An adaptive management method and system for a distributed power distribution network intelligent terminal device

Abstract

This invention discloses an adaptive management method and system for intelligent terminal devices in a distributed power distribution network. The method constructs a self-organizing peer-to-peer communication network covering the entire network, enabling neighbor discovery and information sharing among intelligent terminal devices. Each terminal independently calculates control commands based on local electrical parameters and preset multi-objective functions, and optimizes regional control effects through a collaborative mechanism. An adaptive adjustment strategy for the control algorithm is introduced. When an operational anomaly is detected, the terminal can autonomously locate and isolate the fault, and restore power supply based on backup path judgment or microgrid switching. The central management unit periodically conducts capacity assessments and aggregates terminal model parameters through federated learning. The terminal possesses edge autonomy capabilities, maintaining local control and data retention even during communication interruptions. This invention features rapid response, flexible structure, intelligent collaboration, and strong adaptability, making it suitable for intelligent operation and management scenarios in various types of power distribution systems.

Description

Technical Field

[0001] This invention belongs to the field of intelligent operation and control technology of power distribution networks, and in particular relates to an adaptive management method and system for intelligent terminal equipment in distributed distribution networks, which realizes multi-objective optimized operation and fault self-healing control of distribution networks. Background Technology

[0002] Modern power distribution networks are gradually evolving into smart grids, with a large number of distributed energy resources (DERs) such as photovoltaics, wind power, and electric vehicles being connected to the distribution network. On the one hand, the integration of DERs improves the utilization rate of clean energy, but on the other hand, it also brings new challenges to the voltage stability and power balance of the distribution network. Traditional distribution network control and management mainly rely on centralized monitoring by distribution automation systems (DAS) or distribution management systems (DMS) and local regulation by fixed parameter control devices. For example, on-load tap changers (OLTCs) of distribution transformers and capacitor banks are usually regulated according to pre-set values ​​or curves, making it difficult to respond in a timely manner to rapidly changing photovoltaic output and load fluctuations, which may lead to adverse situations such as voltage exceeding limits or frequent operation. Furthermore, for the active power output of distributed sources, simple fixed power factor or segmented curves are often used, which lack the ability to flexibly adjust according to global demand. This results in severe local voltage rises during periods of high sunlight and forced abandonment of DER output during periods of low load, leading to low grid operating efficiency.

[0003] Meanwhile, the need for resilience in handling distribution network faults is becoming increasingly prominent. Power outages caused by extreme weather and aging equipment occur frequently, making the ability to quickly isolate faults and restore power to non-faulty areas a key indicator of distribution network reliability. Existing technologies have implemented fault indicators (FPIs) and automatic sectionalizing switches in some distribution networks, achieving a degree of self-healing capability. However, most Fault Location, Isolation, and Recovery (FLISR) processes still require decision-making by dispatch center personnel or rely on centralized self-healing systems. This model is highly dependent on communication networks and has long response times. When large-scale disasters disrupt communications or cause dispatching failures, the distribution network's lack of autonomy can easily lead to prolonged power outages.

[0004] On the other hand, the communication and information infrastructure of distribution networks is becoming increasingly widespread, with a large number of smart terminals (smart meters, feeder terminals (FTUs), transformer terminals (TTUs), etc.) distributed throughout the network. This provides the basic conditions for realizing distributed autonomous control. However, currently, the functions of these smart terminals are relatively independent and fixed, lacking a coordination mechanism. For example, sectionalizing switch controllers trip only based on local current criteria, making it difficult to coordinate the isolation range with adjacent equipment in a timely manner; reactive power compensation devices operate independently, which may lead to overcompensation or undercompensation due to their overlapping effects. How to enable smart terminals scattered in the field to work collaboratively and be managed in a unified manner, fully leveraging their sensing and control capabilities, is an important direction for the development of smart distribution networks.

[0005] Furthermore, distribution network planning and real-time operation management need to be considered together. Taking DER access as an example, the distribution network's capacity to accept new DERs needs to be assessed during the planning phase to identify weaknesses in advance; during the operation phase, the DER output should be appropriately constrained based on real-time operating conditions to ensure safety. Traditionally, planning and operation have been two separate systems, lacking real-time interaction and dynamic adjustment, resulting in either missing the best opportunity for transformation or measures lagging behind changes on site.

[0006] Therefore, there is an urgent need for a new technical solution that can integrate communication, autonomous control, and intelligent decision-making, enabling various smart terminals in the distribution network to work together to form an organic whole and achieve adaptive management of the distribution network: it can optimize economic indicators such as voltage and loss under normal conditions, and can quickly isolate faults, reconstruct the network and restore power supply in abnormal situations, and continuously learn and improve its performance as the environment and data accumulate. Summary of the Invention

[0007] The purpose of this invention is to address the shortcomings of the aforementioned background technology by providing an adaptive management method and system for intelligent terminal devices in distributed power distribution networks. Through this invention, intelligent terminal devices in the power distribution network can automatically establish communication networks, coordinate and control operating parameters in real time, respond quickly and self-heal from faults without centralized human intervention, and continuously optimize strategies through data-driven analysis, thereby achieving highly reliable, efficient, and flexible management of the power distribution network.

[0008] To achieve the above objectives, the present invention employs the following technical solution:

[0009] An adaptive management method for smart terminal devices in a distributed power distribution network, the method comprising the following steps:

[0010] Constructing a self-organizing communication network: Each smart terminal device discovers neighboring nodes through its communication module and establishes communication connections with them, thereby forming a decentralized peer-to-peer communication topology. This communication network supports information exchange and state synchronization between terminals and has the ability to adaptively evolve its topology.

[0011] Collect and share operating parameters and status information: Smart terminal devices collect local operating parameters, including electrical quantities such as voltage, current, active power, reactive power, and frequency, as well as equipment status information such as switch status and fault status. This information is then synchronously uploaded to the central management unit and adjacent smart terminal devices via a communication network.

[0012] Execution of local control command calculation and collaborative exchange: The smart terminal device performs control calculations based on the collected state parameters and the preset objective function model. The objective function may include voltage control, power regulation, loss optimization, etc. The control results are exchanged between terminals in the form of control commands, providing support for local optimization and collaboration.

[0013] Adaptive adjustment of control algorithm parameters: Each smart terminal device adjusts the step size, weight, or selects different optimization paths of its local control algorithm based on the direction of change of control variables in the current control cycle and the previous cycle, so as to improve control stability and convergence performance.

[0014] Fault identification and isolation control: Multiple smart terminal devices work together to determine operational anomalies based on locally collected data and information uploaded by neighboring terminals, identify faulty sections, and send control commands to circuit breakers or switching equipment according to preset criteria to achieve fault isolation.

[0015] Backup path identification and power supply restoration: When a local fault causes a power outage, the smart terminal device detects whether there is a backup power supply path in the network and whether the voltage conditions match. If the conditions are met, it sends a control signal to the tie switch or distributed power control module to perform closing, or switches to islanded operation mode to maintain power supply capacity.

[0016] The central management unit performs power flow simulation: It periodically collects information such as network topology, node operating parameters, system load and distributed power output, and performs capacity assessment based on power flow analysis methods to provide decision-making basis for network operation scheduling.

[0017] Model parameter exchange and update: Smart terminal devices and the central management unit can exchange control model parameters periodically or on demand to achieve collaborative model updates. Parameter exchange methods may include central aggregation, local weight feedback, etc.

[0018] Edge control capabilities in communication interruption scenarios: When the communication of a smart terminal device is interrupted, it can still perform control calculations based on locally stored operating data, record its control behavior and collected data, and upload the data to the central management unit to complete state alignment after communication is restored.

[0019] Preferably, the construction of the self-organizing peer-to-peer communication network adopts a topology optimization method based on Hamiltonian functions, specifically including:

[0020] Define the global Hamiltonian function H of the communication network as follows:

[0021]

[0022] Where: N is the set of all smart terminal device nodes; H i Let k be the Hamiltonian function of node i, representing the communication cost of the node; i Let r be the degree of node i (i.e., the number of its direct neighbors); i Let A be the communication radius of node i (i.e., the signal coverage area of ​​the node); ij Given an adjacency matrix, if node i and node j can communicate directly, then A ij =1, otherwise 0; d ij α is the physical spatial distance between node i and node j. conn α over α energy α dist These are adjustable weighting coefficients, representing the cost of increasing the number of node connections, the penalty for excessive node connections, the energy cost of communication, and the reward for long-distance communication, respectively.

[0023] The function uses the connection status, communication radius, and adjacency relationship of each smart terminal device as parameters to evaluate network communication costs.

[0024] The optimization goal is to reduce overall communication energy consumption and avoid excessive connections by adjusting the communication radius and adjacency of each smart terminal device, while maintaining network connectivity and improving communication efficiency.

[0025] The parameters of the Hamiltonian function include the degree of the node, physical distance, adjacency matrix, and multiple adjustable weight coefficients, which are used to characterize the node connection cost, energy consumption cost, and topological stability, respectively.

[0026] Preferably, the process by which the smart terminal device calculates local control commands based on collected local status data and a preset objective function includes:

[0027] Multiple operational objectives are preset, and mathematical evaluation functions and weighting coefficients are set for each objective. The objectives include, but are not limited to, minimizing voltage deviation, frequency stability, reducing power loss, and load balancing.

[0028] Based on the collected local electrical parameters, a weighted optimization problem is constructed, and the joint control variables are solved through a control optimization model. The model supports the definition of constraints and the control of solution accuracy.

[0029] The control variables include, but are not limited to: adjusting the transformer tap position, adjusting active / reactive power output, switching capacitor banks, and controlling the inverter operating status.

[0030] When a change in operating status causes a control objective to approach its constraint boundary, the smart terminal device or central management unit adjusts the weight coefficient of the corresponding objective based on the magnitude of the change in status and risk indicators, reorders the control priorities, and executes the updated control calculation process.

[0031] Preferably, when the smart terminal device executes the local preset objective function to calculate the local control command, it further includes adaptive dynamic adjustment of the control step size, specifically including:

[0032] Analyze the changing trends of control variables within two consecutive control periods, calculate the consistency between the direction of change of control variables in the current period and that in the previous period, and construct a similarity index;

[0033] The similarity index is constructed based on the angular relationship between the vectors of the changes in control variables, reflecting the continuity of the control direction;

[0034] Adjust the control step size according to this indicator: increase the step size to speed up the response when the control direction is relatively continuous; decrease the step size to avoid control oscillation when the control direction is unstable.

[0035] The step size adjustment follows a segmented adjustment mechanism, using different step size adjustment coefficients based on the similarity index value to maintain the responsiveness and convergence of the control algorithm.

[0036] Preferably, the step of sending the disconnect command to control the connected circuit breaker or switch specifically includes:

[0037] When a fault occurs, the smart terminal devices located upstream and downstream of the faulty segment detect the voltage amplitude, current direction, and zero-sequence current change electrical parameters in their respective areas.

[0038] The smart terminal device sends a fault identification signal via communication and determines the line segment where the fault is located based on the received result;

[0039] When a transient fault is identified, the smart terminal device configured with reclosing function will output a closing control command after a set delay and notify the adjacent smart terminal to prepare for power restoration.

[0040] If the fault is still detected after the reclosing operation, it is determined to be a permanent fault. The adjacent smart terminal device issues a disconnect control command to control the circuit breaker or switch to perform electrical isolation operation.

[0041] The fault detection and isolation operation is automatically completed based on the electrical parameters collected by the smart terminal device, with a response time between 0.5 seconds and 1 second after the fault occurs, and the process does not rely on manual operation.

[0042] Preferably, after fault isolation is completed, the smart terminal device performs power restoration control, specifically including:

[0043] After the faulty section is isolated, the smart terminal equipment at the end of the power outage area determines whether there is a backup power path and detects the status information and voltage conditions of the interconnection switch on the backup path.

[0044] When it is confirmed that there is a backup path and the voltage values ​​on both sides of the path are similar or synchronized, a closing request is sent to the tie switch control terminal.

[0045] After receiving the request, the interconnecting switch control terminal performs voltage matching verification, and executes the closing control of the interconnecting switch when the conditions are met, so that the power loss area can be restored to power supply through the adjacent feeder.

[0046] If there is no backup path, the distributed power intelligent terminal equipment in the power failure area switches to island operation mode according to the operation mode switching command to establish a local microgrid to maintain power supply to critical loads.

[0047] The smart terminal equipment executes rotational shutdown, peak shaving, or power limiting control according to the preset load priority sequence, thereby suppressing voltage over-limit or load overload;

[0048] The power restoration process is completed through peer-to-peer communication between smart terminal devices. After the power supply status is stable, the smart terminal devices upload the restoration process status information to the central management unit.

[0049] Preferably, the central management unit uses a rolling simulation analysis method to assess the capacity, specifically including:

[0050] The central management unit regularly collects data on the topology of the distribution network, node equipment parameters, load data, and distributed power generation output to establish a power flow calculation model for the distribution network.

[0051] Multiple distributed power source access scenarios are set up, and the access capacity is gradually increased in each scenario. Power flow simulation calculations are performed, and the minimum access value that causes the node voltage or line current to exceed the limit is recorded as the critical value of the node's capacity.

[0052] Repeat the above process for multiple target nodes to generate a data table of the acceptance capacity distribution of each node under the current operating conditions;

[0053] When the output of a distributed power source at a node approaches its critical capacity, the central management unit issues an output adjustment command to the smart terminal device corresponding to that node, controlling it to reduce active power output or adjust reactive power output.

[0054] If the overall capacity of a certain area is insufficient, the central management unit will adjust the power flow scheduling strategy of adjacent areas and output expansion suggestions to the power supply planning system when necessary.

[0055] The capacity assessment process is executed periodically to update the capacity boundary parameters of each node.

[0056] Preferably, the control model upon which the capacity assessment relies achieves parameter sharing through a federated learning mechanism, specifically including:

[0057] Each smart terminal device trains a machine learning model based on local historical data. The model is used for load forecasting, voltage trend judgment, or abnormal state identification.

[0058] After each training round, the smart terminal device extracts model parameters or gradient information and sends them to the central management unit via peer-to-peer communication.

[0059] After receiving the model parameter information, the central management unit performs an aggregation operation. The aggregation method includes weighted average, maximum likelihood estimation, or adaptive fusion strategy to generate unified global model parameters.

[0060] The global model parameters are sent to the smart terminal device to be used as initial parameters for the next round of local training or to replace the current model parameters.

[0061] The model training and aggregation process is executed iteratively according to a set number of rounds or convergence criteria.

[0062] The original runtime data is not transmitted during the federated learning process, thereby avoiding the leakage of original data.

[0063] Preferably, the smart terminal device has edge autonomy capabilities to maintain local model training and control functions during communication interruptions, specifically including:

[0064] When communication with the central management unit is interrupted, the smart terminal device performs node data acquisition, status monitoring, local control and fault handling operations based on the locally stored operating policies and configuration parameters.

[0065] During communication interruption, the smart terminal device continuously calculates control variables and outputs control commands according to the set control cycle, and completes the operation status judgment and control on its own.

[0066] After communication is restored, the smart terminal devices will upload the measurement data, control behaviors and fault events recorded during the interruption to the central management unit.

[0067] The uploaded data includes timestamps, changes in control variables, equipment action records, and fault identifiers, which are used by the central management unit to supplement data and align with historical records.

[0068] An adaptive management system for distributed power distribution network smart terminal devices includes multiple smart terminal devices deployed at power distribution network nodes and a central management unit, wherein:

[0069] The smart terminal device includes:

[0070] The data acquisition module is used to collect the voltage, current, active power, reactive power, and equipment operating status of this node.

[0071] The communication module is used for bidirectional data communication between terminals and between terminals and the central management unit, supporting the establishment and dynamic maintenance of self-organizing network topology;

[0072] The control execution module is used to operate the connected circuit breaker, voltage regulator, inverter or other power execution device according to control commands;

[0073] The local control and decision module is used to receive collected data and communication information, generate control decisions, and output control commands.

[0074] The central management unit includes:

[0075] The communication interface is used to establish data connections with various smart terminal devices and exchange operational information.

[0076] A global database is used to store power distribution network topology information, equipment parameters, operating data, and control records;

[0077] The scheduling decision module is used to perform tasks such as power flow analysis, capacity assessment, control strategy generation, and model parameter aggregation, and to issue scheduling instructions to each smart terminal device.

[0078] The central management unit and the smart terminal devices constitute a hierarchical and collaborative distributed control system. The central management unit provides global scheduling and optimization instructions, while each smart terminal device independently performs local control, communication coordination, status monitoring, fault detection and recovery processing.

[0079] Compared with existing technologies, the beneficial effects of this invention are as follows: By establishing a self-organizing peer-to-peer communication network, this invention enables each smart terminal to automatically discover its neighbors and establish connections, constructing a decentralized communication architecture. Even when communication links are interrupted, local control tasks can still be maintained, significantly improving the system's fault tolerance and operational stability. Each smart terminal collects local electrical operating parameters and synchronously uploads them to the central management unit and adjacent terminals, forming a state data sharing mechanism. This not only enhances the consistency of boundary coordination judgments between terminals but also improves data redundancy capabilities, providing support for abnormal state recovery and strategy correction. Each terminal performs multi-objective optimization control calculations based on local real-time operating conditions. Through collaborative interaction with neighboring terminals, it achieves multi-objective coordinated control such as voltage regulation, power loss minimization, and load balancing, taking into account both power quality and operating efficiency. During the execution of the control strategy, the terminal can dynamically adjust optimization model parameters, such as control step size and target weights, based on feedback results to cope with complex operating conditions such as load fluctuations and power generation disturbances, achieving adaptive strategy adjustment and improving control robustness. In the event of a fault, the intelligent terminal can quickly locate and isolate the fault based on local abnormal electrical quantities and neighbor communication. It then automatically switches power supply paths or enters islanded operation mode. The entire response process is controlled within 0.5 to 1 second, significantly reducing power outage time and improving power supply reliability. The central management unit periodically assesses the distributed power acceptance margin of each node and dynamically adjusts access strategies and operation and maintenance configurations based on operating status, thereby achieving optimized allocation and economical operation of overall system resources. The system further introduces a federated learning mechanism. Each terminal completes model training based on local data and uploads the parameters to the central aggregation point to continuously optimize the global control model. This improves prediction accuracy and control performance without leaking original data, enabling the control strategy to have long-term evolution capabilities and supporting the intelligent evolution goal of becoming more efficient the longer the system operates. Attached Figure Description

[0080] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0081] in:

[0082] Figure 1 This is a schematic diagram of the method of the present invention;

[0083] Figure 2 This is a schematic diagram of the overall system architecture of the present invention. Detailed Implementation

[0084] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, 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, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the described embodiments of the present invention are within the scope of protection of the present invention.

[0085] This invention addresses the problems of slow terminal control response, lack of coordination mechanisms, strong reliance on centralized computing, and poor adaptability in current power distribution networks. It proposes an adaptive management method based on the fusion of self-organizing communication, local control, edge collaboration, and intelligent learning. This method mainly includes the following eight steps:

[0086] Example 1

[0087] like Figure 1 As shown, this is an embodiment of the present invention, which provides an adaptive management method for smart terminal devices in a distributed power distribution network, including:

[0088] Step 1: Construct a self-organizing peer-to-peer communication network

[0089] After system deployment, each smart terminal device initiates the node discovery process through its communication module. The terminal broadcasts its own address information, device identifier, and access capabilities to detect nearby communicable terminals, establishing an initial communication link. The terminal then optimizes the connection based on link quality (RSSI, link latency, etc.), forming a self-organizing peer-to-peer communication network covering the entire network. This network structure does not rely on a central coordinator and offers the following advantages:

[0090] Multi-path routing mechanisms improve immunity to interference;

[0091] The network topology is dynamically adjustable and scalable.

[0092] Each terminal node serves as a basic unit for communication and control coordination, possessing the capability for the entire process of "discovery-connection-reconstruction".

[0093] Step 2: Local Status Collection and Information Sharing

[0094] In this embodiment, each smart terminal device is equipped with a data acquisition module, which is used to periodically collect the electrical operating parameters of the local power distribution node, including voltage, current, active power, reactive power, frequency, voltage harmonics, etc., as well as the operating status of the connected equipment, such as switch position signals, energy storage status information, etc.

[0095] Data is collected and uploaded to the following location at set intervals:

[0096] Central Management Unit: Used for network-wide status analysis, capacity assessment, and policy generation;

[0097] Adjacent smart terminal devices: used for boundary state perception and collaborative control coordination.

[0098] This mechanism integrates spatial redundancy of state awareness with communication coordination, significantly improving the system's sensitivity and local perception accuracy.

[0099] Step 3: Independent Control Calculation and Collaborative Control Execution

[0100] Based on the local status information, the smart terminal device independently constructs the objective function and performs control optimization calculations according to the preset operating objectives (such as voltage control, power balance, frequency stability, etc.).

[0101] Typical control objectives may include:

[0102] The node voltage is maintained within a set range (e.g., ±5%).

[0103] Prioritize the use of local energy storage resources to reduce losses;

[0104] The control strategy takes into account both load characteristics and time period priority.

[0105] After the control calculation, the terminal executes the following control commands:

[0106] Local execution: Drives circuit breakers, voltage regulators, and inverters through the control module;

[0107] Collaborative execution: Through peer-to-peer communication protocols, boundary variables are sent to adjacent terminals to collaboratively adjust strategies such as load shifting and reactive power response.

[0108] This mechanism is significantly different from the traditional centralized strategy distribution model, emphasizing horizontal communication between terminals and local intelligent collaborative control mechanisms to form a regional autonomous closed loop.

[0109] Step 4: Evaluation of Control Effect and Adaptive Adjustment of Algorithm

[0110] After each round of control, the terminal evaluates the state changes before and after control execution. Key evaluation metrics include: the degree to which the control objective is achieved; the convergence rate of state variables; and the magnitude of control variable adjustments.

[0111] If the control behavior is found to be inefficient or causes local oscillations, the terminal will automatically adjust the control algorithm parameters (such as target weights, control step size, and switching strategy thresholds) to achieve adaptive optimization of the control algorithm. This mechanism gives the terminal a certain degree of policy self-evolution capability, improving the long-term operating efficiency of the system.

[0112] Step 5: Fault Detection and Isolation Control

[0113] During system operation, all smart terminals continuously monitor local fault symptom parameters, such as: sudden voltage drop or drop to zero; sudden change in zero-sequence current; abnormal current direction;

[0114] If an anomaly is detected by a terminal, a fault identification signal is broadcast to nearby terminals via the communication module. After comparing multiple terminals, the faulty segment is accurately located through location inference and cross-verification of anomaly characteristics. The system follows preset logic: instructs the upstream terminal of the fault to disconnect its switch; synchronously instructs the downstream terminal of the fault to isolate and control it.

[0115] Complete the electrical isolation operation for the faulty section.

[0116] This mechanism is entirely completed by the terminal without the need for central intervention, has high responsiveness, and the isolation processing delay is controlled within 1 second.

[0117] Step 6: Power restoration and transfer control

[0118] After the fault isolation is completed, the terminals in the power outage area begin to execute the power restoration judgment process, which includes:

[0119] Check if a backup power supply path exists (such as a tie switch or distributed power source);

[0120] Check whether the voltage status of the backup path meets the conditions (synchronization, amplitude difference, etc.);

[0121] Assess the operating status of energy storage terminals (SOC, power output capability, etc.);

[0122] Recovery strategies include:

[0123] Start the interconnection switch to close, enabling power transfer to adjacent feeders;

[0124] Initiate local microgrid islanding control to maintain critical load operation;

[0125] Implement load rotation control (rotational shutdown, peak shaving, and power restriction) to ensure local voltage stability.

[0126] The entire recovery process is coordinated through peer-to-peer communication between terminals, and the terminals report the process to the central system for subsequent maintenance and statistical purposes.

[0127] Step 7: Capacity Assessment and Access Strategy Optimization

[0128] The central management unit uses rolling simulation analysis to periodically assess the connectable capacity of distributed power sources (such as photovoltaics) at each node of the distribution network. The assessment methods include:

[0129] Model the current distribution network topology;

[0130] Set up access scenarios to simulate different power increases;

[0131] Perform power flow simulation and record the points where node voltage / line current exceeds the limit;

[0132] Obtain the critical capacity of each node;

[0133] If a node is found to be approaching its access limit, the central management unit issues a power limiting command to the terminal to which the node belongs, and simultaneously adjusts the regional power flow strategy to achieve source-load balance optimization configuration.

[0134] Step 8: Federated Learning Drives Control Policy Evolution

[0135] To improve the adaptability of the terminal control strategy to changes in operating conditions, the system introduces a federated learning mechanism. The execution flow is as follows:

[0136] Each terminal trains a control model (such as a neural network or regression model) based on local historical operating data;

[0137] After the model training is completed, the parameters (not the original data) are extracted and sent to the central management unit;

[0138] The central government performs aggregation (such as weighted average, maximum likelihood estimation) on parameters uploaded from all terminals;

[0139] After aggregation, the parameters are redistributed to the terminal to replace or initialize the next round of model training.

[0140] This process is iterative, enabling the system to achieve continuous optimization and improve its local generalization capabilities, while avoiding the transmission of raw data and ensuring data privacy and security on the terminal side.

[0141] The adaptive management method proposed in this invention constructs a complete intelligent power distribution network control mechanism through self-organizing communication, autonomous optimization control, local fault handling, system capacity simulation, and machine learning optimization. It has high autonomy, high responsiveness, high intelligence, and high adaptability, and has extremely high engineering application value in current and future complex power distribution network environments.

[0142] To further enhance the stability, scalability, and local autonomy of the communication network of the smart terminal device of the present invention, in a preferred embodiment of the present invention, the establishment of the self-organizing peer-to-peer communication network does not depend on a fixed central controller or a preset topology, but rather on a topology evolution mechanism based on the Hamiltonian optimization algorithm, in which each terminal device autonomously completes connection optimization and network structure reconstruction.

[0143] I. Initial Discovery and Connection Establishment

[0144] During the initial deployment or operation phase of the system, after each smart terminal device is powered on, it initiates the neighbor discovery and connection evaluation process. Each terminal broadcasts its own identifier, geographical coordinates, communication power level, and the number of connections it can support through its communication module, forming a candidate set of neighbors.

[0145] Without connection rules to constrain them, all terminals may tend to prioritize connecting to nodes with high power and central location (such as terminal A, the master station), forming an unbalanced "star topology". This can lead to problems such as communication congestion, high master station load, and prominent single point of failure risk in large-scale networks.

[0146] To overcome the above problems, this invention introduces the Hamiltonian optimization function as the basis for evaluating communication connections, with each terminal autonomously calculating its current connection structure's "system cost function":

[0147]

[0148] Where: N is the set of all smart terminal device nodes; H i Let k be the Hamiltonian function of node i, representing the communication cost of the node; i Let r be the degree of node i, which is the number of its direct neighbors; i Let A be the communication radius of node i, i.e., the signal coverage area of ​​the node; ij Given an adjacency matrix, if node i and node j can communicate directly, then A ij =1, otherwise 0; d ij α is the physical spatial distance between node i and node j. conn α over α energy α dist These are adjustable weighting coefficients, representing the cost of increasing the number of node connections, the penalty for excessive node connections, the energy cost of communication, and the reward for long-distance communication, respectively.

[0149] The function uses the connection status, communication radius, and adjacency relationship of each smart terminal device as parameters to evaluate network communication costs.

[0150] The optimization goal is to reduce overall communication energy consumption and avoid excessive connections by adjusting the communication radius and adjacency of each smart terminal device, while maintaining network connectivity and improving communication efficiency.

[0151] The parameters of the Hamiltonian function include the degree of the node, physical distance, adjacency matrix, and multiple adjustable weight coefficients, which are used to characterize the node connection cost, energy consumption cost, and topological stability, respectively.

[0152] This function comprehensively evaluates the connection quality of the terminal from four dimensions: distance minimization, connection quantity balance, load balancing, and path redundancy control.

[0153] II. Connection Optimization and Evolution Mechanism

[0154] Each terminal periodically evaluates the Hamiltonian value corresponding to the current connection structure and executes the following connection adjustment actions through local decision logic:

[0155] 1. Connection establishment strategy:

[0156] If a new neighbor j is found to satisfy:

[0157] d ij <d 阈值 ;

[0158] ρ j <ρ 上限 ;

[0159] After adding a connection Then the connection will be established.

[0160] 2. Connection disconnection strategy:

[0161] For the current connection j, if the following conditions are met:

[0162] There are alternative paths among the redundant paths;

[0163] Or the connection leads to H i If the threshold is exceeded, the terminal will actively disconnect the connection.

[0164] 3. Connection power adjustment:

[0165] If a terminal (e.g., terminal A) detects that the number of connections is too high (e.g., >20), it will automatically reduce the broadcast power and reduce the number of non-critical terminal connections.

[0166] Terminals located at the network edge or contact points (such as Class C terminals) increase their transmission power to enhance connectivity and coverage.

[0167] This mechanism achieves self-optimization of the network's communication topology based on independent decision-making at each terminal. The final network structure is as follows: a "chain + network" fusion structure is formed within each region, with the average node connectivity maintained between 2 and 4.

[0168] The main station terminal is directly connected to the beginning and end of each feeder, which facilitates data aggregation;

[0169] Each feeder is horizontally connected by a communication terminal to form a single connected cluster.

[0170] III. Analysis of Technological Advantages

[0171] The Hamiltonian optimization mechanism for communication networks used in this invention has the following technical innovations compared to the fixed communication topology or centralized configuration method in traditional power distribution automation:

[0172] Distributed evolution and autonomous decision-making: Terminals can dynamically evaluate connection quality based on local information without the need for centralized scheduling, thus improving the adaptability and autonomy of the topology;

[0173] Quantitative connectivity evaluation metrics: By integrating Hamiltonian functions, we consider communication quality, load balancing, and path diversity to ensure optimal network performance;

[0174] Supports dynamic reconfiguration and redundancy adjustment: The terminal can automatically adjust the connection when the neighbor changes, a faulty node appears or the channel conditions fluctuate, and has strong fault tolerance capability;

[0175] Modular deployment and parameter flexibility: The connection evaluation weights can be dynamically configured according to factors such as regional structure, operating load, and communication capacity to adapt to different power grid structures.

[0176] The communication network constructed by this mechanism exhibited good connectivity (connectivity rate > 99%), low average hop count (< 3 hops), and fault tolerance (maintaining network connectivity even with the failure of 1-2 nodes) in actual tests, providing a reliable foundation for subsequent control coordination and data sharing.

[0177] In a preferred embodiment of the present invention, when the smart terminal device executes control command calculations, it adopts a multi-objective weighted optimization model, which simultaneously considers multiple operational objectives such as voltage stability, frequency regulation, power loss reduction, and load balancing.

[0178] The specific implementation steps are as follows:

[0179] 1. Preset set of objective functions:

[0180] Voltage deviation target: Minimize |V i -V ref |;

[0181] Frequency stability objective: Minimize |f i -f nom |;

[0182] Loss control objective: Minimize power loss P loss ;

[0183] Load balancing objective: To balance the current distribution of adjacent nodes;

[0184] 2. Assign weight coefficients w1, w2, w3, and w4 to each of the above objective functions to form the overall objective function:

[0185] min(w1F v +w2F f +w3F loss +w4F load )

[0186] 3. The optimization problem is a constrained problem, and the constraints include:

[0187] Control variable boundaries (such as transformer tap limits, inverter output range); node voltage constraints (such as ±5%); load power factor and other operating limitations;

[0188] 4. Control variables include: transformer tap position; active / reactive power output; capacitor bank switching status; energy storage / photovoltaic inverter operating mode, etc.

[0189] 5. In actual operation, if the system operating state changes abruptly, such as when a target approaches its physical constraint boundary (e.g., the voltage drops below 0.95 pu), the system can dynamically adjust the weight coefficients and reorder the control priorities, for example, by increasing the weight of the voltage stability target and executing the updated control optimization.

[0190] This mechanism enables the terminal to flexibly reconstruct control strategies based on actual operation without relying on a central authority, significantly improving the adaptive capability and precision of control.

[0191] To improve control convergence efficiency and avoid oscillations, this invention further introduces an adaptive control step size adjustment mechanism into the terminal local control algorithm to replace the fixed step size setting method used in traditional control algorithms, thereby overcoming the problems of slow response or oscillations when facing rapid fluctuations or disturbances.

[0192] The basic idea behind this adjustment mechanism is that during each round of local control optimization, the smart terminal dynamically analyzes the changing trends of control variables (such as voltage setpoints, active / reactive power reference values, etc.) and evaluates the consistency in direction between the current control action and the control action of the previous cycle. This consistency can be reflected by calculating the angle between the change vectors of control variables in two consecutive cycles; the closer the value is to 1, the higher the consistency in control direction, and the closer the value is to -1, the opposite the direction.

[0193] Based on the above control direction similarity index, the smart terminal adjusts the control step size in segments according to its numerical range: when the control direction change is highly consistent, the current optimization direction is considered reliable, and the step size can be appropriately increased to improve the convergence speed.

[0194] When a significant reversal or drastic change in direction is detected, it is considered that the current control strategy may have overshoot or disturbance effects, and the step size should be reduced to stabilize the system.

[0195] If the direction consistency is in an intermediate state, then keep the current step size unchanged and maintain the control rhythm.

[0196] To achieve the above objectives, the system presets step size increase factor, decrease factor, and similarity threshold parameters. After each control cycle, the smart terminal automatically performs the above judgments and step size updates based on local status data, thereby achieving local control parameter adjustment capabilities that do not rely on central coordination and adapt to actual operating conditions.

[0197] This mechanism has the following technical advantages:

[0198] Highly responsive and adaptive: The control step size can be dynamically adjusted according to the running trend without manual intervention;

[0199] Improve convergence efficiency: Under continuous operation, the control rhythm can be rapidly expanded to improve efficiency;

[0200] Suppressing oscillations: Reducing the step size in a timely manner can effectively prevent problems such as overcorrection and oscillations;

[0201] Strong edge feasibility: The entire strategy can be executed by relying on local variables and historical information, making it suitable for edge smart terminal deployment environments.

[0202] This improvement solves the problem that fixed control step size in traditional distributed control systems can lead to a tradeoff between slow response and poor stability. It represents a technological breakthrough at the control strategy level and has a significant effect on enhancing the stable operation capability of the system under dynamic conditions.

[0203] When a fault occurs on a line in the distribution network (such as single-phase grounding, short circuit, etc.), multiple smart terminal devices work together based on local electrical parameters to locate and isolate the faulty section. The specific process is as follows:

[0204] After the fault occurred, the upstream and downstream terminals of the faulty section detected the following abnormal signals: voltage sudden drop or drop to 0; abnormal current direction or sudden increase; sudden change in zero-sequence current, etc.

[0205] The terminal that detects the anomaly immediately sends a fault identification signal to the nearest terminal via peer-to-peer communication.

[0206] After receiving identification signals from upstream and downstream terminals, each terminal determines that the fault is located in the line segment between them;

[0207] If the fault is determined to be transient, the terminal configured with reclosing function will issue a closing command after a delay.

[0208] If reclosing fails, it is determined to be a permanent fault, and the corresponding terminal executes a disconnect command to control the circuit breaker or switch to isolate the faulty section.

[0209] The entire process is completed automatically by the terminal, with a response time controlled within 0.5 to 1 second, requiring no manual intervention.

[0210] This mechanism supports a three-level closed-loop structure of local judgment, regional collaboration, and rapid isolation, which significantly improves the system's self-healing capability and operational continuity.

[0211] After fault isolation is completed, the terminal automatically initiates the power restoration control process, which has the ability to make hierarchical judgments and autonomous coordination.

[0212] When the terminal detects a power outage on its side, it determines whether there is a backup power supply path (such as a tie switch).

[0213] If a path exists, the terminal checks the status of the handover switch, the voltage amplitude and phase conditions on both sides;

[0214] When the voltage is determined to be synchronized or approximately synchronized, the terminal sends a closing request to the interconnecting switch control terminal.

[0215] The control terminal completes the voltage matching verification. If the conditions are met, remote closing control is executed.

[0216] If the power supply cannot be restored through communication, a backup power source (such as a local distributed power source) will be called in, and the system will switch to islanded operation mode according to the operation strategy to build a local microgrid to power critical loads.

[0217] Strategies such as peak shaving and valley filling, and rotating power outages are implemented based on load priority to control voltage over-limit and overload.

[0218] Once power is restored, the terminal uploads the restoration process information to the central management unit for dispatch reference and data archiving.

[0219] This process features "autonomous judgment, collaborative execution, and diversified paths," adapting to various network faults and recovery path selection needs.

[0220] In this embodiment, the central management unit periodically collects the following operational information: network topology; node device parameters; real-time load data; and distributed power supply output.

[0221] Based on the above information, a power flow model of the distribution network is constructed, and the following evaluation process is performed:

[0222] Define multiple new distributed power supply access scenarios;

[0223] In each scenario, the access power is gradually increased (e.g., by 0.1MW each time), and power flow simulation is performed.

[0224] Record the minimum access capacity when the node voltage exceeds the limit or the line current exceeds the limit, and use it as the upper limit of the node's access capacity.

[0225] Repeated simulations were performed on all key nodes in the network to generate a complete capacity distribution table.

[0226] If the output of a node approaches its capacity limit, a power adjustment command is sent to its corresponding terminal.

[0227] If the overall regional capacity is insufficient, adjust the power flow scheduling or provide capacity expansion suggestions to the power planning department.

[0228] This mechanism ensures that the security boundary between power generation, load and grid in the distribution network is not breached, and supports the friendly access of distributed energy resources.

[0229] This invention employs a federated learning mechanism to achieve privacy-preserving collaborative training of the terminal control model. The process is as follows:

[0230] Each smart terminal trains a model based on local data. The tasks include: load forecasting; voltage trend analysis; and fault mode identification. After each round of training, the terminal extracts model parameters or gradient information (excluding raw data) and sends them to the central management unit through peer-to-peer communication.

[0231] The central management unit aggregates and calculates the parameters uploaded by each terminal using methods including: weighted average (FedAvg); maximum likelihood estimation; adaptive fusion strategy; generating unified global model parameters, which are then redistributed to each terminal as initialization parameters for the next round of local training.

[0232] The process is executed iteratively according to round settings or performance convergence criteria.

[0233] This method enhances the model's generalization ability and the terminal's control intelligence while ensuring privacy and security.

[0234] To ensure high system availability, the smart terminal of this invention possesses edge autonomy capabilities, supporting continuous control and model training tasks even during communication interruptions, specifically including:

[0235] After the communication disconnection detection, the terminal enters autonomous operation mode;

[0236] Continue to collect node status according to the set period and execute control actions based on the local storage policy;

[0237] It can independently complete the fault identification and isolation process without relying on central instructions;

[0238] The local model training task continues, recording the training rounds and parameter changes;

[0239] After communication is restored, the terminal will upload all operation records during the interruption period, including timestamps, changes in control variables, switch action records, and fault handling logs.

[0240] The central management unit completes data supplementation and alignment with historical data to maintain system continuity.

[0241] This mechanism significantly enhances the system's disaster recovery capabilities, control independence, and data integrity.

[0242] like Figure 2 As shown, another embodiment of the present invention provides an adaptive management system for distributed power distribution network smart terminal devices, including multiple smart terminal devices deployed at power distribution network nodes and a central management unit, wherein: the smart terminal devices include:

[0243] The data acquisition module is used to collect the voltage, current, active power, reactive power, and equipment operating status of this node.

[0244] The communication module is used for bidirectional data communication between terminals and between terminals and the central management unit, supporting the establishment and dynamic maintenance of self-organizing network topology;

[0245] The control execution module is used to operate the connected circuit breakers, voltage regulators, inverters or other power execution devices according to control commands; the local control and decision module is used to receive collected data and communication information, generate control decisions and output control commands.

[0246] The central management unit includes:

[0247] The communication interface is used to establish data connections with various smart terminal devices and exchange operational information.

[0248] A global database is used to store power distribution network topology information, equipment parameters, operating data, and control records;

[0249] The scheduling decision module is used to perform tasks such as power flow analysis, capacity assessment, control strategy generation, and model parameter aggregation, and to issue scheduling instructions to each smart terminal device.

[0250] The central management unit and the smart terminal devices constitute a hierarchical and collaborative distributed control system. The central management unit provides global scheduling and optimization instructions, while each smart terminal device independently performs local control, communication coordination, status monitoring, fault detection and recovery processing.

[0251] Example 2

[0252] Application of step-size adaptive adjustment based on control direction similarity in feeder voltage control

[0253] This embodiment takes the intelligent terminal equipment at the end of a 10kV distribution network branch as the object to illustrate the application process of the adaptive adjustment method of control step size proposed in this invention in the local control of feeder voltage.

[0254] This intelligent terminal is deployed at the end of the downstream branch of the distribution transformer. It has local voltage measurement, communication, and control capabilities, and the control objective is to stabilize the voltage at the access point between 10.5kV and 11kV. The terminal performs local voltage optimization adjustment with a control cycle of 1 second, and automatically adjusts the control step size in each adjustment cycle according to the directional relationship between the preceding and following control actions.

[0255] The specific process is as follows:

[0256] Historical variable acquisition: Before the start of control cycle t, the smart terminal has recorded the control variable values ​​v of the previous two cycles. t-2 v t-1 , used for subsequent difference calculations;

[0257] Current control quantity calculation: The terminal collects the current voltage v t Calculate the change in control variables between adjacent periods:

[0258] △v t-1 =v t-1 -v t-2 , △v t =v t -v t-1 ;

[0259] Similarity index calculation: Based on the directional relationship between the two difference vectors mentioned above, the terminal calculates the control direction similarity index:

[0260]

[0261] In the example, if S cos =0.92, indicating that the control direction is highly consistent;

[0262] Step size adaptive adjustment: Terminal preset: γ up =1.05, γ down =0.7, θ high =0.85, θ low =0.15;

[0263] Current S cos =0.92>θ high If the step size acceleration condition is met, the terminal will adjust the current step size to: α t+1 =γ up ×α t =1.05×α t Output control quantity: based on the updated step size α t+1 The system recalculates the next voltage target and generates control commands to act on local voltage regulation devices (such as voltage regulators or reactive power compensation equipment).

[0264] Technical effects: In typical load areas with frequent voltage fluctuations but stable direction, this method can quickly expand the control step size and improve the response rate; in scenarios with unstable direction or intense disturbances (such as sudden load changes), the step size can be reduced in time to prevent overshoot and oscillation; the entire mechanism is executed based on local state data, does not rely on central server scheduling, and has good edge autonomy capabilities.

[0265] In summary, the method of this invention covers the entire lifecycle of distribution network operation, from optimized control in normal conditions to isolation and recovery in fault conditions, and is continuously improved through a learning mechanism. Each step can operate independently or cooperate with each other, forming a closed-loop autonomous management system. This significantly improves the safety (rapid fault response, preventing accident spread), reliability (fewer or zero power outages), economy (reduced losses, reduced voltage drop), and flexibility (adapting to changing supply and demand and equipment conditions) of the distribution network. In particular, this invention fully utilizes existing intelligent terminal resources on-site, achieving "on-site decision-making and on-site execution," greatly reducing reliance on centralized communication and computing, exhibiting better robustness and scalability, and adapting to the future development trend of highly distributed distribution networks and high-proportion renewable energy integration.

[0266] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this application. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of those different embodiments or examples.

[0267] Any process or method description in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or more executable instructions for implementing a particular logical function or process. Furthermore, the scope of the preferred embodiments of this application includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the functionality involved.

[0268] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any person skilled in the art can easily conceive of various variations or substitutions within the technical scope disclosed in this application, and these should all be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A self-adaptive management method of a distributed power distribution network intelligent terminal device, characterized in that, The method comprises the following steps: The intelligent terminal device discovers neighbor nodes through a communication module, establishes a communication connection, and forms a self-organizing peer-to-peer communication network; The intelligent terminal device collects local operating parameters and device states, and uploads them to a central management unit and neighboring intelligent terminal devices through the communication network; The intelligent terminal device calculates local control instructions according to the collected local state data and a preset target function, and transmits control-related data to neighboring intelligent terminal devices through the communication network; The intelligent terminal device further comprises self-adaptive dynamic adjustment of a control step when executing the preset target function to calculate the local control instructions, and specifically comprises: The similarity index is constructed based on the vector angle relationship of the control variable change, and reflects the continuity of the control direction. According to the index, the step size is adjusted: when the control direction continuity is strong, the step size is enlarged to speed up the response; and when the control direction is unstable, the step size is reduced to avoid control oscillation. The step size adjustment follows a segmented adjustment mechanism, and different step size adjustment coefficients are used according to the similarity index value to maintain the responsiveness and convergence of the control algorithm. Based on the collected operating parameters and received neighbor information, the multiple intelligent terminal devices identify fault sections and send opening and closing instructions to control the connected circuit breakers or switches; After completing fault isolation, the intelligent terminal device performs power supply recovery control, specifically comprising: After the fault section is isolated, the intelligent terminal device at the end of the power-off area judges whether there is a standby power supply path, and detects the state information and voltage condition of the tie switch on the standby path; When it is confirmed that there is a standby path and the voltage values on both sides of the path are similar or synchronized, a closing request is sent to the tie switch control terminal; The tie switch control terminal performs voltage matching verification after receiving the request, and performs closing control of the tie switch when the condition is met, so that the power-off area restores power supply through the adjacent feeder; If there is no standby path, the distributed power intelligent terminal device in the power-off area switches to an island operation mode according to the operation mode switching instruction, and establishes a local microgrid to maintain key load power supply; The intelligent terminal device executes load shedding or load control according to the preset load priority order, thereby suppressing voltage over-limit or load overload; The power supply recovery process is coordinated through peer-to-peer communication between intelligent terminal devices, and after the power supply state is stable, the intelligent terminal device uploads the recovery process state information to the central management unit; The central management unit collects network topology, node parameters, load data and distributed power output information, performs power flow simulation and evaluates the receiving capacity; The control model relied on by the capacity evaluation realizes parameter sharing through a federated learning mechanism, specifically comprising: Each intelligent terminal device trains a machine learning model based on local historical data, and the model is used for load prediction, voltage trend judgment or abnormal state recognition; After each round of training is completed, the intelligent terminal device extracts model parameters or gradient information and sends them to the central management unit through peer-to-peer communication. ​ The central management unit receives the model parameter information and performs an aggregation operation, which includes weighted average, maximum likelihood estimation or adaptive fusion strategy, to generate a unified global model parameter; The global model parameter is sent to the smart terminal device and used as the initial parameter for the next round of local training or to replace the current model parameter; The model training and aggregation process is iteratively executed according to a set number of rounds or convergence criteria; In the federated learning process, no raw operation data is transmitted, thereby avoiding leakage of raw data; During a communication interruption, the smart terminal device performs control calculation based on locally stored data and records measurement data and control actions, and sends the recorded data to the central management unit after the communication is restored.

2. The adaptive management method of the distributed power distribution network intelligent terminal device according to claim 1, characterized in that, The construction of the self-organizing peer-to-peer communication network adopts a topology optimization method based on a Hamilton function, which specifically includes: Defining a global hamiltonian function of a communication network with the expression ； wherein: is the set of all intelligent terminal device nodes; is the node Hamiltonian function of the node, representing the communication cost of the node; is the degree of the node , i.e. the number of direct neighbors of the node; is the communication radius of the node , i.e. the signal coverage range of the node; is the adjacency matrix, if the node can directly communicate with the node , then , otherwise 0; is the physical space distance between the node and the node ; , , , are adjustable weight coefficients, respectively representing the cost of the number of node connections, the punishment of node over-connection, the communication energy consumption cost and the reward of long-distance communication; The function takes the connection state, communication radius and adjacency relationship of each smart terminal device as parameters and is used to evaluate the network communication cost.

3. The adaptive management method of distributed power distribution network smart terminal device according to claim 1, characterized in that, The process of calculating the local control instruction by the smart terminal device based on the collected local state data and the preset target function includes: A plurality of operation targets are preset, and a mathematical form evaluation function and a weight coefficient are set for each target. The targets include, but are not limited to, minimization of voltage deviation, frequency stability, power loss reduction and load balancing. Based on the collected local electrical parameters, a weighted optimization problem is constructed, and joint control variables are solved by a control optimization model. The model supports constraint condition definition and solution precision control. The control variables include, but are not limited to, adjusting the transformer tap position, adjusting the active / reactive power output, switching the capacitor bank and controlling the inverter operating state. When the operating state changes cause a control target to approach its constraint boundary, the smart terminal device or the central management unit adjusts the weight coefficient of the corresponding target according to the state change amplitude and the risk index, and reorders the control priority to perform the updated control calculation process.

4. The adaptive management method of distributed power distribution network smart terminal device according to claim 1, characterized in that, The sending opening instruction controls the connected circuit breaker or switch, specifically including: When a fault occurs, the smart terminal devices located upstream and downstream of the fault segment detect the voltage amplitude, current direction and zero sequence current mutation electrical parameters of their respective regions; The smart terminal device sends a fault identification signal through communication and judges the fault location based on the received results; When a transient fault is judged, the smart terminal device configured with a reclosing function outputs a closing control instruction after a set delay and notifies the adjacent smart terminal to prepare for power restoration operation; If the fault still exists after the reclosing operation, it is judged as a permanent fault, and the adjacent smart terminal device issues an opening control instruction to control the connected circuit breaker or switch to perform electrical isolation operation; The fault detection and isolation operation is automatically completed based on the electrical parameters collected by the smart terminal device, with a response time of 0.5 to 1 second after the fault occurs. The process does not rely on manual operation.

5. The adaptive management method of distributed power distribution network smart terminal device according to claim 1, characterized in that, The central management unit uses a rolling simulation analysis method to evaluate the admission capacity, specifically including: The central management unit periodically collects the topology structure, node device parameters, load data and distributed power output of the distribution network, and establishes a power flow calculation model of the distribution network; Set multiple distributed power access scenarios, gradually increase the access capacity in each scenario and perform power flow simulation calculation, record the minimum access value that makes the node voltage or line current exceed the limit as the node capacity acceptance critical value; Repeat the above process for multiple target nodes to generate a table of node capacity acceptance distribution data under the current operating conditions; When the output of a certain node distributed power approaches its capacity acceptance critical value, the central management unit issues an output adjustment instruction to the smart terminal device corresponding to the node to control the reduction of active power output or the adjustment of reactive power output; If the overall acceptance margin of a certain area is insufficient, the central management unit adjusts the power flow dispatching strategy of the adjacent area and outputs expansion suggestions to the power supply planning system when necessary; The capacity evaluation process is performed periodically to update the capacity boundary parameters of each node.

6. The adaptive management method of distributed power distribution network smart terminal device according to claim 1, characterized in that, The smart terminal device has edge autonomy capability to maintain local model training and control functions when communication is interrupted, specifically including: When communication with the central management unit is interrupted, the smart terminal device performs node data acquisition, state monitoring, local control, and fault handling operations based on the locally stored operating strategy and configuration parameters; During communication interruption, the smart terminal device continuously calculates control variables and outputs control instructions according to the set control period to complete the operation state determination and control independently; After communication is restored, the smart terminal device uploads the measurement data, control actions, and fault events recorded during the interruption to the central management unit; The uploaded data includes timestamps, control variable changes, device action records, and fault identifiers for the central management unit to supplement data and align historical records.

7. The adaptive management system of the adaptive management method of the distributed power distribution network intelligent terminal device according to any one of claims 1-6, characterized in that, It includes multiple smart terminal devices deployed at power distribution network nodes and a central management unit, wherein: The smart terminal device includes: A data acquisition module for acquiring voltage, current, active power, reactive power, and device operating status of the node; A communication module for bidirectional data communication between terminals and between terminals and the central management unit, supporting self-organizing network topology establishment and dynamic maintenance; A control execution module for operating connected circuit breakers, voltage regulators, inverters, or other power execution devices according to control instructions; A local control and decision module for receiving acquisition data and communication information, generating control decisions, and outputting control instructions; The central management unit includes: A communication interface for establishing data connection with each smart terminal device and exchanging operating information; A global database for storing power distribution network topology information, device parameters, operating data, and control records; A dispatching decision module for performing power flow analysis, capacity acceptance evaluation, control strategy generation, model parameter aggregation, and other tasks, and issuing dispatching instructions to each smart terminal device; The central management unit and smart terminal device form a hierarchical collaborative distributed control system, with the central management unit providing global dispatching and optimization instructions, and each smart terminal device independently executing local control, communication coordination, state monitoring, fault detection, and recovery processing.

Images