Power supply cluster division method and system

Abstract

The present invention belongs to the field of power supply cluster management, and discloses a power supply cluster division method and system, the method comprising: S1, obtaining a distribution map containing all nodes to be divided; S2, determining the number of clusters based on the distribution map; S3, determining the function to be optimized; S4, using a particle swarm algorithm to obtain a first optimization result; S5, changing the initial value of the particle of the particle swarm algorithm to obtain a second optimization result; S6, calculating the gap coefficient between the first optimization result and the second optimization result; S7, judging whether the gap coefficient is greater than a set gap coefficient threshold, if so, entering S8, if not, entering S9; S8, using a genetic algorithm to obtain the final optimization result, entering S10; S9, selecting the party with a smaller fitness function value as the final optimization result, entering S10; S10, outputting the power supply cluster division result based on the final optimization result. The present invention improves the effectiveness of the number of clusters obtained, so that the cluster can be closer to the optimal cluster result.

Description

Technical Field

[0001] The present invention relates to the field of power supply cluster management, and in particular to a power supply cluster division method and system. Background Art

[0002] Power supply cluster division means that in the power system, the entire power area is divided into several relatively independent areas (clusters) according to factors such as geographical location, power load, and power demand. Each cluster contains multiple nodes, which are divided into power supply nodes and power consumption nodes. In the event of an emergency (such as natural disasters, equipment failures, etc.), the faulty area can be quickly isolated to ensure normal power supply to other areas, thereby ensuring the stability and reliability of power supply.

[0003] In the process of dividing power supply clusters in the prior art, the number of clusters is usually specified based on the industry experience of relevant personnel, which easily leads to an inaccurate number of clusters, resulting in an inaccurate final clustering result, which is too different from the optimal clustering result and makes it impossible to obtain a reasonable clustering result. Summary of the invention

[0004] The purpose of the present invention is to disclose a power supply cluster division method and system to solve the technical problems raised in the background technology.

[0005] In order to achieve the above object, the present invention adopts the following technical solution:

[0006] In one aspect, the present invention provides a method for dividing a power supply cluster, comprising:

[0007] S1, obtain a distribution map containing all nodes to be divided;

[0008] S2, determine the number of clusters based on the distribution map;

[0009] S3, determine the function to be optimized;

[0010] S4, using a particle swarm algorithm to optimize the function to be optimized, and obtaining a first optimization result;

[0011] S5, changing the initial value of the particle of the particle swarm algorithm, and optimizing the function to be optimized by using the particle swarm algorithm again to obtain a second optimization result;

[0012] S6, calculating the difference coefficient between the first optimization result and the second optimization result;

[0013] S7, determining whether the gap coefficient is greater than a set gap coefficient threshold, if so, proceeding to S8, if not, proceeding to S9;

[0014] S8, using a genetic algorithm to optimize the function to be optimized, obtaining the final optimization result, and entering S10;

[0015] S9, from the first optimization result and the second optimization result, select the one with the smaller fitness function value as the final optimization result, and enter S10;

[0016] S10: outputting a power supply cluster division result based on the final optimization result.

[0017] On the other hand, the present invention provides a power supply cluster division system, including an acquisition module, a first determination module, a second determination module, a first optimization module, a second optimization module, a calculation module, a judgment module, a third optimization module, a selection module and an output module;

[0018] The acquisition module is used to obtain a distribution map containing all nodes to be divided;

[0019] The first determination module is used to determine the number of clusters based on the distribution map;

[0020] The second determination module is used to determine the function to be optimized;

[0021] The first optimization module is used to optimize the function to be optimized using a particle swarm algorithm to obtain a first optimization result;

[0022] The second optimization module is used to change the initial value of the particles of the particle swarm algorithm, and use the particle swarm algorithm again to optimize the function to be optimized to obtain a second optimization result;

[0023] The calculation module is used to calculate the gap coefficient between the first optimization result and the second optimization result;

[0024] The judgment module is used to judge whether the gap coefficient is greater than a set gap coefficient threshold;

[0025] The third optimization module is used to optimize the optimization function using a genetic algorithm to obtain a final optimization result when the gap coefficient is greater than a set gap coefficient threshold;

[0026] The selection module is used for selecting the one with a smaller fitness function value from the first optimization result and the second optimization result as the final optimization result when the gap coefficient is less than or equal to the set gap coefficient threshold;

[0027] The output module is used to output the power supply cluster division result based on the final optimization result selected by the selection module when the gap coefficient is less than or equal to the set gap coefficient threshold;

[0028] And it is used to output the power supply cluster division result based on the final optimization result obtained by the third optimization module when the gap coefficient is greater than the set gap coefficient threshold.

[0029] Beneficial effects:

[0030] Compared with the prior art, the present invention does not specify the number of clusters based on the industry experience of personnel when determining the number of clusters, because different personnel have different industry experiences, which makes the number of clusters have a high degree of uncertainty. The present invention determines the number of clusters through the distribution map of nodes, which not only allows the size of the area where the cluster is located to be controlled according to actual needs, but also allows the cluster scale to be controlled, effectively improving the stability of the number of clusters obtained. Moreover, the number of clusters can change adaptively with the changes in the actual distribution of nodes, thereby improving the effectiveness of the number of clusters obtained, so that the cluster can be closer to the optimal cluster result under the given cluster area size and cluster scale constraints, and obtain a more reasonable cluster result. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments are briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without creative work.

[0032] Figure 1 A schematic diagram of a power supply cluster division method of the present invention.

[0033] Figure 2 A schematic diagram of a power supply cluster division system of the present invention. DETAILED DESCRIPTION

[0034] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. The components of the embodiments of the present invention generally described and shown in the drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the drawings is not intended to limit the scope of the claimed invention, but merely represents the selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without making creative work belong to the scope of protection of the present invention.

[0035] like Figure 1 In one embodiment shown, the present invention provides a power supply cluster division method, comprising:

[0036] S1, obtain a distribution map containing all nodes to be divided.

[0037] Specifically, the distribution map is a two-dimensional plane map, on which each node to be divided is marked.

[0038] Preferably, the nodes to be divided include power supply nodes and power consumption nodes.

[0039] Specifically, the power supply nodes include wind power stations, hydropower stations, photovoltaic power stations, etc.

[0040] Power-consuming nodes include residential buildings, schools, hospitals, etc.

[0041] S2, determine the number of clusters based on the distribution map.

[0042] The method used in this step to determine the number of clusters can perform a certain degree of adaptive control on the size of a single cluster. This is related to the accuracy of cluster control. The higher the accuracy, the smaller the size of a single cluster.

[0043] Preferably, determining the number of clusters based on the distribution map includes:

[0044] S21, establishing a rectangular coordinate system in the distribution map;

[0045] Specifically, a point can be randomly selected in the distribution map as the origin of the rectangular coordinate system, and an X-axis and a Y-axis passing through the origin and perpendicular to each other can be drawn;

[0046] S22, respectively determining the coordinates of each node to be divided in a rectangular coordinate system;

[0047] S23, respectively, with x min and x max Indicates the maximum and minimum values ​​of the coordinates of all nodes to be divided on the X-axis, respectively represented by y min and max Indicates the maximum and minimum values ​​of the coordinates of all nodes to be divided on the Y axis; defines a local range Q, and the coordinates x on the X axis and the coordinates y on the Y axis of the points in Q meet the following restrictions:

[0048]

[0049] S24, in a local range, determining the number of clusters by means of a circular calculation.

[0050] In the local range, a control parameter M of the size of the local area is set as required, and a suitable formula for controlling the loop calculation is set, so that the determined number of clusters can include the control information of the cluster size.

[0051] Preferably, in a local range, the number of clusters is determined by a cyclic calculation method, including:

[0052] The first round of calculation:

[0053] Divide Q into M×M local regions of the same size, and calculate the cycle coefficient of each local region obtained in this round of cycle calculation;

[0054] The local areas whose cyclic coefficients are greater than or equal to the set cyclic coefficient comparison value are saved to the set K1, and the local areas whose cyclic coefficients are less than the set cyclic coefficient comparison value are saved to the set FN;

[0055] The hth round of loop calculation, h is greater than or equal to 2:

[0056] K h-1 Each local area in is divided into M×M local areas of the same size again, and the circulation coefficient of each local area obtained by this round of circulation calculation is calculated respectively;

[0057] The local areas whose cyclic coefficients are greater than or equal to the set cyclic coefficient comparison value are saved to the set K h , save the local area whose cyclic coefficient is less than the set cyclic coefficient comparison value to the set FN;

[0058] Judge K h Is it an empty set? If so, filter FN to obtain the number of clusters. If not, perform the next round of calculation.

[0059] Specifically, in the cyclic calculation process of the present invention, except for the first cyclic calculation, the remaining cyclic calculations are further calculated based on the local area where the cyclic coefficient obtained in the previous round is greater than or equal to the set cyclic coefficient comparison value. In this way, when the size of the population is too large, further cyclic calculations can be performed to make the size of all populations meet the set requirements.

[0060] Fn is screened because in the process of the loop, there will be a situation where the area of ​​the local area is larger, but the number of nodes included is smaller. If the total number of local areas finally obtained is directly used as the total number of the population, it will obviously cause the average size of the final population to be smaller than the actual need.

[0061] Specifically, the value of M can be 2, 3, 4, 5, etc.

[0062] Preferably, the value of M is 2.

[0063] Preferably, the calculation formula of the circulation coefficient is:

[0064]

[0065] cycef b represents the circulation coefficient of the local area b, snode b represents the total number of nodes in the local region b, allnode represents the total number of nodes in Q, and sareab represents the area of ​​the local area b, allarea represents the area of ​​Q, w1 and w2 represent the weighted proportion of quantity and area respectively.

[0066] The circulation coefficient of the present invention is obtained by comprehensive calculation based on the total number of nodes and the area of ​​the local area. The more nodes are included in the local area and the larger the area of ​​the local area is, the larger the circulation coefficient is; the fewer nodes are included in the local area and the smaller the area of ​​the local area is, the smaller the circulation coefficient is. In this way, when determining the number of clusters, the total cluster size and the size of the local area can be controlled, which can avoid the cluster from being too large. Because when the cluster is too large, the power supply reliability will be reduced and the line loss will increase.

[0067] A cluster that is too large means that a single power supply unit covers more areas and users. Once a fault occurs somewhere, it may affect users in a wider range, resulting in reduced power supply reliability. Repairs and troubleshooting may be more complicated and time-consuming because a wider range of power supply problems need to be dealt with. Larger clusters usually mean longer transmission lines, which will lead to increased losses (line losses) in power transmission. The energy loss when current passes through the line will increase with the increase of transmission distance, resulting in a decrease in overall efficiency. Increased line losses not only affect energy utilization efficiency, but may also increase operating costs.

[0068] Preferably, the weighted ratio of quantity and the weighted ratio of area are respectively and

[0069] These two ratios can be adjusted according to actual needs. If a smaller cluster size is preferred, the weighted ratio of quantity can be set larger.

[0070] Specifically, the set cycle coefficient comparison value can be

[0071] This comparison value can further control the cluster size. The smaller the cluster size, the smaller the loop coefficient comparison value is set. The range of the loop coefficient comparison value is (0,1).

[0072] Preferably, the FN is screened to obtain the number of clusters, including:

[0073] The number of local areas in FN that contain nodes greater than FNth is taken as the number of clusters, and FNth represents the adaptive number threshold.

[0074] Specifically, by screening, local areas containing too few nodes can be excluded, thereby improving the accuracy of the number of clusters obtained.

[0075] Preferably, the calculation formula of the adaptive quantity threshold is:

[0076]

[0077] Avenum represents the average number of nodes contained in the local area in FN.

[0078] S3, determine the function to be optimized.

[0079] In this step, the function to be optimized is related to the actual consideration. For example, when the stability of the power supply within the cluster needs to be considered, the power supply of the power supply nodes and the power consumption of the power consumption nodes within the cluster need to be considered from the parameters. If the cluster cost needs to be controlled, the distance relationship between the nodes within the cluster needs to be considered.

[0080] Preferably, determining the function to be optimized includes:

[0081] Use the following function as the function to be optimized:

[0082]

[0083] cyceF represents the function to be optimized, optcef c represents the clustering coefficient of the nodes in the cth cluster;

[0084]

[0085] bua represents the set of all nodes in the cth cluster, and bub represents the cth cluster dist i,j Represents the distance between node i and node j, dms b represents the maximum distance between two nodes in bua, spu represents the set of power supply nodes in local area b, csu represents the set of power consumption nodes in local area b, enr u Indicates the amount of power that the power supply node u can provide within a set time length, csue v represents the power consumed by the power-consuming node v within the set time length, w3 and w4 represent the weighted proportion of distance and power respectively, sc represents the total number of all nodes in the cth cluster, and nclst represents the total number of clusters.

[0086] In the present invention, the function to be optimized is the sum of the cluster coefficients of each cluster, and the cluster coefficient is calculated from the distance between the nodes in the cluster and the difference between the power supply and power consumption in the cluster. The smaller the distance between the nodes in the cluster and the smaller the difference between the power supply and power consumption in the cluster, the smaller the cluster coefficient, indicating that the cluster division effect is better. Since the distance between the nodes is further considered when designing the function to be optimized, the correspondence with the determination stage of the number of clusters of the present invention is achieved, and the scale of the cluster can also be considered in the optimization stage to ensure the control of the scale of the cluster. In addition, since the relationship between power supply and power consumption in the cluster is taken into account, the divided cluster can have a stronger self-power supply capacity. When the power connection with the outside is cut off due to a disaster, the number of users affected in the cluster can be more effectively reduced without relying on external power supply.

[0087] Specifically, the set time length can be 24 hours or 12 hours.

[0088] When the set time length is 12 hours, the corresponding time interval may be [9:00, 21:00].

[0089] Specifically, the weighted ratio of distance and the weighted ratio of power can be respectively and

[0090] These two ratios can also be determined according to actual needs. When a smaller distance is preferred, the weighted ratio of the distance can be set larger.

[0091] S4, using a particle swarm algorithm to optimize the function to be optimized to obtain a first optimization result.

[0092] In this step, the present invention encodes each node by integer coding. The function to be optimized is used as the fitness function of the particle swarm algorithm for particle swarm optimization.

[0093] Preferably, the particle swarm algorithm is used to optimize the function to be optimized to obtain a first optimization result, including:

[0094] According to the set number of iterations H, the function to be optimized is used as the fitness function of the particle swarm algorithm, and the particle swarm algorithm is used to minimize the function to be optimized to obtain H optimization results;

[0095] Among the H optimization results, the optimization result with the smallest fitness function value is taken as the first optimization result.

[0096] Specifically, in the initialization phase, each cluster is encoded separately using integer coding, the value range of the coding is [1, NL], and the coding is an integer, and each node is randomly assigned to one of the NL clusters; NL is the total number of clusters;

[0097] The code of the cluster corresponding to each node is used as the initial position of the particle;

[0098] Randomly generate initial velocities for each particle in the value range (-NL, NL).

[0099] During the iterative optimization process, after each update of the particle position, the value corresponding to the position is rounded up so that each particle has a clear corresponding cluster.

[0100] The optimization result is the encoding of the cluster corresponding to each node.

[0101] Preferably, the number of iterations H is set to 100.

[0102] S5, changing the initial value of the particle of the particle swarm algorithm, and optimizing the function to be optimized by using the particle swarm algorithm again to obtain a second optimization result.

[0103] Specifically, change the initial value of the particle, including:

[0104] Change the initial position and initial velocity of the particle;

[0105] The initial position and initial velocity of each particle are processed as follows:

[0106] For particle e, the initial position of particle e in S4 is nl. After the change, the initial position of the particle is NL-nl;

[0107] The initial velocity of particle e in S4 is v. After the change, the initial velocity of the particle is vc.

[0108] Specifically, the present invention significantly changes the position of the initial value of the particle, so that when the optimization process of S4 falls into a local optimum, it can be detected by the subsequent S6, so that the genetic algorithm with better global optimization ability can be used to optimize the optimization function.

[0109] As for why the genetic algorithm is not directly selected, it is mainly because the genetic algorithm usually involves operations such as selection, crossover (mating), and mutation, and multiple individuals may need to be reorganized and mutated in each generation. These operations increase its computational complexity, and crossover and mutation operations introduce complexity, especially when dealing with complex coding (such as binary or symbolic coding). These operations will affect the overall speed of the algorithm. More iterations are required to converge to a better solution, because the randomness of its operation is higher, and it may rely more on the later evolutionary process to find a better solution. In other words, the optimization speed of the particle swarm algorithm is much faster than that of the genetic algorithm, and even if the particle swarm algorithm is used twice, it is much faster than running the genetic algorithm once. Therefore, the present invention first uses the particle swarm algorithm to find the best solution. When the result of the best solution meets the requirements, the party with the better value of the fitness function in the two best solutions is used as the final optimization result. When the optimization result does not meet the requirements, a genetic algorithm is used to solve it. Since the types of parameters involved in the present invention are relatively small, a better balance can be achieved in terms of optimization speed and optimization effect. It can reduce the probability of falling into the local optimum caused by only using the particle swarm algorithm, and can ensure the accuracy of the optimization result.

[0110] S6, calculating the difference coefficient between the first optimization result and the second optimization result.

[0111] Specifically, the gap coefficient is calculated by the following formula:

[0112] def=|cyceF1-cyceF2|

[0113] def represents the gap coefficient, cyceF1 and cyceF2 represent the function values ​​of the function to be optimized calculated based on the first optimization result and the second optimization result, respectively.

[0114] In addition, the gap coefficient can also be 1-siml;

[0115] siml represents the similarity between the first optimization result and the second optimization result.

[0116] The similarity can be calculated by using cosine similarity, Pearson correlation coefficient, etc.

[0117] S7, determine whether the gap coefficient is greater than the set gap coefficient threshold, if so, enter S8, if not, enter S9.

[0118] Specifically, the gap coefficient threshold is one tenth of the value of cyceF1.

[0119] When similarity is used for calculation, the difference coefficient threshold is the set threshold, for example 0.2.

[0120] S8, using a genetic algorithm to optimize the function to be optimized, obtaining the final optimization result, and entering S10.

[0121] Specifically, when the genetic algorithm is used to optimize the function to be optimized, the cluster to which the node belongs is encoded in a binary manner.

[0122] When initializing the population, each node is randomly assigned to a cluster.

[0123] The optimization result obtained is the encoding of the cluster corresponding to each node.

[0124] S9, from the first optimization result and the second optimization result, select the one with the smaller fitness function value as the final optimization result, and enter S10.

[0125] Since the function to be optimized in the present invention is to obtain a cluster with short distances between nodes and small differences in power supply and power consumption within the cluster, the smaller the fitness function value is, the better the cluster result obtained is.

[0126] S10: outputting a power supply cluster division result based on the final optimization result.

[0127] Specifically, the power supply cluster division result is a set of nodes included in each cluster.

[0128] In the process of determining the number of clusters, the above scheme does not specify based on the industry experience of personnel, because different personnel have different industry experiences, which makes the number of clusters have a high degree of uncertainty.

[0129] The above scheme determines the number of clusters through the node distribution map, which not only allows the size of the cluster area to be controlled according to actual needs, but also allows the cluster scale to be controlled, effectively improving the stability of the number of clusters obtained. Moreover, the number of clusters can change adaptively with the change of the actual node distribution, thereby improving the effectiveness of the number of clusters obtained, so that the cluster can be closer to the optimal cluster result under the given cluster area size and cluster scale constraints, and obtain a more reasonable cluster result.

[0130] like Figure 2 , the present invention also provides a power supply cluster division system, including an acquisition module, a first determination module, a second determination module, a first optimization module, a second optimization module, a calculation module, a judgment module, a third optimization module, a selection module and an output module;

[0131] The acquisition module is used to obtain a distribution map containing all nodes to be divided;

[0132] The first determination module is used to determine the number of clusters based on the distribution map;

[0133] The second determination module is used to determine the function to be optimized;

[0134] The first optimization module is used to optimize the function to be optimized using a particle swarm algorithm to obtain a first optimization result;

[0135] The second optimization module is used to change the initial value of the particles of the particle swarm algorithm, and use the particle swarm algorithm again to optimize the function to be optimized to obtain a second optimization result;

[0136] The calculation module is used to calculate the gap coefficient between the first optimization result and the second optimization result;

[0137] The judgment module is used to judge whether the gap coefficient is greater than a set gap coefficient threshold;

[0138] The third optimization module is used to optimize the optimization function using a genetic algorithm to obtain a final optimization result when the gap coefficient is greater than a set gap coefficient threshold;

[0139] The selection module is used for selecting the one with a smaller fitness function value from the first optimization result and the second optimization result as the final optimization result when the gap coefficient is less than or equal to the set gap coefficient threshold;

[0140] The output module is used to output the power supply cluster division result based on the final optimization result selected by the selection module when the gap coefficient is less than or equal to the set gap coefficient threshold;

[0141] And it is used to output the power supply cluster division result based on the final optimization result obtained by the third optimization module when the gap coefficient is greater than the set gap coefficient threshold.

[0142] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A method for dividing a power supply cluster, characterized in that: include: S1, obtain a distribution map containing all nodes to be divided; S2, determine the number of clusters based on the distribution map; S3, determine the function to be optimized; S4, using a particle swarm algorithm to optimize the function to be optimized, and obtaining a first optimization result; S5, changing the initial value of the particle of the particle swarm algorithm, and optimizing the function to be optimized by using the particle swarm algorithm again to obtain a second optimization result; S6, calculating the difference coefficient between the first optimization result and the second optimization result; S7, determining whether the gap coefficient is greater than a set gap coefficient threshold, if so, proceeding to S8, if not, proceeding to S9; S8, using a genetic algorithm to optimize the function to be optimized, obtaining the final optimization result, and entering S10; S9, from the first optimization result and the second optimization result, select the one with the smaller fitness function value as the final optimization result, and enter S10; S10, outputting a power supply cluster division result based on the final optimization result; The nodes to be divided include power supply nodes and power consumption nodes; Determine the number of clusters based on the distribution map, including: S21, establishing a rectangular coordinate system in the distribution map; S22, respectively determining the coordinates of each node to be divided in a rectangular coordinate system; S23, respectively, with x min and x max Indicates the maximum and minimum values ​​of the coordinates of all nodes to be divided on the X-axis, respectively represented by y min and max Indicates the maximum and minimum values ​​of the coordinates of all nodes to be divided on the Y axis; defines a local range Q, and the coordinates x on the X axis and the coordinates y on the Y axis of the points in Q meet the following restrictions: S24, in a local range, determining the number of clusters by means of a circular calculation; In the local scope, the number of clusters is determined by a loop calculation method, including: The first round of calculation: Divide Q into M×M local regions of the same size, and calculate the cycle coefficient of each local region obtained in this round of cycle calculation; The local areas whose cyclic coefficients are greater than or equal to the set cyclic coefficient comparison value are saved to the set K1, and the local areas whose cyclic coefficients are less than the set cyclic coefficient comparison value are saved to the set FN; The hth round of loop calculation, h is greater than or equal to 2: K h-1 Each local area in is divided into M×M local areas of the same size again, and the circulation coefficient of each local area obtained by this round of circulation calculation is calculated respectively; The local areas whose cyclic coefficients are greater than or equal to the set cyclic coefficient comparison value are saved to the set K h , save the local area whose cyclic coefficient is less than the set cyclic coefficient comparison value to the set FN; Judge K h Is it an empty set? If so, filter FN to obtain the number of clusters. If not, perform the next round of calculation.

2. A power supply cluster division method according to claim 1, characterized in that: The calculation formula of the circulation coefficient is: cycef b represents the circulation coefficient of the local area b, snode b represents the total number of nodes in the local region b, allnode represents the total number of nodes in Q, and sarea b represents the area of ​​the local area b, allarea represents the area of ​​Q, w1 and w2 represent the weighted proportion of quantity and area respectively.

3. A method for dividing a power supply cluster according to claim 1, characterized in that: Filter FN to obtain the number of clusters, including: The number of local areas in FN that contain nodes greater than FNth is taken as the number of clusters, and FNth represents the adaptive number threshold.

4. A method for dividing power supply clusters according to claim 3, characterized in that: The calculation formula of the adaptive quantity threshold is: Avenum represents the average number of nodes contained in the local area in FN.

5. The method for dividing a power supply cluster according to claim 1, characterized in that: Determine the functions to be optimized, including: Use the following function as the function to be optimized: cyceF represents the function to be optimized, optcef c represents the clustering coefficient of the nodes in the cth cluster; bua represents the set of all nodes in the cth cluster, bub represents the set of nodes in the cth cluster except node i, and dist i,j Represents the distance between node i and node j, dms b represents the maximum distance between two nodes in bua, spu represents the set of power supply nodes in local area b, csu represents the set of power consumption nodes in local area b, spe u Indicates the amount of power that the power supply node u can provide within a set time length, csue v represents the power consumed by the power-consuming node v within the set time length, w3 and w4 represent the weighted proportion of distance and power respectively, sc represents the total number of all nodes in the cth cluster, and nclst represents the total number of clusters.

6. A method for dividing power supply clusters according to claim 1, characterized in that: The particle swarm algorithm is used to optimize the function to be optimized and obtain the first optimization result, including: According to the set number of iterations H, the function to be optimized is used as the fitness function of the particle swarm algorithm, and the particle swarm algorithm is used to minimize the function to be optimized to obtain H optimization results; Among the H optimization results, the optimization result with the smallest fitness function value is taken as the first optimization result.

7. A power supply cluster division system, characterized in that: It includes an acquisition module, a first determination module, a second determination module, a first optimization module, a second optimization module, a calculation module, a judgment module, a third optimization module, a selection module and an output module; The acquisition module is used to obtain a distribution map containing all nodes to be divided; The first determination module is used to determine the number of clusters based on the distribution map; The second determination module is used to determine the function to be optimized; The first optimization module is used to optimize the function to be optimized using a particle swarm algorithm to obtain a first optimization result; The second optimization module is used to change the initial value of the particles of the particle swarm algorithm, and use the particle swarm algorithm again to optimize the function to be optimized to obtain a second optimization result; The calculation module is used to calculate the gap coefficient between the first optimization result and the second optimization result; The judgment module is used to judge whether the gap coefficient is greater than a set gap coefficient threshold; The third optimization module is used to optimize the optimization function using a genetic algorithm to obtain a final optimization result when the gap coefficient is greater than a set gap coefficient threshold; The selection module is used for selecting the one with a smaller fitness function value from the first optimization result and the second optimization result as the final optimization result when the gap coefficient is less than or equal to the set gap coefficient threshold; The output module is used to output the power supply cluster division result based on the final optimization result selected by the selection module when the gap coefficient is less than or equal to the set gap coefficient threshold; and for outputting a power supply cluster division result based on a final optimization result obtained by the third optimization module when the gap coefficient is greater than a set gap coefficient threshold; The nodes to be divided include power supply nodes and power consumption nodes; Determine the number of clusters based on the distribution map, including: S21, establishing a rectangular coordinate system in the distribution map; S22, respectively determining the coordinates of each node to be divided in a rectangular coordinate system; S23, respectively, with x min and x max Indicates the maximum and minimum values ​​of the coordinates of all nodes to be divided on the X-axis, respectively represented by y min and max Indicates the maximum and minimum values ​​of the coordinates of all nodes to be divided on the Y axis; defines a local range Q, and the coordinates x on the X axis and the coordinates y on the Y axis of the points in Q meet the following restrictions: S24, in a local range, determining the number of clusters by means of a circular calculation; In the local scope, the number of clusters is determined by a loop calculation method, including: The first round of calculation: Divide Q into M×M local regions of the same size, and calculate the cycle coefficient of each local region obtained in this round of cycle calculation; The local areas whose cyclic coefficients are greater than or equal to the set cyclic coefficient comparison value are saved to the set K1, and the local areas whose cyclic coefficients are less than the set cyclic coefficient comparison value are saved to the set FN; The hth round of loop calculation, h is greater than or equal to 2: K h-1 Each local area in is divided into M×M local areas of the same size again, and the circulation coefficient of each local area obtained by this round of circulation calculation is calculated respectively; The local areas whose cyclic coefficients are greater than or equal to the set cyclic coefficient comparison value are saved to the set K h , save the local area whose cyclic coefficient is less than the set cyclic coefficient comparison value to the set FN; Judge K h Is it an empty set? If so, filter FN to obtain the number of clusters. If not, perform the next round of calculation.

Images