Self-adaptive management method and system for intelligent terminal equipment of distributed power distribution network

Abstract

The invention discloses a self-adaptive management method and system for intelligent terminal equipment of a distributed power distribution network. According to the method, a self-organizing peer-to-peer communication network covering the whole network is constructed, and neighbor discovery and information sharing among intelligent terminal devices are achieved; the terminal independently completes control instruction calculation based on the local electrical parameters and a preset multi-objective function and optimizes an area control effect through a cooperation mechanism; introducing a self-adaptive adjustment strategy of a control algorithm; when abnormal operation is detected, the terminal can autonomously complete fault positioning and isolation, and power supply recovery is realized based on standby path judgment or micro-grid switching; the central management unit periodically carries out acceptance capacity evaluation, and terminal model parameters are aggregated in a federated learning mode; the terminal has an edge autonomous capability, and can maintain local control and data retention when communication is interrupted. The system has the characteristics of quick response, flexible structure, intelligent cooperation, high adaptability and the like, and is suitable for intelligent operation and management and control scenes of various types of power distribution systems.

Description

Technical Field

[0001] The present invention belongs to the technical field of intelligent operation control of power distribution networks, and particularly relates to an adaptive management method and system for intelligent terminal devices of a distributed power distribution network, realizing multi-objective optimal operation and fault self-healing control of the power distribution network. Background Art

[0002] Modern distribution power grids are gradually evolving towards smart grids, and a large number of distributed energy resources (DERs), such as photovoltaic, wind power, and electric vehicles, are connected to the power distribution network. On the one hand, the access 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 power distribution network. Traditional power distribution network control and management mainly rely on centralized monitoring of a distribution automation system (DAS) or a distribution management system (DMS) and local adjustment of fixed-parameter control devices. For example, on-load tap changers (OLTCs) of distribution transformers and capacitor banks are usually adjusted according to preset values or curves, and it is difficult to respond in a timely manner to the rapidly changing photovoltaic output and load fluctuations, which may result in adverse situations such as voltage over-limit or frequent operation. Another example is that for the active power output of distributed power sources, simple power factor fixing or segmented curves are mostly adopted, and they do not have the ability to flexibly adjust according to global demands, resulting in serious local voltage rise during high-light periods and forced curtailment of DER output during low-load periods, and the power grid operation efficiency is not high.

[0003] Meanwhile, the resilience requirement for power distribution network fault handling is prominent. Fault power outages caused by extreme weather and equipment aging occur frequently, and how to quickly isolate faults and restore power supply to non-fault areas has become a key indicator for measuring the reliability of power distribution networks. In the prior art, some power distribution networks are equipped with devices such as fault indicators (FPIs) and automatic sectional switches, achieving a certain degree of self-healing function. However, most fault location, isolation, and restoration (FLISR) processes still require decision-making by personnel at the distribution control center or rely on centralized self-healing systems. This mode has a strong dependence on the communication network and a long response time. When communication is damaged or dispatching and command fails due to large-scale disasters, the power distribution network lacks autonomous capabilities and is prone to long-term power outages.

[0004] On the other hand, the communication and information infrastructure of the distribution network is becoming increasingly popular, and a large number of intelligent terminals (such as smart meters, feeder terminal units FTU, transformer terminal units TTU, etc.) are distributed throughout the network. This provides the basic conditions for realizing distributed autonomous control. However, at present, the functions of these intelligent terminals are relatively independent and fixed, lacking a coordination mechanism. For example, the sectionalizing switch controller only trips according to the local current criterion, and it is difficult to coordinate the isolation range with adjacent devices in a timely manner; the reactive power compensation devices operate automatically, and may be superimposed on each other, causing over-compensation or under-compensation. How to make the intelligent terminals scattered on-site work collaboratively and be uniformly managed, and give full play to their sensing and control capabilities, is an important direction for the development of intelligent distribution networks.

[0005] In addition, the planning and real-time operation management of the distribution network need to be considered in combination. Taking the access of DER as an example, in the planning stage, it is necessary to evaluate the ability of the distribution network to accommodate new DER and discover weak links in advance; in the operation stage, it is necessary to appropriately restrict the output of DER according to the real-time working conditions to ensure safety. Traditionally, planning and operation are two separate systems, lacking real-time interaction and dynamic adjustment, resulting in either missing the best transformation opportunity or the measures lagging behind the on-site changes.

[0006] Therefore, there is an urgent need for a new technical solution that can integrate communication, autonomous control, and intelligent decision-making, enabling the intelligent terminals in the distribution network to collaborate 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, quickly isolate faults and reconstruct the network to restore power supply in case of abnormalities, and continuously improve its performance through self-learning as the environment and data accumulate. Summary of the Invention

[0007] The object of the present invention is to address the deficiencies in the above background technology and provide an adaptive management method and system for intelligent terminal devices in a distributed distribution network. Through this invention, the intelligent terminal devices in the distribution network can automatically form a communication network, coordinate and control operation parameters in real time, quickly respond to and self-heal faults without centralized manual intervention, and continuously optimize strategies by combining data-driven analysis, thereby achieving high-reliability, high-efficiency, and high-flexibility management of the distribution network operation.

[0008] To achieve the above object, the present invention is realized through the following technical solutions:

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

[0010] Construct a self-organizing communication network: Each intelligent terminal device discovers neighbor nodes through its communication module and establishes a communication connection with the neighbors, thereby forming a decentralized peer-to-peer communication topology. This communication network can support information exchange and status synchronization between terminals and has the ability of topological adaptive evolution.

[0011] Collect and share operating parameters and status information: The intelligent terminal device collects local operating parameters, including electrical quantities such as voltage, current, active power, reactive power, frequency, and device status information such as switch status and fault status. The above information is synchronously uploaded to the central management unit and adjacent intelligent terminal devices through the communication network.

[0012] Execute local control instruction calculation and collaborative exchange: The intelligent terminal device performs control calculations based on the collected status parameters and a preset objective function model. The objective function can include voltage control, power regulation, loss optimization, etc. The control results are exchanged between terminals in the form of control instructions to support local optimization collaboration.

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

[0014] Fault identification and isolation control: Multiple intelligent terminal devices cooperate to judge abnormal operation based on the locally collected data and the information uploaded by neighboring terminals, identify the fault section, and send control instructions to circuit breakers or switch devices according to preset criteria to achieve fault isolation.

[0015] Standby path judgment and power supply restoration: When a local fault causes a power supply interruption, the intelligent terminal device detects whether there is a standby 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 the island operation mode to maintain the power supply capacity.

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

[0017] Model parameter exchange and update: The intelligent terminal device and the central management unit can exchange control model parameters regularly or on demand to achieve model collaborative update. The parameter exchange methods can include central aggregation, local weight feedback, etc.

[0018] Edge control ability in the communication interruption scenario: When the communication of the intelligent terminal device is interrupted, it can still perform control calculations based on the locally stored operation data, record its control behavior and the collected data at the same time, and upload the data to the central management unit after the communication is restored to complete status alignment.

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

[0020] Define the global Hamiltonian function \(H\) of the communication network, and the expression is as follows:

[0021]

[0022] Where: \(N\) is the set of all intelligent terminal device nodes; \(H\) i is the Hamiltonian function of node \(i\), representing the communication cost of the node; \(k\) i is the degree of node \(i\) (i.e., the number of direct neighbors of this node); \(r\) i is the communication radius of node \(i\) (i.e., the signal coverage range of the node); \(A\) ij is the adjacency matrix. If node \(i\) can communicate directly with node \(j\), then \(A\) ij = 1, otherwise it is 0; \(d\) ij is the physical space distance between node \(i\) and node \(j\); \(\alpha\) conn and \(\alpha\) over and \(\alpha\) energy and \(\alpha\) dist are adjustable weight coefficients, respectively representing the cost of the number of node connections, the penalty for over-connection of nodes, the communication energy consumption cost, and the reward for long-distance communication;

[0023] The function takes the connection status, communication radius, and adjacency relationship of each intelligent terminal device as parameters and is used to evaluate the network communication cost;

[0024] The optimization objective is to reduce the overall communication energy consumption and avoid over-connection by adjusting the communication radius and adjacency relationship of each intelligent terminal device, and improve the communication efficiency while maintaining the network connectivity;

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

[0026] Preferably, the process of the intelligent terminal device calculating the local control instruction according to the collected local state data and the preset objective function includes:

[0027] Preset multiple operation objectives, and set an evaluation function and a weight coefficient in mathematical form for each objective. The objectives include but are not limited to minimizing the voltage deviation, frequency stability, reducing power loss, and load balancing;

[0028] Based on the collected local electrical parameters, construct a weighted optimization problem, and solve the joint control variables through a control optimization model. The model supports the definition of constraint conditions and the control of the solution accuracy;

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

[0030] When a change in the operating state causes a certain control target to approach its constraint boundary, the intelligent 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, reorders the control priorities, and executes the updated control calculation process.

[0031] Preferably, when the intelligent terminal device executes the local preset objective function to calculate the local control instruction, it further includes adaptively and dynamically adjusting the control step size, specifically including:

[0032] Analyze the change trend of the control variable in two consecutive control cycles, calculate the consistency between the change direction of the control variable in the current cycle and that in the previous cycle, and construct a similarity index;

[0033] The similarity index is constructed based on the vector angle relationship of the control variable change, reflecting the continuity of the control direction;

[0034] Adjust the control step size according to this index: when the control direction continuity is strong, enlarge the step size to accelerate the response speed; when the control direction is unstable, reduce the step size to avoid control oscillation;

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

[0036] Preferably, the sending of the opening and closing instruction controls the connected circuit breaker or switch, specifically including:

[0037] When a fault occurs, the intelligent terminal devices located upstream and downstream of the fault section respectively detect the electrical parameters such as the voltage amplitude, current direction, and zero-sequence current mutation in their respective areas;

[0038] The intelligent terminal device sends a fault identification signal through communication, and judges the line section where the fault is located based on the received result;

[0039] When it is judged as an instantaneous fault, the intelligent terminal device configured with the reclosing function outputs a closing control instruction after a set delay, and notifies the adjacent intelligent terminals to prepare for the power restoration operation;

[0040] If a fault is still detected after the reclosing operation, it is judged as a permanent fault, and the adjacent intelligent terminal device issues a disconnection control instruction to control the affiliated 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 intelligent terminal device, and the response time is 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 intelligent terminal device performs power restoration control, which specifically includes:

[0043] After the fault section is isolated, the intelligent terminal device at the end of the power outage area determines whether there is a standby power path and detects the status information and voltage conditions of the tie switch on the standby path;

[0044] 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;

[0045] After receiving the request, the tie switch control terminal performs voltage matching verification and executes the closing control of the tie switch when the conditions are met, so that the power outage area restores power supply through the adjacent feeder;

[0046] If there is no standby path, the intelligent terminal device of the distributed power source in the power outage area switches to the island operation mode according to the operation mode switching instruction, and establishes a local microgrid to maintain the power supply of critical loads;

[0047] The intelligent terminal device performs load shedding, peak shaving or power limiting control according to the preset load priority order, so as to suppress voltage over-limit or load overload;

[0048] The power restoration process is coordinated and completed through peer-to-peer communication between intelligent terminal devices. After the power supply state is stable, the intelligent terminal device uploads the state information of the restoration process to the central management unit.

[0049] Preferably, the central management unit uses the rolling simulation analysis method to evaluate the acceptance capacity, which specifically includes:

[0050] The central management unit regularly collects the topological structure, node device parameters, load data and distributed power output of the distribution network, and establishes a distribution network power flow calculation model;

[0051] Set multiple distributed power access scenarios, gradually increase the access capacity in each scenario and perform power flow simulation calculations, and record the minimum access value that causes the node voltage or line current to exceed the limit as the node acceptance capacity critical value;

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

[0053] When the distributed power output of a certain node is close to its acceptance capacity critical value, the central management unit issues an output adjustment instruction to the intelligent terminal device corresponding to the node to control it to reduce the active power output or adjust the reactive power output;

[0054] 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 an expansion suggestion to the power supply planning system when necessary;

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

[0056] Preferably, the control model relied on by the capacity evaluation realizes parameter sharing through a federated learning mechanism, specifically including:

[0057] 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 identification;

[0058] After each round of training, the intelligent terminal device extracts the model parameters or gradient information and sends them to the central management unit through peer-to-peer communication;

[0059] After receiving the model parameter information, the central management unit performs an aggregation operation, and the aggregation methods include weighted average, maximum likelihood estimation or adaptive fusion strategy, which are used to generate unified global model parameters;

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

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

[0062] In the federated learning process, the original operation data is not transmitted, thus avoiding the leakage of the original data.

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

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

[0065] During the communication interruption, the intelligent terminal device continuously calculates the control variables according to the set control period and outputs control instructions, and independently completes the operation state determination and control;

[0066] After the communication is restored, the intelligent terminal device uploads the measurement data, control behaviors and fault events recorded during the interruption to the central management unit in a unified manner;

[0067] The uploaded data includes timestamps, changes in control variables, device action records and fault flags, etc., which are used for the central management unit to supplement data and align historical records.

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

[0069] The intelligent terminal device includes:

[0070] A data acquisition module, configured to acquire the voltage, current, active power, reactive power, and device operation status of this node;

[0071] A communication module, used for two-way data communication between terminals and between the terminal and the central management unit, supporting the establishment and dynamic maintenance of an ad hoc network topology;

[0072] A control execution module, configured to operate the connected circuit breaker, voltage regulator, inverter, or other power execution devices according to control instructions;

[0073] A local control and decision-making module, configured to receive the acquired data and communication information, generate control decisions, and output control instructions;

[0074] The central management unit includes:

[0075] A communication interface, used to establish a data connection with each intelligent terminal device and exchange operation information;

[0076] A global database, used to store the distribution network topology information, device parameters, operation data, and control records;

[0077] A dispatching decision-making module, configured to perform tasks such as power flow analysis, acceptance capacity assessment, control strategy generation, and model parameter aggregation, and issue dispatching instructions to each intelligent terminal device;

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

[0079] Compared with the prior art, the beneficial effects of the present invention are as follows: By establishing a self-organizing peer-to-peer communication network, the intelligent terminals can automatically discover neighbors and establish connections, constructing a decentralized communication architecture. Even in the case of communication link interruption, local control tasks can still be maintained, significantly improving the fault tolerance and operation stability of the system. Each intelligent terminal collects local electrical operation parameters and synchronously uploads them to the central management unit and adjacent terminals, forming a status data sharing mechanism, which not only enhances the judgment consistency of boundary coordination between terminals but also improves the data redundancy ability, providing support for abnormal state recovery and strategy correction. Each terminal performs multi-objective optimization control calculations based on the local real-time working conditions, and through collaborative interaction with neighboring terminals, realizes multi-objective coordinated control such as voltage regulation, power loss minimization, and load balancing, taking into account both power quality and operation efficiency. During the execution of the control strategy, the terminal can dynamically adjust the parameters of the optimization model according to the feedback results, such as control step size, target weight, etc., to cope with complex working conditions such as load fluctuations and power generation disturbances, realizing strategy self-adaptation adjustment and improving control robustness. When a fault occurs, the intelligent terminal can quickly complete fault location and isolation based on local abnormal electrical quantities and neighbor communication, and then automatically execute power supply path switching or enter the island operation mode. The entire response process is controlled within 0.5 to 1 second, greatly shortening the power outage time and improving power supply reliability. The central management unit periodically evaluates the distributed power acceptance margin of each node and dynamically adjusts the access strategy and operation and maintenance configuration in combination with the operation status, so as to realize the optimal allocation and economic operation of the 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 to continuously optimize the global control model, improving the prediction accuracy and control effect without revealing the original data, enabling the control strategy to have the ability of long-term evolution and supporting the intelligent evolution goal of the system to be more efficient as it runs longer. BRIEF DESCRIPTION OF THE DRAWINGS

[0080] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for the description of the embodiments will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0081] Among them:

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

[0083] Figure 2 is a schematic diagram of the overall system architecture of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[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 of the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the described embodiments of the present invention fall within the scope of protection of the present invention.

[0085] In view of the problems in the current distribution network such as slow terminal control response, lack of coordination mechanism, strong dependence on central computing, and poor adaptability, the present invention proposes an adaptive management method based on the integration of self-organizing communication, local control, edge coordination, and intelligent learning. This method mainly includes the following eight steps:

[0086] Embodiment 1

[0087] As Figure 1 shown, it is an embodiment of the present invention, which provides an adaptive management method for a distributed distribution network intelligent terminal device, including:

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

[0089] After the system is deployed, each intelligent terminal device starts the node discovery process through the communication module. The terminal detects neighboring communicable terminals by broadcasting its own address information, device identifier, access capabilities, etc., and completes the establishment of a preliminary communication link. The terminal adjusts the optimized connection according to the link quality (RSSI, link delay, etc.) to form a self-organizing peer-to-peer communication network covering the entire network. This network structure does not rely on a central coordinator and has the following advantages:

[0090] The multi-path routing mechanism improves the anti-interference ability;

[0091] The network topology is dynamically adjustable and has scalability;

[0092] Each terminal node serves as a basic unit for communication and control coordination and has the full-process ability of "discovery - connection - reconstruction".

[0093] Step 2: Local status collection and information sharing

[0094] In this embodiment, each intelligent terminal device is equipped with a data collection module for regularly collecting the electrical operation parameters of the local distribution node, including voltage, current, active power, reactive power, frequency, voltage harmonics, etc., as well as the operation status of the connected devices, such as switch position signals, energy storage status information, etc.

[0095] The collected data is uploaded to the following at a set period:

[0096] The central management unit: for analyzing the status of the entire network, evaluating the capacity, and generating strategies;

[0097] Adjacent intelligent terminal devices: used for boundary state perception and coordinated control.

[0098] This mechanism realizes the dual-channel fusion of spatial redundancy of state perception and communication coordination, significantly improving the sensitivity and local perception accuracy of the system.

[0099] Step 3: Independent control calculation and coordinated control execution

[0100] Based on the acquired local state information, the intelligent terminal device independently constructs an objective function and executes control optimization calculations according to preset operation goals (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 (such as ±5%);

[0103] Preferentially use local energy storage resources to reduce losses;

[0104] The control strategy takes into account the load characteristics and time period priorities.

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

[0106] Local execution: Drive the circuit breaker, voltage regulator, and inverter through the control module;

[0107] Coordinated execution: Send boundary variables to adjacent terminals through the peer-to-peer communication protocol to coordinately adjust strategies such as load transfer and reactive power response.

[0108] This mechanism is significantly different from the traditional centralized strategy distribution mode, emphasizing the horizontal communication between terminals and the local intelligent coordinated control mechanism, forming a regional autonomous closed loop.

[0109] Step 4: Control effect evaluation and algorithm adaptive adjustment

[0110] After each round of control, the terminal will evaluate the state changes before and after the control execution. The main evaluation indicators include: the achievement degree of the control objective; the convergence speed of the state variable; the adjustment range of the control quantity.

[0111] If it is found that the control behavior has low efficiency or causes local oscillations, the terminal will automatically adjust the control algorithm parameters (such as target weight, control step size, switching strategy threshold, etc.) to achieve adaptive optimization of the control algorithm. This mechanism endows the terminal with a certain degree of strategy self-evolution ability, improving the long-term operation efficiency of the system.

[0112] Step 5: Fault detection and isolation control

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

[0114] If a terminal detects an abnormality, it broadcasts a fault identification signal to neighboring terminals through the communication module. After multi-terminal comparison, through cross-verification of location inference and abnormal characteristics, the fault section is accurately located. The system follows the preset logic: commands the upstream terminal of the fault to disconnect the switch; synchronously commands the downstream terminal of the fault to isolate and control;

[0115] Complete the electrical isolation operation of the fault section.

[0116] This mechanism is completely completed by the terminals without the participation of the center, has high responsiveness, and the isolation processing delay is controlled within 1 second.

[0117] Step 6: Power supply restoration and transfer control

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

[0119] Detect whether there is a standby power supply path (such as tie switches, distributed power sources);

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

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

[0122] The restoration strategies include:

[0123] Start the closing of the tie switch to achieve power transfer from adjacent feeders;

[0124] Start the local microgrid islanding control to maintain the operation of critical loads;

[0125] Implement load cycling control (load shedding, peak shaving, power rationing) to ensure local voltage stability.

[0126] The entire restoration process is coordinated and completed through peer-to-peer communication between terminals, and the terminals report the process to the center for subsequent maintenance and statistics.

[0127] Step 7: Capacity assessment and access strategy optimization

[0128] The central management unit regularly evaluates the accessible capacity of distributed power sources (such as photovoltaic) at each node of the distribution network using the rolling simulation analysis method. The evaluation methods include:

[0129] Model the current topology of the distribution network;

[0130] Set access scenarios to simulate different power growths;

[0131] Execute power flow simulation and record the over-limit points of node voltages / line currents;

[0132] Obtain the critical value of the acceptance capacity for each node;

[0133] If it is found that the access of a certain node is close to the upper limit, the central management unit sends a power limit instruction to the terminal to which the node belongs, and synchronously adjusts the regional power flow strategy to achieve the optimal configuration of source-load balance.

[0134] Step 8: Federal learning-driven control strategy evolution

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

[0136] Each terminal trains a control model (such as a neural network, regression model, etc.) based on local historical operation data;

[0137] After the model training is completed, extract the parameters (not the original data) and send them to the central management unit;

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

[0139] The aggregated parameters are redistributed to the terminals to replace or initialize the next round of model training;

[0140] This process is iterated to enable the system to achieve continuous optimization and improvement of local generalization ability, and to avoid the transmission of original data, ensuring the privacy and security of data on the terminal side.

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

[0142] To further improve the stability, scalability, and local autonomy of the communication network of the intelligent 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 rely on a fixed central controller or a preset topology structure, but through a topology evolution mechanism based on the Hamiltonian optimization algorithm, and each terminal device independently completes connection optimization and network structure reconstruction.

[0143] I. Initial discovery and connection establishment

[0144] In the initial deployment of the system or the initial stage of operation, after each intelligent terminal device is powered on, it starts the neighbor discovery and connection evaluation process. Each terminal broadcasts information such as its own identifier, geographical location coordinates, communication power level, and the number of connections that can be carried through its communication module to form a neighbor candidate set.

[0145] In the case of no connection rule constraints, all terminals may tend to preferentially connect to nodes with strong power and central positions (such as the master station A terminal), forming an unbalanced "star structure", which will lead to problems such as communication congestion, high master station load, and prominent single-point failure risks in large-scale networks.

[0146] To overcome the above problems, the present invention introduces a Hamiltonian optimization function as the basis for evaluating communication connections, and each terminal independently calculates the "system cost function" of its current connection structure:

[0147]

[0148] Where: N is the set of all intelligent terminal device nodes; H i is the Hamiltonian function of node i, representing the communication cost of the node; k i is the degree of node i, that is, the number of direct neighbors of the node; r i is the communication radius of node i, that is, the signal coverage range of the node; A ij is the adjacency matrix. If node i and node j can communicate directly, then A ij = 1, otherwise it is 0; d ij is the physical space distance between node i and node j; α conn 、α over 、α energy 、α dist are adjustable weight coefficients, respectively representing the cost of the number of node connections, the penalty for over-connection of nodes, the communication energy consumption cost, and the reward for long-distance communication;

[0149] The function takes the connection status, communication radius, and adjacency relationship of each intelligent terminal device as parameters and is used to evaluate the network communication cost;

[0150] The optimization goal is to reduce the overall communication energy consumption and avoid over-connection by adjusting the communication radius and adjacency relationship of each intelligent terminal device, and to improve the communication efficiency while maintaining network connectivity;

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

[0152] This function comprehensively evaluates the connection quality of the terminal from four dimensions: distance minimization, connection number balance, load balance, 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 performs the following connection adjustment actions through the local decision-making 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 the connection Then perform connection establishment.

[0160] 2. Connection disconnection strategy:

[0161] For the current connection j, if it satisfies:

[0162] There is an alternative path in its redundant path;

[0163] Or the connection causes H i To exceed the set threshold; then the terminal actively disconnects the connection.

[0164] 3. Connection power adjustment:

[0165] If a certain terminal (such as terminal A) detects that the number of connections is too large (such as > 20), it automatically reduces the broadcast power and reduces the connections to non-critical terminals;

[0166] While the terminals located at the network edge or connection points (such as type C) increase the transmission power to enhance the connection coverage.

[0167] Based on the independent decision-making of each terminal, this mechanism realizes the self-optimization of the communication topology of the entire network. The final network form is as follows: a "chain + network" fusion structure is formed inside the region, and the average connection degree of nodes is maintained between 2 and 4;

[0168] The master station terminal is directly connected to the beginning and end of each feeder for easy aggregation;

[0169] Horizontal connections are built between each feeder by connection terminals to achieve a single connected cluster.

[0170] III. Technical advantage analysis

[0171] The Hamiltonian optimization mechanism of the communication network adopted by the present invention has the following technical innovation points compared with the fixed communication topology or centralized configuration method in traditional distribution automation:

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

[0173] Quantify connection evaluation indicators: Integrate communication quality, load balancing, and path diversity through the Hamiltonian function to ensure the optimization of network performance;

[0174] Support dynamic reconstruction and redundancy adjustment: The terminal can automatically adjust the connection when neighbors change, faulty nodes appear, or channel conditions fluctuate, with strong fault tolerance;

[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 shows good connectivity (connectivity rate > 99%), low average hop count (< 3 hops), and fault tolerance (the network remains connected under the failure of 1 - 2 nodes) in actual measurements, providing a reliable basic support for subsequent control collaboration and data sharing.

[0177] In a preferred embodiment of the present invention, when the intelligent terminal device executes control instruction calculation, a multi-objective weighted optimization model is adopted, considering multiple operating objectives such as voltage stability, frequency regulation, power loss reduction, and load balancing.

[0178] The specific implementation steps are as follows:

[0179] 1. Preset the set of objective functions:

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

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

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

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

[0184] 2. Set weight coefficients w1, w2, w3, w4 for each of the above objective functions to form the total objective function:

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

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

[0187] Control variable boundaries (such as the upper and lower limits of transformer tap positions, the output range of inverters); node voltage constraints (such as ±5%); operation limits such as load power factor;

[0188] 4. The control variables include: the position of the transformer tap; active / reactive power output; the switching state of capacitor banks; the operation mode of energy storage / photovoltaic inverters, etc.

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

[0190] This mechanism enables the terminal to flexibly reconstruct the control strategy according to the actual operation without relying on the center, significantly improving the adaptive ability and refinement level of the control.

[0191] To improve the control convergence efficiency and avoid oscillations, the present invention further introduces a control step size adaptive adjustment mechanism into the terminal local control algorithm to replace the fixed step size setting method used in traditional control algorithms, and overcome the problems of slow response or easy oscillation when facing a fast-changing or disturbed environment.

[0192] The basic idea of this adjustment mechanism is: during each round of local control optimization executed by the intelligent terminal, the change trend of control variables (such as voltage setting values, active / reactive power reference values, etc.) is dynamically analyzed to evaluate the consistency of the current control action with the previous cycle's control action in terms of direction. This consistency can be reflected by calculating the angular relationship between the change vectors of control variables in two consecutive cycles. The closer the value is to 1, the higher the control direction consistency, and the closer the value is to -1, the opposite the direction.

[0193] Based on the above control direction similarity index, the intelligent terminal makes a segmented adjustment of the control step size according to its numerical range: when it detects a high degree of consistency in the control direction change, it is considered that the current optimization direction is reliable, and the step size can be appropriately increased to improve the convergence speed;

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

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

[0196] To achieve the above goals, the system preset step size increase factor, decrease factor, and similarity threshold parameters. After each control cycle ends, the intelligent terminal automatically performs the above judgment and step size update according to the local state data, thereby realizing the local control parameter adjustment ability that does not depend on the central coordination and adapts to the actual operating state.

[0197] This mechanism has the following technical advantages:

[0198] Strong response adaptability: The control step size can be dynamically adjusted according to the running trend without manual intervention;

[0199] Improve the convergence efficiency: It can quickly expand the control rhythm under continuous operating conditions and improve the efficiency;

[0200] Suppress oscillation behavior: By appropriately reducing the step size, problems such as overshoot and oscillation can be effectively prevented;

[0201] Strong edge implementability: The entire strategy can be executed relying on local variables and historical information, and is suitable for the edge intelligent terminal deployment environment.

[0202] This improvement solves the problem that the fixed control step size in the traditional distributed control system is likely to cause the inability to balance "slow response" and "poor stability", reflects the technical breakthrough at the control strategy level, and has a significant effect on enhancing the stable operation ability of the system of the present invention under dynamic working conditions.

[0203] When a fault occurs in a certain line of the distribution network (such as single-phase grounding, short circuit, etc.), multiple intelligent terminal devices cooperate to complete the fault section location and isolation operation based on local electrical parameters. The specific process is as follows:

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

[0205] The terminal that detects the abnormality immediately sends a fault identification signal to the adjacent terminal through peer-to-peer communication;

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

[0207] If it is judged as an instantaneous fault, the terminal configured with the reclosing function issues a closing command with a time delay;

[0208] If the reclosing fails, it is determined as a permanent fault, and the corresponding terminal executes a disconnection command to control the circuit breaker or switch to isolate the fault section;

[0209] The whole process is automatically completed by the terminal, and the response time is controlled within 0.5 to 1 second without manual intervention.

[0210] This mechanism supports a three - level closed - loop structure of local judgment + regional collaboration + rapid isolation, significantly enhancing the system's self - healing ability and operation continuity.

[0211] After fault isolation is completed, the terminal automatically starts the power supply restoration control process, and the restoration process has the capabilities of hierarchical judgment and autonomous collaboration:

[0212] The terminal at the end detects the power loss on its own side and determines whether there is an alternative power supply path (such as a tie switch);

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

[0214] When it is judged that the voltages are synchronized or approximately the same, the terminal sends a closing request to the tie - switch control terminal;

[0215] The control terminal completes the voltage matching verification. If the conditions are met, it executes the remote closing control;

[0216] If restoration cannot be achieved through the tie - line method, the backup power supply (such as a local distributed power source) is called, and according to the operation strategy, it switches to the island operation mode, constructs a local micro - grid, and supplies power to critical loads;

[0217] According to the load priority, strategies such as peak shaving and valley filling, load shedding and power rationing are executed to control voltage over - limit and overload;

[0218] After the power supply restoration is completed, the terminal uploads the restoration process information to the central management unit for dispatching reference and data archiving.

[0219] This process has the characteristics of "autonomous judgment, collaborative execution, and diverse paths", adapting to various network faults and restoration path selection requirements.

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

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

[0222] Set multiple scenarios for new distributed power source access;

[0223] In each scenario, gradually increase the access power (such as increasing by 0.1 MW each time), and execute power flow simulation;

[0224] Record the minimum access capacity when the node voltage exceeds the limit or the line current is overloaded as the upper limit of the acceptance capacity of this node;

[0225] Repeat the simulation for all key nodes in the network to form a complete acceptance capacity distribution table;

[0226] If the output of a certain node is close to its acceptance capacity boundary, a power adjustment instruction is sent to its corresponding terminal.

[0227] If the overall margin of the region is insufficient, adjust the power flow dispatching or provide capacity expansion suggestions to the power supply planning department.

[0228] This mechanism ensures that the security boundaries of the power source-load-network of the distribution network are not breached, and supports the friendly access of distributed energy.

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

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

[0231] The central management unit performs aggregation calculations on the parameters uploaded by each terminal. The methods include: weighted average (FedAvg); maximum likelihood estimation; adaptive fusion strategy; generate unified global model parameters and reissue them to each terminal for use as the initialization parameters for the next round of local training.

[0232] This process is iteratively executed according to the round setting or performance convergence criterion.

[0233] This method improves the generalization ability of the model and the control intelligence of the terminal while ensuring privacy and security.

[0234] To ensure the high availability of the system, the intelligent terminal of the present invention has edge autonomy capabilities and supports continuously completing control and model training tasks during communication interruption, specifically including:

[0235] After detecting communication disconnection, the terminal enters the autonomous operation state;

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

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

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

[0239] After communication is restored, the terminal uploads the operation records during the interruption period uniformly, including timestamps, control variable changes, switch action records, and fault handling logs;

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

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

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

[0243] A data acquisition module for collecting the voltage, current, active power, reactive power, and device operating status of this node;

[0244] A communication module for performing two-way data communication between terminals and between terminals and the central management unit, supporting the establishment and dynamic maintenance of self-organizing network topologies;

[0245] The control execution module is used to operate the connected circuit breaker, voltage regulator, inverter, or other power execution devices according to control instructions; The local control and decision-making module is used to receive the collected data and communication information, generate control decisions, and output control instructions;

[0246] The central management unit includes:

[0247] A communication interface for establishing a data connection with each intelligent terminal device and exchanging operation information;

[0248] A global database for storing distribution network topology information, device parameters, operation data, and control records;

[0249] A scheduling decision-making module for performing tasks such as power flow analysis, acceptance capacity assessment, control strategy generation, and model parameter aggregation, and issuing scheduling instructions to each intelligent terminal device;

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

[0251] Embodiment 2

[0252] Application of Step Size Adaptive Regulation Based on Control Direction Similarity in Feeder Voltage Control

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

[0254] The intelligent terminal is deployed at the end of the downstream branch of the distribution transformer, with the capabilities of local voltage measurement, communication, and control. The control objective is to stabilize the access point voltage between 10.5 kV and 11 kV. The terminal executes local voltage optimization regulation with a control period of 1 second and automatically adjusts the control step size according to the direction relationship of the front and back control actions in each round of regulation.

[0255] The specific process is as follows:

[0256] Historical variable acquisition: Before the start of the t-th control period, the intelligent terminal has recorded the control variable values v t-2 , v t-1 for subsequent differential calculation;

[0257] Current control quantity calculation: The terminal collects the current voltage v t , and calculates the change in the control variable between adjacent periods:

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

[0259] Similarity index calculation: The terminal calculates the control direction similarity index according to the direction relationship of the above two differential vectors:

[0260]

[0261] In the example, if S cos = 0.92, it means that the control directions are highly consistent;

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

[0263] The current S cos = 0.92 > θ high , meeting the condition for accelerating the step size. The terminal adjusts the current step size to: α t+1 = γ up × α t = 1.05 × α t ; Output control quantity: According to the updated step size α t+1 , recalculate the next round of voltage setting target and generate a control instruction to act on the local voltage regulation device (such as a voltage regulator or reactive power compensation equipment).

[0264] Technical effects: In a typical load area with frequent but stable-direction voltage fluctuations, this method can quickly expand the control step size and improve the response rate; in scenarios where the direction is unstable or the disturbance is intense (such as load mutation), the step size can be promptly contracted to prevent overshoot and oscillation during regulation; the entire mechanism is executed based on local state data and does not rely on the scheduling of a central server, having good edge autonomy capabilities.

[0265] In summary, the method of the present invention covers the entire cycle of distribution network operation, from the optimal control in the normal state to the isolation and restoration in the fault state, and is continuously improved through a learning mechanism. Each step can operate independently or cooperate with each other to form a closed-loop autonomous management system, significantly improving the safety (fast fault response, preventing the spread of accidents), reliability (less or zero power outages), economy (reduced losses, reduced voltage drop), and flexibility (adapting to changing supply and demand and equipment status) of distribution network operation. In particular, the present invention makes full use of the existing intelligent terminal resources on-site to achieve "local decision-making and local execution", greatly reducing the dependence on centralized communication and computing, having better robustness and scalability, and being able to adapt to the development trend of future distribution networks with high distribution and high proportion of new energy access.

[0266] In the description of this specification, the description with reference to terms such as "one embodiment", "some embodiments", "example", "specific example", or "some examples" means that the specific features, structures, materials, or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present application. Moreover, the specific features, structures, materials, or characteristics described can be combined in a suitable manner in any one or more embodiments or examples. In addition, without contradiction, those skilled in the art can combine and combine the different embodiments or examples described in this specification and the features of different embodiments or examples.

[0267] Any process or method description shown in the flowchart or described in other ways herein can be understood as representing a module, segment, or part of code including one or more executable instructions for implementing a specific logical function or process. And the scope of the preferred embodiments of the present application includes additional implementations, where the functions can be executed in a substantially simultaneous manner or in the reverse order according to the involved functions, rather than in the order shown or discussed.

[0268] The above is only the specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present application can easily think of various changes or substitutions, and these should all be covered within the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. An adaptive management method for intelligent terminal devices of a distributed distribution network, characterized in that, Including the following steps: The intelligent terminal device discovers neighbor nodes through the communication module, establishes a communication connection, and forms a self-organizing peer-to-peer communication network; the intelligent terminal device collects local operation parameters and device status, and uploads them to the 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 status data and the preset objective function, and transmits control-related data to adjacent intelligent terminal devices through the communication network; The intelligent terminal device adjusts the control algorithm parameters according to the change direction of the control variables in the current control period and the previous control period; multiple intelligent terminal devices identify the fault section based on their respective collected operation parameters and the received neighbor information, and send opening and closing instructions to control the connected circuit breaker or switch; The intelligent terminal device detects the standby power supply path information and voltage conditions, and sends control signals to the tie switch terminal or distributed power source terminal to control closing or operation mode switching; The central management unit collects network topology, node parameters, load data, and distributed power output information, performs power flow simulation, and evaluates the acceptance capacity; The central management unit exchanges model parameters with the intelligent terminal device to update the control model; During the communication interruption, the intelligent terminal device performs control calculations based on the locally stored data and records the 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 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 the Hamiltonian function, specifically including: Defining the global Hamiltonian function H of the communication network, and the expression is: Where: N is the set of all intelligent terminal device nodes; H i is the Hamiltonian function of node i, representing the communication cost of the node; k i is the degree of node i, that is, the number of direct neighbors of the node; r i is the communication radius of node i, that is, the signal coverage range of the node; A ij is the adjacency matrix. If node i and node j can communicate directly, then A ij = 1, otherwise it is 0; d ij is the physical space distance between node i and node j; α conn 、α over 、α energy 、α dist are adjustable weight coefficients, representing the cost of the number of node connections, the penalty for over-connected nodes, the communication energy consumption cost, and the reward for long-distance communication respectively; The function takes the connection status, communication radius, and adjacency relationship of each intelligent terminal device as parameters, and is used to evaluate the network communication cost.

3. The adaptive management method of the distributed distribution network intelligent terminal device according to claim 1, characterized in that The process that the intelligent terminal device calculates local control instructions according to the collected local status data and the preset objective function includes: Presetting multiple operation objectives, and setting an evaluation function and a weight coefficient in mathematical form for each objective. The objectives include but are not limited to voltage deviation minimization, frequency stability, power loss reduction, and load balancing; Based on the collected local electrical parameters, constructing a weighted optimization problem, and solving the joint control variables through a control optimization model. The model supports constraint condition definition and solution accuracy control; The control variables include but are not limited to: adjusting the tap position of the transformer, adjusting the active / reactive power output, switching the capacitor bank, and controlling the operation state of the inverter; When the change in the operation state causes a certain control objective to approach its constraint boundary, the intelligent terminal device or the central management unit adjusts the weight coefficient of the corresponding objective according to the change amplitude of the state and the risk index, reorders the control priorities, and executes the updated control calculation process.

4. The adaptive management method of the distributed distribution network intelligent terminal device according to claim 3, characterized in that, When the intelligent terminal device executes the local preset objective function to calculate local control instructions, it further includes adaptively dynamically adjusting the control step size, specifically including: Analyzing the change trend of the control variables in two consecutive control periods, calculating the consistency between the change direction of the control variables in the current period and the previous period, and constructing a similarity index; The similarity index is constructed based on the vector included angle relationship of the control variable changes, and reflects the continuity of the control direction; Adjust the control step size according to this index: when the continuity of the control direction is strong, enlarge the step size to accelerate the response speed; when the control direction is unstable, reduce the step size to avoid control oscillation; The step size adjustment follows a segmented adjustment mechanism, and different step size adjustment coefficients are adopted according to the similarity index value to maintain the responsiveness and convergence of the control algorithm.

5. The adaptive management method of the distributed distribution network intelligent terminal device according to claim 1, characterized in that The sending of the opening and closing instruction controls the connected circuit breaker or switch, specifically including: When a fault occurs, the intelligent terminal devices located upstream and downstream of the fault section respectively detect the voltage amplitude, current direction, and sudden change of zero-sequence current electrical parameters in their respective regions; The intelligent terminal device sends a fault identification signal through communication and judges the line section where the fault is located based on the received result; When it is judged as an instantaneous fault, the intelligent terminal device with the reclosing function outputs a closing control instruction after a set delay and notifies the adjacent intelligent terminal to prepare for the power restoration operation; If the fault is still detected after the reclosing operation, it is judged as a permanent fault, and the adjacent intelligent terminal device issues a disconnection control instruction to control the affiliated 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 intelligent terminal device, and the response time is between 0.5 seconds and 1 second after the fault occurs, and the process does not rely on manual operation.

6. The adaptive management method of the distributed distribution network intelligent terminal device according to claim 5, characterized in that, After the fault isolation is completed, the intelligent terminal device performs power restoration control, specifically including: After the fault section is isolated, the intelligent terminal device at the end of the power outage area judges whether there is a standby power path and detects the status information and voltage conditions 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; after receiving the request, the tie switch control terminal performs voltage matching verification and executes the closing control of the tie switch when the conditions are met, so that the power outage area restores power through the adjacent feeder; If there is no standby path, the intelligent terminal device of the distributed power source in the power outage area switches to the island operation mode according to the operation mode switching instruction, and establishes a local microgrid to maintain the power supply of critical loads; The intelligent terminal device performs load shedding, peak shaving or power limit control according to the preset load priority order, so as to suppress voltage over-limit or load overload; The power restoration process is coordinated and completed through peer-to-peer communication between intelligent terminal devices. After the power supply state is stable, the intelligent terminal device uploads the power restoration process status information to the central management unit.

7. The adaptive management method of the distributed distribution network intelligent terminal device according to claim 1, characterized in that, The central management unit uses the rolling simulation analysis method to evaluate the acceptance capacity, specifically including: The central management unit regularly collects the topological structure, node device parameters, load data and distributed power output of the distribution network, and establishes a distribution network power flow calculation model; Set multiple distributed power access scenarios, gradually increase the access capacity in each scenario and perform power flow simulation calculations, and record the minimum access value that causes the node voltage or line current to exceed the limit as the node acceptance capacity critical value; Repeat the above process for multiple target nodes to generate an acceptance capacity distribution data table of each node under the current operating conditions; When the output of a distributed power source at a certain node approaches the critical value of its acceptance capacity, the central management unit issues an output adjustment instruction to the intelligent terminal device corresponding to that node to control it to reduce the active power output or adjust the 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 adjacent areas and outputs an expansion suggestion to the power supply planning system when necessary; The capacity assessment process is executed periodically to update the capacity boundary parameters of each node.

8. The adaptive management method of the distributed distribution network intelligent terminal device according to claim 7, characterized in that The control model relied on by the capacity assessment realizes parameter sharing through the federated learning mechanism, specifically including: Each intelligent terminal device trains a machine learning model based on local historical data, and the model is used for load forecasting, voltage trend judgment or abnormal state identification; After each round of training, the intelligent terminal device extracts the model parameters or gradient information and sends them to the central management unit through peer-to-peer communication; After receiving the model parameter information, the central management unit performs an aggregation operation, and the aggregation methods include weighted average, maximum likelihood estimation or adaptive fusion strategy, which are used to generate unified global model parameters; The global model parameters are sent to the intelligent terminal device and used as the initial parameters for the next round of local training or replace the current model parameters; The model training and aggregation process are iteratively executed according to the set number of rounds or convergence criteria; In the federated learning process, the original operation data is not transmitted, thus avoiding the leakage of the original data.

9. The adaptive management method of the distributed distribution network intelligent terminal device according to claim 8, wherein The intelligent terminal device has edge autonomy capabilities to maintain local model training and control functions during communication interruptions, specifically including: When the communication with the central management unit is interrupted, the intelligent terminal device performs node data collection, status monitoring, local control and fault handling operations based on the operation strategies and configuration parameters stored locally; During the communication interruption, the intelligent terminal device continuously calculates the control variables according to the set control period and outputs control instructions to complete the operation state judgment and control by itself; After the communication is restored, the intelligent terminal device uploads the measurement data, control behaviors and fault events recorded during the interruption to the central management unit in a unified manner; The uploaded data includes timestamps, control variable changes, device action records and fault identifiers, which are used by the central management unit for data supplementation and historical record alignment.

10. An adaptive management system for the adaptive management method of a distributed distribution network intelligent terminal device according to any one of claims 1-9, characterized in that, It includes multiple intelligent terminal devices deployed at the nodes of the distribution network and a central management unit, where: The intelligent terminal device includes: A data collection module for collecting the voltage, current, active power, reactive power and device operation status of this node; A communication module for performing two-way data communication between terminals and between the terminal and the central management unit, supporting the establishment and dynamic maintenance of a self-organizing network topology; A control execution module for operating the connected circuit breaker, voltage regulator, inverter or other power execution devices according to control instructions; A local control and decision-making module for receiving the collected data and communication information, generating control decisions and outputting control instructions; The central management unit includes: A communication interface for establishing a data connection with each intelligent terminal device and exchanging operation information; A global database for storing distribution network topology information, device parameters, operation data and control records; The dispatching decision-making module is used to perform tasks such as power flow analysis, acceptance capacity assessment, control strategy generation, and model parameter aggregation, and issue dispatching instructions to each intelligent terminal device; The central management unit and the intelligent terminal devices form a hierarchical and collaborative distributed control system. The central management unit provides global dispatching and optimization instructions, and each intelligent terminal device independently executes local control, communication coordination, status monitoring, fault detection, and recovery processing.

Images