Search

CN-121997401-A - Heat supply operation visualization system based on MCP protocol and large model

CN121997401ACN 121997401 ACN121997401 ACN 121997401ACN-121997401-A

Abstract

The invention discloses a heat supply operation visualization system based on an MCP protocol and a large model, which comprises a data resource layer, an intelligent processing layer, a user interaction layer and a feedback information processing layer, wherein the data resource layer comprises different heat supply operation data visualization rules, scene adaptation rules, heat supply operation function module interface construction tools, data visualization tools and core heat supply function module interfaces, the intelligent processing layer analyzes the large model through visual display intention to identify visual display intention, after a plurality of visual display subtasks are obtained, the visual display service resource matching large model determines corresponding data interfaces and capability units which are called according to the subtasks, and the related information is obtained through interaction of an MCP host, a client and a server to complete visual display of heat supply function, and the user interaction layer is used for obtaining visual display intention of a user through a multi-round dialogue mechanism and calling intention analysis large model and is also used for displaying visual effect of the heat supply function of the user and performing iterative optimization according to feedback information.

Inventors

  • ZHOU YI
  • WU YANLING
  • JIN HEFENG
  • Huang Mengken

Assignees

  • 浙江英集动力科技有限公司

Dates

Publication Date
20260508
Application Date
20260128

Claims (10)

  1. 1. A heating operation visualization system based on MCP protocol and large model, characterized in that it comprises: The data resource layer comprises different heat supply service data visualization rules, scene adaptation rules, heat supply operation function module interface building tools, data visualization tools and core heat supply function module interfaces; The intelligent processing layer comprises a trained visual display intention analysis large model, a visual display business resource matching large model, an MCP host deployed based on an MCP protocol, an MCP client and an MCP server; The MCP server is used for packaging interfaces of the core heat supply service function modules in the data resource layer into standardized data interfaces and packaging various rules and tools into callable capacity units through an MCP protocol; the MCP host is an application program integrating the capacity of the visual display business resource matching large model, and is used for managing the scheduling of the visual display intention of a user and the visual display business resource matching large model, transmitting an instruction generated by the visual display business resource matching large model to a corresponding MCP client, receiving information fed back by the MCP client and transmitting the information to the visual display business resource matching large model; The MCP client is used for converting an instruction generated by matching the visual display service resources with the large model into a structured request, and transmitting the structured request to a corresponding MCP server to call various rules, tools and service function module interfaces; The visual display intention analysis large model is used for identifying and analyzing the visual display intention of a user at a user interaction layer, after a plurality of heat supply visual display subtasks are obtained, the visual display business resource matching large model autonomously determines a called corresponding data interface and a capability unit according to each subtask, and performs multi-round interaction through the MCP host, the MCP client and the MCP server to obtain related information and automatically complete visual display of each heat supply business function; the user interaction layer is used for obtaining the visual display intention of the user by a preset multi-turn dialogue mechanism and calling the visual display intention analysis large model, and is also used for displaying the visual effect of each service function of heat supply to the user and carrying out iterative optimization according to the feedback information of the user.
  2. 2. The heat supply operation visualization system according to claim 1, wherein the different heat supply operation data visualization rules comprise displaying heat supply operation numerical value data through a digital billboard and a tone scale marking rule, highlighting key indexes, displaying operation states through a GIS map and a 3D model and combining a thermodynamic diagram and a flow diagram, displaying heat supply operation trend data through a line diagram, marking a highest value and a lowest value, summarizing heat supply operation statistic data through a histogram or a card; The scene adaptation rule comprises that a heating scene mainly comprises basic colors and a technological sense style, and meanwhile, a 3D model is required to be attached to physical forms of an actual heat source, a pipe network and a heating power station, and the technological sense is enhanced by using environmental light and particle effects; the interface construction tool for each service function module for heat supply operation comprises an easy V and a DataV, and can directly drag the assembly; the data visualization tool comprises a chart tool ECharts and a D3.Js, a GIS map visual configuration professional tool, a 3D effect tool three.js and a CSS3, wherein the GIS map visual configuration professional tool can import heat supply city geographic base map data and then correlate the position, pipe diameter and trend of a heat supply network, the coordinates and attribute information of a heat source and a heat station to a map; the core heat supply service function module interfaces comprise a heat source monitoring interface, a heat network monitoring interface, a heat station monitoring interface, a weather forecast interface, a load prediction model interface, an alarm monitoring interface, a video monitoring interface and a scene modeling interface.
  3. 3. The heat supply operation visualization system according to claim 1, wherein the visual display intention analysis large model is used for identifying and analyzing the user visual display intention of the user interaction layer to obtain a plurality of heat supply visual display subtasks, wherein the heat supply visual display intention of the user interaction layer is obtained by performing multiple rounds of information interaction with the user interaction layer through the intelligent processing layer, and the intention identification analysis is performed in each round of information interaction through the visual display intention analysis large model, and the plurality of heat supply visual display subtasks are obtained round by round, and the heat supply visual display subtasks specifically comprise: In the first round of interaction stage, a resolution drop-down frame, a terminal type selection frame, a physical size input frame, a service function type selection button and an interface for self-defining display requirements are designed, after a visual display request for displaying the size, the type and the core service functions of a terminal on site is initiated to a user, relevant basic information is input by the user; In the second interactive stage, based on relevant basic information input by a user, recommending a plurality of general visual layout templates for heat supply operation for the user, and selecting the templates or the custom layout requirements by the user; In the third interaction stage, based on layout parameters of each service function module to be displayed, initiating a device number request of a heat source, a heat network and a heat station to be displayed, a device number request of a heat supply data visual point position interface number request, and a modeling parameter request comprising a 3D model modeling requirement, a model number, modeling fineness and a rendering effect to a user, and feeding back the modeling parameter by a text, picture and video multi-mode; In the fourth interactive stage, the large model is identified through visual demonstration intention to integrate structured basic parameters, layout parameters, docking parameters and modeling requirement parameters, and after information confirmation and complementation are carried out on the large model fed back to a user through a user interaction layer, the large model is identified through visual demonstration intention to integrate the distributed layout parameters, docking parameters, modeling requirement parameters, demonstration styles and demonstration effects of each service function module through a preset service field mapping table, so that a heat supply visual demonstration subtask of each service function module is formed.
  4. 4. A heating operation visualization system as in claim 3, wherein the visual presentation intent analysis training process of the large model comprises: Acquiring a historical dialogue interaction corpus of a user about a heat supply operation visual demand, marking intention labels and entity parameters in the corpus, acquiring a heat supply operation visual layout diagram, a 3D model reference diagram and a GIS map reference diagram which are drawn by the user by hand, marking positions and model characteristics of various business equipment components in the diagram, acquiring a reference video of a heat supply operation visual display effect, extracting a key frame, marking rendering effect, dynamic interaction, display style and chart type characteristics, and acquiring a heat supply parameter rule base for checking whether a display terminal accords with the display demand; And fusing the acquired various information to construct a visual display corpus, pre-training the selected base large model, and fine-tuning the visual display cases typical of the heat supply service function modules to form a visual display intention analysis large model.
  5. 5. The heating operation visualization system of claim 1, wherein the MCP server is configured to package, via an MCP protocol, a core heating service function module interface in a data resource layer as a standardized data interface, and package various rules and tools as callable capability units, including: The carding data resource layer comprises input parameters, output formats and calling authorities of core service module interfaces of heat source monitoring, heat network monitoring, heating power station monitoring, weather forecast, load prediction models, alarm monitoring, video monitoring and scene modeling, interface adapters are arranged for different types of service function module interfaces, unified standardized interface request and response formats are customized based on an MCP protocol, and an MCP server gateway is built to perform unified management on standardized interfaces of the service function modules for heat supply; Splitting the data resource layer comprising different heat supply service data visualization rules, scene adaptation rules, heat supply operation function module interface building tools and data visualization tools into independent and reusable sub-capability units, defining functions, input and output parameters of the sub-capability units, defining a unified protocol format and setting a protocol adapter for each sub-capability unit based on an MCP protocol, mapping an MCP server request and the sub-capability unit, outputting a format responded by the sub-capability unit and the MCP server, registering the adapted sub-capability units to a registration center of the MCP server, setting a calling flow and a calling strategy of the sub-capability unit.
  6. 6. The heating operation visualization system of claim 5, wherein the building of the MCP server gateway to uniformly manage the standardized interfaces of the various service functional modules for heating comprises: Registering the standardized interface of each service function module for heat supply to the MCP server gateway to generate a unique MCP interface address of each service function module; Setting the interface access authority and interface request number limit per second of each service function module; the call volume, response time and error rate of the interfaces of the service function modules are monitored and managed.
  7. 7. The heating operation visualization system of claim 5, wherein registering each adapted sub-capability unit with the registration center of the MCP server includes registering each adapted sub-capability unit's ID, type, function description, input parameters, output parameters, call permissions, and request per second limits with the MCP server gateway; setting a calling flow of a sub-capability unit, wherein the calling party sends an MCP protocol request to an MCP server gateway, the MCP server gateway routes the request to the corresponding sub-capability unit according to the ID of the sub-capability unit, and a protocol adapter analyzes the request parameter and calls the corresponding sub-capability unit to execute logic; Setting a calling strategy of the sub-capability unit, wherein the calling strategy comprises the steps of rejecting sub-capability unit requests exceeding a limiting threshold based on the limiting of the number of requests per second configured by a MCP server gateway, reducing repeated calculation on a sub-capability unit result of high-frequency calling through buffering, and calling according to a dependency sequence automatically if the sub-capability unit has a coupling dependency scheduling relationship.
  8. 8. The heating operation visualization system of claim 1, wherein the number of MCP hosts is 2, one is a primary MCP host and the other is a standby MCP host, the number of MCP clients is set according to the number of the service function modules for heating, the number of MCP servers is set according to the number of the interfaces of the service function modules for heating, the rules, tools, and cores in the data resource layer, and 1 MCP client can bind a plurality of MCP servers while a binding number limit is set.
  9. 9. A heating operation visualization system as in claim 1, wherein the training process of visually exposing business resource matching large models comprises: In the first stage, a pre-trained general large model is selected as a visual display business resource matching large model, and the general large model is trained in an incremental pre-training mode through general visual field data, so that the visual display business resource matching large model learns the mapping logic of operation data and a chart, equipment data and a GIS map and a 3D model; In the second stage, a heat supply service knowledge graph comprising functions and association rules of each heat supply service module, a resource list comprising interfaces and capability units in an MCP server, a heat supply operation visual demand document and a history heat supply operation visual task case are formed into a heat supply operation visual knowledge base, and a visual display service resource matching large model trained in the first stage is finely tuned to enable the visual display service resource matching large model to learn the field knowledge of the heat supply operation visualization; in the third stage, various intentions of the heat supply operation visualization are obtained, after the heat supply visualization display subtasks of the corresponding business function modules are marked, corresponding MCP client and server resources are marked for each subtask, subtask analysis and MCP resource matching training are carried out on the visualized display business resource matching large model in the second stage, and the visualized display business resource matching large model outputs structured subtask description and a corresponding MCP resource calling list according to the visualized intentions of the subtasks; In the fourth stage, simulating communication links of the visual display business resource matching large model, the MCP host, the MCP client and the MCP server, constructing a multi-round dialogue thinking chain data set, guiding the visual display business resource matching large model in the third stage to master interaction logic with the MCP host, the MCP client and the MCP server in a multi-round dialogue fine tuning mode through the multi-round dialogue data set, constructing a simulation MCP environment, simulating the visual display business resource matching large model in the simulation environment, training and optimizing the visual display business resource matching large model through the simulated interaction result, and finally forming the optimized visual display business resource matching large model.
  10. 10. The heating operation visualization system according to claim 9, wherein the service resource matching large model for visual presentation autonomously determines the invoked corresponding data interface and capability unit according to each subtask, and performs multiple rounds of interaction through the MCP host, the MCP client and the MCP server to obtain related information and automatically complete visual presentation of each service function of heating, and the system comprises: Analyzing the service type of each subtask and extracting the core elements of the subtask by using a visual display service resource matching large model, and outputting a structured subtask description; The visual display business resource matching large model matches corresponding MCP resource call lists for each subtask according to the subtask description based on the structuring and the preset MCP client and MCP server resources, and then generates communication links of the MCP host-MCP client-MCP server of each subtask according to the subtask description and the MCP resource call lists; The method comprises the steps of packaging an MCP resource call list of each subtask into an instruction by a visual display service resource matching large model, sending the instruction to an MCP host, routing the instruction to a corresponding MCP client by the MCP host according to the service type of the subtask, converting the instruction into a structured request by the MCP client, calling an MCP server interface and a capacity unit corresponding to the resource list, acquiring service data and matched rules and tools corresponding to each subtask after multiple rounds of interaction of the visual display service resource matching large model, the MCP host, the MCP client and the MCP server, and completing visual display of each service function of heat supply by data integration, service module layout, display effect rendering, abnormal labeling and model establishment.

Description

Heat supply operation visualization system based on MCP protocol and large model Technical Field The invention belongs to the technical field of heat supply system operation, and particularly relates to a heat supply operation visualization system based on an MCP protocol and a large model. Background The heat supply operation visualization is a core display platform of the intelligent heat supply system, complex operation data indexes are displayed in a multi-dimensional mode in the form of a histogram, a line graph, a thermodynamic diagram and the like, heat supply facilities such as a heat source, a heat supply network, a heat station and the like are subjected to full-flow monitoring and data presentation, various abnormal data are captured, alarm information is marked, meanwhile, a heat supply network GIS map and a 3D (three-dimensional) model of equipment are constructed to realize the fine reduction of the heat supply network and the equipment, the spatial distribution of the heat supply network and the equipment is clearly represented, and the heat load difference and the heat source geographical distribution characteristics of an area are assisted to be analyzed, so that dynamic interactive updating is supported. However, more and more service function modules related to the current heat supply operation visualization are needed, and some of the service function modules also need customized high-demand display, and need to be called into a plurality of databases, calculation models, GIS maps and process simulation models of a background of a heat supply system, and different visualization tools exist, but communication calling protocols, calling modes, data formats and the like of a data interface, a service module interface and the like can be different, so that each newly added module needs to customize and develop an adaptive interface, the development cost is high, the integration period is long, the flexibility is poor, the customized visualization display requirements are difficult to meet, and the unified management of each interface, module and the like is also impossible. Based on the technical problems, a new heating operation visualization system based on the MCP protocol and a large model needs to be designed. Disclosure of Invention The invention aims to solve the technical problems of overcoming the defects of the prior art, providing a heat supply operation visualization system based on an MCP protocol and a large model, solving the heterogeneous problems of non-uniform re-interface format and different tool calling modes of the traditional heat supply system by analyzing the large model through visual display intention, matching the large model with visual display service resources, and arranging an MCP host, an MCP client and an MCP server based on the MCP protocol, and being capable of identifying the visual display intention, automatically calling corresponding interface resources, tool resources and the like to realize the heat supply operation. In order to solve the technical problems, the technical scheme of the invention is as follows: the invention provides a heating operation visualization system based on MCP protocol and large model, comprising: The data resource layer comprises different heat supply service data visualization rules, scene adaptation rules, heat supply operation function module interface building tools, data visualization tools and core heat supply function module interfaces; The intelligent processing layer comprises a trained visual display intention analysis large model, a visual display business resource matching large model, an MCP host deployed based on an MCP protocol, an MCP client and an MCP server; The MCP server is used for packaging interfaces of the core heat supply service function modules in the data resource layer into standardized data interfaces and packaging various rules and tools into callable capacity units through an MCP protocol; the MCP host is an application program integrating the capacity of the visual display business resource matching large model, and is used for managing the scheduling of the visual display intention of a user and the visual display business resource matching large model, transmitting an instruction generated by the visual display business resource matching large model to a corresponding MCP client, receiving information fed back by the MCP client and transmitting the information to the visual display business resource matching large model; The MCP client is used for converting an instruction generated by matching the visual display service resources with the large model into a structured request, and transmitting the structured request to a corresponding MCP server to call various rules, tools and service function module interfaces; The visual display intention analysis large model is used for identifying and analyzing the visual display intention of a user at a user interaction layer, after a plurality of h