Wind Farm Frequency Control Method and System Based on Real-Time Dynamic Estimation of Adjustment Capability

Abstract

The present invention discloses a wind farm frequency control method and system based on real-time regulation capability dynamic estimation, which solves the problem that when the existing method relies on a long short-term memory network to predict the wind farm power, the dynamic changes of its weight and gating mechanism make the prediction evaluation performance poor when facing highly uncertain and noisy multi-source heterogeneous data. The method comprises: obtaining multi-source heterogeneous sequential wind farm operating parameters, identifying and analyzing the sequential wind farm operating parameters with a frequency dynamic estimation model, outputting a wind farm frequency prediction sequence, and calculating the frequency regulation support control parameters corresponding to the wind farm frequency prediction sequence based on a cuckoo algorithm combined with an inertia safety constraint; the present invention constructs a frequency dynamic estimation model capable of predicting the wind farm power grid frequency based on the multi-source heterogeneous sequential wind farm operating parameters, and the construction and training of the frequency dynamic estimation model can ensure the rapid response capability of the wind turbine generator set based on its own frequency regulation responsibility and frequency conversion support capability.

Description

Technical Field

[0001] The present invention belongs to the technical field of wind power generation, and in particular relates to a wind farm frequency control method and system based on real-time regulation capability dynamic estimation. Background Art

[0002] As a clean and renewable energy source, wind power has been rapidly developed around the world, and wind power cluster development and grid connection have become the mainstream model of wind power construction. Under this model, wind farms are no longer isolated units, but participate in the active power control and energy management of the power grid as a whole. Therefore, the wind power cluster control system becomes the key. The core lies in how to achieve precise control and optimization of wind farm output through effective control strategies.

[0003] The wind power AGC system is the abbreviation of automatic generation control, which refers to the system that connects to the active power dispatch instructions of the power grid in the wind farm to automatically adjust the output of the energy station. The wind power AGC system can not only promote the optimized operation of the wind farm, but also significantly improve the competitiveness of wind power energy in the electricity market.

[0004] Chinese patent CN114336592B discloses a wind farm AGC control system based on model predictive control, including a power grid control center, a wind farm automatic power generation control module, and a wind farm unit, characterized in that: the wind farm automatic power generation control module includes a model prediction unit and a rolling optimization unit; the power grid control center inputs a given wind farm reference power into the wind farm automatic power generation control module, and outputs an active power control instruction to the wind turbine units in the wind farm; the wind farm automatic power generation control module outputs an active power control instruction to control the output of the wind farm units through model prediction and rolling optimization; through an automatic power generation control strategy based on model predictive control, multiple wind turbines in the wind farm are coordinated to uniformly participate in active power control. However, when the existing method relies on a long short-term memory network to predict the wind farm power, when faced with highly uncertain and noisy multi-source heterogeneous data, the dynamic changes in its weight and gating mechanism make the model's multi-source heterogeneous data prediction and evaluation performance poor. In response to the above problems, we propose a wind farm frequency control method and system based on real-time regulation capability dynamic estimation. Summary of the invention

[0005] The purpose of the present invention is to address the shortcomings of the prior art and provide a wind farm frequency control method and system based on dynamic estimation of real-time regulation capability, which solves the problem that when the existing method relies on long short-term memory network to predict wind farm power, the dynamic changes of its weights and gating mechanism make the multi-source heterogeneous data prediction evaluation performance of the model poor when faced with highly uncertain and noisy multi-source heterogeneous data.

[0006] When existing methods rely on long short-term memory networks to predict wind farm power, when faced with highly uncertain and noisy multi-source heterogeneous data, the dynamic changes in its weights and gating mechanisms make the model's multi-source heterogeneous data prediction and evaluation performance poor. To address the above problems, we propose a wind farm frequency control method and system based on dynamic estimation of real-time regulation capability. In short, when the method is executed, the multi-source heterogeneous sequential wind farm operating parameters are first obtained, and a frequency dynamic estimation model is constructed by combining fuzzy control algorithm and back propagation neural network. The frequency dynamic estimation model then identifies and analyzes the sequential wind farm operating parameters and outputs a wind farm frequency prediction sequence. Finally, the frequency support control parameters corresponding to the wind farm frequency prediction sequence are calculated based on the cuckoo algorithm combined with inertia safety constraints, and the wind turbines execute the grid frequency active support instructions based on the frequency support control parameters. In an embodiment of the present invention, a frequency dynamic estimation model capable of predicting the wind farm power grid frequency is constructed based on multi-source heterogeneous sequential wind farm operating parameters. The construction and training of the frequency dynamic estimation model can ensure the rapid response capability of the wind turbines based on their own frequency regulation responsibility and frequency conversion support capability, thereby overcoming the problem that when the existing method relies on long short-term memory networks to predict wind farm power, when faced with highly uncertain and noisy multi-source heterogeneous data, the dynamic changes in its weights and gating mechanisms result in poor multi-source heterogeneous data prediction and evaluation performance of the model.

[0007] The present invention is implemented by a wind farm frequency control method based on real-time regulation capability dynamic estimation, the wind farm frequency control method based on real-time regulation capability dynamic estimation comprising:

[0008] Acquire multi-source heterogeneous sequential wind farm operating parameters, wherein the wind farm operating parameters include power grid frequency signals, wind farm environmental data, wind turbine power, wind turbine load data, wind turbine rotor speed, wind turbine rotor current and voltage data, standardize the multi-source heterogeneous wind farm operating parameters, and upload the standardized wind farm operating parameters to the station database;

[0009] Traversing the station database, randomly extracting historical working parameters from the station database, building a frequency dynamic estimation model by combining fuzzy control algorithm and back propagation neural network, and iteratively training the frequency dynamic estimation model through historical working parameters;

[0010] The sequential wind farm operating parameters are loaded, and the frequency dynamic estimation model is executed with the sequential wind farm operating parameters as input. The frequency dynamic estimation model identifies and analyzes the sequential wind farm operating parameters and outputs the wind farm frequency prediction sequence.

[0011] The wind farm frequency prediction sequence is obtained, and the frequency regulation support control parameters corresponding to the wind farm frequency prediction sequence are calculated based on the cuckoo algorithm combined with inertia safety constraints. The wind turbines execute the grid frequency active support instructions based on the frequency regulation support control parameters.

[0012] Preferably, the method of constructing a frequency dynamic estimation model by combining a fuzzy control algorithm and a back propagation neural network specifically includes:

[0013] The back propagation neural network model is used as the initial model of the frequency dynamic estimation model, wherein the initial model includes an input layer, a hidden layer, and an output layer. A fuzzy control mechanism is incorporated into the input layer of the initial model, and the number of neurons in the output layer, hidden layer, and output layer are 5, 4, and 4 respectively;

[0014] The lightweight prediction structure Fire Module structure is used to improve the hidden layer of the initial model, and the loss function and activation function of the hidden layer are set to obtain the improved initial model;

[0015] Load the historical working parameters, decompose the historical working parameters based on the singular spectrum decomposition method, obtain the parameter decomposition set, and divide the parameter decomposition set into a training set and a test set according to the proportion;

[0016] Set the initial weight and judgment threshold, optimize the initial weight and judgment threshold based on the fuzzy control algorithm, and obtain the optimal weight and optimal threshold;

[0017] Obtain the training set. The initial model uses the Levenberg-Marquardt algorithm to iteratively train the training set until the initial model converges. The maximum number of training times is 800, and the initial learning rate is set to 0.01.

[0018] A test set is obtained, and the evaluation accuracy of the initial model is tested based on the test set. If the evaluation accuracy meets a preset accuracy threshold, a frequency dynamic estimation model is output.

[0019] Preferably, the method for the frequency dynamic estimation model to identify and analyze the sequential wind farm operating parameters and output the wind farm frequency prediction sequence specifically includes:

[0020] Obtaining the sequential wind farm operating parameters, decomposing the sequential wind farm operating parameters using the singular spectrum decomposition method, and obtaining a parameter decomposition set;

[0021] Among them, the parameter decomposition set is calculated by the following formula:

[0022] (1)

[0023] in, is the parameter decomposition vector, To use the singular spectrum decomposition method to decompose the time-series wind farm operating parameters, is the singular spectrum decomposition window length, It is the sequential wind farm operating parameter;

[0024] (2)

[0025] in, is the data dimension of the time-series wind farm operating parameters, is the initial weight of the sequential wind farm operating parameters, It is the time for collecting the working parameters of the sequential wind farm;

[0026] Load the parameter decomposition set, the fuzzy control mechanism in the input layer constructs the parameter fuzzy set based on the parameter decomposition set, and calculates the prediction confidence of the parameter fuzzy set;

[0027] The hidden layer obtains the parameter fuzzy set and the corresponding prediction confidence. The lightweight prediction structure FireModule structure in the hidden layer constructs a frequency parameter conversion function based on the parameter fuzzy set and prediction confidence.

[0028] Among them, the frequency parameter conversion function is expressed as:

[0029] (5)

[0030] in, To predict the frequency conversion coefficient at the next moment, is the transformation coefficient of the frequency parameter conversion function, is the number of data in the parameter fuzzy set, for frequency discrete scenarios, is the input value of the parameter fuzzy set;

[0031] Outputting the predicted frequency of the wind farm at the next moment based on the frequency conversion coefficient, integrating the predicted frequency of the wind farm, and outputting a wind farm frequency prediction sequence;

[0032] Among them, the wind farm prediction frequency is calculated by the following formula:

[0033] (6)

[0034] in, Predict frequency for wind farms, To predict the frequency conversion coefficient at the next moment.

[0035] Preferably, when constructing the parameter fuzzy set, it is constructed based on the confidence constraints at different deviation levels, and the confidence constraints are:

[0036] (3)

[0037] in, is the confidence clustering constraint, is the number of data of parameter fuzzy set, for frequency discrete scenarios, is the scene probability corresponding to the parameter decomposition set, is the initial scene probability value of the parameter decomposition set, is the upper limit of the total deviation of the scene probability distribution, The upper limit of the deviation of the scene probability distribution allowed;

[0038] Among them, the prediction confidence of the parameter fuzzy set is calculated by the following formula:

[0039] (4)

[0040] in, is the prediction confidence of the parameter fuzzy set.

[0041] Preferably, the method for calculating the frequency regulation support control parameters corresponding to the wind farm frequency prediction sequence based on the cuckoo algorithm combined with the inertia safety constraint specifically includes:

[0042] Obtain a wind farm frequency prediction sequence, and calculate the wind turbine constrained power corresponding to the wind farm frequency prediction sequence based on inertia safety constraints;

[0043] Load the constrained power of the wind turbine set, and optimize the wind power contribution rate and frequency regulation capability factor parameters of the wind turbine set based on the cuckoo algorithm and the constrained power of the wind turbine set;

[0044] Among them, the optimization function of wind power contribution rate is expressed as:

[0045] (10)

[0046] in, is the wind power contribution rate, is the weight coefficient of the current wind turbine, To find the optimal constant, is the distance between the current wind turbine and the central unit of the wind farm, is the spatial radius of the wind farm center, is the distance between the current wind turbine and the center of the wind farm in terms of spatial radius;

[0047] The optimization function of frequency modulation capability factor is expressed as:

[0048] (11)

[0049] in, is the frequency modulation capability factor.

[0050] Obtain the optimal wind power contribution rate and frequency regulation capability factor of the wind turbine, and calculate the frequency regulation responsibility coefficient corresponding to the wind turbine based on the wind power contribution rate and frequency regulation capability factor;

[0051] The frequency regulation responsibility coefficient is obtained, and primary frequency regulation and reactive voltage regulation instructions are generated based on the frequency regulation responsibility coefficient. The wind turbines respond to the primary frequency regulation and reactive voltage regulation instructions and execute the grid frequency active support instructions.

[0052] Preferably, the wind turbine generator set power constraint satisfies the following inertia safety constraint:

[0053] (7)

[0054] in, This is the frequency regulation period of wind turbines. is the actual power adjustment, The wind turbine power constraint during the frequency modulation period. They are the minimum power adjustment amount and the maximum power adjustment amount respectively;

[0055] (8)

[0056] in, For the previous moment Wind turbine constraint power, is the number of wind turbines, is the frequency regulation power coefficient of the wind turbine;

[0057] (9)

[0058] in, is the frequency modulation power coefficient of the wind turbine, is the angular velocity of the wind turbine rotor, is the blade radius of the wind turbine, is the number of wind turbine blades, is the wind speed, is the wind turbine rotor moment of inertia.

[0059] The present invention also provides a wind farm frequency control system based on real-time regulation capability dynamic estimation, the wind farm frequency control system based on real-time regulation capability dynamic estimation specifically comprises:

[0060] The parameter acquisition module is used to acquire the time-sequential wind farm operating parameters of the multi-source heterogeneous wind farm, standardize the multi-source heterogeneous wind farm operating parameters, and upload the standardized wind farm operating parameters to the station database;

[0061] A frequency dynamic estimation module is used to traverse the station database, randomly extract historical working parameters from the station database, build a frequency dynamic estimation model by combining fuzzy control algorithm and back propagation neural network, and iteratively train the frequency dynamic estimation model through historical working parameters;

[0062] The frequency prediction module is used to load the sequential wind farm operating parameters, take the sequential wind farm operating parameters as input, execute the frequency dynamic estimation model, the frequency dynamic estimation model identifies and analyzes the sequential wind farm operating parameters, and outputs the wind farm frequency prediction sequence;

[0063] The support control module is used to obtain the wind farm frequency prediction sequence, calculate the frequency support control parameters corresponding to the wind farm frequency prediction sequence based on the cuckoo algorithm combined with inertia safety constraints, and the wind turbines execute the grid frequency active support instructions based on the frequency support control parameters.

[0064] Preferably, the support control module comprises:

[0065] A power prediction unit, used to obtain a wind farm frequency prediction sequence, and calculate the wind turbine constrained power corresponding to the wind farm frequency prediction sequence based on inertia safety constraints;

[0066] A parameter optimization unit is used to load the constrained power of the wind turbine set and optimize the wind power contribution rate and frequency regulation capability factor parameters of the wind turbine set based on the cuckoo algorithm and the constrained power of the wind turbine set;

[0067] A responsibility coefficient calculation unit is used to obtain the optimal wind power contribution rate and frequency regulation capability factor of the wind turbine generator set, and calculate the frequency regulation responsibility coefficient corresponding to the wind turbine generator set based on the wind power contribution rate and the frequency regulation capability factor;

[0068] The support instruction execution unit is used to obtain the frequency regulation responsibility coefficient, generate primary frequency regulation and reactive voltage regulation instructions based on the frequency regulation responsibility coefficient, and the wind turbine group responds to the primary frequency regulation and reactive voltage regulation instructions to execute the grid frequency active support instructions.

[0069] Compared with the prior art, the embodiments of the present application have the following beneficial effects:

[0070] In an embodiment of the present invention, a frequency dynamic estimation model capable of predicting the wind farm power grid frequency is constructed based on multi-source heterogeneous sequential wind farm operating parameters. The construction and training of the frequency dynamic estimation model can ensure the rapid response capability of the wind turbines based on their own frequency regulation responsibility and frequency conversion support capability, thereby overcoming the problem that when the existing method relies on long short-term memory networks to predict wind farm power, when faced with highly uncertain and noisy multi-source heterogeneous data, the dynamic changes in its weights and gating mechanisms result in poor multi-source heterogeneous data prediction and evaluation performance of the model.

[0071] In an embodiment of the present invention, a frequency dynamic estimation model and a construction and training method thereof are provided. The frequency dynamic estimation model is constructed by combining a fuzzy control algorithm and a back propagation neural network. The fuzzy control mechanism can effectively handle the uncertainty and noise in the wind farm parameters by fuzzifying the processing parameters, thereby enhancing the robustness of the model. The fuzzy control mechanism allows the model to dynamically adjust the prediction strategy according to real-time data. This adaptive capability makes BPNN more advantageous in dealing with rapidly changing wind farm environments. In addition, the use of a lightweight prediction structure, the Fire Module structure, to improve the hidden layer of the initial model can achieve efficient prediction through simplified calculation steps.

[0072] In the embodiment of the present invention, the frequency dynamic estimation model can accurately identify and analyze the time-series wind farm operating parameters, thereby accurately outputting the wind farm frequency prediction sequence. Compared with the existing technology, it can dynamically adjust the prediction strategy according to real-time data. This adaptive capability makes the frequency dynamic estimation model more advantageous when dealing with rapidly changing wind farm environments.

[0073] In an embodiment of the present invention, the frequency regulation support control parameters corresponding to the wind farm frequency prediction sequence are calculated by combining the cuckoo algorithm with the inertia safety constraint. As an effective optimization tool, the cuckoo algorithm can play a key role in calculating the frequency regulation support control parameters. By simulating the parasitic behavior of the cuckoo, the cuckoo algorithm can effectively search for the optimal solution of the wind power contribution rate and the frequency regulation capability factor, thereby optimizing the frequency regulation responsibility coefficient of the wind farm. This optimization helps to improve the performance of wind turbines in the wind farm in the grid frequency regulation and ensure the stable operation of the power system. BRIEF DESCRIPTION OF THE DRAWINGS

[0074] Figure 1 It is a schematic diagram of the implementation flow of the wind farm frequency control method based on real-time regulation capability dynamic estimation provided by the present invention.

[0075] Figure 2 It is a schematic diagram of the implementation flow of the method for constructing a frequency dynamic estimation model by combining a fuzzy control algorithm and a back propagation neural network provided by the present invention.

[0076] Figure 3 The present invention provides a method for realizing a flow chart of a method for identifying and analyzing sequential wind farm operating parameters using a frequency dynamic estimation model and outputting a wind farm frequency prediction sequence.

[0077] Figure 4 It is a schematic diagram of the implementation flow of the method provided by the present invention for calculating the frequency modulation support control parameters corresponding to the wind farm frequency prediction sequence based on the cuckoo algorithm combined with inertia safety constraints.

[0078] Figure 5 It is a schematic diagram of the control principle of a wind farm frequency control system based on real-time regulation capability dynamic estimation provided by the present invention. DETAILED DESCRIPTION

[0079] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by technicians in the technical field of this application; the terms used in the specification of the application herein are only for the purpose of describing specific embodiments and are not intended to limit this application; the terms "including" and "having" and any variations thereof in the specification and claims of this application and the above-mentioned drawings are intended to cover non-exclusive inclusions. The terms "first", "second", etc. in the specification and claims of this application or the above-mentioned drawings are used to distinguish different objects, not to describe a specific order.

[0080] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0081] When existing methods rely on long short-term memory networks to predict wind farm power, when faced with highly uncertain and noisy multi-source heterogeneous data, the dynamic changes in its weights and gating mechanisms make the model's multi-source heterogeneous data prediction and evaluation performance poor. To address the above problems, we propose a wind farm frequency control method and system based on dynamic estimation of real-time regulation capability. In short, when the method is executed, the multi-source heterogeneous sequential wind farm operating parameters are first obtained, and a frequency dynamic estimation model is constructed by combining fuzzy control algorithm and back propagation neural network. The frequency dynamic estimation model then identifies and analyzes the sequential wind farm operating parameters and outputs a wind farm frequency prediction sequence. Finally, the frequency support control parameters corresponding to the wind farm frequency prediction sequence are calculated based on the cuckoo algorithm combined with inertia safety constraints, and the wind turbines execute the grid frequency active support instructions based on the frequency support control parameters. In an embodiment of the present invention, a frequency dynamic estimation model capable of predicting the wind farm power grid frequency is constructed based on multi-source heterogeneous sequential wind farm operating parameters. The construction and training of the frequency dynamic estimation model can ensure the rapid response capability of the wind turbines based on their own frequency regulation responsibility and frequency conversion support capability, thereby overcoming the problem that when the existing method relies on long short-term memory networks to predict wind farm power, when faced with highly uncertain and noisy multi-source heterogeneous data, the dynamic changes in its weights and gating mechanisms result in poor multi-source heterogeneous data prediction and evaluation performance of the model.

[0082] The embodiment of the present invention provides a wind farm frequency control method based on real-time regulation capability dynamic estimation. Figure 1A schematic diagram of a wind farm frequency control method based on real-time dynamic estimation of regulation capability is shown. The wind farm frequency control method based on real-time dynamic estimation of regulation capability specifically includes:

[0083] Step S10, obtaining multi-source heterogeneous sequential wind farm operating parameters, standardizing the multi-source heterogeneous wind farm operating parameters, and uploading the standardized wind farm operating parameters to a station database;

[0084] It should be noted that the wind farm operating parameters include but are not limited to power grid frequency signals, wind farm environmental data, wind turbine power, wind turbine load data, wind turbine rotor speed, wind turbine rotor current and voltage data. In this embodiment, the method for standardizing the operating parameters of multi-source heterogeneous wind farms can adopt minimum-maximum value standardization processing, decimal calibration standardization processing or power function conversion processing.

[0085] In this embodiment, when the station database is constructed, the logic design is a multidimensional database structure, and the storage engine of the station database can be InnoDB or MyISAM. When the standardized wind farm operating parameters are continuously uploaded, the operating parameters are converted into historical operating parameters, thereby facilitating the construction of the model.

[0086] Step S20, traversing the station database, randomly extracting historical working parameters from the station database, building a frequency dynamic estimation model by combining fuzzy control algorithm and back propagation neural network, and iteratively training the frequency dynamic estimation model through historical working parameters;

[0087] Step S30, loading the sequential wind farm operating parameters, taking the sequential wind farm operating parameters as input, executing the frequency dynamic estimation model, the frequency dynamic estimation model identifies and analyzes the sequential wind farm operating parameters, and outputs the wind farm frequency prediction sequence.

[0088] Step S40, obtaining the wind farm frequency prediction sequence, calculating the frequency modulation support control parameters corresponding to the wind farm frequency prediction sequence based on the cuckoo algorithm combined with the inertia safety constraint, and the wind turbines executing the grid frequency active support instructions based on the frequency modulation support control parameters.

[0089] In an embodiment of the present invention, a frequency dynamic estimation model capable of predicting the wind farm power grid frequency is constructed based on multi-source heterogeneous sequential wind farm operating parameters. The construction and training of the frequency dynamic estimation model can ensure the rapid response capability of the wind turbines based on their own frequency regulation responsibility and frequency conversion support capability, thereby overcoming the problem that when the existing method relies on long short-term memory networks to predict wind farm power, when faced with highly uncertain and noisy multi-source heterogeneous data, the dynamic changes in its weights and gating mechanisms result in poor multi-source heterogeneous data prediction and evaluation performance of the model.

[0090] The embodiment of the present invention provides a method for constructing a frequency dynamic estimation model by combining a fuzzy control algorithm and a back propagation neural network. Figure 2 The present invention shows a schematic flow chart of a method for constructing a frequency dynamic estimation model by combining a fuzzy control algorithm and a back propagation neural network. The method for constructing a frequency dynamic estimation model by combining a fuzzy control algorithm and a back propagation neural network specifically includes:

[0091] Step S101, using a back propagation neural network model as an initial model of a frequency dynamic estimation model, wherein the initial model includes an input layer, a hidden layer, and an output layer, a fuzzy control mechanism is incorporated into the input layer of the initial model, and the number of neurons in the output layer, the hidden layer, and the output layer are 5, 4, and 4 respectively;

[0092] It should be noted that the back-propagation neural network model (BPNN) has significant advantages in parameter prediction. Through the design of multiple hidden layers, BPNN can capture deep features and patterns in the data, which is crucial to improving the accuracy of prediction.

[0093] Step S102, using a lightweight prediction structure Fire Module structure to improve the hidden layer of the initial model, setting the loss function and activation function of the hidden layer, and obtaining an improved initial model;

[0094] It should be noted that, in this embodiment, the loss function of the hidden layer may be a cross entropy loss function, and the activation function may be a logsig function.

[0095] Step S103, loading historical working parameters, decomposing the historical working parameters based on the singular spectrum decomposition method to obtain a parameter decomposition set, and dividing the parameter decomposition set into a training set and a test set in proportion. In this embodiment, the distribution ratio of the training set to the test set may be 7:3;

[0096] Step S104, setting initial weights and decision thresholds, optimizing the initial weights and decision thresholds based on a fuzzy control algorithm, and obtaining optimal weights and optimal thresholds;

[0097] For example, when optimizing the initial weight and the decision threshold based on the fuzzy control algorithm, a series of fuzzy rules can be formulated to describe the relationship between the input and the output according to expert experience or system characteristics. These rules are usually in the form of "if...then...", for example: "if the error is large and the error change rate is positive, then increase the weight adjustment amount."

[0098] Step S105, obtaining a training set, and the initial model uses the Levenberg-Marquardt algorithm to iteratively train the training set until the initial model converges, with a maximum number of training times of 800 times and an initial learning rate set to 0.01;

[0099] Step S106, obtaining a test set, and testing the evaluation accuracy of the initial model based on the test set. In this embodiment, the evaluation accuracy can be set to 0.95 or 0.97;

[0100] Step S107, if the evaluation accuracy meets the preset accuracy threshold, output the frequency dynamic estimation model;

[0101] If it does not meet the preset accuracy threshold, return to step S104.

[0102] In an embodiment of the present invention, a frequency dynamic estimation model and a construction and training method thereof are provided. The frequency dynamic estimation model is constructed by combining a fuzzy control algorithm and a back propagation neural network. The fuzzy control mechanism can effectively handle the uncertainty and noise in the wind farm parameters by fuzzifying the processing parameters, thereby enhancing the robustness of the model. The fuzzy control mechanism allows the model to dynamically adjust the prediction strategy according to real-time data. This adaptive capability makes BPNN more advantageous in dealing with rapidly changing wind farm environments. In addition, the use of a lightweight prediction structure, the Fire Module structure, to improve the hidden layer of the initial model can achieve efficient prediction through simplified calculation steps.

[0103] The embodiment of the present invention provides a method for identifying and analyzing the sequential wind farm operating parameters using a frequency dynamic estimation model and outputting a wind farm frequency prediction sequence. Figure 3 The present invention shows a method for realizing a method of identifying and analyzing the working parameters of a wind farm in a sequential manner using a frequency dynamic estimation model and outputting a wind farm frequency prediction sequence. The method of identifying and analyzing the working parameters of a wind farm in a sequential manner using a frequency dynamic estimation model and outputting a wind farm frequency prediction sequence specifically includes:

[0104] Step S201, obtaining the sequential wind farm operating parameters, and using the singular spectrum decomposition method to decompose the sequential wind farm operating parameters to obtain a parameter decomposition set;

[0105] It should be noted that in this embodiment, the singular spectrum decomposition method is used to decompose the time-series wind farm operating parameters, so that the parameter decomposition set can effectively separate the noise components in the signal, thereby improving the signal-to-noise ratio of the signal, and by removing the noise, the real fluctuation characteristics of the wind power time series can be more accurately captured. This is of great significance for subsequent analysis and prediction, because the interference of noise often affects the performance of the prediction model and reduces the accuracy of the prediction.

[0106] Among them, the parameter decomposition set is calculated by the following formula:

[0107] (1)

[0108] in, is the parameter decomposition vector, To use the singular spectrum decomposition method to decompose the time-series wind farm operating parameters, is the singular spectrum decomposition window length. In this embodiment, the singular spectrum decomposition window length may be 2-10. It is the sequential wind farm operating parameter;

[0109] (2)

[0110] in, is the data dimension of the sequential wind farm operating parameters. In this embodiment, the data dimension may be 1-8. is the initial weight of the sequential wind farm operating parameters. It should be noted that the initial weight of the sequential wind farm operating parameters can be set by normal distribution initialization, orthogonal initialization or LeCun initialization method. It is the time for collecting the working parameters of the sequential wind farm;

[0111] Step S202, loading the parameter decomposition set, the fuzzy control mechanism in the input layer constructs the parameter fuzzy set based on the parameter decomposition set, and calculates the prediction confidence of the parameter fuzzy set;

[0112] In this embodiment, the prediction confidence of the parameter fuzzy set is used to characterize the correlation between the parameter fuzzy set and the parameter scenario probability distribution. By calculating the prediction confidence, the parameters can be easily predicted.

[0113] It should be noted that when constructing parameter fuzzy sets, they are constructed based on confidence constraints at different deviation levels, and the confidence constraints are:

[0114] (3)

[0115] in, is the confidence clustering constraint, is the number of data of parameter fuzzy set, for frequency discrete scenarios, is the scene probability corresponding to the parameter decomposition set, is the initial scene probability value of the parameter decomposition set, is the upper limit of the total deviation of the scene probability distribution, The upper limit of the deviation of the probability distribution of the scene;

[0116] Among them, the prediction confidence of the parameter fuzzy set is calculated by the following formula:

[0117] (4)

[0118] in, is the prediction confidence of the parameter fuzzy set.

[0119] Step S203, the hidden layer obtains the parameter fuzzy set and the corresponding prediction confidence, and the lightweight prediction structure Fire Module structure in the hidden layer constructs a frequency parameter conversion function based on the parameter fuzzy set and the prediction confidence;

[0120] Among them, the frequency parameter conversion function is expressed as:

[0121] (5)

[0122] in, To predict the frequency conversion coefficient at the next moment, is a transformation coefficient of the frequency parameter conversion function. In this embodiment, the frequency parameter conversion function may be an S transformation function, and the transformation coefficient of the frequency parameter conversion function may be 0.1-0.3. is the number of data in the parameter fuzzy set, for frequency discrete scenarios, is the input value of the parameter fuzzy set;

[0123] Step S204, outputting the predicted frequency of the wind farm at the next moment based on the frequency conversion coefficient, integrating the predicted frequency of the wind farm, and outputting a wind farm frequency prediction sequence;

[0124] Among them, the wind farm prediction frequency is calculated by the following formula:

[0125] (6)

[0126] in, Predicting frequency for a wind farm, To predict the frequency conversion coefficient at the next moment.

[0127] In the embodiment of the present invention, the frequency dynamic estimation model can accurately identify and analyze the time-series wind farm operating parameters, thereby accurately outputting the wind farm frequency prediction sequence. Compared with the existing technology, it can dynamically adjust the prediction strategy according to real-time data. This adaptive capability makes the frequency dynamic estimation model more advantageous when dealing with rapidly changing wind farm environments.

[0128] The embodiment of the present invention provides a method for calculating the frequency modulation support control parameters corresponding to the wind farm frequency prediction sequence based on the cuckoo algorithm combined with the inertia safety constraint. Figure 4The present invention shows a schematic flow chart of a method for calculating the frequency modulation support control parameters corresponding to the wind farm frequency prediction sequence based on the cuckoo algorithm combined with the inertia safety constraint. The method for calculating the frequency modulation support control parameters corresponding to the wind farm frequency prediction sequence based on the cuckoo algorithm combined with the inertia safety constraint specifically includes:

[0129] Step S301, obtaining a wind farm frequency prediction sequence, and calculating the wind turbine constrained power corresponding to the wind farm frequency prediction sequence based on inertia safety constraints;

[0130] In this embodiment, the wind turbine generator set constrained power satisfies the following inertia safety constraints:

[0131] (7)

[0132] in, The frequency regulation period of the wind turbine is in seconds or milliseconds. is the actual power adjustment, The wind turbine power constraint during the frequency modulation period. They are the minimum power adjustment amount and the maximum power adjustment amount respectively;

[0133] (8)

[0134] in, For the previous moment Wind turbine constraint power, is the number of wind turbines. In this embodiment, the number of wind turbines may be 10-80. is the frequency regulation power coefficient of the wind turbine;

[0135] (9)

[0136] in, is the frequency modulation power coefficient of the wind turbine, is the angular velocity of the wind turbine rotor, is the blade radius of the wind turbine, is the number of wind turbine blades. In this embodiment, the number of wind turbine blades may be 3 groups. is the wind speed, is the wind turbine rotor moment of inertia.

[0137] Step S302, loading the constrained power of the wind turbine group, and optimizing the wind power contribution rate and frequency regulation capability factor parameters of the wind turbine group based on the cuckoo algorithm and the constrained power of the wind turbine group;

[0138] It should be noted that the cuckoo algorithm is a heuristic optimization algorithm that simulates the breeding behavior of cuckoos to find the optimal solution. The cuckoo algorithm uses Levy flights for global search, which can effectively explore the solution space and avoid falling into the local optimum. As an effective optimization tool, the cuckoo algorithm has been widely used in many fields. Its simplicity and powerful global search capabilities make it a strong candidate for solving complex optimization problems.

[0139] Among them, the optimization function of wind power contribution rate is expressed as:

[0140] (10)

[0141] in, The wind power contribution rate is the wind power contribution rate. It should be noted that the wind power contribution rate is a representation of the frequency active support contribution capacity of a single wind turbine group. is the weight coefficient of the current wind turbine. The weight coefficient of the current wind turbine can also be set by normal distribution initialization, orthogonal initialization or LeCun initialization method. To find the optimal constant, in this embodiment, the optimal constant is greater than 2. is the distance between the current wind turbine and the central unit of the wind farm, is the spatial radius of the wind farm center, is the distance between the current wind turbine and the center of the wind farm in terms of spatial radius;

[0142] The optimization function of frequency modulation capability factor is expressed as:

[0143] (11)

[0144] in, is the frequency modulation capability factor.

[0145] Step S303, obtaining the optimal wind power contribution rate and frequency regulation capability factor of the wind turbine generator set, and calculating the frequency regulation responsibility coefficient corresponding to the wind turbine generator set based on the wind power contribution rate and the frequency regulation capability factor;

[0146] It should be noted that the frequency modulation responsibility coefficient can be calculated by using a fuzzy control algorithm or a particle swarm algorithm.

[0147] Step S304, obtaining the frequency regulation responsibility coefficient, generating a primary frequency regulation and reactive voltage regulation instruction based on the frequency regulation responsibility coefficient, and the wind turbine generator system responds to the primary frequency regulation and reactive voltage regulation instruction and executes the grid frequency active support instruction.

[0148] In an embodiment of the present invention, the frequency regulation support control parameters corresponding to the wind farm frequency prediction sequence are calculated by combining the cuckoo algorithm with the inertia safety constraint. As an effective optimization tool, the cuckoo algorithm can play a key role in calculating the frequency regulation support control parameters. By simulating the parasitic behavior of the cuckoo, the cuckoo algorithm can effectively search for the optimal solution of the wind power contribution rate and the frequency regulation capability factor, thereby optimizing the frequency regulation responsibility coefficient of the wind farm. This optimization helps to improve the performance of wind turbines in the wind farm in the grid frequency regulation and ensure the stable operation of the power system.

[0149] The embodiment of the present invention provides a wind farm frequency control system based on real-time regulation capability dynamic estimation. Figure 5 A schematic diagram of the control principle of a wind farm frequency control system based on real-time dynamic estimation of regulation capability is shown. The wind farm frequency control system based on real-time dynamic estimation of regulation capability specifically includes:

[0150] The parameter acquisition module 100 is used to acquire the time-sequential wind farm operating parameters of the multi-source heterogeneous wind farm, standardize the multi-source heterogeneous wind farm operating parameters, and upload the standardized wind farm operating parameters to the station database;

[0151] The frequency dynamic estimation module 200 is used to traverse the station database, randomly extract historical working parameters from the station database, build a frequency dynamic estimation model by combining fuzzy control algorithm and back propagation neural network, and iteratively train the frequency dynamic estimation model through historical working parameters;

[0152] The frequency prediction module 300 is used to load the sequential wind farm operating parameters, take the sequential wind farm operating parameters as input, execute the frequency dynamic estimation model, the frequency dynamic estimation model identifies and analyzes the sequential wind farm operating parameters, and outputs the wind farm frequency prediction sequence;

[0153] The support control module 400 is used to obtain the wind farm frequency prediction sequence, calculate the frequency support control parameters corresponding to the wind farm frequency prediction sequence based on the cuckoo algorithm combined with the inertia safety constraint, and the wind turbine executes the grid frequency active support instruction based on the frequency support control parameters.

[0154] It should be noted that the parameter acquisition module 100, the frequency dynamic estimation module 200, the frequency prediction module 300, and the support control module 400 realize data interaction through a local area network or a DTU communication connection. It can be understood that the wind farm frequency control system based on real-time adjustment capability dynamic estimation provided in the embodiment of the present invention corresponds to the above-mentioned wind farm frequency control method based on real-time adjustment capability dynamic estimation. The explanations, examples, beneficial effects and other parts of the relevant contents can refer to the corresponding contents in the wind farm frequency control method based on real-time adjustment capability dynamic estimation, and will not be repeated here.

[0155] In this embodiment, the support control module 400 includes:

[0156] The power prediction unit 410 is used to obtain a wind farm frequency prediction sequence, and calculate the wind turbine constrained power corresponding to the wind farm frequency prediction sequence based on the inertia safety constraint;

[0157] The parameter optimization unit 420 is used to load the constrained power of the wind turbine group and optimize the wind power contribution rate and frequency regulation capability factor parameters of the wind turbine group based on the cuckoo algorithm and the constrained power of the wind turbine group;

[0158] The responsibility coefficient calculation unit 430 is used to obtain the optimal wind power contribution rate and frequency regulation capability factor of the wind turbine generator set, and calculate the frequency regulation responsibility coefficient corresponding to the wind turbine generator set based on the wind power contribution rate and the frequency regulation capability factor;

[0159] The support instruction execution unit 440 is used to obtain the frequency regulation responsibility coefficient, generate primary frequency regulation and reactive voltage regulation instructions based on the frequency regulation responsibility coefficient, and the wind turbine executes the grid frequency active support instruction in response to the primary frequency regulation and reactive voltage regulation instructions.

[0160] On the other hand, an embodiment of the present invention further provides a computer-readable storage medium, which stores computer program instructions, and the computer program instructions can be executed by a processor. When the computer program instructions are executed, the method of any of the above embodiments is implemented.

[0161] The computer-readable storage medium (e.g., memory) herein may be a volatile memory or a nonvolatile memory, or may include both a volatile memory and a nonvolatile memory. As an example and not by way of limitation, a nonvolatile memory may include a read-only memory (ROM), a programmable ROM (PROM), an electrically programmable ROM (EPROM), an electrically erasable programmable ROM (EEPROM), or a flash memory. A volatile memory may include a random access memory (RAM), which may act as an external cache memory. As an example and not by way of limitation, RAM may be obtained in a variety of forms, such as synchronous RAM (DRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), and direct Rambus RAM (DRRAM). The storage devices of the disclosed aspects are intended to include, but are not limited to, these and other suitable types of memory.

[0162] In summary, the present invention provides a wind farm frequency control method and system based on dynamic estimation of real-time regulation capability. In an embodiment of the present invention, a frequency dynamic estimation model capable of predicting the wind farm power grid frequency is constructed based on multi-source heterogeneous sequential wind farm operating parameters. The construction and training of the frequency dynamic estimation model can ensure the rapid response capability of the wind turbines based on their own frequency regulation responsibilities and frequency conversion support capabilities, thereby overcoming the problem that when the existing method relies on long short-term memory networks to predict wind farm power, when faced with highly uncertain and noisy multi-source heterogeneous data, the dynamic changes in its weights and gating mechanisms make the model's multi-source heterogeneous data prediction and evaluation performance poor.

[0163] It should be noted that, for the above-mentioned embodiments, for the sake of simplicity, they are all described as a series of action combinations, but those skilled in the art should know that the present invention is not limited by the described order of actions, because according to the present invention, some steps may be performed in other orders or simultaneously. Secondly, those skilled in the art should also know that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by the present invention.

[0164] In the several embodiments provided in the present application, it should be understood that the disclosed device can be implemented in other ways. For example, the device embodiments described above are only schematic, such as the division of the above-mentioned units, which is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or communication connection shown or discussed can be through some interfaces, and the indirect coupling or communication connection between devices or units can be in the form of telecommunication or other forms.

[0165] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the scope of protection of the invention. Obviously, the described embodiments are only some embodiments of the present invention, rather than all embodiments. Based on these embodiments, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in this field can still combine, add, delete or make other adjustments to the features in the various embodiments of the present invention according to the circumstances without conflict, without making creative work, so as to obtain different other technical solutions that do not deviate from the concept of the present invention in essence, and these technical solutions also belong to the scope of protection of the present invention.

Claims

1. Wind farm frequency control method based on real-time regulation capability dynamic estimation, It is characterized in that The wind farm frequency control method based on real-time regulation capability dynamic estimation includes: Acquire multi-source heterogeneous sequential wind farm operating parameters, wherein the wind farm operating parameters include power grid frequency signals, wind farm environmental data, wind turbine power, wind turbine load data, wind turbine rotor speed, wind turbine rotor current and voltage data, standardize the multi-source heterogeneous wind farm operating parameters, and upload the standardized wind farm operating parameters to the station database; Traversing the station database, randomly extracting historical working parameters from the station database, building a frequency dynamic estimation model by combining fuzzy control algorithm and back propagation neural network, and iteratively training the frequency dynamic estimation model through historical working parameters; Load the sequential wind farm operating parameters, take the sequential wind farm operating parameters as input, execute the frequency dynamic estimation model, the frequency dynamic estimation model identifies and analyzes the sequential wind farm operating parameters, and outputs the wind farm frequency prediction sequence; Obtain the wind farm frequency prediction sequence, calculate the frequency regulation support control parameters corresponding to the wind farm frequency prediction sequence based on the cuckoo algorithm combined with the inertia safety constraint, and the wind turbines execute the grid frequency active support instructions based on the frequency regulation support control parameters; The method of identifying and analyzing the time-series wind farm operating parameters with the frequency dynamic estimation model and outputting the wind farm frequency prediction sequence specifically includes: Obtaining the sequential wind farm operating parameters, decomposing the sequential wind farm operating parameters using the singular spectrum decomposition method, and obtaining a parameter decomposition set; Among them, the parameter decomposition set is calculated by the following formula: （1） in, is the parameter decomposition vector, To use the singular spectrum decomposition method to decompose the time-series wind farm operating parameters, is the singular spectrum decomposition window length, It is the sequential wind farm operating parameter; （2） in, is the data dimension of the time-series wind farm operating parameters, is the initial weight of the sequential wind farm operating parameters, It is the time for collecting the working parameters of the sequential wind farm; Load the parameter decomposition set, the fuzzy control mechanism in the input layer constructs the parameter fuzzy set based on the parameter decomposition set, and calculates the prediction confidence of the parameter fuzzy set; The hidden layer obtains the parameter fuzzy set and the corresponding prediction confidence. The lightweight prediction structure FireModule structure in the hidden layer constructs a frequency parameter conversion function based on the parameter fuzzy set and prediction confidence. Among them, the frequency parameter conversion function is expressed as: （5） in, To predict the frequency conversion coefficient at the next moment, is the transformation coefficient of the frequency parameter conversion function, is the number of data in the parameter fuzzy set, for frequency discrete scenarios, is the input value of the parameter fuzzy set; Outputting the predicted frequency of the wind farm at the next moment based on the frequency conversion coefficient, integrating the predicted frequency of the wind farm, and outputting a wind farm frequency prediction sequence; Among them, the wind farm prediction frequency is calculated by the following formula: （6） in, Predict frequency for wind farms, To predict the frequency conversion coefficient at the next moment; The method for calculating the frequency regulation support control parameters corresponding to the wind farm frequency prediction sequence based on the cuckoo algorithm combined with the inertia safety constraint specifically includes: Obtain a wind farm frequency prediction sequence, and calculate the wind turbine constrained power corresponding to the wind farm frequency prediction sequence based on inertia safety constraints; Load the constrained power of the wind turbine set, and optimize the wind power contribution rate and frequency regulation capability factor parameters of the wind turbine set based on the cuckoo algorithm and the constrained power of the wind turbine set; Among them, the optimization function of wind power contribution rate is expressed as: （10） in, is the contribution rate of wind power, is the weight coefficient of the current wind turbine, To find the optimal constant, is the distance between the current wind turbine and the central unit of the wind farm, is the spatial radius of the wind farm center, is the distance between the current wind turbine and the center of the wind farm in terms of spatial radius; The optimization function of frequency modulation capability factor is expressed as: （11） in, is the frequency modulation capability factor; Obtain the optimal wind power contribution rate and frequency regulation capability factor of the wind turbine, and calculate the frequency regulation responsibility coefficient corresponding to the wind turbine based on the wind power contribution rate and frequency regulation capability factor; Obtaining a frequency regulation responsibility coefficient, generating a primary frequency regulation and reactive voltage regulation instruction based on the frequency regulation responsibility coefficient, and the wind turbine generator set responds to the primary frequency regulation and reactive voltage regulation instruction and executes the grid frequency active support instruction; When constructing parameter fuzzy sets, they are constructed based on confidence constraints at different deviation levels. The confidence constraints are: （3） in, is the confidence clustering constraint, is the number of data of parameter fuzzy set, for frequency discrete scenarios, is the scene probability corresponding to the parameter decomposition set, is the initial scene probability value of the parameter decomposition set, is the upper limit of the total deviation of the scene probability distribution, The upper limit of the deviation of the probability distribution of the scene; Among them, the prediction confidence of the parameter fuzzy set is calculated by the following formula: （4） in, is the prediction confidence of the parameter fuzzy set.

2. The wind farm frequency control method based on real-time regulation capability dynamic estimation according to claim 1, Features: The method of combining the fuzzy control algorithm and the back propagation neural network to construct a frequency dynamic estimation model specifically includes: The back propagation neural network model is used as the initial model of the frequency dynamic estimation model, wherein the initial model includes an input layer, a hidden layer, and an output layer. A fuzzy control mechanism is incorporated into the input layer of the initial model, and the number of neurons in the output layer, hidden layer, and output layer are 5, 4, and 4 respectively; The lightweight prediction structure Fire Module structure is used to improve the hidden layer of the initial model, and the loss function and activation function of the hidden layer are set to obtain the improved initial model; Load the historical working parameters, decompose the historical working parameters based on the singular spectrum decomposition method, obtain the parameter decomposition set, and divide the parameter decomposition set into a training set and a test set according to the proportion; Set the initial weight and judgment threshold, optimize the initial weight and judgment threshold based on the fuzzy control algorithm, and obtain the optimal weight and optimal threshold; Obtain the training set. The initial model uses the Levenberg-Marquardt algorithm to iteratively train the training set until the initial model converges. The maximum number of training times is 800, and the initial learning rate is set to 0.

01. A test set is obtained, and the evaluation accuracy of the initial model is tested based on the test set. If the evaluation accuracy meets a preset accuracy threshold, a frequency dynamic estimation model is output.

3. The wind farm frequency control method based on real-time regulation capability dynamic estimation according to claim 2, Features: The wind turbine power constraint meets the following inertia safety constraints: （7） in, This is the frequency regulation period of wind turbines. is the actual power adjustment, The wind turbine power constraint during the frequency modulation period. They are the minimum power adjustment amount and the maximum power adjustment amount respectively; （8） in, For the previous moment Wind turbine constrained power, is the number of wind turbines, is the frequency regulation power coefficient of the wind turbine; （9） in, is the frequency modulation power coefficient of the wind turbine, is the angular velocity of the wind turbine rotor, is the blade radius of the wind turbine, is the number of wind turbine blades, is the wind speed, is the wind turbine rotor moment of inertia.

4. A wind farm frequency control system based on real-time regulation capability dynamic estimation, used to implement a wind farm frequency control method based on real-time regulation capability dynamic estimation as claimed in any one of claims 1 to 3, Features: The wind farm frequency control system based on real-time regulation capability dynamic estimation specifically includes: The parameter acquisition module is used to acquire the time-sequential wind farm operating parameters of the multi-source heterogeneous wind farm, standardize the multi-source heterogeneous wind farm operating parameters, and upload the standardized wind farm operating parameters to the station database; A frequency dynamic estimation module is used to traverse the station database, randomly extract historical working parameters from the station database, build a frequency dynamic estimation model by combining fuzzy control algorithm and back propagation neural network, and iteratively train the frequency dynamic estimation model through historical working parameters; The frequency prediction module is used to load the sequential wind farm operating parameters, take the sequential wind farm operating parameters as input, execute the frequency dynamic estimation model, the frequency dynamic estimation model identifies and analyzes the sequential wind farm operating parameters, and outputs the wind farm frequency prediction sequence; The support control module is used to obtain the wind farm frequency prediction sequence, calculate the frequency support control parameters corresponding to the wind farm frequency prediction sequence based on the cuckoo algorithm combined with inertia safety constraints, and the wind turbines execute the grid frequency active support instructions based on the frequency support control parameters.

5. The wind farm frequency control system based on real-time regulation capability dynamic estimation as claimed in claim 4, Features: The support control module comprises: A power prediction unit, used to obtain a wind farm frequency prediction sequence, and calculate the wind turbine constrained power corresponding to the wind farm frequency prediction sequence based on inertia safety constraints; A parameter optimization unit is used to load the constrained power of the wind turbine set and optimize the wind power contribution rate and frequency regulation capability factor parameters of the wind turbine set based on the cuckoo algorithm and the constrained power of the wind turbine set; A responsibility coefficient calculation unit is used to obtain the optimal wind power contribution rate and frequency regulation capability factor of the wind turbine generator set, and calculate the frequency regulation responsibility coefficient corresponding to the wind turbine generator set based on the wind power contribution rate and the frequency regulation capability factor; The support instruction execution unit is used to obtain the frequency regulation responsibility coefficient, generate primary frequency regulation and reactive voltage regulation instructions based on the frequency regulation responsibility coefficient, and the wind turbine group responds to the primary frequency regulation and reactive voltage regulation instructions to execute the grid frequency active support instructions.

Images