A wind farm power dynamic optimization method and system based on spectral clustering partitioning

Abstract

The application discloses a wind farm power dynamic optimization method and system based on spectral clustering partitioning, the method comprises the following steps: constructing a long-time scale wind farm power maximization steady-state optimization problem based on an engineering wake model according to average wind speed and wind direction, and solving an optimal yaw angle; calculating a wake coupling relationship of the wind farm under the current wind direction with the optimal yaw angle to construct a wake diagram; clustering the wake diagram based on a spectral clustering method, and dividing the wind farm into multiple sub-regions; constructing an optimization objective for different sub-regions, including a short-time scale wind farm power dynamic optimization problem of a wind farm total power maximization and a wind turbine thrust fluctuation minimization, and solving the short-time scale wind farm power dynamic optimization problem to obtain optimal control quantities of each unit. The wind farm maximum power output optimization method aims to make up for the deficiency of a traditional static optimization strategy under a dynamic environment, and dynamically improves the overall power generation of the wind farm.

Description

Technical Field

[0001] The present invention relates to the technical field of wind farm power dynamic optimization, and in particular to a wind farm power dynamic optimization method and system based on spectral clustering partitioning. Background Art

[0002] As a vital component of clean energy, wind power generation continues to see rapid technological advancement and growth in installed capacity. After years of development, wind power technology is moving toward large-scale, intelligent operation, but this brings with it increasingly complex operational and control challenges. In large-scale wind farms in particular, the complex aerodynamic coupling between wind turbines, caused by the wake effect, results in significant dynamic variations in key parameters such as wind speed, power output, and mechanical loads. This places higher demands on the optimal control of the wind farm's overall power. Traditional wind farm optimization and control methods are often based on static models, typically assuming constant wind speed or simply using deterministic models to mathematically represent turbine power, thereby transforming the power generation optimization problem into a high-dimensional, nonlinear, and nonconvex problem. However, this approach often overlooks the uncertainty of wind speed and the complex dynamic coupling between wind turbines, making it difficult to achieve the expected efficiency gains in actual operation and potentially exacerbating turbine fatigue loads. Summary of the Invention

[0003] Technical problem to be solved by the present invention: In response to the above-mentioned problems in the prior art, a method and system for dynamic optimization of wind farm power based on spectral clustering partitioning are provided. The present invention aims to make up for the shortcomings of traditional static optimization strategies in dynamic environments and dynamically improve the overall power generation of wind farms.

[0004] In order to solve the above technical problems, the technical solution adopted by the present invention is:

[0005] A method for dynamic optimization of wind farm power based on spectral clustering partitioning includes the following steps:

[0006] According to the average wind speed and direction, a long-term wind farm power maximization steady-state optimization problem is constructed based on the engineering wake model and the optimal yaw angle is solved;

[0007] The wake coupling relationship of the wind farm under the current wind direction is calculated using the optimal yaw angle to construct a wake map;

[0008] The wake map is clustered based on the spectral clustering method, and the wind farm is divided into multiple sub-areas;

[0009] A short-time-scale wind farm power dynamic optimization problem with optimization objectives including maximizing the total wind farm power and minimizing the thrust fluctuation of wind turbines is constructed for different sub-areas respectively. The short-time-scale wind farm power dynamic optimization problem is solved to obtain the optimal control quantity of each unit.

[0010] Optionally, the construction of the engineering wake model includes:

[0011] S101, for wind turbines i For fans n The wake effect generated by the wind turbine i The yaw action will cause the center of its wake to deflect, which is defined as yaw deflection. The wake deflection distance of the yaw deflection is determined. for:

[0012] ,

[0013] in, Representative fan i The initial deflection angle, represents the impeller radius of the fan, represents the deflection expansion coefficient, Representative fan i For fans n The deflection constant that produces the wake is calculated as:

[0014] ,

[0015] ,

[0016] ,

[0017] in, Representative fan i With fan n The geographical distance between them along the wind direction, Representative fan i The yaw angle, Representative fan i The thrust coefficient, Representative fan i Axial induction factor;

[0018] S102, define the deflection of the wake center caused by the impeller rotation itself as rotational deflection, calculate the fan i For fans n The resulting rotational deflection for:

[0019] ,

[0020] in, and is a constant;

[0021] S103, calculate the fan according to the following formula i For fans n The coordinate of the center of the wake effect on the y-axis :

[0022] ,

[0023] wherein, represents the wind turbine i wherein the geographical position is located y axis coordinate;

[0024] S104, the wake effect will gradually expand with the propagation process, the wind turbine i to the wind turbine n generated by the wake expansion radius :

[0025] ,

[0026] wherein, represents the maximum value, represents the wake expansion coefficient, represents the scale constant of the wake area q , wherein, q represents the wake area, q =1,2,3 respectively represent the near wake area, far wake area and mixed area;

[0027] S105, according to the following formula to calculate the wind turbine i to the wind turbine n generated by the wake wind speed :

[0028] ,

[0029] ,

[0030] ,

[0031] ,

[0032] ,

[0033] ,

[0034] ,

[0035] ,

[0036] wherein, represents the wind turbine i inflow wind speed, represents the wind turbine i to the wind turbine n wind speed attenuation rate, represents the wake attenuation coefficient, represents the wind turbinei wake region induced by the wind turbine n wake overlap area of the wake region induced by the wind turbine q is the swept area of the rotor; is a wake decay correction parameter, , , and are constants; is the y-coordinate of the wind turbine i wake region induced by the wind turbine n is the expansion radius of the wake region induced by the wind turbine q and are two angles used to calculate the area of the wake overlap area of the wake region induced by the wind turbine is the distance between the coordinate of the wind turbine n in the y-axis direction and the wake center induced by the wind turbine i wake region induced by the wind turbine n is the y-coordinate of the wind turbine q i wake center induced by the wind turbine n is the distance between the coordinate of the wind turbine n in the y-axis direction and the wake center induced by the wind turbine is the y-coordinate of the wind turbine n

[0037] Optionally, the constructing a long-time-scale wind farm power maximization steady-state optimization problem and solving an optimal yaw angle based on an engineering wake model according to an average wind speed and a wind direction comprises:

[0038] S201, calculating the power output of the wind turbine according to the following formula: i

[0039] ,

[0040] ,

[0041] wherein, is an axial induction factor of the wind turbine i is a yaw angle of the wind turbine is an inflow wind speed of the wind turbine i is an air density is a swept area of the rotor i is a power coefficient of the wind turbine is a generator efficiency is a yaw-induced power loss coefficient; i p

[0042] ​​​​​​​​​S202, considering the wake effect inside the wind farm, the inflow wind speed of each wind turbine is related to the free wind speed, the axial induction factor and the yaw angle of the upstream wind turbine, and the total power output function of the wind farm is constructed. :

[0043] ,

[0044] in, is the inflow wind speed of the wind farm, represents the set of all fan axial induction factors, represents the set of all wind turbine yaw angles, Represents the number of fans, represents the power of fan i;

[0045] S203, construct the long-time-scale wind farm power maximization steady-state optimization problem shown in the following equation:

[0046] ,

[0047] ,

[0048] in, Representative fan The upper and lower limits of the axial induction factor, Representative fan The upper and lower limits of the yaw angle, Representative fan The upper and lower limits of active output;

[0049] S204, using a heuristic algorithm such as a particle swarm optimization algorithm to solve the long-term steady-state optimization problem of wind farm power maximization to obtain the optimal yaw angle.

[0050] Optionally, the step of calculating the wake coupling relationship of the wind farm in the current wind direction with the optimal yaw angle to construct a wake map includes: constructing the wind speed attenuation rate of each wind turbine in the wind farm calculated when solving the optimal yaw angle into the following formula: The matrix:

[0051] ,

[0052] Any element in row i and column j of the matrix Representative fan i For fans j The wind speed attenuation rate; the wind turbines are respectively regarded as weighted undirected graphs The wind speed attenuation rate between the nodes and wind turbines in the graph is a weighted undirected graph Construct a weighted undirected graph using the edges in :

[0053] ,

[0054] in, Represents a weighted undirected graph If the fan i and fan j If there is wake coupling between ,otherwise = 0; n is the number of fans.

[0055] Optionally, clustering the wake map based on a spectral clustering method to divide the wind farm into multiple sub-areas includes:

[0056] S301, define the fan according to the following formula i The degree of the corresponding node is a weighted undirected graph The sum of all edges of :

[0057] ,

[0058] in, n Represents the number of fans, Represents a weighted undirected graph edge;

[0059] According to all fans i The degree of the corresponding node Defining the degree matrix :

[0060] ;

[0061] S302, calculate the Laplace matrix according to the following formula :

[0062] ,

[0063] in, Represents a weighted undirected graph; the normalized Laplace matrix is ​​calculated according to the following formula :

[0064] ,

[0065] S303, calculate the normalized Laplace matrix The eigenvalue of e , and calculate e The second-order difference of :

[0066] ,

[0067] in, Representative i +2 eigenvalues ​​andi +1 eigenvalue difference, Representative i +1 eigenvalue and i The difference of eigenvalues, Representative i +2 eigenvalues, Representative i eigenvalues, Representative i +1 eigenvalue;

[0068] S304, design condition threshold, from i =3, when the eigenvalue The number of clusters is determined when the following conditions are met K = i :

[0069] ,

[0070] in, Representative i -1 second-order difference, is an adjustable parameter, Representative j A second-order difference, Representative j -1 second-order difference;

[0071] S305, based on the clustering idea, establish the minimum segmentation problem:

[0072] ,

[0073] in, Represents segmentation, ~ Respectively represent belonging to subset 1~ K A collection of nodes, represent and The split weight between Represents a subset The set of points, represent The complement of ; choose to solve this minimum segmentation problem through the "normalized segmentation" method:

[0074] ,

[0075] in, stands for the "normalized segmentation" method, represent The weights and edges of , through the spectral clustering method, the above minimum partitioning problem is transformed into the following form:

[0076] ,

[0077] in, Represents the matrix to be solved, the superscript represents transpose, represent n -dimensional identity matrix, represents the normalized Laplace matrix;

[0078] S306, by solving the normalized Laplace matrix Before K eigenvectors to obtain the matrix to be solved , and then based on the k-means clustering method to finally obtain the clustering results of wind turbine nodes in the wind farm.

[0079] Optionally, constructing a short-time-scale wind farm power dynamic optimization problem with optimization objectives including maximizing the total wind farm power and minimizing the thrust fluctuation of wind turbines for different sub-areas includes:

[0080] S401, determine the fan according to the following formula i The thrust :

[0081] ,

[0082] in, For fans i The axial induction factor, For fans i The yaw angle, For fans i The inflow wind speed, is the air density, is the impeller swept area, For fans i The thrust coefficient, is the thrust loss coefficient of the wind turbine in yaw state;

[0083] S402: Considering that the thrust on the wind turbine changes with the wind speed in a dynamic state, the fluctuation of the thrust on the wind turbine is suppressed while increasing the total power of the wind farm. The following short-time-scale dynamic optimization problem of wind farm power is constructed:

[0084] ,

[0085] in, represents the weight of the optimization objective, Represents the number of fans, Representative fan i The optimal yaw angle, Representative fani Power output at the optimal yaw angle, Representative fan i The thrust at the optimal yaw angle; Representative fan i The initial value of the thrust;

[0086] S403, based on the spectral clustering results, the wind farm has been divided N c The optimization problem of each sub-region can be split into separate sub-regions for separate calculations. The short-time-scale wind farm power dynamic optimization problem can be rewritten into the following sub-region form:

[0087] ,

[0088] st ,

[0089] in, Representatives belong to k The axial induction factor of the fans in the subset, For the k Sub-centralized fans i The axial induction factor, Representative k The number of fans in the subset, Representative fan i The upper and lower limits of the axial induction factor, Representative fan i The upper and lower limits of active output.

[0090] Optionally, when solving the short-time-scale wind farm power dynamic optimization problem to obtain the optimal control variable of each unit, the optimal control variable includes the axial induction factor of each wind turbine.

[0091] In addition, the present invention also provides a wind farm power dynamic optimization system based on spectral clustering partitioning, comprising a microprocessor and a memory connected to each other, wherein the microprocessor is programmed or configured to execute the wind farm power dynamic optimization method based on spectral clustering partitioning.

[0092] In addition, the present invention also provides a computer-readable storage medium, which stores a computer program or instruction. The computer program or instruction is programmed or configured to execute the wind farm power dynamic optimization method based on spectral clustering partitioning through a processor.

[0093] In addition, the present invention also provides a computer program product, including a computer program or instructions, which is programmed or configured to execute the wind farm power dynamic optimization method based on spectral clustering partitioning through a processor.

[0094] Compared with the existing technology, the present invention can achieve the following beneficial effects: the wind farm power dynamic optimization method based on spectral clustering partitioning of the present invention first constructs a long-time scale wind farm total power maximization optimization problem based on the average wind speed and wind direction, calculates the optimal yaw angle, and then considers the change of inflow wind speed, combines the two optimization objectives of maximizing the total power of the wind farm and minimizing the thrust fluctuation of the wind turbine to construct a short-time scale wind farm power dynamic optimization problem, and introduces the spectral clustering method to divide the wind farm into multiple sub-regions according to the wake coupling relationship, and divides the original optimization problem into multiple sub-problems with smaller dimensions for parallel solution, greatly improving the solution speed to adapt to the time-varying wind speed environment. The wind farm power dynamic optimization method based on spectral clustering partitioning of the present invention uses spectral clustering technology to partition the wind turbines in the wind farm, divides the turbines with strong wake coupling into the same sub-region, realizes the coordinated control of the turbines in the sub-region, and can significantly reduce the calculation time of the original optimization problem. This method can not only make up for the shortcomings of traditional static optimization strategies in dynamic environments, but also provide a new technical path for the intelligent control of large-scale wind farms. BRIEF DESCRIPTION OF THE DRAWINGS

[0095] Figure 1 Schematic diagram of the basic process of the method of the embodiment of the present invention.

[0096] Figure 2 Schematic diagram of the structure of the engineering wake model in an embodiment of the present invention.

[0097] Figure 3 Flowchart of a wind farm partitioning method based on spectral clustering in an embodiment of the present invention.

[0098] Figure 4 These are the clustering result diagram and partitioning result diagram of a wind farm in a wind direction of 0 degrees in an embodiment of the present invention, where (a) is the clustering result diagram and (b) is a schematic diagram of wind farm partitioning.

[0099] Figure 5 These are the clustering result diagram and partitioning result diagram of a wind farm in a wind direction of 10 degrees according to an embodiment of the present invention, where (a) is the clustering result diagram and (b) is a schematic diagram of wind farm partitioning.

[0100] Figure 6 These are the clustering result diagram and partitioning result diagram of a wind farm in a wind direction of 20 degrees according to an embodiment of the present invention, where (a) is the clustering result diagram and (b) is a schematic diagram of wind farm partitioning.

[0101] Figure 7 This is a simulation verification diagram of the total power increase of a wind farm under a wind direction of 0 degrees in an embodiment of the present invention.

[0102] Figure 8 This is a simulation verification diagram of the total power increase of a wind farm under a wind direction of 20 degrees in an embodiment of the present invention.

[0103] Figure 9 This is a simulation verification diagram of the total thrust fluctuation suppression of a wind farm at a wind direction of 0 degrees in an embodiment of the present invention.

[0104] Figure 10 This is a simulation verification diagram of the total thrust fluctuation suppression of a wind farm at a wind direction of 20 degrees in an embodiment of the present invention.

[0105] Figure 11 These are simulation verification diagrams of thrust fluctuation suppression of fan #1 and fan #36 in an embodiment of the present invention, where (a) is a simulation verification diagram of thrust fluctuation suppression of fan #1, and (b) is a simulation verification diagram of thrust fluctuation suppression of fan #36.

[0106] Figure 12 This is a comparison chart of the average time between the method of the embodiment of the present invention and the centralized control.

[0107] Figure 13 This is a simulation verification diagram of the computing time saving effect of the method of the embodiment of the present invention compared with centralized control. DETAILED DESCRIPTION

[0108] In order to enable those skilled in the art to better understand the technical solution of the present invention, the technical solution of the present invention will be further described in detail below with reference to the accompanying drawings in the embodiments of the present invention.

[0109] like Figure 1 As shown, the wind farm power dynamic optimization method based on spectral clustering partitioning in this embodiment includes the following steps:

[0110] S1, based on the average wind speed and wind direction, constructs a long-term wind farm power maximization steady-state optimization problem based on the engineering wake model and solves the optimal yaw angle;

[0111] S2, calculates the wake coupling relationship of the wind farm under the current wind direction using the optimal yaw angle to construct a wake map;

[0112] S3, clustering the wake map based on the spectral clustering method to divide the wind farm into multiple sub-areas;

[0113] S4, construct the short-time-scale wind farm power dynamic optimization problem for different sub-areas, including the optimization objectives of maximizing the total wind farm power and minimizing the thrust fluctuation of the wind turbines, and solve the short-time-scale wind farm power dynamic optimization problem to obtain the optimal control quantity of each unit.

[0114] like Figure 2 As shown, for any fan i For example, three wake regions will be formed on the leeward side: the near wake region, the far wake region and the mixing region, q=1,2,3 represent; the wake effect will gradually expand with the propagation process, the fan i to the fan n generated by the wake expansion radius . represent the fan i to the fan n caused by the wake area q wake overlap area. Due to the yawing action of the fan i will cause the wake center deflection defined as yaw deflection, yaw deflection of the wake deflection distance In step S1 of the embodiment, the construction of the engineering wake model includes:

[0115] S101, for the fan i to the fan n generated by the wake effect, the yawing action of the fan i will cause the wake center deflection defined as yaw deflection, yaw deflection of the wake deflection distance :

[0116] ,

[0117] wherein, represent the initial deflection angle of the fan i , represent the impeller radius of the fan, represent the deflection expansion coefficient, represent the fan i to the fan n wake deflection calculation constant, and has:

[0118] ,

[0119] ,

[0120] ,

[0121] wherein, represent the geographical distance between the fan i and the fan n along the wind direction, represent the yaw angle of the fan i , represent the thrust coefficient of the fan i (relevant to the axial induction factor), represent the axial induction factor of the fan i ;

[0122] S102, define the deflection of the wake center caused by the impeller rotation itself as the rotation deflection, calculate the fani For fans n The resulting rotational deflection for:

[0123] ,

[0124] in, and is a constant;

[0125] S103, calculate the fan according to the following formula i For fans n The coordinate of the center of the wake effect on the y-axis :

[0126] ,

[0127] in, Representative fan i Geographical location y axis coordinates;

[0128] S104, the wake effect will gradually expand as the propagation process, the wind turbine i For fans n The resulting wake expansion radius :

[0129] ,

[0130] in, represents the maximum value, represents the wake expansion coefficient, Represents the wake area q The scaling constant, q represents the wake region, q =1, 2, 3 represent the near wake region, far wake region, and mixing region, respectively;

[0131] S105, calculate the fan according to the following formula i For fans n The resulting wake wind speed :

[0132] ,

[0133] ,

[0134] ,

[0135] ,

[0136] ,

[0137] ,

[0138] ,

[0139] ,

[0140] in, Representative fan i The inflow wind speed, Representative fan i For fans n The wind speed attenuation rate, represents the wake attenuation coefficient, Representative fan i For fans n The wake area caused q The wake overlap area, is the impeller swept area; represents the wake attenuation correction parameter, 、 and are all constants; Representative fan i For fans n The wake area caused q The expansion radius; and Represents fans i For fans n The wake area caused q The two angles used to calculate the area of ​​the wake overlap area are: Represents the coordinates of fan n in the y-axis direction and the fan i For fans n The distance between the centers of the wake induced by Representative fan n The y-axis coordinate of .

[0141] In step S1 of this embodiment, according to the average wind speed and wind direction, constructing a long-time-scale wind farm power maximization steady-state optimization problem based on an engineering wake model and solving the optimal yaw angle includes:

[0142] S201, calculate the fan according to the following formula i Power output :

[0143] ,

[0144] ,

[0145] in, Representative fan i The axial induction factor, Representative fani The yaw angle, Representative fan i The inflow wind speed, represents the air density, represents the impeller swept area, Representative fan i The power coefficient, represents the generator efficiency, p represents the power loss coefficient caused by yaw, which is a constant parameter and can be set according to actual needs. For example, in this embodiment, the value is 1.88;

[0146] S202, considering the wake effect inside the wind farm, the inflow wind speed of each wind turbine is related to the free wind speed, the axial induction factor and the yaw angle of the upstream wind turbine, and the total power output function of the wind farm is constructed. :

[0147] ,

[0148] in, is the inflow wind speed of the wind farm, represents the set of all fan axial induction factors, represents the set of all wind turbine yaw angles, Represents the number of fans, represents the power of fan i;

[0149] S203, construct the long-time-scale wind farm power maximization steady-state optimization problem shown in the following equation:

[0150] ,

[0151] ,

[0152] in, Representative fan The upper and lower limits of the axial induction factor, Representative fan The upper and lower limits of the yaw angle, Representative fan The upper and lower limits of active output;

[0153] S204: A heuristic algorithm, such as a particle swarm optimization algorithm, is used to solve the long-term steady-state optimization problem for maximizing wind farm power to determine the optimal yaw angle. Since the wake model is discontinuous when the yaw angle is uncertain, the optimal yaw angle can be obtained by solving the above optimization problem using a heuristic algorithm, such as a particle swarm optimization algorithm.

[0154] In step S2 of this embodiment, the wake coupling relationship of the wind farm under the current wind direction is calculated by the optimal yaw angle to construct the wake map, which includes: constructing the wind speed attenuation rate of each wind turbine in the wind farm calculated when solving the optimal yaw angle into the following formula: The matrix:

[0155] ,

[0156] Any element in row i and column j of the matrix Representative fan i For fans j The wind speed attenuation rate; the wind turbines are respectively regarded as weighted undirected graphs The wind speed attenuation rate between the nodes and wind turbines in the graph is a weighted undirected graph Construct a weighted undirected graph using the edges in :

[0157] ,

[0158] in, Represents a weighted undirected graph If the fan i and fan j If there is wake coupling between ,otherwise = 0; n is the number of fans.

[0159] like Figure 3 As shown, in step S3 of this embodiment, the wake map is clustered based on the spectral clustering method, and the wind farm is divided into multiple sub-areas including:

[0160] S301, define the fan according to the following formula i The degree of the corresponding node is a weighted undirected graph The sum of all edges of :

[0161] ,

[0162] in, n Represents the number of fans, Represents a weighted undirected graph edge;

[0163] According to all fans i The degree of the corresponding node Defining the degree matrix :

[0164] ;

[0165] S302, calculate the Laplace matrix according to the following formula :

[0166] ,

[0167] in, Represents a weighted undirected graph; the normalized Laplace matrix is ​​calculated according to the following formula :

[0168] ,

[0169] S303, calculate the normalized Laplace matrix The eigenvalue of e , and calculate e The second-order difference of :

[0170] ,

[0171] in, Representative i +2 eigenvalues ​​and i +1 eigenvalue difference, Representative i +1 eigenvalue and i The difference of eigenvalues, Representative i +2 eigenvalues, Representative i eigenvalues, Representative i +1 eigenvalue;

[0172] S304, design condition threshold, from i =3, when the eigenvalue The number of clusters is determined when the following conditions are met K = i :

[0173] ,

[0174] in, Representative i -1 second-order difference, is an adjustable parameter, Representative j A second-order difference, Representative j -1 second-order difference;

[0175] S305, based on the clustering idea, establish the minimum segmentation problem:

[0176] ,

[0177] in, Represents segmentation, ~ Respectively represent belonging to subset 1~ K A collection of nodes, represent and The split weight between Represents a subset The set of points, represent The complement of ; choose to solve this minimum segmentation problem through the "normalized segmentation" method:

[0178] ,

[0179] in, stands for the "normalized segmentation" method, represent The weights and edges of , through the spectral clustering method, the above minimum partitioning problem is transformed into the following form:

[0180] ,

[0181] in, Represents the matrix to be solved, the superscript represents transpose, represent n -dimensional identity matrix, represents the normalized Laplace matrix;

[0182] S306, by solving the normalized Laplace matrix Before K eigenvectors to obtain the matrix to be solved , and then based on the k-means clustering method to finally obtain the clustering results of wind turbine nodes in the wind farm.

[0183] In step S4 of this embodiment, the short-time-scale wind farm power dynamic optimization problem with optimization objectives including maximizing the total wind farm power and minimizing the thrust fluctuation of wind turbines is constructed for each sub-area, including:

[0184] S401, determine the fan according to the following formula i The thrust :

[0185] ,

[0186] in, For fans i The axial induction factor, For fans i The yaw angle, For fans ithe incoming wind speed, for air density, for the impeller swept area, for the fan i thrust coefficient, for the thrust loss coefficient of the fan yaw state, the thrust loss coefficient is a constant parameter, which can be valued according to actual needs, for example, the value is 2 in the embodiment;

[0187] S402, considering that the thrust on the fan changes with the wind speed, the fluctuation of the thrust on the fan is suppressed while the total power of the wind farm is improved, and the following short-time-scale wind farm power dynamic optimization problem is constructed:

[0188] ,

[0189] wherein, represent the weight of the optimization target, represent the number of fans, represent the optimal yaw angle of the fan i , represent the power output of the fan i under the optimal yaw angle, represent the thrust on the fan i under the optimal yaw angle; represent the initial value of the thrust on the fan i ;

[0190] S403, according to the number of sub-regions in which the wind farm has been divided in the spectral clustering result, the optimization problem of each sub-region can be split to be calculated separately, and the short-time-scale wind farm power dynamic optimization problem is rewritten in the following sub-region form: N c

[0191] ,

[0192] s.t. ,

[0193] wherein, represent the axial induction factor of the fan belonging to the k th subset, is the axial induction factor of the fan k in the i th subset, represent the number of fans in the k th subset, represent the upper and lower limits of the axial induction factor of the fan i , represent the axial induction factor of the fan i ​In this way, the original optimization problem is broken down into multiple smaller sub-problems that can be solved in parallel, significantly reducing the solution time and making it suitable for real-time dynamic control.

[0194] In step S4 of this embodiment, when solving the short-time-scale wind farm power dynamic optimization problem to obtain the optimal control quantity of each unit, the optimal control quantity includes the axial induction factor of each wind turbine, that is, the axial induction factor of each wind turbine i Axial induction factor .

[0195] Figure 4 、 Figure 5 and Figure 6 The results of wind farm zoning using the method proposed in this embodiment (referred to as the “proposed method”) at different wind directions (0°, 10°, and 20° wind directions) are shown respectively. Figure 4 、 Figure 5 and Figure 6 In sub-figure (a), the short solid lines of the same color represent wind turbines grouped into the same zone. In the corresponding sub-figure (b), the black dashed boxes represent wind turbines grouped into the same zone. It can be seen that changes in wind direction can lead to different wake distributions within the wind farm, thus changing the zoning results. Compared to centralized optimization methods that simultaneously optimize all 36 wind turbines in the farm, the proposed method optimizes wind turbines in different, independent sub-zones, significantly reducing computation time.

[0196] Figure 7 and Figure 8 The total active power of the wind farm under the method proposed in this embodiment and other existing control methods is shown for wind directions of 0° and 20°, respectively. Among the existing control methods, the "centralized method" optimizes all 36 wind turbines simultaneously, the "static control method" uses the optimal value obtained from a long-term steady-state optimization problem to maximize wind farm power as the control variable, and the "greedy control" method maintains the maximum axial induction factor for all wind turbines in the farm without yaw. It can be seen that under different wind directions, the method proposed in this embodiment, the centralized method, and the static control method all achieve higher total active power than the greedy control method. Because the optimization objective is only to maximize wind farm power, the total active power of the wind farm under the static control method is slightly higher than that of the method proposed in this embodiment and the centralized method. The method proposed in this embodiment and the centralized method add the objective of minimizing wind turbine thrust fluctuation, resulting in slightly lower total active power. Furthermore, because the method proposed in this embodiment partitions the wind farm, potential wake coupling between wind turbines in different sub-areas is ignored, resulting in slightly lower total active power compared to the centralized method.

[0197] Figure 9 and Figure 10The total thrust of the wind farm under different control methods is shown at 0° and 20° wind directions respectively; Figure 11 The thrust of wind turbine #1 and wind turbine #36 (all 36 wind turbines are recorded as wind turbine #1 to wind turbine #36) at 0° wind speed is shown. Figures 9 to 11 As can be seen, due to the addition of the objective of minimizing wind turbine thrust fluctuations, both the total wind farm thrust and the thrust of individual wind turbines are smoother and less volatile under the control of the method proposed in this embodiment and the centralized method. This helps to extend the service life of wind turbines and, ultimately, the wind farm, while reducing operation and maintenance costs. Furthermore, the centralized method demonstrates better thrust fluctuation suppression, which is also due to the method proposed in this embodiment ignoring the potential wake coupling between wind turbines in different sub-areas.

[0198] Figure 12 and Figure 13 The time required for each optimization calculation of the method proposed in this embodiment and the centralized method under different wind directions is demonstrated, as well as the effect of the method proposed in this embodiment on saving calculation time compared to the centralized method. According to the previous analysis, although the control performance of the method proposed in this embodiment is not as good as that of the centralized method, the calculation time of the method proposed in this embodiment under all wind directions is less than 0.5 seconds, while the centralized method requires a longer calculation time, and some wind directions can even reach 4 seconds. At a wind direction of 20°, the calculation time required for the method proposed in this embodiment is only 3.42% of that of the centralized method, and at the worst wind direction of 70°, it is only 18.37%, which undoubtedly greatly improves the performance of real-time control.

[0199] In summary, the wind farm power dynamic optimization method based on spectral clustering partitioning in this embodiment includes: constructing a long-time-scale wind farm power maximization steady-state optimization problem based on the engineering wake model according to the average wind speed and wind direction, and solving the optimal yaw angle; calculating the wake coupling relationship of the wind farm under the current wind direction with the optimal yaw angle, and constructing a wake map; clustering the wake map based on the spectral clustering method, and dividing the wind farm into multiple sub-areas; constructing a short-time-scale wind farm power dynamic optimization problem for each sub-area, with the optimization objectives including maximizing the total power of the wind farm and minimizing the thrust fluctuation of the wind turbines, and obtaining the optimal control quantity of each unit. This embodiment can dynamically improve the overall power generation of the wind farm while considering the influence of the wake effect, thereby bringing higher economic benefits to the operator, and has a calculation speed of seconds to ensure the timeliness of control.

[0200] In addition, this embodiment also provides a wind farm power dynamic optimization system based on spectral clustering and partitioning, comprising a microprocessor and a memory connected to each other, wherein the microprocessor is programmed or configured to execute the wind farm power dynamic optimization method based on spectral clustering and partitioning.

[0201] In addition, this embodiment further provides a computer-readable storage medium, in which a computer program or instruction is stored. The computer program or instruction is programmed or configured to execute the wind farm power dynamic optimization method based on spectral clustering partitioning through a processor.

[0202] In addition, this embodiment further provides a computer program product, including a computer program or instructions, which are programmed or configured to execute the wind farm power dynamic optimization method based on spectral clustering and partitioning through a processor.

[0203] Those skilled in the art should understand that the technical solution provided by the present invention may be in the form of a method, a system, or a computer program product. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-readable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code. The present invention is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present invention. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of processes and / or boxes in the flowchart and / or block diagram, may be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the functions described in the process. Figure 1 a process or multiple processes and / or boxes Figure 1 These computer program instructions can also be stored in a computer-readable memory that can guide a computer or other programmable data processing device to work in a specific way, so that the instructions stored in the computer-readable memory produce a product including the instruction device, which implements the function specified in the process. Figure 1 a process or multiple processes and / or boxes Figure 1 These computer program instructions can also be loaded onto a computer or other programmable data processing device, so that a series of operation steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process in the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0204] The above description is merely a preferred embodiment of the present invention. The scope of protection of the present invention is not limited to the above embodiment. All technical solutions based on the concept of the present invention are within the scope of protection of the present invention. It should be noted that for those skilled in the art, various improvements and modifications that do not depart from the principles of the present invention should also be considered within the scope of protection of the present invention.

Claims

1. A wind farm power dynamic optimization method based on spectral clustering partitioning, characterized in that: The steps include: According to the average wind speed and direction, a long-term wind farm power maximization steady-state optimization problem is constructed based on the engineering wake model and the optimal yaw angle is solved; The wake coupling relationship of the wind farm under the current wind direction is calculated using the optimal yaw angle to construct a wake map; The wake map is clustered based on the spectral clustering method, and the wind farm is divided into multiple sub-areas; A short-time-scale wind farm power dynamic optimization problem is constructed for each sub-area, with optimization objectives including maximizing the total wind farm power and minimizing the thrust fluctuation of the wind turbines. The short-time-scale wind farm power dynamic optimization problem is solved to obtain the optimal control quantity of each unit. The short-time-scale wind farm power dynamic optimization problem is constructed for each sub-area, with optimization objectives including maximizing the total wind farm power and minimizing the thrust fluctuation of the wind turbines, including: S401, determine the fan according to the following formula i The thrust : ， in, For fans i The axial induction factor, For fans i The yaw angle, For fans i The inflow wind speed, is the air density, is the impeller swept area, For fans i The thrust coefficient, is the thrust loss coefficient of the wind turbine in yaw state; S402: Considering that the thrust on the wind turbine changes with the wind speed in a dynamic state, the fluctuation of the thrust on the wind turbine is suppressed while increasing the total power of the wind farm. The following short-time-scale dynamic optimization problem of wind farm power is constructed: ， in, represents the weight of the optimization objective, Represents the number of fans, Representative fan i The optimal yaw angle, Representative fan i Power output at the optimal yaw angle, Representative fan i The thrust at the optimal yaw angle; Representative fan i The initial value of the thrust; S403, based on the spectral clustering results, the wind farm has been divided N c The optimization problem of each sub-region is split into separate sub-regions for separate calculations, and the short-time-scale wind farm power dynamic optimization problem is rewritten into the following sub-region form: ， s.t. , in, Representatives belong to k The axial induction factor of the fans in the subset, For the k Sub-centralized fans i The axial induction factor, Representative k The number of fans in the subset, Representative fan i The upper and lower limits of the axial induction factor, Representative fan i The upper and lower limits of active output.

2. The method for dynamic optimization of wind farm power based on spectral clustering partitioning according to claim 1, characterized in that: The construction of the engineering wake model includes: S101, for wind turbines i For fans n The wake effect generated by the wind turbine i The yaw action will cause the center of its wake to deflect, which is defined as yaw deflection. The wake deflection distance of the yaw deflection is determined. for: ， in, Representative fan i The initial deflection angle, represents the impeller radius of the fan, represents the deflection expansion coefficient, Representative fan i For fans n The deflection constant that produces the wake is calculated as: ， ， ， in, Representative fan i With fan n The geographical distance between them along the wind direction, Representative fan i The yaw angle, Representative fan i The thrust coefficient, Representative fan i Axial induction factor; S102, define the deflection of the wake center caused by the impeller rotation itself as rotational deflection, calculate the fan i For fans n The resulting rotational deflection for: ， in, and is a constant; S103, calculate the fan according to the following formula i For fans n The coordinate of the center of the wake effect on the y-axis : ， in, Representative fan i Geographical location y axis coordinates; S104, the wake effect will gradually expand as the propagation process, the wind turbine For fans The resulting wake expansion radius : ， in, represents the maximum value, represents the wake expansion coefficient, Represents the wake area q The scaling constant, q represents the wake region, q =1, 2, 3 represent the near wake region, far wake region, and mixing region, respectively; S105, calculate the fan according to the following formula For fans The resulting wake wind speed : ， ， ， ， ， ， ， ， in, Representative fan i The inflow wind speed, Representative fan i For fans n The wind speed attenuation rate, represents the wake attenuation coefficient, Representative fan i For fans n The wake area caused q The wake overlap area, is the impeller swept area; represents the wake attenuation correction parameter, 、 and are all constants; Representative fan i For fans n The wake area caused q The expansion radius; and Represents fans i For fans n The wake area caused q The two angles used to calculate the area of ​​the wake overlap area are: Represents the coordinates of fan n in the y-axis direction and the fan i For fans n The distance between the centers of the wake induced by Representative fan n The y-axis coordinate of .

3. The method for dynamic optimization of wind farm power based on spectral clustering partitioning according to claim 2, characterized in that: The method of constructing a long-term wind farm power maximization steady-state optimization problem based on an engineering wake model according to the average wind speed and wind direction and solving the optimal yaw angle includes: S201, calculate the fan according to the following formula i Power output : ， ， in, Representative fan i The axial induction factor, Representative fan i The yaw angle, Representative fan i The inflow wind speed, represents the air density, represents the impeller swept area, Representative fan i The power coefficient, represents the generator efficiency, p represents the power loss coefficient caused by yaw; S202, considering the wake effect inside the wind farm, the inflow wind speed of each wind turbine is related to the free wind speed, the axial induction factor and the yaw angle of the upstream wind turbine, and the total power output function of the wind farm is constructed. : ， in, is the inflow wind speed of the wind farm, represents the set of all fan axial induction factors, represents the set of all wind turbine yaw angles, Represents the number of fans, represents the power of fan i; S203, construct the long-time-scale wind farm power maximization steady-state optimization problem shown in the following equation: ， ， in, Representative fan The upper and lower limits of the axial induction factor, Representative fan The upper and lower limits of the yaw angle, Representative fan The upper and lower limits of active output; S204, using a heuristic algorithm to solve the long-time-scale steady-state optimization problem of wind farm power maximization to obtain the optimal yaw angle.

4. The method for dynamic optimization of wind farm power based on spectral clustering partitioning according to claim 1, characterized in that: The method of calculating the wake coupling relationship of the wind farm under the current wind direction by using the optimal yaw angle to construct the wake diagram includes: constructing the wind speed attenuation rate of each wind turbine in the wind farm calculated when solving the optimal yaw angle into the following formula: The matrix: ， Any element in row i and column j of the matrix Representative fan i For fans j The wind speed attenuation rate; the wind turbines are respectively regarded as weighted undirected graphs The wind speed attenuation rate between the nodes and wind turbines in the graph is a weighted undirected graph Construct a weighted undirected graph using the edges in : ， in, Represents a weighted undirected graph If the fan i and fan j If there is wake coupling between ,otherwise = 0; n is the number of fans.

5. The method for dynamic optimization of wind farm power based on spectral clustering partitioning according to claim 4, characterized in that: The wake map is clustered based on the spectral clustering method to divide the wind farm into multiple sub-areas, including: S301, define the fan according to the following formula i The degree of the corresponding node is a weighted undirected graph The sum of all edges of : ， in, n Represents the number of fans, Represents a weighted undirected graph edge; According to all fans i The degree of the corresponding node Defining the degree matrix : ； S302, calculate the Laplace matrix according to the following formula : ， in, Represents a weighted undirected graph; the normalized Laplace matrix is ​​calculated according to the following formula : ， S303, calculate the normalized Laplace matrix The eigenvalue of e , and calculate e The second-order difference of : ， in, Representative i +2 eigenvalues ​​and i +1 eigenvalue difference, Representative i +1 eigenvalue and i The difference of eigenvalues, Representative i +2 eigenvalues, Representative i eigenvalues, Representative i +1 eigenvalue; S304, design condition threshold, from i =3, when the eigenvalue The number of clusters is determined when the following conditions are met K = i : ， in, Representative i -1 second-order difference, is an adjustable parameter, Representative j A second-order difference, Representative j -1 second-order difference; S305, based on the clustering idea, establish the minimum segmentation problem: ， in, Represents segmentation, ~ Respectively represent belonging to subset 1~ K A collection of nodes, represent and The split weight between Represents a subset The set of points, represent The complement of ; choose to solve this minimum segmentation problem by the "normalized segmentation" method: ， in, stands for the "normalized segmentation" method, represent The weights and edges of , through the spectral clustering method, the above minimum partitioning problem is transformed into the following form: ， in, Represents the matrix to be solved, the superscript represents transpose, represent n -dimensional identity matrix, represents the normalized Laplace matrix; S306, by solving the normalized Laplace matrix Before K eigenvectors to obtain the matrix to be solved , and then based on the k-means clustering method to finally obtain the clustering results of wind turbine nodes in the wind farm.

6. The method for dynamic optimization of wind farm power based on spectral clustering partitioning according to claim 1, characterized in that: When solving the short-time-scale wind farm power dynamic optimization problem to obtain the optimal control quantity of each unit, the optimal control quantity includes the axial induction factor of each wind turbine.

7. A wind farm power dynamic optimization system based on spectral clustering partitioning, comprising a microprocessor and a memory connected to each other, characterized in that: The microprocessor is programmed or configured to execute the wind farm power dynamic optimization method based on spectral clustering and partitioning as recited in any one of claims 1 to 6.

8. A computer-readable storage medium having a computer program or instruction stored therein, characterized in that: The computer program or instruction is programmed or configured to execute the wind farm power dynamic optimization method based on spectral clustering partitioning according to any one of claims 1 to 6 through a processor.

9. A computer program product comprising a computer program or instructions, characterized in that The computer program or instruction is programmed or configured to execute the wind farm power dynamic optimization method based on spectral clustering partitioning according to any one of claims 1 to 6 through a processor.

Images