Method, system, equipment and medium for toughness partitioning and fault recovery of power distribution cooperative network

Abstract

The invention provides a method, a system, equipment and a medium for toughness partitioning and fault recovery of a power distribution cooperative network, and the method comprises the steps: constructing a dynamic weighted undirected graph based on a node admittance matrix, and introducing the physical connection state and operation characteristics of the power distribution cooperative network into a partitioning modeling process; and the partitioning result can truly reflect the line coupling strength and the change of the equipment operation state. Initial population construction is carried out by adopting a position-based node coding strategy and combining historical operation data, so that an invalid search space is reduced. The internal stability of the power supply island and the partition adjustment cost are taken as optimization objectives, and a partition scheme is iteratively optimized by using a multi-objective non-dominated sorting genetic algorithm NSGA-II, so that the operation and maintenance cost is reduced. According to the method, the dynamic toughness partition map is generated from the Pareto optimal partition scheme, and the scheduling scheme is generated, so that emergency power supply scheduling and line maintenance decisions can be cooperatively generated with partition results, and the pertinence and overall efficiency of fault recovery are improved.

Description

Technical Field

[0001] This invention belongs to the field of power distribution cooperative network technology, and specifically relates to a method, system, equipment and medium for resilient zoning and fault recovery in power distribution cooperative networks. Background Technology

[0002] Distribution collaborative network refers to an intelligent network that, based on smart distribution network, deeply integrates physical distribution system with information and communication system through advanced sensing, communication and control technologies, forming a network with comprehensive perception, dynamic analysis, autonomous decision-making and collaborative execution capabilities.

[0003] Resilient recovery technology for distribution network coordination is a crucial foundation for ensuring the stability of coordination performance. Currently, mainstream methods still rely primarily on static zoning and manual experience. Typically, fixed configuration areas are pre-defined based on the geographical distribution or load type of participating elements, and redundant paths and backup resources are configured to cope with sudden failures. These methods depend on static analysis of historical data; however, in actual operation, the topology of the coordination network often changes dynamically due to extreme interference or load fluctuations, making it difficult for pre-defined zoning to match real-time demands. Furthermore, existing technologies often focus on single-resource scheduling, limiting resource utilization.

[0004] The main shortcomings of existing technologies lie in three aspects: dynamic adaptability, collaborative optimization capability, and intelligence level. Static partitioning cannot respond in real time to changes in the collaborative network structure, and under extreme events, the fault isolation range often exceeds actual needs, causing unnecessary losses. Meanwhile, traditional methods rely on human experience to formulate recovery strategies, resulting in slow response times and difficulty in dealing with uncertainties such as fluctuations in resource capacity or communication repair delays. For example, when the collaborative effects between dispersed elements change abruptly, the lag in human decision-making leads to low efficiency in critical load recovery. Furthermore, existing resilience assessment models neglect the dynamic correlation characteristics within the collaborative network, failing to accurately identify fault propagation paths, leading to unreasonable resource allocation and further increasing operation and maintenance costs. Summary of the Invention

[0005] To overcome the problems in existing technologies where static partitioning of power distribution cooperative networks cannot respond in real time to changes in the cooperative network structure, and where the fault isolation range often exceeds actual needs under extreme events, this invention provides a method, system, device, and medium for resilient partitioning and fault recovery in power distribution cooperative networks.

[0006] To achieve the above objectives, the present invention provides the following technical solution:

[0007] In a first aspect, embodiments of this disclosure provide a method for resilient zoning and fault recovery in a power distribution cooperative network, comprising the following steps:

[0008] Step S1: Based on the node admittance matrix, abstract the power distribution coordination network at time t into a dynamic weighted undirected graph;

[0009] Step S2: Based on the dynamic weighted undirected graph, a location-based node coding strategy is adopted to transform the partitioning problem of the power distribution cooperative network into a chromosome coding problem, and an initial population representing different partitioning schemes is generated.

[0010] Step S3: With the optimization objectives of maximizing the internal stability of the power supply island and minimizing the partitioning adjustment cost, a multi-objective fitness function is constructed, and the initial population representing different partitioning schemes is iteratively optimized using the multi-objective non-dominated sorting genetic algorithm NSGA-II to select the Pareto optimal set of partitioning schemes.

[0011] Step S4: Select a compromise solution from the Pareto optimal partitioning scheme set, generate a dynamic resilience partitioning map, and formulate switching operation instructions and resource scheduling schemes based on the map.

[0012] Furthermore, in step S1, the dynamic weighted undirected graph In the formula, For a set of nodes, This represents the critical load nodes and source nodes in the power distribution network; Let edge be the set of edges, representing the physical connections between nodes; This is a weight matrix, representing the line at time [time]. The overall connection strength;

[0013] Based on nodal admittance matrix The physical topology of the distribution network is quantified, including the mutual admittance. : Represents a node and The negative value of the branch admittance reflects the physical coupling strength between nodes;

[0014] Based on mutual admittance Initialize the weight matrix edge weight :

[0015] ;

[0016] In the formula, the weight matrix , For the first Line number Column elements are edges At any moment The weight value; when the node With nodes When there is no physical connection between them, let ; The health status of the device is between 0 and 1; and All are normalized coefficients.

[0017] Furthermore, in step S2, generating the initial population representing different partitioning schemes includes the following steps:

[0018] Step A1: Employ a location-based node coding strategy to encode each distribution node in the dynamic weighted undirected graph. For a gene locus on a chromosome, the node encoding strategy includes:

[0019] Gene definition: Value of each gene locus Represents a node In a distribution network, topology adjacency selection is used to determine the relationship between nodes. Physical neighbor nodes belonging to the same power supply area;

[0020] Encoding rules: Gene values Must start from node The dynamic weighted undirected graph physical neighbor set Selected from;

[0021] Step A2: Process historical fault data using a sliding window mechanism, assuming the current time is... Select historical time window The spatiotemporal correlation weights were calculated. Based on spatiotemporal correlation weights And the initialization results of gene values ​​for fusing historical admittance features obtained from the dynamic weighted undirected graph:

[0022] Define historical similarity index :

[0023] ;

[0024] In the formula, and They are time points and time Based on the weight matrix of the dynamic weighted undirected graph Binarized adjacency matrix; For Hadamah accumulation; This is a structural feature extraction function used to extract the topological features of node connections from the adjacency matrix, and the output reflects the structural feature pattern of the network at the corresponding time.

[0025] Select similarity index The highest historical moment If the current node The mutual admittance between it and its neighboring nodes in the historical optimal solution It remains at a high value, and the nodal autoadmittance If no significant abnormal changes are observed, and the physical connection of the line is considered to be good, then probability is used. Inherit the historical gene value; otherwise, use the current physical neighbor set. A node is randomly selected as the historical gene value, and all obtained historical gene values ​​are used as the gene value initialization result; wherein, the mutual admittance "Still high" means that the current mutual admittance level, after comparing with historical operating data statistics, is comparable to the normal mutual admittance. Consistent horizontality; self-admittance Represents a node The sum of the admittances of all connected branches is used to reflect the node's own load throughput capacity.

[0026] Step A3: Use quasi-randomization at the initial time. The sequence generates a uniformly distributed initial population. At non-initial times, a reverse learning mechanism is introduced to process the gene value initialization results and generate reverse individuals. The initial population and reverse individuals are used as initial populations representing different partitioning schemes.

[0027] Furthermore, the multi-objective fitness function in step S3 includes:

[0028] ;

[0029] in:

[0030] ; ;

[0031] In the formula, This indicates the independence of the power supply island. The higher the value, the tighter the internal connections within the defined distribution community and the sparser the connections between communities; For the community The sum of the weights of the internal power lines is obtained based on the dynamic weighted undirected graph. It is the sum of the node degrees within the community. This represents the total network line weight. Used to measure the smoothness of the recovery strategy. Based on the current community division results, The result of dividing historical moments. To standardize mutual information.

[0032] Furthermore, step S3, which involves iteratively optimizing the initial populations representing different partitioning schemes based on a multi-objective fitness function and utilizing the multi-objective non-dominated sorting genetic algorithm NSGA-II to select the Pareto optimal set of partitioning schemes, includes the following steps:

[0033] The multi-objective non-dominated sorting genetic algorithm NSGA-II includes crossover and mutation operations. The crossover operation is configured to retain frequently occurring strong connection links based on a historical disaster scenario database. The strong connection links are lines that remain connected in multiple failures.

[0034] The mutation operation is configured based on the network state equation: In the formula, It represents the node voltage state vector, reflecting the operating voltage or energy level of each distribution node; : Represents the node injected current vector, indicating whether the node is injecting or emitting power; represents the node voltage state vector. As feedback, the node voltage state vector is improved. The mutation probability of a node with an abnormal value makes the node more likely to leave the current community to seek new power sources; and it also reduces the node's voltage state vector. The mutation probability of a value-stable node;

[0035] All individuals in the population are based on The values ​​and Adp-HoNMI values ​​are sorted non-dominated, and individuals located at the Pareto front are retained first. These individuals represent the partitioning schemes that achieve the best balance between "fault isolation effect" and "operational adjustment cost". The selected individuals are used as the set of Pareto optimal partitioning schemes.

[0036] Furthermore, the multi-objective non-dominated sorting genetic algorithm NSGA-II includes:

[0037] In the non-dominated sorting stage, the multi-objective non-dominated sorting genetic algorithm NSGA-II introduces line weight information from a dynamic weighted undirected graph as an auxiliary criterion when multiple individuals are at the same non-dominated level. It prioritizes retaining partitioning schemes with higher line weights and closer connections to critical loads within the power supply island.

[0038] In the process of calculating crowding distance, the distribution density of individuals is corrected by combining the similarity information between the current partitioning scheme and the historical partitioning scheme;

[0039] During the crossover and mutation phases, the crossover and mutation probabilities of different nodes are adaptively adjusted based on the real-time operating status of the power distribution network. This makes it easier for nodes with abnormal operating status or large voltage fluctuations to break out of their original community structure and obtain feasible partitioning schemes that meet the constraints of the current fault scenario through search.

[0040] Furthermore, in step S4, a compromise solution is selected to generate a dynamic resilience partitioning map. In the dynamic resilience partitioning map, the interconnection switches within the community remain closed, while the interconnection switches at the community boundary are open, forming an independent power supply island at the physical level.

[0041] Based on the dynamic resilience partitioning map, characteristic indicators of each community are extracted to generate a resource scheduling list, including:

[0042] Emergency Power Vehicle Dispatch: Based on the network state equation, the "source-load" imbalance within each community is calculated. If the imbalance is too large, it indicates insufficient power generation capacity within the community. An automatic instruction is generated to dispatch a mobile emergency power vehicle to connect to the critical nodes within that community. These critical nodes are those with low voltage levels or high load proportions within the community. The network state equation is as follows: In the formula, It represents the node voltage state vector, reflecting the operating voltage or energy level of each distribution node; : Represents the node injected current vector, as the power injection or outflow of the node; the large imbalance refers to the source-load imbalance in the community being significantly higher than the source-load balance level of the community under normal operating conditions;

[0043] Maintenance team assignment: Based on the weight matrix of a dynamic weighted undirected graph, high-weight faulty lines severed on community boundaries are identified. Combined with the geographic center coordinates of the community, the optimal path for the maintenance team is planned, prioritizing the repair of the connecting trunk lines located at the community boundaries that connect two high-resilience communities. Here, a high-resilience community refers to a community with high internal stability and good operating status in the Pareto optimal partitioning scheme. The connecting trunk line refers to a physical connection line located between different communities, with the corresponding edge weight being at a high level among the community boundary lines.

[0044] In a second aspect, embodiments of this disclosure provide a system for resilient zoning and fault recovery in a power distribution cooperative network, comprising:

[0045] The preprocessing unit is configured to abstract the power distribution coordination network at time t into a dynamic weighted undirected graph based on the node admittance matrix;

[0046] The initialization unit is configured to: based on the dynamic weighted undirected graph, adopt a position-based node encoding strategy to transform the partitioning problem of the power distribution cooperative network into a chromosome encoding problem, and generate an initial population representing different partitioning schemes;

[0047] The screening unit is configured to: construct a multi-objective fitness function with the optimization objectives of maximizing the internal stability of the power supply island and minimizing the partitioning adjustment cost, and use the multi-objective non-dominated sorting genetic algorithm NSGA-II to iteratively optimize the initial population representing different partitioning schemes to screen out the Pareto optimal set of partitioning schemes.

[0048] The output unit is configured to: select a compromise solution from the Pareto optimal partitioning scheme set, generate a dynamic resilience partitioning map, and formulate switching operation instructions and resource scheduling schemes based on the map.

[0049] In a third aspect, embodiments of this disclosure provide an electronic device, characterized in that the electronic device comprises:

[0050] At least one processor; and,

[0051] The memory is communicatively connected to the at least one processor; wherein,

[0052] The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the method of distribution network resilience partitioning and fault recovery.

[0053] In a fourth aspect, embodiments of this disclosure provide a non-transitory computer-readable storage medium, characterized in that the non-transitory computer-readable storage medium stores computer instructions for causing the computer to execute the method of distribution network resilience partitioning and fault recovery.

[0054] Compared with the prior art, the present invention has the following beneficial technical effects:

[0055] This invention provides a method, system, device, and medium for resilient partitioning and fault recovery in distribution network coordination. The method constructs a dynamic weighted undirected graph based on the node admittance matrix, incorporating the physical connection state and operational characteristics of the distribution network into the partitioning modeling process. This ensures that the partitioning results accurately reflect changes in line coupling strength and equipment operating status. By employing a location-based node encoding strategy and combining it with historical operational data for initial population construction, highly feasible candidate schemes are introduced in the early stages of partitioning optimization, reducing the ineffective search space and improving the convergence efficiency of partitioning calculations. Secondly, by using the internal stability of power supply islands and the cost of partitioning adjustments as optimization objectives, the multi-objective non-dominated sorting genetic algorithm NSGA-II is used to iteratively optimize the partitioning scheme. This ensures fault isolation while reducing the maintenance costs caused by frequent partitioning adjustments at adjacent times. Finally, by generating a dynamic resilient partitioning map from the Pareto optimal partitioning scheme and formulating switching operation instructions and resource scheduling schemes accordingly, emergency power dispatching and line maintenance decisions can be generated collaboratively with the partitioning results, improving the targeting and overall efficiency of fault recovery. Attached Figure Description

[0056] Figure 1 A flowchart illustrating a method for resilient zoning and fault recovery in a power distribution cooperative network according to an embodiment of this disclosure is shown.

[0057] Figure 2 The generated dynamic resilience partitioning map according to an embodiment of this disclosure is shown;

[0058] Figure 3 A diagram of a distribution network resilience partitioning and fault recovery apparatus according to an embodiment of the present disclosure is shown. Detailed Implementation

[0059] The embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0060] This disclosure uses a "distribution collaborative network" as a specific application scenario for illustration. In this scenario, nodes represent substations, ring main units, or distributed power sources, and edges represent power lines or communication optical cables. It aims to address the problems of traditional static partitioning being unable to adapt and resource scheduling lagging during fault recovery when the distribution collaborative network encounters frequent topology changes due to extreme natural disasters or network attacks.

[0061] Figure 1 A flowchart 100 of a method for resilient zoning and fault recovery of a power distribution cooperative network according to an embodiment of this disclosure is shown, including the following steps:

[0062] In step S101, the power distribution cooperative network at time t is abstracted into a dynamic weighted undirected graph based on the node admittance matrix.

[0063] Specifically, the dynamic weighted undirected graph in step S101 In the formula, For a set of nodes, This represents the critical load nodes and source nodes in the power distribution network; Let edge be the set of edges, representing the physical connections between nodes; This is a weight matrix, representing the line at time [time]. The overall connection strength;

[0064] Based on nodal admittance matrix The physical topology of the distribution network is quantified, including the mutual admittance. : Represents a node and The negative value of the branch admittance reflects the physical coupling strength between nodes;

[0065] Based on mutual admittance Initialize the weight matrix edge weight :

[0066] ;

[0067] In the formula, the weight matrix , For the first Line number Column elements are edges At any moment The weight value; when the node With nodes When there is no physical connection between them, let ; The health status of the device is between 0 and 1; and All are normalized coefficients.

[0068] This step maps the physical world's power distribution network into a mathematical model that the algorithm can process, and initializes the network weights using physical parameters to generate a dynamic weighted undirected graph, providing a data foundation for subsequent community partitioning. In other words, the admittance parameters of the physical layer are transformed into edge weights in the dynamic weighted undirected graph model; lines with high admittance, i.e., low impedance, are more likely to be partitioned into the same community in the subsequent community partitioning.

[0069] Next, proceed to step S102.

[0070] In step S102, based on the dynamic weighted undirected graph, a location-based node encoding strategy is adopted to transform the partitioning problem of the power distribution cooperative network into a chromosome encoding problem, and an initial population representing different partitioning schemes is generated.

[0071] Specifically, when generating the initial population representing different partitioning schemes in step S2, the following steps are included:

[0072] Step A1: Employ a location-based node coding strategy to encode each distribution node in the dynamic weighted undirected graph. For a gene locus on a chromosome, the node encoding strategy includes:

[0073] Gene definition: Value of each gene locus Represents a node In a distribution network, topology adjacency selection is used to determine the relationship between nodes. Physical neighbor nodes belonging to the same power supply area;

[0074] Encoding rules: Gene values Must start from node The dynamic weighted undirected graph physical neighbor set Selected from;

[0075] Step A2: Process historical fault data using a sliding window mechanism, assuming the current time is... Select historical time window The spatiotemporal correlation weights were calculated. Based on spatiotemporal correlation weights And the initialization results of gene values ​​for fusing historical admittance features obtained from the dynamic weighted undirected graph:

[0076] Define historical similarity index :

[0077] ;

[0078] In the formula, and They are time points and time Based on the weight matrix of the dynamic weighted undirected graph Binarized adjacency matrix; For Hadamah accumulation; This is a structural feature extraction function used to extract the topological features of node connections from the adjacency matrix, and the output reflects the structural feature pattern of the network at the corresponding time.

[0079] Select similarity index The highest historical moment If the current node The mutual admittance between it and its neighboring nodes in the historical optimal solution It remains at a high value, and the nodal autoadmittance If no significant abnormal changes are observed, and the physical connection of the line is considered to be good, then probability is used. Inherit the historical gene value; otherwise, use the current physical neighbor set. A node is randomly selected as the historical gene value, and all obtained historical gene values ​​are used as the gene value initialization result; wherein, the mutual admittance "Still high" means that the current mutual admittance level, after comparing with historical operating data statistics, is comparable to the normal mutual admittance. Consistent horizontality; self-admittance Represents a node The sum of the admittances of all connected branches is used to reflect the node's own load throughput capacity.

[0080] Step A3: Use quasi-randomization at the initial time. A uniformly distributed initial population is generated from the sequence. At non-initial time points, a reverse learning mechanism is introduced to process the gene value initialization results, generating reverse individuals. The initial population and reverse individuals are used as initial populations representing different partitioning schemes. The reverse learning mechanism specifically involves: if a node in a partitioning scheme represented by a certain individual... Belongs to the community Then, a reverse individual is generated, whose corresponding community ID is mapped as follows: This increases population diversity and prevents the algorithm from getting trapped in local optima.

[0081] Next, proceed to step S103.

[0082] In step S103, with the optimization objectives of maximizing the internal stability of the power supply island and minimizing the partitioning adjustment cost, a multi-objective fitness function is constructed, and an evolutionary algorithm is used to iteratively optimize the initial population representing different partitioning schemes to select the Pareto optimal set of partitioning schemes.

[0083] Specifically, the multi-objective fitness function in step S103 includes:

[0084] ;

[0085] in:

[0086] ; ;

[0087] In the formula, This indicates the independence of the power supply island. The higher the value, the tighter the internal connections within the defined distribution community and the sparser the connections between communities; For the community The sum of the weights of the internal power lines is obtained based on the dynamic weighted undirected graph. It is the sum of the node degrees within the community. This represents the total network line weight. Used to measure the smoothness of the recovery strategy. Based on the current community division results, The result of dividing historical moments. To standardize mutual information;

[0088] Furthermore, it should be explained that the genetic operator design step in the multi-objective non-dominated sorting genetic algorithm NSGA-II is used to guide the population to evolve in a direction that satisfies real-time operational constraints while maintaining the stability of critical power supply paths. Step S103, based on a multi-objective fitness function and using the multi-objective non-dominated sorting genetic algorithm NSGA-II to iteratively optimize the initial population representing different partitioning schemes and select the Pareto optimal set of partitioning schemes, includes the following steps:

[0089] The multi-objective non-dominated sorting genetic algorithm NSGA-II includes crossover and mutation operations. The crossover operation is configured to retain frequently occurring strong connection links based on a historical disaster scenario database. The strong connection links are lines that remain connected in multiple failures.

[0090] The mutation operation is configured based on the network state equation: In the formula, It represents the node voltage state vector, reflecting the operating voltage or energy level of each distribution node; : Represents the node injected current vector, indicating whether the node is injecting or emitting power; represents the node voltage state vector. As feedback, the node voltage state vector is improved. The mutation probability of a node with an abnormal value makes the node more likely to leave the current community to seek new power sources; and it also reduces the node's voltage state vector. The mutation probability of a value-stable node;

[0091] All individuals in the population are based on The values ​​and Adp-HoNMI values ​​are sorted non-dominated, and individuals located at the Pareto front are retained first. These individuals represent the partitioning schemes that achieve the best balance between "fault isolation effect" and "operational adjustment cost". The selected individuals are used as the set of Pareto optimal partitioning schemes.

[0092] The multi-objective non-dominated sorting genetic algorithm NSGA-II improves upon the NSGA-II algorithm. The improved NSGA-II algorithm does not change the algorithm framework, but rather, while maintaining the original non-dominated sorting and elite retention mechanism of NSGA-II, it makes targeted improvements to individual evaluation and genetic operators based on the physical characteristics and operational features of the power distribution cooperative network.

[0093] Specifically, in the non-dominated sorting stage, when multiple individuals are at the same non-dominated level, the line weight information in the dynamic weighted undirected graph is introduced as an auxiliary criterion to prioritize the retention of partition schemes with higher line weights and closer connections to critical loads within the power supply island, thereby enhancing the power supply stability after fault isolation.

[0094] In the process of calculating crowding distance, the distribution density of individuals is corrected by combining the similarity information between the current partitioning scheme and the historical partitioning scheme. This maintains population diversity while avoiding drastic changes in partitioning results at adjacent time points, thus improving the smoothness of the evolution of resilient partitioning over time.

[0095] Furthermore, during the crossover and mutation phases, the crossover and mutation probabilities of different nodes are adaptively adjusted based on the real-time operating status of the power distribution network. This makes it easier for nodes with abnormal operating status or large voltage fluctuations to break out of their original community structure, enabling the algorithm to find feasible partitioning schemes that meet the constraints of the current fault scenario more quickly.

[0096] Next, proceed to step S104.

[0097] In step S104, a compromise solution is selected from the Pareto optimal partitioning scheme set to generate a dynamic resilience partitioning map, and switching operation instructions and resource scheduling schemes are formulated based on the map.

[0098] Specifically, in step S104, a compromise solution is selected to generate a dynamic resilience partition map. In the dynamic resilience partition map, the interconnection switches inside the community remain closed, while the interconnection switches at the community boundary are opened, forming an independent power supply island at the physical level. Figure 2 The generated dynamic resilience partitioning map of this disclosure embodiment is shown, such as Figure 2 As shown in the figure, the different colored node sets represent different resilient communities, denoted as community A, community B, and community C. Nodes within the same community maintain strong physical connections, while electrical isolation between different communities is achieved by disconnecting them with switches.

[0099] Based on the dynamic resilience partitioning map, characteristic indicators of each community are extracted to generate a resource scheduling list, including:

[0100] Emergency Power Vehicle Dispatch: Based on the network state equation, the "source-load" imbalance within each community is calculated. If the imbalance is too large, it indicates insufficient power generation capacity within the community. An automatic instruction is generated to dispatch a mobile emergency power vehicle to connect to the critical nodes within that community. These critical nodes are those with low voltage levels or high load proportions within the community. The network state equation is as follows: In the formula, It represents the node voltage state vector, reflecting the operating voltage or energy level of each distribution node; : Represents the node injected current vector, as the power injection or outflow of the node; the large imbalance refers to the source-load imbalance in the community being significantly higher than the source-load balance level of the community under normal operating conditions;

[0101] Maintenance team assignment: Based on the weight matrix of a dynamic weighted undirected graph, high-weight faulty lines severed on community boundaries are identified. Combined with the geographic center coordinates of the community, the optimal path for the maintenance team is planned, prioritizing the repair of the connecting trunk lines located at the community boundaries that connect two high-resilience communities. Here, a high-resilience community refers to a community with high internal stability and good operating status in the Pareto optimal partitioning scheme. The connecting trunk line refers to a physical connection line located between different communities, with the corresponding edge weight being at a high level among the community boundary lines.

[0102] The second embodiment of the present invention also provides a system for resilient zoning and fault recovery of a power distribution cooperative network, comprising:

[0103] The preprocessing unit is configured to abstract the power distribution coordination network at time t into a dynamic weighted undirected graph based on the node admittance matrix;

[0104] The initialization unit is configured to: based on the dynamic weighted undirected graph, adopt a position-based node encoding strategy to transform the partitioning problem of the power distribution cooperative network into a chromosome encoding problem, and generate an initial population representing different partitioning schemes;

[0105] The screening unit is configured to: construct a multi-objective fitness function with the optimization objectives of maximizing the internal stability of the power supply island and minimizing the partitioning adjustment cost, and use the multi-objective non-dominated sorting genetic algorithm NSGA-II to iteratively optimize the initial population representing different partitioning schemes to screen out the Pareto optimal set of partitioning schemes.

[0106] The output unit is configured to: select a compromise solution from the Pareto optimal partitioning scheme set, generate a dynamic resilience partitioning map, and formulate switching operation instructions and resource scheduling schemes based on the map.

[0107] The third embodiment of the present invention also provides an electronic device, the electronic device comprising:

[0108] At least one processor; and,

[0109] The memory is communicatively connected to the at least one processor; wherein,

[0110] The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the distribution cooperative network resilience partitioning and fault recovery method of any of the foregoing embodiments.

[0111] The fourth embodiment of the present invention also provides a non-transitory computer-readable storage medium storing computer instructions for causing the computer to execute the distribution network resilience partitioning and fault recovery method described in any of the foregoing embodiments.

[0112] The fifth embodiment of the present invention also provides a computer program product, which includes a computing program stored on a non-transitory computer-readable storage medium. The computer program includes program instructions that, when executed by a computer, cause the computer to perform the distribution network resilience partitioning and fault recovery method of any of the foregoing embodiments.

[0113] Figure 3 The illustration shows a method or device 1000 implementing an embodiment of the present invention. In some embodiments, more or fewer devices may be included than illustrated. In some embodiments, it may be implemented using a single or multiple devices. In some embodiments, it may be implemented using cloud-based or distributed devices.

[0114] like Figure 3 As shown, device 1000 includes a processor 1001, which can perform various appropriate operations and processes based on programs and / or data stored in read-only memory (ROM) 1002 or programs and / or data loaded from storage portion 1008 into random access memory (RAM) 1003. Processor 1001 may be a multi-core processor or may contain multiple processors. In some embodiments, processor 1001 may include a general-purpose main processor and one or more special coprocessors, such as a central processing unit (CPU), graphics processing unit (GPU), neural network processor (NPU), digital signal processor (DSP), etc. Various programs and data required for the operation of device 1000 are also stored in RAM 1003. Processor 1001, ROM 1002, and RAM 1003 are interconnected via bus 1004. Input / output (I / O) interface 1005 is also connected to bus 1004.

[0115] The processor and memory described above are used together to execute a program stored in the memory. When the program is executed by a computer, it can implement the methods, steps, or functions described in the above embodiments.

[0116] The following components are connected to I / O interface 1005: an input section 1006 including a keyboard, mouse, touchscreen, etc.; an output section 1007 including a cathode ray tube (CRT), liquid crystal display (LCD), etc., and speakers, etc.; a storage section 1008 including a hard disk, etc.; and a communication section 1009 including a network interface card such as a LAN card, modem, etc. The communication section 1009 performs communication processing via a network such as the Internet. A drive 1010 is also connected to I / O interface 1005 as needed. A removable medium 1011, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., is installed on drive 1010 as needed so that computer programs read from it can be installed into storage section 1008 as needed. Figure 3 The diagram only shows a portion of the components and does not imply that the device 1000 only includes... Figure 3 The components shown.

[0117] The systems, devices, modules, or units described in the above embodiments can be implemented by a computer or its associated components. The computer may be, for example, a mobile terminal, smartphone, personal computer, laptop computer, in-vehicle human-machine interface device, personal digital assistant, media player, navigation device, game console, tablet computer, wearable device, smart TV, Internet of Things system, smart home, industrial computer, server, or a combination thereof.

[0118] Although not shown, in this embodiment of the invention, a computer-readable storage medium is provided having a computer program / instruction stored thereon, which, when executed by a processor, implements the method for resilient partitioning and fault recovery of the power distribution cooperative network described in the embodiment.

[0119] Storage media in embodiments of the present invention include articles that are permanent and non-permanent, removable and non-removable, capable of storing information by any method or technology. Examples of storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic magnetic disk storage or other magnetic storage devices, or any other non-transfer medium that can be used to store information accessible by a computing device.

[0120] Although not shown, embodiments of the present invention also provide a computer program product, including: a computer program / instructions that, when executed by a processor, implement the method for resilient partitioning and fault recovery of a power distribution cooperative network as described in the embodiments.

[0121] The methods, programs, systems, apparatuses, etc., in embodiments of the present invention can be executed or implemented in one or more networked computers, or practiced in a distributed computing environment. In the embodiments of this specification, in these distributed computing environments, tasks can be performed by remote processing devices connected via a communication network.

Claims

1. A method for resilient zoning and fault recovery in a power distribution cooperative network, characterized in that, Includes the following steps: Step S1: Based on the node admittance matrix, abstract the power distribution coordination network at time t into a dynamic weighted undirected graph; Step S2: Based on the dynamic weighted undirected graph, a location-based node coding strategy is adopted to transform the partitioning problem of the power distribution cooperative network into a chromosome coding problem, and an initial population representing different partitioning schemes is generated. Step S3: With the optimization objectives of maximizing the internal stability of the power supply island and minimizing the partitioning adjustment cost, a multi-objective fitness function is constructed, and the initial population representing different partitioning schemes is iteratively optimized using the multi-objective non-dominated sorting genetic algorithm NSGA-II to select the Pareto optimal set of partitioning schemes. Step S4: Select a compromise solution from the Pareto optimal partitioning scheme set, generate a dynamic resilience partitioning map, and formulate switching operation instructions and resource scheduling schemes based on the map.

2. The method for resilient zoning and fault recovery of power distribution cooperative networks according to claim 1, characterized in that, The dynamic weighted undirected graph in step S1 In the formula, For a set of nodes, This represents the critical load nodes and source nodes in the power distribution network; Let edge be the set of edges, representing the physical connections between nodes; This is a weight matrix, representing the line at time [time]. The overall connection strength; Based on nodal admittance matrix The physical topology of the distribution network is quantified, including the mutual admittance. : Represents a node and The negative value of the branch admittance reflects the physical coupling strength between nodes; Based on mutual admittance Initialize the weight matrix edge weight : ； In the formula, the weight matrix , For the first Line number Column elements are edges At any moment The weight value; when the node With nodes When there is no physical connection between them, let ; The health status of the device is between 0 and 1; and All are normalized coefficients.

3. The method for resilient zoning and fault recovery of power distribution cooperative networks according to claim 1, characterized in that, When generating the initial population representing different partitioning schemes in step S2, the following steps are included: Step A1: Employ a location-based node coding strategy to encode each distribution node in the dynamic weighted undirected graph. For a gene locus on a chromosome, the node encoding strategy includes: Gene definition: Value of each gene locus Represents a node In a distribution network, topology adjacency selection is used to determine the relationship between nodes. Physical neighbor nodes belonging to the same power supply area; Encoding rules: Gene values Must start from node The dynamic weighted undirected graph physical neighbor set Selected from; Step A2: Process historical fault data using a sliding window mechanism, assuming the current time is... Select historical time window The spatiotemporal correlation weights were calculated. Based on spatiotemporal correlation weights And the initialization results of gene values ​​for fusing historical admittance features obtained from the dynamic weighted undirected graph: Define historical similarity index : ； In the formula, and They are time points and time Based on the weight matrix of the dynamic weighted undirected graph Binarized adjacency matrix; For Hadamah accumulation; This is a structural feature extraction function used to extract the topological features of node connections from the adjacency matrix, and the output reflects the structural feature pattern of the network at the corresponding time. Select similarity index The highest historical moment If the current node The mutual admittance between it and its neighboring nodes in the historical optimal solution It remains at a high value, and the nodal autoadmittance If no significant abnormal changes are observed, and the physical connection of the line is considered to be good, then probability is used. Inherit the historical gene value; otherwise, use the current physical neighbor set. A node is randomly selected as the historical gene value, and all obtained historical gene values ​​are used as the gene value initialization result; wherein, the mutual admittance "Still high" means that the current mutual admittance level, after comparing with historical operating data statistics, is comparable to the normal mutual admittance. Consistent horizontality; self-admittance Represents a node The sum of the admittances of all connected branches is used to reflect the node's own load throughput capacity. Step A3: Use quasi-randomization at the initial time. The sequence generates a uniformly distributed initial population. At non-initial times, a reverse learning mechanism is introduced to process the gene value initialization results and generate reverse individuals. The initial population and reverse individuals are used as initial populations representing different partitioning schemes.

4. The method for resilient zoning and fault recovery of power distribution cooperative networks according to claim 1, characterized in that, The multi-objective fitness function in step S3 includes: ； in: ； ； In the formula, This indicates the independence of the power supply island. The higher the value, the tighter the internal connections within the defined distribution community and the sparser the connections between communities; For the community The sum of the weights of the internal power lines is obtained based on the dynamic weighted undirected graph. It is the sum of the node degrees within the community. This represents the total network line weight. Used to measure the smoothness of the recovery strategy. Based on the current community division results, The result of dividing historical moments. To standardize mutual information.

5. The method for resilient zoning and fault recovery of power distribution cooperative networks according to claim 4, characterized in that, Step S3, which uses a multi-objective fitness function and the multi-objective non-dominated sorting genetic algorithm NSGA-II to iteratively optimize the initial population representing different partitioning schemes and select the Pareto optimal set of partitioning schemes, includes the following steps: The multi-objective non-dominated sorting genetic algorithm NSGA-II includes crossover and mutation operations. The crossover operation is configured to retain frequently occurring strong connection links based on a historical disaster scenario database. The strong connection links are lines that remain connected in multiple failures. The mutation operation is configured based on the network state equation: In the formula, It represents the node voltage state vector, reflecting the operating voltage or energy level of each distribution node; : Represents the node injected current vector, indicating whether the node is injecting or emitting power; represents the node voltage state vector. As feedback, the node voltage state vector is improved. The mutation probability of a node with an abnormal value makes the node more likely to leave the current community to seek new power sources; and it also reduces the node's voltage state vector. The mutation probability of a value-stable node; All individuals in the population are based on The values ​​and Adp-HoNMI values ​​are sorted non-dominated, and individuals located at the Pareto front are retained first. These individuals represent the partitioning schemes that achieve the best balance between "fault isolation effect" and "operational adjustment cost". The selected individuals are used as the set of Pareto optimal partitioning schemes.

6. The method for resilient zoning and fault recovery of power distribution cooperative networks according to claim 5, characterized in that, The multi-objective non-dominated sorting genetic algorithm NSGA-II includes: In the non-dominated sorting stage, the multi-objective non-dominated sorting genetic algorithm NSGA-II introduces line weight information from a dynamic weighted undirected graph as an auxiliary criterion when multiple individuals are at the same non-dominated level. It prioritizes retaining partitioning schemes with higher line weights and closer connections to critical loads within the power supply island. In the process of calculating crowding distance, the distribution density of individuals is corrected by combining the similarity information between the current partitioning scheme and the historical partitioning scheme; During the crossover and mutation phases, the crossover and mutation probabilities of different nodes are adaptively adjusted based on the real-time operating status of the power distribution network. This makes it easier for nodes with abnormal operating status or large voltage fluctuations to break out of their original community structure and obtain feasible partitioning schemes that meet the constraints of the current fault scenario through search.

7. The method for resilient zoning and fault recovery of power distribution cooperative networks according to claim 1, characterized in that, In step S4, a compromise solution is selected to generate a dynamic resilience partitioning map. The interconnection switches within the community in the dynamic resilience partitioning map remain closed, while the interconnection switches at the community boundary are open, forming an independent power supply island at the physical level. Based on the dynamic resilience partitioning map, characteristic indicators of each community are extracted to generate a resource scheduling list, including: Emergency Power Vehicle Dispatch: Based on the network state equation, the "source-load" imbalance within each community is calculated. If the imbalance is too large, it indicates insufficient power generation capacity within the community. An automatic instruction is generated to dispatch a mobile emergency power vehicle to connect to the critical nodes within that community. These critical nodes are those with low voltage levels or high load proportions within the community. The network state equation is as follows: In the formula, It represents the node voltage state vector, reflecting the operating voltage or energy level of each distribution node; : Represents the node injected current vector, as the power injection or outflow of the node; the large imbalance refers to the source-load imbalance in the community being significantly higher than the source-load balance level of the community under normal operating conditions; Maintenance team assignment: Based on the weight matrix of a dynamic weighted undirected graph, high-weight faulty lines severed on community boundaries are identified. Combined with the geographic center coordinates of the community, the optimal path for the maintenance team is planned, prioritizing the repair of the connecting trunk lines located at the community boundaries that connect two high-resilience communities. Here, a high-resilience community refers to a community with high internal stability and good operating status in the Pareto optimal partitioning scheme. The connecting trunk line refers to a physical connection line located between different communities, with the corresponding edge weight being at a high level among the community boundary lines.

8. A system for resilient zoning and fault recovery in a power distribution cooperative network, characterized in that, The method for resilient zoning and fault recovery of distribution cooperative networks according to any one of claims 1-7 includes: The preprocessing unit is configured to abstract the power distribution coordination network at time t into a dynamic weighted undirected graph based on the node admittance matrix; The initialization unit is configured to: based on the dynamic weighted undirected graph, adopt a position-based node encoding strategy to transform the partitioning problem of the power distribution cooperative network into a chromosome encoding problem, and generate an initial population representing different partitioning schemes; The screening unit is configured to: construct a multi-objective fitness function with the optimization objectives of maximizing the internal stability of the power supply island and minimizing the partitioning adjustment cost, and use the multi-objective non-dominated sorting genetic algorithm NSGA-II to iteratively optimize the initial population representing different partitioning schemes to screen out the Pareto optimal set of partitioning schemes. The output unit is configured to: select a compromise solution from the Pareto optimal partitioning scheme set, generate a dynamic resilience partitioning map, and formulate switching operation instructions and resource scheduling schemes based on the map.

9. An electronic device, characterized in that, The electronic device includes: At least one processor; and, The memory is communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the method of distribution cooperative network resilience partitioning and fault recovery as described in any one of claims 1 to 7.

10. A non-transitory computer-readable storage medium, characterized in that, The non-transitory computer-readable storage medium stores computer instructions for causing the computer to perform the method of distribution cooperative network resilience partitioning and fault recovery as described in any one of claims 1 to 7.

Images