Wind power plant wake flow feature extraction method, device, equipment, medium and product

Abstract

The invention relates to the technical field of wind power generation, and discloses a wind power plant wake flow feature extraction method, device and equipment, a medium and a product, and the method comprises the steps: obtaining a target cabin radar data set of a plurality of first fans in a wind power plant; according to the target cabin radar data set, a global wake flow influence potential matrix is constructed, and each element in the global wake flow influence potential matrix is used for representing the potential maximum single wake flow path influence intensity of the upstream fan on the downstream target fan; based on the global wake flow influence potential matrix, carrying out dynamic identification and effective length judgment on a main flow wake flow path to obtain a first ordered sequence containing a plurality of second fans; constructing a two-dimensional wake flow influence matrix according to the first ordered sequence; and on the basis of the two-dimensional wake flow influence matrix, velocity loss and turbulence enhancement of the multiple second fans are analyzed, a condition average wake flow map reflecting the wake flow space structure of the fans is generated, and the wake flow space-time dynamic development rule under the short-time sampling and complex wind conditions can be accurately represented.

Description

Technical Field

[0001] This invention relates to the field of wind power generation technology, specifically to a method, apparatus, equipment, medium, and product for extracting wake features from wind farms. Background Technology

[0002] Existing technologies generally employ full-field averaging models or static models to characterize wake features. However, these methods have inherent limitations in natural environments where wind direction and speed change rapidly, especially under short-term sampling conditions. When the wind flow only passes over a portion of the wind farm, averaging can lead to severe data distortion, failing to capture the spatiotemporal dynamics of the wake and making it difficult to quantify its precise impact at different locations. This, in turn, limits the realization of accurate optimization and control of wind farms. Summary of the Invention

[0003] This invention provides a method, apparatus, equipment, medium, and product for extracting wake features from wind farms, in order to solve the problem that existing technologies, due to their use of whole-field averaging and static analysis, cannot accurately characterize the spatiotemporal dynamic development of wakes under short-term sampling and complex wind conditions.

[0004] In a first aspect, the present invention provides a method for extracting wake features from a wind farm, the method comprising: Obtain target nacelle radar datasets for multiple first wind turbines in a wind farm; construct a global wake influence potential matrix based on the target nacelle radar datasets, where each element characterizes the potential maximum single wake path influence intensity of the upstream wind turbine on the downstream target wind turbine; based on the global wake influence potential matrix, dynamically identify and determine the effective length of the mainstream wake path to obtain a first ordered sequence containing multiple second wind turbines; construct a two-dimensional wake influence matrix based on the first ordered sequence; analyze the velocity deficit and turbulence enhancement of multiple second wind turbines based on the two-dimensional wake influence matrix, and generate a conditionally averaged wake map reflecting the spatial structure of the wind turbine wake.

[0005] The wind farm wake feature extraction method provided by this invention quantifies the potential single wake influence intensity between wind turbines by constructing a global wake influence potential matrix, and determines the core influence relationship between upstream and downstream wind turbines. This provides a theoretical basis for subsequent wake path identification and solves the problem that traditional methods cannot accurately locate the source of wake propagation. Furthermore, based on the global wake influence potential matrix, dynamic identification and effective length determination of the mainstream wake path are performed, enabling dynamic capture of the true wake propagation path, elimination of invalid wake influence, and clarification of the effective range of wake action. This overcomes the limitation of static analysis in adapting to dynamic wind conditions and improves the accuracy of wake path characterization. Furthermore, by constructing a two-dimensional wake influence matrix, a correspondence between the wake chain and the number of wind turbine rows within the chain is established, realizing the structured quantification of wake effects. This provides an ordered framework for the accurate calculation of subsequent velocity deficit and turbulence enhancement, avoiding the disorder of wake influence analysis. Furthermore, based on the two-dimensional wake influence matrix, the velocity deficit and turbulence enhancement of multiple second wind turbines are analyzed, which can accurately extract the core features of the wake. In turn, it can generate a conditionally averaged wake map that intuitively presents the spatial structure of the wake under specific wind conditions, providing a visual basis for wind farm optimization and solving the problem that traditional methods cannot quantify the spatiotemporal distribution law of the wake.

[0006] In one optional implementation, a target nacelle radar dataset of multiple first wind turbines in a wind farm is acquired, including: Obtain the initial nacelle radar dataset of multiple first wind turbines in the wind farm; perform spatiotemporal alignment processing on the initial nacelle radar dataset to obtain the first nacelle radar dataset; perform wind flow direction coordinate system one processing on the wind turbine geographic coordinates corresponding to the first nacelle radar dataset to obtain the target nacelle radar dataset.

[0007] The wind farm wake feature extraction method provided by this invention can eliminate time synchronization deviations of data from different wind turbines through spatiotemporal alignment processing, ensuring data consistency in the time dimension and avoiding misjudgments caused by wake effects due to time misalignment. Furthermore, by converting the wind turbine's geographical coordinates into dynamic coordinates consistent with the wind flow direction, dynamic adaptation between wind turbine location and wind conditions is achieved, eliminating interference from coordinate system differences on wake path identification and distance calculation.

[0008] In one optional implementation, based on the global wake impact potential matrix, dynamic identification and effective length determination of the mainstream wake path are performed to obtain a first ordered sequence containing multiple second wind turbines, including: The global wake impact potential matrix is ​​evaluated, and multiple major impact sources for multiple third wind turbines among multiple first wind turbines are identified. A major impact source is defined as the upstream wind turbine with the greatest potential impact on the third wind turbine among all upstream wind turbines corresponding to it. Based on the multiple major impact sources, multiple first wind turbines are clustered, and dominant wake paths are constructed. According to the downwind coordinates of the first wind turbines, multiple first wind turbines in the dominant wake paths are sorted to obtain a second ordered sequence. Based on the wake superposition model, the effective length of the second ordered sequence is determined to obtain a first ordered sequence containing multiple second wind turbines.

[0009] The wind farm wake feature extraction method provided by this invention, by evaluating the global wake influence potential matrix, can accurately locate the core wake influence source of each wind turbine, eliminate secondary interference, provide accurate basis for wake path clustering, and improve the identification of wake propagation relationships. Furthermore, based on multiple main influence sources, multiple first wind turbines are clustered, and a dominant wake path is constructed. Clustering according to physical influence relationships rather than geometric positions restores the true correlation of wake propagation and solves the problem that traditional geometric clustering cannot reflect the wake superposition effect. Furthermore, according to the downwind coordinates of the first wind turbines, the wind turbines in the dominant wake path are ordered, clarifying the upstream and downstream positional relationships of the turbines within the chain, providing an ordered basis for calculating superposition velocity loss, and ensuring the logic and accuracy of wake influence superposition calculation. Furthermore, based on the wake superposition model, the effective length of the second ordered sequence is determined, which can dynamically filter the effective range of the wake, eliminate invalid influences exceeding the threshold, and thus ensure that the wake chain length matches the actual wake effect, improving the targeting of wake feature analysis.

[0010] In one optional implementation, based on the wake superposition model, the effective length of the second ordered sequence is determined to obtain a first ordered sequence containing multiple second wind turbines, including: Based on the wake superposition model, multiple superposition velocity loss values ​​of multiple first wind turbines in the second ordered sequence are calculated; based on the multiple superposition velocity loss values ​​and the preset velocity loss threshold, the effective length value of the dominant wake path is determined; based on the effective length value and the second ordered sequence, the first ordered sequence is determined.

[0011] The wind farm wake feature extraction method provided by this invention calculates multiple superimposed velocity loss values ​​of wind turbines in a second ordered sequence based on a wake superposition model. This comprehensively considers the superimposed influence of multiple upstream wind turbines, thereby accurately quantifying the actual wake loss experienced by the turbines and overcoming the limitation of a single wake model failing to reflect the wake superposition effect. Furthermore, based on multiple superimposed velocity loss values ​​and a preset velocity loss threshold, the effective length of the dominant wake path is determined, dynamically defining the effective propagation distance of the wake and avoiding the inclusion of turbines with no actual impact in the analysis, thus improving the accuracy of wake path characterization. Further, based on the effective length value and the second ordered sequence, a first ordered sequence is determined, filtering out wind turbines within the effective wake influence range while maintaining order. This provides a high-quality data sequence for subsequent matrix construction and feature extraction, reducing the interference of invalid data on the analysis results.

[0012] In one optional implementation, based on a two-dimensional wake influence matrix, the velocity deficit and turbulence enhancement of multiple second wind turbines are analyzed, and a conditionally averaged wake map reflecting the spatial structure of the wind turbine wake is generated, including: Based on the two-dimensional wake influence matrix, multiple velocity deficit values ​​and multiple turbulence enhancement values ​​of multiple second fans are calculated. Based on the multiple velocity deficit values ​​and multiple turbulence enhancement values, the wake development law of multiple second fans is analyzed, and a conditionally averaged wake spectrum reflecting the spatial structure of the fan wake is generated.

[0013] The wind farm wake feature extraction method provided by this invention calculates multiple velocity deficit values ​​and multiple turbulence enhancement values ​​for multiple second wind turbines based on a two-dimensional wake influence matrix. This quantifies the specific damage degree and turbulence changes of each turbine in the wake, providing core quantitative indicators for wake feature analysis and achieving accurate quantification of the wake effect. Furthermore, based on multiple velocity deficit values ​​and multiple turbulence enhancement values, the wake development law of multiple second wind turbines is analyzed, which can uncover the variation law of wake with downwind distance. This transforms discrete data into structured wake spatial distribution characteristics and generates corresponding conditionally averaged wake maps, providing data support for wind farm optimization and model verification.

[0014] In one optional implementation, the wake development patterns of multiple second wind turbines are analyzed based on multiple velocity deficit values ​​and multiple turbulence enhancement values, and a conditionally averaged wake spectrum reflecting the spatial structure of the wind turbine wake is generated, including: Based on multiple velocity deficit values ​​and multiple turbulence enhancement values, multiple average velocity deficit values ​​and multiple average turbulence intensity values ​​are calculated respectively. Using a Gaussian function, the distribution of multiple velocity deficit values ​​across the wind direction coordinates is analyzed and fitted, and multiple target deficit values ​​are calculated. The target deficit values ​​are used to characterize the deficit between the wake width and the centerline velocity. Based on multiple average velocity deficit values, multiple average turbulence intensity values, and multiple target deficit values, a conditionally averaged wake map reflecting the wake spatial structure is generated.

[0015] The wind farm wake feature extraction method provided by this invention calculates multiple average velocity deficit values ​​and multiple average turbulence intensity values ​​based on multiple velocity deficit values ​​and multiple turbulence enhancement values, eliminating the random fluctuations in data from a single wind turbine, extracting the statistical regularities of wake features, and improving the universality and reliability of wake development patterns. Furthermore, by using a Gaussian function to analyze and fit the distribution of multiple velocity deficit values ​​across wind direction coordinates, the wake width and centerline velocity deficit can be accurately quantified, thereby achieving the quantification of the lateral distribution characteristics of the wake and solving the problem that traditional methods cannot accurately describe the lateral expansion patterns of the wake. Furthermore, by integrating the longitudinal development and lateral distribution characteristics of the wake, a comprehensive and accurate wake spatial structure map is generated, providing a high-quality reference for wake model verification and wind farm optimization.

[0016] Secondly, the present invention provides a wind farm wake feature extraction device, the device comprising: The system comprises the following modules: an acquisition module for acquiring target nacelle radar datasets of multiple first wind turbines in a wind farm; a first construction module for constructing a global wake influence potential matrix based on the target nacelle radar dataset, where each element characterizes the potential maximum single wake path influence intensity of an upstream wind turbine on a downstream target wind turbine; a processing module for dynamically identifying and determining the effective length of the mainstream wake path based on the global wake influence potential matrix, resulting in a first ordered sequence containing multiple second wind turbines; a second construction module for constructing a two-dimensional wake influence matrix based on the first ordered sequence; and an analysis module for analyzing the velocity deficit and turbulence enhancement of multiple second wind turbines based on the two-dimensional wake influence matrix, and generating a conditionally averaged wake map reflecting the spatial structure of the wind turbine wake.

[0017] Thirdly, the present invention provides an electronic device, comprising: a memory and a processor, wherein the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the computer instructions to perform the wind farm wake feature extraction method described in the first aspect or any corresponding embodiment thereof.

[0018] Fourthly, the present invention provides a computer-readable storage medium storing computer instructions for causing a computer to execute the wind farm wake feature extraction method of the first aspect or any corresponding embodiment described above.

[0019] Fifthly, the present invention provides a computer program product, including computer instructions for causing a computer to execute the wind farm wake feature extraction method described in the first aspect or any corresponding embodiment thereof. Attached Figure Description

[0020] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0021] Figure 1 This is a schematic diagram of an application scenario according to an embodiment of the present invention; Figure 2 This is a flowchart illustrating the wind farm wake feature extraction method according to an embodiment of the present invention; Figure 3 This is a flowchart illustrating a wind farm wake feature extraction method based on dynamic wake chain meshing according to an embodiment of the present invention. Figure 4 This is a schematic diagram illustrating the principle of converting geographical coordinates to wind direction coordinates according to an embodiment of the present invention; Figure 5 This is a flowchart of the dynamic wake chain partitioning and clustering algorithm according to an embodiment of the present invention; Figure 6 This is a schematic diagram of the wake influence matrix (velocity deficit / turbulence enhancement) structure according to an embodiment of the present invention; Figure 7 This is a structural block diagram of a wind farm wake feature extraction device according to an embodiment of the present invention; Figure 8 This is a schematic diagram of the hardware structure of an electronic device according to an embodiment of the present invention. Detailed Implementation

[0022] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0023] It is understood that before using the technical solutions disclosed in the various embodiments of the present invention, users should be informed of the types, scope of use, and usage scenarios of the personal information involved in the present invention and their authorization should be obtained in accordance with relevant laws and regulations through appropriate means.

[0024] The terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0025] As an optional application scenario of this invention, the specific application environment architecture or specific hardware architecture on which the wind farm wake feature extraction method depends is described here. For example... Figure 1 As shown, the architecture system may include at least one terminal device and at least one server. Figure 1 The system is illustrated in the example, which includes a computer 101, a mobile terminal 102, and a server 103, and the terminal devices such as the computer 101 and the mobile terminal 102 are connected to the server 103 through a network 110.

[0026] Specifically, the terminal device can be a smartphone, tablet, laptop, PDA, desktop computer, game console, smart TV, smart wearable device, in-vehicle terminal, VR (Virtual Reality) device, AR (Augmented Reality) device, etc. Server 103 can be a standalone physical server, a server cluster, a distributed system, or a cloud server providing cloud services. Network 110 can be a wired or wireless network, examples of which include, but are not limited to, the Internet, corporate intranet, local area network, wide area network, mobile communication network, and combinations thereof.

[0027] According to an embodiment of the present invention, a method for extracting wake features of a wind farm is provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0028] This embodiment provides a method for extracting wake features from wind farms, which can be used on the aforementioned mobile terminals, such as mobile phones and tablets. Figure 2This is a flowchart of a wind farm wake feature extraction method according to an embodiment of the present invention, such as... Figure 2 As shown, the process includes the following steps: Step S201: Obtain the target nacelle radar dataset of multiple first wind turbines in the wind farm.

[0029] In one optional embodiment, the multiple first wind turbines are all wind turbines in the wind farm that participate in wake feature extraction and analysis; the target nacelle radar dataset represents the nacelle radar data of all wind turbines in the wind farm after standardized preprocessing, which may include wind speed, wind direction and turbulence intensity, and is used to reflect the wind environment and operating status of the wind turbines.

[0030] In one optional embodiment, the core of wake feature extraction is based on the mutual influence of wind environment between wind turbines, and the accuracy and consistency of wind environment data directly determine the reliability of wake impact quantification. Furthermore, the original nacelle radar data has problems such as time synchronization deviation (data acquisition time of different wind turbines is not synchronized) and inconsistent coordinate systems (the geographical coordinates of wind turbines are unrelated to the wind flow direction), which cannot be directly used for wake path identification and impact quantification.

[0031] Therefore, in this embodiment, the original nacelle radar data can be transformed into a target nacelle radar dataset through preprocessing, thereby eliminating interference factors in the time and space dimensions of the data and enabling the data to truly reflect the relative positional relationship of the wind turbine and the wind environment state under transient wind conditions.

[0032] Step S202: Construct a global wake impact potential matrix based on the target cabin radar dataset.

[0033] In one alternative embodiment, the global wake impact potential matrix is ​​a theoretical prediction matrix constructed based on the target nacelle radar dataset, where each element is used to characterize the potential maximum single wake path impact intensity of the upstream wind turbine on the downstream target wind turbine.

[0034] In one alternative embodiment, the core of the wake effect is the airflow disturbance (velocity loss) of the upstream wind turbine on the downstream wind turbine, and the physical wake effect is only possible when the downstream wind turbine is within the theoretical wake boundary of the upstream wind turbine.

[0035] Therefore, in this embodiment, based on information such as wind speed, wind direction, and wind turbine coordinates in the target nacelle radar data, the theoretical speed loss of the upstream wind turbine to the downstream wind turbine is calculated using a single wake model. The physically feasible impact path is then selected by combining the wake envelope judgment function. Finally, the theoretical speed loss × path feasibility is quantified into matrix elements to form a global wake impact potential matrix, thereby systematically characterizing the potential wake impact relationship between wind turbines.

[0036] In one optional embodiment, by constructing a global wake impact potential matrix, the complex multi-wind turbine wake impact relationship within the wind farm is transformed into structured matrix data. Each matrix element precisely corresponds to the potential maximum single wake impact intensity of a pair of upstream and downstream wind turbines, which not only eliminates physically impossible wake paths but also quantifies the impact potential of feasible paths.

[0037] In an alternative embodiment, the global wake influence potential matrix Matrix elements The following relation (1) is shown: (1) In the formula: This indicates the calculation using a single wake model, in the case of no wake superposition, for the wind turbine. In the fan The theoretical speed loss caused by the location; Let represent a wake envelope judgment function. The function value is 1 if and only if the target wind turbine t is located within the theoretical wake boundary of the upstream wind turbine m; otherwise, it is 0. This function ensures that only physically possible wake influence paths are considered, as shown in the following relation (2): (2) In the formula: The thrust coefficient represents the wind turbine at the upwind direction, reflecting the turbine's ability to obstruct and disturb airflow. It indicates the intensity of environmental turbulence, used to characterize the degree of turbulence in natural wind flow, and affects the spread range of the wake; This indicates the rotor diameter of the wind turbine located upwind. Indicates downstream target wind turbine Compared to upstream wind turbines The cross-wind distance, i.e., the positional deviation perpendicular to the prevailing wind direction in the wind flow coordinate system; This represents a constant coefficient used for range correction of the wake boundary; Indicates downstream target wind turbine Compared to upstream wind turbines The downwind distance is the distance along the prevailing wind direction in the wind direction coordinate system.

[0038] Among them, for each pair of "upstream wind turbines" - Downstream target wind turbine "Using a single wake model to calculate the wind turbine in the case of no wake superposition." In the wind turbine Theoretical speed loss caused by location Furthermore, the function is determined by the wake envelope. Select physically feasible wake paths only if the target wind turbine Located upstream wind turbine When the theoretical wake boundary is within the boundary, the function value is 1 (path feasible), otherwise it is 0 (path infeasible).

[0039] Furthermore, each pair of "upstream wind turbines" - Downstream target wind turbine The result of multiplying the theoretical velocity loss by the wake envelope judgment is used as a matrix element. And construct a global wake impact potential matrix. Furthermore, the matrix row index corresponds to the upstream wind turbine, and the column index corresponds to the downstream target wind turbine.

[0040] Step S203: Based on the global wake influence potential matrix, perform dynamic identification and effective length determination of the mainstream wake path to obtain a first ordered sequence containing multiple second wind turbines.

[0041] In one optional embodiment, the effective length represents the number of consecutive rows of wind turbines in the dominant wake path where the wake effect is still effective, and is a key indicator for measuring the actual propagation range of the wake.

[0042] Specifically, as the wake propagates from the upstream fan to the downstream, the velocity loss gradually decreases due to energy dissipation. When the cumulative velocity loss borne by a certain row of fans is lower than a preset threshold, the wake effect can be ignored. At this time, the effective length of the wake is defined from the root fan of the path to the fan in front of that row.

[0043] In one optional embodiment, multiple second wind turbines refer to a collection of wind turbines that are within the effective influence range of the wake after dynamic identification and effective length determination of the mainstream wake path.

[0044] In an alternative embodiment, the first ordered sequence represents, within the effective length range of the dominant wake path, the second wind turbine, arranged according to its downwind coordinates. An ordered set of wind turbines formed by sorting them from smallest to largest.

[0045] In one optional embodiment, the wake propagation has dynamic attenuation characteristics, that is, when the wake generated by the upstream wind turbine propagates downstream, the velocity loss will gradually weaken due to the influence of atmospheric turbulence, terrain and other factors, and will only have an effective impact on the downstream wind turbine within a certain range.

[0046] Furthermore, in this embodiment, based on the above characteristics and combined with the physical influence relationship between wind turbines predicted by the global wake influence potential matrix, a first ordered sequence that includes both effective wake influence and conforms to the spatial logic of wake propagation is formed through dynamic identification and effective length determination of the mainstream wake path.

[0047] Step S204: Construct a two-dimensional wake influence matrix based on the first ordered sequence.

[0048] In one alternative embodiment, the two-dimensional wake effect matrix represents a structured matrix constructed based on a first ordered sequence, used to quantify the wake effect of the wind turbine within the wake chain.

[0049] Furthermore, the row indices in the matrix represent the different dominant wake chains identified within the wind farm, and each chain corresponds to a group of wind turbines with a core wake influence relationship; the column indices represent the row number of the wind turbines within the same wake chain, sorted from closest to furthest downwind distance, i.e., the downwind distance. Furthermore, the number of columns is consistent with the number of effective wind turbine rows within the chain.

[0050] Furthermore, the matrix elements store the core quantitative indicators of the wake corresponding to "a certain wake chain - a certain row of fans", which directly reflect the degree of wake impact actually experienced by the fans at that location. Among them, the core quantitative indicators of the wake can include velocity deficit and turbulence enhancement.

[0051] In one alternative embodiment, the analysis of the wake effect requires simultaneously identifying the source of the influence (which wake chain) and the location of the influence (which row of fans within the chain) in order to accurately quantify the spatial distribution and dynamic attenuation pattern of the wake.

[0052] Furthermore, in this embodiment, a two-dimensional wake influence matrix is ​​constructed based on the following characteristics: (1) Independence of wake chains: Different dominant wake chains are caused by different root-source wind turbines, and the wake propagation paths are independent. They need to be classified and analyzed according to the chain to avoid cross-chain interference. (2) Orderliness of wind turbines within the chain: The first ordered sequence has been sorted from smallest to largest by the downwind coordinate, corresponding to the natural path of the wake propagation from the root wind turbine to the downstream. The column index (row number) can be directly associated with the wake propagation distance. (3) The structural requirements of quantitative indicators: wake indicators such as velocity loss and turbulence enhancement need to be bound to the wake chain-row number in order to subsequently statistically analyze the average pattern of different row numbers (such as the decay trend of wake with row number). The matrix structure just meets the requirements of dual-dimensional positioning and indicator storage.

[0053] Step S205: Based on the two-dimensional wake influence matrix, analyze the velocity deficit and turbulence enhancement of multiple second wind turbines, and generate a conditionally averaged wake map that reflects the spatial structure of the wind turbine wake.

[0054] In one optional embodiment, velocity deficit is the core quantitative indicator for measuring the wake effect. It represents the difference between the measured wind speed of a certain fan in the wake chain and the measured wind speed of the head fan (root fan) in the wake chain. It is presented in a dimensionless form, that is, the velocity deficit value is a relative ratio, eliminating the interference of absolute wind speed magnitude.

[0055] Furthermore, the velocity deficit reflects the degree to which the wake generated by the upstream wind turbine weakens the wind speed of the downstream wind turbine. That is, the larger the velocity deficit value, the more significant the impact of the wake on the downstream wind turbine, and the more obvious the reduction in actual power generation efficiency. In this embodiment, the velocity deficit ensures that the wake impact intensity of different wake chains and wind turbines with different numbers of rows can be compared horizontally.

[0056] In an optional embodiment, turbulence enhancement is an indicator characterizing the impact of the wake on the stability of the airflow around the fan, representing the difference between the measured turbulence intensity of a certain row of fans in the wake chain and the measured turbulence intensity of the head fan in the wake chain.

[0057] Furthermore, turbulence intensity reflects the severity of airflow velocity fluctuations. Wakes cause downstream airflow turbulence, therefore, turbulence enhancement values ​​are positive, and the larger the value, the more severe the airflow fluctuations caused by the wake. This not only reduces the stability of wind turbine power generation but may also increase the risk of fatigue damage to the wind turbine structure. In this embodiment, turbulence enhancement can compensate for the impact of the wake, which is not covered by velocity deficit, on the safety of the wind turbine structure, thereby forming a complete wake effect quantification system.

[0058] In one optional embodiment, the conditionally averaged wake map represents a visual map generated by aggregating statistical results based on a two-dimensional wake influence matrix according to specific inflow conditions. These specific inflow conditions can be wind speed ranges, atmospheric stability categories, etc. Its core features are conditionalization and spatialization: on the one hand, it aggregates only wake data under the same inflow conditions, ensuring the map reflects the wake patterns under specific wind conditions; on the other hand, it uses downwind distance as the vertical axis and cross-wind distance as the horizontal axis, superimposing core features such as velocity deficit distribution, turbulence enhancement distribution, and wake width to intuitively present the wake's propagation pattern, attenuation trend, and lateral expansion range in space.

[0059] In one optional embodiment, the spatial structure and dynamic patterns of wakes are directly influenced by inflow conditions (wind speed, atmospheric stability, etc.), and the data from a single sampling period or a single wake chain exhibit randomness, failing to directly reflect universal patterns. In this embodiment, by transforming the discrete data in the two-dimensional wake influence matrix into an intuitive and accurate spatial map, the application of wake patterns from data quantification to visualization is realized.

[0060] In one optional embodiment, by analyzing the velocity deficit and turbulence enhancement data in the two-dimensional wake influence matrix, not only can the degree of wake influence of a single wind turbine be obtained, but also the attenuation law of the wake with downwind distance can be extracted through statistical aggregation (such as calculating the average velocity deficit of each row). Furthermore, the expansion law of the wake with cross-wind distance can be extracted through Gaussian fitting, which solves the problem that traditional methods can only analyze a single wind turbine in isolation and cannot capture the global law of the wake.

[0061] Furthermore, generating conditionally averaged wake maps can provide customized references for different engineering scenarios.

[0062] For example, maps of low wind speed and stable atmospheric conditions can be used for micro-site selection of wind farms to avoid wake coverage areas under these conditions; furthermore, maps of high wind speed and strong turbulence conditions can be used for optimization of wind turbine collaborative control strategies to reduce wake superposition under these conditions.

[0063] Meanwhile, this map and the high-precision data behind it (such as wake width expansion rate) can also serve as a gold standard to calibrate wake numerical models such as FLORIS and CFD, promote the iteration of wake simulation technology, and thus achieve two-way verification between theory and practice.

[0064] The wind farm wake feature extraction method provided in this embodiment quantifies the potential single wake influence intensity between wind turbines by constructing a global wake influence potential matrix, and determines the core influence relationship between upstream and downstream wind turbines. This provides a theoretical basis for subsequent wake path identification and solves the problem that traditional methods cannot accurately locate the source of wake propagation. Furthermore, based on the global wake influence potential matrix, dynamic identification and effective length determination of the mainstream wake path are performed. This dynamically captures the true path of wake propagation, eliminates invalid wake influences, and clarifies the effective range of wake action, overcoming the shortcomings of static analysis in adapting to dynamic wind conditions and improving the accuracy of wake path characterization. Furthermore, by constructing a two-dimensional wake influence matrix, a correspondence between the wake chain and the number of wind turbine rows within the chain is established, realizing the structured quantification of wake effects. This provides an ordered framework for the accurate calculation of subsequent velocity deficit and turbulence enhancement, avoiding the disorder of wake influence analysis. Furthermore, based on the two-dimensional wake influence matrix, the velocity deficit and turbulence enhancement of multiple second wind turbines are analyzed, which can accurately extract the core features of the wake. In turn, it can generate a conditionally averaged wake map that intuitively presents the spatial structure of the wake under specific wind conditions, providing a visual basis for wind farm optimization and solving the problem that traditional methods cannot quantify the spatiotemporal distribution law of the wake.

[0065] In some optional implementations, step S201 above includes: Step S2011: Obtain the initial nacelle radar dataset of multiple first wind turbines in the wind farm.

[0066] In one alternative embodiment, the initial nacelle radar dataset consists of unprocessed nacelle radar data for all wind turbines in the wind farm, which may include wind speed, wind direction, and turbulence intensity.

[0067] In one optional embodiment, the core of wake feature extraction is based on the interaction law between wind turbines and the wind environment, and the wind environment and wind turbine operating status data are the basis for quantifying this law. Therefore, in this embodiment, the nacelle radar equipment can directly collect real-time wind condition parameters at the location of the wind turbine, and can obtain raw data covering all wind turbines involved in the analysis of the wind farm, thereby completely capturing the wake influence relationship between all wind turbines in the entire field.

[0068] Step S2012: Perform spatiotemporal alignment processing on the initial cabin radar dataset to obtain the first cabin radar dataset.

[0069] In one optional embodiment, the spatiotemporal alignment process can map the raw data of all wind turbines to the same time scale by unifying the time base, thereby ensuring that all wind turbine data within the same sampling interval correspond to the same transient wind conditions and eliminating the interference of time deviation on the wake analysis results.

[0070] In one optional embodiment, the nacelle radar equipment of different wind turbines may have issues with asynchronous data acquisition times and inconsistent data recording timestamps. If this data is directly used for wake analysis, it will lead to incorrect matching between the wind condition data of the upstream wind turbine at a certain moment and the wind condition data of the downstream wind turbine at different moments, failing to truly reflect the wake influence relationship between wind turbines at the same time dimension.

[0071] Furthermore, in this embodiment, by calibrating the timestamps and unifying the sampling intervals, the radar data of all the first wind turbines are synchronized and aligned in time.

[0072] For example, the data for each wind turbine is calibrated according to a unified time base, matching data from different timestamps to the corresponding standard time segments. Simultaneously, for time segments with missing data, interpolation with adjacent valid data can be used to supplement them; data outside the time range or with abnormal timestamps are discarded, thus forming a calibrated, time-synchronized first nacelle radar dataset.

[0073] Step S2013: Perform wind direction coordinate system processing on the wind turbine geographic coordinates corresponding to the first nacelle radar dataset to obtain the target nacelle radar dataset.

[0074] In one optional embodiment, the principle of wind direction coordinate system is as follows: based on the dominant wind direction of each sampling interval, the absolute latitude and longitude geographical coordinates are converted into relative downwind distance-cross-wind distance coordinates, so that the converted coordinates directly reflect the relative position of the wind turbine in the wake propagation path (upstream / downstream, cross-wind offset), thereby mapping all wind turbines into a dynamic grid consistent with the wind direction.

[0075] In one alternative embodiment, the geographical coordinates (latitude and longitude) of the wind turbine can only reflect its absolute spatial location and cannot directly reflect the relative positional relationship between the wind turbine and the airflow direction, while the wake propagation is carried out along the airflow direction (diffusion from the upstream wind turbine to the downstream wind turbine).

[0076] Furthermore, in this embodiment, the absolute geographical location of the wind turbine is converted into relative position coordinates related to the airflow direction, thereby realizing dynamic adaptation of spatial coordinates and transient wind conditions.

[0077] For example, the calculation formula is shown in the following relation (3): (3) In the formula: This indicates the downwind distance of the wind turbine in the wind direction coordinate system; This indicates the cross-wind direction coordinates of the wind turbine in the wind direction coordinate system; Indicates the prevailing wind direction at the current sampling interval; This represents a constant additional rotation angle. It is usually taken as 90° or -90° because the natural wind direction (the direction from which the wind comes) differs from the mathematical positive X-axis direction (the direction from which the wind goes) by 90 degrees. , This represents the latitude and longitude coordinates of a wind turbine in a wind farm within a geographic coordinate system. , This represents the geographical coordinates of any selected reference wind turbine within the wind farm.

[0078] In some optional implementations, step S203 above includes: Step S2031: Evaluate the global wake impact potential matrix and identify multiple major impact sources for multiple third wind turbines among multiple first wind turbines.

[0079] In an alternative embodiment, the primary source of influence refers to the wind turbine that has the greatest potential single wake influence intensity among all its upstream wind turbines for each target wind turbine (the third wind turbine) in the wind farm.

[0080] In one optional embodiment, the global wake impact potential matrix quantifies the potential maximum single wake impact intensity of each upstream wind turbine on each downstream target wind turbine, and the matrix elements have been filtered to identify physically feasible impact paths using a wake envelope judgment function. Therefore, for each target wind turbine (the third wind turbine), focusing on its corresponding matrix column data—that is, the potential impact of all upstream wind turbines on it—and locking in the maximum impact source through extreme value filtering ensures that the determination of the main impact source is both physically feasible and reliable.

[0081] For example, for each target wind turbine in the wind farm (Third wind turbine), by assessing the global wake impact potential matrix The Assign a primary source of influence to each column, as shown in the following relation (4): (4) In the formula: Indicates the first The main source of impact from the third wind turbine; Which variable should be returned to make the function reach its maximum value? This represents a constraint condition used to limit consideration to only the target wind turbine located in the wind direction coordinate system. The upstream wind turbines, i.e., limited The third fan is located in the airflow direction coordinate system. Upstream wind turbines; This represents the optimization objective function, a quantitative indicator that, if all other wind turbines are ignored and only the upstream wind turbines are considered... For the target wind turbine The theoretical maximum impact (speed loss) that can be caused.

[0082] Step S2032: Based on multiple major influencing sources, cluster multiple first wind turbines and construct the dominant wake path.

[0083] In one optional embodiment, the dominant wake path represents the wake propagation path formed by grouping all downstream target wind turbines that have the same wind turbine as the main source of influence into the same set as the main source wind turbine (root source wind turbine). It can truly reflect the core path of the wake propagation from the root source wind turbine to the downstream and is the basis for quantifying the wake superposition effect and determining the effective length.

[0084] In one optional embodiment, the wake propagation has a homogeneity characteristic, that is, the wake generated by the same root source wind turbine will propagate downstream in a specific direction and affect multiple wind turbines. The main source of influence for these affected wind turbines is the same root source wind turbine.

[0085] Furthermore, in this embodiment, the aforementioned homology feature is utilized to cluster wind turbines with the same primary source of influence and root source wind turbines to form dominant wake paths, so that the paths can centrally reflect the propagation range of the same root source wake.

[0086] For example, all of them will be from the same upstream wind turbine Wind turbines identified as the primary source of impact were grouped into the same set, along with... Together they form a dominant wake path, as shown in the following equation (5): (5) In the formula: Indicates the first A dominant wake path.

[0087] Step S2033: Sort the multiple first wind turbines in the dominant wake path according to the coordinates of the first wind turbine in the downwind direction to obtain the second ordered sequence.

[0088] In one optional embodiment, the wind turbines within the dominant wake path are spatially distributed along the wake propagation direction, with downwind coordinates... It directly reflects the relative distance between the blower and the root blower, that is The larger the size, the farther away it is from the root wind turbine.

[0089] In an alternative embodiment, the set All wind turbines, according to their downwind coordinates Sort the data from smallest to largest to form an ordered sequence, i.e., a second ordered sequence. .in, Indicates the root wind turbine of this path. .

[0090] Step S2034: Based on the wake superposition model, determine the effective length of the second ordered sequence to obtain a first ordered sequence containing multiple second wind turbines.

[0091] In one optional embodiment, the wake superposition model represents a mathematical model used to calculate the combined wake impact (superimposed velocity loss) actually experienced by downstream wind turbines, such as an energy loss superposition model. Its core logic is that downstream wind turbines may be simultaneously affected by the wake effects of multiple upstream wind turbines. Therefore, it is necessary to integrate the wake effects of all contributing sources, rather than considering only the impact of a single wind turbine, thereby accurately quantifying the actual wake loss experienced by the wind turbine. This differs from the limitation of a single wake model, which only calculates the impact of a single upstream wind turbine.

[0092] Specifically, step S2034 above includes: Step a1: Based on the wake superposition model, calculate the superposition velocity loss values ​​of multiple first wind turbines in the second ordered sequence.

[0093] In one optional embodiment, the second ordered sequence has already defined the upstream and downstream order of the wind turbines within the wake path, and the downstream wind turbine may be simultaneously affected by the wake superposition effect of all upstream wind turbines within the path. Therefore, this embodiment, based on the wake superposition model, comprehensively considers the wake effects of all upstream contributing sources for each wind turbine in the sequence to calculate its actual superposition velocity loss, rather than only considering the influence of a single source wind turbine, thus ensuring the authenticity of the wake loss quantification.

[0094] In an alternative embodiment, the superposition velocity loss value is calculated using the following relationship (6): (6) in, Indicates the second ordered sequence, the first... individual fan The actual cumulative rate of loss suffered; This represents the wake superposition model.

[0095] Step a2: Determine the effective length of the dominant wake path based on multiple superimposed velocity loss values ​​and a preset velocity loss threshold.

[0096] In one optional embodiment, as the wake propagates downstream from the root wind turbine, it gradually attenuates due to factors such as energy dissipation and atmospheric turbulence. When the cumulative velocity loss is lower than a preset threshold, the impact of the wake on the downstream wind turbine can be ignored.

[0097] Furthermore, in this embodiment, the second ordered sequence is traversed downstream from the root wind turbine to find the maximum number of consecutive rows with a superimposed velocity loss ≥ a preset velocity loss threshold. This number of rows is the effective length of the wake, thereby ensuring that the effective length can accurately reflect the actual influence range of the wake, as shown in the following relationship (7): (7) In the formula: Indicates the effective length value; This indicates a preset speed loss threshold; This indicates the row number of the wind turbine in the second ordered sequence.

[0098] Step a3: Determine the first ordered sequence based on the effective length value and the second ordered sequence.

[0099] In an alternative embodiment, The wind turbines in the middle are located according to their downwind coordinates. Sort the data from smallest to largest to form an ordered sequence, namely the first ordered sequence, as shown in the following relation (8): (8) The constraints are as follows: (9) In the formula: Represents the first ordered sequence; Indicates the first The total number of fans in the tail flow chain; Indicates the wake chain In China, according to The number after sorting from smallest to largest One fan; Indicates the number of rows of the fan within the fan chain; Indicates the first The head fan of the tail flow chain.

[0100] In some optional implementations, step S205 above includes: Step S2051: Based on the two-dimensional wake influence matrix, calculate multiple velocity loss values ​​and multiple turbulence enhancement values ​​for multiple second fans.

[0101] In one optional embodiment, the two-dimensional wake effect matrix stores the core data of the second wind turbine in a wake chain-row structure, while velocity deficit and turbulence enhancement are the core indicators characterizing the wake effect. The principle is as follows: using the first wind turbine in the wake chain as a benchmark (rather than the entire field average), by comparing the measured wind speed and turbulence intensity of the second wind turbine in the chain with that of the first wind turbine, the benchmark deviation caused by the complex flow field of the wind farm can be eliminated, and the actual wake effect experienced by each second wind turbine can be accurately quantified.

[0102] In an alternative embodiment, for each element in the two-dimensional wake influence matrix, i.e., a specific wake chain The first in For the exhaust fan, the velocity deficit and turbulence enhancement are calculated using the following equations (10) and (11): (10) (11) In the formula: Indicates the first Tail chain, first The dimensionless velocity loss of the exhaust fan; Indicates wind turbine The measured wind speed; This indicates the wake chain to which the wind turbine belongs. Chain head fan The measured wind speed; Indicates the first Tail chain, first The dimensionless turbulence of the exhaust fan is enhanced. Indicates wind turbine Measured turbulence intensity; This indicates the wake chain to which the wind turbine belongs. Chain head fan The measured turbulence intensity.

[0103] Step S2052: Based on multiple velocity deficit values ​​and multiple turbulence enhancement values, analyze the wake development law of multiple second fans and generate a conditionally averaged wake map that reflects the spatial structure of the fan wake.

[0104] Specifically, step S2052 includes: Step b1: Calculate multiple average velocity deficit values ​​and multiple average turbulence intensity values ​​based on multiple velocity deficit values ​​and multiple turbulence enhancement values.

[0105] In one optional embodiment, the length (number of rows) of different wake chains changes dynamically due to wind conditions. Directly calculating the data of all wind turbines will result in deviations because some wake chains lack corresponding number of wind turbines.

[0106] Furthermore, this embodiment employs an effective ranking aggregation strategy, filtering only those containing the first... Averaging the wake chain data of the exhaust fan ensures that the statistical results accurately reflect the general development pattern of the wake at the same downwind distance (number of rows), eliminating statistical bias caused by missing data.

[0107] Furthermore, the mean velocity deficit and mean turbulence intensity are calculated using the following relationships (12) and (13): (12) (13) In the formula: Indicates the first Average speed loss of the exhaust fan; Indicates the first Average turbulence intensity value of the exhaust fan; Indicates the first The actual length of the tail flow chain, i.e. the total number of rows of the fan. Represents an indicator function, when the wake chain... The length is greater than or equal to (That is, the wake chain has the first) When the function is set to "exhaust fan", the function value is 1; otherwise, it is 0. Furthermore, this function filters all instances containing at least... The logic of the exhaust fan's wake chain instance; Indicates satisfaction The total number of wake chain instances is used to calculate the average value.

[0108] Step b2: Using the Gaussian function, analyze and fit the distribution of multiple velocity loss values ​​across the wind direction coordinates, and calculate multiple target loss values.

[0109] In an alternative embodiment, the target deficit value is used to characterize the deficit between the wake width and the centerline velocity.

[0110] In one optional embodiment, the velocity loss of the second wind turbine in the same row (downwind distance) exhibits a continuous normal distribution across the cross-wind direction coordinates. The principle is as follows: by fitting this distribution with a Gaussian function, the discrete wind turbine velocity loss data is transformed into a continuous wake lateral profile model. The amplitude, standard deviation, and other parameters of the fitting result can directly quantify the wake centerline velocity loss and wake width, achieving a quantitative characterization of the wake lateral structure.

[0111] In an alternative embodiment, for the same row number Analyze the velocity deficit of all wake chains in this row. In cross-wind coordinates The distribution on the upper surface is fitted with a Gaussian function, and the wake width and centerline velocity loss at that downwind distance can be quantitatively calculated, as shown in the following relationship (14): (14) In the formula: Indicates the first The speed loss of the exhaust fan varies with the crosswind coordinate. A changing fitted function; This represents the amplitude of the Gaussian curve; This indicates the center of the Gaussian curve and represents the directional position of the wake centerline; The standard deviation of the Gaussian curve determines the width of the wake; This represents the background deviation, which is usually close to 0.

[0112] Step b3: Based on multiple average velocity deficit values, multiple average turbulence intensity values, and multiple target deficit values, generate a conditionally averaged wake map that reflects the wake spatial structure.

[0113] In one alternative embodiment, the spatial structure and development pattern of the wake are significantly affected by inflow conditions (such as wind speed range and atmospheric stability), and a single map cannot be adapted to different engineering scenarios.

[0114] Furthermore, the average velocity deficit, average turbulence intensity, and target deficit value are classified and aggregated according to the inflow conditions. The longitudinal development pattern (downwind) and lateral structural characteristics (across wind directions) of the wake are integrated to generate a visual map that can intuitively present the spatiotemporal distribution of the wake under specific conditions.

[0115] Furthermore, the wake effect matrix of continuous time slices can be animated to dynamically demonstrate the spatiotemporal evolution of the wake region.

[0116] In one instance, such as Figure 3 As shown, a method for extracting wind farm wake features based on dynamic wake chain meshing is presented. Its core lies in dynamically combining transient wind conditions with the spatial location of the wind turbine to construct and analyze the wake influence matrix. The specific implementation plan is as follows: Step 1: Data spatiotemporal alignment and wind direction coordinate system 1.

[0117] First, radar data from the nacelles of all wind turbines in the wind farm are collected, including wind speed, wind direction, and turbulence intensity. All data are strictly aligned according to their timestamps to ensure time synchronization. Then, for each independent data sampling interval, the latitude and longitude geographic coordinates of all wind turbines are uniformly transformed to a wind flow direction coordinate system with "downwind distance" and "cross-wind distance" as the coordinate axes, based on the prevailing wind direction determined in that time period. This step maps all wind turbines into a dynamic grid consistent with the wind flow direction. The calculation method is shown in the above relation (3). Furthermore, the coordinate transformation method is as follows: Figure 4 As shown.

[0118] Step 2: Dynamic identification and effective length determination of the main wake path.

[0119] This step abandons the wake chain division method based on simple geometric clustering and instead adopts a dynamic "dominant wake path" identification method based on physical influences to accurately construct the wake propagation path. This is the key innovation of this invention when considering the wake superposition effect.

[0120] First, to quantify the mutual influence between wind turbines, a global wake influence potential matrix is ​​constructed. The elements of this matrix Characterized by upstream wind turbines For downstream target wind turbines The potential maximum single wake path influence intensity is shown in the above relationship (1).

[0121] For each target wind turbine in the wind farm We evaluate the matrix The Columns, assigning them a primary source of influence. The primary source of influence is the one with the greatest potential impact among all possible upstream wind turbines, as shown in equation (4) above.

[0122] Subsequently, the wind turbines were clustered based on their primary sources of influence. All turbines located upstream of the same source were clustered together. Wind turbines identified as the primary source of impact were grouped into the same set, along with... Together, they form a dominant wake path. As shown in the above relation (5).

[0123] For each identified dominant wake path , will set All wind turbines, according to their downwind coordinates Sort the data from smallest to largest to form an ordered sequence. .in This refers to the root wind turbine in this path. The actual effective length of the path. The dynamic decay caused by the wake effect is determined. From Start, along the sequence Inspecting downstream, for the first... individual fan Calculate the actual cumulative rate of loss it suffers. This calculation requires the use of a wake superposition model, which comprehensively considers the influence of all upstream wind turbines that contribute to it, as shown in the above relationship (6).

[0124] Set a negligible speed loss threshold. The effective length of the path Defined as satisfying Maximum number of consecutive rows As shown in the above relation (7).

[0125] Furthermore, The wind turbines in the middle are located according to their downwind coordinates. Sort the data from smallest to largest to form an ordered sequence, as shown in relation (8) above. The constraints are shown in relation (9) above.

[0126] Furthermore, the wake chain determination in the above steps is as follows: Figure 5 As shown.

[0127] Step 3: Construction and quantification of the wake influence matrix.

[0128] To quantify the wake effect, a two-dimensional "wake effect matrix" is constructed for each sampling interval. The row index of this matrix represents different wake chains, and the column index represents the number of wind turbines in the same wake chain from near to far (i.e., downwind distance).

[0129] For each element in the matrix, i.e. a specific wake chain The first in For the exhaust fan, two core calculations are performed: velocity deficit and turbulence enhancement, as shown in equations (10) and (11) above.

[0130] Furthermore, taking the inflow at time T as an example, the matrix structure is illustrated as follows: Figure 6 As shown.

[0131] Step 4: Statistical extraction of wake development patterns based on effective row number.

[0132] Due to the dynamic changes in the coverage area of ​​the airflow, the length of each wake chain (i.e., the number of rows it contains) varies at different times. To solve this problem, the present invention adopts an analysis strategy of aggregation based on the effective number of rows to extract unbiased wake development patterns, as shown in the above relationships (12) and (13).

[0133] Furthermore, for the same row number Analyze the velocity deficit of all wake chains in this row. In cross-wind coordinates The distribution on the upper surface is fitted with a Gaussian function, and the wake width and centerline velocity loss at the downwind distance can be quantitatively calculated, i.e., the target loss value, as shown in the above relationship (14).

[0134] Step 5: Conditional wake map generation and model validation.

[0135] Finally, based on the above processing results, a "conditionally averaged wake map" can be generated according to different inflow conditions (such as wind speed range and atmospheric stability category). This map can intuitively display the wake spatial structure under specific conditions. Furthermore, the wake influence matrix of continuous time slices can be animated to dynamically demonstrate the spatiotemporal evolution of the wake region. The high-precision wake data output by this method (such as wake development curves and width expansion rates) can serve as a gold standard for calibrating and validating advanced wake numerical models such as FLORIS and CFD.

[0136] The wind farm wake feature extraction method based on dynamic wake chain meshing provided in this example has the following effects: 1. Dynamism and accuracy: By dynamically determining the dominant wind direction and reference wind turbine for each sampling interval and constructing the wake chain, the distortion problem of traditional averaging methods in partial wind flow coverage scenarios is effectively solved, and the spatiotemporal dynamic evolution of the wake is accurately captured.

[0137] 2. Intra-chain benchmark error elimination: The benchmark wind turbine inside the wake chain is used instead of the average of the entire field as a reference, which eliminates the error of inaccurate benchmark wind speed caused by the complex flow field inside the wind farm, making the calculation results of velocity deficit and turbulence enhancement more realistic and reliable.

[0138] 3. Powerful statistical representation capabilities: Through the "effective ranking" aggregation strategy, it can extract clear and universal statistical patterns of wake development with downwind distance from massive discrete data, and quantify wake width, providing direct basis for wind farm design and optimization.

[0139] 4. Universality and practicality of results: The final generated "conditionally averaged wake map" and dynamic data can not only deepen the understanding of wake physics, but also serve as a valuable resource for verifying and calibrating various wake models, directly serving engineering practice.

[0140] This embodiment also provides a wind farm wake feature extraction device, which is used to implement the above embodiments and preferred embodiments; details already described will not be repeated. As used below, the term "module" can refer to a combination of software and / or hardware that performs a predetermined function. Although the device described in the following embodiments is preferably implemented in software, hardware implementation, or a combination of software and hardware, is also possible and contemplated.

[0141] This embodiment provides a wind farm wake feature extraction device, such as... Figure 7 As shown, the device includes: The acquisition module 701 is used to acquire the target nacelle radar dataset of multiple first wind turbines in the wind farm.

[0142] The first construction module 702 is used to construct a global wake impact potential matrix based on the target nacelle radar dataset. Each element in the global wake impact potential matrix is ​​used to characterize the intensity of the potential maximum single wake path impact of the upstream wind turbine on the downstream target wind turbine.

[0143] The processing module 703 is used to dynamically identify the mainstream wake path and determine its effective length based on the global wake influence potential matrix, so as to obtain a first ordered sequence containing multiple second wind turbines.

[0144] The second construction module 704 is used to construct a two-dimensional wake influence matrix based on the first ordered sequence.

[0145] Analysis module 705 is used to analyze the velocity deficit and turbulence enhancement of multiple second wind turbines based on a two-dimensional wake influence matrix, and generate a conditionally averaged wake map that reflects the spatial structure of the wind turbine wake.

[0146] The wind farm wake feature extraction device provided in this embodiment of the invention can execute the wind farm wake feature extraction method provided in any embodiment of the invention, and has the corresponding functional modules and beneficial effects of the method. Further functional descriptions of the above modules are the same as in the corresponding embodiments described above, and will not be repeated here.

[0147] Figure 8 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention.

[0148] The following is a detailed reference. Figure 8This diagram illustrates a suitable structural schematic for implementing an electronic device according to embodiments of the present invention. The electronic device may include a processor (e.g., a central processing unit, graphics processor, etc.) 801, which can perform various appropriate actions and processes based on a program stored in read-only memory (ROM) 802 or a program loaded from memory 808 into random access memory (RAM) 803. The RAM 803 also stores various programs and data required for the operation of the electronic device. The processor 801, ROM 802, and RAM 803 are interconnected via a bus 804. An input / output (I / O) interface 805 is also connected to the bus 804.

[0149] Typically, the following devices can be connected to I / O interface 805: input devices 806 including, for example, touchscreens, touchpads, keyboards, mice, cameras, microphones, accelerometers, gyroscopes, etc.; output devices 807 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; memory devices 808 including, for example, magnetic tapes, hard disks, etc.; and communication devices 809. Communication device 809 allows electronic devices to communicate wirelessly or wiredly with other devices to exchange data. Although Figure 8 Electronic devices with various devices are shown, but it should be understood that it is not required to implement or have all of the devices shown, and more or fewer devices may be implemented or have instead.

[0150] In particular, according to embodiments of the present invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of the present invention include a computer program product comprising a computer program carried on a non-transitory 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 809, or installed from a memory 808, or installed from a ROM 802. When the computer program is executed by the processor 801, it performs the functions defined in the wind farm wake feature extraction method of the embodiments of the present invention.

[0151] Figure 8 The electronic device shown is merely an example and should not be construed as limiting the functionality and scope of use of the embodiments of the present invention.

[0152] This invention also provides a computer-readable storage medium. The methods described above according to embodiments of the invention can be implemented in hardware or firmware, or implemented as computer code that can be recorded on a storage medium, or implemented as computer code downloaded via a network and originally stored on a remote storage medium or a non-transitory machine-readable storage medium and then stored on a local storage medium. Thus, the methods described herein can be processed by software stored on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. The storage medium can be a magnetic disk, optical disk, read-only memory, random access memory, flash memory, hard disk, or solid-state drive, etc.; further, the storage medium can also include combinations of the above types of memory. It is understood that computers, processors, microprocessor controllers, or programmable hardware include storage components capable of storing or receiving software or computer code. When the software or computer code is accessed and executed by the computer, processor, or hardware, the wind farm wake feature extraction method shown in the above embodiments is implemented.

[0153] A portion of this invention can be applied as a computer program product, such as computer program instructions, which, when executed by a computer, can invoke or provide the methods and / or technical solutions according to the invention through the operation of the computer. Those skilled in the art will understand that the forms in which computer program instructions exist in a computer-readable medium include, but are not limited to, source files, executable files, installation package files, etc. Correspondingly, the ways in which computer program instructions are executed by a computer include, but are not limited to: the computer directly executing the instructions, or the computer compiling the instructions and then executing the corresponding compiled program, or the computer reading and executing the instructions, or the computer reading and installing the instructions and then executing the corresponding installed program. Here, the computer-readable medium can be any available computer-readable storage medium or communication medium accessible to a computer.

[0154] Although embodiments of the invention have been described in conjunction with the accompanying drawings, those skilled in the art can make various modifications and variations without departing from the spirit and scope of the invention, and such modifications and variations all fall within the scope defined by the appended claims.

Claims

1. A method for extracting wake features from a wind farm, characterized in that, The method includes: Acquire target nacelle radar datasets for multiple first-stage wind turbines in a wind farm; Based on the target nacelle radar dataset, a global wake impact potential matrix is ​​constructed. Each element in the global wake impact potential matrix is ​​used to characterize the potential maximum single wake path impact intensity of the upstream wind turbine on the downstream target wind turbine. Based on the global wake impact potential matrix, the mainstream wake path is dynamically identified and its effective length is determined to obtain a first ordered sequence containing multiple second wind turbines. Based on the first ordered sequence, construct a two-dimensional wake influence matrix; Based on the two-dimensional wake influence matrix, the velocity deficit and turbulence enhancement of multiple second wind turbines are analyzed, and a conditionally averaged wake spectrum reflecting the spatial structure of the wind turbine wake is generated.

2. The method according to claim 1, characterized in that, Acquire target nacelle radar datasets for multiple first-stage wind turbines in a wind farm, including: Obtain the initial nacelle radar dataset of the plurality of first wind turbines in the wind farm; The initial cabin radar dataset is spatiotemporally aligned to obtain the first cabin radar dataset. The wind turbine geographic coordinates corresponding to the first nacelle radar dataset are processed using the wind direction coordinate system to obtain the target nacelle radar dataset.

3. The method according to claim 1, characterized in that, Based on the global wake impact potential matrix, dynamic identification and effective length determination of the mainstream wake path are performed to obtain a first ordered sequence containing multiple second wind turbines, including: The global wake impact potential matrix is ​​evaluated, and multiple major impact sources of multiple third wind turbines among the multiple first wind turbines are identified, wherein the major impact source refers to the upstream wind turbine with the greatest potential impact on the third wind turbine among all upstream wind turbines corresponding to the third wind turbine. Based on the multiple main sources of influence, the multiple first wind turbines are clustered and a dominant wake path is constructed; According to the coordinates of the first wind turbine in the downwind direction, the multiple first wind turbines in the dominant wake path are sorted to obtain a second ordered sequence; Based on the wake superposition model, the effective length of the second ordered sequence is determined to obtain the first ordered sequence containing the plurality of second wind turbines.

4. The method according to claim 3, characterized in that, Based on the aforementioned wake superposition model, the effective length of the second ordered sequence is determined to obtain the first ordered sequence containing the plurality of second wind turbines, including: Based on the aforementioned wake superposition model, calculate multiple superposition velocity loss values ​​for the multiple first wind turbines in the second ordered sequence; The effective length of the dominant wake path is determined based on the multiple superimposed velocity loss values ​​and the preset velocity loss threshold. The first ordered sequence is determined based on the effective length value and the second ordered sequence.

5. The method according to claim 1, characterized in that, Based on the aforementioned two-dimensional wake influence matrix, the velocity deficit and turbulence enhancement of multiple second wind turbines are analyzed, and a conditionally averaged wake spectrum reflecting the spatial structure of the wind turbine wake is generated, including: Based on the two-dimensional wake influence matrix, multiple velocity deficit values ​​and multiple turbulence enhancement values ​​of the multiple second wind turbines are calculated; Based on the multiple velocity deficit values ​​and multiple turbulence enhancement values, the wake development law of the multiple second wind turbines is analyzed, and a conditionally averaged wake spectrum reflecting the spatial structure of the wind turbine wake is generated.

6. The method according to claim 5, characterized in that, Based on the multiple velocity deficit values ​​and multiple turbulence enhancement values, the wake development patterns of the multiple second wind turbines are analyzed, and a conditionally averaged wake spectrum reflecting the spatial structure of the wind turbine wake is generated, including: Based on the multiple velocity deficit values ​​and the multiple turbulence enhancement values, multiple average velocity deficit values ​​and multiple average turbulence intensity values ​​are calculated respectively; Using a Gaussian function, the distribution of the multiple velocity deficit values ​​across the wind direction coordinates is analyzed and fitted, and multiple target deficit values ​​are calculated. The target deficit values ​​are used to characterize the deficit between the wake width and the centerline velocity. Based on the plurality of average velocity deficit values, the plurality of average turbulence intensity values, and the plurality of target deficit values, a conditionally averaged wake map reflecting the wake spatial structure is generated.

7. A wind farm wake feature extraction device, characterized in that, The device includes: The acquisition module is used to acquire the target nacelle radar dataset of multiple first wind turbines in a wind farm; The first construction module is used to construct a global wake impact potential matrix based on the target nacelle radar dataset. Each element in the global wake impact potential matrix is ​​used to characterize the potential maximum single wake path impact intensity of the upstream wind turbine on the downstream target wind turbine. The processing module is used to dynamically identify the mainstream wake path and determine its effective length based on the global wake influence potential matrix, so as to obtain a first ordered sequence containing multiple second wind turbines. The second construction module is used to construct a two-dimensional wake influence matrix based on the first ordered sequence; The analysis module is used to analyze the velocity deficit and turbulence enhancement of multiple second wind turbines based on the two-dimensional wake influence matrix, and generate a conditionally averaged wake map that reflects the spatial structure of the wind turbine wake.

8. An electronic device, characterized in that, include: The system includes a memory and a processor, which are interconnected. The memory stores computer instructions, and the processor executes the computer instructions to perform the wind farm wake feature extraction method according to any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions for causing the computer to execute the wind farm wake feature extraction method according to any one of claims 1 to 6.

10. A computer program product, characterized in that, The method includes computer instructions for causing a computer to execute the wind farm wake feature extraction method according to any one of claims 1 to 6.

Images