Wind power plant arrangement processing method and electronic equipment

Abstract

The invention provides a wind power plant arrangement processing method and electronic equipment, and the method comprises the steps: determining an initial point sequence corresponding to each grid parameter set according to a plurality of grid parameter sets in wind power plant data, determining a plurality of first grid parameter sets based on a preset base vector length range and a preset to-be-selected angle range, determining a plurality of second grid parameter sets from the plurality of first grid parameter sets according to the estimated annual energy output, selecting a plurality of to-be-selected intersection points from the initial point sequence corresponding to each second grid parameter set, and performing neighborhood operation on each to-be-selected intersection point according to a neighborhood operation result of each to-be-selected intersection point; and determining a plurality of target intersection points in the target network corresponding to each second grid parameter set, determining a target grid parameter set in the plurality of second grid parameter sets according to the plurality of target intersection points corresponding to each second grid parameter set, and performing wind power plant arrangement based on the plurality of target intersection points of the target grid parameter set. Through two-stage arrangement and site selection, the calculation complexity is low, and the efficiency is high.

Description

Technical Field

[0001] This application relates to the field of renewable energy and power system technology, and more specifically, to a wind farm layout processing method and electronic equipment. Background Technology

[0002] With the accelerated global energy transition, wind power is playing an increasingly crucial role in the renewable energy sector. Due to the dwindling availability of high-quality onshore wind resources, offshore wind power, with its advantages of stable wind speeds, high wind energy density, and large-scale construction, is gradually becoming the core direction of wind power development, especially in deep-sea areas, which are more suitable for large-scale wind farms. The arrangement of wind turbines in offshore wind farms directly affects key indicators such as annual power generation, wake loss, cable routing, and operation and maintenance efficiency; therefore, its optimized design has significant engineering value.

[0003] Existing research on wind farm layout optimization has formed two main directions: one is the wind turbine free layout optimization method based on genetic algorithms, which can explore better solutions in continuous space; the other is to carry out optimization under alignment constraints to meet the actual requirements of shipping safety and operation and maintenance management.

[0004] However, traditional global optimization methods have extremely high computational complexity in wind farms with hundreds of wind turbines, making it difficult to meet the timeliness requirements of engineering projects. Summary of the Invention

[0005] The purpose of this application is to provide a wind farm layout processing method and electronic equipment to address the shortcomings of the prior art and solve the problem of high computational complexity in the prior art.

[0006] To achieve the above objectives, the technical solution adopted in this application is as follows: In a first aspect, this application provides a method for wind farm layout processing, the method comprising: Based on multiple grid parameter sets in the wind farm data, an initial point sequence corresponding to each grid parameter set is determined. Each grid parameter set is used to characterize a target network with a target shape. The target network includes multiple grid intersections. The initial point sequence includes multiple initial points ordered in order of wake loss. Each initial point is one of the multiple grid intersections. Based on a preset range of base vector lengths and a preset range of candidate angles, multiple first grid parameter sets are determined from the multiple grid parameter sets. Based on a preset rapid power generation assessment model, the estimated annual power generation of each first grid parameter set is determined. Based on the estimated annual power generation of each first grid parameter set, multiple second grid parameter sets are determined from the multiple first grid parameter sets. Multiple candidate intersection points are selected from the initial point sequence corresponding to each of the second grid parameter sets, and multiple target intersection points in the target network corresponding to each of the second grid parameter sets are determined based on the neighborhood operation results of each candidate intersection point. Based on a preset objective function, and according to multiple target intersections in the target network corresponding to each second grid parameter set, a target grid parameter set is determined from the multiple second grid parameter sets, and the wind farm is arranged based on the multiple target intersections of the target grid parameter set.

[0007] Optionally, determining the initial point sequence corresponding to each grid parameter set based on multiple grid parameter sets in the wind farm data includes: Based on the set of grid parameters and the preset row and column index values, determine the coordinates of multiple grid intersection points corresponding to the set of grid parameters; Based on the critical range in the wind farm data and the coordinates of multiple grid intersections corresponding to the grid parameter set, multiple intermediate grid intersections among the multiple grid nodes are determined, wherein the critical range includes at least one of the following: boundary range, no-layout range, and minimum safe distance; Determine the wake loss at each intermediate grid intersection point, and determine multiple initial points based on the wake loss at each intermediate grid intersection point. Then, add each initial point to the initial point sequence in sequence according to the wake loss.

[0008] Optionally, each of the mesh parameter sets includes: a first candidate basis vector, a second candidate basis vector, the length of the first candidate basis vector, the length of the second candidate basis vector, a first candidate angle, and a second candidate angle; The determination of multiple first grid parameter sets from the multiple grid parameter sets based on preset basis vector length ranges and preset candidate angle ranges includes: Traverse the multiple sets of mesh parameters. For the current set of mesh parameters, if the lengths of the first candidate basis vector and the second candidate basis vector in the current set of mesh parameters are both within the range of the basis vector lengths, and the first candidate angle and the second candidate angle in the current set of mesh parameters are both within the range of the candidate angles, then the current set of mesh parameters is taken as a first set of mesh parameters.

[0009] Optionally, determining a plurality of second grid parameter sets from the plurality of first grid parameter sets based on the estimated annual power generation of each of the first grid parameters includes: According to the estimated annual power generation of each first grid parameter set, the first grid parameter sets are sorted to obtain a sequence of first grid parameter sets; The first preset number of first grid parameter sets in the first grid parameter set sequence are used as the second grid parameter set.

[0010] Optionally, based on a greedy placement algorithm, a preset number of candidate points are sequentially selected from the initial point sequences corresponding to each of the second grid parameter sets, including: Determine the initial current candidate intersection point from the initial point sequence; Traverse each initial point in the initial point sequence. For the current initial point, determine the boundary lift of the current initial point based on the wind power data of the current initial point and the wind power data of the current candidate intersection point. Based on the boundary lift of the current initial point, determine whether to use the current initial point as a candidate intersection point. If so, use the current initial point as a new current candidate intersection point.

[0011] Optionally, determining the boundary lift of the current initial point based on the wind power data of the current initial point and the wind power data of the current candidate intersection point includes: Based on the wind power data of the current candidate intersection point, calculate the annual power generation of the current candidate intersection point, and based on the wind power data of the current initial point, calculate the annual power generation of the current initial point. The difference between the annual power generation of the current initial point and the annual power generation of the current candidate intersection point is taken as the annual power generation difference value. Calculate the cost of setting up the current candidate intersection points; The quotient of the annual power generation difference of the current candidate intersection point and the cost value is used as the boundary improvement amount of the current candidate intersection point.

[0012] Optionally, determining multiple target intersections in the target network corresponding to each second grid parameter set based on the neighborhood operation results of each candidate intersection point includes: Perform at least one neighborhood operation on the candidate intersection point to obtain the operation-triggered intersection point corresponding to the candidate intersection point. The neighborhood operation includes: point swapping operation, micro-shifting operation, or small-range rotation operation. If the total annual power generation corresponding to the intersection point after the operation is greater than the total annual power generation corresponding to the candidate intersection point, then the intersection point after the operation will be used as a target intersection point. If the total annual power generation of the intersection points obtained after performing all neighborhood operations on the candidate intersection points is less than the total annual power generation of the candidate intersection points, then the candidate intersection points will be selected as the target intersection points.

[0013] Optionally, the objective function is AEP is the annual power generation. It is a set of mesh parameters. It is a vector of decision variables; The step of determining the target grid parameter set from the multiple second grid parameter sets based on the preset objective function and according to multiple target intersections in the target network corresponding to each second grid parameter set includes: Multiple target intersections in the target network corresponding to each of the second grid parameter sets are converted into decision variable vectors, and each of the second grid parameter sets and its corresponding decision variable vectors are input into the objective function to determine the target network parameter set.

[0014] Optionally, the objective function is AEP is the annual power generation. It is a weighting factor for the total cable length. It is the total length of the cable. It is the weighting coefficient of operation and maintenance costs. It's the cost of operation and maintenance. It is the weighting coefficient of the alignment reward. It is an alignment reward; The step of determining the target grid parameter set from the multiple second grid parameter sets based on the preset objective function and according to multiple target intersections in the target network corresponding to each second grid parameter set includes: Based on multiple target intersections in the target network corresponding to each of the second grid parameter sets, the total cable length, maintenance cost, and alignment bonus are determined. Multiple target intersections in the target network corresponding to each second grid parameter set are converted into decision variable vectors. Each second grid parameter set, its corresponding decision variable vector, the total cable length, the operation and maintenance cost, the alignment reward, the weight coefficient of the total cable length, the weight coefficient of the operation and maintenance cost, and the weight coefficient of the alignment reward are input into the objective function to determine the target network parameter set.

[0015] Secondly, this application provides an electronic device, including: a processor, a storage medium, and a bus, wherein the storage medium stores machine-readable instructions executable by the processor. When the electronic device is running, the processor communicates with the storage medium via the bus, and the processor executes the machine-readable instructions to perform the steps of the wind farm layout processing method described above.

[0016] Fourthly, this application provides a computer-readable storage medium storing a computer program, which, when executed by a processor, performs the steps of the wind farm layout processing method described above.

[0017] The beneficial effects of this application are as follows: Based on multiple grid parameter sets in the wind farm data, the initial point sequence corresponding to each grid parameter set is determined, thereby pre-eliminating grid intersections with low power generation potential. This allows subsequent optimization to quickly focus on intersections with low wake loss and power generation potential, avoiding wasting computational resources in low-value areas. Then, based on the range of basis vector lengths and the range of candidate angles, multiple first grid parameter sets are determined from the multiple grid parameter sets, achieving large-scale pruning. Then, based on a fast power generation prediction model, multiple second grid parameter sets are determined from the multiple first grid parameter sets. This avoids the surge in computational load caused by traversing all grid parameters in traditional global optimization, and ensures that the selected grids have high power generation potential through the small-scale model of the fast power generation prediction model. The above method of selecting first and second grid parameter sets quickly narrows the search range of continuous variables, significantly improving site selection efficiency. Furthermore, multiple candidate intersection points are selected from the initial point sequences corresponding to each second grid parameter set. Based on the neighborhood operation results of each candidate intersection point, multiple target intersection points in the target network corresponding to each second grid parameter set are determined. This significantly increases the actual annual power generation of the target intersection points compared to the candidate intersection points. Compared to traditional genetic algorithms and particle swarm optimization algorithms, this scheme significantly improves computational efficiency due to its phased layout optimization, making it suitable for large-scale wind farm deployments. Finally, based on the objective function, the target grid parameter set is determined according to the multiple target intersection points in the target network corresponding to each second grid parameter set. Wind farm deployment is then performed based on these multiple target intersection points, ensuring that the deployment scheme corresponding to the selected target grid parameter set meets power generation and operation and maintenance requirements. The final output deployment scheme requires no additional adjustments and can be directly applied to engineering practice. Additionally, this embodiment improves optimization efficiency by transforming alignment constraints into a parameterized problem compared to direct brute-force point selection and continuous free optimization. Attached Figure Description

[0018] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is a schematic flowchart of a wind farm layout processing method provided in an embodiment of this application; Figure 2 This is a schematic diagram of a target network corresponding to a continuous set of variables provided in an embodiment of this application; Figure 3 This is a flowchart illustrating a method for determining an initial point sequence, as provided in an embodiment of this application. Figure 4 This is a schematic diagram of a process for determining the boundary lift of the current initial point according to an embodiment of this application; Figure 5 This is a flowchart illustrating a process for determining multiple target intersection points, provided in an embodiment of this application. Figure 6 This is a schematic diagram of the structure of a wind farm layout processing device provided in an embodiment of this application; Figure 7 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. It should be understood that the accompanying drawings in this application are for illustrative and descriptive purposes only and are not intended to limit the scope of protection of this application. Furthermore, it should be understood that the schematic drawings are not drawn to scale. The flowcharts used in this application illustrate operations implemented according to some embodiments of this application. It should be understood that the operations in the flowcharts may not be implemented in sequence, and steps without logical contextual relationships may be reversed or implemented simultaneously. In addition, those skilled in the art, guided by the content of this application, may add one or more other operations to the flowcharts, or remove one or more operations from the flowcharts.

[0021] Furthermore, the described embodiments are merely some, not all, of the embodiments of this application. The components of the embodiments of this application described and illustrated herein can typically be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of the application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0022] It should be noted that the term "comprising" will be used in the embodiments of this application to indicate the presence of the features declared thereafter, but does not exclude the addition of other features.

[0023] Traditional global optimization methods suffer from extremely high computational complexity in wind farms with hundreds of turbines, making it difficult to meet the timeliness requirements of engineering projects. Therefore, this application proposes a wind farm layout processing method that performs two-stage layout site selection based on a set of grid parameters. In the first stage, a small set of grid parameters is determined based on a preset range and estimated annual power generation. In the second stage, turbine-by-turbine site selection is performed at the grid intersections corresponding to each set of grid parameters to determine the target intersections for each set. Finally, a target set of grid parameters is determined from the small set of grid parameters according to an objective function, and the wind farm is arranged according to multiple target intersections in the target set. Because this scheme is a two-stage algorithm that gradually reduces the computational range, its computational complexity is low.

[0024] It should be understood that offshore wind farms are located in public waters, and their layout needs to consider ensuring navigational safety and the orderly use of marine resources. Therefore, when the scale of a wind farm reaches hundreds or even thousands of turbines, if the turbine locations are not restricted to grid intersections, the turbine locations become continuous variables, and the search space will expand infinitely. Traditional optimization algorithms will be unable to solve the problem within the project timeframe due to excessive computational complexity. This application utilizes grid intersections to transform continuous variables into grid parameters and a hierarchical design for discrete site selection. Combined with a two-stage layout and site selection, the search space is greatly reduced, enabling efficient solutions for large-scale wind farms.

[0025] Before explaining this plan, let's first introduce the scenario of wind farm layout. In the wind turbine layout process, site selection needs to consider the wind wake generated during energy conversion and engineering feasibility. The wind wake is the area of ​​reduced wind speed downstream after the turbine blades extract energy from the wind; it directly determines the overall power generation of the wind farm. Specifically, the lower the wind wake, the higher the annual power generation. Engineering feasibility can include wiring costs, operation and maintenance costs, etc.

[0026] Figure 1 This is a schematic flowchart illustrating a wind farm layout processing method provided in an embodiment of this application. Figure 1 As shown, the specific steps of the wind farm layout and processing method are as follows.

[0027] S101. Based on multiple grid parameter sets in the wind farm data, determine the initial point sequence corresponding to each grid parameter set. Each grid parameter set is used to characterize a target network with a target shape. The target network includes multiple grid intersections. The initial point sequence includes multiple initial points ordered in order of wake loss. Each initial point is one of the grid intersections among the multiple grid intersections.

[0028] The wind farm data can include multiple grid parameter sets and wind power data. The wind power data includes the overall sea area boundary of the wind farm, the set of no-fly zones, the wind rose dataset, the probability of occurrence in each direction, wind speed distribution parameters, wind turbine power curves, thrust coefficient curves, minimum safe distance, maximum number of wind turbines, and a set of engineering constraints, including cable routing, channel angles, and maintenance paths. Optionally, all wind farm data is projected onto the same plane coordinate system.

[0029] Optionally, the network parameter set can represent a target network with a target shape, where the target shape can be a parallelogram network, a triangular network, or a honeycomb network. This embodiment uses a parallelogram network as an example. It is worth noting that the network shape can be determined based on actual usage; this embodiment does not impose any limitations. The target network can be a constraint boundary for the wind turbine's location, meaning the wind turbine can only fall on grid intersections. Grid intersections are discrete points generated by combining basis vectors in the target network. As an optional implementation, grid intersections can be represented by binary vectors. It means that, among them, Indicates the intersection point Install a fan on top. Indicates the intersection point No fan was installed on it.

[0030] As an optional implementation, the set of mesh parameters can be represented by a set of continuous variables Θ, which can represent a target network. For example, Θ = (r1, r2, θ1, θ2, ν) x , ν ), r1 and r2 are the lengths of the first and second candidate basis vectors, respectively, θ1 and θ2 are the first and second candidate angles, respectively, ν x ν These are the first and second candidate basis vectors, respectively. Figure 2 This is a schematic diagram of a target network corresponding to a continuous set of variables provided in an embodiment of this application. For example... Figure 2 As shown, the set of continuous variables Θ is used to represent a parallelogram network.

[0031] Specifically, a target network is first constructed using a set of grid parameters to obtain all grid intersections. Then, the wake loss at each grid intersection is estimated based on a fast heuristic algorithm, and the points are sorted from largest to smallest wake loss to form an initial point sequence. The fast heuristic algorithm can be, for example, the upwind priority weighting method, the spacing-wake influence mapping method, or the local density-wake superposition coefficient method. It is worth noting that, compared to a precise wake model, the fast heuristic algorithm only considers the key influencing factors of wake loss, using simplified rules or empirical mapping to quickly calculate the wake loss, thus providing a basis for the initial point sequence and, consequently, initial placement criteria during the two-stage site selection. The smaller the wake loss, the less the point is affected by the wake, and the higher the power generation.

[0032] S102. Based on the preset range of base vector lengths and the preset range of candidate angles, determine multiple first grid parameter sets from multiple grid parameter sets, and based on the preset rapid power generation assessment model, determine the estimated annual power generation of each first grid parameter set, and based on the estimated annual power generation of each first grid parameter set, determine multiple second grid parameter sets from multiple first grid parameter sets.

[0033] The range of basis vector lengths can be determined based on the wind turbine impeller diameter to limit the reasonable range of grid edge lengths, avoiding excessive density or sparseness. For example, the range of basis vector lengths can be from 5 times the wind turbine impeller diameter to 12 times the wind turbine impeller diameter.

[0034] The range of angles to be selected is used to limit the directional range of the basis vectors.

[0035] Optionally, this step is the first stage of a two-stage site selection process. Specifically, the first stage consists of two steps. The first step involves determining multiple first grid parameter sets from multiple grid parameter sets based on a preset range, thereby initially narrowing down the range of continuous variables. Site selection at grid intersections within this range has higher power generation potential compared to grid intersections outside this range. The second step involves determining multiple second grid parameter sets from the multiple first grid parameter sets based on the estimated annual power generation of each first grid parameter set, thereby further narrowing down the range of continuous variables. The number of second grid parameter sets is a preset number; for example, three second grid parameter sets are determined from the multiple first grid parameter sets.

[0036] Among them, the length of the basis vector in the first grid parameter set is within a preset range of basis vector lengths, and the selected angle is within a preset range of selected angles.

[0037] The Rapid Energy Production Assessment (AEP) model is a simplified assessment tool designed for efficiently screening wind farm grid patterns. It provides a rapid, conservative upper bound estimate of the annual energy production (AEP) of a wind farm, rather than calculating the precise AEP. The upper bound refers to the theoretically maximum power generation a wind farm can achieve under ideal scenarios, which assume minimal wake interference between turbines and minimal energy loss. It can serve as a relative reference indicator for intuitively comparing the potential power generation capacity of different grid patterns: for example, a regular, compact grid, due to dense turbines and severe wake superposition, will have an actual AEP far lower than this upper bound, while a grid with slightly sparser spacing and orientation aligned with the prevailing wind direction will have an actual AEP closer to the upper bound due to less wake interference. As an optional implementation, the model adopts a balancing strategy in wake processing, using simplified models or engineering experience ratios (such as the matching degree between wind turbine spacing, wind direction and grid direction) for estimation. This method greatly reduces the amount of computation, simplifying the accurate evaluation that originally required millions of calculations to a few thousand, while ensuring the accurate judgment of the trend of power generation potential between different grid patterns, thereby efficiently eliminating low-potential grid parameter sets.

[0038] S103. Select multiple candidate intersection points from the initial point sequence corresponding to each second grid parameter set, and determine multiple target intersection points in the target network corresponding to each second grid parameter set based on the neighborhood operation results of each candidate intersection point.

[0039] Optionally, this step is the second stage of the two-stage site selection process. The second stage also consists of two steps. The first step involves sequentially selecting multiple candidate intersection points from the initial point sequence corresponding to each second grid parameter set. These candidate intersection points can have their precise AEP calculated sequentially according to the order of the initial point sequence, thus refining the selection. Specifically, the precise AEP is further calculated for the initial points sorted using a coarse algorithm, and then greedily placed according to the AEP until the number of selected candidate intersection points reaches the preset number of site selections. This step narrows down the range of discrete site selection. After determining multiple candidate intersection points, the second step is executed: based on the neighborhood operation results of each candidate intersection point, multiple target intersection points in the target network corresponding to each second grid parameter set are determined. The second step addresses the problem of early point selection in the first step leading to subsequent optimization limitations. By performing neighborhood operations on each candidate intersection point, wake loss is further reduced, and the overall AEP is improved. It is worth noting that the number of intersection points remains unchanged in the second step.

[0040] As an alternative implementation, the target intersection point can be determined from the sequence of initial points corresponding to each second grid parameter set based on methods such as mixed integer programming, tabu search, and model predictive control (MPC) rolling.

[0041] Optionally, in the process of determining the initial point sequence in the first stage, the initial point selection can be combined with the minimum safe distance to ensure that the distance between each initial point is greater than the minimum safe distance. In the step of determining the candidate intersection points in the second stage, during the point-by-point selection process, each time a candidate intersection point is added, it is necessary to recheck whether the distance between the candidate intersection point and each of the previously selected candidate intersection points is less than the minimum safe distance, while ensuring that it does not touch constraints such as restricted areas and waterway buffer zones.

[0042] Neighborhood operations can include point swapping, micro-shifting, and / or small-range grid rotation. The specific implementation of neighborhood operations is explained in the following embodiments. The result of a neighborhood operation can be the overall AEP of the current mesh parameter set.

[0043] As an optional implementation, if the growth value of the neighborhood operation result of each candidate intersection point is greater than a preset growth threshold after neighborhood operation, then the intersection point that has undergone neighborhood operation is taken as a new candidate intersection point. This continues until after multiple iterations, the growth value of the neighborhood operation result of each candidate intersection point is no longer greater than the preset growth threshold. Then, each candidate intersection point is taken as the target intersection point. Alternatively, the process continues until the number of iterations exceeds a preset iteration threshold.

[0044] S104. Based on the preset objective function, determine the target grid parameter set in the multiple second grid parameter sets according to the multiple target intersections in the target network corresponding to each second grid parameter set, and arrange the wind farm based on the multiple target intersections of the target grid parameter set.

[0045] Optionally, after determining the discrete variables among the continuous variables in step S103, the optimal variable is further determined from the continuous variables based on the values ​​of the discrete variables. This hierarchical modeling of continuous and discrete variables significantly reduces the dimensionality of the search space, thereby increasing the annual power generation of the wind farm while also significantly reducing computation time. Using the method in this embodiment, the annual power generation of the wind farm is increased by 5%-10% compared to a regular grid, and the computation time is reduced by an order of magnitude.

[0046] As an optional implementation method, multiple target intersection points in the target network corresponding to each second grid parameter set are substituted into the objective function to calculate the comprehensive score, and the parameter set with the best comprehensive score is selected as the target grid parameter set. The corresponding target intersection point is the final wind turbine installation location, and the wind farm layout is completed accordingly.

[0047] Optionally, the objective function may include multiple decision items, such as annual power generation, cable cost, operation and maintenance cost, and alignment reward, and may also include weight coefficients corresponding to each decision item. During implementation, the weight coefficients can be adjusted to adapt to different engineering needs.

[0048] Optionally, after determining the target grid parameter set and the target intersection points therein, the minimum spacing, boundaries, waterway alignment, cable length and branch capacity, and the number and location of booster stations can also be checked.

[0049] As an optional implementation, after determining the target grid parameter set, the target intersection point corresponding to the target grid parameter set is used as the layout coordinate. Based on the annual power generation obtained from the target intersection point, a constraint compliance report and a cable diagram are generated and exported according to a preset format, such as KML / GeoJSON / CSV.

[0050] In this embodiment, based on multiple grid parameter sets in the wind farm data, an initial point sequence corresponding to each grid parameter set is determined, thereby pre-eliminating grid intersections with low power generation potential. This allows subsequent optimization to quickly focus on intersections with low wake loss and power generation potential, avoiding wasting computational resources in low-value areas. Then, based on the range of basis vector lengths and the range of candidate angles, multiple first grid parameter sets are determined from the multiple grid parameter sets, achieving large-scale pruning. Then, based on a fast power generation prediction model, multiple second grid parameter sets are determined from the multiple first grid parameter sets. This avoids the surge in computational load caused by traversing all grid parameters in traditional global optimization, and ensures that the selected grids have high power generation potential through the small-scale model of the fast power generation prediction model. The above method of selecting first and second grid parameter sets quickly narrows the search range of continuous variables, significantly improving site selection efficiency. Furthermore, multiple candidate intersection points are selected from the initial point sequences corresponding to each second grid parameter set. Based on the neighborhood operation results of each candidate intersection point, multiple target intersection points in the target network corresponding to each second grid parameter set are determined. This significantly increases the actual annual power generation of the target intersection points compared to the candidate intersection points. Compared to traditional genetic algorithms and particle swarm optimization algorithms, this scheme significantly improves computational efficiency due to its phased layout optimization, making it suitable for large-scale wind farm deployments. Finally, based on the objective function, the target grid parameter set is determined according to the multiple target intersection points in the target network corresponding to each second grid parameter set. Wind farm deployment is then performed based on these multiple target intersection points, ensuring that the deployment scheme corresponding to the selected target grid parameter set meets power generation and operation and maintenance requirements. The final output deployment scheme requires no additional adjustments and can be directly applied to engineering practice. Additionally, this embodiment improves optimization efficiency by transforming alignment constraints into a parameterized problem compared to direct brute-force point selection and continuous free optimization.

[0051] Next, refer to Figure 3 The specific method for determining the initial point sequence corresponding to each grid parameter set based on multiple grid parameter sets in the wind farm data in step S101 above is described. Figure 3This is a flowchart illustrating a method for determining an initial point sequence, as provided in an embodiment of this application.

[0052] S301. Based on the set of grid parameters and the preset row and column index values, determine the coordinates of multiple grid intersection points corresponding to the set of grid parameters.

[0053] The row and column index values ​​are integer parameters used to locate the grid intersection points, denoted as (k1, k2). k1 corresponds to the row number along the direction of the first candidate basis vector a1, and k2 corresponds to the column number along the direction of the second candidate basis vector a2. For example, if k1=3 and k2=5, it represents the intersection point of the 3rd unit along the direction of a1 and the 5th unit along the direction of a2. The range of values ​​is determined by the overall sea area boundary and the length of the basis vectors. For example, if the length of the sea area along the direction of a1 is 2000m and r1=100m, then the maximum value of k1 is 20.

[0054] The coordinates of the grid intersections are the two-dimensional coordinates of discrete points generated by the set of grid parameters in the target network, and the unit can be, for example, meters. The coordinates can be projected onto the Universal Transverse Mercator (UTM) plane coordinate system.

[0055] Specifically, the coordinates of multiple grid intersection points can be obtained based on the set of grid parameters and through formula (1): (1) in, and These are the first and second candidate basis vectors, respectively. , and These are the lengths of the first and second candidate basis vectors, respectively. and These are the first and second candidate angles, respectively. and These are the row index and column index, respectively, and both are integers. ,in, These are the coordinates of the grid intersections generated by the parameters. It can be the origin of the coordinate system. It can be based on Obtained by offsetting.

[0056] S302. Based on the critical range in the wind farm data and the coordinates of multiple grid intersections corresponding to the grid parameter set, determine multiple intermediate grid intersections among multiple grid nodes. The critical range includes at least one of the following: boundary range, no-layout range, and minimum safe distance.

[0057] The boundary range is the overall sea area boundary in the wind farm data, and the no-deployment range can be determined based on the no-deployment zone set in the wind farm data.

[0058] Specifically, the grid intersections that satisfy the critical range are taken as intermediate grid intersections.

[0059] As an optional implementation, a sparsity parameter can be preset, which controls the minimum distance ratio between candidate points, ranging from 0 to 1. Then, based on the sparsity parameter and the minimum safe distance, intermediate grid intersections are determined. Specifically, if the distance between two points is less than the product of the sparsity parameter and the minimum safe distance, the grid intersection is retained as an intermediate grid intersection according to a priority rule. In this priority rule, points closer to the sea boundary or with a smaller wake have higher priority.

[0060] For example, the intersection of intermediate grids can be represented by constraint modeling: the coordinates of the nth grid intersection. ,in, It is a set of multiple grid intersection points corresponding to a set of grid parameters, the first... With the Euclidean distance of typhoon generator ,in, It is the minimum safe distance, and the coordinates of the nth grid intersection point. ,in, It is the overall sea area boundary of the wind farm. It refers to a collection of restricted areas, such as waterways and protected areas.

[0061] S303. Determine the wake loss at each intermediate grid intersection point, and determine multiple initial points based on the wake loss at each intermediate grid intersection point. Add each initial point to the initial point sequence in sequence according to the wake loss.

[0062] Specifically, first, a rapid assessment method for wake loss is determined. Taking the upwind probability weighted method as an example, if the prevailing wind direction is northeast (40% probability) and east (30% probability), and a point of intersection is simultaneously upwind of both northeast and east winds, then its wake loss score = 1 - (0.4 + 0.3) = 0.3. The lower the score, the smaller the wake loss. Second, the wake loss of each intermediate grid intersection is calculated, and intersections with losses below a preset threshold are selected as initial points. The preset threshold can be adjusted according to the wind field scale. Finally, all initial points are sorted from smallest to largest wake loss. If the losses are the same, they are sorted from closest to furthest from the maintenance base or from closest to furthest from the booster station, ultimately forming an initial point sequence.

[0063] In this embodiment, multiple intermediate grid intersections are determined based on the critical range and the coordinates of the grid intersections. Then, the initial point sequence is determined based on the wake loss of each intermediate grid intersection, thereby rationally screening the discrete addresses that can be addressed. Furthermore, by sorting the initial points according to the wake loss, priority criteria are provided for subsequent fine-grained optimization, which greatly improves the optimization convergence speed.

[0064] Optionally, each set of grid parameters includes: a first candidate basis vector, a second candidate basis vector, the length of the first candidate basis vector, the length of the second candidate basis vector, a first candidate angle, and a second candidate angle.

[0065] The following section describes the first step of the first stage in the two-stage site selection process. This step is used to narrow down multiple sets of grid parameters to achieve coarse-grained surface sweeping within a reasonable range.

[0066] The specific method for determining multiple first grid parameter sets from multiple grid parameter sets in step S102 above, based on the preset range of basis vector lengths and the preset range of candidate angles, is as follows: Traverse multiple grid parameter sets. For the current grid parameter set that has been traversed, if the lengths of the first candidate basis vector and the second candidate basis vector in the current grid parameter set are both within the range of basis vector lengths, and the first candidate angle and the second candidate angle in the current grid parameter set are both within the range of candidate angles, then the current grid parameter set is taken as a first grid parameter set.

[0067] For example, if the length of the basis vector is in the range of 5D to 12D, where D is the diameter of the wind turbine impeller and the range of the selected angle is 60 degrees to 120 degrees, then the set of grid parameters with the length of the first selected basis vector and the length of the second selected basis vector in the range of 5D to 12D, and the first selected angle and the second selected angle in the range of 0 degrees to 180 degrees, is selected as a first grid parameter set.

[0068] In this embodiment, multiple first grid parameter sets are obtained by filtering from multiple grid parameter sets through the range of basis vector lengths and the range of candidate angles, thereby greatly reducing the number of objects to be evaluated. Compared with directly evaluating the power generation of all sets, this greatly improves the site selection efficiency.

[0069] The following section describes the second step in the first stage of the two-stage site selection process. This step is used to further filter out a small number of second grid parameter sets from multiple first grid parameter sets in order to significantly reduce dimensionality through pruning.

[0070] The specific implementation steps for determining multiple sets of second grid parameters from multiple sets of first grid parameters in step S102 above are as follows: Optionally, the first grid parameter sets are sorted according to their estimated annual power generation to obtain a sequence of first grid parameter sets. The first preset number of first grid parameter sets in the sequence are then used as the second grid parameter sets.

[0071] The estimated annual power generation can be obtained based on the rapid power generation assessment model. Although these estimates are not precise AEPs, they accurately reflect the differences in power generation capacity of different grid types. For example, grids adapted to the prevailing wind direction have higher estimated annual power generation, while grids with dense wind turbines have lower estimated annual power generation. Calculating the estimated annual power generation can provide a reliable quantitative standard for ranking.

[0072] Optionally, all first grid parameter sets are sorted in descending order of estimated annual power generation to form a sequence of first grid parameter sets. The first predetermined number of first grid parameter sets are then extracted from this sequence and defined as the second grid parameter set. This step bridges the gap from compliance to high potential, significantly reducing subsequent computational burden.

[0073] Optionally, if multiple sets of first grid parameters have the same estimated annual power generation, a secondary sorting rule can be added to sort the first grid parameter sets. The secondary sorting rule can take into account engineering feasibility, including priority given to engineering cost, alignment, and constraint compliance.

[0074] As an optional implementation, if computational resources are sufficient and the wind field size is relatively small, the preset number can be set to a larger number, thereby retaining more high-potential sets and covering a more comprehensive mesh pattern during subsequent fine-tuning optimization, reducing the risk of missing the optimal solution. If computational resources are limited and the wind field size is large, the preset number can be set to a smaller number, controlling the number of objects to be optimized subsequently to avoid exceeding the project time limit due to excessive computation. If the accuracy of the rapid evaluation model is high, the preset number can be appropriately reduced; if the accuracy is low, the number needs to be increased.

[0075] In this embodiment, multiple sets of second grid parameters are determined from multiple sets of first grid parameters based on the estimated annual power generation of each first grid parameter, thereby further narrowing the scope and providing accurate candidate objects for subsequent refined site selection.

[0076] The following section describes the first step of the second stage in the two-stage site selection process. This step is used to select candidate intersection points from the parameter sets of each second grid.

[0077] The specific steps of selecting multiple candidate intersection points from the initial point sequence corresponding to each second grid parameter set in step S103 are as follows: Optionally, the initial current candidate intersection point can be determined from the initial point sequence.

[0078] Specifically, the initial point sequence is the result of sorting the wake loss from smallest to largest; the smaller the wake loss, the higher the power generation potential. Therefore, the initial current candidate intersection point can be selected from the beginning of the sequence, for example, from the first three initial points.

[0079] As an optional implementation method, points that simultaneously satisfy the basic engineering constraints can be prioritized: if the first point in the initial point sequence simultaneously satisfies "distance from the booster station ≤ preset threshold" and "not on the edge of the channel buffer zone", then it is directly determined as the initial current candidate intersection point. If the first point has a conflict with the engineering constraints, the selection is carried out sequentially until the first point that simultaneously satisfies the minimum wake loss and compliance with the basic constraints is found, and this point is taken as the initial current candidate intersection point.

[0080] Optionally, iterate through each initial point in the initial point sequence. For the current initial point, determine the boundary lift amount based on the wind power data of the current initial point and the wind power data of the current candidate intersection point. Based on the boundary lift amount of the current initial point, determine whether to include the current initial point as a candidate intersection point. If so, include the current initial point as the new current candidate intersection point.

[0081] The traversal order can start from the initial point sequence, excluding the initial current candidate intersection point, and proceed sequentially according to the original order of the sequence. That is, high-potential points are evaluated first to avoid wasting computational resources on low-potential points.

[0082] It is worth noting that the wind power data for the current candidate intersection point includes the wind power data of all currently determined candidate intersection points. For example, if the process is currently iterating to determine the 4th candidate intersection point, then the wind power data for the current candidate intersection point is the wind power data of the first 3 candidate intersection points.

[0083] Optionally, the boundary lift of the current initial point is the boundary lift after adding the current initial point to each of the current candidate intersection points.

[0084] The wind power data includes the wind speed distribution at the current initial point, the wake influence range, the occupied cable capacity, the distance from the current initial point to the booster station, the distance from the current initial point to the operation and maintenance base, and the distance from the current candidate intersection point.

[0085] Specifically, based on the wind power data at the current initial point and the wind power data at the current candidate intersection point, the annual power generation difference and the replacement value are determined, and based on the annual power generation difference and the replacement value, the boundary improvement amount of the current candidate intersection point is determined.

[0086] As an optional implementation, if the boundary lift is greater than 1, it means that adding the current initial point results in an increase in the wind farm's power generation revenue exceeding the increase in constraint costs, thus improving overall efficiency. In this case, the point is added to the candidate intersection set, and the current candidate intersection is updated to a combination of the original candidate intersection and the newly added point. If the boundary lift is less than or equal to 1, it means that adding the point will lead to a decrease in overall efficiency. In this case, the point is skipped, and the process continues to traverse the next initial point.

[0087] In this embodiment, the initial points are traversed according to the initial point sequence to prioritize the evaluation of high-potential points, reduce invalid calculations, adapt to the rapid point selection requirements of hundreds of wind farms, and combine the boundary lift of the initial points to determine the candidate intersection points, ensuring that the selected points are economical for engineering implementation.

[0088] Next, refer to Figure 4 This paper introduces a method for determining the boundary lift of the current initial point based on the wind power data of the current initial point and the wind power data of the current candidate intersection points. Specifically, Figure 4 This is a schematic diagram of a process for determining the boundary lift of the current initial point, provided in an embodiment of this application.

[0089] S401. Based on the wind power data of the current candidate intersection point, calculate the annual power generation of the current candidate intersection point, and based on the wind power data of the current initial point, calculate the annual power generation of the current initial point. The difference between the annual power generation of the current initial point and the annual power generation of the current candidate intersection point is taken as the annual power generation difference value.

[0090] The following section will introduce how to calculate the annual power generation at the current initial point i.

[0091] The layout of the current candidate intersection points is used as... Its total power generation is: When trying to start at the current initial point After installing one fan, the new layout is as follows: The corresponding power generation is: The annual power generation difference is as shown in the following formula (2): (2) in, Let be the annual power generation difference when the wind turbine is installed at intersection point i.

[0092] As an alternative implementation method, the annual power generation can be calculated as follows: First, establish a wake model, then superimpose the wakes of each wind turbine to obtain the superimposed wake value, and then obtain the annual power generation based on the superimposed wake value.

[0093] Specifically, in the wake model, the wind direction... Fan The centerline expansion of the wake is given by formulas (3), (4), and (5): (3) (4) (5) Where x represents the downwind direction, y represents the crosswind direction, and z represents the vertical direction. This indicates free wind speed, expressed in meters. and , , represent the lateral and vertical standard deviations in the Gaussian wake model, respectively, in meters; H is the wind turbine hub height, in meters; and D is the wind turbine impeller diameter, in meters. The wind speed loss due to the wake is expressed in meters per second. The lateral offset of the wake center is expressed in meters. Ky and kz describe the linear growth rate of the wake half-width with downstream distance x, where Ky is the lateral wake expansion coefficient and kz is the vertical wake expansion coefficient. and For the initial wake half-width, The thrust coefficient can be determined based on... get: For the thrust coefficient curve, input the wind speed. Output fan thrust coefficient .

[0094] In the global coordinates of the wind farm, each wind direction Perform a rotation transformation on each, as shown in the following formula (6): (6) in, and Let i be the coordinates of the upstream wind turbine. and Let x be the coordinates of the downstream wind turbine. If x is greater than 0, it means that wind turbine j is located downstream of wind turbine i and may be affected by the wake.

[0095] and It can be calculated using the following formula (7): (7) The result of the wake superposition can be determined by kinetic energy superposition or by the square root method, as shown in the following formula (8): (8) in, It is the effective inflow velocity of the fan, measured in meters per second. The velocity deficit of the j-th wind turbine wake relative to the target wind turbine being calculated is expressed in meters per second.

[0096] Furthermore, the annual power generation can be calculated using formulas (9) and (10): (9) (10) in, The desired power output of the nth wind turbine can be determined based on the wind turbine power curve. Specifically, the wind speed U is input. The power of the fan is obtained. This is for calculating the mathematical expectation of two random variables: wind speed U and wind direction d. Let be the wind speed distribution function, specifically, be the distribution of wind speed in the wind direction. The wind speed probability density is given by, for example, the wind speed distribution function could be a Weibull distribution.

[0097] S402. Calculate the cost of setting up the current candidate intersection points.

[0098] Here, the cost can be the unit constraint cost, representing the cost or constraint penalty incurred if the wind turbine is installed at that intersection.

[0099] As an optional implementation method, the cost can be calculated from the incremental cable cost, the incremental maintenance cost, the weighting coefficient of the incremental cable cost, and the weighting coefficient of the incremental maintenance cost.

[0100] S403. The quotient of the annual power generation difference and the cost value of the current candidate intersection point is used as the boundary improvement amount of the current candidate intersection point.

[0101] For example, the boundary lift can be determined by the following formula (11): (11) in, This represents the boundary lift amount at the i-th intersection point where the wind turbine is deployed. The annual power generation difference for the wind turbines deployed at the i-th intersection point. The cost of deploying a wind turbine at the i-th intersection point.

[0102] In this embodiment, the boundary improvement amount of the current candidate intersection point is determined based on the quotient of the annual power generation difference and the cost value. Thus, while considering the annual power generation, the cost and operation and maintenance costs of the candidate intersection point are also taken into account when selecting the candidate intersection point.

[0103] Next, we will introduce the second step of the second stage in the two-stage site selection process. This step involves fine-tuning the wind turbine location through neighborhood operations.

[0104] Figure 5This is a flowchart illustrating the process of determining multiple target intersection points according to an embodiment of this application. (Refer to...) Figure 5 The specific method for determining multiple target intersection points in the target network corresponding to each second grid parameter set in step S103 above, based on the neighborhood operation results of each candidate intersection point, is as follows: S501. Perform at least one neighborhood operation on the candidate intersection point to obtain the intersection point after the operation corresponding to the candidate intersection point. The neighborhood operation includes: point swapping operation, micro-shifting operation, or small-range rotation operation.

[0105] The point-swapping operation specifically involves selecting a new intersection point from the grid intersection points near the candidate intersection point to replace the original candidate intersection point. These nearby grid intersection points can be, for example, adjacent intersection points 1-2 grid units away from the current point. For instance, if candidate intersection point A is located in the wake region of an upstream wind turbine, it can be replaced by the adjacent intersection point B, where B is not in the wake region, while ensuring that the distance between B and other intersection points still meets safety constraints.

[0106] The fine-shift operation involves subtly adjusting the coordinates of the selected intersection point within a very small range, causing it to deviate from the original grid intersection point while maintaining its relative alignment with surrounding points. This adjustment typically does not exceed a preset small multiple of the grid unit length. For example, if the preset small multiple is 0.5 and the basis vector length is 100m, the fine-shift range is ≤50m. A displacement operation, for instance, might involve shifting an intersection point 30m along a direction perpendicular to the prevailing wind direction to avoid the core region of the upstream wake, without affecting the overall grid regularity.

[0107] Small-scale grid rotation operation specifically involves rotating a local grid containing the candidate intersection points at a small angle to adjust the relative angles of all intersection points within that area to adapt to local wind direction or terrain. A local grid might be a 3×3 grid area centered on that point. For example, if the prevailing wind direction in a certain area deviates by 5° from the overall grid direction, small-scale grid rotation can be used to align the local grid with the wind direction, reducing wake superposition.

[0108] As an optional implementation, for each candidate intersection point, at least one neighborhood operation is performed to generate a preset number of operation-triggered intersection points. The preset number can be adjusted according to computing resources. These intersection points are spatially adjacent to the original candidate intersection points, but their location or local shape is more suitable for wind resource conditions.

[0109] S502. If the total annual power generation corresponding to the intersection point after the operation is greater than the total annual power generation corresponding to the candidate intersection point, then the intersection point after the operation will be used as a target intersection point.

[0110] Optionally, the total annual power generation corresponding to the intersection point after the operation is the result of the neighborhood operation.

[0111] Optionally, if the AEP of the intersection point is higher after the operation, it indicates that the adjustment is effective. In this case, the intersection point after the operation is determined as the target intersection point, replacing the original candidate intersection point.

[0112] S503. If the total annual power generation of the intersection points obtained after performing all neighborhood operations on the candidate intersection points is less than the total annual power generation of the candidate intersection points, then the candidate intersection points will be used as target intersection points.

[0113] Optionally, if after performing all the preset neighborhood operations on a candidate intersection point, the total annual power generation of all the generated intersection points after the operations is lower than that of the original candidate intersection point, it indicates that the location or shape of the current candidate intersection point has adapted to the local wind resources and constraints. Further adjustments would reduce the efficiency. In this case, the original candidate intersection point is directly determined as the target intersection point.

[0114] In this embodiment, the candidate intersection point is optimized by performing a neighborhood operation to obtain the target intersection point, so that the output target intersection point has both compliance and high power generation potential.

[0115] Next, the process of determining the target mesh parameter set will be introduced. Three objective functions will be used as examples.

[0116] As a first optional implementation, the objective function is: AEP is the annual power generation. It is a set of mesh parameters. It is a vector of decision variables.

[0117] Among them, the decision variable vector is a numerical vector that quantifies the target intersection point.

[0118] Optionally, multiple target intersection points in the target network corresponding to each second grid parameter set are converted into decision variable vectors, and each second grid parameter set and its corresponding decision variable vector are input into the objective function to determine the target network parameter set.

[0119] Specifically, the intersection points of each objective are converted into a vector of decision variables, that is, if This indicates the selection of the intersection point. The target intersection point.

[0120] Optionally, each set of second grid parameters and its corresponding decision variable vector are input into the objective function to obtain the AEP corresponding to each set of second grid parameters, and then the set of second grid parameters corresponding to the largest AEP is used as the target network parameter set.

[0121] This objective function is suitable for scenarios where the focus is on annual power generation.

[0122] As a second optional implementation, the objective function is AEP is the annual power generation. It is a weighting factor for the total cable length. It is the total length of the cable. It is the weighting coefficient of operation and maintenance costs. It's the cost of operation and maintenance. It is the weighting coefficient of the alignment reward. It is an alignment reward; Optionally, the total cable length, maintenance cost, and alignment bonus are determined based on multiple target intersections in the target network corresponding to each second grid parameter set.

[0123] In this diagram, the fan and the substation are considered as nodes, and the cable is considered as an edge. The total cable length is determined using formula (12): (12) in, It is the total length of the cable. , The difference in distance between intersection point i and intersection point j. , This indicates whether the cable is laid between intersection point i and intersection point j.

[0124] The operation and maintenance cost can be determined using formula (13): (13) in, For operation and maintenance costs, b is the preset coordinates of the operation and maintenance base, which corresponds to the coordinates of the wind farm data. The closer the wind turbine is to the base, the lower the operation and maintenance cost.

[0125] The alignment bonus can be determined using formula (14): (14) in, This is an alignment bonus, where N is the total number of wind turbines in the wind farm. It is the first candidate basis vector. It is the angle between the line vector connecting wind turbine i to wind turbine j and the first candidate basis vector.

[0126] Optionally, multiple target intersection points in the target network corresponding to each second grid parameter set are converted into decision variable vectors, and each second grid parameter set, its corresponding decision variable vector, total cable length, operation and maintenance cost, alignment reward, weight coefficient of total cable length, weight coefficient of operation and maintenance cost, and weight coefficient of alignment reward are input into the objective function to determine the target network parameter set.

[0127] This objective function is applicable when, in addition to annual power generation, cost and engineering feasibility are also considered when selecting wind turbine sites.

[0128] As a third optional implementation, the objective function is: ,in, It is a set of mesh parameters. This is a vector of decision variables. AEP is the annual power generation, and z is the decision variable for cable laying. If z is 0, the cable will not be laid; if z is 1, the cable will be laid. It is the preset cable power flow. This is a decision variable for the construction of a booster station. If it is 0, a booster station will not be built at that point; if it is 1, a booster station will be built at that point. This is the allocation variable from the wind turbine to the substation, indicating whether the i-th wind turbine is connected to the s-th substation. It is a weighting factor for the total cable length / cost. It is the total cable length / cost. It is a weighting coefficient for the cost of operation and maintenance. It's the cost of maintenance and upkeep. It is the location weight of the booster station. It is the construction cost of the booster station. It is the weighting coefficient of the alignment reward. It is an alignment reward; Optionally, based on multiple target intersections in the target network corresponding to each second grid parameter set, the total cable length / cost, operation and maintenance travel cost, substation construction cost, and alignment bonus are determined.

[0129] Optionally, multiple target intersections in the target network corresponding to each second grid parameter set are converted into decision variable vectors. Each second grid parameter set, its corresponding decision variable vector, total cable length / cost, maintenance travel cost, substation location parameters, alignment reward, weight coefficients of total cable length / cost, weight coefficients of maintenance travel cost, weight coefficients of substation location parameters, and weight coefficients of alignment reward are input into the objective function to determine the target network parameter set.

[0130] Specifically, the method for determining the total cable length / cost is as follows: First, consider the wind turbine and substation as a set of nodes V in graph G=(V,E), and the cables to be laid as a set of edges E. Node V contains all target intersections and the coordinates of the substation. Edge E needs to be determined based on the electrical connection logic. For each edge, the length of a single cable segment... Let i be the straight-line distance between fan i and fan j, specifically, Switch variables This indicates whether the cable section should be laid. If so, then... The value is 1. Based on this, the cable length can be obtained using formula (12). The cost of this cable segment is... ,in, This is the preset material cost per unit length of cable. This refers to the installation cost per unit length of cable. The maximum allowed collection capacity for a single loop. The voltage drop is estimated based on line impedance and power flow. Based on this, a constraint is imposed: the wind turbine at the target intersection point is to be installed. And ensure power conservation: And the capacity is limited to The total cable length / cost is obtained and used as a penalty term in the objective function.

[0131] As an optional implementation, the method for determining the cable routing scheme in the second stage of the two-stage site selection can be as follows: First, obtain the preliminary routing layout that connects all wind turbines and substations with the shortest total cable length through minimum spanning tree. Then, check whether the transmission flow of each cable exceeds its maximum capacity. If there is overload, in simple scenarios, the overload edge is split into parallel loops or branch structures. In complex scenarios, a small-scale mixed integer linear programming is used to solve the problem precisely, and finally, a cable routing scheme that satisfies the capacity constraint and has the optimal total length / cost is achieved.

[0132] The purpose of determining the operation and maintenance travel cost is to ensure that necessary navigation channels are maintained during the wind turbine deployment process. Configurable parameters include: B operation and maintenance base points, and average operation and maintenance speed. Maximum allowed single round trip time Tmax or mileage Lmax, and set of navigation channels. This refers to a polygonal passageway, within which an unobstructed zone width must be maintained. .

[0133] And set constraints: for each fan i, or For each channel The minimum orthogonal distance between the fan and the centerline of the channel is maintained. Alternatively, a soft restricted area can be set up within the target shape of the channel, and no wind turbines can be placed within the soft restricted area, thus obtaining the operation and maintenance travel cost. .

[0134] As an optional implementation, in the second stage of the two-stage site selection process, if there is a conflict between the constraints of neighborhood operations and operation and maintenance costs, the neighborhood intersections involved will not be used as the target intersections.

[0135] The purpose of determining the location parameters of the booster station is to optimize the location and number of booster stations, thereby shortening cables, reducing losses, and meeting capacity requirements.

[0136] The method for determining the location parameters of the booster stations is as follows: based on a pre-determined set of booster station locations, establish relationship variables between the wind turbines at each target intersection point and each booster station. That is, wind turbine i establishes a connection with station s and constrains it. And power is allocated to each station This allows us to determine the location parameters of the booster station.

[0137] As an alternative implementation, in the first stage of the two-stage site selection process, after determining the second grid parameter set, the initial values ​​of candidate sites are obtained by clustering within the sea area using k-medoids or k-means algorithms. In the second stage, the location / number of booster stations and the cable length / cost are alternately optimized.

[0138] As an alternative implementation, during the two-stage site selection process, the grid can be aligned with the channel direction to maintain a minimum angle, or placement can be prohibited within the channel buffer zone. Specifically, a unit vector is set for the channel centerline direction. Buffer half-width and the first candidate basis vector in each second grid parameter set. Second candidate basis vector And set the included angle limit: The included angle restriction can also be: Furthermore, the buffer zone is geometrically expanded into a polygon. As a buffer zone for waterways. Based on the above constraints, when determining the second set of grid parameters in the first stage, grid parameter sets that do not meet the constraints are directly filtered out. Thus, in the second stage, the second set of grid parameters is filtered out based on the constraints.

[0139] Based on the same inventive concept, this application also provides a wind farm layout processing device corresponding to the wind farm layout processing method. Since the principle of the device in this application is similar to the wind farm layout processing method described above, the implementation of the device can refer to the implementation of the method, and the repeated parts will not be described again.

[0140] Reference Figure 6 As shown, this is a wind farm layout processing device provided in an embodiment of this application: The first determining module 601 is used to determine the initial point sequence corresponding to each grid parameter set based on multiple grid parameter sets in the wind farm data. Each grid parameter set is used to characterize a target network with a target shape. The target network includes multiple grid intersections. The initial point sequence includes multiple initial points ordered in order of wake loss. Each initial point is a grid intersection among the multiple grid intersections. The second determining module 602 is used to determine multiple first grid parameter sets from the multiple grid parameter sets based on a preset range of base vector lengths and a preset range of candidate angles, and to determine the estimated annual power generation of each first grid parameter set based on a preset rapid power generation assessment model, and to determine multiple second grid parameter sets from the multiple first grid parameter sets based on the estimated annual power generation of each first grid parameter set. The third determining module 603 is used to select multiple candidate intersection points from the initial point sequence corresponding to each of the second grid parameter sets, and determine multiple target intersection points in the target network corresponding to each of the second grid parameter sets based on the neighborhood operation results of each candidate intersection point; The arrangement module 604 is used to determine the target grid parameter set in the multiple second grid parameter sets based on a preset objective function and multiple target intersections in the target network corresponding to each second grid parameter set, and to arrange the wind farm based on the multiple target intersections of the target grid parameter set.

[0141] The processing flow of each module in the device and the interaction flow between each module can be referred to the relevant descriptions in the above method embodiments, and will not be detailed here.

[0142] This application also provides an electronic device, such as... Figure 7 The diagram shown is a structural schematic of an electronic device provided in an embodiment of this application, including: a processor 701, a memory 702, and a bus. The memory 702 stores machine-readable instructions executable by the processor 701 (e.g., ...). Figure 6 The device includes the execution instructions corresponding to the first determining module 601, the second determining module 602, the third determining module 603, and the arrangement module 604. When the computer device is running, the processor 701 and the memory 702 communicate via a bus. When the machine-readable instructions are executed by the processor 701, the above-mentioned wind farm arrangement processing method is performed.

[0143] This application also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, performs the steps of the wind farm layout processing method described above.

[0144] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems and devices described above can be referred to the corresponding processes in the method embodiments, and will not be repeated here. In the several embodiments provided in this application, it should be understood that the disclosed systems, devices, and methods can be implemented in other ways. The device embodiments described above are merely illustrative. For example, the division of modules is only a logical functional division, and in actual implementation, there may be other division methods. Furthermore, multiple modules or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed mutual coupling or direct coupling or communication connection can be through some communication interfaces; the indirect coupling or communication connection of devices or modules can be electrical, mechanical, or other forms.

[0145] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. If the functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes: USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, optical disks, and other media capable of storing program code.

[0146] The above are merely specific embodiments of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

Claims

1. A method for wind farm layout and processing, characterized in that, The method includes: Based on multiple grid parameter sets in the wind farm data, an initial point sequence corresponding to each grid parameter set is determined. Each grid parameter set is used to characterize a target network with a target shape. The target network includes multiple grid intersections. The initial point sequence includes multiple initial points ordered in order of wake loss. Each initial point is one of the multiple grid intersections. Based on a preset range of base vector lengths and a preset range of candidate angles, multiple first grid parameter sets are determined from the multiple grid parameter sets. Based on a preset rapid power generation assessment model, the estimated annual power generation of each first grid parameter set is determined. Based on the estimated annual power generation of each first grid parameter set, multiple second grid parameter sets are determined from the multiple first grid parameter sets. Multiple candidate intersection points are selected from the initial point sequence corresponding to each of the second grid parameter sets, and multiple target intersection points in the target network corresponding to each of the second grid parameter sets are determined based on the neighborhood operation results of each candidate intersection point. Based on a preset objective function, and according to multiple target intersections in the target network corresponding to each second grid parameter set, a target grid parameter set is determined from the multiple second grid parameter sets, and the wind farm is arranged based on the multiple target intersections of the target grid parameter set.

2. The wind farm layout processing method according to claim 1, characterized in that, The step of determining the initial point sequence corresponding to each grid parameter set based on multiple grid parameter sets in the wind farm data includes: Based on the set of grid parameters and the preset row and column index values, determine the coordinates of multiple grid intersection points corresponding to the set of grid parameters; Based on the critical range in the wind farm data and the coordinates of multiple grid intersections corresponding to the grid parameter set, multiple intermediate grid intersections among multiple grid nodes are determined, wherein the critical range includes at least one of the following: boundary range, no-layout range, and minimum safe distance; Determine the wake loss at each intermediate grid intersection point, and determine multiple initial points based on the wake loss at each intermediate grid intersection point. Then, add each initial point to the initial point sequence in sequence according to the wake loss.

3. The wind farm layout processing method according to claim 1, characterized in that, Each set of mesh parameters includes: a first candidate basis vector, a second candidate basis vector, the length of the first candidate basis vector, the length of the second candidate basis vector, a first candidate angle, and a second candidate angle; The determination of multiple first grid parameter sets from the multiple grid parameter sets based on preset basis vector length ranges and preset candidate angle ranges includes: Traverse the multiple sets of mesh parameters. For the current set of mesh parameters, if the lengths of the first candidate basis vector and the second candidate basis vector in the current set of mesh parameters are both within the range of the basis vector lengths, and the first candidate angle and the second candidate angle in the current set of mesh parameters are both within the range of the candidate angles, then the current set of mesh parameters is taken as a first set of mesh parameters.

4. The wind farm layout processing method according to claim 1, characterized in that, The step of determining multiple sets of second grid parameters from the multiple sets of first grid parameters based on the estimated annual power generation of each of the first grid parameters includes: According to the estimated annual power generation of each first grid parameter set, the first grid parameter sets are sorted to obtain a sequence of first grid parameter sets; The first preset number of first grid parameter sets in the first grid parameter set sequence are used as the second grid parameter set.

5. The wind farm layout processing method according to claim 1, characterized in that, Based on the greedy placement algorithm, a preset number of candidate points are sequentially selected from the initial point sequence corresponding to each of the second grid parameter sets, including: Determine the initial current candidate intersection point from the initial point sequence; Traverse each initial point in the initial point sequence. For the current initial point, determine the boundary lift of the current initial point based on the wind power data of the current initial point and the wind power data of the current candidate intersection point. Based on the boundary lift of the current initial point, determine whether to use the current initial point as a candidate intersection point. If so, use the current initial point as a new current candidate intersection point.

6. The wind farm layout processing method according to claim 5, characterized in that, The step of determining the boundary lift of the current initial point based on the wind power data of the current initial point and the wind power data of the current candidate intersection point includes: Based on the wind power data of the current candidate intersection point, calculate the annual power generation of the current candidate intersection point, and based on the wind power data of the current initial point, calculate the annual power generation of the current initial point. The difference between the annual power generation of the current initial point and the annual power generation of the current candidate intersection point is taken as the annual power generation difference value. Calculate the cost of setting up the current candidate intersection points; The quotient of the annual power generation difference of the current candidate intersection point and the cost value is used as the boundary improvement amount of the current candidate intersection point.

7. The wind farm layout processing method according to claim 1, characterized in that, The step of determining multiple target intersection points in the target network corresponding to each second grid parameter set based on the neighborhood operation results of each candidate intersection point includes: Perform at least one neighborhood operation on the candidate intersection point to obtain the operation-triggered intersection point corresponding to the candidate intersection point. The neighborhood operation includes: point swapping operation, micro-shifting operation, or small-range rotation operation. If the total annual power generation corresponding to the intersection point after the operation is greater than the total annual power generation corresponding to the candidate intersection point, then the intersection point after the operation will be used as a target intersection point. If the total annual power generation of the intersection points obtained after performing all neighborhood operations on the candidate intersection points is less than the total annual power generation of the candidate intersection points, then the candidate intersection points will be selected as the target intersection points.

8. The wind farm layout processing method according to claim 1, characterized in that, The objective function is: AEP is the annual power generation. It is a set of mesh parameters. It is a vector of decision variables; The step of determining the target grid parameter set from the multiple second grid parameter sets based on the preset objective function and according to multiple target intersections in the target network corresponding to each second grid parameter set includes: Multiple target intersections in the target network corresponding to each of the second grid parameter sets are converted into decision variable vectors, and each of the second grid parameter sets and its corresponding decision variable vectors are input into the objective function to determine the target network parameter set.

9. The wind farm layout processing method according to claim 1, characterized in that, The objective function is: AEP is the annual power generation. It is a weighting factor for the total cable length. It is the total length of the cable. It is the weighting coefficient of operation and maintenance costs. It's the cost of operation and maintenance. It is the weighting coefficient of the alignment reward. It is an alignment reward; The step of determining the target grid parameter set from the multiple second grid parameter sets based on the preset objective function and according to multiple target intersections in the target network corresponding to each second grid parameter set includes: Based on multiple target intersections in the target network corresponding to each of the second grid parameter sets, the total cable length, maintenance cost, and alignment bonus are determined. Multiple target intersections in the target network corresponding to each second grid parameter set are converted into decision variable vectors. Each second grid parameter set, its corresponding decision variable vector, the total cable length, the operation and maintenance cost, the alignment reward, the weight coefficient of the total cable length, the weight coefficient of the operation and maintenance cost, and the weight coefficient of the alignment reward are input into the objective function to determine the target network parameter set.

10. An electronic device, characterized in that, include: The processor and memory, the memory storing machine-readable instructions executable by the processor, wherein when the electronic device is running, the processor executes the machine-readable instructions to perform the steps of the wind farm layout processing method as described in any one of claims 1 to 9.

Images