Wake flow loss estimation method and device based on topological graph, terminal equipment and storage medium

Abstract

The invention discloses a wake flow loss estimation method and device based on a topological graph, terminal equipment and a storage medium, and belongs to the technical field of fan wake flow loss estimation. Constructing a current wind power plant aerodynamic interaction topological graph and node features; calculating the weight of the corresponding directional edge according to the longitudinal distance and the transverse offset; then obtaining a current inlet wind speed based on the current wind power plant aerodynamic interaction topological graph, the weight, the node features and a preset inlet wind speed estimation model; and finally, the current wake flow wind speed loss value is calculated according to the difference between the current inlet wind speed and the current environment wind speed. Through the implementation of the method, the problem that in the prior art, when fluid dynamics is used for simulating and calculating the fan inlet wind speed, the requirement for real-time inlet wind speed estimation cannot be met, and then wake flow effect estimation cannot be conducted based on the real-time inlet wind speed can be solved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of wind turbine wake loss estimation, and in particular to a wake loss estimation method and device based on a topological graph, a terminal device and a storage medium. BACKGROUND

[0002] With the increasing demand for clean energy worldwide, wind power has developed rapidly as an important form of renewable energy. In large wind farms, the arrangement density of wind power units (usually referred to as wind turbine generators) is increasing, leading to increasing aerodynamic interaction between wind power units, especially the wake effect. The wake effect refers to the formation of a region of wind speed loss and increased turbulence intensity downstream of an upstream wind power unit. When other wind power units are located in this wake region, the wind energy they can capture is reduced, and the power generation is decreased. At the same time, the unstable airflow also increases the fatigue load of the units, affecting their service life. Therefore, accurately estimating the actual inlet wind speed of each wind power unit and then estimating the wake effect based on the inlet wind speed is crucial for optimizing wind farm control strategies (such as yaw control, power control), improving overall power generation efficiency, and conducting fine performance evaluation and fault diagnosis.

[0003] In the prior art, the inlet wind speed of wind power units is usually estimated using computational fluid dynamics (CFD) simulation. However, this method requires high-precision wake field information, and during the entire calculation process, it needs to capture turbulence details (such as boundary layers, wake vortices), and encrypt grids, etc., resulting in a grid number of up to tens of millions, thus greatly increasing the calculation cost, and unable to meet the demand for real-time inlet wind speed estimation of wind power units in wind farms. Therefore, the prior art has the problem that when using fluid dynamics simulation to calculate the inlet wind speed of the wind turbine, due to the large amount of calculation and high precision requirement of the original data, it cannot meet the demand for real-time inlet wind speed estimation, and thus cannot estimate the wake effect based on real-time inlet wind speed. SUMMARY

[0004] The present application provides a wake loss estimation method and device based on a topological graph, a terminal device and a storage medium, which can solve the problem in the prior art that when using fluid dynamics simulation to calculate the inlet wind speed of the wind turbine, due to the large amount of calculation and high precision requirement of the original data, it cannot meet the demand for real-time inlet wind speed estimation, and thus cannot estimate the wake effect based on real-time inlet wind speed.

[0005] An embodiment of the present application provides a wake loss estimation method based on a topological graph, comprising:

[0006] acquire current environmental data of the wind farm and current operating parameters of each wind power generation unit in the wind farm; wherein the current environmental data includes: current global dominant wind direction and current environmental wind speed; the current operating parameters include: current nacelle wind speed, current active power, current pitch angle and current yaw angle;

[0007] take each wind power generation unit as a node, construct a directed edge for each node according to the current global dominant wind direction, and generate a current wind farm aerodynamic interaction topology graph;

[0008] For each wind power generation unit connected by a directed edge, acquire the longitudinal distance and lateral offset between the wind power generation units;

[0009] According to the longitudinal distance and the lateral offset, the weight of the corresponding directed edge is calculated;

[0010] According to the current operating parameters and the current global dominant wind direction, the node characteristics corresponding to each node are determined;

[0011] The current wind farm aerodynamic interaction topology graph, the weight and the node characteristics are input into a preset inlet wind speed estimation model to estimate the current inlet wind speed of each wind power generation unit in the wind farm;

[0012] For each wind power generation unit, the current wake wind speed loss value of the wind power generation unit is calculated according to the difference between the current inlet wind speed and the current environmental wind speed.

[0013] Further, the directed edge constructed for each node according to the current global dominant wind direction includes:

[0014] Acquire the fan impeller diameter of each wind power generation unit;

[0015] For any two selected wind power generation units, the connecting line between the selected wind power generation units is parallel to the current global dominant wind direction, and the selected wind power generation unit located downstream is within the wake influence distance of the selected wind power generation unit located upstream, a directed edge is constructed from the selected wind power generation unit located upstream to the selected wind power generation unit located downstream; wherein the wake influence distance is the product of the fan impeller diameter of the selected wind power generation unit located upstream and a preset diameter multiple;

[0016] For any two selected wind power generation units, a directed edge is constructed from the wind power generation unit located upstream to the wind power generation unit located downstream, wherein the directed edge is not parallel to the current global prevailing wind direction, and the wind power generation unit located downstream is in the wake influence area of the wind power generation unit located upstream, and the wake influence area is determined according to the fan impeller diameter of the wind power generation unit located upstream, the first longitudinal distance between the two selected wind power generation units, the first transverse offset, and a preset wake model.

[0017] Further, the weight of the directed edge is calculated according to the longitudinal distance and the transverse offset, including:

[0018] Obtain the current environmental turbulence intensity;

[0019] For any directed edge, a longitudinal attenuation factor is calculated according to the current environmental turbulence intensity, the longitudinal distance, and the fan impeller diameter of the wind power generation unit located upstream;

[0020] A transverse overlap factor is calculated according to the transverse offset, the fan impeller diameter of the wind power generation unit located upstream, the fan impeller diameter of the wind power generation unit located downstream, and the longitudinal distance;

[0021] A wake influence indicator function value is calculated according to the transverse offset, the longitudinal distance, and the fan impeller diameter of the wind power generation unit located downstream;

[0022] The weight is calculated according to the product of the longitudinal attenuation factor, the transverse overlap factor, and the wake influence indicator function value.

[0023] Further, the node characteristics corresponding to each node are determined according to the current operating parameters and the current global prevailing wind direction, including:

[0024] For each node, the current nacelle wind speed and the current active power of the wind power generation unit corresponding to the node are denoised, and a current graph signal is generated according to the denoised current nacelle wind speed and the denoised current active power;

[0025] The difference between the current yaw error angle and the current global prevailing wind direction is calculated, and the current yaw error angle is obtained.

[0026] The current graph signal, the current yaw error angle, the current global prevailing wind direction, and the current pitch angle are taken as the node characteristics of the corresponding node.

[0027] Further, the construction of the preset inlet wind speed estimation model includes:

[0028] Obtaining a plurality of training samples with real labels; wherein the training samples include: a plurality of historical aerodynamic interaction topology graphs of wind farms, historical weights of each historical directed edge in the historical aerodynamic interaction topology graphs of wind farms, and historical node features corresponding to each node in the historical aerodynamic interaction topology graphs of wind farms; the real labels are used to represent the real inlet wind speeds corresponding to each node in the historical aerodynamic interaction topology graphs of wind farms;

[0029] Inputting the training sample data with real labels into the inlet wind speed estimation model to be trained for iterative training until the loss function converges, to generate the preset inlet wind speed estimation model;

[0030] Wherein, in each iteration training, according to the current training sample, the node features are aggregated and transformed, and the estimated inlet wind speed of each wind power generation unit in the current historical aerodynamic interaction topology graph of wind farms is estimated; according to the current estimated inlet wind speed and the corresponding real label, the current loss function is calculated, and whether the current loss function converges is determined; if the current loss function converges, the current inlet wind speed estimation model is taken as the preset inlet wind speed estimation model; otherwise, after adjusting the parameters in the current inlet wind speed estimation model, the training is continued.

[0031] Further, the calculation of the current loss function according to the current estimated inlet wind speed and the corresponding real label includes:

[0032] Obtaining a preset theoretical power curve of the wind power generation unit corresponding to the current estimated inlet wind speed, and an active power corresponding to the current estimated inlet wind speed; wherein the horizontal coordinate of the preset theoretical power curve is the inlet wind speed, and the vertical coordinate is the theoretical active power;

[0033] According to the current inlet wind speed and the corresponding preset theoretical power curve, the theoretical active power value under the current inlet wind speed is determined;

[0034] According to the theoretical active power value, the active power corresponding to the current estimated inlet wind speed, the current estimated inlet wind speed and the corresponding real label, the current loss function is calculated.

[0035] Further, after calculating the current wake wind speed loss value of the wind power generation unit, it further includes:

[0036] Obtaining a preset normal wake wind speed loss threshold range of the wind power generation unit;

[0037] In a case where the current wake wind speed loss value exceeds the preset normal wake wind speed loss threshold range, a fault warning is performed.

[0038] On the basis of the above-mentioned method embodiment, the application correspondingly provides a device embodiment;

[0039] The application provides a wake loss estimation device based on a topology graph, comprising:

[0040] The data acquisition module, the topology graph construction module, the wind power generation unit data acquisition module, the weight calculation module, the node feature determination module, the inlet wind speed estimation module and the wake wind speed loss calculation module;

[0041] The data acquisition module is configured to acquire current environmental data of a wind farm and current operating parameters of each wind power generation unit in the wind farm, wherein the current environmental data comprises a current global dominant wind direction and a current environmental wind speed, and the current operating parameters comprise a current nacelle wind speed, a current active power, a current pitch angle and a current yaw angle.

[0042] The topology graph construction module is configured to take each wind power generation unit as a node, construct a directed edge for each node according to the current global dominant wind direction, and generate a current wind farm aerodynamic interaction topology graph.

[0043] The wind power generation unit data acquisition module is configured to acquire a longitudinal distance and a lateral offset between the wind power generation units for each wind power generation unit connected by a directed edge.

[0044] The weight calculation module is configured to calculate a weight of a corresponding directed edge according to the longitudinal distance and the lateral offset.

[0045] The node feature determination module is configured to determine a node feature of each node according to the current operating parameters and the current global dominant wind direction.

[0046] The inlet wind speed estimation module is configured to input the current wind farm aerodynamic interaction topology graph, the weight and the node feature into a preset inlet wind speed estimation model to estimate a current inlet wind speed of each wind power generation unit in the wind farm.

[0047] The wake wind speed loss calculation module is configured to calculate, for each wind power generation unit, a current wake wind speed loss value of the wind power generation unit according to a difference between the current inlet wind speed and the current environmental wind speed.

[0048] On the basis of the above-mentioned method embodiment, the application correspondingly provides a terminal device embodiment;

[0049] The application provides a terminal device, comprising a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, and the processor implements the method for estimating wake loss based on a topological graph according to any one of the embodiments of the application when executing the computer program.

[0050] Based on the above-mentioned method embodiment, the application provides a storage medium.

[0051] The application provides a storage medium, comprising a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, and the processor implements the method for estimating wake loss based on a topological graph according to any one of the embodiments of the application when executing the computer program.

[0052] The embodiments of the application have the following beneficial effects:

[0053] The application provides a topological graph-based wake loss estimation method and device, a terminal equipment and a storage medium. The method comprises the following steps: obtaining current environmental data of a wind farm and current operating parameters of each wind power unit in the wind farm; wherein the current environmental data comprises a current global dominant wind direction and a current environmental wind speed; the current operating parameters comprise a current nacelle wind speed, a current active power, a current pitch angle and a current yaw angle; then, taking each wind power unit as a node, a directed edge is constructed for each node according to the current global dominant wind direction, and a current wind farm aerodynamic interaction topological graph is generated; then, for each wind power unit connected by a directed edge, the longitudinal distance and the lateral offset between the wind power units are obtained; then, the weight of the corresponding directed edge is calculated according to the longitudinal distance and the lateral offset; then, the node characteristics corresponding to each node are determined according to the current operating parameters and the current global dominant wind direction; then, the current wind farm aerodynamic interaction topological graph, the weight and the node characteristics are input into a preset inlet wind speed estimation model to estimate the current inlet wind speed of each wind power unit in the wind farm; finally, for each wind power unit, the current wake wind speed loss value of the wind power unit is calculated according to the difference between the current inlet wind speed and the current environmental wind speed. Therefore, the wind power units in the wind farm and the current environmental data are used to construct a topological graph, and then the node characteristics and the weight of the nodes and the directed edges in the topological graph are respectively assigned according to the current operating parameters of the wind power units, so as to obtain the input data of the preset inlet wind speed estimation model. Finally, the current inlet wind speed is estimated in the preset inlet wind speed estimation model combined with the input data, and then the current wake loss value is calculated. When estimating the inlet wind speed, fluid dynamics simulation is not required, and only the current environmental data of the wind farm and the current operating parameters of the wind power units are required to generate the topological graph and the related information of each node and directed edge in the topological graph, so that the complex calculation process caused by fluid dynamics simulation is greatly reduced, the calculation difficulty is reduced, the real-time estimation of the inlet wind speed is met, and the real-time wake effect estimation result based on the real-time inlet wind speed can be further obtained. BRIEF DESCRIPTION OF DRAWINGS

[0054] In order to more clearly illustrate the technical solutions of the present application, the following will briefly introduce the drawings needed in the embodiments. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without any creative effort.

[0055] Figure 1 FIG. 1 is a flowchart of a topological graph-based wake loss estimation method according to an embodiment of the present application.

[0056] Figure 2 is a wind farm aerodynamic interaction topology diagram provided by an embodiment of the present application.

[0057] Figure 3 is a structural schematic diagram of a wake deficit estimation device based on a topology diagram provided by an embodiment of the present application. DETAILED DESCRIPTION

[0058] In order to make the objects, technical solutions and advantages of the present application clearer, the technical solutions in the present application will be described clearly and completely below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the scope of protection of the present application.

[0059] 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 the present application belongs; the terms used herein are only for the purpose of describing specific embodiments and are not intended to limit the present application; the terms "include" and "have" and any variations thereof in the specification and claims of the present application and the above description of drawings are intended to cover non-exclusive inclusion.

[0060] In the description of the embodiments of the present application, the technical terms "first", "second", etc. are only used to distinguish different objects, and cannot be understood as indicating or implying relative importance or implicitly indicating the number, specific order or primary and secondary relationship of the indicated technical features. In the description of the embodiments of the present application, the meaning of "a plurality of" is two or more, unless otherwise explicitly and specifically limited.

[0061] Reference herein to "an embodiment" means that a particular feature, structure, or characteristic described in connection with the embodiment can be included in at least one embodiment of the present application. The appearance of the phrase in various places in the specification does not necessarily all refer to the same embodiment, nor is it necessarily independent or alternative embodiments to other embodiments. Those skilled in the art explicitly and implicitly understand that the embodiments described herein can be combined with other embodiments.

[0062] In the description of the embodiments of the present application, the term "and / or" is only a description of the association relationship of the associated objects, which means that there can be three relationships, for example, A and / or B, which can represent the three cases of A alone, A and B together, and B alone. In addition, the character " / " in the present application generally represents an "or" relationship between the associated objects.

[0063] In the description of the embodiments of the present application, the term "a plurality of" refers to two or more (including two), and similarly, "a plurality of groups" refers to two or more groups (including two groups), and "a plurality of pieces" refers to two or more pieces (including two pieces).

[0064] In the description of the embodiments of the present application, unless otherwise explicitly specified and limited, the technical terms "mounting", "connection", "connection", "fixing" and the like should be understood in a broad sense, for example, it can be fixedly connected, or it can be detachably connected, or it can be integrated; it can be mechanically connected, or it can be electrically connected; it can be directly connected, or it can be indirectly connected through an intermediate medium; it can be the internal communication of two elements or the interaction relationship between two elements. For those skilled in the art, the specific meanings of the above terms in the embodiments of the present application can be understood according to the specific circumstances.

[0065] Reference Figure 1 To solve the problem in the prior art that when calculating the inlet wind speed of a fan using fluid dynamics simulation, the calculation amount is large, the original data precision requirement is high, the real-time inlet wind speed estimation demand cannot be met, and then the wake effect cannot be estimated based on the real-time inlet wind speed. An embodiment of the present application provides a wake loss estimation method based on a topological graph, comprising:

[0066] Step S101: obtaining current environmental data of a wind farm and current operating parameters of each wind power generation unit in the wind farm; wherein the current environmental data includes: current global dominant wind direction and current environmental wind speed; the current operating parameters include: current nacelle wind speed, current active power, current pitch angle and current yaw angle;

[0067] Specifically, the operating parameters can be obtained from the SCADA system of the wind farm, and these data usually have a certain time resolution, such as 1 second, 1 minute or 10 minutes. The environmental data can be obtained by a wind measurement tower arranged in the wind farm, or based on regional wind condition information provided by a current numerical weather prediction (NWP) system.

[0068] Step S102: taking each wind power generation unit as a node, constructing a directed edge for each node according to the current global dominant wind direction, and generating a current wind farm aerodynamic interaction topological graph;

[0069] Specifically, since the wind farm aerodynamic interaction topological graph is constructed based on the current related data, the wind farm aerodynamic interaction topological graph can be dynamically changed in real time, which can reflect the aerodynamic coupling relationship between the wind power generation units under real-time wind conditions, especially the wake effect.

[0070] Illustratively, the wind farm aerodynamic interaction topological graph is as shown in Figure 2 ​Figure 2 "WTG1", "WTG2", "WTG3", "WTG4" and "WTG5" are all different wind power generation units. Figure 2 The connection between the wind power generation units is a determined directed edge. The parameters marked on the directed edge represent the corresponding directed edge and the weight of the directed edge. For example, the “e” on the directed edge connecting “WTG1” and “WTG2” is 12 (t)” is the directed edge between “WTG1” and “WTG2”, and “w 12 (t)” represents the weight of the directed edge between “WTG1” and “WTG2” at time t.

[0071] In a preferred embodiment, the above-mentioned construction of directed edges for each node according to the current global dominant wind direction includes:

[0072] Obtaining the diameter of the wind turbine impeller of each wind power generation unit;

[0073] For any two selected wind power generation units, if the line connecting the selected wind power generation units is parallel to the current global dominant wind direction and the downstream selected wind power generation unit is within the wake influence distance of the upstream selected wind power generation unit, a directed edge is constructed from the upstream selected wind power generation unit to the downstream selected wind power generation unit; wherein the wake influence distance is the product of the fan impeller diameter of the upstream selected wind power generation unit and a preset diameter multiple;

[0074] Specifically, when facing the incoming wind direction of the current global dominant wind direction, the two selected wind power generation units are divided into an upstream selected wind power generation unit and a downstream selected wind power generation unit along the incoming wind direction.

[0075] Specifically, the above-mentioned preset diameter multiple is set to 8, that is, if the selected wind power generation unit downstream is located within a distance of 8 times the wind turbine impeller diameter of the selected wind power generation unit upstream, then a directed edge is constructed between the two selected wind power generation units, from the selected wind power generation unit upstream to the selected wind power generation unit downstream.

[0076] For any two selected wind power units, a directed edge is constructed from the selected wind power unit located upstream to the selected wind power unit located downstream, if the connecting line between the selected wind power units is not parallel to the current global prevailing wind direction, and the selected wind power unit located downstream is in the wake influence area of the selected wind power unit located upstream; wherein the wake influence area is determined according to the fan impeller diameter of the selected wind power unit located upstream, the first longitudinal distance between the two selected wind power units, the first lateral offset, and a preset wake model.

[0077] Specifically, the first longitudinal distance is the projected longitudinal distance of the selected wind power unit located upstream to the selected wind power unit located downstream along the current global prevailing wind direction; and the first lateral offset is the lateral offset distance of the selected wind power unit located downstream relative to the wake centerline of the selected wind power unit located upstream relative to the current global prevailing wind direction.

[0078] Specifically, the preset wake model can be Jensen model. When the wake influence area is determined based on the Jensen model, the wake radius is calculated based on the fan impeller diameter of the selected wind power unit located upstream and the longitudinal distance of the selected wind power unit located downstream. Wherein the wake radius of the selected wind power unit located upstream is calculated according to the following formula:

[0079]

[0080] In the formula, r i represents the wake radius of the selected wind power unit i located upstream, D′ i represents the fan impeller diameter of the selected wind power unit i located upstream, k represents the preset wake diffusion coefficient, d′ x,ij (t) represents the first longitudinal distance between the selected wind power unit i located upstream and the selected wind power unit j located downstream at time t.

[0081] Subsequently, the lateral offset angle is calculated according to the longitudinal distance of the selected wind power unit located downstream, and the lateral offset of the selected wind power unit located downstream:

[0082]

[0083] In the formula, θ represents the lateral offset angle, d′ y,ij (t) represents the first lateral offset between the selected wind power unit i located upstream and the selected wind power unit j located downstream at time t.

[0084] Then, when the lateral offset angle is less than the preset lateral offset angle threshold, and the absolute value of the lateral offset of the selected wind power generation unit located downstream is less than the wake radius of the selected wind power generation unit located upstream, a sector region is obtained, which is the wake influence region of the selected wind power generation unit located upstream. If the selected wind power generation unit located downstream is in the wake influence region, a directed edge is constructed between the two selected wind power generation units, with the selected wind power generation unit located upstream pointing to the selected wind power generation unit located downstream.

[0085] As can be seen from the schematic diagram, Figure 2 As can be seen from the schematic diagram, "WTG2" and "WTG3" are downstream wind power generation units relative to "WTG1", and correspondingly, "WTG1" is an upstream wind power generation unit. Both "WTG2" and "WTG3" are in the wake influence region of "WTG1", so "WTG1" has a wake influence on "WTG2" and "WTG3", and directed edges are constructed between "WTG1", "WTG2" and "WTG3", with "WTG1" pointing to "WTG2" and "WTG3". "WTG3" has a wake influence on "WTG4", so a directed edge is constructed between "WTG3" and "WTG4", with "WTG3" pointing to "WTG4". "WTG5" is not in the wake influence region or wake influence distance of the other wind power generation units, so no directed edge is constructed.

[0086] In this preferred embodiment, the current global dominant wind direction is used to construct directed edges between nodes in the aerodynamic interaction topology graph of the wind farm.

[0087] Step S103: For each wind power generation unit connected by a directed edge, the longitudinal distance and the lateral offset between the wind power generation units are obtained.

[0088] Specifically, similar to the definition of the first longitudinal distance and the first lateral offset, the longitudinal distance is the projected longitudinal distance of the wind power generation unit at the starting end node of the directed edge (i.e., the upstream wind power generation unit) to the wind power generation unit at the terminal end node of the directed edge (i.e., the downstream wind power generation unit) along the current global dominant wind direction. The lateral offset is the lateral offset distance of the wind power generation unit at the terminal end node of the directed edge relative to the wake centerline of the wind power generation unit at the starting end node of the directed edge relative to the current global dominant wind direction.

[0089] Specifically, the longitudinal distance can be calculated according to the following formula:

[0090]

[0091] In the formula, dx,ij (t) represents the longitudinal distance between the upstream wind power unit i and the downstream wind power unit j at time t, X i represents the abscissa of the position where the upstream wind power unit i is located, Y i represents the abscissa of the position where the upstream wind power unit i is located, X j represents the abscissa of the position where the downstream wind power unit j is located, Y j represents the abscissa of the position where the downstream wind power unit j is located, represents the current global dominant wind direction.

[0092] Specifically, the above-mentioned lateral offset can be calculated according to the following formula:

[0093]

[0094] In the formula, d y,ij (t) represents the lateral offset between the upstream wind power unit i and the downstream wind power unit j at time t.

[0095] Step S104: calculating the weight of the directed edge according to the above-mentioned longitudinal distance and the above-mentioned lateral offset;

[0096] Specifically, the above-mentioned weight is used to represent the influence intensity of the wind turbine unit corresponding to the upstream node on the downstream node wind power unit, so the weight can be designed to be proportional to the thrust coefficient of the upstream wind turbine unit, and related to the decay function constructed by the above-mentioned longitudinal distance and lateral offset.

[0097] In a preferred embodiment, the above-mentioned calculating the weight of the directed edge according to the above-mentioned longitudinal distance and the above-mentioned lateral offset comprises:

[0098] Obtaining the current environmental turbulence intensity;

[0099] For any directed edge, the longitudinal attenuation factor is calculated according to the current environmental turbulence intensity, the longitudinal distance, and the wind turbine impeller diameter of the wind power unit located upstream.

[0100] Specifically, the longitudinal attenuation factor is calculated by the following formula:

[0101]

[0102] In the formula, f decay represents the above-mentioned longitudinal attenuation factor, which is used to represent the effect that the tail flow intensity of the upstream wind power unit i decreases with the increase of the downstream distance, and considers the influence of the environmental turbulence intensity, TI t represents the environmental turbulence intensity at time t, and α and β represent adjustable parameters.

[0103] a lateral overlap factor is calculated according to the lateral offset, the fan impeller diameter of the wind power unit located upstream, the fan impeller diameter of the wind power unit located downstream, and the longitudinal distance;

[0104] Specifically, the lateral overlap factor is calculated according to the following formula:

[0105]

[0106]

[0107] wherein f overlap represents the lateral overlap factor, which is used to describe the overlapping degree of the swept area of the impeller of the downstream wind power unit j and the wake area of the upstream wind power unit i at the longitudinal distance d x,ij (t), k wake (t) represents the wake expansion coefficient at time t, which is used to describe the expansion rate of the wake cone, and is related to the environmental turbulence intensity and the operating state of the upstream wind power unit. According to the existing literature, k wake (t) ≈ 0.04 + 0.5 · TI t , D j represents the fan impeller diameter of the downstream wind power unit j, A overlap represents the overlapping area (i.e. the wake disc of the downstream wind power unit j, which is centered on the wake centerline of the upstream wind power unit i and has a wake radius R wake,i (d x,ij (t) ) at the lateral offset d y,ij (t), R wake,i represents the wake radius of the upstream wind power unit i, and σ wake (d x,ij (t) ) represents the effective width (i.e. the standard deviation) of the wake at the longitudinal distance d x,ij (t), which is proportional to R wake,i (d x,ij (t) ).

[0108] a wake influence indicator function value is calculated according to the lateral offset, the longitudinal distance, and the fan impeller diameter of the downstream wind power unit;

[0109] Specifically, the wake influence indicator function is a binary function, which takes the value 1 if the downstream wind power unit j is indeed located within the wake influence area or the wake influence distance of the upstream wind power unit i, and takes the value 0 otherwise. Therefore, based on this logic, if there is no directed edge between any two wind power units, the corresponding weight is 0. That is, if dx,ij (t)≤0, indicating that the downstream wind power unit j is not within the wake influence area or wake influence distance of the upstream wind power unit i, at which time the wake influence indicator function value is 0; if d x,ij (t) > 0, it is further determined whether the downstream wind power unit j is within the wake cone, and the wake cone half angle can be obtained by tan(0 wake ) = k wake (t), so if the wake influence indicator function value is 1.

[0110] The weight is calculated according to the product of the above longitudinal attenuation factor, the above transverse overlap factor, and the above wake influence indicator function value.

[0111] Specifically, the weight is calculated according to the following formula:

[0112]

[0113] In the formula, w ij (t) represents the weight of the directed edge between the upstream wind power unit i and the downstream wind power unit j at time t, C Ti (t) represents the fan thrust coefficient of the upstream wind power unit i at time t, which is used to represent the ability of the upstream wind power unit i to extract momentum from the airflow, and is directly related to the strength of the wake generated by the upstream wind power unit i, and can be obtained in combination with the SCADA data of the upstream wind power unit i and the performance curve of the upstream wind power unit i, I wake represents the wake influence indicator function value.

[0114] Preferably, this weight integrates the wake generation potential of the upstream fan (i.e., the fan thrust coefficient), the strength attenuation of the wake in the process of propagating downstream (i.e., the longitudinal attenuation factor), and the extent to which the downstream fan is actually covered by the wake (i.e., the transverse overlap factor). Moreover, the wake influence indicator function ensures that the weight is only calculated when there is physically possible wake influence.

[0115] In this preferred embodiment, the weight corresponding to the directed edge is calculated by the longitudinal distance and the transverse offset.

[0116] Step S105: determining the node characteristics corresponding to each node according to the current operating parameters and the current global dominant wind direction;

[0117] In a preferred embodiment, the determination of the node characteristics corresponding to each node according to the current operating parameters and the current global dominant wind direction comprises:

[0118] For each node, the wind power unit corresponding to the node, the current nacelle wind speed and the current active power are denoised, and a current graph signal is generated according to the denoised current nacelle wind speed and the denoised current active power;

[0119] Specifically, the data is denoised by low-pass filtering the current nacelle wind speed and the current active power, and the low-pass filtering helps to remove isolated measurement noise and extract more consistent data components in space.

[0120] A current yaw error angle is obtained by calculating the difference between the current yaw angle of the wind power unit corresponding to the node and the current global dominant wind direction.

[0121] The current graph signal, the current yaw error angle, the current global dominant wind direction and the current pitch angle are taken as the node features of the corresponding node.

[0122] Specifically, for each node, the node features can be a feature vector containing the graph signal, the yaw error angle, the global dominant wind direction and the pitch angle.

[0123] In this preferred embodiment, the current operating parameters and the current global dominant wind direction determine the node features of each node.

[0124] Step S106: inputting the current wind farm aerodynamic interaction topology, the weights and the node features into a preset inlet wind speed estimation model to estimate the current inlet wind speed of each wind power unit in the wind farm.

[0125] Specifically, the preset inlet wind speed estimation model is a graph neural network model that aggregates neighbor node information through its network layers (such as convolution layers and attention layers) and combines its own node features to learn the complex nonlinear mapping from input to target output. Therefore, during model processing, the model can process features according to its multi-layer message passing mechanism, and then estimate the inlet wind speed.

[0126] In a preferred embodiment, the construction of the preset inlet wind speed estimation model includes:

[0127] Obtain a plurality of training samples with real labels; wherein the training samples include: a plurality of historical wind farm aerodynamic interaction topologies, historical weights of each historical directed edge in the historical wind farm aerodynamic interaction topologies, and historical node features corresponding to each node in the historical wind farm aerodynamic interaction topologies; the real labels are used to represent the real inlet wind speed corresponding to each node in the historical wind farm aerodynamic interaction topologies;

[0128] Specifically, the preset inlet wind speed estimation model is trained by a supervised learning method. The training samples can be historical environmental data of a wind farm and historical operating parameters of each wind power unit in the wind farm, historical aerodynamic interaction topology diagrams of the wind farm, historical weights, and historical node features obtained therefrom, or environmental data and operating parameters generated by high-fidelity CFD simulation under various wind conditions and operating conditions.

[0129] The training sample data with real labels are input into the inlet wind speed estimation model to be trained for iterative training until the loss function converges, and the preset inlet wind speed estimation model is generated.

[0130] In each iteration training, the node features are aggregated and transformed according to the current training sample to estimate the estimated inlet wind speed of each wind power unit in the current historical wind farm aerodynamic interaction topology diagram. The current loss function is calculated according to the current estimated inlet wind speed and the corresponding real label, and it is determined whether the current loss function converges. If the current loss function converges, the current inlet wind speed estimation model is taken as the preset inlet wind speed estimation model. Otherwise, the parameters in the current inlet wind speed estimation model are adjusted, and the training is continued.

[0131] In this preferred embodiment, the trained preset inlet wind speed estimation model is obtained by iterative training of the preset inlet wind speed estimation model by supervised learning until the loss function converges.

[0132] In another preferred embodiment, the calculation of the current loss function according to the current estimated inlet wind speed and the corresponding real label includes:

[0133] The preset theoretical power curve of the wind power unit corresponding to the current estimated inlet wind speed and the active power corresponding to the current estimated inlet wind speed are obtained. The abscissa of the preset theoretical power curve is the inlet wind speed, and the ordinate is the theoretical active power.

[0134] The theoretical active power value under the current inlet wind speed is determined according to the current inlet wind speed and the corresponding preset theoretical power curve.

[0135] The current loss function is calculated according to the theoretical active power value, the active power corresponding to the current estimated inlet wind speed, the current estimated inlet wind speed, and the corresponding real label.

[0136] Specifically, the loss function includes a data fitting term and a physical constraint term, and the loss function is represented by the following formula:

[0137] L total =L MSE +λphy L phy

[0138] In the formula, L total represents a loss function, L MSE represents the mean square error between the estimated inlet wind speed and the corresponding true label, L phy represents the norm (for example, the mean square error) of the difference between the theoretical active power value and the actual active power of the wind power generation unit (that is, the active power corresponding to the current estimated inlet wind speed), which is taken as a physical constraint term in this loss function, λ phy represents the weight of the physical constraint term.

[0139] Preferably, in the loss function, the introduction of the physical constraint helps to improve the generalization ability and physical consistency of the model under the condition of insufficient training data or the presence of noise.

[0140] In this preferred embodiment, the current loss function is calculated according to the current estimated inlet wind speed and the corresponding true label.

[0141] Step S107: For each wind power generation unit, the current wake wind speed loss value of the wind power generation unit is calculated according to the difference between the current inlet wind speed and the current ambient wind speed.

[0142] Specifically, the wake wind speed loss can be used to measure the influence of the corresponding wind power generation unit on the wake.

[0143] In a preferred embodiment, after calculating the current wake wind speed loss value of the wind power generation unit, the method further comprises:

[0144] Obtaining a preset normal wake wind speed loss threshold range of the wind power generation unit;

[0145] In the case where the current wake wind speed loss value exceeds the preset normal wake wind speed loss threshold range, a fault warning is performed.

[0146] Specifically, when the wake wind speed loss is abnormal, it may indicate that the fan impeller of the upstream wind power generation unit is stalled, the blade surface is contaminated or damaged, and therefore a fault warning is needed to remind the relevant staff to perform on-site troubleshooting and fault repair.

[0147] In this preferred embodiment, whether the current needs to be warned is determined by comparing the current wake wind speed loss value with the preset normal wake wind speed loss threshold range.

[0148] On the basis of the above-mentioned method embodiment, the present application correspondingly provides a device embodiment.

[0149] As Figure 3As shown, an embodiment of the present application provides a topological graph-based wake deficit estimation device, comprising:

[0150] a data acquisition module, a topological graph construction module, a wind power unit data acquisition module, a weight calculation module, a node feature determination module, an inlet wind speed estimation module, and a wake wind speed deficit calculation module;

[0151] The above-mentioned data acquisition module is configured to acquire current environmental data of a wind farm and current operating parameters of each wind power unit in the wind farm, wherein the current environmental data includes a current global dominant wind direction and a current environmental wind speed, and the current operating parameters include a current nacelle wind speed, a current active power, a current pitch angle, and a current yaw angle.

[0152] The above-mentioned topological graph construction module is configured to take each wind power unit as a node, construct a directed edge for each node according to the current global dominant wind direction, and generate a current wind farm aerodynamic interaction topological graph.

[0153] The above-mentioned wind power unit data acquisition module is configured to acquire a longitudinal distance and a lateral offset between wind power units for each wind power unit connected by a directed edge.

[0154] The above-mentioned weight calculation module is configured to calculate a weight of a corresponding directed edge according to the longitudinal distance and the lateral offset.

[0155] The above-mentioned node feature determination module is configured to determine a node feature corresponding to each node according to the current operating parameters and the current global dominant wind direction.

[0156] The above-mentioned inlet wind speed estimation module is configured to input the current wind farm aerodynamic interaction topological graph, the weight, and the node feature into a preset inlet wind speed estimation model to estimate a current inlet wind speed of each wind power unit in the wind farm.

[0157] The above-mentioned wake wind speed deficit calculation module is configured to calculate a current wake wind speed deficit value of each wind power unit according to a difference between the current inlet wind speed and the current environmental wind speed.

[0158] It should be noted that the apparatus embodiments described above are only illustrative, and the modules described above as separate components can or can not be physically separated, and the components displayed as modules can or can not be physical modules, i.e., can be located in one place, or can be distributed to multiple network units. Part or all of the modules can be selected to achieve the purpose of the embodiment according to actual needs. In addition, the connection relationship between the modules in the apparatus embodiments provided by the present application indicates that there is a communication connection between them, which can be implemented as one or more communication buses or signal lines. Those skilled in the art can understand and implement it without creative labor. The above schematic diagram is only an example of a wake loss estimation device based on a topology graph, and does not constitute a limitation on a wake loss estimation device based on a topology graph, which can include more or fewer components than the diagram, or combine some components, or different components.

[0159] On the basis of the above-mentioned method embodiments, the present application correspondingly provides terminal device embodiments.

[0160] Another embodiment of the present application provides a terminal device, comprising a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein the processor executes the computer program to implement the above-mentioned wake loss estimation method based on a topology graph according to any one of the embodiments of the present application.

[0161] For example, in this embodiment, the computer program can be divided into one or more modules, and the one or more modules are stored in the memory and executed by the processor to complete the present application. The one or more modules can be a series of computer program instruction segments that can complete a specific function, and the instruction segments are used to describe the execution process of the computer program in the device.

[0162] The terminal device can be a desktop computer, a notebook computer, a palm computer, a cloud server, and other computing devices. The device can include, but is not limited to, a processor and a memory.

[0163] The processor can be a central processing module (CPU), and can also be other general-purpose processors, digital signal processors (DSP), application specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or can also be any conventional processor, etc. The processor is a control center of the device, and is connected with various parts of the device through various interfaces and lines.

[0164] The memory can be used to store the computer program and / or the module, and the processor realizes various functions of the device by running or executing the computer program and / or the module stored in the memory and calling the data stored in the memory. The memory can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system, at least one application required by a function, etc.; in addition, the memory can include a high-speed random access memory, and can also include a nonvolatile memory, for example, a hard disk, a memory, a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, at least one disk storage device, a flash memory device, or other volatile solid-state memory devices.

[0165] On the basis of the method embodiment, the application provides a storage medium embodiment.

[0166] Another embodiment of the application provides a storage medium, which includes a stored computer program, wherein the computer program controls a device where the storage medium is located to execute the tail flow loss estimation method based on a topology graph according to any one of the embodiments of the application when the computer program is running.

[0167] In this embodiment, the storage medium is a computer readable storage medium, the computer program includes computer program code, and the computer program code can be in the form of source code, object code, an executable file, or some intermediate form, etc. The computer readable medium can include any entity or device capable of carrying the computer program code, a recording medium, a U disk, a mobile hard disk, a magnetic disk, an optical disk, a computer memory, a read-only memory (ROM), a random access memory (RAM), an electrical carrier signal, a telecommunication signal, a software distribution medium, etc.

[0168] The above is the preferred embodiment of the present application, and it should be pointed out that, for those skilled in the art, without departing from the principles of the present application, a number of improvements and refinements can be made, which are also considered within the scope of protection of the present application.

Claims

1. A method for estimating wake loss based on a topological graph, characterized in that: include: Obtaining current environmental data of the wind farm and current operating parameters of each wind power generation unit in the wind farm; wherein the current environmental data includes: current global dominant wind direction and current ambient wind speed; current operating parameters include: current nacelle wind speed, current active power, current pitch angle, and current yaw angle; Taking each wind power generation unit as a node, a directed edge is constructed for each node according to the current global dominant wind direction to generate the current wind farm aerodynamic interaction topology graph; For each wind power generation unit connected by a directed edge, obtaining the longitudinal distance and the lateral offset between the wind power generation units; Calculate the weight of the corresponding directed edge according to the longitudinal distance and the lateral offset; Determine the node characteristics corresponding to each node based on the current operating parameters and the current global dominant wind direction; Inputting the current wind farm aerodynamic interaction topology, weights, and node characteristics into a preset inlet wind speed estimation model to estimate the current inlet wind speed of each wind power generation unit in the wind farm; For each wind power generation unit, the current wake wind speed loss value of the wind power generation unit is calculated based on the difference between the current inlet wind speed and the current ambient wind speed.

2. A method for estimating wake loss based on a topological graph according to claim 1, characterized in that: The step of constructing directed edges for each node according to the current global dominant wind direction includes: Obtaining the diameter of the wind turbine impeller of each wind power generation unit; For any two selected wind power generation units, if the line connecting the selected wind power generation units is parallel to the current global dominant wind direction and the downstream selected wind power generation unit is within the wake influence distance of the upstream selected wind power generation unit, a directed edge is constructed from the upstream selected wind power generation unit to the downstream selected wind power generation unit; wherein the wake influence distance is the product of the fan impeller diameter of the upstream selected wind power generation unit and a preset diameter multiple; For any two selected wind power generation units, when the line between the selected wind power generation units is not parallel to the current global dominant wind direction, and among the selected wind power generation units, the selected wind power generation unit located downstream is in the wake influence area of ​​the selected wind power generation unit located upstream, a directed edge is constructed from the selected wind power generation unit located upstream to the selected wind power generation unit located downstream; wherein the wake influence area is determined based on the wind turbine impeller diameter of the selected wind power generation unit located upstream, the first longitudinal distance and the first lateral offset between the two selected wind power generation units, and a preset wake model.

3. A method for estimating wake loss based on a topological graph according to claim 2, characterized in that: The calculating the weight of the corresponding directed edge according to the longitudinal distance and the lateral offset includes: Get the current environmental turbulence intensity; For any directed edge, the longitudinal attenuation factor is calculated based on the current ambient turbulence intensity, longitudinal distance, and the diameter of the wind turbine rotor of the upstream wind turbine. Calculating a lateral overlap factor according to the lateral offset, a diameter of a wind turbine impeller of an upstream wind power generation unit, a diameter of a wind turbine impeller of a downstream wind power generation unit, and the longitudinal distance; Calculating a wake influence indicator function value according to the lateral offset, the longitudinal distance, and the diameter of a wind turbine impeller of a downstream wind power generation unit; The weight is calculated according to the product of the longitudinal attenuation factor, the lateral overlap factor, and the wake influence indicator function value.

4. A method for estimating wake loss based on a topological graph according to claim 3, characterized in that: Determining the node characteristics corresponding to each node based on the current operating parameters and the current global dominant wind direction includes: For each node, the wind power generation unit corresponding to the node, the corresponding current nacelle wind speed and the current active power are denoised, and the current graph signal is generated based on the denoised current nacelle wind speed and the denoised current active power; Calculating the difference between the current yaw angle of the wind power generation unit corresponding to the node and the current global dominant wind direction to obtain a current yaw error angle; The current graph signal, the current yaw error angle, the current global dominant wind direction, and the current pitch angle are used as node features of the corresponding node.

5. A method for estimating wake loss based on a topological graph according to claim 4, characterized in that: The construction of the preset inlet wind speed estimation model includes: Acquire a plurality of training samples with true labels, wherein the training samples include: a plurality of historical wind farm aerodynamic interaction topology graphs, the historical weight of each historical directed edge in the historical wind farm aerodynamic interaction topology graphs, and the historical node features corresponding to each node in the historical wind farm aerodynamic interaction topology graphs; the true labels are used to represent the true inlet wind speed corresponding to each node in the historical wind farm aerodynamic interaction topology graphs; Inputting the training sample data with real labels into the inlet wind speed estimation model to be trained for iterative training until the loss function converges, thereby generating the preset inlet wind speed estimation model; Among them, during each iterative training, the node features are aggregated and transformed according to the current training samples, and the estimated inlet wind speed of each wind power generation unit in the current historical wind farm aerodynamic interaction topology diagram is estimated; according to the current estimated inlet wind speed and the corresponding true label, the current loss function is calculated, and it is judged whether the current loss function converges; if the current loss function converges, the current inlet wind speed estimation model is used as the preset inlet wind speed estimation model; otherwise, the parameters in the current inlet wind speed estimation model are adjusted and training continues.

6. The method for estimating wake loss based on a topological graph according to claim 5, wherein: The current loss function is calculated based on the current estimated inlet wind speed and the corresponding true label, including: Obtaining a preset theoretical power curve of the wind power generation unit corresponding to the current estimated inlet wind speed, and the active power corresponding to the current estimated inlet wind speed; wherein the abscissa of the preset theoretical power curve is the inlet wind speed, and the ordinate is the theoretical active power; Determine the theoretical active power value at the current inlet wind speed based on the current inlet wind speed and the corresponding preset theoretical power curve; The current loss function is calculated based on the theoretical active power value, the active power corresponding to the current estimated inlet wind speed, the current estimated inlet wind speed, and the corresponding true label.

7. The method for estimating wake loss based on a topological graph according to claim 6, wherein: After calculating the current wake wind speed loss value of the wind power generation unit, the method further includes: Obtaining a preset normal wake wind speed loss threshold range of the wind power generation unit; When the current wake wind speed loss value exceeds the preset normal wake wind speed loss threshold range, a fault warning is issued.

8. A wake loss estimation device based on a topological map, characterized in that: include: Data acquisition module, topology map construction module, wind power generation unit data acquisition module, weight calculation module, node feature determination module, inlet wind speed estimation module and wake wind speed loss calculation module; The data acquisition module is used to obtain the current environmental data of the wind farm and the current operating parameters of each wind power generation unit in the wind farm; wherein the current environmental data includes: the current global dominant wind direction and the current ambient wind speed; the current operating parameters include: the current nacelle wind speed, the current active power, the current pitch angle and the current yaw angle; The topology map construction module is used to take each wind power generation unit as a node, construct directed edges for each node according to the current global dominant wind direction, and generate the current wind farm aerodynamic interaction topology map; The wind power generation unit data acquisition module is used to acquire the longitudinal distance and lateral offset between the wind power generation units for each wind power generation unit connected by a directed edge; The weight calculation module is used to calculate the weight of the corresponding directed edge according to the longitudinal distance and the lateral offset; The node feature determination module is used to determine the node features corresponding to each node based on the current operating parameters and the current global dominant wind direction; The inlet wind speed estimation module is used to input the current wind farm aerodynamic interaction topology, weights and node characteristics into a preset inlet wind speed estimation model to estimate the current inlet wind speed of each wind power generation unit in the wind farm; The wake wind speed loss calculation module is used to calculate the current wake wind speed loss value of each wind power generation unit according to the difference between the current inlet wind speed and the current ambient wind speed.

9. A terminal device, characterized in that: The invention comprises a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein when the processor executes the computer program, the method for estimating wake loss based on a topological map according to any one of claims 1 to 7 is implemented.

10. A storage medium, characterized in that: The storage medium includes a stored computer program, wherein when the computer program is executed, the device where the storage medium is located is controlled to execute the wake loss estimation method based on a topology map according to any one of claims 1 to 7.

Images