Wind farm wake dynamic simulation method and system based on digital twinning

Abstract

This application relates to the field of digital twin technology, specifically to a method and system for dynamic simulation of wind farm wakes based on digital twins. Specifically, it includes: constructing a three-dimensional geometric model of the area where the wind turbines are located using wind farm monitoring data; dividing the downstream area of ​​each wind turbine in the three-dimensional geometric model into sub-regions; comparing the point cloud height differences and point cloud curvature distribution differences between each sub-region and other sub-regions; and constructing adjustment coefficients for each sub-region based on the point cloud distribution characteristics at different height ranges within the sub-region; adjusting the resolution of each grid based on the adjustment coefficients; discretizing the computational domain using the adjusted grid resolution; and performing dynamic simulation of wind farm wakes using wind farm monitoring data. This improves the simulation accuracy of the core influence area of ​​the wake, helping to enhance the accuracy of subsequent optimization and adjustment of the wind turbines; and avoids the problem of poor adaptability of traditional models to complex terrain, leading to large deviations in the simulation results of wind farm wakes.

Description

Technical Field

[0001] This application relates to the field of digital twin technology, specifically to a method and system for dynamic simulation of wind farm wake based on digital twins. Background Technology

[0002] The wake of a wind farm refers to the region downstream of the wind turbine rotor where wind speed decreases and turbulence intensifies due to the absorption of airflow energy. Its structure is typically divided into near-wake, intermediate, and far-wake regions. Wake generation during wind power generation increases overall power loss in the wind farm and increases fatigue load on downstream turbines. Currently, digital twin models are commonly used for dynamic simulation of wind farm wakes. Because the wake generated by upstream turbines in a wind farm leads to reduced wind speed and increased turbulence intensity at downstream turbines, it accelerates power loss and equipment fatigue damage. Digital twin technology can construct a deeply coupled system of "physical entity-virtual model-data-driven" to achieve accurate judgment and analysis of wake characteristics. This provides high-confidence reference data for wind farm layout optimization, yaw coordination control, and operation and maintenance decisions, serving as a key technological support for the development of smart wind power.

[0003] However, when using digital twin technology to perform dynamic simulation of wind farm wakes, the complex environment in which wind turbines are located and the interference of multiple complex environmental factors can lead to insufficient accuracy of traditional simulation methods, affecting the subsequent optimization of power generation efficiency. Among these factors, the dynamic changes in terrain undulations, differences in surface roughness, atmospheric thermal stratification, turbulence, and other meteorological conditions can cause wake evolution to exhibit strong nonlinear and unsteady characteristics. Traditional models are unable to accurately characterize the multi-physics coupling effect and have poor adaptability to complex terrain, resulting in a large deviation between the simulation results of wind farm wakes and the actual wake state, which affects the optimization of wind turbines. Summary of the Invention

[0004] To address the aforementioned technical problems, the purpose of this application is to provide a method and system for dynamic simulation of wind farm wake based on digital twins. The specific technical solution adopted is as follows: In a first aspect, embodiments of this application provide a method for dynamic simulation of wind farm wake based on digital twins, the method comprising the following steps: Acquire wind farm monitoring data and construct a three-dimensional geometric model of the area where the wind turbines are located; The local features of the wind turbine locations in the 3D geometric model are analyzed, and the mesh resolution of each wind turbine region is adjusted during the discretization of the computational domain. The process includes: (1) Divide the downstream neighborhood of each wind turbine in the three-dimensional geometric model outward from the wind turbine to obtain each sub-region; perform surface fitting on the point cloud in each sub-region to obtain the curvature of each point cloud on the fitted surface; based on the difference in point cloud height between each sub-region and other sub-regions, and compare the degree of disorder in the point cloud curvature distribution between each sub-region and other sub-regions, construct the characteristic coefficients of the geological structure influence of each sub-region. (2) Construct the influence characteristic value of local height change in each sub-region based on the distribution characteristics of point cloud quantity and location within each height range in each sub-region; (3) Construct adjustment coefficients for each sub-region based on the aforementioned characteristic coefficients and the aforementioned influence characteristic values; (4) Adjust the initial grid resolution based on the adjustment coefficient of each sub-region; The computational domain was discretized using an adjusted grid resolution, and dynamic simulation of the wind farm wake was performed in conjunction with wind farm monitoring data.

[0005] In one embodiment, the process of obtaining the feature coefficients is as follows: Calculate the variance of the curvature of all point clouds within each sub-region; calculate the normalized value of the difference between the variances of each sub-region and every other sub-region, denoted as the relative difference value; Calculate the mean height value of all point cloud values ​​in each sub-region, and denot it as the average height value; calculate the normalized value of the difference between the average height value of each sub-region and the average height value of each other sub-region, and denot it as the relative height difference coefficient; The characteristic coefficients of each sub-region are determined based on the relative difference value and the relative height difference coefficient.

[0006] In one embodiment, the feature coefficients are the positive fusion result of all the relative difference values ​​and the relative height difference coefficients of each sub-region.

[0007] In one embodiment, the process of obtaining the influencing feature value is as follows: Calculate the percentage of point cloud instances within each altitude range in each sub-region; Obtain the angle between the normal vector of each point cloud within the sub-region and the horizontal plane on the fitted surface; Calculate the standard deviation of the included angle of all point clouds within each altitude range in each sub-region; The influence characteristic values ​​of each sub-region are constructed based on the stated proportion and the stated standard deviation.

[0008] In one embodiment, the influence feature value is positively correlated with both the proportion and the standard deviation.

[0009] In one embodiment, the adjustment coefficient is a normalized value of the positive fusion value of the feature coefficients of each sub-region and the influencing feature values.

[0010] In one embodiment, the expression for the adjustment coefficient is: In the formula, Indicates the first Adjustment coefficients for each sub-region; and They represent the first The and the first Characteristic coefficients of the influence of geological structure on individual sub-regions; and They represent the first The and the first The influence characteristic value of local height changes in each sub-region; n represents the number of sub-regions downstream of the wind turbine; This is a preset parameter tuning factor.

[0011] In one embodiment, the process of adjusting the initial grid resolution based on the adjustment coefficient of each sub-region is as follows: The scaling ratio of each sub-region is determined based on the adjustment coefficient of each sub-region. The product of the scaling factor of each sub-region and the initial grid resolution is used as the adjusted grid resolution for each sub-region.

[0012] In one embodiment, the scaling ratio of each sub-region is determined based on the adjustment coefficient of each sub-region, expressed as: ,in, Indicates the first Scaling ratio of each sub-region; Indicates the preset maximum scaling ratio; Indicates the first Adjustment coefficients for each sub-region; This represents the difference between the preset maximum scaling ratio and the preset minimum scaling ratio.

[0013] Secondly, embodiments of this application also provide a dynamic simulation system for wind farm wake based on digital twins, including a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor executes the computer program to implement the steps of any of the methods described above.

[0014] The embodiments of this application have at least the following beneficial effects: This application constructs a three-dimensional geometric model of the wind turbine region using wind farm monitoring data. The downstream area of ​​each wind turbine in the three-dimensional geometric model is divided into sub-regions. The differences in point cloud height and curvature distribution between each sub-region and other sub-regions are compared to construct characteristic coefficients of the geological structure influence of each sub-region. Combined with the point cloud distribution characteristics at different height ranges within each sub-region, adjustment coefficients are constructed to reflect the degree of influence of local changes in the downstream terrain structure on the wake analysis. Based on the adjustment coefficients, the resolution of each grid is adjusted, and multi-scale fine-grained grid subdivision is performed in different downstream areas of the wind turbine, improving the simulation accuracy of the core wake influence area. This enables the subsequent digital twin model constructed based on the differentiated discrete processing results of the computational domain to achieve accurate dynamic simulation of the wind farm wake, improving the accuracy of subsequent optimization and adjustment of the wind turbine. This avoids the problem of traditional models having poor adaptability to complex terrain, leading to significant deviations between the simulation results and the actual wake state. Attached Figure Description

[0015] To more clearly illustrate the technical solutions and advantages in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 A flowchart illustrating the steps of a wind farm wake dynamic simulation method based on digital twins provided in one embodiment of this application; Figure 2 A flowchart outlining the steps for adjusting the grid resolution. Detailed Implementation

[0017] To further illustrate the technical means and effects adopted by this application to achieve the intended inventive objective, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of the wind farm wake dynamic simulation method and system based on digital twins proposed in this application. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0018] Unless otherwise defined, 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.

[0019] The following description, in conjunction with the accompanying drawings, details the specific scheme of the wind farm wake dynamic simulation method and system based on digital twin provided in this application.

[0020] Please see Figure 1 The diagram illustrates a flowchart of a method for dynamic simulation of wind farm wake based on digital twins, according to an embodiment of this application. The method includes the following steps: Step S1: Obtain wind farm monitoring data and construct a three-dimensional geometric model of the area where the wind turbine is located.

[0021] To achieve accurate dynamic simulation of wind farm wakes based on digital twin technology, this application collects monitoring data from the wind farm. This monitoring data includes meteorological environmental data, wind turbine equipment operation data, topographic geometric data, and measured wake data. Meteorological environmental data can be collected using a LiDAR array with a detection range of 1-2 km to accurately analyze the spatiotemporal dynamic distribution of wind speed, wind direction, and turbulence intensity. Wind turbine equipment operation data can be collected using a SCADA system to reflect the real-time operating status of the wind turbines. Topographic geometric data can be obtained from a digital elevation model (DEM) and wind farm CAD design drawings to construct a three-dimensional geometric model consistent with the physical scene. Measured wake data can be collected using a scanning LiDAR to obtain detailed data such as wake wind speed attenuation and turbulence intensity changes within a 2D-6D range downstream of the wind turbines.

[0022] Furthermore, since interference during data acquisition can degrade data quality and affect the accuracy of subsequent wind farm geometric modeling and simulation, this application preprocesses the acquired monitoring data to reduce the impact of interference factors on data quality. Specifically, the Kalman filter algorithm is used to smooth the data, reducing the impact of random noise such as sensor noise and electromagnetic interference on data stability; the Moving Least Squares (MLS) algorithm is used to denoise the terrain geometric data, reducing distortion in terrain modeling caused by DEM data elevation anomalies, UAV mapping point cloud noise, and CAD drawing geometric deviations; after the above processing, a monitoring dataset with a data integrity rate of ≥99.5% and an error of ≤5% is obtained, providing core data support for the accuracy of subsequent wind farm geometric modeling and the reliability of wake simulation.

[0023] Furthermore, to achieve accurate dynamic simulation of wind farm wakes, this application constructs a geometric model of the wind farm based on collected monitoring data. Specifically, firstly, preprocessed terrain geometric data is used as input, and the constrained Delaunay triangulation algorithm of ANSYS Space Claim software is used to generate a three-dimensional terrain model with an error of less than or equal to 0.5m from the actual terrain, fully restoring the elevation undulations and surface features of the area where the wind turbine is located. Simultaneously, wind turbine design parameter data is imported, including wind turbine coordinates, hub height, impeller diameter, and blade airfoil. These parameters can be determined based on the actual wind turbine model and layout settings. Then, an accurate wind turbine body model is constructed using SolidWorks parametric modeling methods. The blades are fitted with NURBS surface curves to the actual airfoil curves, while the tower is modeled as a segmented cylinder and matched with the design wall thickness to ensure that the critical dimension error of the wind turbine is less than or equal to 0.02m. Furthermore, Boolean operations are used to integrate the terrain model and the wind turbine model to form a complete three-dimensional geometric model, thereby accurately locating the spatial position of the wind turbine and the local terrain features of each area. The construction process of the above three-dimensional geometric model is a well-known technique, and the specific process will not be described in detail.

[0024] Step S2 involves analyzing the local features of the wind turbine location in the three-dimensional geometric model and adjusting the grid resolution of each wind turbine region during the discretization of the computational domain.

[0025] After the above processing, a three-dimensional geometric model of the area where the wind turbine is located is obtained. The three-dimensional geometric model reflects the spatial morphology and characteristics of the area where the wind turbine is located. In order to further analyze the dynamic characteristics of the wind farm wake in the area where the wind turbine is located, before performing dynamic simulation processing of the wind farm wake based on digital twin technology, it is necessary to divide the computational space into a finite number of grid units through computational domain discretization processing, so as to realize the quantitative simulation analysis of wake generation, diffusion, superposition and turbulence evolution process.

[0026] In the process of wake dynamic simulation analysis of wind farms, due to the complex terrain features of the areas where wind turbines are located, insufficient mesh resolution after discretization of the computational domain for complex terrain may cause uphill or turbulence effects, leading to simulation distortion and affecting the accuracy of wake path analysis. Furthermore, it may fail to capture microstructures such as tip vortices and wake shear layers, resulting in significant errors in downstream turbine power analysis. Therefore, to improve the accuracy of capturing key features in subsequent wake dynamic simulation analysis, while ensuring the real-time performance and stability of digital twin simulation, this application analyzes the local features of the wind turbine location in the constructed 3D geometric model. This achieves accurate computational domain discretization in the wake dynamic simulation of wind farms in complex terrain areas, improving the accuracy of subsequent wake dynamic simulation. The specific analysis and processing process is as follows: (1) First, for the three-dimensional geometric model of the area where the wind turbine is located, the entire field of the three-dimensional geometric model is initially divided using unstructured tetrahedral mesh, and the initial mesh size is set to 50m. Since the wake of the wind farm occurs in the area around the wind turbine, if the terrain around the wind turbine has a large degree of complexity, the error in the analysis of the wake characteristics of the wind farm will be large under the set mesh resolution. Based on the above analysis, the local regional characteristics of the location of each wind turbine are analyzed, and the mesh resolution of the area where each wind turbine is located is optimized and adjusted based on the analysis results.

[0027] Specifically, the location of each wind turbine is determined based on its coordinates in the three-dimensional geometric model. Since wind farm wakes are typically generated within a range of 2 to 10 times the rotor diameter (D) downstream, they affect subsequent downstream turbines. In this application, a judgment and analysis region with a radius of 10D is set with each wind turbine as the center. Furthermore, since the wake generates a gradient decay region from the core strong disturbance area to the outer weak influence area, the region downstream of the wind turbine in this judgment and analysis region is divided into equally spaced sub-regions based on the distance from the wind turbine center. Specifically, in this embodiment, the region is divided at intervals of D radial distance from the wind turbine center, resulting in regions within the range of 0~D, D~2D, ..., 9D~10D from the wind turbine center, which serve as sub-regions in the downstream judgment and analysis region. The purpose is to accurately match the dynamic evolution characteristics of the wake from generation, development to decay, and to configure the grid resolution according to the characteristic differences of different regions.

[0028] Furthermore, based on each sub-region as described above, according to the coverage of the sub-region in the 3D geometric model, all point cloud data corresponding to that sub-region are obtained. Using the point cloud data of each sub-region as input, surface fitting is performed using the least squares method. Based on the surface fitting results, the curvature of each point cloud at its location on the fitted surface is calculated. The variance of all curvatures is calculated, and the normalized value of the absolute difference between the variance of this sub-region and the variance of each other sub-region is calculated and denoted as the relative difference value of the geological structure change within this sub-region. The larger the relative difference value, the more significant the change in structural complexity of the current sub-region compared to other sub-regions, and the greater the impact on the wake distribution. Specifically, for the normalization of the absolute difference value, this embodiment uses the maximum-minimum normalization method to normalize all the absolute difference values ​​of all sub-regions. Many existing normalization methods exist, and implementers can also use other normalization methods to obtain the normalized value of the absolute difference value; this application does not impose specific limitations.

[0029] To further analyze the differences in geological structure downstream of the wind turbine center, the average height of the ground in all point cloud data for each sub-region is calculated, i.e., the average value of the corresponding Z-coordinate data in all point cloud data, denoted as the average height value. The absolute value of the difference between the average height values ​​of the current sub-region and every other sub-region is also calculated. The normalized result of all these absolute values ​​is denoted as the relative height difference coefficient. In this embodiment, the normalization method used is the maximum value normalization method, where the maximum value is the maximum absolute value of the difference between the average height values ​​of all sub-regions. It should be noted that if the maximum absolute value of the difference between the average height values ​​of all sub-regions is 0, then the relative height difference coefficient of each sub-region is set to 0. The larger the relative height difference coefficient, the greater the difference in terrain between the sub-region and other sub-regions, and the greater the impact on the wake diffusion path, wind speed attenuation law, and turbulence intensity distribution.

[0030] It should be noted that there are many existing normalization methods, and implementers may also use other normalization algorithms to normalize the absolute value. This application does not impose any specific restrictions.

[0031] Based on the above analysis, to further analyze the impact of the geological structure complexity of each sub-region on the downstream wake region analysis, characteristic coefficients of the geological structure influence of each sub-region are calculated by comprehensively considering the relative differences in geological structure changes and the relative height difference coefficients. The purpose is to quantify the degree of influence of the geological structure complexity of different sub-regions on the wake simulation accuracy. Specifically, the positive fusion result of the relative differences and relative height difference coefficients between each sub-region and all other sub-regions is calculated as the characteristic coefficients of the geological structure influence of each sub-region. Here, the positive fusion refers to combining two or more indicators through methods such as addition, multiplication, and averaging. Preferably, in this embodiment, the calculation formula for the characteristic coefficients is: ,in Indicates the first The characteristic coefficients of the geological structure influence of each sub-region; n represents the number of sub-regions divided downstream of the wind turbine; Indicates the first Sub-regions and the first The relative difference values ​​between the sub-regions; Indicates the first Sub-regions and the first The relative height difference coefficient between the sub-regions. The larger the calculated characteristic coefficient, the greater the difference in terrain and geological structure between the current sub-region and other sub-regions, and the greater the potential impact on the wake distribution.

[0032] (2) Since the geological structure changes in different areas have different effects on the wake feature analysis when the wake is generated downstream of the wind turbine, the local features of each sub-region are compared and analyzed. Specifically, for the point cloud data in each sub-region, the minimum and maximum values ​​of the Z-axis coordinate values ​​in all point cloud data are obtained to determine the range of terrain height change in the sub-region. Based on the range of terrain height change, the comparison and division at different heights are carried out to analyze the geological change features at different height ranges.

[0033] Specifically, since wind farm wakes originate downstream of the turbine, extending from the core area near the turbine to a distant diffusion area, the terrain in different downstream regions affects the wake's location and thus its influence on the wake's characteristics. Differences in local elevation can cause wake deflection, enhanced superposition effects, or deviations from the ideal wind speed attenuation model. Therefore, each sub-region is uniformly divided based on its elevation variation. The number of elevation ranges can be set between 10 and 20, and the implementer can set it according to actual conditions; this application does not impose a specific limitation. In this embodiment, 15 elevation ranges are used. The proportion of point clouds within each elevation range is statistically analyzed—that is, the ratio of the number of point clouds within each elevation range to the total number of point clouds within the sub-region—and denoted as the first influence coefficient of local terrain elevation coverage. A larger first influence coefficient indicates a higher proportion of terrain within that elevation range within the sub-region, resulting in more significant wake flow obstruction or guiding effects at the corresponding elevation level.

[0034] Specifically, if the terrain height variation range of the sub-region is less than a preset height threshold, then the height range is no longer divided, and the influence characteristic value of the sub-region is directly set to 0. The implementer can set the height threshold according to the actual situation; this application does not impose specific restrictions. In this embodiment, the height threshold is set to 1 meter.

[0035] Furthermore, the angle between the normal vector of each point cloud in the sub-region and the horizontal plane on the fitted surface is obtained, and the standard deviation of the angle of all point clouds in each height range is calculated and denoted as the second influence coefficient of local terrain height change. The larger the second influence coefficient is, the more severe the terrain undulation in that height range, which will aggravate the turbulent disturbance of local airflow, which may lead to the breakup of the wake structure, faster diffusion speed, and greater interference with the wake characteristic simulation.

[0036] Furthermore, based on the above processing, a comprehensive analysis is performed on the influence characteristics of the local height variation range of each sub-region downstream of the wind turbine. The influence characteristic value of the local height variation in each sub-region is calculated. The larger the influence characteristic value, the more significant the influence of the terrain structure within the corresponding sub-region on different height variation ranges. Preferably, in this embodiment, the expression for the influence characteristic value can be: ,in Indicates the first The influence characteristic value of local height change in each sub-region; m is the first... Number of height ranges divided within each sub-region; and They represent the first The division of the sub-regions The first and second influence coefficients are given for each height range. As the first in the sub-region The statistics of the percentage of point cloud quantity under each height range show that the higher the percentage and the greater the drastic change in terrain height within the corresponding height range, the greater the impact on wake characteristics. The larger the calculated impact characteristic value, the more complex the terrain structure of the local area corresponding to the sub-region, and the more significant the impact on the wind farm wake.

[0037] (3) Based on the above judgment and analysis, in the process of dynamic simulation analysis of wind farm wake, the characteristics of the downstream area of ​​the wind turbine that generates wake are analyzed and compared locally, so as to effectively judge the influence of different local areas on the accuracy of wind farm wake characteristic analysis; based on the analysis results, the calculation domain of wind farm wake dynamic simulation is discretized in a differentiated manner to improve the accuracy of subsequent dynamic simulation of wind farm wake.

[0038] Specifically, based on the influence characteristics of local height changes in each sub-region and the relative differences in geological structure in the downstream region, the adjustment coefficient for the dynamic simulation discrete processing of the wind farm wake in each sub-region is calculated. Preferably, in this embodiment, the expression for the adjustment coefficient can be: ,in, Indicates the first The adjustment coefficient for the discrete processing of the wake dynamic simulation of the wind farm in each sub-region; n represents the number of sub-regions divided downstream of the wind turbine; and They represent the first The and the first Characteristic coefficients of the influence of geological structure on individual sub-regions; and They represent the first The and the first The influence of local height changes in individual sub-regions on characteristic values; This is a preset parameter adjustment factor, used to prevent the denominator from being 0.

[0039] The larger the calculated adjustment coefficient, the more significant the influence of the terrain structure features on the wake features in the corresponding local area of ​​the sub-region. In the process of discretizing the computational domain, a denser grid resolution should be used for discretization to accurately capture the interference of terrain undulations, height changes and other features on wake generation, diffusion path and turbulence intensity, reduce wake simulation errors caused by terrain complexity and ensure the overall accuracy of wind farm wake dynamic simulation.

[0040] (4) Further, based on the above analysis and processing, the adjustment coefficients of different sub-regions in the downstream area of ​​each wind turbine are determined. On the basis of the initial discretization and grid division of the computational domain, the grid resolution is adjusted according to the adjustment coefficient of each sub-region. The specific adjustment process is as follows: First, based on the initial grid resolution of the entire field, a correspondence between the adjustment coefficient and the scaling ratio is established. In this embodiment, the grid resolution refers to the grid size, the actual value range of the adjustment coefficient is 0~1, and the grid size scaling ratio is... The range is 1.2 to 0.2, that is... The scaling ratio is determined based on the adjustment coefficient, and the calculation formula is as follows: ,in, Indicates the first The scaling ratio of the grid adjustment within each sub-region; This indicates the maximum scaling ratio, in this embodiment. ; Indicates the first Adjustment coefficients for discrete processing of wind farm wake dynamic simulation in individual sub-regions; This represents the difference between the maximum and minimum scaling ratios.

[0041] Furthermore, the product of the scaling ratio of each sub-region and the initial grid resolution is used as the adjusted grid resolution of each sub-region.

[0042] If the adjustment coefficient is larger, the influence of the terrain structure features in the corresponding local area on the wake features will be more significant. Therefore, a smaller scaling ratio should be set to increase the grid density of the sub-region, thereby improving the overall accuracy of the dynamic simulation of wind farm wake.

[0043] Secondly, based on the scaling relationship determined above, the initial mesh is refined or thinned region by region according to the adjustment coefficient of each sub-region. For sub-regions with high adjustment coefficients, the mesh size is reduced proportionally to achieve local densification, and for sub-regions with low adjustment coefficients, the mesh size is appropriately enlarged to achieve reasonable sparseness. For example, if the adjustment coefficient of a certain sub-region is 0.8, the corresponding scaling ratio is 0.4, reducing its initial 50m mesh to 20m; if the adjustment coefficient of another sub-region is 0.2, the corresponding scaling ratio is 1, keeping the initial 50m mesh size unchanged.

[0044] Step S3: Discretize the computational domain using the adjusted grid resolution, and combine it with wind farm monitoring data to perform dynamic simulation of the wind farm wake.

[0045] Distortion and discontinuity issues may arise during mesh resolution adjustment. Therefore, to address these issues, this application sets a mesh skewness threshold of 0.05. Using the mesh repair function of ANSYSICEM software, meshes with a skewness exceeding 0.05 are locally reconstructed. Specifically, transition meshes are used to smoothly connect adjacent sub-regions where mesh sizes abruptly change, and overlapping meshes are merged. This determines the mesh generation result for each sub-region, ensuring that the mesh density distribution matches the terrain complexity and impact on wake characteristics, significantly improving the accuracy of dynamic simulations of the wake's core influence area. Mesh repair in ANSYSICEM software is well-known, and the specific process will not be elaborated further.

[0046] Furthermore, to construct a digital twin simulation model for the dynamic simulation of wind farm wake, this application uses collected meteorological environmental data, wind turbine equipment operation data, terrain geometry data, and measured wake data as the basic input data for model construction. The empirical Gaussian model integrated in the FLORIS tool is used to calculate the wind farm wake distribution. A CFD model using Large Eddy Simulation (LES) and an improved actuated disk method is used to accurately capture the three-dimensional structure, vorticity distribution, and turbulence intensity changes of the wind farm wake. At the same time, the blade element momentum (BEM) wind turbine dynamics model and the collector network RLC electrical model are used to realize the construction of a full-link linkage simulation model, resulting in a digital twin model for the dynamic simulation of wind farm wake. The detailed model construction process is well known to those skilled in the art and will not be elaborated further.

[0047] Furthermore, using the constructed digital twin model, during the dynamic simulation of wind farm wake, the latest meteorological and environmental data from SCADA and LiDAR, wind turbine equipment operation data, and measured wake data are accessed in real time at 100ms intervals and stored in the twin database. The simulation engine asynchronously retrieves data from the twin database at the target time based on a set evolution time step. Subsequently, the boundary conditions of the model and the equipment operating status are dynamically updated based on the retrieved data. By adopting an asynchronous execution method, continuous simulation of the wind farm wake evolution process under different wind conditions is achieved, and the final simulation results are output. The simulation results include the effective wind speed, turbulent load, and power loss of each wind turbine, as well as vector and cloud maps of the wind farm wake flow field. The acquired simulation results are transmitted to the digital twin visualization platform, where the 3D twin engine reconstructs the terrain, equipment, and dynamic wake distribution in the wind farm. Detailed dynamic simulation and simulation result visualization are well-known, and the specific process will not be elaborated further.

[0048] The flowchart for adjusting the grid resolution is as follows: Figure 2 As shown.

[0049] Based on the same inventive concept as the above methods, this application also provides a wind farm wake dynamic simulation system based on digital twins, including a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, it implements the steps of any one of the above-described wind farm wake dynamic simulation methods based on digital twins.

[0050] In summary, this application provides a method for dynamic simulation of wind farm wakes based on digital twins. A three-dimensional geometric model of the wind turbine region is constructed using wind farm monitoring data. The downstream areas of each wind turbine in the three-dimensional geometric model are divided into sub-regions. The differences in point cloud height and curvature distribution between each sub-region and other sub-regions are compared to construct characteristic coefficients of the geological structure influence of each sub-region. Combined with the point cloud distribution characteristics at different height ranges within each sub-region, adjustment coefficients are constructed for each sub-region, reflecting the degree of influence of local changes in the downstream terrain structure on wake analysis. Based on the adjustment coefficients, the resolution of each grid is adjusted, and multi-scale fine-grained grid division is performed on different downstream areas of the wind turbine, improving the simulation accuracy of the core wake influence area. This enables the subsequent construction of a digital twin model based on the differentiated discrete processing results of the computational domain to achieve accurate dynamic simulation of wind farm wakes, improving the accuracy of subsequent optimization and adjustment of wind turbines. This avoids the problem of traditional models having poor adaptability to complex terrain, leading to significant deviations between the simulation results and the actual wake state.

[0051] It should be noted that the order of the embodiments described above is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. Furthermore, the above description focuses on specific embodiments of this application. Additionally, the processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some implementations, multitasking and parallel processing are possible or may be advantageous.

[0052] The various embodiments in this application are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

[0053] The above description is only a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the principles of this application should be included within the protection scope of this application.

Claims

1. A method for dynamic simulation of wind farm wake based on digital twins, characterized in that, The method includes the following steps: Acquire wind farm monitoring data and construct a three-dimensional geometric model of the area where the wind turbines are located; The local features of the wind turbine locations in the 3D geometric model are analyzed, and the mesh resolution of each wind turbine region is adjusted during the discretization of the computational domain. The process includes: (1) Divide the downstream neighborhood of each wind turbine in the three-dimensional geometric model outward from the wind turbine to obtain each sub-region; perform surface fitting on the point cloud in each sub-region to obtain the curvature of each point cloud on the fitted surface; based on the difference in point cloud height between each sub-region and other sub-regions, and compare the degree of disorder in the point cloud curvature distribution between each sub-region and other sub-regions, construct the characteristic coefficients of the geological structure influence of each sub-region. (2) Construct the influence characteristic value of local height change in each sub-region based on the distribution characteristics of point cloud quantity and location within each height range in each sub-region; (3) Construct adjustment coefficients for each sub-region based on the characteristic coefficients and the influence characteristic values; (4) Adjust the initial grid resolution based on the adjustment coefficient of each sub-region; The computational domain was discretized using an adjusted grid resolution, and dynamic simulation of the wind farm wake was performed in conjunction with wind farm monitoring data.

2. The wind farm wake dynamic simulation method based on digital twin as described in claim 1, characterized in that, The process of obtaining the characteristic coefficients is as follows: Calculate the variance of the curvature of all point clouds within each sub-region; calculate the normalized value of the difference between the variances of each sub-region and every other sub-region, denoted as the relative difference value; Calculate the mean height value of all point cloud values ​​in each sub-region, and denot it as the average height value; calculate the normalized value of the difference between the average height value of each sub-region and the average height value of each other sub-region, and denot it as the relative height difference coefficient; The characteristic coefficients of each sub-region are determined based on the relative difference value and the relative height difference coefficient.

3. The wind farm wake dynamic simulation method based on digital twin as described in claim 2, characterized in that, The feature coefficients are the positive fusion results of all the relative difference values ​​and the relative height difference coefficients of each sub-region.

4. The wind farm wake dynamic simulation method based on digital twin as described in claim 1, characterized in that, The process for obtaining the influencing feature values ​​is as follows: Calculate the percentage of point cloud instances within each altitude range in each sub-region; Obtain the angle between the normal vector of each point cloud within the sub-region and the horizontal plane on the fitted surface; Calculate the standard deviation of the included angle of all point clouds within each altitude range in each sub-region; The influence characteristic values ​​of each sub-region are constructed based on the stated proportion and the stated standard deviation.

5. The wind farm wake dynamic simulation method based on digital twin as described in claim 4, characterized in that, The influencing characteristic values ​​are positively correlated with both the proportion and the standard deviation.

6. The wind farm wake dynamic simulation method based on digital twin as described in claim 1, characterized in that, The adjustment coefficient is the normalized value of the positive fusion value of the feature coefficients of each sub-region and the influencing feature values.

7. The wind farm wake dynamic simulation method based on digital twin as described in claim 6, characterized in that, The expression for the adjustment coefficient is: In the formula, Indicates the first Adjustment coefficients for each sub-region; and They represent the first The and the first Characteristic coefficients of the influence of geological structure on individual sub-regions; and They represent the first The and the first The influence characteristic value of local height changes in each sub-region; n represents the number of sub-regions downstream of the wind turbine; This is a preset parameter tuning factor.

8. The wind farm wake dynamic simulation method based on digital twin as described in claim 1, characterized in that, The process of adjusting the initial grid resolution based on the adjustment coefficient of each sub-region is as follows: The scaling ratio of each sub-region is determined based on the adjustment coefficient of each sub-region. The product of the scaling factor of each sub-region and the initial grid resolution is used as the adjusted grid resolution for each sub-region.

9. The wind farm wake dynamic simulation method based on digital twin as described in claim 8, characterized in that, The scaling ratio of each sub-region is determined based on the adjustment coefficient of each sub-region, and the expression is as follows: ,in, Indicates the first Scaling ratio of each sub-region; Indicates the preset maximum scaling ratio; Indicates the first Adjustment coefficients for each sub-region; This represents the difference between the preset maximum scaling ratio and the preset minimum scaling ratio.

10. A wind farm wake dynamic simulation system based on digital twins, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the method as described in any one of claims 1-9.

Images