Wind farm layout optimization method based on intelligent algorithm

Through intelligent algorithms combined with wake modeling and grid-based methods, the wind farm layout is optimized, and the problem of local optimality and relying on a single search strategy in the existing technology is solved, and efficient optimization of wind farm layout and maximum output power is achieved.

CN118898207BActive Publication Date: 2024-12-03NANJING UNIV OF INFORMATION SCI & TECH
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411400397.3
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-10-09
Publication Date
2024-12-03
Estimated Expiration
2044-10-09

AI Technical Summary

Technical Problem

Existing intelligent algorithms are prone to falling into local optimization in wind farm layout optimization, and their performance relies too much on a single search strategy, resulting in low quality of wind farm layout solutions.

Method used

The wind farm layout optimization method based on intelligent algorithm is adopted to simulate the wind farm through wake modeling and grid-based methods, and the objective function is designed to optimize the wind farm layout. Combined with local and global search strategies, the construction and operation costs of the wind farm are optimized and Output power.

Benefits of technology

It has achieved the steadily and efficient optimization of the wind farm layout under the premise of reducing construction and operation costs, maximize the output power of the wind farm, and improve the quality and economic value of the wind farm layout plan.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118898207B_ABST
    Figure CN118898207B_ABST
Patent Text Reader

Abstract

The present invention discloses a wind farm layout optimization method based on an intelligent algorithm, which aims to optimize the micro-layout adjustment during the construction of the wind farm, so as to increase the output power of the wind farm while reducing the construction and operation costs of the wind farm. The method of the present invention includes: modeling the wake influence existing in the wind farm; designing two types of reasonable wind farm experimental scenarios; establishing a wind farm layout optimization objective function; designing an intelligent optimization algorithm; and optimizing the wind farm layout based on the intelligent optimization algorithm. The present invention can optimize the layout of the wind farm, while reducing the construction and operation costs, determine the best positions of multiple wind turbines within the available area, thereby maximizing the output power of the wind turbines and maximizing the overall power of the wind farm.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to a wind farm layout optimization method based on an intelligent algorithm, and belongs to the technical field of wind farm layout optimization. Background Art

[0002] In an era of energy shortage, we must reduce our dependence on fossil fuels and promote the application of renewable energy. Wind energy, as a pollution-free renewable energy, has been widely used. At present, building wind farms to convert wind energy into electrical energy is one of the main ways to utilize wind energy. Therefore, optimizing the layout of wind farms to improve the efficiency of wind energy utilization is of great significance to accelerate energy transformation and improve the ecological environment.

[0003] The wind farm layout optimization problem involves multiple variables, including the location, number and type of turbines, which makes the combination of decision variables of the wind farm layout optimization problem complex and increases the scale of the problem. In addition, due to the influence of actual terrain and wind speed, the mathematical model involved in the wind farm layout optimization problem is nonlinear. Therefore, the wind farm layout optimization problem is a complex NP-hard problem. As an advanced optimization method, intelligent algorithms can be applied to various optimization problems, including complex combination optimization, continuous optimization, multi-objective optimization, etc. Existing studies have shown that they can solve the wind farm layout optimization problem. However, there are some problems that cannot be ignored in the intelligent algorithms that have been developed. For example, the algorithm is easy to fall into local optimality, the algorithm performance is too dependent on a single search strategy, etc., which leads to the low quality of the wind farm layout solution solved by the algorithm. Therefore, the technology used to optimize the wind farm layout needs to be further improved. Summary of the invention

[0004] Purpose of the invention: In order to improve the performance of the existing technology, further optimize the layout of wind farms, and improve the power generation efficiency of wind farms, the present invention proposes a wind farm layout optimization method based on an intelligent algorithm. This method can not only optimize the layout of wind farms robustly and efficiently while reducing construction and operation costs, but also maximize the output power of wind farms.

[0005] Technical solution: A wind farm layout optimization method based on intelligent algorithm, comprising the following steps:

[0006] Step S1, analyzing the wake impact of the wind farm and performing wake modeling;

[0007] Step S2, based on the wind farm wake analysis, two types of wind farm experimental scenarios that conform to the actual situation are proposed, and simulation experiments that consider the number of wind turbines installed in the wind farm are further designed under two different wind farm layout conditions, and two experiments are designed respectively according to the two situations of certain and uncertain numbers of turbines; a gridding method is used to approximately and reasonably simulate the placement of wind turbines in an actual wind farm; and encoding is performed in combination with the gridding method and different layouts of wind turbines.

[0008] Step S3, comprehensively considering the factors affecting the construction of the wind farm, selecting two most important factors to design an objective function for evaluating the quality of the wind farm layout plan.

[0009] Step S4, design an intelligent algorithm to optimize the wind farm layout. Each wind farm layout method in the algorithm is iteratively updated with the goal of finding the minimum value of the objective function in step S3. When the algorithm meets the iteration stop condition, the layout method that obtains the minimum value of the objective function is retained, and then the information it carries is parsed into the layout position of the wind turbine in the gridded simulated wind farm through a corresponding decoding method, and the wind farm construction, operation cost and output power corresponding to the layout scheme are output.

[0010] Step S5, based on the above results, draw the optimal layout scheme diagram under different wind farm experimental scenario assumptions.

[0011] The analysis of the influence of the wind farm wake in step S1 and the wake modeling include:

[0012] S11, in a wind farm, when the downstream turbine is only affected by the wake from a single upstream turbine, assuming that the initial wind speed flowing to the turbine blades is uniform, after being affected by the wake, the distance from the turbine x The wind wake speed at meters is calculated by the following formula (wake model):

[0013]

[0014] Where V0 is the initial wind speed toward the turbine, D T is the turbine disk diameter, k is the wake decay constant.

[0015] S12, when the downstream turbine is affected by the wake from multiple upstream turbines, assuming that the initial wind speed flowing to the turbine blades is uniform, after being affected by the wake, the distance from the turbine x m The wind wake velocity at is calculated by the following formula (wake model):

[0016]

[0017] In the formula, A coverIt represents the overlapping area affected by the wakes of multiple turbines, and A0 is the area of ​​one rotation of the turbine blades.

[0018] In step S2, a grid method is used to design wind farm simulation experiments under different conditions, and encoding and decoding methods are designed, including:

[0019] S21, the actual wind farm construction area is modeled as a regular shape, such as a K×K square grid, and the turbine placement position is the center of each cell.

[0020] S22, according to the relative positions of the upstream and downstream turbines, the experimental scenarios are divided into two categories according to the influence of the wake. The first category is that the downstream turbine is placed inside the wake coverage area of ​​the upstream turbine, and the wind energy received by the downstream turbine is affected by the full coverage of the upstream turbine wake; the second category is that the downstream turbine is placed at the edge of the wake coverage area of ​​the upstream turbine, and the wind energy received by the downstream turbine is partially affected by the wake of the upstream turbine.

[0021] S23, based on S22, the experiments in each case are divided into two types according to whether the number of turbines is determined when building the wind farm. The first type is when the number of turbines is determined, and in this case, it is only necessary to solve how to arrange the positions of the turbines to maximize the output power of the wind farm; the second type is when the number of turbines is uncertain, and in this case, it is necessary to solve the number and arrangement of turbines that can maximize the output power while minimizing the cost. Therefore, the experimental assumptions in the present invention are the following four cases:

[0022] (1) The downstream turbine is placed inside the upstream turbine wake coverage area. The wind energy received by the downstream turbine is affected by the full coverage of the upstream turbine wake. The number of wind turbines is [1, K 2 ] is an integer between ].

[0023] (2) The downstream turbine is placed at the edge of the upstream turbine wake coverage area. The wind energy received by the downstream turbine is partially affected by the upstream turbine wake. The number of wind turbines is [1, K 2 ] is an integer between ].

[0024] (3) The downstream turbines are placed inside the wake coverage area of ​​the upstream turbines. The wind energy received by the downstream turbines is affected by the full coverage of the upstream turbine wake. The number of wind turbines is set according to actual needs.

[0025] (4) The downstream turbines are placed at the edge of the upstream turbine wake coverage area. The wind energy received by the downstream turbines is partially affected by the upstream turbine wake. The number of wind turbines is set according to actual needs.

[0026] S24, based on S21-S23, encode and decode the turbine layout. When the number of turbines is uncertain, set the variableα 1×K 2 A vector representing K in a K×K square grid 2 positions, all solution vectors are in the range [0,1] and are rounded to integers, i.e. 0 or 1. A solution with a value of 1 represents a location where a turbine is placed, and a solution with a value of 0 represents a location where no turbine is placed. When the number of turbines is determined, α for n ×1 vector, n is the number of turbines, all solution vectors are in [1,K 2 ] and rounded to an integer, which is the cell location where the turbine is installed; then create a G c ={1, 2, ..., K 2}, if α The i-th element in is rounded ( α i ) is lower than G c The number of remaining elements in the grid is G c (round( α i )) is designated as i turbine locations and from the list G c Delete the selected grid position from the G c The last element in is designated as i The location of a turbine is determined and removed from the list.

[0027] In step S3, the factors affecting the construction of the wind farm are comprehensively considered, and two main factors are selected to design an objective function for evaluating the quality of the wind farm layout plan, including:

[0028] S31, wind farm construction and operation costs are calculated by the following formula:

[0029]

[0030] In the formula, N is the number of turbines.

[0031] S32, the wind farm output power is obtained by the following formula:

[0032]

[0033] Where η is the turbine power generation efficiency, ρ is the air density, A0 is the area of ​​the turbine blades during one rotation, V xiis the distance from the upstream turbine x Flow to the i The wind wake velocity of the downstream turbine.

[0034] S33, based on S31 and S32, the objective function for evaluating the quality of wind farm layout scheme is calculated by the following formula:

[0035] Objective Function=min

[0036] The smaller the objective function value is, the higher the quality of the wind farm layout solution corresponding to the solution is, that is, it can maximize the output power while minimizing the cost.

[0037] In step S4, using the designed intelligent algorithm to optimize the layout of the wind farm specifically includes:

[0038] S41, based on steps S31 and S32, input wind farm simulation experiment parameters to improve the modeling of wind farm layout optimization problem, including wind farm area size, hub height, rotor diameter, air density, ground roughness, initial wind speed and wind direction. Initialize the population information of the intelligent algorithm and store the initialized wind farm layout in the following matrix X middle:

[0039]

[0040] Among them, the matrix X Each row in represents a layout of a wind farm. NP is the population size of the intelligent algorithm, d is the dimension of the wind farm layout problem to be solved. Set the algorithm termination condition to reach the maximum number of iterations of the algorithm.

[0041] S42, based on the calculation algorithm in S33, the effect of each wind farm layout mode, i.e., the fitness value, is stored in the following matrix OX middle:

[0042]

[0043] In the formula, OX NP Representative NP The fitness value of the wind farm layout mode. Then, the matrix OX The elements in are sorted by size, and after sorting, the layout modes are saved from best to worst according to the fitness value in the following matrix S middle:

[0044]

[0045] The corresponding sorted and updated layout fitness values ​​are stored in the following matrix OSmiddle:

[0046]

[0047] S43, calculating the probability of being selected for each wind farm layout mode. The probability of being selected for each layout mode is calculated to distinguish between layout modes with good fitness values ​​and layout modes with poor fitness values. The probability of being selected is obtained as follows:

[0048] SP i =

[0049] Where, SP i Represents the first i The probability of selecting a layout mode is i It is the sorting position of the layout after sorting by fitness value. NP is the population size of the intelligent algorithm, SP i The larger the layout, the better, which means that this layout may be closer to the optimal solution. The corresponding wind farm layout is efficiently searched locally according to the following formula:

[0050] X i =|SX i |.e t .cos(2πt)+S

[0051] Among them, S represents the optimal layout method found by the current intelligent algorithm, and X i It is the i-th layout method in the intelligent algorithm, and the parameter t The value range is [ r ,1], r During the iteration process, it decreases linearly in the range of [-2, -1], t The size of X i The distance to S is positively correlated. This update method can make X i In the early stages of the iteration, the entire search space is covered as much as possible. As the iterations proceed, the local search capability of the algorithm gradually increases.

[0052] Otherwise, the wind farm layout is efficiently searched globally according to the following formula:

[0053]

[0054] Among them, NewX i is the i-th layout method newly generated by the intelligent algorithm, X r1 , X r2 and X r3are three layout methods randomly selected from the intelligent algorithm population, and X r1 The fitness value is better than X r2 , X r3 , j represents the jth dimension of the wind farm layout problem to be solved, j=1, 2, …, d; d is the dimension of the wind farm layout problem, j rand is 1 to d A random integer between rand is a random number between 0 and 1. is the value of the jth dimension in the i-th layout mode, New is the value of the jth dimension in the newly generated i-th layout. This search method can help the intelligent algorithm jump out of the local optimal solution, enhance the global exploration ability, and continue to explore better solutions.

[0055] In summary, the adaptive selection mechanism of wind farm layout in the designed intelligent algorithm is embodied in the following formula:

[0056]

[0057] After the search of the current dimension is completed, the fitness value of the new layout is calculated, and the fitness values ​​of the layout modes before and after the update are compared. The survival of the fittest is eliminated, and the matrix is ​​updated. X , OX , S ,and OS , and save the optimal layout mode and optimal value. At this point, the algorithm has gone through one iteration, and the number of iterations increases by one. It is determined whether the iteration stop condition is met at this time. If it is met, the iteration is stopped. If not, S42-S43 is repeated until the iteration stop condition is met.

[0058] S44, outputting the optimal objective function value solved by the algorithm during the search process, and decoding the layout mode to save the layout position of each turbine corresponding to the optimal objective function value.

[0059] In the step S5, optimal layout scheme diagrams under different wind farm experimental scenarios are drawn.

[0060] The visualization software is used to visualize the layout position of each turbine in the different wind farm experimental scenarios saved. The specific results are presented on the corresponding cells in the grid, that is, the location where the turbine is placed.

[0061] A computer device comprises a memory, a processor and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of the wind farm layout optimization method based on the intelligent algorithm as described above are implemented.

[0062] A computer-readable storage medium stores a computer program / instruction, which, when executed by a processor, implements the steps of the wind farm layout optimization method based on an intelligent algorithm.

[0063] Beneficial effects: Compared with the prior art, the wind farm layout optimization method based on intelligent algorithm provided by the present invention has the following advantages:

[0064] (1) Since wind farm layout is a classic NP-hard problem, it is difficult for traditional classical optimization algorithms to effectively solve it. This intelligent algorithm can solve a more reasonable wind farm layout plan. Specifically, under the same conditions, the optimal wind farm layout plan solved by the optimization method provided by the present invention can maximize the wind farm output power while minimizing the wind farm cost.

[0065] (2) The optimization method of the present invention designs more efficient local and global search strategies. Therefore, the optimal wind farm layout solution solved by the proposed intelligent algorithm has a lower computational cost and generates greater economic and social value. BRIEF DESCRIPTION OF THE DRAWINGS

[0066] Figure 1 is a flow chart of the method principle of an embodiment of the present invention;

[0067] Figure 2 The figure is a flow chart of an intelligent algorithm according to an embodiment of the present invention. DETAILED DESCRIPTION

[0068] The present invention is further explained below in conjunction with specific embodiments. It should be understood that these embodiments are only used to illustrate the present invention and are not used to limit the scope of the present invention. After reading the present invention, various equivalent forms of modifications to the present invention by those skilled in the art all fall within the scope defined by the claims attached to this application.

[0069] like Figure 1 As shown, a wind farm layout optimization method based on an intelligent algorithm includes the following steps:

[0070] Step S1, analyzing the wake impact of the wind farm and performing wake modeling;

[0071] Step S2, based on the wind farm wake analysis, two types of wind farm experimental scenarios that conform to the actual situation are proposed, and simulation experiments considering the number of wind turbines installed in the wind farm are further designed under two different wind farm layout conditions, and two experiments are designed respectively according to the two situations of certain and uncertain number of turbines; a gridding method is used to approximately and reasonably simulate the placement of wind turbines in an actual wind farm; and encoding is performed in combination with the gridding method and different layouts of wind turbines.

[0072] Step S3, comprehensively considering the factors affecting the construction of the wind farm, selecting two most important factors to design an objective function for evaluating the quality of the wind farm layout plan.

[0073] Step S4, design an intelligent algorithm to optimize the wind farm layout, and each wind farm layout method in the algorithm is iteratively updated with the goal of finding the minimum value of the objective function in step S3. When the algorithm meets the iteration stop condition, the wind farm layout method that obtains the minimum value of the objective function is retained, and then the information it carries is parsed into the layout position of the wind turbine in the gridded simulated wind farm through a corresponding decoding method, and the wind farm construction, operation cost and output power corresponding to the layout scheme are output.

[0074] Step S5, based on the above results, draw the optimal layout scheme diagram under different wind farm experimental scenario assumptions.

[0075] In step S1, the wake impact analysis of the wind farm and the wake modeling include:

[0076] S11, in a wind farm, when the downstream turbine is only affected by the wake from a single upstream turbine, assuming that the initial wind speed flowing to the turbine blades is uniform, after being affected by the wake, the distance from the turbine x The wind wake velocity at (m) is calculated by the following formula:

[0077]

[0078] Where V0 is the initial wind speed toward the turbine, D T is the turbine disk diameter, k is the wake attenuation constant, which is set to 0.654 based on extensive studies of wind farm layout simulation experiments.

[0079] S12, when the downstream turbine is affected by the wake from multiple upstream turbines, assuming that the initial wind speed flowing to the turbine blades is uniform, after being affected by the wake, the distance from the turbine x The wind wake speed at meters is calculated by the following formula:

[0080]

[0081] In the formula, A cover It represents the overlapping area affected by the wakes of multiple turbines, and A0 is the area of ​​one rotation of the turbine blades.

[0082] In step S2, a grid method is used to design wind farm simulation experiments under different conditions, and encoding and decoding methods are designed, including:

[0083] S21, based on the common wind farm grid design, models the actual wind farm construction area as a regular shape, such as a 10×10 square grid, with the turbine placement at the center of each cell.

[0084] In the simulation experiment, it is assumed that the wind farm area is 2km×2km and the turbine rotor diameter is 40m, so the size of each cell is 200m×200m, and the area of ​​each cell is equivalent to 5 turbine rotor diameters. In order to ensure the normal operation of the turbine, the separation distance between adjacent turbines is 5 times the rotor diameter.

[0085] S22, according to the relative positions of the upstream and downstream turbines, the experimental scenarios are divided into two categories according to the influence of the wake. The first category is that the downstream turbine is placed inside the wake coverage area of ​​the upstream turbine, and the wind energy received by the downstream turbine is affected by the full coverage of the upstream turbine wake; the second category is that the downstream turbine is placed at the edge of the wake coverage area of ​​the upstream turbine, and the wind energy received by the downstream turbine is partially affected by the wake of the upstream turbine.

[0086] S23, based on S22, the experiments in each case are divided into two types according to whether the number of turbines is determined when building the wind farm. The first type is when the number of turbines is determined, and in this case, it is only necessary to solve how to arrange the positions of the turbines to maximize the output power of the wind farm; the second type is when the number of turbines is uncertain, and in this case, it is necessary to solve the number and arrangement of turbines that can maximize the output power while minimizing the cost. Therefore, the experimental assumptions in the present invention are the following four cases:

[0087] (1) The downstream turbine is placed inside the wake coverage area of ​​the upstream turbine. The wind energy received by the downstream turbine is affected by the full coverage of the upstream turbine wake. The number of wind turbines is an integer between [1,100].

[0088] (2) The downstream turbine is placed at the edge of the upstream turbine wake coverage area. The wind energy received by the downstream turbine is partially affected by the upstream turbine wake. The number of wind turbines is an integer between [1,100].

[0089] (3) The downstream turbines are placed inside the wake coverage area of ​​the upstream turbines. The wind energy received by the downstream turbines is affected by the full coverage of the upstream turbine wake. The number of wind turbines is set according to actual needs.

[0090] (4) The downstream turbines are placed at the edge of the upstream turbine wake coverage area. The wind energy received by the downstream turbines is partially affected by the upstream turbine wake. The number of wind turbines is set according to actual needs.

[0091] S24, based on S21-S23, encode and decode the turbine layout. When the number of turbines is uncertain, set the variableα is a 1×100 vector, representing 100 positions in a 10×10 square grid. All solution vectors are in the range [0,1] and are rounded to integers, i.e., 0 or 1. A solution with a value of 1 represents a location where a turbine is placed, and a solution with a value of 0 represents a location where no turbine is placed. When the number of turbines is determined, α for n ×1 vector, n is the number of turbines, all solution vectors are in the range [1,100] and are rounded to integers, which are the cell locations where the turbines are installed; then create a G c = a list of {1, 2, ..., 100}, if α The i-th element in is rounded ( α i ) is lower than G c The number of remaining elements in the grid is G c (round( α i )) is designated as i turbine locations and from the list G c Delete the selected grid position from the G c The last element in is designated as i The location of a turbine is determined and removed from the list.

[0092] In step S3, the factors affecting the construction of wind farms are comprehensively considered, and two main factors are selected to design an objective function for evaluating the quality of wind farm layout schemes, including:

[0093] S31, wind farm construction and operation costs are calculated by the following formula:

[0094]

[0095] In the formula, N is the number of turbines.

[0096] S32, the wind farm output power is obtained by the following formula:

[0097]

[0098] Where η is the turbine power generation efficiency, ρ is the air density, A0 is the area of ​​the turbine blades during one rotation, V xi is the distance from the upstream turbine x (m) flows to iThe wind wake velocity of the downstream turbine.

[0099] S33, based on S31 and S32, the objective function for evaluating the quality of wind farm layout scheme is calculated by the following formula:

[0100] Objective Function=min

[0101] The smaller the objective function value is, the higher the quality of the wind farm layout solution corresponding to the solution is, that is, it can maximize the output power while minimizing the cost.

[0102] like Figure 2 As shown, in step S4, using the designed intelligent algorithm to optimize the layout of the wind farm specifically includes:

[0103] S41, based on steps S31 and S32, input wind farm simulation experiment parameters to improve the modeling of wind farm layout optimization problem, including wind farm area size, hub height, rotor diameter, air density, ground roughness, initial wind speed and wind direction. Initialize the population information of the intelligent algorithm and store the initialized layout mode in the following matrix X middle:

[0104]

[0105] Among them, the matrix X Each row in represents a layout of a wind farm. NP is the population size, which can be set to 50. d is the dimension of the wind farm layout problem to be solved. Based on the S24 simulation experiments (1) and (2) d The size is 100, and the simulation experiments (3) and (4) d The size is 39. Set the algorithm iteration termination condition to reach the maximum number of algorithm iterations, such as setting it to 2500.

[0106] S42, based on S33, calculates the fitness value of each layout mode in the algorithm population and stores it in the following matrix OX middle:

[0107]

[0108] In the formula, OX NP Representative NP The fitness value of the layout mode. Then, OX Sort by size, and then save them in the following matrix from best to worst according to the fitness value. S middle:

[0109]

[0110] The corresponding sorted and updated layout fitness values ​​are stored in the following matrix OS middle:

[0111]

[0112] S43, calculating the probability of each layout being selected. The probability of each layout being selected is calculated to distinguish between layouts with good fitness values ​​and layouts with poor fitness values. The probability of being selected is obtained as follows:

[0113] SP i =

[0114] Where, SP i Represents the algorithm i The probability of selecting a layout mode is i It is the sorting position of the layout after sorting by fitness value. NP is the population size, SP i The larger the layout, the better, which means that this layout may be closer to the optimal solution. Then the corresponding wind farm layout is efficiently searched locally according to the following formula:

[0115] X i =|SX i |.e t .cos(2πt)+S

[0116] Among them, X i is the i-th layout method in the intelligent algorithm population, S represents the optimal layout method found by the current intelligent algorithm, and the parameter t The value range is [ r ,1], r During the iteration process, it decreases linearly in the range of [-2, -1], t The size of X i The distance to S is positively correlated. This update method can make X i In the early stages of the iteration, the entire search space is covered as much as possible. As the iteration proceeds, the local search capability of the algorithm gradually increases.

[0117] Otherwise, the wind farm layout is efficiently searched globally according to the following formula:

[0118]

[0119] Among them, NewX i is the i-th layout method newly generated by the intelligent algorithm, X r1 , X r2 and Xr3 are three layout methods randomly selected from the intelligent algorithm population, and X r1 The fitness value is better than X r2 , X r3 , j represents the jth dimension of the wind farm layout problem to be solved, j=1, 2, …, d; d is the dimension of the wind farm layout problem, j rand is 1 to d A random integer between rand is a random number between 0 and 1. is the value of the jth dimension in the i-th layout mode, New is the value of the jth dimension in the newly generated i-th layout. This search method can help the intelligent algorithm jump out of the local optimal solution, enhance the global exploration ability, and continue to explore better solutions.

[0120] In summary, the adaptive selection mechanism of wind farm layout in the designed intelligent algorithm is embodied in the following formula:

[0121]

[0122] After the search of the current dimension is completed, the fitness value of the new layout is calculated, and the fitness values ​​of the layout modes before and after the update are compared. The survival of the fittest is eliminated, and the matrix is ​​updated. X , OX , S ,and OS , and save the optimal layout mode and optimal value. At this point, the algorithm has gone through one iteration, and the number of iterations increases by one. It is determined whether the iteration stop condition is met at this time. If it is met, the iteration is stopped. If not, S42-S43 is repeated until the iteration stop condition is met.

[0123] S44, outputting the optimal objective function value solved by the algorithm during the search process, and saving the layout position of each turbine corresponding to the optimal objective function value according to the decoding method.

[0124] In step S5, drawing the optimal layout scheme diagram under different wind farm experimental scenarios includes:

[0125] The visualization software is used to visualize the layout of each turbine in the different wind farm experimental scenarios saved. The specific results are presented on the corresponding cells in the grid, that is, the location where the turbine is placed.

[0126] Obviously, those skilled in the art should understand that the various steps of the wind farm layout optimization method based on intelligent algorithms in the above-mentioned embodiments of the present invention can be implemented by a general computing device, they can be concentrated on a single computing device, or distributed on a network composed of multiple computing devices, optionally, they can be implemented by a program code executable by a computing device, so that they can be stored in a storage device and executed by the computing device, and in some cases, the steps shown or described can be executed in a different order than here, or they can be made into individual integrated circuit modules, or multiple modules or steps therein can be made into a single integrated circuit module for implementation. In this way, the embodiments of the present invention are not limited to any specific combination of hardware and software.

Claims

1. A wind farm layout optimization method based on intelligent algorithm, characterized in that: The following steps are involved: Step S1, analyzing the wake impact of the wind farm and performing wake modeling; Step S2: Based on the wind farm wake analysis, two wind farm experimental scenarios that conform to actual conditions are proposed. Simulation experiments are designed under two different wind farm layouts to consider the number of wind turbines in the wind farm. Two experiments are designed for the two cases where the number of turbines is determined and the number of turbines is uncertain. A gridding method is used to simulate the placement of wind turbines in an actual wind farm. Coding is performed by combining the gridding method with the different wind turbine layouts. Step S3, designing an objective function for evaluating the quality of wind farm layout plans; Step S4: Design an intelligent algorithm to optimize the wind farm layout, iteratively update the layout with the minimum value of the objective function for each wind farm layout as the target, retain the layout with the minimum objective function, and then decode the information it carries into the layout position of the wind turbines in the gridded simulated wind farm through a corresponding decoding method, and output the wind farm construction, operation costs and output power corresponding to the layout scheme; In step S2, a gridding method is used to design wind farm simulation experiments under different conditions, and encoding and decoding methods are designed, specifically: S21, the actual wind farm construction area is modeled as a regular shape, the regular shape is a shape consisting of K × K grids, and the turbine is placed at the center of each cell; S22, based on the relative positions of the upstream and downstream turbines, the experimental scenarios are divided into two categories according to the wake effect. The first category is that the downstream turbine is placed inside the upstream turbine wake coverage area, and the wind energy received by the downstream turbine is fully affected by the upstream turbine wake. The second category is that the downstream turbine is placed at the edge of the upstream turbine wake coverage area, and the wind energy received by the downstream turbine is partially affected by the upstream turbine wake. S23, based on whether the number of turbines is certain when constructing the wind farm, the experiments for each case are divided into two types: the first type is when the number of turbines is certain, in which case the only question is how to arrange the turbines to maximize the output power of the wind farm; the second type is when the number of turbines is uncertain, in which case the question is how to arrange the turbines in a way that minimizes cost while maximizing the output power. S24, based on S21-S23, encode and decode the turbine layout. When the number of turbines is uncertain, set the variable α 1×K 2 A vector representing K in a K×K grid 2 positions, all solution vectors are in the range [0,1] and are rounded to integers, i.e. 0 or 1. A solution with a value of 1 represents a position for placing a turbine, and a solution with a value of 0 represents a position where no turbine is placed. When the number of turbines is determined, α for n ×1 vector, n is the number of turbines, all solution vectors are in [1, K 2 ] and rounded to an integer, which is the unit cell location where the turbine is installed; then create a G c ={1, 2, ..., K 2 }, if α The i-th element in is rounded ( α i ) is lower than G c The number of remaining elements in the grid is G c (round( α i )) Designated as i The location of the turbines and from the list G c Delete the selected grid position, otherwise G c The last element in the i The location of the turbine is determined and removed from the list.

2. The wind farm layout optimization method based on intelligent algorithm according to claim 1 is characterized in that: The analysis of the wind farm wake impact and the wake modeling in step S1 include: S11, in a wind farm, assuming that the initial wind speed flowing to the turbine blades is uniform, when the downstream turbine is only affected by the wake from a single upstream turbine, give the calculated distance from the turbine x The formula for the wind wake velocity at meters is used as the wake model; S12, when the downstream turbine is affected by the wake from multiple upstream turbines, assuming that the initial wind speed flowing to the turbine blades is uniform, after the wake is affected, the calculated distance turbine is given x The wind wake speed at meters is used as the wake model.

3. The wind farm layout optimization method based on intelligent algorithm according to claim 1, characterized in that: In step S23, the experiment includes the following four cases: (1) The downstream turbines are placed inside the upstream turbine wake coverage area. The wind energy received by the downstream turbines is affected by the full coverage of the upstream turbine wake. The number of wind turbines is [1, K 2 ]; (2) The downstream turbines are placed at the edge of the upstream turbine wake coverage area. The wind energy received by the downstream turbines is partially affected by the upstream turbine wake. The number of wind turbines is [1, K 2 ]; (3) The downstream turbines are placed inside the wake coverage area of ​​the upstream turbines. The wind energy received by the downstream turbines is affected by the full coverage of the upstream turbine wake. The number of wind turbines is set according to actual needs; (4) The downstream turbines are placed at the edge of the upstream turbine wake coverage area. The wind energy received by the downstream turbines is partially affected by the upstream turbine wake. The number of wind turbines is set according to actual needs.

4. The wind farm layout optimization method based on intelligent algorithm according to claim 1, characterized in that: In step S3, the factors affecting the construction of the wind farm are comprehensively considered, and the construction and operation costs of the wind farm are selected. C total and wind farm output power P total Two main factors are used to design the objective function for evaluating the quality of wind farm layout solutions: Objective Function=min 。 5. The wind farm layout optimization method based on intelligent algorithm according to claim 1, characterized in that: In step S4, optimizing the wind farm layout using the designed intelligent algorithm specifically includes: S41, input wind farm simulation experiment parameters, including wind farm area size, hub height, rotor diameter, air density, ground roughness, initial wind speed and wind direction; initialize the population information of the intelligent algorithm, and store the initialized wind farm layout in the matrix X In the matrix X Each row in represents a wind farm layout method, setting the maximum number of iterations of the intelligent algorithm; S42, based on the objective function of evaluating the quality of wind farm layout scheme, the effect of each wind farm layout method in the algorithm, i.e., the fitness value, is stored in the matrix OX middle; Then, the matrix OX The elements in are sorted by size, and after sorting, the layout is saved in the matrix from best to worst according to the fitness value. S In the , the layout fitness value after sorting and updating is saved in the matrix OS middle; S43, calculating the probability of each wind farm layout being selected. The selection probability is obtained as follows: SP i = Where, SP i Represents the first i The probability of the selected layout is i It is the sorting position of the layout after sorting by fitness value, NP is the population size of the intelligent algorithm; The adaptive selection mechanism of wind farm layout in the designed intelligent algorithm is reflected in the following formula: ; After the search of the current dimension is completed, the fitness value of the new layout method is calculated, and the fitness values ​​of the layout methods before and after the update are compared. The survival of the fittest is eliminated, and the matrix is ​​updated. X , OX , S ,and OS , save the optimal layout mode and optimal value; at this point, the algorithm has gone through one iteration, the number of iterations increases by one, and it is judged whether the iteration stop condition is met at this time. If it is met, the iteration is stopped. If not, S42-S43 are repeated until the iteration stop condition is met; S44, outputting the optimal objective function value solved by the algorithm during the search process, and decoding the layout mode to save the layout position of each turbine corresponding to the optimal objective function value.

6. The wind farm layout optimization method based on intelligent algorithm according to claim 1, characterized in that: After obtaining the optimal wind farm layout, draw the optimal layout plan diagram under different wind farm experimental scenarios; Visualization software is used to visualize the layout position of each turbine in different saved wind farm experimental scenarios. The specific results are presented on the corresponding cells in the grid, that is, the location where the turbine is placed.

7. A computer device, characterized in that: The computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of the wind farm layout optimization method based on the intelligent algorithm as described in any one of claims 1 to 6 are implemented.

8. A computer-readable storage medium having a computer program / instruction stored thereon, characterized in that: When the computer program / instruction is executed by a processor, the steps of the wind farm layout optimization method based on the intelligent algorithm according to any one of claims 1 to 6 are implemented.