Offshore wind farm multi-objective collaborative planning design method, device, equipment and medium

Abstract

The present application belongs to the field of offshore wind farm planning and design, and discloses a kind of offshore wind farm multi-objective collaborative planning and design method, device, equipment and medium, method includes: using orthogonal experiment method to optimize offshore wind farm micro-siting, and combining wake loss model and square and superposition model to calculate the annual power generation of wind farm;Improved elbow method and K-means++ algorithm are used to optimize the layout of offshore wind farm booster station;Improved Prim algorithm is used to calculate the economic evaluation model of the cost of power collection system as the edge weight, and the design of offshore wind farm power collection system is optimized;With the maximum annual power generation of offshore wind farm and the minimum cost of power collection system as the target, NSGA-II genetic algorithm is used to collaboratively optimize offshore wind farm micro-siting, booster station layout and power collection system design, to generate the optimal wind turbine arrangement, booster station layout and power collection line topology.The present application takes into account power generation efficiency and economy, and improves the overall benefit of offshore wind farm.

Description

Technical Field

[0001] This invention belongs to the field of offshore wind farm planning and design technology, and particularly relates to a multi-objective collaborative planning and design method, device, equipment and medium for offshore wind farms. Background Technology

[0002] Wind power is a renewable and clean energy source. Offshore wind resources are more stable and have higher wind speeds than onshore winds, making it easier to scale up and less constrained by environmental issues. Its planning and design mainly includes three key aspects: micro-site selection, substation layout, and power collection system design.

[0003] In traditional design, these steps are often performed independently and sequentially. Micro-site selection aims to maximize power generation, typically increasing turbine spacing to reduce wake, but this can lead to increased collector line length and cost. Substation locations are often determined empirically, without coordination with wind turbine layout. Collector system optimization usually involves cable path planning within a given layout, making it difficult to provide feedback on the impact on wind turbine arrangement. This step-by-step, independent design model results in poor overall economics; the increased power generation may not offset the increased cable costs, making it difficult to balance the trade-off between power generation and cost, especially in large-scale offshore wind farms. Furthermore, existing research often ignores wake interference between wind farms, further affecting the accuracy of power generation prediction. At the algorithmic level, existing wake models (such as Jensen and Larsen models) have limited accuracy under complex wind conditions; substation layout planning lacks automated methods for quantity and location optimization; and collector line optimization often focuses on minimizing length without comprehensively considering line losses and the economics of cable selection. Summary of the Invention

[0004] The purpose of this invention is to overcome the shortcomings of the prior art and provide a multi-objective collaborative planning and design method, device, equipment and medium for offshore wind farms, which takes into account both power generation efficiency and economy, and improves the overall benefits of offshore wind farms.

[0005] To solve the above-mentioned technical problems, the present invention is implemented using the following solution:

[0006] This invention provides a multi-objective collaborative planning and design method for offshore wind farms, comprising:

[0007] The orthogonal experimental design was used to optimize the micro-site selection of offshore wind farms, and the annual power generation of the wind farms was calculated by combining the wake loss model and the sum of squares superposition model.

[0008] The layout of offshore wind farm booster stations was optimized using an improved elbow method and the K-means++ algorithm (the improved K-means algorithm is a clustering algorithm).

[0009] An improved Prim algorithm is used to optimize the design of offshore wind farm power collection systems by employing an economic evaluation model for calculating the cost of power collection systems as edge weights.

[0010] With the goal of maximizing the annual power generation of offshore wind farms and minimizing the cost of the power collection system, the NSGA-II genetic algorithm (non-dominated sorting genetic algorithm II) is used to collaboratively optimize the micro-site selection, substation layout, and power collection system design of offshore wind farms, generating the optimal wind turbine layout, substation layout, and power collection line topology.

[0011] Furthermore, the orthogonal experimental design method was used to optimize the micro-site selection of offshore wind farms, and the annual power generation of the wind farms was calculated by combining the wake loss model and the sum of squares superposition model, including:

[0012] Set the range of values ​​for the five parameters of the minimum unit: α, d1 / D, d2 / D, Δd1 / D, and Δd2 / D;

[0013] The orthogonal experimental method was applied to iteratively optimize the five parameters of the smallest unit to obtain a set of optional machine sites.

[0014] A set of generating regular layout points is selected from the set of optional site locations, and the annual power generation of the wind farm under each layout scheme is calculated by combining the wake loss model and the sum of squares superposition model.

[0015] Where α is the azimuth angle of the longer side in the micro-site selection direction of the wind farm, d1 is the row spacing of the wind turbines (greater than 3D), d2 is the column spacing of the wind turbines (greater than 7D), D is the rotor diameter of the wind turbine, Δd1 and Δd2 are the equal intervals of the row spacing and column spacing of the wind turbines, respectively, and the wake loss model is:

[0016]

[0017]

[0018]

[0019]

[0020] In the formula, When only considering the impact of the i-th upstream wind turbine on the target wind turbine, the downstream... The wake wind speed at the location, The straight-line distance along the wind direction. It is the inflow wind speed. It is the standard deviation, I α It is environmental turbulence, I wake It is mechanical turbulence, Q is synthetic turbulence, C T is the thrust coefficient, r is the radial distance of the calculation point from the centerline of the wake, and k1, k2, and k3 are empirical constants.

[0021] The sum of squares superposition model is:

[0022]

[0023] In the formula, Let i be the wake wind speed of the target wind turbine affected by n upstream wind turbines, where n represents the number of upstream wind turbines affecting the target wind turbine, and i represents the sequence number.

[0024] The formula for calculating the annual power generation of a wind farm is:

[0025]

[0026]

[0027]

[0028] In the formula, p represents the annual power generation of the wind farm. j Let N be the average output power of wind turbine j, N be the number of wind turbines in the wind farm, ω(θ) be the wind direction frequency, and v be the average output power of wind turbine j. cut-out To cut off the wind speed, v cut-in To determine the cutoff wind speed, k(θ) is the shape factor within the wind direction sector θ, and c(θ) is the scale factor within the wind direction sector θ, where θ is the wind direction. The formula for calculating the Weibull distribution is as follows: This is the power curve.

[0029] Furthermore, the layout of offshore wind farm booster stations is optimized using an improved elbow method and the K-means++ algorithm, including:

[0030] An improved elbow method was used to determine the optimal number of booster stations, and the K-means++ algorithm was applied to optimize the location of the booster stations, including:

[0031] Assuming the number of booster stations K equals 1, calculate and determine whether the connection distance from the booster station to the farthest wind turbine is greater than the set distance. If not, the number of booster stations K equals 1. If so, use the improved elbow method to obtain the optimal number of booster stations K.

[0032] The K-means++ algorithm is used to obtain the specific locations of K booster stations, resulting in the location division of the booster stations.

[0033] Furthermore, the optimal number of booster stations K is obtained using the improved elbow method, including:

[0034] Set the range of values ​​for K;

[0035] Calculate the sum of squared errors (SSE) for each K value;

[0036] The starting and ending points are determined based on the range of K and the SSE value, and the internal calculation is performed by equal division.

[0037] Subtract the average value corresponding to each K value from the SSE value;

[0038] The K value corresponding to the maximum difference between the average calculated value of K and the SSE value is taken as the number of booster stations K.

[0039] Furthermore, the K-means++ algorithm includes:

[0040] Step a: Randomly select a data point from the dataset as the first cluster center;

[0041] Step b: Calculate the distance D(X) from each data point to the nearest cluster center and store it in an array. Then, sum the distances stored in the array into Sum(D(X)).

[0042] Step c: Take a random value Random that falls within Sum(D(X)), and use Random = D(X) until Random <= 0 to obtain the next cluster center;

[0043] Step d: Repeat steps b to c until K cluster centers are selected;

[0044] Step e: Run the standard K-means algorithm using K cluster centers.

[0045] Furthermore, an improved Prim algorithm is used with edge weights to optimize the design of offshore wind farm power collection systems, including:

[0046] Using an economic evaluation model that calculates the construction cost and line loss cost of a collector line as edge weights, an improved Prim algorithm is employed to generate the collector line topology, including:

[0047] Step A: Import wind farm data: wind farm turbine location coordinates, cable database, substation location, number of incoming lines to the substation, and number of wind turbines that can be connected to the substation;

[0048] Step B: Assume that the set of turbine locations in the wind farm is H, the set of turbine locations connected to the booster station is H1, the set of turbine locations not connected to the booster station is H2, the set of turbine locations with generated edges is K1, and the set of turbine locations without generated edges is K2.

[0049] Step C: Assuming the number of incoming lines to the substation is M, sort the weights of the edges associated with the substation in ascending order, select the M edges with the smallest weights and connect them, put the generated M edges into K1, put the connected machine points into H1, and delete them from H2.

[0050] Step D: Add wind turbines to the M incoming line sets to connect the submarine cables, based on the principle of adjacent lines with the lowest weight;

[0051] Step E: Calculate the cost of the collector line based on the economic evaluation model;

[0052] Step F: Repeat steps D to E until each incoming line reaches its maximum capacity and no new non-intersecting edges can be added, thus obtaining the collector line topology.

[0053] Among them, economic evaluation model C all for:

[0054]

[0055]

[0056]

[0057] In the formula, C1 is the construction cost of the collector line, C2 is the line loss cost of the collector line, and N g L represents the number of wind turbines in the g-th collector line. gl Let T be the length of the l-th segment of the submarine cable in the g-circuit collector line. gl Let F be the unit cost of constructing the l-th segment of the g-circuit collector line, F be the number of branches of the collector line, R be the resistance of the collector line, and I be the unit cost of constructing the l-th segment of the submarine cable. n It is the operating current in the circuit, p price This refers to the on-grid electricity price.

[0058] Furthermore, aiming to maximize the annual power generation of offshore wind farms and minimize the cost of the power collection system, the NSGA-II genetic algorithm is used to collaboratively optimize the micro-site selection, substation layout, and power collection system design of offshore wind farms, generating optimal wind turbine arrangement, substation layout, and power collection line topology, including:

[0059] With the goals of maximizing the annual net power generation of the entire wind farm and minimizing the cost of the power collection system, the NSGA-II genetic algorithm is used to collaboratively optimize the micro-site selection, substation layout, and power collection system design of the offshore wind farm, outputting a Pareto solution set. From this Pareto solution set, optimized schemes that meet the engineering requirements are selected to generate the optimal wind turbine layout, substation layout, and power collection line topology, including:

[0060] Step G: Randomly generate an initial population of the size of the number of wind turbines, and take the maximum annual power generation of the wind farm and the minimum cost of the power collection system as the objective function;

[0061] Step H: Perform non-dominated sorting and crowding calculation, and record the current iteration number G=1;

[0062] Step I: Obtain the first generation offspring population through selection, crossover, and mutation operations;

[0063] Step J: Starting from the second generation, merge the parent and offspring populations, perform a fast non-dominated sort, and calculate the crowding degree for individuals in each non-dominated layer.

[0064] Step K: Select individuals to form a new parent population, and use a genetic algorithm to generate a new offspring population;

[0065] Step L: Repeat steps I to K, and determine whether the constraints are met. If they are met, proceed to a new round of optimization iterations until the set number of iterations is reached. If they are not met, end the process and output the Pareto solution set.

[0066] Step M: Select an optimal solution that meets the engineering requirements from the Pareto solution set to generate the optimal wind turbine layout, substation layout, and power collection line topology;

[0067] The constraints are as follows: the number of wind turbines is fixed; the wind turbines are located within the planned wind farm; the distance between any two wind turbines is greater than 3D; the wind turbines are arranged in a regular pattern; and D is the diameter of the wind turbine rotor.

[0068] The present invention also provides a multi-objective collaborative planning and design device for offshore wind farms, comprising:

[0069] The micro-site selection optimization module is used to optimize the micro-site selection of offshore wind farms using the orthogonal experimental method, and to calculate the annual power generation of the wind farm by combining the wake loss model and the sum of squares superposition model.

[0070] The substation layout optimization module is used to optimize the layout of offshore wind farm substations using the improved elbow method and K-means++ algorithm.

[0071] The power collection system design optimization module is used to optimize the design of offshore wind farm power collection systems by using an improved Prim algorithm with an economic evaluation model for calculating power collection system costs as edge weights.

[0072] The collaborative optimization module aims to maximize the annual power generation of offshore wind farms and minimize the cost of the power collection system. It uses the NSGA-II genetic algorithm to collaboratively optimize the micro-site selection, substation layout, and power collection system design of offshore wind farms, generating the optimal wind turbine layout, substation layout, and power collection line topology.

[0073] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the steps of the aforementioned multi-objective collaborative planning and design method for offshore wind farms.

[0074] The present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the aforementioned multi-objective collaborative planning and design method for offshore wind farms.

[0075] Beneficial effects

[0076] This invention employs orthogonal experimental design to optimize the micro-site selection of offshore wind farms, uses an improved elbow method and K-means++ algorithm to optimize the layout of booster stations, and uses an improved Prim algorithm with an economic evaluation model for calculating the cost of the power collection system as edge weights to optimize the design of the power collection system. With the goal of maximizing the annual power generation of the offshore wind farm and minimizing the cost of the power collection system, the NSGA-II genetic algorithm is used for collaborative optimization to solve for the wind turbine layout, booster station layout, and power collection line topology, taking into account both power generation efficiency and economy, and improving the overall benefits of the offshore wind farm. Attached Figure Description

[0077] Figure 1 This is a flowchart of a multi-objective collaborative planning and design method for offshore wind farms provided in Embodiment 1 of the present invention.

[0078] Figure 2 This is a schematic diagram of the five parameters of the smallest unit of the orthogonal experimental method provided in Embodiment 1 of the present invention.

[0079] Figure 3 yes Figure 2 A magnified view of a portion of the image.

[0080] Figure 4 This is a flowchart of the micro-site selection optimization process for offshore wind farms based on the orthogonal experimental method provided in Embodiment 1 of the present invention.

[0081] Figure 5 This is a schematic diagram of the improved elbow method provided in Embodiment 1 of the present invention.

[0082] Figure 6 This is a flowchart of the improved elbow method algorithm provided in Embodiment 1 of the present invention.

[0083] Figure 7 This is a flowchart of the K-means++ algorithm provided in Embodiment 1 of the present invention.

[0084] Figure 8 This is a flowchart of the method for dividing multiple substation units in an offshore wind farm, provided in Embodiment 1 of the present invention.

[0085] Figure 9 This is a flowchart of the improved Prim algorithm for collector lines provided in Embodiment 1 of the present invention.

[0086] Figure 10 This is a flowchart of the NSGA-Ⅱ algorithm provided in Embodiment 1 of the present invention.

[0087] Figure 11 This refers to the wind farm range in the case provided in Embodiment 1 of the present invention.

[0088] Figure 12 This is the wind rose diagram of the wind farm in the case provided in Embodiment 1 of the present invention.

[0089] Figure 13 The example provided in Embodiment 1 of this invention shows the power and thrust curves of the wind turbine used in the wind-driven electric field.

[0090] Figure 14 The Pareto solution set is obtained by using the multi-objective collaborative planning and design method for offshore wind farms in the case provided in Embodiment 1 of the present invention.

[0091] Figure 15 This is the first optimized layout and collector line topology scheme obtained from the Pareto solution results provided in Embodiment 1 of the present invention.

[0092] Figure 16 This is the second optimized layout and collector line topology scheme obtained from the Pareto solution results provided in Embodiment 1 of the present invention.

[0093] Figure 17 This is the third optimized layout and collector line topology scheme obtained from the Pareto solution results provided in Embodiment 1 of the present invention. Detailed Implementation

[0094] The present invention will be further described below with reference to the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solution of the present invention, and should not be used to limit the scope of protection of the present invention.

[0095] Example 1

[0096] like Figure 1 As shown in the figure, this embodiment provides a multi-objective collaborative planning and design method for offshore wind farms, with the following steps:

[0097] Step 1: Obtain wind resource parameters (inflow wind speed) for offshore wind farms Wind direction θ and environmental turbulence I α Wind turbine model parameters (thrust coefficient C) T (Wind turbine diameter D), cable database (submarine cable construction cost unit price T) gl ) and wind farm constraints (wind farm boundary range and grid connection price p) price )wait.

[0098] Step 2: Optimize the micro-site selection of offshore wind farms using the orthogonal experimental method, and calculate the wake loss and annual power generation of the wind farm by combining the wake loss model and the sum of squares superposition model.

[0099] like Figure 4 As shown, the specific steps of the orthogonal experimental optimization of the micro-site selection of offshore wind farms, i.e., the layout of wind turbines, are as follows:

[0100] Step 1: As Figure 2 and Figure 3 As shown, the range of values ​​for the five parameters of the minimum unit is set: α, d1 / D, d2 / D, Δd1 / D, and Δd2 / D. The azimuth angle of the longer side in the wind farm layout direction is... Figure 3 The angle α between the longer side of the parallelogram and the x-coordinate ranges from [0° to 130°], with an optimal layout most easily achieved when α is 90°. The row spacing d1 of the wind turbines should be greater than 3D, and the column spacing d2 should be greater than 7D, where D is the diameter of the wind turbine rotor. Δd1 and Δd2 are the equal divisions of the row and column spacings of the wind turbines, respectively, and the number of values ​​is the number of equal divisions of the five parameters of the smallest unit within their respective ranges. To refine the wind turbine layout scheme, the number of values ​​can be set to even greater values.

[0101] Step 2: Parameter optimization using orthogonal experiment method: The five parameters of the smallest unit are used as optimization parameters, and the orthogonal experiment method is used for optimization. The specific parameter settings are shown in Table 1.

[0102] Table 1 shows the five parameter settings for the smallest unit.

[0103]

[0104] Step 3: Optimize to obtain a set of optional machine point locations, and select from them to generate a set of points for regular layout.

[0105] Step 4: Calculate the annual power generation of the current layout plan.

[0106] Step 5: Determine whether the traversal of the orthogonal experimental table (Table 1) has ended, and compare the annual power generation after considering the wake effect. Use the wake loss model and the sum of squares superposition model to calculate the wake loss and annual power generation. Otherwise, continue the traversal and rearrange the turbine positions to improve the utilization rate of the sea area of ​​the site.

[0107] The wake loss model is as follows:

[0108] (1)

[0109] (2)

[0110] In the formula,

[0111] (3)

[0112] (4)

[0113] In the formula, When only considering the impact of the i-th upstream wind turbine on the target wind turbine, the downstream... The wake wind speed at the location, The straight-line distance along the wind direction. It is the inflow wind speed. It is the standard deviation, I α It is environmental turbulence, I wake It is mechanical turbulence, Q is the composite turbulence, D is the rotor diameter, and C is the wind turbine diameter. T is the thrust coefficient, r is the radial distance of the calculation point from the centerline of the wake, and k1, k2, and k3 are empirical constants, with k1=0.27, k2=6.00, and k3=0.004 respectively.

[0114] The sum of squares superposition model is:

[0115] (5)

[0116] In the formula, Let i be the wake wind speed of the target wind turbine affected by n upstream wind turbines, where n represents the number of upstream wind turbines affecting the target wind turbine, and i represents the sequence number.

[0117] The annual power generation of a wind farm is calculated as follows:

[0118] The total average power of the wind farm is calculated using the probability density discretization method. Wind speed follows a Weibull distribution; within a sector θ of a specific wind direction, the inflow wind speed is... The probability density is:

[0119] (6)

[0120] In the formula, k(θ) is the shape factor within the wind direction sector θ, and c(θ) is the scale factor within the wind direction sector θ.

[0121] Therefore, the average output power of wind turbine i can be obtained as follows:

[0122] (7)

[0123] In the formula, ω(θ) is the wind direction frequency, v cut-out To cut off the wind speed, v cut-in To cut into wind speed, This is the power curve.

[0124] The total annual power generation of the wind farm is:

[0125] (8)

[0126] Where: p jdenoted as AEP, where MW is the average output power of wind turbine j, N is the number of wind turbines in the wind farm, 8760 is the annual power generation hours, which is the average annual hours over four years considering leap years, and GWh is the unit of AEP.

[0127] Step 3: Optimize the layout of offshore wind farm booster stations using the improved elbow method and K-means++ algorithm.

[0128] An improved elbow method was used to determine the optimal number of booster stations, and the K-means++ algorithm was applied to optimize the location of these stations. This included optimizing the partitioning of multi-booster turbine units in offshore wind farms based on the improved elbow method and K-means++ clustering algorithm. According to existing enterprise standards, wind farms need to establish multiple booster stations, with the connection distance from each booster station to the farthest wind turbine as a constraint, and 25km as the criterion. Figure 8 As shown, the specific algorithm steps are as follows:

[0129] Step 1: Assuming the number of booster stations K equals 1, calculate the Euclidean (connection) distance from the booster station to each wind turbine.

[0130] (11)

[0131] In the formula, P1 and P2 are two turbine locations within the wind farm; (x1, y1) and (x2, y2) are the horizontal and vertical coordinates of P1 and P2, respectively.

[0132] Step 2: Determine whether the connection distance from the booster station to the farthest wind turbine is less than 25km. If it is, the number of booster stations required for the wind farm is K=1, and proceed directly to Step 5; otherwise, proceed to Step 3.

[0133] Step 3: Reset the range of values ​​for the number of booster stations K. Assume the selected range is [A, B], where B > A ≧ 2.

[0134] Step 4: Use the improved elbow method to obtain the number of booster stations K.

[0135] Step 5: Use the K-means++ algorithm to obtain the specific locations of the K initial booster stations.

[0136] Step 6: Obtain the final substation unit allocation results for the wind farm.

[0137] The number of booster stations K was obtained using the improved elbow method, specifically as follows:

[0138] The improved elbow method determines the optimal K value through linearly dividing the difference, such as... Figure 5 and Figure 6 As shown, it includes the following steps:

[0139] Step 1: Define the range of values ​​for K. Assuming the selected range is [A, B], calculate the sum of squared errors (SSE) for each K value. The corresponding SSE value is then SSE. A and SSE B .

[0140] Step 2: Using (A, SSE) A ) and (B, SSE B The starting and ending points are , and the internal calculation is performed equally.

[0141] (9)

[0142] In the formula, k z Let K be the value of M. z For k z The corresponding average value is calculated, and z = 1, 2, 3, ..., BA.

[0143] Step 3: M z With k z The corresponding sum of squared errors (SSE) z Let the difference be D. z .

[0144] (10)

[0145] In the formula, z=1,2,3,...,BA.

[0146] Step 4: Select D z The maximum value D in max The corresponding k max Then k max That is the optimal K value.

[0147] And by using the K-means++ algorithm for optimization, such as Figure 7 As shown, the specific steps are as follows:

[0148] Step 1: Randomly select a data point from the dataset as the first cluster center.

[0149] Step 2: Calculate the distance D(X) between any data point X and the nearest cluster center, and store it in an array. Then add these distances together to get Sum(D(X)), X=1, 2, 3, 4...

[0150] Step 3: Select a new data point as the new cluster center. The selection principle is: the point with a larger D(X) has a higher probability of being selected as the cluster center. First, take a random value Random that can fall into Sum(D(X)), and then use Random=D(X) until Random<=0. The point at this point is the next cluster center.

[0151] Step 4: Repeat Step 2 and Step 3 until K cluster centers are selected.

[0152] Step 5: Use these K initial cluster centers to run the standard K-means algorithm.

[0153] Step 4: Use the improved Prim algorithm with the economic evaluation model for calculating the cost of the power collection system as the edge weight to optimize the design of the power collection system for offshore wind farms.

[0154] An improved Prim algorithm is used to generate the topology of the power collection line by using an economic evaluation model for calculating the cost of the power collection system as the edge weight.

[0155] The economic evaluation model includes the construction cost and line loss cost of the collector line. Specifically, the calculation model for the collector line construction cost is as follows: The collector cable structure is complex. Once the location of the substation is determined, the construction cost of the collector cable depends on the topology, determined by the length and unit cost of its corresponding branch cables. The formula for calculating the collector line construction cost is:

[0156] (12)

[0157] In the formula, M is the number of incoming lines, and N is the number of incoming lines. g L represents the number of wind turbines in the g-th collector line. gl Let T be the length of the l-th segment of the submarine cable in the g-circuit collector line. gl The unit cost of the l-th section of the g-circuit collector line (including the cost of the submarine cable, transportation and construction expenses).

[0158] Line Loss Cost Calculation Model: During wind farm operation, the electrical energy generated by the wind turbines is transmitted to the substation via collector lines, inevitably resulting in energy losses in these lines. If the energy losses during operation are factored back into the construction cost, the concept of line loss cost is formed. In practice, the calculation of collector line loss cost often uses the rated operating current, without considering the actual operating current of the collector lines, leading to a significant deviation between the actual and actual line loss cost. To obtain a reliable line loss cost, wind resource calculation is introduced, considering the wake loss of the wind farm, thereby obtaining a more reliable wind turbine power generation, which is then used to calculate the instantaneous line loss. The formula for calculating line loss is:

[0159] (13)

[0160] In the formula, F is the number of collector line branches, R is the collector line resistance, and I... n This is the operating current on the collector line, which can be calculated from the power flow of the wind farm's power system; p price The on-grid electricity price is set at 0.25 yuan / kWh. Compared with design methods that do not consider the cost of collector line losses, turbine locations with good wind resources that consider line losses have higher power generation, and the corresponding collector lines have larger operating currents, making them more likely to be directly connected to the busbar to reduce line losses.

[0161] Economic calculation model for collector lines: Combining the construction cost calculation model and the line loss cost calculation model for collector lines, the economic calculation model C for collector lines can be obtained. all for:

[0162] (14)

[0163] In the formula, C1 is the construction cost of the collector line, and C2 is the line loss cost of the collector line.

[0164] An improved Prim algorithm uses the economic efficiency of the power collection line as the edge weight and constrains the maximum number of wind turbines and cable capacity for each incoming line to avoid line crossings. The edge weights in the Prim algorithm are set to consider various factors such as the selection and cost of the power collection line, and an improved algorithm that dynamically adjusts and iteratively optimizes is proposed, such as... Figure 9 As shown, the specific steps are as follows:

[0165] (1) Import wind farm data: including the coordinates of the wind farm turbine locations, cable database, location of the booster station, number of incoming lines to the booster station and the number of wind turbines that can be connected.

[0166] (2) Calculation initialization: Assume that the set of turbine locations in the wind farm is H, the set of turbine locations connected to the booster station is H1, the set of turbine locations not connected to the booster station is H2, the set of turbine locations with generated edges is K1, and the set of turbine locations without generated edges is K2.

[0167] (3) Set the number of incoming lines to the booster station and determine the wind turbines connected to the booster station: Assuming that the number of incoming lines to the booster station is M, sort the weights of the edges associated with the booster station in ascending order, select the M edges with the smallest weights in sequence for connection, put the generated M edges into K1 in sequence, put the connected turbine points into H1 in sequence, and delete them from H2.

[0168] (4) Add wind turbines to the M incoming line sets according to the principle of adjacent and minimum weight: In order to ensure that the number of wind turbines carried by the M incoming line sets is close to the same when the improved Prim algorithm is used for optimization, the machine point in the M incoming line sets is taken as the starting point and the adjacent edge with the minimum weight is found in turn.

[0169] (5) Calculate the cost of the collection line: Taking into account the construction cost and line loss cost of the collection line, calculate the cost of the collection line using the economic calculation model of the collection line.

[0170] (6) Determine whether each incoming line has reached its maximum capacity: During the optimization process, it is necessary to check the number of wind turbines in the set of M incoming lines. If the number of wind turbines has reached the maximum value, then in subsequent optimization, the incoming line set is prohibited from connecting to new edges, so as to meet the limit of the maximum number of wind turbines that the collection line can carry.

[0171] (7) Repeat (4) to (6) until no new edges that do not cross can be added. At this point, the topology of the collector line is formed.

[0172] Step 5: With the goal of maximizing the annual power generation of the entire field and minimizing the cost of the power collection system, the NSGA-II genetic algorithm is used to collaboratively optimize the micro-site selection, substation layout and power collection system design of the offshore wind farm, generating the optimal wind turbine layout, substation layout and power collection line topology.

[0173] like Figure 10 As shown, the collaborative optimization steps are as follows:

[0174] Step 1: Randomly generate an initial population of size W, where W is the number of wind turbines. Calculate the objective function value according to the formula (8) for calculating the annual power generation of the wind farm and the formula (14) for the cost of the wind farm's collection lines.

[0175] Step 2: Perform non-dominated sorting and crowding calculation, and record the current iteration number G=1;

[0176] Step 3: Obtain the first generation offspring population through basic operations of the genetic algorithm (selection, crossover, and selection);

[0177] Step 4: Starting from the second generation, merge the parent and offspring populations, perform fast non-dominated sorting, and calculate the crowding degree of individuals in each non-dominated layer.

[0178] Step 5: Select suitable individuals to form a new parent population;

[0179] Step 6: Use a genetic algorithm to generate a new offspring population, determine whether the constraints are met, if they are met, proceed to a new round of optimization iterations until the limit number of iterations is reached, if they are not met, end the calculation and output the Pareto solution set;

[0180] The constraints include: during the optimization process, the number of wind turbines is fixed, and all wind turbines must be located within the planned wind farm; the distance between any two wind turbines is greater than 3D; and during the optimization process, the wind turbines in the offshore wind farm must be arranged in a regular pattern.

[0181] Step 7: Select the optimal solution that meets the engineering requirements from the Pareto solution set to generate the optimal wind turbine layout, substation layout and power collection line topology.

[0182] The model is then validated using actual data from an offshore wind farm in Jiangsu Province:

[0183] Taking an offshore wind power project in Jiangsu Province as an example, the wind farm measures 6 km × 3.7 km and requires 37 wind turbines of 3 MW each. The wind farm area is as follows: Figure 11 As shown. The wind rose diagram of this wind farm is as follows. Figure 12 As shown in the figure. The selected wind turbine has a hub height of 90 m and a rotor diameter of 110 m. Its power and thrust curves are as follows. Figure 13 As shown.

[0184] (2) Optimization results based on the multi-objective optimization algorithm of NSGA-II

[0185] Since this paper proposes two objective functions—annual power generation and transmission line construction cost—it is difficult to select the optimal solution from the optimization results. Therefore, this paper selects the aforementioned wind farm as the target wind farm and performs optimization using the NSGA-II multi-objective optimization algorithm. Three optimized results are then selected for comparative analysis. Figure 14 This is the Pareto solution set obtained after this optimization.

[0186] Figure 13 This reflects the essential characteristic of multi-objective optimization: the improvement of one optimization objective is accompanied by the deterioration of another. Based on this Pareto solution set, in wind farm development and construction, the most suitable result for the wind farm can be selected from the solution set according to future needs and constraints.

[0187] The results of the multi-objective optimization calculation are shown in Table 2.

[0188] Table 2 shows the results of the multi-objective optimization calculation.

[0189]

[0190] Figure 15 , Figure 16 and Figure 17To optimize the results, three optimized layout and power collection line topology schemes were obtained. In the figure, green dots represent wind turbine units, purple dots represent substations, and different colored line segments represent different cable types: red line segments represent HYJQF41-26 / 35 / mm2 (3×240) cables, blue line segments represent HYJQF41-26 / 35 / mm2 (3×185) cables, green line segments represent HYJQF41-26 / 35 / mm2 (3×120) cables, and black line segments represent HYJQF41-26 / 35 / mm2 (3×70) cables.

[0191] The optimization results of the three wind farm layouts show that the wind turbines are arranged in a regular pattern and evenly distributed within the wind farm area, without dense or sparse arrangements, thus improving the utilization efficiency of the wind farm. Combined with the wind rose diagram, it can be seen that the row and column spacing of these three wind turbine layouts satisfies the following conditions: smaller spacing indicates the direction of arrangement is basically parallel to the incoming flow direction, while larger spacing indicates the direction of arrangement is basically perpendicular to the prevailing wind direction, which meets the design requirements of actual wind farms and industry standards. Similarly, comparing the collector line topologies of these three target wind farms, it can be seen that after optimization using the K-means++ algorithm, the location of the booster station is roughly at the geometric center of the wind farm, and the three collector line topologies obtained using the improved Prim algorithm are all star-chain structures, which are highly consistent with the actual engineering of offshore wind farms.

[0192] As shown in Table 2, Group A is the result selected based on maximizing the annual power generation of the target wind farm; Group B is the optimized result obtained by minimizing the total cost of the wind farm's collector lines when the annual power generation of the wind farm is set at 228.59 GWh; Group C is a compromise non-dominated solution without special requirements. Furthermore, the table also provides the wake loss and collector line length corresponding to these three cases, which can facilitate analysis and comparison by designers.

[0193] Example 2

[0194] Based on the multi-objective collaborative planning and design method for offshore wind farms described in Example 1, this example provides a multi-objective collaborative planning and design device for offshore wind farms, including:

[0195] The micro-site selection optimization module is used to optimize the micro-site selection of offshore wind farms using the orthogonal experimental method, and to calculate the annual power generation of the wind farm by combining the wake loss model and the sum of squares superposition model.

[0196] The substation layout optimization module is used to optimize the layout of offshore wind farm substations using the improved elbow method and K-means++ algorithm.

[0197] The power collection system design optimization module is used to optimize the design of offshore wind farm power collection systems by using an improved Prim algorithm with an economic evaluation model for calculating power collection system costs as edge weights.

[0198] The collaborative optimization module aims to maximize the annual power generation of offshore wind farms and minimize the cost of the power collection system. It uses the NSGA-II genetic algorithm to collaboratively optimize the micro-site selection, substation layout, and power collection system design of offshore wind farms, generating the optimal wind turbine layout, substation layout, and power collection line topology.

[0199] Furthermore, the device is developed based on a B / S architecture, using the Django framework and a MySQL database, and supports multi-user online collaborative design and result export.

[0200] Example 3

[0201] This embodiment provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements the steps of the multi-objective collaborative planning and design method for offshore wind farms described in Embodiment 1.

[0202] Example 4

[0203] This embodiment provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the steps of the multi-objective collaborative planning and design method for offshore wind farms described in Embodiment 1.

[0204] In summary, this invention employs orthogonal experimental design to optimize the micro-site selection of offshore wind farms, uses an improved elbow method and K-means++ algorithm to optimize the layout of booster stations, and uses an improved Prim algorithm with an economic evaluation model for calculating the cost of the power collection system as edge weights to optimize the power collection system design. With the objectives of maximizing the annual power generation of the offshore wind farm and minimizing the cost of the power collection system, the NSGA-II genetic algorithm is used for collaborative optimization to solve for the wind turbine layout, booster station layout, and power collection line topology. This approach balances power generation efficiency and economy, improving the overall benefits of offshore wind farms. It has significant engineering guiding significance for the collaborative optimization design of micro-site selection, booster station layout, and power collection system design for large-scale offshore wind farm clusters, and has excellent application prospects in engineering.

[0205] Those skilled in the art will understand that embodiments of this application can be provided as methods, apparatus, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0206] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0207] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0208] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0209] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the technical principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A multi-objective collaborative planning and design method for offshore wind farms, characterized in that, The application comprises the following steps: The orthogonal experiment method is used to optimize the micro-siting of the offshore wind farm, and the wake loss model and the square and superposition model are used to calculate the annual power generation of the wind farm, including: setting the value range of the five parameters of the minimum unit: α, d1 / D, d2 / D, Δd1 / D, and Δd2 / D; applying the orthogonal experiment method to iteratively optimize the five parameters of the minimum unit to obtain a set of selectable machine sites; selecting a set of regularly arranged points from the set of selectable machine sites, and combining the wake loss model and the square and superposition model to calculate the annual power generation of the wind farm under each arrangement scheme; The elbow method and the K-means++ algorithm are used to optimize the layout of the offshore wind farm booster station; The improved Prim algorithm is used to calculate the economic evaluation model of the cost of the power collection system as the edge weight, and the power collection system design of the offshore wind farm is optimized; The NSGA-II genetic algorithm is used to cooperatively optimize the micro-siting of the offshore wind farm, the layout of the booster station, and the design of the power collection system, so as to generate the optimal wind turbine arrangement, the layout of the booster station, and the topology of the power collection line. Wherein, α is the azimuth angle of the long side in the micro-siting direction of the wind farm, d1 is the row spacing of the wind turbine and is greater than 3D, d2 is the column spacing of the wind turbine and is greater than 7D, D is the diameter of the wind turbine, Δd1 and Δd2 are the equal division intervals of the row spacing and the column spacing of the wind turbine respectively, the wake loss model is: ； ； ； ； In the formula, When only considering the impact of the i-th upstream wind turbine on the target wind turbine, the downstream... The wake wind speed at the location, The straight-line distance along the wind direction. It is the inflow wind speed. It is the standard deviation, I α It is environmental turbulence, I wake It is mechanical turbulence, Q is synthetic turbulence, C T is the thrust coefficient, r is the radial distance of the calculation point from the centerline of the wake, and k1, k2, and k3 are empirical constants. The square and superposition model is: ； In the formula, is the wake wind speed of the target wind turbine affected by n upstream wind turbines, n represents the number of upstream wind turbines affecting the target wind turbine, and i represents the sequence number. The annual power generation calculation formula of the wind farm is: ； ； ； wherein is the annual energy production of the wind farm, p j is the average output power of the wind turbine j, N is the number of wind turbines in the wind farm, ω(θ) is the wind direction frequency, v cut-out is the cut-out wind speed, v cut-in is the cut-in wind speed, k(θ) is the shape factor in the wind direction sector θ, c(θ) is the proportionality factor in the wind direction sector θ, θ is the wind direction, is the Weibull distribution calculation formula, is the power curve.

2. The multi-objective collaborative planning and design method of an offshore wind farm according to claim 1, characterized in that, The elbow method and the K-means++ algorithm are used to optimize the layout of the offshore wind farm booster station, including: The elbow method is used to determine the optimal number of booster stations, and the K-means++ algorithm is used to optimize the position of the booster station, including: Assuming that the number of booster stations K is equal to 1, it is calculated and judged whether the connection distance from the booster station to the farthest wind turbine is greater than the set distance, if not, the number of booster stations K is equal to 1, if yes, the improved elbow method is used to obtain the optimal number of booster stations K; The K-means++ algorithm is used to obtain the specific positions of the K booster stations, and the booster station position division result is obtained.

3. The multi-objective collaborative planning and design method of an offshore wind farm according to claim 2, characterized in that, The improved elbow method is used to obtain the optimal number of booster stations K, including: The value range of K is set; The error sum of squares SSE value corresponding to each K value is calculated; The starting point and the ending point are determined according to the value range of K and the SSE value, and the internal division calculation is performed; The difference between the division calculation value corresponding to each K value and the SSE value is calculated; The K value corresponding to the maximum difference between the division calculation value and the SSE value is taken as the number of booster stations K.

4. The multi-objective collaborative planning and design method of an offshore wind farm according to claim 2, characterized in that, The K-means++ algorithm comprises the following steps: Step a: randomly select a data point from the data set as the first cluster center; Step b: calculate the distance D(X) from each data point to the nearest cluster center and save it in an array, and accumulate the distances saved in the array Sum(D(X)); Step c: take a random value Random falling in Sum(D(X)), and use Random=D(X) until Random<=0, to obtain the next cluster center; Step d: repeat steps b to c until K cluster centers are selected; Step e: Run the standard K-means algorithm with K cluster centers.

5. The multi-objective collaborative planning and design method of an offshore wind farm according to claim 1, characterized in that, An improved Prim algorithm is used to calculate the economic evaluation model of the cost of the power collection system as the edge weight to optimize the design of the power collection system of the offshore wind farm, including: An improved Prim algorithm is used to generate the topology structure of the power collection line with the economic evaluation model of the construction cost and line loss cost of the power collection line as the edge weight, including: Step A: Import wind farm data: wind turbine site coordinates in the wind farm, cable database, substation location, number of incoming lines of the substation, and number of wind turbines that can be connected to the substation; Step B: Assume that the set of wind turbine sites in the wind farm is H, the set of wind turbine sites connected to the substation is H1, the set of wind turbine sites not connected to the substation is H2, the set of generated edges is K1, and the set of ungenerated edges is K2; Step C: Assume that the number of incoming lines of the substation is M, sort the weights of the edges associated with the substation in ascending order, and select the M edges with the smallest weights in turn to connect, put the generated M edges into K1 in turn, put the connected wind turbine sites into H1 in turn, and delete them from H2; Step D: Add wind turbines connected to the submarine cable to the M incoming line set according to the principle of adjacent and minimum weight; Step E: Calculate the cost of the power collection line according to the economic evaluation model; Step F: Repeat steps D to E until each incoming line reaches the maximum capacity and cannot add new non-intersecting edges, and obtain the topology structure of the power collection line; wherein the economic evaluation model C all is: ； ； ； Wherein, C1 is the construction cost of power collection line, C2 is the line loss cost of power collection line, N g is the number of wind turbines in the gth power collection line, L gl is the length of the lth section of submarine cable in the gth power collection line, T gl is the construction cost unit price of the lth section of submarine cable in the gth power collection line, F is the number of branch lines of power collection line, R is the resistance of power collection line, I n is the working current on the power collection line, p price is the on-grid electricity price.

6. The multi-objective collaborative planning and design method of an offshore wind farm according to claim 1, characterized in that, The NSGA-II genetic algorithm is used to optimize the micro-siting of the offshore wind farm, the layout of the substation, and the design of the power collection system to maximize the annual power generation of the offshore wind farm and minimize the cost of the power collection system, and the optimal wind turbine arrangement, substation layout, and power collection line topology are generated, including: The NSGA-II genetic algorithm is used to optimize the micro-siting of the offshore wind farm, the layout of the substation, and the design of the power collection system to maximize the annual net power generation of the entire field and minimize the cost of the power collection system, and the Pareto solution set is output, from which the optimal optimization scheme that meets the engineering requirements is selected to generate the optimal wind turbine arrangement, substation layout, and power collection line topology, including: Step G: Randomly generate an initial population with a size equal to the number of wind turbines, and take the maximum annual power generation of the wind farm and the minimum cost of the power collection system as the objective function; Step H: Perform non-dominated sorting and crowding calculation, and record the current iteration number G = 1; Step I: Obtain the first generation of offspring population through selection, crossover, and mutation operations; Step J: From the second generation, combine the parent and offspring populations, perform fast non-dominated sorting, and calculate the crowding degree of each individual in the non-dominated layer; Step K: Select individuals to form a new parent population, and use the genetic algorithm to generate a new offspring population; Step L: Repeat steps I to K, and judge whether the constraint condition is met. If it is met, perform a new round of optimization iteration until the set iteration step is reached. If it is not met, end and output the Pareto solution set; Step M: Select the optimization scheme that meets the engineering requirements from the Pareto solution set to generate the optimal wind turbine arrangement, substation layout, and power collection line topology; Wherein, the constraint condition is: the number of wind turbines is determined, the wind turbines are inside the planned wind farm, the distance between any two wind turbines is greater than 3D, the wind turbines are arranged in a regular arrangement, and D is the rotor diameter of the wind turbine.

7. An offshore wind farm multi-objective collaborative planning and design apparatus, characterized in that, It comprises: A micro-siting optimization module for optimizing the micro-siting of the offshore wind farm by using an orthogonal experiment method, and calculating the annual power generation of the wind farm in combination with a wake loss model and a square and superposition model, comprising: setting the value range of the minimum unit five parameters: α, d1 / D, d2 / D, Δd1 / D and Δd2 / D; applying the orthogonal experiment method to iteratively optimize the minimum unit five parameters to obtain a set of selectable site points; selecting and generating a set of regularly arranged points from the set of selectable site points, and calculating the annual power generation of the wind farm under each arrangement scheme in combination with the wake loss model and the square and superposition model; A booster station layout optimization module for optimizing the layout of the booster station of the offshore wind farm by using an improved elbow method and a K-means++ algorithm; A power collection system design optimization module for optimizing the design of the power collection system of the offshore wind farm by using an improved Prim algorithm with the economic evaluation model of the cost of the power collection system as the edge weight; A collaborative optimization module for collaboratively optimizing the micro-siting, booster station layout and power collection system design of the offshore wind farm by using an NSGA-II genetic algorithm to maximize the annual power generation of the offshore wind farm and minimize the cost of the power collection system, and generating the optimal wind turbine arrangement, booster station layout and power collection system topology; Wherein, α is the azimuth angle of the long side in the micro-siting direction of the wind farm, d1 is the row spacing of the wind turbines and is greater than 3D, d2 is the column spacing of the wind turbines and is greater than 7D, D is the rotor diameter of the wind turbine, Δd1 and Δd2 are the equal division spacings of the row spacing and column spacing of the wind turbines, respectively, the wake loss model is: ； ； ； ； In the formula, When only considering the impact of the i-th upstream wind turbine on the target wind turbine, the downstream... The wake wind speed at the location, The straight-line distance along the wind direction. It is the inflow wind speed. It is the standard deviation, I α It is environmental turbulence, I wake It is mechanical turbulence, Q is synthetic turbulence, C T is the thrust coefficient, r is the radial distance of the calculation point from the centerline of the wake, and k1, k2, and k3 are empirical constants. The square and superposition model is: ； In the formula, is the wake wind speed of the target wind turbine affected by n upstream wind turbines, n represents the number of upstream wind turbines affecting the target wind turbine, and i represents the sequence number. The annual power generation calculation formula of the wind farm is: ； ； ； wherein is the annual energy production of the wind farm, p j is the average output power of the wind turbine j, N is the number of wind turbines in the wind farm, ω(θ) is the wind direction frequency, v cut-out is the cut-out wind speed, v cut-in is the cut-in wind speed, k(θ) is the shape factor in the wind direction sector θ, c(θ) is the proportionality factor in the wind direction sector θ, θ is the wind direction, is the Weibull distribution calculation formula, is the power curve.

8. An electronic device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, The processor executes the program to realize the steps of the offshore wind farm multi-objective collaborative planning and design method of any one of claims 1 to 6.

9. A computer readable storage medium having stored thereon a computer program, characterized in that The computer program is executed by the processor to realize the steps of the offshore wind farm multi-objective collaborative planning and design method of any one of claims 1 to 6.

Images