Partition division method and device for alternating-current and direct-current hybrid power distribution network

Abstract

The invention discloses a partition division method and device for an AC / DC hybrid power distribution network, and the method comprises the steps: generating an initial gray wolf population adaptive to a power distribution network structure through the topological data of the AC / DC hybrid power distribution network, and laying a foundation for the subsequent optimization; the individual fitness is calculated through a multi-target fitness function, and a target wolf is determined according to the sequence to serve as an iteration guiding direction; in combination with nonlinear control parameter strategy balance search precision and fault constraint guarantee engineering feasibility, residual individuals are iteratively updated, and a closed loop of population initialization-performance evaluation-directional iteration is formed by taking fitness value change as a convergence basis in the process; the power distribution network area divided by the final target scheme is tight in electrical coupling, and power and energy storage are cooperatively matched, so that the cross-regional power interaction loss is reduced, the operation stability in a fault scene is improved, and the comprehensive operation efficiency of the alternating-current and direct-current hybrid power distribution network is effectively enhanced.

Description

Technical Field

[0001] This invention relates to the field of power distribution network zoning technology, and in particular to a method and apparatus for dividing AC / DC hybrid power distribution networks into zones. Background Technology

[0002] With the acceleration of the global energy transition, a large proportion of distributed photovoltaic and other renewable energy sources are being integrated into distribution networks, and traditional AC distribution networks are gradually evolving into hybrid AC / DC structures. Hybrid AC / DC distribution networks interconnect multiple voltage level AC systems with DC systems through DC transformers, which can improve energy absorption efficiency and power supply reliability.

[0003] However, this complex system exhibits diverse fault characteristics, increased uncertainty in power flow distribution, and more complex fault propagation paths. Furthermore, existing distribution network zoning technologies are mostly designed for single voltage levels or purely AC systems, making them ill-suited to the dynamic and uncertain nature of AC / DC hybrid architectures.

[0004] Currently, the main methods for dividing power distribution networks into zones include K-means clustering algorithms or intelligent optimization algorithms such as genetic algorithms and particle swarm optimization. While these methods have made some progress in terms of structural compactness and functional balance, clustering algorithms are sensitive to initial values ​​and are prone to getting trapped in local optima; genetic algorithms are inefficient and time-consuming in the later stages of the search and cannot effectively handle multi-objective optimization problems. In addition, existing schemes do not include energy storage in the zoning indicators and lack robustness to fluctuations in photovoltaic output or sudden load changes. Summary of the Invention

[0005] This invention provides a method and apparatus for dividing AC / DC hybrid distribution networks into zones, which solves the technical problems of existing schemes not including energy storage in the zoning indicators and lacking robustness when photovoltaic output fluctuates or load changes.

[0006] This invention provides a method for dividing AC / DC hybrid distribution networks into zones, comprising:

[0007] Obtain the topology data of the AC / DC hybrid distribution network, and generate an initial gray wolf population according to the topology data;

[0008] The fitness value of each individual gray wolf in the initial gray wolf population is calculated using a preset multi-objective fitness function.

[0009] The gray wolf individuals are sorted according to their fitness values ​​to determine the positions of multiple target wolf individuals at the target sorting positions; each target wolf individual position corresponds to a candidate partitioning scheme.

[0010] Based on the nonlinear control parameter strategy and fault constraints, the positions of the remaining gray wolves, excluding the target wolf individual position, are iteratively updated until the change in the fitness value is less than a preset threshold or the maximum number of iterations is reached. The candidate partitioning scheme corresponding to the target wolf individual position with the maximum fitness value is then determined as the target partitioning scheme.

[0011] The AC / DC hybrid distribution network is divided into multiple distribution network areas according to the target zoning scheme.

[0012] Optionally, the step of acquiring the topology data of the AC / DC hybrid distribution network and generating an initial gray wolf population according to the topology data includes:

[0013] Obtain topology data for AC / DC hybrid distribution networks;

[0014] The number of distribution network nodes and the type of node equipment are determined based on the topology data.

[0015] If the number of node device types is greater than the preset type threshold, then the number of distribution network nodes is determined as a dimension parameter, and the population size is determined according to the number of distribution network nodes combined with the first preset increase.

[0016] If the number of node device types is not greater than a preset type threshold, then the number of distribution network nodes is determined as a dimension parameter, and the population size is determined according to the number of distribution network nodes combined with a second preset increase; the first preset increase is greater than the second preset increase.

[0017] An initial gray wolf population is generated using Tent chaotic mapping based on the stated dimension parameters, the stated population size, and the preset maximum number of iterations.

[0018] Optionally, after performing the step of generating an initial gray wolf population using Tent chaotic mapping according to the dimensional parameters, the population size, and the preset maximum number of iterations, the method further includes:

[0019] Obtain historical source-load uncertainty data for the AC / DC hybrid distribution network;

[0020] Based on Latin hypercube sampling and the historical source load uncertainty data, multiple initial scene matrices are generated;

[0021] The initial scene matrix is ​​reduced using a clustering algorithm to obtain multiple reduced scenes;

[0022] Traverse the reduction scenarios and calculate the cluster size boundary and power constraint respectively;

[0023] The initial gray wolf population is updated according to the cluster size boundary and the power constraint.

[0024] Optionally, the step of calculating the fitness value of each individual gray wolf in the initial gray wolf population using a preset multi-objective fitness function includes:

[0025] Iterate through each individual gray wolf in the initial gray wolf population and calculate the electrical distance modularity, cluster active power balance, cluster reactive power balance, and cluster energy storage supply and demand coordination for each individual gray wolf.

[0026] With the goal of maximizing the energy storage supply and demand coordination of the cluster, the fitness value of each individual gray wolf is obtained by weighting and superimposing the electrical distance modularity, the active power balance of the cluster, the reactive power balance of the cluster, and the energy storage supply and demand coordination of the cluster.

[0027] Optionally, the step of iteratively updating the positions of the remaining gray wolves (excluding the target wolf's position) according to the nonlinear control parameter strategy and fault constraints until the change in the fitness value is less than a preset threshold or the maximum number of iterations is reached, and determining the candidate partitioning scheme corresponding to the target wolf's position with the maximum fitness value as the target partitioning scheme, includes:

[0028] The nonlinear control parameters are calculated according to the nonlinear control parameter strategy, and the search step size coefficient and distance weight coefficient are determined according to the nonlinear control parameters.

[0029] Calculate the actual distance between each of the remaining gray wolf individuals (excluding the target wolf individual's location) and each target wolf individual's location according to the distance weighting coefficient;

[0030] By combining the search step size coefficient with the actual distances, multiple movement vectors corresponding to the position of each individual gray wolf are calculated.

[0031] After updating the position of each individual gray wolf according to all the movement vectors, the position is verified in conjunction with fault constraints.

[0032] If the verification passes, the updated location of the individual gray wolf will be retained;

[0033] If the verification fails, the position of the individual gray wolf will be updated again based on the constraint penalty item corresponding to the fault constraint.

[0034] Jump to the step of calculating the fitness value of each individual gray wolf in the initial gray wolf population using a preset multi-objective fitness function;

[0035] When the change in fitness value is less than a preset threshold or the maximum number of iterations is reached, the candidate partitioning scheme corresponding to the position of the target wolf individual with the maximum fitness value is determined as the target partitioning scheme.

[0036] This invention also provides a method for dividing AC / DC hybrid distribution networks into zones, comprising:

[0037] The population initialization module is used to acquire the topology data of the AC / DC hybrid distribution network and generate an initial gray wolf population according to the topology data.

[0038] The fitness value calculation module is used to calculate the fitness value of each individual gray wolf in the initial gray wolf population using a preset multi-objective fitness function.

[0039] The sorting and filtering module is used to sort the gray wolf individuals according to the fitness value and determine the positions of multiple target wolf individuals at the target sorting position; each target wolf individual position corresponds to a candidate partitioning scheme;

[0040] The position iteration update module is used to perform position iteration update on the remaining gray wolf individual positions other than the target wolf individual position according to the nonlinear control parameter strategy and fault constraints, until the change amplitude of the fitness value is less than a preset threshold or the maximum number of iterations is reached, and the candidate partitioning scheme corresponding to the target wolf individual position with the maximum fitness value is determined as the target partitioning scheme.

[0041] The partitioning module is used to partition the AC / DC hybrid distribution network according to the target partitioning scheme to obtain multiple distribution network areas.

[0042] Optionally, the population initialization module is specifically used for:

[0043] Obtain topology data for AC / DC hybrid distribution networks;

[0044] The number of distribution network nodes and the type of node equipment are determined based on the topology data.

[0045] If the number of node device types is greater than the preset type threshold, then the number of distribution network nodes is determined as a dimension parameter, and the population size is determined according to the number of distribution network nodes combined with the first preset increase.

[0046] If the number of node device types is not greater than a preset type threshold, then the number of distribution network nodes is determined as a dimension parameter, and the population size is determined according to the number of distribution network nodes combined with a second preset increase; the first preset increase is greater than the second preset increase.

[0047] An initial gray wolf population is generated using Tent chaotic mapping based on the stated dimension parameters, the stated population size, and the preset maximum number of iterations.

[0048] Optionally, the population initialization module is further configured to:

[0049] Obtain historical source-load uncertainty data for the AC / DC hybrid distribution network;

[0050] Based on Latin hypercube sampling and the historical source load uncertainty data, multiple initial scene matrices are generated;

[0051] The initial scene matrix is ​​reduced using a clustering algorithm to obtain multiple reduced scenes;

[0052] Traverse the reduction scenarios and calculate the cluster size boundary and power constraint respectively;

[0053] The initial gray wolf population is updated according to the cluster size boundary and the power constraint.

[0054] Optionally, the fitness value calculation module is specifically used for:

[0055] Iterate through each individual gray wolf in the initial gray wolf population and calculate the electrical distance modularity, cluster active power balance, cluster reactive power balance, and cluster energy storage supply and demand coordination for each individual gray wolf.

[0056] With the goal of maximizing the energy storage supply and demand coordination of the cluster, the fitness value of each individual gray wolf is obtained by weighting and superimposing the electrical distance modularity, the active power balance of the cluster, the reactive power balance of the cluster, and the energy storage supply and demand coordination of the cluster.

[0057] Optionally, the position iteration update module is specifically used for:

[0058] The nonlinear control parameters are calculated according to the nonlinear control parameter strategy, and the search step size coefficient and distance weight coefficient are determined according to the nonlinear control parameters.

[0059] Calculate the actual distance between each of the remaining gray wolf individuals (excluding the target wolf individual's location) and each target wolf individual's location according to the distance weighting coefficient;

[0060] By combining the search step size coefficient with the actual distances, multiple movement vectors corresponding to the position of each individual gray wolf are calculated.

[0061] After updating the position of each individual gray wolf according to all the movement vectors, the position is verified in conjunction with fault constraints.

[0062] If the verification passes, the updated location of the individual gray wolf will be retained;

[0063] If the verification fails, the position of the individual gray wolf will be updated again based on the constraint penalty item corresponding to the fault constraint.

[0064] Jump to the step of calculating the fitness value of each individual gray wolf in the initial gray wolf population using a preset multi-objective fitness function;

[0065] When the change in fitness value is less than a preset threshold or the maximum number of iterations is reached, the candidate partitioning scheme corresponding to the position of the target wolf individual with the maximum fitness value is determined as the target partitioning scheme.

[0066] As can be seen from the above technical solutions, the present invention has the following advantages:

[0067] This invention generates an initial gray wolf population adapted to the grid structure using AC / DC hybrid distribution network topology data, laying the foundation for subsequent optimization. Individual fitness is calculated using a multi-objective fitness function, and this ranking determines the target wolves as the direction of iteration. Combining nonlinear control parameter strategies to balance search accuracy and fault constraints to ensure engineering feasibility, the remaining individuals are iteratively updated. This process uses changes in fitness values ​​as the convergence criterion, forming a closed loop of population initialization, performance evaluation, and directional iteration. The final target scheme divides the distribution network area with tight electrical coupling and coordinated power and energy storage matching, reducing cross-regional power exchange losses and improving operational stability under fault scenarios, effectively enhancing the overall operational efficiency of the AC / DC hybrid distribution network. Attached Figure Description

[0068] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0069] Figure 1 A flowchart illustrating the steps of a method for dividing a hybrid AC / DC distribution network according to an embodiment of the present invention;

[0070] Figure 2 This invention provides a schematic diagram of AC / DC hybrid distribution network zoning.

[0071] Figure 3 This is a structural block diagram of an AC / DC hybrid distribution network partitioning device provided in an embodiment of the present invention. Detailed Implementation

[0072] This invention provides a method and apparatus for dividing AC / DC hybrid distribution networks into zones, which addresses the technical problem that existing solutions do not include energy storage in the zoning criteria and lack robustness when photovoltaic output fluctuates or load changes.

[0073] To make the objectives, features, and advantages of this invention more apparent and understandable, the technical solutions of the embodiments of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the embodiments described below are only some embodiments of this invention, and not all embodiments. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0074] Please see Figure 1 , Figure 1 A flowchart illustrating the steps of a method for dividing a hybrid AC / DC distribution network into zones, as provided in an embodiment of the present invention.

[0075] This invention provides a method for dividing AC / DC hybrid distribution networks into zones, comprising:

[0076] Step 101: Obtain the topology data of the AC / DC hybrid distribution network and generate an initial gray wolf population according to the topology data;

[0077] Topology data refers to a dataset that characterizes the physical connections and equipment attributes of AC / DC hybrid distribution networks, including node information (node ​​type, voltage level), connection relationships (line parameters, converter / energy storage device access nodes), etc.

[0078] The initial gray wolf population refers to the initial set of search units constructed based on the gray wolf optimization algorithm. The position of each gray wolf corresponds to a candidate partitioning scheme for the AC / DC hybrid distribution network.

[0079] In this embodiment, information such as node type, voltage level, line connection relationship, and access nodes of converter equipment and energy storage equipment can be collected through the dispatch automation system of the AC / DC hybrid distribution network, such as Supervisory Control and Data Acquisition (SCADA) or Energy Management System (EMS), and integrated to form a structured topology dataset. Based on the number of distribution network nodes in the topology data, the position dimension of each gray wolf in the initial gray wolf population is determined. The initial population is generated using a chaotic mapping method, so that the position vector of each gray wolf corresponds to the cluster affiliation relationship of each node in the topology data, thus completing the construction of the initial gray wolf population.

[0080] In addition, it can synchronize the topology changes after the switching of distribution network equipment and line reconfiguration in real time to ensure that the initial population is adapted to the dynamic operation status of the distribution network. At the same time, it can combine the distribution of key equipment (such as DC transformers) in the topology data to pre-constrain the cluster affiliation of the initial population and improve the engineering feasibility of the initial population.

[0081] In one example of the present invention, step 101 may include the following sub-steps:

[0082] Obtain topology data for AC / DC hybrid distribution networks;

[0083] Determine the number of distribution network nodes and the type of node equipment based on the topology data;

[0084] If the number of node device types is greater than the preset type threshold, the number of distribution network nodes is determined as a dimension parameter, and the population size is determined according to the number of distribution network nodes combined with the first preset increase.

[0085] If the number of node device types is not greater than the preset type threshold, the number of distribution network nodes is determined as a dimension parameter, and the population size is determined according to the number of distribution network nodes combined with the second preset increase; the first preset increase is greater than the second preset increase.

[0086] The initial gray wolf population is generated using Tent chaotic mapping based on the dimension parameters, population size, and preset maximum number of iterations.

[0087] In this embodiment, after parsing the topology data to obtain the number of distribution network nodes and the node equipment types of each node, the population size is determined by statistically analyzing the number of node equipment types. If the number exceeds a preset type threshold (i.e., the distribution network equipment complexity is high), the population size is determined based on the number of distribution network nodes, combined with a first preset increment. If the number does not exceed the type threshold (i.e., the equipment complexity is low), the population size is determined by combining a second preset increment, thus adapting to the search requirements of different equipment complexities through differentiated increments. Based on the determined dimension parameters, population size, and preset maximum number of iterations, a chaotic sequence in the [0,1] interval is generated using the Tent chaotic mapping to ensure a uniform distribution of the solution space and avoid the algorithm getting trapped in local optima. This sequence is mapped to the search space to obtain the position vectors to which each node cluster belongs in the corresponding distribution network. Each position vector corresponds to a candidate partitioning scheme, and all position vectors together form the initial gray wolf population. The parameters of the initial gray wolf population include the gray wolf population size, dimension, maximum number of iterations, initial and final values ​​of control parameters, etc.

[0088] Specifically, the Tent chaotic mapping formula is:

[0089]

[0090] in, The chaotic sequence value generated for iteration number t takes values ​​in the interval [0,1]. The chaotic sequence value generated for iteration number t+1 is determined by the current value. It is generated by the corresponding segmentation rules and its value is in the range of [0,1].

[0091] Furthermore, after performing the step of generating the initial gray wolf population using Tent chaotic mapping according to the dimension parameter, population size, and preset maximum number of iterations, step 101 may also include the following sub-steps:

[0092] Acquire historical source-load uncertainty data for AC / DC hybrid distribution networks;

[0093] Based on Latin hypercube sampling and historical source load uncertainty data, multiple initial scene matrices are generated;

[0094] A clustering algorithm is used to reduce the initial scene matrix, resulting in multiple reduced scenes.

[0095] Iterate through the reduction scenarios and calculate the cluster size boundary and power constraint respectively;

[0096] The initial gray wolf population is updated according to the cluster size boundary and power constraints.

[0097] Historical source-load uncertainty data refers to the fluctuation data of power output (such as distributed photovoltaic and wind power) and load demand (such as AC load and DC load) during the historical operation of AC-DC hybrid distribution networks, reflecting the random characteristics and changing patterns of source and load.

[0098] Latin hypercube sampling is an efficient random sampling method that can cover the probability distribution space of source load variables with fewer samplings, thereby improving the representativeness of the scene.

[0099] The initial scenario matrix refers to a dataset containing multiple source-load fluctuation scenarios. Each scenario corresponds to a specific set of power output and load demand values, reflecting different fluctuation states of the source-load.

[0100] In this embodiment, historical source-load uncertainty data such as distributed photovoltaic output and AC / DC load demand under different time periods and meteorological conditions are extracted from the historical operation database of the distribution network. Based on this data, the probability distribution characteristics of each source-load variable are obtained by fitting. Then, Latin hypercube sampling is used to perform interval-covering sampling on each source-load variable, and the sampling results of different variables are combined into an initial scenario matrix containing multiple sets of source-load fluctuation states. Subsequently, the K-means clustering algorithm is used to cluster the initial scenario matrix based on the Euclidean distance between source-load values ​​of scenarios. The central scenario of each cluster (representing the average characteristics of the scenario) and the extreme scenario where the source-load fluctuation reaches the physical limit are extracted to form a reduction scenario set.

[0101] Next, each reduction scenario is traversed, and the minimum number of clusters that can satisfy the basic power balance within the cluster and the maximum number of clusters that can avoid the clusters being too small and the equipment utilization rate being low are calculated in order to determine the cluster number boundary. The power fluctuation range of each node under different scenarios is also statistically analyzed to obtain the reasonable constraint range of the total power of each cluster, i.e., the power constraint.

[0102] Finally, based on the cluster size boundary and power constraints, the partitioning scheme corresponding to each individual in the initial gray wolf population is verified. For individuals whose cluster size exceeds the boundary, the cluster affiliation of some of their nodes is adjusted to correct the cluster size. For individuals whose total cluster power exceeds the limit, the cluster affiliation of the fine-tuning points is adjusted to ensure that the power of each cluster meets the constraint requirements, thus completing the update of the initial gray wolf population.

[0103] Step 102: Using a preset multi-objective fitness function, calculate the fitness value of each individual gray wolf in the initial gray wolf population;

[0104] A multi-objective fitness function is a function that uses multiple different dimensions of indicators to quantitatively evaluate AC / DC hybrid distribution networks. It is used to transform the comprehensive performance of candidate partitioning schemes into a single comparable value.

[0105] Fitness value is a quantitative indicator used to characterize the overall performance of candidate partitioning schemes for individual gray wolves. The higher the value, the better the performance of the scheme.

[0106] In this embodiment of the invention, the position vector of each individual gray wolf is decoded into a corresponding candidate partitioning scheme, the node set of each cluster under the scheme is extracted, and the cluster-level electrical distance modularity, power balance, energy storage supply and demand coordination and other indicators are calculated based on topology data. Then, the cluster-level indicators are combined into network-wide indicators by combining weights, and finally the fitness value of each individual gray wolf is obtained by weighting.

[0107] It should be noted that the weights of the indicators in the multi-objective fitness function can be dynamically adjusted based on the distribution network operation scenario (such as large-scale renewable energy generation and peak load). For example, the weight of energy storage supply and demand coordination can be increased in the scenario of large-scale renewable energy generation. At the same time, the normalization of fitness values ​​can be introduced to avoid the interference of different indicator dimensions on the evaluation results.

[0108] In one example of the present invention, step 102 may include the following sub-steps:

[0109] Iterate through each individual gray wolf in the initial gray wolf population and calculate the electrical distance modularity, cluster active power balance, cluster reactive power balance, and cluster energy storage supply and demand coordination for each individual gray wolf.

[0110] With the goal of maximizing the coordination of energy storage supply and demand in the cluster, the fitness value of each individual gray wolf is obtained by weighting and superimposing the electrical distance modularity, cluster active power balance, cluster reactive power balance and cluster energy storage supply and demand coordination.

[0111] Electrical distance modularity refers to an index that characterizes the tightness of electrical coupling between nodes within a cluster and the looseness of coupling between clusters in a distribution network partitioning scheme.

[0112] Cluster active / reactive power balance refers to the indicators that reflect the degree of matching between active power output and active load demand, and reactive power output and reactive load demand within a single cluster.

[0113] Cluster energy storage supply and demand coordination refers to an indicator that measures the degree of matching between the output of energy storage equipment and the fluctuation of net source load within a single cluster.

[0114] In this embodiment, each individual gray wolf in the initial gray wolf population is traversed, and the position vector of each individual gray wolf is decoded into the corresponding power distribution network partitioning scheme to determine the cluster set. For each cluster, the electrical distance between nodes within the cluster is calculated based on the power distribution network topology data to obtain the electrical distance modularity of the cluster, which measures the density of the partitioning structure. The formula is as follows:

[0115]

[0116] in, For electrical distance modularity, This represents the edge weight between nodes i and j. Let i be the electrical distance between nodes i and j. This is the cluster affiliation determination function; that is, if node i and node j belong to the same cluster, then... ,otherwise ; It is the sum of the edge weights between all nodes in the distribution network. Let be the sum of the edge weights between node i and other nodes. Let be the sum of the edge weights of node j and other nodes.

[0117] The active power balance of the cluster is obtained by combining source and load data to calculate the deviation between the active power output and load demand within the cluster. The calculation formula is shown below:

[0118]

[0119] in, For the active power balance of the entire distribution network; This represents the total number of clusters. Represents a cluster set The z-th cluster in; The scheduling period; For cluster Cluster active power balance at iteration number t.

[0120] The reactive power balance of the power cluster is obtained by combining the deviation between reactive power output and load demand. The calculation formula is as follows:

[0121]

[0122] For the reactive power balance of the entire distribution network, For cluster Cluster reactive power balance at iteration number t.

[0123] The supply-demand coordination of clustered energy storage is obtained by statistically analyzing the mismatch between the output of energy storage devices and the net load of the source load, reflecting the coordination relationship between energy storage and net load. The calculation formula is as follows:

[0124]

[0125]

[0126]

[0127] In the formula, The maximum absolute value of net power fluctuation over T periods; The net power fluctuation of the z-th cluster at iteration number t is the difference between the energy storage output and the net load power within the cluster. For cluster The coordination of supply and demand in clustered energy storage To coordinate the energy storage supply and demand of the entire distribution network, it reflects the overall matching level of energy storage and source load of the entire power grid.

[0128] The above indicators of each cluster are summarized into the network-wide electrical distance modularity, cluster active power balance, cluster reactive power balance, and cluster energy storage supply and demand coordination. With maximizing the cluster energy storage supply and demand coordination as the core objective, the highest weight is assigned to this indicator, and appropriate auxiliary weights are assigned to the other indicators. The four network-wide indicators are weighted and superimposed to obtain the fitness value corresponding to each individual gray wolf.

[0129] The fitness function aims to maximize the coordination between energy storage supply and demand, and its expression is:

[0130]

[0131] In another example of the present invention, the weight ratio of each indicator can be adjusted according to the real-time operation scenario of the distribution network (such as a surge in new energy generation and peak load). For example, in the scenario of high energy storage utilization, the weight of the supply and demand coordination of cluster energy storage can be further increased. At the same time, a combined weight model of entropy weight method (objective weight) + analytic hierarchy process (subjective weight) can be adopted to reduce the bias of a single weight method and improve the accuracy of the fitness value in evaluating the comprehensive performance of the partitioning scheme.

[0132] Step 103: Sort the individual gray wolves according to their fitness values ​​to determine the positions of multiple target wolf individuals that are in the target sorting position; each target wolf individual position corresponds to a candidate partitioning scheme;

[0133] The target ranking position refers to the sequence position of the gray wolf population after sorting in descending order based on fitness value. It is used to select individuals with better performance in the population, such as the top 3.

[0134] The target wolf individual position refers to the position vector corresponding to the gray wolf individual in the target ranking position, and each position vector corresponds to a candidate partitioning scheme.

[0135] In this embodiment, after calculating the fitness value of each individual gray wolf, the gray wolves in the initial gray wolf population are sorted in descending order according to their fitness values. The gray wolves at the top of the target ranking positions are selected, and their corresponding position vectors are used as the target wolf individual positions. Each target wolf individual position corresponds to a partitioning scheme with high performance, which will guide subsequent population position updates. For example, the position vectors of the top 3 gray wolves are used as the target wolf individual positions, with fitness values ​​from highest to lowest being α wolf, β wolf, and δ wolf, respectively.

[0136] It should be noted that a constraint verification mechanism can be introduced during the sorting process to first filter out gray wolf individuals that do not meet the fault constraint and power balance constraint, and then sort the remaining valid individuals. At the same time, a dynamic number threshold for the target wolf individuals can be set to adjust the number of target sorting positions according to the population size, thereby improving the diversity of population search.

[0137] Step 104: Based on the nonlinear control parameter strategy and fault constraints, iteratively update the positions of the remaining gray wolves except for the target wolf individual position until the change in fitness value is less than the preset threshold or the maximum number of iterations is reached. Then, determine the candidate partitioning scheme corresponding to the target wolf individual position with the maximum fitness value as the target partitioning scheme.

[0138] The nonlinear control parameter strategy refers to the parameter adjustment strategy for adjusting the search characteristics of the Grey Wolf optimization algorithm. By controlling the nonlinear changes of the parameters, the algorithm switches between global search and local exploration, balancing the breadth and accuracy of the search.

[0139] Fault constraints refer to the operational constraints of AC / DC hybrid power distribution under different typical fault scenarios.

[0140] In this embodiment, the search control parameters are adjusted according to a nonlinear control parameter strategy. Based on the target wolf's position, the vector for the remaining gray wolves to move towards it is calculated. This movement vector is then corrected using fault constraints; if the scheme corresponding to the moved position does not meet the fault constraints, a penalty term is introduced to adjust the position, completing the iterative update of the remaining gray wolf positions. After each update, the fitness values ​​of all gray wolves are recalculated, and the above process is repeated until the convergence condition is met. Finally, the target wolf position corresponding to the maximum fitness value is selected, and its corresponding scheme is determined as the target partitioning scheme.

[0141] In one example of the present invention, step 104 may include the following sub-steps:

[0142] The nonlinear control parameters are calculated according to the nonlinear control parameter strategy, and the search step size coefficient and distance weight coefficient are determined according to the nonlinear control parameters.

[0143] Calculate the actual distance between the location of each remaining gray wolf (excluding the target wolf) and the location of each target wolf according to the distance weighting coefficient;

[0144] By combining the search step size coefficient with each actual distance, multiple movement vectors corresponding to the position of each individual gray wolf are calculated;

[0145] After updating the position of each individual gray wolf according to all movement vectors, the position is verified in conjunction with fault constraints.

[0146] If the verification passes, the updated location of the individual gray wolf will be retained;

[0147] If the verification fails, the position of the individual gray wolf will be updated again based on the constraint penalty item corresponding to the fault constraint.

[0148] The jump proceeds to the step of calculating the fitness value of each individual gray wolf in the initial gray wolf population using a preset multi-objective fitness function.

[0149] When the change in fitness value is less than the preset threshold or the maximum number of iterations is reached, the candidate partitioning scheme corresponding to the position of the target wolf individual with the maximum fitness value is determined as the target partitioning scheme.

[0150] The search step size coefficient refers to a coefficient A derived from the nonlinear control parameter, used to control the step size of a gray wolf individual moving towards the target wolf individual. When |A|>1, it enhances global search; when |A|<1, it enhances local exploitation.

[0151] The distance weight coefficient refers to a random coefficient C in the interval [0,2], which is used to randomly amplify or reduce the influence of the target wolf's position to avoid the algorithm getting stuck in local optima.

[0152] The actual distance is used to characterize the degree of difference between the location of the remaining gray wolf individuals and the location of the target wolf individuals. It is calculated based on the element deviation of the location vectors of the two individuals (corresponding to the node cluster affiliation) and reflects the adjustment range of the partitioning scheme.

[0153] The movement vector refers to the vector that guides the direction and magnitude of the position update of an individual gray wolf.

[0154] The constraint penalty term refers to the position correction term (such as the adjustment amount to move closer to the nearest valid position) introduced for the position of gray wolf individuals that do not meet the fault constraints, so that the updated position meets the constraint requirements.

[0155] The fitness value variation refers to the absolute value of the difference in fitness value between two adjacent iterations for the same gray wolf individual.

[0156] In this embodiment of the invention, the nonlinear control parameters for the current iteration are calculated according to a preset nonlinear control parameter strategy (such as a quadratic decay strategy based on the number of iterations). Specifically, the calculation can be performed using the following formula:

[0157]

[0158] in, The nonlinear control parameter for the current iteration number t, To set initial values ​​for control parameters, To control the final value of the parameter, t is the current iteration number. This represents the maximum number of iterations.

[0159] Combined with coefficients randomly generated in the [0,1] interval Determine the search step size coefficient A respectively:

[0160]

[0161] Calculate the distance weighting coefficient: .

[0162] For the remaining gray wolves in the initial gray wolf population (excluding the target wolf), the actual distance between them is calculated based on the element-wise deviation of their position vectors from those of the target wolves, combined with a distance weighting coefficient. Then, the search step size coefficient is multiplied by each actual distance to generate multiple movement vectors for each remaining gray wolf to move towards different target wolves. , , . , , Let represent the vectors that need to be moved towards α, β, and δ wolves during iteration number t, respectively.

[0163] The final movement direction is calculated by vector averaging (or weighting), and this final movement direction vector is used to perform an initial update of the position of each individual gray wolf:

[0164]

[0165] in, This is the final movement direction vector.

[0166] Subsequently, the fault constraints of the AC / DC hybrid distribution network (such as voltage drop limits under DC side faults and power deficit limits under AC side faults) are combined to verify the zoning scheme corresponding to the updated location. If the scheme meets all fault constraints, the location is retained. If it does not meet the constraints, a constraint penalty term that is positively correlated with the degree of constraint violation is introduced (such as adjusting the location to the voltage-compliant cluster affiliation direction when the voltage constraint is violated), and the individual location is updated again.

[0167] After updating the positions of all remaining individuals, proceed to step 102 and repeat the above position update and fitness calculation process until the change in the overall fitness value of the population in two adjacent iterations is less than a preset threshold (or the preset maximum number of iterations is reached). Finally, the candidate partitioning scheme represented by the position of the target wolf individual with the maximum fitness value in the population is determined as the target partitioning scheme for the AC / DC hybrid distribution network. For example, the target wolf individual position corresponding to the maximum fitness value is [1,1,2,3,2]. The number of distribution network nodes is determined by the dimension of this position vector, and the cluster number is determined by the value. The cluster affiliation of each node is directly defined. Therefore, the corresponding target partitioning scheme is that nodes 1-2 belong to cluster 1, nodes 3-5 belong to cluster 2, and node 4 belongs to cluster 3.

[0168] Furthermore, if the maximum fitness value of the population does not improve (or the improvement is less than a small threshold) in multiple consecutive iterations, the remaining number of iterations will be automatically shortened (or the decay rate of the control parameters will be adjusted) to avoid wasting computing power on invalid iterations, while ensuring the optimization accuracy of the target partitioning scheme.

[0169] Step 105: Divide the AC / DC hybrid distribution network into multiple distribution network areas according to the target zoning scheme.

[0170] In this embodiment of the invention, the cluster to which each node of the AC / DC hybrid distribution network belongs is obtained by parsing the target partitioning scheme, and each node is assigned to the corresponding cluster. Each cluster corresponds to a distribution network area. Then, by associating the node composition, boundary range, and the affiliation of devices such as distributed power sources and energy storage devices within each distribution network area, the partitioning of the distribution network is completed. Figure 2As shown in the figure, the application of the optimized partitioning results of the improved Grey Wolf algorithm is visualized. This figure intuitively demonstrates the feasibility and effectiveness of the method in complex distribution network structures, providing a clear basis for partitioned autonomy and fault isolation. The structure in the figure uses a typical AC / DC hybrid distribution network as a background, containing AC and DC systems of multiple voltage levels, interconnected through DC transformers. The partitioning results are clearly marked by blue boundary lines, dividing the distribution network into multiple electrically tightly coupled clusters. Each cluster contains mixed AC and DC nodes, ensuring power self-balancing within the cluster. The distribution network is divided into three distribution network areas (i.e., clusters in the zoning scheme) with substations and 10kV busbars as the core access terminals: Area 1 includes the low-voltage AC part and supporting transformers, and is connected to the medium-voltage DC network through converter equipment to meet the requirements of electrical coupling tightness within the area; Area 2 covers medium-voltage DC and low-voltage DC networks, and integrates converter equipment, distributed power sources and energy storage devices to match the constraints of active / reactive power balance of the cluster and energy storage supply and demand coordination; Area 3 is mainly composed of medium-voltage DC network, equipped with converter equipment and distributed power sources, and the areas are connected through lines and switch stations.

[0171] In addition, regional-level coordination controllers can be configured for each distribution network area in accordance with the needs of distribution network management and control, so as to realize the coordinated control of source, load and storage within the region; at the same time, an inter-regional power interaction mechanism can be established to improve the overall operational reliability of the distribution network through inter-regional power support in extreme scenarios.

[0172] It should be understood that although the steps in the flowcharts of the above embodiments are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the above embodiments may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0173] The AC / DC hybrid distribution network partitioning device provided in the embodiments of the present invention will be described below. The AC / DC hybrid distribution network partitioning device described below can be referred to in correspondence with the AC / DC hybrid distribution network partitioning method described above.

[0174] Please see Figure 3 , Figure 3 The diagram shows a structural block diagram of a hybrid AC / DC distribution network partitioning device provided by an embodiment of the present invention.

[0175] The present invention also provides a device for dividing AC / DC hybrid distribution networks into zones, comprising:

[0176] The population initialization module 301 is used to acquire the topology data of the AC / DC hybrid distribution network and generate an initial gray wolf population according to the topology data.

[0177] The fitness value calculation module 302 is used to calculate the fitness value of each individual gray wolf in the initial gray wolf population using a preset multi-objective fitness function.

[0178] The sorting and filtering module 303 is used to sort gray wolf individuals according to their fitness values ​​and determine the positions of multiple target wolf individuals that are in the target sorting position; each target wolf individual position corresponds to a candidate partitioning scheme;

[0179] The position iteration update module 304 is used to perform position iteration update on the remaining gray wolf individual positions other than the target wolf individual position according to the nonlinear control parameter strategy and fault constraints, until the change amplitude of the fitness value is less than the preset threshold or the maximum number of iterations is reached, and the candidate partitioning scheme corresponding to the target wolf individual position with the maximum fitness value is determined as the target partitioning scheme.

[0180] The partitioning module 305 is used to partition the AC / DC hybrid distribution network according to the target partitioning scheme to obtain multiple distribution network areas.

[0181] Optionally, the population initialization module 301 is specifically used for:

[0182] Obtain topology data for AC / DC hybrid distribution networks;

[0183] Determine the number of distribution network nodes and the type of node equipment based on the topology data;

[0184] If the number of node device types is greater than the preset type threshold, the number of distribution network nodes is determined as a dimension parameter, and the population size is determined according to the number of distribution network nodes combined with the first preset increase.

[0185] If the number of node device types is not greater than the preset type threshold, the number of distribution network nodes is determined as a dimension parameter, and the population size is determined according to the number of distribution network nodes combined with the second preset increase; the first preset increase is greater than the second preset increase.

[0186] The initial gray wolf population is generated using Tent chaotic mapping based on the dimension parameters, population size, and preset maximum number of iterations.

[0187] Optionally, the population initialization module 301 is also specifically used for:

[0188] Acquire historical source-load uncertainty data for AC / DC hybrid distribution networks;

[0189] Based on Latin hypercube sampling and historical source load uncertainty data, multiple initial scene matrices are generated;

[0190] A clustering algorithm is used to reduce the initial scene matrix, resulting in multiple reduced scenes.

[0191] Iterate through the reduction scenarios and calculate the cluster size boundary and power constraint respectively;

[0192] The initial gray wolf population is updated according to the cluster size boundary and power constraints.

[0193] Optionally, the fitness value calculation module 302 is specifically used for:

[0194] Iterate through each individual gray wolf in the initial gray wolf population and calculate the electrical distance modularity, cluster active power balance, cluster reactive power balance, and cluster energy storage supply and demand coordination for each individual gray wolf.

[0195] With the goal of maximizing the coordination of energy storage supply and demand in the cluster, the fitness value of each individual gray wolf is obtained by weighting and superimposing the electrical distance modularity, cluster active power balance, cluster reactive power balance and cluster energy storage supply and demand coordination.

[0196] Optionally, the position iteration update module 304 is specifically used for:

[0197] The nonlinear control parameters are calculated according to the nonlinear control parameter strategy, and the search step size coefficient and distance weight coefficient are determined according to the nonlinear control parameters.

[0198] Calculate the actual distance between the location of each remaining gray wolf (excluding the target wolf) and the location of each target wolf according to the distance weighting coefficient;

[0199] By combining the search step size coefficient with each actual distance, multiple movement vectors corresponding to the position of each individual gray wolf are calculated;

[0200] After updating the position of each individual gray wolf according to all movement vectors, the position is verified in conjunction with fault constraints.

[0201] If the verification passes, the updated location of the individual gray wolf will be retained;

[0202] If the verification fails, the position of the individual gray wolf will be updated again based on the constraint penalty item corresponding to the fault constraint.

[0203] The jump proceeds to the step of calculating the fitness value of each individual gray wolf in the initial gray wolf population using a preset multi-objective fitness function.

[0204] When the change in fitness value is less than the preset threshold or the maximum number of iterations is reached, the candidate partitioning scheme corresponding to the position of the target wolf individual with the maximum fitness value is determined as the target partitioning scheme.

[0205] In the several embodiments provided by this invention, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or modules may be electrical, mechanical, or other forms.

[0206] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical modules; that is, they may be located in one place or distributed across multiple network modules. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0207] The above-described embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for dividing AC / DC hybrid distribution networks into zones, characterized in that, include: Obtain the topology data of the AC / DC hybrid distribution network, and generate an initial gray wolf population according to the topology data; The fitness value of each individual gray wolf in the initial gray wolf population is calculated using a preset multi-objective fitness function. The gray wolf individuals are sorted according to their fitness values ​​to determine the positions of multiple target wolf individuals at the target sorting positions; Each target wolf individual location corresponds to a candidate partitioning scheme; Based on the nonlinear control parameter strategy and fault constraints, the positions of the remaining gray wolves, excluding the target wolf individual position, are iteratively updated until the change in the fitness value is less than a preset threshold or the maximum number of iterations is reached. The candidate partitioning scheme corresponding to the target wolf individual position with the maximum fitness value is then determined as the target partitioning scheme. The AC / DC hybrid distribution network is divided into multiple distribution network areas according to the target zoning scheme.

2. The AC / DC hybrid distribution network zoning method according to claim 1, characterized in that, The step of acquiring the topology data of the AC / DC hybrid distribution network and generating an initial gray wolf population according to the topology data includes: Obtain topology data for AC / DC hybrid distribution networks; The number of distribution network nodes and the type of node equipment are determined based on the topology data. If the number of node device types is greater than the preset type threshold, then the number of distribution network nodes is determined as a dimension parameter, and the population size is determined according to the number of distribution network nodes combined with the first preset increase. If the number of node device types is not greater than a preset type threshold, then the number of distribution network nodes is determined as a dimension parameter, and the population size is determined according to the number of distribution network nodes combined with a second preset increase; the first preset increase is greater than the second preset increase. An initial gray wolf population is generated using Tent chaotic mapping based on the stated dimension parameters, the stated population size, and the preset maximum number of iterations.

3. The AC / DC hybrid distribution network zoning method according to claim 2, characterized in that, After performing the step of generating an initial gray wolf population using Tent chaotic mapping according to the dimensional parameters, the population size, and the preset maximum number of iterations, the method further includes: Obtain historical source-load uncertainty data for the AC / DC hybrid distribution network; Based on Latin hypercube sampling and the historical source load uncertainty data, multiple initial scene matrices are generated; The initial scene matrix is ​​reduced using a clustering algorithm to obtain multiple reduced scenes; Traverse the reduction scenarios and calculate the cluster size boundary and power constraint respectively; The initial gray wolf population is updated according to the cluster size boundary and the power constraint.

4. The AC / DC hybrid distribution network zoning method according to claim 1, characterized in that, The step of calculating the fitness value of each individual gray wolf in the initial gray wolf population using a preset multi-objective fitness function includes: Iterate through each individual gray wolf in the initial gray wolf population and calculate the electrical distance modularity, cluster active power balance, cluster reactive power balance, and cluster energy storage supply and demand coordination for each individual gray wolf. With the goal of maximizing the energy storage supply and demand coordination of the cluster, the fitness value of each individual gray wolf is obtained by weighting and superimposing the electrical distance modularity, the active power balance of the cluster, the reactive power balance of the cluster, and the energy storage supply and demand coordination of the cluster.

5. The AC / DC hybrid distribution network zoning method according to claim 1, characterized in that, The step of iteratively updating the positions of remaining gray wolves (excluding the target wolf) based on a nonlinear control parameter strategy and fault constraints until the change in fitness value is less than a preset threshold or the maximum number of iterations is reached, and determining the candidate partitioning scheme corresponding to the target wolf position with the maximum fitness value as the target partitioning scheme, includes: The nonlinear control parameters are calculated according to the nonlinear control parameter strategy, and the search step size coefficient and distance weight coefficient are determined according to the nonlinear control parameters. Calculate the actual distance between each of the remaining gray wolf individuals (excluding the target wolf individual's location) and each target wolf individual's location according to the distance weighting coefficient; By combining the search step size coefficient with the actual distances, multiple movement vectors corresponding to the position of each individual gray wolf are calculated. After updating the position of each individual gray wolf according to all the movement vectors, the position is verified in conjunction with fault constraints. If the verification passes, the updated location of the individual gray wolf will be retained; If the verification fails, the position of the individual gray wolf will be updated again based on the constraint penalty item corresponding to the fault constraint. Jump to the step of calculating the fitness value of each individual gray wolf in the initial gray wolf population using a preset multi-objective fitness function; When the change in fitness value is less than a preset threshold or the maximum number of iterations is reached, the candidate partitioning scheme corresponding to the position of the target wolf individual with the maximum fitness value is determined as the target partitioning scheme.

6. A device for dividing AC / DC hybrid distribution networks, characterized in that, include: The population initialization module is used to acquire the topology data of the AC / DC hybrid distribution network and generate an initial gray wolf population according to the topology data. The fitness value calculation module is used to calculate the fitness value of each individual gray wolf in the initial gray wolf population using a preset multi-objective fitness function. The sorting and filtering module is used to sort the individual gray wolves according to the fitness value and determine the positions of multiple target wolf individuals at the target sorting position; Each target wolf individual location corresponds to a candidate partitioning scheme; The position iteration update module is used to perform position iteration update on the remaining gray wolf individual positions other than the target wolf individual position according to the nonlinear control parameter strategy and fault constraints, until the change amplitude of the fitness value is less than a preset threshold or the maximum number of iterations is reached, and the candidate partitioning scheme corresponding to the target wolf individual position with the maximum fitness value is determined as the target partitioning scheme. The partitioning module is used to partition the AC / DC hybrid distribution network according to the target partitioning scheme to obtain multiple distribution network areas.

7. The AC / DC hybrid distribution network zoning device according to claim 6, characterized in that, The population initialization module is specifically used for: Obtain topology data for AC / DC hybrid distribution networks; The number of distribution network nodes and the type of node equipment are determined based on the topology data. If the number of node device types is greater than the preset type threshold, then the number of distribution network nodes is determined as a dimension parameter, and the population size is determined according to the number of distribution network nodes combined with the first preset increase. If the number of node device types is not greater than a preset type threshold, then the number of distribution network nodes is determined as a dimension parameter, and the population size is determined according to the number of distribution network nodes combined with a second preset increase; the first preset increase is greater than the second preset increase. An initial gray wolf population is generated using Tent chaotic mapping based on the stated dimension parameters, the stated population size, and the preset maximum number of iterations.

8. The AC / DC hybrid distribution network zoning device according to claim 7, characterized in that, The population initialization module is further used for: Obtain historical source-load uncertainty data for the AC / DC hybrid distribution network; Based on Latin hypercube sampling and the historical source load uncertainty data, multiple initial scene matrices are generated; The initial scene matrix is ​​reduced using a clustering algorithm to obtain multiple reduced scenes; Iterate through the reduction scenarios and calculate the cluster size boundary and power constraint respectively; The initial gray wolf population is updated according to the cluster size boundary and the power constraint.

9. The AC / DC hybrid distribution network zoning device according to claim 6, characterized in that, The fitness value calculation module is specifically used for: Iterate through each individual gray wolf in the initial gray wolf population and calculate the electrical distance modularity, cluster active power balance, cluster reactive power balance, and cluster energy storage supply and demand coordination for each individual gray wolf. With the goal of maximizing the energy storage supply and demand coordination of the cluster, the fitness value of each individual gray wolf is obtained by weighting and superimposing the electrical distance modularity, the active power balance of the cluster, the reactive power balance of the cluster, and the energy storage supply and demand coordination of the cluster.

10. The AC / DC hybrid distribution network zoning device according to claim 6, characterized in that, The position iteration update module is specifically used for: The nonlinear control parameters are calculated according to the nonlinear control parameter strategy, and the search step size coefficient and distance weight coefficient are determined according to the nonlinear control parameters. Calculate the actual distance between each of the remaining gray wolf individuals (excluding the target wolf individual's location) and each target wolf individual's location according to the distance weighting coefficient; By combining the search step size coefficient with each of the actual distances, multiple movement vectors corresponding to the position of each individual gray wolf are calculated. After updating the position of each individual gray wolf according to all the movement vectors, the position is verified in conjunction with fault constraints. If the verification passes, the updated location of the individual gray wolf will be retained; If the verification fails, the position of the individual gray wolf will be updated again based on the constraint penalty item corresponding to the fault constraint. Jump to the step of calculating the fitness value of each individual gray wolf in the initial gray wolf population using a preset multi-objective fitness function; When the change in fitness value is less than a preset threshold or the maximum number of iterations is reached, the candidate partitioning scheme corresponding to the position of the target wolf individual with the maximum fitness value is determined as the target partitioning scheme.

Images