Active power distribution network balance area division method, device and medium

Abstract

The present application relates to a kind of active power distribution network balance zoning method, equipment and medium, comprising: S1, the load of the power distribution network to be studied and the annual daily power data of distributed power are collected;S2, the typical output scene of load and distributed power is extracted from the collected data using improved WGAN algorithm based on scaling dot product attention mechanism;S3, the electrical distance between each feeder block in the power distribution network to be studied is collected;S4, according to the data obtained by S2 and S3, the supply-demand imbalance degree, load similarity and distance module degree of each feeder block are calculated, and the fitness function is formed;S5, based on the fitness function, the power distribution network zoning scheme is determined using genetic algorithm.Compared with the prior art, the present application is based on electrical characteristics, source and load characteristics and electricity behavior factors to divide the active power distribution network balance zone, which can effectively exert the self-balancing ability of the power distribution network, significantly reduce the installation and maintenance cost of the active power distribution network, improve the distributed power consumption rate and the source and load balance level of the grid power distribution network.

Description

Technical Field

[0001] This invention relates to the field of distribution network planning, and in particular to a method, equipment and medium for dividing the balance zone of an active distribution network. Background Technology

[0002] With the large-scale integration of distributed generation and flexible loads, distribution networks are facing unprecedented challenges. In the ever-changing electricity market environment, the characteristics of power sources and loads in different regions are gradually showing significant differentiation, leading to increasingly prominent power balance issues. This phenomenon not only affects the stability of the distribution network but also places higher demands on the economic efficiency of the electricity market. Therefore, how to achieve more economical and wider-ranging power balance in complex power systems has become an important issue that urgently needs to be addressed.

[0003] To achieve an effective balance between power generation and consumption, it is essential to start with the zoning of the distribution network. Current distribution network structures are mostly based on traditional centralized power generation models. However, traditional planning methods are no longer sufficient to meet the demands of emerging distributed power sources.

[0004] Therefore, before conducting joint planning for proactive objects, it is necessary to optimize the existing distribution network to adapt to the dynamic changes in source and load characteristics. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings of the prior art by providing a method, equipment and medium for dividing the balance zone of an active distribution network, which can effectively give full play to the self-balancing capability of the distribution network, significantly reduce the installation and maintenance cost of the active distribution network, and improve the absorption rate of distributed power sources and the source-load balance level of the grid distribution network.

[0006] The objective of this invention can be achieved through the following technical solutions:

[0007] According to a first aspect of the present invention, a method for dividing the balance zone of an active distribution network is provided, comprising:

[0008] S1. Collect daily power data of the distribution network load and distributed power sources throughout the year;

[0009] S2. An improved WGAN algorithm based on scaling dot product attention mechanism is used to extract load and typical output scenarios of distributed power sources from the collected data obtained in S1.

[0010] S3. Collect the electrical distances between feeder blocks in the distribution network under study;

[0011] S4. Based on the data obtained from S2 and S3, calculate the supply-demand imbalance, load similarity, and distance modularity of each feeder block to form a fitness function.

[0012] S5. Based on the fitness function, a genetic algorithm is used to determine the distribution network balance zone division scheme.

[0013] Preferably, in step S1, the daily power data of the distribution network load and distributed power sources under study throughout the year are collected. Specifically, the data is collected every fifteen minutes and then summarized to obtain the daily power data of the distribution network load and distributed power sources under study throughout the year.

[0014] Preferably, step S2 uses an improved WGAN algorithm to extract typical output scenarios of loads and distributed power sources from the collected data obtained in S1, specifically including the following sub-steps:

[0015] S21. Use the improved WGAN algorithm based on the scaling dot product attention mechanism to generate a scene from the collected data obtained in S1.

[0016] S22. Reduce the scene generated in S21 with the goal of retaining the maximum amount of information;

[0017] S23. Using the probabilities of each scenario after reduction as weights, the typical output scenarios of load and distributed power sources are obtained by weighting.

[0018] Preferably, step S21 utilizes an improved WGAN algorithm to generate a scene from the collected data obtained in step S1, specifically including:

[0019] The Wasserstein distance calculation formula and objective function are as follows:

[0020]

[0021] In the formula: V(D,G) is the cross-entropy function for binary classification; E is the expected value of the distribution; G(z) and p G (z) represents the generated sample and its probability distribution; x and p data (x) represents the true sample and its probability distribution; W(·) is the Wasserstein distance between G(z) and x; Ω(·) is the distance between p and x. G (z) and p data The joint distribution set of (x);

[0022] Weights are assigned to the matching degree of the input and output power of the generator and discriminator:

[0023]

[0024] In the formula: x in d is the input matrix; K is the key-value matrix used for learning; K The dimension is K; the softmax(·) function is used to normalize the weights.

[0025] Preferably, in step S22, a scene reduction model is used to reduce the scene generated in step S21 with the goal of retaining the maximum amount of information. The mathematical expression of the scene reduction model is:

[0026]

[0027] P peak =|1-ξ max,j / ξ max,i |

[0028] P trou =|1-ξ min,j / ξ min,i |

[0029]

[0030] In the formula: I R (ξ i ,ξ i ) represents the amount of information about scene j in the typical scene set Z' relative to scene i in the generated scene set Z; N Z' and N Z The number of scenes for Z' and Z are respectively; p i The probability of scenario i occurring is given by C, and their sum is 1. sim P represents the probability similarity between scenes i and j; peak P trou and E tot These represent the peak and valley power information and total power information for scenarios i and j, respectively; ξ i,t Let ξ be the power of scene i at time t; max,i ξ min,i These are the daily peak and valley values ​​for scenario i, respectively.

[0031] Preferably, in step S4, based on the data obtained from S2 and S3, the supply-demand imbalance, load similarity, and distance modularity of each block are calculated to form a fitness function, specifically including:

[0032] Calculate the supply-demand imbalance:

[0033]

[0034] In the formula: and The power values ​​at time t are divided into feeder block i load and typical DG scenario; ε is a sufficiently small positive number;

[0035] The expression for calculating load similarity is:

[0036]

[0037]

[0038] In the formula: P i,t With P j,t t represents the combined load of feeder blocks i and j at time t within the peak-valley overlap period of a day; n represents the number of times the peak and valley overlap in the daily combined load curve; T represents the total time of the daily load curve, which is taken as 24 in this paper; t ij The peak and valley overlap times for blocks i and j within a day;

[0039] Calculate the distance modularity:

[0040]

[0041]

[0042] In the formula: b is the sum of the weights of all edges in the grid; N m e represents the number of blocks in the m-th balanced region; ij k represents the weight of the edge connecting the center points of feeder blocks i and j; i L is the sum of the edge weights connected to feeder block i; ij Let be the electrical distance between feeder blocks i and j.

[0043] Preferably, the fitness function is expressed mathematically as follows:

[0044]

[0045]

[0046]

[0047]

[0048]

[0049] Where: F is the fitness function; F1 is the total aggregation degree in the equilibrium region; F2 is the total coupling degree in the equilibrium region; f m f represents the correlation degree of the m-th equilibrium region after combination; max M is the sum of the correlation between each block before combination; M is the number of balanced blocks; N is the total number of blocks.

[0050] Preferably, step S5, which uses a genetic algorithm based on a fitness function to determine the distribution network balance zone division scheme, specifically includes the following sub-steps:

[0051] S51. Based on the characteristics of electrical connection, the chromosomes are encoded using a network adjacency matrix to generate multiple individuals and form an initial population;

[0052] S52. Calculate the fitness of each individual;

[0053] S53. Calculate the sum of the fitness values ​​of all individuals, and generate a random number based on the ratio of fitness to total fitness, and determine the selected individual using the random number;

[0054] S54. Pair individuals from the first half and the second half of the population to form a crossover parent generation;

[0055] S55. Randomly select certain locations for mutation of each individual, and check whether it meets the connectivity constraints and the N-1 line constraints after mutation. If it does not meet the constraints, trace the relevant blocks and re-examine the possible connection situations.

[0056] S56. Calculate the new fitness value of each individual after mutation and sort them according to fitness value;

[0057] S57. Determine whether the set maximum number of iterations has been reached or the convergence condition is met: if so, sort the individuals in the best set of individuals according to their fitness values ​​and output the optimal partition result; otherwise, return to S52 and continue with the selection, crossover and mutation operations.

[0058] According to a second aspect of the present invention, an electronic device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the program to implement any of the methods described above.

[0059] According to a third aspect of the invention, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements any of the methods described herein.

[0060] Compared with the prior art, the present invention has the following beneficial effects:

[0061] This invention proposes an active distribution network balancing zone division method that considers factors such as electrical characteristics, source-load characteristics, and electricity consumption behavior within the zone. This method can effectively leverage the self-balancing capability of the distribution network, meet the zoning principle of "high cohesion and low coupling," significantly reduce the installation and maintenance costs of active devices, and improve the distributed power absorption rate and the source-load balance level of the grid distribution network. Attached Figure Description

[0062] Figure 1 This is a flowchart of the method of the present invention;

[0063] Figure 2 Here is a flowchart of the genetic algorithm;

[0064] Figure 3 This is a schematic diagram of the power supply grid in the embodiment;

[0065] Figure 4 These are the planning schemes for each scenario in the embodiments;

[0066] Figure 5 The load characteristics of feeder block 1 in the embodiment;

[0067] Figure 6 The time-of-use electricity prices for each scenario in the example;

[0068] Figure 7 The figures show the operating curves of the equilibrium zones for Case 2 and Case 3 in the examples. Detailed Implementation

[0069] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0070] Example

[0071] like Figure 1 As shown in the figure, this embodiment provides a method for dividing the balance zone of an active distribution network, which includes the following steps:

[0072] S1. Collect the daily power curves of the loads at each node of the distribution network under study and the distributed generation (DG) throughout the year.

[0073] In this embodiment, the active power of a point is collected every 15 minutes within a 24-hour period.

[0074] The basic configuration of the power grid in this embodiment is as follows: There are 34 feeder blocks as basic planning units, including six types of loads: residential, industrial, commercial, administrative, educational, and medical. The maximum load capacity reduction is 20% of the node load, and administrative, medical, and educational loads do not participate in load decommissioning (DR). It is assumed that the DG output characteristics of each block within the same grid are the same, differing only in installed capacity, and that the grid is only constructed at industrial, commercial, and residential load nodes. Specific block divisions and source-load characteristics are as follows: Figure 3 As shown.

[0075] S2. An improved WGAN algorithm based on a scaled dot product attention mechanism is used to extract typical output scenarios of loads and distributed power sources from the collected data obtained in S1. The WGAN algorithm is a generative adversarial network based on Wasserstein distance, specifically including:

[0076] S21. Use the improved WGAN algorithm to generate a scene from the data obtained in S1:

[0077] The Wasserstein distance calculation formula and objective function are as follows:

[0078]

[0079] In the formula: V(D,G) is the cross-entropy function for binary classification; E is the expected value of the distribution; G(z) and p G (z) represents the generated sample and its probability distribution; x and p data (x) represents the true sample and its probability distribution; W(·) is the Wasserstein distance between G(z) and x; Ω(·) is the distance between p and x. G (z) and p data The joint distribution set of (x);

[0080] Weights are assigned to the matching degree of the input and output power of the generator and the discriminator, and the weight formula is as follows:

[0081]

[0082] In the formula: x in d is the input matrix; K is the key-value matrix used for learning; K The dimension is K; the softmax(·) function is used to normalize the weights.

[0083] S22. Reduce the generated scene to retain the maximum amount of information. The scene reduction model is shown below:

[0084]

[0085]

[0086]

[0087] P trou =|1-ξ min,j / ξ min,i |

[0088]

[0089] In the formula: I R (ξ i ,ξ i ) represents the amount of information about scene j in the typical scene set Z' relative to scene i in the generated scene set Z; N Z' and N Z The number of scenes for Z' and Z are respectively; p i The probability of scenario i occurring is given by C, and their sum is 1. sim P represents the probability similarity between scenes i and j; peak P trou and E tot These represent the peak and valley power information and total power information for scenarios i and j, respectively; ξ i,t Let ξ be the power of scene i at time t; max,i ξmin,i These are the daily peak and valley values ​​for scenario i, respectively.

[0090] S23. Using the probabilities of each scenario after reduction as weights, the typical output scenarios of load and DG are obtained by weighting.

[0091] S3. Collect electrical distance data between each feeder block.

[0092] S4. Using the load and typical DG scenario curve data obtained in S2, calculate the supply-demand imbalance and load similarity. The formulas for calculating the imbalance and load similarity are as follows:

[0093]

[0094]

[0095]

[0096] In the formula: and The power values ​​at time t are divided into feeder block i load and typical DG scenario; ε is a sufficiently small positive number to avoid a denominator of 0; P i,t With P j,t t represents the combined load of feeder blocks i and j at time t within the peak-valley overlap period of a day; n represents the number of times the peak and valley overlap in the daily combined load curve; T represents the total time of the daily load curve, which is taken as 24 in this paper; t ij The peak and valley overlap times for blocks i and j within a day.

[0097] The distance modularity is calculated using the electrical distance data between each feeder block collected by S3. The calculation formula is as follows:

[0098]

[0099]

[0100] In the formula: b is the sum of the weights of all edges in the grid; N m e represents the number of blocks in the m-th balanced region; ij k represents the weight of the edge connecting the center points of feeder blocks i and j; i L is the sum of the edge weights connected to feeder block i; ij Let be the electrical distance between feeder blocks i and j.

[0101] Forming the fitness function:

[0102]

[0103]

[0104]

[0105]

[0106]

[0107] Where: F is the fitness function; F1 is the total aggregation degree in the equilibrium region; F2 is the total coupling degree in the equilibrium region; f m f represents the correlation degree of the m-th equilibrium region after combination. max M is the sum of the correlation between each block before combination; M is the number of balanced blocks; N is the total number of blocks.

[0108] S5. Based on the obtained fitness function, use a genetic algorithm to determine the equilibrium region partitioning scheme. The flowchart of the genetic algorithm is as follows: Figure 2 As shown, the specific steps are described below:

[0109] S51. Based on the characteristics of electrical connections, a network adjacency matrix is ​​used to encode chromosomes, where 0 represents no connection and 1 represents a connection. Multiple individuals are generated to form an initial population.

[0110] S52. Calculate the fitness of each individual.

[0111] S53. Calculate the sum of the fitness values ​​of all individuals, and generate a random number based on the ratio of fitness to total fitness. Use this random number to determine the selected individual.

[0112] S54. Pair individuals from the first half and the second half of the population to form a crossover parent generation.

[0113] S55. Randomly select certain locations for mutation of each individual, and check whether it meets the connectivity constraints and the N-1 line constraints after mutation. If it does not meet the constraints, trace the relevant blocks and re-examine the possible connection situations.

[0114] S56. Calculate the new fitness value of each individual after mutation and sort them according to fitness value.

[0115] S57. Determine whether the set maximum number of iterations has been reached or the convergence condition has been met. If yes, sort the individuals in the best set of individuals according to their fitness values ​​and output the optimal partition result; otherwise, return to S52 and continue with the selection, crossover, and mutation operations.

[0116] To verify the superiority of the proposed method, simulation examples were conducted in the following three scenarios.

[0117] Table 1 shows the three scenarios to be simulated.

[0118]

[0119] To further demonstrate the impact of each scenario on the active object configuration scheme, energy storage was configured for each scenario with the goal of optimal economic efficiency. Following the steps above, the corresponding planning schemes for each scenario were obtained as follows: Figure 4 As shown in Table 2, the time-of-use pricing strategy is as follows: Figure 6 As shown. Among them, Figure 4 Parts (a), (b), and (c) correspond to Case 1, Case 2, and Case 3, respectively.

[0120] Table 2 Energy Storage Planning Schemes for Various Scenarios

[0121]

[0122]

[0123] Depend on Figure 4 As shown in (a), considering only electrical characteristics for balancing zone division, the result is an independent feeder block 1, within which only residential and industrial loads exist. Further extraction of the load characteristics of this feeder block is as follows... Figure 5 As can be seen, the area is dominated by industrial loads and has a single form, making it impossible to use the temporal complementarity between loads to mitigate peak-valley differences. The overall peak-valley difference rate of the block reaches 60.87%, which is not conducive to the stable operation of the system.

[0124] The planning results for each scenario are shown in Table 3 below.

[0125] Table 3 Planning Results for Each Scenario

[0126] Scene <![CDATA[C ESS (10,000 yuan / day) <![CDATA[I sell( (10,000 yuan / day) Average wind and solar curtailment rate (%) Case 1 10.31 79.46 11.26％ Case 2 10.55 77.36 13.72％ Case 3 1.95 85.23 4.71％

[0127] Comparing Case 1 and Case 3, it can be seen that Case 1, because it only uses active objects to balance the source and load of the grid distribution network, in order to achieve a better source and load balancing effect, firstly, it increases the peak-valley price difference rate of time-of-use electricity pricing, that is, the price difference rate between residential electricity price and industrial electricity price is 3.56 and 4.18 respectively, which is higher than Case 3's 2.55 and 3.42, in order to motivate users to participate in DR; secondly, it increases the configuration scale of ESS, C ESS It reached 103,100 yuan per day. Although this strategy successfully reduced I sell However, Case 1's average daily total cost was 897,700 yuan, higher than Case 3's 871,800 yuan, and its average wind and solar curtailment rate was 11.26%, higher than Case 3's 4.71%. This indicates that relying solely on proactive objects for source-load balancing in the grid distribution network, without utilizing the inherent source-load balancing capabilities of the blocks, leads to increased proactive object operation costs while still maintaining insufficient regulation capacity. Furthermore, the significant impact of time-of-use pricing on user electricity consumption behavior within the region may result in decreased user satisfaction with electricity usage.

[0128] Comparing Case 2 and Case 3, it's clear that while Case 2 initially adopted a balanced distribution approach for power sharing within zones, it still required a high electricity price difference and a large ESS capacity because the zoning criteria only considered the electrical characteristics of the blocks. This was especially true for ESS configuration; Case 2's C... ESS The cost is 105,500 yuan / day, significantly higher than the 19,500 yuan / day in Case 3. Although Case 3, due to its smaller ESS scale and more moderate DR strategy, resulted in a decrease in the active regulation capability and DR response, leading to a higher electricity purchase cost for users compared to Case 2, its total cost is still 7,300 yuan / day lower than Case 2. This indicates that fully considering the supply-demand imbalance and load similarity characteristics during the balance zone division process, and utilizing power mutual assistance within the zone, can effectively reduce the action costs and regulation pressure of active entities. This achieves better wind and solar power absorption rates while minimizing changes to user electricity consumption behavior, thereby improving the source-load balance level of the grid distribution network.

[0129] To further verify the superiority of the proposed method, the operating status of each equilibrium zone in Case 2 and Case 3 is analyzed as shown in Table 4 and... Figure 7 As shown, where Figure 7 (a) to (h) correspond to equilibrium zones 1 to 5 in Case 2 and equilibrium zones 1 to 3 in Case 3, respectively, in the table below.

[0130] Table 4. Operation status of each equilibrium zone in Case 2 and Case 3

[0131]

[0132] Since Case 2 did not consider the source load characteristics of the blocks during the partitioning process, the source load characteristics of each region still show a divergent trend: balanced region 1 and balanced region 3 exhibit power supply characteristics, especially balanced region 1. The value was -3.52, and there was significant wind and solar power curtailment between 12:00 and 17:00; Balance Zones 2 and 4 exhibited load characteristics, while only Balance Zone 5 showed a relatively good source-load balance. It is -1.01. Combined with... Figure 7As can be seen, the main charging period of the ESS is 1:00-10:00, and the discharging period is 12:00-21:00. This means the ESS is mainly used to store excess electricity in balance zones 1 and 3 during off-peak hours and distribute it during peak hours in balance zones 2 and 4 to save electricity purchase costs for users within those zones. It is evident that the method of dividing balance zones based solely on electrical characteristics violates the original intention of setting up balance zones: to utilize the zone's regulation capacity to balance source and load, thereby promoting distributed generation (DG) consumption and reducing the cost of active actions. It fails to change the current source-load differentiation within the grid distribution network. A comparison of Case 1 and Case 2 in Table 3 further verifies this conclusion. Compared to Case 1, Case 2, which pre-divided the balance zones, has a lower C... ESS The average curtailment rate of wind and solar power was actually higher than that of Case 1, meaning that the zoning method not only failed to promote regional source-load balance, but also had the opposite effect.

[0133] Compared to Case 2, the equilibrium regions 2 and 3 obtained in Case 3 are... The values ​​are 1.13 and 1.15 respectively, and the autonomous operation and balance of source loads within the area can be basically achieved through DR alone, resulting in wind and solar curtailment rates of 0.89% and 0.29% respectively; the balance zone 1 The value was -1.16. Due to the severe curtailment of wind and solar power from 13:00 to 18:00 and the high electricity price, the DR regulation was ineffective. Therefore, the ESS was configured to charge during the surplus DG output periods of 1:00-3:00 and 13:00-17:00, and discharge during the insufficient DG output periods of 6:00-8:00 and 21:00-22, thereby further regulating the load curve in the area and reducing the curtailment rate of wind and solar power in this area from the original 20.81% to 12.94%.

[0134] In summary, by applying the active distribution network balance zone division method based on source load and electrical characteristics proposed in this invention, the self-balancing capability of the distribution network can be effectively utilized, the zoning principle of "high cohesion and low coupling" can be met, the installation and maintenance costs of active objects can be significantly reduced, the DG absorption rate can be improved, and the source load balance level of the grid distribution network can be improved.

[0135] The electronic device of this invention includes a central processing unit (CPU), which can perform various appropriate actions and processes according to computer program instructions stored in read-only memory (ROM) or loaded from a storage unit into random access memory (RAM). The RAM may also store various programs and data required for device operation. The CPU, ROM, and RAM are interconnected via a bus. Input / output (I / O) interfaces are also connected to the bus.

[0136] Multiple components in the device are connected to the I / O interface, including: input units such as keyboards and mice; output units such as various types of displays and speakers; storage units such as disks and optical discs; and communication units such as network interface cards (NICs), modems, and wireless transceivers. The communication unit allows the device to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0137] The processing unit executes the various methods and processes described above, such as methods S1 to S5. For example, in some embodiments, methods S1 to S5 may be implemented as computer software programs tangibly contained in a machine-readable medium, such as a storage unit. In some embodiments, part or all of the computer program may be loaded and / or installed on the device via ROM and / or a communication unit. When the computer program is loaded into RAM and executed by the CPU, one or more steps of methods S1 to S5 described above may be performed. Alternatively, in other embodiments, the CPU may be configured to execute methods S1 to S5 by any other suitable means (e.g., by means of firmware).

[0138] The functions described above in this document can be performed at least in part by one or more hardware logic components. For example, exemplary types of hardware logic components that can be used, without limitation, include: field programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), payload programmable logic devices (CPLDs), and so on.

[0139] The program code used to implement the methods of the present invention can be written in any combination of one or more programming languages. This program code can be provided to a processor or controller of a general-purpose computer, special-purpose computer, or other programmable data processing device, such that when executed by the processor or controller, the program code causes the functions / operations specified in the flowcharts and / or block diagrams to be implemented. The program code can be executed entirely on the machine, partially on the machine, as a standalone software package partially on the machine and partially on a remote machine, or entirely on a remote machine or server.

[0140] In the context of this invention, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. Machine-readable media can include, but are not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0141] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in the present invention, and these modifications or substitutions should all be covered within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for dividing the balance zone of an active distribution network, characterized in that, include: S1. Collect daily power data of the distribution network load and distributed power sources throughout the year; S2. The improved WGAN algorithm based on the scaling dot product attention mechanism is used to extract the load and typical output scenarios of distributed power sources from the collected data obtained in S1. This includes the following sub-steps: S21. Use the improved WGAN algorithm based on the scaling dot product attention mechanism to generate a scene from the collected data obtained in S1. S22. Use a scene reduction model to reduce the scene generated in S21 with the goal of retaining the maximum amount of information. The mathematical expression for the scene reduction model is: , , , , , In the formula: Typical scenario set Z’ Mid-scene j For generating scene sets Z Mid-scene i The amount of information; and They are respectively Z’ and Z The number of scenes; For the scene to appear i The probabilities of these probabilities sum to 1; For the scene i and j The probability similarity; , and Scenes i and j The daily peak and valley power information and the total power information; for t Moment Scene i The power; , Scenes i Daily peak and trough values; S23. Using the probabilities of each scenario after reduction as weights, the typical output scenarios of load and distributed power sources are obtained by weighting. S3. Collect the electrical distances between feeder blocks in the distribution network under study; S4. Based on the data obtained from S2 and S3, calculate the supply-demand imbalance, load similarity, and distance modularity of each feeder block to form a fitness function. S5. Based on the fitness function, a genetic algorithm is used to determine the distribution network balance zone division scheme.

2. The method according to claim 1, characterized in that, In step S1, the daily power data of the distribution network load and distributed power sources under study throughout the year are collected. Specifically, the data is collected every fifteen minutes and then summarized to obtain the daily power data of the distribution network load and distributed power sources under study throughout the year.

3. The method according to claim 1, characterized in that, In step S21, the improved WGAN algorithm is used to generate a scene from the collected data obtained in step S1, specifically including: The Wasserstein distance calculation formula and objective function are as follows: , , In the formula: V ( D , G () is the cross-entropy function for binary classification; E The expected value of the distribution; G ( z )and p G ( z ) represents the generated sample and its probability distribution; x and p data ( x () represents the real samples and their probability distribution; for G ( z )and x Wasserstein distance; for p G ( z )and p data ( x The joint distribution set of ). Weights are assigned to the matching degree of the input and output power of the generator and discriminator: , In the formula: x in The input matrix; K A key-value matrix for learning purposes; d K for K The dimension; The function is used to normalize the weights.

4. The method according to claim 1, characterized in that, In step S4, based on the data obtained from S2 and S3, the supply-demand imbalance, load similarity, and distance modularity of each block are calculated to form a fitness function, specifically including: Calculate the supply-demand imbalance: , In the formula: and Divided into feeder blocks i Load and typical DG scenarios in t The power value at any given time; It is a sufficiently small positive number; The expression for calculating load similarity is: , , In the formula: P i,t and P j,t feeder blocks i and j During the period when peaks and troughs overlap in a day t The overall load at any given moment; n This refers to the number of times when the peak and trough of the daily composite load curve coincide. T The total time for the daily load curve is 24, which is taken as the time in this paper; t ij For blocks i and j The time of overlap between peaks and troughs within a day; Calculate the distance modularity: , , In the formula: b It is the sum of the weights of all edges in the grid; N m For the first m The number of blocks in each balanced region; e ij For connecting feeder blocks i , j The weight of the edge at the center point; k i For the feeder block i The sum of the weights of all connected edges; L ij For feeder block i , j The electrical distance between them.

5. The method according to claim 4, characterized in that, The fitness function is expressed mathematically as follows: , , , , , In the formula: F The fitness function; F 1 represents the total degree of aggregation in the equilibrium region; F 2 represents the total coupling degree in the equilibrium region; f m For the combined first m The degree of correlation of each equilibrium region; f max This is the sum of the correlation between each block before combination; M The number of equilibrium zones; N This represents the total number of blocks.

6. The method according to claim 1, characterized in that, S5, based on the fitness function, uses a genetic algorithm to determine the distribution network balance zone division scheme, specifically including the following sub-steps: S51. Based on the characteristics of electrical connection, the chromosomes are encoded using a network adjacency matrix to generate multiple individuals and form an initial population; S52. Calculate the fitness of each individual; S53. Calculate the sum of the fitness values ​​of all individuals, and generate a random number based on the ratio of fitness to total fitness, and determine the selected individual using the random number; S54. Pair individuals from the first half and the second half of the population to form a crossover parent generation; S55. Randomly select certain locations for mutation of each individual, and check whether it meets the connectivity constraints and the N-1 line constraints after mutation. If it does not meet the constraints, trace the relevant blocks and re-examine the connection status. S56. Calculate the new fitness value of each individual after mutation and sort them according to fitness value; S57. Determine whether the set maximum number of iterations has been reached or the convergence condition is met: if so, sort the individuals in the best set of individuals according to their fitness values ​​and output the optimal partition result; otherwise, return to S52 and continue with the selection, crossover and mutation operations.

7. An electronic device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the program, it implements the method as described in any one of claims 1 to 6.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the method as described in any one of claims 1 to 6.

Images