Wind turbine generator unit optimal layout method based on three-level dynamic variation rate

Abstract

The application discloses a kind of wind turbine unit optimization layout method based on three levels dynamic variation rate, comprising the following steps: step one: obtaining target wind farm information, including (1) the annual average wind resource information of target wind farm;(2) unit specification parameter;(3) wind farm planning area boundary;Step two: combined with fan wake wind speed and turbulence model, calculate annual total average power generation and evaluate full life cycle degree electric cost;Step three: the wind farm planning area is gridded, with the optimization target of flat standardization energy cost, carries out unit grid location preliminary optimization using the improved genetic optimization algorithm based on three levels dynamic variation rate, obtains several grid optimization schemes by integration theory;Step four: according to the preliminary optimization result of grid location, constructs the grid boundary of each candidate unit, carries out fine coordinate position optimization in corresponding grid using adaptive weight particle swarm optimization algorithm, and obtains the final unit layout scheme.

Description

Technical Field

[0001] This invention belongs to the field of wind power generation technology, specifically a method for optimizing the layout of wind turbine generator sets based on a three-level dynamic variation rate. Background Technology

[0002] With the depletion of traditional fossil fuels, the demand for wind energy development is increasing. To maximize the power generation efficiency of wind farms, one of the core issues is to rationally optimize the layout of turbine units before construction, thereby reducing the levelized cost of electricity (LCOE) throughout the wind farm's lifecycle. However, due to the wake effects between turbine units, the optimization process for wind turbine layout is extremely complex, making it difficult to obtain a globally optimal layout. Genetic algorithms and particle swarm optimization (PSO) are the most commonly used combined optimization algorithms and have been widely applied to solve wind turbine layout optimization problems. However, existing single genetic and PSO algorithms calculate for all individuals in the population, which is very time-consuming, and the fixed mutation rate makes the algorithms prone to getting trapped in local optima, resulting in non-optimal wind turbine layouts. This leaves significant room for improvement in layout optimization. Therefore, improving the efficiency of optimization algorithms while simultaneously obtaining the globally optimal solution is crucial for enhancing the effectiveness of wind turbine layout optimization. Summary of the Invention

[0003] In view of this, the purpose of this invention is to provide a wind turbine generator set optimization layout method based on a three-level dynamic variation rate, which can achieve reasonable optimization of the generator set layout scheme to reduce the levelized cost of electricity (LCOE) of the wind farm throughout its entire life cycle.

[0004] To achieve the above objectives, the present invention provides the following technical solution:

[0005] A method for optimizing the layout of wind turbine generator sets based on a three-level dynamic variation rate includes the following steps:

[0006] Step 1: Obtain information about the target wind farm, including:

[0007] (1) Annual average wind resource information of the target wind farm, and draw a rose diagram based on the annual statistical wind speed and wind direction information in the target wind farm area;

[0008] (2) Unit specifications, including hub height, impeller diameter and power-thrust curve;

[0009] (3) Boundary of wind farm planning area, delineate the boundary of wind farm planning area including regular or irregular boundaries;

[0010] Step 2: Calculate the annual total average power generation and assess the levelized cost of electricity (LCOE) over the entire life cycle by combining the wind turbine wake velocity and turbulence model;

[0011] Step 3: Grid the wind farm planning area, take levelized energy cost as the optimization objective, and use an improved genetic optimization algorithm based on three-level dynamic mutation rate to perform preliminary optimization of the grid location of the units. Several grid optimization schemes are obtained through ensemble theory.

[0012] Step 4: Based on the preliminary optimization results of the grid positions, construct the grid boundary for each candidate unit, and use the adaptive weighted particle swarm optimization algorithm to perform fine coordinate position optimization within the corresponding grid, and obtain the final unit layout scheme.

[0013] Furthermore, in step two, the method for calculating the annual average total power generation and assessing the levelized cost of electricity (LCOE) over the entire life cycle is as follows:

[0014] 21) Based on the Gaussian wind turbine wake model and turbulence model, calculate the wind speed loss and turbulence enhancement of each wind turbine after being affected by the wake of the upstream wind turbine;

[0015] 22) The inflow wind speed at the hub height of each downstream wind turbine is estimated using the classical wake superposition principle. The annual power generation of each wind turbine is calculated. The annual power generation of each wind turbine is summed to obtain the total annual power generation of the wind farm.

[0016] 23) Calculate the annual average distribution of wind speed across all wind directions and the total power generation of the wind farm over its entire life cycle. Combine this with economic benefit indicators, including the cost of generator components, wind farm operation and maintenance costs, and capital recovery coefficient, to construct a levelized energy cost objective function.

[0017] Furthermore, in step 21), the Gaussian wind turbine wake model is:

[0018]

[0019]

[0020]

[0021]

[0022] Where δ represents the wake wind speed loss; δ hub σ is the wake velocity loss at hub height; σ is the standard deviation of the wake velocity profile; D is the impeller diameter; y and z are the horizontal and vertical coordinates, respectively; z hub It is the height of the wind turbine hub; C T It is the thrust coefficient; k * It is the wake attenuation coefficient; x represents the horizontal distance between the downstream fan and the upstream fan; ε represents a parameter related to the thrust coefficient;

[0023] The calculation method for wind turbine wake turbulence is as follows:

[0024]

[0025] Among them, I + It is an additional wake turbulence; K n These are model constants;

[0026] The superimposed wake of the wind turbine is as follows:

[0027]

[0028] Among them, v i and v j v0 is the inflow velocity at the upstream and downstream fans; v0 is the inflow velocity at infinity; v0 is the inflow velocity at infinity. ij This refers to the inflow velocity caused by the upstream fan at the downstream fan location.

[0029] Furthermore, in step 22), given the incoming flow conditions of velocity v and direction θ, the total power generated by the wind farm can be calculated based on the unit power curve. Combining this with the probability density function of the statistical average wind conditions, the annual average power generation is predicted as follows:

[0030]

[0031] Where, N t N is the number of wind turbines; θ It refers to the number of wind direction zones; N v It refers to the number of wind speed ranges; f j It represents the probability of wind direction range; P i At wind speed v k The power output of a single generating unit is specified; (X,Y) represents the unit's coordinate information; p j The joint distribution probability of wind speed and wind direction, v k It is the dividing point of the wind speed range; θ j This represents the wind direction angle represented by the j-th wind speed.

[0032] Furthermore, in step 23), the levelized energy cost objective function is:

[0033]

[0034] Among them, C total It is the cost of the generator unit; C f It is the capital recovery coefficient; C O&M It is the operation and maintenance cost of the wind farm; AEP is the average annual total power generation of the wind farm.

[0035] Furthermore, step three includes the following steps:

[0036] 31) Divide the wind farm planning area into grid distribution maps of different sizes, and the coordinate position of the candidate units is the grid center by default;

[0037] 32) An improved genetic optimization algorithm with a three-level dynamic mutation rate is proposed, which simultaneously considers the current number of wind turbines installed, the repetition rate of individuals in the population, and the improved mutation rate of the current iteration step, in order to improve the search capability for the global optimum, avoid getting trapped in local optima, and perform preliminary optimization of the gridded layout scheme. Several grid optimization schemes are obtained through ensemble theory.

[0038] Furthermore, in step 32), the method for preliminary optimization of the unit grid location using an improved genetic optimization algorithm based on a three-level dynamic mutation rate is as follows:

[0039] 321) Assume that each grid center has two possibilities: installing a fan or not installing a fan. The corresponding "genes" are encoded as "1" and "0" respectively. Each individual corresponds to a fan arrangement scheme and is composed of a string of "genes".

[0040] 322) Randomly initialize the population: For each "gene", randomly assign an amplitude of "1" or "0", and iterate i times. step =0;

[0041] 323) Calculate the fitness of individuals based on the uniqueness process to obtain a population of completely unique individuals;

[0042] Determine if the individual fitness is less than a set threshold: if yes, then the wind turbine layout grid optimization scheme is obtained; otherwise, i step =i step +1, proceed to step 324);

[0043] 324) Random individual selection is performed using the roulette wheel selection method;

[0044] 325) Random crossover is performed between every two individuals, which means the "genes" at corresponding positions in the two individuals are interchanged;

[0045] 326) Random mutations are performed on the "genes" of each individual;

[0046] 327) Determine i step Is it equal to the set value N? s If yes, then the wind turbine layout grid optimization scheme is obtained; otherwise, step 323 is executed repeatedly.

[0047] Furthermore, in step 326), an improved mutation rate is used to randomly mutate the "genes" of each individual, and the improved mutation rate is:

[0048]

[0049]

[0050] Where p0 is the mutation rate of the 0-encoded position; p1 is the mutation rate of the 1-encoded position; and N is the number of grids. is the average population size; 1 and 2 represent the initial mutation rate and the final mutation rate, respectively; represents the repetition rate of individuals in the population; represents the number of steps in the current iteration; represents the total number of iterations.

[0051] Furthermore, step four includes the following steps:

[0052] 41) Calculate the center coordinates and boundary conditions of the grids containing the candidate units in all grid optimization schemes;

[0053] 42) Use the unit center coordinates of the grid optimization scheme as the initial particle population;

[0054] 43) Using boundary conditions as constraints and leveled energy cost as the optimization objective, calculate the particle fitness function;

[0055] Determine if the particle fitness function is less than a set threshold: if yes, obtain the wind turbine layout scheme; if no, increment the iteration count by 1 and proceed to step 44).

[0056] 44) Determine if the number of iterations has reached the set maximum number of iterations: if yes, obtain the wind turbine generator layout scheme; if no, use the adaptive weighted particle swarm optimization algorithm to update the particle population and execute step 43).

[0057] Furthermore, in step 44), the adaptive weighted particle swarm optimization algorithm is as follows:

[0058]

[0059]

[0060] in, It is the position of the i-th particle at the t-th iteration; r1 and r2 are the velocity vectors of the i-th particle in the t-th iteration; r1 and r2 are random numbers between 0 and 1; p i and p g These are the best candidate solutions found for the i-th particle and the entire population, respectively; β i and β g These are custom parameters that control the particle's ability to develop and explore motion, respectively; ω a It is a nonlinear dynamic inertial weight that is adjusted based on the distance to the global optimal solution during the optimization process.

[0061] The beneficial effects of this invention are as follows:

[0062] This invention presents a wind turbine layout optimization method based on a three-level dynamic variation rate. Addressing the issues of low computational efficiency and susceptibility to local optima in wind turbine layout optimization, it proposes a two-stage optimization approach combining grid and coordinate optimization. This improves computational efficiency while maintaining high accuracy. Furthermore, the three-level dynamic variation rate is used to minimize the risk of getting trapped in local optima during the optimization process. The proposed efficient optimization algorithm can significantly yield better wind turbine optimization schemes than traditional single algorithms. This method greatly improves the computational efficiency of optimization solutions while striving to obtain globally optimal wind turbine layout results. In short, this invention's wind turbine layout optimization method based on a three-level dynamic variation rate enables reasonable optimization of turbine layout schemes, thereby reducing the levelized cost of electricity (LCOE) of wind farms throughout their entire lifecycle. Attached Figure Description

[0063] To make the objectives, technical solutions, and beneficial effects of this invention clearer, the following figures are provided for illustration:

[0064] Figure 1 This is a flowchart of the wind turbine generator layout optimization method disclosed in this invention;

[0065] Figure 2 Flowchart of the genetic particle swarm optimization algorithm;

[0066] Figure 3 Map showing the planned area for the target wind farm;

[0067] Figure 4 Comparison diagrams of wind turbine layout optimization schemes: (a) Wind turbine layout diagram obtained by genetic algorithm; (b) Wind turbine layout diagram obtained by particle swarm optimization algorithm; (c) Wind turbine layout diagram obtained by the fusion algorithm of this invention.

[0068] Figure 5 Box plots were plotted based on statistical calculations and analysis of the results from 10 optimization processes.

[0069] This specific implementation method

[0070] The present invention will be further described below with reference to the accompanying drawings and specific embodiments, so that those skilled in the art can better understand and implement the present invention. However, the embodiments described are not intended to limit the present invention.

[0071] like Figure 1 As shown, this embodiment of the wind turbine generator set optimization layout method based on three-level dynamic variation rate includes the following steps:

[0072] Step 1: Obtain information about the target wind farm, including:

[0073] (1) Annual average wind resource information of the target wind farm, and draw a rose diagram based on the annual statistical wind speed and wind direction information in the target wind farm area;

[0074] (2) Unit specifications, including hub height, impeller diameter and power-thrust curve;

[0075] (3) Boundary of wind farm planning area: Delineate the boundary of wind farm planning area, including regular or irregular boundaries.

[0076] Step Two: Calculate the annual total average power generation and assess the levelized cost of electricity (LCOE) over the entire lifecycle by combining the wind turbine wake velocity and turbulence model. Specifically, the method for calculating the annual total average power generation and assessing the LCOE is as follows:

[0077] 21) Based on the Gaussian wind turbine wake model and turbulence model, calculate the wind speed loss and turbulence enhancement of each wind turbine after being affected by the wake of the upstream wind turbine.

[0078] Specifically, the wake model of a Gaussian wind turbine is as follows:

[0079]

[0080]

[0081]

[0082]

[0083] Where δ represents the wake wind speed loss; δ hub σ is the wake velocity loss at hub height; σ is the standard deviation of the wake velocity profile; D is the impeller diameter; y and z are the horizontal and vertical coordinates, respectively; z hub It is the height of the wind turbine hub; C T It is the thrust coefficient; k * It is the wake attenuation coefficient; x represents the horizontal distance between the downstream fan and the upstream fan; ε represents a parameter related to the thrust coefficient;

[0084] The calculation method for wind turbine wake turbulence is as follows:

[0085]

[0086] Among them, I + It is an additional wake turbulence; K n These are model constants;

[0087] The superimposed wake of the wind turbine is as follows:

[0088]

[0089] Among them, v i and v j v0 is the inflow velocity at the upstream and downstream fans; v0 is the inflow velocity at infinity; v0 is the inflow velocity at infinity. ijThis refers to the inflow velocity caused by the upstream fan at the downstream fan location.

[0090] 22) The inflow wind speed at the hub height of each downstream wind turbine is estimated using the classical wake superposition principle. The annual power generation of each wind turbine is calculated, and the annual power generation of each wind turbine is summed to obtain the total annual power generation of the wind farm.

[0091] Specifically, given the incoming flow conditions of velocity v and direction θ, the total power generated by the wind farm can be calculated based on the unit power curve. Combining this with the probability density function of the statistical average wind conditions, the annual average power generation is predicted as follows:

[0092]

[0093] Where, N t N is the number of wind turbines; θ It refers to the number of wind direction zones; N v It refers to the number of wind speed ranges; f j It represents the probability of wind direction range; P i At wind speed v k The power output of a single generating unit is specified; (X,Y) represents the unit's coordinate information; p j The joint distribution probability of wind speed and wind direction, v k It is the dividing point of the wind speed range; θ j This represents the wind direction angle represented by the j-th wind speed.

[0094] 23) Calculate the annual average distribution of wind speed across all wind directions and the total power generation of the wind farm over its entire life cycle. Combine this with economic benefit indicators, including the cost of generator components, wind farm operation and maintenance costs, and capital recovery coefficient, to construct a levelized energy cost objective function.

[0095] Specifically, the levelized cost of energy objective function is:

[0096]

[0097] Among them, C total It is the cost of the generator unit; C f It is the capital recovery coefficient; C O&M It is the operation and maintenance cost of the wind farm; AEP is the average annual total power generation of the wind farm.

[0098] Step 3: The wind farm planning area is gridded. Using levelized cost of energy (LCOE) as the optimization objective, an improved genetic optimization algorithm based on a three-level dynamic mutation rate is employed for initial optimization of the unit grid locations. Several grid optimization schemes are obtained through ensemble theory. This includes the following steps:

[0099] 31) Divide the wind farm planning area into grid distribution maps of different sizes, and the coordinate position of the candidate unit is the grid center by default.

[0100] 32) An improved genetic optimization algorithm with a three-level dynamic mutation rate is proposed, which simultaneously considers the current number of wind turbines installed, the repetition rate of individuals in the population, and the improved mutation rate of the current iteration step, in order to improve the search capability for the global optimum, avoid getting trapped in local optima, and perform preliminary optimization of the gridded layout scheme. Several grid optimization schemes are obtained through ensemble theory.

[0101] Specifically, such as Figure 2 As shown, the method for preliminary optimization of unit grid location using an improved genetic optimization algorithm based on a three-level dynamic mutation rate is as follows:

[0102] 321) Assume that each grid center has two possibilities: installing a fan or not installing a fan. The corresponding "genes" are encoded as "1" and "0" respectively. Each individual corresponds to a fan arrangement scheme and is composed of a string of "genes".

[0103] 322) Randomly initialize the population: For each "gene", randomly assign an amplitude of "1" or "0", and iterate i times. step =0;

[0104] 323) Calculate the fitness of individuals based on the uniqueness process to obtain a population of completely unique individuals;

[0105] Determine if the individual fitness is less than a set threshold: if yes, then the wind turbine layout grid optimization scheme is obtained; otherwise, i step =i step +1, proceed to step 324);

[0106] 324) Random individual selection is performed using the roulette wheel selection method;

[0107] 325) Random crossover is performed between every two individuals, which means the "genes" at corresponding positions in the two individuals are interchanged;

[0108] 326) Random mutations are performed on the "genes" of each individual;

[0109] 326) Determine i step Is it equal to the set value N? s If yes, then the wind turbine layout grid optimization scheme is obtained; otherwise, step 323 is executed repeatedly.

[0110] In this embodiment, an improved mutation rate is used to randomly mutate the "genes" of each individual. The improved mutation rate is:

[0111]

[0112]

[0113] Where p0 is the mutation rate of the 0-encoded position; p1 is the mutation rate of the 1-encoded position; and N is the number of grids. is the average population size; 1 and 2 represent the initial mutation rate and the final mutation rate, respectively; represents the repetition rate of individuals in the population, obtained by dividing the number of unique individuals by the total population size; represents the number of steps in the current iteration; represents the total number of iteration steps.

[0114] Step 4: Based on the preliminary optimization results of the grid positions, construct the grid boundaries for each candidate unit. Then, use an adaptive weighted particle swarm optimization algorithm to perform fine-grained coordinate position optimization within the corresponding grid, obtaining the final unit layout scheme. For example... Figure 2 As shown, it includes the following steps:

[0115] 41) Calculate the center coordinates and boundary conditions of the grids containing the candidate units in all grid optimization schemes;

[0116] 42) Use the unit center coordinates of the grid optimization scheme as the initial particle population;

[0117] 43) Using boundary conditions as constraints and leveled energy cost as the optimization objective, calculate the particle fitness function;

[0118] Determine if the particle fitness function is less than a set threshold: if yes, obtain the wind turbine layout scheme; if no, increment the iteration count by 1 and proceed to step 44).

[0119] 44) Determine if the number of iterations has reached the set maximum number of iterations: if yes, obtain the wind turbine generator layout scheme; if no, use the adaptive weighted particle swarm optimization algorithm to update the particle population and execute step 43).

[0120] The adaptive weighted particle swarm optimization algorithm is as follows:

[0121]

[0122]

[0123] in, It is the position of the i-th particle at the t-th iteration; r1 and r2 are the velocity vectors of the i-th particle in the t-th iteration; r1 and r2 are random numbers between 0 and 1; p i and p g These are the best candidate solutions found for the i-th particle and the entire population, respectively; β i and β g These are custom parameters that control the particle's ability to develop and explore motion, respectively; ω a It is a nonlinear dynamic inertial weight that is adjusted based on the distance to the global optimal solution during the optimization process.

[0124] Case Study

[0125] Taking a wind farm area with regular boundaries as the research object, the wind turbine generator optimization layout method proposed in this embodiment is used for turbine arrangement. The planned area and reference turbine locations are as follows: Figure 3 As shown.

[0126] By combining long-term observed wind speed data of the target wind farm, the annual average wind speed and direction distribution information is obtained. Then, a single optimization algorithm and a genetic particle swarm optimization algorithm are used to optimize the wind turbine layout. The layout schemes obtained by the two algorithms are as follows: Figure 4 As shown, where Figure 4 (a) and (b) are wind turbine layout diagrams for the genetic algorithm and particle swarm optimization algorithm, respectively. Figure 4 (c) is a wind turbine layout diagram based on the genetic particle swarm fusion algorithm.

[0127] The parameters of the three wind turbine layout schemes are compared, as shown in Table 1. Table 1 shows that, compared with a single optimization algorithm, the proposed fusion optimization algorithm can effectively reduce the number of iterations while ensuring the accuracy of the optimization results. Overall, the proposed efficient algorithm for wind turbine layout schemes can reduce computation time and significantly improve the optimization effect of wind turbine layout.

[0128] Table 1 Comparison of Wind Turbine Layout Schemes Parameters

[0129]

[0130] As shown in Table 1, the fusion optimization algorithm proposed in this embodiment, while simultaneously considering both wind farm power generation and levelized cost of electricity (LCOE), can not only increase total power generation by 6%-10% while maintaining the optimization effect on LCOE, but also reduce iterative calculation time by 25%-40%. Specifically, this example employs a statistical analysis method for 10 optimization processes and results, as illustrated in Table 1. Figure 5 The box plot shown has the horizontal axis representing the method and the vertical axis representing the cost per kilowatt-hour. Figure 5 It can be seen that, compared with the other two methods, the solution obtained by the fusion optimization algorithm is more stable and has better accuracy guarantee.

[0131] In summary, when optimizing wind turbine layout schemes using a single optimization algorithm, it is often difficult to simultaneously guarantee efficiency and accuracy. This embodiment proposes a multi-stage optimization process for layout optimization, and also proposes an improved genetic algorithm mutation rate, which can save computation time and avoid the optimization process getting trapped in local optima. The method proposed in this embodiment combines high efficiency and accuracy, and is suitable for the optimization problem of wind turbine layout schemes.

[0132] The above-described embodiments are merely preferred embodiments provided to fully illustrate the present invention, and the scope of protection of the present invention is not limited thereto. Equivalent substitutions or modifications made by those skilled in the art based on the present invention are all within the scope of protection of the present invention. The scope of protection of the present invention is defined by the claims.

Claims

1. A method for optimizing the layout of wind turbine generator sets based on a three-level dynamic variation rate, characterized in that: Includes the following steps: Step 1: Obtain information about the target wind farm, including: (1) Annual average wind resource information of the target wind farm, and draw rose diagram based on annual statistical wind speed and wind direction information in the target wind farm area; (2) Unit specifications, including hub height, impeller diameter and power-thrust curve; (3) Boundary of wind farm planning area, delineating the boundary of wind farm planning area including regular or irregular boundaries; Step 2: Calculate the annual total average power generation and assess the levelized cost of electricity (LCOE) over the entire life cycle by combining the wind turbine wake velocity and turbulence model; Step 3: Grid the wind farm planning area, take levelized energy cost as the optimization objective, and use an improved genetic optimization algorithm based on three-level dynamic mutation rate to perform preliminary optimization of the grid location of the units. Several grid optimization schemes are obtained through ensemble theory. Step three includes the following steps: 31) Divide the wind farm planning area into grid distribution maps of different sizes, and the coordinate position of the candidate units is the grid center by default; 32) An improved genetic optimization algorithm based on a three-level dynamic mutation rate was used for preliminary optimization of the unit grid location: 321) Assume that each grid center has two possibilities: installing a fan or not installing a fan. The corresponding "genes" are encoded as "1" and "0" respectively. Each individual corresponds to a fan arrangement scheme and is composed of a string of "genes". 322) Randomly initialize the population: For each "gene", randomly assign an amplitude of "1" or "0", and initialize the population to the current iteration step. ; 323) Calculate the fitness of individuals based on the uniqueness process to obtain a population of completely unique individuals; Determine if the individual fitness is less than a set threshold: if yes, then the wind turbine layout grid optimization scheme is obtained; otherwise, (Execute step 324). 324) Random individual selection is performed using the roulette wheel selection method; 325) Random crossover is performed between every two individuals, which means the "genes" at corresponding positions in the two individuals are interchanged; 326) Random mutations are performed on the "genes" of each individual; The "genes" of each individual are randomly mutated using an improved mutation rate, which is: in, The rate of variation at the 0-encoded position; The variation rate at the 1-encoding position; Number of grid cells; This represents the average population size. and These represent the initial mutation rate and the final mutation rate, respectively. This indicates the repetition rate of individuals in a population. Indicates the number of steps in the current iteration; Indicates the total number of iterations; 327) Judgment Is it equal to the set value? If yes, then the wind turbine layout grid optimization scheme is obtained; otherwise, step 323 is executed repeatedly. Step 4: Based on the preliminary optimization results of the grid positions, construct the grid boundary for each candidate unit, and use the adaptive weighted particle swarm optimization algorithm to perform fine coordinate position optimization within the corresponding grid to obtain the final unit layout scheme.

2. The wind turbine generator set optimization layout method based on three-level dynamic variation rate according to claim 1, characterized in that: In step two, the method for calculating the annual average total power generation and assessing the life-cycle cost of electricity is as follows: 21) Based on the Gaussian wind turbine wake model and turbulence model, calculate the wind speed loss and turbulence enhancement of each wind turbine after being affected by the wake of the upstream wind turbine; 22) The inflow wind speed at the hub height of each downstream wind turbine is estimated using the classical wake superposition principle. The annual power generation of each wind turbine is calculated. The annual power generation of each wind turbine is summed to obtain the total annual power generation of the wind farm. 23) Calculate the annual average distribution of wind speed across all wind directions and the total power generation of the wind farm over its entire life cycle. Combine this with economic benefit indicators, including the cost of generator components, wind farm operation and maintenance costs, and capital recovery coefficient, to construct a levelized energy cost objective function.

3. The wind turbine generator set optimization layout method based on three-level dynamic variation rate according to claim 2, characterized in that: In step 21), the Gaussian wind turbine wake model is: in, It is a loss in wake wind speed; The loss is in the wake velocity at the wheel hub height; It is the standard deviation of the wake velocity profile; It is the diameter of the fan impeller; and These are the horizontal and vertical coordinates, respectively. It is the height of the wind turbine hub; It is the thrust coefficient; It is the wake attenuation coefficient; This indicates the horizontal distance between the downstream fan and the upstream fan; Represents parameters related to the thrust coefficient; The calculation method for wind turbine wake turbulence is as follows: in, It is an additional wake turbulence; These are model constants; The superimposed wake of the wind turbine is as follows: in, and These are the inflow velocities at the upstream and downstream fans; The incoming wind speed at infinity; This refers to the inflow velocity caused by the upstream fan at the downstream fan location.

4. The wind turbine generator set optimization layout method based on three-level dynamic variation rate according to claim 2, characterized in that: In step 22), for a given speed and direction Given the incoming flow conditions, the total power generation of the wind farm can be calculated based on the unit power curve. Combining this with the probability density function of the statistical average wind conditions, the annual average power generation is predicted as follows: in, It refers to the number of wind turbines; It refers to the number of wind direction zones. It refers to the number of wind speed zones. It represents the probability of wind direction range; It is in the wind speed Order the power generation capacity of each generating unit; This is the unit's coordinate information; The joint distribution probability of wind speed and wind direction It is the dividing point of wind speed range; Indicates the first The wind speed represents the wind direction angle.

5. The wind turbine generator set optimization layout method based on three-level dynamic variation rate according to claim 2, characterized in that: In step 23), the levelized energy cost objective function is: in, It is the cost of the generator unit; It is the capital recovery coefficient; It is the operation and maintenance cost of the wind farm; It is the average annual total power generation of the wind farm.

6. The wind turbine generator set optimization layout method based on three-level dynamic variation rate according to claim 1, characterized in that: Step four includes the following steps: 41) Calculate the center coordinates and boundary conditions of the grids containing the candidate units in all grid optimization schemes; 42) Use the unit center coordinates of the grid optimization scheme as the initial particle population; 43) Using boundary conditions as constraints and leveled energy cost as the optimization objective, calculate the particle fitness function; Determine if the particle fitness function is less than a set threshold: if yes, obtain the wind turbine layout scheme; if no, increment the iteration count by 1 and proceed to step 44). 44) Determine if the number of iterations has reached the set maximum number of iterations: if yes, obtain the wind turbine generator layout scheme; if no, use the adaptive weighted particle swarm optimization algorithm to update the particle population and execute step 43).

7. The wind turbine generator set optimization layout method based on three-level dynamic variation rate according to claim 6, characterized in that: In step 44), the adaptive weighted particle swarm optimization algorithm is as follows: in, It is the first The particle in the first The position at the next iteration; It is the first The particle in the first The velocity vector at the next iteration; and It is a random number between 0 and 1; and They are respectively the first The best candidate solution found for each particle and the entire population; and These are custom parameters that control the particle's ability to develop and explore motion, respectively. It is a nonlinear dynamic inertial weight that is adjusted based on the distance to the global optimal solution during the optimization process.

Images