Micro-siting and model optimization method for wind farm based on weibull parameter vertical and space-time evolution

Abstract

The present application provides a wind farm micro-siting and model optimization method based on Weibull parameter vertical and space-time evolution, and relates to the field of wind resource analysis and wind power prediction. The steps of the present application are as follows: S1, collecting multi-height layer laser radar observation data; S2, calculating scale parameter A and shape parameter K according to the measured height layer wind speed data, and reconstructing the Weibull parameter field; S3, analyzing the Weibull parameter from the time dimension and the height dimension; S4, establishing a wind resource-wind turbine performance mapping model; S5, developing a collaborative optimization algorithm, optimizing the site and matching the wind turbine type and hub height, and obtaining the final scheme. The beneficial effects of the present application are as follows: the present application can more accurately reflect the real wind speed distribution and energy capture process on the wind turbine swept area, greatly improves the accuracy of the wake loss calculation, breaks the limitations of the traditional process of cutting or sequential processing of these links, and realizes the overall optimization of the overall design scheme of the wind farm.

Description

Technical Field

[0001] This invention relates to the field of wind resource analysis and wind power prediction, specifically to a method for wind farm micro-site selection and turbine optimization based on the vertical and spatiotemporal evolution of Weibull parameters. Background Technology

[0002] As a relatively mature clean and renewable energy source, wind energy's development and utilization hinges on the precise assessment of wind resources, determining the optimal location for wind turbines, and selecting suitable turbine models for specific locations. These three aspects are interconnected, forming a tight decision-making chain. Given the availability of measured data for wind farm resource analysis, rational micro-site selection and appropriate turbine model selection are crucial to determining the economic benefits of a wind farm throughout its entire lifecycle.

[0003] For existing measured wind resource data, the internationally accepted method is to use a two-parameter Weibull distribution to describe the probability distribution of wind speed and assess wind energy resources. The scale parameter A is proportional to the average wind speed and directly characterizes the wind energy potential; the shape parameter K characterizes the stability and concentration of wind speed, affecting the predictability of power generation and extreme loads.

[0004] However, offshore lidar wind speed measurements show that the variations of A and K with altitude are nonlinear and non-monotonic, potentially indicating the presence of significant low-level jet streams, wind speed inflection points, or shear layers. Furthermore, wind parameters vary throughout the year depending on the season and the duration of winds. Using fixed values ​​and simple extrapolation can significantly overestimate or underestimate the energy density and turbulence intensity at certain altitudes and times. This could lead to inappropriate hub height selection, difficulty in capturing the strata with the greatest wind energy, or difficulty in preventing the wind turbine from being placed in unfavorable strata with excessive turbulence, thereby affecting wind farm power generation and wind turbine lifespan. Summary of the Invention

[0005] The main objective of this invention is to provide a method for wind farm micro-site selection and turbine optimization based on the vertical and spatiotemporal evolution of Weibull parameters, thereby solving the problems mentioned in the background art.

[0006] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is: a wind farm micro-site selection and turbine optimization method based on the vertical and spatiotemporal evolution of Weibull parameters, comprising the following steps: S1. Collect multi-height-layer lidar observation data; S2. Calculate the scale parameters based on the measured wind speed data at the measured height level. A and shape parameters K， Reconstruct the Weibull parameter field; S3. Analyze the Weibull parameters from the perspectives of time and height. S4. Establish a wind resource-wind turbine performance mapping model; S5. Develop a collaborative optimization algorithm to select the best locations and match the wind turbine type and hub height to obtain the final solution.

[0007] Furthermore, the data covers the range of wind turbine impeller heights, with a time span of one year and a sampling interval of less than 10 minutes.

[0008] Furthermore, scale parameters A and shape parameters K The expression is: (1); in, It is the first i The wind speed value of each data point. The arithmetic mean of all wind speed values. n For the total number of data, A and K An approximate solution can be obtained iteratively.

[0009] Furthermore, the specific process of step S3 is as follows: scale parameters A and shape parameters K Add a timestamp; The scale parameter corresponding to this point A and shape parameters K Substituting this in, the data dimension is represented as " x , y , h , t , A , K ",in x , y , h These are the coordinates corresponding to the spatial coordinate axes. t For time; The data was divided into 12 sets with a monthly cycle. Within each set, the scaling parameters were applied to the data at the same time on different days. A and shape parameters K The operation of calculating the average; Using the calculated mean data, the scale parameter can be determined. A Shape parameters K With height h Fit the relationship; At this point, the scale parameter A Shape parameters K It can be expressed in the form of a four-dimensional parametric field as A ( x , y , h , t )and K (x , y , h , t ).

[0010] Furthermore, the specific process of step S4 is as follows: Scale parameter in the height dimension A Shape parameters K They can be denoted as respectively A ( h )and K ( h ); For any candidate machine site, its wind resources are determined by a scale parameter that is highly correlated with the wind speed. A ( h ) and shape parameters K ( h The definition is as follows: (2); in, AEP This represents the theoretical annual power generation. Set height layer intervals b Each layer has pre-calculated scale parameters. A ( h Shape parameters K ( h ); The center height of the wheel hub is denoted as H The radius of the impeller is denoted as R Centered on the height of the wheel hub, the interval is... b Then the number of impeller layers m for: (3); in, It is the floor function, i.e., its radius. R The integer part; The sweep area of ​​each layer is calculated, with the bottom layer designated as layer 1, and so on upwards, up to layer 2. m Layer, number i The area of ​​the layer is denoted as S. i ; For any height layer, its Weibull probability density f The expression is: (4); in, v Wind speed; Then the Weibull cumulative distribution function F The expression is: (5); For wind speed, it is necessary to determine its cut-in wind speed. and cut-out wind speed ; The wind speed range is divided into intervals of 100°C and 100°C. Then the lower limit of the total wind speed range is 0, and the upper limit is... Then the first j Each wind speed range is denoted as Its interval range is [ v j-1 , v j ]; For wind speeds within the range probability Essentially, it is the Weibull cumulative distribution function over this interval. According to equation (5), its expression is: (6); For interval The average of its upper and lower limits is recorded as the representative wind speed. ; According to the wind turbine power curve P ( v Find the power corresponding to the wind speed. P ( v ja ), will the first i The representative power of each height level is denoted as P i Its expression is: (7); in, l This represents the total number of wind speed ranges for this unit. The total power of the wind turbine The expression is: (8); in, w i For the first i The weight of each height layer; Discretize the whole year into M The sampling interval of the lidar is [number] equal time intervals. Then we have: (9); but AEP The expression is: (10); in, t k For the first k At any given time period For the first kThe total power of the wind turbine at each time period.

[0011] Furthermore, the calculation process for the swept area is as follows: Scenario 1: In R / b In scenarios where the area is an integer, the upper bound of the lowest layer is tangent to the circular surface swept by the impeller, therefore the area is... S Since 1 is equal to 0, the swept area calculation starts from the second layer. Because the height interval is extremely small compared to the impeller radius, the first layer can be considered equivalent to a triangle with the area: (11); Starting from the 3rd layer, the area of ​​the sweeping airflow can be represented by a trapezoid with the following area: (12); Scenario 2: In R / b In scenarios where the integer value is not an integer, a straight line parallel to the horizontal plane and tangent to the circular surface swept by the impeller can be drawn within the first layer. The distance between this line and the upper boundary of the first layer is denoted as... b 1. The sweeping area of ​​the first layer is also approximately triangular, so the sweeping area of ​​the first layer is: (13); Starting from the second layer, the content is the same as after the third layer in Case 1, and the swept area is: (14).

[0012] Furthermore, weight w i The expression is: (15); Furthermore, the detailed process of step S5 is as follows: First, the optimization objective is to minimize the levelized cost of electricity (LCOE). Give all within the wind farm area C Add a permutation number to each machine site, then the first... c Each camera position is recorded as h c ,for h c Its value can be 1 or 0. When it is 1, it means that a fan is installed at the point, and when it is 0, it means that no fan is installed at the point. For each location where the decision is made to install a fan c Select a wind turbine from a predefined model library. The model library contains a total of [number] wind turbine types. BThe format for storing wind turbine information is "Serial Number-Model-Hub Height-Impeller Diameter-Rated Power-Cost". c The model selected for each camera position y c express, y c The range of values ​​for is [1, ... B The integer in ] when it takes b When, the representative selects the first b Types of fans; The objective function is determined as follows: (16); in, G total For construction costs, G w For operating costs, G z The residual value of a wind farm after it has reached the end of its operating life. Q total Total construction cost and total power generation; For the first t Annual project operating costs, t Indicates the year of operation. For the project's rate of return on investment, For the first t Annual project power generation Given the designed service life, the optimization objective is to minimize the objective function value; The optimal solution is obtained by performing a global search using the particle swarm optimization algorithm.

[0013] Furthermore, the constraints include minimum wind turbine spacing constraints, wind farm boundary constraints, and wake loss constraints. set up i , j For any two wind turbines, Distance along the prevailing wind direction Distance perpendicular to the prevailing wind direction. D ᵢ , Dⱼ Given the diameter of each impeller, the minimum fan spacing constraint is: (17); The boundary constraints of a wind farm are as follows: for a turbine location, its range can only be within the boundary of the wind farm or exactly on the boundary of the wind farm; The wake loss constraint is: (18); in, For the first i The wake loss rate of a typhoon turbine. For the first i The actual output power of a typhoon turbine after being affected by wake. For the first i The theoretical power of a typhoon turbine operating in isolation. This represents the maximum allowable wake loss rate for a single fan. The average wake loss rate of all wind turbines in the field. N This represents the total number of wind turbines. The average allowable wake loss rate for the entire field; The wake loss constraint is a soft constraint, meaning that if the loss exceeds the constraint range, it will not be forcibly eliminated. Instead, a penalty term, Penalty, will be added. The expression for the penalty term is as follows: (19); in, AEP total This represents the theoretical total power generation when the wind turbine reaches its service life. , These are the penalty coefficient for a single fan and the average penalty coefficient for the entire field, respectively. At this point, the expression for the objective function is: (20).

[0014] Furthermore, during the particle swarm optimization algorithm iteration process, in the wake loss calculation, the wake model uses the actual location of that point. A ( h )and k ( h The decreasing wind speed is used as input to perform iterative calculations. For the wake model, the wake generated by the upstream unit will cause a decrease in the representative wind speed of the downstream unit. The decrease in wind speed is denoted as... Then the model expression is: (twenty one); in, v down The incoming wind speed of the upstream unit. C T The thrust coefficient of the upstream wind turbine. k T The wake attenuation coefficient is... L The horizontal distance between the upstream and downstream wind turbines. D T The diameter of the upstream wind turbine impeller. This is a gamma function.

[0015] Beneficial effects: (1) Weibull parameters A and kViewing it as a dynamic field that varies with height, space, and time, rather than a fixed constant, provides a solid data foundation for optimization.

[0016] (2) An AEP calculation model for explicitly coupled Weibull parameter vertical distribution (A(h), k(h)) and wind turbine performance was established. Compared with the simple model of hub height, this model can more accurately reflect the real wind speed distribution and energy capture process on the swept surface of the wind turbine.

[0017] (3) In the wake assessment, the Weibull parameter field, which is dynamically evolved by the upstream wind turbine, is introduced as the inflow condition of the downstream wind turbine, which greatly improves the accuracy of wake loss calculation and makes the results of optimized layout closer to the actual operation.

[0018] (4) The algorithm puts the selection of the site, the matching of the turbine model and the wake effect into the same framework, breaking the limitation of the traditional process of separating or sequentially processing these links, and realizing the overall optimization of the wind farm design scheme. Attached Figure Description

[0019] The present invention will be further described below with reference to the accompanying drawings and embodiments: Figure 1 This is a flowchart of the technical solution of the present invention. Detailed Implementation

[0020] Example 1 like Figure 1 As shown, the wind farm micro-situation and turbine optimization method based on the vertical and spatiotemporal evolution of Weibull parameters includes the following steps: S1. Collect multi-height-layer lidar observation data; LiDAR wind measurement data were collected from multiple potential turbine sites in deep-sea wind farms. The data covered the range of wind turbine rotor height, with a time span of one year and a sampling interval of less than 10 minutes.

[0021] S2. Calculate the scale parameters based on the measured wind speed data at the measured height level. A and shape parameters K， Reconstruct the Weibull parameter field; the calculation method can be the least squares method, the maximum likelihood estimation method (MLE), etc., taking the MLE algorithm as an example: Scale parameters A and shape parameters K The expression is: (1); in, It is the first i The wind speed value of each data point. The arithmetic mean of all wind speed values. nThe total number of data points can be calculated by iterating over formula (1). K An approximate solution is obtained by solving the problem. K Then, it can be calculated. A The value of .

[0022] S3. Analyze the Weibull parameters from the time and height dimensions. The specific process is as follows: scale parameters A and shape parameters K The process of adding a timestamp is as follows: For the data, excluding the wind speed value, the format after adding the timestamp is "location-altitude-year-month-day"; The scale parameter corresponding to this point A and shape parameters K Substituting this into the data, the format is "location-height-year-month-day-scale parameter-shape parameter", which can be represented using dimensions as "". x , y , h , t , A , K ",in x , y , h These are the coordinates corresponding to the spatial coordinate axes. t For time; The data was divided into 12 sets with a monthly cycle. Within each set, the scaling parameters were applied to the data at the same time on different days. A and shape parameters K The operation of calculating the average; By using the average data, the scale parameter can be determined. A Shape parameters K With height h The relationship is fitted; the fitting method can be nonlinear fitting, logarithmic fitting, etc., for example: (2); in, a , d , c , p , q These are the fitting coefficients; For cases where the variation with altitude is complex, piecewise functions can be used to capture inflection points, thus replacing simple constant values ​​and power-law extrapolation. At this point, the scale parameter A Shape parameters K It can be expressed in the form of a four-dimensional parametric field as A ( x , y ,h , t )and K ( x , y , h , t ).

[0023] S4. Establish a wind resource-wind turbine performance mapping model. The specific process is as follows: Scale parameter in the height dimension A Shape parameters K They can be denoted as respectively A ( h )and K ( h ); For any candidate machine site, its wind resources are determined by a scale parameter that is highly correlated with the wind speed. A ( h ) and shape parameters K ( h The definition is as follows: (3); in, AEP This value represents the theoretical annual power generation and thus indicates the abundance of wind resources. For a certain type of wind turbine, the lidar measurement height range is set to 40~300m to cover the height range of the lowest and highest points of the impeller, with height layer intervals set. b The height is 1m. Since the endpoints are included, there are a total of 261 height levels, each with a pre-calculated scale parameter. A ( h Shape parameters K ( h ); The center height of the wheel hub is denoted as H The radius of the impeller is denoted as R With the center height of the hub as the center, and intervals of 1m, the number of impeller layers is... m for: (4); in, It is the integer part of the radius R; Next, we need to calculate the sweep area of ​​each layer. Since the plane swept by the impeller is symmetrical, we will only explain the part below the hub center. The sweep area above it can be obtained based on the symmetry. Starting with the bottom layer as the first layer, and proceeding upwards in sequence, the top layer is the _____. m Layer, number i The area of ​​the layer is denoted as S. i ; Scenario 1: In R / b In scenarios where the area is an integer, the upper bound of the lowest layer is tangent to the circular surface swept by the impeller, therefore the area is... S Since 1 is equal to 0, the swept area calculation starts from the second layer. Because the height interval is extremely small compared to the impeller radius, the first layer can be considered equivalent to a triangle with the area: (5); Starting from the 3rd layer, the area of ​​the sweeping airflow can be represented by a trapezoid with the following area: (6); Scenario 2: In R / b In scenarios where the integer value is not an integer, a straight line parallel to the horizontal plane and tangent to the circular surface swept by the impeller can be drawn within the first layer. The distance between this line and the upper boundary of the first layer is denoted as... b 1. The sweeping area of ​​the first layer is also approximately triangular, so the sweeping area of ​​the first layer is: (7); Starting from the second layer, the content is the same as after the third layer in Case 1, and the swept area is: (8); For any height layer, its Weibull probability density f The expression is: (9); in, v Wind speed; Then the Weibull cumulative distribution function F The expression is: (10) For wind speed, it is necessary to determine its cut-in wind speed. and cut-out wind speed The cut-in wind speed is generally taken as 3 m / s, and the cut-out wind speed is generally taken as 25 m / s; The wind speed range is divided into intervals of 100°C and 100°C. Then the lower limit of the total wind speed range is 0, and the upper limit is... Then the first j Each wind speed range is denoted as Its interval range is [ v j-1 , v j ]; For wind speeds within the range probability Essentially, it is the Weibull cumulative distribution function over this interval, and according to equation (10), its expression is: (11); For interval The average of its upper and lower limits is recorded as the representative wind speed. ; According to the wind turbine power curve P ( v Find the power corresponding to the wind speed. P ( v ja ), will the first i The representative power of each height level is denoted as P i Its expression is: (12); in, l This represents the total number of wind speed ranges for this unit. No. i The weight of each height layer is denoted as w i Its expression is: (13); The total power of the wind turbine The expression is: (14); Discretize the whole year into M The sampling interval of the lidar is [number] equal time intervals. Then we have: (15); but AEP The expression is: (16); in, t k For the first k At any given time period For the first k The total power of the wind turbine at each time period.

[0024] S5. Develop a collaborative optimization algorithm to select optimal locations and match wind turbine types and hub heights to obtain the final solution. The detailed process is as follows: First, the optimization objective is to minimize the levelized cost of electricity (LCOE), which is defined as the average cost incurred in generating one kilowatt-hour of electricity. Give all within the wind farm area C Add a permutation number to each machine site, then the first... c Each camera position is recorded as h c ,for h cIts value can be 1 or 0. When it is 1, it means that a fan is installed at the point, and when it is 0, it means that no fan is installed at the point. For each location where the decision is made to install a fan c Select a wind turbine from a predefined model library. The model library contains a total of [number] wind turbine types. B The format for storing wind turbine information is "Serial Number-Model-Hub Height-Impeller Diameter-Rated Power-Cost". c The model selected for each camera position y c express, y c The range of values ​​for is [1, ... B The integer in ] when it takes b When, the representative selects the first b There are several types of wind turbines; the costs here include construction costs and operating costs. The objective function is determined as follows: (17); in, G total For construction costs, G w For operating costs, G z The residual value of a wind farm after it has reached the end of its operating life. Q total Total construction cost and total power generation; For the first t Annual project operating costs, t Indicates the year of operation. For the project's rate of return on investment, For the first t Annual project power generation Given the designed service life, the optimization objective is to minimize the objective function value; The constraints include minimum turbine spacing constraints, wind farm boundary constraints, and wake loss constraints; set up i , j For any two wind turbines, Distance along the prevailing wind direction Distance perpendicular to the prevailing wind direction. D ᵢ , Dⱼ Given the diameter of each impeller, the minimum fan spacing constraint is: (18); If the minimum wind turbine spacing constraint is not met, the turbine will be directly eliminated. The boundary constraints of the wind farm are as follows: For the turbine location, its range can only be within the boundary of the wind farm or exactly on the boundary of the wind farm. Once it exceeds the boundary range of the wind farm, it will be directly eliminated. For the wake model, the wake generated by the upstream unit will cause a decrease in the representative wind speed of the downstream unit. The decrease in wind speed is denoted as... Then the model expression is: (19); in, v down The incoming wind speed of the upstream unit. C T The thrust coefficient of the upstream wind turbine. k T The wake attenuation coefficient is... L The horizontal distance between the upstream and downstream wind turbines. D T The diameter of the upstream wind turbine impeller. This is the gamma function, and its value can be obtained directly by looking up a table; The wake loss constraint is: (20); in, For the first i The wake loss rate of a typhoon turbine. For the first i The actual output power of a typhoon turbine after being affected by wake. For the first i The theoretical power of a typhoon turbine operating in isolation. This is the maximum allowable wake loss rate for a single fan, typically taken as 40%. The average wake loss rate of all wind turbines in the field. N This represents the total number of wind turbines. The average allowable wake loss rate for the entire field is typically taken as 20%. The wake loss constraint is a soft constraint, meaning that if the loss exceeds the constraint range, it will not be forcibly eliminated. Instead, a penalty term, Penalty, will be added. The expression for the penalty term is as follows: (twenty one); in, AEP total This represents the theoretical total power generation when the wind turbine reaches its service life. , These are the penalty coefficient for a single fan and the average penalty coefficient for the entire field, respectively. At this point, the expression for the objective function is: (twenty two); The particle swarm optimization algorithm is used to perform a global search to obtain the optimal solution, i.e., the final solution. The optimization process of the particle swarm optimization algorithm is as follows: A particle swarm of 50 particles is randomly generated, each representing a wind farm scheme. The generated particles must satisfy constraints; if not, they are regenerated until all 50 particles satisfy the constraints are obtained. The maximum number of iterations is set to 300, and the weights are determined accordingly. w pso The acceleration constant decreases adaptively with iteration. u 1. u Both are set to 2.0, weights w pso The initial value is 0.9, and it decreases by 0.005 in each iteration, with a lower limit of 0.4 for the weights. The position and velocity of the particles are updated according to the weights of the particle swarm. The specific process is as follows: (twenty three); in, , The first i The particle in the first t Velocity and position after the next iteration , A random number in [0,1] For the first i The individual optimal solution for a particle refers to the position where the objective function value is minimized for that particle. The global optimal solution is obtained as follows: after updating the individual optimal solutions of all particles, iterate through the individual optimal solutions of all particles, select the individual optimal solution with the smallest objective function value, and set it as the global optimal solution for the current iteration; it should be noted that step S4... AEP The calculation did not consider the effect of the wake, but in the iterative process of the particle swarm optimization algorithm, the wake is taken into account when calculating the wind speed, thus the calculated value... AEP Higher accuracy; For particles that do not meet the minimum wind turbine spacing constraint or are removed from the wind farm boundary constraint, move the particles to the boundary that is closest to them. To improve the accuracy of wake loss calculation, a uniform background field is not used in each iteration of the wake loss calculation; instead, the actual background field at that point is used. A ( h )and k ( h As input, the decreasing wind speed value is iteratively calculated using equation (19); The minimum number of iterations is set to 100, the maximum number of iterations is set to 300, and the early stopping threshold is set to 0.1; training stops when the maximum number of iterations is reached. When training reaches the minimum number of iterations, a judgment is made: if the improvement of 10 consecutive iterations does not exceed the early stopping threshold, training is stopped early.

[0025] The above embodiments are merely preferred technical solutions of the present invention and should not be considered as limitations on the present invention. The scope of protection of the present invention should be limited to the technical solutions described in the claims, including equivalent substitutions of the technical features described in the claims. That is, equivalent substitutions and improvements within this scope are also within the scope of protection of the present invention.

Claims

1. A wind farm micro-situation and turbine optimization method based on the vertical and spatiotemporal evolution of Weibull parameters, characterized in that: Includes the following steps: S1. Collect multi-height-layer lidar observation data; S2. Calculate the scale parameters based on the measured wind speed data at the measured height level. A and shape parameters K， Reconstruct the Weibull parameter field; S3. Analyze the Weibull parameters from the perspectives of time and height. The specific process of step S3 is as follows: scale parameters A and shape parameters K Add a timestamp; scale parameters corresponding to the points A and shape parameters K Substituting this in, the data dimension is represented as " x , y , h , t , A , K ",in x , y , h These are the coordinates corresponding to the spatial coordinate axes. t For time; The data was divided into 12 sets with a monthly cycle. Within each set, the scaling parameters were applied to the data at the same time on different days. A and shape parameters K The operation of calculating the average; Using the calculated mean data, the scale parameter can be determined. A Shape parameters K With height h Fit the relationship; At this point, the scale parameter A Shape parameters K It can be expressed in the form of a four-dimensional parametric field as A ( x , y , h , t )and K ( x , y , h , t ); S4. Establish a wind resource-wind turbine performance mapping model; The specific process of step S4 is as follows: Scale parameter in the height dimension A Shape parameters K They can be denoted as respectively A ( h )and K ( h ); For any candidate machine site, its wind resources are determined by a scale parameter that is highly correlated with the wind speed. A ( h ) and shape parameters K ( h The definition is as follows: （2）； in, AEP This represents the theoretical annual power generation. Set height layer intervals b Each layer corresponds to a pre-calculated scale parameter. A ( h ), shape parameters K ( h ); The center height of the wheel hub is denoted as H The radius of the impeller is denoted as R Centered on the height of the wheel hub, the interval is... b Then the number of impeller layers m for: （3）； in, It is the floor function, i.e., its radius. R The integer part; The sweep area of ​​each layer is calculated, with the bottom layer designated as layer 1, and so on upwards, up to layer 2. m Layer, number i The area of ​​the layer is denoted as S. i ; For any height layer, its Weibull probability density f The expression is: （4）； in, v Wind speed; Then the Weibull cumulative distribution function F The expression is: （5）； For wind speed, it is necessary to determine its cut-in wind speed. and cut-out wind speed ; The wind speed range is divided into intervals of 100°C and 100°C. Then the lower limit of the total wind speed range is 0, and the upper limit is... Then the first j Each wind speed range is denoted as Its interval range is [ v j-1 , v j ]; For wind speeds within the range probability Essentially, it is the Weibull cumulative distribution function over this interval. According to equation (5), its expression is: （6）； For interval The average of its upper and lower limits is recorded as the representative wind speed. ; According to the wind turbine power curve P ( v Find the power corresponding to the wind speed. P ( v ja ), will the first i The representative power of each height level is denoted as P i Its expression is: （7）； in, l This represents the total number of wind speed ranges for each location. The total power of the wind turbine The expression is: （8）； in, w i For the first i The weight of each height layer; Discretize the whole year into M The sampling interval of the lidar is [number] equal time intervals. Then we have: （9）； but AEP The expression is: （10）； in, t k For the first k At any given time period For the first k The total power of the wind turbine at each time period; S5. Develop a collaborative optimization algorithm to select optimal locations and match wind turbine types and hub heights to obtain the final solution; The detailed process of step S5 is as follows: First, the optimization objective is to minimize the levelized cost of electricity (LCOE). Give all within the wind farm area C Add a permutation number to each machine site, then the first... c Each camera position is recorded as h c ,for h c Its value can be 1 or 0. When it is 1, it means that a fan is installed at the point, and when it is 0, it means that no fan is installed at the point. For each location where the decision is made to install a fan c Select a wind turbine from a predefined model library. The model library contains a total of [number] wind turbine types. B The format for storing wind turbine information is "Serial Number-Model-Hub Height-Impeller Diameter-Rated Power-Cost". c The model selected for each camera position y c express, y c The range of values ​​for is [1, ... B The integer in ] when it takes b When, the representative selects the first b Types of fans; The objective function is determined as follows: （16）； in, G total For construction costs, G w For operating costs, G z The residual value of a wind farm after it has reached the end of its operating life. Q total Total construction cost and total power generation; For the first t Annual project operating costs, t Indicates the year of operation. For the project's rate of return on investment, For the first t Annual project power generation Given the designed service life, the optimization objective is to minimize the objective function value; The optimal solution is obtained by performing a global search using the particle swarm optimization algorithm. The constraints include minimum turbine spacing constraints, wind farm boundary constraints, and wake loss constraints; set up i , j For any two wind turbines, Distance along the prevailing wind direction Distance perpendicular to the prevailing wind direction. D i 、D j Given the diameter of each impeller, the minimum fan spacing constraint is: （17）； The boundary constraints of a wind farm are as follows: for a turbine location, its range can only be within the boundary of the wind farm or exactly on the boundary of the wind farm; The wake loss constraint is: （18）； in, For the first i The wake loss rate of a typhoon turbine. For the first i The actual output power of a typhoon turbine after being affected by wake. For the first i The theoretical power of a typhoon turbine operating in isolation. This represents the maximum allowable wake loss rate for a single fan. The average wake loss rate of all wind turbines in the field. N This represents the total number of wind turbines. The average allowable wake loss rate for the entire field; The wake loss constraint is a soft constraint, meaning that if the loss exceeds the constraint range, it will not be forcibly eliminated. Instead, a penalty term, Penalty, will be added. The expression for the penalty term is as follows: （19）； in, AEP total This represents the theoretical total power generation when the wind turbine reaches its service life. , These are the penalty coefficient for a single fan and the average penalty coefficient for the entire field, respectively. At this point, the expression for the objective function is: （20）。 2. The optimization method according to claim 1, characterized in that, The data covers the range of wind turbine impeller height, with a time span of one year and a sampling interval of less than 10 minutes.

3. The optimization method according to claim 1, characterized in that, Scale parameters A and shape parameters K The expression is: （1）； in, It is the first i The wind speed value of each data point. The arithmetic mean of all wind speed values. n For the total number of data, A and K An approximate solution can be obtained iteratively.

4. The optimization method according to claim 1, characterized in that, The calculation process for the swept area is as follows: Scenario 1: In R / b In scenarios where the area is an integer, the upper bound of the lowest layer is tangent to the circular surface swept by the impeller, therefore the area is... S Since 1 is equal to 0, the swept area calculation starts from the second layer. Because the height interval is extremely small compared to the impeller radius, the first layer can be considered equivalent to a triangle with the area: （11）； Starting from the 3rd layer, the area of ​​the sweeping airflow can be represented by a trapezoid with the following area: （12）； Scenario 2: In R / b In scenarios where the integer value is not an integer, a straight line parallel to the horizontal plane and tangent to the circular surface swept by the impeller can be drawn within the first layer. The distance between this line and the upper boundary of the first layer is denoted as... b 1. The sweeping area of ​​the first layer is also approximately triangular, so the sweeping area of ​​the first layer is: （13）； Starting from the second layer, the content is the same as after the third layer in Case 1, and the swept area is: （14）。 5. The optimization method according to claim 4, characterized in that, Weight w i The expression is: （15）。 6. The optimization method according to claim 1, characterized in that, During the iterative process of the particle swarm optimization algorithm, in the wake loss calculation, the wake model uses the actual location of that point. A ( h )and k ( h The decreasing wind speed is used as input to perform iterative calculations. For the wake model, the wake generated by the upstream unit will cause a decrease in the representative wind speed of the downstream unit. The decrease in wind speed is denoted as... Then the model expression is: （21）； in, v down The incoming wind speed of the upstream unit. C T The thrust coefficient of the upstream wind turbine. k T The wake attenuation coefficient is... L The horizontal distance between the upstream and downstream wind turbines. D T The diameter of the upstream wind turbine impeller. This is a gamma function.

Images