Offshore wind power plant group wake flow regulation and control method and system based on dynamic grouping and sequential optimization

Abstract

The invention relates to the technical field of wind power control, and discloses an offshore wind power plant group wake flow regulation and control method and system based on dynamic grouping and sequential optimization, and the method comprises the steps: building a wake flow loss analysis model based on a Marsen wake flow model; performing dynamic grouping on all the wind turbine generators in the wind power plant group based on the speed loss of the target wind turbine generator output by the wake flow loss analysis model to obtain a plurality of subgroups; constructing a multi-objective optimization function; sequential optimization is carried out in each subgroup, yaw angles are optimized one by one according to a windward sequence, and an optimal yaw angle combination enabling a multi-objective optimization function to be minimum is solved; the dynamic grouping result and the yaw instruction are updated once every preset control period; and taking the solved optimal yaw angle combination of each wind turbine generator corresponding to the next control period as a control instruction to be issued. According to the real-time wake flow regulation and control method, the optimization effect can be guaranteed, the calculation complexity can be remarkably reduced, and the method is suitable for a large-scale offshore wind power plant group.

Description

Technical Field

[0001] This invention belongs to the field of wind power control technology, and relates to a method and system for controlling the wake of offshore wind farm clusters based on dynamic grouping and sequential optimization. Background Technology

[0002] As offshore wind power develops towards clustering and large-scale operations, the wake interference problem between turbines within a wind farm cluster is becoming increasingly prominent. The wake of upstream turbines can reduce the inflow wind speed of downstream turbines, thus significantly reducing the overall power generation efficiency of the wind farm cluster. Studies have shown that by using active yaw control to deflect the wake of upstream turbines to avoid downstream turbines, the total power generation of the wind farm cluster can be effectively improved.

[0003] Applying active yaw wake control to large-scale offshore wind farm clusters faces three major challenges:

[0004] (1) The impact of active yaw wake control on the safety, fatigue and lifespan of wind turbine units.

[0005] Active yaw wake control causes wind turbines to deviate from the direction of the incoming wind, which may threaten the safety of wind turbines, increase fatigue, and reduce their lifespan.

[0006] (2) The wake calculation efficiency of active yaw wake control in large-scale wind farm clusters is low and the optimization object dimension is large.

[0007] In existing active yaw wake control methods, the target of wake control is usually a regularly arranged wind farm. For wind farm clusters, the wake calculation speed is relatively slow because their scale is larger than that of individual wind farms. Furthermore, because there are many wind turbines in a wind farm cluster, the number of objects to be optimized (the yaw angles of all wind turbines) is large, i.e., the dimensionality of the optimization object is high.

[0008] (3) The high-dimensional, nonlinear, and strongly coupled characteristics of the active yaw wake control optimization problem.

[0009] In the optimization problem of active yaw wake control, the common optimization objective is to maximize the power generation of wind farms or wind farm clusters. The calculation of power generation of wind farms or wind farm clusters is characterized by high dimensionality, nonlinearity, and strong coupling, making it difficult to find the global optimum. This makes it difficult for traditional optimization algorithms to find the global optimum in the time priority.

[0010] Existing wind farm-level wake control methods, such as global optimization based on the FLORIS model, are often impractical when scaled up to the scale of wind farm clusters due to computational complexity and time constraints. Summary of the Invention

[0011] The purpose of this invention is to provide a wake control method and system for offshore wind farm clusters based on dynamic clustering and sequential optimization. By dynamically clustering, the problem size is reduced, and by sequential optimization, the decision space is simplified. Thus, while ensuring the optimization effect, the yaw command of all units in the wind farm cluster can be solved quickly within a control cycle. This method is suitable for real-time wake control of large-scale offshore wind farm clusters.

[0012] To solve the above-mentioned technical problems, the present invention is implemented using the following technical solution.

[0013] In a first aspect, this invention proposes a method for controlling the wake of offshore wind farm clusters based on dynamic grouping and sequential optimization, comprising the following steps:

[0014] Based on the Larsen wake model, a wake loss analysis model is constructed; based on the velocity loss of the target wind turbine output by the wake loss analysis model, all wind turbines in the wind farm cluster are dynamically grouped to obtain several subgroups.

[0015] Construct a multi-objective optimization function that considers the power generation of the wind farm cluster, the fatigue loss of the yaw system, and the fatigue and life of the wind turbine units;

[0016] Within each subgroup, sequential optimization is performed, optimizing the yaw angle of each aircraft in the order of windward movement, and solving for the optimal yaw angle combination that minimizes the multi-objective optimization function.

[0017] The dynamic clustering results and yaw commands are updated once every preset control cycle;

[0018] The optimal yaw angle combination of each wind turbine unit obtained from the solution and corresponding to the next control cycle is issued as a control command to achieve the regulation of the wake of the offshore wind farm group.

[0019] In conjunction with the first aspect, the aforementioned construction of an analysis model for wake loss based on the Larsen wake model includes:

[0020] The speed loss of the downstream target wind turbine is the axial distance. and radial distance The function, with the following expression:

[0021] ;

[0022] ;

[0023] ;

[0024] ;

[0025] ;

[0026] ;

[0027] in, Indicates the area of ​​the wind turbine; Indicates the wake radius; Indicates the 0-dimensional mixing length of Prandtl; Indicates the thrust coefficient of the wind turbine; Indicates the diameter of the wind turbine; Represents an approximate number; This represents the wake radius at a point where the wake length is 9.5 times the rotor diameter. An empirical formula for representing the wake radius at a wake length of 9.5 times the rotor diameter; This indicates the intensity of atmospheric turbulence at the hub height; This indicates the axial distance downstream of the wind turbine. Radial distance is The wind speed at the location and within the wake region; Indicates the free-flowing wind speed; This indicates the hub height of the wind turbine.

[0028] In conjunction with the first aspect, further, based on the velocity loss of the target wind turbine output by the wake loss analysis model, all wind turbines within the wind farm cluster are dynamically grouped to obtain several subgroups, including:

[0029] Calculate upstream wind turbines For the target wind turbine located downstream Speed ​​loss caused Preset speed loss threshold ,Will The target wind turbines are grouped into the same subgroup. Velocity loss refers to: assuming only upstream wind turbines... and target wind turbine units Existence, upstream wind turbines The wake of the target wind turbine The percentage of wind speed at the center of the wheel hub relative to the ambient wind speed.

[0030] In conjunction with the first aspect, the dynamic clustering further employs parallel computing, with a sub-cluster size of ≤60 and a maximum sub-cluster size of ≤6 wind turbine units, ensuring that the optimization time for a single sub-cluster is <30 seconds.

[0031] In conjunction with the first aspect, the specific steps of the dynamic clustering are as follows:

[0032] Step S11: Iterate for each prediction time step of the model predictive control;

[0033] Step S12: Initialize the clustering results of the current model's predictive control at prediction time step t. , as in equation (1.1);

[0034] (1.1);

[0035] in, Indicates the number of clusters;

[0036] Step S13: Cycle through each wind turbine unit;

[0037] in, Indicates the serial number of the target wind turbine unit, Indicates the target number of wind turbine units;

[0038] Step S14: Under the condition of no yaw, calculate all upstream wind turbine units. For the target wind turbine The resulting speed loss; setting up measures for upstream wind turbines With the target wind turbine speed loss The velocity loss threshold κ, if If the value is greater than or equal to κ, then the upstream wind turbine is considered to be... With the target wind turbine There is value in active yaw wake control; all The set is j, as shown in equation (1.2);

[0039] (1.2);

[0040] In the formula, It is the ambient wind speed; Indicates the prediction time step upstream wind turbines The wake of the target wind turbine Wind speed at the center of the wheel hub; This indicates the predicted time step. Environmental wind speed;

[0041] Step S15, all upstream wind turbine units satisfying equation (1.2) Included In each group, as in equation (1.3);

[0042] (1.3);

[0043] in, Indicates the prediction time step This includes upstream wind turbines. Groups; Indicates the first The groups at the predicted time step The state; Indicates the index of the group; Indicates the prediction time step The Chinese number is Groups; Represents the set of group IDs one of the;

[0044] Step S16, merge Join a group and add the target wind turbine. , as in equation (1.4);

[0045] (1.4);

[0046] in, Represents a collection of groups;

[0047] Step S17, return to step S13, until the cycle of all wind turbine units is completed;

[0048] Step S18, return to step S11, until the loop for all prediction time steps is completed;

[0049] Step S19: Perform dynamic clustering and initialize the dynamic clustering results. , as in equation (1.5);

[0050] (1.5);

[0051] In the formula, It is the number of prediction time steps in model predictive control; yes The number of groups; express The Middle Groups They represent the first A group of predicted time steps;

[0052] Step S110: Determine whether there are any different groups that contain the same elements, as shown in equation (1.6).

[0053] (1.6);

[0054] in, and This represents any two distinct groups;

[0055] Step S111: If the condition judgment result of step S110 is yes, then the condition is satisfied. The set of groups, as shown in equation (1.7).

[0056] (1.7);

[0057] in, Indicates a group that meets the criteria; Indicates number For groups; Each represents the set of natural numbers;

[0058] Step S112, will Take the union of all elements of the original group to form a new group; delete the old group. A group, as shown in equation (1.8); return to step S110;

[0059] (1.8);

[0060] Step S113: If the condition judgment result of step S110 is negative, the final clustering result G is obtained, and the process ends.

[0061] In conjunction with the first aspect, further, the construction of a multi-objective optimization function considering the power generation of the wind farm cluster, the fatigue loss of the yaw system, and the fatigue and lifespan of the wind turbine units includes:

[0062] The expression for the multi-objective optimization function is:

[0063] ;

[0064] ;

[0065] ;

[0066] ;

[0067] In the formula, It is a multi-objective optimization function for active yaw wake control; yes The term considering power generation capacity is the power generation capacity enhancement term; yes The term that considers fatigue losses in the yaw system is the yaw loss term; yes The term considering wind turbine fatigue and lifespan, namely the power fluctuation term; , , These are the weighting coefficients for the power generation improvement item, the yaw loss item, and the power fluctuation item, respectively. This is the number of prediction time steps for MPC; It is a prediction time step In the matrix, the yaw angles of all wind turbine units; It is a zero matrix; The target wind turbine Rated power; It's an angle. and angle The angle between them; Indicates the target wind turbine unit Method for calculating wind speed at the center of the wheel hub Indicates the target wind turbine unit The method for calculating power generation capacity; Indicates the prediction time step The original angle; Indicates the prediction time step Target wind turbine Yaw angle; The predicted time step is Target wind turbine The yaw angle; RMSE(a) is the standard deviation of set a; It is a matrix representing the yaw angles of all wind turbines within the subgroup across all prediction time steps; It is a prediction time step At that time, the target wind turbine Yaw angle, Indicates the current state. Indicates the prediction time step is All states; The time step for prediction is At that time, the target wind turbine The yaw angle is 0. Indicates the current state. Indicates the prediction time step is All states; Indicates the target number of wind turbine units.

[0068] In conjunction with the first aspect, further, the sequential optimization within each subgroup, optimizing the yaw angle aircraft by aircraft in the order of windward movement, and solving for the optimal yaw angle combination that minimizes the multi-objective optimization function, includes:

[0069] Step S31: Set the number of rounds for sequential optimization and the number of yaw angles per round, as shown in equation (3.1):

[0070] (3.1);

[0071] in, A matrix representing the number of rounds and the number of yaw angles per round; This indicates the number of yaw angles for the current round; Indicates the current optimization round number;

[0072] Step S32: Repeat the process for each round of sequence optimization;

[0073] Step S33: Based on the optimal solution of the yaw angle in the previous round of sequential optimization and the set number of yaw angles for the current round, To form a yaw angle test sequence As shown in equation (3.2); the median of the yaw angle test sequence is the optimal solution of the yaw angle in the previous round of sequential optimization, and there are a total of The number of numbers, where the difference between each number is equal and is equal to the difference between each number in the previous round divided by the number in this round;

[0074] (3.2);

[0075] in, Indicates the first One yaw angle;

[0076] Step S34: Cycle according to the windward sequence of all wind turbines in the subgroup;

[0077] Step S35: Iterate through each prediction time step controlled by the model prediction framework;

[0078] Step S36: In all model prediction framework control prediction time steps, all wind turbines within the subgroup are arranged in windward order. Predicting time step, the first step A wind turbine unit facing the wind is represented as , as in equation (3.3);

[0079] (3.3);

[0080] Step S37: Calculate the current prediction time step, the number of sequential optimization rounds, and the... The target wind turbine is an upstream wind turbine. The power at that time is given by equation (3.4), where the target wind turbine unit The yaw angle is taken from each value in the yaw angle test sequence;

[0081] (3.4);

[0082] in, Indicates the current prediction time step Optimize the number of rounds in sequence Target wind turbine Power at that time; Indicates the current prediction time step Optimize the number of rounds in sequence Target wind turbine The yaw angle taken at that time; Indicates the target wind turbine unit Method for calculating wind speed at the center of a wheel hub; Indicates the current prediction time step At that time, the target wind turbine Yaw angle; Indicates the current order optimization round number as Yaw angle sequence at time; Indicates the current prediction time step At that time, the first Yaw angle of each wind turbine unit; This indicates the method for calculating power generation. Indicates the target number of wind turbine units.

[0083] Step S38, based on simplifying assumption 1; the simplifying assumption 1 is that the yaw angle of the downstream wind turbine is independent of the optimal solution of the yaw angle of the upstream wind turbine;

[0084] Step S39, return to step S35, until the loop for all prediction time steps is completed;

[0085] In step S310, at each time point, only the yaw angle of one wind turbine is adjusted. The yaw angles of the wind turbines adjusted at different times are cross-combined to calculate the objective function, which is shown in equation (3.5).

[0086] (3.5);

[0087] in, express The number of each prediction time step; This represents the loss value under the current yaw angle combination; Indicates the first The typhoon turbine unit in the Power generation at each predicted time step; This represents the matrix indicating yaw angle and power; This represents the objective function value under the yaw angle sequence;

[0088] Step S311, return to step S34, until the cycle of all wind turbine units is completed;

[0089] Step S312: Find the combination of yaw angles that minimizes the value of the multi-objective optimization function. This combination is called the optimal solution of yaw angles under the current sequential optimization round number, as shown in equation (3.6).

[0090] (3.6);

[0091] in, This represents the optimal yaw angle combination under the current optimization round number;

[0092] Step S313: If the current sequence optimization round is not the final sequence optimization round, return to step S32 until the loop of all sequence optimization rounds is completed;

[0093] Step S314: If the current sequence optimization wheel is the final sequence optimization wheel, output the optimal yaw angle solution and end.

[0094] In conjunction with the first aspect, the sequential optimization method further comprises multi-round optimization, including: in the first round of optimization, generating a candidate yaw angle test sequence with coarse resolution around 0 degrees for each wind turbine; in subsequent rounds of optimization, using the optimal yaw angle obtained in the previous round as the median, generating a candidate yaw angle test sequence with finer resolution; in each round, following the windward order from upstream to downstream, fixing the optimized yaw angle of the upstream wind turbine, and sequentially selecting the optimal yaw angle for the downstream wind turbine.

[0095] Secondly, this invention proposes a wake control system for offshore wind farm clusters based on dynamic clustering and sequence optimization, used to implement the aforementioned wake control method for offshore wind farm clusters based on dynamic clustering and sequence optimization, comprising:

[0096] The dynamic clustering module is configured to build an analysis model of wake loss based on the Larsen wake model; based on the velocity loss of the target wind turbine output by the analysis model of wake loss, it dynamically clusters all wind turbines in the wind farm cluster to obtain several sub-clusters.

[0097] The multi-objective optimization function construction module is configured to construct a multi-objective optimization function that considers the power generation of the wind farm cluster, the fatigue loss of the yaw system, and the fatigue and life of the wind turbine.

[0098] The sequential optimization module is configured to perform sequential optimization within each subgroup, optimizing the yaw angle of each aircraft in the order of windward movement, and solving for the optimal combination of yaw angles that minimizes the multi-objective optimization function.

[0099] The rolling execution module is configured to update the dynamic clustering results and yaw commands once every preset control cycle.

[0100] The command issuing module is configured to issue the optimal yaw angle combination of each wind turbine unit obtained by solving the problem for the next control cycle as a control command to achieve the regulation of the wake of the offshore wind farm group.

[0101] Thirdly, the present invention proposes a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the steps of the above-mentioned method for controlling the wake of offshore wind farm clusters based on dynamic clustering and sequential optimization.

[0102] Fourthly, the present invention provides a computer device comprising:

[0103] Memory, used to store computer programs;

[0104] A processor is used to execute the computer program to implement the steps of the above-described method for controlling the wake of offshore wind farm clusters based on dynamic clustering and sequential optimization.

[0105] Fifthly, the present invention proposes a computer program product, including a computer program that, when executed by a processor, implements the steps of the above-described method for controlling the wake of offshore wind farm clusters based on dynamic clustering and sequential optimization.

[0106] This invention proposes a two-layer architecture of "dynamic clustering-sequential optimization": 1. Dynamic clustering layer: Based on the real-time changes in wake interference intensity and wind direction / speed, the wind farm cluster is dynamically divided into several independent control sub-clusters, reducing the optimization dimensionality; 2. Sequential optimization layer: Within each sub-cluster, a "windward sequence-sequence search" strategy is adopted to optimize the yaw angle on a turbine-by-turbine basis, taking into account power generation improvement, yaw system fatigue, and power fluctuation suppression. This invention achieves rapid solution within a 15-minute control cycle (time <3 minutes) through a multi-objective model predictive control (MPC) framework, which is more than 20 times faster than traditional global optimization methods. Actual calculations show that the annualized revenue of a single wind farm cluster is increased by approximately RMB 4500 / MW, wake loss is reduced by 0.65%, and it can be extended to ultra-large offshore wind power bases. It solves the problems of high computational complexity, poor real-time performance, and multi-objective conflicts in wake control of large-scale offshore wind farm clusters.

[0107] This invention is applicable to ultra-large offshore wind power bases (installed capacity ≥ 800MW) and can be extended to deep-sea floating wind farm clusters.

[0108] Compared with the prior art, the beneficial effects achieved by the present invention are as follows:

[0109] (1) This invention decomposes the ultra-large-scale optimization problem into multiple small problems that can be computed in parallel by dynamic clustering, and reduces the decision space complexity from exponential to linear by sequential optimization, so that the entire optimization process can be completed within a single control cycle (15 minutes), which meets the real-time requirements of engineering and has high efficiency.

[0110] (2) The sequence optimization method of the present invention follows the physical sequence of wind flow conduction, and can find a high-quality solution that is close to the global optimum, effectively improving the total power generation of the wind farm group, and is effective.

[0111] (3) This invention achieves a balance between economy and safety by using a multi-objective optimization function to pursue power improvement while taking into account the operating life and stability of the unit, thus possessing comprehensiveness.

[0112] (4) This invention is particularly applicable to large-scale offshore wind farms with slow wake recovery and significant wake effects, and provides a feasible technical solution for solving the problem of their collaborative optimization control. Attached Figure Description

[0113] Figure 1 This is a flowchart illustrating the wake control method for offshore wind farm clusters in Embodiment 1 of the present invention.

[0114] Figure 2 This is a flowchart illustrating the dynamic clustering method in Embodiment 1 of the present invention;

[0115] Figure 3 This is a flowchart illustrating the sequence optimization method in Embodiment 1 of the present invention;

[0116] Figure 4 This is a schematic diagram showing the optimized yaw angle of the turbines in an offshore wind farm group under specific operating conditions after adopting the wake control method of the present invention.

[0117] Figure 5 Velocity cloud map of offshore wind farm clusters for applying the yaw angle optimized by this invention;

[0118] Figure 6 Turbulence intensity cloud map of offshore wind farm clusters, optimized by the present invention for applying the yaw angle. Detailed Implementation

[0119] The technical solution of the present invention will be described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the embodiments of the present invention and the specific features in the embodiments are detailed descriptions of the technical solution of the present invention, rather than limitations thereof. In the absence of conflict, the embodiments of the present invention and the technical features in the embodiments can be combined with each other.

[0120] The term "and / or" simply describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, or B alone. Additionally, the character " / " generally indicates that the preceding and following related objects have an "or" relationship.

[0121] Example 1

[0122] like Figure 1 As shown, the offshore wind farm wake control method based on dynamic clustering and sequential optimization proposed in this embodiment includes the following steps:

[0123] Step S1: Based on the Larsen wake model, construct and analyze the wake loss model; based on the velocity loss of the target wind turbine output by the wake loss model, dynamically group all wind turbines in the wind farm cluster based on the output velocity loss of the target wind turbine, and calculate the upstream wind turbines. For the target wind turbine located downstream Speed ​​loss caused Preset speed loss threshold ,Will The target wind turbines are classified into the same subgroup (that is, if the wake velocity loss caused by an upstream wind turbine to the downstream target wind turbine exceeds the preset velocity loss threshold, it is considered that there is strong wake interference between the upstream wind turbine and the downstream target wind turbine, and these target wind turbines should belong to the same subgroup). The target wind turbine units are grouped into different subgroups;

[0124] Step S2: Construct a multi-objective optimization function that comprehensively considers the power generation of the wind farm cluster, the fatigue loss of the yaw system, and the fatigue and life of the wind turbine units.

[0125] Step S3: Within each subgroup obtained in step S1, sequential optimization is performed, optimizing the yaw angle for each aircraft in the order of windward movement. A sequential search strategy is used to optimize the discrete yaw angle candidate set. A fast solution is found for the optimal combination of yaw angles that minimizes the multi-objective optimization function; where... Indicates the yaw angle; This represents the discrete interval for the yaw angle search; the yaw angle candidate set is set empirically.

[0126] Step S4: Analyze the closed-loop control framework called by the wake loss model, and execute it in a rolling manner using the model predictive control (MPC) framework. Update the grouping results and yaw commands every 15 minutes, and optimize the time consumption to within 3 minutes.

[0127] Step S5: The optimal yaw angle combination of each wind turbine unit corresponding to the next control cycle, obtained in step S3, is issued as a control command.

[0128] In one specific implementation of this embodiment, in step S1, the Larsen wake model utilizes self-similarity to establish an analytical wake loss model that outputs velocity loss and uses the output velocity loss as the grouping criterion. It assumes that the wake region behind the wind turbine can be described by the turbulent boundary layer equation, and that the fluid is stationary and incompressible, with negligible wind shear. The velocity loss of the downstream target wind turbine is the axial distance. and radial distance (i.e., radial offset) The function of ).

[0129] ;

[0130] ;

[0131] ;

[0132] ;

[0133] ;

[0134] ;

[0135] in, Indicates the area of ​​the wind turbine; Indicates the wake radius; Represents the 0-dimensional mixing length of Prandtl (a constant value with no spatial distribution). Indicates the thrust coefficient of the wind turbine; Indicates the diameter of the wind turbine; Represents an approximate number; This represents the wake radius at a point where the wake length is 9.5 times the rotor diameter. An empirical formula for representing the wake radius at a wake length of 9.5 times the rotor diameter. , These are all empirical values ​​obtained from experimental observations; This indicates the intensity of atmospheric turbulence at the hub height; This indicates the axial distance downstream of the wind turbine. Radial offset is At the location, the wind speed within the wake region; Indicates the free-flowing wind speed; This indicates the hub height of the wind turbine.

[0136] In one specific implementation of this embodiment, in step S1, dynamic clustering involves dividing the wind turbines in the entire wind farm cluster into several independent subgroups for each prediction time step, based on directly obtained wind direction, wind speed, and wind farm cluster layout. If the wake velocity loss caused by an upstream wind turbine to a downstream target wind turbine exceeds a preset velocity loss threshold, it is considered that there is strong wake interference between the upstream wind turbine and the downstream target wind turbine, and these target wind turbines should belong to the same subgroup. Velocity loss threshold. Furthermore, it decreases linearly with increasing environmental turbulence intensity. The dynamic clustering process is as follows: Figure 2 The specific steps are as follows:

[0137] Step S11: Cycle through each prediction time step of Model Predictive Control (MPC).

[0138] Step S12: Initialize the clustering results for the current MPC prediction time step t. As shown in equation (1.1).

[0139] (1.1);

[0140] in, Indicates the number of clusters.

[0141] Step S13: Cycle through each wind turbine unit.

[0142] in, Indicates the serial number of the target wind turbine unit, Indicates the target number of wind turbine units.

[0143] Step S14: Under the condition of no yaw, calculate all upstream wind turbine units. For the target wind turbine The resulting speed loss. Speed ​​loss refers to the loss caused by assuming only upstream wind turbines... and target wind turbine units Existence, upstream wind turbines The wake of the target wind turbine The percentage of wind speed at the hub center relative to the ambient wind speed. This setting is for upstream wind turbines. With the target wind turbine speed loss The velocity loss threshold κ, if If the value is greater than or equal to κ, then the upstream wind turbine is considered to be... With the target wind turbine There is value in active yaw wake control. All The set is j, as shown in equation (1.2).

[0144] (1.2);

[0145] In the formula, It is the ambient wind speed; Indicates the prediction time step upstream wind turbines The wake of the target wind turbine Wind speed at the center of the wheel hub; This indicates the predicted time step. The ambient wind speed.

[0146] Step S15, all upstream wind turbine units satisfying equation (1.2) Included In each group, as shown in equation (1.3).

[0147] (1.3);

[0148] in, Indicates the prediction time step This includes upstream wind turbines. Groups; Indicates the first The groups at the predicted time step The state; Indicates the index of the group; Indicates the prediction time step The Chinese number is Groups; Represents the set of group IDs one of the.

[0149] Step S16, merge Join a group and add the target wind turbine. As shown in equation (1.4).

[0150] (1.4);

[0151] in, This represents a collection of groups.

[0152] Step S17, return to step S13, until the cycle of all wind turbine units is completed.

[0153] Step S18: Return to step S11 until all prediction time steps are completed.

[0154] Step S19: Perform dynamic clustering and initialize the dynamic clustering results. As shown in equation (1.5).

[0155] (1.5);

[0156] In the formula, This is the number of prediction time steps for MPC; yes The number of groups; express The Middle Groups They represent the first A group of predicted time steps.

[0157] Step S110: Determine whether there are any different groups that contain the same elements, as shown in equation (1.6).

[0158] (1.6);

[0159] in, and This represents any two distinct groups.

[0160] Step S111: If the condition judgment result of step S110 is yes, then the condition is satisfied. The set of groups is as shown in equation (1.7).

[0161] (1.7);

[0162] in, Indicates a group that meets the criteria; Indicates number For groups; These represent the sets of natural numbers.

[0163] Step S112, will Take the union of all elements of the original group to form a new group; delete the old group. A group, as shown in equation (1.8). Return to step S110.

[0164] (1.8);

[0165] Step S113: If the condition judgment result of step S110 is negative, the final clustering result G is obtained, and the process ends.

[0166] In one specific implementation of this embodiment, the multi-objective optimization function in step S2 is in the form of:

[0167] ;

[0168] ;

[0169] ;

[0170] ;

[0171] In the formula, It is a multi-objective optimization function for active yaw wake control; yes The term considering power generation capacity is the power generation capacity enhancement term; yes The term that considers fatigue losses in the yaw system is the yaw loss term; yes The term considering wind turbine fatigue and lifespan, namely the power fluctuation term; , , These are the weighting coefficients for the power generation improvement item, the yaw loss item, and the power fluctuation item, respectively. This is the number of prediction time steps for MPC; It is a prediction time step In the matrix, the yaw angles of all wind turbine units; It is a zero matrix; The target wind turbine Rated power; It's an angle. and angle The angle between them; Indicates the target wind turbine unit Method for calculating wind speed at the center of the wheel hub Indicates the target wind turbine unit The method for calculating power generation capacity; Indicates the prediction time step The original angle; Indicates the prediction time step Target wind turbine Yaw angle; The predicted time step is Target wind turbine The yaw angle; RMSE(a) is the standard deviation of set a; It is a matrix representing the yaw angles of all wind turbines within the subgroup across all prediction time steps; It is a prediction time step At that time, the target wind turbine Yaw angle, Indicates the current state. Indicates the prediction time step is All states; The time step for prediction is At that time, the target wind turbine The yaw angle is 0. Indicates the current state. Indicates the prediction time step is All states.

[0172] In one specific implementation of this embodiment, the dynamic grouping in step S1 adopts parallel computing, with the number of subgroups ≤60 and the maximum subgroup size ≤6 units, ensuring that the optimization time of a single subgroup is <30 seconds.

[0173] In one specific implementation of this embodiment, the sequential optimization method in step S3 is a multi-round optimization, including: in the first round of optimization, generating a candidate yaw angle sequence with coarse resolution around 0 degrees for each wind turbine; in subsequent rounds of optimization, using the optimal yaw angle obtained in the previous round as the median, generating a candidate yaw angle sequence with finer resolution; in each round, fixing the optimized yaw angle of the upstream wind turbine according to the windward order from upstream to downstream, and sequentially selecting the optimal yaw angle for the downstream turbine.

[0174] In one specific implementation of this embodiment, in step S3, the resolution of the sequentially optimized yaw angle candidate set is... Decreasing with each optimization round (first round) The final round ), to balance calculation accuracy and speed.

[0175] In this embodiment, a specific implementation method is described, and the flowchart of the sequential optimization method in step S3 is as follows: Figure 3 As shown, the specific steps are as follows:

[0176] Step S31: Set the number of rounds for sequential optimization and the number of yaw angles per round, as shown in equation (3.1).

[0177] (3.1);

[0178] in, A matrix representing the number of rounds and the number of yaw angles per round; This indicates the number of yaw angles for the current round; This indicates the current optimization round number.

[0179] Step S32: Repeat the sequence optimization for each round.

[0180] Step S33: Based on the optimal solution of the yaw angle in the previous round of sequential optimization (0° in the first round) and the set number of yaw angles for the current round, To form a yaw angle test sequence As shown in equation (3.2); the median of this yaw angle test sequence is the optimal solution for the yaw angle in the previous round of sequential optimization, and there are a total of The number of yaw angles is determined by the fact that the differences between each number in the previous round are equal, and the difference between each number in the previous round is divided by the number of yaw angles in this round. For example, with a median of 10°, a difference of 5 between each number in the previous round, and a set number of 5, the yaw angle test sequence would be {8°, 9°, 10°, 11°, 12°}. If the sequence exceeds the upper or lower limits of the yaw angle, the entire yaw angle test sequence is shifted to ensure that all yaw angles meet the upper and lower limit requirements.

[0181] (3.2);

[0182] in, Indicates the first Yaw angle.

[0183] Step S34: Cycle according to the windward sequence of all wind turbines in the subgroup.

[0184] Step S35: Loop through each prediction time step of MPC.

[0185] Step S36: In all MPC prediction time steps, all wind turbines within the subgroup are arranged in windward order. Predicting time step, the first step A wind turbine unit facing the wind is represented as As shown in equation (3.3).

[0186] (3.3);

[0187] Step S37: Calculate the current prediction time step, the number of sequential optimization rounds, and the... The target wind turbine is an upstream wind turbine. The power at that time is given by equation (3.4), where the target wind turbine unit The yaw angle is taken from each value in the yaw angle test sequence.

[0188] (3.4);

[0189] in, Indicates the current prediction time step Optimize the number of rounds in sequence Target wind turbine Power at that time; Indicates the current prediction time step Optimize the number of rounds in sequence Target wind turbine The yaw angle taken at that time; Indicates the target wind turbine unit Method for calculating wind speed at the center of a wheel hub; Indicates the current prediction time step At that time, the target wind turbine Yaw angle; Indicates the current order optimization round number as Yaw angle sequence at time; Indicates the current prediction time step At that time, the first Yaw angle of each wind turbine unit; This indicates the method for calculating power generation. Indicates the target number of wind turbine units.

[0190] Step S38, based on simplifying assumption 1: the yaw angle of the downstream wind turbine is independent of the optimal solution of the yaw angle of the upstream wind turbine.

[0191] Step S39, return to step S35, until the loop for all predicted time steps is completed.

[0192] In step S310, at each moment, only the yaw angle of one wind turbine is adjusted. The yaw angles of the wind turbines adjusted at different moments are cross-combined to calculate the objective function, which is shown in equation (3.5).

[0193] (3.5);

[0194] in, express The number of each prediction time step; This represents the loss value under the current yaw angle combination; Indicates the first The typhoon turbine unit in the Power generation at each predicted time step; This represents the matrix indicating yaw angle and power; This represents the objective function value under the yaw angle sequence.

[0195] Assuming the objective function is decoupled across different time points, then when calculating the objective function, the power generation at different prediction time steps, for different wind turbines, and for different yaw angles has already been calculated in step S37.

[0196] Step S311, return to step S34, until the cycle of all wind turbine units is completed.

[0197] Step S312: Find the yaw angle combination that minimizes the multi-objective optimization function value. This combination is called the optimal yaw angle solution under the current sequential optimization round number, i.e., the optimal yaw angle combination, as shown in equation (3.6).

[0198] (3.6);

[0199] in, This represents the optimal yaw angle combination under the current sequential optimization round number.

[0200] Step S313: If the current sequence optimization round is not the final sequence optimization round, return to step S32 until the loop of all sequence optimization rounds is completed.

[0201] Step S314: If the current sequence optimization wheel is the final sequence optimization wheel, output the optimal yaw angle solution and end.

[0202] In one specific implementation of this embodiment, a data acquisition and preprocessing step is included before step S1. Specifically, this involves: acquiring historical and real-time data of the wind farm group, including wind speed, wind direction, temperature, and operating power; performing integrity and rationality checks on the historical and real-time data, and processing unreasonable data (unreasonable data refers to wind direction jumps, missing data interpolation, etc.); and performing normalization and denormalization processing on the historical and real-time data to meet the input requirements for analyzing the wake loss model.

[0203] It should be noted that historical data is used to train and validate the constructed wake loss analysis model; real-time data is used as input to the wake loss analysis model for real-time wake regulation.

[0204] Example 2

[0205] Taking a 800MW offshore wind farm cluster in southeastern coastal China as an example, this wind farm cluster includes 134 wind turbine units. The control cycle is set to 15 minutes.

[0206] Set wake velocity loss threshold Based on the predicted wind direction (e.g., 1.41°) and wind speed (e.g., 8.14 m / s) for the next cycle, the 134 wind turbines are dynamically grouped. In this embodiment's control cycle, the wind turbines are divided into 55 subgroups, with the largest subgroup containing 6 wind turbines.

[0207] The weight coefficients of the multi-objective optimization function are calculated based on economic costs; in this embodiment, they are taken as... , , .

[0208] Parallel optimization is performed on the 55 subgroups. For each subgroup, two rounds of sequential optimization are executed:

[0209] Round 1: The candidate yaw angle sequence is {-15°, -10°, -5°, 0°, 5°, 10°, 15°}.

[0210] Second round: Using the optimal solution obtained in the previous round (e.g., 10°) as the median, generate a new sequence {8°, 9°, 10°, 11°, 12°}.

[0211] Following the windward order, select the multi-objective optimization function for each unit in the subgroup. The minimum yaw angle yields the following: Figure 4 The diagram showing the optimized yaw angle of the wind turbine is illustrated. The optimization process took approximately 18 seconds, significantly less than the 15-minute control cycle. Figure 4 The horizontal axis X represents the horizontal distance, and the vertical axis Y represents the vertical distance.

[0212] The yaw angle command obtained from the first predicted time step was sent to all 134 aircraft units on site, and the result was as follows: Figure 5 This is a velocity contour map of the wind farm cluster. Figure 6 This is a turbulence intensity contour map of the wind farm complex. Among them, Figure 5 The x-axis represents the horizontal distance, and the y-axis represents the vertical distance; Figure 6 The horizontal axis X represents the horizontal distance, and the vertical axis Y represents the vertical distance.

[0213] Calculations show that applying the wake control method of this invention can increase the theoretical annualized power generation of the wind farm group by 0.65%, with an annualized economic benefit of approximately RMB 4,500 per megawatt of installed capacity.

[0214] In summary, the wake control method of the present invention successfully solves the real-time optimization problem of wake control for large-scale offshore wind farm clusters, and has the advantages of high efficiency, effectiveness and engineering applicability.

[0215] Example 2

[0216] Based on the same inventive concept as Embodiment 1, this embodiment introduces a wake control system for offshore wind farm clusters based on dynamic clustering and sequence optimization, used to implement the wake control method for offshore wind farm clusters based on dynamic clustering and sequence optimization in Embodiment 1, including:

[0217] The dynamic clustering module is configured to build an analysis model of wake loss based on the Larsen wake model; based on the velocity loss of the target wind turbine output by the analysis model of wake loss, it dynamically clusters all wind turbines in the wind farm cluster to obtain several sub-clusters.

[0218] The multi-objective optimization function construction module is configured to construct a multi-objective optimization function that considers the power generation of the wind farm cluster, the fatigue loss of the yaw system, and the fatigue and life of the wind turbine.

[0219] The sequential optimization module is configured to perform sequential optimization within each subgroup, optimizing the yaw angle of each aircraft in the order of windward movement, and solving for the optimal combination of yaw angles that minimizes the multi-objective optimization function.

[0220] The rolling execution module is configured to update the dynamic clustering results and yaw commands once every preset control cycle.

[0221] The command issuing module is configured to issue the optimal yaw angle combination of each wind turbine unit obtained by solving the problem for the next control cycle as a control command to achieve the regulation of the wake of the offshore wind farm group.

[0222] Example 3

[0223] Based on the same inventive concept as other embodiments, this embodiment introduces a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the above-described method for controlling the wake of offshore wind farm clusters based on dynamic clustering and sequential optimization.

[0224] Example 4

[0225] Based on the same inventive concept as other embodiments, this embodiment introduces a computer device, including: a memory for storing a computer program; and a processor for executing the computer program to implement the steps of the above-described method for controlling the wake of offshore wind farm clusters based on dynamic clustering and sequential optimization.

[0226] Example 5

[0227] Based on the same inventive concept as other embodiments, this embodiment introduces a computer program product, including a computer program that, when executed by a processor, implements the steps of the above-described method for controlling the wake of offshore wind farm clusters based on dynamic clustering and sequential optimization.

[0228] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0229] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0230] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0231] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0232] The embodiments of the present invention have been described above with reference to the accompanying drawings. However, the present invention is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other modifications under the guidance of the present invention, and these modifications are all within the protection scope of the present invention.

Claims

1. A method for controlling the wake of an offshore wind farm cluster based on dynamic grouping and sequential optimization, characterized in that, Includes the following steps: Based on the Larsen wake model, a wake loss analysis model is constructed; based on the velocity loss of the target wind turbine output by the wake loss analysis model, all wind turbines in the wind farm cluster are dynamically grouped to obtain several subgroups. Construct a multi-objective optimization function that considers the power generation of the wind farm cluster, the fatigue loss of the yaw system, and the fatigue and life of the wind turbine units; Within each subgroup, sequential optimization is performed, optimizing the yaw angle of each aircraft in the order of windward movement, and solving for the optimal yaw angle combination that minimizes the multi-objective optimization function. The dynamic clustering results and yaw commands are updated once every preset control cycle; The optimal yaw angle combination of each wind turbine unit obtained from the solution and corresponding to the next control cycle is issued as a control command to achieve the regulation of the wake of the offshore wind farm group.

2. The offshore wind farm wake control method based on dynamic clustering and sequence optimization as described in claim 1, characterized in that, The aforementioned wake loss analysis model, based on the Larsen wake model, includes: The speed loss of the downstream target wind turbine is the axial distance. and radial distance The function, with the following expression: ； ； ； ； ； ； in, Indicates the area of ​​the wind turbine; Indicates the wake radius; Indicates the 0-dimensional mixing length of Prandtl; Indicates the thrust coefficient of the wind turbine; Indicates the diameter of the wind turbine; Represents an approximate number; This represents the wake radius at a point where the wake length is 9.5 times the rotor diameter. An empirical formula for representing the wake radius at a wake length of 9.5 times the rotor diameter; This indicates the intensity of atmospheric turbulence at the hub height; This indicates the axial distance downstream of the wind turbine. Radial distance is The wind speed at the location and within the wake region; Indicates the free-flowing wind speed; This indicates the hub height of the wind turbine.

3. The offshore wind farm wake control method based on dynamic clustering and sequence optimization as described in claim 1, characterized in that, Based on the velocity loss of the target wind turbine output by the wake loss analysis model, all wind turbines in the wind farm cluster are dynamically grouped to obtain several subgroups, including: Calculate upstream wind turbines For the target wind turbine located downstream Speed ​​loss caused Preset speed loss threshold ,Will The target wind turbine units are grouped into the same subgroup.

4. The offshore wind farm wake control method based on dynamic clustering and sequence optimization as described in claim 3, characterized in that, The specific steps of the dynamic clustering are as follows: Step S11: Iterate for each prediction time step of the model predictive control; Step S12: Initialize the clustering results of the current model's predictive control at prediction time step t. , as in equation (1.1); （1.1）； in, Indicates the number of clusters; Step S13: Cycle through each wind turbine unit; in, Indicates the serial number of the target wind turbine unit, Indicates the target number of wind turbine units; Step S14: Under the condition of no yaw, calculate all upstream wind turbine units. For the target wind turbine The resulting speed loss; setting up measures for upstream wind turbines With the target wind turbine speed loss The velocity loss threshold κ, if If the value is greater than or equal to κ, then the upstream wind turbine is considered to be... With the target wind turbine There is value in active yaw wake control; all The set is j, as shown in equation (1.2); （1.2）； In the formula, It is the ambient wind speed; Indicates the prediction time step upstream wind turbines The wake of the target wind turbine Wind speed at the center of the wheel hub; This indicates the predicted time step. Environmental wind speed; Step S15, all upstream wind turbine units satisfying equation (1.2) Included In each group, as in equation (1.3); （1.3）； in, Indicates the prediction time step This includes upstream wind turbines. The group; Indicates the first The groups at the predicted time step The state; Indicates the index of the group; Indicates the prediction time step The Chinese number is The group; Represents the set of group IDs one of the; Step S16, merge Join a group and add the target wind turbine. , as in equation (1.4); （1.4）； in, Represents a collection of groups; Step S17, return to step S13, until the cycle of all wind turbine units is completed; Step S18, return to step S11, until the loop for all prediction time steps is completed; Step S19: Perform dynamic clustering and initialize the dynamic clustering results. , as in equation (1.5); （1.5）； In the formula, It is the number of prediction time steps in model predictive control; yes The number of groups; express The Middle Groups They represent the first A group of predicted time steps; Step S110: Determine whether there are any different groups that contain the same elements, as shown in equation (1.6). （1.6）； in, and This represents any two distinct groups; Step S111: If the condition judgment result of step S110 is yes, then the condition is satisfied. The set of groups, as shown in equation (1.7). （1.7）； in, Indicates a group that meets the criteria; Indicates number For groups; They represent the sets of natural numbers respectively; Step S112, will Take the union of all elements of the original group to form a new group; delete the old group. A group, as shown in equation (1.8); return to step S110; （1.8）； Step S113: If the condition judgment result of step S110 is negative, the final clustering result G is obtained, and the process ends.

5. The offshore wind farm wake control method based on dynamic clustering and sequence optimization as described in claim 1, characterized in that, The construction of a multi-objective optimization function considering the power generation of the wind farm cluster, the fatigue loss of the yaw system, and the fatigue and lifespan of the wind turbine units includes: The expression for the multi-objective optimization function is: ； ； ； ； In the formula, It is a multi-objective optimization function for active yaw wake control; yes The term considering power generation capacity is the power generation capacity enhancement term; yes The term that considers fatigue losses in the yaw system is the yaw loss term; yes The term considering wind turbine fatigue and lifespan, namely the power fluctuation term; , , These are the weighting coefficients for the power generation improvement item, the yaw loss item, and the power fluctuation item, respectively. This is the number of prediction time steps for MPC; It is a prediction time step In the matrix, the yaw angles of all wind turbine units; It is a zero matrix; The target wind turbine Rated power; It's an angle. and angle The angle between them; Indicates the target wind turbine unit Method for calculating wind speed at the center of the wheel hub Indicates the target wind turbine unit The method for calculating power generation capacity; Indicates the prediction time step The original angle; Indicates the prediction time step Target wind turbine Yaw angle; The predicted time step is Target wind turbine The yaw angle; RMSE(a) is the standard deviation of set a; It is a matrix representing the yaw angles of all wind turbines within the subgroup across all prediction time steps; It is a prediction time step At that time, the target wind turbine Yaw angle, Indicates the current state. Indicates the prediction time step is All states; The time step for prediction is At that time, the target wind turbine The yaw angle is 0. Indicates the current state. Indicates the prediction time step is All states; This indicates the target number of wind turbine units.

6. The offshore wind farm wake control method based on dynamic clustering and sequence optimization as described in claim 1, characterized in that, The process of sequential optimization within each subgroup, optimizing the yaw angle aircraft by aircraft in the order of windward movement, and solving for the optimal yaw angle combination that minimizes the multi-objective optimization function includes: Step S31: Set the number of rounds for sequential optimization and the number of yaw angles per round, as shown in equation (3.1): （3.1）； in, A matrix representing the number of rounds and the number of yaw angles per round; This indicates the number of yaw angles for the current round; Indicates the current optimization round number; Step S32: Repeat the process for each round of sequence optimization; Step S33: Based on the optimal solution of the yaw angle in the previous round of sequential optimization and the set number of yaw angles for the current round, To form a yaw angle test sequence As shown in equation (3.2); the median of the yaw angle test sequence is the optimal solution of the yaw angle in the previous round of sequential optimization, and there are a total of The number of numbers, where the difference between each number is equal and is equal to the difference between each number in the previous round divided by the number in this round; （3.2）； in, Indicates the first One yaw angle; Step S34: Cycle according to the windward sequence of all wind turbines in the subgroup; Step S35: Iterate through each prediction time step controlled by the model prediction framework; Step S36: In all model prediction framework control prediction time steps, all wind turbines within the subgroup are arranged in windward order. Step prediction time step, first A wind turbine unit facing the wind is represented as , as in equation (3.3); （3.3）； Step S37: Calculate the current prediction time step, the number of sequential optimization rounds, and the... The target wind turbine is an upstream wind turbine. The power at that time is given by equation (3.4), where the target wind turbine unit The yaw angle is taken from each value in the yaw angle test sequence; （3.4）； in, Indicates the current prediction time step Optimize the number of rounds in sequence Target wind turbine Power at that time; Indicates the current prediction time step Optimize the number of rounds in sequence Target wind turbine The yaw angle taken at that time; Indicates the target wind turbine unit Method for calculating wind speed at the center of a wheel hub; Indicates the current prediction time step At that time, the target wind turbine Yaw angle; Indicates the current order optimization round number as Yaw angle sequence at time; Indicates the current prediction time step At that time, the first Yaw angle of each wind turbine unit; This indicates the method for calculating power generation. Indicates the target number of wind turbine units; Step S38, based on simplifying assumption 1; the simplifying assumption 1 is that the yaw angle of the downstream wind turbine is independent of the optimal solution of the yaw angle of the upstream wind turbine; Step S39, return to step S35, until the loop for all prediction time steps is completed; In step S310, at each moment, only the yaw angle of one wind turbine is adjusted. The yaw angles of the wind turbines adjusted at different moments are cross-combined to calculate the objective function, which is shown in equation (3.5). （3.5）； in, express The number of each prediction time step; This represents the loss value under the current yaw angle combination; Indicates the first The typhoon turbine unit in the Power generation at each predicted time step; This represents the matrix indicating yaw angle and power; This represents the objective function value under the yaw angle sequence; Step S311, return to step S34, until the cycle of all wind turbine units is completed; Step S312: Find the combination of yaw angles that minimizes the value of the multi-objective optimization function. This combination is called the optimal solution of yaw angles under the current sequential optimization round number, as shown in equation (3.6). （3.6）； in, This represents the optimal yaw angle combination under the current optimization round number; Step S313: If the current sequence optimization round is not the final sequence optimization round, return to step S32 until the loop of all sequence optimization rounds is completed; Step S314: If the current sequence optimization wheel is the final sequence optimization wheel, output the optimal yaw angle solution and end.

7. A wake control system for offshore wind farm clusters based on dynamic grouping and sequential optimization, characterized in that, The method for implementing the wake control of offshore wind farm clusters based on dynamic grouping and sequence optimization as described in any one of claims 1 to 6 includes: The dynamic clustering module is configured to build an analysis model of wake loss based on the Larsen wake model; based on the velocity loss of the target wind turbine output by the analysis model of wake loss, it dynamically clusters all wind turbines in the wind farm cluster to obtain several sub-clusters. The multi-objective optimization function construction module is configured to construct a multi-objective optimization function that considers the power generation of the wind farm cluster, the fatigue loss of the yaw system, and the fatigue and life of the wind turbine. The sequential optimization module is configured to perform sequential optimization within each subgroup, optimizing the yaw angle of each aircraft in the order of windward movement, and solving for the optimal combination of yaw angles that minimizes the multi-objective optimization function. The rolling execution module is configured to update the dynamic clustering results and yaw commands once every preset control cycle. The command issuing module is configured to issue the optimal yaw angle combination of each wind turbine unit obtained by solving the problem for the next control cycle as a control command to achieve the regulation of the wake of the offshore wind farm group.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the processor, it implements the steps of the offshore wind farm wake control method based on dynamic grouping and sequential optimization as described in any one of claims 1 to 6.

9. A computer device, characterized in that, include: Memory, used to store computer programs; A processor is configured to execute the computer program to implement the steps of the offshore wind farm wake control method based on dynamic clustering and sequential optimization as described in any one of claims 1 to 6.

10. A computer program product, comprising a computer program, characterized in that: When the computer program is executed by the processor, it implements the steps of the offshore wind farm wake control method based on dynamic grouping and sequential optimization as described in any one of claims 1 to 6.

Images