Control method and system of wind turbine generator, storage medium and electronic equipment

Abstract

The invention provides a wind turbine generator control method and system, a storage medium and electronic equipment, and relates to the technical field of wind turbine generator energy-saving control. The method comprises the following steps: setting simulation parameters of a wind turbine generator, and constructing a generator dynamical model; historical wind speed data of the wind turbine generator and real-time operation state data of the wind turbine generator are obtained; inputting the historical wind speed data and the real-time operation state data into a pre-constructed wind speed prediction model, wherein the wind speed prediction model outputs wind speed prediction data; and inputting the wind speed prediction data as feedforward data and the real-time operation state data as feedback data into a pre-constructed control model, wherein the control model is combined with the unit kinetic model to obtain an optimal control instruction sequence for controlling the wind turbine generator. According to the scheme, the wind condition can be pre-judged in advance, and the operation state of the wind turbine generator is adjusted in advance based on the wind speed prediction data to improve the wind energy capture efficiency.

Description

Technical Field

[0001] This application relates to the field of wind turbine energy-saving control technology, specifically to a control method, system, storage medium, and electronic equipment for wind turbines. Background Technology

[0002] Traditional control strategies for large wind turbines rely on real-time wind speed as input, combined with the turbine's operating status. However, wind conditions in wind farms change in real time, and controlling turbines solely based on real-time wind speed is inherently lagging. Therefore, current control strategies for large wind turbines lack accurate wind speed prediction capabilities, making it difficult to adapt to changes in operating conditions in advance, resulting in low energy capture efficiency. Summary of the Invention

[0003] The technical problem to be solved by this application is that the existing control schemes for wind turbines are difficult to adapt to changes in wind conditions in advance, resulting in low wind energy utilization efficiency. Therefore, this application provides a control method, system, storage medium and electronic equipment for wind turbines.

[0004] Firstly, the technical solution of this application provides a control method for a wind turbine generator, including: Set the simulation parameters for the wind turbine and construct the turbine dynamics model; Acquire the historical wind speed data and real-time operating status data of the wind turbine; The historical wind speed data and the real-time operating status data are input into a pre-built wind speed prediction model, and the wind speed prediction model outputs wind speed prediction data. The wind speed prediction data is used as feedforward data, and the real-time operating status data is used as feedback data. The data is then input into a pre-built control model, and the control model, combined with the unit dynamics model, yields the optimal control command sequence for controlling the wind turbine unit.

[0005] Preferably, in the wind turbine control method, the wind speed prediction model is a TCN-GRU combined model, and the step of inputting the historical wind speed data and the real-time operating status data into the pre-constructed wind speed prediction model, and the wind speed prediction model outputting wind speed prediction data, includes: The historical wind speed data and the real-time operating status data are processed into a time-series input sequence; The TCN The GRU combined model includes a TCN module and a GRU module; the TCN module receives the time-series input sequence and processes the time-series input sequence into a multi-dimensional time-series feature vector; the GRU module receives the multi-dimensional time-series feature vector and generates a wind speed prediction sequence for several future control cycles based on the multi-dimensional time-series feature vector, and the wind speed prediction sequence serves as the wind speed prediction data.

[0006] Preferably, in the wind turbine control method, the control model is an MPC architecture model. The wind speed prediction data is used as feedforward data, and the real-time operating status data is used as feedback data, which are input into a pre-constructed control model. The control model, combined with the turbine dynamics model, obtains the optimal control command sequence for controlling the wind turbine, including: The MPC architecture model receives the wind speed prediction sequence as a feedforward input and the real-time operating status data as a feedback input. Set constraints based on the simulation parameters; Calculate the difference between the real-time operating status data and the preset target status data; Based on the wind speed prediction sequence, the difference, the constraints, and the unit dynamics model, an optimal control command sequence for controlling the wind turbine is obtained. The optimal control command sequence is used to ensure that the real-time operating status prediction data of the wind turbine in the next control cycle is consistent with the preset target status data.

[0007] Preferably, in the control method for the wind turbine generator, the generator dynamics model includes the rotor motion equations of the generator and the rotor equations of relevant variables: The rotor motion equation of the generator: ;in, Indicates the moment of inertia of the rotor. Represents mechanical torque. Indicates electromagnetic torque. Indicates the damping coefficient. This represents the actual angular velocity of the rotor. Indicates synchronous mechanical angular velocity, t Indicates time; Rotor equations for relevant variables: ;in, Indicates mechanical power. Indicates electromagnetic power. Indicates the power grid frequency. Indicates the number of pole pairs of the generator; Control function: , ;in, This represents the wind speed prediction sequence at time k. This represents the wind speed prediction sequence at time k-1. This represents the increment of the input sequence between two adjacent time points; This represents the sequence of control commands output at time k. This represents the sequence of control commands output at time k-1, where A and B represent coefficient matrices. Objective function: ;in, This represents the frequency deviation weighting coefficient. This represents the difference in angular frequency predicted at time k+i from time k. This represents the control increment weighting coefficient. Let M represent the increment of the input sequence at time k+i predicted at time k, where M represents the control time domain and N represents the prediction time domain. Constraints: ;in, , Indicates the upper and lower limits of the input sequence. This represents the wind speed prediction sequence for time k+i predicted at time k. , Indicates the upper and lower limits of the input sequence increment. , This represents the upper and lower limits of the output control vector. This represents the sequence of control instructions predicted at time k+i in the future.

[0008] Preferably, in the control method for the wind turbine generator, the optimal control command sequence includes a pitch angle adjustment command and an excitation current adjustment command, wherein: The blade pitch angle adjustment command is used to control the blade pitch angle adjustment. The excitation current regulation command is used to control the generator excitation current.

[0009] Preferably, the control method for the wind turbine generator further includes: The wind turbine is controlled according to the optimal control command sequence to obtain the updated operating status data of the wind turbine. If the updated operating status data is consistent with the expected operating status data, then the optimal control command sequence passes verification.

[0010] Secondly, the technical solution of this application provides a control system for a wind turbine generator, comprising: The parameter setting module is configured to set the simulation parameters of the wind turbine and build the dynamic model of the turbine. The data acquisition module is configured to acquire historical wind speed data and real-time operating status data of the wind turbine. The wind speed prediction module is configured to input the historical wind speed data and the real-time operating status data into a pre-built wind speed prediction model, and the wind speed prediction model outputs wind speed prediction data. The instruction generation module is configured to input the wind speed prediction data as feedforward data and the real-time operating status data as feedback data into a pre-constructed control model. The control model, combined with the unit dynamics model, obtains the optimal control instruction sequence for controlling the wind turbine unit.

[0011] Thirdly, the present application provides a computer-readable storage medium storing program information, wherein a computer reads the program information and executes the steps of the wind turbine control method described in any one of the first aspects.

[0012] Fourthly, the present application provides a computer program product, including a computer program / instructions, which, when executed by a processor, implement the steps of the wind turbine control method described in any of the first aspects.

[0013] Fifthly, the present application provides an electronic device, including a memory, a processor, and a computer program stored in the memory, wherein the processor executes the computer program to implement the steps of the wind turbine control method described in any of the first aspects.

[0014] The technical solution provided in this application has the following technical effects compared with the prior art: The wind turbine control method, system, storage medium, and electronic equipment provided in this application construct a turbine dynamics model based on the simulation parameters of the controlled wind turbine. By introducing a wind speed prediction model, wind speed prediction data is obtained by predicting future wind conditions based on historical wind speed data and real-time operating status data. The wind speed prediction data is used as feedforward data, and the real-time operating status data is used as feedback data, which is then input into the pre-constructed control model. The control model, combined with the turbine dynamics model, solves for the optimal control command sequence. This application's solution can predict wind conditions in advance and adjust the wind turbine's operating status based on wind speed prediction data to improve wind energy capture efficiency. Attached Figure Description

[0015] Figure 1 This is a flowchart of a wind turbine control method according to one embodiment of this application; Figure 2 This is a structural block diagram of the wind speed prediction model and control model according to one embodiment of this application; Figure 3 This is a schematic diagram of the framework of a wind speed prediction model according to an embodiment of this application; Figure 4 This is a schematic diagram illustrating the structural principle of the control model described in one embodiment of this application; Figure 5 This is a structural block diagram of the control system of a wind turbine generator according to one embodiment of this application; Figure 6 This is a schematic diagram of the hardware connection relationship of an electronic device that performs the wind turbine control method according to an embodiment of this application. Detailed Implementation

[0016] The specific embodiments of this application will be further described below with reference to the accompanying drawings.

[0017] It is readily understood that, based on the technical solution of this application, various structural and implementation methods can be interchanged by those skilled in the art without altering the essential spirit of this application. Therefore, the following detailed embodiments and accompanying drawings are merely illustrative examples of the technical solution of this application and should not be considered as the entirety of this application or as limitations or restrictions on the technical solution of the application.

[0018] This application provides a control method for a wind turbine generator, applied to a computer system, such as... Figure 1 As shown, the method includes: S100: Set the simulation parameters of the wind turbine and build the dynamic model of the unit.

[0019] In this step, core simulation parameters are set based on the hardware parameters and design specifications of the wind turbine, covering the core fundamental parameters of the wind turbine, the power grid, and the direct-drive wind turbine generator. Specifically, the simulation parameters include the rotor radius (e.g., 35.25m), rotor swept area (e.g., 3904m²), and rotor moment of inertia (e.g., 2.4×10⁻⁶). 6 kg m²), generator rotor rotational inertia (e.g., 70kg) m², air density (e.g., 1.2 kg / m³), number of generator pole pairs (e.g., 3), grid rated frequency (e.g., 50 Hz), unit rated power (e.g., 1.5 MW), damping coefficient (e.g., 50 N). m (s / rad), pitch angle adjustment range (e.g., 0°~90°), and generator torque constraint (e.g., 0~12000 N) Based on the above simulation parameters, a unit dynamics model is established. The unit dynamics model includes rotor motion equations and related variable relationship equations, which describe the mechanical motion and energy conversion laws of the unit through a set of mathematical equations.

[0020] As a preferred embodiment, the unit dynamics model includes the rotor motion equations of the generator and the rotor equations of relevant variables: The rotor motion equation of the generator: ;in, Indicates the moment of inertia of the rotor. Represents mechanical torque. Indicates electromagnetic torque. Indicates the damping coefficient. This represents the actual angular velocity of the rotor. Indicates synchronous mechanical angular velocity, t Indicates time; Rotor equations for relevant variables: ;in, Indicates mechanical power. Indicates electromagnetic power. Indicates the power grid frequency. This indicates the number of pole pairs of the generator.

[0021] Furthermore, when the grid load fluctuates, the wind turbine generator speed will change accordingly. Grid frequency fluctuations can be suppressed by adjusting the rotor speed. By adjusting the generator's input active power to stabilize the grid frequency, the above equations can be rewritten as a discrete model: In the formula, This represents the change in angular frequency; and denoted by , where represents the changes in mechanical torque and electromagnetic torque of the wind turbine; k represents the current control cycle. The system sampling time is 1 second.

[0022] S200: Obtain historical wind speed data and real-time operating status data of the wind turbine.

[0023] In this step, historical wind speed data is collected using wind speed sensors installed in the wind turbine. This historical wind speed data includes a time-series wind speed sequence over a preset time period (e.g., the most recent 72 hours), and the sampling frequency can be consistent with the control cycle (e.g., 10Hz). Real-time operating status data is collected through the turbine's own condition monitoring system. This real-time operating status data includes key state quantities such as generator real-time speed, current pitch angle, output power, grid real-time frequency, and drivetrain component real-time load. The collection cycle is synchronized with the control cycle to ensure the timeliness and accuracy of the data. The collected historical wind speed data and real-time operating status data are preprocessed, including outlier removal, data standardization, and missing value completion, to obtain clean and usable input data.

[0024] S300: Input the historical wind speed data and the real-time operating status data into the pre-built wind speed prediction model, and the wind speed prediction model outputs wind speed prediction data.

[0025] Specifically, this step concatenates the preprocessed historical wind speed data and real-time operational status data into a one-dimensional time-series input sequence to ensure the temporal correlation of the data. Wind speed prediction data is then obtained using methods such as TCN (Temporal Convolutional Network) and GRU (Gated Recurrent Unit). TCN and GRU are pre-trained using one year's worth of historical wind speed time-series data from a wind farm; for example, the sampling interval of the historical wind speed time-series data is 1 second, the training batch size is 128, and the training epochs are 100. The model outputs wind speed prediction sequences for the next 10 control cycles, with a sequence resolution of 1 second per sequence.

[0026] S400: The wind speed prediction data is used as feedforward data, and the real-time operating status data is used as feedback data. The data is input into the pre-built control model, and the control model is combined with the unit dynamics model to obtain the optimal control command sequence for controlling the wind turbine.

[0027] In this step, the control model inputs wind speed prediction data and real-time operating status data into the unit dynamics model to construct system state prediction results for the next 10 cycles. The control model can generate the optimal control command sequence for the next 10 control cycles based on pre-set objective functions such as pitch angle adjustment range of 0-90°, generator power fluctuation amplitude ≤ ±5%, generator torque range: [0, 12000] N·m, and grid frequency fluctuation range: [49.5, 50.5] Hz as constraints, for controlling the wind turbine unit.

[0028] In the above embodiment, a dynamic model of the wind turbine is constructed based on the simulation parameters of the controlled wind turbine. By introducing a wind speed prediction model, wind speed prediction data is obtained by predicting future wind conditions based on historical wind speed data and real-time operating status data. The wind speed prediction data is used as feedforward data, and the real-time operating status data is used as feedback data and input into the pre-constructed control model. The control model, combined with the wind turbine dynamic model, solves for the optimal control command sequence. This application's solution can predict wind conditions in advance and adjust the operating status of the wind turbine in advance based on wind speed prediction data to improve wind energy capture efficiency.

[0029] Preferably, such as Figure 2 As shown, in the above scheme, the wind speed prediction model is a TCN-GRU combined model. The historical wind speed data and the real-time operating status data are input into the pre-constructed wind speed prediction model, and the wind speed prediction model outputs wind speed prediction data, including: The historical wind speed data and the real-time operating status data are processed into a time-series input sequence. Specifically, after outlier removal, standardization, and missing value completion processing of the historical wind speed data and the real-time operating status data, they are spliced ​​together into a one-dimensional time-series input sequence according to the time dimension to ensure the temporal correlation and integrity of the data.

[0030] The TCN The GRU combined model includes a TCN module and a GRU module; the TCN module receives the time-series input sequence and processes the time-series input sequence into a multi-dimensional time-series feature vector; the GRU module receives the multi-dimensional time-series feature vector and generates a wind speed prediction sequence for several future control cycles based on the multi-dimensional time-series feature vector, and the wind speed prediction sequence serves as the wind speed prediction data.

[0031] The TCN module consists of causal convolution, dilated convolution, and residual connections. In this embodiment, it combines... Figure 3 The diagram shows the framework of the wind speed prediction model. The parameter configuration of the TCN module is shown in Table 1. Table 1. Parameter Configuration of TCN Module

[0032] As shown in Table 1, the TCN module in this embodiment: Causal convolution: The kernel size is set to 3, and a unidirectional structure is adopted. The data at time t is only related to the data at time t and previous times in the previous layer. By using zero padding length equal to the kernel radius, the length of the input and output sequences is ensured to be consistent.

[0033] Dilated convolution: By adding a dilation factor, the receptive field of the convolutional layer can be expanded in causal convolution. The dilation factor coefficient takes values ​​of [1, 2, 4], increasing layer by layer. The expression for dilated causal convolution is shown in the following formula: ; In the above formula, The input is a one-dimensional time series sequence. Represents the expansion factor coefficient. K Indicates the size of the convolution kernel. k dilated convolution kernel, Indicates the convolution kernel number 1 One element, This represents an element in the output multidimensional time-series feature vector. This represents the dilated convolution operator.

[0034] Residual Connections: A stacked connection method is used with three residual blocks to address gradient vanishing, overfitting, and performance degradation issues caused by increasing the number of network layers. Each residual block contains a dilated causal convolution, weight normalization, a ReLU activation function, a Dropout layer, and a one-dimensional convolutional network. The random deactivation probability of Dropout is set to 0.2, and L2 normalization is used for weight normalization. The residual block operation expression is shown in the following formula: ; In the formula, Represents the input sequence. For weight normalization operation, This is a one-dimensional convolution operation. This is the output for the residual block.

[0035] Based on the TCN module above, after receiving the temporal input sequence, it extracts deep temporal features and long-distance dependencies in the sequence through dilated causal convolution operations with a kernel size of 3 and dilation factors of 1, 2, and 4 respectively. After L2 weight normalization, ReLU activation function and Dropout regularization, it outputs a multidimensional temporal feature vector with a preset dimension (such as 64 dimensions), realizing feature dimensionality reduction and redundant information removal.

[0036] Furthermore, the GRU module is a simplified model of LSTM, with a relatively simple structure containing only two gating mechanisms: reset gates and update gates. In this embodiment, the GRU module has 3 network layers, and the activation function is tanh; a bidirectional GRU structure is used. The specific calculation formula is as follows: Reset door calculation: ; Update gate calculation: ; Candidate state calculation: ; Output state calculation: ; In the above implementation formulas for the GPU module, and These represent the input information at the current moment and the hidden information at the previous moment, respectively. express function, and To update the door The weight, and To reset the door The weight, and Candidate state information The weight, To reset the door's state, This represents the Hadmard product. This indicates the output information.

[0037] In this embodiment, combined with Figure 3 As shown in Table 2, the parameter configuration of the GRU module is as follows.

[0038] Table 2 Parameter Configuration of GRU Module

[0039] Preferably, as shown in the figure, the GRU module may further include a fully connected layer: a Dense layer, the parameter settings of which are shown in Table 3: Table 3 Parameter Configuration of Dense Layer

[0040] The Dense layer is set after the GRU module and is used to map the high-dimensional temporal features output by the GRU module to a preset output dimension to complete the output of wind speed prediction values.

[0041] Furthermore, such as Figure 2 and Figure 4As shown, the control model is an MPC architecture model. The wind speed prediction data is used as feedforward data, and the real-time operating status data is used as feedback data, inputting these into a pre-built control model. The control model, combined with the turbine dynamics model, obtains the optimal control command sequence for controlling the wind turbine. This includes: the MPC architecture model receiving the wind speed prediction sequence as feedforward input and the real-time operating status data as feedback input; setting constraints according to the simulation parameters; calculating the difference between the real-time operating status data and the preset target state data; and obtaining the optimal control command sequence for controlling the wind turbine based on the wind speed prediction sequence, the difference, the constraints, and the turbine dynamics model. This optimal control command sequence ensures that the real-time operating status prediction data of the wind turbine in the next control cycle remains consistent with the preset target state data. Specifically, the MPC architecture model receives the wind speed prediction sequence as feedforward input and the real-time operating status data as feedback input, while simultaneously calling the pre-built turbine dynamics model; it sets constraints according to the simulation parameters, including control input constraints, control increment constraints, and output constraints, to ensure that the control commands are within the allowable range of the turbine hardware. The difference between the real-time operating status data and the preset target status data is calculated to obtain core control deviations such as frequency deviation, power deviation, and load deviation. Based on the wind speed prediction sequence, the difference, and the constraints, combined with the unit dynamics model, the optimal control command sequence for controlling the wind turbine is obtained through a rolling optimization algorithm. The optimal control command sequence is used to enable the wind turbine to track and approximate the preset target status data in the real-time operating status of the wind turbine in the next control cycle. The key points of the rolling optimization algorithm include: Prediction time domain: At the current time k, based on the current state and model, predict the unit state for the next N control cycles (e.g., the next 10 steps).

[0042] Optimization solution: In the prediction time domain, solve for the optimal control command sequence for the next M control cycles (e.g., the next 3 steps).

[0043] Only the first step is executed: only the first control command in the sequence is sent to the crew for execution, and all subsequent commands are discarded.

[0044] Rolling update: When the next time step k+1 arrives, the latest state is collected again, the prediction is re-made, the optimization is re-executed, and the first step of the new sequence is executed again, and so on.

[0045] Preferably, the core mathematical model in the MPC control architecture includes: Control function: , ;in, This represents the wind speed prediction sequence at time k. This represents the wind speed prediction sequence at time k-1. This represents the increment of the input sequence between two adjacent time points; This represents the sequence of control commands output at time k. This represents the sequence of control commands output at time k-1, where A and B represent coefficient matrices. Objective function: ;in, This represents the frequency deviation weighting coefficient. This represents the difference in angular frequency predicted at time k+i from time k. This represents the control increment weighting coefficient. Let M represent the increment of the input sequence at time k+i predicted at time k, where M represents the control time domain and N represents the prediction time domain; preferably, N=10 and M=3. , .

[0046] Constraints: ;in, , Indicates the upper and lower limits of the input sequence. This represents the wind speed prediction sequence for time k+i predicted at time k. , Indicates the upper and lower limits of the input sequence increment. , This represents the upper and lower limits of the output control vector. This represents the sequence of control commands predicted at time k+i from time k. Preferably, , .

[0047] Preferably, the optimal control command sequence includes a pitch angle adjustment command and an excitation current adjustment command, wherein: the pitch angle adjustment command is used to control the blade pitch angle adjustment; and the excitation current adjustment command is used to control the generator excitation current. Specifically, based on the characteristics of rolling optimization, the first element of the optimal control sequence is applied to the unit: The pitch angle adjustment command is applied to the electric actuator, which drives the blades to adjust the pitch angle with an adjustment accuracy of 0.1. The response time is ≤0.5s. To address the delay characteristics of the pitch angle controller, a pre-set adjustment lead is established based on the wind speed prediction value 1-2 seconds in the future. This time range is selected because the response time of the pitch angle actuator is ≤0.5s. Selecting 1-2s of prediction data can completely cover the control lag time, while reserving sufficient adjustment buffer time to avoid control lag problems caused by sudden changes in wind speed.

[0048] The excitation current regulation command acts on the converter, which adjusts the generator excitation current with an adjustment accuracy of 0.01A. The converter switching frequency is 5kHz. When the predicted wind speed increases, the excitation current is increased to improve the electromagnetic torque; when the predicted wind speed decreases, the excitation current is decreased to avoid a sudden drop in power.

[0049] More preferably, the control method for the wind turbine generator further includes: S500: Control the wind turbine according to the optimal control command sequence and obtain the updated operating status data of the wind turbine; S600: If the updated operating status data is consistent with the expected operating status data, the optimal control command sequence passes verification. Specifically, in this step, the updated operating status data is compared with the preset expected operating status data. If the deviation between the updated operating status data and the expected operating status data is within a preset threshold range, the optimal control command sequence is determined to have passed verification. If the deviation exceeds the preset threshold, the process of S200~S400 is re-executed to generate a new optimal control command sequence and execute it until the operating status data meets the expected requirements.

[0050] This application provides a control system for a wind turbine generator, such as... Figure 5 As shown, it includes: The parameter setting module 100 is configured to set the simulation parameters of the wind turbine and construct the turbine's dynamic model. Specifically, based on the wind turbine's hardware parameters and design specifications, core simulation parameters are set, which can cover the core fundamental parameters of the wind turbine, the power grid, and the direct-drive wind turbine. The turbine's dynamic model includes rotor motion equations and related variable relationship equations, describing the turbine's mechanical motion and energy conversion laws through a set of mathematical equations.

[0051] The data acquisition module 200 is configured to acquire historical wind speed data and real-time operating status data of the wind turbine. Specifically, historical wind speed data is acquired through wind speed sensors installed in the wind turbine, which includes a time-series wind speed sequence over a preset period of time. Real-time operating status data is acquired through the turbine's own status monitoring system, which includes key status variables such as generator real-time speed, current pitch angle, output power, grid real-time frequency, and real-time load on drivetrain components. The acquisition cycle is synchronized with the control cycle to ensure the timeliness and accuracy of the data.

[0052] The wind speed prediction module 300 is configured to input the historical wind speed data and the real-time operating status data into a pre-built wind speed prediction model, which outputs wind speed prediction data. Specifically, the preprocessed historical wind speed data and real-time operating status data are concatenated into a one-dimensional time-series input sequence to ensure the temporal correlation of the data. Processing techniques such as TCN, GRU, and fully connected layers are used to obtain the wind speed prediction data.

[0053] The instruction generation module 400 is configured to input the wind speed prediction data as feedforward data and the real-time operating status data as feedback data into a pre-constructed control model. The control model, combined with the turbine dynamics model, obtains the optimal control instruction sequence for controlling the wind turbine. Specifically, the wind speed prediction data and real-time operating status data are substituted into the turbine dynamics model to construct a future system state prediction. The control model generates the optimal control instruction sequence for future control cycles based on pre-set objective functions such as minimizing frequency error and minimizing output power fluctuation rate as constraints, for controlling the wind turbine.

[0054] The effects of using the scheme provided in the above embodiments of this application to control wind turbine generators in terms of both power control and load regulation are shown in Tables 4 and 5: Table 4 Power Control Effect

[0055] Table 5 Load Adjustment Effect

[0056] As shown in Table 4, under the same conditions, the TCN-GRU combined model used in this application consistently exhibits a smaller power standard deviation compared to the GRU model used in traditional schemes, indicating that it can better capture wind speed temporal characteristics and make the power output more stable. After optimizing the input conditions by predicting wind speed, power fluctuations are further reduced, proving that the wind speed prediction and feedforward adjustment combination method provided in this application can mitigate the power impact caused by wind condition changes.

[0057] As shown in Table 5, compared to the GRU model used in traditional schemes, the TCN-GRU combined model adopted in this application significantly reduces the standard deviation of shaft torque under the same conditions. This indicates that accurate wind speed prediction can adjust control commands in advance, avoiding the impact of sudden wind speed changes on the drivetrain. By optimizing input conditions through wind speed prediction, load fluctuations are reduced, which is more effective in protecting key components of the unit and extending their service life than traditional passive response.

[0058] The test data in the two tables above prove that the proposed solution can simultaneously improve power stability and load suppression capability compared to traditional solutions.

[0059] This application provides a computer-readable storage medium storing program information. After reading the program information, the computer executes the steps of the wind turbine control method described in any one of the method embodiments.

[0060] This application provides a computer program product, including a computer program / instruction, which, when executed by a processor, implements the steps of the wind turbine control method described in any one of the method embodiments.

[0061] This application also provides an electronic device, such as... Figure 6 As shown, the electronic device includes at least one processor 61 and at least one memory 62. The at least one memory 62 stores program information. After reading the program information, the at least one processor 61 executes the wind turbine control method described in any of the above method embodiments. The device may further include an input device 63 and an output device 64. The processor 61, memory 62, input device 63, and output device 64 can be communicatively connected. The memory 62, as a non-volatile computer-readable storage medium, can be used to store non-volatile software programs, non-volatile computer-executable programs, and modules. The processor 61 executes various functional applications and data processing by running the non-volatile software programs, instructions, and modules stored in the memory 62, thereby implementing the wind turbine control method provided in any of the above embodiments. The memory 62 may include a program storage area and a data storage area. The program storage area may store the operating system and at least one application program required for a function; the data storage area may store data created based on the use of the wind turbine control method, etc. Furthermore, memory 62 may include high-speed random access memory and non-volatile memory, such as at least one disk storage device, flash memory device, or other non-volatile solid-state storage device. In some embodiments, memory 62 may optionally include memory remotely located relative to processor 61, which can be connected via a network to means of performing the wind turbine control method. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof. Input device 63 may receive user clicks and generate signal inputs related to user settings and function control of the wind turbine control method. Output device 64 may include a display device such as a display screen. When the one or more modules are stored in memory 62 and are run by the one or more processors 61, the wind turbine control method of any of the above method embodiments is executed.

[0062] As needed, the above technical solutions can be combined to achieve the best technical effect.

[0063] The above are merely the principles and preferred embodiments of this application. It should be noted that, for those skilled in the art, several other modifications can be made based on the principles of this application, and these modifications should also be considered within the scope of protection of this application.

Claims

1. A control method for a wind turbine generator set, characterized in that, include: Set the simulation parameters for the wind turbine and construct the turbine dynamics model; Acquire the historical wind speed data and real-time operating status data of the wind turbine; The historical wind speed data and the real-time operating status data are input into a pre-built wind speed prediction model, and the wind speed prediction model outputs wind speed prediction data. The wind speed prediction data is used as feedforward data, and the real-time operating status data is used as feedback data. The data is then input into a pre-built control model, and the control model, combined with the unit dynamics model, yields the optimal control command sequence for controlling the wind turbine unit.

2. The control method for a wind turbine generator according to claim 1, characterized in that, The wind speed prediction model is a TCN-GRU combined model. The historical wind speed data and the real-time operating status data are input into the pre-constructed wind speed prediction model, and the wind speed prediction model outputs wind speed prediction data, including: The historical wind speed data and the real-time operating status data are processed into a time-series input sequence; The TCN The GRU combined model includes a TCN module and a GRU module; the TCN module receives the time-series input sequence and processes the time-series input sequence into a multi-dimensional time-series feature vector; the GRU module receives the multi-dimensional time-series feature vector and generates a wind speed prediction sequence for several future control cycles based on the multi-dimensional time-series feature vector, and the wind speed prediction sequence serves as the wind speed prediction data.

3. The control method for a wind turbine generator according to claim 2, characterized in that, The control model is an MPC architecture model. The wind speed prediction data is used as feedforward data, and the real-time operating status data is used as feedback data, which are input into the pre-constructed control model. The control model, combined with the unit dynamics model, obtains the optimal control command sequence for controlling the wind turbine unit, including: The MPC architecture model receives the wind speed prediction sequence as a feedforward input and the real-time operating status data as a feedback input. Set constraints based on the simulation parameters; Calculate the difference between the real-time operating status data and the preset target status data; Based on the wind speed prediction sequence, the difference, the constraints, and the unit dynamics model, an optimal control command sequence for controlling the wind turbine is obtained. The optimal control command sequence is used to ensure that the real-time operating status prediction data of the wind turbine in the next control cycle is consistent with the preset target status data.

4. The control method for a wind turbine generator according to claim 3, characterized in that: The unit dynamics model includes the rotor motion equations of the generator and the rotor equations of relevant variables, wherein: The rotor motion equation of the generator: ;in, Indicates the moment of inertia of the rotor. Represents mechanical torque. Indicates electromagnetic torque. Indicates the damping coefficient. This represents the actual angular velocity of the rotor. Indicates synchronous mechanical angular velocity, t Indicates time; Rotor equations for relevant variables: ;in, Indicates mechanical power. Indicates electromagnetic power. Indicates the power grid frequency. Indicates the number of pole pairs of the generator; Control function: , ;in, This represents the wind speed prediction sequence at time k. This represents the wind speed prediction sequence at time k-1. This represents the increment of the input sequence between two adjacent time points; This represents the sequence of control commands output at time k. This represents the sequence of control commands output at time k-1, where A and B represent coefficient matrices. Objective function: ;in, This represents the frequency deviation weighting coefficient. This represents the difference in angular frequency predicted at time k+i from time k. This represents the control increment weighting coefficient. Let M represent the increment of the input sequence at time k+i predicted at time k, where M represents the control time domain and N represents the prediction time domain. Constraints: ;in, , Indicates the upper and lower limits of the input sequence. This represents the wind speed prediction sequence for time k+i predicted at time k. , Indicates the upper and lower limits of the input sequence increment. , This represents the upper and lower limits of the output control vector. This represents the sequence of control instructions predicted at time k+i in the future.

5. The control method for a wind turbine generator according to claim 4, characterized in that, The optimal control command sequence includes pitch angle adjustment command and excitation current adjustment command, wherein: The blade pitch angle adjustment command is used to control the blade pitch angle adjustment. The excitation current regulation command is used to control the generator excitation current.

6. The control method for a wind turbine generator according to any one of claims 1-5, characterized in that, The method further includes: The wind turbine is controlled according to the optimal control command sequence to obtain the updated operating status data of the wind turbine. If the updated operating status data is consistent with the expected operating status data, then the optimal control command sequence passes verification.

7. A control system for a wind turbine generator set, characterized in that, include: The parameter setting module is configured to set the simulation parameters of the wind turbine and build the dynamic model of the turbine. The data acquisition module is configured to acquire historical wind speed data and real-time operating status data of the wind turbine. The wind speed prediction module is configured to input the historical wind speed data and the real-time operating status data into a pre-built wind speed prediction model, and the wind speed prediction model outputs wind speed prediction data. The instruction generation module is configured to input the wind speed prediction data as feedforward data and the real-time operating status data as feedback data into a pre-constructed control model. The control model, combined with the unit dynamics model, obtains the optimal control instruction sequence for controlling the wind turbine unit.

8. A computer-readable storage medium, characterized in that, The storage medium stores program information, and after the computer reads the program information, it executes the steps of the wind turbine control method according to any one of claims 1-6.

9. A computer program product comprising a computer program / instructions, characterized in that, When the computer program / instructions are executed by the processor, they implement the steps of the control method for the wind turbine generator as described in any one of claims 1-6.

10. An electronic device comprising a memory, a processor, and a computer program stored in the memory, characterized in that, The processor executes the computer program to implement the steps of the wind turbine control method according to any one of claims 1-6.

Images