Offshore wind farm modeling method and system considering ocean current influence and clustering algorithm

Abstract

The application belongs to the technical field of wind farm modeling, and provides a sea wind farm modeling method and system considering ocean current influence and clustering algorithm. First, according to marine related data, the friction velocity and sea surface roughness are determined. Then, the axial induction factor of the wind turbine in the preset wake model is dynamically adjusted according to the friction velocity, and the wake diffusion coefficient is dynamically corrected according to the sea surface roughness. The wake center wind speed is determined according to the adjusted axial induction factor and the corrected wake diffusion coefficient. Finally, the wind farm is grouped according to the wake center wind speed and the clustering algorithm, and the sea wind farm modeling is realized. Through the determined sea surface roughness, the dynamic determination of the axial induction factor and the wake diffusion coefficient corrected according to the sea environment, the problem that the existing method does not fully consider the influence of the wake of the wind turbine, the ocean current and the atmospheric boundary layer of the sea wind farm under complex sea conditions is solved.

Description

Technical Field

[0001] This invention belongs to the field of wind farm modeling technology, and in particular relates to a method and system for modeling offshore wind farms that takes into account the influence of ocean currents and clustering algorithms. Background Technology

[0002] With the large-scale grid connection of offshore wind power, its impact on grid stability is increasing day by day. When conducting grid planning and stability analysis, detailed models are too time-consuming, and equivalent models are needed to represent the external characteristics of the entire wind farm.

[0003] Traditional methods of constructing equivalent models for offshore wind farms, which model each turbine individually, result in large computational scales and long processing times. Conversely, modeling an equivalent model for a single turbine suffers from insufficient computational accuracy. Furthermore, offshore wind farms operate in a more complex environment than onshore wind farms. Sea surface wind speeds are not only affected by atmospheric circulation but also closely related to ocean current distribution and air-sea exchange processes. Traditional wake models struggle to accurately depict these physical processes of air-sea interaction. In addition, traditional K-means clustering for wind farms requires manually pre-setting the number of clusters and is sensitive to initial centers, potentially leading to local optima. In practical applications, due to real-time changes in operating conditions, pre-setting clusters limits its applicability. Summary of the Invention

[0004] To address the aforementioned problems, this invention proposes a method and system for modeling offshore wind farms that considers ocean current influences and clustering algorithms. First, based on relevant ocean data, the invention determines the friction velocity and sea surface roughness. Then, it dynamically adjusts the axial induction factor of the wind turbine in the preset wake model based on the friction velocity, and dynamically corrects the wake diffusion coefficient based on the sea surface roughness. Based on the adjusted axial induction factor and the corrected wake diffusion coefficient, the wake center wind speed is determined. Finally, based on the wake center wind speed and a clustering algorithm, the wind farm is grouped to achieve offshore wind farm modeling. By determining the sea surface roughness, and using the dynamic determination of the axial induction factor and the wake diffusion coefficient corrected for the marine environment, this invention solves the problem that existing methods do not fully consider the influence of wind turbine wakes, ocean currents, and the atmospheric boundary layer on offshore wind farms under complex sea conditions.

[0005] To achieve the above objectives, the present invention is implemented through the following technical solution: In a first aspect, the present invention provides a method for modeling offshore wind farms that takes into account the influence of ocean currents and clustering algorithms, including: Acquire ocean-related data; Based on relevant ocean data, determine the friction speed and sea surface roughness; The axial induction factor of the wind turbine in the preset wake model is dynamically adjusted according to the friction speed, and the wake diffusion coefficient is dynamically corrected according to the sea surface roughness. The wake center wind speed is determined based on the adjusted axial induction factor and the corrected wake diffusion coefficient. Wind farms are grouped based on wake center wind speed and clustering algorithms to achieve offshore wind farm modeling.

[0006] Furthermore, the ocean-related data includes air temperature, relative humidity, atmospheric pressure, and relative wind speed at a height of 10 meters.

[0007] Furthermore, the sea surface roughness for: ; in, The friction speed; Charnock coefficient; The viscosity of air in motion; This is the acceleration due to gravity.

[0008] Furthermore, the friction speed is: ; ; ; ; ; ; in, It is a constant; Air temperature; Relative humidity; Atmospheric pressure; This is a virtual temperature. and These are saturated vapor pressure and specific humidity, respectively. The relative wind speed at a height of 10 meters; For Richardson's number; Sea surface temperature; For the stability function: .

[0009] Furthermore, the axial induction factor for: ; in, The friction speed is denoted as .

[0010] Furthermore, the wake diffusion coefficient for: ; in, It is a constant; This refers to the height of the wind turbine hub. This refers to sea surface roughness.

[0011] Furthermore, when grouping wind farms based on wake center wind speed and clustering algorithms, the K-means algorithm is used. Through the density clustering characteristics of DBSCAN, the number of clusters in the K-means algorithm is automatically output.

[0012] Furthermore, a 4-dimensional feature vector is constructed for each wind turbine to form a feature matrix, and each feature dimension is standardized. Based on the 4-dimensional feature dimension, the minimum number of points MinPts is determined, and the neighborhood radius is calculated using the K-nearest neighbor distance method. According to the neighborhood radius, the density clusters are identified by the DBSCAN algorithm, and invalid clusters with fewer than the preset value are removed. The number of remaining valid clusters is the number of clusters K in the K-means algorithm.

[0013] Furthermore, for each valid cluster identified by DBSCAN, the mean of the 4-dimensional features within the cluster is calculated, and the mean point is used as the initial centroid of the K-means algorithm. With a determined K value and initial centroid, the K-means algorithm is iterated until the change in centroid is less than a preset threshold, thus obtaining the final cluster affiliation label for each wind turbine.

[0014] Secondly, the present invention also provides an offshore wind farm modeling system that takes into account ocean current effects and clustering algorithms, comprising: The data acquisition module is configured to acquire ocean-related data. The calculation module is configured to determine the friction velocity and sea surface roughness based on ocean-related data. The correction module is configured to: dynamically adjust the axial induction factor of the wind turbine in the preset wake model according to the friction speed, and dynamically correct the wake diffusion coefficient according to the sea surface roughness. The wind speed determination module is configured to determine the wake center wind speed based on the adjusted axial induction factor and the corrected wake diffusion coefficient. The electric field modeling module is configured to: group wind farms based on the wake center wind speed and clustering algorithm to achieve offshore wind farm modeling.

[0015] Thirdly, the present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the offshore wind farm modeling method that takes into account ocean current effects and clustering algorithms as described in the first aspect.

[0016] Fourthly, the present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and capable of running on the processor, wherein the processor executes the program to implement the steps of the offshore wind farm modeling method taking into account ocean current effects and clustering algorithms described in the first aspect.

[0017] Fifthly, the present invention also provides a computer program product, the computer program product comprising a computer program, which, when executed by a processor, implements the steps of the offshore wind farm modeling method taking into account ocean current influence and clustering algorithm described in the first aspect.

[0018] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. This invention first determines the friction velocity and sea surface roughness based on relevant marine data; then, it dynamically adjusts the axial induction factor of the wind turbine in the preset wake model based on the friction velocity, and dynamically corrects the wake diffusion coefficient based on the sea surface roughness; based on the adjusted axial induction factor and the corrected wake diffusion coefficient, it determines the wake center wind speed; finally, it performs wind farm grouping based on the wake center wind speed and a clustering algorithm to achieve offshore wind farm modeling; by determining the sea surface roughness, and based on the dynamic determination of the axial induction factor and the wake diffusion coefficient corrected by the marine environment, this invention solves the problem that existing methods do not fully consider the influence of wind turbine wakes, ocean currents, and atmospheric boundary layers on offshore wind farms under complex sea conditions.

[0019] 2. This invention utilizes the density clustering characteristics of DBSCAN to automatically output the number of clusters instead of manually preset K values. The cluster structure of DBSCAN guides the selection of initial centroids to avoid local optima, ultimately achieving adaptive and high-precision clustering. Attached Figure Description

[0020] The accompanying drawings, which form part of this embodiment, are used to provide a further understanding of this embodiment. The illustrative embodiments and their descriptions are used to explain this embodiment and do not constitute an improper limitation of this embodiment.

[0021] Figure 1 This is a flowchart of the method in Embodiment 1 of the present invention. Detailed Implementation

[0022] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0023] It should be noted that the following detailed descriptions are exemplary and intended to provide further explanation of this application. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains.

[0024] Example 1: In the process of global energy structure transformation towards clean and low-carbon energy, offshore wind power has become a core development direction in the renewable energy field due to its advantages such as abundant resource reserves, high power generation efficiency, and low environmental impact. With the large-scale grid connection of offshore wind power, its impact on grid stability is increasing daily. Detailed models are too time-consuming for grid planning and stability analysis; equivalent models are needed to represent the external characteristics of the entire wind farm. This facilitates the design of overall control strategies, rapid verification of control effects, evaluation of the feasibility of different wind farm access schemes, and rapid comparison of the technical and economic indicators of different schemes.

[0025] As described in the background section, constructing equivalent models of offshore wind farms involves large computational scale and long processing times when modeling each turbine individually, while equivalence to a single turbine results in insufficient computational accuracy. Furthermore, offshore wind farms operate in a more complex environment than onshore wind farms. Sea surface wind speeds are not only affected by atmospheric circulation but also closely related to ocean current distribution and air-sea exchange processes. Traditional wake models struggle to accurately depict these physical processes of air-sea interaction. Therefore, the accuracy of traditional equivalent methods is insufficient for engineering requirements when applied to offshore wind farms. Simultaneously, traditional K-means clustering for wind farms requires manually pre-setting the number of clusters and is sensitive to initial centers, potentially leading to local optima. In practical applications, due to real-time changes in operating conditions, pre-setting clusters limits its applicability.

[0026] To solve at least one of the above problems, such as Figure 1 As shown, this embodiment provides a method for modeling offshore wind farms that considers the influence of ocean currents and clustering algorithms. First, considering the impact of ocean currents on the wake effect of wind turbines, the wake diffusion coefficient is corrected based on the marine environment by calculating sea surface roughness, addressing the problem that existing methods do not fully consider the influence of wind turbine wakes, ocean currents, and the atmospheric boundary layer under complex sea conditions. Next, the density clustering characteristics of DBSCAN are used to automatically output the number of clusters instead of manually preset K values. The cluster structure of DBSCAN guides the selection of initial centroids to avoid local optima, ultimately achieving adaptive and high-precision clustering. The method includes: S1. Build an ocean current-wake coupling model based on the ocean meteorological physics model and the wind turbine wake effect calculation.

[0027] Traditional Jensen wake models, which neglect marine environmental factors and the influence of ocean currents, exhibit significant biases in offshore wind farm applications. This embodiment constructs a wake effect model suitable for marine environments by introducing dynamic adjustment of the wake diffusion coefficient.

[0028] S1.1 Calculate sea surface roughness using the COARE (Coupled Ocean-Atmosphere Response Experiment) 3.0 model. Sea surface roughness The physical expression is: ; in, Friction speed (m / s); Charnock coefficient; The air velocity viscosity affects the molecular viscous dissipation of the near-sea surface layer and plays a dominant role in roughness calculations under low wind speed conditions. ; Let be the acceleration due to gravity, and take . .

[0029] Given the characteristics of the marine environment, the roughness of the ocean surface is not static; it changes in real time with the intensity of the waves (development and breaking). This change is precisely characterized by dynamically adjusting the axial induction factor α value of the wind turbine. Values ​​are dynamically adjusted: ; Dynamic value acquisition effectively captures the phase transition process at the air-sea interface (wave breaking → spray generation), reflecting the influence of wave development on roughness.

[0030] S1.2, friction speed The solution is obtained through step-by-step iterative calculation, starting with the calculation of each auxiliary parameter.

[0031] The saturated vapor pressure is calculated using the Tetens formula (used to calculate parameters related to air humidity). and wet : ; ; in, Air temperature (°C); Relative humidity; Atmospheric pressure.

[0032] Calculate virtual temperature Adjust for air temperature to account for humidity effects, and introduce a neutral drag coefficient. : ; ; in, The relative wind speed is at a height of 10 meters.

[0033] Calculate the Richardson number (stability parameter). : ; in, Sea surface temperature (°C).

[0034] Then, the friction velocity is solved iteratively. Initial friction velocity for: ; Calculate the stability function : ; in, .

[0035] Iterative update of friction speed: ; in, is von Kármán's constant, taken as 0.4.

[0036] S1.3 The relative motion of wind speed and ocean currents determines the actual wind stress acting on the sea surface, and its vector composition must satisfy the Galilean transformation. Based on the principle of fluid relative motion, the effective wind speed under the influence of ocean currents is established: ; in, The speed of the ocean current; The absolute wind speed at a height of 10 meters; The angle between the ocean current and the wind direction; This method calculates relative wind speed by vector synthesis of wind speed and ocean current. The formula considers the vector superposition of wind and ocean current; when the ocean current and wind direction are opposite, the relative wind speed increases; when they are in the same direction, the relative wind speed decreases. This is a crucial element missing from traditional wake models. Traditional wake models use absolute wind speed only to consider terrestrial environments. However, in marine environments, relative wind speed must be used to calculate wake parameters (such as sea surface roughness, wake diffusion coefficient, velocity loss rate, etc.) because the essence of the wake effect is the velocity loss region formed after airflow is obstructed by wind turbines. Its intensity and diffusion range are determined by the interaction between the airflow and the sea surface (wind stress). The magnitude of wind stress directly depends on the relative velocity between air and seawater, not the absolute velocity of the air.

[0037] S1.4, wake diffusion coefficient Reflects wake propagation velocity and sea surface roughness (This is directly related to) the intensity of sea surface turbulence—the higher the roughness (e.g., strong winds, high sea states), the stronger the turbulence and the faster the wake expands. The larger the value, the greater the difference. The traditional Jensen wake model uses an empirical constant (0.04) considering a stable terrestrial environment, which cannot reflect the differences caused by ocean current variations in the marine environment. Therefore, a dynamic correction of the wake diffusion coefficient based on sea surface roughness is proposed. More in line with the actual conditions of offshore wind farms: ; in, von Kármán's constant, taken as 0.4; This refers to the height of the wind turbine hub. This refers to sea surface roughness.

[0038] The above model incorporates the sea surface roughness calculated by the COARE3.0 model into the wake propagation process by dynamically adjusting the wake diffusion coefficient, and constructs an ocean current-wind field coupling model suitable for the marine environment by correcting the wake diffusion coefficient based on the marine environment.

[0039] The traditional Jensen wake model is a classic model widely used in wind farm wake calculations. Its basic form is as follows: For a single wind turbine, its downstream distance wind speed at the center of the wake It can be represented as: ; in, The upstream wind speed (m / s); The axial induction factor of the wind turbine, and the thrust coefficient The relationship is ; The diameter of the fan rotor (m); This is the wake attenuation coefficient (also known as the wake diffusion coefficient). The downstream distance is (m).

[0040] In traditional onshore wind farm applications, the wake attenuation coefficient It is usually taken as an empirical constant, such as 0.04 (or 0.05). This value is obtained by fitting observational data on a neutral atmospheric stratification over land with relatively stable surface roughness. It implicitly assumes that the spread rate of the wake region is constant and uniform.

[0041] However, in the marine environment, the turbulence intensity that determines the wake propagation rate is strongly influenced by multiple dynamic factors such as sea surface roughness, ocean currents, and waves. Continuing to use fixed... Values ​​(such as 0.04) will fail to reflect the true physical process, resulting in significant deviations in the calculation of wake wind speed. This is the fundamental reason for the dynamic correction in this patent.

[0042] S2. Wind farms are grouped using an improved K-means algorithm based on DBSCAN.

[0043] This step leverages DBSCAN's density clustering capabilities to directly address two major shortcomings of traditional K-means: replacing manually preset K values ​​with DBSCAN's automatic output of cluster numbers, and using DBSCAN's cluster structure to guide the selection of initial centroids to avoid local optima, ultimately achieving adaptive, high-precision clustering.

[0044] S2.1 Data preprocessing.

[0045] Clustering algorithms are sensitive to data units (e.g., power in MW, wind speed in m / s), and outliers can disrupt the density cluster structure, requiring preprocessing. For wind farms... The 4-dimensional feature vector of each typhoon turbine is: Construct a 4×M feature matrix: ; in, For the first Real-time wind speed of the typhoon (m / s, after correction by the wake model). No. Active power (MW) of a typhoon generator. , For the first The planar coordinates of the typhoon turbine (km, relative coordinates based on the center point of the wind farm).

[0046] To prevent features with different dimensions from having a dominant influence on distance calculation, each feature dimension is standardized: ; in, Features Standard deviation; , as a feature The mean.

[0047] S2.2 and DBSCAN identify wind turbine clusters that simultaneously meet the criteria of "similar operating status" and "spatial proximity" through "density reachability" (due to spatial attenuation caused by wake effects, the status of neighboring wind turbines is more relevant). The core is to determine the minimum number of points MinPts and the neighborhood radius. .

[0048] The minimum number of points, MinPts, is determined based on four feature dimensions (wind speed, power, and relative horizontal and vertical coordinates): ; Therefore, the minimum number of points, MinPts, is 9, which means that at least 9 wind turbines must be included in the neighborhood of the core point to balance the density representativeness of the state and space.

[0049] Neighborhood radius A balance needs to be struck between "state similarity" and "spatial proximity". The K-nearest neighbor distance method is used to adapt 4D features and calculate the neighborhood radius. First, the sample... 4D Euclidean distance (including position coordinates) to all samples: ; Then, for each sample, select the one with the smallest distance. A distance, (because) To ensure the neighborhood radius is consistent with the definition of "≥MinPts samples within the neighborhood" of the core point (avoiding distorted cluster structure due to excessively large / small radii), the median of the "8 nearest neighbor distances" of all samples is taken as the neighborhood radius. .

[0050] ; S2.3 Cluster identification and determination of the K value. Initialize all wind turbine points as "unvisited" and assign a cluster number (C=0); iterate through the unvisited samples. : S2.3.1, Mark as "Visited", calculate 4D Neighborhood ; S2.3.2, if These are marked as "noise points" (isolated fans, such as edge fans); S2.3.3, if Create a new cluster ,Will And neighboring samples are added to cluster C; S2.3.4 Recursively expand cluster C: For unvisited samples within the cluster, calculate their neighborhood and merge new samples until no new samples are added; Finally obtained There are 1 initial cluster. After removing invalid clusters with less than 3 samples, the remaining number is K: K = number of valid clusters (each cluster has at least 3 wind turbines).

[0051] S2.4. Initial centroid selection based on DBSCAN clusters to avoid local optima. Initial centroids are selected from valid DBSCAN clusters, ensuring that the centroids are simultaneously located at both the "state center" and the "spatial center" (due to the need to consider both for wind turbine clustering). For each valid cluster... Calculate the cluster mean of the 4-dimensional features as the initial centroid. : ; The initial centroid calculated in this way represents both the average wind speed and power within the cluster and is located at the spatial geometric center of the cluster, which aligns with the "state-space" correlation characteristics of offshore wind turbines.

[0052] S2.5. Using the "effective cluster count K" and "cluster mean centroid" output by DBSCAN as initial conditions, the clustering results are further optimized through an iterative process of "sample allocation - centroid update". The iterative process is as follows: S2.5.1, For each preprocessed sample Calculate its relationship with the current K centroids. The 4-dimensional Euclidean distance (considering the similarity of wind speed and power, as well as the spatial proximity of locations): ; in, Let be the wind speed component of the c-th centroid at the t-th iteration, and so on.

[0053] Sample Assign to the cluster containing the centroid with the smallest distance, i.e.: ; After the allocation is completed, the cluster partitioning result of the t-th iteration is obtained: ( (This refers to the set of samples contained in the c-th cluster).

[0054] S2.5.2, For each cluster Calculate the mean of the 4-dimensional features of all samples within the cluster, and use it as the new centroid. The formula is: ; in, Let be the number of samples in the c-th cluster at the t-th iteration.

[0055] Wind speed component of the new center of mass It is the average wind speed of all fans in the cluster, with the position component. It is the spatial geometric center of the wind turbine within the cluster, ensuring that the center of mass can represent both the "state center" and the "spatial center" of the cluster (fitting the spatial correlation of the offshore wind turbine wake).

[0056] S2.5.3, Iteration Stopping Condition: When the cluster centers no longer change, or the change in the 4D coordinates of all centroids is less than a preset threshold. (usually taken) ),Right now: ; At this point, the centroid is considered to be stable, the clustering results converge, the iteration stops, and K-means outputs "cluster affiliation label for each wind turbine".

[0057] This embodiment introduces a marine meteorological model, making the model more closely resemble the real and complex operating environment of offshore wind farms. It is particularly suitable for evaluating the power output characteristics of wind farm clusters in areas with drastic ocean current changes (such as tidal and current confluence). Simultaneously, this invention proposes to introduce DBSCAN to improve K-means. DBSCAN automatically discovers natural cluster structures in the data, providing accurate K values ​​for K-means. The identified dense regions serve as initial centers, improving convergence quality and speed. This method aims to address the problems in existing technologies where clustering results do not consider the environmental differences between offshore and onshore wind farms, and where wind turbine equivalent modeling has large errors and is computationally complex. It provides a reliable solution for wind turbine clustering under complex sea conditions.

[0058] Example 2: This embodiment provides an offshore wind farm modeling system that takes into account ocean current effects and clustering algorithms, including: The data acquisition module is configured to acquire ocean-related data. The calculation module is configured to determine the friction velocity and sea surface roughness based on ocean-related data. The correction module is configured to: dynamically adjust the axial induction factor of the wind turbine in the preset wake model according to the friction speed, and dynamically correct the wake diffusion coefficient according to the sea surface roughness. The wind speed determination module is configured to determine the wake center wind speed based on the adjusted axial induction factor and the corrected wake diffusion coefficient. The electric field modeling module is configured to: group wind farms based on the wake center wind speed and clustering algorithm to achieve offshore wind farm modeling.

[0059] The working method of the system is the same as that of the offshore wind farm modeling method taking into account the influence of ocean currents and clustering algorithms in Example 1, and will not be repeated here.

[0060] Example 3: This embodiment provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the offshore wind farm modeling method that takes into account ocean current effects and clustering algorithms as described in Embodiment 1.

[0061] Example 4: This embodiment provides an electronic device, including a memory, a processor, and a computer program stored in the memory and capable of running on the processor. When the processor executes the program, it implements the steps of the offshore wind farm modeling method that takes into account ocean current effects and clustering algorithms described in Embodiment 1.

[0062] Example 5: This embodiment provides a computer program product, which includes a computer program. When the computer program is executed by a processor, it implements the steps of the offshore wind farm modeling method that takes into account ocean current influence and clustering algorithm as described in Embodiment 1.

[0063] The above description is merely a preferred embodiment of this practice and is not intended to limit the scope of this practice. Various modifications and variations can be made to this practice by those skilled in the art. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of this practice should be included within the protection scope of this practice.

Claims

1. A method for modeling offshore wind farms that takes into account ocean current effects and clustering algorithms, characterized in that, include: Acquire ocean-related data; Based on relevant ocean data, determine the friction speed and sea surface roughness; The axial induction factor of the wind turbine in the preset wake model is dynamically adjusted according to the friction speed, and the wake diffusion coefficient is dynamically corrected according to the sea surface roughness. The wake center wind speed is determined based on the adjusted axial induction factor and the corrected wake diffusion coefficient. Wind farms are grouped based on wake center wind speed and clustering algorithms to achieve offshore wind farm modeling.

2. The offshore wind farm modeling method considering ocean current effects and clustering algorithms as described in claim 1, characterized in that, The ocean-related data includes air temperature, relative humidity, atmospheric pressure, and relative wind speed at a height of 10 meters.

3. The offshore wind farm modeling method considering ocean current effects and clustering algorithms as described in claim 1, characterized in that, The sea surface roughness for: ； in, The friction speed; Charnock coefficient; The viscosity of air in motion; This is the acceleration due to gravity.

4. The offshore wind farm modeling method considering ocean current influence and clustering algorithm as described in claim 3, characterized in that, The friction speed is: ； ； ； ； ； ； in, It is a constant; Air temperature; Relative humidity; Atmospheric pressure; This is a virtual temperature. and These are saturated vapor pressure and specific humidity, respectively. The relative wind speed at a height of 10 meters; For Richardson's number; Sea surface temperature; For the stability function: .

5. The offshore wind farm modeling method considering ocean current effects and clustering algorithms as described in claim 1, characterized in that, The axial induction factor for: ； in, The friction speed is denoted as .

6. The offshore wind farm modeling method considering ocean current effects and clustering algorithms as described in claim 1, characterized in that, The wake diffusion coefficient for: ； in, It is a constant; This refers to the height of the wind turbine hub. This refers to sea surface roughness.

7. The offshore wind farm modeling method considering ocean current effects and clustering algorithms as described in claim 1, characterized in that, When grouping wind farms based on wake center wind speed and clustering algorithms, the K-means algorithm is used. Through the density clustering characteristics of DBSCAN, the number of clusters in the K-means algorithm is automatically output.

8. The offshore wind farm modeling method considering ocean current effects and clustering algorithms as described in claim 7, characterized in that, A 4-dimensional feature vector is constructed for each wind turbine to form a feature matrix, and each feature dimension is standardized. The minimum number of points MinPts is determined based on the 4-dimensional feature dimension, and the neighborhood radius is calculated using the K-nearest neighbor distance method. Based on the neighborhood radius, the density clusters are identified by the DBSCAN algorithm, and invalid clusters with a sample number less than the preset value are removed. The number of remaining valid clusters is the number of clusters K of the K-means algorithm.

9. The offshore wind farm modeling method considering ocean current effects and clustering algorithms as described in claim 8, characterized in that, For each valid cluster identified by DBSCAN, the mean of the 4-dimensional features within the cluster is calculated, and the mean point is used as the initial centroid of the K-means algorithm. Given a defined K value and initial centroid, the K-means algorithm is used for iteration until the change in centroid is less than a preset threshold, thus obtaining the final cluster affiliation label for each wind turbine.

10. A modeling system for offshore wind farms that takes into account ocean current effects and clustering algorithms, characterized in that, include: The data acquisition module is configured to acquire ocean-related data. The calculation module is configured to determine the friction velocity and sea surface roughness based on ocean-related data. The correction module is configured to: dynamically adjust the axial induction factor of the wind turbine in the preset wake model according to the friction speed, and dynamically correct the wake diffusion coefficient according to the sea surface roughness. The wind speed determination module is configured to determine the wake center wind speed based on the adjusted axial induction factor and the corrected wake diffusion coefficient. The electric field modeling module is configured to: group wind farms based on the wake center wind speed and clustering algorithm to achieve offshore wind farm modeling.

Images