Dynamic Evaluation System for Three-Phase Imbalance in Distribution Networks Based on Distributed Edge Computing

Abstract

This invention discloses a dynamic assessment system for three-phase imbalance in power distribution networks based on distributed edge computing, specifically relating to the field of power system automation technology. It includes a cloud management platform, multiple edge computing nodes, and intelligent monitoring and control terminals deployed in the power distribution network. The intelligent monitoring and control terminals are responsible for collecting data and extracting feature quantities. The edge computing nodes manage the collaborative assessment domain, perform data fusion and dynamic assessment within the domain, and, through the operation of a distributed collaborative decision-making algorithm among the nodes, jointly negotiate and generate mutually coordinated local control commands, which are ultimately issued to controllable power equipment for implementation. This invention, through its distributed edge collaborative architecture, achieves rapid perception, collaborative control, and global optimization of three-phase imbalance states, effectively improving the speed of governance, system reliability, and security capabilities.

Description

Technical Field

[0001] This invention relates to the field of power system automation technology, and more specifically, to a dynamic assessment system for three-phase imbalance in distribution networks based on distributed edge computing. Background Technology

[0002] In the power distribution network, the rapid growth of unbalanced loads such as residential loads, single-phase photovoltaics, and electric vehicle charging piles has led to an increasingly prominent three-phase imbalance problem. Severe imbalance can cause neutral point voltage deviation, a surge in line losses, and transformer overheating, which not only affects the economic efficiency of power supply but also threatens the safety of the power grid.

[0003] However, the current governance solutions have obvious bottlenecks: Centralized governance solution: This solution relies on a main station or cloud platform for centralized data collection, calculation and decision-making. All terminal monitoring data must be uploaded to the center, and the central server will analyze the data before sending control instructions down. This solution has inherent defects such as high response latency, high communication bandwidth pressure and poor system reliability due to single point failure of the central server. It cannot meet the rapid response requirements for dynamic changes in three-phase imbalance. Local compensation scheme: This scheme uses devices such as static var generators and smart capacitors, which are installed in the problem area and perform independent compensation based on local measurement signals. Although the response speed is improved, this type of scheme lacks a global perspective. There is no coordination between the compensation devices, which can easily lead to conflict in the control targets of adjacent equipment or negative transfer effects after treatment. That is, it solves the local imbalance but aggravates the imbalance of adjacent lines, and cannot achieve the global optimization operation of the distribution network. Simple edge computing applications: Some recent technologies have proposed data preprocessing or simple calculations on terminal devices, but their computing power is limited, and their functions are mostly limited to data filtering and single indicator judgment. They cannot execute complex imbalance evaluation algorithms and multi-objective optimization decisions. At the same time, edge nodes are still information silos, lacking effective collaboration mechanisms, and the governance effect is limited.

[0004] Therefore, in view of the limitations of the prior art, the present invention provides a dynamic evaluation system for three-phase imbalance in distribution networks based on distributed edge computing. Summary of the Invention

[0005] To overcome the aforementioned deficiencies of the prior art, this invention provides a dynamic evaluation system for three-phase imbalance in distribution networks based on distributed edge computing, in order to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a dynamic evaluation system for three-phase imbalance in a distribution network based on distributed edge computing, comprising: a cloud management platform, multiple edge computing nodes, and intelligent measurement and control terminals deployed at multiple monitoring points in the distribution network; The intelligent measurement and control terminal is communicatively coupled with the edge computing node, and is used to collect three-phase electrical quantity data of the power distribution network, extract feature quantities, and upload the feature quantity data containing negative sequence components and / or voltage imbalance to the corresponding edge computing node. The edge computing node is deployed in the substation or distribution room of the power distribution network and is communicatively coupled with one or more intelligent measurement and control terminals and at least one controllable power device to manage a physically continuous collaborative evaluation domain. The cloud management platform is communicatively coupled with all edge computing nodes; The system is configured as follows: The edge computing node receives feature data uploaded by all intelligent measurement and control terminals within its collaborative evaluation domain, performs data fusion, and generates a real-time three-phase imbalance dynamic evaluation result for that domain. Multiple edge computing nodes interact with each other through a communication network and run a distributed collaborative decision-making algorithm. Based on the evaluation results of each domain, they negotiate together so that each edge computing node generates its own, mutually collaborative local control instructions. The edge computing nodes issue their respective local control commands to the controllable power equipment under their jurisdiction to perform collaborative governance of three-phase imbalance.

[0007] Preferably, the edge computing node is specifically configured as follows: Based on the physical connection relationship and electrical distance of the intelligent measurement and control terminals it manages, the boundary of its collaborative evaluation domain is dynamically defined or adjusted, and the real-time three-phase imbalance dynamic evaluation results include at least: the global voltage imbalance degree, negative sequence current distribution map and neutral line current over-limit warning obtained based on the data fusion calculation within the domain.

[0008] Preferably, the distributed collaborative decision-making algorithm is a consensus algorithm: During the decision-making process, the edge computing node acts as an intelligent agent in the collaborative network, broadcasting its local imbalance control needs and adjustable resource information to neighboring edge computing nodes and receiving corresponding information from neighboring nodes. Through iterative calculation, all participating edge computing nodes reach a consensus on a global optimization goal, thereby generating collaborative control instructions. The global optimization goal is to minimize the three-phase imbalance across the entire domain while avoiding control conflicts and equipment overload.

[0009] Preferably, the cloud management platform includes a digital twin module, which is used to build and run a virtual model corresponding to the physical distribution network, and predict the three-phase imbalance risk of the distribution network in a specific future period based on historical data, real-time data and external environment data. The cloud management platform is configured to send the risk prediction results as early warning information to relevant edge computing nodes so that the edge computing nodes can initiate preventive control.

[0010] Preferably, the cloud management platform is further configured as follows: Through knowledge distillation technology, pre-trained large-scale artificial intelligence models are transformed into lightweight inference models. These lightweight inference models are then dynamically distributed and deployed to edge computing nodes to support real-time dynamic evaluation of three-phase imbalance and / or distributed collaborative decision-making.

[0011] Preferably, the feature extraction performed by the intelligent measurement and control terminal includes: calculating the fundamental positive sequence, negative sequence, and zero sequence components, and continuously collecting and calculating the instantaneous values ​​of the three-phase electrical quantities at a local sampling frequency higher than the feature quantity data upload frequency.

[0012] Preferably, the controllable power equipment includes any one or more of the following: a static var generator, a smart capacitor bank, a phase-changing switch, an on-load tap changer, and a controllable load.

[0013] Preferably, it also includes a method for dynamic assessment and coordinated control of three-phase imbalance in distribution networks based on distributed edge computing, specifically including: S1. Collect three-phase electrical quantity data through intelligent measurement and control terminal and extract characteristic quantities; S2. Receive characteristic data uploaded by all intelligent measurement and control terminals in its collaborative evaluation domain through edge computing nodes, perform data fusion and dynamic evaluation, and generate real-time three-phase imbalance dynamic evaluation results. S3. Through data interaction and distributed collaborative decision-making among multiple edge computing nodes, collaborative control instructions are jointly negotiated and generated. S4. Distribute collaborative control commands to controllable power equipment through edge computing nodes to execute collaborative governance.

[0014] Preferably, step S3 specifically includes: S3.1 Each edge computing node broadcasts its local imbalance control needs and adjustable resource information to neighboring nodes; S3.2 Each edge computing node runs a consensus algorithm based on local information and received neighbor node information, and seeks the global optimal solution through iterative calculation; S3.3 When a consensus is reached or the convergence condition is met, each edge computing node generates its own final local control instruction, and these local control instructions together constitute the collaborative control instruction.

[0015] Preferably, the method further includes predicting the risk of three-phase imbalance through the digital twin module of the cloud management platform, and distributing the prediction results to relevant edge computing nodes so that the edge computing nodes can integrate the prediction results in dynamic evaluation and collaborative decision-making and initiate preventive control.

[0016] The technical effects and advantages of this invention are as follows: This invention completely changes the high-latency mode of traditional centralized systems—data upload-centralized computing-command issuance—by sinking core computing and decision-making tasks to the edge, and constructing a collaborative architecture of cloud intelligence and edge intelligence. It achieves millisecond-level perception, second-level assessment and collaborative control of three-phase imbalance, transforming from reactive post-event management to rapid in-event response and pre-event prediction and defense, and greatly improves the dynamic power quality management capability of the distribution network. By running a distributed collaborative decision-making algorithm among edge computing nodes, this invention transforms each governance unit from an information silo into a collaborative network, enabling them to jointly negotiate and generate mutually cooperating local control instructions. This effectively solves the negative transfer effect and control conflict problems caused by traditional local compensation devices, achieving a fundamental leap from local optimum to regional global optimum, and significantly improving governance efficiency and overall system operating economy. The distributed system architecture of this invention eliminates the single point of failure risk of centralized master stations. Even if a single edge node fails or communication with the cloud is interrupted, its neighboring nodes can still continue to work through the collaborative network, ensuring that the regional governance function is not paralyzed. By utilizing decentralized autonomy and collaboration capabilities, combined with the predictive early warning of cloud platform digital twins, a highly resilient active defense system that meets the requirements of large-scale power grid security is jointly constructed. Attached Figure Description

[0017] Figure 1 This is a system diagram of the present invention.

[0018] Figure 2 This is a flowchart of the collaborative decision-making process for edge computing nodes in this invention.

[0019] Figure 3 This is a flowchart of the method of the present invention.

[0020] Figure 4 This is a flowchart of step S3 of the present invention.

[0021] The attached diagram is labeled as follows: 10, cloud management platform; 20, edge computing node; 30, intelligent measurement and control terminal; 40, controllable power equipment; 11, digital twin module. Detailed Implementation

[0022] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention. Example

[0023] This invention provides a dynamic assessment system for three-phase imbalance in a distribution network based on distributed edge computing. The specific structure and connection relationships of the system are described in detail below. The system of the present invention mainly includes a cloud management platform 10, an edge computing node 20, and an intelligent measurement and control terminal 30. Specifically, in this embodiment, the edge computing node 20 is illustrated by taking node A and node B as examples, and the intelligent measurement and control terminal 30 is illustrated by taking terminals 1, 2, 3, and 4 as examples. Intelligent Measurement and Control Terminal 30: Deployed at key monitoring points such as distribution transformers and line sectionalizing switches, it has a built-in high-precision synchronous phasor measurement unit that collects instantaneous values ​​of three-phase voltage and current at a frequency of 10kHz. The feature extraction function is executed by the microprocessor in the terminal, which specifically includes: using the fast Fourier transform algorithm to calculate the fundamental positive sequence, negative sequence and zero sequence components, and calculating the voltage imbalance in real time. The intelligent measurement and control terminal 30 does not continuously upload the raw data at 10kHz, but instead packages and uploads the calculated feature data, such as the negative sequence current amplitude / phase angle and voltage imbalance, to its affiliated edge computing node 20 at a frequency of once per second, which greatly saves communication bandwidth. Edge computing node 20: It uses industrial-grade edge servers deployed in the ring network cabinet or power distribution room of a 10kV line. Each edge computing node 20 manages a collaborative evaluation domain. For example, node A manages the line area covered by terminal 1 and terminal 2, and node B manages the area covered by terminal 3 and terminal 4. The boundary of the collaborative evaluation domain is not fixed. The system can dynamically adjust the management scope of each node through the cloud management platform 10 according to changes in network topology (such as switch opening and closing). After node A receives the characteristic data uploaded by terminal 1 and terminal 2, it performs data fusion on the data. This is not just a simple averaging, but rather combines line parameters to generate a real-time dynamic evaluation result of the three-phase imbalance in this domain. The result includes the following: Global voltage imbalance: An indicator representing the overall level of the region, calculated by weighting data from all monitoring points within the domain. Negative sequence current distribution diagram: This diagram graphically displays the magnitude and direction of negative sequence current in each branch within the domain, accurately locating the source of imbalance. Neutral current over-limit warning: When the calculated neutral current exceeds the safety threshold, an early warning is triggered immediately.

[0024] Distributed collaborative decision-making: Suppose that the evaluation results of node A show that its intra-domain imbalance is severely excessive, requiring the initiation of regulation. Node A and node B interact with each other via a 5G dedicated network or fiber optic Ethernet, and they run a consensus algorithm (specifically, an average consensus algorithm or a distributed ADMM algorithm can be used). During this process: Node A broadcasts its local imbalance control needs (such as the need to compensate for -50kVar reactive power) and adjustable resource information (such as the remaining capacity of its managed static var generator being 100kVar) to Node B. Node B also broadcasts its local information. Based on the received information, the two nodes continuously update their decision variables through iterative calculations, and finally reach a consensus on a global optimization goal. This goal is to minimize the three-phase imbalance of the entire AB region while ensuring that no line is overloaded and the control equipment does not exceed its limits. After reaching a consensus, Node A generates its local control command (such as instructing its managed SVG to output -30kVar reactive power), and Node B also generates its local control command (such as instructing its managed smart capacitor bank to put one group into operation). These two commands work together to achieve global optimization, rather than acting independently. Cloud Management Platform 10: Deployed in a cloud data center, it communicates with all edge nodes via encrypted IPsec VPN. Its digital twin module 11 constructs a virtual model corresponding to the physical distribution network at a 1:1 scale based on data such as power grid GIS and SCADA. This model accesses weather forecast data (such as sunshine and temperature for the next 2 hours) and calendar information (such as holidays) from the meteorological bureau. Through machine learning models, it predicts the three-phase imbalance risks that may occur in the next 15-30 minutes. For example, if it predicts that a sudden drop in photovoltaic output in the afternoon may lead to a serious imbalance in a certain area, the platform will issue an early warning to nodes A and B in advance. The platform also uses knowledge distillation technology to transform complex LSTM prediction models trained on massive amounts of data in the cloud into lightweight models with fewer parameters and higher computational efficiency, and dynamically deploys them to edge nodes, enabling edge nodes to also have a certain short-term prediction capability. Controllable electrical equipment 40: Including but not limited to static var generators, smart capacitor banks, phase switching switches, on-load tap changers, and electric vehicle charging pile clusters as controllable loads, edge computing nodes 20 send local control commands to these devices in the form of remote control commands through standard industrial protocols such as IEC 61850 or Modbus to execute the final governance actions. Example

[0025] This embodiment describes in detail the steps of the method based on the above system, specifically including the following steps: S1. Data Acquisition and Feature Extraction: The intelligent measurement and control terminal 30 acquires three-phase electrical quantity data at a high frequency and calculates feature quantities such as negative sequence component, zero sequence component and voltage imbalance in real time. S2. Intra-domain data fusion and dynamic evaluation: Edge computing node 20 receives feature data uploaded by all terminals in its collaborative evaluation domain, performs data fusion, and generates real-time three-phase imbalance dynamic evaluation results for the domain, including global imbalance, negative sequence current distribution, and neutral line current warning. S3, Distributed Collaborative Decision Making: The specific steps include: S3.1 Each edge computing node 20 (such as node A and node B) broadcasts its local imbalance control needs and adjustable resource information to neighboring nodes; S3.2 Each node runs a consensus algorithm based on local information and received neighbor information, and seeks the global optimal solution through multiple iterative calculations; S3.3 When the change in the decision variables of all nodes is less than the preset threshold (i.e., consensus is reached), the iteration stops, and each node generates its own final local control instruction. S4. Command Issuance and Collaborative Governance: Each edge computing node 20 issues its generated local control commands to the controllable power equipment 40 under its jurisdiction to perform collaborative governance. Furthermore, the method also includes predicting the risk of three-phase imbalance through the digital twin module 11 of the cloud management platform 10, and sending the prediction results to the relevant edge computing nodes 20. In steps S2 and S3, the edge nodes can integrate this prediction information and adjust the assessment strategy and decision weight in advance to achieve preventive control.

[0026] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A dynamic assessment system for three-phase imbalance in distribution networks based on distributed edge computing, characterized in that: include: A cloud management platform (10), multiple edge computing nodes (20), and intelligent monitoring and control terminals (30) deployed at multiple monitoring points in the power distribution network; The intelligent measurement and control terminal (30) is communicatively coupled with the edge computing node (20) to collect three-phase electrical quantity data of the power distribution network and extract feature quantities. The feature quantity extraction includes calculating the fundamental positive sequence, negative sequence, and zero sequence components, and uploading the feature quantity data containing the negative sequence component and / or voltage imbalance to the corresponding edge computing node (20). The edge computing node (20) is deployed in a substation or distribution room of the power distribution network and is communicatively coupled with one or more intelligent measurement and control terminals (30) and at least one controllable power device (40) to manage a physically continuous collaborative evaluation domain; the edge computing node (20) is specifically configured to dynamically define or adjust the boundary of its collaborative evaluation domain according to the physical connection relationship and electrical distance of the intelligent measurement and control terminals (30) it manages; The cloud management platform (10) is communicatively coupled to all edge computing nodes (20); The system is configured as follows: The edge computing node (20) receives the feature data uploaded by all intelligent measurement and control terminals (30) in its collaborative evaluation domain, performs data fusion, and generates the real-time three-phase imbalance dynamic evaluation result of the domain. The real-time three-phase imbalance dynamic evaluation result includes at least: global voltage imbalance degree, negative sequence current distribution map and neutral line current over-limit warning obtained based on the data fusion calculation in the domain. Multiple edge computing nodes (20) interact with each other through a communication network and run a distributed collaborative decision-making algorithm. Based on the evaluation results of each domain, they negotiate together so that each edge computing node (20) generates its own local control instructions that are mutually coordinated. The distributed collaborative decision-making algorithm is a consensus algorithm. In the decision-making process, the edge computing node (20) acts as an intelligent agent in the collaborative network, broadcasting its local imbalance control requirements and adjustable resource information to the adjacent edge computing nodes (20) and receiving the corresponding information from the adjacent nodes. Through iterative calculation, all participating edge computing nodes (20) reach a consensus on a global optimization goal, thereby generating collaborative control instructions. The global optimization goal is to minimize the three-phase imbalance of the entire domain while avoiding control conflicts and equipment overload. The edge computing nodes (20) issue their respective local control commands to the controllable power equipment (40) under their jurisdiction to perform collaborative governance of three-phase imbalance.

2. The dynamic assessment system for three-phase imbalance in distribution networks based on distributed edge computing according to claim 1, characterized in that: The cloud management platform (10) includes a digital twin module (11), which is used to build and run a virtual model corresponding to the physical distribution network, and predict the three-phase imbalance risk of the distribution network in a specific future period based on historical data, real-time data and external environment data. The cloud management platform (10) is configured to send the risk prediction results as early warning information to the relevant edge computing nodes (20) so that the edge computing nodes (20) can start preventive control.

3. The dynamic assessment system for three-phase imbalance in distribution networks based on distributed edge computing according to claim 1, characterized in that: The cloud management platform (10) is also configured as follows: Through knowledge distillation technology, pre-trained large-scale artificial intelligence models are transformed into lightweight inference models. The lightweight inference models are dynamically distributed and deployed to edge computing nodes (20) to support them in real-time three-phase imbalance dynamic evaluation and / or distributed collaborative decision-making.

4. The dynamic assessment system for three-phase imbalance in distribution networks based on distributed edge computing according to claim 1, characterized in that: The controllable power equipment (40) includes any one or more of the following: static var generator, intelligent capacitor bank, phase switching switch, on-load tap changer, and controllable load.

5. The dynamic assessment system for three-phase imbalance in distribution networks based on distributed edge computing according to claim 1, characterized in that: It also includes a method for dynamic assessment and coordinated control of three-phase imbalance in distribution networks based on distributed edge computing, specifically including: S1. Collect three-phase electrical quantity data through the intelligent measurement and control terminal (30) and extract characteristic quantities; S2. Receive characteristic data uploaded by all intelligent measurement and control terminals (30) in its collaborative evaluation domain through the edge computing node (20), perform data fusion and dynamic evaluation, and generate real-time three-phase imbalance dynamic evaluation results. S3. Through data interaction and distributed collaborative decision-making among multiple edge computing nodes (20), collaborative control instructions are jointly negotiated and generated. S4. The collaborative control instructions are sent to the controllable power equipment (40) through the edge computing node (20) to perform collaborative governance.

6. The dynamic evaluation system for three-phase imbalance in distribution networks based on distributed edge computing according to claim 5, characterized in that: Step S3 specifically includes: S3.1 Each edge computing node (20) broadcasts its local imbalance control requirements and adjustable resource information to neighboring nodes; S3.2 Each edge computing node (20) runs a consensus algorithm based on local information and received neighbor node information, and seeks the global optimal solution through iterative calculation; S3.3 When a consensus is reached or the convergence condition is met, each edge computing node (20) generates its own final local control instruction, which together constitutes the collaborative control instruction.

7. The dynamic assessment system for three-phase imbalance in distribution networks based on distributed edge computing according to claim 5, characterized in that: The method also includes predicting the risk of three-phase imbalance through the digital twin module (11) of the cloud management platform (10), and sending the prediction results to the relevant edge computing nodes (20) so that the edge computing nodes (20) can integrate the prediction results in dynamic evaluation and collaborative decision-making and initiate preventive control.

Images