategies of heat sources, heat networks, and user sides according to information such as heating demand, energy supply, and equipment status; knowledge scope: heat source heating capacity, heat network hydraulic characteristics, user heating demand curve, energy price, etc.; factual knowledge: the maximum heating capacity of a certain thermal power plant is 200 MW; rule-based knowledge: when the energy price rises by 10%, preferentially dispatch heat source equipment with low energy consumption; procedural knowledge: the heating scheduling process, the process of formulating and implementing a scheduling plan based on load prediction and equipment status assessment.

[0097] 2) Data preprocessing, annotation, and classification

[0098] Data preprocessing: Clean the collected raw data to remove noise and outliers. For example, for the temperature data collected by sensors, if there are values that significantly exceed the reasonable range (such as a temperature of -100 °C), they are judged as outliers and corrected or deleted.

[0099] Data annotation: Annotate the data according to different tasks. For example, in the fault diagnosis task, data with sudden pressure drop and abnormal increase in flow rate of the heat network are annotated as "suspected pipeline leak data"; in the parameter prediction task, historical heating load data are annotated as "actual heating load", and at the same time, relevant information such as the corresponding meteorological conditions is annotated.

[0100] Data classification: Classify and store the data according to the task type. For example, store the operation perception data in the "operation perception data" table, and store the parameter prediction-related data in the "parameter prediction data" table, and so on.

[0101] 3) Knowledge extraction

[0102] Explicit rule and experience extraction: For example, in fault diagnosis, based on long-term accumulated experience, the rule is defined as "When the boiler water level continuously drops below the lowest water level line and exceeds 5 minutes, it is determined as a boiler water shortage fault", and this rule is directly applied to knowledge extraction.

[0103] Knowledge extraction using machine learning algorithms: For example, in the parameter prediction task, the time series analysis algorithm (such as the ARIMA model) is used to discover potential change patterns from historical heating load data, and knowledge such as "The heating load shows different fluctuation patterns on weekdays and weekends" is extracted.

[0104] Knowledge extraction using natural language processing: When processing unstructured text data such as equipment maintenance records, methods of named entity recognition and relationship extraction in natural language processing technology are used. For example, key information and knowledge such as "Boiler No. 1", "Burner components", and "Insufficient combustion" are extracted from "The No. 1 boiler was repaired today, and the burner components were replaced because of insufficient combustion".

[0105] 4) Knowledge representation

[0106] Production rule representation method: For example, in scheduling planning, the rule "If the outdoor temperature is lower than -10°C and the heating load exceeds 80% of the rated load, then start the standby heat source" can be expressed as: IF (outdoor temperature < -10°C) AND (heating load > 0.8 * rated load) THEN start the standby heat source.

[0107] Frame representation method: For the heat source equipment boiler, the following frame can be constructed:

[0108] Frame name: Boiler;

[0109] Slot name 1: Model, value: [specific model];

[0110] Slot name 2: Rated power, value: [X] MW;

[0111] Slot name 3: Operating status, value: [normal / fault];

[0112] Slot name 4: Fault description, value: [if there is a fault, fill in the specific fault information].

[0113] Semantic network representation method: Construct a semantic network for heat network faults. The "heat network pipeline" node is connected to nodes such as "sudden pressure drop", "abnormal increase in flow rate", and "pipe leakage" through relationship edges, clearly showing their semantic relationships.

[0114] 5) Knowledge storage and retrieval

[0115] Storage: Store the extracted and represented knowledge into corresponding database tables or graph structures. For example, store the factual knowledge of operation perception in the "Operation Perception Information Table" of a relational database; store the knowledge of fault diagnosis in a graph database in a graph structure to facilitate the display of the association relationships between knowledge.

[0116] Indexing and Association Establishment: Establish indexes for knowledge. For example, in operation perception data, use the device number and monitoring time as indexes for quick query. At the same time, establish association relationships between knowledge. For example, in fault diagnosis knowledge, establish associations between fault phenomena, causes, and repair methods.

[0117] Knowledge Retrieval: Provide a knowledge retrieval interface that supports multiple retrieval methods. For example, when a user enters the keyword "pipe leakage", relevant fault phenomena, causes, and repair knowledge related to pipe leakage can be quickly obtained through keyword retrieval; through semantic retrieval, when entering "abnormal situation in the heat supply network", the system can understand the semantics and return relevant fault diagnosis knowledge; association retrieval can retrieve relevant scheduling strategies, energy consumption evaluations, etc. based on the operating status of a certain device.

[0118] In this embodiment, the large model is trained according to the diversified task knowledge base of heating intelligent decision-making and the learning strategy of heating intelligent decision-making prompts. The large model captures the knowledge and behavior characteristics of each task in heating intelligent decision-making, and at the same time establishes the internal connection between each prompt word and high-quality answers, including:

[0119] In the large model training stage, the diversified knowledge base of heating intelligent decision-making is input into the base large model as part of the training corpus to learn the heating professional terms, semantic relationships, and logical relationships between various factors in the heating system in the knowledge base, analyze the operating modes, behavior characteristics, and intelligent decision-making processes of the heating system under different working conditions, and according to the characteristics of heating intelligent decision-making tasks, preset the thinking process and rules in the way of chain of thought, decompose the decision-making task problems into multiple sub-problems, and provide logically clear steps in the form of prompts, design different types of prompt templates to guide the large model to analyze and process the input information according to the steps and processes;

[0120] Combine the designed prompt templates with the content of the knowledge base and input them into the large model to enable the model to learn the ability to extract relevant knowledge from the knowledge base according to the prompt information and generate answers, and collect the feedback information of users on the answers to establish the internal connection between the prompt words and the answers.

[0121] In practical applications, for the study of heating-related technical terms and semantic relationships: For example, in a diverse knowledge base for heating intelligent decision-making, there are technical terms such as "primary network", "secondary network", and "heat exchange station". When the base large model is trained, by learning the corpus in the knowledge base, it understands that the "primary network" is a high-temperature hot water transmission pipeline network from the heat source to the heat exchange station, and the "secondary network" is the pipeline network from the heat exchange station to the user end. The two exchange heat and adjust parameters through the "heat exchange station", thus mastering the semantic relationships between these terms.

[0122] For the term "heat medium", the model learns that it is the medium for transferring heat. In a heating system, the common heat media are hot water or steam, and it understands the impact of changes in parameters such as the temperature and pressure of the heat medium on the heating effect, establishing a semantic connection between the term and the actual heating physical quantities.

[0123] For the study of the logical relationships of heating system factors: For example, when there is a description of the logical relationship "when the outdoor temperature drops, the heating load increases" in the knowledge base. After learning this knowledge, when the large model encounters input information about changes in the outdoor temperature, it can infer the possible change trend of the heating load based on the learned logical relationship, thus providing a basis for heating intelligent decision-making.

[0124] Another example is that after learning the logical chain "the output of the heat source equipment determines the heat supply of the heat network, and the flow rate and temperature distribution of the heat network affect the heating effect at the user end", the model can understand the interaction relationships between the key factors in the heating system and conduct a comprehensive analysis when dealing with relevant decision-making problems.

[0125] For the study of operation modes and behavior characteristics: Under different working conditions, the operation modes of the heating system are different. For example, during the severe winter period, the heat source equipment needs to maintain a high output, and the supply water temperature and flow rate of the heat network need to be maintained at a high level to meet the heating needs of users. The large model learns the characteristics of this operation mode during the severe winter period from the relevant cases and data in the knowledge base, including the operating parameter ranges of the equipment, adjustment strategies, etc.

[0126] For example, for the behavior characteristics of the heating system under equipment failure conditions, such as when there is a pipeline leak, the pressure of the heat network will drop and the flow rate will show abnormal changes. After learning these characteristics, when the large model encounters similar input of parameter changes, it can judge the possible fault conditions and provide support for fault diagnosis and emergency decision-making.

[0127] Learning for intelligent decision-making processes: Taking the heating scheduling decision as an example, the decision-making process includes formulating a heating plan for heat sources based on the load prediction results, adjusting the flow distribution of the heat network, and making real-time adjustments according to user feedback. After learning this decision-making process, when faced with a scheduling decision task, the large model can analyze and process according to this process, starting from the load prediction data, and gradually generating a reasonable scheduling strategy.

[0128] It should be noted that the thought chain presetting and problem decomposition: For example, when dealing with the decision-making task of "how to ensure heating quality in extremely low-temperature weather", the thinking process is preset in the way of thought chain. First, the problem is decomposed into several sub-problems: how will the heating load change in extremely low-temperature weather (sub-problem 1)? Whether the current heat source's heating capacity meets the demand (sub-problem 2)? Whether the transmission capacity of the heat network needs to be adjusted (sub-problem 3)? Whether the heating facilities at the user end need special maintenance (sub-problem 4)?

[0129] For each sub-problem, corresponding prompting rules are designed. For example, for sub-problem 1, the prompting rule can be "Refer to the heating load data in historical extremely low-temperature weather, combine the building conditions and population distribution in the current area, and analyze the changing trend of the heating load".

[0130] Prompt template design and application: For example, design a prompt template for fault diagnosis: "When [specific fault phenomenon occurs, such as the sudden drop in the heat network pressure], first check [related equipment or system, such as the connection of the heat network pipeline], then check [other possible influencing factors, such as the status of the pressure sensor], and finally judge the possible cause of the fault according to the inspection results as [list the possible causes of the fault, such as pipeline leakage, sensor failure]".

[0131] Combine such a prompt template with the content about heat network faults in the knowledge base and input it into the large model. When the model encounters the input information of the sudden drop in the heat network pressure, it will extract relevant knowledge about heat network pipeline connections, pressure sensors, etc. from the knowledge base according to the steps of the prompt template, conduct analysis and judgment, and generate an answer for fault diagnosis.

[0132] In this embodiment, the fine-tuning of the large model according to the specific scenario requirements, task objectives, and constraint conditions of each heating task to establish each task large model includes:

[0133] For the heating system operation perception task, its specific scenario requirement is to perceive the operation status of each device in the heat source-network-load-storage of the heating system, the task objective is to comprehensively and accurately obtain the operation data of each device, and the constraint condition is that the data collection is compatible with the data interfaces of different types and different manufacturers' devices;

[0134] For the task of predicting heating system parameters, its specific scenario requirements are to predict the heat load and heat network parameters. The task objective is to predict in advance and accurately the changes in the operating parameters of the heating system. The constraints are to consider the impact factors of weather changes, user heat consumption behavior, and historical data patterns;

[0135] For the task of diagnosing heating system faults, its specific scenario requirements are to diagnose the types of faults occurring in the heating system and determine the causes of the faults. The task objective is to accurately identify the fault types and locate the causes of the faults. The constraints are the complexity of the fault scenarios;

[0136] For the task of scheduling and planning the heating system, its specific scenario requirements are to generate scheduling plans for each device of the source, network, load, and storage at different time scales according to the supply-demand relationship of the heat load. The task objectives are to meet the heating demand, minimize the economic cost, and have low energy consumption. The constraints are the operating parameter constraints of each device and the heat power balance constraint;

[0137] For the task of evaluating the energy consumption of the heating system, its specific scenario requirements are to evaluate the energy consumption of the heating system, analyze the influencing factors of energy consumption, and propose energy-saving suggestions. The task objective is to reduce the energy consumption of the heating system. The constraints are to consider the energy consumption differences under different operating conditions;

[0138] Fine-tune the large model according to the specific scenario requirements, objective functions, and constraints of each task to establish large models for each task, including operation perception large model, parameter prediction large model, fault diagnosis large model, scheduling and planning large model, and energy consumption evaluation large model.

[0139] In this embodiment, during the conversation process of each task of heating intelligent decision-making, the heating dispatch control personnel put forward improvement suggestions to the large model, and the large model optimizes the generated answers according to the suggestions, including:

[0140] During the conversation process of each task of heating intelligent decision-making, the heating dispatch control personnel can input improvement suggestions through a text box or voice, and convert the input improvement suggestions into instructions that can be understood by the large model by using the natural language understanding function;

[0141] The large model adjusts the internal knowledge base and reasoning logic according to the improvement suggestions to optimize the generated answers.

[0142] In actual applications, an example is given to illustrate the optimization of heating dispatch in extreme weather: During an extremely cold weather period, the heating dispatch control center monitored a sharp increase in the heating load, and the existing heating strategies were difficult to meet the demand. At this time, the dispatcher input an improvement suggestion through voice: "It is extremely cold weather now, and the load is increasing too fast. Start all standby heat sources, and give priority to ensuring heating in residential areas. Increase the supply water temperature in residential areas, and at the same time closely monitor the pipe network pressure to avoid overpressure."

[0143] The system utilizes natural language understanding capabilities to convert this voice into instructions that can be understood by the large model:

[0144] 1) Start all standby heat sources;

[0145] 2) Set the heating priority, with residential areas having priority;

[0146] 3) Increase the set value of the water supply temperature in residential areas;

[0147] 4) Monitor the pipe network pressure in real-time. When the pressure approaches the upper limit, issue a warning and adjust the heating strategy.

[0148] Based on these instructions, the large model adjusts the internal knowledge base and reasoning logic. It retrieves knowledge such as the startup process of standby heat sources and the heating parameter standards for residential areas from the knowledge base, adjusts the reasoning logic, and re-plans the heating dispatch plan. In the original plan, the startup conditions of standby heat sources were relatively conservative. After optimization by the large model, the standby heat sources are immediately started, increasing the heating output. At the same time, the heat network flow is reallocated to give priority to heating residential areas and the water supply temperature in residential areas is increased to ensure the warmth inside residents' rooms. During this process, the large model also monitors the pipe network pressure in real-time and avoids the risk of excessive pipe network pressure by adjusting the flow distribution in each area, and finally optimizes and generates a dispatch plan that better meets the heating requirements under extreme weather conditions.

[0149] Illustrate the adjustment of the heating strategy after equipment failure: A key circulation pump in the heating system suddenly fails, affecting the heating in some areas. After discovering the failure, the dispatcher enters improvement suggestions through a text box: "The circulation pump has failed. Immediately switch to the standby pump, calculate the heating gap in the affected areas, allocate heat from other heat sources to make up for it, and at the same time estimate the repair time and adjust the subsequent heating strategy according to the repair progress."

[0150] The natural language understanding function parses the suggestions into instructions that the large model can recognize:

[0151] 1) Execute the standby pump switching operation;

[0152] 2) Calculate the heating load gap in the affected areas;

[0153] 3) Develop a plan to allocate heat from other heat sources based on the heating capacity of heat sources and the heat network transmission capacity;

[0154] 4) Dynamically adjust the subsequent heating strategy based on the estimated repair time feedback from the maintenance department.

[0155] After receiving the instruction, the large model searches for knowledge in the knowledge base, such as the switching process of standby pumps, the calculation method of heating load, and the heat source allocation rules. It adjusts the inference logic, first quickly sends out the instruction to start the standby pump to ensure the uninterrupted circulation of the heat network. Then, it accurately calculates the heating gap in the affected area, combines the real-time heating capacity of other heat sources and the hydraulic conditions of the heat network, and formulates a plan to allocate heat from adjacent heat sources to ensure the basic stability of heating in the affected area. At the same time, the large model is connected to the information system of the maintenance department to obtain the maintenance progress in real time. According to the estimated maintenance time, it plans the subsequent heating strategy in advance. For example, before the maintenance is completed, it continuously optimizes the heat allocation plan, and after the maintenance is completed, it smoothly resumes the normal heating strategy, thus optimizing and generating a comprehensive heating strategy to cope with equipment failures.

[0156] In this embodiment, the heating dispatching control personnel evaluate multiple answers generated by the large model according to the prompt words, and train and optimize the large model parameters according to the evaluation results, including:

[0157] The heating dispatching control personnel generate multiple different answers by using the trained large model according to the prompt words, construct evaluation indicators according to the language understanding index, semantic rationality index, context coherence index, heating knowledge accuracy index, and heating reasoning ability index, and evaluate and assign values to the multiple different answers;

[0158] Taking different questions, answers, and evaluation assignment results as inputs, the large model parameters are updated and optimized by using the reinforcement learning algorithm.

[0159] In actual applications, taking the evaluation and model optimization of the heating system's winter peak dispatching strategy as an example, during the winter heating peak, the heating dispatching control personnel face the problem of how to balance the heating demands of each region and optimize energy distribution. At this time, the dispatcher inputs the prompt words: "In the current extremely cold weather, the heating load remains at a peak, and the heating effect of some old communities is not good. How to formulate a reasonable heating dispatching strategy?"

[0160] Based on the knowledge and logic learned through training, the large model generates the following multiple different answers:

[0161] Answer 1: First, increase the heating output of all heat sources to ensure the overall heating quantity. Then, conduct pressure detection on the heat network of old communities. If the pressure is insufficient, appropriately increase the pressure of the branch pipe network in old communities. At the same time, arrange staff to inspect the heating equipment in old communities and promptly handle possible failures.

[0162] Answer 2: Prioritize improving the heating power of heat sources near old communities and precisely supply heat to old communities. At the same time, install intelligent regulating valves in the heat network to adjust the flow distribution in real time according to the temperature feedback of each area to ensure balanced heating in each area. In addition, activate the emergency plan and prepare standby heat sources to cope with possible emergencies.

[0163] Answer 3: Comprehensively check the hydraulic balance of the heat network, recalculate the flow requirements of each area, and adjust the opening of the valves of each branch according to the calculation results to ensure uniform heat distribution. For old communities, increase the circulating pump to improve the flow rate of the heat medium and enhance the heating effect. Also, maintain close contact with the meteorological department and adjust the heating strategy in advance according to future weather changes.

[0164] The heating dispatching and control personnel evaluate and assign values to these answers according to the established evaluation indicators:

[0165] Language comprehensibility indicator: The language expressions of the three answers are clear, and all can enable the dispatching personnel to understand their core contents, so relatively high scores are given to all of them.

[0166] Semantic rationality indicator: Answer 1 directly increases the output of all heat sources without considering the energy supply and equipment carrying capacity, so the semantic rationality is slightly lower; The ideas of precise heating and adjustment by intelligent regulating valves in Answer 2 are reasonable, but the timing of starting the standby heat source is expressed rather vaguely; The content of comprehensively checking the hydraulic balance and adjusting the strategy according to the weather in Answer 3 is semantically reasonable. After comprehensive evaluation, Answer 3 gets the highest score.

[0167] Context coherence indicator: The three answers all focus on solving the current heating problem and are contextually coherent. However, when Answer 2 mentions the intelligent regulating valve and the standby heat source, the connection with the previous text is not tight enough, while Answer 1 and Answer 3 have better coherence.

[0168] Heating knowledge accuracy indicator: The method of increasing the pressure of the branch pipe network in old communities in Answer 1 may lead to safety hazards, so there are problems with the accuracy of heating knowledge; The knowledge of heating equipment operation and dispatching strategy in Answer 2 and Answer 3 is accurate, and Answer 2 and Answer 3 get higher scores.

[0169] Heating reasoning ability indicator: Answer 3 reasons based on the characteristics of the heat network and the problems of old communities to formulate a comprehensive strategy, with the strongest reasoning ability; Answer 2 has certain reasoning but is not comprehensive enough; Answer 1 has simple and direct reasoning and lacks consideration of complex situations.

[0170] After evaluating and assigning values to various indicators, different questions (under the current extremely cold weather... how to formulate a reasonable heating dispatch strategy?), answers (the above three answers), and the evaluation and assignment results are used as inputs, and the parameters of the large model are updated and optimized using a reinforcement learning algorithm. In this process, the large model learns that in complex situations such as extremely cold weather and poor heating in some areas, the hydraulic balance analysis and flow optimization distribution of the heat network should be prioritized, intelligent devices should be reasonably utilized, and strategies should be adjusted according to weather changes, rather than simply increasing the heat source output. By continuously accumulating such evaluation and optimization processes, the large model can generate more reasonable and accurate answers when facing heating dispatch-related questions, improving the scientificity and effectiveness of heating dispatch decisions.

[0171] In this embodiment, integrating the large models of each task of the optimized heating intelligent decision-making into each set intelligent agent, using the large model to drive the intelligent agent, and guiding each intelligent agent to perform environment perception, collaborative interaction, and autonomous adaptive decision-making in the simulation environment, including:

[0172] Corresponding task intelligent agents are set according to the diverse tasks of heating intelligent decision-making, including operation perception intelligent agents, parameter prediction intelligent agents, fault diagnosis intelligent agents, scheduling planning intelligent agents, and energy consumption evaluation intelligent agents;

[0173] Set bottom-layer device intelligent agents, including heat source intelligent agents, heat network intelligent agents, heat storage intelligent agents, and heat user intelligent agents, which respectively collect data of heat source devices, heat networks, heat storage devices, and heat users and transmit them to each task intelligent agent, and issue and execute decision instructions transmitted by each task intelligent agent;

[0174] Optimize the input and output interfaces of each task large model to dock it with each task intelligent agent of heating intelligent decision-making, and allocate large model call channels for each task intelligent agent to integrate each task large model into the corresponding task intelligent agent;

[0175] Each task intelligent agent obtains the collected data through its perception unit, extracts the key information of the data, and transmits it to the large model of the decision-making unit. At the same time, a collaborative interaction mechanism is established between the task intelligent agent and the bottom-layer device intelligent agent to perform collaborative interaction of data and optimize task execution and task resource allocation. Finally, through in-depth analysis and inference learning of the large model of the decision-making unit, each task intelligent agent makes autonomous decisions to generate decision instructions to adapt to the dynamically changing operating conditions of the heating system.

[0176] In actual applications, taking the heating system encountering extreme weather as an example, the working processes of each intelligent agent include:

[0177] 1) Data collection by bottom-layer device intelligent agents

[0178] Heat source agent: Quickly collect the operation data of heat source equipment such as thermal power plants and boiler houses, including the combustion efficiency of boilers, steam output, heating power, etc. For example, the heat source agent of a thermal power plant monitors a significant increase in the fuel consumption of the boiler to maintain a high heating output.

[0179] Heat network agent: Real-time monitor the temperature, pressure and flow distribution of the heat network. It is found that the supply water temperature of the heat network in some areas has decreased and the flow rate has also fluctuated, because the increase in heating load has led to uneven heat distribution.

[0180] Heat storage agent: Report the heat storage status and the heat that can be released of the heat storage device, providing a reference for subsequent heating scheduling.

[0181] Heat user agent: Collect feedback information from the user side. For example, some users report that the indoor temperature is relatively low and fails to reach the comfortable temperature standard.

[0182] 2) Task agent environment perception and decision-making

[0183] Operation perception agent: Obtain the data transmitted by the underlying device agents through the perception unit, and extract key information, such as the change trend of the operation status of each device, the real-time working conditions of the heat network, etc. Transmit this information to the large model in the decision-making unit. After analysis, the large model judges that the overall heating system is in a high-load operation state, and there are unstable signs in some links.

[0184] Parameter prediction agent: Combine historical data and current meteorological information, and use the large model in its decision-making unit to predict that the heating load will continue to rise for some time, and it is expected that the outdoor temperature will further drop in the next few hours.

[0185] Fault diagnosis agent: Analyze the collected data, and judge that although there is no serious fault in the current heating system, there are potential risks for some equipment due to high-load operation. For example, the motor current of a certain circulating pump is close to the rated value, and an overload fault may occur.

[0186] Scheduling and planning agent: According to the information provided by the operation perception, parameter prediction and fault diagnosis agents, formulate a scheduling strategy with the assistance of the large model in its decision-making unit. Decide to increase the heating output of the thermal power plant and start some standby boiler houses; adjust the valve opening of the heat network to optimize the flow distribution and give priority to ensuring the heating of residential users and key areas; at the same time, control the heat storage agent to release some heat to make up for the heating gap.

[0187] Energy consumption assessment agent: Real-time evaluate the energy consumption situation and find that the energy consumption has increased significantly with the increase in heating load. Through large model analysis, put forward suggestions to optimize the operation parameters of heat source equipment to reduce energy consumption on the premise of ensuring heating quality, such as adjusting the combustion air ratio of the boiler to improve the combustion efficiency.

[0188] 3) Cooperative Interaction and Instruction Execution

[0189] A cooperative interaction mechanism is established among task agents to share information and coordinate actions. For example, when formulating strategies, the scheduling and planning agent refers to the energy-saving suggestions of the energy consumption assessment agent to avoid excessive energy consumption; the fault diagnosis agent informs the scheduling and planning agent of the potential risks of the equipment so that the safe operation of the equipment can be considered in the scheduling strategy.

[0190] The scheduling and planning agent issues decision instructions to the underlying device agents for execution. After receiving the instruction to increase the heating output, the heat source agent adjusts the operating parameters of the equipment in the thermal power plant and the boiler room; the heat network agent adjusts the valve opening according to the instruction to achieve optimal distribution of the heat network flow; the heat user agent feeds back the heating adjustment situation to the user and guides the user to use the heating facilities reasonably.

[0191] In this embodiment, the simulation operation environment for providing intelligent heating decision-making interaction includes: managing the simulation time to analyze the operating status of the heating system at different times, updating the status of the simulation environment, and providing a flexible time control interface to allow immediate operation of the simulation time during the system operation;

[0192] The user interaction interface for providing intelligent heating decision-making interaction includes: providing a graphical parameter configuration interface, setting simulation parameters through slider, dropdown menu, text box interaction elements, supporting loading parameter configuration files, and visualizing and analyzing the simulation results in the form of text, pictures, and videos;

[0193] The simulation of dynamic parameters of the scenario for providing intelligent heating decision-making interaction includes: using a large model to construct a simulation model of the actual heating system and simulating the system in a dynamic environment.

[0194] In this embodiment, constructing performance indicators for evaluating each agent, and evaluating and optimizing the behavior and decision-making effects of each agent according to the performance indicators and decision simulation results includes:

[0195] Constructing performance indicators for evaluating each task agent in intelligent heating intelligent decision-making, including task completion rate, task completion efficiency, decision adaptability, and cooperation among agents;

[0196] According to the behavior trajectories, decision-making processes, and output decision simulation results of each task agent, calculate the performance indicator values of each task agent, and compare them with the preset standards to evaluate the performance of each task agent;

[0197] Optimize and adjust the behaviors and decision-making strategies of each task agent according to the evaluation results, including: improving the algorithms and model structures within the agent, enhancing the agent's environmental perception ability and adaptability, and optimizing the cooperation mechanism and communication protocol among agents.

[0198] 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 from that 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.

[0199] In addition, each functional module in various embodiments of the present invention may be integrated together to form an independent part, or each module may exist separately, or two or more modules may 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 this 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 to execute all or part of the steps of the methods in various embodiments 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.

[0200] Based on the above inspiration from the 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 heating intelligent decision-making system based on an artificial intelligence large model, characterized in that: It includes: Data layer: used to collect, store and manage various types of data on the source, grid, load and storage of the heating system, and as a large-scale data set for knowledge base establishment and large model training layer; Knowledge base establishment and large model training layer: used to perform data preprocessing, data annotation, knowledge extraction and representation on large-scale data sets of the heating system, and then establish a knowledge base for diversified tasks of intelligent decision-making in heating; It is also used to train a large model based on the knowledge base of diversified tasks of intelligent heating decision-making and the prompt learning strategy of intelligent heating decision-making. The large model captures the knowledge and behavioral characteristics of each task of intelligent heating decision-making, and establishes the intrinsic connection between each prompt word and high-quality answer. It also fine-tunes the large model according to the specific scenario requirements, task objectives and constraints of each heating task to establish a large model for each task, so as to adapt to specific business scenarios, decision-making tasks and various intelligent agents. Human-machine enhanced optimization big model layer: used for the heating dispatching and control personnel to make improvement suggestions to the big model during the dialogue process of each task of heating intelligent decision-making, and the big model optimizes the generated answers based on the suggestions; it is also used for the heating dispatching and control personnel to evaluate the multiple answers generated by the big model based on the prompt words, and train and optimize the big model parameters based on the evaluation results; Intelligent agent decision layer: used to integrate the large models of each task of intelligent heating decision-making after training and optimization into each intelligent agent, use the large model to drive the intelligent agent, and guide each intelligent agent to perform environmental perception, collaborative interaction and autonomous adaptive decision-making in the simulation environment; Simulation layer: used to provide the simulation operation environment, user interaction interface and scenario dynamic parameter simulation for heating intelligent decision-making interaction. After the heating dispatching control personnel raise questions, each intelligent agent generates decision simulation data and supports displaying answers to users in the form of text, pictures and videos. Agent evaluation and optimization layer: used to construct performance indicators for evaluating each agent, and evaluate and optimize the behavior and decision-making effects of each agent based on the performance indicators and decision simulation results.

2. The intelligent heating decision-making system according to claim 1 is characterized in that: The heating system source, grid, load and storage multi-type data includes: historical operation data of the heating system source, grid, load and storage, heating operation and maintenance and scheduling management manual, dynamic change parameters, and physical property data of each device in the system; Among them, the historical operation data include: the operating parameters of the source-side heat source equipment, the heat metering data of each pipeline in the heat network, the indoor temperature and heat consumption data, the operating status and storage and release heat of the heat storage equipment, the historical heat load data, and the on-demand scheduling decision strategy, the historical energy consumption analysis data, and the abnormal diagnosis data; the heating operation and maintenance and scheduling management manual includes the heating inspection content, equipment maintenance strategy, fault handling process, scheduling process, and emergency scheduling response processing rules; the dynamically changing parameters include outdoor environmental parameters and sudden abnormal fault parameters; the physical property data of each device in the system include the fixed parameters, operating characteristics, and service life of each source network, load storage equipment, and pipeline network.

3. The intelligent heating decision-making system according to claim 1 is characterized in that: After data preprocessing, data annotation, knowledge extraction and representation of the large-scale data set of the heating system, a knowledge base of diversified tasks of intelligent decision-making for heating is established, including: Clarify the diverse tasks of intelligent heating decision-making, including operation perception, parameter prediction, fault diagnosis, scheduling planning and energy consumption evaluation, and determine the knowledge scope and knowledge types covered by the knowledge base based on the diverse tasks; the knowledge types include factual knowledge, rule knowledge and process knowledge; According to the diverse tasks, knowledge scope and knowledge type of intelligent heating decision-making, data preprocessing, labeling and classification of large-scale data sets of heating systems are carried out; Knowledge extraction: Directly formulate rules for knowledge extraction based on clear rules and experience; use machine learning algorithms to discover potential knowledge from data; use natural language processing technology to extract key information and knowledge from unstructured text data; Knowledge representation: The extracted knowledge is represented based on production rule representation, framework representation and semantic network representation; The extracted and represented knowledge is stored in the corresponding database table or graph structure according to the format and requirements of the knowledge base management system, and indexes and associations between knowledge are established. A knowledge retrieval interface is provided to support keyword retrieval, semantic retrieval and association retrieval methods, and finally a knowledge base for diversified tasks of intelligent heating decision-making is established.

4. The intelligent heating decision-making system according to claim 1, characterized in that: The large model training is performed based on the diversified task knowledge base of heating intelligent decision-making and the prompt learning strategy of heating intelligent decision-making. The large model captures the knowledge and behavior characteristics of each task of heating intelligent decision-making and establishes the intrinsic connection between each prompt word and high-quality answer, including: In the large model training stage, the diversified knowledge base of heating intelligent decision-making is input into the base large model as part of the training corpus, and the heating professional terms, semantic relations, and logical relations between various factors in the heating system in the knowledge base are learned. The operation mode, behavioral characteristics and intelligent decision-making process of the heating system under different working conditions are analyzed. According to the characteristics of the heating intelligent decision-making task, the thinking process and rules are preset through the thinking chain method, and the decision-making task problem is decomposed into multiple sub-problems. Logically clear steps are provided in the form of prompts, and different types of prompt templates are designed to guide the large model to analyze and process the input information according to the steps and processes; The designed prompt template is combined with the content of the knowledge base and input into the large model, so that the model can learn the ability to extract relevant knowledge from the knowledge base and generate answers based on the prompt information, as well as collect user feedback on the answers and establish the intrinsic connection between the prompt words and the answers.

5. The intelligent heating decision-making system according to claim 1 is characterized in that: The large model is fine-tuned according to the specific scenario requirements, task objectives and constraints of each heating task to establish a large model for each task, including: For the task of sensing the operation of the heating system, the specific scenario requirement is to sense the operating status of each device in the heating system, and the task goal is to fully and accurately obtain the operating data of each device. The constraint condition is that the data collection is compatible with the data interfaces of devices of different types and manufacturers. For the task of predicting heating system parameters, the specific scenario requirement is to predict heat load and heating network parameters. The task goal is to predict and accurately predict the changes in heating system operating parameters in advance. The constraints are to consider the influencing factors of weather changes, user heating behavior, and historical data rules. For the heating system fault diagnosis task, the specific scenario requirement is to diagnose the fault type of the heating system and determine the cause of the fault. The task goal is to accurately identify the fault type and locate the fault cause. The constraint condition is the complexity of the fault scenario. For the scheduling and planning task of the heating system, the specific scenario requirement is to generate scheduling plans of different time scales for each source, grid, load and storage device according to the supply and demand relationship of the heat load. The task goal is to meet the heating demand, minimize the economic cost and reduce energy consumption. The constraints are the operating parameter constraints of each device and the thermal power balance constraints. For the task of evaluating the energy consumption of the heating system, the specific scenario requirements are to evaluate the energy consumption of the heating system, analyze the factors affecting energy consumption, and put forward energy-saving suggestions. The task goal is to reduce the energy consumption of the heating system, and the constraint condition is to consider the energy consumption differences under different operating conditions; The big model is fine-tuned according to the specific scenario requirements, objective functions, and constraints of each task, and a big model for each task is established, including an operation perception big model, a parameter prediction big model, a fault diagnosis big model, a scheduling planning big model, and an energy consumption assessment big model.

6. The intelligent heating decision-making system according to claim 1, characterized in that: During the dialogue process of each task of heating intelligent decision-making, the heating dispatch control personnel put forward improvement suggestions to the big model, and the big model generates answers based on the optimization suggestions, including: During the dialogue process of each task of heating intelligent decision-making, the heating dispatch control personnel can input improvement suggestions through text boxes or voice, and convert the input improvement suggestions into instructions that can be understood by the large model by using the natural language understanding function; The big model adjusts its internal knowledge base and reasoning logic based on the improvement suggestions to optimize the generated answers.

7. The intelligent heating decision-making system according to claim 1, characterized in that: The heating dispatch control personnel evaluate the multiple answers generated by the large model according to the prompt words, and train and optimize the parameters of the large model according to the evaluation results, including: The heating dispatching and control personnel use the trained large model to generate multiple different answers based on the prompt words, and construct evaluation indicators according to the language comprehension index, semantic rationality index, context coherence index, heating knowledge accuracy index, and heating reasoning ability index to evaluate and assign values ​​to multiple different answers; Different questions, answers, and evaluation results are used as input, and reinforcement learning algorithms are used to update and optimize large model parameters.

8. The intelligent heating decision-making system according to claim 1, characterized in that: The aforementioned method integrates the large models of each task of intelligent heating decision-making after training and optimization into each intelligent agent set, uses the large model to drive the intelligent agent, and guides each intelligent agent to perform environmental perception, collaborative interaction, and autonomous adaptive decision-making in a simulation environment, including: According to the diversified tasks of intelligent heating decision-making, corresponding task agents are set up, including operation perception agent, parameter prediction agent, fault diagnosis agent, scheduling planning agent and energy consumption evaluation agent; Set up bottom-level equipment agents, including heat source agents, heat network agents, heat storage agents and heat user agents, respectively collect data of heat source equipment, heat network, heat storage devices and heat users and transmit them to each task agent, and issue and execute decision instructions transmitted by each task agent; Optimize the input and output interfaces of each task model to connect it with each task agent of heating intelligent decision-making, allocate a large model calling channel for each task agent, and integrate each task model into the corresponding task agent; Each task agent uses its perception unit to perceive the environment to obtain collected data, extract key information from the data, and transmit it to the big model of the decision-making unit. At the same time, a collaborative interaction mechanism is established between the task agent and the underlying equipment agent to conduct collaborative data interaction and optimize task execution and task resource allocation. Finally, through the big model of the decision-making unit, in-depth analysis and reasoning learning are performed, and each task agent makes autonomous decisions to generate decision instructions to adapt to the dynamically changing operating conditions of the heating system.

9. The intelligent heating decision-making system according to claim 1, characterized in that: The simulation operation environment for providing heating intelligent decision interaction includes: managing simulation time to analyze the operation status of the heating system at different times, updating the status of the simulation environment, and providing a flexible time control interface to allow real-time operation of simulation time during system operation; The user interaction interface for providing heating intelligent decision-making interaction includes: providing a graphical parameter configuration interface, setting simulation parameters and supporting the loading of parameter configuration files through sliders, drop-down menus, and text box interactive elements, and visually displaying and analyzing simulation results in the form of text, pictures, and videos; The dynamic parameter simulation of the scene when providing heating intelligent decision-making interaction includes: using a large model to build a simulation model of the actual heating system and a dynamic environment to perform system simulation.

10. The intelligent heating decision-making system according to claim 1, characterized in that: The construction evaluates the performance indicators of each intelligent agent, and evaluates and optimizes the behavior and decision-making effect of each intelligent agent based on the performance indicators and decision simulation results, including: Construct performance indicators to evaluate the intelligent decision-making tasks of intelligent heating, including task completion, task completion efficiency, decision-making adaptability, and collaboration between intelligent agents; According to the behavior trajectory, decision-making process and output decision simulation results of each task agent, the performance index value of each task agent is calculated and compared with the preset standards to evaluate the performance of each task agent; Based on the evaluation results, the behavior and decision-making strategies of each task agent are optimized and adjusted, including: improving the algorithm and model structure within the agent, enhancing the agent's environmental perception and adaptability, and optimizing the collaboration mechanism and communication protocol between agents.

Images