A method for optimizing the topology of a power collection network of an offshore wind farm based on an improved grey wolf optimization algorithm

Abstract

The application provides a kind of offshore wind farm power collection network topology optimization method based on improved grey wolf optimization algorithm, including S1, obtaining the basic data of offshore wind farm and initializing grey wolf algorithm, generating initial solution set;S2, polar coordinates and minimum spanning tree are used to optimize and expand the initial solution set;S3, the connection mode inside each group is optimized;S4, the optimal connection of each group of wind turbine combination is executed;S5, the optimal solution set exploration is carried out by using operator for grey wolf algorithm;S6, repeat execution until the preset maximum iteration number is reached, generate offshore wind farm power collection system topology model;The application can improve the initial solution quality and convergence efficiency, consider global search ability and local optimization precision, realize the fast and effective processing of geometric constraint.

Description

Technical Field

[0001] This invention relates to the field of offshore wind farm power collection systems, and specifically to a method for optimizing the topology of offshore wind farm power collection networks based on an improved gray wolf optimization algorithm. Background Technology

[0002] Offshore wind farm power collection systems play a crucial role in collecting and transmitting the electrical energy generated by each wind turbine to an offshore substation. Their topology directly determines the investment cost, operational reliability, and ease of maintenance during construction. In engineering practice, power collection systems primarily employ four typical topologies: star topology, ring topology, tree topology, and hybrid topology.

[0003] Star topology, centered on a substation, connects each wind turbine directly to the substation via independent lines. It offers advantages such as simple structure, easy protection configuration, and low investment cost. However, due to the lack of redundancy, a failure in one feeder will cause all turbines on that line to shut down, resulting in relatively low system reliability. Ring topology, by connecting turbines into a closed loop, provides redundant paths for power transmission, significantly improving power supply reliability. However, this topology requires laying more cables, leading to a substantial increase in cable investment costs and poor economic efficiency. Tree topology uses a trunk-line structure with multiple branch lines connecting to the main cable. It offers a good balance between scalability and economy, making it particularly suitable for large offshore wind farms with a large number of turbines and a wide distribution. However, a failure in the main cable can cause all downstream turbines to lose power, and topology optimization is challenging.

[0004] Traditional power collection system topology design often relies on the experience and simple comparisons of engineers. When facing large-scale offshore wind farms (with more than 50 turbines), the number of feasible topology combinations increases exponentially, making it difficult to find a truly optimal solution through manual design. Existing research attempts to use mathematical programming methods (such as integer programming and dynamic programming) to solve these problems, but these methods suffer from excessive computational complexity and long solution times when dealing with nonlinear constraints and large-scale discrete variables. While conventional heuristic algorithms (such as genetic algorithms and particle swarm optimization) have some global search capabilities, they are prone to getting trapped in local optima when dealing with multiple complex constraints such as cable current carrying capacity constraints and line crossing geometric constraints, and they also exhibit slow convergence speed and poor stability.

[0005] Furthermore, existing methods often suffer from randomness in generating initial solutions during topology optimization, leading to low quality of the algorithm's search starting point and low iteration efficiency. For the critical geometric constraint of submarine cable crossings, traditional methods typically employ a successive judgment approach, resulting in high computational costs and significantly slowing down the optimization process. Therefore, how to quickly and stably find the optimal economical topology scheme that satisfies both electrical and geometric constraints within a massive feasible solution space has become a pressing technical challenge in the design of offshore wind farm power collection systems. Summary of the Invention

[0006] The purpose of this invention is to solve the problems mentioned in the background section above, and this invention provides the following technical solution:

[0007] A method for optimizing the topology of offshore wind farm collector networks based on an improved gray wolf optimization algorithm includes the following steps:

[0008] S1. Obtain basic data of the offshore wind farm and set the number of wind turbine nodes, maximum number of connection groups, and maximum cable current carrying capacity of the offshore wind farm.

[0009] Simultaneously initialize the maximum number of iterations Tmax of the Grey Wolf algorithm and the leader wolf. Deputy Leader Wolf And the third leader, the wolf The optimal objective function and location;

[0010] Generate an initial set of wind turbine connection sequences as the initial solution set;

[0011] S2. Based on the initialized system parameters and the initial solution set, the initial solution set is optimized and expanded using polar coordinates and minimum spanning tree to generate no less than two initial topology schemes, each of which includes an initial group connection scheme.

[0012] S3. Optimize the connection method within each group by using a greedy algorithm that adds the best connection one by one to reduce the search difficulty;

[0013] S4. By performing the optimal connection for each wind turbine combination, and performing the wiring judgment for the connection within each group, calculate the submarine cable cost of the current topology based on the optimal connection method obtained in step S3 as the objective function value, and calculate the cost of the current solution set.

[0014] S5. The Grey Wolf algorithm is used to explore the optimal solution set by operators. The obtained solution set is sorted and transformed into discrete position vectors, corresponding to the connection method of the wind turbines.

[0015] S6. Repeat steps S3 to S5 until the preset maximum number of iterations Tmax is reached. Obtain the topology scheme and suboptimal solution corresponding to the current optimal objective function value, and generate the topology model of the offshore wind farm collection system.

[0016] More preferably, step S2 specifically includes: taking the booster station as the pole, converting the Cartesian coordinates of each wind turbine into polar coordinates, then sorting the coordinates by index mapping according to the azimuth angle to obtain the initial group connection order, then applying the minimum spanning tree algorithm to optimize the line based on the distance matrix between wind turbines, and expanding multiple initial topology schemes that satisfy geometric constraints through polar coordinate rotation, which serve as the initial population of the Grey Wolf algorithm.

[0017] A further preferred method is to convert the Cartesian coordinates of each wind turbine to polar coordinates as follows:

[0018] ;

[0019] in and These represent the x-coordinates of the i-th wind turbine and the booster station, respectively. and These represent the ordinates of the i-th wind turbine and the substation, respectively. Let be the radial distance from the i-th wind turbine node to the reference point. It is the azimuth angle;

[0020] The specific steps for route optimization based on the distance matrix between wind turbines using the minimum spanning tree algorithm are as follows:

[0021] ;

[0022] ;

[0023] in This represents the straight-line distance between the i-th wind turbine node and the j-th wind turbine node. and These represent the horizontal and vertical coordinates of the i-th wind turbine, respectively. and These represent the x and y coordinates of the j-th wind turbine, respectively. Represented by distance matrix The minimum spanning tree constructed for the weights. Let f(x) represent the set of all edges that satisfy the "spanning tree structure". Represents a specific set of edges in a spanning tree. This represents the total distance of the resulting minimum spanning tree.

[0024] A further preferred embodiment is that step S3 specifically involves: first, initializing the set of connected nodes and the set of nodes to be connected; then, selecting the optimal connection pair based on the cost matrix; and dynamically updating the node set and the line set until all wind turbines in the group are connected.

[0025] Further preferably, in step S4, calculating the cost of the current solution set specifically involves: using the submarine cable cost as the objective function value; if the objective function value of the current solution is better than the historical best value, then the vector cross product method is used to quickly detect line crossings, and dynamic penalties are applied to solutions that violate the crossing constraints to obtain a corrected fitness value; applying dynamic penalties to solutions that violate the constraints specifically includes: only if the cost function of the solution set... Only when the cable crossing constraint violation is determined, the overall objective with penalty and the constraint violation calculation form are as follows:

[0026] ;

[0027] ;

[0028] ;

[0029] ;

[0030] in For cost function, This refers to the unit price coefficient per unit length of the relevant lines. Let be the length of the i-th submarine cable. A function to determine whether two lines intersect. For the penalty function, Let F be the penalty coefficient, and F be the objective function for optimization. and These represent the cross product and the dot product, respectively. L To plan the total number of lines (submarine cables), i and j For route indexing, The first Article and Section j The geometric segments of a submarine cable line; For the set of lines , For the first The coordinate vectors of the endpoints of the line.

[0031] More preferably, in step S5, the individuals in the gray wolf algorithm include four types, namely, the leader wolf. Deputy Leader Wolf And the third leader, the wolf and ordinary individuals The solution set update of the gray wolf algorithm includes prey location encirclement update and guided search to track the prey.

[0032] Further optimization, the specific process of the Grey Wolf algorithm is as follows:

[0033] ;

[0034] in , and The current generation's optimal / second-best / third-best solution (the encoded wind turbine connection sequence). For ordinary individuals, For the updated set of optimal solutions, , and Each represents a consideration of the alpha wolf Deputy Leader Wolf And the third leader, the wolf The location of the prey encirclement and the lead of the prey tracking, A1, A2, and A3 are the contraction-expansion control coefficient vectors for various wolf types.

[0035] More preferably, the Grey Wolf algorithm further includes an inter-group collaborative optimization mechanism:

[0036] In each iteration, through , , The swarm intelligence search guided by the three-tiered leader wolf enables collaborative optimization among different groups. The leader wolf's position vector is updated in real time, guiding ordinary individuals to move towards a better solution space.

[0037] A further preferred embodiment includes a cable current carrying capacity constraint, which is achieved by limiting the maximum number of connections for each group of wind turbines, ensuring that the current carrying capacity of each submarine cable does not exceed its maximum allowable value.

[0038] Compared with existing technologies, the present invention provides a method for optimizing the topology of offshore wind farm collector networks based on an improved gray wolf optimization algorithm, which can bring the following significant benefits:

[0039] (1) Improving the quality of initial solutions and convergence efficiency: To address the problem of poor search starting points and slow convergence caused by random initialization in traditional heuristic algorithms, this invention proposes an initial solution construction method based on polar coordinate transformation and minimum spanning tree. This method utilizes the geometric distribution characteristics of wind turbines relative to the substation, and through polar angle sorting and minimum tree growth, can generate a high-quality initial population that satisfies basic engineering constraints and has diverse distributions at the algorithm startup stage. This not only avoids the computational waste caused by blind search, but also provides a better starting point for subsequent iterative optimization, significantly improving the overall convergence speed and solution stability of the algorithm.

[0040] (2) Balancing global search capability and local optimization accuracy: To overcome the shortcomings of the traditional gray wolf algorithm in dealing with discrete topology optimization problems, this invention constructs a new framework for collaborative optimization of the improved gray wolf algorithm and dynamic grouping strategy. On the one hand, by simulating the social hierarchy of wolf packs ( , , The algorithm employs a wolf-like (and cooperative hunting) mechanism, preserving and enhancing its global exploration capabilities. Furthermore, it introduces a dynamic grouping strategy, breaking down large-scale complex topology optimization problems into sub-problems involving inter-group collaboration and intra-group fine-tuning. This two-tiered mechanism of "global guidance and local refinement" enables the algorithm to effectively escape local optima in the complex solution space and find more economical global topology solutions.

[0041] (3) Fast and efficient handling of geometric constraints: For the strict "no-intersection" geometric constraints in submarine cable laying, traditional methods suffer from heavy computational burden due to checking each cable individually. This invention introduces a fast cross-validation method based on the vector cross product principle and combines it with a dynamic penalty function. This mechanism only initiates precise cross-validation when the algorithm iterates to a relatively optimal solution region, thereby significantly reducing the ineffective geometric computation overhead while ensuring the engineering feasibility of the solution. This allows the algorithm to maintain high computational efficiency when dealing with complex constraint optimization problems in large-scale wind turbine scenarios, combining engineering practicality with algorithmic efficiency. Attached Figure Description

[0042] Figure 1 This is a flowchart of the initial solution set generation process of the method of the present invention;

[0043] Figure 2 This is a flowchart of the intra-group circuit optimization of the improved Grey Wolf algorithm of the present invention;

[0044] Figure 3 This is the overall optimization framework of the improved Grey Wolf algorithm of the present invention. Detailed Implementation

[0045] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0046] The preferred embodiments of the present invention will now be described in further detail with reference to the accompanying drawings.

[0047] Example 1:

[0048] This invention provides a method for optimizing the topology of offshore wind farm collector networks based on an improved gray wolf optimization algorithm. This invention belongs to the fields of deterministic algorithms, heuristic algorithms, and topology optimization of offshore wind farm collector systems. It relates to an optimization method for wind turbine connection topology and is applicable to research on topology optimization of offshore wind farm collector systems.

[0049] This invention proposes a topology optimization method for offshore wind farm collector networks based on an improved Grey Wolf algorithm. The core concept is to model wind turbine generators as network nodes, collector cables as connecting edges, and the substation as the central aggregation point. This transforms the power transmission path planning problem into a minimum-cost path selection optimization problem constrained by cable current carrying capacity and line crossings, where the edge weights represent the cost of laying submarine cables. In terms of solution strategy, this method first uses polar coordinate transformation and a minimum spanning tree algorithm to quickly generate a high-quality initial solution. Subsequently, the improved Grey Wolf algorithm is used for multiple rounds of iterative optimization, simultaneously handling capacity and geometric constraints during the iteration process. By simulating the social hierarchy and cooperative hunting mechanism of a Grey Wolf pack, this method effectively balances global search and local exploitation capabilities, achieving efficient solutions to topology optimization problems under complex constraints.

[0050] The method of this invention includes the following steps: S1, acquiring basic data of the offshore wind farm, setting the number of wind turbine nodes, the maximum number of connection groups, and the maximum current carrying capacity of the cable; and simultaneously initializing the maximum number of iterations Tmax of the Grey Wolf algorithm and the leader wolf algorithm. Deputy Leader Wolf And the third leader, the wolf Find the optimal objective function and location; generate an initial set of wind turbine connection sequences as the initial solution set.

[0051] See the flowchart for the initial solution set generation process. Figure 1 As shown in the figure, the algorithm starts with the original wind turbine coordinate data. First, it sets the location of the booster station as the polar coordinate reference origin, converts the Cartesian coordinates of each wind turbine to polar coordinate representation, and determines the initial connection order of the wind turbines by sorting them by azimuth angle. Then, it constructs the distance matrix between the wind turbine nodes, applies the minimum spanning tree algorithm to generate the basic connection topology, and expands it through polar coordinate rotation operations to obtain diverse initial solution sets. This process, through a combination of geometric transformation and optimization, ensures that the generated initial solution set possesses good population diversity, providing a high-quality search starting point for subsequent swarm intelligence optimization algorithms.

[0052] S2. Based on the initialized system parameters and initial solution set, the initial solution set is optimized and expanded using polar coordinates and minimum spanning tree algorithm to generate at least two initial topology schemes, each of which includes an initial group connection scheme. Step S2 specifically includes: taking the booster station as the pole, converting the Cartesian coordinates of each wind turbine to polar coordinates; then, sorting the coordinates by azimuth angle index mapping to obtain the initial group connection order; then, applying the minimum spanning tree algorithm based on the distance matrix between wind turbines to optimize the line; and expanding multiple initial topology schemes that satisfy geometric constraints through polar coordinate rotation, serving as the initial population for the Grey Wolf algorithm. Specifically, converting the Cartesian coordinates of each wind turbine to polar coordinates involves:

[0053] ;

[0054] in and These represent the x-coordinates of the i-th wind turbine and the booster station, respectively. and These represent the ordinates of the i-th wind turbine and the substation, respectively. Let be the radial distance from the i-th wind turbine node to the reference point. It is the azimuth angle;

[0055] The specific steps for route optimization based on the distance matrix between wind turbines using the minimum spanning tree algorithm are as follows:

[0056] ;

[0057] ;

[0058] in This represents the straight-line distance between the i-th wind turbine node and the j-th wind turbine node. and These represent the horizontal and vertical coordinates of the i-th wind turbine, respectively. and These represent the x and y coordinates of the j-th wind turbine, respectively. Represented by distance matrix The minimum spanning tree constructed for the weights. Let f(x) represent the set of all edges that satisfy the "spanning tree structure". Represents a specific set of edges in a spanning tree. This represents the total distance of the resulting minimum spanning tree.

[0059] S3. Optimize the connection method within each group by using a greedy algorithm that adds the best connection line by line to reduce the search difficulty. Step S3 specifically involves: first, initializing the set of connected nodes and the set of nodes to be connected; then, selecting the optimal connection pairs based on the cost matrix; and dynamically updating the node set and line set until all wind turbines within the group are connected.

[0060] S4. By executing the optimal connection for each wind turbine combination and performing wiring judgment for connections within each group, the submarine cable cost of the current topology is calculated as the objective function value based on the optimal connection method obtained in step S3, and the cost of the current solution set is calculated. Specifically, calculating the cost of the current solution set in step S4 involves using the submarine cable cost as the objective function value. If the objective function value of the current solution is better than the historical best value, the vector cross product method is used to quickly detect line crossings. Dynamic penalties are applied to solutions that violate crossing constraints to obtain a corrected fitness value. Applying dynamic penalties to solutions that violate constraints specifically includes: only if the cost function of the solution set... Only when the cable crossing constraint violation is determined, the overall objective with penalty and the constraint violation calculation form are as follows:

[0061] ;

[0062] ;

[0063] ;

[0064] ;

[0065] in For cost function, This refers to the unit price coefficient per unit length of the relevant lines. Let be the length of the i-th submarine cable. A function to determine whether two lines intersect. For the penalty function, Let F be the penalty coefficient, and F be the objective function for optimization. and These represent the cross product and the dot product, respectively. L To plan the total number of lines (submarine cables), i and j For route indexing, The first Article and Section j The geometric segments of a submarine cable line; For the set of lines , For the first The coordinate vectors of the endpoints of the line.

[0066] See also Figure 2 The improved Grey Wolf algorithm of this invention provides a flowchart for intra-group route optimization. Starting from the current optimization state, the algorithm first establishes a set of connected nodes and a set of nodes to be connected. Based on the turbine spacing matrix, it calculates the connection cost of candidate routes, filters the optimal connection pairs by minimizing the objective function, dynamically updates the node set and route set, and iterates until all nodes are connected. Finally, it verifies and outputs the optimal intra-group topology through geometric constraints. This process employs a greedy strategy to gradually expand the connection graph, ensuring the optimality of each local decision.

[0067] S5. The Gray Wolf algorithm is used to explore the optimal solution set using operators. The obtained solution set is sorted and transformed into discrete position vectors, corresponding to the connection methods of the wind turbines. In step S5, the individuals in the Gray Wolf algorithm include four types, namely the leader wolf. Deputy Leader Wolf And the third leader, the wolf and ordinary individuals The solution set update of the Gray Wolf algorithm includes prey location encirclement update and guided search to track the prey. The specific process of the Gray Wolf algorithm is as follows:

[0068] ;

[0069] in , and The current generation's optimal / second-best / third-best solution (the encoded wind turbine connection sequence). For ordinary individuals, For the updated set of optimal solutions, , and Each represents a consideration of the alpha wolf Deputy Leader Wolf And the third leader, the wolf The location of the prey encirclement and the lead of the prey tracking, A1, A2, and A3 are the contraction-expansion control coefficient vectors for various wolf types.

[0070] The Grey Wolf algorithm also includes an inter-group collaborative optimization mechanism:

[0071] In each iteration, through , , The swarm intelligence search guided by the three-tiered leader wolf enables collaborative optimization among different groups. The leader wolf's position vector is updated in real time, guiding ordinary individuals to move towards a better solution space.

[0072] See also Figure 3 This paper demonstrates the overall optimization framework of the improved gray wolf algorithm of the present invention. After inputting the spatial coordinates of the wind turbine, initial clustering is first achieved through polar coordinate transformation. The Prim algorithm is then applied to construct the minimum spanning tree basic topology, and the submarine cable cost model and gray wolf population parameters are initialized. The algorithm employs intra-group and inter-group optimization: inter-group optimization is achieved through... , , A swarm intelligence search guided by a three-tiered leader wolf enables collaborative optimization between groups; within each group, dynamic candidate routes are added to complete intra-group optimization. During the iteration process, the algorithm updates the leader wolf's position vector in real time and uses fast constraint verification based on vector cross product, ultimately outputting a globally optimal topology scheme that satisfies the constraints.

[0073] S6. Repeat steps S3 to S5 until the preset maximum number of iterations Tmax is reached. Obtain the topology scheme and suboptimal solution corresponding to the current optimal objective function value, and generate the topology model of the offshore wind farm collection system.

[0074] The method of this invention aims to minimize the investment cost of submarine cables while strictly satisfying cable current-carrying capacity constraints and path geometry constraints. The cable current-carrying capacity constraint is achieved by limiting the maximum number of connections for each wind turbine group, ensuring that the current-carrying capacity of each submarine cable does not exceed its maximum allowable value. The path geometry constraint specifically refers to the constraint that line crossings are not allowed during the submarine cable laying process, and step S4 clearly mentions how to handle this constraint.

[0075] Compared with existing technologies, the present invention provides a method for optimizing the topology of offshore wind farm collector networks based on an improved gray wolf optimization algorithm, which can bring the following significant benefits:

[0076] (1) Improving the quality of initial solutions and convergence efficiency: To address the problem of poor search starting points and slow convergence caused by random initialization in traditional heuristic algorithms, this invention proposes an initial solution construction method based on polar coordinate transformation and minimum spanning tree. This method utilizes the geometric distribution characteristics of wind turbines relative to the substation, and through polar angle sorting and minimum tree growth, can generate a high-quality initial population that satisfies basic engineering constraints and has diverse distributions at the algorithm startup stage. This not only avoids the computational waste caused by blind search, but also provides a better starting point for subsequent iterative optimization, significantly improving the overall convergence speed and solution stability of the algorithm.

[0077] (2) Balancing global search capability and local optimization accuracy: To overcome the shortcomings of the traditional gray wolf algorithm in dealing with discrete topology optimization problems, this invention constructs a new framework for collaborative optimization of the improved gray wolf algorithm and dynamic grouping strategy. On the one hand, by simulating the social hierarchy of wolf packs ( , , The algorithm employs a wolf-like (and cooperative hunting) mechanism, preserving and enhancing its global exploration capabilities. Furthermore, it introduces a dynamic grouping strategy, breaking down large-scale complex topology optimization problems into sub-problems involving inter-group collaboration and intra-group fine-tuning. This two-tiered mechanism of "global guidance and local refinement" enables the algorithm to effectively escape local optima in the complex solution space and find more economical global topology solutions.

[0078] (3) Fast and efficient handling of geometric constraints: For the strict "no-intersection" geometric constraints in submarine cable laying, traditional methods suffer from heavy computational burden due to checking each cable individually. This invention introduces a fast cross-validation method based on the vector cross product principle and combines it with a dynamic penalty function. This mechanism only initiates precise cross-validation when the algorithm iterates to a relatively optimal solution region, thereby significantly reducing the ineffective geometric computation overhead while ensuring the engineering feasibility of the solution. This allows the algorithm to maintain high computational efficiency when dealing with complex constraint optimization problems in large-scale wind turbine scenarios, combining engineering practicality with algorithmic efficiency.

[0079] The above description, in conjunction with specific preferred embodiments, provides a further detailed explanation of the present invention. It should not be construed that the specific implementation of the present invention is limited to these descriptions. For those skilled in the art, various simple deductions or substitutions can be made without departing from the concept of the present invention, and all such modifications and substitutions should be considered within the scope of protection of the present invention.

Claims

1. A method for topology optimization of offshore wind farm collector networks based on an improved gray wolf optimization algorithm, characterized in that, Includes the following steps S1. Obtain basic data of the offshore wind farm and set the number of wind turbine nodes, maximum number of connection groups, and maximum cable current carrying capacity of the offshore wind farm. Simultaneously initialize the maximum number of iterations Tmax of the Grey Wolf algorithm and the leader wolf. Deputy Leader Wolf And the third leader, the wolf The optimal objective function and location; Generate an initial set of wind turbine connection sequences as the initial solution set; S2. Based on the initialized system parameters and the initial solution set, the initial solution set is optimized and expanded using polar coordinates and minimum spanning tree to generate no less than two initial topology schemes, each of which includes an initial group connection scheme. S3. Optimize the connection method within each group by using a greedy algorithm that adds the best connection one by one to reduce the search difficulty; S4. By performing the optimal connection for each wind turbine combination, and performing the wiring judgment for the connection within each group, calculate the submarine cable cost of the current topology based on the optimal connection method obtained in step S3 as the objective function value, and calculate the cost of the current solution set. S5. The Grey Wolf algorithm is used to explore the optimal solution set by operators. The obtained solution set is sorted and transformed into discrete position vectors, corresponding to the connection method of the wind turbines. S6. Repeat steps S3 to S5 until the preset maximum number of iterations Tmax is reached. Obtain the topology scheme and suboptimal solution corresponding to the current optimal objective function value, and generate the topology model of the offshore wind farm collection system.

2. The method for topology optimization of offshore wind farm collector networks based on an improved gray wolf optimization algorithm according to claim 1, characterized in that, Step S2 specifically includes: taking the booster station as the pole, converting the Cartesian coordinates of each wind turbine to polar coordinates, then sorting the coordinates by index mapping according to the azimuth angle to obtain the initial group connection order, then applying the minimum spanning tree algorithm to optimize the line based on the distance matrix between wind turbines, and expanding multiple initial topology schemes that satisfy geometric constraints through polar coordinate rotation, which serve as the initial population of the Grey Wolf algorithm.

3. The method for topology optimization of offshore wind farm collector networks based on an improved gray wolf optimization algorithm according to claim 2, characterized in that, The specific steps for converting the Cartesian coordinates of each wind turbine to polar coordinates are as follows: ； in and These represent the x-coordinates of the i-th wind turbine and the booster station, respectively. and These represent the ordinates of the i-th wind turbine and the substation, respectively. Let be the radial distance from the i-th wind turbine node to the reference point. It is the azimuth angle; The specific steps for route optimization based on the distance matrix between wind turbines using the minimum spanning tree algorithm are as follows: ； ； in This represents the straight-line distance between the i-th wind turbine node and the j-th wind turbine node. and These represent the horizontal and vertical coordinates of the i-th wind turbine, respectively. and These represent the x and y coordinates of the j-th wind turbine, respectively. Represented by distance matrix The minimum spanning tree constructed for the weights. Let represent the set of all edges that satisfy the "spanning tree structure". Represents a specific set of edges in a spanning tree. This represents the total distance of the resulting minimum spanning tree.

4. The method for topology optimization of offshore wind farm collector networks based on an improved gray wolf optimization algorithm according to claim 1, characterized in that, Step S3 specifically involves: first, initializing the set of connected nodes and the set of nodes to be connected; then, selecting the optimal connection pair based on the cost matrix; and dynamically updating the node set and the line set until all wind turbines in the group are connected.

5. The method for topology optimization of offshore wind farm collector networks based on an improved gray wolf optimization algorithm according to claim 1, characterized in that, In step S4, the cost of the current solution set is calculated as follows: the cost of the submarine cable is used as the objective function value. If the objective function value of the current solution is better than the historical best value, the vector cross product method is used to quickly detect line crossings. Dynamic penalties are applied to solutions that violate the crossing constraints to obtain the corrected fitness value. Apply dynamic penalties to solutions that violate constraints, specifically including: only if the cost function of the solution set... Only when the cable crossing constraint violation is determined, the overall objective with penalty and the constraint violation calculation form are as follows: ； ； ； ； in For cost function, This refers to the unit price coefficient per unit length of the relevant lines. Let be the length of the i-th submarine cable. A function to determine whether two lines intersect. For the penalty function, Let F be the penalty coefficient, and F be the objective function for optimization. and These represent the cross product and the dot product, respectively. L To plan the total number of lines, i and j For route indexing, The first Article and Section j The geometric segments of the line; For the set of routes , For the first The coordinate vectors of the endpoints of the line.

6. The method for topology optimization of offshore wind farm collector networks based on an improved gray wolf optimization algorithm according to claim 1, characterized in that, In step S5, the individuals in the Gray Wolf algorithm include four types, namely, the leader wolf. Deputy Leader Wolf And the third leader, the wolf and ordinary individuals The solution set update of the gray wolf algorithm includes prey location encirclement update and guided search to track the prey.

7. The method for topology optimization of offshore wind farm collector networks based on an improved gray wolf optimization algorithm according to claim 6, characterized in that, The specific process of the Grey Wolf Algorithm is as follows: ； in , and This represents the current generation's optimal / second-best / third-best solution. For ordinary individuals, For the updated set of optimal solutions, , and Each represents a consideration of the alpha wolf Deputy Leader Wolf And the third leader, the wolf The location of the prey encirclement and the lead of the prey tracking, A1, A2, and A3 are the contraction-expansion control coefficient vectors for various wolf types.

8. The method for topology optimization of offshore wind farm collector networks based on an improved gray wolf optimization algorithm according to claim 7, characterized in that, The Grey Wolf algorithm also includes an inter-group collaborative optimization mechanism: In each iteration, through , , The swarm intelligence search guided by the three-tiered leader wolf enables collaborative optimization among different groups. The leader wolf's position vector is updated in real time, guiding ordinary individuals to move towards a better solution space.

9. A method for optimizing the topology of offshore wind farm collector networks based on an improved gray wolf optimization algorithm according to claim 1, characterized in that, It also includes cable current carrying capacity constraints, which are achieved by limiting the maximum number of connections for each group of wind turbines, ensuring that the current carrying capacity of each submarine cable does not exceed its maximum allowable value.

Images