A fan blade control method, device, equipment and storage medium

Abstract

The application discloses a wind turbine blade control method, device, equipment and storage medium, and relates to the technical field of wind power generation. The method comprises the following steps: acquiring historical operation state data and historical clearance data of a target wind turbine, and dividing the historical clearance data into a dangerous clearance data set and a safe clearance data set based on a preset dangerous clearance threshold; performing correlation analysis on each dangerous clearance data in the dangerous clearance data set and the corresponding operation state data, and on each safe clearance data in the safe clearance data set and the corresponding operation state data to determine a first training data set and a second training data set; and generating a clearance Bayesian classifier model corresponding to the target wind turbine; and determining the blade clearance state of the target wind turbine according to the clearance Bayesian classifier model and current operation state data to control the operation of the wind turbine blade. In this way, it can be determined whether to start the variable pitch according to the current clearance state of the target wind turbine at any time to protect the safety of the wind turbine blade.

Description

Technical Field

[0001] This invention relates to the field of wind power generation technology, and in particular to a wind turbine blade control method, device, equipment and storage medium. Background Technology

[0002] As wind turbines become larger and the demand for cost reduction continues to increase, the size of wind turbine blades is constantly increasing while their stiffness is gradually decreasing. This leads to a continuous reduction in the minimum distance between the blade tip and the tower surface (minimum blade clearance) during operation. When the blade clearance is too small, blade-to-tower sweep events are highly likely to occur, resulting in blade damage and tower loss.

[0003] Currently, the main approach to addressing blade clearance issues involves real-time monitoring of blade clearance values ​​using hardware sensors such as lidar, video detectors, and millimeter-wave radar. When the clearance value becomes too low, pitch control is initiated to protect the turbine. However, adding hardware sensors incurs significant additional costs, and these sensors are susceptible to external interference, potentially leading to false alarms of low clearance values ​​when the turbine is safe, or even missed alarms when the turbine is in danger. Therefore, a new clearance warning and monitoring solution is urgently needed to address the aforementioned shortcomings and deficiencies of existing technologies. Summary of the Invention

[0004] In view of this, the purpose of this invention is to provide a wind turbine blade control method, device, equipment, and storage medium, capable of determining whether the current clearance of the target wind turbine is dangerous; if dangerous, it activates pitch control to increase the blade clearance value, thereby reducing the probability of blade sweeping. The specific solution is as follows:

[0005] In a first aspect, this application discloses a wind turbine blade control method, including:

[0006] The historical operating status data and corresponding historical air clearance data of the target wind turbine are obtained, and the historical air clearance data are divided into dangerous air clearance data and safe air clearance data based on a preset dangerous air clearance threshold to obtain dangerous air clearance dataset and safe air clearance dataset; the air clearance value of the dangerous air clearance data is less than or equal to the preset dangerous air clearance threshold, and the air clearance value of the safe air clearance data is greater than the preset dangerous air clearance threshold.

[0007] Correlation analysis is performed on each of the hazardous airspace data and corresponding operational status data in the hazardous airspace dataset and each of the safe airspace data and corresponding operational status data in the safe airspace dataset to determine the first training dataset and the second training dataset.

[0008] Generate a Bayesian classifier model for the target wind turbine based on the first training dataset and the second training dataset;

[0009] The blade clearance status of the target wind turbine is determined based on the clearance Bayesian classifier model and the current operating status data of the target wind turbine, and the operation of the wind turbine blades is controlled based on the blade clearance status.

[0010] Optionally, before performing correlation analysis on each of the hazardous airspace data and corresponding operational status data in the hazardous airspace dataset and each of the safe airspace data and corresponding operational status data in the safe airspace dataset to determine the first training dataset and the second training dataset, the method further includes:

[0011] Determine whether the operating status data corresponding to each of the hazardous airspace data in the hazardous airspace dataset meets the preset threshold judgment condition;

[0012] If the preset threshold judgment condition is met, the dangerous airspace data is identified as interference data and the interference data is removed from the dangerous airspace dataset.

[0013] Optionally, determining whether the operational status data corresponding to each hazardous airspace data in the hazardous airspace dataset meets a preset threshold judgment condition includes:

[0014] Obtain the unit power data, generator speed data, and blade pitch angle value corresponding to each of the aforementioned hazardous airspace data;

[0015] Determine whether the unit power data is less than a preset power threshold, and / or whether the generator speed data is less than a preset speed threshold, and / or whether the pitch angle value is greater than a preset pitch angle value threshold.

[0016] Optionally, the step of performing correlation analysis on each of the hazardous airspace data and corresponding operational status data in the hazardous airspace dataset and each of the safe airspace data and corresponding operational status data in the safe airspace dataset to determine the first training dataset and the second training dataset includes:

[0017] Based on the preset correlation coefficient determination formula, a correlation analysis is performed on each of the hazardous airspace data and the corresponding operational status data to obtain the first correlation value;

[0018] Based on the preset correlation coefficient determination formula, a correlation analysis is performed on each of the safety clearance data and the corresponding operating status data to obtain a second correlation value;

[0019] A first initial training clearance dataset is determined based on the first correlation value; the first correlation value between the hazardous clearance data and the corresponding operational status data in the first training clearance dataset is greater than the first correlation value between the hazardous clearance data and the corresponding other operational status data in the hazardous clearance dataset.

[0020] A second initial training data set is selected from the safety clearance dataset based on the second correlation value; the second correlation value between the safety clearance data and the corresponding operating status data in the second training data set is greater than the second correlation value between the other safety clearance data and the corresponding operating status data in the safety clearance dataset.

[0021] The first training net dataset and the second training net dataset are determined based on the first initial training net dataset and the second initial training net dataset.

[0022] Optionally, generating the headroom Bayesian classifier model corresponding to the target wind turbine based on the first training dataset and the second training dataset includes:

[0023] Determine the first probability density distribution function corresponding to the first training dataset and the second probability density distribution function corresponding to the second training dataset;

[0024] A headroom Bayesian classifier model corresponding to the target wind turbine is generated based on the first probability density distribution function and the second probability density distribution function.

[0025] Optionally, determining the first probability density distribution function corresponding to the first training dataset and the second probability density distribution function corresponding to the second training dataset includes:

[0026] A first frequency histogram is generated based on each hazardous clearance data and corresponding operational status data in the first training dataset, and a first probability distribution type is determined based on the type of the first frequency histogram.

[0027] A second frequency histogram is generated based on each safety clearance data and corresponding operational status data in the second training dataset, and a second probability distribution type is determined based on the type of the second frequency histogram.

[0028] A first probability density distribution function is determined using a preset maximum likelihood estimation method and the first probability distribution type, and a second probability density distribution function is determined using the preset maximum likelihood estimation method and the second probability distribution type.

[0029] Optionally, determining the blade clearance state of the target wind turbine based on the clearance Bayesian classifier model and the current operating status data of the target wind turbine, and controlling the operation of the wind turbine blades based on the blade clearance state, includes:

[0030] The current operating status data of the target wind turbine is obtained, and the current dangerous airspace probability and current safe airspace probability of the target wind turbine are determined by the airspace Bayesian classifier model and the current operating status data.

[0031] Determine whether the current dangerous airspace probability is greater than a preset dangerous airspace probability threshold;

[0032] If the current dangerous clearance probability is greater than the preset dangerous clearance probability threshold, then the wind turbine blades are controlled to adjust the pitch angle.

[0033] Secondly, this application discloses a wind turbine blade control device, comprising:

[0034] The airspace data classification module is used to acquire historical operating status data and corresponding historical airspace data of the target wind turbine, and classify the historical airspace data into dangerous airspace data and safe airspace data based on a preset dangerous airspace threshold to obtain dangerous airspace dataset and safe airspace dataset; the airspace value of the dangerous airspace data is less than or equal to the preset dangerous airspace threshold, and the airspace value of the safe airspace data is the preset dangerous airspace threshold.

[0035] The training dataset generation module is used to perform correlation analysis on each of the dangerous airspace data and corresponding operating status data in the dangerous airspace dataset and each of the safe airspace data and corresponding operating status data in the safe airspace dataset, so as to determine the first training dataset and the second training dataset.

[0036] The model generation module is used to generate a headroom Bayesian classifier model for the target wind turbine based on the first training dataset and the second training dataset.

[0037] The blade control module is used to determine the blade clearance status of the target wind turbine based on the clearance Bayesian classifier model and the current operating status data of the target wind turbine, and to control the operation of the wind turbine blades based on the blade clearance status.

[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 aforementioned wind turbine blade control method.

[0041] Fourthly, this application discloses a computer-readable storage medium for storing a computer program, which, when executed by a processor, implements the aforementioned wind turbine blade control method.

[0042] As can be seen, in this application, historical operating status data and corresponding historical clearance data of the target wind turbine are obtained, and the historical clearance data are divided into dangerous clearance data and safe clearance data based on a preset dangerous clearance threshold to obtain dangerous clearance datasets and safe clearance datasets; the clearance value of the dangerous clearance data is less than or equal to the preset dangerous clearance threshold, and the clearance value of the safe clearance data is greater than the preset dangerous clearance threshold; correlation analysis is performed on each dangerous clearance data and corresponding operating status data in the dangerous clearance dataset, and on each safe clearance data and corresponding operating status data in the safe clearance dataset, to determine a first training dataset and a second training dataset; a clearance Bayesian classifier model corresponding to the target wind turbine is generated based on the first training dataset and the second training dataset; the blade clearance state of the target wind turbine is determined according to the clearance Bayesian classifier model and the current operating status data of the target wind turbine, and the wind turbine blade operation is controlled based on the blade clearance state. In this way, a clearance Bayesian classifier model for the target wind turbine is generated based on the selected blade clearance data with high correlation and the corresponding operating status data. During the actual operation of the target wind turbine, current operating status data is collected and fed into the headroom Bayesian classifier model to determine the current blade status of the target wind turbine. If the blade status is dangerous, the headroom status of the blades can be adjusted by controlling the target wind turbine to avoid blade sweeping events due to insufficient blade headroom, which could cause blade damage and tower loss. Attached Figure Description

[0043] 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.

[0044] Figure 1 This is a flowchart of a wind turbine blade control method disclosed in this application;

[0045] Figure 2 This is a flowchart of a specific wind turbine blade control method disclosed in this application;

[0046] Figure 3 This is a flowchart of a specific wind turbine blade control method disclosed in this application;

[0047] Figure 4 This is a schematic diagram of the structure of a wind turbine blade control device disclosed in this application;

[0048] Figure 5This is a structural diagram of an electronic device disclosed in this application. Detailed Implementation

[0049] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0050] Currently, the main approach to addressing blade clearance issues is through hardware sensors such as lidar, video detectors, and millimeter-wave radar to monitor blade clearance values ​​in real time. When the clearance value becomes too low, pitch control is initiated to protect the turbine's safety. This embodiment will specifically introduce a method for predicting the current blade clearance state using a clearance Bayesian classifier model and controlling the turbine blades based on that state.

[0051] See Figure 1 As shown in the figure, this application discloses a wind turbine blade control method, including:

[0052] Step S11: Obtain the historical operating status data and corresponding historical air clearance data of the target wind turbine, and divide the historical air clearance data into dangerous air clearance data and safe air clearance data based on the preset dangerous air clearance threshold to obtain dangerous air clearance dataset and safe air clearance dataset.

[0053] In this embodiment, historical operating status data and corresponding historical air clearance data of the target wind turbine are acquired. These historical air clearance data can be determined based on simulation results of the air clearance values ​​of the entire power plant unit. Then, air clearance data with values ​​less than or equal to a preset dangerous air clearance threshold are identified as dangerous air clearance data, while air clearance data with values ​​greater than the preset dangerous air clearance threshold are identified as safe air clearance data. The set of dangerous air clearance data is called the dangerous air clearance dataset, and the set of safe air clearance data is called the safe air clearance dataset.

[0054] Step S12: Perform correlation analysis on each of the hazardous clearance data and corresponding operational status data in the hazardous clearance dataset and each of the safe clearance data and corresponding operational status data in the safe clearance dataset to determine the first training dataset and the second training dataset.

[0055] In this embodiment, before performing correlation analysis on each of the dangerous airspace data and corresponding operating status data in the dangerous airspace dataset and each of the safe airspace data and corresponding operating status data in the safe airspace dataset to determine the first training dataset and the second training dataset, the method further includes: determining whether the operating status data corresponding to each of the dangerous airspace data in the dangerous airspace dataset meets a preset threshold judgment condition; if the preset threshold judgment condition is met, the dangerous airspace data is identified as interference data and removed from the dangerous airspace dataset. In practical terms, when the unit power data is less than a preset power threshold, or the generator speed data is less than a preset speed threshold, or the pitch angle value is greater than a preset pitch angle value threshold, from the perspective of unit operation mechanism and simulation, it is considered that dangerous airspace values ​​will basically not appear under this state. Therefore, the airspace data appearing in the dangerous airspace dataset under this condition can be identified as interference data and removed from the dangerous airspace dataset. The determination of whether the operating status data corresponding to each of the hazardous airspace data in the hazardous airspace dataset meets the preset threshold judgment conditions includes: acquiring the unit power data, generator speed data, and pitch angle value corresponding to each hazardous airspace data; determining whether the unit power data is less than a preset power threshold, and / or whether the generator speed data is less than a preset speed threshold, and / or whether the pitch angle value is greater than a preset pitch angle value threshold. When the operating status data corresponding to the hazardous airspace data in the hazardous airspace dataset meets any of the above three threshold judgment conditions, the hazardous airspace data is identified as interference data and the interference data is removed from the hazardous airspace dataset. In addition, based on the above judgment, it can also be determined whether interference data still exists in the hazardous airspace dataset based on the maximum rate of decrease of the airspace value. Figure 2 As shown, from the perspective of unit operation mechanism and simulation, the actual air clearance value is unlikely to decrease too rapidly. Therefore, after careful consideration, it can be determined whether the difference between the current dangerous air clearance value and the dangerous air clearance value of the previous second is greater than the preset maximum rate of decrease in air clearance value. If it is greater, the air clearance data appearing in the dangerous air clearance dataset under this condition is identified as interference data and removed from the dangerous air clearance dataset. Then, the... Figure 2 In, the P 实际 For unit power data, P 设定 The preset power threshold; ω 实际 For generator speed data, ω 设定 The preset speed threshold; θ 实际 θ is the pitch angle value. 设定 This is a preset pitch angle threshold. Where L... 实际 (t) represents the current danger clearance data value used for assessment, L实际 (t-1) represents the hazard clearance data value one second before the current hazard clearance data value being assessed, and t represents the time when the current hazard clearance data value is being assessed. 设定 The maximum reduction rate is set to the preset clearance value. By eliminating interfering data, the accuracy of subsequent model construction can be improved, thereby increasing the accuracy of subsequent wind turbine blade clearance status judgment.

[0056] In this embodiment, correlation analysis is performed on each of the hazardous airspace data and corresponding operational status data in the hazardous airspace dataset, and on each of the safe airspace data and corresponding operational status data in the safe airspace dataset, to determine a first training dataset and a second training dataset. Specifically, a correlation analysis is performed on each of the hazardous airspace data and corresponding operational status data in the hazardous airspace dataset using a preset correlation coefficient determination formula, and then a preset number of highly correlated data are selected to generate the first training dataset. Similarly, a correlation analysis is performed on each of the safe airspace data and corresponding operational status data in the safe airspace dataset using a preset correlation coefficient determination formula, and then a preset number of highly correlated data are selected to generate the second training dataset. The preset correlation coefficient determination formula includes, but is not limited to, quantitative correlation analysis methods such as Spearman correlation coefficient, Pearson correlation coefficient, and Kendall correlation coefficient.

[0057] Step S13: Generate the headroom Bayesian classifier model corresponding to the target wind turbine based on the first training dataset and the second training dataset.

[0058] In this embodiment, a headroom Bayesian classifier model corresponding to the target wind turbine is generated based on the first training dataset and the second training dataset. Specifically, a probability distribution function is fitted using the first training dataset to obtain several probability distribution functions similar to those in the first training dataset. Then, these probability distribution functions are summed using a preset maximum likelihood estimation method to obtain a first target probability distribution function. Similarly, a second target probability distribution function corresponding to the second training dataset is obtained. Finally, the headroom Bayesian classifier model corresponding to the target wind turbine is determined based on the first and second target probability distribution functions. The headroom Bayesian classifier model is as follows:

[0059]

[0060]

[0061] C = 0, 1

[0062] Wherein, P(0) is the probability of safe clearance occurrence, obtained by the proportion of safe clearance data to total data; P(1) is the probability of dangerous clearance occurrence, obtained by the proportion of dangerous clearance data to total data; the total number is the historical clearance data obtained in step S11; x is the current operating parameter; P(1|x) is the probability of dangerous clearance; P(0|x) is the probability of safe clearance. Then, the clearance Bayesian classifier model is written into the PLC controller (Programmable Logic Controller) of the target wind turbine using calculation software such as Matlab (programming and numerical calculation platform).

[0063] Step S14: Determine the blade clearance status of the target wind turbine based on the clearance Bayesian classifier model and the current operating status data of the target wind turbine, and control the operation of the wind turbine blades based on the blade clearance status.

[0064] In this embodiment, determining the blade clearance status of the target wind turbine based on the clearance Bayesian classifier model and the current operating status data of the target wind turbine, and controlling the operation of the wind turbine blades based on the blade clearance status, includes: acquiring the current operating status data of the target wind turbine; determining the current dangerous clearance probability and the current safe clearance probability of the target wind turbine using the clearance Bayesian classifier model and the current operating status data; determining whether the current dangerous clearance probability is greater than a preset dangerous clearance probability threshold; if the current dangerous clearance probability is greater than the preset dangerous clearance probability threshold, then controlling the wind turbine blades to adjust the pitch angle. That is, acquiring the current operating status data of the target wind turbine, inputting the current operating status data into the clearance Bayesian classifier model to determine the current dangerous clearance probability and the current safe clearance probability of the target wind turbine, and when the current dangerous clearance probability is greater than the preset dangerous clearance probability threshold, it indicates that there is a high probability that the blade clearance value of the target wind turbine is too small, which may lead to a blade sweeping event, resulting in blade damage and tower loss. Therefore, it is necessary to control the blade pitch angle of the wind turbine to increase the blade clearance of the target wind turbine.

[0065] As can be seen, in this embodiment, historical operating status data and corresponding historical clearance data of the target wind turbine are acquired, and the historical clearance data are divided into dangerous clearance data and safe clearance data based on a preset dangerous clearance threshold to obtain dangerous clearance datasets and safe clearance datasets; the clearance value of the dangerous clearance data is greater than the preset dangerous clearance threshold, and the clearance value of the safe clearance data is less than or equal to the preset dangerous clearance threshold; correlation analysis is performed on each dangerous clearance data and corresponding operating status data in the dangerous clearance dataset, and on each safe clearance data and corresponding operating status data in the safe clearance dataset, to determine a first training dataset and a second training dataset; a clearance Bayesian classifier model corresponding to the target wind turbine is generated based on the first training dataset and the second training dataset; the blade clearance state of the target wind turbine is determined according to the clearance Bayesian classifier model and the current operating status data of the target wind turbine, and the wind turbine blade operation is controlled based on the blade clearance state. In this way, a clearance Bayesian classifier model for the target wind turbine is generated based on the selected blade clearance data with high correlation and the corresponding operating status data. During the actual operation of the target wind turbine, current operating status data is collected and fed into the headroom Bayesian classifier model to determine the current blade status of the target wind turbine. If the blade status is dangerous, the headroom status of the blades can be adjusted by controlling the target wind turbine to avoid blade sweeping events due to insufficient blade headroom, which could cause blade damage and tower loss.

[0066] The above embodiments have described in detail the method of controlling wind turbine blades based on the current blade clearance state. This embodiment will describe in detail the generation process of the Bayesian classifier model.

[0067] See Figure 3 As shown in the figure, this application discloses a specific wind turbine blade control method, including:

[0068] Step S21: Based on the preset correlation coefficient determination formula, perform correlation analysis on each hazardous airspace data and the corresponding operating status data to obtain the first correlation value, and based on the preset correlation coefficient determination formula, perform correlation analysis on each safe airspace data and the corresponding operating status data to obtain the second correlation value.

[0069] In this embodiment, a correlation analysis is performed on each hazardous airspace data and its corresponding operating status data using a preset correlation coefficient determination formula to obtain a first correlation value for each hazardous airspace data and its corresponding operating status data. For example, the operating status data includes unit power data, generator speed data, and pitch angle value. The similarity between the hazardous airspace data and unit power data, the hazardous airspace data and generator speed data, and the hazardous airspace data and pitch angle value are calculated respectively to obtain the first correlation value. Similarly, a correlation analysis is performed on each safe airspace data and its corresponding operating status data using a preset correlation coefficient determination formula to obtain a second correlation value for each safe airspace data and its corresponding operating status data. The preset correlation coefficient determination formula includes, but is not limited to, quantitative correlation analysis methods such as Spearman correlation coefficient, Pearson correlation coefficient, and Kendall correlation coefficient.

[0070] Step S22: Determine the first training net dataset based on the first correlation value, and determine the second training net dataset based on the second correlation value.

[0071] In this embodiment, the hazardous airspace data in the hazardous airspace dataset are sorted according to the first correlation value, and a predetermined number of operating status data are selected from the corresponding operating status data based on the first correlation value to obtain a first initial training airspace dataset. The first correlation value between the hazardous airspace data and the corresponding operating status data in the first initial training airspace dataset is greater than the first correlation value between other hazardous airspace data and their corresponding operating status data in the hazardous airspace dataset. For example, after calculating the similarity between the hazardous airspace data and generator power data, the hazardous airspace data and generator speed data, and the hazardous airspace data and pitch angle value to obtain the first correlation value, the first similarity value between the hazardous airspace data and the generator speed data and the pitch angle value is relatively high. Therefore, the hazardous airspace data and generator power data, and the hazardous airspace data and generator speed data are selected as the first initial training airspace data. Similarly, based on the second correlation value, a preset number of operating status data are selected from the operating status data corresponding to the safe airspace data to obtain the second initial training airspace dataset; wherein, the second correlation value between the safe airspace data and the corresponding operating status data in the second initial training airspace dataset is greater than the second correlation value between other safe airspace data and the corresponding operating status data in the safe airspace dataset. To improve the accuracy of the model's final prediction, this application may consider using the same operating status data when determining the first and second training airspace datasets. That is, the first and second training airspace datasets are determined based on the first and second initial training airspace datasets. For example, the first initial training airspace dataset contains the dangerous airspace data and unit power data, and the dangerous airspace data and generator speed data; the second initial training airspace dataset contains the safe airspace data and unit power data, and the safe airspace data and pitch angle value. The dangerous clearance data and pitch angle values ​​can then be added to the first initial training clearance dataset to determine the final first training clearance dataset, while the safe clearance data and generator speed data can be added to the second initial training clearance dataset to determine the second training clearance dataset. Subsequent operations can then be performed. This approach can improve the accuracy of the model's final predictions.

[0072] Step S23: Generate a first frequency histogram based on each hazardous clearance data and corresponding operational status data in the first training dataset, and determine a first probability distribution type based on the type of the first frequency histogram; and generate a second frequency histogram based on each safe clearance data and corresponding operational status data in the second training dataset, and determine a second probability distribution type based on the type of the second frequency histogram.

[0073] In this embodiment, based on the hazardous clearance data and corresponding operating status data in the first training dataset, a first frequency histogram corresponding to the hazardous clearance data is plotted. The operating status data includes, but is not limited to, unit power data, generator speed data, and blade pitch angle values. After generating each first frequency histogram, the first frequency histogram is judged to determine one or more frequency distribution histogram types corresponding to the first frequency histogram. The frequency distribution histogram types include, but are not limited to, left-skewed distribution, right-skewed distribution, and symmetrical distribution. Then, several probability distribution types similar to the frequency distribution histogram type are selected to obtain one or more first probability distribution types corresponding to the frequency distribution histogram type. The probability distribution types include, but are not limited to, normal distribution, log-normal distribution, Weibull distribution, chi-square distribution, gamma distribution, exponential distribution, etc. For example, the hazardous clearance data and unit power data, and the hazardous clearance data and generator speed data are selected as the first training clearance data. Then, one or more first probability distribution types are determined for the dangerous clearance data and unit power data, and one or more first probability distribution types are determined for the dangerous clearance data and generator speed data. Similarly, based on each safe clearance data and corresponding operating status data in the first training dataset, each second frequency histogram corresponding to the safe clearance data is plotted. The operating status data includes, but is not limited to, unit power data, generator speed data, and pitch angle values. After generating each second frequency histogram, each second frequency histogram is judged to determine one or more frequency distribution histogram types corresponding to the second frequency histogram. The frequency distribution histogram types include, but are not limited to, left-skewed distribution, right-skewed distribution, and symmetrical distribution. Then, several probability distribution types similar to the frequency distribution histogram type are selected to obtain one or more second probability distribution types corresponding to the frequency distribution histogram type. The probability distribution types include, but are not limited to, normal distribution, log-normal distribution, Weibull distribution, chi-square distribution, gamma distribution, exponential distribution, etc.

[0074] Step S24: Determine the first probability density distribution function using the preset maximum likelihood estimation method and the first probability distribution type, and determine the second probability density distribution function using the preset maximum likelihood estimation method and the second probability distribution type.

[0075] In this embodiment, a first probability density distribution function is determined using a preset maximum likelihood estimation method and the first probability distribution type. Specifically, based on the data points corresponding to the hazardous clearance data and the corresponding operational status data under each first probability distribution type, a preset maximum likelihood estimation (MLE) method is used to fit and calculate the probability density function. The accuracy of the probability distribution function is quantified using the root mean square error (RMSE), and the probability density function with the smallest MSE is selected to characterize the probability distribution of the variables to obtain the first probability density distribution function. The first probability density distribution function is:

[0076]

[0077] Wherein, P(1) is the probability of dangerous airspace occurrence, which is obtained by the proportion of dangerous airspace data to the total data; the total number is the historical airspace data obtained in step S11; P(1|x) is the probability of dangerous airspace occurrence.

[0078]

[0079] Where P(0) is the probability of safe airspace occurrence, obtained by the proportion of safe airspace data to the total data; the total number is the historical airspace data obtained in step S11; x is the current operating parameter; P(0|x) is the safe airspace probability. It should be noted that the root mean square error calculation method divides the actual data point variable range into n probability calculation points, statistically obtains the actual cumulative probability at each probability calculation point, and calculates the actual cumulative probability at each probability calculation point based on the fitted probability distribution.

[0080]

[0081] In the formula, Re represents the root mean square error of the fitted distribution, and x i For the i-th probability calculation point, y(x) i The x-th data point is calculated from the actual data points. i The actual cumulative probability at that location. The x-th value calculated for the fitted distribution i The cumulative probability at a given location.

[0082] For example, after determining one or more first probability distribution types of the dangerous airspace data and unit power data, and one or more first probability distribution types of the dangerous airspace data and generator speed data, the first probability density distribution function is obtained by fitting one or more first probability distribution types of the dangerous airspace data and unit power data and one or more first probability distribution types of the dangerous airspace data and generator speed data.

[0083] Step S25: Generate the headroom Bayesian classifier model corresponding to the target wind turbine based on the first probability density distribution function and the second probability density distribution function.

[0084] In this embodiment, since the first probability density distribution function and the second probability density distribution function are calculated separately for safe airspace data and dangerous airspace data, the first probability density distribution function and the second probability density distribution function can be combined to generate the final airspace Bayesian classifier model:

[0085] c = 0, 1;

[0086] Where d is the number of empty data in the first training empty dataset or the second training empty dataset; x is the current running parameter.

[0087] It is evident that by selecting multiple probability distribution types and effectively choosing a suitable probability distribution function, the accuracy of the Bayesian classifier can be improved. By constructing an airspace early warning model, this method can effectively reduce the cost of installing hardware sensors and avoid false alarms caused by environmental interference with the hardware sensors.

[0088] refer to Figure 4 The present application also discloses a wind turbine blade control device, comprising:

[0089] The airspace data classification module 11 is used to acquire the historical operating status data and corresponding historical airspace data of the target wind turbine, and to divide the historical airspace data into dangerous airspace data and safe airspace data based on a preset dangerous airspace threshold to obtain dangerous airspace dataset and safe airspace dataset; the airspace value of the dangerous airspace data is less than or equal to the preset dangerous airspace threshold, and the airspace value of the safe airspace data is greater than the preset dangerous airspace threshold.

[0090] The training dataset generation module 12 is used to perform correlation analysis on each of the dangerous airspace data and corresponding operating status data in the dangerous airspace dataset and each of the safe airspace data and corresponding operating status data in the safe airspace dataset, so as to determine the first training dataset and the second training dataset.

[0091] Model generation module 13 is used to generate a headroom Bayesian classifier model corresponding to the target wind turbine based on the first training dataset and the second training dataset;

[0092] The blade control module 14 is used to determine the blade clearance status of the target wind turbine based on the clearance Bayesian classifier model and the current operating status data of the target wind turbine, and to control the operation of the wind turbine blades based on the blade clearance status.

[0093] As can be seen, in this embodiment, historical operating status data and corresponding historical clearance data of the target wind turbine are acquired, and the historical clearance data are divided into dangerous clearance data and safe clearance data based on a preset dangerous clearance threshold to obtain dangerous clearance datasets and safe clearance datasets; the clearance value of the dangerous clearance data is less than or equal to the preset dangerous clearance threshold, and the clearance value of the safe clearance data is greater than the preset dangerous clearance threshold; correlation analysis is performed on each dangerous clearance data and corresponding operating status data in the dangerous clearance dataset, and on each safe clearance data and corresponding operating status data in the safe clearance dataset, to determine a first training dataset and a second training dataset; a clearance Bayesian classifier model corresponding to the target wind turbine is generated based on the first training dataset and the second training dataset; the blade clearance state of the target wind turbine is determined according to the clearance Bayesian classifier model and the current operating status data of the target wind turbine, and the wind turbine blade operation is controlled based on the blade clearance state. In this way, a clearance Bayesian classifier model for the target wind turbine is generated based on the selected blade clearance data with high correlation and the corresponding operating status data. During the actual operation of the target wind turbine, current operating status data is collected and fed into the headroom Bayesian classifier model to determine the current blade status of the target wind turbine. If the blade status is dangerous, the headroom status of the blades can be adjusted by controlling the target wind turbine to avoid blade sweeping events due to insufficient blade headroom, which could cause blade damage and tower loss.

[0094] In some specific embodiments, the wind turbine blade control device may further include:

[0095] The threshold judgment module is used to determine whether the operating status data corresponding to each of the dangerous airspace data in the dangerous airspace dataset meets the preset threshold judgment conditions.

[0096] The data removal module is used to identify the dangerous airspace data as interference data and remove the interference data from the dangerous airspace dataset if the preset threshold judgment condition is met.

[0097] In some specific embodiments, the threshold determination module may specifically include:

[0098] The data acquisition unit is used to acquire the unit power data, generator speed data and blade pitch angle value corresponding to each of the aforementioned hazardous airspace data;

[0099] The operating status judgment unit is used to determine whether the unit power data is less than a preset power threshold, and / or whether the generator speed data is less than a preset speed threshold, and / or whether the pitch angle value is greater than a preset pitch angle value threshold.

[0100] In some specific embodiments, the training dataset generation module 12 may specifically include:

[0101] The first correlation analysis unit is used to perform correlation analysis on each of the dangerous airspace data and the corresponding operational status data based on a preset correlation coefficient determination formula to obtain a first correlation value.

[0102] The second correlation analysis unit is used to perform correlation analysis on each of the safety clearance data and the corresponding operating status data based on a preset correlation coefficient determination formula to obtain a second correlation value.

[0103] The first dataset determination unit is used to determine the first initial training clearance dataset based on the first correlation value; the first correlation value between the dangerous clearance data and the corresponding operating status data in the first training clearance dataset is greater than the first correlation value between the dangerous clearance data and the corresponding other operating status data in the dangerous clearance dataset.

[0104] The second dataset determination unit is used to determine the second initial training net dataset based on the second correlation value; the second correlation value between the safety net data and the corresponding operating status data in the second training net dataset is greater than the second correlation value between the safety net data and the corresponding other operating status data in the safety net dataset.

[0105] The dataset determination unit is used to determine the first training net dataset and the second training net dataset based on the first initial training net dataset and the second initial training net dataset.

[0106] In some specific embodiments, the model generation module 13 may specifically include:

[0107] The distribution function determination submodule is used to determine the first probability density distribution function corresponding to the first training dataset and the second probability density distribution function corresponding to the second training dataset.

[0108] The model determination unit is used to generate a headroom Bayesian classifier model corresponding to the target wind turbine based on the first probability density distribution function and the second probability density distribution function.

[0109] In some specific embodiments, the distribution function determination submodule may specifically include:

[0110] The first probability distribution type determination unit is used to generate a first frequency histogram based on each dangerous clearance data and corresponding operational status data in the first training dataset, and to determine the first probability distribution type based on the type of the first frequency histogram.

[0111] The second probability distribution type determination unit is used to generate a second frequency histogram based on each safety clearance data and corresponding operating status data in the second training dataset, and to determine the second probability distribution type based on the type of the second frequency histogram.

[0112] The function determination unit is used to determine a first probability density distribution function using a preset maximum likelihood estimation method and the first probability distribution type, and to determine a second probability density distribution function using the preset maximum likelihood estimation method and the second probability distribution type.

[0113] In some specific embodiments, the blade control module 14 may specifically include:

[0114] The operation status data acquisition unit is used to obtain the current operation status data of the target wind turbine, and to determine the current dangerous clearance probability and the current safe clearance probability of the target wind turbine through the clearance Bayesian classifier model and the current operation status data.

[0115] The probability judgment unit is used to determine whether the current dangerous clearance probability is greater than a preset dangerous clearance probability threshold.

[0116] The pitch angle adjustment unit is used to control the wind turbine blades to adjust the pitch angle if the current dangerous clearance probability is greater than a preset dangerous clearance probability threshold.

[0117] Furthermore, embodiments of this application also disclose an electronic device, Figure 5 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.

[0118] Figure 5 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 blade control method disclosed in any of the foregoing embodiments. Alternatively, the electronic device 20 in this embodiment may specifically be an electronic computer.

[0119] 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.

[0120] 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.

[0121] 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 blade control method executed by the electronic device 20 as disclosed in any of the foregoing embodiments, the computer program 222 may further include a computer program capable of performing other specific tasks.

[0122] 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 blade 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.

[0123] 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.

[0124] 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.

[0125] 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.

[0126] 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.

[0127] 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 method for controlling wind turbine blades, characterized in that, include: Acquire historical operating status data and corresponding historical air clearance data of the target wind turbine, and divide the historical air clearance data into dangerous air clearance data and safe air clearance data based on a preset dangerous air clearance threshold to obtain dangerous air clearance dataset and safe air clearance dataset. The airspace value of the hazardous airspace data is less than or equal to the preset hazardous airspace threshold, and the airspace value of the safe airspace data is greater than the preset hazardous airspace threshold. Determine whether the operating status data corresponding to each of the hazardous airspace data in the hazardous airspace dataset meets the preset threshold judgment condition; If the preset threshold judgment condition is met, the dangerous airspace data is identified as interference data and the interference data is removed from the dangerous airspace dataset; Correlation analysis is performed on each of the hazardous airspace data and corresponding operational status data in the hazardous airspace dataset and each of the safe airspace data and corresponding operational status data in the safe airspace dataset to determine the first training dataset and the second training dataset. The process of conducting correlation analysis includes: Based on the preset correlation coefficient determination formula, a correlation analysis is performed on each of the hazardous airspace data and the corresponding operational status data to obtain the first correlation value; Based on the preset correlation coefficient determination formula, a correlation analysis is performed on each of the safety clearance data and the corresponding operating status data to obtain a second correlation value; A first initial training clearance dataset is determined based on the first correlation value; the first correlation value between the hazardous clearance data and the corresponding operational status data in the first initial training clearance dataset is greater than the first correlation value between the hazardous clearance data and the corresponding other operational status data in the hazardous clearance dataset. The second initial training net dataset is determined based on the second correlation value; the second correlation value between the safety net data and the corresponding operating status data in the second initial training net dataset is greater than the second correlation value between the safety net data and the corresponding other operating status data in the safety net dataset. The first training net dataset and the second training net dataset are determined based on the first initial training net dataset and the second initial training net dataset. Generate a Bayesian classifier model for the target wind turbine based on the first training dataset and the second training dataset; The current operating status data of the target wind turbine is obtained, and the current dangerous airspace probability and current safe airspace probability of the target wind turbine are determined by the airspace Bayesian classifier model and the current operating status data. Determine whether the current dangerous airspace probability is greater than a preset dangerous airspace probability threshold; If the current dangerous clearance probability is greater than the preset dangerous clearance probability threshold, then the wind turbine blades are controlled to adjust the pitch angle.

2. The wind turbine blade control method according to claim 1, characterized in that, The step of determining whether the operational status data corresponding to each hazardous airspace data in the hazardous airspace dataset meets the preset threshold judgment condition includes: Obtain the unit power data, generator speed data, and blade pitch angle value corresponding to each of the aforementioned hazardous airspace data; Determine whether the unit power data is less than a preset power threshold, and / or whether the generator speed data is less than a preset speed threshold, and / or whether the pitch angle value is greater than a preset pitch angle value threshold.

3. The wind turbine blade control method according to claim 1, characterized in that, The step of generating the headroom Bayesian classifier model corresponding to the target wind turbine based on the first training dataset and the second training dataset includes: Determine the first probability density distribution function corresponding to the first training dataset and the second probability density distribution function corresponding to the second training dataset; A headroom Bayesian classifier model corresponding to the target wind turbine is generated based on the first probability density distribution function and the second probability density distribution function.

4. The wind turbine blade control method according to claim 3, characterized in that, Determining the first probability density distribution function corresponding to the first training dataset and the second probability density distribution function corresponding to the second training dataset includes: A first frequency histogram is generated based on each hazardous clearance data and corresponding operational status data in the first training dataset, and a first probability distribution type is determined based on the type of the first frequency histogram. A second frequency histogram is generated based on each safety clearance data and corresponding operational status data in the second training dataset, and a second probability distribution type is determined based on the type of the second frequency histogram. A first probability density distribution function is determined using a preset maximum likelihood estimation method and the first probability distribution type, and a second probability density distribution function is determined using the preset maximum likelihood estimation method and the second probability distribution type.

5. A wind turbine blade control device, characterized in that, The wind turbine blade control method according to any one of claims 1 to 4 includes: The airspace data classification module is used to acquire historical operating status data and corresponding historical airspace data of the target wind turbine, and classify the historical airspace data into dangerous airspace data and safe airspace data based on a preset dangerous airspace threshold to obtain dangerous airspace dataset and safe airspace dataset; the airspace value of the dangerous airspace data is less than or equal to the preset dangerous airspace threshold, and the airspace value of the safe airspace data is greater than the preset dangerous airspace threshold. The device further includes: Determine whether the operating status data corresponding to each of the hazardous airspace data in the hazardous airspace dataset meets the preset threshold judgment condition; If the preset threshold judgment condition is met, the dangerous airspace data is identified as interference data and the interference data is removed from the dangerous airspace dataset; The training dataset generation module is used to perform correlation analysis on each of the dangerous airspace data and corresponding operating status data in the dangerous airspace dataset and each of the safe airspace data and corresponding operating status data in the safe airspace dataset, so as to determine the first training dataset and the second training dataset. The process of conducting correlation analysis includes: Based on the preset correlation coefficient determination formula, a correlation analysis is performed on each of the hazardous airspace data and the corresponding operational status data to obtain the first correlation value; Based on the preset correlation coefficient determination formula, a correlation analysis is performed on each of the safety clearance data and the corresponding operating status data to obtain a second correlation value; A first initial training clearance dataset is determined based on the first correlation value; the first correlation value between the hazardous clearance data and the corresponding operational status data in the first initial training clearance dataset is greater than the first correlation value between the hazardous clearance data and the corresponding other operational status data in the hazardous clearance dataset. The second initial training net dataset is determined based on the second correlation value; the second correlation value between the safety net data and the corresponding operating status data in the second initial training net dataset is greater than the second correlation value between the safety net data and the corresponding other operating status data in the safety net dataset. The first training net dataset and the second training net dataset are determined based on the first initial training net dataset and the second initial training net dataset. The model generation module is used to generate a headroom Bayesian classifier model for the target wind turbine based on the first training dataset and the second training dataset. The blade control module is used to acquire the current operating status data of the target wind turbine, determine the current dangerous clearance probability and the current safe clearance probability of the target wind turbine through the clearance Bayesian classifier model and the current operating status data; determine whether the current dangerous clearance probability is greater than a preset dangerous clearance probability threshold; if the current dangerous clearance probability is greater than the preset dangerous clearance probability threshold, control the wind turbine blades to adjust the pitch angle.

6. An electronic device, characterized in that, include: Memory, used to store computer programs; A processor for executing the computer program to implement the wind turbine blade control method as described in any one of claims 1 to 4.

7. A computer-readable storage medium, characterized in that, Used to store a computer program, which, when executed by a processor, implements the wind turbine blade control method as described in any one of claims 1 to 4.

Images