Wind turbine selection methods, equipment, storage media and software products

Abstract

This application discloses a wind turbine selection method, equipment, storage medium, and program product, relating to the technical field of wind power generation. By collecting multi-source data from the target wind farm and performing refined wind resource assessment, candidate turbine models are screened from a wind turbine product database. A multi-objective optimization model is constructed, with turbine model combinations, turbine location allocation schemes, and turbine location coordinates as optimization variables, and including multiple optimization objective functions and constraints. A multi-objective optimization algorithm is used to search for the optimal solution set in the feasible solution space, and an adaptive dynamic weight allocation mechanism is introduced to map the optimal solution set to the user preference space. The output is the optimal solution set containing turbine type and number of installed units, turbine location allocation schemes and turbine location coordinates, and calculated values ​​of various evaluation indicators. This application achieves coordinated optimization of multiple objectives such as power generation, life-cycle cost, grid adaptability, wake loss, and environmental impact, and can be widely applied to the joint optimization of turbine selection and micro-site selection for onshore and offshore wind farms.

Description

Technical Field

[0001] This application relates to the technical field of wind power generation, and in particular to a wind turbine selection method, wind turbine selection equipment, storage medium, and computer program product. Background Technology

[0002] Wind turbine selection is a core aspect of wind farm planning and design, and its rationality directly impacts the total power generation, investment costs, and overall project returns throughout the wind power cycle. With the rapid development of the wind power industry and the significantly accelerated pace of turbine iteration, the resulting surge in the number of candidate turbines has dramatically increased the complexity and difficulty of the selection process.

[0003] Currently, the mainstream methods for wind turbine selection mainly include manual experience-based comparison, single-index optimization, and weighted scoring. Among these existing methods: First, the data acquisition phase relies primarily on anemometer data, lacking comprehensive utilization of multi-source information such as topographic data, environmental constraint data, grid data, and meteorological reanalysis data. The basic data used for selection is not comprehensive or accurate enough to support refined selection decisions. Second, linear models or engineering experience models are often used for wind resource assessment, which have limited accuracy in simulating wind flow fields under complex terrain conditions. Furthermore, there is a lack of effective connection between CFD simulation results and selection decisions, making it difficult to directly input into the selection optimization model. Moreover, relying mainly on manual selection of candidate models from product catalogs is highly subjective and lacks quantifiable evidence. When the number of candidate models is large, the workload is enormous, and the selection results lack objectivity and repeatability. Finally, many methods use maximizing power generation or minimizing LCOE as the single optimization objective, ignoring factors such as grid support capacity, noise environmental impact, and wake loss. While multi-objective optimization methods have been theoretically studied, most employ linear weighted summation to transform multiple objectives into a single objective. This approach suffers from strong subjectivity in weight setting and fails to reflect the coupling relationships between objectives. Furthermore, these methods primarily consider basic constraints such as installed capacity and turbine spacing, lacking systematic modeling and quantitative evaluation methods for grid connection capacity constraints (e.g., short-circuit ratio requirements at grid connection points), environmental constraints (e.g., noise limits, ecological red lines), terrain adaptability constraints, and electromagnetic compatibility constraints when multiple turbine models are used interchangeably. Finally, the method separates turbine selection and turbine placement into steps, failing to integrate turbine selection with micro-site selection and wake optimization within a unified optimization framework. Moreover, it fails to incorporate the impact of differences in impeller diameters between different turbine models on the wake range and extent during the selection optimization process.

[0004] In summary, existing wind turbine selection methods have significant shortcomings in areas such as multi-source data utilization, wind resource assessment accuracy, systematic screening of candidate turbine models, multi-objective collaborative optimization, comprehensive constraint modeling, and joint optimization of selection and micro-site selection, and urgently need improvement.

[0005] The above content is only used to help understand the technical solution of this application and does not represent an admission that the above content is prior art. Summary of the Invention

[0006] The main objective of this application is to provide a wind turbine selection method, wind turbine selection equipment, storage medium, and computer program product, aiming to solve the technical problems in the prior art, such as insufficient data utilization, low evaluation accuracy, reliance on manual screening, one-sided objectives and constraints, and separation of selection and layout, which lead to low selection efficiency and difficulty in obtaining a globally optimal solution that takes into account economy, grid adaptability, and environmental friendliness.

[0007] To achieve the above objectives, this application proposes a wind turbine selection method, which includes: Collect multi-source data from the target wind farm, including at least wind measurement data, topographic and geographic data, environmental constraint data, power grid data, and meteorological reanalysis data; The wind resources are finely assessed using the wind measurement data, the topographic and geographic data, and the meteorological reanalysis data through a wind flow field numerical simulation method. The wind resource characteristic parameters at each candidate machine site are output, including at least wind speed parameters, turbulence parameters, and limiting wind speed parameters. A candidate turbine model set is selected from the wind turbine product database, and the characteristic parameters of the candidate turbine models are extracted. A secondary selection is performed based on the wind resource characteristic parameters to remove turbine models that do not meet the conditions. A multi-objective optimization model is constructed based on the candidate turbine models retained after the secondary selection. The multi-objective optimization model uses the combination of candidate turbine models, the turbine model allocation scheme of each turbine location, and the turbine location coordinates as optimization variables, and includes multiple optimization objective functions and multiple constraints. The multiple constraints include at least environmental constraints generated based on the environmental constraint data and grid access constraints generated based on the grid data. A multi-objective optimization algorithm is used to search for the optimal solution set in the feasible solution space, and an adaptive dynamic weight allocation mechanism is introduced to map the multi-objective optimal solution set to the user preference space. Output the optimal solution set. Each solution should include at least the selected aircraft type and number of units to be installed, the aircraft type allocation scheme for each location, the location coordinates, and the calculated values ​​of various evaluation indicators.

[0008] In one embodiment, the plurality of optimization objective functions include at least: a minimum cost per kilowatt-hour based on total life cycle cost and generation; a minimum value per kilowatt-hour based on matching time-series electricity prices with generation output; a maximum net present value based on the difference between generation revenue and cost; a minimum wake efficiency loss based on wake effect calculation; a grid support performance optimization objective based on the unit's contribution to grid support; and a minimum noise impact objective based on sound propagation calculation.

[0009] In one embodiment, the multiple constraints include at least: a capacity constraint that the total installed capacity is not less than the minimum requirement of the project; a spacing constraint that meets the safety distance requirements between adjacent generator sites; a grid connection constraint that the total capacity of the entire site meets the grid transmission capacity and short-circuit ratio requirements; an environmental constraint that the noise in residential areas does not exceed the environmental protection limits; a terrain adaptability constraint that the generator foundation and hoisting platform adapt to the site's topography and geological conditions; a constraint on the consistency or diversity ratio of generator types specified according to project needs; and a resonance avoidance constraint that meets electromagnetic compatibility requirements when multiple generator types are used together.

[0010] In one embodiment, the multi-objective optimization algorithm is an evolutionary algorithm based on non-dominated sorting, comprising: An initial population is generated using a hybrid encoding strategy, wherein the mapping from aircraft position to aircraft type is encoded discretely, and the aircraft position coordinates and hub height are encoded continuously. The initial population is used as the current population to start the iteration. Based on the wind resource characteristic parameters output by the refined wind resource assessment and the impeller diameter parameter among the characteristic parameters of the candidate turbine models, the wake influence value between each unit is calculated, and the fitness of each individual in the current population is evaluated in combination with each optimization objective function. Based on the fitness assessment results, a selection operator is used to select superior individuals from the current population, and a new generation of individuals is generated using crossover and mutation operators to update the current population. The fitness evaluation, selection, crossover, and mutation operations of the current population are performed repeatedly until the iterative convergence condition is met. After each generation, non-dominated sorting and diversity preservation operations are performed on each individual in the population to update the optimal frontier. The optimization terminates when the convergence condition is met, and the current best frontier is output as the optimal solution set.

[0011] In one embodiment, an adaptive dynamic weight allocation mechanism is also introduced to dynamically adjust the weight coefficient of each optimization objective in the search guidance according to the degree of deviation between the current value of each optimization objective and its optimal value during each iteration, so that the objective with the greater deviation receives a higher weight, thereby guiding the algorithm to converge toward the user's preferred region.

[0012] In one embodiment, the joint optimization includes: The wake influence range of each unit is determined based on the differences in impeller diameter among different models; A quantitative assessment of the wake asymmetry effect caused by different impeller sizes when arranging mixer models is conducted. By optimizing the relationship between the unit layout direction and the prevailing wind direction, the area of ​​wake overlap can be reduced; Furthermore, within the framework of the multi-objective optimization model, the generator position coordinates are used as optimization variables and generator type allocation variables for synchronous iterative optimization, so as to achieve a collaborative solution for generator selection and spatial layout.

[0013] In one embodiment, time-series generation revenue optimization is also included: A typical annual electricity price dataset is constructed based on electricity market price data for different periods of historical years; Based on the power curves of the candidate units and the time series data of wind resources, the power generation output sequence of each candidate unit at each time period of the year was simulated. The power generation output sequence is matched with the electricity price of the same period in the typical electricity price annual dataset on a time-by-time basis, the power generation revenue of each period is calculated and accumulated to obtain the time-series total revenue; Based on the time-series total revenue, the power generation strategy of each candidate unit is optimized in reverse, so as to increase output during high electricity price periods, reduce output during low electricity price periods, or arrange shutdown for maintenance.

[0014] Furthermore, to achieve the above objectives, this application also proposes a wind turbine selection device, which includes: The first module is used to collect multi-source data of the target wind farm. The multi-source data includes at least wind measurement data, topographic and geographic data, environmental constraint data, power grid data, and meteorological reanalysis data. The second module is used to conduct a refined assessment of wind resources using the wind measurement data, the topographic and geographic data and the meteorological reanalysis data by employing a wind flow field numerical simulation method, and outputs wind resource characteristic parameters at each candidate machine location. The wind resource characteristic parameters include at least wind speed parameters, turbulence parameters and limiting wind speed parameters. The third module is used to filter candidate turbine models from the wind turbine product database, extract the feature parameters of the candidate models, perform secondary filtering based on the wind resource feature parameters to eliminate models that do not meet the conditions, and construct a multi-objective optimization model based on the candidate models retained after secondary filtering. The multi-objective optimization model uses the combination of candidate models, the model allocation scheme of each turbine location, and the turbine location coordinates as optimization variables, and includes multiple optimization objective functions and multiple constraints. The multiple constraints include at least environmental constraints generated based on the environmental constraint data and grid access constraints generated based on the grid data. The multi-objective optimization algorithm is used to search for the optimal solution set in the feasible solution space, and an adaptive dynamic weight allocation mechanism is introduced to map the multi-objective optimal solution set to the user preference space. The fourth module is used to output the optimal solution set. Each solution includes at least the selected aircraft type and number of units installed, the aircraft type allocation scheme and coordinates of each unit, and the calculated values ​​of various evaluation indicators.

[0015] In addition, to achieve the above objectives, this application also proposes a wind turbine selection device, the device comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, the computer program being configured to implement the steps of the wind turbine selection method described above.

[0016] In addition, to achieve the above objectives, this application also proposes a storage medium, which is a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, it implements the steps of the wind turbine selection method described above.

[0017] In addition, to achieve the above objectives, this application also provides a computer program product, which includes a computer program that, when executed by a processor, implements the steps of the wind turbine selection method described above.

[0018] One or more technical solutions proposed in this application have at least the following technical effects: First, by collecting multi-source data including wind measurement data, topographic and geographic data, environmental constraint data, power grid data, and meteorological reanalysis data, and employing a wind flow field numerical simulation method (CFD method) for refined wind resource assessment, a leap from single-point wind measurement to area-wide multi-source data fusion assessment has been achieved. Compared with existing methods that rely on single wind tower data and linear engineering empirical models, this significantly improves the accuracy and reliability of wind resource assessment, and directly uses the assessment results as input for subsequent optimization models, effectively solving the problem of the disconnect between assessment and decision-making.

[0019] Secondly, by screening a set of candidate models from the wind turbine product database, a second screening process is conducted based on wind resource characteristic parameters to eliminate models that do not meet the criteria, thus constructing a completely objective, quantifiable, and repeatable automated screening process. Compared to existing methods that rely on engineers manually comparing products one by one from the catalog, this significantly improves screening efficiency and eliminates subjective bias and human omissions.

[0020] Furthermore, an optimization model was constructed, using candidate aircraft type combinations, aircraft type allocation schemes for each aircraft location, and aircraft location coordinates as optimization variables and incorporating multiple objective functions. An adaptive dynamic weight allocation mechanism was introduced to map the multi-objective optimal solution set to the user preference space. Compared to existing methods that use maximizing power generation or minimizing LCOE as a single objective, or employ fixed subjective weights for weighted summation, this approach can obtain a series of non-dominated optimal solutions (Pareto fronts) through multi-objective optimization, allowing decision-makers to flexibly choose according to actual needs. The adaptive weight mechanism further avoids biases caused by subjective weighting.

[0021] Next, the constraints include at least environmental constraints generated based on environmental constraint data and grid access constraints generated based on grid data. Compared with existing methods that only consider basic constraints such as installed capacity and turbine spacing, this method systematically incorporates hard constraints such as grid connection point short-circuit ratio, transmission channel capacity, noise limits, and ecological red lines into the optimization model, ensuring that the selection results meet the actual construction requirements in terms of engineering feasibility and compliance.

[0022] Furthermore, by simultaneously incorporating the aircraft type allocation scheme and aircraft location coordinates of each location as optimization variables into the same multi-objective optimization model for synchronous iterative optimization, joint solutions for aircraft selection and spatial layout are achieved. Compared to existing methods that separate aircraft selection and micro-location into steps, this approach seeks global optima in both aircraft selection and location layout dimensions, effectively reducing overall wake efficiency loss and avoiding local optima in step-by-step optimization.

[0023] Finally, a complete intelligent selection process was constructed through standardized collection of multi-source data, automated CFD evaluation, database-driven objective screening, and automatic optimization based on multi-objective optimization algorithms. Compared with existing methods that rely on manual comparison and selection, this significantly shortens the selection cycle and improves the accuracy and reproducibility of the results. Attached Figure Description

[0024] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0025] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0026] Figure 1 This is a first flowchart illustrating the first embodiment of the wind turbine selection method of this application. Figure 2 This is a second flowchart illustrating the first embodiment of the wind turbine selection method of this application. Figure 3 This is a third flowchart illustrating the first embodiment of the wind turbine selection method of this application. Figure 4 This is a fourth flowchart illustrating the first embodiment of the wind turbine selection method of this application. Figure 5 This is a schematic diagram of the module structure of the wind turbine selection device according to an embodiment of this application; Figure 6This is a schematic diagram of the equipment structure of the hardware operating environment involved in the wind turbine selection method in this application embodiment.

[0027] The purpose, features, and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0028] It should be understood that the specific embodiments described herein are merely illustrative of the technical solutions of this application and are not intended to limit this application.

[0029] To better understand the technical solution of this application, a detailed description will be provided below in conjunction with the accompanying drawings and specific implementation methods.

[0030] It should be noted that the executing entity in this embodiment can be a computing service device with data processing, network communication, and program execution functions, such as a tablet computer, personal computer, or mobile phone, or an electronic device or wind turbine selection device capable of performing the above functions. The following description uses a wind turbine selection device as an example to illustrate this embodiment and the subsequent embodiments.

[0031] Based on this, the embodiments of this application provide a method for selecting wind turbine generators, referring to... Figure 1 , Figure 1 This is a flowchart illustrating the first embodiment of the wind turbine selection method of this application.

[0032] In this embodiment, the wind turbine selection method includes steps S10 to S40: Step S10: Collect multi-source data of the target wind farm. The multi-source data includes at least wind measurement data, topographic and geographic data, environmental constraint data, power grid data, and meteorological reanalysis data. In this embodiment, step S10 specifically includes the following sub-steps.

[0033] S11: Wind Data Acquisition. A wind measurement tower should be erected within the target wind farm area. The height of the tower should be determined based on the hub height of the turbine model, and is generally not lower than the hub height. Preferably, for projects with a hub height of 120m, the wind measurement tower height should be at least 120m, with anemometers and wind direction indicators installed at multiple height levels (10m, 30m, 50m, 70m, 90m, 120m). The wind measurement period should be no less than a full year, including at least one major wind month and one minor wind month to ensure that the wind data covers the wind conditions of different seasons throughout the year. The wind data should include at least wind speed, wind direction, temperature, air pressure, and air density parameters at each height level. The wind speed sampling frequency should be no less than 1Hz, with a recording interval of 10-minute averages. Wind direction data must be magnetically declination corrected to true north. For projects with existing wind measurement towers, existing tower data can be used directly. For projects without towers, temporary towers can be erected, or virtual tower technology can be used to generate wind data through the fusion of mesoscale and microscale meteorological data. The number of wind measurement towers is determined based on the area and terrain complexity of the wind farm: for flat terrain with an area less than 50 km²... 2 For wind farms with complex terrain or large areas, the number of wind measurement towers should be increased appropriately, with an increase of 50km for each additional tower. 2 Add one wind measurement tower.

[0034] S12: Topographic and Geographic Data Acquisition. Obtain digital elevation model (DEM) data of the target wind farm with an accuracy of no less than 1:2000 scale and a grid resolution preferably between 5m×5m and 30m×30m. For mountainous wind farms with complex terrain, high-resolution DEM data of 5m×5m is preferred. DEM data can be obtained through UAV aerial surveying, lidar ranging, or satellite remote sensing. Simultaneously, collect land use type data and surface roughness data of the target wind farm. Land use type data can be obtained from land resources departments or based on satellite imagery interpretation and classification. Surface roughness is classified into levels 0 to 4 according to international standards. The typical surface types corresponding to each level are as follows: Level 0 is water surface (roughness length 0.0002m), Level 1 is open land (roughness length 0.03m), Level 2 is farmland (roughness length 0.1m), Level 3 is forest land (roughness length 0.5m), and Level 4 is urban built-up area (roughness length 1.0m or more).

[0035] S13: Collection of Environmental Constraint Data. Collect environmental constraint data for the target wind farm and its surrounding area, including but not limited to: the location and boundary range of noise-sensitive buildings such as residential areas, schools, and hospitals; the specific scope of ecological protection red lines, the boundaries of nature reserves and their control requirements; the protection scope and construction control zones of cultural relics protection units; and the scope and control requirements of special control areas such as airport airspace and airways. The above data can be obtained through official channels such as local environmental protection departments, natural resources departments, cultural relics departments, and military management agencies, or by querying and verifying publicly available geographic information data. For each environmental constraint area, detailed information such as its location coordinates, protection level, and control requirements needs to be recorded. In particular, for noise-sensitive buildings, the building type (residential, school, hospital, etc.), number of floors, and usage function also need to be recorded to accurately determine the evaluation criteria during subsequent noise impact assessments.

[0036] S14: Grid Data Collection. Collect relevant data on the target wind farm's grid connection, including: grid connection point coordinates, grid voltage level, short-circuit capacity, transmission capacity, grid peak-shaving capacity, and time-of-use electricity price data. For time-of-use price data, historical data from the past three years is preferred for constructing a typical annual electricity price dataset. For regions already involved in electricity market trading, price data can be obtained from the local electricity trading center; for regions not yet involved in market trading, the on-grid price approved by the local price authority can be used, along with electricity price information from surrounding areas for reference. Grid data collection also includes technical documents such as grid company approval documents for the grid connection system and grid connection system design plans. The short-circuit ratio at the grid connection point is a crucial parameter for assessing the wind farm's grid support capability; accurate acquisition of the short-circuit ratio under minimum and maximum operating conditions is necessary.

[0037] S15: Acquisition of Meteorological Reanalysis Data. Meteorological reanalysis data for the target wind farm area will be collected, including satellite Earth observation data and numerical weather prediction (NWP) data. Meteorological reanalysis data is used to supplement and validate wind measurement data, especially in areas with insufficient wind tower coverage, where spatial interpolation of wind resources and correction of the representative year can be performed. Global reanalysis datasets such as ERA5 and MERRA-2 are preferred, with a spatial resolution of at least 0.25° × 0.25° and a temporal resolution of at least 1 hour. The meteorological reanalysis data also includes long-term (typically 20–30 years) monthly and annual average data of meteorological elements such as wind speed, wind direction, temperature, and air pressure for the target area, used to correct short-term wind measurement data of the measurement year to long-term average representative year data.

[0038] S16: Data Quality Inspection and Preprocessing. The collected multi-source data is inspected, corrected, and the representative year is corrected to generate a standardized wind resource input dataset. Data quality inspection includes: (1) Integrity inspection: the data integrity rate of each channel of the wind measurement data is checked. The data integrity rate is not less than 90%. Otherwise, the missing data needs to be interpolated; (2) Reasonableness inspection: the wind speed value should be within a reasonable range of 0 to 60 m / s, the wind direction value should be within a range of 0° to 360°, and the temperature and air pressure values ​​should conform to basic physical laws; (3) Consistency inspection: the vertical change of wind speed at different height layers should conform to the wind shear law, and there should be no abrupt changes in the data of adjacent time periods. For abnormal data and missing data, linear interpolation, correlation analysis of adjacent wind measurement tower data, or correlation analysis based on reanalysis data are used for interpolation and correction. The interpolated data should be marked to distinguish it from the original measured data. The representative year correction adopts the long-sequence reference year method. Specifically, annual wind speed data (usually no less than 20 years) from long-term meteorological stations near the target wind farm area are collected, and the long-term average is calculated. The wind speed data of the year in which the wind was measured is compared with the long-term average, and the correction factor is calculated. The wind speed parameters of the year in which the wind was measured are multiplied by the correction factor to obtain the wind speed data of the representative year. For wind direction distribution, it is generally assumed that the interannual variation is small, and the wind direction data of the year in which the wind was measured can be used directly.

[0039] Step S20: Using the wind measurement data, the topographic data, and the meteorological reanalysis data, a wind flow field numerical simulation method is used to conduct a refined assessment of wind resources, and the wind resource characteristic parameters at each candidate machine location are output. The wind resource characteristic parameters include at least wind speed parameters, turbulence parameters, and limiting wind speed parameters. In this embodiment, step S20 specifically includes the following sub-steps.

[0040] S21: Digital Terrain Model Construction. Based on the digital elevation model data and ground roughness data collected in step S12, combined with the wind measurement data from step S11 and the meteorological reanalysis data from step S15, a digital terrain model of the target wind farm is constructed. The terrain model should cover the entire planned area of ​​the wind farm and extend outward by at least 5 km to eliminate boundary effects. The terrain model is discretized using either a structured or unstructured grid. Structured grids are suitable for areas with relatively simple terrain and have higher computational efficiency; unstructured grids are suitable for areas with complex terrain and can better fit complex terrain boundaries. The grid resolution is refined to 5m–20m in the wind turbine layout area and can be appropriately relaxed to 50m–100m in the boundary area. For locations with wind measurement towers, the spatial coordinates and height layer information of the wind measurement towers are embedded into the digital terrain model as verification points for subsequent simulation results. For cases with multiple wind measurement towers, model calibration is performed using data from all wind measurement towers.

[0041] S22: Numerical Simulation of Wind Flow Field. Computational Fluid Dynamics (CFD) is used in conjunction with the digital terrain model constructed in step S21 to perform a refined numerical simulation of the wind flow field in the wind farm area. The Reynolds-averaged Navier-Stokes (RANS) equations are used as the flow control equations, and the standard k-ε model or the SST k-ω model is used as the turbulence model. For complex mountainous terrain, the SST k-ω model is preferred, as it has higher computational accuracy in the near-wall region and can more accurately simulate the influence of terrain on boundary layer flow. The boundary conditions are set as follows: the inlet boundary uses a wind speed profile as input, given according to an exponential or logarithmic law, with parameters calibrated from measured data from the anemometer tower; the outlet boundary uses a pressure outlet condition, setting the relative pressure to zero; the top and sides use symmetrical boundary conditions with zero normal velocity components; the ground uses wall boundary conditions, with roughness lengths set according to land use type data. A CFD simulation is performed for each set of inflow conditions (including different wind directions and speed ranges). The wind direction is typically divided into 16 sectors (each sector 22.5°), and the wind speed is divided into several intervals (e.g., 4–25 m / s, with each interval being 1 m / s), forming a complete wind flow field database. The numerical solution employs the finite volume method, and the governing equations are discretized using a second-order upwind scheme to ensure computational accuracy. Pressure-velocity coupling is achieved using either the SIMPLE algorithm or the PISO algorithm. The convergence criterion is that the residuals of each equation decrease to 10. -4 The wind speed values ​​at the monitoring points no longer change significantly with the increase of iteration steps. For wind farms with particularly complex terrain or particularly large computational domains, a partitioned simulation strategy is adopted: the entire wind farm is divided into several terrain sub-regions, each sub-region is simulated in detail, overlapping boundaries are set between adjacent sub-regions to ensure the continuity of simulation results, and finally the simulation results of each sub-region are stitched together to form a complete wind resource map of the entire field.

[0042] S23: Output of wind resource characteristic parameters. Based on the numerical simulation results of the wind flow field in step S22, output the wind resource characteristic parameters at each candidate turbine location. The location of the candidate turbine location can be initially determined based on topographic conditions, land use, environmental impact, etc., or based on gridded layout to uniformly cover the planned area. The output wind resource characteristic parameters include the following types: (1) Wind speed parameters: annual average wind speed, monthly average wind speed, average wind speed of each wind sector, and wind speed frequency distribution at the hub height of each turbine location. The annual average wind speed is calculated by weighting the wind speed values ​​of each wind sector and each wind speed interval obtained from the CFD simulation, combined with the wind frequency distribution in that direction. (2) Turbulence parameters: annual average turbulence intensity at the hub height of each turbine location and turbulence intensity distribution of each wind sector. Turbulence intensity is defined as the ratio of the root mean square of wind speed fluctuation to the average wind speed. (3) Limiting wind speed parameters: 50-year return maximum wind speed and 50-year return maximum wind speed at each turbine location. Based on the extreme wind speed sequence of each wind sector obtained by CFD simulation, the maximum wind speed value with a 50-year return period is estimated by extreme value statistical methods (such as Gumbel distribution fitting). (4) Other characteristic parameters: wind shear index at each turbine location (obtained by wind speed fitting at different height layers), inflow angle (the angle between the airflow direction and the horizontal plane), and wind direction distribution rose diagram (wind frequency and wind energy distribution in each direction). For wind farms with high terrain complexity, due to the large differences in wind resource characteristics between each turbine location, it is necessary to output the wind resource characteristic parameters of each candidate turbine location one by one to form a wind resource characteristic parameter table.

[0043] Step S30: Select a set of candidate turbine models from the wind turbine product database, extract the feature parameters of the candidate models, perform secondary screening based on the wind resource feature parameters to remove models that do not meet the conditions, and construct a multi-objective optimization model based on the candidate models retained after secondary screening. The multi-objective optimization model uses the combination of candidate models, the model allocation scheme of each turbine location, and the turbine location coordinates as optimization variables, and includes multiple optimization objective functions and multiple constraints. The multiple constraints include at least environmental constraints generated based on the environmental constraint data and grid access constraints generated based on the grid data. The multi-objective optimization algorithm is used to search for the optimal solution set in the feasible solution space, and an adaptive dynamic weight allocation mechanism is introduced to map the multi-objective optimal solution set to the user preference space. In this embodiment, step S30 specifically includes the following sub-steps.

[0044] S31: Establishment and maintenance of the wind turbine product database. Establish a wind turbine product database, which includes information on models released and to be launched by major wind turbine manufacturers. The database adopts an scalable data structure to support adding new models and updating existing model information. The database contains the following field categories: (1) Basic parameters: Rated power P rated (Unit: MW), Impeller diameter Drotor (Unit: m), Hub Height H hub (Unit: m, including optional hub height range), Cut-in wind speed V cut_in (Unit: m / s) Rated wind speed V rated (Unit: m / s) Cut-out wind speed V cut_out (Unit: m / s). (2) Aerodynamic performance parameters: power curve PV, stored in discrete point form (wind speed interval 0.5 m / s, covering the entire range from cut-in wind speed to cut-out wind speed), and supports linear interpolation to obtain the power value corresponding to any wind speed; thrust coefficient curve C t -V, also stored in discrete point form, is used for subsequent wake effect calculation. (3) Structural parameters: tower type (steel tower / concrete tower / mixed tower / flexible tower), nacelle weight (unit t), blade material, blade length (unit m), number of blades (usually 3). (4) Reliability parameters: design life (usually 25 years on land, 30 years at sea), annual average availability (usually 97% to 98%), expected replacement cycle of major components (generator, gearbox, blades, main shaft, converter, etc.) (unit year). (5) Cost parameters: cost per kilowatt (yuan / kW), foundation construction cost (estimated value given according to different foundation types), hoisting cost (given according to different hoisting schemes), annual operation and maintenance cost estimate (including material cost and maintenance cost, usually calculated per kilowatt per year). (6) Certification Information: IEC type certification level (e.g., IEC IB, IEC IIB, IEC IIIB, etc.), special environmental adaptability certification (high-altitude / low-temperature / typhoon-resistant / salt spray resistant, etc.), certification body and certification validity period. The database should preferably use a relational database (e.g., MySQL, PostgreSQL, SQLite) for storage and management, and be equipped with a database management interface for maintenance personnel to update data regularly. The database should support filtering and querying based on rated power range, impeller diameter range, certification level, etc.

[0045] S32: Initial screening, selecting candidate turbine models based on the basic conditions of the project. Based on the basic conditions of the target wind farm project, a set of turbine models meeting the conditions is initially screened from the wind turbine product database. Initial screening conditions include, but are not limited to: (1) Rated power range: based on the planned installed capacity P of the project. total Given the number of available machine sites N, a reasonable range for single-machine capacity is initially determined. Single-machine capacity is typically taken as P. totalThe screening range is ±20% of / N. For example, if the planned installed capacity is 200MW and the expected number of turbine sites is 30 to 35, the screening range for single turbine capacity is 5.0MW to 7.5MW. (2) Type certification level: Based on the turbulence intensity and the maximum wind speed that occurs once every 50 years output in step S20, determine the IEC wind zone level corresponding to the target wind farm. The IEC certification level of the model during screening shall not be lower than the level requirement of the target wind farm. For example, when the turbulence intensity is 0.14 and the maximum wind speed that occurs once every 50 years is 37.5m / s, the corresponding IEC level is IIB, and the model certification level is required to be IEC IIB or higher (such as IA) during screening. (3) Environmental adaptability: Based on the climate and environmental conditions of the project location, screen models with corresponding special environmental adaptability certifications. For example, high-altitude projects (altitude exceeding 2000m) require screening for high-altitude certified models; low-temperature projects (extreme temperatures below -20℃) require screening for low-temperature certified models; coastal or offshore projects require screening for models with typhoon resistance and salt spray resistance certification; and projects in frigid regions require screening for models with ice-freezing resistance certification.

[0046] S33: Secondary screening, eliminating models that do not meet the conditions based on wind resource characteristic parameters. Based on the wind resource characteristic parameters output in step S20, a secondary screening is performed on the candidate models after the initial screening to eliminate models that have poor operating performance or cannot operate normally under the target wind field conditions.

[0047] The specific judgment conditions for secondary screening include: (1) Cut-in wind speed matching: The cut-in wind speed of the turbine model should be lower than the annual average wind speed of the target turbine location. Usually, the cut-in wind speed is required to be no higher than 50% of the annual average wind speed to ensure that the unit can generate electricity normally for most of the time. For example, if the annual average wind speed of the wind farm is 6.8 m / s, the cut-in wind speed of the turbine model is required to be no higher than 3.4 m / s. (2) Rated wind speed matching: The rated wind speed of the turbine model should match the average wind speed distribution characteristics of the target wind farm. If the rated wind speed is too high (significantly higher than the average wind speed), the unit will not be able to reach the rated power for most of the time, and the full-load hours will be low; if the rated wind speed is too low (close to or lower than the average wind speed), the unit will frequently reach the rated power and then limit the power through pitch control, resulting in a waste of wind energy resources. Preferably, the rated wind speed is 1.4 to 1.8 times the annual average wind speed of the target wind farm. (3) Cut-out wind speed adaptability: The cut-out wind speed of the turbine model should be higher than the maximum wind speed of the target wind field once in 50 years, or at least higher than the maximum sustained wind speed that may occur during normal operation. For turbine models with low cut-out wind speeds (such as 25 m / s), frequent cut-out shutdowns may occur in typhoon-prone areas, affecting power generation, and they should be excluded. (4) Extreme wind speed tolerance: The extreme wind speed tolerance (i.e., design extreme wind speed) of the turbine model should not be lower than the maximum wind speed once in 50 years, and a certain safety margin should be reserved. Preferably, the extreme wind speed tolerance of the turbine model is required to be ≥1.1× the maximum wind speed once in 50 years. For example, if the maximum wind speed once in 50 years is 32.5 m / s, then the design extreme wind speed of the turbine model is required to be ≥35.75 m / s. (5) Turbulence intensity adaptability: The upper limit of turbulence intensity corresponding to the IEC certification level of the turbine model should not be lower than the turbulence intensity at the target turbine location. If the turbulence intensity at the target turbine location exceeds the upper limit corresponding to the turbine's certification level, the fatigue load the unit experiences during operation will exceed the design value, affecting the unit's lifespan and reliability. After initial screening and secondary screening, the set of remaining candidate turbine models is denoted as M={1,2,…,M}, where M is the total number of candidate turbine models, typically 5 to 15.

[0048] S34: Extraction and formatting of candidate model feature parameters. For each candidate model retained after secondary screening, its complete feature parameters are extracted from the database and formatted into the input data format required for the optimization model. Specifically, this includes: (1) Extracting the power curve P j (V) and thrust coefficient curve C t,j (V), and output in the form of a discrete array with the same wind speed interval (preferably 0.5m / s) to ensure the consistency of curve data for each model and facilitate subsequent calculations. (2) Extract impeller diameter D j Rated power P rated,j Wheel hub height range (minimum wheel hub height H) j,min and maximum wheel hub height H j,max(3) Extract the unit cost C per kilowatt. turbine,j Basic Fee C foundation,j Cost parameters such as operation and maintenance costs are used for full life cycle cost calculation. (4) Extract the design life T years j Average annual availability η j Reliability parameters such as replacement cycle of major components are used for power generation reduction and operation and maintenance cost calculation. (5) Extract IEC certification level and short-circuit capacity contribution coefficient (SCR). contrib,j Grid-friendly parameters such as grid-type control capability identifiers are used for power grid support performance evaluation. (6) Extract sound power level L W,j (Usually the overall sound power level, given separately according to different operating conditions), for use in noise impact assessment.

[0049] Step S40: Output the optimal solution set. Each solution includes at least the selected aircraft type and number of units installed, the aircraft type allocation scheme and coordinates of each unit, and the calculated values ​​of various evaluation indicators.

[0050] In this embodiment, step S40 specifically includes the following sub-steps.

[0051] S41: Define optimization variables. The optimization variables of the multi-objective optimization model include: (1) turbine type allocation variable: Suppose that there are N candidate turbine locations in the wind farm, and the turbine type allocated to the i-th turbine location is x. i x i ∈{1,2,…,M}, i=1,2,…,N. x i For discrete variables, integer encoding is used, and their values ​​correspond to the index number of the candidate aircraft type in set M. (2) Aircraft position coordinate variable: the planar coordinate (X) of the i-th aircraft position. i ,Y i ), i=1,2,…,N. X i and Y i As a continuous variable, its value range is limited by factors such as the boundary of the wind farm planning land, terrain conditions, and environmental constraints. The coordinate unit is meters (m), and the coordinate system is consistent with the coordinate system of the DEM data in step S12. (3) Machine location-machine type pairing relationship: optimization variable x i and (X) i ,Y i The variables (x, y, y) together determine the type of aircraft used at the i-th location and its spatial location. A complete optimization scheme consists of the variables (x, y, y) of all locations. i ,X i ,Y i Together, they constitute the complete vector representation of the optimization variables: z = [x1, x2, ..., x...]. N ,X1,Y1,X2,Y2,…,XN ,Y N The vector has a dimension of 3N, with N discrete variables (aircraft type allocation) and 2N continuous variables (aircraft position coordinates).

[0052] S42: Constructing an optimization objective function set. Based on the characteristic parameters of each candidate turbine type retained after secondary screening, the wind resource characteristic parameters of each turbine location output in step S20, and the power grid data and environmental constraint data collected in step S10, an objective function set containing multiple optimization objective functions is constructed. The calculation of each objective function takes the optimization variable z as input, and the specific calculation process is as follows: For any optimization scheme z, firstly, x is allocated according to the turbine type of each turbine location. i Extract the power curve P of the corresponding model xi (V), Thrust coefficient curve C t,xi (V), Impeller diameter D xi Parameters are specified; then, for each turbine location i, the annual power generation per unit is calculated based on the wind speed distribution at that point and the power curve of the turbine type; then, based on the coordinates (X, Y, Z) of each turbine location... i , Yi ) and the impeller diameter D of each model xi The wake effects between each unit are calculated, and the actual power generation of each unit is corrected. Finally, the power generation of each unit, the cost parameters of each unit type, grid data and environmental data are combined to calculate the objective function value.

[0053] In one feasible implementation, the plurality of optimization objective functions include at least: a minimum cost per kilowatt-hour based on total life cycle cost and power generation; a minimum value per kilowatt-hour based on matching time-series electricity prices with power generation output; a maximum net present value based on the difference between power generation revenue and cost; a minimum wake efficiency loss based on wake effect calculation; a grid support performance optimization objective based on the unit's contribution to grid support; and a minimum noise impact objective based on sound propagation calculation.

[0054] The objective function set specifically includes the following objectives: (a) Minimizing the levelized cost of electricity (LCOE): reflecting the economic viability of the wind farm throughout its entire life cycle, using the levelized cost of electricity (LCOE) as the evaluation index. (b) Minimizing the value cost of electricity (LCOV): reflecting the economic viability of the wind farm in the electricity market environment, using the value cost of electricity (LCOV) as the evaluation index, and incorporating the matching degree between time-series electricity prices and power generation output into the evaluation. (c) Maximizing the net present value (NPV): reflecting the net income of the wind farm throughout its entire life cycle, using the net present value (NPV) as the evaluation index, and comprehensively considering power generation revenue, operation and maintenance costs, and initial investment. (d) Minimizing wake efficiency loss: reflecting the degree of power loss of downstream units due to the wake effect of upstream units. (e) Optimizing grid support performance: reflecting the wind farm's voltage and frequency support capability for the grid, comprehensively evaluating the short-circuit capacity contribution and grid configuration control capability of the units. (f) Minimizing noise impact: reflecting the degree of noise impact of wind farm operation on the surrounding environment. The above multiple objective functions constitute the objective vector F(z)=[f1(z),f2(z),…,f K [(z)], where K is the total number of objective functions. There may be conflicting relationships between the objective functions (such as increasing power generation may increase costs and noise). The purpose of multi-objective optimization is to find the optimal trade-off between these conflicting objectives.

[0055] S43: Set constraints. In one feasible implementation, the multiple constraints include at least: a capacity constraint that the total installed capacity is not less than the minimum requirement of the project; a spacing constraint that meets the safety distance requirements between adjacent generator sites; a grid connection constraint that meets the grid transmission capacity and short-circuit ratio requirements for the total capacity of the entire site; an environmental constraint that the noise in residential areas does not exceed the environmental protection limits; a terrain adaptability constraint that the generator foundation and hoisting platform adapt to the site's topography and geological conditions; a constraint on the consistency or diversity ratio of generator types specified according to project needs; and a resonance avoidance constraint that meets electromagnetic compatibility requirements when multiple generator types are used together.

[0056] The constraints of the multi-objective optimization model include the following types: (1) Capacity constraint: The total installed capacity shall not be less than the minimum capacity requirement P of the project. min, to ensure that the installed capacity of the project is met. (2) Spacing constraints: The distance between any two turbine locations shall not be less than the safety spacing requirement. The safety spacing is related to the rotor diameter of the installed turbine model to ensure that there is no structural interference between the units under extreme working conditions and to control the degree of wake influence. (3) Grid access capacity constraints: The total capacity of the entire site shall be within the capacity range of the transmission channel connected to the grid, and the short-circuit ratio of the grid connection point shall meet the minimum requirements of the grid operator to ensure that the safe and stable operation of the grid is not endangered after the wind farm is connected. (4) Environmental constraints: Based on the environmental constraint data collected in step S13, including that the calculated noise value at the boundary of all residential areas does not exceed the limit specified by the local environmental protection standards, and that the site selection and construction of the wind turbines do not occupy the prohibited construction areas such as ecological protection red lines, nature reserves, and cultural relic protection areas. (5) Terrain adaptability constraints: The foundation size and hoisting platform size of the turbines must be adapted to the site topography and geological conditions of each turbine location and shall not exceed the available construction site range of each turbine location. (6) Model consistency / diversity constraints: Specify a lower limit for the proportion of the same model based on the project operation strategy (same model facilitates spare parts management, maintenance personnel training and repair), or an upper and lower limit range for the proportion of mixed models. (7) Electromagnetic compatibility and resonance constraints: When multiple models are used, the switching frequencies of different models of converters do not generate resonance superposition, avoiding the introduction of harmful harmonics or the risk of subsynchronous oscillation in the power grid. The above constraints constitute the feasible solution space of the optimization problem. The values ​​of all optimization variables z must simultaneously satisfy all constraints in order to be considered as feasible solutions for subsequent optimization searches.

[0057] S44: Use a multi-objective optimization algorithm to search for the optimal solution set in the feasible solution space. Use a multi-objective optimization algorithm to search for the Pareto optimal solution set in the feasible solution space defined by steps S41 to S43. The basic process of the algorithm is as follows: (1) Initialize the solution set: Randomly generate an initial population in the feasible solution space. Each individual in the population corresponds to a complete optimization scheme z. The initial population size N popThe number of individuals is usually determined according to the problem size, typically ranging from 100 to 500. To ensure the feasibility and diversity of the initial population, some individuals are generated using heuristic strategies based on the wind rose diagram and terrain conditions (e.g., assigning the generator with the greatest power generation potential to the most efficient generator), while the remaining individuals are randomly generated within the feasible solution space. (2) Feasibility verification: For each generated individual, check whether it meets all the constraints defined in step S43. Individuals that do not meet the constraints are corrected or discarded through constraint handling strategies. Constraint handling can adopt the penalty function method, that is, adding a penalty term proportional to the degree of constraint violation to the objective function value; or it can adopt the constraint dominance principle, that is, giving priority to individuals with smaller constraint violations when sorting non-dominated. (3) Iterative optimization: The current population is iterated through genetic operations such as selection, crossover, and mutation. In each generation, based on the wind resource assessment results of step S20 and the characteristic parameters of each candidate model, the objective function values ​​of each individual are calculated; based on the objective function values, non-dominated sorting is performed to identify the Pareto optimal individuals in the population; excellent individuals are retained through selection operations, and new individuals are generated through crossover and mutation operations, gradually approaching the Pareto optimal frontier. (4) Pareto frontier update: After each generation, the Pareto optimal solution set is updated, old solutions dominated by newly discovered excellent solutions are removed, and all non-dominated solutions are retained. When the Pareto frontier no longer shows significant improvement for several consecutive generations (e.g., 20 to 50 generations), the algorithm converges and the iteration is terminated. (5) Adaptive dynamic weight allocation: In the iterative process of the multi-objective optimization algorithm, an adaptive dynamic weight allocation mechanism is introduced. This mechanism dynamically adjusts the weight coefficients of each objective in the search guidance according to the degree of deviation between the current value of each optimization objective and its expected optimal value. Specifically, during the iteration process, the average or optimal value of each objective in the current population is statistically analyzed and compared with the theoretical optimal value (or the user-defined expected value) of each objective to calculate the deviation rate. Objectives with larger deviation rates receive higher weights in subsequent searches, thus guiding the algorithm to converge towards the user's preferred region. Adaptive weight adjustment is performed every generation or every few generations, with weight changes transitioning smoothly generation by generation to avoid drastic oscillations in the search direction. The adaptive dynamic weight allocation mechanism enables the final Pareto optimal solution set to be optimized in a targeted manner according to the differentiated needs of wind farm projects. For example, in grid parity scenarios, the focus is on reducing LCOE; in electricity market trading scenarios, the focus is on optimizing LCOV; in peak shaving and frequency regulation scenarios, the focus is on grid support performance; and in ecologically sensitive areas, the focus is on reducing noise impact. Users can select the corresponding preference configuration according to the actual needs of the project, and the algorithm promotes the convergence of the optimal solution set towards the user's preferred direction through adaptive weight adjustment. Finally, step S40 outputs a Pareto optimal solution set containing multiple non-dominated solutions, each solution corresponding to a selection scheme that achieves the optimal trade-off among multiple objectives. The subsequent steps involve outputting and comparing these options, allowing decision-makers to select the final implementation plan based on the actual needs of the project.

[0058] In one feasible implementation, refer to Figure 2 The multi-objective optimization algorithm is an evolutionary algorithm based on non-dominated sorting, comprising: Step A1: An initial population is generated using a hybrid encoding strategy, wherein the mapping from aircraft position to aircraft type is encoded discretely, and the aircraft position coordinates and hub height are encoded continuously. The initial population is used as the current population to start the iteration. First, the set of variables to be optimized is determined. Taking a planned 200MW wind farm as an example, 35 candidate turbine locations are set. After screening in step S30, 6 candidate turbine models are retained. The optimization variables include: turbine model allocation variables for the 35 locations (discrete type, values ​​1-6) and turbine coordinate variables for the 70 locations (continuous type, X and Y coordinates). A hybrid coding strategy is adopted. The turbine model allocation variables are encoded with integers, with each gene location taking an integer from 1 to 6 to represent the corresponding turbine model index; the turbine coordinate variables are encoded with real numbers, directly storing the coordinate values. An initial population is generated based on the above coding scheme, with a population size of 200. 10% of the individuals are generated using a heuristic strategy, for example, assigning upstream turbine locations in the prevailing wind direction to high-efficiency large-rotor turbine models and downstream turbine locations to medium-capacity turbine models. The remaining 90% of the individuals are randomly generated within the feasible solution space to ensure diversity. All individuals in the initial population meet the basic constraints such as turbine spacing (not less than 5 times the rotor diameter), land boundary, and environmentally restricted construction zones.

[0059] Step A2: Calculate the wake impact value between each unit based on the wind resource characteristic parameters output by the refined wind resource assessment and the impeller diameter parameter among the characteristic parameters of the candidate turbine models, and evaluate the fitness of each individual in the current population in combination with each optimization objective function; For each individual in the current population, fitness is assessed individually. First, based on the individual's turbine type allocation and coordinate information, the wind speed, turbulence intensity, and other parameters at each turbine location are obtained from the wind resource assessment results in step S20. Combined with the power curves of each turbine type, the theoretical power generation per unit is calculated. Then, based on the rotor diameter parameters of each turbine type and the coordinates of each turbine location, the wake impact values ​​between all upstream and downstream turbine pairs are calculated. Specifically, for any two turbines i and j, the distance and relative azimuth angle are calculated based on their coordinates. Combined with the rotor diameter and thrust coefficient curves of the upstream turbine, the cumulative velocity loss wake model is used to calculate the wind speed loss and power loss of the downstream turbine due to the wake obstruction of the upstream turbine. The actual power generation of each turbine in the entire field is accumulated to obtain the total power generation of the entire field. Based on cost parameters and power generation results, economic target values ​​such as LCOE and NPV are calculated. Based on the wake superposition results, the wake loss rate of the entire field is calculated. Based on the noise propagation model, the noise values ​​at each sensitive point are calculated. All target values ​​constitute the fitness vector of this individual.

[0060] Step A3: Based on the fitness assessment results, select superior individuals from the current population using the selection operator, and generate a new generation of individuals using the crossover and mutation operators, and update the current population with the new generation of individuals; Based on the fitness assessment results of step A2, a tournament selection method is used to select parent individuals from the current population. Two individuals are randomly selected each time, and their Pareto non-dominance levels are compared; the one with the higher level wins. If the levels are the same, crowding distance is compared, and the one with the larger distance wins. This process is repeated until a sufficient number of parent individuals are selected. Then, crossover and mutation operations are performed on the parent individuals: for the integer encoding of the model assignment part, two-point crossover is used, randomly selecting two crossover points to exchange gene segments; for the real number encoding of the coordinate part, simulated binary crossover is used. In the mutation operation, the model assignment gene is randomly transformed into other candidate model indices with a certain probability, and the coordinate gene undergoes polynomial mutation with a certain probability. The crossover probability is set to 0.9, and the mutation probability is set to 1 / total number of variables. Through selection, crossover, and mutation, a new generation population with the same size as the parent population is generated.

[0061] Step A4: Iteratively perform fitness evaluation, selection, crossover, and mutation operations on the current population until the iterative convergence condition is met. After each generation of iteration, perform non-dominated sorting and diversity preservation operations on each individual in the population to update the optimal frontier. After each generation of genetic operations, the parent and offspring populations are merged, and all individuals in the merged population are sorted into non-dominated layers. Specifically, all individuals not dominated by other individuals are identified as the first non-dominated layer, and these individuals are removed. New non-dominated individuals are then identified from the remaining individuals to form the second layer, and this process is repeated until all individuals are sorted. Within the same non-dominated layer, the crowding distance of each individual is calculated, which is the sum of its distances to neighboring individuals in the target space, to maintain solution diversity. The merged population is sorted in ascending order by non-dominated layer level and in descending order within the same layer by crowding distance, and the first N_pop individuals are selected as the next generation population. Simultaneously, all individuals at level 1 in the current non-dominated layer are recorded, and the Pareto optimal front is updated.

[0062] Step A5: Terminate the optimization when the iterative convergence condition is met, and output the current optimal frontier as the optimal solution set.

[0063] After each generation, convergence conditions are checked. This embodiment uses two criteria for joint determination: first, the average hypervolume change rate of the Pareto front between two adjacent generations is less than a threshold of 1×10. -3Second, no new Pareto-dominated solutions are generated for 30 consecutive generations. The algorithm is considered convergent and the iteration terminates when either condition is met. In this embodiment, the algorithm meets the convergence condition at generation 420, and after termination, it outputs all non-dominated individuals on the Pareto optimal front as the optimal solution set, totaling 37 candidate schemes. Each scheme includes the allocation and coordinate information of 35 turbine positions, as well as the calculated values ​​of various evaluation indicators such as LCOE, annual power generation, wake loss, and NPV, for decision-makers to make the final comparison.

[0064] In one feasible implementation, an adaptive dynamic weight allocation mechanism is also introduced to dynamically adjust the weight coefficient of each optimization objective in the search guidance according to the degree of deviation between the current value of each optimization objective and its optimal value during each iteration, so that the objective with the greater deviation receives a higher weight, thereby guiding the algorithm to converge toward the user's preferred region.

[0065] During the iterative process of the multi-objective optimization algorithm, adaptive dynamic weight allocation is performed to guide the search direction towards the user's preferred region. After each iteration, the current optimal value of each optimization objective in the current population is first calculated. Simultaneously, the expected optimal value of each objective is obtained, which is pre-set by the user according to project requirements or determined based on the independent single-objective optimization results of each objective. The current optimal value of each objective is compared with its expected optimal value, and the degree of deviation between the two is calculated. The greater the deviation, the less ideal the objective's performance in the current search process, requiring a higher search priority. The weight adjustment amount is calculated based on the deviation of each objective. Objectives with larger deviations receive positive adjustments, increasing their weights; objectives with smaller deviations have their weights decreased accordingly. The weight adjustment adopts a smooth change strategy, with the single adjustment magnitude controlled by the learning rate parameter to avoid drastic fluctuations in the search direction due to sudden weight changes. After all objectives have completed weight adjustments, normalization is performed to keep the sum of the weights of all objectives constant. The learning rate adaptively changes according to the iteration process: a larger learning rate is used in the early stages of iteration to accelerate the convergence speed towards the user's preferred region; a smaller learning rate is used in the later stages of iteration to ensure search stability and finely adjust the balance between objectives. User preferences are reflected in the setting of expected optimal values. For different wind farm project types, users set different combinations of expected optimal values ​​to achieve differentiated weight configuration guidance. During the iteration process, the weights of each objective are dynamically recorded and visualized, facilitating user monitoring of the evolution of the algorithm's search direction. This adaptive mechanism is executed once in each iteration or every preset number of iterations, continuously guiding the algorithm to focus on optimizing the currently least desirable objective in subsequent iterations until all objectives are close to the user's expectations or the algorithm converges. The final Pareto optimal solution set is distributed near the user's preference region.

[0066] In one feasible implementation, refer to Figure 3The joint optimization includes: Step B1: Determine the wake influence range of each unit based on the differences in impeller diameter among different models; In the evaluation of each generation of optimization schemes, the corresponding impeller diameter parameters are first extracted based on the turbine type assigned to each turbine location. For each pair of upstream and downstream units, the wake influence range of the upstream unit is determined by its impeller diameter and the thrust coefficient at the current wind speed. Specifically, based on the impeller diameter and combined with the wake expansion coefficient (usually taken as 0.04 to 0.08), the lateral expansion range of the wake during its downstream propagation is calculated. For turbines with larger impeller diameters, their wake influence range is correspondingly larger, and the influence distance on downstream units is farther. By traversing all unit pairs in the entire site, it is determined whether the downstream unit falls within the wake influence cone of the upstream unit, thereby establishing a complete wake influence relationship matrix and clarifying the upstream and downstream influence relationships among all units in the entire site.

[0067] Step B2: Quantitatively evaluate the wake asymmetry effect caused by different impeller sizes when arranging mixer models; When the impeller diameters of the upstream and downstream units differ, the wake effect exhibits significant asymmetric characteristics. This step calculates the differences in wake effect for two scenarios: one where the upstream large impeller affects the downstream small impeller, and another where the upstream small impeller affects the downstream large impeller. When the upstream impeller diameter is larger than the downstream impeller diameter, the wake's lateral expansion range upon reaching the downstream unit is relatively large, potentially covering the entire swept area of ​​the downstream unit or even exceeding it, resulting in the downstream unit being entirely within the influence zone of the upstream wake. Conversely, when the upstream impeller diameter is smaller than the downstream impeller diameter, the wake's influence range upon reaching the downstream position may only cover a portion of the downstream swept area, with some blade areas of the downstream unit still in free flow. The actual power loss needs to be determined by integrating and averaging the swept area. This step employs a swept area weighted averaging method to refine the wake velocity loss under different impeller diameter combinations, accurately quantifying the asymmetric wake effect caused by impeller size differences.

[0068] Step B3: Reduce the wake overlap area by optimizing the relationship between the unit layout direction and the prevailing wind direction; Based on the wake impact calculation results in steps B1 and B2, the wake loss distribution among all turbine locations across the entire site is determined. During the search process, the cumulative degree of wake impact on each turbine location in different wind directions is calculated for each candidate scheme. When multiple upstream turbines exist at a turbine location upwind, that location is in a wake superposition area, resulting in a significant increase in power generation loss. By adjusting the arrangement direction of each turbine location, the turbines are staggered along the prevailing wind direction, reducing the number of turbines on the same wake path in the same wind direction. In specific implementation, cross-iteration is applied to the turbine location coordinates in the optimization variables. After each adjustment, the wake loss value across the entire site is recalculated. Driven by the objective function of a multi-objective optimization algorithm, layout schemes with high wake losses are gradually eliminated, retaining superior layouts with more uniform wake distribution and fewer superposition areas.

[0069] Step B4 involves using the turbine location coordinates as optimization variables and turbine type allocation variables within the framework of the multi-objective optimization model to perform synchronous iterative optimization, thereby achieving a collaborative solution for turbine selection and spatial layout.

[0070] In each iteration of the multi-objective optimization algorithm, both the aircraft type allocation variable (discrete value) and the aircraft position coordinate variable (continuous value) participate in selection, crossover, and mutation operations. Each individual contains both types of variables and is uniformly evaluated by the objective function and tested against the constraints. In the non-dominated ranking, the changes in wake loss caused by the adjustment of aircraft position coordinates and the changes in power generation caused by the change in aircraft type allocation are comprehensively evaluated. The aircraft position coordinates are searched in continuous space using simulated binary crossover and polynomial mutation, while the aircraft type allocation is searched in discrete space using two-point crossover and random mutation. The two types of variables evolve alternately in the same generation: coordinate adjustment improves the spatial layout to reduce wake, while aircraft type adjustment changes the output characteristics and wake generation intensity at each point. Through synchronous iteration, each scheme in the final Pareto optimal solution set achieves synergistic optimization of aircraft type configuration and spatial layout, rather than local optima in a single dimension.

[0071] In one feasible implementation, refer to Figure 4 It also includes time-series power generation revenue optimization: Step C1: Construct a typical annual electricity price dataset based on electricity market price data for each period of historical years; Based on the time-of-use electricity price data collected in step S10, a typical annual electricity price dataset is constructed. Specifically, historical electricity market price data for the past three years in the target wind farm's location are collected, organized into a complete annual electricity price sequence of 8760 hours with hourly time granularity. Outliers in the data (such as extremely high or negative prices) are identified and smoothed using statistical methods. If multiple trading instruments exist in the region (such as day-ahead market or real-time balancing market), the corresponding electricity price data source is selected according to the actual trading type the project participates in, or the comprehensive electricity price is calculated by weighting the participation ratio of different trading instruments. For regions that have not yet entered market-based trading, the time-of-use on-grid electricity price approved by the local price authority can be used as substitute data, and appropriate corrections can be made by referring to the electricity price fluctuation characteristics of neighboring market-based regions.

[0072] Step C2: Simulate the power output sequence of each candidate turbine location for each period of the year based on the power curves of the candidate turbines and the wind resource time series data; For each selected option in the Pareto optimal solution set output in step S40, based on the 8760-hour annual wind speed time-series data for each turbine location output in step S20, and combined with the power curves of the turbine types allocated to each location, the hourly power generation sequence for the entire year is calculated. The wind speed time-series data is derived from the fusion of CFD simulation results and wind measurement time-series data in step S20. Specifically, based on the measured wind speed at the wind measurement tower, a wind speed transfer function between each turbine location and the wind measurement tower is established through CFD simulation, mapping the wind speed at the wind measurement tower to the wind speed at each turbine location. For time periods with insufficient wind measurement tower coverage, meteorological reanalysis data collected in step S15 is used for supplementation. Linear interpolation is used for power curve interpolation to ensure that each wind speed value corresponds to a unique power output value.

[0073] Step C3: Match the power generation output sequence with the electricity prices of the same period in the typical electricity price annual dataset on a time-by-time basis, calculate the power generation revenue of each time period, and sum them up to obtain the time-series total revenue; The hourly power generation output sequence obtained in step C2 is matched with the electricity price for the same time period (the same hour in the same year) in step C1. The revenue for each hour is equal to the power generation (output multiplied by 1 hour) multiplied by the corresponding electricity price for that hour. The revenues of all 8760 hours are summed to obtain the time-series total revenue of this selected scheme under the annual time-series electricity price conditions. This calculation method differs from the traditional fixed electricity price revenue calculation, and can accurately reflect the degree of matching between the power generation output time sequence and electricity price fluctuations, identifying the power generation distribution characteristics of each scheme during high-electricity-price periods and low-electricity-price periods.

[0074] Step C4: Based on the time-series total revenue, perform reverse optimization on the power generation strategy of each candidate unit to increase output during periods of high electricity price, reduce output during periods of low electricity price, or schedule shutdown for maintenance.

[0075] Based on the calculation results of step C3, a reverse optimization analysis is performed on the power generation strategies of each candidate unit. Specifically, the top 20% of high-electricity-price periods and the bottom 20% of low-electricity-price periods in the annual electricity price sequence are identified, and the power generation output characteristics of each scheme during these periods are statistically analyzed. For schemes with insufficient output during high-electricity-price periods, the feasibility and marginal benefits of increasing output by adjusting operation control strategies (such as early wind speed optimization and pitch control parameter adjustment) are evaluated. For schemes with excessive output during low-electricity-price periods, the impact of moderate capacity reduction or scheduled maintenance shutdowns on the total annual revenue is evaluated. Finally, time-series revenue optimization suggestions for each scheme are output for decision-makers to refer to and implement during the operation phase after the selected scheme is determined.

[0076] It should be noted that the above examples are only for understanding this application and do not constitute a limitation on the wind turbine selection method of this application. Any simple modifications based on this technical concept are within the protection scope of this application.

[0077] This application also provides a wind turbine selection device; please refer to... Figure 5 The wind turbine selection device includes: The first module 10 is used to collect multi-source data of the target wind farm. The multi-source data includes at least wind measurement data, topographic and geographic data, environmental constraint data, power grid data, and meteorological reanalysis data. The second module 20 is used to perform a refined assessment of wind resources using the wind measurement data, the topographic and geographic data and the meteorological reanalysis data by employing a wind flow field numerical simulation method, and outputs wind resource characteristic parameters at each candidate machine location. The wind resource characteristic parameters include at least wind speed parameters, turbulence parameters and limiting wind speed parameters. The third module 30 is used to filter a set of candidate turbine models from the wind turbine product database, extract the feature parameters of the candidate turbine models, perform secondary filtering based on the wind resource feature parameters to remove models that do not meet the conditions, and construct a multi-objective optimization model based on the candidate turbine models retained after secondary filtering. The multi-objective optimization model uses the combination of candidate turbine models, the turbine model allocation scheme for each turbine location, and the turbine location coordinates as optimization variables, and includes multiple optimization objective functions and multiple constraints. The multiple constraints include at least environmental constraints generated based on the environmental constraint data and grid access constraints generated based on the grid data. The multi-objective optimization algorithm is used to search for the optimal solution set in the feasible solution space, and an adaptive dynamic weight allocation mechanism is introduced to map the multi-objective optimal solution set to the user preference space. The fourth module 40 is used to output the optimal solution set. Each solution includes at least the selected aircraft type and number of units installed, the aircraft type allocation scheme and coordinates of each unit, and the calculated values ​​of various evaluation indicators.

[0078] The wind turbine selection device provided in this application, employing the wind turbine selection method described in the above embodiments, can solve the technical problems in the prior art, such as insufficient data utilization, low evaluation accuracy, reliance on manual screening, one-sided objectives and constraints, and the separation of selection and layout, resulting in low selection efficiency and difficulty in obtaining a globally optimal solution that balances economy, grid adaptability, and environmental friendliness. Compared with the prior art, the beneficial effects of the wind turbine selection device provided in this application are the same as those of the wind turbine selection method provided in the above embodiments, and other technical features in the wind turbine selection device are the same as those disclosed in the methods of the above embodiments, and will not be repeated here.

[0079] This application provides a wind turbine selection device, which includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the wind turbine selection method in the above embodiment 1.

[0080] The following is for reference. Figure 6 The diagram illustrates a structural schematic suitable for implementing wind turbine selection equipment in the embodiments of this application. The wind turbine selection equipment in the embodiments of this application may include, but is not limited to, mobile terminals such as mobile phones, laptops, digital radio receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Description), PMPs (Portable Media Players), vehicle terminals (e.g., vehicle navigation terminals), and fixed terminals such as digital TVs and desktop computers. Figure 6 The wind turbine selection equipment shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of this application.

[0081] like Figure 6As shown, the wind turbine selection device may include a processing unit 1001 (e.g., a central processing unit, a graphics processing unit, etc.), which can perform various appropriate actions and processes according to a program stored in a read-only memory 1002 or a program loaded from a storage device 1003 into a random access memory 1004. The random access memory 1004 also stores various programs and data required for the operation of the wind turbine selection device. The processing unit 1001, the read-only memory 1002, and the random access memory 1004 are interconnected via a bus 1005. An input / output interface 1006 is also connected to the bus. Typically, the following systems can be connected to the input / output interface 1006: input devices 1007 including, for example, a touchscreen, touchpad, keyboard, mouse, image sensor, microphone, accelerometer, gyroscope, etc.; output devices 1008 including, for example, a liquid crystal display (LCD), speaker, vibrator, etc.; storage devices 1003 including, for example, magnetic tape, hard disk, etc.; and communication devices 1009. Communication device 1009 allows the wind turbine selection equipment to communicate wirelessly or wiredly with other equipment to exchange data. Although the figure shows wind turbine selection equipment with various systems, it should be understood that implementation or possession of all the systems shown is not required. More or fewer systems may be implemented alternatively.

[0082] Specifically, according to the embodiments disclosed in this application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments disclosed in this application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device, or installed from storage device 1003, or installed from read-only memory 1002. When the computer program is executed by processing device 1001, it performs the functions defined in the methods of the embodiments disclosed in this application.

[0083] The wind turbine selection equipment provided in this application, employing the wind turbine selection method described in the above embodiments, can solve the technical problems in the prior art where insufficient data utilization, low evaluation accuracy, reliance on manual screening, one-sided objectives and constraints, and a disconnect between selection and layout lead to low selection efficiency and difficulty in obtaining a globally optimal solution that balances economy, grid adaptability, and environmental friendliness. Compared with the prior art, the beneficial effects of the wind turbine selection equipment provided in this application are the same as those of the wind turbine selection method provided in the above embodiments, and other technical features of this wind turbine selection equipment are the same as those disclosed in the previous embodiment method, and will not be repeated here.

[0084] It should be understood that the various parts disclosed in this application can be implemented using hardware, software, firmware, or a combination thereof. In the description of the above embodiments, specific features, structures, materials, or characteristics can be combined in any suitable manner in one or more embodiments or examples.

[0085] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0086] This application provides a computer-readable storage medium having computer-readable program instructions (i.e., a computer program) stored thereon, which are used to execute the wind turbine selection method in the above embodiments.

[0087] The computer-readable storage medium provided in this application may be, for example, a USB flash drive, but is not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems or devices, or any combination thereof. More specific examples of computer-readable storage media may include, but are not limited to: electrical connections having one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this embodiment, the computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system or device. The program code contained on the computer-readable storage medium may be transmitted using any suitable medium, including but not limited to: wires, optical cables, RF (Radio Frequency), etc., or any suitable combination thereof.

[0088] The aforementioned computer-readable storage medium may be included in the wind turbine selection equipment; or it may exist independently and not be assembled into the wind turbine selection equipment.

[0089] The aforementioned computer-readable storage medium carries one or more programs that, when executed by the wind turbine selection equipment, cause the wind turbine selection equipment to: collect multi-source data from the target wind farm, wherein the multi-source data includes at least wind measurement data, topographic and geographic data, environmental constraint data, power grid data, and meteorological reanalysis data; utilize the wind measurement data, the topographic and geographic data, and the meteorological reanalysis data to perform a refined assessment of wind resources using a wind flow field numerical simulation method, and output wind resource characteristic parameters at each candidate turbine location, wherein the wind resource characteristic parameters include at least wind speed parameters, turbulence parameters, and limiting wind speed parameters; filter a set of candidate turbine models from a wind turbine product database, extract characteristic parameters of the candidate turbine models, and perform secondary analysis based on the wind resource characteristic parameters. The first screening eliminates models that do not meet the criteria. Based on the candidate models retained after the second screening, a multi-objective optimization model is constructed. The multi-objective optimization model uses the combination of candidate models, the model allocation scheme for each unit, and the unit coordinates as optimization variables, and includes multiple optimization objective functions and multiple constraints. The multiple constraints include at least environmental constraints generated based on the environmental constraint data and grid access constraints generated based on the grid data. A multi-objective optimization algorithm is used to search for the optimal solution set in the feasible solution space, and an adaptive dynamic weight allocation mechanism is introduced to map the multi-objective optimal solution set to the user preference space. The optimal solution set is output, and each solution includes at least the selected model category and number of units installed, the model allocation scheme for each unit, the unit coordinates, and the calculated values ​​of various evaluation indicators.

[0090] Computer program code for performing the operations of this application can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a Local Area Network (LAN) or a Wide Area Network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0091] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0092] The modules described in the embodiments of this application can be implemented in software or hardware. The names of the modules do not necessarily limit the functionality of the unit itself.

[0093] The readable storage medium provided in this application is a computer-readable storage medium that stores computer-readable program instructions (i.e., a computer program) for executing the above-described wind turbine selection method. This solves the technical problems in the prior art, such as insufficient data utilization, low evaluation accuracy, reliance on manual screening, one-sided objectives and constraints, and a disconnect between selection and layout, leading to low selection efficiency and difficulty in obtaining a globally optimal solution that balances economy, grid adaptability, and environmental friendliness. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided in this application are the same as those of the wind turbine selection method provided in the above embodiments, and will not be repeated here.

[0094] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the wind turbine selection method described above.

[0095] The computer program product provided in this application can solve the technical problems in the prior art, which are low selection efficiency and difficulty in obtaining a globally optimal solution that takes into account economy, grid adaptability and environmental friendliness due to insufficient data utilization, low evaluation accuracy, reliance on manual screening, one-sided objectives and constraints and separation of selection and layout. Compared with the prior art, the beneficial effects of the computer program product provided in this application are the same as those of the wind turbine selection method provided in the above embodiments, and will not be repeated here.

[0096] The above description is only a part of the embodiments of this application and does not limit the patent scope of this application. All equivalent structural transformations made under the technical concept of this application and using the contents of the specification and drawings of this application, or direct / indirect applications in other related technical fields, are included in the patent protection scope of this application.

Claims

1. A method for selecting wind turbine generator sets, characterized in that, The wind turbine selection method includes: Collect multi-source data from the target wind farm, including at least wind measurement data, topographic and geographic data, environmental constraint data, power grid data, and meteorological reanalysis data; The wind resources are finely assessed using the wind measurement data, the topographic and geographic data, and the meteorological reanalysis data through a wind flow field numerical simulation method. The wind resource characteristic parameters at each candidate machine site are output, including at least wind speed parameters, turbulence parameters, and limiting wind speed parameters. A candidate turbine model set is selected from the wind turbine product database, and the characteristic parameters of the candidate turbine models are extracted. A secondary selection is performed based on the wind resource characteristic parameters to remove turbine models that do not meet the conditions. A multi-objective optimization model is constructed based on the candidate turbine models retained after the secondary selection. The multi-objective optimization model uses the combination of candidate turbine models, the turbine model allocation scheme of each turbine location, and the turbine location coordinates as optimization variables, and includes multiple optimization objective functions and multiple constraints. The multiple constraints include at least environmental constraints generated based on the environmental constraint data and grid access constraints generated based on the grid data. A multi-objective optimization algorithm is used to search for the optimal solution set in the feasible solution space, and an adaptive dynamic weight allocation mechanism is introduced to map the multi-objective optimal solution set to the user preference space. Output the optimal solution set. Each solution should include at least the selected aircraft type and number of units to be installed, the aircraft type allocation scheme for each location, the location coordinates, and the calculated values ​​of various evaluation indicators.

2. The wind turbine selection method as described in claim 1, characterized in that, The multiple optimization objective functions include at least the following: minimizing the cost per kilowatt-hour based on the total life cycle cost and the generation output; minimizing the value per kilowatt-hour based on the matching of time-series electricity prices and generation output; maximizing the net present value based on the difference between generation revenue and cost; minimizing the wake efficiency loss based on wake effect calculation; optimizing the grid support performance based on the unit's contribution to grid support; and minimizing the noise impact based on sound propagation calculation.

3. The wind turbine selection method as described in claim 1, characterized in that, The multiple constraints include at least the following: capacity constraint that the total installed capacity is not less than the minimum requirement of the project; spacing constraint that the safety distance between adjacent unit sites is met; grid connection constraint that the total capacity of the entire site meets the grid transmission capacity and short-circuit ratio requirements; environmental constraint that noise in residential areas does not exceed environmental protection limits; terrain adaptability constraint that the unit foundation and hoisting platform are adapted to the site's topography and geological conditions; constraint on the consistency or diversity ratio of unit types specified according to project needs; and resonance avoidance constraint that meets electromagnetic compatibility requirements when multiple unit types are used together.

4. The wind turbine selection method as described in claim 1, characterized in that, The multi-objective optimization algorithm is an evolutionary algorithm based on non-dominated sorting, including: An initial population is generated using a hybrid encoding strategy, wherein the mapping from aircraft position to aircraft type is encoded discretely, and the aircraft position coordinates and hub height are encoded continuously. The initial population is used as the current population to start the iteration. Based on the wind resource characteristic parameters output by the refined wind resource assessment and the impeller diameter parameter among the characteristic parameters of the candidate turbine models, the wake influence value between each unit is calculated, and the fitness of each individual in the current population is evaluated in combination with each optimization objective function. Based on the fitness assessment results, a selection operator is used to select superior individuals from the current population, and a new generation of individuals is generated using crossover and mutation operators to update the current population. The fitness evaluation, selection, crossover, and mutation operations of the current population are performed repeatedly until the iterative convergence condition is met. After each generation, non-dominated sorting and diversity preservation operations are performed on each individual in the population to update the optimal frontier. The optimization terminates when the convergence condition is met, and the current best frontier is output as the optimal solution set.

5. The wind turbine selection method as described in claim 1, characterized in that, It also includes the introduction of an adaptive dynamic weight allocation mechanism, which dynamically adjusts the weight coefficient of each optimization objective in the search guidance according to the degree of deviation between the current value of each optimization objective and its optimal value in each iteration, so that the objective with the greater deviation gets a higher weight, thereby guiding the algorithm to converge toward the user's preferred region.

6. The wind turbine selection method as described in claim 1, characterized in that, The joint optimization includes: The wake influence range of each unit is determined based on the differences in impeller diameter among different models; A quantitative assessment of the wake asymmetry effect caused by different impeller sizes when arranging mixer models is conducted. By optimizing the relationship between the unit layout direction and the prevailing wind direction, the area of ​​wake overlap can be reduced; Furthermore, within the framework of the multi-objective optimization model, the generator position coordinates are used as optimization variables and generator type allocation variables for synchronous iterative optimization, so as to achieve a collaborative solution for generator selection and spatial layout.

7. The wind turbine selection method as described in claim 1, characterized in that, It also includes time-series power generation revenue optimization: A typical annual electricity price dataset is constructed based on electricity market price data for different periods of historical years; Based on the power curves of the candidate units and wind resource time series data, the power generation output sequence of each candidate unit at each time period of the year was simulated. The power generation output sequence is matched with the electricity price of the same period in the typical electricity price annual dataset on a time-by-time basis, the power generation revenue of each period is calculated and accumulated to obtain the time-series total revenue; Based on the time-series total revenue, the power generation strategy of each candidate unit is optimized in reverse, so as to increase output during high electricity price periods, reduce output during low electricity price periods, or arrange shutdown for maintenance.

8. A wind turbine selection device, characterized in that, The device includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, the computer program being configured to implement the steps of the wind turbine selection method as described in any one of claims 1 to 7.

9. A storage medium, characterized in that, The storage medium is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by a processor, it implements the steps of the wind turbine selection method as described in any one of claims 1 to 7.

10. A computer program product, characterized in that, The computer program product includes a computer program that, when executed by a processor, implements the steps of the wind turbine selection method as described in any one of claims 1 to 7.

Images