Active power distribution network operation data processing method, system, device and medium

Abstract

This invention provides a method, system, device, and medium for processing operational data of active power distribution networks, applicable to the field of power grid operation technology. The invention constructs a core component model of a high-proportion renewable energy distribution network and calculates the parameter probability distribution of each component under different operating scenarios. Based on photovoltaic cell and wind turbine models, and combined with parameter probability distribution analysis, it identifies the comprehensive weaknesses of the high-proportion renewable energy distribution network to quantify the structural importance of components in a multi-source network, improving the accuracy of identifying weak links in the distribution network. Then, based on the comprehensive weaknesses, a multi-scenario fault isolation strategy for the high-proportion renewable energy distribution network is generated using a wide-area topology strategy. Finally, based on the multi-scenario fault isolation strategy and the energy storage system model, a hybrid reinforcement learning strategy is used to perform fault recovery on the high-proportion renewable energy distribution network, enabling rapid isolation when a fault occurs and improving the stability of fault recovery.

Description

Technical Field

[0001] This invention relates to the field of power grid operation technology, and in particular to a method, system, device and medium for processing operation data of an active distribution network. Background Technology

[0002] Currently, the widespread local integration of distributed power sources has transformed traditional single-source radial distribution networks into complex systems with multiple power sources and variable power flow directions, necessitating significant adjustments to the safe operation of distribution networks. Existing monitoring methods for distribution network operation safety suffer from low accuracy in identifying weak points because the identification process relies on the transient stability of the transmission network and only identifies weak points after a fault has occurred. Furthermore, existing methods are inefficient in fault isolation and have poor adaptability to multi-source networks. In addition, they offer inconsistent fault recovery performance and involve high complexity in control. Summary of the Invention

[0003] The main objective of this invention is to provide a method, system, device, and medium for processing operational data of an active power distribution network, which can effectively improve the accuracy of identifying weak links in the power distribution network, quickly isolate faults when they occur, and improve the stability of fault recovery.

[0004] To achieve the above objectives, the present invention provides a method for processing operational data of an active power distribution network, the method comprising the following steps:

[0005] Construct core component models for a high-proportion renewable energy distribution network, including photovoltaic cell models, wind turbine models, and energy storage system models;

[0006] Calculate the parameter probability distribution of each component in the high-proportion new energy distribution network under different operating scenarios. The parameter probability distribution includes load power distribution, photovoltaic irradiance distribution and wind speed distribution. The load power distribution is a normal distribution, the photovoltaic irradiance distribution is a beta distribution and the wind speed distribution is a Weiber distribution.

[0007] Based on the photovoltaic cell model and the wind turbine model, and combined with the parameter probability distribution, the comprehensive weakness of the high-proportion new energy distribution network is analyzed. The comprehensive weakness includes state weakness and structural weakness.

[0008] Based on the comprehensive weaknesses, a multi-scenario fault isolation strategy for the high-proportion renewable energy distribution network is generated based on a wide-area topology strategy.

[0009] Based on the multi-scenario fault isolation strategy and the energy storage system model, a hybrid reinforcement learning strategy is used to perform fault recovery on the high-proportion renewable energy distribution network.

[0010] In some embodiments, the photovoltaic cell model adopts a simplified dual-diode engineering model, and the output power of the simplified dual-diode engineering model is calculated by the following formula:

[0011] ;

[0012] In the formula, This represents the output power of the i-th photovoltaic cell. Indicates photoelectric conversion efficiency. This represents the light-receiving area of ​​the photovoltaic module in the photovoltaic cell model. Indicates irradiation intensity;

[0013] The wind turbine model adopts a doubly-fed induction generator model, and the output power of the doubly-fed induction generator model is calculated using the following formula:

[0014] ;

[0015] In the formula, Indicates the first Each fan is at all times The active power output; , , These represent the measured wind speed, cut-in wind speed, rated wind speed, and cut-out wind speed, respectively. All represent the fitting coefficients. Indicates rated power;

[0016] The energy storage system model adopts a composite energy storage model, and the power balance formula of the composite energy storage model is as follows:

[0017] ;

[0018] The capacity constraint of the composite energy storage model is as follows:

[0019] ;

[0020] In the formula, Indicates charge / discharge efficiency. These represent charging power and discharging power, respectively. This represents the energy storage capacity at time t. and These represent the upper and lower limits of capacity, respectively.

[0021] In some embodiments, the short-circuit current of the photovoltaic cell model is corrected using the following formula:

[0022] ;

[0023] The open-circuit voltage of the photovoltaic cell model is corrected using the following formula:

[0024] ;

[0025] In the formula, and These represent the short-circuit current after correction for illumination and temperature, and the open-circuit voltage after correction for temperature and illumination, respectively. and These represent the short-circuit current and open-circuit voltage before correction, respectively. b and c represent correction factors, and S represents the actual light intensity. Indicates the reference light intensity. This represents the change in light intensity, where T represents the actual temperature. This indicates temperature deviation.

[0026] In some embodiments, the step of analyzing the comprehensive weaknesses of the high-proportion renewable energy distribution network based on the photovoltaic cell model and the wind turbine model, combined with the parameter probability distribution, includes:

[0027] The state vulnerability of the high-proportion renewable energy distribution network is analyzed by the probability distribution of the parameters.

[0028] The structural weaknesses of the high-proportion renewable energy distribution network are analyzed using the photovoltaic cell model and the wind turbine model.

[0029] The overall vulnerability of the high-proportion renewable energy distribution network is obtained by weighted summation of the state vulnerability and the structural vulnerability.

[0030] In some embodiments, the analysis of the state vulnerability of the high-proportion renewable energy distribution network through the parameter probability distribution includes:

[0031] The uncertainty of the output power corresponding to the photovoltaic cell model and the wind turbine model is described by the beta distribution and the Weibull distribution, and the multi-scenario operation data corresponding to the high proportion of new energy distribution network is generated by combining the normal distribution.

[0032] Based on the multi-scenario operation data, probabilistic power flow calculations are performed to obtain the over-limit probabilities of node voltage and line current in the high-proportion new energy distribution network.

[0033] Obtain the maximum power supply capacity model of the high-proportion renewable energy distribution network;

[0034] The vulnerability of the high-proportion renewable energy distribution network is analyzed based on the over-limit probability and the maximum power supply capacity model.

[0035] In some embodiments, the analysis of the structural weaknesses of the high-proportion renewable energy distribution network using the photovoltaic cell model and the wind turbine model includes:

[0036] The nodes and lines of the high-proportion new energy distribution network are analyzed based on the photovoltaic cell model and the wind turbine model.

[0037] Calculate the degree weakness of nodes based on the connection relationships between nodes;

[0038] The dielectric weakness of the node is calculated based on the scenario voltage and scenario current in the multi-scenario operation data;

[0039] The degree weakness of the node and the medium weakness of the node are weighted and summed to obtain the overall structural weakness of the node.

[0040] The degree of weakness of the line is calculated based on the degree of weakness of the node;

[0041] The dielectric weakness of the line is calculated based on the scenario power in the multi-scenario operation data;

[0042] The overall structural weakness of the line is obtained by weighted summing of the line's degree weakness and its dielectric weakness.

[0043] The structural weakness of the high-proportion new energy distribution network is analyzed based on the comprehensive structural weakness of the nodes and the comprehensive structural weakness of the lines.

[0044] In some embodiments, generating a multi-scenario fault isolation strategy for the high-proportion renewable energy distribution network based on a wide-area topology strategy according to the comprehensive weaknesses includes:

[0045] A topology model is performed on the high-proportion new energy distribution network to obtain a connected undirected graph. The vertex set of the connected undirected graph includes buses, lines, transformers, and circuit breakers. The edge set of the connected undirected graph includes bidirectional connections between components. The bidirectional connections are quantified using an adjacency matrix.

[0046] When the high-proportion new energy distribution network undergoes structural or operational changes, the adjacency matrix is ​​updated based on a topology dynamic update mechanism.

[0047] Based on the updated adjacency matrix, the shortest tripping path of the high-proportion renewable energy distribution network is pre-generated;

[0048] The multi-scenario fault isolation strategy for the high-proportion renewable energy distribution network is generated based on the fault type and the shortest tripping path.

[0049] Another aspect of the present invention provides an operation data processing system for an active power distribution network, the system comprising:

[0050] The construction module is used to build core component models of a high-proportion renewable energy distribution network. The core component models include photovoltaic cell models, wind turbine models, and energy storage system models.

[0051] The calculation module is used to calculate the parameter probability distribution of each component in the high-proportion new energy distribution network under different operating scenarios. The parameter probability distribution includes load power distribution, photovoltaic irradiance distribution and wind speed distribution. The load power distribution is a normal distribution, the photovoltaic irradiance distribution is a beta distribution and the wind speed distribution is a Weiber distribution.

[0052] The analysis module is used to analyze the comprehensive weakness of the high-proportion new energy distribution network based on the photovoltaic cell model and the wind turbine model, combined with the parameter probability distribution. The comprehensive weakness includes state weakness and structural weakness.

[0053] The generation module is used to generate a multi-scenario fault isolation strategy for the high-proportion renewable energy distribution network based on the comprehensive vulnerability and a wide-area topology strategy.

[0054] The recovery module is used to perform fault recovery on the high-proportion renewable energy distribution network based on the multi-scenario fault isolation strategy and the energy storage system model, using a hybrid reinforcement learning strategy.

[0055] In another aspect, the present invention provides an electronic device, the electronic device / computer device including a memory and a processor, the memory storing a computer program, the processor executing the computer program to implement the above-described method.

[0056] In another aspect, the present invention provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method.

[0057] The present invention provides the following beneficial effects: By constructing core component models of a high-proportion renewable energy distribution network, including photovoltaic cell models, wind turbine models, and energy storage system models, and calculating the parameter probability distributions of each component in the high-proportion renewable energy distribution network under different operating scenarios, the comprehensive weakness of the high-proportion renewable energy distribution network is analyzed based on the photovoltaic cell model and wind turbine model, combined with the parameter probability distribution. This quantifies the structural importance of components in the multi-source network, avoids single-perspective bias, and improves the accuracy of identifying weak links in the distribution network. Then, based on the comprehensive weakness, a multi-scenario fault isolation strategy for the high-proportion renewable energy distribution network is generated based on a wide-area topology strategy. Finally, based on the multi-scenario fault isolation strategy and the energy storage system model, a hybrid reinforcement learning strategy is used to perform fault recovery on the high-proportion renewable energy distribution network, thereby enabling rapid isolation when a fault occurs and improving the stability of fault recovery. Attached Figure Description

[0058] To more clearly illustrate the technical solutions in this 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 some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0059] Figure 1 This is a flowchart of an active power distribution network operation data processing method according to this application;

[0060] Figure 2 This is a schematic diagram of the improved IEEE 33-node power distribution system proposed in this application;

[0061] Figure 3 This is a schematic diagram of the structural weakness calculation results of the node in this application;

[0062] Figure 4 This is a schematic diagram showing the structural weakness calculation results of the circuit in this application;

[0063] Figure 5 This is a flowchart of the fault isolation process based on wide-area protection in this application;

[0064] Figure 6 This is a schematic diagram of the Dueling neural network structure in this application;

[0065] Figure 7 This is a flowchart of the hybrid reinforcement learning algorithm of this application;

[0066] Figure 8 This is a comparative diagram of the fault recovery effects of this application;

[0067] Figure 9 This is a schematic diagram of the module of an active power distribution network operation data processing system according to this application. Detailed Implementation

[0068] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0069] It should be noted that, in the description of this application, the terms "comprising," "including," or any other variations thereof are intended to cover a 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 a process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element. The terms "upper," "lower," etc., indicating orientation or positional relationships based on the orientation or positional relationships shown in the accompanying drawings, are used only for the convenience of describing this application and for simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of this application. Unless otherwise expressly specified and limited, the terms "installed," "connected," and "linked" should be interpreted broadly, for example, as a fixed connection, a detachable connection, or an integral connection; a mechanical connection or an electrical connection; a direct connection or an indirect connection through an intermediate medium; or a connection within two elements. Those skilled in the art can understand the specific meaning of the above terms in this application according to the specific circumstances.

[0070] The terms "first," "second," etc., used in this application are used to distinguish similar objects and not to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that embodiments of this application can be implemented in orders other than those illustrated or described herein, and the objects distinguished by "first," "second," etc., are generally of the same class, without limiting the number of objects; for example, a first object can be one or more. Furthermore, "and / or" indicates at least one of the connected objects, and the character " / " generally indicates that the preceding and following objects have an "or" relationship.

[0071] Before providing a detailed description of the embodiments of this application, the terms used in this application are explained as follows:

[0072] Power flow refers to the situation in which current or power flows from the power source into the load and is distributed throughout the power grid under the excitation of the power source's electromotive force during the operation of the power system.

[0073] A high-proportion renewable energy distribution network refers to a distribution network in which renewable energy sources such as wind power and photovoltaics and their power electronic equipment occupy a significant share in the power supply structure. This accelerates the transformation of the distribution network from a traditional "passive, unidirectional" network to an "active, bidirectional" interactive system. It has the ability to efficiently aggregate and locally absorb distributed power sources, energy storage, and flexible loads, and achieves coordinated operation of power generation, grid, load, and storage through digitalization and intelligence.

[0074] Wind turbines are a key component of wind power systems, primarily responsible for converting wind energy into electrical energy and distributing it through the power grid. Their core function is to capture wind energy through the rotor, drive generators to produce electricity, and ultimately deliver clean energy.

[0075] A photovoltaic cell is the smallest functional unit that directly converts solar radiation energy into direct current using the photovoltaic effect of semiconductors. Multiple cells are packaged to form a photovoltaic module, and then the modules are connected in series and parallel to form a photovoltaic array. On the distribution network side, photovoltaics are usually connected to the user side or transformer substation side in the form of distributed photovoltaics, generating and consuming electricity locally, with surplus electricity fed into the grid, becoming a distributed power source for the distribution system.

[0076] Photovoltaic output refers to the power output capability of a distributed photovoltaic system to the power distribution network within a specific time period. Its characteristics are affected by the power generation time, spatial distribution, and grid operation mode.

[0077] In related technologies, the local integration of numerous distributed power sources transforms traditional single-source radial distribution networks into complex systems with multiple power sources and variable power flow directions, thus necessitating significant adjustments to the safe operation of distribution networks. Existing monitoring methods for distribution network operation safety have the following shortcomings when identifying weak points:

[0078] In identifying weak links, there are issues of insufficient adaptability and a single perspective. Existing methods are mostly derived from transmission networks and biased towards transient stability, failing to build an indicator system around the core objective of distribution networks to "ensure power supply capacity." They often only consider single dimensions of state or structural weakness and ignore the randomness of distributed power generation output, making it difficult to accurately locate weak nodes and lines, resulting in insufficient pre-emptive prevention capabilities. In actual engineering projects, distribution network companies rely more on the experience of maintenance personnel and historical fault statistics to determine "problem lines" and "problem transformer areas." Often, renovations and reinforcements are only initiated after a feeder has repeatedly exceeded limits, tripped, or even experienced an increase in complaints, lacking unified quantitative indicators to support the early stages. Distribution networks at the same voltage level in different regions vary significantly in load type and distributed power generation access methods. Using the transmission network indicator system can easily lead to situations where "the evaluation results look good, but the actual problems persist." New distribution network construction and expansion projects also struggle to identify potential weak links in advance during the planning stage.

[0079] The fault isolation process is inefficient and poorly adaptable to multi-power supply networks. Existing methods rely on local measurement backup protection, which requires a long delay to ensure selectivity and regenerates tripping paths after a fault occurs. This prolongs fault clearing time and can easily expand the power outage area. After distributed power generation is connected to the grid, the network is no longer equivalent to a radial structure, which can easily lead to protection maloperation or failure to operate, further amplifying the impact. On current 10 kV feeders, for single-phase grounding or short-circuit faults, the coordination time between main protection and backup protection is generally between several hundred milliseconds and several seconds. Many tie switches and sectionalizing switches still rely on dispatcher orders or manual operation on-site. In densely populated urban load areas, a single fault can cause the entire feeder or even multiple switching stations to lose power. After distributed photovoltaic and wind power are connected through inverters, the fault current distribution is significantly different from that of traditional radial networks. Some areas with lagging upgrades have experienced typical problems such as "small faults, large power outages" and "frequent blowouts of DG-side fuses and failure of upstream switches to operate." The fault boundaries are blurred, and maloperation and failure to operate occur frequently, affecting users' acceptance of new energy grid connection.

[0080] The fault recovery phase is characterized by unstable performance and high complexity. Classical particle swarm optimization and integer programming are prone to dimensionality curse and significantly reduced solution efficiency under conditions of high proportion of renewable energy sources, especially when faced with strong coupling of multiple variables. Some studies employ single reinforcement learning methods that only cover discrete actions and struggle to handle continuous actions such as power generation and energy storage output simultaneously, failing to fully utilize the support capabilities of distributed power sources. Renewable energy sources are significantly affected by sunlight and wind speed, exhibiting randomness that traditional strategies do not incorporate, thus increasing the risk of voltage and power exceeding limits during the recovery process. In actual dispatching, after a distribution network fault is cleared, dispatchers often rely on operational experience to formulate switching plans and then issue instructions via telephone or feeder automation to complete the phased power restoration. The coordinated utilization of high proportion of distributed power sources, energy storage, and flexible loads mainly remains at the experience level of "maximizing local consumption and minimizing backfeeding to the upper level," making it difficult to simultaneously achieve multiple objectives such as phased power restoration for users, avoiding overload of public transformers, and ensuring qualified branch voltages while guaranteeing power quality. If the weather changes suddenly or the output of distributed power sources deviates from expectations during the recovery process, the original plan may face risks such as excessive node voltage, local line overload, or even relay protection reactivation. The dispatch center will need to revise the plan temporarily, further increasing the recovery time and complexity.

[0081] In view of this, embodiments of this application provide a method, system, device, and medium for processing operational data of an active power distribution network, which can effectively improve the accuracy of identifying weak links in the power distribution network, quickly isolate faults when they occur, and improve the stability of fault recovery.

[0082] The embodiments of this application will be described in detail below with reference to the accompanying drawings:

[0083] Reference Figure 1 This application provides a method for processing operational data of an active power distribution network, which includes, but is not limited to, the following steps:

[0084] Step S110: Construct a core component model for a high-proportion renewable energy distribution network, wherein the core component model includes a photovoltaic cell model, a wind turbine model, and an energy storage system model;

[0085] Step S120: Calculate the parameter probability distribution of each component in the high-proportion new energy distribution network under different operating scenarios. The parameter probability distribution includes load power distribution, photovoltaic irradiance distribution and wind speed distribution. The load power distribution is a normal distribution, the photovoltaic irradiance distribution is a beta distribution and the wind speed distribution is a Weiber distribution.

[0086] Step S130: Based on the photovoltaic cell model and the wind turbine model, analyze the comprehensive weakness of the high-proportion new energy distribution network by combining the parameter probability distribution. The comprehensive weakness includes state weakness and structural weakness.

[0087] Step S140: Based on the comprehensive vulnerability, generate a multi-scenario fault isolation strategy for a high proportion of new energy distribution networks based on the wide-area topology strategy;

[0088] Step S150: Based on the multi-scenario fault isolation strategy and energy storage system model, perform fault recovery on the high-proportion new energy distribution network using a hybrid reinforcement learning strategy.

[0089] It is understandable that this embodiment provides data support for identifying weak links in the distribution network by constructing a core component model of a high-proportion renewable energy distribution network. Specifically, in the photovoltaic cell model, this embodiment adopts a simplified dual-diode engineering model. Considering the influence of light intensity and temperature on power output, the output power of the simplified dual-diode engineering model in this embodiment is calculated using the following formula:

[0090] (1);

[0091] In the formula, This represents the output power of the i-th photovoltaic cell. Indicates photoelectric conversion efficiency. This represents the light-receiving area of ​​the photovoltaic module in the photovoltaic cell model. Indicates the intensity of radiation.

[0092] Formula (1) describes the output power of the photovoltaic cell. The photovoltaic output is calculated based on the light intensity and the conversion efficiency of the photovoltaic module, which provides an uncertainty model for the photovoltaic power output for subsequent identification of weak links.

[0093] In this embodiment, to adapt to the short-circuit current and open-circuit voltage in the power distribution network under different environmental conditions, the short-circuit current of the photovoltaic cell model is corrected using the following formula:

[0094] (2-1);

[0095] The open-circuit voltage of the photovoltaic cell model is corrected using the following formula:

[0096] (2-2);

[0097] In the formula, and These represent the short-circuit current after correction for illumination and temperature, and the open-circuit voltage after correction for temperature and illumination, respectively. and These represent the short-circuit current and open-circuit voltage before correction, respectively. b and c represent correction factors, and S represents the actual light intensity. Indicates the reference light intensity. This represents the change in light intensity, where T represents the actual temperature. This indicates temperature deviation.

[0098] Formulas (2-1) and (2-2) correct the current and voltage of the photovoltaic cell under different light and temperature conditions to adapt to environmental changes. Since the calculation of photovoltaic output depends not only on light intensity but also on the impact of temperature changes on cell performance, this embodiment enhances the adaptability of the fault recovery model in complex environments through correction.

[0099] It is understood that the wind turbine model in this embodiment adopts a doubly-fed induction generator model, and the output power of the doubly-fed induction generator model is calculated using the following formula:

[0100] (3);

[0101] In the formula, Indicates the first Each fan is at all times The active power output; , , These represent the measured wind speed, cut-in wind speed, rated wind speed, and cut-out wind speed, respectively. All represent the fitting coefficients. Indicates the rated power.

[0102] Formula (3) describes the active power output of the wind turbine generator set, which is a piecewise function of wind speed. This formula takes into account the influence of different wind speeds on the wind turbine output power. The change in wind speed directly affects the wind power output power, providing necessary stochastic modeling for the utilization of renewable energy in the fault recovery of the distribution network.

[0103] It is understood that the energy storage system model in this embodiment adopts a composite energy storage model, and the power balance formula of the composite energy storage model is as follows:

[0104] (4-1);

[0105] The capacity constraint of the composite energy storage model is as follows:

[0106] (4-2);

[0107] In the formula, Indicates charge / discharge efficiency. These represent charging power and discharging power, respectively. This represents the energy storage capacity at time t. and These represent the upper and lower limits of capacity, respectively.

[0108] Formulas (4-1) and (4-2) calculate the charging and discharging power of the energy storage system and define the upper and lower limits of the energy storage capacity. The charging and discharging capacity of the energy storage system is crucial to the power balance during the fault recovery process, which limits the operating range of the energy storage system and affects the execution of the recovery strategy.

[0109] Understandably, due to the uncertainty in the output power of new energy sources, this embodiment introduces a probabilistic power flow method to calculate the parameter probability distribution of each component under different operating scenarios. The load power follows a normal distribution, and the photovoltaic irradiance follows a beta distribution. The probability density function corresponding to the photovoltaic irradiance is given by the following formula:

[0110] (5);

[0111] In the formula, Photovoltaic irradiance, This is the upper limit for calibrating irradiance. Let be the shape parameter of the beta distribution. Let be the probability density function.

[0112] Wind speed follows a Weibull distribution, and its probability density function is as follows:

[0113] (6);

[0114] In the formula, For wind speed, Let be the shape parameter of the Weibull distribution. is the scaling parameter of the Weiber distribution.

[0115] Formulas (5) and (6) take into account the impact of environmental factors on the stability of the power distribution network by introducing the probability distribution of photovoltaic irradiance and wind speed, thus providing a theoretical basis for the identification of weak links.

[0116] This embodiment obtains the probability distribution of node voltage and line current in multiple scenarios through sampling, thereby providing a quantitative basis for identifying weak links.

[0117] Understandably, this embodiment also constructs a weak link evaluation index system to analyze the comprehensive weakness of a high-proportion renewable energy distribution network based on photovoltaic cell models and wind turbine models, combined with parameter probability distribution. Specifically, this embodiment analyzes the state weakness of a high-proportion renewable energy distribution network through parameter probability distribution and the structural weakness through photovoltaic cell models and wind turbine models. The overall weakness of the high-proportion renewable energy distribution network is then obtained by weighted summation of the state and structural weaknesses. This embodiment constructs indicators from both state and structural dimensions to comprehensively characterize the degree of weakness in each link of the distribution network.

[0118] Understandably, for state vulnerability, this embodiment describes the uncertainty of the output power corresponding to the photovoltaic cell model and the wind turbine model through beta distribution and Weiber distribution, and generates multi-scenario operation data corresponding to the high-proportion renewable energy distribution network by combining normal distribution. Based on the multi-scenario operation data, probabilistic power flow calculation is performed to obtain the over-limit probability of node voltage and line current of the high-proportion renewable energy distribution network. At the same time, after obtaining the maximum power supply capacity model of the high-proportion renewable energy distribution network, the state vulnerability of the high-proportion renewable energy distribution network is analyzed based on the over-limit probability and the maximum power supply capacity model.

[0119] Specifically, the vulnerability index quantifies the vulnerability of a component's operating state by multiplying the "component over-limit probability × over-limit consequence," thus yielding a risk value. The over-limit probability is calculated using a probabilistic power flow method to obtain the node voltage over-limit probability and the line current over-limit probability. The over-limit consequence assessment is based on the maximum power supply capacity model, utilizing the impact of component parameter changes on the distribution network's maximum power supply capacity to characterize the severity of the consequences. The maximum power supply capacity model formula is as follows:

[0120] Formula (7);

[0121] In the formula, For nodes The average annual load, For load The growth base, N is the load growth factor, and N is the total number of nodes. In this embodiment, the maximum power supply capacity of each node is evaluated using formula (7), and the maximum power supply capacity of the entire distribution network is calculated. This formula is used to quantify the importance of each component in a multi-power supply network and to assist in assessing its impact on the overall system capacity when identifying weak links.

[0122] The change in maximum power supply capacity is calculated by gradually altering the component constraints using the following formula:

[0123] Formula (8-1);

[0124] Formula (8-2);

[0125] In the formula, To change only the components After constraints, the change in maximum power supply capacity relative to the reference; To change the components simultaneously and The change in maximum power supply capacity after constraints.

[0126] Risk value calculation quantifies the degree of weakness by combining a risk-taking utility function:

[0127] Formula (9-1);

[0128] Formula (9-2);

[0129] Formula (9-3);

[0130] In the formula, Let x be the initial value of the consequence quantity, and a, b, and c be the shape, translation, and scale parameters, respectively. , The risk values ​​for the corresponding nodes and lines. , This represents the probability of exceeding the limit for the corresponding node and line.

[0131] Based on the probability and consequences of exceeding limits, the risk value of a node or line is calculated using formulas (9-1), (9-2), and (9-3) to assess its potential impact on the power grid when a fault occurs. By calculating the risk value, weak links can be further analyzed and fault recovery strategies can be optimized.

[0132] Understandably, this embodiment addresses structural vulnerability by introducing complex network concepts to characterize the topological vulnerability of the distribution network, defining two types of structural vulnerability: nodes and lines. Specifically, after analyzing the nodes and lines of a high-proportion renewable energy distribution network based on photovoltaic cell and wind turbine models, this embodiment calculates the degree vulnerability of nodes based on the connection relationships between nodes. The formula for calculating the degree vulnerability of node i is as follows:

[0133] Formula (10);

[0134] In the formula, For nodes Connectivity Let j be the connectivity degree of adjacent node j. The average connectivity of the network. For nodes The higher the degree of the adjacent node set and the denser the neighborhood, the stronger the connection breadth.

[0135] The dielectric weakness of the node is calculated based on the scenario voltage and scenario current from multi-scenario operation data. The formula for calculating the dielectric weakness of the node is as follows:

[0136] Formula (11);

[0137] In the formula, n represents the total number of scenes. For the scene Next node Voltage, As the reference voltage, The weights of scene s, Let i be the set of branches connected to node i. For the branch path in scene s The larger the current value, the stronger the control over the power flow.

[0138] After obtaining the degree weakness and the dielectric weakness of the nodes, a weighted sum of these two values ​​is performed to obtain the overall structural weakness of the node. The overall structural weakness of the node is... The calculation formula is as follows:

[0139] Formula (12);

[0140] In the formula, , These are the weighting coefficients for the node degree metric and the node betweenness metric, respectively.

[0141] Regarding line weakness, this embodiment first calculates the line's degree weakness based on the degree weakness of the nodes, using the following formula:

[0142] Formula (13);

[0143] In the formula, They are respectively with nodes The weakness of the node degree of the two connected nodes. For the line The reactance value reflects the importance of the line connection.

[0144] The dielectric weakness of the line is calculated based on the scenario power data from multiple operational scenarios. The calculation formula is as follows:

[0145] Formula (14);

[0146] In the formula, In the scene Downline The effective transmission power, For the scene Downline Maximum permissible transmission power For the scene Downline The current.

[0147] Then, the line's linear weakness and dielectric weakness are weighted and summed to obtain the line's overall structural weakness, calculated using the following formula:

[0148] Formula (15);

[0149] In the formula, , These are the weighting coefficients for the line degree index and the line betweenness index, respectively.

[0150] Next, the structural weaknesses of high-proportion renewable energy distribution networks are analyzed based on the comprehensive structural weaknesses of nodes and lines. For example, using... Figure 2 Taking the improved IEEE 33-node distribution system as an example, we will evaluate the weak links in the distribution system. Several important nodes are configured as follows: node 1 is a distribution transformer, nodes 13 and 24 are connected to photovoltaic (PV) power, node 8 is connected to wind power, and node 30 is connected to energy storage. A model incorporating PV and energy storage is constructed, and probabilistic power flow simulation is performed to obtain the over-limit probabilities of voltage and line current at each node. Based on this, the state vulnerability risk value and structural vulnerability quantification value are calculated. The structural vulnerability calculation results for the nodes are shown below. Figure 3 As shown, the calculation results of the structural weakness of the line are as follows: Figure 4 As shown.

[0151] This embodiment obtains the overall vulnerability by using weighted summation, with the weights set to 0.6 for state and 0.4 for structure, forming an overall vulnerability ranking and identifying key weak links, which can be used as the basis for subsequent fault isolation and recovery.

[0152] It is understandable that, after obtaining the comprehensive weaknesses of the distribution network, this embodiment generates a multi-scenario fault isolation strategy for a high-proportion renewable energy distribution network based on a wide-area topology strategy. Specifically, this embodiment first performs topology modeling on the high-proportion renewable energy distribution network to obtain a connected undirected graph, G=(V,E), where the vertex set V contains buses, lines, transformers, and circuit breakers, and the edge set E represents the bidirectional connection relationship between components. The topology structure is quantified through an adjacency matrix.

[0153] Among them, the component-circuit breaker adjacency matrix This is used to describe the connection relationship between the protected element and the circuit breaker, and takes the form of the following formula:

[0154] Formula (16);

[0155] In the formula, A1 is a P×P matrix, where P is the number of circuit breakers, the main diagonal is 0, and the rest are ∞, indicating no direct connection between circuit breakers; B1 is a Q×Q matrix, where Q is the number of protected components, the main diagonal is 0, and the rest are ∞, indicating no direct connection between components; C1 is a P×Q binary matrix, C... pq =1 indicates that circuit breaker p is directly connected to component q, C pq =∞ indicates no connection.

[0156] Circuit breaker - Circuit breaker adjacency matrix This is used to handle the interconnection of circuit breakers in the event of failure to operate, and takes the form of the following formula:

[0157] Formula (17);

[0158] In the formula, A2 is a P×P matrix, a mn =0 indicates that circuit breaker m and circuit breaker n are the same circuit breaker, a mn =1 indicates that circuit breaker m and circuit breaker n are directly connected, a mn =∞ indicates that circuit breaker m and circuit breaker n are not directly connected; B2=B1; C2 is an all-∞ matrix, indicating that the circuit breaker and the component are not connected.

[0159] When structural or operational changes occur in a high-proportion renewable energy distribution network, the adjacency matrix is ​​updated based on a topology dynamic update mechanism. This mechanism refreshes the adjacency matrix in real time when structural or operational changes occur in the distribution network. Typical triggers include changes in the status of circuit breakers and disconnectors, substation outages or restorations, commissioning or decommissioning of new equipment, and adjustments to the energized state of the source side.

[0160] For example, when the state of a circuit breaker or disconnector changes, if both are closed simultaneously, then... and The corresponding elements remain unchanged, and if any end is disconnected, it is set to ∞. When a substation experiences a power outage, all circuit breakers within the substation are merged into a single vertex, corresponding to a. mn Zero indicates the same node. When a new device is connected, the corresponding row and column are added to the adjacency matrix, and values ​​are assigned according to its connectivity with existing components. For example, if a new circuit breaker is connected to an existing line, then C is set to zero. pq =1.

[0161] Then, based on the updated adjacency matrix, the shortest tripping path for a high-proportion renewable energy distribution network is pre-generated. Specifically, this embodiment uses the Floyd-Warshall algorithm to solve the adjacency matrix, pre-generating the shortest path between all components, eliminating the need for repeated calculations after a fault. The distance matrix D stores the shortest path length between any two vertices, and the calculation logic is as follows:

[0162] Formula (18);

[0163] In the formula, The elements of the adjacency matrix are the weights of the direct edges from i to j. If there are no edges, they are set to ∞. Self-loops are usually 0. This represents the current shortest path length from i to j; Where N is the intermediate node index and N is the total number of vertices, when When the distance is equal to N, the final distance matrix D is obtained.

[0164] The predecessor matrix R stores the sequence of nodes for the shortest path, and its calculation logic is as follows:

[0165] ;

[0166] In the formula, For each element of the predecessor matrix, the predecessor of j on the path from i to j under the current shortest path is recorded. The circuit breaker operation sequence of the shortest path can be traced back through the predecessor matrix.

[0167] Next, based on the fault type and the shortest tripping path, a multi-scenario fault isolation strategy for a high-proportion renewable energy distribution network is generated. This allows for the development of targeted fault isolation schemes, ensuring rapid and minimal-scope fault isolation. Specifically, when a single component fault occurs, the vertex V corresponding to the faulty component is identified. i In the search distance matrix D, select those directly connected to it and The set of circuit breakers V = 1 j Fault isolation is completed by tripping circuit breakers sequentially according to distance sequence. If a circuit breaker fails to operate on the tripping path, the circuit breaker that fails to operate is taken as the new node V. iContinue searching for circuit breakers with a distance of 2 until the fault is isolated, avoiding the repeated generation of paths. In scenarios involving tie lines with distributed power sources, first set the adjacency relationship of the faulty line to unreachable and recalculate the shortest tripping path to ensure reliable separation between the source-end circuit breaker and the load-end circuit breaker, preventing the outage area from being too large.

[0168] like Figure 5 As shown, during normal operation of the distribution network, an adjacency matrix W is formed, and after constructing the distribution network topology, a distance matrix D and a predecessor matrix R are obtained. Then, based on the adjacency matrix W, the distance matrix D, and the predecessor matrix R, a tripping strategy is determined during a fault, and fault isolation is performed based on the tripping strategy when a distribution network fault occurs.

[0169] It is understood that, after obtaining the multi-scenario fault isolation strategy, this embodiment combines the energy storage system model and uses a hybrid reinforcement learning strategy to perform fault recovery on a high-proportion renewable energy distribution network. Specifically, the fault recovery process in this embodiment includes, but is not limited to, the following steps:

[0170] Step 1: Fault Recovery Model Construction

[0171] Based on the principle of "safety first, minimum loss," a recovery model is established, including an objective function and constraints. The objective function comprehensively considers six key indicators, including power loss load, number of switching operations, and recovery cost, with the goal of minimizing the overall fault loss.

[0172] Formula (20);

[0173] Formula (21);

[0174] Formula (22);

[0175] Formula (23);

[0176] Formula (24);

[0177] Formula (25);

[0178] Formula (26);

[0179] In the formula, Indicates power loss load. Indicates the number of times the switch has been activated. Indicates recovery costs, Indicates network loss. Indicates voltage fluctuation. This indicates the duration of the user's power outage; the weighting coefficients k1−k6 are determined using the analytic hierarchy process (AHP). This is the weighting coefficient for the importance of the load. The load time-varying coefficient, The power supply status is indicated by 1, which means that the load is not powered at time t, and 0 means that it is powered. For the switching element state, closed = 1, open = 0. Each state reversal is counted, and coefficient 2 converts one combination / division into one action count. , For equipment operating costs, , Let n be the number of devices, n1, n2, n3, and n4 corresponding to the number of devices that are activated or scheduled; M is the set of tributaries. Let Ps be the branch impedance, and Qs and Us be the branch power and voltage, respectively. This represents the actual node voltage amplitude. Rated voltage; This is the power outage time loss coefficient.

[0180] In this embodiment, formula (20) integrates various key indicators in the recovery process, such as power loss load, number of switching actions, and recovery cost, in order to achieve the optimal goal of fault recovery. The fault recovery objective function defines a comprehensive evaluation index that combines multiple factors to minimize the overall loss in the recovery process.

[0181] The constraint set covers the key boundaries of the safe operation of the distribution network, including power flow constraints, distributed generation constraints, energy storage constraints, and recovery time constraints. Among them, power flow constraints include voltage, power, and frequency constraints, as shown in equations (27) to (29); distributed generation constraints include photovoltaic power output and wind power output constraints, as shown in equation (30); energy storage constraints include charging and discharging power and energy storage capacity constraints, as shown in equations (31) and (32); the recovery process is controlled by a time limit as the maximum acceptable outage duration for users, as shown in equation (33).

[0182] Formula (27);

[0183] Formula (28);

[0184] Formula (29);

[0185] Formula (30);

[0186] Formula (31);

[0187] Formula (32);

[0188] Formula (33);

[0189] In the formula, This indicates the upper and lower limits of the allowed voltage at node i. This represents the actual voltage at time t; , This represents the upper and lower limits of the allowed active power at node i. This represents the actual active power at time t. , This represents the upper and lower limits of the reactive power allowed at node i. This represents the actual reactive power at time t; , Indicates the upper and lower limits of the frequency allowed by the system. The system frequency at time t; This indicates the maximum available output of the photovoltaic device. This indicates the maximum available output of the fan unit; This indicates the upper limit of energy storage charging and discharging power. Indicates the lower limit of the corresponding power; This indicates the upper and lower limits of the allowable energy storage range; T represents the duration of this fault recovery process. The maximum acceptable power outage duration for the user.

[0190] In this embodiment, formulas (27) to (33) set constraints on key variables such as voltage and power during the recovery process, ensuring that the distribution network will not exceed safety limits during the recovery process. These constraints provide a safety boundary for the fault recovery process, ensuring that the system can avoid problems such as overload or voltage exceeding limits during the recovery process.

[0191] Step 2: Hybrid Reinforcement Learning Modeling

[0192] To address the "discrete + continuous" hybrid action space requirements of fault recovery, a hybrid algorithm combining D3QN and DDPG is employed to construct an MDP model, and... Figure 3 The structure of the Dueling neural network shown is designed to improve the stability and convergence efficiency of the estimation.

[0193] The action space A is divided into discrete actions and continuous actions, covering circuit breaker operation and power supply regulation. The discrete action set is used for circuit breaker opening and closing control, employing 0-1 variables. express, =0 indicates that the state is maintained. =1 indicates a state switch, as shown in the formula:

[0194] Formula (34);

[0195] In the formula, =1 indicates that the operation of closing or opening the i-th circuit breaker is performed. =0 indicates that it remains unchanged. The number of operable circuit breakers. This is a set of circuit breakers.

[0196] The continuous action set is used for active and reactive power regulation of conventional power sources, photovoltaic, wind power, and energy storage, as shown in the following formula:

[0197] Formula (35);

[0198] In the formula, , Let represent the active and reactive power of the controllable power source at node i. For the number of controllable power supplies, It is a collection of controllable power sources.

[0199] The state space S includes observed states and constraint states, comprehensively reflecting the operating state of the distribution network. The observed states consist of power flow, renewable energy output, voltage, energy storage status, and switching status, denoted by the following formula:

[0200] Formula (36);

[0201] The voltage and power limits, frequency limits, and time limit are taken as constant inputs and denoted by the following formula:

[0202] Formula (37);

[0203] The reward uses a combination of positive and negative factors to guide the strategy toward convergence in a fast, safe, and economical direction, as shown in the following formula:

[0204] Formula (38);

[0205] Formula (39);

[0206] Formula (40);

[0207] Formula (41);

[0208] In the formula, As a stage reward, , , As a reward and punishment constant, A large penalty coefficient represents the punishment for violating hard constraints; As a positive reward, As a soft constraint penalty, As a hard constraint and punishment; These are soft constraint weights; , , The number of actions that received rewards, the number of actions that received soft constraint penalties, and the number of actions that received hard constraint penalties; For agent actions, The constant coefficient, This is the deviation metric for the i-th type of soft constraint.

[0209] In this embodiment, formula (34) is a discrete action set, mainly used for the opening and closing control of the circuit breaker, that is, to isolate the fault area or restore power supply by switching the circuit breaker. Formula (35) is a continuous action set used to adjust the power output of photovoltaic, wind power and energy storage, and to optimize the power restoration process by adjusting the output of these power sources. Formula (36) defines the observation space, which provides data on the current state of the distribution network. This data will be used as input to the reinforcement learning algorithm to help the agent make optimization decisions. Formula (37) defines the constraint space, which provides the safety boundary that should be followed during the restoration process, ensuring that the restoration process does not exceed the safe operating range of the system. Formula (38) defines the reward function in hybrid reinforcement learning, which comprehensively considers positive rewards, soft constraint penalties and hard constraint penalties. Its goal is to guide the agent to converge in a fast, safe and economical direction.

[0210] Step 3: Training and solving the hybrid reinforcement learning algorithm, the specific process of which is as follows: Figure 7 As shown:

[0211] First, set the core parameters of D3QN and DDPG. The state dimension of D3QN is 14, the discount factor γ=0.9, and the maximum training step size is 120000. The learning rate of the actor and critic networks in DDPG is the same, which is 1e-4, and the capacity of the experience replay pool is 1e6.

[0212] Subsequently, alternating collaborative training is employed. First, the DDPG parameters are fixed and treated as part of the environment, allowing D3QN to interact with the environment and update the discrete strategy with the Q-value, thereby optimizing the circuit breaker closing / opening sequence and source switching. The target Q-value formula is as follows:

[0213] Formula (42);

[0214] In the formula, This represents the Q-value of the discrete action target network in the algorithm. Indicates the state at time t. This represents the discrete action performed at time t. For immediate rewards, γ is the discount factor. *; For online Q network, the parameters are: , For the parameters of the target Q network, Choose an action for the online network in the next moment.

[0215] Secondly, the D3QN parameters are fixed, allowing DDPG to interact with the environment and update the continuous policy with policy gradients, thereby optimizing the active and reactive power output of each power source. Both types of networks use target networks and mini-batch sampling, alternating iteratively until convergence. The gradient formula for the actor network is as follows:

[0216] Formula (43);

[0217] In the formula, This indicates that in optimizing the objective function China gradient, Indicates the agent's policy in the state Continuous output action , ) represents the critic's action value function, with parameters This represents the gradient of Q with respect to the action. This represents the gradient of the policy with respect to its parameters; This represents the network parameters of the actors.

[0218] The model verification and deployment in this embodiment are based on the improved IEEE 33-node distribution network dataset, with training set convergence as the criterion, including the increase of the moving average of the reward and the stabilization of the single-step loss. After training, the agent directly outputs the circuit breaker operation sequence and power output scheduling under actual operating conditions to complete fault isolation and rapid recovery. In this embodiment, formula (42) is used to update the Q network and optimize the control action of the circuit breaker. Formula (43) enables the agent to dynamically adjust the output of photovoltaic, wind power and energy storage by updating the continuous action strategy, thereby achieving optimal fault recovery.

[0219] As described above, this embodiment primarily focuses on modeling the discrete and continuous action spaces in reinforcement learning. By defining the discrete and continuous action sets, a unified framework is provided for circuit breaker operation and power regulation during fault recovery. The reward and penalty functions, combined with constraints, guide the agent to optimize as much as possible during recovery while adhering to system safety constraints. The Q-value update formulas for discrete and continuous actions ensure the effectiveness and convergence of the algorithm, thereby achieving fast and safe fault recovery.

[0220] As can be seen from the above, the method of this application embodiment has the following beneficial effects:

[0221] First, it enables more accurate identification of weak points and more effective fault prevention. For the first time, it jointly models "state vulnerability" and "structural vulnerability," covering issues such as voltage fluctuations and power flow reversals after the integration of new energy sources. With probabilistic power flow and maximum power supply capacity assessment as the core, it provides visualized vulnerability levels and risk grades. Based on this, operation and maintenance personnel can formulate preventive upgrade plans and reinforcement plans.

[0222] Secondly, fault isolation is faster and power outage time is shorter. The Floyd-Warshall algorithm is used to pre-generate tripping paths, and the sequence can be issued directly by querying the matrix after a fault occurs; combined with dynamic topology updates and fault tolerance strategies, it supports multiple scenarios and handling of failure to operate; the isolation range is smaller, and the power outage time of traditional methods can be reduced by 40% to 60%.

[0223] Third, it offers superior recovery decision-making and higher utilization of renewable energy. Hybrid reinforcement learning handles both discrete and continuous actions simultaneously, improving computational speed by approximately three orders of magnitude compared to heuristic algorithms such as particle swarm optimization and genetics, and reducing the number of iterations required to achieve the same goal by 40-50%. Continuous control based on DDPG allows for refined adjustment of photovoltaic, wind power, and energy storage output, increasing renewable energy consumption by approximately 10-20%, while reducing grid losses and power fluctuations. The rate of exceeding limits during the recovery process is significantly reduced, such as... Figure 8 The comparison of the recovery effects of different algorithms shown in the simulation demonstrates that the simulation has verified the effectiveness and stability in multiple sets of examples.

[0224] Fourth, it has stronger engineering applicability and more comprehensive scenario coverage. The model is designed for typical distribution networks including photovoltaic, wind power and energy storage, and is adapted to different penetration rates and wiring methods; it can be connected to existing dispatching processes without offline fine-tuning, forming an integrated closed loop of "prevention-isolation-recovery"; it supports the entire life cycle application from planned maintenance to sudden failures, and has promotional value.

[0225] Reference Figure 9 This application provides an active power distribution network operation data processing system, the system comprising:

[0226] The building module is used to construct core component models for high-proportion renewable energy distribution networks. These core component models include photovoltaic cell models, wind turbine models, and energy storage system models.

[0227] The calculation module is used to calculate the parameter probability distribution of each component in a high-proportion renewable energy distribution network under different operating scenarios. The parameter probability distribution includes load power distribution, photovoltaic irradiance distribution and wind speed distribution. The load power distribution is a normal distribution, the photovoltaic irradiance distribution is a beta distribution and the wind speed distribution is a Weiber distribution.

[0228] The analysis module is used to analyze the comprehensive vulnerability of the high-proportion new energy distribution network based on the photovoltaic cell model and the wind turbine model, combined with the parameter probability distribution. The comprehensive vulnerability includes state vulnerability and structural vulnerability.

[0229] The generation module is used to generate multi-scenario fault isolation strategies for high-proportion renewable energy distribution networks based on comprehensive weaknesses and wide-area topology strategies.

[0230] The recovery module is used to perform fault recovery on high-proportion renewable energy distribution networks based on a hybrid reinforcement learning strategy, according to multi-scenario fault isolation strategies and energy storage system models.

[0231] It is understood that the content of the above method embodiments is applicable to this system embodiment. The specific functions implemented in this system embodiment are the same as those in the above method embodiments, and the beneficial effects achieved are also the same as those achieved in the above method embodiments.

[0232] This application provides an electronic device / computer apparatus, which includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement... Figure 1 The method shown.

[0233] It is understood that the content of the above method embodiments is applicable to this device embodiment. The specific functions implemented by this device embodiment are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0234] This application provides a computer-readable storage medium storing a computer program, which is implemented when executed by a processor. Figure 1 The method shown.

[0235] It is understood that the content of the above method embodiments is applicable to this medium embodiment. The specific functions implemented in this medium embodiment are the same as those in the above method embodiments, and the beneficial effects achieved are also the same as those achieved in the above method embodiments.

[0236] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0237] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0238] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical coding feature maps; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for processing operational data of an active power distribution network, characterized in that, The method includes the following steps: Construct core component models for a high-proportion renewable energy distribution network, including photovoltaic cell models, wind turbine models, and energy storage system models; Calculate the parameter probability distribution of each component in the high-proportion new energy distribution network under different operating scenarios. The parameter probability distribution includes load power distribution, photovoltaic irradiance distribution and wind speed distribution. The load power distribution is a normal distribution, the photovoltaic irradiance distribution is a beta distribution and the wind speed distribution is a Weiber distribution. Based on the photovoltaic cell model and the wind turbine model, and combined with the parameter probability distribution, the comprehensive weakness of the high-proportion new energy distribution network is analyzed. The comprehensive weakness includes state weakness and structural weakness. Based on the comprehensive weaknesses, a multi-scenario fault isolation strategy for the high-proportion renewable energy distribution network is generated based on a wide-area topology strategy. Based on the multi-scenario fault isolation strategy and the energy storage system model, a hybrid reinforcement learning strategy is used to perform fault recovery on the high-proportion renewable energy distribution network.

2. The method according to claim 1, characterized in that, The photovoltaic cell model adopts a simplified dual-diode engineering model, and the output power of the simplified dual-diode engineering model is calculated using the following formula: ； In the formula, This represents the output power of the i-th photovoltaic cell. Indicates photoelectric conversion efficiency. This represents the light-receiving area of ​​the photovoltaic module in the photovoltaic cell model. Indicates irradiation intensity; The wind turbine model adopts a doubly-fed induction generator model, and the output power of the doubly-fed induction generator model is calculated using the following formula: ； In the formula, Indicates the first Each fan is at all times The active power output; , , These represent the measured wind speed, cut-in wind speed, rated wind speed, and cut-out wind speed, respectively. All represent the fitting coefficients. Indicates rated power; The energy storage system model adopts a composite energy storage model, and the power balance formula of the composite energy storage model is as follows: ； The capacity constraint of the composite energy storage model is as follows: ； In the formula, Indicates charge / discharge efficiency. These represent charging power and discharging power, respectively. This represents the energy storage capacity at time t. and These represent the upper and lower limits of capacity, respectively.

3. The method according to claim 2, characterized in that, The short-circuit current of the photovoltaic cell model is corrected using the following formula: ； The open-circuit voltage of the photovoltaic cell model is corrected using the following formula: ； In the formula, and These represent the short-circuit current after correction for illumination and temperature, and the open-circuit voltage after correction for temperature and illumination, respectively. and These represent the short-circuit current and open-circuit voltage before correction, respectively. b and c represent correction factors, and S represents the actual light intensity. Indicates the reference light intensity. This represents the change in light intensity, where T represents the actual temperature. This indicates temperature deviation.

4. The method according to claim 1, characterized in that, The analysis of the comprehensive weaknesses of the high-proportion renewable energy distribution network based on the photovoltaic cell model and the wind turbine model, combined with the parameter probability distribution, includes: The state vulnerability of the high-proportion renewable energy distribution network is analyzed by the probability distribution of the parameters. The structural weaknesses of the high-proportion renewable energy distribution network are analyzed using the photovoltaic cell model and the wind turbine model. The overall vulnerability of the high-proportion renewable energy distribution network is obtained by weighted summation of the state vulnerability and the structural vulnerability.

5. The method according to claim 4, characterized in that, The analysis of the state vulnerability of the high-proportion renewable energy distribution network through the parameter probability distribution includes: The uncertainty of the output power corresponding to the photovoltaic cell model and the wind turbine model is described by the beta distribution and the Weibull distribution, and the multi-scenario operation data corresponding to the high proportion of new energy distribution network is generated by combining the normal distribution. Based on the multi-scenario operation data, probabilistic power flow calculations are performed to obtain the over-limit probabilities of node voltage and line current in the high-proportion new energy distribution network. Obtain the maximum power supply capacity model of the high-proportion renewable energy distribution network; The vulnerability of the high-proportion renewable energy distribution network is analyzed based on the over-limit probability and the maximum power supply capacity model.

6. The method according to claim 5, characterized in that, The analysis of the structural weaknesses of the high-proportion renewable energy distribution network using the photovoltaic cell model and the wind turbine model includes: The nodes and lines of the high-proportion new energy distribution network are analyzed based on the photovoltaic cell model and the wind turbine model. Calculate the degree weakness of nodes based on the connection relationships between nodes; The dielectric weakness of the node is calculated based on the scenario voltage and scenario current in the multi-scenario operation data; The degree weakness of the node and the medium weakness of the node are weighted and summed to obtain the overall structural weakness of the node. The degree of weakness of the line is calculated based on the degree of weakness of the node; The dielectric weakness of the line is calculated based on the scenario power in the multi-scenario operation data; The overall structural weakness of the line is obtained by weighted summing of the line's degree weakness and its dielectric weakness. The structural weakness of the high-proportion new energy distribution network is analyzed based on the comprehensive structural weakness of the nodes and the comprehensive structural weakness of the lines.

7. The method according to claim 1, characterized in that, The step of generating a multi-scenario fault isolation strategy for the high-proportion renewable energy distribution network based on the comprehensive weaknesses and a wide-area topology strategy includes: A topology model is performed on the high-proportion new energy distribution network to obtain a connected undirected graph. The vertex set of the connected undirected graph includes buses, lines, transformers, and circuit breakers. The edge set of the connected undirected graph includes bidirectional connections between components. The bidirectional connections are quantified using an adjacency matrix. When the high-proportion new energy distribution network undergoes structural or operational changes, the adjacency matrix is ​​updated based on a topology dynamic update mechanism. Based on the updated adjacency matrix, the shortest tripping path of the high-proportion renewable energy distribution network is pre-generated; The multi-scenario fault isolation strategy for the high-proportion renewable energy distribution network is generated based on the fault type and the shortest tripping path.

8. A data processing system for the operation of an active power distribution network, characterized in that, The system includes: The construction module is used to build core component models of a high-proportion renewable energy distribution network. The core component models include photovoltaic cell models, wind turbine models, and energy storage system models. The calculation module is used to calculate the parameter probability distribution of each component in the high-proportion new energy distribution network under different operating scenarios. The parameter probability distribution includes load power distribution, photovoltaic irradiance distribution and wind speed distribution. The load power distribution is a normal distribution, the photovoltaic irradiance distribution is a beta distribution and the wind speed distribution is a Weiber distribution. The analysis module is used to analyze the comprehensive weakness of the high-proportion new energy distribution network based on the photovoltaic cell model and the wind turbine model, combined with the parameter probability distribution. The comprehensive weakness includes state weakness and structural weakness. The generation module is used to generate a multi-scenario fault isolation strategy for the high-proportion renewable energy distribution network based on the comprehensive vulnerability and a wide-area topology strategy. The recovery module is used to perform fault recovery on the high-proportion renewable energy distribution network based on the multi-scenario fault isolation strategy and the energy storage system model, using a hybrid reinforcement learning strategy.

9. An electronic device, characterized in that, The electronic device / computer apparatus includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the method according to any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it implements the method of any one of claims 1 to 7.

Images