Torque control method, device and equipment for wind turbine generator and medium
By optimizing the sliding window mean smoothing and target threshold dynamic coupling model, the adaptability and accuracy problems of traditional wind turbine torque control methods are solved, achieving accurate quantification of wind turbine torque imbalance and fault early warning, and extending the service life of the blades.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-01-07
- Publication Date
- 2026-03-10
AI Technical Summary
Traditional wind turbine torque control methods are based on fixed thresholds, which are difficult to adapt to different wind farms and operating conditions, resulting in high false alarm rates, insufficient diagnostic accuracy, poor adaptability and insufficient real-time performance, and inability to effectively protect the blades.
The difference norm between blades is calculated using a sliding window mean smoothing technique, and the threshold is adaptively optimized through a target threshold dynamic coupling model. This is combined with a fault identification model for precise torque control.
It enables precise quantification of torque imbalance in wind turbine units, improves the accuracy of fault warning and torque control precision, and extends the availability of blades.
Smart Images
Figure CN121630640A_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The present application relates to the technical field of wind power, in particular to a wind turbine torque control method, device, equipment and medium. BACKGROUND
[0002] With the rapid growth of wind power installed capacity, the safety of wind turbine operation is increasingly prominent. The damage of the blade, as the core component of the wind turbine, may cause serious accidents. The traditional protection method is mainly based on fixed threshold, which is difficult to adapt to the differences of different wind farm operation conditions, and has problems such as high false alarm rate and poor adaptability.
[0003] The traditional protection method mainly has the following technical defects: relying on fixed threshold: the existing technology mainly adopts the protection strategy based on fixed threshold, which is difficult to adapt to the needs of different wind farms, different models and changing operation conditions, resulting in high false alarm rate; insufficient diagnostic accuracy: the traditional method has limited ability to identify early blade faults, and cannot realize early warning of faults; poor self-adaptability: lacking of parameter self-adaptive optimization mechanism for changing operation environment, the protection effect is limited by the initial parameter setting; insufficient real-time performance: long data processing and analysis period, difficult to timely discover and handle sudden faults. SUMMARY
[0004] Therefore, the purpose of the present application is to provide a wind turbine torque control method, device, equipment and medium, which can realize accurate quantification of torque imbalance based on sliding window mean smoothing, and self-adaptive parameter optimization mechanism, so that the determined threshold is more accurate, thereby improving the accuracy of torque control, and effectively prolonging the availability of the blades in the wind turbine. The specific scheme is as follows:
[0005] In a first aspect, the present application discloses a wind turbine torque control method, comprising:
[0006] Collecting the variable pitch torque data of the target wind turbine, and smoothing the variable pitch torque data according to a preset sliding window to obtain processed variable pitch torque data;
[0007] Calculating the cumulative average value of each blade corresponding to the target wind turbine based on the processed variable pitch torque data, and determining the difference norm between the blades according to the cumulative average value;
[0008] Calculating the target threshold of the target wind turbine at present by a target threshold dynamic coupling model based on the model parameters of itself and the current equipment parameters of the target wind turbine;
[0009] Comparing the difference norm and the target threshold to determine whether to perform fault warning according to the comparison result, and if so, identifying the fault type of the target wind turbine by a preset fault identification model.
[0010] A corresponding torque control command is generated based on the fault type, and the torque control command is sent to the target wind turbine so that the target wind turbine can adjust its torque by executing the torque control command.
[0011] Optionally, the step of collecting the pitch torque data of the target wind turbine and smoothing the pitch torque data according to a preset sliding window to obtain processed pitch torque data includes:
[0012] The initial pitch torque data of the target wind turbine is collected based on a preset sampling frequency, and outliers and invalid values in the initial pitch torque data are removed to obtain the pitch torque data after removal.
[0013] The filtered pitch torque data is smoothed using a preset sliding window with a preset window length to obtain processed pitch torque data.
[0014] Optionally, the step of calculating the cumulative average value of each blade corresponding to the target wind turbine based on the processed pitch torque data, and determining the difference norm between each blade based on the cumulative average value, includes:
[0015] The cumulative average value of each blade corresponding to the target wind turbine is calculated using a preset sliding window algorithm.
[0016] The absolute value of the cumulative average is taken as the target cumulative average of each blade, and the Euclidean distance of each blade is calculated based on the target cumulative average, so as to use the obtained target Euclidean distance as the difference norm between each blade.
[0017] Optionally, the step of calculating the current target threshold of the target wind turbine based on its own model parameters and the current equipment parameters of the target wind turbine through the target threshold dynamic coupling model includes:
[0018] Obtain the current equipment parameters of the target wind turbine; the equipment parameters include the average wind speed, generator speed, turbulence intensity, and overall impact value of the target wind turbine.
[0019] The target threshold dynamic coupling model calculates the target optimal solution of the preset threshold function within the range of the several model parameters and the device parameters based on several model parameters and the device parameters; the several model parameters are the parameters corresponding to the target threshold dynamic coupling model.
[0020] The target optimal solution is used as the current target threshold for the target wind turbine.
[0021] Optionally, the step of calculating the target optimal solution of the preset threshold function within the range of the several model parameters and the device parameters based on a target threshold dynamic coupling model includes:
[0022] A grid search is performed on the target threshold dynamic coupling model based on the aforementioned model parameters and the device parameters to determine a combination of several model parameters for the target threshold dynamic coupling model; the combination of several model parameters is a combination of the aforementioned model parameters and the device parameters.
[0023] The application traversal of the aforementioned combinations of model parameters is performed, and the model performance of the aforementioned combinations of model parameters in the target threshold dynamic coupling model is determined.
[0024] Several target model parameter combinations to be optimized are selected from the aforementioned combinations of model parameters, and particle swarm optimization is performed on these combinations to obtain target model parameter combinations. The target model parameter combinations are the combinations of parameters that enable the model performance of the target threshold dynamic coupling model to reach the optimal level among the aforementioned combinations of model parameters.
[0025] The combination of the target model parameters is taken as the optimal solution for the target.
[0026] Optionally, the comparison of the difference norm and the target threshold is used to determine whether to issue a fault warning based on the comparison result. If a fault warning is issued, the fault type of the target wind turbine is identified through a preset fault identification model, including:
[0027] By comparing the difference norm with the target threshold, if the difference norm is greater than the target threshold, a fault warning is determined.
[0028] The fault type of the target wind turbine is determined based on the equipment parameters by a preset fault identification model; the preset fault identification model is a model obtained by training an initial fault identification model with historical fault data.
[0029] Optionally, the step of generating a corresponding torque control command based on the fault type and sending the torque control command to the target wind turbine so that the target wind turbine can adjust its torque by executing the torque control command includes:
[0030] Based on the fault type, a corresponding torque control command is generated, and the torque control commands are prioritized to obtain the target torque control command.
[0031] The target torque control command is sent to the target wind turbine, so that the target wind turbine executes the target torque control command based on the priority corresponding to the target torque control command to complete the torque adjustment.
[0032] Secondly, this application discloses a wind turbine torque control device, comprising:
[0033] The data processing module is used to collect the pitch torque data of the target wind turbine and smooth the pitch torque data according to a preset sliding window to obtain the processed pitch torque data.
[0034] The norm calculation module is used to calculate the cumulative average value of each blade corresponding to the target wind turbine based on the processed pitch torque data, and to determine the difference norm between each blade based on the cumulative average value.
[0035] The threshold determination module is used to calculate the current target threshold of the target wind turbine based on its own model parameters and the current equipment parameters of the target wind turbine through a target threshold dynamic coupling model.
[0036] The fault type determination module is used to compare the difference norm and the target threshold to determine whether to issue a fault warning based on the comparison result. If a fault warning is issued, the fault type of the target wind turbine is identified by a preset fault identification model.
[0037] The torque adjustment module is used to generate a corresponding torque control command according to the fault type, and send the torque control command to the target wind turbine, so that the target wind turbine can adjust its torque by executing the torque control command.
[0038] Thirdly, this application discloses an electronic device, including:
[0039] Memory, used to store computer programs;
[0040] A processor is used to execute the computer program to implement the wind turbine torque control method as described above.
[0041] Fourthly, this application discloses a computer-readable storage medium for storing a computer program, wherein the computer program, when executed by a processor, implements the aforementioned wind turbine torque control method.
[0042] In this application, pitch torque data of the target wind turbine can be collected and smoothed according to a preset sliding window to obtain processed pitch torque data. Based on the processed pitch torque data, the cumulative average value of each blade corresponding to the target wind turbine is calculated, and the difference norm between the blades is determined based on the cumulative average value. A target threshold dynamic coupling model is used to calculate the current target threshold of the target wind turbine based on its own model parameters and the current equipment parameters of the target wind turbine. The difference norm and the target threshold are compared to determine whether a fault warning should be issued. If a fault warning is issued, the fault type of the target wind turbine is identified using a preset fault identification model. A corresponding torque control command is generated based on the fault type and sent to the target wind turbine so that the target wind turbine can adjust its torque by executing the torque control command.
[0043] Therefore, the method of this application can collect pitch torque data of the target wind turbine and smooth the pitch torque data according to a preset sliding window; calculate the cumulative average value of each blade of the target wind turbine based on the obtained processed pitch torque data, and determine the difference norm between each blade based on the cumulative average value; calculate the current target threshold of the target wind turbine based on its own model parameters and the current equipment parameters of the target wind turbine through a target threshold dynamic coupling model; compare the difference norm and the target threshold to determine whether to issue a fault warning based on the comparison result; if a fault warning is issued, identify the fault type of the target wind turbine; generate corresponding torque control commands according to the fault type so that the target wind turbine can adjust its torque by executing the torque control commands. In this way, the sliding window mean smoothing can be used to achieve accurate quantification of torque imbalance, and the adaptive parameter optimization mechanism makes the determined threshold more accurate, effectively solving the key technical problems in the torque imbalance protection of wind turbines, and thus effectively extending the availability of the blades in the wind turbine. Attached Figure Description
[0044] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.
[0045] Figure 1 This is a flowchart of a wind turbine torque control method disclosed in this application;
[0046] Figure 2 This application discloses a specific flowchart of a wind turbine torque control method.
[0047] Figure 3 This application discloses a parameter optimization flowchart;
[0048] Figure 4 This is a schematic diagram of the architecture for torque control of a wind turbine generator disclosed in this application;
[0049] Figure 5 This application discloses a flowchart of a wind turbine torque adjustment process.
[0050] Figure 6 This application discloses a SCADA server deployment architecture diagram;
[0051] Figure 7 This is a schematic diagram of the structure of a wind turbine torque control device disclosed in this application;
[0052] Figure 8 This is a structural diagram of an electronic device disclosed in this application. Detailed Implementation
[0053] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0054] Currently, the blades are the core component of wind turbines, and damage to them can lead to serious accidents. Traditional blade protection methods are mainly based on fixed torque thresholds, which are difficult to adapt to the differences in operating conditions of different wind farms, resulting in problems such as high false alarm rates and poor adaptability.
[0055] To overcome the aforementioned technical problems, this application discloses a method, device, equipment, and medium for controlling the torque of a wind turbine. It can achieve precise quantification of torque imbalance based on sliding window mean smoothing, and an adaptive parameter optimization mechanism makes the determined threshold more accurate, thereby improving the accuracy of torque control and effectively extending the availability of the blades in the wind turbine.
[0056] See Figure 1 As shown in the figure, an embodiment of the present invention discloses a wind turbine torque control method, including:
[0057] Step S11: Collect the pitch torque data of the target wind turbine, and smooth the pitch torque data according to a preset sliding window to obtain the processed pitch torque data.
[0058] In this embodiment, pitch torque data from the wind turbine's SCADA (Supervisory Control And Data Acquisition) system needs to be collected first, and then smoothed. Specifically, initial pitch torque data of the target wind turbine needs to be collected based on a preset sampling frequency, and outliers and invalid values in the initial pitch torque data need to be removed to obtain the removed pitch torque data. The preset sampling frequency can be set as needed; in this embodiment, it is set to 1Hz to 10Hz. Then, the initial pitch torque data is quality checked, and outliers and invalid values are removed. Further, a preset sliding window based on a preset window length is used to smooth the removed pitch torque data to obtain the processed pitch torque data. It should be noted that the window length can be adaptively adjusted according to the wind turbine's operating status, and the preferred range of the preset window length is 30 seconds to 300 seconds. In this way, by smoothing the collected pitch torque data, data quality can be guaranteed, thereby improving the accuracy of wind turbine torque control.
[0059] Step S12: Calculate the cumulative average value of each blade corresponding to the target wind turbine based on the processed pitch torque data, and determine the difference norm between each blade according to the cumulative average value.
[0060] In this embodiment, it is necessary to extract the feature values of the processed pitch torque data. These features can be represented by the cumulative average value and the difference norm of each blade corresponding to the target wind turbine. Specifically, the cumulative average value of each blade corresponding to the target wind turbine needs to be calculated using a preset sliding window algorithm. The formula for the cumulative average value can be expressed as follows: Where N is the number of data points within the window. This represents the torque value at a single sampling point.
[0061] Furthermore, the absolute value of the cumulative average needs to be used as the target cumulative average for each blade, and the Euclidean distance of each blade needs to be calculated based on the target cumulative average. This target Euclidean distance is then used as the difference norm between the blades. The quantification formula for the difference norm can be expressed as follows: ,in, Let M1, M2, and M3 be the target cumulative average values corresponding to the three blades, denoted by the square root. This allows for the combination of sliding window mean smoothing and Euclidean distance quantization, achieving precise quantification of torque imbalance and thus improving the accuracy of wind turbine torque control.
[0062] Step S13: Calculate the current target threshold of the target wind turbine based on its own model parameters and the current equipment parameters of the target wind turbine through the target threshold dynamic coupling model.
[0063] In this embodiment, it is necessary to obtain the current equipment parameters of the target wind turbine. These parameters include the average wind speed C1, generator speed C2, turbulence intensity C3, and overall turbine impact value C4. Then, based on the target model parameter combination and the equipment parameters, a target threshold dynamic coupling model is used to calculate the optimal solution of a preset threshold function within the range of the target model parameter combination and the equipment parameters. This optimal solution is then used as the current target threshold for the target wind turbine. It should be noted that an alarm is triggered when the difference norm exceeds the threshold. The threshold is not independent but rather a function of the model's internal parameters. For a given fault characteristic value F, its alarm threshold... and shutdown threshold Dynamically determined by the target threshold dynamic coupling model:
[0064] ;
[0065] Where Th is the calculated dynamic threshold (which can be further subdivided into alarm thresholds). and shutdown threshold ); This is a basic threshold constant obtained from historical health data statistics; This is an adjustable parameter adjustment factor that reflects how the model's own structure adjusts the threshold. This is a runtime context adaptation factor that enables the threshold to be intelligently adjusted according to the current runtime environment. It is a time evolution factor used to handle the persistence or mutation of features.
[0066] For adjustable parameter adjustment factor This includes the sliding window length P1, the Euclidean distance weighting coefficient P2, and the sliding window speed P3. These need to be mapped to a systematic adjustment of the threshold, expressed by the following formula:
[0067] ;
[0068] in, , , These are reference values for each parameter under standard operating conditions. , , This represents the standard deviation of each parameter in historical data, used for normalization. , , The weighting coefficients for each parameter, obtained through training with historical data, characterize the degree of influence of that parameter on threshold sensitivity. This factor ensures that when the model parameters are adjusted to pursue higher sensitivity (e.g., shortening the sliding window P1), the threshold will automatically tighten to match its higher resolution.
[0069] For runtime context adaptation factors These include average wind speed C1, generator speed C2, turbulence intensity C3, and overall impact value C4, which need to be dynamically adapted. The formula is expressed as follows:
[0070] ;
[0071] in, , , , This factor acts as an adjustment coefficient for various environmental factors, determining their impact range. When wind speed (C1) deviates from the rated value, turbulence intensity (C3) increases, or the overall machine vibration and shock value (C4) increases, this factor will moderately relax the threshold to avoid false alarms triggered by normal operating condition fluctuations. When the generator speed (C2) is abnormal, it may also trigger threshold adjustment and participate in system protection.
[0072] For time evolution factor It targets the temporal characteristics of fault features, realizing the change of threshold over time. For sudden anomalies (such as instantaneous impacts), its formula is expressed as follows:
[0073] ;
[0074] Among them, when Upon detecting a transient impact, the threshold will instantly rise by a certain magnitude. Then according to the time constant The index decays back to normal levels, which effectively prevents multiple false alarms for the same event.
[0075] For progressive failures (such as crack propagation), the formula is expressed as follows:
[0076] ;
[0077] Among them, Once a suspected fault characteristic is identified, if the characteristic persists, the threshold will slowly decrease over time (the decay rate increases from...). (Control), thereby gradually improving the system's sensitivity and achieving gradual triggering of early warning.
[0078] And all the above weight coefficients ( , , , , , , , , All data were obtained through training on historical data. The optimization process typically employs multi-objective particle swarm optimization, aiming to minimize a comprehensive performance index J, expressed as follows:
[0079] ;
[0080] Where FPR is the false positive rate, FNR is the false negative rate, and ResponseTime is the model response time; a, b, and c are weighting coefficients used to balance the importance of different indicators. By optimizing the model parameters to minimize J, the optimal balance between accuracy and sensitivity can be achieved. Finally, Th, which minimizes J, is taken as the current target threshold for the target wind turbine.
[0081] Step S14: Compare the difference norm and the target threshold to determine whether to issue a fault warning based on the comparison result. If a fault warning is issued, the fault type of the target wind turbine is identified by a preset fault identification model.
[0082] In this embodiment, it is necessary to compare the difference norm and the target threshold to determine whether to issue a fault warning based on the comparison result. Specifically, it is necessary to compare the difference norm and the target threshold; if the difference norm is greater than the target threshold, a fault warning is determined. Then, the fault type of the target wind turbine is determined based on the equipment parameters using a preset fault identification model. The preset fault identification model is a model trained on an initial fault identification model using historical fault data. Furthermore, a tiered warning mechanism can be set to issue alarms according to the severity of the fault.
[0083] Step S15: Generate a corresponding torque control command according to the fault type, and send the torque control command to the target wind turbine so that the target wind turbine can adjust its torque by executing the torque control command.
[0084] In this embodiment, corresponding torque control commands need to be generated based on the fault type, and these commands are prioritized to obtain the target torque control command. This target torque control command is then sent to the target wind turbine, allowing it to execute the command based on its priority to achieve torque adjustment. It should be noted that the diagnostic results need to be transmitted to the turbine controller in real time via the field-side communication network. Furthermore, a command priority mechanism needs to be established to ensure that important commands are processed first, and a command confirmation mechanism needs to be set up to guarantee the reliable execution of control commands. In this way, the efficiency of wind turbine torque control can be ensured through the established command priority, and the real-time execution of commands can be ensured through the command confirmation mechanism.
[0085] Therefore, the method of this application can collect pitch torque data of the target wind turbine and smooth the pitch torque data according to a preset sliding window; calculate the cumulative average value of each blade of the target wind turbine based on the obtained processed pitch torque data, and determine the difference norm between each blade based on the cumulative average value; calculate the current target threshold of the target wind turbine based on its own model parameters and the current equipment parameters of the target wind turbine through a target threshold dynamic coupling model; compare the difference norm and the target threshold to determine whether to issue a fault warning based on the comparison result; if a fault warning is issued, identify the fault type of the target wind turbine; generate corresponding torque control commands according to the fault type so that the target wind turbine can adjust its torque by executing the torque control commands. In this way, the sliding window mean smoothing can be used to achieve accurate quantification of torque imbalance, and the adaptive parameter optimization mechanism makes the determined threshold more accurate, effectively solving the key technical problems in the torque imbalance protection of wind turbines, and thus effectively extending the availability of the blades in the wind turbine.
[0086] As can be seen from the foregoing embodiments, the method of this application can calculate the target optimal solution of a preset threshold function within the range of several model parameters and device parameters using a target threshold dynamic coupling model based on several model parameters and device parameters. Therefore, this embodiment provides a detailed explanation of how to calculate the optimal solution. See [link to documentation]. Figure 2 As shown in the figure, an embodiment of the present invention discloses a wind turbine torque control method, including:
[0087] Step S21: Perform a grid search on the target threshold dynamic coupling model based on several model parameters and device parameters to determine several model parameter combinations of the target threshold dynamic coupling model; the several model parameter combinations are combinations of the several model parameters and the device parameters.
[0088] In this embodiment, as Figure 3 As shown, the first step is to perform a grid search on the target threshold dynamic coupling model based on several model parameters and device parameters to determine several combinations of model parameters for the target threshold dynamic coupling model. The goal of the grid search stage is to conduct a global exploration within a broad parameter space and quickly locate potential regions. First, a possible value range and step size need to be defined for each parameter to be optimized. For example, the sliding window size W can range from [30, 300] seconds, with a step size of 30 seconds, and the model parameter combinations include the sliding window length, Euclidean distance weighting coefficient, and sliding window speed. This forms a multi-dimensional grid parameter space.
[0089] Step S22: Apply the several model parameter combinations to the model performance of the target threshold dynamic coupling model.
[0090] In this embodiment, it is necessary to apply and iterate through several combinations of model parameters to determine the model performance of each combination of model parameters applied to the target threshold dynamic coupling model. Specifically, a global traversal calculation is required, automatically traversing each parameter combination in the grid. Then, the application and iteration of several combinations of model parameters is performed to determine the model performance of each combination of model parameters applied to the target threshold dynamic coupling model. For each set of parameters, a period of historical data (e.g., one month) can be used for simulation and its comprehensive performance index J can be calculated.
[0091] Step S23: Select several target model parameter combinations to be optimized from the several model parameter combinations, and perform particle swarm optimization on the several target model parameter combinations to be optimized to obtain target model parameter combinations; the target model parameter combinations are the parameter combinations among the several model parameter combinations that make the model performance of the target threshold dynamic coupling model reach the optimal level.
[0092] In this embodiment, it is necessary to select several target model parameter combinations to be optimized from the several model parameter combinations, and as follows: Figure 3 As shown, particle swarm optimization is performed on the several target model parameter combinations to be optimized. It should be noted that this step is equivalent to locating the parameter region with excellent model performance. The model performance corresponding to each parameter combination can be sorted, and then the top K groups of parameters can be selected as the target model parameter combinations to be optimized. Further, particle swarm optimization is performed on the target model parameter combinations to obtain the target model parameter combinations. It should be noted that the goal of particle swarm optimization on the target model parameter combinations is to perform a highly efficient local fine-grained search based on the target model parameter combinations to find the globally optimal or near-optimal solution. First, the obtained several target model parameter combinations to be optimized are used as the initial positions of the particle swarm. Each particle represents a parameter combination Pi=(P1,P2,...,Pn), and its velocity vi is randomly initialized. Then, it is iteratively updated. In each iteration, each example updates its own velocity and position by tracking two extreme values, where the two extreme values are the individual extreme value (pbest): the best position experienced by the particle itself, and the global extreme value (gbest): the best position found by the entire population so far. The update formula is shown below:
[0093] ;
[0094] in, It is inertial weight. , It is a learning factor. , The iteration stops when the maximum number of iterations is reached or the global extremum gbest no longer improves significantly within several consecutive iterations. gbest is then output as the optimal parameter configuration, i.e., the target model parameter combination. This target model parameter combination is then applied to the target threshold dynamic coupling model to obtain the corresponding optimal solution.
[0095] Step S24: Combine the target model parameters as the target optimal solution.
[0096] Therefore, in this embodiment, a grid search can be performed on the target threshold dynamic coupling model based on several model parameters and equipment parameters to determine several combinations of model parameters for the target threshold dynamic coupling model. Then, these combinations of model parameters are applied and traversed to determine their respective performance on the target threshold dynamic coupling model. From these combinations, several target model parameter combinations to be optimized are selected, and particle swarm optimization is performed on these target model parameter combinations to obtain the target model parameter combination. Finally, the target model parameter combination is taken as the target optimal solution. In this way, the model generalization problem can be solved through its hierarchical parameter optimization mechanism, ensuring the accuracy of the threshold solution and thus guaranteeing the precision of wind turbine torque control.
[0097] As a preferred embodiment, such as Figure 4 The diagram shown illustrates the architecture for wind turbine torque control in this application. The turbine's SCADA system collects operational data from each wind turbine in the wind farm cluster. The SCADA server corresponding to the turbine's SCADA system then aggregates the data and performs other processing as needed. Figure 5 The preprocessing shown includes sliding window smoothing and Euclidean quantization. Then, the intelligent diagnostic engine calculates the current target threshold for the target wind turbine using a dynamically coupled model pushed by the cloud-based collaborative platform. It should be noted that, as... Figure 6 The diagram shows the SCADA server deployment architecture, which adopts a cloud-edge collaborative architecture. This architecture utilizes a combined cloud computing and edge computing working mode to achieve distributed data processing, reducing the resource consumption for local model training. Then, the cumulative average value of each blade corresponding to the target wind turbine is calculated through a target threshold dynamic coupling model, and the difference norm between each blade is determined based on the cumulative average value. Next, the target threshold dynamic coupling model calculates the current target threshold of the target wind turbine based on its own model parameters and the current equipment parameters of the target wind turbine. The fault diagnosis module compares the difference norm with the target threshold to determine whether a fault warning should be issued. If a fault warning is issued, a preset fault identification model identifies the fault type of the target wind turbine, generates a corresponding torque control command based on the fault type, and sends the torque control command to the appropriate server. Figure 6The target wind turbine unit that has malfunctioned is shown so that the target wind turbine unit can adjust its torque by executing torque control commands.
[0098] See Figure 7 As shown, an embodiment of the present invention discloses a wind turbine torque control device, comprising:
[0099] The data processing module 11 is used to collect the pitch torque data of the target wind turbine and smooth the pitch torque data according to a preset sliding window to obtain the processed pitch torque data.
[0100] Norm calculation module 12 is used to calculate the cumulative average value of each blade corresponding to the target wind turbine based on the processed pitch torque data, and to determine the difference norm between each blade based on the cumulative average value.
[0101] Threshold determination module 13 is used to calculate the current target threshold of the target wind turbine based on its own model parameters and the current equipment parameters of the target wind turbine through the target threshold dynamic coupling model;
[0102] The fault type determination module 14 is used to compare the difference norm and the target threshold to determine whether to issue a fault warning based on the comparison result. If a fault warning is issued, the fault type of the target wind turbine is identified by a preset fault identification model.
[0103] The torque adjustment module 15 is used to generate a corresponding torque control command according to the fault type and send the torque control command to the target wind turbine, so that the target wind turbine can adjust its torque by executing the torque control command.
[0104] In some embodiments, the data processing module 11 may specifically include:
[0105] The data removal unit is used to collect the initial pitch torque data of the target wind turbine based on a preset sampling frequency, and remove outliers and invalid values in the initial pitch torque data to obtain the pitch torque data after removal.
[0106] The smoothing unit is used to smooth the rejected pitch torque data through a preset sliding window based on a preset window length, so as to obtain the processed pitch torque data.
[0107] In some embodiments, the norm calculation module 12 may specifically include:
[0108] The average value calculation unit is used to calculate the cumulative average value of each blade corresponding to the target wind turbine through a preset sliding window algorithm.
[0109] The norm calculation unit is used to take the absolute value of the cumulative average as the target cumulative average of each blade, and calculate the Euclidean distance of each blade based on the target cumulative average, so as to take the obtained target Euclidean distance as the difference norm between each blade.
[0110] In some embodiments, the threshold determination module 13 may specifically include:
[0111] The equipment parameter acquisition submodule is used to acquire the current equipment parameters of the target wind turbine; the equipment parameters include the average wind speed, generator speed, turbulence intensity, and overall impact value of the target wind turbine.
[0112] The optimal solution solving submodule is used to calculate the target optimal solution of the preset threshold function within the range of the several model parameters and the device parameters based on several model parameters and the device parameters through the target threshold dynamic coupling model; the several model parameters are the parameters corresponding to the target threshold dynamic coupling model;
[0113] The target threshold determination submodule is used to take the target optimal solution as the current target threshold of the target wind turbine.
[0114] In some embodiments, the optimal solution solving submodule may specifically include:
[0115] A grid search unit is used to perform a grid search on the target threshold dynamic coupling model based on the plurality of model parameters and the device parameters to determine a plurality of model parameter combinations of the target threshold dynamic coupling model; the plurality of model parameter combinations are combinations of the plurality of model parameters and the device parameters;
[0116] The parameter application unit is used to perform an application traversal on the several model parameter combinations and determine the model performance of the several model parameter combinations applied to the target threshold dynamic coupling model respectively.
[0117] The parameter filtering unit is used to filter out several target model parameter combinations to be optimized from the several model parameter combinations, and to perform particle swarm optimization on the several target model parameter combinations to be optimized to obtain the target model parameter combination; the target model parameter combination is the parameter combination that makes the model performance of the target threshold dynamic coupling model reach the optimal among the several model parameter combinations.
[0118] The optimal solution determination unit is used to combine the target model parameters as the target optimal solution.
[0119] In some embodiments, the fault type determination module 14 may specifically include:
[0120] A data comparison unit is used to compare the difference norm and the target threshold. If the difference norm is greater than the target threshold, a fault warning is determined.
[0121] The fault type determination unit is used to determine the fault type of the target wind turbine based on the equipment parameters using a preset fault identification model; the preset fault identification model is a model obtained by training an initial fault identification model using historical fault data.
[0122] In some embodiments, the torque adjustment module 15 may specifically include:
[0123] The instruction sorting unit is used to generate corresponding torque control instructions based on the fault type and sort the torque control instructions by priority to obtain the target torque control instruction.
[0124] The instruction sending unit is used to send the target torque control instruction to the target wind turbine, so that the target wind turbine executes the target torque control instruction based on the priority corresponding to the target torque control instruction to complete the torque adjustment.
[0125] Furthermore, embodiments of this application also disclose an electronic device, Figure 8 This is a structural diagram of an electronic device 20 according to an exemplary embodiment. The content of the diagram should not be construed as limiting the scope of this application.
[0126] Figure 8 This is a schematic diagram of the structure of an electronic device 20 provided in an embodiment of this application. Specifically, the electronic device 20 may include: at least one processor 21, at least one memory 22, a power supply 23, a communication interface 24, an input / output interface 25, and a communication bus 26. The memory 22 stores a computer program, which is loaded and executed by the processor 21 to implement the relevant steps in the wind turbine torque control method disclosed in any of the foregoing embodiments. Alternatively, the electronic device 20 in this embodiment may specifically be an electronic computer.
[0127] In this embodiment, the power supply 23 is used to provide operating voltage for each hardware device on the electronic device 20; the communication interface 24 can create a data transmission channel between the electronic device 20 and external devices, and the communication protocol it follows can be any communication protocol applicable to the technical solution of this application, and is not specifically limited here; the input / output interface 25 is used to acquire external input data or output data to the outside world, and its specific interface type can be selected according to specific application needs, and is not specifically limited here.
[0128] In addition, the memory 22, as a carrier for resource storage, can be a read-only memory, random access memory, disk or optical disk, etc. The resources stored thereon can include operating system 221, computer program 222, etc., and the storage method can be temporary storage or permanent storage.
[0129] The operating system 221 is used to manage and control the various hardware devices on the electronic device 20 and the computer program 222, which may be Windows Server, Netware, Unix, Linux, etc. In addition to including a computer program capable of performing the wind turbine torque control method executed by the electronic device 20 as disclosed in any of the foregoing embodiments, the computer program 222 may further include computer programs capable of performing other specific tasks.
[0130] Furthermore, this application also discloses a computer-readable storage medium for storing a computer program; wherein, when the computer program is executed by a processor, it implements the aforementioned wind turbine torque control method. Specific steps of this method can be found in the corresponding content disclosed in the foregoing embodiments, and will not be repeated here.
[0131] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the apparatus disclosed in the embodiments, since it corresponds to the method disclosed in the embodiments, the description is relatively simple; relevant parts can be referred to in the method section.
[0132] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.
[0133] The steps of the methods or algorithms described in conjunction with the embodiments disclosed herein can be implemented directly by hardware, a software module executed by a processor, or a combination of both. The software module can be located in random access memory (RAM), main memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, removable disk, CD-ROM, or any other form of storage medium known in the art.
[0134] Finally, it should be noted that in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.
[0135] The technical solutions provided in this application have been described in detail above. Specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this application. Therefore, the content of this specification should not be construed as a limitation of this application.
Claims
1. A wind turbine generator torque control method characterized by, The method comprises the following steps: collecting variable pitch torque data of a target wind turbine, and performing smoothing processing on the variable pitch torque data according to a preset sliding window to obtain processed variable pitch torque data; calculating cumulative average values of each blade corresponding to the target wind turbine based on the processed variable pitch torque data, and determining a difference norm between the blades according to the cumulative average values; calculating a current target threshold of the target wind turbine based on model parameters of a target threshold dynamic coupling model and current equipment parameters of the target wind turbine; comparing the difference norm and the target threshold to determine whether to perform fault early warning according to a comparison result, and if so, identifying a fault type of the target wind turbine through a preset fault identification model; generating a corresponding torque control instruction according to the fault type, and sending the torque control instruction to the target wind turbine so that the target wind turbine performs torque adjustment by executing the torque control instruction.
2. The wind turbine generator torque control method according to claim 1, characterized by, The method of collecting variable pitch torque data of a target wind turbine, and performing smoothing processing on the variable pitch torque data according to a preset sliding window to obtain processed variable pitch torque data comprises the following steps: collecting initial variable pitch torque data of a target wind turbine based on a preset sampling frequency, and removing abnormal values and invalid values in the initial variable pitch torque data to obtain removed variable pitch torque data; performing smoothing processing on the removed variable pitch torque data based on a preset window length through a preset sliding window to obtain processed variable pitch torque data.
3. The wind turbine generator torque control method according to claim 1, characterized by, The method of calculating cumulative average values of each blade corresponding to the target wind turbine based on the processed variable pitch torque data, and determining a difference norm between the blades according to the cumulative average values comprises the following steps: calculating the cumulative average values of each blade corresponding to the target wind turbine through a preset sliding window algorithm; taking an absolute value of the cumulative average value as a target cumulative average value of each blade, and calculating an Euclidean distance of each blade based on the target cumulative average value to obtain a target Euclidean distance as the difference norm between the blades.
4. The wind turbine generator torque control method according to claim 1, characterized by, The method of calculating a current target threshold of the target wind turbine based on model parameters of a target threshold dynamic coupling model and current equipment parameters of the target wind turbine comprises the following steps: obtaining current equipment parameters of the target wind turbine; the equipment parameters include average wind speed, generator speed, turbulence intensity, and overall impact value of the target wind turbine; calculating a target optimal solution of a preset threshold function within a range of a plurality of model parameters and the equipment parameters based on the plurality of model parameters and the equipment parameters through a target threshold dynamic coupling model; the plurality of model parameters are parameters corresponding to the target threshold dynamic coupling model; taking the target optimal solution as the current target threshold of the target wind turbine.
5. The wind turbine generator torque control method according to claim 4, characterized by, The method of calculating a target optimal solution of a preset threshold function within a range of a plurality of model parameters and the equipment parameters based on the plurality of model parameters and the equipment parameters comprises the following steps: perform a grid search on the target threshold dynamic coupling model based on the model parameters and the device parameters to determine a plurality of model parameter combinations of the target threshold dynamic coupling model; the plurality of model parameter combinations are combinations of the model parameters and the device parameters; perform application traversal on the plurality of model parameter combinations and determine model performances of the target threshold dynamic coupling model when the plurality of model parameter combinations are applied respectively; select a plurality of target model parameter combinations to be optimized from the plurality of model parameter combinations and perform particle swarm optimization on the plurality of target model parameter combinations to obtain a target model parameter combination; the target model parameter combination is a parameter combination that makes the model performance of the target threshold dynamic coupling model optimal among the plurality of model parameter combinations; use the target model parameter combination as the target optimal solution.
6. The wind turbine generator torque control method according to claim 1, characterized by, comparing the difference norm and the target threshold to determine whether to perform fault warning according to a comparison result, and if fault warning is performed, identifying a fault type of the target wind turbine through a preset fault identification model, including: comparing the difference norm and the target threshold, and if the difference norm is greater than the target threshold, determining to perform fault warning; determining the fault type of the target wind turbine according to the device parameters through a preset fault identification model; the preset fault identification model is a model obtained by training an initial fault identification model through historical fault data.
7. A wind turbine generator torque control method according to any one of claims 1 to 6, characterized by, generating a corresponding torque control instruction according to the fault type, and sending the torque control instruction to the target wind turbine, so that the target wind turbine performs torque adjustment by executing the torque control instruction, including: generating a corresponding torque control instruction based on the fault type, and performing priority sorting on the torque control instruction to obtain a target torque control instruction; sending the target torque control instruction to the target wind turbine, so that the target wind turbine executes the target torque control instruction based on the priority corresponding to the target torque control instruction to complete torque adjustment.
8. A wind turbine generator torque control device characterized by, including: a data processing module configured to collect pitch torque data of a target wind turbine and perform smoothing processing on the pitch torque data according to a preset sliding window to obtain processed pitch torque data; a norm calculation module configured to calculate cumulative average values of each blade corresponding to the target wind turbine based on the processed pitch torque data, and determine a difference norm between the blades according to the cumulative average values; a threshold determination module configured to calculate a current target threshold of the target wind turbine based on model parameters of a target threshold dynamic coupling model and current device parameters of the target wind turbine; a fault type determination module configured to compare the difference norm and the target threshold to determine whether to perform fault warning according to a comparison result, and if fault warning is performed, identify a fault type of the target wind turbine through a preset fault identification model; A torque adjustment module is configured to generate a corresponding torque control instruction according to the fault type, and send the torque control instruction to the target wind turbine, so that the target wind turbine performs torque adjustment by executing the torque control instruction.
9. An electronic device, comprising: The method comprises: a memory for storing a computer program; a processor for executing the computer program to implement the wind turbine torque control method according to any one of claims 1 to 7.
10. A computer-readable storage medium, characterized in that, A computer program product is configured to implement the wind turbine torque control method according to any one of claims 1 to 7 when executed by a processor.