A fluctuation coordinated distributed smart healing power distribution network method and system

Abstract

The application provides a fluctuation coordination distributed intelligent healing power distribution network method and system, the method comprises the following steps: collecting power distribution network operation data and distributed energy dynamic characteristic data in real time; calculating distributed energy real-time output power and line voltage deviation based on a distributed energy dynamic output model; triggering a multi-agent negotiation mechanism if the line voltage deviation exceeds a threshold; each agent selects the optimal coordination strategy through message passing and strategy updating; and performing self-healing control operation. By collecting distributed energy dynamic characteristic data in real time and establishing an accurate dynamic output model, the power output of intermittent power sources can be accurately predicted, and the problem of insufficient consideration of the fluctuation of distributed energy in the prior art is effectively improved.

Description

Technical Field

[0001] This invention relates to the field of power system control technology, specifically to a distributed intelligent distribution network method and system for fluctuation coordination. Background Technology

[0002] Existing self-healing technologies for power distribution systems are primarily based on traditional centralized control architectures, achieving automated fault handling through a fault detection, location, isolation, and recovery (FLISR) process. For example, traditional self-healing schemes based on reclosers and sectionalizers can achieve basic fault isolation functions, but their location accuracy is limited in complex fault scenarios. In recent years, with the development of artificial intelligence technology, deep learning-based fault prediction methods have been widely used. These methods can train models using historical data to predict potential faults and formulate preventative maintenance strategies in advance.

[0003] However, while existing self-healing technologies for power distribution systems have achieved fault prediction based on deep learning, centralized self-healing control, and distributed intelligent terminal collaboration, these solutions are mainly designed for traditional power distribution networks and are not fully adapted to scenarios with a high proportion of distributed energy resources. Specifically, these issues manifest in: insufficient consideration of the dynamic characteristics of intermittent power sources such as solar and wind power, including power fluctuations and output uncertainties; a lack of effective coordination mechanisms between distributed photovoltaic and energy storage systems and the main grid; fault detection and location algorithms that do not incorporate the real-time output characteristics of distributed energy resources; and communication latency and computational bottlenecks in centralized control architectures when handling large-scale distributed equipment. These problems lead to a disconnect between self-healing strategies and actual operating conditions, limiting fault handling efficiency in scenarios with a high proportion of renewable energy integration, making it difficult to fully leverage the regulation potential of distributed energy resources, and affecting the overall power supply reliability and renewable energy absorption capacity of the power distribution network. Summary of the Invention

[0004] The purpose of this invention is to solve the technical problems in the prior art, such as the lack of integration of distributed energy output characteristics, large positioning errors under complex topologies, and the failure to combine the dynamic characteristics of power fluctuations and curtailment of solar and wind energy, which make it difficult for self-healing strategies to match the actual output status.

[0005] A first aspect of the present invention provides a distributed intelligent healing distribution network method with fluctuation coordination, comprising:

[0006] Real-time acquisition of power distribution network operation data and distributed energy dynamic characteristic data;

[0007] Calculate the real-time output power and line voltage deviation of distributed energy based on the dynamic output model of distributed energy.

[0008] If the line voltage deviation exceeds the threshold, a multi-agent negotiation mechanism is triggered.

[0009] Each agent selects the optimal cooperative strategy through message passing and policy updates;

[0010] Perform self-healing control operations.

[0011] Furthermore, the real-time acquisition of power distribution network operation data and distributed energy dynamic characteristic data includes line fault information data, line voltage and current data, light intensity, solar panel temperature and remaining capacity data of energy storage system.

[0012] Furthermore, the real-time output power of the distributed energy source is:

[0013] ;

[0014] Where Ppv(t) is the photovoltaic output power at time t, η is the photovoltaic module conversion efficiency, A is the photovoltaic array area, G(t) is the irradiance at time t, γ is the temperature coefficient, Tcell(t) is the solar panel temperature at time t, and Tref is the reference temperature.

[0015] Furthermore, when the light intensity changes abruptly or the temperature of the solar panel rises, the distributed energy dynamic output model can adjust the real-time output power of the distributed energy in real time and perform self-healing control operations.

[0016] Furthermore, it also includes fault detection and anomaly assessment steps:

[0017] Fault location and type identification based on distribution network operation data;

[0018] Determine whether the problem is not a device malfunction by combining dynamic characteristic data of distributed energy resources;

[0019] When a non-equipment fault is confirmed, the voltage deviation will be used as the primary trigger condition to activate the self-healing mechanism.

[0020] Furthermore, the process by which each agent selects the optimal cooperative strategy through message passing and policy updates includes:

[0021] Message passing process:

[0022] ;

[0023] Strategy update process:

[0024] ;

[0025] Where mij(t) is the message from agent i to agent j, and πi is the candidate policy of agent i. Let Ui be the policy space set of agent i, λ be the utility function of agent i, λ be the cooperation coefficient, Ni be the set of neighboring agents of agent i, Uk be the utility function of agent k, and πk be the candidate policies of agent k. Let be the optimal policy of agent i in time t.

[0026] Furthermore, the self-healing control operations include fault isolation operations; power restoration operations; and topology reconfiguration operations.

[0027] Furthermore, the execution of the self-healing control operation includes:

[0028] When the output of distributed energy sources is abnormal, the energy storage system will be activated first to compensate.

[0029] When energy storage compensation is insufficient, adjust the main grid power supply parameters;

[0030] When the power deficit persists, perform topology reconfiguration.

[0031] Furthermore, the multi-agent system includes a main grid agent, a distributed photovoltaic agent, and an energy storage agent.

[0032] A second aspect of the present invention provides a fluctuation-coordinated distributed intelligent healing distribution network system, employing the fluctuation-coordinated distributed intelligent healing distribution network method as described in any of the preceding claims, comprising:

[0033] The sensing layer collects real-time data on power distribution network operation and dynamic characteristics of distributed energy resources.

[0034] The edge computing layer calculates the real-time output power of distributed energy based on the dynamic output model of distributed energy, and performs fault detection and anomaly judgment.

[0035] The cloud-based collaboration layer triggers a multi-agent negotiation mechanism, where each agent selects the optimal collaboration strategy through message passing and policy updates.

[0036] The execution control layer performs self-healing control operations.

[0037] Compared with existing technologies, this invention has at least the following beneficial effects: by collecting dynamic characteristic data of distributed energy in real time and establishing an accurate dynamic output model, it can accurately predict the power output of intermittent power sources, effectively improving the problem of insufficient consideration of the volatility of distributed energy in existing technologies; at the same time, by adopting a multi-agent negotiation mechanism to realize the intelligent coordination of multiple power sources such as the main grid, photovoltaic, and energy storage, and by selecting the optimal self-healing scheme through message passing and strategy optimization, it improves the efficiency of fault handling and power supply reliability, and enables the self-healing strategy to fully adapt to the actual operating state of scenarios with a high proportion of distributed energy access. Attached Figure Description

[0038] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained as provided without creative effort.

[0039] Figure 1 This is a schematic diagram of the steps of a distributed intelligent healing distribution network method with fluctuation coordination in one embodiment of the present invention;

[0040] Figure 2 This is a flowchart illustrating a distributed intelligent healing distribution network method with fluctuation coordination in one embodiment of the present invention. Detailed Implementation

[0041] The present invention will now be described in more detail with reference to the accompanying drawings, which illustrate preferred embodiments of the invention. It should be understood that those skilled in the art can modify the invention described herein while still achieving its advantageous effects. Therefore, the following description should be understood as being broadly known to those skilled in the art and is not intended to limit the invention.

[0042] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0043] The invention is described more specifically by way of example in the following paragraphs with reference to the accompanying drawings. The advantages and features of the invention will become clearer as explained below. It should be noted that the drawings are in a very simplified form and use non-precise proportions, and are only used to facilitate and clarify the illustration of the embodiments of the invention.

[0044] Example 1

[0045] This embodiment provides a distributed intelligent healing distribution network method with fluctuation coordination. Please refer to [reference needed]. Figures 1-2 ,include:

[0046] Real-time acquisition of power distribution network operation data and distributed energy dynamic characteristic data;

[0047] Calculate the real-time output power and line voltage deviation of distributed energy based on the dynamic output model of distributed energy.

[0048] If the line voltage deviation exceeds the threshold, a multi-agent negotiation mechanism is triggered.

[0049] Each agent selects the optimal cooperative strategy through message passing and policy updates;

[0050] Perform self-healing control operations.

[0051] Specifically, by establishing a complete closed-loop process of data perception, analysis, decision-making, and execution, intelligent self-healing control of a high proportion of distributed energy resources connected to the distribution network is achieved. This method first collects real-time data on the operating status of the distribution network and the dynamic characteristics of distributed energy resources. Based on an accurate dynamic output model, it calculates the real-time output power of each distributed energy resource, providing an accurate data foundation for subsequent decision-making.

[0052] Based on this, when a fault occurs, it is determined whether the line voltage deviation exceeds a threshold. If it does, a multi-agent negotiation mechanism is triggered, and a dynamic model is used to confirm that the fault is not equipment-related. Each agent represents different distributed energy sources and power distribution equipment. Through distributed message passing and iterative policy updates, they collaboratively seek the optimal self-healing control strategy. This negotiation mechanism fully considers the real-time output capacity and dynamic characteristics of each distributed energy source, maximizing the utilization of distributed energy sources for load transfer and power compensation while isolating faults. Ultimately, it executes optimized self-healing control operations, improving the fault handling efficiency and power supply reliability of the distribution network in scenarios with a high proportion of renewable energy access.

[0053] Furthermore, the real-time acquisition of power distribution network operation data and distributed energy dynamic characteristic data includes line fault information data, line voltage and current data, light intensity, solar panel temperature and remaining capacity data of energy storage system.

[0054] Specifically, line fault information data is collected through fault indicators installed at key nodes of the distribution network, and transient waveform recording technology is used to record the voltage and current waveform characteristics at the time of the fault. Line voltage and current data are synchronously measured by smart meters at a 1-second interval and uploaded to edge computing nodes via power line carrier communication. Irradiance data is acquired through irradiance sensors deployed on the surface of the photovoltaic array, with a sampling frequency of no less than 10Hz. The temperature of the solar panels is monitored in real time using patch-type temperature sensors, with a temperature measurement accuracy of ±0.5℃. The remaining capacity data of the energy storage system is collected through the battery management system (BMS) and verified using both coulomb counting and open-circuit voltage methods.

[0055] As a preferred implementation, the light intensity sensor uses a standard secondary photovoltaic cell as the photosensitive element, and its spectral responsivity is matched to that of the photovoltaic module. The temperature monitoring system evenly distributes three temperature measurement points on the back of each photovoltaic module, and takes the median value as the representative temperature. Line fault detection employs the transient energy method, extracting the fault characteristic frequency band through wavelet transform. Energy storage capacity monitoring incorporates a Kalman filter algorithm to eliminate the cumulative error caused by current sampling noise.

[0056] Therefore, through the collaborative acquisition of multi-source heterogeneous data, the operating status of the distribution network and the dynamic characteristics of distributed energy resources can be comprehensively perceived. High-precision environmental parameter monitoring provides reliable input for subsequent power calculations, while high-frequency sampling of electrical parameters ensures real-time fault detection. Compared with existing technologies, by optimizing sensor layout and data processing algorithms, the problems of incomplete and insufficient data acquisition in distributed energy resource integration scenarios are effectively solved, providing an accurate data foundation for subsequent self-healing control.

[0057] Furthermore, the real-time output power of the distributed energy source is:

[0058] ;

[0059] Where Ppv(t) is the photovoltaic output power at time t, η is the photovoltaic module conversion efficiency, A is the photovoltaic array area, G(t) is the irradiance at time t, γ is the temperature coefficient, Tcell(t) is the solar panel temperature at time t, and Tref is the reference temperature.

[0060] Specifically, the photovoltaic module conversion efficiency η can be obtained through laboratory calibration or technical parameters provided by the manufacturer, with a typical range of 15%-22%. The photovoltaic array area A is determined based on the actual installation scale, usually measured in square meters. The irradiance G(t) is collected in real time by an irradiance sensor, with units of W / m². The temperature coefficient γ reflects the sensitivity of the photovoltaic module output to temperature changes, typically ranging from -0.0045 / ℃ to -0.0035 / ℃. The panel temperature Tcell(t) is measured by a temperature sensor, with the reference temperature Tref usually set to 25℃. As a preferred implementation, when there are sudden changes in irradiance or panel temperature, this formula can dynamically adjust the photovoltaic output calculation results, providing an accurate power reference for subsequent self-healing control.

[0061] By establishing a precise mathematical model for photovoltaic power output, the problem of quantifying the dynamic characteristics of distributed energy output has been solved. Compared with existing technologies, this model comprehensively considers the dual influences of light intensity and temperature, and can more accurately reflect the real-time power output characteristics of photovoltaic systems, avoiding the power calculation errors caused by neglecting temperature effects in traditional methods.

[0062] Furthermore, when the light intensity changes abruptly or the temperature of the solar panel rises, the distributed energy dynamic output model can adjust the real-time output power of the distributed energy in real time and perform self-healing control operations.

[0063] Specifically, sudden changes in light intensity refer to a significant shift in light intensity within a short period, such as a rapid movement of clouds causing a drop in irradiance received by the photovoltaic array of more than 30% within seconds. Increased panel temperature is typically caused by rising ambient temperature or continuous operation of the photovoltaic modules, resulting in a temperature rise exceeding the rated operating temperature by more than 5°C. By dynamically responding to changes in environmental parameters, the power imbalance in the distribution network caused by sudden changes in distributed energy output is mitigated.

[0064] Furthermore, it also includes fault detection and anomaly assessment steps:

[0065] Fault location and type identification based on distribution network operation data;

[0066] Determine whether the problem is not a device malfunction by combining dynamic characteristic data of distributed energy resources;

[0067] When a non-equipment fault is confirmed, the voltage deviation will be used as the primary trigger condition to activate the self-healing mechanism.

[0068] This embodiment establishes a three-layer detection mechanism: "sensor data fault identification - dynamic output model anomaly cause judgment - fault or anomaly triggering self-healing." Through distribution network fault identification based on sensor data, the location and type of physical faults can be accurately pinpointed. Simultaneously, the distributed energy dynamic output model is used to analyze the root cause of power anomalies, distinguishing between equipment failures and anomalies caused by fluctuations in renewable energy output. When a real fault or power anomaly is detected, the self-healing mechanism is triggered promptly, ensuring the accuracy and timeliness of self-healing control, avoiding false triggering and missed triggering issues, and providing a reliable decision-making basis for subsequent multi-agent negotiation and self-healing control.

[0069] Furthermore, the process by which each agent selects the optimal cooperative strategy through message passing and policy updates includes:

[0070] Message passing process:

[0071] ;

[0072] Strategy update process:

[0073] ;

[0074] Where mij(t) is the message from agent i to agent j, and πi is the candidate policy of agent i. Let Ui be the policy space set of agent i, λ be the utility function of agent i, λ be the cooperation coefficient, Ni be the set of neighboring agents of agent i, Uk be the utility function of agent k, and πk be the candidate policies of agent k. Let be the optimal policy of agent i in time t.

[0075] Specifically, message passing can be implemented using a TCP / IP-based communication method. The message content includes the agent's current state, set of executable actions, and priority information. Policy updates can employ the Nash equilibrium algorithm from game theory, iteratively calculating to bring each agent's policy to a stable state. As a preferred implementation, the neighbor agent set Ni can be automatically maintained through a dynamic topology discovery protocol, updating the neighbor relationship table in real time when the network structure changes. The utility function Uk can be designed as a composite function including power supply reliability indicators, economic indicators, and renewable energy absorption rate.

[0076] A distributed negotiation mechanism enables collaborative decision-making among multiple agents. Compared to centralized control, it avoids the risk of single points of failure and reduces communication latency. Compared to fully distributed control, it ensures system-level optimization goals through message passing and policy update mechanisms. Specifically, when photovoltaic output changes abruptly, the photovoltaic agent can immediately notify the energy storage agent to initiate compensation via message, while the main grid agent synchronously adjusts power supply parameters. The three parties quickly reach a collaborative solution through policy updates, thereby effectively addressing the volatility issues of distributed energy resources.

[0077] Furthermore, the self-healing control operations include fault isolation operations; power restoration operations; and topology reconfiguration operations.

[0078] Furthermore, the execution of the self-healing control operation includes:

[0079] When the output of distributed energy sources is abnormal, the energy storage system should be activated first to compensate.

[0080] When energy storage compensation is insufficient, adjust the main grid power supply parameters.

[0081] When the power deficit persists, perform topology reconfiguration.

[0082] In fault isolation operations, when a distribution network fault is detected, the faulty section is quickly disconnected via smart circuit breakers or sectionalizing switches to prevent the fault from spreading. Specifically, a fast isolation algorithm based on differential protection can be used, combined with dual verification based on voltage and current surge characteristics. Power restoration operations restore power to non-faulty power-loss areas via tie switches or backup power supplies. During implementation, the grid synchronization conditions of distributed power sources must be considered. Topology reconfiguration operations dynamically adjust switch combinations based on real-time network conditions. The reconfiguration scheme must meet radial operation constraints and power flow safety limits; a heuristic search algorithm can be used to generate the optimal topology.

[0083] Specifically, fault isolation operations can combine impedance ranging and traveling wave positioning techniques to improve positioning accuracy, and the isolation time can be controlled within 100 milliseconds. Power restoration operations prioritize the use of energy storage systems as temporary power sources, automatically switching to the main grid's backup feeder when energy storage capacity is insufficient. Topology reconfiguration operations are implemented using a multi-objective optimization algorithm, simultaneously considering factors such as minimum network loss, load balancing, and the number of switching operations; the reconfiguration process can be completed within 30 seconds.

[0084] By employing a layered and progressive self-healing control strategy, this solution effectively addresses the issue of insufficient timeliness in fault handling during distributed energy resource integration scenarios. Compared to existing centralized self-healing control systems, this scheme forms a closed-loop operation chain encompassing fault isolation, power restoration, and topology reconfiguration, with each stage seamlessly integrated using standardized interfaces. Through dynamic topology reconfiguration capabilities, the flexible adjustment characteristics of distributed power sources can be fully utilized, reducing network losses while ensuring power supply reliability.

[0085] Furthermore, the multi-agent system includes a main grid agent, a distributed photovoltaic agent, and an energy storage agent.

[0086] Specifically, the main grid intelligent agent is responsible for monitoring the operational status and issuing control commands on the main grid side of the distribution network, interacting with distributed devices through communication modules. Distributed photovoltaic (PV) intelligent agents are deployed in PV power generation units, collecting real-time data such as irradiance and panel temperature, and calculating PV output power. Energy storage intelligent agents manage the charging and discharging status of the energy storage system, monitoring remaining capacity and charging / discharging power. All intelligent agents interact with each other through standardized communication protocols, such as using the IEC 61850 protocol to transmit real-time operational data. Point-to-point communication links are established between intelligent agents, which can be achieved using fiber optics or a dedicated wireless network. The main grid intelligent agent, acting as a coordinator, sends coordination requests to the PV and energy storage intelligent agents when a fault occurs. The PV intelligent agent provides feedback on adjustable capacity based on the current output status, while the energy storage intelligent agent reports available discharge power and duration.

[0087] The main grid agent monitors the voltage and power flow distribution of the distribution network bus to determine whether coordinated control needs to be initiated. When voltage exceedances or power imbalances are detected, the main grid agent sends adjustment commands to the photovoltaic (PV) and energy storage agents. The PV agent calculates the maximum adjustable power based on a dynamic output model, considering the temperature coefficient and the rate of change of irradiance. The energy storage agent determines the instantaneous power support capacity based on the state of charge (SOC) and charge / discharge efficiency. The three agents determine the optimal power allocation scheme through iterative negotiation. The main grid agent uses weighted least squares to calculate the power adjustment amount of each node, the PV agent executes the maximum power point tracking algorithm to adjust its output, and the energy storage agent controls the charging and discharging power through a bidirectional converter.

[0088] Through the division of labor and cooperation among three types of intelligent agents, real-time coordination between the main grid and distributed energy resources is achieved. The main grid intelligent agent provides global optimization goals, the photovoltaic intelligent agent ensures efficient utilization of new energy sources, and the energy storage intelligent agent provides rapid power support. This architecture solves the problem that centralized control is difficult to adapt to the dynamic characteristics of distributed energy resources and avoids control lag caused by communication delays. Each intelligent agent responds quickly based on local information, and at the same time, the negotiation mechanism ensures the overall optimization of the system, improving the adaptability of the distribution network to intermittent power fluctuations.

[0089] The following is an explanation using specific examples:

[0090] The existing parameters are as follows:

[0091] Table 1

[0092] Parameter type symbol numerical values illustrate Line impedance R 0.2Ω Resistance component (2km × 0.1Ω / km) X 0.3Ω Reactance component (2km × 0.15Ω / km) Rated voltage <![CDATA[V rated ]]> 10kV Distribution network line voltage reference value Load power <![CDATA[P load ]]> 50kW Total active load of the line <![CDATA[Q load ]]> 30kvar Total reactive load of the line

[0093] 1. Fault Trigger (10:00:00):

[0094] Light intensity from 800W / m 2 The sudden drop to 200W / m 2 The solar panel temperature is 35℃. The photovoltaic output is calculated according to the model.

[0095] Ppv(t)=0.18×100×200×[1-0.0045×(35-25)]=2880W;

[0096] The voltage drop of approximately 70% compared to the previous 9600W resulted in a line voltage deviation ΔV > 5% (conservative threshold).

[0097] In addition, when the photovoltaic power consumption drops sharply, the line current increases. If the line impedance effect is taken into account:

[0098] Net active power deficit: P_net = P_load - P_pv_new = 50 kW - 2.88 kW = 47.12 kW;

[0099] Net reactive power deficit: Q_net = Q_load - Q_pv_new = 30 kVar - 0 = 30 kVar;

[0100] The final line voltage deviation is even greater:

[0101] The value far exceeded the national standard limit (±7%), which also triggered the self-healing system.

[0102] 2. The self-healing sequence is as follows:

[0103] 10:00:01: The edge node detected abnormal photovoltaic output (deviation exceeding the threshold) through real-time data. Combined with the dynamic model, it was confirmed that the problem was not equipment failure, but rather caused by a sudden change in sunlight.

[0104] 10:00:02: Edge layer triggers multi-agent negotiation:

[0105] The photovoltaic intelligent agent sends a message to the photovoltaic-storage system: "Output has dropped sharply by 2880W, and energy storage is needed for compensation."

[0106] The energy storage agent responds to the solar energy storage system: "Current SOC = 30% > 20%, can provide 30kW compensation." The mainnet agent, after integrating the information, updates its strategy to the mainnet: "Do not intervene for now, prioritize energy storage compensation." Edge nodes control the energy storage system to start up and compensate for the power shortfall.

[0107] 10:00:05: Line voltage still deviates by 3% (not yet recovered to within the threshold), cloud-based multi-agent secondary negotiation:

[0108] Mainnet intelligent agent adjustment strategy: "Increase voltage to 10.5kV to enhance power supply capacity."

[0109] Edge nodes execute main grid voltage adjustment commands.

[0110] 10:00:08: Voltage returned to normal, but photovoltaic output remained low. The cloud platform decided to switch to the backup feeder to ensure long-term stability.

[0111] 10:00:10: The edge node controls the smart switch to disconnect the faulty associated line, close the backup feeder switch, complete the topology reconstruction, and complete the self-healing.

[0112] Example 2

[0113] This embodiment provides a distributed intelligent healing distribution network system with fluctuation coordination, employing the distributed intelligent healing distribution network method with fluctuation coordination as described in Embodiment 1, including:

[0114] The sensing layer collects real-time data on the operation of the power distribution network and the dynamic characteristics of distributed energy resources.

[0115] The edge computing layer calculates the real-time output power of distributed energy sources based on the dynamic output model of distributed energy, and performs fault detection and anomaly judgment.

[0116] The cloud-based collaboration layer triggers a multi-agent negotiation mechanism, where each agent selects the optimal collaboration strategy through message passing and policy updates.

[0117] The execution control layer performs self-healing control operations.

[0118] Specifically, the sensing layer collects real-time data on line faults, line voltage and current, solar irradiance, solar panel temperature, and remaining capacity of the energy storage system through sensors deployed in the power distribution network. Specifically, the sensing layer can use devices such as smart meters, fault indicators, and photovoltaic array monitoring terminals to collect data. For example, solar irradiance sensors can use silicon photovoltaic cells or thermopile sensors, and solar panel temperature monitoring can use PT100 temperature sensors.

[0119] The edge computing layer calculates real-time photovoltaic (PV) output power based on a distributed energy dynamic output model. The calculation formula for PV output power considers parameters such as irradiance, panel temperature, and PV module conversion efficiency. The edge computing layer also identifies power distribution network faults by analyzing sensor data and determines the causes of power anomalies based on the dynamic output model. As a preferred implementation, edge computing nodes (such as NVIDIA Jetson AGX Xavier) are deployed, incorporating the dynamic output model and multi-agent negotiation algorithm package. The nodes communicate with the cloud master station via a 5G network with a latency of ≤30ms.

[0120] The cloud-based collaboration layer comprises a multi-agent system consisting of main network agents, distributed photovoltaic agents, and energy storage agents. Each agent exchanges operational status information through a message passing mechanism and updates its strategy based on utility functions and collaboration coefficients, ultimately selecting the optimal collaboration strategy. Specifically, the message passing process can be implemented using a publish / subscribe model, and the strategy update process can be optimized using game theory or reinforcement learning algorithms.

[0121] The execution control layer performs self-healing control operations based on the collaborative strategy, including fault isolation, power restoration, and topology reconfiguration. When distributed energy output is abnormal, the energy storage system is activated first for power compensation; if the energy storage compensation is insufficient, the main grid power supply parameters are adjusted; when the power gap persists, network topology reconfiguration is performed. The execution control layer can implement control functions through devices such as smart circuit breakers, energy storage converters, and automatic sectionalizing switches.

[0122] The above examples illustrate the present invention only to aid in understanding it and are not intended to limit the scope of the invention. Those skilled in the art can make various simple deductions, modifications, or substitutions based on the principles of this invention.

Claims

1. A distributed intelligent healing distribution network method with fluctuation coordination, characterized in that, include: Real-time acquisition of power distribution network operation data and distributed energy dynamic characteristic data; Calculate the real-time output power and line voltage deviation of distributed energy based on the dynamic output model of distributed energy. If the line voltage deviation exceeds the threshold, a multi-agent negotiation mechanism is triggered. Each agent selects the optimal cooperative strategy through message passing and policy updates; Perform self-healing control operations; The multi-agent system includes a main grid agent, a distributed photovoltaic agent, and an energy storage agent. The main grid agent monitors the voltage and power flow distribution of the distribution network bus and determines whether coordinated control needs to be initiated. The distributed photovoltaic agent calculates the maximum adjustable power based on a dynamic output model and provides feedback on the adjustable capacity based on the current output status. The energy storage agent determines the instantaneous power support capability it can provide based on the remaining capacity status and charging / discharging efficiency of the energy storage system, and reports the available discharge power and duration.

2. The distributed intelligent healing distribution network method with fluctuation coordination as described in claim 1, characterized in that, The real-time data collected on power distribution network operation and distributed energy dynamic characteristics include line fault information data, line voltage and current data, light intensity, solar panel temperature, and remaining capacity data of energy storage systems.

3. The distributed intelligent healing distribution network method with fluctuation coordination as described in claim 1, characterized in that, The real-time output power of the distributed energy source is: ； Where Ppv(t) is the photovoltaic output power at time t, η is the photovoltaic module conversion efficiency, A is the photovoltaic array area, G(t) is the irradiance at time t, γ is the temperature coefficient, Tcell(t) is the solar panel temperature at time t, and Tref is the reference temperature.

4. The distributed intelligent healing distribution network method with fluctuation coordination as described in claim 3, characterized in that, When the light intensity changes abruptly or the temperature of the solar panel rises, the real-time output power of the distributed energy source is adjusted in real time, and a self-healing control operation is performed.

5. The distributed intelligent healing distribution network method with fluctuation coordination as described in claim 1, characterized in that, It also includes fault detection and anomaly assessment steps: Fault location and type identification based on distribution network operation data; Determine whether the problem is not a device malfunction by combining dynamic characteristic data of distributed energy resources; When a non-equipment fault is confirmed, the voltage deviation will be used as the primary trigger condition to activate the self-healing mechanism.

6. The distributed intelligent healing distribution network method with fluctuation coordination as described in claim 1, characterized in that, The process by which each agent selects the optimal cooperative strategy through message passing and policy updates includes: Message passing process: ； Strategy update process: ； Where mij(t) is the message from agent i to agent j, and πi is the candidate policy of agent i. Let Ui be the policy space set of agent i, λ be the utility function of agent i, λ be the cooperation coefficient, Ni be the set of neighboring agents of agent i, Uk be the utility function of agent k, and πk be the candidate policies of agent k. Let be the optimal policy of agent i in time t.

7. The distributed intelligent healing distribution network method with fluctuation coordination as described in claim 1, characterized in that, The self-healing control operations include fault isolation, power restoration, and topology reconfiguration.

8. The distributed intelligent healing distribution network method with fluctuation coordination as described in claim 7, characterized in that, The self-healing control operation includes: When the output of distributed energy sources is abnormal, the energy storage system will be activated first to compensate. When energy storage compensation is insufficient, adjust the main grid power supply parameters; When the power deficit persists, perform topology reconfiguration.

9. A distributed intelligent healing distribution network system with fluctuation coordination, employing the distributed intelligent healing distribution network method with fluctuation coordination as described in any one of claims 1-8, characterized in that, include: The sensing layer collects real-time data on power distribution network operation and dynamic characteristics of distributed energy resources. The edge computing layer calculates the real-time output power of distributed energy based on the dynamic output model of distributed energy, and performs fault detection and anomaly judgment. The cloud-based collaboration layer triggers a multi-agent negotiation mechanism, where each agent selects the optimal collaboration strategy through message passing and policy updates. The execution control layer performs self-healing control operations.

Images