Heat supply operation visualization system based on MCP protocol and large model

Abstract

The invention discloses a heat supply operation visualization system based on an MCP protocol and a large model. A data resource layer comprises different heat supply business data visualization rules, a scene adaptation rule, a heat supply operation business function module interface building tool, a data visualization tool and a core heat supply business function module interface; the intelligent processing layer identifies a visual display intention through a visual display intention analysis large model, obtains a plurality of visual display sub-tasks, and then determines called corresponding data interfaces and capability units according to the sub-tasks through a visual display business resource matching large model; related information is obtained through interaction of the MCP host, the client and the server, and visual display of each heat supply service function is completed; the user interaction layer is used for obtaining a user visual display intention through a multi-round dialogue mechanism and calling an intention analysis large model; and the server is also used for displaying the visualization effect of each heat supply business function to the user and carrying out iterative optimization according to feedback information.

Description

Technical Field

[0001] This invention belongs to the field of heating system operation technology, specifically relating to a heating operation visualization system based on the MCP protocol and a large model. Background Technology

[0002] Visualized heating operation is the core display platform of a smart heating system. It displays complex operational data indicators in multiple dimensions, such as bar charts, line charts, and heat maps. It enables full-process monitoring and data presentation of heating facilities such as heat sources, heating networks, and heating stations, captures various abnormal data, and marks alarm information. At the same time, it constructs a GIS map of the heating network and a 3D model of the equipment to achieve a detailed reconstruction of the heating network and equipment, clearly showing the spatial distribution of the heating network and equipment, assisting in the analysis of regional heat load differences and the geographical distribution characteristics of heat sources, and supporting dynamic interactive updates.

[0003] However, the current visualization of heating operation involves more and more business function modules, some of which require customized high-demand displays. These modules need to call multiple databases, calculation models, GIS maps, and process simulation models in the heating system backend, as well as different visualization tools. However, the communication call protocols, call methods, and data formats of data interfaces and business module interfaces will differ. This means that each new module requires custom development of an adaptation interface, resulting in high development costs, long integration cycles, and poor flexibility. It is difficult to meet the customized visualization display needs and also impossible to manage the various interfaces and modules in a unified manner.

[0004] Based on the aforementioned technical issues, a new heating operation visualization system based on the MCP protocol and a large model needs to be designed. Summary of the Invention

[0005] The technical problem to be solved by this invention is to overcome the shortcomings of the prior art and provide a heating operation visualization system based on the MCP protocol and a large model. By visually displaying the intent analysis large model, visually displaying the business resource matching large model, and deploying the MCP host, MCP client and MCP server based on the MCP protocol, this invention solves the heterogeneity problem of inconsistent interface formats and different tool calling methods in traditional heating systems. It can also identify the visualization intent and automatically call the corresponding interface resources and tool resources to realize the visualization display of heating operation.

[0006] To solve the above-mentioned technical problems, the technical solution of the present invention is as follows: This invention provides a heating operation visualization system based on the MCP protocol and a large model, comprising: The data resource layer includes data visualization rules for different heating services, scenario adaptation rules, interface building tools for various business function modules of heating operation, data visualization tools, and interfaces for core heating business function modules. The intelligent processing layer includes a trained large-scale model for analyzing visualization intent and a large-scale model for matching visualization business resources, as well as an MCP host, MCP client, and MCP server deployed based on the MCP protocol. The MCP server is used to encapsulate the interfaces of the core heating business function modules in the data resource layer into standardized data interfaces and to encapsulate various rules and tools into callable capability units through the MCP protocol. The MCP host is an application that integrates the ability to match large models of visualization display business resources. It is used to manage the user's visualization display intent and the scheduling of large models of visualization display business resources, transmit the instructions generated by the large models of visualization display business resources to the corresponding MCP client, and receive the information fed back by the MCP client and transmit it to the large models of visualization display business resources. The MCP client is used to convert the instructions generated by the visualization display business resource matching large model into structured requests and send them to the corresponding MCP server to call various rules, tools and business function module interfaces; The visualization intent analysis model is used to identify and analyze the user's visualization intent in the user interaction layer. After obtaining multiple heating visualization sub-tasks, the visualization business resource matching model autonomously determines the corresponding data interface and capability unit to be called according to each sub-task. It also obtains relevant information through multiple rounds of interaction with the MCP host, MCP client and MCP server and automatically completes the visualization display of each heating business function. The user interaction layer is used to obtain the user's visualization intent through a preset multi-turn dialogue mechanism and by calling the visualization intent analysis big model; it is also used to display the visualization effects of various heating business functions to users and to iteratively optimize based on user feedback.

[0007] Furthermore, the visualization rules for different heating service data include: displaying numerical data of heating services through digital dashboards and color-coded annotation rules to highlight key indicators; displaying spatial data of heating services through GIS maps and 3D models, combined with heat maps and flow maps to show the operational status; displaying trend data of heating services through line charts, marking the highest and lowest values; and summarizing statistical data of heating services through bar charts or card-based formats. The scene adaptation rules include: the heating scene should be based on a base color and a technological style, while the 3D model should conform to the physical form of the actual heat source, pipe network and heating station, and use ambient light and particle effects to enhance the technological feel. The interface building tools for each business function module of the heating operation include EasyV and DataV, which allow for direct drag-and-drop of components. The data visualization tools include: chart tools ECharts and D3.js; GIS map visualization configuration tools that can import geographic base map data of heating cities and then associate the location, pipe diameter, and direction of the heating network, as well as the coordinates and attribute information of heat sources and heating stations, onto the map; and 3D effect tools Three.js and CSS3. The interfaces for the core heating business function modules include: heat source monitoring interface, heating network monitoring interface, heating station monitoring interface, weather forecast interface, load prediction model interface, alarm monitoring interface, video monitoring interface, and scene modeling interface.

[0008] Furthermore, the visualization intent analysis model is used to identify and analyze the user's visualization intent at the user interaction layer, obtaining multiple heating visualization sub-tasks: Through multiple rounds of information interaction between the intelligent processing layer and the user interaction layer, the user's visualization intent at the user interaction layer is obtained. In each round of information interaction, the visualization intent analysis model is used to perform intent recognition and analysis, obtaining multiple heating visualization sub-tasks round by round, specifically including: In the first round of interaction: an interface with a resolution drop-down box, a terminal type selection box, a physical size input box, a business function type selection button, and custom display requirements is designed. After initiating a visual display request to the user regarding the size, type, and core display business functions of the on-site display terminal, the user inputs relevant basic information. The core entities of the relevant basic information are identified through the large model of visual display intent analysis and transformed into structured basic parameters. In the second round of interaction: Based on the relevant basic information input by the user, various general visual layout templates for heating operation are recommended to the user, who can choose a template or customize their layout requirements; The core elements in the templates or customized layout requirements are identified through the large model of visual display intent recognition, including the layout position and display rules of each business function module, and transformed into structured layout parameters; In the third round of interaction: Based on the layout parameters of each business function module to be displayed, the system initiates requests to users regarding the number of devices (heat sources, heating networks, heating stations), the number of interfaces for heating data visualization points, and modeling parameters (including 3D modeling requirements, number of models, modeling detail, and rendering effects). Users provide feedback through multimodal methods such as text, images, and videos. After recognizing the multimodal feature parameters from the large model based on the visualization intent and performing feature fusion, the system generates device docking parameters and modeling requirements, which are then transformed into structured docking parameters and modeling requirements. In the fourth round of interaction: the visualization intent recognition model integrates structured basic parameters, layout parameters, docking parameters and modeling requirement parameters, and provides feedback to the user through the user interaction layer for information confirmation and completion. Then, the visualization intent recognition model integrates the layout parameters, docking parameters, modeling requirement parameters, display style and display effect of the scattered heating business function modules through the preset business field mapping table, forming the heating visualization display sub-task of each business function module.

[0009] Furthermore, the training process of the visualization intent analysis large model includes: The system acquires historical dialogue and interaction data regarding user requests for visualization of heating operation, and annotates the intent tags and entity parameters in the data; it acquires user-drawn visualization layout diagrams, 3D model reference diagrams, and GIS map reference diagrams, and annotates the location and model features of each business equipment component in the diagrams; it acquires reference videos of the visualization display effect of heating operation, extracts keyframes and annotates rendering effects, dynamic interactions, display style, and chart type features; and it acquires a heating parameter rule base to verify whether the display terminal meets the display requirements. The acquired information is integrated to build a visualization corpus. After pre-training the selected base model, and fine-tuning it using typical visualization cases of various heating business functional modules, a large visualization intent analysis model is formed.

[0010] Furthermore, the MCP server is used to encapsulate the interfaces of the core heating business function modules in the data resource layer into standardized data interfaces and to encapsulate various rules and tools into callable capability units through the MCP protocol, including: The system analyzes the input parameters, output formats, and access permissions of the core business modules interfaces in the data resource layer, including heat source monitoring, heat network monitoring, heating station monitoring, weather forecasting, load prediction models, alarm monitoring, video surveillance, and scene modeling. It also sets up interface adapters for different types of business function module interfaces, customizes unified and standardized interface request and response formats based on the MCP protocol, and builds an MCP server gateway to uniformly manage the standardized interfaces of various heating business function modules. The data resource layer, including data visualization rules for different heating services, scenario adaptation rules, interface building tools for various business function modules of heating operation, and data visualization tools, is broken down into independent and reusable sub-capability units. The functions, input and output parameters of each sub-capability unit are clearly defined. Based on the MCP protocol, a unified protocol format is defined for each sub-capability unit and a protocol adapter is set. The mapping between MCP server requests and sub-capability unit input parameters, the format of sub-capability unit output and MCP server response are configured, and the adapted sub-capability units are registered to the MCP server's registration center and the calling process and calling strategy of the sub-capability units are set.

[0011] Furthermore, the establishment of the MCP server gateway enables unified management of standardized interfaces for various heating service function modules, including: Register the standardized interfaces of each heating business function module to the MCP server gateway to generate a unique MCP interface address for each business function module. Set access permissions and requests per second limit for each business function module's interface; Monitor and manage the call volume, response time, and error rate of the interfaces of each business function module.

[0012] Furthermore, registering each adapted sub-capability unit to the registration center of the MCP server includes: registering the ID, type, function description, input parameters, output parameters, calling permissions, and request limit per second of each adapted sub-capability unit to the MCP server gateway. The process of setting up the sub-capability unit invocation flow includes: the caller sending an MCP protocol request to the MCP server gateway; the MCP server gateway routing to the corresponding sub-capability unit based on the sub-capability unit's ID; the protocol adapter parsing the request parameters and calling the corresponding sub-capability unit to execute the logic; and the protocol adapter standardizing the execution result format and returning it to the caller via the MCP protocol. Configure the calling strategy for sub-capability units, including: based on the request limit per second configured in the MCP server gateway, reject sub-capability unit requests that exceed the limit threshold; reduce redundant calculations by caching the results of frequently called sub-capability units; and automatically call sub-capability units in the order of dependency if there are coupled dependency scheduling relationships.

[0013] Furthermore, the number of MCP hosts is two, one as the primary MCP host and the other as a backup MCP host; the number of MCP clients is set according to the number of each heating business function module, and the number of MCP servers is set according to the number of rules, tools, and core heating business function module interfaces in the data resource layer; one MCP client can bind to multiple MCP servers, and there is a limit to the number of bindings.

[0014] Furthermore, the training process of the visualization display business resource matching large model includes: In the first stage, a pre-trained general large model is selected as the large model for matching business resources for visualization. The general large model is trained by incremental pre-training using general visualization domain data, so that the large model for matching business resources for visualization learns the mapping logic between running data and charts, device data and GIS maps, and 3D models. In the second phase, a heating operation visualization knowledge base will be formed, which includes the functions of each heating business module, the association rules of the heating business knowledge graph, the resource list of interfaces and capability units in the MCP server, the heating operation visualization requirement document, and historical heating operation visualization task cases. The visualization display business resource matching model trained in the first phase will be fine-tuned so that the visualization display business resource matching model can learn the domain knowledge of heating operation visualization. In the third stage, after obtaining various intentions for the visualization of heating operation and labeling the corresponding sub-tasks of the heating visualization display of each business function module, the corresponding MCP client and server resources are labeled for each sub-task. The visualization display business resource matching big model of the second stage is subjected to sub-task parsing and MCP resource matching training, so that the visualization display business resource matching big model outputs a structured sub-task description and a corresponding MCP resource call list based on the visualization intention of each sub-task. In the fourth stage, the communication links between the large-scale visualization display business resource matching model, the MCP host, the MCP client, and the MCP server are simulated. A multi-turn dialogue thought chain dataset is constructed, and the large-scale visualization display business resource matching model from the third stage is guided to master the interaction logic with the MCP host, MCP client, and MCP server through multi-turn dialogue fine-tuning using the multi-turn dialogue dataset. A simulated MCP environment is also built to allow the large-scale visualization display business resource matching model to be simulated in the simulation environment. The simulation interaction results are used to train and optimize the large-scale visualization display business resource matching model, ultimately forming an optimized large-scale visualization display business resource matching model.

[0015] Furthermore, the visualization display business resource matching model autonomously determines the corresponding data interfaces and capability units to be invoked based on each sub-task, and automatically completes the visualization display of various heating business functions through multiple rounds of interaction between the MCP host, MCP client, and MCP server, including: The visualization-based business resource matching model analyzes the business type of each subtask and extracts the core elements of the subtask, outputting a structured description of the subtask. The visualization-based business resource matching model matches the corresponding MCP resource call list for each subtask based on the structured subtask description and the preset MCP client and MCP server resources. Then, it generates the communication link of MCP host-MCP client-MCP server for each subtask based on the subtask description and the MCP resource call list. The visualization display business resource matching big model encapsulates the MCP resource call list of each subtask into instructions and sends them to the MCP host. The MCP host routes the instructions to the corresponding MCP client according to the business type of the subtask. The MCP client converts the instructions into structured requests and calls the MCP server interface and capability unit corresponding to the resource list. Through multiple rounds of interaction between the visualization display business resource matching big model, the MCP host, the MCP client and the MCP server, the business data, matching rules and tools corresponding to each subtask are obtained. Through data integration, business module layout, display effect rendering, anomaly annotation and model building, the visualization display of each heating business function is completed.

[0016] The beneficial effects of this invention are: (1) This invention encapsulates the scattered heating business interfaces, visualization tools, and rules into standardized interfaces and capability units through the MCP protocol, which solves the heterogeneity problem of inconsistent interface formats and different tool calling methods in traditional heating systems, reduces the integration cost across modules and systems, and allows for quick access when adding new business functions in the future by simply encapsulating them according to the MCP standard, thereby improving the scalability and maintainability of the system. (2) Through the layered architecture of MCP host, client and server, this invention realizes the efficient flow of visualization display business resource matching large model instructions, structured requests and resource calls, avoiding the cumbersome process of manual resource docking and manual configuration calls in traditional systems, and allowing the instructions generated by the visualization display business resource matching large model to be directly converted into actual tools and interface calls, thereby improving the automation level and response speed of the technology link; (3) The present invention can automatically identify the user’s visualization needs through the visualization display intent analysis big model and break them down into multiple sub-tasks; then, through the business resource matching big model, it can autonomously select the corresponding interface, tools and rules, and complete the resource scheduling of complex visualization tasks without manual intervention, thereby improving the accuracy and efficiency of business response and heating operation visualization.

[0017] Other features and advantages will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention are realized and obtained through the structures particularly pointed out in the description and the drawings.

[0018] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description

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

[0020] Figure 1 This is a schematic diagram of the structure of a heating operation visualization system based on the MCP protocol and a large model according to the present invention; Figure 2 This invention provides a flowchart of the user visualization intent multi-round interaction and recognition analysis process; Figure 3 This invention provides a visual representation of the training flowchart for a large-scale business resource matching model. Detailed Implementation

[0021] 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 accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0022] like Figure 1 As shown, this embodiment provides a heating operation visualization system based on the MCP protocol and a large model, which includes: The data resource layer includes data visualization rules for different heating services, scenario adaptation rules, interface building tools for various business function modules of heating operation, data visualization tools, and interfaces for core heating business function modules. The intelligent processing layer includes a trained large-scale model for analyzing visualization intent and a large-scale model for matching visualization business resources, as well as an MCP host, MCP client, and MCP server deployed based on the MCP protocol. The MCP server is used to encapsulate the interfaces of the core heating business function modules in the data resource layer into standardized data interfaces and to encapsulate various rules and tools into callable capability units through the MCP protocol. The MCP host is an application that integrates the ability to match large models of visualization display business resources. It is used to manage the user's visualization display intent and the scheduling of large models of visualization display business resources, transmit the instructions generated by the large models of visualization display business resources to the corresponding MCP client, and receive the information fed back by the MCP client and transmit it to the large models of visualization display business resources. The MCP client is used to convert the instructions generated by the visualization display business resource matching large model into structured requests and send them to the corresponding MCP server to call various rules, tools and business function module interfaces; The visualization intent analysis model is used to identify and analyze the user's visualization intent in the user interaction layer. After obtaining multiple heating visualization sub-tasks, the visualization business resource matching model autonomously determines the corresponding data interface and capability unit to be called according to each sub-task. It also obtains relevant information through multiple rounds of interaction with the MCP host, MCP client and MCP server and automatically completes the visualization display of each heating business function. The user interaction layer is used to obtain the user's visualization intent through a preset multi-turn dialogue mechanism and by calling the visualization intent analysis big model; it is also used to display the visualization effects of various heating business functions to users and to iteratively optimize based on user feedback.

[0023] In this embodiment, the visualization rules for different heating service data include: displaying numerical data of heating services through digital dashboards and color-coded annotation rules to highlight key indicators; displaying spatial data of heating services through GIS maps and 3D models, combined with heat maps and flow maps to show the operational status; displaying trend data of heating services through line charts, marking the highest and lowest values; and summarizing statistical data of heating services through bar charts or card-based formats. The scene adaptation rules include: the heating scene should be based on a base color and a technological style, while the 3D model should conform to the physical form of the actual heat source, pipe network and heating station, and use ambient light and particle effects to enhance the technological feel. The interface building tools for each business function module of the heating operation include EasyV and DataV, which allow for direct drag-and-drop of components. The data visualization tools include: chart tools ECharts and D3.js; GIS map visualization configuration tools that can import geographic base map data of heating cities and then associate the location, pipe diameter, and direction of the heating network, as well as the coordinates and attribute information of heat sources and heating stations, onto the map; and 3D effect tools Three.js and CSS3. The interfaces for the core heating business function modules include: heat source monitoring interface, heating network monitoring interface, heating station monitoring interface, weather forecast interface, load prediction model interface, alarm monitoring interface, video monitoring interface, and scene modeling interface.

[0024] It should be noted that the scene modeling interface includes interfaces related to GIS maps, 3D pipeline models, 3D equipment models, and 3D building models. EasyV is the EasyV visualization platform, a drag-and-drop visualization interface building tool. DataV is the DataV data visualization dashboard tool, a drag-and-drop dashboard building tool. ECharts is an enterprise-level chart library, a front-end chart visualization tool. D3.js is a data-driven documentation, a custom visualization development library. Three.js is a front-end 3D visualization development library. CSS3 (Cascading Style Sheets, version 3) is a technical standard for web page style design, enabling basic visualization effects.

[0025] like Figure 2 As shown, in this embodiment, the visualization intent analysis model is used to identify and analyze the user's visualization intent in the user interaction layer, obtaining multiple heating visualization sub-tasks: Through multiple rounds of information interaction between the intelligent processing layer and the user interaction layer, the user's visualization intent in the user interaction layer is obtained, and in each round of information interaction, the visualization intent analysis model is used to perform intent recognition and analysis, obtaining multiple heating visualization sub-tasks round by round, specifically including: In the first round of interaction: an interface with a resolution drop-down box, a terminal type selection box, a physical size input box, a business function type selection button, and custom display requirements is designed. After initiating a visual display request to the user regarding the size, type, and core display business functions of the on-site display terminal, the user inputs relevant basic information. The core entities of the relevant basic information are identified through the large model of visual display intent analysis and transformed into structured basic parameters. In the second round of interaction: Based on the relevant basic information input by the user, various general visual layout templates for heating operation are recommended to the user, who can choose a template or customize their layout requirements; The core elements in the templates or customized layout requirements are identified through the large model of visual display intent recognition, including the layout position and display rules of each business function module, and transformed into structured layout parameters; In the third round of interaction: Based on the layout parameters of each business function module to be displayed, the system initiates requests to users regarding the number of devices (heat sources, heating networks, heating stations), the number of interfaces for heating data visualization points, and modeling parameters (including 3D modeling requirements, number of models, modeling detail, and rendering effects). Users provide feedback through multimodal methods such as text, images, and videos. After recognizing the multimodal feature parameters from the large model based on the visualization intent and performing feature fusion, the system generates device docking parameters and modeling requirements, which are then transformed into structured docking parameters and modeling requirements. In the fourth round of interaction: the visualization intent recognition model integrates structured basic parameters, layout parameters, docking parameters and modeling requirement parameters, and provides feedback to the user through the user interaction layer for information confirmation and completion. Then, the visualization intent recognition model integrates the layout parameters, docking parameters, modeling requirement parameters, display style and display effect of the scattered heating business function modules through the preset business field mapping table, forming the heating visualization display sub-task of each business function module.

[0026] In this embodiment, the training process of the visualization intent analysis large model includes: The system acquires historical dialogue and interaction data regarding user requests for visualization of heating operation, and annotates the intent tags and entity parameters in the data; it acquires user-drawn visualization layout diagrams, 3D model reference diagrams, and GIS map reference diagrams, and annotates the location and model features of each business equipment component in the diagrams; it acquires reference videos of the visualization display effect of heating operation, extracts keyframes and annotates rendering effects, dynamic interactions, display style, and chart type features; and it acquires a heating parameter rule base to verify whether the display terminal meets the display requirements. The acquired information is integrated to build a visualization corpus. After pre-training the selected base model, and fine-tuning it using typical visualization cases of various heating business functional modules, a large visualization intent analysis model is formed.

[0027] In this embodiment, the MCP server is used to encapsulate the interfaces of the core heating business function modules in the data resource layer into standardized data interfaces and to encapsulate various rules and tools into callable capability units through the MCP protocol, including: The system analyzes the input parameters, output formats, and access permissions of the core business modules interfaces in the data resource layer, including heat source monitoring, heat network monitoring, heating station monitoring, weather forecasting, load prediction models, alarm monitoring, video surveillance, and scene modeling. It also sets up interface adapters for different types of business function module interfaces, customizes unified and standardized interface request and response formats based on the MCP protocol, and builds an MCP server gateway to uniformly manage the standardized interfaces of various heating business function modules. The data resource layer, including data visualization rules for different heating services, scenario adaptation rules, interface building tools for various business function modules of heating operation, and data visualization tools, is broken down into independent and reusable sub-capability units. The functions, input and output parameters of each sub-capability unit are clearly defined. Based on the MCP protocol, a unified protocol format is defined for each sub-capability unit and a protocol adapter is set. The mapping between MCP server requests and sub-capability unit input parameters, the format of sub-capability unit output and MCP server response are configured, and the adapted sub-capability units are registered to the MCP server's registration center and the calling process and calling strategy of the sub-capability units are set.

[0028] In practical applications, the interface adapter setting can solve the interface incompatibility problem and realize the conversion between heterogeneous interfaces and standard interfaces. Among them, input adaptation: converts external MCP standard requests into the private input parameter format of the corresponding business module, and performs parameter verification at the same time; output adaptation: converts the private output of the business module into the MCP standard response format.

[0029] In practical applications, visualization rules, scene adaptation rules, interface building tools, and data visualization tools are broken down into numerical data visualization units, spatial data visualization units, trend data visualization units, statistical data visualization units, scene adaptation rendering units, and interface building and arrangement units.

[0030] The input parameters for the numerical data visualization unit are numerical data and color scale rules, and the output parameters are digital dashboard rendering configuration and visualization component ID. The input parameters for the spatial data visualization unit are spatial data, display type, and GIS base map, and the output parameters are GIS / 3D model rendering configuration and visualization preview link. The input parameters for the trend data visualization unit are trend data and annotation rules, and the output parameters are line chart rendering configuration and data update frequency. The input parameters for the statistical data visualization unit are statistical data and display type, and the output parameters are statistical chart rendering configuration. The input parameters for the scene adaptation rendering unit are basic visualization configuration, scene type, and 3D model ID, and the output parameters are scene-adapted rendering configuration and technological effect parameters. The input parameters for the interface building and arrangement unit are module ID, component list, and layout parameters, and the output parameters are interface rendering and component interaction logic configuration.

[0031] In this embodiment, the establishment of the MCP server gateway for unified management of standardized interfaces for various heating service function modules includes: Register the standardized interfaces of each heating business function module to the MCP server gateway to generate a unique MCP interface address for each business function module. Set access permissions and requests per second limit for each business function module's interface; Monitor and manage the call volume, response time, and error rate of the interfaces of each business function module.

[0032] It's important to note that all heating service interfaces are managed centrally through a gateway. Users do not need to individually interface with different business function modules; they can access all functions simply through the gateway. Furthermore, access permission configurations prevent unauthorized parties from calling core business interfaces, ensuring data and business security. The limit on requests per second for each interface prevents overload due to sudden surges in requests. By monitoring the interface's call volume, response time, and error rate, performance bottlenecks or faults can be quickly identified.

[0033] In this embodiment, registering each adapted sub-capability unit to the registration center of the MCP server includes: registering the ID, type, function description, input parameters, output parameters, calling permissions, and request per second limit of each adapted sub-capability unit to the MCP server gateway. The process of setting up the sub-capability unit invocation flow includes: the caller sending an MCP protocol request to the MCP server gateway; the MCP server gateway routing to the corresponding sub-capability unit based on the sub-capability unit's ID; the protocol adapter parsing the request parameters and calling the corresponding sub-capability unit to execute the logic; and the protocol adapter standardizing the execution result format and returning it to the caller via the MCP protocol. Configure the calling strategy for sub-capability units, including: based on the request limit per second configured in the MCP server gateway, reject sub-capability unit requests that exceed the limit threshold; reduce redundant calculations by caching the results of frequently called sub-capability units; and automatically call sub-capability units in the order of dependency if there are coupled dependency scheduling relationships.

[0034] It's important to note that all sub-capability unit attributes and rules are centrally stored in the gateway registry center. This allows for unified viewing and modification of unit configurations, avoiding the tedious maintenance of each unit individually. Furthermore, the call permission configuration during registration restricts unauthorized access to core units, preventing sensitive data leakage. Protocol adapters resolve interface differences between units; large models only need to be called uniformly according to the MCP protocol, eliminating the need to adapt to different unit parameter formats. The output format is standardized and can be directly used for visualization. Rate limiting strategies prevent individual sub-capability units from being overwhelmed by high-frequency requests, preventing the entire visualization business chain from paralyzing due to a single sub-capability unit failure. Caching strategies directly return cached results for high-frequency requests, reducing redundant calculations and improving the real-time performance of visualizations. Dependency scheduling strategies automatically schedule coupled units according to dependency order, eliminating the need for callers to manually orchestrate the call order and preventing execution failures due to incorrect dependency order.

[0035] In this embodiment, there are two MCP hosts, one as the primary MCP host and the other as a backup MCP host. The number of MCP clients is set according to the number of each heating business function module. The number of MCP servers is set according to the number of rules, tools, and core heating business function module interfaces in the data resource layer. One MCP client can bind to multiple MCP servers, and there is a limit to the number of bindings.

[0036] like Figure 3 As shown in this embodiment, the training process of the visualization display business resource matching large model includes: In the first stage, a pre-trained general large model is selected as the large model for matching business resources for visualization. The general large model is trained by incremental pre-training using general visualization domain data, so that the large model for matching business resources for visualization learns the mapping logic between running data and charts, device data and GIS maps, and 3D models. In the second phase, a heating operation visualization knowledge base will be formed, which includes the functions of each heating business module, the association rules of the heating business knowledge graph, the resource list of interfaces and capability units in the MCP server, the heating operation visualization requirement document, and historical heating operation visualization task cases. The visualization display business resource matching model trained in the first phase will be fine-tuned so that the visualization display business resource matching model can learn the domain knowledge of heating operation visualization. In the third stage, after obtaining various intentions for the visualization of heating operation and labeling the corresponding sub-tasks of the heating visualization display of each business function module, the corresponding MCP client and server resources are labeled for each sub-task. The visualization display business resource matching big model of the second stage is subjected to sub-task parsing and MCP resource matching training, so that the visualization display business resource matching big model outputs a structured sub-task description and a corresponding MCP resource call list based on the visualization intention of each sub-task. In the fourth stage, the communication links between the large-scale visualization display business resource matching model, the MCP host, the MCP client, and the MCP server are simulated. A multi-turn dialogue thought chain dataset is constructed, and the large-scale visualization display business resource matching model from the third stage is guided to master the interaction logic with the MCP host, MCP client, and MCP server through multi-turn dialogue fine-tuning using the multi-turn dialogue dataset. A simulated MCP environment is also built to allow the large-scale visualization display business resource matching model to be simulated in the simulation environment. The simulation interaction results are used to train and optimize the large-scale visualization display business resource matching model, ultimately forming an optimized large-scale visualization display business resource matching model.

[0037] It should be noted that the selection of a general large model in the first stage may include Llama and ChatGLM, enabling the large model to master the basic mapping logic and possess the underlying capabilities for cross-scenario visualization; in the second stage, the large model will master the system of heating business and MCP architecture, completing the domain adaptation from general visualization to heating visualization; in the third stage, the large model will possess the ability to parse subtasks and match MCP resources; in the fourth stage, the large model will master the communication and feedback logic with the MCP architecture; and the final trained visualization display business resource matching large model can autonomously complete the entire process of heating visualization requirements, subtask analysis, MCP resource invocation, multi-round interaction, and visualization output.

[0038] In this embodiment, the visualization display business resource matching big model autonomously determines the corresponding data interface and capability unit to be called based on each sub-task, and automatically completes the visualization display of each heating business function through multiple rounds of interaction between the MCP host, MCP client, and MCP server to obtain relevant information. The visualization-based business resource matching model analyzes the business type of each subtask and extracts the core elements of the subtask, outputting a structured description of the subtask. The visualization-based business resource matching model matches the corresponding MCP resource call list for each subtask based on the structured subtask description and the preset MCP client and MCP server resources. Then, it generates the communication link of MCP host-MCP client-MCP server for each subtask based on the subtask description and the MCP resource call list. The visualization display business resource matching big model encapsulates the MCP resource call list of each subtask into instructions and sends them to the MCP host. The MCP host routes the instructions to the corresponding MCP client according to the business type of the subtask. The MCP client converts the instructions into structured requests and calls the MCP server interface and capability unit corresponding to the resource list. Through multiple rounds of interaction between the visualization display business resource matching big model, the MCP host, the MCP client and the MCP server, the business data, matching rules and tools corresponding to each subtask are obtained. Through data integration, business module layout, display effect rendering, anomaly annotation and model building, the visualization display of each heating business function is completed.

[0039] It should be noted that during the multi-round interaction between the visualization display business resource matching big model, MCP host, MCP client and MCP server, if an MCP server reports an anomaly, the MCP client will send the anomaly information back to the visualization display business resource matching big model through the MCP host. The big model will then adjust the resource call and repeat the interaction process until the complete data of all sub-tasks is obtained.

[0040] The visualization showcases the multi-round interactions between the large model of business resource matching, the MCP host, the MCP client, and the MCP server to obtain core resources, including: The MCP host uses a visual model to match business resources. Based on the subtask type, it determines the interfaces and capability units that each subtask needs to call. Then, the MCP host creates a corresponding MCP client instance, and each client connects to an MCP server that provides the corresponding interfaces and capability units. The MCP client transforms the requests from the MCP host into protocol information that the MCP server can recognize, and initiates calls to the MCP server for interfaces and capability units. After the MCP server completes data extraction and tool and rule preparation, it transmits the data back to the MCP host via the MCP client. The MCP host then aggregates the information returned by multiple MCP clients and synchronizes it to the visualization display business resource matching big model. The visualization display business resource matching big model integrates scattered business data and uses corresponding tools and rules to visualize data, GIS maps, 3D models, etc.

[0041] Visualization can include: business module layout: planning display areas for each business module on the visualization interface; display effect rendering and model building: presenting data using different visualization rules and forms, such as using line charts to show the trend of heat load changes over time, using GIS maps to present different heating areas and related operational data, using equipment icons combined with numerical labels to show the operating status of individual heating stations, using 3D models to build detailed models of buildings in key areas, supporting the creation of 3D models of pipelines along the pipeline, including valves, steam traps, etc., and supporting the access of real-time data to display the actual operating conditions of on-site equipment for objects such as on-site heating stations, energy stations, or indoor equipment; anomaly labeling: anomalies such as equipment temperature exceeding the threshold and pipeline pressure fluctuations being too large are labeled on the interface using flashing red icons, alarm pop-ups, etc.

[0042] In the several embodiments provided in this 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 illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of the present invention. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code, which contains one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order than those marked in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram and / or flowchart, and combinations of blocks in block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions.

[0043] Furthermore, the functional modules in the various embodiments of this invention can be integrated together to form an independent part, or each module can exist independently, or two or more modules can be integrated to form an independent part. If the function is implemented as 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 this invention, or the part that contributes to the prior art, or a part of the 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 to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory, random access memory, magnetic disks, or optical disks.

[0044] Based on the above-described preferred embodiments of the present invention, and through the foregoing description, those skilled in the art can make various changes and modifications without departing from the inventive concept. The technical scope of this invention is not limited to the contents of the specification, but must be determined according to the scope of the claims.

Claims

1. A heating operation visualization system based on the MCP protocol and a large model, characterized in that, It includes: The data resource layer includes data visualization rules for different heating services, scenario adaptation rules, interface building tools for various business function modules of heating operation, data visualization tools, and interfaces for core heating business function modules. The intelligent processing layer includes a trained large-scale model for analyzing visualization intent and a large-scale model for matching visualization business resources, as well as an MCP host, MCP client, and MCP server deployed based on the MCP protocol. The MCP server is used to encapsulate the interfaces of the core heating business function modules in the data resource layer into standardized data interfaces and to encapsulate various rules and tools into callable capability units through the MCP protocol. The MCP host is an application that integrates the ability to match large models of visualization display business resources. It is used to manage the user's visualization display intent and the scheduling of large models of visualization display business resources, transmit the instructions generated by the large models of visualization display business resources to the corresponding MCP client, and receive the information fed back by the MCP client and transmit it to the large models of visualization display business resources. The MCP client is used to convert the instructions generated by the visualization display business resource matching large model into structured requests and send them to the corresponding MCP server to call various rules, tools and business function module interfaces; The visualization intent analysis model is used to identify and analyze the user's visualization intent in the user interaction layer. After obtaining multiple heating visualization sub-tasks, the visualization business resource matching model autonomously determines the corresponding data interface and capability unit to be called according to each sub-task. It also obtains relevant information through multiple rounds of interaction with the MCP host, MCP client and MCP server and automatically completes the visualization display of each heating business function. The user interaction layer is used to obtain the user's visualization intent through a preset multi-turn dialogue mechanism and by calling the visualization intent analysis big model; it is also used to display the visualization effects of various heating business functions to users and to iteratively optimize based on user feedback.

2. The heating operation visualization system according to claim 1, characterized in that, The visualization rules for different heating service data include: displaying numerical data of heating services through digital dashboards and color-coded labels to highlight key indicators; displaying spatial data of heating services through GIS maps and 3D models, combined with heat maps and flow maps to show the operational status; displaying trend data of heating services through line charts, marking the highest and lowest values; and summarizing statistical data of heating services through bar charts or card-based formats. The scene adaptation rules include: the heating scene should be based on a base color and a technological style, while the 3D model should conform to the physical form of the actual heat source, pipe network and heating station, and use ambient light and particle effects to enhance the technological feel. The interface building tools for each business function module of the heating operation include EasyV and DataV, which allow for direct drag-and-drop of components. The data visualization tools include: chart tools ECharts and D3.js; GIS map visualization configuration tools that can import geographic base map data of heating cities and then associate the location, pipe diameter, and direction of the heating network, as well as the coordinates and attribute information of heat sources and heating stations, onto the map; and 3D effect tools Three.js and CSS3. The interfaces for the core heating business function modules include: heat source monitoring interface, heating network monitoring interface, heating station monitoring interface, weather forecast interface, load prediction model interface, alarm monitoring interface, video monitoring interface, and scene modeling interface.

3. The heating operation visualization system according to claim 1, characterized in that, The visualization intent analysis model is used to identify and analyze the user's visualization intent at the user interaction layer, obtaining multiple heating visualization sub-tasks: Through multiple rounds of information interaction between the intelligent processing layer and the user interaction layer, the user's visualization intent at the user interaction layer is obtained. In each round of information interaction, the visualization intent analysis model is used to perform intent recognition and analysis, obtaining multiple heating visualization sub-tasks round by round, specifically including: In the first round of interaction: an interface with a resolution drop-down box, a terminal type selection box, a physical size input box, a business function type selection button, and custom display requirements is designed. After initiating a visual display request to the user regarding the size, type, and core display business functions of the on-site display terminal, the user inputs relevant basic information. The core entities of the relevant basic information are identified through the large model of visual display intent analysis and transformed into structured basic parameters. In the second round of interaction: Based on the relevant basic information input by the user, various general visual layout templates for heating operation are recommended to the user, who can choose a template or customize their layout requirements; The core elements in the templates or customized layout requirements are identified through the large model of visual display intent recognition, including the layout position and display rules of each business function module, and transformed into structured layout parameters; In the third round of interaction: Based on the layout parameters of each business function module to be displayed, the system initiates requests to users regarding the number of devices (heat sources, heating networks, heating stations), the number of interfaces for heating data visualization points, and modeling parameters (including 3D modeling requirements, number of models, modeling detail, and rendering effects). Users provide feedback through multimodal methods such as text, images, and videos. After recognizing the multimodal feature parameters from the large model based on the visualization intent and performing feature fusion, the system generates device docking parameters and modeling requirements, which are then transformed into structured docking parameters and modeling requirements. In the fourth round of interaction: the visualization intent recognition model integrates structured basic parameters, layout parameters, docking parameters and modeling requirement parameters, and provides feedback to the user through the user interaction layer for information confirmation and completion. Then, the visualization intent recognition model integrates the layout parameters, docking parameters, modeling requirement parameters, display style and display effect of the scattered heating business function modules through the preset business field mapping table, forming the heating visualization display sub-task of each business function module.

4. The heating operation visualization system according to claim 3, characterized in that, The training process of the visualization intent analysis large model includes: The system acquires historical dialogue and interaction data regarding user requests for visualization of heating operation, and annotates the intent tags and entity parameters in the data; it acquires user-drawn visualization layout diagrams, 3D model reference diagrams, and GIS map reference diagrams, and annotates the location and model features of each business equipment component in the diagrams; it acquires reference videos of the visualization display effect of heating operation, extracts keyframes and annotates rendering effects, dynamic interactions, display style, and chart type features; and it acquires a heating parameter rule base to verify whether the display terminal meets the display requirements. The acquired information is integrated to build a visualization corpus. After pre-training the selected base model, and fine-tuning it using typical visualization cases of various heating business functional modules, a large visualization intent analysis model is formed.

5. The heating operation visualization system according to claim 1, characterized in that, The MCP server is used to encapsulate the interfaces of the core heating business function modules in the data resource layer into standardized data interfaces and to encapsulate various rules and tools into callable capability units through the MCP protocol, including: The system analyzes the input parameters, output formats, and access permissions of the core business modules interfaces in the data resource layer, including heat source monitoring, heat network monitoring, heating station monitoring, weather forecasting, load prediction models, alarm monitoring, video surveillance, and scene modeling. It also sets up interface adapters for different types of business function module interfaces, customizes unified and standardized interface request and response formats based on the MCP protocol, and builds an MCP server gateway to uniformly manage the standardized interfaces of various heating business function modules. The data resource layer, including data visualization rules for different heating services, scenario adaptation rules, interface building tools for various business function modules of heating operation, and data visualization tools, is broken down into independent and reusable sub-capability units. The functions, input and output parameters of each sub-capability unit are clearly defined. Based on the MCP protocol, a unified protocol format is defined for each sub-capability unit and a protocol adapter is set. The mapping between MCP server requests and sub-capability unit input parameters, the format of sub-capability unit output and MCP server response are configured, and the adapted sub-capability units are registered to the MCP server's registration center and the calling process and calling strategy of the sub-capability units are set.

6. The heating operation visualization system according to claim 5, characterized in that, The establishment of the MCP server gateway enables unified management of standardized interfaces for various heating service function modules, including: Register the standardized interfaces of each heating business function module to the MCP server gateway to generate a unique MCP interface address for each business function module. Set access permissions and requests per second limit for each business function module's interface; Monitor and manage the call volume, response time, and error rate of the interfaces of each business function module.

7. The heating operation visualization system according to claim 5, characterized in that, The process of registering each adapted sub-capability unit to the registration center of the MCP server includes: registering the ID, type, function description, input parameters, output parameters, calling permissions, and request limit per second of each adapted sub-capability unit to the MCP server gateway. The process of setting up the sub-capability unit invocation flow includes: the caller sending an MCP protocol request to the MCP server gateway; the MCP server gateway routing to the corresponding sub-capability unit based on the sub-capability unit's ID; the protocol adapter parsing the request parameters and calling the corresponding sub-capability unit to execute the logic; and the protocol adapter standardizing the execution result format and returning it to the caller via the MCP protocol. Configure the calling strategy for sub-capability units, including: based on the request limit per second configured in the MCP server gateway, reject sub-capability unit requests that exceed the limit threshold; reduce redundant calculations by caching the results of frequently called sub-capability units; and automatically call sub-capability units in the order of dependency if there are coupled dependency scheduling relationships.

8. The heating operation visualization system according to claim 1, characterized in that, The number of MCP hosts is two, one as the primary MCP host and the other as the backup MCP host; the number of MCP clients is set according to the number of each heating business function module; the number of MCP servers is set according to the number of rules, tools, and core heating business function module interfaces in the data resource layer; one MCP client can bind to multiple MCP servers, and there is a limit to the number of bindings.

9. The heating operation visualization system according to claim 1, characterized in that, The training process of the visualization display business resource matching large model includes: In the first stage, a pre-trained general large model is selected as the large model for matching business resources for visualization. The general large model is trained by incremental pre-training using general visualization domain data, so that the large model for matching business resources for visualization learns the mapping logic between running data and charts, device data and GIS maps, and 3D models. In the second phase, a heating operation visualization knowledge base will be formed, which includes the functions of each heating business module, the association rules of the heating business knowledge graph, the resource list of interfaces and capability units in the MCP server, the heating operation visualization requirement document, and historical heating operation visualization task cases. The visualization display business resource matching model trained in the first phase will be fine-tuned so that the visualization display business resource matching model can learn the domain knowledge of heating operation visualization. In the third stage, after obtaining various intentions for the visualization of heating operation and labeling the corresponding sub-tasks of the heating visualization display of each business function module, the corresponding MCP client and server resources are labeled for each sub-task. The visualization display business resource matching big model of the second stage is subjected to sub-task parsing and MCP resource matching training, so that the visualization display business resource matching big model outputs a structured sub-task description and a corresponding MCP resource call list based on the visualization intention of each sub-task. In the fourth stage, the communication links between the large-scale visualization display business resource matching model, the MCP host, the MCP client, and the MCP server are simulated. A multi-turn dialogue thought chain dataset is constructed, and the large-scale visualization display business resource matching model from the third stage is guided to master the interaction logic with the MCP host, MCP client, and MCP server through multi-turn dialogue fine-tuning using the multi-turn dialogue dataset. A simulated MCP environment is also built to allow the large-scale visualization display business resource matching model to be simulated in the simulation environment. The simulation interaction results are used to train and optimize the large-scale visualization display business resource matching model, ultimately forming an optimized large-scale visualization display business resource matching model.

10. The heating operation visualization system according to claim 9, characterized in that, The visualization display business resource matching big model autonomously determines the corresponding data interfaces and capability units to be invoked based on each sub-task, and automatically completes the visualization display of various heating business functions through multiple rounds of interaction between the MCP host, MCP client, and MCP server, including: The visualization-based business resource matching model analyzes the business type of each subtask and extracts the core elements of the subtask, outputting a structured description of the subtask. The visualization-based business resource matching model matches the corresponding MCP resource call list for each subtask based on the structured subtask description and the preset MCP client and MCP server resources. Then, it generates the communication link of MCP host-MCP client-MCP server for each subtask based on the subtask description and the MCP resource call list. The visualization display business resource matching big model encapsulates the MCP resource call list of each subtask into instructions and sends them to the MCP host. The MCP host routes the instructions to the corresponding MCP client according to the business type of the subtask. The MCP client converts the instructions into structured requests and calls the MCP server interface and capability unit corresponding to the resource list. Through multiple rounds of interaction between the visualization display business resource matching big model, the MCP host, the MCP client and the MCP server, the business data, matching rules and tools corresponding to each subtask are obtained. Through data integration, business module layout, display effect rendering, anomaly annotation and model building, the visualization display of each heating business function is completed.

Images