Multi-level perception-control-computing power fusion model construction method and system for medium and low voltage distribution network

Abstract

The invention provides a method and a system for constructing a multi-level perception-control-computing power fusion model of a medium and low voltage power distribution network. The method comprises the following steps: constructing a perception layer knowledge graph model; constructing a control layer knowledge graph model; constructing a computing power layer knowledge graph model; and based on the constructed perception layer knowledge graph model, the constructed control layer knowledge graph model and the constructed computing power layer knowledge graph model, constructing a medium and low voltage distribution network perception-control-computing power fusion model based on the knowledge graph model. The model constructed by the invention realizes real-time monitoring, intelligent analysis and optimal control of the operation state of the power distribution network through cooperative work of the sensing layer, the control layer and the computing power layer, and can significantly improve the operation efficiency, reliability and adaptive ability of the power distribution network.

Description

Technical Field

[0001] The present invention relates to the field of medium and low voltage distribution networks, and specifically to a method and system for constructing a multi-level perception-control-computing power fusion model of medium and low voltage distribution networks. Background Art

[0002] With the rapid growth of global energy demand and the increasing severity of climate change, the sustainable development of energy systems has become a global focus. Against this background, the structure and operation mode of the power system are undergoing profound changes. Especially in the field of distribution networks, the traditional centralized power supply mode is gradually evolving towards distributed energy management and intelligent power grids. This transformation is not only aimed at improving energy efficiency, but also at enhancing the reliability, flexibility and adaptability of the power grid.

[0003] As an important part of the power system, the medium and low voltage distribution network connects the high voltage transmission network with the end users (including residential, commercial and industrial users). The stability of its operation directly affects the power quality and power supply security of users. The traditional distribution network mainly relies on centralized control, which is uniformly dispatched and managed by the central control system. However, with the complexity of the power system structure and the diversification of load demand, this centralized control method faces more and more challenges.

[0004] The introduction of distributed energy has further exacerbated the complexity of the distribution network. The widespread application of distributed energy such as photovoltaic power generation, wind power generation, electric vehicles and energy storage systems has made the power flow in the distribution network no longer unidirectional, but presents the characteristics of bidirectional flow. This bidirectional flow has increased the complexity of the operation of the power grid and increased the difficulty of grid dispatching. In addition, the intermittent and volatile characteristics of distributed energy have brought new challenges to the stable operation of the distribution network. Traditional dispatching and control methods are difficult to cope with this new complex situation.

[0005] In traditional distribution networks, a single-level control mode is mainly adopted, that is, the central control system is responsible for the dispatching and management of the entire distribution network. The main feature of this mode is that it relies on a centralized monitoring and decision-making system, and all monitoring data and control commands must be processed and issued by the central control system. However, with the expansion of the scale of distribution networks and the complexity of the operating environment, this single-level control mode has gradually exposed some drawbacks, facing challenges such as slow response speed, low resource utilization efficiency, lack of adaptive capabilities, and inability to meet the needs of distributed energy access.

[0006] In order to meet these challenges, modern distribution networks need a more intelligent management model, which should be able to make full use of advanced information and communication technologies, artificial intelligence technologies, and big data technologies to achieve refined management and intelligent dispatch of various resources in the distribution network. Summary of the invention

[0007] The present invention provides a method and system for constructing a multi-level perception-control-computing power fusion model of a medium and low voltage distribution network. The constructed model realizes real-time monitoring, intelligent analysis and optimization control of the operating status of the distribution network through the collaborative work of the perception layer, control layer and computing power layer, and can significantly improve the operating efficiency, reliability and adaptability of the distribution network.

[0008] A method for constructing a multi-level perception-control-computing power fusion model for a medium and low voltage distribution network, comprising:

[0009] Build a perception layer knowledge graph model;

[0010] Build a control layer knowledge graph model;

[0011] Build a knowledge graph model for the computing power layer;

[0012] Based on the constructed perception layer knowledge graph model, control layer knowledge graph model and computing power layer knowledge graph model, a perception-control-computing power fusion model of medium and low voltage distribution networks based on the knowledge graph model is constructed.

[0013] Furthermore, the construction of the perception layer knowledge graph model includes:

[0014] Collecting multi-source data: The perception layer collects data from each node in the distribution network in real time through distributed sensors, including electrical parameters and environmental parameters;

[0015] Data semantic processing: By semantically annotating multi-source data, the relationship between data and devices is established. The semantic information of the data is associated with the relevant entities in the knowledge graph;

[0016] Data preprocessing: The perception layer preprocesses the collected multi-source data, including data cleaning, outlier detection, and data fusion. It performs association analysis and consistency check on the multi-source data by calling the rules and relationships stored in the knowledge graph.

[0017] Real-time data transmission: The preprocessed data is transmitted to the control layer and computing power layer in real time through the communication network.

[0018] Furthermore, the construction of the control layer knowledge graph model includes:

[0019] Data feature extraction: The control layer receives data from the perception layer through feature extraction and dimensionality reduction technology to extract representative and relevant features from high-dimensional data;

[0020] Control strategy formulation and control strategy library establishment: build control strategies based on knowledge graphs

[0021] The control strategy library stores various control strategies and rules of the distribution network, including load regulation, reactive power compensation, and fault handling. The rules in the control strategy library are formed by combining expert experience and operation data and are managed and updated through knowledge graphs;

[0022] Decision execution and feedback: After receiving the data from the perception layer, the control layer calls the rule base in the knowledge graph and uses the inference engine to analyze the current state of the distribution network. Based on the analysis results, it automatically generates control strategies and optimizes and adjusts the strategies. The control layer converts the generated control strategies into control instructions and realizes the actual control of the distribution network through the actuators. At the same time, the control layer monitors the execution results in real time, feeds back the monitoring data to the knowledge graph, and updates the rule base and control strategy base.

[0023] Dynamic adjustment and adaptive control: The control layer uses the dynamic semantic analysis capability provided by the knowledge graph to adaptively adjust emergencies that occur during the operation of the distribution network.

[0024] Furthermore, the construction of the computing power layer knowledge graph model includes:

[0025] Data storage and management: Build a large-scale data storage system at the computing layer to manage the real-time and historical data transmitted by the perception layer;

[0026] Scheduling and optimization of computing resources: Evaluate the computing power of each task, obtain the optimization order of tasks, and schedule and optimize computing resources according to task requirements;

[0027] Collaboration between edge computing and cloud computing: Tasks are processed in layers according to the real-time requirements and complexity of computing tasks. Tasks with high real-time requirements and small computational workloads are completed in edge computing units, while tasks with high complexity and requiring large-scale computing are handled by cloud computing.

[0028] Furthermore, based on the constructed perception layer knowledge graph model, control layer knowledge graph model and computing power layer knowledge graph model, a medium and low voltage distribution network perception-control-computing power fusion model based on the knowledge graph model is constructed, including:

[0029] Construction of a unified knowledge base: Integrate the data and rules of the perception layer, control layer, and computing layer to build a unified knowledge graph. The knowledge graph includes the equipment entities, operating status, control strategies, historical data, and fault models of the distribution network. Each layer can share data, rules, and strategies by calling the unified knowledge graph.

[0030] Cross-level knowledge sharing and collaboration: In a multi-level fusion model, each level shares and collaborates through knowledge graphs. The perception layer obtains the standard operating parameters of the equipment through the knowledge graph, the control layer performs rule reasoning based on the knowledge graph, and the computing power layer uses the knowledge graph to optimize the allocation of computing resources.

[0031] Knowledge graph-driven system optimization and evolution: The data of the perception layer is transmitted to the control layer through the knowledge graph, and the decision rules of the control layer are fed back to the computing layer for optimization. The analysis results of the computing layer are updated to the knowledge graph for reference by other layers. The control layer continuously optimizes the control strategy by analyzing the historical data and rules in the knowledge graph, and the computing layer improves the data processing and calculation methods based on the new knowledge in the knowledge graph.

[0032] Update and maintenance of dynamic knowledge graphs: When new devices are connected to the system or the status of existing devices changes, the perception layer updates the relevant information to the knowledge graph, allowing the control layer and computing layer to make decisions based on the latest knowledge.

[0033] A multi-level perception-control-computing power fusion model construction system for medium and low voltage distribution networks, comprising:

[0034] The perception layer knowledge graph model construction module is used to construct the perception layer knowledge graph model;

[0035] A control layer knowledge graph model construction module is used to construct a control layer knowledge graph model;

[0036] The computing power layer knowledge graph model construction module is used to build the computing power layer knowledge graph model;

[0037] The fusion model construction module is used to build a perception-control-computing power fusion model of medium and low voltage distribution networks based on the constructed perception layer knowledge graph model, control layer knowledge graph model and computing power layer knowledge graph model.

[0038] Furthermore, the perception layer knowledge graph model construction module is specifically used to:

[0039] Collecting multi-source data: The perception layer collects data from each node in the distribution network in real time through distributed sensors, including electrical parameters and environmental parameters;

[0040] Data semantic processing: By semantically annotating multi-source data, the relationship between data and devices is established. The semantic information of the data is associated with the relevant entities in the knowledge graph;

[0041] Data preprocessing: The perception layer preprocesses the collected multi-source data, including data cleaning, outlier detection, and data fusion. It performs association analysis and consistency check on the multi-source data by calling the rules and relationships stored in the knowledge graph.

[0042] Real-time data transmission: The preprocessed data is transmitted to the control layer and computing power layer in real time through the communication network.

[0043] Furthermore, the control layer knowledge graph model construction module is specifically used to:

[0044] Data feature extraction: The control layer receives data from the perception layer through feature extraction and dimensionality reduction technology to extract representative and relevant features from high-dimensional data;

[0045] Control strategy formulation and control strategy library establishment: build control strategies based on knowledge graphs

[0046] The control strategy library stores various control strategies and rules of the distribution network, including load regulation, reactive power compensation, and fault handling. The rules in the control strategy library are formed by combining expert experience and operation data and are managed and updated through knowledge graphs;

[0047] Decision execution and feedback: After receiving the data from the perception layer, the control layer calls the rule base in the knowledge graph and uses the inference engine to analyze the current state of the distribution network. Based on the analysis results, it automatically generates control strategies and optimizes and adjusts the strategies. The control layer converts the generated control strategies into control instructions and realizes the actual control of the distribution network through the actuators. At the same time, the control layer monitors the execution results in real time, feeds back the monitoring data to the knowledge graph, and updates the rule base and control strategy base.

[0048] Dynamic adjustment and adaptive control: The control layer uses the dynamic semantic analysis capability provided by the knowledge graph to adaptively adjust emergencies that occur during the operation of the distribution network.

[0049] Furthermore, the computing power layer knowledge graph model construction module is specifically used to:

[0050] Data storage and management: Build a large-scale data storage system at the computing layer to manage the real-time and historical data transmitted by the perception layer;

[0051] Scheduling and optimization of computing resources: Evaluate the computing power of each task, obtain the optimization order of tasks, and schedule and optimize computing resources according to task requirements;

[0052] Collaboration between edge computing and cloud computing: Tasks are processed in layers according to the real-time requirements and complexity of computing tasks. Tasks with high real-time requirements and small computational workloads are completed in edge computing units, while tasks with high complexity and requiring large-scale computing are handled by cloud computing.

[0053] Furthermore, the fusion model construction module is specifically used for:

[0054] Construction of a unified knowledge base: Integrate the data and rules of the perception layer, control layer, and computing layer to build a unified knowledge graph. The knowledge graph includes the equipment entities, operating status, control strategies, historical data, and fault models of the distribution network. Each layer can share data, rules, and strategies by calling the unified knowledge graph.

[0055] Cross-level knowledge sharing and collaboration: In a multi-level fusion model, each level shares and collaborates through knowledge graphs. The perception layer obtains the standard operating parameters of the equipment through the knowledge graph, the control layer performs rule reasoning based on the knowledge graph, and the computing power layer uses the knowledge graph to optimize the allocation of computing resources.

[0056] Knowledge graph-driven system optimization and evolution: The data of the perception layer is transmitted to the control layer through the knowledge graph, and the decision rules of the control layer are fed back to the computing layer for optimization. The analysis results of the computing layer are updated to the knowledge graph for reference by other layers. The control layer continuously optimizes the control strategy by analyzing the historical data and rules in the knowledge graph, and the computing layer improves the data processing and calculation methods based on the new knowledge in the knowledge graph.

[0057] Update and maintenance of dynamic knowledge graph: When new devices are connected to the system or the status of existing devices changes,

[0058] When the state changes, the perception layer updates the relevant information to the knowledge graph, allowing the control layer and computing power layer to make decisions based on the latest knowledge.

[0059] The present invention has the following effects:

[0060] The model constructed by the present invention realizes real-time monitoring, intelligent analysis and optimization control of the operation status of the distribution network through the organic combination of the perception layer, control layer and computing power layer. At the same time, combined with the knowledge graph technology, the data of the perception layer, control layer and computing power layer are calculated, and the control layer is omitted, and the computing power of each part of the distribution network is calculated, so as to realize the visualization of the distribution network status information and the intelligent and refined management of the distribution network. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] Figure 1 It is a schematic diagram of the multi-level perception-control-computing power fusion model of the medium and low voltage distribution network constructed by the present invention. DETAILED DESCRIPTION

[0062] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0063] See also Figure 1 The embodiment of the present invention provides a method for constructing a multi-level perception-control-computing power fusion model of a medium and low voltage distribution network. The specific implementation steps are as follows:

[0064] Step 1: Build a perception layer knowledge graph model

[0065] The perception layer is the basic layer in the multi-level perception-control-computing power fusion model. It is responsible for real-time data collection, monitoring and preliminary processing of various source and load points (power nodes and load nodes) in the distribution network, and provides accurate and reliable data information for the upper control layer and computing power layer.

[0066] The perception layer is the basic data source of the entire system. It consists of a distributed sensor network, a data transmission module, a data preprocessing module, etc. The main function of the perception layer is to achieve all-round monitoring of the physical state (such as voltage, current, power, frequency, etc.) and environmental state (such as temperature, humidity, etc.) of each node in the distribution network, and reliably transmit this information to the upper layer for analysis and decision-making. Step 1 specifically includes:

[0067] Step 1.1, collect multi-source data: The perception layer collects data from each node in the distribution network in real time through a large number of distributed sensors, including electrical parameters and environmental parameters. According to the needs of distribution network monitoring, the sensors in the perception layer mainly include electrical sensors (such as voltage sensors, current sensors, power sensors, etc.) and environmental sensors (such as temperature sensors, humidity sensors, vibration sensors, etc.). In terms of sensor layout, it is necessary to reasonably select the installation location and number of sensors according to the actual situation of the distribution network. For example, electrical sensors are deployed at transformers and distribution boxes, and environmental sensors are deployed in outdoor environments. In areas with high load density, the density of sensors is also increased accordingly to ensure the comprehensiveness of the data. The accuracy and reliability of sensors directly affect the quality of perception layer data. Therefore, regular sensor calibration and maintenance are essential. This requires professional calibration equipment and maintenance teams to regularly check and calibrate sensors to ensure the accuracy of their measurement data.

[0068] Step 1.2, data semantic processing: Use the knowledge graph to semantically process the data collected by the perception layer. First, establish the association between data and equipment by semantically annotating multi-source data. Second, associate the semantic information of the data with relevant entities in the knowledge graph (such as equipment, operating status, environmental parameters, etc.).

[0069] Step 1.3, data preprocessing: The perception layer preprocesses the collected multi-source data, including data cleaning, outlier detection and data fusion. By calling the rules and relationships stored in the knowledge graph, the data is analyzed for association and consistency to ensure the accuracy and integrity of the data. At the same time, edge computing units can be deployed at key nodes of the distribution network, such as substations and distribution rooms. These units are equipped with high-performance computing processors and storage devices, and can process data from various sensors in real time, such as data cleaning, abnormal data marking, etc. At the same time, edge computing can also perform simple data analysis, such as trend prediction, anomaly detection, etc., to provide early warning information for the upper system.

[0070] The specific data fusion method is to unify and fuse the pre-processed PMU, SCADA, and AMI signals in time scale, virtually collect data such as phase angle that the SCADA system cannot collect, and estimate the current state of the distribution network and predict the future state. Through state estimation and prediction, the control layer can identify potential risks in advance and formulate corresponding control strategies to avoid system abnormalities.

[0071] Step 1.4, real-time data transmission: The preprocessed data is transmitted to the control layer and computing layer in real time through the communication network. During the data transmission process, the perception layer dynamically adjusts the data according to the priority rules in the knowledge graph to ensure that key data is transmitted first.

[0072] Step 2: Build a control layer knowledge graph model

[0073] The control layer is the core part of the multi-level perception-control-computing power fusion model. It is responsible for receiving data from the perception layer, conducting in-depth analysis and processing, interacting with the computing power layer, allocating system computing power and executing corresponding control decisions. The main task of the control layer is to achieve real-time monitoring, optimal scheduling and emergency response of the distribution network, so as to ensure the stable and efficient operation of the distribution network. The construction of the control layer involves data analysis, control strategy formulation, decision execution, interaction with the computing power layer and other aspects. Step 2 specifically includes:

[0074] Step 2.1, data feature extraction: The control layer studies the data features of the large amount of data received from the perception layer. The control layer extracts representative and relevant features from high-dimensional data through feature extraction and dimensionality reduction techniques, such as principal component analysis (PCA) and feature selection algorithms, to simplify the data structure and improve computing efficiency.

[0075] Step 2.2, control strategy formulation and control strategy library establishment: The formulation of control strategies is the core task of the control layer, which is related to the overall operating efficiency and stability of the distribution network. Based on the knowledge graph, a control strategy library is constructed to store various control strategies and rules of the distribution network, including load regulation, reactive power compensation, fault handling, etc. The rules in the control strategy library are formed by combining expert experience and operation data, and are managed and updated through the knowledge graph.

[0076] The key to control strategy formulation is to comprehensively consider multiple factors such as the operating status, load demand, and power supply capacity of the distribution network to design the optimal control solution. In order to cope with the complex and changeable operating environment, the control layer can introduce optimization algorithms (such as genetic algorithms, particle swarm optimization, linear programming, etc.) to formulate control strategies. The optimization algorithm solves multiple objective functions (such as minimizing power loss and maximizing reliability) to find the optimal control solution and improve the overall performance of the system. At the same time, the control layer can use machine learning and deep learning models to formulate more intelligent control strategies. For example, through the reinforcement learning model, the control system can automatically learn the optimal control strategy in continuous interaction with the environment; through deep neural networks, the system can achieve modeling and real-time decision-making of complex nonlinear relationships. The operating environment of the distribution network changes dynamically, and the load demand and power supply capacity may fluctuate at any time. Therefore, the control strategy needs to have the ability to adjust dynamically. Adaptive control technology can automatically adjust control parameters and strategies based on real-time data to ensure that the system can maintain the best state under different operating conditions.

[0077] Step 2.3, decision execution and feedback: After receiving the data from the perception layer, the control layer calls the rule base in the knowledge graph and uses the inference engine to analyze the current state of the distribution network. Based on the analysis results, the control strategy is automatically generated and optimized. The control layer converts the generated control strategy into specific control instructions and implements actual control of the distribution network through the actuator (such as adjusting switches, adjusting load distribution, etc.). At the same time, the control layer monitors the execution results in real time, and feeds back the monitoring data to the knowledge graph to update the rule base and control strategy base.

[0078] The strategy formulated by the control layer needs to be implemented through the decision-making execution module, and the feedback of the system operation results is the key to further optimize the control strategy. The decision-making execution module converts the control strategy into specific control instructions and sends them to each actuator through the communication network. The generation of control instructions needs to consider the real-time performance of the system and the accuracy of the instructions to avoid system failures caused by delays in instruction execution or misoperation. After receiving the control instructions, each actuator (such as switchgear, voltage regulating equipment, etc.) needs to perform the corresponding operations quickly and accurately. The control layer needs to monitor the operating status of the actuator in real time through the perception layer to ensure the smooth execution of the control operation. After the actuator executes the control instruction, the control layer needs to obtain the execution results through the feedback mechanism and evaluate the control effect. If the control effect is not good or an abnormal situation occurs, the control layer needs to adjust the strategy in time or recalculate the optimization plan to ensure the stable operation of the system.

[0079] Step 2.4, Dynamic Adjustment and Adaptive Control: The control layer uses the dynamic semantic analysis capabilities provided by the knowledge graph to adaptively adjust emergencies that occur during the operation of the distribution network. For example, when the load increases sharply, the control layer automatically infers the optimal load adjustment plan and quickly implements it to ensure the stable operation of the power grid.

[0080] Step 3: Build a knowledge graph model for the computing power layer

[0081] The computing power layer is a key component of the multi-level perception-control-computing power fusion model, providing the high-performance computing resources, complex algorithm support and big data processing capabilities required by the system. The main task of the computing power layer is to conduct in-depth analysis, model training and optimization of the data collected by the perception layer, and support decision-making and execution of the control layer. It is not just a simple stacking of computing resources, but also involves multiple technical links such as cloud computing, edge computing, big data processing, artificial intelligence algorithms, and real-time assurance. Its overall architecture includes the following modules: data storage and management module, computing resource scheduling module, artificial intelligence and algorithm module, edge computing and cloud computing collaboration module, and interface module with the control layer and perception layer. Step 3 includes:

[0082] Step 3.1, Data storage and management: Build a large-scale data storage system at the computing layer to manage the real-time and historical data transmitted by the perception layer. The data storage system uses the knowledge graph to perform semantic management of data, achieving efficient organization and rapid retrieval of data.

[0083] The data storage and management of the computing layer is the basis for ensuring the efficient operation of the system. The module is responsible for managing large-scale, multi-source heterogeneous data from the perception layer, including real-time data, historical data, structured and unstructured data. The data storage and management module needs to have efficient data storage capabilities, fast data retrieval functions, and support data preprocessing and cleaning. Due to the large amount of data and its wide distribution, the computing layer needs to adopt a distributed data storage system. The distributed storage system can store data on multiple nodes, improving the reliability of data storage and retrieval efficiency. The computing layer needs to process the real-time data stream from the perception layer, which can realize efficient processing and analysis of real-time data and provide timely decision support for the control layer.

[0084] Step 3.2, Scheduling and optimization of computing resources: The computing tasks of the computing power layer are complex and diverse, and the demand for computing resources is changeable, so the scheduling and optimization of computing resources is crucial. This model evaluates the computing power of each task, obtains the optimization order of tasks, and implements the scheduling and optimization of computing resources according to task requirements. The computing power layer optimizes the allocation of computing resources based on the relationship between computing resources and tasks recorded in the knowledge graph. The knowledge graph is used to infer the dependencies between computing tasks, optimize task scheduling, and reduce idle computing resources and computing delays.

[0085] Step 3.3, Collaboration between edge computing and cloud computing: The computing resources of the computing power layer are distributed on the local side, the cloud side, and the edge side, and the coordinated optimization of the three sides needs to be considered. The computing power layer processes tasks in layers according to the real-time requirements and complexity of the computing tasks. Tasks with high real-time requirements and small computational workloads can be completed in edge computing units, while tasks with high complexity and requiring large-scale computing are handled by cloud computing. Edge computing units are located at the edge nodes of the distribution network and are easily affected by environmental factors. Therefore, the computing power layer needs to build a fault-tolerant mechanism for edge computing, and ensure the reliability and stability of edge computing through redundant computing, task migration, and other means.

[0086] Step 4: Construct a medium and low voltage distribution network perception-control-computing power fusion model based on the knowledge graph model

[0087] The core of the multi-level perception-control-computing power fusion model is to organically integrate the perception layer, control layer, and computing power layer to form an efficient and intelligent distribution network management system. Seamless collaboration is required between the various layers to improve the overall operating efficiency and safety of the distribution network through real-time data transmission, analysis, and decision-making. The following will elaborate on how to integrate the perception layer, control layer, and computing power layer into a multi-level fusion model and ensure its efficient operation. This section uses the knowledge graph modeling method to build a perception-control-computing power fusion model for medium and low voltage distribution networks.

[0088] The knowledge graph is a graph structure that represents entities and their relationships through nodes and edges, where nodes represent entities or concepts and edges represent the relationships between nodes. The knowledge graph can store and manage large-scale knowledge data in a structured manner and support complex reasoning and queries. In the application of distribution networks, the knowledge graph can be used to represent various entities (such as power sources, loads, equipment, etc.) and their relationships (such as power supply relationships, control relationships, etc.) in the power system, and perform intelligent analysis and decision support on this basis. The specific implementation steps of step 4 include:

[0089] Step 4.1: Construction of a unified knowledge base: Integrate the data and rules of the perception layer, control layer, and computing layer to build a unified knowledge graph. The knowledge graph includes the equipment entities, operating status, control strategies, historical data, fault models, etc. of the distribution network. Each level can share data, rules, and strategies by calling the unified knowledge graph.

[0090] Step 4.2, cross-level knowledge sharing and collaboration: In a multi-level fusion model, each level needs to share and collaborate through knowledge graphs. For example, the perception layer can obtain the standard operating parameters of the equipment through the knowledge graph, the control layer can perform rule reasoning based on the knowledge graph, and the computing layer can use the knowledge graph to optimize the allocation of computing resources. The computing layer mainly provides powerful computing resources and support for complex algorithms, while the control layer is responsible for converting the calculation results of the computing layer into actual control operations. When formulating control strategies, the control layer may need to call the high-performance computing resources of the computing layer, such as performing in-depth analysis of large amounts of data or solving complex optimization problems. The control layer needs to have the ability to allocate tasks and call computing resources to ensure that computing tasks can be completed efficiently. After the computing layer completes the computing task, it feeds the results back to the control layer. The control layer needs to update the control strategy or adjust the control parameters based on these results to ensure that the system runs in the optimal state. At the same time, the control layer can also feed back the actual operation data to the computing layer for further optimization of the algorithm and training of the model. In some complex scenarios, the control layer and the computing layer need to be optimized collaboratively. For example, in the scenario of distributed power access, the computing layer can be responsible for global optimization, while the control layer is responsible for local adjustments. The two can ensure the improvement of the overall performance of the system through joint control. The various levels are closely connected through knowledge graphs to achieve the coordination of data, knowledge and decision-making.

[0091] Step 4.3, knowledge graph-driven system optimization and evolution (knowledge reasoning): In the multi-level fusion model, the perception layer, control layer, and computing layer achieve knowledge sharing and collaborative work through a unified knowledge graph. For example, the data of the perception layer can be passed to the control layer through the knowledge graph, the decision rules of the control layer can be fed back to the computing layer for optimization, and the analysis results of the computing layer can also be updated to the knowledge graph for reference by other layers.

[0092] As the system runs and data accumulates, the knowledge graph can be continuously updated and expanded to help the system optimize and evolve. For example, the control layer can continuously optimize the control strategy by analyzing the historical data and rules in the knowledge graph; the computing layer can improve data processing and calculation methods based on the new knowledge in the knowledge graph.

[0093] Step 4.4, Update and maintain the dynamic knowledge graph (knowledge fusion, new data knowledge extraction): In the multi-level fusion model, the knowledge graph needs to be dynamically updated and maintained to adapt to changes in the distribution network. For example, when a new device is connected to the system or the status of an existing device changes, the perception layer can update the relevant information to the knowledge graph, so that the control layer and the computing layer can make decisions based on the latest knowledge.

[0094] The embodiment of the present invention also provides a multi-level perception-control-computing power fusion model construction system for medium and low voltage distribution networks, including:

[0095] A perception layer knowledge graph model construction module is used to construct a perception layer knowledge graph model; a control layer knowledge graph model construction module is used to construct a control layer knowledge graph model;

[0096] The computing power layer knowledge graph model construction module is used to build the computing power layer knowledge graph model;

[0097] The fusion model construction module is used to build a perception-control-computing power fusion model of medium and low voltage distribution networks based on the constructed perception layer knowledge graph model, control layer knowledge graph model and computing power layer knowledge graph model.

[0098] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed by the present invention should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A method for constructing a multi-level perception-control-computing power fusion model for medium and low voltage distribution networks, characterized in that: The steps include: Build a perception layer knowledge graph model; Build a control layer knowledge graph model; Build a knowledge graph model for the computing power layer; Based on the constructed perception layer knowledge graph model, control layer knowledge graph model and computing power layer knowledge graph model, a perception-control-computing power fusion model of medium and low voltage distribution networks based on the knowledge graph model is constructed.

2. The method for constructing a multi-level perception-control-computing power fusion model for a medium and low voltage distribution network according to claim 1, characterized in that: The construction of the perception layer knowledge graph model includes: Collecting multi-source data: The perception layer collects data from each node in the distribution network in real time through distributed sensors, including electrical parameters and environmental parameters; Data semantic processing: By semantically annotating multi-source data, the relationship between data and devices is established. The semantic information of the data is associated with the relevant entities in the knowledge graph; Data preprocessing: The perception layer preprocesses the collected multi-source data, including data cleaning, outlier detection, and data fusion. It performs association analysis and consistency check on the multi-source data by calling the rules and relationships stored in the knowledge graph. Real-time data transmission: The preprocessed data is transmitted to the control layer and computing power layer in real time through the communication network.

3. The method for constructing a multi-level perception-control-computing power fusion model for a medium and low voltage distribution network according to claim 1, characterized in that: The construction of the control layer knowledge graph model includes: Data feature extraction: The control layer receives data from the perception layer through feature extraction and dimensionality reduction technology to extract representative and relevant features from high-dimensional data; Control strategy formulation and control strategy library establishment: Based on the knowledge graph, a control strategy library is constructed to store various control strategies and rules of the distribution network, including load regulation, reactive power compensation, and fault handling. The rules in the control strategy library are formed by combining expert experience and operation data, and are managed and updated through the knowledge graph; Decision execution and feedback: After receiving the data from the perception layer, the control layer calls the rule base in the knowledge graph and uses the inference engine to analyze the current state of the distribution network. Based on the analysis results, it automatically generates control strategies and optimizes and adjusts the strategies. The control layer converts the generated control strategies into control instructions and realizes the actual control of the distribution network through the actuators. At the same time, the control layer monitors the execution results in real time, feeds back the monitoring data to the knowledge graph, and updates the rule base and control strategy base. Dynamic adjustment and adaptive control: The control layer uses the dynamic semantic analysis capability provided by the knowledge graph to adaptively adjust emergencies that occur during the operation of the distribution network.

4. The method for constructing a multi-level perception-control-computing power fusion model for a medium and low voltage distribution network according to claim 1, characterized in that: The construction of the computing power layer knowledge graph model includes: Data storage and management: Build a large-scale data storage system at the computing layer to manage the real-time and historical data transmitted by the perception layer; Scheduling and optimization of computing resources: Evaluate the computing power of each task, obtain the optimization order of tasks, and schedule and optimize computing resources according to task requirements; Collaboration between edge computing and cloud computing: Tasks are processed in layers according to the real-time requirements and complexity of computing tasks. Tasks with high real-time requirements and small computational workloads are completed in edge computing units, while tasks with high complexity and requiring large-scale computing are handled by cloud computing.

5. The method for constructing a multi-level perception-control-computing power fusion model for a medium and low voltage distribution network according to claim 1, characterized in that: Based on the constructed perception layer knowledge graph model, control layer knowledge graph model and computing power layer knowledge graph model, a medium and low voltage distribution network perception-control-computing power fusion model based on the knowledge graph model is constructed, including: Construction of a unified knowledge base: Integrate the data and rules of the perception layer, control layer, and computing layer to build a unified knowledge graph. The knowledge graph includes the equipment entities, operating status, control strategies, historical data, and fault models of the distribution network. Each layer can share data, rules, and strategies by calling the unified knowledge graph. Cross-level knowledge sharing and collaboration: In a multi-level fusion model, each level shares and collaborates through knowledge graphs. The perception layer obtains the standard operating parameters of the equipment through the knowledge graph, the control layer performs rule reasoning based on the knowledge graph, and the computing power layer uses the knowledge graph to optimize the allocation of computing resources. Knowledge graph-driven system optimization and evolution: The data of the perception layer is transmitted to the control layer through the knowledge graph, and the decision rules of the control layer are fed back to the computing layer for optimization. The analysis results of the computing layer are updated to the knowledge graph for reference by other layers. The control layer continuously optimizes the control strategy by analyzing the historical data and rules in the knowledge graph, and the computing layer improves the data processing and calculation methods based on the new knowledge in the knowledge graph. Update and maintenance of dynamic knowledge graphs: When new devices are connected to the system or the status of existing devices changes, the perception layer updates the relevant information to the knowledge graph, allowing the control layer and computing layer to make decisions based on the latest knowledge.

6. A multi-level perception-control-computing power fusion model construction system for medium and low voltage distribution networks, characterized in that: include: The perception layer knowledge graph model construction module is used to construct the perception layer knowledge graph model; A control layer knowledge graph model construction module is used to construct a control layer knowledge graph model; The computing power layer knowledge graph model construction module is used to build the computing power layer knowledge graph model; The fusion model construction module is used to build a perception-control-computing power fusion model of medium and low voltage distribution networks based on the constructed perception layer knowledge graph model, control layer knowledge graph model and computing power layer knowledge graph model.

7. The multi-level perception-control-computing power fusion model construction system for medium and low voltage distribution networks according to claim 6, characterized in that: The perception layer knowledge graph model construction module is specifically used for: Collecting multi-source data: The perception layer collects data from each node in the distribution network in real time through distributed sensors, including electrical parameters and environmental parameters; Data semantic processing: By semantically annotating multi-source data, the relationship between data and devices is established. The semantic information of the data is associated with the relevant entities in the knowledge graph; Data preprocessing: The perception layer preprocesses the collected multi-source data, including data cleaning, outlier detection, and data fusion. It performs association analysis and consistency check on the multi-source data by calling the rules and relationships stored in the knowledge graph. Real-time data transmission: The preprocessed data is transmitted to the control layer and computing power layer in real time through the communication network.

8. The multi-level perception-control-computing power fusion model construction system for medium and low voltage distribution networks according to claim 6, characterized in that: The control layer knowledge graph model construction module is specifically used for: Data feature extraction: The control layer receives data from the perception layer through feature extraction and dimensionality reduction technology to extract representative and relevant features from high-dimensional data; Control strategy formulation and control strategy library establishment: Based on the knowledge graph, a control strategy library is constructed to store various control strategies and rules of the distribution network, including load regulation, reactive power compensation, and fault handling. The rules in the control strategy library are formed by combining expert experience and operation data, and are managed and updated through the knowledge graph; Decision execution and feedback: After receiving the data from the perception layer, the control layer calls the rule base in the knowledge graph and uses the inference engine to analyze the current state of the distribution network. Based on the analysis results, it automatically generates control strategies and optimizes and adjusts the strategies. The control layer converts the generated control strategies into control instructions and realizes the actual control of the distribution network through the actuators. At the same time, the control layer monitors the execution results in real time, feeds back the monitoring data to the knowledge graph, and updates the rule base and control strategy base. Dynamic adjustment and adaptive control: The control layer uses the dynamic semantic analysis capability provided by the knowledge graph to adaptively adjust emergencies that occur during the operation of the distribution network.

9. The multi-level perception-control-computing power fusion model construction system for medium and low voltage distribution networks according to claim 6, characterized in that: The computing power layer knowledge graph model construction module is specifically used to: Data storage and management: Build a large-scale data storage system at the computing layer to manage the real-time and historical data transmitted by the perception layer; Scheduling and optimization of computing resources: Evaluate the computing power of each task, obtain the optimization order of tasks, and schedule and optimize computing resources according to task requirements; Collaboration between edge computing and cloud computing: Tasks are processed in layers according to the real-time requirements and complexity of computing tasks. Tasks with high real-time requirements and small computational workloads are completed in edge computing units, while tasks with high complexity and requiring large-scale computing are handled by cloud computing.

10. The multi-level perception-control-computing power fusion model construction system for medium and low voltage distribution networks according to claim 6, characterized in that: The fusion model construction module is specifically used for: Construction of a unified knowledge base: Integrate the data and rules of the perception layer, control layer, and computing layer to build a unified knowledge graph. The knowledge graph includes the equipment entities, operating status, control strategies, historical data, and fault models of the distribution network. Each layer can share data, rules, and strategies by calling the unified knowledge graph. Cross-level knowledge sharing and collaboration: In a multi-level fusion model, each level shares and collaborates through knowledge graphs. The perception layer obtains the standard operating parameters of the equipment through the knowledge graph, the control layer performs rule reasoning based on the knowledge graph, and the computing power layer uses the knowledge graph to optimize the allocation of computing resources. Knowledge graph-driven system optimization and evolution: Data from the perception layer is passed to the control layer through the knowledge graph, and the decision rules of the control layer are fed back to the computing layer for optimization. The analysis results of the computing layer are updated to the knowledge graph for reference by other layers. The control layer continuously optimizes the control strategy by analyzing the historical data and rules in the knowledge graph, and the computing layer improves the data processing and calculation methods based on the new knowledge in the knowledge graph; Update and maintenance of dynamic knowledge graphs: When new devices are connected to the system or the status of existing devices changes, the perception layer updates the relevant information to the knowledge graph, allowing the control layer and computing layer to make decisions based on the latest knowledge.

Images