Intelligent management and control system based on large model and integrated multi-dimensional spatio-temporal data fusion

By building an intelligent management and control system based on large models and integrated multi-dimensional spatiotemporal data fusion, we have solved the problems of weak spatiotemporal correlation processing capabilities, difficulty in fusing multi-source heterogeneous data, delayed business response and difficulty in implementing AI technology in the engineering field, achieved efficient data fusion and intelligent services, and improved the level of intelligent management and control in the engineering field.

CN120653646AActive Publication Date: 2025-09-16BEIJING LONGRUAN TECHNOLOGIES INC +1

Patent Information

Application Number
CN202510743281.8
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2025-06-05
Publication Date
2025-09-16
Estimated Expiration
2045-06-05

AI Technical Summary

Technical Problem

The existing engineering management and control systems have weak spatiotemporal correlation processing capabilities, difficulty in fusing multi-source heterogeneous data, insufficient data representativeness and integrity, delayed business response and insufficient decision-making intelligence, and difficulty in integrating AI technology with business, resulting in poor intelligent management and control effects.

Method used

Build an intelligent management and control system based on large models and integrated multi-dimensional spatiotemporal data fusion, including a large spatiotemporal model base, a three-dimensional or four-dimensional integrated model of the engineering object, an intelligent body collaboration module, and an autonomous analysis, decision-making, and control module. Through adaptation technology, the engineering field understanding ability of the large model is improved, and multi-source data is integrated under a unified spatiotemporal benchmark to realize intelligent task planning and execution.

Benefits of technology

It significantly enhances spatiotemporal cognition and processing capabilities, realizes efficient data fusion and intelligent services, improves business response speed and decision-making intelligence, promotes the deep integration of AI and engineering business, and realizes adaptive and continuous optimization management and control.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN120653646A_ABST
    Figure CN120653646A_ABST
Patent Text Reader

Abstract

The invention discloses an intelligent management and control system based on a large model and integrated multi-dimensional spatio-temporal data fusion, and belongs to the field of artificial intelligence and engineering application. The system comprises a space-time large model base, an engineering object three-dimensional or four-dimensional integrated model, an intelligent agent cooperation module and an autonomous analysis decision and control module. The method comprises the following steps: embedding engineering field knowledge type public domain data based on a large model, forming a space-time large model base, fusing three-dimensional or four-dimensional geometry and attribute data of an engineering object and dynamic business data such as monitoring and equipment automation, constructing a private domain data source which is integrally expressed, stored and managed, and autonomously analyzing and deciding by utilizing the space-time large model reasoning capability. And the scheduling agent collaboratively completes task planning and arrangement, safety production analysis and decision support and control, and automatically generates various engineering image-text reports. The intelligent management and control problem in the engineering fields of mines, factories, traffic, water conservancy, buildings, machinery and the like can be solved, and the practicability of an artificial intelligence system in the engineering field is remarkably improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the fields of artificial intelligence and engineering, and in particular to an intelligent management and control system based on a large model and integrated multi-dimensional spatiotemporal data fusion. Background Art

[0002] With the rapid development of artificial intelligence (AI) large-scale modeling technology, its application is becoming increasingly widespread across various industries. However, in engineering fields (such as mining, chemical engineering, water conservancy, urban construction, and environmental governance), the in-depth application and value realization of large-scale modeling technology still face many challenges.

[0003] Existing engineering management and control platforms or systems often have the following problems: (1) Weak ability to process spatiotemporal correlations. Engineering objects (such as geological bodies, mine tunnels, buildings, and pipeline networks) and their environments have complex three-dimensional spatial structures and temporal evolution characteristics. Existing systems have difficulty effectively processing and understanding this inherent spatiotemporal correlation, especially for multidimensional vector graphics commonly used in the engineering field (such as CAD, BIM, and GIS data). The system is unable to perform deep spatiotemporal reasoning, and the application of large models is limited.

[0004] (2) Data integration and utilization are difficult. The entire life cycle of project management (covering planning, design, construction, operation, maintenance, and even abandonment) generates massive amounts of multi-source heterogeneous data, including three-dimensional geometry and attribute models in the design phase, progress and material data in the construction phase, monitoring and operation data in the operation phase, and management data such as safety, production, and operations. These dynamic data streams have different standards, diverse formats, non-standard storage directory systems, and inconsistent time and space benchmarks. It is difficult to effectively integrate them to form a globally consistent cognitive foundation, resulting in a serious data silo phenomenon.

[0005] (3) The data lacks representativeness and integrity, which is not conducive to improving the accuracy of model training and management applications. Spatiotemporal data is often a single object or a single type of two-dimensional or three-dimensional graphics or images, lacking the most complete integrated expression of the engineering object's geometry and associated attribute data. The expression and retrieval of data are discrete, rather than derived from a unified data base.

[0006] (4) Delayed business response and insufficient intelligent decision-making. Engineering management and control face a dynamically changing environment and complex business needs. Existing systems often respond based on fixed rules or simple models, lacking a deep understanding of complex business logic and the ability to accurately predict future trends, resulting in delayed business response. The decision-making process still relies heavily on manual experience and lacks optimized decision support based on global information and intelligent algorithms, making it difficult to cope with complexity and uncertainty.

[0007] (5) It is difficult to integrate AI technology with business. Although general AI big models are powerful, they lack professional knowledge in the engineering field and understanding of business processes, and their direct application in engineering scenarios has limited effect. How to deeply integrate big model capabilities with core engineering business (such as planning and design, risk assessment, schedule optimization, and safety management) to form an end-to-end intelligent solution is a technical challenge that needs to be solved urgently.

[0008] Therefore, there is an urgent need for a new technical solution that can overcome the above-mentioned defects, fully utilize the capabilities of AI big models, combine the spatiotemporal data characteristics of the engineering field, and realize intelligent analysis, decision-making and adaptive management of the entire life cycle of engineering fields such as smart mines, smart factories, water conservancy projects, urban projects, and environmental projects. Summary of the Invention

[0009] In view of the above problems, the present invention aims to overcome the shortcomings of the existing technology and provide an intelligent management and control system based on a large model and integrated multi-dimensional spatiotemporal data fusion.

[0010] The embodiment of the present invention provides an intelligent management and control system based on a large model and integrated multi-dimensional spatiotemporal data fusion, including: A spatiotemporal large model base, configured to receive engineering-related input information and perform intelligent task planning and orchestration based on its deep understanding, knowledge association, and logical reasoning of data provided by its engineering domain expertise, the engineering-related input information, and the three-dimensional or four-dimensional integrated model of the engineering object. The input information includes at least business requirements and multimodal data. The three-dimensional or four-dimensional integrated model of the engineering object, based on the full element, three-dimensional or four-dimensional geometry and attribute data of the engineering object under a unified spatiotemporal reference, integrates the dynamic private domain data related to the engineering business, and constructs and dynamically updates a spatiotemporal data volume that reflects the true overall picture and life cycle of the project through integrated spatiotemporal expression and processing. This spatiotemporal data volume serves as the three-dimensional or four-dimensional integrated model of the engineering object, which is used to respond to data requests and, through processing, generation or synthesis, forms global spatiotemporal situation information; An agent collaboration module, comprising multiple functional agents and providing a communication and coordination mechanism between the functional agents, is used to receive task instructions corresponding to intelligent task planning and orchestration issued by the spatiotemporal large-scale model base, and, under the scheduling and monitoring of the autonomous analysis, decision-making and control module, obtain the data, analysis results or generated content required to execute the task from the three-dimensional or four-dimensional integrated model of the engineering object, thereby driving the execution of the functional agents; The autonomous analysis, decision-making and control module is used to obtain global spatiotemporal situation information from the three-dimensional or four-dimensional integrated model of the engineering object, and receive execution feedback results from the functional intelligent body collaboration module, and utilize the reasoning ability of the spatiotemporal large model base for the engineering field, combined with the global spatiotemporal situation information, the execution feedback results and preset or dynamically generated rules and goals, to perform autonomous analysis and judgment, and based on the results of the autonomous analysis and judgment, autonomously schedule the intelligent body collaboration module to perform subsequent tasks, and generate intelligent decision results or control instructions.

[0011] Optionally, the spatiotemporal large model base is an intelligent model capable of deeply understanding and processing unique problems in the engineering field, wherein the ability to deeply understand and process unique problems in the engineering field is obtained through an adaptation method; The adaptation method includes a combination of one or more of the following techniques: Domain fine-tuning technology, which refers to the use of large-scale text corpora, code corpora, and various public domain multimodal data in the engineering field to perform supervised fine-tuning, instruction fine-tuning, or continuous pre-training of intelligent models; Knowledge enhancement technology, which refers to: combining a spatiotemporal data volume constructed from a three-dimensional or four-dimensional integrated model of the engineering object or an external professional knowledge base, using retrieval-enhanced generation, knowledge graph embedding, or arbitrary knowledge injection, and extracting knowledge related to specific problems in the engineering field and providing it as additional background knowledge to the spatiotemporal large model base; Spatiotemporal capability enhancement technology, which involves training based on the spatiotemporal characteristics of engineering domains. Specifically, this includes designing tasks that incorporate spatiotemporal concepts, constructing datasets with precise spatiotemporal labels, and training on a large-scale base model to enhance the base model's ability to understand spatiotemporal queries, reason about spatiotemporal relationships, and predict spatiotemporal evolution patterns. Prompt engineering technology refers to designing optimized prompt word templates for engineering tasks, which contain clear instructions, context information, constraints, and expected output formats.

[0012] Optionally, the three-dimensional or four-dimensional integrated model of the engineering object is a collection of advanced data and knowledge in the form of a knowledge graph or digital twin, constructed based on integrated engineering spatiotemporal data, containing rich semantic relationships and business information, and supporting intelligent reasoning and analysis, and is obtained by the following method: According to the project scenario, select relative time or absolute time as the time reference, select local coordinate system or geographic coordinate system as the spatial reference, and define a unified time and space benchmark; Under a unified spatiotemporal benchmark, data from different sources and formats are accessed, associated, and integrated to form a consistent, analyzable data set. These data from different sources and formats are divided into two core categories: the first is the data corresponding to the integrated 3D geometry and attributes model that serves as the core spatiotemporal foundation; the second is various dynamic data streams covering the entire project lifecycle. Based on the consistent and analyzable data set, and taking the geographic environment and physical engineering entities as the foundation, a spatial geometry and attribute model of the engineering project is established; based on the spatial model, the engineering object is used as the element unit, and at least one time series data including environmental monitoring and automated operation is integrated; knowledge engineering technology is used to explicitly express the engineering entities and their relationships, and a three-dimensional or four-dimensional integrated model of the engineering object is constructed and dynamically maintained; The dynamic private domain data related to engineering business include: monitoring and supervision, equipment automation data, and dynamic full life cycle data that reflects the true picture of the project.

[0013] Optionally, the three-dimensional or four-dimensional integrated model of the engineering object responds to traditional query requests and directly returns the geometric data and attribute data contained in the three-dimensional or four-dimensional engineering object according to the requirements of the upper-level module; or The three-dimensional or four-dimensional integrated model of the engineering object extracts the geometric data or attribute data from the three-dimensional or four-dimensional integrated model of the engineering object in response to traditional query requests and based on the needs of the upper-level modules through the understanding and generation capabilities of the spatiotemporal large model base, wherein the extraction includes at least one of: object attribute query, local graphic output, and three-dimensional model sectioning; Based on the geometric data or attribute data, the spatiotemporal big model conducts comprehensive analysis or calls external tools for processing to output new generative content; the generative content refers to: output that goes beyond the original data or simple query results, is generated after intelligent processing, and has business value, and the type is at least one of text results, two-dimensional graphics, and three-dimensional models.

[0014] Optionally, the agent collaboration module provides management, communication, and collaboration of agents; it ensures the normal operation of each functional agent through registration, discovery, and management; The plurality of functional agents interact with each other on data and status and execute the operations through the communication channel provided by the agent collaboration module; If any functional agent encounters an abnormality during execution, based on the analysis and understanding of the spatiotemporal model base and the autonomous decision-making and control module, the agent collaboration module provides multiple rounds of task allocation execution and abnormal state resolution.

[0015] Optionally, the autonomous analysis, decision-making and control module is also used for autonomous analysis and judgment of engineering problems, specifically including: The autonomous analysis, decision-making and control module uses the large space-time model base to output the engineering knowledge background related to the engineering problem and the list of steps to solve the problem based on the engineering problem, and obtains relevant space-time situation data from the three-dimensional or four-dimensional integrated model of the engineering object to form complete knowledge, methods and data for problem analysis and solution. The large space-time model base conducts autonomous analysis and judgment, and calls the intelligent body collaboration module to execute, and outputs the decision-making results or control instructions of the engineering problem.

[0016] Optionally, the spatiotemporal large model base is also used to process and parse the multimodal data, where the multimodal data includes at least one of text, sound, image, video, vector graphics and structured data.

[0017] Optionally, the specific implementation form of the three-dimensional or four-dimensional integrated model of the engineering object is: an engineering spatiotemporal knowledge graph or an engineering digital twin, which is a platform that integrates multi-source spatiotemporal data of the engineering and organizes them into spatiotemporal data and knowledge with rich semantic relationships for reasoning and understanding by the spatiotemporal large model; The engineering spatiotemporal knowledge graph uses a graph structure to explicitly express engineering entities and the complex relationships between them, and contains and represents a large amount of business data; The engineering digital twin constructs a high-fidelity virtual model of the physical engineering entity, and uses fused dynamic data to achieve real-time or near real-time synchronization of the virtual model and the physical entity state, supporting simulation and interaction.

[0018] Optionally, the full-element, multi-dimensional geometric and attribute data of the engineering object include but are not limited to: a full-stratum geological model, a building information model BIM, a city information model CIM, or a three-dimensional factory model, a three-dimensional infrastructure model.

[0019] Optionally, the various types of dynamic data streams covering the entire life cycle of the project include: personnel activity data, safety monitoring data, equipment operation status data, environmental monitoring data, construction progress data, material consumption data, cost data or at least one of production and operation management data.

[0020] Optionally, the functional agents in the agent collaboration module include at least one of: a data acquisition and preprocessing agent, a domain model calculation agent, a risk identification and assessment agent, a resource scheduling optimization agent, a report and visualization generation agent, or a human-computer interaction interface agent; or The functional intelligent agent includes at least one combination of a spatiotemporal analysis intelligent agent, a process execution intelligent agent, a decision optimization intelligent agent and an industrial control intelligent agent.

[0021] Optionally, the intelligent decision-making results or control instructions include: safety risk level assessment, abnormal event warning, performance prediction analysis, resource allocation plan, operation optimization suggestion, emergency response plan, or at least one of adaptive control parameter set.

[0022] Optionally, the intelligent management and control system is further configured to receive feedback information from the execution layer, and integrate the feedback information into the three-dimensional or four-dimensional integrated model of the engineering object for updating, and trigger adaptive adjustment of the strategy of the spatiotemporal large model base or the scheduling of the autonomous analysis, decision-making and control module; The adaptive adjustment is used to dynamically update the spatiotemporal data body, and automatically call and execute functional intelligent bodies based on the spatiotemporal large model base to regularly or automatically generate intelligent decision-making results or control instructions.

[0023] Optionally, the application method of the intelligent management and control system includes the following steps: Step T1: Utilizing the spatiotemporal large model base, deeply analyzing the input engineering business requirements, combining the state and spatiotemporal characteristics of the complete spatiotemporal data volume in the current three-dimensional or four-dimensional integrated model of the engineering object, the engineering business requirements are automatically converted into a series of subtasks executed in the intelligent agent collaboration module; Step T2: The autonomous analysis, decision-making, and control module dynamically schedules the relevant functional agents in the agent collaboration module based on the subtasks and real-time status. The relevant functional agents utilize the communication and collaboration mechanism to exchange information and allocate tasks. When necessary, they obtain the processed, generated, or synthesized spatiotemporal data, analysis results, or generated content required to perform the tasks from the three-dimensional or four-dimensional integrated model of the engineering object, so as to collaboratively perform the subtasks. Step T3: The autonomous analysis, decision-making, and control module obtains the global spatiotemporal situation information from the three-dimensional or four-dimensional integrated model of the engineering object, combines the execution results and status feedback of the functional agents, and utilizes the reasoning capabilities of the spatiotemporal large model base for the engineering field and preset or dynamically generated rules and goals to conduct autonomous analysis and judgment, and based on this, generates intelligent decision results or control instructions; Step T4: Output the decision result or control instruction to the execution level, receive execution effect feedback or environmental change information, and use the execution effect feedback or environmental change information to update the three-dimensional or four-dimensional integrated model of the engineering object, and trigger adaptive adjustment of the strategy of the spatiotemporal large model base or the scheduling of the autonomous analysis decision and control module.

[0024] The present invention aims to address existing issues in the application of artificial intelligence in engineering, such as weak spatiotemporal correlation processing capabilities, difficulty fusing multi-source heterogeneous data, delayed business response, insufficient intelligent decision-making, and the shallow integration of AI technology with engineering business. By introducing a spatiotemporal large model optimized for the engineering field as the intelligent core, the present invention constructs a three-dimensional or four-dimensional integrated model of the engineering object with a multidimensional data model of all elements of the engineering object as the core. Through intelligent agent collaboration and autonomous decision-making, the intelligent management and control system provided by the present invention has one or more of the following beneficial effects compared to existing technologies: (1) Significantly enhance the spatiotemporal cognition and processing capabilities: With the engineering native full-factor multidimensional data model as the core, the system integrates dynamic data and manages it under a unified spatiotemporal benchmark. Combined with the understanding of spatiotemporal correlation and evolution patterns of the spatiotemporal large model, it greatly enhances the system's spatiotemporal perception and analysis capabilities for complex engineering environments and can effectively process multidimensional vector graphics information.

[0025] (2) Realize efficient data fusion and intelligent services: The three-dimensional or four-dimensional integrated model of the engineering object not only solves the problem of data islands, but also can respond to requests to actively process, generate and integrate new spatiotemporal data, analysis results or generative content (for example, risk assessment results after spatiotemporal analysis, equipment health index generated based on multi-source monitoring data, daily summary automatically generated based on templates and real-time data, preliminary optimization layout suggestions or design elements, etc.), providing a powerful and intelligent new data support paradigm for upper-level intelligent applications, which is different from the traditional data query mode.

[0026] (3) Improve business response speed and decision-making intelligence: The spatiotemporal big model can deeply understand business needs and quickly decompose and plan tasks. The autonomous decision-making module can make accurate judgments and optimize scheduling based on the global spatiotemporal situation and model reasoning, which significantly improves business response speed and the scientific and forward-looking nature of decision-making.

[0027] (4) Promote the deep integration of AI and engineering business: Closely combine the understanding and reasoning capabilities of the domain-adapted spatiotemporal large model with engineering spatiotemporal data and specific business processes (executed by intelligent agents), forming a closed loop from demand to execution to feedback, so that AI technology can be truly implemented and empower the core business of engineering.

[0028] (5) Achieve adaptive and continuous optimization management and control: Through the closed-loop feedback mechanism, the system can continuously adjust strategies and optimize models according to actual execution results and environmental changes, achieving adaptive and intelligent management and control of the entire life cycle of the project.

[0029] (6) Broad application prospects: The system and method can be flexibly applied to multiple engineering fields such as smart mines, smart cities, smart factories, smart transportation, smart water conservancy, smart energy, and smart environment, and has good versatility and scalability.

[0030] The intelligent management and control system proposed in the present invention, based on the fusion of large models and integrated multidimensional spatiotemporal data, builds a multidimensional spatiotemporal data platform with multidimensional geometry and attribute models as its core by constructing a large spatiotemporal model base optimized for the engineering field. This platform provides intelligent data services and information products, and uses intelligent agents to collaboratively execute tasks and autonomous decision-making modules to conduct command, dispatch, and closed-loop control. This effectively solves many of the pain points of the existing technology in applying AI large models in the engineering field, and can significantly improve the level of intelligent management and control throughout the entire life cycle of the project. The present invention can solve the problems of intelligent management and control in engineering fields such as mining, chemical plants, water conservancy projects, and construction, significantly improving the practicality of artificial intelligence systems in engineering fields and having high practicality. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Various other advantages and benefits will become apparent to those skilled in the art upon reading the detailed description of the preferred embodiment below. The accompanying drawings are for illustration purposes only and are not to be considered as limiting the present invention. The same reference symbols are used throughout the drawings to represent the same components. In the drawings: Figure 1 This is an architectural diagram of an intelligent management and control system based on a large model and integrated multi-dimensional spatiotemporal data fusion proposed in an embodiment of the present application; Figure 2 It is an application method of the intelligent management and control system based on large models and integrated multi-dimensional spatiotemporal data fusion in the embodiment of the present application. DETAILED DESCRIPTION

[0032] In order to make the above-mentioned objects, features and advantages of the present invention more clearly understood, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only intended to explain the present invention, are only part of the embodiments of the present invention, not all of the embodiments, and are not intended to limit the present invention.

[0033] The present invention proposes an intelligent management and control system based on a large model and integrated multi-dimensional spatiotemporal data fusion, referring to Figure 1 The structural schematic diagram shown includes: a large space-time model base, a three-dimensional or four-dimensional integrated model of the engineering object, an intelligent body collaboration module, and an autonomous decision-making and control module.

[0034] The spatiotemporal large model base is used to receive engineering-related input information and perform intelligent task planning and orchestration based on its deep understanding, knowledge association and logical reasoning of engineering domain expertise, engineering-related input information and data provided by the three-dimensional or four-dimensional integrated model of the engineering object. The input information at least includes business needs and multimodal data.

[0035] The "space-time large model base" in the present invention refers to a large-scale artificial intelligence model with powerful general capabilities and adapted and optimized for the engineering field. The "space-time large model base" here does not refer to an unmodified general large model, but it must have the ability to deeply understand and handle problems unique to the engineering field. This ability is obtained through an adaptation process, which is one of the key features of the present invention that distinguishes it from simply calling a general large model API. This adaptation is intended to enable it to deeply understand the professional knowledge, terminology, and specifications of the public domain of the engineering industry, the business data and business processes of the private domain of engineering, and the complex space-time relationships contained therein.

[0036] Adaptation methods may include, but are not limited to, a combination of one or more of the following techniques: Domain fine-tuning technology: This is a key adaptation method. This technology utilizes a variety of public-domain data, including large-scale text corpora from the engineering field (such as industry reports, design specifications, construction plans, safety regulations, maintenance manuals, and project management documents), code corpora (such as engineering calculation scripts and simulation model code), and multimodal data (described in detail below), for supervised fine-tuning, instruction fine-tuning, or continuous pre-training. This enables the spatiotemporal large-scale model base to more accurately understand engineering language, follow engineering instructions, and generate text or code that conforms to engineering specifications. For example, using a large number of mine safety accident reports to fine-tune the model enables it to better understand the terminology and causal relationships related to mine risks.

[0037] Knowledge Enhancement technology: Another important adaptation method. Combining spatiotemporal data bodies (such as knowledge graphs) constructed from three-dimensional or four-dimensional integrated models of engineering objects or external professional knowledge bases (such as material databases and equipment knowledge bases), retrieval-augmented generation (RAG), knowledge graph embedding, or other knowledge injection technologies are used, as well as knowledge related to specific problems in the engineering field, and provided as additional background knowledge to the spatiotemporal large model base. When the spatiotemporal large model base is reasoning, relevant structured engineering knowledge is dynamically retrieved and injected to improve the accuracy, professionalism, and factual consistency of the spatiotemporal large model base's answers and reduce "hallucinations." For example, when the spatiotemporal large model base answers a question about the properties of a certain material, accurate data can be retrieved from the material database and integrated into the answer.

[0038] Spatiotemporal capability enhancement technology: Training targets the strong spatiotemporal characteristics of the engineering field. By designing tasks that incorporate spatiotemporal concepts (such as understanding topological relationships, time series prediction, and spatial pattern recognition), and using data with precise spatiotemporal labels for training, the spatiotemporal large-scale model base is enhanced in its ability to understand spatiotemporal queries, reason about spatiotemporal relationships, and predict spatiotemporal evolution patterns. For example, the spatiotemporal large-scale model base is trained to understand spatiotemporal queries such as "Find all microseismic events within 500 meters of point A that occurred within the past 24 hours."

[0039] Prompt Engineering: Design optimized prompt templates for engineering tasks, including clear instructions, contextual information, constraints, and desired output formats, to better guide the adapted spatiotemporal model base to complete tasks.

[0040] Regarding the aforementioned multimodal data: The spatiotemporal large model base is used to process multimodal data commonly found in engineering fields. "Multimodal data" here refers to data types that encompass a variety of different information forms. The specific content varies depending on the engineering application scenario, but generally includes at least the following core types: (1) Text: various engineering documents, reports, records, standards and specifications, etc.

[0041] (2) Images and vector graphics: engineering drawings (GIS, CAD, scans), visualization of BIM / CIM models, on-site photos, satellite / UAV remote sensing images, surveillance video screenshots, etc. The model needs to have the ability to understand graphic symbols, recognize image content, and associate graphic and text information.

[0042] (3) Structured data: time series data collected by sensors, monitoring reports, equipment operating parameter tables, bills of materials, cost data, project schedules, etc.

[0043] In some preferred embodiments, the spatiotemporal model base can also be configured to process other unstructured data according to application requirements, such as: (1) Sound: such as the operating sound of key equipment, used for condition monitoring or abnormality diagnosis (such as abnormal bearing noise, rock cracking sound) through sound pattern recognition.

[0044] (2) Video: such as continuous monitoring video of the construction site, which is used for personnel behavior identification, safety violation detection, automatic progress tracking, equipment motion analysis, etc.

[0045] The spatiotemporal large model base needs to have the ability to integrate and process these different modal information. For example, it needs to be able to understand a text describing a device failure and associate it with the corresponding device operation sound clip or monitoring video image.

[0046] Core function summary: Receive engineering business requirements; use its adapted engineering understanding and spatiotemporal cognition capabilities to parse requirements; query the three-dimensional or four-dimensional integrated model of the engineering object to obtain relevant information; perform task planning; generate instructions and send them to the intelligent agent collaboration module; understand intelligent agent feedback; provide analytical and judgment support based on engineering knowledge and spatiotemporal reasoning for the autonomous decision-making and control module.

[0047] A 3D or 4D integrated model of an engineering object is constructed and dynamically updated based on the full-element, 3D or 4D geometry and attribute data of the engineering object, integrated with dynamic private domain data related to the engineering business, under a unified spatiotemporal reference. Through integrated spatiotemporal expression and processing, a spatiotemporal data volume is constructed and dynamically updated, reflecting the true full picture of the project and its entire lifecycle. This spatiotemporal data volume serves as the 3D or 4D integrated model of the engineering object. This data volume is used to respond to data requests (from the spatiotemporal large-scale model base, the intelligent agent collaboration module, or the autonomous analysis, decision-making, and control module) by processing, generating, or synthesizing global spatiotemporal situational information to provide the spatiotemporal data, analysis results, or generated content required for training the spatiotemporal large-scale model base, intelligent agent collaboration execution, and analysis, decision-making, and control. This full-element, multi-dimensional geometry and attribute data of the engineering object includes, but is not limited to, full-stratum geological models, building information models (BIMs), city information models (CIMs), or 3D plant and infrastructure models.

[0048] The "3D or 4D integrated model of the engineering object" in this invention is a core data hub. It is not only a data repository but also a provider of data fusion, knowledge construction, and intelligent services. The core of the intelligent management and control system of this invention is to fuse data into a spatiotemporal data volume based on a unified benchmark, and to provide intelligent analysis in conjunction with the spatiotemporal large-scale model foundation.

[0049] Unified space-time benchmark: Defining and implementing a globally consistent space-time reference framework to provide a unified space-time "anchor" for all data is the basis for data fusion and space-time analysis.

[0050] Data Access and Fusion: This is responsible for accessing two core data types: 1. Integrated 3D geometry and attribute models (such as BIM and geological models) that serve as the core spatiotemporal foundation; 2. Various dynamic data streams covering the entire project lifecycle (from planning and design to operations and maintenance) (such as monitoring, status, cost, and activity data). "Fusion" here refers to the process of correlating and integrating these data from diverse sources and formats under a unified spatiotemporal benchmark to form a consistent, analyzable data set. This includes a comprehensive, multi-dimensional integration of the entire project object, including models of all mine strata, tunnels, pipelines, and equipment. These dynamic data streams covering the entire project lifecycle include at least one of: personnel activity data, safety monitoring data, equipment operating status data, environmental monitoring data, construction progress data, material consumption data, cost data, or production and operations management data.

[0051] Construction and management of integrated 3D or 4D models of engineering objects: This is a key step in transforming fused data into knowledge. Based on the unified data set provided by the data access and fusion layer, this layer utilizes knowledge engineering techniques (such as ontology modeling, entity recognition, relationship extraction, semantic linking, and rule-based reasoning) to construct and dynamically maintain an integrated 3D or 4D model of the engineering object. This "integrated model" emphasizes the structured, semantic, and reasonable nature of its content, far surpassing traditional databases. Preferred implementations include: Engineering spatiotemporal knowledge graph: It uses a graph structure to explicitly express engineering entities (such as equipment, components, measurement points, events) and the complex relationships between them (such as spatial inclusion, topological connection, temporal sequence, causal influence, attribute association, etc.), and contains and represents a large amount of business data (for example, the ID, attributes, and historical status changes of a specific device as nodes and attribute values ​​in the graph).

[0052] Engineering digital twin: It builds a high-fidelity virtual model of a physical engineering entity and uses fused dynamic data to achieve real-time or near real-time synchronization between the virtual model and the physical entity state, supporting simulation and interaction.

[0053] Therefore, in this context, a "3D or 4D integrated model of an engineering object" refers to a collection of advanced data and knowledge, such as a knowledge graph or digital twin, built from integrated engineering spatiotemporal data. This collection contains rich semantic relationships and business information, and supports intelligent reasoning and analysis. Specifically, it is a platform that integrates multi-source spatiotemporal engineering data, organizing it into a rich set of semantically rich spatiotemporal data and knowledge that supports reasoning and understanding within the large spatiotemporal model.

[0054] Data processing and services (embodying a new input-output paradigm): Provide intelligent data services. It not only responds to traditional query requests but also proactively utilizes integrated 3D or 4D models of engineering objects to perform data processing (such as spatiotemporal aggregation and feature extraction), information synthesis (such as weighted assessment of multi-source risk factors), or content generation (such as automated report summaries and preliminary design elements) based on the needs of upper-level modules. These services then provide this new "generative content." "Generative content" here refers to outputs that go beyond raw data or simple query results and are intelligently processed to create business-valued outputs. For example, this could be: a comprehensive risk assessment including risk level, impact scope, and confidence level; a spatiotemporal heat map predicting traffic congestion for the next hour; a list of optimized material distribution recommendations based on current inventory and production plans; or a preliminary draft of a compliant equipment installation flowchart containing key parameters. This model enables the data platform to become a crucial support for intelligent decision-making, rather than simply a passive data source.

[0055] The three-dimensional or four-dimensional integrated model of the engineering object responds to traditional query requests and, based on the needs of the upper-level modules, directly returns the geometric data and attribute data contained in the three-dimensional or four-dimensional engineering object; or, the three-dimensional or four-dimensional integrated model of the engineering object responds to traditional query requests and, based on the needs of the upper-level modules, extracts geometric data or attribute data from the three-dimensional or four-dimensional integrated model of the engineering object through the understanding and generation capabilities of the spatiotemporal large model base, where the extraction includes at least one of: object attribute query, local graphic output, and three-dimensional model sectioning.

[0056] Based on geometric data or attribute data, the spatiotemporal model conducts comprehensive analysis or calls external tools for processing to output new generative content. Generative content refers to output that goes beyond the original data or simple query results, is generated after intelligent processing, and has business value. The type is at least one of text results, two-dimensional graphics, and three-dimensional models.

[0057] The intelligent agent collaboration module contains multiple (two or more) functional intelligent agents and provides a communication and collaboration mechanism between functional intelligent agents. The intelligent agent collaboration module is used to receive task instructions corresponding to intelligent task planning and orchestration issued by the spatiotemporal large model base, and under the scheduling and monitoring of the autonomous analysis, decision-making and control module, drive the functional intelligent agents to use the communication and collaboration mechanism to conduct information exchange and task allocation, and obtain the spatiotemporal data, analysis results or generative content required to perform tasks from the three-dimensional or four-dimensional integrated model of the engineering object, so as to collaboratively execute task instructions to process data, perform calculations or interact with external systems, and feedback the status information or final results of the execution process to the autonomous analysis, decision-making and control module or the spatiotemporal large model base.

[0058] The agent collaboration module provides agent management, communication infrastructure, and collaborative mechanisms and protocol frameworks. Agent management is responsible for agent registration, discovery, and lifecycle management. Communication infrastructure ensures reliable and efficient information exchange between agents. Collaboration mechanisms and protocols implement collaborative logic such as task allocation, information sharing, and conflict resolution.

[0059] The intelligent agent collaboration module provides management, communication, and collaboration of intelligent agents; it ensures the normal operation of each functional intelligent agent through registration, discovery, and management; multiple functional intelligent agents interact and execute each other's data and status through the communication channel provided by the intelligent agent collaboration module; if any functional intelligent agent encounters an abnormality in its execution, the intelligent agent collaboration module provides multiple rounds of task allocation execution and abnormal status resolution based on the analysis and understanding of the spatiotemporal large model base and the autonomous decision-making and control module.

[0060] Functional Agent Instances: These include at least two functional agents. These can be general-purpose (data acquisition, model calculation, risk assessment, optimization, reporting, interaction, etc.) or more engineering-process-focused (spatiotemporal analysis, process execution, decision optimization, industrial control, etc.). Specifically, the functional agents in the agent collaboration module include at least one of a data acquisition and preprocessing agent, a domain model calculation agent, a risk identification and assessment agent, a resource scheduling and optimization agent, a reporting and visualization generation agent, or a human-computer interface agent. Alternatively, the functional agents include at least one combination of a spatiotemporal analysis agent, a process execution agent, a decision optimization agent, and an industrial control agent. These functional agents serve as the fundamental units for executing specific tasks.

[0061] The autonomous analysis, decision-making and control module is used to obtain global spatiotemporal situation information from the three-dimensional or four-dimensional integrated model of the engineering object, receive execution feedback results from the functional intelligent agent collaboration module, and use the reasoning ability of the spatiotemporal large model base for the engineering field, combined with the global spatiotemporal situation information, execution feedback results and preset or dynamically generated rules and goals, to conduct autonomous analysis and judgment, and based on the results of autonomous analysis and judgment, autonomously schedule the intelligent agent collaboration module to perform subsequent tasks, and generate intelligent decision results or control instructions.

[0062] The autonomous analysis, decision-making, and control module is responsible for high-level situational awareness, analysis, judgment, and command and dispatch. It includes: information acquisition, core decision-making logic, autonomous dispatch, and decision or control output.

[0063] Among them, information acquisition: obtaining processed, generated or integrated global spatiotemporal situation information from the three-dimensional or four-dimensional integrated model of the engineering object, as well as the execution feedback of the intelligent body collaboration module.

[0064] Core decision-making logic (embodying autonomy): The key to its autonomy lies in leveraging the engineering-specific reasoning capabilities of the spatiotemporal big model. This logic inputs comprehensive information, rules, and goals into the big model, which then conducts autonomous analysis and judgment, ultimately outputting decision recommendations or conclusions.

[0065] Autonomous scheduling: Based on the analysis and judgment results of the large model, this module autonomously issues scheduling instructions to the intelligent agent collaboration module 130.

[0066] Decision or control output: Generates the final intelligent decision result or control instruction. Intelligent decision results or control instructions include: security risk level assessment, abnormal event warning, performance prediction analysis, resource allocation plan, operation optimization suggestion, emergency response plan, or at least one of the adaptive control parameter set.

[0067] The autonomous analysis, decision-making and control module is also used for autonomous analysis and judgment of engineering problems. Specifically, the autonomous analysis, decision-making and control module uses the spatiotemporal large model base to output the engineering knowledge background related to the engineering problem and the list of steps to solve the problem based on the engineering problem, and obtains relevant spatiotemporal situation data from the three-dimensional or four-dimensional integrated model of the engineering object to form complete knowledge, methods and data for problem analysis and solution. The spatiotemporal large model base conducts autonomous analysis and judgment, and calls the intelligent body collaboration module to execute, and outputs the decision results or control instructions of the engineering problem.

[0068] Furthermore, the intelligent management and control system receives feedback from the execution layer and integrates it into the integrated three-dimensional or four-dimensional model of the engineering object, triggering adaptive adjustments to the strategies of the spatiotemporal model base or the scheduling of autonomous analysis, decision-making, and control modules. Adaptive adjustments are used to dynamically update the spatiotemporal data volume and automatically call and execute functional intelligent agents based on the spatiotemporal model base to periodically or automatically generate intelligent decision results or control instructions.

[0069] Based on the large model and integrated multi-dimensional spatiotemporal data fusion intelligent management and control system proposed in this invention, refer to Figure 2 The method of applying the system shown includes: Step 201: Using the spatiotemporal large model base, deeply analyze the input engineering business requirements, combine the status and spatiotemporal characteristics of the complete spatiotemporal data volume in the current three-dimensional or four-dimensional integrated model of the engineering object, and automatically convert the engineering business requirements into a series of subtasks executed in the intelligent agent collaboration module; Step 202: The autonomous analysis, decision-making, and control module dynamically schedules the relevant functional agents in the agent collaboration module based on the subtasks and real-time status. The relevant functional agents utilize communication and collaboration mechanisms to exchange information and allocate tasks. When necessary, they obtain the processed, generated, or synthesized spatiotemporal data, analysis results, or generated content required to perform the tasks from the three-dimensional or four-dimensional integrated model of the engineering object to collaboratively perform the subtasks. Step 203: The autonomous analysis, decision-making, and control module obtains global spatiotemporal situation information from the three-dimensional or four-dimensional integrated model of the engineering object, combines the execution results and status feedback of the aggregated functional agents, and utilizes the spatiotemporal large model base's reasoning capabilities for the engineering domain and preset or dynamically generated rules and goals to conduct autonomous analysis and judgment, and based on this, generates intelligent decision results or control instructions. Step 204: Output the decision result or control instruction to the execution level, receive feedback on the execution effect or environmental change information, and use the feedback on the execution effect or environmental change information to update the three-dimensional or four-dimensional integrated model of the engineering object, and trigger adaptive adjustment of the strategy of the spatiotemporal large model base or the scheduling of the autonomous analysis decision and control module.

[0070] In summary, the intelligent management and control system of the present invention, based on a large model and integrated multi-dimensional spatiotemporal data fusion, aims to address the existing problems of weak spatiotemporal correlation processing capabilities, difficulty in fusing multi-source heterogeneous data, delayed business response, insufficient decision-making intelligence, and shallow integration of AI technology with engineering business when applying artificial intelligence in the engineering field. The present invention introduces a large spatiotemporal model optimized for the engineering field as the intelligent core, constructs a three-dimensional or four-dimensional integrated model of the engineering object with a multi-dimensional data model of all elements of the engineering object as the core, and through intelligent agent collaboration and autonomous decision-making closed loop, the intelligent management and control system provided by the present invention has one or more of the following beneficial effects compared to the existing technology: (1) Significantly enhance the spatiotemporal cognition and processing capabilities: With the engineering native full-factor multidimensional data model as the core, the system integrates dynamic data and manages it under a unified spatiotemporal benchmark. Combined with the understanding of spatiotemporal correlation and evolution patterns of the spatiotemporal large model, it greatly enhances the system's spatiotemporal perception and analysis capabilities for complex engineering environments and can effectively process multidimensional vector graphics information.

[0071] (2) Realize efficient data fusion and intelligent services: The three-dimensional or four-dimensional integrated model of the engineering object not only solves the problem of data islands, but also can respond to requests to actively process, generate and integrate new spatiotemporal data, analysis results or generative content (for example, risk assessment results after spatiotemporal analysis, equipment health index generated based on multi-source monitoring data, daily summary automatically generated based on templates and real-time data, preliminary optimization layout suggestions or design elements, etc.), providing a powerful and intelligent new data support paradigm for upper-level intelligent applications, which is different from the traditional data query mode.

[0072] (3) Improve business response speed and decision-making intelligence: The spatiotemporal big model can deeply understand business needs and quickly decompose and plan tasks. The autonomous decision-making module can make accurate judgments and optimize scheduling based on the global spatiotemporal situation and model reasoning, which significantly improves business response speed and the scientific and forward-looking nature of decision-making.

[0073] (4) Promote the deep integration of AI and engineering business: Closely combine the understanding and reasoning capabilities of the domain-adapted spatiotemporal large model with engineering spatiotemporal data and specific business processes (executed by intelligent agents), forming a closed loop from demand to execution to feedback, so that AI technology can be truly implemented and empower the core business of engineering.

[0074] (5) Achieve adaptive and continuous optimization management and control: Through the closed-loop feedback mechanism, the system can continuously adjust strategies and optimize models according to actual execution results and environmental changes, achieving adaptive and intelligent management and control of the entire life cycle of the project.

[0075] (6) Broad application prospects: The system and method can be flexibly applied to multiple engineering fields such as smart mines, smart cities, smart factories, smart transportation, smart water conservancy, smart energy, and smart environment, and has good versatility and scalability.

[0076] The intelligent management and control system proposed in the present invention, based on the fusion of large models and integrated multidimensional spatiotemporal data, builds a multidimensional spatiotemporal data platform with multidimensional geometry and attribute models as its core by constructing a large spatiotemporal model base optimized for the engineering field. This platform provides intelligent data services and information products, and uses intelligent agents to collaboratively execute tasks and autonomous decision-making modules to conduct command, dispatch, and closed-loop control. This effectively solves many of the pain points of the existing technology in applying AI large models in the engineering field, and can significantly improve the level of intelligent management and control throughout the entire life cycle of the project. The present invention can solve the problems of intelligent management and control in engineering fields such as mining, chemical plants, water conservancy projects, and construction, significantly improving the practicality of artificial intelligence systems in engineering fields and having high practicality.

[0077] Although the preferred embodiments of the present invention have been described, those skilled in the art may make additional changes and modifications to these embodiments once they become aware of the basic creative concepts. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the embodiments of the present invention.

[0078] Finally, it should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or terminal device that includes a series of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such process, method, article, or terminal device. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of additional identical elements in the process, method, article, or terminal device that includes the element.

[0079] The embodiments of the present invention are described above in conjunction with the accompanying drawings, but the present invention is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Under the guidance of the present invention, ordinary technicians in this field can also make many forms without departing from the scope of protection of the present invention and the claims, all of which are protected by the present invention.

Claims

1. An intelligent management and control system based on a large model and integrated multi-dimensional spatiotemporal data fusion, characterized by: include: A spatiotemporal large model base, configured to receive engineering-related input information and perform intelligent task planning and orchestration based on its deep understanding, knowledge association, and logical reasoning of data provided by its engineering domain expertise, the engineering-related input information, and the three-dimensional or four-dimensional integrated model of the engineering object. The input information includes at least business requirements and multimodal data. The three-dimensional or four-dimensional integrated model of the engineering object, based on the full element, three-dimensional or four-dimensional geometry and attribute data of the engineering object under a unified spatiotemporal reference, integrates the dynamic private domain data related to the engineering business, and constructs and dynamically updates a spatiotemporal data volume that reflects the true overall picture and life cycle of the project through integrated spatiotemporal expression and processing. This spatiotemporal data volume serves as the three-dimensional or four-dimensional integrated model of the engineering object, which is used to respond to data requests and, through processing, generation or synthesis, forms global spatiotemporal situation information; An agent collaboration module, comprising multiple functional agents and providing a communication and coordination mechanism between the functional agents, is used to receive task instructions corresponding to intelligent task planning and orchestration issued by the spatiotemporal large-scale model base, and, under the scheduling and monitoring of the autonomous analysis, decision-making and control module, obtain the data, analysis results or generated content required to execute the task from the three-dimensional or four-dimensional integrated model of the engineering object, thereby driving the execution of the functional agents; The autonomous analysis, decision-making and control module is used to obtain global spatiotemporal situation information from the three-dimensional or four-dimensional integrated model of the engineering object, and receive execution feedback results from the functional intelligent body collaboration module, and utilize the reasoning ability of the spatiotemporal large model base for the engineering field, combined with the global spatiotemporal situation information, the execution feedback results and preset or dynamically generated rules and goals, to perform autonomous analysis and judgment, and based on the results of the autonomous analysis and judgment, autonomously schedule the intelligent body collaboration module to perform subsequent tasks, and generate intelligent decision results or control instructions.

2. The intelligent management and control system according to claim 1, characterized in that: The spatiotemporal large model base is an intelligent model capable of deeply understanding and handling problems unique to the engineering field, wherein the ability to deeply understand and handle problems unique to the engineering field is obtained through an adaptation method; The adaptation method includes a combination of one or more of the following techniques: Domain fine-tuning technology, which refers to the use of large-scale text corpora, code corpora, and various public domain multimodal data in the engineering field to perform supervised fine-tuning, instruction fine-tuning, or continuous pre-training of intelligent models; Knowledge enhancement technology, which refers to: combining a spatiotemporal data volume constructed from a three-dimensional or four-dimensional integrated model of the engineering object or an external professional knowledge base, using retrieval-enhanced generation, knowledge graph embedding, or arbitrary knowledge injection, and extracting knowledge related to specific problems in the engineering field and providing it as additional background knowledge to the spatiotemporal large model base; Spatiotemporal capability enhancement technology, which involves training based on the spatiotemporal characteristics of engineering domains. Specifically, this includes designing tasks that incorporate spatiotemporal concepts, constructing datasets with precise spatiotemporal labels, and training on a large-scale base model to enhance the base model's ability to understand spatiotemporal queries, reason about spatiotemporal relationships, and predict spatiotemporal evolution patterns. Prompt engineering technology refers to designing optimized prompt word templates for engineering tasks, which contain clear instructions, context information, constraints, and expected output formats.

3. The intelligent management and control system according to claim 1, characterized in that: The three-dimensional or four-dimensional integrated model of the engineering object is a collection of advanced data and knowledge in the form of a knowledge graph or digital twin, built based on integrated engineering spatiotemporal data, containing rich semantic relationships and business information, and supporting intelligent reasoning and analysis. It is obtained through the following method: According to the project scenario, select relative time or absolute time as the time reference, select local coordinate system or geographic coordinate system as the spatial reference, and define a unified time and space benchmark; Under a unified spatiotemporal benchmark, data from different sources and formats are accessed, associated, and integrated to form a consistent, analyzable data set. These data from different sources and formats are divided into two core categories: the first is the data corresponding to the integrated 3D geometry and attributes model that serves as the core spatiotemporal foundation; the second is various dynamic data streams covering the entire project lifecycle. Based on the consistent and analyzable data set, and taking the geographic environment and physical engineering entities as the foundation, a spatial geometry and attribute model of the engineering project is established; based on the spatial model, the engineering object is used as the element unit, and at least one time series data including environmental monitoring and automated operation is integrated; knowledge engineering technology is used to explicitly express the engineering entities and their relationships, and a three-dimensional or four-dimensional integrated model of the engineering object is constructed and dynamically maintained; The dynamic private domain data related to engineering business include: monitoring and supervision, equipment automation data, and dynamic full life cycle data that reflects the true picture of the project.

4. The intelligent management and control system according to claim 3, characterized in that: The three-dimensional or four-dimensional integrated model of the engineering object responds to traditional query requests and directly returns the geometric data and attribute data contained in the three-dimensional or four-dimensional engineering object according to the requirements of the upper-level module; or The three-dimensional or four-dimensional integrated model of the engineering object extracts the geometric data or attribute data from the three-dimensional or four-dimensional integrated model of the engineering object in response to traditional query requests and based on the needs of the upper-level modules through the understanding and generation capabilities of the spatiotemporal large model base, wherein the extraction includes at least one of: object attribute query, local graphic output, and three-dimensional model sectioning; Based on the geometric data or attribute data, the spatiotemporal big model conducts comprehensive analysis or calls external tools for processing to output new generative content; the generative content refers to: output that goes beyond the original data or simple query results, is generated after intelligent processing, and has business value, and the type is at least one of text results, two-dimensional graphics, and three-dimensional models.

5. The intelligent management and control system according to claim 1, characterized in that: The agent collaboration module provides management, communication, and collaboration of agents; it ensures the normal operation of each functional agent through registration, discovery, and management; The plurality of functional agents interact with each other on data and status and execute the interaction through the communication channel provided by the agent collaboration module; If any functional agent encounters an abnormality during execution, based on the analysis and understanding of the spatiotemporal model base and the autonomous decision-making and control module, the agent collaboration module provides multiple rounds of task allocation execution and abnormal state resolution.

6. The intelligent management and control system according to claim 1, characterized in that: The autonomous analysis, decision-making and control module is also used for autonomous analysis and judgment of engineering problems, specifically including: The autonomous analysis, decision-making and control module uses the large space-time model base to output the engineering knowledge background related to the engineering problem and the list of steps to solve the problem based on the engineering problem, and obtains relevant space-time situation data from the three-dimensional or four-dimensional integrated model of the engineering object to form complete knowledge, methods and data for problem analysis and solution. The large space-time model base conducts autonomous analysis and judgment, and calls the intelligent body collaboration module to execute, and outputs the decision-making results or control instructions of the engineering problem.

7. The intelligent management and control system according to claim 1, characterized in that: The spatiotemporal large model base is also used to process and parse the multimodal data, which includes at least one of text, sound, image, video, vector graphics and structured data.

8. The intelligent management and control system according to claim 3, characterized in that: The specific implementation form of the three-dimensional or four-dimensional integrated model of the engineering object is: an engineering spatiotemporal knowledge graph or an engineering digital twin, which is a platform for integrating multi-source spatiotemporal data of the engineering project and organizing it into spatiotemporal data and knowledge with rich semantic relationships for reasoning and understanding by the spatiotemporal large model; The engineering spatiotemporal knowledge graph uses a graph structure to explicitly express engineering entities and the complex relationships between them, and contains and represents a large amount of business data; The engineering digital twin constructs a high-fidelity virtual model of the physical engineering entity, and uses fused dynamic data to achieve real-time or near real-time synchronization of the virtual model and the physical entity state, supporting simulation and interaction.

9. The intelligent management and control system according to claim 8, characterized in that: The full-element, multi-dimensional geometric and attribute data of the engineering object include but are not limited to: full-stratum geological model, building information model BIM, city information model CIM, or factory 3D model, infrastructure 3D model.

10. The intelligent management and control system according to claim 3, characterized in that: The various types of dynamic data streams covering the entire life cycle of the project include: personnel activity data, safety monitoring data, equipment operation status data, environmental monitoring data, construction progress data, material consumption data, cost data or at least one of production and operation management data.

11. The intelligent management and control system according to claim 1, characterized in that: The functional agents in the agent collaboration module include at least one of: a data acquisition and preprocessing agent, a domain model calculation agent, a risk identification and assessment agent, a resource scheduling optimization agent, a report and visualization generation agent, or a human-computer interaction interface agent; or The functional intelligent agent includes at least one combination of a spatiotemporal analysis intelligent agent, a process execution intelligent agent, a decision optimization intelligent agent and an industrial control intelligent agent.

12. The intelligent management and control system according to claim 11, characterized in that: The intelligent decision-making results or control instructions include: safety risk level assessment, abnormal event warning, performance prediction analysis, resource allocation plan, operation optimization suggestion, emergency response plan, or at least one of adaptive control parameter set.

13. The intelligent management and control system according to claim 12, characterized in that: The intelligent management and control system is further configured to receive feedback information from the execution layer, integrate the feedback information into the three-dimensional or four-dimensional integrated model of the engineering object for updating, and trigger adaptive adjustment of the strategy of the spatiotemporal large model base or the scheduling of the autonomous analysis, decision-making and control module; The adaptive adjustment is used to dynamically update the spatiotemporal data body, and automatically call and execute functional intelligent bodies based on the spatiotemporal large model base to regularly or automatically generate intelligent decision-making results or control instructions.

14. The intelligent management and control system according to claim 13, characterized in that: The application method of the intelligent management and control system includes the following steps: Step T1: Utilizing the spatiotemporal large model base, deeply analyzing the input engineering business requirements, combining the state and spatiotemporal characteristics of the complete spatiotemporal data volume in the current three-dimensional or four-dimensional integrated model of the engineering object, the engineering business requirements are automatically converted into a series of subtasks executed in the intelligent agent collaboration module; Step T2: The autonomous analysis, decision-making, and control module dynamically schedules the relevant functional agents in the agent collaboration module based on the subtasks and real-time status. The relevant functional agents utilize the communication and collaboration mechanism to exchange information and allocate tasks. When necessary, they obtain the processed, generated, or synthesized spatiotemporal data, analysis results, or generated content required to perform the tasks from the three-dimensional or four-dimensional integrated model of the engineering object, so as to collaboratively perform the subtasks. Step T3: The autonomous analysis, decision-making, and control module obtains the global spatiotemporal situation information from the three-dimensional or four-dimensional integrated model of the engineering object, combines the execution results and status feedback of the functional agents, and utilizes the reasoning capabilities of the spatiotemporal large model base for the engineering field and preset or dynamically generated rules and goals to conduct autonomous analysis and judgment, and based on this, generates intelligent decision results or control instructions; Step T4: Output the decision result or control instruction to the execution level, receive execution effect feedback or environmental change information, and use the execution effect feedback or environmental change information to update the three-dimensional or four-dimensional integrated model of the engineering object, and trigger adaptive adjustment of the strategy of the spatiotemporal large model base or the scheduling of the autonomous analysis decision and control module.

Citation Information

Patent Citations

  • Method and system for generating unique identifier of device

    CN114064732A

  • Mutual generation type artificial intelligence system based on multi-dimensional spatio-temporal information vector graph

    CN118484551A

  • Construction method of general artificial space-time intelligent large model

    CN118966161A

  • Method for performing engineering data governance and application by using large model knowledge base

    CN119884389A

  • Large-model digital twin system

    CN119885874A

Cited By

  • Equipment control method, device and equipment of beam-pumping unit and medium

    CN120990541A

  • Big reasoning model-based cooperative control method and system for intelligent agent with body

    CN121276986A

  • Intelligent management scheduling system and method based on multi-agent and retrieval enhancement generation

    CN121329086A

  • Intelligent management scheduling system and method based on multi-agent and retrieval enhanced generation

    CN121329086B

  • Industrial carrier space management information intelligent generation method and system based on multi-modal large model

    CN121882557A