Model selection method based on wind generating set and related equipment

Abstract

The invention discloses a model selection method based on a wind generating set and related equipment, and relates to the technical field of wind power generation, and the method comprises the steps: obtaining power generation environment information to determine multiple groups of model selection schemes corresponding to the wind generating set, and determining the equipment cost corresponding to each group of model selection scheme, predicting generating capacity corresponding to different model selection schemes by using a preset prediction model, determining maintenance cost and service life corresponding to different model selection schemes, and determining a recommendation score corresponding to each group of model selection schemes based on the equipment cost, the predicted generating capacity, the maintenance cost, the service life and a preset model selection strategy, and determining a target type selection scheme in the multiple groups of type selection schemes. According to the method, the recommendation score corresponding to each type selection scheme is determined based on the power generation environment information by considering the equipment cost, the maintenance cost, the service life and the predicted generating capacity corresponding to the wind generating sets of different type selection schemes and combining the large model, so that the optimal type selection scheme is determined, and the overall income is maximized.

Description

Technical Field

[0001] This application relates to the field of wind power generation technology, and in particular to selection methods and related equipment based on wind turbine generator sets. Background Technology

[0002] With the rapid development of the wind power industry, wind turbine selection has become a crucial aspect of wind farm planning and construction. Traditional selection methods rely heavily on manual experience, making decisions by simply comparing the technical parameters of different models. This approach suffers from problems such as a single evaluation dimension, strong subjectivity, and difficulty in achieving global optimization. While related technologies consider the impact of environmental factors on power generation, they remain insufficient in comprehensive cost-benefit analysis, lacking a systematic and quantitative assessment of equipment procurement costs, operation and maintenance costs, and the overall life-cycle power generation revenue.

[0003] Meanwhile, traditional methods struggle to handle complex constraints involving multiple variables and nonlinearity, failing to dynamically adapt to the individualized needs of different sites. This results in poor economic efficiency in unit selection and extended investment payback periods. Therefore, there is an urgent need for a multi-dimensional intelligent unit selection method that comprehensively considers environmental adaptability, cost input, and power generation revenue, enabling precise unit selection decisions and maximizing overall benefits through big data and artificial intelligence technologies. Summary of the Invention

[0004] The main purpose of this application is to provide a method and related equipment for selecting wind turbine generator sets, aiming to solve the technical problem of how to comprehensively consider multiple factors when selecting wind turbine generator sets.

[0005] To achieve the above objectives, this application proposes a selection method based on wind turbine generator sets, which includes: In response to wind turbine generator selection instructions, obtain the corresponding power generation environment information for the wind turbine generator; Based on the power generation environment information, multiple selection schemes corresponding to the wind turbine generator set are determined, and the equipment cost corresponding to each selection scheme is determined. Based on the power generation environment information, a preset prediction model is used to predict the power generation corresponding to different selection schemes, obtain the predicted power generation, and determine the maintenance cost and service life corresponding to different selection schemes. Based on the equipment cost, the predicted power generation, the maintenance cost, the service life, and the preset selection strategy, a recommended score is determined for each selection scheme to identify the target selection scheme among the multiple selection schemes.

[0006] In one embodiment, the step of obtaining the power generation environment information corresponding to the wind turbine generator in response to the wind turbine generator selection command further includes: In response to wind turbine generator selection instructions, acquire corresponding geographical environment data, climate and meteorological data, power grid condition data, and geological condition data for wind turbine generators; The geographic environment data, the climate and meteorological data, the power grid condition data, and the geological condition data are all preprocessed to obtain preprocessed geographic environment data, climate and meteorological data, power grid condition data, and geological condition data. The data preprocessing includes data cleaning and data completion. Feature extraction operations are performed on the preprocessed geographic environment data, climate and meteorological data, power grid condition data, and geological condition data to obtain the power generation environment information corresponding to the wind turbine generator.

[0007] In one embodiment, the step of determining multiple selection schemes corresponding to the wind turbine generator set based on the power generation environment information further includes: Based on the wind power generation environment information and the preset wind turbine generator selection rules, the selection constraint information is determined; Based on the selection constraint information, multiple candidate models are selected from the preset unit database, and the parameters corresponding to each candidate model are determined. Based on the parameters and the various candidate models, multiple selection schemes are generated.

[0008] In one embodiment, the step of predicting the power generation corresponding to different selection schemes using a preset prediction model based on the power generation environment information, and obtaining the predicted power generation, further includes: Based on the power generation environment information, the wind frequency distribution is determined, and the wind turbine power curves corresponding to different selection schemes are determined; Based on the wind frequency distribution, the wind turbine power curve, and the Weibull distribution fitting, a preset prediction model is used to predict the power generation corresponding to different selection schemes, and the predicted power generation is obtained.

[0009] In one embodiment, the step of predicting the power generation corresponding to different selection schemes based on the wind frequency distribution, the wind turbine power curve, and the Weibull distribution fitting, and obtaining the predicted power generation, further includes: Based on the wind frequency distribution, the wind turbine power curve and the Weibull distribution fitting, the power generation corresponding to different selection schemes is predicted using a preset prediction model to obtain the initial predicted power generation. Determine the electrical and environmental losses corresponding to different selection schemes during the power generation process, and predict the performance degradation factor of wind turbines corresponding to different selection schemes; Based on the electrical losses, the environmental losses, and the annual performance degradation factor of the wind turbine, the initial predicted power generation is corrected to obtain the corrected predicted power generation.

[0010] In one embodiment, before the step of predicting the power generation corresponding to different selection schemes using a preset prediction model, the method further includes: Acquire sample data, which corresponds to a first prediction result; The sample data is processed using the current prediction model to obtain a second prediction result; Determine whether the first prediction result is consistent with the second prediction result; If the first prediction result is inconsistent with the second prediction result, the parameters of the current prediction model are adjusted. Based on the current prediction model with adjusted parameters, the process of using the current prediction model to process the sample data and obtain the second prediction result is repeated until the first prediction result is consistent with the second prediction result, thus obtaining the preset prediction model.

[0011] Furthermore, to achieve the above objectives, this application also proposes a selection device based on wind turbine generator sets, the selection device based on wind turbine generator sets comprising: The acquisition module is used to acquire the power generation environment information corresponding to the wind turbine in response to the wind turbine selection instruction. The first determining module is used to determine multiple selection schemes corresponding to the wind turbine generator set based on the power generation environment information, and to determine the equipment cost corresponding to each selection scheme. The prediction module is used to predict the power generation corresponding to different selection schemes based on the power generation environment information using a preset prediction model, obtain the predicted power generation, and determine the maintenance cost and service life corresponding to different selection schemes. The second determining module is used to determine the recommended score corresponding to each selection scheme based on the equipment cost, the predicted power generation, the maintenance cost, the service life and the preset selection strategy, so as to determine the target selection scheme among the multiple selection schemes.

[0012] In one embodiment, the acquisition module further includes: The first acquisition unit is used to acquire geographical environment data, climate and meteorological data, power grid condition data, and geological condition data corresponding to the wind turbine generator set in response to the wind turbine generator set selection instruction. The data preprocessing unit is used to preprocess the geographic environment data, the climate and meteorological data, the power grid condition data, and the geological condition data to obtain preprocessed geographic environment data, climate and meteorological data, power grid condition data, and geological condition data. The data preprocessing includes data cleaning and data completion. The feature extraction unit is used to perform feature extraction operations on the preprocessed geographic environment data, climate and meteorological data, power grid condition data, and geological condition data to obtain the power generation environment information corresponding to the wind turbine generator.

[0013] In one embodiment, the first determining module further includes: The first determining unit is used to determine the selection constraint information based on the wind power generation environment information and the preset wind turbine generator set selection rules; The filtering unit is used to filter out multiple candidate models from the preset unit database based on the selection constraint information, and determine the parameters corresponding to each candidate model; The generation unit is used to generate multiple selection schemes based on the parameters and the multiple candidate models.

[0014] In one embodiment, the prediction module further includes: The second determining unit is used to determine the wind frequency distribution based on the power generation environment information, and to determine the wind turbine power curves corresponding to different selection schemes; The first prediction unit is used to predict the power generation corresponding to different selection schemes based on the wind frequency distribution, the wind turbine power curve and the Weibull distribution fitting, and obtain the predicted power generation.

[0015] In one embodiment, the prediction module further includes: The second prediction unit is used to predict the power generation corresponding to different selection schemes based on the wind frequency distribution, the wind turbine power curve and the Weibull distribution fitting, and to obtain the initial predicted power generation. The third determining unit is used to determine the electrical and environmental losses corresponding to different selection schemes during the power generation process, and to predict the wind turbine performance degradation factor corresponding to different selection schemes; The correction unit is used to correct the initial predicted power generation based on the electrical losses, the environmental losses, and the annual performance degradation factor of the wind turbine, so as to obtain the corrected predicted power generation.

[0016] In one embodiment, the wind turbine generator selection device further includes a model training module, which further includes: The second acquisition unit is used to acquire sample data, which corresponds to the first prediction result; A data processing unit is used to process the sample data using the current prediction model to obtain a second prediction result; A judgment unit is used to determine whether the first prediction result is consistent with the second prediction result; The iterative training unit is used to adjust the parameters of the current prediction model if the first prediction result is inconsistent with the second prediction result, and based on the current prediction model with adjusted parameters, return to the step of using the current prediction model to process the sample data to obtain the second prediction result, until the first prediction result is consistent with the second prediction result, and obtain the preset prediction model.

[0017] In addition, to achieve the above objectives, this application also proposes a selection device based on wind turbine generator sets, 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 selection method based on wind turbine generator sets as described above.

[0018] 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 generator selection method described above.

[0019] 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 generator selection method described above.

[0020] One or more technical solutions proposed in this application have at least the following technical effects: This application proposes a selection method and related equipment for wind turbine generator sets, relating to the field of wind power generation technology. Traditional selection methods rely heavily on manual experience, making decisions by simply comparing the technical parameters of different models. This approach suffers from problems such as a single evaluation dimension, strong subjectivity, and difficulty in achieving global optimization. While related technologies consider the impact of environmental factors on power generation, they are still insufficient in comprehensive cost-benefit analysis, lacking a systematic quantitative assessment of equipment procurement costs, operation and maintenance costs, and the entire life-cycle power generation revenue. Furthermore, traditional methods struggle to handle complex constraints involving multiple variables and nonlinearity, failing to dynamically adapt to the personalized needs of different sites, resulting in poor economic efficiency of the selection scheme and a prolonged investment payback period. In this application, firstly, in response to a wind turbine generator selection instruction, the power generation environment information corresponding to the wind turbine generator is obtained. Then, based on the power generation environment information, multiple selection schemes corresponding to the wind turbine generator are determined, and the equipment cost corresponding to each selection scheme is determined. Further, based on the power generation environment information, a preset prediction model is used to predict the power generation corresponding to different selection schemes, and the predicted power generation is obtained. The maintenance cost and service life corresponding to different selection schemes are also determined. Finally, based on the equipment cost, the predicted power generation, the maintenance cost, the service life, and the preset selection strategy, a recommended score is determined for each selection scheme to determine the target selection scheme among the multiple selection schemes.

[0021] Understandably, this application considers four aspects—equipment cost, maintenance cost, service life, and predicted power generation—for wind turbine generator sets with different selection schemes based on power generation environment information. Combining this with a large model, it determines the recommended score for each selection scheme and then determines the optimal selection scheme to maximize overall benefits. Attached Figure Description

[0022] 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.

[0023] To more clearly illustrate the technical solutions in the embodiments of this application or related technologies, the accompanying drawings used in the description of the embodiments or related technologies will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0024] Figure 1 This is a flowchart illustrating an embodiment of the wind turbine generator selection method of this application. Figure 2 This is a flowchart illustrating Embodiment 2 of the wind turbine generator selection method provided in this application; Figure 3This is a flowchart illustrating Embodiment 3 of the wind turbine generator selection method provided in this application; Figure 4 This is a schematic diagram of the module structure of the wind turbine generator selection device according to an embodiment of this application; Figure 5 This is a schematic diagram of the equipment structure of the hardware operating environment involved in the wind turbine generator selection method in the embodiments of this application.

[0025] 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

[0026] 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.

[0027] 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.

[0028] The main solution in this application's embodiments is: In this embodiment, for ease of description, the following description will focus on the selection of equipment based on wind turbine generator sets.

[0029] Due to the rapid development of the wind power industry, wind turbine selection has become a crucial aspect of wind farm planning and construction. Traditional selection methods rely heavily on manual experience, making decisions by simply comparing the technical parameters of different models. This approach suffers from problems such as a single evaluation dimension, strong subjectivity, and difficulty in achieving global optimization. While related technologies consider the impact of environmental factors on power generation, they remain insufficient in comprehensive cost-benefit analysis, lacking a systematic and quantitative assessment of equipment procurement costs, operation and maintenance costs, and the overall life-cycle power generation revenue.

[0030] Meanwhile, traditional methods struggle to handle complex constraints involving multiple variables and nonlinearity, failing to dynamically adapt to the individualized needs of different sites. This results in poor economic efficiency in unit selection and extended investment payback periods. Therefore, there is an urgent need for a multi-dimensional intelligent unit selection method that comprehensively considers environmental adaptability, cost input, and power generation revenue, enabling precise unit selection decisions and maximizing overall benefits through big data and artificial intelligence technologies.

[0031] This application provides a solution in which: first, in response to a wind turbine generator selection instruction, power generation environment information corresponding to the wind turbine generator is obtained; then, based on the power generation environment information, multiple selection schemes corresponding to the wind turbine generator are determined, and the equipment cost corresponding to each selection scheme is determined; further, based on the power generation environment information, a preset prediction model is used to predict the power generation corresponding to different selection schemes, and the predicted power generation is obtained, and the maintenance cost and service life corresponding to different selection schemes are determined; finally, based on the equipment cost, the predicted power generation, the maintenance cost, the service life, and the preset selection strategy, a recommended score corresponding to each selection scheme is determined, so as to determine the target selection scheme among the multiple selection schemes.

[0032] Understandably, this application considers four aspects—equipment cost, maintenance cost, service life, and predicted power generation—for wind turbine generator sets with different selection schemes based on power generation environment information. Combining this with a large model, it determines the recommended score for each selection scheme and then determines the optimal selection scheme to maximize overall benefits.

[0033] 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 capable of performing the above functions, such as a wind turbine generator selection device. The following description uses a wind turbine generator selection device as an example to illustrate this embodiment and the subsequent embodiments.

[0034] Based on this, the embodiments of this application provide a method for selecting wind turbine generator sets, referring to... Figure 1 , Figure 1 This is a flowchart illustrating the first embodiment of the wind turbine generator selection method of this application.

[0035] In this embodiment, the selection method based on wind turbine generator sets includes steps S10 to S40: Step S10: In response to the wind turbine generator selection command, obtain the power generation environment information corresponding to the wind turbine generator; It should be noted that the wind turbine generator selection instruction refers to the selection task request triggered by the user or system. It usually includes basic information such as site location, project scale, and investment budget, and is the trigger signal to start the intelligent selection process.

[0036] Power generation environmental information refers to all the data that affect the power generation performance of the unit, including meteorological resource data (wind speed, wind direction, turbulence intensity, air density, temperature), geographical condition data (altitude, terrain slope, surface roughness, obstacle distribution), grid connection conditions (grid connection capacity, voltage level, distance), and policy and regulatory constraints (noise limits, environmental protection red lines, land use restrictions).

[0037] In this embodiment, the system first receives the selection instruction and parses the site coordinate range and project requirements; then it automatically calls the multi-source data interface to capture raw data from meteorological databases (such as ERA5 reanalysis data, wind tower measured data), geographic information systems (GIS elevation data), power grid planning databases, and other channels; finally, it performs spatiotemporal alignment, missing value imputation, and quality inspection on the multi-source heterogeneous data to form a structured environmental information dataset.

[0038] Understandably, this step establishes a data-driven, objective decision-making basis, completely replacing the inefficient traditional model of manual data collection. By automating the acquisition of comprehensive environmental information, it ensures that subsequent selection analysis is based on complete, accurate, and traceable objective data, avoiding decision-making biases caused by information omissions or subjective assumptions, and laying a solid foundation for accurate selection.

[0039] Specifically, the step of obtaining the power generation environment information corresponding to the wind turbine generator set in response to the wind turbine generator set selection command further includes steps S11 to S13: Step S11: In response to the wind turbine generator selection command, obtain the geographical environment data, climate and meteorological data, power grid condition data, and geological condition data corresponding to the wind turbine generator. It should be noted that the geographic environment data refers to the topographic features of the site area, including elevation (DEM data), slope and aspect, surface roughness (vegetation cover type), distribution of obstacles (buildings, mountains), and the vector boundary of the site (inflection point coordinates).

[0040] It should be noted that climate and meteorological data refers to atmospheric parameters observed or simulated over a long period of time in the site area, including hourly wind speed and direction time series (usually requiring more than 2 years of historical data), air density, turbulence intensity, extreme wind speed (once every 5 years), temperature, humidity, air pressure, and the probability of occurrence of severe weather (typhoons, lightning, icing).

[0041] It should be noted that grid condition data refers to the power system parameters at the wind farm's grid connection point, including grid structure (voltage level, short-circuit capacity), connection distance (line length), curtailment policy (historical data on wind curtailment rate), electricity pricing mechanism (benchmark price, trading price), and the grid company's requirements for power forecast accuracy.

[0042] It should be noted that geological condition data refers to the geotechnical engineering characteristics of the site foundation, including soil structure, bearing capacity, groundwater level, seismic intensity, frost depth, and corrosivity (soil chloride ion content). These data directly determine the foundation type and construction cost.

[0043] In this embodiment, after receiving the selection instruction, the system initiates a multi-source heterogeneous data acquisition process: First, it extracts geographic environment data from the National Geographic Information Public Service Platform or UAV mapping results to generate a 50m×50m precision digital terrain model; then, it connects to the meteorological reanalysis database or local wind tower to download and parse climate and meteorological data, performing time-series stitching and quality control; further, it obtains power grid condition data from the power grid company's planning department and parses the system access approval documents; finally, it retrieves the site geological survey report and extracts borehole data and geotechnical test parameters as geological condition data. All data is stored aligned according to a unified spatiotemporal reference (WGS84 coordinate system, UTC time).

[0044] Understandably, this step achieves structured collection and classified management of all-element data, solving the problems of scattered data sources, inconsistent formats, and omissions of key information in traditional methods. By clearly defining four categories of basic data, a standardized data foundation is laid for subsequent targeted processing, ensuring that the selection analysis covers the four core influencing factors of wind resources, topography, power grid, and geology. This avoids the failure of a solution due to a single data deficiency, and improves the completeness and reliability of decision-making.

[0045] Step S12: Perform data preprocessing on the geographic environment data, the climate and meteorological data, the power grid condition data, and the geological condition data to obtain preprocessed geographic environment data, climate and meteorological data, power grid condition data, and geological condition data. The data preprocessing includes data cleaning and data completion. It should be noted that data cleaning refers to identifying and correcting outliers, errors, and noise in the original data. For example, removing abnormal wind speed fluctuations caused by sensor failures in wind measurement towers, correcting geographical offsets caused by inconsistent coordinate systems, and unifying the units of measurement for data from different sources.

[0046] It should be noted that data completion refers to the reasonable interpolation or reconstruction of missing data. For example, for wind speed data of missing periods from meteorological towers, interpolation can be performed using the simulation results of mesoscale meteorological models or data from neighboring stations with high correlation to ensure the continuity of the time series.

[0047] In this embodiment, preprocessing is performed on four types of data respectively: Climate and meteorological data cleaning: Abnormal wind speed values ​​are identified using the 3σ criterion, and errors are corrected by combining data consistency checks of adjacent height layers; during completion, three-dimensional spatiotemporal kriging interpolation is performed using the regression relationship between MERRA-2 reanalysis data and field data.

[0048] Geographic environment data cleaning: Use a raster calculator to remove abnormal elevation points in DEM data and use neighborhood mean to smooth noise; use inverse distance weighted interpolation to complete the terrain of missing areas.

[0049] Data cleaning for power grid conditions: standardize the electricity price file format for different provinces, verify the matching of short-circuit capacity and voltage level at the access point; when supplementing missing wind curtailment rate data, use statistical values ​​of similar sites within the province.

[0050] Geological condition data cleaning: Standardize the units of geotechnical test data (e.g., kPa to MPa conversion), and remove abnormal bearing capacity values ​​that do not conform to physical laws; when supplementing data for un-drilled areas, geological sequence models are used for estimation. All preprocessing operations are logged to ensure traceability.

[0051] Understandably, this step achieves quantitative control and reliability improvement of data quality, effectively solving the common problems of incompleteness, inconsistency, and inaccuracy in raw data. By cleaning and removing "dirty data," outliers are prevented from interfering with the prediction accuracy of subsequent models; by completing the data, data continuity is ensured, making time series analysis possible. It provides clean, complete, and reliable input for feature extraction and model training, significantly improving the robustness of the final selection results.

[0052] Step S13: Perform feature extraction on the preprocessed geographic environment data, climate and meteorological data, power grid condition data, and geological condition data to obtain the power generation environment information corresponding to the wind turbine generator set.

[0053] It should be noted that feature extraction refers to the process of transforming high-dimensional, redundant, and unstructured information from raw data into low-dimensional, refined, and model-recognizable feature vectors. Essentially, it is about extracting key indicators that have a substantial impact on unit selection decisions.

[0054] It should be noted that the power generation environment information specifically refers to the extracted structured feature set, which is the quantitative expression of the generalized power generation environment information in step S10, and can be directly input into the prediction model and optimization algorithm.

[0055] In this embodiment, the system performs differentiated feature engineering on each type of data: For climate and meteorological data: extract Weibull distribution parameters (shape parameter k, scale parameter A) to characterize wind frequency characteristics, calculate annual average wind speed, turbulence intensity I15 (turbulence value at 15 m / s), wind shear index, effective wind energy density, extreme wind speed exceedance probability, and generate sector-specific wind rose diagram vectors.

[0056] For geographic environment data: extract average altitude, terrain complexity index (standard deviation / mean), maximum slope, local terrain acceleration ratio of the aircraft site, and wake interaction sensitivity (based on the distance between aircraft sites and wind direction distribution).

[0057] For grid condition data: extract unit investment of access lines (ten thousand yuan / km), wind curtailment rate estimate, time-of-use weighted average of electricity price, and power factor constraint.

[0058] For geological condition data: extract weighted average bearing capacity, foundation type adaptation index (pile foundation / spread foundation), and special geological risk level (such as frozen soil, soft soil).

[0059] In this embodiment, all features are normalized and concatenated into a feature vector of uniform dimension (e.g., 82-dimensional), and the feature importance weights are labeled.

[0060] Understandably, this step achieves an intelligent transformation from raw data to decision-making information, compressing terabytes of raw monitoring data into kilobytes of key feature vectors, achieving a dimensionality reduction of over 90%. The extracted features directly correspond to the core concerns of unit selection (such as the impact of turbulence on load and wake on power generation), enabling subsequent models to focus on effective information and avoid the curse of dimensionality and overfitting. The generated standardized power generation environment information serves as a unified interface, allowing for rapid adaptation to different optimization algorithms and unit databases, significantly improving computational efficiency and cross-project reusability.

[0061] Step S20: Based on the power generation environment information, determine multiple selection schemes corresponding to the wind turbine generator set, and determine the equipment cost corresponding to each selection scheme; It should be noted that multiple selection schemes refer to combinations of candidate turbine models selected from the turbine database. Each scheme includes a complete configuration of specific turbine models (e.g., power 2.5MW / impeller diameter 140m), configuration quantity (e.g., 40 units installed), key parameters (tower height, cut-in / cut-out wind speed), and other elements.

[0062] It should be noted that the equipment cost refers to the purchase cost of a single wind turbine generator set, including the price of the main unit, the price of the tower, and the cost of the foundation embedded parts, but excluding construction and installation costs.

[0063] In this embodiment, the system first performs an initial screening based on parameters such as average wind speed, turbulence intensity, and extreme wind speed in the environmental information, from its built-in turbine database (covering 50+ manufacturers and 200+ models): models that are unsuitable for local wind resources or exceed safety levels are eliminated; then, according to the total project capacity requirement (e.g., 100MW), it automatically generates multiple combination schemes (e.g., Scheme A: 40 2.5MW models; Scheme B: 33 3.0MW models); next, it calls the cost prediction model (based on the random forest algorithm, inputting parameters such as turbine power, impeller diameter, and technical route) to predict the current market purchase price of each model in each scheme; finally, it summarizes and calculates the total equipment investment cost for each scheme.

[0064] Understandably, this step enables intelligent generation and rapid comparison of candidate solutions, overcoming the limitations of traditional manual comparison of only 3-5 models. By automatically matching environmental adaptability and technical parameters through algorithms, it ensures that all candidate solutions are technically feasible. At the same time, quantifying equipment costs avoids information asymmetry during the price inquiry process, significantly expanding the scope of comparison and providing the possibility of finding the globally optimal solution.

[0065] Specifically, the step of determining multiple selection schemes corresponding to the wind turbine generator set based on the power generation environment information further includes steps S21 to S23: Step S21: Based on the wind power generation environment information and the preset wind turbine generator selection rules, determine the selection constraint information; It should be noted that the preset wind turbine selection rules refer to calculable judgment criteria that are transformed from industry standards and project requirements. These include mandatory rules (such as safety level matching: the extreme wind speed must be less than the turbine's design wind speed), recommended rules (such as turbulence adaptability: the actual turbulence intensity should be less than 90% of the turbine's design turbulence value), and project-customized rules (such as single-unit capacity preference and shortlist of manufacturers).

[0066] It should be noted that the selection constraint information refers to the set of mathematical constraint expressions formed after the above rules are instantiated. It includes hard constraints (which must be met, otherwise the solution is not feasible, such as a safety distance ≥ 5 times the impeller diameter) and soft constraints (which are recommended to be met, and can be subject to quantifiable penalties, such as road construction costs not exceeding 5% of the total investment).

[0067] In this embodiment, the system first analyzes the key feature values ​​in the power generation environment information (such as a 50-year return period extreme wind speed of 52 m / s and a turbulence intensity I15 of 0.16); then it matches the preset selection rules one by one: for example, rule 1: safety-related rules generate hard constraints - filtering out models with a design wind speed ≥ 52 m / s and a design turbulence ≥ 0.16; rule 2: efficiency-related rules generate soft constraints - such as requiring models with a capacity factor ≥ 25% to enter the candidate list; rule 3: project restriction-related rules generate boundary constraints - such as the total installed capacity must be between 99-101 MW and the single unit capacity must not be less than 2.0 MW; rule 4: manufacturer preference rules generate filtering conditions - such as only considering the top 10 turbine manufacturers.

[0068] In this embodiment, all constraints are converted into Boolean expressions or inequalities and stored in the constraint manager to form structured constraint information.

[0069] Understandably, this step enables a preliminary quantitative assessment of the feasibility of the selection process, transforming vague empirical rules into clear computational constraints and avoiding the waste of computational resources on ineffective solutions. By quickly eliminating insecure models through hard constraints and guiding the search direction through soft constraints, the candidate solution generation process becomes systematic, compliant, and controllable, fundamentally eliminating the significant risk of "selected solutions failing to be implemented" and improving the rigor of decision-making and project security.

[0070] Step S22: Based on the selection constraint information, select multiple candidate models from the preset unit database and determine the parameters corresponding to each candidate model; It should be noted that the preset unit database refers to the built-in structured model information database, which contains complete technical parameters and business data of mainstream models on the market. Each record corresponds to one model, and the fields cover more than 50 dimensions such as power, impeller diameter, tower height, design wind speed, design turbulence, certification information, reference price, reliability index (MTBF), and power curve (wind speed-power mapping table).

[0071] It should be noted that the candidate models refer to the set of qualified models retained after screening through the S21 constraints. Each model meets the technical and safety requirements of the site environment and is qualified to participate in the subsequent optimization and selection.

[0072] It should be noted that parameters specifically refer to the attributes of candidate models that have quantitative value for selection decisions, and are divided into performance parameters (such as rated power, swept area, and design turbulence), cost parameters (such as unit price of equipment and unit operation and maintenance cost), and risk parameters (such as failure rate and availability).

[0073] In this embodiment, the system performs multiple rounds of screening: Initial screening: Iterates through a preset turbine database (assuming 200 models are stored), applying hard constraints for rapid filtering, such as executing an SQL query to obtain 30 passing models; Fine screening: Apply soft constraints to the initial screening results, such as calculating the theoretical capacity coefficient of each model (based on power curves and wind frequency distribution), eliminating models below 25%, and retaining 15 models; Parameter extraction: Extract a core parameter dictionary for each candidate model, including static parameters (such as impeller diameter D, hub height range 80-120m), dynamic parameters (such as power curve function P=f(V)), and cost parameters (such as equipment unit price ¥4500 / kW, unit maintenance cost ¥0.08 / kWh). The parameters are standardized (units unified, outlier correction) and then stored in the candidate model list.

[0074] Understandably, this step enables intelligent pruning of a massive pool of device models, precisely focusing the computational attention from 200 models across the entire market to 15 highly compatible models, improving screening efficiency by over 90%. Through hierarchical screening logic, both the technical feasibility of the candidate set and sufficient diversity (covering different technical approaches and power levels) are ensured, preventing premature entrapment in local optima. The extracted parameters form standardized inputs, providing a plug-and-play data structure for subsequent scheme combination and optimization calculations, significantly improving engineering processing efficiency.

[0075] Step S23: Based on the parameters and the multiple candidate models, generate multiple selection schemes.

[0076] It should be noted that the selection scheme refers to a complete configuration combination that is further refined based on S22. It not only includes the selection of the model, but also determines a complete set of parameters such as single unit capacity, tower height, number of turbine positions, total installed capacity, and deployment method, forming a feasible engineering plan. For example: "Scheme X: Use 3.0MW model from Manufacturer A, hub height 100m, a total of 33 units, total capacity 99MW, arranged in 3 rows in a staggered manner."

[0077] In this embodiment, the system implements parameterized combination enumeration: Single-unit configuration generation: For each candidate model, iterate through its selectable tower height range (e.g., 80m, 90m, 100m) to generate different configuration variants, such as "Model A-80m" and "Model A-90m". Assuming that each of the 15 candidate models has 3 tower heights, this will generate 45 configuration options.

[0078] Quantity matching: Based on the total capacity requirement of the project (e.g., 100MW) and the capacity of a single unit, calculate the theoretical number of units corresponding to each configuration (e.g., 33 units are required for the 3.0MW model), allowing a fluctuation of ±1 unit to form a capacity combination (e.g., 32 units = 96MW, 34 units = 102MW).

[0079] Scheme instantiation: Combine configuration and quantity to generate specific scheme objects. Each scheme contains a complete set of parameters {model ID, single unit capacity, tower height, number of units, total capacity, total equipment cost, and estimated annual power generation}.

[0080] Deduplication and pruning: Eliminate schemes whose total capacity deviates from the project requirements by more than 5%, and eliminate schemes with significantly deteriorated technical and economic efficiency (such as those with less than 2000 hours of power generation). Finally, generate 50-100 differentiated selection schemes to enter the next round of evaluation.

[0081] Understandably, this step achieves systematic exhaustive search and intelligent reduction of the solution space, ensuring that the globally optimal solution exists within the candidate solution set. Through parameterized combination, the discrete machine selection is transformed into a continuous solution generation process, overcoming the limitation of manually designing only 3-5 solutions. Automated combination logic avoids solution omissions caused by human cognitive biases, while early pruning strategies control the computational scale within a reasonable range, improving generation efficiency while ensuring completeness, and laying a high-quality foundation for subsequent multi-objective optimization.

[0082] Step S30: Based on the power generation environment information, use a preset prediction model to predict the power generation corresponding to different selection schemes, obtain the predicted power generation, and determine the maintenance cost and service life corresponding to different selection schemes; It should be noted that the preset prediction model refers to the pre-trained power generation prediction algorithm, which is usually an integrated model (such as a hybrid architecture of XGBoost+LSTM+physical model), which can output the annual power generation prediction value based on inputs such as wind speed time series data, unit modeling power curves, and terrain wake effects.

[0083] It should be noted that the predicted power generation refers to the total power generation (unit: MWh) of a certain selected scheme over its entire life cycle (e.g., 20 years) under specific environmental conditions, and factors such as available hours, wind curtailment rate, power curve compliance, and aging degradation need to be considered.

[0084] It should be noted that maintenance costs refer to the average annual operation and maintenance expenditures during the operation of the unit, including the present value of costs for regular maintenance (lubricating oil replacement, component inspection), unplanned repairs (fault repair), and replacement of major components (gearbox, blades).

[0085] It should be noted that service life refers to the economic operating years of the unit, which is usually a reliability life prediction value that is corrected based on the design life (20-25 years) and environmental severity (such as high turbulence, high corrosion).

[0086] In this embodiment, wind speed time-series data from the environmental information is first input into a preset prediction model. Combined with the actual power curve (not the theoretical curve, considering air density correction) of each turbine model, wake loss model (based on Jensen or LARS model), availability (98%-99%), and other parameters, the annual net power generation of each turbine site is calculated and multiplied by the total number of turbines in the plan to obtain the predicted power generation. Then, the operation and maintenance cost database is called, and based on the turbine model reliability data, failure rate statistics, and spare parts price information, a quantile regression model is used to predict the cost range of the turbine model at the target site from 25% to 75%. Finally, combined with the environmental corrosion level (such as sea salt spray, land dust) and turbulence intensity, a fatigue damage accumulation model is used to correct the equivalent life of the unit and output the predicted service life.

[0087] Understandably, in this embodiment, this step achieves precise quantification of the benefits and risks throughout the entire lifecycle. Traditional methods only estimate power generation, while this step simultaneously predicts maintenance costs and lifespan, making implicit costs explicit and avoiding the situation of "low procurement costs but high operation and maintenance costs." Through data-driven cost and lifespan prediction, economic assessment is upgraded from static estimation to dynamic simulation, truly reflecting the long-term value differences of different models and providing key input for scientific decision-making.

[0088] Step S40: Based on the equipment cost, the predicted power generation, the maintenance cost, the service life, and the preset selection strategy, determine the recommended score corresponding to each selection scheme, so as to determine the target selection scheme among the multiple selection schemes.

[0089] It should be noted that the preset selection strategy refers to the decision preference rules that users can configure. It usually includes three types: economic strategy (NPV maximization), robust strategy (risk minimization), and balanced strategy (NPV and risk weighted optimality). The weights can also be customized.

[0090] It should be noted that the recommendation score refers to a quantitative decision score calculated by comprehensively considering multiple dimensions of indicators, which is used to rank candidate solutions.

[0091] It should be noted that the target selection scheme refers to the final selection decision result with the highest recommendation score among all candidate schemes, which includes complete information such as specific model, quantity, layout coordinates, and expected benefits.

[0092] In this embodiment, the system first calculates the net present value (NPV) of each scheme based on the outputs of S20 and S30: discounting the predicted power generation at the current electricity price and subtracting the present value of equipment costs, maintenance costs, and residual value; then it calculates risk assessment indicators, such as the standard deviation of power generation (reflecting wind speed uncertainty) and the width of the cost quantile interval (reflecting cost fluctuation risk); next, it automatically configures weighting coefficients according to the user's selected preset selection strategy (e.g., economic strategy: NPV weight 70%, risk weight 20%, and recovery option weight 10%); finally, it performs a TOPSIS comprehensive evaluation on multiple schemes, obtains a recommended score and ranks them, outputs the top-ranked target selection scheme, and provides sensitivity analysis (e.g., the impact of ±5% wind speed change on NPV) and a decision report.

[0093] Understandably, this step achieves a closed loop from data analysis to intelligent decision-making, transforming complex multi-objective decision problems into quantifiable and interpretable recommendation results. By pre-setting strategies to meet the personalized preferences of different users (e.g., risk-seeking investors vs. conservative state-owned enterprises), the TOPSIS method ensures the objectivity and transparency of the decision-making process. The final output solution is not only the "optimal solution" but also an interpretable, auditable, and traceable scientific decision, significantly improving the quality and persuasiveness of investment decisions.

[0094] This application proposes a selection method and related equipment for wind turbine generator sets, relating to the field of wind power generation technology. Traditional selection methods rely heavily on manual experience, making decisions by simply comparing the technical parameters of different models. This approach suffers from problems such as a single evaluation dimension, strong subjectivity, and difficulty in achieving global optimization. While related technologies consider the impact of environmental factors on power generation, they are still insufficient in comprehensive cost-benefit analysis, lacking a systematic quantitative assessment of equipment procurement costs, operation and maintenance costs, and the entire life-cycle power generation revenue. Furthermore, traditional methods struggle to handle complex constraints involving multiple variables and nonlinearity, failing to dynamically adapt to the personalized needs of different sites, resulting in poor economic efficiency of the selection scheme and a prolonged investment payback period. In this application, firstly, in response to a wind turbine generator selection instruction, the power generation environment information corresponding to the wind turbine generator is obtained. Then, based on the power generation environment information, multiple selection schemes corresponding to the wind turbine generator are determined, and the equipment cost corresponding to each selection scheme is determined. Further, based on the power generation environment information, a preset prediction model is used to predict the power generation corresponding to different selection schemes, and the predicted power generation is obtained. The maintenance cost and service life corresponding to different selection schemes are also determined. Finally, based on the equipment cost, the predicted power generation, the maintenance cost, the service life, and the preset selection strategy, a recommended score is determined for each selection scheme to determine the target selection scheme among the multiple selection schemes.

[0095] Understandably, this application considers four aspects—equipment cost, maintenance cost, service life, and predicted power generation—for wind turbine generator sets with different selection schemes based on power generation environment information. Combining this with a large model, it determines the recommended score for each selection scheme and then determines the optimal selection scheme to maximize overall benefits.

[0096] Based on the first embodiment of this application, in the second embodiment of this application, the content that is the same as or similar to that in Embodiment 1 above can be referred to the above description, and will not be repeated hereafter. Based on this, please refer to... Figure 2 The step of predicting the power generation corresponding to different selection schemes based on the power generation environment information and obtaining the predicted power generation further includes steps A10 to A20: Step A10: Based on the power generation environment information, determine the wind frequency distribution and the wind turbine power curves corresponding to different selection schemes; It should be noted that wind frequency distribution refers to the probability distribution of wind speeds in different numerical ranges. It is usually expressed as the annual cumulative hours or percentage of each wind speed segment (such as 0.5 m / s intervals), reflecting the temporal statistical characteristics of wind resources (such as whether the dominant wind speed is concentrated in the 6-8 m / s range or dispersed).

[0097] It should be noted that the wind turbine power curve refers to the wind speed-power output mapping relationship of a single wind turbine generator set, that is, the electrical power value (kW) that the unit can stably output at different wind speeds. It is the core characteristic curve for measuring the wind energy capture efficiency of the unit, and it is usually S-shaped (no output below the cut-in wind speed, constant power after the rated wind speed, and shutdown above the cut-out wind speed).

[0098] In this embodiment, the system first parses the preprocessed and feature-extracted hourly wind speed time-series data from the power generation environment information; then it performs wind speed segmentation statistics: dividing the wind speed into 0-50 m / s intervals with 0.5 m / s intervals, and calculating the cumulative number of hours each interval occurs in one year (or many years) to generate a wind frequency histogram; next, it performs power curve matching for candidate models: extracting the rated power curve data table of each candidate model under standard air density (1.225 kg / m³) from the preset unit database; finally, it performs environmental adaptability correction: based on the actual air density (lower in high-altitude areas), turbulence intensity (high turbulence reduces dynamic efficiency) and blade pollution coefficient in the power generation environment information, it uses empirical formulas to perform dynamic loss correction on the standard power curve (usually high turbulence will shift the power curve down by 2%-5%), generating a localized power curve suitable for this site.

[0099] Understandably, this step achieves a precise coupling between wind resource characteristics and turbine characteristics. Wind frequency distribution quantifies the "wind supply pattern," and localized power curves quantify the "conversion efficiency of wind turbines." By using box-based statistics instead of simple averaging, the skewed characteristics of wind speed distribution (usually positively skewed) are preserved, avoiding overestimation of power generation. By correcting the power curve through environmental parameters, the turbine performance prediction is made more consistent with actual operating conditions, significantly improving the physical accuracy of power generation potential assessment and laying an accurate dual-factor input foundation (wind conditions + turbine performance) for subsequent calculations.

[0100] Step A20: Based on the wind frequency distribution, the wind turbine power curve, and the Weibull distribution fitting, a preset prediction model is used to predict the power generation corresponding to different selection schemes, and the predicted power generation is obtained.

[0101] It should be noted that Weibull distribution fitting refers to using the Weibull probability density function (two parameters: shape parameter k and scale parameter A) to mathematically fit the measured wind frequency histogram, transforming discrete wind speed statistics into a continuous and smooth probability distribution model, which facilitates analytical integration calculations.

[0102] It should be noted that the preset prediction model refers to a pre-built power generation calculation model that integrates theoretical analysis and data-driven approaches.

[0103] It should be noted that the predicted power generation refers to the total power generation capacity of a certain selected scheme over its entire life cycle, in MWh, and needs to take into account multiple factors such as wind energy resources, unit performance, terrain influence, equipment reliability, and grid curtailment.

[0104] In this embodiment, the system performs four layers of calculations sequentially: Probabilistic modeling layer: The wind frequency distribution data is fitted with a Weibull distribution, and the parameters k and A are solved by the least squares method or the maximum likelihood estimation to obtain the wind speed probability density function f(V).

[0105] Theoretical power generation layer: The fitted power curves P(V) and f(V) are integrated to calculate the theoretical annual power generation E_theoretical=8760×∫[P(V)×f(V)]dV (8760 is the number of hours throughout the year).

[0106] Loss correction layer: The wake loss rate caused by the arrangement of the power plants is calculated by calling the wake model (such as the Jensen model) (usually 5%-15%), and the power curve conformity correction (difference between the manufacturer's theoretical curve and the measured curve, about -2%), availability correction (98%-99%), and grid curtailment rate correction (based on historical wind curtailment data) are added to obtain the net annual power generation E_net = E_theoretical × (1-wake loss) × availability × (1-wind curtailment rate).

[0107] Full-lifecycle prediction layer: The net annual power generation is multiplied by the project's lifespan (e.g., 20 years), and the aging degradation of the units (typically 0.5% per year) is taken into account. The sum is then used to obtain the predicted power generation for the entire lifecycle. All calculations are repeated for each selected project.

[0108] Understandably, this step achieves precise quantification of the entire wind resource-to-power generation chain, breaking through the traditional method's crude estimation model of simply applying "annual utilization hours." Weibull fitting analyzes discrete data, making integral calculations possible and improving theoretical rigor; the multi-layered correction factor system systematically incorporates key losses in actual engineering, such as wake, reliability, and grid curtailment, ensuring that the prediction results have an error of less than 5% compared to production and operation data. The final output of predicted power generation is the core benefit input for subsequent economic evaluation, directly determining the accuracy of NPV calculation and serving as the quantitative cornerstone of the entire selection scheme's value assessment.

[0109] Specifically, the step of predicting the power generation corresponding to different selection schemes based on the wind frequency distribution, the wind turbine power curve, and the Weibull distribution fitting, and obtaining the predicted power generation, further includes steps A21 to A23: Step A21: Based on the wind frequency distribution, the wind turbine power curve and the Weibull distribution fitting, use a preset prediction model to predict the power generation corresponding to different selection schemes, and obtain the initial predicted power generation. It should be noted that the initial predicted power generation refers to the theoretical power generation calculated under idealized assumptions (the unit is always in standard condition, with no external losses and no performance degradation). It serves as a benchmark value for subsequent corrections and reflects the theoretical matching potential between wind resources and the technical performance of the unit.

[0110] In this embodiment, the system performs purely theoretical calculations: Weibull distribution parameters are fitted to the measured wind frequency distribution to obtain the wind speed probability density function f(V); the localized power curve P(V) of the candidate turbine model is numerically integrated with f(V) (discreteized and summed at 0.5 m / s wind speed intervals) to calculate the theoretical annual generating hours; multiplied by the number of turbines and the rated power of a single unit, the gross annual power generation (Gross AEP) of the selected scheme is obtained; without considering any engineering losses or aging factors, it is directly multiplied by the service life (e.g., 20 years) to generate the initial predicted power generation over the entire life cycle. This calculation assumes 100% unit availability, no wake interference, and no performance degradation.

[0111] Understandably, this step achieves a "pure" quantification of power generation potential, stripping away all external interference factors and focusing on assessing the essential match between "wind energy resources and turbine performance." This benchmark provides a clear reference for subsequent corrections, facilitating the identification of the relative impact of various losses, making the prediction process decomposable and diagnosable, and avoiding benchmark ambiguity caused by mixed correction factors.

[0112] Step A22: Determine the electrical and environmental losses corresponding to different selection schemes during the power generation process, and predict the wind turbine performance degradation factor corresponding to different selection schemes; It should be noted that electrical loss refers to the energy loss caused by resistive heating, hysteresis loss, reactive power, etc., between the wind turbine outlet and the grid connection point. It is usually expressed as a loss rate (%).

[0113] It should be noted that environmental losses refer to the reduction in energy capture caused by site environmental conditions, including wake losses (obstruction by upstream units), blade contamination losses (dust and oil reduce aerodynamic efficiency), icing losses (icing changes airfoil), and air density deviation losses (low density at high altitudes reduces thrust), etc.

[0114] It should be noted that the wind turbine performance degradation factor refers to the cumulative effect coefficient of the output power decreasing year by year due to factors such as mechanical wear, blade surface corrosion, transmission efficiency reduction, and aging control strategy after long-term operation of the unit. Typically, the performance in the 20th year is about 85%-90% of that in the first year.

[0115] In this embodiment, the system quantizes losses into three categories: Electrical loss calculation: Calculate the total length of the collection line according to the layout diagram of the unit location, and estimate the line loss (about 1%-3%) based on the cable cross-sectional area and current; add the transformer substation loss (about 1%-2%) and the main transformer loss of the substation (about 0.5%), and combine them into the total electrical loss rate.

[0116] Environmental loss calculation: Use a wake model (such as FLORIS) to simulate the wake loss rate (5%-15%) based on the machine location coordinates and wind direction distribution; query the blade pollution loss database based on the site dust concentration and rainfall frequency (usually 2%-5%); if there is a risk of icing, calculate the icing loss rate (up to 10%-20%) using the icing time and the efficiency of the de-icing system.

[0117] Performance degradation factor prediction: Based on the historical operating data of the model, a Weibull aging model is established, the annual performance degradation curve is fitted, and the degradation coefficient in year t is output; the average degradation factor is obtained by integrating the power generation degradation over the entire life cycle.

[0118] Understandably, this step achieves the systematic identification and parameterization of loss factors, decomposing the traditionally empirically estimated "comprehensive reduction factor" into three categories of losses: quantifiable, verifiable, and traceable. Through classification modeling, it accurately captures the personalized loss characteristics of different sites (such as significant wake in mountainous areas and severe pollution along the coast), avoiding estimation biases caused by a "one-size-fits-all" approach. This refined breakdown provides clear input for subsequent accurate corrections, and the generated loss analysis report can directly guide engineering design optimization (such as adjusting turbine location to reduce wake and configuring automatic cleaning systems), significantly enhancing the engineering guidance value of the prediction results.

[0119] Step A23: Based on the electrical losses, the environmental losses, and the annual performance degradation factor of the wind turbine, the initial predicted power generation is corrected to obtain the corrected predicted power generation.

[0120] It should be noted that the revised predicted power generation refers to the net power generation after deducting electrical losses and environmental losses and incorporating the performance degradation over the entire life cycle, based on the initial theoretical power generation. It is the final predicted power generation value that reflects the actual operating status of the project and is directly used for economic evaluation.

[0121] In this embodiment, the system performs chained multiplication correction, a step that achieves a leap from theory to practice, making the power generation prediction results closer to real-world engineering operation scenarios. The prediction error is reduced from the traditional 15%-20% to less than 5%. The chained correction structure is transparent and interpretable, allowing decision-makers to clearly identify the degree to which various losses diminish the final return (e.g., "wake loss accounts for 5%, higher than expected, requiring turbine location optimization"), providing accurate feedback for scheme iteration. The corrected power generation, as the core input for NPV calculation, directly affects the reliability of the investment payback period and rate of return. It is a crucial link in ensuring the credibility, usability, and decision-making of the entire selection scheme's economic evaluation, significantly improving the scientific nature of investment decisions.

[0122] Based on the first and second embodiments of this application, in the third embodiment of this application, the content that is the same as or similar to that in embodiments one and two above can be referred to the above description, and will not be repeated hereafter. Based on this, please refer to... Figure 3 Before the step of using a preset prediction model to predict the power generation corresponding to different selection schemes, steps B10 to B40 are also included: Step B10: Obtain sample data, which corresponds to the first prediction result; It should be noted that sample data refers to the historical observation dataset used to train the prediction model. It contains input features (such as historical environmental information: wind speed, wind direction, temperature, etc.) and output labels (corresponding historical actual power generation), and is the basic material for supervised learning.

[0123] It should be noted that the first prediction result specifically refers to the actual power generation label (i.e., the historical actual value) in the sample data, which serves as the ground truth for model training and is used to measure the accuracy of the model's predictions.

[0124] Understandably, by obtaining input-output mapping pairs in real-world scenarios, a "standard answer" is provided for the model, enabling it to learn complex nonlinear relationships (such as the dynamic changes of power curves under high turbulence). This is a necessary prerequisite for building a high-precision prediction model, ensuring that the training process is based on evidence and has measurable standards.

[0125] Step B20: Process the sample data using the current prediction model to obtain a second prediction result; It should be noted that the current prediction model refers to a machine learning model (such as a deep neural network, random forest, etc.) that is in a certain iterative state during the training process. Its internal parameters (weights, biases) have not yet been optimized to the optimal level, representing the current cognitive level.

[0126] The second prediction result refers to the predicted power generation calculated by the current model based on the input features (environmental information) of the sample data. It is the output of the model under the current parameter state and is used to compare with the first prediction result (true value).

[0127] Understandably, quantifying the prediction error provides a clear direction for subsequent parameter adjustments, avoiding blind parameter tuning. Simultaneously, this process outputs a learning curve (error changing with iteration), helping to determine whether the model has converged or overfitted, making it a core feedback loop for controllable and visualized training.

[0128] Step B30: Determine whether the first prediction result is consistent with the second prediction result; It should be noted that in the context of machine learning, "consistency" does not mean that the exponential values ​​are absolutely equal, but rather that the error between the two (such as root mean square error RMSE or mean absolute percentage error MAPE) is lower than a preset threshold (such as MAPE < 3%), that is, the model's predicted value and the actual value are sufficiently close within the acceptable range for engineering.

[0129] Understandably, by setting a reasonable error threshold, the model's prediction accuracy is ensured to meet the requirements of engineering applications (rather than endlessly pursuing mathematical optimality), thus balancing model complexity and generalization ability. The automatic judgment mechanism eliminates the need for manual intervention in the training process, significantly improving model development efficiency and standardization.

[0130] Step B40: If the first prediction result is inconsistent with the second prediction result, adjust the parameters of the current prediction model, and based on the current prediction model with adjusted parameters, return to the step of using the current prediction model to process the sample data to obtain the second prediction result, until the first prediction result is consistent with the second prediction result, and obtain the preset prediction model.

[0131] It should be noted that adjusting the parameters of the current prediction model refers to automatically updating the model weights along the negative gradient direction of the error using optimization algorithms (such as gradient descent or Adam optimizer) to reduce the error of the next round of prediction. This is the core learning mechanism of machine learning.

[0132] It should be noted that the preset prediction model refers to the final model that has undergone a complete training cycle, whose error has converged to within the threshold, and whose parameters have been fixed. This is a production-grade model that can be deployed and used to predict the power generation of new sites (as used in step A20).

[0133] Understandably, parameter optimization driven by error feedback allows the model to gradually approximate the true nature of power generation from its initial stochastic state. The iterative mechanism ensures the model can capture the complex nonlinear relationship between the environment and power generation (such as the differentiated impact of varying turbulence intensities on the power curve), something traditional physical models struggle to achieve. The resulting pre-defined prediction model possesses high accuracy (error <3%), strong generalization (adapting to different sites), and automatic updating (online learning) capabilities. It is the core carrier of the "intelligence" in the entire intelligent selection method, directly determining the reliability of power generation prediction and economic assessment, and providing a trustworthy quantitative engine for optimal selection.

[0134] It should be noted that the above examples are only for understanding this application and do not constitute a limitation on the selection method of wind turbine generator sets in this application. Any simple modifications based on this technical concept are within the protection scope of this application.

[0135] This application also provides a selection device based on wind turbine generator sets; please refer to... Figure 4 The selection device based on wind turbine generator sets includes: The acquisition module 10 is used to acquire the power generation environment information corresponding to the wind turbine generator set in response to the wind turbine generator set selection instruction. The first determining module 20 is used to determine multiple selection schemes corresponding to the wind turbine generator set based on the power generation environment information, and to determine the equipment cost corresponding to each selection scheme. The prediction module 30 is used to predict the power generation corresponding to different selection schemes based on the power generation environment information using a preset prediction model, obtain the predicted power generation, and determine the maintenance cost and service life corresponding to different selection schemes. The second determining module 40 is used to determine the recommended score corresponding to each selection scheme based on the equipment cost, the predicted power generation, the maintenance cost, the service life and the preset selection strategy, so as to determine the target selection scheme among the multiple selection schemes.

[0136] In one embodiment, the acquisition module further includes: The first acquisition unit is used to acquire geographical environment data, climate and meteorological data, power grid condition data, and geological condition data corresponding to the wind turbine generator set in response to the wind turbine generator set selection instruction. The data preprocessing unit is used to preprocess the geographic environment data, the climate and meteorological data, the power grid condition data, and the geological condition data to obtain preprocessed geographic environment data, climate and meteorological data, power grid condition data, and geological condition data. The data preprocessing includes data cleaning and data completion. The feature extraction unit is used to perform feature extraction operations on the preprocessed geographic environment data, climate and meteorological data, power grid condition data, and geological condition data to obtain the power generation environment information corresponding to the wind turbine generator.

[0137] In one embodiment, the first determining module further includes: The first determining unit is used to determine the selection constraint information based on the wind power generation environment information and the preset wind turbine generator set selection rules; The filtering unit is used to filter out multiple candidate models from the preset unit database based on the selection constraint information, and determine the parameters corresponding to each candidate model; The generation unit is used to generate multiple selection schemes based on the parameters and the multiple candidate models.

[0138] In one embodiment, the prediction module further includes: The second determining unit is used to determine the wind frequency distribution based on the power generation environment information, and to determine the wind turbine power curves corresponding to different selection schemes; The first prediction unit is used to predict the power generation corresponding to different selection schemes based on the wind frequency distribution, the wind turbine power curve and the Weibull distribution fitting, and obtain the predicted power generation.

[0139] In one embodiment, the prediction module further includes: The second prediction unit is used to predict the power generation corresponding to different selection schemes based on the wind frequency distribution, the wind turbine power curve and the Weibull distribution fitting, and to obtain the initial predicted power generation. The third determining unit is used to determine the electrical and environmental losses corresponding to different selection schemes during the power generation process, and to predict the wind turbine performance degradation factor corresponding to different selection schemes; The correction unit is used to correct the initial predicted power generation based on the electrical losses, the environmental losses, and the annual performance degradation factor of the wind turbine, so as to obtain the corrected predicted power generation.

[0140] In one embodiment, the wind turbine generator selection device further includes a model training module, which further includes: The second acquisition unit is used to acquire sample data, which corresponds to the first prediction result; A data processing unit is used to process the sample data using the current prediction model to obtain a second prediction result; A judgment unit is used to determine whether the first prediction result is consistent with the second prediction result; The iterative training unit is used to adjust the parameters of the current prediction model if the first prediction result is inconsistent with the second prediction result, and based on the current prediction model with adjusted parameters, return to the step of using the current prediction model to process the sample data to obtain the second prediction result, until the first prediction result is consistent with the second prediction result, and obtain the preset prediction model.

[0141] The wind turbine generator selection device provided in this application, employing the wind turbine generator selection method described in the above embodiments, can solve the technical problem of wind turbine generator selection. Compared with related technologies, the beneficial effects of the wind turbine generator selection device provided in this application are the same as those of the wind turbine generator selection method provided in the above embodiments, and other technical features in the wind turbine generator selection device are the same as those disclosed in the methods of the above embodiments, and will not be repeated here.

[0142] This application provides a wind turbine generator 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 execute the wind turbine generator selection method in the first embodiment described above.

[0143] The following is for reference. Figure 5 This document illustrates a structural schematic diagram of a wind turbine generator-based selection device suitable for implementing embodiments of this application. The wind turbine generator-based selection device in these embodiments 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), and in-vehicle terminals (e.g., in-vehicle navigation terminals), as well as fixed terminals such as digital TVs and desktop computers. Figure 5 The selected equipment based on wind turbine generators shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of this application.

[0144] like Figure 5As shown, the wind turbine generator 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 (ROM) 1002 or a program loaded from a storage device 1003 into a random access memory (RAM) 1004. The RAM 1004 also stores various programs and data required for the operation of the wind turbine generator selection device. The processing unit 1001, ROM 1002, and RAM 1004 are interconnected via a bus 1005. An input / output (I / O) interface 1006 is also connected to the bus. Typically, the following systems can be connected to I / O interface 1006: input devices 1007 including, for example, touchscreens, touchpads, keyboards, mice, image sensors, microphones, accelerometers, gyroscopes, etc.; output devices 1008 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; storage devices 1003 including, for example, magnetic tapes, hard disks, etc.; and communication devices 1009. Communication device 1009 allows the wind turbine-based equipment to communicate wirelessly or wiredly with other devices to exchange data. Although the figure shows wind turbine-based equipment with various systems, it should be understood that it is not required to implement or possess all the systems shown. More or fewer systems can be implemented alternatively.

[0145] 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 ROM 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.

[0146] The wind turbine generator selection device provided in this application, employing the wind turbine generator selection method described in the above embodiments, can solve the technical problem of wind turbine generator selection. Compared with related technologies, the beneficial effects of the wind turbine generator selection device provided in this application are the same as those of the wind turbine generator selection method provided in the above embodiments, and other technical features of this wind turbine generator selection device are the same as those disclosed in the previous embodiment method, and will not be repeated here.

[0147] 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.

[0148] 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.

[0149] 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 generator selection method in the above embodiments.

[0150] 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.

[0151] The aforementioned computer-readable storage medium may be included in the selection equipment based on the wind turbine generator set; or it may exist independently and not be assembled into the selection equipment based on the wind turbine generator set.

[0152] The aforementioned computer-readable storage medium carries one or more programs that, when executed by the wind turbine generator selection device, cause the wind turbine generator selection device to: In response to wind turbine generator selection instructions, obtain the corresponding power generation environment information for the wind turbine generator; Based on the power generation environment information, multiple selection schemes corresponding to the wind turbine generator set are determined, and the equipment cost corresponding to each selection scheme is determined. Based on the power generation environment information, a preset prediction model is used to predict the power generation corresponding to different selection schemes, obtain the predicted power generation, and determine the maintenance cost and service life corresponding to different selection schemes. Based on the equipment cost, the predicted power generation, the maintenance cost, the service life, and the preset selection strategy, a recommended score is determined for each selection scheme to identify the target selection scheme among the multiple selection schemes.

[0153] 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).

[0154] 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.

[0155] 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.

[0156] 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 generator selection method, thereby solving the technical problem of wind turbine generator selection. Compared with related technologies, the beneficial effects of the computer-readable storage medium provided in this application are the same as those of the wind turbine generator selection method provided in the above embodiments, and will not be repeated here.

[0157] 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 generator selection method described above.

[0158] The computer program product provided in this application can solve the technical problem of selecting wind turbine generator sets. Compared with related technologies, the beneficial effects of the computer program product provided in this application are the same as those of the wind turbine generator set selection method provided in the above embodiments, and will not be repeated here.

[0159] The above description is only a part of the embodiments of this application and does not limit the scope of protection of this application. All equivalent structural transformations made under the technical concept of this application and using the content of this application specification and drawings, or direct / indirect applications in other related technical fields, are included in the scope of protection of this application.

Claims

1. A method for selecting wind turbine generator sets, characterized in that, The selection method based on wind turbine generator sets includes: In response to wind turbine generator selection instructions, obtain the corresponding power generation environment information for the wind turbine generator; Based on the power generation environment information, multiple selection schemes corresponding to the wind turbine generator set are determined, and the equipment cost corresponding to each selection scheme is determined. Based on the power generation environment information, a preset prediction model is used to predict the power generation corresponding to different selection schemes, obtain the predicted power generation, and determine the maintenance cost and service life corresponding to different selection schemes. Based on the equipment cost, the predicted power generation, the maintenance cost, the service life, and the preset selection strategy, a recommended score is determined for each selection scheme to identify the target selection scheme among the multiple selection schemes.

2. The selection method based on wind turbine generator sets as described in claim 1, characterized in that, The step of obtaining the power generation environment information corresponding to the wind turbine generator in response to the wind turbine generator selection command further includes: In response to wind turbine generator selection instructions, acquire corresponding geographical environment data, climate and meteorological data, power grid condition data, and geological condition data for wind turbine generators; The geographic environment data, the climate and meteorological data, the power grid condition data, and the geological condition data are all preprocessed to obtain preprocessed geographic environment data, climate and meteorological data, power grid condition data, and geological condition data. The data preprocessing includes data cleaning and data completion. Feature extraction operations are performed on the preprocessed geographic environment data, climate and meteorological data, power grid condition data, and geological condition data to obtain the power generation environment information corresponding to the wind turbine generator.

3. The selection method based on wind turbine generator sets as described in claim 1, characterized in that, The step of determining multiple selection schemes corresponding to the wind turbine generator set based on the power generation environment information further includes: Based on the wind power generation environment information and the preset wind turbine generator selection rules, the selection constraint information is determined; Based on the selection constraint information, multiple candidate models are selected from the preset unit database, and the parameters corresponding to each candidate model are determined. Based on the parameters and the various candidate models, multiple selection schemes are generated.

4. The selection method based on wind turbine generator sets as described in claim 1, characterized in that, The step of predicting the power generation corresponding to different selection schemes based on the power generation environment information using a preset prediction model to obtain the predicted power generation further includes: Based on the power generation environment information, the wind frequency distribution is determined, and the wind turbine power curves corresponding to different selection schemes are determined; Based on the wind frequency distribution, the wind turbine power curve, and the Weibull distribution fitting, a preset prediction model is used to predict the power generation corresponding to different selection schemes, and the predicted power generation is obtained.

5. The selection method based on wind turbine generator sets as described in claim 4, characterized in that, The step of predicting the power generation corresponding to different selection schemes based on the wind frequency distribution, the wind turbine power curve, and the Weibull distribution fitting, and obtaining the predicted power generation, further includes: Based on the wind frequency distribution, the wind turbine power curve and the Weibull distribution fitting, the power generation corresponding to different selection schemes is predicted using a preset prediction model to obtain the initial predicted power generation. Determine the electrical and environmental losses corresponding to different selection schemes during the power generation process, and predict the performance degradation factor of wind turbines corresponding to different selection schemes; Based on the electrical losses, the environmental losses, and the annual performance degradation factor of the wind turbine, the initial predicted power generation is corrected to obtain the corrected predicted power generation.

6. The selection method based on wind turbine generator sets as described in claim 1, characterized in that, Before the step of predicting the power generation corresponding to different selection schemes using a preset prediction model, the method further includes: Acquire sample data, which corresponds to a first prediction result; The sample data is processed using the current prediction model to obtain a second prediction result; Determine whether the first prediction result is consistent with the second prediction result; If the first prediction result is inconsistent with the second prediction result, the parameters of the current prediction model are adjusted. Based on the current prediction model with adjusted parameters, the process of using the current prediction model to process the sample data and obtain the second prediction result is repeated until the first prediction result is consistent with the second prediction result, thus obtaining the preset prediction model.

7. A selection device based on a wind turbine generator set, characterized in that, The wind turbine generator selection device includes: The acquisition module is used to acquire the power generation environment information corresponding to the wind turbine in response to the wind turbine selection instruction. The first determining module is used to determine multiple selection schemes corresponding to the wind turbine generator set based on the power generation environment information, and to determine the equipment cost corresponding to each selection scheme. The prediction module is used to predict the power generation corresponding to different selection schemes based on the power generation environment information using a preset prediction model, obtain the predicted power generation, and determine the maintenance cost and service life corresponding to different selection schemes. The second determining module is used to determine the recommended score corresponding to each selection scheme based on the equipment cost, the predicted power generation, the maintenance cost, the service life and the preset selection strategy, so as to determine the target selection scheme among the multiple selection schemes.

8. A selection device based on wind turbine generator sets, 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 generator selection method as described in any one of claims 1 to 6.

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 generator selection method as described in any one of claims 1 to 6.

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 generator selection method as described in any one of claims 1 to 6.

Images