Optimization method for arrangement of wind power plant

Through genetic algorithms, the problem of wind farm fan layout depends on experience in the traditional method of wind farm fan layout is solved, and the automation and efficiency improvement of wind farm fan layout is achieved. It is suitable for complex environments such as offshore, onshore and Shago wasteland.

CN118966413APending Publication Date: 2024-11-15CENT SOUTHERN CHINA ELECTRIC POWER DESIGN INST CHINA POWER ENG CONSULTING GROUP CORP
View PDF 0 Cites 5 Cited by

Patent Information

Application Number
CN202410988671.7
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-23
Publication Date
2024-11-15

AI Technical Summary

Technical Problem

The traditional wind farm fan layout method depends on the experience of engineers and cannot realize the optimization layout of large-scale wind farm fans such as offshore, land plains and Shago wasteland, which makes it difficult to resolve the contradiction between wake effect and floor area. The existing software Unable to realize automated fan layout calculations.

Method used

The wind farm layout optimization method based on genetic algorithm is adopted. By constructing the wind farm area model, setting the fan layout node, using the genetic algorithm to perform multiple iteration calculations, automatically optimizing the fan layout plan, and combining the basic information and constraint parameters of the wind farm, the wind farm is calculated. The full-field power generation and wake effect achieve the optimal configuration of fan spacing.

Benefits of technology

It realizes automatic optimization of wind farm fan layout, reduces manual intervention, improves calculation efficiency and result reliability, and can optimize fan layout in complex environments such as offshore, onshore and Shago wasteland, and improves power generation and power generation efficiency per unit area.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118966413A_ABST
    Figure CN118966413A_ABST
Patent Text Reader

Abstract

The invention belongs to the technical field of wind power generation, and particularly relates to a wind power plant arrangement optimization method which comprises the following steps: constructing a wind power plant regional model, and setting a plurality of fan arrangement nodes on the wind power plant regional model; generating a plurality of individuals as an initial population according to the number of fans required by the wind power plant and the wind power plant regional model; determining constraint parameters of the genetic algorithm; in combination with the basic information and the constraint parameters of the wind power plant, performing multiple times of population iterative calculation on the initial population through a genetic algorithm, and performing fitness calculation on each individual in a filial generation population generated after each time of population iterative calculation; stopping calculation until the number of times of population iterative calculation reaches a convergence condition, and taking a filial generation population after the last iteration as a target population; and taking a first number of individuals with the maximum target fitness value in the target population as an optimal individual set, and obtaining a first number of fan arrangement schemes of the wind power plant according to the optimal individual set.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention belongs to the technical field of wind power generation, and in particular relates to a method for optimizing the layout of a wind farm. Background Art

[0002] During the construction of large-scale wind farms such as offshore wind farms, onshore plain wind farms and Shagohuang wind farms, the main issues that the builders consider are construction investment and power generation.

[0003] The equipment cost of wind turbines is an important part of the investment cost of wind farm construction. The equipment cost of wind turbines is generally determined according to the bidding results, while the power generation is generally related to the layout plan of wind turbines. In the traditional design method, the layout of wind turbines in wind farms is usually carried out according to the experience of engineers. In the design process of wind turbine layout plan, it is mainly necessary to solve the contradiction between the distance of wind turbines, wake effect and land area. On the one hand: the existence of the wake effect of wind farms requires sufficient spacing between wind turbines to ensure that the mutual influence between the positions is minimized and the power generation is high. On the other hand, the planned land area constraint requires that the spacing between wind turbines cannot be too large. For the array layout of wind farms, the traditional solution is to ensure that the spacing in the main wind direction is large and the spacing perpendicular to the main wind direction is small. This empirical method suppresses the power generation loss caused by the wake to a certain extent, but it relies too much on the experience of engineers and cannot be exhausted and optimized.

[0004] There are many software related to power generation calculation in the field of wind power, such as WAsP, WindPRO, WindFarmer, WindSim and Meteodyn WT, but the existing software cannot realize the optimal layout calculation of wind turbines in large-scale wind farms such as offshore wind power, onshore plain wind farms and Shagohuang. At present, most survey and design companies still mainly use traditional manual optimization or local module automatic optimization methods in the micro-site selection stage of wind farms, which cannot reflect the design superiority. Summary of the invention

[0005] In order to solve the defects of the above-mentioned prior art, the present invention provides a method for optimizing the layout of wind farms. The method is improved based on a genetic algorithm and is a method for fully automatic iterative calculation that does not require excessive human intervention. The method arranges wind turbines according to the overall power generation of the wind farm. Applying the method to the layout of offshore wind turbines can better solve the problems mentioned above.

[0006] In order to achieve the above object, the technical solution of the present invention is as follows:

[0007] A method for optimizing the layout of a wind farm comprises the following steps:

[0008] Step 1: construct a wind farm regional model, and set a plurality of wind turbine layout nodes on the wind farm regional model;

[0009] Step 2: Generate a number of individuals as an initial population according to the number of wind turbines required by the wind farm and the regional model of the wind farm; wherein the individuals include wind turbine layout information of each wind turbine layout node;

[0010] Step 3: Determine the constraint parameters of the genetic algorithm, wherein the constraint parameters include a preset number of iterations, a target fitness value, and the number of wind turbines required in the wind farm;

[0011] Step 4: Combine the basic information of the wind farm and the constraint parameters, perform multiple population iterative calculations on the initial population through a genetic algorithm, and perform fitness calculations on each individual in the offspring population generated after each population iterative calculation; stop the calculation until the number of population iterative calculations reaches the preset number of iterations, or the number of individuals in the offspring population obtained after a certain population iterative calculation whose first fitness value is not less than the target fitness value is not less than the first number, and use the offspring population after the last iteration as the target population;

[0012] Step 5: Taking a first number of individuals with the largest target fitness values ​​in the target population as an optimal individual set, and obtaining a first number of wind turbine layout schemes for the wind farm according to the optimal individual set.

[0013] The fitness can be set to various parameters of the wind farm, including but not limited to: the total power generation of the wind farm, the total wind speed lost by each motor in the wind farm due to the wake effect, and the power generation efficiency per unit area of ​​the wind farm.

[0014] The above technical solution is improved based on genetic algorithm. It does not require much human intervention in the automatic iterative calculation process, and can exhaustively enumerate the wind turbine layout schemes of the wind farm to a certain extent. By calculating the fitness of each individual, the advantages and disadvantages of the total wind farm power generation corresponding to each individual are compared to select the optimal wind turbine layout scheme. The wind farm wind turbine optimization layout calculation can be performed fully automatically. Applying it to wind turbine layout can reduce manual operations and liberate productivity.

[0015] Furthermore, the basic information of the wind farm includes wind direction, wind speed, wind turbine power curve and wake loss coefficient of the wind farm;

[0016] The method for calculating the fitness of any individual in the child population generated after each population iterative calculation includes: obtaining the total power generation value of all wind turbines in the individual according to the wind turbine layout information of each wind turbine layout node in the individual and the basic information of the wind farm, and taking the total power generation value corresponding to the individual as the first fitness value of the individual.

[0017] By making the basic information of the wind farm include wind direction, wind speed, wind turbine power curve and wake loss coefficient, it can be combined with various wake calculation models such as Jensen to obtain the total power generation value of all wind turbines in any individual, and output parameters such as wind turbine coordinates required by users. The calculation results have good reliability and strong operability, and provide a good theoretical basis.

[0018] Among them, wind farm wind direction, wind farm wind speed, wind turbine power curve, wake loss coefficient, theoretical power generation, and power generation considering wake benefits can be output as calculation results in the wind turbine layout plan.

[0019] Furthermore, the constraint parameters also include crossover probability and mutation probability;

[0020] The step 4 includes the following sub-steps:

[0021] S401: performing a crossover operation or a mutation operation of a genetic algorithm on each individual in the initial population to generate a progeny population, and completing a population iterative calculation;

[0022] S402: Calculate the fitness of each individual in the offspring population. If the number of individuals in the offspring population whose first fitness value is not less than the target fitness value is not less than the first number, take the offspring population as the target population and proceed to step 5; otherwise, proceed to sub-step S403;

[0023] S403: Detect whether the number of iterative calculations of the population reaches the preset number of iterations. If so, use the offspring population as the target population and proceed to step 5; otherwise, use the offspring population as the new initial population and return to sub-step S401 to perform the next population iterative calculation.

[0024] Compared with the implementation of genetic algorithms in the prior art, each iterative calculation usually performs a crossover operation and a mutation operation before a fitness calculation. In the wind farm layout optimization method provided by the present invention, each population iterative calculation only performs a crossover operation or a mutation operation before a fitness calculation. This can not only retain the advantage of jumping out of the local optimal solution through the mutation operation, but also more comprehensively evaluate the fitness of the population after each operation. When the genetic algorithm is implemented in the prior art, the situation in which the optimal individual in the previous operation is covered by the next operation will not occur, thereby improving the comprehensiveness and efficiency of the fitness calculation.

[0025] Further, in S402, if the number of individuals in the offspring population whose first fitness value is not less than the target fitness value is less than the first number, the individuals in the offspring population are sorted in descending order according to the first fitness value, and a preset number of individuals in the front order are set as elite individuals, and then the process proceeds to sub-step S403;

[0026] In S401, if there are elite individuals among the individuals in the initial population, the crossover operation or the mutation operation is performed only on the individuals in the initial population except the elite individuals, and the new individuals generated by the crossover operation or the mutation operation and the elite individuals together form the offspring population.

[0027] Furthermore, the method of setting a plurality of wind turbine arrangement nodes on the wind farm regional model includes:

[0028] The wind farm regional model is discretized into plane grid units, the grid nodes in the restricted area of ​​the plane grid units are removed, and the remaining grid nodes in the plane grid units are used as wind turbine layout nodes.

[0029] Furthermore, the horizontal coordinate direction or the vertical coordinate direction of the plane grid unit is arranged parallel to the wind direction of the wind farm.

[0030] By making the horizontal coordinate direction or the vertical coordinate direction of the plane grid unit parallel to the wind direction of the wind farm, constraints are set in subsequent iterative calculation processes to control the distance between adjacent wind turbines in a direction parallel to the wind direction of the wind farm, and to control the distance between adjacent wind turbines in a direction perpendicular to the wind direction of the wind farm.

[0031] Furthermore, the restricted areas include shipwreck areas, reef areas, buried mine areas and ecological red line areas.

[0032] Furthermore, the spacing value between any two adjacent grid nodes in the same row in the plane grid unit is greater than the preset row spacing value, and the spacing value between any two adjacent grid nodes in the same column in the plane grid unit is greater than the preset column spacing value.

[0033] In another technical solution, the constraint parameters further include a minimum row spacing value between two fans in the same row in the plane grid unit, and a minimum column spacing value between two fans in the same column in the plane grid unit.

[0034] Furthermore, the wind farm layout optimization method further includes the following steps:

[0035] Step 6: selecting the layout location of the booster station in the wind farm regional model;

[0036] Step 7: Based on the layout position of the booster station and all the fan layout schemes, the fans are sorted according to the number of loops using a scanning method, including setting an initial baseline, calculating the angle between the line connecting each fan and the booster station and the baseline, and sorting the fans according to the absolute value of the angle;

[0037] Step 8: Use the scanning method to group the sorted fans according to the number of circuits;

[0038] Step 9: Execute the minimum spanning tree method in each group until all wind turbines in each group are connected, and obtain the topology scheme of the collector line of each group;

[0039] Step 10: Calculate the estimated current carrying capacity of each collector line in each group of collector line topology schemes, and complete the selection of each collector line in each group of collector line topology schemes;

[0040] Step 11: After the optimal topology solution in each group of collector line topology solutions is saved in the variable Y, return to step 7 until the convergence condition is reached, and then output the saved optimal topology solution;

[0041] Among them, when the fan sorting process is cyclically executed, the connection between the fan and the booster station is taken as the new baseline in each cycle, and the fans are re-sorted using the scanning method; when the fan grouping process is cyclically executed, the adjacent mutation method is used to re-group the fans to obtain adjacent mutation offspring groupings.

[0042] Based on the wind farm's wind turbine layout scheme, the minimum spanning tree method and scanning method combined with iterative calculations are used to perform the wind farm's collector line topology optimization design. The wind farm's wind turbine layout and collector line comprehensive optimization calculations can be fully automatically performed, saving labor. When the number of wind turbines in a wind farm increases significantly, the workload of the wind turbine layout and collector line comprehensive optimization has obviously exceeded the acceptable range of manpower. At this time, this method can be used to conveniently complete it.

[0043] Further, the use of the scanning method to sort the fans includes:

[0044] Take the horizontal line pointing to the right as the initial reference line;

[0045] Calculate the angle between the line connecting each fan and the booster station and the reference line, where counterclockwise is positive and clockwise is negative, and the angle range is -180° to +180°;

[0046] Sort the fans according to the size of the angle.

[0047] Furthermore, the method of grouping the sorted fans according to the number of circuits by using the scanning method includes:

[0048] According to the maximum number of fans N allowed to be connected to the collector line max、Total number of fans N tot , calculate the minimum required for the collector line layout

[0049] Pre-generate random vector A1(N1, N2, ..., N n ),in, x<N i <N max , and is an integer, n∈[m,l], and n is an integer, m is the minimum number of loops required for the collector line layout, l is the maximum number of loops required for the collector line layout, x is a non-negative integer, m, l, x are all assigned by the user;

[0050] According to the random vector A1 in N i The numerical values ​​group the sorted fans.

[0051] Further, the performing of the minimum spanning tree method in each group includes:

[0052] Generate a variable Cable_Distance for counting the total length of the collector line, set the initial value of the variable Cable_Distance to 0, and generate an empty vector Set_Cable for counting the connection information of the collector line, which is used to store the information of the two end points of the collector line connection;

[0053] Calculate the distance between each fan and the booster station, and the distance between the fans to form an N tot +1-order square matrix Distance, the first column or the first row stores the distance from the fan to the booster station, and the remaining matrix cells store the distance between the fans;

[0054] Put the distance between each wind turbine and the booster station in the group into the row vector distance, select the smallest element distance(min) in the row vector, and update the variable Cable_Distance=Cable_Distance+distance(min); update the empty vector Set_Cable, including putting the wind turbine number and booster station information corresponding to the smallest element into the empty vector Set_Cable, and delete the information of the smallest element in the row vector distance;

[0055] Calculate the distance between the remaining fans in the group and all points in the empty vector Set_Cable, select the fan number and point position in the empty vector Set_Cable corresponding to the smallest element, put them into the empty vector Set_Cable, update the variable Cable_Distance, and delete the fan information corresponding to the smallest element.

[0056] Furthermore, the estimated current carrying capacity of the collector line is calculated according to the following formula:

[0057]

[0058] Where I1 is the expected current carrying capacity of the collector line; P j is the rated power of the jth wind turbine connected to the collector line; U is the rated voltage of the wind turbine; is the current coefficient;

[0059] The rated current carrying capacity of the selected collector circuit is greater than the expected current carrying capacity of the collector circuit.

[0060] Furthermore, in the step 10, the estimated current carrying capacity of each collector line in each group of collector line topology schemes is calculated, and the selection of each collector line in each group of collector line topology schemes is completed;

[0061] According to the selection of each collector line in each group of collector line topology schemes, the cross-sectional area of ​​each collector line and the length of each collector line in each group of collector line topology schemes are obtained, and then according to the cross-sectional area of ​​each collector line and the length of each collector line in each group of collector line topology schemes, the collector line voltage drop loss corresponding to each group of collector line topology schemes is obtained;

[0062] In step 11, the collector line voltage drop loss corresponding to each group of collector line topology schemes is detected. If the collector line voltage drop loss does not meet the requirements, the corresponding collector line topology scheme is deleted, and the optimal topology scheme among the remaining collector line topology schemes is saved in variable Y, and then return to step 7 until the convergence condition is reached, and the saved optimal topology scheme is output.

[0063] Furthermore, the optimal topology solution in each group of collector line topology solutions is saved in the variable Y, including:

[0064] Calculate the second fitness value corresponding to each group of collector line topology schemes, where the second fitness value is the sum of the products of the cross-sectional area and the length of each collector line in the collector line topology scheme, and save the collector line topology scheme corresponding to the smallest second fitness value as the optimal scheme in the variable Y;

[0065] The information saved in the variable Y includes the machine position coordinates on each loop, the loop number, the actual current on the loop, the length of the collector line, the selected collector line model, the voltage drop loss, and whether the collector line passes through the restricted area.

[0066] Furthermore, the regrouping of wind turbines by using the adjacent variation method includes:

[0067] The two adjacent vector elements of each grouping scheme are increased by 1 or decreased by 1 to form a new vector A2;

[0068] Check each element in vector A2 and remove N i >Nmax The grouping scheme is used to obtain the adjacent mutant offspring grouping.

[0069] Furthermore, in step 6, boundary conditions are set, including the wind turbine layout plan, the layout location of the booster station in the wind farm area model, and the direct buried current carrying capacity, current coefficient, length coefficient, resistance and reactance of various cross-section collector lines.

[0070] The optimization method for wind farm layout provided by the present invention has a clear process, strong operability, a good theoretical basis, and a fast calculation speed. Based on the genetic algorithm, the minimum spanning tree method, and the scanning method, the improved algorithm can output the machine position coordinates, the circuit number, the real current on the circuit, the length of the collector line, the selected collector line conductor model, the voltage drop loss, whether the collector line passes through the area where the machine cannot be laid, and other information on each circuit of each scheme. The calculation results are comprehensive, more schemes are obtained, and the reliability of the calculation results is good. BRIEF DESCRIPTION OF THE DRAWINGS

[0071] In order to more clearly illustrate the technical solution of the present invention, the drawings required for use in the implementation mode will be briefly introduced below. Obviously, the drawings described below are only some implementation modes of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0072] Figure 1 is an execution flow chart of the optimization method for wind farm layout in the embodiment;

[0073] Figure 2 Schematic diagram of the optimal layout of the wind farm in the embodiment DETAILED DESCRIPTION

[0074] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments.

[0075] like Figure 1 As shown, this embodiment provides a method for optimizing the layout of a wind farm, comprising the following steps:

[0076] Step 1: construct a wind farm regional model, and set multiple wind turbine layout nodes on the wind farm regional model;

[0077] Step 2: Generate a number of individuals as the initial population according to the number of wind turbines required in the wind farm and the regional model of the wind farm; wherein the individuals include the wind turbine layout information of each wind turbine layout node;

[0078] Step 3: Determine the constraint parameters of the genetic algorithm, which include the preset number of iterations, the target fitness value, and the number of wind turbines required in the wind farm;

[0079] Step 4: Combine the basic information and constraint parameters of the wind farm, perform multiple population iteration calculations on the initial population through the genetic algorithm, and perform fitness calculations on each individual in the offspring population generated after each population iteration calculation; stop the calculation until the number of population iteration calculations reaches the preset number of iterations, or the number of individuals in the offspring population obtained after a certain population iteration calculation whose first fitness value is not less than the target fitness value is not less than the first number, and use the offspring population after the last iteration as the target population;

[0080] Step 5: Taking the first number of individuals with the largest target fitness values ​​in the target population as the optimal individual set, and obtaining the first number of wind turbine layout plans for the wind farm according to the optimal individual set.

[0081] The fitness can be set to various parameters of the wind farm, including but not limited to: the total power generation of the wind farm, the total wind speed lost by each motor in the wind farm due to the wake effect, and the power generation efficiency per unit area of ​​the wind farm.

[0082] The above technical solution is improved based on genetic algorithm. It does not require much human intervention in the automatic iterative calculation process, and can exhaustively enumerate the wind turbine layout schemes of the wind farm to a certain extent. By calculating the fitness of each individual, the advantages and disadvantages of the total wind farm power generation corresponding to each individual are compared to select the optimal wind turbine layout scheme. The wind farm wind turbine optimization layout calculation can be performed fully automatically. Applying it to wind turbine layout can reduce manual operations and liberate productivity.

[0083] Specifically, in this embodiment, the basic information of the wind farm includes the wind direction of the wind farm, the wind speed of the wind farm, the wind turbine power curve and the wake loss coefficient;

[0084] The method for calculating the fitness of any individual in the child population generated after each population iteration calculation includes: obtaining the total power generation value of all wind turbines in the individual according to the wind turbine layout information of each wind turbine layout node in the individual and the basic information of the wind farm, and taking the total power generation value corresponding to the individual as the first fitness value of the individual.

[0085] By making the basic information of the wind farm include wind direction, wind speed, wind turbine power curve and wake loss coefficient, it can be combined with various wake calculation models such as Jensen to obtain the total power generation value of all wind turbines in any individual, and output parameters such as wind turbine coordinates required by users. The calculation results have good reliability and strong operability, providing a good theoretical basis.

[0086] Among them, preferably, the wind direction of the wind farm, the wind speed of the wind farm, the wind turbine power curve, the wake loss coefficient, the theoretical power generation, and the power generation considering the wake benefit can be output as calculation results in the wind turbine layout plan.

[0087] The improved Jensen model (this is a wind farm power generation calculation model, and other models can also be selected) is used to calculate the energy density, theoretical power generation, average wind speed considering the wake effect, power generation considering the wake, and wake loss coefficient of each scheme in the group. The power generation of the entire wind farm considering the wake is used as the fitness value, and the fitness of various layout schemes is calculated respectively, and sorted in descending order according to the fitness, and the results of the optimal scheme, such as the position coordinates, position number, energy density, theoretical power generation, average wind speed considering the wake effect, power generation considering the wake, and wake loss coefficient, are stored.

[0088] Specifically, in this embodiment, the method of generating a plurality of individuals according to the number of wind turbines required by the wind farm and the wind farm regional model includes:

[0089] According to the number of wind turbines N required by the wind farm, N nodes are selected from multiple wind turbine layout nodes on the wind farm regional model, and the wind turbines are arranged at the selected wind turbine layout nodes.

[0090] The multiple wind turbine layout nodes on the wind farm regional model are encoded. The wind turbine layout nodes with wind turbines are encoded with binary digits "1", and the wind turbine layout nodes without wind turbines are encoded with binary digits "0". The encoded nodes are called genes. When all gene encodings are completed, a wind turbine layout scheme S1 is formed, also called an individual S1. For example, the gene encoding of an individual S1 is (1 0 1 0 1······1 1). The encoding "1" indicates that a wind turbine is arranged at the wind turbine layout node corresponding to the gene, and the encoding "0" indicates that a wind turbine is not arranged at the wind turbine layout node corresponding to the gene. The total number of genes encoded as "1" in the individual is the number of wind turbines N required for the wind farm.

[0091] Select x fan layout schemes, and the x fan layout schemes form an initial population. For example, the initial population X = (S1, S2, S3...Sx).

[0092] Specifically, Figure 1 As shown, in this embodiment, the constraint parameters also include crossover probability and mutation probability;

[0093] Step 4 includes the following sub-steps:

[0094] S401: Perform a crossover operation or a mutation operation of a genetic algorithm on each individual in the initial population to generate a progeny population, and complete a population iterative calculation;

[0095] S402: Calculate the fitness of each individual in the offspring population. If the number of individuals in the offspring population whose first fitness value is not less than the target fitness value is not less than the first number, the offspring population is used as the target population and the process proceeds to step 5; otherwise, the process proceeds to sub-step S403;

[0096] S403: Check whether the number of population iterative calculations reaches the preset number of iterations. If so, use the offspring population as the target population and proceed to step 5; otherwise, use the offspring population as the new initial population and return to sub-step S401 to perform the next population iterative calculation.

[0097] Compared with the implementation of genetic algorithms in the prior art, each iterative calculation usually performs a crossover operation and a mutation operation before a fitness calculation. In the wind farm layout optimization method provided by the present invention, each population iterative calculation only performs a crossover operation or a mutation operation before a fitness calculation. This can not only retain the advantage of jumping out of the local optimal solution through the mutation operation, but also more comprehensively evaluate the fitness of the population after each operation. When the genetic algorithm is implemented in the prior art, the situation in which the optimal individual in the previous operation is covered by the next operation will not occur, thereby improving the comprehensiveness and efficiency of the fitness calculation.

[0098] Specifically, in this embodiment, in S402, if the number of individuals in the offspring population whose first fitness value is not less than the target fitness value is less than the first number, the individuals in the offspring population are sorted in descending order according to the first fitness value, and after a preset number of individuals in the front order are set as elite individuals, sub-step S403 is entered;

[0099] In S401, if there are elite individuals among the individuals in the initial population, a crossover operation or a mutation operation is performed only on the individuals in the initial population except the elite individuals, and the new individuals generated by the crossover operation or the mutation operation form a descendant population together with the elite individuals.

[0100] Specifically, in this embodiment, if Figure 1 As shown, in any two adjacent population iteration calculations, one population iteration calculation implements the crossover operation of the genetic algorithm, and the other population iteration calculation implements the mutation operation of the genetic algorithm.

[0101] Before the crossover calculation, the elite retention strategy is first executed to pass the optimal layout plan to the next generation, and the remaining plans are crossed in pairs. During the crossover process, if the genes at the corresponding positions in the two individuals are the same, that is, the gene points are both equipped with fans or neither is equipped with fans, they will be retained. For different genes, that is, one individual has a fan and the other individual does not have a fan at the gene point, a random number is generated by a random function. When the random number is less than the crossover probability Pc (user-defined), the corresponding position gene is crossover calculated. During the crossover calculation process, the total number of fans in each plan must be guaranteed to remain unchanged, and the fan spacing in each plan must be calculated, and the plans that do not meet the fan spacing requirements are deleted.

[0102] In order to escape from the local optimal solution, mutation calculation is performed in each scheme formed after the crossover calculation. Traverse the genes of each layout scheme and generate random numbers through a random function. When the random number is less than the mutation probability Pm (user-defined), the gene at the corresponding position mutates, and the "0" gene mutates to "1" or the "1" gene mutates to "0". In the process of mutation calculation, the total number of fans in each scheme must be kept unchanged, and the fan spacing in each scheme must be calculated, and the schemes with fan spacing that do not meet the requirements must be deleted. The schemes after the mutation and the schemes that implement the elite retention strategy form a new offspring group.

[0103] Specifically, in this embodiment, the method for setting a plurality of wind turbine arrangement nodes on the wind farm regional model includes:

[0104] The wind farm regional model is discretized into plane grid units, the grid nodes in the restricted area of ​​the plane grid units are removed, and the remaining grid nodes in the plane grid units are used as wind turbine layout nodes.

[0105] Furthermore, the horizontal coordinate direction or the vertical coordinate direction of the plane grid unit is arranged parallel to the wind direction of the wind farm.

[0106] By making the horizontal or vertical coordinate direction of the plane grid unit parallel to the wind direction of the wind farm, constraints are set in the subsequent iterative calculation process to control the distance between adjacent wind turbines in the direction parallel to the wind direction of the wind farm, and to control the distance between adjacent wind turbines in the direction perpendicular to the wind direction of the wind farm.

[0107] Specifically, in this embodiment, the restricted areas include shipwreck areas, reef areas, buried mine areas and ecological red line areas.

[0108] In order to control the distance between adjacent wind turbines in a direction parallel to the wind direction of the wind farm, and control the distance between adjacent wind turbines in a direction perpendicular to the wind direction of the wind farm, this embodiment provides the following solutions:

[0109] Solution 1: The spacing value between any two adjacent grid nodes in the same row in the plane grid unit is greater than the preset row spacing value, and the spacing value between any two adjacent grid nodes in the same column in the plane grid unit is greater than the preset column spacing value.

[0110] Solution 2: The constraint parameters also include a minimum row spacing value between two fans in the same row in a plane grid unit, and a minimum column spacing value between two fans in the same column in a plane grid unit.

[0111] Through the above two solutions, the distance between adjacent wind turbines in a direction parallel to the wind direction of the wind farm and the distance between adjacent wind turbines in a direction perpendicular to the wind direction of the wind farm can be well controlled.

[0112] In another embodiment, Figure 1 As shown, the optimization method for wind farm layout also includes the following steps:

[0113] Step 6: Select the location of the booster station in the wind farm regional model;

[0114] Step 7: Based on the layout location of the booster station and the layout plan of all wind turbines, the wind turbines are sorted according to the number of loops using the scanning method, including setting an initial baseline, calculating the angle between the line connecting each wind turbine and the booster station and the baseline, and sorting the wind turbines according to the absolute value of the angle;

[0115] Step 8: Use the scanning method to group the sorted fans according to the number of circuits;

[0116] Step 9: Execute the minimum spanning tree method in each group until all wind turbines in each group are connected, and obtain the topology scheme of the collector line of each group;

[0117] Step 10: Calculate the estimated current carrying capacity of each collector line in each group of collector line topology schemes, and complete the selection of each collector line in each group of collector line topology schemes;

[0118] Step 11: After the optimal topology solution in each group of collector line topology solutions is saved in the variable Y, return to step 7 until the convergence condition is reached, and then output the saved optimal topology solution;

[0119] Among them, when the fan sorting process is cyclically executed, the connection between the fan and the booster station is taken as the new baseline in each cycle, and the fans are re-sorted using the scanning method; when the fan grouping process is cyclically executed, the adjacent mutation method is used to re-group the fans to obtain adjacent mutation offspring groupings.

[0120] Based on the wind farm's wind turbine layout scheme, the minimum spanning tree method and scanning method combined with iterative calculations are used to perform the wind farm's collector line topology optimization design. The wind farm's wind turbine layout and collector line comprehensive optimization calculations can be fully automatically performed, saving labor. When the number of wind turbines in a wind farm increases significantly, the workload of the wind turbine layout and collector line comprehensive optimization has obviously exceeded the acceptable range of manpower. At this time, this method can be used to conveniently complete it.

[0121] Specifically, in this embodiment, sorting the fans by using the scanning method includes:

[0122] Take the horizontal line pointing to the right as the initial reference line;

[0123] Calculate the angle between the line connecting each fan and the booster station and the reference line, where counterclockwise is positive and clockwise is negative, and the angle range is -180° to +180°;

[0124] Sort the fans according to the size of the angle.

[0125] Furthermore, the scanning method is used to group the sorted fans according to the number of circuits, including:

[0126] According to the maximum number of fans N allowed to be connected to the collector line max 、Total number of fans N tot , calculate the minimum required for the collector line layout

[0127] Pre-generate random vector A1(N1, N2, ..., N n ),in, x<N i <N max , and is an integer, n∈[m,l], and n is an integer, m is the minimum number of loops required for the collector line layout, l is the maximum number of loops required for the collector line layout, x is a non-negative integer, m, l, x are all assigned by the user;

[0128] According to the random vector A1 in N i The numerical values ​​group the sorted fans.

[0129] Specifically, in this embodiment, executing the minimum spanning tree method in each group includes:

[0130] Generate a variable Cable_Distance for counting the total length of the collector line, set the initial value of the variable Cable_Distance to 0, and generate an empty vector Set_Cable for counting the connection information of the collector line, which is used to store the information of the two end points of the collector line connection;

[0131] Calculate the distance between each fan and the booster station, and the distance between the fans to form an Ntot +1-order square matrix Distance, the first column or the first row stores the distance from the fan to the booster station, and the remaining matrix cells store the distance between the fans;

[0132] Put the distance between each wind turbine and the booster station in the group into the row vector distance, select the smallest element distance(min) in the row vector, and update the variable Cable_Distance=Cable_Distance+distance(min); update the empty vector Set_Cable, including putting the wind turbine number and booster station information corresponding to the smallest element into the empty vector Set_Cable, and delete the information of the smallest element in the row vector distance;

[0133] Calculate the distance between the remaining fans in the group and all points in the empty vector Set_Cable, select the fan number and point position in the empty vector Set_Cable corresponding to the smallest element, put them into the empty vector Set_Cable, update the variable Cable_Distance, and delete the fan information corresponding to the smallest element.

[0134] Specifically, in this embodiment, the estimated current carrying capacity of the collector line is calculated according to the following formula:

[0135]

[0136] Where I1 is the expected current carrying capacity of the collector line; P j is the rated power of the jth wind turbine connected to the collector line; U is the rated voltage of the wind turbine; is the current coefficient (generally 0.96);

[0137] The rated current carrying capacity of the selected collector circuit is greater than the expected current carrying capacity of the collector circuit.

[0138] Specifically, in this embodiment, in step 10, the estimated current carrying capacity of each collector line in each group of collector line topology schemes is calculated, and the selection of each collector line in each group of collector line topology schemes is completed;

[0139] According to the selection of each collector line in each group of collector line topology schemes, the cross-sectional area of ​​each collector line and the length of each collector line in each group of collector line topology schemes are obtained, and then according to the cross-sectional area of ​​each collector line and the length of each collector line in each group of collector line topology schemes, the collector line voltage drop loss corresponding to each group of collector line topology schemes is obtained;

[0140] In step 11, the collector line voltage drop loss corresponding to each group of collector line topology schemes is detected. If the collector line voltage drop loss does not meet the requirements, the corresponding collector line topology scheme is deleted, and the optimal topology scheme among the remaining collector line topology schemes is saved in variable Y, and then returns to step 7 until the convergence condition is met (that is, the second convergence condition is met), and then the saved optimal topology scheme is output.

[0141] Specifically, in this embodiment, saving the optimal topology solution in each group of collector line topology solutions into the variable Y includes:

[0142] Calculate the second fitness value corresponding to each group of collector line topology schemes, the second fitness value is the ratio between the sum of the products of the cross-sectional area and the length of each collector line in the collector line topology scheme and the total power generation value of all wind turbines in the wind turbine arrangement scheme corresponding to the collector line topology scheme, and save the collector line topology scheme corresponding to the smallest second fitness value as the optimal scheme in the variable Y;

[0143] The second convergence condition is that the number of times the optimal topology solution is saved in the variable Y reaches a preset number, or the second fitness value corresponding to the optimal topology solution stored in the variable Y is less than a preset fitness value.

[0144] The information saved in the variable Y includes the machine position coordinates on each loop, the loop number, the actual current on the loop, the length of the collector line, the selected collector line model, the voltage drop loss, and whether the collector line passes through the restricted area.

[0145] Specifically, in this embodiment, regrouping the wind turbines using the adjacent variation method includes:

[0146] The two adjacent vector elements of each grouping scheme are increased by 1 or decreased by 1 to form a new vector A2;

[0147] Check each element in vector A2 and remove N i >N max The grouping scheme is used to obtain the adjacent mutant offspring grouping.

[0148] Furthermore, in step 6, boundary conditions are set, including the wind turbine layout plan, the layout location of the booster station in the wind farm area model, and the direct buried current carrying capacity, current coefficient, length coefficient, resistance and reactance of various cross-section collector lines.

[0149] The optimization method for wind farm layout provided by the present invention has a clear process, strong operability, a good theoretical basis, and a fast calculation speed. Based on the genetic algorithm, the minimum spanning tree method, and the scanning method, the improved algorithm can output the machine position coordinates, the circuit number, the real current on the circuit, the length of the collector line, the selected collector line conductor model, the voltage drop loss, whether the collector line passes through the area where the machine cannot be laid, and other information on each circuit of each scheme. The calculation results are comprehensive, more schemes are obtained, and the reliability of the calculation results is good.

[0150] The wind farm layout optimization method provided by the present invention has at least the following technical effects or advantages:

[0151] 1. The optimization method for wind farm layout provided by the present invention is improved based on genetic algorithm. It does not require much human intervention in the automatic iterative calculation process, and can exhaustively enumerate the wind turbine layout schemes of the wind farm to a certain extent. By calculating the fitness of each individual, the advantages and disadvantages of the total wind farm power generation corresponding to each individual are compared to select the optimal wind turbine layout scheme. The wind farm wind turbine optimization layout calculation can be fully automatically performed. Applying it to wind turbine layout can reduce manual operations and liberate productivity.

[0152] 2. By making the basic information of the wind farm including the wind direction, wind speed, wind turbine power curve and wake loss coefficient of the wind farm, it can be combined with various wake calculation models such as Jensen to obtain the total power generation value of all wind turbines in any individual, and output the parameters such as wind turbine coordinates required by the user. The calculation results have good reliability and strong operability, providing a good theoretical basis.

[0153] 3. Compared with the implementation of genetic algorithms in the prior art, each iterative calculation usually performs a crossover operation and a mutation operation before a fitness calculation. In the optimization method for wind farm layout provided by the present invention, each population iterative calculation only performs a crossover operation or a mutation operation before a fitness calculation. This can not only retain the advantage of jumping out of the local optimal solution through the mutation operation, but also more comprehensively evaluate the fitness of the population after each operation. When the genetic algorithm is implemented in the prior art, the optimal individual in the previous operation may be covered by the next operation, which improves the comprehensiveness and efficiency of the fitness calculation.

[0154] 4. By making the horizontal or vertical coordinate direction of the plane grid unit parallel to the wind direction of the wind farm, constraints are set in the subsequent iterative calculation process to control the distance between adjacent wind turbines in the direction parallel to the wind direction of the wind farm, and to control the distance between adjacent wind turbines in the direction perpendicular to the wind direction of the wind farm.

[0155] 5. Based on the wind farm wind turbine layout scheme obtained, the minimum spanning tree method and scanning method combined with iterative calculations are used to perform the wind farm collector line topology optimization design. The wind farm wind turbine layout and collector line comprehensive optimization calculation can be fully automatically performed, saving labor. When the number of wind turbines in the wind farm increases significantly, the workload of the wind turbine layout and collector line comprehensive optimization has obviously exceeded the acceptable range of manpower. At this time, this method can be used to complete it conveniently.

[0156] 6. The optimization method for wind farm layout provided by the present invention has a clear process, strong operability, a good theoretical basis, and a fast calculation speed. Based on the genetic algorithm, the minimum spanning tree method, and the scanning method, the improved algorithm can output the coordinates of the machine position on each circuit of each scheme, the circuit number, the actual current on the circuit, the length of the collector line, the selected collector line conductor model, the voltage drop loss, whether the collector line passes through the area where the machine cannot be laid, and other information. The calculation results are comprehensive, more schemes are obtained, and the reliability of the calculation results is good.

[0157] The above are only specific application examples of the present invention and do not constitute any limitation to the protection scope of the present invention. Any technical solutions formed by equivalent transformation or equivalent replacement shall fall within the protection scope of the present invention.

Claims

1. A method for optimizing the layout of a wind farm, characterized in that: The following steps are involved: Step 1: construct a wind farm regional model, and set a plurality of wind turbine layout nodes on the wind farm regional model; Step 2: Generate a number of individuals as an initial population according to the number of wind turbines required by the wind farm and the regional model of the wind farm; wherein the individuals include wind turbine layout information of each wind turbine layout node; Step 3: Determine the constraint parameters of the genetic algorithm, wherein the constraint parameters include a preset number of iterations, a target fitness value, and the number of wind turbines required in the wind farm; Step 4: Combine the basic information of the wind farm and the constraint parameters, perform multiple population iterative calculations on the initial population through a genetic algorithm, and perform fitness calculations on each individual in the offspring population generated after each population iterative calculation; stop the calculation until the number of population iterative calculations reaches the preset number of iterations, or the number of individuals in the offspring population obtained after a certain population iterative calculation whose first fitness value is not less than the target fitness value is not less than the first number, and use the offspring population after the last iteration as the target population; Step 5: Taking a first number of individuals with the largest target fitness values ​​in the target population as an optimal individual set, and obtaining a first number of wind turbine layout schemes for the wind farm according to the optimal individual set.

2. The method for optimizing the layout of a wind farm according to claim 1, characterized in that: The basic information of the wind farm includes wind direction, wind speed, wind turbine power curve and wake loss coefficient; The method for calculating the fitness of any individual in the child population generated after each population iterative calculation includes: obtaining the total power generation value of all wind turbines in the individual according to the wind turbine layout information of each wind turbine layout node in the individual and the basic information of the wind farm, and taking the total power generation value corresponding to the individual as the first fitness value of the individual.

3. The method for optimizing the layout of a wind farm according to claim 1, characterized in that: The constraint parameters also include crossover probability and mutation probability; The step 4 includes the following sub-steps: S401: performing a crossover operation or a mutation operation of a genetic algorithm on each individual in the initial population to generate a progeny population, and completing a population iterative calculation; S402: Calculate the fitness of each individual in the offspring population. If the number of individuals in the offspring population whose first fitness value is not less than the target fitness value is not less than the first number, take the offspring population as the target population and proceed to step 5; otherwise, proceed to sub-step S403; S403: Detect whether the number of iterative calculations of the population reaches the preset number of iterations. If so, use the offspring population as the target population and proceed to step 5; otherwise, use the offspring population as the new initial population and return to sub-step S401 to perform the next population iterative calculation.

4. The method for optimizing the layout of a wind farm according to claim 3, characterized in that: In S402, if the number of individuals in the offspring population whose first fitness value is not less than the target fitness value is less than the first number, the individuals in the offspring population are sorted in descending order according to the first fitness value, and a preset number of individuals in the front order are set as elite individuals, and then the process proceeds to sub-step S403; In S401, if there are elite individuals among the individuals in the initial population, the crossover operation or the mutation operation is performed only on the individuals in the initial population except the elite individuals, and the new individuals generated by the crossover operation or the mutation operation and the elite individuals together form the offspring population.

5. The method for optimizing the layout of a wind farm according to claim 1, characterized in that: The method for setting a plurality of wind turbine arrangement nodes on the wind farm regional model comprises: The wind farm regional model is discretized into plane grid units, the grid nodes in the restricted area of ​​the plane grid units are removed, and the remaining grid nodes in the plane grid units are used as wind turbine layout nodes.

6. The method for optimizing the layout of a wind farm according to claim 5, characterized in that: The spacing value between any two adjacent grid nodes in the same row in the plane grid unit is greater than the preset row spacing value, and the spacing value between any two adjacent grid nodes in the same column in the plane grid unit is greater than the preset column spacing value.

7. The method for optimizing the layout of a wind farm according to claim 5, characterized in that: The constraint parameters also include a minimum row spacing value between two fans in a row in the plane grid unit, and a minimum column spacing value between two fans in a column in the plane grid unit.

8. A method for optimizing the layout of a wind farm, characterized in that: The following steps are also included: Step 6: selecting the layout location of the booster station in the wind farm regional model; Step 7: Based on the layout position of the booster station and all the fan layout schemes, the fans are sorted according to the number of loops using a scanning method, including setting an initial baseline, calculating the angle between the line connecting each fan and the booster station and the baseline, and sorting the fans according to the absolute value of the angle; Step 8: Use the scanning method to group the sorted fans according to the number of circuits; Step 9: Execute the minimum spanning tree method in each group until all wind turbines in each group are connected, and obtain the topology scheme of the collector line of each group; Step 10: Calculate the estimated current carrying capacity of each collector line in each group of collector line topology schemes, and complete the selection of each collector line in each group of collector line topology schemes; Step 11: After the optimal topology solution in each group of collector line topology solutions is saved in the variable Y, return to step 7 until the convergence condition is reached, and then output the saved optimal topology solution; Among them, when the fan sorting process is cyclically executed, the connection between the fan and the booster station is taken as the new baseline in each cycle, and the fans are re-sorted using the scanning method; when the fan grouping process is cyclically executed, the adjacent mutation method is used to re-group the fans to obtain adjacent mutation offspring groupings.

9. The method for optimizing the layout of a wind farm according to claim 8, characterized in that: The scanning method for ranking fans includes: Take the horizontal line pointing to the right as the initial reference line; Calculate the angle between the line connecting each fan and the booster station and the reference line, where counterclockwise is positive and clockwise is negative, and the angle range is -180° to +180°; Sort the fans according to the size of the angle.

10. The method for optimizing the layout of a wind farm according to claim 8, characterized in that: The method of grouping the sorted fans according to the number of circuits by using the scanning method comprises: According to the maximum number of fans Nmax and the total number of fans Ntot allowed to be connected to the collector line, calculate the minimum number of fans required for the collector line layout. Pregenerate a random vector A1(N1, N2, ..., Nn), where x<Ni<Nmax, and is an integer, n∈[m,l], and n is an integer, m is the minimum number of loops required for the collector line layout, l is the maximum number of loops required for the collector line layout, x is a non-negative integer, and m, l, and x are all assigned by the user; The sorted wind turbines are grouped according to the Ni values ​​in the random vector A1.

Citation Information

Cited By

  • Wind turbine generator arrangement optimization method and system considering area limitation

    CN119273102A

  • A method and system for optimizing the layout of wind turbines considering regional restrictions

    CN119273102B

  • Offshore wind farm arrangement optimization design method considering booster station

    CN120068727A

  • Optimal design method for offshore wind farm layout considering booster stations

    CN120068727B

  • Parallelogram rule-based wind power plant layout optimization method and device, and medium

    CN120124433A